From 7b76bfc86757ca5e7b0a62f50003752da77c115b Mon Sep 17 00:00:00 2001 From: Dick Streefland Date: Tue, 6 Oct 2009 22:35:40 +0200 Subject: kconfig: simplification of scripts/extract-ikconfig I've rewritten the extract-ikconfig script to extract the kernel configuration from a kernel compiled with CONFIG_IKCONFIG. The main motivation for the rewrite was to remove the dependency on the external C program binoffset.c, which is compiled on the initial run. The binoffset executable is invoked with a relative path, which means that the old script can only be run from the top of the kernel tree, and only when you have write permission in the scripts directory. The new script uses tr/grep/tail/zcat only, and can be invoked from anywhere. The binoffset.c program has been removed. This script requires GNU grep 2.5 (released 2002-03-13) or higher, because the -o option was introduced in that version. Signed-off-by: Dick Streefland LKML-Reference: <20091006203540.GA14634@streefland.net> Tested-by: Steven Rostedt Cc: Sam Ravnborg Signed-off-by: Steven Rostedt diff --git a/Documentation/dontdiff b/Documentation/dontdiff index e1efc40..bc01a5a 100644 --- a/Documentation/dontdiff +++ b/Documentation/dontdiff @@ -68,7 +68,6 @@ autoconf.h* bbootsect bin2c binkernel.spec -binoffset bootsect bounds.h bsetup diff --git a/scripts/.gitignore b/scripts/.gitignore index 52cab46a..c5d5db5 100644 --- a/scripts/.gitignore +++ b/scripts/.gitignore @@ -6,5 +6,4 @@ kallsyms pnmtologo bin2c unifdef -binoffset ihex2fw diff --git a/scripts/binoffset.c b/scripts/binoffset.c deleted file mode 100644 index 1a2e39b..0000000 --- a/scripts/binoffset.c +++ /dev/null @@ -1,163 +0,0 @@ -/*************************************************************************** - * binoffset.c - * (C) 2002 Randy Dunlap - -# This program is free software; you can redistribute it and/or modify -# it under the terms of the GNU General Public License as published by -# the Free Software Foundation; either version 2 of the License, or -# (at your option) any later version. -# -# This program is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the -# GNU General Public License for more details. -# -# You should have received a copy of the GNU General Public License -# along with this program; if not, write to the Free Software -# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - -# binoffset.c: -# - searches a (binary) file for a specified (binary) pattern -# - returns the offset of the located pattern or ~0 if not found -# - exits with exit status 0 normally or non-0 if pattern is not found -# or any other error occurs. - -****************************************************************/ - -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#define VERSION "0.1" -#define BUF_SIZE (16 * 1024) -#define PAT_SIZE 100 - -char *progname; -char *inputname; -int inputfd; -unsigned int bix; /* buf index */ -unsigned char patterns [PAT_SIZE] = {0}; /* byte-sized pattern array */ -int pat_len; /* actual number of pattern bytes */ -unsigned char *madr; /* mmap address */ -size_t filesize; -int num_matches = 0; -off_t firstloc = 0; - -void usage (void) -{ - fprintf (stderr, "%s ver. %s\n", progname, VERSION); - fprintf (stderr, "usage: %s filename pattern_bytes\n", - progname); - fprintf (stderr, " [prints location of pattern_bytes in file]\n"); - exit (1); -} - -void get_pattern (int pat_count, char *pats []) -{ - int ix, err, tmp; - -#ifdef DEBUG - fprintf (stderr,"get_pattern: count = %d\n", pat_count); - for (ix = 0; ix < pat_count; ix++) - fprintf (stderr, " pat # %d: [%s]\n", ix, pats[ix]); -#endif - - for (ix = 0; ix < pat_count; ix++) { - tmp = 0; - err = sscanf (pats[ix], "%5i", &tmp); - if (err != 1 || tmp > 0xff) { - fprintf (stderr, "pattern or value error in pattern # %d [%s]\n", - ix, pats[ix]); - usage (); - } - patterns [ix] = tmp; - } - pat_len = pat_count; -} - -void search_pattern (void) -{ - for (bix = 0; bix < filesize; bix++) { - if (madr[bix] == patterns[0]) { - if (memcmp (&madr[bix], patterns, pat_len) == 0) { - if (num_matches == 0) - firstloc = bix; - num_matches++; - } - } - } -} - -#ifdef NOTDEF -size_t get_filesize (int fd) -{ - off_t end_off = lseek (fd, 0, SEEK_END); - lseek (fd, 0, SEEK_SET); - return (size_t) end_off; -} -#endif - -size_t get_filesize (int fd) -{ - int err; - struct stat stat; - - err = fstat (fd, &stat); - fprintf (stderr, "filesize: %ld\n", err < 0 ? (long)err : stat.st_size); - if (err < 0) - return err; - return (size_t) stat.st_size; -} - -int main (int argc, char *argv []) -{ - progname = argv[0]; - - if (argc < 3) - usage (); - - get_pattern (argc - 2, argv + 2); - - inputname = argv[1]; - - inputfd = open (inputname, O_RDONLY); - if (inputfd == -1) { - fprintf (stderr, "%s: cannot open '%s'\n", - progname, inputname); - exit (3); - } - - filesize = get_filesize (inputfd); - - madr = mmap (0, filesize, PROT_READ, MAP_PRIVATE, inputfd, 0); - if (madr == MAP_FAILED) { - fprintf (stderr, "mmap error = %d\n", errno); - close (inputfd); - exit (4); - } - - search_pattern (); - - if (munmap (madr, filesize)) - fprintf (stderr, "munmap error = %d\n", errno); - - if (close (inputfd)) - fprintf (stderr, "%s: error %d closing '%s'\n", - progname, errno, inputname); - - fprintf (stderr, "number of pattern matches = %d\n", num_matches); - if (num_matches == 0) - firstloc = ~0; - printf ("%ld\n", firstloc); - fprintf (stderr, "%ld\n", firstloc); - - exit (num_matches ? 0 : 2); -} - -/* end binoffset.c */ diff --git a/scripts/extract-ikconfig b/scripts/extract-ikconfig index de233ff..37f30d3 100755 --- a/scripts/extract-ikconfig +++ b/scripts/extract-ikconfig @@ -1,92 +1,53 @@ #!/bin/sh -# extracts .config info from a [b]zImage file -# uses: binoffset (new), dd, zcat, strings, grep -# $arg1 is [b]zImage filename - -binoffset="./scripts/binoffset" -test -e $binoffset || cc -o $binoffset ./scripts/binoffset.c || exit 1 - -IKCFG_ST="0x49 0x4b 0x43 0x46 0x47 0x5f 0x53 0x54" -IKCFG_ED="0x49 0x4b 0x43 0x46 0x47 0x5f 0x45 0x44" -dump_config() { - file="$1" - - start=`$binoffset $file $IKCFG_ST 2>/dev/null` - [ "$?" != "0" ] && start="-1" - if [ "$start" -eq "-1" ]; then - return - fi - end=`$binoffset $file $IKCFG_ED 2>/dev/null` - [ "$?" != "0" ] && end="-1" - if [ "$end" -eq "-1" ]; then - return - fi - - start=`expr $start + 8` - size=`expr $end - $start` - - dd if="$file" ibs=1 skip="$start" count="$size" 2>/dev/null | zcat - - clean_up - exit 0 -} - - -usage() -{ - echo " usage: extract-ikconfig [b]zImage_filename" -} - -clean_up() +# ---------------------------------------------------------------------- +# extract-ikconfig - Extract the .config file from a kernel image +# +# This will only work when the kernel was compiled with CONFIG_IKCONFIG. +# +# The obscure use of the "tr" filter is to work around older versions of +# "grep" that report the byte offset of the line instead of the pattern. +# +# (c) 2009, Dick Streefland +# Licensed under the terms of the GNU General Public License. +# ---------------------------------------------------------------------- + +gz1='\037\213\010' +gz2='01' +cf1='IKCFG_ST\037\213\010' +cf2='0123456789' + +dump_config() { - if [ "$TMPFILE" != "" ]; then - rm -f $TMPFILE + if pos=`tr "$cf1\n$cf2" "\n$cf2=" < "$1" | grep -abo "^$cf2"` + then + pos=${pos%%:*} + tail -c+$(($pos+8)) "$1" | zcat -q + exit 0 fi } -if [ $# -lt 1 ] +# Check invocation: +me=${0##*/} +img=$1 +if [ $# -ne 1 -o ! -s "$img" ] then - usage - exit 1 + echo "Usage: $me " >&2 + exit 2 fi -TMPFILE=`mktemp -t ikconfig-XXXXXX` || exit 1 -image="$1" - -# vmlinux: Attempt to dump the configuration from the file directly -dump_config "$image" - -GZHDR1="0x1f 0x8b 0x08 0x00" -GZHDR2="0x1f 0x8b 0x08 0x08" - -ELFHDR="0x7f 0x45 0x4c 0x46" - -# vmlinux.gz: Check for a compressed images -off=`$binoffset "$image" $GZHDR1 2>/dev/null` -[ "$?" != "0" ] && off="-1" -if [ "$off" -eq "-1" ]; then - off=`$binoffset "$image" $GZHDR2 2>/dev/null` - [ "$?" != "0" ] && off="-1" -fi -if [ "$off" -eq "0" ]; then - zcat <"$image" >"$TMPFILE" - dump_config "$TMPFILE" -elif [ "$off" -ne "-1" ]; then - (dd ibs="$off" skip=1 count=0 && dd bs=512k) <"$image" 2>/dev/null | \ - zcat >"$TMPFILE" - dump_config "$TMPFILE" - -# check if this is simply an ELF file -else - off=`$binoffset "$image" $ELFHDR 2>/dev/null` - [ "$?" != "0" ] && off="-1" - if [ "$off" -eq "0" ]; then - dump_config "$image" - fi -fi - -echo "ERROR: Unable to extract kernel configuration information." -echo " This kernel image may not have the config info." - -clean_up +# Initial attempt for uncompressed images or objects: +dump_config "$img" + +# That didn't work, so decompress and try again: +tmp=/tmp/ikconfig$$ +trap "rm -f $tmp" 0 +for pos in `tr "$gz1\n$gz2" "\n$gz2=" < "$img" | grep -abo "^$gz2"` +do + pos=${pos%%:*} + tail -c+$pos "$img" | zcat 2> /dev/null > $tmp + dump_config $tmp +done + +# Bail out: +echo "$me: Cannot find kernel config." >&2 exit 1 -- cgit v0.10.2 From e169cfbef46d62e042614ffafa8880eed1d894bb Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 23 Nov 2009 14:53:09 -0700 Subject: of/flattree: merge find_flat_dt_string and initial_boot_params Merge common code between Microblaze and PowerPC. Signed-off-by: Grant Likely Reviewed-by: Wolfram Sang Tested-by: Michal Simek diff --git a/arch/microblaze/Kconfig b/arch/microblaze/Kconfig index bbd8327..f39c927 100644 --- a/arch/microblaze/Kconfig +++ b/arch/microblaze/Kconfig @@ -111,6 +111,7 @@ config CMDLINE_FORCE config OF def_bool y + select OF_FLATTREE config PROC_DEVICETREE bool "Support for device tree in /proc" diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index b817df1..06d620a 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -47,17 +47,9 @@ static int __initdata dt_root_size_cells; typedef u32 cell_t; -static struct boot_param_header *initial_boot_params; - /* export that to outside world */ struct device_node *of_chosen; -static inline char *find_flat_dt_string(u32 offset) -{ - return ((char *)initial_boot_params) + - initial_boot_params->off_dt_strings + offset; -} - /** * This function is used to scan the flattened device-tree, it is * used to extract the memory informations at boot before we can diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 2ba14e7..2a75c6ae 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -163,6 +163,7 @@ config PPC_OF config OF def_bool y + select OF_FLATTREE config PPC_UDBG_16550 bool diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 4ec3008..fccf7e4 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -73,12 +73,6 @@ unsigned long tce_alloc_start, tce_alloc_end; typedef u32 cell_t; -#if 0 -static struct boot_param_header *initial_boot_params __initdata; -#else -struct boot_param_header *initial_boot_params; -#endif - extern struct device_node *allnodes; /* temporary while merging */ extern rwlock_t devtree_lock; /* temporary while merging */ @@ -86,12 +80,6 @@ extern rwlock_t devtree_lock; /* temporary while merging */ /* export that to outside world */ struct device_node *of_chosen; -static inline char *find_flat_dt_string(u32 offset) -{ - return ((char *)initial_boot_params) + - initial_boot_params->off_dt_strings + offset; -} - /** * This function is used to scan the flattened device-tree, it is * used to extract the memory informations at boot before we can diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig index d2fa27c..462825e 100644 --- a/drivers/of/Kconfig +++ b/drivers/of/Kconfig @@ -1,3 +1,7 @@ +config OF_FLATTREE + bool + depends on OF + config OF_DEVICE def_bool y depends on OF && (SPARC || PPC_OF || MICROBLAZE) diff --git a/drivers/of/Makefile b/drivers/of/Makefile index bdfb5f5..f232cc9 100644 --- a/drivers/of/Makefile +++ b/drivers/of/Makefile @@ -1,4 +1,5 @@ obj-y = base.o +obj-$(CONFIG_OF_FLATTREE) += fdt.o obj-$(CONFIG_OF_DEVICE) += device.o platform.o obj-$(CONFIG_OF_GPIO) += gpio.o obj-$(CONFIG_OF_I2C) += of_i2c.o diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c new file mode 100644 index 0000000..9faa9a5 --- /dev/null +++ b/drivers/of/fdt.c @@ -0,0 +1,21 @@ +/* + * Functions for working with the Flattened Device Tree data format + * + * Copyright 2009 Benjamin Herrenschmidt, IBM Corp + * benh@kernel.crashing.org + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * version 2 as published by the Free Software Foundation. + */ + +#include +#include + +struct boot_param_header *initial_boot_params; + +char *find_flat_dt_string(u32 offset) +{ + return ((char *)initial_boot_params) + + initial_boot_params->off_dt_strings + offset; +} diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index 41d432b..d1a79f3 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -57,7 +57,11 @@ struct boot_param_header { u32 dt_struct_size; /* size of the DT structure block */ }; +/* TBD: Temporary export of fdt globals - remove when code fully merged */ +extern struct boot_param_header *initial_boot_params; + /* For scanning the flat device-tree at boot time */ +extern char *find_flat_dt_string(u32 offset); extern int __init of_scan_flat_dt(int (*it)(unsigned long node, const char *uname, int depth, void *data), -- cgit v0.10.2 From c8cb7a59842c0b512b44f6f818cdb0b5a3ddc89e Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 23 Nov 2009 18:54:23 -0700 Subject: of/flattree: merge of_scan_flat_dt Merge common code between PowerPC and Microblaze Signed-off-by: Grant Likely Reviewed-by: Wolfram Sang Tested-by: Michal Simek diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 06d620a..0db8ee6 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -50,67 +50,6 @@ typedef u32 cell_t; /* export that to outside world */ struct device_node *of_chosen; -/** - * This function is used to scan the flattened device-tree, it is - * used to extract the memory informations at boot before we can - * unflatten the tree - */ -int __init of_scan_flat_dt(int (*it)(unsigned long node, - const char *uname, int depth, - void *data), - void *data) -{ - unsigned long p = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; - int rc = 0; - int depth = -1; - - do { - u32 tag = *((u32 *)p); - char *pathp; - - p += 4; - if (tag == OF_DT_END_NODE) { - depth--; - continue; - } - if (tag == OF_DT_NOP) - continue; - if (tag == OF_DT_END) - break; - if (tag == OF_DT_PROP) { - u32 sz = *((u32 *)p); - p += 8; - if (initial_boot_params->version < 0x10) - p = _ALIGN(p, sz >= 8 ? 8 : 4); - p += sz; - p = _ALIGN(p, 4); - continue; - } - if (tag != OF_DT_BEGIN_NODE) { - printk(KERN_WARNING "Invalid tag %x scanning flattened" - " device tree !\n", tag); - return -EINVAL; - } - depth++; - pathp = (char *)p; - p = _ALIGN(p + strlen(pathp) + 1, 4); - if ((*pathp) == '/') { - char *lp, *np; - for (lp = NULL, np = pathp; *np; np++) - if ((*np) == '/') - lp = np+1; - if (lp != NULL) - pathp = lp; - } - rc = it(p, pathp, depth, data); - if (rc != 0) - break; - } while (1); - - return rc; -} - unsigned long __init of_get_flat_dt_root(void) { unsigned long p = ((unsigned long)initial_boot_params) + diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index fccf7e4..14a07b9 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -80,67 +80,6 @@ extern rwlock_t devtree_lock; /* temporary while merging */ /* export that to outside world */ struct device_node *of_chosen; -/** - * This function is used to scan the flattened device-tree, it is - * used to extract the memory informations at boot before we can - * unflatten the tree - */ -int __init of_scan_flat_dt(int (*it)(unsigned long node, - const char *uname, int depth, - void *data), - void *data) -{ - unsigned long p = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; - int rc = 0; - int depth = -1; - - do { - u32 tag = *((u32 *)p); - char *pathp; - - p += 4; - if (tag == OF_DT_END_NODE) { - depth --; - continue; - } - if (tag == OF_DT_NOP) - continue; - if (tag == OF_DT_END) - break; - if (tag == OF_DT_PROP) { - u32 sz = *((u32 *)p); - p += 8; - if (initial_boot_params->version < 0x10) - p = _ALIGN(p, sz >= 8 ? 8 : 4); - p += sz; - p = _ALIGN(p, 4); - continue; - } - if (tag != OF_DT_BEGIN_NODE) { - printk(KERN_WARNING "Invalid tag %x scanning flattened" - " device tree !\n", tag); - return -EINVAL; - } - depth++; - pathp = (char *)p; - p = _ALIGN(p + strlen(pathp) + 1, 4); - if ((*pathp) == '/') { - char *lp, *np; - for (lp = NULL, np = pathp; *np; np++) - if ((*np) == '/') - lp = np+1; - if (lp != NULL) - pathp = lp; - } - rc = it(p, pathp, depth, data); - if (rc != 0) - break; - } while(1); - - return rc; -} - unsigned long __init of_get_flat_dt_root(void) { unsigned long p = ((unsigned long)initial_boot_params) + diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 9faa9a5..dd9057c 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -19,3 +19,67 @@ char *find_flat_dt_string(u32 offset) return ((char *)initial_boot_params) + initial_boot_params->off_dt_strings + offset; } + +/** + * of_scan_flat_dt - scan flattened tree blob and call callback on each. + * @it: callback function + * @data: context data pointer + * + * This function is used to scan the flattened device-tree, it is + * used to extract the memory information at boot before we can + * unflatten the tree + */ +int __init of_scan_flat_dt(int (*it)(unsigned long node, + const char *uname, int depth, + void *data), + void *data) +{ + unsigned long p = ((unsigned long)initial_boot_params) + + initial_boot_params->off_dt_struct; + int rc = 0; + int depth = -1; + + do { + u32 tag = *((u32 *)p); + char *pathp; + + p += 4; + if (tag == OF_DT_END_NODE) { + depth--; + continue; + } + if (tag == OF_DT_NOP) + continue; + if (tag == OF_DT_END) + break; + if (tag == OF_DT_PROP) { + u32 sz = *((u32 *)p); + p += 8; + if (initial_boot_params->version < 0x10) + p = _ALIGN(p, sz >= 8 ? 8 : 4); + p += sz; + p = _ALIGN(p, 4); + continue; + } + if (tag != OF_DT_BEGIN_NODE) { + pr_err("Invalid tag %x in flat device tree!\n", tag); + return -EINVAL; + } + depth++; + pathp = (char *)p; + p = _ALIGN(p + strlen(pathp) + 1, 4); + if ((*pathp) == '/') { + char *lp, *np; + for (lp = NULL, np = pathp; *np; np++) + if ((*np) == '/') + lp = np+1; + if (lp != NULL) + pathp = lp; + } + rc = it(p, pathp, depth, data); + if (rc != 0) + break; + } while (1); + + return rc; +} -- cgit v0.10.2 From 819d2819303654c6829d572e698e2d0021c08599 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 23 Nov 2009 19:44:23 -0700 Subject: of/flattree: merge of_get_flat_dt_root Merge common code between PowerPC and MicroBlaze Signed-off-by: Grant Likely Reviewed-by: Wolfram Sang Tested-by: Michal Simek diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 0db8ee6..7eb6f8b 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -50,18 +50,6 @@ typedef u32 cell_t; /* export that to outside world */ struct device_node *of_chosen; -unsigned long __init of_get_flat_dt_root(void) -{ - unsigned long p = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; - - while (*((u32 *)p) == OF_DT_NOP) - p += 4; - BUG_ON(*((u32 *)p) != OF_DT_BEGIN_NODE); - p += 4; - return _ALIGN(p + strlen((char *)p) + 1, 4); -} - /** * This function can be used within scan_flattened_dt callback to get * access to properties diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 14a07b9..b5d5f85 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -80,18 +80,6 @@ extern rwlock_t devtree_lock; /* temporary while merging */ /* export that to outside world */ struct device_node *of_chosen; -unsigned long __init of_get_flat_dt_root(void) -{ - unsigned long p = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; - - while(*((u32 *)p) == OF_DT_NOP) - p += 4; - BUG_ON (*((u32 *)p) != OF_DT_BEGIN_NODE); - p += 4; - return _ALIGN(p + strlen((char *)p) + 1, 4); -} - /** * This function can be used within scan_flattened_dt callback to get * access to properties diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index dd9057c..f41d739 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -83,3 +83,19 @@ int __init of_scan_flat_dt(int (*it)(unsigned long node, return rc; } + +/** + * of_get_flat_dt_root - find the root node in the flat blob + */ +unsigned long __init of_get_flat_dt_root(void) +{ + unsigned long p = ((unsigned long)initial_boot_params) + + initial_boot_params->off_dt_struct; + + while (*((u32 *)p) == OF_DT_NOP) + p += 4; + BUG_ON(*((u32 *)p) != OF_DT_BEGIN_NODE); + p += 4; + return _ALIGN(p + strlen((char *)p) + 1, 4); +} + -- cgit v0.10.2 From 31a6a87dfc34fbf02aef9a160adf558ec56d3ccd Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 23 Nov 2009 19:49:38 -0700 Subject: of/flattree: remove __init annotations from the header file __init annotation belongs in the .c file, not the header. Signed-off-by: Grant Likely Reviewed-by: Wolfram Sang Tested-by: Michal Simek diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index d1a79f3..81231e0 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -62,15 +62,13 @@ extern struct boot_param_header *initial_boot_params; /* For scanning the flat device-tree at boot time */ extern char *find_flat_dt_string(u32 offset); -extern int __init of_scan_flat_dt(int (*it)(unsigned long node, - const char *uname, int depth, - void *data), - void *data); -extern void __init *of_get_flat_dt_prop(unsigned long node, const char *name, - unsigned long *size); -extern int __init of_flat_dt_is_compatible(unsigned long node, - const char *name); -extern unsigned long __init of_get_flat_dt_root(void); +extern int of_scan_flat_dt(int (*it)(unsigned long node, const char *uname, + int depth, void *data), + void *data); +extern void *of_get_flat_dt_prop(unsigned long node, const char *name, + unsigned long *size); +extern int of_flat_dt_is_compatible(unsigned long node, const char *name); +extern unsigned long of_get_flat_dt_root(void); /* Other Prototypes */ extern void finish_device_tree(void); -- cgit v0.10.2 From ca900cfa2944448bdb76e1246f282e59bc65f472 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 23 Nov 2009 20:06:59 -0700 Subject: of/flattree: merge of_get_flat_dt_prop Merge common code between PowerPC and Microblaze Signed-off-by: Grant Likely Reviewed-by: Wolfram Sang Tested-by: Michal Simek diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 7eb6f8b..d75c625 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -50,48 +50,6 @@ typedef u32 cell_t; /* export that to outside world */ struct device_node *of_chosen; -/** - * This function can be used within scan_flattened_dt callback to get - * access to properties - */ -void *__init of_get_flat_dt_prop(unsigned long node, const char *name, - unsigned long *size) -{ - unsigned long p = node; - - do { - u32 tag = *((u32 *)p); - u32 sz, noff; - const char *nstr; - - p += 4; - if (tag == OF_DT_NOP) - continue; - if (tag != OF_DT_PROP) - return NULL; - - sz = *((u32 *)p); - noff = *((u32 *)(p + 4)); - p += 8; - if (initial_boot_params->version < 0x10) - p = _ALIGN(p, sz >= 8 ? 8 : 4); - - nstr = find_flat_dt_string(noff); - if (nstr == NULL) { - printk(KERN_WARNING "Can't find property index" - " name !\n"); - return NULL; - } - if (strcmp(name, nstr) == 0) { - if (size) - *size = sz; - return (void *)p; - } - p += sz; - p = _ALIGN(p, 4); - } while (1); -} - int __init of_flat_dt_is_compatible(unsigned long node, const char *compat) { const char *cp; diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index b5d5f85..cd0a2bfc 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -80,48 +80,6 @@ extern rwlock_t devtree_lock; /* temporary while merging */ /* export that to outside world */ struct device_node *of_chosen; -/** - * This function can be used within scan_flattened_dt callback to get - * access to properties - */ -void* __init of_get_flat_dt_prop(unsigned long node, const char *name, - unsigned long *size) -{ - unsigned long p = node; - - do { - u32 tag = *((u32 *)p); - u32 sz, noff; - const char *nstr; - - p += 4; - if (tag == OF_DT_NOP) - continue; - if (tag != OF_DT_PROP) - return NULL; - - sz = *((u32 *)p); - noff = *((u32 *)(p + 4)); - p += 8; - if (initial_boot_params->version < 0x10) - p = _ALIGN(p, sz >= 8 ? 8 : 4); - - nstr = find_flat_dt_string(noff); - if (nstr == NULL) { - printk(KERN_WARNING "Can't find property index" - " name !\n"); - return NULL; - } - if (strcmp(name, nstr) == 0) { - if (size) - *size = sz; - return (void *)p; - } - p += sz; - p = _ALIGN(p, 4); - } while(1); -} - int __init of_flat_dt_is_compatible(unsigned long node, const char *compat) { const char* cp; diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index f41d739..b17a908 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -99,3 +99,46 @@ unsigned long __init of_get_flat_dt_root(void) return _ALIGN(p + strlen((char *)p) + 1, 4); } +/** + * of_get_flat_dt_prop - Given a node in the flat blob, return the property ptr + * + * This function can be used within scan_flattened_dt callback to get + * access to properties + */ +void *__init of_get_flat_dt_prop(unsigned long node, const char *name, + unsigned long *size) +{ + unsigned long p = node; + + do { + u32 tag = *((u32 *)p); + u32 sz, noff; + const char *nstr; + + p += 4; + if (tag == OF_DT_NOP) + continue; + if (tag != OF_DT_PROP) + return NULL; + + sz = *((u32 *)p); + noff = *((u32 *)(p + 4)); + p += 8; + if (initial_boot_params->version < 0x10) + p = _ALIGN(p, sz >= 8 ? 8 : 4); + + nstr = find_flat_dt_string(noff); + if (nstr == NULL) { + pr_warning("Can't find property index name !\n"); + return NULL; + } + if (strcmp(name, nstr) == 0) { + if (size) + *size = sz; + return (void *)p; + } + p += sz; + p = _ALIGN(p, 4); + } while (1); +} + -- cgit v0.10.2 From 00e38efd90f27518ec96b37b1c7773e3ac529966 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 23 Nov 2009 20:07:00 -0700 Subject: of/flattree: Merge of_flat_dt_is_compatible Merge common code between PowerPC and Microblaze Signed-off-by: Grant Likely Reviewed-by: Wolfram Sang Tested-by: Michal Simek diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index d75c625..eb27bd3 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -50,25 +50,6 @@ typedef u32 cell_t; /* export that to outside world */ struct device_node *of_chosen; -int __init of_flat_dt_is_compatible(unsigned long node, const char *compat) -{ - const char *cp; - unsigned long cplen, l; - - cp = of_get_flat_dt_prop(node, "compatible", &cplen); - if (cp == NULL) - return 0; - while (cplen > 0) { - if (strncasecmp(cp, compat, strlen(compat)) == 0) - return 1; - l = strlen(cp) + 1; - cp += l; - cplen -= l; - } - - return 0; -} - static void *__init unflatten_dt_alloc(unsigned long *mem, unsigned long size, unsigned long align) { diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index cd0a2bfc..413e608 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -80,25 +80,6 @@ extern rwlock_t devtree_lock; /* temporary while merging */ /* export that to outside world */ struct device_node *of_chosen; -int __init of_flat_dt_is_compatible(unsigned long node, const char *compat) -{ - const char* cp; - unsigned long cplen, l; - - cp = of_get_flat_dt_prop(node, "compatible", &cplen); - if (cp == NULL) - return 0; - while (cplen > 0) { - if (strncasecmp(cp, compat, strlen(compat)) == 0) - return 1; - l = strlen(cp) + 1; - cp += l; - cplen -= l; - } - - return 0; -} - static void *__init unflatten_dt_alloc(unsigned long *mem, unsigned long size, unsigned long align) { diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index b17a908..5cdd958 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -142,3 +142,27 @@ void *__init of_get_flat_dt_prop(unsigned long node, const char *name, } while (1); } +/** + * of_flat_dt_is_compatible - Return true if given node has compat in compatible list + * @node: node to test + * @compat: compatible string to compare with compatible list. + */ +int __init of_flat_dt_is_compatible(unsigned long node, const char *compat) +{ + const char *cp; + unsigned long cplen, l; + + cp = of_get_flat_dt_prop(node, "compatible", &cplen); + if (cp == NULL) + return 0; + while (cplen > 0) { + if (strncasecmp(cp, compat, strlen(compat)) == 0) + return 1; + l = strlen(cp) + 1; + cp += l; + cplen -= l; + } + + return 0; +} + -- cgit v0.10.2 From bbd33931a08362f78266a4016211a35947b91041 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 23 Nov 2009 20:07:00 -0700 Subject: of/flattree: Merge unflatten_dt_node Merge common code between PowerPC and MicroBlaze Signed-off-by: Grant Likely Reviewed-by: Wolfram Sang Tested-by: Michal Simek diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index eb27bd3..021770a 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -50,201 +50,6 @@ typedef u32 cell_t; /* export that to outside world */ struct device_node *of_chosen; -static void *__init unflatten_dt_alloc(unsigned long *mem, unsigned long size, - unsigned long align) -{ - void *res; - - *mem = _ALIGN(*mem, align); - res = (void *)*mem; - *mem += size; - - return res; -} - -static unsigned long __init unflatten_dt_node(unsigned long mem, - unsigned long *p, - struct device_node *dad, - struct device_node ***allnextpp, - unsigned long fpsize) -{ - struct device_node *np; - struct property *pp, **prev_pp = NULL; - char *pathp; - u32 tag; - unsigned int l, allocl; - int has_name = 0; - int new_format = 0; - - tag = *((u32 *)(*p)); - if (tag != OF_DT_BEGIN_NODE) { - printk("Weird tag at start of node: %x\n", tag); - return mem; - } - *p += 4; - pathp = (char *)*p; - l = allocl = strlen(pathp) + 1; - *p = _ALIGN(*p + l, 4); - - /* version 0x10 has a more compact unit name here instead of the full - * path. we accumulate the full path size using "fpsize", we'll rebuild - * it later. We detect this because the first character of the name is - * not '/'. - */ - if ((*pathp) != '/') { - new_format = 1; - if (fpsize == 0) { - /* root node: special case. fpsize accounts for path - * plus terminating zero. root node only has '/', so - * fpsize should be 2, but we want to avoid the first - * level nodes to have two '/' so we use fpsize 1 here - */ - fpsize = 1; - allocl = 2; - } else { - /* account for '/' and path size minus terminal 0 - * already in 'l' - */ - fpsize += l; - allocl = fpsize; - } - } - - np = unflatten_dt_alloc(&mem, sizeof(struct device_node) + allocl, - __alignof__(struct device_node)); - if (allnextpp) { - memset(np, 0, sizeof(*np)); - np->full_name = ((char *)np) + sizeof(struct device_node); - if (new_format) { - char *p2 = np->full_name; - /* rebuild full path for new format */ - if (dad && dad->parent) { - strcpy(p2, dad->full_name); -#ifdef DEBUG - if ((strlen(p2) + l + 1) != allocl) { - pr_debug("%s: p: %d, l: %d, a: %d\n", - pathp, (int)strlen(p2), - l, allocl); - } -#endif - p2 += strlen(p2); - } - *(p2++) = '/'; - memcpy(p2, pathp, l); - } else - memcpy(np->full_name, pathp, l); - prev_pp = &np->properties; - **allnextpp = np; - *allnextpp = &np->allnext; - if (dad != NULL) { - np->parent = dad; - /* we temporarily use the next field as `last_child'*/ - if (dad->next == NULL) - dad->child = np; - else - dad->next->sibling = np; - dad->next = np; - } - kref_init(&np->kref); - } - while (1) { - u32 sz, noff; - char *pname; - - tag = *((u32 *)(*p)); - if (tag == OF_DT_NOP) { - *p += 4; - continue; - } - if (tag != OF_DT_PROP) - break; - *p += 4; - sz = *((u32 *)(*p)); - noff = *((u32 *)((*p) + 4)); - *p += 8; - if (initial_boot_params->version < 0x10) - *p = _ALIGN(*p, sz >= 8 ? 8 : 4); - - pname = find_flat_dt_string(noff); - if (pname == NULL) { - printk(KERN_INFO - "Can't find property name in list !\n"); - break; - } - if (strcmp(pname, "name") == 0) - has_name = 1; - l = strlen(pname) + 1; - pp = unflatten_dt_alloc(&mem, sizeof(struct property), - __alignof__(struct property)); - if (allnextpp) { - if (strcmp(pname, "linux,phandle") == 0) { - np->node = *((u32 *)*p); - if (np->linux_phandle == 0) - np->linux_phandle = np->node; - } - if (strcmp(pname, "ibm,phandle") == 0) - np->linux_phandle = *((u32 *)*p); - pp->name = pname; - pp->length = sz; - pp->value = (void *)*p; - *prev_pp = pp; - prev_pp = &pp->next; - } - *p = _ALIGN((*p) + sz, 4); - } - /* with version 0x10 we may not have the name property, recreate - * it here from the unit name if absent - */ - if (!has_name) { - char *p1 = pathp, *ps = pathp, *pa = NULL; - int sz; - - while (*p1) { - if ((*p1) == '@') - pa = p1; - if ((*p1) == '/') - ps = p1 + 1; - p1++; - } - if (pa < ps) - pa = p1; - sz = (pa - ps) + 1; - pp = unflatten_dt_alloc(&mem, sizeof(struct property) + sz, - __alignof__(struct property)); - if (allnextpp) { - pp->name = "name"; - pp->length = sz; - pp->value = pp + 1; - *prev_pp = pp; - prev_pp = &pp->next; - memcpy(pp->value, ps, sz - 1); - ((char *)pp->value)[sz - 1] = 0; - pr_debug("fixed up name for %s -> %s\n", pathp, - (char *)pp->value); - } - } - if (allnextpp) { - *prev_pp = NULL; - np->name = of_get_property(np, "name", NULL); - np->type = of_get_property(np, "device_type", NULL); - - if (!np->name) - np->name = ""; - if (!np->type) - np->type = ""; - } - while (tag == OF_DT_BEGIN_NODE) { - mem = unflatten_dt_node(mem, p, np, allnextpp, fpsize); - tag = *((u32 *)(*p)); - } - if (tag != OF_DT_END_NODE) { - printk(KERN_INFO "Weird tag at end of node: %x\n", tag); - return mem; - } - *p += 4; - return mem; -} - /** * unflattens the device-tree passed by the firmware, creating the * tree of struct device_node. It also fills the "name" and "type" diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 413e608..a102a0a 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -80,200 +80,6 @@ extern rwlock_t devtree_lock; /* temporary while merging */ /* export that to outside world */ struct device_node *of_chosen; -static void *__init unflatten_dt_alloc(unsigned long *mem, unsigned long size, - unsigned long align) -{ - void *res; - - *mem = _ALIGN(*mem, align); - res = (void *)*mem; - *mem += size; - - return res; -} - -static unsigned long __init unflatten_dt_node(unsigned long mem, - unsigned long *p, - struct device_node *dad, - struct device_node ***allnextpp, - unsigned long fpsize) -{ - struct device_node *np; - struct property *pp, **prev_pp = NULL; - char *pathp; - u32 tag; - unsigned int l, allocl; - int has_name = 0; - int new_format = 0; - - tag = *((u32 *)(*p)); - if (tag != OF_DT_BEGIN_NODE) { - printk("Weird tag at start of node: %x\n", tag); - return mem; - } - *p += 4; - pathp = (char *)*p; - l = allocl = strlen(pathp) + 1; - *p = _ALIGN(*p + l, 4); - - /* version 0x10 has a more compact unit name here instead of the full - * path. we accumulate the full path size using "fpsize", we'll rebuild - * it later. We detect this because the first character of the name is - * not '/'. - */ - if ((*pathp) != '/') { - new_format = 1; - if (fpsize == 0) { - /* root node: special case. fpsize accounts for path - * plus terminating zero. root node only has '/', so - * fpsize should be 2, but we want to avoid the first - * level nodes to have two '/' so we use fpsize 1 here - */ - fpsize = 1; - allocl = 2; - } else { - /* account for '/' and path size minus terminal 0 - * already in 'l' - */ - fpsize += l; - allocl = fpsize; - } - } - - - np = unflatten_dt_alloc(&mem, sizeof(struct device_node) + allocl, - __alignof__(struct device_node)); - if (allnextpp) { - memset(np, 0, sizeof(*np)); - np->full_name = ((char*)np) + sizeof(struct device_node); - if (new_format) { - char *p = np->full_name; - /* rebuild full path for new format */ - if (dad && dad->parent) { - strcpy(p, dad->full_name); -#ifdef DEBUG - if ((strlen(p) + l + 1) != allocl) { - DBG("%s: p: %d, l: %d, a: %d\n", - pathp, (int)strlen(p), l, allocl); - } -#endif - p += strlen(p); - } - *(p++) = '/'; - memcpy(p, pathp, l); - } else - memcpy(np->full_name, pathp, l); - prev_pp = &np->properties; - **allnextpp = np; - *allnextpp = &np->allnext; - if (dad != NULL) { - np->parent = dad; - /* we temporarily use the next field as `last_child'*/ - if (dad->next == 0) - dad->child = np; - else - dad->next->sibling = np; - dad->next = np; - } - kref_init(&np->kref); - } - while(1) { - u32 sz, noff; - char *pname; - - tag = *((u32 *)(*p)); - if (tag == OF_DT_NOP) { - *p += 4; - continue; - } - if (tag != OF_DT_PROP) - break; - *p += 4; - sz = *((u32 *)(*p)); - noff = *((u32 *)((*p) + 4)); - *p += 8; - if (initial_boot_params->version < 0x10) - *p = _ALIGN(*p, sz >= 8 ? 8 : 4); - - pname = find_flat_dt_string(noff); - if (pname == NULL) { - printk("Can't find property name in list !\n"); - break; - } - if (strcmp(pname, "name") == 0) - has_name = 1; - l = strlen(pname) + 1; - pp = unflatten_dt_alloc(&mem, sizeof(struct property), - __alignof__(struct property)); - if (allnextpp) { - if (strcmp(pname, "linux,phandle") == 0) { - np->node = *((u32 *)*p); - if (np->linux_phandle == 0) - np->linux_phandle = np->node; - } - if (strcmp(pname, "ibm,phandle") == 0) - np->linux_phandle = *((u32 *)*p); - pp->name = pname; - pp->length = sz; - pp->value = (void *)*p; - *prev_pp = pp; - prev_pp = &pp->next; - } - *p = _ALIGN((*p) + sz, 4); - } - /* with version 0x10 we may not have the name property, recreate - * it here from the unit name if absent - */ - if (!has_name) { - char *p = pathp, *ps = pathp, *pa = NULL; - int sz; - - while (*p) { - if ((*p) == '@') - pa = p; - if ((*p) == '/') - ps = p + 1; - p++; - } - if (pa < ps) - pa = p; - sz = (pa - ps) + 1; - pp = unflatten_dt_alloc(&mem, sizeof(struct property) + sz, - __alignof__(struct property)); - if (allnextpp) { - pp->name = "name"; - pp->length = sz; - pp->value = pp + 1; - *prev_pp = pp; - prev_pp = &pp->next; - memcpy(pp->value, ps, sz - 1); - ((char *)pp->value)[sz - 1] = 0; - DBG("fixed up name for %s -> %s\n", pathp, - (char *)pp->value); - } - } - if (allnextpp) { - *prev_pp = NULL; - np->name = of_get_property(np, "name", NULL); - np->type = of_get_property(np, "device_type", NULL); - - if (!np->name) - np->name = ""; - if (!np->type) - np->type = ""; - } - while (tag == OF_DT_BEGIN_NODE) { - mem = unflatten_dt_node(mem, p, np, allnextpp, fpsize); - tag = *((u32 *)(*p)); - } - if (tag != OF_DT_END_NODE) { - printk("Weird tag at end of node: %x\n", tag); - return mem; - } - *p += 4; - return mem; -} - static int __init early_parse_mem(char *p) { if (!p) diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 5cdd958..6852ecf 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -166,3 +166,203 @@ int __init of_flat_dt_is_compatible(unsigned long node, const char *compat) return 0; } +static void *__init unflatten_dt_alloc(unsigned long *mem, unsigned long size, + unsigned long align) +{ + void *res; + + *mem = _ALIGN(*mem, align); + res = (void *)*mem; + *mem += size; + + return res; +} + +/** + * unflatten_dt_node - Alloc and populate a device_node from the flat tree + * @p: pointer to node in flat tree + * @dad: Parent struct device_node + * @allnextpp: pointer to ->allnext from last allocated device_node + * @fpsize: Size of the node path up at the current depth. + */ +unsigned long __init unflatten_dt_node(unsigned long mem, + unsigned long *p, + struct device_node *dad, + struct device_node ***allnextpp, + unsigned long fpsize) +{ + struct device_node *np; + struct property *pp, **prev_pp = NULL; + char *pathp; + u32 tag; + unsigned int l, allocl; + int has_name = 0; + int new_format = 0; + + tag = *((u32 *)(*p)); + if (tag != OF_DT_BEGIN_NODE) { + pr_err("Weird tag at start of node: %x\n", tag); + return mem; + } + *p += 4; + pathp = (char *)*p; + l = allocl = strlen(pathp) + 1; + *p = _ALIGN(*p + l, 4); + + /* version 0x10 has a more compact unit name here instead of the full + * path. we accumulate the full path size using "fpsize", we'll rebuild + * it later. We detect this because the first character of the name is + * not '/'. + */ + if ((*pathp) != '/') { + new_format = 1; + if (fpsize == 0) { + /* root node: special case. fpsize accounts for path + * plus terminating zero. root node only has '/', so + * fpsize should be 2, but we want to avoid the first + * level nodes to have two '/' so we use fpsize 1 here + */ + fpsize = 1; + allocl = 2; + } else { + /* account for '/' and path size minus terminal 0 + * already in 'l' + */ + fpsize += l; + allocl = fpsize; + } + } + + np = unflatten_dt_alloc(&mem, sizeof(struct device_node) + allocl, + __alignof__(struct device_node)); + if (allnextpp) { + memset(np, 0, sizeof(*np)); + np->full_name = ((char *)np) + sizeof(struct device_node); + if (new_format) { + char *fn = np->full_name; + /* rebuild full path for new format */ + if (dad && dad->parent) { + strcpy(fn, dad->full_name); +#ifdef DEBUG + if ((strlen(fn) + l + 1) != allocl) { + pr_debug("%s: p: %d, l: %d, a: %d\n", + pathp, (int)strlen(fn), + l, allocl); + } +#endif + fn += strlen(fn); + } + *(fn++) = '/'; + memcpy(fn, pathp, l); + } else + memcpy(np->full_name, pathp, l); + prev_pp = &np->properties; + **allnextpp = np; + *allnextpp = &np->allnext; + if (dad != NULL) { + np->parent = dad; + /* we temporarily use the next field as `last_child'*/ + if (dad->next == NULL) + dad->child = np; + else + dad->next->sibling = np; + dad->next = np; + } + kref_init(&np->kref); + } + while (1) { + u32 sz, noff; + char *pname; + + tag = *((u32 *)(*p)); + if (tag == OF_DT_NOP) { + *p += 4; + continue; + } + if (tag != OF_DT_PROP) + break; + *p += 4; + sz = *((u32 *)(*p)); + noff = *((u32 *)((*p) + 4)); + *p += 8; + if (initial_boot_params->version < 0x10) + *p = _ALIGN(*p, sz >= 8 ? 8 : 4); + + pname = find_flat_dt_string(noff); + if (pname == NULL) { + pr_info("Can't find property name in list !\n"); + break; + } + if (strcmp(pname, "name") == 0) + has_name = 1; + l = strlen(pname) + 1; + pp = unflatten_dt_alloc(&mem, sizeof(struct property), + __alignof__(struct property)); + if (allnextpp) { + if (strcmp(pname, "linux,phandle") == 0) { + np->node = *((u32 *)*p); + if (np->linux_phandle == 0) + np->linux_phandle = np->node; + } + if (strcmp(pname, "ibm,phandle") == 0) + np->linux_phandle = *((u32 *)*p); + pp->name = pname; + pp->length = sz; + pp->value = (void *)*p; + *prev_pp = pp; + prev_pp = &pp->next; + } + *p = _ALIGN((*p) + sz, 4); + } + /* with version 0x10 we may not have the name property, recreate + * it here from the unit name if absent + */ + if (!has_name) { + char *p1 = pathp, *ps = pathp, *pa = NULL; + int sz; + + while (*p1) { + if ((*p1) == '@') + pa = p1; + if ((*p1) == '/') + ps = p1 + 1; + p1++; + } + if (pa < ps) + pa = p1; + sz = (pa - ps) + 1; + pp = unflatten_dt_alloc(&mem, sizeof(struct property) + sz, + __alignof__(struct property)); + if (allnextpp) { + pp->name = "name"; + pp->length = sz; + pp->value = pp + 1; + *prev_pp = pp; + prev_pp = &pp->next; + memcpy(pp->value, ps, sz - 1); + ((char *)pp->value)[sz - 1] = 0; + pr_debug("fixed up name for %s -> %s\n", pathp, + (char *)pp->value); + } + } + if (allnextpp) { + *prev_pp = NULL; + np->name = of_get_property(np, "name", NULL); + np->type = of_get_property(np, "device_type", NULL); + + if (!np->name) + np->name = ""; + if (!np->type) + np->type = ""; + } + while (tag == OF_DT_BEGIN_NODE) { + mem = unflatten_dt_node(mem, p, np, allnextpp, fpsize); + tag = *((u32 *)(*p)); + } + if (tag != OF_DT_END_NODE) { + pr_err("Weird tag at end of node: %x\n", tag); + return mem; + } + *p += 4; + return mem; +} diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index 81231e0..ace9068 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -69,6 +69,10 @@ extern void *of_get_flat_dt_prop(unsigned long node, const char *name, unsigned long *size); extern int of_flat_dt_is_compatible(unsigned long node, const char *name); extern unsigned long of_get_flat_dt_root(void); +extern unsigned long unflatten_dt_node(unsigned long mem, unsigned long *p, + struct device_node *dad, + struct device_node ***allnextpp, + unsigned long fpsize); /* Other Prototypes */ extern void finish_device_tree(void); -- cgit v0.10.2 From 41f880091c15b039ffcc8b3d831656b81517a6d3 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 23 Nov 2009 20:07:01 -0700 Subject: of/flattree: Merge unflatten_device_tree Merge common code between PowerPC and MicroBlaze Signed-off-by: Grant Likely Reviewed-by: Wolfram Sang Tested-by: Michal Simek diff --git a/arch/microblaze/include/asm/prom.h b/arch/microblaze/include/asm/prom.h index ef3ec1d..07d1063 100644 --- a/arch/microblaze/include/asm/prom.h +++ b/arch/microblaze/include/asm/prom.h @@ -37,7 +37,6 @@ extern struct device_node *of_chosen; #define HAVE_ARCH_DEVTREE_FIXUPS -extern struct device_node *allnodes; /* temporary while merging */ extern rwlock_t devtree_lock; /* temporary while merging */ /* For updating the device tree at runtime */ diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 021770a..901d538 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -50,55 +50,6 @@ typedef u32 cell_t; /* export that to outside world */ struct device_node *of_chosen; -/** - * unflattens the device-tree passed by the firmware, creating the - * tree of struct device_node. It also fills the "name" and "type" - * pointers of the nodes so the normal device-tree walking functions - * can be used (this used to be done by finish_device_tree) - */ -void __init unflatten_device_tree(void) -{ - unsigned long start, mem, size; - struct device_node **allnextp = &allnodes; - - pr_debug(" -> unflatten_device_tree()\n"); - - /* First pass, scan for size */ - start = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; - size = unflatten_dt_node(0, &start, NULL, NULL, 0); - size = (size | 3) + 1; - - pr_debug(" size is %lx, allocating...\n", size); - - /* Allocate memory for the expanded device tree */ - mem = lmb_alloc(size + 4, __alignof__(struct device_node)); - mem = (unsigned long) __va(mem); - - ((u32 *)mem)[size / 4] = 0xdeadbeef; - - pr_debug(" unflattening %lx...\n", mem); - - /* Second pass, do actual unflattening */ - start = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; - unflatten_dt_node(mem, &start, NULL, &allnextp, 0); - if (*((u32 *)start) != OF_DT_END) - printk(KERN_WARNING "Weird tag at end of tree: %08x\n", - *((u32 *)start)); - if (((u32 *)mem)[size / 4] != 0xdeadbeef) - printk(KERN_WARNING "End of tree marker overwritten: %08x\n", - ((u32 *)mem)[size / 4]); - *allnextp = NULL; - - /* Get pointer to OF "/chosen" node for use everywhere */ - of_chosen = of_find_node_by_path("/chosen"); - if (of_chosen == NULL) - of_chosen = of_find_node_by_path("/chosen@0"); - - pr_debug(" <- unflatten_device_tree()\n"); -} - #define early_init_dt_scan_drconf_memory(node) 0 static int __init early_init_dt_scan_cpus(unsigned long node, diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index a102a0a..1280f34 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -73,8 +73,6 @@ unsigned long tce_alloc_start, tce_alloc_end; typedef u32 cell_t; -extern struct device_node *allnodes; /* temporary while merging */ - extern rwlock_t devtree_lock; /* temporary while merging */ /* export that to outside world */ @@ -119,54 +117,6 @@ static void __init move_device_tree(void) DBG("<- move_device_tree\n"); } -/** - * unflattens the device-tree passed by the firmware, creating the - * tree of struct device_node. It also fills the "name" and "type" - * pointers of the nodes so the normal device-tree walking functions - * can be used (this used to be done by finish_device_tree) - */ -void __init unflatten_device_tree(void) -{ - unsigned long start, mem, size; - struct device_node **allnextp = &allnodes; - - DBG(" -> unflatten_device_tree()\n"); - - /* First pass, scan for size */ - start = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; - size = unflatten_dt_node(0, &start, NULL, NULL, 0); - size = (size | 3) + 1; - - DBG(" size is %lx, allocating...\n", size); - - /* Allocate memory for the expanded device tree */ - mem = lmb_alloc(size + 4, __alignof__(struct device_node)); - mem = (unsigned long) __va(mem); - - ((u32 *)mem)[size / 4] = 0xdeadbeef; - - DBG(" unflattening %lx...\n", mem); - - /* Second pass, do actual unflattening */ - start = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; - unflatten_dt_node(mem, &start, NULL, &allnextp, 0); - if (*((u32 *)start) != OF_DT_END) - printk(KERN_WARNING "Weird tag at end of tree: %08x\n", *((u32 *)start)); - if (((u32 *)mem)[size / 4] != 0xdeadbeef) - printk(KERN_WARNING "End of tree marker overwritten: %08x\n", - ((u32 *)mem)[size / 4] ); - *allnextp = NULL; - - /* Get pointer to OF "/chosen" node for use everywhere */ - of_chosen = of_find_node_by_path("/chosen"); - if (of_chosen == NULL) - of_chosen = of_find_node_by_path("/chosen@0"); - - DBG(" <- unflatten_device_tree()\n"); -} - /* * ibm,pa-features is a per-cpu property that contains a string of * attribute descriptors, each of which has a 2 byte header plus up diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 6852ecf..43d236c 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -9,6 +9,8 @@ * version 2 as published by the Free Software Foundation. */ +#include +#include #include #include @@ -366,3 +368,53 @@ unsigned long __init unflatten_dt_node(unsigned long mem, *p += 4; return mem; } + +/** + * unflatten_device_tree - create tree of device_nodes from flat blob + * + * unflattens the device-tree passed by the firmware, creating the + * tree of struct device_node. It also fills the "name" and "type" + * pointers of the nodes so the normal device-tree walking functions + * can be used. + */ +void __init unflatten_device_tree(void) +{ + unsigned long start, mem, size; + struct device_node **allnextp = &allnodes; + + pr_debug(" -> unflatten_device_tree()\n"); + + /* First pass, scan for size */ + start = ((unsigned long)initial_boot_params) + + initial_boot_params->off_dt_struct; + size = unflatten_dt_node(0, &start, NULL, NULL, 0); + size = (size | 3) + 1; + + pr_debug(" size is %lx, allocating...\n", size); + + /* Allocate memory for the expanded device tree */ + mem = lmb_alloc(size + 4, __alignof__(struct device_node)); + mem = (unsigned long) __va(mem); + + ((u32 *)mem)[size / 4] = 0xdeadbeef; + + pr_debug(" unflattening %lx...\n", mem); + + /* Second pass, do actual unflattening */ + start = ((unsigned long)initial_boot_params) + + initial_boot_params->off_dt_struct; + unflatten_dt_node(mem, &start, NULL, &allnextp, 0); + if (*((u32 *)start) != OF_DT_END) + pr_warning("Weird tag at end of tree: %08x\n", *((u32 *)start)); + if (((u32 *)mem)[size / 4] != 0xdeadbeef) + pr_warning("End of tree marker overwritten: %08x\n", + ((u32 *)mem)[size / 4]); + *allnextp = NULL; + + /* Get pointer to OF "/chosen" node for use everywhere */ + of_chosen = of_find_node_by_path("/chosen"); + if (of_chosen == NULL) + of_chosen = of_find_node_by_path("/chosen@0"); + + pr_debug(" <- unflatten_device_tree()\n"); +} diff --git a/include/linux/of.h b/include/linux/of.h index e7facd8..bec2157 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -63,6 +63,9 @@ struct device_node { #endif }; +/* Pointer for first entry in chain of all nodes. */ +extern struct device_node *allnodes; + static inline int of_node_check_flag(struct device_node *n, unsigned long flag) { return test_bit(flag, &n->_flags); diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index ace9068..81231e0 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -69,10 +69,6 @@ extern void *of_get_flat_dt_prop(unsigned long node, const char *name, unsigned long *size); extern int of_flat_dt_is_compatible(unsigned long node, const char *name); extern unsigned long of_get_flat_dt_root(void); -extern unsigned long unflatten_dt_node(unsigned long mem, unsigned long *p, - struct device_node *dad, - struct device_node ***allnextpp, - unsigned long fpsize); /* Other Prototypes */ extern void finish_device_tree(void); -- cgit v0.10.2 From 02af11b03fce3ddb264d7873d7a2e295e697938c Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 23 Nov 2009 20:16:45 -0700 Subject: of: merge prom_{add,remove,modify}_property Merge common code between PowerPC and MicroBlaze Signed-off-by: Grant Likely Reviewed-by: Wolfram Sang Tested-by: Michal Simek diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 901d538..a38e373 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -606,119 +606,6 @@ out_unlock: write_unlock_irqrestore(&devtree_lock, flags); } -/* - * Add a property to a node - */ -int prom_add_property(struct device_node *np, struct property *prop) -{ - struct property **next; - unsigned long flags; - - prop->next = NULL; - write_lock_irqsave(&devtree_lock, flags); - next = &np->properties; - while (*next) { - if (strcmp(prop->name, (*next)->name) == 0) { - /* duplicate ! don't insert it */ - write_unlock_irqrestore(&devtree_lock, flags); - return -1; - } - next = &(*next)->next; - } - *next = prop; - write_unlock_irqrestore(&devtree_lock, flags); - -#ifdef CONFIG_PROC_DEVICETREE - /* try to add to proc as well if it was initialized */ - if (np->pde) - proc_device_tree_add_prop(np->pde, prop); -#endif /* CONFIG_PROC_DEVICETREE */ - - return 0; -} - -/* - * Remove a property from a node. Note that we don't actually - * remove it, since we have given out who-knows-how-many pointers - * to the data using get-property. Instead we just move the property - * to the "dead properties" list, so it won't be found any more. - */ -int prom_remove_property(struct device_node *np, struct property *prop) -{ - struct property **next; - unsigned long flags; - int found = 0; - - write_lock_irqsave(&devtree_lock, flags); - next = &np->properties; - while (*next) { - if (*next == prop) { - /* found the node */ - *next = prop->next; - prop->next = np->deadprops; - np->deadprops = prop; - found = 1; - break; - } - next = &(*next)->next; - } - write_unlock_irqrestore(&devtree_lock, flags); - - if (!found) - return -ENODEV; - -#ifdef CONFIG_PROC_DEVICETREE - /* try to remove the proc node as well */ - if (np->pde) - proc_device_tree_remove_prop(np->pde, prop); -#endif /* CONFIG_PROC_DEVICETREE */ - - return 0; -} - -/* - * Update a property in a node. Note that we don't actually - * remove it, since we have given out who-knows-how-many pointers - * to the data using get-property. Instead we just move the property - * to the "dead properties" list, and add the new property to the - * property list - */ -int prom_update_property(struct device_node *np, - struct property *newprop, - struct property *oldprop) -{ - struct property **next; - unsigned long flags; - int found = 0; - - write_lock_irqsave(&devtree_lock, flags); - next = &np->properties; - while (*next) { - if (*next == oldprop) { - /* found the node */ - newprop->next = oldprop->next; - *next = newprop; - oldprop->next = np->deadprops; - np->deadprops = oldprop; - found = 1; - break; - } - next = &(*next)->next; - } - write_unlock_irqrestore(&devtree_lock, flags); - - if (!found) - return -ENODEV; - -#ifdef CONFIG_PROC_DEVICETREE - /* try to add to proc as well if it was initialized */ - if (np->pde) - proc_device_tree_update_prop(np->pde, newprop, oldprop); -#endif /* CONFIG_PROC_DEVICETREE */ - - return 0; -} - #if defined(CONFIG_DEBUG_FS) && defined(DEBUG) static struct debugfs_blob_wrapper flat_dt_blob; diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 1280f34..7f88566 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -1130,120 +1130,6 @@ static int __init prom_reconfig_setup(void) __initcall(prom_reconfig_setup); #endif -/* - * Add a property to a node - */ -int prom_add_property(struct device_node* np, struct property* prop) -{ - struct property **next; - unsigned long flags; - - prop->next = NULL; - write_lock_irqsave(&devtree_lock, flags); - next = &np->properties; - while (*next) { - if (strcmp(prop->name, (*next)->name) == 0) { - /* duplicate ! don't insert it */ - write_unlock_irqrestore(&devtree_lock, flags); - return -1; - } - next = &(*next)->next; - } - *next = prop; - write_unlock_irqrestore(&devtree_lock, flags); - -#ifdef CONFIG_PROC_DEVICETREE - /* try to add to proc as well if it was initialized */ - if (np->pde) - proc_device_tree_add_prop(np->pde, prop); -#endif /* CONFIG_PROC_DEVICETREE */ - - return 0; -} - -/* - * Remove a property from a node. Note that we don't actually - * remove it, since we have given out who-knows-how-many pointers - * to the data using get-property. Instead we just move the property - * to the "dead properties" list, so it won't be found any more. - */ -int prom_remove_property(struct device_node *np, struct property *prop) -{ - struct property **next; - unsigned long flags; - int found = 0; - - write_lock_irqsave(&devtree_lock, flags); - next = &np->properties; - while (*next) { - if (*next == prop) { - /* found the node */ - *next = prop->next; - prop->next = np->deadprops; - np->deadprops = prop; - found = 1; - break; - } - next = &(*next)->next; - } - write_unlock_irqrestore(&devtree_lock, flags); - - if (!found) - return -ENODEV; - -#ifdef CONFIG_PROC_DEVICETREE - /* try to remove the proc node as well */ - if (np->pde) - proc_device_tree_remove_prop(np->pde, prop); -#endif /* CONFIG_PROC_DEVICETREE */ - - return 0; -} - -/* - * Update a property in a node. Note that we don't actually - * remove it, since we have given out who-knows-how-many pointers - * to the data using get-property. Instead we just move the property - * to the "dead properties" list, and add the new property to the - * property list - */ -int prom_update_property(struct device_node *np, - struct property *newprop, - struct property *oldprop) -{ - struct property **next; - unsigned long flags; - int found = 0; - - write_lock_irqsave(&devtree_lock, flags); - next = &np->properties; - while (*next) { - if (*next == oldprop) { - /* found the node */ - newprop->next = oldprop->next; - *next = newprop; - oldprop->next = np->deadprops; - np->deadprops = oldprop; - found = 1; - break; - } - next = &(*next)->next; - } - write_unlock_irqrestore(&devtree_lock, flags); - - if (!found) - return -ENODEV; - -#ifdef CONFIG_PROC_DEVICETREE - /* try to add to proc as well if it was initialized */ - if (np->pde) - proc_device_tree_update_prop(np->pde, newprop, oldprop); -#endif /* CONFIG_PROC_DEVICETREE */ - - return 0; -} - - /* Find the device node for a given logical cpu number, also returns the cpu * local thread number (index in ibm,interrupt-server#s) if relevant and * asked for (non NULL) diff --git a/drivers/of/base.c b/drivers/of/base.c index e6627b2..ec56739 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -658,3 +658,119 @@ err0: return ret; } EXPORT_SYMBOL(of_parse_phandles_with_args); + +/** + * prom_add_property - Add a property to a node + */ +int prom_add_property(struct device_node *np, struct property *prop) +{ + struct property **next; + unsigned long flags; + + prop->next = NULL; + write_lock_irqsave(&devtree_lock, flags); + next = &np->properties; + while (*next) { + if (strcmp(prop->name, (*next)->name) == 0) { + /* duplicate ! don't insert it */ + write_unlock_irqrestore(&devtree_lock, flags); + return -1; + } + next = &(*next)->next; + } + *next = prop; + write_unlock_irqrestore(&devtree_lock, flags); + +#ifdef CONFIG_PROC_DEVICETREE + /* try to add to proc as well if it was initialized */ + if (np->pde) + proc_device_tree_add_prop(np->pde, prop); +#endif /* CONFIG_PROC_DEVICETREE */ + + return 0; +} + +/** + * prom_remove_property - Remove a property from a node. + * + * Note that we don't actually remove it, since we have given out + * who-knows-how-many pointers to the data using get-property. + * Instead we just move the property to the "dead properties" + * list, so it won't be found any more. + */ +int prom_remove_property(struct device_node *np, struct property *prop) +{ + struct property **next; + unsigned long flags; + int found = 0; + + write_lock_irqsave(&devtree_lock, flags); + next = &np->properties; + while (*next) { + if (*next == prop) { + /* found the node */ + *next = prop->next; + prop->next = np->deadprops; + np->deadprops = prop; + found = 1; + break; + } + next = &(*next)->next; + } + write_unlock_irqrestore(&devtree_lock, flags); + + if (!found) + return -ENODEV; + +#ifdef CONFIG_PROC_DEVICETREE + /* try to remove the proc node as well */ + if (np->pde) + proc_device_tree_remove_prop(np->pde, prop); +#endif /* CONFIG_PROC_DEVICETREE */ + + return 0; +} + +/* + * prom_update_property - Update a property in a node. + * + * Note that we don't actually remove it, since we have given out + * who-knows-how-many pointers to the data using get-property. + * Instead we just move the property to the "dead properties" list, + * and add the new property to the property list + */ +int prom_update_property(struct device_node *np, + struct property *newprop, + struct property *oldprop) +{ + struct property **next; + unsigned long flags; + int found = 0; + + write_lock_irqsave(&devtree_lock, flags); + next = &np->properties; + while (*next) { + if (*next == oldprop) { + /* found the node */ + newprop->next = oldprop->next; + *next = newprop; + oldprop->next = np->deadprops; + np->deadprops = oldprop; + found = 1; + break; + } + next = &(*next)->next; + } + write_unlock_irqrestore(&devtree_lock, flags); + + if (!found) + return -ENODEV; + +#ifdef CONFIG_PROC_DEVICETREE + /* try to add to proc as well if it was initialized */ + if (np->pde) + proc_device_tree_update_prop(np->pde, newprop, oldprop); +#endif /* CONFIG_PROC_DEVICETREE */ + + return 0; +} -- cgit v0.10.2 From 2be09cb993826b52c9fc1d44747c20dd43a50038 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 23 Nov 2009 20:16:46 -0700 Subject: of: remove special case definition of of_read_ulong() Special case of of_read_ulong() was defined for PPC32 to toss away all but the last 32 bits when a large number value was read, and the 'normal' version for ppc64 just #defined of_read_ulong to of_read_number which causes compiler warnings on MicroBlaze and other 32 bit architectures because it returns a u64 instead of a ulong. This patch fixes the problem by defining a common implementation of of_read_ulong() that works everywhere. Signed-off-by: Grant Likely Reviewed-by: Wolfram Sang Tested-by: Michal Simek diff --git a/include/linux/of.h b/include/linux/of.h index bec2157..d4c014a 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -113,14 +113,11 @@ static inline u64 of_read_number(const u32 *cell, int size) } /* Like of_read_number, but we want an unsigned long result */ -#ifdef CONFIG_PPC32 static inline unsigned long of_read_ulong(const u32 *cell, int size) { - return cell[size-1]; + /* toss away upper bits if unsigned long is smaller than u64 */ + return of_read_number(cell, size); } -#else -#define of_read_ulong(cell, size) of_read_number(cell, size) -#endif #include -- cgit v0.10.2 From 09a072947791088b88ae15111cf68fc5aaaf758d Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 9 Nov 2009 16:27:40 +0900 Subject: sh: hw-breakpoints: Add preliminary support for SH-4A UBC. This adds preliminary support for the SH-4A UBC to the hw-breakpoints API. Presently only a single channel is implemented, and the ptrace interface still needs to be converted. This is the first step to cleaning up the long-standing UBC mess, making the UBC more generally accessible, and finally making it SMP safe. An additional abstraction will be layered on top of this as with the perf events code to permit the various CPU families to wire up support for their own specific UBCs, as many variations exist. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 88cdeb9..d563884 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -39,6 +39,7 @@ config SUPERH32 select HAVE_FUNCTION_TRACE_MCOUNT_TEST select HAVE_FUNCTION_GRAPH_TRACER select HAVE_ARCH_KGDB + select HAVE_HW_BREAKPOINT if CPU_SH4A select ARCH_HIBERNATION_POSSIBLE if MMU config SUPERH64 diff --git a/arch/sh/include/asm/Kbuild b/arch/sh/include/asm/Kbuild index e121c30..46cb934 100644 --- a/arch/sh/include/asm/Kbuild +++ b/arch/sh/include/asm/Kbuild @@ -1,6 +1,8 @@ include include/asm-generic/Kbuild.asm -header-y += cachectl.h cpu-features.h +header-y += cachectl.h +header-y += cpu-features.h +header-y += hw_breakpoint.h unifdef-y += unistd_32.h unifdef-y += unistd_64.h diff --git a/arch/sh/include/asm/hw_breakpoint.h b/arch/sh/include/asm/hw_breakpoint.h new file mode 100644 index 0000000..0f4a00f --- /dev/null +++ b/arch/sh/include/asm/hw_breakpoint.h @@ -0,0 +1,53 @@ +#ifndef __ASM_SH_HW_BREAKPOINT_H +#define __ASM_SH_HW_BREAKPOINT_H + +#include +#include +#include + +#ifdef __KERNEL__ +#define __ARCH_HW_BREAKPOINT_H + +struct arch_hw_breakpoint { + char *name; /* Contains name of the symbol to set bkpt */ + unsigned long address; + unsigned long asid; + u16 len; + u16 type; +}; + +enum { + SH_BREAKPOINT_READ = (1 << 1), + SH_BREAKPOINT_WRITE = (1 << 2), + SH_BREAKPOINT_RW = SH_BREAKPOINT_READ | SH_BREAKPOINT_WRITE, + + SH_BREAKPOINT_LEN_1 = (1 << 12), + SH_BREAKPOINT_LEN_2 = (1 << 13), + SH_BREAKPOINT_LEN_4 = SH_BREAKPOINT_LEN_1 | SH_BREAKPOINT_LEN_2, + SH_BREAKPOINT_LEN_8 = (1 << 14), +}; + +/* Total number of available UBC channels */ +#define HBP_NUM 1 /* XXX */ + +struct perf_event; +struct task_struct; +struct pmu; + +extern int arch_check_va_in_userspace(unsigned long va, u16 hbp_len); +extern int arch_validate_hwbkpt_settings(struct perf_event *bp, + struct task_struct *tsk); +extern int hw_breakpoint_exceptions_notify(struct notifier_block *unused, + unsigned long val, void *data); + +int arch_install_hw_breakpoint(struct perf_event *bp); +void arch_uninstall_hw_breakpoint(struct perf_event *bp); +void hw_breakpoint_pmu_read(struct perf_event *bp); +void hw_breakpoint_pmu_unthrottle(struct perf_event *bp); + +extern void arch_fill_perf_breakpoint(struct perf_event *bp); + +extern struct pmu perf_ops_bp; + +#endif /* __KERNEL__ */ +#endif /* __ASM_SH_HW_BREAKPOINT_H */ diff --git a/arch/sh/include/asm/kdebug.h b/arch/sh/include/asm/kdebug.h index 985219f..5f6d2e9 100644 --- a/arch/sh/include/asm/kdebug.h +++ b/arch/sh/include/asm/kdebug.h @@ -6,6 +6,8 @@ enum die_val { DIE_TRAP, DIE_NMI, DIE_OOPS, + DIE_BREAKPOINT, + DIE_SSTEP, }; #endif /* __ASM_SH_KDEBUG_H */ diff --git a/arch/sh/include/asm/processor_32.h b/arch/sh/include/asm/processor_32.h index 9a87149..f4b5404 100644 --- a/arch/sh/include/asm/processor_32.h +++ b/arch/sh/include/asm/processor_32.h @@ -14,6 +14,7 @@ #include #include #include +#include /* * Default implementation of macro that returns current @@ -99,8 +100,8 @@ struct thread_struct { unsigned long sp; unsigned long pc; - /* Hardware debugging registers */ - unsigned long ubc_pc; + /* Save middle states of ptrace breakpoints */ + struct perf_event *ptrace_bps[NR_UBC_CHANNELS]; /* floating point info */ union sh_fpu_union fpu; @@ -111,9 +112,6 @@ struct thread_struct { #endif }; -/* Count of active tasks with UBC settings */ -extern int ubc_usercnt; - #define INIT_THREAD { \ .sp = sizeof(init_stack) + (long) &init_stack, \ } diff --git a/arch/sh/include/asm/system.h b/arch/sh/include/asm/system.h index b5c5acd..1014da8 100644 --- a/arch/sh/include/asm/system.h +++ b/arch/sh/include/asm/system.h @@ -144,8 +144,6 @@ void per_cpu_trap_init(void); void default_idle(void); void cpu_idle_wait(void); -asmlinkage void break_point_trap(void); - #ifdef CONFIG_SUPERH32 #define BUILD_TRAP_HANDLER(name) \ asmlinkage void name##_trap_handler(unsigned long r4, unsigned long r5, \ diff --git a/arch/sh/include/asm/thread_info.h b/arch/sh/include/asm/thread_info.h index bdeb9d4..8ab9145 100644 --- a/arch/sh/include/asm/thread_info.h +++ b/arch/sh/include/asm/thread_info.h @@ -117,6 +117,7 @@ extern void free_thread_info(struct thread_info *ti); #define TIF_SECCOMP 6 /* secure computing */ #define TIF_NOTIFY_RESUME 7 /* callback before returning to user */ #define TIF_SYSCALL_TRACEPOINT 8 /* for ftrace syscall instrumentation */ +#define TIF_DEBUG 9 /* uses UBC */ #define TIF_USEDFPU 16 /* FPU was used by this task this quantum (SMP) */ #define TIF_POLLING_NRFLAG 17 /* true if poll_idle() is polling TIF_NEED_RESCHED */ #define TIF_MEMDIE 18 @@ -131,6 +132,7 @@ extern void free_thread_info(struct thread_info *ti); #define _TIF_SECCOMP (1 << TIF_SECCOMP) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) #define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT) +#define _TIF_DEBUG (1 << TIF_DEBUG) #define _TIF_USEDFPU (1 << TIF_USEDFPU) #define _TIF_POLLING_NRFLAG (1 << TIF_POLLING_NRFLAG) #define _TIF_FREEZE (1 << TIF_FREEZE) diff --git a/arch/sh/include/asm/ubc.h b/arch/sh/include/asm/ubc.h index 4ca4b77..dd78781 100644 --- a/arch/sh/include/asm/ubc.h +++ b/arch/sh/include/asm/ubc.h @@ -10,8 +10,8 @@ */ #ifndef __ASM_SH_UBC_H #define __ASM_SH_UBC_H -#ifdef __KERNEL__ +#ifdef __KERNEL__ #include /* User Break Controller */ @@ -60,6 +60,12 @@ #define BRCR_UBDE (1 << 0) #endif +/* + * All SH parts have 2 UBC channels. I defy any hardware designer to + * invalidate this assertion. + */ +#define NR_UBC_CHANNELS 2 + #ifndef __ASSEMBLY__ /* arch/sh/kernel/cpu/ubc.S */ extern void ubc_sleep(void); diff --git a/arch/sh/kernel/Makefile b/arch/sh/kernel/Makefile index a2d0a40..649daad 100644 --- a/arch/sh/kernel/Makefile +++ b/arch/sh/kernel/Makefile @@ -36,6 +36,7 @@ obj-$(CONFIG_DUMP_CODE) += disassemble.o obj-$(CONFIG_HIBERNATION) += swsusp.o obj-$(CONFIG_DWARF_UNWINDER) += dwarf.o +obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o obj-$(CONFIG_GENERIC_CLOCKEVENTS_BROADCAST) += localtimer.o EXTRA_CFLAGS += -Werror diff --git a/arch/sh/kernel/cpu/sh3/ex.S b/arch/sh/kernel/cpu/sh3/ex.S index 46610c3..99b4d02 100644 --- a/arch/sh/kernel/cpu/sh3/ex.S +++ b/arch/sh/kernel/cpu/sh3/ex.S @@ -49,7 +49,7 @@ ENTRY(exception_handling_table) .long exception_error ! reserved_instruction (filled by trap_init) /* 180 */ .long exception_error ! illegal_slot_instruction (filled by trap_init) /*1A0*/ .long nmi_trap_handler /* 1C0 */ ! Allow trap to debugger - .long break_point_trap /* 1E0 */ + .long breakpoint_trap_handler /* 1E0 */ /* * Pad the remainder of the table out, exceptions residing in far diff --git a/arch/sh/kernel/hw_breakpoint.c b/arch/sh/kernel/hw_breakpoint.c new file mode 100644 index 0000000..ff3cb3d --- /dev/null +++ b/arch/sh/kernel/hw_breakpoint.c @@ -0,0 +1,416 @@ +/* + * arch/sh/kernel/hw_breakpoint.c + * + * Unified kernel/user-space hardware breakpoint facility for the on-chip UBC. + * + * Copyright (C) 2009 Paul Mundt + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +struct ubc_context { + unsigned long pc; + unsigned long state; +}; + +/* Per cpu ubc channel state */ +static DEFINE_PER_CPU(struct ubc_context, ubc_ctx[HBP_NUM]); + +/* + * Stores the breakpoints currently in use on each breakpoint address + * register for each cpus + */ +static DEFINE_PER_CPU(struct perf_event *, bp_per_reg[HBP_NUM]); + +static int __init ubc_init(void) +{ + __raw_writel(0, UBC_CAMR0); + __raw_writel(0, UBC_CBR0); + __raw_writel(0, UBC_CBCR); + + __raw_writel(UBC_CRR_BIE | UBC_CRR_PCB, UBC_CRR0); + + /* dummy read for write posting */ + (void)__raw_readl(UBC_CRR0); + + return 0; +} +arch_initcall(ubc_init); + +/* + * Install a perf counter breakpoint. + * + * We seek a free UBC channel and use it for this breakpoint. + * + * Atomic: we hold the counter->ctx->lock and we only handle variables + * and registers local to this cpu. + */ +int arch_install_hw_breakpoint(struct perf_event *bp) +{ + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + struct ubc_context *ubc_ctx; + int i; + + for (i = 0; i < HBP_NUM; i++) { + struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); + + if (!*slot) { + *slot = bp; + break; + } + } + + if (WARN_ONCE(i == HBP_NUM, "Can't find any breakpoint slot")) + return -EBUSY; + + ubc_ctx = &__get_cpu_var(ubc_ctx[i]); + + ubc_ctx->pc = info->address; + ubc_ctx->state = info->len | info->type; + + __raw_writel(UBC_CBR_CE | ubc_ctx->state, UBC_CBR0); + __raw_writel(ubc_ctx->pc, UBC_CAR0); + + return 0; +} + +/* + * Uninstall the breakpoint contained in the given counter. + * + * First we search the debug address register it uses and then we disable + * it. + * + * Atomic: we hold the counter->ctx->lock and we only handle variables + * and registers local to this cpu. + */ +void arch_uninstall_hw_breakpoint(struct perf_event *bp) +{ + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + struct ubc_context *ubc_ctx; + int i; + + for (i = 0; i < HBP_NUM; i++) { + struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); + + if (*slot == bp) { + *slot = NULL; + break; + } + } + + if (WARN_ONCE(i == HBP_NUM, "Can't find any breakpoint slot")) + return; + + ubc_ctx = &__get_cpu_var(ubc_ctx[i]); + ubc_ctx->pc = 0; + ubc_ctx->state &= ~(info->len | info->type); + + __raw_writel(ubc_ctx->pc, UBC_CBR0); + __raw_writel(ubc_ctx->state, UBC_CAR0); +} + +static int get_hbp_len(u16 hbp_len) +{ + unsigned int len_in_bytes = 0; + + switch (hbp_len) { + case SH_BREAKPOINT_LEN_1: + len_in_bytes = 1; + break; + case SH_BREAKPOINT_LEN_2: + len_in_bytes = 2; + break; + case SH_BREAKPOINT_LEN_4: + len_in_bytes = 4; + break; + case SH_BREAKPOINT_LEN_8: + len_in_bytes = 8; + break; + } + return len_in_bytes; +} + +/* + * Check for virtual address in user space. + */ +int arch_check_va_in_userspace(unsigned long va, u16 hbp_len) +{ + unsigned int len; + + len = get_hbp_len(hbp_len); + + return (va <= TASK_SIZE - len); +} + +/* + * Check for virtual address in kernel space. + */ +static int arch_check_va_in_kernelspace(unsigned long va, u8 hbp_len) +{ + unsigned int len; + + len = get_hbp_len(hbp_len); + + return (va >= TASK_SIZE) && ((va + len - 1) >= TASK_SIZE); +} + +/* + * Store a breakpoint's encoded address, length, and type. + */ +static int arch_store_info(struct perf_event *bp) +{ + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + + /* + * User-space requests will always have the address field populated + * For kernel-addresses, either the address or symbol name can be + * specified. + */ + if (info->name) + info->address = (unsigned long)kallsyms_lookup_name(info->name); + if (info->address) { + info->asid = get_asid(); + return 0; + } + + return -EINVAL; +} + +int arch_bp_generic_fields(int sh_len, int sh_type, + int *gen_len, int *gen_type) +{ + /* Len */ + switch (sh_len) { + case SH_BREAKPOINT_LEN_1: + *gen_len = HW_BREAKPOINT_LEN_1; + break; + case SH_BREAKPOINT_LEN_2: + *gen_len = HW_BREAKPOINT_LEN_2; + break; + case SH_BREAKPOINT_LEN_4: + *gen_len = HW_BREAKPOINT_LEN_4; + break; + case SH_BREAKPOINT_LEN_8: + *gen_len = HW_BREAKPOINT_LEN_8; + break; + default: + return -EINVAL; + } + + /* Type */ + switch (sh_type) { + case SH_BREAKPOINT_READ: + *gen_type = HW_BREAKPOINT_R; + case SH_BREAKPOINT_WRITE: + *gen_type = HW_BREAKPOINT_W; + break; + case SH_BREAKPOINT_RW: + *gen_type = HW_BREAKPOINT_W | HW_BREAKPOINT_R; + break; + default: + return -EINVAL; + } + + return 0; +} + +static int arch_build_bp_info(struct perf_event *bp) +{ + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + + info->address = bp->attr.bp_addr; + + /* Len */ + switch (bp->attr.bp_len) { + case HW_BREAKPOINT_LEN_1: + info->len = SH_BREAKPOINT_LEN_1; + break; + case HW_BREAKPOINT_LEN_2: + info->len = SH_BREAKPOINT_LEN_2; + break; + case HW_BREAKPOINT_LEN_4: + info->len = SH_BREAKPOINT_LEN_4; + break; + case HW_BREAKPOINT_LEN_8: + info->len = SH_BREAKPOINT_LEN_8; + break; + default: + return -EINVAL; + } + + /* Type */ + switch (bp->attr.bp_type) { + case HW_BREAKPOINT_R: + info->type = SH_BREAKPOINT_READ; + break; + case HW_BREAKPOINT_W: + info->type = SH_BREAKPOINT_WRITE; + break; + case HW_BREAKPOINT_W | HW_BREAKPOINT_R: + info->type = SH_BREAKPOINT_RW; + break; + default: + return -EINVAL; + } + + return 0; +} + +/* + * Validate the arch-specific HW Breakpoint register settings + */ +int arch_validate_hwbkpt_settings(struct perf_event *bp, + struct task_struct *tsk) +{ + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + unsigned int align; + int ret; + + ret = arch_build_bp_info(bp); + if (ret) + return ret; + + ret = -EINVAL; + + switch (info->len) { + case SH_BREAKPOINT_LEN_1: + align = 0; + break; + case SH_BREAKPOINT_LEN_2: + align = 1; + break; + case SH_BREAKPOINT_LEN_4: + align = 3; + break; + case SH_BREAKPOINT_LEN_8: + align = 7; + break; + default: + return ret; + } + + if (bp->callback) + ret = arch_store_info(bp); + + if (ret < 0) + return ret; + + /* + * Check that the low-order bits of the address are appropriate + * for the alignment implied by len. + */ + if (info->address & align) + return -EINVAL; + + /* Check that the virtual address is in the proper range */ + if (tsk) { + if (!arch_check_va_in_userspace(info->address, info->len)) + return -EFAULT; + } else { + if (!arch_check_va_in_kernelspace(info->address, info->len)) + return -EFAULT; + } + + return 0; +} + +/* + * Release the user breakpoints used by ptrace + */ +void flush_ptrace_hw_breakpoint(struct task_struct *tsk) +{ + int i; + struct thread_struct *t = &tsk->thread; + + for (i = 0; i < HBP_NUM; i++) { + unregister_hw_breakpoint(t->ptrace_bps[i]); + t->ptrace_bps[i] = NULL; + } +} + +static int __kprobes hw_breakpoint_handler(struct die_args *args) +{ + int cpu, i, rc = NOTIFY_STOP; + struct perf_event *bp; + unsigned long val; + + val = __raw_readl(UBC_CBR0); + __raw_writel(val & ~UBC_CBR_CE, UBC_CBR0); + + cpu = get_cpu(); + for (i = 0; i < HBP_NUM; i++) { + /* + * The counter may be concurrently released but that can only + * occur from a call_rcu() path. We can then safely fetch + * the breakpoint, use its callback, touch its counter + * while we are in an rcu_read_lock() path. + */ + rcu_read_lock(); + + bp = per_cpu(bp_per_reg[i], cpu); + if (bp) { + rc = NOTIFY_DONE; + } else { + rcu_read_unlock(); + break; + } + + (bp->callback)(bp, args->regs); + + rcu_read_unlock(); + } + + if (bp) { + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + + __raw_writel(UBC_CBR_CE | info->len | info->type, UBC_CBR0); + __raw_writel(info->address, UBC_CAR0); + } + + put_cpu(); + + return rc; +} + +BUILD_TRAP_HANDLER(breakpoint) +{ + unsigned long ex = lookup_exception_vector(); + TRAP_HANDLER_DECL; + + notify_die(DIE_BREAKPOINT, "breakpoint", regs, 0, ex, SIGTRAP); +} + +/* + * Handle debug exception notifications. + */ +int __kprobes hw_breakpoint_exceptions_notify(struct notifier_block *unused, + unsigned long val, void *data) +{ + if (val != DIE_BREAKPOINT) + return NOTIFY_DONE; + + return hw_breakpoint_handler(data); +} + +void hw_breakpoint_pmu_read(struct perf_event *bp) +{ + /* TODO */ +} + +void hw_breakpoint_pmu_unthrottle(struct perf_event *bp) +{ + /* TODO */ +} diff --git a/arch/sh/kernel/process_32.c b/arch/sh/kernel/process_32.c index 0673c47..4a2c866 100644 --- a/arch/sh/kernel/process_32.c +++ b/arch/sh/kernel/process_32.c @@ -25,6 +25,7 @@ #include #include #include +#include #include #include #include @@ -34,8 +35,6 @@ #include #include -int ubc_usercnt = 0; - #ifdef CONFIG_32BIT static void watchdog_trigger_immediate(void) { @@ -148,16 +147,15 @@ int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags) */ void exit_thread(void) { - if (current->thread.ubc_pc) { - current->thread.ubc_pc = 0; - ubc_usercnt -= 1; - } } void flush_thread(void) { -#if defined(CONFIG_SH_FPU) struct task_struct *tsk = current; + + flush_ptrace_hw_breakpoint(tsk); + +#if defined(CONFIG_SH_FPU) /* Forget lazy FPU state */ clear_fpu(tsk, task_pt_regs(tsk)); clear_used_math(); @@ -195,9 +193,7 @@ int copy_thread(unsigned long clone_flags, unsigned long usp, { struct thread_info *ti = task_thread_info(p); struct pt_regs *childregs; -#if defined(CONFIG_SH_FPU) || defined(CONFIG_SH_DSP) struct task_struct *tsk = current; -#endif #if defined(CONFIG_SH_FPU) unlazy_fpu(tsk, regs); @@ -234,53 +230,11 @@ int copy_thread(unsigned long clone_flags, unsigned long usp, p->thread.sp = (unsigned long) childregs; p->thread.pc = (unsigned long) ret_from_fork; - p->thread.ubc_pc = 0; + memset(p->thread.ptrace_bps, 0, sizeof(p->thread.ptrace_bps)); return 0; } -/* Tracing by user break controller. */ -static void ubc_set_tracing(int asid, unsigned long pc) -{ -#if defined(CONFIG_CPU_SH4A) - unsigned long val; - - val = (UBC_CBR_ID_INST | UBC_CBR_RW_READ | UBC_CBR_CE); - val |= (UBC_CBR_AIE | UBC_CBR_AIV_SET(asid)); - - ctrl_outl(val, UBC_CBR0); - ctrl_outl(pc, UBC_CAR0); - ctrl_outl(0x0, UBC_CAMR0); - ctrl_outl(0x0, UBC_CBCR); - - val = (UBC_CRR_RES | UBC_CRR_PCB | UBC_CRR_BIE); - ctrl_outl(val, UBC_CRR0); - - /* Read UBC register that we wrote last, for checking update */ - val = ctrl_inl(UBC_CRR0); - -#else /* CONFIG_CPU_SH4A */ - ctrl_outl(pc, UBC_BARA); - -#ifdef CONFIG_MMU - ctrl_outb(asid, UBC_BASRA); -#endif - - ctrl_outl(0, UBC_BAMRA); - - if (current_cpu_data.type == CPU_SH7729 || - current_cpu_data.type == CPU_SH7710 || - current_cpu_data.type == CPU_SH7712 || - current_cpu_data.type == CPU_SH7203){ - ctrl_outw(BBR_INST | BBR_READ | BBR_CPU, UBC_BBRA); - ctrl_outl(BRCR_PCBA | BRCR_PCTE, UBC_BRCR); - } else { - ctrl_outw(BBR_INST | BBR_READ, UBC_BBRA); - ctrl_outw(BRCR_PCBA, UBC_BRCR); - } -#endif /* CONFIG_CPU_SH4A */ -} - /* * switch_to(x,y) should switch tasks from x to y. * @@ -302,25 +256,6 @@ __switch_to(struct task_struct *prev, struct task_struct *next) : "r" (task_thread_info(next))); #endif - /* If no tasks are using the UBC, we're done */ - if (ubc_usercnt == 0) - /* If no tasks are using the UBC, we're done */; - else if (next->thread.ubc_pc && next->mm) { - int asid = 0; -#ifdef CONFIG_MMU - asid |= cpu_asid(smp_processor_id(), next->mm); -#endif - ubc_set_tracing(asid, next->thread.ubc_pc); - } else { -#if defined(CONFIG_CPU_SH4A) - ctrl_outl(UBC_CBR_INIT, UBC_CBR0); - ctrl_outl(UBC_CRR_INIT, UBC_CRR0); -#else - ctrl_outw(0, UBC_BBRA); - ctrl_outw(0, UBC_BBRB); -#endif - } - return prev; } @@ -412,20 +347,3 @@ unsigned long get_wchan(struct task_struct *p) return pc; } - -asmlinkage void break_point_trap(void) -{ - /* Clear tracing. */ -#if defined(CONFIG_CPU_SH4A) - ctrl_outl(UBC_CBR_INIT, UBC_CBR0); - ctrl_outl(UBC_CRR_INIT, UBC_CRR0); -#else - ctrl_outw(0, UBC_BBRA); - ctrl_outw(0, UBC_BBRB); - ctrl_outl(0, UBC_BRCR); -#endif - current->thread.ubc_pc = 0; - ubc_usercnt -= 1; - - force_sig(SIGTRAP, current); -} diff --git a/arch/sh/kernel/ptrace_32.c b/arch/sh/kernel/ptrace_32.c index 9be35f3..bdb1044 100644 --- a/arch/sh/kernel/ptrace_32.c +++ b/arch/sh/kernel/ptrace_32.c @@ -65,31 +65,12 @@ static inline int put_stack_long(struct task_struct *task, int offset, void user_enable_single_step(struct task_struct *child) { - /* Next scheduling will set up UBC */ - if (child->thread.ubc_pc == 0) - ubc_usercnt += 1; - - child->thread.ubc_pc = get_stack_long(child, - offsetof(struct pt_regs, pc)); - set_tsk_thread_flag(child, TIF_SINGLESTEP); } void user_disable_single_step(struct task_struct *child) { clear_tsk_thread_flag(child, TIF_SINGLESTEP); - - /* - * Ensure the UBC is not programmed at the next context switch. - * - * Normally this is not needed but there are sequences such as - * singlestep, signal delivery, and continue that leave the - * ubc_pc non-zero leading to spurious SIGTRAPs. - */ - if (child->thread.ubc_pc != 0) { - ubc_usercnt -= 1; - child->thread.ubc_pc = 0; - } } /* -- cgit v0.10.2 From f7b3a8355ba6cad251297844a0bdd08898ea36e0 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Tue, 24 Nov 2009 03:26:58 -0700 Subject: of/flattree: Merge early_init_dt_check_for_initrd() Merge common code between PowerPC and Microblaze Signed-off-by: Grant Likely Tested-by: Wolfram Sang Acked-by: Benjamin Herrenschmidt diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index a38e373..7959495b 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -113,38 +113,6 @@ static int __init early_init_dt_scan_cpus(unsigned long node, return 0; } -#ifdef CONFIG_BLK_DEV_INITRD -static void __init early_init_dt_check_for_initrd(unsigned long node) -{ - unsigned long l; - u32 *prop; - - pr_debug("Looking for initrd properties... "); - - prop = of_get_flat_dt_prop(node, "linux,initrd-start", &l); - if (prop) { - initrd_start = (unsigned long) - __va((u32)of_read_ulong(prop, l/4)); - - prop = of_get_flat_dt_prop(node, "linux,initrd-end", &l); - if (prop) { - initrd_end = (unsigned long) - __va((u32)of_read_ulong(prop, 1/4)); - initrd_below_start_ok = 1; - } else { - initrd_start = 0; - } - } - - pr_debug("initrd_start=0x%lx initrd_end=0x%lx\n", - initrd_start, initrd_end); -} -#else -static inline void early_init_dt_check_for_initrd(unsigned long node) -{ -} -#endif /* CONFIG_BLK_DEV_INITRD */ - static int __init early_init_dt_scan_chosen(unsigned long node, const char *uname, int depth, void *data) { diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 7f88566..1ecd6c6 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -373,36 +373,6 @@ static int __init early_init_dt_scan_cpus(unsigned long node, return 0; } -#ifdef CONFIG_BLK_DEV_INITRD -static void __init early_init_dt_check_for_initrd(unsigned long node) -{ - unsigned long l; - u32 *prop; - - DBG("Looking for initrd properties... "); - - prop = of_get_flat_dt_prop(node, "linux,initrd-start", &l); - if (prop) { - initrd_start = (unsigned long)__va(of_read_ulong(prop, l/4)); - - prop = of_get_flat_dt_prop(node, "linux,initrd-end", &l); - if (prop) { - initrd_end = (unsigned long) - __va(of_read_ulong(prop, l/4)); - initrd_below_start_ok = 1; - } else { - initrd_start = 0; - } - } - - DBG("initrd_start=0x%lx initrd_end=0x%lx\n", initrd_start, initrd_end); -} -#else -static inline void early_init_dt_check_for_initrd(unsigned long node) -{ -} -#endif /* CONFIG_BLK_DEV_INITRD */ - static int __init early_init_dt_scan_chosen(unsigned long node, const char *uname, int depth, void *data) { diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 43d236c..6ad98e8 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -11,6 +11,7 @@ #include #include +#include #include #include @@ -369,6 +370,42 @@ unsigned long __init unflatten_dt_node(unsigned long mem, return mem; } +#ifdef CONFIG_BLK_DEV_INITRD +/** + * early_init_dt_check_for_initrd - Decode initrd location from flat tree + * @node: reference to node containing initrd location ('chosen') + */ +void __init early_init_dt_check_for_initrd(unsigned long node) +{ + unsigned long len; + u32 *prop; + + pr_debug("Looking for initrd properties... "); + + prop = of_get_flat_dt_prop(node, "linux,initrd-start", &len); + if (prop) { + initrd_start = (unsigned long) + __va(of_read_ulong(prop, len/4)); + + prop = of_get_flat_dt_prop(node, "linux,initrd-end", &len); + if (prop) { + initrd_end = (unsigned long) + __va(of_read_ulong(prop, len/4)); + initrd_below_start_ok = 1; + } else { + initrd_start = 0; + } + } + + pr_debug("initrd_start=0x%lx initrd_end=0x%lx\n", + initrd_start, initrd_end); +} +#else +inline void early_init_dt_check_for_initrd(unsigned long node) +{ +} +#endif /* CONFIG_BLK_DEV_INITRD */ + /** * unflatten_device_tree - create tree of device_nodes from flat blob * diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index 81231e0..ec2db82 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -69,6 +69,7 @@ extern void *of_get_flat_dt_prop(unsigned long node, const char *name, unsigned long *size); extern int of_flat_dt_is_compatible(unsigned long node, const char *name); extern unsigned long of_get_flat_dt_root(void); +extern void early_init_dt_check_for_initrd(unsigned long node); /* Other Prototypes */ extern void finish_device_tree(void); -- cgit v0.10.2 From f00abd94918c9780f9d2d961fc0e419c11457922 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Tue, 24 Nov 2009 03:27:10 -0700 Subject: of/flattree: Merge earlyinit_dt_scan_root() Merge common code between PowerPC and Microblaze Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Tested-by: Wolfram Sang diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 7959495b..189179a 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -42,9 +42,6 @@ #include #include -static int __initdata dt_root_addr_cells; -static int __initdata dt_root_size_cells; - typedef u32 cell_t; /* export that to outside world */ @@ -158,26 +155,6 @@ static int __init early_init_dt_scan_chosen(unsigned long node, return 1; } -static int __init early_init_dt_scan_root(unsigned long node, - const char *uname, int depth, void *data) -{ - u32 *prop; - - if (depth != 0) - return 0; - - prop = of_get_flat_dt_prop(node, "#size-cells", NULL); - dt_root_size_cells = (prop == NULL) ? 1 : *prop; - pr_debug("dt_root_size_cells = %x\n", dt_root_size_cells); - - prop = of_get_flat_dt_prop(node, "#address-cells", NULL); - dt_root_addr_cells = (prop == NULL) ? 2 : *prop; - pr_debug("dt_root_addr_cells = %x\n", dt_root_addr_cells); - - /* break now */ - return 1; -} - static u64 __init dt_mem_next_cell(int s, cell_t **cellp) { cell_t *p = *cellp; diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 1ecd6c6..78f65a4 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -61,10 +61,6 @@ #define DBG(fmt...) #endif - -static int __initdata dt_root_addr_cells; -static int __initdata dt_root_size_cells; - #ifdef CONFIG_PPC64 int __initdata iommu_is_off; int __initdata iommu_force_on; @@ -436,26 +432,6 @@ static int __init early_init_dt_scan_chosen(unsigned long node, return 1; } -static int __init early_init_dt_scan_root(unsigned long node, - const char *uname, int depth, void *data) -{ - u32 *prop; - - if (depth != 0) - return 0; - - prop = of_get_flat_dt_prop(node, "#size-cells", NULL); - dt_root_size_cells = (prop == NULL) ? 1 : *prop; - DBG("dt_root_size_cells = %x\n", dt_root_size_cells); - - prop = of_get_flat_dt_prop(node, "#address-cells", NULL); - dt_root_addr_cells = (prop == NULL) ? 2 : *prop; - DBG("dt_root_addr_cells = %x\n", dt_root_addr_cells); - - /* break now */ - return 1; -} - static u64 __init dt_mem_next_cell(int s, cell_t **cellp) { cell_t *p = *cellp; diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 6ad98e8..be200be 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -15,6 +15,9 @@ #include #include +int __initdata dt_root_addr_cells; +int __initdata dt_root_size_cells; + struct boot_param_header *initial_boot_params; char *find_flat_dt_string(u32 offset) @@ -407,6 +410,29 @@ inline void early_init_dt_check_for_initrd(unsigned long node) #endif /* CONFIG_BLK_DEV_INITRD */ /** + * early_init_dt_scan_root - fetch the top level address and size cells + */ +int __init early_init_dt_scan_root(unsigned long node, const char *uname, + int depth, void *data) +{ + u32 *prop; + + if (depth != 0) + return 0; + + prop = of_get_flat_dt_prop(node, "#size-cells", NULL); + dt_root_size_cells = (prop == NULL) ? 1 : *prop; + pr_debug("dt_root_size_cells = %x\n", dt_root_size_cells); + + prop = of_get_flat_dt_prop(node, "#address-cells", NULL); + dt_root_addr_cells = (prop == NULL) ? 2 : *prop; + pr_debug("dt_root_addr_cells = %x\n", dt_root_addr_cells); + + /* break now */ + return 1; +} + +/** * unflatten_device_tree - create tree of device_nodes from flat blob * * unflattens the device-tree passed by the firmware, creating the diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index ec2db82..828c3cd 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -58,6 +58,8 @@ struct boot_param_header { }; /* TBD: Temporary export of fdt globals - remove when code fully merged */ +extern int __initdata dt_root_addr_cells; +extern int __initdata dt_root_size_cells; extern struct boot_param_header *initial_boot_params; /* For scanning the flat device-tree at boot time */ @@ -71,6 +73,10 @@ extern int of_flat_dt_is_compatible(unsigned long node, const char *name); extern unsigned long of_get_flat_dt_root(void); extern void early_init_dt_check_for_initrd(unsigned long node); +/* Early flat tree scan hooks */ +extern int early_init_dt_scan_root(unsigned long node, const char *uname, + int depth, void *data); + /* Other Prototypes */ extern void finish_device_tree(void); extern void unflatten_device_tree(void); -- cgit v0.10.2 From 83f7a06eb479e2aeb83536e77a2cb14cc2285e32 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Tue, 24 Nov 2009 03:37:56 -0700 Subject: of/flattree: merge dt_mem_next_cell Merge common code between PowerPC and Microblaze Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Tested-by: Wolfram Sang diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 189179a..e0f4c34 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -155,14 +155,6 @@ static int __init early_init_dt_scan_chosen(unsigned long node, return 1; } -static u64 __init dt_mem_next_cell(int s, cell_t **cellp) -{ - cell_t *p = *cellp; - - *cellp = p + s; - return of_read_number(p, s); -} - static int __init early_init_dt_scan_memory(unsigned long node, const char *uname, int depth, void *data) { diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 78f65a4..048e3a3 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -432,14 +432,6 @@ static int __init early_init_dt_scan_chosen(unsigned long node, return 1; } -static u64 __init dt_mem_next_cell(int s, cell_t **cellp) -{ - cell_t *p = *cellp; - - *cellp = p + s; - return of_read_number(p, s); -} - #ifdef CONFIG_PPC_PSERIES /* * Interpret the ibm,dynamic-memory property in the diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index be200be..ebce509 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -432,6 +432,14 @@ int __init early_init_dt_scan_root(unsigned long node, const char *uname, return 1; } +u64 __init dt_mem_next_cell(int s, u32 **cellp) +{ + u32 *p = *cellp; + + *cellp = p + s; + return of_read_number(p, s); +} + /** * unflatten_device_tree - create tree of device_nodes from flat blob * diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index 828c3cd..d1a37e5 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -72,6 +72,7 @@ extern void *of_get_flat_dt_prop(unsigned long node, const char *name, extern int of_flat_dt_is_compatible(unsigned long node, const char *name); extern unsigned long of_get_flat_dt_root(void); extern void early_init_dt_check_for_initrd(unsigned long node); +extern u64 dt_mem_next_cell(int s, u32 **cellp); /* Early flat tree scan hooks */ extern int early_init_dt_scan_root(unsigned long node, const char *uname, -- cgit v0.10.2 From 0f0b56c3f2df4a083fc9e934266e5bab1710e286 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 10 Dec 2009 23:42:17 -0700 Subject: of/flattree: eliminate cell_t typedef A cell is firmly established as a u32. No need to do an ugly typedef to redefine it to cell_t. Eliminate the unnecessary typedef so that it doesn't have to be added to the of_fdt header file Signed-off-by: Grant Likely Tested-by: Wolfram Sang Acked-by: Benjamin Herrenschmidt diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index e0f4c34..50d8b09 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -42,8 +42,6 @@ #include #include -typedef u32 cell_t; - /* export that to outside world */ struct device_node *of_chosen; @@ -159,7 +157,7 @@ static int __init early_init_dt_scan_memory(unsigned long node, const char *uname, int depth, void *data) { char *type = of_get_flat_dt_prop(node, "device_type", NULL); - cell_t *reg, *endp; + __be32 *reg, *endp; unsigned long l; /* Look for the ibm,dynamic-reconfiguration-memory node */ @@ -178,13 +176,13 @@ static int __init early_init_dt_scan_memory(unsigned long node, } else if (strcmp(type, "memory") != 0) return 0; - reg = (cell_t *)of_get_flat_dt_prop(node, "linux,usable-memory", &l); + reg = (__be32 *)of_get_flat_dt_prop(node, "linux,usable-memory", &l); if (reg == NULL) - reg = (cell_t *)of_get_flat_dt_prop(node, "reg", &l); + reg = (__be32 *)of_get_flat_dt_prop(node, "reg", &l); if (reg == NULL) return 0; - endp = reg + (l / sizeof(cell_t)); + endp = reg + (l / sizeof(__be32)); pr_debug("memory scan node %s, reg size %ld, data: %x %x %x %x,\n", uname, l, reg[0], reg[1], reg[2], reg[3]); diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 048e3a3..4e3181c 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -67,8 +67,6 @@ int __initdata iommu_force_on; unsigned long tce_alloc_start, tce_alloc_end; #endif -typedef u32 cell_t; - extern rwlock_t devtree_lock; /* temporary while merging */ /* export that to outside world */ @@ -441,22 +439,22 @@ static int __init early_init_dt_scan_chosen(unsigned long node, */ static int __init early_init_dt_scan_drconf_memory(unsigned long node) { - cell_t *dm, *ls, *usm; + __be32 *dm, *ls, *usm; unsigned long l, n, flags; u64 base, size, lmb_size; unsigned int is_kexec_kdump = 0, rngs; ls = of_get_flat_dt_prop(node, "ibm,lmb-size", &l); - if (ls == NULL || l < dt_root_size_cells * sizeof(cell_t)) + if (ls == NULL || l < dt_root_size_cells * sizeof(__be32)) return 0; lmb_size = dt_mem_next_cell(dt_root_size_cells, &ls); dm = of_get_flat_dt_prop(node, "ibm,dynamic-memory", &l); - if (dm == NULL || l < sizeof(cell_t)) + if (dm == NULL || l < sizeof(__be32)) return 0; n = *dm++; /* number of entries */ - if (l < (n * (dt_root_addr_cells + 4) + 1) * sizeof(cell_t)) + if (l < (n * (dt_root_addr_cells + 4) + 1) * sizeof(__be32)) return 0; /* check if this is a kexec/kdump kernel. */ @@ -515,7 +513,7 @@ static int __init early_init_dt_scan_memory(unsigned long node, const char *uname, int depth, void *data) { char *type = of_get_flat_dt_prop(node, "device_type", NULL); - cell_t *reg, *endp; + __be32 *reg, *endp; unsigned long l; /* Look for the ibm,dynamic-reconfiguration-memory node */ @@ -540,7 +538,7 @@ static int __init early_init_dt_scan_memory(unsigned long node, if (reg == NULL) return 0; - endp = reg + (l / sizeof(cell_t)); + endp = reg + (l / sizeof(__be32)); DBG("memory scan node %s, reg size %ld, data: %x %x %x %x,\n", uname, l, reg[0], reg[1], reg[2], reg[3]); -- cgit v0.10.2 From 86e032213424958b45564d0cc96b3316641a49d3 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 10 Dec 2009 23:42:21 -0700 Subject: of/flattree: merge early_init_dt_scan_chosen() Merge common code between PowerPC and Microblaze. This patch splits the arch-specific stuff out into a new function, early_init_dt_scan_chosen_arch(). Signed-off-by: Grant Likely Tested-by: Wolfram Sang Acked-by: Benjamin Herrenschmidt diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 50d8b09..5505bcf 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -108,49 +108,9 @@ static int __init early_init_dt_scan_cpus(unsigned long node, return 0; } -static int __init early_init_dt_scan_chosen(unsigned long node, - const char *uname, int depth, void *data) +void __init early_init_dt_scan_chosen_arch(unsigned long node) { - unsigned long l; - char *p; - - pr_debug("search \"chosen\", depth: %d, uname: %s\n", depth, uname); - - if (depth != 1 || - (strcmp(uname, "chosen") != 0 && - strcmp(uname, "chosen@0") != 0)) - return 0; - -#ifdef CONFIG_KEXEC - lprop = (u64 *)of_get_flat_dt_prop(node, - "linux,crashkernel-base", NULL); - if (lprop) - crashk_res.start = *lprop; - - lprop = (u64 *)of_get_flat_dt_prop(node, - "linux,crashkernel-size", NULL); - if (lprop) - crashk_res.end = crashk_res.start + *lprop - 1; -#endif - - early_init_dt_check_for_initrd(node); - - /* Retreive command line */ - p = of_get_flat_dt_prop(node, "bootargs", &l); - if (p != NULL && l > 0) - strlcpy(cmd_line, p, min((int)l, COMMAND_LINE_SIZE)); - -#ifdef CONFIG_CMDLINE -#ifndef CONFIG_CMDLINE_FORCE - if (p == NULL || l == 0 || (l == 1 && (*p) == 0)) -#endif - strlcpy(cmd_line, CONFIG_CMDLINE, COMMAND_LINE_SIZE); -#endif /* CONFIG_CMDLINE */ - - pr_debug("Command line is: %s\n", cmd_line); - - /* break now */ - return 1; + /* No Microblaze specific code here */ } static int __init early_init_dt_scan_memory(unsigned long node, diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 4e3181c..877fad9 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -367,18 +367,9 @@ static int __init early_init_dt_scan_cpus(unsigned long node, return 0; } -static int __init early_init_dt_scan_chosen(unsigned long node, - const char *uname, int depth, void *data) +void __init early_init_dt_scan_chosen_arch(unsigned long node) { unsigned long *lprop; - unsigned long l; - char *p; - - DBG("search \"chosen\", depth: %d, uname: %s\n", depth, uname); - - if (depth != 1 || - (strcmp(uname, "chosen") != 0 && strcmp(uname, "chosen@0") != 0)) - return 0; #ifdef CONFIG_PPC64 /* check if iommu is forced on or off */ @@ -389,17 +380,17 @@ static int __init early_init_dt_scan_chosen(unsigned long node, #endif /* mem=x on the command line is the preferred mechanism */ - lprop = of_get_flat_dt_prop(node, "linux,memory-limit", NULL); - if (lprop) - memory_limit = *lprop; + lprop = of_get_flat_dt_prop(node, "linux,memory-limit", NULL); + if (lprop) + memory_limit = *lprop; #ifdef CONFIG_PPC64 - lprop = of_get_flat_dt_prop(node, "linux,tce-alloc-start", NULL); - if (lprop) - tce_alloc_start = *lprop; - lprop = of_get_flat_dt_prop(node, "linux,tce-alloc-end", NULL); - if (lprop) - tce_alloc_end = *lprop; + lprop = of_get_flat_dt_prop(node, "linux,tce-alloc-start", NULL); + if (lprop) + tce_alloc_start = *lprop; + lprop = of_get_flat_dt_prop(node, "linux,tce-alloc-end", NULL); + if (lprop) + tce_alloc_end = *lprop; #endif #ifdef CONFIG_KEXEC @@ -411,23 +402,6 @@ static int __init early_init_dt_scan_chosen(unsigned long node, if (lprop) crashk_res.end = crashk_res.start + *lprop - 1; #endif - - early_init_dt_check_for_initrd(node); - - /* Retreive command line */ - p = of_get_flat_dt_prop(node, "bootargs", &l); - if (p != NULL && l > 0) - strlcpy(cmd_line, p, min((int)l, COMMAND_LINE_SIZE)); - -#ifdef CONFIG_CMDLINE - if (p == NULL || l == 0 || (l == 1 && (*p) == 0)) - strlcpy(cmd_line, CONFIG_CMDLINE, COMMAND_LINE_SIZE); -#endif /* CONFIG_CMDLINE */ - - DBG("Command line is: %s\n", cmd_line); - - /* break now */ - return 1; } #ifdef CONFIG_PPC_PSERIES diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index ebce509..616a476 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -15,6 +15,10 @@ #include #include +#ifdef CONFIG_PPC +#include +#endif /* CONFIG_PPC */ + int __initdata dt_root_addr_cells; int __initdata dt_root_size_cells; @@ -440,6 +444,40 @@ u64 __init dt_mem_next_cell(int s, u32 **cellp) return of_read_number(p, s); } +int __init early_init_dt_scan_chosen(unsigned long node, const char *uname, + int depth, void *data) +{ + unsigned long l; + char *p; + + pr_debug("search \"chosen\", depth: %d, uname: %s\n", depth, uname); + + if (depth != 1 || + (strcmp(uname, "chosen") != 0 && strcmp(uname, "chosen@0") != 0)) + return 0; + + early_init_dt_check_for_initrd(node); + + /* Retreive command line */ + p = of_get_flat_dt_prop(node, "bootargs", &l); + if (p != NULL && l > 0) + strlcpy(cmd_line, p, min((int)l, COMMAND_LINE_SIZE)); + +#ifdef CONFIG_CMDLINE +#ifndef CONFIG_CMDLINE_FORCE + if (p == NULL || l == 0 || (l == 1 && (*p) == 0)) +#endif + strlcpy(cmd_line, CONFIG_CMDLINE, COMMAND_LINE_SIZE); +#endif /* CONFIG_CMDLINE */ + + early_init_dt_scan_chosen_arch(node); + + pr_debug("Command line is: %s\n", cmd_line); + + /* break now */ + return 1; +} + /** * unflatten_device_tree - create tree of device_nodes from flat blob * diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index d1a37e5..8118d45 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -71,6 +71,9 @@ extern void *of_get_flat_dt_prop(unsigned long node, const char *name, unsigned long *size); extern int of_flat_dt_is_compatible(unsigned long node, const char *name); extern unsigned long of_get_flat_dt_root(void); +extern void early_init_dt_scan_chosen_arch(unsigned long node); +extern int early_init_dt_scan_chosen(unsigned long node, const char *uname, + int depth, void *data); extern void early_init_dt_check_for_initrd(unsigned long node); extern u64 dt_mem_next_cell(int s, u32 **cellp); -- cgit v0.10.2 From b73c806341cfc7492ede6a2ce713cb579547d0ab Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Wed, 25 Nov 2009 22:00:08 +0000 Subject: sh: Abstract the number of page table levels Keep the dimensions of the page tables in a separate header file in preparation for allowing a three level page table structure. Signed-off-by: Matt Fleming Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/pgalloc.h b/arch/sh/include/asm/pgalloc.h index 63ca37b..fe9f037 100644 --- a/arch/sh/include/asm/pgalloc.h +++ b/arch/sh/include/asm/pgalloc.h @@ -4,9 +4,10 @@ #include #include -#define QUICK_PGD 0 /* We preserve special mappings over free */ #define QUICK_PT 1 /* Other page table pages that are zero on free */ +#include + static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd, pte_t *pte) { @@ -20,28 +21,9 @@ static inline void pmd_populate(struct mm_struct *mm, pmd_t *pmd, } #define pmd_pgtable(pmd) pmd_page(pmd) -static inline void pgd_ctor(void *x) -{ - pgd_t *pgd = x; - - memcpy(pgd + USER_PTRS_PER_PGD, - swapper_pg_dir + USER_PTRS_PER_PGD, - (PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t)); -} - /* * Allocate and free page tables. */ -static inline pgd_t *pgd_alloc(struct mm_struct *mm) -{ - return quicklist_alloc(QUICK_PGD, GFP_KERNEL | __GFP_REPEAT, pgd_ctor); -} - -static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd) -{ - quicklist_free(QUICK_PGD, NULL, pgd); -} - static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address) { @@ -81,7 +63,7 @@ do { \ static inline void check_pgt_cache(void) { - quicklist_trim(QUICK_PGD, NULL, 25, 16); + __check_pgt_cache(); quicklist_trim(QUICK_PT, NULL, 25, 16); } diff --git a/arch/sh/include/asm/pgalloc_nopmd.h b/arch/sh/include/asm/pgalloc_nopmd.h new file mode 100644 index 0000000..e4b344c --- /dev/null +++ b/arch/sh/include/asm/pgalloc_nopmd.h @@ -0,0 +1,30 @@ +#ifndef __ASM_SH_PGALLOC_NOPMD_H +#define __ASM_SH_PGALLOC_NOPMD_H + +#define QUICK_PGD 0 /* We preserve special mappings over free */ + +static inline void pgd_ctor(void *x) +{ + pgd_t *pgd = x; + + memcpy(pgd + USER_PTRS_PER_PGD, + swapper_pg_dir + USER_PTRS_PER_PGD, + (PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t)); +} + +static inline pgd_t *pgd_alloc(struct mm_struct *mm) +{ + return quicklist_alloc(QUICK_PGD, GFP_KERNEL | __GFP_REPEAT, pgd_ctor); +} + +static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd) +{ + quicklist_free(QUICK_PGD, NULL, pgd); +} + +static inline void __check_pgt_cache(void) +{ + quicklist_trim(QUICK_PGD, NULL, 25, 16); +} + +#endif /* __ASM_SH_PGALLOC_NOPMD_H */ diff --git a/arch/sh/include/asm/pgtable.h b/arch/sh/include/asm/pgtable.h index ba3046e..9a0f66c 100644 --- a/arch/sh/include/asm/pgtable.h +++ b/arch/sh/include/asm/pgtable.h @@ -12,7 +12,7 @@ #ifndef __ASM_SH_PGTABLE_H #define __ASM_SH_PGTABLE_H -#include +#include #include #ifndef __ASSEMBLY__ @@ -51,28 +51,12 @@ static inline unsigned long long neff_sign_extend(unsigned long val) #define NPHYS_SIGN (1LL << (NPHYS - 1)) #define NPHYS_MASK (-1LL << NPHYS) -/* - * traditional two-level paging structure - */ -/* PTE bits */ -#if defined(CONFIG_X2TLB) || defined(CONFIG_SUPERH64) -# define PTE_MAGNITUDE 3 /* 64-bit PTEs on extended mode SH-X2 TLB */ -#else -# define PTE_MAGNITUDE 2 /* 32-bit PTEs */ -#endif -#define PTE_SHIFT PAGE_SHIFT -#define PTE_BITS (PTE_SHIFT - PTE_MAGNITUDE) - -/* PGD bits */ -#define PGDIR_SHIFT (PTE_SHIFT + PTE_BITS) #define PGDIR_SIZE (1UL << PGDIR_SHIFT) #define PGDIR_MASK (~(PGDIR_SIZE-1)) /* Entries per level */ #define PTRS_PER_PTE (PAGE_SIZE / (1 << PTE_MAGNITUDE)) -#define PTRS_PER_PGD (PAGE_SIZE / sizeof(pgd_t)) -#define USER_PTRS_PER_PGD (TASK_SIZE/PGDIR_SIZE) #define FIRST_USER_ADDRESS 0 #define PHYS_ADDR_MASK29 0x1fffffff diff --git a/arch/sh/include/asm/pgtable_nopmd.h b/arch/sh/include/asm/pgtable_nopmd.h new file mode 100644 index 0000000..f0b525b --- /dev/null +++ b/arch/sh/include/asm/pgtable_nopmd.h @@ -0,0 +1,22 @@ +#ifndef __ASM_SH_PGTABLE_NOPMD_H +#define __ASM_SH_PGTABLE_NOPMD_H + +#include + +/* + * traditional two-level paging structure + */ + +/* PTE bits */ +#define PTE_MAGNITUDE 2 /* 32-bit PTEs */ + +#define PTE_SHIFT PAGE_SHIFT +#define PTE_BITS (PTE_SHIFT - PTE_MAGNITUDE) + +/* PGD bits */ +#define PGDIR_SHIFT (PTE_SHIFT + PTE_BITS) + +#define PTRS_PER_PGD (PAGE_SIZE / (1 << PTE_MAGNITUDE)) +#define USER_PTRS_PER_PGD (TASK_SIZE/PGDIR_SIZE) + +#endif /* __ASM_SH_PGTABLE_NOPMD_H */ -- cgit v0.10.2 From 5d9b4b19f118abfb75e352841f7bf74580d7e427 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Sun, 13 Dec 2009 14:38:50 +0000 Subject: sh: Definitions for 3-level page table layout If using 64-bit PTEs and 4K pages then each page table has 512 entries (as opposed to 1024 entries with 32-bit PTEs). Unlike MIPS, SH follows the convention that all structures in the page table (pgd_t, pmd_t, pgprot_t, etc) must be the same size. Therefore, 64-bit PTEs require 64-bit PGD entries, etc. Using 2-levels of page tables and 64-bit PTEs it is only possible to map 1GB of virtual address space. In order to map all 4GB of virtual address space we need to adopt a 3-level page table layout. This actually works out better for CONFIG_SUPERH32 because we only waste 2 PGD entries on the P1 and P2 areas (which are untranslated) instead of 256. Signed-off-by: Matt Fleming Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/pgalloc.h b/arch/sh/include/asm/pgalloc.h index fe9f037..4ea2785 100644 --- a/arch/sh/include/asm/pgalloc.h +++ b/arch/sh/include/asm/pgalloc.h @@ -6,7 +6,11 @@ #define QUICK_PT 1 /* Other page table pages that are zero on free */ +#ifdef CONFIG_PGTABLE_LEVELS_3 +#include +#else #include +#endif static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd, pte_t *pte) diff --git a/arch/sh/include/asm/pgalloc_pmd.h b/arch/sh/include/asm/pgalloc_pmd.h new file mode 100644 index 0000000..20f75cc --- /dev/null +++ b/arch/sh/include/asm/pgalloc_pmd.h @@ -0,0 +1,41 @@ +#ifndef __ASM_SH_PGALLOC_PMD_H +#define __ASM_SH_PGALLOC_PMD_H + +static inline pgd_t *pgd_alloc(struct mm_struct *mm) +{ + pgd_t *pgd; + int i; + + pgd = kzalloc(sizeof(*pgd) * PTRS_PER_PGD, GFP_KERNEL | __GFP_REPEAT); + + for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) + pgd[i] = swapper_pg_dir[i]; + + return pgd; +} + +static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd) +{ + kfree(pgd); +} + +static inline void __check_pgt_cache(void) +{ +} + +static inline void pud_populate(struct mm_struct *mm, pud_t *pud, pmd_t *pmd) +{ + set_pud(pud, __pud((unsigned long)pmd)); +} + +static inline pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long address) +{ + return quicklist_alloc(QUICK_PT, GFP_KERNEL | __GFP_REPEAT, NULL); +} + +static inline void pmd_free(struct mm_struct *mm, pmd_t *pmd) +{ + quicklist_free(QUICK_PT, NULL, pmd); +} + +#endif /* __ASM_SH_PGALLOC_PMD_H */ diff --git a/arch/sh/include/asm/pgtable.h b/arch/sh/include/asm/pgtable.h index 9a0f66c..9effcc3 100644 --- a/arch/sh/include/asm/pgtable.h +++ b/arch/sh/include/asm/pgtable.h @@ -12,7 +12,11 @@ #ifndef __ASM_SH_PGTABLE_H #define __ASM_SH_PGTABLE_H +#ifdef CONFIG_PGTABLE_LEVELS_3 +#include +#else #include +#endif #include #ifndef __ASSEMBLY__ diff --git a/arch/sh/include/asm/pgtable_pmd.h b/arch/sh/include/asm/pgtable_pmd.h new file mode 100644 index 0000000..78dc36e --- /dev/null +++ b/arch/sh/include/asm/pgtable_pmd.h @@ -0,0 +1,55 @@ +#ifndef __ASM_SH_PGTABLE_PMD_H +#define __ASM_SH_PGTABLE_PMD_H + +#include + +/* + * Some cores need a 3-level page table layout, for example when using + * 64-bit PTEs and 4K pages. + */ + +#define PTE_MAGNITUDE 3 /* 64-bit PTEs on extended mode SH-X2 TLB */ + +/* PGD bits */ +#define PGDIR_SHIFT 30 + +#define PTRS_PER_PGD 4 +#define USER_PTRS_PER_PGD 2 + +/* PMD bits */ +#define PMD_SHIFT (PAGE_SHIFT + (PAGE_SHIFT - 3)) +#define PMD_SIZE (1UL << PMD_SHIFT) +#define PMD_MASK (~(PMD_SIZE-1)) + +#define PTRS_PER_PMD (PAGE_SIZE / sizeof(pmd_t)) + +#define pmd_ERROR(e) \ + printk("%s:%d: bad pmd %016llx.\n", __FILE__, __LINE__, pmd_val(e)) + +typedef struct { unsigned long long pmd; } pmd_t; +#define pmd_val(x) ((x).pmd) +#define __pmd(x) ((pmd_t) { (x) } ) + +static inline unsigned long pud_page_vaddr(pud_t pud) +{ + return pud_val(pud); +} + +#define pmd_index(address) (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1)) +static inline pmd_t *pmd_offset(pud_t *pud, unsigned long address) +{ + return (pmd_t *)pud_page_vaddr(*pud) + pmd_index(address); +} + +#define pud_none(x) (!pud_val(x)) +#define pud_present(x) (pud_val(x)) +#define pud_clear(xp) do { set_pud(xp, __pud(0)); } while (0) +#define pud_bad(x) (pud_val(x) & ~PAGE_MASK) + +/* + * (puds are folded into pgds so this doesn't get actually called, + * but the define is needed for a generic inline function.) + */ +#define set_pud(pudptr, pudval) do { *(pudptr) = (pudval); } while(0) + +#endif /* __ASM_SH_PGTABLE_PMD_H */ diff --git a/arch/sh/mm/Kconfig b/arch/sh/mm/Kconfig index 0e7ba8e..b3f6c1a 100644 --- a/arch/sh/mm/Kconfig +++ b/arch/sh/mm/Kconfig @@ -190,19 +190,37 @@ config ARCH_MEMORY_PROBE depends on MEMORY_HOTPLUG choice + prompt "Page table layout" + default PGTABLE_LEVELS_3 if X2TLB + default PGTABLE_LEVELS_2 + +config PGTABLE_LEVELS_2 + bool "2 Levels" + help + This is the default page table layout for all SuperH CPUs. + +config PGTABLE_LEVELS_3 + bool "3 Levels" + depends on X2TLB + help + This enables a 3 level page table structure. + +endchoice + +choice prompt "Kernel page size" default PAGE_SIZE_8KB if X2TLB default PAGE_SIZE_4KB config PAGE_SIZE_4KB bool "4kB" - depends on !MMU || !X2TLB + depends on !MMU || !X2TLB || PGTABLE_LEVELS_3 help This is the default page size used by all SuperH CPUs. config PAGE_SIZE_8KB bool "8kB" - depends on !MMU || X2TLB + depends on !MMU || X2TLB && !PGTABLE_LEVELS_3 help This enables 8kB pages as supported by SH-X2 and later MMUs. @@ -214,7 +232,7 @@ config PAGE_SIZE_16KB config PAGE_SIZE_64KB bool "64kB" - depends on !MMU || CPU_SH4 || CPU_SH5 + depends on !MMU || CPU_SH4 && !PGTABLE_LEVELS_3 || CPU_SH5 help This enables support for 64kB pages, possible on all SH-4 CPUs and later. diff --git a/arch/sh/mm/fault_32.c b/arch/sh/mm/fault_32.c index 4753010..28e2283 100644 --- a/arch/sh/mm/fault_32.c +++ b/arch/sh/mm/fault_32.c @@ -53,6 +53,9 @@ static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address) if (!pud_present(*pud_k)) return NULL; + if (!pud_present(*pud)) + set_pud(pud, *pud_k); + pmd = pmd_offset(pud, address); pmd_k = pmd_offset(pud_k, address); if (!pmd_present(*pmd_k)) diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c index 432acd0..761910d 100644 --- a/arch/sh/mm/init.c +++ b/arch/sh/mm/init.c @@ -120,7 +120,13 @@ void __init page_table_range_init(unsigned long start, unsigned long end, for ( ; (i < PTRS_PER_PGD) && (vaddr != end); pgd++, i++) { pud = (pud_t *)pgd; for ( ; (j < PTRS_PER_PUD) && (vaddr != end); pud++, j++) { +#ifdef __PAGETABLE_PMD_FOLDED pmd = (pmd_t *)pud; +#else + pmd = (pmd_t *)alloc_bootmem_low_pages(PAGE_SIZE); + pud_populate(&init_mm, pud, pmd); + pmd += k; +#endif for (; (k < PTRS_PER_PMD) && (vaddr != end); pmd++, k++) { if (pmd_none(*pmd)) { pte = (pte_t *) alloc_bootmem_low_pages(PAGE_SIZE); -- cgit v0.10.2 From b76365a18f7593c9df32a74bf2b4a39741b67bc6 Mon Sep 17 00:00:00 2001 From: Russ Anderson Date: Thu, 17 Dec 2009 10:53:25 -0600 Subject: x86, uv: Add serial number parameter to uv_bios_get_sn_info() Add system_serial_number to the information returned by uv_bios_get_sn_info() UV BIOS call. Signed-off-by: Russ Anderson LKML-Reference: <20091217165323.GA30774@sgi.com> Cc: Jack Steiner Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/uv/bios.h b/arch/x86/include/asm/uv/bios.h index 2751f30..3fbc1f3 100644 --- a/arch/x86/include/asm/uv/bios.h +++ b/arch/x86/include/asm/uv/bios.h @@ -18,8 +18,8 @@ * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA * - * Copyright (c) 2008 Silicon Graphics, Inc. All Rights Reserved. - * Copyright (c) Russ Anderson + * Copyright (c) 2008-2009 Silicon Graphics, Inc. All Rights Reserved. + * Copyright (c) Russ Anderson */ #include @@ -89,7 +89,7 @@ extern s64 uv_bios_call(enum uv_bios_cmd, u64, u64, u64, u64, u64); extern s64 uv_bios_call_irqsave(enum uv_bios_cmd, u64, u64, u64, u64, u64); extern s64 uv_bios_call_reentrant(enum uv_bios_cmd, u64, u64, u64, u64, u64); -extern s64 uv_bios_get_sn_info(int, int *, long *, long *, long *); +extern s64 uv_bios_get_sn_info(int, int *, long *, long *, long *, long *); extern s64 uv_bios_freq_base(u64, u64 *); extern int uv_bios_mq_watchlist_alloc(unsigned long, unsigned int, unsigned long *); @@ -104,6 +104,7 @@ extern int uv_type; extern long sn_partition_id; extern long sn_coherency_id; extern long sn_region_size; +extern long system_serial_number; #define partition_coherence_id() (sn_coherency_id) extern struct kobject *sgi_uv_kobj; /* /sys/firmware/sgi_uv */ diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index b684bb3..af5d103 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -5,7 +5,7 @@ * * SGI UV APIC functions (note: not an Intel compatible APIC) * - * Copyright (C) 2007-2008 Silicon Graphics, Inc. All rights reserved. + * Copyright (C) 2007-2009 Silicon Graphics, Inc. All rights reserved. */ #include #include @@ -627,8 +627,8 @@ void __init uv_system_init(void) } uv_bios_init(); - uv_bios_get_sn_info(0, &uv_type, &sn_partition_id, - &sn_coherency_id, &sn_region_size); + uv_bios_get_sn_info(0, &uv_type, &sn_partition_id, &sn_coherency_id, + &sn_region_size, &system_serial_number); uv_rtc_init(); for_each_present_cpu(cpu) { diff --git a/arch/x86/kernel/bios_uv.c b/arch/x86/kernel/bios_uv.c index b0206a2..c918eba 100644 --- a/arch/x86/kernel/bios_uv.c +++ b/arch/x86/kernel/bios_uv.c @@ -15,8 +15,8 @@ * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA * - * Copyright (c) 2008 Silicon Graphics, Inc. All Rights Reserved. - * Copyright (c) Russ Anderson + * Copyright (c) 2008-2009 Silicon Graphics, Inc. All Rights Reserved. + * Copyright (c) Russ Anderson */ #include @@ -30,6 +30,7 @@ static struct uv_systab uv_systab; s64 uv_bios_call(enum uv_bios_cmd which, u64 a1, u64 a2, u64 a3, u64 a4, u64 a5) { struct uv_systab *tab = &uv_systab; + s64 ret; if (!tab->function) /* @@ -37,9 +38,11 @@ s64 uv_bios_call(enum uv_bios_cmd which, u64 a1, u64 a2, u64 a3, u64 a4, u64 a5) */ return BIOS_STATUS_UNIMPLEMENTED; - return efi_call6((void *)__va(tab->function), - (u64)which, a1, a2, a3, a4, a5); + ret = efi_call6((void *)__va(tab->function), (u64)which, + a1, a2, a3, a4, a5); + return ret; } +EXPORT_SYMBOL_GPL(uv_bios_call); s64 uv_bios_call_irqsave(enum uv_bios_cmd which, u64 a1, u64 a2, u64 a3, u64 a4, u64 a5) @@ -73,11 +76,14 @@ long sn_coherency_id; EXPORT_SYMBOL_GPL(sn_coherency_id); long sn_region_size; EXPORT_SYMBOL_GPL(sn_region_size); +long system_serial_number; +EXPORT_SYMBOL_GPL(system_serial_number); int uv_type; +EXPORT_SYMBOL_GPL(uv_type); s64 uv_bios_get_sn_info(int fc, int *uvtype, long *partid, long *coher, - long *region) + long *region, long *ssn) { s64 ret; u64 v0, v1; @@ -97,8 +103,11 @@ s64 uv_bios_get_sn_info(int fc, int *uvtype, long *partid, long *coher, *coher = part.coherence_id; if (region) *region = part.region_size; + if (ssn) + *ssn = v1; return ret; } +EXPORT_SYMBOL_GPL(uv_bios_get_sn_info); int uv_bios_mq_watchlist_alloc(unsigned long addr, unsigned int mq_size, @@ -185,4 +194,3 @@ void uv_bios_init(void) void uv_bios_init(void) { } #endif - -- cgit v0.10.2 From bc85b25e5de17d714e8001cb3dc0feb66eac2750 Mon Sep 17 00:00:00 2001 From: Alessandro Rubini Date: Sat, 19 Dec 2009 19:45:43 +0800 Subject: hwrng: nomadik - Add hardware RNG driver The hardware random number generator by ST is used in both the Nomadik 8815 SoC and the U8500. It returns 16 bits every 400ns with automatic delay if a read is issued too early. It depends on PLAT_NOMADIK. Signed-off-by: Alessandro Rubini Acked-by: Andrea Gallo Acked-by: Matt Mackall Signed-off-by: Herbert Xu diff --git a/drivers/char/hw_random/Kconfig b/drivers/char/hw_random/Kconfig index 8706026..6ea1014 100644 --- a/drivers/char/hw_random/Kconfig +++ b/drivers/char/hw_random/Kconfig @@ -186,3 +186,15 @@ config HW_RANDOM_MXC_RNGA module will be called mxc-rnga. If unsure, say Y. + +config HW_RANDOM_NOMADIK + tristate "ST-Ericsson Nomadik Random Number Generator support" + depends on HW_RANDOM && PLAT_NOMADIK + ---help--- + This driver provides kernel-side support for the Random Number + Generator hardware found on ST-Ericsson SoCs (8815 and 8500). + + To compile this driver as a module, choose M here: the + module will be called nomadik-rng. + + If unsure, say Y. diff --git a/drivers/char/hw_random/Makefile b/drivers/char/hw_random/Makefile index 5eeb130..4273308 100644 --- a/drivers/char/hw_random/Makefile +++ b/drivers/char/hw_random/Makefile @@ -18,3 +18,4 @@ obj-$(CONFIG_HW_RANDOM_VIRTIO) += virtio-rng.o obj-$(CONFIG_HW_RANDOM_TX4939) += tx4939-rng.o obj-$(CONFIG_HW_RANDOM_MXC_RNGA) += mxc-rnga.o obj-$(CONFIG_HW_RANDOM_OCTEON) += octeon-rng.o +obj-$(CONFIG_HW_RANDOM_NOMADIK) += nomadik-rng.o diff --git a/drivers/char/hw_random/nomadik-rng.c b/drivers/char/hw_random/nomadik-rng.c new file mode 100644 index 0000000..a8b4c40 --- /dev/null +++ b/drivers/char/hw_random/nomadik-rng.c @@ -0,0 +1,103 @@ +/* + * Nomadik RNG support + * Copyright 2009 Alessandro Rubini + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + */ + +#include +#include +#include +#include +#include +#include +#include + +static int nmk_rng_read(struct hwrng *rng, void *data, size_t max, bool wait) +{ + void __iomem *base = (void __iomem *)rng->priv; + + /* + * The register is 32 bits and gives 16 random bits (low half). + * A subsequent read will delay the core for 400ns, so we just read + * once and accept the very unlikely very small delay, even if wait==0. + */ + *(u16 *)data = __raw_readl(base + 8) & 0xffff; + return 2; +} + +/* we have at most one RNG per machine, granted */ +static struct hwrng nmk_rng = { + .name = "nomadik", + .read = nmk_rng_read, +}; + +static int nmk_rng_probe(struct amba_device *dev, struct amba_id *id) +{ + void __iomem *base; + int ret; + + ret = amba_request_regions(dev, dev->dev.init_name); + if (ret) + return ret; + ret = -ENOMEM; + base = ioremap(dev->res.start, resource_size(&dev->res)); + if (!base) + goto out_release; + nmk_rng.priv = (unsigned long)base; + ret = hwrng_register(&nmk_rng); + if (ret) + goto out_unmap; + return 0; + +out_unmap: + iounmap(base); +out_release: + amba_release_regions(dev); + return ret; +} + +static int nmk_rng_remove(struct amba_device *dev) +{ + void __iomem *base = (void __iomem *)nmk_rng.priv; + hwrng_unregister(&nmk_rng); + iounmap(base); + amba_release_regions(dev); + return 0; +} + +static struct amba_id nmk_rng_ids[] = { + { + .id = 0x000805e1, + .mask = 0x000fffff, /* top bits are rev and cfg: accept all */ + }, + {0, 0}, +}; + +static struct amba_driver nmk_rng_driver = { + .drv = { + .owner = THIS_MODULE, + .name = "rng", + }, + .probe = nmk_rng_probe, + .remove = nmk_rng_remove, + .id_table = nmk_rng_ids, +}; + +static int __init nmk_rng_init(void) +{ + return amba_driver_register(&nmk_rng_driver); +} + +static void __devexit nmk_rng_exit(void) +{ + amba_driver_unregister(&nmk_rng_driver); +} + +module_init(nmk_rng_init); +module_exit(nmk_rng_exit); + +MODULE_LICENSE("GPL"); -- cgit v0.10.2 From d2392ba06a477cf1876ce3d505b7cd77e295cfd4 Mon Sep 17 00:00:00 2001 From: Alessandro Rubini Date: Sat, 19 Dec 2009 19:46:33 +0800 Subject: hwrng: nomadik - Add platform device for rng in cpu-8815 Signed-off-by: Alessandro Rubini Acked-by: Andrea Gallo Acked-by: Matt Mackall Signed-off-by: Herbert Xu diff --git a/arch/arm/mach-nomadik/cpu-8815.c b/arch/arm/mach-nomadik/cpu-8815.c index f93c596..9bf33b3 100644 --- a/arch/arm/mach-nomadik/cpu-8815.c +++ b/arch/arm/mach-nomadik/cpu-8815.c @@ -86,11 +86,19 @@ static struct amba_device cpu8815_amba_gpio[] = { }, }; +static struct amba_device cpu8815_amba_rng = { + .dev = { + .init_name = "rng", + }, + __MEM_4K_RESOURCE(NOMADIK_RNG_BASE), +}; + static struct amba_device *amba_devs[] __initdata = { cpu8815_amba_gpio + 0, cpu8815_amba_gpio + 1, cpu8815_amba_gpio + 2, cpu8815_amba_gpio + 3, + &cpu8815_amba_rng }; static int __init cpu8815_init(void) -- cgit v0.10.2 From a28b460ec7968f430a2d6ea2809b249ee9fe1d28 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 21 Dec 2009 15:56:24 +0900 Subject: sh: Fix up hw-breakpoints build for API changes. The event callback handling has been removed in favour of going through a generic event handler to handle overflows. Follows the x86 change. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/hw_breakpoint.c b/arch/sh/kernel/hw_breakpoint.c index ff3cb3d..75b8c47 100644 --- a/arch/sh/kernel/hw_breakpoint.c +++ b/arch/sh/kernel/hw_breakpoint.c @@ -302,8 +302,7 @@ int arch_validate_hwbkpt_settings(struct perf_event *bp, return ret; } - if (bp->callback) - ret = arch_store_info(bp); + ret = arch_store_info(bp); if (ret < 0) return ret; @@ -368,7 +367,7 @@ static int __kprobes hw_breakpoint_handler(struct die_args *args) break; } - (bp->callback)(bp, args->regs); + perf_bp_event(bp, args->regs); rcu_read_unlock(); } -- cgit v0.10.2 From 6c0ee6a4bba0e21eb0fb4de5bdac03c11b74242d Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 21 Dec 2009 17:35:00 +0900 Subject: sh: hw-breakpoint support requires perf events. This follows the x86 change to select perf events when hw_breakpoint support is enabled. This fixes up build issues where perf events can otherwise be disabled on their own. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 185e9af..fa76d62 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -43,6 +43,7 @@ config SUPERH32 select HAVE_FUNCTION_GRAPH_TRACER select HAVE_ARCH_KGDB select HAVE_HW_BREAKPOINT if CPU_SH4A + select PERF_EVENTS if HAVE_HW_BREAKPOINT select ARCH_HIBERNATION_POSSIBLE if MMU config SUPERH64 -- cgit v0.10.2 From b74ab703b1326aafadb5604ff6200da62c61df49 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 22 Dec 2009 12:40:38 +0900 Subject: sh: Don't NOTIFY_STOP for non-UBC breakpoint traps. Presently the hw_breakpoint code is the primary notifier dispatch for breakpoint traps, although it's only UBC traps that are of particular interest there. This patches in a check to allow non-UBC generated breakpoints to pass through down the remainder of the notifier chain, giving things like kgdb a chance at getting notified. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/hw_breakpoint.c b/arch/sh/kernel/hw_breakpoint.c index 75b8c47..022d8ed 100644 --- a/arch/sh/kernel/hw_breakpoint.c +++ b/arch/sh/kernel/hw_breakpoint.c @@ -398,9 +398,18 @@ BUILD_TRAP_HANDLER(breakpoint) int __kprobes hw_breakpoint_exceptions_notify(struct notifier_block *unused, unsigned long val, void *data) { + struct die_args *args = data; + if (val != DIE_BREAKPOINT) return NOTIFY_DONE; + /* + * If the breakpoint hasn't been triggered by the UBC, it's + * probably from a debugger, so don't do anything more here. + */ + if (args->trapnr != 0x1e0) + return NOTIFY_DONE; + return hw_breakpoint_handler(data); } -- cgit v0.10.2 From 22648735405f73299b717bb5933767e9a9c335ca Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 22 Dec 2009 12:44:14 +0900 Subject: sh: kgdb: Rework breakpoint handling on top of notifier chain. This kills off kgdb's breakpoint handler and ties in to the notifier chain instead. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/debugtraps.S b/arch/sh/kernel/debugtraps.S index 5917413..7a1b46f 100644 --- a/arch/sh/kernel/debugtraps.S +++ b/arch/sh/kernel/debugtraps.S @@ -13,7 +13,6 @@ #include #if !defined(CONFIG_KGDB) -#define breakpoint_trap_handler debug_trap_handler #define singlestep_trap_handler debug_trap_handler #endif diff --git a/arch/sh/kernel/kgdb.c b/arch/sh/kernel/kgdb.c index 3e532d0..70c6965 100644 --- a/arch/sh/kernel/kgdb.c +++ b/arch/sh/kernel/kgdb.c @@ -1,7 +1,7 @@ /* * SuperH KGDB support * - * Copyright (C) 2008 Paul Mundt + * Copyright (C) 2008 - 2009 Paul Mundt * * Single stepping taken from the old stub by Henry Bell and Jeremy Siegel. * @@ -251,24 +251,60 @@ BUILD_TRAP_HANDLER(singlestep) local_irq_restore(flags); } +static int __kgdb_notify(struct die_args *args, unsigned long cmd) +{ + int ret; + + switch (cmd) { + case DIE_BREAKPOINT: + /* + * This means a user thread is single stepping + * a system call which should be ignored + */ + if (test_thread_flag(TIF_SINGLESTEP)) + return NOTIFY_DONE; + + ret = kgdb_handle_exception(args->trapnr & 0xff, args->signr, + args->err, args->regs); + if (ret) + return NOTIFY_DONE; + + break; + } -BUILD_TRAP_HANDLER(breakpoint) + return NOTIFY_STOP; +} + +static int +kgdb_notify(struct notifier_block *self, unsigned long cmd, void *ptr) { unsigned long flags; - TRAP_HANDLER_DECL; + int ret; local_irq_save(flags); - kgdb_handle_exception(vec >> 2, SIGTRAP, 0, regs); + ret = __kgdb_notify(ptr, cmd); local_irq_restore(flags); + + return ret; } +static struct notifier_block kgdb_notifier = { + .notifier_call = kgdb_notify, + + /* + * Lowest-prio notifier priority, we want to be notified last: + */ + .priority = -INT_MAX, +}; + int kgdb_arch_init(void) { - return 0; + return register_die_notifier(&kgdb_notifier); } void kgdb_arch_exit(void) { + unregister_die_notifier(&kgdb_notifier); } struct kgdb_arch arch_kgdb_ops = { -- cgit v0.10.2 From b6353f4f36f03a12edaf3fa5365b475a28106035 Mon Sep 17 00:00:00 2001 From: Stephane Chatty Date: Tue, 22 Dec 2009 23:04:17 +0100 Subject: HID: Support for 3M multitouch panel Add support for 3M multitouch panels. Signed-off-by: Stephane Chatty [jkosina@suse.cz: fix build failure because of inconsistent 3M/MMM defines] Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index 24d90ea..a05133f 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -55,6 +55,13 @@ source "drivers/hid/usbhid/Kconfig" menu "Special HID drivers" depends on HID +config HID_3M_PCT + tristate "3M PCT" if EMBEDDED + depends on USB_HID + default !EMBEDDED + ---help--- + Support for 3M PCT touch screens. + config HID_A4TECH tristate "A4 tech" if EMBEDDED depends on USB_HID diff --git a/drivers/hid/Makefile b/drivers/hid/Makefile index 0de2dff..1bdfb97 100644 --- a/drivers/hid/Makefile +++ b/drivers/hid/Makefile @@ -19,6 +19,7 @@ ifdef CONFIG_LOGIRUMBLEPAD2_FF hid-logitech-objs += hid-lg2ff.o endif +obj-$(CONFIG_HID_3M_PCT) += hid-3m-pct.o obj-$(CONFIG_HID_A4TECH) += hid-a4tech.o obj-$(CONFIG_HID_APPLE) += hid-apple.o obj-$(CONFIG_HID_BELKIN) += hid-belkin.o diff --git a/drivers/hid/hid-3m-pct.c b/drivers/hid/hid-3m-pct.c new file mode 100644 index 0000000..6d11e3d --- /dev/null +++ b/drivers/hid/hid-3m-pct.c @@ -0,0 +1,291 @@ +/* + * HID driver for 3M PCT multitouch panels + * + * Copyright (c) 2009 Stephane Chatty + * + */ + +/* + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include +#include +#include +#include + +MODULE_VERSION("0.6"); +MODULE_AUTHOR("Stephane Chatty "); +MODULE_DESCRIPTION("3M PCT multitouch panels"); +MODULE_LICENSE("GPL"); + +#include "hid-ids.h" + +struct mmm_finger { + __s32 x, y; + __u8 rank; + bool touch, valid; +}; + +struct mmm_data { + struct mmm_finger f[10]; + __u8 curid, num; + bool touch, valid; +}; + +static int mmm_input_mapping(struct hid_device *hdev, struct hid_input *hi, + struct hid_field *field, struct hid_usage *usage, + unsigned long **bit, int *max) +{ + switch (usage->hid & HID_USAGE_PAGE) { + + case HID_UP_BUTTON: + return -1; + + case HID_UP_GENDESK: + switch (usage->hid) { + case HID_GD_X: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_POSITION_X); + /* touchscreen emulation */ + input_set_abs_params(hi->input, ABS_X, + field->logical_minimum, + field->logical_maximum, 0, 0); + return 1; + case HID_GD_Y: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_POSITION_Y); + /* touchscreen emulation */ + input_set_abs_params(hi->input, ABS_Y, + field->logical_minimum, + field->logical_maximum, 0, 0); + return 1; + } + return 0; + + case HID_UP_DIGITIZER: + switch (usage->hid) { + /* we do not want to map these: no input-oriented meaning */ + case 0x14: + case 0x23: + case HID_DG_INPUTMODE: + case HID_DG_DEVICEINDEX: + case HID_DG_CONTACTCOUNT: + case HID_DG_CONTACTMAX: + case HID_DG_INRANGE: + case HID_DG_CONFIDENCE: + return -1; + case HID_DG_TIPSWITCH: + /* touchscreen emulation */ + hid_map_usage(hi, usage, bit, max, EV_KEY, BTN_TOUCH); + return 1; + case HID_DG_CONTACTID: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_TRACKING_ID); + return 1; + } + /* let hid-input decide for the others */ + return 0; + + case 0xff000000: + /* we do not want to map these: no input-oriented meaning */ + return -1; + } + + return 0; +} + +static int mmm_input_mapped(struct hid_device *hdev, struct hid_input *hi, + struct hid_field *field, struct hid_usage *usage, + unsigned long **bit, int *max) +{ + if (usage->type == EV_KEY || usage->type == EV_ABS) + clear_bit(usage->code, *bit); + + return 0; +} + +/* + * this function is called when a whole packet has been received and processed, + * so that it can decide what to send to the input layer. + */ +static void mmm_filter_event(struct mmm_data *md, struct input_dev *input) +{ + struct mmm_finger *oldest = 0; + bool pressed = false, released = false; + int i; + + /* + * we need to iterate on all fingers to decide if we have a press + * or a release event in our touchscreen emulation. + */ + for (i = 0; i < 10; ++i) { + struct mmm_finger *f = &md->f[i]; + if (!f->valid) { + /* this finger is just placeholder data, ignore */ + } else if (f->touch) { + /* this finger is on the screen */ + input_event(input, EV_ABS, ABS_MT_TRACKING_ID, i); + input_event(input, EV_ABS, ABS_MT_POSITION_X, f->x); + input_event(input, EV_ABS, ABS_MT_POSITION_Y, f->y); + input_mt_sync(input); + /* + * touchscreen emulation: maintain the age rank + * of this finger, decide if we have a press + */ + if (f->rank == 0) { + f->rank = ++(md->num); + if (f->rank == 1) + pressed = true; + } + if (f->rank == 1) + oldest = f; + } else { + /* this finger took off the screen */ + /* touchscreen emulation: maintain age rank of others */ + int j; + + for (j = 0; j < 10; ++j) { + struct mmm_finger *g = &md->f[j]; + if (g->rank > f->rank) { + g->rank--; + if (g->rank == 1) + oldest = g; + } + } + f->rank = 0; + --(md->num); + if (md->num == 0) + released = true; + } + f->valid = 0; + } + + /* touchscreen emulation */ + if (oldest) { + if (pressed) + input_event(input, EV_KEY, BTN_TOUCH, 1); + input_event(input, EV_ABS, ABS_X, oldest->x); + input_event(input, EV_ABS, ABS_Y, oldest->y); + } else if (released) { + input_event(input, EV_KEY, BTN_TOUCH, 0); + } +} + +/* + * this function is called upon all reports + * so that we can accumulate contact point information, + * and call input_mt_sync after each point. + */ +static int mmm_event(struct hid_device *hid, struct hid_field *field, + struct hid_usage *usage, __s32 value) +{ + struct mmm_data *md = hid_get_drvdata(hid); + /* + * strangely, this function can be called before + * field->hidinput is initialized! + */ + if (hid->claimed & HID_CLAIMED_INPUT) { + struct input_dev *input = field->hidinput->input; + switch (usage->hid) { + case HID_DG_TIPSWITCH: + md->touch = value; + break; + case HID_DG_CONFIDENCE: + md->valid = value; + break; + case HID_DG_CONTACTID: + if (md->valid) { + md->curid = value; + md->f[value].touch = md->touch; + md->f[value].valid = 1; + } + break; + case HID_GD_X: + if (md->valid) + md->f[md->curid].x = value; + break; + case HID_GD_Y: + if (md->valid) + md->f[md->curid].y = value; + break; + case HID_DG_CONTACTCOUNT: + mmm_filter_event(md, input); + break; + } + } + + /* we have handled the hidinput part, now remains hiddev */ + if (hid->claimed & HID_CLAIMED_HIDDEV && hid->hiddev_hid_event) + hid->hiddev_hid_event(hid, field, usage, value); + + return 1; +} + +static int mmm_probe(struct hid_device *hdev, const struct hid_device_id *id) +{ + int ret; + struct mmm_data *md; + + md = kzalloc(sizeof(struct mmm_data), GFP_KERNEL); + if (!md) { + dev_err(&hdev->dev, "cannot allocate 3M data\n"); + return -ENOMEM; + } + hid_set_drvdata(hdev, md); + + ret = hid_parse(hdev); + if (!ret) + ret = hid_hw_start(hdev, HID_CONNECT_DEFAULT); + + if (ret) + kfree(md); + return ret; +} + +static void mmm_remove(struct hid_device *hdev) +{ + hid_hw_stop(hdev); + kfree(hid_get_drvdata(hdev)); + hid_set_drvdata(hdev, NULL); +} + +static const struct hid_device_id mmm_devices[] = { + { HID_USB_DEVICE(USB_VENDOR_ID_3M, USB_DEVICE_ID_3M1968) }, + { } +}; +MODULE_DEVICE_TABLE(hid, mmm_devices); + +static const struct hid_usage_id mmm_grabbed_usages[] = { + { HID_ANY_ID, HID_ANY_ID, HID_ANY_ID }, + { HID_ANY_ID - 1, HID_ANY_ID - 1, HID_ANY_ID - 1} +}; + +static struct hid_driver mmm_driver = { + .name = "3m-pct", + .id_table = mmm_devices, + .probe = mmm_probe, + .remove = mmm_remove, + .input_mapping = mmm_input_mapping, + .input_mapped = mmm_input_mapped, + .usage_table = mmm_grabbed_usages, + .event = mmm_event, +}; + +static int __init mmm_init(void) +{ + return hid_register_driver(&mmm_driver); +} + +static void __exit mmm_exit(void) +{ + hid_unregister_driver(&mmm_driver); +} + +module_init(mmm_init); +module_exit(mmm_exit); +MODULE_LICENSE("GPL"); + diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index 80792d3..70c25a0 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1248,6 +1248,7 @@ EXPORT_SYMBOL_GPL(hid_disconnect); /* a list of devices for which there is a specialized driver on HID bus */ static const struct hid_device_id hid_blacklist[] = { + { HID_USB_DEVICE(USB_VENDOR_ID_3M, USB_DEVICE_ID_3M1968) }, { HID_USB_DEVICE(USB_VENDOR_ID_A4TECH, USB_DEVICE_ID_A4TECH_WCP32PU) }, { HID_USB_DEVICE(USB_VENDOR_ID_A4TECH, USB_DEVICE_ID_A4TECH_X5_005D) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_ATV_IRCONTROL) }, diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index 3839340..3d55f98 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -18,6 +18,9 @@ #ifndef HID_IDS_H_FILE #define HID_IDS_H_FILE +#define USB_VENDOR_ID_3M 0x0596 +#define USB_DEVICE_ID_3M1968 0x0500 + #define USB_VENDOR_ID_A4TECH 0x09da #define USB_DEVICE_ID_A4TECH_WCP32PU 0x0006 #define USB_DEVICE_ID_A4TECH_X5_005D 0x000a -- cgit v0.10.2 From 863b557a88f8c033f7419fabafef4712a5055f85 Mon Sep 17 00:00:00 2001 From: "Youquan, Song" Date: Wed, 23 Dec 2009 19:45:20 +0800 Subject: crypto: testmgr - Fix complain about lack test for internal used algorithm When load aesni-intel and ghash_clmulni-intel driver,kernel will complain no test for some internal used algorithm. The strange information as following: alg: No test for __aes-aesni (__driver-aes-aesni) alg: No test for __ecb-aes-aesni (__driver-ecb-aes-aesni) alg: No test for __cbc-aes-aesni (__driver-cbc-aes-aesni) alg: No test for __ecb-aes-aesni (cryptd(__driver-ecb-aes-aesni) alg: No test for __ghash (__ghash-pclmulqdqni) alg: No test for __ghash (cryptd(__ghash-pclmulqdqni)) This patch add NULL test entries for these algorithm and driver. Signed-off-by: Youquan, Song Signed-off-by: Ying, Huang Signed-off-by: Herbert Xu diff --git a/crypto/testmgr.c b/crypto/testmgr.c index 7620bfc..c494d76 100644 --- a/crypto/testmgr.c +++ b/crypto/testmgr.c @@ -1477,9 +1477,54 @@ static int alg_test_cprng(const struct alg_test_desc *desc, const char *driver, return err; } +static int alg_test_null(const struct alg_test_desc *desc, + const char *driver, u32 type, u32 mask) +{ + return 0; +} + /* Please keep this list sorted by algorithm name. */ static const struct alg_test_desc alg_test_descs[] = { { + .alg = "__driver-cbc-aes-aesni", + .test = alg_test_null, + .suite = { + .cipher = { + .enc = { + .vecs = NULL, + .count = 0 + }, + .dec = { + .vecs = NULL, + .count = 0 + } + } + } + }, { + .alg = "__driver-ecb-aes-aesni", + .test = alg_test_null, + .suite = { + .cipher = { + .enc = { + .vecs = NULL, + .count = 0 + }, + .dec = { + .vecs = NULL, + .count = 0 + } + } + } + }, { + .alg = "__ghash-pclmulqdqni", + .test = alg_test_null, + .suite = { + .hash = { + .vecs = NULL, + .count = 0 + } + } + }, { .alg = "ansi_cprng", .test = alg_test_cprng, .fips_allowed = 1, @@ -1623,6 +1668,30 @@ static const struct alg_test_desc alg_test_descs[] = { } } }, { + .alg = "cryptd(__driver-ecb-aes-aesni)", + .test = alg_test_null, + .suite = { + .cipher = { + .enc = { + .vecs = NULL, + .count = 0 + }, + .dec = { + .vecs = NULL, + .count = 0 + } + } + } + }, { + .alg = "cryptd(__ghash-pclmulqdqni)", + .test = alg_test_null, + .suite = { + .hash = { + .vecs = NULL, + .count = 0 + } + } + }, { .alg = "ctr(aes)", .test = alg_test_skcipher, .fips_allowed = 1, @@ -1669,6 +1738,21 @@ static const struct alg_test_desc alg_test_descs[] = { } } }, { + .alg = "ecb(__aes-aesni)", + .test = alg_test_null, + .suite = { + .cipher = { + .enc = { + .vecs = NULL, + .count = 0 + }, + .dec = { + .vecs = NULL, + .count = 0 + } + } + } + }, { .alg = "ecb(aes)", .test = alg_test_skcipher, .fips_allowed = 1, -- cgit v0.10.2 From 4b186f72033611c2b526c7341534e71ee4afd222 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Wed, 23 Dec 2009 13:12:32 +0100 Subject: HID: make 3M PCT touchscreen driver standalone config option The point behind 'default !EMBEDDED' for certain HID drivers that simple and straightforward quirks for HID devices (which are implemented as drivers on HID bus) wouldn't have to be enabled separately, if the device is otherwise more-or-less HID standard compliant. But this driver is rather standalone driver, so we'd want to have it normally selectable. Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index a05133f..0c8ce3a 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -56,9 +56,8 @@ menu "Special HID drivers" depends on HID config HID_3M_PCT - tristate "3M PCT" if EMBEDDED + tristate "3M PCT" depends on USB_HID - default !EMBEDDED ---help--- Support for 3M PCT touch screens. -- cgit v0.10.2 From 3912f2abc942a002ef611fc973add5e5eadb3432 Mon Sep 17 00:00:00 2001 From: Amerigo Wang Date: Mon, 14 Dec 2009 03:23:56 -0500 Subject: perf: Use format string of printf to align strings Instead of filling whitespaces to do alignment, use printf's format string. This simplifies the code a bit. Signed-off-by: WANG Cong Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <20091214082700.4224.57640.sendpatchset@localhost.localdomain> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-help.c b/tools/perf/builtin-help.c index 9f810b1..e427d69 100644 --- a/tools/perf/builtin-help.c +++ b/tools/perf/builtin-help.c @@ -286,8 +286,7 @@ void list_common_cmds_help(void) puts(" The most commonly used perf commands are:"); for (i = 0; i < ARRAY_SIZE(common_cmds); i++) { - printf(" %s ", common_cmds[i].name); - mput_char(' ', longest - strlen(common_cmds[i].name)); + printf(" %-*s ", longest, common_cmds[i].name); puts(common_cmds[i].help); } } -- cgit v0.10.2 From 06aae590033d1ae3c35b2920ef950cfc603e2a2d Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 27 Dec 2009 21:36:59 -0200 Subject: perf session: Move the event processing routines to session.c MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit No need for an extra "data_map" file since the routines there operate mainly on a perf_session instance. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1261957026-15580-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 652a470..4172c3b 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -423,7 +423,6 @@ LIB_OBJS += util/trace-event-perl.o LIB_OBJS += util/svghelper.o LIB_OBJS += util/sort.o LIB_OBJS += util/hist.o -LIB_OBJS += util/data_map.o LIB_OBJS += util/probe-event.o BUILTIN_OBJS += builtin-annotate.o diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c deleted file mode 100644 index b557b83..0000000 --- a/tools/perf/util/data_map.c +++ /dev/null @@ -1,252 +0,0 @@ -#include "symbol.h" -#include "util.h" -#include "debug.h" -#include "thread.h" -#include "session.h" - -static int process_event_stub(event_t *event __used, - struct perf_session *session __used) -{ - dump_printf(": unhandled!\n"); - return 0; -} - -static void perf_event_ops__fill_defaults(struct perf_event_ops *handler) -{ - if (!handler->process_sample_event) - handler->process_sample_event = process_event_stub; - if (!handler->process_mmap_event) - handler->process_mmap_event = process_event_stub; - if (!handler->process_comm_event) - handler->process_comm_event = process_event_stub; - if (!handler->process_fork_event) - handler->process_fork_event = process_event_stub; - if (!handler->process_exit_event) - handler->process_exit_event = process_event_stub; - if (!handler->process_lost_event) - handler->process_lost_event = process_event_stub; - if (!handler->process_read_event) - handler->process_read_event = process_event_stub; - if (!handler->process_throttle_event) - handler->process_throttle_event = process_event_stub; - if (!handler->process_unthrottle_event) - handler->process_unthrottle_event = process_event_stub; -} - -static const char *event__name[] = { - [0] = "TOTAL", - [PERF_RECORD_MMAP] = "MMAP", - [PERF_RECORD_LOST] = "LOST", - [PERF_RECORD_COMM] = "COMM", - [PERF_RECORD_EXIT] = "EXIT", - [PERF_RECORD_THROTTLE] = "THROTTLE", - [PERF_RECORD_UNTHROTTLE] = "UNTHROTTLE", - [PERF_RECORD_FORK] = "FORK", - [PERF_RECORD_READ] = "READ", - [PERF_RECORD_SAMPLE] = "SAMPLE", -}; - -unsigned long event__total[PERF_RECORD_MAX]; - -void event__print_totals(void) -{ - int i; - for (i = 0; i < PERF_RECORD_MAX; ++i) - pr_info("%10s events: %10ld\n", - event__name[i], event__total[i]); -} - -static int process_event(event_t *event, struct perf_session *session, - struct perf_event_ops *ops, - unsigned long offset, unsigned long head) -{ - trace_event(event); - - if (event->header.type < PERF_RECORD_MAX) { - dump_printf("%p [%p]: PERF_RECORD_%s", - (void *)(offset + head), - (void *)(long)(event->header.size), - event__name[event->header.type]); - ++event__total[0]; - ++event__total[event->header.type]; - } - - switch (event->header.type) { - case PERF_RECORD_SAMPLE: - return ops->process_sample_event(event, session); - case PERF_RECORD_MMAP: - return ops->process_mmap_event(event, session); - case PERF_RECORD_COMM: - return ops->process_comm_event(event, session); - case PERF_RECORD_FORK: - return ops->process_fork_event(event, session); - case PERF_RECORD_EXIT: - return ops->process_exit_event(event, session); - case PERF_RECORD_LOST: - return ops->process_lost_event(event, session); - case PERF_RECORD_READ: - return ops->process_read_event(event, session); - case PERF_RECORD_THROTTLE: - return ops->process_throttle_event(event, session); - case PERF_RECORD_UNTHROTTLE: - return ops->process_unthrottle_event(event, session); - default: - ops->total_unknown++; - return -1; - } -} - -int perf_header__read_build_ids(int input, u64 offset, u64 size) -{ - struct build_id_event bev; - char filename[PATH_MAX]; - u64 limit = offset + size; - int err = -1; - - while (offset < limit) { - struct dso *dso; - ssize_t len; - - if (read(input, &bev, sizeof(bev)) != sizeof(bev)) - goto out; - - len = bev.header.size - sizeof(bev); - if (read(input, filename, len) != len) - goto out; - - dso = dsos__findnew(filename); - if (dso != NULL) - dso__set_build_id(dso, &bev.build_id); - - offset += bev.header.size; - } - err = 0; -out: - return err; -} - -static struct thread *perf_session__register_idle_thread(struct perf_session *self) -{ - struct thread *thread = perf_session__findnew(self, 0); - - if (!thread || thread__set_comm(thread, "swapper")) { - pr_err("problem inserting idle task.\n"); - thread = NULL; - } - - return thread; -} - -int perf_session__process_events(struct perf_session *self, - struct perf_event_ops *ops) -{ - int err; - unsigned long head, shift; - unsigned long offset = 0; - size_t page_size; - event_t *event; - uint32_t size; - char *buf; - - if (perf_session__register_idle_thread(self) == NULL) - return -ENOMEM; - - perf_event_ops__fill_defaults(ops); - - page_size = getpagesize(); - - head = self->header.data_offset; - self->sample_type = perf_header__sample_type(&self->header); - - err = -EINVAL; - if (ops->sample_type_check && ops->sample_type_check(self) < 0) - goto out_err; - - if (!ops->full_paths) { - char bf[PATH_MAX]; - - if (getcwd(bf, sizeof(bf)) == NULL) { - err = -errno; -out_getcwd_err: - pr_err("failed to get the current directory\n"); - goto out_err; - } - self->cwd = strdup(bf); - if (self->cwd == NULL) { - err = -ENOMEM; - goto out_getcwd_err; - } - self->cwdlen = strlen(self->cwd); - } - - shift = page_size * (head / page_size); - offset += shift; - head -= shift; - -remap: - buf = mmap(NULL, page_size * self->mmap_window, PROT_READ, - MAP_SHARED, self->fd, offset); - if (buf == MAP_FAILED) { - pr_err("failed to mmap file\n"); - err = -errno; - goto out_err; - } - -more: - event = (event_t *)(buf + head); - - size = event->header.size; - if (!size) - size = 8; - - if (head + event->header.size >= page_size * self->mmap_window) { - int munmap_ret; - - shift = page_size * (head / page_size); - - munmap_ret = munmap(buf, page_size * self->mmap_window); - assert(munmap_ret == 0); - - offset += shift; - head -= shift; - goto remap; - } - - size = event->header.size; - - dump_printf("\n%p [%p]: event: %d\n", - (void *)(offset + head), - (void *)(long)event->header.size, - event->header.type); - - if (!size || process_event(event, self, ops, offset, head) < 0) { - - dump_printf("%p [%p]: skipping unknown header type: %d\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->header.type); - - /* - * assume we lost track of the stream, check alignment, and - * increment a single u64 in the hope to catch on again 'soon'. - */ - - if (unlikely(head & 7)) - head &= ~7ULL; - - size = 8; - } - - head += size; - - if (offset + head >= self->header.data_offset + self->header.data_size) - goto done; - - if (offset + head < self->size) - goto more; - -done: - err = 0; -out_err: - return err; -} diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index ce3a6c8..736d4fd 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -148,3 +148,248 @@ struct symbol **perf_session__resolve_callchain(struct perf_session *self, return syms; } + +static int process_event_stub(event_t *event __used, + struct perf_session *session __used) +{ + dump_printf(": unhandled!\n"); + return 0; +} + +static void perf_event_ops__fill_defaults(struct perf_event_ops *handler) +{ + if (handler->process_sample_event == NULL) + handler->process_sample_event = process_event_stub; + if (handler->process_mmap_event == NULL) + handler->process_mmap_event = process_event_stub; + if (handler->process_comm_event == NULL) + handler->process_comm_event = process_event_stub; + if (handler->process_fork_event == NULL) + handler->process_fork_event = process_event_stub; + if (handler->process_exit_event == NULL) + handler->process_exit_event = process_event_stub; + if (handler->process_lost_event == NULL) + handler->process_lost_event = process_event_stub; + if (handler->process_read_event == NULL) + handler->process_read_event = process_event_stub; + if (handler->process_throttle_event == NULL) + handler->process_throttle_event = process_event_stub; + if (handler->process_unthrottle_event == NULL) + handler->process_unthrottle_event = process_event_stub; +} + +static const char *event__name[] = { + [0] = "TOTAL", + [PERF_RECORD_MMAP] = "MMAP", + [PERF_RECORD_LOST] = "LOST", + [PERF_RECORD_COMM] = "COMM", + [PERF_RECORD_EXIT] = "EXIT", + [PERF_RECORD_THROTTLE] = "THROTTLE", + [PERF_RECORD_UNTHROTTLE] = "UNTHROTTLE", + [PERF_RECORD_FORK] = "FORK", + [PERF_RECORD_READ] = "READ", + [PERF_RECORD_SAMPLE] = "SAMPLE", +}; + +unsigned long event__total[PERF_RECORD_MAX]; + +void event__print_totals(void) +{ + int i; + for (i = 0; i < PERF_RECORD_MAX; ++i) + pr_info("%10s events: %10ld\n", + event__name[i], event__total[i]); +} + +static int perf_session__process_event(struct perf_session *self, + event_t *event, + struct perf_event_ops *ops, + unsigned long offset, unsigned long head) +{ + trace_event(event); + + if (event->header.type < PERF_RECORD_MAX) { + dump_printf("%p [%p]: PERF_RECORD_%s", + (void *)(offset + head), + (void *)(long)(event->header.size), + event__name[event->header.type]); + ++event__total[0]; + ++event__total[event->header.type]; + } + + switch (event->header.type) { + case PERF_RECORD_SAMPLE: + return ops->process_sample_event(event, self); + case PERF_RECORD_MMAP: + return ops->process_mmap_event(event, self); + case PERF_RECORD_COMM: + return ops->process_comm_event(event, self); + case PERF_RECORD_FORK: + return ops->process_fork_event(event, self); + case PERF_RECORD_EXIT: + return ops->process_exit_event(event, self); + case PERF_RECORD_LOST: + return ops->process_lost_event(event, self); + case PERF_RECORD_READ: + return ops->process_read_event(event, self); + case PERF_RECORD_THROTTLE: + return ops->process_throttle_event(event, self); + case PERF_RECORD_UNTHROTTLE: + return ops->process_unthrottle_event(event, self); + default: + ops->total_unknown++; + return -1; + } +} + +int perf_header__read_build_ids(int input, u64 offset, u64 size) +{ + struct build_id_event bev; + char filename[PATH_MAX]; + u64 limit = offset + size; + int err = -1; + + while (offset < limit) { + struct dso *dso; + ssize_t len; + + if (read(input, &bev, sizeof(bev)) != sizeof(bev)) + goto out; + + len = bev.header.size - sizeof(bev); + if (read(input, filename, len) != len) + goto out; + + dso = dsos__findnew(filename); + if (dso != NULL) + dso__set_build_id(dso, &bev.build_id); + + offset += bev.header.size; + } + err = 0; +out: + return err; +} + +static struct thread *perf_session__register_idle_thread(struct perf_session *self) +{ + struct thread *thread = perf_session__findnew(self, 0); + + if (thread == NULL || thread__set_comm(thread, "swapper")) { + pr_err("problem inserting idle task.\n"); + thread = NULL; + } + + return thread; +} + +int perf_session__process_events(struct perf_session *self, + struct perf_event_ops *ops) +{ + int err; + unsigned long head, shift; + unsigned long offset = 0; + size_t page_size; + event_t *event; + uint32_t size; + char *buf; + + if (perf_session__register_idle_thread(self) == NULL) + return -ENOMEM; + + perf_event_ops__fill_defaults(ops); + + page_size = getpagesize(); + + head = self->header.data_offset; + self->sample_type = perf_header__sample_type(&self->header); + + err = -EINVAL; + if (ops->sample_type_check && ops->sample_type_check(self) < 0) + goto out_err; + + if (!ops->full_paths) { + char bf[PATH_MAX]; + + if (getcwd(bf, sizeof(bf)) == NULL) { + err = -errno; +out_getcwd_err: + pr_err("failed to get the current directory\n"); + goto out_err; + } + self->cwd = strdup(bf); + if (self->cwd == NULL) { + err = -ENOMEM; + goto out_getcwd_err; + } + self->cwdlen = strlen(self->cwd); + } + + shift = page_size * (head / page_size); + offset += shift; + head -= shift; + +remap: + buf = mmap(NULL, page_size * self->mmap_window, PROT_READ, + MAP_SHARED, self->fd, offset); + if (buf == MAP_FAILED) { + pr_err("failed to mmap file\n"); + err = -errno; + goto out_err; + } + +more: + event = (event_t *)(buf + head); + + size = event->header.size; + if (size == 0) + size = 8; + + if (head + event->header.size >= page_size * self->mmap_window) { + int munmap_ret; + + shift = page_size * (head / page_size); + + munmap_ret = munmap(buf, page_size * self->mmap_window); + assert(munmap_ret == 0); + + offset += shift; + head -= shift; + goto remap; + } + + size = event->header.size; + + dump_printf("\n%p [%p]: event: %d\n", + (void *)(offset + head), + (void *)(long)event->header.size, + event->header.type); + + if (size == 0 || + perf_session__process_event(self, event, ops, offset, head) < 0) { + dump_printf("%p [%p]: skipping unknown header type: %d\n", + (void *)(offset + head), + (void *)(long)(event->header.size), + event->header.type); + /* + * assume we lost track of the stream, check alignment, and + * increment a single u64 in the hope to catch on again 'soon'. + */ + if (unlikely(head & 7)) + head &= ~7ULL; + + size = 8; + } + + head += size; + + if (offset + head >= self->header.data_offset + self->header.data_size) + goto done; + + if (offset + head < self->size) + goto more; +done: + err = 0; +out_err: + return err; +} -- cgit v0.10.2 From 4a58e61161074776aa34187ea369414ce4852394 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 27 Dec 2009 21:37:00 -0200 Subject: perf tools: Move the map class definition to a separate header MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit And this resulted in the need for adding some missing includes in some places that were getting the definitions needed out of sheer luck. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1261957026-15580-4-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 4172c3b..fafea0b 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -357,6 +357,7 @@ LIB_H += util/event.h LIB_H += util/exec_cmd.h LIB_H += util/types.h LIB_H += util/levenshtein.h +LIB_H += util/map.h LIB_H += util/parse-options.h LIB_H += util/parse-events.h LIB_H += util/quote.h diff --git a/tools/perf/util/debug.c b/tools/perf/util/debug.c index 28d520d..0905600 100644 --- a/tools/perf/util/debug.c +++ b/tools/perf/util/debug.c @@ -9,6 +9,7 @@ #include "color.h" #include "event.h" #include "debug.h" +#include "util.h" int verbose = 0; int dump_trace = 0; diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 690a96d..80fb365 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -1,10 +1,10 @@ #ifndef __PERF_RECORD_H #define __PERF_RECORD_H +#include + #include "../perf.h" -#include "util.h" -#include -#include +#include "map.h" /* * PERF_SAMPLE_IP | PERF_SAMPLE_TID | * @@ -101,67 +101,8 @@ struct events_stats { void event__print_totals(void); -enum map_type { - MAP__FUNCTION = 0, - MAP__VARIABLE, -}; - -#define MAP__NR_TYPES (MAP__VARIABLE + 1) - -struct map { - union { - struct rb_node rb_node; - struct list_head node; - }; - u64 start; - u64 end; - enum map_type type; - u64 pgoff; - u64 (*map_ip)(struct map *, u64); - u64 (*unmap_ip)(struct map *, u64); - struct dso *dso; -}; - -static inline u64 map__map_ip(struct map *map, u64 ip) -{ - return ip - map->start + map->pgoff; -} - -static inline u64 map__unmap_ip(struct map *map, u64 ip) -{ - return ip + map->start - map->pgoff; -} - -static inline u64 identity__map_ip(struct map *map __used, u64 ip) -{ - return ip; -} - -struct symbol; - -typedef int (*symbol_filter_t)(struct map *map, struct symbol *sym); - -void map__init(struct map *self, enum map_type type, - u64 start, u64 end, u64 pgoff, struct dso *dso); -struct map *map__new(struct mmap_event *event, enum map_type, - char *cwd, int cwdlen); -void map__delete(struct map *self); -struct map *map__clone(struct map *self); -int map__overlap(struct map *l, struct map *r); -size_t map__fprintf(struct map *self, FILE *fp); - struct perf_session; -int map__load(struct map *self, struct perf_session *session, - symbol_filter_t filter); -struct symbol *map__find_symbol(struct map *self, struct perf_session *session, - u64 addr, symbol_filter_t filter); -struct symbol *map__find_symbol_by_name(struct map *self, const char *name, - struct perf_session *session, - symbol_filter_t filter); -void map__fixup_start(struct map *self); -void map__fixup_end(struct map *self); - int event__synthesize_thread(pid_t pid, int (*process)(event_t *event, struct perf_session *session), diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h new file mode 100644 index 0000000..72f0b6a --- /dev/null +++ b/tools/perf/util/map.h @@ -0,0 +1,73 @@ +#ifndef __PERF_MAP_H +#define __PERF_MAP_H + +#include +#include +#include +#include + +enum map_type { + MAP__FUNCTION = 0, + MAP__VARIABLE, +}; + +#define MAP__NR_TYPES (MAP__VARIABLE + 1) + +struct dso; + +struct map { + union { + struct rb_node rb_node; + struct list_head node; + }; + u64 start; + u64 end; + enum map_type type; + u64 pgoff; + u64 (*map_ip)(struct map *, u64); + u64 (*unmap_ip)(struct map *, u64); + struct dso *dso; +}; + +static inline u64 map__map_ip(struct map *map, u64 ip) +{ + return ip - map->start + map->pgoff; +} + +static inline u64 map__unmap_ip(struct map *map, u64 ip) +{ + return ip + map->start - map->pgoff; +} + +static inline u64 identity__map_ip(struct map *map __used, u64 ip) +{ + return ip; +} + +struct symbol; +struct mmap_event; + +typedef int (*symbol_filter_t)(struct map *map, struct symbol *sym); + +void map__init(struct map *self, enum map_type type, + u64 start, u64 end, u64 pgoff, struct dso *dso); +struct map *map__new(struct mmap_event *event, enum map_type, + char *cwd, int cwdlen); +void map__delete(struct map *self); +struct map *map__clone(struct map *self); +int map__overlap(struct map *l, struct map *r); +size_t map__fprintf(struct map *self, FILE *fp); + +struct perf_session; + +int map__load(struct map *self, struct perf_session *session, + symbol_filter_t filter); +struct symbol *map__find_symbol(struct map *self, struct perf_session *session, + u64 addr, symbol_filter_t filter); +struct symbol *map__find_symbol_by_name(struct map *self, const char *name, + struct perf_session *session, + symbol_filter_t filter); +void map__fixup_start(struct map *self); +void map__fixup_end(struct map *self); + +#endif /* __PERF_MAP_H */ diff --git a/tools/perf/util/probe-finder.h b/tools/perf/util/probe-finder.h index a4086aad..e3f3968 100644 --- a/tools/perf/util/probe-finder.h +++ b/tools/perf/util/probe-finder.h @@ -1,6 +1,8 @@ #ifndef _PROBE_FINDER_H #define _PROBE_FINDER_H +#include "util.h" + #define MAX_PATH_LEN 256 #define MAX_PROBE_BUFFER 1024 #define MAX_PROBES 128 -- cgit v0.10.2 From 27295592c22e71bbd38110c302da8dbb43912a60 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 27 Dec 2009 21:37:01 -0200 Subject: perf session: Share the common trace sample_check routine as perf_session__has_traces MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1261957026-15580-5-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index fc21ad7..a85936f 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -342,22 +342,10 @@ static int process_sample_event(event_t *event, struct perf_session *session) return 0; } -static int sample_type_check(struct perf_session *session) -{ - if (!(session->sample_type & PERF_SAMPLE_RAW)) { - fprintf(stderr, - "No trace sample to read. Did you call perf record " - "without -R?"); - return -1; - } - - return 0; -} - static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_comm_event = event__process_comm, - .sample_type_check = sample_type_check, + .sample_type_check = perf_session__has_traces, }; static double fragmentation(unsigned long n_req, unsigned long n_alloc) diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 80209df..d65098c 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1653,23 +1653,11 @@ static int process_lost_event(event_t *event __used, return 0; } -static int sample_type_check(struct perf_session *session __used) -{ - if (!(session->sample_type & PERF_SAMPLE_RAW)) { - fprintf(stderr, - "No trace sample to read. Did you call perf record " - "without -R?"); - return -1; - } - - return 0; -} - static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_comm_event = event__process_comm, .process_lost_event = process_lost_event, - .sample_type_check = sample_type_check, + .sample_type_check = perf_session__has_traces, }; static int read_events(void) diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index a589a43..b42f337 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -1029,23 +1029,12 @@ static void process_samples(struct perf_session *session) } } -static int sample_type_check(struct perf_session *session) -{ - if (!(session->sample_type & PERF_SAMPLE_RAW)) { - fprintf(stderr, "No trace samples found in the file.\n" - "Have you used 'perf timechart record' to record it?\n"); - return -1; - } - - return 0; -} - static struct perf_event_ops event_ops = { .process_comm_event = process_comm_event, .process_fork_event = process_fork_event, .process_exit_event = process_exit_event, .process_sample_event = queue_sample_event, - .sample_type_check = sample_type_check, + .sample_type_check = perf_session__has_traces, }; static int __cmd_timechart(void) diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 574a215..b0ba2ac 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -103,22 +103,10 @@ static int process_sample_event(event_t *event, struct perf_session *session) return 0; } -static int sample_type_check(struct perf_session *session) -{ - if (!(session->sample_type & PERF_SAMPLE_RAW)) { - fprintf(stderr, - "No trace sample to read. Did you call perf record " - "without -R?"); - return -1; - } - - return 0; -} - static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_comm_event = event__process_comm, - .sample_type_check = sample_type_check, + .sample_type_check = perf_session__has_traces, }; static int __cmd_trace(struct perf_session *session) diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 736d4fd..60eab8b 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -393,3 +393,14 @@ done: out_err: return err; } + +int perf_session__has_traces(struct perf_session *self) +{ + if (!(self->sample_type & PERF_SAMPLE_RAW)) { + pr_err("No trace sample to read. Did you call perf record " + "without -R?"); + return -1; + } + + return 0; +} diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 32eaa1b..a6951d2 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -56,6 +56,8 @@ struct symbol **perf_session__resolve_callchain(struct perf_session *self, struct ip_callchain *chain, struct symbol **parent); +int perf_session__has_traces(struct perf_session *self); + int perf_header__read_build_ids(int input, u64 offset, u64 file_size); #endif /* __PERF_SESSION_H */ -- cgit v0.10.2 From d549c7690190d9739005e19604faad6da4b802ac Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 27 Dec 2009 21:37:02 -0200 Subject: perf session: Remove sample_type_check from event_ops MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This is really something tools need to do before asking for the events to be processed, leaving perf_session__process_events to do just that, process events. Also add a msg parameter to perf_session__has_traces() so that the right message can be printed, fixing a regression added by me in the previous cset (right timechart message) and also fixing 'perf kmem', that was not asking if 'perf kmem record' was ran. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1261957026-15580-6-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index a85936f..73b0650 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -345,7 +345,6 @@ static int process_sample_event(event_t *event, struct perf_session *session) static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_comm_event = event__process_comm, - .sample_type_check = perf_session__has_traces, }; static double fragmentation(unsigned long n_req, unsigned long n_alloc) @@ -492,11 +491,14 @@ static void sort_result(void) static int __cmd_kmem(void) { - int err; + int err = -EINVAL; struct perf_session *session = perf_session__new(input_name, O_RDONLY, 0); if (session == NULL) return -ENOMEM; + if (!perf_session__has_traces(session, "kmem record")) + goto out_delete; + setup_pager(); err = perf_session__process_events(session, &event_ops); if (err != 0) diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index db10c0e..0825918 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -156,14 +156,14 @@ static int process_read_event(event_t *event, struct perf_session *session __use return 0; } -static int sample_type_check(struct perf_session *session) +static int perf_session__setup_sample_type(struct perf_session *self) { - if (!(session->sample_type & PERF_SAMPLE_CALLCHAIN)) { + if (!(self->sample_type & PERF_SAMPLE_CALLCHAIN)) { if (sort__has_parent) { fprintf(stderr, "selected --sort parent, but no" " callchain data. Did you call" " perf record without -g?\n"); - return -1; + return -EINVAL; } if (symbol_conf.use_callchain) { fprintf(stderr, "selected -g but no callchain data." @@ -176,7 +176,7 @@ static int sample_type_check(struct perf_session *session) if (register_callchain_param(&callchain_param) < 0) { fprintf(stderr, "Can't register callchain" " params\n"); - return -1; + return -EINVAL; } } @@ -191,13 +191,11 @@ static struct perf_event_ops event_ops = { .process_fork_event = event__process_task, .process_lost_event = event__process_lost, .process_read_event = process_read_event, - .sample_type_check = sample_type_check, }; - static int __cmd_report(void) { - int ret; + int ret = -EINVAL; struct perf_session *session; session = perf_session__new(input_name, O_RDONLY, force); @@ -207,6 +205,10 @@ static int __cmd_report(void) if (show_threads) perf_read_values_init(&show_threads_values); + ret = perf_session__setup_sample_type(session); + if (ret) + goto out_delete; + ret = perf_session__process_events(session, &event_ops); if (ret) goto out_delete; diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index d65098c..e862e71 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1657,17 +1657,18 @@ static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_comm_event = event__process_comm, .process_lost_event = process_lost_event, - .sample_type_check = perf_session__has_traces, }; static int read_events(void) { - int err; + int err = -EINVAL; struct perf_session *session = perf_session__new(input_name, O_RDONLY, 0); if (session == NULL) return -ENOMEM; - err = perf_session__process_events(session, &event_ops); + if (perf_session__has_traces(session, "record -R")) + err = perf_session__process_events(session, &event_ops); + perf_session__delete(session); return err; } diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index b42f337..8252837 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -1034,17 +1034,19 @@ static struct perf_event_ops event_ops = { .process_fork_event = process_fork_event, .process_exit_event = process_exit_event, .process_sample_event = queue_sample_event, - .sample_type_check = perf_session__has_traces, }; static int __cmd_timechart(void) { struct perf_session *session = perf_session__new(input_name, O_RDONLY, 0); - int ret; + int ret = -EINVAL; if (session == NULL) return -ENOMEM; + if (!perf_session__has_traces(session, "timechart record")) + goto out_delete; + ret = perf_session__process_events(session, &event_ops); if (ret) goto out_delete; diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index b0ba2ac..e94f346 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -106,7 +106,6 @@ static int process_sample_event(event_t *event, struct perf_session *session) static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_comm_event = event__process_comm, - .sample_type_check = perf_session__has_traces, }; static int __cmd_trace(struct perf_session *session) @@ -580,6 +579,9 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used) if (session == NULL) return -ENOMEM; + if (!perf_session__has_traces(session, "record -R")) + return -EINVAL; + if (generate_script_lang) { struct stat perf_stat; diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 60eab8b..bc84a52 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -73,6 +73,8 @@ struct perf_session *perf_session__new(const char *filename, int mode, bool forc if (mode == O_RDONLY && perf_session__open(self, force) < 0) goto out_delete; + + self->sample_type = perf_header__sample_type(&self->header); out: return self; out_free: @@ -302,11 +304,6 @@ int perf_session__process_events(struct perf_session *self, page_size = getpagesize(); head = self->header.data_offset; - self->sample_type = perf_header__sample_type(&self->header); - - err = -EINVAL; - if (ops->sample_type_check && ops->sample_type_check(self) < 0) - goto out_err; if (!ops->full_paths) { char bf[PATH_MAX]; @@ -394,13 +391,12 @@ out_err: return err; } -int perf_session__has_traces(struct perf_session *self) +bool perf_session__has_traces(struct perf_session *self, const char *msg) { if (!(self->sample_type & PERF_SAMPLE_RAW)) { - pr_err("No trace sample to read. Did you call perf record " - "without -R?"); - return -1; + pr_err("No trace sample to read. Did you call 'perf %s'?\n", msg); + return false; } - return 0; + return true; } diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index a6951d2..5771ccb 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -40,7 +40,6 @@ struct perf_event_ops { event_op process_read_event; event_op process_throttle_event; event_op process_unthrottle_event; - int (*sample_type_check)(struct perf_session *session); unsigned long total_unknown; bool full_paths; }; @@ -56,7 +55,7 @@ struct symbol **perf_session__resolve_callchain(struct perf_session *self, struct ip_callchain *chain, struct symbol **parent); -int perf_session__has_traces(struct perf_session *self); +bool perf_session__has_traces(struct perf_session *self, const char *msg); int perf_header__read_build_ids(int input, u64 offset, u64 file_size); -- cgit v0.10.2 From 31d337c4ee3152b7271897eae576251643f5a3b5 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 27 Dec 2009 21:37:03 -0200 Subject: perf session: Move total_unknown to perf_session->unknown events MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As this is a session property, not belonging to perf_event_ops, that can be shared by many perf_session instances. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1261957026-15580-7-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index bc84a52..4ca427f 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -66,6 +66,7 @@ struct perf_session *perf_session__new(const char *filename, int mode, bool forc self->mmap_window = 32; self->cwd = NULL; self->cwdlen = 0; + self->unknown_events = 0; map_groups__init(&self->kmaps); if (perf_session__create_kernel_maps(self) < 0) @@ -239,7 +240,7 @@ static int perf_session__process_event(struct perf_session *self, case PERF_RECORD_UNTHROTTLE: return ops->process_unthrottle_event(event, self); default: - ops->total_unknown++; + self->unknown_events++; return -1; } } diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 5771ccb..585937b 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -20,6 +20,7 @@ struct perf_session { struct thread *last_match; struct events_stats events_stats; unsigned long event_total[PERF_RECORD_MAX]; + unsigned long unknown_events; struct rb_root hists; u64 sample_type; int fd; @@ -40,7 +41,6 @@ struct perf_event_ops { event_op process_read_event; event_op process_throttle_event; event_op process_unthrottle_event; - unsigned long total_unknown; bool full_paths; }; -- cgit v0.10.2 From f7d87444e6ee6f4a19634e5412664c1c529a2370 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 27 Dec 2009 21:37:04 -0200 Subject: perf session: Move full_paths config to symbol_conf MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Now perf_event_ops has just that, event handlers. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1261957026-15580-8-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c index bd71b8c..e164b3d 100644 --- a/tools/perf/builtin-diff.c +++ b/tools/perf/builtin-diff.c @@ -204,7 +204,7 @@ static const struct option options[] = { OPT_BOOLEAN('f', "force", &force, "don't complain, do it"), OPT_BOOLEAN('m', "modules", &symbol_conf.use_modules, "load module symbols - WARNING: use only with -k and LIVE kernel"), - OPT_BOOLEAN('P', "full-paths", &event_ops.full_paths, + OPT_BOOLEAN('P', "full-paths", &symbol_conf.full_paths, "Don't shorten the pathnames taking into account the cwd"), OPT_STRING('d', "dsos", &symbol_conf.dso_list_str, "dso[,dso...]", "only consider symbols in these dsos"), diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 0825918..f695084 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -321,7 +321,7 @@ static const struct option options[] = { "pretty printing style key: normal raw"), OPT_STRING('s', "sort", &sort_order, "key[,key2...]", "sort by key(s): pid, comm, dso, symbol, parent"), - OPT_BOOLEAN('P', "full-paths", &event_ops.full_paths, + OPT_BOOLEAN('P', "full-paths", &symbol_conf.full_paths, "Don't shorten the pathnames taking into account the cwd"), OPT_STRING('p', "parent", &parent_pattern, "regex", "regex filter to identify parent, see: '--sort parent'"), diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 4ca427f..4f2eeb58 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -306,7 +306,7 @@ int perf_session__process_events(struct perf_session *self, head = self->header.data_offset; - if (!ops->full_paths) { + if (!symbol_conf.full_paths) { char bf[PATH_MAX]; if (getcwd(bf, sizeof(bf)) == NULL) { diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 585937b..2ff77fe 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -41,7 +41,6 @@ struct perf_event_ops { event_op process_read_event; event_op process_throttle_event; event_op process_unthrottle_event; - bool full_paths; }; struct perf_session *perf_session__new(const char *filename, int mode, bool force); diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 8aded23..9eabd60 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -58,7 +58,8 @@ struct symbol_conf { sort_by_name, show_nr_samples, use_callchain, - exclude_other; + exclude_other, + full_paths; const char *vmlinux_name, *field_sep; char *dso_list_str, -- cgit v0.10.2 From 55aa640f54280da25046acd2075842d464f451e6 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 27 Dec 2009 21:37:05 -0200 Subject: perf session: Remove redundant prefix & suffix from perf_event_ops MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since now all that we have are perf event handlers, leave just the name of the event. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1261957026-15580-9-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 593ff25..117bbae 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -451,10 +451,10 @@ static void perf_session__find_annotations(struct perf_session *self) } static struct perf_event_ops event_ops = { - .process_sample_event = process_sample_event, - .process_mmap_event = event__process_mmap, - .process_comm_event = event__process_comm, - .process_fork_event = event__process_task, + .sample = process_sample_event, + .mmap = event__process_mmap, + .comm = event__process_comm, + .fork = event__process_task, }; static int __cmd_annotate(void) diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c index e164b3d..1cbecaf 100644 --- a/tools/perf/builtin-diff.c +++ b/tools/perf/builtin-diff.c @@ -66,12 +66,12 @@ static int diff__process_sample_event(event_t *event, struct perf_session *sessi } static struct perf_event_ops event_ops = { - .process_sample_event = diff__process_sample_event, - .process_mmap_event = event__process_mmap, - .process_comm_event = event__process_comm, - .process_exit_event = event__process_task, - .process_fork_event = event__process_task, - .process_lost_event = event__process_lost, + .sample = diff__process_sample_event, + .mmap = event__process_mmap, + .comm = event__process_comm, + .exit = event__process_task, + .fork = event__process_task, + .lost = event__process_lost, }; static void perf_session__insert_hist_entry_by_name(struct rb_root *root, diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 73b0650..4c06828 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -343,8 +343,8 @@ static int process_sample_event(event_t *event, struct perf_session *session) } static struct perf_event_ops event_ops = { - .process_sample_event = process_sample_event, - .process_comm_event = event__process_comm, + .sample = process_sample_event, + .comm = event__process_comm, }; static double fragmentation(unsigned long n_req, unsigned long n_alloc) diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index f695084..508934b 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -184,13 +184,13 @@ static int perf_session__setup_sample_type(struct perf_session *self) } static struct perf_event_ops event_ops = { - .process_sample_event = process_sample_event, - .process_mmap_event = event__process_mmap, - .process_comm_event = event__process_comm, - .process_exit_event = event__process_task, - .process_fork_event = event__process_task, - .process_lost_event = event__process_lost, - .process_read_event = process_read_event, + .sample = process_sample_event, + .mmap = event__process_mmap, + .comm = event__process_comm, + .exit = event__process_task, + .fork = event__process_task, + .lost = event__process_lost, + .read = process_read_event, }; static int __cmd_report(void) diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index e862e71..702322f 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1654,9 +1654,9 @@ static int process_lost_event(event_t *event __used, } static struct perf_event_ops event_ops = { - .process_sample_event = process_sample_event, - .process_comm_event = event__process_comm, - .process_lost_event = process_lost_event, + .sample = process_sample_event, + .comm = event__process_comm, + .lost = process_lost_event, }; static int read_events(void) diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 8252837..5b68d81 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -1030,10 +1030,10 @@ static void process_samples(struct perf_session *session) } static struct perf_event_ops event_ops = { - .process_comm_event = process_comm_event, - .process_fork_event = process_fork_event, - .process_exit_event = process_exit_event, - .process_sample_event = queue_sample_event, + .comm = process_comm_event, + .fork = process_fork_event, + .exit = process_exit_event, + .sample = queue_sample_event, }; static int __cmd_timechart(void) diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index e94f346..1831434 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -104,8 +104,8 @@ static int process_sample_event(event_t *event, struct perf_session *session) } static struct perf_event_ops event_ops = { - .process_sample_event = process_sample_event, - .process_comm_event = event__process_comm, + .sample = process_sample_event, + .comm = event__process_comm, }; static int __cmd_trace(struct perf_session *session) diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 4f2eeb58..7f0537d 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -161,24 +161,24 @@ static int process_event_stub(event_t *event __used, static void perf_event_ops__fill_defaults(struct perf_event_ops *handler) { - if (handler->process_sample_event == NULL) - handler->process_sample_event = process_event_stub; - if (handler->process_mmap_event == NULL) - handler->process_mmap_event = process_event_stub; - if (handler->process_comm_event == NULL) - handler->process_comm_event = process_event_stub; - if (handler->process_fork_event == NULL) - handler->process_fork_event = process_event_stub; - if (handler->process_exit_event == NULL) - handler->process_exit_event = process_event_stub; - if (handler->process_lost_event == NULL) - handler->process_lost_event = process_event_stub; - if (handler->process_read_event == NULL) - handler->process_read_event = process_event_stub; - if (handler->process_throttle_event == NULL) - handler->process_throttle_event = process_event_stub; - if (handler->process_unthrottle_event == NULL) - handler->process_unthrottle_event = process_event_stub; + if (handler->sample == NULL) + handler->sample = process_event_stub; + if (handler->mmap == NULL) + handler->mmap = process_event_stub; + if (handler->comm == NULL) + handler->comm = process_event_stub; + if (handler->fork == NULL) + handler->fork = process_event_stub; + if (handler->exit == NULL) + handler->exit = process_event_stub; + if (handler->lost == NULL) + handler->lost = process_event_stub; + if (handler->read == NULL) + handler->read = process_event_stub; + if (handler->throttle == NULL) + handler->throttle = process_event_stub; + if (handler->unthrottle == NULL) + handler->unthrottle = process_event_stub; } static const char *event__name[] = { @@ -222,23 +222,23 @@ static int perf_session__process_event(struct perf_session *self, switch (event->header.type) { case PERF_RECORD_SAMPLE: - return ops->process_sample_event(event, self); + return ops->sample(event, self); case PERF_RECORD_MMAP: - return ops->process_mmap_event(event, self); + return ops->mmap(event, self); case PERF_RECORD_COMM: - return ops->process_comm_event(event, self); + return ops->comm(event, self); case PERF_RECORD_FORK: - return ops->process_fork_event(event, self); + return ops->fork(event, self); case PERF_RECORD_EXIT: - return ops->process_exit_event(event, self); + return ops->exit(event, self); case PERF_RECORD_LOST: - return ops->process_lost_event(event, self); + return ops->lost(event, self); case PERF_RECORD_READ: - return ops->process_read_event(event, self); + return ops->read(event, self); case PERF_RECORD_THROTTLE: - return ops->process_throttle_event(event, self); + return ops->throttle(event, self); case PERF_RECORD_UNTHROTTLE: - return ops->process_unthrottle_event(event, self); + return ops->unthrottle(event, self); default: self->unknown_events++; return -1; diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 2ff77fe..77c5ee29 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -32,15 +32,15 @@ struct perf_session { typedef int (*event_op)(event_t *self, struct perf_session *session); struct perf_event_ops { - event_op process_sample_event; - event_op process_mmap_event; - event_op process_comm_event; - event_op process_fork_event; - event_op process_exit_event; - event_op process_lost_event; - event_op process_read_event; - event_op process_throttle_event; - event_op process_unthrottle_event; + event_op sample, + mmap, + comm, + fork, + exit, + lost, + read, + throttle, + unthrottle; }; struct perf_session *perf_session__new(const char *filename, int mode, bool force); -- cgit v0.10.2 From 4cf40131a5cf4918e83b3756e58a1fc9e984f8ef Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 27 Dec 2009 21:37:06 -0200 Subject: perf record: Introduce a symtab cache MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Now a cache will be created in a ~/.debug debuginfo like hierarchy, so that at the end of a 'perf record' session all the binaries (with build-ids) involved get collected and indexed by their build-ids, so that perf report can find them. This is interesting when developing software where you want to do a 'perf diff' with the previous build and opens avenues for lots more interesting tools, like a 'perf diff --graph' that takes more than two binaries into account. Tunables for collecting just the symtabs can be added if one doesn't want to have the full binary, but having the full binary allows things like 'perf rerecord' or other tools that can re-run the tests by having access to the exact binary in some perf.data file, so it may well be interesting to keep the full binary there. Space consumption is minimised by trying to use hard links, a 'perf cache' tool to manage the space used, a la ccache is required to purge older entries. With this in place it will be possible also to introduce new commands, 'perf archive' and 'perf restore' (or some more suitable and future proof names) to create a cpio/tar file with the perf data and the files in the cache that _had_ perf hits of interest. There are more aspects to polish, like finding the right vmlinux file to cache, etc, but this is enough for a first step. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1261957026-15580-10-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index fafea0b..7c84642 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -425,6 +425,7 @@ LIB_OBJS += util/svghelper.o LIB_OBJS += util/sort.o LIB_OBJS += util/hist.o LIB_OBJS += util/probe-event.o +LIB_OBJS += util/util.o BUILTIN_OBJS += builtin-annotate.o diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 8a0bca5..df237c3 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -169,20 +169,23 @@ static int do_write(int fd, const void *buf, size_t size) return 0; } +#define dsos__for_each_with_build_id(pos, head) \ + list_for_each_entry(pos, head, node) \ + if (!pos->has_build_id) \ + continue; \ + else + static int __dsos__write_buildid_table(struct list_head *head, int fd) { #define NAME_ALIGN 64 struct dso *pos; static const char zero_buf[NAME_ALIGN]; - list_for_each_entry(pos, head, node) { + dsos__for_each_with_build_id(pos, head) { int err; struct build_id_event b; - size_t len; + size_t len = pos->long_name_len + 1; - if (!pos->has_build_id) - continue; - len = pos->long_name_len + 1; len = ALIGN(len, NAME_ALIGN); memset(&b, 0, sizeof(b)); memcpy(&b.build_id, pos->build_id, sizeof(pos->build_id)); @@ -209,6 +212,74 @@ static int dsos__write_buildid_table(int fd) return err; } +static int dso__cache_build_id(struct dso *self, const char *debugdir) +{ + const size_t size = PATH_MAX; + char *filename = malloc(size), + *linkname = malloc(size), *targetname, *sbuild_id; + int len, err = -1; + + if (filename == NULL || linkname == NULL) + goto out_free; + + len = snprintf(filename, size, "%s%s", debugdir, self->long_name); + if (mkdir_p(filename, 0755)) + goto out_free; + + len += snprintf(filename + len, sizeof(filename) - len, "/"); + sbuild_id = filename + len; + build_id__sprintf(self->build_id, sizeof(self->build_id), sbuild_id); + + if (access(filename, F_OK) && link(self->long_name, filename) && + copyfile(self->long_name, filename)) + goto out_free; + + len = snprintf(linkname, size, "%s/.build-id/%.2s", + debugdir, sbuild_id); + + if (access(linkname, X_OK) && mkdir_p(linkname, 0755)) + goto out_free; + + snprintf(linkname + len, size - len, "/%s", sbuild_id + 2); + targetname = filename + strlen(debugdir) - 5; + memcpy(targetname, "../..", 5); + + if (symlink(targetname, linkname) == 0) + err = 0; +out_free: + free(filename); + free(linkname); + return err; +} + +static int __dsos__cache_build_ids(struct list_head *head, const char *debugdir) +{ + struct dso *pos; + int err = 0; + + dsos__for_each_with_build_id(pos, head) + if (dso__cache_build_id(pos, debugdir)) + err = -1; + + return err; +} + +static int dsos__cache_build_ids(void) +{ + int err_kernel, err_user; + char debugdir[PATH_MAX]; + + snprintf(debugdir, sizeof(debugdir), "%s/%s", getenv("HOME"), + DEBUG_CACHE_DIR); + + if (mkdir(debugdir, 0755) != 0 && errno != EEXIST) + return -1; + + err_kernel = __dsos__cache_build_ids(&dsos__kernel, debugdir); + err_user = __dsos__cache_build_ids(&dsos__user, debugdir); + return err_kernel || err_user ? -1 : 0; +} + static int perf_header__adds_write(struct perf_header *self, int fd) { int nr_sections; @@ -258,6 +329,7 @@ static int perf_header__adds_write(struct perf_header *self, int fd) goto out_free; } buildid_sec->size = lseek(fd, 0, SEEK_CUR) - buildid_sec->offset; + dsos__cache_build_ids(); } lseek(fd, sec_start, SEEK_SET); diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index ab92763..79ca6a0 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -22,6 +22,7 @@ enum dso_origin { DSO__ORIG_KERNEL = 0, DSO__ORIG_JAVA_JIT, + DSO__ORIG_BUILD_ID_CACHE, DSO__ORIG_FEDORA, DSO__ORIG_UBUNTU, DSO__ORIG_BUILDID, @@ -1191,6 +1192,7 @@ char dso__symtab_origin(const struct dso *self) static const char origin[] = { [DSO__ORIG_KERNEL] = 'k', [DSO__ORIG_JAVA_JIT] = 'j', + [DSO__ORIG_BUILD_ID_CACHE] = 'B', [DSO__ORIG_FEDORA] = 'f', [DSO__ORIG_UBUNTU] = 'u', [DSO__ORIG_BUILDID] = 'b', @@ -1209,6 +1211,7 @@ int dso__load(struct dso *self, struct map *map, struct perf_session *session, int size = PATH_MAX; char *name; u8 build_id[BUILD_ID_SIZE]; + char build_id_hex[BUILD_ID_SIZE * 2 + 1]; int ret = -1; int fd; @@ -1230,8 +1233,16 @@ int dso__load(struct dso *self, struct map *map, struct perf_session *session, return ret; } - self->origin = DSO__ORIG_FEDORA - 1; + self->origin = DSO__ORIG_BUILD_ID_CACHE; + if (self->has_build_id) { + build_id__sprintf(self->build_id, sizeof(self->build_id), + build_id_hex); + snprintf(name, size, "%s/%s/.build-id/%.2s/%s", + getenv("HOME"), DEBUG_CACHE_DIR, + build_id_hex, build_id_hex + 2); + goto open_file; + } more: do { self->origin++; @@ -1247,8 +1258,6 @@ more: case DSO__ORIG_BUILDID: if (filename__read_build_id(self->long_name, build_id, sizeof(build_id))) { - char build_id_hex[BUILD_ID_SIZE * 2 + 1]; - build_id__sprintf(build_id, sizeof(build_id), build_id_hex); snprintf(name, size, @@ -1276,7 +1285,7 @@ compare_build_id: if (!dso__build_id_equal(self, build_id)) goto more; } - +open_file: fd = open(name, O_RDONLY); } while (fd < 0); diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 9eabd60..f27e158 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -8,6 +8,8 @@ #include #include "event.h" +#define DEBUG_CACHE_DIR ".debug" + #ifdef HAVE_CPLUS_DEMANGLE extern char *cplus_demangle(const char *, int); diff --git a/tools/perf/util/util.c b/tools/perf/util/util.c new file mode 100644 index 0000000..f3c0798 --- /dev/null +++ b/tools/perf/util/util.c @@ -0,0 +1,69 @@ +#include +#include +#include +#include +#include +#include +#include "util.h" + +int mkdir_p(char *path, mode_t mode) +{ + struct stat st; + int err; + char *d = path; + + if (*d != '/') + return -1; + + if (stat(path, &st) == 0) + return 0; + + while (*++d == '/'); + + while ((d = strchr(d, '/'))) { + *d = '\0'; + err = stat(path, &st) && mkdir(path, mode); + *d++ = '/'; + if (err) + return -1; + while (*d == '/') + ++d; + } + return (stat(path, &st) && mkdir(path, mode)) ? -1 : 0; +} + +int copyfile(const char *from, const char *to) +{ + int fromfd, tofd; + struct stat st; + void *addr; + int err = -1; + + if (stat(from, &st)) + goto out; + + fromfd = open(from, O_RDONLY); + if (fromfd < 0) + goto out; + + tofd = creat(to, 0755); + if (tofd < 0) + goto out_close_from; + + addr = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fromfd, 0); + if (addr == MAP_FAILED) + goto out_close_to; + + if (write(tofd, addr, st.st_size) == st.st_size) + err = 0; + + munmap(addr, st.st_size); +out_close_to: + close(tofd); + if (err) + unlink(to); +out_close_from: + close(fromfd); +out: + return err; +} diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h index c673d88..0f5b2a6 100644 --- a/tools/perf/util/util.h +++ b/tools/perf/util/util.h @@ -403,4 +403,7 @@ void git_qsort(void *base, size_t nmemb, size_t size, #endif #endif +int mkdir_p(char *path, mode_t mode); +int copyfile(const char *from, const char *to); + #endif -- cgit v0.10.2 From 49f474331e563a6ecf3b1e87ec27ec5482b3e4f1 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Sun, 27 Dec 2009 11:51:52 +0100 Subject: perf events: Remove arg from perf sched hooks Since we only ever schedule the local cpu, there is no need to pass the cpu number to the perf sched hooks. This micro-optimizes things a bit. Signed-off-by: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index c66b34f..a494e75 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -746,10 +746,10 @@ extern int perf_max_events; extern const struct pmu *hw_perf_event_init(struct perf_event *event); -extern void perf_event_task_sched_in(struct task_struct *task, int cpu); +extern void perf_event_task_sched_in(struct task_struct *task); extern void perf_event_task_sched_out(struct task_struct *task, - struct task_struct *next, int cpu); -extern void perf_event_task_tick(struct task_struct *task, int cpu); + struct task_struct *next); +extern void perf_event_task_tick(struct task_struct *task); extern int perf_event_init_task(struct task_struct *child); extern void perf_event_exit_task(struct task_struct *child); extern void perf_event_free_task(struct task_struct *task); @@ -870,12 +870,12 @@ extern void perf_event_enable(struct perf_event *event); extern void perf_event_disable(struct perf_event *event); #else static inline void -perf_event_task_sched_in(struct task_struct *task, int cpu) { } +perf_event_task_sched_in(struct task_struct *task) { } static inline void perf_event_task_sched_out(struct task_struct *task, - struct task_struct *next, int cpu) { } + struct task_struct *next) { } static inline void -perf_event_task_tick(struct task_struct *task, int cpu) { } +perf_event_task_tick(struct task_struct *task) { } static inline int perf_event_init_task(struct task_struct *child) { return 0; } static inline void perf_event_exit_task(struct task_struct *child) { } static inline void perf_event_free_task(struct task_struct *task) { } diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 03cc061..099bd66 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1170,9 +1170,9 @@ static void perf_event_sync_stat(struct perf_event_context *ctx, * not restart the event. */ void perf_event_task_sched_out(struct task_struct *task, - struct task_struct *next, int cpu) + struct task_struct *next) { - struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_event_context *ctx = task->perf_event_ctxp; struct perf_event_context *next_ctx; struct perf_event_context *parent; @@ -1252,8 +1252,9 @@ static void perf_event_cpu_sched_out(struct perf_cpu_context *cpuctx) static void __perf_event_sched_in(struct perf_event_context *ctx, - struct perf_cpu_context *cpuctx, int cpu) + struct perf_cpu_context *cpuctx) { + int cpu = smp_processor_id(); struct perf_event *event; int can_add_hw = 1; @@ -1326,24 +1327,24 @@ __perf_event_sched_in(struct perf_event_context *ctx, * accessing the event control register. If a NMI hits, then it will * keep the event running. */ -void perf_event_task_sched_in(struct task_struct *task, int cpu) +void perf_event_task_sched_in(struct task_struct *task) { - struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_event_context *ctx = task->perf_event_ctxp; if (likely(!ctx)) return; if (cpuctx->task_ctx == ctx) return; - __perf_event_sched_in(ctx, cpuctx, cpu); + __perf_event_sched_in(ctx, cpuctx); cpuctx->task_ctx = ctx; } -static void perf_event_cpu_sched_in(struct perf_cpu_context *cpuctx, int cpu) +static void perf_event_cpu_sched_in(struct perf_cpu_context *cpuctx) { struct perf_event_context *ctx = &cpuctx->ctx; - __perf_event_sched_in(ctx, cpuctx, cpu); + __perf_event_sched_in(ctx, cpuctx); } #define MAX_INTERRUPTS (~0ULL) @@ -1461,7 +1462,7 @@ static void rotate_ctx(struct perf_event_context *ctx) raw_spin_unlock(&ctx->lock); } -void perf_event_task_tick(struct task_struct *curr, int cpu) +void perf_event_task_tick(struct task_struct *curr) { struct perf_cpu_context *cpuctx; struct perf_event_context *ctx; @@ -1469,7 +1470,7 @@ void perf_event_task_tick(struct task_struct *curr, int cpu) if (!atomic_read(&nr_events)) return; - cpuctx = &per_cpu(perf_cpu_context, cpu); + cpuctx = &__get_cpu_var(perf_cpu_context); ctx = curr->perf_event_ctxp; perf_ctx_adjust_freq(&cpuctx->ctx); @@ -1484,9 +1485,9 @@ void perf_event_task_tick(struct task_struct *curr, int cpu) if (ctx) rotate_ctx(ctx); - perf_event_cpu_sched_in(cpuctx, cpu); + perf_event_cpu_sched_in(cpuctx); if (ctx) - perf_event_task_sched_in(curr, cpu); + perf_event_task_sched_in(curr); } /* @@ -1527,7 +1528,7 @@ static void perf_event_enable_on_exec(struct task_struct *task) raw_spin_unlock(&ctx->lock); - perf_event_task_sched_in(task, smp_processor_id()); + perf_event_task_sched_in(task); out: local_irq_restore(flags); } diff --git a/kernel/sched.c b/kernel/sched.c index 18cceee..d6527ac 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2752,7 +2752,7 @@ static void finish_task_switch(struct rq *rq, struct task_struct *prev) */ prev_state = prev->state; finish_arch_switch(prev); - perf_event_task_sched_in(current, cpu_of(rq)); + perf_event_task_sched_in(current); finish_lock_switch(rq, prev); fire_sched_in_preempt_notifiers(current); @@ -5266,7 +5266,7 @@ void scheduler_tick(void) curr->sched_class->task_tick(rq, curr, 0); raw_spin_unlock(&rq->lock); - perf_event_task_tick(curr, cpu); + perf_event_task_tick(curr); #ifdef CONFIG_SMP rq->idle_at_tick = idle_cpu(cpu); @@ -5480,7 +5480,7 @@ need_resched_nonpreemptible: if (likely(prev != next)) { sched_info_switch(prev, next); - perf_event_task_sched_out(prev, next, cpu); + perf_event_task_sched_out(prev, next); rq->nr_switches++; rq->curr = next; -- cgit v0.10.2 From fd2a50a0240f5f5b59070474eabd83a85720a406 Mon Sep 17 00:00:00 2001 From: Naga Chumbalkar Date: Thu, 24 Dec 2009 01:54:47 +0000 Subject: x86, perfctr: Remove unused func avail_to_resrv_perfctr_nmi() avail_to_resrv_perfctr_nmi() is neither EXPORT'd, nor used in the file. So remove it. Signed-off-by: Naga Chumbalkar Acked-by: Cyrill Gorcunov Cc: oprofile-list@lists.sf.net LKML-Reference: <20091224015441.6005.4408.sendpatchset@localhost.localdomain> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/nmi.h b/arch/x86/include/asm/nmi.h index 139d4c1..93da9c3 100644 --- a/arch/x86/include/asm/nmi.h +++ b/arch/x86/include/asm/nmi.h @@ -19,7 +19,6 @@ extern void die_nmi(char *str, struct pt_regs *regs, int do_panic); extern int check_nmi_watchdog(void); extern int nmi_watchdog_enabled; extern int avail_to_resrv_perfctr_nmi_bit(unsigned int); -extern int avail_to_resrv_perfctr_nmi(unsigned int); extern int reserve_perfctr_nmi(unsigned int); extern void release_perfctr_nmi(unsigned int); extern int reserve_evntsel_nmi(unsigned int); diff --git a/arch/x86/kernel/cpu/perfctr-watchdog.c b/arch/x86/kernel/cpu/perfctr-watchdog.c index 898df97..74f4e85 100644 --- a/arch/x86/kernel/cpu/perfctr-watchdog.c +++ b/arch/x86/kernel/cpu/perfctr-watchdog.c @@ -115,17 +115,6 @@ int avail_to_resrv_perfctr_nmi_bit(unsigned int counter) return !test_bit(counter, perfctr_nmi_owner); } - -/* checks the an msr for availability */ -int avail_to_resrv_perfctr_nmi(unsigned int msr) -{ - unsigned int counter; - - counter = nmi_perfctr_msr_to_bit(msr); - BUG_ON(counter > NMI_MAX_COUNTER_BITS); - - return !test_bit(counter, perfctr_nmi_owner); -} EXPORT_SYMBOL(avail_to_resrv_perfctr_nmi_bit); int reserve_perfctr_nmi(unsigned int msr) -- cgit v0.10.2 From 34d0b5af50a063cded842716633501b38ff815fb Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 28 Dec 2009 17:53:47 +0900 Subject: sh: Convert ptrace to hw_breakpoint API. This is the initial step for converting singlestep handling via ptrace over to hw_breakpoints. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/ptrace.h b/arch/sh/include/asm/ptrace.h index 1dc12cb..201d11e 100644 --- a/arch/sh/include/asm/ptrace.h +++ b/arch/sh/include/asm/ptrace.h @@ -124,6 +124,12 @@ struct task_struct; extern void user_enable_single_step(struct task_struct *); extern void user_disable_single_step(struct task_struct *); +struct perf_event; +struct perf_sample_data; + +extern void ptrace_triggered(struct perf_event *bp, int nmi, + struct perf_sample_data *data, struct pt_regs *regs); + #define task_pt_regs(task) \ ((struct pt_regs *) (task_stack_page(task) + THREAD_SIZE) - 1) diff --git a/arch/sh/kernel/hw_breakpoint.c b/arch/sh/kernel/hw_breakpoint.c index 022d8ed..c515a3e 100644 --- a/arch/sh/kernel/hw_breakpoint.c +++ b/arch/sh/kernel/hw_breakpoint.c @@ -20,6 +20,7 @@ #include #include #include +#include struct ubc_context { unsigned long pc; @@ -372,7 +373,7 @@ static int __kprobes hw_breakpoint_handler(struct die_args *args) rcu_read_unlock(); } - if (bp) { + if (bp && bp->overflow_handler != ptrace_triggered) { struct arch_hw_breakpoint *info = counter_arch_bp(bp); __raw_writel(UBC_CBR_CE | info->len | info->type, UBC_CBR0); @@ -387,9 +388,19 @@ static int __kprobes hw_breakpoint_handler(struct die_args *args) BUILD_TRAP_HANDLER(breakpoint) { unsigned long ex = lookup_exception_vector(); + siginfo_t info; + int err; TRAP_HANDLER_DECL; - notify_die(DIE_BREAKPOINT, "breakpoint", regs, 0, ex, SIGTRAP); + err = notify_die(DIE_BREAKPOINT, "breakpoint", regs, 0, ex, SIGTRAP); + if (err == NOTIFY_STOP) + return; + + /* Deliver the signal to userspace */ + info.si_signo = SIGTRAP; + info.si_errno = 0; + info.si_code = TRAP_HWBKPT; + force_sig_info(SIGTRAP, &info, current); } /* diff --git a/arch/sh/kernel/ptrace_32.c b/arch/sh/kernel/ptrace_32.c index bdb1044..8e094c4 100644 --- a/arch/sh/kernel/ptrace_32.c +++ b/arch/sh/kernel/ptrace_32.c @@ -2,7 +2,7 @@ * SuperH process tracing * * Copyright (C) 1999, 2000 Kaz Kojima & Niibe Yutaka - * Copyright (C) 2002 - 2008 Paul Mundt + * Copyright (C) 2002 - 2009 Paul Mundt * * Audit support by Yuichi Nakamura * @@ -26,6 +26,7 @@ #include #include #include +#include #include #include #include @@ -63,9 +64,59 @@ static inline int put_stack_long(struct task_struct *task, int offset, return 0; } +void ptrace_triggered(struct perf_event *bp, int nmi, + struct perf_sample_data *data, struct pt_regs *regs) +{ + struct perf_event_attr attr; + + /* + * Disable the breakpoint request here since ptrace has defined a + * one-shot behaviour for breakpoint exceptions. + */ + attr = bp->attr; + attr.disabled = true; + modify_user_hw_breakpoint(bp, &attr); +} + +static int set_single_step(struct task_struct *tsk, unsigned long addr) +{ + struct thread_struct *thread = &tsk->thread; + struct perf_event *bp; + struct perf_event_attr attr; + + bp = thread->ptrace_bps[0]; + if (!bp) { + hw_breakpoint_init(&attr); + + attr.bp_addr = addr; + attr.bp_len = HW_BREAKPOINT_LEN_2; + attr.bp_type = HW_BREAKPOINT_R; + + bp = register_user_hw_breakpoint(&attr, ptrace_triggered, tsk); + if (IS_ERR(bp)) + return PTR_ERR(bp); + + thread->ptrace_bps[0] = bp; + } else { + int err; + + attr = bp->attr; + attr.bp_addr = addr; + err = modify_user_hw_breakpoint(bp, &attr); + if (unlikely(err)) + return err; + } + + return 0; +} + void user_enable_single_step(struct task_struct *child) { + unsigned long pc = get_stack_long(child, offsetof(struct pt_regs, pc)); + set_tsk_thread_flag(child, TIF_SINGLESTEP); + + set_single_step(child, pc); } void user_disable_single_step(struct task_struct *child) -- cgit v0.10.2 From 659d8cfbb225f1fa5a4f8671a847ef3ab5a89660 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Sat, 19 Dec 2009 16:40:28 -0500 Subject: perf tools: Do a few more directory handling optimizations A few more optimizations for perf when dealing with directories. Some of them significantly cut down the work which has to be done. d_type should always be set; otherwise fix the kernel code. And there are functions available to parse fstab-like files, so use them. Signed-off-by: Ulrich Drepper Acked-by: Pekka Enberg Cc: a.p.zijlstra@chello.nl Cc: acme@redhat.com Cc: eranian@google.com Cc: fweisbec@gmail.com Cc: lizf@cn.fujitsu.com Cc: paulus@samba.org Cc: xiaoguangrong@cn.fujitsu.com LKML-Reference: <200912192140.nBJLeSfA028905@hs20-bc2-1.build.redhat.com> [ v2: two small stylistic fixlets ] Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 4c06828..05dc5a7 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -92,23 +92,18 @@ static void setup_cpunode_map(void) if (!dir1) return; - while (true) { - dent1 = readdir(dir1); - if (!dent1) - break; - - if (sscanf(dent1->d_name, "node%u", &mem) < 1) + while ((dent1 = readdir(dir1)) != NULL) { + if (dent1->d_type != DT_DIR || + sscanf(dent1->d_name, "node%u", &mem) < 1) continue; snprintf(buf, PATH_MAX, "%s/%s", PATH_SYS_NODE, dent1->d_name); dir2 = opendir(buf); if (!dir2) continue; - while (true) { - dent2 = readdir(dir2); - if (!dent2) - break; - if (sscanf(dent2->d_name, "cpu%u", &cpu) < 1) + while ((dent2 = readdir(dir2)) != NULL) { + if (dent2->d_type != DT_LNK || + sscanf(dent2->d_name, "cpu%u", &cpu) < 1) continue; cpunode_map[cpu] = mem; } diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c index cace355..dfef238 100644 --- a/tools/perf/util/trace-event-info.c +++ b/tools/perf/util/trace-event-info.c @@ -20,6 +20,7 @@ */ #define _GNU_SOURCE #include +#include #include #include #include @@ -103,28 +104,28 @@ static const char *find_debugfs(void) { static char debugfs[MAX_PATH+1]; static int debugfs_found; - char type[100]; FILE *fp; + struct mntent *m; if (debugfs_found) return debugfs; - if ((fp = fopen("/proc/mounts","r")) == NULL) + fp = setmntent("/proc/mounts", "r"); + if (!fp) die("Can't open /proc/mounts for read"); - while (fscanf(fp, "%*s %" - STR(MAX_PATH) - "s %99s %*s %*d %*d\n", - debugfs, type) == 2) { - if (strcmp(type, "debugfs") == 0) + while ((m = getmntent(fp)) != NULL) { + if (strcmp(m->mnt_type, "debugfs") == 0) { + strcpy(debugfs, m->mnt_dir); + debugfs_found = 1; break; + } } - fclose(fp); - if (strcmp(type, "debugfs") != 0) - die("debugfs not mounted, please mount"); + endmntent(fp); - debugfs_found = 1; + if (!debugfs_found) + die("debugfs not mounted, please mount"); return debugfs; } @@ -317,7 +318,8 @@ static void copy_event_system(const char *sys, struct tracepoint_path *tps) die("can't read directory '%s'", sys); while ((dent = readdir(dir))) { - if (strcmp(dent->d_name, ".") == 0 || + if (dent->d_type != DT_DIR || + strcmp(dent->d_name, ".") == 0 || strcmp(dent->d_name, "..") == 0 || !name_in_tp_list(dent->d_name, tps)) continue; @@ -334,7 +336,8 @@ static void copy_event_system(const char *sys, struct tracepoint_path *tps) rewinddir(dir); while ((dent = readdir(dir))) { - if (strcmp(dent->d_name, ".") == 0 || + if (dent->d_type != DT_DIR || + strcmp(dent->d_name, ".") == 0 || strcmp(dent->d_name, "..") == 0 || !name_in_tp_list(dent->d_name, tps)) continue; @@ -394,26 +397,21 @@ static void read_event_files(struct tracepoint_path *tps) die("can't read directory '%s'", path); while ((dent = readdir(dir))) { - if (strcmp(dent->d_name, ".") == 0 || + if (dent->d_type != DT_DIR || + strcmp(dent->d_name, ".") == 0 || strcmp(dent->d_name, "..") == 0 || strcmp(dent->d_name, "ftrace") == 0 || !system_in_tp_list(dent->d_name, tps)) continue; - sys = malloc_or_die(strlen(path) + strlen(dent->d_name) + 2); - sprintf(sys, "%s/%s", path, dent->d_name); - ret = stat(sys, &st); - free(sys); - if (ret < 0) - continue; - if (S_ISDIR(st.st_mode)) - count++; + count++; } write_or_die(&count, 4); rewinddir(dir); while ((dent = readdir(dir))) { - if (strcmp(dent->d_name, ".") == 0 || + if (dent->d_type != DT_DIR || + strcmp(dent->d_name, ".") == 0 || strcmp(dent->d_name, "..") == 0 || strcmp(dent->d_name, "ftrace") == 0 || !system_in_tp_list(dent->d_name, tps)) @@ -422,10 +420,8 @@ static void read_event_files(struct tracepoint_path *tps) sprintf(sys, "%s/%s", path, dent->d_name); ret = stat(sys, &st); if (ret >= 0) { - if (S_ISDIR(st.st_mode)) { - write_or_die(dent->d_name, strlen(dent->d_name) + 1); - copy_event_system(sys, tps); - } + write_or_die(dent->d_name, strlen(dent->d_name) + 1); + copy_event_system(sys, tps); } free(sys); } -- cgit v0.10.2 From 07b139c8c81b97bbe55c68daf0cbeca8b1c609ca Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Mon, 21 Dec 2009 14:27:35 +0800 Subject: perf events: Remove CONFIG_EVENT_PROFILE Quoted from Ingo: | This reminds me - i think we should eliminate CONFIG_EVENT_PROFILE - | it's an unnecessary Kconfig complication. If both PERF_EVENTS and | EVENT_TRACING is enabled we should expose generic tracepoints. | | Nor is it limited to event 'profiling', so it has become a misnomer as | well. Signed-off-by: Li Zefan Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <4B2F1557.2050705@cn.fujitsu.com> Signed-off-by: Ingo Molnar diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h index 2233c98..0a09e75 100644 --- a/include/linux/ftrace_event.h +++ b/include/linux/ftrace_event.h @@ -188,7 +188,7 @@ do { \ __trace_printk(ip, fmt, ##args); \ } while (0) -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS struct perf_event; extern int ftrace_profile_enable(int event_id); extern void ftrace_profile_disable(int event_id); diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index a494e75..9a1d276 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -658,7 +658,7 @@ struct perf_event { perf_overflow_handler_t overflow_handler; -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_EVENT_TRACING struct event_filter *filter; #endif diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index 65793e9..b7c7fcf 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -99,7 +99,7 @@ struct perf_event_attr; #define __SC_TEST5(t5, a5, ...) __SC_TEST(t5); __SC_TEST4(__VA_ARGS__) #define __SC_TEST6(t6, a6, ...) __SC_TEST(t6); __SC_TEST5(__VA_ARGS__) -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS #define TRACE_SYS_ENTER_PROFILE_INIT(sname) \ .profile_enable = prof_sysenter_enable, \ @@ -113,7 +113,7 @@ struct perf_event_attr; #define TRACE_SYS_ENTER_PROFILE_INIT(sname) #define TRACE_SYS_EXIT_PROFILE(sname) #define TRACE_SYS_EXIT_PROFILE_INIT(sname) -#endif +#endif /* CONFIG_PERF_EVENTS */ #ifdef CONFIG_FTRACE_SYSCALLS #define __SC_STR_ADECL1(t, a) #a diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h index 73523151..2fdd36d 100644 --- a/include/trace/ftrace.h +++ b/include/trace/ftrace.h @@ -498,7 +498,7 @@ static inline int ftrace_get_offsets_##call( \ #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS /* * Generate the functions needed for tracepoint perf_event support. @@ -541,7 +541,7 @@ static void ftrace_profile_disable_##name(struct ftrace_event_call *unused)\ #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) -#endif +#endif /* CONFIG_PERF_EVENTS */ /* * Stage 4 of the trace events. @@ -626,7 +626,7 @@ static void ftrace_profile_disable_##name(struct ftrace_event_call *unused)\ * */ -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS #define _TRACE_PROFILE_INIT(call) \ .profile_enable = ftrace_profile_enable_##call, \ @@ -634,7 +634,7 @@ static void ftrace_profile_disable_##name(struct ftrace_event_call *unused)\ #else #define _TRACE_PROFILE_INIT(call) -#endif +#endif /* CONFIG_PERF_EVENTS */ #undef __entry #define __entry entry @@ -834,7 +834,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ * } */ -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS #undef __perf_addr #define __perf_addr(a) __addr = (a) @@ -926,7 +926,7 @@ static void ftrace_profile_##call(proto) \ DEFINE_EVENT(template, name, PARAMS(proto), PARAMS(args)) #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) -#endif /* CONFIG_EVENT_PROFILE */ +#endif /* CONFIG_PERF_EVENTS */ #undef _TRACE_PROFILE_INIT diff --git a/include/trace/syscall.h b/include/trace/syscall.h index 961fda3..3d463dc 100644 --- a/include/trace/syscall.h +++ b/include/trace/syscall.h @@ -49,12 +49,12 @@ ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s); enum print_line_t print_syscall_enter(struct trace_iterator *iter, int flags); enum print_line_t print_syscall_exit(struct trace_iterator *iter, int flags); #endif -#ifdef CONFIG_EVENT_PROFILE + +#ifdef CONFIG_PERF_EVENTS int prof_sysenter_enable(struct ftrace_event_call *call); void prof_sysenter_disable(struct ftrace_event_call *call); int prof_sysexit_enable(struct ftrace_event_call *call); void prof_sysexit_disable(struct ftrace_event_call *call); - #endif #endif /* _TRACE_SYSCALL_H */ diff --git a/init/Kconfig b/init/Kconfig index a23da9f..06dab27 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -966,19 +966,6 @@ config PERF_EVENTS Say Y if unsure. -config EVENT_PROFILE - bool "Tracepoint profiling sources" - depends on PERF_EVENTS && EVENT_TRACING - default y - help - Allow the use of tracepoints as software performance events. - - When this is enabled, you can create perf events based on - tracepoints using PERF_TYPE_TRACEPOINT and the tracepoint ID - found in debugfs://tracing/events/*/*/id. (The -e/--events - option to the perf tool can parse and interpret symbolic - tracepoints, in the subsystem:tracepoint_name format.) - config PERF_COUNTERS bool "Kernel performance counters (old config option)" depends on HAVE_PERF_EVENTS diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 099bd66..5b987b4 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -4177,7 +4177,7 @@ static const struct pmu perf_ops_task_clock = { .read = task_clock_perf_event_read, }; -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_EVENT_TRACING void perf_tp_event(int event_id, u64 addr, u64 count, void *record, int entry_size) @@ -4282,7 +4282,7 @@ static void perf_event_free_filter(struct perf_event *event) { } -#endif /* CONFIG_EVENT_PROFILE */ +#endif /* CONFIG_EVENT_TRACING */ #ifdef CONFIG_HAVE_HW_BREAKPOINT static void bp_perf_event_destroy(struct perf_event *event) diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index cd9ecd8..d00c6fe 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -51,7 +51,9 @@ endif obj-$(CONFIG_EVENT_TRACING) += trace_events.o obj-$(CONFIG_EVENT_TRACING) += trace_export.o obj-$(CONFIG_FTRACE_SYSCALLS) += trace_syscalls.o -obj-$(CONFIG_EVENT_PROFILE) += trace_event_profile.o +ifeq ($(CONFIG_PERF_EVENTS),y) +obj-$(CONFIG_EVENT_TRACING) += trace_event_profile.o +endif obj-$(CONFIG_EVENT_TRACING) += trace_events_filter.o obj-$(CONFIG_KPROBE_EVENT) += trace_kprobe.o obj-$(CONFIG_KSYM_TRACER) += trace_ksym.o diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 50504cb..74563d7 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -1360,7 +1360,7 @@ out_unlock: return err; } -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS void ftrace_profile_free_filter(struct perf_event *event) { @@ -1428,5 +1428,5 @@ out_unlock: return err; } -#endif /* CONFIG_EVENT_PROFILE */ +#endif /* CONFIG_PERF_EVENTS */ diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 375f81a..75d75de 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1249,7 +1249,7 @@ static int kretprobe_event_show_format(struct ftrace_event_call *call, ", REC->" FIELD_STRING_RETIP); } -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS /* Kprobe profile handler */ static __kprobes int kprobe_profile_func(struct kprobe *kp, @@ -1407,7 +1407,7 @@ static void probe_profile_disable(struct ftrace_event_call *call) disable_kprobe(&tp->rp.kp); } } -#endif /* CONFIG_EVENT_PROFILE */ +#endif /* CONFIG_PERF_EVENTS */ static __kprobes @@ -1417,10 +1417,10 @@ int kprobe_dispatcher(struct kprobe *kp, struct pt_regs *regs) if (tp->flags & TP_FLAG_TRACE) kprobe_trace_func(kp, regs); -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS if (tp->flags & TP_FLAG_PROFILE) kprobe_profile_func(kp, regs); -#endif /* CONFIG_EVENT_PROFILE */ +#endif return 0; /* We don't tweek kernel, so just return 0 */ } @@ -1431,10 +1431,10 @@ int kretprobe_dispatcher(struct kretprobe_instance *ri, struct pt_regs *regs) if (tp->flags & TP_FLAG_TRACE) kretprobe_trace_func(ri, regs); -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS if (tp->flags & TP_FLAG_PROFILE) kretprobe_profile_func(ri, regs); -#endif /* CONFIG_EVENT_PROFILE */ +#endif return 0; /* We don't tweek kernel, so just return 0 */ } @@ -1463,7 +1463,7 @@ static int register_probe_event(struct trace_probe *tp) call->regfunc = probe_event_enable; call->unregfunc = probe_event_disable; -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS call->profile_enable = probe_profile_enable; call->profile_disable = probe_profile_disable; #endif diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 75289f3..f694f66 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -421,7 +421,7 @@ int __init init_ftrace_syscalls(void) } core_initcall(init_ftrace_syscalls); -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS static DECLARE_BITMAP(enabled_prof_enter_syscalls, NR_syscalls); static DECLARE_BITMAP(enabled_prof_exit_syscalls, NR_syscalls); @@ -626,6 +626,5 @@ void prof_sysexit_disable(struct ftrace_event_call *call) mutex_unlock(&syscall_trace_lock); } -#endif - +#endif /* CONFIG_PERF_EVENTS */ -- cgit v0.10.2 From 29c52aa2300173dd45df04dae1f5acc81a2c93b1 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Mon, 28 Dec 2009 16:47:12 +0800 Subject: perf tools: Mount debugfs automatically Mount debugfs filesystem under '/sys/kernel/debug', if it's not mounted. Signed-off-by: Xiao Guangrong Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker Cc: Clark Williams Cc: John Kacur LKML-Reference: <4B387090.7080407@cn.fujitsu.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/perf.c b/tools/perf/perf.c index 873e55f..fc89005 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -388,7 +388,7 @@ static int run_argv(int *argcp, const char ***argv) /* mini /proc/mounts parser: searching for "^blah /mount/point debugfs" */ static void get_debugfs_mntpt(void) { - const char *path = debugfs_find_mountpoint(); + const char *path = debugfs_mount(NULL); if (path) strncpy(debugfs_mntpt, path, sizeof(debugfs_mntpt)); diff --git a/tools/perf/util/debugfs.c b/tools/perf/util/debugfs.c index 06b73ee..1f805fd 100644 --- a/tools/perf/util/debugfs.c +++ b/tools/perf/util/debugfs.c @@ -106,16 +106,14 @@ int debugfs_valid_entry(const char *path) return 0; } -/* mount the debugfs somewhere */ +/* mount the debugfs somewhere if it's not mounted */ -int debugfs_mount(const char *mountpoint) +char *debugfs_mount(const char *mountpoint) { - char mountcmd[128]; - /* see if it's already mounted */ if (debugfs_find_mountpoint()) { debugfs_premounted = 1; - return 0; + return debugfs_mountpoint; } /* if not mounted and no argument */ @@ -127,13 +125,13 @@ int debugfs_mount(const char *mountpoint) mountpoint = "/sys/kernel/debug"; } + if (mount(NULL, mountpoint, "debugfs", 0, NULL) < 0) + return NULL; + /* save the mountpoint */ strncpy(debugfs_mountpoint, mountpoint, sizeof(debugfs_mountpoint)); - /* mount it */ - snprintf(mountcmd, sizeof(mountcmd), - "/bin/mount -t debugfs debugfs %s", mountpoint); - return system(mountcmd); + return debugfs_mountpoint; } /* umount the debugfs */ diff --git a/tools/perf/util/debugfs.h b/tools/perf/util/debugfs.h index 3cd14f9..83a0287 100644 --- a/tools/perf/util/debugfs.h +++ b/tools/perf/util/debugfs.h @@ -15,7 +15,7 @@ extern const char *debugfs_find_mountpoint(void); extern int debugfs_valid_mountpoint(const char *debugfs); extern int debugfs_valid_entry(const char *path); -extern int debugfs_mount(const char *mountpoint); +extern char *debugfs_mount(const char *mountpoint); extern int debugfs_umount(void); extern int debugfs_write(const char *entry, const char *value); extern int debugfs_read(const char *entry, char *buffer, size_t size); -- cgit v0.10.2 From 61be3e59ba7a6dbd39f92fd1f107285a0caeb008 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Mon, 28 Dec 2009 16:48:30 +0800 Subject: perf trace: Clean up find_debugfs() Remove redundant code for 'perf trace' Signed-off-by: Xiao Guangrong Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker Cc: Clark Williams Cc: John Kacur LKML-Reference: <4B3870DE.7090500@cn.fujitsu.com> [ v2: resolved conflicts with recent changes ] Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/debugfs.c b/tools/perf/util/debugfs.c index 1f805fd..a88fefc 100644 --- a/tools/perf/util/debugfs.c +++ b/tools/perf/util/debugfs.c @@ -130,6 +130,7 @@ char *debugfs_mount(const char *mountpoint) /* save the mountpoint */ strncpy(debugfs_mountpoint, mountpoint, sizeof(debugfs_mountpoint)); + debugfs_found = 1; return debugfs_mountpoint; } diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c index dfef238..535176d 100644 --- a/tools/perf/util/trace-event-info.c +++ b/tools/perf/util/trace-event-info.c @@ -38,6 +38,7 @@ #include "../perf.h" #include "trace-event.h" +#include "debugfs.h" #define VERSION "0.5" @@ -102,32 +103,12 @@ void *malloc_or_die(unsigned int size) static const char *find_debugfs(void) { - static char debugfs[MAX_PATH+1]; - static int debugfs_found; - FILE *fp; - struct mntent *m; - - if (debugfs_found) - return debugfs; - - fp = setmntent("/proc/mounts", "r"); - if (!fp) - die("Can't open /proc/mounts for read"); - - while ((m = getmntent(fp)) != NULL) { - if (strcmp(m->mnt_type, "debugfs") == 0) { - strcpy(debugfs, m->mnt_dir); - debugfs_found = 1; - break; - } - } - - endmntent(fp); + const char *path = debugfs_mount(NULL); - if (!debugfs_found) - die("debugfs not mounted, please mount"); + if (!path) + die("Your kernel not support debugfs filesystem"); - return debugfs; + return path; } /* -- cgit v0.10.2 From 9967411e5b324a908e344d6ce66b77bd5d372c3e Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Mon, 28 Dec 2009 16:49:38 +0800 Subject: perf trace: Fix forgotten close of file/dir Signed-off-by: Xiao Guangrong Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker Cc: Clark Williams Cc: John Kacur LKML-Reference: <4B387122.7090801@cn.fujitsu.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c index 535176d..407fd65 100644 --- a/tools/perf/util/trace-event-info.c +++ b/tools/perf/util/trace-event-info.c @@ -253,6 +253,8 @@ static void read_header_files(void) write_or_die("header_page", 12); write_or_die(&size, 8); check_size = copy_file_fd(fd); + close(fd); + if (size != check_size) die("wrong size for '%s' size=%lld read=%lld", path, size, check_size); @@ -271,6 +273,7 @@ static void read_header_files(void) if (size != check_size) die("wrong size for '%s'", path); put_tracing_file(path); + close(fd); } static bool name_in_tp_list(char *sys, struct tracepoint_path *tps) @@ -337,6 +340,7 @@ static void copy_event_system(const char *sys, struct tracepoint_path *tps) free(format); } + closedir(dir); } static void read_ftrace_files(struct tracepoint_path *tps) @@ -407,6 +411,7 @@ static void read_event_files(struct tracepoint_path *tps) free(sys); } + closedir(dir); put_tracing_file(path); } -- cgit v0.10.2 From 39d997b514e12d5aff0dca206eb8996b3957927e Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Mon, 21 Dec 2009 16:20:16 -0800 Subject: x86, core: Optimize hweight32() Optimize hweight32 by using the same technique in hweight64. The proof of this technique can be found in the commit log for f9b4192923fa6e38331e88214b1fe5fc21583fcc ("bitops: hweight() speedup"). The userspace benchmark on x86_32 showed 20% speedup with bitmap_weight() which uses hweight32 to count bits for each unsigned long on 32bit architectures. int main(void) { #define SZ (1024 * 1024 * 512) static DECLARE_BITMAP(bitmap, SZ) = { [0 ... 100] = 1, }; return bitmap_weight(bitmap, SZ); } Signed-off-by: Akinobu Mita Signed-off-by: Andrew Morton Cc: Linus Torvalds LKML-Reference: <1258603932-4590-1-git-send-email-akinobu.mita@gmail.com> [ only x86 sets ARCH_HAS_FAST_MULTIPLIER so we do this via the x86 tree] Signed-off-by: Ingo Molnar diff --git a/lib/hweight.c b/lib/hweight.c index 389424e..63ee4eb 100644 --- a/lib/hweight.c +++ b/lib/hweight.c @@ -11,11 +11,18 @@ unsigned int hweight32(unsigned int w) { +#ifdef ARCH_HAS_FAST_MULTIPLIER + w -= (w >> 1) & 0x55555555; + w = (w & 0x33333333) + ((w >> 2) & 0x33333333); + w = (w + (w >> 4)) & 0x0f0f0f0f; + return (w * 0x01010101) >> 24; +#else unsigned int res = w - ((w >> 1) & 0x55555555); res = (res & 0x33333333) + ((res >> 2) & 0x33333333); res = (res + (res >> 4)) & 0x0F0F0F0F; res = res + (res >> 8); return (res + (res >> 16)) & 0x000000FF; +#endif } EXPORT_SYMBOL(hweight32); -- cgit v0.10.2 From d894837f23f491aa7ed167aae767fc07cfe6e6e6 Mon Sep 17 00:00:00 2001 From: Simon Kagstrom Date: Wed, 23 Dec 2009 11:08:18 +0100 Subject: sched: might_sleep(): Make file parameter const char * Fixes a warning when building with g++: warning: deprecated conversion from string constant to 'char*' And the file parameter use is constant, so mark it as such. Signed-off-by: Simon Kagstrom Cc: peterz@infradead.org LKML-Reference: <20091223110818.442d848e@marrow.netinsight.se> Signed-off-by: Ingo Molnar diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 3fc9f5a..785d7d1 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h @@ -124,7 +124,7 @@ extern int _cond_resched(void); #endif #ifdef CONFIG_DEBUG_SPINLOCK_SLEEP - void __might_sleep(char *file, int line, int preempt_offset); + void __might_sleep(const char *file, int line, int preempt_offset); /** * might_sleep - annotation for functions that can sleep * @@ -138,7 +138,8 @@ extern int _cond_resched(void); # define might_sleep() \ do { __might_sleep(__FILE__, __LINE__, 0); might_resched(); } while (0) #else - static inline void __might_sleep(char *file, int line, int preempt_offset) { } + static inline void __might_sleep(const char *file, int line, + int preempt_offset) { } # define might_sleep() do { might_resched(); } while (0) #endif diff --git a/kernel/sched.c b/kernel/sched.c index c535cc4..64298a5 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -9694,7 +9694,7 @@ static inline int preempt_count_equals(int preempt_offset) return (nested == PREEMPT_INATOMIC_BASE + preempt_offset); } -void __might_sleep(char *file, int line, int preempt_offset) +void __might_sleep(const char *file, int line, int preempt_offset) { #ifdef in_atomic static unsigned long prev_jiffy; /* ratelimiting */ -- cgit v0.10.2 From 499a5f1efa0b0ac56ec5d060412aed84ae68e63e Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Fri, 18 Dec 2009 16:05:51 +0000 Subject: x86: Lift restriction on the location of FIX_BTMAP_* The early ioremap fixmap entries cover half (or for 32-bit non-PAE, a quarter) of a page table, yet they got uncondtitionally aligned so far to a 256-entry boundary. This is not necessary if the range of page table entries anyway falls into a single page table. This buys back, for (theoretically) 50% of all configurations (25% of all non-PAE ones), at least some of the lowmem necessarily lost with commit e621bd18958ef5dbace3129ebe17a0a475e127d9. Signed-off-by: Jan Beulich Cc: Linus Torvalds Cc: Andrew Morton LKML-Reference: <4B2BB66F0200007800026AD6@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/fixmap.h b/arch/x86/include/asm/fixmap.h index 14f9890..635f03b 100644 --- a/arch/x86/include/asm/fixmap.h +++ b/arch/x86/include/asm/fixmap.h @@ -118,14 +118,20 @@ enum fixed_addresses { * 256 temporary boot-time mappings, used by early_ioremap(), * before ioremap() is functional. * - * We round it up to the next 256 pages boundary so that we - * can have a single pgd entry and a single pte table: + * If necessary we round it up to the next 256 pages boundary so + * that we can have a single pgd entry and a single pte table: */ #define NR_FIX_BTMAPS 64 #define FIX_BTMAPS_SLOTS 4 - FIX_BTMAP_END = __end_of_permanent_fixed_addresses + 256 - - (__end_of_permanent_fixed_addresses & 255), - FIX_BTMAP_BEGIN = FIX_BTMAP_END + NR_FIX_BTMAPS*FIX_BTMAPS_SLOTS - 1, +#define TOTAL_FIX_BTMAPS (NR_FIX_BTMAPS * FIX_BTMAPS_SLOTS) + FIX_BTMAP_END = + (__end_of_permanent_fixed_addresses ^ + (__end_of_permanent_fixed_addresses + TOTAL_FIX_BTMAPS - 1)) & + -PTRS_PER_PTE + ? __end_of_permanent_fixed_addresses + TOTAL_FIX_BTMAPS - + (__end_of_permanent_fixed_addresses & (TOTAL_FIX_BTMAPS - 1)) + : __end_of_permanent_fixed_addresses, + FIX_BTMAP_BEGIN = FIX_BTMAP_END + TOTAL_FIX_BTMAPS - 1, #ifdef CONFIG_PROVIDE_OHCI1394_DMA_INIT FIX_OHCI1394_BASE, #endif diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c index c246d25..03c75ff 100644 --- a/arch/x86/mm/ioremap.c +++ b/arch/x86/mm/ioremap.c @@ -422,6 +422,10 @@ void __init early_ioremap_init(void) * The boot-ioremap range spans multiple pmds, for which * we are not prepared: */ +#define __FIXADDR_TOP (-PAGE_SIZE) + BUILD_BUG_ON((__fix_to_virt(FIX_BTMAP_BEGIN) >> PMD_SHIFT) + != (__fix_to_virt(FIX_BTMAP_END) >> PMD_SHIFT)); +#undef __FIXADDR_TOP if (pmd != early_ioremap_pmd(fix_to_virt(FIX_BTMAP_END))) { WARN_ON(1); printk(KERN_WARNING "pmd %p != %p\n", -- cgit v0.10.2 From 1b1d9258181bae199dc940f4bd0298126b9a73d9 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Fri, 18 Dec 2009 16:12:56 +0000 Subject: x86-64: Modify copy_user_generic() alternatives mechanism In order to avoid unnecessary chains of branches, rather than implementing copy_user_generic() as a function consisting of just a single (possibly patched) branch, instead properly deal with patching call instructions in the alternative instructions framework, and move the patching into the callers. As a follow-on, one could also introduce something like __EXPORT_SYMBOL_ALT() to avoid patching call sites in modules. Signed-off-by: Jan Beulich Cc: Nick Piggin Cc: Linus Torvalds Cc: Andrew Morton LKML-Reference: <4B2BB8180200007800026AE7@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/alternative.h b/arch/x86/include/asm/alternative.h index 69b74a7..3b5b828 100644 --- a/arch/x86/include/asm/alternative.h +++ b/arch/x86/include/asm/alternative.h @@ -125,11 +125,16 @@ static inline void alternatives_smp_switch(int smp) {} asm volatile (ALTERNATIVE(oldinstr, newinstr, feature) \ : output : "i" (0), ## input) +/* Like alternative_io, but for replacing a direct call with another one. */ +#define alternative_call(oldfunc, newfunc, feature, output, input...) \ + asm volatile (ALTERNATIVE("call %P[old]", "call %P[new]", feature) \ + : output : [old] "i" (oldfunc), [new] "i" (newfunc), ## input) + /* * use this macro(s) if you need more than one output parameter * in alternative_io */ -#define ASM_OUTPUT2(a, b) a, b +#define ASM_OUTPUT2(a...) a struct paravirt_patch_site; #ifdef CONFIG_PARAVIRT diff --git a/arch/x86/include/asm/uaccess_64.h b/arch/x86/include/asm/uaccess_64.h index 46324c6..a78c403 100644 --- a/arch/x86/include/asm/uaccess_64.h +++ b/arch/x86/include/asm/uaccess_64.h @@ -8,6 +8,8 @@ #include #include #include +#include +#include #include /* @@ -16,7 +18,24 @@ /* Handles exceptions in both to and from, but doesn't do access_ok */ __must_check unsigned long -copy_user_generic(void *to, const void *from, unsigned len); +copy_user_generic_string(void *to, const void *from, unsigned len); +__must_check unsigned long +copy_user_generic_unrolled(void *to, const void *from, unsigned len); + +static __always_inline __must_check unsigned long +copy_user_generic(void *to, const void *from, unsigned len) +{ + unsigned ret; + + alternative_call(copy_user_generic_unrolled, + copy_user_generic_string, + X86_FEATURE_REP_GOOD, + ASM_OUTPUT2("=a" (ret), "=D" (to), "=S" (from), + "=d" (len)), + "1" (to), "2" (from), "3" (len) + : "memory", "rcx", "r8", "r9", "r10", "r11"); + return ret; +} __must_check unsigned long _copy_to_user(void __user *to, const void *from, unsigned len); diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index de7353c..2589ea4 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -205,7 +205,7 @@ void __init_or_module apply_alternatives(struct alt_instr *start, struct alt_instr *end) { struct alt_instr *a; - char insnbuf[MAX_PATCH_LEN]; + u8 insnbuf[MAX_PATCH_LEN]; DPRINTK("%s: alt table %p -> %p\n", __func__, start, end); for (a = start; a < end; a++) { @@ -223,6 +223,8 @@ void __init_or_module apply_alternatives(struct alt_instr *start, } #endif memcpy(insnbuf, a->replacement, a->replacementlen); + if (*insnbuf == 0xe8 && a->replacementlen == 5) + *(s32 *)(insnbuf + 1) += a->replacement - a->instr; add_nops(insnbuf + a->replacementlen, a->instrlen - a->replacementlen); text_poke_early(instr, insnbuf, a->instrlen); diff --git a/arch/x86/kernel/x8664_ksyms_64.c b/arch/x86/kernel/x8664_ksyms_64.c index 619f7f8..693920b 100644 --- a/arch/x86/kernel/x8664_ksyms_64.c +++ b/arch/x86/kernel/x8664_ksyms_64.c @@ -26,7 +26,8 @@ EXPORT_SYMBOL(__put_user_2); EXPORT_SYMBOL(__put_user_4); EXPORT_SYMBOL(__put_user_8); -EXPORT_SYMBOL(copy_user_generic); +EXPORT_SYMBOL(copy_user_generic_string); +EXPORT_SYMBOL(copy_user_generic_unrolled); EXPORT_SYMBOL(__copy_user_nocache); EXPORT_SYMBOL(_copy_from_user); EXPORT_SYMBOL(_copy_to_user); diff --git a/arch/x86/lib/copy_user_64.S b/arch/x86/lib/copy_user_64.S index cf889d4..71100c9 100644 --- a/arch/x86/lib/copy_user_64.S +++ b/arch/x86/lib/copy_user_64.S @@ -90,12 +90,6 @@ ENTRY(_copy_from_user) CFI_ENDPROC ENDPROC(_copy_from_user) -ENTRY(copy_user_generic) - CFI_STARTPROC - ALTERNATIVE_JUMP X86_FEATURE_REP_GOOD,copy_user_generic_unrolled,copy_user_generic_string - CFI_ENDPROC -ENDPROC(copy_user_generic) - .section .fixup,"ax" /* must zero dest */ ENTRY(bad_from_user) -- cgit v0.10.2 From 7269e8812a59f74fb1ce134465d0bcf5683b93a1 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Fri, 18 Dec 2009 16:16:03 +0000 Subject: x86-64: Modify memcpy()/memset() alternatives mechanism In order to avoid unnecessary chains of branches, rather than implementing memcpy()/memset()'s access to their alternative implementations via a jump, patch the (larger) original function directly. The memcpy() part of this is slightly subtle: while alternative instruction patching does itself use memcpy(), with the replacement block being less than 64-bytes in size the main loop of the original function doesn't get used for copying memcpy_c() over memcpy(), and hence we can safely write over its beginning. Also note that the CFI annotations are fine for both variants of each of the functions. Signed-off-by: Jan Beulich Cc: Nick Piggin Cc: Linus Torvalds Cc: Andrew Morton LKML-Reference: <4B2BB8D30200007800026AF2@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/lib/memcpy_64.S b/arch/x86/lib/memcpy_64.S index ad5441e..f82e884 100644 --- a/arch/x86/lib/memcpy_64.S +++ b/arch/x86/lib/memcpy_64.S @@ -20,12 +20,11 @@ /* * memcpy_c() - fast string ops (REP MOVSQ) based variant. * - * Calls to this get patched into the kernel image via the + * This gets patched over the unrolled variant (below) via the * alternative instructions framework: */ - ALIGN -memcpy_c: - CFI_STARTPROC + .section .altinstr_replacement, "ax", @progbits +.Lmemcpy_c: movq %rdi, %rax movl %edx, %ecx @@ -35,8 +34,8 @@ memcpy_c: movl %edx, %ecx rep movsb ret - CFI_ENDPROC -ENDPROC(memcpy_c) +.Lmemcpy_e: + .previous ENTRY(__memcpy) ENTRY(memcpy) @@ -128,16 +127,10 @@ ENDPROC(__memcpy) * It is also a lot simpler. Use this when possible: */ - .section .altinstr_replacement, "ax" -1: .byte 0xeb /* jmp */ - .byte (memcpy_c - memcpy) - (2f - 1b) /* offset */ -2: - .previous - .section .altinstructions, "a" .align 8 .quad memcpy - .quad 1b + .quad .Lmemcpy_c .byte X86_FEATURE_REP_GOOD /* @@ -145,6 +138,6 @@ ENDPROC(__memcpy) * so it is silly to overwrite itself with nops - reboot is the * only outcome... */ - .byte 2b - 1b - .byte 2b - 1b + .byte .Lmemcpy_e - .Lmemcpy_c + .byte .Lmemcpy_e - .Lmemcpy_c .previous diff --git a/arch/x86/lib/memset_64.S b/arch/x86/lib/memset_64.S index 2c59481..e88d3b8 100644 --- a/arch/x86/lib/memset_64.S +++ b/arch/x86/lib/memset_64.S @@ -12,9 +12,8 @@ * * rax original destination */ - ALIGN -memset_c: - CFI_STARTPROC + .section .altinstr_replacement, "ax", @progbits +.Lmemset_c: movq %rdi,%r9 movl %edx,%r8d andl $7,%r8d @@ -29,8 +28,8 @@ memset_c: rep stosb movq %r9,%rax ret - CFI_ENDPROC -ENDPROC(memset_c) +.Lmemset_e: + .previous ENTRY(memset) ENTRY(__memset) @@ -118,16 +117,11 @@ ENDPROC(__memset) #include - .section .altinstr_replacement,"ax" -1: .byte 0xeb /* jmp */ - .byte (memset_c - memset) - (2f - 1b) /* offset */ -2: - .previous .section .altinstructions,"a" .align 8 .quad memset - .quad 1b + .quad .Lmemset_c .byte X86_FEATURE_REP_GOOD .byte .Lfinal - memset - .byte 2b - 1b + .byte .Lmemset_e - .Lmemset_c .previous -- cgit v0.10.2 From 41bdcb23dab22bf27361c5f2d89fe895d8904915 Mon Sep 17 00:00:00 2001 From: Liming Wang Date: Tue, 29 Dec 2009 16:37:07 +0800 Subject: perf tools: Unify event type description make event type description to a unified array and the array index consistent to perf_type_id. Signed-off-by: Liming Wang Cc: Frederic Weisbecker Cc: Masami Hiramatsu Cc: Paul Mackerras Cc: Peter Zijlstra LKML-Reference: <1262075829-16257-1-git-send-email-liming.wang@windriver.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index e5bc0fb..dc585a8 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -835,11 +835,12 @@ int parse_filter(const struct option *opt __used, const char *str, } static const char * const event_type_descriptors[] = { - "", "Hardware event", "Software event", "Tracepoint event", "Hardware cache event", + "Raw hardware event descriptor", + "Hardware breakpoint", }; /* @@ -872,7 +873,7 @@ static void print_tracepoint_events(void) snprintf(evt_path, MAXPATHLEN, "%s:%s", sys_dirent.d_name, evt_dirent.d_name); printf(" %-42s [%s]\n", evt_path, - event_type_descriptors[PERF_TYPE_TRACEPOINT+1]); + event_type_descriptors[PERF_TYPE_TRACEPOINT]); } closedir(evt_dir); } @@ -892,9 +893,7 @@ void print_events(void) printf("List of pre-defined events (to be used in -e):\n"); for (i = 0; i < ARRAY_SIZE(event_symbols); i++, syms++) { - type = syms->type + 1; - if (type >= ARRAY_SIZE(event_type_descriptors)) - type = 0; + type = syms->type; if (type != prev_type) printf("\n"); @@ -919,17 +918,19 @@ void print_events(void) for (i = 0; i < PERF_COUNT_HW_CACHE_RESULT_MAX; i++) { printf(" %-42s [%s]\n", event_cache_name(type, op, i), - event_type_descriptors[4]); + event_type_descriptors[PERF_TYPE_HW_CACHE]); } } } printf("\n"); - printf(" %-42s [raw hardware event descriptor]\n", - "rNNN"); + printf(" %-42s [%s]\n", + "rNNN", event_type_descriptors[PERF_TYPE_RAW]); printf("\n"); - printf(" %-42s [hardware breakpoint]\n", "mem:[:access]"); + printf(" %-42s [%s]\n", + "mem:[:access]", + event_type_descriptors[PERF_TYPE_BREAKPOINT]); printf("\n"); print_tracepoint_events(); -- cgit v0.10.2 From 63bbd5e2d539c9290b229c832f62d42aac23db94 Mon Sep 17 00:00:00 2001 From: Liming Wang Date: Tue, 29 Dec 2009 16:37:09 +0800 Subject: perf probe: Change CONFIG_KPROBE_TRACER to CONFIG_KPROBE_EVENT make the config name consistent Signed-off-by: Liming Wang Acked-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: Peter Zijlstra LKML-Reference: <1262075829-16257-3-git-send-email-liming.wang@windriver.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 29465d4..8e532d9 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -368,7 +368,7 @@ static int open_kprobe_events(int flags, int mode) if (ret < 0) { if (errno == ENOENT) die("kprobe_events file does not exist -" - " please rebuild with CONFIG_KPROBE_TRACER."); + " please rebuild with CONFIG_KPROBE_EVENT."); else die("Could not open kprobe_events file: %s", strerror(errno)); -- cgit v0.10.2 From 769885f372300a7fcfb9e54e4e2990718d40b529 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 28 Dec 2009 22:48:32 -0200 Subject: perf header: Do_read shouldn't die MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Propagate the errors instead, its callers already propagate other errors. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262047716-23171-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index df237c3..6b3cb94 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -432,19 +432,19 @@ int perf_header__write(struct perf_header *self, int fd, bool at_exit) return 0; } -static void do_read(int fd, void *buf, size_t size) +static int do_read(int fd, void *buf, size_t size) { while (size) { int ret = read(fd, buf, size); - if (ret < 0) - die("failed to read"); - if (ret == 0) - die("failed to read: missing data"); + if (ret <= 0) + return -1; size -= ret; buf += ret; } + + return 0; } int perf_header__process_sections(struct perf_header *self, int fd, @@ -455,7 +455,7 @@ int perf_header__process_sections(struct perf_header *self, int fd, int nr_sections; int sec_size; int idx = 0; - int err = 0, feat = 1; + int err = -1, feat = 1; nr_sections = bitmap_weight(self->adds_features, HEADER_FEAT_BITS); if (!nr_sections) @@ -469,8 +469,10 @@ int perf_header__process_sections(struct perf_header *self, int fd, lseek(fd, self->data_offset + self->data_size, SEEK_SET); - do_read(fd, feat_sec, sec_size); + if (do_read(fd, feat_sec, sec_size)) + goto out_free; + err = 0; while (idx < nr_sections && feat < HEADER_LAST_FEATURE) { if (perf_header__has_feat(self, feat)) { struct perf_file_section *sec = &feat_sec[idx++]; @@ -481,18 +483,18 @@ int perf_header__process_sections(struct perf_header *self, int fd, } ++feat; } - +out_free: free(feat_sec); return err; -}; +} int perf_file_header__read(struct perf_file_header *self, struct perf_header *ph, int fd) { lseek(fd, 0, SEEK_SET); - do_read(fd, self, sizeof(*self)); - if (self->magic != PERF_MAGIC || + if (do_read(fd, self, sizeof(*self)) || + self->magic != PERF_MAGIC || self->attr_size != sizeof(struct perf_file_attr)) return -1; @@ -558,7 +560,8 @@ int perf_header__read(struct perf_header *self, int fd) struct perf_header_attr *attr; off_t tmp; - do_read(fd, &f_attr, sizeof(f_attr)); + if (do_read(fd, &f_attr, sizeof(f_attr))) + goto out_errno; tmp = lseek(fd, 0, SEEK_CUR); attr = perf_header_attr__new(&f_attr.attr); @@ -569,7 +572,8 @@ int perf_header__read(struct perf_header *self, int fd) lseek(fd, f_attr.ids.offset, SEEK_SET); for (j = 0; j < nr_ids; j++) { - do_read(fd, &f_id, sizeof(f_id)); + if (do_read(fd, &f_id, sizeof(f_id))) + goto out_errno; if (perf_header_attr__add_id(attr, f_id) < 0) { perf_header_attr__delete(attr); @@ -589,7 +593,8 @@ int perf_header__read(struct perf_header *self, int fd) events = malloc(f_header.event_types.size); if (events == NULL) return -ENOMEM; - do_read(fd, events, f_header.event_types.size); + if (do_read(fd, events, f_header.event_types.size)) + goto out_errno; event_count = f_header.event_types.size / sizeof(struct perf_trace_event_type); } @@ -599,6 +604,8 @@ int perf_header__read(struct perf_header *self, int fd) self->frozen = 1; return 0; +out_errno: + return -errno; } u64 perf_header__sample_type(struct perf_header *header) -- cgit v0.10.2 From ae99fb2c335ef018520950ddc9692faacab39cf2 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 28 Dec 2009 22:48:33 -0200 Subject: perf header: perf_header__push_event() shouldn't die MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Just propagate eventual errors. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262047716-23171-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 6b3cb94..709e325 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -105,24 +105,28 @@ struct perf_trace_event_type { static int event_count; static struct perf_trace_event_type *events; -void perf_header__push_event(u64 id, const char *name) +int perf_header__push_event(u64 id, const char *name) { if (strlen(name) > MAX_EVENT_NAME) pr_warning("Event %s will be truncated\n", name); if (!events) { events = malloc(sizeof(struct perf_trace_event_type)); - if (!events) - die("nomem"); + if (events == NULL) + return -ENOMEM; } else { - events = realloc(events, (event_count + 1) * sizeof(struct perf_trace_event_type)); - if (!events) - die("nomem"); + struct perf_trace_event_type *nevents; + + nevents = realloc(events, (event_count + 1) * sizeof(*events)); + if (nevents == NULL) + return -ENOMEM; + events = nevents; } memset(&events[event_count], 0, sizeof(struct perf_trace_event_type)); events[event_count].event_id = id; strncpy(events[event_count].name, name, MAX_EVENT_NAME - 1); event_count++; + return 0; } char *perf_header__find_event(u64 id) diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h index d118d05..2b69aab 100644 --- a/tools/perf/util/header.h +++ b/tools/perf/util/header.h @@ -64,7 +64,7 @@ int perf_header__write(struct perf_header *self, int fd, bool at_exit); int perf_header__add_attr(struct perf_header *self, struct perf_header_attr *attr); -void perf_header__push_event(u64 id, const char *name); +int perf_header__push_event(u64 id, const char *name); char *perf_header__find_event(u64 id); struct perf_header_attr *perf_header_attr__new(struct perf_event_attr *attr); diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index dc585a8..609d5a9 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -753,11 +753,11 @@ modifier: return ret; } -static void store_event_type(const char *orgname) +static int store_event_type(const char *orgname) { char filename[PATH_MAX], *c; FILE *file; - int id; + int id, n; sprintf(filename, "%s/", debugfs_path); strncat(filename, orgname, strlen(orgname)); @@ -769,11 +769,14 @@ static void store_event_type(const char *orgname) file = fopen(filename, "r"); if (!file) - return; - if (fscanf(file, "%i", &id) < 1) - die("cannot store event ID"); + return 0; + n = fscanf(file, "%i", &id); fclose(file); - perf_header__push_event(id, orgname); + if (n < 1) { + pr_err("cannot store event ID\n"); + return -EINVAL; + } + return perf_header__push_event(id, orgname); } int parse_events(const struct option *opt __used, const char *str, int unset __used) @@ -782,7 +785,8 @@ int parse_events(const struct option *opt __used, const char *str, int unset __u enum event_result ret; if (strchr(str, ':')) - store_event_type(str); + if (store_event_type(str) < 0) + return -1; for (;;) { if (nr_counters == MAX_COUNTERS) -- cgit v0.10.2 From 71289be7630fb97f2de6bb2e18a50289dc869f9d Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 28 Dec 2009 22:48:34 -0200 Subject: perf report: Add --hide-unresolved/-U command line option MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Useful to match the 'overhead' column in 'perf report' with the 'baseline' one in 'perf diff'. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262047716-23171-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 508934b..4292d7a 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -34,6 +34,7 @@ static char const *input_name = "perf.data"; static int force; +static bool hide_unresolved; static int show_threads; static struct perf_read_values show_threads_values; @@ -121,7 +122,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) return -1; } - if (al.filtered) + if (al.filtered || (hide_unresolved && al.sym == NULL)) return 0; if (perf_session__add_hist_entry(session, &al, data.callchain, data.period)) { @@ -342,6 +343,8 @@ static const struct option options[] = { OPT_STRING('t', "field-separator", &symbol_conf.field_sep, "separator", "separator for columns, no spaces will be added between " "columns '.' is reserved."), + OPT_BOOLEAN('U', "hide-unresolved", &hide_unresolved, + "Only display entries resolved to a symbol"), OPT_END() }; -- cgit v0.10.2 From cdbae31408cf39372402076cf2e189ec693daa71 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 28 Dec 2009 22:48:35 -0200 Subject: perf diff: Don't add the period for unresolved symbols MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since we don't add histograms buckets for them, this way the sum of baselines should be 100%. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262047716-23171-4-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c index 1cbecaf..876a4b9 100644 --- a/tools/perf/builtin-diff.c +++ b/tools/perf/builtin-diff.c @@ -51,12 +51,12 @@ static int diff__process_sample_event(event_t *event, struct perf_session *sessi return -1; } - if (al.filtered) + if (al.filtered || al.sym == NULL) return 0; event__parse_sample(event, session->sample_type, &data); - if (al.sym && perf_session__add_hist_entry(session, &al, data.period)) { + if (perf_session__add_hist_entry(session, &al, data.period)) { pr_warning("problem incrementing symbol count, skipping event\n"); return -1; } -- cgit v0.10.2 From 9c443dfdd31eddea6cbe6ee0ca469fbcc4e1dc3b Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 28 Dec 2009 22:48:36 -0200 Subject: perf diff: Fix support for all --sort combinations MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When we finish creating the hist_entries we _already_ have them sorted "by name", in fact by what is in --sort, that is exactly how we can find the pairs in perf_session__match_hists as 'comm', 'dso' & 'symbol' all are strings we need to find the matches in the baseline session. So only do the sort by hits followed by a resort by --sort if we need to find the position for shwowing the --displacement of hist entries. Now all these modes work correctly: Example is a simple 'perf record -f find / > /dev/null' ran twice then followed by the following commands: $ perf diff -f --sort comm # Baseline Delta Command # ........ .......... ....... # 0.00% +100.00% find $ perf diff -f --sort dso # Baseline Delta Shared Object # ........ .......... .................. # 59.97% -0.44% [kernel] 21.17% +0.28% libc-2.5.so 18.49% +0.16% [ext3] 0.37% find $ perf diff -f --sort symbol | head -8 # Baseline Delta Symbol # ........ .......... ...... # 6.21% +0.36% [k] ext3fs_dirhash 3.43% +0.41% [.] __GI_strlen 3.53% +0.16% [k] __kmalloc 3.17% +0.49% [k] system_call 3.06% +0.37% [k] ext3_htree_store_dirent $ perf diff -f --sort dso,symbol | head -8 # Baseline Delta Shared Object Symbol # ........ .......... .................. ...... # 6.21% +0.36% [ext3] [k] ext3fs_dirhash 3.43% +0.41% libc-2.5.so [.] __GI_strlen 3.53% +0.16% [kernel] [k] __kmalloc 3.17% +0.49% [kernel] [k] system_call 3.06% +0.37% [ext3] [k] ext3_htree_store_dirent $ And we don't have to do two expensive resorts in the common, non --displacement case. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262047716-23171-5-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c index 876a4b9..924bfb7 100644 --- a/tools/perf/builtin-diff.c +++ b/tools/perf/builtin-diff.c @@ -82,29 +82,19 @@ static void perf_session__insert_hist_entry_by_name(struct rb_root *root, struct hist_entry *iter; while (*p != NULL) { - int cmp; parent = *p; iter = rb_entry(parent, struct hist_entry, rb_node); - - cmp = strcmp(he->map->dso->name, iter->map->dso->name); - if (cmp > 0) + if (hist_entry__cmp(he, iter) < 0) p = &(*p)->rb_left; - else if (cmp < 0) + else p = &(*p)->rb_right; - else { - cmp = strcmp(he->sym->name, iter->sym->name); - if (cmp > 0) - p = &(*p)->rb_left; - else - p = &(*p)->rb_right; - } } rb_link_node(&he->rb_node, parent, p); rb_insert_color(&he->rb_node, root); } -static void perf_session__resort_by_name(struct perf_session *self) +static void perf_session__resort_hist_entries(struct perf_session *self) { unsigned long position = 1; struct rb_root tmp = RB_ROOT; @@ -122,29 +112,28 @@ static void perf_session__resort_by_name(struct perf_session *self) self->hists = tmp; } +static void perf_session__set_hist_entries_positions(struct perf_session *self) +{ + perf_session__output_resort(self, self->events_stats.total); + perf_session__resort_hist_entries(self); +} + static struct hist_entry * -perf_session__find_hist_entry_by_name(struct perf_session *self, - struct hist_entry *he) +perf_session__find_hist_entry(struct perf_session *self, + struct hist_entry *he) { struct rb_node *n = self->hists.rb_node; while (n) { struct hist_entry *iter = rb_entry(n, struct hist_entry, rb_node); - int cmp = strcmp(he->map->dso->name, iter->map->dso->name); + int64_t cmp = hist_entry__cmp(he, iter); - if (cmp > 0) + if (cmp < 0) n = n->rb_left; - else if (cmp < 0) + else if (cmp > 0) n = n->rb_right; - else { - cmp = strcmp(he->sym->name, iter->sym->name); - if (cmp > 0) - n = n->rb_left; - else if (cmp < 0) - n = n->rb_right; - else - return iter; - } + else + return iter; } return NULL; @@ -155,11 +144,9 @@ static void perf_session__match_hists(struct perf_session *old_session, { struct rb_node *nd; - perf_session__resort_by_name(old_session); - for (nd = rb_first(&new_session->hists); nd; nd = rb_next(nd)) { struct hist_entry *pos = rb_entry(nd, struct hist_entry, rb_node); - pos->pair = perf_session__find_hist_entry_by_name(old_session, pos); + pos->pair = perf_session__find_hist_entry(old_session, pos); } } @@ -177,9 +164,12 @@ static int __cmd_diff(void) ret = perf_session__process_events(session[i], &event_ops); if (ret) goto out_delete; - perf_session__output_resort(session[i], session[i]->events_stats.total); } + perf_session__output_resort(session[1], session[1]->events_stats.total); + if (show_displacement) + perf_session__set_hist_entries_positions(session[0]); + perf_session__match_hists(session[0], session[1]); perf_session__fprintf_hists(session[1], session[0], show_displacement, stdout); -- cgit v0.10.2 From e591a517230a97e93f5edd19e7bef7ec8d22ac2b Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Sun, 13 Dec 2009 22:18:09 +0000 Subject: sh: Remove unused functions Apply some TLC to the SH64 header files and remove some functions that are not used anymore. Signed-off-by: Matt Fleming diff --git a/arch/sh/include/asm/pgtable_64.h b/arch/sh/include/asm/pgtable_64.h index 17cdbec..dd38158 100644 --- a/arch/sh/include/asm/pgtable_64.h +++ b/arch/sh/include/asm/pgtable_64.h @@ -43,11 +43,6 @@ static __inline__ void set_pte(pte_t *pteptr, pte_t pteval) } #define set_pte_at(mm,addr,ptep,pteval) set_pte(ptep,pteval) -static __inline__ void pmd_set(pmd_t *pmdp,pte_t *ptep) -{ - pmd_val(*pmdp) = (unsigned long) ptep; -} - /* * PGD defines. Top level. */ @@ -203,12 +198,6 @@ static __inline__ void pmd_set(pmd_t *pmdp,pte_t *ptep) #define pgprot_writecombine(prot) __pgprot(pgprot_val(prot) & ~_PAGE_CACHABLE) /* - * Handling allocation failures during page table setup. - */ -extern void __handle_bad_pmd_kernel(pmd_t * pmd); -#define __handle_bad_pmd(x) __handle_bad_pmd_kernel(x) - -/* * PTE level access routines. * * Note1: -- cgit v0.10.2 From 3f5ab768164df9a44721660b96e0accb92eb2c24 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Thu, 24 Dec 2009 20:38:45 +0000 Subject: sh: Correct the PTRS_PER_PMD and PMD_SHIFT values The previous expressions were wrong which made free_pmd_range() explode when using anything other than 4KB pages (which is why 8KB and 64KB pages were disabled with the 3-level page table layout). The problem was that pmd_offset() was returning an index of non-zero when it should have been returning 0. This non-zero offset was used to calculate the address of the pmd table to free in free_pmd_range(), which ended up trying to free an object that was not aligned on a page boundary. Now 3-level page tables should work with 4KB, 8KB and 64KB pages. Signed-off-by: Matt Fleming diff --git a/arch/sh/include/asm/pgtable_pmd.h b/arch/sh/include/asm/pgtable_pmd.h index 78dc36e..42a180e 100644 --- a/arch/sh/include/asm/pgtable_pmd.h +++ b/arch/sh/include/asm/pgtable_pmd.h @@ -17,11 +17,11 @@ #define USER_PTRS_PER_PGD 2 /* PMD bits */ -#define PMD_SHIFT (PAGE_SHIFT + (PAGE_SHIFT - 3)) +#define PMD_SHIFT (PAGE_SHIFT + (PAGE_SHIFT - PTE_MAGNITUDE)) #define PMD_SIZE (1UL << PMD_SHIFT) #define PMD_MASK (~(PMD_SIZE-1)) -#define PTRS_PER_PMD (PAGE_SIZE / sizeof(pmd_t)) +#define PTRS_PER_PMD ((1 << PGDIR_SHIFT) / PMD_SIZE) #define pmd_ERROR(e) \ printk("%s:%d: bad pmd %016llx.\n", __FILE__, __LINE__, pmd_val(e)) diff --git a/arch/sh/mm/Kconfig b/arch/sh/mm/Kconfig index b3f6c1a..1b43648 100644 --- a/arch/sh/mm/Kconfig +++ b/arch/sh/mm/Kconfig @@ -220,7 +220,7 @@ config PAGE_SIZE_4KB config PAGE_SIZE_8KB bool "8kB" - depends on !MMU || X2TLB && !PGTABLE_LEVELS_3 + depends on !MMU || X2TLB help This enables 8kB pages as supported by SH-X2 and later MMUs. @@ -232,7 +232,7 @@ config PAGE_SIZE_16KB config PAGE_SIZE_64KB bool "64kB" - depends on !MMU || CPU_SH4 && !PGTABLE_LEVELS_3 || CPU_SH5 + depends on !MMU || CPU_SH4 || CPU_SH5 help This enables support for 64kB pages, possible on all SH-4 CPUs and later. -- cgit v0.10.2 From b4c892762373c5e59c7e8db35f5f9a7658602bda Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Thu, 24 Dec 2009 22:17:35 +0000 Subject: sh: Optimise flush_dcache_page() on SH4 If the page is not mapped into any process's address space then aliases cannot exist in the cache. So reduce the amount of flushing we perform. Signed-off-by: Matt Fleming diff --git a/arch/sh/mm/cache-sh4.c b/arch/sh/mm/cache-sh4.c index 560ddb6..a2301da 100644 --- a/arch/sh/mm/cache-sh4.c +++ b/arch/sh/mm/cache-sh4.c @@ -109,6 +109,7 @@ static inline void flush_cache_one(unsigned long start, unsigned long phys) static void sh4_flush_dcache_page(void *arg) { struct page *page = arg; + unsigned long addr = (unsigned long)page_address(page); #ifndef CONFIG_SMP struct address_space *mapping = page_mapping(page); @@ -116,16 +117,8 @@ static void sh4_flush_dcache_page(void *arg) set_bit(PG_dcache_dirty, &page->flags); else #endif - { - unsigned long phys = page_to_phys(page); - unsigned long addr = CACHE_OC_ADDRESS_ARRAY; - int i, n; - - /* Loop all the D-cache */ - n = boot_cpu_data.dcache.n_aliases; - for (i = 0; i < n; i++, addr += PAGE_SIZE) - flush_cache_one(addr, phys); - } + flush_cache_one(CACHE_OC_ADDRESS_ARRAY | + (addr & shm_align_mask), page_to_phys(page)); wmb(); } -- cgit v0.10.2 From 2a5eacca85d39d8b6dffae821d7d260f05584dc7 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Thu, 31 Dec 2009 12:19:24 +0000 Subject: sh: Move page table allocation out of line We also switched away from quicklists and instead moved to slab caches. After benchmarking both implementations the difference is negligible. The slab caches suit us better though because the size of a pgd table is just 4 entries when we're using a 3-level page table layout and quicklists always deal with pages. Signed-off-by: Matt Fleming diff --git a/arch/sh/include/asm/pgalloc.h b/arch/sh/include/asm/pgalloc.h index 4ea2785..e106474 100644 --- a/arch/sh/include/asm/pgalloc.h +++ b/arch/sh/include/asm/pgalloc.h @@ -6,10 +6,13 @@ #define QUICK_PT 1 /* Other page table pages that are zero on free */ +extern pgd_t *pgd_alloc(struct mm_struct *); +extern void pgd_free(struct mm_struct *mm, pgd_t *pgd); + #ifdef CONFIG_PGTABLE_LEVELS_3 -#include -#else -#include +extern void pud_populate(struct mm_struct *mm, pud_t *pudp, pmd_t *pmd); +extern pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long address); +extern void pmd_free(struct mm_struct *mm, pmd_t *pmd); #endif static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd, @@ -67,7 +70,6 @@ do { \ static inline void check_pgt_cache(void) { - __check_pgt_cache(); quicklist_trim(QUICK_PT, NULL, 25, 16); } diff --git a/arch/sh/include/asm/pgalloc_nopmd.h b/arch/sh/include/asm/pgalloc_nopmd.h deleted file mode 100644 index e4b344c..0000000 --- a/arch/sh/include/asm/pgalloc_nopmd.h +++ /dev/null @@ -1,30 +0,0 @@ -#ifndef __ASM_SH_PGALLOC_NOPMD_H -#define __ASM_SH_PGALLOC_NOPMD_H - -#define QUICK_PGD 0 /* We preserve special mappings over free */ - -static inline void pgd_ctor(void *x) -{ - pgd_t *pgd = x; - - memcpy(pgd + USER_PTRS_PER_PGD, - swapper_pg_dir + USER_PTRS_PER_PGD, - (PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t)); -} - -static inline pgd_t *pgd_alloc(struct mm_struct *mm) -{ - return quicklist_alloc(QUICK_PGD, GFP_KERNEL | __GFP_REPEAT, pgd_ctor); -} - -static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd) -{ - quicklist_free(QUICK_PGD, NULL, pgd); -} - -static inline void __check_pgt_cache(void) -{ - quicklist_trim(QUICK_PGD, NULL, 25, 16); -} - -#endif /* __ASM_SH_PGALLOC_NOPMD_H */ diff --git a/arch/sh/include/asm/pgalloc_pmd.h b/arch/sh/include/asm/pgalloc_pmd.h deleted file mode 100644 index 20f75cc..0000000 --- a/arch/sh/include/asm/pgalloc_pmd.h +++ /dev/null @@ -1,41 +0,0 @@ -#ifndef __ASM_SH_PGALLOC_PMD_H -#define __ASM_SH_PGALLOC_PMD_H - -static inline pgd_t *pgd_alloc(struct mm_struct *mm) -{ - pgd_t *pgd; - int i; - - pgd = kzalloc(sizeof(*pgd) * PTRS_PER_PGD, GFP_KERNEL | __GFP_REPEAT); - - for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) - pgd[i] = swapper_pg_dir[i]; - - return pgd; -} - -static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd) -{ - kfree(pgd); -} - -static inline void __check_pgt_cache(void) -{ -} - -static inline void pud_populate(struct mm_struct *mm, pud_t *pud, pmd_t *pmd) -{ - set_pud(pud, __pud((unsigned long)pmd)); -} - -static inline pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long address) -{ - return quicklist_alloc(QUICK_PT, GFP_KERNEL | __GFP_REPEAT, NULL); -} - -static inline void pmd_free(struct mm_struct *mm, pmd_t *pmd) -{ - quicklist_free(QUICK_PT, NULL, pmd); -} - -#endif /* __ASM_SH_PGALLOC_PMD_H */ diff --git a/arch/sh/include/asm/pgtable.h b/arch/sh/include/asm/pgtable.h index 9effcc3..78598ec 100644 --- a/arch/sh/include/asm/pgtable.h +++ b/arch/sh/include/asm/pgtable.h @@ -141,9 +141,9 @@ typedef pte_t *pte_addr_t; #define pte_pfn(x) ((unsigned long)(((x).pte_low >> PAGE_SHIFT))) /* - * No page table caches to initialise + * Initialise the page table caches */ -#define pgtable_cache_init() do { } while (0) +extern void pgtable_cache_init(void); struct vm_area_struct; diff --git a/arch/sh/mm/Makefile b/arch/sh/mm/Makefile index 8a70535..dd5010c 100644 --- a/arch/sh/mm/Makefile +++ b/arch/sh/mm/Makefile @@ -15,7 +15,7 @@ obj-y += $(cacheops-y) mmu-y := nommu.o extable_32.o mmu-$(CONFIG_MMU) := extable_$(BITS).o fault_$(BITS).o \ - ioremap_$(BITS).o kmap.o tlbflush_$(BITS).o + ioremap_$(BITS).o kmap.o pgtable.o tlbflush_$(BITS).o obj-y += $(mmu-y) obj-$(CONFIG_DEBUG_FS) += asids-debugfs.o diff --git a/arch/sh/mm/pgtable.c b/arch/sh/mm/pgtable.c new file mode 100644 index 0000000..e1bc548 --- /dev/null +++ b/arch/sh/mm/pgtable.c @@ -0,0 +1,57 @@ +#include + +#define PGALLOC_GFP GFP_KERNEL | __GFP_REPEAT | __GFP_ZERO + +static struct kmem_cache *pgd_cachep; + +#ifdef CONFIG_PGTABLE_LEVELS_3 +static struct kmem_cache *pmd_cachep; +#endif + +void pgd_ctor(void *x) +{ + pgd_t *pgd = x; + + memcpy(pgd + USER_PTRS_PER_PGD, + swapper_pg_dir + USER_PTRS_PER_PGD, + (PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t)); +} + +void pgtable_cache_init(void) +{ + pgd_cachep = kmem_cache_create("pgd_cache", + PTRS_PER_PGD * (1< Date: Mon, 4 Jan 2010 12:04:08 +0100 Subject: HID: add support for Stantum multitouch panel Added support for the Stantum multitouch panel. Signed-off-by: Stephane Chatty Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index 0c8ce3a..84272c6 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -247,6 +247,12 @@ config HID_SONY ---help--- Support for Sony PS3 controller. +config HID_STANTUM + tristate "Stantum" if EMBEDDED + depends on USB_HID + ---help--- + Support for Stantum multitouch panel. + config HID_SUNPLUS tristate "Sunplus" if EMBEDDED depends on USB_HID diff --git a/drivers/hid/Makefile b/drivers/hid/Makefile index 1bdfb97..f8dc5ba 100644 --- a/drivers/hid/Makefile +++ b/drivers/hid/Makefile @@ -40,6 +40,7 @@ obj-$(CONFIG_HID_PETALYNX) += hid-petalynx.o obj-$(CONFIG_HID_SAMSUNG) += hid-samsung.o obj-$(CONFIG_HID_SMARTJOYPLUS) += hid-sjoy.o obj-$(CONFIG_HID_SONY) += hid-sony.o +obj-$(CONFIG_HID_STANTUM) += hid-stantum.o obj-$(CONFIG_HID_SUNPLUS) += hid-sunplus.o obj-$(CONFIG_HID_GREENASIA) += hid-gaff.o obj-$(CONFIG_HID_THRUSTMASTER) += hid-tmff.o diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index 70c25a0..6462c92 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1339,6 +1339,7 @@ static const struct hid_device_id hid_blacklist[] = { { HID_USB_DEVICE(USB_VENDOR_ID_SAMSUNG, USB_DEVICE_ID_SAMSUNG_IR_REMOTE) }, { HID_USB_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_PS3_CONTROLLER) }, { HID_USB_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_VAIO_VGX_MOUSE) }, + { HID_USB_DEVICE(USB_VENDOR_ID_STANTUM, USB_DEVICE_ID_MTP) }, { HID_USB_DEVICE(USB_VENDOR_ID_SUNPLUS, USB_DEVICE_ID_SUNPLUS_WDESKTOP) }, { HID_USB_DEVICE(USB_VENDOR_ID_THRUSTMASTER, 0xb300) }, { HID_USB_DEVICE(USB_VENDOR_ID_THRUSTMASTER, 0xb304) }, diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index 3d55f98..07e056b 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -393,6 +393,9 @@ #define USB_DEVICE_ID_SOUNDGRAPH_IMON_FIRST 0x0034 #define USB_DEVICE_ID_SOUNDGRAPH_IMON_LAST 0x0046 +#define USB_VENDOR_ID_STANTUM 0x1f87 +#define USB_DEVICE_ID_MTP 0x0002 + #define USB_VENDOR_ID_SUN 0x0430 #define USB_DEVICE_ID_RARITAN_KVM_DONGLE 0xcdab diff --git a/drivers/hid/hid-stantum.c b/drivers/hid/hid-stantum.c new file mode 100644 index 0000000..add965d --- /dev/null +++ b/drivers/hid/hid-stantum.c @@ -0,0 +1,283 @@ +/* + * HID driver for Stantum multitouch panels + * + * Copyright (c) 2009 Stephane Chatty + * + */ + +/* + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include +#include +#include + +MODULE_VERSION("0.6"); +MODULE_AUTHOR("Stephane Chatty "); +MODULE_DESCRIPTION("Stantum HID multitouch panels"); +MODULE_LICENSE("GPL"); + +#include "hid-ids.h" + +struct stantum_data { + __s32 x, y, z, w, h; /* x, y, pressure, width, height */ + __u16 id; /* touch id */ + bool valid; /* valid finger data, or just placeholder? */ + bool first; /* first finger in the HID packet? */ + bool activity; /* at least one active finger so far? */ +}; + +static int stantum_input_mapping(struct hid_device *hdev, struct hid_input *hi, + struct hid_field *field, struct hid_usage *usage, + unsigned long **bit, int *max) +{ + switch (usage->hid & HID_USAGE_PAGE) { + + case HID_UP_GENDESK: + switch (usage->hid) { + case HID_GD_X: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_POSITION_X); + /* touchscreen emulation */ + input_set_abs_params(hi->input, ABS_X, + field->logical_minimum, + field->logical_maximum, 0, 0); + return 1; + case HID_GD_Y: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_POSITION_Y); + /* touchscreen emulation */ + input_set_abs_params(hi->input, ABS_Y, + field->logical_minimum, + field->logical_maximum, 0, 0); + return 1; + } + return 0; + + case HID_UP_DIGITIZER: + switch (usage->hid) { + case HID_DG_INRANGE: + case HID_DG_CONFIDENCE: + case HID_DG_INPUTMODE: + case HID_DG_DEVICEINDEX: + case HID_DG_CONTACTCOUNT: + case HID_DG_CONTACTMAX: + case HID_DG_TIPPRESSURE: + return -1; + + case HID_DG_TIPSWITCH: + /* touchscreen emulation */ + hid_map_usage(hi, usage, bit, max, EV_KEY, BTN_TOUCH); + return 1; + + case HID_DG_WIDTH: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_TOUCH_MAJOR); + return 1; + case HID_DG_HEIGHT: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_TOUCH_MINOR); + input_set_abs_params(hi->input, ABS_MT_ORIENTATION, + 1, 1, 0, 0); + return 1; + case HID_DG_CONTACTID: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_TRACKING_ID); + return 1; + + } + return 0; + + case 0xff000000: + /* no input-oriented meaning */ + return -1; + } + + return 0; +} + +static int stantum_input_mapped(struct hid_device *hdev, struct hid_input *hi, + struct hid_field *field, struct hid_usage *usage, + unsigned long **bit, int *max) +{ + if (usage->type == EV_KEY || usage->type == EV_ABS) + clear_bit(usage->code, *bit); + + return 0; +} + +/* + * this function is called when a whole finger has been parsed, + * so that it can decide what to send to the input layer. + */ +static void stantum_filter_event(struct stantum_data *sd, + struct input_dev *input) +{ + bool wide; + + if (!sd->valid) { + /* + * touchscreen emulation: if the first finger is not valid and + * there previously was finger activity, this is a release + */ + if (sd->first && sd->activity) { + input_event(input, EV_KEY, BTN_TOUCH, 0); + sd->activity = false; + } + return; + } + + input_event(input, EV_ABS, ABS_MT_TRACKING_ID, sd->id); + input_event(input, EV_ABS, ABS_MT_POSITION_X, sd->x); + input_event(input, EV_ABS, ABS_MT_POSITION_Y, sd->y); + + wide = (sd->w > sd->h); + input_event(input, EV_ABS, ABS_MT_ORIENTATION, wide); + input_event(input, EV_ABS, ABS_MT_TOUCH_MAJOR, wide ? sd->w : sd->h); + input_event(input, EV_ABS, ABS_MT_TOUCH_MINOR, wide ? sd->h : sd->w); + +#if 0 + /* MT_PRESSURE does not exist yet */ + input_event(input, EV_ABS, ABS_MT_PRESSURE, sd->z); +#endif + + input_mt_sync(input); + sd->valid = false; + sd->first = false; + + /* touchscreen emulation */ + if (sd->first) { + if (!sd->activity) { + input_event(input, EV_KEY, BTN_TOUCH, 1); + sd->activity = true; + } + input_event(input, EV_ABS, ABS_X, sd->x); + input_event(input, EV_ABS, ABS_Y, sd->y); + } +} + + +static int stantum_event(struct hid_device *hid, struct hid_field *field, + struct hid_usage *usage, __s32 value) +{ + struct stantum_data *sd = hid_get_drvdata(hid); + + if (hid->claimed & HID_CLAIMED_INPUT) { + struct input_dev *input = field->hidinput->input; + + switch (usage->hid) { + case HID_DG_INRANGE: + /* this is the last field in a finger */ + stantum_filter_event(sd, input); + break; + case HID_DG_WIDTH: + sd->w = value; + break; + case HID_DG_HEIGHT: + sd->h = value; + break; + case HID_GD_X: + sd->x = value; + break; + case HID_GD_Y: + sd->y = value; + break; + case HID_DG_TIPPRESSURE: + sd->z = value; + break; + case HID_DG_CONTACTID: + sd->id = value; + break; + case HID_DG_CONFIDENCE: + sd->valid = !!value; + break; + case 0xff000002: + /* this comes only before the first finger */ + sd->first = true; + break; + + default: + /* ignore the others */ + return 1; + } + } + + /* we have handled the hidinput part, now remains hiddev */ + if (hid->claimed & HID_CLAIMED_HIDDEV && hid->hiddev_hid_event) + hid->hiddev_hid_event(hid, field, usage, value); + + return 1; +} + +static int stantum_probe(struct hid_device *hdev, + const struct hid_device_id *id) +{ + int ret; + struct stantum_data *sd; + + sd = kmalloc(sizeof(struct stantum_data), GFP_KERNEL); + if (!sd) { + dev_err(&hdev->dev, "cannot allocate Stantum data\n"); + return -ENOMEM; + } + sd->valid = false; + sd->first = false; + sd->activity = false; + hid_set_drvdata(hdev, sd); + + ret = hid_parse(hdev); + if (!ret) + ret = hid_hw_start(hdev, HID_CONNECT_DEFAULT); + + if (ret) + kfree(sd); + + return ret; +} + +static void stantum_remove(struct hid_device *hdev) +{ + hid_hw_stop(hdev); + kfree(hid_get_drvdata(hdev)); + hid_set_drvdata(hdev, NULL); +} + +static const struct hid_device_id stantum_devices[] = { + { HID_USB_DEVICE(USB_VENDOR_ID_STANTUM, USB_DEVICE_ID_MTP) }, + { } +}; +MODULE_DEVICE_TABLE(hid, stantum_devices); + +static const struct hid_usage_id stantum_grabbed_usages[] = { + { HID_ANY_ID, HID_ANY_ID, HID_ANY_ID }, + { HID_ANY_ID - 1, HID_ANY_ID - 1, HID_ANY_ID - 1} +}; + +static struct hid_driver stantum_driver = { + .name = "stantum", + .id_table = stantum_devices, + .probe = stantum_probe, + .remove = stantum_remove, + .input_mapping = stantum_input_mapping, + .input_mapped = stantum_input_mapped, + .usage_table = stantum_grabbed_usages, + .event = stantum_event, +}; + +static int __init stantum_init(void) +{ + return hid_register_driver(&stantum_driver); +} + +static void __exit stantum_exit(void) +{ + hid_unregister_driver(&stantum_driver); +} + +module_init(stantum_init); +module_exit(stantum_exit); + -- cgit v0.10.2 From 92688c0c3c1c9e2daf705d307e8fda1b5a180d26 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Mon, 4 Jan 2010 12:04:59 +0100 Subject: HID: make Stantum driver standalone config option Analogically to commit "HID: make 3M PCT touchscreen driver standalone config option", remove the dependency of Stantum driver on CONFIG_EMBEDDED, as it is a standalone driver rather than device quirk. Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index 84272c6..5f73774 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -248,7 +248,7 @@ config HID_SONY Support for Sony PS3 controller. config HID_STANTUM - tristate "Stantum" if EMBEDDED + tristate "Stantum" depends on USB_HID ---help--- Support for Stantum multitouch panel. -- cgit v0.10.2 From cf2f765f1896064e34c6f0f2ef896ff058dd5c06 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Mon, 4 Jan 2010 12:20:56 +0100 Subject: HID: handle joysticks with large number of buttons Current HID code doesn't properly handle HID joysticks which have larger number of buttons than what fits into current range reserved for BTN_JOYSTICK. One such joystick reported to not work properly is Saitek X52 Pro Flight System. We can't extend the range to fit more buttons in, because of backwards compatibility reasons. Therefore this patch introduces a new BTN_TRIGGER_HAPPY range, and uses these to map the buttons which are over BTN_JOYSTICK limit. Acked-by: Dmitry Torokhov [for the input.h part] Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-input.c b/drivers/hid/hid-input.c index 5862b0f..dad7aae 100644 --- a/drivers/hid/hid-input.c +++ b/drivers/hid/hid-input.c @@ -198,7 +198,12 @@ static void hidinput_configure_usage(struct hid_input *hidinput, struct hid_fiel switch (field->application) { case HID_GD_MOUSE: case HID_GD_POINTER: code += 0x110; break; - case HID_GD_JOYSTICK: code += 0x120; break; + case HID_GD_JOYSTICK: + if (code <= 0xf) + code += BTN_JOYSTICK; + else + code += BTN_TRIGGER_HAPPY; + break; case HID_GD_GAMEPAD: code += 0x130; break; default: switch (field->physical) { diff --git a/include/linux/input.h b/include/linux/input.h index 7be8a65..97f98ca 100644 --- a/include/linux/input.h +++ b/include/linux/input.h @@ -597,6 +597,48 @@ struct input_absinfo { #define KEY_CAMERA_FOCUS 0x210 +#define BTN_TRIGGER_HAPPY 0x2c0 +#define BTN_TRIGGER_HAPPY1 0x2c0 +#define BTN_TRIGGER_HAPPY2 0x2c1 +#define BTN_TRIGGER_HAPPY3 0x2c2 +#define BTN_TRIGGER_HAPPY4 0x2c3 +#define BTN_TRIGGER_HAPPY5 0x2c4 +#define BTN_TRIGGER_HAPPY6 0x2c5 +#define BTN_TRIGGER_HAPPY7 0x2c6 +#define BTN_TRIGGER_HAPPY8 0x2c7 +#define BTN_TRIGGER_HAPPY9 0x2c8 +#define BTN_TRIGGER_HAPPY10 0x2c9 +#define BTN_TRIGGER_HAPPY11 0x2ca +#define BTN_TRIGGER_HAPPY12 0x2cb +#define BTN_TRIGGER_HAPPY13 0x2cc +#define BTN_TRIGGER_HAPPY14 0x2cd +#define BTN_TRIGGER_HAPPY15 0x2ce +#define BTN_TRIGGER_HAPPY16 0x2cf +#define BTN_TRIGGER_HAPPY17 0x2d0 +#define BTN_TRIGGER_HAPPY18 0x2d1 +#define BTN_TRIGGER_HAPPY19 0x2d2 +#define BTN_TRIGGER_HAPPY20 0x2d3 +#define BTN_TRIGGER_HAPPY21 0x2d4 +#define BTN_TRIGGER_HAPPY22 0x2d5 +#define BTN_TRIGGER_HAPPY23 0x2d6 +#define BTN_TRIGGER_HAPPY24 0x2d7 +#define BTN_TRIGGER_HAPPY25 0x2d8 +#define BTN_TRIGGER_HAPPY26 0x2d9 +#define BTN_TRIGGER_HAPPY27 0x2da +#define BTN_TRIGGER_HAPPY28 0x2db +#define BTN_TRIGGER_HAPPY29 0x2dc +#define BTN_TRIGGER_HAPPY30 0x2dd +#define BTN_TRIGGER_HAPPY31 0x2de +#define BTN_TRIGGER_HAPPY32 0x2df +#define BTN_TRIGGER_HAPPY33 0x2e0 +#define BTN_TRIGGER_HAPPY34 0x2e1 +#define BTN_TRIGGER_HAPPY35 0x2e2 +#define BTN_TRIGGER_HAPPY36 0x2e3 +#define BTN_TRIGGER_HAPPY37 0x2e4 +#define BTN_TRIGGER_HAPPY38 0x2e5 +#define BTN_TRIGGER_HAPPY39 0x2e6 +#define BTN_TRIGGER_HAPPY40 0x2e7 + /* We avoid low common keys in module aliases so they don't get huge. */ #define KEY_MIN_INTERESTING KEY_MUTE #define KEY_MAX 0x2ff -- cgit v0.10.2 From 0176bd3dab4fe522bfb6ceab9e3c441fe0305738 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 5 Jan 2010 12:35:00 +0900 Subject: sh: Drop down to a single quicklist. We previously had 2 quicklists, one for the PGD case and one for PTEs. Now that the PGD/PMD cases are handled through slab caches due to the multi-level configurability, only the PTE quicklist remains. As such, reduce NR_QUICK to its appropriate size and bump down the PTE quicklist index. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/pgalloc.h b/arch/sh/include/asm/pgalloc.h index e106474..f8982f4 100644 --- a/arch/sh/include/asm/pgalloc.h +++ b/arch/sh/include/asm/pgalloc.h @@ -4,7 +4,7 @@ #include #include -#define QUICK_PT 1 /* Other page table pages that are zero on free */ +#define QUICK_PT 0 /* Other page table pages that are zero on free */ extern pgd_t *pgd_alloc(struct mm_struct *); extern void pgd_free(struct mm_struct *mm, pgd_t *pgd); diff --git a/mm/Kconfig b/mm/Kconfig index 17b8947..d34c2b9 100644 --- a/mm/Kconfig +++ b/mm/Kconfig @@ -195,7 +195,7 @@ config BOUNCE config NR_QUICK int depends on QUICKLIST - default "2" if SUPERH || AVR32 + default "2" if AVR32 default "1" config VIRT_TO_BUS -- cgit v0.10.2 From c4761815ab49feca904776dec464046bc7138d3a Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 5 Jan 2010 12:44:02 +0900 Subject: sh: Fix up breakpoint trap handler patching on SH-2A. SH-2A was referencing the old handler that no longer exists, fix it up. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/traps_32.c b/arch/sh/kernel/traps_32.c index 86639be..706c1c1 100644 --- a/arch/sh/kernel/traps_32.c +++ b/arch/sh/kernel/traps_32.c @@ -956,7 +956,7 @@ void __init trap_init(void) #endif #ifdef TRAP_UBC - set_exception_table_vec(TRAP_UBC, break_point_trap); + set_exception_table_vec(TRAP_UBC, breakpoint_trap_handler); #endif /* Setup VBR for boot cpu */ -- cgit v0.10.2 From 4352fc1b12fae4c753a063a2f162ddf9277af774 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 5 Jan 2010 19:06:45 +0900 Subject: sh: Abstracted SH-4A UBC support on hw-breakpoint core. This is the next big chunk of hw_breakpoint support. This decouples the SH-4A support from the core and moves it out in to its own stub, following many of the conventions established with the perf events layering. In addition to extending SH-4A support to encapsulate the remainder of the UBC channels, clock framework support for handling the UBC interface clock is added as well, allowing for dynamic clock gating. This also fixes up a regression introduced by the SIGTRAP handling that broke the ksym_tracer, to the extent that the current support works well with all of the ksym_tracer/ptrace/kgdb. The kprobes singlestep code will follow in turn. With this in place, the remaining UBC variants (SH-2A and SH-4) can now be trivially plugged in. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/hw_breakpoint.h b/arch/sh/include/asm/hw_breakpoint.h index 0f4a00f..7295d62 100644 --- a/arch/sh/include/asm/hw_breakpoint.h +++ b/arch/sh/include/asm/hw_breakpoint.h @@ -3,7 +3,6 @@ #include #include -#include #ifdef __KERNEL__ #define __ARCH_HW_BREAKPOINT_H @@ -11,7 +10,6 @@ struct arch_hw_breakpoint { char *name; /* Contains name of the symbol to set bkpt */ unsigned long address; - unsigned long asid; u16 len; u16 type; }; @@ -27,13 +25,28 @@ enum { SH_BREAKPOINT_LEN_8 = (1 << 14), }; -/* Total number of available UBC channels */ -#define HBP_NUM 1 /* XXX */ +struct sh_ubc { + const char *name; + unsigned int num_events; + unsigned int trap_nr; + void (*enable)(struct arch_hw_breakpoint *, int); + void (*disable)(struct arch_hw_breakpoint *, int); + void (*enable_all)(unsigned long); + void (*disable_all)(void); + unsigned long (*active_mask)(void); + unsigned long (*triggered_mask)(void); + void (*clear_triggered_mask)(unsigned long); + struct clk *clk; /* optional interface clock / MSTP bit */ +}; struct perf_event; struct task_struct; struct pmu; +/* Maximum number of UBC channels */ +#define HBP_NUM 2 + +/* arch/sh/kernel/hw_breakpoint.c */ extern int arch_check_va_in_userspace(unsigned long va, u16 hbp_len); extern int arch_validate_hwbkpt_settings(struct perf_event *bp, struct task_struct *tsk); @@ -46,6 +59,7 @@ void hw_breakpoint_pmu_read(struct perf_event *bp); void hw_breakpoint_pmu_unthrottle(struct perf_event *bp); extern void arch_fill_perf_breakpoint(struct perf_event *bp); +extern int register_sh_ubc(struct sh_ubc *); extern struct pmu perf_ops_bp; diff --git a/arch/sh/include/asm/processor_32.h b/arch/sh/include/asm/processor_32.h index d60b282..259112c 100644 --- a/arch/sh/include/asm/processor_32.h +++ b/arch/sh/include/asm/processor_32.h @@ -14,7 +14,7 @@ #include #include #include -#include +#include /* * Default implementation of macro that returns current @@ -102,7 +102,7 @@ struct thread_struct { unsigned long pc; /* Save middle states of ptrace breakpoints */ - struct perf_event *ptrace_bps[NR_UBC_CHANNELS]; + struct perf_event *ptrace_bps[HBP_NUM]; /* floating point info */ union sh_fpu_union fpu; diff --git a/arch/sh/kernel/cpu/sh4a/Makefile b/arch/sh/kernel/cpu/sh4a/Makefile index 33bab47..b144e8a 100644 --- a/arch/sh/kernel/cpu/sh4a/Makefile +++ b/arch/sh/kernel/cpu/sh4a/Makefile @@ -41,7 +41,8 @@ pinmux-$(CONFIG_CPU_SUBTYPE_SH7757) := pinmux-sh7757.o pinmux-$(CONFIG_CPU_SUBTYPE_SH7785) := pinmux-sh7785.o pinmux-$(CONFIG_CPU_SUBTYPE_SH7786) := pinmux-sh7786.o -obj-y += $(clock-y) -obj-$(CONFIG_SMP) += $(smp-y) -obj-$(CONFIG_GENERIC_GPIO) += $(pinmux-y) -obj-$(CONFIG_PERF_EVENTS) += perf_event.o +obj-y += $(clock-y) +obj-$(CONFIG_SMP) += $(smp-y) +obj-$(CONFIG_GENERIC_GPIO) += $(pinmux-y) +obj-$(CONFIG_PERF_EVENTS) += perf_event.o +obj-$(CONFIG_HAVE_HW_BREAKPOINT) += ubc.o diff --git a/arch/sh/kernel/cpu/sh4a/ubc.c b/arch/sh/kernel/cpu/sh4a/ubc.c new file mode 100644 index 0000000..efb2745 --- /dev/null +++ b/arch/sh/kernel/cpu/sh4a/ubc.c @@ -0,0 +1,133 @@ +/* + * arch/sh/kernel/cpu/sh4a/ubc.c + * + * On-chip UBC support for SH-4A CPUs. + * + * Copyright (C) 2009 - 2010 Paul Mundt + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ +#include +#include +#include +#include +#include + +#define UBC_CBR(idx) (0xff200000 + (0x20 * idx)) +#define UBC_CRR(idx) (0xff200004 + (0x20 * idx)) +#define UBC_CAR(idx) (0xff200008 + (0x20 * idx)) +#define UBC_CAMR(idx) (0xff20000c + (0x20 * idx)) + +#define UBC_CCMFR 0xff200600 +#define UBC_CBCR 0xff200620 + +/* CRR */ +#define UBC_CRR_PCB (1 << 1) +#define UBC_CRR_BIE (1 << 0) + +/* CBR */ +#define UBC_CBR_CE (1 << 0) + +static struct sh_ubc sh4a_ubc; + +static void sh4a_ubc_enable(struct arch_hw_breakpoint *info, int idx) +{ + __raw_writel(UBC_CBR_CE | info->len | info->type, UBC_CBR(idx)); + __raw_writel(info->address, UBC_CAR(idx)); +} + +static void sh4a_ubc_disable(struct arch_hw_breakpoint *info, int idx) +{ + __raw_writel(0, UBC_CBR(idx)); + __raw_writel(0, UBC_CAR(idx)); +} + +static void sh4a_ubc_enable_all(unsigned long mask) +{ + int i; + + for (i = 0; i < sh4a_ubc.num_events; i++) + if (mask & (1 << i)) + __raw_writel(__raw_readl(UBC_CBR(i)) | UBC_CBR_CE, + UBC_CBR(i)); +} + +static void sh4a_ubc_disable_all(void) +{ + int i; + + for (i = 0; i < sh4a_ubc.num_events; i++) + __raw_writel(__raw_readl(UBC_CBR(i)) & ~UBC_CBR_CE, + UBC_CBR(i)); +} + +static unsigned long sh4a_ubc_active_mask(void) +{ + unsigned long active = 0; + int i; + + for (i = 0; i < sh4a_ubc.num_events; i++) + if (__raw_readl(UBC_CBR(i)) & UBC_CBR_CE) + active |= (1 << i); + + return active; +} + +static unsigned long sh4a_ubc_triggered_mask(void) +{ + return __raw_readl(UBC_CCMFR); +} + +static void sh4a_ubc_clear_triggered_mask(unsigned long mask) +{ + __raw_writel(__raw_readl(UBC_CCMFR) & ~mask, UBC_CCMFR); +} + +static struct sh_ubc sh4a_ubc = { + .name = "SH-4A", + .num_events = 2, + .trap_nr = 0x1e0, + .enable = sh4a_ubc_enable, + .disable = sh4a_ubc_disable, + .enable_all = sh4a_ubc_enable_all, + .disable_all = sh4a_ubc_disable_all, + .active_mask = sh4a_ubc_active_mask, + .triggered_mask = sh4a_ubc_triggered_mask, + .clear_triggered_mask = sh4a_ubc_clear_triggered_mask, +}; + +static int __init sh4a_ubc_init(void) +{ + struct clk *ubc_iclk = clk_get(NULL, "ubc0"); + int i; + + /* + * The UBC MSTP bit is optional, as not all platforms will have + * it. Just ignore it if we can't find it. + */ + if (IS_ERR(ubc_iclk)) + ubc_iclk = NULL; + + clk_enable(ubc_iclk); + + __raw_writel(0, UBC_CBCR); + + for (i = 0; i < sh4a_ubc.num_events; i++) { + __raw_writel(0, UBC_CAMR(i)); + __raw_writel(0, UBC_CBR(i)); + + __raw_writel(UBC_CRR_BIE | UBC_CRR_PCB, UBC_CRR(i)); + + /* dummy read for write posting */ + (void)__raw_readl(UBC_CRR(i)); + } + + clk_disable(ubc_iclk); + + sh4a_ubc.clk = ubc_iclk; + + return register_sh_ubc(&sh4a_ubc); +} +arch_initcall(sh4a_ubc_init); diff --git a/arch/sh/kernel/hw_breakpoint.c b/arch/sh/kernel/hw_breakpoint.c index c515a3e..e2f1753 100644 --- a/arch/sh/kernel/hw_breakpoint.c +++ b/arch/sh/kernel/hw_breakpoint.c @@ -3,7 +3,7 @@ * * Unified kernel/user-space hardware breakpoint facility for the on-chip UBC. * - * Copyright (C) 2009 Paul Mundt + * Copyright (C) 2009 - 2010 Paul Mundt * * This file is subject to the terms and conditions of the GNU General Public * License. See the file "COPYING" in the main directory of this archive @@ -18,38 +18,24 @@ #include #include #include +#include #include #include #include -struct ubc_context { - unsigned long pc; - unsigned long state; -}; - -/* Per cpu ubc channel state */ -static DEFINE_PER_CPU(struct ubc_context, ubc_ctx[HBP_NUM]); - /* * Stores the breakpoints currently in use on each breakpoint address * register for each cpus */ static DEFINE_PER_CPU(struct perf_event *, bp_per_reg[HBP_NUM]); -static int __init ubc_init(void) -{ - __raw_writel(0, UBC_CAMR0); - __raw_writel(0, UBC_CBR0); - __raw_writel(0, UBC_CBCR); - - __raw_writel(UBC_CRR_BIE | UBC_CRR_PCB, UBC_CRR0); - - /* dummy read for write posting */ - (void)__raw_readl(UBC_CRR0); +/* + * A dummy placeholder for early accesses until the CPUs get a chance to + * register their UBCs later in the boot process. + */ +static struct sh_ubc ubc_dummy = { .num_events = 0 }; - return 0; -} -arch_initcall(ubc_init); +static struct sh_ubc *sh_ubc __read_mostly = &ubc_dummy; /* * Install a perf counter breakpoint. @@ -62,10 +48,9 @@ arch_initcall(ubc_init); int arch_install_hw_breakpoint(struct perf_event *bp) { struct arch_hw_breakpoint *info = counter_arch_bp(bp); - struct ubc_context *ubc_ctx; int i; - for (i = 0; i < HBP_NUM; i++) { + for (i = 0; i < sh_ubc->num_events; i++) { struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); if (!*slot) { @@ -74,16 +59,11 @@ int arch_install_hw_breakpoint(struct perf_event *bp) } } - if (WARN_ONCE(i == HBP_NUM, "Can't find any breakpoint slot")) + if (WARN_ONCE(i == sh_ubc->num_events, "Can't find any breakpoint slot")) return -EBUSY; - ubc_ctx = &__get_cpu_var(ubc_ctx[i]); - - ubc_ctx->pc = info->address; - ubc_ctx->state = info->len | info->type; - - __raw_writel(UBC_CBR_CE | ubc_ctx->state, UBC_CBR0); - __raw_writel(ubc_ctx->pc, UBC_CAR0); + clk_enable(sh_ubc->clk); + sh_ubc->enable(info, i); return 0; } @@ -100,10 +80,9 @@ int arch_install_hw_breakpoint(struct perf_event *bp) void arch_uninstall_hw_breakpoint(struct perf_event *bp) { struct arch_hw_breakpoint *info = counter_arch_bp(bp); - struct ubc_context *ubc_ctx; int i; - for (i = 0; i < HBP_NUM; i++) { + for (i = 0; i < sh_ubc->num_events; i++) { struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); if (*slot == bp) { @@ -112,15 +91,11 @@ void arch_uninstall_hw_breakpoint(struct perf_event *bp) } } - if (WARN_ONCE(i == HBP_NUM, "Can't find any breakpoint slot")) + if (WARN_ONCE(i == sh_ubc->num_events, "Can't find any breakpoint slot")) return; - ubc_ctx = &__get_cpu_var(ubc_ctx[i]); - ubc_ctx->pc = 0; - ubc_ctx->state &= ~(info->len | info->type); - - __raw_writel(ubc_ctx->pc, UBC_CBR0); - __raw_writel(ubc_ctx->state, UBC_CAR0); + sh_ubc->disable(info, i); + clk_disable(sh_ubc->clk); } static int get_hbp_len(u16 hbp_len) @@ -182,10 +157,8 @@ static int arch_store_info(struct perf_event *bp) */ if (info->name) info->address = (unsigned long)kallsyms_lookup_name(info->name); - if (info->address) { - info->asid = get_asid(); + if (info->address) return 0; - } return -EINVAL; } @@ -335,7 +308,7 @@ void flush_ptrace_hw_breakpoint(struct task_struct *tsk) int i; struct thread_struct *t = &tsk->thread; - for (i = 0; i < HBP_NUM; i++) { + for (i = 0; i < sh_ubc->num_events; i++) { unregister_hw_breakpoint(t->ptrace_bps[i]); t->ptrace_bps[i] = NULL; } @@ -345,13 +318,32 @@ static int __kprobes hw_breakpoint_handler(struct die_args *args) { int cpu, i, rc = NOTIFY_STOP; struct perf_event *bp; - unsigned long val; + unsigned int cmf, resume_mask; + + /* + * Do an early return if none of the channels triggered. + */ + cmf = sh_ubc->triggered_mask(); + if (unlikely(!cmf)) + return NOTIFY_DONE; + + /* + * By default, resume all of the active channels. + */ + resume_mask = sh_ubc->active_mask(); - val = __raw_readl(UBC_CBR0); - __raw_writel(val & ~UBC_CBR_CE, UBC_CBR0); + /* + * Disable breakpoints during exception handling. + */ + sh_ubc->disable_all(); cpu = get_cpu(); - for (i = 0; i < HBP_NUM; i++) { + for (i = 0; i < sh_ubc->num_events; i++) { + unsigned long event_mask = (1 << i); + + if (likely(!(cmf & event_mask))) + continue; + /* * The counter may be concurrently released but that can only * occur from a call_rcu() path. We can then safely fetch @@ -361,24 +353,52 @@ static int __kprobes hw_breakpoint_handler(struct die_args *args) rcu_read_lock(); bp = per_cpu(bp_per_reg[i], cpu); - if (bp) { + if (bp) rc = NOTIFY_DONE; - } else { + + /* + * Reset the condition match flag to denote completion of + * exception handling. + */ + sh_ubc->clear_triggered_mask(event_mask); + + /* + * bp can be NULL due to concurrent perf counter + * removing. + */ + if (!bp) { rcu_read_unlock(); break; } + /* + * Don't restore the channel if the breakpoint is from + * ptrace, as it always operates in one-shot mode. + */ + if (bp->overflow_handler == ptrace_triggered) + resume_mask &= ~(1 << i); + perf_bp_event(bp, args->regs); + /* Deliver the signal to userspace */ + if (arch_check_va_in_userspace(bp->attr.bp_addr, + bp->attr.bp_len)) { + siginfo_t info; + + info.si_signo = args->signr; + info.si_errno = notifier_to_errno(rc); + info.si_code = TRAP_HWBKPT; + + force_sig_info(args->signr, &info, current); + } + rcu_read_unlock(); } - if (bp && bp->overflow_handler != ptrace_triggered) { - struct arch_hw_breakpoint *info = counter_arch_bp(bp); + if (cmf == 0) + rc = NOTIFY_DONE; - __raw_writel(UBC_CBR_CE | info->len | info->type, UBC_CBR0); - __raw_writel(info->address, UBC_CAR0); - } + sh_ubc->enable_all(resume_mask); put_cpu(); @@ -388,19 +408,9 @@ static int __kprobes hw_breakpoint_handler(struct die_args *args) BUILD_TRAP_HANDLER(breakpoint) { unsigned long ex = lookup_exception_vector(); - siginfo_t info; - int err; TRAP_HANDLER_DECL; - err = notify_die(DIE_BREAKPOINT, "breakpoint", regs, 0, ex, SIGTRAP); - if (err == NOTIFY_STOP) - return; - - /* Deliver the signal to userspace */ - info.si_signo = SIGTRAP; - info.si_errno = 0; - info.si_code = TRAP_HWBKPT; - force_sig_info(SIGTRAP, &info, current); + notify_die(DIE_BREAKPOINT, "breakpoint", regs, 0, ex, SIGTRAP); } /* @@ -417,8 +427,12 @@ int __kprobes hw_breakpoint_exceptions_notify(struct notifier_block *unused, /* * If the breakpoint hasn't been triggered by the UBC, it's * probably from a debugger, so don't do anything more here. + * + * This also permits the UBC interface clock to remain off for + * non-UBC breakpoints, as we don't need to check the triggered + * or active channel masks. */ - if (args->trapnr != 0x1e0) + if (args->trapnr != sh_ubc->trap_nr) return NOTIFY_DONE; return hw_breakpoint_handler(data); @@ -433,3 +447,17 @@ void hw_breakpoint_pmu_unthrottle(struct perf_event *bp) { /* TODO */ } + +int register_sh_ubc(struct sh_ubc *ubc) +{ + /* Bail if it's already assigned */ + if (sh_ubc != &ubc_dummy) + return -EBUSY; + sh_ubc = ubc; + + pr_info("HW Breakpoints: %s UBC support registered\n", ubc->name); + + WARN_ON(ubc->num_events > HBP_NUM); + + return 0; +} -- cgit v0.10.2 From 7025bec9125b0a02edcaf22c2dce753bf2c95480 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 5 Jan 2010 19:16:35 +0900 Subject: sh: Kill off dead UBC headers. Nothing is using these now, so kill them all off. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/ubc.h b/arch/sh/include/asm/ubc.h deleted file mode 100644 index e3c2968..0000000 --- a/arch/sh/include/asm/ubc.h +++ /dev/null @@ -1,70 +0,0 @@ -/* - * include/asm-sh/ubc.h - * - * Copyright (C) 1999 Niibe Yutaka - * Copyright (C) 2002, 2003 Paul Mundt - * - * This file is subject to the terms and conditions of the GNU General Public - * License. See the file "COPYING" in the main directory of this archive - * for more details. - */ -#ifndef __ASM_SH_UBC_H -#define __ASM_SH_UBC_H - -#ifdef __KERNEL__ -#include - -/* User Break Controller */ -#if defined(CONFIG_CPU_SUBTYPE_SH7707) || defined(CONFIG_CPU_SUBTYPE_SH7709) -#define UBC_TYPE_SH7729 (current_cpu_data.type == CPU_SH7729) -#else -#define UBC_TYPE_SH7729 0 -#endif - -#define BAMR_ASID (1 << 2) -#define BAMR_NONE 0 -#define BAMR_10 0x1 -#define BAMR_12 0x2 -#define BAMR_ALL 0x3 -#define BAMR_16 0x8 -#define BAMR_20 0x9 - -#define BBR_INST (1 << 4) -#define BBR_DATA (2 << 4) -#define BBR_READ (1 << 2) -#define BBR_WRITE (2 << 2) -#define BBR_BYTE 0x1 -#define BBR_HALF 0x2 -#define BBR_LONG 0x3 -#define BBR_QUAD (1 << 6) /* SH7750 */ -#define BBR_CPU (1 << 6) /* SH7709A,SH7729 */ -#define BBR_DMA (2 << 6) /* SH7709A,SH7729 */ - -#define BRCR_CMFA (1 << 15) -#define BRCR_CMFB (1 << 14) - -#if defined CONFIG_CPU_SH2A -#define BRCR_CMFCA (1 << 15) -#define BRCR_CMFCB (1 << 14) -#define BRCR_CMFDA (1 << 13) -#define BRCR_CMFDB (1 << 12) -#define BRCR_PCBB (1 << 6) /* 1: after execution */ -#define BRCR_PCBA (1 << 5) /* 1: after execution */ -#define BRCR_PCTE 0 -#else -#define BRCR_PCTE (1 << 11) -#define BRCR_PCBA (1 << 10) /* 1: after execution */ -#define BRCR_DBEB (1 << 7) -#define BRCR_PCBB (1 << 6) -#define BRCR_SEQ (1 << 3) -#define BRCR_UBDE (1 << 0) -#endif - -/* - * All SH parts have 2 UBC channels. I defy any hardware designer to - * invalidate this assertion. - */ -#define NR_UBC_CHANNELS 2 - -#endif /* __KERNEL__ */ -#endif /* __ASM_SH_UBC_H */ diff --git a/arch/sh/include/cpu-sh2/cpu/ubc.h b/arch/sh/include/cpu-sh2/cpu/ubc.h deleted file mode 100644 index ba0e87f..0000000 --- a/arch/sh/include/cpu-sh2/cpu/ubc.h +++ /dev/null @@ -1,32 +0,0 @@ -/* - * include/asm-sh/cpu-sh2/ubc.h - * - * Copyright (C) 2003 Paul Mundt - * - * This file is subject to the terms and conditions of the GNU General Public - * License. See the file "COPYING" in the main directory of this archive - * for more details. - */ -#ifndef __ASM_CPU_SH2_UBC_H -#define __ASM_CPU_SH2_UBC_H - -#define UBC_BARA 0xffffff40 -#define UBC_BAMRA 0xffffff44 -#define UBC_BBRA 0xffffff48 -#define UBC_BARB 0xffffff60 -#define UBC_BAMRB 0xffffff64 -#define UBC_BBRB 0xffffff68 -#define UBC_BDRB 0xffffff70 -#define UBC_BDMRB 0xffffff74 -#define UBC_BRCR 0xffffff78 - -/* - * We don't have any ASID changes to make in the UBC on the SH-2. - * - * Make these purposely invalid to track misuse. - */ -#define UBC_BASRA 0x00000000 -#define UBC_BASRB 0x00000000 - -#endif /* __ASM_CPU_SH2_UBC_H */ - diff --git a/arch/sh/include/cpu-sh3/cpu/ubc.h b/arch/sh/include/cpu-sh3/cpu/ubc.h deleted file mode 100644 index 4e6381d..0000000 --- a/arch/sh/include/cpu-sh3/cpu/ubc.h +++ /dev/null @@ -1,42 +0,0 @@ -/* - * include/asm-sh/cpu-sh3/ubc.h - * - * Copyright (C) 1999 Niibe Yutaka - * Copyright (C) 2003 Paul Mundt - * - * This file is subject to the terms and conditions of the GNU General Public - * License. See the file "COPYING" in the main directory of this archive - * for more details. - */ -#ifndef __ASM_CPU_SH3_UBC_H -#define __ASM_CPU_SH3_UBC_H - -#if defined(CONFIG_CPU_SUBTYPE_SH7710) || \ - defined(CONFIG_CPU_SUBTYPE_SH7720) || \ - defined(CONFIG_CPU_SUBTYPE_SH7721) -#define UBC_BARA 0xa4ffffb0 -#define UBC_BAMRA 0xa4ffffb4 -#define UBC_BBRA 0xa4ffffb8 -#define UBC_BASRA 0xffffffe4 -#define UBC_BARB 0xa4ffffa0 -#define UBC_BAMRB 0xa4ffffa4 -#define UBC_BBRB 0xa4ffffa8 -#define UBC_BASRB 0xffffffe8 -#define UBC_BDRB 0xa4ffff90 -#define UBC_BDMRB 0xa4ffff94 -#define UBC_BRCR 0xa4ffff98 -#else -#define UBC_BARA 0xffffffb0 -#define UBC_BAMRA 0xffffffb4 -#define UBC_BBRA 0xffffffb8 -#define UBC_BASRA 0xffffffe4 -#define UBC_BARB 0xffffffa0 -#define UBC_BAMRB 0xffffffa4 -#define UBC_BBRB 0xffffffa8 -#define UBC_BASRB 0xffffffe8 -#define UBC_BDRB 0xffffff90 -#define UBC_BDMRB 0xffffff94 -#define UBC_BRCR 0xffffff98 -#endif - -#endif /* __ASM_CPU_SH3_UBC_H */ diff --git a/arch/sh/include/cpu-sh4/cpu/ubc.h b/arch/sh/include/cpu-sh4/cpu/ubc.h deleted file mode 100644 index c86e170..0000000 --- a/arch/sh/include/cpu-sh4/cpu/ubc.h +++ /dev/null @@ -1,64 +0,0 @@ -/* - * include/asm-sh/cpu-sh4/ubc.h - * - * Copyright (C) 1999 Niibe Yutaka - * Copyright (C) 2003 Paul Mundt - * Copyright (C) 2006 Lineo Solutions Inc. support SH4A UBC - * - * This file is subject to the terms and conditions of the GNU General Public - * License. See the file "COPYING" in the main directory of this archive - * for more details. - */ -#ifndef __ASM_CPU_SH4_UBC_H -#define __ASM_CPU_SH4_UBC_H - -#if defined(CONFIG_CPU_SH4A) -#define UBC_CBR0 0xff200000 -#define UBC_CRR0 0xff200004 -#define UBC_CAR0 0xff200008 -#define UBC_CAMR0 0xff20000c -#define UBC_CBR1 0xff200020 -#define UBC_CRR1 0xff200024 -#define UBC_CAR1 0xff200028 -#define UBC_CAMR1 0xff20002c -#define UBC_CDR1 0xff200030 -#define UBC_CDMR1 0xff200034 -#define UBC_CETR1 0xff200038 -#define UBC_CCMFR 0xff200600 -#define UBC_CBCR 0xff200620 - -/* CBR */ -#define UBC_CBR_AIE (0x01<<30) -#define UBC_CBR_ID_INST (0x01<<4) -#define UBC_CBR_RW_READ (0x01<<1) -#define UBC_CBR_CE (0x01) - -#define UBC_CBR_AIV_MASK (0x00FF0000) -#define UBC_CBR_AIV_SHIFT (16) -#define UBC_CBR_AIV_SET(asid) (((asid)< #include #include -#ifdef CONFIG_SUPERH32 -#include -#endif /* * Generic wrapper for command line arguments to disable on-chip @@ -252,20 +249,19 @@ static void __init dsp_init(void) /** * sh_cpu_init * - * This is our initial entry point for each CPU, and is invoked on the boot - * CPU prior to calling start_kernel(). For SMP, a combination of this and - * start_secondary() will bring up each processor to a ready state prior - * to hand forking the idle loop. + * This is our initial entry point for each CPU, and is invoked on the + * boot CPU prior to calling start_kernel(). For SMP, a combination of + * this and start_secondary() will bring up each processor to a ready + * state prior to hand forking the idle loop. * - * We do all of the basic processor init here, including setting up the - * caches, FPU, DSP, kicking the UBC, etc. By the time start_kernel() is - * hit (and subsequently platform_setup()) things like determining the - * CPU subtype and initial configuration will all be done. + * We do all of the basic processor init here, including setting up + * the caches, FPU, DSP, etc. By the time start_kernel() is hit (and + * subsequently platform_setup()) things like determining the CPU + * subtype and initial configuration will all be done. * * Each processor family is still responsible for doing its own probing * and cache configuration in detect_cpu_and_cache_system(). */ - asmlinkage void __init sh_cpu_init(void) { current_thread_info()->cpu = hard_smp_processor_id(); diff --git a/arch/sh/kernel/process_32.c b/arch/sh/kernel/process_32.c index 7399d78..b94f9d9 100644 --- a/arch/sh/kernel/process_32.c +++ b/arch/sh/kernel/process_32.c @@ -30,7 +30,6 @@ #include #include #include -#include #include #include #include -- cgit v0.10.2 From 9fae4fb3ce38cc657ff0aba570ae875aae2806d4 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 5 Jan 2010 19:30:18 +0900 Subject: sh: Reclaim TIF_DEBUG. This was used by the old hw-breakpoints API, but now there is nothing is using it anymore, so just kill it off. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/thread_info.h b/arch/sh/include/asm/thread_info.h index e0856b2..1f3d927 100644 --- a/arch/sh/include/asm/thread_info.h +++ b/arch/sh/include/asm/thread_info.h @@ -118,7 +118,6 @@ extern void free_thread_info(struct thread_info *ti); #define TIF_SECCOMP 6 /* secure computing */ #define TIF_NOTIFY_RESUME 7 /* callback before returning to user */ #define TIF_SYSCALL_TRACEPOINT 8 /* for ftrace syscall instrumentation */ -#define TIF_DEBUG 9 /* uses UBC */ #define TIF_POLLING_NRFLAG 17 /* true if poll_idle() is polling TIF_NEED_RESCHED */ #define TIF_MEMDIE 18 #define TIF_FREEZE 19 /* Freezing for suspend */ @@ -131,7 +130,6 @@ extern void free_thread_info(struct thread_info *ti); #define _TIF_SECCOMP (1 << TIF_SECCOMP) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) #define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT) -#define _TIF_DEBUG (1 << TIF_DEBUG) #define _TIF_POLLING_NRFLAG (1 << TIF_POLLING_NRFLAG) #define _TIF_FREEZE (1 << TIF_FREEZE) -- cgit v0.10.2 From 722612cd51cf1b574c89dff57cc5dbedf1f645bb Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Tue, 5 Jan 2010 11:45:52 +0100 Subject: HID: fix parsing of local delimiter with size 0 Acording to HID standard 1.11, value 0 allows for size being 0. Local delimiter tag has has 0 one of the possible values. Therefore we need to handle this case properly, to be fully compliant with the specification. Reported-by: Marcin Tolysz Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index 6462c92..a4b496c 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -387,7 +387,8 @@ static int hid_parser_local(struct hid_parser *parser, struct hid_item *item) __u32 data; unsigned n; - if (item->size == 0) { + /* Local delimiter could have value 0, which allows size to be 0 */ + if (item->size == 0 && item->tag != HID_LOCAL_ITEM_TAG_DELIMITER) { dbg_hid("item data expected for local item\n"); return -1; } -- cgit v0.10.2 From 56d45b62ce622a003da972428fdbba2b42102efb Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 6 Jan 2010 14:45:14 +0900 Subject: sh: Fix up nommu build for out-of-line pgtable changes. pgtable_cache_init() has been moved out-of-line, so we also need a dummy definition for it on nommu to fix up the build. Signed-off-by: Paul Mundt diff --git a/arch/sh/mm/nommu.c b/arch/sh/mm/nommu.c index ac16c05..7694f50 100644 --- a/arch/sh/mm/nommu.c +++ b/arch/sh/mm/nommu.c @@ -94,3 +94,7 @@ void __init page_table_range_init(unsigned long start, unsigned long end, void __set_fixmap(enum fixed_addresses idx, unsigned long phys, pgprot_t prot) { } + +void pgtable_cache_init(void) +{ +} -- cgit v0.10.2 From 6fbfe8d7cd7e71ceb281c5a1b9de6e5e0dfbf1c8 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 6 Jan 2010 15:46:53 +0900 Subject: sh: select HAVE_HW_BREAKPOINT for all SUPERH32 CPUs. All SH CPUs (with the exception of sh64) support the UBC, so select HW_BREAKPOINT support by default. This fixes up the build for non-SH4A targets. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index fa76d62..f971cb7 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -42,7 +42,7 @@ config SUPERH32 select HAVE_FTRACE_NMI_ENTER if DYNAMIC_FTRACE select HAVE_FUNCTION_GRAPH_TRACER select HAVE_ARCH_KGDB - select HAVE_HW_BREAKPOINT if CPU_SH4A + select HAVE_HW_BREAKPOINT select PERF_EVENTS if HAVE_HW_BREAKPOINT select ARCH_HIBERNATION_POSSIBLE if MMU -- cgit v0.10.2 From 16295bec6398a3eedc9377e1af6ff4c71b98c300 Mon Sep 17 00:00:00 2001 From: Steffen Klassert Date: Wed, 6 Jan 2010 19:47:10 +1100 Subject: padata: Generic parallelization/serialization interface This patch introduces an interface to process data objects in parallel. The parallelized objects return after serialization in the same order as they were before the parallelization. Signed-off-by: Steffen Klassert Signed-off-by: Herbert Xu diff --git a/include/linux/padata.h b/include/linux/padata.h new file mode 100644 index 0000000..51611da --- /dev/null +++ b/include/linux/padata.h @@ -0,0 +1,88 @@ +/* + * padata.h - header for the padata parallelization interface + * + * Copyright (C) 2008, 2009 secunet Security Networks AG + * Copyright (C) 2008, 2009 Steffen Klassert + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifndef PADATA_H +#define PADATA_H + +#include +#include +#include + +struct padata_priv { + struct list_head list; + struct parallel_data *pd; + int cb_cpu; + int seq_nr; + int info; + void (*parallel)(struct padata_priv *padata); + void (*serial)(struct padata_priv *padata); +}; + +struct padata_list { + struct list_head list; + spinlock_t lock; +}; + +struct padata_queue { + struct padata_list parallel; + struct padata_list reorder; + struct padata_list serial; + struct work_struct pwork; + struct work_struct swork; + struct parallel_data *pd; + atomic_t num_obj; + int cpu_index; +}; + +struct parallel_data { + struct padata_instance *pinst; + struct padata_queue *queue; + atomic_t seq_nr; + atomic_t reorder_objects; + atomic_t refcnt; + unsigned int max_seq_nr; + cpumask_var_t cpumask; + spinlock_t lock; +}; + +struct padata_instance { + struct notifier_block cpu_notifier; + struct workqueue_struct *wq; + struct parallel_data *pd; + cpumask_var_t cpumask; + struct mutex lock; + u8 flags; +#define PADATA_INIT 1 +#define PADATA_RESET 2 +}; + +extern struct padata_instance *padata_alloc(const struct cpumask *cpumask, + struct workqueue_struct *wq); +extern void padata_free(struct padata_instance *pinst); +extern int padata_do_parallel(struct padata_instance *pinst, + struct padata_priv *padata, int cb_cpu); +extern void padata_do_serial(struct padata_priv *padata); +extern int padata_set_cpumask(struct padata_instance *pinst, + cpumask_var_t cpumask); +extern int padata_add_cpu(struct padata_instance *pinst, int cpu); +extern int padata_remove_cpu(struct padata_instance *pinst, int cpu); +extern void padata_start(struct padata_instance *pinst); +extern void padata_stop(struct padata_instance *pinst); +#endif diff --git a/init/Kconfig b/init/Kconfig index a23da9f..9fd23bc 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -1252,4 +1252,8 @@ source "block/Kconfig" config PREEMPT_NOTIFIERS bool +config PADATA + depends on SMP + bool + source "kernel/Kconfig.locks" diff --git a/kernel/Makefile b/kernel/Makefile index 864ff75..6aebdeb 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -100,6 +100,7 @@ obj-$(CONFIG_SLOW_WORK_DEBUG) += slow-work-debugfs.o obj-$(CONFIG_PERF_EVENTS) += perf_event.o obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o obj-$(CONFIG_USER_RETURN_NOTIFIER) += user-return-notifier.o +obj-$(CONFIG_PADATA) += padata.o ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER),y) # According to Alan Modra , the -fno-omit-frame-pointer is diff --git a/kernel/padata.c b/kernel/padata.c new file mode 100644 index 0000000..6f9bcb8 --- /dev/null +++ b/kernel/padata.c @@ -0,0 +1,690 @@ +/* + * padata.c - generic interface to process data streams in parallel + * + * Copyright (C) 2008, 2009 secunet Security Networks AG + * Copyright (C) 2008, 2009 Steffen Klassert + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include +#include +#include +#include +#include +#include +#include +#include + +#define MAX_SEQ_NR INT_MAX - NR_CPUS +#define MAX_OBJ_NUM 10000 * NR_CPUS + +static int padata_index_to_cpu(struct parallel_data *pd, int cpu_index) +{ + int cpu, target_cpu; + + target_cpu = cpumask_first(pd->cpumask); + for (cpu = 0; cpu < cpu_index; cpu++) + target_cpu = cpumask_next(target_cpu, pd->cpumask); + + return target_cpu; +} + +static int padata_cpu_hash(struct padata_priv *padata) +{ + int cpu_index; + struct parallel_data *pd; + + pd = padata->pd; + + /* + * Hash the sequence numbers to the cpus by taking + * seq_nr mod. number of cpus in use. + */ + cpu_index = padata->seq_nr % cpumask_weight(pd->cpumask); + + return padata_index_to_cpu(pd, cpu_index); +} + +static void padata_parallel_worker(struct work_struct *work) +{ + struct padata_queue *queue; + struct parallel_data *pd; + struct padata_instance *pinst; + LIST_HEAD(local_list); + + local_bh_disable(); + queue = container_of(work, struct padata_queue, pwork); + pd = queue->pd; + pinst = pd->pinst; + + spin_lock(&queue->parallel.lock); + list_replace_init(&queue->parallel.list, &local_list); + spin_unlock(&queue->parallel.lock); + + while (!list_empty(&local_list)) { + struct padata_priv *padata; + + padata = list_entry(local_list.next, + struct padata_priv, list); + + list_del_init(&padata->list); + + padata->parallel(padata); + } + + local_bh_enable(); +} + +/* + * padata_do_parallel - padata parallelization function + * + * @pinst: padata instance + * @padata: object to be parallelized + * @cb_cpu: cpu the serialization callback function will run on, + * must be in the cpumask of padata. + * + * The parallelization callback function will run with BHs off. + * Note: Every object which is parallelized by padata_do_parallel + * must be seen by padata_do_serial. + */ +int padata_do_parallel(struct padata_instance *pinst, + struct padata_priv *padata, int cb_cpu) +{ + int target_cpu, err; + struct padata_queue *queue; + struct parallel_data *pd; + + rcu_read_lock_bh(); + + pd = rcu_dereference(pinst->pd); + + err = 0; + if (!(pinst->flags & PADATA_INIT)) + goto out; + + err = -EBUSY; + if ((pinst->flags & PADATA_RESET)) + goto out; + + if (atomic_read(&pd->refcnt) >= MAX_OBJ_NUM) + goto out; + + err = -EINVAL; + if (!cpumask_test_cpu(cb_cpu, pd->cpumask)) + goto out; + + err = -EINPROGRESS; + atomic_inc(&pd->refcnt); + padata->pd = pd; + padata->cb_cpu = cb_cpu; + + if (unlikely(atomic_read(&pd->seq_nr) == pd->max_seq_nr)) + atomic_set(&pd->seq_nr, -1); + + padata->seq_nr = atomic_inc_return(&pd->seq_nr); + + target_cpu = padata_cpu_hash(padata); + queue = per_cpu_ptr(pd->queue, target_cpu); + + spin_lock(&queue->parallel.lock); + list_add_tail(&padata->list, &queue->parallel.list); + spin_unlock(&queue->parallel.lock); + + queue_work_on(target_cpu, pinst->wq, &queue->pwork); + +out: + rcu_read_unlock_bh(); + + return err; +} +EXPORT_SYMBOL(padata_do_parallel); + +static struct padata_priv *padata_get_next(struct parallel_data *pd) +{ + int cpu, num_cpus, empty, calc_seq_nr; + int seq_nr, next_nr, overrun, next_overrun; + struct padata_queue *queue, *next_queue; + struct padata_priv *padata; + struct padata_list *reorder; + + empty = 0; + next_nr = -1; + next_overrun = 0; + next_queue = NULL; + + num_cpus = cpumask_weight(pd->cpumask); + + for_each_cpu(cpu, pd->cpumask) { + queue = per_cpu_ptr(pd->queue, cpu); + reorder = &queue->reorder; + + /* + * Calculate the seq_nr of the object that should be + * next in this queue. + */ + overrun = 0; + calc_seq_nr = (atomic_read(&queue->num_obj) * num_cpus) + + queue->cpu_index; + + if (unlikely(calc_seq_nr > pd->max_seq_nr)) { + calc_seq_nr = calc_seq_nr - pd->max_seq_nr - 1; + overrun = 1; + } + + if (!list_empty(&reorder->list)) { + padata = list_entry(reorder->list.next, + struct padata_priv, list); + + seq_nr = padata->seq_nr; + BUG_ON(calc_seq_nr != seq_nr); + } else { + seq_nr = calc_seq_nr; + empty++; + } + + if (next_nr < 0 || seq_nr < next_nr + || (next_overrun && !overrun)) { + next_nr = seq_nr; + next_overrun = overrun; + next_queue = queue; + } + } + + padata = NULL; + + if (empty == num_cpus) + goto out; + + reorder = &next_queue->reorder; + + if (!list_empty(&reorder->list)) { + padata = list_entry(reorder->list.next, + struct padata_priv, list); + + if (unlikely(next_overrun)) { + for_each_cpu(cpu, pd->cpumask) { + queue = per_cpu_ptr(pd->queue, cpu); + atomic_set(&queue->num_obj, 0); + } + } + + spin_lock(&reorder->lock); + list_del_init(&padata->list); + atomic_dec(&pd->reorder_objects); + spin_unlock(&reorder->lock); + + atomic_inc(&next_queue->num_obj); + + goto out; + } + + if (next_nr % num_cpus == next_queue->cpu_index) { + padata = ERR_PTR(-ENODATA); + goto out; + } + + padata = ERR_PTR(-EINPROGRESS); +out: + return padata; +} + +static void padata_reorder(struct parallel_data *pd) +{ + struct padata_priv *padata; + struct padata_queue *queue; + struct padata_instance *pinst = pd->pinst; + +try_again: + if (!spin_trylock_bh(&pd->lock)) + goto out; + + while (1) { + padata = padata_get_next(pd); + + if (!padata || PTR_ERR(padata) == -EINPROGRESS) + break; + + if (PTR_ERR(padata) == -ENODATA) { + spin_unlock_bh(&pd->lock); + goto out; + } + + queue = per_cpu_ptr(pd->queue, padata->cb_cpu); + + spin_lock(&queue->serial.lock); + list_add_tail(&padata->list, &queue->serial.list); + spin_unlock(&queue->serial.lock); + + queue_work_on(padata->cb_cpu, pinst->wq, &queue->swork); + } + + spin_unlock_bh(&pd->lock); + + if (atomic_read(&pd->reorder_objects)) + goto try_again; + +out: + return; +} + +static void padata_serial_worker(struct work_struct *work) +{ + struct padata_queue *queue; + struct parallel_data *pd; + LIST_HEAD(local_list); + + local_bh_disable(); + queue = container_of(work, struct padata_queue, swork); + pd = queue->pd; + + spin_lock(&queue->serial.lock); + list_replace_init(&queue->serial.list, &local_list); + spin_unlock(&queue->serial.lock); + + while (!list_empty(&local_list)) { + struct padata_priv *padata; + + padata = list_entry(local_list.next, + struct padata_priv, list); + + list_del_init(&padata->list); + + padata->serial(padata); + atomic_dec(&pd->refcnt); + } + local_bh_enable(); +} + +/* + * padata_do_serial - padata serialization function + * + * @padata: object to be serialized. + * + * padata_do_serial must be called for every parallelized object. + * The serialization callback function will run with BHs off. + */ +void padata_do_serial(struct padata_priv *padata) +{ + int cpu; + struct padata_queue *queue; + struct parallel_data *pd; + + pd = padata->pd; + + cpu = get_cpu(); + queue = per_cpu_ptr(pd->queue, cpu); + + spin_lock(&queue->reorder.lock); + atomic_inc(&pd->reorder_objects); + list_add_tail(&padata->list, &queue->reorder.list); + spin_unlock(&queue->reorder.lock); + + put_cpu(); + + padata_reorder(pd); +} +EXPORT_SYMBOL(padata_do_serial); + +static struct parallel_data *padata_alloc_pd(struct padata_instance *pinst, + const struct cpumask *cpumask) +{ + int cpu, cpu_index, num_cpus; + struct padata_queue *queue; + struct parallel_data *pd; + + cpu_index = 0; + + pd = kzalloc(sizeof(struct parallel_data), GFP_KERNEL); + if (!pd) + goto err; + + pd->queue = alloc_percpu(struct padata_queue); + if (!pd->queue) + goto err_free_pd; + + if (!alloc_cpumask_var(&pd->cpumask, GFP_KERNEL)) + goto err_free_queue; + + for_each_possible_cpu(cpu) { + queue = per_cpu_ptr(pd->queue, cpu); + + queue->pd = pd; + + if (cpumask_test_cpu(cpu, cpumask) + && cpumask_test_cpu(cpu, cpu_active_mask)) { + queue->cpu_index = cpu_index; + cpu_index++; + } else + queue->cpu_index = -1; + + INIT_LIST_HEAD(&queue->reorder.list); + INIT_LIST_HEAD(&queue->parallel.list); + INIT_LIST_HEAD(&queue->serial.list); + spin_lock_init(&queue->reorder.lock); + spin_lock_init(&queue->parallel.lock); + spin_lock_init(&queue->serial.lock); + + INIT_WORK(&queue->pwork, padata_parallel_worker); + INIT_WORK(&queue->swork, padata_serial_worker); + atomic_set(&queue->num_obj, 0); + } + + cpumask_and(pd->cpumask, cpumask, cpu_active_mask); + + num_cpus = cpumask_weight(pd->cpumask); + pd->max_seq_nr = (MAX_SEQ_NR / num_cpus) * num_cpus - 1; + + atomic_set(&pd->seq_nr, -1); + atomic_set(&pd->reorder_objects, 0); + atomic_set(&pd->refcnt, 0); + pd->pinst = pinst; + spin_lock_init(&pd->lock); + + return pd; + +err_free_queue: + free_percpu(pd->queue); +err_free_pd: + kfree(pd); +err: + return NULL; +} + +static void padata_free_pd(struct parallel_data *pd) +{ + free_cpumask_var(pd->cpumask); + free_percpu(pd->queue); + kfree(pd); +} + +static void padata_replace(struct padata_instance *pinst, + struct parallel_data *pd_new) +{ + struct parallel_data *pd_old = pinst->pd; + + pinst->flags |= PADATA_RESET; + + rcu_assign_pointer(pinst->pd, pd_new); + + synchronize_rcu(); + + while (atomic_read(&pd_old->refcnt) != 0) + yield(); + + flush_workqueue(pinst->wq); + + padata_free_pd(pd_old); + + pinst->flags &= ~PADATA_RESET; +} + +/* + * padata_set_cpumask - set the cpumask that padata should use + * + * @pinst: padata instance + * @cpumask: the cpumask to use + */ +int padata_set_cpumask(struct padata_instance *pinst, + cpumask_var_t cpumask) +{ + struct parallel_data *pd; + int err = 0; + + might_sleep(); + + mutex_lock(&pinst->lock); + + pd = padata_alloc_pd(pinst, cpumask); + if (!pd) { + err = -ENOMEM; + goto out; + } + + cpumask_copy(pinst->cpumask, cpumask); + + padata_replace(pinst, pd); + +out: + mutex_unlock(&pinst->lock); + + return err; +} +EXPORT_SYMBOL(padata_set_cpumask); + +static int __padata_add_cpu(struct padata_instance *pinst, int cpu) +{ + struct parallel_data *pd; + + if (cpumask_test_cpu(cpu, cpu_active_mask)) { + pd = padata_alloc_pd(pinst, pinst->cpumask); + if (!pd) + return -ENOMEM; + + padata_replace(pinst, pd); + } + + return 0; +} + +/* + * padata_add_cpu - add a cpu to the padata cpumask + * + * @pinst: padata instance + * @cpu: cpu to add + */ +int padata_add_cpu(struct padata_instance *pinst, int cpu) +{ + int err; + + might_sleep(); + + mutex_lock(&pinst->lock); + + cpumask_set_cpu(cpu, pinst->cpumask); + err = __padata_add_cpu(pinst, cpu); + + mutex_unlock(&pinst->lock); + + return err; +} +EXPORT_SYMBOL(padata_add_cpu); + +static int __padata_remove_cpu(struct padata_instance *pinst, int cpu) +{ + struct parallel_data *pd; + + if (cpumask_test_cpu(cpu, cpu_online_mask)) { + pd = padata_alloc_pd(pinst, pinst->cpumask); + if (!pd) + return -ENOMEM; + + padata_replace(pinst, pd); + } + + return 0; +} + +/* + * padata_remove_cpu - remove a cpu from the padata cpumask + * + * @pinst: padata instance + * @cpu: cpu to remove + */ +int padata_remove_cpu(struct padata_instance *pinst, int cpu) +{ + int err; + + might_sleep(); + + mutex_lock(&pinst->lock); + + cpumask_clear_cpu(cpu, pinst->cpumask); + err = __padata_remove_cpu(pinst, cpu); + + mutex_unlock(&pinst->lock); + + return err; +} +EXPORT_SYMBOL(padata_remove_cpu); + +/* + * padata_start - start the parallel processing + * + * @pinst: padata instance to start + */ +void padata_start(struct padata_instance *pinst) +{ + might_sleep(); + + mutex_lock(&pinst->lock); + pinst->flags |= PADATA_INIT; + mutex_unlock(&pinst->lock); +} +EXPORT_SYMBOL(padata_start); + +/* + * padata_stop - stop the parallel processing + * + * @pinst: padata instance to stop + */ +void padata_stop(struct padata_instance *pinst) +{ + might_sleep(); + + mutex_lock(&pinst->lock); + pinst->flags &= ~PADATA_INIT; + mutex_unlock(&pinst->lock); +} +EXPORT_SYMBOL(padata_stop); + +static int __cpuinit padata_cpu_callback(struct notifier_block *nfb, + unsigned long action, void *hcpu) +{ + int err; + struct padata_instance *pinst; + int cpu = (unsigned long)hcpu; + + pinst = container_of(nfb, struct padata_instance, cpu_notifier); + + switch (action) { + case CPU_ONLINE: + case CPU_ONLINE_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + err = __padata_add_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + if (err) + return NOTIFY_BAD; + break; + + case CPU_DOWN_PREPARE: + case CPU_DOWN_PREPARE_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + err = __padata_remove_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + if (err) + return NOTIFY_BAD; + break; + + case CPU_UP_CANCELED: + case CPU_UP_CANCELED_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + __padata_remove_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + + case CPU_DOWN_FAILED: + case CPU_DOWN_FAILED_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + __padata_add_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + } + + return NOTIFY_OK; +} + +/* + * padata_alloc - allocate and initialize a padata instance + * + * @cpumask: cpumask that padata uses for parallelization + * @wq: workqueue to use for the allocated padata instance + */ +struct padata_instance *padata_alloc(const struct cpumask *cpumask, + struct workqueue_struct *wq) +{ + int err; + struct padata_instance *pinst; + struct parallel_data *pd; + + pinst = kzalloc(sizeof(struct padata_instance), GFP_KERNEL); + if (!pinst) + goto err; + + pd = padata_alloc_pd(pinst, cpumask); + if (!pd) + goto err_free_inst; + + rcu_assign_pointer(pinst->pd, pd); + + pinst->wq = wq; + + cpumask_copy(pinst->cpumask, cpumask); + + pinst->flags = 0; + + pinst->cpu_notifier.notifier_call = padata_cpu_callback; + pinst->cpu_notifier.priority = 0; + err = register_hotcpu_notifier(&pinst->cpu_notifier); + if (err) + goto err_free_pd; + + mutex_init(&pinst->lock); + + return pinst; + +err_free_pd: + padata_free_pd(pd); +err_free_inst: + kfree(pinst); +err: + return NULL; +} +EXPORT_SYMBOL(padata_alloc); + +/* + * padata_free - free a padata instance + * + * @ padata_inst: padata instance to free + */ +void padata_free(struct padata_instance *pinst) +{ + padata_stop(pinst); + + synchronize_rcu(); + + while (atomic_read(&pinst->pd->refcnt) != 0) + yield(); + + unregister_hotcpu_notifier(&pinst->cpu_notifier); + padata_free_pd(pinst->pd); + kfree(pinst); +} +EXPORT_SYMBOL(padata_free); -- cgit v0.10.2 From 54a6593d65e638ad7e1e8cc986159d76054dab4b Mon Sep 17 00:00:00 2001 From: Stefan Glasenhardt Date: Tue, 5 Jan 2010 23:30:30 +0100 Subject: HID: allow disabling hard-coded ISO-layout for Apple keyboards This patch adds a new option named "iso_layout" to the driver "hid-apple.ko", to allow disabling of the hard-coded ISO-layout. Disabling the hard-coded layout solves the problem that the kernel-module only works perfectly for the english/american version of the Apple aluminum keyboard. Other versions have swapped keys, e.g. the "<"-key is swapped with "^"-key on the german keyboard. There is a very long bug-entry on Launchpad to this problem: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/214786 Signed-off-by: Stefan Glasenhardt Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-apple.c b/drivers/hid/hid-apple.c index 4b96e7a..1433cbb 100644 --- a/drivers/hid/hid-apple.c +++ b/drivers/hid/hid-apple.c @@ -40,6 +40,11 @@ module_param(fnmode, uint, 0644); MODULE_PARM_DESC(fnmode, "Mode of fn key on Apple keyboards (0 = disabled, " "[1] = fkeyslast, 2 = fkeysfirst)"); +static unsigned int iso_layout = 1; +module_param(iso_layout, uint, 0644); +MODULE_PARM_DESC(iso_layout, "Enable/Disable hardcoded ISO-layout of the keyboard. " + "(0 = disabled, [1] = enabled)"); + struct apple_sc { unsigned long quirks; unsigned int fn_on; @@ -199,11 +204,13 @@ static int hidinput_apple_event(struct hid_device *hid, struct input_dev *input, } } - if (asc->quirks & APPLE_ISO_KEYBOARD) { - trans = apple_find_translation(apple_iso_keyboard, usage->code); - if (trans) { - input_event(input, usage->type, trans->to, value); - return 1; + if (iso_layout) { + if (asc->quirks & APPLE_ISO_KEYBOARD) { + trans = apple_find_translation(apple_iso_keyboard, usage->code); + if (trans) { + input_event(input, usage->type, trans->to, value); + return 1; + } } } -- cgit v0.10.2 From 809826a389040e0ad9d646b587bccc0e34691afd Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:34 +0800 Subject: tracing: Have __dynamic_array() define a field This is part of a patch set that removes the show_format method in the ftrace event macros. This patch set requires that all fields are added to the ftrace_event_call->fields. This patch changes __dynamic_array() to call trace_define_field() to include fields that use __dynamic_array(). Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D36.8090100@cn.fujitsu.com> Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index d4fa5dc..9978a4f 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -175,7 +175,12 @@ ftrace_format_##name(struct ftrace_event_call *unused, \ return ret; #undef __dynamic_array -#define __dynamic_array(type, item) +#define __dynamic_array(type, item) \ + ret = trace_define_field(event_call, #type, #item, \ + offsetof(typeof(field), item), \ + 0, is_signed_type(type), FILTER_OTHER);\ + if (ret) \ + return ret; #undef FTRACE_ENTRY #define FTRACE_ENTRY(name, struct_name, id, tstruct, print) \ -- cgit v0.10.2 From 509e760cd91c831983097ae174cb6c0b8c6c8e6b Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:42 +0800 Subject: tracing: Add print_fmt field This is part of a patch set that removes the show_format method in the ftrace event macros. The print_fmt field is added to hold the string that shows the print_fmt in the event format files. This patch only adds the field but it is currently not used. Later patches will use this field to enable us to remove the show_format field and function. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D3E.2000704@cn.fujitsu.com> Signed-off-by: Steven Rostedt diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h index 2233c98..bd23d8e 100644 --- a/include/linux/ftrace_event.h +++ b/include/linux/ftrace_event.h @@ -121,6 +121,7 @@ struct ftrace_event_call { int (*regfunc)(struct ftrace_event_call *); void (*unregfunc)(struct ftrace_event_call *); int id; + const char *print_fmt; int (*raw_init)(struct ftrace_event_call *); int (*show_format)(struct ftrace_event_call *, struct trace_seq *); diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h index c6fe03e..3351d85 100644 --- a/include/trace/ftrace.h +++ b/include/trace/ftrace.h @@ -722,8 +722,20 @@ static struct trace_event ftrace_event_type_##call = { \ #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) +#undef __entry +#define __entry REC + +#undef __print_flags +#undef __print_symbolic +#undef __get_dynamic_array +#undef __get_str + +#undef TP_printk +#define TP_printk(fmt, args...) "\"" fmt "\", " __stringify(args) + #undef DECLARE_EVENT_CLASS -#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) \ +static const char print_fmt_##call[] = print; #undef DEFINE_EVENT #define DEFINE_EVENT(template, call, proto, args) \ @@ -737,6 +749,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .raw_init = trace_event_raw_init, \ .regfunc = ftrace_raw_reg_event_##call, \ .unregfunc = ftrace_raw_unreg_event_##call, \ + .print_fmt = print_fmt_##template, \ .show_format = ftrace_format_##template, \ .define_fields = ftrace_define_fields_##template, \ _TRACE_PROFILE_INIT(call) \ @@ -745,6 +758,8 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ #undef DEFINE_EVENT_PRINT #define DEFINE_EVENT_PRINT(template, call, proto, args, print) \ \ +static const char print_fmt_##call[] = print; \ + \ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ @@ -754,6 +769,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .raw_init = trace_event_raw_init, \ .regfunc = ftrace_raw_reg_event_##call, \ .unregfunc = ftrace_raw_unreg_event_##call, \ + .print_fmt = print_fmt_##call, \ .show_format = ftrace_format_##call, \ .define_fields = ftrace_define_fields_##template, \ _TRACE_PROFILE_INIT(call) \ @@ -837,6 +853,16 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ #ifdef CONFIG_EVENT_PROFILE +#undef __entry +#define __entry entry + +#undef __get_dynamic_array +#define __get_dynamic_array(field) \ + ((void *)__entry + (__entry->__data_loc_##field & 0xffff)) + +#undef __get_str +#define __get_str(field) (char *)__get_dynamic_array(field) + #undef __perf_addr #define __perf_addr(a) __addr = (a) diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index 9978a4f..95d14b6 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -203,6 +203,9 @@ static int ftrace_raw_init_event(struct ftrace_event_call *call) return 0; } +#undef __entry +#define __entry REC + #undef __field #define __field(type, item) @@ -218,6 +221,9 @@ static int ftrace_raw_init_event(struct ftrace_event_call *call) #undef __dynamic_array #define __dynamic_array(type, item) +#undef F_printk +#define F_printk(fmt, args...) #fmt ", " __stringify(args) + #undef FTRACE_ENTRY #define FTRACE_ENTRY(call, struct_name, type, tstruct, print) \ \ @@ -228,6 +234,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .id = type, \ .system = __stringify(TRACE_SYSTEM), \ .raw_init = ftrace_raw_init_event, \ + .print_fmt = print, \ .show_format = ftrace_format_##call, \ .define_fields = ftrace_define_fields_##call, \ }; \ -- cgit v0.10.2 From 50307a45f8515f6244e3b08e6b19824b9fbfe293 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:45 +0800 Subject: tracing/syscalls: Init print_fmt for syscall events This is part of a patch set that removes the show_format method in the ftrace event macros. Add the print_fmt initialization to the syscall events. The print_fmt is still not used, but will be in the follow up patches. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D41.609@cn.fujitsu.com> Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 75289f3..1352b0a 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -191,6 +191,67 @@ int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s) return trace_seq_putc(s, '\n'); } +static +int __set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len) +{ + int i; + int pos = 0; + + /* When len=0, we just calculate the needed length */ +#define LEN_OR_ZERO (len ? len - pos : 0) + + pos += snprintf(buf + pos, LEN_OR_ZERO, "\""); + for (i = 0; i < entry->nb_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, "%s: 0x%%0%zulx%s", + entry->args[i], sizeof(unsigned long), + i == entry->nb_args - 1 ? "" : ", "); + } + pos += snprintf(buf + pos, LEN_OR_ZERO, "\""); + + for (i = 0; i < entry->nb_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, + ", ((unsigned long)(REC->%s))", entry->args[i]); + } + +#undef LEN_OR_ZERO + + /* return the length of print_fmt */ + return pos; +} + +static int set_syscall_print_fmt(struct ftrace_event_call *call) +{ + char *print_fmt; + int len; + struct syscall_metadata *entry = call->data; + + if (entry->enter_event != call) { + call->print_fmt = "\"0x%lx\", REC->ret"; + return 0; + } + + /* First: called with 0 length to calculate the needed length */ + len = __set_enter_print_fmt(entry, NULL, 0); + + print_fmt = kmalloc(len + 1, GFP_KERNEL); + if (!print_fmt) + return -ENOMEM; + + /* Second: actually write the @print_fmt */ + __set_enter_print_fmt(entry, print_fmt, len + 1); + call->print_fmt = print_fmt; + + return 0; +} + +static void free_syscall_print_fmt(struct ftrace_event_call *call) +{ + struct syscall_metadata *entry = call->data; + + if (entry->enter_event == call) + kfree(call->print_fmt); +} + int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s) { int ret; @@ -386,9 +447,14 @@ int init_syscall_trace(struct ftrace_event_call *call) { int id; + if (set_syscall_print_fmt(call) < 0) + return -ENOMEM; + id = register_ftrace_event(call->event); - if (!id) + if (!id) { + free_syscall_print_fmt(call); return -ENODEV; + } call->id = id; INIT_LIST_HEAD(&call->fields); return 0; -- cgit v0.10.2 From a342a0280b981c130e32dbb94dbd3a57959c4d04 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:49 +0800 Subject: tracing/kprobes: Init print_fmt for kprobe events This is part of a patch set that removes the show_format method in the ftrace event macros. Add the print_fmt initialization to the kprobe events. The print_fmt is still not used, but will be in the follow up patches. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D45.3080100@cn.fujitsu.com> Acked-by: Masami Hiramatsu Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 6ea90c0..147491d 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1250,6 +1250,62 @@ static int kretprobe_event_show_format(struct ftrace_event_call *call, ", REC->" FIELD_STRING_RETIP); } +static int __set_print_fmt(struct trace_probe *tp, char *buf, int len) +{ + int i; + int pos = 0; + + const char *fmt, *arg; + + if (!probe_is_return(tp)) { + fmt = "(%lx)"; + arg = "REC->" FIELD_STRING_IP; + } else { + fmt = "(%lx <- %lx)"; + arg = "REC->" FIELD_STRING_FUNC ", REC->" FIELD_STRING_RETIP; + } + + /* When len=0, we just calculate the needed length */ +#define LEN_OR_ZERO (len ? len - pos : 0) + + pos += snprintf(buf + pos, LEN_OR_ZERO, "\"%s", fmt); + + for (i = 0; i < tp->nr_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, " %s=%%lx", + tp->args[i].name); + } + + pos += snprintf(buf + pos, LEN_OR_ZERO, "\", %s", arg); + + for (i = 0; i < tp->nr_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, ", REC->%s", + tp->args[i].name); + } + +#undef LEN_OR_ZERO + + /* return the length of print_fmt */ + return pos; +} + +static int set_print_fmt(struct trace_probe *tp) +{ + int len; + char *print_fmt; + + /* First: called with 0 length to calculate the needed length */ + len = __set_print_fmt(tp, NULL, 0); + print_fmt = kmalloc(len + 1, GFP_KERNEL); + if (!print_fmt) + return -ENOMEM; + + /* Second: actually write the @print_fmt */ + __set_print_fmt(tp, print_fmt, len + 1); + tp->call.print_fmt = print_fmt; + + return 0; +} + #ifdef CONFIG_EVENT_PROFILE /* Kprobe profile handler */ @@ -1456,10 +1512,14 @@ static int register_probe_event(struct trace_probe *tp) call->show_format = kprobe_event_show_format; call->define_fields = kprobe_event_define_fields; } + if (set_print_fmt(tp) < 0) + return -ENOMEM; call->event = &tp->event; call->id = register_ftrace_event(&tp->event); - if (!call->id) + if (!call->id) { + kfree(call->print_fmt); return -ENODEV; + } call->enabled = 0; call->regfunc = probe_event_enable; call->unregfunc = probe_event_disable; @@ -1472,6 +1532,7 @@ static int register_probe_event(struct trace_probe *tp) ret = trace_add_event_call(call); if (ret) { pr_info("Failed to register kprobe event: %s\n", call->name); + kfree(call->print_fmt); unregister_ftrace_event(&tp->event); } return ret; @@ -1481,6 +1542,7 @@ static void unregister_probe_event(struct trace_probe *tp) { /* tp->event is unregistered in trace_remove_event_call() */ trace_remove_event_call(&tp->call); + kfree(tp->call.print_fmt); } /* Make a debugfs interface for controling probe points */ -- cgit v0.10.2 From c7ef3a9004201bca90626db246a19dadd2c29c9b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 28 Dec 2009 21:13:59 -0500 Subject: tracing: Have syscall tracing call its own init function In the clean up of having all events call one specific function, the syscall event init was changed to call this helper function. With the new print_fmt updates, the syscalls need to do special initializations. This patch converts the syscall events to call its own init function again. Cc: Lai Jiangshan Cc: Li Zefan Signed-off-by: Steven Rostedt diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index 207466a..ed353d2 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -143,7 +143,7 @@ struct perf_event_attr; .name = "sys_enter"#sname, \ .system = "syscalls", \ .event = &enter_syscall_print_##sname, \ - .raw_init = trace_event_raw_init, \ + .raw_init = init_syscall_trace, \ .show_format = syscall_enter_format, \ .define_fields = syscall_enter_define_fields, \ .regfunc = reg_event_syscall_enter, \ @@ -165,7 +165,7 @@ struct perf_event_attr; .name = "sys_exit"#sname, \ .system = "syscalls", \ .event = &exit_syscall_print_##sname, \ - .raw_init = trace_event_raw_init, \ + .raw_init = init_syscall_trace, \ .show_format = syscall_exit_format, \ .define_fields = syscall_exit_define_fields, \ .regfunc = reg_event_syscall_exit, \ diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 1352b0a..a78e863 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -450,14 +450,14 @@ int init_syscall_trace(struct ftrace_event_call *call) if (set_syscall_print_fmt(call) < 0) return -ENOMEM; - id = register_ftrace_event(call->event); - if (!id) { + id = trace_event_raw_init(call); + + if (id < 0) { free_syscall_print_fmt(call); - return -ENODEV; + return id; } - call->id = id; - INIT_LIST_HEAD(&call->fields); - return 0; + + return id; } int __init init_ftrace_syscalls(void) -- cgit v0.10.2 From 5a65e956220efc2421e21ee56d6153fd5c533a95 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:53 +0800 Subject: tracing: Use defined fields and print_fmt to print formats The calls ftrace_format_##call() and ftrace_define_fields_##call() are almost duplicate in functionality. With the addition of the print_fmt in previous patches, these two functions can be merged into one. The trace_define_field() defines the fields and links them into the struct ftrace_event_call. The previous patches introduced the print_fmt field and this can now be used with the trace_define_field() to create the event format file fields and print_fmt field. The struct ftrace_event_call->fields are used to print the fields The struct ftrace_event_call->print_fmt is used to print the "print fmt: XXXXXXXXXXX" line. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D49.5000006@cn.fujitsu.com> Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 189b09b..250ec86 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -528,33 +528,16 @@ extern char *__bad_type_size(void); #type, "common_" #name, offsetof(typeof(field), name), \ sizeof(field.name), is_signed_type(type) -static int trace_write_header(struct trace_seq *s) -{ - struct trace_entry field; - - /* struct trace_entry */ - return trace_seq_printf(s, - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\n", - FIELD(unsigned short, type), - FIELD(unsigned char, flags), - FIELD(unsigned char, preempt_count), - FIELD(int, pid), - FIELD(int, lock_depth)); -} - static ssize_t event_format_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { struct ftrace_event_call *call = filp->private_data; + struct ftrace_event_field *field; struct trace_seq *s; + int common_field_count = 5; char *buf; - int r; + int r = 0; if (*ppos) return 0; @@ -565,14 +548,48 @@ event_format_read(struct file *filp, char __user *ubuf, size_t cnt, trace_seq_init(s); - /* If any of the first writes fail, so will the show_format. */ - trace_seq_printf(s, "name: %s\n", call->name); trace_seq_printf(s, "ID: %d\n", call->id); trace_seq_printf(s, "format:\n"); - trace_write_header(s); - r = call->show_format(call, s); + list_for_each_entry_reverse(field, &call->fields, link) { + /* + * Smartly shows the array type(except dynamic array). + * Normal: + * field:TYPE VAR + * If TYPE := TYPE[LEN], it is shown: + * field:TYPE VAR[LEN] + */ + const char *array_descriptor = strchr(field->type, '['); + + if (!strncmp(field->type, "__data_loc", 10)) + array_descriptor = NULL; + + if (!array_descriptor) { + r = trace_seq_printf(s, "\tfield:%s %s;\toffset:%u;" + "\tsize:%u;\tsigned:%d;\n", + field->type, field->name, field->offset, + field->size, !!field->is_signed); + } else { + r = trace_seq_printf(s, "\tfield:%.*s %s%s;\toffset:%u;" + "\tsize:%u;\tsigned:%d;\n", + (int)(array_descriptor - field->type), + field->type, field->name, + array_descriptor, field->offset, + field->size, !!field->is_signed); + } + + if (--common_field_count == 0) + r = trace_seq_printf(s, "\n"); + + if (!r) + break; + } + + if (r) + r = trace_seq_printf(s, "\nprint fmt: %s\n", + call->print_fmt); + if (!r) { /* * ug! The format output is bigger than a PAGE!! -- cgit v0.10.2 From 0fa0edaf32b9a78b9854f1da98d4511a501089b0 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 15 Dec 2009 15:39:57 +0800 Subject: tracing: Remove show_format and related macros from TRACE_EVENT The previous patches added the use of print_fmt string and changes the trace_define_field() function to also create the fields and format output for the event format files. text data bss dec hex filename 5857201 1355780 9336808 16549789 fc879d vmlinux 5884589 1351684 9337896 16574169 fce6d9 vmlinux-orig The above shows the size of the vmlinux after this patch set compared to the vmlinux-orig which is before the patch set. This saves us 27k on text, 1k on bss and adds just 4k of data. The total savings of 24k in size. Signed-off-by: Lai Jiangshan LKML-Reference: <4B273D4D.40604@cn.fujitsu.com> Acked-by: Masami Hiramatsu Signed-off-by: Steven Rostedt diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h index bd23d8e..84a5629 100644 --- a/include/linux/ftrace_event.h +++ b/include/linux/ftrace_event.h @@ -123,8 +123,6 @@ struct ftrace_event_call { int id; const char *print_fmt; int (*raw_init)(struct ftrace_event_call *); - int (*show_format)(struct ftrace_event_call *, - struct trace_seq *); int (*define_fields)(struct ftrace_event_call *); struct list_head fields; int filter_active; diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index ed353d2..7b21969 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -144,7 +144,6 @@ struct perf_event_attr; .system = "syscalls", \ .event = &enter_syscall_print_##sname, \ .raw_init = init_syscall_trace, \ - .show_format = syscall_enter_format, \ .define_fields = syscall_enter_define_fields, \ .regfunc = reg_event_syscall_enter, \ .unregfunc = unreg_event_syscall_enter, \ @@ -166,7 +165,6 @@ struct perf_event_attr; .system = "syscalls", \ .event = &exit_syscall_print_##sname, \ .raw_init = init_syscall_trace, \ - .show_format = syscall_exit_format, \ .define_fields = syscall_exit_define_fields, \ .regfunc = reg_event_syscall_exit, \ .unregfunc = unreg_event_syscall_exit, \ diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h index 3351d85..df65b99 100644 --- a/include/trace/ftrace.h +++ b/include/trace/ftrace.h @@ -131,130 +131,6 @@ #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) /* - * Setup the showing format of trace point. - * - * int - * ftrace_format_##call(struct trace_seq *s) - * { - * struct ftrace_raw_##call field; - * int ret; - * - * ret = trace_seq_printf(s, #type " " #item ";" - * " offset:%u; size:%u;\n", - * offsetof(struct ftrace_raw_##call, item), - * sizeof(field.type)); - * - * } - */ - -#undef TP_STRUCT__entry -#define TP_STRUCT__entry(args...) args - -#undef __field -#define __field(type, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%u;\tsize:%u;\tsigned:%u;\n", \ - (unsigned int)offsetof(typeof(field), item), \ - (unsigned int)sizeof(field.item), \ - (unsigned int)is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __field_ext -#define __field_ext(type, item, filter_type) __field(type, item) - -#undef __array -#define __array(type, item, len) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item "[" #len "];\t" \ - "offset:%u;\tsize:%u;\tsigned:%u;\n", \ - (unsigned int)offsetof(typeof(field), item), \ - (unsigned int)sizeof(field.item), \ - (unsigned int)is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __dynamic_array -#define __dynamic_array(type, item, len) \ - ret = trace_seq_printf(s, "\tfield:__data_loc " #type "[] " #item ";\t"\ - "offset:%u;\tsize:%u;\tsigned:%u;\n", \ - (unsigned int)offsetof(typeof(field), \ - __data_loc_##item), \ - (unsigned int)sizeof(field.__data_loc_##item), \ - (unsigned int)is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __string -#define __string(item, src) __dynamic_array(char, item, -1) - -#undef __entry -#define __entry REC - -#undef __print_symbolic -#undef __get_dynamic_array -#undef __get_str - -#undef TP_printk -#define TP_printk(fmt, args...) "\"%s\", %s\n", fmt, __stringify(args) - -#undef TP_fast_assign -#define TP_fast_assign(args...) args - -#undef TP_perf_assign -#define TP_perf_assign(args...) - -#undef DECLARE_EVENT_CLASS -#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, func, print) \ -static int \ -ftrace_format_setup_##call(struct ftrace_event_call *unused, \ - struct trace_seq *s) \ -{ \ - struct ftrace_raw_##call field __attribute__((unused)); \ - int ret = 0; \ - \ - tstruct; \ - \ - return ret; \ -} \ - \ -static int \ -ftrace_format_##call(struct ftrace_event_call *unused, \ - struct trace_seq *s) \ -{ \ - int ret = 0; \ - \ - ret = ftrace_format_setup_##call(unused, s); \ - if (!ret) \ - return ret; \ - \ - ret = trace_seq_printf(s, "\nprint fmt: " print); \ - \ - return ret; \ -} - -#undef DEFINE_EVENT -#define DEFINE_EVENT(template, name, proto, args) - -#undef DEFINE_EVENT_PRINT -#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ -static int \ -ftrace_format_##name(struct ftrace_event_call *unused, \ - struct trace_seq *s) \ -{ \ - int ret = 0; \ - \ - ret = ftrace_format_setup_##template(unused, s); \ - if (!ret) \ - return ret; \ - \ - trace_seq_printf(s, "\nprint fmt: " print); \ - \ - return ret; \ -} - -#include TRACE_INCLUDE(TRACE_INCLUDE_FILE) - -/* * Stage 3 of the trace events. * * Override the macros in to include the following: @@ -622,7 +498,6 @@ static void ftrace_profile_disable_##name(struct ftrace_event_call *unused)\ * .raw_init = trace_event_raw_init, * .regfunc = ftrace_reg_event_, * .unregfunc = ftrace_unreg_event_, - * .show_format = ftrace_format_, * } * */ @@ -657,6 +532,12 @@ static void ftrace_profile_disable_##name(struct ftrace_event_call *unused)\ #define __assign_str(dst, src) \ strcpy(__get_str(dst), src); +#undef TP_fast_assign +#define TP_fast_assign(args...) args + +#undef TP_perf_assign +#define TP_perf_assign(args...) + #undef DECLARE_EVENT_CLASS #define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) \ \ @@ -750,7 +631,6 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .regfunc = ftrace_raw_reg_event_##call, \ .unregfunc = ftrace_raw_unreg_event_##call, \ .print_fmt = print_fmt_##template, \ - .show_format = ftrace_format_##template, \ .define_fields = ftrace_define_fields_##template, \ _TRACE_PROFILE_INIT(call) \ } @@ -770,7 +650,6 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .regfunc = ftrace_raw_reg_event_##call, \ .unregfunc = ftrace_raw_unreg_event_##call, \ .print_fmt = print_fmt_##call, \ - .show_format = ftrace_format_##call, \ .define_fields = ftrace_define_fields_##template, \ _TRACE_PROFILE_INIT(call) \ } diff --git a/include/trace/syscall.h b/include/trace/syscall.h index 961fda3..8cd4102 100644 --- a/include/trace/syscall.h +++ b/include/trace/syscall.h @@ -34,10 +34,6 @@ struct syscall_metadata { extern unsigned long arch_syscall_addr(int nr); extern int init_syscall_trace(struct ftrace_event_call *call); -extern int syscall_enter_format(struct ftrace_event_call *call, - struct trace_seq *s); -extern int syscall_exit_format(struct ftrace_event_call *call, - struct trace_seq *s); extern int syscall_enter_define_fields(struct ftrace_event_call *call); extern int syscall_exit_define_fields(struct ftrace_event_call *call); extern int reg_event_syscall_enter(struct ftrace_event_call *call); diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 250ec86..c2a3077 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -520,14 +520,6 @@ out: return ret; } -extern char *__bad_type_size(void); - -#undef FIELD -#define FIELD(type, name) \ - sizeof(type) != sizeof(field.name) ? __bad_type_size() : \ - #type, "common_" #name, offsetof(typeof(field), name), \ - sizeof(field.name), is_signed_type(type) - static ssize_t event_format_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) @@ -965,10 +957,6 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events, filter); } - /* A trace may not want to export its format */ - if (!call->show_format) - return 0; - trace_create_file("format", 0444, call->dir, call, format); diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index 95d14b6..e091f64 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -62,78 +62,6 @@ static void __always_unused ____ftrace_check_##name(void) \ #include "trace_entries.h" - -#undef __field -#define __field(type, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item), is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __field_desc -#define __field_desc(type, container, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), container.item), \ - sizeof(field.container.item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __array -#define __array(type, item, len) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item "[" #len "];\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item), is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __array_desc -#define __array_desc(type, container, item, len) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item "[" #len "];\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), container.item), \ - sizeof(field.container.item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __dynamic_array -#define __dynamic_array(type, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:0;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef F_printk -#define F_printk(fmt, args...) "%s, %s\n", #fmt, __stringify(args) - -#undef __entry -#define __entry REC - -#undef FTRACE_ENTRY -#define FTRACE_ENTRY(name, struct_name, id, tstruct, print) \ -static int \ -ftrace_format_##name(struct ftrace_event_call *unused, \ - struct trace_seq *s) \ -{ \ - struct struct_name field __attribute__((unused)); \ - int ret = 0; \ - \ - tstruct; \ - \ - trace_seq_printf(s, "\nprint fmt: " print); \ - \ - return ret; \ -} - -#include "trace_entries.h" - #undef __field #define __field(type, item) \ ret = trace_define_field(event_call, #type, #item, \ @@ -235,7 +163,6 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .system = __stringify(TRACE_SYSTEM), \ .raw_init = ftrace_raw_init_event, \ .print_fmt = print, \ - .show_format = ftrace_format_##call, \ .define_fields = ftrace_define_fields_##call, \ }; \ diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 147491d..c990299 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1174,82 +1174,6 @@ static int kretprobe_event_define_fields(struct ftrace_event_call *event_call) return 0; } -static int __probe_event_show_format(struct trace_seq *s, - struct trace_probe *tp, const char *fmt, - const char *arg) -{ - int i; - - /* Show format */ - if (!trace_seq_printf(s, "\nprint fmt: \"%s", fmt)) - return 0; - - for (i = 0; i < tp->nr_args; i++) - if (!trace_seq_printf(s, " %s=%%lx", tp->args[i].name)) - return 0; - - if (!trace_seq_printf(s, "\", %s", arg)) - return 0; - - for (i = 0; i < tp->nr_args; i++) - if (!trace_seq_printf(s, ", REC->%s", tp->args[i].name)) - return 0; - - return trace_seq_puts(s, "\n"); -} - -#undef SHOW_FIELD -#define SHOW_FIELD(type, item, name) \ - do { \ - ret = trace_seq_printf(s, "\tfield:" #type " %s;\t" \ - "offset:%u;\tsize:%u;\tsigned:%d;\n", name,\ - (unsigned int)offsetof(typeof(field), item),\ - (unsigned int)sizeof(type), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; \ - } while (0) - -static int kprobe_event_show_format(struct ftrace_event_call *call, - struct trace_seq *s) -{ - struct kprobe_trace_entry field __attribute__((unused)); - int ret, i; - struct trace_probe *tp = (struct trace_probe *)call->data; - - SHOW_FIELD(unsigned long, ip, FIELD_STRING_IP); - SHOW_FIELD(int, nargs, FIELD_STRING_NARGS); - - /* Show fields */ - for (i = 0; i < tp->nr_args; i++) - SHOW_FIELD(unsigned long, args[i], tp->args[i].name); - trace_seq_puts(s, "\n"); - - return __probe_event_show_format(s, tp, "(%lx)", - "REC->" FIELD_STRING_IP); -} - -static int kretprobe_event_show_format(struct ftrace_event_call *call, - struct trace_seq *s) -{ - struct kretprobe_trace_entry field __attribute__((unused)); - int ret, i; - struct trace_probe *tp = (struct trace_probe *)call->data; - - SHOW_FIELD(unsigned long, func, FIELD_STRING_FUNC); - SHOW_FIELD(unsigned long, ret_ip, FIELD_STRING_RETIP); - SHOW_FIELD(int, nargs, FIELD_STRING_NARGS); - - /* Show fields */ - for (i = 0; i < tp->nr_args; i++) - SHOW_FIELD(unsigned long, args[i], tp->args[i].name); - trace_seq_puts(s, "\n"); - - return __probe_event_show_format(s, tp, "(%lx <- %lx)", - "REC->" FIELD_STRING_FUNC - ", REC->" FIELD_STRING_RETIP); -} - static int __set_print_fmt(struct trace_probe *tp, char *buf, int len) { int i; @@ -1504,12 +1428,10 @@ static int register_probe_event(struct trace_probe *tp) if (probe_is_return(tp)) { tp->event.trace = print_kretprobe_event; call->raw_init = probe_event_raw_init; - call->show_format = kretprobe_event_show_format; call->define_fields = kretprobe_event_define_fields; } else { tp->event.trace = print_kprobe_event; call->raw_init = probe_event_raw_init; - call->show_format = kprobe_event_show_format; call->define_fields = kprobe_event_define_fields; } if (set_print_fmt(tp) < 0) diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index a78e863..49cea70 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -143,54 +143,6 @@ extern char *__bad_type_size(void); #type, #name, offsetof(typeof(trace), name), \ sizeof(trace.name), is_signed_type(type) -int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s) -{ - int i; - int ret; - struct syscall_metadata *entry = call->data; - struct syscall_trace_enter trace; - int offset = offsetof(struct syscall_trace_enter, args); - - ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n", - SYSCALL_FIELD(int, nr)); - if (!ret) - return 0; - - for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, "\tfield:%s %s;", entry->types[i], - entry->args[i]); - if (!ret) - return 0; - ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;" - "\tsigned:%u;\n", offset, - sizeof(unsigned long), - is_signed_type(unsigned long)); - if (!ret) - return 0; - offset += sizeof(unsigned long); - } - - trace_seq_puts(s, "\nprint fmt: \""); - for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, "%s: 0x%%0%zulx%s", entry->args[i], - sizeof(unsigned long), - i == entry->nb_args - 1 ? "" : ", "); - if (!ret) - return 0; - } - trace_seq_putc(s, '"'); - - for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, ", ((unsigned long)(REC->%s))", - entry->args[i]); - if (!ret) - return 0; - } - - return trace_seq_putc(s, '\n'); -} - static int __set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len) { @@ -252,24 +204,6 @@ static void free_syscall_print_fmt(struct ftrace_event_call *call) kfree(call->print_fmt); } -int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s) -{ - int ret; - struct syscall_trace_exit trace; - - ret = trace_seq_printf(s, - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n", - SYSCALL_FIELD(int, nr), - SYSCALL_FIELD(long, ret)); - if (!ret) - return 0; - - return trace_seq_printf(s, "\nprint fmt: \"0x%%lx\", REC->ret\n"); -} - int syscall_enter_define_fields(struct ftrace_event_call *call) { struct syscall_trace_enter trace; -- cgit v0.10.2 From 7e53bd42d14c75192b99674c40fcc359392da59d Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 6 Jan 2010 20:08:50 +0800 Subject: tracing: Consolidate protection of reader access to the ring buffer At the beginning, access to the ring buffer was fully serialized by trace_types_lock. Patch d7350c3f4569 gives more freedom to readers, and patch b04cc6b1f6 adds code to protect trace_pipe and cpu#/trace_pipe. But actually it is not enough, ring buffer readers are not always read-only, they may consume data. This patch makes accesses to trace, trace_pipe, trace_pipe_raw cpu#/trace, cpu#/trace_pipe and cpu#/trace_pipe_raw serialized. And removes tracing_reader_cpumask which is used to protect trace_pipe. Details: Ring buffer serializes readers, but it is low level protection. The validity of the events (which returns by ring_buffer_peek() ..etc) are not protected by ring buffer. The content of events may become garbage if we allow another process to consume these events concurrently: A) the page of the consumed events may become a normal page (not reader page) in ring buffer, and this page will be rewritten by the events producer. B) The page of the consumed events may become a page for splice_read, and this page will be returned to system. This patch adds trace_access_lock() and trace_access_unlock() primitives. These primitives allow multi process access to different cpu ring buffers concurrently. These primitives don't distinguish read-only and read-consume access. Multi read-only access is also serialized. And we don't use these primitives when we open files, we only use them when we read files. Signed-off-by: Lai Jiangshan LKML-Reference: <4B447D52.1050602@cn.fujitsu.com> Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 0df1b0f..abdd333 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -32,6 +32,7 @@ #include #include #include +#include #include #include #include @@ -102,9 +103,6 @@ static inline void ftrace_enable_cpu(void) static cpumask_var_t __read_mostly tracing_buffer_mask; -/* Define which cpu buffers are currently read in trace_pipe */ -static cpumask_var_t tracing_reader_cpumask; - #define for_each_tracing_cpu(cpu) \ for_each_cpu(cpu, tracing_buffer_mask) @@ -243,12 +241,91 @@ static struct tracer *current_trace __read_mostly; /* * trace_types_lock is used to protect the trace_types list. - * This lock is also used to keep user access serialized. - * Accesses from userspace will grab this lock while userspace - * activities happen inside the kernel. */ static DEFINE_MUTEX(trace_types_lock); +/* + * serialize the access of the ring buffer + * + * ring buffer serializes readers, but it is low level protection. + * The validity of the events (which returns by ring_buffer_peek() ..etc) + * are not protected by ring buffer. + * + * The content of events may become garbage if we allow other process consumes + * these events concurrently: + * A) the page of the consumed events may become a normal page + * (not reader page) in ring buffer, and this page will be rewrited + * by events producer. + * B) The page of the consumed events may become a page for splice_read, + * and this page will be returned to system. + * + * These primitives allow multi process access to different cpu ring buffer + * concurrently. + * + * These primitives don't distinguish read-only and read-consume access. + * Multi read-only access are also serialized. + */ + +#ifdef CONFIG_SMP +static DECLARE_RWSEM(all_cpu_access_lock); +static DEFINE_PER_CPU(struct mutex, cpu_access_lock); + +static inline void trace_access_lock(int cpu) +{ + if (cpu == TRACE_PIPE_ALL_CPU) { + /* gain it for accessing the whole ring buffer. */ + down_write(&all_cpu_access_lock); + } else { + /* gain it for accessing a cpu ring buffer. */ + + /* Firstly block other trace_access_lock(TRACE_PIPE_ALL_CPU). */ + down_read(&all_cpu_access_lock); + + /* Secondly block other access to this @cpu ring buffer. */ + mutex_lock(&per_cpu(cpu_access_lock, cpu)); + } +} + +static inline void trace_access_unlock(int cpu) +{ + if (cpu == TRACE_PIPE_ALL_CPU) { + up_write(&all_cpu_access_lock); + } else { + mutex_unlock(&per_cpu(cpu_access_lock, cpu)); + up_read(&all_cpu_access_lock); + } +} + +static inline void trace_access_lock_init(void) +{ + int cpu; + + for_each_possible_cpu(cpu) + mutex_init(&per_cpu(cpu_access_lock, cpu)); +} + +#else + +static DEFINE_MUTEX(access_lock); + +static inline void trace_access_lock(int cpu) +{ + (void)cpu; + mutex_lock(&access_lock); +} + +static inline void trace_access_unlock(int cpu) +{ + (void)cpu; + mutex_unlock(&access_lock); +} + +static inline void trace_access_lock_init(void) +{ +} + +#endif + /* trace_wait is a waitqueue for tasks blocked on trace_poll */ static DECLARE_WAIT_QUEUE_HEAD(trace_wait); @@ -1580,12 +1657,6 @@ static void tracing_iter_reset(struct trace_iterator *iter, int cpu) } /* - * No necessary locking here. The worst thing which can - * happen is loosing events consumed at the same time - * by a trace_pipe reader. - * Other than that, we don't risk to crash the ring buffer - * because it serializes the readers. - * * The current tracer is copied to avoid a global locking * all around. */ @@ -1640,12 +1711,16 @@ static void *s_start(struct seq_file *m, loff_t *pos) } trace_event_read_lock(); + trace_access_lock(cpu_file); return p; } static void s_stop(struct seq_file *m, void *p) { + struct trace_iterator *iter = m->private; + atomic_dec(&trace_record_cmdline_disabled); + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); } @@ -2836,22 +2911,6 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) mutex_lock(&trace_types_lock); - /* We only allow one reader per cpu */ - if (cpu_file == TRACE_PIPE_ALL_CPU) { - if (!cpumask_empty(tracing_reader_cpumask)) { - ret = -EBUSY; - goto out; - } - cpumask_setall(tracing_reader_cpumask); - } else { - if (!cpumask_test_cpu(cpu_file, tracing_reader_cpumask)) - cpumask_set_cpu(cpu_file, tracing_reader_cpumask); - else { - ret = -EBUSY; - goto out; - } - } - /* create a buffer to store the information to pass to userspace */ iter = kzalloc(sizeof(*iter), GFP_KERNEL); if (!iter) { @@ -2907,12 +2966,6 @@ static int tracing_release_pipe(struct inode *inode, struct file *file) mutex_lock(&trace_types_lock); - if (iter->cpu_file == TRACE_PIPE_ALL_CPU) - cpumask_clear(tracing_reader_cpumask); - else - cpumask_clear_cpu(iter->cpu_file, tracing_reader_cpumask); - - if (iter->trace->pipe_close) iter->trace->pipe_close(iter); @@ -3074,6 +3127,7 @@ waitagain: iter->pos = -1; trace_event_read_lock(); + trace_access_lock(iter->cpu_file); while (find_next_entry_inc(iter) != NULL) { enum print_line_t ret; int len = iter->seq.len; @@ -3090,6 +3144,7 @@ waitagain: if (iter->seq.len >= cnt) break; } + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); /* Now copy what we have to the user */ @@ -3215,6 +3270,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, } trace_event_read_lock(); + trace_access_lock(iter->cpu_file); /* Fill as many pages as possible. */ for (i = 0, rem = len; i < PIPE_BUFFERS && rem; i++) { @@ -3238,6 +3294,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, trace_seq_init(&iter->seq); } + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); mutex_unlock(&iter->mutex); @@ -3539,10 +3596,12 @@ tracing_buffers_read(struct file *filp, char __user *ubuf, info->read = 0; + trace_access_lock(info->cpu); ret = ring_buffer_read_page(info->tr->buffer, &info->spare, count, info->cpu, 0); + trace_access_unlock(info->cpu); if (ret < 0) return 0; @@ -3670,6 +3729,7 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos, len &= PAGE_MASK; } + trace_access_lock(info->cpu); entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu); for (i = 0; i < PIPE_BUFFERS && len && entries; i++, len -= PAGE_SIZE) { @@ -3717,6 +3777,7 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos, entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu); } + trace_access_unlock(info->cpu); spd.nr_pages = i; /* did we read anything? */ @@ -4153,6 +4214,8 @@ static __init int tracer_init_debugfs(void) struct dentry *d_tracer; int cpu; + trace_access_lock_init(); + d_tracer = tracing_init_dentry(); trace_create_file("tracing_enabled", 0644, d_tracer, @@ -4387,9 +4450,6 @@ __init static int tracer_alloc_buffers(void) if (!alloc_cpumask_var(&tracing_cpumask, GFP_KERNEL)) goto out_free_buffer_mask; - if (!zalloc_cpumask_var(&tracing_reader_cpumask, GFP_KERNEL)) - goto out_free_tracing_cpumask; - /* To save memory, keep the ring buffer size to its minimum */ if (ring_buffer_expanded) ring_buf_size = trace_buf_size; @@ -4447,8 +4507,6 @@ __init static int tracer_alloc_buffers(void) return 0; out_free_cpumask: - free_cpumask_var(tracing_reader_cpumask); -out_free_tracing_cpumask: free_cpumask_var(tracing_cpumask); out_free_buffer_mask: free_cpumask_var(tracing_buffer_mask); -- cgit v0.10.2 From dc4f8845ee2ca39fe054a2d911729ffd269b4b66 Mon Sep 17 00:00:00 2001 From: Wolfram Sang Date: Tue, 5 Jan 2010 19:27:51 +0100 Subject: tracing: optimize recordmcount.pl for offsets-handling - move check for open file in front of the writing loop - use perl-constructs to access the array Signed-off-by: Wolfram Sang LKML-Reference: <1262716072-14414-2-git-send-email-w.sang@pengutronix.de> Signed-off-by: Steven Rostedt diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl index 92f09fe..5de12c7 100755 --- a/scripts/recordmcount.pl +++ b/scripts/recordmcount.pl @@ -432,14 +432,14 @@ sub update_funcs # Loop through all the mcount caller offsets and print a reference # to the caller based from the ref_func. - for (my $i=0; $i <= $#offsets; $i++) { - if (!$opened) { - open(FILE, ">$mcount_s") || die "can't create $mcount_s\n"; - $opened = 1; - print FILE "\t.section $mcount_section,\"a\",$section_type\n"; - print FILE "\t.align $alignment\n" if (defined($alignment)); - } - printf FILE "\t%s %s + %d\n", $type, $ref_func, $offsets[$i] - $offset; + if (!$opened) { + open(FILE, ">$mcount_s") || die "can't create $mcount_s\n"; + $opened = 1; + print FILE "\t.section $mcount_section,\"a\",$section_type\n"; + print FILE "\t.align $alignment\n" if (defined($alignment)); + } + foreach my $cur_offset (@offsets) { + printf FILE "\t%s %s + %d\n", $type, $ref_func, $cur_offset - $offset; } } @@ -514,7 +514,7 @@ while () { } # is this a call site to mcount? If so, record it to print later if ($text_found && /$mcount_regex/) { - $offsets[$#offsets + 1] = hex $1; + push(@offsets, hex $1); } } -- cgit v0.10.2 From 17263baf958b7ab1d8c60445f412a1080362c88c Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 6 Jan 2010 16:43:08 -0500 Subject: kconfig: Create include/generated for localmodconfig If someone downloads a brand new kernel and runs localmodconfig or localyesconfig, the ending result will report: *** Error during update of the kernel configuration. This is because localmodconfig and localyesconfig must create the include/generated directory to place the autoconf.h file. Signed-off-by: Steven Rostedt diff --git a/scripts/kconfig/Makefile b/scripts/kconfig/Makefile index 999e8a7..006c96f 100644 --- a/scripts/kconfig/Makefile +++ b/scripts/kconfig/Makefile @@ -31,6 +31,7 @@ silentoldconfig: $(obj)/conf $< -s $(Kconfig) localmodconfig: $(obj)/streamline_config.pl $(obj)/conf + $(Q)mkdir -p include/generated $(Q)perl $< $(srctree) $(Kconfig) > .tmp.config $(Q)if [ -f .config ]; then \ cmp -s .tmp.config .config || \ @@ -45,6 +46,7 @@ localmodconfig: $(obj)/streamline_config.pl $(obj)/conf $(Q)rm -f .tmp.config localyesconfig: $(obj)/streamline_config.pl $(obj)/conf + $(Q)mkdir -p include/generated $(Q)perl $< $(srctree) $(Kconfig) > .tmp.config $(Q)sed -i s/=m/=y/ .tmp.config $(Q)if [ -f .config ]; then \ -- cgit v0.10.2 From 13d7e9385644d376a0816ad663ba3dfc45359f2f Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 6 Jan 2010 17:56:12 -0500 Subject: kconfig: Check for if conditions in Kconfig for localmodconfig The streamline_config.pl misses the if conditions for checking dependencies. For Kconfigs with the following construct: if MEDIA_SUPPORT config VIDEO_DEV [...] If VIDEO_DEV was enabled, the script will miss the fact that MEDIA_SUPPORT is also needed. This patch changes streamline_config.pl to include if conditions into the dependencies of configs. Reported-by: Anton Blanchard Tested-by: Anton Blanchard Signed-off-by: Steven Rostedt diff --git a/scripts/kconfig/streamline_config.pl b/scripts/kconfig/streamline_config.pl index 0d80082..9e66fa8 100644 --- a/scripts/kconfig/streamline_config.pl +++ b/scripts/kconfig/streamline_config.pl @@ -121,6 +121,8 @@ my %prompts; my %objects; my $var; my $cont = 0; +my $iflevel = 0; +my @ifdeps; # prevent recursion my %read_kconfigs; @@ -146,6 +148,15 @@ sub read_kconfig { $state = "NEW"; $config = $1; + for (my $i = 0; $i < $iflevel; $i++) { + if ($i) { + $depends{$config} .= " " . $ifdeps[$i]; + } else { + $depends{$config} = $ifdeps[$i]; + } + $state = "DEP"; + } + # collect the depends for the config } elsif ($state eq "NEW" && /^\s*depends\s+on\s+(.*)$/) { $state = "DEP"; @@ -166,6 +177,21 @@ sub read_kconfig { # note if the config has a prompt $prompt{$config} = 1; + # Check for if statements + } elsif (/^if\s+(.*\S)\s*$/) { + my $deps = $1; + # remove beginning and ending non text + $deps =~ s/^[^a-zA-Z0-9_]*//; + $deps =~ s/[^a-zA-Z0-9_]*$//; + + my @deps = split /[^a-zA-Z0-9_]+/, $deps; + + $ifdeps[$iflevel++] = join ':', @deps; + + } elsif (/^endif/) { + + $iflevel-- if ($iflevel); + # stop on "help" } elsif (/^\s*help\s*$/) { $state = "NONE"; -- cgit v0.10.2 From dfaa9e2c5707b2c217c0121aac796e0fa3051482 Mon Sep 17 00:00:00 2001 From: Wolfram Sang Date: Tue, 5 Jan 2010 21:41:22 +0100 Subject: tracing: Use appropriate perl constructs in recordmcount.pl Modified recordmcount.pl to use perl constructs that are still understandable by C hackers that are not perl programmers. Signed-off-by: Wolfram Sang LKML-Reference: <1262724082-9517-1-git-send-email-w.sang@pengutronix.de> Signed-off-by: Steven Rostedt diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl index 5de12c7..545fe71 100755 --- a/scripts/recordmcount.pl +++ b/scripts/recordmcount.pl @@ -136,13 +136,14 @@ my %text_sections = ( ".text.unlikely" => 1, ); -$objdump = "objdump" if ((length $objdump) == 0); -$objcopy = "objcopy" if ((length $objcopy) == 0); -$cc = "gcc" if ((length $cc) == 0); -$ld = "ld" if ((length $ld) == 0); -$nm = "nm" if ((length $nm) == 0); -$rm = "rm" if ((length $rm) == 0); -$mv = "mv" if ((length $mv) == 0); +# Note: we are nice to C-programmers here, thus we skip the '||='-idiom. +$objdump = 'objdump' if (!$objdump); +$objcopy = 'objcopy' if (!$objcopy); +$cc = 'gcc' if (!$cc); +$ld = 'ld' if (!$ld); +$nm = 'nm' if (!$nm); +$rm = 'rm' if (!$rm); +$mv = 'mv' if (!$mv); #print STDERR "running: $P '$arch' '$objdump' '$objcopy' '$cc' '$ld' " . # "'$nm' '$rm' '$mv' '$inputfile'\n"; @@ -194,12 +195,8 @@ sub check_objcopy } } -if ($arch eq "x86") { - if ($bits == 64) { - $arch = "x86_64"; - } else { - $arch = "i386"; - } +if ($arch eq 'x86') { + $arch = ($bits == 64) ? 'x86_64' : 'i386'; } # @@ -476,11 +473,7 @@ while () { $read_headers = 0; # Only record text sections that we know are safe - if (defined($text_sections{$1})) { - $read_function = 1; - } else { - $read_function = 0; - } + $read_function = defined($text_sections{$1}); # print out any recorded offsets update_funcs(); -- cgit v0.10.2 From d931369b74b3d6f2044f595af6f3dd074f65d9cf Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 6 Jan 2010 17:27:11 -0500 Subject: tracing: Add stack dump to trace_printk if stacktrace option is set If the ftrace stacktrace option is set, then add the stack dumps to trace_printk. Requested-by: Peter Zijlstra Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index abdd333..5314c90 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1392,8 +1392,10 @@ int trace_vbprintk(unsigned long ip, const char *fmt, va_list args) entry->fmt = fmt; memcpy(entry->buf, trace_buf, sizeof(u32) * len); - if (!filter_check_discard(call, entry, buffer, event)) + if (!filter_check_discard(call, entry, buffer, event)) { ring_buffer_unlock_commit(buffer, event); + ftrace_trace_stack(buffer, flags, 6, pc); + } out_unlock: arch_spin_unlock(&trace_buf_lock); @@ -1466,8 +1468,10 @@ int trace_array_vprintk(struct trace_array *tr, memcpy(&entry->buf, trace_buf, len); entry->buf[len] = '\0'; - if (!filter_check_discard(call, entry, buffer, event)) + if (!filter_check_discard(call, entry, buffer, event)) { ring_buffer_unlock_commit(buffer, event); + ftrace_trace_stack(buffer, irq_flags, 6, pc); + } out_unlock: arch_spin_unlock(&trace_buf_lock); -- cgit v0.10.2 From 5068c7a883d1694f0fb3631b664827644153cd08 Mon Sep 17 00:00:00 2001 From: Steffen Klassert Date: Thu, 7 Jan 2010 15:57:19 +1100 Subject: crypto: pcrypt - Add pcrypt crypto parallelization wrapper This patch adds a parallel crypto template that takes a crypto algorithm and converts it to process the crypto transforms in parallel. For the moment only aead algorithms are supported. Signed-off-by: Steffen Klassert Signed-off-by: Herbert Xu diff --git a/crypto/Kconfig b/crypto/Kconfig index 81c185a..6a2e295 100644 --- a/crypto/Kconfig +++ b/crypto/Kconfig @@ -114,6 +114,16 @@ config CRYPTO_NULL help These are 'Null' algorithms, used by IPsec, which do nothing. +config CRYPTO_PCRYPT + tristate "Parallel crypto engine (EXPERIMENTAL)" + depends on SMP && EXPERIMENTAL + select PADATA + select CRYPTO_MANAGER + select CRYPTO_AEAD + help + This converts an arbitrary crypto algorithm into a parallel + algorithm that executes in kernel threads. + config CRYPTO_WORKQUEUE tristate diff --git a/crypto/Makefile b/crypto/Makefile index 9e8f619..d7e6441 100644 --- a/crypto/Makefile +++ b/crypto/Makefile @@ -56,6 +56,7 @@ obj-$(CONFIG_CRYPTO_XTS) += xts.o obj-$(CONFIG_CRYPTO_CTR) += ctr.o obj-$(CONFIG_CRYPTO_GCM) += gcm.o obj-$(CONFIG_CRYPTO_CCM) += ccm.o +obj-$(CONFIG_CRYPTO_PCRYPT) += pcrypt.o obj-$(CONFIG_CRYPTO_CRYPTD) += cryptd.o obj-$(CONFIG_CRYPTO_DES) += des_generic.o obj-$(CONFIG_CRYPTO_FCRYPT) += fcrypt.o diff --git a/crypto/pcrypt.c b/crypto/pcrypt.c new file mode 100644 index 0000000..b9527d0 --- /dev/null +++ b/crypto/pcrypt.c @@ -0,0 +1,445 @@ +/* + * pcrypt - Parallel crypto wrapper. + * + * Copyright (C) 2009 secunet Security Networks AG + * Copyright (C) 2009 Steffen Klassert + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include +#include +#include +#include +#include +#include +#include + +static struct padata_instance *pcrypt_enc_padata; +static struct padata_instance *pcrypt_dec_padata; +static struct workqueue_struct *encwq; +static struct workqueue_struct *decwq; + +struct pcrypt_instance_ctx { + struct crypto_spawn spawn; + unsigned int tfm_count; +}; + +struct pcrypt_aead_ctx { + struct crypto_aead *child; + unsigned int cb_cpu; +}; + +static int pcrypt_do_parallel(struct padata_priv *padata, unsigned int *cb_cpu, + struct padata_instance *pinst) +{ + unsigned int cpu_index, cpu, i; + + cpu = *cb_cpu; + + if (cpumask_test_cpu(cpu, cpu_active_mask)) + goto out; + + cpu_index = cpu % cpumask_weight(cpu_active_mask); + + cpu = cpumask_first(cpu_active_mask); + for (i = 0; i < cpu_index; i++) + cpu = cpumask_next(cpu, cpu_active_mask); + + *cb_cpu = cpu; + +out: + return padata_do_parallel(pinst, padata, cpu); +} + +static int pcrypt_aead_setkey(struct crypto_aead *parent, + const u8 *key, unsigned int keylen) +{ + struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(parent); + + return crypto_aead_setkey(ctx->child, key, keylen); +} + +static int pcrypt_aead_setauthsize(struct crypto_aead *parent, + unsigned int authsize) +{ + struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(parent); + + return crypto_aead_setauthsize(ctx->child, authsize); +} + +static void pcrypt_aead_serial(struct padata_priv *padata) +{ + struct pcrypt_request *preq = pcrypt_padata_request(padata); + struct aead_request *req = pcrypt_request_ctx(preq); + + aead_request_complete(req->base.data, padata->info); +} + +static void pcrypt_aead_giv_serial(struct padata_priv *padata) +{ + struct pcrypt_request *preq = pcrypt_padata_request(padata); + struct aead_givcrypt_request *req = pcrypt_request_ctx(preq); + + aead_request_complete(req->areq.base.data, padata->info); +} + +static void pcrypt_aead_done(struct crypto_async_request *areq, int err) +{ + struct aead_request *req = areq->data; + struct pcrypt_request *preq = aead_request_ctx(req); + struct padata_priv *padata = pcrypt_request_padata(preq); + + padata->info = err; + req->base.flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; + + padata_do_serial(padata); +} + +static void pcrypt_aead_enc(struct padata_priv *padata) +{ + struct pcrypt_request *preq = pcrypt_padata_request(padata); + struct aead_request *req = pcrypt_request_ctx(preq); + + padata->info = crypto_aead_encrypt(req); + + if (padata->info) + return; + + padata_do_serial(padata); +} + +static int pcrypt_aead_encrypt(struct aead_request *req) +{ + int err; + struct pcrypt_request *preq = aead_request_ctx(req); + struct aead_request *creq = pcrypt_request_ctx(preq); + struct padata_priv *padata = pcrypt_request_padata(preq); + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(aead); + u32 flags = aead_request_flags(req); + + memset(padata, 0, sizeof(struct padata_priv)); + + padata->parallel = pcrypt_aead_enc; + padata->serial = pcrypt_aead_serial; + + aead_request_set_tfm(creq, ctx->child); + aead_request_set_callback(creq, flags & ~CRYPTO_TFM_REQ_MAY_SLEEP, + pcrypt_aead_done, req); + aead_request_set_crypt(creq, req->src, req->dst, + req->cryptlen, req->iv); + aead_request_set_assoc(creq, req->assoc, req->assoclen); + + err = pcrypt_do_parallel(padata, &ctx->cb_cpu, pcrypt_enc_padata); + if (err) + return err; + else + err = crypto_aead_encrypt(creq); + + return err; +} + +static void pcrypt_aead_dec(struct padata_priv *padata) +{ + struct pcrypt_request *preq = pcrypt_padata_request(padata); + struct aead_request *req = pcrypt_request_ctx(preq); + + padata->info = crypto_aead_decrypt(req); + + if (padata->info) + return; + + padata_do_serial(padata); +} + +static int pcrypt_aead_decrypt(struct aead_request *req) +{ + int err; + struct pcrypt_request *preq = aead_request_ctx(req); + struct aead_request *creq = pcrypt_request_ctx(preq); + struct padata_priv *padata = pcrypt_request_padata(preq); + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(aead); + u32 flags = aead_request_flags(req); + + memset(padata, 0, sizeof(struct padata_priv)); + + padata->parallel = pcrypt_aead_dec; + padata->serial = pcrypt_aead_serial; + + aead_request_set_tfm(creq, ctx->child); + aead_request_set_callback(creq, flags & ~CRYPTO_TFM_REQ_MAY_SLEEP, + pcrypt_aead_done, req); + aead_request_set_crypt(creq, req->src, req->dst, + req->cryptlen, req->iv); + aead_request_set_assoc(creq, req->assoc, req->assoclen); + + err = pcrypt_do_parallel(padata, &ctx->cb_cpu, pcrypt_dec_padata); + if (err) + return err; + else + err = crypto_aead_decrypt(creq); + + return err; +} + +static void pcrypt_aead_givenc(struct padata_priv *padata) +{ + struct pcrypt_request *preq = pcrypt_padata_request(padata); + struct aead_givcrypt_request *req = pcrypt_request_ctx(preq); + + padata->info = crypto_aead_givencrypt(req); + + if (padata->info) + return; + + padata_do_serial(padata); +} + +static int pcrypt_aead_givencrypt(struct aead_givcrypt_request *req) +{ + int err; + struct aead_request *areq = &req->areq; + struct pcrypt_request *preq = aead_request_ctx(areq); + struct aead_givcrypt_request *creq = pcrypt_request_ctx(preq); + struct padata_priv *padata = pcrypt_request_padata(preq); + struct crypto_aead *aead = aead_givcrypt_reqtfm(req); + struct pcrypt_aead_ctx *ctx = crypto_aead_ctx(aead); + u32 flags = aead_request_flags(areq); + + memset(padata, 0, sizeof(struct padata_priv)); + + padata->parallel = pcrypt_aead_givenc; + padata->serial = pcrypt_aead_giv_serial; + + aead_givcrypt_set_tfm(creq, ctx->child); + aead_givcrypt_set_callback(creq, flags & ~CRYPTO_TFM_REQ_MAY_SLEEP, + pcrypt_aead_done, areq); + aead_givcrypt_set_crypt(creq, areq->src, areq->dst, + areq->cryptlen, areq->iv); + aead_givcrypt_set_assoc(creq, areq->assoc, areq->assoclen); + aead_givcrypt_set_giv(creq, req->giv, req->seq); + + err = pcrypt_do_parallel(padata, &ctx->cb_cpu, pcrypt_enc_padata); + if (err) + return err; + else + err = crypto_aead_givencrypt(creq); + + return err; +} + +static int pcrypt_aead_init_tfm(struct crypto_tfm *tfm) +{ + int cpu, cpu_index; + struct crypto_instance *inst = crypto_tfm_alg_instance(tfm); + struct pcrypt_instance_ctx *ictx = crypto_instance_ctx(inst); + struct pcrypt_aead_ctx *ctx = crypto_tfm_ctx(tfm); + struct crypto_aead *cipher; + + ictx->tfm_count++; + + cpu_index = ictx->tfm_count % cpumask_weight(cpu_active_mask); + + ctx->cb_cpu = cpumask_first(cpu_active_mask); + for (cpu = 0; cpu < cpu_index; cpu++) + ctx->cb_cpu = cpumask_next(ctx->cb_cpu, cpu_active_mask); + + cipher = crypto_spawn_aead(crypto_instance_ctx(inst)); + + if (IS_ERR(cipher)) + return PTR_ERR(cipher); + + ctx->child = cipher; + tfm->crt_aead.reqsize = sizeof(struct pcrypt_request) + + sizeof(struct aead_givcrypt_request) + + crypto_aead_reqsize(cipher); + + return 0; +} + +static void pcrypt_aead_exit_tfm(struct crypto_tfm *tfm) +{ + struct pcrypt_aead_ctx *ctx = crypto_tfm_ctx(tfm); + + crypto_free_aead(ctx->child); +} + +static struct crypto_instance *pcrypt_alloc_instance(struct crypto_alg *alg) +{ + struct crypto_instance *inst; + struct pcrypt_instance_ctx *ctx; + int err; + + inst = kzalloc(sizeof(*inst) + sizeof(*ctx), GFP_KERNEL); + if (!inst) { + inst = ERR_PTR(-ENOMEM); + goto out; + } + + err = -ENAMETOOLONG; + if (snprintf(inst->alg.cra_driver_name, CRYPTO_MAX_ALG_NAME, + "pcrypt(%s)", alg->cra_driver_name) >= CRYPTO_MAX_ALG_NAME) + goto out_free_inst; + + memcpy(inst->alg.cra_name, alg->cra_name, CRYPTO_MAX_ALG_NAME); + + ctx = crypto_instance_ctx(inst); + err = crypto_init_spawn(&ctx->spawn, alg, inst, + CRYPTO_ALG_TYPE_MASK); + if (err) + goto out_free_inst; + + inst->alg.cra_priority = alg->cra_priority + 100; + inst->alg.cra_blocksize = alg->cra_blocksize; + inst->alg.cra_alignmask = alg->cra_alignmask; + +out: + return inst; + +out_free_inst: + kfree(inst); + inst = ERR_PTR(err); + goto out; +} + +static struct crypto_instance *pcrypt_alloc_aead(struct rtattr **tb) +{ + struct crypto_instance *inst; + struct crypto_alg *alg; + struct crypto_attr_type *algt; + + algt = crypto_get_attr_type(tb); + + alg = crypto_get_attr_alg(tb, algt->type, + (algt->mask & CRYPTO_ALG_TYPE_MASK)); + if (IS_ERR(alg)) + return ERR_CAST(alg); + + inst = pcrypt_alloc_instance(alg); + if (IS_ERR(inst)) + goto out_put_alg; + + inst->alg.cra_flags = CRYPTO_ALG_TYPE_AEAD | CRYPTO_ALG_ASYNC; + inst->alg.cra_type = &crypto_aead_type; + + inst->alg.cra_aead.ivsize = alg->cra_aead.ivsize; + inst->alg.cra_aead.geniv = alg->cra_aead.geniv; + inst->alg.cra_aead.maxauthsize = alg->cra_aead.maxauthsize; + + inst->alg.cra_ctxsize = sizeof(struct pcrypt_aead_ctx); + + inst->alg.cra_init = pcrypt_aead_init_tfm; + inst->alg.cra_exit = pcrypt_aead_exit_tfm; + + inst->alg.cra_aead.setkey = pcrypt_aead_setkey; + inst->alg.cra_aead.setauthsize = pcrypt_aead_setauthsize; + inst->alg.cra_aead.encrypt = pcrypt_aead_encrypt; + inst->alg.cra_aead.decrypt = pcrypt_aead_decrypt; + inst->alg.cra_aead.givencrypt = pcrypt_aead_givencrypt; + +out_put_alg: + crypto_mod_put(alg); + return inst; +} + +static struct crypto_instance *pcrypt_alloc(struct rtattr **tb) +{ + struct crypto_attr_type *algt; + + algt = crypto_get_attr_type(tb); + if (IS_ERR(algt)) + return ERR_CAST(algt); + + switch (algt->type & algt->mask & CRYPTO_ALG_TYPE_MASK) { + case CRYPTO_ALG_TYPE_AEAD: + return pcrypt_alloc_aead(tb); + } + + return ERR_PTR(-EINVAL); +} + +static void pcrypt_free(struct crypto_instance *inst) +{ + struct pcrypt_instance_ctx *ctx = crypto_instance_ctx(inst); + + crypto_drop_spawn(&ctx->spawn); + kfree(inst); +} + +static struct crypto_template pcrypt_tmpl = { + .name = "pcrypt", + .alloc = pcrypt_alloc, + .free = pcrypt_free, + .module = THIS_MODULE, +}; + +static int __init pcrypt_init(void) +{ + encwq = create_workqueue("pencrypt"); + if (!encwq) + goto err; + + decwq = create_workqueue("pdecrypt"); + if (!decwq) + goto err_destroy_encwq; + + + pcrypt_enc_padata = padata_alloc(cpu_possible_mask, encwq); + if (!pcrypt_enc_padata) + goto err_destroy_decwq; + + pcrypt_dec_padata = padata_alloc(cpu_possible_mask, decwq); + if (!pcrypt_dec_padata) + goto err_free_padata; + + padata_start(pcrypt_enc_padata); + padata_start(pcrypt_dec_padata); + + return crypto_register_template(&pcrypt_tmpl); + +err_free_padata: + padata_free(pcrypt_enc_padata); + +err_destroy_decwq: + destroy_workqueue(decwq); + +err_destroy_encwq: + destroy_workqueue(encwq); + +err: + return -ENOMEM; +} + +static void __exit pcrypt_exit(void) +{ + padata_stop(pcrypt_enc_padata); + padata_stop(pcrypt_dec_padata); + + destroy_workqueue(encwq); + destroy_workqueue(decwq); + + padata_free(pcrypt_enc_padata); + padata_free(pcrypt_dec_padata); + + crypto_unregister_template(&pcrypt_tmpl); +} + +module_init(pcrypt_init); +module_exit(pcrypt_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Steffen Klassert "); +MODULE_DESCRIPTION("Parallel crypto wrapper"); diff --git a/include/crypto/pcrypt.h b/include/crypto/pcrypt.h new file mode 100644 index 0000000..d7d8bd8 --- /dev/null +++ b/include/crypto/pcrypt.h @@ -0,0 +1,51 @@ +/* + * pcrypt - Parallel crypto engine. + * + * Copyright (C) 2009 secunet Security Networks AG + * Copyright (C) 2009 Steffen Klassert + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifndef _CRYPTO_PCRYPT_H +#define _CRYPTO_PCRYPT_H + +#include +#include +#include + +struct pcrypt_request { + struct padata_priv padata; + void *data; + void *__ctx[] CRYPTO_MINALIGN_ATTR; +}; + +static inline void *pcrypt_request_ctx(struct pcrypt_request *req) +{ + return req->__ctx; +} + +static inline +struct padata_priv *pcrypt_request_padata(struct pcrypt_request *req) +{ + return &req->padata; +} + +static inline +struct pcrypt_request *pcrypt_padata_request(struct padata_priv *padata) +{ + return container_of(padata, struct pcrypt_request, padata); +} + +#endif -- cgit v0.10.2 From 1a3b1d89eded68d64e5ea409ad37827310059441 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Thu, 7 Jan 2010 11:53:33 -0500 Subject: x86: Split atomic64_t functions into seperate headers Split atomic64_t functions out into separate headers, since they will not be practical to merge between 32 and 64 bits. Signed-off-by: Brian Gerst LKML-Reference: <1262883215-4034-2-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/atomic64_32.h b/arch/x86/include/asm/atomic64_32.h new file mode 100644 index 0000000..03027bf --- /dev/null +++ b/arch/x86/include/asm/atomic64_32.h @@ -0,0 +1,160 @@ +#ifndef _ASM_X86_ATOMIC64_32_H +#define _ASM_X86_ATOMIC64_32_H + +#include +#include +#include +//#include + +/* An 64bit atomic type */ + +typedef struct { + u64 __aligned(8) counter; +} atomic64_t; + +#define ATOMIC64_INIT(val) { (val) } + +extern u64 atomic64_cmpxchg(atomic64_t *ptr, u64 old_val, u64 new_val); + +/** + * atomic64_xchg - xchg atomic64 variable + * @ptr: pointer to type atomic64_t + * @new_val: value to assign + * + * Atomically xchgs the value of @ptr to @new_val and returns + * the old value. + */ +extern u64 atomic64_xchg(atomic64_t *ptr, u64 new_val); + +/** + * atomic64_set - set atomic64 variable + * @ptr: pointer to type atomic64_t + * @new_val: value to assign + * + * Atomically sets the value of @ptr to @new_val. + */ +extern void atomic64_set(atomic64_t *ptr, u64 new_val); + +/** + * atomic64_read - read atomic64 variable + * @ptr: pointer to type atomic64_t + * + * Atomically reads the value of @ptr and returns it. + */ +static inline u64 atomic64_read(atomic64_t *ptr) +{ + u64 res; + + /* + * Note, we inline this atomic64_t primitive because + * it only clobbers EAX/EDX and leaves the others + * untouched. We also (somewhat subtly) rely on the + * fact that cmpxchg8b returns the current 64-bit value + * of the memory location we are touching: + */ + asm volatile( + "mov %%ebx, %%eax\n\t" + "mov %%ecx, %%edx\n\t" + LOCK_PREFIX "cmpxchg8b %1\n" + : "=&A" (res) + : "m" (*ptr) + ); + + return res; +} + +extern u64 atomic64_read(atomic64_t *ptr); + +/** + * atomic64_add_return - add and return + * @delta: integer value to add + * @ptr: pointer to type atomic64_t + * + * Atomically adds @delta to @ptr and returns @delta + *@ptr + */ +extern u64 atomic64_add_return(u64 delta, atomic64_t *ptr); + +/* + * Other variants with different arithmetic operators: + */ +extern u64 atomic64_sub_return(u64 delta, atomic64_t *ptr); +extern u64 atomic64_inc_return(atomic64_t *ptr); +extern u64 atomic64_dec_return(atomic64_t *ptr); + +/** + * atomic64_add - add integer to atomic64 variable + * @delta: integer value to add + * @ptr: pointer to type atomic64_t + * + * Atomically adds @delta to @ptr. + */ +extern void atomic64_add(u64 delta, atomic64_t *ptr); + +/** + * atomic64_sub - subtract the atomic64 variable + * @delta: integer value to subtract + * @ptr: pointer to type atomic64_t + * + * Atomically subtracts @delta from @ptr. + */ +extern void atomic64_sub(u64 delta, atomic64_t *ptr); + +/** + * atomic64_sub_and_test - subtract value from variable and test result + * @delta: integer value to subtract + * @ptr: pointer to type atomic64_t + * + * Atomically subtracts @delta from @ptr and returns + * true if the result is zero, or false for all + * other cases. + */ +extern int atomic64_sub_and_test(u64 delta, atomic64_t *ptr); + +/** + * atomic64_inc - increment atomic64 variable + * @ptr: pointer to type atomic64_t + * + * Atomically increments @ptr by 1. + */ +extern void atomic64_inc(atomic64_t *ptr); + +/** + * atomic64_dec - decrement atomic64 variable + * @ptr: pointer to type atomic64_t + * + * Atomically decrements @ptr by 1. + */ +extern void atomic64_dec(atomic64_t *ptr); + +/** + * atomic64_dec_and_test - decrement and test + * @ptr: pointer to type atomic64_t + * + * Atomically decrements @ptr by 1 and + * returns true if the result is 0, or false for all other + * cases. + */ +extern int atomic64_dec_and_test(atomic64_t *ptr); + +/** + * atomic64_inc_and_test - increment and test + * @ptr: pointer to type atomic64_t + * + * Atomically increments @ptr by 1 + * and returns true if the result is zero, or false for all + * other cases. + */ +extern int atomic64_inc_and_test(atomic64_t *ptr); + +/** + * atomic64_add_negative - add and test if negative + * @delta: integer value to add + * @ptr: pointer to type atomic64_t + * + * Atomically adds @delta to @ptr and returns true + * if the result is negative, or false when + * result is greater than or equal to zero. + */ +extern int atomic64_add_negative(u64 delta, atomic64_t *ptr); + +#endif /* _ASM_X86_ATOMIC64_32_H */ diff --git a/arch/x86/include/asm/atomic64_64.h b/arch/x86/include/asm/atomic64_64.h new file mode 100644 index 0000000..51c5b40 --- /dev/null +++ b/arch/x86/include/asm/atomic64_64.h @@ -0,0 +1,224 @@ +#ifndef _ASM_X86_ATOMIC64_64_H +#define _ASM_X86_ATOMIC64_64_H + +#include +#include +#include + +/* The 64-bit atomic type */ + +#define ATOMIC64_INIT(i) { (i) } + +/** + * atomic64_read - read atomic64 variable + * @v: pointer of type atomic64_t + * + * Atomically reads the value of @v. + * Doesn't imply a read memory barrier. + */ +static inline long atomic64_read(const atomic64_t *v) +{ + return v->counter; +} + +/** + * atomic64_set - set atomic64 variable + * @v: pointer to type atomic64_t + * @i: required value + * + * Atomically sets the value of @v to @i. + */ +static inline void atomic64_set(atomic64_t *v, long i) +{ + v->counter = i; +} + +/** + * atomic64_add - add integer to atomic64 variable + * @i: integer value to add + * @v: pointer to type atomic64_t + * + * Atomically adds @i to @v. + */ +static inline void atomic64_add(long i, atomic64_t *v) +{ + asm volatile(LOCK_PREFIX "addq %1,%0" + : "=m" (v->counter) + : "er" (i), "m" (v->counter)); +} + +/** + * atomic64_sub - subtract the atomic64 variable + * @i: integer value to subtract + * @v: pointer to type atomic64_t + * + * Atomically subtracts @i from @v. + */ +static inline void atomic64_sub(long i, atomic64_t *v) +{ + asm volatile(LOCK_PREFIX "subq %1,%0" + : "=m" (v->counter) + : "er" (i), "m" (v->counter)); +} + +/** + * atomic64_sub_and_test - subtract value from variable and test result + * @i: integer value to subtract + * @v: pointer to type atomic64_t + * + * Atomically subtracts @i from @v and returns + * true if the result is zero, or false for all + * other cases. + */ +static inline int atomic64_sub_and_test(long i, atomic64_t *v) +{ + unsigned char c; + + asm volatile(LOCK_PREFIX "subq %2,%0; sete %1" + : "=m" (v->counter), "=qm" (c) + : "er" (i), "m" (v->counter) : "memory"); + return c; +} + +/** + * atomic64_inc - increment atomic64 variable + * @v: pointer to type atomic64_t + * + * Atomically increments @v by 1. + */ +static inline void atomic64_inc(atomic64_t *v) +{ + asm volatile(LOCK_PREFIX "incq %0" + : "=m" (v->counter) + : "m" (v->counter)); +} + +/** + * atomic64_dec - decrement atomic64 variable + * @v: pointer to type atomic64_t + * + * Atomically decrements @v by 1. + */ +static inline void atomic64_dec(atomic64_t *v) +{ + asm volatile(LOCK_PREFIX "decq %0" + : "=m" (v->counter) + : "m" (v->counter)); +} + +/** + * atomic64_dec_and_test - decrement and test + * @v: pointer to type atomic64_t + * + * Atomically decrements @v by 1 and + * returns true if the result is 0, or false for all other + * cases. + */ +static inline int atomic64_dec_and_test(atomic64_t *v) +{ + unsigned char c; + + asm volatile(LOCK_PREFIX "decq %0; sete %1" + : "=m" (v->counter), "=qm" (c) + : "m" (v->counter) : "memory"); + return c != 0; +} + +/** + * atomic64_inc_and_test - increment and test + * @v: pointer to type atomic64_t + * + * Atomically increments @v by 1 + * and returns true if the result is zero, or false for all + * other cases. + */ +static inline int atomic64_inc_and_test(atomic64_t *v) +{ + unsigned char c; + + asm volatile(LOCK_PREFIX "incq %0; sete %1" + : "=m" (v->counter), "=qm" (c) + : "m" (v->counter) : "memory"); + return c != 0; +} + +/** + * atomic64_add_negative - add and test if negative + * @i: integer value to add + * @v: pointer to type atomic64_t + * + * Atomically adds @i to @v and returns true + * if the result is negative, or false when + * result is greater than or equal to zero. + */ +static inline int atomic64_add_negative(long i, atomic64_t *v) +{ + unsigned char c; + + asm volatile(LOCK_PREFIX "addq %2,%0; sets %1" + : "=m" (v->counter), "=qm" (c) + : "er" (i), "m" (v->counter) : "memory"); + return c; +} + +/** + * atomic64_add_return - add and return + * @i: integer value to add + * @v: pointer to type atomic64_t + * + * Atomically adds @i to @v and returns @i + @v + */ +static inline long atomic64_add_return(long i, atomic64_t *v) +{ + long __i = i; + asm volatile(LOCK_PREFIX "xaddq %0, %1;" + : "+r" (i), "+m" (v->counter) + : : "memory"); + return i + __i; +} + +static inline long atomic64_sub_return(long i, atomic64_t *v) +{ + return atomic64_add_return(-i, v); +} + +#define atomic64_inc_return(v) (atomic64_add_return(1, (v))) +#define atomic64_dec_return(v) (atomic64_sub_return(1, (v))) + +static inline long atomic64_cmpxchg(atomic64_t *v, long old, long new) +{ + return cmpxchg(&v->counter, old, new); +} + +static inline long atomic64_xchg(atomic64_t *v, long new) +{ + return xchg(&v->counter, new); +} + +/** + * atomic64_add_unless - add unless the number is a given value + * @v: pointer of type atomic64_t + * @a: the amount to add to v... + * @u: ...unless v is equal to u. + * + * Atomically adds @a to @v, so long as it was not @u. + * Returns non-zero if @v was not @u, and zero otherwise. + */ +static inline int atomic64_add_unless(atomic64_t *v, long a, long u) +{ + long c, old; + c = atomic64_read(v); + for (;;) { + if (unlikely(c == (u))) + break; + old = atomic64_cmpxchg((v), c, c + (a)); + if (likely(old == c)) + break; + c = old; + } + return c != (u); +} + +#define atomic64_inc_not_zero(v) atomic64_add_unless((v), 1, 0) + +#endif /* _ASM_X86_ATOMIC64_64_H */ diff --git a/arch/x86/include/asm/atomic_32.h b/arch/x86/include/asm/atomic_32.h index dc5a667..e128ae9 100644 --- a/arch/x86/include/asm/atomic_32.h +++ b/arch/x86/include/asm/atomic_32.h @@ -260,156 +260,6 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u) #define smp_mb__before_atomic_inc() barrier() #define smp_mb__after_atomic_inc() barrier() -/* An 64bit atomic type */ - -typedef struct { - u64 __aligned(8) counter; -} atomic64_t; - -#define ATOMIC64_INIT(val) { (val) } - -extern u64 atomic64_cmpxchg(atomic64_t *ptr, u64 old_val, u64 new_val); - -/** - * atomic64_xchg - xchg atomic64 variable - * @ptr: pointer to type atomic64_t - * @new_val: value to assign - * - * Atomically xchgs the value of @ptr to @new_val and returns - * the old value. - */ -extern u64 atomic64_xchg(atomic64_t *ptr, u64 new_val); - -/** - * atomic64_set - set atomic64 variable - * @ptr: pointer to type atomic64_t - * @new_val: value to assign - * - * Atomically sets the value of @ptr to @new_val. - */ -extern void atomic64_set(atomic64_t *ptr, u64 new_val); - -/** - * atomic64_read - read atomic64 variable - * @ptr: pointer to type atomic64_t - * - * Atomically reads the value of @ptr and returns it. - */ -static inline u64 atomic64_read(atomic64_t *ptr) -{ - u64 res; - - /* - * Note, we inline this atomic64_t primitive because - * it only clobbers EAX/EDX and leaves the others - * untouched. We also (somewhat subtly) rely on the - * fact that cmpxchg8b returns the current 64-bit value - * of the memory location we are touching: - */ - asm volatile( - "mov %%ebx, %%eax\n\t" - "mov %%ecx, %%edx\n\t" - LOCK_PREFIX "cmpxchg8b %1\n" - : "=&A" (res) - : "m" (*ptr) - ); - - return res; -} - -extern u64 atomic64_read(atomic64_t *ptr); - -/** - * atomic64_add_return - add and return - * @delta: integer value to add - * @ptr: pointer to type atomic64_t - * - * Atomically adds @delta to @ptr and returns @delta + *@ptr - */ -extern u64 atomic64_add_return(u64 delta, atomic64_t *ptr); - -/* - * Other variants with different arithmetic operators: - */ -extern u64 atomic64_sub_return(u64 delta, atomic64_t *ptr); -extern u64 atomic64_inc_return(atomic64_t *ptr); -extern u64 atomic64_dec_return(atomic64_t *ptr); - -/** - * atomic64_add - add integer to atomic64 variable - * @delta: integer value to add - * @ptr: pointer to type atomic64_t - * - * Atomically adds @delta to @ptr. - */ -extern void atomic64_add(u64 delta, atomic64_t *ptr); - -/** - * atomic64_sub - subtract the atomic64 variable - * @delta: integer value to subtract - * @ptr: pointer to type atomic64_t - * - * Atomically subtracts @delta from @ptr. - */ -extern void atomic64_sub(u64 delta, atomic64_t *ptr); - -/** - * atomic64_sub_and_test - subtract value from variable and test result - * @delta: integer value to subtract - * @ptr: pointer to type atomic64_t - * - * Atomically subtracts @delta from @ptr and returns - * true if the result is zero, or false for all - * other cases. - */ -extern int atomic64_sub_and_test(u64 delta, atomic64_t *ptr); - -/** - * atomic64_inc - increment atomic64 variable - * @ptr: pointer to type atomic64_t - * - * Atomically increments @ptr by 1. - */ -extern void atomic64_inc(atomic64_t *ptr); - -/** - * atomic64_dec - decrement atomic64 variable - * @ptr: pointer to type atomic64_t - * - * Atomically decrements @ptr by 1. - */ -extern void atomic64_dec(atomic64_t *ptr); - -/** - * atomic64_dec_and_test - decrement and test - * @ptr: pointer to type atomic64_t - * - * Atomically decrements @ptr by 1 and - * returns true if the result is 0, or false for all other - * cases. - */ -extern int atomic64_dec_and_test(atomic64_t *ptr); - -/** - * atomic64_inc_and_test - increment and test - * @ptr: pointer to type atomic64_t - * - * Atomically increments @ptr by 1 - * and returns true if the result is zero, or false for all - * other cases. - */ -extern int atomic64_inc_and_test(atomic64_t *ptr); - -/** - * atomic64_add_negative - add and test if negative - * @delta: integer value to add - * @ptr: pointer to type atomic64_t - * - * Atomically adds @delta to @ptr and returns true - * if the result is negative, or false when - * result is greater than or equal to zero. - */ -extern int atomic64_add_negative(u64 delta, atomic64_t *ptr); - +#include #include #endif /* _ASM_X86_ATOMIC_32_H */ diff --git a/arch/x86/include/asm/atomic_64.h b/arch/x86/include/asm/atomic_64.h index d605dc2..042c331 100644 --- a/arch/x86/include/asm/atomic_64.h +++ b/arch/x86/include/asm/atomic_64.h @@ -187,196 +187,6 @@ static inline int atomic_sub_return(int i, atomic_t *v) #define atomic_inc_return(v) (atomic_add_return(1, v)) #define atomic_dec_return(v) (atomic_sub_return(1, v)) -/* The 64-bit atomic type */ - -#define ATOMIC64_INIT(i) { (i) } - -/** - * atomic64_read - read atomic64 variable - * @v: pointer of type atomic64_t - * - * Atomically reads the value of @v. - * Doesn't imply a read memory barrier. - */ -static inline long atomic64_read(const atomic64_t *v) -{ - return v->counter; -} - -/** - * atomic64_set - set atomic64 variable - * @v: pointer to type atomic64_t - * @i: required value - * - * Atomically sets the value of @v to @i. - */ -static inline void atomic64_set(atomic64_t *v, long i) -{ - v->counter = i; -} - -/** - * atomic64_add - add integer to atomic64 variable - * @i: integer value to add - * @v: pointer to type atomic64_t - * - * Atomically adds @i to @v. - */ -static inline void atomic64_add(long i, atomic64_t *v) -{ - asm volatile(LOCK_PREFIX "addq %1,%0" - : "=m" (v->counter) - : "er" (i), "m" (v->counter)); -} - -/** - * atomic64_sub - subtract the atomic64 variable - * @i: integer value to subtract - * @v: pointer to type atomic64_t - * - * Atomically subtracts @i from @v. - */ -static inline void atomic64_sub(long i, atomic64_t *v) -{ - asm volatile(LOCK_PREFIX "subq %1,%0" - : "=m" (v->counter) - : "er" (i), "m" (v->counter)); -} - -/** - * atomic64_sub_and_test - subtract value from variable and test result - * @i: integer value to subtract - * @v: pointer to type atomic64_t - * - * Atomically subtracts @i from @v and returns - * true if the result is zero, or false for all - * other cases. - */ -static inline int atomic64_sub_and_test(long i, atomic64_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "subq %2,%0; sete %1" - : "=m" (v->counter), "=qm" (c) - : "er" (i), "m" (v->counter) : "memory"); - return c; -} - -/** - * atomic64_inc - increment atomic64 variable - * @v: pointer to type atomic64_t - * - * Atomically increments @v by 1. - */ -static inline void atomic64_inc(atomic64_t *v) -{ - asm volatile(LOCK_PREFIX "incq %0" - : "=m" (v->counter) - : "m" (v->counter)); -} - -/** - * atomic64_dec - decrement atomic64 variable - * @v: pointer to type atomic64_t - * - * Atomically decrements @v by 1. - */ -static inline void atomic64_dec(atomic64_t *v) -{ - asm volatile(LOCK_PREFIX "decq %0" - : "=m" (v->counter) - : "m" (v->counter)); -} - -/** - * atomic64_dec_and_test - decrement and test - * @v: pointer to type atomic64_t - * - * Atomically decrements @v by 1 and - * returns true if the result is 0, or false for all other - * cases. - */ -static inline int atomic64_dec_and_test(atomic64_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "decq %0; sete %1" - : "=m" (v->counter), "=qm" (c) - : "m" (v->counter) : "memory"); - return c != 0; -} - -/** - * atomic64_inc_and_test - increment and test - * @v: pointer to type atomic64_t - * - * Atomically increments @v by 1 - * and returns true if the result is zero, or false for all - * other cases. - */ -static inline int atomic64_inc_and_test(atomic64_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "incq %0; sete %1" - : "=m" (v->counter), "=qm" (c) - : "m" (v->counter) : "memory"); - return c != 0; -} - -/** - * atomic64_add_negative - add and test if negative - * @i: integer value to add - * @v: pointer to type atomic64_t - * - * Atomically adds @i to @v and returns true - * if the result is negative, or false when - * result is greater than or equal to zero. - */ -static inline int atomic64_add_negative(long i, atomic64_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "addq %2,%0; sets %1" - : "=m" (v->counter), "=qm" (c) - : "er" (i), "m" (v->counter) : "memory"); - return c; -} - -/** - * atomic64_add_return - add and return - * @i: integer value to add - * @v: pointer to type atomic64_t - * - * Atomically adds @i to @v and returns @i + @v - */ -static inline long atomic64_add_return(long i, atomic64_t *v) -{ - long __i = i; - asm volatile(LOCK_PREFIX "xaddq %0, %1;" - : "+r" (i), "+m" (v->counter) - : : "memory"); - return i + __i; -} - -static inline long atomic64_sub_return(long i, atomic64_t *v) -{ - return atomic64_add_return(-i, v); -} - -#define atomic64_inc_return(v) (atomic64_add_return(1, (v))) -#define atomic64_dec_return(v) (atomic64_sub_return(1, (v))) - -static inline long atomic64_cmpxchg(atomic64_t *v, long old, long new) -{ - return cmpxchg(&v->counter, old, new); -} - -static inline long atomic64_xchg(atomic64_t *v, long new) -{ - return xchg(&v->counter, new); -} - static inline long atomic_cmpxchg(atomic_t *v, int old, int new) { return cmpxchg(&v->counter, old, new); @@ -414,30 +224,6 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u) #define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) /** - * atomic64_add_unless - add unless the number is a given value - * @v: pointer of type atomic64_t - * @a: the amount to add to v... - * @u: ...unless v is equal to u. - * - * Atomically adds @a to @v, so long as it was not @u. - * Returns non-zero if @v was not @u, and zero otherwise. - */ -static inline int atomic64_add_unless(atomic64_t *v, long a, long u) -{ - long c, old; - c = atomic64_read(v); - for (;;) { - if (unlikely(c == (u))) - break; - old = atomic64_cmpxchg((v), c, c + (a)); - if (likely(old == c)) - break; - c = old; - } - return c != (u); -} - -/** * atomic_inc_short - increment of a short integer * @v: pointer to type int * @@ -463,8 +249,6 @@ static inline void atomic_or_long(unsigned long *v1, unsigned long v2) asm(LOCK_PREFIX "orq %1, %0" : "+m" (*v1) : "r" (v2)); } -#define atomic64_inc_not_zero(v) atomic64_add_unless((v), 1, 0) - /* These are x86-specific, used by some header files */ #define atomic_clear_mask(mask, addr) \ asm volatile(LOCK_PREFIX "andl %0,%1" \ @@ -481,5 +265,6 @@ static inline void atomic_or_long(unsigned long *v1, unsigned long v2) #define smp_mb__before_atomic_inc() barrier() #define smp_mb__after_atomic_inc() barrier() +#include #include #endif /* _ASM_X86_ATOMIC_64_H */ -- cgit v0.10.2 From 3ce59bb8352e1c53446bef1ead1c63956dfef64a Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Thu, 7 Jan 2010 11:53:34 -0500 Subject: x86: Sync asm/atomic_32.h and asm/atomic_64.h Prepare for merging into asm/atomic.h. Signed-off-by: Brian Gerst LKML-Reference: <1262883215-4034-3-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/atomic_32.h b/arch/x86/include/asm/atomic_32.h index e128ae9..036962e 100644 --- a/arch/x86/include/asm/atomic_32.h +++ b/arch/x86/include/asm/atomic_32.h @@ -4,6 +4,7 @@ #include #include #include +#include #include /* @@ -145,8 +146,8 @@ static inline int atomic_inc_and_test(atomic_t *v) /** * atomic_add_negative - add and test if negative - * @v: pointer of type atomic_t * @i: integer value to add + * @v: pointer of type atomic_t * * Atomically adds @i to @v and returns true * if the result is negative, or false when @@ -164,8 +165,8 @@ static inline int atomic_add_negative(int i, atomic_t *v) /** * atomic_add_return - add integer and return - * @v: pointer of type atomic_t * @i: integer value to add + * @v: pointer of type atomic_t * * Atomically adds @i to @v and returns @i + @v */ @@ -206,6 +207,9 @@ static inline int atomic_sub_return(int i, atomic_t *v) return atomic_add_return(-i, v); } +#define atomic_inc_return(v) (atomic_add_return(1, v)) +#define atomic_dec_return(v) (atomic_sub_return(1, v)) + static inline int atomic_cmpxchg(atomic_t *v, int old, int new) { return cmpxchg(&v->counter, old, new); @@ -242,8 +246,33 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u) #define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) -#define atomic_inc_return(v) (atomic_add_return(1, v)) -#define atomic_dec_return(v) (atomic_sub_return(1, v)) +/** + * atomic_inc_short - increment of a short integer + * @v: pointer to type int + * + * Atomically adds 1 to @v + * Returns the new value of @u + */ +static inline short int atomic_inc_short(short int *v) +{ + asm(LOCK_PREFIX "addw $1, %0" : "+m" (*v)); + return *v; +} + +#ifdef CONFIG_X86_64 +/** + * atomic_or_long - OR of two long integers + * @v1: pointer to type unsigned long + * @v2: pointer to type unsigned long + * + * Atomically ORs @v1 and @v2 + * Returns the result of the OR + */ +static inline void atomic_or_long(unsigned long *v1, unsigned long v2) +{ + asm(LOCK_PREFIX "orq %1, %0" : "+m" (*v1) : "r" (v2)); +} +#endif /* These are x86-specific, used by some header files */ #define atomic_clear_mask(mask, addr) \ @@ -251,8 +280,9 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u) : : "r" (~(mask)), "m" (*(addr)) : "memory") #define atomic_set_mask(mask, addr) \ - asm volatile(LOCK_PREFIX "orl %0,%1" \ - : : "r" (mask), "m" (*(addr)) : "memory") + asm volatile(LOCK_PREFIX "orl %0,%1" \ + : : "r" ((unsigned)(mask)), "m" (*(addr)) \ + : "memory") /* Atomic operations are already serializing on x86 */ #define smp_mb__before_atomic_dec() barrier() diff --git a/arch/x86/include/asm/atomic_64.h b/arch/x86/include/asm/atomic_64.h index 042c331..7740788 100644 --- a/arch/x86/include/asm/atomic_64.h +++ b/arch/x86/include/asm/atomic_64.h @@ -1,7 +1,9 @@ #ifndef _ASM_X86_ATOMIC_64_H #define _ASM_X86_ATOMIC_64_H +#include #include +#include #include #include @@ -45,12 +47,12 @@ static inline void atomic_set(atomic_t *v, int i) static inline void atomic_add(int i, atomic_t *v) { asm volatile(LOCK_PREFIX "addl %1,%0" - : "=m" (v->counter) - : "ir" (i), "m" (v->counter)); + : "+m" (v->counter) + : "ir" (i)); } /** - * atomic_sub - subtract the atomic variable + * atomic_sub - subtract integer from atomic variable * @i: integer value to subtract * @v: pointer of type atomic_t * @@ -59,8 +61,8 @@ static inline void atomic_add(int i, atomic_t *v) static inline void atomic_sub(int i, atomic_t *v) { asm volatile(LOCK_PREFIX "subl %1,%0" - : "=m" (v->counter) - : "ir" (i), "m" (v->counter)); + : "+m" (v->counter) + : "ir" (i)); } /** @@ -77,8 +79,8 @@ static inline int atomic_sub_and_test(int i, atomic_t *v) unsigned char c; asm volatile(LOCK_PREFIX "subl %2,%0; sete %1" - : "=m" (v->counter), "=qm" (c) - : "ir" (i), "m" (v->counter) : "memory"); + : "+m" (v->counter), "=qm" (c) + : "ir" (i) : "memory"); return c; } @@ -91,8 +93,7 @@ static inline int atomic_sub_and_test(int i, atomic_t *v) static inline void atomic_inc(atomic_t *v) { asm volatile(LOCK_PREFIX "incl %0" - : "=m" (v->counter) - : "m" (v->counter)); + : "+m" (v->counter)); } /** @@ -104,8 +105,7 @@ static inline void atomic_inc(atomic_t *v) static inline void atomic_dec(atomic_t *v) { asm volatile(LOCK_PREFIX "decl %0" - : "=m" (v->counter) - : "m" (v->counter)); + : "+m" (v->counter)); } /** @@ -121,8 +121,8 @@ static inline int atomic_dec_and_test(atomic_t *v) unsigned char c; asm volatile(LOCK_PREFIX "decl %0; sete %1" - : "=m" (v->counter), "=qm" (c) - : "m" (v->counter) : "memory"); + : "+m" (v->counter), "=qm" (c) + : : "memory"); return c != 0; } @@ -139,8 +139,8 @@ static inline int atomic_inc_and_test(atomic_t *v) unsigned char c; asm volatile(LOCK_PREFIX "incl %0; sete %1" - : "=m" (v->counter), "=qm" (c) - : "m" (v->counter) : "memory"); + : "+m" (v->counter), "=qm" (c) + : : "memory"); return c != 0; } @@ -158,13 +158,13 @@ static inline int atomic_add_negative(int i, atomic_t *v) unsigned char c; asm volatile(LOCK_PREFIX "addl %2,%0; sets %1" - : "=m" (v->counter), "=qm" (c) - : "ir" (i), "m" (v->counter) : "memory"); + : "+m" (v->counter), "=qm" (c) + : "ir" (i) : "memory"); return c; } /** - * atomic_add_return - add and return + * atomic_add_return - add integer and return * @i: integer value to add * @v: pointer of type atomic_t * @@ -172,13 +172,36 @@ static inline int atomic_add_negative(int i, atomic_t *v) */ static inline int atomic_add_return(int i, atomic_t *v) { - int __i = i; + int __i; +#ifdef CONFIG_M386 + unsigned long flags; + if (unlikely(boot_cpu_data.x86 <= 3)) + goto no_xadd; +#endif + /* Modern 486+ processor */ + __i = i; asm volatile(LOCK_PREFIX "xaddl %0, %1" : "+r" (i), "+m" (v->counter) : : "memory"); return i + __i; + +#ifdef CONFIG_M386 +no_xadd: /* Legacy 386 processor */ + local_irq_save(flags); + __i = atomic_read(v); + atomic_set(v, i + __i); + local_irq_restore(flags); + return i + __i; +#endif } +/** + * atomic_sub_return - subtract integer and return + * @v: pointer of type atomic_t + * @i: integer value to subtract + * + * Atomically subtracts @i from @v and returns @v - @i + */ static inline int atomic_sub_return(int i, atomic_t *v) { return atomic_add_return(-i, v); @@ -187,23 +210,23 @@ static inline int atomic_sub_return(int i, atomic_t *v) #define atomic_inc_return(v) (atomic_add_return(1, v)) #define atomic_dec_return(v) (atomic_sub_return(1, v)) -static inline long atomic_cmpxchg(atomic_t *v, int old, int new) +static inline int atomic_cmpxchg(atomic_t *v, int old, int new) { return cmpxchg(&v->counter, old, new); } -static inline long atomic_xchg(atomic_t *v, int new) +static inline int atomic_xchg(atomic_t *v, int new) { return xchg(&v->counter, new); } /** - * atomic_add_unless - add unless the number is a given value + * atomic_add_unless - add unless the number is already a given value * @v: pointer of type atomic_t * @a: the amount to add to v... * @u: ...unless v is equal to u. * - * Atomically adds @a to @v, so long as it was not @u. + * Atomically adds @a to @v, so long as @v was not already @u. * Returns non-zero if @v was not @u, and zero otherwise. */ static inline int atomic_add_unless(atomic_t *v, int a, int u) @@ -236,6 +259,7 @@ static inline short int atomic_inc_short(short int *v) return *v; } +#ifdef CONFIG_X86_64 /** * atomic_or_long - OR of two long integers * @v1: pointer to type unsigned long @@ -248,15 +272,16 @@ static inline void atomic_or_long(unsigned long *v1, unsigned long v2) { asm(LOCK_PREFIX "orq %1, %0" : "+m" (*v1) : "r" (v2)); } +#endif /* These are x86-specific, used by some header files */ -#define atomic_clear_mask(mask, addr) \ - asm volatile(LOCK_PREFIX "andl %0,%1" \ +#define atomic_clear_mask(mask, addr) \ + asm volatile(LOCK_PREFIX "andl %0,%1" \ : : "r" (~(mask)), "m" (*(addr)) : "memory") -#define atomic_set_mask(mask, addr) \ - asm volatile(LOCK_PREFIX "orl %0,%1" \ - : : "r" ((unsigned)(mask)), "m" (*(addr)) \ +#define atomic_set_mask(mask, addr) \ + asm volatile(LOCK_PREFIX "orl %0,%1" \ + : : "r" ((unsigned)(mask)), "m" (*(addr)) \ : "memory") /* Atomic operations are already serializing on x86 */ -- cgit v0.10.2 From 5abbbbf0b0cd4abf5898136d0c345dc99b859c8c Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Thu, 7 Jan 2010 11:53:35 -0500 Subject: x86: Merge asm/atomic_{32,64}.h Merge the now identical code from asm/atomic_32.h and asm/atomic_64.h into asm/atomic.h. Signed-off-by: Brian Gerst LKML-Reference: <1262883215-4034-4-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/atomic.h b/arch/x86/include/asm/atomic.h index 4e1b887..8baaa71 100644 --- a/arch/x86/include/asm/atomic.h +++ b/arch/x86/include/asm/atomic.h @@ -1,5 +1,300 @@ +#ifndef _ASM_X86_ATOMIC_H +#define _ASM_X86_ATOMIC_H + +#include +#include +#include +#include +#include + +/* + * Atomic operations that C can't guarantee us. Useful for + * resource counting etc.. + */ + +#define ATOMIC_INIT(i) { (i) } + +/** + * atomic_read - read atomic variable + * @v: pointer of type atomic_t + * + * Atomically reads the value of @v. + */ +static inline int atomic_read(const atomic_t *v) +{ + return v->counter; +} + +/** + * atomic_set - set atomic variable + * @v: pointer of type atomic_t + * @i: required value + * + * Atomically sets the value of @v to @i. + */ +static inline void atomic_set(atomic_t *v, int i) +{ + v->counter = i; +} + +/** + * atomic_add - add integer to atomic variable + * @i: integer value to add + * @v: pointer of type atomic_t + * + * Atomically adds @i to @v. + */ +static inline void atomic_add(int i, atomic_t *v) +{ + asm volatile(LOCK_PREFIX "addl %1,%0" + : "+m" (v->counter) + : "ir" (i)); +} + +/** + * atomic_sub - subtract integer from atomic variable + * @i: integer value to subtract + * @v: pointer of type atomic_t + * + * Atomically subtracts @i from @v. + */ +static inline void atomic_sub(int i, atomic_t *v) +{ + asm volatile(LOCK_PREFIX "subl %1,%0" + : "+m" (v->counter) + : "ir" (i)); +} + +/** + * atomic_sub_and_test - subtract value from variable and test result + * @i: integer value to subtract + * @v: pointer of type atomic_t + * + * Atomically subtracts @i from @v and returns + * true if the result is zero, or false for all + * other cases. + */ +static inline int atomic_sub_and_test(int i, atomic_t *v) +{ + unsigned char c; + + asm volatile(LOCK_PREFIX "subl %2,%0; sete %1" + : "+m" (v->counter), "=qm" (c) + : "ir" (i) : "memory"); + return c; +} + +/** + * atomic_inc - increment atomic variable + * @v: pointer of type atomic_t + * + * Atomically increments @v by 1. + */ +static inline void atomic_inc(atomic_t *v) +{ + asm volatile(LOCK_PREFIX "incl %0" + : "+m" (v->counter)); +} + +/** + * atomic_dec - decrement atomic variable + * @v: pointer of type atomic_t + * + * Atomically decrements @v by 1. + */ +static inline void atomic_dec(atomic_t *v) +{ + asm volatile(LOCK_PREFIX "decl %0" + : "+m" (v->counter)); +} + +/** + * atomic_dec_and_test - decrement and test + * @v: pointer of type atomic_t + * + * Atomically decrements @v by 1 and + * returns true if the result is 0, or false for all other + * cases. + */ +static inline int atomic_dec_and_test(atomic_t *v) +{ + unsigned char c; + + asm volatile(LOCK_PREFIX "decl %0; sete %1" + : "+m" (v->counter), "=qm" (c) + : : "memory"); + return c != 0; +} + +/** + * atomic_inc_and_test - increment and test + * @v: pointer of type atomic_t + * + * Atomically increments @v by 1 + * and returns true if the result is zero, or false for all + * other cases. + */ +static inline int atomic_inc_and_test(atomic_t *v) +{ + unsigned char c; + + asm volatile(LOCK_PREFIX "incl %0; sete %1" + : "+m" (v->counter), "=qm" (c) + : : "memory"); + return c != 0; +} + +/** + * atomic_add_negative - add and test if negative + * @i: integer value to add + * @v: pointer of type atomic_t + * + * Atomically adds @i to @v and returns true + * if the result is negative, or false when + * result is greater than or equal to zero. + */ +static inline int atomic_add_negative(int i, atomic_t *v) +{ + unsigned char c; + + asm volatile(LOCK_PREFIX "addl %2,%0; sets %1" + : "+m" (v->counter), "=qm" (c) + : "ir" (i) : "memory"); + return c; +} + +/** + * atomic_add_return - add integer and return + * @i: integer value to add + * @v: pointer of type atomic_t + * + * Atomically adds @i to @v and returns @i + @v + */ +static inline int atomic_add_return(int i, atomic_t *v) +{ + int __i; +#ifdef CONFIG_M386 + unsigned long flags; + if (unlikely(boot_cpu_data.x86 <= 3)) + goto no_xadd; +#endif + /* Modern 486+ processor */ + __i = i; + asm volatile(LOCK_PREFIX "xaddl %0, %1" + : "+r" (i), "+m" (v->counter) + : : "memory"); + return i + __i; + +#ifdef CONFIG_M386 +no_xadd: /* Legacy 386 processor */ + local_irq_save(flags); + __i = atomic_read(v); + atomic_set(v, i + __i); + local_irq_restore(flags); + return i + __i; +#endif +} + +/** + * atomic_sub_return - subtract integer and return + * @v: pointer of type atomic_t + * @i: integer value to subtract + * + * Atomically subtracts @i from @v and returns @v - @i + */ +static inline int atomic_sub_return(int i, atomic_t *v) +{ + return atomic_add_return(-i, v); +} + +#define atomic_inc_return(v) (atomic_add_return(1, v)) +#define atomic_dec_return(v) (atomic_sub_return(1, v)) + +static inline int atomic_cmpxchg(atomic_t *v, int old, int new) +{ + return cmpxchg(&v->counter, old, new); +} + +static inline int atomic_xchg(atomic_t *v, int new) +{ + return xchg(&v->counter, new); +} + +/** + * atomic_add_unless - add unless the number is already a given value + * @v: pointer of type atomic_t + * @a: the amount to add to v... + * @u: ...unless v is equal to u. + * + * Atomically adds @a to @v, so long as @v was not already @u. + * Returns non-zero if @v was not @u, and zero otherwise. + */ +static inline int atomic_add_unless(atomic_t *v, int a, int u) +{ + int c, old; + c = atomic_read(v); + for (;;) { + if (unlikely(c == (u))) + break; + old = atomic_cmpxchg((v), c, c + (a)); + if (likely(old == c)) + break; + c = old; + } + return c != (u); +} + +#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) + +/** + * atomic_inc_short - increment of a short integer + * @v: pointer to type int + * + * Atomically adds 1 to @v + * Returns the new value of @u + */ +static inline short int atomic_inc_short(short int *v) +{ + asm(LOCK_PREFIX "addw $1, %0" : "+m" (*v)); + return *v; +} + +#ifdef CONFIG_X86_64 +/** + * atomic_or_long - OR of two long integers + * @v1: pointer to type unsigned long + * @v2: pointer to type unsigned long + * + * Atomically ORs @v1 and @v2 + * Returns the result of the OR + */ +static inline void atomic_or_long(unsigned long *v1, unsigned long v2) +{ + asm(LOCK_PREFIX "orq %1, %0" : "+m" (*v1) : "r" (v2)); +} +#endif + +/* These are x86-specific, used by some header files */ +#define atomic_clear_mask(mask, addr) \ + asm volatile(LOCK_PREFIX "andl %0,%1" \ + : : "r" (~(mask)), "m" (*(addr)) : "memory") + +#define atomic_set_mask(mask, addr) \ + asm volatile(LOCK_PREFIX "orl %0,%1" \ + : : "r" ((unsigned)(mask)), "m" (*(addr)) \ + : "memory") + +/* Atomic operations are already serializing on x86 */ +#define smp_mb__before_atomic_dec() barrier() +#define smp_mb__after_atomic_dec() barrier() +#define smp_mb__before_atomic_inc() barrier() +#define smp_mb__after_atomic_inc() barrier() + #ifdef CONFIG_X86_32 -# include "atomic_32.h" +# include "atomic64_32.h" #else -# include "atomic_64.h" +# include "atomic64_64.h" #endif + +#include +#endif /* _ASM_X86_ATOMIC_H */ diff --git a/arch/x86/include/asm/atomic_32.h b/arch/x86/include/asm/atomic_32.h deleted file mode 100644 index 036962e..0000000 --- a/arch/x86/include/asm/atomic_32.h +++ /dev/null @@ -1,295 +0,0 @@ -#ifndef _ASM_X86_ATOMIC_32_H -#define _ASM_X86_ATOMIC_32_H - -#include -#include -#include -#include -#include - -/* - * Atomic operations that C can't guarantee us. Useful for - * resource counting etc.. - */ - -#define ATOMIC_INIT(i) { (i) } - -/** - * atomic_read - read atomic variable - * @v: pointer of type atomic_t - * - * Atomically reads the value of @v. - */ -static inline int atomic_read(const atomic_t *v) -{ - return v->counter; -} - -/** - * atomic_set - set atomic variable - * @v: pointer of type atomic_t - * @i: required value - * - * Atomically sets the value of @v to @i. - */ -static inline void atomic_set(atomic_t *v, int i) -{ - v->counter = i; -} - -/** - * atomic_add - add integer to atomic variable - * @i: integer value to add - * @v: pointer of type atomic_t - * - * Atomically adds @i to @v. - */ -static inline void atomic_add(int i, atomic_t *v) -{ - asm volatile(LOCK_PREFIX "addl %1,%0" - : "+m" (v->counter) - : "ir" (i)); -} - -/** - * atomic_sub - subtract integer from atomic variable - * @i: integer value to subtract - * @v: pointer of type atomic_t - * - * Atomically subtracts @i from @v. - */ -static inline void atomic_sub(int i, atomic_t *v) -{ - asm volatile(LOCK_PREFIX "subl %1,%0" - : "+m" (v->counter) - : "ir" (i)); -} - -/** - * atomic_sub_and_test - subtract value from variable and test result - * @i: integer value to subtract - * @v: pointer of type atomic_t - * - * Atomically subtracts @i from @v and returns - * true if the result is zero, or false for all - * other cases. - */ -static inline int atomic_sub_and_test(int i, atomic_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "subl %2,%0; sete %1" - : "+m" (v->counter), "=qm" (c) - : "ir" (i) : "memory"); - return c; -} - -/** - * atomic_inc - increment atomic variable - * @v: pointer of type atomic_t - * - * Atomically increments @v by 1. - */ -static inline void atomic_inc(atomic_t *v) -{ - asm volatile(LOCK_PREFIX "incl %0" - : "+m" (v->counter)); -} - -/** - * atomic_dec - decrement atomic variable - * @v: pointer of type atomic_t - * - * Atomically decrements @v by 1. - */ -static inline void atomic_dec(atomic_t *v) -{ - asm volatile(LOCK_PREFIX "decl %0" - : "+m" (v->counter)); -} - -/** - * atomic_dec_and_test - decrement and test - * @v: pointer of type atomic_t - * - * Atomically decrements @v by 1 and - * returns true if the result is 0, or false for all other - * cases. - */ -static inline int atomic_dec_and_test(atomic_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "decl %0; sete %1" - : "+m" (v->counter), "=qm" (c) - : : "memory"); - return c != 0; -} - -/** - * atomic_inc_and_test - increment and test - * @v: pointer of type atomic_t - * - * Atomically increments @v by 1 - * and returns true if the result is zero, or false for all - * other cases. - */ -static inline int atomic_inc_and_test(atomic_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "incl %0; sete %1" - : "+m" (v->counter), "=qm" (c) - : : "memory"); - return c != 0; -} - -/** - * atomic_add_negative - add and test if negative - * @i: integer value to add - * @v: pointer of type atomic_t - * - * Atomically adds @i to @v and returns true - * if the result is negative, or false when - * result is greater than or equal to zero. - */ -static inline int atomic_add_negative(int i, atomic_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "addl %2,%0; sets %1" - : "+m" (v->counter), "=qm" (c) - : "ir" (i) : "memory"); - return c; -} - -/** - * atomic_add_return - add integer and return - * @i: integer value to add - * @v: pointer of type atomic_t - * - * Atomically adds @i to @v and returns @i + @v - */ -static inline int atomic_add_return(int i, atomic_t *v) -{ - int __i; -#ifdef CONFIG_M386 - unsigned long flags; - if (unlikely(boot_cpu_data.x86 <= 3)) - goto no_xadd; -#endif - /* Modern 486+ processor */ - __i = i; - asm volatile(LOCK_PREFIX "xaddl %0, %1" - : "+r" (i), "+m" (v->counter) - : : "memory"); - return i + __i; - -#ifdef CONFIG_M386 -no_xadd: /* Legacy 386 processor */ - local_irq_save(flags); - __i = atomic_read(v); - atomic_set(v, i + __i); - local_irq_restore(flags); - return i + __i; -#endif -} - -/** - * atomic_sub_return - subtract integer and return - * @v: pointer of type atomic_t - * @i: integer value to subtract - * - * Atomically subtracts @i from @v and returns @v - @i - */ -static inline int atomic_sub_return(int i, atomic_t *v) -{ - return atomic_add_return(-i, v); -} - -#define atomic_inc_return(v) (atomic_add_return(1, v)) -#define atomic_dec_return(v) (atomic_sub_return(1, v)) - -static inline int atomic_cmpxchg(atomic_t *v, int old, int new) -{ - return cmpxchg(&v->counter, old, new); -} - -static inline int atomic_xchg(atomic_t *v, int new) -{ - return xchg(&v->counter, new); -} - -/** - * atomic_add_unless - add unless the number is already a given value - * @v: pointer of type atomic_t - * @a: the amount to add to v... - * @u: ...unless v is equal to u. - * - * Atomically adds @a to @v, so long as @v was not already @u. - * Returns non-zero if @v was not @u, and zero otherwise. - */ -static inline int atomic_add_unless(atomic_t *v, int a, int u) -{ - int c, old; - c = atomic_read(v); - for (;;) { - if (unlikely(c == (u))) - break; - old = atomic_cmpxchg((v), c, c + (a)); - if (likely(old == c)) - break; - c = old; - } - return c != (u); -} - -#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) - -/** - * atomic_inc_short - increment of a short integer - * @v: pointer to type int - * - * Atomically adds 1 to @v - * Returns the new value of @u - */ -static inline short int atomic_inc_short(short int *v) -{ - asm(LOCK_PREFIX "addw $1, %0" : "+m" (*v)); - return *v; -} - -#ifdef CONFIG_X86_64 -/** - * atomic_or_long - OR of two long integers - * @v1: pointer to type unsigned long - * @v2: pointer to type unsigned long - * - * Atomically ORs @v1 and @v2 - * Returns the result of the OR - */ -static inline void atomic_or_long(unsigned long *v1, unsigned long v2) -{ - asm(LOCK_PREFIX "orq %1, %0" : "+m" (*v1) : "r" (v2)); -} -#endif - -/* These are x86-specific, used by some header files */ -#define atomic_clear_mask(mask, addr) \ - asm volatile(LOCK_PREFIX "andl %0,%1" \ - : : "r" (~(mask)), "m" (*(addr)) : "memory") - -#define atomic_set_mask(mask, addr) \ - asm volatile(LOCK_PREFIX "orl %0,%1" \ - : : "r" ((unsigned)(mask)), "m" (*(addr)) \ - : "memory") - -/* Atomic operations are already serializing on x86 */ -#define smp_mb__before_atomic_dec() barrier() -#define smp_mb__after_atomic_dec() barrier() -#define smp_mb__before_atomic_inc() barrier() -#define smp_mb__after_atomic_inc() barrier() - -#include -#include -#endif /* _ASM_X86_ATOMIC_32_H */ diff --git a/arch/x86/include/asm/atomic_64.h b/arch/x86/include/asm/atomic_64.h deleted file mode 100644 index 7740788..0000000 --- a/arch/x86/include/asm/atomic_64.h +++ /dev/null @@ -1,295 +0,0 @@ -#ifndef _ASM_X86_ATOMIC_64_H -#define _ASM_X86_ATOMIC_64_H - -#include -#include -#include -#include -#include - -/* - * Atomic operations that C can't guarantee us. Useful for - * resource counting etc.. - */ - -#define ATOMIC_INIT(i) { (i) } - -/** - * atomic_read - read atomic variable - * @v: pointer of type atomic_t - * - * Atomically reads the value of @v. - */ -static inline int atomic_read(const atomic_t *v) -{ - return v->counter; -} - -/** - * atomic_set - set atomic variable - * @v: pointer of type atomic_t - * @i: required value - * - * Atomically sets the value of @v to @i. - */ -static inline void atomic_set(atomic_t *v, int i) -{ - v->counter = i; -} - -/** - * atomic_add - add integer to atomic variable - * @i: integer value to add - * @v: pointer of type atomic_t - * - * Atomically adds @i to @v. - */ -static inline void atomic_add(int i, atomic_t *v) -{ - asm volatile(LOCK_PREFIX "addl %1,%0" - : "+m" (v->counter) - : "ir" (i)); -} - -/** - * atomic_sub - subtract integer from atomic variable - * @i: integer value to subtract - * @v: pointer of type atomic_t - * - * Atomically subtracts @i from @v. - */ -static inline void atomic_sub(int i, atomic_t *v) -{ - asm volatile(LOCK_PREFIX "subl %1,%0" - : "+m" (v->counter) - : "ir" (i)); -} - -/** - * atomic_sub_and_test - subtract value from variable and test result - * @i: integer value to subtract - * @v: pointer of type atomic_t - * - * Atomically subtracts @i from @v and returns - * true if the result is zero, or false for all - * other cases. - */ -static inline int atomic_sub_and_test(int i, atomic_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "subl %2,%0; sete %1" - : "+m" (v->counter), "=qm" (c) - : "ir" (i) : "memory"); - return c; -} - -/** - * atomic_inc - increment atomic variable - * @v: pointer of type atomic_t - * - * Atomically increments @v by 1. - */ -static inline void atomic_inc(atomic_t *v) -{ - asm volatile(LOCK_PREFIX "incl %0" - : "+m" (v->counter)); -} - -/** - * atomic_dec - decrement atomic variable - * @v: pointer of type atomic_t - * - * Atomically decrements @v by 1. - */ -static inline void atomic_dec(atomic_t *v) -{ - asm volatile(LOCK_PREFIX "decl %0" - : "+m" (v->counter)); -} - -/** - * atomic_dec_and_test - decrement and test - * @v: pointer of type atomic_t - * - * Atomically decrements @v by 1 and - * returns true if the result is 0, or false for all other - * cases. - */ -static inline int atomic_dec_and_test(atomic_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "decl %0; sete %1" - : "+m" (v->counter), "=qm" (c) - : : "memory"); - return c != 0; -} - -/** - * atomic_inc_and_test - increment and test - * @v: pointer of type atomic_t - * - * Atomically increments @v by 1 - * and returns true if the result is zero, or false for all - * other cases. - */ -static inline int atomic_inc_and_test(atomic_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "incl %0; sete %1" - : "+m" (v->counter), "=qm" (c) - : : "memory"); - return c != 0; -} - -/** - * atomic_add_negative - add and test if negative - * @i: integer value to add - * @v: pointer of type atomic_t - * - * Atomically adds @i to @v and returns true - * if the result is negative, or false when - * result is greater than or equal to zero. - */ -static inline int atomic_add_negative(int i, atomic_t *v) -{ - unsigned char c; - - asm volatile(LOCK_PREFIX "addl %2,%0; sets %1" - : "+m" (v->counter), "=qm" (c) - : "ir" (i) : "memory"); - return c; -} - -/** - * atomic_add_return - add integer and return - * @i: integer value to add - * @v: pointer of type atomic_t - * - * Atomically adds @i to @v and returns @i + @v - */ -static inline int atomic_add_return(int i, atomic_t *v) -{ - int __i; -#ifdef CONFIG_M386 - unsigned long flags; - if (unlikely(boot_cpu_data.x86 <= 3)) - goto no_xadd; -#endif - /* Modern 486+ processor */ - __i = i; - asm volatile(LOCK_PREFIX "xaddl %0, %1" - : "+r" (i), "+m" (v->counter) - : : "memory"); - return i + __i; - -#ifdef CONFIG_M386 -no_xadd: /* Legacy 386 processor */ - local_irq_save(flags); - __i = atomic_read(v); - atomic_set(v, i + __i); - local_irq_restore(flags); - return i + __i; -#endif -} - -/** - * atomic_sub_return - subtract integer and return - * @v: pointer of type atomic_t - * @i: integer value to subtract - * - * Atomically subtracts @i from @v and returns @v - @i - */ -static inline int atomic_sub_return(int i, atomic_t *v) -{ - return atomic_add_return(-i, v); -} - -#define atomic_inc_return(v) (atomic_add_return(1, v)) -#define atomic_dec_return(v) (atomic_sub_return(1, v)) - -static inline int atomic_cmpxchg(atomic_t *v, int old, int new) -{ - return cmpxchg(&v->counter, old, new); -} - -static inline int atomic_xchg(atomic_t *v, int new) -{ - return xchg(&v->counter, new); -} - -/** - * atomic_add_unless - add unless the number is already a given value - * @v: pointer of type atomic_t - * @a: the amount to add to v... - * @u: ...unless v is equal to u. - * - * Atomically adds @a to @v, so long as @v was not already @u. - * Returns non-zero if @v was not @u, and zero otherwise. - */ -static inline int atomic_add_unless(atomic_t *v, int a, int u) -{ - int c, old; - c = atomic_read(v); - for (;;) { - if (unlikely(c == (u))) - break; - old = atomic_cmpxchg((v), c, c + (a)); - if (likely(old == c)) - break; - c = old; - } - return c != (u); -} - -#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) - -/** - * atomic_inc_short - increment of a short integer - * @v: pointer to type int - * - * Atomically adds 1 to @v - * Returns the new value of @u - */ -static inline short int atomic_inc_short(short int *v) -{ - asm(LOCK_PREFIX "addw $1, %0" : "+m" (*v)); - return *v; -} - -#ifdef CONFIG_X86_64 -/** - * atomic_or_long - OR of two long integers - * @v1: pointer to type unsigned long - * @v2: pointer to type unsigned long - * - * Atomically ORs @v1 and @v2 - * Returns the result of the OR - */ -static inline void atomic_or_long(unsigned long *v1, unsigned long v2) -{ - asm(LOCK_PREFIX "orq %1, %0" : "+m" (*v1) : "r" (v2)); -} -#endif - -/* These are x86-specific, used by some header files */ -#define atomic_clear_mask(mask, addr) \ - asm volatile(LOCK_PREFIX "andl %0,%1" \ - : : "r" (~(mask)), "m" (*(addr)) : "memory") - -#define atomic_set_mask(mask, addr) \ - asm volatile(LOCK_PREFIX "orl %0,%1" \ - : : "r" ((unsigned)(mask)), "m" (*(addr)) \ - : "memory") - -/* Atomic operations are already serializing on x86 */ -#define smp_mb__before_atomic_dec() barrier() -#define smp_mb__after_atomic_dec() barrier() -#define smp_mb__before_atomic_inc() barrier() -#define smp_mb__after_atomic_inc() barrier() - -#include -#include -#endif /* _ASM_X86_ATOMIC_64_H */ -- cgit v0.10.2 From e1e0138d7d10fd447c71cc70f367eac514bd3ce4 Mon Sep 17 00:00:00 2001 From: Jack Steiner Date: Thu, 7 Jan 2010 10:12:40 -0600 Subject: x86, uv: uv_global_gru_mmr_address() macro fix Fix bug in uv_global_gru_mmr_address macro. Macro failed to cast an int value to a long prior to a left shift > 32. Signed-off-by: Jack Steiner LKML-Reference: <20100107161240.GA2610@sgi.com> Cc: Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h index 811bfab..7a81d9d 100644 --- a/arch/x86/include/asm/uv/uv_hub.h +++ b/arch/x86/include/asm/uv/uv_hub.h @@ -335,7 +335,8 @@ static inline unsigned long uv_read_global_mmr64(int pnode, */ static inline unsigned long uv_global_gru_mmr_address(int pnode, unsigned long offset) { - return UV_GLOBAL_GRU_MMR_BASE | offset | (pnode << uv_hub_info->m_val); + return UV_GLOBAL_GRU_MMR_BASE | offset | + ((unsigned long)pnode << uv_hub_info->m_val); } /* -- cgit v0.10.2 From 99659a929d653d0c9ce458091870544768add871 Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Thu, 7 Jan 2010 15:35:42 +0100 Subject: x86, uv: Remove recursion in uv_heartbeat_enable() The recursion is not needed and does not improve readability. Signed-off-by: Roel Kluin LKML-Reference: <4B45F13E.3040202@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index af5d103..d199dc3 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -475,7 +475,7 @@ static void uv_heartbeat(unsigned long ignored) static void __cpuinit uv_heartbeat_enable(int cpu) { - if (!uv_cpu_hub_info(cpu)->scir.enabled) { + while (!uv_cpu_hub_info(cpu)->scir.enabled) { struct timer_list *timer = &uv_cpu_hub_info(cpu)->scir.timer; uv_set_cpu_scir_bits(cpu, SCIR_CPU_HEARTBEAT|SCIR_CPU_ACTIVITY); @@ -483,11 +483,10 @@ static void __cpuinit uv_heartbeat_enable(int cpu) timer->expires = jiffies + SCIR_CPU_HB_INTERVAL; add_timer_on(timer, cpu); uv_cpu_hub_info(cpu)->scir.enabled = 1; - } - /* check boot cpu */ - if (!uv_cpu_hub_info(0)->scir.enabled) - uv_heartbeat_enable(0); + /* also ensure that boot cpu is enabled */ + cpu = 0; + } } #ifdef CONFIG_HOTPLUG_CPU -- cgit v0.10.2 From d7ac769068df87ca8c7f72d99cf67ead16739f18 Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Fri, 8 Jan 2010 14:18:34 +1100 Subject: crypto: aes_s390 - access .cip instead of .blk in cipher mode The fallback code in cipher mode touch the union fallback.blk instead of fallback.cip. This is wrong because we use the cipher and not the blockcipher. This did not show any side effects yet because both types / structs contain the same element right now. Signed-off-by: Roel Kluin Signed-off-by: Sebastian Andrzej Siewior Signed-off-by: Herbert Xu diff --git a/arch/s390/crypto/aes_s390.c b/arch/s390/crypto/aes_s390.c index 6118890..9ed05bc 100644 --- a/arch/s390/crypto/aes_s390.c +++ b/arch/s390/crypto/aes_s390.c @@ -78,14 +78,14 @@ static int setkey_fallback_cip(struct crypto_tfm *tfm, const u8 *in_key, struct s390_aes_ctx *sctx = crypto_tfm_ctx(tfm); int ret; - sctx->fallback.blk->base.crt_flags &= ~CRYPTO_TFM_REQ_MASK; - sctx->fallback.blk->base.crt_flags |= (tfm->crt_flags & + sctx->fallback.cip->base.crt_flags &= ~CRYPTO_TFM_REQ_MASK; + sctx->fallback.cip->base.crt_flags |= (tfm->crt_flags & CRYPTO_TFM_REQ_MASK); ret = crypto_cipher_setkey(sctx->fallback.cip, in_key, key_len); if (ret) { tfm->crt_flags &= ~CRYPTO_TFM_RES_MASK; - tfm->crt_flags |= (sctx->fallback.blk->base.crt_flags & + tfm->crt_flags |= (sctx->fallback.cip->base.crt_flags & CRYPTO_TFM_RES_MASK); } return ret; -- cgit v0.10.2 From faad98f29606d9d3c6bddae7c88693be37d2fb43 Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Fri, 8 Jan 2010 14:19:21 +1100 Subject: crypto: geode-aes - access .cip instead of .blk in cipher mode The fallback code in cipher mode touch the union fallback.blk instead of fallback.cip. This is wrong because we use the cipher and not the blockcipher. This did not show any side effects yet because both types / structs contain the same element right now. Signed-off-by: Roel Kluin Signed-off-by: Sebastian Andrzej Siewior Signed-off-by: Herbert Xu diff --git a/drivers/crypto/geode-aes.c b/drivers/crypto/geode-aes.c index 4801162..03e71b1 100644 --- a/drivers/crypto/geode-aes.c +++ b/drivers/crypto/geode-aes.c @@ -135,8 +135,8 @@ static int geode_setkey_cip(struct crypto_tfm *tfm, const u8 *key, /* * The requested key size is not supported by HW, do a fallback */ - op->fallback.blk->base.crt_flags &= ~CRYPTO_TFM_REQ_MASK; - op->fallback.blk->base.crt_flags |= (tfm->crt_flags & CRYPTO_TFM_REQ_MASK); + op->fallback.cip->base.crt_flags &= ~CRYPTO_TFM_REQ_MASK; + op->fallback.cip->base.crt_flags |= (tfm->crt_flags & CRYPTO_TFM_REQ_MASK); ret = crypto_cipher_setkey(op->fallback.cip, key, len); if (ret) { @@ -263,7 +263,7 @@ static int fallback_init_cip(struct crypto_tfm *tfm) if (IS_ERR(op->fallback.cip)) { printk(KERN_ERR "Error allocating fallback algo %s\n", name); - return PTR_ERR(op->fallback.blk); + return PTR_ERR(op->fallback.cip); } return 0; -- cgit v0.10.2 From 8c0b8139c87cfe8b95c6e763b4ca3190aa9b1ad0 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 8 Jan 2010 17:02:17 +0900 Subject: sh: consolidate atomic_cmpxchg()/atomic_add_unless() definitions. The LL/SC and IRQ versions were using generic stubs while the GRB version was just reimplementing what it already had for the standard cmpxchg() code. As we have optimized cmpxchg() implementations that are decoupled from the atomic code, simply falling back on the generic wrapper does the right thing. With this in place the GRB case is unaffected while the LL/SC case gets to use its optimized cmpxchg(). Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/atomic-grb.h b/arch/sh/include/asm/atomic-grb.h index 4c5b7db..a273c88 100644 --- a/arch/sh/include/asm/atomic-grb.h +++ b/arch/sh/include/asm/atomic-grb.h @@ -120,50 +120,4 @@ static inline void atomic_set_mask(unsigned int mask, atomic_t *v) : "memory" , "r0", "r1"); } -static inline int atomic_cmpxchg(atomic_t *v, int old, int new) -{ - int ret; - - __asm__ __volatile__ ( - " .align 2 \n\t" - " mova 1f, r0 \n\t" - " nop \n\t" - " mov r15, r1 \n\t" - " mov #-8, r15 \n\t" - " mov.l @%1, %0 \n\t" - " cmp/eq %2, %0 \n\t" - " bf 1f \n\t" - " mov.l %3, @%1 \n\t" - "1: mov r1, r15 \n\t" - : "=&r" (ret) - : "r" (v), "r" (old), "r" (new) - : "memory" , "r0", "r1" , "t"); - - return ret; -} - -static inline int atomic_add_unless(atomic_t *v, int a, int u) -{ - int ret; - unsigned long tmp; - - __asm__ __volatile__ ( - " .align 2 \n\t" - " mova 1f, r0 \n\t" - " nop \n\t" - " mov r15, r1 \n\t" - " mov #-12, r15 \n\t" - " mov.l @%2, %1 \n\t" - " mov %1, %0 \n\t" - " cmp/eq %4, %0 \n\t" - " bt/s 1f \n\t" - " add %3, %1 \n\t" - " mov.l %1, @%2 \n\t" - "1: mov r1, r15 \n\t" - : "=&r" (ret), "=&r" (tmp) - : "r" (v), "r" (a), "r" (u) - : "memory" , "r0", "r1" , "t"); - - return ret != u; -} #endif /* __ASM_SH_ATOMIC_GRB_H */ diff --git a/arch/sh/include/asm/atomic-llsc.h b/arch/sh/include/asm/atomic-llsc.h index b040e1e..4b00b78 100644 --- a/arch/sh/include/asm/atomic-llsc.h +++ b/arch/sh/include/asm/atomic-llsc.h @@ -104,31 +104,4 @@ static inline void atomic_set_mask(unsigned int mask, atomic_t *v) : "t"); } -#define atomic_cmpxchg(v, o, n) (cmpxchg(&((v)->counter), (o), (n))) - -/** - * atomic_add_unless - add unless the number is a given value - * @v: pointer of type atomic_t - * @a: the amount to add to v... - * @u: ...unless v is equal to u. - * - * Atomically adds @a to @v, so long as it was not @u. - * Returns non-zero if @v was not @u, and zero otherwise. - */ -static inline int atomic_add_unless(atomic_t *v, int a, int u) -{ - int c, old; - c = atomic_read(v); - for (;;) { - if (unlikely(c == (u))) - break; - old = atomic_cmpxchg((v), c, c + (a)); - if (likely(old == c)) - break; - c = old; - } - - return c != (u); -} - #endif /* __ASM_SH_ATOMIC_LLSC_H */ diff --git a/arch/sh/include/asm/atomic.h b/arch/sh/include/asm/atomic.h index b16388d..275a448 100644 --- a/arch/sh/include/asm/atomic.h +++ b/arch/sh/include/asm/atomic.h @@ -25,58 +25,43 @@ #endif #define atomic_add_negative(a, v) (atomic_add_return((a), (v)) < 0) +#define atomic_dec_return(v) atomic_sub_return(1, (v)) +#define atomic_inc_return(v) atomic_add_return(1, (v)) +#define atomic_inc_and_test(v) (atomic_inc_return(v) == 0) +#define atomic_sub_and_test(i,v) (atomic_sub_return((i), (v)) == 0) +#define atomic_dec_and_test(v) (atomic_sub_return(1, (v)) == 0) +#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) -#define atomic_dec_return(v) atomic_sub_return(1,(v)) -#define atomic_inc_return(v) atomic_add_return(1,(v)) +#define atomic_inc(v) atomic_add(1, (v)) +#define atomic_dec(v) atomic_sub(1, (v)) -/* - * atomic_inc_and_test - increment and test +#define atomic_xchg(v, new) (xchg(&((v)->counter), new)) +#define atomic_cmpxchg(v, o, n) (cmpxchg(&((v)->counter), (o), (n))) + +/** + * atomic_add_unless - add unless the number is a given value * @v: pointer of type atomic_t + * @a: the amount to add to v... + * @u: ...unless v is equal to u. * - * Atomically increments @v by 1 - * and returns true if the result is zero, or false for all - * other cases. + * Atomically adds @a to @v, so long as it was not @u. + * Returns non-zero if @v was not @u, and zero otherwise. */ -#define atomic_inc_and_test(v) (atomic_inc_return(v) == 0) - -#define atomic_sub_and_test(i,v) (atomic_sub_return((i), (v)) == 0) -#define atomic_dec_and_test(v) (atomic_sub_return(1, (v)) == 0) - -#define atomic_inc(v) atomic_add(1,(v)) -#define atomic_dec(v) atomic_sub(1,(v)) - -#if !defined(CONFIG_GUSA_RB) && !defined(CONFIG_CPU_SH4A) -static inline int atomic_cmpxchg(atomic_t *v, int old, int new) -{ - int ret; - unsigned long flags; - - local_irq_save(flags); - ret = v->counter; - if (likely(ret == old)) - v->counter = new; - local_irq_restore(flags); - - return ret; -} - static inline int atomic_add_unless(atomic_t *v, int a, int u) { - int ret; - unsigned long flags; - - local_irq_save(flags); - ret = v->counter; - if (ret != u) - v->counter += a; - local_irq_restore(flags); - - return ret != u; + int c, old; + c = atomic_read(v); + for (;;) { + if (unlikely(c == (u))) + break; + old = atomic_cmpxchg((v), c, c + (a)); + if (likely(old == c)) + break; + c = old; + } + + return c != (u); } -#endif /* !CONFIG_GUSA_RB && !CONFIG_CPU_SH4A */ - -#define atomic_xchg(v, new) (xchg(&((v)->counter), new)) -#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0) #define smp_mb__before_atomic_dec() smp_mb() #define smp_mb__after_atomic_dec() smp_mb() -- cgit v0.10.2 From 53e6d8e0060fe2bb9b11238f8250fdfbb0589425 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 13:37:04 +0900 Subject: sh: mach-se: Convert SE7343 FPGA to dynamic IRQ allocation. This gets rid of the arbitrary set of vectors used by the SE7722 FPGA interrupt controller and switches over to a completely dynamic set. No assumptions regarding a contiguous range are made, and the platform resources themselves need to be filled in lazily. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-se/7343/irq.c b/arch/sh/boards/mach-se/7343/irq.c index 051c29d..c60fd13 100644 --- a/arch/sh/boards/mach-se/7343/irq.c +++ b/arch/sh/boards/mach-se/7343/irq.c @@ -16,15 +16,17 @@ #include #include +unsigned int se7343_fpga_irq[SE7343_FPGA_IRQ_NR] = { 0, }; + static void disable_se7343_irq(unsigned int irq) { - unsigned int bit = irq - SE7343_FPGA_IRQ_BASE; + unsigned int bit = (unsigned int)get_irq_chip_data(irq); ctrl_outw(ctrl_inw(PA_CPLD_IMSK) | 1 << bit, PA_CPLD_IMSK); } static void enable_se7343_irq(unsigned int irq) { - unsigned int bit = irq - SE7343_FPGA_IRQ_BASE; + unsigned int bit = (unsigned int)get_irq_chip_data(irq); ctrl_outw(ctrl_inw(PA_CPLD_IMSK) & ~(1 << bit), PA_CPLD_IMSK); } @@ -38,18 +40,15 @@ static struct irq_chip se7343_irq_chip __read_mostly = { static void se7343_irq_demux(unsigned int irq, struct irq_desc *desc) { unsigned short intv = ctrl_inw(PA_CPLD_ST); - struct irq_desc *ext_desc; - unsigned int ext_irq = SE7343_FPGA_IRQ_BASE; + unsigned int ext_irq = 0; intv &= (1 << SE7343_FPGA_IRQ_NR) - 1; - while (intv) { - if (intv & 1) { - ext_desc = irq_desc + ext_irq; - handle_level_irq(ext_irq, ext_desc); - } - intv >>= 1; - ext_irq++; + for (; intv; intv >>= 1, ext_irq++) { + if (!(intv & 1)) + continue; + + generic_handle_irq(se7343_fpga_irq[ext_irq]); } } @@ -58,16 +57,24 @@ static void se7343_irq_demux(unsigned int irq, struct irq_desc *desc) */ void __init init_7343se_IRQ(void) { - int i; + int i, irq; ctrl_outw(0, PA_CPLD_IMSK); /* disable all irqs */ ctrl_outw(0x2000, 0xb03fffec); /* mrshpc irq enable */ - for (i = 0; i < SE7343_FPGA_IRQ_NR; i++) - set_irq_chip_and_handler_name(SE7343_FPGA_IRQ_BASE + i, + for (i = 0; i < SE7343_FPGA_IRQ_NR; i++) { + irq = create_irq(); + if (irq < 0) + return; + se7343_fpga_irq[i] = irq; + + set_irq_chip_and_handler_name(se7343_fpga_irq[i], &se7343_irq_chip, handle_level_irq, "level"); + set_irq_chip_data(se7343_fpga_irq[i], (void *)i); + } + set_irq_chained_handler(IRQ0_IRQ, se7343_irq_demux); set_irq_type(IRQ0_IRQ, IRQ_TYPE_LEVEL_LOW); set_irq_chained_handler(IRQ1_IRQ, se7343_irq_demux); diff --git a/arch/sh/boards/mach-se/7343/setup.c b/arch/sh/boards/mach-se/7343/setup.c index 4de56f3..292cc47 100644 --- a/arch/sh/boards/mach-se/7343/setup.c +++ b/arch/sh/boards/mach-se/7343/setup.c @@ -82,7 +82,6 @@ static struct plat_serial8250_port serial_platform_data[] = { .mapbase = 0x16000000, .regshift = 1, .flags = ST16C2550C_FLAGS, - .irq = UARTA_IRQ, .uartclk = 7372800, }, [1] = { @@ -90,7 +89,6 @@ static struct plat_serial8250_port serial_platform_data[] = { .mapbase = 0x17000000, .regshift = 1, .flags = ST16C2550C_FLAGS, - .irq = UARTB_IRQ, .uartclk = 7372800, }, { }, @@ -121,7 +119,7 @@ static struct resource usb_resources[] = { .flags = IORESOURCE_MEM, }, [2] = { - .start = USB_IRQ, + /* Filled in later */ .flags = IORESOURCE_IRQ, }, }; @@ -138,8 +136,8 @@ static struct isp116x_platform_data usb_platform_data = { static struct platform_device usb_device = { .name = "isp116x-hcd", .id = -1, - .num_resources = ARRAY_SIZE(usb_resources), - .resource = usb_resources, + .num_resources = ARRAY_SIZE(usb_resources), + .resource = usb_resources, .dev = { .platform_data = &usb_platform_data, }, @@ -155,6 +153,13 @@ static struct platform_device *sh7343se_platform_devices[] __initdata = { static int __init sh7343se_devices_setup(void) { + /* Wire-up dynamic vectors */ + serial_platform_data[0].irq = se7343_fpga_irq[SE7343_FPGA_IRQ_UARTA]; + serial_platform_data[1].irq = se7343_fpga_irq[SE7343_FPGA_IRQ_UARTB]; + + usb_resources[2].start = usb_resources[2].end = + se7343_fpga_irq[SE7343_FPGA_IRQ_USB]; + return platform_add_devices(sh7343se_platform_devices, ARRAY_SIZE(sh7343se_platform_devices)); } @@ -179,6 +184,5 @@ static void __init sh7343se_setup(char **cmdline_p) static struct sh_machine_vector mv_7343se __initmv = { .mv_name = "SolutionEngine 7343", .mv_setup = sh7343se_setup, - .mv_nr_irqs = SE7343_FPGA_IRQ_BASE + SE7343_FPGA_IRQ_NR, .mv_init_irq = init_7343se_IRQ, }; diff --git a/arch/sh/include/mach-se/mach/se7343.h b/arch/sh/include/mach-se/mach/se7343.h index 749914b..8d8170d 100644 --- a/arch/sh/include/mach-se/mach/se7343.h +++ b/arch/sh/include/mach-se/mach/se7343.h @@ -94,26 +94,26 @@ #define PORT_DRVCR 0xA4050180 -#define PORT_PADR 0xA4050120 -#define PORT_PBDR 0xA4050122 -#define PORT_PCDR 0xA4050124 -#define PORT_PDDR 0xA4050126 -#define PORT_PEDR 0xA4050128 -#define PORT_PFDR 0xA405012A -#define PORT_PGDR 0xA405012C -#define PORT_PHDR 0xA405012E -#define PORT_PJDR 0xA4050130 -#define PORT_PKDR 0xA4050132 -#define PORT_PLDR 0xA4050134 -#define PORT_PMDR 0xA4050136 -#define PORT_PNDR 0xA4050138 -#define PORT_PQDR 0xA405013A -#define PORT_PRDR 0xA405013C -#define PORT_PTDR 0xA4050160 -#define PORT_PUDR 0xA4050162 -#define PORT_PVDR 0xA4050164 -#define PORT_PWDR 0xA4050166 -#define PORT_PYDR 0xA4050168 +#define PORT_PADR 0xA4050120 +#define PORT_PBDR 0xA4050122 +#define PORT_PCDR 0xA4050124 +#define PORT_PDDR 0xA4050126 +#define PORT_PEDR 0xA4050128 +#define PORT_PFDR 0xA405012A +#define PORT_PGDR 0xA405012C +#define PORT_PHDR 0xA405012E +#define PORT_PJDR 0xA4050130 +#define PORT_PKDR 0xA4050132 +#define PORT_PLDR 0xA4050134 +#define PORT_PMDR 0xA4050136 +#define PORT_PNDR 0xA4050138 +#define PORT_PQDR 0xA405013A +#define PORT_PRDR 0xA405013C +#define PORT_PTDR 0xA4050160 +#define PORT_PUDR 0xA4050162 +#define PORT_PVDR 0xA4050164 +#define PORT_PWDR 0xA4050166 +#define PORT_PYDR 0xA4050168 #define FPGA_IN 0xb1400000 #define FPGA_OUT 0xb1400002 @@ -133,18 +133,10 @@ #define SE7343_FPGA_IRQ_UARTB 11 #define SE7343_FPGA_IRQ_NR 12 -#define SE7343_FPGA_IRQ_BASE 120 - -#define MRSHPC_IRQ3 (SE7343_FPGA_IRQ_BASE + SE7343_FPGA_IRQ_MRSHPC3) -#define MRSHPC_IRQ2 (SE7343_FPGA_IRQ_BASE + SE7343_FPGA_IRQ_MRSHPC2) -#define MRSHPC_IRQ1 (SE7343_FPGA_IRQ_BASE + SE7343_FPGA_IRQ_MRSHPC1) -#define MRSHPC_IRQ0 (SE7343_FPGA_IRQ_BASE + SE7343_FPGA_IRQ_MRSHPC0) -#define SMC_IRQ (SE7343_FPGA_IRQ_BASE + SE7343_FPGA_IRQ_SMC) -#define USB_IRQ (SE7343_FPGA_IRQ_BASE + SE7343_FPGA_IRQ_USB) -#define UARTA_IRQ (SE7343_FPGA_IRQ_BASE + SE7343_FPGA_IRQ_UARTA) -#define UARTB_IRQ (SE7343_FPGA_IRQ_BASE + SE7343_FPGA_IRQ_UARTB) /* arch/sh/boards/se/7343/irq.c */ +extern unsigned int se7343_fpga_irq[]; + void init_7343se_IRQ(void); #endif /* __ASM_SH_HITACHI_SE7343_H */ -- cgit v0.10.2 From ee2760ea58d81fc00bcc2137232ed9bc28202aec Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 13:48:27 +0900 Subject: sh: default to sparseirq. As SH has a very sparse IRQ map by default, all new CPUs and boards benefit from using sparseirq by default. Despite this, there are still a few stragglers (mostly due to using a fixed IRQ range for their FPGA IRQ mappings), and these still need to be converted over one by one. As these are now in the minority, and we do not want to encourage this sort of brain-damage in newer board ports, we force sparseirq on. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 2121fbb..ae6c736 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -726,8 +726,9 @@ config GUSA_RB disabling interrupts around the atomic sequence. config SPARSE_IRQ - bool "Support sparse irq numbering" - depends on EXPERIMENTAL + def_bool y + depends on SUPERH32 && !SH_DREAMCAST && !SH_HIGHLANDER && \ + !SH_RTS7751R2D && !HD64461 && !SH_7724_SOLUTION_ENGINE help This enables support for sparse irqs. This is useful in general as most CPUs have a fairly sparse array of IRQ vectors, which -- cgit v0.10.2 From 191d0d24b632eb69767705acded5cbf7449ad457 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 14:50:43 +0900 Subject: sh: Tidy up the sh bios VBR handling. This moves the VBR handling out of the main trap handling code and in to the sh-bios helper code. A couple of accessors are added in order to permit other kernel code to get at the VBR value for state save/restore paths. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/sh_bios.h b/arch/sh/include/asm/sh_bios.h index d9c96d7..d8a67a1 100644 --- a/arch/sh/include/asm/sh_bios.h +++ b/arch/sh/include/asm/sh_bios.h @@ -1,13 +1,13 @@ #ifndef __ASM_SH_BIOS_H #define __ASM_SH_BIOS_H +#ifdef CONFIG_SH_STANDARD_BIOS + /* * Copyright (C) 2000 Greg Banks, Mitch Davis * C API to interface to the standard LinuxSH BIOS * usually from within the early stages of kernel boot. */ - - extern void sh_bios_console_write(const char *buf, unsigned int len); extern void sh_bios_char_out(char ch); extern void sh_bios_gdb_detach(void); @@ -15,4 +15,14 @@ extern void sh_bios_gdb_detach(void); extern void sh_bios_get_node_addr(unsigned char *node_addr); extern void sh_bios_shutdown(unsigned int how); +extern void sh_bios_vbr_init(void); +extern void sh_bios_vbr_reload(void); + +#else + +static inline void sh_bios_vbr_init(void) { } +static inline void sh_bios_vbr_reload(void) { } + +#endif /* CONFIG_SH_STANDARD_BIOS */ + #endif /* __ASM_SH_BIOS_H */ diff --git a/arch/sh/kernel/machine_kexec.c b/arch/sh/kernel/machine_kexec.c index 76f2802..f52d8ed 100644 --- a/arch/sh/kernel/machine_kexec.c +++ b/arch/sh/kernel/machine_kexec.c @@ -21,6 +21,7 @@ #include #include #include +#include typedef void (*relocate_new_kernel_t)(unsigned long indirection_page, unsigned long reboot_code_buffer, @@ -28,7 +29,6 @@ typedef void (*relocate_new_kernel_t)(unsigned long indirection_page, extern const unsigned char relocate_new_kernel[]; extern const unsigned int relocate_new_kernel_size; -extern void *gdb_vbr_vector; extern void *vbr_base; void machine_shutdown(void) @@ -117,11 +117,7 @@ void machine_kexec(struct kimage *image) kexec_info(image); flush_cache_all(); -#if defined(CONFIG_SH_STANDARD_BIOS) - asm volatile("ldc %0, vbr" : - : "r" (((unsigned long) gdb_vbr_vector) - 0x100) - : "memory"); -#endif + sh_bios_vbr_reload(); /* now call it */ rnk = (relocate_new_kernel_t) reboot_code_buffer; diff --git a/arch/sh/kernel/sh_bios.c b/arch/sh/kernel/sh_bios.c index c852f78..2a5f2e0 100644 --- a/arch/sh/kernel/sh_bios.c +++ b/arch/sh/kernel/sh_bios.c @@ -55,3 +55,40 @@ void sh_bios_shutdown(unsigned int how) { sh_bios_call(BIOS_CALL_SHUTDOWN, how, 0, 0, 0); } + +void *gdb_vbr_vector = NULL; + +/* + * Read the old value of the VBR register to initialise the vector + * through which debug and BIOS traps are delegated by the Linux trap + * handler. + */ +void sh_bios_vbr_init(void) +{ + unsigned long vbr; + + if (unlikely(gdb_vbr_vector)) + return; + + __asm__ __volatile__ ("stc vbr, %0" : "=r" (vbr)); + + gdb_vbr_vector = (void *)(vbr + 0x100); + printk(KERN_NOTICE "Setting GDB trap vector to %p\n", gdb_vbr_vector); +} + +/** + * sh_bios_vbr_reload - Re-load the system VBR from the BIOS vector. + * + * This can be used by save/restore code to reinitialize the system VBR + * from the fixed BIOS VBR. A no-op if no BIOS VBR is known. + */ +void sh_bios_vbr_reload(void) +{ + if (gdb_vbr_vector) + __asm__ __volatile__ ( + "ldc %0, vbr" + : + : "r" (((unsigned long) gdb_vbr_vector) - 0x100) + : "memory" + ); +} diff --git a/arch/sh/kernel/traps_32.c b/arch/sh/kernel/traps_32.c index 86639be..efcbdfe 100644 --- a/arch/sh/kernel/traps_32.c +++ b/arch/sh/kernel/traps_32.c @@ -31,6 +31,7 @@ #include #include #include +#include #ifdef CONFIG_CPU_SH2 # define TRAP_RESERVED_INST 4 @@ -876,35 +877,10 @@ asmlinkage void do_exception_error(unsigned long r4, unsigned long r5, die_if_kernel("exception", regs, ex); } -#if defined(CONFIG_SH_STANDARD_BIOS) -void *gdb_vbr_vector; - -static inline void __init gdb_vbr_init(void) -{ - register unsigned long vbr; - - /* - * Read the old value of the VBR register to initialise - * the vector through which debug and BIOS traps are - * delegated by the Linux trap handler. - */ - asm volatile("stc vbr, %0" : "=r" (vbr)); - - gdb_vbr_vector = (void *)(vbr + 0x100); - printk("Setting GDB trap vector to 0x%08lx\n", - (unsigned long)gdb_vbr_vector); -} -#endif - void __cpuinit per_cpu_trap_init(void) { extern void *vbr_base; -#ifdef CONFIG_SH_STANDARD_BIOS - if (raw_smp_processor_id() == 0) - gdb_vbr_init(); -#endif - /* NOTE: The VBR value should be at P1 (or P2, virtural "fixed" address space). It's definitely should not in physical address. */ @@ -959,6 +935,9 @@ void __init trap_init(void) set_exception_table_vec(TRAP_UBC, break_point_trap); #endif + /* Save off the BIOS VBR, if there is one */ + sh_bios_vbr_init(); + /* Setup VBR for boot cpu */ per_cpu_trap_init(); } -- cgit v0.10.2 From 65fedbbef8404be5e1dbd18bdad252616b49a24b Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 15:22:26 +0900 Subject: sh64: Fix up early serial fixmap. This was conditionalized on CONFIG_EARLY_PRINTK, which has subsequently gone away. Now that the serial driver always supports the early console, make sure we always establish the mapping. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/head_64.S b/arch/sh/kernel/head_64.S index 3ea7658..defd851 100644 --- a/arch/sh/kernel/head_64.S +++ b/arch/sh/kernel/head_64.S @@ -220,7 +220,6 @@ clear_DTLB: add.l r22, r63, r22 /* Sign extend */ putcfg r21, 0, r22 /* Set MMUDR[0].PTEH */ -#ifdef CONFIG_EARLY_PRINTK /* * Setup a DTLB translation for SCIF phys. */ @@ -231,7 +230,6 @@ clear_DTLB: movi 0xfa03, r22 /* 0xfa030000, fixed SCIF virt */ shori 0x0003, r22 putcfg r21, 0, r22 /* PTEH last */ -#endif /* * Set cache behaviours. -- cgit v0.10.2 From b9303a79567d4a45b015dff7e71dd24923332d8d Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 15:26:11 +0900 Subject: sh: Kill off more unused sh_bios callbacks. sh_bios_char_out() is not used by anything in-tree these days, so just get rid of it. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/sh_bios.h b/arch/sh/include/asm/sh_bios.h index d8a67a1..95714c2 100644 --- a/arch/sh/include/asm/sh_bios.h +++ b/arch/sh/include/asm/sh_bios.h @@ -9,7 +9,6 @@ * usually from within the early stages of kernel boot. */ extern void sh_bios_console_write(const char *buf, unsigned int len); -extern void sh_bios_char_out(char ch); extern void sh_bios_gdb_detach(void); extern void sh_bios_get_node_addr(unsigned char *node_addr); diff --git a/arch/sh/kernel/sh_bios.c b/arch/sh/kernel/sh_bios.c index 2a5f2e0..2a9c6d5 100644 --- a/arch/sh/kernel/sh_bios.c +++ b/arch/sh/kernel/sh_bios.c @@ -34,11 +34,6 @@ void sh_bios_console_write(const char *buf, unsigned int len) sh_bios_call(BIOS_CALL_CONSOLE_WRITE, (long)buf, (long)len, 0, 0); } -void sh_bios_char_out(char ch) -{ - sh_bios_call(BIOS_CALL_CHAR_OUT, ch, 0, 0, 0); -} - void sh_bios_gdb_detach(void) { sh_bios_call(BIOS_CALL_GDB_DETACH, 0, 0, 0, 0); -- cgit v0.10.2 From 776258df925acd0563f471ee4b3f19bbffb3c04f Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 15:31:20 +0900 Subject: sh: Consolidate the sh_bios earlyprintk code. Now that the sh-sci earlyprintk is taken care of by the sh-sci driver directly, there's no longer any reason for having a split-out early_printk framework. sh_bios is the only other thing that uses it, so we just migrate the leftovers in to there. As it's possible to have multiple early_param()'s for the same string, there's not much point in having this split out anymore anyways, particularly since the sh_bios dependencies are still special-cased within sh-sci itself. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/setup.h b/arch/sh/include/asm/setup.h index ce37435..4758325 100644 --- a/arch/sh/include/asm/setup.h +++ b/arch/sh/include/asm/setup.h @@ -18,7 +18,6 @@ /* ... */ #define COMMAND_LINE ((char *) (PARAM+0x100)) -int setup_early_printk(char *); void sh_mv_setup(void); #endif /* __KERNEL__ */ diff --git a/arch/sh/kernel/Makefile b/arch/sh/kernel/Makefile index 0d587da..5bec10c 100644 --- a/arch/sh/kernel/Makefile +++ b/arch/sh/kernel/Makefile @@ -22,7 +22,7 @@ obj-y := debugtraps.o dma-nommu.o dumpstack.o \ obj-y += cpu/ obj-$(CONFIG_VSYSCALL) += vsyscall/ obj-$(CONFIG_SMP) += smp.o -obj-$(CONFIG_SH_STANDARD_BIOS) += sh_bios.o early_printk.o +obj-$(CONFIG_SH_STANDARD_BIOS) += sh_bios.o obj-$(CONFIG_KGDB) += kgdb.o obj-$(CONFIG_SH_CPU_FREQ) += cpufreq.o obj-$(CONFIG_MODULES) += sh_ksyms_$(BITS).o module.o diff --git a/arch/sh/kernel/early_printk.c b/arch/sh/kernel/early_printk.c deleted file mode 100644 index f8bb50c..0000000 --- a/arch/sh/kernel/early_printk.c +++ /dev/null @@ -1,85 +0,0 @@ -/* - * arch/sh/kernel/early_printk.c - * - * Copyright (C) 1999, 2000 Niibe Yutaka - * Copyright (C) 2002 M. R. Brown - * Copyright (C) 2004 - 2007 Paul Mundt - * - * This file is subject to the terms and conditions of the GNU General Public - * License. See the file "COPYING" in the main directory of this archive - * for more details. - */ -#include -#include -#include -#include -#include - -#include - -/* - * Print a string through the BIOS - */ -static void sh_console_write(struct console *co, const char *s, - unsigned count) -{ - sh_bios_console_write(s, count); -} - -/* - * Setup initial baud/bits/parity. We do two things here: - * - construct a cflag setting for the first rs_open() - * - initialize the serial port - * Return non-zero if we didn't find a serial port. - */ -static int __init sh_console_setup(struct console *co, char *options) -{ - int cflag = CREAD | HUPCL | CLOCAL; - - /* - * Now construct a cflag setting. - * TODO: this is a totally bogus cflag, as we have - * no idea what serial settings the BIOS is using, or - * even if its using the serial port at all. - */ - cflag |= B115200 | CS8 | /*no parity*/0; - - co->cflag = cflag; - - return 0; -} - -static struct console bios_console = { - .name = "bios", - .write = sh_console_write, - .setup = sh_console_setup, - .flags = CON_PRINTBUFFER, - .index = -1, -}; - -static struct console *early_console; - -static int __init setup_early_printk(char *buf) -{ - int keep_early = 0; - - if (!buf) - return 0; - - if (strstr(buf, "keep")) - keep_early = 1; - - if (!strncmp(buf, "bios", 4)) - early_console = &bios_console; - - if (likely(early_console)) { - if (keep_early) - early_console->flags &= ~CON_BOOT; - else - early_console->flags |= CON_BOOT; - register_console(early_console); - } - - return 0; -} -early_param("earlyprintk", setup_early_printk); diff --git a/arch/sh/kernel/sh_bios.c b/arch/sh/kernel/sh_bios.c index 2a9c6d5..29cd252 100644 --- a/arch/sh/kernel/sh_bios.c +++ b/arch/sh/kernel/sh_bios.c @@ -1,17 +1,26 @@ /* - * linux/arch/sh/kernel/sh_bios.c * C interface for trapping into the standard LinuxSH BIOS. * * Copyright (C) 2000 Greg Banks, Mitch Davis + * Copyright (C) 1999, 2000 Niibe Yutaka + * Copyright (C) 2002 M. R. Brown + * Copyright (C) 2004 - 2010 Paul Mundt * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. */ #include +#include +#include +#include +#include +#include #include #define BIOS_CALL_CONSOLE_WRITE 0 #define BIOS_CALL_ETH_NODE_ADDR 10 #define BIOS_CALL_SHUTDOWN 11 -#define BIOS_CALL_CHAR_OUT 0x1f /* TODO: hack */ #define BIOS_CALL_GDB_DETACH 0xff static inline long sh_bios_call(long func, long arg0, long arg1, long arg2, @@ -87,3 +96,70 @@ void sh_bios_vbr_reload(void) : "memory" ); } + +/* + * Print a string through the BIOS + */ +static void sh_console_write(struct console *co, const char *s, + unsigned count) +{ + sh_bios_console_write(s, count); +} + +/* + * Setup initial baud/bits/parity. We do two things here: + * - construct a cflag setting for the first rs_open() + * - initialize the serial port + * Return non-zero if we didn't find a serial port. + */ +static int __init sh_console_setup(struct console *co, char *options) +{ + int cflag = CREAD | HUPCL | CLOCAL; + + /* + * Now construct a cflag setting. + * TODO: this is a totally bogus cflag, as we have + * no idea what serial settings the BIOS is using, or + * even if its using the serial port at all. + */ + cflag |= B115200 | CS8 | /*no parity*/0; + + co->cflag = cflag; + + return 0; +} + +static struct console bios_console = { + .name = "bios", + .write = sh_console_write, + .setup = sh_console_setup, + .flags = CON_PRINTBUFFER, + .index = -1, +}; + +static struct console *early_console; + +static int __init setup_early_printk(char *buf) +{ + int keep_early = 0; + + if (!buf) + return 0; + + if (strstr(buf, "keep")) + keep_early = 1; + + if (!strncmp(buf, "bios", 4)) + early_console = &bios_console; + + if (likely(early_console)) { + if (keep_early) + early_console->flags &= ~CON_BOOT; + else + early_console->flags |= CON_BOOT; + register_console(early_console); + } + + return 0; +} +early_param("earlyprintk", setup_early_printk); -- cgit v0.10.2 From a99eae5417a09e0be66bf574a9a79a2a7388c967 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 16:12:25 +0900 Subject: sh: Split out the unaligned counters and user bits. This splits out the unaligned access counters and userspace bits in to their own generic interface, which will allow them to be wired up on sh64 too. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/alignment.h b/arch/sh/include/asm/alignment.h new file mode 100644 index 0000000..b12efec --- /dev/null +++ b/arch/sh/include/asm/alignment.h @@ -0,0 +1,21 @@ +#ifndef __ASM_SH_ALIGNMENT_H +#define __ASM_SH_ALIGNMENT_H + +#include + +extern void inc_unaligned_byte_access(void); +extern void inc_unaligned_word_access(void); +extern void inc_unaligned_dword_access(void); +extern void inc_unaligned_multi_access(void); +extern void inc_unaligned_user_access(void); +extern void inc_unaligned_kernel_access(void); + +#define UM_WARN (1 << 0) +#define UM_FIXUP (1 << 1) +#define UM_SIGNAL (1 << 2) + +extern unsigned int unaligned_user_action(void); + +extern void unaligned_fixups_notify(struct task_struct *, insn_size_t, struct pt_regs *); + +#endif /* __ASM_SH_ALIGNMENT_H */ diff --git a/arch/sh/kernel/traps_32.c b/arch/sh/kernel/traps_32.c index efcbdfe..204def6 100644 --- a/arch/sh/kernel/traps_32.c +++ b/arch/sh/kernel/traps_32.c @@ -24,11 +24,10 @@ #include #include #include -#include -#include #include +#include #include -#include +#include #include #include #include @@ -48,73 +47,6 @@ #define TRAP_ILLEGAL_SLOT_INST 13 #endif -static unsigned long se_user; -static unsigned long se_sys; -static unsigned long se_half; -static unsigned long se_word; -static unsigned long se_dword; -static unsigned long se_multi; -/* bitfield: 1: warn 2: fixup 4: signal -> combinations 2|4 && 1|2|4 are not - valid! */ -static int se_usermode = 3; -/* 0: no warning 1: print a warning message, disabled by default */ -static int se_kernmode_warn; - -#ifdef CONFIG_PROC_FS -static const char *se_usermode_action[] = { - "ignored", - "warn", - "fixup", - "fixup+warn", - "signal", - "signal+warn" -}; - -static int alignment_proc_show(struct seq_file *m, void *v) -{ - seq_printf(m, "User:\t\t%lu\n", se_user); - seq_printf(m, "System:\t\t%lu\n", se_sys); - seq_printf(m, "Half:\t\t%lu\n", se_half); - seq_printf(m, "Word:\t\t%lu\n", se_word); - seq_printf(m, "DWord:\t\t%lu\n", se_dword); - seq_printf(m, "Multi:\t\t%lu\n", se_multi); - seq_printf(m, "User faults:\t%i (%s)\n", se_usermode, - se_usermode_action[se_usermode]); - seq_printf(m, "Kernel faults:\t%i (fixup%s)\n", se_kernmode_warn, - se_kernmode_warn ? "+warn" : ""); - return 0; -} - -static int alignment_proc_open(struct inode *inode, struct file *file) -{ - return single_open(file, alignment_proc_show, NULL); -} - -static ssize_t alignment_proc_write(struct file *file, - const char __user *buffer, size_t count, loff_t *pos) -{ - int *data = PDE(file->f_path.dentry->d_inode)->data; - char mode; - - if (count > 0) { - if (get_user(mode, buffer)) - return -EFAULT; - if (mode >= '0' && mode <= '5') - *data = mode - '0'; - } - return count; -} - -static const struct file_operations alignment_proc_fops = { - .owner = THIS_MODULE, - .open = alignment_proc_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, - .write = alignment_proc_write, -}; -#endif - static void dump_mem(const char *str, unsigned long bottom, unsigned long top) { unsigned long p; @@ -266,10 +198,10 @@ static int handle_unaligned_ins(insn_size_t instruction, struct pt_regs *regs, count = 1<<(instruction&3); switch (count) { - case 1: se_half += 1; break; - case 2: se_word += 1; break; - case 4: se_dword += 1; break; - case 8: se_multi += 1; break; /* ??? */ + case 1: inc_unaligned_byte_access(); break; + case 2: inc_unaligned_word_access(); break; + case 4: inc_unaligned_dword_access(); break; + case 8: inc_unaligned_multi_access(); break; } ret = -EFAULT; @@ -453,18 +385,8 @@ int handle_unaligned_access(insn_size_t instruction, struct pt_regs *regs, rm = regs->regs[index]; /* shout about fixups */ - if (!expected) { - if (user_mode(regs) && (se_usermode & 1) && printk_ratelimit()) - pr_notice("Fixing up unaligned userspace access " - "in \"%s\" pid=%d pc=0x%p ins=0x%04hx\n", - current->comm, task_pid_nr(current), - (void *)regs->pc, instruction); - else if (se_kernmode_warn && printk_ratelimit()) - pr_notice("Fixing up unaligned kernel access " - "in \"%s\" pid=%d pc=0x%p ins=0x%04hx\n", - current->comm, task_pid_nr(current), - (void *)regs->pc, instruction); - } + if (!expected) + unaligned_fixups_notify(current, instruction, regs); ret = -EFAULT; switch (instruction&0xF000) { @@ -617,10 +539,10 @@ asmlinkage void do_address_error(struct pt_regs *regs, if (user_mode(regs)) { int si_code = BUS_ADRERR; + unsigned int user_action; local_irq_enable(); - - se_user += 1; + inc_unaligned_user_access(); set_fs(USER_DS); if (copy_from_user(&instruction, (insn_size_t *)(regs->pc & ~1), @@ -631,16 +553,12 @@ asmlinkage void do_address_error(struct pt_regs *regs, set_fs(oldfs); /* shout about userspace fixups */ - if (se_usermode & 1) - printk(KERN_NOTICE "Unaligned userspace access " - "in \"%s\" pid=%d pc=0x%p ins=0x%04hx\n", - current->comm, current->pid, (void *)regs->pc, - instruction); + unaligned_fixups_notify(current, instruction, regs); - if (se_usermode & 2) + user_action = unaligned_user_action(); + if (user_action & UM_FIXUP) goto fixup; - - if (se_usermode & 4) + if (user_action & UM_SIGNAL) goto uspace_segv; else { /* ignore */ @@ -660,7 +578,7 @@ fixup: &user_mem_access, 0); set_fs(oldfs); - if (tmp==0) + if (tmp == 0) return; /* sorted */ uspace_segv: printk(KERN_NOTICE "Sending SIGBUS to \"%s\" due to unaligned " @@ -673,7 +591,7 @@ uspace_segv: info.si_addr = (void __user *)address; force_sig_info(SIGBUS, &info, current); } else { - se_sys += 1; + inc_unaligned_kernel_access(); if (regs->pc & 1) die("unaligned program counter", regs, error_code); @@ -688,11 +606,7 @@ uspace_segv: die("insn faulting in do_address_error", regs, 0); } - if (se_kernmode_warn) - printk(KERN_NOTICE "Unaligned kernel access " - "on behalf of \"%s\" pid=%d pc=0x%p ins=0x%04hx\n", - current->comm, current->pid, (void *)regs->pc, - instruction); + unaligned_fixups_notify(current, instruction, regs); handle_unaligned_access(instruction, regs, &user_mem_access, 0); @@ -964,34 +878,3 @@ void dump_stack(void) show_stack(NULL, NULL); } EXPORT_SYMBOL(dump_stack); - -#ifdef CONFIG_PROC_FS -/* - * This needs to be done after sysctl_init, otherwise sys/ will be - * overwritten. Actually, this shouldn't be in sys/ at all since - * it isn't a sysctl, and it doesn't contain sysctl information. - * We now locate it in /proc/cpu/alignment instead. - */ -static int __init alignment_init(void) -{ - struct proc_dir_entry *dir, *res; - - dir = proc_mkdir("cpu", NULL); - if (!dir) - return -ENOMEM; - - res = proc_create_data("alignment", S_IWUSR | S_IRUGO, dir, - &alignment_proc_fops, &se_usermode); - if (!res) - return -ENOMEM; - - res = proc_create_data("kernel_alignment", S_IWUSR | S_IRUGO, dir, - &alignment_proc_fops, &se_kernmode_warn); - if (!res) - return -ENOMEM; - - return 0; -} - -fs_initcall(alignment_init); -#endif diff --git a/arch/sh/mm/Makefile b/arch/sh/mm/Makefile index dd5010c..9fa11d6 100644 --- a/arch/sh/mm/Makefile +++ b/arch/sh/mm/Makefile @@ -2,7 +2,7 @@ # Makefile for the Linux SuperH-specific parts of the memory manager. # -obj-y := cache.o init.o consistent.o mmap.o +obj-y := alignment.o cache.o init.o consistent.o mmap.o cacheops-$(CONFIG_CPU_SH2) := cache-sh2.o cacheops-$(CONFIG_CPU_SH2A) := cache-sh2a.o diff --git a/arch/sh/mm/alignment.c b/arch/sh/mm/alignment.c new file mode 100644 index 0000000..e615151 --- /dev/null +++ b/arch/sh/mm/alignment.c @@ -0,0 +1,159 @@ +/* + * Alignment access counters and corresponding user-space interfaces. + * + * Copyright (C) 2009 ST Microelectronics + * Copyright (C) 2009 - 2010 Paul Mundt + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ +#include +#include +#include +#include +#include +#include + +static unsigned long se_user; +static unsigned long se_sys; +static unsigned long se_half; +static unsigned long se_word; +static unsigned long se_dword; +static unsigned long se_multi; +/* bitfield: 1: warn 2: fixup 4: signal -> combinations 2|4 && 1|2|4 are not + valid! */ +static int se_usermode = UM_WARN | UM_FIXUP; +/* 0: no warning 1: print a warning message, disabled by default */ +static int se_kernmode_warn; + +void inc_unaligned_byte_access(void) +{ + se_half++; +} + +void inc_unaligned_word_access(void) +{ + se_word++; +} + +void inc_unaligned_dword_access(void) +{ + se_dword++; +} + +void inc_unaligned_multi_access(void) +{ + se_multi++; +} + +void inc_unaligned_user_access(void) +{ + se_user++; +} + +void inc_unaligned_kernel_access(void) +{ + se_sys++; +} + +unsigned int unaligned_user_action(void) +{ + return se_usermode; +} + +void unaligned_fixups_notify(struct task_struct *tsk, insn_size_t insn, + struct pt_regs *regs) +{ + if (user_mode(regs) && (se_usermode & UM_WARN) && printk_ratelimit()) + pr_notice("Fixing up unaligned userspace access " + "in \"%s\" pid=%d pc=0x%p ins=0x%04hx\n", + tsk->comm, task_pid_nr(tsk), + (void *)regs->pc, insn); + else if (se_kernmode_warn && printk_ratelimit()) + pr_notice("Fixing up unaligned kernel access " + "in \"%s\" pid=%d pc=0x%p ins=0x%04hx\n", + tsk->comm, task_pid_nr(tsk), + (void *)regs->pc, insn); +} + +static const char *se_usermode_action[] = { + "ignored", + "warn", + "fixup", + "fixup+warn", + "signal", + "signal+warn" +}; + +static int alignment_proc_show(struct seq_file *m, void *v) +{ + seq_printf(m, "User:\t\t%lu\n", se_user); + seq_printf(m, "System:\t\t%lu\n", se_sys); + seq_printf(m, "Half:\t\t%lu\n", se_half); + seq_printf(m, "Word:\t\t%lu\n", se_word); + seq_printf(m, "DWord:\t\t%lu\n", se_dword); + seq_printf(m, "Multi:\t\t%lu\n", se_multi); + seq_printf(m, "User faults:\t%i (%s)\n", se_usermode, + se_usermode_action[se_usermode]); + seq_printf(m, "Kernel faults:\t%i (fixup%s)\n", se_kernmode_warn, + se_kernmode_warn ? "+warn" : ""); + return 0; +} + +static int alignment_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, alignment_proc_show, NULL); +} + +static ssize_t alignment_proc_write(struct file *file, + const char __user *buffer, size_t count, loff_t *pos) +{ + int *data = PDE(file->f_path.dentry->d_inode)->data; + char mode; + + if (count > 0) { + if (get_user(mode, buffer)) + return -EFAULT; + if (mode >= '0' && mode <= '5') + *data = mode - '0'; + } + return count; +} + +static const struct file_operations alignment_proc_fops = { + .owner = THIS_MODULE, + .open = alignment_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, + .write = alignment_proc_write, +}; + +/* + * This needs to be done after sysctl_init, otherwise sys/ will be + * overwritten. Actually, this shouldn't be in sys/ at all since + * it isn't a sysctl, and it doesn't contain sysctl information. + * We now locate it in /proc/cpu/alignment instead. + */ +static int __init alignment_init(void) +{ + struct proc_dir_entry *dir, *res; + + dir = proc_mkdir("cpu", NULL); + if (!dir) + return -ENOMEM; + + res = proc_create_data("alignment", S_IWUSR | S_IRUGO, dir, + &alignment_proc_fops, &se_usermode); + if (!res) + return -ENOMEM; + + res = proc_create_data("kernel_alignment", S_IWUSR | S_IRUGO, dir, + &alignment_proc_fops, &se_kernmode_warn); + if (!res) + return -ENOMEM; + + return 0; +} +fs_initcall(alignment_init); -- cgit v0.10.2 From 94cd049522136c2f3bbe063d2e98b2b8d4286fd3 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 18:42:52 +0900 Subject: sh: sh_bios detection. This adds some VBR sanity checks in the sh_bios code to ensure that the BIOS VBR is in range before blindly trapping in to it. This permits boards with varying boot loader configurations to always leave support for sh-bios enabled and it will just be disabled at run-time if not found. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/sh_bios.c b/arch/sh/kernel/sh_bios.c index 29cd252..47475cc 100644 --- a/arch/sh/kernel/sh_bios.c +++ b/arch/sh/kernel/sh_bios.c @@ -23,6 +23,8 @@ #define BIOS_CALL_SHUTDOWN 11 #define BIOS_CALL_GDB_DETACH 0xff +void *gdb_vbr_vector = NULL; + static inline long sh_bios_call(long func, long arg0, long arg1, long arg2, long arg3) { @@ -32,6 +34,9 @@ static inline long sh_bios_call(long func, long arg0, long arg1, long arg2, register long r6 __asm__("r6") = arg2; register long r7 __asm__("r7") = arg3; + if (!gdb_vbr_vector) + return -ENOSYS; + __asm__ __volatile__("trapa #0x3f":"=z"(r0) :"0"(r0), "r"(r4), "r"(r5), "r"(r6), "r"(r7) :"memory"); @@ -60,8 +65,6 @@ void sh_bios_shutdown(unsigned int how) sh_bios_call(BIOS_CALL_SHUTDOWN, how, 0, 0, 0); } -void *gdb_vbr_vector = NULL; - /* * Read the old value of the VBR register to initialise the vector * through which debug and BIOS traps are delegated by the Linux trap @@ -76,8 +79,12 @@ void sh_bios_vbr_init(void) __asm__ __volatile__ ("stc vbr, %0" : "=r" (vbr)); - gdb_vbr_vector = (void *)(vbr + 0x100); - printk(KERN_NOTICE "Setting GDB trap vector to %p\n", gdb_vbr_vector); + if (vbr) { + gdb_vbr_vector = (void *)(vbr + 0x100); + printk(KERN_NOTICE "Setting GDB trap vector to %p\n", + gdb_vbr_vector); + } else + printk(KERN_NOTICE "SH-BIOS not detected\n"); } /** -- cgit v0.10.2 From 70e068eef97d05c97c3512f82352f39fdadfa8cb Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 18:52:00 +0900 Subject: sh: Move start_thread() out of line. start_thread() will become a bit heavier with the xstate freeing to be added in, so move it out-of-line in preparation. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/processor_32.h b/arch/sh/include/asm/processor_32.h index 1f3d6fa..50b8c9c 100644 --- a/arch/sh/include/asm/processor_32.h +++ b/arch/sh/include/asm/processor_32.h @@ -119,19 +119,10 @@ extern int ubc_usercnt; .sp = sizeof(init_stack) + (long) &init_stack, \ } -/* - * Do necessary setup to start up a newly executed thread. - */ -#define start_thread(_regs, new_pc, new_sp) \ - set_fs(USER_DS); \ - _regs->pr = 0; \ - _regs->sr = SR_FD; /* User mode. */ \ - _regs->pc = new_pc; \ - _regs->regs[15] = new_sp - /* Forward declaration, a strange C thing */ struct task_struct; -struct mm_struct; + +extern void start_thread(struct pt_regs *regs, unsigned long new_pc, unsigned long new_sp); /* Free all resources held by a thread. */ extern void release_thread(struct task_struct *); diff --git a/arch/sh/kernel/process_32.c b/arch/sh/kernel/process_32.c index d8af889..c436140 100644 --- a/arch/sh/kernel/process_32.c +++ b/arch/sh/kernel/process_32.c @@ -147,6 +147,18 @@ int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags) } EXPORT_SYMBOL(kernel_thread); +void start_thread(struct pt_regs *regs, unsigned long new_pc, + unsigned long new_sp) +{ + set_fs(USER_DS); + + regs->pr = 0; + regs->sr = SR_FD; + regs->pc = new_pc; + regs->regs[15] = new_sp; +} +EXPORT_SYMBOL(start_thread); + /* * Free current thread data structures etc.. */ -- cgit v0.10.2 From cbf6b1ba7ae12b3f7cb6b0d060b88d44649f9eda Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 19:01:11 +0900 Subject: sh: Always provide thread_info allocators. Presently the thread_info allocators are special cased, depending on THREAD_SHIFT < PAGE_SHIFT. This provides a sensible definition for them regardless of configuration, in preparation for extended CPU state. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/thread_info.h b/arch/sh/include/asm/thread_info.h index 1f3d927..2c5b48e 100644 --- a/arch/sh/include/asm/thread_info.h +++ b/arch/sh/include/asm/thread_info.h @@ -93,14 +93,12 @@ static inline struct thread_info *current_thread_info(void) #define THREAD_SIZE_ORDER (THREAD_SHIFT - PAGE_SHIFT) -#else /* THREAD_SHIFT < PAGE_SHIFT */ - -#define __HAVE_ARCH_THREAD_INFO_ALLOCATOR +#endif extern struct thread_info *alloc_thread_info(struct task_struct *tsk); extern void free_thread_info(struct thread_info *ti); -#endif /* THREAD_SHIFT < PAGE_SHIFT */ +#define __HAVE_ARCH_THREAD_INFO_ALLOCATOR #endif /* __ASSEMBLY__ */ diff --git a/arch/sh/kernel/Makefile b/arch/sh/kernel/Makefile index 5bec10c..379053c 100644 --- a/arch/sh/kernel/Makefile +++ b/arch/sh/kernel/Makefile @@ -13,8 +13,8 @@ CFLAGS_REMOVE_return_address.o = -pg obj-y := debugtraps.o dma-nommu.o dumpstack.o \ idle.o io.o io_generic.o irq.o \ - irq_$(BITS).o machvec.o nmi_debug.o process_$(BITS).o \ - ptrace_$(BITS).o return_address.o \ + irq_$(BITS).o machvec.o nmi_debug.o process.o \ + process_$(BITS).o ptrace_$(BITS).o return_address.o \ setup.o signal_$(BITS).o sys_sh.o sys_sh$(BITS).o \ syscalls_$(BITS).o time.o topology.o traps.o \ traps_$(BITS).o unwinder.o diff --git a/arch/sh/kernel/process.c b/arch/sh/kernel/process.c new file mode 100644 index 0000000..b2bda83b --- /dev/null +++ b/arch/sh/kernel/process.c @@ -0,0 +1,47 @@ +#include +#include +#include + +#if THREAD_SHIFT < PAGE_SHIFT +static struct kmem_cache *thread_info_cache; + +struct thread_info *alloc_thread_info(struct task_struct *tsk) +{ + struct thread_info *ti; + + ti = kmem_cache_alloc(thread_info_cache, GFP_KERNEL); + if (unlikely(ti == NULL)) + return NULL; +#ifdef CONFIG_DEBUG_STACK_USAGE + memset(ti, 0, THREAD_SIZE); +#endif + return ti; +} + +void free_thread_info(struct thread_info *ti) +{ + kmem_cache_free(thread_info_cache, ti); +} + +void thread_info_cache_init(void) +{ + thread_info_cache = kmem_cache_create("thread_info", THREAD_SIZE, + THREAD_SIZE, 0, NULL); + BUG_ON(thread_info_cache == NULL); +} +#else +struct thread_info *alloc_thread_info(struct task_struct *tsk) +{ +#ifdef CONFIG_DEBUG_STACK_USAGE + gfp_t mask = GFP_KERNEL | __GFP_ZERO; +#else + gfp_t mask = GFP_KERNEL; +#endif + return (struct thread_info *)__get_free_pages(mask, THREAD_SIZE_ORDER); +} + +void free_thread_info(struct thread_info *ti) +{ + free_pages((unsigned long)ti, THREAD_SIZE_ORDER); +} +#endif /* THREAD_SHIFT < PAGE_SHIFT */ diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c index 761910d..d5fb014 100644 --- a/arch/sh/mm/init.c +++ b/arch/sh/mm/init.c @@ -283,35 +283,6 @@ void free_initrd_mem(unsigned long start, unsigned long end) } #endif -#if THREAD_SHIFT < PAGE_SHIFT -static struct kmem_cache *thread_info_cache; - -struct thread_info *alloc_thread_info(struct task_struct *tsk) -{ - struct thread_info *ti; - - ti = kmem_cache_alloc(thread_info_cache, GFP_KERNEL); - if (unlikely(ti == NULL)) - return NULL; -#ifdef CONFIG_DEBUG_STACK_USAGE - memset(ti, 0, THREAD_SIZE); -#endif - return ti; -} - -void free_thread_info(struct thread_info *ti) -{ - kmem_cache_free(thread_info_cache, ti); -} - -void thread_info_cache_init(void) -{ - thread_info_cache = kmem_cache_create("thread_info", THREAD_SIZE, - THREAD_SIZE, 0, NULL); - BUG_ON(thread_info_cache == NULL); -} -#endif /* THREAD_SHIFT < PAGE_SHIFT */ - #ifdef CONFIG_MEMORY_HOTPLUG int arch_add_memory(int nid, u64 start, u64 size) { -- cgit v0.10.2 From a3705799e2cc5fb69d88ad6a7f317a8f5597f18d Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 12 Jan 2010 19:10:06 +0900 Subject: sh: Use SLAB_PANIC for thread_info slab cache. Presently this has a BUG_ON() for failure cases, as powerpc does. Switch this over to a SLAB_PANIC instead. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/process.c b/arch/sh/kernel/process.c index b2bda83b..077e06e 100644 --- a/arch/sh/kernel/process.c +++ b/arch/sh/kernel/process.c @@ -26,8 +26,7 @@ void free_thread_info(struct thread_info *ti) void thread_info_cache_init(void) { thread_info_cache = kmem_cache_create("thread_info", THREAD_SIZE, - THREAD_SIZE, 0, NULL); - BUG_ON(thread_info_cache == NULL); + THREAD_SIZE, SLAB_PANIC, NULL); } #else struct thread_info *alloc_thread_info(struct task_struct *tsk) -- cgit v0.10.2 From d67dec5b2cc208215de21dc7806945bf6a6e85d0 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?M=C3=A1rton=20N=C3=A9meth?= Date: Sun, 10 Jan 2010 17:59:22 +0100 Subject: HID: make USB device id constant MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The id_table field of the struct usb_device_id is constant in so it is worth to make the initialization data also constant. The semantic match that finds this kind of pattern is as follows: (http://coccinelle.lip6.fr/) // @r@ disable decl_init,const_decl_init; identifier I1, I2, x; @@ struct I1 { ... const struct I2 *x; ... }; @s@ identifier r.I1, y; identifier r.x, E; @@ struct I1 y = { .x = E, }; @c@ identifier r.I2; identifier s.E; @@ const struct I2 E[] = ... ; @depends on !c@ identifier r.I2; identifier s.E; @@ + const struct I2 E[] = ...; // Signed-off-by: Márton Németh Cc: Julia Lawall Cc: cocci@diku.dk Signed-off-by: Jiri Kosina diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c index 0258289..e72751d 100644 --- a/drivers/hid/usbhid/hid-core.c +++ b/drivers/hid/usbhid/hid-core.c @@ -1342,7 +1342,7 @@ static int hid_reset_resume(struct usb_interface *intf) #endif /* CONFIG_PM */ -static struct usb_device_id hid_usb_ids [] = { +static const struct usb_device_id hid_usb_ids[] = { { .match_flags = USB_DEVICE_ID_MATCH_INT_CLASS, .bInterfaceClass = USB_INTERFACE_CLASS_HID }, { } /* Terminating entry */ -- cgit v0.10.2 From 74f292ca8c7a2b9370f80d97a49e48174f4c7635 Mon Sep 17 00:00:00 2001 From: Gary Stein Date: Wed, 13 Jan 2010 00:25:58 +0100 Subject: HID: add driver for the Logitech Flight System G940 Implements a new USB-HID for Force Feedback based on the normal Logitech Force Feedback code and FF-Memless. Currently only supports the FF_CONSTANT effect although the joystick appears to support additional non-standard ones. Signed-off-by: Gary Stein Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index 5f73774..317049b 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -189,6 +189,14 @@ config LOGIRUMBLEPAD2_FF Say Y here if you want to enable force feedback support for Logitech Rumblepad 2 devices. +config LOGIG940_FF + bool "Logitech Flight System G940 force feedback support" + depends on HID_LOGITECH + select INPUT_FF_MEMLESS + help + Say Y here if you want to enable force feedback support for Logitech + Flight System G940 devices. + config HID_MICROSOFT tristate "Microsoft" if EMBEDDED depends on USB_HID diff --git a/drivers/hid/Makefile b/drivers/hid/Makefile index f8dc5ba..ce6c8da 100644 --- a/drivers/hid/Makefile +++ b/drivers/hid/Makefile @@ -18,6 +18,9 @@ endif ifdef CONFIG_LOGIRUMBLEPAD2_FF hid-logitech-objs += hid-lg2ff.o endif +ifdef CONFIG_LOGIG940_FF + hid-logitech-objs += hid-lg3ff.o +endif obj-$(CONFIG_HID_3M_PCT) += hid-3m-pct.o obj-$(CONFIG_HID_A4TECH) += hid-a4tech.o diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index a4b496c..a0c0c49 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1323,6 +1323,7 @@ static const struct hid_device_id hid_blacklist[] = { { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_LOGITECH_WINGMAN_F3D) }, { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_LOGITECH_WINGMAN_FFG ) }, { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_LOGITECH_FORCE3D_PRO) }, + { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_LOGITECH_FLIGHT_SYSTEM_G940) }, { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_LOGITECH_MOMO_WHEEL) }, { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_LOGITECH_MOMO_WHEEL2) }, { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_LOGITECH_G25_WHEEL) }, diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index 07e056b..c7c9cbf 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -300,6 +300,7 @@ #define USB_DEVICE_ID_LOGITECH_RUMBLEPAD2_2 0xc219 #define USB_DEVICE_ID_LOGITECH_WINGMAN_F3D 0xc283 #define USB_DEVICE_ID_LOGITECH_FORCE3D_PRO 0xc286 +#define USB_DEVICE_ID_LOGITECH_FLIGHT_SYSTEM_G940 0xc287 #define USB_DEVICE_ID_LOGITECH_WHEEL 0xc294 #define USB_DEVICE_ID_LOGITECH_WINGMAN_FFG 0xc293 #define USB_DEVICE_ID_LOGITECH_MOMO_WHEEL 0xc295 diff --git a/drivers/hid/hid-lg.c b/drivers/hid/hid-lg.c index 9fcd3d0..3677c90 100644 --- a/drivers/hid/hid-lg.c +++ b/drivers/hid/hid-lg.c @@ -34,6 +34,7 @@ #define LG_FF 0x200 #define LG_FF2 0x400 #define LG_RDESC_REL_ABS 0x800 +#define LG_FF3 0x1000 /* * Certain Logitech keyboards send in report #3 keys which are far @@ -266,7 +267,7 @@ static int lg_probe(struct hid_device *hdev, const struct hid_device_id *id) goto err_free; } - if (quirks & (LG_FF | LG_FF2)) + if (quirks & (LG_FF | LG_FF2 | LG_FF3)) connect_mask &= ~HID_CONNECT_FF; ret = hid_hw_start(hdev, connect_mask); @@ -279,6 +280,8 @@ static int lg_probe(struct hid_device *hdev, const struct hid_device_id *id) lgff_init(hdev); if (quirks & LG_FF2) lg2ff_init(hdev); + if (quirks & LG_FF3) + lg3ff_init(hdev); return 0; err_free: @@ -331,6 +334,8 @@ static const struct hid_device_id lg_devices[] = { .driver_data = LG_FF }, { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_LOGITECH_RUMBLEPAD2), .driver_data = LG_FF2 }, + { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_LOGITECH_FLIGHT_SYSTEM_G940), + .driver_data = LG_FF3 }, { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_SPACENAVIGATOR), .driver_data = LG_RDESC_REL_ABS }, { HID_USB_DEVICE(USB_VENDOR_ID_LOGITECH, USB_DEVICE_ID_SPACETRAVELLER), diff --git a/drivers/hid/hid-lg.h b/drivers/hid/hid-lg.h index 27ae750..8069c3e 100644 --- a/drivers/hid/hid-lg.h +++ b/drivers/hid/hid-lg.h @@ -15,4 +15,10 @@ int lg2ff_init(struct hid_device *hdev); static inline int lg2ff_init(struct hid_device *hdev) { return -1; } #endif +#ifdef CONFIG_LOGIG940_FF +int lg3ff_init(struct hid_device *hdev); +#else +static inline int lg3ff_init(struct hid_device *hdev) { return -1; } +#endif + #endif diff --git a/drivers/hid/hid-lg3ff.c b/drivers/hid/hid-lg3ff.c new file mode 100644 index 0000000..4002832 --- /dev/null +++ b/drivers/hid/hid-lg3ff.c @@ -0,0 +1,176 @@ +/* + * Force feedback support for Logitech Flight System G940 + * + * Copyright (c) 2009 Gary Stein + */ + +/* + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + + +#include +#include +#include + +#include "usbhid/usbhid.h" +#include "hid-lg.h" + +/* + * G940 Theory of Operation (from experimentation) + * + * There are 63 fields (only 3 of them currently used) + * 0 - seems to be command field + * 1 - 30 deal with the x axis + * 31 -60 deal with the y axis + * + * Field 1 is x axis constant force + * Field 31 is y axis constant force + * + * other interesting fields 1,2,3,4 on x axis + * (same for 31,32,33,34 on y axis) + * + * 0 0 127 127 makes the joystick autocenter hard + * + * 127 0 127 127 makes the joystick loose on the right, + * but stops all movemnt left + * + * -127 0 -127 -127 makes the joystick loose on the left, + * but stops all movement right + * + * 0 0 -127 -127 makes the joystick rattle very hard + * + * I'm sure these are effects that I don't know enough about them + */ + +struct lg3ff_device { + struct hid_report *report; +}; + +static int hid_lg3ff_play(struct input_dev *dev, void *data, + struct ff_effect *effect) +{ + struct hid_device *hid = input_get_drvdata(dev); + struct list_head *report_list = &hid->report_enum[HID_OUTPUT_REPORT].report_list; + struct hid_report *report = list_entry(report_list->next, struct hid_report, list); + int x, y; + +/* + * Maxusage should always be 63 (maximum fields) + * likely a better way to ensure this data is clean + */ + memset(report->field[0]->value, 0, sizeof(__s32)*report->field[0]->maxusage); + + switch (effect->type) { + case FF_CONSTANT: +/* + * Already clamped in ff_memless + * 0 is center (different then other logitech) + */ + x = effect->u.ramp.start_level; + y = effect->u.ramp.end_level; + + /* send command byte */ + report->field[0]->value[0] = 0x51; + +/* + * Sign backwards from other Force3d pro + * which get recast here in two's complement 8 bits + */ + report->field[0]->value[1] = (unsigned char)(-x); + report->field[0]->value[31] = (unsigned char)(-y); + + usbhid_submit_report(hid, report, USB_DIR_OUT); + break; + } + return 0; +} +static void hid_lg3ff_set_autocenter(struct input_dev *dev, u16 magnitude) +{ + struct hid_device *hid = input_get_drvdata(dev); + struct list_head *report_list = &hid->report_enum[HID_OUTPUT_REPORT].report_list; + struct hid_report *report = list_entry(report_list->next, struct hid_report, list); + +/* + * Auto Centering probed from device + * NOTE: deadman's switch on G940 must be covered + * for effects to work + */ + report->field[0]->value[0] = 0x51; + report->field[0]->value[1] = 0x00; + report->field[0]->value[2] = 0x00; + report->field[0]->value[3] = 0x7F; + report->field[0]->value[4] = 0x7F; + report->field[0]->value[31] = 0x00; + report->field[0]->value[32] = 0x00; + report->field[0]->value[33] = 0x7F; + report->field[0]->value[34] = 0x7F; + + usbhid_submit_report(hid, report, USB_DIR_OUT); +} + + +static const signed short ff3_joystick_ac[] = { + FF_CONSTANT, + FF_AUTOCENTER, + -1 +}; + +int lg3ff_init(struct hid_device *hid) +{ + struct hid_input *hidinput = list_entry(hid->inputs.next, struct hid_input, list); + struct list_head *report_list = &hid->report_enum[HID_OUTPUT_REPORT].report_list; + struct input_dev *dev = hidinput->input; + struct hid_report *report; + struct hid_field *field; + const signed short *ff_bits = ff3_joystick_ac; + int error; + int i; + + /* Find the report to use */ + if (list_empty(report_list)) { + err_hid("No output report found"); + return -1; + } + + /* Check that the report looks ok */ + report = list_entry(report_list->next, struct hid_report, list); + if (!report) { + err_hid("NULL output report"); + return -1; + } + + field = report->field[0]; + if (!field) { + err_hid("NULL field"); + return -1; + } + + /* Assume single fixed device G940 */ + for (i = 0; ff_bits[i] >= 0; i++) + set_bit(ff_bits[i], dev->ffbit); + + error = input_ff_create_memless(dev, NULL, hid_lg3ff_play); + if (error) + return error; + + if (test_bit(FF_AUTOCENTER, dev->ffbit)) + dev->ff->set_autocenter = hid_lg3ff_set_autocenter; + + dev_info(&hid->dev, "Force feedback for Logitech Flight System G940 by " + "Gary Stein \n"); + return 0; +} + diff --git a/drivers/hid/hid-lgff.c b/drivers/hid/hid-lgff.c index 987abeb..61142b7 100644 --- a/drivers/hid/hid-lgff.c +++ b/drivers/hid/hid-lgff.c @@ -67,6 +67,7 @@ static const struct dev_type devices[] = { { 0x046d, 0xc219, ff_rumble }, { 0x046d, 0xc283, ff_joystick }, { 0x046d, 0xc286, ff_joystick_ac }, + { 0x046d, 0xc287, ff_joystick_ac }, { 0x046d, 0xc293, ff_joystick }, { 0x046d, 0xc294, ff_wheel }, { 0x046d, 0xc295, ff_joystick }, -- cgit v0.10.2 From 49e4739a0cf681cbfe08c72232c1dcc130b66dde Mon Sep 17 00:00:00 2001 From: Stephane Chatty Date: Wed, 13 Jan 2010 00:29:16 +0100 Subject: HID: add support for Acer T230H multitouch Add support for the Quanta Optical Touch dual-touch panel, present in the Acer T230H monitor, HP L2105tm, and Packard-Bell Video 200t. Signed-off-by: Stephane Chatty Tested-by: Jerome Vidal Tested-by: Cedric Berthier Acked-by: Dmitry Torokhov Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index 317049b..38e9692 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -241,6 +241,12 @@ config HID_PETALYNX ---help--- Support for Petalynx Maxter remote control. +config HID_QUANTA + tristate "Quanta Optical Touch" + depends on USB_HID + ---help--- + Support for Quanta Optical Touch dual-touch panels. + config HID_SAMSUNG tristate "Samsung" if EMBEDDED depends on USB_HID diff --git a/drivers/hid/Makefile b/drivers/hid/Makefile index ce6c8da..15541c4 100644 --- a/drivers/hid/Makefile +++ b/drivers/hid/Makefile @@ -38,6 +38,7 @@ obj-$(CONFIG_HID_LOGITECH) += hid-logitech.o obj-$(CONFIG_HID_MICROSOFT) += hid-microsoft.o obj-$(CONFIG_HID_MONTEREY) += hid-monterey.o obj-$(CONFIG_HID_NTRIG) += hid-ntrig.o +obj-$(CONFIG_HID_QUANTA) += hid-quanta.o obj-$(CONFIG_HID_PANTHERLORD) += hid-pl.o obj-$(CONFIG_HID_PETALYNX) += hid-petalynx.o obj-$(CONFIG_HID_SAMSUNG) += hid-samsung.o diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index a0c0c49..056384c 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1338,6 +1338,7 @@ static const struct hid_device_id hid_blacklist[] = { { HID_USB_DEVICE(USB_VENDOR_ID_MONTEREY, USB_DEVICE_ID_GENIUS_KB29E) }, { HID_USB_DEVICE(USB_VENDOR_ID_NTRIG, USB_DEVICE_ID_NTRIG_TOUCH_SCREEN) }, { HID_USB_DEVICE(USB_VENDOR_ID_PETALYNX, USB_DEVICE_ID_PETALYNX_MAXTER_REMOTE) }, + { HID_USB_DEVICE(USB_VENDOR_ID_QUANTA, USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH) }, { HID_USB_DEVICE(USB_VENDOR_ID_SAMSUNG, USB_DEVICE_ID_SAMSUNG_IR_REMOTE) }, { HID_USB_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_PS3_CONTROLLER) }, { HID_USB_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_VAIO_VGX_MOUSE) }, diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index c7c9cbf..b009fd6 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -383,6 +383,9 @@ #define USB_VENDOR_ID_SAITEK 0x06a3 #define USB_DEVICE_ID_SAITEK_RUMBLEPAD 0xff17 +#define USB_VENDOR_ID_QUANTA 0x0408 +#define USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH 0x3000 + #define USB_VENDOR_ID_SAMSUNG 0x0419 #define USB_DEVICE_ID_SAMSUNG_IR_REMOTE 0x0001 diff --git a/drivers/hid/hid-quanta.c b/drivers/hid/hid-quanta.c new file mode 100644 index 0000000..244d61c --- /dev/null +++ b/drivers/hid/hid-quanta.c @@ -0,0 +1,259 @@ +/* + * HID driver for Quanta Optical Touch dual-touch panels + * + * Copyright (c) 2009-2010 Stephane Chatty + * + */ + +/* + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include +#include +#include + +MODULE_VERSION("1.00"); +MODULE_AUTHOR("Stephane Chatty "); +MODULE_DESCRIPTION("Quanta dual-touch panel"); +MODULE_LICENSE("GPL"); + +#include "hid-ids.h" + +struct quanta_data { + __u16 x, y; + __u8 id; + bool valid; /* valid finger data, or just placeholder? */ + bool first; /* is this the first finger in this frame? */ + bool activity_now; /* at least one active finger in this frame? */ + bool activity; /* at least one active finger previously? */ +}; + +static int quanta_input_mapping(struct hid_device *hdev, struct hid_input *hi, + struct hid_field *field, struct hid_usage *usage, + unsigned long **bit, int *max) +{ + switch (usage->hid & HID_USAGE_PAGE) { + + case HID_UP_GENDESK: + switch (usage->hid) { + case HID_GD_X: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_POSITION_X); + /* touchscreen emulation */ + input_set_abs_params(hi->input, ABS_X, + field->logical_minimum, + field->logical_maximum, 0, 0); + return 1; + case HID_GD_Y: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_POSITION_Y); + /* touchscreen emulation */ + input_set_abs_params(hi->input, ABS_Y, + field->logical_minimum, + field->logical_maximum, 0, 0); + return 1; + } + return 0; + + case HID_UP_DIGITIZER: + switch (usage->hid) { + case HID_DG_CONFIDENCE: + case HID_DG_TIPSWITCH: + case HID_DG_INPUTMODE: + case HID_DG_DEVICEINDEX: + case HID_DG_CONTACTCOUNT: + case HID_DG_CONTACTMAX: + case HID_DG_TIPPRESSURE: + case HID_DG_WIDTH: + case HID_DG_HEIGHT: + return -1; + case HID_DG_INRANGE: + /* touchscreen emulation */ + hid_map_usage(hi, usage, bit, max, EV_KEY, BTN_TOUCH); + return 1; + case HID_DG_CONTACTID: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_TRACKING_ID); + return 1; + } + return 0; + + case 0xff000000: + /* ignore vendor-specific features */ + return -1; + } + + return 0; +} + +static int quanta_input_mapped(struct hid_device *hdev, struct hid_input *hi, + struct hid_field *field, struct hid_usage *usage, + unsigned long **bit, int *max) +{ + if (usage->type == EV_KEY || usage->type == EV_ABS) + clear_bit(usage->code, *bit); + + return 0; +} + +/* + * this function is called when a whole finger has been parsed, + * so that it can decide what to send to the input layer. + */ +static void quanta_filter_event(struct quanta_data *td, struct input_dev *input) +{ + + td->first = !td->first; /* touchscreen emulation */ + + if (!td->valid) { + /* + * touchscreen emulation: if no finger in this frame is valid + * and there previously was finger activity, this is a release + */ + if (!td->first && !td->activity_now && td->activity) { + input_event(input, EV_KEY, BTN_TOUCH, 0); + td->activity = false; + } + return; + } + + input_event(input, EV_ABS, ABS_MT_TRACKING_ID, td->id); + input_event(input, EV_ABS, ABS_MT_POSITION_X, td->x); + input_event(input, EV_ABS, ABS_MT_POSITION_Y, td->y); + + input_mt_sync(input); + td->valid = false; + + /* touchscreen emulation: if first active finger in this frame... */ + if (!td->activity_now) { + /* if there was no previous activity, emit touch event */ + if (!td->activity) { + input_event(input, EV_KEY, BTN_TOUCH, 1); + td->activity = true; + } + td->activity_now = true; + /* and in any case this is our preferred finger */ + input_event(input, EV_ABS, ABS_X, td->x); + input_event(input, EV_ABS, ABS_Y, td->y); + } +} + + +static int quanta_event(struct hid_device *hid, struct hid_field *field, + struct hid_usage *usage, __s32 value) +{ + struct quanta_data *td = hid_get_drvdata(hid); + + if (hid->claimed & HID_CLAIMED_INPUT) { + struct input_dev *input = field->hidinput->input; + + switch (usage->hid) { + case HID_DG_INRANGE: + td->valid = !!value; + break; + case HID_GD_X: + td->x = value; + break; + case HID_GD_Y: + td->y = value; + quanta_filter_event(td, input); + break; + case HID_DG_CONTACTID: + td->id = value; + break; + case HID_DG_CONTACTCOUNT: + /* touch emulation: this is the last field in a frame */ + td->first = false; + td->activity_now = false; + break; + case HID_DG_CONFIDENCE: + case HID_DG_TIPSWITCH: + /* avoid interference from generic hidinput handling */ + break; + + default: + /* fallback to the generic hidinput handling */ + return 0; + } + } + + /* we have handled the hidinput part, now remains hiddev */ + if (hid->claimed & HID_CLAIMED_HIDDEV && hid->hiddev_hid_event) + hid->hiddev_hid_event(hid, field, usage, value); + + return 1; +} + +static int quanta_probe(struct hid_device *hdev, const struct hid_device_id *id) +{ + int ret; + struct quanta_data *td; + + td = kmalloc(sizeof(struct quanta_data), GFP_KERNEL); + if (!td) { + dev_err(&hdev->dev, "cannot allocate Quanta Touch data\n"); + return -ENOMEM; + } + td->valid = false; + td->activity = false; + td->activity_now = false; + td->first = false; + hid_set_drvdata(hdev, td); + + ret = hid_parse(hdev); + if (!ret) + ret = hid_hw_start(hdev, HID_CONNECT_DEFAULT); + + if (ret) + kfree(td); + + return ret; +} + +static void quanta_remove(struct hid_device *hdev) +{ + hid_hw_stop(hdev); + kfree(hid_get_drvdata(hdev)); + hid_set_drvdata(hdev, NULL); +} + +static const struct hid_device_id quanta_devices[] = { + { HID_USB_DEVICE(USB_VENDOR_ID_QUANTA, + USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH) }, + { } +}; +MODULE_DEVICE_TABLE(hid, quanta_devices); + +static const struct hid_usage_id quanta_grabbed_usages[] = { + { HID_ANY_ID, HID_ANY_ID, HID_ANY_ID }, + { HID_ANY_ID - 1, HID_ANY_ID - 1, HID_ANY_ID - 1} +}; + +static struct hid_driver quanta_driver = { + .name = "quanta-touch", + .id_table = quanta_devices, + .probe = quanta_probe, + .remove = quanta_remove, + .input_mapping = quanta_input_mapping, + .input_mapped = quanta_input_mapped, + .usage_table = quanta_grabbed_usages, + .event = quanta_event, +}; + +static int __init quanta_init(void) +{ + return hid_register_driver(&quanta_driver); +} + +static void __exit quanta_exit(void) +{ + hid_unregister_driver(&quanta_driver); +} + +module_init(quanta_init); +module_exit(quanta_exit); + -- cgit v0.10.2 From 0ea820cf9bf58f735ed40ec67947159c4f170012 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 13 Jan 2010 12:51:40 +0900 Subject: sh: Move over to dynamically allocated FPU context. This follows the x86 xstate changes and implements a task_xstate slab cache that is dynamically sized to match one of hard FP/soft FP/FPU-less. This also tidies up and consolidates some of the SH-2A/SH-4 FPU fragmentation. Now fpu state restorers are commonly defined, with the init_fpu()/fpu_init() mess reworked to follow the x86 convention. The fpu_init() register initialization has been replaced by xstate setup followed by writing out to hardware via the standard restore path. As init_fpu() now performs a slab allocation a secondary lighterweight restorer is also introduced for the context switch. In the future the DSP state will be rolled in here, too. More work remains for math emulation and the SH-5 FPU, which presently uses its own special (UP-only) interfaces. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/fpu.h b/arch/sh/include/asm/fpu.h index fb6bbb9..06c4281 100644 --- a/arch/sh/include/asm/fpu.h +++ b/arch/sh/include/asm/fpu.h @@ -2,8 +2,8 @@ #define __ASM_SH_FPU_H #ifndef __ASSEMBLY__ -#include -#include + +struct task_struct; #ifdef CONFIG_SH_FPU static inline void release_fpu(struct pt_regs *regs) @@ -16,22 +16,23 @@ static inline void grab_fpu(struct pt_regs *regs) regs->sr &= ~SR_FD; } -struct task_struct; - extern void save_fpu(struct task_struct *__tsk); -void fpu_state_restore(struct pt_regs *regs); +extern void restore_fpu(struct task_struct *__tsk); +extern void fpu_state_restore(struct pt_regs *regs); +extern void __fpu_state_restore(void); #else - -#define save_fpu(tsk) do { } while (0) -#define release_fpu(regs) do { } while (0) -#define grab_fpu(regs) do { } while (0) -#define fpu_state_restore(regs) do { } while (0) - +#define save_fpu(tsk) do { } while (0) +#define restore_fpu(tsk) do { } while (0) +#define release_fpu(regs) do { } while (0) +#define grab_fpu(regs) do { } while (0) +#define fpu_state_restore(regs) do { } while (0) +#define __fpu_state_restore(regs) do { } while (0) #endif struct user_regset; extern int do_fpu_inst(unsigned short, struct pt_regs *); +extern int init_fpu(struct task_struct *); extern int fpregs_get(struct task_struct *target, const struct user_regset *regset, @@ -65,18 +66,6 @@ static inline void clear_fpu(struct task_struct *tsk, struct pt_regs *regs) preempt_enable(); } -static inline int init_fpu(struct task_struct *tsk) -{ - if (tsk_used_math(tsk)) { - if ((boot_cpu_data.flags & CPU_HAS_FPU) && tsk == current) - unlazy_fpu(tsk, task_pt_regs(tsk)); - return 0; - } - - set_stopped_child_used_math(tsk); - return 0; -} - #endif /* __ASSEMBLY__ */ #endif /* __ASM_SH_FPU_H */ diff --git a/arch/sh/include/asm/processor_32.h b/arch/sh/include/asm/processor_32.h index 50b8c9c..a359898 100644 --- a/arch/sh/include/asm/processor_32.h +++ b/arch/sh/include/asm/processor_32.h @@ -90,11 +90,15 @@ struct sh_fpu_soft_struct { unsigned long entry_pc; }; -union sh_fpu_union { - struct sh_fpu_hard_struct hard; - struct sh_fpu_soft_struct soft; +union thread_xstate { + struct sh_fpu_hard_struct hardfpu; + struct sh_fpu_soft_struct softfpu; }; +extern unsigned int xstate_size; +extern void free_thread_xstate(struct task_struct *); +extern struct kmem_cache *task_xstate_cachep; + struct thread_struct { /* Saved registers when thread is descheduled */ unsigned long sp; @@ -103,13 +107,13 @@ struct thread_struct { /* Hardware debugging registers */ unsigned long ubc_pc; - /* floating point info */ - union sh_fpu_union fpu; - #ifdef CONFIG_SH_DSP /* Dsp status information */ struct sh_dsp_struct dsp_status; #endif + + /* Extended processor state */ + union thread_xstate *xstate; }; /* Count of active tasks with UBC settings */ diff --git a/arch/sh/include/asm/thread_info.h b/arch/sh/include/asm/thread_info.h index 2c5b48e..55a36fe 100644 --- a/arch/sh/include/asm/thread_info.h +++ b/arch/sh/include/asm/thread_info.h @@ -97,6 +97,10 @@ static inline struct thread_info *current_thread_info(void) extern struct thread_info *alloc_thread_info(struct task_struct *tsk); extern void free_thread_info(struct thread_info *ti); +extern void arch_task_cache_init(void); +#define arch_task_cache_init arch_task_cache_init +extern int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src); +extern void init_thread_xstate(void); #define __HAVE_ARCH_THREAD_INFO_ALLOCATOR diff --git a/arch/sh/kernel/cpu/Makefile b/arch/sh/kernel/cpu/Makefile index d97c803..0e48bc6 100644 --- a/arch/sh/kernel/cpu/Makefile +++ b/arch/sh/kernel/cpu/Makefile @@ -17,5 +17,7 @@ obj-$(CONFIG_ARCH_SHMOBILE) += shmobile/ obj-$(CONFIG_SH_ADC) += adc.o obj-$(CONFIG_SH_CLK_CPG) += clock-cpg.o +obj-$(CONFIG_SH_FPU) += fpu.o +obj-$(CONFIG_SH_FPU_EMU) += fpu.o obj-y += irq/ init.o clock.o hwblk.o diff --git a/arch/sh/kernel/cpu/fpu.c b/arch/sh/kernel/cpu/fpu.c new file mode 100644 index 0000000..c23e672 --- /dev/null +++ b/arch/sh/kernel/cpu/fpu.c @@ -0,0 +1,82 @@ +#include +#include +#include + +int init_fpu(struct task_struct *tsk) +{ + if (tsk_used_math(tsk)) { + if ((boot_cpu_data.flags & CPU_HAS_FPU) && tsk == current) + unlazy_fpu(tsk, task_pt_regs(tsk)); + return 0; + } + + /* + * Memory allocation at the first usage of the FPU and other state. + */ + if (!tsk->thread.xstate) { + tsk->thread.xstate = kmem_cache_alloc(task_xstate_cachep, + GFP_KERNEL); + if (!tsk->thread.xstate) + return -ENOMEM; + } + + if (boot_cpu_data.flags & CPU_HAS_FPU) { + struct sh_fpu_hard_struct *fp = &tsk->thread.xstate->hardfpu; + memset(fp, 0, xstate_size); + fp->fpscr = FPSCR_INIT; + } else { + struct sh_fpu_soft_struct *fp = &tsk->thread.xstate->softfpu; + memset(fp, 0, xstate_size); + fp->fpscr = FPSCR_INIT; + } + + set_stopped_child_used_math(tsk); + return 0; +} + +#ifdef CONFIG_SH_FPU +void __fpu_state_restore(void) +{ + struct task_struct *tsk = current; + + restore_fpu(tsk); + + task_thread_info(tsk)->status |= TS_USEDFPU; + tsk->fpu_counter++; +} + +void fpu_state_restore(struct pt_regs *regs) +{ + struct task_struct *tsk = current; + + if (unlikely(!user_mode(regs))) { + printk(KERN_ERR "BUG: FPU is used in kernel mode.\n"); + BUG(); + return; + } + + if (!tsk_used_math(tsk)) { + /* + * does a slab alloc which can sleep + */ + if (init_fpu(tsk)) { + /* + * ran out of memory! + */ + do_group_exit(SIGKILL); + return; + } + } + + grab_fpu(regs); + + __fpu_state_restore(); +} + +BUILD_TRAP_HANDLER(fpu_state_restore) +{ + TRAP_HANDLER_DECL; + + fpu_state_restore(regs); +} +#endif /* CONFIG_SH_FPU */ diff --git a/arch/sh/kernel/cpu/init.c b/arch/sh/kernel/cpu/init.c index 89b4b76..2e23422 100644 --- a/arch/sh/kernel/cpu/init.c +++ b/arch/sh/kernel/cpu/init.c @@ -28,18 +28,30 @@ #include #endif +#ifdef CONFIG_SH_FPU +#define cpu_has_fpu 1 +#else +#define cpu_has_fpu 0 +#endif + +#ifdef CONFIG_SH_DSP +#define cpu_has_dsp 1 +#else +#define cpu_has_dsp 0 +#endif + /* * Generic wrapper for command line arguments to disable on-chip * peripherals (nofpu, nodsp, and so forth). */ -#define onchip_setup(x) \ -static int x##_disabled __initdata = 0; \ - \ -static int __init x##_setup(char *opts) \ -{ \ - x##_disabled = 1; \ - return 1; \ -} \ +#define onchip_setup(x) \ +static int x##_disabled __initdata = !cpu_has_##x; \ + \ +static int __init x##_setup(char *opts) \ +{ \ + x##_disabled = 1; \ + return 1; \ +} \ __setup("no" __stringify(x), x##_setup); onchip_setup(fpu); @@ -207,6 +219,18 @@ static void detect_cache_shape(void) l2_cache_shape = -1; /* No S-cache */ } +static void __init fpu_init(void) +{ + /* Disable the FPU */ + if (fpu_disabled && (current_cpu_data.flags & CPU_HAS_FPU)) { + printk("FPU Disabled\n"); + current_cpu_data.flags &= ~CPU_HAS_FPU; + } + + disable_fpu(); + clear_used_math(); +} + #ifdef CONFIG_SH_DSP static void __init release_dsp(void) { @@ -244,9 +268,17 @@ static void __init dsp_init(void) if (sr & SR_DSP) current_cpu_data.flags |= CPU_HAS_DSP; + /* Disable the DSP */ + if (dsp_disabled && (current_cpu_data.flags & CPU_HAS_DSP)) { + printk("DSP Disabled\n"); + current_cpu_data.flags &= ~CPU_HAS_DSP; + } + /* Now that we've determined the DSP status, clear the DSP bit. */ release_dsp(); } +#else +static inline void __init dsp_init(void) { } #endif /* CONFIG_SH_DSP */ /** @@ -302,18 +334,8 @@ asmlinkage void __init sh_cpu_init(void) detect_cache_shape(); } - /* Disable the FPU */ - if (fpu_disabled) { - printk("FPU Disabled\n"); - current_cpu_data.flags &= ~CPU_HAS_FPU; - } - - /* FPU initialization */ - disable_fpu(); - if ((current_cpu_data.flags & CPU_HAS_FPU)) { - current_thread_info()->status &= ~TS_USEDFPU; - clear_used_math(); - } + fpu_init(); + dsp_init(); /* * Initialize the per-CPU ASID cache very early, since the @@ -321,18 +343,12 @@ asmlinkage void __init sh_cpu_init(void) */ current_cpu_data.asid_cache = NO_CONTEXT; -#ifdef CONFIG_SH_DSP - /* Probe for DSP */ - dsp_init(); - - /* Disable the DSP */ - if (dsp_disabled) { - printk("DSP Disabled\n"); - current_cpu_data.flags &= ~CPU_HAS_DSP; - release_dsp(); - } -#endif - speculative_execution_init(); expmask_init(); + + /* + * Boot processor to setup the FP and extended state context info. + */ + if (raw_smp_processor_id() == 0) + init_thread_xstate(); } diff --git a/arch/sh/kernel/cpu/sh2a/fpu.c b/arch/sh/kernel/cpu/sh2a/fpu.c index d395ce5..488d24e 100644 --- a/arch/sh/kernel/cpu/sh2a/fpu.c +++ b/arch/sh/kernel/cpu/sh2a/fpu.c @@ -26,8 +26,7 @@ /* * Save FPU registers onto task structure. */ -void -save_fpu(struct task_struct *tsk) +void save_fpu(struct task_struct *tsk) { unsigned long dummy; @@ -52,7 +51,7 @@ save_fpu(struct task_struct *tsk) "fmov.s fr0, @-%0\n\t" "lds %3, fpscr\n\t" : "=r" (dummy) - : "0" ((char *)(&tsk->thread.fpu.hard.status)), + : "0" ((char *)(&tsk->thread.xstate->hardfpu.status)), "r" (FPSCR_RCHG), "r" (FPSCR_INIT) : "memory"); @@ -60,8 +59,7 @@ save_fpu(struct task_struct *tsk) disable_fpu(); } -static void -restore_fpu(struct task_struct *tsk) +void restore_fpu(struct task_struct *tsk) { unsigned long dummy; @@ -85,45 +83,12 @@ restore_fpu(struct task_struct *tsk) "lds.l @%0+, fpscr\n\t" "lds.l @%0+, fpul\n\t" : "=r" (dummy) - : "0" (&tsk->thread.fpu), "r" (FPSCR_RCHG) + : "0" (tsk->thread.xstate), "r" (FPSCR_RCHG) : "memory"); disable_fpu(); } /* - * Load the FPU with signalling NANS. This bit pattern we're using - * has the property that no matter wether considered as single or as - * double precission represents signaling NANS. - */ - -static void -fpu_init(void) -{ - enable_fpu(); - asm volatile("lds %0, fpul\n\t" - "fsts fpul, fr0\n\t" - "fsts fpul, fr1\n\t" - "fsts fpul, fr2\n\t" - "fsts fpul, fr3\n\t" - "fsts fpul, fr4\n\t" - "fsts fpul, fr5\n\t" - "fsts fpul, fr6\n\t" - "fsts fpul, fr7\n\t" - "fsts fpul, fr8\n\t" - "fsts fpul, fr9\n\t" - "fsts fpul, fr10\n\t" - "fsts fpul, fr11\n\t" - "fsts fpul, fr12\n\t" - "fsts fpul, fr13\n\t" - "fsts fpul, fr14\n\t" - "fsts fpul, fr15\n\t" - "lds %2, fpscr\n\t" - : /* no output */ - : "r" (0), "r" (FPSCR_RCHG), "r" (FPSCR_INIT)); - disable_fpu(); -} - -/* * Emulate arithmetic ops on denormalized number for some FPU insns. */ @@ -490,9 +455,9 @@ ieee_fpe_handler (struct pt_regs *regs) if ((finsn & 0xf1ff) == 0xf0ad) { /* fcnvsd */ struct task_struct *tsk = current; - if ((tsk->thread.fpu.hard.fpscr & FPSCR_FPU_ERROR)) { + if ((tsk->thread.xstate->hardfpu.fpscr & FPSCR_FPU_ERROR)) { /* FPU error */ - denormal_to_double (&tsk->thread.fpu.hard, + denormal_to_double (&tsk->thread.xstate->hardfpu, (finsn >> 8) & 0xf); } else return 0; @@ -507,9 +472,9 @@ ieee_fpe_handler (struct pt_regs *regs) n = (finsn >> 8) & 0xf; m = (finsn >> 4) & 0xf; - hx = tsk->thread.fpu.hard.fp_regs[n]; - hy = tsk->thread.fpu.hard.fp_regs[m]; - fpscr = tsk->thread.fpu.hard.fpscr; + hx = tsk->thread.xstate->hardfpu.fp_regs[n]; + hy = tsk->thread.xstate->hardfpu.fp_regs[m]; + fpscr = tsk->thread.xstate->hardfpu.fpscr; prec = fpscr & (1 << 19); if ((fpscr & FPSCR_FPU_ERROR) @@ -519,15 +484,15 @@ ieee_fpe_handler (struct pt_regs *regs) /* FPU error because of denormal */ llx = ((long long) hx << 32) - | tsk->thread.fpu.hard.fp_regs[n+1]; + | tsk->thread.xstate->hardfpu.fp_regs[n+1]; lly = ((long long) hy << 32) - | tsk->thread.fpu.hard.fp_regs[m+1]; + | tsk->thread.xstate->hardfpu.fp_regs[m+1]; if ((hx & 0x7fffffff) >= 0x00100000) llx = denormal_muld(lly, llx); else llx = denormal_muld(llx, lly); - tsk->thread.fpu.hard.fp_regs[n] = llx >> 32; - tsk->thread.fpu.hard.fp_regs[n+1] = llx & 0xffffffff; + tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32; + tsk->thread.xstate->hardfpu.fp_regs[n+1] = llx & 0xffffffff; } else if ((fpscr & FPSCR_FPU_ERROR) && (!prec && ((hx & 0x7fffffff) < 0x00800000 || (hy & 0x7fffffff) < 0x00800000))) { @@ -536,7 +501,7 @@ ieee_fpe_handler (struct pt_regs *regs) hx = denormal_mulf(hy, hx); else hx = denormal_mulf(hx, hy); - tsk->thread.fpu.hard.fp_regs[n] = hx; + tsk->thread.xstate->hardfpu.fp_regs[n] = hx; } else return 0; @@ -550,9 +515,9 @@ ieee_fpe_handler (struct pt_regs *regs) n = (finsn >> 8) & 0xf; m = (finsn >> 4) & 0xf; - hx = tsk->thread.fpu.hard.fp_regs[n]; - hy = tsk->thread.fpu.hard.fp_regs[m]; - fpscr = tsk->thread.fpu.hard.fpscr; + hx = tsk->thread.xstate->hardfpu.fp_regs[n]; + hy = tsk->thread.xstate->hardfpu.fp_regs[m]; + fpscr = tsk->thread.xstate->hardfpu.fpscr; prec = fpscr & (1 << 19); if ((fpscr & FPSCR_FPU_ERROR) @@ -562,15 +527,15 @@ ieee_fpe_handler (struct pt_regs *regs) /* FPU error because of denormal */ llx = ((long long) hx << 32) - | tsk->thread.fpu.hard.fp_regs[n+1]; + | tsk->thread.xstate->hardfpu.fp_regs[n+1]; lly = ((long long) hy << 32) - | tsk->thread.fpu.hard.fp_regs[m+1]; + | tsk->thread.xstate->hardfpu.fp_regs[m+1]; if ((finsn & 0xf00f) == 0xf000) llx = denormal_addd(llx, lly); else llx = denormal_addd(llx, lly ^ (1LL << 63)); - tsk->thread.fpu.hard.fp_regs[n] = llx >> 32; - tsk->thread.fpu.hard.fp_regs[n+1] = llx & 0xffffffff; + tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32; + tsk->thread.xstate->hardfpu.fp_regs[n+1] = llx & 0xffffffff; } else if ((fpscr & FPSCR_FPU_ERROR) && (!prec && ((hx & 0x7fffffff) < 0x00800000 || (hy & 0x7fffffff) < 0x00800000))) { @@ -579,7 +544,7 @@ ieee_fpe_handler (struct pt_regs *regs) hx = denormal_addf(hx, hy); else hx = denormal_addf(hx, hy ^ 0x80000000); - tsk->thread.fpu.hard.fp_regs[n] = hx; + tsk->thread.xstate->hardfpu.fp_regs[n] = hx; } else return 0; @@ -597,7 +562,7 @@ BUILD_TRAP_HANDLER(fpu_error) __unlazy_fpu(tsk, regs); if (ieee_fpe_handler(regs)) { - tsk->thread.fpu.hard.fpscr &= + tsk->thread.xstate->hardfpu.fpscr &= ~(FPSCR_CAUSE_MASK | FPSCR_FLAG_MASK); grab_fpu(regs); restore_fpu(tsk); @@ -607,33 +572,3 @@ BUILD_TRAP_HANDLER(fpu_error) force_sig(SIGFPE, tsk); } - -void fpu_state_restore(struct pt_regs *regs) -{ - struct task_struct *tsk = current; - - grab_fpu(regs); - if (unlikely(!user_mode(regs))) { - printk(KERN_ERR "BUG: FPU is used in kernel mode.\n"); - BUG(); - return; - } - - if (likely(used_math())) { - /* Using the FPU again. */ - restore_fpu(tsk); - } else { - /* First time FPU user. */ - fpu_init(); - set_used_math(); - } - task_thread_info(tsk)->status |= TS_USEDFPU; - tsk->fpu_counter++; -} - -BUILD_TRAP_HANDLER(fpu_state_restore) -{ - TRAP_HANDLER_DECL; - - fpu_state_restore(regs); -} diff --git a/arch/sh/kernel/cpu/sh4/fpu.c b/arch/sh/kernel/cpu/sh4/fpu.c index e97857a..447482d 100644 --- a/arch/sh/kernel/cpu/sh4/fpu.c +++ b/arch/sh/kernel/cpu/sh4/fpu.c @@ -85,14 +85,14 @@ void save_fpu(struct task_struct *tsk) "fmov.s fr1, @-%0\n\t" "fmov.s fr0, @-%0\n\t" "lds %3, fpscr\n\t":"=r" (dummy) - :"0"((char *)(&tsk->thread.fpu.hard.status)), + :"0"((char *)(&tsk->thread.xstate->hardfpu.status)), "r"(FPSCR_RCHG), "r"(FPSCR_INIT) :"memory"); disable_fpu(); } -static void restore_fpu(struct task_struct *tsk) +void restore_fpu(struct task_struct *tsk) { unsigned long dummy; @@ -135,62 +135,11 @@ static void restore_fpu(struct task_struct *tsk) "lds.l @%0+, fpscr\n\t" "lds.l @%0+, fpul\n\t" :"=r" (dummy) - :"0"(&tsk->thread.fpu), "r"(FPSCR_RCHG) + :"0" (tsk->thread.xstate), "r" (FPSCR_RCHG) :"memory"); disable_fpu(); } -/* - * Load the FPU with signalling NANS. This bit pattern we're using - * has the property that no matter wether considered as single or as - * double precision represents signaling NANS. - */ - -static void fpu_init(void) -{ - enable_fpu(); - asm volatile ( "lds %0, fpul\n\t" - "lds %1, fpscr\n\t" - "fsts fpul, fr0\n\t" - "fsts fpul, fr1\n\t" - "fsts fpul, fr2\n\t" - "fsts fpul, fr3\n\t" - "fsts fpul, fr4\n\t" - "fsts fpul, fr5\n\t" - "fsts fpul, fr6\n\t" - "fsts fpul, fr7\n\t" - "fsts fpul, fr8\n\t" - "fsts fpul, fr9\n\t" - "fsts fpul, fr10\n\t" - "fsts fpul, fr11\n\t" - "fsts fpul, fr12\n\t" - "fsts fpul, fr13\n\t" - "fsts fpul, fr14\n\t" - "fsts fpul, fr15\n\t" - "frchg\n\t" - "fsts fpul, fr0\n\t" - "fsts fpul, fr1\n\t" - "fsts fpul, fr2\n\t" - "fsts fpul, fr3\n\t" - "fsts fpul, fr4\n\t" - "fsts fpul, fr5\n\t" - "fsts fpul, fr6\n\t" - "fsts fpul, fr7\n\t" - "fsts fpul, fr8\n\t" - "fsts fpul, fr9\n\t" - "fsts fpul, fr10\n\t" - "fsts fpul, fr11\n\t" - "fsts fpul, fr12\n\t" - "fsts fpul, fr13\n\t" - "fsts fpul, fr14\n\t" - "fsts fpul, fr15\n\t" - "frchg\n\t" - "lds %2, fpscr\n\t" - : /* no output */ - :"r" (0), "r"(FPSCR_RCHG), "r"(FPSCR_INIT)); - disable_fpu(); -} - /** * denormal_to_double - Given denormalized float number, * store double float @@ -282,9 +231,9 @@ static int ieee_fpe_handler(struct pt_regs *regs) /* fcnvsd */ struct task_struct *tsk = current; - if ((tsk->thread.fpu.hard.fpscr & FPSCR_CAUSE_ERROR)) + if ((tsk->thread.xstate->hardfpu.fpscr & FPSCR_CAUSE_ERROR)) /* FPU error */ - denormal_to_double(&tsk->thread.fpu.hard, + denormal_to_double(&tsk->thread.xstate->hardfpu, (finsn >> 8) & 0xf); else return 0; @@ -300,9 +249,9 @@ static int ieee_fpe_handler(struct pt_regs *regs) n = (finsn >> 8) & 0xf; m = (finsn >> 4) & 0xf; - hx = tsk->thread.fpu.hard.fp_regs[n]; - hy = tsk->thread.fpu.hard.fp_regs[m]; - fpscr = tsk->thread.fpu.hard.fpscr; + hx = tsk->thread.xstate->hardfpu.fp_regs[n]; + hy = tsk->thread.xstate->hardfpu.fp_regs[m]; + fpscr = tsk->thread.xstate->hardfpu.fpscr; prec = fpscr & FPSCR_DBL_PRECISION; if ((fpscr & FPSCR_CAUSE_ERROR) @@ -312,18 +261,18 @@ static int ieee_fpe_handler(struct pt_regs *regs) /* FPU error because of denormal (doubles) */ llx = ((long long)hx << 32) - | tsk->thread.fpu.hard.fp_regs[n + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[n + 1]; lly = ((long long)hy << 32) - | tsk->thread.fpu.hard.fp_regs[m + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[m + 1]; llx = float64_mul(llx, lly); - tsk->thread.fpu.hard.fp_regs[n] = llx >> 32; - tsk->thread.fpu.hard.fp_regs[n + 1] = llx & 0xffffffff; + tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32; + tsk->thread.xstate->hardfpu.fp_regs[n + 1] = llx & 0xffffffff; } else if ((fpscr & FPSCR_CAUSE_ERROR) && (!prec && ((hx & 0x7fffffff) < 0x00800000 || (hy & 0x7fffffff) < 0x00800000))) { /* FPU error because of denormal (floats) */ hx = float32_mul(hx, hy); - tsk->thread.fpu.hard.fp_regs[n] = hx; + tsk->thread.xstate->hardfpu.fp_regs[n] = hx; } else return 0; @@ -338,9 +287,9 @@ static int ieee_fpe_handler(struct pt_regs *regs) n = (finsn >> 8) & 0xf; m = (finsn >> 4) & 0xf; - hx = tsk->thread.fpu.hard.fp_regs[n]; - hy = tsk->thread.fpu.hard.fp_regs[m]; - fpscr = tsk->thread.fpu.hard.fpscr; + hx = tsk->thread.xstate->hardfpu.fp_regs[n]; + hy = tsk->thread.xstate->hardfpu.fp_regs[m]; + fpscr = tsk->thread.xstate->hardfpu.fpscr; prec = fpscr & FPSCR_DBL_PRECISION; if ((fpscr & FPSCR_CAUSE_ERROR) @@ -350,15 +299,15 @@ static int ieee_fpe_handler(struct pt_regs *regs) /* FPU error because of denormal (doubles) */ llx = ((long long)hx << 32) - | tsk->thread.fpu.hard.fp_regs[n + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[n + 1]; lly = ((long long)hy << 32) - | tsk->thread.fpu.hard.fp_regs[m + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[m + 1]; if ((finsn & 0xf00f) == 0xf000) llx = float64_add(llx, lly); else llx = float64_sub(llx, lly); - tsk->thread.fpu.hard.fp_regs[n] = llx >> 32; - tsk->thread.fpu.hard.fp_regs[n + 1] = llx & 0xffffffff; + tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32; + tsk->thread.xstate->hardfpu.fp_regs[n + 1] = llx & 0xffffffff; } else if ((fpscr & FPSCR_CAUSE_ERROR) && (!prec && ((hx & 0x7fffffff) < 0x00800000 || (hy & 0x7fffffff) < 0x00800000))) { @@ -367,7 +316,7 @@ static int ieee_fpe_handler(struct pt_regs *regs) hx = float32_add(hx, hy); else hx = float32_sub(hx, hy); - tsk->thread.fpu.hard.fp_regs[n] = hx; + tsk->thread.xstate->hardfpu.fp_regs[n] = hx; } else return 0; @@ -382,9 +331,9 @@ static int ieee_fpe_handler(struct pt_regs *regs) n = (finsn >> 8) & 0xf; m = (finsn >> 4) & 0xf; - hx = tsk->thread.fpu.hard.fp_regs[n]; - hy = tsk->thread.fpu.hard.fp_regs[m]; - fpscr = tsk->thread.fpu.hard.fpscr; + hx = tsk->thread.xstate->hardfpu.fp_regs[n]; + hy = tsk->thread.xstate->hardfpu.fp_regs[m]; + fpscr = tsk->thread.xstate->hardfpu.fpscr; prec = fpscr & FPSCR_DBL_PRECISION; if ((fpscr & FPSCR_CAUSE_ERROR) @@ -394,20 +343,20 @@ static int ieee_fpe_handler(struct pt_regs *regs) /* FPU error because of denormal (doubles) */ llx = ((long long)hx << 32) - | tsk->thread.fpu.hard.fp_regs[n + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[n + 1]; lly = ((long long)hy << 32) - | tsk->thread.fpu.hard.fp_regs[m + 1]; + | tsk->thread.xstate->hardfpu.fp_regs[m + 1]; llx = float64_div(llx, lly); - tsk->thread.fpu.hard.fp_regs[n] = llx >> 32; - tsk->thread.fpu.hard.fp_regs[n + 1] = llx & 0xffffffff; + tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32; + tsk->thread.xstate->hardfpu.fp_regs[n + 1] = llx & 0xffffffff; } else if ((fpscr & FPSCR_CAUSE_ERROR) && (!prec && ((hx & 0x7fffffff) < 0x00800000 || (hy & 0x7fffffff) < 0x00800000))) { /* FPU error because of denormal (floats) */ hx = float32_div(hx, hy); - tsk->thread.fpu.hard.fp_regs[n] = hx; + tsk->thread.xstate->hardfpu.fp_regs[n] = hx; } else return 0; @@ -420,17 +369,17 @@ static int ieee_fpe_handler(struct pt_regs *regs) unsigned int hx; m = (finsn >> 8) & 0x7; - hx = tsk->thread.fpu.hard.fp_regs[m]; + hx = tsk->thread.xstate->hardfpu.fp_regs[m]; - if ((tsk->thread.fpu.hard.fpscr & FPSCR_CAUSE_ERROR) + if ((tsk->thread.xstate->hardfpu.fpscr & FPSCR_CAUSE_ERROR) && ((hx & 0x7fffffff) < 0x00100000)) { /* subnormal double to float conversion */ long long llx; - llx = ((long long)tsk->thread.fpu.hard.fp_regs[m] << 32) - | tsk->thread.fpu.hard.fp_regs[m + 1]; + llx = ((long long)tsk->thread.xstate->hardfpu.fp_regs[m] << 32) + | tsk->thread.xstate->hardfpu.fp_regs[m + 1]; - tsk->thread.fpu.hard.fpul = float64_to_float32(llx); + tsk->thread.xstate->hardfpu.fpul = float64_to_float32(llx); } else return 0; @@ -449,7 +398,7 @@ void float_raise(unsigned int flags) int float_rounding_mode(void) { struct task_struct *tsk = current; - int roundingMode = FPSCR_ROUNDING_MODE(tsk->thread.fpu.hard.fpscr); + int roundingMode = FPSCR_ROUNDING_MODE(tsk->thread.xstate->hardfpu.fpscr); return roundingMode; } @@ -461,16 +410,16 @@ BUILD_TRAP_HANDLER(fpu_error) __unlazy_fpu(tsk, regs); fpu_exception_flags = 0; if (ieee_fpe_handler(regs)) { - tsk->thread.fpu.hard.fpscr &= + tsk->thread.xstate->hardfpu.fpscr &= ~(FPSCR_CAUSE_MASK | FPSCR_FLAG_MASK); - tsk->thread.fpu.hard.fpscr |= fpu_exception_flags; + tsk->thread.xstate->hardfpu.fpscr |= fpu_exception_flags; /* Set the FPSCR flag as well as cause bits - simply * replicate the cause */ - tsk->thread.fpu.hard.fpscr |= (fpu_exception_flags >> 10); + tsk->thread.xstate->hardfpu.fpscr |= (fpu_exception_flags >> 10); grab_fpu(regs); restore_fpu(tsk); task_thread_info(tsk)->status |= TS_USEDFPU; - if ((((tsk->thread.fpu.hard.fpscr & FPSCR_ENABLE_MASK) >> 7) & + if ((((tsk->thread.xstate->hardfpu.fpscr & FPSCR_ENABLE_MASK) >> 7) & (fpu_exception_flags >> 2)) == 0) { return; } @@ -478,33 +427,3 @@ BUILD_TRAP_HANDLER(fpu_error) force_sig(SIGFPE, tsk); } - -void fpu_state_restore(struct pt_regs *regs) -{ - struct task_struct *tsk = current; - - grab_fpu(regs); - if (unlikely(!user_mode(regs))) { - printk(KERN_ERR "BUG: FPU is used in kernel mode.\n"); - BUG(); - return; - } - - if (likely(used_math())) { - /* Using the FPU again. */ - restore_fpu(tsk); - } else { - /* First time FPU user. */ - fpu_init(); - set_used_math(); - } - task_thread_info(tsk)->status |= TS_USEDFPU; - tsk->fpu_counter++; -} - -BUILD_TRAP_HANDLER(fpu_state_restore) -{ - TRAP_HANDLER_DECL; - - fpu_state_restore(regs); -} diff --git a/arch/sh/kernel/process.c b/arch/sh/kernel/process.c index 077e06e..81add9b 100644 --- a/arch/sh/kernel/process.c +++ b/arch/sh/kernel/process.c @@ -2,6 +2,32 @@ #include #include +struct kmem_cache *task_xstate_cachep = NULL; +unsigned int xstate_size; + +int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src) +{ + *dst = *src; + + if (src->thread.xstate) { + dst->thread.xstate = kmem_cache_alloc(task_xstate_cachep, + GFP_KERNEL); + if (!dst->thread.xstate) + return -ENOMEM; + memcpy(dst->thread.xstate, src->thread.xstate, xstate_size); + } + + return 0; +} + +void free_thread_xstate(struct task_struct *tsk) +{ + if (tsk->thread.xstate) { + kmem_cache_free(task_xstate_cachep, tsk->thread.xstate); + tsk->thread.xstate = NULL; + } +} + #if THREAD_SHIFT < PAGE_SHIFT static struct kmem_cache *thread_info_cache; @@ -20,6 +46,7 @@ struct thread_info *alloc_thread_info(struct task_struct *tsk) void free_thread_info(struct thread_info *ti) { + free_thread_xstate(ti->task); kmem_cache_free(thread_info_cache, ti); } @@ -41,6 +68,33 @@ struct thread_info *alloc_thread_info(struct task_struct *tsk) void free_thread_info(struct thread_info *ti) { + free_thread_xstate(ti->task); free_pages((unsigned long)ti, THREAD_SIZE_ORDER); } #endif /* THREAD_SHIFT < PAGE_SHIFT */ + +void arch_task_cache_init(void) +{ + if (!xstate_size) + return; + + task_xstate_cachep = kmem_cache_create("task_xstate", xstate_size, + __alignof__(union thread_xstate), + SLAB_PANIC | SLAB_NOTRACK, NULL); +} + +#ifdef CONFIG_SH_FPU_EMU +# define HAVE_SOFTFP 1 +#else +# define HAVE_SOFTFP 0 +#endif + +void init_thread_xstate(void) +{ + if (boot_cpu_data.flags & CPU_HAS_FPU) + xstate_size = sizeof(struct sh_fpu_hard_struct); + else if (HAVE_SOFTFP) + xstate_size = sizeof(struct sh_fpu_soft_struct); + else + xstate_size = 0; +} diff --git a/arch/sh/kernel/process_32.c b/arch/sh/kernel/process_32.c index c436140..03de657 100644 --- a/arch/sh/kernel/process_32.c +++ b/arch/sh/kernel/process_32.c @@ -156,6 +156,8 @@ void start_thread(struct pt_regs *regs, unsigned long new_pc, regs->sr = SR_FD; regs->pc = new_pc; regs->regs[15] = new_sp; + + free_thread_xstate(current); } EXPORT_SYMBOL(start_thread); @@ -316,7 +318,7 @@ __switch_to(struct task_struct *prev, struct task_struct *next) /* we're going to use this soon, after a few expensive things */ if (next->fpu_counter > 5) - prefetch(&next_t->fpu.hard); + prefetch(next_t->xstate); #ifdef CONFIG_MMU /* @@ -353,7 +355,7 @@ __switch_to(struct task_struct *prev, struct task_struct *next) * chances of needing FPU soon are obviously high now */ if (next->fpu_counter > 5) - fpu_state_restore(task_pt_regs(next)); + __fpu_state_restore(); return prev; } diff --git a/arch/sh/kernel/ptrace_32.c b/arch/sh/kernel/ptrace_32.c index 9be35f3..be9b5dc 100644 --- a/arch/sh/kernel/ptrace_32.c +++ b/arch/sh/kernel/ptrace_32.c @@ -163,10 +163,10 @@ int fpregs_get(struct task_struct *target, if ((boot_cpu_data.flags & CPU_HAS_FPU)) return user_regset_copyout(&pos, &count, &kbuf, &ubuf, - &target->thread.fpu.hard, 0, -1); + &target->thread.xstate->hardfpu, 0, -1); return user_regset_copyout(&pos, &count, &kbuf, &ubuf, - &target->thread.fpu.soft, 0, -1); + &target->thread.xstate->softfpu, 0, -1); } static int fpregs_set(struct task_struct *target, @@ -184,10 +184,10 @@ static int fpregs_set(struct task_struct *target, if ((boot_cpu_data.flags & CPU_HAS_FPU)) return user_regset_copyin(&pos, &count, &kbuf, &ubuf, - &target->thread.fpu.hard, 0, -1); + &target->thread.xstate->hardfpu, 0, -1); return user_regset_copyin(&pos, &count, &kbuf, &ubuf, - &target->thread.fpu.soft, 0, -1); + &target->thread.xstate->softfpu, 0, -1); } static int fpregs_active(struct task_struct *target, @@ -333,7 +333,7 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) else tmp = 0; } else - tmp = ((long *)&child->thread.fpu) + tmp = ((long *)child->thread.xstate) [(addr - (long)&dummy->fpu) >> 2]; } else if (addr == (long) &dummy->u_fpvalid) tmp = !!tsk_used_math(child); @@ -362,7 +362,7 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) else if (addr >= (long) &dummy->fpu && addr < (long) &dummy->u_fpvalid) { set_stopped_child_used_math(child); - ((long *)&child->thread.fpu) + ((long *)child->thread.xstate) [(addr - (long)&dummy->fpu) >> 2] = data; ret = 0; } else if (addr == (long) &dummy->u_fpvalid) { diff --git a/arch/sh/kernel/signal_32.c b/arch/sh/kernel/signal_32.c index 12815ce..6a7cce7 100644 --- a/arch/sh/kernel/signal_32.c +++ b/arch/sh/kernel/signal_32.c @@ -150,7 +150,7 @@ static inline int restore_sigcontext_fpu(struct sigcontext __user *sc) return 0; set_used_math(); - return __copy_from_user(&tsk->thread.fpu.hard, &sc->sc_fpregs[0], + return __copy_from_user(&tsk->thread.xstate->hardfpu, &sc->sc_fpregs[0], sizeof(long)*(16*2+2)); } @@ -175,7 +175,7 @@ static inline int save_sigcontext_fpu(struct sigcontext __user *sc, clear_used_math(); unlazy_fpu(tsk, regs); - return __copy_to_user(&sc->sc_fpregs[0], &tsk->thread.fpu.hard, + return __copy_to_user(&sc->sc_fpregs[0], &tsk->thread.xstate->hardfpu, sizeof(long)*(16*2+2)); } #endif /* CONFIG_SH_FPU */ diff --git a/arch/sh/math-emu/math.c b/arch/sh/math-emu/math.c index d6c15ca..1fcdb12 100644 --- a/arch/sh/math-emu/math.c +++ b/arch/sh/math-emu/math.c @@ -471,10 +471,10 @@ static int fpu_emulate(u16 code, struct sh_fpu_soft_struct *fregs, struct pt_reg * denormal_to_double - Given denormalized float number, * store double float * - * @fpu: Pointer to sh_fpu_hard structure + * @fpu: Pointer to sh_fpu_soft structure * @n: Index to FP register */ -static void denormal_to_double(struct sh_fpu_hard_struct *fpu, int n) +static void denormal_to_double(struct sh_fpu_soft_struct *fpu, int n) { unsigned long du, dl; unsigned long x = fpu->fpul; @@ -552,11 +552,11 @@ static int ieee_fpe_handler(struct pt_regs *regs) if ((finsn & 0xf1ff) == 0xf0ad) { /* fcnvsd */ struct task_struct *tsk = current; - if ((tsk->thread.fpu.hard.fpscr & (1 << 17))) { + if ((tsk->thread.xstate->softfpu.fpscr & (1 << 17))) { /* FPU error */ - denormal_to_double (&tsk->thread.fpu.hard, + denormal_to_double (&tsk->thread.xstate->softfpu, (finsn >> 8) & 0xf); - tsk->thread.fpu.hard.fpscr &= + tsk->thread.xstate->softfpu.fpscr &= ~(FPSCR_CAUSE_MASK | FPSCR_FLAG_MASK); task_thread_info(tsk)->status |= TS_USEDFPU; } else { @@ -617,7 +617,7 @@ static void fpu_init(struct sh_fpu_soft_struct *fpu) int do_fpu_inst(unsigned short inst, struct pt_regs *regs) { struct task_struct *tsk = current; - struct sh_fpu_soft_struct *fpu = &(tsk->thread.fpu.soft); + struct sh_fpu_soft_struct *fpu = &(tsk->thread.xstate->softfpu); if (!(task_thread_info(tsk)->status & TS_USEDFPU)) { /* initialize once. */ -- cgit v0.10.2 From 6430a5987f8bb9a7c5187b343956aa8ef404d571 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Wed, 13 Jan 2010 12:59:24 +0900 Subject: sh: Don't perform an icbi on a P2 address The legacy P2 area may not always be mapped (for example when using PMB). So perform an icbi on an address that we know will always be mapped. Signed-off-by: Matt Fleming Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/system.h b/arch/sh/include/asm/system.h index c15415b..b74d253 100644 --- a/arch/sh/include/asm/system.h +++ b/arch/sh/include/asm/system.h @@ -32,7 +32,7 @@ #define mb() __asm__ __volatile__ ("synco": : :"memory") #define rmb() mb() #define wmb() __asm__ __volatile__ ("synco": : :"memory") -#define ctrl_barrier() __icbi(0xa8000000) +#define ctrl_barrier() __icbi(PAGE_OFFSET) #define read_barrier_depends() do { } while(0) #else #define mb() __asm__ __volatile__ ("": : :"memory") -- cgit v0.10.2 From c7b16efb7d0c7fce218b2cdafa2432c5fbf57314 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 13 Jan 2010 13:29:19 +0900 Subject: sh: Add support for LZO-compressed kernels. Plugs in LZO along with the others. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 4eaf340..5ac0115 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -22,6 +22,7 @@ config SUPERH select HAVE_KERNEL_GZIP select HAVE_KERNEL_BZIP2 select HAVE_KERNEL_LZMA + select HAVE_KERNEL_LZO select HAVE_SYSCALL_TRACEPOINTS select RTC_LIB select GENERIC_ATOMIC64 diff --git a/arch/sh/Makefile b/arch/sh/Makefile index db91925..fd57676 100644 --- a/arch/sh/Makefile +++ b/arch/sh/Makefile @@ -203,8 +203,8 @@ endif libs-$(CONFIG_SUPERH32) := arch/sh/lib/ $(libs-y) libs-$(CONFIG_SUPERH64) := arch/sh/lib64/ $(libs-y) -BOOT_TARGETS = uImage uImage.bz2 uImage.gz uImage.lzma uImage.srec uImage.bin \ - zImage vmlinux.srec romImage +BOOT_TARGETS = uImage uImage.bz2 uImage.gz uImage.lzma uImage.lzo \ + uImage.srec uImage.bin zImage vmlinux.srec romImage PHONY += $(BOOT_TARGETS) all: $(KBUILD_IMAGE) @@ -231,4 +231,5 @@ define archhelp @echo '* uImage.gz - Kernel-only image for U-Boot (gzip)' @echo ' uImage.bz2 - Kernel-only image for U-Boot (bzip2)' @echo ' uImage.lzma - Kernel-only image for U-Boot (lzma)' + @echo ' uImage.lzo - Kernel-only image for U-Boot (lzo)' endef diff --git a/arch/sh/boot/Makefile b/arch/sh/boot/Makefile index cb8cf55..743ce0c 100644 --- a/arch/sh/boot/Makefile +++ b/arch/sh/boot/Makefile @@ -21,12 +21,15 @@ CONFIG_ZERO_PAGE_OFFSET ?= 0x00001000 CONFIG_ENTRY_OFFSET ?= 0x00001000 suffix-y := bin -suffix-$(CONFIG_KERNEL_GZIP) := gz -suffix-$(CONFIG_KERNEL_BZIP2) := bz2 -suffix-$(CONFIG_KERNEL_LZMA) := lzma - -targets := zImage vmlinux.srec romImage uImage uImage.srec uImage.gz uImage.bz2 uImage.lzma uImage.bin -extra-y += vmlinux.bin vmlinux.bin.gz vmlinux.bin.bz2 vmlinux.bin.lzma +suffix-$(CONFIG_KERNEL_GZIP) := gz +suffix-$(CONFIG_KERNEL_BZIP2) := bz2 +suffix-$(CONFIG_KERNEL_LZMA) := lzma +suffix-$(CONFIG_KERNEL_LZO) := lzo + +targets := zImage vmlinux.srec romImage uImage uImage.srec uImage.gz \ + uImage.bz2 uImage.lzma uImage.lzo uImage.bin +extra-y += vmlinux.bin vmlinux.bin.gz vmlinux.bin.bz2 vmlinux.bin.lzma \ + vmlinux.bin.lzo subdir- := compressed romimage $(obj)/zImage: $(obj)/compressed/vmlinux FORCE @@ -80,6 +83,9 @@ $(obj)/vmlinux.bin.bz2: $(obj)/vmlinux.bin FORCE $(obj)/vmlinux.bin.lzma: $(obj)/vmlinux.bin FORCE $(call if_changed,lzma) +$(obj)/vmlinux.bin.lzo: $(obj)/vmlinux.bin FORCE + $(call if_changed,lzo) + $(obj)/uImage.bz2: $(obj)/vmlinux.bin.bz2 $(call if_changed,uimage,bzip2) @@ -89,6 +95,9 @@ $(obj)/uImage.gz: $(obj)/vmlinux.bin.gz $(obj)/uImage.lzma: $(obj)/vmlinux.bin.lzma $(call if_changed,uimage,lzma) +$(obj)/uImage.lzo: $(obj)/vmlinux.bin.lzo + $(call if_changed,uimage,lzo) + $(obj)/uImage.bin: $(obj)/vmlinux.bin $(call if_changed,uimage,none) diff --git a/arch/sh/boot/compressed/Makefile b/arch/sh/boot/compressed/Makefile index 6182eca..6b95a2a 100644 --- a/arch/sh/boot/compressed/Makefile +++ b/arch/sh/boot/compressed/Makefile @@ -6,6 +6,7 @@ targets := vmlinux vmlinux.bin vmlinux.bin.gz \ vmlinux.bin.bz2 vmlinux.bin.lzma \ + vmlinux.bin.lzo \ head_$(BITS).o misc.o piggy.o OBJECTS = $(obj)/head_$(BITS).o $(obj)/misc.o $(obj)/cache.o @@ -47,6 +48,8 @@ $(obj)/vmlinux.bin.bz2: $(vmlinux.bin.all-y) FORCE $(call if_changed,bzip2) $(obj)/vmlinux.bin.lzma: $(vmlinux.bin.all-y) FORCE $(call if_changed,lzma) +$(obj)/vmlinux.bin.lzo: $(vmlinux.bin.all-y) FORCE + $(call if_changed,lzo) OBJCOPYFLAGS += -R .empty_zero_page diff --git a/arch/sh/boot/compressed/misc.c b/arch/sh/boot/compressed/misc.c index b51b1fc..4b319e9 100644 --- a/arch/sh/boot/compressed/misc.c +++ b/arch/sh/boot/compressed/misc.c @@ -62,6 +62,10 @@ static unsigned long free_mem_end_ptr; #include "../../../../lib/decompress_unlzma.c" #endif +#ifdef CONFIG_KERNEL_LZO +#include "../../../../lib/decompress_unlzo.c" +#endif + #ifdef CONFIG_SH_STANDARD_BIOS size_t strlen(const char *s) { -- cgit v0.10.2 From 59c33fa7791e9948ba467c2b83e307a0d087ab49 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Tue, 12 Jan 2010 16:21:09 -0800 Subject: x86-32: clean up rwsem inline asm statements This makes gcc use the right register names and instruction operand sizes automatically for the rwsem inline asm statements. So instead of using "(%%eax)" to specify the memory address that is the semaphore, we use "(%1)" or similar. And instead of forcing the operation to always be 32-bit, we use "%z0", taking the size from the actual semaphore data structure itself. This doesn't actually matter on x86-32, but if we want to use the same inline asm for x86-64, we'll need to have the compiler generate the proper 64-bit names for the registers (%rax instead of %eax), and if we want to use a 64-bit counter too (in order to avoid the 15-bit limit on the write counter that limits concurrent users to 32767 threads), we'll need to be able to generate instructions with "q" accesses rather than "l". Since this header currently isn't enabled on x86-64, none of that matters, but we do want to use the xadd version of the semaphores rather than have to take spinlocks to do a rwsem. The mm->mmap_sem can be heavily contended when you have lots of threads all taking page faults, and the fallback rwsem code that uses a spinlock performs abysmally badly in that case. [ hpa: modified the patch to skip size suffixes entirely when they are redundant due to register operands. ] Signed-off-by: Linus Torvalds LKML-Reference: Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/rwsem.h b/arch/x86/include/asm/rwsem.h index ca7517d..4136200 100644 --- a/arch/x86/include/asm/rwsem.h +++ b/arch/x86/include/asm/rwsem.h @@ -105,7 +105,7 @@ do { \ static inline void __down_read(struct rw_semaphore *sem) { asm volatile("# beginning down_read\n\t" - LOCK_PREFIX " incl (%%eax)\n\t" + LOCK_PREFIX " inc%z0 (%1)\n\t" /* adds 0x00000001, returns the old value */ " jns 1f\n" " call call_rwsem_down_read_failed\n" @@ -123,12 +123,12 @@ static inline int __down_read_trylock(struct rw_semaphore *sem) { __s32 result, tmp; asm volatile("# beginning __down_read_trylock\n\t" - " movl %0,%1\n\t" + " mov %0,%1\n\t" "1:\n\t" - " movl %1,%2\n\t" - " addl %3,%2\n\t" + " mov %1,%2\n\t" + " add %3,%2\n\t" " jle 2f\n\t" - LOCK_PREFIX " cmpxchgl %2,%0\n\t" + LOCK_PREFIX " cmpxchg %2,%0\n\t" " jnz 1b\n\t" "2:\n\t" "# ending __down_read_trylock\n\t" @@ -147,9 +147,9 @@ static inline void __down_write_nested(struct rw_semaphore *sem, int subclass) tmp = RWSEM_ACTIVE_WRITE_BIAS; asm volatile("# beginning down_write\n\t" - LOCK_PREFIX " xadd %%edx,(%%eax)\n\t" + LOCK_PREFIX " xadd %1,(%2)\n\t" /* subtract 0x0000ffff, returns the old value */ - " testl %%edx,%%edx\n\t" + " test %1,%1\n\t" /* was the count 0 before? */ " jz 1f\n" " call call_rwsem_down_write_failed\n" @@ -185,7 +185,7 @@ static inline void __up_read(struct rw_semaphore *sem) { __s32 tmp = -RWSEM_ACTIVE_READ_BIAS; asm volatile("# beginning __up_read\n\t" - LOCK_PREFIX " xadd %%edx,(%%eax)\n\t" + LOCK_PREFIX " xadd %1,(%2)\n\t" /* subtracts 1, returns the old value */ " jns 1f\n\t" " call call_rwsem_wake\n" @@ -201,18 +201,18 @@ static inline void __up_read(struct rw_semaphore *sem) */ static inline void __up_write(struct rw_semaphore *sem) { + unsigned long tmp; asm volatile("# beginning __up_write\n\t" - " movl %2,%%edx\n\t" - LOCK_PREFIX " xaddl %%edx,(%%eax)\n\t" + LOCK_PREFIX " xadd %1,(%2)\n\t" /* tries to transition 0xffff0001 -> 0x00000000 */ " jz 1f\n" " call call_rwsem_wake\n" "1:\n\t" "# ending __up_write\n" - : "+m" (sem->count) - : "a" (sem), "i" (-RWSEM_ACTIVE_WRITE_BIAS) - : "memory", "cc", "edx"); + : "+m" (sem->count), "=d" (tmp) + : "a" (sem), "1" (-RWSEM_ACTIVE_WRITE_BIAS) + : "memory", "cc"); } /* @@ -221,7 +221,7 @@ static inline void __up_write(struct rw_semaphore *sem) static inline void __downgrade_write(struct rw_semaphore *sem) { asm volatile("# beginning __downgrade_write\n\t" - LOCK_PREFIX " addl %2,(%%eax)\n\t" + LOCK_PREFIX " add%z0 %2,(%1)\n\t" /* transitions 0xZZZZ0001 -> 0xYYYY0001 */ " jns 1f\n\t" " call call_rwsem_downgrade_wake\n" @@ -237,7 +237,7 @@ static inline void __downgrade_write(struct rw_semaphore *sem) */ static inline void rwsem_atomic_add(int delta, struct rw_semaphore *sem) { - asm volatile(LOCK_PREFIX "addl %1,%0" + asm volatile(LOCK_PREFIX "add%z0 %1,%0" : "+m" (sem->count) : "ir" (delta)); } -- cgit v0.10.2 From eca50f14b88315bd31183a17c89e53534f2ed2fa Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 13 Jan 2010 16:28:47 +0900 Subject: sh: Add a vmlinux.bin target. This makes vmlinux.bin generation an explicit make target, as opposed to just a dependency for some of the other targets. Signed-off-by: Paul Mundt diff --git a/arch/sh/Makefile b/arch/sh/Makefile index fd57676..f7d5f42 100644 --- a/arch/sh/Makefile +++ b/arch/sh/Makefile @@ -204,7 +204,8 @@ libs-$(CONFIG_SUPERH32) := arch/sh/lib/ $(libs-y) libs-$(CONFIG_SUPERH64) := arch/sh/lib64/ $(libs-y) BOOT_TARGETS = uImage uImage.bz2 uImage.gz uImage.lzma uImage.lzo \ - uImage.srec uImage.bin zImage vmlinux.srec romImage + uImage.srec uImage.bin zImage vmlinux.bin vmlinux.srec \ + romImage PHONY += $(BOOT_TARGETS) all: $(KBUILD_IMAGE) @@ -225,6 +226,7 @@ define archhelp @echo ' zImage - Compressed kernel image' @echo ' romImage - Compressed ROM image, if supported' @echo ' vmlinux.srec - Create an ELF S-record' + @echo ' vmlinux.bin - Create an uncompressed binary image' @echo '* uImage - Alias to bootable U-Boot image' @echo ' uImage.srec - Create an S-record for U-Boot' @echo ' uImage.bin - Kernel-only image for U-Boot (bin)' -- cgit v0.10.2 From 7f33306ee57bce9c79825e89c457a91025aa5aad Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Wed, 13 Jan 2010 16:35:58 +0900 Subject: sh: PVR detection for 2nd cut SH7786. The mass produced cuts use an updated PVR value, add them to the list. Signed-off-by: Matt Fleming Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/cpu/sh4/probe.c b/arch/sh/kernel/cpu/sh4/probe.c index d36f0c4..75189c4 100644 --- a/arch/sh/kernel/cpu/sh4/probe.c +++ b/arch/sh/kernel/cpu/sh4/probe.c @@ -124,6 +124,7 @@ int __init detect_cpu_and_cache_system(void) boot_cpu_data.type = CPU_SH7785; break; case 0x4004: + case 0x4005: boot_cpu_data.type = CPU_SH7786; boot_cpu_data.flags |= CPU_HAS_PTEAEX | CPU_HAS_L2_CACHE; break; -- cgit v0.10.2 From 2ca49b2fcf5813571663c3c4c894b78148c43690 Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Mon, 4 Jan 2010 09:47:35 -0500 Subject: x86: Macroise x86 cache descriptors Use a macro to define the cache sizes when cachesize > 1 MB. This is less typing, and less prone to introducing bugs like we saw in e02e0e1a130b9ca37c5186d38ad4b3aaf58bb149, and means we don't have to do maths when adding new non-power-of-2 updates like those seen recently. Signed-off-by: Dave Jones LKML-Reference: <20100104144735.GA18390@redhat.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index fc6c8ef..c2b722d 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -31,6 +31,8 @@ struct _cache_table { short size; }; +#define MB(x) ((x) * 1024) + /* All the cache descriptor types we care about (no TLB or trace cache entries) */ @@ -44,9 +46,9 @@ static const struct _cache_table __cpuinitconst cache_table[] = { 0x0d, LVL_1_DATA, 16 }, /* 4-way set assoc, 64 byte line size */ { 0x21, LVL_2, 256 }, /* 8-way set assoc, 64 byte line size */ { 0x22, LVL_3, 512 }, /* 4-way set assoc, sectored cache, 64 byte line size */ - { 0x23, LVL_3, 1024 }, /* 8-way set assoc, sectored cache, 64 byte line size */ - { 0x25, LVL_3, 2048 }, /* 8-way set assoc, sectored cache, 64 byte line size */ - { 0x29, LVL_3, 4096 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x23, LVL_3, MB(1) }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x25, LVL_3, MB(2) }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x29, LVL_3, MB(4) }, /* 8-way set assoc, sectored cache, 64 byte line size */ { 0x2c, LVL_1_DATA, 32 }, /* 8-way set assoc, 64 byte line size */ { 0x30, LVL_1_INST, 32 }, /* 8-way set assoc, 64 byte line size */ { 0x39, LVL_2, 128 }, /* 4-way set assoc, sectored cache, 64 byte line size */ @@ -59,16 +61,16 @@ static const struct _cache_table __cpuinitconst cache_table[] = { 0x41, LVL_2, 128 }, /* 4-way set assoc, 32 byte line size */ { 0x42, LVL_2, 256 }, /* 4-way set assoc, 32 byte line size */ { 0x43, LVL_2, 512 }, /* 4-way set assoc, 32 byte line size */ - { 0x44, LVL_2, 1024 }, /* 4-way set assoc, 32 byte line size */ - { 0x45, LVL_2, 2048 }, /* 4-way set assoc, 32 byte line size */ - { 0x46, LVL_3, 4096 }, /* 4-way set assoc, 64 byte line size */ - { 0x47, LVL_3, 8192 }, /* 8-way set assoc, 64 byte line size */ - { 0x49, LVL_3, 4096 }, /* 16-way set assoc, 64 byte line size */ - { 0x4a, LVL_3, 6144 }, /* 12-way set assoc, 64 byte line size */ - { 0x4b, LVL_3, 8192 }, /* 16-way set assoc, 64 byte line size */ - { 0x4c, LVL_3, 12288 }, /* 12-way set assoc, 64 byte line size */ - { 0x4d, LVL_3, 16384 }, /* 16-way set assoc, 64 byte line size */ - { 0x4e, LVL_2, 6144 }, /* 24-way set assoc, 64 byte line size */ + { 0x44, LVL_2, MB(1) }, /* 4-way set assoc, 32 byte line size */ + { 0x45, LVL_2, MB(2) }, /* 4-way set assoc, 32 byte line size */ + { 0x46, LVL_3, MB(4) }, /* 4-way set assoc, 64 byte line size */ + { 0x47, LVL_3, MB(8) }, /* 8-way set assoc, 64 byte line size */ + { 0x49, LVL_3, MB(4) }, /* 16-way set assoc, 64 byte line size */ + { 0x4a, LVL_3, MB(6) }, /* 12-way set assoc, 64 byte line size */ + { 0x4b, LVL_3, MB(8) }, /* 16-way set assoc, 64 byte line size */ + { 0x4c, LVL_3, MB(12) }, /* 12-way set assoc, 64 byte line size */ + { 0x4d, LVL_3, MB(16) }, /* 16-way set assoc, 64 byte line size */ + { 0x4e, LVL_2, MB(6) }, /* 24-way set assoc, 64 byte line size */ { 0x60, LVL_1_DATA, 16 }, /* 8-way set assoc, sectored cache, 64 byte line size */ { 0x66, LVL_1_DATA, 8 }, /* 4-way set assoc, sectored cache, 64 byte line size */ { 0x67, LVL_1_DATA, 16 }, /* 4-way set assoc, sectored cache, 64 byte line size */ @@ -77,34 +79,34 @@ static const struct _cache_table __cpuinitconst cache_table[] = { 0x71, LVL_TRACE, 16 }, /* 8-way set assoc */ { 0x72, LVL_TRACE, 32 }, /* 8-way set assoc */ { 0x73, LVL_TRACE, 64 }, /* 8-way set assoc */ - { 0x78, LVL_2, 1024 }, /* 4-way set assoc, 64 byte line size */ - { 0x79, LVL_2, 128 }, /* 8-way set assoc, sectored cache, 64 byte line size */ - { 0x7a, LVL_2, 256 }, /* 8-way set assoc, sectored cache, 64 byte line size */ - { 0x7b, LVL_2, 512 }, /* 8-way set assoc, sectored cache, 64 byte line size */ - { 0x7c, LVL_2, 1024 }, /* 8-way set assoc, sectored cache, 64 byte line size */ - { 0x7d, LVL_2, 2048 }, /* 8-way set assoc, 64 byte line size */ - { 0x7f, LVL_2, 512 }, /* 2-way set assoc, 64 byte line size */ - { 0x82, LVL_2, 256 }, /* 8-way set assoc, 32 byte line size */ - { 0x83, LVL_2, 512 }, /* 8-way set assoc, 32 byte line size */ - { 0x84, LVL_2, 1024 }, /* 8-way set assoc, 32 byte line size */ - { 0x85, LVL_2, 2048 }, /* 8-way set assoc, 32 byte line size */ - { 0x86, LVL_2, 512 }, /* 4-way set assoc, 64 byte line size */ - { 0x87, LVL_2, 1024 }, /* 8-way set assoc, 64 byte line size */ - { 0xd0, LVL_3, 512 }, /* 4-way set assoc, 64 byte line size */ - { 0xd1, LVL_3, 1024 }, /* 4-way set assoc, 64 byte line size */ - { 0xd2, LVL_3, 2048 }, /* 4-way set assoc, 64 byte line size */ - { 0xd6, LVL_3, 1024 }, /* 8-way set assoc, 64 byte line size */ - { 0xd7, LVL_3, 2048 }, /* 8-way set assoc, 64 byte line size */ - { 0xd8, LVL_3, 4096 }, /* 12-way set assoc, 64 byte line size */ - { 0xdc, LVL_3, 2048 }, /* 12-way set assoc, 64 byte line size */ - { 0xdd, LVL_3, 4096 }, /* 12-way set assoc, 64 byte line size */ - { 0xde, LVL_3, 8192 }, /* 12-way set assoc, 64 byte line size */ - { 0xe2, LVL_3, 2048 }, /* 16-way set assoc, 64 byte line size */ - { 0xe3, LVL_3, 4096 }, /* 16-way set assoc, 64 byte line size */ - { 0xe4, LVL_3, 8192 }, /* 16-way set assoc, 64 byte line size */ - { 0xea, LVL_3, 12288 }, /* 24-way set assoc, 64 byte line size */ - { 0xeb, LVL_3, 18432 }, /* 24-way set assoc, 64 byte line size */ - { 0xec, LVL_3, 24576 }, /* 24-way set assoc, 64 byte line size */ + { 0x78, LVL_2, MB(1) }, /* 4-way set assoc, 64 byte line size */ + { 0x79, LVL_2, 128 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x7a, LVL_2, 256 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x7b, LVL_2, 512 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x7c, LVL_2, MB(1) }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x7d, LVL_2, MB(2) }, /* 8-way set assoc, 64 byte line size */ + { 0x7f, LVL_2, 512 }, /* 2-way set assoc, 64 byte line size */ + { 0x82, LVL_2, 256 }, /* 8-way set assoc, 32 byte line size */ + { 0x83, LVL_2, 512 }, /* 8-way set assoc, 32 byte line size */ + { 0x84, LVL_2, MB(1) }, /* 8-way set assoc, 32 byte line size */ + { 0x85, LVL_2, MB(2) }, /* 8-way set assoc, 32 byte line size */ + { 0x86, LVL_2, 512 }, /* 4-way set assoc, 64 byte line size */ + { 0x87, LVL_2, MB(1) }, /* 8-way set assoc, 64 byte line size */ + { 0xd0, LVL_3, 512 }, /* 4-way set assoc, 64 byte line size */ + { 0xd1, LVL_3, MB(1) }, /* 4-way set assoc, 64 byte line size */ + { 0xd2, LVL_3, MB(2) }, /* 4-way set assoc, 64 byte line size */ + { 0xd6, LVL_3, MB(1) }, /* 8-way set assoc, 64 byte line size */ + { 0xd7, LVL_3, MB(2) }, /* 8-way set assoc, 64 byte line size */ + { 0xd8, LVL_3, MB(4) }, /* 12-way set assoc, 64 byte line size */ + { 0xdc, LVL_3, MB(2) }, /* 12-way set assoc, 64 byte line size */ + { 0xdd, LVL_3, MB(4) }, /* 12-way set assoc, 64 byte line size */ + { 0xde, LVL_3, MB(8) }, /* 12-way set assoc, 64 byte line size */ + { 0xe2, LVL_3, MB(2) }, /* 16-way set assoc, 64 byte line size */ + { 0xe3, LVL_3, MB(4) }, /* 16-way set assoc, 64 byte line size */ + { 0xe4, LVL_3, MB(8) }, /* 16-way set assoc, 64 byte line size */ + { 0xea, LVL_3, MB(12) }, /* 24-way set assoc, 64 byte line size */ + { 0xeb, LVL_3, MB(18) }, /* 24-way set assoc, 64 byte line size */ + { 0xec, LVL_3, MB(24) }, /* 24-way set assoc, 64 byte line size */ { 0x00, 0, 0} }; -- cgit v0.10.2 From f96e9232e04856c781d4f71923a46dd3f7b429fa Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:00 -0800 Subject: rcu: Adjust force_quiescent_state() locking, step 1 This causes rnp->lock to be held on entry to force_quiescent_state()'s switch statement. This is a first step towards prohibiting starting grace periods while force_quiescent_state() is executing, which will reduce the number and complexity of races that force_quiescent_state() is involved in. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465501455-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 53ae959..eae331d 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1204,7 +1204,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) } if (relaxed && (long)(rsp->jiffies_force_qs - jiffies) >= 0) - goto unlock_ret; /* no emergency and done recently. */ + goto unlock_fqs_ret; /* no emergency and done recently. */ rsp->n_force_qs++; spin_lock(&rnp->lock); lastcomp = rsp->gpnum - 1; @@ -1213,31 +1213,32 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) if(!rcu_gp_in_progress(rsp)) { rsp->n_force_qs_ngp++; spin_unlock(&rnp->lock); - goto unlock_ret; /* no GP in progress, time updated. */ + goto unlock_fqs_ret; /* no GP in progress, time updated. */ } - spin_unlock(&rnp->lock); switch (signaled) { case RCU_GP_IDLE: case RCU_GP_INIT: + spin_unlock(&rnp->lock); break; /* grace period idle or initializing, ignore. */ case RCU_SAVE_DYNTICK: + spin_unlock(&rnp->lock); if (RCU_SIGNAL_INIT != RCU_SAVE_DYNTICK) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ if (rcu_process_dyntick(rsp, lastcomp, dyntick_save_progress_counter)) - goto unlock_ret; + goto unlock_fqs_ret; + spin_lock(&rnp->lock); /* fall into next case. */ case RCU_SAVE_COMPLETED: /* Update state, record completion counter. */ forcenow = 0; - spin_lock(&rnp->lock); if (lastcomp + 1 == rsp->gpnum && lastcomp == rsp->completed && rsp->signaled == signaled) { @@ -1245,23 +1246,31 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) rsp->completed_fqs = lastcomp; forcenow = signaled == RCU_SAVE_COMPLETED; } - spin_unlock(&rnp->lock); - if (!forcenow) + if (!forcenow) { + spin_unlock(&rnp->lock); break; + } /* fall into next case. */ case RCU_FORCE_QS: /* Check dyntick-idle state, send IPI to laggarts. */ + spin_unlock(&rnp->lock); if (rcu_process_dyntick(rsp, rsp->completed_fqs, rcu_implicit_dynticks_qs)) - goto unlock_ret; + goto unlock_fqs_ret; /* Leave state in case more forcing is required. */ break; + + default: + + spin_unlock(&rnp->lock); + WARN_ON_ONCE(1); + break; } -unlock_ret: +unlock_fqs_ret: spin_unlock_irqrestore(&rsp->fqslock, flags); } -- cgit v0.10.2 From 559569acf94f538b56bd6eead80b439d6a78cdff Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:01 -0800 Subject: rcu: Adjust force_quiescent_state() locking, step 2 This patch releases rnp->lock after the end of force_quiescent_state()'s switch statement. This is a second step towards prohibiting starting grace periods while force_quiescent_state() is executing, which will reduce the number and complexity of races that force_quiescent_state() is involved in. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465501994-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index eae331d..d42ad30 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1219,7 +1219,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) case RCU_GP_IDLE: case RCU_GP_INIT: - spin_unlock(&rnp->lock); break; /* grace period idle or initializing, ignore. */ case RCU_SAVE_DYNTICK: @@ -1246,10 +1245,8 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) rsp->completed_fqs = lastcomp; forcenow = signaled == RCU_SAVE_COMPLETED; } - if (!forcenow) { - spin_unlock(&rnp->lock); + if (!forcenow) break; - } /* fall into next case. */ case RCU_FORCE_QS: @@ -1262,14 +1259,10 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Leave state in case more forcing is required. */ - break; - - default: - - spin_unlock(&rnp->lock); - WARN_ON_ONCE(1); + spin_lock(&rnp->lock); break; } + spin_unlock(&rnp->lock); unlock_fqs_ret: spin_unlock_irqrestore(&rsp->fqslock, flags); } -- cgit v0.10.2 From 07079d5357a4d53c2b13126c4a38fb40e6e04966 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:02 -0800 Subject: rcu: Prohibit starting new grace periods while forcing quiescent states Reduce the number and variety of race conditions by prohibiting the start of a new grace period while force_quiescent_state() is active. A new fqs_active flag in the rcu_state structure is used to trace whether or not force_quiescent_state() is active, and this new flag is tested by rcu_start_gp(). If the CPU that closed out the last grace period needs another grace period, this new grace period may be delayed up to one scheduling-clock tick, but it will eventually get started. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <126264655052-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index d42ad30..41688ff 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -659,7 +659,7 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) struct rcu_data *rdp = rsp->rda[smp_processor_id()]; struct rcu_node *rnp = rcu_get_root(rsp); - if (!cpu_needs_another_gp(rsp, rdp)) { + if (!cpu_needs_another_gp(rsp, rdp) || rsp->fqs_active) { if (rnp->completed == rsp->completed) { spin_unlock_irqrestore(&rnp->lock, flags); return; @@ -1195,6 +1195,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) struct rcu_node *rnp = rcu_get_root(rsp); u8 signaled; u8 forcenow; + u8 gpdone; if (!rcu_gp_in_progress(rsp)) return; /* No grace period in progress, nothing to force. */ @@ -1206,15 +1207,16 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) (long)(rsp->jiffies_force_qs - jiffies) >= 0) goto unlock_fqs_ret; /* no emergency and done recently. */ rsp->n_force_qs++; - spin_lock(&rnp->lock); + spin_lock(&rnp->lock); /* irqs already disabled */ lastcomp = rsp->gpnum - 1; signaled = rsp->signaled; rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; if(!rcu_gp_in_progress(rsp)) { rsp->n_force_qs_ngp++; - spin_unlock(&rnp->lock); + spin_unlock(&rnp->lock); /* irqs remain disabled */ goto unlock_fqs_ret; /* no GP in progress, time updated. */ } + rsp->fqs_active = 1; switch (signaled) { case RCU_GP_IDLE: case RCU_GP_INIT: @@ -1223,15 +1225,16 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) case RCU_SAVE_DYNTICK: - spin_unlock(&rnp->lock); + spin_unlock(&rnp->lock); /* irqs remain disabled */ if (RCU_SIGNAL_INIT != RCU_SAVE_DYNTICK) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - if (rcu_process_dyntick(rsp, lastcomp, - dyntick_save_progress_counter)) - goto unlock_fqs_ret; - spin_lock(&rnp->lock); + gpdone = rcu_process_dyntick(rsp, lastcomp, + dyntick_save_progress_counter); + spin_lock(&rnp->lock); /* irqs already disabled */ + if (gpdone) + break; /* fall into next case. */ case RCU_SAVE_COMPLETED: @@ -1252,17 +1255,17 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) case RCU_FORCE_QS: /* Check dyntick-idle state, send IPI to laggarts. */ - spin_unlock(&rnp->lock); - if (rcu_process_dyntick(rsp, rsp->completed_fqs, - rcu_implicit_dynticks_qs)) - goto unlock_fqs_ret; + spin_unlock(&rnp->lock); /* irqs remain disabled */ + gpdone = rcu_process_dyntick(rsp, rsp->completed_fqs, + rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ - spin_lock(&rnp->lock); + spin_lock(&rnp->lock); /* irqs already disabled */ break; } - spin_unlock(&rnp->lock); + rsp->fqs_active = 0; + spin_unlock(&rnp->lock); /* irqs remain disabled */ unlock_fqs_ret: spin_unlock_irqrestore(&rsp->fqslock, flags); } diff --git a/kernel/rcutree.h b/kernel/rcutree.h index d2a0046..dc386a7 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -277,6 +277,8 @@ struct rcu_state { u8 signaled ____cacheline_internodealigned_in_smp; /* Force QS state. */ + u8 fqs_active; /* force_quiescent_state() */ + /* is running. */ long gpnum; /* Current gp number. */ long completed; /* # of last completed gp. */ -- cgit v0.10.2 From f3a8b5c6aa543bd87764418d63632eb65b80e2f6 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:03 -0800 Subject: rcu: Eliminate local variable signaled from force_quiescent_state() Because the root rcu_node lock is held across entry to the switch statement in force_quiescent_state(), it is no longer necessary to snapshot rsp->signaled to a local variable. Eliminate both the snapshotting and the local variable. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1262646550602-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 41688ff..1d8cfb1 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1193,7 +1193,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) unsigned long flags; long lastcomp; struct rcu_node *rnp = rcu_get_root(rsp); - u8 signaled; u8 forcenow; u8 gpdone; @@ -1209,7 +1208,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) rsp->n_force_qs++; spin_lock(&rnp->lock); /* irqs already disabled */ lastcomp = rsp->gpnum - 1; - signaled = rsp->signaled; rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; if(!rcu_gp_in_progress(rsp)) { rsp->n_force_qs_ngp++; @@ -1217,7 +1215,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) goto unlock_fqs_ret; /* no GP in progress, time updated. */ } rsp->fqs_active = 1; - switch (signaled) { + switch (rsp->signaled) { case RCU_GP_IDLE: case RCU_GP_INIT: @@ -1242,11 +1240,10 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Update state, record completion counter. */ forcenow = 0; if (lastcomp + 1 == rsp->gpnum && - lastcomp == rsp->completed && - rsp->signaled == signaled) { + lastcomp == rsp->completed) { + forcenow = rsp->signaled == RCU_SAVE_COMPLETED; rsp->signaled = RCU_FORCE_QS; rsp->completed_fqs = lastcomp; - forcenow = signaled == RCU_SAVE_COMPLETED; } if (!forcenow) break; -- cgit v0.10.2 From 39c0bbfc07c6e28db7346d0e11106f2d045d3035 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:04 -0800 Subject: rcu: Eliminate local variable lastcomp from force_quiescent_state() Because rsp->fqs_active is set to 1 across force_quiescent_state()'s switch statement, rcu_start_gp() will refrain from starting a new grace period during this time. Therefore, rsp->gpnum is constant, and can be propagated to all uses of lastcomp, eliminating this local variable. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465502985-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 1d8cfb1..62b6433 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1191,7 +1191,6 @@ static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, static void force_quiescent_state(struct rcu_state *rsp, int relaxed) { unsigned long flags; - long lastcomp; struct rcu_node *rnp = rcu_get_root(rsp); u8 forcenow; u8 gpdone; @@ -1207,7 +1206,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) goto unlock_fqs_ret; /* no emergency and done recently. */ rsp->n_force_qs++; spin_lock(&rnp->lock); /* irqs already disabled */ - lastcomp = rsp->gpnum - 1; rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; if(!rcu_gp_in_progress(rsp)) { rsp->n_force_qs_ngp++; @@ -1228,7 +1226,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - gpdone = rcu_process_dyntick(rsp, lastcomp, + gpdone = rcu_process_dyntick(rsp, rsp->gpnum - 1, dyntick_save_progress_counter); spin_lock(&rnp->lock); /* irqs already disabled */ if (gpdone) @@ -1239,11 +1237,9 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Update state, record completion counter. */ forcenow = 0; - if (lastcomp + 1 == rsp->gpnum && - lastcomp == rsp->completed) { + if (rsp->gpnum - 1 == rsp->completed) { forcenow = rsp->signaled == RCU_SAVE_COMPLETED; rsp->signaled = RCU_FORCE_QS; - rsp->completed_fqs = lastcomp; } if (!forcenow) break; @@ -1253,7 +1249,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Check dyntick-idle state, send IPI to laggarts. */ spin_unlock(&rnp->lock); /* irqs remain disabled */ - gpdone = rcu_process_dyntick(rsp, rsp->completed_fqs, + gpdone = rcu_process_dyntick(rsp, rsp->gpnum - 1, rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ diff --git a/kernel/rcutree.h b/kernel/rcutree.h index dc386a7..5348561 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -296,8 +296,6 @@ struct rcu_state { long orphan_qlen; /* Number of orphaned cbs. */ spinlock_t fqslock; /* Only one task forcing */ /* quiescent states. */ - long completed_fqs; /* Value of completed @ snap. */ - /* Protected by fqslock. */ unsigned long jiffies_force_qs; /* Time at which to invoke */ /* force_quiescent_state(). */ unsigned long n_force_qs; /* Number of calls to */ -- cgit v0.10.2 From eb1ba45f1e7f6e626fefc063b340c7cbec9bd8c7 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:05 -0800 Subject: rcu: Eliminate second argument of rcu_process_dyntick() At this point, the second argument to all calls to rcu_process_dyntick() is a function of the same field of the structure passed in as the first argument, namely, rsp->gpnum-1. So propagate rsp->gpnum-1 to all uses of the second argument within rcu_process_dyntick() and then eliminate the second argument. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465503786-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 62b6433..c7d0070 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1147,7 +1147,7 @@ void rcu_check_callbacks(int cpu, int user) * Returns 1 if the current grace period ends while scanning (possibly * because we made it end). */ -static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, +static int rcu_process_dyntick(struct rcu_state *rsp, int (*f)(struct rcu_data *)) { unsigned long bit; @@ -1159,7 +1159,7 @@ static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, rcu_for_each_leaf_node(rsp, rnp) { mask = 0; spin_lock_irqsave(&rnp->lock, flags); - if (rnp->completed != lastcomp) { + if (rnp->completed != rsp->gpnum - 1) { spin_unlock_irqrestore(&rnp->lock, flags); return 1; } @@ -1173,7 +1173,7 @@ static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, if ((rnp->qsmask & bit) != 0 && f(rsp->rda[cpu])) mask |= bit; } - if (mask != 0 && rnp->completed == lastcomp) { + if (mask != 0 && rnp->completed == rsp->gpnum - 1) { /* rcu_report_qs_rnp() releases rnp->lock. */ rcu_report_qs_rnp(mask, rsp, rnp, flags); @@ -1226,7 +1226,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - gpdone = rcu_process_dyntick(rsp, rsp->gpnum - 1, + gpdone = rcu_process_dyntick(rsp, dyntick_save_progress_counter); spin_lock(&rnp->lock); /* irqs already disabled */ if (gpdone) @@ -1249,8 +1249,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Check dyntick-idle state, send IPI to laggarts. */ spin_unlock(&rnp->lock); /* irqs remain disabled */ - gpdone = rcu_process_dyntick(rsp, rsp->gpnum - 1, - rcu_implicit_dynticks_qs); + gpdone = rcu_process_dyntick(rsp, rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ -- cgit v0.10.2 From 0f10dc826646134dce3e5751512b87d30f3903e4 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:06 -0800 Subject: rcu: Eliminate rcu_process_dyntick() return value Because a new grace period cannot start while we are executing within the force_quiescent_state() function's switch statement, if any test within that switch statement or within any function called from that switch statement shows that the current grace period has ended, we can safely re-do that test any time before we leave the switch statement. This means that we no longer need a return value from rcu_process_dyntick(), as we can simply invoke rcu_gp_in_progress() to check whether the old grace period has finished -- there is no longer any need to worry about whether or not a new grace period has been started. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465501857-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index c7d0070..e497119 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1144,11 +1144,9 @@ void rcu_check_callbacks(int cpu, int user) /* * Scan the leaf rcu_node structures, processing dyntick state for any that * have not yet encountered a quiescent state, using the function specified. - * Returns 1 if the current grace period ends while scanning (possibly - * because we made it end). */ -static int rcu_process_dyntick(struct rcu_state *rsp, - int (*f)(struct rcu_data *)) +static void rcu_process_dyntick(struct rcu_state *rsp, + int (*f)(struct rcu_data *)) { unsigned long bit; int cpu; @@ -1161,7 +1159,7 @@ static int rcu_process_dyntick(struct rcu_state *rsp, spin_lock_irqsave(&rnp->lock, flags); if (rnp->completed != rsp->gpnum - 1) { spin_unlock_irqrestore(&rnp->lock, flags); - return 1; + return; } if (rnp->qsmask == 0) { spin_unlock_irqrestore(&rnp->lock, flags); @@ -1181,7 +1179,6 @@ static int rcu_process_dyntick(struct rcu_state *rsp, } spin_unlock_irqrestore(&rnp->lock, flags); } - return 0; } /* @@ -1193,7 +1190,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) unsigned long flags; struct rcu_node *rnp = rcu_get_root(rsp); u8 forcenow; - u8 gpdone; if (!rcu_gp_in_progress(rsp)) return; /* No grace period in progress, nothing to force. */ @@ -1226,10 +1222,9 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - gpdone = rcu_process_dyntick(rsp, - dyntick_save_progress_counter); + rcu_process_dyntick(rsp, dyntick_save_progress_counter); spin_lock(&rnp->lock); /* irqs already disabled */ - if (gpdone) + if (!rcu_gp_in_progress(rsp)) break; /* fall into next case. */ @@ -1249,7 +1244,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Check dyntick-idle state, send IPI to laggarts. */ spin_unlock(&rnp->lock); /* irqs remain disabled */ - gpdone = rcu_process_dyntick(rsp, rcu_implicit_dynticks_qs); + rcu_process_dyntick(rsp, rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ -- cgit v0.10.2 From ee47eb9f4da6f44af965d6d049e77ee8c8a4b822 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:07 -0800 Subject: rcu: Remove leg of force_quiescent_state() switch statement The comparisons of rsp->gpnum nad rsp->completed in rcu_process_dyntick() and force_quiescent_state() can be replaced by the much more clear rcu_gp_in_progress() predicate function. After doing this, it becomes clear that the RCU_SAVE_COMPLETED leg of the force_quiescent_state() function's switch statement is almost completely a no-op. A small change to the RCU_SAVE_DYNTICK leg renders it a complete no-op, after which it can be removed. Doing so also eliminates the forcenow local variable from force_quiescent_state(). Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465501781-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index e497119..6268f37 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1144,6 +1144,7 @@ void rcu_check_callbacks(int cpu, int user) /* * Scan the leaf rcu_node structures, processing dyntick state for any that * have not yet encountered a quiescent state, using the function specified. + * The caller must have suppressed start of new grace periods. */ static void rcu_process_dyntick(struct rcu_state *rsp, int (*f)(struct rcu_data *)) @@ -1157,7 +1158,7 @@ static void rcu_process_dyntick(struct rcu_state *rsp, rcu_for_each_leaf_node(rsp, rnp) { mask = 0; spin_lock_irqsave(&rnp->lock, flags); - if (rnp->completed != rsp->gpnum - 1) { + if (!rcu_gp_in_progress(rsp)) { spin_unlock_irqrestore(&rnp->lock, flags); return; } @@ -1171,7 +1172,7 @@ static void rcu_process_dyntick(struct rcu_state *rsp, if ((rnp->qsmask & bit) != 0 && f(rsp->rda[cpu])) mask |= bit; } - if (mask != 0 && rnp->completed == rsp->gpnum - 1) { + if (mask != 0 && rcu_gp_in_progress(rsp)) { /* rcu_report_qs_rnp() releases rnp->lock. */ rcu_report_qs_rnp(mask, rsp, rnp, flags); @@ -1189,7 +1190,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) { unsigned long flags; struct rcu_node *rnp = rcu_get_root(rsp); - u8 forcenow; if (!rcu_gp_in_progress(rsp)) return; /* No grace period in progress, nothing to force. */ @@ -1224,21 +1224,9 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Record dyntick-idle state. */ rcu_process_dyntick(rsp, dyntick_save_progress_counter); spin_lock(&rnp->lock); /* irqs already disabled */ - if (!rcu_gp_in_progress(rsp)) - break; - /* fall into next case. */ - - case RCU_SAVE_COMPLETED: - - /* Update state, record completion counter. */ - forcenow = 0; - if (rsp->gpnum - 1 == rsp->completed) { - forcenow = rsp->signaled == RCU_SAVE_COMPLETED; + if (rcu_gp_in_progress(rsp)) rsp->signaled = RCU_FORCE_QS; - } - if (!forcenow) - break; - /* fall into next case. */ + break; case RCU_FORCE_QS: diff --git a/kernel/rcutree.h b/kernel/rcutree.h index 5348561..edb6fae 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -237,12 +237,11 @@ struct rcu_data { #define RCU_GP_IDLE 0 /* No grace period in progress. */ #define RCU_GP_INIT 1 /* Grace period being initialized. */ #define RCU_SAVE_DYNTICK 2 /* Need to scan dyntick state. */ -#define RCU_SAVE_COMPLETED 3 /* Need to save rsp->completed. */ -#define RCU_FORCE_QS 4 /* Need to force quiescent state. */ +#define RCU_FORCE_QS 3 /* Need to force quiescent state. */ #ifdef CONFIG_NO_HZ #define RCU_SIGNAL_INIT RCU_SAVE_DYNTICK #else /* #ifdef CONFIG_NO_HZ */ -#define RCU_SIGNAL_INIT RCU_SAVE_COMPLETED +#define RCU_SIGNAL_INIT RCU_FORCE_QS #endif /* #else #ifdef CONFIG_NO_HZ */ #define RCU_JIFFIES_TILL_FORCE_QS 3 /* for rsp->jiffies_force_qs */ -- cgit v0.10.2 From 45f014c52eef022873b19d6a20eb0ec9668f2b09 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:08 -0800 Subject: rcu: Remove redundant grace-period check The rcu_process_dyntick() function checks twice for the end of the current grace period. However, it holds the current rcu_node structure's ->lock field throughout, and doesn't get to the second call to rcu_gp_in_progress() unless there is at least one CPU corresponding to this rcu_node structure that has not yet checked in for the current grace period, which would prevent the current grace period from ending. So the current grace period cannot have ended, and the second check is redundant, so remove it. Also, given that this function is used even with !CONFIG_NO_HZ, its name is quite misleading. Change from rcu_process_dyntick() to force_qs_rnp(). Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1262646550562-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 6268f37..d920285 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1146,8 +1146,7 @@ void rcu_check_callbacks(int cpu, int user) * have not yet encountered a quiescent state, using the function specified. * The caller must have suppressed start of new grace periods. */ -static void rcu_process_dyntick(struct rcu_state *rsp, - int (*f)(struct rcu_data *)) +static void force_qs_rnp(struct rcu_state *rsp, int (*f)(struct rcu_data *)) { unsigned long bit; int cpu; @@ -1172,7 +1171,7 @@ static void rcu_process_dyntick(struct rcu_state *rsp, if ((rnp->qsmask & bit) != 0 && f(rsp->rda[cpu])) mask |= bit; } - if (mask != 0 && rcu_gp_in_progress(rsp)) { + if (mask != 0) { /* rcu_report_qs_rnp() releases rnp->lock. */ rcu_report_qs_rnp(mask, rsp, rnp, flags); @@ -1222,7 +1221,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - rcu_process_dyntick(rsp, dyntick_save_progress_counter); + force_qs_rnp(rsp, dyntick_save_progress_counter); spin_lock(&rnp->lock); /* irqs already disabled */ if (rcu_gp_in_progress(rsp)) rsp->signaled = RCU_FORCE_QS; @@ -1232,7 +1231,7 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) /* Check dyntick-idle state, send IPI to laggarts. */ spin_unlock(&rnp->lock); /* irqs remain disabled */ - rcu_process_dyntick(rsp, rcu_implicit_dynticks_qs); + force_qs_rnp(rsp, rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ -- cgit v0.10.2 From 46a1e34eda805501a8b32f26394faa435149f6d1 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:09 -0800 Subject: rcu: Make force_quiescent_state() start grace period if needed Grace periods cannot be started while force_quiescent_state() is active. This is OK in that the affected CPUs will try again later, but it does induce needless grace-period delays. This patch causes rcu_start_gp() to record a failed attempt to start a grace period. When force_quiescent_state() prepares to return, it then starts the grace period if there was such a failed attempt. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626465501854-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index d920285..55e8f6e 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -660,6 +660,8 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) struct rcu_node *rnp = rcu_get_root(rsp); if (!cpu_needs_another_gp(rsp, rdp) || rsp->fqs_active) { + if (cpu_needs_another_gp(rsp, rdp)) + rsp->fqs_need_gp = 1; if (rnp->completed == rsp->completed) { spin_unlock_irqrestore(&rnp->lock, flags); return; @@ -1239,6 +1241,12 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) break; } rsp->fqs_active = 0; + if (rsp->fqs_need_gp) { + spin_unlock(&rsp->fqslock); /* irqs remain disabled */ + rsp->fqs_need_gp = 0; + rcu_start_gp(rsp, flags); /* releases rnp->lock */ + return; + } spin_unlock(&rnp->lock); /* irqs remain disabled */ unlock_fqs_ret: spin_unlock_irqrestore(&rsp->fqslock, flags); diff --git a/kernel/rcutree.h b/kernel/rcutree.h index edb6fae..bd5d78ad 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -278,6 +278,11 @@ struct rcu_state { /* Force QS state. */ u8 fqs_active; /* force_quiescent_state() */ /* is running. */ + u8 fqs_need_gp; /* A CPU was prevented from */ + /* starting a new grace */ + /* period because */ + /* force_quiescent_state() */ + /* was running. */ long gpnum; /* Current gp number. */ long completed; /* # of last completed gp. */ -- cgit v0.10.2 From bf66f18e79e34c421bbd8f6511e2c556b779df2f Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 15:09:10 -0800 Subject: rcu: Add force_quiescent_state() testing to rcutorture Add force_quiescent_state() testing to rcutorture, with a separate thread that repeatedly invokes force_quiescent_state() in bursts. This can greatly increase the probability of encountering certain types of race conditions. Suggested-by: Josh Triplett Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <1262646551116-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h index 96cc307..2b70d4e 100644 --- a/include/linux/rcutiny.h +++ b/include/linux/rcutiny.h @@ -62,6 +62,18 @@ static inline long rcu_batches_completed_bh(void) extern int rcu_expedited_torture_stats(char *page); +static inline void rcu_force_quiescent_state(void) +{ +} + +static inline void rcu_bh_force_quiescent_state(void) +{ +} + +static inline void rcu_sched_force_quiescent_state(void) +{ +} + #define synchronize_rcu synchronize_sched static inline void synchronize_rcu_expedited(void) diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h index 8044b1b..704a010 100644 --- a/include/linux/rcutree.h +++ b/include/linux/rcutree.h @@ -99,6 +99,9 @@ extern void rcu_check_callbacks(int cpu, int user); extern long rcu_batches_completed(void); extern long rcu_batches_completed_bh(void); extern long rcu_batches_completed_sched(void); +extern void rcu_force_quiescent_state(void); +extern void rcu_bh_force_quiescent_state(void); +extern void rcu_sched_force_quiescent_state(void); #ifdef CONFIG_NO_HZ void rcu_enter_nohz(void); diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index 9bb5217..adda92b 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -61,6 +61,9 @@ static int test_no_idle_hz; /* Test RCU's support for tickless idle CPUs. */ static int shuffle_interval = 3; /* Interval between shuffles (in sec)*/ static int stutter = 5; /* Start/stop testing interval (in sec) */ static int irqreader = 1; /* RCU readers from irq (timers). */ +static int fqs_duration = 0; /* Duration of bursts (us), 0 to disable. */ +static int fqs_holdoff = 0; /* Hold time within burst (us). */ +static int fqs_stutter = 3; /* Wait time between bursts (s). */ static char *torture_type = "rcu"; /* What RCU implementation to torture. */ module_param(nreaders, int, 0444); @@ -79,6 +82,12 @@ module_param(stutter, int, 0444); MODULE_PARM_DESC(stutter, "Number of seconds to run/halt test"); module_param(irqreader, int, 0444); MODULE_PARM_DESC(irqreader, "Allow RCU readers from irq handlers"); +module_param(fqs_duration, int, 0444); +MODULE_PARM_DESC(fqs_duration, "Duration of fqs bursts (us)"); +module_param(fqs_holdoff, int, 0444); +MODULE_PARM_DESC(fqs_holdoff, "Holdoff time within fqs bursts (us)"); +module_param(fqs_stutter, int, 0444); +MODULE_PARM_DESC(fqs_stutter, "Wait time between fqs bursts (s)"); module_param(torture_type, charp, 0444); MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, rcu_bh, srcu)"); @@ -99,6 +108,7 @@ static struct task_struct **reader_tasks; static struct task_struct *stats_task; static struct task_struct *shuffler_task; static struct task_struct *stutter_task; +static struct task_struct *fqs_task; #define RCU_TORTURE_PIPE_LEN 10 @@ -263,6 +273,7 @@ struct rcu_torture_ops { void (*deferred_free)(struct rcu_torture *p); void (*sync)(void); void (*cb_barrier)(void); + void (*fqs)(void); int (*stats)(char *page); int irq_capable; char *name; @@ -347,6 +358,7 @@ static struct rcu_torture_ops rcu_ops = { .deferred_free = rcu_torture_deferred_free, .sync = synchronize_rcu, .cb_barrier = rcu_barrier, + .fqs = rcu_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu" @@ -388,6 +400,7 @@ static struct rcu_torture_ops rcu_sync_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = synchronize_rcu, .cb_barrier = NULL, + .fqs = rcu_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_sync" @@ -403,6 +416,7 @@ static struct rcu_torture_ops rcu_expedited_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = synchronize_rcu_expedited, .cb_barrier = NULL, + .fqs = rcu_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_expedited" @@ -465,6 +479,7 @@ static struct rcu_torture_ops rcu_bh_ops = { .deferred_free = rcu_bh_torture_deferred_free, .sync = rcu_bh_torture_synchronize, .cb_barrier = rcu_barrier_bh, + .fqs = rcu_bh_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_bh" @@ -480,6 +495,7 @@ static struct rcu_torture_ops rcu_bh_sync_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = rcu_bh_torture_synchronize, .cb_barrier = NULL, + .fqs = rcu_bh_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_bh_sync" @@ -621,6 +637,7 @@ static struct rcu_torture_ops sched_ops = { .deferred_free = rcu_sched_torture_deferred_free, .sync = sched_torture_synchronize, .cb_barrier = rcu_barrier_sched, + .fqs = rcu_sched_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "sched" @@ -636,6 +653,7 @@ static struct rcu_torture_ops sched_sync_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = sched_torture_synchronize, .cb_barrier = NULL, + .fqs = rcu_sched_force_quiescent_state, .stats = NULL, .name = "sched_sync" }; @@ -650,12 +668,45 @@ static struct rcu_torture_ops sched_expedited_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = synchronize_sched_expedited, .cb_barrier = NULL, + .fqs = rcu_sched_force_quiescent_state, .stats = rcu_expedited_torture_stats, .irq_capable = 1, .name = "sched_expedited" }; /* + * RCU torture force-quiescent-state kthread. Repeatedly induces + * bursts of calls to force_quiescent_state(), increasing the probability + * of occurrence of some important types of race conditions. + */ +static int +rcu_torture_fqs(void *arg) +{ + unsigned long fqs_resume_time; + int fqs_burst_remaining; + + VERBOSE_PRINTK_STRING("rcu_torture_fqs task started"); + do { + fqs_resume_time = jiffies + fqs_stutter * HZ; + while (jiffies - fqs_resume_time > LONG_MAX) { + schedule_timeout_interruptible(1); + } + fqs_burst_remaining = fqs_duration; + while (fqs_burst_remaining > 0) { + cur_ops->fqs(); + udelay(fqs_holdoff); + fqs_burst_remaining -= fqs_holdoff; + } + rcu_stutter_wait("rcu_torture_fqs"); + } while (!kthread_should_stop() && fullstop == FULLSTOP_DONTSTOP); + VERBOSE_PRINTK_STRING("rcu_torture_fqs task stopping"); + rcutorture_shutdown_absorb("rcu_torture_fqs"); + while (!kthread_should_stop()) + schedule_timeout_uninterruptible(1); + return 0; +} + +/* * RCU torture writer kthread. Repeatedly substitutes a new structure * for that pointed to by rcu_torture_current, freeing the old structure * after a series of grace periods (the "pipeline"). @@ -1030,10 +1081,11 @@ rcu_torture_print_module_parms(char *tag) printk(KERN_ALERT "%s" TORTURE_FLAG "--- %s: nreaders=%d nfakewriters=%d " "stat_interval=%d verbose=%d test_no_idle_hz=%d " - "shuffle_interval=%d stutter=%d irqreader=%d\n", + "shuffle_interval=%d stutter=%d irqreader=%d " + "fqs_duration=%d fqs_holdoff=%d fqs_stutter=%d\n", torture_type, tag, nrealreaders, nfakewriters, stat_interval, verbose, test_no_idle_hz, shuffle_interval, - stutter, irqreader); + stutter, irqreader, fqs_duration, fqs_holdoff, fqs_stutter); } static struct notifier_block rcutorture_nb = { @@ -1109,6 +1161,12 @@ rcu_torture_cleanup(void) } stats_task = NULL; + if (fqs_task) { + VERBOSE_PRINTK_STRING("Stopping rcu_torture_fqs task"); + kthread_stop(fqs_task); + } + fqs_task = NULL; + /* Wait for all RCU callbacks to fire. */ if (cur_ops->cb_barrier != NULL) @@ -1154,6 +1212,11 @@ rcu_torture_init(void) mutex_unlock(&fullstop_mutex); return -EINVAL; } + if (cur_ops->fqs == NULL && fqs_duration != 0) { + printk(KERN_ALERT "rcu-torture: ->fqs NULL and non-zero " + "fqs_duration, fqs disabled.\n"); + fqs_duration = 0; + } if (cur_ops->init) cur_ops->init(); /* no "goto unwind" prior to this point!!! */ @@ -1282,6 +1345,19 @@ rcu_torture_init(void) goto unwind; } } + if (fqs_duration < 0) + fqs_duration = 0; + if (fqs_duration) { + /* Create the stutter thread */ + fqs_task = kthread_run(rcu_torture_fqs, NULL, + "rcu_torture_fqs"); + if (IS_ERR(fqs_task)) { + firsterr = PTR_ERR(fqs_task); + VERBOSE_PRINTK_ERRSTRING("Failed to create fqs"); + fqs_task = NULL; + goto unwind; + } + } register_reboot_notifier(&rcutorture_nb); mutex_unlock(&fullstop_mutex); return 0; diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 55e8f6e..0a4c328 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -157,6 +157,24 @@ long rcu_batches_completed_bh(void) EXPORT_SYMBOL_GPL(rcu_batches_completed_bh); /* + * Force a quiescent state for RCU BH. + */ +void rcu_bh_force_quiescent_state(void) +{ + force_quiescent_state(&rcu_bh_state, 0); +} +EXPORT_SYMBOL_GPL(rcu_bh_force_quiescent_state); + +/* + * Force a quiescent state for RCU-sched. + */ +void rcu_sched_force_quiescent_state(void) +{ + force_quiescent_state(&rcu_sched_state, 0); +} +EXPORT_SYMBOL_GPL(rcu_sched_force_quiescent_state); + +/* * Does the CPU have callbacks ready to be invoked? */ static int diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 37fbccd..f11ebd4 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -62,6 +62,15 @@ long rcu_batches_completed(void) EXPORT_SYMBOL_GPL(rcu_batches_completed); /* + * Force a quiescent state for preemptible RCU. + */ +void rcu_force_quiescent_state(void) +{ + force_quiescent_state(&rcu_preempt_state, 0); +} +EXPORT_SYMBOL_GPL(rcu_force_quiescent_state); + +/* * Record a preemptable-RCU quiescent state for the specified CPU. Note * that this just means that the task currently running on the CPU is * not in a quiescent state. There might be any number of tasks blocked @@ -713,6 +722,16 @@ long rcu_batches_completed(void) EXPORT_SYMBOL_GPL(rcu_batches_completed); /* + * Force a quiescent state for RCU, which, because there is no preemptible + * RCU, becomes the same as rcu-sched. + */ +void rcu_force_quiescent_state(void) +{ + rcu_sched_force_quiescent_state(); +} +EXPORT_SYMBOL_GPL(rcu_force_quiescent_state); + +/* * Because preemptable RCU does not exist, we never have to check for * CPUs being in quiescent states. */ -- cgit v0.10.2 From f9094d8e5587cf21091a9516628147c0b55e4264 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 16:04:00 -0800 Subject: rcu: Make MAINTAINERS file match new RCU reality Both rcutorture and RCU are supported. Also, update the files to match the new layout. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626498422334-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/MAINTAINERS b/MAINTAINERS index c8f47bf..175da7c 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4509,7 +4509,7 @@ F: drivers/net/wireless/ray* RCUTORTURE MODULE M: Josh Triplett M: "Paul E. McKenney" -S: Maintained +S: Supported F: Documentation/RCU/torture.txt F: kernel/rcutorture.c @@ -4534,11 +4534,12 @@ M: Dipankar Sarma M: "Paul E. McKenney" W: http://www.rdrop.com/users/paulmck/rclock/ S: Supported -F: Documentation/RCU/rcu.txt -F: Documentation/RCU/rcuref.txt -F: include/linux/rcupdate.h -F: include/linux/srcu.h -F: kernel/rcupdate.c +F: Documentation/RCU/ +F: include/linux/rcu* +F: include/linux/srcu* +F: kernel/rcu* +F: kernel/srcu* +X: kernel/rcutorture.c REAL TIME CLOCK DRIVER M: Paul Gortmaker -- cgit v0.10.2 From cba8244a0f1c277b6b1e48ed6504fa434119e24d Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 16:04:01 -0800 Subject: rcu: Add debug check for too many rcu_read_unlock() TREE_PREEMPT_RCU maintains an rcu_read_lock_nesting counter in the task structure, which happens to be a signed int. So this patch adds a check for this counter being negative at the end of __rcu_read_unlock(). This check is under CONFIG_PROVE_LOCKING, so can be thought of as being part of lockdep. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626498423064-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index f11ebd4..e77cdf3 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -304,6 +304,9 @@ void __rcu_read_unlock(void) if (--ACCESS_ONCE(t->rcu_read_lock_nesting) == 0 && unlikely(ACCESS_ONCE(t->rcu_read_unlock_special))) rcu_read_unlock_special(t); +#ifdef CONFIG_PROVE_LOCKING + WARN_ON_ONCE(ACCESS_ONCE(t->rcu_read_lock_nesting) < 0); +#endif /* #ifdef CONFIG_PROVE_LOCKING */ } EXPORT_SYMBOL_GPL(__rcu_read_unlock); -- cgit v0.10.2 From b6407e863934965cdc66cbc244d811ceeb6f4d77 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Mon, 4 Jan 2010 16:04:02 -0800 Subject: rcu: Give different levels of the rcu_node hierarchy distinct lockdep names Previously, each level of the rcu_node hierarchy had the same rather unimaginative name: "&rcu_node_class[i]". This makes lockdep diagnostics involving these lockdep classes less helpful than would be nice. This patch fixes this by giving each level of the rcu_node hierarchy a distinct name: "rcu_node_level_0", "rcu_node_level_1", and so on. This version of the patch includes improved diagnostics suggested by Josh Triplett and Peter Zijlstra. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12626498421830-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 0a4c328..3b13d64 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1811,11 +1811,17 @@ static void __init rcu_init_levelspread(struct rcu_state *rsp) */ static void __init rcu_init_one(struct rcu_state *rsp) { + static char *buf[] = { "rcu_node_level_0", + "rcu_node_level_1", + "rcu_node_level_2", + "rcu_node_level_3" }; /* Match MAX_RCU_LVLS */ int cpustride = 1; int i; int j; struct rcu_node *rnp; + BUILD_BUG_ON(MAX_RCU_LVLS > ARRAY_SIZE(buf)); /* Fix buf[] init! */ + /* Initialize the level-tracking arrays. */ for (i = 1; i < NUM_RCU_LVLS; i++) @@ -1829,7 +1835,8 @@ static void __init rcu_init_one(struct rcu_state *rsp) rnp = rsp->level[i]; for (j = 0; j < rsp->levelcnt[i]; j++, rnp++) { spin_lock_init(&rnp->lock); - lockdep_set_class(&rnp->lock, &rcu_node_class[i]); + lockdep_set_class_and_name(&rnp->lock, + &rcu_node_class[i], buf[i]); rnp->gpnum = 0; rnp->qsmask = 0; rnp->qsmaskinit = 0; -- cgit v0.10.2 From 0fb8ee48d9dfff6a0913ceb0be2068d8be203763 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 31 Dec 2009 05:53:03 +0100 Subject: perf: Drop useless check for ignored frame The check that ignores the debug and nmi stack frames is useless now that we have a frame pointer that makes us start at the right place. We don't anymore have to deal with these. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras LKML-Reference: <1262235183-5320-2-git-send-regression-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/stacktrace.h b/arch/x86/include/asm/stacktrace.h index 35e8912..4dab78e 100644 --- a/arch/x86/include/asm/stacktrace.h +++ b/arch/x86/include/asm/stacktrace.h @@ -3,8 +3,6 @@ extern int kstack_depth_to_print; -int x86_is_stack_id(int id, char *name); - struct thread_info; struct stacktrace_ops; diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index d616c06..b1bb8c5 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -2297,7 +2297,6 @@ void callchain_store(struct perf_callchain_entry *entry, u64 ip) static DEFINE_PER_CPU(struct perf_callchain_entry, pmc_irq_entry); static DEFINE_PER_CPU(struct perf_callchain_entry, pmc_nmi_entry); -static DEFINE_PER_CPU(int, in_ignored_frame); static void @@ -2313,10 +2312,6 @@ static void backtrace_warning(void *data, char *msg) static int backtrace_stack(void *data, char *name) { - per_cpu(in_ignored_frame, smp_processor_id()) = - x86_is_stack_id(NMI_STACK, name) || - x86_is_stack_id(DEBUG_STACK, name); - return 0; } @@ -2324,9 +2319,6 @@ static void backtrace_address(void *data, unsigned long addr, int reliable) { struct perf_callchain_entry *entry = data; - if (per_cpu(in_ignored_frame, smp_processor_id())) - return; - if (reliable) callchain_store(entry, addr); } diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c index ae775ca..11540a1 100644 --- a/arch/x86/kernel/dumpstack_32.c +++ b/arch/x86/kernel/dumpstack_32.c @@ -18,11 +18,6 @@ #include "dumpstack.h" -/* Just a stub for now */ -int x86_is_stack_id(int id, char *name) -{ - return 0; -} void dump_trace(struct task_struct *task, struct pt_regs *regs, unsigned long *stack, unsigned long bp, diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index 0ad9597..676bc05 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -33,11 +33,6 @@ static char x86_stack_ids[][8] = { #endif }; -int x86_is_stack_id(int id, char *name) -{ - return x86_stack_ids[id - 1] == name; -} - static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack, unsigned *usedp, char **idp) { -- cgit v0.10.2 From 60666c630bdb33983a894b050b588b663f38f368 Mon Sep 17 00:00:00 2001 From: Liming Wang Date: Thu, 31 Dec 2009 16:05:50 +0800 Subject: perf tools: Fix --pid option for stat current pid option doesn't work for perf stat. Change it to what perf record --pid acts as. Signed-off-by: Liming Wang Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: Peter Zijlstra LKML-Reference: <1262246750-2191-1-git-send-email-liming.wang@windriver.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index c70d720..e8c85d5 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -44,6 +44,7 @@ #include "util/parse-events.h" #include "util/event.h" #include "util/debug.h" +#include "util/header.h" #include #include @@ -79,6 +80,8 @@ static int fd[MAX_NR_CPUS][MAX_COUNTERS]; static int event_scaled[MAX_COUNTERS]; +static volatile int done = 0; + struct stats { double n, mean, M2; @@ -247,61 +250,64 @@ static int run_perf_stat(int argc __used, const char **argv) unsigned long long t0, t1; int status = 0; int counter; - int pid; + int pid = target_pid; int child_ready_pipe[2], go_pipe[2]; + const bool forks = (target_pid == -1 && argc > 0); char buf; if (!system_wide) nr_cpus = 1; - if (pipe(child_ready_pipe) < 0 || pipe(go_pipe) < 0) { + if (forks && (pipe(child_ready_pipe) < 0 || pipe(go_pipe) < 0)) { perror("failed to create pipes"); exit(1); } - if ((pid = fork()) < 0) - perror("failed to fork"); + if (forks) { + if ((pid = fork()) < 0) + perror("failed to fork"); + + if (!pid) { + close(child_ready_pipe[0]); + close(go_pipe[1]); + fcntl(go_pipe[0], F_SETFD, FD_CLOEXEC); + + /* + * Do a dummy execvp to get the PLT entry resolved, + * so we avoid the resolver overhead on the real + * execvp call. + */ + execvp("", (char **)argv); + + /* + * Tell the parent we're ready to go + */ + close(child_ready_pipe[1]); + + /* + * Wait until the parent tells us to go. + */ + if (read(go_pipe[0], &buf, 1) == -1) + perror("unable to read pipe"); + + execvp(argv[0], (char **)argv); + + perror(argv[0]); + exit(-1); + } - if (!pid) { - close(child_ready_pipe[0]); - close(go_pipe[1]); - fcntl(go_pipe[0], F_SETFD, FD_CLOEXEC); + child_pid = pid; /* - * Do a dummy execvp to get the PLT entry resolved, - * so we avoid the resolver overhead on the real - * execvp call. - */ - execvp("", (char **)argv); - - /* - * Tell the parent we're ready to go + * Wait for the child to be ready to exec. */ close(child_ready_pipe[1]); - - /* - * Wait until the parent tells us to go. - */ - if (read(go_pipe[0], &buf, 1) == -1) + close(go_pipe[0]); + if (read(child_ready_pipe[0], &buf, 1) == -1) perror("unable to read pipe"); - - execvp(argv[0], (char **)argv); - - perror(argv[0]); - exit(-1); + close(child_ready_pipe[0]); } - child_pid = pid; - - /* - * Wait for the child to be ready to exec. - */ - close(child_ready_pipe[1]); - close(go_pipe[0]); - if (read(child_ready_pipe[0], &buf, 1) == -1) - perror("unable to read pipe"); - close(child_ready_pipe[0]); - for (counter = 0; counter < nr_counters; counter++) create_perf_stat_counter(counter, pid); @@ -310,8 +316,12 @@ static int run_perf_stat(int argc __used, const char **argv) */ t0 = rdclock(); - close(go_pipe[1]); - wait(&status); + if (forks) { + close(go_pipe[1]); + wait(&status); + } else { + while(!done); + } t1 = rdclock(); @@ -417,10 +427,13 @@ static void print_stat(int argc, const char **argv) fflush(stdout); fprintf(stderr, "\n"); - fprintf(stderr, " Performance counter stats for \'%s", argv[0]); - - for (i = 1; i < argc; i++) - fprintf(stderr, " %s", argv[i]); + fprintf(stderr, " Performance counter stats for "); + if(target_pid == -1) { + fprintf(stderr, "\'%s", argv[0]); + for (i = 1; i < argc; i++) + fprintf(stderr, " %s", argv[i]); + }else + fprintf(stderr, "task pid \'%d", target_pid); fprintf(stderr, "\'"); if (run_count > 1) @@ -445,6 +458,9 @@ static volatile int signr = -1; static void skip_signal(int signo) { + if(target_pid != -1) + done = 1; + signr = signo; } @@ -461,7 +477,7 @@ static void sig_atexit(void) } static const char * const stat_usage[] = { - "perf stat [] ", + "perf stat [] []", NULL }; @@ -492,7 +508,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used) argc = parse_options(argc, argv, options, stat_usage, PARSE_OPT_STOP_AT_NON_OPTION); - if (!argc) + if (!argc && target_pid == -1) usage_with_options(stat_usage, options); if (run_count <= 0) usage_with_options(stat_usage, options); -- cgit v0.10.2 From 682b335a5bccf9e5b7e74380784aa2f145d04444 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 4 Jan 2010 16:19:26 -0200 Subject: perf symbols: Generalise the kallsyms parsing routine MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Will be used to find an specific symbol by name on 'perf record' to support relocation reference symbols to support relocatable kernels. Still have to conver the perf trace tools to use it instead of their current reimplementation. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262629169-22797-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 79ca6a0..b9e0da5 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -383,16 +383,12 @@ size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp) return ret; } -/* - * Loads the function entries in /proc/kallsyms into kernel_map->dso, - * so that we can in the next step set the symbol ->end address and then - * call kernel_maps__split_kallsyms. - */ -static int dso__load_all_kallsyms(struct dso *self, struct map *map) +int kallsyms__parse(void *arg, int (*process_symbol)(void *arg, const char *name, + char type, u64 start)) { char *line = NULL; size_t n; - struct rb_root *root = &self->symbols[map->type]; + int err = 0; FILE *file = fopen("/proc/kallsyms", "r"); if (file == NULL) @@ -400,7 +396,6 @@ static int dso__load_all_kallsyms(struct dso *self, struct map *map) while (!feof(file)) { u64 start; - struct symbol *sym; int line_len, len; char symbol_type; char *symbol_name; @@ -421,35 +416,62 @@ static int dso__load_all_kallsyms(struct dso *self, struct map *map) continue; symbol_type = toupper(line[len]); - if (!symbol_type__is_a(symbol_type, map->type)) - continue; - symbol_name = line + len + 2; - /* - * Will fix up the end later, when we have all symbols sorted. - */ - sym = symbol__new(start, 0, symbol_name); - if (sym == NULL) - goto out_delete_line; - /* - * We will pass the symbols to the filter later, in - * map__split_kallsyms, when we have split the maps per module - */ - symbols__insert(root, sym); + err = process_symbol(arg, symbol_name, symbol_type, start); + if (err) + break; } free(line); fclose(file); + return err; - return 0; - -out_delete_line: - free(line); out_failure: return -1; } +struct process_kallsyms_args { + struct map *map; + struct dso *dso; +}; + +static int map__process_kallsym_symbol(void *arg, const char *name, + char type, u64 start) +{ + struct symbol *sym; + struct process_kallsyms_args *a = arg; + struct rb_root *root = &a->dso->symbols[a->map->type]; + + if (!symbol_type__is_a(type, a->map->type)) + return 0; + + /* + * Will fix up the end later, when we have all symbols sorted. + */ + sym = symbol__new(start, 0, name); + + if (sym == NULL) + return -ENOMEM; + /* + * We will pass the symbols to the filter later, in + * map__split_kallsyms, when we have split the maps per module + */ + symbols__insert(root, sym); + return 0; +} + +/* + * Loads the function entries in /proc/kallsyms into kernel_map->dso, + * so that we can in the next step set the symbol ->end address and then + * call kernel_maps__split_kallsyms. + */ +static int dso__load_all_kallsyms(struct dso *self, struct map *map) +{ + struct process_kallsyms_args args = { .map = map, .dso = self, }; + return kallsyms__parse(&args, map__process_kallsym_symbol); +} + /* * Split the symbols into maps, making sure there are no overlaps, i.e. the * kernel range is broken in several maps, named [kernel].N, as we don't have diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index f27e158..21313e8 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -135,6 +135,8 @@ int filename__read_build_id(const char *filename, void *bf, size_t size); int sysfs__read_build_id(const char *filename, void *bf, size_t size); bool dsos__read_build_ids(void); int build_id__sprintf(u8 *self, int len, char *bf); +int kallsyms__parse(void *arg, int (*process_symbol)(void *arg, const char *name, + char type, u64 start)); int symbol__init(void); int perf_session__create_kernel_maps(struct perf_session *self); -- cgit v0.10.2 From 36a3e6461a0dac8e84b8c94877365324010c151b Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 4 Jan 2010 16:19:27 -0200 Subject: perf symbols: Export symbol_type__is_a MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Will be needed by the new HEADER_DSO_INFO feature that will be a HEADER_BUILD_ID superset, replacing it. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262629169-22797-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index b9e0da5..5dffcd1 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -64,7 +64,7 @@ static void dso__set_sorted_by_name(struct dso *self, enum map_type type) self->sorted_by_name |= (1 << type); } -static bool symbol_type__is_a(char symbol_type, enum map_type map_type) +bool symbol_type__is_a(char symbol_type, enum map_type map_type) { switch (map_type) { case MAP__FUNCTION: diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 21313e8..b2b5330 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -139,6 +139,8 @@ int kallsyms__parse(void *arg, int (*process_symbol)(void *arg, const char *name char type, u64 start)); int symbol__init(void); +bool symbol_type__is_a(char symbol_type, enum map_type map_type); + int perf_session__create_kernel_maps(struct perf_session *self); extern struct list_head dsos__user, dsos__kernel; -- cgit v0.10.2 From f92cb24c78a7c853435e46a20d1bd5c894378132 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 4 Jan 2010 16:19:28 -0200 Subject: perf tools: Create write_padded routine out of __dsos__write_buildid_table MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Will be used by other options where padding is needed. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262629169-22797-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 709e325..942f7da 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -173,6 +173,20 @@ static int do_write(int fd, const void *buf, size_t size) return 0; } +#define NAME_ALIGN 64 + +static int write_padded(int fd, const void *bf, size_t count, + size_t count_aligned) +{ + static const char zero_buf[NAME_ALIGN]; + int err = do_write(fd, bf, count); + + if (!err) + err = do_write(fd, zero_buf, count_aligned - count); + + return err; +} + #define dsos__for_each_with_build_id(pos, head) \ list_for_each_entry(pos, head, node) \ if (!pos->has_build_id) \ @@ -181,9 +195,7 @@ static int do_write(int fd, const void *buf, size_t size) static int __dsos__write_buildid_table(struct list_head *head, int fd) { -#define NAME_ALIGN 64 struct dso *pos; - static const char zero_buf[NAME_ALIGN]; dsos__for_each_with_build_id(pos, head) { int err; @@ -197,10 +209,8 @@ static int __dsos__write_buildid_table(struct list_head *head, int fd) err = do_write(fd, &b, sizeof(b)); if (err < 0) return err; - err = do_write(fd, pos->long_name, pos->long_name_len + 1); - if (err < 0) - return err; - err = do_write(fd, zero_buf, len - pos->long_name_len - 1); + err = write_padded(fd, pos->long_name, + pos->long_name_len + 1, len); if (err < 0) return err; } -- cgit v0.10.2 From de1764892a61a3ed212973cc028c80dd083179dd Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 4 Jan 2010 16:19:29 -0200 Subject: perf session: Keep pointers to the vmlinux maps MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit So that tools such as 'perf probe' don't have to lookup '[kernel.kallsyms]' but instead access them directly after perf_session__create_kernel_maps or map_groups__create_kernel_maps. Signed-off-by: Arnaldo Carvalho de Melo Cc: Masami Hiramatsu Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262629169-22797-4-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index c1e6774..ffdd3fe 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -235,9 +235,7 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) session.psession = perf_session__new(NULL, O_WRONLY, false); if (session.psession == NULL) die("Failed to init perf_session."); - session.kmap = map_groups__find_by_name(&session.psession->kmaps, - MAP__FUNCTION, - "[kernel.kallsyms]"); + session.kmap = session.psession->vmlinux_maps[MAP__FUNCTION]; if (!session.kmap) die("Could not find kernel map.\n"); diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 77c5ee29..8db37bb 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -18,6 +18,7 @@ struct perf_session { struct map_groups kmaps; struct rb_root threads; struct thread *last_match; + struct map *vmlinux_maps[MAP__NR_TYPES]; struct events_stats events_stats; unsigned long event_total[PERF_RECORD_MAX]; unsigned long unknown_events; diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 5dffcd1..e290429 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1662,7 +1662,7 @@ size_t dsos__fprintf_buildid(FILE *fp) __dsos__fprintf_buildid(&dsos__user, fp)); } -static struct dso *dsos__create_kernel( const char *vmlinux) +static struct dso *dsos__create_kernel(const char *vmlinux) { struct dso *kernel = dso__new(vmlinux ?: "[kernel.kallsyms]"); @@ -1691,29 +1691,26 @@ out_delete_kernel_dso: return NULL; } -static int map_groups__create_kernel_maps(struct map_groups *self, const char *vmlinux) +static int map_groups__create_kernel_maps(struct map_groups *self, + struct map *vmlinux_maps[MAP__NR_TYPES], + const char *vmlinux) { - struct map *functions, *variables; struct dso *kernel = dsos__create_kernel(vmlinux); + enum map_type type; if (kernel == NULL) return -1; - functions = map__new2(0, kernel, MAP__FUNCTION); - if (functions == NULL) - return -1; + for (type = 0; type < MAP__NR_TYPES; ++type) { + vmlinux_maps[type] = map__new2(0, kernel, type); + if (vmlinux_maps[type] == NULL) + return -1; - variables = map__new2(0, kernel, MAP__VARIABLE); - if (variables == NULL) { - map__delete(functions); - return -1; + vmlinux_maps[type]->map_ip = + vmlinux_maps[type]->unmap_ip = identity__map_ip; + map_groups__insert(self, vmlinux_maps[type]); } - functions->map_ip = functions->unmap_ip = - variables->map_ip = variables->unmap_ip = identity__map_ip; - map_groups__insert(self, functions); - map_groups__insert(self, variables); - return 0; } @@ -1824,7 +1821,7 @@ out_free_comm_list: int perf_session__create_kernel_maps(struct perf_session *self) { - if (map_groups__create_kernel_maps(&self->kmaps, + if (map_groups__create_kernel_maps(&self->kmaps, self->vmlinux_maps, symbol_conf.vmlinux_name) < 0) return -1; -- cgit v0.10.2 From b9a63b9b56d6910a25e3d4905525aef150420a9b Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 5 Jan 2010 11:54:45 -0200 Subject: perf report: Fix --no-call-chain option handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit To avoid the funny: [root@doppio ~]# perf record -a -f sleep 2s [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.334 MB perf.data (~14572 samples) ] [root@doppio ~]# perf report --no-call-graph selected -g but no callchain data. Did you call perf record without -g? And fix the bug reported by peterz when we do indeed record with callchains and then ask for a report without: [root@doppio ~]# perf record -a -g -f sleep 2s [root@doppio ~]# perf report --no-call-graph Segmentation fault [root@doppio ~]# Reported-by: Peter Zijlstra Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262699685-27820-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 4292d7a..80d691a 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -35,6 +35,7 @@ static char const *input_name = "perf.data"; static int force; static bool hide_unresolved; +static bool dont_use_callchains; static int show_threads; static struct perf_read_values show_threads_values; @@ -172,7 +173,8 @@ static int perf_session__setup_sample_type(struct perf_session *self) " -g?\n"); return -1; } - } else if (callchain_param.mode != CHAIN_NONE && !symbol_conf.use_callchain) { + } else if (!dont_use_callchains && callchain_param.mode != CHAIN_NONE && + !symbol_conf.use_callchain) { symbol_conf.use_callchain = true; if (register_callchain_param(&callchain_param) < 0) { fprintf(stderr, "Can't register callchain" @@ -246,11 +248,19 @@ out_delete: static int parse_callchain_opt(const struct option *opt __used, const char *arg, - int unset __used) + int unset) { char *tok; char *endptr; + /* + * --no-call-graph + */ + if (unset) { + dont_use_callchains = true; + return 0; + } + symbol_conf.use_callchain = true; if (!arg) -- cgit v0.10.2 From 56b03f3c4d641dbdbce2e52a2969712e85b0e030 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 5 Jan 2010 16:50:31 -0200 Subject: perf tools: Handle relocatable kernels MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit DSOs don't have this problem because the kernel emits a PERF_MMAP for each new executable mapping it performs on monitored threads. To fix the kernel case we simulate the same behaviour, by having 'perf record' to synthesize a PERF_MMAP for the kernel, encoded like this: [root@doppio ~]# perf record -a -f sleep 1 [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.344 MB perf.data (~15038 samples) ] [root@doppio ~]# perf report -D | head -10 0xd0 [0x40]: event: 1 . . ... raw event: size 64 bytes . 0000: 01 00 00 00 00 00 40 00 00 00 00 00 00 00 00 00 ......@........ . 0010: 00 00 00 81 ff ff ff ff 00 00 00 00 00 00 00 00 ............... . 0020: 00 00 00 00 00 00 00 00 5b 6b 65 72 6e 65 6c 2e ........ [kernel . 0030: 6b 61 6c 6c 73 79 6d 73 2e 5f 74 65 78 74 5d 00 kallsyms._text] . 0xd0 [0x40]: PERF_RECORD_MMAP 0/0: [0xffffffff81000000((nil)) @ (nil)]: [kernel.kallsyms._text] I.e. we identify such event as having: .pid = 0 .filename = [kernel.kallsyms.REFNAME] .start = REFNAME addr in /proc/kallsyms at 'perf record' time and use now a hardcoded value of '.text' for REFNAME. Then, later, in 'perf report', if there are any kernel hits and thus we need to resolve kernel symbols, we search for REFNAME and if its address changed, relocation happened and we thus must change the kernel mapping routines to one that uses .pgoff as the relocation to apply. This way we use the same mechanism used for the other DSOs and don't have to do a two pass in all the kernel symbols. Reported-by: Xiao Guangrong Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: "H. Peter Anvin" Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Xiao Guangrong LKML-Reference: <1262717431-1246-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 2654253..8f88420 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -551,6 +551,13 @@ static int __cmd_record(int argc, const char **argv) return err; } + err = event__synthesize_kernel_mmap(process_synthesized_event, + session, "_text"); + if (err < 0) { + pr_err("Couldn't record kernel reference relocation symbol.\n"); + return err; + } + if (!system_wide && profile_cpu == -1) event__synthesize_thread(pid, process_synthesized_event, session); diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index bb0fd6d..1a31feb 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -189,6 +189,50 @@ void event__synthesize_threads(int (*process)(event_t *event, closedir(proc); } +struct process_symbol_args { + const char *name; + u64 start; +}; + +static int find_symbol_cb(void *arg, const char *name, char type, u64 start) +{ + struct process_symbol_args *args = arg; + + if (!symbol_type__is_a(type, MAP__FUNCTION) || strcmp(name, args->name)) + return 0; + + args->start = start; + return 1; +} + +int event__synthesize_kernel_mmap(int (*process)(event_t *event, + struct perf_session *session), + struct perf_session *session, + const char *symbol_name) +{ + size_t size; + event_t ev = { + .header = { .type = PERF_RECORD_MMAP }, + }; + /* + * We should get this from /sys/kernel/sections/.text, but till that is + * available use this, and after it is use this as a fallback for older + * kernels. + */ + struct process_symbol_args args = { .name = symbol_name, }; + + if (kallsyms__parse(&args, find_symbol_cb) <= 0) + return -ENOENT; + + size = snprintf(ev.mmap.filename, sizeof(ev.mmap.filename), + "[kernel.kallsyms.%s]", symbol_name) + 1; + size = ALIGN(size, sizeof(u64)); + ev.mmap.header.size = (sizeof(ev.mmap) - (sizeof(ev.mmap.filename) - size)); + ev.mmap.start = args.start; + + return process(&ev, session); +} + static void thread__comm_adjust(struct thread *self) { char *comm = self->comm; @@ -240,9 +284,9 @@ int event__process_lost(event_t *self, struct perf_session *session) int event__process_mmap(event_t *self, struct perf_session *session) { - struct thread *thread = perf_session__findnew(session, self->mmap.pid); - struct map *map = map__new(&self->mmap, MAP__FUNCTION, - session->cwd, session->cwdlen); + struct thread *thread; + struct map *map; + static const char kmmap_prefix[] = "[kernel.kallsyms."; dump_printf(" %d/%d: [%p(%p) @ %p]: %s\n", self->mmap.pid, self->mmap.tid, @@ -251,6 +295,20 @@ int event__process_mmap(event_t *self, struct perf_session *session) (void *)(long)self->mmap.pgoff, self->mmap.filename); + if (self->mmap.pid == 0 && + memcmp(self->mmap.filename, kmmap_prefix, + sizeof(kmmap_prefix) - 1) == 0) { + const char *symbol_name = (self->mmap.filename + + sizeof(kmmap_prefix) - 1); + perf_session__set_kallsyms_ref_reloc_sym(session, symbol_name, + self->mmap.start); + return 0; + } + + thread = perf_session__findnew(session, self->mmap.pid); + map = map__new(&self->mmap, MAP__FUNCTION, + session->cwd, session->cwdlen); + if (thread == NULL || map == NULL) dump_printf("problem processing PERF_RECORD_MMAP, skipping event.\n"); else diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 80fb365..61fc0dc 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -110,6 +110,10 @@ int event__synthesize_thread(pid_t pid, void event__synthesize_threads(int (*process)(event_t *event, struct perf_session *session), struct perf_session *session); +int event__synthesize_kernel_mmap(int (*process)(event_t *event, + struct perf_session *session), + struct perf_session *session, + const char *symbol_name); int event__process_comm(event_t *self, struct perf_session *session); int event__process_lost(event_t *self, struct perf_session *session); diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 7f0537d..e0e6a07 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -401,3 +401,49 @@ bool perf_session__has_traces(struct perf_session *self, const char *msg) return true; } + +int perf_session__set_kallsyms_ref_reloc_sym(struct perf_session *self, + const char *symbol_name, + u64 addr) +{ + char *bracket; + + self->ref_reloc_sym.name = strdup(symbol_name); + if (self->ref_reloc_sym.name == NULL) + return -ENOMEM; + + bracket = strchr(self->ref_reloc_sym.name, ']'); + if (bracket) + *bracket = '\0'; + + self->ref_reloc_sym.addr = addr; + return 0; +} + +static u64 map__reloc_map_ip(struct map *map, u64 ip) +{ + return ip + (s64)map->pgoff; +} + +static u64 map__reloc_unmap_ip(struct map *map, u64 ip) +{ + return ip - (s64)map->pgoff; +} + +void perf_session__reloc_vmlinux_maps(struct perf_session *self, + u64 unrelocated_addr) +{ + enum map_type type; + s64 reloc = unrelocated_addr - self->ref_reloc_sym.addr; + + if (!reloc) + return; + + for (type = 0; type < MAP__NR_TYPES; ++type) { + struct map *map = self->vmlinux_maps[type]; + + map->map_ip = map__reloc_map_ip; + map->unmap_ip = map__reloc_unmap_ip; + map->pgoff = reloc; + } +} diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 8db37bb..d4a9d20 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -24,6 +24,10 @@ struct perf_session { unsigned long unknown_events; struct rb_root hists; u64 sample_type; + struct { + const char *name; + u64 addr; + } ref_reloc_sym; int fd; int cwdlen; char *cwd; @@ -59,4 +63,10 @@ bool perf_session__has_traces(struct perf_session *self, const char *msg); int perf_header__read_build_ids(int input, u64 offset, u64 file_size); +int perf_session__set_kallsyms_ref_reloc_sym(struct perf_session *self, + const char *symbol_name, + u64 addr); +void perf_session__reloc_vmlinux_maps(struct perf_session *self, + u64 unrelocated_addr); + #endif /* __PERF_SESSION_H */ diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index e290429..da2f07f 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -956,11 +956,15 @@ static int dso__load_sym(struct dso *self, struct map *map, elf_symtab__for_each_symbol(syms, nr_syms, idx, sym) { struct symbol *f; - const char *elf_name; + const char *elf_name = elf_sym__name(&sym, symstrs); char *demangled = NULL; int is_label = elf_sym__is_label(&sym); const char *section_name; + if (kernel && session->ref_reloc_sym.name != NULL && + strcmp(elf_name, session->ref_reloc_sym.name) == 0) + perf_session__reloc_vmlinux_maps(session, sym.st_value); + if (!is_label && !elf_sym__is_a(&sym, map->type)) continue; @@ -973,7 +977,6 @@ static int dso__load_sym(struct dso *self, struct map *map, if (is_label && !elf_sec__is_a(&shdr, secstrs, map->type)) continue; - elf_name = elf_sym__name(&sym, symstrs); section_name = elf_sec__name(&shdr, secstrs); if (kernel || kmodule) { -- cgit v0.10.2 From ec3a9039601af210fca4650d229621fe5a21df0b Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 5 Jan 2010 17:46:41 -0500 Subject: tracing/kprobe: Update example output in documentation Update example output in documentation according to current implementation. Signed-off-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <20100105224641.19431.34967.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/Documentation/trace/kprobetrace.txt b/Documentation/trace/kprobetrace.txt index 47aabee..c3eff6f 100644 --- a/Documentation/trace/kprobetrace.txt +++ b/Documentation/trace/kprobetrace.txt @@ -97,23 +97,24 @@ recording return value as "myretprobe" event. cat /sys/kernel/debug/tracing/events/kprobes/myprobe/format name: myprobe -ID: 75 +ID: 780 format: - field:unsigned short common_type; offset:0; size:2; - field:unsigned char common_flags; offset:2; size:1; - field:unsigned char common_preempt_count; offset:3; size:1; - field:int common_pid; offset:4; size:4; - field:int common_tgid; offset:8; size:4; - - field: unsigned long ip; offset:16;tsize:8; - field: int nargs; offset:24;tsize:4; - field: unsigned long dfd; offset:32;tsize:8; - field: unsigned long filename; offset:40;tsize:8; - field: unsigned long flags; offset:48;tsize:8; - field: unsigned long mode; offset:56;tsize:8; - -print fmt: "(%lx) dfd=%lx filename=%lx flags=%lx mode=%lx", REC->ip, REC->dfd, REC->filename, REC->flags, REC->mode - + field:unsigned short common_type; offset:0; size:2; signed:0; + field:unsigned char common_flags; offset:2; size:1; signed:0; + field:unsigned char common_preempt_count; offset:3; size:1;signed:0; + field:int common_pid; offset:4; size:4; signed:1; + field:int common_lock_depth; offset:8; size:4; signed:1; + + field:unsigned long __probe_ip; offset:12; size:4; signed:0; + field:int __probe_nargs; offset:16; size:4; signed:1; + field:unsigned long dfd; offset:20; size:4; signed:0; + field:unsigned long filename; offset:24; size:4; signed:0; + field:unsigned long flags; offset:28; size:4; signed:0; + field:unsigned long mode; offset:32; size:4; signed:0; + + +print fmt: "(%lx) dfd=%lx filename=%lx flags=%lx mode=%lx", REC->__probe_ip, +REC->dfd, REC->filename, REC->flags, REC->mode You can see that the event has 4 arguments as in the expressions you specified. -- cgit v0.10.2 From 14640106f243a3b29944d7198569090fa6546f2d Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 5 Jan 2010 17:46:48 -0500 Subject: tracing/kprobe: Drop function argument access syntax Drop function argument access syntax, because the function arguments depend on not only architecture but also compile-options and function API. And now, we have perf-probe for finding register/memory assigned to each argument. Signed-off-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Roland McGrath Cc: Oleg Nesterov Cc: Mahesh Salgaonkar Cc: Benjamin Herrenschmidt Cc: Michael Neuling Cc: linuxppc-dev@ozlabs.org LKML-Reference: <20100105224648.19431.52309.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/Documentation/trace/kprobetrace.txt b/Documentation/trace/kprobetrace.txt index c3eff6f..f30978e 100644 --- a/Documentation/trace/kprobetrace.txt +++ b/Documentation/trace/kprobetrace.txt @@ -37,15 +37,12 @@ Synopsis of kprobe_events @SYM[+|-offs] : Fetch memory at SYM +|- offs (SYM should be a data symbol) $stackN : Fetch Nth entry of stack (N >= 0) $stack : Fetch stack address. - $argN : Fetch function argument. (N >= 0)(*) - $retval : Fetch return value.(**) - +|-offs(FETCHARG) : Fetch memory at FETCHARG +|- offs address.(***) + $retval : Fetch return value.(*) + +|-offs(FETCHARG) : Fetch memory at FETCHARG +|- offs address.(**) NAME=FETCHARG: Set NAME as the argument name of FETCHARG. - (*) aN may not correct on asmlinkaged functions and at the middle of - function body. - (**) only for return probe. - (***) this is useful for fetching a field of data structures. + (*) only for return probe. + (**) this is useful for fetching a field of data structures. Per-Probe Event Filtering @@ -82,11 +79,14 @@ Usage examples To add a probe as a new event, write a new definition to kprobe_events as below. - echo p:myprobe do_sys_open dfd=$arg0 filename=$arg1 flags=$arg2 mode=$arg3 > /sys/kernel/debug/tracing/kprobe_events + echo p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack) > /sys/kernel/debug/tracing/kprobe_events This sets a kprobe on the top of do_sys_open() function with recording -1st to 4th arguments as "myprobe" event. As this example shows, users can -choose more familiar names for each arguments. +1st to 4th arguments as "myprobe" event. Note, which register/stack entry is +assigned to each function argument depends on arch-specific ABI. If you unsure +the ABI, please try to use probe subcommand of perf-tools (you can find it +under tools/perf/). +As this example shows, users can choose more familiar names for each arguments. echo r:myretprobe do_sys_open $retval >> /sys/kernel/debug/tracing/kprobe_events @@ -147,4 +147,3 @@ events, you need to enable it. returns from SYMBOL(e.g. "sys_open+0x1b/0x1d <- do_sys_open" means kernel returns from do_sys_open to sys_open+0x1b). - diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 47f54ab..7ac728d 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -91,11 +91,6 @@ static __kprobes unsigned long fetch_memory(struct pt_regs *regs, void *addr) return retval; } -static __kprobes unsigned long fetch_argument(struct pt_regs *regs, void *num) -{ - return regs_get_argument_nth(regs, (unsigned int)((unsigned long)num)); -} - static __kprobes unsigned long fetch_retvalue(struct pt_regs *regs, void *dummy) { @@ -231,9 +226,7 @@ static int probe_arg_string(char *buf, size_t n, struct fetch_func *ff) { int ret = -EINVAL; - if (ff->func == fetch_argument) - ret = snprintf(buf, n, "$arg%lu", (unsigned long)ff->data); - else if (ff->func == fetch_register) { + if (ff->func == fetch_register) { const char *name; name = regs_query_register_name((unsigned int)((long)ff->data)); ret = snprintf(buf, n, "%%%s", name); @@ -489,14 +482,6 @@ static int parse_probe_vars(char *arg, struct fetch_func *ff, int is_return) } } else ret = -EINVAL; - } else if (strncmp(arg, "arg", 3) == 0 && isdigit(arg[3])) { - ret = strict_strtoul(arg + 3, 10, ¶m); - if (ret || param > PARAM_MAX_ARGS) - ret = -EINVAL; - else { - ff->func = fetch_argument; - ff->data = (void *)param; - } } else ret = -EINVAL; return ret; @@ -611,7 +596,6 @@ static int create_trace_probe(int argc, char **argv) * - Add kprobe: p[:[GRP/]EVENT] KSYM[+OFFS]|KADDR [FETCHARGS] * - Add kretprobe: r[:[GRP/]EVENT] KSYM[+0] [FETCHARGS] * Fetch args: - * $argN : fetch Nth of function argument. (N:0-) * $retval : fetch return value * $stack : fetch stack address * $stackN : fetch Nth of stack (N:0-) -- cgit v0.10.2 From aa5add93e92019018e905146f8c3d3f8e3c08300 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 5 Jan 2010 17:46:56 -0500 Subject: x86/ptrace: Remove unused regs_get_argument_nth API Because of dropping function argument syntax from kprobe-tracer, we don't need this API anymore. Signed-off-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Roland McGrath Cc: Oleg Nesterov Cc: Mahesh Salgaonkar Cc: Benjamin Herrenschmidt Cc: Michael Neuling Cc: Steven Rostedt Cc: linuxppc-dev@ozlabs.org LKML-Reference: <20100105224656.19431.92588.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/ptrace.h b/arch/x86/include/asm/ptrace.h index 9d369f6..2010280 100644 --- a/arch/x86/include/asm/ptrace.h +++ b/arch/x86/include/asm/ptrace.h @@ -274,10 +274,6 @@ static inline unsigned long regs_get_kernel_stack_nth(struct pt_regs *regs, return 0; } -/* Get Nth argument at function call */ -extern unsigned long regs_get_argument_nth(struct pt_regs *regs, - unsigned int n); - /* * These are defined as per linux/ptrace.h, which see. */ diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 017d937..73554a3 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -140,30 +140,6 @@ static const int arg_offs_table[] = { #endif }; -/** - * regs_get_argument_nth() - get Nth argument at function call - * @regs: pt_regs which contains registers at function entry. - * @n: argument number. - * - * regs_get_argument_nth() returns @n th argument of a function call. - * Since usually the kernel stack will be changed right after function entry, - * you must use this at function entry. If the @n th entry is NOT in the - * kernel stack or pt_regs, this returns 0. - */ -unsigned long regs_get_argument_nth(struct pt_regs *regs, unsigned int n) -{ - if (n < ARRAY_SIZE(arg_offs_table)) - return *(unsigned long *)((char *)regs + arg_offs_table[n]); - else { - /* - * The typical case: arg n is on the stack. - * (Note: stack[0] = return address, so skip it) - */ - n -= ARRAY_SIZE(arg_offs_table); - return regs_get_kernel_stack_nth(regs, 1 + n); - } -} - /* * does not yet catch signals sent when the child dies. * in exit.c or in signal.c. -- cgit v0.10.2 From bbaa46fac6d1c652bfa6282420d36a44bdc53b64 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 5 Jan 2010 17:47:03 -0500 Subject: perf probe: Remove newline from die() Remove newline from die(), because it is automatically added. Signed-off-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Mike Galbraith LKML-Reference: <20100105224703.19431.42475.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c index 4b852c0..6402798 100644 --- a/tools/perf/util/probe-finder.c +++ b/tools/perf/util/probe-finder.c @@ -402,11 +402,11 @@ static void show_location(Dwarf_Loc *loc, struct probe_finder *pf) } else if (op == DW_OP_regx) { regn = loc->lr_number; } else - die("Dwarf_OP %d is not supported.\n", op); + die("Dwarf_OP %d is not supported.", op); regs = get_arch_regstr(regn); if (!regs) - die("%lld exceeds max register number.\n", regn); + die("%lld exceeds max register number.", regn); if (deref) ret = snprintf(pf->buf, pf->len, @@ -438,7 +438,7 @@ static void show_variable(Dwarf_Die vr_die, struct probe_finder *pf) return ; error: die("Failed to find the location of %s at this address.\n" - " Perhaps, it has been optimized out.\n", pf->var); + " Perhaps, it has been optimized out.", pf->var); } static int variable_callback(struct die_link *dlink, void *data) @@ -476,7 +476,7 @@ static void find_variable(Dwarf_Die sp_die, struct probe_finder *pf) /* Search child die for local variables and parameters. */ ret = search_die_from_children(sp_die, variable_callback, pf); if (!ret) - die("Failed to find '%s' in this function.\n", pf->var); + die("Failed to find '%s' in this function.", pf->var); } /* Get a frame base on the address */ @@ -602,7 +602,7 @@ static void find_by_line(struct probe_finder *pf) ret = search_die_from_children(pf->cu_die, probeaddr_callback, pf); if (ret == 0) - die("Probe point is not found in subprograms.\n"); + die("Probe point is not found in subprograms."); /* Continuing, because target line might be inlined. */ } dwarf_srclines_dealloc(__dw_debug, lines, cnt); @@ -661,7 +661,7 @@ static int probefunc_callback(struct die_link *dlink, void *data) !die_inlined_subprogram(lk->die)) goto found; } - die("Failed to find real subprogram.\n"); + die("Failed to find real subprogram."); found: /* Get offset from subprogram */ ret = die_within_subprogram(lk->die, pf->addr, &offs); -- cgit v0.10.2 From 72041334b8c75ae7e1da2f17ba2b7afee8f2abd7 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 5 Jan 2010 17:47:10 -0500 Subject: perf probe: Show probe list in pager Show probe list in pager, because the list can be longer than a page. Signed-off-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Mike Galbraith LKML-Reference: <20100105224710.19431.61542.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 8e532d9..a22141a 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -37,6 +37,7 @@ #include "string.h" #include "strlist.h" #include "debug.h" +#include "cache.h" #include "parse-events.h" /* For debugfs_path */ #include "probe-event.h" @@ -455,6 +456,8 @@ void show_perf_probe_events(void) struct strlist *rawlist; struct str_node *ent; + setup_pager(); + fd = open_kprobe_events(O_RDONLY, 0); rawlist = get_trace_kprobe_event_rawlist(fd); close(fd); -- cgit v0.10.2 From fb1d2edf7ee25a26ad0b238d0ee335a3b28b7aa3 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 5 Jan 2010 17:47:17 -0500 Subject: perf tools: Support tracepoint glob matching Support glob wildcard when selecting tracepoint events by -e option. Without this patch, perf-tools supports 'GROUP:*:record' syntax for selecting all tracepoints under GROUP group. With this patch, user can choose tracepoints more flexibly by using partial wildcards, e.g. 'block:*bio*:record'. Signed-off-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Mike Galbraith LKML-Reference: <20100105224717.19431.68972.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 609d5a9..05d0c5c 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -450,7 +450,8 @@ parse_single_tracepoint_event(char *sys_name, /* sys + ':' + event + ':' + flags*/ #define MAX_EVOPT_LEN (MAX_EVENT_LENGTH * 2 + 2 + 128) static enum event_result -parse_subsystem_tracepoint_event(char *sys_name, char *flags) +parse_multiple_tracepoint_event(char *sys_name, const char *evt_exp, + char *flags) { char evt_path[MAXPATHLEN]; struct dirent *evt_ent; @@ -474,6 +475,9 @@ parse_subsystem_tracepoint_event(char *sys_name, char *flags) || !strcmp(evt_ent->d_name, "filter")) continue; + if (!strglobmatch(evt_ent->d_name, evt_exp)) + continue; + len = snprintf(event_opt, MAX_EVOPT_LEN, "%s:%s%s%s", sys_name, evt_ent->d_name, flags ? ":" : "", flags ?: ""); @@ -522,9 +526,10 @@ static enum event_result parse_tracepoint_event(const char **strp, if (evt_length >= MAX_EVENT_LENGTH) return EVT_FAILED; - if (!strcmp(evt_name, "*")) { + if (strpbrk(evt_name, "*?")) { *strp = evt_name + evt_length; - return parse_subsystem_tracepoint_event(sys_name, flags); + return parse_multiple_tracepoint_event(sys_name, evt_name, + flags); } else return parse_single_tracepoint_event(sys_name, evt_name, evt_length, flags, -- cgit v0.10.2 From 6964cd2c8efe6e048401f1fe3952a06c563c34c1 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 5 Jan 2010 17:47:24 -0500 Subject: perf tools: Enhance glob string matching Enhance strglobmatch() for supporting character classes([CHARS], complementation and ranges are also supported) and escaped special characters (\*, \? etc). Signed-off-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Mike Galbraith LKML-Reference: <20100105224724.19431.56271.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/string.c b/tools/perf/util/string.c index 5352d7d..c397d4f 100644 --- a/tools/perf/util/string.c +++ b/tools/perf/util/string.c @@ -227,16 +227,73 @@ fail: return NULL; } -/* Glob expression pattern matching */ +/* Character class matching */ +static bool __match_charclass(const char *pat, char c, const char **npat) +{ + bool complement = false, ret = true; + + if (*pat == '!') { + complement = true; + pat++; + } + if (*pat++ == c) /* First character is special */ + goto end; + + while (*pat && *pat != ']') { /* Matching */ + if (*pat == '-' && *(pat + 1) != ']') { /* Range */ + if (*(pat - 1) <= c && c <= *(pat + 1)) + goto end; + if (*(pat - 1) > *(pat + 1)) + goto error; + pat += 2; + } else if (*pat++ == c) + goto end; + } + if (!*pat) + goto error; + ret = false; + +end: + while (*pat && *pat != ']') /* Searching closing */ + pat++; + if (!*pat) + goto error; + *npat = pat + 1; + return complement ? !ret : ret; + +error: + return false; +} + +/** + * strglobmatch - glob expression pattern matching + * @str: the target string to match + * @pat: the pattern string to match + * + * This returns true if the @str matches @pat. @pat can includes wildcards + * ('*','?') and character classes ([CHARS], complementation and ranges are + * also supported). Also, this supports escape character ('\') to use special + * characters as normal character. + * + * Note: if @pat syntax is broken, this always returns false. + */ bool strglobmatch(const char *str, const char *pat) { while (*str && *pat && *pat != '*') { - if (*pat == '?') { + if (*pat == '?') { /* Matches any single character */ str++; pat++; - } else - if (*str++ != *pat++) + continue; + } else if (*pat == '[') /* Character classes/Ranges */ + if (__match_charclass(pat + 1, *str, &pat)) { + str++; + continue; + } else return false; + else if (*pat == '\\') /* Escaped char match as normal char */ + pat++; + if (*str++ != *pat++) + return false; } /* Check wild card */ if (*pat == '*') { -- cgit v0.10.2 From 631c9def804b2c92b5cca04fb9ff7b5df9e35094 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Wed, 6 Jan 2010 09:45:34 -0500 Subject: perf probe: Support --line option to show probable source-code lines Add --line option to support showing probable source-code lines. perf probe --line SRC:LN[-LN|+NUM] or perf probe --line FUNC[:LN[-LN|+NUM]] This option shows source-code with line number if the line can be probed. Lines without line number (and blue color) means that the line can not be probed, because debuginfo doesn't have the information of those lines. The argument specifies the range of lines, "source.c:100-120" shows lines between 100th to l20th in source.c file. And "func:10+20" shows 20 lines from 10th line of func function. e.g. # ./perf probe --line kernel/sched.c:1080 * * called with rq->lock held and irqs disabled */ static void hrtick_start(struct rq *rq, u64 delay) { struct hrtimer *timer = &rq->hrtick_timer; 1086 ktime_t time = ktime_add_ns(timer->base->get_time(), delay); hrtimer_set_expires(timer, time); 1090 if (rq == this_rq()) { 1091 hrtimer_restart(timer); 1092 } else if (!rq->hrtick_csd_pending) { 1093 __smp_call_function_single(cpu_of(rq), &rq->hrtick_csd, 1094 rq->hrtick_csd_pending = 1; If you specifying function name, this shows function-relative line number. # ./perf probe --line schedule asmlinkage void __sched schedule(void) 1 { struct task_struct *prev, *next; unsigned long *switch_count; struct rq *rq; int cpu; need_resched: preempt_disable(); 9 cpu = smp_processor_id(); 10 rq = cpu_rq(cpu); 11 rcu_sched_qs(cpu); 12 prev = rq->curr; 13 switch_count = &prev->nivcsw; Signed-off-by: Masami Hiramatsu Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Mike Galbraith LKML-Reference: <20100106144534.27218.77939.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Documentation/perf-probe.txt b/tools/perf/Documentation/perf-probe.txt index 250e391..2de3407 100644 --- a/tools/perf/Documentation/perf-probe.txt +++ b/tools/perf/Documentation/perf-probe.txt @@ -15,6 +15,8 @@ or 'perf probe' [options] --del='[GROUP:]EVENT' [...] or 'perf probe' --list +or +'perf probe' --line='FUNC[:RLN[+NUM|:RLN2]]|SRC:ALN[+NUM|:ALN2]' DESCRIPTION ----------- @@ -45,6 +47,11 @@ OPTIONS --list:: List up current probe events. +-L:: +--line=:: + Show source code lines which can be probed. This needs an argument + which specifies a range of the source code. + PROBE SYNTAX ------------ Probe points are defined by following syntax. @@ -56,6 +63,19 @@ Probe points are defined by following syntax. It is also possible to specify a probe point by the source line number by using 'SRC:ALN' syntax, where 'SRC' is the source file path and 'ALN' is the line number. 'ARG' specifies the arguments of this probe point. You can use the name of local variable, or kprobe-tracer argument format (e.g. $retval, %ax, etc). +LINE SYNTAX +----------- +Line range is descripted by following syntax. + + "FUNC[:RLN[+NUM|:RLN2]]|SRC:ALN[+NUM|:ALN2]" + +FUNC specifies the function name of showing lines. 'RLN' is the start line +number from function entry line, and 'RLN2' is the end line number. As same as +probe syntax, 'SRC' means the source file path, 'ALN' is start line number, +and 'ALN2' is end line number in the file. It is also possible to specify how +many lines to show by using 'NUM'. +So, "source.c:100-120" shows lines between 100th to l20th in source.c file. And "func:10+20" shows 20 lines from 10th line of func function. + SEE ALSO -------- linkperf:perf-trace[1], linkperf:perf-record[1] diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index ffdd3fe..1d3a99e 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -55,11 +55,13 @@ static struct { bool need_dwarf; bool list_events; bool force_add; + bool show_lines; int nr_probe; struct probe_point probes[MAX_PROBES]; struct strlist *dellist; struct perf_session *psession; struct map *kmap; + struct line_range line_range; } session; @@ -116,6 +118,15 @@ static int opt_del_probe_event(const struct option *opt __used, return 0; } +static int opt_show_lines(const struct option *opt __used, + const char *str, int unset __used) +{ + if (str) + parse_line_range_desc(str, &session.line_range); + INIT_LIST_HEAD(&session.line_range.line_list); + session.show_lines = true; + return 0; +} /* Currently just checking function name from symbol map */ static void evaluate_probe_point(struct probe_point *pp) { @@ -144,6 +155,7 @@ static const char * const probe_usage[] = { "perf probe [] --add 'PROBEDEF' [--add 'PROBEDEF' ...]", "perf probe [] --del '[GROUP:]EVENT' ...", "perf probe --list", + "perf probe --line 'LINEDESC'", NULL }; @@ -182,9 +194,32 @@ static const struct option options[] = { opt_add_probe_event), OPT_BOOLEAN('f', "force", &session.force_add, "forcibly add events" " with existing name"), +#ifndef NO_LIBDWARF + OPT_CALLBACK('L', "line", NULL, + "FUNC[:RLN[+NUM|:RLN2]]|SRC:ALN[+NUM|:ALN2]", + "Show source code lines.", opt_show_lines), +#endif OPT_END() }; +/* Initialize symbol maps for vmlinux */ +static void init_vmlinux(void) +{ + symbol_conf.sort_by_name = true; + if (symbol_conf.vmlinux_name == NULL) + symbol_conf.try_vmlinux_path = true; + else + pr_debug("Use vmlinux: %s\n", symbol_conf.vmlinux_name); + if (symbol__init() < 0) + die("Failed to init symbol map."); + session.psession = perf_session__new(NULL, O_WRONLY, false); + if (session.psession == NULL) + die("Failed to init perf_session."); + session.kmap = session.psession->vmlinux_maps[MAP__FUNCTION]; + if (!session.kmap) + die("Could not find kernel map.\n"); +} + int cmd_probe(int argc, const char **argv, const char *prefix __used) { int i, ret; @@ -203,7 +238,8 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) parse_probe_event_argv(argc, argv); } - if ((!session.nr_probe && !session.dellist && !session.list_events)) + if ((!session.nr_probe && !session.dellist && !session.list_events && + !session.show_lines)) usage_with_options(probe_usage, options); if (debugfs_valid_mountpoint(debugfs_path) < 0) @@ -215,10 +251,34 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) " --add/--del.\n"); usage_with_options(probe_usage, options); } + if (session.show_lines) { + pr_warning(" Error: Don't use --list with --line.\n"); + usage_with_options(probe_usage, options); + } show_perf_probe_events(); return 0; } +#ifndef NO_LIBDWARF + if (session.show_lines) { + if (session.nr_probe != 0 || session.dellist) { + pr_warning(" Error: Don't use --line with" + " --add/--del.\n"); + usage_with_options(probe_usage, options); + } + init_vmlinux(); + fd = open_vmlinux(); + if (fd < 0) + die("Could not open debuginfo file."); + ret = find_line_range(fd, &session.line_range); + if (ret <= 0) + die("Source line is not found.\n"); + close(fd); + show_line_range(&session.line_range); + return 0; + } +#endif + if (session.dellist) { del_trace_kprobe_events(session.dellist); strlist__delete(session.dellist); @@ -226,18 +286,8 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) return 0; } - /* Initialize symbol maps for vmlinux */ - symbol_conf.sort_by_name = true; - if (symbol_conf.vmlinux_name == NULL) - symbol_conf.try_vmlinux_path = true; - if (symbol__init() < 0) - die("Failed to init symbol map."); - session.psession = perf_session__new(NULL, O_WRONLY, false); - if (session.psession == NULL) - die("Failed to init perf_session."); - session.kmap = session.psession->vmlinux_maps[MAP__FUNCTION]; - if (!session.kmap) - die("Could not find kernel map.\n"); + /* Add probes */ + init_vmlinux(); if (session.need_dwarf) #ifdef NO_LIBDWARF diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index a22141a..71b0dd5 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -38,6 +38,7 @@ #include "strlist.h" #include "debug.h" #include "cache.h" +#include "color.h" #include "parse-events.h" /* For debugfs_path */ #include "probe-event.h" @@ -63,6 +64,42 @@ static int e_snprintf(char *str, size_t size, const char *format, ...) return ret; } +void parse_line_range_desc(const char *arg, struct line_range *lr) +{ + const char *ptr; + char *tmp; + /* + * + * SRC:SLN[+NUM|-ELN] + * FUNC[:SLN[+NUM|-ELN]] + */ + ptr = strchr(arg, ':'); + if (ptr) { + lr->start = (unsigned int)strtoul(ptr + 1, &tmp, 0); + if (*tmp == '+') + lr->end = lr->start + (unsigned int)strtoul(tmp + 1, + &tmp, 0); + else if (*tmp == '-') + lr->end = (unsigned int)strtoul(tmp + 1, &tmp, 0); + else + lr->end = 0; + pr_debug("Line range is %u to %u\n", lr->start, lr->end); + if (lr->end && lr->start > lr->end) + semantic_error("Start line must be smaller" + " than end line."); + if (*tmp != '\0') + semantic_error("Tailing with invalid character '%d'.", + *tmp); + tmp = strndup(arg, (ptr - arg)); + } else + tmp = strdup(arg); + + if (strchr(tmp, '.')) + lr->file = tmp; + else + lr->function = tmp; +} + /* Check the name is good for event/group */ static bool check_event_name(const char *name) { @@ -678,3 +715,66 @@ void del_trace_kprobe_events(struct strlist *dellist) close(fd); } +#define LINEBUF_SIZE 256 + +static void show_one_line(FILE *fp, unsigned int l, bool skip, bool show_num) +{ + char buf[LINEBUF_SIZE]; + const char *color = PERF_COLOR_BLUE; + + if (fgets(buf, LINEBUF_SIZE, fp) == NULL) + goto error; + if (!skip) { + if (show_num) + fprintf(stdout, "%7u %s", l, buf); + else + color_fprintf(stdout, color, " %s", buf); + } + + while (strlen(buf) == LINEBUF_SIZE - 1 && + buf[LINEBUF_SIZE - 2] != '\n') { + if (fgets(buf, LINEBUF_SIZE, fp) == NULL) + goto error; + if (!skip) { + if (show_num) + fprintf(stdout, "%s", buf); + else + color_fprintf(stdout, color, "%s", buf); + } + } + return; +error: + if (feof(fp)) + die("Source file is shorter than expected."); + else + die("File read error: %s", strerror(errno)); +} + +void show_line_range(struct line_range *lr) +{ + unsigned int l = 1; + struct line_node *ln; + FILE *fp; + + setup_pager(); + + if (lr->function) + fprintf(stdout, "<%s:%d>\n", lr->function, + lr->start - lr->offset); + else + fprintf(stdout, "<%s:%d>\n", lr->file, lr->start); + + fp = fopen(lr->path, "r"); + if (fp == NULL) + die("Failed to open %s: %s", lr->path, strerror(errno)); + /* Skip to starting line number */ + while (l < lr->start) + show_one_line(fp, l++, true, false); + + list_for_each_entry(ln, &lr->line_list, list) { + while (ln->line > l) + show_one_line(fp, (l++) - lr->offset, false, false); + show_one_line(fp, (l++) - lr->offset, false, true); + } + fclose(fp); +} diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h index 7f1d499..711287d 100644 --- a/tools/perf/util/probe-event.h +++ b/tools/perf/util/probe-event.h @@ -5,6 +5,7 @@ #include "probe-finder.h" #include "strlist.h" +extern void parse_line_range_desc(const char *arg, struct line_range *lr); extern void parse_perf_probe_event(const char *str, struct probe_point *pp, bool *need_dwarf); extern int synthesize_perf_probe_point(struct probe_point *pp); @@ -15,6 +16,7 @@ extern void add_trace_kprobe_events(struct probe_point *probes, int nr_probes, bool force_add); extern void del_trace_kprobe_events(struct strlist *dellist); extern void show_perf_probe_events(void); +extern void show_line_range(struct line_range *lr); /* Maximum index number of event-name postfix */ #define MAX_EVENT_INDEX 1024 diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c index 6402798..1b2124d 100644 --- a/tools/perf/util/probe-finder.c +++ b/tools/perf/util/probe-finder.c @@ -140,6 +140,31 @@ static Dwarf_Unsigned cu_find_fileno(Dwarf_Die cu_die, const char *fname) return found; } +static int cu_get_filename(Dwarf_Die cu_die, Dwarf_Unsigned fno, char **buf) +{ + Dwarf_Signed cnt, i; + char **srcs; + int ret = 0; + + if (!buf || !fno) + return -EINVAL; + + ret = dwarf_srcfiles(cu_die, &srcs, &cnt, &__dw_error); + if (ret == DW_DLV_OK) { + if ((Dwarf_Unsigned)cnt > fno - 1) { + *buf = strdup(srcs[fno - 1]); + ret = 0; + pr_debug("found filename: %s\n", *buf); + } else + ret = -ENOENT; + for (i = 0; i < cnt; i++) + dwarf_dealloc(__dw_debug, srcs[i], DW_DLA_STRING); + dwarf_dealloc(__dw_debug, srcs, DW_DLA_LIST); + } else + ret = -EINVAL; + return ret; +} + /* Compare diename and tname */ static int die_compare_name(Dwarf_Die dw_die, const char *tname) { @@ -567,7 +592,7 @@ static int probeaddr_callback(struct die_link *dlink, void *data) } /* Find probe point from its line number */ -static void find_by_line(struct probe_finder *pf) +static void find_probe_point_by_line(struct probe_finder *pf) { Dwarf_Signed cnt, i, clm; Dwarf_Line *lines; @@ -626,7 +651,7 @@ static int probefunc_callback(struct die_link *dlink, void *data) pf->fno = die_get_decl_file(dlink->die); pf->lno = die_get_decl_line(dlink->die) + pp->line; - find_by_line(pf); + find_probe_point_by_line(pf); return 1; } if (die_inlined_subprogram(dlink->die)) { @@ -673,7 +698,7 @@ found: return 0; } -static void find_by_func(struct probe_finder *pf) +static void find_probe_point_by_func(struct probe_finder *pf) { search_die_from_children(pf->cu_die, probefunc_callback, pf); } @@ -714,10 +739,10 @@ int find_probepoint(int fd, struct probe_point *pp) if (ret == DW_DLV_NO_ENTRY) pf.cu_base = 0; if (pp->function) - find_by_func(&pf); + find_probe_point_by_func(&pf); else { pf.lno = pp->line; - find_by_line(&pf); + find_probe_point_by_line(&pf); } } dwarf_dealloc(__dw_debug, pf.cu_die, DW_DLA_DIE); @@ -728,3 +753,159 @@ int find_probepoint(int fd, struct probe_point *pp) return pp->found; } + +static void line_range_add_line(struct line_range *lr, unsigned int line) +{ + struct line_node *ln; + struct list_head *p; + + /* Reverse search, because new line will be the last one */ + list_for_each_entry_reverse(ln, &lr->line_list, list) { + if (ln->line < line) { + p = &ln->list; + goto found; + } else if (ln->line == line) /* Already exist */ + return ; + } + /* List is empty, or the smallest entry */ + p = &lr->line_list; +found: + pr_debug("Debug: add a line %u\n", line); + ln = zalloc(sizeof(struct line_node)); + DIE_IF(ln == NULL); + ln->line = line; + INIT_LIST_HEAD(&ln->list); + list_add(&ln->list, p); +} + +/* Find line range from its line number */ +static void find_line_range_by_line(struct line_finder *lf) +{ + Dwarf_Signed cnt, i; + Dwarf_Line *lines; + Dwarf_Unsigned lineno = 0; + Dwarf_Unsigned fno; + Dwarf_Addr addr; + int ret; + + ret = dwarf_srclines(lf->cu_die, &lines, &cnt, &__dw_error); + DIE_IF(ret != DW_DLV_OK); + + for (i = 0; i < cnt; i++) { + ret = dwarf_line_srcfileno(lines[i], &fno, &__dw_error); + DIE_IF(ret != DW_DLV_OK); + if (fno != lf->fno) + continue; + + ret = dwarf_lineno(lines[i], &lineno, &__dw_error); + DIE_IF(ret != DW_DLV_OK); + if (lf->lno_s > lineno || lf->lno_e < lineno) + continue; + + /* Filter line in the function address range */ + if (lf->addr_s && lf->addr_e) { + ret = dwarf_lineaddr(lines[i], &addr, &__dw_error); + DIE_IF(ret != DW_DLV_OK); + if (lf->addr_s > addr || lf->addr_e <= addr) + continue; + } + line_range_add_line(lf->lr, (unsigned int)lineno); + } + dwarf_srclines_dealloc(__dw_debug, lines, cnt); + if (!list_empty(&lf->lr->line_list)) + lf->found = 1; +} + +/* Search function from function name */ +static int linefunc_callback(struct die_link *dlink, void *data) +{ + struct line_finder *lf = (struct line_finder *)data; + struct line_range *lr = lf->lr; + Dwarf_Half tag; + int ret; + + ret = dwarf_tag(dlink->die, &tag, &__dw_error); + DIE_IF(ret == DW_DLV_ERROR); + if (tag == DW_TAG_subprogram && + die_compare_name(dlink->die, lr->function) == 0) { + /* Get the address range of this function */ + ret = dwarf_highpc(dlink->die, &lf->addr_e, &__dw_error); + if (ret == DW_DLV_OK) + ret = dwarf_lowpc(dlink->die, &lf->addr_s, &__dw_error); + DIE_IF(ret == DW_DLV_ERROR); + if (ret == DW_DLV_NO_ENTRY) { + lf->addr_s = 0; + lf->addr_e = 0; + } + + lf->fno = die_get_decl_file(dlink->die); + lr->offset = die_get_decl_line(dlink->die);; + lf->lno_s = lr->offset + lr->start; + if (!lr->end) + lf->lno_e = (Dwarf_Unsigned)-1; + else + lf->lno_e = lr->offset + lr->end; + lr->start = lf->lno_s; + lr->end = lf->lno_e; + find_line_range_by_line(lf); + /* If we find a target function, this should be end. */ + lf->found = 1; + return 1; + } + return 0; +} + +static void find_line_range_by_func(struct line_finder *lf) +{ + search_die_from_children(lf->cu_die, linefunc_callback, lf); +} + +int find_line_range(int fd, struct line_range *lr) +{ + Dwarf_Half addr_size = 0; + Dwarf_Unsigned next_cuh = 0; + int ret; + struct line_finder lf = {.lr = lr}; + + ret = dwarf_init(fd, DW_DLC_READ, 0, 0, &__dw_debug, &__dw_error); + if (ret != DW_DLV_OK) + return -ENOENT; + + while (!lf.found) { + /* Search CU (Compilation Unit) */ + ret = dwarf_next_cu_header(__dw_debug, NULL, NULL, NULL, + &addr_size, &next_cuh, &__dw_error); + DIE_IF(ret == DW_DLV_ERROR); + if (ret == DW_DLV_NO_ENTRY) + break; + + /* Get the DIE(Debugging Information Entry) of this CU */ + ret = dwarf_siblingof(__dw_debug, 0, &lf.cu_die, &__dw_error); + DIE_IF(ret != DW_DLV_OK); + + /* Check if target file is included. */ + if (lr->file) + lf.fno = cu_find_fileno(lf.cu_die, lr->file); + + if (!lr->file || lf.fno) { + if (lr->function) + find_line_range_by_func(&lf); + else { + lf.lno_s = lr->start; + if (!lr->end) + lf.lno_e = (Dwarf_Unsigned)-1; + else + lf.lno_e = lr->end; + find_line_range_by_line(&lf); + } + /* Get the real file path */ + if (lf.found) + cu_get_filename(lf.cu_die, lf.fno, &lr->path); + } + dwarf_dealloc(__dw_debug, lf.cu_die, DW_DLA_DIE); + } + ret = dwarf_finish(__dw_debug, &__dw_error); + DIE_IF(ret != DW_DLV_OK); + return lf.found; +} + diff --git a/tools/perf/util/probe-finder.h b/tools/perf/util/probe-finder.h index e3f3968..972b386 100644 --- a/tools/perf/util/probe-finder.h +++ b/tools/perf/util/probe-finder.h @@ -34,8 +34,26 @@ struct probe_point { char *probes[MAX_PROBES]; /* Output buffers (will be allocated)*/ }; +/* Line number container */ +struct line_node { + struct list_head list; + unsigned int line; +}; + +/* Line range */ +struct line_range { + char *file; /* File name */ + char *function; /* Function name */ + unsigned int start; /* Start line number */ + unsigned int end; /* End line number */ + unsigned int offset; /* Start line offset */ + char *path; /* Real path name */ + struct list_head line_list; /* Visible lines */ +}; + #ifndef NO_LIBDWARF extern int find_probepoint(int fd, struct probe_point *pp); +extern int find_line_range(int fd, struct line_range *lr); /* Workaround for undefined _MIPS_SZLONG bug in libdwarf.h: */ #ifndef _MIPS_SZLONG @@ -62,6 +80,19 @@ struct probe_finder { char *buf; /* Current output buffer */ int len; /* Length of output buffer */ }; + +struct line_finder { + struct line_range *lr; /* Target line range */ + + Dwarf_Unsigned fno; /* File number */ + Dwarf_Unsigned lno_s; /* Start line number */ + Dwarf_Unsigned lno_e; /* End line number */ + Dwarf_Addr addr_s; /* Start address */ + Dwarf_Addr addr_e; /* End address */ + Dwarf_Die cu_die; /* Current CU */ + int found; +}; + #endif /* NO_LIBDWARF */ #endif /*_PROBE_FINDER_H */ -- cgit v0.10.2 From 8d9e503928638fc95317be42c416fb7907322aff Mon Sep 17 00:00:00 2001 From: Alexander Beregalov Date: Thu, 7 Jan 2010 19:40:47 +0300 Subject: perf: Fix memory leak: counterwidth Signed-off-by: Alexander Beregalov Cc: a.p.zijlstra@chello.nl Cc: paulus@samba.org LKML-Reference: <1262882447-23776-2-git-send-email-a.beregalov@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/values.c b/tools/perf/util/values.c index 1c15e39..cfa55d6 100644 --- a/tools/perf/util/values.c +++ b/tools/perf/util/values.c @@ -169,6 +169,7 @@ static void perf_read_values__display_pretty(FILE *fp, counterwidth[j], values->value[i][j]); fprintf(fp, "\n"); } + free(counterwidth); } static void perf_read_values__display_raw(FILE *fp, -- cgit v0.10.2 From fed5af61dc0d9402d26e7fb8fb9731a60a8e05ca Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 7 Jan 2010 19:59:38 -0200 Subject: perf buildid-list: No need to process the header sections again MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As it is already processed by: perf_session__new perf_session__open perf_session__read This was harmless, because we use dsos__findnew, that would already find it, but is unnecessary work and removing it makes builtin-buildid-list.c even shorter. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262901583-8074-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-buildid-list.c b/tools/perf/builtin-buildid-list.c index 1e99ac8..4229c2c 100644 --- a/tools/perf/builtin-buildid-list.c +++ b/tools/perf/builtin-buildid-list.c @@ -31,26 +31,6 @@ static const struct option options[] = { OPT_END() }; -static int perf_file_section__process_buildids(struct perf_file_section *self, - int feat, int fd) -{ - if (feat != HEADER_BUILD_ID) - return 0; - - if (lseek(fd, self->offset, SEEK_SET) < 0) { - pr_warning("Failed to lseek to %Ld offset for buildids!\n", - self->offset); - return -1; - } - - if (perf_header__read_build_ids(fd, self->offset, self->size)) { - pr_warning("Failed to read buildids!\n"); - return -1; - } - - return 0; -} - static int __cmd_buildid_list(void) { int err = -1; @@ -60,10 +40,7 @@ static int __cmd_buildid_list(void) if (session == NULL) return -1; - err = perf_header__process_sections(&session->header, session->fd, - perf_file_section__process_buildids); - if (err >= 0) - dsos__fprintf_buildid(stdout); + dsos__fprintf_buildid(stdout); perf_session__delete(session); return err; -- cgit v0.10.2 From a89e5abe3efcc7facc666d3985769278937f86b0 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 7 Jan 2010 19:59:39 -0200 Subject: perf symbols: Record the domain of DSOs in HEADER_BUILD_ID header table MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit So that we can restore them to the right DSO list (either dsos__kernel or dsos__user). We do that just like the kernel does for the other events, encoding PERF_RECORD_MISC_{KERNEL,USER} in perf_event_header. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262901583-8074-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 942f7da..ec96321 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -193,7 +193,7 @@ static int write_padded(int fd, const void *bf, size_t count, continue; \ else -static int __dsos__write_buildid_table(struct list_head *head, int fd) +static int __dsos__write_buildid_table(struct list_head *head, u16 misc, int fd) { struct dso *pos; @@ -205,6 +205,7 @@ static int __dsos__write_buildid_table(struct list_head *head, int fd) len = ALIGN(len, NAME_ALIGN); memset(&b, 0, sizeof(b)); memcpy(&b.build_id, pos->build_id, sizeof(pos->build_id)); + b.header.misc = misc; b.header.size = sizeof(b) + len; err = do_write(fd, &b, sizeof(b)); if (err < 0) @@ -220,9 +221,11 @@ static int __dsos__write_buildid_table(struct list_head *head, int fd) static int dsos__write_buildid_table(int fd) { - int err = __dsos__write_buildid_table(&dsos__kernel, fd); + int err = __dsos__write_buildid_table(&dsos__kernel, + PERF_RECORD_MISC_KERNEL, fd); if (err == 0) - err = __dsos__write_buildid_table(&dsos__user, fd); + err = __dsos__write_buildid_table(&dsos__user, + PERF_RECORD_MISC_USER, fd); return err; } diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index e0e6a07..378ac54 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -255,6 +255,7 @@ int perf_header__read_build_ids(int input, u64 offset, u64 size) while (offset < limit) { struct dso *dso; ssize_t len; + struct list_head *head = &dsos__user; if (read(input, &bev, sizeof(bev)) != sizeof(bev)) goto out; @@ -263,7 +264,10 @@ int perf_header__read_build_ids(int input, u64 offset, u64 size) if (read(input, filename, len) != len) goto out; - dso = dsos__findnew(filename); + if (bev.header.misc & PERF_RECORD_MISC_KERNEL) + head = &dsos__kernel; + + dso = __dsos__findnew(head, filename); if (dso != NULL) dso__set_build_id(dso, &bev.build_id); diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index da2f07f..8e6627e 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1615,14 +1615,14 @@ static struct dso *dsos__find(struct list_head *head, const char *name) return NULL; } -struct dso *dsos__findnew(const char *name) +struct dso *__dsos__findnew(struct list_head *head, const char *name) { - struct dso *dso = dsos__find(&dsos__user, name); + struct dso *dso = dsos__find(head, name); if (!dso) { dso = dso__new(name); if (dso != NULL) { - dsos__add(&dsos__user, dso); + dsos__add(head, dso); dso__set_basename(dso); } } diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index b2b5330..ee0b459 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -115,9 +115,17 @@ bool dso__sorted_by_name(const struct dso *self, enum map_type type); void dso__sort_by_name(struct dso *self, enum map_type type); +extern struct list_head dsos__user, dsos__kernel; + +struct dso *__dsos__findnew(struct list_head *head, const char *name); + +static inline struct dso *dsos__findnew(const char *name) +{ + return __dsos__findnew(&dsos__user, name); +} + struct perf_session; -struct dso *dsos__findnew(const char *name); int dso__load(struct dso *self, struct map *map, struct perf_session *session, symbol_filter_t filter); void dsos__fprintf(FILE *fp); @@ -143,6 +151,5 @@ bool symbol_type__is_a(char symbol_type, enum map_type map_type); int perf_session__create_kernel_maps(struct perf_session *self); -extern struct list_head dsos__user, dsos__kernel; extern struct dso *vdso; #endif /* __PERF_SYMBOL */ -- cgit v0.10.2 From cf5531148ff34938840d6da775c0a4ace442d573 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 7 Jan 2010 19:59:40 -0200 Subject: perf tools: Create typedef for common event synthesizing callback MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1262901583-8074-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 1a31feb..bfb3d87 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -8,8 +8,7 @@ #include "thread.h" static pid_t event__synthesize_comm(pid_t pid, int full, - int (*process)(event_t *event, - struct perf_session *session), + event__handler_t process, struct perf_session *session) { event_t ev; @@ -91,8 +90,7 @@ out_failure: } static int event__synthesize_mmap_events(pid_t pid, pid_t tgid, - int (*process)(event_t *event, - struct perf_session *session), + event__handler_t process, struct perf_session *session) { char filename[PATH_MAX]; @@ -156,9 +154,7 @@ static int event__synthesize_mmap_events(pid_t pid, pid_t tgid, return 0; } -int event__synthesize_thread(pid_t pid, - int (*process)(event_t *event, - struct perf_session *session), +int event__synthesize_thread(pid_t pid, event__handler_t process, struct perf_session *session) { pid_t tgid = event__synthesize_comm(pid, 1, process, session); @@ -167,8 +163,7 @@ int event__synthesize_thread(pid_t pid, return event__synthesize_mmap_events(pid, tgid, process, session); } -void event__synthesize_threads(int (*process)(event_t *event, - struct perf_session *session), +void event__synthesize_threads(event__handler_t process, struct perf_session *session) { DIR *proc; @@ -205,8 +200,7 @@ static int find_symbol_cb(void *arg, const char *name, char type, u64 start) return 1; } -int event__synthesize_kernel_mmap(int (*process)(event_t *event, - struct perf_session *session), +int event__synthesize_kernel_mmap(event__handler_t process, struct perf_session *session, const char *symbol_name) { diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 61fc0dc..80356da 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -103,15 +103,13 @@ void event__print_totals(void); struct perf_session; -int event__synthesize_thread(pid_t pid, - int (*process)(event_t *event, - struct perf_session *session), +typedef int (*event__handler_t)(event_t *event, struct perf_session *session); + +int event__synthesize_thread(pid_t pid, event__handler_t process, struct perf_session *session); -void event__synthesize_threads(int (*process)(event_t *event, - struct perf_session *session), +void event__synthesize_threads(event__handler_t process, struct perf_session *session); -int event__synthesize_kernel_mmap(int (*process)(event_t *event, - struct perf_session *session), +int event__synthesize_kernel_mmap(event__handler_t process, struct perf_session *session, const char *symbol_name); -- cgit v0.10.2 From a0ab36689a36e583b6e736f1c99ac8c9aebdad59 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 13 Jan 2010 18:31:48 +0900 Subject: sh: fixed PMB mode refactoring. This introduces some much overdue chainsawing of the fixed PMB support. fixed PMB was introduced initially to work around the fact that dynamic PMB mode was relatively broken, though they were never intended to converge. The main areas where there are differences are whether the system is booted in 29-bit mode or 32-bit mode, and whether legacy mappings are to be preserved. Any system booting in true 32-bit mode will not care about legacy mappings, so these are roughly decoupled. Regardless of the entry point, PMB and 32BIT are directly related as far as the kernel is concerned, so we also switch back to having one select the other. With legacy mappings iterated through and applied in the initialization path it's now possible to finally merge the two implementations and permit dynamic remapping overtop of remaining entries regardless of whether boot mappings are crafted by hand or inherited from the boot loader. Signed-off-by: Paul Mundt diff --git a/arch/sh/boot/Makefile b/arch/sh/boot/Makefile index 743ce0c..1ce6362 100644 --- a/arch/sh/boot/Makefile +++ b/arch/sh/boot/Makefile @@ -46,15 +46,8 @@ $(obj)/romImage: $(obj)/romimage/vmlinux FORCE $(obj)/romimage/vmlinux: $(obj)/zImage FORCE $(Q)$(MAKE) $(build)=$(obj)/romimage $@ -KERNEL_MEMORY := 0x00000000 -ifeq ($(CONFIG_PMB_FIXED),y) -KERNEL_MEMORY := $(shell /bin/bash -c 'printf "0x%08x" \ +KERNEL_MEMORY := $(shell /bin/bash -c 'printf "0x%08x" \ $$[$(CONFIG_MEMORY_START) & 0x1fffffff]') -endif -ifeq ($(CONFIG_29BIT),y) -KERNEL_MEMORY := $(shell /bin/bash -c 'printf "0x%08x" \ - $$[$(CONFIG_MEMORY_START)]') -endif KERNEL_LOAD := $(shell /bin/bash -c 'printf "0x%08x" \ $$[$(CONFIG_PAGE_OFFSET) + \ diff --git a/arch/sh/include/asm/addrspace.h b/arch/sh/include/asm/addrspace.h index 99d6b3e..bcd7d4d 100644 --- a/arch/sh/include/asm/addrspace.h +++ b/arch/sh/include/asm/addrspace.h @@ -28,7 +28,7 @@ /* Returns the privileged segment base of a given address */ #define PXSEG(a) (((unsigned long)(a)) & 0xe0000000) -#if defined(CONFIG_29BIT) || defined(CONFIG_PMB_FIXED) +#ifdef CONFIG_29BIT /* * Map an address to a certain privileged segment */ @@ -40,7 +40,7 @@ ((__typeof__(a))(((unsigned long)(a) & 0x1fffffff) | P3SEG)) #define P4SEGADDR(a) \ ((__typeof__(a))(((unsigned long)(a) & 0x1fffffff) | P4SEG)) -#endif /* 29BIT || PMB_FIXED */ +#endif /* 29BIT */ #endif /* P1SEG */ /* Check if an address can be reached in 29 bits */ diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index 026dd65..f4314d8 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -244,18 +244,11 @@ __ioremap(unsigned long offset, unsigned long size, unsigned long flags) } static inline void __iomem * -__ioremap_mode(unsigned long offset, unsigned long size, unsigned long flags) +__ioremap_29bit(unsigned long offset, unsigned long size, unsigned long flags) { -#if defined(CONFIG_SUPERH32) && !defined(CONFIG_PMB_FIXED) && !defined(CONFIG_PMB) +#ifdef CONFIG_29BIT unsigned long last_addr = offset + size - 1; -#endif - void __iomem *ret; - ret = __ioremap_trapped(offset, size); - if (ret) - return ret; - -#if defined(CONFIG_SUPERH32) && !defined(CONFIG_PMB_FIXED) && !defined(CONFIG_PMB) /* * For P1 and P2 space this is trivial, as everything is already * mapped. Uncached access for P1 addresses are done through P2. @@ -274,6 +267,22 @@ __ioremap_mode(unsigned long offset, unsigned long size, unsigned long flags) return (void __iomem *)P4SEGADDR(offset); #endif + return NULL; +} + +static inline void __iomem * +__ioremap_mode(unsigned long offset, unsigned long size, unsigned long flags) +{ + void __iomem *ret; + + ret = __ioremap_trapped(offset, size); + if (ret) + return ret; + + ret = __ioremap_29bit(offset, size, flags); + if (ret) + return ret; + return __ioremap(offset, size, flags); } #else diff --git a/arch/sh/include/asm/mmu.h b/arch/sh/include/asm/mmu.h index c7426ad..4b0882b 100644 --- a/arch/sh/include/asm/mmu.h +++ b/arch/sh/include/asm/mmu.h @@ -65,11 +65,29 @@ struct pmb_entry { struct pmb_entry *link; }; +#ifdef CONFIG_PMB /* arch/sh/mm/pmb.c */ long pmb_remap(unsigned long virt, unsigned long phys, unsigned long size, unsigned long flags); void pmb_unmap(unsigned long addr); int pmb_init(void); +#else +static inline long pmb_remap(unsigned long virt, unsigned long phys, + unsigned long size, unsigned long flags) +{ + return -EINVAL +} + +static inline void pmb_unmap(unsigned long addr) +{ +} + +static inline int pmb_init(void) +{ + return -ENODEV; +} +#endif /* CONFIG_PMB */ + #endif /* __ASSEMBLY__ */ #endif /* __MMU_H */ diff --git a/arch/sh/include/asm/page.h b/arch/sh/include/asm/page.h index 81bffc0..a86c0f1 100644 --- a/arch/sh/include/asm/page.h +++ b/arch/sh/include/asm/page.h @@ -127,7 +127,7 @@ typedef struct page *pgtable_t; * is not visible (it is part of the PMB mapping) and so needs to be * added or subtracted as required. */ -#if defined(CONFIG_PMB_FIXED) +#if defined(CONFIG_PMB_LEGACY) /* phys = virt - PAGE_OFFSET - (__MEMORY_START & 0xe0000000) */ #define PMB_OFFSET (PAGE_OFFSET - PXSEG(__MEMORY_START)) #define __pa(x) ((unsigned long)(x) - PMB_OFFSET) diff --git a/arch/sh/kernel/head_32.S b/arch/sh/kernel/head_32.S index 1151ecd..e5d421db 100644 --- a/arch/sh/kernel/head_32.S +++ b/arch/sh/kernel/head_32.S @@ -13,6 +13,8 @@ #include #include #include +#include +#include #ifdef CONFIG_CPU_SH4A #define SYNCO() synco @@ -33,7 +35,7 @@ ENTRY(empty_zero_page) .long 1 /* LOADER_TYPE */ .long 0x00000000 /* INITRD_START */ .long 0x00000000 /* INITRD_SIZE */ -#if defined(CONFIG_32BIT) && defined(CONFIG_PMB_FIXED) +#ifdef CONFIG_32BIT .long 0x53453f00 + 32 /* "SE?" = 32 bit */ #else .long 0x53453f00 + 29 /* "SE?" = 29 bit */ diff --git a/arch/sh/kernel/setup.c b/arch/sh/kernel/setup.c index 8b0e697..f79ebe3 100644 --- a/arch/sh/kernel/setup.c +++ b/arch/sh/kernel/setup.c @@ -455,10 +455,7 @@ void __init setup_arch(char **cmdline_p) sh_mv.mv_setup(cmdline_p); paging_init(); - -#ifdef CONFIG_PMB_ENABLE pmb_init(); -#endif #ifdef CONFIG_SMP plat_smp_setup(); diff --git a/arch/sh/kernel/vmlinux.lds.S b/arch/sh/kernel/vmlinux.lds.S index a1e4ec2..9e5a587 100644 --- a/arch/sh/kernel/vmlinux.lds.S +++ b/arch/sh/kernel/vmlinux.lds.S @@ -14,17 +14,16 @@ OUTPUT_ARCH(sh) #include #include +#if defined(CONFIG_32BIT) && !defined(CONFIG_PMB_LEGACY) +#define MEMORY_OFFSET 0 +#else +#define MEMORY_OFFSET (CONFIG_MEMORY_START & 0x1fffffff) +#endif + ENTRY(_start) SECTIONS { -#ifdef CONFIG_PMB_FIXED - . = CONFIG_PAGE_OFFSET + (CONFIG_MEMORY_START & 0x1fffffff) + - CONFIG_ZERO_PAGE_OFFSET; -#elif defined(CONFIG_32BIT) - . = CONFIG_PAGE_OFFSET + CONFIG_ZERO_PAGE_OFFSET; -#else - . = CONFIG_PAGE_OFFSET + CONFIG_MEMORY_START + CONFIG_ZERO_PAGE_OFFSET; -#endif + . = CONFIG_PAGE_OFFSET + MEMORY_OFFSET + CONFIG_ZERO_PAGE_OFFSET; _text = .; /* Text and read-only data */ diff --git a/arch/sh/mm/Kconfig b/arch/sh/mm/Kconfig index 358c860..860cd24 100644 --- a/arch/sh/mm/Kconfig +++ b/arch/sh/mm/Kconfig @@ -80,30 +80,18 @@ config 32BIT bool default y if CPU_SH5 -config PMB_ENABLE - bool "Support 32-bit physical addressing through PMB" - depends on MMU && EXPERIMENTAL && CPU_SH4A && !CPU_SH4AL_DSP - help - If you say Y here, physical addressing will be extended to - 32-bits through the SH-4A PMB. If this is not set, legacy - 29-bit physical addressing will be used. - -choice - prompt "PMB handling type" - depends on PMB_ENABLE - default PMB_FIXED - config PMB - bool "PMB" + bool "Support 32-bit physical addressing through PMB" depends on MMU && EXPERIMENTAL && CPU_SH4A && !CPU_SH4AL_DSP + select 32BIT help If you say Y here, physical addressing will be extended to 32-bits through the SH-4A PMB. If this is not set, legacy 29-bit physical addressing will be used. -config PMB_FIXED - bool "fixed PMB" - depends on MMU && EXPERIMENTAL && CPU_SH4A && !CPU_SH4AL_DSP +config PMB_LEGACY + bool "Support legacy boot mappings for PMB" + depends on PMB select 32BIT help If this option is enabled, fixed PMB mappings are inherited @@ -111,8 +99,6 @@ config PMB_FIXED management. This is the closest to legacy 29-bit physical mode, and allows systems to support up to 512MiB of system memory. -endchoice - config X2TLB bool "Enable extended TLB mode" depends on (CPU_SHX2 || CPU_SHX3) && MMU && EXPERIMENTAL diff --git a/arch/sh/mm/Makefile b/arch/sh/mm/Makefile index 9fa11d6..edde8bd 100644 --- a/arch/sh/mm/Makefile +++ b/arch/sh/mm/Makefile @@ -33,7 +33,7 @@ obj-y += $(tlb-y) endif obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o -obj-$(CONFIG_PMB_ENABLE) += pmb.o +obj-$(CONFIG_PMB) += pmb.o obj-$(CONFIG_NUMA) += numa.o # Special flags for fault_64.o. This puts restrictions on the number of diff --git a/arch/sh/mm/pmb.c b/arch/sh/mm/pmb.c index 280f6a1..8f7dbf1 100644 --- a/arch/sh/mm/pmb.c +++ b/arch/sh/mm/pmb.c @@ -3,7 +3,7 @@ * * Privileged Space Mapping Buffer (PMB) Support. * - * Copyright (C) 2005, 2006, 2007 Paul Mundt + * Copyright (C) 2005 - 2010 Paul Mundt * * P1/P2 Section mapping definitions from map32.h, which was: * @@ -279,51 +279,12 @@ static void __pmb_unmap(struct pmb_entry *pmbe) } while (pmbe); } -#ifdef CONFIG_PMB -int __uses_jump_to_uncached pmb_init(void) -{ - unsigned int i; - long size, ret; - - jump_to_uncached(); - - /* - * Insert PMB entries for the P1 and P2 areas so that, after - * we've switched the MMU to 32-bit mode, the semantics of P1 - * and P2 are the same as in 29-bit mode, e.g. - * - * P1 - provides a cached window onto physical memory - * P2 - provides an uncached window onto physical memory - */ - size = __MEMORY_START + __MEMORY_SIZE; - - ret = pmb_remap(P1SEG, 0x00000000, size, PMB_C); - BUG_ON(ret != size); - - ret = pmb_remap(P2SEG, 0x00000000, size, PMB_WT | PMB_UB); - BUG_ON(ret != size); - - ctrl_outl(0, PMB_IRMCR); - - /* PMB.SE and UB[7] */ - ctrl_outl(PASCR_SE | (1 << 7), PMB_PASCR); - - /* Flush out the TLB */ - i = ctrl_inl(MMUCR); - i |= MMUCR_TI; - ctrl_outl(i, MMUCR); - - back_to_cached(); - - return 0; -} -#else -int __uses_jump_to_uncached pmb_init(void) +#ifdef CONFIG_PMB_LEGACY +static int pmb_apply_legacy_mappings(void) { int i; unsigned long addr, data; - - jump_to_uncached(); + unsigned int applied = 0; for (i = 0; i < PMB_ENTRY_MAX; i++) { struct pmb_entry *pmbe; @@ -357,13 +318,69 @@ int __uses_jump_to_uncached pmb_init(void) pmbe = pmb_alloc(vpn, ppn, flags, i); WARN_ON(IS_ERR(pmbe)); + + applied++; + } + + return (applied == 0); +} +#else +static inline int pmb_apply_legacy_mappings(void) +{ + return 1; +} +#endif + +int __uses_jump_to_uncached pmb_init(void) +{ + unsigned int i; + unsigned long size, ret; + + jump_to_uncached(); + + /* + * Attempt to apply the legacy boot mappings if configured. If + * this is successful then we simply carry on with those and + * don't bother establishing additional memory mappings. Dynamic + * device mappings through pmb_remap() can still be bolted on + * after this. + */ + ret = pmb_apply_legacy_mappings(); + if (ret == 0) { + back_to_cached(); + return 0; } + /* + * Insert PMB entries for the P1 and P2 areas so that, after + * we've switched the MMU to 32-bit mode, the semantics of P1 + * and P2 are the same as in 29-bit mode, e.g. + * + * P1 - provides a cached window onto physical memory + * P2 - provides an uncached window onto physical memory + */ + size = (unsigned long)__MEMORY_START + __MEMORY_SIZE; + + ret = pmb_remap(P1SEG, 0x00000000, size, PMB_C); + BUG_ON(ret != size); + + ret = pmb_remap(P2SEG, 0x00000000, size, PMB_WT | PMB_UB); + BUG_ON(ret != size); + + ctrl_outl(0, PMB_IRMCR); + + /* PMB.SE and UB[7] */ + ctrl_outl(PASCR_SE | (1 << 7), PMB_PASCR); + + /* Flush out the TLB */ + i = ctrl_inl(MMUCR); + i |= MMUCR_TI; + ctrl_outl(i, MMUCR); + back_to_cached(); return 0; } -#endif /* CONFIG_PMB */ static int pmb_seq_show(struct seq_file *file, void *iter) { @@ -462,6 +479,5 @@ static int __init pmb_sysdev_init(void) { return sysdev_driver_register(&cpu_sysdev_class, &pmb_sysdev_driver); } - subsys_initcall(pmb_sysdev_init); #endif -- cgit v0.10.2 From 88f73d22853cb080432e91d68759323477a8b835 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 13 Jan 2010 18:37:19 +0900 Subject: sh: Fix up L2 cache comment typo. Valid sizes include 256kB, not 258kB. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/cpu/sh4/probe.c b/arch/sh/kernel/cpu/sh4/probe.c index 75189c4..21371f8 100644 --- a/arch/sh/kernel/cpu/sh4/probe.c +++ b/arch/sh/kernel/cpu/sh4/probe.c @@ -228,7 +228,7 @@ int __init detect_cpu_and_cache_system(void) * Size calculation is much more sensible * than it is for the L1. * - * Sizes are 128KB, 258KB, 512KB, and 1MB. + * Sizes are 128KB, 256KB, 512KB, and 1MB. */ size = (cvr & 0xf) << 17; -- cgit v0.10.2 From 8381f65d097dad90416808314737dd7d3ae38ea9 Mon Sep 17 00:00:00 2001 From: Jamie Iles Date: Fri, 8 Jan 2010 15:27:33 +0000 Subject: sched/perf: Make sure irqs are disabled for perf_event_task_sched_in() perf_event_task_sched_in() expects interrupts to be disabled, but on architectures with __ARCH_WANT_INTERRUPTS_ON_CTXSW defined, this isn't true. If this is defined, disable irqs around the call in finish_task_switch(). Signed-off-by: Jamie Iles Acked-by: Peter Zijlstra Cc: Russell King - ARM Linux LKML-Reference: <1262964453-27370-1-git-send-email-jamie.iles@picochip.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index e507af0..c3ad342 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2783,7 +2783,13 @@ static void finish_task_switch(struct rq *rq, struct task_struct *prev) */ prev_state = prev->state; finish_arch_switch(prev); +#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW + local_irq_disable(); +#endif /* __ARCH_WANT_INTERRUPTS_ON_CTXSW */ perf_event_task_sched_in(current); +#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW + local_irq_enable(); +#endif /* __ARCH_WANT_INTERRUPTS_ON_CTXSW */ finish_lock_switch(rq, prev); fire_sched_in_preempt_notifiers(current); -- cgit v0.10.2 From 206582c3161f165f5bf49ececa962c5f95fdf0a3 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 13 Jan 2010 18:45:12 +0900 Subject: sh: Make all PxSEGADDR() calls fatal for non-legacy configs. This stubs out all of the PxSEGADDR() wrappers for non-legacy code. 29-bit will continue to work with these, while 32-bit code will now blow up on compile rather than at runtime. The vast majority of the in-tree offenders are gone, with the only remaining culprits being unable to support 32-bit mode. Hopefully this will prevent anyone from ever using these again. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/addrspace.h b/arch/sh/include/asm/addrspace.h index bcd7d4d..268efd6 100644 --- a/arch/sh/include/asm/addrspace.h +++ b/arch/sh/include/asm/addrspace.h @@ -40,7 +40,15 @@ ((__typeof__(a))(((unsigned long)(a) & 0x1fffffff) | P3SEG)) #define P4SEGADDR(a) \ ((__typeof__(a))(((unsigned long)(a) & 0x1fffffff) | P4SEG)) -#endif /* 29BIT */ +#else +/* + * These will never work in 32-bit, don't even bother. + */ +#define P1SEGADDR(a) __futile_remapping_attempt +#define P2SEGADDR(a) __futile_remapping_attempt +#define P3SEGADDR(a) __futile_remapping_attempt +#define P4SEGADDR(a) __futile_remapping_attempt +#endif #endif /* P1SEG */ /* Check if an address can be reached in 29 bits */ -- cgit v0.10.2 From 9ca94d7c016130f9ed77f142424ace9c19742809 Mon Sep 17 00:00:00 2001 From: John Kacur Date: Mon, 11 Jan 2010 21:21:06 +0100 Subject: plist: Fix grammar mistake, and c-style mistake Signed-off-by: John Kacur Acked-by: Peter Zijlstra LKML-Reference: <1263241267-25204-2-git-send-email-jkacur@redhat.com> Signed-off-by: Ingo Molnar diff --git a/include/linux/plist.h b/include/linux/plist.h index 8227f71..6898985 100644 --- a/include/linux/plist.h +++ b/include/linux/plist.h @@ -45,7 +45,7 @@ * the insertion of new nodes. There are no nodes with duplicate * priorites on the list. * - * The nodes on the node_list is ordered by priority and can contain + * The nodes on the node_list are ordered by priority and can contain * entries which have the same priority. Those entries are ordered * FIFO * @@ -265,7 +265,7 @@ static inline int plist_node_empty(const struct plist_node *node) * * Assumes the plist is _not_ empty. */ -static inline struct plist_node* plist_first(const struct plist_head *head) +static inline struct plist_node *plist_first(const struct plist_head *head) { return list_entry(head->node_list.next, struct plist_node, plist.node_list); -- cgit v0.10.2 From ff314d3903c2843de65c2148f66f277f2440ed26 Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Wed, 13 Jan 2010 17:01:38 +0800 Subject: perf: Make cmd_to_page() function more compact Remove branch for is_perf_command. Signed-off-by: Wenji Huang Cc: fweisbec@gmail.com Cc: jkacur@redhat.com Cc: acme@redhat.com LKML-Reference: <1263373298-13282-1-git-send-email-wenji.huang@oracle.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-help.c b/tools/perf/builtin-help.c index e427d69..215b584 100644 --- a/tools/perf/builtin-help.c +++ b/tools/perf/builtin-help.c @@ -313,8 +313,6 @@ static const char *cmd_to_page(const char *perf_cmd) return "perf"; else if (!prefixcmp(perf_cmd, "perf")) return perf_cmd; - else if (is_perf_command(perf_cmd)) - return prepend("perf-", perf_cmd); else return prepend("perf-", perf_cmd); } -- cgit v0.10.2 From 782bb5a532f883540bf403afb19f735a4eefd95b Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 13 Jan 2010 19:11:14 +0900 Subject: sh: default to extended TLB support. All SH-X2 and SH-X3 parts support an extended TLB mode, which has been left as experimental since support was originally merged. Now that it's had some time to stabilize and get some exposure to various platforms, we can drop it as an option and default enable it across the board. This is also good future proofing for newer parts that will drop support for the legacy TLB mode completely. This will also force 3-level page tables for all newer parts, which is necessary both for the varying page sizes and larger memories. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/pgalloc.h b/arch/sh/include/asm/pgalloc.h index f8982f4..8c00785 100644 --- a/arch/sh/include/asm/pgalloc.h +++ b/arch/sh/include/asm/pgalloc.h @@ -9,7 +9,7 @@ extern pgd_t *pgd_alloc(struct mm_struct *); extern void pgd_free(struct mm_struct *mm, pgd_t *pgd); -#ifdef CONFIG_PGTABLE_LEVELS_3 +#if PAGETABLE_LEVELS > 2 extern void pud_populate(struct mm_struct *mm, pud_t *pudp, pmd_t *pmd); extern pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long address); extern void pmd_free(struct mm_struct *mm, pmd_t *pmd); diff --git a/arch/sh/include/asm/pgtable.h b/arch/sh/include/asm/pgtable.h index 78598ec..856ece0 100644 --- a/arch/sh/include/asm/pgtable.h +++ b/arch/sh/include/asm/pgtable.h @@ -12,7 +12,7 @@ #ifndef __ASM_SH_PGTABLE_H #define __ASM_SH_PGTABLE_H -#ifdef CONFIG_PGTABLE_LEVELS_3 +#ifdef CONFIG_X2TLB #include #else #include diff --git a/arch/sh/include/asm/pgtable_nopmd.h b/arch/sh/include/asm/pgtable_nopmd.h index f0b525b..b8355e4 100644 --- a/arch/sh/include/asm/pgtable_nopmd.h +++ b/arch/sh/include/asm/pgtable_nopmd.h @@ -6,17 +6,18 @@ /* * traditional two-level paging structure */ +#define PAGETABLE_LEVELS 2 /* PTE bits */ -#define PTE_MAGNITUDE 2 /* 32-bit PTEs */ +#define PTE_MAGNITUDE 2 /* 32-bit PTEs */ -#define PTE_SHIFT PAGE_SHIFT -#define PTE_BITS (PTE_SHIFT - PTE_MAGNITUDE) +#define PTE_SHIFT PAGE_SHIFT +#define PTE_BITS (PTE_SHIFT - PTE_MAGNITUDE) /* PGD bits */ -#define PGDIR_SHIFT (PTE_SHIFT + PTE_BITS) +#define PGDIR_SHIFT (PTE_SHIFT + PTE_BITS) -#define PTRS_PER_PGD (PAGE_SIZE / (1 << PTE_MAGNITUDE)) +#define PTRS_PER_PGD (PAGE_SIZE / (1 << PTE_MAGNITUDE)) #define USER_PTRS_PER_PGD (TASK_SIZE/PGDIR_SIZE) #endif /* __ASM_SH_PGTABLE_NOPMD_H */ diff --git a/arch/sh/include/asm/pgtable_pmd.h b/arch/sh/include/asm/pgtable_pmd.h index 42a180e..587b05e 100644 --- a/arch/sh/include/asm/pgtable_pmd.h +++ b/arch/sh/include/asm/pgtable_pmd.h @@ -7,11 +7,12 @@ * Some cores need a 3-level page table layout, for example when using * 64-bit PTEs and 4K pages. */ +#define PAGETABLE_LEVELS 3 -#define PTE_MAGNITUDE 3 /* 64-bit PTEs on extended mode SH-X2 TLB */ +#define PTE_MAGNITUDE 3 /* 64-bit PTEs on SH-X2 TLB */ /* PGD bits */ -#define PGDIR_SHIFT 30 +#define PGDIR_SHIFT 30 #define PTRS_PER_PGD 4 #define USER_PTRS_PER_PGD 2 diff --git a/arch/sh/mm/Kconfig b/arch/sh/mm/Kconfig index 860cd24..7a4ebc8 100644 --- a/arch/sh/mm/Kconfig +++ b/arch/sh/mm/Kconfig @@ -100,13 +100,8 @@ config PMB_LEGACY and allows systems to support up to 512MiB of system memory. config X2TLB - bool "Enable extended TLB mode" - depends on (CPU_SHX2 || CPU_SHX3) && MMU && EXPERIMENTAL - help - Selecting this option will enable the extended mode of the SH-X2 - TLB. For legacy SH-X behaviour and interoperability, say N. For - all of the fun new features and a willingless to submit bug reports, - say Y. + def_bool y + depends on (CPU_SHX2 || CPU_SHX3) && MMU config VSYSCALL bool "Support vsyscall page" @@ -175,31 +170,11 @@ config ARCH_MEMORY_PROBE depends on MEMORY_HOTPLUG choice - prompt "Page table layout" - default PGTABLE_LEVELS_3 if X2TLB - default PGTABLE_LEVELS_2 - -config PGTABLE_LEVELS_2 - bool "2 Levels" - help - This is the default page table layout for all SuperH CPUs. - -config PGTABLE_LEVELS_3 - bool "3 Levels" - depends on X2TLB - help - This enables a 3 level page table structure. - -endchoice - -choice prompt "Kernel page size" - default PAGE_SIZE_8KB if X2TLB default PAGE_SIZE_4KB config PAGE_SIZE_4KB bool "4kB" - depends on !MMU || !X2TLB || PGTABLE_LEVELS_3 help This is the default page size used by all SuperH CPUs. diff --git a/arch/sh/mm/pgtable.c b/arch/sh/mm/pgtable.c index e1bc548..6f21fb1 100644 --- a/arch/sh/mm/pgtable.c +++ b/arch/sh/mm/pgtable.c @@ -3,8 +3,7 @@ #define PGALLOC_GFP GFP_KERNEL | __GFP_REPEAT | __GFP_ZERO static struct kmem_cache *pgd_cachep; - -#ifdef CONFIG_PGTABLE_LEVELS_3 +#if PAGETABLE_LEVELS > 2 static struct kmem_cache *pmd_cachep; #endif @@ -22,7 +21,7 @@ void pgtable_cache_init(void) pgd_cachep = kmem_cache_create("pgd_cache", PTRS_PER_PGD * (1< 2 pmd_cachep = kmem_cache_create("pmd_cache", PTRS_PER_PMD * (1< 2 void pud_populate(struct mm_struct *mm, pud_t *pud, pmd_t *pmd) { set_pud(pud, __pud((unsigned long)pmd)); @@ -54,4 +53,4 @@ void pmd_free(struct mm_struct *mm, pmd_t *pmd) { kmem_cache_free(pmd_cachep, pmd); } -#endif /* CONFIG_PGTABLE_LEVELS_3 */ +#endif /* PAGETABLE_LEVELS > 2 */ -- cgit v0.10.2 From e44d6c4010a4c84369013932eea4721d16cbc08d Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 13 Jan 2010 19:18:39 +0900 Subject: sh: Rename split-level pgtable headers. These were originally named _nopmd and _pmd to follow their asm-generic counterparts, but we rename them to -2level and -3level for general consistency. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/pgtable-2level.h b/arch/sh/include/asm/pgtable-2level.h new file mode 100644 index 0000000..19bd89d --- /dev/null +++ b/arch/sh/include/asm/pgtable-2level.h @@ -0,0 +1,23 @@ +#ifndef __ASM_SH_PGTABLE_2LEVEL_H +#define __ASM_SH_PGTABLE_2LEVEL_H + +#include + +/* + * traditional two-level paging structure + */ +#define PAGETABLE_LEVELS 2 + +/* PTE bits */ +#define PTE_MAGNITUDE 2 /* 32-bit PTEs */ + +#define PTE_SHIFT PAGE_SHIFT +#define PTE_BITS (PTE_SHIFT - PTE_MAGNITUDE) + +/* PGD bits */ +#define PGDIR_SHIFT (PTE_SHIFT + PTE_BITS) + +#define PTRS_PER_PGD (PAGE_SIZE / (1 << PTE_MAGNITUDE)) +#define USER_PTRS_PER_PGD (TASK_SIZE/PGDIR_SIZE) + +#endif /* __ASM_SH_PGTABLE_2LEVEL_H */ diff --git a/arch/sh/include/asm/pgtable-3level.h b/arch/sh/include/asm/pgtable-3level.h new file mode 100644 index 0000000..249a985 --- /dev/null +++ b/arch/sh/include/asm/pgtable-3level.h @@ -0,0 +1,56 @@ +#ifndef __ASM_SH_PGTABLE_3LEVEL_H +#define __ASM_SH_PGTABLE_3LEVEL_H + +#include + +/* + * Some cores need a 3-level page table layout, for example when using + * 64-bit PTEs and 4K pages. + */ +#define PAGETABLE_LEVELS 3 + +#define PTE_MAGNITUDE 3 /* 64-bit PTEs on SH-X2 TLB */ + +/* PGD bits */ +#define PGDIR_SHIFT 30 + +#define PTRS_PER_PGD 4 +#define USER_PTRS_PER_PGD 2 + +/* PMD bits */ +#define PMD_SHIFT (PAGE_SHIFT + (PAGE_SHIFT - PTE_MAGNITUDE)) +#define PMD_SIZE (1UL << PMD_SHIFT) +#define PMD_MASK (~(PMD_SIZE-1)) + +#define PTRS_PER_PMD ((1 << PGDIR_SHIFT) / PMD_SIZE) + +#define pmd_ERROR(e) \ + printk("%s:%d: bad pmd %016llx.\n", __FILE__, __LINE__, pmd_val(e)) + +typedef struct { unsigned long long pmd; } pmd_t; +#define pmd_val(x) ((x).pmd) +#define __pmd(x) ((pmd_t) { (x) } ) + +static inline unsigned long pud_page_vaddr(pud_t pud) +{ + return pud_val(pud); +} + +#define pmd_index(address) (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1)) +static inline pmd_t *pmd_offset(pud_t *pud, unsigned long address) +{ + return (pmd_t *)pud_page_vaddr(*pud) + pmd_index(address); +} + +#define pud_none(x) (!pud_val(x)) +#define pud_present(x) (pud_val(x)) +#define pud_clear(xp) do { set_pud(xp, __pud(0)); } while (0) +#define pud_bad(x) (pud_val(x) & ~PAGE_MASK) + +/* + * (puds are folded into pgds so this doesn't get actually called, + * but the define is needed for a generic inline function.) + */ +#define set_pud(pudptr, pudval) do { *(pudptr) = (pudval); } while(0) + +#endif /* __ASM_SH_PGTABLE_3LEVEL_H */ diff --git a/arch/sh/include/asm/pgtable.h b/arch/sh/include/asm/pgtable.h index 856ece0..aab7652 100644 --- a/arch/sh/include/asm/pgtable.h +++ b/arch/sh/include/asm/pgtable.h @@ -13,9 +13,9 @@ #define __ASM_SH_PGTABLE_H #ifdef CONFIG_X2TLB -#include +#include #else -#include +#include #endif #include diff --git a/arch/sh/include/asm/pgtable_nopmd.h b/arch/sh/include/asm/pgtable_nopmd.h deleted file mode 100644 index b8355e4..0000000 --- a/arch/sh/include/asm/pgtable_nopmd.h +++ /dev/null @@ -1,23 +0,0 @@ -#ifndef __ASM_SH_PGTABLE_NOPMD_H -#define __ASM_SH_PGTABLE_NOPMD_H - -#include - -/* - * traditional two-level paging structure - */ -#define PAGETABLE_LEVELS 2 - -/* PTE bits */ -#define PTE_MAGNITUDE 2 /* 32-bit PTEs */ - -#define PTE_SHIFT PAGE_SHIFT -#define PTE_BITS (PTE_SHIFT - PTE_MAGNITUDE) - -/* PGD bits */ -#define PGDIR_SHIFT (PTE_SHIFT + PTE_BITS) - -#define PTRS_PER_PGD (PAGE_SIZE / (1 << PTE_MAGNITUDE)) -#define USER_PTRS_PER_PGD (TASK_SIZE/PGDIR_SIZE) - -#endif /* __ASM_SH_PGTABLE_NOPMD_H */ diff --git a/arch/sh/include/asm/pgtable_pmd.h b/arch/sh/include/asm/pgtable_pmd.h deleted file mode 100644 index 587b05e..0000000 --- a/arch/sh/include/asm/pgtable_pmd.h +++ /dev/null @@ -1,56 +0,0 @@ -#ifndef __ASM_SH_PGTABLE_PMD_H -#define __ASM_SH_PGTABLE_PMD_H - -#include - -/* - * Some cores need a 3-level page table layout, for example when using - * 64-bit PTEs and 4K pages. - */ -#define PAGETABLE_LEVELS 3 - -#define PTE_MAGNITUDE 3 /* 64-bit PTEs on SH-X2 TLB */ - -/* PGD bits */ -#define PGDIR_SHIFT 30 - -#define PTRS_PER_PGD 4 -#define USER_PTRS_PER_PGD 2 - -/* PMD bits */ -#define PMD_SHIFT (PAGE_SHIFT + (PAGE_SHIFT - PTE_MAGNITUDE)) -#define PMD_SIZE (1UL << PMD_SHIFT) -#define PMD_MASK (~(PMD_SIZE-1)) - -#define PTRS_PER_PMD ((1 << PGDIR_SHIFT) / PMD_SIZE) - -#define pmd_ERROR(e) \ - printk("%s:%d: bad pmd %016llx.\n", __FILE__, __LINE__, pmd_val(e)) - -typedef struct { unsigned long long pmd; } pmd_t; -#define pmd_val(x) ((x).pmd) -#define __pmd(x) ((pmd_t) { (x) } ) - -static inline unsigned long pud_page_vaddr(pud_t pud) -{ - return pud_val(pud); -} - -#define pmd_index(address) (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1)) -static inline pmd_t *pmd_offset(pud_t *pud, unsigned long address) -{ - return (pmd_t *)pud_page_vaddr(*pud) + pmd_index(address); -} - -#define pud_none(x) (!pud_val(x)) -#define pud_present(x) (pud_val(x)) -#define pud_clear(xp) do { set_pud(xp, __pud(0)); } while (0) -#define pud_bad(x) (pud_val(x) & ~PAGE_MASK) - -/* - * (puds are folded into pgds so this doesn't get actually called, - * but the define is needed for a generic inline function.) - */ -#define set_pud(pudptr, pudval) do { *(pudptr) = (pudval); } while(0) - -#endif /* __ASM_SH_PGTABLE_PMD_H */ -- cgit v0.10.2 From 4bb9508bbbb06f10bc3e249dd34375b4a4d6bfc0 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Wed, 23 Dec 2009 14:13:46 +0100 Subject: HID: remove TENX iBuddy from blacklist There were multiple reports which indicate that vendor messed up horribly and the same VID/PID combination is used for completely different devices, some of them requiring the blacklist entry and other not. Remove the blacklist entry for this combination of VID/PID completely, and let the user decide and unbind the driver via sysfs eventually, if needed. Proper fix would be fixing the vendor. References: http://lkml.org/lkml/2009/2/10/434 http://bugzilla.kernel.org/show_bug.cgi?id=13411 Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index 056384c..116a346 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1662,8 +1662,6 @@ static const struct hid_device_id hid_ignore_list[] = { { HID_USB_DEVICE(USB_VENDOR_ID_PANJIT, 0x0004) }, { HID_USB_DEVICE(USB_VENDOR_ID_PHILIPS, USB_DEVICE_ID_PHILIPS_IEEE802154_DONGLE) }, { HID_USB_DEVICE(USB_VENDOR_ID_POWERCOM, USB_DEVICE_ID_POWERCOM_UPS) }, - { HID_USB_DEVICE(USB_VENDOR_ID_TENX, USB_DEVICE_ID_TENX_IBUDDY1) }, - { HID_USB_DEVICE(USB_VENDOR_ID_TENX, USB_DEVICE_ID_TENX_IBUDDY2) }, { HID_USB_DEVICE(USB_VENDOR_ID_VERNIER, USB_DEVICE_ID_VERNIER_LABPRO) }, { HID_USB_DEVICE(USB_VENDOR_ID_VERNIER, USB_DEVICE_ID_VERNIER_GOTEMP) }, { HID_USB_DEVICE(USB_VENDOR_ID_VERNIER, USB_DEVICE_ID_VERNIER_SKIP) }, diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index b009fd6..064c09a 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -406,10 +406,6 @@ #define USB_VENDOR_ID_SUNPLUS 0x04fc #define USB_DEVICE_ID_SUNPLUS_WDESKTOP 0x05d8 -#define USB_VENDOR_ID_TENX 0x1130 -#define USB_DEVICE_ID_TENX_IBUDDY1 0x0001 -#define USB_DEVICE_ID_TENX_IBUDDY2 0x0002 - #define USB_VENDOR_ID_THRUSTMASTER 0x044f #define USB_VENDOR_ID_TOPMAX 0x0663 -- cgit v0.10.2 From b7cece76783c68fb391f9882235b4b0c9c300c46 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 13 Jan 2010 13:22:17 -0200 Subject: perf tools: Encode kernel module mappings in perf.data MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We were always looking at the running machine /proc/modules, even when processing a perf.data file, which only makes sense when we're doing 'perf record' and 'perf report' on the same machine, and in close sucession, or if we don't use modules at all, right Peter? ;-) Now, at 'perf record' time we read /proc/modules, find the long path for modules, and put them as PERF_MMAP events, just like we did to encode the reloc reference symbol for vmlinux. Talking about that now it is encoded in .pgoff, so that we can use .{start,len} to store the address boundaries for the kernel so that when we reconstruct the kmaps tree we can do lookups right away, without having to fixup the end of the kernel maps like we did in the past (and now only in perf record). One more step in the 'perf archive' direction when we'll finally be able to collect data in one machine and analyse in another. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263396139-4798-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 88c570c..4af7199 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -494,6 +494,11 @@ static int __cmd_kmem(void) if (!perf_session__has_traces(session, "kmem record")) goto out_delete; + if (perf_session__create_kernel_maps(session) < 0) { + pr_err("Problems creating kernel maps\n"); + return -1; + } + setup_pager(); err = perf_session__process_events(session, &event_ops); if (err != 0) diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 8f88420..c130df2 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -465,6 +465,11 @@ static int __cmd_record(int argc, const char **argv) return -1; } + if (perf_session__create_kernel_maps(session) < 0) { + pr_err("Problems creating kernel maps\n"); + return -1; + } + if (!file_new) { err = perf_header__read(&session->header, output); if (err < 0) @@ -558,6 +563,12 @@ static int __cmd_record(int argc, const char **argv) return err; } + err = event__synthesize_modules(process_synthesized_event, session); + if (err < 0) { + pr_err("Couldn't record kernel reference relocation symbol.\n"); + return err; + } + if (!system_wide && profile_cpu == -1) event__synthesize_thread(pid, process_synthesized_event, session); diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index ddc584b..6822b44 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -1165,6 +1165,11 @@ static int __cmd_top(void) if (session == NULL) return -ENOMEM; + if (perf_session__create_kernel_maps(session) < 0) { + pr_err("Problems creating kernel maps\n"); + return -1; + } + if (target_pid != -1) event__synthesize_thread(target_pid, event__process, session); else diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index bfb3d87..4f3e7ef 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -154,6 +154,36 @@ static int event__synthesize_mmap_events(pid_t pid, pid_t tgid, return 0; } +int event__synthesize_modules(event__handler_t process, + struct perf_session *session) +{ + struct rb_node *nd; + + for (nd = rb_first(&session->kmaps.maps[MAP__FUNCTION]); + nd; nd = rb_next(nd)) { + event_t ev; + size_t size; + struct map *pos = rb_entry(nd, struct map, rb_node); + + if (pos->dso->kernel) + continue; + + size = ALIGN(pos->dso->long_name_len + 1, sizeof(u64)); + memset(&ev, 0, sizeof(ev)); + ev.mmap.header.type = PERF_RECORD_MMAP; + ev.mmap.header.size = (sizeof(ev.mmap) - + (sizeof(ev.mmap.filename) - size)); + ev.mmap.start = pos->start; + ev.mmap.len = pos->end - pos->start; + + memcpy(ev.mmap.filename, pos->dso->long_name, + pos->dso->long_name_len + 1); + process(&ev, session); + } + + return 0; +} + int event__synthesize_thread(pid_t pid, event__handler_t process, struct perf_session *session) { @@ -222,7 +252,9 @@ int event__synthesize_kernel_mmap(event__handler_t process, "[kernel.kallsyms.%s]", symbol_name) + 1; size = ALIGN(size, sizeof(u64)); ev.mmap.header.size = (sizeof(ev.mmap) - (sizeof(ev.mmap.filename) - size)); - ev.mmap.start = args.start; + ev.mmap.pgoff = args.start; + ev.mmap.start = session->vmlinux_maps[MAP__FUNCTION]->start; + ev.mmap.len = session->vmlinux_maps[MAP__FUNCTION]->end - ev.mmap.start ; return process(&ev, session); } @@ -280,7 +312,6 @@ int event__process_mmap(event_t *self, struct perf_session *session) { struct thread *thread; struct map *map; - static const char kmmap_prefix[] = "[kernel.kallsyms."; dump_printf(" %d/%d: [%p(%p) @ %p]: %s\n", self->mmap.pid, self->mmap.tid, @@ -289,13 +320,61 @@ int event__process_mmap(event_t *self, struct perf_session *session) (void *)(long)self->mmap.pgoff, self->mmap.filename); - if (self->mmap.pid == 0 && - memcmp(self->mmap.filename, kmmap_prefix, - sizeof(kmmap_prefix) - 1) == 0) { - const char *symbol_name = (self->mmap.filename + - sizeof(kmmap_prefix) - 1); - perf_session__set_kallsyms_ref_reloc_sym(session, symbol_name, - self->mmap.start); + if (self->mmap.pid == 0) { + static const char kmmap_prefix[] = "[kernel.kallsyms."; + + if (self->mmap.filename[0] == '/') { + char short_module_name[1024]; + char *name = strrchr(self->mmap.filename, '/'), *dot; + + if (name == NULL) + goto out_problem; + + ++name; /* skip / */ + dot = strrchr(name, '.'); + if (dot == NULL) + goto out_problem; + + snprintf(short_module_name, sizeof(short_module_name), + "[%.*s]", (int)(dot - name), name); + strxfrchar(short_module_name, '-', '_'); + + map = perf_session__new_module_map(session, + self->mmap.start, + short_module_name); + if (map == NULL) + goto out_problem; + + name = strdup(self->mmap.filename); + if (name == NULL) + goto out_problem; + + dso__set_long_name(map->dso, name); + map->end = map->start + self->mmap.len; + } else if (memcmp(self->mmap.filename, kmmap_prefix, + sizeof(kmmap_prefix) - 1) == 0) { + const char *symbol_name = (self->mmap.filename + + sizeof(kmmap_prefix) - 1); + /* + * Should be there already, from the build-id table in + * the header. + */ + struct dso *kernel = __dsos__findnew(&dsos__kernel, + "[kernel.kallsyms]"); + if (kernel == NULL) + goto out_problem; + + if (__map_groups__create_kernel_maps(&session->kmaps, + session->vmlinux_maps, + kernel) < 0) + goto out_problem; + + session->vmlinux_maps[MAP__FUNCTION]->start = self->mmap.start; + session->vmlinux_maps[MAP__FUNCTION]->end = self->mmap.start + self->mmap.len; + + perf_session__set_kallsyms_ref_reloc_sym(session, symbol_name, + self->mmap.pgoff); + } return 0; } @@ -304,10 +383,13 @@ int event__process_mmap(event_t *self, struct perf_session *session) session->cwd, session->cwdlen); if (thread == NULL || map == NULL) - dump_printf("problem processing PERF_RECORD_MMAP, skipping event.\n"); - else - thread__insert_map(thread, map); + goto out_problem; + + thread__insert_map(thread, map); + return 0; +out_problem: + dump_printf("problem processing PERF_RECORD_MMAP, skipping event.\n"); return 0; } diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 80356da..50a7132 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -112,6 +112,8 @@ void event__synthesize_threads(event__handler_t process, int event__synthesize_kernel_mmap(event__handler_t process, struct perf_session *session, const char *symbol_name); +int event__synthesize_modules(event__handler_t process, + struct perf_session *session); int event__process_comm(event_t *self, struct perf_session *session); int event__process_lost(event_t *self, struct perf_session *session); diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 378ac54..fd1c5a3 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -69,9 +69,6 @@ struct perf_session *perf_session__new(const char *filename, int mode, bool forc self->unknown_events = 0; map_groups__init(&self->kmaps); - if (perf_session__create_kernel_maps(self) < 0) - goto out_delete; - if (mode == O_RDONLY && perf_session__open(self, force) < 0) goto out_delete; @@ -268,8 +265,11 @@ int perf_header__read_build_ids(int input, u64 offset, u64 size) head = &dsos__kernel; dso = __dsos__findnew(head, filename); - if (dso != NULL) + if (dso != NULL) { dso__set_build_id(dso, &bev.build_id); + if (head == &dsos__kernel && filename[0] == '[') + dso->kernel = 1; + } offset += bev.header.size; } diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 8e6627e..381999d 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -161,7 +161,7 @@ static size_t symbol__fprintf(struct symbol *self, FILE *fp) self->start, self->end, self->name); } -static void dso__set_long_name(struct dso *self, char *name) +void dso__set_long_name(struct dso *self, char *name) { if (name == NULL) return; @@ -176,7 +176,7 @@ static void dso__set_basename(struct dso *self) struct dso *dso__new(const char *name) { - struct dso *self = malloc(sizeof(*self) + strlen(name) + 1); + struct dso *self = zalloc(sizeof(*self) + strlen(name) + 1); if (self != NULL) { int i; @@ -500,13 +500,17 @@ static int dso__split_kallsyms(struct dso *self, struct map *map, *module++ = '\0'; - if (strcmp(self->name, module)) { + if (strcmp(curr_map->dso->short_name, module)) { curr_map = map_groups__find_by_name(&session->kmaps, map->type, module); if (curr_map == NULL) { pr_debug("/proc/{kallsyms,modules} " - "inconsistency!\n"); + "inconsistency while looking " + "for \"%s\" module!\n", module); return -1; } + + if (curr_map->dso->loaded) + goto discard_symbol; } /* * So that we look just like we get from .ko files, @@ -1343,13 +1347,33 @@ struct map *map_groups__find_by_name(struct map_groups *self, for (nd = rb_first(&self->maps[type]); nd; nd = rb_next(nd)) { struct map *map = rb_entry(nd, struct map, rb_node); - if (map->dso && strcmp(map->dso->name, name) == 0) + if (map->dso && strcmp(map->dso->short_name, name) == 0) return map; } return NULL; } +static int dso__kernel_module_get_build_id(struct dso *self) +{ + char filename[PATH_MAX]; + /* + * kernel module short names are of the form "[module]" and + * we need just "module" here. + */ + const char *name = self->short_name + 1; + + snprintf(filename, sizeof(filename), + "/sys/module/%.*s/notes/.note.gnu.build-id", + (int)strlen(name - 1), name); + + if (sysfs__read_build_id(filename, self->build_id, + sizeof(self->build_id)) == 0) + self->has_build_id = true; + + return 0; +} + static int perf_session__set_modules_path_dir(struct perf_session *self, char *dirname) { struct dirent *dent; @@ -1395,6 +1419,7 @@ static int perf_session__set_modules_path_dir(struct perf_session *self, char *d if (long_name == NULL) goto failure; dso__set_long_name(map->dso, long_name); + dso__kernel_module_get_build_id(map->dso); } } @@ -1437,6 +1462,24 @@ static struct map *map__new2(u64 start, struct dso *dso, enum map_type type) return self; } +struct map *perf_session__new_module_map(struct perf_session *self, u64 start, + const char *filename) +{ + struct map *map; + struct dso *dso = __dsos__findnew(&dsos__kernel, filename); + + if (dso == NULL) + return NULL; + + map = map__new2(start, dso, MAP__FUNCTION); + if (map == NULL) + return NULL; + + dso->origin = DSO__ORIG_KMODULE; + map_groups__insert(&self->kmaps, map); + return map; +} + static int perf_session__create_module_maps(struct perf_session *self) { char *line = NULL; @@ -1450,7 +1493,6 @@ static int perf_session__create_module_maps(struct perf_session *self) while (!feof(file)) { char name[PATH_MAX]; u64 start; - struct dso *dso; char *sep; int line_len; @@ -1476,26 +1518,10 @@ static int perf_session__create_module_maps(struct perf_session *self) *sep = '\0'; snprintf(name, sizeof(name), "[%s]", line); - dso = dso__new(name); - - if (dso == NULL) - goto out_delete_line; - - map = map__new2(start, dso, MAP__FUNCTION); - if (map == NULL) { - dso__delete(dso); + map = perf_session__new_module_map(self, start, name); + if (map == NULL) goto out_delete_line; - } - - snprintf(name, sizeof(name), - "/sys/module/%s/notes/.note.gnu.build-id", line); - if (sysfs__read_build_id(name, dso->build_id, - sizeof(dso->build_id)) == 0) - dso->has_build_id = true; - - dso->origin = DSO__ORIG_KMODULE; - map_groups__insert(&self->kmaps, map); - dsos__add(&dsos__kernel, dso); + dso__kernel_module_get_build_id(map->dso); } free(line); @@ -1573,10 +1599,28 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, } } + /* + * Say the kernel DSO was created when processing the build-id header table, + * we have a build-id, so check if it is the same as the running kernel, + * using it if it is. + */ + if (self->has_build_id) { + u8 kallsyms_build_id[BUILD_ID_SIZE]; + + if (sysfs__read_build_id("/sys/kernel/notes", kallsyms_build_id, + sizeof(kallsyms_build_id)) == 0) + + is_kallsyms = dso__build_id_equal(self, kallsyms_build_id); + if (is_kallsyms) + goto do_kallsyms; + goto do_vmlinux; + } + is_kallsyms = self->long_name[0] == '['; if (is_kallsyms) goto do_kallsyms; +do_vmlinux: err = dso__load_vmlinux(self, map, session, self->long_name, filter); if (err <= 0) { pr_info("The file %s cannot be used, " @@ -1694,16 +1738,12 @@ out_delete_kernel_dso: return NULL; } -static int map_groups__create_kernel_maps(struct map_groups *self, - struct map *vmlinux_maps[MAP__NR_TYPES], - const char *vmlinux) +int __map_groups__create_kernel_maps(struct map_groups *self, + struct map *vmlinux_maps[MAP__NR_TYPES], + struct dso *kernel) { - struct dso *kernel = dsos__create_kernel(vmlinux); enum map_type type; - if (kernel == NULL) - return -1; - for (type = 0; type < MAP__NR_TYPES; ++type) { vmlinux_maps[type] = map__new2(0, kernel, type); if (vmlinux_maps[type] == NULL) @@ -1717,6 +1757,18 @@ static int map_groups__create_kernel_maps(struct map_groups *self, return 0; } +static int map_groups__create_kernel_maps(struct map_groups *self, + struct map *vmlinux_maps[MAP__NR_TYPES], + const char *vmlinux) +{ + struct dso *kernel = dsos__create_kernel(vmlinux); + + if (kernel == NULL) + return -1; + + return __map_groups__create_kernel_maps(self, vmlinux_maps, kernel); +} + static void vmlinux_path__exit(void) { while (--vmlinux_path__nr_entries >= 0) { diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index ee0b459..594156e 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -134,6 +134,7 @@ size_t dsos__fprintf_buildid(FILE *fp); size_t dso__fprintf_buildid(struct dso *self, FILE *fp); size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp); char dso__symtab_origin(const struct dso *self); +void dso__set_long_name(struct dso *self, char *name); void dso__set_build_id(struct dso *self, void *build_id); struct symbol *dso__find_symbol(struct dso *self, enum map_type type, u64 addr); struct symbol *dso__find_symbol_by_name(struct dso *self, enum map_type type, @@ -151,5 +152,7 @@ bool symbol_type__is_a(char symbol_type, enum map_type map_type); int perf_session__create_kernel_maps(struct perf_session *self); +struct map *perf_session__new_module_map(struct perf_session *self, u64 start, + const char *filename); extern struct dso *vdso; #endif /* __PERF_SYMBOL */ diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index c206f72..c06c135 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -67,4 +67,8 @@ map_groups__find_function(struct map_groups *self, struct perf_session *session, struct map *map_groups__find_by_name(struct map_groups *self, enum map_type type, const char *name); + +int __map_groups__create_kernel_maps(struct map_groups *self, + struct map *vmlinux_maps[MAP__NR_TYPES], + struct dso *kernel); #endif /* __PERF_THREAD_H */ -- cgit v0.10.2 From 0895cf0a823e03ea6d79736611e90186006c805e Mon Sep 17 00:00:00 2001 From: Kirill Smelkov Date: Wed, 13 Jan 2010 13:22:18 -0200 Subject: perf: Fix few typos + cosmetics Signed-off-by: Kirill Smelkov Signed-off-by: Arnaldo Carvalho de Melo LKML-Reference: <1263396139-4798-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Documentation/perf.txt b/tools/perf/Documentation/perf.txt index 69c8325..0eeb247 100644 --- a/tools/perf/Documentation/perf.txt +++ b/tools/perf/Documentation/perf.txt @@ -12,7 +12,7 @@ SYNOPSIS DESCRIPTION ----------- -Performance counters for Linux are are a new kernel-based subsystem +Performance counters for Linux are a new kernel-based subsystem that provide a framework for all things performance analysis. It covers hardware level (CPU/PMU, Performance Monitoring Unit) features and software features (software counters, tracepoints) as well. diff --git a/tools/perf/design.txt b/tools/perf/design.txt index 8d0de51..bd0bb1b 100644 --- a/tools/perf/design.txt +++ b/tools/perf/design.txt @@ -101,10 +101,10 @@ enum hw_event_ids { */ PERF_COUNT_HW_CPU_CYCLES = 0, PERF_COUNT_HW_INSTRUCTIONS = 1, - PERF_COUNT_HW_CACHE_REFERENCES = 2, + PERF_COUNT_HW_CACHE_REFERENCES = 2, PERF_COUNT_HW_CACHE_MISSES = 3, PERF_COUNT_HW_BRANCH_INSTRUCTIONS = 4, - PERF_COUNT_HW_BRANCH_MISSES = 5, + PERF_COUNT_HW_BRANCH_MISSES = 5, PERF_COUNT_HW_BUS_CYCLES = 6, }; @@ -131,8 +131,8 @@ software events, selected by 'event_id': */ enum sw_event_ids { PERF_COUNT_SW_CPU_CLOCK = 0, - PERF_COUNT_SW_TASK_CLOCK = 1, - PERF_COUNT_SW_PAGE_FAULTS = 2, + PERF_COUNT_SW_TASK_CLOCK = 1, + PERF_COUNT_SW_PAGE_FAULTS = 2, PERF_COUNT_SW_CONTEXT_SWITCHES = 3, PERF_COUNT_SW_CPU_MIGRATIONS = 4, PERF_COUNT_SW_PAGE_FAULTS_MIN = 5, -- cgit v0.10.2 From 66aeb6d5cb701aedd508187e08612bfd1e108e2e Mon Sep 17 00:00:00 2001 From: Kirill Smelkov Date: Wed, 13 Jan 2010 13:22:19 -0200 Subject: perf top: Fix code typo in prompt_symbol() sym_filter is what was (if ever) passed with -s option. What was typed by user, and what we were looking for, is in buf. Signed-off-by: Kirill Smelkov Signed-off-by: Arnaldo Carvalho de Melo LKML-Reference: <1263396139-4798-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 6822b44..7a8a77e 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -667,7 +667,7 @@ static void prompt_symbol(struct sym_entry **target, const char *msg) } if (!found) { - fprintf(stderr, "Sorry, %s is not active.\n", sym_filter); + fprintf(stderr, "Sorry, %s is not active.\n", buf); sleep(1); return; } else -- cgit v0.10.2 From 3bef444797f7624f8fbd27f4e0334ce96a108725 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Wed, 13 Jan 2010 10:45:55 -0500 Subject: x86: Merge show_regs() Using kernel_stack_pointer() allows 32-bit and 64-bit versions to be merged. This is more correct for 64-bit, since the old %rsp is always saved on the stack. Signed-off-by: Brian Gerst LKML-Reference: <1263397555-27695-1-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index 98c2cde..cf1e04b 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -92,6 +92,13 @@ void exit_thread(void) } } +void show_regs(struct pt_regs *regs) +{ + show_registers(regs); + show_trace(NULL, regs, (unsigned long *)kernel_stack_pointer(regs), + regs->bp); +} + void show_regs_common(void) { const char *board, *product; diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 9c517b5..fe6a34e 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -174,12 +174,6 @@ void __show_regs(struct pt_regs *regs, int all) d6, d7); } -void show_regs(struct pt_regs *regs) -{ - show_registers(regs); - show_trace(NULL, regs, ®s->sp, regs->bp); -} - void release_thread(struct task_struct *dead_task) { BUG_ON(dead_task->mm); diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index 52fbd0c..418f860 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -211,12 +211,6 @@ void __show_regs(struct pt_regs *regs, int all) printk(KERN_INFO "DR3: %016lx DR6: %016lx DR7: %016lx\n", d3, d6, d7); } -void show_regs(struct pt_regs *regs) -{ - show_registers(regs); - show_trace(NULL, regs, (void *)(regs + 1), regs->bp); -} - void release_thread(struct task_struct *dead_task) { if (dead_task->mm) { -- cgit v0.10.2 From 5d0b7235d83eefdafda300656e97d368afcafc9a Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Tue, 12 Jan 2010 17:57:35 -0800 Subject: x86: clean up rwsem type system The fast version of the rwsems (the code that uses xadd) has traditionally only worked on x86-32, and as a result it mixes different kinds of types wildly - they just all happen to be 32-bit. We have "long", we have "__s32", and we have "int". To make it work on x86-64, the types suddenly matter a lot more. It can be either a 32-bit or 64-bit signed type, and both work (with the caveat that a 32-bit counter will only have 15 bits of effective write counters, so it's limited to 32767 users). But whatever type you choose, it needs to be used consistently. This makes a new 'rwsem_counter_t', that is a 32-bit signed type. For a 64-bit type, you'd need to also update the BIAS values. Signed-off-by: Linus Torvalds LKML-Reference: Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/rwsem.h b/arch/x86/include/asm/rwsem.h index 4136200..5f9af30 100644 --- a/arch/x86/include/asm/rwsem.h +++ b/arch/x86/include/asm/rwsem.h @@ -55,6 +55,9 @@ extern asmregparm struct rw_semaphore * /* * the semaphore definition + * + * The bias values and the counter type needs to be extended to 64 bits + * if we want to have more than 32767 potential readers/writers */ #define RWSEM_UNLOCKED_VALUE 0x00000000 @@ -64,8 +67,10 @@ extern asmregparm struct rw_semaphore * #define RWSEM_ACTIVE_READ_BIAS RWSEM_ACTIVE_BIAS #define RWSEM_ACTIVE_WRITE_BIAS (RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS) +typedef signed int rwsem_count_t; + struct rw_semaphore { - signed long count; + rwsem_count_t count; spinlock_t wait_lock; struct list_head wait_list; #ifdef CONFIG_DEBUG_LOCK_ALLOC @@ -121,7 +126,7 @@ static inline void __down_read(struct rw_semaphore *sem) */ static inline int __down_read_trylock(struct rw_semaphore *sem) { - __s32 result, tmp; + rwsem_count_t result, tmp; asm volatile("# beginning __down_read_trylock\n\t" " mov %0,%1\n\t" "1:\n\t" @@ -143,7 +148,7 @@ static inline int __down_read_trylock(struct rw_semaphore *sem) */ static inline void __down_write_nested(struct rw_semaphore *sem, int subclass) { - int tmp; + rwsem_count_t tmp; tmp = RWSEM_ACTIVE_WRITE_BIAS; asm volatile("# beginning down_write\n\t" @@ -170,9 +175,9 @@ static inline void __down_write(struct rw_semaphore *sem) */ static inline int __down_write_trylock(struct rw_semaphore *sem) { - signed long ret = cmpxchg(&sem->count, - RWSEM_UNLOCKED_VALUE, - RWSEM_ACTIVE_WRITE_BIAS); + rwsem_count_t ret = cmpxchg(&sem->count, + RWSEM_UNLOCKED_VALUE, + RWSEM_ACTIVE_WRITE_BIAS); if (ret == RWSEM_UNLOCKED_VALUE) return 1; return 0; @@ -183,7 +188,7 @@ static inline int __down_write_trylock(struct rw_semaphore *sem) */ static inline void __up_read(struct rw_semaphore *sem) { - __s32 tmp = -RWSEM_ACTIVE_READ_BIAS; + rwsem_count_t tmp = -RWSEM_ACTIVE_READ_BIAS; asm volatile("# beginning __up_read\n\t" LOCK_PREFIX " xadd %1,(%2)\n\t" /* subtracts 1, returns the old value */ @@ -201,7 +206,7 @@ static inline void __up_read(struct rw_semaphore *sem) */ static inline void __up_write(struct rw_semaphore *sem) { - unsigned long tmp; + rwsem_count_t tmp; asm volatile("# beginning __up_write\n\t" LOCK_PREFIX " xadd %1,(%2)\n\t" /* tries to transition @@ -245,9 +250,9 @@ static inline void rwsem_atomic_add(int delta, struct rw_semaphore *sem) /* * implement exchange and add functionality */ -static inline int rwsem_atomic_update(int delta, struct rw_semaphore *sem) +static inline rwsem_count_t rwsem_atomic_update(int delta, struct rw_semaphore *sem) { - int tmp = delta; + rwsem_count_t tmp = delta; asm volatile(LOCK_PREFIX "xadd %0,%1" : "+r" (tmp), "+m" (sem->count) -- cgit v0.10.2 From bafaecd11df15ad5b1e598adc7736afcd38ee13d Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Tue, 12 Jan 2010 18:16:42 -0800 Subject: x86-64: support native xadd rwsem implementation This one is much faster than the spinlock based fallback rwsem code, with certain artifical benchmarks having shown 300%+ improvement on threaded page faults etc. Again, note the 32767-thread limit here. So this really does need that whole "make rwsem_count_t be 64-bit and fix the BIAS values to match" extension on top of it, but that is conceptually a totally independent issue. NOT TESTED! The original patch that this all was based on were tested by KAMEZAWA Hiroyuki, but maybe I screwed up something when I created the cleaned-up series, so caveat emptor.. Also note that it _may_ be a good idea to mark some more registers clobbered on x86-64 in the inline asms instead of saving/restoring them. They are inline functions, but they are only used in places where there are not a lot of live registers _anyway_, so doing for example the clobbers of %r8-%r11 in the asm wouldn't make the fast-path code any worse, and would make the slow-path code smaller. (Not that the slow-path really matters to that degree. Saving a few unnecessary registers is the _least_ of our problems when we hit the slow path. The instruction/cycle counting really only matters in the fast path). Signed-off-by: Linus Torvalds LKML-Reference: Signed-off-by: H. Peter Anvin diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu index 08e442b..9d38a13 100644 --- a/arch/x86/Kconfig.cpu +++ b/arch/x86/Kconfig.cpu @@ -319,7 +319,7 @@ config X86_L1_CACHE_SHIFT config X86_XADD def_bool y - depends on X86_32 && !M386 + depends on X86_64 || !M386 config X86_PPRO_FENCE bool "PentiumPro memory ordering errata workaround" diff --git a/arch/x86/lib/Makefile b/arch/x86/lib/Makefile index cffd754f..c802451 100644 --- a/arch/x86/lib/Makefile +++ b/arch/x86/lib/Makefile @@ -39,4 +39,5 @@ else lib-y += thunk_64.o clear_page_64.o copy_page_64.o lib-y += memmove_64.o memset_64.o lib-y += copy_user_64.o rwlock_64.o copy_user_nocache_64.o + lib-$(CONFIG_RWSEM_XCHGADD_ALGORITHM) += rwsem_64.o endif diff --git a/arch/x86/lib/rwsem_64.S b/arch/x86/lib/rwsem_64.S new file mode 100644 index 0000000..15acecf --- /dev/null +++ b/arch/x86/lib/rwsem_64.S @@ -0,0 +1,81 @@ +/* + * x86-64 rwsem wrappers + * + * This interfaces the inline asm code to the slow-path + * C routines. We need to save the call-clobbered regs + * that the asm does not mark as clobbered, and move the + * argument from %rax to %rdi. + * + * NOTE! We don't need to save %rax, because the functions + * will always return the semaphore pointer in %rax (which + * is also the input argument to these helpers) + * + * The following can clobber %rdx because the asm clobbers it: + * call_rwsem_down_write_failed + * call_rwsem_wake + * but %rdi, %rsi, %rcx, %r8-r11 always need saving. + */ + +#include +#include +#include +#include +#include + +#define save_common_regs \ + pushq %rdi; \ + pushq %rsi; \ + pushq %rcx; \ + pushq %r8; \ + pushq %r9; \ + pushq %r10; \ + pushq %r11 + +#define restore_common_regs \ + popq %r11; \ + popq %r10; \ + popq %r9; \ + popq %r8; \ + popq %rcx; \ + popq %rsi; \ + popq %rdi + +/* Fix up special calling conventions */ +ENTRY(call_rwsem_down_read_failed) + save_common_regs + pushq %rdx + movq %rax,%rdi + call rwsem_down_read_failed + popq %rdx + restore_common_regs + ret + ENDPROC(call_rwsem_down_read_failed) + +ENTRY(call_rwsem_down_write_failed) + save_common_regs + movq %rax,%rdi + call rwsem_down_write_failed + restore_common_regs + ret + ENDPROC(call_rwsem_down_write_failed) + +ENTRY(call_rwsem_wake) + decw %dx /* do nothing if still outstanding active readers */ + jnz 1f + save_common_regs + movq %rax,%rdi + call rwsem_wake + restore_common_regs +1: ret + ENDPROC(call_rwsem_wake) + +/* Fix up special calling conventions */ +ENTRY(call_rwsem_downgrade_wake) + save_common_regs + pushq %rdx + movq %rax,%rdi + call rwsem_downgrade_wake + popq %rdx + restore_common_regs + ret + ENDPROC(call_rwsem_downgrade_wake) -- cgit v0.10.2 From 02bf6cc72cc2a6258411ddf1649f33a65fc9a06e Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 14 Jan 2010 20:58:58 +0900 Subject: sh: Preliminary SDK7786 board support. This stubs in some preliminary board support for the RTE SDK7786. This is quite stunted at the moment, and primarily builds on top of the system FPGA. FPGA IRQs are handled via CPU IRL masking for simplicity, with initial peripheral support restricted to the debug ethernet. Signed-off-by: Matt Fleming Signed-off-by: Paul Mundt diff --git a/arch/sh/Makefile b/arch/sh/Makefile index f7d5f42..30c80fd 100644 --- a/arch/sh/Makefile +++ b/arch/sh/Makefile @@ -83,6 +83,7 @@ defaultimage-$(CONFIG_SH_AP325RXA) := uImage defaultimage-$(CONFIG_SH_7724_SOLUTION_ENGINE) := uImage defaultimage-$(CONFIG_SH_7206_SOLUTION_ENGINE) := vmlinux defaultimage-$(CONFIG_SH_7619_SOLUTION_ENGINE) := vmlinux +defaultimage-$(CONFIG_SH_SDK7786) := vmlinux.bin # Set some sensible Kbuild defaults KBUILD_IMAGE := $(defaultimage-y) @@ -143,6 +144,7 @@ machdir-$(CONFIG_SH_AP325RXA) += mach-ap325rxa machdir-$(CONFIG_SH_KFR2R09) += mach-kfr2r09 machdir-$(CONFIG_SH_ECOVEC) += mach-ecovec24 machdir-$(CONFIG_SH_SDK7780) += mach-sdk7780 +machdir-$(CONFIG_SH_SDK7786) += mach-sdk7786 machdir-$(CONFIG_SH_X3PROTO) += mach-x3proto machdir-$(CONFIG_SH_SH7763RDP) += mach-sh7763rdp machdir-$(CONFIG_SH_SH4202_MICRODEV) += mach-microdev diff --git a/arch/sh/boards/Kconfig b/arch/sh/boards/Kconfig index aedd9de..c0a1992 100644 --- a/arch/sh/boards/Kconfig +++ b/arch/sh/boards/Kconfig @@ -150,6 +150,13 @@ config SH_SDK7780 Select SDK7780 if configuring for a Renesas SH7780 SDK7780R3 evaluation board. +config SH_SDK7786 + bool "SDK7786" + depends on CPU_SUBTYPE_SH7786 + help + Select SDK7786 if configuring for a Renesas Technology Europe + SH7786-65nm board. + config SH_HIGHLANDER bool "Highlander" depends on CPU_SUBTYPE_SH7780 || CPU_SUBTYPE_SH7785 diff --git a/arch/sh/boards/mach-sdk7786/Makefile b/arch/sh/boards/mach-sdk7786/Makefile new file mode 100644 index 0000000..f663768 --- /dev/null +++ b/arch/sh/boards/mach-sdk7786/Makefile @@ -0,0 +1 @@ +obj-y := setup.o diff --git a/arch/sh/boards/mach-sdk7786/setup.c b/arch/sh/boards/mach-sdk7786/setup.c new file mode 100644 index 0000000..b460dcc --- /dev/null +++ b/arch/sh/boards/mach-sdk7786/setup.c @@ -0,0 +1,219 @@ +/* + * Renesas Technology Europe SDK7786 Support. + * + * Copyright (C) 2010 Matt Fleming + * Copyright (C) 2010 Paul Mundt + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ +#include +#include +#include +#include +#include +#include +#include +#include + +static struct resource smsc911x_resources[] = { + [0] = { + .name = "smsc911x-memory", + .start = 0x07ffff00, + .end = 0x07ffff00 + SZ_256 - 1, + .flags = IORESOURCE_MEM, + }, + [1] = { + .name = "smsc911x-irq", + .start = evt2irq(0x2c0), + .end = evt2irq(0x2c0), + .flags = IORESOURCE_IRQ, + }, +}; + +static struct smsc911x_platform_config smsc911x_config = { + .irq_polarity = SMSC911X_IRQ_POLARITY_ACTIVE_LOW, + .irq_type = SMSC911X_IRQ_TYPE_OPEN_DRAIN, + .flags = SMSC911X_USE_32BIT, + .phy_interface = PHY_INTERFACE_MODE_MII, +}; + +static struct platform_device smsc911x_device = { + .name = "smsc911x", + .id = -1, + .num_resources = ARRAY_SIZE(smsc911x_resources), + .resource = smsc911x_resources, + .dev = { + .platform_data = &smsc911x_config, + }, +}; + +static struct resource smbus_fpga_resource = { + .start = 0x07fff9e0, + .end = 0x07fff9e0 + SZ_32 - 1, + .flags = IORESOURCE_MEM, +}; + +static struct platform_device smbus_fpga_device = { + .name = "i2c-sdk7786", + .id = 0, + .num_resources = 1, + .resource = &smbus_fpga_resource, +}; + +static struct resource smbus_pcie_resource = { + .start = 0x07fffc30, + .end = 0x07fffc30 + SZ_32 - 1, + .flags = IORESOURCE_MEM, +}; + +static struct platform_device smbus_pcie_device = { + .name = "i2c-sdk7786", + .id = 1, + .num_resources = 1, + .resource = &smbus_pcie_resource, +}; + +static struct i2c_board_info __initdata sdk7786_i2c_devices[] = { + { + I2C_BOARD_INFO("max6900", 0x68), + }, +}; + +static struct platform_device *sh7786_devices[] __initdata = { + &smsc911x_device, + &smbus_fpga_device, + &smbus_pcie_device, +}; + +#define SBCR_REGS_BASE 0x07fff990 + +#define SCBR_I2CMEN (1 << 0) /* FPGA I2C master enable */ +#define SCBR_I2CCEN (1 << 1) /* CPU I2C master enable */ + +static int sdk7786_i2c_setup(void) +{ + void __iomem *sbcr; + unsigned int tmp; + + sbcr = ioremap_nocache(SBCR_REGS_BASE, SZ_16); + + /* + * Hand over I2C control to the FPGA. + */ + tmp = ioread16(sbcr); + tmp &= ~SCBR_I2CCEN; + tmp |= SCBR_I2CMEN; + iowrite16(tmp, sbcr); + + iounmap(sbcr); + + return i2c_register_board_info(0, sdk7786_i2c_devices, + ARRAY_SIZE(sdk7786_i2c_devices)); +} + +static int __init sdk7786_devices_setup(void) +{ + int ret; + + ret = platform_add_devices(sh7786_devices, ARRAY_SIZE(sh7786_devices)); + if (unlikely(ret != 0)) + return ret; + + return sdk7786_i2c_setup(); +} +__initcall(sdk7786_devices_setup); + +#define FPGA_REGS_BASE 0x07fff800 +#define FPGA_REGS_SIZE 1152 + +#define INTASR 0x010 +#define INTAMR 0x020 +#define INTBSR 0x090 +#define INTBMR 0x0a0 +#define INTMSR 0x130 + +#define IASELR1 0x210 +#define IASELR2 0x220 +#define IASELR3 0x230 +#define IASELR4 0x240 +#define IASELR5 0x250 +#define IASELR6 0x260 +#define IASELR7 0x270 +#define IASELR8 0x280 +#define IASELR9 0x290 +#define IASELR10 0x2a0 +#define IASELR11 0x2b0 +#define IASELR12 0x2c0 +#define IASELR13 0x2d0 +#define IASELR14 0x2e0 +#define IASELR15 0x2f0 + +static void __iomem *fpga_regs; + +static u16 fpga_read_reg(unsigned int reg) +{ + return __raw_readw(fpga_regs + reg); +} + +static void fpga_write_reg(u16 val, unsigned int reg) +{ + __raw_writew(val, fpga_regs + reg); +} + +enum { + ATA_IRQ_BIT = 1, + SPI_BUSY_BIT = 2, + LIRQ5_BIT = 3, + LIRQ6_BIT = 4, + LIRQ7_BIT = 5, + LIRQ8_BIT = 6, + KEY_IRQ_BIT = 7, + PEN_IRQ_BIT = 8, + ETH_IRQ_BIT = 9, + RTC_ALARM_BIT = 10, + CRYSTAL_FAIL_BIT = 12, + ETH_PME_BIT = 14, +}; + +static void __init init_sdk7786_IRQ(void) +{ + unsigned int tmp; + + fpga_regs = ioremap_nocache(FPGA_REGS_BASE, FPGA_REGS_SIZE); + if (!fpga_regs) { + printk(KERN_ERR "Couldn't map FPGA registers\n"); + return; + } + + /* Enable priority encoding for all IRLs */ + fpga_write_reg(fpga_read_reg(INTMSR) | 0x0303, INTMSR); + + /* Clear FPGA interrupt status registers */ + fpga_write_reg(0x0000, INTASR); + fpga_write_reg(0x0000, INTBSR); + + /* Unmask FPGA interrupts */ + tmp = fpga_read_reg(INTAMR); + tmp &= ~(1 << ETH_IRQ_BIT); + fpga_write_reg(tmp, INTAMR); + + plat_irq_setup_pins(IRQ_MODE_IRL7654_MASK); + plat_irq_setup_pins(IRQ_MODE_IRL3210_MASK); +} + +/* Initialize the board */ +static void __init sdk7786_setup(char **cmdline_p) +{ + printk(KERN_INFO "Renesas Technology Corp. SDK7786 support.\n"); +} + +/* + * The Machine Vector + */ +static struct sh_machine_vector mv_sdk7786 __initmv = { + .mv_name = "SDK7786", + .mv_setup = sdk7786_setup, + .mv_init_irq = init_sdk7786_IRQ, +}; diff --git a/arch/sh/configs/sdk7786_defconfig b/arch/sh/configs/sdk7786_defconfig new file mode 100644 index 0000000..1721d2c --- /dev/null +++ b/arch/sh/configs/sdk7786_defconfig @@ -0,0 +1,1440 @@ +# +# Automatically generated make config: don't edit +# Linux kernel version: 2.6.33-rc3 +# Thu Jan 14 20:37:44 2010 +# +CONFIG_SUPERH=y +CONFIG_SUPERH32=y +# CONFIG_SUPERH64 is not set +CONFIG_ARCH_DEFCONFIG="arch/sh/configs/shx3_defconfig" +CONFIG_RWSEM_GENERIC_SPINLOCK=y +CONFIG_GENERIC_BUG=y +CONFIG_GENERIC_FIND_NEXT_BIT=y +CONFIG_GENERIC_HWEIGHT=y +CONFIG_GENERIC_HARDIRQS=y +CONFIG_GENERIC_HARDIRQS_NO__DO_IRQ=y +CONFIG_GENERIC_IRQ_PROBE=y +CONFIG_IRQ_PER_CPU=y +# CONFIG_GENERIC_GPIO is not set +CONFIG_GENERIC_TIME=y +CONFIG_GENERIC_CLOCKEVENTS=y +# CONFIG_ARCH_SUSPEND_POSSIBLE is not set +CONFIG_ARCH_HIBERNATION_POSSIBLE=y +CONFIG_SYS_SUPPORTS_HUGETLBFS=y +CONFIG_SYS_SUPPORTS_SMP=y +CONFIG_SYS_SUPPORTS_NUMA=y +CONFIG_SYS_SUPPORTS_TMU=y +CONFIG_STACKTRACE_SUPPORT=y +CONFIG_LOCKDEP_SUPPORT=y +CONFIG_HAVE_LATENCYTOP_SUPPORT=y +# CONFIG_ARCH_HAS_ILOG2_U32 is not set +# CONFIG_ARCH_HAS_ILOG2_U64 is not set +CONFIG_ARCH_NO_VIRT_TO_BUS=y +CONFIG_ARCH_HAS_DEFAULT_IDLE=y +CONFIG_ARCH_HAS_CPU_IDLE_WAIT=y +CONFIG_DMA_COHERENT=y +# CONFIG_DMA_NONCOHERENT is not set +CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config" +CONFIG_CONSTRUCTORS=y + +# +# General setup +# +CONFIG_EXPERIMENTAL=y +CONFIG_BROKEN_ON_SMP=y +CONFIG_LOCK_KERNEL=y +CONFIG_INIT_ENV_ARG_LIMIT=32 +CONFIG_LOCALVERSION="" +CONFIG_LOCALVERSION_AUTO=y +CONFIG_HAVE_KERNEL_GZIP=y +CONFIG_HAVE_KERNEL_BZIP2=y +CONFIG_HAVE_KERNEL_LZMA=y +CONFIG_HAVE_KERNEL_LZO=y +CONFIG_KERNEL_GZIP=y +# CONFIG_KERNEL_BZIP2 is not set +# CONFIG_KERNEL_LZMA is not set +# CONFIG_KERNEL_LZO is not set +CONFIG_SWAP=y +CONFIG_SYSVIPC=y +CONFIG_SYSVIPC_SYSCTL=y +CONFIG_POSIX_MQUEUE=y +CONFIG_POSIX_MQUEUE_SYSCTL=y +CONFIG_BSD_PROCESS_ACCT=y +# CONFIG_BSD_PROCESS_ACCT_V3 is not set +# CONFIG_TASKSTATS is not set +# CONFIG_AUDIT is not set + +# +# RCU Subsystem +# +CONFIG_TREE_RCU=y +# CONFIG_TREE_PREEMPT_RCU is not set +# CONFIG_TINY_RCU is not set +CONFIG_RCU_TRACE=y +CONFIG_RCU_FANOUT=32 +# CONFIG_RCU_FANOUT_EXACT is not set +CONFIG_TREE_RCU_TRACE=y +CONFIG_IKCONFIG=y +CONFIG_IKCONFIG_PROC=y +CONFIG_LOG_BUF_SHIFT=14 +CONFIG_GROUP_SCHED=y +CONFIG_FAIR_GROUP_SCHED=y +CONFIG_RT_GROUP_SCHED=y +CONFIG_USER_SCHED=y +# CONFIG_CGROUP_SCHED is not set +CONFIG_CGROUPS=y +# CONFIG_CGROUP_DEBUG is not set +CONFIG_CGROUP_NS=y +CONFIG_CGROUP_FREEZER=y +CONFIG_CGROUP_DEVICE=y +# CONFIG_CPUSETS is not set +CONFIG_CGROUP_CPUACCT=y +CONFIG_RESOURCE_COUNTERS=y +CONFIG_CGROUP_MEM_RES_CTLR=y +# CONFIG_CGROUP_MEM_RES_CTLR_SWAP is not set +CONFIG_MM_OWNER=y +# CONFIG_SYSFS_DEPRECATED_V2 is not set +# CONFIG_RELAY is not set +CONFIG_NAMESPACES=y +CONFIG_UTS_NS=y +CONFIG_IPC_NS=y +CONFIG_USER_NS=y +CONFIG_PID_NS=y +CONFIG_NET_NS=y +# CONFIG_BLK_DEV_INITRD is not set +CONFIG_CC_OPTIMIZE_FOR_SIZE=y +CONFIG_SYSCTL=y +CONFIG_ANON_INODES=y +CONFIG_EMBEDDED=y +CONFIG_UID16=y +CONFIG_SYSCTL_SYSCALL=y +CONFIG_KALLSYMS=y +CONFIG_KALLSYMS_ALL=y +# CONFIG_KALLSYMS_EXTRA_PASS is not set +CONFIG_HOTPLUG=y +CONFIG_PRINTK=y +CONFIG_BUG=y +CONFIG_ELF_CORE=y +CONFIG_BASE_FULL=y +CONFIG_FUTEX=y +CONFIG_EPOLL=y +CONFIG_SIGNALFD=y +CONFIG_TIMERFD=y +CONFIG_EVENTFD=y +CONFIG_SHMEM=y +CONFIG_AIO=y +CONFIG_HAVE_PERF_EVENTS=y +CONFIG_PERF_USE_VMALLOC=y + +# +# Kernel Performance Events And Counters +# +CONFIG_PERF_EVENTS=y +CONFIG_EVENT_PROFILE=y +# CONFIG_PERF_COUNTERS is not set +# CONFIG_DEBUG_PERF_USE_VMALLOC is not set +CONFIG_VM_EVENT_COUNTERS=y +# CONFIG_COMPAT_BRK is not set +CONFIG_SLAB=y +# CONFIG_SLUB is not set +# CONFIG_SLOB is not set +CONFIG_PROFILING=y +CONFIG_TRACEPOINTS=y +# CONFIG_OPROFILE is not set +CONFIG_HAVE_OPROFILE=y +# CONFIG_KPROBES is not set +CONFIG_HAVE_IOREMAP_PROT=y +CONFIG_HAVE_KPROBES=y +CONFIG_HAVE_KRETPROBES=y +CONFIG_HAVE_ARCH_TRACEHOOK=y +CONFIG_HAVE_DMA_ATTRS=y +CONFIG_HAVE_CLK=y +CONFIG_HAVE_DMA_API_DEBUG=y +CONFIG_HAVE_HW_BREAKPOINT=y + +# +# GCOV-based kernel profiling +# +# CONFIG_GCOV_KERNEL is not set +# CONFIG_SLOW_WORK is not set +CONFIG_HAVE_GENERIC_DMA_COHERENT=y +CONFIG_SLABINFO=y +CONFIG_RT_MUTEXES=y +CONFIG_BASE_SMALL=0 +CONFIG_MODULES=y +# CONFIG_MODULE_FORCE_LOAD is not set +CONFIG_MODULE_UNLOAD=y +# CONFIG_MODULE_FORCE_UNLOAD is not set +# CONFIG_MODVERSIONS is not set +# CONFIG_MODULE_SRCVERSION_ALL is not set +CONFIG_BLOCK=y +# CONFIG_LBDAF is not set +# CONFIG_BLK_DEV_BSG is not set +# CONFIG_BLK_DEV_INTEGRITY is not set +CONFIG_BLK_CGROUP=y +# CONFIG_DEBUG_BLK_CGROUP is not set + +# +# IO Schedulers +# +CONFIG_IOSCHED_NOOP=y +CONFIG_IOSCHED_DEADLINE=y +CONFIG_IOSCHED_CFQ=y +CONFIG_CFQ_GROUP_IOSCHED=y +# CONFIG_DEBUG_CFQ_IOSCHED is not set +# CONFIG_DEFAULT_DEADLINE is not set +CONFIG_DEFAULT_CFQ=y +# CONFIG_DEFAULT_NOOP is not set +CONFIG_DEFAULT_IOSCHED="cfq" +# CONFIG_INLINE_SPIN_TRYLOCK is not set +# CONFIG_INLINE_SPIN_TRYLOCK_BH is not set +# CONFIG_INLINE_SPIN_LOCK is not set +# CONFIG_INLINE_SPIN_LOCK_BH is not set +# CONFIG_INLINE_SPIN_LOCK_IRQ is not set +# CONFIG_INLINE_SPIN_LOCK_IRQSAVE is not set +# CONFIG_INLINE_SPIN_UNLOCK is not set +# CONFIG_INLINE_SPIN_UNLOCK_BH is not set +# CONFIG_INLINE_SPIN_UNLOCK_IRQ is not set +# CONFIG_INLINE_SPIN_UNLOCK_IRQRESTORE is not set +# CONFIG_INLINE_READ_TRYLOCK is not set +# CONFIG_INLINE_READ_LOCK is not set +# CONFIG_INLINE_READ_LOCK_BH is not set +# CONFIG_INLINE_READ_LOCK_IRQ is not set +# CONFIG_INLINE_READ_LOCK_IRQSAVE is not set +# CONFIG_INLINE_READ_UNLOCK is not set +# CONFIG_INLINE_READ_UNLOCK_BH is not set +# CONFIG_INLINE_READ_UNLOCK_IRQ is not set +# CONFIG_INLINE_READ_UNLOCK_IRQRESTORE is not set +# CONFIG_INLINE_WRITE_TRYLOCK is not set +# CONFIG_INLINE_WRITE_LOCK is not set +# CONFIG_INLINE_WRITE_LOCK_BH is not set +# CONFIG_INLINE_WRITE_LOCK_IRQ is not set +# CONFIG_INLINE_WRITE_LOCK_IRQSAVE is not set +# CONFIG_INLINE_WRITE_UNLOCK is not set +# CONFIG_INLINE_WRITE_UNLOCK_BH is not set +# CONFIG_INLINE_WRITE_UNLOCK_IRQ is not set +# CONFIG_INLINE_WRITE_UNLOCK_IRQRESTORE is not set +# CONFIG_MUTEX_SPIN_ON_OWNER is not set +CONFIG_FREEZER=y + +# +# System type +# +CONFIG_CPU_SH4=y +CONFIG_CPU_SH4A=y +CONFIG_CPU_SHX3=y +# CONFIG_CPU_SUBTYPE_SH7619 is not set +# CONFIG_CPU_SUBTYPE_SH7201 is not set +# CONFIG_CPU_SUBTYPE_SH7203 is not set +# CONFIG_CPU_SUBTYPE_SH7206 is not set +# CONFIG_CPU_SUBTYPE_SH7263 is not set +# CONFIG_CPU_SUBTYPE_MXG is not set +# CONFIG_CPU_SUBTYPE_SH7705 is not set +# CONFIG_CPU_SUBTYPE_SH7706 is not set +# CONFIG_CPU_SUBTYPE_SH7707 is not set +# CONFIG_CPU_SUBTYPE_SH7708 is not set +# CONFIG_CPU_SUBTYPE_SH7709 is not set +# CONFIG_CPU_SUBTYPE_SH7710 is not set +# CONFIG_CPU_SUBTYPE_SH7712 is not set +# CONFIG_CPU_SUBTYPE_SH7720 is not set +# CONFIG_CPU_SUBTYPE_SH7721 is not set +# CONFIG_CPU_SUBTYPE_SH7750 is not set +# CONFIG_CPU_SUBTYPE_SH7091 is not set +# CONFIG_CPU_SUBTYPE_SH7750R is not set +# CONFIG_CPU_SUBTYPE_SH7750S is not set +# CONFIG_CPU_SUBTYPE_SH7751 is not set +# CONFIG_CPU_SUBTYPE_SH7751R is not set +# CONFIG_CPU_SUBTYPE_SH7760 is not set +# CONFIG_CPU_SUBTYPE_SH4_202 is not set +# CONFIG_CPU_SUBTYPE_SH7723 is not set +# CONFIG_CPU_SUBTYPE_SH7724 is not set +# CONFIG_CPU_SUBTYPE_SH7757 is not set +# CONFIG_CPU_SUBTYPE_SH7763 is not set +# CONFIG_CPU_SUBTYPE_SH7770 is not set +# CONFIG_CPU_SUBTYPE_SH7780 is not set +# CONFIG_CPU_SUBTYPE_SH7785 is not set +CONFIG_CPU_SUBTYPE_SH7786=y +# CONFIG_CPU_SUBTYPE_SHX3 is not set +# CONFIG_CPU_SUBTYPE_SH7343 is not set +# CONFIG_CPU_SUBTYPE_SH7722 is not set +# CONFIG_CPU_SUBTYPE_SH7366 is not set + +# +# Memory management options +# +CONFIG_QUICKLIST=y +CONFIG_MMU=y +CONFIG_PAGE_OFFSET=0x80000000 +CONFIG_FORCE_MAX_ZONEORDER=11 +CONFIG_MEMORY_START=0x60000000 +CONFIG_MEMORY_SIZE=0x08000000 +# CONFIG_29BIT is not set +CONFIG_32BIT=y +CONFIG_PMB=y +# CONFIG_PMB_LEGACY is not set +CONFIG_X2TLB=y +CONFIG_VSYSCALL=y +# CONFIG_NUMA is not set +CONFIG_ARCH_FLATMEM_ENABLE=y +CONFIG_ARCH_SPARSEMEM_ENABLE=y +CONFIG_ARCH_SPARSEMEM_DEFAULT=y +CONFIG_MAX_ACTIVE_REGIONS=1 +CONFIG_ARCH_POPULATES_NODE_MAP=y +CONFIG_ARCH_SELECT_MEMORY_MODEL=y +CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y +CONFIG_ARCH_ENABLE_MEMORY_HOTREMOVE=y +CONFIG_ARCH_MEMORY_PROBE=y +CONFIG_PAGE_SIZE_4KB=y +# CONFIG_PAGE_SIZE_8KB is not set +# CONFIG_PAGE_SIZE_16KB is not set +# CONFIG_PAGE_SIZE_64KB is not set +# CONFIG_HUGETLB_PAGE_SIZE_64K is not set +# CONFIG_HUGETLB_PAGE_SIZE_256K is not set +CONFIG_HUGETLB_PAGE_SIZE_1MB=y +# CONFIG_HUGETLB_PAGE_SIZE_4MB is not set +# CONFIG_HUGETLB_PAGE_SIZE_64MB is not set +# CONFIG_HUGETLB_PAGE_SIZE_512MB is not set +CONFIG_SELECT_MEMORY_MODEL=y +# CONFIG_FLATMEM_MANUAL is not set +# CONFIG_DISCONTIGMEM_MANUAL is not set +CONFIG_SPARSEMEM_MANUAL=y +CONFIG_SPARSEMEM=y +CONFIG_HAVE_MEMORY_PRESENT=y +CONFIG_SPARSEMEM_STATIC=y +CONFIG_MEMORY_HOTPLUG=y +CONFIG_MEMORY_HOTPLUG_SPARSE=y +CONFIG_MEMORY_HOTREMOVE=y +CONFIG_SPLIT_PTLOCK_CPUS=4 +CONFIG_MIGRATION=y +# CONFIG_PHYS_ADDR_T_64BIT is not set +CONFIG_ZONE_DMA_FLAG=0 +CONFIG_NR_QUICK=1 +CONFIG_KSM=y +CONFIG_DEFAULT_MMAP_MIN_ADDR=4096 + +# +# Cache configuration +# +CONFIG_CACHE_WRITEBACK=y +# CONFIG_CACHE_WRITETHROUGH is not set +# CONFIG_CACHE_OFF is not set + +# +# Processor features +# +CONFIG_CPU_LITTLE_ENDIAN=y +# CONFIG_CPU_BIG_ENDIAN is not set +CONFIG_SH_FPU=y +CONFIG_SH_STORE_QUEUES=y +CONFIG_CPU_HAS_INTEVT=y +CONFIG_CPU_HAS_SR_RB=y +CONFIG_CPU_HAS_PTEAEX=y +CONFIG_CPU_HAS_FPU=y + +# +# Board support +# +CONFIG_SH_SDK7786=y +# CONFIG_SH_URQUELL is not set + +# +# Timer and clock configuration +# +CONFIG_SH_TIMER_TMU=y +CONFIG_SH_PCLK_FREQ=50000000 +CONFIG_SH_CLK_CPG=y +CONFIG_SH_CLK_CPG_LEGACY=y +CONFIG_TICK_ONESHOT=y +CONFIG_NO_HZ=y +CONFIG_HIGH_RES_TIMERS=y +CONFIG_GENERIC_CLOCKEVENTS_BUILD=y + +# +# CPU Frequency scaling +# +CONFIG_CPU_FREQ=y +CONFIG_CPU_FREQ_TABLE=y +# CONFIG_CPU_FREQ_DEBUG is not set +CONFIG_CPU_FREQ_STAT=y +# CONFIG_CPU_FREQ_STAT_DETAILS is not set +CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE=y +# CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE is not set +# CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE is not set +# CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND is not set +# CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE is not set +CONFIG_CPU_FREQ_GOV_PERFORMANCE=y +CONFIG_CPU_FREQ_GOV_POWERSAVE=m +CONFIG_CPU_FREQ_GOV_USERSPACE=m +CONFIG_CPU_FREQ_GOV_ONDEMAND=m +CONFIG_CPU_FREQ_GOV_CONSERVATIVE=m +CONFIG_SH_CPU_FREQ=y + +# +# DMA support +# +# CONFIG_SH_DMA is not set + +# +# Companion Chips +# + +# +# Additional SuperH Device Drivers +# +CONFIG_HEARTBEAT=y +# CONFIG_PUSH_SWITCH is not set + +# +# Kernel features +# +# CONFIG_HZ_100 is not set +CONFIG_HZ_250=y +# CONFIG_HZ_300 is not set +# CONFIG_HZ_1000 is not set +CONFIG_HZ=250 +CONFIG_SCHED_HRTICK=y +CONFIG_KEXEC=y +# CONFIG_CRASH_DUMP is not set +CONFIG_SECCOMP=y +# CONFIG_SMP is not set +# CONFIG_PREEMPT_NONE is not set +# CONFIG_PREEMPT_VOLUNTARY is not set +CONFIG_PREEMPT=y +CONFIG_GUSA=y +CONFIG_SPARSE_IRQ=y + +# +# Boot options +# +CONFIG_ZERO_PAGE_OFFSET=0x00001000 +CONFIG_BOOT_LINK_OFFSET=0x00800000 +CONFIG_ENTRY_OFFSET=0x00001000 +CONFIG_CMDLINE_OVERWRITE=y +# CONFIG_CMDLINE_EXTEND is not set +CONFIG_CMDLINE="console=ttySC1,115200 earlyprintk=sh-sci.1,115200 root=/dev/nfs ip=dhcp" + +# +# Bus options +# +# CONFIG_ARCH_SUPPORTS_MSI is not set +# CONFIG_PCCARD is not set + +# +# Executable file formats +# +CONFIG_BINFMT_ELF=y +# CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS is not set +# CONFIG_HAVE_AOUT is not set +CONFIG_BINFMT_MISC=y + +# +# Power management options (EXPERIMENTAL) +# +# CONFIG_PM is not set +# CONFIG_CPU_IDLE is not set +CONFIG_NET=y + +# +# Networking options +# +CONFIG_PACKET=y +CONFIG_PACKET_MMAP=y +CONFIG_UNIX=y +CONFIG_XFRM=y +# CONFIG_XFRM_USER is not set +# CONFIG_XFRM_SUB_POLICY is not set +# CONFIG_XFRM_MIGRATE is not set +# CONFIG_XFRM_STATISTICS is not set +CONFIG_NET_KEY=y +# CONFIG_NET_KEY_MIGRATE is not set +CONFIG_INET=y +# CONFIG_IP_MULTICAST is not set +# CONFIG_IP_ADVANCED_ROUTER is not set +CONFIG_IP_FIB_HASH=y +CONFIG_IP_PNP=y +CONFIG_IP_PNP_DHCP=y +# CONFIG_IP_PNP_BOOTP is not set +# CONFIG_IP_PNP_RARP is not set +# CONFIG_NET_IPIP is not set +# CONFIG_NET_IPGRE is not set +# CONFIG_ARPD is not set +# CONFIG_SYN_COOKIES is not set +# CONFIG_INET_AH is not set +# CONFIG_INET_ESP is not set +# CONFIG_INET_IPCOMP is not set +# CONFIG_INET_XFRM_TUNNEL is not set +# CONFIG_INET_TUNNEL is not set +CONFIG_INET_XFRM_MODE_TRANSPORT=y +CONFIG_INET_XFRM_MODE_TUNNEL=y +CONFIG_INET_XFRM_MODE_BEET=y +# CONFIG_INET_LRO is not set +CONFIG_INET_DIAG=y +CONFIG_INET_TCP_DIAG=y +# CONFIG_TCP_CONG_ADVANCED is not set +CONFIG_TCP_CONG_CUBIC=y +CONFIG_DEFAULT_TCP_CONG="cubic" +# CONFIG_TCP_MD5SIG is not set +# CONFIG_IPV6 is not set +# CONFIG_NETWORK_SECMARK is not set +# CONFIG_NETFILTER is not set +# CONFIG_IP_DCCP is not set +# CONFIG_IP_SCTP is not set +# CONFIG_RDS is not set +# CONFIG_TIPC is not set +# CONFIG_ATM is not set +# CONFIG_BRIDGE is not set +# CONFIG_NET_DSA is not set +# CONFIG_VLAN_8021Q is not set +# CONFIG_DECNET is not set +# CONFIG_LLC2 is not set +# CONFIG_IPX is not set +# CONFIG_ATALK is not set +# CONFIG_X25 is not set +# CONFIG_LAPB is not set +# CONFIG_ECONET is not set +# CONFIG_WAN_ROUTER is not set +# CONFIG_PHONET is not set +# CONFIG_IEEE802154 is not set +# CONFIG_NET_SCHED is not set +# CONFIG_DCB is not set + +# +# Network testing +# +# CONFIG_NET_PKTGEN is not set +# CONFIG_NET_DROP_MONITOR is not set +# CONFIG_HAMRADIO is not set +# CONFIG_CAN is not set +# CONFIG_IRDA is not set +# CONFIG_BT is not set +# CONFIG_AF_RXRPC is not set +CONFIG_WIRELESS=y +# CONFIG_CFG80211 is not set +# CONFIG_LIB80211 is not set + +# +# CFG80211 needs to be enabled for MAC80211 +# +# CONFIG_WIMAX is not set +# CONFIG_RFKILL is not set +# CONFIG_NET_9P is not set + +# +# Device Drivers +# + +# +# Generic Driver Options +# +CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug" +# CONFIG_DEVTMPFS is not set +CONFIG_STANDALONE=y +CONFIG_PREVENT_FIRMWARE_BUILD=y +# CONFIG_FW_LOADER is not set +# CONFIG_DEBUG_DRIVER is not set +# CONFIG_DEBUG_DEVRES is not set +# CONFIG_SYS_HYPERVISOR is not set +# CONFIG_CONNECTOR is not set +# CONFIG_MTD is not set +# CONFIG_PARPORT is not set +CONFIG_BLK_DEV=y +# CONFIG_BLK_DEV_COW_COMMON is not set +# CONFIG_BLK_DEV_LOOP is not set + +# +# DRBD disabled because PROC_FS, INET or CONNECTOR not selected +# +# CONFIG_BLK_DEV_NBD is not set +# CONFIG_BLK_DEV_UB is not set +CONFIG_BLK_DEV_RAM=y +CONFIG_BLK_DEV_RAM_COUNT=16 +CONFIG_BLK_DEV_RAM_SIZE=4096 +# CONFIG_BLK_DEV_XIP is not set +# CONFIG_CDROM_PKTCDVD is not set +# CONFIG_ATA_OVER_ETH is not set +# CONFIG_BLK_DEV_HD is not set +CONFIG_MISC_DEVICES=y +# CONFIG_AD525X_DPOT is not set +# CONFIG_ICS932S401 is not set +# CONFIG_ENCLOSURE_SERVICES is not set +# CONFIG_ISL29003 is not set +# CONFIG_DS1682 is not set +# CONFIG_TI_DAC7512 is not set +# CONFIG_C2PORT is not set + +# +# EEPROM support +# +# CONFIG_EEPROM_AT24 is not set +# CONFIG_EEPROM_AT25 is not set +# CONFIG_EEPROM_LEGACY is not set +# CONFIG_EEPROM_MAX6875 is not set +# CONFIG_EEPROM_93CX6 is not set +CONFIG_HAVE_IDE=y +# CONFIG_IDE is not set + +# +# SCSI device support +# +# CONFIG_RAID_ATTRS is not set +CONFIG_SCSI=y +CONFIG_SCSI_DMA=y +# CONFIG_SCSI_TGT is not set +# CONFIG_SCSI_NETLINK is not set +CONFIG_SCSI_PROC_FS=y + +# +# SCSI support type (disk, tape, CD-ROM) +# +CONFIG_BLK_DEV_SD=y +# CONFIG_CHR_DEV_ST is not set +# CONFIG_CHR_DEV_OSST is not set +# CONFIG_BLK_DEV_SR is not set +# CONFIG_CHR_DEV_SG is not set +# CONFIG_CHR_DEV_SCH is not set +# CONFIG_SCSI_MULTI_LUN is not set +# CONFIG_SCSI_CONSTANTS is not set +# CONFIG_SCSI_LOGGING is not set +# CONFIG_SCSI_SCAN_ASYNC is not set +CONFIG_SCSI_WAIT_SCAN=m + +# +# SCSI Transports +# +# CONFIG_SCSI_SPI_ATTRS is not set +# CONFIG_SCSI_FC_ATTRS is not set +# CONFIG_SCSI_ISCSI_ATTRS is not set +# CONFIG_SCSI_SAS_LIBSAS is not set +# CONFIG_SCSI_SRP_ATTRS is not set +CONFIG_SCSI_LOWLEVEL=y +# CONFIG_ISCSI_TCP is not set +# CONFIG_LIBFC is not set +# CONFIG_LIBFCOE is not set +# CONFIG_SCSI_DEBUG is not set +# CONFIG_SCSI_DH is not set +# CONFIG_SCSI_OSD_INITIATOR is not set +CONFIG_ATA=y +# CONFIG_ATA_NONSTANDARD is not set +CONFIG_ATA_VERBOSE_ERROR=y +CONFIG_SATA_PMP=y +CONFIG_ATA_SFF=y +# CONFIG_SATA_MV is not set +CONFIG_PATA_PLATFORM=y +# CONFIG_MD is not set +CONFIG_NETDEVICES=y +# CONFIG_DUMMY is not set +# CONFIG_BONDING is not set +# CONFIG_MACVLAN is not set +# CONFIG_EQUALIZER is not set +# CONFIG_TUN is not set +# CONFIG_VETH is not set +CONFIG_PHYLIB=y + +# +# MII PHY device drivers +# +# CONFIG_MARVELL_PHY is not set +# CONFIG_DAVICOM_PHY is not set +# CONFIG_QSEMI_PHY is not set +# CONFIG_LXT_PHY is not set +# CONFIG_CICADA_PHY is not set +# CONFIG_VITESSE_PHY is not set +# CONFIG_SMSC_PHY is not set +# CONFIG_BROADCOM_PHY is not set +# CONFIG_ICPLUS_PHY is not set +# CONFIG_REALTEK_PHY is not set +# CONFIG_NATIONAL_PHY is not set +# CONFIG_STE10XP is not set +# CONFIG_LSI_ET1011C_PHY is not set +# CONFIG_FIXED_PHY is not set +CONFIG_MDIO_BITBANG=y +CONFIG_NET_ETHERNET=y +CONFIG_MII=y +# CONFIG_AX88796 is not set +# CONFIG_STNIC is not set +CONFIG_SMC91X=y +# CONFIG_ENC28J60 is not set +# CONFIG_ETHOC is not set +# CONFIG_SMC911X is not set +CONFIG_SMSC911X=y +# CONFIG_DNET is not set +# CONFIG_IBM_NEW_EMAC_ZMII is not set +# CONFIG_IBM_NEW_EMAC_RGMII is not set +# CONFIG_IBM_NEW_EMAC_TAH is not set +# CONFIG_IBM_NEW_EMAC_EMAC4 is not set +# CONFIG_IBM_NEW_EMAC_NO_FLOW_CTRL is not set +# CONFIG_IBM_NEW_EMAC_MAL_CLR_ICINTSTAT is not set +# CONFIG_IBM_NEW_EMAC_MAL_COMMON_ERR is not set +# CONFIG_B44 is not set +# CONFIG_KS8842 is not set +# CONFIG_KS8851 is not set +# CONFIG_KS8851_MLL is not set +# CONFIG_NETDEV_1000 is not set +# CONFIG_NETDEV_10000 is not set +CONFIG_WLAN=y +# CONFIG_USB_ZD1201 is not set +# CONFIG_HOSTAP is not set + +# +# Enable WiMAX (Networking options) to see the WiMAX drivers +# + +# +# USB Network Adapters +# +# CONFIG_USB_CATC is not set +# CONFIG_USB_KAWETH is not set +# CONFIG_USB_PEGASUS is not set +# CONFIG_USB_RTL8150 is not set +# CONFIG_USB_USBNET is not set +# CONFIG_WAN is not set +# CONFIG_PPP is not set +# CONFIG_SLIP is not set +# CONFIG_NETCONSOLE is not set +# CONFIG_NETPOLL is not set +# CONFIG_NET_POLL_CONTROLLER is not set +# CONFIG_ISDN is not set +# CONFIG_PHONE is not set + +# +# Input device support +# +# CONFIG_INPUT is not set + +# +# Hardware I/O ports +# +# CONFIG_SERIO is not set +# CONFIG_GAMEPORT is not set + +# +# Character devices +# +# CONFIG_VT is not set +CONFIG_DEVKMEM=y +# CONFIG_SERIAL_NONSTANDARD is not set + +# +# Serial drivers +# +# CONFIG_SERIAL_8250 is not set + +# +# Non-8250 serial port support +# +# CONFIG_SERIAL_MAX3100 is not set +CONFIG_SERIAL_SH_SCI=y +CONFIG_SERIAL_SH_SCI_NR_UARTS=6 +CONFIG_SERIAL_SH_SCI_CONSOLE=y +CONFIG_SERIAL_CORE=y +CONFIG_SERIAL_CORE_CONSOLE=y +CONFIG_UNIX98_PTYS=y +# CONFIG_DEVPTS_MULTIPLE_INSTANCES is not set +# CONFIG_LEGACY_PTYS is not set +# CONFIG_IPMI_HANDLER is not set +# CONFIG_HW_RANDOM is not set +# CONFIG_R3964 is not set +# CONFIG_RAW_DRIVER is not set +# CONFIG_TCG_TPM is not set +CONFIG_I2C=y +CONFIG_I2C_BOARDINFO=y +# CONFIG_I2C_COMPAT is not set +CONFIG_I2C_CHARDEV=y +CONFIG_I2C_HELPER_AUTO=y + +# +# I2C Hardware Bus support +# + +# +# I2C system bus drivers (mostly embedded / system-on-chip) +# +# CONFIG_I2C_DESIGNWARE is not set +# CONFIG_I2C_OCORES is not set +CONFIG_I2C_SDK7786=y +# CONFIG_I2C_SH_MOBILE is not set +# CONFIG_I2C_SIMTEC is not set + +# +# External I2C/SMBus adapter drivers +# +# CONFIG_I2C_PARPORT_LIGHT is not set +# CONFIG_I2C_TAOS_EVM is not set +# CONFIG_I2C_TINY_USB is not set + +# +# Other I2C/SMBus bus drivers +# +# CONFIG_I2C_PCA_PLATFORM is not set +# CONFIG_I2C_STUB is not set + +# +# Miscellaneous I2C Chip support +# +# CONFIG_SENSORS_TSL2550 is not set +CONFIG_I2C_DEBUG_CORE=y +CONFIG_I2C_DEBUG_ALGO=y +CONFIG_I2C_DEBUG_BUS=y +CONFIG_I2C_DEBUG_CHIP=y +CONFIG_SPI=y +# CONFIG_SPI_DEBUG is not set +CONFIG_SPI_MASTER=y + +# +# SPI Master Controller Drivers +# +# CONFIG_SPI_BITBANG is not set +# CONFIG_SPI_SH_MSIOF is not set +# CONFIG_SPI_SH_SCI is not set +# CONFIG_SPI_XILINX is not set +# CONFIG_SPI_DESIGNWARE is not set + +# +# SPI Protocol Masters +# +# CONFIG_SPI_SPIDEV is not set +# CONFIG_SPI_TLE62X0 is not set + +# +# PPS support +# +# CONFIG_PPS is not set +# CONFIG_W1 is not set +# CONFIG_POWER_SUPPLY is not set +# CONFIG_HWMON is not set +# CONFIG_THERMAL is not set +CONFIG_WATCHDOG=y +# CONFIG_WATCHDOG_NOWAYOUT is not set + +# +# Watchdog Device Drivers +# +# CONFIG_SOFT_WATCHDOG is not set +# CONFIG_SH_WDT is not set + +# +# USB-based Watchdog Cards +# +# CONFIG_USBPCWATCHDOG is not set +CONFIG_SSB_POSSIBLE=y + +# +# Sonics Silicon Backplane +# +# CONFIG_SSB is not set + +# +# Multifunction device drivers +# +# CONFIG_MFD_CORE is not set +# CONFIG_MFD_SM501 is not set +# CONFIG_MFD_SH_MOBILE_SDHI is not set +# CONFIG_HTC_PASIC3 is not set +# CONFIG_TWL4030_CORE is not set +# CONFIG_MFD_TMIO is not set +# CONFIG_PMIC_DA903X is not set +# CONFIG_PMIC_ADP5520 is not set +# CONFIG_MFD_WM8400 is not set +# CONFIG_MFD_WM831X is not set +# CONFIG_MFD_WM8350_I2C is not set +# CONFIG_MFD_PCF50633 is not set +# CONFIG_MFD_MC13783 is not set +# CONFIG_AB3100_CORE is not set +# CONFIG_EZX_PCAP is not set +# CONFIG_MFD_88PM8607 is not set +# CONFIG_AB4500_CORE is not set +# CONFIG_REGULATOR is not set +# CONFIG_MEDIA_SUPPORT is not set + +# +# Graphics support +# +# CONFIG_VGASTATE is not set +CONFIG_VIDEO_OUTPUT_CONTROL=m +# CONFIG_FB is not set +# CONFIG_BACKLIGHT_LCD_SUPPORT is not set + +# +# Display device support +# +# CONFIG_DISPLAY_SUPPORT is not set +# CONFIG_SOUND is not set +CONFIG_USB_SUPPORT=y +CONFIG_USB_ARCH_HAS_HCD=y +CONFIG_USB_ARCH_HAS_OHCI=y +# CONFIG_USB_ARCH_HAS_EHCI is not set +CONFIG_USB=y +# CONFIG_USB_DEBUG is not set +# CONFIG_USB_ANNOUNCE_NEW_DEVICES is not set + +# +# Miscellaneous USB options +# +# CONFIG_USB_DEVICEFS is not set +CONFIG_USB_DEVICE_CLASS=y +# CONFIG_USB_DYNAMIC_MINORS is not set +# CONFIG_USB_OTG is not set +# CONFIG_USB_OTG_WHITELIST is not set +# CONFIG_USB_OTG_BLACKLIST_HUB is not set +CONFIG_USB_MON=y +# CONFIG_USB_WUSB is not set +# CONFIG_USB_WUSB_CBAF is not set + +# +# USB Host Controller Drivers +# +# CONFIG_USB_C67X00_HCD is not set +# CONFIG_USB_OXU210HP_HCD is not set +# CONFIG_USB_ISP116X_HCD is not set +# CONFIG_USB_ISP1760_HCD is not set +# CONFIG_USB_ISP1362_HCD is not set +# CONFIG_USB_OHCI_HCD is not set +# CONFIG_USB_SL811_HCD is not set +CONFIG_USB_R8A66597_HCD=m +# CONFIG_USB_HWA_HCD is not set +# CONFIG_USB_GADGET_MUSB_HDRC is not set + +# +# USB Device Class drivers +# +# CONFIG_USB_ACM is not set +# CONFIG_USB_PRINTER is not set +# CONFIG_USB_WDM is not set +# CONFIG_USB_TMC is not set + +# +# NOTE: USB_STORAGE depends on SCSI but BLK_DEV_SD may +# + +# +# also be needed; see USB_STORAGE Help for more info +# +# CONFIG_USB_STORAGE is not set +# CONFIG_USB_LIBUSUAL is not set + +# +# USB Imaging devices +# +# CONFIG_USB_MDC800 is not set +# CONFIG_USB_MICROTEK is not set + +# +# USB port drivers +# +# CONFIG_USB_SERIAL is not set + +# +# USB Miscellaneous drivers +# +# CONFIG_USB_EMI62 is not set +# CONFIG_USB_EMI26 is not set +# CONFIG_USB_ADUTUX is not set +# CONFIG_USB_SEVSEG is not set +# CONFIG_USB_RIO500 is not set +# CONFIG_USB_LEGOTOWER is not set +# CONFIG_USB_LCD is not set +# CONFIG_USB_BERRY_CHARGE is not set +# CONFIG_USB_LED is not set +# CONFIG_USB_CYPRESS_CY7C63 is not set +# CONFIG_USB_CYTHERM is not set +# CONFIG_USB_IDMOUSE is not set +# CONFIG_USB_FTDI_ELAN is not set +# CONFIG_USB_APPLEDISPLAY is not set +# CONFIG_USB_LD is not set +# CONFIG_USB_TRANCEVIBRATOR is not set +# CONFIG_USB_IOWARRIOR is not set +# CONFIG_USB_TEST is not set +# CONFIG_USB_ISIGHTFW is not set +# CONFIG_USB_VST is not set +CONFIG_USB_GADGET=y +# CONFIG_USB_GADGET_DEBUG is not set +# CONFIG_USB_GADGET_DEBUG_FILES is not set +# CONFIG_USB_GADGET_DEBUG_FS is not set +CONFIG_USB_GADGET_VBUS_DRAW=2 +CONFIG_USB_GADGET_SELECTED=y +# CONFIG_USB_GADGET_AT91 is not set +# CONFIG_USB_GADGET_ATMEL_USBA is not set +# CONFIG_USB_GADGET_FSL_USB2 is not set +# CONFIG_USB_GADGET_LH7A40X is not set +# CONFIG_USB_GADGET_OMAP is not set +# CONFIG_USB_GADGET_PXA25X is not set +# CONFIG_USB_GADGET_R8A66597 is not set +# CONFIG_USB_GADGET_PXA27X is not set +# CONFIG_USB_GADGET_S3C_HSOTG is not set +# CONFIG_USB_GADGET_IMX is not set +# CONFIG_USB_GADGET_S3C2410 is not set +CONFIG_USB_GADGET_M66592=y +CONFIG_USB_M66592=y +# CONFIG_USB_GADGET_AMD5536UDC is not set +# CONFIG_USB_GADGET_FSL_QE is not set +# CONFIG_USB_GADGET_CI13XXX is not set +# CONFIG_USB_GADGET_NET2280 is not set +# CONFIG_USB_GADGET_GOKU is not set +# CONFIG_USB_GADGET_LANGWELL is not set +# CONFIG_USB_GADGET_DUMMY_HCD is not set +CONFIG_USB_GADGET_DUALSPEED=y +# CONFIG_USB_ZERO is not set +# CONFIG_USB_AUDIO is not set +# CONFIG_USB_ETH is not set +# CONFIG_USB_GADGETFS is not set +# CONFIG_USB_FILE_STORAGE is not set +# CONFIG_USB_MASS_STORAGE is not set +# CONFIG_USB_G_SERIAL is not set +# CONFIG_USB_MIDI_GADGET is not set +# CONFIG_USB_G_PRINTER is not set +# CONFIG_USB_CDC_COMPOSITE is not set +# CONFIG_USB_G_MULTI is not set + +# +# OTG and related infrastructure +# +# CONFIG_NOP_USB_XCEIV is not set +# CONFIG_MMC is not set +# CONFIG_MEMSTICK is not set +# CONFIG_NEW_LEDS is not set +# CONFIG_ACCESSIBILITY is not set +CONFIG_RTC_LIB=y +CONFIG_RTC_CLASS=y +CONFIG_RTC_HCTOSYS=y +CONFIG_RTC_HCTOSYS_DEVICE="rtc0" +# CONFIG_RTC_DEBUG is not set + +# +# RTC interfaces +# +CONFIG_RTC_INTF_SYSFS=y +CONFIG_RTC_INTF_PROC=y +CONFIG_RTC_INTF_DEV=y +# CONFIG_RTC_INTF_DEV_UIE_EMUL is not set +# CONFIG_RTC_DRV_TEST is not set + +# +# I2C RTC drivers +# +# CONFIG_RTC_DRV_DS1307 is not set +# CONFIG_RTC_DRV_DS1374 is not set +# CONFIG_RTC_DRV_DS1672 is not set +CONFIG_RTC_DRV_MAX6900=y +# CONFIG_RTC_DRV_RS5C372 is not set +# CONFIG_RTC_DRV_ISL1208 is not set +# CONFIG_RTC_DRV_X1205 is not set +# CONFIG_RTC_DRV_PCF8563 is not set +# CONFIG_RTC_DRV_PCF8583 is not set +# CONFIG_RTC_DRV_M41T80 is not set +# CONFIG_RTC_DRV_BQ32K is not set +# CONFIG_RTC_DRV_S35390A is not set +# CONFIG_RTC_DRV_FM3130 is not set +# CONFIG_RTC_DRV_RX8581 is not set +# CONFIG_RTC_DRV_RX8025 is not set + +# +# SPI RTC drivers +# +# CONFIG_RTC_DRV_M41T94 is not set +# CONFIG_RTC_DRV_DS1305 is not set +# CONFIG_RTC_DRV_DS1390 is not set +# CONFIG_RTC_DRV_MAX6902 is not set +# CONFIG_RTC_DRV_R9701 is not set +# CONFIG_RTC_DRV_RS5C348 is not set +# CONFIG_RTC_DRV_DS3234 is not set +# CONFIG_RTC_DRV_PCF2123 is not set + +# +# Platform RTC drivers +# +# CONFIG_RTC_DRV_DS1286 is not set +# CONFIG_RTC_DRV_DS1511 is not set +# CONFIG_RTC_DRV_DS1553 is not set +# CONFIG_RTC_DRV_DS1742 is not set +# CONFIG_RTC_DRV_STK17TA8 is not set +# CONFIG_RTC_DRV_M48T86 is not set +# CONFIG_RTC_DRV_M48T35 is not set +# CONFIG_RTC_DRV_M48T59 is not set +# CONFIG_RTC_DRV_MSM6242 is not set +# CONFIG_RTC_DRV_BQ4802 is not set +# CONFIG_RTC_DRV_RP5C01 is not set +# CONFIG_RTC_DRV_V3020 is not set + +# +# on-CPU RTC drivers +# +CONFIG_RTC_DRV_SH=y +# CONFIG_RTC_DRV_GENERIC is not set +# CONFIG_DMADEVICES is not set +# CONFIG_AUXDISPLAY is not set +CONFIG_UIO=m +# CONFIG_UIO_PDRV is not set +# CONFIG_UIO_PDRV_GENIRQ is not set +# CONFIG_UIO_SMX is not set +# CONFIG_UIO_SERCOS3 is not set + +# +# TI VLYNQ +# +# CONFIG_STAGING is not set + +# +# File systems +# +CONFIG_EXT2_FS=y +# CONFIG_EXT2_FS_XATTR is not set +# CONFIG_EXT2_FS_XIP is not set +CONFIG_EXT3_FS=y +# CONFIG_EXT3_DEFAULTS_TO_ORDERED is not set +CONFIG_EXT3_FS_XATTR=y +# CONFIG_EXT3_FS_POSIX_ACL is not set +# CONFIG_EXT3_FS_SECURITY is not set +# CONFIG_EXT4_FS is not set +CONFIG_JBD=y +# CONFIG_JBD_DEBUG is not set +CONFIG_FS_MBCACHE=y +# CONFIG_REISERFS_FS is not set +# CONFIG_JFS_FS is not set +# CONFIG_FS_POSIX_ACL is not set +# CONFIG_XFS_FS is not set +# CONFIG_OCFS2_FS is not set +# CONFIG_BTRFS_FS is not set +# CONFIG_NILFS2_FS is not set +CONFIG_FILE_LOCKING=y +CONFIG_FSNOTIFY=y +CONFIG_DNOTIFY=y +CONFIG_INOTIFY=y +CONFIG_INOTIFY_USER=y +# CONFIG_QUOTA is not set +# CONFIG_AUTOFS_FS is not set +# CONFIG_AUTOFS4_FS is not set +# CONFIG_FUSE_FS is not set + +# +# Caches +# +# CONFIG_FSCACHE is not set + +# +# CD-ROM/DVD Filesystems +# +# CONFIG_ISO9660_FS is not set +# CONFIG_UDF_FS is not set + +# +# DOS/FAT/NT Filesystems +# +# CONFIG_MSDOS_FS is not set +# CONFIG_VFAT_FS is not set +# CONFIG_NTFS_FS is not set + +# +# Pseudo filesystems +# +CONFIG_PROC_FS=y +CONFIG_PROC_KCORE=y +CONFIG_PROC_SYSCTL=y +CONFIG_PROC_PAGE_MONITOR=y +CONFIG_SYSFS=y +CONFIG_TMPFS=y +# CONFIG_TMPFS_POSIX_ACL is not set +CONFIG_HUGETLBFS=y +CONFIG_HUGETLB_PAGE=y +# CONFIG_CONFIGFS_FS is not set +CONFIG_MISC_FILESYSTEMS=y +# CONFIG_ADFS_FS is not set +# CONFIG_AFFS_FS is not set +# CONFIG_HFS_FS is not set +# CONFIG_HFSPLUS_FS is not set +# CONFIG_BEFS_FS is not set +# CONFIG_BFS_FS is not set +# CONFIG_EFS_FS is not set +# CONFIG_CRAMFS is not set +# CONFIG_SQUASHFS is not set +# CONFIG_VXFS_FS is not set +# CONFIG_MINIX_FS is not set +# CONFIG_OMFS_FS is not set +# CONFIG_HPFS_FS is not set +# CONFIG_QNX4FS_FS is not set +# CONFIG_ROMFS_FS is not set +# CONFIG_SYSV_FS is not set +# CONFIG_UFS_FS is not set +CONFIG_NETWORK_FILESYSTEMS=y +CONFIG_NFS_FS=y +CONFIG_NFS_V3=y +# CONFIG_NFS_V3_ACL is not set +# CONFIG_NFS_V4 is not set +CONFIG_ROOT_NFS=y +# CONFIG_NFSD is not set +CONFIG_LOCKD=y +CONFIG_LOCKD_V4=y +CONFIG_NFS_COMMON=y +CONFIG_SUNRPC=y +# CONFIG_RPCSEC_GSS_KRB5 is not set +# CONFIG_RPCSEC_GSS_SPKM3 is not set +# CONFIG_SMB_FS is not set +# CONFIG_CIFS is not set +# CONFIG_NCP_FS is not set +# CONFIG_CODA_FS is not set +# CONFIG_AFS_FS is not set + +# +# Partition Types +# +# CONFIG_PARTITION_ADVANCED is not set +CONFIG_MSDOS_PARTITION=y +CONFIG_NLS=y +CONFIG_NLS_DEFAULT="iso8859-1" +# CONFIG_NLS_CODEPAGE_437 is not set +# CONFIG_NLS_CODEPAGE_737 is not set +# CONFIG_NLS_CODEPAGE_775 is not set +# CONFIG_NLS_CODEPAGE_850 is not set +# CONFIG_NLS_CODEPAGE_852 is not set +# CONFIG_NLS_CODEPAGE_855 is not set +# CONFIG_NLS_CODEPAGE_857 is not set +# CONFIG_NLS_CODEPAGE_860 is not set +# CONFIG_NLS_CODEPAGE_861 is not set +# CONFIG_NLS_CODEPAGE_862 is not set +# CONFIG_NLS_CODEPAGE_863 is not set +# CONFIG_NLS_CODEPAGE_864 is not set +# CONFIG_NLS_CODEPAGE_865 is not set +# CONFIG_NLS_CODEPAGE_866 is not set +# CONFIG_NLS_CODEPAGE_869 is not set +# CONFIG_NLS_CODEPAGE_936 is not set +# CONFIG_NLS_CODEPAGE_950 is not set +# CONFIG_NLS_CODEPAGE_932 is not set +# CONFIG_NLS_CODEPAGE_949 is not set +# CONFIG_NLS_CODEPAGE_874 is not set +# CONFIG_NLS_ISO8859_8 is not set +# CONFIG_NLS_CODEPAGE_1250 is not set +# CONFIG_NLS_CODEPAGE_1251 is not set +# CONFIG_NLS_ASCII is not set +# CONFIG_NLS_ISO8859_1 is not set +# CONFIG_NLS_ISO8859_2 is not set +# CONFIG_NLS_ISO8859_3 is not set +# CONFIG_NLS_ISO8859_4 is not set +# CONFIG_NLS_ISO8859_5 is not set +# CONFIG_NLS_ISO8859_6 is not set +# CONFIG_NLS_ISO8859_7 is not set +# CONFIG_NLS_ISO8859_9 is not set +# CONFIG_NLS_ISO8859_13 is not set +# CONFIG_NLS_ISO8859_14 is not set +# CONFIG_NLS_ISO8859_15 is not set +# CONFIG_NLS_KOI8_R is not set +# CONFIG_NLS_KOI8_U is not set +# CONFIG_NLS_UTF8 is not set +# CONFIG_DLM is not set + +# +# Kernel hacking +# +CONFIG_TRACE_IRQFLAGS_SUPPORT=y +CONFIG_PRINTK_TIME=y +CONFIG_ENABLE_WARN_DEPRECATED=y +# CONFIG_ENABLE_MUST_CHECK is not set +CONFIG_FRAME_WARN=1024 +CONFIG_MAGIC_SYSRQ=y +# CONFIG_STRIP_ASM_SYMS is not set +# CONFIG_UNUSED_SYMBOLS is not set +CONFIG_DEBUG_FS=y +# CONFIG_HEADERS_CHECK is not set +CONFIG_DEBUG_KERNEL=y +CONFIG_DEBUG_SHIRQ=y +CONFIG_DETECT_SOFTLOCKUP=y +# CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC is not set +CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC_VALUE=0 +CONFIG_DETECT_HUNG_TASK=y +# CONFIG_BOOTPARAM_HUNG_TASK_PANIC is not set +CONFIG_BOOTPARAM_HUNG_TASK_PANIC_VALUE=0 +CONFIG_SCHED_DEBUG=y +# CONFIG_SCHEDSTATS is not set +# CONFIG_TIMER_STATS is not set +# CONFIG_DEBUG_OBJECTS is not set +# CONFIG_DEBUG_SLAB is not set +CONFIG_DEBUG_PREEMPT=y +# CONFIG_DEBUG_RT_MUTEXES is not set +# CONFIG_RT_MUTEX_TESTER is not set +# CONFIG_DEBUG_SPINLOCK is not set +# CONFIG_DEBUG_MUTEXES is not set +# CONFIG_DEBUG_LOCK_ALLOC is not set +# CONFIG_PROVE_LOCKING is not set +# CONFIG_LOCK_STAT is not set +# CONFIG_DEBUG_SPINLOCK_SLEEP is not set +# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set +CONFIG_STACKTRACE=y +# CONFIG_DEBUG_KOBJECT is not set +CONFIG_DEBUG_BUGVERBOSE=y +# CONFIG_DEBUG_INFO is not set +CONFIG_DEBUG_VM=y +# CONFIG_DEBUG_WRITECOUNT is not set +# CONFIG_DEBUG_MEMORY_INIT is not set +# CONFIG_DEBUG_LIST is not set +# CONFIG_DEBUG_SG is not set +# CONFIG_DEBUG_NOTIFIERS is not set +# CONFIG_DEBUG_CREDENTIALS is not set +CONFIG_FRAME_POINTER=y +# CONFIG_RCU_TORTURE_TEST is not set +# CONFIG_RCU_CPU_STALL_DETECTOR is not set +# CONFIG_BACKTRACE_SELF_TEST is not set +# CONFIG_DEBUG_BLOCK_EXT_DEVT is not set +# CONFIG_DEBUG_FORCE_WEAK_PER_CPU is not set +# CONFIG_FAULT_INJECTION is not set +# CONFIG_LATENCYTOP is not set +# CONFIG_SYSCTL_SYSCALL_CHECK is not set +# CONFIG_PAGE_POISONING is not set +CONFIG_NOP_TRACER=y +CONFIG_HAVE_FUNCTION_TRACER=y +CONFIG_HAVE_FUNCTION_GRAPH_TRACER=y +CONFIG_HAVE_FUNCTION_TRACE_MCOUNT_TEST=y +CONFIG_HAVE_DYNAMIC_FTRACE=y +CONFIG_HAVE_FTRACE_MCOUNT_RECORD=y +CONFIG_HAVE_SYSCALL_TRACEPOINTS=y +CONFIG_RING_BUFFER=y +CONFIG_EVENT_TRACING=y +CONFIG_CONTEXT_SWITCH_TRACER=y +CONFIG_TRACING=y +CONFIG_TRACING_SUPPORT=y +CONFIG_FTRACE=y +# CONFIG_FUNCTION_TRACER is not set +# CONFIG_IRQSOFF_TRACER is not set +# CONFIG_PREEMPT_TRACER is not set +# CONFIG_SCHED_TRACER is not set +# CONFIG_ENABLE_DEFAULT_TRACERS is not set +# CONFIG_FTRACE_SYSCALLS is not set +# CONFIG_BOOT_TRACER is not set +CONFIG_BRANCH_PROFILE_NONE=y +# CONFIG_PROFILE_ANNOTATED_BRANCHES is not set +# CONFIG_PROFILE_ALL_BRANCHES is not set +CONFIG_KSYM_TRACER=y +# CONFIG_PROFILE_KSYM_TRACER is not set +# CONFIG_STACK_TRACER is not set +# CONFIG_KMEMTRACE is not set +# CONFIG_WORKQUEUE_TRACER is not set +# CONFIG_BLK_DEV_IO_TRACE is not set +# CONFIG_RING_BUFFER_BENCHMARK is not set +# CONFIG_DYNAMIC_DEBUG is not set +# CONFIG_DMA_API_DEBUG is not set +# CONFIG_SAMPLES is not set +CONFIG_HAVE_ARCH_KGDB=y +# CONFIG_KGDB is not set +# CONFIG_SH_STANDARD_BIOS is not set +# CONFIG_STACK_DEBUG is not set +CONFIG_DEBUG_STACK_USAGE=y +# CONFIG_4KSTACKS is not set +CONFIG_DUMP_CODE=y +CONFIG_DWARF_UNWINDER=y +# CONFIG_SH_NO_BSS_INIT is not set + +# +# Security options +# +# CONFIG_KEYS is not set +# CONFIG_SECURITY is not set +# CONFIG_SECURITYFS is not set +# CONFIG_DEFAULT_SECURITY_SELINUX is not set +# CONFIG_DEFAULT_SECURITY_SMACK is not set +# CONFIG_DEFAULT_SECURITY_TOMOYO is not set +CONFIG_DEFAULT_SECURITY_DAC=y +CONFIG_DEFAULT_SECURITY="" +CONFIG_CRYPTO=y + +# +# Crypto core or helper +# +# CONFIG_CRYPTO_MANAGER is not set +# CONFIG_CRYPTO_MANAGER2 is not set +# CONFIG_CRYPTO_GF128MUL is not set +# CONFIG_CRYPTO_NULL is not set +# CONFIG_CRYPTO_CRYPTD is not set +# CONFIG_CRYPTO_AUTHENC is not set +# CONFIG_CRYPTO_TEST is not set + +# +# Authenticated Encryption with Associated Data +# +# CONFIG_CRYPTO_CCM is not set +# CONFIG_CRYPTO_GCM is not set +# CONFIG_CRYPTO_SEQIV is not set + +# +# Block modes +# +# CONFIG_CRYPTO_CBC is not set +# CONFIG_CRYPTO_CTR is not set +# CONFIG_CRYPTO_CTS is not set +# CONFIG_CRYPTO_ECB is not set +# CONFIG_CRYPTO_LRW is not set +# CONFIG_CRYPTO_PCBC is not set +# CONFIG_CRYPTO_XTS is not set + +# +# Hash modes +# +# CONFIG_CRYPTO_HMAC is not set +# CONFIG_CRYPTO_XCBC is not set +# CONFIG_CRYPTO_VMAC is not set + +# +# Digest +# +# CONFIG_CRYPTO_CRC32C is not set +# CONFIG_CRYPTO_GHASH is not set +# CONFIG_CRYPTO_MD4 is not set +# CONFIG_CRYPTO_MD5 is not set +# CONFIG_CRYPTO_MICHAEL_MIC is not set +# CONFIG_CRYPTO_RMD128 is not set +# CONFIG_CRYPTO_RMD160 is not set +# CONFIG_CRYPTO_RMD256 is not set +# CONFIG_CRYPTO_RMD320 is not set +# CONFIG_CRYPTO_SHA1 is not set +# CONFIG_CRYPTO_SHA256 is not set +# CONFIG_CRYPTO_SHA512 is not set +# CONFIG_CRYPTO_TGR192 is not set +# CONFIG_CRYPTO_WP512 is not set + +# +# Ciphers +# +# CONFIG_CRYPTO_AES is not set +# CONFIG_CRYPTO_ANUBIS is not set +# CONFIG_CRYPTO_ARC4 is not set +# CONFIG_CRYPTO_BLOWFISH is not set +# CONFIG_CRYPTO_CAMELLIA is not set +# CONFIG_CRYPTO_CAST5 is not set +# CONFIG_CRYPTO_CAST6 is not set +# CONFIG_CRYPTO_DES is not set +# CONFIG_CRYPTO_FCRYPT is not set +# CONFIG_CRYPTO_KHAZAD is not set +# CONFIG_CRYPTO_SALSA20 is not set +# CONFIG_CRYPTO_SEED is not set +# CONFIG_CRYPTO_SERPENT is not set +# CONFIG_CRYPTO_TEA is not set +# CONFIG_CRYPTO_TWOFISH is not set + +# +# Compression +# +# CONFIG_CRYPTO_DEFLATE is not set +# CONFIG_CRYPTO_ZLIB is not set +# CONFIG_CRYPTO_LZO is not set + +# +# Random Number Generation +# +# CONFIG_CRYPTO_ANSI_CPRNG is not set +CONFIG_CRYPTO_HW=y +CONFIG_BINARY_PRINTF=y + +# +# Library routines +# +CONFIG_BITREVERSE=y +CONFIG_GENERIC_FIND_LAST_BIT=y +# CONFIG_CRC_CCITT is not set +# CONFIG_CRC16 is not set +# CONFIG_CRC_T10DIF is not set +# CONFIG_CRC_ITU_T is not set +CONFIG_CRC32=y +# CONFIG_CRC7 is not set +# CONFIG_LIBCRC32C is not set +CONFIG_HAS_IOMEM=y +CONFIG_HAS_IOPORT=y +CONFIG_HAS_DMA=y +CONFIG_HAVE_LMB=y +CONFIG_NLATTR=y +CONFIG_GENERIC_ATOMIC64=y diff --git a/arch/sh/tools/mach-types b/arch/sh/tools/mach-types index 6639b25..b25aa55 100644 --- a/arch/sh/tools/mach-types +++ b/arch/sh/tools/mach-types @@ -32,6 +32,7 @@ DREAMCAST SH_DREAMCAST SNAPGEAR SH_SECUREEDGE5410 EDOSK7705 SH_EDOSK7705 EDOSK7760 SH_EDOSK7760 +SDK7786 SH_SDK7786 SH4202_MICRODEV SH_SH4202_MICRODEV SH03 SH_SH03 LANDISK SH_LANDISK -- cgit v0.10.2 From 46c4e5daea3d5df06e27bf5a49a0c42274db6725 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Fri, 15 Jan 2010 08:00:45 +0900 Subject: sh: Fix CONFIG_PMB=n build. The last commit introduced the following breakage arch/sh/include/asm/mmu.h: In function 'pmb_remap': arch/sh/include/asm/mmu.h:79: error: expected ';' before '}' token and... arch/sh/include/asm/mmu.h:78: error: 'EINVAL' undeclared (first use in this function) arch/sh/include/asm/mmu.h:78: error: (Each undeclared identifier is reported only once arch/sh/include/asm/mmu.h:78: error: for each function it appears in.) arch/sh/include/asm/mmu.h: In function 'pmb_init': arch/sh/include/asm/mmu.h:87: error: 'ENODEV' undeclared (first use in this function) Signed-off-by: Matt Fleming Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/mmu.h b/arch/sh/include/asm/mmu.h index 4b0882b..e5e8f48 100644 --- a/arch/sh/include/asm/mmu.h +++ b/arch/sh/include/asm/mmu.h @@ -30,6 +30,7 @@ #define PMB_NO_ENTRY (-1) #ifndef __ASSEMBLY__ +#include /* Default "unsigned long" context */ typedef unsigned long mm_context_id_t[NR_CPUS]; @@ -75,7 +76,7 @@ int pmb_init(void); static inline long pmb_remap(unsigned long virt, unsigned long phys, unsigned long size, unsigned long flags) { - return -EINVAL + return -EINVAL; } static inline void pmb_unmap(unsigned long addr) -- cgit v0.10.2 From 10ab92d8c336c25af1fce46113ed85856e02e95d Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 15 Jan 2010 12:08:31 +0900 Subject: sh: heartbeat: Support access size specification via resource flags. This permits the resource access size to be handed off through the resource flags, which saves platforms from having to establish platform data only to specify the register width. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/heartbeat.c b/arch/sh/drivers/heartbeat.c index a9339a6..2acbc79 100644 --- a/arch/sh/drivers/heartbeat.c +++ b/arch/sh/drivers/heartbeat.c @@ -1,7 +1,7 @@ /* * Generic heartbeat driver for regular LED banks * - * Copyright (C) 2007 Paul Mundt + * Copyright (C) 2007 - 2010 Paul Mundt * * Most SH reference boards include a number of individual LEDs that can * be independently controlled (either via a pre-defined hardware @@ -27,7 +27,7 @@ #include #define DRV_NAME "heartbeat" -#define DRV_VERSION "0.1.1" +#define DRV_VERSION "0.1.2" static unsigned char default_bit_pos[] = { 0, 1, 2, 3, 4, 5, 6, 7 }; @@ -98,7 +98,7 @@ static int heartbeat_drv_probe(struct platform_device *pdev) return -ENOMEM; } - hd->base = ioremap_nocache(res->start, res->end - res->start + 1); + hd->base = ioremap_nocache(res->start, resource_size(res)); if (unlikely(!hd->base)) { dev_err(&pdev->dev, "ioremap failed\n"); @@ -117,8 +117,20 @@ static int heartbeat_drv_probe(struct platform_device *pdev) for (i = 0; i < hd->nr_bits; i++) hd->mask |= (1 << hd->bit_pos[i]); - if (!hd->regsize) - hd->regsize = 8; /* default access size */ + if (!hd->regsize) { + switch (res->flags & IORESOURCE_MEM_TYPE_MASK) { + case IORESOURCE_MEM_32BIT: + hd->regsize = 32; + break; + case IORESOURCE_MEM_16BIT: + hd->regsize = 16; + break; + case IORESOURCE_MEM_8BIT: + default: + hd->regsize = 8; + break; + } + } setup_timer(&hd->timer, heartbeat_timer, (unsigned long)hd); platform_set_drvdata(pdev, hd); -- cgit v0.10.2 From 2267c7875b3a3a41b7a8cc9659d75ec934c98ffd Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 15 Jan 2010 12:11:30 +0900 Subject: sh: mach-sdk7786: heartbeat support. Hand off the user LEDs to the heartbeat driver. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-sdk7786/setup.c b/arch/sh/boards/mach-sdk7786/setup.c index b460dcc..f2d4b75 100644 --- a/arch/sh/boards/mach-sdk7786/setup.c +++ b/arch/sh/boards/mach-sdk7786/setup.c @@ -15,8 +15,22 @@ #include #include #include +#include #include +static struct resource heartbeat_resource = { + .start = 0x07fff8b0, + .end = 0x07fff8b0 + sizeof(u16) - 1, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_16BIT, +}; + +static struct platform_device heartbeat_device = { + .name = "heartbeat", + .id = -1, + .num_resources = 1, + .resource = &heartbeat_resource, +}; + static struct resource smsc911x_resources[] = { [0] = { .name = "smsc911x-memory", @@ -82,6 +96,7 @@ static struct i2c_board_info __initdata sdk7786_i2c_devices[] = { }; static struct platform_device *sh7786_devices[] __initdata = { + &heartbeat_device, &smsc911x_device, &smbus_fpga_device, &smbus_pcie_device, -- cgit v0.10.2 From a09d2831b3457a92a58d67bddfed370516acfe28 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 15 Jan 2010 12:24:34 +0900 Subject: sh: heartbeat: Update boards for access size hinting. This updates the existing boards that specify the register width through platform data to use the resource flags instead. This eliminates platform data completely in most cases, and permits further improvement in the heartbeat driver as well as shrinking the overall private data size. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/board-polaris.c b/arch/sh/boards/board-polaris.c index 62607eb..5bc1269 100644 --- a/arch/sh/boards/board-polaris.c +++ b/arch/sh/boards/board-polaris.c @@ -59,15 +59,12 @@ static unsigned char heartbeat_bit_pos[] = { 0, 1, 2, 3 }; static struct heartbeat_data heartbeat_data = { .bit_pos = heartbeat_bit_pos, .nr_bits = ARRAY_SIZE(heartbeat_bit_pos), - .regsize = 8, }; -static struct resource heartbeat_resources[] = { - [0] = { - .start = PORT_PCDR, - .end = PORT_PCDR, - .flags = IORESOURCE_MEM, - }, +static struct resource heartbeat_resource = { + .start = PORT_PCDR, + .end = PORT_PCDR, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_8BIT, }; static struct platform_device heartbeat_device = { @@ -76,8 +73,8 @@ static struct platform_device heartbeat_device = { .dev = { .platform_data = &heartbeat_data, }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1, + .resource = &heartbeat_resource, }; static struct platform_device *polaris_devices[] __initdata = { diff --git a/arch/sh/boards/board-sh7785lcr.c b/arch/sh/boards/board-sh7785lcr.c index e5a8a2f..7eea90d 100644 --- a/arch/sh/boards/board-sh7785lcr.c +++ b/arch/sh/boards/board-sh7785lcr.c @@ -32,26 +32,17 @@ * NOTE: This board has 2 physical memory maps. * Please look at include/asm-sh/sh7785lcr.h or hardware manual. */ -static struct resource heartbeat_resources[] = { - [0] = { - .start = PLD_LEDCR, - .end = PLD_LEDCR, - .flags = IORESOURCE_MEM, - }, -}; - -static struct heartbeat_data heartbeat_data = { - .regsize = 8, +static struct resource heartbeat_resource = { + .start = PLD_LEDCR, + .end = PLD_LEDCR, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_8BIT, }; static struct platform_device heartbeat_device = { .name = "heartbeat", .id = -1, - .dev = { - .platform_data = &heartbeat_data, - }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1, + .resource = &heartbeat_resource, }; static struct mtd_partition nor_flash_partitions[] = { diff --git a/arch/sh/boards/board-urquell.c b/arch/sh/boards/board-urquell.c index 36b8bac..268f09e 100644 --- a/arch/sh/boards/board-urquell.c +++ b/arch/sh/boards/board-urquell.c @@ -50,26 +50,17 @@ */ /* HeartBeat */ -static struct resource heartbeat_resources[] = { - [0] = { - .start = BOARDREG(SLEDR), - .end = BOARDREG(SLEDR), - .flags = IORESOURCE_MEM, - }, -}; - -static struct heartbeat_data heartbeat_data = { - .regsize = 16, +static struct resource heartbeat_resource = { + .start = BOARDREG(SLEDR), + .end = BOARDREG(SLEDR), + .flags = IORESOURCE_MEM | IORESOURCE_MEM_16BIT, }; static struct platform_device heartbeat_device = { .name = "heartbeat", .id = -1, - .dev = { - .platform_data = &heartbeat_data, - }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1 + .resource = &heartbeat_resource, }; /* LAN91C111 */ diff --git a/arch/sh/boards/mach-ecovec24/setup.c b/arch/sh/boards/mach-ecovec24/setup.c index 6a8861b..a6b2233 100644 --- a/arch/sh/boards/mach-ecovec24/setup.c +++ b/arch/sh/boards/mach-ecovec24/setup.c @@ -64,18 +64,16 @@ /* Heartbeat */ static unsigned char led_pos[] = { 0, 1, 2, 3 }; + static struct heartbeat_data heartbeat_data = { - .regsize = 8, .nr_bits = 4, .bit_pos = led_pos, }; -static struct resource heartbeat_resources[] = { - [0] = { - .start = 0xA405012C, /* PTG */ - .end = 0xA405012E - 1, - .flags = IORESOURCE_MEM, - }, +static struct resource heartbeat_resource = { + .start = 0xA405012C, /* PTG */ + .end = 0xA405012E - 1, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_8BIT, }; static struct platform_device heartbeat_device = { @@ -84,8 +82,8 @@ static struct platform_device heartbeat_device = { .dev = { .platform_data = &heartbeat_data, }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1, + .resource = &heartbeat_resource, }; /* MTD */ diff --git a/arch/sh/boards/mach-sdk7780/setup.c b/arch/sh/boards/mach-sdk7780/setup.c index aad94a7..1d04657 100644 --- a/arch/sh/boards/mach-sdk7780/setup.c +++ b/arch/sh/boards/mach-sdk7780/setup.c @@ -20,27 +20,18 @@ #define GPIO_PECR 0xFFEA0008 -//* Heartbeat */ -static struct heartbeat_data heartbeat_data = { - .regsize = 16, -}; - -static struct resource heartbeat_resources[] = { - [0] = { - .start = PA_LED, - .end = PA_LED, - .flags = IORESOURCE_MEM, - }, +/* Heartbeat */ +static struct resource heartbeat_resource = { + .start = PA_LED, + .end = PA_LED, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_16BIT, }; static struct platform_device heartbeat_device = { .name = "heartbeat", .id = -1, - .dev = { - .platform_data = &heartbeat_data, - }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1 + .resource = &heartbeat_resource, }; /* SMC91x */ diff --git a/arch/sh/boards/mach-se/7206/setup.c b/arch/sh/boards/mach-se/7206/setup.c index f5466384..8f5c65d 100644 --- a/arch/sh/boards/mach-se/7206/setup.c +++ b/arch/sh/boards/mach-se/7206/setup.c @@ -50,15 +50,12 @@ static unsigned char heartbeat_bit_pos[] = { 8, 9, 10, 11, 12, 13, 14, 15 }; static struct heartbeat_data heartbeat_data = { .bit_pos = heartbeat_bit_pos, .nr_bits = ARRAY_SIZE(heartbeat_bit_pos), - .regsize = 32, }; -static struct resource heartbeat_resources[] = { - [0] = { - .start = PA_LED, - .end = PA_LED, - .flags = IORESOURCE_MEM, - }, +static struct resource heartbeat_resource = { + .start = PA_LED, + .end = PA_LED, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_32BIT, }; static struct platform_device heartbeat_device = { @@ -67,8 +64,8 @@ static struct platform_device heartbeat_device = { .dev = { .platform_data = &heartbeat_data, }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1, + .resource = &heartbeat_resource, }; static struct platform_device *se7206_devices[] __initdata = { diff --git a/arch/sh/boards/mach-se/7343/setup.c b/arch/sh/boards/mach-se/7343/setup.c index 292cc47..3412bb2 100644 --- a/arch/sh/boards/mach-se/7343/setup.c +++ b/arch/sh/boards/mach-se/7343/setup.c @@ -11,26 +11,17 @@ #include #include -static struct resource heartbeat_resources[] = { - [0] = { - .start = PA_LED, - .end = PA_LED, - .flags = IORESOURCE_MEM, - }, -}; - -static struct heartbeat_data heartbeat_data = { - .regsize = 16, +static struct resource heartbeat_resource = { + .start = PA_LED, + .end = PA_LED, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_16BIT, }; static struct platform_device heartbeat_device = { .name = "heartbeat", .id = -1, - .dev = { - .platform_data = &heartbeat_data, - }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1, + .resource = &heartbeat_resource, }; static struct mtd_partition nor_flash_partitions[] = { diff --git a/arch/sh/boards/mach-se/770x/setup.c b/arch/sh/boards/mach-se/770x/setup.c index 527eb6b..66d39d1 100644 --- a/arch/sh/boards/mach-se/770x/setup.c +++ b/arch/sh/boards/mach-se/770x/setup.c @@ -93,15 +93,12 @@ static unsigned char heartbeat_bit_pos[] = { 8, 9, 10, 11, 12, 13, 14, 15 }; static struct heartbeat_data heartbeat_data = { .bit_pos = heartbeat_bit_pos, .nr_bits = ARRAY_SIZE(heartbeat_bit_pos), - .regsize = 16, }; -static struct resource heartbeat_resources[] = { - [0] = { - .start = PA_LED, - .end = PA_LED, - .flags = IORESOURCE_MEM, - }, +static struct resource heartbeat_resource = { + .start = PA_LED, + .end = PA_LED, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_16BIT, }; static struct platform_device heartbeat_device = { @@ -110,8 +107,8 @@ static struct platform_device heartbeat_device = { .dev = { .platform_data = &heartbeat_data, }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1, + .resource = &heartbeat_resource, }; #if defined(CONFIG_CPU_SUBTYPE_SH7710) ||\ diff --git a/arch/sh/boards/mach-se/7721/setup.c b/arch/sh/boards/mach-se/7721/setup.c index 55af4c3..460da53 100644 --- a/arch/sh/boards/mach-se/7721/setup.c +++ b/arch/sh/boards/mach-se/7721/setup.c @@ -23,15 +23,12 @@ static unsigned char heartbeat_bit_pos[] = { 8, 9, 10, 11, 12, 13, 14, 15 }; static struct heartbeat_data heartbeat_data = { .bit_pos = heartbeat_bit_pos, .nr_bits = ARRAY_SIZE(heartbeat_bit_pos), - .regsize = 16, }; -static struct resource heartbeat_resources[] = { - [0] = { - .start = PA_LED, - .end = PA_LED, - .flags = IORESOURCE_MEM, - }, +static struct resource heartbeat_resource = { + .start = PA_LED, + .end = PA_LED, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_16BIT, }; static struct platform_device heartbeat_device = { @@ -40,8 +37,8 @@ static struct platform_device heartbeat_device = { .dev = { .platform_data = &heartbeat_data, }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1, + .resource = &heartbeat_resource, }; static struct resource cf_ide_resources[] = { diff --git a/arch/sh/boards/mach-se/7722/setup.c b/arch/sh/boards/mach-se/7722/setup.c index b1cb942..9367541 100644 --- a/arch/sh/boards/mach-se/7722/setup.c +++ b/arch/sh/boards/mach-se/7722/setup.c @@ -25,26 +25,17 @@ #include /* Heartbeat */ -static struct heartbeat_data heartbeat_data = { - .regsize = 16, -}; - -static struct resource heartbeat_resources[] = { - [0] = { - .start = PA_LED, - .end = PA_LED, - .flags = IORESOURCE_MEM, - }, +static struct resource heartbeat_resource = { + .start = PA_LED, + .end = PA_LED, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_16BIT, }; static struct platform_device heartbeat_device = { .name = "heartbeat", .id = -1, - .dev = { - .platform_data = &heartbeat_data, - }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1, + .resource = &heartbeat_resource, }; /* SMC91x */ diff --git a/arch/sh/boards/mach-se/7724/setup.c b/arch/sh/boards/mach-se/7724/setup.c index 5d0f70b..cbfba78 100644 --- a/arch/sh/boards/mach-se/7724/setup.c +++ b/arch/sh/boards/mach-se/7724/setup.c @@ -53,26 +53,17 @@ */ /* Heartbeat */ -static struct heartbeat_data heartbeat_data = { - .regsize = 16, -}; - -static struct resource heartbeat_resources[] = { - [0] = { - .start = PA_LED, - .end = PA_LED, - .flags = IORESOURCE_MEM, - }, +static struct resource heartbeat_resource = { + .start = PA_LED, + .end = PA_LED, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_16BIT, }; static struct platform_device heartbeat_device = { .name = "heartbeat", .id = -1, - .dev = { - .platform_data = &heartbeat_data, - }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1, + .resource = &heartbeat_resource, }; /* LAN91C111 */ diff --git a/arch/sh/boards/mach-se/7780/setup.c b/arch/sh/boards/mach-se/7780/setup.c index 1d3a867..f7bfb3f 100644 --- a/arch/sh/boards/mach-se/7780/setup.c +++ b/arch/sh/boards/mach-se/7780/setup.c @@ -17,26 +17,17 @@ #include /* Heartbeat */ -static struct heartbeat_data heartbeat_data = { - .regsize = 16, -}; - -static struct resource heartbeat_resources[] = { - [0] = { - .start = PA_LED, - .end = PA_LED, - .flags = IORESOURCE_MEM, - }, +static struct resource heartbeat_resource = { + .start = PA_LED, + .end = PA_LED, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_16BIT, }; static struct platform_device heartbeat_device = { .name = "heartbeat", .id = -1, - .dev = { - .platform_data = &heartbeat_data, - }, - .num_resources = ARRAY_SIZE(heartbeat_resources), - .resource = heartbeat_resources, + .num_resources = 1, + .resource = &heartbeat_resource, }; /* SMC91x */ -- cgit v0.10.2 From a6198a238baceae9d4e0ce3915f6d239c89b5c08 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 15 Jan 2010 14:21:37 +0900 Subject: sh: Guard against early IPIs in flush_cache_all(). flush_cache_all() gets called in to when we do some early ioremapping. Unfortunately on SDK7786 the interrupt controller itself requires ioremapping, leading to a bit of a chicken and egg scenario. For now, don't bother with IPI crosscalls if there aren't any other CPUs online. Signed-off-by: Paul Mundt diff --git a/arch/sh/mm/cache.c b/arch/sh/mm/cache.c index b8607fa..0f4095d 100644 --- a/arch/sh/mm/cache.c +++ b/arch/sh/mm/cache.c @@ -2,7 +2,7 @@ * arch/sh/mm/cache.c * * Copyright (C) 1999, 2000, 2002 Niibe Yutaka - * Copyright (C) 2002 - 2009 Paul Mundt + * Copyright (C) 2002 - 2010 Paul Mundt * * Released under the terms of the GNU GPL v2.0. */ @@ -41,8 +41,17 @@ static inline void cacheop_on_each_cpu(void (*func) (void *info), void *info, int wait) { preempt_disable(); - smp_call_function(func, info, wait); + + /* + * It's possible that this gets called early on when IRQs are + * still disabled due to ioremapping by the boot CPU, so don't + * even attempt IPIs unless there are other CPUs online. + */ + if (num_online_cpus() > 1) + smp_call_function(func, info, wait); + func(info); + preempt_enable(); } -- cgit v0.10.2 From f0cb77372c3cf8c0cb17bbfb30a62506ea119286 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 15 Jan 2010 15:13:48 +0900 Subject: sh: Fix up the secondary CPU entry point for 32bit mode. Presently the secondary CPU entry point is only aimed at 29bit phys mode, causing it to point to a stray virtual address in 32bit mode. Fix it up after consulting with our shiny new __in_29bit_mode(). Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/cpu/sh4a/smp-shx3.c b/arch/sh/kernel/cpu/sh4a/smp-shx3.c index 5863e0c..11bf4c1 100644 --- a/arch/sh/kernel/cpu/sh4a/smp-shx3.c +++ b/arch/sh/kernel/cpu/sh4a/smp-shx3.c @@ -78,7 +78,10 @@ void __init plat_prepare_cpus(unsigned int max_cpus) void plat_start_cpu(unsigned int cpu, unsigned long entry_point) { - __raw_writel(entry_point, RESET_REG(cpu)); + if (__in_29bit_mode()) + __raw_writel(entry_point, RESET_REG(cpu)); + else + __raw_writel(virt_to_phys(entry_point), RESET_REG(cpu)); if (!(__raw_readl(STBCR_REG(cpu)) & STBCR_MSTP)) __raw_writel(STBCR_MSTP, STBCR_REG(cpu)); -- cgit v0.10.2 From 6f832e8aab8c4c6d28629e970e4a85e74a8d0be6 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 15 Jan 2010 16:31:04 +0900 Subject: sh: mach-sdk7786: mode pins support. This wires up the mode pins support on the SDK7786. The pins are standard SH7786 pins, and all are fixed in software. Needed for the clock framework, PCIe, and so forth. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-sdk7786/setup.c b/arch/sh/boards/mach-sdk7786/setup.c index f2d4b75..c38c6cc 100644 --- a/arch/sh/boards/mach-sdk7786/setup.c +++ b/arch/sh/boards/mach-sdk7786/setup.c @@ -218,6 +218,23 @@ static void __init init_sdk7786_IRQ(void) plat_irq_setup_pins(IRQ_MODE_IRL3210_MASK); } +#define MODSWR_REGS 0x07fff830 + +static int sdk7786_mode_pins(void) +{ + void __iomem *modswr; + int pin_states; + + modswr = ioremap_nocache(MODSWR_REGS, SZ_16); + if (!modswr) + return -ENXIO; + + pin_states = ioread16(modswr); + iounmap(modswr); + + return pin_states; +} + /* Initialize the board */ static void __init sdk7786_setup(char **cmdline_p) { @@ -230,5 +247,6 @@ static void __init sdk7786_setup(char **cmdline_p) static struct sh_machine_vector mv_sdk7786 __initmv = { .mv_name = "SDK7786", .mv_setup = sdk7786_setup, + .mv_mode_pins = sdk7786_mode_pins, .mv_init_irq = init_sdk7786_IRQ, }; -- cgit v0.10.2 From 7dcaa8e8e67b2cfbe0097c9bb52e23aed5443b8b Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 15 Jan 2010 19:13:56 +0900 Subject: sh: Generalize SH7786 PCIe support. Previously this was only built in for Urquell boards, but the same approach can be used on SDK7786 now that the mode pin reading is supported, so make it generic to SH7786. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/Kconfig b/arch/sh/boards/Kconfig index c0a1992..938e87d 100644 --- a/arch/sh/boards/Kconfig +++ b/arch/sh/boards/Kconfig @@ -153,6 +153,7 @@ config SH_SDK7780 config SH_SDK7786 bool "SDK7786" depends on CPU_SUBTYPE_SH7786 + select SYS_SUPPORTS_PCI help Select SDK7786 if configuring for a Renesas Technology Europe SH7786-65nm board. diff --git a/arch/sh/drivers/pci/Makefile b/arch/sh/drivers/pci/Makefile index 08af1f4..2c458b6 100644 --- a/arch/sh/drivers/pci/Makefile +++ b/arch/sh/drivers/pci/Makefile @@ -8,7 +8,7 @@ obj-$(CONFIG_CPU_SUBTYPE_SH7751R) += pci-sh7751.o ops-sh4.o obj-$(CONFIG_CPU_SUBTYPE_SH7763) += pci-sh7780.o ops-sh4.o obj-$(CONFIG_CPU_SUBTYPE_SH7780) += pci-sh7780.o ops-sh4.o obj-$(CONFIG_CPU_SUBTYPE_SH7785) += pci-sh7780.o ops-sh4.o -obj-$(CONFIG_CPU_SUBTYPE_SH7786) += ops-sh7786.o +obj-$(CONFIG_CPU_SUBTYPE_SH7786) += pcie-sh7786.o ops-sh7786.o obj-$(CONFIG_CPU_SH5) += pci-sh5.o ops-sh5.o obj-$(CONFIG_SH_DREAMCAST) += ops-dreamcast.o fixups-dreamcast.o \ @@ -25,4 +25,3 @@ obj-$(CONFIG_SH_TITAN) += fixups-titan.o obj-$(CONFIG_SH_LANDISK) += fixups-landisk.o obj-$(CONFIG_SH_LBOX_RE2) += fixups-rts7751r2d.o obj-$(CONFIG_SH_CAYMAN) += fixups-cayman.o -obj-$(CONFIG_SH_URQUELL) += pcie-sh7786.o -- cgit v0.10.2 From 62e62da856dba2edb897b672cbd05a69edd4485c Mon Sep 17 00:00:00 2001 From: H Hartley Sweeten Date: Thu, 14 Jan 2010 19:10:07 -0700 Subject: HID: hid-debug.c: make local symbols static hid-debug.c: make local symbols static The symbols hid_resolv_event and hid_dump_input_mapping are only used locally in this file. Make them static to prevent the following sparse warnings: warning: symbol 'hid_resolv_event' was not declared. Should it be static? warning: symbol 'hid_dump_input_mapping' was not declared. Should it be static? Signed-off-by: H Hartley Sweeten Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-debug.c b/drivers/hid/hid-debug.c index 6abd036..cd4ece6 100644 --- a/drivers/hid/hid-debug.c +++ b/drivers/hid/hid-debug.c @@ -864,13 +864,13 @@ static const char **names[EV_MAX + 1] = { [EV_SND] = sounds, [EV_REP] = repeats, }; -void hid_resolv_event(__u8 type, __u16 code, struct seq_file *f) { - +static void hid_resolv_event(__u8 type, __u16 code, struct seq_file *f) +{ seq_printf(f, "%s.%s", events[type] ? events[type] : "?", names[type] ? (names[type][code] ? names[type][code] : "?") : "?"); } -void hid_dump_input_mapping(struct hid_device *hid, struct seq_file *f) +static void hid_dump_input_mapping(struct hid_device *hid, struct seq_file *f) { int i, j, k; struct hid_report *report; -- cgit v0.10.2 From 3a85cd96d3ab3c6dcf88b81fc6eaddb84e565a43 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 14 Jan 2010 01:33:55 +0000 Subject: xfs: add tracing to xfs_swap_extents To be able to diagnose whether the swap extents function is detecting compatible inode data fork configurations for swapping extents, add tracing points to the code to allow us to see the format of the inode forks before and after the swap. Signed-off-by: Dave Chinner Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_trace.h b/fs/xfs/linux-2.6/xfs_trace.h index c22a608..3353aef 100644 --- a/fs/xfs/linux-2.6/xfs_trace.h +++ b/fs/xfs/linux-2.6/xfs_trace.h @@ -1414,6 +1414,59 @@ TRACE_EVENT(xfs_dir2_leafn_moveents, __entry->count) ); +#define XFS_SWAPEXT_INODES \ + { 0, "target" }, \ + { 1, "temp" } + +#define XFS_INODE_FORMAT_STR \ + { 0, "invalid" }, \ + { 1, "local" }, \ + { 2, "extent" }, \ + { 3, "btree" } + +DECLARE_EVENT_CLASS(xfs_swap_extent_class, + TP_PROTO(struct xfs_inode *ip, int which), + TP_ARGS(ip, which), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(int, which) + __field(xfs_ino_t, ino) + __field(int, format) + __field(int, nex) + __field(int, max_nex) + __field(int, broot_size) + __field(int, fork_off) + ), + TP_fast_assign( + __entry->dev = VFS_I(ip)->i_sb->s_dev; + __entry->which = which; + __entry->ino = ip->i_ino; + __entry->format = ip->i_d.di_format; + __entry->nex = ip->i_d.di_nextents; + __entry->max_nex = ip->i_df.if_ext_max; + __entry->broot_size = ip->i_df.if_broot_bytes; + __entry->fork_off = XFS_IFORK_BOFF(ip); + ), + TP_printk("dev %d:%d ino 0x%llx (%s), %s format, num_extents %d, " + "Max in-fork extents %d, broot size %d, fork offset %d", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __print_symbolic(__entry->which, XFS_SWAPEXT_INODES), + __print_symbolic(__entry->format, XFS_INODE_FORMAT_STR), + __entry->nex, + __entry->max_nex, + __entry->broot_size, + __entry->fork_off) +) + +#define DEFINE_SWAPEXT_EVENT(name) \ +DEFINE_EVENT(xfs_swap_extent_class, name, \ + TP_PROTO(struct xfs_inode *ip, int which), \ + TP_ARGS(ip, which)) + +DEFINE_SWAPEXT_EVENT(xfs_swap_extent_before); +DEFINE_SWAPEXT_EVENT(xfs_swap_extent_after); + #endif /* _TRACE_XFS_H */ #undef TRACE_INCLUDE_PATH diff --git a/fs/xfs/xfs_dfrag.c b/fs/xfs/xfs_dfrag.c index 84ca1cf..f25e540 100644 --- a/fs/xfs/xfs_dfrag.c +++ b/fs/xfs/xfs_dfrag.c @@ -254,6 +254,9 @@ xfs_swap_extents( goto out_unlock; } + trace_xfs_swap_extent_before(ip, 0); + trace_xfs_swap_extent_before(tip, 1); + /* check inode formats now that data is flushed */ error = xfs_swap_extents_check_format(ip, tip); if (error) { @@ -421,6 +424,8 @@ xfs_swap_extents( error = xfs_trans_commit(tp, XFS_TRANS_SWAPEXT); + trace_xfs_swap_extent_after(ip, 0); + trace_xfs_swap_extent_after(tip, 1); out: kmem_free(tempifp); return error; -- cgit v0.10.2 From 6bded0f383fd7971b76ad6c194dda7d5b814b871 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 14 Jan 2010 01:33:56 +0000 Subject: xfs: clean up inconsistent variable naming in xfs_swap_extent The swap extent ioctl passes in a target inode and a temporary inode which are clearly named in the ioctl structure. The code then assigns temp to target and vice versa, making it extremely difficult to work out which inode is which later in the code. Make this consistent throughout the code. Also make xfs_swap_extent static as there are no external users of the function. Signed-off-by: Dave Chinner Signed-off-by: Alex Elder diff --git a/fs/xfs/xfs_dfrag.c b/fs/xfs/xfs_dfrag.c index f25e540..cd27c9d 100644 --- a/fs/xfs/xfs_dfrag.c +++ b/fs/xfs/xfs_dfrag.c @@ -45,15 +45,21 @@ #include "xfs_vnodeops.h" #include "xfs_trace.h" + +static int xfs_swap_extents( + xfs_inode_t *ip, /* target inode */ + xfs_inode_t *tip, /* tmp inode */ + xfs_swapext_t *sxp); + /* - * Syssgi interface for swapext + * ioctl interface for swapext */ int xfs_swapext( xfs_swapext_t *sxp) { xfs_inode_t *ip, *tip; - struct file *file, *target_file; + struct file *file, *tmp_file; int error = 0; /* Pull information for the target fd */ @@ -68,46 +74,46 @@ xfs_swapext( goto out_put_file; } - target_file = fget((int)sxp->sx_fdtmp); - if (!target_file) { + tmp_file = fget((int)sxp->sx_fdtmp); + if (!tmp_file) { error = XFS_ERROR(EINVAL); goto out_put_file; } - if (!(target_file->f_mode & FMODE_WRITE) || - (target_file->f_flags & O_APPEND)) { + if (!(tmp_file->f_mode & FMODE_WRITE) || + (tmp_file->f_flags & O_APPEND)) { error = XFS_ERROR(EBADF); - goto out_put_target_file; + goto out_put_tmp_file; } if (IS_SWAPFILE(file->f_path.dentry->d_inode) || - IS_SWAPFILE(target_file->f_path.dentry->d_inode)) { + IS_SWAPFILE(tmp_file->f_path.dentry->d_inode)) { error = XFS_ERROR(EINVAL); - goto out_put_target_file; + goto out_put_tmp_file; } ip = XFS_I(file->f_path.dentry->d_inode); - tip = XFS_I(target_file->f_path.dentry->d_inode); + tip = XFS_I(tmp_file->f_path.dentry->d_inode); if (ip->i_mount != tip->i_mount) { error = XFS_ERROR(EINVAL); - goto out_put_target_file; + goto out_put_tmp_file; } if (ip->i_ino == tip->i_ino) { error = XFS_ERROR(EINVAL); - goto out_put_target_file; + goto out_put_tmp_file; } if (XFS_FORCED_SHUTDOWN(ip->i_mount)) { error = XFS_ERROR(EIO); - goto out_put_target_file; + goto out_put_tmp_file; } error = xfs_swap_extents(ip, tip, sxp); - out_put_target_file: - fput(target_file); + out_put_tmp_file: + fput(tmp_file); out_put_file: fput(file); out: @@ -186,7 +192,7 @@ xfs_swap_extents_check_format( return 0; } -int +static int xfs_swap_extents( xfs_inode_t *ip, /* target inode */ xfs_inode_t *tip, /* tmp inode */ diff --git a/fs/xfs/xfs_dfrag.h b/fs/xfs/xfs_dfrag.h index 4f55a63..20bdd93 100644 --- a/fs/xfs/xfs_dfrag.h +++ b/fs/xfs/xfs_dfrag.h @@ -48,9 +48,6 @@ typedef struct xfs_swapext */ int xfs_swapext(struct xfs_swapext *sx); -int xfs_swap_extents(struct xfs_inode *ip, struct xfs_inode *tip, - struct xfs_swapext *sxp); - #endif /* __KERNEL__ */ #endif /* __XFS_DFRAG_H__ */ -- cgit v0.10.2 From 5d77c0dc0c05c2c65aee16149fae06831a118730 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 19 Nov 2009 15:52:00 +0000 Subject: xfs: make several more functions static Just minor housekeeping, a lot more functions can be trivially made static; others could if we reordered things a bit... Signed-off-by: Eric Sandeen Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c index 1f5e4bb..0f90bfe 100644 --- a/fs/xfs/linux-2.6/xfs_sync.c +++ b/fs/xfs/linux-2.6/xfs_sync.c @@ -351,7 +351,7 @@ xfs_commit_dummy_trans( return error; } -int +STATIC int xfs_sync_fsdata( struct xfs_mount *mp, int flags) diff --git a/fs/xfs/linux-2.6/xfs_sync.h b/fs/xfs/linux-2.6/xfs_sync.h index ea932b4..d480c34 100644 --- a/fs/xfs/linux-2.6/xfs_sync.h +++ b/fs/xfs/linux-2.6/xfs_sync.h @@ -37,7 +37,6 @@ void xfs_syncd_stop(struct xfs_mount *mp); int xfs_sync_attr(struct xfs_mount *mp, int flags); int xfs_sync_data(struct xfs_mount *mp, int flags); -int xfs_sync_fsdata(struct xfs_mount *mp, int flags); int xfs_quiesce_data(struct xfs_mount *mp); void xfs_quiesce_attr(struct xfs_mount *mp); diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index e953b6cf..9d11eba 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -197,7 +197,7 @@ xfs_attr_get( /* * Calculate how many blocks we need for the new attribute, */ -int +STATIC int xfs_attr_calc_size( struct xfs_inode *ip, int namelen, diff --git a/fs/xfs/xfs_attr.h b/fs/xfs/xfs_attr.h index 59b410c..9c3a243 100644 --- a/fs/xfs/xfs_attr.h +++ b/fs/xfs/xfs_attr.h @@ -139,7 +139,6 @@ typedef struct xfs_attr_list_context { /* * Overall external interface routines. */ -int xfs_attr_calc_size(struct xfs_inode *, int, int, int *); int xfs_attr_inactive(struct xfs_inode *dp); int xfs_attr_rmtval_get(struct xfs_da_args *args); int xfs_attr_list_int(struct xfs_attr_list_context *); diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c index 38751d5..416e47e 100644 --- a/fs/xfs/xfs_bmap_btree.c +++ b/fs/xfs/xfs_bmap_btree.c @@ -334,7 +334,7 @@ xfs_bmbt_disk_set_allf( /* * Set all the fields in a bmap extent record from the uncompressed form. */ -void +STATIC void xfs_bmbt_disk_set_all( xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s) diff --git a/fs/xfs/xfs_bmap_btree.h b/fs/xfs/xfs_bmap_btree.h index cf07ca7..0e66c4e 100644 --- a/fs/xfs/xfs_bmap_btree.h +++ b/fs/xfs/xfs_bmap_btree.h @@ -223,7 +223,6 @@ extern void xfs_bmbt_set_startblock(xfs_bmbt_rec_host_t *r, xfs_fsblock_t v); extern void xfs_bmbt_set_startoff(xfs_bmbt_rec_host_t *r, xfs_fileoff_t v); extern void xfs_bmbt_set_state(xfs_bmbt_rec_host_t *r, xfs_exntst_t v); -extern void xfs_bmbt_disk_set_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s); extern void xfs_bmbt_disk_set_allf(xfs_bmbt_rec_t *r, xfs_fileoff_t o, xfs_fsblock_t b, xfs_filblks_t c, xfs_exntst_t v); diff --git a/fs/xfs/xfs_dir2_node.c b/fs/xfs/xfs_dir2_node.c index ce6e355..78fc4d9 100644 --- a/fs/xfs/xfs_dir2_node.c +++ b/fs/xfs/xfs_dir2_node.c @@ -65,7 +65,7 @@ static int xfs_dir2_node_addname_int(xfs_da_args_t *args, /* * Log entries from a freespace block. */ -void +STATIC void xfs_dir2_free_log_bests( xfs_trans_t *tp, /* transaction pointer */ xfs_dabuf_t *bp, /* freespace buffer */ diff --git a/fs/xfs/xfs_dir2_node.h b/fs/xfs/xfs_dir2_node.h index dde72db..82dfe71 100644 --- a/fs/xfs/xfs_dir2_node.h +++ b/fs/xfs/xfs_dir2_node.h @@ -75,8 +75,6 @@ xfs_dir2_db_to_fdindex(struct xfs_mount *mp, xfs_dir2_db_t db) return ((db) % XFS_DIR2_MAX_FREE_BESTS(mp)); } -extern void xfs_dir2_free_log_bests(struct xfs_trans *tp, struct xfs_dabuf *bp, - int first, int last); extern int xfs_dir2_leaf_to_node(struct xfs_da_args *args, struct xfs_dabuf *lbp); extern xfs_dahash_t xfs_dir2_leafn_lasthash(struct xfs_dabuf *bp, int *count); diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index d55662d..fd02a18 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -443,14 +443,9 @@ typedef struct log { /* common routines */ extern xfs_lsn_t xlog_assign_tail_lsn(struct xfs_mount *mp); -extern int xlog_find_tail(xlog_t *log, - xfs_daddr_t *head_blk, - xfs_daddr_t *tail_blk); extern int xlog_recover(xlog_t *log); extern int xlog_recover_finish(xlog_t *log); extern void xlog_pack_data(xlog_t *log, xlog_in_core_t *iclog, int); -extern struct xfs_buf *xlog_get_bp(xlog_t *, int); -extern void xlog_put_bp(struct xfs_buf *); extern kmem_zone_t *xfs_log_ticket_zone; diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 69ac2e5..48a7ab1 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -68,7 +68,7 @@ STATIC void xlog_recover_check_summary(xlog_t *); ((bbs + (log)->l_sectbb_mask + 1) & ~(log)->l_sectbb_mask) : (bbs) ) #define XLOG_SECTOR_ROUNDDOWN_BLKNO(log, bno) ((bno) & ~(log)->l_sectbb_mask) -xfs_buf_t * +STATIC xfs_buf_t * xlog_get_bp( xlog_t *log, int nbblks) @@ -88,7 +88,7 @@ xlog_get_bp( return xfs_buf_get_noaddr(BBTOB(nbblks), log->l_mp->m_logdev_targp); } -void +STATIC void xlog_put_bp( xfs_buf_t *bp) { @@ -805,7 +805,7 @@ xlog_find_head( * We could speed up search by using current head_blk buffer, but it is not * available. */ -int +STATIC int xlog_find_tail( xlog_t *log, xfs_daddr_t *head_blk, -- cgit v0.10.2 From f0a7695380efa31cd281730917f7e907a724d5cb Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:49:57 +0000 Subject: xfs: Use list_heads for log recovery item lists Remove the roll-your-own linked list operations. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 48a7ab1..65f1f13 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -50,8 +50,6 @@ STATIC int xlog_find_zeroed(xlog_t *, xfs_daddr_t *); STATIC int xlog_clear_stale_blocks(xlog_t *, xfs_lsn_t); -STATIC void xlog_recover_insert_item_backq(xlog_recover_item_t **q, - xlog_recover_item_t *item); #if defined(DEBUG) STATIC void xlog_recover_check_summary(xlog_t *); #else @@ -1367,36 +1365,45 @@ xlog_clear_stale_blocks( STATIC xlog_recover_t * xlog_recover_find_tid( - xlog_recover_t *q, + struct hlist_head *head, xlog_tid_t tid) { - xlog_recover_t *p = q; + xlog_recover_t *trans; + struct hlist_node *n; - while (p != NULL) { - if (p->r_log_tid == tid) - break; - p = p->r_next; + hlist_for_each_entry(trans, n, head, r_list) { + if (trans->r_log_tid == tid) + return trans; } - return p; + return NULL; } STATIC void -xlog_recover_put_hashq( - xlog_recover_t **q, - xlog_recover_t *trans) +xlog_recover_new_tid( + struct hlist_head *head, + xlog_tid_t tid, + xfs_lsn_t lsn) { - trans->r_next = *q; - *q = trans; + xlog_recover_t *trans; + + trans = kmem_zalloc(sizeof(xlog_recover_t), KM_SLEEP); + trans->r_log_tid = tid; + trans->r_lsn = lsn; + INIT_LIST_HEAD(&trans->r_itemq); + + INIT_HLIST_NODE(&trans->r_list); + hlist_add_head(&trans->r_list, head); } STATIC void xlog_recover_add_item( - xlog_recover_item_t **itemq) + struct list_head *head) { xlog_recover_item_t *item; item = kmem_zalloc(sizeof(xlog_recover_item_t), KM_SLEEP); - xlog_recover_insert_item_backq(itemq, item); + INIT_LIST_HEAD(&item->ri_list); + list_add_tail(&item->ri_list, head); } STATIC int @@ -1409,8 +1416,7 @@ xlog_recover_add_to_cont_trans( xfs_caddr_t ptr, old_ptr; int old_len; - item = trans->r_itemq; - if (item == NULL) { + if (list_empty(&trans->r_itemq)) { /* finish copying rest of trans header */ xlog_recover_add_item(&trans->r_itemq); ptr = (xfs_caddr_t) &trans->r_theader + @@ -1418,7 +1424,8 @@ xlog_recover_add_to_cont_trans( memcpy(ptr, dp, len); /* d, s, l */ return 0; } - item = item->ri_prev; + /* take the tail entry */ + item = list_entry(trans->r_itemq.prev, xlog_recover_item_t, ri_list); old_ptr = item->ri_buf[item->ri_cnt-1].i_addr; old_len = item->ri_buf[item->ri_cnt-1].i_len; @@ -1455,8 +1462,7 @@ xlog_recover_add_to_trans( if (!len) return 0; - item = trans->r_itemq; - if (item == NULL) { + if (list_empty(&trans->r_itemq)) { /* we need to catch log corruptions here */ if (*(uint *)dp != XFS_TRANS_HEADER_MAGIC) { xlog_warn("XFS: xlog_recover_add_to_trans: " @@ -1474,12 +1480,15 @@ xlog_recover_add_to_trans( memcpy(ptr, dp, len); in_f = (xfs_inode_log_format_t *)ptr; - if (item->ri_prev->ri_total != 0 && - item->ri_prev->ri_total == item->ri_prev->ri_cnt) { + /* take the tail entry */ + item = list_entry(trans->r_itemq.prev, xlog_recover_item_t, ri_list); + if (item->ri_total != 0 && + item->ri_total == item->ri_cnt) { + /* tail item is in use, get a new one */ xlog_recover_add_item(&trans->r_itemq); + item = list_entry(trans->r_itemq.prev, + xlog_recover_item_t, ri_list); } - item = trans->r_itemq; - item = item->ri_prev; if (item->ri_total == 0) { /* first region to be added */ if (in_f->ilf_size == 0 || @@ -1504,96 +1513,29 @@ xlog_recover_add_to_trans( return 0; } -STATIC void -xlog_recover_new_tid( - xlog_recover_t **q, - xlog_tid_t tid, - xfs_lsn_t lsn) -{ - xlog_recover_t *trans; - - trans = kmem_zalloc(sizeof(xlog_recover_t), KM_SLEEP); - trans->r_log_tid = tid; - trans->r_lsn = lsn; - xlog_recover_put_hashq(q, trans); -} - -STATIC int -xlog_recover_unlink_tid( - xlog_recover_t **q, - xlog_recover_t *trans) -{ - xlog_recover_t *tp; - int found = 0; - - ASSERT(trans != NULL); - if (trans == *q) { - *q = (*q)->r_next; - } else { - tp = *q; - while (tp) { - if (tp->r_next == trans) { - found = 1; - break; - } - tp = tp->r_next; - } - if (!found) { - xlog_warn( - "XFS: xlog_recover_unlink_tid: trans not found"); - ASSERT(0); - return XFS_ERROR(EIO); - } - tp->r_next = tp->r_next->r_next; - } - return 0; -} - -STATIC void -xlog_recover_insert_item_backq( - xlog_recover_item_t **q, - xlog_recover_item_t *item) -{ - if (*q == NULL) { - item->ri_prev = item->ri_next = item; - *q = item; - } else { - item->ri_next = *q; - item->ri_prev = (*q)->ri_prev; - (*q)->ri_prev = item; - item->ri_prev->ri_next = item; - } -} - -STATIC void -xlog_recover_insert_item_frontq( - xlog_recover_item_t **q, - xlog_recover_item_t *item) -{ - xlog_recover_insert_item_backq(q, item); - *q = item; -} - +/* + * Sort the log items in the transaction. Cancelled buffers need + * to be put first so they are processed before any items that might + * modify the buffers. If they are cancelled, then the modifications + * don't need to be replayed. + */ STATIC int xlog_recover_reorder_trans( xlog_recover_t *trans) { - xlog_recover_item_t *first_item, *itemq, *itemq_next; - xfs_buf_log_format_t *buf_f; - ushort flags = 0; + xlog_recover_item_t *item, *n; + LIST_HEAD(sort_list); + + list_splice_init(&trans->r_itemq, &sort_list); + list_for_each_entry_safe(item, n, &sort_list, ri_list) { + xfs_buf_log_format_t *buf_f; - first_item = itemq = trans->r_itemq; - trans->r_itemq = NULL; - do { - itemq_next = itemq->ri_next; - buf_f = (xfs_buf_log_format_t *)itemq->ri_buf[0].i_addr; + buf_f = (xfs_buf_log_format_t *)item->ri_buf[0].i_addr; - switch (ITEM_TYPE(itemq)) { + switch (ITEM_TYPE(item)) { case XFS_LI_BUF: - flags = buf_f->blf_flags; - if (!(flags & XFS_BLI_CANCEL)) { - xlog_recover_insert_item_frontq(&trans->r_itemq, - itemq); + if (!(buf_f->blf_flags & XFS_BLI_CANCEL)) { + list_move(&item->ri_list, &trans->r_itemq); break; } case XFS_LI_INODE: @@ -1601,7 +1543,7 @@ xlog_recover_reorder_trans( case XFS_LI_QUOTAOFF: case XFS_LI_EFD: case XFS_LI_EFI: - xlog_recover_insert_item_backq(&trans->r_itemq, itemq); + list_move_tail(&item->ri_list, &trans->r_itemq); break; default: xlog_warn( @@ -1609,8 +1551,8 @@ xlog_recover_reorder_trans( ASSERT(0); return XFS_ERROR(EIO); } - itemq = itemq_next; - } while (first_item != itemq); + } + ASSERT(list_empty(&sort_list)); return 0; } @@ -2814,14 +2756,13 @@ xlog_recover_do_trans( int pass) { int error = 0; - xlog_recover_item_t *item, *first_item; + xlog_recover_item_t *item; error = xlog_recover_reorder_trans(trans); if (error) return error; - first_item = item = trans->r_itemq; - do { + list_for_each_entry(item, &trans->r_itemq, ri_list) { switch (ITEM_TYPE(item)) { case XFS_LI_BUF: error = xlog_recover_do_buffer_trans(log, item, pass); @@ -2854,8 +2795,7 @@ xlog_recover_do_trans( if (error) return error; - item = item->ri_next; - } while (first_item != item); + } return 0; } @@ -2869,21 +2809,18 @@ STATIC void xlog_recover_free_trans( xlog_recover_t *trans) { - xlog_recover_item_t *first_item, *item, *free_item; + xlog_recover_item_t *item, *n; int i; - item = first_item = trans->r_itemq; - do { - free_item = item; - item = item->ri_next; - /* Free the regions in the item. */ - for (i = 0; i < free_item->ri_cnt; i++) { - kmem_free(free_item->ri_buf[i].i_addr); - } + list_for_each_entry_safe(item, n, &trans->r_itemq, ri_list) { + /* Free the regions in the item. */ + list_del(&item->ri_list); + for (i = 0; i < item->ri_cnt; i++) + kmem_free(item->ri_buf[i].i_addr); /* Free the item itself */ - kmem_free(free_item->ri_buf); - kmem_free(free_item); - } while (first_item != item); + kmem_free(item->ri_buf); + kmem_free(item); + } /* Free the transaction recover structure */ kmem_free(trans); } @@ -2891,14 +2828,12 @@ xlog_recover_free_trans( STATIC int xlog_recover_commit_trans( xlog_t *log, - xlog_recover_t **q, xlog_recover_t *trans, int pass) { int error; - if ((error = xlog_recover_unlink_tid(q, trans))) - return error; + hlist_del(&trans->r_list); if ((error = xlog_recover_do_trans(log, trans, pass))) return error; xlog_recover_free_trans(trans); /* no error */ @@ -2926,7 +2861,7 @@ xlog_recover_unmount_trans( STATIC int xlog_recover_process_data( xlog_t *log, - xlog_recover_t *rhash[], + struct hlist_head rhash[], xlog_rec_header_t *rhead, xfs_caddr_t dp, int pass) @@ -2960,7 +2895,7 @@ xlog_recover_process_data( } tid = be32_to_cpu(ohead->oh_tid); hash = XLOG_RHASH(tid); - trans = xlog_recover_find_tid(rhash[hash], tid); + trans = xlog_recover_find_tid(&rhash[hash], tid); if (trans == NULL) { /* not found; add new tid */ if (ohead->oh_flags & XLOG_START_TRANS) xlog_recover_new_tid(&rhash[hash], tid, @@ -2978,7 +2913,7 @@ xlog_recover_process_data( switch (flags) { case XLOG_COMMIT_TRANS: error = xlog_recover_commit_trans(log, - &rhash[hash], trans, pass); + trans, pass); break; case XLOG_UNMOUNT_TRANS: error = xlog_recover_unmount_trans(trans); @@ -3517,7 +3452,7 @@ xlog_do_recovery_pass( int error = 0, h_size; int bblks, split_bblks; int hblks, split_hblks, wrapped_hblks; - xlog_recover_t *rhash[XLOG_RHASH_SIZE]; + struct hlist_head rhash[XLOG_RHASH_SIZE]; ASSERT(head_blk != tail_blk); diff --git a/fs/xfs/xfs_log_recover.h b/fs/xfs/xfs_log_recover.h index b225455..75d7492 100644 --- a/fs/xfs/xfs_log_recover.h +++ b/fs/xfs/xfs_log_recover.h @@ -35,22 +35,21 @@ * item headers are in ri_buf[0]. Additional buffers follow. */ typedef struct xlog_recover_item { - struct xlog_recover_item *ri_next; - struct xlog_recover_item *ri_prev; - int ri_type; - int ri_cnt; /* count of regions found */ - int ri_total; /* total regions */ - xfs_log_iovec_t *ri_buf; /* ptr to regions buffer */ + struct list_head ri_list; + int ri_type; + int ri_cnt; /* count of regions found */ + int ri_total; /* total regions */ + xfs_log_iovec_t *ri_buf; /* ptr to regions buffer */ } xlog_recover_item_t; struct xlog_tid; typedef struct xlog_recover { - struct xlog_recover *r_next; - xlog_tid_t r_log_tid; /* log's transaction id */ - xfs_trans_header_t r_theader; /* trans header for partial */ - int r_state; /* not needed */ - xfs_lsn_t r_lsn; /* xact lsn */ - xlog_recover_item_t *r_itemq; /* q for items */ + struct hlist_node r_list; + xlog_tid_t r_log_tid; /* log's transaction id */ + xfs_trans_header_t r_theader; /* trans header for partial */ + int r_state; /* not needed */ + xfs_lsn_t r_lsn; /* xact lsn */ + struct list_head r_itemq; /* q for items */ } xlog_recover_t; #define ITEM_TYPE(i) (*(ushort *)(i)->ri_buf[0].i_addr) -- cgit v0.10.2 From 453eac8a9aa417878a38bdfbccafd5f7ce4e8e4e Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:49:58 +0000 Subject: xfs: Don't wake the aild once per second Now that the AIL push algorithm is traversal safe, we don't need a watchdog function in the xfsaild to catch pushes that fail to make progress. Remove the watchdog timeout and make pushes purely driven by demand. This will remove the once-per-second wakeup that is seen when the filesystem is idle and make laptop power misers happy. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index 77414db..9f2e398 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -877,12 +877,11 @@ xfsaild( { struct xfs_ail *ailp = data; xfs_lsn_t last_pushed_lsn = 0; - long tout = 0; + long tout = 0; /* milliseconds */ while (!kthread_should_stop()) { - if (tout) - schedule_timeout_interruptible(msecs_to_jiffies(tout)); - tout = 1000; + schedule_timeout_interruptible(tout ? + msecs_to_jiffies(tout) : MAX_SCHEDULE_TIMEOUT); /* swsusp */ try_to_freeze(); diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c index 2ffc570..063dfbd 100644 --- a/fs/xfs/xfs_trans_ail.c +++ b/fs/xfs/xfs_trans_ail.c @@ -237,14 +237,15 @@ out: } /* - * Function that does the work of pushing on the AIL + * xfsaild_push does the work of pushing on the AIL. Returning a timeout of + * zero indicates that the caller should sleep until woken. */ long xfsaild_push( struct xfs_ail *ailp, xfs_lsn_t *last_lsn) { - long tout = 1000; /* milliseconds */ + long tout = 0; xfs_lsn_t last_pushed_lsn = *last_lsn; xfs_lsn_t target = ailp->xa_target; xfs_lsn_t lsn; @@ -262,7 +263,7 @@ xfsaild_push( */ xfs_trans_ail_cursor_done(ailp, cur); spin_unlock(&ailp->xa_lock); - last_pushed_lsn = 0; + *last_lsn = 0; return tout; } @@ -279,7 +280,6 @@ xfsaild_push( * prevents use from spinning when we can't do anything or there is * lots of contention on the AIL lists. */ - tout = 10; lsn = lip->li_lsn; flush_log = stuck = count = 0; while ((XFS_LSN_CMP(lip->li_lsn, target) < 0)) { @@ -376,14 +376,14 @@ xfsaild_push( if (!count) { /* We're past our target or empty, so idle */ - tout = 1000; + last_pushed_lsn = 0; } else if (XFS_LSN_CMP(lsn, target) >= 0) { /* * We reached the target so wait a bit longer for I/O to * complete and remove pushed items from the AIL before we * start the next scan from the start of the AIL. */ - tout += 20; + tout = 50; last_pushed_lsn = 0; } else if ((stuck * 100) / count > 90) { /* @@ -395,11 +395,14 @@ xfsaild_push( * Backoff a bit more to allow some I/O to complete before * continuing from where we were. */ - tout += 10; + tout = 20; + } else { + /* more to do, but wait a short while before continuing */ + tout = 10; } *last_lsn = last_pushed_lsn; return tout; -} /* xfsaild_push */ +} /* -- cgit v0.10.2 From c9c129714e71c890bed1bd5b61697a896c3c2d54 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:49:59 +0000 Subject: xfs: Don't wake xfsbufd when idle The xfsbufd wakes every xfsbufd_centisecs (once per second by default) for each filesystem even when the filesystem is idle. If the xfsbufd has nothing to do, put it into a long term sleep and only wake it up when there is work pending (i.e. dirty buffers to flush soon). This will make laptop power misers happy. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index 77b8be8..18ae3ba 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -1595,6 +1595,11 @@ xfs_buf_delwri_queue( list_del(&bp->b_list); } + if (list_empty(dwq)) { + /* start xfsbufd as it is about to have something to do */ + wake_up_process(bp->b_target->bt_task); + } + bp->b_flags |= _XBF_DELWRI_Q; list_add_tail(&bp->b_list, dwq); bp->b_queuetime = jiffies; @@ -1644,6 +1649,8 @@ xfsbufd_wakeup( list_for_each_entry(btp, &xfs_buftarg_list, bt_list) { if (test_bit(XBT_FORCE_SLEEP, &btp->bt_flags)) continue; + if (list_empty(&btp->bt_delwrite_queue)) + continue; set_bit(XBT_FORCE_FLUSH, &btp->bt_flags); wake_up_process(btp->bt_task); } @@ -1708,6 +1715,9 @@ xfsbufd( set_freezable(); do { + long age = xfs_buf_age_centisecs * msecs_to_jiffies(10); + long tout = xfs_buf_timer_centisecs * msecs_to_jiffies(10); + if (unlikely(freezing(current))) { set_bit(XBT_FORCE_SLEEP, &target->bt_flags); refrigerator(); @@ -1715,12 +1725,12 @@ xfsbufd( clear_bit(XBT_FORCE_SLEEP, &target->bt_flags); } - schedule_timeout_interruptible( - xfs_buf_timer_centisecs * msecs_to_jiffies(10)); - - xfs_buf_delwri_split(target, &tmp, - xfs_buf_age_centisecs * msecs_to_jiffies(10)); + /* sleep for a long time if there is nothing to do. */ + if (list_empty(&target->bt_delwrite_queue)) + tout = MAX_SCHEDULE_TIMEOUT; + schedule_timeout_interruptible(tout); + xfs_buf_delwri_split(target, &tmp, age); count = 0; while (!list_empty(&tmp)) { bp = list_entry(tmp.next, xfs_buf_t, b_list); -- cgit v0.10.2 From 5017e97d52628fb8ae56e434e86ac2e72ddaac2b Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:47:40 +0000 Subject: xfs: rename xfs_get_perag xfs_get_perag is really getting the perag that an inode belongs to based on it's inode number. Convert the use of this function to just get the perag from a provided ag number. Use this new function to obtain the per-ag structure when traversing the per AG inode trees for sync and reclaim. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c index 0f90bfe..cc964fa 100644 --- a/fs/xfs/linux-2.6/xfs_sync.c +++ b/fs/xfs/linux-2.6/xfs_sync.c @@ -90,14 +90,13 @@ xfs_inode_ag_lookup( STATIC int xfs_inode_ag_walk( struct xfs_mount *mp, - xfs_agnumber_t ag, + struct xfs_perag *pag, int (*execute)(struct xfs_inode *ip, struct xfs_perag *pag, int flags), int flags, int tag, int exclusive) { - struct xfs_perag *pag = &mp->m_perag[ag]; uint32_t first_index; int last_error = 0; int skipped; @@ -141,8 +140,6 @@ restart: delay(1); goto restart; } - - xfs_put_perag(mp, pag); return last_error; } @@ -160,10 +157,16 @@ xfs_inode_ag_iterator( xfs_agnumber_t ag; for (ag = 0; ag < mp->m_sb.sb_agcount; ag++) { - if (!mp->m_perag[ag].pag_ici_init) + struct xfs_perag *pag; + + pag = xfs_perag_get(mp, ag); + if (!pag->pag_ici_init) { + xfs_perag_put(pag); continue; - error = xfs_inode_ag_walk(mp, ag, execute, flags, tag, + } + error = xfs_inode_ag_walk(mp, pag, execute, flags, tag, exclusive); + xfs_perag_put(pag); if (error) { last_error = error; if (error == EFSCORRUPTED) @@ -690,16 +693,17 @@ void xfs_inode_set_reclaim_tag( xfs_inode_t *ip) { - xfs_mount_t *mp = ip->i_mount; - xfs_perag_t *pag = xfs_get_perag(mp, ip->i_ino); + struct xfs_mount *mp = ip->i_mount; + struct xfs_perag *pag; + pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); read_lock(&pag->pag_ici_lock); spin_lock(&ip->i_flags_lock); __xfs_inode_set_reclaim_tag(pag, ip); __xfs_iflags_set(ip, XFS_IRECLAIMABLE); spin_unlock(&ip->i_flags_lock); read_unlock(&pag->pag_ici_lock); - xfs_put_perag(mp, pag); + xfs_perag_put(pag); } void diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c index 155e798..e281eb4 100644 --- a/fs/xfs/xfs_iget.c +++ b/fs/xfs/xfs_iget.c @@ -374,7 +374,7 @@ xfs_iget( return EINVAL; /* get the perag structure and ensure that it's inode capable */ - pag = xfs_get_perag(mp, ino); + pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ino)); if (!pag->pagi_inodeok) return EINVAL; ASSERT(pag->pag_ici_init); @@ -398,7 +398,7 @@ again: if (error) goto out_error_or_again; } - xfs_put_perag(mp, pag); + xfs_perag_put(pag); *ipp = ip; @@ -417,7 +417,7 @@ out_error_or_again: delay(1); goto again; } - xfs_put_perag(mp, pag); + xfs_perag_put(pag); return error; } @@ -488,12 +488,12 @@ xfs_ireclaim( * added to the tree assert that it's been there before to catch * problems with the inode life time early on. */ - pag = xfs_get_perag(mp, ip->i_ino); + pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); write_lock(&pag->pag_ici_lock); if (!radix_tree_delete(&pag->pag_ici_root, agino)) ASSERT(0); write_unlock(&pag->pag_ici_lock); - xfs_put_perag(mp, pag); + xfs_perag_put(pag); /* * Here we do an (almost) spurious inode lock in order to coordinate diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index ef77fd8..bd3d816 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1946,8 +1946,9 @@ xfs_ifree_cluster( xfs_inode_t *ip, **ip_found; xfs_inode_log_item_t *iip; xfs_log_item_t *lip; - xfs_perag_t *pag = xfs_get_perag(mp, inum); + struct xfs_perag *pag; + pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, inum)); if (mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp)) { blks_per_cluster = 1; ninodes = mp->m_sb.sb_inopblock; @@ -2088,7 +2089,7 @@ xfs_ifree_cluster( } kmem_free(ip_found); - xfs_put_perag(mp, pag); + xfs_perag_put(pag); } /* @@ -2675,7 +2676,7 @@ xfs_iflush_cluster( xfs_buf_t *bp) { xfs_mount_t *mp = ip->i_mount; - xfs_perag_t *pag = xfs_get_perag(mp, ip->i_ino); + struct xfs_perag *pag; unsigned long first_index, mask; unsigned long inodes_per_cluster; int ilist_size; @@ -2686,6 +2687,7 @@ xfs_iflush_cluster( int bufwasdelwri; int i; + pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); ASSERT(pag->pagi_inodeok); ASSERT(pag->pag_ici_init); diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 1df7e45..f8a68a2 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -386,14 +386,14 @@ xfs_daddr_to_agbno(struct xfs_mount *mp, xfs_daddr_t d) /* * perag get/put wrappers for eventual ref counting */ -static inline xfs_perag_t * -xfs_get_perag(struct xfs_mount *mp, xfs_ino_t ino) +static inline struct xfs_perag * +xfs_perag_get(struct xfs_mount *mp, xfs_agnumber_t agno) { - return &mp->m_perag[XFS_INO_TO_AGNO(mp, ino)]; + return &mp->m_perag[agno]; } static inline void -xfs_put_perag(struct xfs_mount *mp, xfs_perag_t *pag) +xfs_perag_put(struct xfs_perag *pag) { /* nothing to see here, move along */ } -- cgit v0.10.2 From a862e0fdcb8862aab2538ec2fc2f0dc07a625c59 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:47:41 +0000 Subject: xfs: Don't directly reference m_perag in allocation code Start abstracting the perag references so that the indexing of the structures is not directly coded into all the places that uses the perag structures. This will allow us to separate the use of the perag structure and the way it is indexed and hence avoid the known deadlocks related to growing a busy filesystem. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index 275b1f4..84070f2 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -1662,11 +1662,13 @@ xfs_free_ag_extent( xfs_agf_t *agf; xfs_perag_t *pag; /* per allocation group data */ + pag = xfs_perag_get(mp, agno); + pag->pagf_freeblks += len; + xfs_perag_put(pag); + agf = XFS_BUF_TO_AGF(agbp); - pag = &mp->m_perag[agno]; be32_add_cpu(&agf->agf_freeblks, len); xfs_trans_agblocks_delta(tp, len); - pag->pagf_freeblks += len; XFS_WANT_CORRUPTED_GOTO( be32_to_cpu(agf->agf_freeblks) <= be32_to_cpu(agf->agf_length), @@ -1969,10 +1971,12 @@ xfs_alloc_get_freelist( xfs_trans_brelse(tp, agflbp); if (be32_to_cpu(agf->agf_flfirst) == XFS_AGFL_SIZE(mp)) agf->agf_flfirst = 0; - pag = &mp->m_perag[be32_to_cpu(agf->agf_seqno)]; + + pag = xfs_perag_get(mp, be32_to_cpu(agf->agf_seqno)); be32_add_cpu(&agf->agf_flcount, -1); xfs_trans_agflist_delta(tp, -1); pag->pagf_flcount--; + xfs_perag_put(pag); logflags = XFS_AGF_FLFIRST | XFS_AGF_FLCOUNT; if (btreeblk) { @@ -2078,7 +2082,8 @@ xfs_alloc_put_freelist( be32_add_cpu(&agf->agf_fllast, 1); if (be32_to_cpu(agf->agf_fllast) == XFS_AGFL_SIZE(mp)) agf->agf_fllast = 0; - pag = &mp->m_perag[be32_to_cpu(agf->agf_seqno)]; + + pag = xfs_perag_get(mp, be32_to_cpu(agf->agf_seqno)); be32_add_cpu(&agf->agf_flcount, 1); xfs_trans_agflist_delta(tp, 1); pag->pagf_flcount++; @@ -2089,6 +2094,7 @@ xfs_alloc_put_freelist( pag->pagf_btreeblks--; logflags |= XFS_AGF_BTREEBLKS; } + xfs_perag_put(pag); xfs_alloc_log_agf(tp, agbp, logflags); @@ -2152,7 +2158,6 @@ xfs_read_agf( xfs_trans_brelse(tp, *bpp); return XFS_ERROR(EFSCORRUPTED); } - XFS_BUF_SET_VTYPE_REF(*bpp, B_FS_AGF, XFS_AGF_REF); return 0; } @@ -2184,7 +2189,7 @@ xfs_alloc_read_agf( ASSERT(!XFS_BUF_GETERROR(*bpp)); agf = XFS_BUF_TO_AGF(*bpp); - pag = &mp->m_perag[agno]; + pag = xfs_perag_get(mp, agno); if (!pag->pagf_init) { pag->pagf_freeblks = be32_to_cpu(agf->agf_freeblks); pag->pagf_btreeblks = be32_to_cpu(agf->agf_btreeblks); @@ -2211,6 +2216,7 @@ xfs_alloc_read_agf( be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi])); } #endif + xfs_perag_put(pag); return 0; } @@ -2271,7 +2277,7 @@ xfs_alloc_vextent( */ args->agno = XFS_FSB_TO_AGNO(mp, args->fsbno); down_read(&mp->m_peraglock); - args->pag = &mp->m_perag[args->agno]; + args->pag = xfs_perag_get(mp, args->agno); args->minleft = 0; error = xfs_alloc_fix_freelist(args, 0); args->minleft = minleft; @@ -2341,7 +2347,7 @@ xfs_alloc_vextent( */ down_read(&mp->m_peraglock); for (;;) { - args->pag = &mp->m_perag[args->agno]; + args->pag = xfs_perag_get(mp, args->agno); if (no_min) args->minleft = 0; error = xfs_alloc_fix_freelist(args, flags); args->minleft = minleft; @@ -2400,6 +2406,7 @@ xfs_alloc_vextent( } } } + xfs_perag_put(args->pag); } up_read(&mp->m_peraglock); if (bump_rotor || (type == XFS_ALLOCTYPE_ANY_AG)) { @@ -2427,8 +2434,10 @@ xfs_alloc_vextent( args->len); #endif } + xfs_perag_put(args->pag); return 0; error0: + xfs_perag_put(args->pag); up_read(&mp->m_peraglock); return error; } @@ -2455,7 +2464,7 @@ xfs_free_extent( ASSERT(args.agno < args.mp->m_sb.sb_agcount); args.agbno = XFS_FSB_TO_AGBNO(args.mp, bno); down_read(&args.mp->m_peraglock); - args.pag = &args.mp->m_perag[args.agno]; + args.pag = xfs_perag_get(args.mp, args.agno); if ((error = xfs_alloc_fix_freelist(&args, XFS_ALLOC_FLAG_FREEING))) goto error0; #ifdef DEBUG @@ -2465,6 +2474,7 @@ xfs_free_extent( #endif error = xfs_free_ag_extent(tp, args.agbp, args.agno, args.agbno, len, 0); error0: + xfs_perag_put(args.pag); up_read(&args.mp->m_peraglock); return error; } @@ -2486,15 +2496,15 @@ xfs_alloc_mark_busy(xfs_trans_t *tp, xfs_agblock_t bno, xfs_extlen_t len) { - xfs_mount_t *mp; xfs_perag_busy_t *bsy; + struct xfs_perag *pag; int n; - mp = tp->t_mountp; - spin_lock(&mp->m_perag[agno].pagb_lock); + pag = xfs_perag_get(tp->t_mountp, agno); + spin_lock(&pag->pagb_lock); /* search pagb_list for an open slot */ - for (bsy = mp->m_perag[agno].pagb_list, n = 0; + for (bsy = pag->pagb_list, n = 0; n < XFS_PAGB_NUM_SLOTS; bsy++, n++) { if (bsy->busy_tp == NULL) { @@ -2502,11 +2512,11 @@ xfs_alloc_mark_busy(xfs_trans_t *tp, } } - trace_xfs_alloc_busy(mp, agno, bno, len, n); + trace_xfs_alloc_busy(tp->t_mountp, agno, bno, len, n); if (n < XFS_PAGB_NUM_SLOTS) { - bsy = &mp->m_perag[agno].pagb_list[n]; - mp->m_perag[agno].pagb_count++; + bsy = &pag->pagb_list[n]; + pag->pagb_count++; bsy->busy_start = bno; bsy->busy_length = len; bsy->busy_tp = tp; @@ -2521,7 +2531,8 @@ xfs_alloc_mark_busy(xfs_trans_t *tp, xfs_trans_set_sync(tp); } - spin_unlock(&mp->m_perag[agno].pagb_lock); + spin_unlock(&pag->pagb_lock); + xfs_perag_put(pag); } void @@ -2529,24 +2540,23 @@ xfs_alloc_clear_busy(xfs_trans_t *tp, xfs_agnumber_t agno, int idx) { - xfs_mount_t *mp; + struct xfs_perag *pag; xfs_perag_busy_t *list; - mp = tp->t_mountp; - - spin_lock(&mp->m_perag[agno].pagb_lock); - list = mp->m_perag[agno].pagb_list; - ASSERT(idx < XFS_PAGB_NUM_SLOTS); + pag = xfs_perag_get(tp->t_mountp, agno); + spin_lock(&pag->pagb_lock); + list = pag->pagb_list; - trace_xfs_alloc_unbusy(mp, agno, idx, list[idx].busy_tp == tp); + trace_xfs_alloc_unbusy(tp->t_mountp, agno, idx, list[idx].busy_tp == tp); if (list[idx].busy_tp == tp) { list[idx].busy_tp = NULL; - mp->m_perag[agno].pagb_count--; + pag->pagb_count--; } - spin_unlock(&mp->m_perag[agno].pagb_lock); + spin_unlock(&pag->pagb_lock); + xfs_perag_put(pag); } @@ -2560,17 +2570,15 @@ xfs_alloc_search_busy(xfs_trans_t *tp, xfs_agblock_t bno, xfs_extlen_t len) { - xfs_mount_t *mp; + struct xfs_perag *pag; xfs_perag_busy_t *bsy; xfs_agblock_t uend, bend; xfs_lsn_t lsn = 0; int cnt; - mp = tp->t_mountp; - - spin_lock(&mp->m_perag[agno].pagb_lock); - - uend = bno + len - 1; + pag = xfs_perag_get(tp->t_mountp, agno); + spin_lock(&pag->pagb_lock); + cnt = pag->pagb_count; /* * search pagb_list for this slot, skipping open slots. We have to @@ -2578,8 +2586,9 @@ xfs_alloc_search_busy(xfs_trans_t *tp, * we have to get the most recent LSN for the log force to push out * all the transactions that span the range. */ - for (cnt = 0; cnt < mp->m_perag[agno].pagb_count; cnt++) { - bsy = &mp->m_perag[agno].pagb_list[cnt]; + uend = bno + len - 1; + for (cnt = 0; cnt < pag->pagb_count; cnt++) { + bsy = &pag->pagb_list[cnt]; if (!bsy->busy_tp) continue; @@ -2591,7 +2600,8 @@ xfs_alloc_search_busy(xfs_trans_t *tp, if (XFS_LSN_CMP(bsy->busy_tp->t_commit_lsn, lsn) > 0) lsn = bsy->busy_tp->t_commit_lsn; } - spin_unlock(&mp->m_perag[agno].pagb_lock); + spin_unlock(&pag->pagb_lock); + xfs_perag_put(pag); trace_xfs_alloc_busysearch(tp->t_mountp, agno, bno, len, lsn); /* @@ -2599,5 +2609,5 @@ xfs_alloc_search_busy(xfs_trans_t *tp, * transaction that freed the block */ if (lsn) - xfs_log_force(mp, lsn, XFS_LOG_FORCE|XFS_LOG_SYNC); + xfs_log_force(tp->t_mountp, lsn, XFS_LOG_FORCE|XFS_LOG_SYNC); } diff --git a/fs/xfs/xfs_alloc_btree.c b/fs/xfs/xfs_alloc_btree.c index adbd914..b726e10 100644 --- a/fs/xfs/xfs_alloc_btree.c +++ b/fs/xfs/xfs_alloc_btree.c @@ -61,12 +61,14 @@ xfs_allocbt_set_root( struct xfs_agf *agf = XFS_BUF_TO_AGF(agbp); xfs_agnumber_t seqno = be32_to_cpu(agf->agf_seqno); int btnum = cur->bc_btnum; + struct xfs_perag *pag = xfs_perag_get(cur->bc_mp, seqno); ASSERT(ptr->s != 0); agf->agf_roots[btnum] = ptr->s; be32_add_cpu(&agf->agf_levels[btnum], inc); - cur->bc_mp->m_perag[seqno].pagf_levels[btnum] += inc; + pag->pagf_levels[btnum] += inc; + xfs_perag_put(pag); xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS); } @@ -150,6 +152,7 @@ xfs_allocbt_update_lastrec( { struct xfs_agf *agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp); xfs_agnumber_t seqno = be32_to_cpu(agf->agf_seqno); + struct xfs_perag *pag; __be32 len; int numrecs; @@ -193,7 +196,9 @@ xfs_allocbt_update_lastrec( } agf->agf_longest = len; - cur->bc_mp->m_perag[seqno].pagf_longest = be32_to_cpu(len); + pag = xfs_perag_get(cur->bc_mp, seqno); + pag->pagf_longest = be32_to_cpu(len); + xfs_perag_put(pag); xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp, XFS_AGF_LONGEST); } -- cgit v0.10.2 From 4196ac08c023c6dab90c3fa460d9c06deaa304c4 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:47:42 +0000 Subject: xfs: Convert filestreams code to use per-ag get/put routines Use xfs_perag_get() and xfs_perag_put() in the filestreams code. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c index a631e14..e61f2aa 100644 --- a/fs/xfs/xfs_filestream.c +++ b/fs/xfs/xfs_filestream.c @@ -140,6 +140,7 @@ _xfs_filestream_pick_ag( int flags, xfs_extlen_t minlen) { + int streams, max_streams; int err, trylock, nscan; xfs_extlen_t longest, free, minfree, maxfree = 0; xfs_agnumber_t ag, max_ag = NULLAGNUMBER; @@ -155,15 +156,15 @@ _xfs_filestream_pick_ag( trylock = XFS_ALLOC_FLAG_TRYLOCK; for (nscan = 0; 1; nscan++) { - - TRACE_AG_SCAN(mp, ag, xfs_filestream_peek_ag(mp, ag)); - - pag = mp->m_perag + ag; + pag = xfs_perag_get(mp, ag); + TRACE_AG_SCAN(mp, ag, atomic_read(&pag->pagf_fstrms)); if (!pag->pagf_init) { err = xfs_alloc_pagf_init(mp, NULL, ag, trylock); - if (err && !trylock) + if (err && !trylock) { + xfs_perag_put(pag); return err; + } } /* Might fail sometimes during the 1st pass with trylock set. */ @@ -173,6 +174,7 @@ _xfs_filestream_pick_ag( /* Keep track of the AG with the most free blocks. */ if (pag->pagf_freeblks > maxfree) { maxfree = pag->pagf_freeblks; + max_streams = atomic_read(&pag->pagf_fstrms); max_ag = ag; } @@ -195,6 +197,8 @@ _xfs_filestream_pick_ag( /* Break out, retaining the reference on the AG. */ free = pag->pagf_freeblks; + streams = atomic_read(&pag->pagf_fstrms); + xfs_perag_put(pag); *agp = ag; break; } @@ -202,6 +206,7 @@ _xfs_filestream_pick_ag( /* Drop the reference on this AG, it's not usable. */ xfs_filestream_put_ag(mp, ag); next_ag: + xfs_perag_put(pag); /* Move to the next AG, wrapping to AG 0 if necessary. */ if (++ag >= mp->m_sb.sb_agcount) ag = 0; @@ -229,6 +234,7 @@ next_ag: if (max_ag != NULLAGNUMBER) { xfs_filestream_get_ag(mp, max_ag); TRACE_AG_PICK1(mp, max_ag, maxfree); + streams = max_streams; free = maxfree; *agp = max_ag; break; @@ -240,8 +246,7 @@ next_ag: return 0; } - TRACE_AG_PICK2(mp, startag, *agp, xfs_filestream_peek_ag(mp, *agp), - free, nscan, flags); + TRACE_AG_PICK2(mp, startag, *agp, streams, free, nscan, flags); return 0; } diff --git a/fs/xfs/xfs_filestream.h b/fs/xfs/xfs_filestream.h index 4aba67c..58378b2 100644 --- a/fs/xfs/xfs_filestream.h +++ b/fs/xfs/xfs_filestream.h @@ -79,12 +79,21 @@ extern ktrace_t *xfs_filestreams_trace_buf; * the cache that reference per-ag array elements that have since been * reallocated. */ +/* + * xfs_filestream_peek_ag is only used in tracing code + */ static inline int xfs_filestream_peek_ag( xfs_mount_t *mp, xfs_agnumber_t agno) { - return atomic_read(&mp->m_perag[agno].pagf_fstrms); + struct xfs_perag *pag; + int ret; + + pag = xfs_perag_get(mp, agno); + ret = atomic_read(&pag->pagf_fstrms); + xfs_perag_put(pag); + return ret; } static inline int @@ -92,7 +101,13 @@ xfs_filestream_get_ag( xfs_mount_t *mp, xfs_agnumber_t agno) { - return atomic_inc_return(&mp->m_perag[agno].pagf_fstrms); + struct xfs_perag *pag; + int ret; + + pag = xfs_perag_get(mp, agno); + ret = atomic_inc_return(&pag->pagf_fstrms); + xfs_perag_put(pag); + return ret; } static inline int @@ -100,7 +115,13 @@ xfs_filestream_put_ag( xfs_mount_t *mp, xfs_agnumber_t agno) { - return atomic_dec_return(&mp->m_perag[agno].pagf_fstrms); + struct xfs_perag *pag; + int ret; + + pag = xfs_perag_get(mp, agno); + ret = atomic_dec_return(&pag->pagf_fstrms); + xfs_perag_put(pag); + return ret; } /* allocation selection flags */ -- cgit v0.10.2 From 44b56e0a1aed522a10051645e85d300e10926fd3 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:47:43 +0000 Subject: xfs: convert remaining direct references to m_perag Convert the remaining direct lookups of the per ag structures to use get/put accesses. Ensure that the loops across AGs and prior users of the interface balance gets and puts correctly. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 98251cd..a9b95d9 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -2630,11 +2630,12 @@ xfs_bmap_btalloc( startag = ag = 0; notinit = 0; down_read(&mp->m_peraglock); + pag = xfs_perag_get(mp, ag); while (blen < ap->alen) { - pag = &mp->m_perag[ag]; if (!pag->pagf_init && (error = xfs_alloc_pagf_init(mp, args.tp, ag, XFS_ALLOC_FLAG_TRYLOCK))) { + xfs_perag_put(pag); up_read(&mp->m_peraglock); return error; } @@ -2667,6 +2668,7 @@ xfs_bmap_btalloc( break; error = xfs_filestream_new_ag(ap, &ag); + xfs_perag_put(pag); if (error) { up_read(&mp->m_peraglock); return error; @@ -2674,6 +2676,7 @@ xfs_bmap_btalloc( /* loop again to set 'blen'*/ startag = NULLAGNUMBER; + pag = xfs_perag_get(mp, ag); continue; } } @@ -2681,7 +2684,10 @@ xfs_bmap_btalloc( ag = 0; if (ag == startag) break; + xfs_perag_put(pag); + pag = xfs_perag_get(mp, ag); } + xfs_perag_put(pag); up_read(&mp->m_peraglock); /* * Since the above loop did a BUF_TRYLOCK, it is diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index cb907ba..884ee13 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -253,6 +253,7 @@ xfs_ialloc_ag_alloc( xfs_agino_t thisino; /* current inode number, for loop */ int isaligned = 0; /* inode allocation at stripe unit */ /* boundary */ + struct xfs_perag *pag; args.tp = tp; args.mp = tp->t_mountp; @@ -383,7 +384,9 @@ xfs_ialloc_ag_alloc( be32_add_cpu(&agi->agi_count, newlen); be32_add_cpu(&agi->agi_freecount, newlen); down_read(&args.mp->m_peraglock); - args.mp->m_perag[agno].pagi_freecount += newlen; + pag = xfs_perag_get(args.mp, agno); + pag->pagi_freecount += newlen; + xfs_perag_put(pag); up_read(&args.mp->m_peraglock); agi->agi_newino = cpu_to_be32(newino); @@ -488,7 +491,7 @@ xfs_ialloc_ag_select( flags = XFS_ALLOC_FLAG_TRYLOCK; down_read(&mp->m_peraglock); for (;;) { - pag = &mp->m_perag[agno]; + pag = xfs_perag_get(mp, agno); if (!pag->pagi_init) { if (xfs_ialloc_read_agi(mp, tp, agno, &agbp)) { agbp = NULL; @@ -527,6 +530,7 @@ xfs_ialloc_ag_select( agbp = NULL; goto nextag; } + xfs_perag_put(pag); up_read(&mp->m_peraglock); return agbp; } @@ -535,6 +539,7 @@ unlock_nextag: if (agbp) xfs_trans_brelse(tp, agbp); nextag: + xfs_perag_put(pag); /* * No point in iterating over the rest, if we're shutting * down. @@ -672,6 +677,7 @@ xfs_dialloc( xfs_agnumber_t tagno; /* testing allocation group number */ xfs_btree_cur_t *tcur; /* temp cursor */ xfs_inobt_rec_incore_t trec; /* temp inode allocation record */ + struct xfs_perag *pag; if (*IO_agbp == NULL) { @@ -772,11 +778,14 @@ nextag: return noroom ? ENOSPC : 0; } down_read(&mp->m_peraglock); - if (mp->m_perag[tagno].pagi_inodeok == 0) { + pag = xfs_perag_get(mp, tagno); + if (pag->pagi_inodeok == 0) { + xfs_perag_put(pag); up_read(&mp->m_peraglock); goto nextag; } error = xfs_ialloc_read_agi(mp, tp, tagno, &agbp); + xfs_perag_put(pag); up_read(&mp->m_peraglock); if (error) goto nextag; @@ -790,6 +799,7 @@ nextag: */ agno = tagno; *IO_agbp = NULL; + pag = xfs_perag_get(mp, agno); restart_pagno: cur = xfs_inobt_init_cursor(mp, tp, agbp, be32_to_cpu(agi->agi_seqno)); @@ -808,7 +818,6 @@ nextag: * If in the same AG as the parent, try to get near the parent. */ if (pagno == agno) { - xfs_perag_t *pag = &mp->m_perag[agno]; int doneleft; /* done, to the left */ int doneright; /* done, to the right */ int searchdistance = 10; @@ -1007,7 +1016,7 @@ alloc_inode: be32_add_cpu(&agi->agi_freecount, -1); xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT); down_read(&mp->m_peraglock); - mp->m_perag[tagno].pagi_freecount--; + pag->pagi_freecount--; up_read(&mp->m_peraglock); error = xfs_check_agi_freecount(cur, agi); @@ -1016,12 +1025,14 @@ alloc_inode: xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR); xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1); + xfs_perag_put(pag); *inop = ino; return 0; error1: xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR); error0: xfs_btree_del_cursor(cur, XFS_BTREE_ERROR); + xfs_perag_put(pag); return error; } @@ -1052,6 +1063,7 @@ xfs_difree( xfs_mount_t *mp; /* mount structure for filesystem */ int off; /* offset of inode in inode chunk */ xfs_inobt_rec_incore_t rec; /* btree record */ + struct xfs_perag *pag; mp = tp->t_mountp; @@ -1158,7 +1170,9 @@ xfs_difree( be32_add_cpu(&agi->agi_freecount, -(ilen - 1)); xfs_ialloc_log_agi(tp, agbp, XFS_AGI_COUNT | XFS_AGI_FREECOUNT); down_read(&mp->m_peraglock); - mp->m_perag[agno].pagi_freecount -= ilen - 1; + pag = xfs_perag_get(mp, agno); + pag->pagi_freecount -= ilen - 1; + xfs_perag_put(pag); up_read(&mp->m_peraglock); xfs_trans_mod_sb(tp, XFS_TRANS_SB_ICOUNT, -ilen); xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -(ilen - 1)); @@ -1189,7 +1203,9 @@ xfs_difree( be32_add_cpu(&agi->agi_freecount, 1); xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT); down_read(&mp->m_peraglock); - mp->m_perag[agno].pagi_freecount++; + pag = xfs_perag_get(mp, agno); + pag->pagi_freecount++; + xfs_perag_put(pag); up_read(&mp->m_peraglock); xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, 1); } @@ -1379,7 +1395,6 @@ xfs_imap( XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks)); return XFS_ERROR(EINVAL); } - return 0; } @@ -1523,8 +1538,7 @@ xfs_ialloc_read_agi( return error; agi = XFS_BUF_TO_AGI(*bpp); - pag = &mp->m_perag[agno]; - + pag = xfs_perag_get(mp, agno); if (!pag->pagi_init) { pag->pagi_freecount = be32_to_cpu(agi->agi_freecount); pag->pagi_count = be32_to_cpu(agi->agi_count); @@ -1537,6 +1551,7 @@ xfs_ialloc_read_agi( */ ASSERT(pag->pagi_freecount == be32_to_cpu(agi->agi_freecount) || XFS_FORCED_SHUTDOWN(mp)); + xfs_perag_put(pag); return 0; } diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index bd3d816..0317b00 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2695,7 +2695,7 @@ xfs_iflush_cluster( ilist_size = inodes_per_cluster * sizeof(xfs_inode_t *); ilist = kmem_alloc(ilist_size, KM_MAYFAIL|KM_NOFS); if (!ilist) - return 0; + goto out_put; mask = ~(((XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog)) - 1); first_index = XFS_INO_TO_AGINO(mp, ip->i_ino) & mask; @@ -2764,6 +2764,8 @@ xfs_iflush_cluster( out_free: read_unlock(&pag->pag_ici_lock); kmem_free(ilist); +out_put: + xfs_perag_put(pag); return 0; @@ -2807,6 +2809,7 @@ cluster_corrupt_out: */ xfs_iflush_abort(iq); kmem_free(ilist); + xfs_perag_put(pag); return XFS_ERROR(EFSCORRUPTED); } diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index eb403b4..9055b60 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -438,18 +438,20 @@ xfs_initialize_perag( } /* This ag is preferred for inodes */ - pag = &mp->m_perag[index]; + pag = xfs_perag_get(mp, index); pag->pagi_inodeok = 1; if (index < max_metadata) pag->pagf_metadata = 1; xfs_initialize_perag_icache(pag); + xfs_perag_put(pag); } } else { /* Setup default behavior for smaller filesystems */ for (index = 0; index < agcount; index++) { - pag = &mp->m_perag[index]; + pag = xfs_perag_get(mp, index); pag->pagi_inodeok = 1; xfs_initialize_perag_icache(pag); + xfs_perag_put(pag); } } return index; @@ -731,12 +733,13 @@ xfs_initialize_perag_data(xfs_mount_t *mp, xfs_agnumber_t agcount) error = xfs_ialloc_pagi_init(mp, NULL, index); if (error) return error; - pag = &mp->m_perag[index]; + pag = xfs_perag_get(mp, index); ifree += pag->pagi_freecount; ialloc += pag->pagi_count; bfree += pag->pagf_freeblks; bfreelst += pag->pagf_flcount; btree += pag->pagf_btreeblks; + xfs_perag_put(pag); } /* * Overwrite incore superblock counters with just-read data -- cgit v0.10.2 From 1c1c6ebcf5284aee4910f3b906ac90c20e510c82 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:47:44 +0000 Subject: xfs: Replace per-ag array with a radix tree The use of an array for the per-ag structures requires reallocation of the array when growing the filesystem. This requires locking access to the array to avoid use after free situations, and the locking is difficult to get right. To avoid needing to reallocate an array, change the per-ag structures to an allocated object per ag and index them using a tree structure. The AGs are always densely indexed (hence the use of an array), but the number supported is 2^32 and lookups tend to be random and hence indexing needs to scale. A simple choice is a radix tree - it works well with this sort of index. This change also removes another large contiguous allocation from the mount/growfs path in XFS. The growing process now needs to change to only initialise the new AGs required for the extra space, and as such only needs to exclusively lock the tree for inserts. The rest of the code only needs to lock the tree while doing lookups, and hence this will remove all the deadlocks that currently occur on the m_perag_lock as it is now an innermost lock. The lock is also changed to a spinlock from a read/write lock as the hold time is now extremely short. To complete the picture, the per-ag structures will need to be reference counted to ensure that we don't free/modify them while they are still in use. This will be done in subsequent patch. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index 84070f2..4d66bb7 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -2276,7 +2276,6 @@ xfs_alloc_vextent( * These three force us into a single a.g. */ args->agno = XFS_FSB_TO_AGNO(mp, args->fsbno); - down_read(&mp->m_peraglock); args->pag = xfs_perag_get(mp, args->agno); args->minleft = 0; error = xfs_alloc_fix_freelist(args, 0); @@ -2286,14 +2285,12 @@ xfs_alloc_vextent( goto error0; } if (!args->agbp) { - up_read(&mp->m_peraglock); trace_xfs_alloc_vextent_noagbp(args); break; } args->agbno = XFS_FSB_TO_AGBNO(mp, args->fsbno); if ((error = xfs_alloc_ag_vextent(args))) goto error0; - up_read(&mp->m_peraglock); break; case XFS_ALLOCTYPE_START_BNO: /* @@ -2345,7 +2342,6 @@ xfs_alloc_vextent( * Loop over allocation groups twice; first time with * trylock set, second time without. */ - down_read(&mp->m_peraglock); for (;;) { args->pag = xfs_perag_get(mp, args->agno); if (no_min) args->minleft = 0; @@ -2408,7 +2404,6 @@ xfs_alloc_vextent( } xfs_perag_put(args->pag); } - up_read(&mp->m_peraglock); if (bump_rotor || (type == XFS_ALLOCTYPE_ANY_AG)) { if (args->agno == sagno) mp->m_agfrotor = (mp->m_agfrotor + 1) % @@ -2438,7 +2433,6 @@ xfs_alloc_vextent( return 0; error0: xfs_perag_put(args->pag); - up_read(&mp->m_peraglock); return error; } @@ -2463,7 +2457,6 @@ xfs_free_extent( args.agno = XFS_FSB_TO_AGNO(args.mp, bno); ASSERT(args.agno < args.mp->m_sb.sb_agcount); args.agbno = XFS_FSB_TO_AGBNO(args.mp, bno); - down_read(&args.mp->m_peraglock); args.pag = xfs_perag_get(args.mp, args.agno); if ((error = xfs_alloc_fix_freelist(&args, XFS_ALLOC_FLAG_FREEING))) goto error0; @@ -2475,7 +2468,6 @@ xfs_free_extent( error = xfs_free_ag_extent(tp, args.agbp, args.agno, args.agbno, len, 0); error0: xfs_perag_put(args.pag); - up_read(&args.mp->m_peraglock); return error; } diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index a9b95d9..7c6d9ac 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -2629,14 +2629,12 @@ xfs_bmap_btalloc( if (startag == NULLAGNUMBER) startag = ag = 0; notinit = 0; - down_read(&mp->m_peraglock); pag = xfs_perag_get(mp, ag); while (blen < ap->alen) { if (!pag->pagf_init && (error = xfs_alloc_pagf_init(mp, args.tp, ag, XFS_ALLOC_FLAG_TRYLOCK))) { xfs_perag_put(pag); - up_read(&mp->m_peraglock); return error; } /* @@ -2669,10 +2667,8 @@ xfs_bmap_btalloc( error = xfs_filestream_new_ag(ap, &ag); xfs_perag_put(pag); - if (error) { - up_read(&mp->m_peraglock); + if (error) return error; - } /* loop again to set 'blen'*/ startag = NULLAGNUMBER; @@ -2688,7 +2684,6 @@ xfs_bmap_btalloc( pag = xfs_perag_get(mp, ag); } xfs_perag_put(pag); - up_read(&mp->m_peraglock); /* * Since the above loop did a BUF_TRYLOCK, it is * possible that there is space for this request. diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c index e61f2aa..914d00d 100644 --- a/fs/xfs/xfs_filestream.c +++ b/fs/xfs/xfs_filestream.c @@ -253,8 +253,7 @@ next_ag: /* * Set the allocation group number for a file or a directory, updating inode - * references and per-AG references as appropriate. Must be called with the - * m_peraglock held in read mode. + * references and per-AG references as appropriate. */ static int _xfs_filestream_update_ag( @@ -456,10 +455,10 @@ xfs_filestream_unmount( } /* - * If the mount point's m_perag array is going to be reallocated, all + * If the mount point's m_perag tree is going to be modified, all * outstanding cache entries must be flushed to avoid accessing reference count * addresses that have been freed. The call to xfs_filestream_flush() must be - * made inside the block that holds the m_peraglock in write mode to do the + * made inside the block that holds the m_perag_lock in write mode to do the * reallocation. */ void @@ -531,7 +530,6 @@ xfs_filestream_associate( mp = pip->i_mount; cache = mp->m_filestream; - down_read(&mp->m_peraglock); /* * We have a problem, Houston. @@ -548,10 +546,8 @@ xfs_filestream_associate( * * So, if we can't get the iolock without sleeping then just give up */ - if (!xfs_ilock_nowait(pip, XFS_IOLOCK_EXCL)) { - up_read(&mp->m_peraglock); + if (!xfs_ilock_nowait(pip, XFS_IOLOCK_EXCL)) return 1; - } /* If the parent directory is already in the cache, use its AG. */ item = xfs_mru_cache_lookup(cache, pip->i_ino); @@ -606,7 +602,6 @@ exit_did_pick: exit: xfs_iunlock(pip, XFS_IOLOCK_EXCL); - up_read(&mp->m_peraglock); return -err; } diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index a13919a..37a6f62 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -167,27 +167,14 @@ xfs_growfs_data_private( } new = nb - mp->m_sb.sb_dblocks; oagcount = mp->m_sb.sb_agcount; - if (nagcount > oagcount) { - void *new_perag, *old_perag; - - xfs_filestream_flush(mp); - - new_perag = kmem_zalloc(sizeof(xfs_perag_t) * nagcount, - KM_MAYFAIL); - if (!new_perag) - return XFS_ERROR(ENOMEM); - - down_write(&mp->m_peraglock); - memcpy(new_perag, mp->m_perag, sizeof(xfs_perag_t) * oagcount); - old_perag = mp->m_perag; - mp->m_perag = new_perag; - - mp->m_flags |= XFS_MOUNT_32BITINODES; - nagimax = xfs_initialize_perag(mp, nagcount); - up_write(&mp->m_peraglock); - kmem_free(old_perag); + /* allocate the new per-ag structures */ + if (nagcount > oagcount) { + error = xfs_initialize_perag(mp, nagcount, &nagimax); + if (error) + return error; } + tp = xfs_trans_alloc(mp, XFS_TRANS_GROWFS); tp->t_flags |= XFS_TRANS_RESERVE; if ((error = xfs_trans_reserve(tp, XFS_GROWFS_SPACE_RES(mp), @@ -196,6 +183,11 @@ xfs_growfs_data_private( return error; } + /* + * Write new AG headers to disk. Non-transactional, but written + * synchronously so they are completed prior to the growfs transaction + * being logged. + */ nfree = 0; for (agno = nagcount - 1; agno >= oagcount; agno--, new -= agsize) { /* @@ -359,6 +351,12 @@ xfs_growfs_data_private( goto error0; } } + + /* + * Update changed superblock fields transactionally. These are not + * seen by the rest of the world until the transaction commit applies + * them atomically to the superblock. + */ if (nagcount > oagcount) xfs_trans_mod_sb(tp, XFS_TRANS_SB_AGCOUNT, nagcount - oagcount); if (nb > mp->m_sb.sb_dblocks) @@ -369,9 +367,9 @@ xfs_growfs_data_private( if (dpct) xfs_trans_mod_sb(tp, XFS_TRANS_SB_IMAXPCT, dpct); error = xfs_trans_commit(tp, 0); - if (error) { + if (error) return error; - } + /* New allocation groups fully initialized, so update mount struct */ if (nagimax) mp->m_maxagi = nagimax; @@ -381,6 +379,8 @@ xfs_growfs_data_private( mp->m_maxicount = icount << mp->m_sb.sb_inopblog; } else mp->m_maxicount = 0; + + /* update secondary superblocks. */ for (agno = 1; agno < nagcount; agno++) { error = xfs_read_buf(mp, mp->m_ddev_targp, XFS_AGB_TO_DADDR(mp, agno, XFS_SB_BLOCK(mp)), diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index 884ee13..52c9d00 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -383,11 +383,9 @@ xfs_ialloc_ag_alloc( newino = XFS_OFFBNO_TO_AGINO(args.mp, args.agbno, 0); be32_add_cpu(&agi->agi_count, newlen); be32_add_cpu(&agi->agi_freecount, newlen); - down_read(&args.mp->m_peraglock); pag = xfs_perag_get(args.mp, agno); pag->pagi_freecount += newlen; xfs_perag_put(pag); - up_read(&args.mp->m_peraglock); agi->agi_newino = cpu_to_be32(newino); /* @@ -489,7 +487,6 @@ xfs_ialloc_ag_select( */ agno = pagno; flags = XFS_ALLOC_FLAG_TRYLOCK; - down_read(&mp->m_peraglock); for (;;) { pag = xfs_perag_get(mp, agno); if (!pag->pagi_init) { @@ -531,7 +528,6 @@ xfs_ialloc_ag_select( goto nextag; } xfs_perag_put(pag); - up_read(&mp->m_peraglock); return agbp; } } @@ -544,18 +540,14 @@ nextag: * No point in iterating over the rest, if we're shutting * down. */ - if (XFS_FORCED_SHUTDOWN(mp)) { - up_read(&mp->m_peraglock); + if (XFS_FORCED_SHUTDOWN(mp)) return NULL; - } agno++; if (agno >= agcount) agno = 0; if (agno == pagno) { - if (flags == 0) { - up_read(&mp->m_peraglock); + if (flags == 0) return NULL; - } flags = 0; } } @@ -777,16 +769,13 @@ nextag: *inop = NULLFSINO; return noroom ? ENOSPC : 0; } - down_read(&mp->m_peraglock); pag = xfs_perag_get(mp, tagno); if (pag->pagi_inodeok == 0) { xfs_perag_put(pag); - up_read(&mp->m_peraglock); goto nextag; } error = xfs_ialloc_read_agi(mp, tp, tagno, &agbp); xfs_perag_put(pag); - up_read(&mp->m_peraglock); if (error) goto nextag; agi = XFS_BUF_TO_AGI(agbp); @@ -1015,9 +1004,7 @@ alloc_inode: goto error0; be32_add_cpu(&agi->agi_freecount, -1); xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT); - down_read(&mp->m_peraglock); pag->pagi_freecount--; - up_read(&mp->m_peraglock); error = xfs_check_agi_freecount(cur, agi); if (error) @@ -1100,9 +1087,7 @@ xfs_difree( /* * Get the allocation group header. */ - down_read(&mp->m_peraglock); error = xfs_ialloc_read_agi(mp, tp, agno, &agbp); - up_read(&mp->m_peraglock); if (error) { cmn_err(CE_WARN, "xfs_difree: xfs_ialloc_read_agi() returned an error %d on %s. Returning error.", @@ -1169,11 +1154,9 @@ xfs_difree( be32_add_cpu(&agi->agi_count, -ilen); be32_add_cpu(&agi->agi_freecount, -(ilen - 1)); xfs_ialloc_log_agi(tp, agbp, XFS_AGI_COUNT | XFS_AGI_FREECOUNT); - down_read(&mp->m_peraglock); pag = xfs_perag_get(mp, agno); pag->pagi_freecount -= ilen - 1; xfs_perag_put(pag); - up_read(&mp->m_peraglock); xfs_trans_mod_sb(tp, XFS_TRANS_SB_ICOUNT, -ilen); xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -(ilen - 1)); @@ -1202,11 +1185,9 @@ xfs_difree( */ be32_add_cpu(&agi->agi_freecount, 1); xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT); - down_read(&mp->m_peraglock); pag = xfs_perag_get(mp, agno); pag->pagi_freecount++; xfs_perag_put(pag); - up_read(&mp->m_peraglock); xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, 1); } @@ -1328,9 +1309,7 @@ xfs_imap( xfs_buf_t *agbp; /* agi buffer */ int i; /* temp state */ - down_read(&mp->m_peraglock); error = xfs_ialloc_read_agi(mp, tp, agno, &agbp); - up_read(&mp->m_peraglock); if (error) { xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: " "xfs_ialloc_read_agi() returned " diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index 62efab2..940307a 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -420,9 +420,7 @@ xfs_bulkstat( while (XFS_BULKSTAT_UBLEFT(ubleft) && agno < mp->m_sb.sb_agcount) { cond_resched(); bp = NULL; - down_read(&mp->m_peraglock); error = xfs_ialloc_read_agi(mp, NULL, agno, &agbp); - up_read(&mp->m_peraglock); if (error) { /* * Skip this allocation group and go to the next one. @@ -849,9 +847,7 @@ xfs_inumbers( agbp = NULL; while (left > 0 && agno < mp->m_sb.sb_agcount) { if (agbp == NULL) { - down_read(&mp->m_peraglock); error = xfs_ialloc_read_agi(mp, NULL, agno, &agbp); - up_read(&mp->m_peraglock); if (error) { /* * If we can't read the AGI of this ag, diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 9055b60..c04dd83 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -209,13 +209,16 @@ STATIC void xfs_free_perag( xfs_mount_t *mp) { - if (mp->m_perag) { - int agno; + xfs_agnumber_t agno; + struct xfs_perag *pag; - for (agno = 0; agno < mp->m_maxagi; agno++) - if (mp->m_perag[agno].pagb_list) - kmem_free(mp->m_perag[agno].pagb_list); - kmem_free(mp->m_perag); + for (agno = 0; agno < mp->m_sb.sb_agcount; agno++) { + spin_lock(&mp->m_perag_lock); + pag = radix_tree_delete(&mp->m_perag_tree, agno); + spin_unlock(&mp->m_perag_lock); + ASSERT(pag); + kmem_free(pag->pagb_list); + kmem_free(pag); } } @@ -389,10 +392,11 @@ xfs_initialize_perag_icache( } } -xfs_agnumber_t +int xfs_initialize_perag( xfs_mount_t *mp, - xfs_agnumber_t agcount) + xfs_agnumber_t agcount, + xfs_agnumber_t *maxagi) { xfs_agnumber_t index, max_metadata; xfs_perag_t *pag; @@ -405,6 +409,33 @@ xfs_initialize_perag( agino = XFS_OFFBNO_TO_AGINO(mp, sbp->sb_agblocks - 1, 0); ino = XFS_AGINO_TO_INO(mp, agcount - 1, agino); + /* + * Walk the current per-ag tree so we don't try to initialise AGs + * that already exist (growfs case). Allocate and insert all the + * AGs we don't find ready for initialisation. + */ + for (index = 0; index < agcount; index++) { + pag = xfs_perag_get(mp, index); + if (pag) { + xfs_perag_put(pag); + continue; + } + pag = kmem_zalloc(sizeof(*pag), KM_MAYFAIL); + if (!pag) + return -ENOMEM; + if (radix_tree_preload(GFP_NOFS)) + return -ENOMEM; + spin_lock(&mp->m_perag_lock); + if (radix_tree_insert(&mp->m_perag_tree, index, pag)) { + BUG(); + spin_unlock(&mp->m_perag_lock); + kmem_free(pag); + return -EEXIST; + } + spin_unlock(&mp->m_perag_lock); + radix_tree_preload_end(); + } + /* Clear the mount flag if no inode can overflow 32 bits * on this filesystem, or if specifically requested.. */ @@ -454,7 +485,9 @@ xfs_initialize_perag( xfs_perag_put(pag); } } - return index; + if (maxagi) + *maxagi = index; + return 0; } void @@ -1155,13 +1188,13 @@ xfs_mountfs( /* * Allocate and initialize the per-ag data. */ - init_rwsem(&mp->m_peraglock); - mp->m_perag = kmem_zalloc(sbp->sb_agcount * sizeof(xfs_perag_t), - KM_MAYFAIL); - if (!mp->m_perag) + spin_lock_init(&mp->m_perag_lock); + INIT_RADIX_TREE(&mp->m_perag_tree, GFP_NOFS); + error = xfs_initialize_perag(mp, sbp->sb_agcount, &mp->m_maxagi); + if (error) { + cmn_err(CE_WARN, "XFS: Failed per-ag init: %d", error); goto out_remove_uuid; - - mp->m_maxagi = xfs_initialize_perag(mp, sbp->sb_agcount); + } if (!sbp->sb_logblocks) { cmn_err(CE_WARN, "XFS: no log defined"); diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index f8a68a2..cfa7a5d 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -207,8 +207,8 @@ typedef struct xfs_mount { uint m_ag_maxlevels; /* XFS_AG_MAXLEVELS */ uint m_bm_maxlevels[2]; /* XFS_BM_MAXLEVELS */ uint m_in_maxlevels; /* max inobt btree levels. */ - struct xfs_perag *m_perag; /* per-ag accounting info */ - struct rw_semaphore m_peraglock; /* lock for m_perag (pointer) */ + struct radix_tree_root m_perag_tree; /* per-ag accounting info */ + spinlock_t m_perag_lock; /* lock for m_perag_tree */ struct mutex m_growlock; /* growfs mutex */ int m_fixedfsid[2]; /* unchanged for life of FS */ uint m_dmevmask; /* DMI events for this FS */ @@ -389,7 +389,12 @@ xfs_daddr_to_agbno(struct xfs_mount *mp, xfs_daddr_t d) static inline struct xfs_perag * xfs_perag_get(struct xfs_mount *mp, xfs_agnumber_t agno) { - return &mp->m_perag[agno]; + struct xfs_perag *pag; + + spin_lock(&mp->m_perag_lock); + pag = radix_tree_lookup(&mp->m_perag_tree, agno); + spin_unlock(&mp->m_perag_lock); + return pag; } static inline void @@ -450,7 +455,8 @@ extern struct xfs_dmops xfs_dmcore_xfs; #endif /* __KERNEL__ */ extern void xfs_mod_sb(struct xfs_trans *, __int64_t); -extern xfs_agnumber_t xfs_initialize_perag(struct xfs_mount *, xfs_agnumber_t); +extern int xfs_initialize_perag(struct xfs_mount *, xfs_agnumber_t, + xfs_agnumber_t *); extern void xfs_sb_from_disk(struct xfs_sb *, struct xfs_dsb *); extern void xfs_sb_to_disk(struct xfs_dsb *, struct xfs_sb *, __int64_t); -- cgit v0.10.2 From aed3bb90abaf0b42e8c8747e192f7bb97f445279 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:47:45 +0000 Subject: xfs: Reference count per-ag structures Reference count the per-ag structures to ensure that we keep get/put pairs balanced. Assert that the reference counts are zero at unmount time to catch leaks. In future, reference counts will enable us to safely remove perag structures by allowing us to detect when they are no longer in use. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/xfs_ag.h b/fs/xfs/xfs_ag.h index 6702bd8..18ae43f 100644 --- a/fs/xfs/xfs_ag.h +++ b/fs/xfs/xfs_ag.h @@ -196,8 +196,8 @@ typedef struct xfs_perag_busy { #define XFS_PAGB_NUM_SLOTS 128 #endif -typedef struct xfs_perag -{ +typedef struct xfs_perag { + atomic_t pag_ref; /* perag reference count */ char pagf_init; /* this agf's entry is initialized */ char pagi_init; /* this agi's entry is initialized */ char pagf_metadata; /* the agf is preferred to be metadata */ diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index c04dd83..f241fec 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -215,6 +215,7 @@ xfs_free_perag( for (agno = 0; agno < mp->m_sb.sb_agcount; agno++) { spin_lock(&mp->m_perag_lock); pag = radix_tree_delete(&mp->m_perag_tree, agno); + ASSERT(atomic_read(&pag->pag_ref) == 0); spin_unlock(&mp->m_perag_lock); ASSERT(pag); kmem_free(pag->pagb_list); diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index cfa7a5d..16b2212 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -384,7 +384,7 @@ xfs_daddr_to_agbno(struct xfs_mount *mp, xfs_daddr_t d) } /* - * perag get/put wrappers for eventual ref counting + * perag get/put wrappers for ref counting */ static inline struct xfs_perag * xfs_perag_get(struct xfs_mount *mp, xfs_agnumber_t agno) @@ -393,6 +393,12 @@ xfs_perag_get(struct xfs_mount *mp, xfs_agnumber_t agno) spin_lock(&mp->m_perag_lock); pag = radix_tree_lookup(&mp->m_perag_tree, agno); + if (pag) { + ASSERT(atomic_read(&pag->pag_ref) >= 0); + /* catch leaks in the positive direction during testing */ + ASSERT(atomic_read(&pag->pag_ref) < 1000); + atomic_inc(&pag->pag_ref); + } spin_unlock(&mp->m_perag_lock); return pag; } @@ -400,7 +406,8 @@ xfs_perag_get(struct xfs_mount *mp, xfs_agnumber_t agno) static inline void xfs_perag_put(struct xfs_perag *pag) { - /* nothing to see here, move along */ + ASSERT(atomic_read(&pag->pag_ref) > 0); + atomic_dec(&pag->pag_ref); } /* -- cgit v0.10.2 From 0fa800fbd549736dfdc1d7761f87e33dc8cd973b Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:47:46 +0000 Subject: xfs: Add trace points for per-ag refcount debugging. Uninline xfs_perag_{get,put} so that tracepoints can be inserted into them to speed debugging of reference count problems. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_trace.h b/fs/xfs/linux-2.6/xfs_trace.h index 3353aef..1bb09e7 100644 --- a/fs/xfs/linux-2.6/xfs_trace.h +++ b/fs/xfs/linux-2.6/xfs_trace.h @@ -78,6 +78,33 @@ DECLARE_EVENT_CLASS(xfs_attr_list_class, ) ) +#define DEFINE_PERAG_REF_EVENT(name) \ +TRACE_EVENT(name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, int refcount, \ + unsigned long caller_ip), \ + TP_ARGS(mp, agno, refcount, caller_ip), \ + TP_STRUCT__entry( \ + __field(dev_t, dev) \ + __field(xfs_agnumber_t, agno) \ + __field(int, refcount) \ + __field(unsigned long, caller_ip) \ + ), \ + TP_fast_assign( \ + __entry->dev = mp->m_super->s_dev; \ + __entry->agno = agno; \ + __entry->refcount = refcount; \ + __entry->caller_ip = caller_ip; \ + ), \ + TP_printk("dev %d:%d agno %u refcount %d caller %pf", \ + MAJOR(__entry->dev), MINOR(__entry->dev), \ + __entry->agno, \ + __entry->refcount, \ + (char *)__entry->caller_ip) \ +); + +DEFINE_PERAG_REF_EVENT(xfs_perag_get) +DEFINE_PERAG_REF_EVENT(xfs_perag_put) + #define DEFINE_ATTR_LIST_EVENT(name) \ DEFINE_EVENT(xfs_attr_list_class, name, \ TP_PROTO(struct xfs_attr_list_context *ctx), \ diff --git a/fs/xfs/xfs_ag.h b/fs/xfs/xfs_ag.h index 18ae43f..963bc27 100644 --- a/fs/xfs/xfs_ag.h +++ b/fs/xfs/xfs_ag.h @@ -197,6 +197,8 @@ typedef struct xfs_perag_busy { #endif typedef struct xfs_perag { + struct xfs_mount *pag_mount; /* owner filesystem */ + xfs_agnumber_t pag_agno; /* AG this structure belongs to */ atomic_t pag_ref; /* perag reference count */ char pagf_init; /* this agf's entry is initialized */ char pagi_init; /* this agi's entry is initialized */ diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index f241fec..049dbc7 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -201,6 +201,38 @@ xfs_uuid_unmount( /* + * Reference counting access wrappers to the perag structures. + */ +struct xfs_perag * +xfs_perag_get(struct xfs_mount *mp, xfs_agnumber_t agno) +{ + struct xfs_perag *pag; + int ref = 0; + + spin_lock(&mp->m_perag_lock); + pag = radix_tree_lookup(&mp->m_perag_tree, agno); + if (pag) { + ASSERT(atomic_read(&pag->pag_ref) >= 0); + /* catch leaks in the positive direction during testing */ + ASSERT(atomic_read(&pag->pag_ref) < 1000); + ref = atomic_inc_return(&pag->pag_ref); + } + spin_unlock(&mp->m_perag_lock); + trace_xfs_perag_get(mp, agno, ref, _RET_IP_); + return pag; +} + +void +xfs_perag_put(struct xfs_perag *pag) +{ + int ref; + + ASSERT(atomic_read(&pag->pag_ref) > 0); + ref = atomic_dec_return(&pag->pag_ref); + trace_xfs_perag_put(pag->pag_mount, pag->pag_agno, ref, _RET_IP_); +} + +/* * Free up the resources associated with a mount structure. Assume that * the structure was initially zeroed, so we can tell which fields got * initialized. @@ -433,6 +465,8 @@ xfs_initialize_perag( kmem_free(pag); return -EEXIST; } + pag->pag_agno = index; + pag->pag_mount = mp; spin_unlock(&mp->m_perag_lock); radix_tree_preload_end(); } diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 16b2212..e62fd1c 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -386,29 +386,8 @@ xfs_daddr_to_agbno(struct xfs_mount *mp, xfs_daddr_t d) /* * perag get/put wrappers for ref counting */ -static inline struct xfs_perag * -xfs_perag_get(struct xfs_mount *mp, xfs_agnumber_t agno) -{ - struct xfs_perag *pag; - - spin_lock(&mp->m_perag_lock); - pag = radix_tree_lookup(&mp->m_perag_tree, agno); - if (pag) { - ASSERT(atomic_read(&pag->pag_ref) >= 0); - /* catch leaks in the positive direction during testing */ - ASSERT(atomic_read(&pag->pag_ref) < 1000); - atomic_inc(&pag->pag_ref); - } - spin_unlock(&mp->m_perag_lock); - return pag; -} - -static inline void -xfs_perag_put(struct xfs_perag *pag) -{ - ASSERT(atomic_read(&pag->pag_ref) > 0); - atomic_dec(&pag->pag_ref); -} +struct xfs_perag *xfs_perag_get(struct xfs_mount *mp, xfs_agnumber_t agno); +void xfs_perag_put(struct xfs_perag *pag); /* * Per-cpu superblock locking functions -- cgit v0.10.2 From b657fc82a3ca6d7ad16a59e81765f0fb0e86cdbb Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:47:47 +0000 Subject: xfs: Kill filestreams cache flush The filestreams cache flush is not needed in the sync code as it does not affect data writeback, and it is now not used by the growfs code, either, so kill it. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c index cc964fa..b58f841 100644 --- a/fs/xfs/linux-2.6/xfs_sync.c +++ b/fs/xfs/linux-2.6/xfs_sync.c @@ -451,9 +451,6 @@ xfs_quiesce_data( xfs_sync_data(mp, SYNC_WAIT); xfs_qm_sync(mp, SYNC_WAIT); - /* drop inode references pinned by filestreams */ - xfs_filestream_flush(mp); - /* write superblock and hoover up shutdown errors */ error = xfs_sync_fsdata(mp, SYNC_WAIT); diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c index 914d00d..390850e 100644 --- a/fs/xfs/xfs_filestream.c +++ b/fs/xfs/xfs_filestream.c @@ -455,20 +455,6 @@ xfs_filestream_unmount( } /* - * If the mount point's m_perag tree is going to be modified, all - * outstanding cache entries must be flushed to avoid accessing reference count - * addresses that have been freed. The call to xfs_filestream_flush() must be - * made inside the block that holds the m_perag_lock in write mode to do the - * reallocation. - */ -void -xfs_filestream_flush( - xfs_mount_t *mp) -{ - xfs_mru_cache_flush(mp->m_filestream); -} - -/* * Return the AG of the filestream the file or directory belongs to, or * NULLAGNUMBER otherwise. */ diff --git a/fs/xfs/xfs_filestream.h b/fs/xfs/xfs_filestream.h index 58378b2..260f757 100644 --- a/fs/xfs/xfs_filestream.h +++ b/fs/xfs/xfs_filestream.h @@ -135,7 +135,6 @@ int xfs_filestream_init(void); void xfs_filestream_uninit(void); int xfs_filestream_mount(struct xfs_mount *mp); void xfs_filestream_unmount(struct xfs_mount *mp); -void xfs_filestream_flush(struct xfs_mount *mp); xfs_agnumber_t xfs_filestream_lookup_ag(struct xfs_inode *ip); int xfs_filestream_associate(struct xfs_inode *dip, struct xfs_inode *ip); void xfs_filestream_deassociate(struct xfs_inode *ip); diff --git a/fs/xfs/xfs_mru_cache.c b/fs/xfs/xfs_mru_cache.c index 4b0613d..45ce15d 100644 --- a/fs/xfs/xfs_mru_cache.c +++ b/fs/xfs/xfs_mru_cache.c @@ -398,7 +398,7 @@ exit: * guaranteed that all the free functions for all the elements have finished * executing and the reaper is not running. */ -void +static void xfs_mru_cache_flush( xfs_mru_cache_t *mru) { diff --git a/fs/xfs/xfs_mru_cache.h b/fs/xfs/xfs_mru_cache.h index 5d439f3..36dd3ec 100644 --- a/fs/xfs/xfs_mru_cache.h +++ b/fs/xfs/xfs_mru_cache.h @@ -42,7 +42,6 @@ void xfs_mru_cache_uninit(void); int xfs_mru_cache_create(struct xfs_mru_cache **mrup, unsigned int lifetime_ms, unsigned int grp_count, xfs_mru_cache_free_func_t free_func); -void xfs_mru_cache_flush(xfs_mru_cache_t *mru); void xfs_mru_cache_destroy(struct xfs_mru_cache *mru); int xfs_mru_cache_insert(struct xfs_mru_cache *mru, unsigned long key, void *value); -- cgit v0.10.2 From 8b26c5825e023b1bccac7afd174ebe55b8905cb1 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:47:48 +0000 Subject: xfs: handle ENOMEM correctly during initialisation of perag structures Add proper error handling in case an error occurs while initializing new perag structures for a mount point. The mount structure is restored to its previous state by deleting and freeing any perag structures added during the call. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 049dbc7..be643e5 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -432,11 +432,13 @@ xfs_initialize_perag( xfs_agnumber_t *maxagi) { xfs_agnumber_t index, max_metadata; + xfs_agnumber_t first_initialised = 0; xfs_perag_t *pag; xfs_agino_t agino; xfs_ino_t ino; xfs_sb_t *sbp = &mp->m_sb; xfs_ino_t max_inum = XFS_MAXINUMBER_32; + int error = -ENOMEM; /* Check to see if the filesystem can overflow 32 bit inodes */ agino = XFS_OFFBNO_TO_AGINO(mp, sbp->sb_agblocks - 1, 0); @@ -453,17 +455,20 @@ xfs_initialize_perag( xfs_perag_put(pag); continue; } + if (!first_initialised) + first_initialised = index; pag = kmem_zalloc(sizeof(*pag), KM_MAYFAIL); if (!pag) - return -ENOMEM; + goto out_unwind; if (radix_tree_preload(GFP_NOFS)) - return -ENOMEM; + goto out_unwind; spin_lock(&mp->m_perag_lock); if (radix_tree_insert(&mp->m_perag_tree, index, pag)) { BUG(); spin_unlock(&mp->m_perag_lock); - kmem_free(pag); - return -EEXIST; + radix_tree_preload_end(); + error = -EEXIST; + goto out_unwind; } pag->pag_agno = index; pag->pag_mount = mp; @@ -523,6 +528,14 @@ xfs_initialize_perag( if (maxagi) *maxagi = index; return 0; + +out_unwind: + kmem_free(pag); + for (; index > first_initialised; index--) { + pag = radix_tree_delete(&mp->m_perag_tree, index); + kmem_free(pag); + } + return error; } void -- cgit v0.10.2 From e57336ff7fc7520bec7b3a7741043bdebaf622ea Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Mon, 11 Jan 2010 11:47:49 +0000 Subject: xfs: embed the pagb_list array in the perag structure Now that the perag structure is allocated memory rather than held in an array, we don't need to have the busy extent array external to the structure. Embed it into the perag structure to avoid needing an extra allocation when setting up. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/xfs_ag.h b/fs/xfs/xfs_ag.h index 963bc27..b1a5a1f 100644 --- a/fs/xfs/xfs_ag.h +++ b/fs/xfs/xfs_ag.h @@ -187,14 +187,8 @@ typedef struct xfs_perag_busy { /* * Per-ag incore structure, copies of information in agf and agi, * to improve the performance of allocation group selection. - * - * pick sizes which fit in allocation buckets well */ -#if (BITS_PER_LONG == 32) -#define XFS_PAGB_NUM_SLOTS 84 -#elif (BITS_PER_LONG == 64) #define XFS_PAGB_NUM_SLOTS 128 -#endif typedef struct xfs_perag { struct xfs_mount *pag_mount; /* owner filesystem */ @@ -212,8 +206,6 @@ typedef struct xfs_perag { __uint32_t pagf_btreeblks; /* # of blocks held in AGF btrees */ xfs_agino_t pagi_freecount; /* number of free inodes */ xfs_agino_t pagi_count; /* number of allocated inodes */ - int pagb_count; /* pagb slots in use */ - xfs_perag_busy_t *pagb_list; /* unstable blocks */ /* * Inode allocation search lookup optimisation. @@ -232,6 +224,8 @@ typedef struct xfs_perag { rwlock_t pag_ici_lock; /* incore inode lock */ struct radix_tree_root pag_ici_root; /* incore inode cache root */ #endif + int pagb_count; /* pagb slots in use */ + xfs_perag_busy_t pagb_list[XFS_PAGB_NUM_SLOTS]; /* unstable blocks */ } xfs_perag_t; /* diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index 4d66bb7..8aa181d 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -2200,8 +2200,8 @@ xfs_alloc_read_agf( pag->pagf_levels[XFS_BTNUM_CNTi] = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]); spin_lock_init(&pag->pagb_lock); - pag->pagb_list = kmem_zalloc(XFS_PAGB_NUM_SLOTS * - sizeof(xfs_perag_busy_t), KM_SLEEP); + pag->pagb_count = 0; + memset(pag->pagb_list, 0, sizeof(pag->pagb_list)); pag->pagf_init = 1; } #ifdef DEBUG diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index be643e5..0df5045 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -247,10 +247,9 @@ xfs_free_perag( for (agno = 0; agno < mp->m_sb.sb_agcount; agno++) { spin_lock(&mp->m_perag_lock); pag = radix_tree_delete(&mp->m_perag_tree, agno); + ASSERT(pag); ASSERT(atomic_read(&pag->pag_ref) == 0); spin_unlock(&mp->m_perag_lock); - ASSERT(pag); - kmem_free(pag->pagb_list); kmem_free(pag); } } -- cgit v0.10.2 From 873ff5501d8cd1a21045d6c1da34f0c3876bc235 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 13 Jan 2010 22:17:57 +0000 Subject: xfs: clean up log buffer writes Don't bother using XFS_bwrite as it doesn't provide much code for our use case. Instead opencode it and fold xlog_bdstrat_cb into the new xlog_bdstrat helper. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h index a34c7b5..c20a760 100644 --- a/fs/xfs/linux-2.6/xfs_buf.h +++ b/fs/xfs/linux-2.6/xfs_buf.h @@ -408,8 +408,6 @@ static inline int XFS_bwrite(xfs_buf_t *bp) return error; } -#define XFS_bdstrat(bp) xfs_buf_iorequest(bp) - #define xfs_iowait(bp) xfs_buf_iowait(bp) #define xfs_baread(target, rablkno, ralen) \ diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 600b5b0..0d17516 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -50,7 +50,6 @@ kmem_zone_t *xfs_log_ticket_zone; (off) += (bytes);} /* Local miscellaneous function prototypes */ -STATIC int xlog_bdstrat_cb(struct xfs_buf *); STATIC int xlog_commit_record(xfs_mount_t *mp, xlog_ticket_t *ticket, xlog_in_core_t **, xfs_lsn_t *); STATIC xlog_t * xlog_alloc_log(xfs_mount_t *mp, @@ -988,35 +987,6 @@ xlog_iodone(xfs_buf_t *bp) } /* xlog_iodone */ /* - * The bdstrat callback function for log bufs. This gives us a central - * place to trap bufs in case we get hit by a log I/O error and need to - * shutdown. Actually, in practice, even when we didn't get a log error, - * we transition the iclogs to IOERROR state *after* flushing all existing - * iclogs to disk. This is because we don't want anymore new transactions to be - * started or completed afterwards. - */ -STATIC int -xlog_bdstrat_cb(struct xfs_buf *bp) -{ - xlog_in_core_t *iclog; - - iclog = XFS_BUF_FSPRIVATE(bp, xlog_in_core_t *); - - if ((iclog->ic_state & XLOG_STATE_IOERROR) == 0) { - /* note for irix bstrat will need struct bdevsw passed - * Fix the following macro if the code ever is merged - */ - XFS_bdstrat(bp); - return 0; - } - - XFS_BUF_ERROR(bp, EIO); - XFS_BUF_STALE(bp); - xfs_biodone(bp); - return XFS_ERROR(EIO); -} - -/* * Return size of each in-core log record buffer. * * All machines get 8 x 32kB buffers by default, unless tuned otherwise. @@ -1158,7 +1128,6 @@ xlog_alloc_log(xfs_mount_t *mp, if (!bp) goto out_free_log; XFS_BUF_SET_IODONE_FUNC(bp, xlog_iodone); - XFS_BUF_SET_BDSTRAT_FUNC(bp, xlog_bdstrat_cb); XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)1); ASSERT(XFS_BUF_ISBUSY(bp)); ASSERT(XFS_BUF_VALUSEMA(bp) <= 0); @@ -1196,7 +1165,6 @@ xlog_alloc_log(xfs_mount_t *mp, if (!XFS_BUF_CPSEMA(bp)) ASSERT(0); XFS_BUF_SET_IODONE_FUNC(bp, xlog_iodone); - XFS_BUF_SET_BDSTRAT_FUNC(bp, xlog_bdstrat_cb); XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)1); iclog->ic_bp = bp; iclog->ic_data = bp->b_addr; @@ -1343,6 +1311,37 @@ xlog_grant_push_ail(xfs_mount_t *mp, xfs_trans_ail_push(log->l_ailp, threshold_lsn); } /* xlog_grant_push_ail */ +/* + * The bdstrat callback function for log bufs. This gives us a central + * place to trap bufs in case we get hit by a log I/O error and need to + * shutdown. Actually, in practice, even when we didn't get a log error, + * we transition the iclogs to IOERROR state *after* flushing all existing + * iclogs to disk. This is because we don't want anymore new transactions to be + * started or completed afterwards. + */ +STATIC int +xlog_bdstrat( + struct xfs_buf *bp) +{ + struct xlog_in_core *iclog; + + iclog = XFS_BUF_FSPRIVATE(bp, xlog_in_core_t *); + if (iclog->ic_state & XLOG_STATE_IOERROR) { + XFS_BUF_ERROR(bp, EIO); + XFS_BUF_STALE(bp); + xfs_biodone(bp); + /* + * It would seem logical to return EIO here, but we rely on + * the log state machine to propagate I/O errors instead of + * doing it here. + */ + return 0; + } + + bp->b_flags |= _XBF_RUN_QUEUES; + xfs_buf_iorequest(bp); + return 0; +} /* * Flush out the in-core log (iclog) to the on-disk log in an asynchronous @@ -1462,7 +1461,7 @@ xlog_sync(xlog_t *log, */ XFS_BUF_WRITE(bp); - if ((error = XFS_bwrite(bp))) { + if ((error = xlog_bdstrat(bp))) { xfs_ioerror_alert("xlog_sync", log->l_mp, bp, XFS_BUF_ADDR(bp)); return error; @@ -1502,7 +1501,7 @@ xlog_sync(xlog_t *log, /* account for internal log which doesn't start at block #0 */ XFS_BUF_SET_ADDR(bp, XFS_BUF_ADDR(bp) + log->l_logBBstart); XFS_BUF_WRITE(bp); - if ((error = XFS_bwrite(bp))) { + if ((error = xlog_bdstrat(bp))) { xfs_ioerror_alert("xlog_sync (split)", log->l_mp, bp, XFS_BUF_ADDR(bp)); return error; -- cgit v0.10.2 From 64e0bc7d2a6609ad265757a600e2a0d93c8adb47 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 13 Jan 2010 22:17:58 +0000 Subject: xfs: clean up xfs_bwrite Fold XFS_bwrite into it's only caller, xfs_bwrite and move it into xfs_buf.c instead of leaving it as a fairly large inline function. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index 18ae3ba..492465c 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -1051,6 +1051,33 @@ xfs_buf_ioerror( } int +xfs_bwrite( + struct xfs_mount *mp, + struct xfs_buf *bp) +{ + int iowait = (bp->b_flags & XBF_ASYNC) == 0; + int error = 0; + + bp->b_strat = xfs_bdstrat_cb; + bp->b_mount = mp; + bp->b_flags |= XBF_WRITE; + if (!iowait) + bp->b_flags |= _XBF_RUN_QUEUES; + + xfs_buf_delwri_dequeue(bp); + xfs_buf_iostrategy(bp); + + if (iowait) { + error = xfs_buf_iowait(bp); + if (error) + xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR); + xfs_buf_relse(bp); + } + + return error; +} + +int xfs_bawrite( void *mp, struct xfs_buf *bp) diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h index c20a760..f69b8e7 100644 --- a/fs/xfs/linux-2.6/xfs_buf.h +++ b/fs/xfs/linux-2.6/xfs_buf.h @@ -232,6 +232,7 @@ extern void xfs_buf_lock(xfs_buf_t *); extern void xfs_buf_unlock(xfs_buf_t *); /* Buffer Read and Write Routines */ +extern int xfs_bwrite(struct xfs_mount *mp, struct xfs_buf *bp); extern int xfs_bawrite(void *mp, xfs_buf_t *bp); extern void xfs_bdwrite(void *mp, xfs_buf_t *bp); extern void xfs_buf_ioend(xfs_buf_t *, int); @@ -390,24 +391,6 @@ static inline void xfs_buf_relse(xfs_buf_t *bp) #define xfs_biozero(bp, off, len) \ xfs_buf_iomove((bp), (off), (len), NULL, XBRW_ZERO) - -static inline int XFS_bwrite(xfs_buf_t *bp) -{ - int iowait = (bp->b_flags & XBF_ASYNC) == 0; - int error = 0; - - if (!iowait) - bp->b_flags |= _XBF_RUN_QUEUES; - - xfs_buf_delwri_dequeue(bp); - xfs_buf_iostrategy(bp); - if (iowait) { - error = xfs_buf_iowait(bp); - xfs_buf_relse(bp); - } - return error; -} - #define xfs_iowait(bp) xfs_buf_iowait(bp) #define xfs_baread(target, rablkno, ralen) \ diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c index 5aa07ca..9d933a1 100644 --- a/fs/xfs/xfs_rw.c +++ b/fs/xfs/xfs_rw.c @@ -306,37 +306,6 @@ xfs_read_buf( } /* - * Wrapper around bwrite() so that we can trap - * write errors, and act accordingly. - */ -int -xfs_bwrite( - struct xfs_mount *mp, - struct xfs_buf *bp) -{ - int error; - - /* - * XXXsup how does this work for quotas. - */ - XFS_BUF_SET_BDSTRAT_FUNC(bp, xfs_bdstrat_cb); - bp->b_mount = mp; - XFS_BUF_WRITE(bp); - - if ((error = XFS_bwrite(bp))) { - ASSERT(mp); - /* - * Cannot put a buftrace here since if the buffer is not - * B_HOLD then we will brelse() the buffer before returning - * from bwrite and we could be tracing a buffer that has - * been reused. - */ - xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR); - } - return (error); -} - -/* * helper function to extract extent size hint from inode */ xfs_extlen_t diff --git a/fs/xfs/xfs_rw.h b/fs/xfs/xfs_rw.h index 571f217..ff68eb5 100644 --- a/fs/xfs/xfs_rw.h +++ b/fs/xfs/xfs_rw.h @@ -40,7 +40,6 @@ xfs_fsb_to_db(struct xfs_inode *ip, xfs_fsblock_t fsb) * Prototypes for functions in xfs_rw.c. */ extern int xfs_write_clear_setuid(struct xfs_inode *ip); -extern int xfs_bwrite(struct xfs_mount *mp, struct xfs_buf *bp); extern int xfs_bioerror(struct xfs_buf *bp); extern int xfs_bioerror_relse(struct xfs_buf *bp); extern int xfs_read_buf(struct xfs_mount *mp, xfs_buftarg_t *btp, -- cgit v0.10.2 From 4e23471a3f3aba885ea70100db47ccacb5f069f6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 13 Jan 2010 22:17:56 +0000 Subject: xfs: move more buffer helpers into xfs_buf.c Move xfsbdstrat and xfs_bdstrat_cb from xfs_lrw.c and xfs_bioerror and xfs_bioerror_relse from xfs_rw.c into xfs_buf.c. This also means xfs_bioerror and xfs_bioerror_relse can be marked static now. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index 492465c..158fad4 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -1112,6 +1112,126 @@ xfs_bdwrite( xfs_buf_delwri_queue(bp, 1); } +/* + * Called when we want to stop a buffer from getting written or read. + * We attach the EIO error, muck with its flags, and call biodone + * so that the proper iodone callbacks get called. + */ +STATIC int +xfs_bioerror( + xfs_buf_t *bp) +{ +#ifdef XFSERRORDEBUG + ASSERT(XFS_BUF_ISREAD(bp) || bp->b_iodone); +#endif + + /* + * No need to wait until the buffer is unpinned, we aren't flushing it. + */ + XFS_BUF_ERROR(bp, EIO); + + /* + * We're calling biodone, so delete XBF_DONE flag. + */ + XFS_BUF_UNREAD(bp); + XFS_BUF_UNDELAYWRITE(bp); + XFS_BUF_UNDONE(bp); + XFS_BUF_STALE(bp); + + XFS_BUF_CLR_BDSTRAT_FUNC(bp); + xfs_biodone(bp); + + return EIO; +} + +/* + * Same as xfs_bioerror, except that we are releasing the buffer + * here ourselves, and avoiding the biodone call. + * This is meant for userdata errors; metadata bufs come with + * iodone functions attached, so that we can track down errors. + */ +STATIC int +xfs_bioerror_relse( + struct xfs_buf *bp) +{ + int64_t fl = XFS_BUF_BFLAGS(bp); + /* + * No need to wait until the buffer is unpinned. + * We aren't flushing it. + * + * chunkhold expects B_DONE to be set, whether + * we actually finish the I/O or not. We don't want to + * change that interface. + */ + XFS_BUF_UNREAD(bp); + XFS_BUF_UNDELAYWRITE(bp); + XFS_BUF_DONE(bp); + XFS_BUF_STALE(bp); + XFS_BUF_CLR_IODONE_FUNC(bp); + XFS_BUF_CLR_BDSTRAT_FUNC(bp); + if (!(fl & XFS_B_ASYNC)) { + /* + * Mark b_error and B_ERROR _both_. + * Lot's of chunkcache code assumes that. + * There's no reason to mark error for + * ASYNC buffers. + */ + XFS_BUF_ERROR(bp, EIO); + XFS_BUF_FINISH_IOWAIT(bp); + } else { + xfs_buf_relse(bp); + } + + return EIO; +} + + +/* + * All xfs metadata buffers except log state machine buffers + * get this attached as their b_bdstrat callback function. + * This is so that we can catch a buffer + * after prematurely unpinning it to forcibly shutdown the filesystem. + */ +int +xfs_bdstrat_cb( + struct xfs_buf *bp) +{ + if (XFS_FORCED_SHUTDOWN(bp->b_mount)) { + trace_xfs_bdstrat_shut(bp, _RET_IP_); + /* + * Metadata write that didn't get logged but + * written delayed anyway. These aren't associated + * with a transaction, and can be ignored. + */ + if (!bp->b_iodone && !XFS_BUF_ISREAD(bp)) + return xfs_bioerror_relse(bp); + else + return xfs_bioerror(bp); + } + + xfs_buf_iorequest(bp); + return 0; +} + +/* + * Wrapper around bdstrat so that we can stop data from going to disk in case + * we are shutting down the filesystem. Typically user data goes thru this + * path; one of the exceptions is the superblock. + */ +void +xfsbdstrat( + struct xfs_mount *mp, + struct xfs_buf *bp) +{ + if (XFS_FORCED_SHUTDOWN(mp)) { + trace_xfs_bdstrat_shut(bp, _RET_IP_); + xfs_bioerror_relse(bp); + return; + } + + xfs_buf_iorequest(bp); +} + STATIC void _xfs_buf_ioend( xfs_buf_t *bp, diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h index f69b8e7..9a29d18 100644 --- a/fs/xfs/linux-2.6/xfs_buf.h +++ b/fs/xfs/linux-2.6/xfs_buf.h @@ -235,6 +235,10 @@ extern void xfs_buf_unlock(xfs_buf_t *); extern int xfs_bwrite(struct xfs_mount *mp, struct xfs_buf *bp); extern int xfs_bawrite(void *mp, xfs_buf_t *bp); extern void xfs_bdwrite(void *mp, xfs_buf_t *bp); + +extern void xfsbdstrat(struct xfs_mount *, struct xfs_buf *); +extern int xfs_bdstrat_cb(struct xfs_buf *); + extern void xfs_buf_ioend(xfs_buf_t *, int); extern void xfs_buf_ioerror(xfs_buf_t *, int); extern int xfs_buf_iorequest(xfs_buf_t *); diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c index 0d32457..c80fa00d 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.c +++ b/fs/xfs/linux-2.6/xfs_lrw.c @@ -784,53 +784,6 @@ write_retry: } /* - * All xfs metadata buffers except log state machine buffers - * get this attached as their b_bdstrat callback function. - * This is so that we can catch a buffer - * after prematurely unpinning it to forcibly shutdown the filesystem. - */ -int -xfs_bdstrat_cb(struct xfs_buf *bp) -{ - if (XFS_FORCED_SHUTDOWN(bp->b_mount)) { - trace_xfs_bdstrat_shut(bp, _RET_IP_); - /* - * Metadata write that didn't get logged but - * written delayed anyway. These aren't associated - * with a transaction, and can be ignored. - */ - if (XFS_BUF_IODONE_FUNC(bp) == NULL && - (XFS_BUF_ISREAD(bp)) == 0) - return (xfs_bioerror_relse(bp)); - else - return (xfs_bioerror(bp)); - } - - xfs_buf_iorequest(bp); - return 0; -} - -/* - * Wrapper around bdstrat so that we can stop data from going to disk in case - * we are shutting down the filesystem. Typically user data goes thru this - * path; one of the exceptions is the superblock. - */ -void -xfsbdstrat( - struct xfs_mount *mp, - struct xfs_buf *bp) -{ - ASSERT(mp); - if (!XFS_FORCED_SHUTDOWN(mp)) { - xfs_buf_iorequest(bp); - return; - } - - trace_xfs_bdstrat_shut(bp, _RET_IP_); - xfs_bioerror_relse(bp); -} - -/* * If the underlying (data/log/rt) device is readonly, there are some * operations that cannot proceed. */ diff --git a/fs/xfs/linux-2.6/xfs_lrw.h b/fs/xfs/linux-2.6/xfs_lrw.h index d1f7789..342ae8c 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.h +++ b/fs/xfs/linux-2.6/xfs_lrw.h @@ -22,9 +22,6 @@ struct xfs_mount; struct xfs_inode; struct xfs_buf; -/* errors from xfsbdstrat() must be extracted from the buffer */ -extern void xfsbdstrat(struct xfs_mount *, struct xfs_buf *); -extern int xfs_bdstrat_cb(struct xfs_buf *); extern int xfs_dev_is_read_only(struct xfs_mount *, char *); extern int xfs_zero_eof(struct xfs_inode *, xfs_off_t, xfs_fsize_t); diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c index 9d933a1..abb2c45 100644 --- a/fs/xfs/xfs_rw.c +++ b/fs/xfs/xfs_rw.c @@ -153,88 +153,6 @@ xfs_do_force_shutdown( } } - -/* - * Called when we want to stop a buffer from getting written or read. - * We attach the EIO error, muck with its flags, and call biodone - * so that the proper iodone callbacks get called. - */ -int -xfs_bioerror( - xfs_buf_t *bp) -{ - -#ifdef XFSERRORDEBUG - ASSERT(XFS_BUF_ISREAD(bp) || bp->b_iodone); -#endif - - /* - * No need to wait until the buffer is unpinned. - * We aren't flushing it. - */ - XFS_BUF_ERROR(bp, EIO); - /* - * We're calling biodone, so delete B_DONE flag. Either way - * we have to call the iodone callback, and calling biodone - * probably is the best way since it takes care of - * GRIO as well. - */ - XFS_BUF_UNREAD(bp); - XFS_BUF_UNDELAYWRITE(bp); - XFS_BUF_UNDONE(bp); - XFS_BUF_STALE(bp); - - XFS_BUF_CLR_BDSTRAT_FUNC(bp); - xfs_biodone(bp); - - return (EIO); -} - -/* - * Same as xfs_bioerror, except that we are releasing the buffer - * here ourselves, and avoiding the biodone call. - * This is meant for userdata errors; metadata bufs come with - * iodone functions attached, so that we can track down errors. - */ -int -xfs_bioerror_relse( - xfs_buf_t *bp) -{ - int64_t fl; - - ASSERT(XFS_BUF_IODONE_FUNC(bp) != xfs_buf_iodone_callbacks); - ASSERT(XFS_BUF_IODONE_FUNC(bp) != xlog_iodone); - - fl = XFS_BUF_BFLAGS(bp); - /* - * No need to wait until the buffer is unpinned. - * We aren't flushing it. - * - * chunkhold expects B_DONE to be set, whether - * we actually finish the I/O or not. We don't want to - * change that interface. - */ - XFS_BUF_UNREAD(bp); - XFS_BUF_UNDELAYWRITE(bp); - XFS_BUF_DONE(bp); - XFS_BUF_STALE(bp); - XFS_BUF_CLR_IODONE_FUNC(bp); - XFS_BUF_CLR_BDSTRAT_FUNC(bp); - if (!(fl & XFS_B_ASYNC)) { - /* - * Mark b_error and B_ERROR _both_. - * Lot's of chunkcache code assumes that. - * There's no reason to mark error for - * ASYNC buffers. - */ - XFS_BUF_ERROR(bp, EIO); - XFS_BUF_FINISH_IOWAIT(bp); - } else { - xfs_buf_relse(bp); - } - return (EIO); -} - /* * Prints out an ALERT message about I/O error. */ diff --git a/fs/xfs/xfs_rw.h b/fs/xfs/xfs_rw.h index ff68eb5..a54c3b7 100644 --- a/fs/xfs/xfs_rw.h +++ b/fs/xfs/xfs_rw.h @@ -40,8 +40,6 @@ xfs_fsb_to_db(struct xfs_inode *ip, xfs_fsblock_t fsb) * Prototypes for functions in xfs_rw.c. */ extern int xfs_write_clear_setuid(struct xfs_inode *ip); -extern int xfs_bioerror(struct xfs_buf *bp); -extern int xfs_bioerror_relse(struct xfs_buf *bp); extern int xfs_read_buf(struct xfs_mount *mp, xfs_buftarg_t *btp, xfs_daddr_t blkno, int len, uint flags, struct xfs_buf **bpp); -- cgit v0.10.2 From 4c54005ca438a8b46dd542b497d4f0dc2ca375e8 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 14 Jan 2010 16:10:57 -0800 Subject: rcu: 1Q2010 update for RCU documentation Add expedited functions. Review documentation and update obsolete verbiage. Also fix the advice for the RCU CPU-stall kernel configuration parameter, and document RCU CPU-stall warnings. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12635142581866-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/Documentation/RCU/00-INDEX b/Documentation/RCU/00-INDEX index 9bb62f7..0a27ea9 100644 --- a/Documentation/RCU/00-INDEX +++ b/Documentation/RCU/00-INDEX @@ -8,14 +8,18 @@ listRCU.txt - Using RCU to Protect Read-Mostly Linked Lists NMI-RCU.txt - Using RCU to Protect Dynamic NMI Handlers +rcubarrier.txt + - RCU and Unloadable Modules +rculist_nulls.txt + - RCU list primitives for use with SLAB_DESTROY_BY_RCU rcuref.txt - Reference-count design for elements of lists/arrays protected by RCU rcu.txt - RCU Concepts -rcubarrier.txt - - Unloading modules that use RCU callbacks RTFP.txt - List of RCU papers (bibliography) going back to 1980. +stallwarn.txt + - RCU CPU stall warnings (CONFIG_RCU_CPU_STALL_DETECTOR) torture.txt - RCU Torture Test Operation (CONFIG_RCU_TORTURE_TEST) trace.txt diff --git a/Documentation/RCU/RTFP.txt b/Documentation/RCU/RTFP.txt index d2b8523..5051209 100644 --- a/Documentation/RCU/RTFP.txt +++ b/Documentation/RCU/RTFP.txt @@ -25,10 +25,10 @@ to be referencing the data structure. However, this mechanism was not optimized for modern computer systems, which is not surprising given that these overheads were not so expensive in the mid-80s. Nonetheless, passive serialization appears to be the first deferred-destruction -mechanism to be used in production. Furthermore, the relevant patent has -lapsed, so this approach may be used in non-GPL software, if desired. -(In contrast, use of RCU is permitted only in software licensed under -GPL. Sorry!!!) +mechanism to be used in production. Furthermore, the relevant patent +has lapsed, so this approach may be used in non-GPL software, if desired. +(In contrast, implementation of RCU is permitted only in software licensed +under either GPL or LGPL. Sorry!!!) In 1990, Pugh [Pugh90] noted that explicitly tracking which threads were reading a given data structure permitted deferred free to operate @@ -150,6 +150,18 @@ preemptible RCU [PaulEMcKenney2007PreemptibleRCU], and the three-part LWN "What is RCU?" series [PaulEMcKenney2007WhatIsRCUFundamentally, PaulEMcKenney2008WhatIsRCUUsage, and PaulEMcKenney2008WhatIsRCUAPI]. +2008 saw a journal paper on real-time RCU [DinakarGuniguntala2008IBMSysJ], +a history of how Linux changed RCU more than RCU changed Linux +[PaulEMcKenney2008RCUOSR], and a design overview of hierarchical RCU +[PaulEMcKenney2008HierarchicalRCU]. + +2009 introduced user-level RCU algorithms [PaulEMcKenney2009MaliciousURCU], +which Mathieu Desnoyers is now maintaining [MathieuDesnoyers2009URCU] +[MathieuDesnoyersPhD]. TINY_RCU [PaulEMcKenney2009BloatWatchRCU] made +its appearance, as did expedited RCU [PaulEMcKenney2009expeditedRCU]. +The problem of resizeable RCU-protected hash tables may now be on a path +to a solution [JoshTriplett2009RPHash]. + Bibtex Entries @article{Kung80 @@ -730,6 +742,11 @@ Revised: " } +# +# "What is RCU?" LWN series. +# +######################################################################## + @article{DinakarGuniguntala2008IBMSysJ ,author="D. Guniguntala and P. E. McKenney and J. Triplett and J. Walpole" ,title="The read-copy-update mechanism for supporting real-time applications on shared-memory multiprocessor systems with {Linux}" @@ -820,3 +837,36 @@ Revised: Uniprocessor assumptions allow simplified RCU implementation. " } + +@unpublished{PaulEMcKenney2009expeditedRCU +,Author="Paul E. McKenney" +,Title="[{PATCH} -tip 0/3] expedited 'big hammer' {RCU} grace periods" +,month="June" +,day="25" +,year="2009" +,note="Available: +\url{http://lkml.org/lkml/2009/6/25/306} +[Viewed August 16, 2009]" +,annotation=" + First posting of expedited RCU to be accepted into -tip. +" +} + +@unpublished{JoshTriplett2009RPHash +,Author="Josh Triplett" +,Title="Scalable concurrent hash tables via relativistic programming" +,month="September" +,year="2009" +,note="Linux Plumbers Conference presentation" +,annotation=" + RP fun with hash tables. +" +} + +@phdthesis{MathieuDesnoyersPhD +, title = "Low-impact Operating System Tracing" +, author = "Mathieu Desnoyers" +, school = "Ecole Polytechnique de Montr\'{e}al" +, month = "December" +, year = 2009 +} diff --git a/Documentation/RCU/checklist.txt b/Documentation/RCU/checklist.txt index 51525a3..767cf06 100644 --- a/Documentation/RCU/checklist.txt +++ b/Documentation/RCU/checklist.txt @@ -8,13 +8,12 @@ would cause. This list is based on experiences reviewing such patches over a rather long period of time, but improvements are always welcome! 0. Is RCU being applied to a read-mostly situation? If the data - structure is updated more than about 10% of the time, then - you should strongly consider some other approach, unless - detailed performance measurements show that RCU is nonetheless - the right tool for the job. Yes, you might think of RCU - as simply cutting overhead off of the readers and imposing it - on the writers. That is exactly why normal uses of RCU will - do much more reading than updating. + structure is updated more than about 10% of the time, then you + should strongly consider some other approach, unless detailed + performance measurements show that RCU is nonetheless the right + tool for the job. Yes, RCU does reduce read-side overhead by + increasing write-side overhead, which is exactly why normal uses + of RCU will do much more reading than updating. Another exception is where performance is not an issue, and RCU provides a simpler implementation. An example of this situation @@ -35,13 +34,13 @@ over a rather long period of time, but improvements are always welcome! If you choose #b, be prepared to describe how you have handled memory barriers on weakly ordered machines (pretty much all of - them -- even x86 allows reads to be reordered), and be prepared - to explain why this added complexity is worthwhile. If you - choose #c, be prepared to explain how this single task does not - become a major bottleneck on big multiprocessor machines (for - example, if the task is updating information relating to itself - that other tasks can read, there by definition can be no - bottleneck). + them -- even x86 allows later loads to be reordered to precede + earlier stores), and be prepared to explain why this added + complexity is worthwhile. If you choose #c, be prepared to + explain how this single task does not become a major bottleneck on + big multiprocessor machines (for example, if the task is updating + information relating to itself that other tasks can read, there + by definition can be no bottleneck). 2. Do the RCU read-side critical sections make proper use of rcu_read_lock() and friends? These primitives are needed @@ -51,8 +50,10 @@ over a rather long period of time, but improvements are always welcome! actuarial risk of your kernel. As a rough rule of thumb, any dereference of an RCU-protected - pointer must be covered by rcu_read_lock() or rcu_read_lock_bh() - or by the appropriate update-side lock. + pointer must be covered by rcu_read_lock(), rcu_read_lock_bh(), + rcu_read_lock_sched(), or by the appropriate update-side lock. + Disabling of preemption can serve as rcu_read_lock_sched(), but + is less readable. 3. Does the update code tolerate concurrent accesses? @@ -62,25 +63,27 @@ over a rather long period of time, but improvements are always welcome! of ways to handle this concurrency, depending on the situation: a. Use the RCU variants of the list and hlist update - primitives to add, remove, and replace elements on an - RCU-protected list. Alternatively, use the RCU-protected - trees that have been added to the Linux kernel. + primitives to add, remove, and replace elements on + an RCU-protected list. Alternatively, use the other + RCU-protected data structures that have been added to + the Linux kernel. This is almost always the best approach. b. Proceed as in (a) above, but also maintain per-element locks (that are acquired by both readers and writers) that guard per-element state. Of course, fields that - the readers refrain from accessing can be guarded by the - update-side lock. + the readers refrain from accessing can be guarded by + some other lock acquired only by updaters, if desired. This works quite well, also. c. Make updates appear atomic to readers. For example, - pointer updates to properly aligned fields will appear - atomic, as will individual atomic primitives. Operations - performed under a lock and sequences of multiple atomic - primitives will -not- appear to be atomic. + pointer updates to properly aligned fields will + appear atomic, as will individual atomic primitives. + Sequences of perations performed under a lock will -not- + appear to be atomic to RCU readers, nor will sequences + of multiple atomic primitives. This can work, but is starting to get a bit tricky. @@ -98,9 +101,9 @@ over a rather long period of time, but improvements are always welcome! a new structure containing updated values. 4. Weakly ordered CPUs pose special challenges. Almost all CPUs - are weakly ordered -- even i386 CPUs allow reads to be reordered. - RCU code must take all of the following measures to prevent - memory-corruption problems: + are weakly ordered -- even x86 CPUs allow later loads to be + reordered to precede earlier stores. RCU code must take all of + the following measures to prevent memory-corruption problems: a. Readers must maintain proper ordering of their memory accesses. The rcu_dereference() primitive ensures that @@ -113,14 +116,21 @@ over a rather long period of time, but improvements are always welcome! The rcu_dereference() primitive is also an excellent documentation aid, letting the person reading the code know exactly which pointers are protected by RCU. - - The rcu_dereference() primitive is used by the various - "_rcu()" list-traversal primitives, such as the - list_for_each_entry_rcu(). Note that it is perfectly - legal (if redundant) for update-side code to use - rcu_dereference() and the "_rcu()" list-traversal - primitives. This is particularly useful in code - that is common to readers and updaters. + Please note that compilers can also reorder code, and + they are becoming increasingly aggressive about doing + just that. The rcu_dereference() primitive therefore + also prevents destructive compiler optimizations. + + The rcu_dereference() primitive is used by the + various "_rcu()" list-traversal primitives, such + as the list_for_each_entry_rcu(). Note that it is + perfectly legal (if redundant) for update-side code to + use rcu_dereference() and the "_rcu()" list-traversal + primitives. This is particularly useful in code that + is common to readers and updaters. However, neither + rcu_dereference() nor the "_rcu()" list-traversal + primitives can substitute for a good concurrency design + coordinating among multiple updaters. b. If the list macros are being used, the list_add_tail_rcu() and list_add_rcu() primitives must be used in order @@ -135,11 +145,14 @@ over a rather long period of time, but improvements are always welcome! readers. Similarly, if the hlist macros are being used, the hlist_del_rcu() primitive is required. - The list_replace_rcu() primitive may be used to - replace an old structure with a new one in an - RCU-protected list. + The list_replace_rcu() and hlist_replace_rcu() primitives + may be used to replace an old structure with a new one + in their respective types of RCU-protected lists. + + d. Rules similar to (4b) and (4c) apply to the "hlist_nulls" + type of RCU-protected linked lists. - d. Updates must ensure that initialization of a given + e. Updates must ensure that initialization of a given structure happens before pointers to that structure are publicized. Use the rcu_assign_pointer() primitive when publicizing a pointer to a structure that can @@ -151,16 +164,31 @@ over a rather long period of time, but improvements are always welcome! it cannot block. 6. Since synchronize_rcu() can block, it cannot be called from - any sort of irq context. Ditto for synchronize_sched() and - synchronize_srcu(). - -7. If the updater uses call_rcu(), then the corresponding readers - must use rcu_read_lock() and rcu_read_unlock(). If the updater - uses call_rcu_bh(), then the corresponding readers must use - rcu_read_lock_bh() and rcu_read_unlock_bh(). If the updater - uses call_rcu_sched(), then the corresponding readers must - disable preemption. Mixing things up will result in confusion - and broken kernels. + any sort of irq context. The same rule applies for + synchronize_rcu_bh(), synchronize_sched(), synchronize_srcu(), + synchronize_rcu_expedited(), synchronize_rcu_bh_expedited(), + synchronize_sched_expedite(), and synchronize_srcu_expedited(). + + The expedited forms of these primitives have the same semantics + as the non-expedited forms, but expediting is both expensive + and unfriendly to real-time workloads. Use of the expedited + primitives should be restricted to rare configuration-change + operations that would not normally be undertaken while a real-time + workload is running. + +7. If the updater uses call_rcu() or synchronize_rcu(), then the + corresponding readers must use rcu_read_lock() and + rcu_read_unlock(). If the updater uses call_rcu_bh() or + synchronize_rcu_bh(), then the corresponding readers must + use rcu_read_lock_bh() and rcu_read_unlock_bh(). If the + updater uses call_rcu_sched() or synchronize_sched(), then + the corresponding readers must disable preemption, possibly + by calling rcu_read_lock_sched() and rcu_read_unlock_sched(). + If the updater uses synchronize_srcu(), the the corresponding + readers must use srcu_read_lock() and srcu_read_unlock(), + and with the same srcu_struct. The rules for the expedited + primitives are the same as for their non-expedited counterparts. + Mixing things up will result in confusion and broken kernels. One exception to this rule: rcu_read_lock() and rcu_read_unlock() may be substituted for rcu_read_lock_bh() and rcu_read_unlock_bh() @@ -212,6 +240,8 @@ over a rather long period of time, but improvements are always welcome! e. Periodically invoke synchronize_rcu(), permitting a limited number of updates per grace period. + The same cautions apply to call_rcu_bh() and call_rcu_sched(). + 9. All RCU list-traversal primitives, which include rcu_dereference(), list_for_each_entry_rcu(), list_for_each_continue_rcu(), and list_for_each_safe_rcu(), @@ -229,7 +259,8 @@ over a rather long period of time, but improvements are always welcome! 10. Conversely, if you are in an RCU read-side critical section, and you don't hold the appropriate update-side lock, you -must- use the "_rcu()" variants of the list macros. Failing to do so - will break Alpha and confuse people reading your code. + will break Alpha, cause aggressive compilers to generate bad code, + and confuse people trying to read your code. 11. Note that synchronize_rcu() -only- guarantees to wait until all currently executing rcu_read_lock()-protected RCU read-side @@ -239,15 +270,21 @@ over a rather long period of time, but improvements are always welcome! rcu_read_lock()-protected read-side critical sections, do -not- use synchronize_rcu(). - If you want to wait for some of these other things, you might - instead need to use synchronize_irq() or synchronize_sched(). + Similarly, disabling preemption is not an acceptable substitute + for rcu_read_lock(). Code that attempts to use preemption + disabling where it should be using rcu_read_lock() will break + in real-time kernel builds. + + If you want to wait for interrupt handlers, NMI handlers, and + code under the influence of preempt_disable(), you instead + need to use synchronize_irq() or synchronize_sched(). 12. Any lock acquired by an RCU callback must be acquired elsewhere with softirq disabled, e.g., via spin_lock_irqsave(), spin_lock_bh(), etc. Failing to disable irq on a given - acquisition of that lock will result in deadlock as soon as the - RCU callback happens to interrupt that acquisition's critical - section. + acquisition of that lock will result in deadlock as soon as + the RCU softirq handler happens to run your RCU callback while + interrupting that acquisition's critical section. 13. RCU callbacks can be and are executed in parallel. In many cases, the callback code simply wrappers around kfree(), so that this @@ -265,29 +302,30 @@ over a rather long period of time, but improvements are always welcome! not the case, a self-spawning RCU callback would prevent the victim CPU from ever going offline.) -14. SRCU (srcu_read_lock(), srcu_read_unlock(), and synchronize_srcu()) - may only be invoked from process context. Unlike other forms of - RCU, it -is- permissible to block in an SRCU read-side critical - section (demarked by srcu_read_lock() and srcu_read_unlock()), - hence the "SRCU": "sleepable RCU". Please note that if you - don't need to sleep in read-side critical sections, you should - be using RCU rather than SRCU, because RCU is almost always - faster and easier to use than is SRCU. +14. SRCU (srcu_read_lock(), srcu_read_unlock(), synchronize_srcu(), + and synchronize_srcu_expedited()) may only be invoked from + process context. Unlike other forms of RCU, it -is- permissible + to block in an SRCU read-side critical section (demarked by + srcu_read_lock() and srcu_read_unlock()), hence the "SRCU": + "sleepable RCU". Please note that if you don't need to sleep + in read-side critical sections, you should be using RCU rather + than SRCU, because RCU is almost always faster and easier to + use than is SRCU. Also unlike other forms of RCU, explicit initialization and cleanup is required via init_srcu_struct() and cleanup_srcu_struct(). These are passed a "struct srcu_struct" that defines the scope of a given SRCU domain. Once initialized, the srcu_struct is passed to srcu_read_lock(), srcu_read_unlock() - and synchronize_srcu(). A given synchronize_srcu() waits only - for SRCU read-side critical sections governed by srcu_read_lock() - and srcu_read_unlock() calls that have been passd the same - srcu_struct. This property is what makes sleeping read-side - critical sections tolerable -- a given subsystem delays only - its own updates, not those of other subsystems using SRCU. - Therefore, SRCU is less prone to OOM the system than RCU would - be if RCU's read-side critical sections were permitted to - sleep. + synchronize_srcu(), and synchronize_srcu_expedited(). A given + synchronize_srcu() waits only for SRCU read-side critical + sections governed by srcu_read_lock() and srcu_read_unlock() + calls that have been passed the same srcu_struct. This property + is what makes sleeping read-side critical sections tolerable -- + a given subsystem delays only its own updates, not those of other + subsystems using SRCU. Therefore, SRCU is less prone to OOM the + system than RCU would be if RCU's read-side critical sections + were permitted to sleep. The ability to sleep in read-side critical sections does not come for free. First, corresponding srcu_read_lock() and @@ -311,12 +349,12 @@ over a rather long period of time, but improvements are always welcome! destructive operation, and -only- -then- invoke call_rcu(), synchronize_rcu(), or friends. - Because these primitives only wait for pre-existing readers, - it is the caller's responsibility to guarantee safety to - any subsequent readers. + Because these primitives only wait for pre-existing readers, it + is the caller's responsibility to guarantee that any subsequent + readers will execute safely. -16. The various RCU read-side primitives do -not- contain memory - barriers. The CPU (and in some cases, the compiler) is free - to reorder code into and out of RCU read-side critical sections. - It is the responsibility of the RCU update-side primitives to - deal with this. +16. The various RCU read-side primitives do -not- necessarily contain + memory barriers. You should therefore plan for the CPU + and the compiler to freely reorder code into and out of RCU + read-side critical sections. It is the responsibility of the + RCU update-side primitives to deal with this. diff --git a/Documentation/RCU/rcu.txt b/Documentation/RCU/rcu.txt index 2a23523..3185270 100644 --- a/Documentation/RCU/rcu.txt +++ b/Documentation/RCU/rcu.txt @@ -75,6 +75,8 @@ o I hear that RCU is patented? What is with that? search for the string "Patent" in RTFP.txt to find them. Of these, one was allowed to lapse by the assignee, and the others have been contributed to the Linux kernel under GPL. + There are now also LGPL implementations of user-level RCU + available (http://lttng.org/?q=node/18). o I hear that RCU needs work in order to support realtime kernels? @@ -91,48 +93,4 @@ o Where can I find more information on RCU? o What are all these files in this directory? - - NMI-RCU.txt - - Describes how to use RCU to implement dynamic - NMI handlers, which can be revectored on the fly, - without rebooting. - - RTFP.txt - - List of RCU-related publications and web sites. - - UP.txt - - Discussion of RCU usage in UP kernels. - - arrayRCU.txt - - Describes how to use RCU to protect arrays, with - resizeable arrays whose elements reference other - data structures being of the most interest. - - checklist.txt - - Lists things to check for when inspecting code that - uses RCU. - - listRCU.txt - - Describes how to use RCU to protect linked lists. - This is the simplest and most common use of RCU - in the Linux kernel. - - rcu.txt - - You are reading it! - - rcuref.txt - - Describes how to combine use of reference counts - with RCU. - - whatisRCU.txt - - Overview of how the RCU implementation works. Along - the way, presents a conceptual view of RCU. + See 00-INDEX for the list. diff --git a/Documentation/RCU/stallwarn.txt b/Documentation/RCU/stallwarn.txt new file mode 100644 index 0000000..1423d25 --- /dev/null +++ b/Documentation/RCU/stallwarn.txt @@ -0,0 +1,58 @@ +Using RCU's CPU Stall Detector + +The CONFIG_RCU_CPU_STALL_DETECTOR kernel config parameter enables +RCU's CPU stall detector, which detects conditions that unduly delay +RCU grace periods. The stall detector's idea of what constitutes +"unduly delayed" is controlled by a pair of C preprocessor macros: + +RCU_SECONDS_TILL_STALL_CHECK + + This macro defines the period of time that RCU will wait from + the beginning of a grace period until it issues an RCU CPU + stall warning. It is normally ten seconds. + +RCU_SECONDS_TILL_STALL_RECHECK + + This macro defines the period of time that RCU will wait after + issuing a stall warning until it issues another stall warning. + It is normally set to thirty seconds. + +RCU_STALL_RAT_DELAY + + The CPU stall detector tries to make the offending CPU rat on itself, + as this often gives better-quality stack traces. However, if + the offending CPU does not detect its own stall in the number + of jiffies specified by RCU_STALL_RAT_DELAY, then other CPUs will + complain. This is normally set to two jiffies. + +The following problems can result in an RCU CPU stall warning: + +o A CPU looping in an RCU read-side critical section. + +o A CPU looping with interrupts disabled. + +o A CPU looping with preemption disabled. + +o For !CONFIG_PREEMPT kernels, a CPU looping anywhere in the kernel + without invoking schedule(). + +o A bug in the RCU implementation. + +o A hardware failure. This is quite unlikely, but has occurred + at least once in a former life. A CPU failed in a running system, + becoming unresponsive, but not causing an immediate crash. + This resulted in a series of RCU CPU stall warnings, eventually + leading the realization that the CPU had failed. + +The RCU, RCU-sched, and RCU-bh implementations have CPU stall warning. +SRCU does not do so directly, but its calls to synchronize_sched() will +result in RCU-sched detecting any CPU stalls that might be occurring. + +To diagnose the cause of the stall, inspect the stack traces. The offending +function will usually be near the top of the stack. If you have a series +of stall warnings from a single extended stall, comparing the stack traces +can often help determine where the stall is occurring, which will usually +be in the function nearest the top of the stack that stays the same from +trace to trace. + +RCU bugs can often be debugged with the help of CONFIG_RCU_TRACE. diff --git a/Documentation/RCU/torture.txt b/Documentation/RCU/torture.txt index 9dba3bb..0e50bc2 100644 --- a/Documentation/RCU/torture.txt +++ b/Documentation/RCU/torture.txt @@ -30,6 +30,18 @@ MODULE PARAMETERS This module has the following parameters: +fqs_duration Duration (in microseconds) of artificially induced bursts + of force_quiescent_state() invocations. In RCU + implementations having force_quiescent_state(), these + bursts help force races between forcing a given grace + period and that grace period ending on its own. + +fqs_holdoff Holdoff time (in microseconds) between consecutive calls + to force_quiescent_state() within a burst. + +fqs_stutter Wait time (in seconds) between consecutive bursts + of calls to force_quiescent_state(). + irqreaders Says to invoke RCU readers from irq level. This is currently done via timers. Defaults to "1" for variants of RCU that permit this. (Or, more accurately, variants of RCU that do diff --git a/Documentation/RCU/whatisRCU.txt b/Documentation/RCU/whatisRCU.txt index d542ca2..469a58b 100644 --- a/Documentation/RCU/whatisRCU.txt +++ b/Documentation/RCU/whatisRCU.txt @@ -327,7 +327,8 @@ a. synchronize_rcu() rcu_read_lock() / rcu_read_unlock() b. call_rcu_bh() rcu_read_lock_bh() / rcu_read_unlock_bh() -c. synchronize_sched() preempt_disable() / preempt_enable() +c. synchronize_sched() rcu_read_lock_sched() / rcu_read_unlock_sched() + preempt_disable() / preempt_enable() local_irq_save() / local_irq_restore() hardirq enter / hardirq exit NMI enter / NMI exit diff --git a/Documentation/filesystems/dentry-locking.txt b/Documentation/filesystems/dentry-locking.txt index 4c0c575..79334ed 100644 --- a/Documentation/filesystems/dentry-locking.txt +++ b/Documentation/filesystems/dentry-locking.txt @@ -62,7 +62,8 @@ changes are : 2. Insertion of a dentry into the hash table is done using hlist_add_head_rcu() which take care of ordering the writes - the writes to the dentry must be visible before the dentry is - inserted. This works in conjunction with hlist_for_each_rcu() while + inserted. This works in conjunction with hlist_for_each_rcu(), + which has since been replaced by hlist_for_each_entry_rcu(), while walking the hash chain. The only requirement is that all initialization to the dentry must be done before hlist_add_head_rcu() since we don't have dcache_lock protection diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 25c3ed5..6bf97d1 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -765,9 +765,9 @@ config RCU_CPU_STALL_DETECTOR CPUs are delaying the current grace period, but only when the grace period extends for excessive time periods. - Say Y if you want RCU to perform such checks. + Say N if you want to disable such checks. - Say N if you are unsure. + Say Y if you are unsure. config KPROBES_SANITY_TEST bool "Kprobes sanity tests" -- cgit v0.10.2 From 017c426138122c8e9b9f5057fbd0567c37b35247 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 14 Jan 2010 16:10:58 -0800 Subject: rcu: Fix sparse warnings Rename local variable "i" in rcu_init() to avoid conflict with RCU_INIT_FLAVOR(), restrict the scope of RCU_TREE_NONCORE, and make __synchronize_srcu() static. Signed-off-by: Paul E. McKenney Cc: laijs@cn.fujitsu.com Cc: dipankar@in.ibm.com Cc: mathieu.desnoyers@polymtl.ca Cc: josh@joshtriplett.org Cc: dvhltc@us.ibm.com Cc: niv@us.ibm.com Cc: peterz@infradead.org Cc: rostedt@goodmis.org Cc: Valdis.Kletnieks@vt.edu Cc: dhowells@redhat.com LKML-Reference: <12635142581560-git-send-email-> Signed-off-by: Ingo Molnar diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 3b13d64..099a255 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -1888,7 +1888,7 @@ do { \ void __init rcu_init(void) { - int i; + int cpu; rcu_bootup_announce(); #ifdef CONFIG_RCU_CPU_STALL_DETECTOR @@ -1908,8 +1908,8 @@ void __init rcu_init(void) * or the scheduler are operational. */ cpu_notifier(rcu_cpu_notify, 0); - for_each_online_cpu(i) - rcu_cpu_notify(NULL, CPU_UP_PREPARE, (void *)(long)i); + for_each_online_cpu(cpu) + rcu_cpu_notify(NULL, CPU_UP_PREPARE, (void *)(long)cpu); } #include "rcutree_plugin.h" diff --git a/kernel/rcutree.h b/kernel/rcutree.h index bd5d78ad..d9d032a 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -323,8 +323,6 @@ struct rcu_state { #define RCU_OFL_TASKS_EXP_GP 0x2 /* Tasks blocking expedited */ /* GP were moved to root. */ -#ifdef RCU_TREE_NONCORE - /* * RCU implementation internal declarations: */ @@ -339,7 +337,7 @@ extern struct rcu_state rcu_preempt_state; DECLARE_PER_CPU(struct rcu_data, rcu_preempt_data); #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ -#else /* #ifdef RCU_TREE_NONCORE */ +#ifndef RCU_TREE_NONCORE /* Forward declarations for rcutree_plugin.h */ static void rcu_bootup_announce(void); @@ -372,4 +370,4 @@ static void __cpuinit rcu_preempt_init_percpu_data(int cpu); static void rcu_preempt_send_cbs_to_orphanage(void); static void __init __rcu_init_preempt(void); -#endif /* #else #ifdef RCU_TREE_NONCORE */ +#endif /* #ifndef RCU_TREE_NONCORE */ diff --git a/kernel/srcu.c b/kernel/srcu.c index 818d7d9..31b275b 100644 --- a/kernel/srcu.c +++ b/kernel/srcu.c @@ -144,7 +144,7 @@ EXPORT_SYMBOL_GPL(srcu_read_unlock); /* * Helper function for synchronize_srcu() and synchronize_srcu_expedited(). */ -void __synchronize_srcu(struct srcu_struct *sp, void (*sync_func)(void)) +static void __synchronize_srcu(struct srcu_struct *sp, void (*sync_func)(void)) { int idx; -- cgit v0.10.2 From 0d755034dbd01e240eadf2d31f4f75d3088ccd21 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2010 12:23:09 -0200 Subject: perf tools: Don't cast RIP to pointers MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since they can come from another architecture with bigger pointers, i.e. processing a 64-bit perf.data on a 32-bit arch. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263478990-8200-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 117bbae..73c202e 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -132,8 +132,8 @@ static int process_sample_event(event_t *event, struct perf_session *session) { struct addr_location al; - dump_printf("(IP, %d): %d: %p\n", event->header.misc, - event->ip.pid, (void *)(long)event->ip.ip); + dump_printf("(IP, %d): %d: %#Lx\n", event->header.misc, + event->ip.pid, event->ip.ip); if (event__preprocess_sample(event, session, &al, symbol_filter) < 0) { fprintf(stderr, "problem processing %d event, skipping it.\n", diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c index 924bfb7..18b3f50 100644 --- a/tools/perf/builtin-diff.c +++ b/tools/perf/builtin-diff.c @@ -42,8 +42,8 @@ static int diff__process_sample_event(event_t *event, struct perf_session *sessi struct addr_location al; struct sample_data data = { .period = 1, }; - dump_printf("(IP, %d): %d: %p\n", event->header.misc, - event->ip.pid, (void *)(long)event->ip.ip); + dump_printf("(IP, %d): %d: %#Lx\n", event->header.misc, + event->ip.pid, event->ip.ip); if (event__preprocess_sample(event, session, &al, NULL) < 0) { pr_warning("problem processing %d event, skipping it.\n", diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 4af7199..7323d9d 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -316,11 +316,8 @@ static int process_sample_event(event_t *event, struct perf_session *session) event__parse_sample(event, session->sample_type, &data); - dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", - event->header.misc, - data.pid, data.tid, - (void *)(long)data.ip, - (long long)data.period); + dump_printf("(IP, %d): %d/%d: %#Lx period: %Ld\n", event->header.misc, + data.pid, data.tid, data.ip, data.period); thread = perf_session__findnew(session, event->ip.pid); if (thread == NULL) { diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 80d691a..4c3d699 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -93,11 +93,8 @@ static int process_sample_event(event_t *event, struct perf_session *session) event__parse_sample(event, session->sample_type, &data); - dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", - event->header.misc, - data.pid, data.tid, - (void *)(long)data.ip, - (long long)data.period); + dump_printf("(IP, %d): %d/%d: %#Lx period: %Ld\n", event->header.misc, + data.pid, data.tid, data.ip, data.period); if (session->sample_type & PERF_SAMPLE_CALLCHAIN) { unsigned int i; diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 702322f..4f5a03e 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1621,11 +1621,8 @@ static int process_sample_event(event_t *event, struct perf_session *session) event__parse_sample(event, session->sample_type, &data); - dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", - event->header.misc, - data.pid, data.tid, - (void *)(long)data.ip, - (long long)data.period); + dump_printf("(IP, %d): %d/%d: %#Lx period: %Ld\n", event->header.misc, + data.pid, data.tid, data.ip, data.period); thread = perf_session__findnew(session, data.pid); if (thread == NULL) { diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 1831434..8e9cbfe 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -75,11 +75,8 @@ static int process_sample_event(event_t *event, struct perf_session *session) event__parse_sample(event, session->sample_type, &data); - dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", - event->header.misc, - data.pid, data.tid, - (void *)(long)data.ip, - (long long)data.period); + dump_printf("(IP, %d): %d/%d: %#Lx period: %Ld\n", event->header.misc, + data.pid, data.tid, data.ip, data.period); thread = perf_session__findnew(session, event->ip.pid); if (thread == NULL) { diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 4f3e7ef..24ec5be 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -313,12 +313,9 @@ int event__process_mmap(event_t *self, struct perf_session *session) struct thread *thread; struct map *map; - dump_printf(" %d/%d: [%p(%p) @ %p]: %s\n", - self->mmap.pid, self->mmap.tid, - (void *)(long)self->mmap.start, - (void *)(long)self->mmap.len, - (void *)(long)self->mmap.pgoff, - self->mmap.filename); + dump_printf(" %d/%d: [%#Lx(%#Lx) @ %#Lx]: %s\n", + self->mmap.pid, self->mmap.tid, self->mmap.start, + self->mmap.len, self->mmap.pgoff, self->mmap.filename); if (self->mmap.pid == 0) { static const char kmmap_prefix[] = "[kernel.kallsyms."; diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index fd1c5a3..e3ccdb4 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -209,9 +209,8 @@ static int perf_session__process_event(struct perf_session *self, trace_event(event); if (event->header.type < PERF_RECORD_MAX) { - dump_printf("%p [%p]: PERF_RECORD_%s", - (void *)(offset + head), - (void *)(long)(event->header.size), + dump_printf("%#lx [%#x]: PERF_RECORD_%s", + offset + head, event->header.size, event__name[event->header.type]); ++event__total[0]; ++event__total[event->header.type]; @@ -362,16 +361,13 @@ more: size = event->header.size; - dump_printf("\n%p [%p]: event: %d\n", - (void *)(offset + head), - (void *)(long)event->header.size, - event->header.type); + dump_printf("\n%#lx [%#x]: event: %d\n", + offset + head, event->header.size, event->header.type); if (size == 0 || perf_session__process_event(self, event, ops, offset, head) < 0) { - dump_printf("%p [%p]: skipping unknown header type: %d\n", - (void *)(offset + head), - (void *)(long)(event->header.size), + dump_printf("%#lx [%#x]: skipping unknown header type: %d\n", + offset + head, event->header.size, event->header.type); /* * assume we lost track of the stream, check alignment, and -- cgit v0.10.2 From ba21594cddee0a3af582971656702b1c4509d8f5 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2010 12:23:10 -0200 Subject: perf tools: Cross platform perf.data analysis support MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There are still some problems related to loading vmlinux files, but those are unrelated to the feature implemented in this patch, so will get fixed in the next patches, but here are some results: 1. collect perf.data file on a Fedora 12 machine, x86_64, 64-bit userland 2. transfer it to a Debian Testing machine, PARISC64, 32-bit userland acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | head -5 74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms] 55fdd56670453ea66c011158c4b9d30179c1d049 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/ipt_MASQUERADE.ko 41adff63c730890480980d5d8ba513f1c216a858 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/iptable_nat.ko 90a33def1077bb8e97b8a78546dc96c2de62df46 /lib/modules/2.6.33-rc4-tip+/kernel/net/ipv4/netfilter/nf_nat.ko 984c7bea90ce1376d5c8e7ef43a781801286e62d /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/tun.ko acme@parisc:~/git/linux-2.6-tip$ perf buildid-list | tail -5 22492f3753c6a67de5c7ccbd6b863390c92c0723 /usr/lib64/libXt.so.6.0.0 353802bb7e1b895ba43507cc678f951e778e4c6f /usr/lib64/libMagickCore.so.2.0.0 d10c2897558595efe7be8b0584cf7e6398bc776c /usr/lib64/libfprint.so.0.0.0 a83ecfb519a788774a84d5ddde633c9ba56c03ab /home/acme/bin/perf d3ca765a8ecf257d263801d7ad8c49c189082317 /usr/lib64/libdwarf.so.0.0 acme@parisc:~/git/linux-2.6-tip$ acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm The file [kernel.kallsyms] cannot be used, trying to use /proc/kallsyms... ^^^^ The problem related to vmlinux handling, it shouldn't be trying this ^^^^ rather alien /proc/kallsyms at all... /lib64/libpthread-2.10.2.so with build id 5c68f7afeb33309c78037e374b0deee84dd441f6 not found, continuing without symbols /lib64/libc-2.10.2.so with build id eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 not found, continuing without symbols /home/acme/bin/perf with build id a83ecfb519a788774a84d5ddde633c9ba56c03ab not found, continuing without symbols /usr/sbin/openvpn with build id f2037a091ef36b591187a858d75e203690ea9409 not found, continuing without symbols Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko, continuing without symbols Failed to open /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwlcore.ko, continuing without symbols # Samples: 293085637 # # Overhead Command # ........ ............... # 61.70% find 23.50% perf 5.86% swapper 3.12% sshd 2.39% init 0.87% bash 0.86% sleep 0.59% dbus-daemon 0.25% hald 0.24% NetworkManager 0.19% hald-addon-rfki 0.15% openvpn 0.07% phy0 0.07% events/0 0.05% iwl3945 0.05% events/1 0.03% kondemand/0 acme@parisc:~/git/linux-2.6-tip$ Which matches what we get when running the same command for the same perf.data file on the F12, x86_64, source machine: [root@doppio linux-2.6-tip]# perf report --sort comm # Samples: 293085637 # # Overhead Command # ........ ............... # 61.70% find 23.50% perf 5.86% swapper 3.12% sshd 2.39% init 0.87% bash 0.86% sleep 0.59% dbus-daemon 0.25% hald 0.24% NetworkManager 0.19% hald-addon-rfki 0.15% openvpn 0.07% phy0 0.07% events/0 0.05% iwl3945 0.05% events/1 0.03% kondemand/0 [root@doppio linux-2.6-tip]# The other modes work as well, modulo the problem with vmlinux: acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso 2> /dev/null | head -15 # Samples: 293085637 # # Overhead Command Shared Object # ........ ............... ................................. # 35.11% find ffffffff81002b5a 18.25% perf ffffffff8102235f 16.17% find libc-2.10.2.so 9.07% find find 5.80% swapper ffffffff8102235f 3.95% perf libc-2.10.2.so 2.33% init ffffffff810091b9 1.65% sshd libcrypto.so.0.9.8k 1.35% find [e1000e] 0.68% sleep libc-2.10.2.so acme@parisc:~/git/linux-2.6-tip$ And the lack of the right buildids: acme@parisc:~/git/linux-2.6-tip$ perf report --sort comm,dso,symbol 2> /dev/null | head -15 # Samples: 293085637 # # Overhead Command Shared Object Symbol # ........ ............... ................................. ...... # 35.11% find ffffffff81002b5a [k] 0xffffffff81002b5a 18.25% perf ffffffff8102235f [k] 0xffffffff8102235f 16.17% find libc-2.10.2.so [.] 0x00000000045782 9.07% find find [.] 0x0000000000fb0e 5.80% swapper ffffffff8102235f [k] 0xffffffff8102235f 3.95% perf libc-2.10.2.so [.] 0x0000000007f398 2.33% init ffffffff810091b9 [k] 0xffffffff810091b9 1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440 1.35% find [e1000e] [k] 0x00000000010948 0.68% sleep libc-2.10.2.so [.] 0x0000000011ad5b acme@parisc:~/git/linux-2.6-tip$ But if we: acme@parisc:~/git/linux-2.6-tip$ ls ~/.debug ls: cannot access /home/acme/.debug: No such file or directory acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/lib64/libc-2.10.2.so/ acme@parisc:~/git/linux-2.6-tip$ scp doppio:.debug/lib64/libc-2.10.2.so/* ~/.debug/lib64/libc-2.10.2.so/ acme@doppio's password: eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 100% 1783KB 714.7KB/s 00:02 acme@parisc:~/git/linux-2.6-tip$ mkdir -p ~/.debug/.build-id/eb acme@parisc:~/git/linux-2.6-tip$ ln -s ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 ~/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 acme@parisc:~/git/linux-2.6-tip$ perf report --dsos libc-2.10.2.so 2> /dev/null # dso: libc-2.10.2.so # Samples: 64281170 # # Overhead Command Symbol # ........ ............... ...... # 14.98% perf [.] __GI_strcmp 12.30% find [.] __GI_memmove 9.25% find [.] _int_malloc 7.60% find [.] _IO_vfprintf_internal 6.10% find [.] _IO_new_file_xsputn 6.02% find [.] __GI_close 3.08% find [.] _IO_file_overflow_internal 3.08% find [.] malloc_consolidate 3.08% find [.] _int_free 3.08% find [.] __strchrnul 3.08% find [.] __getdents64 3.08% find [.] __write_nocancel 3.08% sleep [.] __GI__dl_addr 3.08% sshd [.] __libc_select 3.08% find [.] _IO_new_file_write 3.07% find [.] _IO_new_do_write 3.06% find [.] __GI___errno_location 3.05% find [.] __GI___libc_malloc 3.04% perf [.] __GI_memcpy 1.71% find [.] __fprintf_chk 1.29% bash [.] __gconv_transform_utf8_internal 0.79% dbus-daemon [.] __GI_strlen # # (For a higher level overview, try: perf report --sort comm,dso) # acme@parisc:~/git/linux-2.6-tip$ Which matches what we get on the source, F12, x86_64 machine: [root@doppio linux-2.6-tip]# perf report --dsos libc-2.10.2.so # dso: libc-2.10.2.so # Samples: 64281170 # # Overhead Command Symbol # ........ ............... ...... # 14.98% perf [.] __GI_strcmp 12.30% find [.] __GI_memmove 9.25% find [.] _int_malloc 7.60% find [.] _IO_vfprintf_internal 6.10% find [.] _IO_new_file_xsputn 6.02% find [.] __GI_close 3.08% find [.] _IO_file_overflow_internal 3.08% find [.] malloc_consolidate 3.08% find [.] _int_free 3.08% find [.] __strchrnul 3.08% find [.] __getdents64 3.08% find [.] __write_nocancel 3.08% sleep [.] __GI__dl_addr 3.08% sshd [.] __libc_select 3.08% find [.] _IO_new_file_write 3.07% find [.] _IO_new_do_write 3.06% find [.] __GI___errno_location 3.05% find [.] __GI___libc_malloc 3.04% perf [.] __GI_memcpy 1.71% find [.] __fprintf_chk 1.29% bash [.] __gconv_transform_utf8_internal 0.79% dbus-daemon [.] __GI_strlen # # (For a higher level overview, try: perf report --sort comm,dso) # [root@doppio linux-2.6-tip]# So I think this is really, really nice in that it demonstrates the portability of perf.data files and the use of build-ids accross such aliens worlds :-) There are some things to fix tho, like the bitmap on the header, but things are looking good. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263478990-8200-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index ec96321..b31e0ae 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -1,8 +1,10 @@ #include +#include #include #include #include #include +#include #include "util.h" #include "header.h" @@ -464,8 +466,21 @@ static int do_read(int fd, void *buf, size_t size) return 0; } +static int perf_header__getbuffer64(struct perf_header *self, + int fd, void *buf, size_t size) +{ + if (do_read(fd, buf, size)) + return -1; + + if (self->needs_swap) + mem_bswap_64(buf, size); + + return 0; +} + int perf_header__process_sections(struct perf_header *self, int fd, int (*process)(struct perf_file_section *self, + struct perf_header *ph, int feat, int fd)) { struct perf_file_section *feat_sec; @@ -486,7 +501,7 @@ int perf_header__process_sections(struct perf_header *self, int fd, lseek(fd, self->data_offset + self->data_size, SEEK_SET); - if (do_read(fd, feat_sec, sec_size)) + if (perf_header__getbuffer64(self, fd, feat_sec, sec_size)) goto out_free; err = 0; @@ -494,7 +509,7 @@ int perf_header__process_sections(struct perf_header *self, int fd, if (perf_header__has_feat(self, feat)) { struct perf_file_section *sec = &feat_sec[idx++]; - err = process(sec, feat, fd); + err = process(sec, self, feat, fd); if (err < 0) break; } @@ -511,10 +526,20 @@ int perf_file_header__read(struct perf_file_header *self, lseek(fd, 0, SEEK_SET); if (do_read(fd, self, sizeof(*self)) || - self->magic != PERF_MAGIC || - self->attr_size != sizeof(struct perf_file_attr)) + memcmp(&self->magic, __perf_magic, sizeof(self->magic))) return -1; + if (self->attr_size != sizeof(struct perf_file_attr)) { + u64 attr_size = bswap_64(self->attr_size); + + if (attr_size != sizeof(struct perf_file_attr)) + return -1; + + mem_bswap_64(self, offsetof(struct perf_file_header, + adds_features)); + ph->needs_swap = true; + } + if (self->size != sizeof(*self)) { /* Support the previous format */ if (self->size == offsetof(typeof(*self), adds_features)) @@ -524,16 +549,28 @@ int perf_file_header__read(struct perf_file_header *self, } memcpy(&ph->adds_features, &self->adds_features, - sizeof(self->adds_features)); + sizeof(ph->adds_features)); + /* + * FIXME: hack that assumes that if we need swap the perf.data file + * may be coming from an arch with a different word-size, ergo different + * DEFINE_BITMAP format, investigate more later, but for now its mostly + * safe to assume that we have a build-id section. Trace files probably + * have several other issues in this realm anyway... + */ + if (ph->needs_swap) { + memset(&ph->adds_features, 0, sizeof(ph->adds_features)); + perf_header__set_feat(ph, HEADER_BUILD_ID); + } ph->event_offset = self->event_types.offset; - ph->event_size = self->event_types.size; - ph->data_offset = self->data.offset; + ph->event_size = self->event_types.size; + ph->data_offset = self->data.offset; ph->data_size = self->data.size; return 0; } static int perf_file_section__process(struct perf_file_section *self, + struct perf_header *ph, int feat, int fd) { if (lseek(fd, self->offset, SEEK_SET) < 0) { @@ -548,7 +585,7 @@ static int perf_file_section__process(struct perf_file_section *self, break; case HEADER_BUILD_ID: - if (perf_header__read_build_ids(fd, self->offset, self->size)) + if (perf_header__read_build_ids(ph, fd, self->offset, self->size)) pr_debug("Failed to read buildids, continuing...\n"); break; default: @@ -560,7 +597,7 @@ static int perf_file_section__process(struct perf_file_section *self, int perf_header__read(struct perf_header *self, int fd) { - struct perf_file_header f_header; + struct perf_file_header f_header; struct perf_file_attr f_attr; u64 f_id; int nr_attrs, nr_ids, i, j; @@ -577,8 +614,9 @@ int perf_header__read(struct perf_header *self, int fd) struct perf_header_attr *attr; off_t tmp; - if (do_read(fd, &f_attr, sizeof(f_attr))) + if (perf_header__getbuffer64(self, fd, &f_attr, sizeof(f_attr))) goto out_errno; + tmp = lseek(fd, 0, SEEK_CUR); attr = perf_header_attr__new(&f_attr.attr); @@ -589,7 +627,7 @@ int perf_header__read(struct perf_header *self, int fd) lseek(fd, f_attr.ids.offset, SEEK_SET); for (j = 0; j < nr_ids; j++) { - if (do_read(fd, &f_id, sizeof(f_id))) + if (perf_header__getbuffer64(self, fd, &f_id, sizeof(f_id))) goto out_errno; if (perf_header_attr__add_id(attr, f_id) < 0) { @@ -610,7 +648,8 @@ int perf_header__read(struct perf_header *self, int fd) events = malloc(f_header.event_types.size); if (events == NULL) return -ENOMEM; - if (do_read(fd, events, f_header.event_types.size)) + if (perf_header__getbuffer64(self, fd, events, + f_header.event_types.size)) goto out_errno; event_count = f_header.event_types.size / sizeof(struct perf_trace_event_type); } diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h index 2b69aab..ccc8540 100644 --- a/tools/perf/util/header.h +++ b/tools/perf/util/header.h @@ -52,6 +52,7 @@ struct perf_header { u64 data_size; u64 event_offset; u64 event_size; + bool needs_swap; DECLARE_BITMAP(adds_features, HEADER_FEAT_BITS); }; @@ -80,6 +81,7 @@ bool perf_header__has_feat(const struct perf_header *self, int feat); int perf_header__process_sections(struct perf_header *self, int fd, int (*process)(struct perf_file_section *self, + struct perf_header *ph, int feat, int fd)); #endif /* __PERF_HEADER_H */ diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index e3ccdb4..604e14f 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1,5 +1,6 @@ #include +#include #include #include @@ -201,21 +202,88 @@ void event__print_totals(void) event__name[i], event__total[i]); } +void mem_bswap_64(void *src, int byte_size) +{ + u64 *m = src; + + while (byte_size > 0) { + *m = bswap_64(*m); + byte_size -= sizeof(u64); + ++m; + } +} + +static void event__all64_swap(event_t *self) +{ + struct perf_event_header *hdr = &self->header; + mem_bswap_64(hdr + 1, self->header.size - sizeof(*hdr)); +} + +static void event__comm_swap(event_t *self) +{ + self->comm.pid = bswap_32(self->comm.pid); + self->comm.tid = bswap_32(self->comm.tid); +} + +static void event__mmap_swap(event_t *self) +{ + self->mmap.pid = bswap_32(self->mmap.pid); + self->mmap.tid = bswap_32(self->mmap.tid); + self->mmap.start = bswap_64(self->mmap.start); + self->mmap.len = bswap_64(self->mmap.len); + self->mmap.pgoff = bswap_64(self->mmap.pgoff); +} + +static void event__task_swap(event_t *self) +{ + self->fork.pid = bswap_32(self->fork.pid); + self->fork.tid = bswap_32(self->fork.tid); + self->fork.ppid = bswap_32(self->fork.ppid); + self->fork.ptid = bswap_32(self->fork.ptid); + self->fork.time = bswap_64(self->fork.time); +} + +static void event__read_swap(event_t *self) +{ + self->read.pid = bswap_32(self->read.pid); + self->read.tid = bswap_32(self->read.tid); + self->read.value = bswap_64(self->read.value); + self->read.time_enabled = bswap_64(self->read.time_enabled); + self->read.time_running = bswap_64(self->read.time_running); + self->read.id = bswap_64(self->read.id); +} + +typedef void (*event__swap_op)(event_t *self); + +static event__swap_op event__swap_ops[] = { + [PERF_RECORD_MMAP] = event__mmap_swap, + [PERF_RECORD_COMM] = event__comm_swap, + [PERF_RECORD_FORK] = event__task_swap, + [PERF_RECORD_EXIT] = event__task_swap, + [PERF_RECORD_LOST] = event__all64_swap, + [PERF_RECORD_READ] = event__read_swap, + [PERF_RECORD_SAMPLE] = event__all64_swap, + [PERF_RECORD_MAX] = NULL, +}; + static int perf_session__process_event(struct perf_session *self, event_t *event, struct perf_event_ops *ops, - unsigned long offset, unsigned long head) + u64 offset, u64 head) { trace_event(event); if (event->header.type < PERF_RECORD_MAX) { - dump_printf("%#lx [%#x]: PERF_RECORD_%s", + dump_printf("%#Lx [%#x]: PERF_RECORD_%s", offset + head, event->header.size, event__name[event->header.type]); ++event__total[0]; ++event__total[event->header.type]; } + if (self->header.needs_swap && event__swap_ops[event->header.type]) + event__swap_ops[event->header.type](event); + switch (event->header.type) { case PERF_RECORD_SAMPLE: return ops->sample(event, self); @@ -241,7 +309,15 @@ static int perf_session__process_event(struct perf_session *self, } } -int perf_header__read_build_ids(int input, u64 offset, u64 size) +void perf_event_header__bswap(struct perf_event_header *self) +{ + self->type = bswap_32(self->type); + self->misc = bswap_16(self->misc); + self->size = bswap_16(self->size); +} + +int perf_header__read_build_ids(struct perf_header *self, + int input, u64 offset, u64 size) { struct build_id_event bev; char filename[PATH_MAX]; @@ -256,6 +332,9 @@ int perf_header__read_build_ids(int input, u64 offset, u64 size) if (read(input, &bev, sizeof(bev)) != sizeof(bev)) goto out; + if (self->needs_swap) + perf_event_header__bswap(&bev.header); + len = bev.header.size - sizeof(bev); if (read(input, filename, len) != len) goto out; @@ -292,9 +371,9 @@ static struct thread *perf_session__register_idle_thread(struct perf_session *se int perf_session__process_events(struct perf_session *self, struct perf_event_ops *ops) { - int err; - unsigned long head, shift; - unsigned long offset = 0; + int err, mmap_prot, mmap_flags; + u64 head, shift; + u64 offset = 0; size_t page_size; event_t *event; uint32_t size; @@ -330,9 +409,16 @@ out_getcwd_err: offset += shift; head -= shift; + mmap_prot = PROT_READ; + mmap_flags = MAP_SHARED; + + if (self->header.needs_swap) { + mmap_prot |= PROT_WRITE; + mmap_flags = MAP_PRIVATE; + } remap: - buf = mmap(NULL, page_size * self->mmap_window, PROT_READ, - MAP_SHARED, self->fd, offset); + buf = mmap(NULL, page_size * self->mmap_window, mmap_prot, + mmap_flags, self->fd, offset); if (buf == MAP_FAILED) { pr_err("failed to mmap file\n"); err = -errno; @@ -342,6 +428,8 @@ remap: more: event = (event_t *)(buf + head); + if (self->header.needs_swap) + perf_event_header__bswap(&event->header); size = event->header.size; if (size == 0) size = 8; @@ -361,12 +449,12 @@ more: size = event->header.size; - dump_printf("\n%#lx [%#x]: event: %d\n", + dump_printf("\n%#Lx [%#x]: event: %d\n", offset + head, event->header.size, event->header.type); if (size == 0 || perf_session__process_event(self, event, ops, offset, head) < 0) { - dump_printf("%#lx [%#x]: skipping unknown header type: %d\n", + dump_printf("%#Lx [%#x]: skipping unknown header type: %d\n", offset + head, event->header.size, event->header.type); /* diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index d4a9d20..36d1a80 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -51,6 +51,8 @@ struct perf_event_ops { struct perf_session *perf_session__new(const char *filename, int mode, bool force); void perf_session__delete(struct perf_session *self); +void perf_event_header__bswap(struct perf_event_header *self); + int perf_session__process_events(struct perf_session *self, struct perf_event_ops *event_ops); @@ -61,7 +63,8 @@ struct symbol **perf_session__resolve_callchain(struct perf_session *self, bool perf_session__has_traces(struct perf_session *self, const char *msg); -int perf_header__read_build_ids(int input, u64 offset, u64 file_size); +int perf_header__read_build_ids(struct perf_header *self, int input, + u64 offset, u64 file_size); int perf_session__set_kallsyms_ref_reloc_sym(struct perf_session *self, const char *symbol_name, @@ -69,4 +72,6 @@ int perf_session__set_kallsyms_ref_reloc_sym(struct perf_session *self, void perf_session__reloc_vmlinux_maps(struct perf_session *self, u64 unrelocated_addr); +void mem_bswap_64(void *src, int byte_size); + #endif /* __PERF_SESSION_H */ -- cgit v0.10.2 From 1b75962e92d48a41019d4b440e221638aa2a7238 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2010 18:30:04 -0200 Subject: perf tools: Convert getpagesize() uses to sysconf(_SC_GETPAGESIZE) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Using the more portable and equivalent sysconf call. Reported-by: Aristeu Rozanski Reported-by: Ulrich Drepper Signed-off-by: Arnaldo Carvalho de Melo Cc: Aristeu Rozanski Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Ulrich Drepper LKML-Reference: <1263501006-14185-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 604e14f..1951e33 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -384,7 +384,7 @@ int perf_session__process_events(struct perf_session *self, perf_event_ops__fill_defaults(ops); - page_size = getpagesize(); + page_size = sysconf(_SC_PAGESIZE); head = self->header.data_offset; diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c index 407fd65..5ea8973 100644 --- a/tools/perf/util/trace-event-info.c +++ b/tools/perf/util/trace-event-info.c @@ -515,7 +515,7 @@ int read_tracing_data(int fd, struct perf_event_attr *pattrs, int nb_events) write_or_die(buf, 1); /* save page_size */ - page_size = getpagesize(); + page_size = sysconf(_SC_PAGESIZE); write_or_die(&page_size, 4); read_header_files(); -- cgit v0.10.2 From 8d0591f6ad9edf66697ce29de176fb6f3213b9e3 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2010 18:30:05 -0200 Subject: perf symbols: Don't try to load kallsyms if doesn't match the record build-id MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Now a perf.data file collected on a x86_64 fedora 12 machine gets properly parsed on a Debian testing PARISC64 machine with 32-bit userland: acme@parisc:~/git/linux-2.6-tip$ perf report 2> /dev/null | head -15 # Samples: 293085637 # # Overhead Command Shared Object Symbol # ........ ............... ................................. ...... # 35.11% find [kernel.kallsyms] [k] 0xffffffff81002b5a 18.25% perf [kernel.kallsyms] [k] 0xffffffff8102235f 9.07% find find [.] 0x0000000000fb0e 5.80% swapper [kernel.kallsyms] [k] 0xffffffff8102235f 3.29% perf libc-2.10.2.so [.] __GI_strcmp 2.70% find libc-2.10.2.so [.] __GI_memmove 2.33% init [kernel.kallsyms] [k] 0xffffffff810091b9 2.03% find libc-2.10.2.so [.] _int_malloc 1.67% find libc-2.10.2.so [.] _IO_vfprintf_internal 1.65% sshd libcrypto.so.0.9.8k [.] 0x00000000105440 acme@parisc:~/git/linux-2.6-tip$ Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263501006-14185-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 381999d..71d23e1 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1608,11 +1608,11 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, u8 kallsyms_build_id[BUILD_ID_SIZE]; if (sysfs__read_build_id("/sys/kernel/notes", kallsyms_build_id, - sizeof(kallsyms_build_id)) == 0) - - is_kallsyms = dso__build_id_equal(self, kallsyms_build_id); - if (is_kallsyms) - goto do_kallsyms; + sizeof(kallsyms_build_id)) == 0) { + is_kallsyms = dso__build_id_equal(self, kallsyms_build_id); + if (is_kallsyms) + goto do_kallsyms; + } goto do_vmlinux; } @@ -1623,6 +1623,9 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, do_vmlinux: err = dso__load_vmlinux(self, map, session, self->long_name, filter); if (err <= 0) { + if (self->has_build_id) + return -1; + pr_info("The file %s cannot be used, " "trying to use /proc/kallsyms...", self->long_name); do_kallsyms: -- cgit v0.10.2 From 9e201442de7c954f03710ac76f28c1927d07550c Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2010 18:30:06 -0200 Subject: perf symbols: Cache /proc/kallsyms files by build-id MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit So that when we don't have a vmlinux handy we can store the kallsyms for later use by 'perf report'. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263501006-14185-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 24ec5be..0e9820a 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -245,7 +245,7 @@ int event__synthesize_kernel_mmap(event__handler_t process, */ struct process_symbol_args args = { .name = symbol_name, }; - if (kallsyms__parse(&args, find_symbol_cb) <= 0) + if (kallsyms__parse("/proc/kallsyms", &args, find_symbol_cb) <= 0) return -ENOENT; size = snprintf(ev.mmap.filename, sizeof(ev.mmap.filename), diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index b31e0ae..1b65fed 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -237,11 +237,13 @@ static int dso__cache_build_id(struct dso *self, const char *debugdir) char *filename = malloc(size), *linkname = malloc(size), *targetname, *sbuild_id; int len, err = -1; + bool is_kallsyms = self->kernel && self->long_name[0] != '/'; if (filename == NULL || linkname == NULL) goto out_free; - len = snprintf(filename, size, "%s%s", debugdir, self->long_name); + len = snprintf(filename, size, "%s%s%s", + debugdir, is_kallsyms ? "/" : "", self->long_name); if (mkdir_p(filename, 0755)) goto out_free; @@ -249,9 +251,14 @@ static int dso__cache_build_id(struct dso *self, const char *debugdir) sbuild_id = filename + len; build_id__sprintf(self->build_id, sizeof(self->build_id), sbuild_id); - if (access(filename, F_OK) && link(self->long_name, filename) && - copyfile(self->long_name, filename)) - goto out_free; + if (access(filename, F_OK)) { + if (is_kallsyms) { + if (copyfile("/proc/kallsyms", filename)) + goto out_free; + } else if (link(self->long_name, filename) && + copyfile(self->long_name, filename)) + goto out_free; + } len = snprintf(linkname, size, "%s/.build-id/%.2s", debugdir, sbuild_id); diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 71d23e1..ae61e9f 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -383,13 +383,14 @@ size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp) return ret; } -int kallsyms__parse(void *arg, int (*process_symbol)(void *arg, const char *name, +int kallsyms__parse(const char *filename, void *arg, + int (*process_symbol)(void *arg, const char *name, char type, u64 start)) { char *line = NULL; size_t n; int err = 0; - FILE *file = fopen("/proc/kallsyms", "r"); + FILE *file = fopen(filename, "r"); if (file == NULL) goto out_failure; @@ -466,10 +467,11 @@ static int map__process_kallsym_symbol(void *arg, const char *name, * so that we can in the next step set the symbol ->end address and then * call kernel_maps__split_kallsyms. */ -static int dso__load_all_kallsyms(struct dso *self, struct map *map) +static int dso__load_all_kallsyms(struct dso *self, const char *filename, + struct map *map) { struct process_kallsyms_args args = { .map = map, .dso = self, }; - return kallsyms__parse(&args, map__process_kallsym_symbol); + return kallsyms__parse(filename, &args, map__process_kallsym_symbol); } /* @@ -556,10 +558,10 @@ discard_symbol: rb_erase(&pos->rb_node, root); } -static int dso__load_kallsyms(struct dso *self, struct map *map, +static int dso__load_kallsyms(struct dso *self, const char *filename, struct map *map, struct perf_session *session, symbol_filter_t filter) { - if (dso__load_all_kallsyms(self, map) < 0) + if (dso__load_all_kallsyms(self, filename, map) < 0) return -1; symbols__fixup_end(&self->symbols[map->type]); @@ -1580,7 +1582,8 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, struct perf_session *session, symbol_filter_t filter) { int err; - bool is_kallsyms; + const char *kallsyms_filename = NULL; + char *kallsyms_allocated_filename = NULL; if (vmlinux_path != NULL) { int i; @@ -1606,19 +1609,37 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, */ if (self->has_build_id) { u8 kallsyms_build_id[BUILD_ID_SIZE]; + char sbuild_id[BUILD_ID_SIZE * 2 + 1]; if (sysfs__read_build_id("/sys/kernel/notes", kallsyms_build_id, sizeof(kallsyms_build_id)) == 0) { - is_kallsyms = dso__build_id_equal(self, kallsyms_build_id); - if (is_kallsyms) + if (dso__build_id_equal(self, kallsyms_build_id)) { + kallsyms_filename = "/proc/kallsyms"; goto do_kallsyms; + } } + + build_id__sprintf(self->build_id, sizeof(self->build_id), + sbuild_id); + + if (asprintf(&kallsyms_allocated_filename, + "%s/.debug/[kernel.kallsyms]/%s", + getenv("HOME"), sbuild_id) != -1) { + if (access(kallsyms_filename, F_OK)) { + kallsyms_filename = kallsyms_allocated_filename; + goto do_kallsyms; + } + free(kallsyms_allocated_filename); + kallsyms_allocated_filename = NULL; + } + goto do_vmlinux; } - is_kallsyms = self->long_name[0] == '['; - if (is_kallsyms) + if (self->long_name[0] == '[') { + kallsyms_filename = "/proc/kallsyms"; goto do_kallsyms; + } do_vmlinux: err = dso__load_vmlinux(self, map, session, self->long_name, filter); @@ -1629,9 +1650,10 @@ do_vmlinux: pr_info("The file %s cannot be used, " "trying to use /proc/kallsyms...", self->long_name); do_kallsyms: - err = dso__load_kallsyms(self, map, session, filter); - if (err > 0 && !is_kallsyms) + err = dso__load_kallsyms(self, kallsyms_filename, map, session, filter); + if (err > 0 && kallsyms_filename == NULL) dso__set_long_name(self, strdup("[kernel.kallsyms]")); + free(kallsyms_allocated_filename); } if (err > 0) { diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 594156e..36b7c71 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -144,8 +144,9 @@ int filename__read_build_id(const char *filename, void *bf, size_t size); int sysfs__read_build_id(const char *filename, void *bf, size_t size); bool dsos__read_build_ids(void); int build_id__sprintf(u8 *self, int len, char *bf); -int kallsyms__parse(void *arg, int (*process_symbol)(void *arg, const char *name, - char type, u64 start)); +int kallsyms__parse(const char *filename, void *arg, + int (*process_symbol)(void *arg, const char *name, + char type, u64 start)); int symbol__init(void); bool symbol_type__is_a(char symbol_type, enum map_type map_type); diff --git a/tools/perf/util/util.c b/tools/perf/util/util.c index f3c0798..f068584 100644 --- a/tools/perf/util/util.c +++ b/tools/perf/util/util.c @@ -32,6 +32,33 @@ int mkdir_p(char *path, mode_t mode) return (stat(path, &st) && mkdir(path, mode)) ? -1 : 0; } +static int slow_copyfile(const char *from, const char *to) +{ + int err = 0; + char *line = NULL; + size_t n; + FILE *from_fp = fopen(from, "r"), *to_fp; + + if (from_fp == NULL) + goto out; + + to_fp = fopen(to, "w"); + if (to_fp == NULL) + goto out_fclose_from; + + while (getline(&line, &n, from_fp) > 0) + if (fputs(line, to_fp) == EOF) + goto out_fclose_to; + err = 0; +out_fclose_to: + fclose(to_fp); + free(line); +out_fclose_from: + fclose(from_fp); +out: + return err; +} + int copyfile(const char *from, const char *to) { int fromfd, tofd; @@ -42,6 +69,9 @@ int copyfile(const char *from, const char *to) if (stat(from, &st)) goto out; + if (st.st_size == 0) /* /proc? do it slowly... */ + return slow_copyfile(from, to); + fromfd = open(from, O_RDONLY); if (fromfd < 0) goto out; -- cgit v0.10.2 From cf4e5b0838e822dd404638ad00d35b63fffe8191 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2010 23:45:27 -0200 Subject: perf symbols: Use dso->long_name in dsos__find() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If not we end up duplicating the module DSOs because first we insert them using the short name found in /proc/modules, then, when processing synthesized MMAP events we add them again. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263519930-22803-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index ae61e9f..4267138 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1679,7 +1679,7 @@ static struct dso *dsos__find(struct list_head *head, const char *name) struct dso *pos; list_for_each_entry(pos, head, node) - if (strcmp(pos->name, name) == 0) + if (strcmp(pos->long_name, name) == 0) return pos; return NULL; } -- cgit v0.10.2 From 18c3daa4961b9fa1f2db0711d93c0acf0c39fd12 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2010 23:45:28 -0200 Subject: perf record: Encode the domain while synthesizing MMAP events MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In the past 'perf record' had to process only userspace MMAP events, the ones generated in the kernel, but after we reused the MMAP events to encode the module mapings we ended up adding them first to the list of userspace DSOs (dsos__user) and to the kernel one (dsos__kernel). Fix this by encoding the header.misc field and then using it, like other parts to decide the right DSOs list to insert/find. The gotcha here is that since the kernel puts zero in .misc, which isn't PERF_RECORD_MISC_KERNEL (1 << 1), to differentiate, we put 1 in .misc. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263519930-22803-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index c130df2..614fa9a 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -117,8 +117,12 @@ static void write_event(event_t *buf, size_t size) * Add it to the list of DSOs, so that when we finish this * record session we can pick the available build-ids. */ - if (buf->header.type == PERF_RECORD_MMAP) - dsos__findnew(buf->mmap.filename); + if (buf->header.type == PERF_RECORD_MMAP) { + struct list_head *head = &dsos__user; + if (buf->mmap.header.misc == 1) + head = &dsos__kernel; + __dsos__findnew(head, buf->mmap.filename); + } write_output(buf, size); } diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 0e9820a..1abaefc 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -110,7 +110,10 @@ static int event__synthesize_mmap_events(pid_t pid, pid_t tgid, while (1) { char bf[BUFSIZ], *pbf = bf; event_t ev = { - .header = { .type = PERF_RECORD_MMAP }, + .header = { + .type = PERF_RECORD_MMAP, + .misc = 0, /* Just like the kernel, see kernel/perf_event.c __perf_event_mmap */ + }, }; int n; size_t size; @@ -170,6 +173,7 @@ int event__synthesize_modules(event__handler_t process, size = ALIGN(pos->dso->long_name_len + 1, sizeof(u64)); memset(&ev, 0, sizeof(ev)); + ev.mmap.header.misc = 1; /* kernel uses 0 for user space maps, see kernel/perf_event.c __perf_event_mmap */ ev.mmap.header.type = PERF_RECORD_MMAP; ev.mmap.header.size = (sizeof(ev.mmap) - (sizeof(ev.mmap.filename) - size)); @@ -236,7 +240,10 @@ int event__synthesize_kernel_mmap(event__handler_t process, { size_t size; event_t ev = { - .header = { .type = PERF_RECORD_MMAP }, + .header = { + .type = PERF_RECORD_MMAP, + .misc = 1, /* kernel uses 0 for user space maps, see kernel/perf_event.c __perf_event_mmap */ + }, }; /* * We should get this from /sys/kernel/sections/.text, but till that is -- cgit v0.10.2 From 59ee68ecd1561a233fb6ad351980bea8402533e7 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2010 23:45:29 -0200 Subject: perf symbols: Create thread__find_addr_map from thread__find_addr_location MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Because some tools will only want to know with maps had hits, not needing the full symbol resolution done by thread__find_addr_location. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263519930-22803-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 1abaefc..5a6e827 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -422,11 +422,10 @@ int event__process_task(event_t *self, struct perf_session *session) return 0; } -void thread__find_addr_location(struct thread *self, - struct perf_session *session, u8 cpumode, - enum map_type type, u64 addr, - struct addr_location *al, - symbol_filter_t filter) +void thread__find_addr_map(struct thread *self, + struct perf_session *session, u8 cpumode, + enum map_type type, u64 addr, + struct addr_location *al) { struct map_groups *mg = &self->mg; @@ -441,7 +440,6 @@ void thread__find_addr_location(struct thread *self, else { al->level = 'H'; al->map = NULL; - al->sym = NULL; return; } try_again: @@ -460,11 +458,21 @@ try_again: mg = &session->kmaps; goto try_again; } - al->sym = NULL; - } else { + } else al->addr = al->map->map_ip(al->map, al->addr); +} + +void thread__find_addr_location(struct thread *self, + struct perf_session *session, u8 cpumode, + enum map_type type, u64 addr, + struct addr_location *al, + symbol_filter_t filter) +{ + thread__find_addr_map(self, session, cpumode, type, addr, al); + if (al->map != NULL) al->sym = map__find_symbol(al->map, session, al->addr, filter); - } + else + al->sym = NULL; } static void dso__calc_col_width(struct dso *self) diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index c06c135..e35653c 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -48,6 +48,11 @@ static inline struct map *thread__find_map(struct thread *self, return self ? map_groups__find(&self->mg, type, addr) : NULL; } +void thread__find_addr_map(struct thread *self, + struct perf_session *session, u8 cpumode, + enum map_type type, u64 addr, + struct addr_location *al); + void thread__find_addr_location(struct thread *self, struct perf_session *session, u8 cpumode, enum map_type type, u64 addr, -- cgit v0.10.2 From 88d3d9b7c843a42cb73c55a2d13cd1041da31fb9 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 14 Jan 2010 23:45:30 -0200 Subject: perf buildid-list: Introduce --with-hits option MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Using this option 'perf buildid-list' will process all samples, marking the DSOs that had some hits to list just them. This in turn will be used by a new porcelain, 'perf archive', that will be just a shell script to create a tarball from the 'perf buildid-list --with-hits' output and the files cached by 'perf record' in ~/.debug. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263519930-22803-4-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-buildid-list.c b/tools/perf/builtin-buildid-list.c index 4229c2c..431f204 100644 --- a/tools/perf/builtin-buildid-list.c +++ b/tools/perf/builtin-buildid-list.c @@ -16,6 +16,7 @@ static char const *input_name = "perf.data"; static int force; +static bool with_hits; static const char * const buildid_list_usage[] = { "perf buildid-list []", @@ -23,6 +24,7 @@ static const char * const buildid_list_usage[] = { }; static const struct option options[] = { + OPT_BOOLEAN('H', "with-hits", &with_hits, "Show only DSOs with hits"), OPT_STRING('i', "input", &input_name, "file", "input file name"), OPT_BOOLEAN('f', "force", &force, "don't complain, do it"), @@ -31,6 +33,34 @@ static const struct option options[] = { OPT_END() }; +static int build_id_list__process_event(event_t *event, + struct perf_session *session) +{ + struct addr_location al; + u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; + struct thread *thread = perf_session__findnew(session, event->ip.pid); + + if (thread == NULL) { + pr_err("problem processing %d event, skipping it.\n", + event->header.type); + return -1; + } + + thread__find_addr_map(thread, session, cpumode, MAP__FUNCTION, + event->ip.ip, &al); + + if (al.map != NULL) + al.map->dso->hit = 1; + + return 0; +} + +static struct perf_event_ops build_id_list__event_ops = { + .sample = build_id_list__process_event, + .mmap = event__process_mmap, + .fork = event__process_task, +}; + static int __cmd_buildid_list(void) { int err = -1; @@ -40,7 +70,10 @@ static int __cmd_buildid_list(void) if (session == NULL) return -1; - dsos__fprintf_buildid(stdout); + if (with_hits) + perf_session__process_events(session, &build_id_list__event_ops); + + dsos__fprintf_buildid(stdout, with_hits); perf_session__delete(session); return err; diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 4267138..a4e7459 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1716,22 +1716,25 @@ void dsos__fprintf(FILE *fp) __dsos__fprintf(&dsos__user, fp); } -static size_t __dsos__fprintf_buildid(struct list_head *head, FILE *fp) +static size_t __dsos__fprintf_buildid(struct list_head *head, FILE *fp, + bool with_hits) { struct dso *pos; size_t ret = 0; list_for_each_entry(pos, head, node) { + if (with_hits && !pos->hit) + continue; ret += dso__fprintf_buildid(pos, fp); ret += fprintf(fp, " %s\n", pos->long_name); } return ret; } -size_t dsos__fprintf_buildid(FILE *fp) +size_t dsos__fprintf_buildid(FILE *fp, bool with_hits) { - return (__dsos__fprintf_buildid(&dsos__kernel, fp) + - __dsos__fprintf_buildid(&dsos__user, fp)); + return (__dsos__fprintf_buildid(&dsos__kernel, fp, with_hits) + + __dsos__fprintf_buildid(&dsos__user, fp, with_hits)); } static struct dso *dsos__create_kernel(const char *vmlinux) diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 36b7c71..525085fd 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -97,6 +97,7 @@ struct dso { u8 slen_calculated:1; u8 has_build_id:1; u8 kernel:1; + u8 hit:1; unsigned char origin; u8 sorted_by_name; u8 loaded; @@ -129,7 +130,7 @@ struct perf_session; int dso__load(struct dso *self, struct map *map, struct perf_session *session, symbol_filter_t filter); void dsos__fprintf(FILE *fp); -size_t dsos__fprintf_buildid(FILE *fp); +size_t dsos__fprintf_buildid(FILE *fp, bool with_hits); size_t dso__fprintf_buildid(struct dso *self, FILE *fp); size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp); -- cgit v0.10.2 From 460848fceffc91652b2d36d19db4ac40d12fb607 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 15 Jan 2010 13:17:51 -0200 Subject: perf symbols: The synthesized kernel modules MMAP must use the pathnames MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since we use ->long_name in dsos__find now. Now 'perf buildid_list' is not duplicating those and managing to show the proper build-ids for the DSOs with hits: [root@doppio linux-2.6-tip]# perf buildid-list -H 74f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms] 9ffdcac0a7935922d1f04b6cc9029dfef0f066ef /lib/modules/2.6.33-rc4-tip+/kernel/arch/x86/crypto/aes-x86_64.ko 3aaf89c32ebfc438ff546c93597d41788e3e65f3 /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwl3945.ko 19f46033f73e1ec612937189bb118c5daba5a0c8 /lib/modules/2.6.33-rc4-tip+/kernel/net/mac80211/mac80211.ko 1772f014a7a7272859655acb0c64a20ab20b75ee /lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 /lib64/libc-2.10.2.so 5c68f7afeb33309c78037e374b0deee84dd441f6 /lib64/libpthread-2.10.2.so e9c9ad5c138ef882e4507d2605645b597da43873 /bin/dbus-daemon bcda7d09eb6c9ee380dae0ed3d591d4311decc31 /lib64/libdbus-1.so.3.4.0 7cc449a77f48b85d6088114000e970ced613bed8 /usr/lib64/libcrypto.so.0.9.8k fdd1ccd1ff7917ab020653147ab3bacf0a85b5b9 /lib64/libglib-2.0.so.0.2000.5 e4417ebb8762e5f2eee93c8011a71115ff5edad8 /lib64/libgobject-2.0.so.0.2000.5 931e49461f6df99104f0febcc52f6fed5e2efce6 /usr/sbin/sshd dab5f724c088f89fbd8304da553ed6cb30bbec96 /usr/lib64/libgdk-x11-2.0.so.0.1600.6 f2037a091ef36b591187a858d75e203690ea9409 /usr/sbin/openvpn a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f /bin/find 81120aada06e68b1e85882925a0fc6d7345ef59a /home/acme/bin/perf [root@doppio linux-2.6-tip]# Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263568672-30323-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 5a6e827..966d207 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -345,15 +345,15 @@ int event__process_mmap(event_t *self, struct perf_session *session) map = perf_session__new_module_map(session, self->mmap.start, - short_module_name); + self->mmap.filename); if (map == NULL) goto out_problem; - name = strdup(self->mmap.filename); + name = strdup(short_module_name); if (name == NULL) goto out_problem; - dso__set_long_name(map->dso, name); + map->dso->short_name = name; map->end = map->start + self->mmap.len; } else if (memcmp(self->mmap.filename, kmmap_prefix, sizeof(kmmap_prefix) - 1) == 0) { -- cgit v0.10.2 From 2c5851747bcf751908c02e253cb7582d342b4612 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 15 Jan 2010 13:17:52 -0200 Subject: perf archive: Add helper script to package files needed to do analysis MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit It uses 'perf buildid-list --with-hits' to create a tarball with what is needed to have in the destination machine ~/.debug hierarchy to properly decode the perf.data file specified. Here is an example where a perf.data file collected on a x86-64 machine running Fedora 12 is used and then the data is packaged, transferred and decoded on a PARISC64 machine running Debian Testing, 32-bit userspace: [root@doppio linux-2.6-tip]# uname -a Linux doppio.ghostprotocols.net 2.6.33-rc4-tip+ #3 SMP Wed Jan 13 11:58:15 BRST 2010 x86_64 x86_64 x86_64 GNU/Linux [root@doppio linux-2.6-tip]# perf archive [root@doppio linux-2.6-tip]# ls -la perf.data* -rw------- 1 root root 737696 2010-01-14 23:36 perf.data -rw-r--r-- 1 root root 8840025 2010-01-15 12:27 perf.data.tar.bz2 [root@doppio linux-2.6-tip]# scp perf.data.* parisc64:. Password: perf.data.tar.bz2 100% 8633KB 1.4MB/s 00:06 [root@doppio linux-2.6-tip]# ssh parisc64 Password: Linux parisc 2.6.19-g2bbf29ac-dirty #1 Sun Dec 3 17:24:04 BRST 2006 parisc64 The programs included with the Debian GNU/Linux system are free software; the exact distribution terms for each program are described in the individual files in /usr/share/doc/*/copyright. Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent permitted by applicable law. Last login: Thu Jan 14 11:23:24 2010 from d parisc:~# uname -a Linux parisc 2.6.19-g2bbf29ac-dirty #1 Sun Dec 3 17:24:04 BRST 2006 parisc64 GNU/Linux parisc:~# mkdir .debug parisc:~# tar xvf perf.data.tar.bz2 -C ~/.debug tar: Record size = 8 blocks .build-id/74/f9930ee94475b6b3238caf3725a50d59cb994b [kernel.kallsyms]/74f9930ee94475b6b3238caf3725a50d59cb994b .build-id/9f/fdcac0a7935922d1f04b6cc9029dfef0f066ef lib/modules/2.6.33-rc4-tip+/kernel/arch/x86/crypto/aes-x86_64.ko/9ffdcac0a7935922d1f04b6cc9029dfef0f066ef .build-id/3a/af89c32ebfc438ff546c93597d41788e3e65f3 lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/wireless/iwlwifi/iwl3945.ko/3aaf89c32ebfc438ff546c93597d41788e3e65f3 .build-id/19/f46033f73e1ec612937189bb118c5daba5a0c8 lib/modules/2.6.33-rc4-tip+/kernel/net/mac80211/mac80211.ko/19f46033f73e1ec612937189bb118c5daba5a0c8 .build-id/17/72f014a7a7272859655acb0c64a20ab20b75ee lib/modules/2.6.33-rc4-tip+/kernel/drivers/net/e1000e/e1000e.ko/1772f014a7a7272859655acb0c64a20ab20b75ee .build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 .build-id/5c/68f7afeb33309c78037e374b0deee84dd441f6 lib64/libpthread-2.10.2.so/5c68f7afeb33309c78037e374b0deee84dd441f6 .build-id/e9/c9ad5c138ef882e4507d2605645b597da43873 bin/dbus-daemon/e9c9ad5c138ef882e4507d2605645b597da43873 .build-id/bc/da7d09eb6c9ee380dae0ed3d591d4311decc31 lib64/libdbus-1.so.3.4.0/bcda7d09eb6c9ee380dae0ed3d591d4311decc31 .build-id/7c/c449a77f48b85d6088114000e970ced613bed8 usr/lib64/libcrypto.so.0.9.8k/7cc449a77f48b85d6088114000e970ced613bed8 .build-id/fd/d1ccd1ff7917ab020653147ab3bacf0a85b5b9 lib64/libglib-2.0.so.0.2000.5/fdd1ccd1ff7917ab020653147ab3bacf0a85b5b9 .build-id/e4/417ebb8762e5f2eee93c8011a71115ff5edad8 lib64/libgobject-2.0.so.0.2000.5/e4417ebb8762e5f2eee93c8011a71115ff5edad8 .build-id/93/1e49461f6df99104f0febcc52f6fed5e2efce6 usr/sbin/sshd/931e49461f6df99104f0febcc52f6fed5e2efce6 .build-id/da/b5f724c088f89fbd8304da553ed6cb30bbec96 usr/lib64/libgdk-x11-2.0.so.0.1600.6/dab5f724c088f89fbd8304da553ed6cb30bbec96 .build-id/f2/037a091ef36b591187a858d75e203690ea9409 usr/sbin/openvpn/f2037a091ef36b591187a858d75e203690ea9409 .build-id/a8/e4f743b40fb1fd8b85e2f9b88d93b661472b8f bin/find/a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f .build-id/81/120aada06e68b1e85882925a0fc6d7345ef59a home/acme/bin/perf/81120aada06e68b1e85882925a0fc6d7345ef59a parisc:~# perf report 2> /dev/null | head -25 9.07% find find [.] 0x0000000000fb0e 3.29% perf libc-2.10.2.so [.] __GI_strcmp 3.19% find [kernel.kallsyms] [k] _raw_spin_unlock_irqrestore 2.70% find libc-2.10.2.so [.] __GI_memmove 2.62% perf [kernel.kallsyms] [k] vsnprintf 2.03% find libc-2.10.2.so [.] _int_malloc 2.02% perf [kernel.kallsyms] [k] format_decode 1.70% find [kernel.kallsyms] [k] n_tty_write 1.70% find [kernel.kallsyms] [k] half_md4_transform 1.67% find libc-2.10.2.so [.] _IO_vfprintf_internal 1.66% perf [kernel.kallsyms] [k] audit_free_aux 1.62% swapper [kernel.kallsyms] [k] mwait_idle_with_hints 1.58% find [kernel.kallsyms] [k] __kmalloc 1.35% find [kernel.kallsyms] [k] sched_clock_local 1.35% find [kernel.kallsyms] [k] ext4_check_dir_entry 1.35% find [kernel.kallsyms] [k] ext4_htree_store_dirent 1.35% find [kernel.kallsyms] [k] sys_write 1.35% find [e1000e] [k] e1000_clean 1.35% find [kernel.kallsyms] [k] _atomic_dec_and_lock 1.34% find [kernel.kallsyms] [k] __d_lookup parisc:~# Probably the next step is to have 'perf report' notice that there is a perf.data.tar.bz2 file in the same directory and look if it was already added to ~/.debug/. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263568672-30323-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 2c03a94..d739552 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -286,11 +286,7 @@ SCRIPT_PERL = SCRIPT_SH = TEST_PROGRAMS = -# -# No scripts right now: -# - -# SCRIPT_SH += perf-am.sh +SCRIPT_SH += perf-archive.sh # # No Perl scripts right now: @@ -315,9 +311,7 @@ PROGRAMS += perf # List built-in command $C whose implementation cmd_$C() is not in # builtin-$C.o but is linked in as part of some other command. # -# None right now: -# -# BUILT_INS += perf-init $X +BUILT_INS += perf-archive # what 'all' will build and 'install' will install, in perfexecdir ALL_PROGRAMS = $(PROGRAMS) $(SCRIPTS) diff --git a/tools/perf/command-list.txt b/tools/perf/command-list.txt index 71dc7c3..f73d1d9 100644 --- a/tools/perf/command-list.txt +++ b/tools/perf/command-list.txt @@ -3,6 +3,7 @@ # command name category [deprecated] [common] # perf-annotate mainporcelain common +perf-archive mainporcelain perf-bench mainporcelain common perf-buildid-list mainporcelain common perf-diff mainporcelain common diff --git a/tools/perf/perf-archive.sh b/tools/perf/perf-archive.sh new file mode 100644 index 0000000..45fbe2f --- /dev/null +++ b/tools/perf/perf-archive.sh @@ -0,0 +1,32 @@ +#!/bin/bash +# perf archive +# Arnaldo Carvalho de Melo + +PERF_DATA=perf.data +if [ $# -ne 0 ] ; then + PERF_DATA=$1 +fi + +DEBUGDIR=~/.debug/ +BUILDIDS=$(mktemp /tmp/perf-archive-buildids.XXXXXX) + +perf buildid-list -i $PERF_DATA --with-hits > $BUILDIDS +if [ ! -s $BUILDIDS ] ; then + echo "perf archive: no build-ids found" + rm -f $BUILDIDS + exit 1 +fi + +MANIFEST=$(mktemp /tmp/perf-archive-manifest.XXXXXX) + +cut -d ' ' -f 1 $BUILDIDS | \ +while read build_id ; do + linkname=$DEBUGDIR.build-id/${build_id:0:2}/${build_id:2} + filename=$(readlink -f $linkname) + echo ${linkname#$DEBUGDIR} >> $MANIFEST + echo ${filename#$DEBUGDIR} >> $MANIFEST +done + +tar cfj $PERF_DATA.tar.bz2 -C $DEBUGDIR -T $MANIFEST +rm -f $MANIFEST $BUILDIDS +exit 0 -- cgit v0.10.2 From f5a2c3dce03621b55f84496f58adc2d1a87ca16f Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 15 Jan 2010 18:08:26 -0200 Subject: perf record: Intercept all events MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The event interception we need to do in 'perf record' to create a list of all DSOs in PERF_RECORD_MMAP events wasn't seeing all events, make sure that happens by checking size agains event_t->header.size. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263586107-1756-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 614fa9a..7bb9ca1 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -113,16 +113,24 @@ static void write_output(void *buf, size_t size) static void write_event(event_t *buf, size_t size) { - /* - * Add it to the list of DSOs, so that when we finish this - * record session we can pick the available build-ids. - */ - if (buf->header.type == PERF_RECORD_MMAP) { - struct list_head *head = &dsos__user; - if (buf->mmap.header.misc == 1) - head = &dsos__kernel; - __dsos__findnew(head, buf->mmap.filename); - } + size_t processed_size = buf->header.size; + event_t *ev = buf; + + do { + /* + * Add it to the list of DSOs, so that when we finish this + * record session we can pick the available build-ids. + */ + if (ev->header.type == PERF_RECORD_MMAP) { + struct list_head *head = &dsos__user; + if (ev->header.misc == 1) + head = &dsos__kernel; + __dsos__findnew(head, ev->mmap.filename); + } + + ev = ((void *)ev) + ev->header.size; + processed_size += ev->header.size; + } while (processed_size < size); write_output(buf, size); } -- cgit v0.10.2 From 881516eb828a3f7276c378bcef96b7788fc99016 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 15 Jan 2010 18:08:27 -0200 Subject: perf symbols: Accept an alias when looking for "_text" MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As it is in PARISC64: parisc:~# uname -a Linux parisc 2.6.33-rc4-tip+ #1 SMP Thu Jan 14 13:33:34 BRST 2010 parisc64 GNU/Linux parisc:~# grep -w _text /proc/kallsyms 0000000040100000 A _text parisc:~# grep 0000000040100000 /proc/kallsyms 0000000040100000 T stext 0000000040100000 T _stext 0000000040100000 A _text parisc:~# Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263586107-1756-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 966d207..dc13cad 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -227,7 +227,12 @@ static int find_symbol_cb(void *arg, const char *name, char type, u64 start) { struct process_symbol_args *args = arg; - if (!symbol_type__is_a(type, MAP__FUNCTION) || strcmp(name, args->name)) + /* + * Must be a function or at least an alias, as in PARISC64, where "_text" is + * an 'A' to the same address as "_stext". + */ + if (!(symbol_type__is_a(type, MAP__FUNCTION) || + type == 'A') || strcmp(name, args->name)) return 0; args->start = start; -- cgit v0.10.2 From 889ff0150661512d79484219612b7e2e024b6c07 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 9 Jan 2010 20:04:47 +0100 Subject: perf/core: Split context's event group list into pinned and non-pinned lists Split-up struct perf_event_context::group_list into pinned_groups and flexible_groups (non-pinned). This first appears to be useless as it duplicates various loops around the group list handlings. But it scales better in the fast-path in perf_sched_in(). We don't anymore iterate twice through the entire list to separate pinned and non-pinned scheduling. Instead we interate through two distinct lists. The another desired effect is that it makes easier to define distinct scheduling rules on both. Changes in v2: - Respectively rename pinned_grp_list and volatile_grp_list into pinned_groups and flexible_groups as per Ingo suggestion. - Various cleanups Signed-off-by: Frederic Weisbecker Acked-by: Peter Zijlstra Cc: Paul Mackerras Cc: Ingo Molnar Cc: Arnaldo Carvalho de Melo diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 9a1d276..cdbc2aa 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -683,7 +683,8 @@ struct perf_event_context { */ struct mutex mutex; - struct list_head group_list; + struct list_head pinned_groups; + struct list_head flexible_groups; struct list_head event_list; int nr_events; int nr_active; diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 27f69a0..c9f8a75 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -289,6 +289,15 @@ static void update_event_times(struct perf_event *event) event->total_time_running = run_end - event->tstamp_running; } +static struct list_head * +ctx_group_list(struct perf_event *event, struct perf_event_context *ctx) +{ + if (event->attr.pinned) + return &ctx->pinned_groups; + else + return &ctx->flexible_groups; +} + /* * Add a event from the lists for its context. * Must be called with ctx->mutex and ctx->lock held. @@ -303,9 +312,12 @@ list_add_event(struct perf_event *event, struct perf_event_context *ctx) * add it straight to the context's event list, or to the group * leader's sibling list: */ - if (group_leader == event) - list_add_tail(&event->group_entry, &ctx->group_list); - else { + if (group_leader == event) { + struct list_head *list; + + list = ctx_group_list(event, ctx); + list_add_tail(&event->group_entry, list); + } else { list_add_tail(&event->group_entry, &group_leader->sibling_list); group_leader->nr_siblings++; } @@ -355,8 +367,10 @@ list_del_event(struct perf_event *event, struct perf_event_context *ctx) * to the context list directly: */ list_for_each_entry_safe(sibling, tmp, &event->sibling_list, group_entry) { + struct list_head *list; - list_move_tail(&sibling->group_entry, &ctx->group_list); + list = ctx_group_list(event, ctx); + list_move_tail(&sibling->group_entry, list); sibling->group_leader = sibling; } } @@ -1056,7 +1070,10 @@ void __perf_event_sched_out(struct perf_event_context *ctx, perf_disable(); if (ctx->nr_active) { - list_for_each_entry(event, &ctx->group_list, group_entry) + list_for_each_entry(event, &ctx->pinned_groups, group_entry) + group_sched_out(event, cpuctx, ctx); + + list_for_each_entry(event, &ctx->flexible_groups, group_entry) group_sched_out(event, cpuctx, ctx); } perf_enable(); @@ -1271,9 +1288,8 @@ __perf_event_sched_in(struct perf_event_context *ctx, * First go through the list and put on any pinned groups * in order to give them the best chance of going on. */ - list_for_each_entry(event, &ctx->group_list, group_entry) { - if (event->state <= PERF_EVENT_STATE_OFF || - !event->attr.pinned) + list_for_each_entry(event, &ctx->pinned_groups, group_entry) { + if (event->state <= PERF_EVENT_STATE_OFF) continue; if (event->cpu != -1 && event->cpu != cpu) continue; @@ -1291,15 +1307,10 @@ __perf_event_sched_in(struct perf_event_context *ctx, } } - list_for_each_entry(event, &ctx->group_list, group_entry) { - /* - * Ignore events in OFF or ERROR state, and - * ignore pinned events since we did them already. - */ - if (event->state <= PERF_EVENT_STATE_OFF || - event->attr.pinned) + list_for_each_entry(event, &ctx->flexible_groups, group_entry) { + /* Ignore events in OFF or ERROR state */ + if (event->state <= PERF_EVENT_STATE_OFF) continue; - /* * Listen to the 'cpu' scheduling filter constraint * of events: @@ -1453,8 +1464,13 @@ static void rotate_ctx(struct perf_event_context *ctx) * Rotate the first entry last (works just fine for group events too): */ perf_disable(); - list_for_each_entry(event, &ctx->group_list, group_entry) { - list_move_tail(&event->group_entry, &ctx->group_list); + list_for_each_entry(event, &ctx->pinned_groups, group_entry) { + list_move_tail(&event->group_entry, &ctx->pinned_groups); + break; + } + + list_for_each_entry(event, &ctx->flexible_groups, group_entry) { + list_move_tail(&event->group_entry, &ctx->flexible_groups); break; } perf_enable(); @@ -1490,6 +1506,21 @@ void perf_event_task_tick(struct task_struct *curr) perf_event_task_sched_in(curr); } +static int event_enable_on_exec(struct perf_event *event, + struct perf_event_context *ctx) +{ + if (!event->attr.enable_on_exec) + return 0; + + event->attr.enable_on_exec = 0; + if (event->state >= PERF_EVENT_STATE_INACTIVE) + return 0; + + __perf_event_mark_enabled(event, ctx); + + return 1; +} + /* * Enable all of a task's events that have been marked enable-on-exec. * This expects task == current. @@ -1500,6 +1531,7 @@ static void perf_event_enable_on_exec(struct task_struct *task) struct perf_event *event; unsigned long flags; int enabled = 0; + int ret; local_irq_save(flags); ctx = task->perf_event_ctxp; @@ -1510,14 +1542,16 @@ static void perf_event_enable_on_exec(struct task_struct *task) raw_spin_lock(&ctx->lock); - list_for_each_entry(event, &ctx->group_list, group_entry) { - if (!event->attr.enable_on_exec) - continue; - event->attr.enable_on_exec = 0; - if (event->state >= PERF_EVENT_STATE_INACTIVE) - continue; - __perf_event_mark_enabled(event, ctx); - enabled = 1; + list_for_each_entry(event, &ctx->pinned_groups, group_entry) { + ret = event_enable_on_exec(event, ctx); + if (ret) + enabled = 1; + } + + list_for_each_entry(event, &ctx->flexible_groups, group_entry) { + ret = event_enable_on_exec(event, ctx); + if (ret) + enabled = 1; } /* @@ -1591,7 +1625,8 @@ __perf_event_init_context(struct perf_event_context *ctx, { raw_spin_lock_init(&ctx->lock); mutex_init(&ctx->mutex); - INIT_LIST_HEAD(&ctx->group_list); + INIT_LIST_HEAD(&ctx->pinned_groups); + INIT_LIST_HEAD(&ctx->flexible_groups); INIT_LIST_HEAD(&ctx->event_list); atomic_set(&ctx->refcount, 1); ctx->task = task; @@ -5032,7 +5067,11 @@ void perf_event_exit_task(struct task_struct *child) mutex_lock_nested(&child_ctx->mutex, SINGLE_DEPTH_NESTING); again: - list_for_each_entry_safe(child_event, tmp, &child_ctx->group_list, + list_for_each_entry_safe(child_event, tmp, &child_ctx->pinned_groups, + group_entry) + __perf_event_exit_task(child_event, child_ctx, child); + + list_for_each_entry_safe(child_event, tmp, &child_ctx->flexible_groups, group_entry) __perf_event_exit_task(child_event, child_ctx, child); @@ -5041,7 +5080,8 @@ again: * its siblings to the list, but we obtained 'tmp' before that which * will still point to the list head terminating the iteration. */ - if (!list_empty(&child_ctx->group_list)) + if (!list_empty(&child_ctx->pinned_groups) || + !list_empty(&child_ctx->flexible_groups)) goto again; mutex_unlock(&child_ctx->mutex); @@ -5049,6 +5089,24 @@ again: put_ctx(child_ctx); } +static void perf_free_event(struct perf_event *event, + struct perf_event_context *ctx) +{ + struct perf_event *parent = event->parent; + + if (WARN_ON_ONCE(!parent)) + return; + + mutex_lock(&parent->child_mutex); + list_del_init(&event->child_list); + mutex_unlock(&parent->child_mutex); + + fput(parent->filp); + + list_del_event(event, ctx); + free_event(event); +} + /* * free an unexposed, unused context as created by inheritance by * init_task below, used by fork() in case of fail. @@ -5063,36 +5121,70 @@ void perf_event_free_task(struct task_struct *task) mutex_lock(&ctx->mutex); again: - list_for_each_entry_safe(event, tmp, &ctx->group_list, group_entry) { - struct perf_event *parent = event->parent; + list_for_each_entry_safe(event, tmp, &ctx->pinned_groups, group_entry) + perf_free_event(event, ctx); - if (WARN_ON_ONCE(!parent)) - continue; + list_for_each_entry_safe(event, tmp, &ctx->flexible_groups, + group_entry) + perf_free_event(event, ctx); - mutex_lock(&parent->child_mutex); - list_del_init(&event->child_list); - mutex_unlock(&parent->child_mutex); + if (!list_empty(&ctx->pinned_groups) || + !list_empty(&ctx->flexible_groups)) + goto again; - fput(parent->filp); + mutex_unlock(&ctx->mutex); - list_del_event(event, ctx); - free_event(event); + put_ctx(ctx); +} + +static int +inherit_task_group(struct perf_event *event, struct task_struct *parent, + struct perf_event_context *parent_ctx, + struct task_struct *child, + int *inherited_all) +{ + int ret; + struct perf_event_context *child_ctx = child->perf_event_ctxp; + + if (!event->attr.inherit) { + *inherited_all = 0; + return 0; } - if (!list_empty(&ctx->group_list)) - goto again; + if (!child_ctx) { + /* + * This is executed from the parent task context, so + * inherit events that have been marked for cloning. + * First allocate and initialize a context for the + * child. + */ - mutex_unlock(&ctx->mutex); + child_ctx = kzalloc(sizeof(struct perf_event_context), + GFP_KERNEL); + if (!child_ctx) + return -ENOMEM; - put_ctx(ctx); + __perf_event_init_context(child_ctx, child); + child->perf_event_ctxp = child_ctx; + get_task_struct(child); + } + + ret = inherit_group(event, parent, parent_ctx, + child, child_ctx); + + if (ret) + *inherited_all = 0; + + return ret; } + /* * Initialize the perf_event context in task_struct */ int perf_event_init_task(struct task_struct *child) { - struct perf_event_context *child_ctx = NULL, *parent_ctx; + struct perf_event_context *child_ctx, *parent_ctx; struct perf_event_context *cloned_ctx; struct perf_event *event; struct task_struct *parent = current; @@ -5130,41 +5222,22 @@ int perf_event_init_task(struct task_struct *child) * We dont have to disable NMIs - we are only looking at * the list, not manipulating it: */ - list_for_each_entry(event, &parent_ctx->group_list, group_entry) { - - if (!event->attr.inherit) { - inherited_all = 0; - continue; - } - - if (!child->perf_event_ctxp) { - /* - * This is executed from the parent task context, so - * inherit events that have been marked for cloning. - * First allocate and initialize a context for the - * child. - */ - - child_ctx = kzalloc(sizeof(struct perf_event_context), - GFP_KERNEL); - if (!child_ctx) { - ret = -ENOMEM; - break; - } - - __perf_event_init_context(child_ctx, child); - child->perf_event_ctxp = child_ctx; - get_task_struct(child); - } + list_for_each_entry(event, &parent_ctx->pinned_groups, group_entry) { + ret = inherit_task_group(event, parent, parent_ctx, child, + &inherited_all); + if (ret) + break; + } - ret = inherit_group(event, parent, parent_ctx, - child, child_ctx); - if (ret) { - inherited_all = 0; + list_for_each_entry(event, &parent_ctx->flexible_groups, group_entry) { + ret = inherit_task_group(event, parent, parent_ctx, child, + &inherited_all); + if (ret) break; - } } + child_ctx = child->perf_event_ctxp; + if (child_ctx && inherited_all) { /* * Mark the child context as a clone of the parent @@ -5213,7 +5286,9 @@ static void __perf_event_exit_cpu(void *info) struct perf_event_context *ctx = &cpuctx->ctx; struct perf_event *event, *tmp; - list_for_each_entry_safe(event, tmp, &ctx->group_list, group_entry) + list_for_each_entry_safe(event, tmp, &ctx->pinned_groups, group_entry) + __perf_event_remove_from_context(event); + list_for_each_entry_safe(event, tmp, &ctx->flexible_groups, group_entry) __perf_event_remove_from_context(event); } static void perf_event_exit_cpu(int cpu) -- cgit v0.10.2 From 5908cdc85eb30f8d07f2cb11d4a62334d7229048 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 9 Jan 2010 20:53:14 +0100 Subject: list: Introduce list_rotate_left() Bring a new list_rotate_left() helper that rotates a list to the left. This is useful for codes that need to round roubin elements which queue priority increases from tail to head. Signed-off-by: Frederic Weisbecker Acked-by: Peter Zijlstra Cc: Paul Mackerras Cc: Ingo Molnar Cc: Arnaldo Carvalho de Melo diff --git a/include/linux/list.h b/include/linux/list.h index 969f6e9..5d9c655 100644 --- a/include/linux/list.h +++ b/include/linux/list.h @@ -206,6 +206,20 @@ static inline int list_empty_careful(const struct list_head *head) } /** + * list_rotate_left - rotate the list to the left + * @head: the head of the list + */ +static inline void list_rotate_left(struct list_head *head) +{ + struct list_head *first; + + if (!list_empty(head)) { + first = head->next; + list_move_tail(first, head); + } +} + +/** * list_is_singular - tests whether a list has just one entry. * @head: the list to test. */ -- cgit v0.10.2 From e286417378b4f9ce6e473b556193465ab22e12ab Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 9 Jan 2010 21:05:28 +0100 Subject: perf: Round robin flexible groups of events using list_rotate_left() This is more proper that doing it through a list_for_each_entry() that breaks after the first entry. v2: Don't rotate pinned groups as its not needed to time share them. Signed-off-by: Frederic Weisbecker Acked-by: Peter Zijlstra Cc: Paul Mackerras Cc: Ingo Molnar Cc: Arnaldo Carvalho de Melo diff --git a/kernel/perf_event.c b/kernel/perf_event.c index c9f8a75..bbebe28 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1454,25 +1454,16 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx) */ static void rotate_ctx(struct perf_event_context *ctx) { - struct perf_event *event; - if (!ctx->nr_events) return; raw_spin_lock(&ctx->lock); - /* - * Rotate the first entry last (works just fine for group events too): - */ + + /* Rotate the first entry last of non-pinned groups */ perf_disable(); - list_for_each_entry(event, &ctx->pinned_groups, group_entry) { - list_move_tail(&event->group_entry, &ctx->pinned_groups); - break; - } - list_for_each_entry(event, &ctx->flexible_groups, group_entry) { - list_move_tail(&event->group_entry, &ctx->flexible_groups); - break; - } + list_rotate_left(&ctx->flexible_groups); + perf_enable(); raw_spin_unlock(&ctx->lock); -- cgit v0.10.2 From d6f962b57bfaab62891c7abbf1469212a56d6103 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 10 Jan 2010 01:25:51 +0100 Subject: perf: Export software-only event group characteristic as a flag Before scheduling an event group, we first check if a group can go on. We first check if the group is made of software only events first, in which case it is enough to know if the group can be scheduled in. For that purpose, we iterate through the whole group, which is wasteful as we could do this check when we add/delete an event to a group. So we create a group_flags field in perf event that can host characteristics from a group of events, starting with a first PERF_GROUP_SOFTWARE flag that reduces the check on the fast path. Signed-off-by: Frederic Weisbecker Acked-by: Peter Zijlstra Cc: Paul Mackerras Cc: Ingo Molnar Cc: Arnaldo Carvalho de Melo diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index cdbc2aa..c6f812e 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -565,6 +565,10 @@ typedef void (*perf_overflow_handler_t)(struct perf_event *, int, struct perf_sample_data *, struct pt_regs *regs); +enum perf_group_flag { + PERF_GROUP_SOFTWARE = 0x1, +}; + /** * struct perf_event - performance event kernel representation: */ @@ -574,6 +578,7 @@ struct perf_event { struct list_head event_entry; struct list_head sibling_list; int nr_siblings; + int group_flags; struct perf_event *group_leader; struct perf_event *output; const struct pmu *pmu; diff --git a/kernel/perf_event.c b/kernel/perf_event.c index bbebe28..eae6ff6 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -315,9 +315,16 @@ list_add_event(struct perf_event *event, struct perf_event_context *ctx) if (group_leader == event) { struct list_head *list; + if (is_software_event(event)) + event->group_flags |= PERF_GROUP_SOFTWARE; + list = ctx_group_list(event, ctx); list_add_tail(&event->group_entry, list); } else { + if (group_leader->group_flags & PERF_GROUP_SOFTWARE && + !is_software_event(event)) + group_leader->group_flags &= ~PERF_GROUP_SOFTWARE; + list_add_tail(&event->group_entry, &group_leader->sibling_list); group_leader->nr_siblings++; } @@ -372,6 +379,9 @@ list_del_event(struct perf_event *event, struct perf_event_context *ctx) list = ctx_group_list(event, ctx); list_move_tail(&sibling->group_entry, list); sibling->group_leader = sibling; + + /* Inherit group flags from the previous leader */ + sibling->group_flags = event->group_flags; } } @@ -700,24 +710,6 @@ group_error: } /* - * Return 1 for a group consisting entirely of software events, - * 0 if the group contains any hardware events. - */ -static int is_software_only_group(struct perf_event *leader) -{ - struct perf_event *event; - - if (!is_software_event(leader)) - return 0; - - list_for_each_entry(event, &leader->sibling_list, group_entry) - if (!is_software_event(event)) - return 0; - - return 1; -} - -/* * Work out whether we can put this event group on the CPU now. */ static int group_can_go_on(struct perf_event *event, @@ -727,7 +719,7 @@ static int group_can_go_on(struct perf_event *event, /* * Groups consisting entirely of software events can always go on. */ - if (is_software_only_group(event)) + if (event->group_flags & PERF_GROUP_SOFTWARE) return 1; /* * If an exclusive group is already on, no other hardware -- cgit v0.10.2 From 8eda55142080f0373b1f0268fe6d6807f193e713 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Tue, 17 Nov 2009 21:05:31 +0000 Subject: sh: New extended page flag to wire/unwire TLB entries Provide a new extended page flag, _PAGE_WIRED and an SH4 implementation for wiring TLB entries and use it in the fixmap code path so that we can wire the fixmap TLB entry. Signed-off-by: Matt Fleming diff --git a/arch/sh/include/asm/pgtable_32.h b/arch/sh/include/asm/pgtable_32.h index 5003ee8..c573d45 100644 --- a/arch/sh/include/asm/pgtable_32.h +++ b/arch/sh/include/asm/pgtable_32.h @@ -71,6 +71,8 @@ #define _PAGE_EXT_KERN_WRITE 0x1000 /* EPR4-bit: Kernel space writable */ #define _PAGE_EXT_KERN_READ 0x2000 /* EPR5-bit: Kernel space readable */ +#define _PAGE_EXT_WIRED 0x4000 /* software: Wire TLB entry */ + /* Wrapper for extended mode pgprot twiddling */ #define _PAGE_EXT(x) ((unsigned long long)(x) << 32) @@ -164,6 +166,8 @@ static inline unsigned long copy_ptea_attributes(unsigned long x) (PTE_MASK | _PAGE_ACCESSED | _PAGE_CACHABLE | \ _PAGE_DIRTY | _PAGE_SPECIAL) +#define _PAGE_WIRED (_PAGE_EXT(_PAGE_EXT_WIRED)) + #ifndef __ASSEMBLY__ #if defined(CONFIG_X2TLB) /* SH-X2 TLB */ diff --git a/arch/sh/include/asm/tlb.h b/arch/sh/include/asm/tlb.h index da8fe7a..3ed2f7a 100644 --- a/arch/sh/include/asm/tlb.h +++ b/arch/sh/include/asm/tlb.h @@ -97,6 +97,22 @@ tlb_end_vma(struct mmu_gather *tlb, struct vm_area_struct *vma) #define tlb_migrate_finish(mm) do { } while (0) +#ifdef CONFIG_CPU_SH4 +extern void tlb_wire_entry(struct vm_area_struct *, unsigned long, pte_t); +extern void tlb_unwire_entry(void); +#else +static inline void tlb_wire_entry(struct vm_area_struct *vma , + unsigned long addr, pte_t pte) +{ + BUG(); +} + +static inline void tlb_unwire_entry(void) +{ + BUG(); +} +#endif /* CONFIG_CPU_SH4 */ + #else /* CONFIG_MMU */ #define tlb_start_vma(tlb, vma) do { } while (0) diff --git a/arch/sh/include/cpu-sh4/cpu/mmu_context.h b/arch/sh/include/cpu-sh4/cpu/mmu_context.h index 3ce7ef6..03ea75c 100644 --- a/arch/sh/include/cpu-sh4/cpu/mmu_context.h +++ b/arch/sh/include/cpu-sh4/cpu/mmu_context.h @@ -25,6 +25,10 @@ #define MMUCR_TI (1<<2) +#define MMUCR_URB 0x00FC0000 +#define MMUCR_URB_SHIFT 18 +#define MMUCR_URB_NENTRIES 64 + #if defined(CONFIG_32BIT) && defined(CONFIG_CPU_SUBTYPE_ST40) #define MMUCR_SE (1 << 4) #else diff --git a/arch/sh/mm/tlb-pteaex.c b/arch/sh/mm/tlb-pteaex.c index 409b7c2..88c8bb0 100644 --- a/arch/sh/mm/tlb-pteaex.c +++ b/arch/sh/mm/tlb-pteaex.c @@ -76,3 +76,69 @@ void __uses_jump_to_uncached local_flush_tlb_one(unsigned long asid, __raw_writel(asid, MMU_UTLB_ADDRESS_ARRAY2 | MMU_PAGE_ASSOC_BIT); back_to_cached(); } + +/* + * Load the entry for 'addr' into the TLB and wire the entry. + */ +void tlb_wire_entry(struct vm_area_struct *vma, unsigned long addr, pte_t pte) +{ + unsigned long status, flags; + int urb; + + local_irq_save(flags); + + /* Load the entry into the TLB */ + __update_tlb(vma, addr, pte); + + /* ... and wire it up. */ + status = ctrl_inl(MMUCR); + urb = (status & MMUCR_URB) >> MMUCR_URB_SHIFT; + status &= ~MMUCR_URB; + + /* + * Make sure we're not trying to wire the last TLB entry slot. + */ + BUG_ON(!--urb); + + urb = urb % MMUCR_URB_NENTRIES; + + status |= (urb << MMUCR_URB_SHIFT); + ctrl_outl(status, MMUCR); + ctrl_barrier(); + + local_irq_restore(flags); +} + +/* + * Unwire the last wired TLB entry. + * + * It should also be noted that it is not possible to wire and unwire + * TLB entries in an arbitrary order. If you wire TLB entry N, followed + * by entry N+1, you must unwire entry N+1 first, then entry N. In this + * respect, it works like a stack or LIFO queue. + */ +void tlb_unwire_entry(void) +{ + unsigned long status, flags; + int urb; + + local_irq_save(flags); + + status = ctrl_inl(MMUCR); + urb = (status & MMUCR_URB) >> MMUCR_URB_SHIFT; + status &= ~MMUCR_URB; + + /* + * Make sure we're not trying to unwire a TLB entry when none + * have been wired. + */ + BUG_ON(urb++ == MMUCR_URB_NENTRIES); + + urb = urb % MMUCR_URB_NENTRIES; + + status |= (urb << MMUCR_URB_SHIFT); + ctrl_outl(status, MMUCR); + ctrl_barrier(); + + local_irq_restore(flags); +} diff --git a/arch/sh/mm/tlb-sh4.c b/arch/sh/mm/tlb-sh4.c index 8cf550e..4c62347 100644 --- a/arch/sh/mm/tlb-sh4.c +++ b/arch/sh/mm/tlb-sh4.c @@ -81,3 +81,69 @@ void __uses_jump_to_uncached local_flush_tlb_one(unsigned long asid, ctrl_outl(data, addr); back_to_cached(); } + +/* + * Load the entry for 'addr' into the TLB and wire the entry. + */ +void tlb_wire_entry(struct vm_area_struct *vma, unsigned long addr, pte_t pte) +{ + unsigned long status, flags; + int urb; + + local_irq_save(flags); + + /* Load the entry into the TLB */ + __update_tlb(vma, addr, pte); + + /* ... and wire it up. */ + status = ctrl_inl(MMUCR); + urb = (status & MMUCR_URB) >> MMUCR_URB_SHIFT; + status &= ~MMUCR_URB; + + /* + * Make sure we're not trying to wire the last TLB entry slot. + */ + BUG_ON(!--urb); + + urb = urb % MMUCR_URB_NENTRIES; + + status |= (urb << MMUCR_URB_SHIFT); + ctrl_outl(status, MMUCR); + ctrl_barrier(); + + local_irq_restore(flags); +} + +/* + * Unwire the last wired TLB entry. + * + * It should also be noted that it is not possible to wire and unwire + * TLB entries in an arbitrary order. If you wire TLB entry N, followed + * by entry N+1, you must unwire entry N+1 first, then entry N. In this + * respect, it works like a stack or LIFO queue. + */ +void tlb_unwire_entry(void) +{ + unsigned long status, flags; + int urb; + + local_irq_save(flags); + + status = ctrl_inl(MMUCR); + urb = (status & MMUCR_URB) >> MMUCR_URB_SHIFT; + status &= ~MMUCR_URB; + + /* + * Make sure we're not trying to unwire a TLB entry when none + * have been wired. + */ + BUG_ON(urb++ == MMUCR_URB_NENTRIES); + + urb = urb % MMUCR_URB_NENTRIES; + + status |= (urb << MMUCR_URB_SHIFT); + ctrl_outl(status, MMUCR); + ctrl_barrier(); + + local_irq_restore(flags); +} -- cgit v0.10.2 From 24ef7fc4dcc57afa0c33166c25bfe7676ffd4296 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Thu, 19 Nov 2009 21:11:05 +0000 Subject: sh: Acquire some more page flags for SH-5. We need some more page flags to hook up _PAGE_WIRED (and eventually other things). So use the unused PTE bits above the PPN field as no implementations use these for anything currently. Now that we have _PAGE_WIRED let's provide the SH-5 functions for wiring up TLB entries. Signed-off-by: Matt Fleming diff --git a/arch/sh/include/asm/page.h b/arch/sh/include/asm/page.h index a86c0f1..61e5810 100644 --- a/arch/sh/include/asm/page.h +++ b/arch/sh/include/asm/page.h @@ -88,7 +88,7 @@ typedef struct { unsigned long pgd; } pgd_t; #define __pte(x) ((pte_t) { (x) } ) #else typedef struct { unsigned long long pte_low; } pte_t; -typedef struct { unsigned long pgprot; } pgprot_t; +typedef struct { unsigned long long pgprot; } pgprot_t; typedef struct { unsigned long pgd; } pgd_t; #define pte_val(x) ((x).pte_low) #define __pte(x) ((pte_t) { (x) } ) diff --git a/arch/sh/include/asm/pgtable_64.h b/arch/sh/include/asm/pgtable_64.h index dd38158..0ee4677 100644 --- a/arch/sh/include/asm/pgtable_64.h +++ b/arch/sh/include/asm/pgtable_64.h @@ -123,8 +123,21 @@ static __inline__ void set_pte(pte_t *pteptr, pte_t pteval) #define _PAGE_DIRTY 0x400 /* software: page accessed in write */ #define _PAGE_ACCESSED 0x800 /* software: page referenced */ +/* Wrapper for extended mode pgprot twiddling */ +#define _PAGE_EXT(x) ((unsigned long long)(x) << 32) + +/* + * We can use the sign-extended bits in the PTEL to get 32 bits of + * software flags. This works for now because no implementations uses + * anything above the PPN field. + */ +#define _PAGE_WIRED _PAGE_EXT(0x001) /* software: wire the tlb entry */ + +#define _PAGE_CLEAR_FLAGS (_PAGE_PRESENT | _PAGE_FILE | _PAGE_SHARED | \ + _PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_WIRED) + /* Mask which drops software flags */ -#define _PAGE_FLAGS_HARDWARE_MASK 0xfffffffffffff3dbLL +#define _PAGE_FLAGS_HARDWARE_MASK (NEFF_MASK & ~(_PAGE_CLEAR_FLAGS)) /* * HugeTLB support diff --git a/arch/sh/include/asm/tlb.h b/arch/sh/include/asm/tlb.h index 3ed2f7a..dfc8fcd 100644 --- a/arch/sh/include/asm/tlb.h +++ b/arch/sh/include/asm/tlb.h @@ -11,6 +11,7 @@ #ifdef CONFIG_MMU #include #include +#include /* * TLB handling. This allows us to remove pages from the page @@ -100,6 +101,46 @@ tlb_end_vma(struct mmu_gather *tlb, struct vm_area_struct *vma) #ifdef CONFIG_CPU_SH4 extern void tlb_wire_entry(struct vm_area_struct *, unsigned long, pte_t); extern void tlb_unwire_entry(void); +#elif defined(CONFIG_SUPERH64) +static int dtlb_entry; +static unsigned long long dtlb_entries[64]; + +static inline void tlb_wire_entry(struct vm_area_struct *vma, + unsigned long addr, pte_t pte) +{ + unsigned long long entry; + unsigned long paddr, flags; + + BUG_ON(dtlb_entry == 64); + + local_irq_save(flags); + + entry = sh64_get_wired_dtlb_entry(); + dtlb_entries[dtlb_entry++] = entry; + + paddr = pte_val(pte) & _PAGE_FLAGS_HARDWARE_MASK; + paddr &= ~PAGE_MASK; + + sh64_setup_tlb_slot(entry, addr, get_asid(), paddr); + + local_irq_restore(flags); +} + +static inline void tlb_unwire_entry(void) +{ + unsigned long long entry; + unsigned long flags; + + BUG_ON(!dtlb_entry); + + local_irq_save(flags); + entry = dtlb_entries[dtlb_entry--]; + + sh64_teardown_tlb_slot(entry); + sh64_put_wired_dtlb_entry(entry); + + local_irq_restore(flags); +} #else static inline void tlb_wire_entry(struct vm_area_struct *vma , unsigned long addr, pte_t pte) diff --git a/arch/sh/mm/tlbflush_64.c b/arch/sh/mm/tlbflush_64.c index de0b0e88..706da1d 100644 --- a/arch/sh/mm/tlbflush_64.c +++ b/arch/sh/mm/tlbflush_64.c @@ -36,7 +36,7 @@ extern void die(const char *,struct pt_regs *,long); static inline void print_prots(pgprot_t prot) { - printk("prot is 0x%08lx\n",pgprot_val(prot)); + printk("prot is 0x%016llx\n",pgprot_val(prot)); printk("%s %s %s %s %s\n",PPROT(_PAGE_SHARED),PPROT(_PAGE_READ), PPROT(_PAGE_EXECUTE),PPROT(_PAGE_WRITE),PPROT(_PAGE_USER)); -- cgit v0.10.2 From 07cad4dc1bfdaefd20c6329e9d8179ad1c600e92 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Tue, 17 Nov 2009 22:03:41 +0000 Subject: sh: Generalise the pte handling code for the fixmap path Generalise the code for setting and clearing pte's and allow TLB entries to be pinned and unpinned if the _PAGE_WIRED flag is present. Signed-off-by: Matt Fleming diff --git a/arch/sh/include/asm/fixmap.h b/arch/sh/include/asm/fixmap.h index 5ac1e40..1566d33 100644 --- a/arch/sh/include/asm/fixmap.h +++ b/arch/sh/include/asm/fixmap.h @@ -65,6 +65,7 @@ enum fixed_addresses { extern void __set_fixmap(enum fixed_addresses idx, unsigned long phys, pgprot_t flags); +extern void __clear_fixmap(enum fixed_addresses idx, pgprot_t flags); #define set_fixmap(idx, phys) \ __set_fixmap(idx, phys, PAGE_KERNEL) diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c index d5fb014..30a9b53 100644 --- a/arch/sh/mm/init.c +++ b/arch/sh/mm/init.c @@ -39,7 +39,7 @@ unsigned long cached_to_uncached = P2SEG - P1SEG; #endif #ifdef CONFIG_MMU -static void set_pte_phys(unsigned long addr, unsigned long phys, pgprot_t prot) +static pte_t *__get_pte_phys(unsigned long addr) { pgd_t *pgd; pud_t *pud; @@ -49,22 +49,30 @@ static void set_pte_phys(unsigned long addr, unsigned long phys, pgprot_t prot) pgd = pgd_offset_k(addr); if (pgd_none(*pgd)) { pgd_ERROR(*pgd); - return; + return NULL; } pud = pud_alloc(NULL, pgd, addr); if (unlikely(!pud)) { pud_ERROR(*pud); - return; + return NULL; } pmd = pmd_alloc(NULL, pud, addr); if (unlikely(!pmd)) { pmd_ERROR(*pmd); - return; + return NULL; } pte = pte_offset_kernel(pmd, addr); + return pte; +} + +static void set_pte_phys(unsigned long addr, unsigned long phys, pgprot_t prot) +{ + pte_t *pte; + + pte = __get_pte_phys(addr); if (!pte_none(*pte)) { pte_ERROR(*pte); return; @@ -72,6 +80,22 @@ static void set_pte_phys(unsigned long addr, unsigned long phys, pgprot_t prot) set_pte(pte, pfn_pte(phys >> PAGE_SHIFT, prot)); local_flush_tlb_one(get_asid(), addr); + + if (pgprot_val(prot) & _PAGE_WIRED) + tlb_wire_entry(NULL, addr, *pte); +} + +static void clear_pte_phys(unsigned long addr, pgprot_t prot) +{ + pte_t *pte; + + pte = __get_pte_phys(addr); + + if (pgprot_val(prot) & _PAGE_WIRED) + tlb_unwire_entry(); + + set_pte(pte, pfn_pte(0, __pgprot(0))); + local_flush_tlb_one(get_asid(), addr); } /* @@ -101,6 +125,18 @@ void __set_fixmap(enum fixed_addresses idx, unsigned long phys, pgprot_t prot) set_pte_phys(address, phys, prot); } +void __clear_fixmap(enum fixed_addresses idx, pgprot_t prot) +{ + unsigned long address = __fix_to_virt(idx); + + if (idx >= __end_of_fixed_addresses) { + BUG(); + return; + } + + clear_pte_phys(address, prot); +} + void __init page_table_range_init(unsigned long start, unsigned long end, pgd_t *pgd_base) { -- cgit v0.10.2 From 4d35b93a66e9b87df20784fcf130d2e8760be53f Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Thu, 5 Nov 2009 07:54:17 +0000 Subject: sh: Add fixed ioremap support Some devices need to be ioremap'd and accessed very early in the boot process. It is not possible to use the standard ioremap() function in this case because that requires kmalloc()'ing some virtual address space and kmalloc() may not be available so early in boot. This patch provides fixmap mappings that allow physical address ranges to be remapped into the kernel address space during the early boot stages. Signed-off-by: Matt Fleming diff --git a/arch/sh/include/asm/fixmap.h b/arch/sh/include/asm/fixmap.h index 1566d33..38a1de8 100644 --- a/arch/sh/include/asm/fixmap.h +++ b/arch/sh/include/asm/fixmap.h @@ -60,6 +60,14 @@ enum fixed_addresses { FIX_KMAP_BEGIN, /* reserved pte's for temporary kernel mappings */ FIX_KMAP_END = FIX_KMAP_BEGIN+(KM_TYPE_NR*NR_CPUS)-1, #endif + /* + * FIX_IOREMAP entries are useful for mapping physical address + * space before ioremap() is useable, e.g. really early in boot + * before kmalloc() is working. + */ +#define FIX_N_IOREMAPS 32 + FIX_IOREMAP_BEGIN, + FIX_IOREMAP_END = FIX_IOREMAP_BEGIN + FIX_N_IOREMAPS, __end_of_fixed_addresses }; diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index f4314d8..bee5965 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -237,6 +237,12 @@ void __iomem *__ioremap_caller(unsigned long offset, unsigned long size, unsigned long flags, void *caller); void __iounmap(void __iomem *addr); +#ifdef CONFIG_IOREMAP_FIXED +extern void __iomem *ioremap_fixed(resource_size_t, unsigned long, pgprot_t); +extern void iounmap_fixed(void __iomem *); +extern void ioremap_fixed_init(void); +#endif + static inline void __iomem * __ioremap(unsigned long offset, unsigned long size, unsigned long flags) { diff --git a/arch/sh/kernel/setup.c b/arch/sh/kernel/setup.c index f79ebe3..e187750 100644 --- a/arch/sh/kernel/setup.c +++ b/arch/sh/kernel/setup.c @@ -449,14 +449,15 @@ void __init setup_arch(char **cmdline_p) #ifdef CONFIG_DUMMY_CONSOLE conswitchp = &dummy_con; #endif + paging_init(); + pmb_init(); + + ioremap_fixed_init(); /* Perform the machine specific initialisation */ if (likely(sh_mv.mv_setup)) sh_mv.mv_setup(cmdline_p); - paging_init(); - pmb_init(); - #ifdef CONFIG_SMP plat_smp_setup(); #endif diff --git a/arch/sh/mm/Kconfig b/arch/sh/mm/Kconfig index 7a4ebc8..b890752 100644 --- a/arch/sh/mm/Kconfig +++ b/arch/sh/mm/Kconfig @@ -169,6 +169,10 @@ config ARCH_MEMORY_PROBE def_bool y depends on MEMORY_HOTPLUG +config IOREMAP_FIXED + def_bool y + depends on X2TLB || SUPERH64 + choice prompt "Kernel page size" default PAGE_SIZE_4KB diff --git a/arch/sh/mm/Makefile b/arch/sh/mm/Makefile index edde8bd..89ba56c 100644 --- a/arch/sh/mm/Makefile +++ b/arch/sh/mm/Makefile @@ -35,6 +35,7 @@ endif obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o obj-$(CONFIG_PMB) += pmb.o obj-$(CONFIG_NUMA) += numa.o +obj-$(CONFIG_IOREMAP_FIXED) += ioremap_fixed.o # Special flags for fault_64.o. This puts restrictions on the number of # caller-save registers that the compiler can target when building this file. diff --git a/arch/sh/mm/ioremap_fixed.c b/arch/sh/mm/ioremap_fixed.c new file mode 100644 index 0000000..3a9d3d8 --- /dev/null +++ b/arch/sh/mm/ioremap_fixed.c @@ -0,0 +1,144 @@ +/* + * Re-map IO memory to kernel address space so that we can access it. + * + * These functions should only be used when it is necessary to map a + * physical address space into the kernel address space before ioremap() + * can be used, e.g. early in boot before paging_init(). + * + * Copyright (C) 2009 Matt Fleming + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +struct ioremap_map { + void __iomem *addr; + unsigned long size; + unsigned long fixmap_addr; +}; + +static struct ioremap_map ioremap_maps[FIX_N_IOREMAPS]; + +void __init ioremap_fixed_init(void) +{ + struct ioremap_map *map; + int i; + + for (i = 0; i < FIX_N_IOREMAPS; i++) { + map = &ioremap_maps[i]; + map->fixmap_addr = __fix_to_virt(FIX_IOREMAP_BEGIN + i); + } +} + +void __init __iomem * +ioremap_fixed(resource_size_t phys_addr, unsigned long size, pgprot_t prot) +{ + enum fixed_addresses idx0, idx; + resource_size_t last_addr; + struct ioremap_map *map; + unsigned long offset; + unsigned int nrpages; + int i, slot; + + slot = -1; + for (i = 0; i < FIX_N_IOREMAPS; i++) { + map = &ioremap_maps[i]; + if (!map->addr) { + map->size = size; + slot = i; + break; + } + } + + if (slot < 0) + return NULL; + + /* Don't allow wraparound or zero size */ + last_addr = phys_addr + size - 1; + if (!size || last_addr < phys_addr) + return NULL; + + /* + * Fixmap mappings have to be page-aligned + */ + offset = phys_addr & ~PAGE_MASK; + phys_addr &= PAGE_MASK; + size = PAGE_ALIGN(last_addr + 1) - phys_addr; + + /* + * Mappings have to fit in the FIX_IOREMAP area. + */ + nrpages = size >> PAGE_SHIFT; + if (nrpages > FIX_N_IOREMAPS) + return NULL; + + /* + * Ok, go for it.. + */ + idx0 = FIX_IOREMAP_BEGIN + slot; + idx = idx0; + while (nrpages > 0) { + pgprot_val(prot) |= _PAGE_WIRED; + __set_fixmap(idx, phys_addr, prot); + phys_addr += PAGE_SIZE; + idx++; + --nrpages; + } + + map->addr = (void __iomem *)(offset + map->fixmap_addr); + return map->addr; +} + +void __init iounmap_fixed(void __iomem *addr) +{ + enum fixed_addresses idx; + unsigned long virt_addr; + struct ioremap_map *map; + unsigned long offset; + unsigned int nrpages; + int i, slot; + pgprot_t prot; + + slot = -1; + for (i = 0; i < FIX_N_IOREMAPS; i++) { + map = &ioremap_maps[i]; + if (map->addr == addr) { + slot = i; + break; + } + } + + if (slot < 0) + return; + + virt_addr = (unsigned long)addr; + + offset = virt_addr & ~PAGE_MASK; + nrpages = PAGE_ALIGN(offset + map->size - 1) >> PAGE_SHIFT; + + pgprot_val(prot) = _PAGE_WIRED; + + idx = FIX_IOREMAP_BEGIN + slot + nrpages; + while (nrpages > 0) { + __clear_fixmap(idx, prot); + --idx; + --nrpages; + } + + map->size = 0; + map->addr = NULL; +} -- cgit v0.10.2 From 6f82b6ebb1dcc0c3c23820a70a5d79a60c4f5e84 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Thu, 5 Nov 2009 08:01:10 +0000 Subject: sh: Use ioremap_fixmed to map the SM501 DRAM config register We need to write to the DRAM config register very early and at such an early stage ioremap() is not available. So use ioremap_fixed() to map the register. The reason that we are avoiding using the legacy P2 mapping is that there will come a day when the legacy P2 mappings no longer exist. Signed-off-by: Matt Fleming diff --git a/arch/sh/boards/board-sh7785lcr.c b/arch/sh/boards/board-sh7785lcr.c index 7eea90d..511de38 100644 --- a/arch/sh/boards/board-sh7785lcr.c +++ b/arch/sh/boards/board-sh7785lcr.c @@ -332,8 +332,15 @@ static void __init sh7785lcr_setup(char **cmdline_p) pm_power_off = sh7785lcr_power_off; /* sm501 DRAM configuration */ - sm501_reg = (void __iomem *)0xb3e00000 + SM501_DRAM_CONTROL; - writel(0x000307c2, sm501_reg); + sm501_reg = ioremap_fixed(SM107_REG_ADDR, SM501_DRAM_CONTROL, + PAGE_KERNEL); + if (!sm501_reg) { + printk(KERN_ERR "%s: ioremap error.\n", __func__); + return; + } + + writel(0x000307c2, sm501_reg + SM501_DRAM_CONTROL); + iounmap_fixed(sm501_reg); } /* Return the board specific boot mode pin configuration */ -- cgit v0.10.2 From 597fe76ec36e782aa45b8dac5e12e4dbb0588967 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Thu, 5 Nov 2009 20:49:05 +0000 Subject: sh: Use ioremap_fixed() to implement SH-5 ioremap() Use the fixmap-based memory mapping implementation for SH-5's ioremap() functions and delete the old static allocator that was borrowed from sparc. Signed-off-by: Matt Fleming diff --git a/arch/sh/mm/ioremap_64.c b/arch/sh/mm/ioremap_64.c index ef43465..fb0aa45 100644 --- a/arch/sh/mm/ioremap_64.c +++ b/arch/sh/mm/ioremap_64.c @@ -28,299 +28,20 @@ #include #include -static struct resource shmedia_iomap = { - .name = "shmedia_iomap", - .start = IOBASE_VADDR + PAGE_SIZE, - .end = IOBASE_END - 1, -}; - -static void shmedia_mapioaddr(unsigned long pa, unsigned long va, - unsigned long flags); -static void shmedia_unmapioaddr(unsigned long vaddr); -static void __iomem *shmedia_ioremap(struct resource *res, u32 pa, - int sz, unsigned long flags); - -/* - * We have the same problem as the SPARC, so lets have the same comment: - * Our mini-allocator... - * Boy this is gross! We need it because we must map I/O for - * timers and interrupt controller before the kmalloc is available. - */ - -#define XNMLN 15 -#define XNRES 10 - -struct xresource { - struct resource xres; /* Must be first */ - int xflag; /* 1 == used */ - char xname[XNMLN+1]; -}; - -static struct xresource xresv[XNRES]; - -static struct xresource *xres_alloc(void) -{ - struct xresource *xrp; - int n; - - xrp = xresv; - for (n = 0; n < XNRES; n++) { - if (xrp->xflag == 0) { - xrp->xflag = 1; - return xrp; - } - xrp++; - } - return NULL; -} - -static void xres_free(struct xresource *xrp) -{ - xrp->xflag = 0; -} - -static struct resource *shmedia_find_resource(struct resource *root, - unsigned long vaddr) -{ - struct resource *res; - - for (res = root->child; res; res = res->sibling) - if (res->start <= vaddr && res->end >= vaddr) - return res; - - return NULL; -} - -static void __iomem *shmedia_alloc_io(unsigned long phys, unsigned long size, - const char *name, unsigned long flags) -{ - struct xresource *xres; - struct resource *res; - char *tack; - int tlen; - - if (name == NULL) - name = "???"; - - xres = xres_alloc(); - if (xres != 0) { - tack = xres->xname; - res = &xres->xres; - } else { - printk_once(KERN_NOTICE "%s: done with statics, " - "switching to kmalloc\n", __func__); - tlen = strlen(name); - tack = kmalloc(sizeof(struct resource) + tlen + 1, GFP_KERNEL); - if (!tack) - return NULL; - memset(tack, 0, sizeof(struct resource)); - res = (struct resource *) tack; - tack += sizeof(struct resource); - } - - strncpy(tack, name, XNMLN); - tack[XNMLN] = 0; - res->name = tack; - - return shmedia_ioremap(res, phys, size, flags); -} - -static void __iomem *shmedia_ioremap(struct resource *res, u32 pa, int sz, - unsigned long flags) -{ - unsigned long offset = ((unsigned long) pa) & (~PAGE_MASK); - unsigned long round_sz = (offset + sz + PAGE_SIZE-1) & PAGE_MASK; - unsigned long va; - unsigned int psz; - - if (allocate_resource(&shmedia_iomap, res, round_sz, - shmedia_iomap.start, shmedia_iomap.end, - PAGE_SIZE, NULL, NULL) != 0) { - panic("alloc_io_res(%s): cannot occupy\n", - (res->name != NULL) ? res->name : "???"); - } - - va = res->start; - pa &= PAGE_MASK; - - psz = (res->end - res->start + (PAGE_SIZE - 1)) / PAGE_SIZE; - - for (psz = res->end - res->start + 1; psz != 0; psz -= PAGE_SIZE) { - shmedia_mapioaddr(pa, va, flags); - va += PAGE_SIZE; - pa += PAGE_SIZE; - } - - return (void __iomem *)(unsigned long)(res->start + offset); -} - -static void shmedia_free_io(struct resource *res) -{ - unsigned long len = res->end - res->start + 1; - - BUG_ON((len & (PAGE_SIZE - 1)) != 0); - - while (len) { - len -= PAGE_SIZE; - shmedia_unmapioaddr(res->start + len); - } - - release_resource(res); -} - -static __init_refok void *sh64_get_page(void) -{ - void *page; - - if (slab_is_available()) - page = (void *)get_zeroed_page(GFP_KERNEL); - else - page = alloc_bootmem_pages(PAGE_SIZE); - - if (!page || ((unsigned long)page & ~PAGE_MASK)) - panic("sh64_get_page: Out of memory already?\n"); - - return page; -} - -static void shmedia_mapioaddr(unsigned long pa, unsigned long va, - unsigned long flags) +void __iomem *__ioremap_caller(unsigned long offset, unsigned long size, + unsigned long flags, void *caller) { - pgd_t *pgdp; - pud_t *pudp; - pmd_t *pmdp; - pte_t *ptep, pte; pgprot_t prot; - pr_debug("shmedia_mapiopage pa %08lx va %08lx\n", pa, va); - - if (!flags) - flags = 1; /* 1 = CB0-1 device */ - - pgdp = pgd_offset_k(va); - if (pgd_none(*pgdp) || !pgd_present(*pgdp)) { - pudp = (pud_t *)sh64_get_page(); - set_pgd(pgdp, __pgd((unsigned long)pudp | _KERNPG_TABLE)); - } - - pudp = pud_offset(pgdp, va); - if (pud_none(*pudp) || !pud_present(*pudp)) { - pmdp = (pmd_t *)sh64_get_page(); - set_pud(pudp, __pud((unsigned long)pmdp | _KERNPG_TABLE)); - } - - pmdp = pmd_offset(pudp, va); - if (pmd_none(*pmdp) || !pmd_present(*pmdp)) { - ptep = (pte_t *)sh64_get_page(); - set_pmd(pmdp, __pmd((unsigned long)ptep + _PAGE_TABLE)); - } - prot = __pgprot(_PAGE_PRESENT | _PAGE_READ | _PAGE_WRITE | _PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_SHARED | flags); - pte = pfn_pte(pa >> PAGE_SHIFT, prot); - ptep = pte_offset_kernel(pmdp, va); - - if (!pte_none(*ptep) && - pte_val(*ptep) != pte_val(pte)) - pte_ERROR(*ptep); - - set_pte(ptep, pte); - - flush_tlb_kernel_range(va, PAGE_SIZE); -} - -static void shmedia_unmapioaddr(unsigned long vaddr) -{ - pgd_t *pgdp; - pud_t *pudp; - pmd_t *pmdp; - pte_t *ptep; - - pgdp = pgd_offset_k(vaddr); - if (pgd_none(*pgdp) || pgd_bad(*pgdp)) - return; - - pudp = pud_offset(pgdp, vaddr); - if (pud_none(*pudp) || pud_bad(*pudp)) - return; - - pmdp = pmd_offset(pudp, vaddr); - if (pmd_none(*pmdp) || pmd_bad(*pmdp)) - return; - - ptep = pte_offset_kernel(pmdp, vaddr); - - if (pte_none(*ptep) || !pte_present(*ptep)) - return; - - clear_page((void *)ptep); - pte_clear(&init_mm, vaddr, ptep); -} - -void __iomem *__ioremap_caller(unsigned long offset, unsigned long size, - unsigned long flags, void *caller) -{ - char name[14]; - - sprintf(name, "phys_%08x", (u32)offset); - return shmedia_alloc_io(offset, size, name, flags); + return ioremap_fixed(offset, size, prot); } EXPORT_SYMBOL(__ioremap_caller); void __iounmap(void __iomem *virtual) { - unsigned long vaddr = (unsigned long)virtual & PAGE_MASK; - struct resource *res; - unsigned int psz; - - res = shmedia_find_resource(&shmedia_iomap, vaddr); - if (!res) { - printk(KERN_ERR "%s: Failed to free 0x%08lx\n", - __func__, vaddr); - return; - } - - psz = (res->end - res->start + (PAGE_SIZE - 1)) / PAGE_SIZE; - - shmedia_free_io(res); - - if ((char *)res >= (char *)xresv && - (char *)res < (char *)&xresv[XNRES]) { - xres_free((struct xresource *)res); - } else { - kfree(res); - } + iounmap_fixed(virtual); } EXPORT_SYMBOL(__iounmap); - -static int -ioremap_proc_info(char *buf, char **start, off_t fpos, int length, int *eof, - void *data) -{ - char *p = buf, *e = buf + length; - struct resource *r; - const char *nm; - - for (r = ((struct resource *)data)->child; r != NULL; r = r->sibling) { - if (p + 32 >= e) /* Better than nothing */ - break; - nm = r->name; - if (nm == NULL) - nm = "???"; - - p += sprintf(p, "%08lx-%08lx: %s\n", - (unsigned long)r->start, - (unsigned long)r->end, nm); - } - - return p-buf; -} - -static int __init register_proc_onchip(void) -{ - create_proc_read_entry("io_map", 0, 0, ioremap_proc_info, - &shmedia_iomap); - return 0; -} -late_initcall(register_proc_onchip); -- cgit v0.10.2 From 78bf04fc96f509474c6b443b515d6b79bb7bf584 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Sun, 17 Jan 2010 01:45:26 +0900 Subject: sh: Tidy up non-translatable checks in iounmap path. This tidies up the iounmap path with consolidated checks for nontranslatable mappings. This is in preparation of unifying the implementations. Signed-off-by: Paul Mundt diff --git a/arch/sh/mm/ioremap_32.c b/arch/sh/mm/ioremap_32.c index 2141bef..c80a816 100644 --- a/arch/sh/mm/ioremap_32.c +++ b/arch/sh/mm/ioremap_32.c @@ -105,15 +105,35 @@ void __iomem *__ioremap_caller(unsigned long phys_addr, unsigned long size, } EXPORT_SYMBOL(__ioremap_caller); +/* + * Simple checks for non-translatable mappings. + */ +static inline int iomapping_nontranslatable(unsigned long offset) +{ +#ifdef CONFIG_29BIT + /* + * In 29-bit mode this includes the fixed P1/P2 areas, as well as + * parts of P3. + */ + if (PXSEG(offset) < P3SEG || offset >= P3_ADDR_MAX) + return 1; +#endif + + if (is_pci_memory_fixed_range(offset, 0)) + return 1; + + return 0; +} + void __iounmap(void __iomem *addr) { unsigned long vaddr = (unsigned long __force)addr; - unsigned long seg = PXSEG(vaddr); struct vm_struct *p; - if (seg < P3SEG || vaddr >= P3_ADDR_MAX) - return; - if (is_pci_memory_fixed_range(vaddr, 0)) + /* + * Nothing to do if there is no translatable mapping. + */ + if (iomapping_nontranslatable(vaddr)) return; #ifdef CONFIG_PMB -- cgit v0.10.2 From 69e3f52d1b1a3ed4390bb8a09bb1324265af7fbf Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 16 Jan 2010 14:21:15 +0100 Subject: perf: Fix implicit declaration of getline in util.c getline() is considered as undeclared in util/util.c because it includes string.h, that in turn includes stdio.h, without having defined _GNU_SOURCE. But util.c also includes util.h that handles the _GNU_SOURCE and all the needed inclusions already. Let's include only util.h and sys/mman.h which is the only one header not handled by util.h This fixes the following build error: util/util.c: In function 'slow_copyfile': util/util.c:49: erreur: implicit declaration of function 'getline' util/util.c:49: erreur: nested extern declaration of 'getline' Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras LKML-Reference: <1263648075-3858-1-git-send-regression-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/util.c b/tools/perf/util/util.c index f068584..f9b890f 100644 --- a/tools/perf/util/util.c +++ b/tools/perf/util/util.c @@ -1,10 +1,5 @@ -#include -#include -#include -#include -#include -#include #include "util.h" +#include int mkdir_p(char *path, mode_t mode) { -- cgit v0.10.2 From 0eda7385db1f30271ade830a231006938a76fb53 Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Sat, 16 Jan 2010 21:31:16 +0900 Subject: perf probe: Fix build error of builtin-probe.c I got this build error when building tip tree: | cc1: warnings being treated as errors | builtin-probe.c:123: error: 'opt_show_lines' defined but not used This error is caused by: | #ifndef NO_LIBDWARF | OPT_CALLBACK('L', "line", NULL, | "FUNC[:RLN[+NUM|:RLN2]]|SRC:ALN[+NUM|:ALN2]", | "Show source code lines.", opt_show_lines), | #endif My environment defines NO_LIBDWARF, so gcc treated opt_show_lines() as garbage. So I moved opt_show_lines() into #ifndef NO_LIBDWARF ... #endif block. Signed-off-by: Hitoshi Mitake Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker Cc: Masami Hiramatsu Cc: Mike Galbraith LKML-Reference: <1263645076-9993-1-git-send-email-mitake@dcl.info.waseda.ac.jp> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index 1d3a99e..34f2acb 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -118,15 +118,6 @@ static int opt_del_probe_event(const struct option *opt __used, return 0; } -static int opt_show_lines(const struct option *opt __used, - const char *str, int unset __used) -{ - if (str) - parse_line_range_desc(str, &session.line_range); - INIT_LIST_HEAD(&session.line_range.line_list); - session.show_lines = true; - return 0; -} /* Currently just checking function name from symbol map */ static void evaluate_probe_point(struct probe_point *pp) { @@ -148,6 +139,16 @@ static int open_vmlinux(void) pr_debug("Try to open %s\n", session.kmap->dso->long_name); return open(session.kmap->dso->long_name, O_RDONLY); } + +static int opt_show_lines(const struct option *opt __used, + const char *str, int unset __used) +{ + if (str) + parse_line_range_desc(str, &session.line_range); + INIT_LIST_HEAD(&session.line_range.line_list); + session.show_lines = true; + return 0; +} #endif static const char * const probe_usage[] = { -- cgit v0.10.2 From 24a53652e361321b09df5040711e69387344ce09 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 14 Jan 2010 08:23:15 +0100 Subject: tracing: Drop the tr check from the graph tracing path Each time we save a function entry from the function graph tracer, we check if the trace array is set, which is wasteful because it is set anyway before we start the tracer. All we need is to ensure we have good read and write orderings. When we set the trace array, we just need to guarantee it to be visible before starting tracing. Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Cc: Lai Jiangshan Cc: Paul E. McKenney LKML-Reference: <1263453795-7496-1-git-send-regression-fweisbec@gmail.com> Signed-off-by: Ingo Molnar diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index b1342c5..f225229 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -212,9 +212,6 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) int cpu; int pc; - if (unlikely(!tr)) - return 0; - if (!ftrace_trace_task(current)) return 0; @@ -287,11 +284,20 @@ void trace_graph_return(struct ftrace_graph_ret *trace) local_irq_restore(flags); } +void set_graph_array(struct trace_array *tr) +{ + graph_array = tr; + + /* Make graph_array visible before we start tracing */ + + smp_mb(); +} + static int graph_trace_init(struct trace_array *tr) { int ret; - graph_array = tr; + set_graph_array(tr); ret = register_ftrace_graph(&trace_graph_return, &trace_graph_entry); if (ret) @@ -301,11 +307,6 @@ static int graph_trace_init(struct trace_array *tr) return 0; } -void set_graph_array(struct trace_array *tr) -{ - graph_array = tr; -} - static void graph_trace_reset(struct trace_array *tr) { tracing_stop_cmdline_record(); -- cgit v0.10.2 From 6d686f4564f3fc7c6e678852919e48ad331d276b Mon Sep 17 00:00:00 2001 From: H Hartley Sweeten Date: Wed, 13 Jan 2010 20:21:52 -0700 Subject: sched: Don't expose local functions kernel/sched: don't expose local functions The get_rr_interval_* functions are all class methods of struct sched_class. They are not exported so make them static. Signed-off-by: H Hartley Sweeten Cc: Peter Zijlstra LKML-Reference: <201001132021.53253.hartleys@visionengravers.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 42ac3c9..7177860 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2076,7 +2076,7 @@ static void moved_group_fair(struct task_struct *p, int on_rq) } #endif -unsigned int get_rr_interval_fair(struct rq *rq, struct task_struct *task) +static unsigned int get_rr_interval_fair(struct rq *rq, struct task_struct *task) { struct sched_entity *se = &task->se; unsigned int rr_interval = 0; diff --git a/kernel/sched_idletask.c b/kernel/sched_idletask.c index 5f93b57..01332bf 100644 --- a/kernel/sched_idletask.c +++ b/kernel/sched_idletask.c @@ -97,7 +97,7 @@ static void prio_changed_idle(struct rq *rq, struct task_struct *p, check_preempt_curr(rq, p, 0); } -unsigned int get_rr_interval_idle(struct rq *rq, struct task_struct *task) +static unsigned int get_rr_interval_idle(struct rq *rq, struct task_struct *task) { return 0; } diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index f48328a..072b3fc 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -1721,7 +1721,7 @@ static void set_curr_task_rt(struct rq *rq) dequeue_pushable_task(rq, p); } -unsigned int get_rr_interval_rt(struct rq *rq, struct task_struct *task) +static unsigned int get_rr_interval_rt(struct rq *rq, struct task_struct *task) { /* * Time slice is 0 for SCHED_FIFO tasks -- cgit v0.10.2 From 231e36f4d2e63dd770db80b9f5113310c2bcfcfd Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Thu, 14 Jan 2010 00:12:12 -0500 Subject: tracing/kprobe: Update kprobe tracing self test for new syntax Update kprobe tracing self test for new syntax (it supports deleting individual probes, and drops $argN support) and behavior change (new probes are disabled in default). This selftest includes the following checks: - Adding function-entry probe and return probe with arguments. - Enabling these probes. - Deleting it individually. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <20100114051211.7814.29436.stgit@localhost6.localdomain6> Signed-off-by: Ingo Molnar diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 7ac728d..d6266ca 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1507,28 +1507,67 @@ static int kprobe_trace_selftest_target(int a1, int a2, int a3, static __init int kprobe_trace_self_tests_init(void) { - int ret; + int ret, warn = 0; int (*target)(int, int, int, int, int, int); + struct trace_probe *tp; target = kprobe_trace_selftest_target; pr_info("Testing kprobe tracing: "); ret = command_trace_probe("p:testprobe kprobe_trace_selftest_target " - "$arg1 $arg2 $arg3 $arg4 $stack $stack0"); - if (WARN_ON_ONCE(ret)) - pr_warning("error enabling function entry\n"); + "$stack $stack0 +0($stack)"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on probing function entry.\n"); + warn++; + } else { + /* Enable trace point */ + tp = find_probe_event("testprobe", KPROBE_EVENT_SYSTEM); + if (WARN_ON_ONCE(tp == NULL)) { + pr_warning("error on getting new probe.\n"); + warn++; + } else + probe_event_enable(&tp->call); + } ret = command_trace_probe("r:testprobe2 kprobe_trace_selftest_target " "$retval"); - if (WARN_ON_ONCE(ret)) - pr_warning("error enabling function return\n"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on probing function return.\n"); + warn++; + } else { + /* Enable trace point */ + tp = find_probe_event("testprobe2", KPROBE_EVENT_SYSTEM); + if (WARN_ON_ONCE(tp == NULL)) { + pr_warning("error on getting new probe.\n"); + warn++; + } else + probe_event_enable(&tp->call); + } + + if (warn) + goto end; ret = target(1, 2, 3, 4, 5, 6); - cleanup_all_probes(); + ret = command_trace_probe("-:testprobe"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on deleting a probe.\n"); + warn++; + } + + ret = command_trace_probe("-:testprobe2"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on deleting a probe.\n"); + warn++; + } - pr_cont("OK\n"); +end: + cleanup_all_probes(); + if (warn) + pr_cont("NG: Some tests are failed. Please check them.\n"); + else + pr_cont("OK\n"); return 0; } -- cgit v0.10.2 From 73c89c15b959adf06366722c4be8d2eddec0a529 Mon Sep 17 00:00:00 2001 From: Tobias Brunner Date: Sun, 17 Jan 2010 21:52:11 +1100 Subject: crypto: gcm - Add RFC4543 wrapper for GCM This patch adds the RFC4543 (GMAC) wrapper for GCM similar to the existing RFC4106 wrapper. The main differences between GCM and GMAC are the contents of the AAD and that the plaintext is empty for the latter. Signed-off-by: Tobias Brunner Signed-off-by: Herbert Xu diff --git a/crypto/gcm.c b/crypto/gcm.c index c654713..2f5fbba 100644 --- a/crypto/gcm.c +++ b/crypto/gcm.c @@ -37,6 +37,19 @@ struct crypto_rfc4106_ctx { u8 nonce[4]; }; +struct crypto_rfc4543_ctx { + struct crypto_aead *child; + u8 nonce[4]; +}; + +struct crypto_rfc4543_req_ctx { + u8 auth_tag[16]; + struct scatterlist cipher[1]; + struct scatterlist payload[2]; + struct scatterlist assoc[2]; + struct aead_request subreq; +}; + struct crypto_gcm_ghash_ctx { unsigned int cryptlen; struct scatterlist *src; @@ -1047,6 +1060,272 @@ static struct crypto_template crypto_rfc4106_tmpl = { .module = THIS_MODULE, }; +static inline struct crypto_rfc4543_req_ctx *crypto_rfc4543_reqctx( + struct aead_request *req) +{ + unsigned long align = crypto_aead_alignmask(crypto_aead_reqtfm(req)); + + return (void *)PTR_ALIGN((u8 *)aead_request_ctx(req), align + 1); +} + +static int crypto_rfc4543_setkey(struct crypto_aead *parent, const u8 *key, + unsigned int keylen) +{ + struct crypto_rfc4543_ctx *ctx = crypto_aead_ctx(parent); + struct crypto_aead *child = ctx->child; + int err; + + if (keylen < 4) + return -EINVAL; + + keylen -= 4; + memcpy(ctx->nonce, key + keylen, 4); + + crypto_aead_clear_flags(child, CRYPTO_TFM_REQ_MASK); + crypto_aead_set_flags(child, crypto_aead_get_flags(parent) & + CRYPTO_TFM_REQ_MASK); + err = crypto_aead_setkey(child, key, keylen); + crypto_aead_set_flags(parent, crypto_aead_get_flags(child) & + CRYPTO_TFM_RES_MASK); + + return err; +} + +static int crypto_rfc4543_setauthsize(struct crypto_aead *parent, + unsigned int authsize) +{ + struct crypto_rfc4543_ctx *ctx = crypto_aead_ctx(parent); + + if (authsize != 16) + return -EINVAL; + + return crypto_aead_setauthsize(ctx->child, authsize); +} + +/* this is the same as crypto_authenc_chain */ +static void crypto_rfc4543_chain(struct scatterlist *head, + struct scatterlist *sg, int chain) +{ + if (chain) { + head->length += sg->length; + sg = scatterwalk_sg_next(sg); + } + + if (sg) + scatterwalk_sg_chain(head, 2, sg); + else + sg_mark_end(head); +} + +static struct aead_request *crypto_rfc4543_crypt(struct aead_request *req, + int enc) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct crypto_rfc4543_ctx *ctx = crypto_aead_ctx(aead); + struct crypto_rfc4543_req_ctx *rctx = crypto_rfc4543_reqctx(req); + struct aead_request *subreq = &rctx->subreq; + struct scatterlist *dst = req->dst; + struct scatterlist *cipher = rctx->cipher; + struct scatterlist *payload = rctx->payload; + struct scatterlist *assoc = rctx->assoc; + unsigned int authsize = crypto_aead_authsize(aead); + unsigned int assoclen = req->assoclen; + struct page *dstp; + u8 *vdst; + u8 *iv = PTR_ALIGN((u8 *)(rctx + 1) + crypto_aead_reqsize(ctx->child), + crypto_aead_alignmask(ctx->child) + 1); + + memcpy(iv, ctx->nonce, 4); + memcpy(iv + 4, req->iv, 8); + + /* construct cipher/plaintext */ + if (enc) + memset(rctx->auth_tag, 0, authsize); + else + scatterwalk_map_and_copy(rctx->auth_tag, dst, + req->cryptlen - authsize, + authsize, 0); + + sg_init_one(cipher, rctx->auth_tag, authsize); + + /* construct the aad */ + dstp = sg_page(dst); + vdst = PageHighMem(dstp) ? NULL : page_address(dstp) + dst->offset; + + sg_init_table(payload, 2); + sg_set_buf(payload, req->iv, 8); + crypto_rfc4543_chain(payload, dst, vdst == req->iv + 8); + assoclen += 8 + req->cryptlen - (enc ? 0 : authsize); + + sg_init_table(assoc, 2); + sg_set_page(assoc, sg_page(req->assoc), req->assoc->length, + req->assoc->offset); + crypto_rfc4543_chain(assoc, payload, 0); + + aead_request_set_tfm(subreq, ctx->child); + aead_request_set_callback(subreq, req->base.flags, req->base.complete, + req->base.data); + aead_request_set_crypt(subreq, cipher, cipher, enc ? 0 : authsize, iv); + aead_request_set_assoc(subreq, assoc, assoclen); + + return subreq; +} + +static int crypto_rfc4543_encrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct crypto_rfc4543_req_ctx *rctx = crypto_rfc4543_reqctx(req); + struct aead_request *subreq; + int err; + + subreq = crypto_rfc4543_crypt(req, 1); + err = crypto_aead_encrypt(subreq); + if (err) + return err; + + scatterwalk_map_and_copy(rctx->auth_tag, req->dst, req->cryptlen, + crypto_aead_authsize(aead), 1); + + return 0; +} + +static int crypto_rfc4543_decrypt(struct aead_request *req) +{ + req = crypto_rfc4543_crypt(req, 0); + + return crypto_aead_decrypt(req); +} + +static int crypto_rfc4543_init_tfm(struct crypto_tfm *tfm) +{ + struct crypto_instance *inst = (void *)tfm->__crt_alg; + struct crypto_aead_spawn *spawn = crypto_instance_ctx(inst); + struct crypto_rfc4543_ctx *ctx = crypto_tfm_ctx(tfm); + struct crypto_aead *aead; + unsigned long align; + + aead = crypto_spawn_aead(spawn); + if (IS_ERR(aead)) + return PTR_ERR(aead); + + ctx->child = aead; + + align = crypto_aead_alignmask(aead); + align &= ~(crypto_tfm_ctx_alignment() - 1); + tfm->crt_aead.reqsize = sizeof(struct crypto_rfc4543_req_ctx) + + ALIGN(crypto_aead_reqsize(aead), + crypto_tfm_ctx_alignment()) + + align + 16; + + return 0; +} + +static void crypto_rfc4543_exit_tfm(struct crypto_tfm *tfm) +{ + struct crypto_rfc4543_ctx *ctx = crypto_tfm_ctx(tfm); + + crypto_free_aead(ctx->child); +} + +static struct crypto_instance *crypto_rfc4543_alloc(struct rtattr **tb) +{ + struct crypto_attr_type *algt; + struct crypto_instance *inst; + struct crypto_aead_spawn *spawn; + struct crypto_alg *alg; + const char *ccm_name; + int err; + + algt = crypto_get_attr_type(tb); + err = PTR_ERR(algt); + if (IS_ERR(algt)) + return ERR_PTR(err); + + if ((algt->type ^ CRYPTO_ALG_TYPE_AEAD) & algt->mask) + return ERR_PTR(-EINVAL); + + ccm_name = crypto_attr_alg_name(tb[1]); + err = PTR_ERR(ccm_name); + if (IS_ERR(ccm_name)) + return ERR_PTR(err); + + inst = kzalloc(sizeof(*inst) + sizeof(*spawn), GFP_KERNEL); + if (!inst) + return ERR_PTR(-ENOMEM); + + spawn = crypto_instance_ctx(inst); + crypto_set_aead_spawn(spawn, inst); + err = crypto_grab_aead(spawn, ccm_name, 0, + crypto_requires_sync(algt->type, algt->mask)); + if (err) + goto out_free_inst; + + alg = crypto_aead_spawn_alg(spawn); + + err = -EINVAL; + + /* We only support 16-byte blocks. */ + if (alg->cra_aead.ivsize != 16) + goto out_drop_alg; + + /* Not a stream cipher? */ + if (alg->cra_blocksize != 1) + goto out_drop_alg; + + err = -ENAMETOOLONG; + if (snprintf(inst->alg.cra_name, CRYPTO_MAX_ALG_NAME, + "rfc4543(%s)", alg->cra_name) >= CRYPTO_MAX_ALG_NAME || + snprintf(inst->alg.cra_driver_name, CRYPTO_MAX_ALG_NAME, + "rfc4543(%s)", alg->cra_driver_name) >= + CRYPTO_MAX_ALG_NAME) + goto out_drop_alg; + + inst->alg.cra_flags = CRYPTO_ALG_TYPE_AEAD; + inst->alg.cra_flags |= alg->cra_flags & CRYPTO_ALG_ASYNC; + inst->alg.cra_priority = alg->cra_priority; + inst->alg.cra_blocksize = 1; + inst->alg.cra_alignmask = alg->cra_alignmask; + inst->alg.cra_type = &crypto_nivaead_type; + + inst->alg.cra_aead.ivsize = 8; + inst->alg.cra_aead.maxauthsize = 16; + + inst->alg.cra_ctxsize = sizeof(struct crypto_rfc4543_ctx); + + inst->alg.cra_init = crypto_rfc4543_init_tfm; + inst->alg.cra_exit = crypto_rfc4543_exit_tfm; + + inst->alg.cra_aead.setkey = crypto_rfc4543_setkey; + inst->alg.cra_aead.setauthsize = crypto_rfc4543_setauthsize; + inst->alg.cra_aead.encrypt = crypto_rfc4543_encrypt; + inst->alg.cra_aead.decrypt = crypto_rfc4543_decrypt; + + inst->alg.cra_aead.geniv = "seqiv"; + +out: + return inst; + +out_drop_alg: + crypto_drop_aead(spawn); +out_free_inst: + kfree(inst); + inst = ERR_PTR(err); + goto out; +} + +static void crypto_rfc4543_free(struct crypto_instance *inst) +{ + crypto_drop_spawn(crypto_instance_ctx(inst)); + kfree(inst); +} + +static struct crypto_template crypto_rfc4543_tmpl = { + .name = "rfc4543", + .alloc = crypto_rfc4543_alloc, + .free = crypto_rfc4543_free, + .module = THIS_MODULE, +}; + static int __init crypto_gcm_module_init(void) { int err; @@ -1067,8 +1346,14 @@ static int __init crypto_gcm_module_init(void) if (err) goto out_undo_gcm; + err = crypto_register_template(&crypto_rfc4543_tmpl); + if (err) + goto out_undo_rfc4106; + return 0; +out_undo_rfc4106: + crypto_unregister_template(&crypto_rfc4106_tmpl); out_undo_gcm: crypto_unregister_template(&crypto_gcm_tmpl); out_undo_base: @@ -1081,6 +1366,7 @@ out: static void __exit crypto_gcm_module_exit(void) { kfree(gcm_zeroes); + crypto_unregister_template(&crypto_rfc4543_tmpl); crypto_unregister_template(&crypto_rfc4106_tmpl); crypto_unregister_template(&crypto_gcm_tmpl); crypto_unregister_template(&crypto_gcm_base_tmpl); @@ -1094,3 +1380,4 @@ MODULE_DESCRIPTION("Galois/Counter Mode"); MODULE_AUTHOR("Mikko Herranen "); MODULE_ALIAS("gcm_base"); MODULE_ALIAS("rfc4106"); +MODULE_ALIAS("rfc4543"); diff --git a/include/linux/pfkeyv2.h b/include/linux/pfkeyv2.h index 228b0b6..0b80c80 100644 --- a/include/linux/pfkeyv2.h +++ b/include/linux/pfkeyv2.h @@ -315,6 +315,7 @@ struct sadb_x_kmaddress { #define SADB_X_EALG_AES_GCM_ICV12 19 #define SADB_X_EALG_AES_GCM_ICV16 20 #define SADB_X_EALG_CAMELLIACBC 22 +#define SADB_X_EALG_NULL_AES_GMAC 23 #define SADB_EALG_MAX 253 /* last EALG */ /* private allocations should use 249-255 (RFC2407) */ #define SADB_X_EALG_SERPENTCBC 252 /* draft-ietf-ipsec-ciph-aes-cbc-00 */ diff --git a/net/xfrm/xfrm_algo.c b/net/xfrm/xfrm_algo.c index 743c013..8b4d6e3 100644 --- a/net/xfrm/xfrm_algo.c +++ b/net/xfrm/xfrm_algo.c @@ -125,6 +125,22 @@ static struct xfrm_algo_desc aead_list[] = { .sadb_alg_maxbits = 256 } }, +{ + .name = "rfc4543(gcm(aes))", + + .uinfo = { + .aead = { + .icv_truncbits = 128, + } + }, + + .desc = { + .sadb_alg_id = SADB_X_EALG_NULL_AES_GMAC, + .sadb_alg_ivlen = 8, + .sadb_alg_minbits = 128, + .sadb_alg_maxbits = 256 + } +}, }; static struct xfrm_algo_desc aalg_list[] = { -- cgit v0.10.2 From 6c3f975a4cafaf4dcc9795385da6d42caa37ddeb Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?M=C3=A1rton=20N=C3=A9meth?= Date: Sun, 17 Jan 2010 21:54:01 +1100 Subject: crypto: Make Open Firmware device id constant MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit From: Márton Németh The match_table field of the struct of_device_id is constant in so it is worth to make the initialization data also constant. The semantic match that finds this kind of pattern is as follows: (http://coccinelle.lip6.fr/) // @r@ disable decl_init,const_decl_init; identifier I1, I2, x; @@ struct I1 { ... const struct I2 *x; ... }; @s@ identifier r.I1, y; identifier r.x, E; @@ struct I1 y = { .x = E, }; @c@ identifier r.I2; identifier s.E; @@ const struct I2 E[] = ... ; @depends on !c@ identifier r.I2; identifier s.E; @@ + const struct I2 E[] = ...; // Signed-off-by: Márton Németh Cc: Julia Lawall Cc: cocci@diku.dk Signed-off-by: Herbert Xu diff --git a/drivers/crypto/amcc/crypto4xx_core.c b/drivers/crypto/amcc/crypto4xx_core.c index 46e899a..1c3849f 100644 --- a/drivers/crypto/amcc/crypto4xx_core.c +++ b/drivers/crypto/amcc/crypto4xx_core.c @@ -1274,7 +1274,7 @@ static int __exit crypto4xx_remove(struct of_device *ofdev) return 0; } -static struct of_device_id crypto4xx_match[] = { +static const struct of_device_id crypto4xx_match[] = { { .compatible = "amcc,ppc4xx-crypto",}, { }, }; diff --git a/drivers/crypto/talitos.c b/drivers/crypto/talitos.c index c47ffe8..fd529d6 100644 --- a/drivers/crypto/talitos.c +++ b/drivers/crypto/talitos.c @@ -1958,7 +1958,7 @@ err_out: return err; } -static struct of_device_id talitos_match[] = { +static const struct of_device_id talitos_match[] = { { .compatible = "fsl,sec2.0", }, -- cgit v0.10.2 From 7d6f75eb21b84cdc5dfb09789974f02b42a89058 Mon Sep 17 00:00:00 2001 From: Max Vozeler Date: Sun, 17 Jan 2010 21:55:31 +1100 Subject: crypto: md5 - Add export support This patch adds export/import support to md5. The exported type is defined by struct md5_state. This is modeled after the equivalent change to sha1_generic. Signed-off-by: Max Vozeler Signed-off-by: Herbert Xu diff --git a/crypto/md5.c b/crypto/md5.c index 83eb529..9fda213 100644 --- a/crypto/md5.c +++ b/crypto/md5.c @@ -16,17 +16,13 @@ * */ #include +#include #include #include #include #include #include -#define MD5_DIGEST_SIZE 16 -#define MD5_HMAC_BLOCK_SIZE 64 -#define MD5_BLOCK_WORDS 16 -#define MD5_HASH_WORDS 4 - #define F1(x, y, z) (z ^ (x & (y ^ z))) #define F2(x, y, z) F1(z, x, y) #define F3(x, y, z) (x ^ y ^ z) @@ -35,12 +31,6 @@ #define MD5STEP(f, w, x, y, z, in, s) \ (w += f(x, y, z) + in, w = (w<>(32-s)) + x) -struct md5_ctx { - u32 hash[MD5_HASH_WORDS]; - u32 block[MD5_BLOCK_WORDS]; - u64 byte_count; -}; - static void md5_transform(u32 *hash, u32 const *in) { u32 a, b, c, d; @@ -141,7 +131,7 @@ static inline void cpu_to_le32_array(u32 *buf, unsigned int words) } } -static inline void md5_transform_helper(struct md5_ctx *ctx) +static inline void md5_transform_helper(struct md5_state *ctx) { le32_to_cpu_array(ctx->block, sizeof(ctx->block) / sizeof(u32)); md5_transform(ctx->hash, ctx->block); @@ -149,7 +139,7 @@ static inline void md5_transform_helper(struct md5_ctx *ctx) static int md5_init(struct shash_desc *desc) { - struct md5_ctx *mctx = shash_desc_ctx(desc); + struct md5_state *mctx = shash_desc_ctx(desc); mctx->hash[0] = 0x67452301; mctx->hash[1] = 0xefcdab89; @@ -162,7 +152,7 @@ static int md5_init(struct shash_desc *desc) static int md5_update(struct shash_desc *desc, const u8 *data, unsigned int len) { - struct md5_ctx *mctx = shash_desc_ctx(desc); + struct md5_state *mctx = shash_desc_ctx(desc); const u32 avail = sizeof(mctx->block) - (mctx->byte_count & 0x3f); mctx->byte_count += len; @@ -194,7 +184,7 @@ static int md5_update(struct shash_desc *desc, const u8 *data, unsigned int len) static int md5_final(struct shash_desc *desc, u8 *out) { - struct md5_ctx *mctx = shash_desc_ctx(desc); + struct md5_state *mctx = shash_desc_ctx(desc); const unsigned int offset = mctx->byte_count & 0x3f; char *p = (char *)mctx->block + offset; int padding = 56 - (offset + 1); @@ -220,12 +210,30 @@ static int md5_final(struct shash_desc *desc, u8 *out) return 0; } +static int md5_export(struct shash_desc *desc, void *out) +{ + struct md5_state *ctx = shash_desc_ctx(desc); + + memcpy(out, ctx, sizeof(*ctx)); + return 0; +} + +static int md5_import(struct shash_desc *desc, const void *in) +{ + struct md5_state *ctx = shash_desc_ctx(desc); + + memcpy(ctx, in, sizeof(*ctx)); + return 0; +} + static struct shash_alg alg = { .digestsize = MD5_DIGEST_SIZE, .init = md5_init, .update = md5_update, .final = md5_final, - .descsize = sizeof(struct md5_ctx), + .export = md5_export, + .import = md5_import, + .descsize = sizeof(struct md5_state), .base = { .cra_name = "md5", .cra_flags = CRYPTO_ALG_TYPE_SHASH, diff --git a/include/crypto/md5.h b/include/crypto/md5.h new file mode 100644 index 0000000..65f299b --- /dev/null +++ b/include/crypto/md5.h @@ -0,0 +1,17 @@ +#ifndef _CRYPTO_MD5_H +#define _CRYPTO_MD5_H + +#include + +#define MD5_DIGEST_SIZE 16 +#define MD5_HMAC_BLOCK_SIZE 64 +#define MD5_BLOCK_WORDS 16 +#define MD5_HASH_WORDS 4 + +struct md5_state { + u32 hash[MD5_HASH_WORDS]; + u32 block[MD5_BLOCK_WORDS]; + u64 byte_count; +}; + +#endif -- cgit v0.10.2 From 42cce92f4ddfa41e2dfe26fdcad4887943c032f2 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 17 Jan 2010 10:36:08 +0100 Subject: perf: Make __perf_event_sched_out static __perf_event_sched_out doesn't need to be globally available, make it static. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Ingo Molnar Cc: Arnaldo Carvalho de Melo diff --git a/kernel/perf_event.c b/kernel/perf_event.c index eae6ff6..c4e90b8 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1049,8 +1049,8 @@ static int perf_event_refresh(struct perf_event *event, int refresh) return 0; } -void __perf_event_sched_out(struct perf_event_context *ctx, - struct perf_cpu_context *cpuctx) +static void __perf_event_sched_out(struct perf_event_context *ctx, + struct perf_cpu_context *cpuctx) { struct perf_event *event; -- cgit v0.10.2 From 5b0311e1f2464547fc6f17a82d7ea2538c8c7a70 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 17 Jan 2010 11:59:13 +0100 Subject: perf: Allow pinned and flexible groups to be scheduled separately Tune the scheduling helpers so that we can choose to schedule either pinned and/or flexible groups from a context. And while at it, refactor a bit the naming of these helpers to make these more consistent and flexible. There is no (intended) change in scheduling behaviour in this patch. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Ingo Molnar Cc: Arnaldo Carvalho de Melo diff --git a/kernel/perf_event.c b/kernel/perf_event.c index c4e90b8..bfc4ee0 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1049,8 +1049,15 @@ static int perf_event_refresh(struct perf_event *event, int refresh) return 0; } -static void __perf_event_sched_out(struct perf_event_context *ctx, - struct perf_cpu_context *cpuctx) +enum event_type_t { + EVENT_FLEXIBLE = 0x1, + EVENT_PINNED = 0x2, + EVENT_ALL = EVENT_FLEXIBLE | EVENT_PINNED, +}; + +static void ctx_sched_out(struct perf_event_context *ctx, + struct perf_cpu_context *cpuctx, + enum event_type_t event_type) { struct perf_event *event; @@ -1061,13 +1068,18 @@ static void __perf_event_sched_out(struct perf_event_context *ctx, update_context_time(ctx); perf_disable(); - if (ctx->nr_active) { + if (!ctx->nr_active) + goto out_enable; + + if (event_type & EVENT_PINNED) list_for_each_entry(event, &ctx->pinned_groups, group_entry) group_sched_out(event, cpuctx, ctx); + if (event_type & EVENT_FLEXIBLE) list_for_each_entry(event, &ctx->flexible_groups, group_entry) group_sched_out(event, cpuctx, ctx); - } + + out_enable: perf_enable(); out: raw_spin_unlock(&ctx->lock); @@ -1229,15 +1241,13 @@ void perf_event_task_sched_out(struct task_struct *task, rcu_read_unlock(); if (do_switch) { - __perf_event_sched_out(ctx, cpuctx); + ctx_sched_out(ctx, cpuctx, EVENT_ALL); cpuctx->task_ctx = NULL; } } -/* - * Called with IRQs disabled - */ -static void __perf_event_task_sched_out(struct perf_event_context *ctx) +static void task_ctx_sched_out(struct perf_event_context *ctx, + enum event_type_t event_type) { struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); @@ -1247,39 +1257,34 @@ static void __perf_event_task_sched_out(struct perf_event_context *ctx) if (WARN_ON_ONCE(ctx != cpuctx->task_ctx)) return; - __perf_event_sched_out(ctx, cpuctx); + ctx_sched_out(ctx, cpuctx, event_type); cpuctx->task_ctx = NULL; } /* * Called with IRQs disabled */ -static void perf_event_cpu_sched_out(struct perf_cpu_context *cpuctx) +static void __perf_event_task_sched_out(struct perf_event_context *ctx) { - __perf_event_sched_out(&cpuctx->ctx, cpuctx); + task_ctx_sched_out(ctx, EVENT_ALL); +} + +/* + * Called with IRQs disabled + */ +static void cpu_ctx_sched_out(struct perf_cpu_context *cpuctx, + enum event_type_t event_type) +{ + ctx_sched_out(&cpuctx->ctx, cpuctx, event_type); } static void -__perf_event_sched_in(struct perf_event_context *ctx, - struct perf_cpu_context *cpuctx) +ctx_pinned_sched_in(struct perf_event_context *ctx, + struct perf_cpu_context *cpuctx, + int cpu) { - int cpu = smp_processor_id(); struct perf_event *event; - int can_add_hw = 1; - - raw_spin_lock(&ctx->lock); - ctx->is_active = 1; - if (likely(!ctx->nr_events)) - goto out; - - ctx->timestamp = perf_clock(); - - perf_disable(); - /* - * First go through the list and put on any pinned groups - * in order to give them the best chance of going on. - */ list_for_each_entry(event, &ctx->pinned_groups, group_entry) { if (event->state <= PERF_EVENT_STATE_OFF) continue; @@ -1298,6 +1303,15 @@ __perf_event_sched_in(struct perf_event_context *ctx, event->state = PERF_EVENT_STATE_ERROR; } } +} + +static void +ctx_flexible_sched_in(struct perf_event_context *ctx, + struct perf_cpu_context *cpuctx, + int cpu) +{ + struct perf_event *event; + int can_add_hw = 1; list_for_each_entry(event, &ctx->flexible_groups, group_entry) { /* Ignore events in OFF or ERROR state */ @@ -1314,11 +1328,53 @@ __perf_event_sched_in(struct perf_event_context *ctx, if (group_sched_in(event, cpuctx, ctx, cpu)) can_add_hw = 0; } +} + +static void +ctx_sched_in(struct perf_event_context *ctx, + struct perf_cpu_context *cpuctx, + enum event_type_t event_type) +{ + int cpu = smp_processor_id(); + + raw_spin_lock(&ctx->lock); + ctx->is_active = 1; + if (likely(!ctx->nr_events)) + goto out; + + ctx->timestamp = perf_clock(); + + perf_disable(); + + /* + * First go through the list and put on any pinned groups + * in order to give them the best chance of going on. + */ + if (event_type & EVENT_PINNED) + ctx_pinned_sched_in(ctx, cpuctx, cpu); + + /* Then walk through the lower prio flexible groups */ + if (event_type & EVENT_FLEXIBLE) + ctx_flexible_sched_in(ctx, cpuctx, cpu); + perf_enable(); out: raw_spin_unlock(&ctx->lock); } +static void task_ctx_sched_in(struct task_struct *task, + enum event_type_t event_type) +{ + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + struct perf_event_context *ctx = task->perf_event_ctxp; + + if (likely(!ctx)) + return; + if (cpuctx->task_ctx == ctx) + return; + ctx_sched_in(ctx, cpuctx, event_type); + cpuctx->task_ctx = ctx; +} /* * Called from scheduler to add the events of the current task * with interrupts disabled. @@ -1332,22 +1388,15 @@ __perf_event_sched_in(struct perf_event_context *ctx, */ void perf_event_task_sched_in(struct task_struct *task) { - struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); - struct perf_event_context *ctx = task->perf_event_ctxp; - - if (likely(!ctx)) - return; - if (cpuctx->task_ctx == ctx) - return; - __perf_event_sched_in(ctx, cpuctx); - cpuctx->task_ctx = ctx; + task_ctx_sched_in(task, EVENT_ALL); } -static void perf_event_cpu_sched_in(struct perf_cpu_context *cpuctx) +static void cpu_ctx_sched_in(struct perf_cpu_context *cpuctx, + enum event_type_t event_type) { struct perf_event_context *ctx = &cpuctx->ctx; - __perf_event_sched_in(ctx, cpuctx); + ctx_sched_in(ctx, cpuctx, event_type); } #define MAX_INTERRUPTS (~0ULL) @@ -1476,17 +1525,17 @@ void perf_event_task_tick(struct task_struct *curr) if (ctx) perf_ctx_adjust_freq(ctx); - perf_event_cpu_sched_out(cpuctx); + cpu_ctx_sched_out(cpuctx, EVENT_ALL); if (ctx) - __perf_event_task_sched_out(ctx); + task_ctx_sched_out(ctx, EVENT_ALL); rotate_ctx(&cpuctx->ctx); if (ctx) rotate_ctx(ctx); - perf_event_cpu_sched_in(cpuctx); + cpu_ctx_sched_in(cpuctx, EVENT_ALL); if (ctx) - perf_event_task_sched_in(curr); + task_ctx_sched_in(curr, EVENT_ALL); } static int event_enable_on_exec(struct perf_event *event, -- cgit v0.10.2 From 7defb0f879bbcfe29e3c6f29d685d4f29b7a0700 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 17 Jan 2010 12:15:31 +0100 Subject: perf: Don't schedule out/in pinned events on task tick We don't need to schedule in/out pinned events on task tick, now that pinned and flexible groups can be scheduled separately. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Ingo Molnar Cc: Arnaldo Carvalho de Melo diff --git a/kernel/perf_event.c b/kernel/perf_event.c index bfc4ee0..a90ae69 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1525,17 +1525,17 @@ void perf_event_task_tick(struct task_struct *curr) if (ctx) perf_ctx_adjust_freq(ctx); - cpu_ctx_sched_out(cpuctx, EVENT_ALL); + cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE); if (ctx) - task_ctx_sched_out(ctx, EVENT_ALL); + task_ctx_sched_out(ctx, EVENT_FLEXIBLE); rotate_ctx(&cpuctx->ctx); if (ctx) rotate_ctx(ctx); - cpu_ctx_sched_in(cpuctx, EVENT_ALL); + cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE); if (ctx) - task_ctx_sched_in(curr, EVENT_ALL); + task_ctx_sched_in(curr, EVENT_FLEXIBLE); } static int event_enable_on_exec(struct perf_event *event, -- cgit v0.10.2 From 329c0e012b99fa2325a0be205c052e4aba690f16 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 17 Jan 2010 12:56:05 +0100 Subject: perf: Better order flexible and pinned scheduling When a task gets scheduled in. We don't touch the cpu bound events so the priority order becomes: cpu pinned, cpu flexible, task pinned, task flexible. So schedule out cpu flexibles when a new task context gets in and correctly order the groups to schedule in: task pinned, cpu flexible, task flexible. Cpu pinned groups don't need to be touched at this time. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Ingo Molnar Cc: Arnaldo Carvalho de Melo diff --git a/kernel/perf_event.c b/kernel/perf_event.c index a90ae69..edc46b9 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1362,6 +1362,14 @@ ctx_sched_in(struct perf_event_context *ctx, raw_spin_unlock(&ctx->lock); } +static void cpu_ctx_sched_in(struct perf_cpu_context *cpuctx, + enum event_type_t event_type) +{ + struct perf_event_context *ctx = &cpuctx->ctx; + + ctx_sched_in(ctx, cpuctx, event_type); +} + static void task_ctx_sched_in(struct task_struct *task, enum event_type_t event_type) { @@ -1388,15 +1396,27 @@ static void task_ctx_sched_in(struct task_struct *task, */ void perf_event_task_sched_in(struct task_struct *task) { - task_ctx_sched_in(task, EVENT_ALL); -} + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + struct perf_event_context *ctx = task->perf_event_ctxp; -static void cpu_ctx_sched_in(struct perf_cpu_context *cpuctx, - enum event_type_t event_type) -{ - struct perf_event_context *ctx = &cpuctx->ctx; + if (likely(!ctx)) + return; - ctx_sched_in(ctx, cpuctx, event_type); + if (cpuctx->task_ctx == ctx) + return; + + /* + * We want to keep the following priority order: + * cpu pinned (that don't need to move), task pinned, + * cpu flexible, task flexible. + */ + cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE); + + ctx_sched_in(ctx, cpuctx, EVENT_PINNED); + cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE); + ctx_sched_in(ctx, cpuctx, EVENT_FLEXIBLE); + + cpuctx->task_ctx = ctx; } #define MAX_INTERRUPTS (~0ULL) -- cgit v0.10.2 From 933a838aa1aae8388438bb002fbdaf6fca519a5c Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Tue, 29 Dec 2009 18:21:39 +0100 Subject: pcmcia: make use of pcmcia_dev_resume() return value In runtime_resume(), do not throw away the return value of pcmcia_dev_resume(), for we can use it (at least) in pcmcia_store_pm_state(). This also fixes the pointless assignment previosly seen there, as noted by Dan Carpenter. CC: Dan Carpenter Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/ds.c b/drivers/pcmcia/ds.c index 1a4a3c4..defa44c 100644 --- a/drivers/pcmcia/ds.c +++ b/drivers/pcmcia/ds.c @@ -970,13 +970,14 @@ static int runtime_suspend(struct device *dev) return rc; } -static void runtime_resume(struct device *dev) +static int runtime_resume(struct device *dev) { int rc; down(&dev->sem); rc = pcmcia_dev_resume(dev); up(&dev->sem); + return rc; } /************************ per-device sysfs output ***************************/ @@ -1027,7 +1028,7 @@ static ssize_t pcmcia_store_pm_state(struct device *dev, struct device_attribute if ((!p_dev->suspended) && !strncmp(buf, "off", 3)) ret = runtime_suspend(dev); else if (p_dev->suspended && !strncmp(buf, "on", 2)) - runtime_resume(dev); + ret = runtime_resume(dev); return ret ? ret : count; } -- cgit v0.10.2 From cd2e18fe785082b132d960063df83a02acc9a4f2 Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Fri, 1 Jan 2010 17:43:08 +0100 Subject: pcmcia: remove remaining unused IRQ_FIRST_SHARED parameter Komuro pointed out correctly that I missed one IRQ_FIRST_SHARED parameter in smc91c92_cs.c, and that another line could be writter more beautifully. CC: netdev@vger.kernel.org CC: Komuro Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/net/pcmcia/smc91c92_cs.c b/drivers/net/pcmcia/smc91c92_cs.c index 6dd486d..aa57cfd 100644 --- a/drivers/net/pcmcia/smc91c92_cs.c +++ b/drivers/net/pcmcia/smc91c92_cs.c @@ -453,8 +453,7 @@ static int mhz_mfc_config(struct pcmcia_device *link) link->conf.Attributes |= CONF_ENABLE_SPKR; link->conf.Status = CCSR_AUDIO_ENA; - link->irq.Attributes = - IRQ_TYPE_DYNAMIC_SHARING; + link->irq.Attributes = IRQ_TYPE_DYNAMIC_SHARING; link->io.IOAddrLines = 16; link->io.Attributes2 = IO_DATA_PATH_WIDTH_8; link->io.NumPorts2 = 8; @@ -652,8 +651,7 @@ static int osi_config(struct pcmcia_device *link) link->conf.Attributes |= CONF_ENABLE_SPKR; link->conf.Status = CCSR_AUDIO_ENA; - link->irq.Attributes = - IRQ_TYPE_DYNAMIC_SHARING|IRQ_FIRST_SHARED; + link->irq.Attributes = IRQ_TYPE_DYNAMIC_SHARING; link->io.NumPorts1 = 64; link->io.Attributes2 = IO_DATA_PATH_WIDTH_8; link->io.NumPorts2 = 8; -- cgit v0.10.2 From fa0b3bc504ff813cc05988bb30bbb6c6a0263eb4 Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Sat, 2 Jan 2010 12:08:22 +0100 Subject: pcmcia: do not meddle with already assigned resources Do not release any iomem resources already in use. Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/rsrc_nonstatic.c b/drivers/pcmcia/rsrc_nonstatic.c index 9b0dc43..4f93889 100644 --- a/drivers/pcmcia/rsrc_nonstatic.c +++ b/drivers/pcmcia/rsrc_nonstatic.c @@ -727,13 +727,6 @@ static int adjust_memory(struct pcmcia_socket *s, unsigned int action, unsigned break; case REMOVE_MANAGED_RESOURCE: ret = sub_interval(&data->mem_db, start, size); - if (!ret) { - struct pcmcia_socket *socket; - down_read(&pcmcia_socket_list_rwsem); - list_for_each_entry(socket, &pcmcia_socket_list, socket_list) - release_cis_mem(socket); - up_read(&pcmcia_socket_list_rwsem); - } break; default: ret = -EINVAL; -- cgit v0.10.2 From 904e377744bfdcea276c27167fa6a609929f39dc Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Sat, 2 Jan 2010 12:28:04 +0100 Subject: pcmcia: validate CIS, not CIS cache. In pccard_validate_cis(), validate the card CIS, not the CIS cache. Also, destroy the CIS cache if pccard_validate_cis fails. Furthermore, do not remove the fake CIS in destroy_cis_cache() but do so explicitely in the code paths where it makes sense. Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/cistpl.c b/drivers/pcmcia/cistpl.c index 25b1cd2..41ec772 100644 --- a/drivers/pcmcia/cistpl.c +++ b/drivers/pcmcia/cistpl.c @@ -319,22 +319,23 @@ remove_cis_cache(struct pcmcia_socket *s, int attr, u_int addr, u_int len) } } +/** + * destroy_cis_cache() - destroy the CIS cache + * @s: pcmcia_socket for which CIS cache shall be destroyed + * + * This destroys the CIS cache but keeps any fake CIS alive. + */ + void destroy_cis_cache(struct pcmcia_socket *s) { struct list_head *l, *n; + struct cis_cache_entry *cis; list_for_each_safe(l, n, &s->cis_cache) { - struct cis_cache_entry *cis = list_entry(l, struct cis_cache_entry, node); - + cis = list_entry(l, struct cis_cache_entry, node); list_del(&cis->node); kfree(cis); } - - /* - * If there was a fake CIS, destroy that as well. - */ - kfree(s->fake_cis); - s->fake_cis = NULL; } EXPORT_SYMBOL(destroy_cis_cache); @@ -1596,6 +1597,9 @@ int pccard_validate_cis(struct pcmcia_socket *s, unsigned int *info) if (!s) return -EINVAL; + /* We do not want to validate the CIS cache... */ + destroy_cis_cache(s); + tuple = kmalloc(sizeof(*tuple), GFP_KERNEL); if (tuple == NULL) { dev_printk(KERN_WARNING, &s->dev, "no memory to validate CIS\n"); @@ -1647,6 +1651,10 @@ int pccard_validate_cis(struct pcmcia_socket *s, unsigned int *info) count = 0; done: + /* invalidate CIS cache on failure */ + if (!dev_ok || !ident_ok || !count) + destroy_cis_cache(s); + if (info) *info = count; kfree(tuple); diff --git a/drivers/pcmcia/cs.c b/drivers/pcmcia/cs.c index 6d6f82b..96d8d25 100644 --- a/drivers/pcmcia/cs.c +++ b/drivers/pcmcia/cs.c @@ -407,6 +407,8 @@ static void socket_shutdown(struct pcmcia_socket *s) s->irq.AssignedIRQ = s->irq.Config = 0; s->lock_count = 0; destroy_cis_cache(s); + kfree(s->fake_cis); + s->fake_cis = NULL; #ifdef CONFIG_CARDBUS cb_free(s); #endif @@ -577,6 +579,8 @@ static int socket_late_resume(struct pcmcia_socket *skt) dev_dbg(&skt->dev, "cis mismatch - different card\n"); socket_remove_drivers(skt); destroy_cis_cache(skt); + kfree(skt->fake_cis); + skt->fake_cis = NULL; /* * Workaround: give DS time to schedule removal. * Remove me once the 100ms delay is eliminated diff --git a/drivers/pcmcia/rsrc_nonstatic.c b/drivers/pcmcia/rsrc_nonstatic.c index 4f93889..b886385 100644 --- a/drivers/pcmcia/rsrc_nonstatic.c +++ b/drivers/pcmcia/rsrc_nonstatic.c @@ -281,10 +281,9 @@ static int readable(struct pcmcia_socket *s, struct resource *res, s->cis_virt = ioremap(res->start, s->map_size); if (s->cis_virt) { ret = pccard_validate_cis(s, count); - /* invalidate mapping and CIS cache */ + /* invalidate mapping */ iounmap(s->cis_virt); s->cis_virt = NULL; - destroy_cis_cache(s); } s->cis_mem.res = NULL; if ((ret != 0) || (*count == 0)) -- cgit v0.10.2 From f131ddc4bd1713385c70606555d4d63bed5ec3fd Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Sat, 2 Jan 2010 12:58:10 +0100 Subject: pcmcia: cleanup pccard_validate_cis() Cleanup pccard_validate_cis() and make it return an error code on all failures, not merely on some failures. Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/cistpl.c b/drivers/pcmcia/cistpl.c index 41ec772..04bf1ba 100644 --- a/drivers/pcmcia/cistpl.c +++ b/drivers/pcmcia/cistpl.c @@ -1577,88 +1577,95 @@ next_entry: EXPORT_SYMBOL(pccard_loop_tuple); -/*====================================================================== - - This tries to determine if a card has a sensible CIS. It returns - the number of tuples in the CIS, or 0 if the CIS looks bad. The - checks include making sure several critical tuples are present and - valid; seeing if the total number of tuples is reasonable; and - looking for tuples that use reserved codes. - -======================================================================*/ - +/** + * pccard_validate_cis() - check whether card has a sensible CIS + * @s: the struct pcmcia_socket we are to check + * @info: returns the number of tuples in the (valid) CIS, or 0 + * + * This tries to determine if a card has a sensible CIS. In @info, it + * returns the number of tuples in the CIS, or 0 if the CIS looks bad. The + * checks include making sure several critical tuples are present and + * valid; seeing if the total number of tuples is reasonable; and + * looking for tuples that use reserved codes. + * + * The function returns 0 on success. + */ int pccard_validate_cis(struct pcmcia_socket *s, unsigned int *info) { - tuple_t *tuple; - cisparse_t *p; - unsigned int count = 0; - int ret, reserved, dev_ok = 0, ident_ok = 0; + tuple_t *tuple; + cisparse_t *p; + unsigned int count = 0; + int ret, reserved, dev_ok = 0, ident_ok = 0; - if (!s) - return -EINVAL; + if (!s) + return -EINVAL; - /* We do not want to validate the CIS cache... */ - destroy_cis_cache(s); + /* We do not want to validate the CIS cache... */ + destroy_cis_cache(s); - tuple = kmalloc(sizeof(*tuple), GFP_KERNEL); - if (tuple == NULL) { - dev_printk(KERN_WARNING, &s->dev, "no memory to validate CIS\n"); - return -ENOMEM; - } - p = kmalloc(sizeof(*p), GFP_KERNEL); - if (p == NULL) { - kfree(tuple); - dev_printk(KERN_WARNING, &s->dev, "no memory to validate CIS\n"); - return -ENOMEM; - } + tuple = kmalloc(sizeof(*tuple), GFP_KERNEL); + if (tuple == NULL) { + dev_warn(&s->dev, "no memory to validate CIS\n"); + return -ENOMEM; + } + p = kmalloc(sizeof(*p), GFP_KERNEL); + if (p == NULL) { + kfree(tuple); + dev_warn(&s->dev, "no memory to validate CIS\n"); + return -ENOMEM; + } - count = reserved = 0; - tuple->DesiredTuple = RETURN_FIRST_TUPLE; - tuple->Attributes = TUPLE_RETURN_COMMON; - ret = pccard_get_first_tuple(s, BIND_FN_ALL, tuple); - if (ret != 0) - goto done; - - /* First tuple should be DEVICE; we should really have either that - or a CFTABLE_ENTRY of some sort */ - if ((tuple->TupleCode == CISTPL_DEVICE) || - (pccard_read_tuple(s, BIND_FN_ALL, CISTPL_CFTABLE_ENTRY, p) == 0) || - (pccard_read_tuple(s, BIND_FN_ALL, CISTPL_CFTABLE_ENTRY_CB, p) == 0)) - dev_ok++; - - /* All cards should have a MANFID tuple, and/or a VERS_1 or VERS_2 - tuple, for card identification. Certain old D-Link and Linksys - cards have only a broken VERS_2 tuple; hence the bogus test. */ - if ((pccard_read_tuple(s, BIND_FN_ALL, CISTPL_MANFID, p) == 0) || - (pccard_read_tuple(s, BIND_FN_ALL, CISTPL_VERS_1, p) == 0) || - (pccard_read_tuple(s, BIND_FN_ALL, CISTPL_VERS_2, p) != -ENOSPC)) - ident_ok++; - - if (!dev_ok && !ident_ok) - goto done; - - for (count = 1; count < MAX_TUPLES; count++) { - ret = pccard_get_next_tuple(s, BIND_FN_ALL, tuple); + count = reserved = 0; + tuple->DesiredTuple = RETURN_FIRST_TUPLE; + tuple->Attributes = TUPLE_RETURN_COMMON; + ret = pccard_get_first_tuple(s, BIND_FN_ALL, tuple); if (ret != 0) - break; - if (((tuple->TupleCode > 0x23) && (tuple->TupleCode < 0x40)) || - ((tuple->TupleCode > 0x47) && (tuple->TupleCode < 0x80)) || - ((tuple->TupleCode > 0x90) && (tuple->TupleCode < 0xff))) - reserved++; - } - if ((count == MAX_TUPLES) || (reserved > 5) || - ((!dev_ok || !ident_ok) && (count > 10))) - count = 0; + goto done; + + /* First tuple should be DEVICE; we should really have either that + or a CFTABLE_ENTRY of some sort */ + if ((tuple->TupleCode == CISTPL_DEVICE) || + (!pccard_read_tuple(s, BIND_FN_ALL, CISTPL_CFTABLE_ENTRY, p)) || + (!pccard_read_tuple(s, BIND_FN_ALL, CISTPL_CFTABLE_ENTRY_CB, p))) + dev_ok++; + + /* All cards should have a MANFID tuple, and/or a VERS_1 or VERS_2 + tuple, for card identification. Certain old D-Link and Linksys + cards have only a broken VERS_2 tuple; hence the bogus test. */ + if ((pccard_read_tuple(s, BIND_FN_ALL, CISTPL_MANFID, p) == 0) || + (pccard_read_tuple(s, BIND_FN_ALL, CISTPL_VERS_1, p) == 0) || + (pccard_read_tuple(s, BIND_FN_ALL, CISTPL_VERS_2, p) != -ENOSPC)) + ident_ok++; + + if (!dev_ok && !ident_ok) + goto done; + + for (count = 1; count < MAX_TUPLES; count++) { + ret = pccard_get_next_tuple(s, BIND_FN_ALL, tuple); + if (ret != 0) + break; + if (((tuple->TupleCode > 0x23) && (tuple->TupleCode < 0x40)) || + ((tuple->TupleCode > 0x47) && (tuple->TupleCode < 0x80)) || + ((tuple->TupleCode > 0x90) && (tuple->TupleCode < 0xff))) + reserved++; + } + if ((count == MAX_TUPLES) || (reserved > 5) || + ((!dev_ok || !ident_ok) && (count > 10))) + count = 0; + + ret = 0; done: - /* invalidate CIS cache on failure */ - if (!dev_ok || !ident_ok || !count) - destroy_cis_cache(s); - - if (info) - *info = count; - kfree(tuple); - kfree(p); - return 0; + /* invalidate CIS cache on failure */ + if (!dev_ok || !ident_ok || !count) { + destroy_cis_cache(s); + ret = -EIO; + } + + if (info) + *info = count; + kfree(tuple); + kfree(p); + return ret; } EXPORT_SYMBOL(pccard_validate_cis); -- cgit v0.10.2 From 88b060d6c03fcb9e4d2018b4349954c4242a5c7f Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Sat, 2 Jan 2010 14:14:23 +0100 Subject: pcmcia: improve check for same card in slot after resume During a suspend/resume cycle, an user may change the card in the PCMCIA/CardBus slot. The pcmcia_core can at least look at the socket state to check whether it is the same. For PCMCIA devices, move the detection and handling of such a change to ds.c. For CardBus devices, the PCI hotplug interface doesn't offer a "rescan" facility which also _removes_ devices no longer to be found behind a bridge. Therefore, remove and re-add all devices unconditionally. CC: Jesse Barnes CC: Linus Torvalds Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/cistpl.c b/drivers/pcmcia/cistpl.c index 04bf1ba..a8323cb 100644 --- a/drivers/pcmcia/cistpl.c +++ b/drivers/pcmcia/cistpl.c @@ -377,6 +377,7 @@ int verify_cis_cache(struct pcmcia_socket *s) kfree(buf); return 0; } +EXPORT_SYMBOL(verify_cis_cache); /*====================================================================== diff --git a/drivers/pcmcia/cs.c b/drivers/pcmcia/cs.c index 96d8d25..8c51493 100644 --- a/drivers/pcmcia/cs.c +++ b/drivers/pcmcia/cs.c @@ -328,7 +328,7 @@ static int send_event(struct pcmcia_socket *s, event_t event, int priority) { int ret; - if (s->state & SOCKET_CARDBUS) + if ((s->state & SOCKET_CARDBUS) && (event != CS_EVENT_CARD_REMOVAL)) return 0; dev_dbg(&s->dev, "send_event(event %d, pri %d, callback 0x%p)\n", @@ -346,13 +346,6 @@ static int send_event(struct pcmcia_socket *s, event_t event, int priority) return ret; } -static void socket_remove_drivers(struct pcmcia_socket *skt) -{ - dev_dbg(&skt->dev, "remove_drivers\n"); - - send_event(skt, CS_EVENT_CARD_REMOVAL, CS_EVENT_PRI_HIGH); -} - static int socket_reset(struct pcmcia_socket *skt) { int status, i; @@ -395,7 +388,7 @@ static void socket_shutdown(struct pcmcia_socket *s) dev_dbg(&s->dev, "shutdown\n"); - socket_remove_drivers(s); + send_event(s, CS_EVENT_CARD_REMOVAL, CS_EVENT_PRI_HIGH); s->state &= SOCKET_INUSE | SOCKET_PRESENT; msleep(shutdown_delay * 10); s->state &= SOCKET_INUSE; @@ -462,7 +455,8 @@ static int socket_setup(struct pcmcia_socket *skt, int initial_delay) return -EINVAL; } skt->state |= SOCKET_CARDBUS; - } + } else + skt->state &= ~SOCKET_CARDBUS; /* * Decode the card voltage requirements, and apply power to the card. @@ -544,6 +538,8 @@ static int socket_suspend(struct pcmcia_socket *skt) if (skt->state & SOCKET_SUSPEND) return -EBUSY; + skt->suspended_state = skt->state; + send_event(skt, CS_EVENT_PM_SUSPEND, CS_EVENT_PRI_LOW); skt->socket = dead_socket; skt->ops->set_socket(skt, &skt->socket); @@ -566,38 +562,37 @@ static int socket_early_resume(struct pcmcia_socket *skt) static int socket_late_resume(struct pcmcia_socket *skt) { - if (!(skt->state & SOCKET_PRESENT)) { - skt->state &= ~SOCKET_SUSPEND; + skt->state &= ~SOCKET_SUSPEND; + + if (!(skt->state & SOCKET_PRESENT)) return socket_insert(skt); + + if (skt->resume_status) { + socket_shutdown(skt); + return 0; } - if (skt->resume_status == 0) { - /* - * FIXME: need a better check here for cardbus cards. - */ - if (verify_cis_cache(skt) != 0) { - dev_dbg(&skt->dev, "cis mismatch - different card\n"); - socket_remove_drivers(skt); - destroy_cis_cache(skt); - kfree(skt->fake_cis); - skt->fake_cis = NULL; - /* - * Workaround: give DS time to schedule removal. - * Remove me once the 100ms delay is eliminated - * in ds.c - */ - msleep(200); - send_event(skt, CS_EVENT_CARD_INSERTION, CS_EVENT_PRI_LOW); - } else { - dev_dbg(&skt->dev, "cis matches cache\n"); - send_event(skt, CS_EVENT_PM_RESUME, CS_EVENT_PRI_LOW); - } - } else { + if (skt->suspended_state != skt->state) { + dev_dbg(&skt->dev, + "suspend state 0x%x != resume state 0x%x\n", + skt->suspended_state, skt->state); + socket_shutdown(skt); + return socket_insert(skt); } - skt->state &= ~SOCKET_SUSPEND; +#ifdef CONFIG_CARDBUS + if (skt->state & SOCKET_CARDBUS) { + /* We can't be sure the CardBus card is the same + * as the one previously inserted. Therefore, remove + * and re-add... */ + cb_free(skt); + cb_alloc(skt); + return 0; + } +#endif + send_event(skt, CS_EVENT_PM_RESUME, CS_EVENT_PRI_LOW); return 0; } diff --git a/drivers/pcmcia/ds.c b/drivers/pcmcia/ds.c index defa44c..87e0639 100644 --- a/drivers/pcmcia/ds.c +++ b/drivers/pcmcia/ds.c @@ -1252,8 +1252,22 @@ static int ds_event(struct pcmcia_socket *skt, event_t event, int priority) case CS_EVENT_EJECTION_REQUEST: break; - case CS_EVENT_PM_SUSPEND: case CS_EVENT_PM_RESUME: + if (verify_cis_cache(skt) != 0) { + dev_dbg(&skt->dev, "cis mismatch - different card\n"); + /* first, remove the card */ + ds_event(skt, CS_EVENT_CARD_REMOVAL, CS_EVENT_PRI_HIGH); + destroy_cis_cache(skt); + kfree(skt->fake_cis); + skt->fake_cis = NULL; + /* now, add the new card */ + ds_event(skt, CS_EVENT_CARD_INSERTION, + CS_EVENT_PRI_LOW); + } + handle_event(skt, event); + break; + + case CS_EVENT_PM_SUSPEND: case CS_EVENT_RESET_PHYSICAL: case CS_EVENT_CARD_RESET: default: diff --git a/include/pcmcia/ss.h b/include/pcmcia/ss.h index cbfba88..b0ebd11 100644 --- a/include/pcmcia/ss.h +++ b/include/pcmcia/ss.h @@ -137,6 +137,7 @@ struct pcmcia_socket { spinlock_t lock; socket_state_t socket; u_int state; + u_int suspended_state; /* state before suspend */ u_short functions; u_short lock_count; pccard_mem_map cis_mem; -- cgit v0.10.2 From 57197b9b7712eb19f5344c466e8aefbac1adbe55 Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Sat, 2 Jan 2010 17:27:33 +0100 Subject: pcmcia: CardBus doesn't need CIS access At least no in-kernel CardBus-capable PCI driver makes use of the CIS access functions. Therefore, it seems sensible to remove this unused code, and cleanup cardbus.c a lot. CC: Jesse Barnes CC: Linus Torvalds Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/cardbus.c b/drivers/pcmcia/cardbus.c index d99f846..ac0686e 100644 --- a/drivers/pcmcia/cardbus.c +++ b/drivers/pcmcia/cardbus.c @@ -20,170 +20,12 @@ */ -#include #include -#include -#include -#include +#include #include -#include -#include -#include -#include #include -#include -#include -#include "cs_internal.h" - -/*====================================================================*/ - -/* Offsets in the Expansion ROM Image Header */ -#define ROM_SIGNATURE 0x0000 /* 2 bytes */ -#define ROM_DATA_PTR 0x0018 /* 2 bytes */ - -/* Offsets in the CardBus PC Card Data Structure */ -#define PCDATA_SIGNATURE 0x0000 /* 4 bytes */ -#define PCDATA_VPD_PTR 0x0008 /* 2 bytes */ -#define PCDATA_LENGTH 0x000a /* 2 bytes */ -#define PCDATA_REVISION 0x000c -#define PCDATA_IMAGE_SZ 0x0010 /* 2 bytes */ -#define PCDATA_ROM_LEVEL 0x0012 /* 2 bytes */ -#define PCDATA_CODE_TYPE 0x0014 -#define PCDATA_INDICATOR 0x0015 - -/*===================================================================== - - Expansion ROM's have a special layout, and pointers specify an - image number and an offset within that image. xlate_rom_addr() - converts an image/offset address to an absolute offset from the - ROM's base address. - -=====================================================================*/ - -static u_int xlate_rom_addr(void __iomem *b, u_int addr) -{ - u_int img = 0, ofs = 0, sz; - u_short data; - while ((readb(b) == 0x55) && (readb(b + 1) == 0xaa)) { - if (img == (addr >> 28)) - return (addr & 0x0fffffff) + ofs; - data = readb(b + ROM_DATA_PTR) + (readb(b + ROM_DATA_PTR + 1) << 8); - sz = 512 * (readb(b + data + PCDATA_IMAGE_SZ) + - (readb(b + data + PCDATA_IMAGE_SZ + 1) << 8)); - if ((sz == 0) || (readb(b + data + PCDATA_INDICATOR) & 0x80)) - break; - b += sz; - ofs += sz; - img++; - } - return 0; -} - -/*===================================================================== - - These are similar to setup_cis_mem and release_cis_mem for 16-bit - cards. The "result" that is used externally is the cb_cis_virt - pointer in the struct pcmcia_socket structure. - -=====================================================================*/ - -static void cb_release_cis_mem(struct pcmcia_socket *s) -{ - if (s->cb_cis_virt) { - dev_dbg(&s->dev, "cb_release_cis_mem()\n"); - iounmap(s->cb_cis_virt); - s->cb_cis_virt = NULL; - s->cb_cis_res = NULL; - } -} - -static int cb_setup_cis_mem(struct pcmcia_socket *s, struct resource *res) -{ - unsigned int start, size; - - if (res == s->cb_cis_res) - return 0; - - if (s->cb_cis_res) - cb_release_cis_mem(s); - - start = res->start; - size = res->end - start + 1; - s->cb_cis_virt = ioremap(start, size); - - if (!s->cb_cis_virt) - return -1; - - s->cb_cis_res = res; - - return 0; -} - -/*===================================================================== - - This is used by the CIS processing code to read CIS information - from a CardBus device. - -=====================================================================*/ - -int read_cb_mem(struct pcmcia_socket *s, int space, u_int addr, u_int len, - void *ptr) -{ - struct pci_dev *dev; - struct resource *res; - - dev_dbg(&s->dev, "read_cb_mem(%d, %#x, %u)\n", space, addr, len); - dev = pci_get_slot(s->cb_dev->subordinate, 0); - if (!dev) - goto fail; - - /* Config space? */ - if (space == 0) { - if (addr + len > 0x100) - goto failput; - for (; len; addr++, ptr++, len--) - pci_read_config_byte(dev, addr, ptr); - return 0; - } - - res = dev->resource + space - 1; - - pci_dev_put(dev); - - if (!res->flags) - goto fail; - - if (cb_setup_cis_mem(s, res) != 0) - goto fail; - - if (space == 7) { - addr = xlate_rom_addr(s->cb_cis_virt, addr); - if (addr == 0) - goto fail; - } - - if (addr + len > res->end - res->start) - goto fail; - - memcpy_fromio(ptr, s->cb_cis_virt + addr, len); - return 0; - -failput: - pci_dev_put(dev); -fail: - memset(ptr, 0xff, len); - return -1; -} - -/*===================================================================== - - cb_alloc() and cb_free() allocate and free the kernel data - structures for a Cardbus device, and handle the lowest level PCI - device setup issues. - -=====================================================================*/ static void cardbus_config_irq_and_cls(struct pci_bus *bus, int irq) { @@ -215,6 +57,13 @@ static void cardbus_config_irq_and_cls(struct pci_bus *bus, int irq) } } +/** + * cb_alloc() - add CardBus device + * @s: the pcmcia_socket where the CardBus device is located + * + * cb_alloc() allocates the kernel data structures for a Cardbus device + * and handles the lowest level PCI device setup issues. + */ int __ref cb_alloc(struct pcmcia_socket *s) { struct pci_bus *bus = s->cb_dev->subordinate; @@ -249,12 +98,16 @@ int __ref cb_alloc(struct pcmcia_socket *s) return 0; } +/** + * cb_free() - remove CardBus device + * @s: the pcmcia_socket where the CardBus device was located + * + * cb_free() handles the lowest level PCI device cleanup. + */ void cb_free(struct pcmcia_socket *s) { struct pci_dev *bridge = s->cb_dev; - cb_release_cis_mem(s); - if (bridge) pci_remove_behind_bridge(bridge); } diff --git a/drivers/pcmcia/cistpl.c b/drivers/pcmcia/cistpl.c index a8323cb..368367c 100644 --- a/drivers/pcmcia/cistpl.c +++ b/drivers/pcmcia/cistpl.c @@ -268,29 +268,27 @@ EXPORT_SYMBOL(pcmcia_write_cis_mem); static void read_cis_cache(struct pcmcia_socket *s, int attr, u_int addr, size_t len, void *ptr) { - struct cis_cache_entry *cis; - int ret; + struct cis_cache_entry *cis; + int ret; - if (s->fake_cis) { - if (s->fake_cis_len >= addr+len) - memcpy(ptr, s->fake_cis+addr, len); - else - memset(ptr, 0xff, len); - return; - } + if (s->state & SOCKET_CARDBUS) + return; - list_for_each_entry(cis, &s->cis_cache, node) { - if (cis->addr == addr && cis->len == len && cis->attr == attr) { - memcpy(ptr, cis->cache, len); - return; + if (s->fake_cis) { + if (s->fake_cis_len >= addr+len) + memcpy(ptr, s->fake_cis+addr, len); + else + memset(ptr, 0xff, len); + return; + } + + list_for_each_entry(cis, &s->cis_cache, node) { + if (cis->addr == addr && cis->len == len && cis->attr == attr) { + memcpy(ptr, cis->cache, len); + return; + } } - } -#ifdef CONFIG_CARDBUS - if (s->state & SOCKET_CARDBUS) - ret = read_cb_mem(s, attr, addr, len, ptr); - else -#endif ret = pcmcia_read_cis_mem(s, attr, addr, len, ptr); if (ret == 0) { @@ -351,6 +349,9 @@ int verify_cis_cache(struct pcmcia_socket *s) struct cis_cache_entry *cis; char *buf; + if (s->state & SOCKET_CARDBUS) + return -EINVAL; + buf = kmalloc(256, GFP_KERNEL); if (buf == NULL) { dev_printk(KERN_WARNING, &s->dev, @@ -362,12 +363,8 @@ int verify_cis_cache(struct pcmcia_socket *s) if (len > 256) len = 256; -#ifdef CONFIG_CARDBUS - if (s->state & SOCKET_CARDBUS) - read_cb_mem(s, cis->attr, cis->addr, len, buf); - else -#endif - pcmcia_read_cis_mem(s, cis->attr, cis->addr, len, buf); + + pcmcia_read_cis_mem(s, cis->attr, cis->addr, len, buf); if (memcmp(buf, cis->cache, len) != 0) { kfree(buf); @@ -427,25 +424,16 @@ int pccard_get_first_tuple(struct pcmcia_socket *s, unsigned int function, tuple { if (!s) return -EINVAL; - if (!(s->state & SOCKET_PRESENT)) + + if (!(s->state & SOCKET_PRESENT) || (s->state & SOCKET_CARDBUS)) return -ENODEV; tuple->TupleLink = tuple->Flags = 0; -#ifdef CONFIG_CARDBUS - if (s->state & SOCKET_CARDBUS) { - struct pci_dev *dev = s->cb_dev; - u_int ptr; - pci_bus_read_config_dword(dev->subordinate, 0, PCI_CARDBUS_CIS, &ptr); - tuple->CISOffset = ptr & ~7; - SPACE(tuple->Flags) = (ptr & 7); - } else -#endif - { - /* Assume presence of a LONGLINK_C to address 0 */ - tuple->CISOffset = tuple->LinkOffset = 0; - SPACE(tuple->Flags) = HAS_LINK(tuple->Flags) = 1; - } - if (!(s->state & SOCKET_CARDBUS) && (s->functions > 1) && - !(tuple->Attributes & TUPLE_RETURN_COMMON)) { + + /* Assume presence of a LONGLINK_C to address 0 */ + tuple->CISOffset = tuple->LinkOffset = 0; + SPACE(tuple->Flags) = HAS_LINK(tuple->Flags) = 1; + + if ((s->functions > 1) && !(tuple->Attributes & TUPLE_RETURN_COMMON)) { cisdata_t req = tuple->DesiredTuple; tuple->DesiredTuple = CISTPL_LONGLINK_MFC; if (pccard_get_next_tuple(s, function, tuple) == 0) { @@ -481,7 +469,7 @@ static int follow_link(struct pcmcia_socket *s, tuple_t *tuple) } else { return -1; } - if (!(s->state & SOCKET_CARDBUS) && SPACE(tuple->Flags)) { + if (SPACE(tuple->Flags)) { /* This is ugly, but a common CIS error is to code the long link offset incorrectly, so we check the right spot... */ read_cis_cache(s, SPACE(tuple->Flags), ofs, 5, link); @@ -507,7 +495,7 @@ int pccard_get_next_tuple(struct pcmcia_socket *s, unsigned int function, tuple_ if (!s) return -EINVAL; - if (!(s->state & SOCKET_PRESENT)) + if (!(s->state & SOCKET_PRESENT) || (s->state & SOCKET_CARDBUS)) return -ENODEV; link[1] = tuple->TupleLink; @@ -1192,119 +1180,6 @@ static int parse_cftable_entry(tuple_t *tuple, /*====================================================================*/ -#ifdef CONFIG_CARDBUS - -static int parse_bar(tuple_t *tuple, cistpl_bar_t *bar) -{ - u_char *p; - if (tuple->TupleDataLen < 6) - return -EINVAL; - p = (u_char *)tuple->TupleData; - bar->attr = *p; - p += 2; - bar->size = get_unaligned_le32(p); - return 0; -} - -static int parse_config_cb(tuple_t *tuple, cistpl_config_t *config) -{ - u_char *p; - - p = (u_char *)tuple->TupleData; - if ((*p != 3) || (tuple->TupleDataLen < 6)) - return -EINVAL; - config->last_idx = *(++p); - p++; - config->base = get_unaligned_le32(p); - config->subtuples = tuple->TupleDataLen - 6; - return 0; -} - -static int parse_cftable_entry_cb(tuple_t *tuple, - cistpl_cftable_entry_cb_t *entry) -{ - u_char *p, *q, features; - - p = tuple->TupleData; - q = p + tuple->TupleDataLen; - entry->index = *p & 0x3f; - entry->flags = 0; - if (*p & 0x40) - entry->flags |= CISTPL_CFTABLE_DEFAULT; - - /* Process optional features */ - if (++p == q) - return -EINVAL; - features = *p; p++; - - /* Power options */ - if ((features & 3) > 0) { - p = parse_power(p, q, &entry->vcc); - if (p == NULL) - return -EINVAL; - } else - entry->vcc.present = 0; - if ((features & 3) > 1) { - p = parse_power(p, q, &entry->vpp1); - if (p == NULL) - return -EINVAL; - } else - entry->vpp1.present = 0; - if ((features & 3) > 2) { - p = parse_power(p, q, &entry->vpp2); - if (p == NULL) - return -EINVAL; - } else - entry->vpp2.present = 0; - - /* I/O window options */ - if (features & 0x08) { - if (p == q) - return -EINVAL; - entry->io = *p; p++; - } else - entry->io = 0; - - /* Interrupt options */ - if (features & 0x10) { - p = parse_irq(p, q, &entry->irq); - if (p == NULL) - return -EINVAL; - } else - entry->irq.IRQInfo1 = 0; - - if (features & 0x20) { - if (p == q) - return -EINVAL; - entry->mem = *p; p++; - } else - entry->mem = 0; - - /* Misc features */ - if (features & 0x80) { - if (p == q) - return -EINVAL; - entry->flags |= (*p << 8); - if (*p & 0x80) { - if (++p == q) - return -EINVAL; - entry->flags |= (*p << 16); - } - while (*p & 0x80) - if (++p == q) - return -EINVAL; - p++; - } - - entry->subtuples = q-p; - - return 0; -} - -#endif - -/*====================================================================*/ - static int parse_device_geo(tuple_t *tuple, cistpl_device_geo_t *geo) { u_char *p, *q; @@ -1406,17 +1281,6 @@ int pcmcia_parse_tuple(tuple_t *tuple, cisparse_t *parse) case CISTPL_DEVICE_A: ret = parse_device(tuple, &parse->device); break; -#ifdef CONFIG_CARDBUS - case CISTPL_BAR: - ret = parse_bar(tuple, &parse->bar); - break; - case CISTPL_CONFIG_CB: - ret = parse_config_cb(tuple, &parse->config); - break; - case CISTPL_CFTABLE_ENTRY_CB: - ret = parse_cftable_entry_cb(tuple, &parse->cftable_entry_cb); - break; -#endif case CISTPL_CHECKSUM: ret = parse_checksum(tuple, &parse->checksum); break; diff --git a/include/pcmcia/ss.h b/include/pcmcia/ss.h index b0ebd11..b4e5cfd 100644 --- a/include/pcmcia/ss.h +++ b/include/pcmcia/ss.h @@ -244,12 +244,6 @@ struct pcmcia_socket { #endif /* CONFIG_PCMCIA_IOCTL */ #endif /* CONFIG_PCMCIA */ - /* cardbus (32-bit) */ -#ifdef CONFIG_CARDBUS - struct resource *cb_cis_res; - void __iomem *cb_cis_virt; -#endif /* CONFIG_CARDBUS */ - /* socket device */ struct device dev; /* data internal to the socket driver */ -- cgit v0.10.2 From 180c33ee409eb3ed605d4ad9884e4a526a7655ff Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Sat, 2 Jan 2010 17:34:09 +0100 Subject: pcmcia: call CIS cleanup from ds.c As ds.c is the only real user of CIS access functions, call the cleanup functions from ds.c, too. Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/cs.c b/drivers/pcmcia/cs.c index 8c51493..9d8b9c1 100644 --- a/drivers/pcmcia/cs.c +++ b/drivers/pcmcia/cs.c @@ -283,8 +283,6 @@ void pcmcia_unregister_socket(struct pcmcia_socket *socket) if (socket->thread) kthread_stop(socket->thread); - release_cis_mem(socket); - /* remove from our own list */ down_write(&pcmcia_socket_list_rwsem); list_del(&socket->socket_list); @@ -399,7 +397,6 @@ static void socket_shutdown(struct pcmcia_socket *s) s->ops->set_socket(s, &s->socket); s->irq.AssignedIRQ = s->irq.Config = 0; s->lock_count = 0; - destroy_cis_cache(s); kfree(s->fake_cis); s->fake_cis = NULL; #ifdef CONFIG_CARDBUS diff --git a/drivers/pcmcia/ds.c b/drivers/pcmcia/ds.c index 87e0639..7bb52b0 100644 --- a/drivers/pcmcia/ds.c +++ b/drivers/pcmcia/ds.c @@ -1241,10 +1241,12 @@ static int ds_event(struct pcmcia_socket *skt, event_t event, int priority) s->pcmcia_state.present = 0; pcmcia_card_remove(skt, NULL); handle_event(skt, event); + destroy_cis_cache(s); break; case CS_EVENT_CARD_INSERTION: s->pcmcia_state.present = 1; + destroy_cis_cache(s); /* to be on the safe side... */ pcmcia_card_add(skt); handle_event(skt, event); break; @@ -1366,6 +1368,7 @@ static void pcmcia_bus_remove_socket(struct device *dev, /* unregister any unbound devices */ mutex_lock(&socket->skt_mutex); pcmcia_card_remove(socket, NULL); + release_cis_mem(socket); mutex_unlock(&socket->skt_mutex); pcmcia_put_socket(socket); -- cgit v0.10.2 From 3f32b3c093eddc03ed477ae0d7a6938db6b94a05 Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Sat, 2 Jan 2010 22:22:50 +0100 Subject: pcmcia: rsrc_nonstatic io memory probe improvements Add a lot of documentation to the rsrc_nonstatic io memory probe functions. Also, add a first memory probe call -- just checking whether request_resource() succeeds -- upon adding of resources. Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/rsrc_nonstatic.c b/drivers/pcmcia/rsrc_nonstatic.c index b886385..120d5ad 100644 --- a/drivers/pcmcia/rsrc_nonstatic.c +++ b/drivers/pcmcia/rsrc_nonstatic.c @@ -264,18 +264,15 @@ static void do_io_probe(struct pcmcia_socket *s, unsigned int base, } #endif -/*====================================================================== - - This is tricky... when we set up CIS memory, we try to validate - the memory window space allocations. - -======================================================================*/ +/*======================================================================*/ -/* Validation function for cards with a valid CIS */ +/** + * readable() - iomem validation function for cards with a valid CIS + */ static int readable(struct pcmcia_socket *s, struct resource *res, unsigned int *count) { - int ret = -1; + int ret = -EINVAL; s->cis_mem.res = res; s->cis_virt = ioremap(res->start, s->map_size); @@ -286,13 +283,16 @@ static int readable(struct pcmcia_socket *s, struct resource *res, s->cis_virt = NULL; } s->cis_mem.res = NULL; - if ((ret != 0) || (*count == 0)) - return 0; - return 1; + if ((ret) || (*count == 0)) + return -EINVAL; + return 0; } -/* Validation function for simple memory cards */ -static int checksum(struct pcmcia_socket *s, struct resource *res) +/** + * checksum() - iomem validation function for simple memory cards + */ +static int checksum(struct pcmcia_socket *s, struct resource *res, + unsigned int *value) { pccard_mem_map map; int i, a = 0, b = -1, d; @@ -320,61 +320,83 @@ static int checksum(struct pcmcia_socket *s, struct resource *res) iounmap(virt); } - return (b == -1) ? -1 : (a>>1); -} - -static int -cis_readable(struct pcmcia_socket *s, unsigned long base, unsigned long size) -{ - struct resource *res1, *res2; - unsigned int info1, info2; - int ret = 0; - - res1 = claim_region(s, base, size/2, IORESOURCE_MEM, "PCMCIA memprobe"); - res2 = claim_region(s, base + size/2, size/2, IORESOURCE_MEM, - "PCMCIA memprobe"); - - if (res1 && res2) { - ret = readable(s, res1, &info1); - ret += readable(s, res2, &info2); - } + if (b == -1) + return -EINVAL; - free_region(res2); - free_region(res1); + *value = a; - return (ret == 2) && (info1 == info2); + return 0; } -static int -checksum_match(struct pcmcia_socket *s, unsigned long base, unsigned long size) +/** + * do_validate_mem() - low level validate a memory region for PCMCIA use + * @s: PCMCIA socket to validate + * @base: start address of resource to check + * @size: size of resource to check + * @validate: validation function to use + * + * do_validate_mem() splits up the memory region which is to be checked + * into two parts. Both are passed to the @validate() function. If + * @validate() returns non-zero, or the value parameter to @validate() + * is zero, or the value parameter is different between both calls, + * the check fails, and -EINVAL is returned. Else, 0 is returned. + */ +static int do_validate_mem(struct pcmcia_socket *s, + unsigned long base, unsigned long size, + int validate (struct pcmcia_socket *s, + struct resource *res, + unsigned int *value)) { struct resource *res1, *res2; - int a = -1, b = -1; + unsigned int info1 = 1, info2 = 1; + int ret = -EINVAL; res1 = claim_region(s, base, size/2, IORESOURCE_MEM, "PCMCIA memprobe"); res2 = claim_region(s, base + size/2, size/2, IORESOURCE_MEM, "PCMCIA memprobe"); if (res1 && res2) { - a = checksum(s, res1); - b = checksum(s, res2); + ret = 0; + if (validate) { + ret = validate(s, res1, &info1); + ret += validate(s, res2, &info2); + } } free_region(res2); free_region(res1); - return (a == b) && (a >= 0); -} + dev_dbg(&s->dev, "cs: memory probe 0x%06lx-0x%06lx: %p %p %u %u %u", + base, base+size-1, res1, res2, ret, info1, info2); -/*====================================================================== + if ((ret) || (info1 != info2) || (info1 == 0)) + return -EINVAL; - The memory probe. If the memory list includes a 64K-aligned block - below 1MB, we probe in 64K chunks, and as soon as we accumulate at - least mem_limit free space, we quit. + return 0; +} -======================================================================*/ -static int do_mem_probe(u_long base, u_long num, struct pcmcia_socket *s) +/** + * do_mem_probe() - validate a memory region for PCMCIA use + * @s: PCMCIA socket to validate + * @base: start address of resource to check + * @num: size of resource to check + * @validate: validation function to use + * @fallback: validation function to use if validate fails + * + * do_mem_probe() checks a memory region for use by the PCMCIA subsystem. + * To do so, the area is split up into sensible parts, and then passed + * into the @validate() function. Only if @validate() and @fallback() fail, + * the area is marked as unavaibale for use by the PCMCIA subsystem. The + * function returns the size of the usable memory area. + */ +static int do_mem_probe(struct pcmcia_socket *s, u_long base, u_long num, + int validate (struct pcmcia_socket *s, + struct resource *res, + unsigned int *value), + int fallback (struct pcmcia_socket *s, + struct resource *res, + unsigned int *value)) { struct socket_data *s_data = s->resource_data; u_long i, j, bad, fail, step; @@ -392,15 +414,14 @@ static int do_mem_probe(u_long base, u_long num, struct pcmcia_socket *s) for (i = j = base; i < base+num; i = j + step) { if (!fail) { for (j = i; j < base+num; j += step) { - if (cis_readable(s, j, step)) + if (!do_validate_mem(s, j, step, validate)) break; } fail = ((i == base) && (j == base+num)); } - if (fail) { - for (j = i; j < base+num; j += 2*step) - if (checksum_match(s, j, step) && - checksum_match(s, j + step, step)) + if ((fail) && (fallback)) { + for (j = i; j < base+num; j += step) + if (!do_validate_mem(s, j, step, fallback)) break; } if (i != j) { @@ -415,8 +436,14 @@ static int do_mem_probe(u_long base, u_long num, struct pcmcia_socket *s) return num - bad; } + #ifdef CONFIG_PCMCIA_PROBE +/** + * inv_probe() - top-to-bottom search for one usuable high memory area + * @s: PCMCIA socket to validate + * @m: resource_map to check + */ static u_long inv_probe(struct resource_map *m, struct pcmcia_socket *s) { struct socket_data *s_data = s->resource_data; @@ -431,9 +458,18 @@ static u_long inv_probe(struct resource_map *m, struct pcmcia_socket *s) } if (m->base < 0x100000) return 0; - return do_mem_probe(m->base, m->num, s); + return do_mem_probe(s, m->base, m->num, readable, checksum); } +/** + * validate_mem() - memory probe function + * @s: PCMCIA socket to validate + * @probe_mask: MEM_PROBE_LOW | MEM_PROBE_HIGH + * + * The memory probe. If the memory list includes a 64K-aligned block + * below 1MB, we probe in 64K chunks, and as soon as we accumulate at + * least mem_limit free space, we quit. Returns 0 on usuable ports. + */ static int validate_mem(struct pcmcia_socket *s, unsigned int probe_mask) { struct resource_map *m, mm; @@ -456,7 +492,8 @@ static int validate_mem(struct pcmcia_socket *s, unsigned int probe_mask) if (mm.base >= 0x100000) continue; if ((mm.base | mm.num) & 0xffff) { - ok += do_mem_probe(mm.base, mm.num, s); + ok += do_mem_probe(s, mm.base, mm.num, readable, + checksum); continue; } /* Special probe for 64K-aligned block */ @@ -466,7 +503,8 @@ static int validate_mem(struct pcmcia_socket *s, unsigned int probe_mask) if (ok >= mem_limit) sub_interval(&s_data->mem_db, b, 0x10000); else - ok += do_mem_probe(b, 0x10000, s); + ok += do_mem_probe(s, b, 0x10000, + readable, checksum); } } } @@ -479,6 +517,13 @@ static int validate_mem(struct pcmcia_socket *s, unsigned int probe_mask) #else /* CONFIG_PCMCIA_PROBE */ +/** + * validate_mem() - memory probe function + * @s: PCMCIA socket to validate + * @probe_mask: ignored + * + * Returns 0 on usuable ports. + */ static int validate_mem(struct pcmcia_socket *s, unsigned int probe_mask) { struct resource_map *m, mm; @@ -487,7 +532,7 @@ static int validate_mem(struct pcmcia_socket *s, unsigned int probe_mask) for (m = s_data->mem_db.next; m != &s_data->mem_db; m = mm.next) { mm = *m; - ok += do_mem_probe(mm.base, mm.num, s); + ok += do_mem_probe(s, mm.base, mm.num, readable, checksum); } if (ok > 0) return 0; @@ -497,7 +542,13 @@ static int validate_mem(struct pcmcia_socket *s, unsigned int probe_mask) #endif /* CONFIG_PCMCIA_PROBE */ -/* +/** + * pcmcia_nonstatic_validate_mem() - try to validate iomem for PCMCIA use + * @s: PCMCIA socket to validate + * + * This is tricky... when we set up CIS memory, we try to validate + * the memory window space allocations. + * * Locking note: Must be called with skt_mutex held! */ static int pcmcia_nonstatic_validate_mem(struct pcmcia_socket *s) @@ -515,10 +566,11 @@ static int pcmcia_nonstatic_validate_mem(struct pcmcia_socket *s) probe_mask = MEM_PROBE_HIGH; if (probe_mask & ~s_data->rsrc_mem_probe) { - if (s->state & SOCKET_PRESENT) + if (s->state & SOCKET_PRESENT) { ret = validate_mem(s, probe_mask); - if (!ret) - s_data->rsrc_mem_probe |= probe_mask; + if (!ret) + s_data->rsrc_mem_probe |= probe_mask; + } } mutex_unlock(&rsrc_mutex); @@ -723,6 +775,8 @@ static int adjust_memory(struct pcmcia_socket *s, unsigned int action, unsigned switch (action) { case ADD_MANAGED_RESOURCE: ret = add_interval(&data->mem_db, start, size); + if (!ret) + do_mem_probe(s, start, size, NULL, NULL); break; case REMOVE_MANAGED_RESOURCE: ret = sub_interval(&data->mem_db, start, size); -- cgit v0.10.2 From 593f010bc0d8f7fde2ce948cac3f77f6a3d9db2b Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Sat, 2 Jan 2010 22:59:15 +0100 Subject: pcmcia: do not lock socket driver module in pcmcia_get_socket() Do not lock the socket driver module in pcmcia_get_socket(), as the PCMCIA core can handle a socket module removal: In pcmcia_unregister_socket(), we explicitely wait for the last put_device() to succeed. Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/cs.c b/drivers/pcmcia/cs.c index 9d8b9c1..f0630a6 100644 --- a/drivers/pcmcia/cs.c +++ b/drivers/pcmcia/cs.c @@ -140,19 +140,13 @@ struct pcmcia_socket *pcmcia_get_socket(struct pcmcia_socket *skt) struct device *dev = get_device(&skt->dev); if (!dev) return NULL; - skt = dev_get_drvdata(dev); - if (!try_module_get(skt->owner)) { - put_device(&skt->dev); - return NULL; - } - return skt; + return dev_get_drvdata(dev); } EXPORT_SYMBOL(pcmcia_get_socket); void pcmcia_put_socket(struct pcmcia_socket *skt) { - module_put(skt->owner); put_device(&skt->dev); } EXPORT_SYMBOL(pcmcia_put_socket); -- cgit v0.10.2 From 8b0e58a70a7a41443c779de074288035b014cb94 Mon Sep 17 00:00:00 2001 From: Stephane Chatty Date: Wed, 13 Jan 2010 21:52:34 +0100 Subject: HID: let hid-input accept digitizers Extended IS_INPUT_APPLICATION to accept digitzers that are actual input devices (touchscreens, light pens, touch pads, white boards) Signed-off-by: Stephane Chatty Signed-off-by: Jiri Kosina diff --git a/include/linux/hid.h b/include/linux/hid.h index 8709365..b978c1e 100644 --- a/include/linux/hid.h +++ b/include/linux/hid.h @@ -663,7 +663,7 @@ struct hid_ll_driver { /* Applications from HID Usage Tables 4/8/99 Version 1.1 */ /* We ignore a few input applications that are not widely used */ -#define IS_INPUT_APPLICATION(a) (((a >= 0x00010000) && (a <= 0x00010008)) || (a == 0x00010080) || (a == 0x000c0001) || (a == 0x000d0002)) +#define IS_INPUT_APPLICATION(a) (((a >= 0x00010000) && (a <= 0x00010008)) || (a == 0x00010080) || (a == 0x000c0001) || ((a >= 0x000d0002) && (a <= 0x000d0006))) /* HID core API */ -- cgit v0.10.2 From 4126faf0ab7417fbc6eb99fb0fd407e01e9e9dfe Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 17 Jan 2010 10:24:07 -0800 Subject: x86: Fix breakage of UML from the changes in the rwsem system The patches 5d0b7235d83eefdafda300656e97d368afcafc9a and bafaecd11df15ad5b1e598adc7736afcd38ee13d broke the UML build: On Sun, 17 Jan 2010, Ingo Molnar wrote: > > FYI, -tip testing found that these changes break the UML build: > > kernel/built-in.o: In function `__up_read': > /home/mingo/tip/arch/x86/include/asm/rwsem.h:192: undefined reference to `call_rwsem_wake' > kernel/built-in.o: In function `__up_write': > /home/mingo/tip/arch/x86/include/asm/rwsem.h:210: undefined reference to `call_rwsem_wake' > kernel/built-in.o: In function `__downgrade_write': > /home/mingo/tip/arch/x86/include/asm/rwsem.h:228: undefined reference to `call_rwsem_downgrade_wake' > kernel/built-in.o: In function `__down_read': > /home/mingo/tip/arch/x86/include/asm/rwsem.h:112: undefined reference to `call_rwsem_down_read_failed' > kernel/built-in.o: In function `__down_write_nested': > /home/mingo/tip/arch/x86/include/asm/rwsem.h:154: undefined reference to `call_rwsem_down_write_failed' > collect2: ld returned 1 exit status Add lib/rwsem_64.o to the UML subarch objects to fix. LKML-Reference: Signed-off-by: H. Peter Anvin diff --git a/arch/um/sys-x86_64/Makefile b/arch/um/sys-x86_64/Makefile index 2201e9c..c1ea9eb 100644 --- a/arch/um/sys-x86_64/Makefile +++ b/arch/um/sys-x86_64/Makefile @@ -8,7 +8,8 @@ obj-y = bug.o bugs.o delay.o fault.o ldt.o mem.o ptrace.o ptrace_user.o \ setjmp.o signal.o stub.o stub_segv.o syscalls.o syscall_table.o \ sysrq.o ksyms.o tls.o -subarch-obj-y = lib/csum-partial_64.o lib/memcpy_64.o lib/thunk_64.o +subarch-obj-y = lib/csum-partial_64.o lib/memcpy_64.o lib/thunk_64.o \ + lib/rwsem_64.o subarch-obj-$(CONFIG_MODULES) += kernel/module.o ldt-y = ../sys-i386/ldt.o -- cgit v0.10.2 From 00097c4fdf117d9845d772f571a987ae95523f8c Mon Sep 17 00:00:00 2001 From: Thadeu Lima de Souza Cascardo Date: Sun, 17 Jan 2010 19:44:44 -0200 Subject: x86, trivial: Fix grammo in tsc comment about Geode TSC reliability Signed-off-by: Thadeu Lima de Souza Cascardo Cc: marcelo@kvack.org Cc: dilinger@collabora.co.uk Cc: trivial@kernel.org LKML-Reference: <1263764685-9871-1-git-send-email-cascardo@holoscopio.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c index 597683a..23066ec 100644 --- a/arch/x86/kernel/tsc.c +++ b/arch/x86/kernel/tsc.c @@ -806,7 +806,7 @@ static void __init check_system_tsc_reliable(void) unsigned long res_low, res_high; rdmsr_safe(MSR_GEODE_BUSCONT_CONF0, &res_low, &res_high); - /* Geode_LX - the OLPC CPU has a possibly a very reliable TSC */ + /* Geode_LX - the OLPC CPU has a very reliable TSC */ if (res_low & RTSC_SUSP) tsc_clocksource_reliable = 1; #endif -- cgit v0.10.2 From e03bcb68629c7f0728c95f1afe06ce48565c7713 Mon Sep 17 00:00:00 2001 From: Milton Miller Date: Mon, 18 Jan 2010 13:00:51 +1100 Subject: generic-ipi: Optimize accesses by using DEFINE_PER_CPU_SHARED_ALIGNED for IPI data The smp ipi data is passed around and given write access by other cpus and should be separated from per-cpu data consumed by this cpu. Looking for hot lines, I saw call_function_data shared with tick_cpu_sched. Signed-off-by: Milton Miller Acked-by: Anton Blanchard Acked-by: Jens Axboe Cc: Andrew Morton Cc: Linus Torvalds Cc: Peter Zijlstra Cc: : Nick Piggin LKML-Reference: <20100118020051.GR12666@kryten> Signed-off-by: Ingo Molnar diff --git a/kernel/smp.c b/kernel/smp.c index f104084..9867b6b 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -12,8 +12,6 @@ #include #include -static DEFINE_PER_CPU(struct call_single_queue, call_single_queue); - static struct { struct list_head queue; raw_spinlock_t lock; @@ -33,12 +31,14 @@ struct call_function_data { cpumask_var_t cpumask; }; +static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_function_data, cfd_data); + struct call_single_queue { struct list_head list; raw_spinlock_t lock; }; -static DEFINE_PER_CPU(struct call_function_data, cfd_data); +static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_single_queue, call_single_queue); static int hotplug_cfd(struct notifier_block *nfb, unsigned long action, void *hcpu) @@ -256,7 +256,7 @@ void generic_smp_call_function_single_interrupt(void) } } -static DEFINE_PER_CPU(struct call_single_data, csd_data); +static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_single_data, csd_data); /* * smp_call_function_single - Run a function on a specific CPU -- cgit v0.10.2 From 3d467676abf5f01f5ee99056273a58486968e252 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Mon, 18 Jan 2010 19:33:10 +0900 Subject: sh: Setup early PMB mappings. More and more boards are going to start shipping that boot with the MMU in 32BIT mode by default. Previously we relied on the bootloader to setup PMB mappings for use by the kernel but we also need to cater for boards whose bootloaders don't set them up. If CONFIG_PMB_LEGACY is not enabled we have full control over our PMB mappings and can compress our address space. Usually, the distance between the the cached and uncached mappings of RAM is always 512MB, however we can compress the distance to be the amount of RAM on the board. pmb_init() now becomes much simpler. It no longer has to calculate any mappings, it just has to synchronise the software PMB table with the hardware. Tested on SDK7786 and SH7785LCR. Signed-off-by: Matt Fleming Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/head_32.S b/arch/sh/kernel/head_32.S index e5d421db..8ee31a0 100644 --- a/arch/sh/kernel/head_32.S +++ b/arch/sh/kernel/head_32.S @@ -3,6 +3,7 @@ * arch/sh/kernel/head.S * * Copyright (C) 1999, 2000 Niibe Yutaka & Kaz Kojima + * Copyright (C) 2010 Matt Fleming * * This file is subject to the terms and conditions of the GNU General Public * License. See the file "COPYING" in the main directory of this archive @@ -84,6 +85,236 @@ ENTRY(_stext) ldc r0, r7_bank ! ... and initial thread_info #endif +#if defined(CONFIG_PMB) && !defined(CONFIG_PMB_LEGACY) + /* + * Reconfigure the initial PMB mappings setup by the hardware. + * + * When we boot in 32-bit MMU mode there are 2 PMB entries already + * setup for us. + * + * Entry VPN PPN V SZ C UB WT + * --------------------------------------------------------------- + * 0 0x80000000 0x00000000 1 512MB 1 0 1 + * 1 0xA0000000 0x00000000 1 512MB 0 0 0 + * + * But we reprogram them here because we want complete control over + * our address space and the initial mappings may not map PAGE_OFFSET + * to __MEMORY_START (or even map all of our RAM). + * + * Once we've setup cached and uncached mappings for all of RAM we + * clear the rest of the PMB entries. + * + * This clearing also deals with the fact that PMB entries can persist + * across reboots. The PMB could have been left in any state when the + * reboot occurred, so to be safe we clear all entries and start with + * with a clean slate. + */ + + mov.l .LMMUCR, r1 /* Flush the TLB */ + mov.l @r1, r0 + or #MMUCR_TI, r0 + mov.l r0, @r1 + + mov.l .LMEMORY_SIZE, r5 + mov r5, r7 + + mov #PMB_E_SHIFT, r0 + mov #0x1, r4 + shld r0, r4 + + mov.l .LFIRST_DATA_ENTRY, r0 + mov.l .LPMB_DATA, r1 + mov.l .LFIRST_ADDR_ENTRY, r2 + mov.l .LPMB_ADDR, r3 + + mov #0, r10 + + /* + * r0 = PMB_DATA data field + * r1 = PMB_DATA address field + * r2 = PMB_ADDR data field + * r3 = PMB_ADDR address field + * r4 = PMB_E_SHIFT + * r5 = remaining amount of RAM to map + * r6 = PMB mapping size we're trying to use + * r7 = cached_to_uncached + * r8 = scratch register + * r9 = scratch register + * r10 = number of PMB entries we've setup + */ +.L512: + mov #(512 >> 4), r6 + shll16 r6 + shll8 r6 + + cmp/hi r5, r6 + bt .L128 + + mov #(PMB_SZ_512M >> 2), r9 + shll2 r9 + + /* + * Cached mapping + */ + mov #PMB_C, r8 + or r0, r8 + or r9, r8 + mov.l r8, @r1 + mov.l r2, @r3 + + add r4, r1 /* Increment to the next PMB_DATA entry */ + add r4, r3 /* Increment to the next PMB_ADDR entry */ + + add #1, r10 /* Increment number of PMB entries */ + + /* + * Uncached mapping + */ + mov #(PMB_UB >> 8), r8 + shll8 r8 + + or r0, r8 + or r9, r8 + mov.l r8, @r1 + mov r2, r8 + add r7, r8 + mov.l r8, @r3 + + add r4, r1 /* Increment to the next PMB_DATA entry */ + add r4, r3 /* Increment to the next PMB_ADDR entry */ + + add #1, r10 /* Increment number of PMB entries */ + + sub r6, r5 + add r6, r0 + add r6, r2 + + bra .L512 + +.L128: + mov #(128 >> 4), r6 + shll16 r6 + shll8 r6 + + cmp/hi r5, r6 + bt .L64 + + mov #(PMB_SZ_128M >> 2), r9 + shll2 r9 + + /* + * Cached mapping + */ + mov #PMB_C, r8 + or r0, r8 + or r9, r8 + mov.l r8, @r1 + mov.l r2, @r3 + + add r4, r1 /* Increment to the next PMB_DATA entry */ + add r4, r3 /* Increment to the next PMB_ADDR entry */ + + add #1, r10 /* Increment number of PMB entries */ + + /* + * Uncached mapping + */ + mov #(PMB_UB >> 8), r8 + shll8 r8 + + or r0, r8 + or r9, r8 + mov.l r8, @r1 + mov r2, r8 + add r7, r8 + mov.l r8, @r3 + + add r4, r1 /* Increment to the next PMB_DATA entry */ + add r4, r3 /* Increment to the next PMB_ADDR entry */ + + add #1, r10 /* Increment number of PMB entries */ + + sub r6, r5 + add r6, r0 + add r6, r2 + + bra .L128 + +.L64: + mov #(64 >> 4), r6 + shll16 r6 + shll8 r6 + + cmp/hi r5, r6 + bt .Ldone + + mov #(PMB_SZ_64M >> 2), r9 + shll2 r9 + + /* + * Cached mapping + */ + mov #PMB_C, r8 + or r0, r8 + or r9, r8 + mov.l r8, @r1 + mov.l r2, @r3 + + add r4, r1 /* Increment to the next PMB_DATA entry */ + add r4, r3 /* Increment to the next PMB_ADDR entry */ + + add #1, r10 /* Increment number of PMB entries */ + + /* + * Uncached mapping + */ + mov #(PMB_UB >> 8), r8 + shll8 r8 + + or r0, r8 + or r9, r8 + mov.l r8, @r1 + mov r2, r8 + add r7, r8 + mov.l r8, @r3 + + add r4, r1 /* Increment to the next PMB_DATA entry */ + add r4, r3 /* Increment to the next PMB_ADDR entry */ + + add #1, r10 /* Increment number of PMB entries */ + + sub r6, r5 + add r6, r0 + add r6, r2 + + bra .L64 + +.Ldone: + /* Update cached_to_uncached */ + mov.l .Lcached_to_uncached, r0 + mov.l r7, @r0 + + /* + * Clear the remaining PMB entries. + * + * r3 = entry to begin clearing from + * r10 = number of entries we've setup so far + */ + mov #0, r1 + mov #PMB_ENTRY_MAX, r0 + +.Lagain: + mov.l r1, @r3 /* Clear PMB_ADDR entry */ + add #1, r10 /* Increment the loop counter */ + cmp/eq r0, r10 + bf/s .Lagain + add r4, r3 /* Increment to the next PMB_ADDR entry */ + + mov.l 6f, r0 + icbi @r0 + +#endif /* !CONFIG_PMB_LEGACY */ + #ifndef CONFIG_SH_NO_BSS_INIT /* * Don't clear BSS if running on slow platforms such as an RTL simulation, @@ -133,3 +364,13 @@ ENTRY(stack_start) 5: .long start_kernel 6: .long sh_cpu_init 7: .long init_thread_union + +#if defined(CONFIG_PMB) && !defined(CONFIG_PMB_LEGACY) +.LPMB_ADDR: .long PMB_ADDR +.LPMB_DATA: .long PMB_DATA +.LFIRST_ADDR_ENTRY: .long PAGE_OFFSET | PMB_V +.LFIRST_DATA_ENTRY: .long __MEMORY_START | PMB_V +.LMMUCR: .long MMUCR +.Lcached_to_uncached: .long cached_to_uncached +.LMEMORY_SIZE: .long __MEMORY_SIZE +#endif diff --git a/arch/sh/mm/pmb.c b/arch/sh/mm/pmb.c index 8f7dbf1..b796b6c 100644 --- a/arch/sh/mm/pmb.c +++ b/arch/sh/mm/pmb.c @@ -3,11 +3,8 @@ * * Privileged Space Mapping Buffer (PMB) Support. * - * Copyright (C) 2005 - 2010 Paul Mundt - * - * P1/P2 Section mapping definitions from map32.h, which was: - * - * Copyright 2003 (c) Lineo Solutions,Inc. + * Copyright (C) 2005 - 2010 Paul Mundt + * Copyright (C) 2010 Matt Fleming * * This file is subject to the terms and conditions of the GNU General Public * License. See the file "COPYING" in the main directory of this archive @@ -280,46 +277,82 @@ static void __pmb_unmap(struct pmb_entry *pmbe) } #ifdef CONFIG_PMB_LEGACY +static inline unsigned int pmb_ppn_in_range(unsigned long ppn) +{ + return ppn >= __MEMORY_START && ppn < __MEMORY_START + __MEMORY_SIZE; +} + static int pmb_apply_legacy_mappings(void) { - int i; - unsigned long addr, data; unsigned int applied = 0; + int i; - for (i = 0; i < PMB_ENTRY_MAX; i++) { - struct pmb_entry *pmbe; - unsigned long vpn, ppn, flags; - - addr = PMB_DATA + (i << PMB_E_SHIFT); - data = ctrl_inl(addr); - if (!(data & PMB_V)) - continue; + pr_info("PMB: Preserving legacy mappings:\n"); - if (data & PMB_C) { -#if defined(CONFIG_CACHE_WRITETHROUGH) - data |= PMB_WT; -#elif defined(CONFIG_CACHE_WRITEBACK) - data &= ~PMB_WT; -#else - data &= ~(PMB_C | PMB_WT); -#endif - } - ctrl_outl(data, addr); - - ppn = data & PMB_PFN_MASK; + /* + * The following entries are setup by the bootloader. + * + * Entry VPN PPN V SZ C UB + * -------------------------------------------------------- + * 0 0xA0000000 0x00000000 1 64MB 0 0 + * 1 0xA4000000 0x04000000 1 16MB 0 0 + * 2 0xA6000000 0x08000000 1 16MB 0 0 + * 9 0x88000000 0x48000000 1 128MB 1 1 + * 10 0x90000000 0x50000000 1 128MB 1 1 + * 11 0x98000000 0x58000000 1 128MB 1 1 + * 13 0xA8000000 0x48000000 1 128MB 0 0 + * 14 0xB0000000 0x50000000 1 128MB 0 0 + * 15 0xB8000000 0x58000000 1 128MB 0 0 + * + * The only entries the we need are the ones that map the kernel + * at the cached and uncached addresses. + */ + for (i = 0; i < PMB_ENTRY_MAX; i++) { + unsigned long addr, data; + unsigned long addr_val, data_val; + unsigned long ppn, vpn; - flags = data & (PMB_C | PMB_WT | PMB_UB); - flags |= data & PMB_SZ_MASK; + addr = mk_pmb_addr(i); + data = mk_pmb_data(i); - addr = PMB_ADDR + (i << PMB_E_SHIFT); - data = ctrl_inl(addr); + addr_val = __raw_readl(addr); + data_val = __raw_readl(data); - vpn = data & PMB_PFN_MASK; + /* + * Skip over any bogus entries + */ + if (!(data_val & PMB_V) || !(addr_val & PMB_V)) + continue; - pmbe = pmb_alloc(vpn, ppn, flags, i); - WARN_ON(IS_ERR(pmbe)); + ppn = data_val & PMB_PFN_MASK; + vpn = addr_val & PMB_PFN_MASK; - applied++; + /* + * Only preserve in-range mappings. + */ + if (pmb_ppn_in_range(ppn)) { + unsigned int size; + char *sz_str = NULL; + + size = data_val & PMB_SZ_MASK; + + sz_str = (size == PMB_SZ_16M) ? " 16MB": + (size == PMB_SZ_64M) ? " 64MB": + (size == PMB_SZ_128M) ? "128MB": + "512MB"; + + pr_info("\t0x%08lx -> 0x%08lx [ %s %scached ]\n", + vpn >> PAGE_SHIFT, ppn >> PAGE_SHIFT, sz_str, + (data_val & PMB_C) ? "" : "un"); + + applied++; + } else { + /* + * Invalidate anything out of bounds. + */ + __raw_writel(addr_val & ~PMB_V, addr); + __raw_writel(data_val & ~PMB_V, data); + } } return (applied == 0); @@ -333,8 +366,9 @@ static inline int pmb_apply_legacy_mappings(void) int __uses_jump_to_uncached pmb_init(void) { - unsigned int i; - unsigned long size, ret; + int i; + unsigned long addr, data; + unsigned long ret; jump_to_uncached(); @@ -352,25 +386,45 @@ int __uses_jump_to_uncached pmb_init(void) } /* - * Insert PMB entries for the P1 and P2 areas so that, after - * we've switched the MMU to 32-bit mode, the semantics of P1 - * and P2 are the same as in 29-bit mode, e.g. - * - * P1 - provides a cached window onto physical memory - * P2 - provides an uncached window onto physical memory + * Sync our software copy of the PMB mappings with those in + * hardware. The mappings in the hardware PMB were either set up + * by the bootloader or very early on by the kernel. */ - size = (unsigned long)__MEMORY_START + __MEMORY_SIZE; + for (i = 0; i < PMB_ENTRY_MAX; i++) { + struct pmb_entry *pmbe; + unsigned long vpn, ppn, flags; - ret = pmb_remap(P1SEG, 0x00000000, size, PMB_C); - BUG_ON(ret != size); + addr = PMB_DATA + (i << PMB_E_SHIFT); + data = ctrl_inl(addr); + if (!(data & PMB_V)) + continue; - ret = pmb_remap(P2SEG, 0x00000000, size, PMB_WT | PMB_UB); - BUG_ON(ret != size); + if (data & PMB_C) { +#if defined(CONFIG_CACHE_WRITETHROUGH) + data |= PMB_WT; +#elif defined(CONFIG_CACHE_WRITEBACK) + data &= ~PMB_WT; +#else + data &= ~(PMB_C | PMB_WT); +#endif + } + ctrl_outl(data, addr); - ctrl_outl(0, PMB_IRMCR); + ppn = data & PMB_PFN_MASK; + + flags = data & (PMB_C | PMB_WT | PMB_UB); + flags |= data & PMB_SZ_MASK; - /* PMB.SE and UB[7] */ - ctrl_outl(PASCR_SE | (1 << 7), PMB_PASCR); + addr = PMB_ADDR + (i << PMB_E_SHIFT); + data = ctrl_inl(addr); + + vpn = data & PMB_PFN_MASK; + + pmbe = pmb_alloc(vpn, ppn, flags, i); + WARN_ON(IS_ERR(pmbe)); + } + + ctrl_outl(0, PMB_IRMCR); /* Flush out the TLB */ i = ctrl_inl(MMUCR); -- cgit v0.10.2 From 4291b730cd0f0cf98a90d946b6cabbd804397350 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 18 Jan 2010 20:39:49 +0900 Subject: sh: Need IRQs enabled for init_fpu(). This tosses in a local_irq_enable()/disable() pair around the init_fpu() callsite in the FPU state restore exception handler. Fixes up a slab BUG triggered by making a slab cache allocation that can sleep whilst irqs_disabled(). This follows the behaviour undertaken by the x86 implementation. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/cpu/fpu.c b/arch/sh/kernel/cpu/fpu.c index c23e672..f059ed6 100644 --- a/arch/sh/kernel/cpu/fpu.c +++ b/arch/sh/kernel/cpu/fpu.c @@ -56,6 +56,7 @@ void fpu_state_restore(struct pt_regs *regs) } if (!tsk_used_math(tsk)) { + local_irq_enable(); /* * does a slab alloc which can sleep */ @@ -66,6 +67,7 @@ void fpu_state_restore(struct pt_regs *regs) do_group_exit(SIGKILL); return; } + local_irq_disable(); } grab_fpu(regs); -- cgit v0.10.2 From d9b9487af79955a8e8fcddc963f56697e020cfed Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 18 Jan 2010 21:08:32 +0900 Subject: sh: Handle early ioremaps through fixed mappings. This adds in a mem_init_done to work out when a standard ioremap() is possible, falling back to the fixmap based ioremap otherwise. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/processor.h b/arch/sh/include/asm/processor.h index 017e0c1..a522e5d 100644 --- a/arch/sh/include/asm/processor.h +++ b/arch/sh/include/asm/processor.h @@ -101,6 +101,9 @@ struct seq_operations; extern struct pt_regs fake_swapper_regs; +/* arch/sh/mm/init.c */ +extern unsigned int mem_init_done; + /* arch/sh/kernel/setup.c */ const char *get_cpu_subtype(struct sh_cpuinfo *c); extern const struct seq_operations cpuinfo_op; diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c index 30a9b53..fe578a2 100644 --- a/arch/sh/mm/init.c +++ b/arch/sh/mm/init.c @@ -237,6 +237,8 @@ static void __init iommu_init(void) no_iommu_init(); } +unsigned int mem_init_done = 0; + void __init mem_init(void) { int codesize, datasize, initsize; @@ -287,6 +289,8 @@ void __init mem_init(void) /* Initialize the vDSO */ vsyscall_init(); + + mem_init_done = 1; } void free_initmem(void) diff --git a/arch/sh/mm/ioremap_32.c b/arch/sh/mm/ioremap_32.c index c80a816..24f6ba6 100644 --- a/arch/sh/mm/ioremap_32.c +++ b/arch/sh/mm/ioremap_32.c @@ -1,13 +1,13 @@ /* * arch/sh/mm/ioremap.c * + * (C) Copyright 1995 1996 Linus Torvalds + * (C) Copyright 2005 - 2010 Paul Mundt + * * Re-map IO memory to kernel address space so that we can access it. * This is needed for high PCI addresses that aren't mapped in the * 640k-1MB IO memory area on PC's * - * (C) Copyright 1995 1996 Linus Torvalds - * (C) Copyright 2005, 2006 Paul Mundt - * * This file is subject to the terms and conditions of the GNU General * Public License. See the file "COPYING" in the main directory of this * archive for more details. @@ -65,6 +65,12 @@ void __iomem *__ioremap_caller(unsigned long phys_addr, unsigned long size, size = PAGE_ALIGN(last_addr+1) - phys_addr; /* + * If we can't yet use the regular approach, go the fixmap route. + */ + if (!mem_init_done) + return ioremap_fixed(phys_addr, size, __pgprot(flags)); + + /* * Ok, go for it.. */ area = get_vm_area_caller(size, VM_IOREMAP, caller); -- cgit v0.10.2 From edf711b66f2a8a9629aa253ad8b3a5a03b5d7af5 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 18 Jan 2010 21:20:13 +0900 Subject: sh: Fixup the IOREMAP_FIXED=n build. Presently the fixed ioremap API is only defined when CONFIG_IOREMAP_FIXED is set. As we want to call in to it unconditionally, provide a stubbed out interface. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index bee5965..9aa9438 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -241,6 +241,15 @@ void __iounmap(void __iomem *addr); extern void __iomem *ioremap_fixed(resource_size_t, unsigned long, pgprot_t); extern void iounmap_fixed(void __iomem *); extern void ioremap_fixed_init(void); +#else +static inline void __iomem * +ioremap_fixed(resource_size t phys_addr, unsigned long size, pgprot_t prot) +{ + BUG(); +} + +static inline void ioremap_fixed_init(void) { } +static inline void iounmap_fixed(void __iomem *addr) { } #endif static inline void __iomem * -- cgit v0.10.2 From 0b59e38ffaf7b201ff6afe5b736365d16848c7e3 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 18 Jan 2010 21:21:32 +0900 Subject: sh: Merge _32/_64 ioremap implementations. There is nothing of interest in the _64 version anymore, so the _32 one can be renamed and used unconditionally. Signed-off-by: Paul Mundt diff --git a/arch/sh/mm/Makefile b/arch/sh/mm/Makefile index 89ba56c..0027cde 100644 --- a/arch/sh/mm/Makefile +++ b/arch/sh/mm/Makefile @@ -15,7 +15,7 @@ obj-y += $(cacheops-y) mmu-y := nommu.o extable_32.o mmu-$(CONFIG_MMU) := extable_$(BITS).o fault_$(BITS).o \ - ioremap_$(BITS).o kmap.o pgtable.o tlbflush_$(BITS).o + ioremap.o kmap.o pgtable.o tlbflush_$(BITS).o obj-y += $(mmu-y) obj-$(CONFIG_DEBUG_FS) += asids-debugfs.o diff --git a/arch/sh/mm/ioremap.c b/arch/sh/mm/ioremap.c new file mode 100644 index 0000000..24f6ba6 --- /dev/null +++ b/arch/sh/mm/ioremap.c @@ -0,0 +1,171 @@ +/* + * arch/sh/mm/ioremap.c + * + * (C) Copyright 1995 1996 Linus Torvalds + * (C) Copyright 2005 - 2010 Paul Mundt + * + * Re-map IO memory to kernel address space so that we can access it. + * This is needed for high PCI addresses that aren't mapped in the + * 640k-1MB IO memory area on PC's + * + * This file is subject to the terms and conditions of the GNU General + * Public License. See the file "COPYING" in the main directory of this + * archive for more details. + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +/* + * Remap an arbitrary physical address space into the kernel virtual + * address space. Needed when the kernel wants to access high addresses + * directly. + * + * NOTE! We need to allow non-page-aligned mappings too: we will obviously + * have to convert them into an offset in a page-aligned mapping, but the + * caller shouldn't need to know that small detail. + */ +void __iomem *__ioremap_caller(unsigned long phys_addr, unsigned long size, + unsigned long flags, void *caller) +{ + struct vm_struct *area; + unsigned long offset, last_addr, addr, orig_addr; + pgprot_t pgprot; + + /* Don't allow wraparound or zero size */ + last_addr = phys_addr + size - 1; + if (!size || last_addr < phys_addr) + return NULL; + + /* + * If we're in the fixed PCI memory range, mapping through page + * tables is not only pointless, but also fundamentally broken. + * Just return the physical address instead. + * + * For boards that map a small PCI memory aperture somewhere in + * P1/P2 space, ioremap() will already do the right thing, + * and we'll never get this far. + */ + if (is_pci_memory_fixed_range(phys_addr, size)) + return (void __iomem *)phys_addr; + + /* + * Mappings have to be page-aligned + */ + offset = phys_addr & ~PAGE_MASK; + phys_addr &= PAGE_MASK; + size = PAGE_ALIGN(last_addr+1) - phys_addr; + + /* + * If we can't yet use the regular approach, go the fixmap route. + */ + if (!mem_init_done) + return ioremap_fixed(phys_addr, size, __pgprot(flags)); + + /* + * Ok, go for it.. + */ + area = get_vm_area_caller(size, VM_IOREMAP, caller); + if (!area) + return NULL; + area->phys_addr = phys_addr; + orig_addr = addr = (unsigned long)area->addr; + +#ifdef CONFIG_PMB + /* + * First try to remap through the PMB once a valid VMA has been + * established. Smaller allocations (or the rest of the size + * remaining after a PMB mapping due to the size not being + * perfectly aligned on a PMB size boundary) are then mapped + * through the UTLB using conventional page tables. + * + * PMB entries are all pre-faulted. + */ + if (unlikely(phys_addr >= P1SEG)) { + unsigned long mapped = pmb_remap(addr, phys_addr, size, flags); + + if (likely(mapped)) { + addr += mapped; + phys_addr += mapped; + size -= mapped; + } + } +#endif + + pgprot = __pgprot(pgprot_val(PAGE_KERNEL_NOCACHE) | flags); + if (likely(size)) + if (ioremap_page_range(addr, addr + size, phys_addr, pgprot)) { + vunmap((void *)orig_addr); + return NULL; + } + + return (void __iomem *)(offset + (char *)orig_addr); +} +EXPORT_SYMBOL(__ioremap_caller); + +/* + * Simple checks for non-translatable mappings. + */ +static inline int iomapping_nontranslatable(unsigned long offset) +{ +#ifdef CONFIG_29BIT + /* + * In 29-bit mode this includes the fixed P1/P2 areas, as well as + * parts of P3. + */ + if (PXSEG(offset) < P3SEG || offset >= P3_ADDR_MAX) + return 1; +#endif + + if (is_pci_memory_fixed_range(offset, 0)) + return 1; + + return 0; +} + +void __iounmap(void __iomem *addr) +{ + unsigned long vaddr = (unsigned long __force)addr; + struct vm_struct *p; + + /* + * Nothing to do if there is no translatable mapping. + */ + if (iomapping_nontranslatable(vaddr)) + return; + +#ifdef CONFIG_PMB + /* + * Purge any PMB entries that may have been established for this + * mapping, then proceed with conventional VMA teardown. + * + * XXX: Note that due to the way that remove_vm_area() does + * matching of the resultant VMA, we aren't able to fast-forward + * the address past the PMB space until the end of the VMA where + * the page tables reside. As such, unmap_vm_area() will be + * forced to linearly scan over the area until it finds the page + * tables where PTEs that need to be unmapped actually reside, + * which is far from optimal. Perhaps we need to use a separate + * VMA for the PMB mappings? + * -- PFM. + */ + pmb_unmap(vaddr); +#endif + + p = remove_vm_area((void *)(vaddr & PAGE_MASK)); + if (!p) { + printk(KERN_ERR "%s: bad address %p\n", __func__, addr); + return; + } + + kfree(p); +} +EXPORT_SYMBOL(__iounmap); diff --git a/arch/sh/mm/ioremap_32.c b/arch/sh/mm/ioremap_32.c deleted file mode 100644 index 24f6ba6..0000000 --- a/arch/sh/mm/ioremap_32.c +++ /dev/null @@ -1,171 +0,0 @@ -/* - * arch/sh/mm/ioremap.c - * - * (C) Copyright 1995 1996 Linus Torvalds - * (C) Copyright 2005 - 2010 Paul Mundt - * - * Re-map IO memory to kernel address space so that we can access it. - * This is needed for high PCI addresses that aren't mapped in the - * 640k-1MB IO memory area on PC's - * - * This file is subject to the terms and conditions of the GNU General - * Public License. See the file "COPYING" in the main directory of this - * archive for more details. - */ -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -/* - * Remap an arbitrary physical address space into the kernel virtual - * address space. Needed when the kernel wants to access high addresses - * directly. - * - * NOTE! We need to allow non-page-aligned mappings too: we will obviously - * have to convert them into an offset in a page-aligned mapping, but the - * caller shouldn't need to know that small detail. - */ -void __iomem *__ioremap_caller(unsigned long phys_addr, unsigned long size, - unsigned long flags, void *caller) -{ - struct vm_struct *area; - unsigned long offset, last_addr, addr, orig_addr; - pgprot_t pgprot; - - /* Don't allow wraparound or zero size */ - last_addr = phys_addr + size - 1; - if (!size || last_addr < phys_addr) - return NULL; - - /* - * If we're in the fixed PCI memory range, mapping through page - * tables is not only pointless, but also fundamentally broken. - * Just return the physical address instead. - * - * For boards that map a small PCI memory aperture somewhere in - * P1/P2 space, ioremap() will already do the right thing, - * and we'll never get this far. - */ - if (is_pci_memory_fixed_range(phys_addr, size)) - return (void __iomem *)phys_addr; - - /* - * Mappings have to be page-aligned - */ - offset = phys_addr & ~PAGE_MASK; - phys_addr &= PAGE_MASK; - size = PAGE_ALIGN(last_addr+1) - phys_addr; - - /* - * If we can't yet use the regular approach, go the fixmap route. - */ - if (!mem_init_done) - return ioremap_fixed(phys_addr, size, __pgprot(flags)); - - /* - * Ok, go for it.. - */ - area = get_vm_area_caller(size, VM_IOREMAP, caller); - if (!area) - return NULL; - area->phys_addr = phys_addr; - orig_addr = addr = (unsigned long)area->addr; - -#ifdef CONFIG_PMB - /* - * First try to remap through the PMB once a valid VMA has been - * established. Smaller allocations (or the rest of the size - * remaining after a PMB mapping due to the size not being - * perfectly aligned on a PMB size boundary) are then mapped - * through the UTLB using conventional page tables. - * - * PMB entries are all pre-faulted. - */ - if (unlikely(phys_addr >= P1SEG)) { - unsigned long mapped = pmb_remap(addr, phys_addr, size, flags); - - if (likely(mapped)) { - addr += mapped; - phys_addr += mapped; - size -= mapped; - } - } -#endif - - pgprot = __pgprot(pgprot_val(PAGE_KERNEL_NOCACHE) | flags); - if (likely(size)) - if (ioremap_page_range(addr, addr + size, phys_addr, pgprot)) { - vunmap((void *)orig_addr); - return NULL; - } - - return (void __iomem *)(offset + (char *)orig_addr); -} -EXPORT_SYMBOL(__ioremap_caller); - -/* - * Simple checks for non-translatable mappings. - */ -static inline int iomapping_nontranslatable(unsigned long offset) -{ -#ifdef CONFIG_29BIT - /* - * In 29-bit mode this includes the fixed P1/P2 areas, as well as - * parts of P3. - */ - if (PXSEG(offset) < P3SEG || offset >= P3_ADDR_MAX) - return 1; -#endif - - if (is_pci_memory_fixed_range(offset, 0)) - return 1; - - return 0; -} - -void __iounmap(void __iomem *addr) -{ - unsigned long vaddr = (unsigned long __force)addr; - struct vm_struct *p; - - /* - * Nothing to do if there is no translatable mapping. - */ - if (iomapping_nontranslatable(vaddr)) - return; - -#ifdef CONFIG_PMB - /* - * Purge any PMB entries that may have been established for this - * mapping, then proceed with conventional VMA teardown. - * - * XXX: Note that due to the way that remove_vm_area() does - * matching of the resultant VMA, we aren't able to fast-forward - * the address past the PMB space until the end of the VMA where - * the page tables reside. As such, unmap_vm_area() will be - * forced to linearly scan over the area until it finds the page - * tables where PTEs that need to be unmapped actually reside, - * which is far from optimal. Perhaps we need to use a separate - * VMA for the PMB mappings? - * -- PFM. - */ - pmb_unmap(vaddr); -#endif - - p = remove_vm_area((void *)(vaddr & PAGE_MASK)); - if (!p) { - printk(KERN_ERR "%s: bad address %p\n", __func__, addr); - return; - } - - kfree(p); -} -EXPORT_SYMBOL(__iounmap); diff --git a/arch/sh/mm/ioremap_64.c b/arch/sh/mm/ioremap_64.c deleted file mode 100644 index fb0aa45..0000000 --- a/arch/sh/mm/ioremap_64.c +++ /dev/null @@ -1,47 +0,0 @@ -/* - * arch/sh/mm/ioremap_64.c - * - * Copyright (C) 2000, 2001 Paolo Alberelli - * Copyright (C) 2003 - 2007 Paul Mundt - * - * Mostly derived from arch/sh/mm/ioremap.c which, in turn is mostly - * derived from arch/i386/mm/ioremap.c . - * - * (C) Copyright 1995 1996 Linus Torvalds - * - * This file is subject to the terms and conditions of the GNU General Public - * License. See the file "COPYING" in the main directory of this archive - * for more details. - */ -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -void __iomem *__ioremap_caller(unsigned long offset, unsigned long size, - unsigned long flags, void *caller) -{ - pgprot_t prot; - - prot = __pgprot(_PAGE_PRESENT | _PAGE_READ | _PAGE_WRITE | - _PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_SHARED | flags); - - return ioremap_fixed(offset, size, prot); -} -EXPORT_SYMBOL(__ioremap_caller); - -void __iounmap(void __iomem *virtual) -{ - iounmap_fixed(virtual); -} -EXPORT_SYMBOL(__iounmap); -- cgit v0.10.2 From 4f744affc345f8b158615e0cdd01d1f4985837c3 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 18 Jan 2010 21:30:29 +0900 Subject: sh: Make iounmap_fixed() return success/failure for iounmap() path. This converts iounmap_fixed() to return success/error if it handled the unmap request or not. At the same time, drop the __init label, as this can be called in to later. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index 9aa9438..6a0dd8c 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -22,6 +22,7 @@ * for old compat code for I/O offseting to SuperIOs, all of which are * better handled through the machvec ioport mapping routines these days. */ +#include #include #include #include @@ -239,7 +240,7 @@ void __iounmap(void __iomem *addr); #ifdef CONFIG_IOREMAP_FIXED extern void __iomem *ioremap_fixed(resource_size_t, unsigned long, pgprot_t); -extern void iounmap_fixed(void __iomem *); +extern int iounmap_fixed(void __iomem *); extern void ioremap_fixed_init(void); #else static inline void __iomem * @@ -249,7 +250,7 @@ ioremap_fixed(resource_size t phys_addr, unsigned long size, pgprot_t prot) } static inline void ioremap_fixed_init(void) { } -static inline void iounmap_fixed(void __iomem *addr) { } +static inline int iounmap_fixed(void __iomem *addr) { return -EINVAL; } #endif static inline void __iomem * diff --git a/arch/sh/mm/ioremap_fixed.c b/arch/sh/mm/ioremap_fixed.c index 3a9d3d8..425f6c6 100644 --- a/arch/sh/mm/ioremap_fixed.c +++ b/arch/sh/mm/ioremap_fixed.c @@ -103,7 +103,7 @@ ioremap_fixed(resource_size_t phys_addr, unsigned long size, pgprot_t prot) return map->addr; } -void __init iounmap_fixed(void __iomem *addr) +int iounmap_fixed(void __iomem *addr) { enum fixed_addresses idx; unsigned long virt_addr; @@ -122,8 +122,11 @@ void __init iounmap_fixed(void __iomem *addr) } } + /* + * If we don't match, it's not for us. + */ if (slot < 0) - return; + return -EINVAL; virt_addr = (unsigned long)addr; @@ -141,4 +144,6 @@ void __init iounmap_fixed(void __iomem *addr) map->size = 0; map->addr = NULL; + + return 0; } -- cgit v0.10.2 From 12b6b01cb47dc3eefbef866592193661dad7afb9 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 18 Jan 2010 21:33:08 +0900 Subject: sh: Handle unmapping of fixed slots transparently in iounmap(). iounmap() should balance whatever is done by ioremap(). Presently ioremap() can do any of fixed mappings, PMB mappings, or page table mappings. Presently only the latter two are handled through the standard unmap path, so tie in the fixed unmapping, too. Signed-off-by: Paul Mundt diff --git a/arch/sh/mm/ioremap.c b/arch/sh/mm/ioremap.c index 24f6ba6..e8b65f6 100644 --- a/arch/sh/mm/ioremap.c +++ b/arch/sh/mm/ioremap.c @@ -142,6 +142,12 @@ void __iounmap(void __iomem *addr) if (iomapping_nontranslatable(vaddr)) return; + /* + * There's no VMA if it's from an early fixed mapping. + */ + if (iounmap_fixed(addr) == 0) + return; + #ifdef CONFIG_PMB /* * Purge any PMB entries that may have been established for this -- cgit v0.10.2 From af1415314a4190b8ea06e53808d392fcf91555af Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 18 Jan 2010 21:45:00 +0900 Subject: sh: Flag __ioremap_caller() __init_refok. The mem_init_done test makes sure that this path is only entered in __init cases, so leaving ioremap_fixed() as __init and flagging the caller __init_refok is sufficient. Signed-off-by: Paul Mundt diff --git a/arch/sh/mm/ioremap.c b/arch/sh/mm/ioremap.c index e8b65f6..a130b22 100644 --- a/arch/sh/mm/ioremap.c +++ b/arch/sh/mm/ioremap.c @@ -33,8 +33,9 @@ * have to convert them into an offset in a page-aligned mapping, but the * caller shouldn't need to know that small detail. */ -void __iomem *__ioremap_caller(unsigned long phys_addr, unsigned long size, - unsigned long flags, void *caller) +void __iomem * __init_refok +__ioremap_caller(unsigned long phys_addr, unsigned long size, + unsigned long flags, void *caller) { struct vm_struct *area; unsigned long offset, last_addr, addr, orig_addr; -- cgit v0.10.2 From e3deec090558d5cb5ffdc574e5560f3ed9723394 Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Tue, 3 Nov 2009 12:33:07 -0600 Subject: [SCSI] eliminate potential kmalloc failure in scsi_get_vpd_page() The best way to fix this is to eliminate the intenal kmalloc() and make the caller allocate the required amount of storage. Signed-off-by: James Bottomley diff --git a/drivers/scsi/scsi.c b/drivers/scsi/scsi.c index a60da55..513661f 100644 --- a/drivers/scsi/scsi.c +++ b/drivers/scsi/scsi.c @@ -1026,55 +1026,39 @@ static int scsi_vpd_inquiry(struct scsi_device *sdev, unsigned char *buffer, * responsible for calling kfree() on this pointer when it is no longer * needed. If we cannot retrieve the VPD page this routine returns %NULL. */ -unsigned char *scsi_get_vpd_page(struct scsi_device *sdev, u8 page) +int scsi_get_vpd_page(struct scsi_device *sdev, u8 page, unsigned char *buf, + int buf_len) { int i, result; - unsigned int len; - const unsigned int init_vpd_len = 255; - unsigned char *buf = kmalloc(init_vpd_len, GFP_KERNEL); - - if (!buf) - return NULL; /* Ask for all the pages supported by this device */ - result = scsi_vpd_inquiry(sdev, buf, 0, init_vpd_len); + result = scsi_vpd_inquiry(sdev, buf, 0, buf_len); if (result) goto fail; /* If the user actually wanted this page, we can skip the rest */ if (page == 0) - return buf; + return -EINVAL; - for (i = 0; i < buf[3]; i++) + for (i = 0; i < min((int)buf[3], buf_len - 4); i++) if (buf[i + 4] == page) goto found; + + if (i < buf[3] && i > buf_len) + /* ran off the end of the buffer, give us benefit of doubt */ + goto found; /* The device claims it doesn't support the requested page */ goto fail; found: - result = scsi_vpd_inquiry(sdev, buf, page, 255); + result = scsi_vpd_inquiry(sdev, buf, page, buf_len); if (result) goto fail; - /* - * Some pages are longer than 255 bytes. The actual length of - * the page is returned in the header. - */ - len = ((buf[2] << 8) | buf[3]) + 4; - if (len <= init_vpd_len) - return buf; - - kfree(buf); - buf = kmalloc(len, GFP_KERNEL); - result = scsi_vpd_inquiry(sdev, buf, page, len); - if (result) - goto fail; - - return buf; + return 0; fail: - kfree(buf); - return NULL; + return -EINVAL; } EXPORT_SYMBOL_GPL(scsi_get_vpd_page); diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c index 255da53..c5e9a99 100644 --- a/drivers/scsi/sd.c +++ b/drivers/scsi/sd.c @@ -1946,13 +1946,13 @@ static void sd_read_block_limits(struct scsi_disk *sdkp) { struct request_queue *q = sdkp->disk->queue; unsigned int sector_sz = sdkp->device->sector_size; - char *buffer; + const int vpd_len = 32; + unsigned char *buffer = kmalloc(vpd_len, GFP_KERNEL); - /* Block Limits VPD */ - buffer = scsi_get_vpd_page(sdkp->device, 0xb0); - - if (buffer == NULL) - return; + if (!buffer || + /* Block Limits VPD */ + scsi_get_vpd_page(sdkp->device, 0xb0, buffer, vpd_len)) + goto out; blk_queue_io_min(sdkp->disk->queue, get_unaligned_be16(&buffer[6]) * sector_sz); @@ -1984,6 +1984,7 @@ static void sd_read_block_limits(struct scsi_disk *sdkp) get_unaligned_be32(&buffer[32]) & ~(1 << 31); } + out: kfree(buffer); } @@ -1993,20 +1994,23 @@ static void sd_read_block_limits(struct scsi_disk *sdkp) */ static void sd_read_block_characteristics(struct scsi_disk *sdkp) { - char *buffer; + unsigned char *buffer; u16 rot; + const int vpd_len = 32; - /* Block Device Characteristics VPD */ - buffer = scsi_get_vpd_page(sdkp->device, 0xb1); + buffer = kmalloc(vpd_len, GFP_KERNEL); - if (buffer == NULL) - return; + if (!buffer || + /* Block Device Characteristics VPD */ + scsi_get_vpd_page(sdkp->device, 0xb1, buffer, vpd_len)) + goto out; rot = get_unaligned_be16(&buffer[4]); if (rot == 1) queue_flag_set_unlocked(QUEUE_FLAG_NONROT, sdkp->disk->queue); + out: kfree(buffer); } diff --git a/drivers/scsi/ses.c b/drivers/scsi/ses.c index 55b034b..1d7a878 100644 --- a/drivers/scsi/ses.c +++ b/drivers/scsi/ses.c @@ -448,13 +448,17 @@ static void ses_match_to_enclosure(struct enclosure_device *edev, .addr = 0, }; - buf = scsi_get_vpd_page(sdev, 0x83); - if (!buf) - return; + buf = kmalloc(INIT_ALLOC_SIZE, GFP_KERNEL); + if (!buf || scsi_get_vpd_page(sdev, 0x83, buf, INIT_ALLOC_SIZE)) + goto free; ses_enclosure_data_process(edev, to_scsi_device(edev->edev.parent), 0); vpd_len = ((buf[2] << 8) | buf[3]) + 4; + kfree(buf); + buf = kmalloc(vpd_len, GFP_KERNEL); + if (!buf ||scsi_get_vpd_page(sdev, 0x83, buf, vpd_len)) + goto free; desc = buf + 4; while (desc < buf + vpd_len) { diff --git a/include/scsi/scsi_device.h b/include/scsi/scsi_device.h index 7c44499..d80b6db 100644 --- a/include/scsi/scsi_device.h +++ b/include/scsi/scsi_device.h @@ -348,7 +348,8 @@ extern int scsi_mode_select(struct scsi_device *sdev, int pf, int sp, struct scsi_sense_hdr *); extern int scsi_test_unit_ready(struct scsi_device *sdev, int timeout, int retries, struct scsi_sense_hdr *sshdr); -extern unsigned char *scsi_get_vpd_page(struct scsi_device *, u8 page); +extern int scsi_get_vpd_page(struct scsi_device *, u8 page, unsigned char *buf, + int buf_len); extern int scsi_device_set_state(struct scsi_device *sdev, enum scsi_device_state state); extern struct scsi_event *sdev_evt_alloc(enum scsi_device_event evt_type, -- cgit v0.10.2 From 40c4f3e4eaa7eda2b0a00abd4d69778808d26f77 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Fri, 4 Dec 2009 20:38:49 +0100 Subject: [SCSI] libsrp: fix typo -- replace RDAM by RDMA Fixed a typo in libsrp.c: replaced two occurrences of 'RDAM' by 'RDMA'. Signed-off-by: Bart Van Assche Acked-by: FUJITA Tomonori Signed-off-by: James Bottomley diff --git a/drivers/scsi/libsrp.c b/drivers/scsi/libsrp.c index ab19b3b..f79602f 100644 --- a/drivers/scsi/libsrp.c +++ b/drivers/scsi/libsrp.c @@ -1,5 +1,5 @@ /* - * SCSI RDAM Protocol lib functions + * SCSI RDMA Protocol lib functions * * Copyright (C) 2006 FUJITA Tomonori * @@ -440,6 +440,6 @@ int srp_cmd_queue(struct Scsi_Host *shost, struct srp_cmd *cmd, void *info, } EXPORT_SYMBOL_GPL(srp_cmd_queue); -MODULE_DESCRIPTION("SCSI RDAM Protocol lib functions"); +MODULE_DESCRIPTION("SCSI RDMA Protocol lib functions"); MODULE_AUTHOR("FUJITA Tomonori"); MODULE_LICENSE("GPL"); -- cgit v0.10.2 From 9b7dac086bec7c71722d6e79464609e17f0996e7 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Fri, 4 Dec 2009 20:43:37 +0100 Subject: [SCSI] ibmvscsi: fix a typo in a source code comment Signed-off-by: Bart Van Assche Acked-by: Brian King Signed-off-by: James Bottomley diff --git a/drivers/scsi/ibmvscsi/ibmvscsi.c b/drivers/scsi/ibmvscsi/ibmvscsi.c index e475b79..e3a18e0 100644 --- a/drivers/scsi/ibmvscsi/ibmvscsi.c +++ b/drivers/scsi/ibmvscsi/ibmvscsi.c @@ -40,7 +40,7 @@ * (CRQ), which is just a buffer of 16 byte entries in the receiver's * Senders cannot access the buffer directly, but send messages by * making a hypervisor call and passing in the 16 bytes. The hypervisor - * puts the message in the next 16 byte space in round-robbin fashion, + * puts the message in the next 16 byte space in round-robin fashion, * turns on the high order bit of the message (the valid bit), and * generates an interrupt to the receiver (if interrupts are turned on.) * The receiver just turns off the valid bit when they have copied out -- cgit v0.10.2 From 8fe79162a6807bf120140d64e96da54fc273b88b Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Tue, 8 Dec 2009 14:08:57 -0800 Subject: [SCSI] eata: fix buffer overflow Allows i == MAX_INT_PARAM, which is out of range. Signed-off-by: Roel Kluin Signed-off-by: Andrew Morton Signed-off-by: James Bottomley diff --git a/drivers/scsi/eata.c b/drivers/scsi/eata.c index c7076ce..3c5abf7 100644 --- a/drivers/scsi/eata.c +++ b/drivers/scsi/eata.c @@ -1509,7 +1509,7 @@ static int option_setup(char *str) char *cur = str; int i = 1; - while (cur && isdigit(*cur) && i <= MAX_INT_PARAM) { + while (cur && isdigit(*cur) && i < MAX_INT_PARAM) { ints[i++] = simple_strtoul(cur, NULL, 0); if ((cur = strchr(cur, ',')) != NULL) -- cgit v0.10.2 From 4a02462af1f4c3498e6a330ffd5c063118309b31 Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Tue, 8 Dec 2009 14:08:59 -0800 Subject: [SCSI] u14-34f: fix buffer overflow This allows i == MAX_INT_PARAM, which is out of range for ints[] Signed-off-by: Roel Kluin Signed-off-by: Andrew Morton Signed-off-by: James Bottomley diff --git a/drivers/scsi/u14-34f.c b/drivers/scsi/u14-34f.c index 54023d41..26e8e0e 100644 --- a/drivers/scsi/u14-34f.c +++ b/drivers/scsi/u14-34f.c @@ -1070,7 +1070,7 @@ static int option_setup(char *str) { char *cur = str; int i = 1; - while (cur && isdigit(*cur) && i <= MAX_INT_PARAM) { + while (cur && isdigit(*cur) && i < MAX_INT_PARAM) { ints[i++] = simple_strtoul(cur, NULL, 0); if ((cur = strchr(cur, ',')) != NULL) cur++; -- cgit v0.10.2 From 340f052001d46aff9e7e853c492e0d3f5554d42f Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Tue, 8 Dec 2009 14:08:56 -0800 Subject: [SCSI] ibmmca: fix buffer overflow Allows i == IM_MAX_HOSTS, which is out of range. Signed-off-by: Roel Kluin Signed-off-by: Andrew Morton Signed-off-by: James Bottomley diff --git a/drivers/scsi/ibmmca.c b/drivers/scsi/ibmmca.c index 9c1e6a5..9a4b69d 100644 --- a/drivers/scsi/ibmmca.c +++ b/drivers/scsi/ibmmca.c @@ -2336,7 +2336,7 @@ static int option_setup(char *str) char *cur = str; int i = 1; - while (cur && isdigit(*cur) && i <= IM_MAX_HOSTS) { + while (cur && isdigit(*cur) && i < IM_MAX_HOSTS) { ints[i++] = simple_strtoul(cur, NULL, 0); if ((cur = strchr(cur, ',')) != NULL) cur++; -- cgit v0.10.2 From e47c11c7a402a054a85cb917a6ed020f6b5fae04 Mon Sep 17 00:00:00 2001 From: Erik Ekman Date: Mon, 14 Dec 2009 21:21:56 +0100 Subject: [SCSI] fusion: fix warning when not using procfs Fixes the following warning: drivers/message/fusion/mptbase.c:129: warning: 'mpt_proc_root_dir' defined but not used also moves it from public data section since it is static. Signed-off-by: Erik Ekman Acked-by: "Desai, Kashyap" Signed-off-by: James Bottomley diff --git a/drivers/message/fusion/mptbase.c b/drivers/message/fusion/mptbase.c index 44d2037..5382b5a 100644 --- a/drivers/message/fusion/mptbase.c +++ b/drivers/message/fusion/mptbase.c @@ -126,8 +126,6 @@ static int mfcounter = 0; * Public data... */ -static struct proc_dir_entry *mpt_proc_root_dir; - #define WHOINIT_UNKNOWN 0xAA /*=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=*/ @@ -146,6 +144,9 @@ static MPT_EVHANDLER MptEvHandlers[MPT_MAX_PROTOCOL_DRIVERS]; static MPT_RESETHANDLER MptResetHandlers[MPT_MAX_PROTOCOL_DRIVERS]; static struct mpt_pci_driver *MptDeviceDriverHandlers[MPT_MAX_PROTOCOL_DRIVERS]; +#ifdef CONFIG_PROC_FS +static struct proc_dir_entry *mpt_proc_root_dir; +#endif /* * Driver Callback Index's -- cgit v0.10.2 From f2818663c82b7297ff4aa38cbddb870dc02f7104 Mon Sep 17 00:00:00 2001 From: Hannes Reinecke Date: Tue, 15 Dec 2009 09:26:06 +0100 Subject: [SCSI] scsi_transport_fc: Remove capping from dev_loss_tmo Currently dev_loss_tmo is capped by SCSI_DEVICE_BLOCK_MAX_TIMEOUT. This causes problem with multipathing when the 'no_path_retry' setting exceeds the dev_loss_tmo setting, as then the system might run into a deadlock when all paths have been removed temporarily for longer than dev_loss_tmo. The principal reasons for the capping has been that we should not allow a remote port to remain in status 'blocked' indefinitely, so the capping is there to ensure that the port status is being reset eventually. However, the fast_io_fail_tmo will also move the remote port out of the 'blocked' state, so for any HBA driver implementing both the capping should really be on the fast_io_fail_tmo, and not on the dev_loss_tmo. This patch implements just that, ie the fast_io_fail_tmo is capped to SCSI_DEVICE_BLOCK_TIMEOUT and the capping is removed from dev_loss_tmo when fast_io_fail_tmo is set. This allows us to synchronize the dev_loss_tmo setting to the 'no_path_retry' setting from multipathing thus avoiding the deadlock. Signed-off-by: Hannes Reinecke Acked-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/scsi_transport_fc.c b/drivers/scsi/scsi_transport_fc.c index 653f22a..79660ee 100644 --- a/drivers/scsi/scsi_transport_fc.c +++ b/drivers/scsi/scsi_transport_fc.c @@ -475,7 +475,8 @@ MODULE_PARM_DESC(dev_loss_tmo, "Maximum number of seconds that the FC transport should" " insulate the loss of a remote port. Once this value is" " exceeded, the scsi target is removed. Value should be" - " between 1 and SCSI_DEVICE_BLOCK_MAX_TIMEOUT."); + " between 1 and SCSI_DEVICE_BLOCK_MAX_TIMEOUT if" + " fast_io_fail_tmo is not set."); /* * Netlink Infrastructure @@ -842,9 +843,17 @@ store_fc_rport_dev_loss_tmo(struct device *dev, struct device_attribute *attr, (rport->port_state == FC_PORTSTATE_NOTPRESENT)) return -EBUSY; val = simple_strtoul(buf, &cp, 0); - if ((*cp && (*cp != '\n')) || - (val < 0) || (val > SCSI_DEVICE_BLOCK_MAX_TIMEOUT)) + if ((*cp && (*cp != '\n')) || (val < 0)) return -EINVAL; + + /* + * If fast_io_fail is off we have to cap + * dev_loss_tmo at SCSI_DEVICE_BLOCK_MAX_TIMEOUT + */ + if (rport->fast_io_fail_tmo == -1 && + val > SCSI_DEVICE_BLOCK_MAX_TIMEOUT) + return -EINVAL; + i->f->set_rport_dev_loss_tmo(rport, val); return count; } @@ -925,9 +934,16 @@ store_fc_rport_fast_io_fail_tmo(struct device *dev, rport->fast_io_fail_tmo = -1; else { val = simple_strtoul(buf, &cp, 0); - if ((*cp && (*cp != '\n')) || - (val < 0) || (val >= rport->dev_loss_tmo)) + if ((*cp && (*cp != '\n')) || (val < 0)) return -EINVAL; + /* + * Cap fast_io_fail by dev_loss_tmo or + * SCSI_DEVICE_BLOCK_MAX_TIMEOUT. + */ + if ((val >= rport->dev_loss_tmo) || + (val > SCSI_DEVICE_BLOCK_MAX_TIMEOUT)) + return -EINVAL; + rport->fast_io_fail_tmo = val; } return count; -- cgit v0.10.2 From 02507a80b35edd720480540d917e9f92cc371009 Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Sat, 5 Dec 2009 12:30:42 +1100 Subject: [SCSI] mac_esp: fix PIO mode, take 2 The mac_esp PIO algorithm no longer works in 2.6.31 and crashes my Centris 660av. So here's a better one. Also, force async with esp_set_offset() rather than esp_slave_configure(). One of the SCSI drives I tested still doesn't like the PIO mode and fails with "esp: esp0: Reconnect IRQ2 timeout" (the same drive works fine in PDMA mode). This failure happens when esp_reconnect_with_tag() tries to read in two tag bytes but the chip only provides one (0x20). I don't know what causes this. I decided not to waste any more time trying to fix it because the best solution is to rip out the PIO mode altogether and use the DMA engine. Signed-off-by: Finn Thain Signed-off-by: James Bottomley diff --git a/drivers/scsi/esp_scsi.c b/drivers/scsi/esp_scsi.c index a680e18..e2bc779 100644 --- a/drivers/scsi/esp_scsi.c +++ b/drivers/scsi/esp_scsi.c @@ -1449,9 +1449,6 @@ static void esp_msgin_sdtr(struct esp *esp, struct esp_target_data *tp) if (offset > 15) goto do_reject; - if (esp->flags & ESP_FLAG_DISABLE_SYNC) - offset = 0; - if (offset) { int one_clock; @@ -2405,12 +2402,6 @@ static int esp_slave_configure(struct scsi_device *dev) struct esp_target_data *tp = &esp->target[dev->id]; int goal_tags, queue_depth; - if (esp->flags & ESP_FLAG_DISABLE_SYNC) { - /* Bypass async domain validation */ - dev->ppr = 0; - dev->sdtr = 0; - } - goal_tags = 0; if (dev->tagged_supported) { @@ -2660,7 +2651,10 @@ static void esp_set_offset(struct scsi_target *target, int offset) struct esp *esp = shost_priv(host); struct esp_target_data *tp = &esp->target[target->id]; - tp->nego_goal_offset = offset; + if (esp->flags & ESP_FLAG_DISABLE_SYNC) + tp->nego_goal_offset = 0; + else + tp->nego_goal_offset = offset; tp->flags |= ESP_TGT_CHECK_NEGO; } diff --git a/drivers/scsi/mac_esp.c b/drivers/scsi/mac_esp.c index c24e86f..dd808ae 100644 --- a/drivers/scsi/mac_esp.c +++ b/drivers/scsi/mac_esp.c @@ -22,7 +22,6 @@ #include #include - #include #include @@ -279,24 +278,27 @@ static void mac_esp_send_pdma_cmd(struct esp *esp, u32 addr, u32 esp_count, * Programmed IO routines follow. */ -static inline int mac_esp_wait_for_fifo(struct esp *esp) +static inline unsigned int mac_esp_wait_for_fifo(struct esp *esp) { int i = 500000; do { - if (esp_read8(ESP_FFLAGS) & ESP_FF_FBYTES) - return 0; + unsigned int fbytes = esp_read8(ESP_FFLAGS) & ESP_FF_FBYTES; + + if (fbytes) + return fbytes; udelay(2); } while (--i); printk(KERN_ERR PFX "FIFO is empty (sreg %02x)\n", esp_read8(ESP_STATUS)); - return 1; + return 0; } static inline int mac_esp_wait_for_intr(struct esp *esp) { + struct mac_esp_priv *mep = MAC_ESP_GET_PRIV(esp); int i = 500000; do { @@ -308,6 +310,7 @@ static inline int mac_esp_wait_for_intr(struct esp *esp) } while (--i); printk(KERN_ERR PFX "IRQ timeout (sreg %02x)\n", esp->sreg); + mep->error = 1; return 1; } @@ -347,11 +350,10 @@ static inline int mac_esp_wait_for_intr(struct esp *esp) static void mac_esp_send_pio_cmd(struct esp *esp, u32 addr, u32 esp_count, u32 dma_count, int write, u8 cmd) { - unsigned long flags; struct mac_esp_priv *mep = MAC_ESP_GET_PRIV(esp); u8 *fifo = esp->regs + ESP_FDATA * 16; - local_irq_save(flags); + disable_irq(esp->host->irq); cmd &= ~ESP_CMD_DMA; mep->error = 0; @@ -359,11 +361,35 @@ static void mac_esp_send_pio_cmd(struct esp *esp, u32 addr, u32 esp_count, if (write) { scsi_esp_cmd(esp, cmd); - if (!mac_esp_wait_for_intr(esp)) { - if (mac_esp_wait_for_fifo(esp)) - esp_count = 0; - } else { - esp_count = 0; + while (1) { + unsigned int n; + + n = mac_esp_wait_for_fifo(esp); + if (!n) + break; + + if (n > esp_count) + n = esp_count; + esp_count -= n; + + MAC_ESP_PIO_LOOP("%2@,%0@+", n); + + if (!esp_count) + break; + + if (mac_esp_wait_for_intr(esp)) + break; + + if (((esp->sreg & ESP_STAT_PMASK) != ESP_DIP) && + ((esp->sreg & ESP_STAT_PMASK) != ESP_MIP)) + break; + + esp->ireg = esp_read8(ESP_INTRPT); + if ((esp->ireg & (ESP_INTR_DC | ESP_INTR_BSERV)) != + ESP_INTR_BSERV) + break; + + scsi_esp_cmd(esp, ESP_CMD_TI); } } else { scsi_esp_cmd(esp, ESP_CMD_FLUSH); @@ -374,47 +400,24 @@ static void mac_esp_send_pio_cmd(struct esp *esp, u32 addr, u32 esp_count, MAC_ESP_PIO_LOOP("%0@+,%2@", esp_count); scsi_esp_cmd(esp, cmd); - } - - while (esp_count) { - unsigned int n; - - if (mac_esp_wait_for_intr(esp)) { - mep->error = 1; - break; - } - - if (esp->sreg & ESP_STAT_SPAM) { - printk(KERN_ERR PFX "gross error\n"); - mep->error = 1; - break; - } - n = esp_read8(ESP_FFLAGS) & ESP_FF_FBYTES; - - if (write) { - if (n > esp_count) - n = esp_count; - esp_count -= n; - - MAC_ESP_PIO_LOOP("%2@,%0@+", n); + while (esp_count) { + unsigned int n; - if ((esp->sreg & ESP_STAT_PMASK) == ESP_STATP) + if (mac_esp_wait_for_intr(esp)) break; - if (esp_count) { - esp->ireg = esp_read8(ESP_INTRPT); - if (esp->ireg & ESP_INTR_DC) - break; + if (((esp->sreg & ESP_STAT_PMASK) != ESP_DOP) && + ((esp->sreg & ESP_STAT_PMASK) != ESP_MOP)) + break; - scsi_esp_cmd(esp, ESP_CMD_TI); - } - } else { esp->ireg = esp_read8(ESP_INTRPT); - if (esp->ireg & ESP_INTR_DC) + if ((esp->ireg & (ESP_INTR_DC | ESP_INTR_BSERV)) != + ESP_INTR_BSERV) break; - n = MAC_ESP_FIFO_SIZE - n; + n = MAC_ESP_FIFO_SIZE - + (esp_read8(ESP_FFLAGS) & ESP_FF_FBYTES); if (n > esp_count) n = esp_count; @@ -429,7 +432,7 @@ static void mac_esp_send_pio_cmd(struct esp *esp, u32 addr, u32 esp_count, } } - local_irq_restore(flags); + enable_irq(esp->host->irq); } static int mac_esp_irq_pending(struct esp *esp) -- cgit v0.10.2 From 65f89c2396aa113a06fe7e2f6ba46f0712cb4806 Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 19:01:13 +0530 Subject: [SCSI] mptfusion: Added MPI_SCSIIO_CONTROL_HEADOFQ priority There is a 'ioprio' field in the BIO and the Request structure. check this priority field and set MPI_SCSIIO_CONTROL_HEADOFQ to pass down I/O priority. An enhancement to the LSI Disk Array Controller firmware is being developed to look at the Head Of Queue bit to allow I/Os with the HOQ bit set to be processed before I/Os which do not have the HOQ bit set. In order to set the HOQ bit, the mpt fusion driver needs to look at the 'ioprio' field in the request structure associated with the scsi command. Signed-off-by: Kashyap Desai Signed-off-by: James Bottomley diff --git a/drivers/message/fusion/mptscsih.c b/drivers/message/fusion/mptscsih.c index 5775275..8128d30 100644 --- a/drivers/message/fusion/mptscsih.c +++ b/drivers/message/fusion/mptscsih.c @@ -1438,9 +1438,14 @@ mptscsih_qcmd(struct scsi_cmnd *SCpnt, void (*done)(struct scsi_cmnd *)) && (vdevice->vtarget->tflags & MPT_TARGET_FLAGS_Q_YES) && (SCpnt->device->tagged_supported)) { scsictl = scsidir | MPI_SCSIIO_CONTROL_SIMPLEQ; - } else { + if (SCpnt->request && SCpnt->request->ioprio) { + if (((SCpnt->request->ioprio & 0x7) == 1) || + !(SCpnt->request->ioprio & 0x7)) + scsictl |= MPI_SCSIIO_CONTROL_HEADOFQ; + } + } else scsictl = scsidir | MPI_SCSIIO_CONTROL_UNTAGGED; - } + /* Use the above information to set up the message frame */ -- cgit v0.10.2 From e0f553ab58f478321717100e44b28f765bd2a045 Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 19:01:58 +0530 Subject: [SCSI] mptfusion: Added sysfs expander manufacture information at the time of expander add. Added new function mptsas_exp_manufacture_info, which will obtain the REPORT_MANUFACTURING, and fill the details into the sas_expander_device object when the expander port is created. Signed-off-by: Kashyap Desai Signed-off-by: James Bottomley diff --git a/drivers/message/fusion/mptsas.c b/drivers/message/fusion/mptsas.c index 83873e3..1b7f550 100644 --- a/drivers/message/fusion/mptsas.c +++ b/drivers/message/fusion/mptsas.c @@ -2686,6 +2686,187 @@ mptsas_sas_expander_pg1(MPT_ADAPTER *ioc, struct mptsas_phyinfo *phy_info, return error; } +struct rep_manu_request{ + u8 smp_frame_type; + u8 function; + u8 reserved; + u8 request_length; +}; + +struct rep_manu_reply{ + u8 smp_frame_type; /* 0x41 */ + u8 function; /* 0x01 */ + u8 function_result; + u8 response_length; + u16 expander_change_count; + u8 reserved0[2]; + u8 sas_format:1; + u8 reserved1:7; + u8 reserved2[3]; + u8 vendor_id[SAS_EXPANDER_VENDOR_ID_LEN]; + u8 product_id[SAS_EXPANDER_PRODUCT_ID_LEN]; + u8 product_rev[SAS_EXPANDER_PRODUCT_REV_LEN]; + u8 component_vendor_id[SAS_EXPANDER_COMPONENT_VENDOR_ID_LEN]; + u16 component_id; + u8 component_revision_id; + u8 reserved3; + u8 vendor_specific[8]; +}; + +/** + * mptsas_exp_repmanufacture_info - + * @ioc: per adapter object + * @sas_address: expander sas address + * @edev: the sas_expander_device object + * + * Fills in the sas_expander_device object when SMP port is created. + * + * Returns 0 for success, non-zero for failure. + */ +static int +mptsas_exp_repmanufacture_info(MPT_ADAPTER *ioc, + u64 sas_address, struct sas_expander_device *edev) +{ + MPT_FRAME_HDR *mf; + SmpPassthroughRequest_t *smpreq; + SmpPassthroughReply_t *smprep; + struct rep_manu_reply *manufacture_reply; + struct rep_manu_request *manufacture_request; + int ret; + int flagsLength; + unsigned long timeleft; + char *psge; + unsigned long flags; + void *data_out = NULL; + dma_addr_t data_out_dma = 0; + u32 sz; + + spin_lock_irqsave(&ioc->taskmgmt_lock, flags); + if (ioc->ioc_reset_in_progress) { + spin_unlock_irqrestore(&ioc->taskmgmt_lock, flags); + printk(MYIOC_s_INFO_FMT "%s: host reset in progress!\n", + __func__, ioc->name); + return -EFAULT; + } + spin_unlock_irqrestore(&ioc->taskmgmt_lock, flags); + + ret = mutex_lock_interruptible(&ioc->sas_mgmt.mutex); + if (ret) + goto out; + + mf = mpt_get_msg_frame(mptsasMgmtCtx, ioc); + if (!mf) { + ret = -ENOMEM; + goto out_unlock; + } + + smpreq = (SmpPassthroughRequest_t *)mf; + memset(smpreq, 0, sizeof(*smpreq)); + + sz = sizeof(struct rep_manu_request) + sizeof(struct rep_manu_reply); + + data_out = pci_alloc_consistent(ioc->pcidev, sz, &data_out_dma); + if (!data_out) { + printk(KERN_ERR "Memory allocation failure at %s:%d/%s()!\n", + __FILE__, __LINE__, __func__); + ret = -ENOMEM; + goto put_mf; + } + + manufacture_request = data_out; + manufacture_request->smp_frame_type = 0x40; + manufacture_request->function = 1; + manufacture_request->reserved = 0; + manufacture_request->request_length = 0; + + smpreq->Function = MPI_FUNCTION_SMP_PASSTHROUGH; + smpreq->PhysicalPort = 0xFF; + *((u64 *)&smpreq->SASAddress) = cpu_to_le64(sas_address); + smpreq->RequestDataLength = sizeof(struct rep_manu_request); + + psge = (char *) + (((int *) mf) + (offsetof(SmpPassthroughRequest_t, SGL) / 4)); + + flagsLength = MPI_SGE_FLAGS_SIMPLE_ELEMENT | + MPI_SGE_FLAGS_SYSTEM_ADDRESS | + MPI_SGE_FLAGS_HOST_TO_IOC | + MPI_SGE_FLAGS_END_OF_BUFFER; + flagsLength = flagsLength << MPI_SGE_FLAGS_SHIFT; + flagsLength |= sizeof(struct rep_manu_request); + + ioc->add_sge(psge, flagsLength, data_out_dma); + psge += ioc->SGE_size; + + flagsLength = MPI_SGE_FLAGS_SIMPLE_ELEMENT | + MPI_SGE_FLAGS_SYSTEM_ADDRESS | + MPI_SGE_FLAGS_IOC_TO_HOST | + MPI_SGE_FLAGS_END_OF_BUFFER; + flagsLength = flagsLength << MPI_SGE_FLAGS_SHIFT; + flagsLength |= sizeof(struct rep_manu_reply); + ioc->add_sge(psge, flagsLength, data_out_dma + + sizeof(struct rep_manu_request)); + + INITIALIZE_MGMT_STATUS(ioc->sas_mgmt.status) + mpt_put_msg_frame(mptsasMgmtCtx, ioc, mf); + + timeleft = wait_for_completion_timeout(&ioc->sas_mgmt.done, 10 * HZ); + if (!(ioc->sas_mgmt.status & MPT_MGMT_STATUS_COMMAND_GOOD)) { + ret = -ETIME; + mpt_free_msg_frame(ioc, mf); + mf = NULL; + if (ioc->sas_mgmt.status & MPT_MGMT_STATUS_DID_IOCRESET) + goto out_free; + if (!timeleft) + mpt_HardResetHandler(ioc, CAN_SLEEP); + goto out_free; + } + + mf = NULL; + + if (ioc->sas_mgmt.status & MPT_MGMT_STATUS_RF_VALID) { + u8 *tmp; + + smprep = (SmpPassthroughReply_t *)ioc->sas_mgmt.reply; + if (le16_to_cpu(smprep->ResponseDataLength) != + sizeof(struct rep_manu_reply)) + goto out_free; + + manufacture_reply = data_out + sizeof(struct rep_manu_request); + strncpy(edev->vendor_id, manufacture_reply->vendor_id, + SAS_EXPANDER_VENDOR_ID_LEN); + strncpy(edev->product_id, manufacture_reply->product_id, + SAS_EXPANDER_PRODUCT_ID_LEN); + strncpy(edev->product_rev, manufacture_reply->product_rev, + SAS_EXPANDER_PRODUCT_REV_LEN); + edev->level = manufacture_reply->sas_format; + if (manufacture_reply->sas_format) { + strncpy(edev->component_vendor_id, + manufacture_reply->component_vendor_id, + SAS_EXPANDER_COMPONENT_VENDOR_ID_LEN); + tmp = (u8 *)&manufacture_reply->component_id; + edev->component_id = tmp[0] << 8 | tmp[1]; + edev->component_revision_id = + manufacture_reply->component_revision_id; + } + } else { + printk(MYIOC_s_ERR_FMT + "%s: smp passthru reply failed to be returned\n", + ioc->name, __func__); + ret = -ENXIO; + } +out_free: + if (data_out_dma) + pci_free_consistent(ioc->pcidev, sz, data_out, data_out_dma); +put_mf: + if (mf) + mpt_free_msg_frame(ioc, mf); +out_unlock: + CLEAR_MGMT_STATUS(ioc->sas_mgmt.status) + mutex_unlock(&ioc->sas_mgmt.mutex); +out: + return ret; + } + static void mptsas_parse_device_info(struct sas_identify *identify, struct mptsas_devinfo *device_info) @@ -2967,6 +3148,11 @@ static int mptsas_probe_one_phy(struct device *dev, goto out; } mptsas_set_rphy(ioc, phy_info, rphy); + if (identify.device_type == SAS_EDGE_EXPANDER_DEVICE || + identify.device_type == SAS_FANOUT_EXPANDER_DEVICE) + mptsas_exp_repmanufacture_info(ioc, + identify.sas_address, + rphy_to_expander_device(rphy)); } out: -- cgit v0.10.2 From 64e155adc250ab68255b761c1faa77799aa1e41a Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 19:02:29 +0530 Subject: [SCSI] mptfusion: block device when target is being removed by FW Add support to set the sdev state to SDEV_BLOCK during device removal to stop IOs comming to the deleting driver immediately. Signed-off-by: Kashyap Desai Signed-off-by: James Bottomley diff --git a/drivers/message/fusion/mptsas.c b/drivers/message/fusion/mptsas.c index 1b7f550..c20bbe4 100644 --- a/drivers/message/fusion/mptsas.c +++ b/drivers/message/fusion/mptsas.c @@ -1075,6 +1075,19 @@ mptsas_target_reset(MPT_ADAPTER *ioc, u8 channel, u8 id) return 0; } +static void +mptsas_block_io_sdev(struct scsi_device *sdev, void *data) +{ + scsi_device_set_state(sdev, SDEV_BLOCK); +} + +static void +mptsas_block_io_starget(struct scsi_target *starget) +{ + if (starget) + starget_for_each_device(starget, NULL, mptsas_block_io_sdev); +} + /** * mptsas_target_reset_queue * @@ -1098,10 +1111,11 @@ mptsas_target_reset_queue(MPT_ADAPTER *ioc, id = sas_event_data->TargetID; channel = sas_event_data->Bus; - if (!(vtarget = mptsas_find_vtarget(ioc, channel, id))) - return; - - vtarget->deleted = 1; /* block IO */ + vtarget = mptsas_find_vtarget(ioc, channel, id); + if (vtarget) { + mptsas_block_io_starget(vtarget->starget); + vtarget->deleted = 1; /* block IO */ + } target_reset_list = kzalloc(sizeof(struct mptsas_target_reset_event), GFP_ATOMIC); @@ -1868,7 +1882,8 @@ mptsas_qcmd(struct scsi_cmnd *SCpnt, void (*done)(struct scsi_cmnd *)) if (ioc->sas_discovery_quiesce_io) return SCSI_MLQUEUE_HOST_BUSY; -// scsi_print_command(SCpnt); + if (ioc->debug_level & MPT_DEBUG_SCSI) + scsi_print_command(SCpnt); return mptscsih_qcmd(SCpnt,done); } -- cgit v0.10.2 From de81562f2f9852a1f2c609ede1e26c07ed457c60 Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 19:02:59 +0530 Subject: [SCSI] mptfusion: corrected if condition check for SCSIIO and PASSTHROUGH commands Modified the function type check to verify it is not MPI_FUNCTION_RAID_SCSI_IO_PASSTHROUGH or MPI_FUNCTION_SCSI_IO_REQUEST. Signed-off-by: Kashyap Desai Signed-off-by: James Bottomley diff --git a/drivers/message/fusion/mptctl.c b/drivers/message/fusion/mptctl.c index 352acd0..caa8f56 100644 --- a/drivers/message/fusion/mptctl.c +++ b/drivers/message/fusion/mptctl.c @@ -360,8 +360,8 @@ static int mptctl_bus_reset(MPT_ADAPTER *ioc, u8 function) u16 iocstatus; /* bus reset is only good for SCSI IO, RAID PASSTHRU */ - if (!(function == MPI_FUNCTION_RAID_SCSI_IO_PASSTHROUGH) || - (function == MPI_FUNCTION_SCSI_IO_REQUEST)) { + if (!(function == MPI_FUNCTION_RAID_SCSI_IO_PASSTHROUGH || + function == MPI_FUNCTION_SCSI_IO_REQUEST)) { dtmprintk(ioc, printk(MYIOC_s_WARN_FMT "TaskMgmt, not SCSI_IO!!\n", ioc->name)); return -EPERM; -- cgit v0.10.2 From 4ffd005a4dedf536330756742595ec904f740f48 Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 19:04:49 +0530 Subject: [SCSI] mptfusion: Bump version 03.04.14 Version upgrade to 3.04.14. Signed-off-by: Kashyap Desai Signed-off-by: James Bottomley diff --git a/drivers/message/fusion/mptbase.h b/drivers/message/fusion/mptbase.h index b494867..9718c8f 100644 --- a/drivers/message/fusion/mptbase.h +++ b/drivers/message/fusion/mptbase.h @@ -76,8 +76,8 @@ #define COPYRIGHT "Copyright (c) 1999-2008 " MODULEAUTHOR #endif -#define MPT_LINUX_VERSION_COMMON "3.04.13" -#define MPT_LINUX_PACKAGE_NAME "@(#)mptlinux-3.04.13" +#define MPT_LINUX_VERSION_COMMON "3.04.14" +#define MPT_LINUX_PACKAGE_NAME "@(#)mptlinux-3.04.14" #define WHAT_MAGIC_STRING "@" "(" "#" ")" #define show_mptmod_ver(s,ver) \ -- cgit v0.10.2 From 3dbf6c0012d12473461b7485006db373e8192fa5 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Sat, 19 Dec 2009 08:17:27 +0100 Subject: [SCSI] pm8001: Use kzalloc for allocating only one thing Use kzalloc rather than kcalloc(1,...) The semantic patch that makes this change is as follows: (http://coccinelle.lip6.fr/) // @@ @@ - kcalloc(1, + kzalloc( ...) // Signed-off-by: Julia Lawall Acked-by:Jack Wang Signed-off-by: James Bottomley diff --git a/drivers/scsi/pm8001/pm8001_init.c b/drivers/scsi/pm8001/pm8001_init.c index c2f1032..f80c1da8 100644 --- a/drivers/scsi/pm8001/pm8001_init.c +++ b/drivers/scsi/pm8001/pm8001_init.c @@ -654,7 +654,7 @@ static int __devinit pm8001_pci_probe(struct pci_dev *pdev, } chip = &pm8001_chips[ent->driver_data]; SHOST_TO_SAS_HA(shost) = - kcalloc(1, sizeof(struct sas_ha_struct), GFP_KERNEL); + kzalloc(sizeof(struct sas_ha_struct), GFP_KERNEL); if (!SHOST_TO_SAS_HA(shost)) { rc = -ENOMEM; goto err_out_free_host; -- cgit v0.10.2 From 0ed8570ef4d51bb46e2028b2e3c3ddd1f860e193 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Mon, 21 Dec 2009 16:27:55 -0800 Subject: [SCSI] cxgbi3: remove unnecessary NULL test Stanse found that c3cn is poked many times around in cxgb3i_conn_pdu_ready, there is no need to check if it is NULL. Remove the test. Signed-off-by: Jiri Slaby Reviewed-by: Mike Christie Acked-by: Karen Xie Signed-off-by: Andrew Morton Signed-off-by: James Bottomley diff --git a/drivers/scsi/cxgb3i/cxgb3i_pdu.c b/drivers/scsi/cxgb3i/cxgb3i_pdu.c index 1fe3b0f..9c38539 100644 --- a/drivers/scsi/cxgb3i/cxgb3i_pdu.c +++ b/drivers/scsi/cxgb3i/cxgb3i_pdu.c @@ -461,10 +461,8 @@ void cxgb3i_conn_pdu_ready(struct s3_conn *c3cn) skb = skb_peek(&c3cn->receive_queue); } read_unlock(&c3cn->callback_lock); - if (c3cn) { - c3cn->copied_seq += read; - cxgb3i_c3cn_rx_credits(c3cn, read); - } + c3cn->copied_seq += read; + cxgb3i_c3cn_rx_credits(c3cn, read); conn->rxdata_octets += read; if (err) { -- cgit v0.10.2 From 5d7ebb9c7a04d29efce1099024944dfd94d9f63c Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Mon, 28 Dec 2009 20:08:39 +0200 Subject: [SCSI] FlashPoint: fix off by one tests The check on MAX_SCSI_TAR should be >= instead of > or we could go past the end of the array. Joe Eykholt aslo correctly points out that the check on MAX_LUN should be >= as well. That matches with how it is used in the rest of the file. Signed-off-by: Dan Carpenter Signed-off-by: James Bottomley diff --git a/drivers/scsi/FlashPoint.c b/drivers/scsi/FlashPoint.c index b898d38..e40cdfb 100644 --- a/drivers/scsi/FlashPoint.c +++ b/drivers/scsi/FlashPoint.c @@ -3924,7 +3924,7 @@ static void FPT_sinits(struct sccb *p_sccb, unsigned char p_card) { struct sccb_mgr_tar_info *currTar_Info; - if ((p_sccb->TargID > MAX_SCSI_TAR) || (p_sccb->Lun > MAX_LUN)) { + if ((p_sccb->TargID >= MAX_SCSI_TAR) || (p_sccb->Lun >= MAX_LUN)) { return; } currTar_Info = &FPT_sccbMgrTbl[p_card][p_sccb->TargID]; -- cgit v0.10.2 From e7efe5932b1d3916c79326a4221693ea90a900e2 Mon Sep 17 00:00:00 2001 From: Douglas Gilbert Date: Sun, 3 Jan 2010 13:51:15 -0500 Subject: [SCSI] skip sense logging for some ATA PASS-THROUGH cdbs Further to the lsml thread titled: "does scsi_io_completion need to dump sense data for ata pass through (ck_cond = 1) ?" This is a patch to skip logging when the sense data is associated with a SENSE_KEY of "RECOVERED_ERROR" and the additional sense code is "ATA PASS-THROUGH INFORMATION AVAILABLE". This only occurs with the SAT ATA PASS-THROUGH commands when CK_COND=1 (in the cdb). It indicates that the sense data contains ATA registers. Smartmontools uses such commands on ATA disks connected via SAT. Periodic checks such as those done by smartd cause nuisance entries into logs that are: - neither errors nor warnings - pointless unless the cdb that caused them are also logged Signed-off-by: Douglas Gilbert Signed-off-by: James Bottomley diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c index c664242..5697709 100644 --- a/drivers/scsi/scsi_lib.c +++ b/drivers/scsi/scsi_lib.c @@ -773,8 +773,14 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) * we already took a copy of the original into rq->errors which * is what gets returned to the user */ - if (sense_valid && sshdr.sense_key == RECOVERED_ERROR) { - if (!(req->cmd_flags & REQ_QUIET)) + if (sense_valid && (sshdr.sense_key == RECOVERED_ERROR)) { + /* if ATA PASS-THROUGH INFORMATION AVAILABLE skip + * print since caller wants ATA registers. Only occurs on + * SCSI ATA PASS_THROUGH commands when CK_COND=1 + */ + if ((sshdr.asc == 0x0) && (sshdr.ascq == 0x1d)) + ; + else if (!(req->cmd_flags & REQ_QUIET)) scsi_print_sense("", cmd); result = 0; /* BLOCK_PC may have set error */ -- cgit v0.10.2 From 1fe6dbf4d0afba52ad0249f398e6296a1433a004 Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Mon, 4 Jan 2010 10:19:34 -0500 Subject: [SCSI] gdth: Convert to use regular kernel types. converted using this script.. perl -p -i -e 's|ulong32|u32|g' drivers/scsi/gdth* perl -p -i -e 's|ulong64|u64|g' drivers/scsi/gdth* perl -p -i -e 's|ushort|u16|g' drivers/scsi/gdth* perl -p -i -e 's|unchar|u8|g' drivers/scsi/gdth* perl -p -i -e 's|ulong|unsigned long|g' drivers/scsi/gdth* perl -p -i -e 's|PACKED|__attribute__((packed))|g' drivers/scsi/gdth* sha1sum of the generated code was identical before and after. Signed-off-by: Dave Jones Signed-off-by: James Bottomley diff --git a/drivers/scsi/gdth.c b/drivers/scsi/gdth.c index 9e8fce0..ba3c94c 100644 --- a/drivers/scsi/gdth.c +++ b/drivers/scsi/gdth.c @@ -140,40 +140,40 @@ #include "gdth.h" static void gdth_delay(int milliseconds); -static void gdth_eval_mapping(ulong32 size, ulong32 *cyls, int *heads, int *secs); +static void gdth_eval_mapping(u32 size, u32 *cyls, int *heads, int *secs); static irqreturn_t gdth_interrupt(int irq, void *dev_id); static irqreturn_t __gdth_interrupt(gdth_ha_str *ha, int gdth_from_wait, int* pIndex); -static int gdth_sync_event(gdth_ha_str *ha, int service, unchar index, +static int gdth_sync_event(gdth_ha_str *ha, int service, u8 index, Scsi_Cmnd *scp); static int gdth_async_event(gdth_ha_str *ha); static void gdth_log_event(gdth_evt_data *dvr, char *buffer); -static void gdth_putq(gdth_ha_str *ha, Scsi_Cmnd *scp, unchar priority); +static void gdth_putq(gdth_ha_str *ha, Scsi_Cmnd *scp, u8 priority); static void gdth_next(gdth_ha_str *ha); -static int gdth_fill_raw_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, unchar b); +static int gdth_fill_raw_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, u8 b); static int gdth_special_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp); -static gdth_evt_str *gdth_store_event(gdth_ha_str *ha, ushort source, - ushort idx, gdth_evt_data *evt); +static gdth_evt_str *gdth_store_event(gdth_ha_str *ha, u16 source, + u16 idx, gdth_evt_data *evt); static int gdth_read_event(gdth_ha_str *ha, int handle, gdth_evt_str *estr); -static void gdth_readapp_event(gdth_ha_str *ha, unchar application, +static void gdth_readapp_event(gdth_ha_str *ha, u8 application, gdth_evt_str *estr); static void gdth_clear_events(void); static void gdth_copy_internal_data(gdth_ha_str *ha, Scsi_Cmnd *scp, - char *buffer, ushort count); + char *buffer, u16 count); static int gdth_internal_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp); -static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, ushort hdrive); +static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, u16 hdrive); static void gdth_enable_int(gdth_ha_str *ha); static int gdth_test_busy(gdth_ha_str *ha); static int gdth_get_cmd_index(gdth_ha_str *ha); static void gdth_release_event(gdth_ha_str *ha); -static int gdth_wait(gdth_ha_str *ha, int index,ulong32 time); -static int gdth_internal_cmd(gdth_ha_str *ha, unchar service, ushort opcode, - ulong32 p1, ulong64 p2,ulong64 p3); +static int gdth_wait(gdth_ha_str *ha, int index,u32 time); +static int gdth_internal_cmd(gdth_ha_str *ha, u8 service, u16 opcode, + u32 p1, u64 p2,u64 p3); static int gdth_search_drives(gdth_ha_str *ha); -static int gdth_analyse_hdrive(gdth_ha_str *ha, ushort hdrive); +static int gdth_analyse_hdrive(gdth_ha_str *ha, u16 hdrive); static const char *gdth_ctr_name(gdth_ha_str *ha); @@ -189,7 +189,7 @@ static int __gdth_queuecommand(gdth_ha_str *ha, struct scsi_cmnd *scp, static void gdth_scsi_done(struct scsi_cmnd *scp); #ifdef DEBUG_GDTH -static unchar DebugState = DEBUG_GDTH; +static u8 DebugState = DEBUG_GDTH; #ifdef __SERIAL__ #define MAX_SERBUF 160 @@ -270,30 +270,30 @@ static int ser_printk(const char *fmt, ...) #endif #ifdef GDTH_STATISTICS -static ulong32 max_rq=0, max_index=0, max_sg=0; +static u32 max_rq=0, max_index=0, max_sg=0; #ifdef INT_COAL -static ulong32 max_int_coal=0; +static u32 max_int_coal=0; #endif -static ulong32 act_ints=0, act_ios=0, act_stats=0, act_rq=0; +static u32 act_ints=0, act_ios=0, act_stats=0, act_rq=0; static struct timer_list gdth_timer; #endif -#define PTR2USHORT(a) (ushort)(ulong)(a) +#define PTR2USHORT(a) (u16)(unsigned long)(a) #define GDTOFFSOF(a,b) (size_t)&(((a*)0)->b) #define INDEX_OK(i,t) ((i)(a)->virt_bus ? (b-1):(b)) #ifdef CONFIG_ISA -static unchar gdth_drq_tab[4] = {5,6,7,7}; /* DRQ table */ +static u8 gdth_drq_tab[4] = {5,6,7,7}; /* DRQ table */ #endif #if defined(CONFIG_EISA) || defined(CONFIG_ISA) -static unchar gdth_irq_tab[6] = {0,10,11,12,14,0}; /* IRQ table */ +static u8 gdth_irq_tab[6] = {0,10,11,12,14,0}; /* IRQ table */ #endif -static unchar gdth_polling; /* polling if TRUE */ +static u8 gdth_polling; /* polling if TRUE */ static int gdth_ctr_count = 0; /* controller count */ static LIST_HEAD(gdth_instances); /* controller list */ -static unchar gdth_write_through = FALSE; /* write through */ +static u8 gdth_write_through = FALSE; /* write through */ static gdth_evt_str ebuffer[MAX_EVENTS]; /* event buffer */ static int elastidx; static int eoldidx; @@ -303,7 +303,7 @@ static int major; #define DOU 2 /* OUT data direction */ #define DNO DIN /* no data transfer */ #define DUN DIN /* unknown data direction */ -static unchar gdth_direction_tab[0x100] = { +static u8 gdth_direction_tab[0x100] = { DNO,DNO,DIN,DIN,DOU,DIN,DIN,DOU,DIN,DUN,DOU,DOU,DUN,DUN,DUN,DIN, DNO,DIN,DIN,DOU,DIN,DOU,DNO,DNO,DOU,DNO,DIN,DNO,DIN,DOU,DNO,DUN, DIN,DUN,DIN,DUN,DOU,DIN,DUN,DUN,DIN,DIN,DOU,DNO,DUN,DIN,DOU,DOU, @@ -390,7 +390,7 @@ static gdth_ha_str *gdth_find_ha(int hanum) static struct gdth_cmndinfo *gdth_get_cmndinfo(gdth_ha_str *ha) { struct gdth_cmndinfo *priv = NULL; - ulong flags; + unsigned long flags; int i; spin_lock_irqsave(&ha->smp_lock, flags); @@ -493,7 +493,7 @@ int gdth_execute(struct Scsi_Host *shost, gdth_cmd_str *gdtcmd, char *cmnd, return rval; } -static void gdth_eval_mapping(ulong32 size, ulong32 *cyls, int *heads, int *secs) +static void gdth_eval_mapping(u32 size, u32 *cyls, int *heads, int *secs) { *cyls = size /HEADS/SECS; if (*cyls <= MAXCYLS) { @@ -514,9 +514,9 @@ static void gdth_eval_mapping(ulong32 size, ulong32 *cyls, int *heads, int *secs /* controller search and initialization functions */ #ifdef CONFIG_EISA -static int __init gdth_search_eisa(ushort eisa_adr) +static int __init gdth_search_eisa(u16 eisa_adr) { - ulong32 id; + u32 id; TRACE(("gdth_search_eisa() adr. %x\n",eisa_adr)); id = inl(eisa_adr+ID0REG); @@ -533,13 +533,13 @@ static int __init gdth_search_eisa(ushort eisa_adr) #endif /* CONFIG_EISA */ #ifdef CONFIG_ISA -static int __init gdth_search_isa(ulong32 bios_adr) +static int __init gdth_search_isa(u32 bios_adr) { void __iomem *addr; - ulong32 id; + u32 id; TRACE(("gdth_search_isa() bios adr. %x\n",bios_adr)); - if ((addr = ioremap(bios_adr+BIOS_ID_OFFS, sizeof(ulong32))) != NULL) { + if ((addr = ioremap(bios_adr+BIOS_ID_OFFS, sizeof(u32))) != NULL) { id = readl(addr); iounmap(addr); if (id == GDT2_ID) /* GDT2000 */ @@ -551,7 +551,7 @@ static int __init gdth_search_isa(ulong32 bios_adr) #ifdef CONFIG_PCI -static bool gdth_search_vortex(ushort device) +static bool gdth_search_vortex(u16 device) { if (device <= PCI_DEVICE_ID_VORTEX_GDT6555) return true; @@ -603,9 +603,9 @@ static void __devexit gdth_pci_remove_one(struct pci_dev *pdev) static int __devinit gdth_pci_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) { - ushort vendor = pdev->vendor; - ushort device = pdev->device; - ulong base0, base1, base2; + u16 vendor = pdev->vendor; + u16 device = pdev->device; + unsigned long base0, base1, base2; int rc; gdth_pci_str gdth_pcistr; gdth_ha_str *ha = NULL; @@ -658,10 +658,10 @@ static int __devinit gdth_pci_init_one(struct pci_dev *pdev, #endif /* CONFIG_PCI */ #ifdef CONFIG_EISA -static int __init gdth_init_eisa(ushort eisa_adr,gdth_ha_str *ha) +static int __init gdth_init_eisa(u16 eisa_adr,gdth_ha_str *ha) { - ulong32 retries,id; - unchar prot_ver,eisacf,i,irq_found; + u32 retries,id; + u8 prot_ver,eisacf,i,irq_found; TRACE(("gdth_init_eisa() adr. %x\n",eisa_adr)); @@ -688,7 +688,7 @@ static int __init gdth_init_eisa(ushort eisa_adr,gdth_ha_str *ha) return 0; } ha->bmic = eisa_adr; - ha->brd_phys = (ulong32)eisa_adr >> 12; + ha->brd_phys = (u32)eisa_adr >> 12; outl(0,eisa_adr+MAILBOXREG); outl(0,eisa_adr+MAILBOXREG+4); @@ -752,12 +752,12 @@ static int __init gdth_init_eisa(ushort eisa_adr,gdth_ha_str *ha) #endif /* CONFIG_EISA */ #ifdef CONFIG_ISA -static int __init gdth_init_isa(ulong32 bios_adr,gdth_ha_str *ha) +static int __init gdth_init_isa(u32 bios_adr,gdth_ha_str *ha) { register gdt2_dpram_str __iomem *dp2_ptr; int i; - unchar irq_drq,prot_ver; - ulong32 retries; + u8 irq_drq,prot_ver; + u32 retries; TRACE(("gdth_init_isa() bios adr. %x\n",bios_adr)); @@ -812,7 +812,7 @@ static int __init gdth_init_isa(ulong32 bios_adr,gdth_ha_str *ha) } gdth_delay(1); } - prot_ver = (unchar)readl(&dp2_ptr->u.ic.S_Info[0]); + prot_ver = (u8)readl(&dp2_ptr->u.ic.S_Info[0]); writeb(0, &dp2_ptr->u.ic.Status); writeb(0xff, &dp2_ptr->io.irqdel); if (prot_ver != PROTOCOL_VERSION) { @@ -859,9 +859,9 @@ static int __devinit gdth_init_pci(struct pci_dev *pdev, gdth_pci_str *pcistr, register gdt6_dpram_str __iomem *dp6_ptr; register gdt6c_dpram_str __iomem *dp6c_ptr; register gdt6m_dpram_str __iomem *dp6m_ptr; - ulong32 retries; - unchar prot_ver; - ushort command; + u32 retries; + u8 prot_ver; + u16 command; int i, found = FALSE; TRACE(("gdth_init_pci()\n")); @@ -871,7 +871,7 @@ static int __devinit gdth_init_pci(struct pci_dev *pdev, gdth_pci_str *pcistr, else ha->oem_id = OEM_ID_ICP; ha->brd_phys = (pdev->bus->number << 8) | (pdev->devfn & 0xf8); - ha->stype = (ulong32)pdev->device; + ha->stype = (u32)pdev->device; ha->irq = pdev->irq; ha->pdev = pdev; @@ -891,7 +891,7 @@ static int __devinit gdth_init_pci(struct pci_dev *pdev, gdth_pci_str *pcistr, found = FALSE; for (i = 0xC8000; i < 0xE8000; i += 0x4000) { iounmap(ha->brd); - ha->brd = ioremap(i, sizeof(ushort)); + ha->brd = ioremap(i, sizeof(u16)); if (ha->brd == NULL) { printk("GDT-PCI: Initialization error (DPMEM remap error)\n"); return 0; @@ -947,7 +947,7 @@ static int __devinit gdth_init_pci(struct pci_dev *pdev, gdth_pci_str *pcistr, } gdth_delay(1); } - prot_ver = (unchar)readl(&dp6_ptr->u.ic.S_Info[0]); + prot_ver = (u8)readl(&dp6_ptr->u.ic.S_Info[0]); writeb(0, &dp6_ptr->u.ic.S_Status); writeb(0xff, &dp6_ptr->io.irqdel); if (prot_ver != PROTOCOL_VERSION) { @@ -1000,7 +1000,7 @@ static int __devinit gdth_init_pci(struct pci_dev *pdev, gdth_pci_str *pcistr, found = FALSE; for (i = 0xC8000; i < 0xE8000; i += 0x4000) { iounmap(ha->brd); - ha->brd = ioremap(i, sizeof(ushort)); + ha->brd = ioremap(i, sizeof(u16)); if (ha->brd == NULL) { printk("GDT-PCI: Initialization error (DPMEM remap error)\n"); return 0; @@ -1059,7 +1059,7 @@ static int __devinit gdth_init_pci(struct pci_dev *pdev, gdth_pci_str *pcistr, } gdth_delay(1); } - prot_ver = (unchar)readl(&dp6c_ptr->u.ic.S_Info[0]); + prot_ver = (u8)readl(&dp6c_ptr->u.ic.S_Info[0]); writeb(0, &dp6c_ptr->u.ic.Status); if (prot_ver != PROTOCOL_VERSION) { printk("GDT-PCI: Illegal protocol version\n"); @@ -1128,7 +1128,7 @@ static int __devinit gdth_init_pci(struct pci_dev *pdev, gdth_pci_str *pcistr, found = FALSE; for (i = 0xC8000; i < 0xE8000; i += 0x4000) { iounmap(ha->brd); - ha->brd = ioremap(i, sizeof(ushort)); + ha->brd = ioremap(i, sizeof(u16)); if (ha->brd == NULL) { printk("GDT-PCI: Initialization error (DPMEM remap error)\n"); return 0; @@ -1180,7 +1180,7 @@ static int __devinit gdth_init_pci(struct pci_dev *pdev, gdth_pci_str *pcistr, } gdth_delay(1); } - prot_ver = (unchar)readl(&dp6m_ptr->u.ic.S_Info[0]); + prot_ver = (u8)readl(&dp6m_ptr->u.ic.S_Info[0]); writeb(0, &dp6m_ptr->u.ic.S_Status); if (prot_ver != PROTOCOL_VERSION) { printk("GDT-PCI: Illegal protocol version\n"); @@ -1223,7 +1223,7 @@ static int __devinit gdth_init_pci(struct pci_dev *pdev, gdth_pci_str *pcistr, } gdth_delay(1); } - prot_ver = (unchar)(readl(&dp6m_ptr->u.ic.S_Info[0]) >> 16); + prot_ver = (u8)(readl(&dp6m_ptr->u.ic.S_Info[0]) >> 16); writeb(0, &dp6m_ptr->u.ic.S_Status); if (prot_ver < 0x2b) /* FW < x.43: no 64-bit DMA support */ ha->dma64_support = 0; @@ -1239,7 +1239,7 @@ static int __devinit gdth_init_pci(struct pci_dev *pdev, gdth_pci_str *pcistr, static void __devinit gdth_enable_int(gdth_ha_str *ha) { - ulong flags; + unsigned long flags; gdt2_dpram_str __iomem *dp2_ptr; gdt6_dpram_str __iomem *dp6_ptr; gdt6m_dpram_str __iomem *dp6m_ptr; @@ -1274,14 +1274,14 @@ static void __devinit gdth_enable_int(gdth_ha_str *ha) } /* return IStatus if interrupt was from this card else 0 */ -static unchar gdth_get_status(gdth_ha_str *ha) +static u8 gdth_get_status(gdth_ha_str *ha) { - unchar IStatus = 0; + u8 IStatus = 0; TRACE(("gdth_get_status() irq %d ctr_count %d\n", ha->irq, gdth_ctr_count)); if (ha->type == GDT_EISA) - IStatus = inb((ushort)ha->bmic + EDOORREG); + IStatus = inb((u16)ha->bmic + EDOORREG); else if (ha->type == GDT_ISA) IStatus = readb(&((gdt2_dpram_str __iomem *)ha->brd)->u.ic.Cmd_Index); @@ -1329,7 +1329,7 @@ static int gdth_get_cmd_index(gdth_ha_str *ha) if (ha->cmd_tab[i].cmnd == UNUSED_CMND) { ha->cmd_tab[i].cmnd = ha->pccb->RequestBuffer; ha->cmd_tab[i].service = ha->pccb->Service; - ha->pccb->CommandIndex = (ulong32)i+2; + ha->pccb->CommandIndex = (u32)i+2; return (i+2); } } @@ -1362,7 +1362,7 @@ static void gdth_copy_command(gdth_ha_str *ha) register gdt6c_dpram_str __iomem *dp6c_ptr; gdt6_dpram_str __iomem *dp6_ptr; gdt2_dpram_str __iomem *dp2_ptr; - ushort cp_count,dp_offset,cmd_no; + u16 cp_count,dp_offset,cmd_no; TRACE(("gdth_copy_command() hanum %d\n", ha->hanum)); @@ -1386,28 +1386,28 @@ static void gdth_copy_command(gdth_ha_str *ha) dp2_ptr = ha->brd; writew(dp_offset + DPMEM_COMMAND_OFFSET, &dp2_ptr->u.ic.comm_queue[cmd_no].offset); - writew((ushort)cmd_ptr->Service, + writew((u16)cmd_ptr->Service, &dp2_ptr->u.ic.comm_queue[cmd_no].serv_id); memcpy_toio(&dp2_ptr->u.ic.gdt_dpr_cmd[dp_offset],cmd_ptr,cp_count); } else if (ha->type == GDT_PCI) { dp6_ptr = ha->brd; writew(dp_offset + DPMEM_COMMAND_OFFSET, &dp6_ptr->u.ic.comm_queue[cmd_no].offset); - writew((ushort)cmd_ptr->Service, + writew((u16)cmd_ptr->Service, &dp6_ptr->u.ic.comm_queue[cmd_no].serv_id); memcpy_toio(&dp6_ptr->u.ic.gdt_dpr_cmd[dp_offset],cmd_ptr,cp_count); } else if (ha->type == GDT_PCINEW) { dp6c_ptr = ha->brd; writew(dp_offset + DPMEM_COMMAND_OFFSET, &dp6c_ptr->u.ic.comm_queue[cmd_no].offset); - writew((ushort)cmd_ptr->Service, + writew((u16)cmd_ptr->Service, &dp6c_ptr->u.ic.comm_queue[cmd_no].serv_id); memcpy_toio(&dp6c_ptr->u.ic.gdt_dpr_cmd[dp_offset],cmd_ptr,cp_count); } else if (ha->type == GDT_PCIMPR) { dp6m_ptr = ha->brd; writew(dp_offset + DPMEM_COMMAND_OFFSET, &dp6m_ptr->u.ic.comm_queue[cmd_no].offset); - writew((ushort)cmd_ptr->Service, + writew((u16)cmd_ptr->Service, &dp6m_ptr->u.ic.comm_queue[cmd_no].serv_id); memcpy_toio(&dp6m_ptr->u.ic.gdt_dpr_cmd[dp_offset],cmd_ptr,cp_count); } @@ -1420,14 +1420,14 @@ static void gdth_release_event(gdth_ha_str *ha) #ifdef GDTH_STATISTICS { - ulong32 i,j; + u32 i,j; for (i=0,j=0; jcmd_tab[j].cmnd != UNUSED_CMND) ++i; } if (max_index < i) { max_index = i; - TRACE3(("GDT: max_index = %d\n",(ushort)i)); + TRACE3(("GDT: max_index = %d\n",(u16)i)); } } #endif @@ -1450,7 +1450,7 @@ static void gdth_release_event(gdth_ha_str *ha) } } -static int gdth_wait(gdth_ha_str *ha, int index, ulong32 time) +static int gdth_wait(gdth_ha_str *ha, int index, u32 time) { int answer_found = FALSE; int wait_index = 0; @@ -1476,8 +1476,8 @@ static int gdth_wait(gdth_ha_str *ha, int index, ulong32 time) } -static int gdth_internal_cmd(gdth_ha_str *ha, unchar service, ushort opcode, - ulong32 p1, ulong64 p2, ulong64 p3) +static int gdth_internal_cmd(gdth_ha_str *ha, u8 service, u16 opcode, + u32 p1, u64 p2, u64 p3) { register gdth_cmd_str *cmd_ptr; int retries,index; @@ -1501,35 +1501,35 @@ static int gdth_internal_cmd(gdth_ha_str *ha, unchar service, ushort opcode, if (service == CACHESERVICE) { if (opcode == GDT_IOCTL) { cmd_ptr->u.ioctl.subfunc = p1; - cmd_ptr->u.ioctl.channel = (ulong32)p2; - cmd_ptr->u.ioctl.param_size = (ushort)p3; + cmd_ptr->u.ioctl.channel = (u32)p2; + cmd_ptr->u.ioctl.param_size = (u16)p3; cmd_ptr->u.ioctl.p_param = ha->scratch_phys; } else { if (ha->cache_feat & GDT_64BIT) { - cmd_ptr->u.cache64.DeviceNo = (ushort)p1; + cmd_ptr->u.cache64.DeviceNo = (u16)p1; cmd_ptr->u.cache64.BlockNo = p2; } else { - cmd_ptr->u.cache.DeviceNo = (ushort)p1; - cmd_ptr->u.cache.BlockNo = (ulong32)p2; + cmd_ptr->u.cache.DeviceNo = (u16)p1; + cmd_ptr->u.cache.BlockNo = (u32)p2; } } } else if (service == SCSIRAWSERVICE) { if (ha->raw_feat & GDT_64BIT) { cmd_ptr->u.raw64.direction = p1; - cmd_ptr->u.raw64.bus = (unchar)p2; - cmd_ptr->u.raw64.target = (unchar)p3; - cmd_ptr->u.raw64.lun = (unchar)(p3 >> 8); + cmd_ptr->u.raw64.bus = (u8)p2; + cmd_ptr->u.raw64.target = (u8)p3; + cmd_ptr->u.raw64.lun = (u8)(p3 >> 8); } else { cmd_ptr->u.raw.direction = p1; - cmd_ptr->u.raw.bus = (unchar)p2; - cmd_ptr->u.raw.target = (unchar)p3; - cmd_ptr->u.raw.lun = (unchar)(p3 >> 8); + cmd_ptr->u.raw.bus = (u8)p2; + cmd_ptr->u.raw.target = (u8)p3; + cmd_ptr->u.raw.lun = (u8)(p3 >> 8); } } else if (service == SCREENSERVICE) { if (opcode == GDT_REALTIME) { - *(ulong32 *)&cmd_ptr->u.screen.su.data[0] = p1; - *(ulong32 *)&cmd_ptr->u.screen.su.data[4] = (ulong32)p2; - *(ulong32 *)&cmd_ptr->u.screen.su.data[8] = (ulong32)p3; + *(u32 *)&cmd_ptr->u.screen.su.data[0] = p1; + *(u32 *)&cmd_ptr->u.screen.su.data[4] = (u32)p2; + *(u32 *)&cmd_ptr->u.screen.su.data[8] = (u32)p3; } } ha->cmd_len = sizeof(gdth_cmd_str); @@ -1555,9 +1555,9 @@ static int gdth_internal_cmd(gdth_ha_str *ha, unchar service, ushort opcode, static int __devinit gdth_search_drives(gdth_ha_str *ha) { - ushort cdev_cnt, i; + u16 cdev_cnt, i; int ok; - ulong32 bus_no, drv_cnt, drv_no, j; + u32 bus_no, drv_cnt, drv_no, j; gdth_getch_str *chn; gdth_drlist_str *drl; gdth_iochan_str *ioc; @@ -1570,8 +1570,8 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) #endif #ifdef GDTH_RTC - unchar rtc[12]; - ulong flags; + u8 rtc[12]; + unsigned long flags; #endif TRACE(("gdth_search_drives() hanum %d\n", ha->hanum)); @@ -1584,7 +1584,7 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) if (ok) ha->screen_feat = GDT_64BIT; } - if (force_dma32 || (!ok && ha->status == (ushort)S_NOFUNC)) + if (force_dma32 || (!ok && ha->status == (u16)S_NOFUNC)) ok = gdth_internal_cmd(ha, SCREENSERVICE, GDT_INIT, 0, 0, 0); if (!ok) { printk("GDT-HA %d: Initialization error screen service (code %d)\n", @@ -1609,11 +1609,11 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) rtc[j] = CMOS_READ(j); } while (rtc[0] != CMOS_READ(0)); spin_unlock_irqrestore(&rtc_lock, flags); - TRACE2(("gdth_search_drives(): RTC: %x/%x/%x\n",*(ulong32 *)&rtc[0], - *(ulong32 *)&rtc[4], *(ulong32 *)&rtc[8])); + TRACE2(("gdth_search_drives(): RTC: %x/%x/%x\n",*(u32 *)&rtc[0], + *(u32 *)&rtc[4], *(u32 *)&rtc[8])); /* 3. send to controller firmware */ - gdth_internal_cmd(ha, SCREENSERVICE, GDT_REALTIME, *(ulong32 *)&rtc[0], - *(ulong32 *)&rtc[4], *(ulong32 *)&rtc[8]); + gdth_internal_cmd(ha, SCREENSERVICE, GDT_REALTIME, *(u32 *)&rtc[0], + *(u32 *)&rtc[4], *(u32 *)&rtc[8]); #endif /* unfreeze all IOs */ @@ -1627,7 +1627,7 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) if (ok) ha->cache_feat = GDT_64BIT; } - if (force_dma32 || (!ok && ha->status == (ushort)S_NOFUNC)) + if (force_dma32 || (!ok && ha->status == (u16)S_NOFUNC)) ok = gdth_internal_cmd(ha, CACHESERVICE, GDT_INIT, LINUX_OS, 0, 0); if (!ok) { printk("GDT-HA %d: Initialization error cache service (code %d)\n", @@ -1635,7 +1635,7 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) return 0; } TRACE2(("gdth_search_drives(): CACHESERVICE initialized\n")); - cdev_cnt = (ushort)ha->info; + cdev_cnt = (u16)ha->info; ha->fw_vers = ha->service; #ifdef INT_COAL @@ -1644,7 +1644,7 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) pmod = (gdth_perf_modes *)ha->pscratch; pmod->version = 1; pmod->st_mode = 1; /* enable one status buffer */ - *((ulong64 *)&pmod->st_buff_addr1) = ha->coal_stat_phys; + *((u64 *)&pmod->st_buff_addr1) = ha->coal_stat_phys; pmod->st_buff_indx1 = COALINDEX; pmod->st_buff_addr2 = 0; pmod->st_buff_u_addr2 = 0; @@ -1705,7 +1705,7 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) else ha->bus_id[bus_no] = 0xff; } - ha->bus_cnt = (unchar)bus_no; + ha->bus_cnt = (u8)bus_no; } TRACE2(("gdth_search_drives() %d channels\n",ha->bus_cnt)); @@ -1789,12 +1789,12 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) /* logical drives */ if (gdth_internal_cmd(ha, CACHESERVICE, GDT_IOCTL, CACHE_DRV_CNT, - INVALID_CHANNEL,sizeof(ulong32))) { - drv_cnt = *(ulong32 *)ha->pscratch; + INVALID_CHANNEL,sizeof(u32))) { + drv_cnt = *(u32 *)ha->pscratch; if (gdth_internal_cmd(ha, CACHESERVICE, GDT_IOCTL, CACHE_DRV_LIST, - INVALID_CHANNEL,drv_cnt * sizeof(ulong32))) { + INVALID_CHANNEL,drv_cnt * sizeof(u32))) { for (j = 0; j < drv_cnt; ++j) { - drv_no = ((ulong32 *)ha->pscratch)[j]; + drv_no = ((u32 *)ha->pscratch)[j]; if (drv_no < MAX_LDRIVES) { ha->hdr[drv_no].is_logdrv = TRUE; TRACE2(("Drive %d is log. drive\n",drv_no)); @@ -1838,7 +1838,7 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) if (ok) ha->raw_feat = GDT_64BIT; } - if (force_dma32 || (!ok && ha->status == (ushort)S_NOFUNC)) + if (force_dma32 || (!ok && ha->status == (u16)S_NOFUNC)) ok = gdth_internal_cmd(ha, SCSIRAWSERVICE, GDT_INIT, 0, 0, 0); if (!ok) { printk("GDT-HA %d: Initialization error raw service (code %d)\n", @@ -1854,7 +1854,7 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) if (gdth_internal_cmd(ha, SCSIRAWSERVICE, GDT_GET_FEAT, 0, 0, 0)) { TRACE2(("gdth_search_dr(): get feat RAWSERVICE %d\n", ha->info)); - ha->raw_feat |= (ushort)ha->info; + ha->raw_feat |= (u16)ha->info; } } @@ -1865,7 +1865,7 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) if (gdth_internal_cmd(ha, CACHESERVICE, GDT_GET_FEAT, 0, 0, 0)) { TRACE2(("gdth_search_dr(): get feat CACHESERV. %d\n", ha->info)); - ha->cache_feat |= (ushort)ha->info; + ha->cache_feat |= (u16)ha->info; } } @@ -1923,9 +1923,9 @@ static int __devinit gdth_search_drives(gdth_ha_str *ha) return 1; } -static int gdth_analyse_hdrive(gdth_ha_str *ha, ushort hdrive) +static int gdth_analyse_hdrive(gdth_ha_str *ha, u16 hdrive) { - ulong32 drv_cyls; + u32 drv_cyls; int drv_hds, drv_secs; TRACE(("gdth_analyse_hdrive() hanum %d drive %d\n", ha->hanum, hdrive)); @@ -1944,17 +1944,17 @@ static int gdth_analyse_hdrive(gdth_ha_str *ha, ushort hdrive) } else { drv_hds = ha->info2 & 0xff; drv_secs = (ha->info2 >> 8) & 0xff; - drv_cyls = (ulong32)ha->hdr[hdrive].size / drv_hds / drv_secs; + drv_cyls = (u32)ha->hdr[hdrive].size / drv_hds / drv_secs; } - ha->hdr[hdrive].heads = (unchar)drv_hds; - ha->hdr[hdrive].secs = (unchar)drv_secs; + ha->hdr[hdrive].heads = (u8)drv_hds; + ha->hdr[hdrive].secs = (u8)drv_secs; /* round size */ ha->hdr[hdrive].size = drv_cyls * drv_hds * drv_secs; if (ha->cache_feat & GDT_64BIT) { if (gdth_internal_cmd(ha, CACHESERVICE, GDT_X_INFO, hdrive, 0, 0) && ha->info2 != 0) { - ha->hdr[hdrive].size = ((ulong64)ha->info2 << 32) | ha->info; + ha->hdr[hdrive].size = ((u64)ha->info2 << 32) | ha->info; } } TRACE2(("gdth_search_dr() cdr. %d size %d hds %d scs %d\n", @@ -1964,7 +1964,7 @@ static int gdth_analyse_hdrive(gdth_ha_str *ha, ushort hdrive) if (gdth_internal_cmd(ha, CACHESERVICE, GDT_DEVTYPE, hdrive, 0, 0)) { TRACE2(("gdth_search_dr() cache drive %d devtype %d\n", hdrive,ha->info)); - ha->hdr[hdrive].devtype = (ushort)ha->info; + ha->hdr[hdrive].devtype = (u16)ha->info; } /* cluster info */ @@ -1972,14 +1972,14 @@ static int gdth_analyse_hdrive(gdth_ha_str *ha, ushort hdrive) TRACE2(("gdth_search_dr() cache drive %d cluster info %d\n", hdrive,ha->info)); if (!shared_access) - ha->hdr[hdrive].cluster_type = (unchar)ha->info; + ha->hdr[hdrive].cluster_type = (u8)ha->info; } /* R/W attributes */ if (gdth_internal_cmd(ha, CACHESERVICE, GDT_RW_ATTRIBS, hdrive, 0, 0)) { TRACE2(("gdth_search_dr() cache drive %d r/w attrib. %d\n", hdrive,ha->info)); - ha->hdr[hdrive].rw_attribs = (unchar)ha->info; + ha->hdr[hdrive].rw_attribs = (u8)ha->info; } return 1; @@ -1988,12 +1988,12 @@ static int gdth_analyse_hdrive(gdth_ha_str *ha, ushort hdrive) /* command queueing/sending functions */ -static void gdth_putq(gdth_ha_str *ha, Scsi_Cmnd *scp, unchar priority) +static void gdth_putq(gdth_ha_str *ha, Scsi_Cmnd *scp, u8 priority) { struct gdth_cmndinfo *cmndinfo = gdth_cmnd_priv(scp); register Scsi_Cmnd *pscp; register Scsi_Cmnd *nscp; - ulong flags; + unsigned long flags; TRACE(("gdth_putq() priority %d\n",priority)); spin_lock_irqsave(&ha->smp_lock, flags); @@ -2023,7 +2023,7 @@ static void gdth_putq(gdth_ha_str *ha, Scsi_Cmnd *scp, unchar priority) ++flags; if (max_rq < flags) { max_rq = flags; - TRACE3(("GDT: max_rq = %d\n",(ushort)max_rq)); + TRACE3(("GDT: max_rq = %d\n",(u16)max_rq)); } #endif } @@ -2032,9 +2032,9 @@ static void gdth_next(gdth_ha_str *ha) { register Scsi_Cmnd *pscp; register Scsi_Cmnd *nscp; - unchar b, t, l, firsttime; - unchar this_cmd, next_cmd; - ulong flags = 0; + u8 b, t, l, firsttime; + u8 this_cmd, next_cmd; + unsigned long flags = 0; int cmd_index; TRACE(("gdth_next() hanum %d\n", ha->hanum)); @@ -2282,20 +2282,20 @@ static void gdth_next(gdth_ha_str *ha) * buffers, kmap_atomic() as needed. */ static void gdth_copy_internal_data(gdth_ha_str *ha, Scsi_Cmnd *scp, - char *buffer, ushort count) + char *buffer, u16 count) { - ushort cpcount,i, max_sg = scsi_sg_count(scp); - ushort cpsum,cpnow; + u16 cpcount,i, max_sg = scsi_sg_count(scp); + u16 cpsum,cpnow; struct scatterlist *sl; char *address; - cpcount = min_t(ushort, count, scsi_bufflen(scp)); + cpcount = min_t(u16, count, scsi_bufflen(scp)); if (cpcount) { cpsum=0; scsi_for_each_sg(scp, sl, max_sg, i) { unsigned long flags; - cpnow = (ushort)sl->length; + cpnow = (u16)sl->length; TRACE(("copy_internal() now %d sum %d count %d %d\n", cpnow, cpsum, cpcount, scsi_bufflen(scp))); if (cpsum+cpnow > cpcount) @@ -2325,7 +2325,7 @@ static void gdth_copy_internal_data(gdth_ha_str *ha, Scsi_Cmnd *scp, static int gdth_internal_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp) { - unchar t; + u8 t; gdth_inq_data inq; gdth_rdcap_data rdc; gdth_sense_data sd; @@ -2389,7 +2389,7 @@ static int gdth_internal_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp) case READ_CAPACITY: TRACE2(("Read capacity hdrive %d\n",t)); - if (ha->hdr[t].size > (ulong64)0xffffffff) + if (ha->hdr[t].size > (u64)0xffffffff) rdc.last_block_no = 0xffffffff; else rdc.last_block_no = cpu_to_be32(ha->hdr[t].size-1); @@ -2425,12 +2425,12 @@ static int gdth_internal_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp) return 0; } -static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, ushort hdrive) +static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, u16 hdrive) { register gdth_cmd_str *cmdp; struct gdth_cmndinfo *cmndinfo = gdth_cmnd_priv(scp); - ulong32 cnt, blockcnt; - ulong64 no, blockno; + u32 cnt, blockcnt; + u64 no, blockno; int i, cmd_index, read_write, sgcnt, mode64; cmdp = ha->pccb; @@ -2498,17 +2498,17 @@ static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, ushort hdrive) if (read_write) { if (scp->cmd_len == 16) { - memcpy(&no, &scp->cmnd[2], sizeof(ulong64)); + memcpy(&no, &scp->cmnd[2], sizeof(u64)); blockno = be64_to_cpu(no); - memcpy(&cnt, &scp->cmnd[10], sizeof(ulong32)); + memcpy(&cnt, &scp->cmnd[10], sizeof(u32)); blockcnt = be32_to_cpu(cnt); } else if (scp->cmd_len == 10) { - memcpy(&no, &scp->cmnd[2], sizeof(ulong32)); + memcpy(&no, &scp->cmnd[2], sizeof(u32)); blockno = be32_to_cpu(no); - memcpy(&cnt, &scp->cmnd[7], sizeof(ushort)); + memcpy(&cnt, &scp->cmnd[7], sizeof(u16)); blockcnt = be16_to_cpu(cnt); } else { - memcpy(&no, &scp->cmnd[0], sizeof(ulong32)); + memcpy(&no, &scp->cmnd[0], sizeof(u32)); blockno = be32_to_cpu(no) & 0x001fffffUL; blockcnt= scp->cmnd[4]==0 ? 0x100 : scp->cmnd[4]; } @@ -2516,7 +2516,7 @@ static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, ushort hdrive) cmdp->u.cache64.BlockNo = blockno; cmdp->u.cache64.BlockCnt = blockcnt; } else { - cmdp->u.cache.BlockNo = (ulong32)blockno; + cmdp->u.cache.BlockNo = (u32)blockno; cmdp->u.cache.BlockCnt = blockcnt; } @@ -2528,12 +2528,12 @@ static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, ushort hdrive) if (mode64) { struct scatterlist *sl; - cmdp->u.cache64.DestAddr= (ulong64)-1; + cmdp->u.cache64.DestAddr= (u64)-1; cmdp->u.cache64.sg_canz = sgcnt; scsi_for_each_sg(scp, sl, sgcnt, i) { cmdp->u.cache64.sg_lst[i].sg_ptr = sg_dma_address(sl); #ifdef GDTH_DMA_STATISTICS - if (cmdp->u.cache64.sg_lst[i].sg_ptr > (ulong64)0xffffffff) + if (cmdp->u.cache64.sg_lst[i].sg_ptr > (u64)0xffffffff) ha->dma64_cnt++; else ha->dma32_cnt++; @@ -2555,8 +2555,8 @@ static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, ushort hdrive) } #ifdef GDTH_STATISTICS - if (max_sg < (ulong32)sgcnt) { - max_sg = (ulong32)sgcnt; + if (max_sg < (u32)sgcnt) { + max_sg = (u32)sgcnt; TRACE3(("GDT: max_sg = %d\n",max_sg)); } #endif @@ -2572,7 +2572,7 @@ static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, ushort hdrive) TRACE(("cache cmd: cmd %d blockno. %d, blockcnt %d\n", cmdp->OpCode,cmdp->u.cache64.BlockNo,cmdp->u.cache64.BlockCnt)); ha->cmd_len = GDTOFFSOF(gdth_cmd_str,u.cache64.sg_lst) + - (ushort)cmdp->u.cache64.sg_canz * sizeof(gdth_sg64_str); + (u16)cmdp->u.cache64.sg_canz * sizeof(gdth_sg64_str); } else { TRACE(("cache cmd: addr. %x sganz %x sgptr0 %x sglen0 %x\n", cmdp->u.cache.DestAddr,cmdp->u.cache.sg_canz, @@ -2581,7 +2581,7 @@ static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, ushort hdrive) TRACE(("cache cmd: cmd %d blockno. %d, blockcnt %d\n", cmdp->OpCode,cmdp->u.cache.BlockNo,cmdp->u.cache.BlockCnt)); ha->cmd_len = GDTOFFSOF(gdth_cmd_str,u.cache.sg_lst) + - (ushort)cmdp->u.cache.sg_canz * sizeof(gdth_sg_str); + (u16)cmdp->u.cache.sg_canz * sizeof(gdth_sg_str); } if (ha->cmd_len & 3) ha->cmd_len += (4 - (ha->cmd_len & 3)); @@ -2600,15 +2600,15 @@ static int gdth_fill_cache_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, ushort hdrive) return cmd_index; } -static int gdth_fill_raw_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, unchar b) +static int gdth_fill_raw_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, u8 b) { register gdth_cmd_str *cmdp; - ushort i; + u16 i; dma_addr_t sense_paddr; int cmd_index, sgcnt, mode64; - unchar t,l; + u8 t,l; struct page *page; - ulong offset; + unsigned long offset; struct gdth_cmndinfo *cmndinfo; t = scp->device->id; @@ -2654,7 +2654,7 @@ static int gdth_fill_raw_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, unchar b) } else { page = virt_to_page(scp->sense_buffer); - offset = (ulong)scp->sense_buffer & ~PAGE_MASK; + offset = (unsigned long)scp->sense_buffer & ~PAGE_MASK; sense_paddr = pci_map_page(ha->pdev,page,offset, 16,PCI_DMA_FROMDEVICE); @@ -2703,12 +2703,12 @@ static int gdth_fill_raw_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, unchar b) if (mode64) { struct scatterlist *sl; - cmdp->u.raw64.sdata = (ulong64)-1; + cmdp->u.raw64.sdata = (u64)-1; cmdp->u.raw64.sg_ranz = sgcnt; scsi_for_each_sg(scp, sl, sgcnt, i) { cmdp->u.raw64.sg_lst[i].sg_ptr = sg_dma_address(sl); #ifdef GDTH_DMA_STATISTICS - if (cmdp->u.raw64.sg_lst[i].sg_ptr > (ulong64)0xffffffff) + if (cmdp->u.raw64.sg_lst[i].sg_ptr > (u64)0xffffffff) ha->dma64_cnt++; else ha->dma32_cnt++; @@ -2744,7 +2744,7 @@ static int gdth_fill_raw_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, unchar b) cmdp->u.raw64.sg_lst[0].sg_len)); /* evaluate command size */ ha->cmd_len = GDTOFFSOF(gdth_cmd_str,u.raw64.sg_lst) + - (ushort)cmdp->u.raw64.sg_ranz * sizeof(gdth_sg64_str); + (u16)cmdp->u.raw64.sg_ranz * sizeof(gdth_sg64_str); } else { TRACE(("raw cmd: addr. %x sganz %x sgptr0 %x sglen0 %x\n", cmdp->u.raw.sdata,cmdp->u.raw.sg_ranz, @@ -2752,7 +2752,7 @@ static int gdth_fill_raw_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp, unchar b) cmdp->u.raw.sg_lst[0].sg_len)); /* evaluate command size */ ha->cmd_len = GDTOFFSOF(gdth_cmd_str,u.raw.sg_lst) + - (ushort)cmdp->u.raw.sg_ranz * sizeof(gdth_sg_str); + (u16)cmdp->u.raw.sg_ranz * sizeof(gdth_sg_str); } } /* check space */ @@ -2802,7 +2802,7 @@ static int gdth_special_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp) if (cmdp->OpCode == GDT_IOCTL) { TRACE2(("IOCTL\n")); ha->cmd_len = - GDTOFFSOF(gdth_cmd_str,u.ioctl.p_param) + sizeof(ulong64); + GDTOFFSOF(gdth_cmd_str,u.ioctl.p_param) + sizeof(u64); } else if (cmdp->Service == CACHESERVICE) { TRACE2(("cache command %d\n",cmdp->OpCode)); if (ha->cache_feat & GDT_64BIT) @@ -2840,8 +2840,8 @@ static int gdth_special_cmd(gdth_ha_str *ha, Scsi_Cmnd *scp) /* Controller event handling functions */ -static gdth_evt_str *gdth_store_event(gdth_ha_str *ha, ushort source, - ushort idx, gdth_evt_data *evt) +static gdth_evt_str *gdth_store_event(gdth_ha_str *ha, u16 source, + u16 idx, gdth_evt_data *evt) { gdth_evt_str *e; struct timeval tv; @@ -2890,7 +2890,7 @@ static int gdth_read_event(gdth_ha_str *ha, int handle, gdth_evt_str *estr) { gdth_evt_str *e; int eindex; - ulong flags; + unsigned long flags; TRACE2(("gdth_read_event() handle %d\n", handle)); spin_lock_irqsave(&ha->smp_lock, flags); @@ -2919,12 +2919,12 @@ static int gdth_read_event(gdth_ha_str *ha, int handle, gdth_evt_str *estr) } static void gdth_readapp_event(gdth_ha_str *ha, - unchar application, gdth_evt_str *estr) + u8 application, gdth_evt_str *estr) { gdth_evt_str *e; int eindex; - ulong flags; - unchar found = FALSE; + unsigned long flags; + u8 found = FALSE; TRACE2(("gdth_readapp_event() app. %d\n", application)); spin_lock_irqsave(&ha->smp_lock, flags); @@ -2969,9 +2969,9 @@ static irqreturn_t __gdth_interrupt(gdth_ha_str *ha, gdt2_dpram_str __iomem *dp2_ptr; Scsi_Cmnd *scp; int rval, i; - unchar IStatus; - ushort Service; - ulong flags = 0; + u8 IStatus; + u16 Service; + unsigned long flags = 0; #ifdef INT_COAL int coalesced = FALSE; int next = FALSE; @@ -3018,7 +3018,7 @@ static irqreturn_t __gdth_interrupt(gdth_ha_str *ha, if (coalesced) { /* For coalesced requests all status information is found in the status buffer */ - IStatus = (unchar)(pcs->status & 0xff); + IStatus = (u8)(pcs->status & 0xff); } #endif @@ -3197,7 +3197,7 @@ static irqreturn_t __gdth_interrupt(gdth_ha_str *ha, ++act_int_coal; if (act_int_coal > max_int_coal) { max_int_coal = act_int_coal; - printk("GDT: max_int_coal = %d\n",(ushort)max_int_coal); + printk("GDT: max_int_coal = %d\n",(u16)max_int_coal); } #endif /* see if there is another status */ @@ -3225,12 +3225,12 @@ static irqreturn_t gdth_interrupt(int irq, void *dev_id) return __gdth_interrupt(ha, false, NULL); } -static int gdth_sync_event(gdth_ha_str *ha, int service, unchar index, +static int gdth_sync_event(gdth_ha_str *ha, int service, u8 index, Scsi_Cmnd *scp) { gdth_msg_str *msg; gdth_cmd_str *cmdp; - unchar b, t; + u8 b, t; struct gdth_cmndinfo *cmndinfo = gdth_cmnd_priv(scp); cmdp = ha->pccb; @@ -3263,7 +3263,7 @@ static int gdth_sync_event(gdth_ha_str *ha, int service, unchar index, cmdp->u.screen.su.msg.msg_addr = ha->msg_phys; ha->cmd_offs_dpmem = 0; ha->cmd_len = GDTOFFSOF(gdth_cmd_str,u.screen.su.msg.msg_addr) - + sizeof(ulong64); + + sizeof(u64); ha->cmd_cnt = 0; gdth_copy_command(ha); gdth_release_event(ha); @@ -3297,7 +3297,7 @@ static int gdth_sync_event(gdth_ha_str *ha, int service, unchar index, cmdp->u.screen.su.msg.msg_addr = ha->msg_phys; ha->cmd_offs_dpmem = 0; ha->cmd_len = GDTOFFSOF(gdth_cmd_str,u.screen.su.msg.msg_addr) - + sizeof(ulong64); + + sizeof(u64); ha->cmd_cnt = 0; gdth_copy_command(ha); gdth_release_event(ha); @@ -3335,7 +3335,7 @@ static int gdth_sync_event(gdth_ha_str *ha, int service, unchar index, cmndinfo->OpCode)); /* special commands GDT_CLUST_INFO/GDT_MOUNT ? */ if (cmndinfo->OpCode == GDT_CLUST_INFO) { - ha->hdr[t].cluster_type = (unchar)ha->info; + ha->hdr[t].cluster_type = (u8)ha->info; if (!(ha->hdr[t].cluster_type & CLUSTER_MOUNTED)) { /* NOT MOUNTED -> MOUNT */ @@ -3397,7 +3397,7 @@ static int gdth_sync_event(gdth_ha_str *ha, int service, unchar index, ha->hdr[t].cluster_type &= ~CLUSTER_RESERVED; } memset((char*)scp->sense_buffer,0,16); - if (ha->status == (ushort)S_CACHE_RESERV) { + if (ha->status == (u16)S_CACHE_RESERV) { scp->result = (DID_OK << 16) | (RESERVATION_CONFLICT << 1); } else { scp->sense_buffer[0] = 0x70; @@ -3614,16 +3614,16 @@ static int gdth_async_event(gdth_ha_str *ha) cmdp->u.screen.su.msg.msg_addr = ha->msg_phys; ha->cmd_offs_dpmem = 0; ha->cmd_len = GDTOFFSOF(gdth_cmd_str,u.screen.su.msg.msg_addr) - + sizeof(ulong64); + + sizeof(u64); ha->cmd_cnt = 0; gdth_copy_command(ha); if (ha->type == GDT_EISA) - printk("[EISA slot %d] ",(ushort)ha->brd_phys); + printk("[EISA slot %d] ",(u16)ha->brd_phys); else if (ha->type == GDT_ISA) - printk("[DPMEM 0x%4X] ",(ushort)ha->brd_phys); + printk("[DPMEM 0x%4X] ",(u16)ha->brd_phys); else - printk("[PCI %d/%d] ",(ushort)(ha->brd_phys>>8), - (ushort)((ha->brd_phys>>3)&0x1f)); + printk("[PCI %d/%d] ",(u16)(ha->brd_phys>>8), + (u16)((ha->brd_phys>>3)&0x1f)); gdth_release_event(ha); } @@ -3640,7 +3640,7 @@ static int gdth_async_event(gdth_ha_str *ha) ha->dvr.eu.async.service = ha->service; ha->dvr.eu.async.status = ha->status; ha->dvr.eu.async.info = ha->info; - *(ulong32 *)ha->dvr.eu.async.scsi_coord = ha->info2; + *(u32 *)ha->dvr.eu.async.scsi_coord = ha->info2; } gdth_store_event( ha, ES_ASYNC, ha->service, &ha->dvr ); gdth_log_event( &ha->dvr, NULL ); @@ -3648,8 +3648,8 @@ static int gdth_async_event(gdth_ha_str *ha) /* new host drive from expand? */ if (ha->service == CACHESERVICE && ha->status == 56) { TRACE2(("gdth_async_event(): new host drive %d created\n", - (ushort)ha->info)); - /* gdth_analyse_hdrive(hanum, (ushort)ha->info); */ + (u16)ha->info)); + /* gdth_analyse_hdrive(hanum, (u16)ha->info); */ } } return 1; @@ -3680,13 +3680,13 @@ static void gdth_log_event(gdth_evt_data *dvr, char *buffer) for (j=0,i=1; i < f[0]; i+=2) { switch (f[i+1]) { case 4: - stack.b[j++] = *(ulong32*)&dvr->eu.stream[(int)f[i]]; + stack.b[j++] = *(u32*)&dvr->eu.stream[(int)f[i]]; break; case 2: - stack.b[j++] = *(ushort*)&dvr->eu.stream[(int)f[i]]; + stack.b[j++] = *(u16*)&dvr->eu.stream[(int)f[i]]; break; case 1: - stack.b[j++] = *(unchar*)&dvr->eu.stream[(int)f[i]]; + stack.b[j++] = *(u8*)&dvr->eu.stream[(int)f[i]]; break; default: break; @@ -3712,14 +3712,14 @@ static void gdth_log_event(gdth_evt_data *dvr, char *buffer) } #ifdef GDTH_STATISTICS -static unchar gdth_timer_running; +static u8 gdth_timer_running; -static void gdth_timeout(ulong data) +static void gdth_timeout(unsigned long data) { - ulong32 i; + u32 i; Scsi_Cmnd *nscp; gdth_ha_str *ha; - ulong flags; + unsigned long flags; if(unlikely(list_empty(&gdth_instances))) { gdth_timer_running = 0; @@ -3891,8 +3891,8 @@ static enum blk_eh_timer_return gdth_timed_out(struct scsi_cmnd *scp) { gdth_ha_str *ha = shost_priv(scp->device->host); struct gdth_cmndinfo *cmndinfo = gdth_cmnd_priv(scp); - unchar b, t; - ulong flags; + u8 b, t; + unsigned long flags; enum blk_eh_timer_return retval = BLK_EH_NOT_HANDLED; TRACE(("%s() cmd 0x%x\n", scp->cmnd[0], __func__)); @@ -3924,9 +3924,9 @@ static int gdth_eh_bus_reset(Scsi_Cmnd *scp) { gdth_ha_str *ha = shost_priv(scp->device->host); int i; - ulong flags; + unsigned long flags; Scsi_Cmnd *cmnd; - unchar b; + u8 b; TRACE2(("gdth_eh_bus_reset()\n")); @@ -3974,7 +3974,7 @@ static int gdth_eh_bus_reset(Scsi_Cmnd *scp) static int gdth_bios_param(struct scsi_device *sdev,struct block_device *bdev,sector_t cap,int *ip) { - unchar b, t; + u8 b, t; gdth_ha_str *ha = shost_priv(sdev->host); struct scsi_device *sd; unsigned capacity; @@ -4062,7 +4062,7 @@ static int ioc_event(void __user *arg) { gdth_ioctl_event evt; gdth_ha_str *ha; - ulong flags; + unsigned long flags; if (copy_from_user(&evt, arg, sizeof(gdth_ioctl_event))) return -EFAULT; @@ -4098,8 +4098,8 @@ static int ioc_event(void __user *arg) static int ioc_lockdrv(void __user *arg) { gdth_ioctl_lockdrv ldrv; - unchar i, j; - ulong flags; + u8 i, j; + unsigned long flags; gdth_ha_str *ha; if (copy_from_user(&ldrv, arg, sizeof(gdth_ioctl_lockdrv))) @@ -4165,7 +4165,7 @@ static int ioc_general(void __user *arg, char *cmnd) { gdth_ioctl_general gen; char *buf = NULL; - ulong64 paddr; + u64 paddr; gdth_ha_str *ha; int rval; @@ -4194,7 +4194,7 @@ static int ioc_general(void __user *arg, char *cmnd) gen.command.u.cache64.DeviceNo = gen.command.u.cache.DeviceNo; /* addresses */ if (ha->cache_feat & SCATTER_GATHER) { - gen.command.u.cache64.DestAddr = (ulong64)-1; + gen.command.u.cache64.DestAddr = (u64)-1; gen.command.u.cache64.sg_canz = 1; gen.command.u.cache64.sg_lst[0].sg_ptr = paddr; gen.command.u.cache64.sg_lst[0].sg_len = gen.data_len; @@ -4207,7 +4207,7 @@ static int ioc_general(void __user *arg, char *cmnd) if (ha->cache_feat & SCATTER_GATHER) { gen.command.u.cache.DestAddr = 0xffffffff; gen.command.u.cache.sg_canz = 1; - gen.command.u.cache.sg_lst[0].sg_ptr = (ulong32)paddr; + gen.command.u.cache.sg_lst[0].sg_ptr = (u32)paddr; gen.command.u.cache.sg_lst[0].sg_len = gen.data_len; gen.command.u.cache.sg_lst[1].sg_len = 0; } else { @@ -4230,7 +4230,7 @@ static int ioc_general(void __user *arg, char *cmnd) gen.command.u.raw64.direction = gen.command.u.raw.direction; /* addresses */ if (ha->raw_feat & SCATTER_GATHER) { - gen.command.u.raw64.sdata = (ulong64)-1; + gen.command.u.raw64.sdata = (u64)-1; gen.command.u.raw64.sg_ranz = 1; gen.command.u.raw64.sg_lst[0].sg_ptr = paddr; gen.command.u.raw64.sg_lst[0].sg_len = gen.data_len; @@ -4244,14 +4244,14 @@ static int ioc_general(void __user *arg, char *cmnd) if (ha->raw_feat & SCATTER_GATHER) { gen.command.u.raw.sdata = 0xffffffff; gen.command.u.raw.sg_ranz = 1; - gen.command.u.raw.sg_lst[0].sg_ptr = (ulong32)paddr; + gen.command.u.raw.sg_lst[0].sg_ptr = (u32)paddr; gen.command.u.raw.sg_lst[0].sg_len = gen.data_len; gen.command.u.raw.sg_lst[1].sg_len = 0; } else { gen.command.u.raw.sdata = paddr; gen.command.u.raw.sg_ranz = 0; } - gen.command.u.raw.sense_data = (ulong32)paddr + gen.data_len; + gen.command.u.raw.sense_data = (u32)paddr + gen.data_len; } } else { gdth_ioctl_free(ha, gen.data_len+gen.sense_len, buf, paddr); @@ -4283,7 +4283,7 @@ static int ioc_hdrlist(void __user *arg, char *cmnd) gdth_ioctl_rescan *rsc; gdth_cmd_str *cmd; gdth_ha_str *ha; - unchar i; + u8 i; int rc = -ENOMEM; u32 cluster_type = 0; @@ -4335,11 +4335,11 @@ static int ioc_rescan(void __user *arg, char *cmnd) { gdth_ioctl_rescan *rsc; gdth_cmd_str *cmd; - ushort i, status, hdr_cnt; - ulong32 info; + u16 i, status, hdr_cnt; + u32 info; int cyls, hds, secs; int rc = -ENOMEM; - ulong flags; + unsigned long flags; gdth_ha_str *ha; rsc = kmalloc(sizeof(*rsc), GFP_KERNEL); @@ -4367,7 +4367,7 @@ static int ioc_rescan(void __user *arg, char *cmnd) status = __gdth_execute(ha->sdev, cmd, cmnd, 30, &info); i = 0; - hdr_cnt = (status == S_OK ? (ushort)info : 0); + hdr_cnt = (status == S_OK ? (u16)info : 0); } else { i = rsc->hdr_no; hdr_cnt = i + 1; @@ -4418,7 +4418,7 @@ static int ioc_rescan(void __user *arg, char *cmnd) status = __gdth_execute(ha->sdev, cmd, cmnd, 30, &info); spin_lock_irqsave(&ha->smp_lock, flags); - ha->hdr[i].devtype = (status == S_OK ? (ushort)info : 0); + ha->hdr[i].devtype = (status == S_OK ? (u16)info : 0); spin_unlock_irqrestore(&ha->smp_lock, flags); cmd->Service = CACHESERVICE; @@ -4432,7 +4432,7 @@ static int ioc_rescan(void __user *arg, char *cmnd) spin_lock_irqsave(&ha->smp_lock, flags); ha->hdr[i].cluster_type = - ((status == S_OK && !shared_access) ? (ushort)info : 0); + ((status == S_OK && !shared_access) ? (u16)info : 0); spin_unlock_irqrestore(&ha->smp_lock, flags); rsc->hdr_list[i].cluster_type = ha->hdr[i].cluster_type; @@ -4446,7 +4446,7 @@ static int ioc_rescan(void __user *arg, char *cmnd) status = __gdth_execute(ha->sdev, cmd, cmnd, 30, &info); spin_lock_irqsave(&ha->smp_lock, flags); - ha->hdr[i].rw_attribs = (status == S_OK ? (ushort)info : 0); + ha->hdr[i].rw_attribs = (status == S_OK ? (u16)info : 0); spin_unlock_irqrestore(&ha->smp_lock, flags); } @@ -4466,7 +4466,7 @@ static int gdth_ioctl(struct inode *inode, struct file *filep, { gdth_ha_str *ha; Scsi_Cmnd *scp; - ulong flags; + unsigned long flags; char cmnd[MAX_COMMAND_SIZE]; void __user *argp = (void __user *)arg; @@ -4495,9 +4495,9 @@ static int gdth_ioctl(struct inode *inode, struct file *filep, { gdth_ioctl_osvers osv; - osv.version = (unchar)(LINUX_VERSION_CODE >> 16); - osv.subversion = (unchar)(LINUX_VERSION_CODE >> 8); - osv.revision = (ushort)(LINUX_VERSION_CODE & 0xff); + osv.version = (u8)(LINUX_VERSION_CODE >> 16); + osv.subversion = (u8)(LINUX_VERSION_CODE >> 8); + osv.revision = (u16)(LINUX_VERSION_CODE & 0xff); if (copy_to_user(argp, &osv, sizeof(gdth_ioctl_osvers))) return -EFAULT; break; @@ -4512,10 +4512,10 @@ static int gdth_ioctl(struct inode *inode, struct file *filep, return -EFAULT; if (ha->type == GDT_ISA || ha->type == GDT_EISA) { - ctrt.type = (unchar)((ha->stype>>20) - 0x10); + ctrt.type = (u8)((ha->stype>>20) - 0x10); } else { if (ha->type != GDT_PCIMPR) { - ctrt.type = (unchar)((ha->stype<<4) + 6); + ctrt.type = (u8)((ha->stype<<4) + 6); } else { ctrt.type = (ha->oem_id == OEM_ID_INTEL ? 0xfd : 0xfe); @@ -4546,7 +4546,7 @@ static int gdth_ioctl(struct inode *inode, struct file *filep, case GDTIOCTL_LOCKCHN: { gdth_ioctl_lockchn lchn; - unchar i, j; + u8 i, j; if (copy_from_user(&lchn, argp, sizeof(gdth_ioctl_lockchn)) || (NULL == (ha = gdth_find_ha(lchn.ionode)))) @@ -4670,7 +4670,7 @@ static struct scsi_host_template gdth_template = { }; #ifdef CONFIG_ISA -static int __init gdth_isa_probe_one(ulong32 isa_bios) +static int __init gdth_isa_probe_one(u32 isa_bios) { struct Scsi_Host *shp; gdth_ha_str *ha; @@ -4802,7 +4802,7 @@ static int __init gdth_isa_probe_one(ulong32 isa_bios) #endif /* CONFIG_ISA */ #ifdef CONFIG_EISA -static int __init gdth_eisa_probe_one(ushort eisa_slot) +static int __init gdth_eisa_probe_one(u16 eisa_slot) { struct Scsi_Host *shp; gdth_ha_str *ha; @@ -5120,7 +5120,7 @@ static void gdth_remove_one(gdth_ha_str *ha) scsi_host_put(shp); } -static int gdth_halt(struct notifier_block *nb, ulong event, void *buf) +static int gdth_halt(struct notifier_block *nb, unsigned long event, void *buf) { gdth_ha_str *ha; @@ -5158,14 +5158,14 @@ static int __init gdth_init(void) if (probe_eisa_isa) { /* scanning for controllers, at first: ISA controller */ #ifdef CONFIG_ISA - ulong32 isa_bios; + u32 isa_bios; for (isa_bios = 0xc8000UL; isa_bios <= 0xd8000UL; isa_bios += 0x8000UL) gdth_isa_probe_one(isa_bios); #endif #ifdef CONFIG_EISA { - ushort eisa_slot; + u16 eisa_slot; for (eisa_slot = 0x1000; eisa_slot <= 0x8000; eisa_slot += 0x1000) gdth_eisa_probe_one(eisa_slot); diff --git a/drivers/scsi/gdth.h b/drivers/scsi/gdth.h index 1646444..120a062 100644 --- a/drivers/scsi/gdth.h +++ b/drivers/scsi/gdth.h @@ -321,524 +321,524 @@ /* screenservice message */ typedef struct { - ulong32 msg_handle; /* message handle */ - ulong32 msg_len; /* size of message */ - ulong32 msg_alen; /* answer length */ - unchar msg_answer; /* answer flag */ - unchar msg_ext; /* more messages */ - unchar msg_reserved[2]; + u32 msg_handle; /* message handle */ + u32 msg_len; /* size of message */ + u32 msg_alen; /* answer length */ + u8 msg_answer; /* answer flag */ + u8 msg_ext; /* more messages */ + u8 msg_reserved[2]; char msg_text[MSGLEN+2]; /* the message text */ -} PACKED gdth_msg_str; +} __attribute__((packed)) gdth_msg_str; /* IOCTL data structures */ /* Status coalescing buffer for returning multiple requests per interrupt */ typedef struct { - ulong32 status; - ulong32 ext_status; - ulong32 info0; - ulong32 info1; -} PACKED gdth_coal_status; + u32 status; + u32 ext_status; + u32 info0; + u32 info1; +} __attribute__((packed)) gdth_coal_status; /* performance mode data structure */ typedef struct { - ulong32 version; /* The version of this IOCTL structure. */ - ulong32 st_mode; /* 0=dis., 1=st_buf_addr1 valid, 2=both */ - ulong32 st_buff_addr1; /* physical address of status buffer 1 */ - ulong32 st_buff_u_addr1; /* reserved for 64 bit addressing */ - ulong32 st_buff_indx1; /* reserved command idx. for this buffer */ - ulong32 st_buff_addr2; /* physical address of status buffer 1 */ - ulong32 st_buff_u_addr2; /* reserved for 64 bit addressing */ - ulong32 st_buff_indx2; /* reserved command idx. for this buffer */ - ulong32 st_buff_size; /* size of each buffer in bytes */ - ulong32 cmd_mode; /* 0 = mode disabled, 1 = cmd_buff_addr1 */ - ulong32 cmd_buff_addr1; /* physical address of cmd buffer 1 */ - ulong32 cmd_buff_u_addr1; /* reserved for 64 bit addressing */ - ulong32 cmd_buff_indx1; /* cmd buf addr1 unique identifier */ - ulong32 cmd_buff_addr2; /* physical address of cmd buffer 1 */ - ulong32 cmd_buff_u_addr2; /* reserved for 64 bit addressing */ - ulong32 cmd_buff_indx2; /* cmd buf addr1 unique identifier */ - ulong32 cmd_buff_size; /* size of each cmd bufer in bytes */ - ulong32 reserved1; - ulong32 reserved2; -} PACKED gdth_perf_modes; + u32 version; /* The version of this IOCTL structure. */ + u32 st_mode; /* 0=dis., 1=st_buf_addr1 valid, 2=both */ + u32 st_buff_addr1; /* physical address of status buffer 1 */ + u32 st_buff_u_addr1; /* reserved for 64 bit addressing */ + u32 st_buff_indx1; /* reserved command idx. for this buffer */ + u32 st_buff_addr2; /* physical address of status buffer 1 */ + u32 st_buff_u_addr2; /* reserved for 64 bit addressing */ + u32 st_buff_indx2; /* reserved command idx. for this buffer */ + u32 st_buff_size; /* size of each buffer in bytes */ + u32 cmd_mode; /* 0 = mode disabled, 1 = cmd_buff_addr1 */ + u32 cmd_buff_addr1; /* physical address of cmd buffer 1 */ + u32 cmd_buff_u_addr1; /* reserved for 64 bit addressing */ + u32 cmd_buff_indx1; /* cmd buf addr1 unique identifier */ + u32 cmd_buff_addr2; /* physical address of cmd buffer 1 */ + u32 cmd_buff_u_addr2; /* reserved for 64 bit addressing */ + u32 cmd_buff_indx2; /* cmd buf addr1 unique identifier */ + u32 cmd_buff_size; /* size of each cmd bufer in bytes */ + u32 reserved1; + u32 reserved2; +} __attribute__((packed)) gdth_perf_modes; /* SCSI drive info */ typedef struct { - unchar vendor[8]; /* vendor string */ - unchar product[16]; /* product string */ - unchar revision[4]; /* revision */ - ulong32 sy_rate; /* current rate for sync. tr. */ - ulong32 sy_max_rate; /* max. rate for sync. tr. */ - ulong32 no_ldrive; /* belongs to this log. drv.*/ - ulong32 blkcnt; /* number of blocks */ - ushort blksize; /* size of block in bytes */ - unchar available; /* flag: access is available */ - unchar init; /* medium is initialized */ - unchar devtype; /* SCSI devicetype */ - unchar rm_medium; /* medium is removable */ - unchar wp_medium; /* medium is write protected */ - unchar ansi; /* SCSI I/II or III? */ - unchar protocol; /* same as ansi */ - unchar sync; /* flag: sync. transfer enab. */ - unchar disc; /* flag: disconnect enabled */ - unchar queueing; /* flag: command queing enab. */ - unchar cached; /* flag: caching enabled */ - unchar target_id; /* target ID of device */ - unchar lun; /* LUN id of device */ - unchar orphan; /* flag: drive fragment */ - ulong32 last_error; /* sense key or drive state */ - ulong32 last_result; /* result of last command */ - ulong32 check_errors; /* err. in last surface check */ - unchar percent; /* progress for surface check */ - unchar last_check; /* IOCTRL operation */ - unchar res[2]; - ulong32 flags; /* from 1.19/2.19: raw reserv.*/ - unchar multi_bus; /* multi bus dev? (fibre ch.) */ - unchar mb_status; /* status: available? */ - unchar res2[2]; - unchar mb_alt_status; /* status on second bus */ - unchar mb_alt_bid; /* number of second bus */ - unchar mb_alt_tid; /* target id on second bus */ - unchar res3; - unchar fc_flag; /* from 1.22/2.22: info valid?*/ - unchar res4; - ushort fc_frame_size; /* frame size (bytes) */ + u8 vendor[8]; /* vendor string */ + u8 product[16]; /* product string */ + u8 revision[4]; /* revision */ + u32 sy_rate; /* current rate for sync. tr. */ + u32 sy_max_rate; /* max. rate for sync. tr. */ + u32 no_ldrive; /* belongs to this log. drv.*/ + u32 blkcnt; /* number of blocks */ + u16 blksize; /* size of block in bytes */ + u8 available; /* flag: access is available */ + u8 init; /* medium is initialized */ + u8 devtype; /* SCSI devicetype */ + u8 rm_medium; /* medium is removable */ + u8 wp_medium; /* medium is write protected */ + u8 ansi; /* SCSI I/II or III? */ + u8 protocol; /* same as ansi */ + u8 sync; /* flag: sync. transfer enab. */ + u8 disc; /* flag: disconnect enabled */ + u8 queueing; /* flag: command queing enab. */ + u8 cached; /* flag: caching enabled */ + u8 target_id; /* target ID of device */ + u8 lun; /* LUN id of device */ + u8 orphan; /* flag: drive fragment */ + u32 last_error; /* sense key or drive state */ + u32 last_result; /* result of last command */ + u32 check_errors; /* err. in last surface check */ + u8 percent; /* progress for surface check */ + u8 last_check; /* IOCTRL operation */ + u8 res[2]; + u32 flags; /* from 1.19/2.19: raw reserv.*/ + u8 multi_bus; /* multi bus dev? (fibre ch.) */ + u8 mb_status; /* status: available? */ + u8 res2[2]; + u8 mb_alt_status; /* status on second bus */ + u8 mb_alt_bid; /* number of second bus */ + u8 mb_alt_tid; /* target id on second bus */ + u8 res3; + u8 fc_flag; /* from 1.22/2.22: info valid?*/ + u8 res4; + u16 fc_frame_size; /* frame size (bytes) */ char wwn[8]; /* world wide name */ -} PACKED gdth_diskinfo_str; +} __attribute__((packed)) gdth_diskinfo_str; /* get SCSI channel count */ typedef struct { - ulong32 channel_no; /* number of channel */ - ulong32 drive_cnt; /* drive count */ - unchar siop_id; /* SCSI processor ID */ - unchar siop_state; /* SCSI processor state */ -} PACKED gdth_getch_str; + u32 channel_no; /* number of channel */ + u32 drive_cnt; /* drive count */ + u8 siop_id; /* SCSI processor ID */ + u8 siop_state; /* SCSI processor state */ +} __attribute__((packed)) gdth_getch_str; /* get SCSI drive numbers */ typedef struct { - ulong32 sc_no; /* SCSI channel */ - ulong32 sc_cnt; /* sc_list[] elements */ - ulong32 sc_list[MAXID]; /* minor device numbers */ -} PACKED gdth_drlist_str; + u32 sc_no; /* SCSI channel */ + u32 sc_cnt; /* sc_list[] elements */ + u32 sc_list[MAXID]; /* minor device numbers */ +} __attribute__((packed)) gdth_drlist_str; /* get grown/primary defect count */ typedef struct { - unchar sddc_type; /* 0x08: grown, 0x10: prim. */ - unchar sddc_format; /* list entry format */ - unchar sddc_len; /* list entry length */ - unchar sddc_res; - ulong32 sddc_cnt; /* entry count */ -} PACKED gdth_defcnt_str; + u8 sddc_type; /* 0x08: grown, 0x10: prim. */ + u8 sddc_format; /* list entry format */ + u8 sddc_len; /* list entry length */ + u8 sddc_res; + u32 sddc_cnt; /* entry count */ +} __attribute__((packed)) gdth_defcnt_str; /* disk statistics */ typedef struct { - ulong32 bid; /* SCSI channel */ - ulong32 first; /* first SCSI disk */ - ulong32 entries; /* number of elements */ - ulong32 count; /* (R) number of init. el. */ - ulong32 mon_time; /* time stamp */ + u32 bid; /* SCSI channel */ + u32 first; /* first SCSI disk */ + u32 entries; /* number of elements */ + u32 count; /* (R) number of init. el. */ + u32 mon_time; /* time stamp */ struct { - unchar tid; /* target ID */ - unchar lun; /* LUN */ - unchar res[2]; - ulong32 blk_size; /* block size in bytes */ - ulong32 rd_count; /* bytes read */ - ulong32 wr_count; /* bytes written */ - ulong32 rd_blk_count; /* blocks read */ - ulong32 wr_blk_count; /* blocks written */ - ulong32 retries; /* retries */ - ulong32 reassigns; /* reassigns */ - } PACKED list[1]; -} PACKED gdth_dskstat_str; + u8 tid; /* target ID */ + u8 lun; /* LUN */ + u8 res[2]; + u32 blk_size; /* block size in bytes */ + u32 rd_count; /* bytes read */ + u32 wr_count; /* bytes written */ + u32 rd_blk_count; /* blocks read */ + u32 wr_blk_count; /* blocks written */ + u32 retries; /* retries */ + u32 reassigns; /* reassigns */ + } __attribute__((packed)) list[1]; +} __attribute__((packed)) gdth_dskstat_str; /* IO channel header */ typedef struct { - ulong32 version; /* version (-1UL: newest) */ - unchar list_entries; /* list entry count */ - unchar first_chan; /* first channel number */ - unchar last_chan; /* last channel number */ - unchar chan_count; /* (R) channel count */ - ulong32 list_offset; /* offset of list[0] */ -} PACKED gdth_iochan_header; + u32 version; /* version (-1UL: newest) */ + u8 list_entries; /* list entry count */ + u8 first_chan; /* first channel number */ + u8 last_chan; /* last channel number */ + u8 chan_count; /* (R) channel count */ + u32 list_offset; /* offset of list[0] */ +} __attribute__((packed)) gdth_iochan_header; /* get IO channel description */ typedef struct { gdth_iochan_header hdr; struct { - ulong32 address; /* channel address */ - unchar type; /* type (SCSI, FCAL) */ - unchar local_no; /* local number */ - ushort features; /* channel features */ - } PACKED list[MAXBUS]; -} PACKED gdth_iochan_str; + u32 address; /* channel address */ + u8 type; /* type (SCSI, FCAL) */ + u8 local_no; /* local number */ + u16 features; /* channel features */ + } __attribute__((packed)) list[MAXBUS]; +} __attribute__((packed)) gdth_iochan_str; /* get raw IO channel description */ typedef struct { gdth_iochan_header hdr; struct { - unchar proc_id; /* processor id */ - unchar proc_defect; /* defect ? */ - unchar reserved[2]; - } PACKED list[MAXBUS]; -} PACKED gdth_raw_iochan_str; + u8 proc_id; /* processor id */ + u8 proc_defect; /* defect ? */ + u8 reserved[2]; + } __attribute__((packed)) list[MAXBUS]; +} __attribute__((packed)) gdth_raw_iochan_str; /* array drive component */ typedef struct { - ulong32 al_controller; /* controller ID */ - unchar al_cache_drive; /* cache drive number */ - unchar al_status; /* cache drive state */ - unchar al_res[2]; -} PACKED gdth_arraycomp_str; + u32 al_controller; /* controller ID */ + u8 al_cache_drive; /* cache drive number */ + u8 al_status; /* cache drive state */ + u8 al_res[2]; +} __attribute__((packed)) gdth_arraycomp_str; /* array drive information */ typedef struct { - unchar ai_type; /* array type (RAID0,4,5) */ - unchar ai_cache_drive_cnt; /* active cachedrives */ - unchar ai_state; /* array drive state */ - unchar ai_master_cd; /* master cachedrive */ - ulong32 ai_master_controller; /* ID of master controller */ - ulong32 ai_size; /* user capacity [sectors] */ - ulong32 ai_striping_size; /* striping size [sectors] */ - ulong32 ai_secsize; /* sector size [bytes] */ - ulong32 ai_err_info; /* failed cache drive */ - unchar ai_name[8]; /* name of the array drive */ - unchar ai_controller_cnt; /* number of controllers */ - unchar ai_removable; /* flag: removable */ - unchar ai_write_protected; /* flag: write protected */ - unchar ai_devtype; /* type: always direct access */ + u8 ai_type; /* array type (RAID0,4,5) */ + u8 ai_cache_drive_cnt; /* active cachedrives */ + u8 ai_state; /* array drive state */ + u8 ai_master_cd; /* master cachedrive */ + u32 ai_master_controller; /* ID of master controller */ + u32 ai_size; /* user capacity [sectors] */ + u32 ai_striping_size; /* striping size [sectors] */ + u32 ai_secsize; /* sector size [bytes] */ + u32 ai_err_info; /* failed cache drive */ + u8 ai_name[8]; /* name of the array drive */ + u8 ai_controller_cnt; /* number of controllers */ + u8 ai_removable; /* flag: removable */ + u8 ai_write_protected; /* flag: write protected */ + u8 ai_devtype; /* type: always direct access */ gdth_arraycomp_str ai_drives[35]; /* drive components: */ - unchar ai_drive_entries; /* number of drive components */ - unchar ai_protected; /* protection flag */ - unchar ai_verify_state; /* state of a parity verify */ - unchar ai_ext_state; /* extended array drive state */ - unchar ai_expand_state; /* array expand state (>=2.18)*/ - unchar ai_reserved[3]; -} PACKED gdth_arrayinf_str; + u8 ai_drive_entries; /* number of drive components */ + u8 ai_protected; /* protection flag */ + u8 ai_verify_state; /* state of a parity verify */ + u8 ai_ext_state; /* extended array drive state */ + u8 ai_expand_state; /* array expand state (>=2.18)*/ + u8 ai_reserved[3]; +} __attribute__((packed)) gdth_arrayinf_str; /* get array drive list */ typedef struct { - ulong32 controller_no; /* controller no. */ - unchar cd_handle; /* master cachedrive */ - unchar is_arrayd; /* Flag: is array drive? */ - unchar is_master; /* Flag: is array master? */ - unchar is_parity; /* Flag: is parity drive? */ - unchar is_hotfix; /* Flag: is hotfix drive? */ - unchar res[3]; -} PACKED gdth_alist_str; + u32 controller_no; /* controller no. */ + u8 cd_handle; /* master cachedrive */ + u8 is_arrayd; /* Flag: is array drive? */ + u8 is_master; /* Flag: is array master? */ + u8 is_parity; /* Flag: is parity drive? */ + u8 is_hotfix; /* Flag: is hotfix drive? */ + u8 res[3]; +} __attribute__((packed)) gdth_alist_str; typedef struct { - ulong32 entries_avail; /* allocated entries */ - ulong32 entries_init; /* returned entries */ - ulong32 first_entry; /* first entry number */ - ulong32 list_offset; /* offset of following list */ + u32 entries_avail; /* allocated entries */ + u32 entries_init; /* returned entries */ + u32 first_entry; /* first entry number */ + u32 list_offset; /* offset of following list */ gdth_alist_str list[1]; /* list */ -} PACKED gdth_arcdl_str; +} __attribute__((packed)) gdth_arcdl_str; /* cache info/config IOCTL */ typedef struct { - ulong32 version; /* firmware version */ - ushort state; /* cache state (on/off) */ - ushort strategy; /* cache strategy */ - ushort write_back; /* write back state (on/off) */ - ushort block_size; /* cache block size */ -} PACKED gdth_cpar_str; + u32 version; /* firmware version */ + u16 state; /* cache state (on/off) */ + u16 strategy; /* cache strategy */ + u16 write_back; /* write back state (on/off) */ + u16 block_size; /* cache block size */ +} __attribute__((packed)) gdth_cpar_str; typedef struct { - ulong32 csize; /* cache size */ - ulong32 read_cnt; /* read/write counter */ - ulong32 write_cnt; - ulong32 tr_hits; /* hits */ - ulong32 sec_hits; - ulong32 sec_miss; /* misses */ -} PACKED gdth_cstat_str; + u32 csize; /* cache size */ + u32 read_cnt; /* read/write counter */ + u32 write_cnt; + u32 tr_hits; /* hits */ + u32 sec_hits; + u32 sec_miss; /* misses */ +} __attribute__((packed)) gdth_cstat_str; typedef struct { gdth_cpar_str cpar; gdth_cstat_str cstat; -} PACKED gdth_cinfo_str; +} __attribute__((packed)) gdth_cinfo_str; /* cache drive info */ typedef struct { - unchar cd_name[8]; /* cache drive name */ - ulong32 cd_devtype; /* SCSI devicetype */ - ulong32 cd_ldcnt; /* number of log. drives */ - ulong32 cd_last_error; /* last error */ - unchar cd_initialized; /* drive is initialized */ - unchar cd_removable; /* media is removable */ - unchar cd_write_protected; /* write protected */ - unchar cd_flags; /* Pool Hot Fix? */ - ulong32 ld_blkcnt; /* number of blocks */ - ulong32 ld_blksize; /* blocksize */ - ulong32 ld_dcnt; /* number of disks */ - ulong32 ld_slave; /* log. drive index */ - ulong32 ld_dtype; /* type of logical drive */ - ulong32 ld_last_error; /* last error */ - unchar ld_name[8]; /* log. drive name */ - unchar ld_error; /* error */ -} PACKED gdth_cdrinfo_str; + u8 cd_name[8]; /* cache drive name */ + u32 cd_devtype; /* SCSI devicetype */ + u32 cd_ldcnt; /* number of log. drives */ + u32 cd_last_error; /* last error */ + u8 cd_initialized; /* drive is initialized */ + u8 cd_removable; /* media is removable */ + u8 cd_write_protected; /* write protected */ + u8 cd_flags; /* Pool Hot Fix? */ + u32 ld_blkcnt; /* number of blocks */ + u32 ld_blksize; /* blocksize */ + u32 ld_dcnt; /* number of disks */ + u32 ld_slave; /* log. drive index */ + u32 ld_dtype; /* type of logical drive */ + u32 ld_last_error; /* last error */ + u8 ld_name[8]; /* log. drive name */ + u8 ld_error; /* error */ +} __attribute__((packed)) gdth_cdrinfo_str; /* OEM string */ typedef struct { - ulong32 ctl_version; - ulong32 file_major_version; - ulong32 file_minor_version; - ulong32 buffer_size; - ulong32 cpy_count; - ulong32 ext_error; - ulong32 oem_id; - ulong32 board_id; -} PACKED gdth_oem_str_params; - -typedef struct { - unchar product_0_1_name[16]; - unchar product_4_5_name[16]; - unchar product_cluster_name[16]; - unchar product_reserved[16]; - unchar scsi_cluster_target_vendor_id[16]; - unchar cluster_raid_fw_name[16]; - unchar oem_brand_name[16]; - unchar oem_raid_type[16]; - unchar bios_type[13]; - unchar bios_title[50]; - unchar oem_company_name[37]; - ulong32 pci_id_1; - ulong32 pci_id_2; - unchar validation_status[80]; - unchar reserved_1[4]; - unchar scsi_host_drive_inquiry_vendor_id[16]; - unchar library_file_template[16]; - unchar reserved_2[16]; - unchar tool_name_1[32]; - unchar tool_name_2[32]; - unchar tool_name_3[32]; - unchar oem_contact_1[84]; - unchar oem_contact_2[84]; - unchar oem_contact_3[84]; -} PACKED gdth_oem_str; + u32 ctl_version; + u32 file_major_version; + u32 file_minor_version; + u32 buffer_size; + u32 cpy_count; + u32 ext_error; + u32 oem_id; + u32 board_id; +} __attribute__((packed)) gdth_oem_str_params; + +typedef struct { + u8 product_0_1_name[16]; + u8 product_4_5_name[16]; + u8 product_cluster_name[16]; + u8 product_reserved[16]; + u8 scsi_cluster_target_vendor_id[16]; + u8 cluster_raid_fw_name[16]; + u8 oem_brand_name[16]; + u8 oem_raid_type[16]; + u8 bios_type[13]; + u8 bios_title[50]; + u8 oem_company_name[37]; + u32 pci_id_1; + u32 pci_id_2; + u8 validation_status[80]; + u8 reserved_1[4]; + u8 scsi_host_drive_inquiry_vendor_id[16]; + u8 library_file_template[16]; + u8 reserved_2[16]; + u8 tool_name_1[32]; + u8 tool_name_2[32]; + u8 tool_name_3[32]; + u8 oem_contact_1[84]; + u8 oem_contact_2[84]; + u8 oem_contact_3[84]; +} __attribute__((packed)) gdth_oem_str; typedef struct { gdth_oem_str_params params; gdth_oem_str text; -} PACKED gdth_oem_str_ioctl; +} __attribute__((packed)) gdth_oem_str_ioctl; /* board features */ typedef struct { - unchar chaining; /* Chaining supported */ - unchar striping; /* Striping (RAID-0) supp. */ - unchar mirroring; /* Mirroring (RAID-1) supp. */ - unchar raid; /* RAID-4/5/10 supported */ -} PACKED gdth_bfeat_str; + u8 chaining; /* Chaining supported */ + u8 striping; /* Striping (RAID-0) supp. */ + u8 mirroring; /* Mirroring (RAID-1) supp. */ + u8 raid; /* RAID-4/5/10 supported */ +} __attribute__((packed)) gdth_bfeat_str; /* board info IOCTL */ typedef struct { - ulong32 ser_no; /* serial no. */ - unchar oem_id[2]; /* OEM ID */ - ushort ep_flags; /* eprom flags */ - ulong32 proc_id; /* processor ID */ - ulong32 memsize; /* memory size (bytes) */ - unchar mem_banks; /* memory banks */ - unchar chan_type; /* channel type */ - unchar chan_count; /* channel count */ - unchar rdongle_pres; /* dongle present? */ - ulong32 epr_fw_ver; /* (eprom) firmware version */ - ulong32 upd_fw_ver; /* (update) firmware version */ - ulong32 upd_revision; /* update revision */ + u32 ser_no; /* serial no. */ + u8 oem_id[2]; /* OEM ID */ + u16 ep_flags; /* eprom flags */ + u32 proc_id; /* processor ID */ + u32 memsize; /* memory size (bytes) */ + u8 mem_banks; /* memory banks */ + u8 chan_type; /* channel type */ + u8 chan_count; /* channel count */ + u8 rdongle_pres; /* dongle present? */ + u32 epr_fw_ver; /* (eprom) firmware version */ + u32 upd_fw_ver; /* (update) firmware version */ + u32 upd_revision; /* update revision */ char type_string[16]; /* controller name */ char raid_string[16]; /* RAID firmware name */ - unchar update_pres; /* update present? */ - unchar xor_pres; /* XOR engine present? */ - unchar prom_type; /* ROM type (eprom/flash) */ - unchar prom_count; /* number of ROM devices */ - ulong32 dup_pres; /* duplexing module present? */ - ulong32 chan_pres; /* number of expansion chn. */ - ulong32 mem_pres; /* memory expansion inst. ? */ - unchar ft_bus_system; /* fault bus supported? */ - unchar subtype_valid; /* board_subtype valid? */ - unchar board_subtype; /* subtype/hardware level */ - unchar ramparity_pres; /* RAM parity check hardware? */ -} PACKED gdth_binfo_str; + u8 update_pres; /* update present? */ + u8 xor_pres; /* XOR engine present? */ + u8 prom_type; /* ROM type (eprom/flash) */ + u8 prom_count; /* number of ROM devices */ + u32 dup_pres; /* duplexing module present? */ + u32 chan_pres; /* number of expansion chn. */ + u32 mem_pres; /* memory expansion inst. ? */ + u8 ft_bus_system; /* fault bus supported? */ + u8 subtype_valid; /* board_subtype valid? */ + u8 board_subtype; /* subtype/hardware level */ + u8 ramparity_pres; /* RAM parity check hardware? */ +} __attribute__((packed)) gdth_binfo_str; /* get host drive info */ typedef struct { char name[8]; /* host drive name */ - ulong32 size; /* size (sectors) */ - unchar host_drive; /* host drive number */ - unchar log_drive; /* log. drive (master) */ - unchar reserved; - unchar rw_attribs; /* r/w attribs */ - ulong32 start_sec; /* start sector */ -} PACKED gdth_hentry_str; - -typedef struct { - ulong32 entries; /* entry count */ - ulong32 offset; /* offset of entries */ - unchar secs_p_head; /* sectors/head */ - unchar heads_p_cyl; /* heads/cylinder */ - unchar reserved; - unchar clust_drvtype; /* cluster drive type */ - ulong32 location; /* controller number */ + u32 size; /* size (sectors) */ + u8 host_drive; /* host drive number */ + u8 log_drive; /* log. drive (master) */ + u8 reserved; + u8 rw_attribs; /* r/w attribs */ + u32 start_sec; /* start sector */ +} __attribute__((packed)) gdth_hentry_str; + +typedef struct { + u32 entries; /* entry count */ + u32 offset; /* offset of entries */ + u8 secs_p_head; /* sectors/head */ + u8 heads_p_cyl; /* heads/cylinder */ + u8 reserved; + u8 clust_drvtype; /* cluster drive type */ + u32 location; /* controller number */ gdth_hentry_str entry[MAX_HDRIVES]; /* entries */ -} PACKED gdth_hget_str; +} __attribute__((packed)) gdth_hget_str; /* DPRAM structures */ /* interface area ISA/PCI */ typedef struct { - unchar S_Cmd_Indx; /* special command */ - unchar volatile S_Status; /* status special command */ - ushort reserved1; - ulong32 S_Info[4]; /* add. info special command */ - unchar volatile Sema0; /* command semaphore */ - unchar reserved2[3]; - unchar Cmd_Index; /* command number */ - unchar reserved3[3]; - ushort volatile Status; /* command status */ - ushort Service; /* service(for async.events) */ - ulong32 Info[2]; /* additional info */ + u8 S_Cmd_Indx; /* special command */ + u8 volatile S_Status; /* status special command */ + u16 reserved1; + u32 S_Info[4]; /* add. info special command */ + u8 volatile Sema0; /* command semaphore */ + u8 reserved2[3]; + u8 Cmd_Index; /* command number */ + u8 reserved3[3]; + u16 volatile Status; /* command status */ + u16 Service; /* service(for async.events) */ + u32 Info[2]; /* additional info */ struct { - ushort offset; /* command offs. in the DPRAM*/ - ushort serv_id; /* service */ - } PACKED comm_queue[MAXOFFSETS]; /* command queue */ - ulong32 bios_reserved[2]; - unchar gdt_dpr_cmd[1]; /* commands */ -} PACKED gdt_dpr_if; + u16 offset; /* command offs. in the DPRAM*/ + u16 serv_id; /* service */ + } __attribute__((packed)) comm_queue[MAXOFFSETS]; /* command queue */ + u32 bios_reserved[2]; + u8 gdt_dpr_cmd[1]; /* commands */ +} __attribute__((packed)) gdt_dpr_if; /* SRAM structure PCI controllers */ typedef struct { - ulong32 magic; /* controller ID from BIOS */ - ushort need_deinit; /* switch betw. BIOS/driver */ - unchar switch_support; /* see need_deinit */ - unchar padding[9]; - unchar os_used[16]; /* OS code per service */ - unchar unused[28]; - unchar fw_magic; /* contr. ID from firmware */ -} PACKED gdt_pci_sram; + u32 magic; /* controller ID from BIOS */ + u16 need_deinit; /* switch betw. BIOS/driver */ + u8 switch_support; /* see need_deinit */ + u8 padding[9]; + u8 os_used[16]; /* OS code per service */ + u8 unused[28]; + u8 fw_magic; /* contr. ID from firmware */ +} __attribute__((packed)) gdt_pci_sram; /* SRAM structure EISA controllers (but NOT GDT3000/3020) */ typedef struct { - unchar os_used[16]; /* OS code per service */ - ushort need_deinit; /* switch betw. BIOS/driver */ - unchar switch_support; /* see need_deinit */ - unchar padding; -} PACKED gdt_eisa_sram; + u8 os_used[16]; /* OS code per service */ + u16 need_deinit; /* switch betw. BIOS/driver */ + u8 switch_support; /* see need_deinit */ + u8 padding; +} __attribute__((packed)) gdt_eisa_sram; /* DPRAM ISA controllers */ typedef struct { union { struct { - unchar bios_used[0x3c00-32]; /* 15KB - 32Bytes BIOS */ - ulong32 magic; /* controller (EISA) ID */ - ushort need_deinit; /* switch betw. BIOS/driver */ - unchar switch_support; /* see need_deinit */ - unchar padding[9]; - unchar os_used[16]; /* OS code per service */ - } PACKED dp_sram; - unchar bios_area[0x4000]; /* 16KB reserved for BIOS */ + u8 bios_used[0x3c00-32]; /* 15KB - 32Bytes BIOS */ + u32 magic; /* controller (EISA) ID */ + u16 need_deinit; /* switch betw. BIOS/driver */ + u8 switch_support; /* see need_deinit */ + u8 padding[9]; + u8 os_used[16]; /* OS code per service */ + } __attribute__((packed)) dp_sram; + u8 bios_area[0x4000]; /* 16KB reserved for BIOS */ } bu; union { gdt_dpr_if ic; /* interface area */ - unchar if_area[0x3000]; /* 12KB for interface */ + u8 if_area[0x3000]; /* 12KB for interface */ } u; struct { - unchar memlock; /* write protection DPRAM */ - unchar event; /* release event */ - unchar irqen; /* board interrupts enable */ - unchar irqdel; /* acknowledge board int. */ - unchar volatile Sema1; /* status semaphore */ - unchar rq; /* IRQ/DRQ configuration */ - } PACKED io; -} PACKED gdt2_dpram_str; + u8 memlock; /* write protection DPRAM */ + u8 event; /* release event */ + u8 irqen; /* board interrupts enable */ + u8 irqdel; /* acknowledge board int. */ + u8 volatile Sema1; /* status semaphore */ + u8 rq; /* IRQ/DRQ configuration */ + } __attribute__((packed)) io; +} __attribute__((packed)) gdt2_dpram_str; /* DPRAM PCI controllers */ typedef struct { union { gdt_dpr_if ic; /* interface area */ - unchar if_area[0xff0-sizeof(gdt_pci_sram)]; + u8 if_area[0xff0-sizeof(gdt_pci_sram)]; } u; gdt_pci_sram gdt6sr; /* SRAM structure */ struct { - unchar unused0[1]; - unchar volatile Sema1; /* command semaphore */ - unchar unused1[3]; - unchar irqen; /* board interrupts enable */ - unchar unused2[2]; - unchar event; /* release event */ - unchar unused3[3]; - unchar irqdel; /* acknowledge board int. */ - unchar unused4[3]; - } PACKED io; -} PACKED gdt6_dpram_str; + u8 unused0[1]; + u8 volatile Sema1; /* command semaphore */ + u8 unused1[3]; + u8 irqen; /* board interrupts enable */ + u8 unused2[2]; + u8 event; /* release event */ + u8 unused3[3]; + u8 irqdel; /* acknowledge board int. */ + u8 unused4[3]; + } __attribute__((packed)) io; +} __attribute__((packed)) gdt6_dpram_str; /* PLX register structure (new PCI controllers) */ typedef struct { - unchar cfg_reg; /* DPRAM cfg.(2:below 1MB,0:anywhere)*/ - unchar unused1[0x3f]; - unchar volatile sema0_reg; /* command semaphore */ - unchar volatile sema1_reg; /* status semaphore */ - unchar unused2[2]; - ushort volatile status; /* command status */ - ushort service; /* service */ - ulong32 info[2]; /* additional info */ - unchar unused3[0x10]; - unchar ldoor_reg; /* PCI to local doorbell */ - unchar unused4[3]; - unchar volatile edoor_reg; /* local to PCI doorbell */ - unchar unused5[3]; - unchar control0; /* control0 register(unused) */ - unchar control1; /* board interrupts enable */ - unchar unused6[0x16]; -} PACKED gdt6c_plx_regs; + u8 cfg_reg; /* DPRAM cfg.(2:below 1MB,0:anywhere)*/ + u8 unused1[0x3f]; + u8 volatile sema0_reg; /* command semaphore */ + u8 volatile sema1_reg; /* status semaphore */ + u8 unused2[2]; + u16 volatile status; /* command status */ + u16 service; /* service */ + u32 info[2]; /* additional info */ + u8 unused3[0x10]; + u8 ldoor_reg; /* PCI to local doorbell */ + u8 unused4[3]; + u8 volatile edoor_reg; /* local to PCI doorbell */ + u8 unused5[3]; + u8 control0; /* control0 register(unused) */ + u8 control1; /* board interrupts enable */ + u8 unused6[0x16]; +} __attribute__((packed)) gdt6c_plx_regs; /* DPRAM new PCI controllers */ typedef struct { union { gdt_dpr_if ic; /* interface area */ - unchar if_area[0x4000-sizeof(gdt_pci_sram)]; + u8 if_area[0x4000-sizeof(gdt_pci_sram)]; } u; gdt_pci_sram gdt6sr; /* SRAM structure */ -} PACKED gdt6c_dpram_str; +} __attribute__((packed)) gdt6c_dpram_str; /* i960 register structure (PCI MPR controllers) */ typedef struct { - unchar unused1[16]; - unchar volatile sema0_reg; /* command semaphore */ - unchar unused2; - unchar volatile sema1_reg; /* status semaphore */ - unchar unused3; - ushort volatile status; /* command status */ - ushort service; /* service */ - ulong32 info[2]; /* additional info */ - unchar ldoor_reg; /* PCI to local doorbell */ - unchar unused4[11]; - unchar volatile edoor_reg; /* local to PCI doorbell */ - unchar unused5[7]; - unchar edoor_en_reg; /* board interrupts enable */ - unchar unused6[27]; - ulong32 unused7[939]; - ulong32 severity; + u8 unused1[16]; + u8 volatile sema0_reg; /* command semaphore */ + u8 unused2; + u8 volatile sema1_reg; /* status semaphore */ + u8 unused3; + u16 volatile status; /* command status */ + u16 service; /* service */ + u32 info[2]; /* additional info */ + u8 ldoor_reg; /* PCI to local doorbell */ + u8 unused4[11]; + u8 volatile edoor_reg; /* local to PCI doorbell */ + u8 unused5[7]; + u8 edoor_en_reg; /* board interrupts enable */ + u8 unused6[27]; + u32 unused7[939]; + u32 severity; char evt_str[256]; /* event string */ -} PACKED gdt6m_i960_regs; +} __attribute__((packed)) gdt6m_i960_regs; /* DPRAM PCI MPR controllers */ typedef struct { gdt6m_i960_regs i960r; /* 4KB i960 registers */ union { gdt_dpr_if ic; /* interface area */ - unchar if_area[0x3000-sizeof(gdt_pci_sram)]; + u8 if_area[0x3000-sizeof(gdt_pci_sram)]; } u; gdt_pci_sram gdt6sr; /* SRAM structure */ -} PACKED gdt6m_dpram_str; +} __attribute__((packed)) gdt6m_dpram_str; /* PCI resources */ typedef struct { struct pci_dev *pdev; - ulong dpmem; /* DPRAM address */ - ulong io; /* IO address */ + unsigned long dpmem; /* DPRAM address */ + unsigned long io; /* IO address */ } gdth_pci_str; @@ -846,93 +846,93 @@ typedef struct { typedef struct { struct Scsi_Host *shost; struct list_head list; - ushort hanum; - ushort oem_id; /* OEM */ - ushort type; /* controller class */ - ulong32 stype; /* subtype (PCI: device ID) */ - ushort fw_vers; /* firmware version */ - ushort cache_feat; /* feat. cache serv. (s/g,..)*/ - ushort raw_feat; /* feat. raw service (s/g,..)*/ - ushort screen_feat; /* feat. raw service (s/g,..)*/ - ushort bmic; /* BMIC address (EISA) */ + u16 hanum; + u16 oem_id; /* OEM */ + u16 type; /* controller class */ + u32 stype; /* subtype (PCI: device ID) */ + u16 fw_vers; /* firmware version */ + u16 cache_feat; /* feat. cache serv. (s/g,..)*/ + u16 raw_feat; /* feat. raw service (s/g,..)*/ + u16 screen_feat; /* feat. raw service (s/g,..)*/ + u16 bmic; /* BMIC address (EISA) */ void __iomem *brd; /* DPRAM address */ - ulong32 brd_phys; /* slot number/BIOS address */ + u32 brd_phys; /* slot number/BIOS address */ gdt6c_plx_regs *plx; /* PLX regs (new PCI contr.) */ gdth_cmd_str cmdext; gdth_cmd_str *pccb; /* address command structure */ - ulong32 ccb_phys; /* phys. address */ + u32 ccb_phys; /* phys. address */ #ifdef INT_COAL gdth_coal_status *coal_stat; /* buffer for coalescing int.*/ - ulong64 coal_stat_phys; /* phys. address */ + u64 coal_stat_phys; /* phys. address */ #endif char *pscratch; /* scratch (DMA) buffer */ - ulong64 scratch_phys; /* phys. address */ - unchar scratch_busy; /* in use? */ - unchar dma64_support; /* 64-bit DMA supported? */ + u64 scratch_phys; /* phys. address */ + u8 scratch_busy; /* in use? */ + u8 dma64_support; /* 64-bit DMA supported? */ gdth_msg_str *pmsg; /* message buffer */ - ulong64 msg_phys; /* phys. address */ - unchar scan_mode; /* current scan mode */ - unchar irq; /* IRQ */ - unchar drq; /* DRQ (ISA controllers) */ - ushort status; /* command status */ - ushort service; /* service/firmware ver./.. */ - ulong32 info; - ulong32 info2; /* additional info */ + u64 msg_phys; /* phys. address */ + u8 scan_mode; /* current scan mode */ + u8 irq; /* IRQ */ + u8 drq; /* DRQ (ISA controllers) */ + u16 status; /* command status */ + u16 service; /* service/firmware ver./.. */ + u32 info; + u32 info2; /* additional info */ Scsi_Cmnd *req_first; /* top of request queue */ struct { - unchar present; /* Flag: host drive present? */ - unchar is_logdrv; /* Flag: log. drive (master)? */ - unchar is_arraydrv; /* Flag: array drive? */ - unchar is_master; /* Flag: array drive master? */ - unchar is_parity; /* Flag: parity drive? */ - unchar is_hotfix; /* Flag: hotfix drive? */ - unchar master_no; /* number of master drive */ - unchar lock; /* drive locked? (hot plug) */ - unchar heads; /* mapping */ - unchar secs; - ushort devtype; /* further information */ - ulong64 size; /* capacity */ - unchar ldr_no; /* log. drive no. */ - unchar rw_attribs; /* r/w attributes */ - unchar cluster_type; /* cluster properties */ - unchar media_changed; /* Flag:MOUNT/UNMOUNT occured */ - ulong32 start_sec; /* start sector */ + u8 present; /* Flag: host drive present? */ + u8 is_logdrv; /* Flag: log. drive (master)? */ + u8 is_arraydrv; /* Flag: array drive? */ + u8 is_master; /* Flag: array drive master? */ + u8 is_parity; /* Flag: parity drive? */ + u8 is_hotfix; /* Flag: hotfix drive? */ + u8 master_no; /* number of master drive */ + u8 lock; /* drive locked? (hot plug) */ + u8 heads; /* mapping */ + u8 secs; + u16 devtype; /* further information */ + u64 size; /* capacity */ + u8 ldr_no; /* log. drive no. */ + u8 rw_attribs; /* r/w attributes */ + u8 cluster_type; /* cluster properties */ + u8 media_changed; /* Flag:MOUNT/UNMOUNT occured */ + u32 start_sec; /* start sector */ } hdr[MAX_LDRIVES]; /* host drives */ struct { - unchar lock; /* channel locked? (hot plug) */ - unchar pdev_cnt; /* physical device count */ - unchar local_no; /* local channel number */ - unchar io_cnt[MAXID]; /* current IO count */ - ulong32 address; /* channel address */ - ulong32 id_list[MAXID]; /* IDs of the phys. devices */ + u8 lock; /* channel locked? (hot plug) */ + u8 pdev_cnt; /* physical device count */ + u8 local_no; /* local channel number */ + u8 io_cnt[MAXID]; /* current IO count */ + u32 address; /* channel address */ + u32 id_list[MAXID]; /* IDs of the phys. devices */ } raw[MAXBUS]; /* SCSI channels */ struct { Scsi_Cmnd *cmnd; /* pending request */ - ushort service; /* service */ + u16 service; /* service */ } cmd_tab[GDTH_MAXCMDS]; /* table of pend. requests */ struct gdth_cmndinfo { /* per-command private info */ int index; int internal_command; /* don't call scsi_done */ gdth_cmd_str *internal_cmd_str; /* crier for internal messages*/ dma_addr_t sense_paddr; /* sense dma-addr */ - unchar priority; + u8 priority; int timeout_count; /* # of timeout calls */ volatile int wait_for_completion; - ushort status; - ulong32 info; + u16 status; + u32 info; enum dma_data_direction dma_dir; int phase; /* ???? */ int OpCode; } cmndinfo[GDTH_MAXCMDS]; /* index==0 is free */ - unchar bus_cnt; /* SCSI bus count */ - unchar tid_cnt; /* Target ID count */ - unchar bus_id[MAXBUS]; /* IOP IDs */ - unchar virt_bus; /* number of virtual bus */ - unchar more_proc; /* more /proc info supported */ - ushort cmd_cnt; /* command count in DPRAM */ - ushort cmd_len; /* length of actual command */ - ushort cmd_offs_dpmem; /* actual offset in DPRAM */ - ushort ic_all_size; /* sizeof DPRAM interf. area */ + u8 bus_cnt; /* SCSI bus count */ + u8 tid_cnt; /* Target ID count */ + u8 bus_id[MAXBUS]; /* IOP IDs */ + u8 virt_bus; /* number of virtual bus */ + u8 more_proc; /* more /proc info supported */ + u16 cmd_cnt; /* command count in DPRAM */ + u16 cmd_len; /* length of actual command */ + u16 cmd_offs_dpmem; /* actual offset in DPRAM */ + u16 ic_all_size; /* sizeof DPRAM interf. area */ gdth_cpar_str cpar; /* controller cache par. */ gdth_bfeat_str bfeat; /* controller features */ gdth_binfo_str binfo; /* controller info */ @@ -941,7 +941,7 @@ typedef struct { struct pci_dev *pdev; char oem_name[8]; #ifdef GDTH_DMA_STATISTICS - ulong dma32_cnt, dma64_cnt; /* statistics: DMA buffer */ + unsigned long dma32_cnt, dma64_cnt; /* statistics: DMA buffer */ #endif struct scsi_device *sdev; } gdth_ha_str; @@ -953,65 +953,65 @@ static inline struct gdth_cmndinfo *gdth_cmnd_priv(struct scsi_cmnd* cmd) /* INQUIRY data format */ typedef struct { - unchar type_qual; - unchar modif_rmb; - unchar version; - unchar resp_aenc; - unchar add_length; - unchar reserved1; - unchar reserved2; - unchar misc; - unchar vendor[8]; - unchar product[16]; - unchar revision[4]; -} PACKED gdth_inq_data; + u8 type_qual; + u8 modif_rmb; + u8 version; + u8 resp_aenc; + u8 add_length; + u8 reserved1; + u8 reserved2; + u8 misc; + u8 vendor[8]; + u8 product[16]; + u8 revision[4]; +} __attribute__((packed)) gdth_inq_data; /* READ_CAPACITY data format */ typedef struct { - ulong32 last_block_no; - ulong32 block_length; -} PACKED gdth_rdcap_data; + u32 last_block_no; + u32 block_length; +} __attribute__((packed)) gdth_rdcap_data; /* READ_CAPACITY (16) data format */ typedef struct { - ulong64 last_block_no; - ulong32 block_length; -} PACKED gdth_rdcap16_data; + u64 last_block_no; + u32 block_length; +} __attribute__((packed)) gdth_rdcap16_data; /* REQUEST_SENSE data format */ typedef struct { - unchar errorcode; - unchar segno; - unchar key; - ulong32 info; - unchar add_length; - ulong32 cmd_info; - unchar adsc; - unchar adsq; - unchar fruc; - unchar key_spec[3]; -} PACKED gdth_sense_data; + u8 errorcode; + u8 segno; + u8 key; + u32 info; + u8 add_length; + u32 cmd_info; + u8 adsc; + u8 adsq; + u8 fruc; + u8 key_spec[3]; +} __attribute__((packed)) gdth_sense_data; /* MODE_SENSE data format */ typedef struct { struct { - unchar data_length; - unchar med_type; - unchar dev_par; - unchar bd_length; - } PACKED hd; + u8 data_length; + u8 med_type; + u8 dev_par; + u8 bd_length; + } __attribute__((packed)) hd; struct { - unchar dens_code; - unchar block_count[3]; - unchar reserved; - unchar block_length[3]; - } PACKED bd; -} PACKED gdth_modep_data; + u8 dens_code; + u8 block_count[3]; + u8 reserved; + u8 block_length[3]; + } __attribute__((packed)) bd; +} __attribute__((packed)) gdth_modep_data; /* stack frame */ typedef struct { - ulong b[10]; /* 32/64 bit compiler ! */ -} PACKED gdth_stackframe; + unsigned long b[10]; /* 32/64 bit compiler ! */ +} __attribute__((packed)) gdth_stackframe; /* function prototyping */ diff --git a/drivers/scsi/gdth_ioctl.h b/drivers/scsi/gdth_ioctl.h index 783fae7..b004c61 100644 --- a/drivers/scsi/gdth_ioctl.h +++ b/drivers/scsi/gdth_ioctl.h @@ -32,109 +32,101 @@ #define MAX_HDRIVES MAX_LDRIVES /* max. host drive count */ #endif -/* typedefs */ -#ifdef __KERNEL__ -typedef u32 ulong32; -typedef u64 ulong64; -#endif - -#define PACKED __attribute__((packed)) - /* scatter/gather element */ typedef struct { - ulong32 sg_ptr; /* address */ - ulong32 sg_len; /* length */ -} PACKED gdth_sg_str; + u32 sg_ptr; /* address */ + u32 sg_len; /* length */ +} __attribute__((packed)) gdth_sg_str; /* scatter/gather element - 64bit addresses */ typedef struct { - ulong64 sg_ptr; /* address */ - ulong32 sg_len; /* length */ -} PACKED gdth_sg64_str; + u64 sg_ptr; /* address */ + u32 sg_len; /* length */ +} __attribute__((packed)) gdth_sg64_str; /* command structure */ typedef struct { - ulong32 BoardNode; /* board node (always 0) */ - ulong32 CommandIndex; /* command number */ - ushort OpCode; /* the command (READ,..) */ + u32 BoardNode; /* board node (always 0) */ + u32 CommandIndex; /* command number */ + u16 OpCode; /* the command (READ,..) */ union { struct { - ushort DeviceNo; /* number of cache drive */ - ulong32 BlockNo; /* block number */ - ulong32 BlockCnt; /* block count */ - ulong32 DestAddr; /* dest. addr. (if s/g: -1) */ - ulong32 sg_canz; /* s/g element count */ + u16 DeviceNo; /* number of cache drive */ + u32 BlockNo; /* block number */ + u32 BlockCnt; /* block count */ + u32 DestAddr; /* dest. addr. (if s/g: -1) */ + u32 sg_canz; /* s/g element count */ gdth_sg_str sg_lst[GDTH_MAXSG]; /* s/g list */ - } PACKED cache; /* cache service cmd. str. */ + } __attribute__((packed)) cache; /* cache service cmd. str. */ struct { - ushort DeviceNo; /* number of cache drive */ - ulong64 BlockNo; /* block number */ - ulong32 BlockCnt; /* block count */ - ulong64 DestAddr; /* dest. addr. (if s/g: -1) */ - ulong32 sg_canz; /* s/g element count */ + u16 DeviceNo; /* number of cache drive */ + u64 BlockNo; /* block number */ + u32 BlockCnt; /* block count */ + u64 DestAddr; /* dest. addr. (if s/g: -1) */ + u32 sg_canz; /* s/g element count */ gdth_sg64_str sg_lst[GDTH_MAXSG]; /* s/g list */ - } PACKED cache64; /* cache service cmd. str. */ + } __attribute__((packed)) cache64; /* cache service cmd. str. */ struct { - ushort param_size; /* size of p_param buffer */ - ulong32 subfunc; /* IOCTL function */ - ulong32 channel; /* device */ - ulong64 p_param; /* buffer */ - } PACKED ioctl; /* IOCTL command structure */ + u16 param_size; /* size of p_param buffer */ + u32 subfunc; /* IOCTL function */ + u32 channel; /* device */ + u64 p_param; /* buffer */ + } __attribute__((packed)) ioctl; /* IOCTL command structure */ struct { - ushort reserved; + u16 reserved; union { struct { - ulong32 msg_handle; /* message handle */ - ulong64 msg_addr; /* message buffer address */ - } PACKED msg; - unchar data[12]; /* buffer for rtc data, ... */ + u32 msg_handle; /* message handle */ + u64 msg_addr; /* message buffer address */ + } __attribute__((packed)) msg; + u8 data[12]; /* buffer for rtc data, ... */ } su; - } PACKED screen; /* screen service cmd. str. */ + } __attribute__((packed)) screen; /* screen service cmd. str. */ struct { - ushort reserved; - ulong32 direction; /* data direction */ - ulong32 mdisc_time; /* disc. time (0: no timeout)*/ - ulong32 mcon_time; /* connect time(0: no to.) */ - ulong32 sdata; /* dest. addr. (if s/g: -1) */ - ulong32 sdlen; /* data length (bytes) */ - ulong32 clen; /* SCSI cmd. length(6,10,12) */ - unchar cmd[12]; /* SCSI command */ - unchar target; /* target ID */ - unchar lun; /* LUN */ - unchar bus; /* SCSI bus number */ - unchar priority; /* only 0 used */ - ulong32 sense_len; /* sense data length */ - ulong32 sense_data; /* sense data addr. */ - ulong32 link_p; /* linked cmds (not supp.) */ - ulong32 sg_ranz; /* s/g element count */ + u16 reserved; + u32 direction; /* data direction */ + u32 mdisc_time; /* disc. time (0: no timeout)*/ + u32 mcon_time; /* connect time(0: no to.) */ + u32 sdata; /* dest. addr. (if s/g: -1) */ + u32 sdlen; /* data length (bytes) */ + u32 clen; /* SCSI cmd. length(6,10,12) */ + u8 cmd[12]; /* SCSI command */ + u8 target; /* target ID */ + u8 lun; /* LUN */ + u8 bus; /* SCSI bus number */ + u8 priority; /* only 0 used */ + u32 sense_len; /* sense data length */ + u32 sense_data; /* sense data addr. */ + u32 link_p; /* linked cmds (not supp.) */ + u32 sg_ranz; /* s/g element count */ gdth_sg_str sg_lst[GDTH_MAXSG]; /* s/g list */ - } PACKED raw; /* raw service cmd. struct. */ + } __attribute__((packed)) raw; /* raw service cmd. struct. */ struct { - ushort reserved; - ulong32 direction; /* data direction */ - ulong32 mdisc_time; /* disc. time (0: no timeout)*/ - ulong32 mcon_time; /* connect time(0: no to.) */ - ulong64 sdata; /* dest. addr. (if s/g: -1) */ - ulong32 sdlen; /* data length (bytes) */ - ulong32 clen; /* SCSI cmd. length(6,..,16) */ - unchar cmd[16]; /* SCSI command */ - unchar target; /* target ID */ - unchar lun; /* LUN */ - unchar bus; /* SCSI bus number */ - unchar priority; /* only 0 used */ - ulong32 sense_len; /* sense data length */ - ulong64 sense_data; /* sense data addr. */ - ulong32 sg_ranz; /* s/g element count */ + u16 reserved; + u32 direction; /* data direction */ + u32 mdisc_time; /* disc. time (0: no timeout)*/ + u32 mcon_time; /* connect time(0: no to.) */ + u64 sdata; /* dest. addr. (if s/g: -1) */ + u32 sdlen; /* data length (bytes) */ + u32 clen; /* SCSI cmd. length(6,..,16) */ + u8 cmd[16]; /* SCSI command */ + u8 target; /* target ID */ + u8 lun; /* LUN */ + u8 bus; /* SCSI bus number */ + u8 priority; /* only 0 used */ + u32 sense_len; /* sense data length */ + u64 sense_data; /* sense data addr. */ + u32 sg_ranz; /* s/g element count */ gdth_sg64_str sg_lst[GDTH_MAXSG]; /* s/g list */ - } PACKED raw64; /* raw service cmd. struct. */ + } __attribute__((packed)) raw64; /* raw service cmd. struct. */ } u; /* additional variables */ - unchar Service; /* controller service */ - unchar reserved; - ushort Status; /* command result */ - ulong32 Info; /* additional information */ + u8 Service; /* controller service */ + u8 reserved; + u16 Status; /* command result */ + u32 Info; /* additional information */ void *RequestBuffer; /* request buffer */ -} PACKED gdth_cmd_str; +} __attribute__((packed)) gdth_cmd_str; /* controller event structure */ #define ES_ASYNC 1 @@ -142,129 +134,129 @@ typedef struct { #define ES_TEST 3 #define ES_SYNC 4 typedef struct { - ushort size; /* size of structure */ + u16 size; /* size of structure */ union { char stream[16]; struct { - ushort ionode; - ushort service; - ulong32 index; - } PACKED driver; + u16 ionode; + u16 service; + u32 index; + } __attribute__((packed)) driver; struct { - ushort ionode; - ushort service; - ushort status; - ulong32 info; - unchar scsi_coord[3]; - } PACKED async; + u16 ionode; + u16 service; + u16 status; + u32 info; + u8 scsi_coord[3]; + } __attribute__((packed)) async; struct { - ushort ionode; - ushort service; - ushort status; - ulong32 info; - ushort hostdrive; - unchar scsi_coord[3]; - unchar sense_key; - } PACKED sync; + u16 ionode; + u16 service; + u16 status; + u32 info; + u16 hostdrive; + u8 scsi_coord[3]; + u8 sense_key; + } __attribute__((packed)) sync; struct { - ulong32 l1, l2, l3, l4; - } PACKED test; + u32 l1, l2, l3, l4; + } __attribute__((packed)) test; } eu; - ulong32 severity; - unchar event_string[256]; -} PACKED gdth_evt_data; + u32 severity; + u8 event_string[256]; +} __attribute__((packed)) gdth_evt_data; typedef struct { - ulong32 first_stamp; - ulong32 last_stamp; - ushort same_count; - ushort event_source; - ushort event_idx; - unchar application; - unchar reserved; + u32 first_stamp; + u32 last_stamp; + u16 same_count; + u16 event_source; + u16 event_idx; + u8 application; + u8 reserved; gdth_evt_data event_data; -} PACKED gdth_evt_str; +} __attribute__((packed)) gdth_evt_str; #ifdef GDTH_IOCTL_PROC /* IOCTL structure (write) */ typedef struct { - ulong32 magic; /* IOCTL magic */ - ushort ioctl; /* IOCTL */ - ushort ionode; /* controller number */ - ushort service; /* controller service */ - ushort timeout; /* timeout */ + u32 magic; /* IOCTL magic */ + u16 ioctl; /* IOCTL */ + u16 ionode; /* controller number */ + u16 service; /* controller service */ + u16 timeout; /* timeout */ union { struct { - unchar command[512]; /* controller command */ - unchar data[1]; /* add. data */ + u8 command[512]; /* controller command */ + u8 data[1]; /* add. data */ } general; struct { - unchar lock; /* lock/unlock */ - unchar drive_cnt; /* drive count */ - ushort drives[MAX_HDRIVES];/* drives */ + u8 lock; /* lock/unlock */ + u8 drive_cnt; /* drive count */ + u16 drives[MAX_HDRIVES];/* drives */ } lockdrv; struct { - unchar lock; /* lock/unlock */ - unchar channel; /* channel */ + u8 lock; /* lock/unlock */ + u8 channel; /* channel */ } lockchn; struct { int erase; /* erase event ? */ int handle; - unchar evt[EVENT_SIZE]; /* event structure */ + u8 evt[EVENT_SIZE]; /* event structure */ } event; struct { - unchar bus; /* SCSI bus */ - unchar target; /* target ID */ - unchar lun; /* LUN */ - unchar cmd_len; /* command length */ - unchar cmd[12]; /* SCSI command */ + u8 bus; /* SCSI bus */ + u8 target; /* target ID */ + u8 lun; /* LUN */ + u8 cmd_len; /* command length */ + u8 cmd[12]; /* SCSI command */ } scsi; struct { - ushort hdr_no; /* host drive number */ - unchar flag; /* old meth./add/remove */ + u16 hdr_no; /* host drive number */ + u8 flag; /* old meth./add/remove */ } rescan; } iu; } gdth_iowr_str; /* IOCTL structure (read) */ typedef struct { - ulong32 size; /* buffer size */ - ulong32 status; /* IOCTL error code */ + u32 size; /* buffer size */ + u32 status; /* IOCTL error code */ union { struct { - unchar data[1]; /* data */ + u8 data[1]; /* data */ } general; struct { - ushort version; /* driver version */ + u16 version; /* driver version */ } drvers; struct { - unchar type; /* controller type */ - ushort info; /* slot etc. */ - ushort oem_id; /* OEM ID */ - ushort bios_ver; /* not used */ - ushort access; /* not used */ - ushort ext_type; /* extended type */ - ushort device_id; /* device ID */ - ushort sub_device_id; /* sub device ID */ + u8 type; /* controller type */ + u16 info; /* slot etc. */ + u16 oem_id; /* OEM ID */ + u16 bios_ver; /* not used */ + u16 access; /* not used */ + u16 ext_type; /* extended type */ + u16 device_id; /* device ID */ + u16 sub_device_id; /* sub device ID */ } ctrtype; struct { - unchar version; /* OS version */ - unchar subversion; /* OS subversion */ - ushort revision; /* revision */ + u8 version; /* OS version */ + u8 subversion; /* OS subversion */ + u16 revision; /* revision */ } osvers; struct { - ushort count; /* controller count */ + u16 count; /* controller count */ } ctrcnt; struct { int handle; - unchar evt[EVENT_SIZE]; /* event structure */ + u8 evt[EVENT_SIZE]; /* event structure */ } event; struct { - unchar bus; /* SCSI bus, 0xff: invalid */ - unchar target; /* target ID */ - unchar lun; /* LUN */ - unchar cluster_type; /* cluster properties */ + u8 bus; /* SCSI bus, 0xff: invalid */ + u8 target; /* target ID */ + u8 lun; /* LUN */ + u8 cluster_type; /* cluster properties */ } hdr_list[MAX_HDRIVES]; /* index is host drive number */ } iu; } gdth_iord_str; @@ -272,53 +264,53 @@ typedef struct { /* GDTIOCTL_GENERAL */ typedef struct { - ushort ionode; /* controller number */ - ushort timeout; /* timeout */ - ulong32 info; /* error info */ - ushort status; /* status */ - ulong data_len; /* data buffer size */ - ulong sense_len; /* sense buffer size */ + u16 ionode; /* controller number */ + u16 timeout; /* timeout */ + u32 info; /* error info */ + u16 status; /* status */ + unsigned long data_len; /* data buffer size */ + unsigned long sense_len; /* sense buffer size */ gdth_cmd_str command; /* command */ } gdth_ioctl_general; /* GDTIOCTL_LOCKDRV */ typedef struct { - ushort ionode; /* controller number */ - unchar lock; /* lock/unlock */ - unchar drive_cnt; /* drive count */ - ushort drives[MAX_HDRIVES]; /* drives */ + u16 ionode; /* controller number */ + u8 lock; /* lock/unlock */ + u8 drive_cnt; /* drive count */ + u16 drives[MAX_HDRIVES]; /* drives */ } gdth_ioctl_lockdrv; /* GDTIOCTL_LOCKCHN */ typedef struct { - ushort ionode; /* controller number */ - unchar lock; /* lock/unlock */ - unchar channel; /* channel */ + u16 ionode; /* controller number */ + u8 lock; /* lock/unlock */ + u8 channel; /* channel */ } gdth_ioctl_lockchn; /* GDTIOCTL_OSVERS */ typedef struct { - unchar version; /* OS version */ - unchar subversion; /* OS subversion */ - ushort revision; /* revision */ + u8 version; /* OS version */ + u8 subversion; /* OS subversion */ + u16 revision; /* revision */ } gdth_ioctl_osvers; /* GDTIOCTL_CTRTYPE */ typedef struct { - ushort ionode; /* controller number */ - unchar type; /* controller type */ - ushort info; /* slot etc. */ - ushort oem_id; /* OEM ID */ - ushort bios_ver; /* not used */ - ushort access; /* not used */ - ushort ext_type; /* extended type */ - ushort device_id; /* device ID */ - ushort sub_device_id; /* sub device ID */ + u16 ionode; /* controller number */ + u8 type; /* controller type */ + u16 info; /* slot etc. */ + u16 oem_id; /* OEM ID */ + u16 bios_ver; /* not used */ + u16 access; /* not used */ + u16 ext_type; /* extended type */ + u16 device_id; /* device ID */ + u16 sub_device_id; /* sub device ID */ } gdth_ioctl_ctrtype; /* GDTIOCTL_EVENT */ typedef struct { - ushort ionode; + u16 ionode; int erase; /* erase event? */ int handle; /* event handle */ gdth_evt_str event; @@ -326,22 +318,22 @@ typedef struct { /* GDTIOCTL_RESCAN/GDTIOCTL_HDRLIST */ typedef struct { - ushort ionode; /* controller number */ - unchar flag; /* add/remove */ - ushort hdr_no; /* drive no. */ + u16 ionode; /* controller number */ + u8 flag; /* add/remove */ + u16 hdr_no; /* drive no. */ struct { - unchar bus; /* SCSI bus */ - unchar target; /* target ID */ - unchar lun; /* LUN */ - unchar cluster_type; /* cluster properties */ + u8 bus; /* SCSI bus */ + u8 target; /* target ID */ + u8 lun; /* LUN */ + u8 cluster_type; /* cluster properties */ } hdr_list[MAX_HDRIVES]; /* index is host drive number */ } gdth_ioctl_rescan; /* GDTIOCTL_RESET_BUS/GDTIOCTL_RESET_DRV */ typedef struct { - ushort ionode; /* controller number */ - ushort number; /* bus/host drive number */ - ushort status; /* status */ + u16 ionode; /* controller number */ + u16 number; /* bus/host drive number */ + u16 status; /* status */ } gdth_ioctl_reset; #endif diff --git a/drivers/scsi/gdth_proc.c b/drivers/scsi/gdth_proc.c index 1258da3..ffb2b21 100644 --- a/drivers/scsi/gdth_proc.c +++ b/drivers/scsi/gdth_proc.c @@ -43,7 +43,7 @@ static int gdth_set_asc_info(struct Scsi_Host *host, char *buffer, int i, found; gdth_cmd_str gdtcmd; gdth_cpar_str *pcpar; - ulong64 paddr; + u64 paddr; char cmnd[MAX_COMMAND_SIZE]; memset(cmnd, 0xff, 12); @@ -156,8 +156,8 @@ static int gdth_get_info(char *buffer,char **start,off_t offset,int length, off_t begin = 0,pos = 0; int id, i, j, k, sec, flag; int no_mdrv = 0, drv_no, is_mirr; - ulong32 cnt; - ulong64 paddr; + u32 cnt; + u64 paddr; int rc = -ENOMEM; gdth_cmd_str *gdtcmd; @@ -220,14 +220,14 @@ static int gdth_get_info(char *buffer,char **start,off_t offset,int length, if (ha->more_proc) sprintf(hrec, "%d.%02d.%02d-%c%03X", - (unchar)(ha->binfo.upd_fw_ver>>24), - (unchar)(ha->binfo.upd_fw_ver>>16), - (unchar)(ha->binfo.upd_fw_ver), + (u8)(ha->binfo.upd_fw_ver>>24), + (u8)(ha->binfo.upd_fw_ver>>16), + (u8)(ha->binfo.upd_fw_ver), ha->bfeat.raid ? 'R':'N', ha->binfo.upd_revision); else - sprintf(hrec, "%d.%02d", (unchar)(ha->cpar.version>>8), - (unchar)(ha->cpar.version)); + sprintf(hrec, "%d.%02d", (u8)(ha->cpar.version>>8), + (u8)(ha->cpar.version)); size = sprintf(buffer+len, " Driver Ver.: \t%-10s\tFirmware Ver.: \t%s\n", @@ -281,7 +281,7 @@ static int gdth_get_info(char *buffer,char **start,off_t offset,int length, pds->bid = ha->raw[i].local_no; pds->first = 0; pds->entries = ha->raw[i].pdev_cnt; - cnt = (3*GDTH_SCRATCH/4 - 5 * sizeof(ulong32)) / + cnt = (3*GDTH_SCRATCH/4 - 5 * sizeof(u32)) / sizeof(pds->list[0]); if (pds->entries > cnt) pds->entries = cnt; @@ -604,7 +604,7 @@ static int gdth_get_info(char *buffer,char **start,off_t offset,int length, size = sprintf(buffer+len, " Capacity [MB]:\t%-6d \tStart Sector: \t%d\n", - (ulong32)(ha->hdr[i].size/2048), ha->hdr[i].start_sec); + (u32)(ha->hdr[i].size/2048), ha->hdr[i].start_sec); len += size; pos = begin + len; if (pos < offset) { len = 0; @@ -664,9 +664,9 @@ free_fail: } static char *gdth_ioctl_alloc(gdth_ha_str *ha, int size, int scratch, - ulong64 *paddr) + u64 *paddr) { - ulong flags; + unsigned long flags; char *ret_val; if (size == 0) @@ -691,9 +691,9 @@ static char *gdth_ioctl_alloc(gdth_ha_str *ha, int size, int scratch, return ret_val; } -static void gdth_ioctl_free(gdth_ha_str *ha, int size, char *buf, ulong64 paddr) +static void gdth_ioctl_free(gdth_ha_str *ha, int size, char *buf, u64 paddr) { - ulong flags; + unsigned long flags; if (buf == ha->pscratch) { spin_lock_irqsave(&ha->smp_lock, flags); @@ -705,16 +705,16 @@ static void gdth_ioctl_free(gdth_ha_str *ha, int size, char *buf, ulong64 paddr) } #ifdef GDTH_IOCTL_PROC -static int gdth_ioctl_check_bin(gdth_ha_str *ha, ushort size) +static int gdth_ioctl_check_bin(gdth_ha_str *ha, u16 size) { - ulong flags; + unsigned long flags; int ret_val; spin_lock_irqsave(&ha->smp_lock, flags); ret_val = FALSE; if (ha->scratch_busy) { - if (((gdth_iord_str *)ha->pscratch)->size == (ulong32)size) + if (((gdth_iord_str *)ha->pscratch)->size == (u32)size) ret_val = TRUE; } spin_unlock_irqrestore(&ha->smp_lock, flags); @@ -724,11 +724,11 @@ static int gdth_ioctl_check_bin(gdth_ha_str *ha, ushort size) static void gdth_wait_completion(gdth_ha_str *ha, int busnum, int id) { - ulong flags; + unsigned long flags; int i; Scsi_Cmnd *scp; struct gdth_cmndinfo *cmndinfo; - unchar b, t; + u8 b, t; spin_lock_irqsave(&ha->smp_lock, flags); @@ -738,8 +738,8 @@ static void gdth_wait_completion(gdth_ha_str *ha, int busnum, int id) b = scp->device->channel; t = scp->device->id; - if (!SPECIAL_SCP(scp) && t == (unchar)id && - b == (unchar)busnum) { + if (!SPECIAL_SCP(scp) && t == (u8)id && + b == (u8)busnum) { cmndinfo->wait_for_completion = 0; spin_unlock_irqrestore(&ha->smp_lock, flags); while (!cmndinfo->wait_for_completion) diff --git a/drivers/scsi/gdth_proc.h b/drivers/scsi/gdth_proc.h index 9b900cc..dab15f5 100644 --- a/drivers/scsi/gdth_proc.h +++ b/drivers/scsi/gdth_proc.h @@ -17,8 +17,8 @@ static int gdth_set_asc_info(struct Scsi_Host *host, char *buffer, int length, gdth_ha_str *ha); static char *gdth_ioctl_alloc(gdth_ha_str *ha, int size, int scratch, - ulong64 *paddr); -static void gdth_ioctl_free(gdth_ha_str *ha, int size, char *buf, ulong64 paddr); + u64 *paddr); +static void gdth_ioctl_free(gdth_ha_str *ha, int size, char *buf, u64 paddr); static void gdth_wait_completion(gdth_ha_str *ha, int busnum, int id); #endif -- cgit v0.10.2 From 7da5087971b1a187f92be4efb74a991ac9ccb0a3 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:04:12 +0530 Subject: [SCSI] be2iscsi: Use start cid and number of cid and icd from FW This patch enablesi be2iscsi to use the start number and number of cids/icd provided by FW rather than hard coded values. Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_iscsi.c b/drivers/scsi/be2iscsi/be_iscsi.c index d587b03..2b3c58b 100644 --- a/drivers/scsi/be2iscsi/be_iscsi.c +++ b/drivers/scsi/be2iscsi/be_iscsi.c @@ -431,9 +431,10 @@ static int beiscsi_open_conn(struct iscsi_endpoint *ep, } SE_DEBUG(DBG_LVL_8, "In beiscsi_open_conn, ep_cid=%d ", beiscsi_ep->ep_cid); - phba->ep_array[beiscsi_ep->ep_cid] = ep; - if (beiscsi_ep->ep_cid > - (phba->fw_config.iscsi_cid_start + phba->params.cxns_per_ctrl)) { + phba->ep_array[beiscsi_ep->ep_cid - + phba->fw_config.iscsi_cid_start] = ep; + if (beiscsi_ep->ep_cid > (phba->fw_config.iscsi_cid_start + + phba->params.cxns_per_ctrl * 2)) { SE_DEBUG(DBG_LVL_1, "Failed in allocate iscsi cid\n"); return ret; } diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 1a557fa..6c512b6 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -230,29 +230,27 @@ static int be_ctrl_init(struct beiscsi_hba *phba, struct pci_dev *pdev) static void beiscsi_get_params(struct beiscsi_hba *phba) { - phba->params.ios_per_ctrl = BE2_IO_DEPTH; - phba->params.cxns_per_ctrl = BE2_MAX_SESSIONS; - phba->params.asyncpdus_per_ctrl = BE2_ASYNCPDUS; - phba->params.icds_per_ctrl = BE2_MAX_ICDS / 2; + phba->params.ios_per_ctrl = (phba->fw_config.iscsi_icd_count + - (phba->fw_config.iscsi_cid_count + + BE2_TMFS + + BE2_NOPOUT_REQ)); + phba->params.cxns_per_ctrl = phba->fw_config.iscsi_cid_count; + phba->params.asyncpdus_per_ctrl = phba->fw_config.iscsi_cid_count;; + phba->params.icds_per_ctrl = phba->fw_config.iscsi_icd_count;; phba->params.num_sge_per_io = BE2_SGE; phba->params.defpdu_hdr_sz = BE2_DEFPDU_HDR_SZ; phba->params.defpdu_data_sz = BE2_DEFPDU_DATA_SZ; phba->params.eq_timer = 64; phba->params.num_eq_entries = - (((BE2_CMDS_PER_CXN * 2 + BE2_LOGOUTS + BE2_TMFS + BE2_ASYNCPDUS) / - 512) + 1) * 512; + (((BE2_CMDS_PER_CXN * 2 + phba->fw_config.iscsi_cid_count * 2 + + BE2_TMFS) / 512) + 1) * 512; phba->params.num_eq_entries = (phba->params.num_eq_entries < 1024) ? 1024 : phba->params.num_eq_entries; SE_DEBUG(DBG_LVL_8, "phba->params.num_eq_entries=%d \n", - phba->params.num_eq_entries); + phba->params.num_eq_entries); phba->params.num_cq_entries = - (((BE2_CMDS_PER_CXN * 2 + BE2_LOGOUTS + BE2_TMFS + BE2_ASYNCPDUS) / - 512) + 1) * 512; - SE_DEBUG(DBG_LVL_8, - "phba->params.num_cq_entries=%d BE2_CMDS_PER_CXN=%d" - "BE2_LOGOUTS=%d BE2_TMFS=%d BE2_ASYNCPDUS=%d \n", - phba->params.num_cq_entries, BE2_CMDS_PER_CXN, - BE2_LOGOUTS, BE2_TMFS, BE2_ASYNCPDUS); + (((BE2_CMDS_PER_CXN * 2 + phba->fw_config.iscsi_cid_count * 2 + + BE2_TMFS) / 512) + 1) * 512; phba->params.wrbs_per_cxn = 256; } @@ -877,7 +875,8 @@ hwi_complete_drvr_msgs(struct beiscsi_conn *beiscsi_conn, } else { pwrb_context = &phwi_ctrlr->wrb_context[((psol-> dw[offsetof(struct amap_sol_cqe, cid) / 32] & - SOL_CID_MASK) >> 6)]; + SOL_CID_MASK) >> 6) - + phba->fw_config.iscsi_cid_start]; pwrb_handle = pwrb_context->pwrb_handle_basestd[((psol-> dw[offsetof(struct amap_sol_cqe, wrb_index) / 32] & SOL_WRB_INDEX_MASK) >> 16)]; @@ -939,7 +938,8 @@ static void hwi_complete_cmd(struct beiscsi_conn *beiscsi_conn, pwrb_context = &phwi_ctrlr-> wrb_context[((psol->dw[offsetof (struct amap_sol_cqe, cid) / 32] - & SOL_CID_MASK) >> 6)]; + & SOL_CID_MASK) >> 6) - + phba->fw_config.iscsi_cid_start]; pwrb_handle = pwrb_context->pwrb_handle_basestd[((psol-> dw[offsetof(struct amap_sol_cqe, wrb_index) / 32] & SOL_WRB_INDEX_MASK) >> 16)]; @@ -1077,7 +1077,8 @@ hwi_get_async_handle(struct beiscsi_hba *phba, WARN_ON(!pasync_handle); - pasync_handle->cri = (unsigned short)beiscsi_conn->beiscsi_conn_cid; + pasync_handle->cri = (unsigned short)beiscsi_conn->beiscsi_conn_cid - + phba->fw_config.iscsi_cid_start; pasync_handle->is_header = is_header; pasync_handle->buffer_len = ((pdpdu_cqe-> dw[offsetof(struct amap_i_t_dpdu_cqe, dpl) / 32] @@ -1327,9 +1328,10 @@ hwi_fwd_async_msg(struct beiscsi_conn *beiscsi_conn, } status = beiscsi_process_async_pdu(beiscsi_conn, phba, - beiscsi_conn->beiscsi_conn_cid, - phdr, hdr_len, pfirst_buffer, - buf_len); + (beiscsi_conn->beiscsi_conn_cid - + phba->fw_config.iscsi_cid_start), + phdr, hdr_len, pfirst_buffer, + buf_len); if (status == 0) hwi_free_async_msg(phba, cri); @@ -1456,10 +1458,10 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) } } else { - beiscsi_conn = phba->conn_table[(u32) (sol-> + beiscsi_conn = phba->conn_table[(u32) ((sol-> dw[offsetof(struct amap_sol_cqe, cid) / 32] & - SOL_CID_MASK) >> 6]; - + SOL_CID_MASK) >> 6) - + phba->fw_config.iscsi_cid_start]; if (!beiscsi_conn || !beiscsi_conn->ep) { shost_printk(KERN_WARNING, phba->shost, "Connection table empty for cid = %d\n", @@ -1557,8 +1559,8 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) "0x%x...\n", sol->dw[offsetof(struct amap_sol_cqe, code) / 32] & CQE_CODE_MASK, - sol->dw[offsetof(struct amap_sol_cqe, cid) / - 32] & CQE_CID_MASK); + (sol->dw[offsetof(struct amap_sol_cqe, cid) / + 32] & CQE_CID_MASK)); } iscsi_conn_failure(beiscsi_conn->conn, ISCSI_ERR_CONN_FAILED); @@ -1575,8 +1577,8 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) "received/sent on CID 0x%x...\n", sol->dw[offsetof(struct amap_sol_cqe, code) / 32] & CQE_CODE_MASK, - sol->dw[offsetof(struct amap_sol_cqe, cid) / - 32] & CQE_CID_MASK); + (sol->dw[offsetof(struct amap_sol_cqe, cid) / + 32] & CQE_CID_MASK)); } iscsi_conn_failure(beiscsi_conn->conn, ISCSI_ERR_CONN_FAILED); @@ -1586,8 +1588,8 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) "received on CID 0x%x...\n", sol->dw[offsetof(struct amap_sol_cqe, code) / 32] & CQE_CODE_MASK, - sol->dw[offsetof(struct amap_sol_cqe, cid) / - 32] & CQE_CID_MASK); + (sol->dw[offsetof(struct amap_sol_cqe, cid) / + 32] & CQE_CID_MASK)); break; } @@ -2383,7 +2385,7 @@ static int beiscsi_create_cqs(struct beiscsi_hba *phba, &paddr); if (!cq_vaddress) goto create_cq_error; - ret = be_fill_queue(cq, phba->params.icds_per_ctrl / 2, + ret = be_fill_queue(cq, phba->params.num_cq_entries, sizeof(struct sol_cqe), cq_vaddress); if (ret) { shost_printk(KERN_ERR, phba->shost, @@ -2634,7 +2636,8 @@ beiscsi_create_wrb_rings(struct beiscsi_hba *phba, "wrbq create failed."); return status; } - phwi_ctrlr->wrb_context[i].cid = phwi_context->be_wrbq[i].id; + phwi_ctrlr->wrb_context[i * 2].cid = phwi_context->be_wrbq[i]. + id; } kfree(pwrb_arr); return 0; @@ -2807,12 +2810,6 @@ static int hwi_init_port(struct beiscsi_hba *phba) ring_mode = 1; else ring_mode = 0; - status = mgmt_get_fw_config(ctrl, phba); - if (status != 0) { - shost_printk(KERN_ERR, phba->shost, - "Error getting fw config\n"); - goto error; - } status = beiscsi_create_cqs(phba, phwi_context); if (status != 0) { @@ -3032,7 +3029,7 @@ static int beiscsi_init_sgl_handle(struct beiscsi_hba *phba) AMAP_SET_BITS(struct amap_iscsi_sge, addr_lo, pfrag, 0); pfrag += phba->params.num_sge_per_io; psgl_handle->sgl_index = - phba->fw_config.iscsi_cid_start + arr_index++; + phba->fw_config.iscsi_icd_start + arr_index++; } idx++; } @@ -3064,7 +3061,7 @@ static int hba_setup_cid_tbls(struct beiscsi_hba *phba) kfree(phba->cid_array); return -ENOMEM; } - new_cid = phba->fw_config.iscsi_icd_start; + new_cid = phba->fw_config.iscsi_cid_start; for (i = 0; i < phba->params.cxns_per_ctrl; i++) { phba->cid_array[i] = new_cid; new_cid += 2; @@ -3219,7 +3216,8 @@ beiscsi_offload_connection(struct beiscsi_conn *beiscsi_conn, * We can always use 0 here because it is reserved by libiscsi for * login/startup related tasks. */ - pwrb_handle = alloc_wrb_handle(phba, beiscsi_conn->beiscsi_conn_cid, 0); + pwrb_handle = alloc_wrb_handle(phba, (beiscsi_conn->beiscsi_conn_cid - + phba->fw_config.iscsi_cid_start), 0); pwrb = (struct iscsi_target_context_update_wrb *)pwrb_handle->pwrb; memset(pwrb, 0, sizeof(*pwrb)); AMAP_SET_BITS(struct amap_iscsi_target_context_update_wrb, @@ -3328,7 +3326,8 @@ static int beiscsi_alloc_pdu(struct iscsi_task *task, uint8_t opcode) io_task->bhs_pa.u.a64.address = paddr; io_task->libiscsi_itt = (itt_t)task->itt; io_task->pwrb_handle = alloc_wrb_handle(phba, - beiscsi_conn->beiscsi_conn_cid, + beiscsi_conn->beiscsi_conn_cid - + phba->fw_config.iscsi_cid_start, task->itt); io_task->conn = beiscsi_conn; @@ -3372,10 +3371,11 @@ static int beiscsi_alloc_pdu(struct iscsi_task *task, uint8_t opcode) (io_task->psgl_handle->sgl_index)); if (ring_mode) { phba->sgl_hndl_array[io_task->psgl_handle->sgl_index - - phba->fw_config.iscsi_cid_start] = + phba->fw_config.iscsi_icd_start] = io_task->psgl_handle; io_task->psgl_handle->task = task; - io_task->psgl_handle->cid = beiscsi_conn->beiscsi_conn_cid; + io_task->psgl_handle->cid = beiscsi_conn->beiscsi_conn_cid - + phba->fw_config.iscsi_cid_start; } else io_task->pwrb_handle->pio_handle = task; @@ -3384,7 +3384,9 @@ static int beiscsi_alloc_pdu(struct iscsi_task *task, uint8_t opcode) free_hndls: phwi_ctrlr = phba->phwi_ctrlr; - pwrb_context = &phwi_ctrlr->wrb_context[beiscsi_conn->beiscsi_conn_cid]; + pwrb_context = &phwi_ctrlr->wrb_context[ + beiscsi_conn->beiscsi_conn_cid - + phba->fw_config.iscsi_cid_start]; free_wrb_handle(phba, pwrb_context, io_task->pwrb_handle); io_task->pwrb_handle = NULL; pci_pool_free(beiscsi_sess->bhs_pool, io_task->cmd_bhs, @@ -3404,7 +3406,8 @@ static void beiscsi_cleanup_task(struct iscsi_task *task) struct hwi_controller *phwi_ctrlr; phwi_ctrlr = phba->phwi_ctrlr; - pwrb_context = &phwi_ctrlr->wrb_context[beiscsi_conn->beiscsi_conn_cid]; + pwrb_context = &phwi_ctrlr->wrb_context[beiscsi_conn->beiscsi_conn_cid + - phba->fw_config.iscsi_cid_start]; if (io_task->pwrb_handle) { free_wrb_handle(phba, pwrb_context, io_task->pwrb_handle); io_task->pwrb_handle = NULL; @@ -3561,7 +3564,8 @@ static int beiscsi_mtask(struct iscsi_task *task) session = conn->session; i = ((struct iscsi_tm *)task->hdr)->rtt; phwi_ctrlr = phba->phwi_ctrlr; - pwrb_context = &phwi_ctrlr->wrb_context[cid]; + pwrb_context = &phwi_ctrlr->wrb_context[cid - + phba->fw_config.iscsi_cid_start]; pwrb_handle = pwrb_context->pwrb_handle_basestd[be32_to_cpu(i) >> 16]; aborted_task = pwrb_handle->pio_handle; @@ -3754,6 +3758,14 @@ static int __devinit beiscsi_dev_probe(struct pci_dev *pcidev, spin_lock_init(&phba->io_sgl_lock); spin_lock_init(&phba->mgmt_sgl_lock); spin_lock_init(&phba->isr_lock); + ret = mgmt_get_fw_config(&phba->ctrl, phba); + if (ret != 0) { + shost_printk(KERN_ERR, phba->shost, + "Error getting fw config\n"); + goto free_port; + } + phba->shost->max_id = phba->fw_config.iscsi_cid_count; + phba->shost->can_queue = phba->params.ios_per_ctrl; beiscsi_get_params(phba); ret = beiscsi_init_port(phba); if (ret < 0) { @@ -3859,7 +3871,7 @@ struct iscsi_transport beiscsi_iscsi_transport = { ISCSI_USERNAME | ISCSI_PASSWORD | ISCSI_USERNAME_IN | ISCSI_PASSWORD_IN | ISCSI_FAST_ABORT | ISCSI_ABORT_TMO | - ISCSI_LU_RESET_TMO | ISCSI_TGT_RESET_TMO | + ISCSI_LU_RESET_TMO | ISCSI_PING_TMO | ISCSI_RECV_TMO | ISCSI_IFACE_NAME | ISCSI_INITIATOR_NAME, .host_param_mask = ISCSI_HOST_HWADDRESS | ISCSI_HOST_IPADDRESS | diff --git a/drivers/scsi/be2iscsi/be_main.h b/drivers/scsi/be2iscsi/be_main.h index 25e6b20..0e2eac6 100644 --- a/drivers/scsi/be2iscsi/be_main.h +++ b/drivers/scsi/be2iscsi/be_main.h @@ -46,23 +46,18 @@ #define OC_DEVICE_ID3 0x712 #define OC_DEVICE_ID4 0x222 -#define BE2_MAX_SESSIONS 64 +#define BE2_IO_DEPTH 1024 +#define BE2_MAX_SESSIONS 256 #define BE2_CMDS_PER_CXN 128 -#define BE2_LOGOUTS BE2_MAX_SESSIONS #define BE2_TMFS 16 #define BE2_NOPOUT_REQ 16 -#define BE2_ASYNCPDUS BE2_MAX_SESSIONS -#define BE2_MAX_ICDS 2048 #define BE2_SGE 32 #define BE2_DEFPDU_HDR_SZ 64 #define BE2_DEFPDU_DATA_SZ 8192 -#define BE2_IO_DEPTH \ - (BE2_MAX_ICDS / 2 - (BE2_LOGOUTS + BE2_TMFS + BE2_NOPOUT_REQ)) #define MAX_CPUS 31 #define BEISCSI_SGLIST_ELEMENTS BE2_SGE -#define BEISCSI_MAX_CMNDS 1024 /* Max IO's per Ctrlr sht->can_queue */ #define BEISCSI_CMD_PER_LUN 128 /* scsi_host->cmd_per_lun */ #define BEISCSI_MAX_SECTORS 2048 /* scsi_host->max_sectors */ @@ -802,7 +797,6 @@ struct hwi_controller { struct be_ring default_pdu_hdr; struct be_ring default_pdu_data; struct hwi_context_memory *phwi_ctxt; - unsigned short cq_errors[CXN_KILLED_CMND_DATA_NOT_ON_SAME_CONN]; }; enum hwh_type_enum { diff --git a/drivers/scsi/be2iscsi/be_mgmt.c b/drivers/scsi/be2iscsi/be_mgmt.c index 79c2bd5..df1b327 100644 --- a/drivers/scsi/be2iscsi/be_mgmt.c +++ b/drivers/scsi/be2iscsi/be_mgmt.c @@ -48,6 +48,14 @@ unsigned char mgmt_get_fw_config(struct be_ctrl_info *ctrl, pfw_cfg->ulp[0].sq_base; phba->fw_config.iscsi_cid_count = pfw_cfg->ulp[0].sq_count; + if (phba->fw_config.iscsi_cid_count > (BE2_MAX_SESSIONS / 2)) { + status = 1; + shost_printk(KERN_WARNING, phba->shost, + "FW reported MAX CXNS as %d \t" + "Max Supported = %d. Failing to load \n", + phba->fw_config.iscsi_cid_count, + BE2_MAX_SESSIONS); + } } else { shost_printk(KERN_WARNING, phba->shost, "Failed in mgmt_get_fw_config \n"); @@ -317,7 +325,8 @@ int mgmt_open_connection(struct beiscsi_hba *phba, struct tcp_connect_and_offload_out *ptcpcnct_out = embedded_payload(wrb); - ep = phba->ep_array[ptcpcnct_out->cid]; + ep = phba->ep_array[ptcpcnct_out->cid - + phba->fw_config.iscsi_cid_start]; beiscsi_ep = ep->dd_data; beiscsi_ep->fw_handle = ptcpcnct_out->connection_handle; beiscsi_ep->cid_vld = 1; -- cgit v0.10.2 From c24622886fb934313a2a43ea1f516cbf1ddd947b Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:05:34 +0530 Subject: [SCSI] be2iscsi: Move freeing of resources to stop_conn We need to hold on to ep resources untill invalidate and close connection are completed Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_iscsi.c b/drivers/scsi/be2iscsi/be_iscsi.c index 2b3c58b..f229184 100644 --- a/drivers/scsi/be2iscsi/be_iscsi.c +++ b/drivers/scsi/be2iscsi/be_iscsi.c @@ -460,14 +460,12 @@ static void beiscsi_put_cid(struct beiscsi_hba *phba, unsigned short cid) * beiscsi_free_ep - free endpoint * @ep: pointer to iscsi endpoint structure */ -static void beiscsi_free_ep(struct iscsi_endpoint *ep) +static void beiscsi_free_ep(struct beiscsi_endpoint *beiscsi_ep) { - struct beiscsi_endpoint *beiscsi_ep = ep->dd_data; struct beiscsi_hba *phba = beiscsi_ep->phba; beiscsi_put_cid(phba, beiscsi_ep->ep_cid); beiscsi_ep->phba = NULL; - iscsi_destroy_endpoint(ep); } /** @@ -498,7 +496,7 @@ beiscsi_ep_connect(struct Scsi_Host *shost, struct sockaddr *dst_addr, if (phba->state) { ret = -EBUSY; - SE_DEBUG(DBG_LVL_1, "The Adapet state is Not UP \n"); + SE_DEBUG(DBG_LVL_1, "The Adapter state is Not UP \n"); return ERR_PTR(ret); } @@ -510,9 +508,10 @@ beiscsi_ep_connect(struct Scsi_Host *shost, struct sockaddr *dst_addr, beiscsi_ep = ep->dd_data; beiscsi_ep->phba = phba; + beiscsi_ep->openiscsi_ep = ep; if (beiscsi_open_conn(ep, NULL, dst_addr, non_blocking)) { - SE_DEBUG(DBG_LVL_1, "Failed in allocate iscsi cid\n"); + SE_DEBUG(DBG_LVL_1, "Failed in beiscsi_open_conn \n"); ret = -ENOMEM; goto free_ep; } @@ -520,7 +519,7 @@ beiscsi_ep_connect(struct Scsi_Host *shost, struct sockaddr *dst_addr, return ep; free_ep: - beiscsi_free_ep(ep); + beiscsi_free_ep(beiscsi_ep); return ERR_PTR(ret); } @@ -547,15 +546,14 @@ int beiscsi_ep_poll(struct iscsi_endpoint *ep, int timeout_ms) * @ep: The iscsi endpoint * @flag: The type of connection closure */ -static int beiscsi_close_conn(struct iscsi_endpoint *ep, int flag) +static int beiscsi_close_conn(struct beiscsi_endpoint *beiscsi_ep, int flag) { int ret = 0; - struct beiscsi_endpoint *beiscsi_ep = ep->dd_data; struct beiscsi_hba *phba = beiscsi_ep->phba; if (MGMT_STATUS_SUCCESS != mgmt_upload_connection(phba, beiscsi_ep->ep_cid, - CONNECTION_UPLOAD_GRACEFUL)) { + flag)) { SE_DEBUG(DBG_LVL_8, "upload failed for cid 0x%x", beiscsi_ep->ep_cid); ret = -1; @@ -575,19 +573,15 @@ void beiscsi_ep_disconnect(struct iscsi_endpoint *ep) struct beiscsi_conn *beiscsi_conn; struct beiscsi_endpoint *beiscsi_ep; struct beiscsi_hba *phba; - int flag = 0; beiscsi_ep = ep->dd_data; phba = beiscsi_ep->phba; - SE_DEBUG(DBG_LVL_8, "In beiscsi_ep_disconnect\n"); if (beiscsi_ep->conn) { beiscsi_conn = beiscsi_ep->conn; iscsi_suspend_queue(beiscsi_conn->conn); - beiscsi_close_conn(ep, flag); } - beiscsi_free_ep(ep); } /** @@ -637,6 +631,9 @@ void beiscsi_conn_stop(struct iscsi_cls_conn *cls_conn, int flag) "mgmt_invalidate_connection Failed for cid=%d \n", beiscsi_ep->ep_cid); } + beiscsi_close_conn(beiscsi_ep, CONNECTION_UPLOAD_GRACEFUL); + beiscsi_free_ep(beiscsi_ep); + iscsi_destroy_endpoint(beiscsi_ep->openiscsi_ep); beiscsi_unbind_conn_to_cid(phba, beiscsi_ep->ep_cid); iscsi_conn_stop(cls_conn, flag); } diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 6c512b6..1390023 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -1434,6 +1434,8 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) unsigned int tot_nump = 0; struct beiscsi_conn *beiscsi_conn; struct sgl_handle *psgl_handle = NULL; + struct beiscsi_endpoint *beiscsi_ep; + struct iscsi_endpoint *ep; struct beiscsi_hba *phba; cq = pbe_eq->cq; @@ -1449,28 +1451,15 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) dw[offsetof(struct amap_sol_cqe_ring, icd_index) / 32] & SOL_ICD_INDEX_MASK) >> 6)]; - beiscsi_conn = phba->conn_table[psgl_handle->cid]; - if (!beiscsi_conn || !beiscsi_conn->ep) { - shost_printk(KERN_WARNING, phba->shost, - "Connection table empty for cid = %d\n", - psgl_handle->cid); - return 0; - } - + ep = phba->ep_array[psgl_handle->cid]; } else { - beiscsi_conn = phba->conn_table[(u32) ((sol-> - dw[offsetof(struct amap_sol_cqe, cid) / 32] & - SOL_CID_MASK) >> 6) - + ep = phba->ep_array[(u32) ((sol-> + dw[offsetof(struct amap_sol_cqe, cid) / 32] & + SOL_CID_MASK) >> 6) - phba->fw_config.iscsi_cid_start]; - if (!beiscsi_conn || !beiscsi_conn->ep) { - shost_printk(KERN_WARNING, phba->shost, - "Connection table empty for cid = %d\n", - (u32)(sol->dw[offsetof(struct amap_sol_cqe, - cid) / 32] & SOL_CID_MASK) >> 6); - return 0; - } } - + beiscsi_ep = ep->dd_data; + beiscsi_conn = beiscsi_ep->conn; if (num_processed >= 32) { hwi_ring_cq_db(phba, cq->id, num_processed, 0, 0); @@ -3044,7 +3033,7 @@ static int hba_setup_cid_tbls(struct beiscsi_hba *phba) { int i, new_cid; - phba->cid_array = kmalloc(sizeof(void *) * phba->params.cxns_per_ctrl, + phba->cid_array = kzalloc(sizeof(void *) * phba->params.cxns_per_ctrl, GFP_KERNEL); if (!phba->cid_array) { shost_printk(KERN_ERR, phba->shost, @@ -3052,7 +3041,7 @@ static int hba_setup_cid_tbls(struct beiscsi_hba *phba) "hba_setup_cid_tbls\n"); return -ENOMEM; } - phba->ep_array = kmalloc(sizeof(struct iscsi_endpoint *) * + phba->ep_array = kzalloc(sizeof(struct iscsi_endpoint *) * phba->params.cxns_per_ctrl * 2, GFP_KERNEL); if (!phba->ep_array) { shost_printk(KERN_ERR, phba->shost, diff --git a/drivers/scsi/be2iscsi/be_mgmt.h b/drivers/scsi/be2iscsi/be_mgmt.h index 24eaff9..6bc59e8 100644 --- a/drivers/scsi/be2iscsi/be_mgmt.h +++ b/drivers/scsi/be2iscsi/be_mgmt.h @@ -231,6 +231,7 @@ struct beiscsi_endpoint { struct beiscsi_hba *phba; struct beiscsi_sess *sess; struct beiscsi_conn *conn; + struct iscsi_endpoint *openiscsi_ep; unsigned short ip_type; char dst6_addr[ISCSI_ADDRESS_BUF_LEN]; unsigned long dst_addr; -- cgit v0.10.2 From d543148883f65c34e6cd54c5e9ed0592dfbb6acb Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:06:21 +0530 Subject: [SCSI] be2iscsi: Link Wrb with next Wrb This patch will link the current allocated wrb with the next wrb that will be allocated. This is a requirement from the chip. Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 1390023..ac76b14 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -629,29 +629,30 @@ free_io_sgl_handle(struct beiscsi_hba *phba, struct sgl_handle *psgl_handle) * alloc_wrb_handle - To allocate a wrb handle * @phba: The hba pointer * @cid: The cid to use for allocation - * @index: index allocation and wrb index * * This happens under session_lock until submission to chip */ -struct wrb_handle *alloc_wrb_handle(struct beiscsi_hba *phba, unsigned int cid, - int index) +struct wrb_handle *alloc_wrb_handle(struct beiscsi_hba *phba, unsigned int cid) { struct hwi_wrb_context *pwrb_context; struct hwi_controller *phwi_ctrlr; - struct wrb_handle *pwrb_handle; + struct wrb_handle *pwrb_handle, *pwrb_handle_tmp; phwi_ctrlr = phba->phwi_ctrlr; pwrb_context = &phwi_ctrlr->wrb_context[cid]; - if (pwrb_context->wrb_handles_available) { + if (pwrb_context->wrb_handles_available >= 2) { pwrb_handle = pwrb_context->pwrb_handle_base[ pwrb_context->alloc_index]; pwrb_context->wrb_handles_available--; - pwrb_handle->nxt_wrb_index = pwrb_handle->wrb_index; if (pwrb_context->alloc_index == (phba->params.wrbs_per_cxn - 1)) pwrb_context->alloc_index = 0; else pwrb_context->alloc_index++; + + pwrb_handle_tmp = pwrb_context->pwrb_handle_base[ + pwrb_context->alloc_index]; + pwrb_handle->nxt_wrb_index = pwrb_handle_tmp->wrb_index; } else pwrb_handle = NULL; return pwrb_handle; @@ -3206,7 +3207,7 @@ beiscsi_offload_connection(struct beiscsi_conn *beiscsi_conn, * login/startup related tasks. */ pwrb_handle = alloc_wrb_handle(phba, (beiscsi_conn->beiscsi_conn_cid - - phba->fw_config.iscsi_cid_start), 0); + phba->fw_config.iscsi_cid_start)); pwrb = (struct iscsi_target_context_update_wrb *)pwrb_handle->pwrb; memset(pwrb, 0, sizeof(*pwrb)); AMAP_SET_BITS(struct amap_iscsi_target_context_update_wrb, @@ -3316,8 +3317,8 @@ static int beiscsi_alloc_pdu(struct iscsi_task *task, uint8_t opcode) io_task->libiscsi_itt = (itt_t)task->itt; io_task->pwrb_handle = alloc_wrb_handle(phba, beiscsi_conn->beiscsi_conn_cid - - phba->fw_config.iscsi_cid_start, - task->itt); + phba->fw_config.iscsi_cid_start + ); io_task->conn = beiscsi_conn; task->hdr = (struct iscsi_hdr *)&io_task->cmd_bhs->iscsi_hdr; diff --git a/drivers/scsi/be2iscsi/be_main.h b/drivers/scsi/be2iscsi/be_main.h index 0e2eac6..0553f49 100644 --- a/drivers/scsi/be2iscsi/be_main.h +++ b/drivers/scsi/be2iscsi/be_main.h @@ -651,8 +651,7 @@ struct amap_iscsi_wrb { } __packed; -struct wrb_handle *alloc_wrb_handle(struct beiscsi_hba *phba, unsigned int cid, - int index); +struct wrb_handle *alloc_wrb_handle(struct beiscsi_hba *phba, unsigned int cid); void free_mgmt_sgl_handle(struct beiscsi_hba *phba, struct sgl_handle *psgl_handle); -- cgit v0.10.2 From 7bd6e25cdbee7d4f6bc4946dc914310220e637b8 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:07:02 +0530 Subject: [SCSI] be2iscsi: Added opcode for LOGOUT_RSP, TEXT_RESP, TMFUNC_RSP This patch adds opcodes in thecompletion path that were missed out earlier Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index ac76b14..4a855a9 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -559,6 +559,7 @@ beiscsi_process_async_pdu(struct beiscsi_conn *beiscsi_conn, SE_DEBUG(DBG_LVL_1, "In ISCSI_OP_REJECT\n"); break; case ISCSI_OP_LOGIN_RSP: + case ISCSI_OP_TEXT_RSP: task = conn->login_task; io_task = task->dd_data; login_hdr = (struct iscsi_hdr *)ppdu; @@ -810,6 +811,7 @@ be_complete_logout(struct beiscsi_conn *beiscsi_conn, struct iscsi_conn *conn = beiscsi_conn->conn; hdr = (struct iscsi_logout_rsp *)task->hdr; + hdr->opcode = ISCSI_OP_LOGOUT_RSP; hdr->t2wait = 5; hdr->t2retain = 0; hdr->flags = ((psol->dw[offsetof(struct amap_sol_cqe, i_flags) / 32] @@ -824,6 +826,9 @@ be_complete_logout(struct beiscsi_conn *beiscsi_conn, & SOL_EXP_CMD_SN_MASK) + ((psol->dw[offsetof(struct amap_sol_cqe, i_cmd_wnd) / 32] & SOL_CMD_WND_MASK) >> 24) - 1); + hdr->dlength[0] = 0; + hdr->dlength[1] = 0; + hdr->dlength[2] = 0; hdr->hlength = 0; hdr->itt = io_task->libiscsi_itt; __iscsi_complete_pdu(conn, (struct iscsi_hdr *)hdr, NULL, 0); @@ -838,6 +843,7 @@ be_complete_tmf(struct beiscsi_conn *beiscsi_conn, struct beiscsi_io_task *io_task = task->dd_data; hdr = (struct iscsi_tm_rsp *)task->hdr; + hdr->opcode = ISCSI_OP_SCSI_TMFUNC_RSP; hdr->flags = ((psol->dw[offsetof(struct amap_sol_cqe, i_flags) / 32] & SOL_FLAGS_MASK) >> 24) | 0x80; hdr->response = (psol->dw[offsetof(struct amap_sol_cqe, i_resp) / -- cgit v0.10.2 From 2807afb7411b97834fc7338f3735f3d152443551 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:07:49 +0530 Subject: [SCSI] be2iscsi:moved pci_set_drvdata to inside beiscsi_hba_alloc This patch moves pci_set_drvdata to inside beiscsi_hba_alloc Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 4a855a9..18471ed 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -112,6 +112,7 @@ static struct beiscsi_hba *beiscsi_hba_alloc(struct pci_dev *pcidev) memset(phba, 0, sizeof(*phba)); phba->shost = shost; phba->pcidev = pci_dev_get(pcidev); + pci_set_drvdata(pcidev, phba); if (iscsi_host_add(shost, &phba->pcidev->dev)) goto free_devices; @@ -3734,7 +3735,6 @@ static int __devinit beiscsi_dev_probe(struct pci_dev *pcidev, } SE_DEBUG(DBG_LVL_8, " phba = %p \n", phba); - pci_set_drvdata(pcidev, phba); if (enable_msix) num_cpus = find_num_cpus(); else -- cgit v0.10.2 From d7aea67b8a7665fe5e53cdf59ba76c9b8d67b751 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:08:39 +0530 Subject: [SCSI] be2iscsi: Use of opcode in beiscsi_alloc_pdu This patch enables use of opcode that is passed in Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 18471ed..0fae264 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -3339,7 +3339,7 @@ static int beiscsi_alloc_pdu(struct iscsi_task *task, uint8_t opcode) goto free_hndls; } else { io_task->scsi_cmnd = NULL; - if ((task->hdr->opcode & ISCSI_OPCODE_MASK) == ISCSI_OP_LOGIN) { + if ((opcode & ISCSI_OPCODE_MASK) == ISCSI_OP_LOGIN) { if (!beiscsi_conn->login_in_progress) { spin_lock(&phba->mgmt_sgl_lock); io_task->psgl_handle = (struct sgl_handle *) -- cgit v0.10.2 From 0ecb0b45f22df911c564070b64af21db36934f0f Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:09:19 +0530 Subject: [SCSI] be2iscsi: decide which requests need completion This patch decides whether ack based completion is required or not Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 0fae264..26d7016 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -3546,6 +3546,11 @@ static int beiscsi_mtask(struct iscsi_task *task) else AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, INI_RD_CMD); + if (task->hdr->ttt == ISCSI_RESERVED_TAG) + AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 0); + else + AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 1); + hwi_write_buffer(pwrb, task); break; case ISCSI_OP_TEXT: @@ -3554,6 +3559,7 @@ static int beiscsi_mtask(struct iscsi_task *task) else AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, INI_WR_CMD); + AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 0); AMAP_SET_BITS(struct amap_iscsi_wrb, dsp, pwrb, 1); hwi_write_buffer(pwrb, task); break; -- cgit v0.10.2 From 51a462500fbed4a1e8110dc60a421a3f12b9580b Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:10:01 +0530 Subject: [SCSI] be2iscsi: No requirement for endianess change for data_count This patch removes the endianess change that was wrongly added for data_count Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 26d7016..6170548 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -3607,7 +3607,7 @@ static int beiscsi_mtask(struct iscsi_task *task) } AMAP_SET_BITS(struct amap_iscsi_wrb, r2t_exp_dtl, pwrb, - be32_to_cpu(task->data_count)); + task->data_count); AMAP_SET_BITS(struct amap_iscsi_wrb, ptr2nextwrb, pwrb, io_task->pwrb_handle->nxt_wrb_index); be_dws_le_to_cpu(pwrb, sizeof(struct iscsi_wrb)); -- cgit v0.10.2 From 756d29c8c7ed8887ed7d752371ce2f6d12399267 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:10:46 +0530 Subject: [SCSI] be2iscsi: Enable async mode for mcc rings This patches enables async mode for mcc rings so that multiple requests can be queued. Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be.h b/drivers/scsi/be2iscsi/be.h index a93a504..3861cf4 100644 --- a/drivers/scsi/be2iscsi/be.h +++ b/drivers/scsi/be2iscsi/be.h @@ -24,6 +24,7 @@ #define FW_VER_LEN 32 #define MCC_Q_LEN 128 #define MCC_CQ_LEN 256 +#define MAX_MCC_CMD 16 struct be_dma_mem { void *va; @@ -57,6 +58,11 @@ static inline void *queue_head_node(struct be_queue_info *q) return q->dma_mem.va + q->head * q->entry_size; } +static inline void *queue_get_wrb(struct be_queue_info *q, unsigned int wrb_num) +{ + return q->dma_mem.va + wrb_num * q->entry_size; +} + static inline void *queue_tail_node(struct be_queue_info *q) { return q->dma_mem.va + q->tail * q->entry_size; @@ -104,15 +110,19 @@ struct be_ctrl_info { spinlock_t mcc_lock; /* For serializing mcc cmds to BE card */ spinlock_t mcc_cq_lock; - /* MCC Async callback */ - void (*async_cb) (void *adapter, bool link_up); - void *adapter_ctxt; + wait_queue_head_t mcc_wait[MAX_MCC_CMD + 1]; + unsigned int mcc_tag[MAX_MCC_CMD]; + unsigned int mcc_numtag[MAX_MCC_CMD + 1]; + unsigned short mcc_alloc_index; + unsigned short mcc_free_index; + unsigned int mcc_tag_available; }; #include "be_cmds.h" #define PAGE_SHIFT_4K 12 #define PAGE_SIZE_4K (1 << PAGE_SHIFT_4K) +#define mcc_timeout 120000 /* 5s timeout */ /* Returns number of pages spanned by the data starting at the given addr */ #define PAGES_4K_SPANNED(_address, size) \ diff --git a/drivers/scsi/be2iscsi/be_cmds.c b/drivers/scsi/be2iscsi/be_cmds.c index f008708..d4a0d1d 100644 --- a/drivers/scsi/be2iscsi/be_cmds.c +++ b/drivers/scsi/be2iscsi/be_cmds.c @@ -19,7 +19,7 @@ #include "be_mgmt.h" #include "be_main.h" -static void be_mcc_notify(struct beiscsi_hba *phba) +void be_mcc_notify(struct beiscsi_hba *phba) { struct be_queue_info *mccq = &phba->ctrl.mcc_obj.q; u32 val = 0; @@ -29,6 +29,52 @@ static void be_mcc_notify(struct beiscsi_hba *phba) iowrite32(val, phba->db_va + DB_MCCQ_OFFSET); } +unsigned int alloc_mcc_tag(struct beiscsi_hba *phba) +{ + unsigned int tag = 0; + unsigned int num = 0; + +mcc_tag_rdy: + if (phba->ctrl.mcc_tag_available) { + tag = phba->ctrl.mcc_tag[phba->ctrl.mcc_alloc_index]; + phba->ctrl.mcc_tag[phba->ctrl.mcc_alloc_index] = 0; + phba->ctrl.mcc_numtag[tag] = 0; + } else { + udelay(100); + num++; + if (num < mcc_timeout) + goto mcc_tag_rdy; + } + if (tag) { + phba->ctrl.mcc_tag_available--; + if (phba->ctrl.mcc_alloc_index == (MAX_MCC_CMD - 1)) + phba->ctrl.mcc_alloc_index = 0; + else + phba->ctrl.mcc_alloc_index++; + } + return tag; +} + +void free_mcc_tag(struct be_ctrl_info *ctrl, unsigned int tag) +{ + spin_lock(&ctrl->mbox_lock); + tag = tag & 0x000000FF; + ctrl->mcc_tag[ctrl->mcc_free_index] = tag; + if (ctrl->mcc_free_index == (MAX_MCC_CMD - 1)) + ctrl->mcc_free_index = 0; + else + ctrl->mcc_free_index++; + ctrl->mcc_tag_available++; + spin_unlock(&ctrl->mbox_lock); +} + +bool is_link_state_evt(u32 trailer) +{ + return (((trailer >> ASYNC_TRAILER_EVENT_CODE_SHIFT) & + ASYNC_TRAILER_EVENT_CODE_MASK) == + ASYNC_EVENT_CODE_LINK_STATE); +} + static inline bool be_mcc_compl_is_new(struct be_mcc_compl *compl) { if (compl->flags != 0) { @@ -64,12 +110,30 @@ static int be_mcc_compl_process(struct be_ctrl_info *ctrl, return 0; } - -static inline bool is_link_state_evt(u32 trailer) +int be_mcc_compl_process_isr(struct be_ctrl_info *ctrl, + struct be_mcc_compl *compl) { - return (((trailer >> ASYNC_TRAILER_EVENT_CODE_SHIFT) & - ASYNC_TRAILER_EVENT_CODE_MASK) == - ASYNC_EVENT_CODE_LINK_STATE); + u16 compl_status, extd_status; + unsigned short tag; + + be_dws_le_to_cpu(compl, 4); + + compl_status = (compl->status >> CQE_STATUS_COMPL_SHIFT) & + CQE_STATUS_COMPL_MASK; + /* The ctrl.mcc_numtag[tag] is filled with + * [31] = valid, [30:24] = Rsvd, [23:16] = wrb, [15:8] = extd_status, + * [7:0] = compl_status + */ + tag = (compl->tag0 & 0x000000FF); + extd_status = (compl->status >> CQE_STATUS_EXTD_SHIFT) & + CQE_STATUS_EXTD_MASK; + + ctrl->mcc_numtag[tag] = 0x80000000; + ctrl->mcc_numtag[tag] |= (compl->tag0 & 0x00FF0000); + ctrl->mcc_numtag[tag] |= (extd_status & 0x000000FF) << 8; + ctrl->mcc_numtag[tag] |= (compl_status & 0x000000FF); + wake_up_interruptible(&ctrl->mcc_wait[tag]); + return 0; } static struct be_mcc_compl *be_mcc_compl_get(struct beiscsi_hba *phba) @@ -89,7 +153,7 @@ static void be2iscsi_fail_session(struct iscsi_cls_session *cls_session) iscsi_session_failure(cls_session->dd_data, ISCSI_ERR_CONN_FAILED); } -static void beiscsi_async_link_state_process(struct beiscsi_hba *phba, +void beiscsi_async_link_state_process(struct beiscsi_hba *phba, struct be_async_event_link_state *evt) { switch (evt->port_link_status) { @@ -162,7 +226,6 @@ int beiscsi_process_mcc(struct beiscsi_hba *phba) /* Wait till no more pending mcc requests are present */ static int be_mcc_wait_compl(struct beiscsi_hba *phba) { -#define mcc_timeout 120000 /* 5s timeout */ int i, status; for (i = 0; i < mcc_timeout; i++) { status = beiscsi_process_mcc(phba); @@ -372,9 +435,10 @@ struct be_mcc_wrb *wrb_from_mccq(struct beiscsi_hba *phba) BUG_ON(atomic_read(&mccq->used) >= mccq->len); wrb = queue_head_node(mccq); + memset(wrb, 0, sizeof(*wrb)); + wrb->tag0 = (mccq->head & 0x000000FF) << 16; queue_head_inc(mccq); atomic_inc(&mccq->used); - memset(wrb, 0, sizeof(*wrb)); return wrb; } diff --git a/drivers/scsi/be2iscsi/be_cmds.h b/drivers/scsi/be2iscsi/be_cmds.h index 5de8acb..69dddfa 100644 --- a/drivers/scsi/be2iscsi/be_cmds.h +++ b/drivers/scsi/be2iscsi/be_cmds.h @@ -425,14 +425,20 @@ int beiscsi_cmd_mccq_create(struct beiscsi_hba *phba, int be_poll_mcc(struct be_ctrl_info *ctrl); unsigned char mgmt_check_supported_fw(struct be_ctrl_info *ctrl, struct beiscsi_hba *phba); -int be_cmd_get_mac_addr(struct beiscsi_hba *phba, u8 *mac_addr); - +unsigned int be_cmd_get_mac_addr(struct beiscsi_hba *phba); +void free_mcc_tag(struct be_ctrl_info *ctrl, unsigned int tag); /*ISCSI Functuions */ int be_cmd_fw_initialize(struct be_ctrl_info *ctrl); struct be_mcc_wrb *wrb_from_mbox(struct be_dma_mem *mbox_mem); struct be_mcc_wrb *wrb_from_mccq(struct beiscsi_hba *phba); int be_mcc_notify_wait(struct beiscsi_hba *phba); +void be_mcc_notify(struct beiscsi_hba *phba); +unsigned int alloc_mcc_tag(struct beiscsi_hba *phba); +void beiscsi_async_link_state_process(struct beiscsi_hba *phba, + struct be_async_event_link_state *evt); +int be_mcc_compl_process_isr(struct be_ctrl_info *ctrl, + struct be_mcc_compl *compl); int be_mbox_notify(struct be_ctrl_info *ctrl); @@ -448,6 +454,8 @@ int be_cmd_iscsi_post_sgl_pages(struct be_ctrl_info *ctrl, int be_cmd_wrbq_create(struct be_ctrl_info *ctrl, struct be_dma_mem *q_mem, struct be_queue_info *wrbq); +bool is_link_state_evt(u32 trailer); + struct be_default_pdu_context { u32 dw[4]; } __packed; diff --git a/drivers/scsi/be2iscsi/be_iscsi.c b/drivers/scsi/be2iscsi/be_iscsi.c index f229184..95694d3 100644 --- a/drivers/scsi/be2iscsi/be_iscsi.c +++ b/drivers/scsi/be2iscsi/be_iscsi.c @@ -101,6 +101,7 @@ void beiscsi_session_destroy(struct iscsi_cls_session *cls_session) struct iscsi_session *sess = cls_session->dd_data; struct beiscsi_session *beiscsi_sess = sess->dd_data; + SE_DEBUG(DBG_LVL_8, "In beiscsi_session_destroy\n"); pci_pool_destroy(beiscsi_sess->bhs_pool); iscsi_session_teardown(cls_session); } @@ -224,6 +225,7 @@ int beiscsi_conn_get_param(struct iscsi_cls_conn *cls_conn, struct beiscsi_conn *beiscsi_conn = conn->dd_data; int len = 0; + SE_DEBUG(DBG_LVL_8, "In beiscsi_conn_get_param, param= %d\n", param); beiscsi_ep = beiscsi_conn->ep; if (!beiscsi_ep) { SE_DEBUG(DBG_LVL_1, @@ -254,6 +256,7 @@ int beiscsi_set_param(struct iscsi_cls_conn *cls_conn, struct iscsi_session *session = conn->session; int ret; + SE_DEBUG(DBG_LVL_8, "In beiscsi_conn_set_param, param= %d\n", param); ret = iscsi_set_param(cls_conn, param, buf, buflen); if (ret) return ret; @@ -293,12 +296,41 @@ int beiscsi_get_host_param(struct Scsi_Host *shost, enum iscsi_host_param param, char *buf) { struct beiscsi_hba *phba = (struct beiscsi_hba *)iscsi_host_priv(shost); + struct be_cmd_resp_get_mac_addr *resp; + struct be_mcc_wrb *wrb; + unsigned int tag, wrb_num; int len = 0; + unsigned short status, extd_status; + struct be_queue_info *mccq = &phba->ctrl.mcc_obj.q; + SE_DEBUG(DBG_LVL_8, "In beiscsi_get_host_param, param= %d\n", param); switch (param) { case ISCSI_HOST_PARAM_HWADDRESS: - be_cmd_get_mac_addr(phba, phba->mac_address); - len = sysfs_format_mac(buf, phba->mac_address, ETH_ALEN); + tag = be_cmd_get_mac_addr(phba); + if (!tag) { + SE_DEBUG(DBG_LVL_1, "be_cmd_get_mac_addr Failed \n"); + return -1; + } else + wait_event_interruptible(phba->ctrl.mcc_wait[tag], + phba->ctrl.mcc_numtag[tag]); + + wrb_num = (phba->ctrl.mcc_numtag[tag] & 0x00FF0000) >> 16; + extd_status = (phba->ctrl.mcc_numtag[tag] & 0x0000FF00) >> 8; + status = phba->ctrl.mcc_numtag[tag] & 0x000000FF; + if (status || extd_status) { + SE_DEBUG(DBG_LVL_1, "be_cmd_get_mac_addr Failed" + " status = %d extd_status = %d \n", + status, extd_status); + free_mcc_tag(&phba->ctrl, tag); + return -1; + } else { + wrb = queue_get_wrb(mccq, wrb_num); + free_mcc_tag(&phba->ctrl, tag); + resp = embedded_payload(wrb); + memcpy(phba->mac_address, resp->mac_address, ETH_ALEN); + len = sysfs_format_mac(buf, phba->mac_address, + ETH_ALEN); + } break; default: return iscsi_host_get_param(shost, param, buf); @@ -378,6 +410,7 @@ int beiscsi_conn_start(struct iscsi_cls_conn *cls_conn) struct beiscsi_endpoint *beiscsi_ep; struct beiscsi_offload_params params; + SE_DEBUG(DBG_LVL_8, "In beiscsi_conn_start\n"); memset(¶ms, 0, sizeof(struct beiscsi_offload_params)); beiscsi_ep = beiscsi_conn->ep; if (!beiscsi_ep) @@ -422,8 +455,14 @@ static int beiscsi_open_conn(struct iscsi_endpoint *ep, { struct beiscsi_endpoint *beiscsi_ep = ep->dd_data; struct beiscsi_hba *phba = beiscsi_ep->phba; + struct be_queue_info *mccq = &phba->ctrl.mcc_obj.q; + struct be_mcc_wrb *wrb; + struct tcp_connect_and_offload_out *ptcpcnct_out; + unsigned short status, extd_status; + unsigned int tag, wrb_num; int ret = -1; + SE_DEBUG(DBG_LVL_8, "In beiscsi_open_conn\n"); beiscsi_ep->ep_cid = beiscsi_get_cid(phba); if (beiscsi_ep->ep_cid == 0xFFFF) { SE_DEBUG(DBG_LVL_1, "No free cid available\n"); @@ -440,7 +479,35 @@ static int beiscsi_open_conn(struct iscsi_endpoint *ep, } beiscsi_ep->cid_vld = 0; - return mgmt_open_connection(phba, dst_addr, beiscsi_ep); + tag = mgmt_open_connection(phba, dst_addr, beiscsi_ep); + if (!tag) { + SE_DEBUG(DBG_LVL_1, + "mgmt_invalidate_connection Failed for cid=%d \n", + beiscsi_ep->ep_cid); + } else { + wait_event_interruptible(phba->ctrl.mcc_wait[tag], + phba->ctrl.mcc_numtag[tag]); + } + wrb_num = (phba->ctrl.mcc_numtag[tag] & 0x00FF0000) >> 16; + extd_status = (phba->ctrl.mcc_numtag[tag] & 0x0000FF00) >> 8; + status = phba->ctrl.mcc_numtag[tag] & 0x000000FF; + if (status || extd_status) { + SE_DEBUG(DBG_LVL_1, "mgmt_open_connection Failed" + " status = %d extd_status = %d \n", + status, extd_status); + free_mcc_tag(&phba->ctrl, tag); + return -1; + } else { + wrb = queue_get_wrb(mccq, wrb_num); + free_mcc_tag(&phba->ctrl, tag); + + ptcpcnct_out = embedded_payload(wrb); + beiscsi_ep = ep->dd_data; + beiscsi_ep->fw_handle = ptcpcnct_out->connection_handle; + beiscsi_ep->cid_vld = 1; + SE_DEBUG(DBG_LVL_8, "mgmt_open_connection Success\n"); + } + return 0; } /** @@ -509,7 +576,6 @@ beiscsi_ep_connect(struct Scsi_Host *shost, struct sockaddr *dst_addr, beiscsi_ep = ep->dd_data; beiscsi_ep->phba = phba; beiscsi_ep->openiscsi_ep = ep; - if (beiscsi_open_conn(ep, NULL, dst_addr, non_blocking)) { SE_DEBUG(DBG_LVL_1, "Failed in beiscsi_open_conn \n"); ret = -ENOMEM; @@ -549,16 +615,19 @@ int beiscsi_ep_poll(struct iscsi_endpoint *ep, int timeout_ms) static int beiscsi_close_conn(struct beiscsi_endpoint *beiscsi_ep, int flag) { int ret = 0; + unsigned int tag; struct beiscsi_hba *phba = beiscsi_ep->phba; - if (MGMT_STATUS_SUCCESS != - mgmt_upload_connection(phba, beiscsi_ep->ep_cid, - flag)) { + tag = mgmt_upload_connection(phba, beiscsi_ep->ep_cid, flag); + if (!tag) { SE_DEBUG(DBG_LVL_8, "upload failed for cid 0x%x", beiscsi_ep->ep_cid); ret = -1; + } else { + wait_event_interruptible(phba->ctrl.mcc_wait[tag], + phba->ctrl.mcc_numtag[tag]); + free_mcc_tag(&phba->ctrl, tag); } - return ret; } @@ -576,6 +645,8 @@ void beiscsi_ep_disconnect(struct iscsi_endpoint *ep) beiscsi_ep = ep->dd_data; phba = beiscsi_ep->phba; + SE_DEBUG(DBG_LVL_8, "In beiscsi_ep_disconnect for ep_cid = %d\n", + beiscsi_ep->ep_cid); if (beiscsi_ep->conn) { beiscsi_conn = beiscsi_ep->conn; @@ -614,22 +685,27 @@ void beiscsi_conn_stop(struct iscsi_cls_conn *cls_conn, int flag) struct iscsi_session *session = conn->session; struct Scsi_Host *shost = iscsi_session_to_shost(session->cls_session); struct beiscsi_hba *phba = iscsi_host_priv(shost); - unsigned int status; + unsigned int tag; unsigned short savecfg_flag = CMD_ISCSI_SESSION_SAVE_CFG_ON_FLASH; - SE_DEBUG(DBG_LVL_8, "In beiscsi_conn_stop\n"); beiscsi_ep = beiscsi_conn->ep; if (!beiscsi_ep) { SE_DEBUG(DBG_LVL_8, "In beiscsi_conn_stop , no beiscsi_ep\n"); return; } - status = mgmt_invalidate_connection(phba, beiscsi_ep, + SE_DEBUG(DBG_LVL_8, "In beiscsi_conn_stop ep_cid = %d\n", + beiscsi_ep->ep_cid); + tag = mgmt_invalidate_connection(phba, beiscsi_ep, beiscsi_ep->ep_cid, 1, savecfg_flag); - if (status != MGMT_STATUS_SUCCESS) { + if (!tag) { SE_DEBUG(DBG_LVL_1, "mgmt_invalidate_connection Failed for cid=%d \n", beiscsi_ep->ep_cid); + } else { + wait_event_interruptible(phba->ctrl.mcc_wait[tag], + phba->ctrl.mcc_numtag[tag]); + free_mcc_tag(&phba->ctrl, tag); } beiscsi_close_conn(beiscsi_ep, CONNECTION_UPLOAD_GRACEFUL); beiscsi_free_ep(beiscsi_ep); diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 6170548..a6a2c64 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -442,7 +442,7 @@ static irqreturn_t be_isr(int irq, void *dev_id) if (phba->todo_mcc_cq) queue_work(phba->wq, &phba->work_cqs); - if ((num_mcceq_processed) && (!num_ioeq_processed)) + if ((num_mcceq_processed) && (!num_ioeq_processed)) hwi_ring_eq_db(phba, eq->id, 0, (num_ioeq_processed + num_mcceq_processed) , 1, 1); @@ -651,7 +651,6 @@ struct wrb_handle *alloc_wrb_handle(struct beiscsi_hba *phba, unsigned int cid) pwrb_context->alloc_index = 0; else pwrb_context->alloc_index++; - pwrb_handle_tmp = pwrb_context->pwrb_handle_base[ pwrb_context->alloc_index]; pwrb_handle->nxt_wrb_index = pwrb_handle_tmp->wrb_index; @@ -791,6 +790,7 @@ be_complete_io(struct beiscsi_conn *beiscsi_conn, memcpy(task->sc->sense_buffer, sense, min_t(u16, sense_len, SCSI_SENSE_BUFFERSIZE)); } + if (io_task->cmd_bhs->iscsi_hdr.flags & ISCSI_FLAG_CMD_READ) { if (psol->dw[offsetof(struct amap_sol_cqe, i_res_cnt) / 32] & SOL_RES_CNT_MASK) @@ -1432,6 +1432,48 @@ static void hwi_process_default_pdu_ring(struct beiscsi_conn *beiscsi_conn, hwi_post_async_buffers(phba, pasync_handle->is_header); } +static void beiscsi_process_mcc_isr(struct beiscsi_hba *phba) +{ + struct be_queue_info *mcc_cq; + struct be_mcc_compl *mcc_compl; + unsigned int num_processed = 0; + + mcc_cq = &phba->ctrl.mcc_obj.cq; + mcc_compl = queue_tail_node(mcc_cq); + mcc_compl->flags = le32_to_cpu(mcc_compl->flags); + while (mcc_compl->flags & CQE_FLAGS_VALID_MASK) { + + if (num_processed >= 32) { + hwi_ring_cq_db(phba, mcc_cq->id, + num_processed, 0, 0); + num_processed = 0; + } + if (mcc_compl->flags & CQE_FLAGS_ASYNC_MASK) { + /* Interpret flags as an async trailer */ + if (is_link_state_evt(mcc_compl->flags)) + /* Interpret compl as a async link evt */ + beiscsi_async_link_state_process(phba, + (struct be_async_event_link_state *) mcc_compl); + else + SE_DEBUG(DBG_LVL_1, + " Unsupported Async Event, flags" + " = 0x%08x \n", mcc_compl->flags); + } else if (mcc_compl->flags & CQE_FLAGS_COMPLETED_MASK) { + be_mcc_compl_process_isr(&phba->ctrl, mcc_compl); + atomic_dec(&phba->ctrl.mcc_obj.q.used); + } + + mcc_compl->flags = 0; + queue_tail_inc(mcc_cq); + mcc_compl = queue_tail_node(mcc_cq); + mcc_compl->flags = le32_to_cpu(mcc_compl->flags); + num_processed++; + } + + if (num_processed > 0) + hwi_ring_cq_db(phba, mcc_cq->id, num_processed, 1, 0); + +} static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) { @@ -1468,6 +1510,7 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) } beiscsi_ep = ep->dd_data; beiscsi_conn = beiscsi_ep->conn; + if (num_processed >= 32) { hwi_ring_cq_db(phba, cq->id, num_processed, 0, 0); @@ -1603,7 +1646,7 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) return tot_nump; } -static void beiscsi_process_all_cqs(struct work_struct *work) +void beiscsi_process_all_cqs(struct work_struct *work) { unsigned long flags; struct hwi_controller *phwi_ctrlr; @@ -1623,6 +1666,7 @@ static void beiscsi_process_all_cqs(struct work_struct *work) spin_lock_irqsave(&phba->isr_lock, flags); phba->todo_mcc_cq = 0; spin_unlock_irqrestore(&phba->isr_lock, flags); + beiscsi_process_mcc_isr(phba); } if (phba->todo_cq) { @@ -3160,6 +3204,7 @@ static void hwi_purge_eq(struct beiscsi_hba *phba) struct be_queue_info *eq; struct be_eq_entry *eqe = NULL; int i, eq_msix; + unsigned int num_processed; phwi_ctrlr = phba->phwi_ctrlr; phwi_context = phwi_ctrlr->phwi_ctxt; @@ -3171,13 +3216,17 @@ static void hwi_purge_eq(struct beiscsi_hba *phba) for (i = 0; i < (phba->num_cpus + eq_msix); i++) { eq = &phwi_context->be_eq[i].q; eqe = queue_tail_node(eq); - + num_processed = 0; while (eqe->dw[offsetof(struct amap_eq_entry, valid) / 32] & EQE_VALID_MASK) { AMAP_SET_BITS(struct amap_eq_entry, valid, eqe, 0); queue_tail_inc(eq); eqe = queue_tail_node(eq); + num_processed++; } + + if (num_processed) + hwi_ring_eq_db(phba, eq->id, 1, num_processed, 1, 1); } } @@ -3189,8 +3238,9 @@ static void beiscsi_clean_port(struct beiscsi_hba *phba) if (mgmt_status) shost_printk(KERN_WARNING, phba->shost, "mgmt_epfw_cleanup FAILED \n"); - hwi_cleanup(phba); + hwi_purge_eq(phba); + hwi_cleanup(phba); if (ring_mode) kfree(phba->sgl_hndl_array); kfree(phba->io_sgl_hndl_base); @@ -3519,6 +3569,7 @@ static int beiscsi_mtask(struct iscsi_task *task) unsigned int doorbell = 0; unsigned int i, cid; struct iscsi_task *aborted_task; + unsigned int tag; cid = beiscsi_conn->beiscsi_conn_cid; pwrb = io_task->pwrb_handle->pwrb; @@ -3550,7 +3601,6 @@ static int beiscsi_mtask(struct iscsi_task *task) AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 0); else AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 1); - hwi_write_buffer(pwrb, task); break; case ISCSI_OP_TEXT: @@ -3579,9 +3629,18 @@ static int beiscsi_mtask(struct iscsi_task *task) if (!aborted_io_task->scsi_cmnd) return 0; - mgmt_invalidate_icds(phba, + tag = mgmt_invalidate_icds(phba, aborted_io_task->psgl_handle->sgl_index, cid); + if (!tag) { + shost_printk(KERN_WARNING, phba->shost, + "mgmt_invalidate_icds could not be" + " submitted\n"); + } else { + wait_event_interruptible(phba->ctrl.mcc_wait[tag], + phba->ctrl.mcc_numtag[tag]); + free_mcc_tag(&phba->ctrl, tag); + } if (ring_mode) io_task->psgl_handle->type = INI_TMF_CMD; else @@ -3656,7 +3715,6 @@ static int beiscsi_task_xmit(struct iscsi_task *task) return beiscsi_iotask(task, sg, num_sg, xferlen, writedir); } - static void beiscsi_remove(struct pci_dev *pcidev) { struct beiscsi_hba *phba = NULL; @@ -3776,6 +3834,15 @@ static int __devinit beiscsi_dev_probe(struct pci_dev *pcidev, goto free_port; } + for (i = 0; i < MAX_MCC_CMD ; i++) { + init_waitqueue_head(&phba->ctrl.mcc_wait[i + 1]); + phba->ctrl.mcc_tag[i] = i + 1; + phba->ctrl.mcc_numtag[i + 1] = 0; + phba->ctrl.mcc_tag_available++; + } + + phba->ctrl.mcc_alloc_index = phba->ctrl.mcc_free_index = 0; + snprintf(phba->wq_name, sizeof(phba->wq_name), "beiscsi_q_irq%u", phba->shost->host_no); phba->wq = create_workqueue(phba->wq_name); diff --git a/drivers/scsi/be2iscsi/be_main.h b/drivers/scsi/be2iscsi/be_main.h index 0553f49..4cde8f6 100644 --- a/drivers/scsi/be2iscsi/be_main.h +++ b/drivers/scsi/be2iscsi/be_main.h @@ -655,6 +655,8 @@ struct wrb_handle *alloc_wrb_handle(struct beiscsi_hba *phba, unsigned int cid); void free_mgmt_sgl_handle(struct beiscsi_hba *phba, struct sgl_handle *psgl_handle); +void beiscsi_process_all_cqs(struct work_struct *work); + struct pdu_nop_out { u32 dw[12]; }; diff --git a/drivers/scsi/be2iscsi/be_mgmt.c b/drivers/scsi/be2iscsi/be_mgmt.c index df1b327..18f8041 100644 --- a/drivers/scsi/be2iscsi/be_mgmt.c +++ b/drivers/scsi/be2iscsi/be_mgmt.c @@ -148,10 +148,17 @@ unsigned char mgmt_invalidate_icds(struct beiscsi_hba *phba, { struct be_dma_mem nonemb_cmd; struct be_ctrl_info *ctrl = &phba->ctrl; - struct be_mcc_wrb *wrb = wrb_from_mccq(phba); - struct be_sge *sge = nonembedded_sgl(wrb); + struct be_mcc_wrb *wrb; + struct be_sge *sge; struct invalidate_commands_params_in *req; - int status = 0; + unsigned int tag = 0; + + spin_lock(&ctrl->mbox_lock); + tag = alloc_mcc_tag(phba); + if (!tag) { + spin_unlock(&ctrl->mbox_lock); + return tag; + } nonemb_cmd.va = pci_alloc_consistent(ctrl->pdev, sizeof(struct invalidate_commands_params_in), @@ -164,8 +171,9 @@ unsigned char mgmt_invalidate_icds(struct beiscsi_hba *phba, } nonemb_cmd.size = sizeof(struct invalidate_commands_params_in); req = nonemb_cmd.va; - spin_lock(&ctrl->mbox_lock); - memset(wrb, 0, sizeof(*wrb)); + wrb = wrb_from_mccq(phba); + sge = nonembedded_sgl(wrb); + wrb->tag0 |= tag; be_wrb_hdr_prepare(wrb, sizeof(*req), false, 1); be_cmd_hdr_prepare(&req->hdr, CMD_SUBSYSTEM_ISCSI, @@ -180,14 +188,12 @@ unsigned char mgmt_invalidate_icds(struct beiscsi_hba *phba, sge->pa_lo = cpu_to_le32(nonemb_cmd.dma & 0xFFFFFFFF); sge->len = cpu_to_le32(nonemb_cmd.size); - status = be_mcc_notify_wait(phba); - if (status) - SE_DEBUG(DBG_LVL_1, "ICDS Invalidation Failed\n"); + be_mcc_notify(phba); spin_unlock(&ctrl->mbox_lock); if (nonemb_cmd.va) pci_free_consistent(ctrl->pdev, nonemb_cmd.size, nonemb_cmd.va, nonemb_cmd.dma); - return status; + return tag; } unsigned char mgmt_invalidate_connection(struct beiscsi_hba *phba, @@ -197,13 +203,19 @@ unsigned char mgmt_invalidate_connection(struct beiscsi_hba *phba, unsigned short savecfg_flag) { struct be_ctrl_info *ctrl = &phba->ctrl; - struct be_mcc_wrb *wrb = wrb_from_mccq(phba); - struct iscsi_invalidate_connection_params_in *req = - embedded_payload(wrb); - int status = 0; + struct be_mcc_wrb *wrb; + struct iscsi_invalidate_connection_params_in *req; + unsigned int tag = 0; spin_lock(&ctrl->mbox_lock); - memset(wrb, 0, sizeof(*wrb)); + tag = alloc_mcc_tag(phba); + if (!tag) { + spin_unlock(&ctrl->mbox_lock); + return tag; + } + wrb = wrb_from_mccq(phba); + wrb->tag0 |= tag; + req = embedded_payload(wrb); be_wrb_hdr_prepare(wrb, sizeof(*req), true, 0); be_cmd_hdr_prepare(&req->hdr, CMD_SUBSYSTEM_ISCSI_INI, @@ -216,35 +228,37 @@ unsigned char mgmt_invalidate_connection(struct beiscsi_hba *phba, else req->cleanup_type = CMD_ISCSI_CONNECTION_INVALIDATE; req->save_cfg = savecfg_flag; - status = be_mcc_notify_wait(phba); - if (status) - SE_DEBUG(DBG_LVL_1, "Invalidation Failed\n"); - + be_mcc_notify(phba); spin_unlock(&ctrl->mbox_lock); - return status; + return tag; } unsigned char mgmt_upload_connection(struct beiscsi_hba *phba, unsigned short cid, unsigned int upload_flag) { struct be_ctrl_info *ctrl = &phba->ctrl; - struct be_mcc_wrb *wrb = wrb_from_mccq(phba); - struct tcp_upload_params_in *req = embedded_payload(wrb); - int status = 0; + struct be_mcc_wrb *wrb; + struct tcp_upload_params_in *req; + unsigned int tag = 0; spin_lock(&ctrl->mbox_lock); - memset(wrb, 0, sizeof(*wrb)); + tag = alloc_mcc_tag(phba); + if (!tag) { + spin_unlock(&ctrl->mbox_lock); + return tag; + } + wrb = wrb_from_mccq(phba); + req = embedded_payload(wrb); + wrb->tag0 |= tag; be_wrb_hdr_prepare(wrb, sizeof(*req), true, 0); be_cmd_hdr_prepare(&req->hdr, CMD_COMMON_TCP_UPLOAD, OPCODE_COMMON_TCP_UPLOAD, sizeof(*req)); req->id = (unsigned short)cid; req->upload_type = (unsigned char)upload_flag; - status = be_mcc_notify_wait(phba); - if (status) - SE_DEBUG(DBG_LVL_1, "mgmt_upload_connection Failed\n"); + be_mcc_notify(phba); spin_unlock(&ctrl->mbox_lock); - return status; + return tag; } int mgmt_open_connection(struct beiscsi_hba *phba, @@ -256,13 +270,13 @@ int mgmt_open_connection(struct beiscsi_hba *phba, struct sockaddr_in *daddr_in = (struct sockaddr_in *)dst_addr; struct sockaddr_in6 *daddr_in6 = (struct sockaddr_in6 *)dst_addr; struct be_ctrl_info *ctrl = &phba->ctrl; - struct be_mcc_wrb *wrb = wrb_from_mccq(phba); - struct tcp_connect_and_offload_in *req = embedded_payload(wrb); + struct be_mcc_wrb *wrb; + struct tcp_connect_and_offload_in *req; unsigned short def_hdr_id; unsigned short def_data_id; struct phys_addr template_address = { 0, 0 }; struct phys_addr *ptemplate_address; - int status = 0; + unsigned int tag = 0; unsigned int i; unsigned short cid = beiscsi_ep->ep_cid; @@ -274,7 +288,14 @@ int mgmt_open_connection(struct beiscsi_hba *phba, ptemplate_address = &template_address; ISCSI_GET_PDU_TEMPLATE_ADDRESS(phba, ptemplate_address); spin_lock(&ctrl->mbox_lock); - memset(wrb, 0, sizeof(*wrb)); + tag = alloc_mcc_tag(phba); + if (!tag) { + spin_unlock(&ctrl->mbox_lock); + return tag; + } + wrb = wrb_from_mccq(phba); + req = embedded_payload(wrb); + wrb->tag0 |= tag; be_wrb_hdr_prepare(wrb, sizeof(*req), true, 0); be_cmd_hdr_prepare(&req->hdr, CMD_SUBSYSTEM_ISCSI, @@ -319,47 +340,36 @@ int mgmt_open_connection(struct beiscsi_hba *phba, req->do_offload = 1; req->dataout_template_pa.lo = ptemplate_address->lo; req->dataout_template_pa.hi = ptemplate_address->hi; - status = be_mcc_notify_wait(phba); - if (!status) { - struct iscsi_endpoint *ep; - struct tcp_connect_and_offload_out *ptcpcnct_out = - embedded_payload(wrb); - - ep = phba->ep_array[ptcpcnct_out->cid - - phba->fw_config.iscsi_cid_start]; - beiscsi_ep = ep->dd_data; - beiscsi_ep->fw_handle = ptcpcnct_out->connection_handle; - beiscsi_ep->cid_vld = 1; - SE_DEBUG(DBG_LVL_8, "mgmt_open_connection Success\n"); - } else - SE_DEBUG(DBG_LVL_1, "mgmt_open_connection Failed\n"); + be_mcc_notify(phba); spin_unlock(&ctrl->mbox_lock); - return status; + return tag; } -int be_cmd_get_mac_addr(struct beiscsi_hba *phba, u8 *mac_addr) +unsigned int be_cmd_get_mac_addr(struct beiscsi_hba *phba) { struct be_ctrl_info *ctrl = &phba->ctrl; - struct be_mcc_wrb *wrb = wrb_from_mccq(phba); - struct be_cmd_req_get_mac_addr *req = embedded_payload(wrb); - int status; + struct be_mcc_wrb *wrb; + struct be_cmd_req_get_mac_addr *req; + unsigned int tag = 0; SE_DEBUG(DBG_LVL_8, "In be_cmd_get_mac_addr\n"); spin_lock(&ctrl->mbox_lock); - memset(wrb, 0, sizeof(*wrb)); + tag = alloc_mcc_tag(phba); + if (!tag) { + spin_unlock(&ctrl->mbox_lock); + return tag; + } + + wrb = wrb_from_mccq(phba); + req = embedded_payload(wrb); + wrb->tag0 |= tag; be_wrb_hdr_prepare(wrb, sizeof(*req), true, 0); be_cmd_hdr_prepare(&req->hdr, CMD_SUBSYSTEM_ISCSI, OPCODE_COMMON_ISCSI_NTWK_GET_NIC_CONFIG, sizeof(*req)); - status = be_mcc_notify_wait(phba); - if (!status) { - struct be_cmd_resp_get_mac_addr *resp = embedded_payload(wrb); - - memcpy(mac_addr, resp->mac_address, ETH_ALEN); - } - + be_mcc_notify(phba); spin_unlock(&ctrl->mbox_lock); - return status; + return tag; } diff --git a/drivers/scsi/be2iscsi/be_mgmt.h b/drivers/scsi/be2iscsi/be_mgmt.h index 6bc59e8..f873a66 100644 --- a/drivers/scsi/be2iscsi/be_mgmt.h +++ b/drivers/scsi/be2iscsi/be_mgmt.h @@ -250,7 +250,4 @@ unsigned char mgmt_invalidate_connection(struct beiscsi_hba *phba, unsigned short issue_reset, unsigned short savecfg_flag); -unsigned char mgmt_fw_cmd(struct be_ctrl_info *ctrl, - struct beiscsi_hba *phba, - char *buf, unsigned int len); #endif -- cgit v0.10.2 From da7408c800e3ae293275f52497d0ef7a9b09c9e4 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:11:23 +0530 Subject: [SCSI] be2iscsi: The session failure only when Link Goes down This fixes a situation where the sessions were being killed whenever LinkUP is notified rather than LinkDown Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_cmds.c b/drivers/scsi/be2iscsi/be_cmds.c index d4a0d1d..44e69bd 100644 --- a/drivers/scsi/be2iscsi/be_cmds.c +++ b/drivers/scsi/be2iscsi/be_cmds.c @@ -161,13 +161,13 @@ void beiscsi_async_link_state_process(struct beiscsi_hba *phba, SE_DEBUG(DBG_LVL_1, "Link Down on Physical Port %d \n", evt->physical_port); phba->state |= BE_ADAPTER_LINK_DOWN; + iscsi_host_for_each_session(phba->shost, + be2iscsi_fail_session); break; case ASYNC_EVENT_LINK_UP: phba->state = BE_ADAPTER_UP; SE_DEBUG(DBG_LVL_1, "Link UP on Physical Port %d \n", evt->physical_port); - iscsi_host_for_each_session(phba->shost, - be2iscsi_fail_session); break; default: SE_DEBUG(DBG_LVL_1, "Unexpected Async Notification %d on" -- cgit v0.10.2 From aa874f0738ecac7e5ac13b4d808a1026dcb5367d Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:12:03 +0530 Subject: [SCSI] be2iscsi: Fixing initialization of can_queue This patch fixes can_queue being uninitiallized since it was done before beiscsi_get_params was called. Thanks to Mike Christie for identifying this Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index a6a2c64..b1c8975 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -3825,8 +3825,8 @@ static int __devinit beiscsi_dev_probe(struct pci_dev *pcidev, goto free_port; } phba->shost->max_id = phba->fw_config.iscsi_cid_count; - phba->shost->can_queue = phba->params.ios_per_ctrl; beiscsi_get_params(phba); + phba->shost->can_queue = phba->params.ios_per_ctrl; ret = beiscsi_init_port(phba); if (ret < 0) { shost_printk(KERN_ERR, phba->shost, "beiscsi_dev_probe-" -- cgit v0.10.2 From 9db0fb3aa4cc4e42241e194ef64931321fa72196 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Tue, 5 Jan 2010 05:12:43 +0530 Subject: [SCSI] be2iscsi: Enable TEXT req resp This patch enables TEXT Request / Response for the driver Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index b1c8975..9fc4446 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -3917,7 +3917,7 @@ disable_pci: struct iscsi_transport beiscsi_iscsi_transport = { .owner = THIS_MODULE, .name = DRV_NAME, - .caps = CAP_RECOVERY_L0 | CAP_HDRDGST | + .caps = CAP_RECOVERY_L0 | CAP_HDRDGST | CAP_TEXT_NEGO | CAP_MULTI_R2T | CAP_DATADGST | CAP_DATA_PATH_OFFLOAD, .param_mask = ISCSI_MAX_RECV_DLENGTH | ISCSI_MAX_XMIT_DLENGTH | -- cgit v0.10.2 From 90a86fc05ffefe48581c88106d0b9cc37f6e060c Mon Sep 17 00:00:00 2001 From: Joe Carnuccio Date: Tue, 12 Jan 2010 13:02:46 -0800 Subject: [SCSI] qla2xxx: Enhance EEH support and enable AER support. qla2xxx: EEH added call to pci_restore_state. qla2xxx: EEH added delay in slot reset routine. qla2xxx: EEH moved call to pci_save_state(), see (1). qla2xxx: EEH additional changes for RHEL5.5. qla2xxx: EEH added function call, removed function call, see (2). (1) In qla2xxx_probe_one the call to pci_save_state() has been moved to after the call to qla2xxx_request_irqs(). (2) Add call to pci_disable_pcie_error_reporting() in remove_one. Delete call to pci_cleanup_aer_uncorrect_error_status() in pci_resume. Signed-off-by: Giridhar Malavali Signed-off-by: James Bottomley diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c index ffd0efd..5613456 100644 --- a/drivers/scsi/qla2xxx/qla_isr.c +++ b/drivers/scsi/qla2xxx/qla_isr.c @@ -2252,10 +2252,11 @@ qla2x00_free_irqs(scsi_qla_host_t *vha) if (ha->flags.msix_enabled) qla24xx_disable_msix(ha); - else if (ha->flags.inta_enabled) { + else if (ha->flags.msi_enabled) { free_irq(ha->pdev->irq, rsp); pci_disable_msi(ha->pdev); - } + } else + free_irq(ha->pdev->irq, rsp); } diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c index 8529eb1..fe34b7f 100644 --- a/drivers/scsi/qla2xxx/qla_os.c +++ b/drivers/scsi/qla2xxx/qla_os.c @@ -1818,7 +1818,6 @@ qla2x00_probe_one(struct pci_dev *pdev, const struct pci_device_id *id) /* Set EEH reset type to fundamental if required by hba */ if ( IS_QLA24XX(ha) || IS_QLA25XX(ha) || IS_QLA81XX(ha)) { pdev->needs_freset = 1; - pci_save_state(pdev); } /* Configure PCI I/O space */ @@ -1975,6 +1974,9 @@ qla2x00_probe_one(struct pci_dev *pdev, const struct pci_device_id *id) ret = qla2x00_request_irqs(ha, rsp); if (ret) goto probe_init_failed; + + pci_save_state(pdev); + /* Alloc arrays of request and response ring ptrs */ que_init: if (!qla2x00_alloc_queues(ha)) { @@ -2176,6 +2178,8 @@ qla2x00_remove_one(struct pci_dev *pdev) kfree(ha); ha = NULL; + pci_disable_pcie_error_reporting(pdev); + pci_disable_device(pdev); pci_set_drvdata(pdev, NULL); } @@ -3310,6 +3314,7 @@ qla2xxx_pci_error_detected(struct pci_dev *pdev, pci_channel_state_t state) return PCI_ERS_RESULT_CAN_RECOVER; case pci_channel_io_frozen: ha->flags.eeh_busy = 1; + qla2x00_free_irqs(vha); pci_disable_device(pdev); return PCI_ERS_RESULT_NEED_RESET; case pci_channel_io_perm_failure: @@ -3363,10 +3368,19 @@ qla2xxx_pci_slot_reset(struct pci_dev *pdev) pci_ers_result_t ret = PCI_ERS_RESULT_DISCONNECT; scsi_qla_host_t *base_vha = pci_get_drvdata(pdev); struct qla_hw_data *ha = base_vha->hw; - int rc; + struct rsp_que *rsp; + int rc, retries = 10; DEBUG17(qla_printk(KERN_WARNING, ha, "slot_reset\n")); + /* Workaround: qla2xxx driver which access hardware earlier + * needs error state to be pci_channel_io_online. + * Otherwise mailbox command timesout. + */ + pdev->error_state = pci_channel_io_normal; + + pci_restore_state(pdev); + if (ha->mem_only) rc = pci_enable_device_mem(pdev); else @@ -3378,27 +3392,23 @@ qla2xxx_pci_slot_reset(struct pci_dev *pdev) return ret; } + rsp = ha->rsp_q_map[0]; + if (qla2x00_request_irqs(ha, rsp)) + return ret; + if (ha->isp_ops->pci_config(base_vha)) return ret; -#ifdef QL_DEBUG_LEVEL_17 - { - uint8_t b; - uint32_t i; + while (ha->flags.mbox_busy && retries--) + msleep(1000); - printk("slot_reset_1: "); - for (i = 0; i < 256; i++) { - pci_read_config_byte(ha->pdev, i, &b); - printk("%s%02x", (i%16) ? " " : "\n", b); - } - printk("\n"); - } -#endif set_bit(ABORT_ISP_ACTIVE, &base_vha->dpc_flags); if (qla2x00_abort_isp(base_vha) == QLA_SUCCESS) ret = PCI_ERS_RESULT_RECOVERED; clear_bit(ABORT_ISP_ACTIVE, &base_vha->dpc_flags); + pci_cleanup_aer_uncorrect_error_status(pdev); + DEBUG17(qla_printk(KERN_WARNING, ha, "slot_reset-return:ret=%x\n", ret)); @@ -3422,8 +3432,6 @@ qla2xxx_pci_resume(struct pci_dev *pdev) } ha->flags.eeh_busy = 0; - - pci_cleanup_aer_uncorrect_error_status(pdev); } static struct pci_error_handlers qla2xxx_err_handler = { -- cgit v0.10.2 From 1838ef1d782f7527e6defe87e180598622d2d071 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Mon, 18 Jan 2010 14:00:34 -0800 Subject: x86-64, rwsem: 64-bit xadd rwsem implementation For x86-64, 32767 threads really is not enough. Change rwsem_count_t to a signed long, so that it is 64 bits on x86-64. This required the following changes to the assembly code: a) %z0 doesn't work on all versions of gcc! At least gcc 4.4.2 as shipped with Fedora 12 emits "ll" not "q" for 64 bits, even for integer operands. Newer gccs apparently do this correctly, but avoid this problem by using the _ASM_ macros instead of %z. b) 64 bits immediates are only allowed in "movq $imm,%reg" constructs... no others. Change some of the constraints to "e", and fix the one case where we would have had to use an invalid immediate -- in that case, we only care about the upper half anyway, so just access the upper half. Signed-off-by: H. Peter Anvin Cc: Linus Torvalds LKML-Reference: diff --git a/arch/x86/include/asm/rwsem.h b/arch/x86/include/asm/rwsem.h index 5f9af30..10204a2 100644 --- a/arch/x86/include/asm/rwsem.h +++ b/arch/x86/include/asm/rwsem.h @@ -41,6 +41,7 @@ #include #include #include +#include struct rwsem_waiter; @@ -56,18 +57,24 @@ extern asmregparm struct rw_semaphore * /* * the semaphore definition * - * The bias values and the counter type needs to be extended to 64 bits - * if we want to have more than 32767 potential readers/writers + * The bias values and the counter type limits the number of + * potential readers/writers to 32767 for 32 bits and 2147483647 + * for 64 bits. */ -#define RWSEM_UNLOCKED_VALUE 0x00000000 -#define RWSEM_ACTIVE_BIAS 0x00000001 -#define RWSEM_ACTIVE_MASK 0x0000ffff -#define RWSEM_WAITING_BIAS (-0x00010000) +#ifdef CONFIG_X86_64 +# define RWSEM_ACTIVE_MASK 0xffffffffL +#else +# define RWSEM_ACTIVE_MASK 0x0000ffffL +#endif + +#define RWSEM_UNLOCKED_VALUE 0x00000000L +#define RWSEM_ACTIVE_BIAS 0x00000001L +#define RWSEM_WAITING_BIAS (-RWSEM_ACTIVE_MASK-1) #define RWSEM_ACTIVE_READ_BIAS RWSEM_ACTIVE_BIAS #define RWSEM_ACTIVE_WRITE_BIAS (RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS) -typedef signed int rwsem_count_t; +typedef signed long rwsem_count_t; struct rw_semaphore { rwsem_count_t count; @@ -110,7 +117,7 @@ do { \ static inline void __down_read(struct rw_semaphore *sem) { asm volatile("# beginning down_read\n\t" - LOCK_PREFIX " inc%z0 (%1)\n\t" + LOCK_PREFIX _ASM_INC "(%1)\n\t" /* adds 0x00000001, returns the old value */ " jns 1f\n" " call call_rwsem_down_read_failed\n" @@ -225,8 +232,25 @@ static inline void __up_write(struct rw_semaphore *sem) */ static inline void __downgrade_write(struct rw_semaphore *sem) { +#ifdef CONFIG_X86_64 +# if RWSEM_WAITING_BIAS != -0x100000000 +# error "This code assumes RWSEM_WAITING_BIAS == -2^32" +# endif + + /* 64-bit immediates are special and expensive, and not needed here */ + asm volatile("# beginning __downgrade_write\n\t" + LOCK_PREFIX "incl 4(%1)\n\t" + /* transitions 0xZZZZZZZZ00000001 -> 0xYYYYYYYY00000001 */ + " jns 1f\n\t" + " call call_rwsem_downgrade_wake\n" + "1:\n\t" + "# ending __downgrade_write\n" + : "+m" (sem->count) + : "a" (sem) + : "memory", "cc"); +#else asm volatile("# beginning __downgrade_write\n\t" - LOCK_PREFIX " add%z0 %2,(%1)\n\t" + LOCK_PREFIX _ASM_ADD "%2,(%1)\n\t" /* transitions 0xZZZZ0001 -> 0xYYYY0001 */ " jns 1f\n\t" " call call_rwsem_downgrade_wake\n" @@ -235,22 +259,25 @@ static inline void __downgrade_write(struct rw_semaphore *sem) : "+m" (sem->count) : "a" (sem), "i" (-RWSEM_WAITING_BIAS) : "memory", "cc"); +#endif } /* * implement atomic add functionality */ -static inline void rwsem_atomic_add(int delta, struct rw_semaphore *sem) +static inline void rwsem_atomic_add(rwsem_count_t delta, + struct rw_semaphore *sem) { - asm volatile(LOCK_PREFIX "add%z0 %1,%0" + asm volatile(LOCK_PREFIX _ASM_ADD "%1,%0" : "+m" (sem->count) - : "ir" (delta)); + : "er" (delta)); } /* * implement exchange and add functionality */ -static inline rwsem_count_t rwsem_atomic_update(int delta, struct rw_semaphore *sem) +static inline rwsem_count_t rwsem_atomic_update(rwsem_count_t delta, + struct rw_semaphore *sem) { rwsem_count_t tmp = delta; -- cgit v0.10.2 From 9a069e196767d7b87184fd8d8211d22bb5b9c0b8 Mon Sep 17 00:00:00 2001 From: Giridhar Malavali Date: Tue, 12 Jan 2010 13:02:47 -0800 Subject: [SCSI] qla2xxx: Add BSG support for FC ELS/CT passthrough and vendor commands. [jejb: fixed printk casting issues] Signed-off-by: Sarang Radke Signed-off-by: Giridhar Malavali Signed-off-by: James Bottomley diff --git a/drivers/scsi/qla2xxx/qla_attr.c b/drivers/scsi/qla2xxx/qla_attr.c index 3a9f5b2..5a19aea 100644 --- a/drivers/scsi/qla2xxx/qla_attr.c +++ b/drivers/scsi/qla2xxx/qla_attr.c @@ -11,7 +11,9 @@ #include static int qla24xx_vport_disable(struct fc_vport *, bool); - +static int qla84xx_reset(scsi_qla_host_t *, struct msg_echo_lb *, struct fc_bsg_job *); +int qla84xx_reset_chip(scsi_qla_host_t *, uint16_t, uint16_t *); +static int qla84xx_mgmt_cmd(scsi_qla_host_t *, struct msg_echo_lb *, struct fc_bsg_job *); /* SYSFS attributes --------------------------------------------------------- */ static ssize_t @@ -1168,6 +1170,28 @@ qla2x00_total_isp_aborts_show(struct device *dev, } static ssize_t +qla24xx_84xx_fw_version_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + int rval = QLA_SUCCESS; + uint16_t status[2] = {0, 0}; + scsi_qla_host_t *vha = shost_priv(class_to_shost(dev)); + struct qla_hw_data *ha = vha->hw; + + if (IS_QLA84XX(ha) && ha->cs84xx) { + if (ha->cs84xx->op_fw_version == 0) { + rval = qla84xx_verify_chip(vha, status); + } + + if ((rval == QLA_SUCCESS) && (status[0] == 0)) + return snprintf(buf, PAGE_SIZE, "%u\n", + (uint32_t)ha->cs84xx->op_fw_version); + } + + return snprintf(buf, PAGE_SIZE, "\n"); +} + +static ssize_t qla2x00_mpi_version_show(struct device *dev, struct device_attribute *attr, char *buf) { @@ -1281,6 +1305,8 @@ static DEVICE_ATTR(optrom_fcode_version, S_IRUGO, qla2x00_optrom_fcode_version_show, NULL); static DEVICE_ATTR(optrom_fw_version, S_IRUGO, qla2x00_optrom_fw_version_show, NULL); +static DEVICE_ATTR(84xx_fw_version, S_IRUGO, qla24xx_84xx_fw_version_show, + NULL); static DEVICE_ATTR(total_isp_aborts, S_IRUGO, qla2x00_total_isp_aborts_show, NULL); static DEVICE_ATTR(mpi_version, S_IRUGO, qla2x00_mpi_version_show, NULL); @@ -1310,6 +1336,7 @@ struct device_attribute *qla2x00_host_attrs[] = { &dev_attr_optrom_efi_version, &dev_attr_optrom_fcode_version, &dev_attr_optrom_fw_version, + &dev_attr_84xx_fw_version, &dev_attr_total_isp_aborts, &dev_attr_mpi_version, &dev_attr_phy_version, @@ -1795,6 +1822,597 @@ qla24xx_vport_disable(struct fc_vport *fc_vport, bool disable) return 0; } +/* BSG support for ELS/CT pass through */ +inline srb_t * +qla2x00_get_ctx_bsg_sp(scsi_qla_host_t *vha, fc_port_t *fcport, size_t size) +{ + srb_t *sp; + struct qla_hw_data *ha = vha->hw; + struct srb_bsg_ctx *ctx; + + sp = mempool_alloc(ha->srb_mempool, GFP_KERNEL); + if (!sp) + goto done; + ctx = kzalloc(size, GFP_KERNEL); + if (!ctx) { + mempool_free(sp, ha->srb_mempool); + goto done; + } + + memset(sp, 0, sizeof(*sp)); + sp->fcport = fcport; + sp->ctx = ctx; +done: + return sp; +} + +static int +qla2x00_process_els(struct fc_bsg_job *bsg_job) +{ + struct fc_rport *rport; + fc_port_t *fcport; + struct Scsi_Host *host; + scsi_qla_host_t *vha; + struct qla_hw_data *ha; + srb_t *sp; + const char *type; + int req_sg_cnt, rsp_sg_cnt; + int rval = (DRIVER_ERROR << 16); + uint16_t nextlid = 0; + struct srb_bsg *els; + + /* Multiple SG's are not supported for ELS requests */ + if (bsg_job->request_payload.sg_cnt > 1 || + bsg_job->reply_payload.sg_cnt > 1) { + DEBUG2(printk(KERN_INFO + "multiple SG's are not supported for ELS requests" + " [request_sg_cnt: %x reply_sg_cnt: %x]\n", + bsg_job->request_payload.sg_cnt, + bsg_job->reply_payload.sg_cnt)); + rval = -EPERM; + goto done; + } + + /* ELS request for rport */ + if (bsg_job->request->msgcode == FC_BSG_RPT_ELS) { + rport = bsg_job->rport; + fcport = *(fc_port_t **) rport->dd_data; + host = rport_to_shost(rport); + vha = shost_priv(host); + ha = vha->hw; + type = "FC_BSG_RPT_ELS"; + + DEBUG2(printk(KERN_INFO + "scsi(%ld): loop-id=%x portid=%02x%02x%02x.\n", + fcport->vha->host_no, fcport->loop_id, + fcport->d_id.b.domain, fcport->d_id.b.area, + fcport->d_id.b.al_pa)); + + /* make sure the rport is logged in, + * if not perform fabric login + */ + if (qla2x00_fabric_login(vha, fcport, &nextlid)) { + DEBUG2(qla_printk(KERN_WARNING, ha, + "failed to login port %06X for ELS passthru\n", + fcport->d_id.b24)); + rval = -EIO; + goto done; + } + } else { + host = bsg_job->shost; + vha = shost_priv(host); + ha = vha->hw; + type = "FC_BSG_HST_ELS_NOLOGIN"; + + DEBUG2(printk(KERN_INFO + "scsi(%ld): loop-id=%x portid=%02x%02x%02x.\n", + vha->host_no, vha->loop_id, + vha->d_id.b.domain, vha->d_id.b.area, vha->d_id.b.al_pa)); + + /* Allocate a dummy fcport structure, since functions + * preparing the IOCB and mailbox command retrieves port + * specific information from fcport structure. For Host based + * ELS commands there will be no fcport structure allocated + */ + fcport = qla2x00_alloc_fcport(vha, GFP_KERNEL); + if (!fcport) { + rval = -ENOMEM; + goto done; + } + + /* Initialize all required fields of fcport */ + fcport->vha = vha; + fcport->vp_idx = vha->vp_idx; + fcport->d_id.b.al_pa = + bsg_job->request->rqst_data.h_els.port_id[0]; + fcport->d_id.b.area = + bsg_job->request->rqst_data.h_els.port_id[1]; + fcport->d_id.b.domain = + bsg_job->request->rqst_data.h_els.port_id[2]; + fcport->loop_id = + (fcport->d_id.b.al_pa == 0xFD) ? + NPH_FABRIC_CONTROLLER : NPH_F_PORT; + } + + DEBUG2(printk(KERN_INFO + "scsi(%ld): vendor-id = %llu\n", + vha->host_no, host->hostt->vendor_id)); + + req_sg_cnt = + dma_map_sg(&ha->pdev->dev, bsg_job->request_payload.sg_list, + bsg_job->request_payload.sg_cnt, DMA_TO_DEVICE); + if (!req_sg_cnt) { + rval = -ENOMEM; + goto done_free_fcport; + } + rsp_sg_cnt = dma_map_sg(&ha->pdev->dev, bsg_job->reply_payload.sg_list, + bsg_job->reply_payload.sg_cnt, DMA_FROM_DEVICE); + if (!rsp_sg_cnt) { + rval = -ENOMEM; + goto done_free_fcport; + } + + if ((req_sg_cnt != bsg_job->request_payload.sg_cnt) || + (rsp_sg_cnt != bsg_job->reply_payload.sg_cnt)) + { + DEBUG2(printk(KERN_INFO + "dma mapping resulted in different sg counts \ + [request_sg_cnt: %x dma_request_sg_cnt: %x\ + reply_sg_cnt: %x dma_reply_sg_cnt: %x]\n", + bsg_job->request_payload.sg_cnt, req_sg_cnt, + bsg_job->reply_payload.sg_cnt, rsp_sg_cnt)); + rval = -EAGAIN; + goto done_unmap_sg; + } + + /* Alloc SRB structure */ + sp = qla2x00_get_ctx_bsg_sp(vha, fcport, sizeof(struct srb_bsg)); + if (!sp) { + rval = -ENOMEM; + goto done_unmap_sg; + } + + els = sp->ctx; + els->ctx.type = + (bsg_job->request->msgcode == FC_BSG_RPT_ELS ? + SRB_ELS_CMD_RPT : SRB_ELS_CMD_HST); + els->bsg_job = bsg_job; + + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld:%x): bsg rqst type: %s els type: %x - loop-id=%x " + "portid=%02x%02x%02x.\n", vha->host_no, sp->handle, type, + bsg_job->request->rqst_data.h_els.command_code, + fcport->loop_id, fcport->d_id.b.domain, fcport->d_id.b.area, + fcport->d_id.b.al_pa)); + + rval = qla2x00_start_sp(sp); + if (rval != QLA_SUCCESS) { + kfree(sp->ctx); + mempool_free(sp, ha->srb_mempool); + rval = -EIO; + goto done_unmap_sg; + } + return rval; + +done_unmap_sg: + dma_unmap_sg(&ha->pdev->dev, bsg_job->request_payload.sg_list, + bsg_job->request_payload.sg_cnt, DMA_TO_DEVICE); + dma_unmap_sg(&ha->pdev->dev, bsg_job->reply_payload.sg_list, + bsg_job->reply_payload.sg_cnt, DMA_FROM_DEVICE); + goto done_free_fcport; + +done_free_fcport: + if (bsg_job->request->msgcode == FC_BSG_HST_ELS_NOLOGIN) + kfree(fcport); +done: + return rval; +} + +static int +qla2x00_process_ct(struct fc_bsg_job *bsg_job) +{ + srb_t *sp; + struct Scsi_Host *host = bsg_job->shost; + scsi_qla_host_t *vha = shost_priv(host); + struct qla_hw_data *ha = vha->hw; + int rval = (DRIVER_ERROR << 16); + int req_sg_cnt, rsp_sg_cnt; + uint16_t loop_id; + struct fc_port *fcport; + char *type = "FC_BSG_HST_CT"; + struct srb_bsg *ct; + + /* pass through is supported only for ISP 4Gb or higher */ + if (!IS_FWI2_CAPABLE(ha)) { + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld):Firmware is not capable to support FC " + "CT pass thru\n", vha->host_no)); + rval = -EPERM; + goto done; + } + + req_sg_cnt = + dma_map_sg(&ha->pdev->dev, bsg_job->request_payload.sg_list, + bsg_job->request_payload.sg_cnt, DMA_TO_DEVICE); + if (!req_sg_cnt) { + rval = -ENOMEM; + goto done; + } + + rsp_sg_cnt = dma_map_sg(&ha->pdev->dev, bsg_job->reply_payload.sg_list, + bsg_job->reply_payload.sg_cnt, DMA_FROM_DEVICE); + if (!rsp_sg_cnt) { + rval = -ENOMEM; + goto done; + } + + if ((req_sg_cnt != bsg_job->request_payload.sg_cnt) || + (rsp_sg_cnt != bsg_job->reply_payload.sg_cnt)) + { + DEBUG2(qla_printk(KERN_WARNING, ha, + "dma mapping resulted in different sg counts \ + [request_sg_cnt: %x dma_request_sg_cnt: %x\ + reply_sg_cnt: %x dma_reply_sg_cnt: %x]\n", + bsg_job->request_payload.sg_cnt, req_sg_cnt, + bsg_job->reply_payload.sg_cnt, rsp_sg_cnt)); + rval = -EAGAIN; + goto done_unmap_sg; + } + + loop_id = + (bsg_job->request->rqst_data.h_ct.preamble_word1 & 0xFF000000) + >> 24; + switch (loop_id) { + case 0xFC: + loop_id = cpu_to_le16(NPH_SNS); + break; + case 0xFA: + loop_id = vha->mgmt_svr_loop_id; + break; + default: + DEBUG2(qla_printk(KERN_INFO, ha, + "Unknown loop id: %x\n", loop_id)); + rval = -EINVAL; + goto done_unmap_sg; + } + + /* Allocate a dummy fcport structure, since functions preparing the + * IOCB and mailbox command retrieves port specific information + * from fcport structure. For Host based ELS commands there will be + * no fcport structure allocated + */ + fcport = qla2x00_alloc_fcport(vha, GFP_KERNEL); + if (!fcport) + { + rval = -ENOMEM; + goto done_unmap_sg; + } + + /* Initialize all required fields of fcport */ + fcport->vha = vha; + fcport->vp_idx = vha->vp_idx; + fcport->d_id.b.al_pa = bsg_job->request->rqst_data.h_ct.port_id[0]; + fcport->d_id.b.area = bsg_job->request->rqst_data.h_ct.port_id[1]; + fcport->d_id.b.domain = bsg_job->request->rqst_data.h_ct.port_id[2]; + fcport->loop_id = loop_id; + + /* Alloc SRB structure */ + sp = qla2x00_get_ctx_bsg_sp(vha, fcport, sizeof(struct srb_bsg)); + if (!sp) { + rval = -ENOMEM; + goto done_free_fcport; + } + + ct = sp->ctx; + ct->ctx.type = SRB_CT_CMD; + ct->bsg_job = bsg_job; + + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld:%x): bsg rqst type: %s els type: %x - loop-id=%x " + "portid=%02x%02x%02x.\n", vha->host_no, sp->handle, type, + (bsg_job->request->rqst_data.h_ct.preamble_word2 >> 16), + fcport->loop_id, fcport->d_id.b.domain, fcport->d_id.b.area, + fcport->d_id.b.al_pa)); + + rval = qla2x00_start_sp(sp); + if (rval != QLA_SUCCESS) { + kfree(sp->ctx); + mempool_free(sp, ha->srb_mempool); + rval = -EIO; + goto done_free_fcport; + } + return rval; + +done_free_fcport: + kfree(fcport); +done_unmap_sg: + dma_unmap_sg(&ha->pdev->dev, bsg_job->request_payload.sg_list, + bsg_job->request_payload.sg_cnt, DMA_TO_DEVICE); + dma_unmap_sg(&ha->pdev->dev, bsg_job->reply_payload.sg_list, + bsg_job->reply_payload.sg_cnt, DMA_FROM_DEVICE); +done: + return rval; +} + +static int +qla2x00_process_vendor_specific(struct fc_bsg_job *bsg_job) +{ + struct Scsi_Host *host = bsg_job->shost; + scsi_qla_host_t *vha = shost_priv(host); + struct qla_hw_data *ha = vha->hw; + int rval; + uint8_t command_sent; + uint32_t vendor_cmd; + char *type; + struct msg_echo_lb elreq; + uint16_t response[MAILBOX_REGISTER_COUNT]; + uint8_t* fw_sts_ptr; + uint8_t *req_data; + dma_addr_t req_data_dma; + uint32_t req_data_len; + uint8_t *rsp_data; + dma_addr_t rsp_data_dma; + uint32_t rsp_data_len; + + if (test_bit(ISP_ABORT_NEEDED, &vha->dpc_flags) || + test_bit(ABORT_ISP_ACTIVE, &vha->dpc_flags) || + test_bit(ISP_ABORT_RETRY, &vha->dpc_flags)) { + rval = -EBUSY; + goto done; + } + + elreq.req_sg_cnt = + dma_map_sg(&ha->pdev->dev, bsg_job->request_payload.sg_list, + bsg_job->request_payload.sg_cnt, DMA_TO_DEVICE); + if (!elreq.req_sg_cnt) { + rval = -ENOMEM; + goto done; + } + elreq.rsp_sg_cnt = + dma_map_sg(&ha->pdev->dev, bsg_job->reply_payload.sg_list, + bsg_job->reply_payload.sg_cnt, DMA_FROM_DEVICE); + if (!elreq.rsp_sg_cnt) { + rval = -ENOMEM; + goto done; + } + + if ((elreq.req_sg_cnt != bsg_job->request_payload.sg_cnt) || + (elreq.rsp_sg_cnt != bsg_job->reply_payload.sg_cnt)) + { + DEBUG2(printk(KERN_INFO + "dma mapping resulted in different sg counts \ + [request_sg_cnt: %x dma_request_sg_cnt: %x\ + reply_sg_cnt: %x dma_reply_sg_cnt: %x]\n", + bsg_job->request_payload.sg_cnt, elreq.req_sg_cnt, + bsg_job->reply_payload.sg_cnt, elreq.rsp_sg_cnt)); + rval = -EAGAIN; + goto done_unmap_sg; + } + req_data_len = rsp_data_len = bsg_job->request_payload.payload_len; + req_data = dma_alloc_coherent(&ha->pdev->dev, req_data_len, + &req_data_dma, GFP_KERNEL); + + rsp_data = dma_alloc_coherent(&ha->pdev->dev, rsp_data_len, + &rsp_data_dma, GFP_KERNEL); + + /* Copy the request buffer in req_data now */ + sg_copy_to_buffer(bsg_job->request_payload.sg_list, + bsg_job->request_payload.sg_cnt, req_data, + req_data_len); + + elreq.send_dma = req_data_dma; + elreq.rcv_dma = rsp_data_dma; + elreq.transfer_size = req_data_len; + + /* Vendor cmd : loopback or ECHO diagnostic + * Options: + * Loopback : Either internal or external loopback + * ECHO: ECHO ELS or Vendor specific FC4 link data + */ + vendor_cmd = bsg_job->request->rqst_data.h_vendor.vendor_cmd[0]; + elreq.options = + *(((uint32_t *)bsg_job->request->rqst_data.h_vendor.vendor_cmd) + + 1); + + switch (bsg_job->request->rqst_data.h_vendor.vendor_cmd[0]) { + case QL_VND_LOOPBACK: + if (ha->current_topology != ISP_CFG_F) { + type = "FC_BSG_HST_VENDOR_LOOPBACK"; + + if ((IS_QLA81XX(ha)) && + ((elreq.options == 0) || (elreq.options == 2))) { + DEBUG2(qla_printk(KERN_INFO, ha, "scsi(%ld)" + "loopback option:0x%x not supported\n", vha->host_no, elreq.options)); + rval = -EINVAL; + goto done_unmap_sg; + } + + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld) bsg rqst type: %s vendor rqst type: %x options: %x.\n", + vha->host_no, type, vendor_cmd, elreq.options)); + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld) tx_addr: 0x%llx rx_addr: 0x%llx tx_sg_cnt: %x rx_sg_cnt: %x\n", + vha->host_no, (unsigned long long)elreq.send_dma, (unsigned long long)elreq.rcv_dma, elreq.req_sg_cnt, elreq.rsp_sg_cnt)); + command_sent = INT_DEF_LB_LOOPBACK_CMD; + rval = qla2x00_loopback_test(vha, &elreq, response); + if (IS_QLA81XX(ha)) { + if (response[0] == MBS_COMMAND_ERROR && response[1] == MBS_LB_RESET) { + DEBUG2(printk(KERN_ERR "%s(%ld): ABORTing " + "ISP\n", __func__, vha->host_no)); + set_bit(ISP_ABORT_NEEDED, &vha->dpc_flags); + qla2xxx_wake_dpc(vha); + } + } + } else { + type = "FC_BSG_HST_VENDOR_ECHO_DIAG"; + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld) bsg rqst type: %s vendor rqst type: %x options: %x.\n", + vha->host_no, type, vendor_cmd, elreq.options)); + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld) tx_addr: 0x%llx rx_addr: 0x%llx tx_sg_cnt: %x rx_sg_cnt: %x\n", + vha->host_no, (unsigned long long)elreq.send_dma, (unsigned long long)elreq.rcv_dma, elreq.req_sg_cnt, elreq.rsp_sg_cnt)); + command_sent = INT_DEF_LB_ECHO_CMD; + rval = qla2x00_echo_test(vha, &elreq, response); + } + break; + case QLA84_RESET: + if (!IS_QLA84XX(vha->hw)) { + rval = -EINVAL; + DEBUG16(printk( + "%s(%ld): 8xxx exiting.\n", + __func__, vha->host_no)); + return rval; + } + rval = qla84xx_reset(vha, &elreq, bsg_job); + break; + case QLA84_MGMT_CMD: + if (!IS_QLA84XX(vha->hw)) { + rval = -EINVAL; + DEBUG16(printk( + "%s(%ld): 8xxx exiting.\n", + __func__, vha->host_no)); + return rval; + } + rval = qla84xx_mgmt_cmd(vha, &elreq, bsg_job); + break; + default: + rval = -ENOSYS; + } + + if (rval != QLA_SUCCESS) { + DEBUG2(qla_printk(KERN_WARNING, ha, + "scsi(%ld) Vendor request %s failed\n", vha->host_no, type)); + rval = 0; + bsg_job->reply->result = (DID_ERROR << 16); + fw_sts_ptr = ((uint8_t*)bsg_job->req->sense) + sizeof(struct fc_bsg_reply); + memcpy( fw_sts_ptr, response, sizeof(response)); + fw_sts_ptr += sizeof(response); + *fw_sts_ptr = command_sent; + } else { + DEBUG2(qla_printk(KERN_WARNING, ha, + "scsi(%ld) Vendor request %s completed\n", vha->host_no, type)); + rval = bsg_job->reply->result = 0; + bsg_job->reply_len = sizeof(struct fc_bsg_reply) + sizeof(response) + sizeof(uint8_t); + bsg_job->reply->reply_payload_rcv_len = bsg_job->reply_payload.payload_len; + fw_sts_ptr = ((uint8_t*)bsg_job->req->sense) + sizeof(struct fc_bsg_reply); + memcpy(fw_sts_ptr, response, sizeof(response)); + fw_sts_ptr += sizeof(response); + *fw_sts_ptr = command_sent; + sg_copy_from_buffer(bsg_job->reply_payload.sg_list, + bsg_job->reply_payload.sg_cnt, rsp_data, + rsp_data_len); + } + bsg_job->job_done(bsg_job); + +done_unmap_sg: + + if(req_data) + dma_free_coherent(&ha->pdev->dev, req_data_len, + req_data, req_data_dma); + dma_unmap_sg(&ha->pdev->dev, + bsg_job->request_payload.sg_list, + bsg_job->request_payload.sg_cnt, DMA_TO_DEVICE); + dma_unmap_sg(&ha->pdev->dev, + bsg_job->reply_payload.sg_list, + bsg_job->reply_payload.sg_cnt, DMA_FROM_DEVICE); + +done: + return rval; +} + +static int +qla24xx_bsg_request(struct fc_bsg_job *bsg_job) +{ + int ret = -EINVAL; + + switch (bsg_job->request->msgcode) { + case FC_BSG_RPT_ELS: + case FC_BSG_HST_ELS_NOLOGIN: + ret = qla2x00_process_els(bsg_job); + break; + case FC_BSG_HST_CT: + ret = qla2x00_process_ct(bsg_job); + break; + case FC_BSG_HST_VENDOR: + ret = qla2x00_process_vendor_specific(bsg_job); + break; + case FC_BSG_HST_ADD_RPORT: + case FC_BSG_HST_DEL_RPORT: + case FC_BSG_RPT_CT: + default: + DEBUG2(printk("qla2xxx: unsupported BSG request\n")); + break; + } + return ret; +} + +static int +qla24xx_bsg_timeout(struct fc_bsg_job *bsg_job) +{ + scsi_qla_host_t *vha = shost_priv(bsg_job->shost); + struct qla_hw_data *ha = vha->hw; + srb_t *sp; + int i; + unsigned long flags; + uint16_t que_id; + struct req_que *req; + struct rsp_que *rsp; + int found = 0; + struct srb_bsg *sp_bsg; + + /* find the bsg job from the active list of commands */ + spin_lock_irqsave(&ha->hardware_lock, flags); + req = ha->req_q_map[0]; + que_id = req->id; + if (req->rsp) + rsp = req->rsp; + else + rsp = ha->rsp_q_map[que_id]; + + for (i = 1; i < MAX_OUTSTANDING_COMMANDS; i++ ) { + sp = req->outstanding_cmds[i]; + + if (sp == NULL) + continue; + + sp_bsg = (struct srb_bsg*)sp->ctx; + + if (((sp_bsg->ctx.type == SRB_CT_CMD) || + (sp_bsg->ctx.type == SRB_ELS_CMD_RPT) + || ( sp_bsg->ctx.type == SRB_ELS_CMD_HST)) && + (sp_bsg->bsg_job == bsg_job)) { + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld) req_q: %p rsp_q: %p que_id: %x sp: %p\n", + vha->host_no, req, rsp, que_id, sp)); + found = 1; + break; + } + } + spin_unlock_irqrestore(&ha->hardware_lock, flags); + if (!found) { + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld) SRB not found to abort\n", vha->host_no)); + bsg_job->req->errors = bsg_job->reply->result = -ENXIO; + return 0; + } + + if (ha->isp_ops->abort_command(sp)) { + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld): mbx abort_command failed\n", vha->host_no)); + bsg_job->req->errors = bsg_job->reply->result = -EIO; + } else { + DEBUG2(qla_printk(KERN_INFO, ha, + "scsi(%ld): mbx abort_command success\n", vha->host_no)); + bsg_job->req->errors = bsg_job->reply->result = 0; + } + + if (bsg_job->request->msgcode == FC_BSG_HST_CT) + kfree(sp->fcport); + kfree(sp->ctx); + mempool_free(sp, ha->srb_mempool); + return 0; +} + struct fc_function_template qla2xxx_transport_functions = { .show_host_node_name = 1, @@ -1838,6 +2456,8 @@ struct fc_function_template qla2xxx_transport_functions = { .vport_create = qla24xx_vport_create, .vport_disable = qla24xx_vport_disable, .vport_delete = qla24xx_vport_delete, + .bsg_request = qla24xx_bsg_request, + .bsg_timeout = qla24xx_bsg_timeout, }; struct fc_function_template qla2xxx_transport_vport_functions = { @@ -1878,6 +2498,8 @@ struct fc_function_template qla2xxx_transport_vport_functions = { .dev_loss_tmo_callbk = qla2x00_dev_loss_tmo_callbk, .terminate_rport_io = qla2x00_terminate_rport_io, .get_fc_host_stats = qla2x00_get_fc_host_stats, + .bsg_request = qla24xx_bsg_request, + .bsg_timeout = qla24xx_bsg_timeout, }; void @@ -1906,3 +2528,125 @@ qla2x00_init_host_attr(scsi_qla_host_t *vha) speed = FC_PORTSPEED_1GBIT; fc_host_supported_speeds(vha->host) = speed; } +static int +qla84xx_reset(scsi_qla_host_t *ha, struct msg_echo_lb *mreq, struct fc_bsg_job *bsg_job) +{ + int ret = 0; + int cmd; + uint16_t cmd_status; + + DEBUG16(printk("%s(%ld): entered.\n", __func__, ha->host_no)); + + cmd = (*((bsg_job->request->rqst_data.h_vendor.vendor_cmd) + 2)) + == A84_RESET_FLAG_ENABLE_DIAG_FW ? + A84_ISSUE_RESET_DIAG_FW : A84_ISSUE_RESET_OP_FW; + ret = qla84xx_reset_chip(ha, cmd == A84_ISSUE_RESET_DIAG_FW, + &cmd_status); + return ret; +} + +static int +qla84xx_mgmt_cmd(scsi_qla_host_t *ha, struct msg_echo_lb *mreq, struct fc_bsg_job *bsg_job) +{ + struct access_chip_84xx *mn; + dma_addr_t mn_dma, mgmt_dma; + void *mgmt_b = NULL; + int ret = 0; + int rsp_hdr_len, len = 0; + struct qla84_msg_mgmt *ql84_mgmt; + + ql84_mgmt = (struct qla84_msg_mgmt *) vmalloc(sizeof(struct qla84_msg_mgmt)); + ql84_mgmt->cmd = + *((uint16_t *)(bsg_job->request->rqst_data.h_vendor.vendor_cmd + 2)); + ql84_mgmt->mgmtp.u.mem.start_addr = + *((uint32_t *)(bsg_job->request->rqst_data.h_vendor.vendor_cmd + 3)); + ql84_mgmt->len = + *((uint32_t *)(bsg_job->request->rqst_data.h_vendor.vendor_cmd + 4)); + ql84_mgmt->mgmtp.u.config.id = + *((uint32_t *)(bsg_job->request->rqst_data.h_vendor.vendor_cmd + 5)); + ql84_mgmt->mgmtp.u.config.param0 = + *((uint32_t *)(bsg_job->request->rqst_data.h_vendor.vendor_cmd + 6)); + ql84_mgmt->mgmtp.u.config.param1 = + *((uint32_t *)(bsg_job->request->rqst_data.h_vendor.vendor_cmd + 7)); + ql84_mgmt->mgmtp.u.info.type = + *((uint32_t *)(bsg_job->request->rqst_data.h_vendor.vendor_cmd + 8)); + ql84_mgmt->mgmtp.u.info.context = + *((uint32_t *)(bsg_job->request->rqst_data.h_vendor.vendor_cmd + 9)); + + rsp_hdr_len = bsg_job->request_payload.payload_len; + + mn = dma_pool_alloc(ha->hw->s_dma_pool, GFP_KERNEL, &mn_dma); + if (mn == NULL) { + DEBUG2(printk(KERN_ERR "%s: dma alloc for fw buffer " + "failed%lu\n", __func__, ha->host_no)); + return -ENOMEM; + } + + memset(mn, 0, sizeof (struct access_chip_84xx)); + + mn->entry_type = ACCESS_CHIP_IOCB_TYPE; + mn->entry_count = 1; + + switch (ql84_mgmt->cmd) { + case QLA84_MGMT_READ_MEM: + mn->options = cpu_to_le16(ACO_DUMP_MEMORY); + mn->parameter1 = cpu_to_le32(ql84_mgmt->mgmtp.u.mem.start_addr); + break; + case QLA84_MGMT_WRITE_MEM: + mn->options = cpu_to_le16(ACO_LOAD_MEMORY); + mn->parameter1 = cpu_to_le32(ql84_mgmt->mgmtp.u.mem.start_addr); + break; + case QLA84_MGMT_CHNG_CONFIG: + mn->options = cpu_to_le16(ACO_CHANGE_CONFIG_PARAM); + mn->parameter1 = cpu_to_le32(ql84_mgmt->mgmtp.u.config.id); + mn->parameter2 = cpu_to_le32(ql84_mgmt->mgmtp.u.config.param0); + mn->parameter3 = cpu_to_le32(ql84_mgmt->mgmtp.u.config.param1); + break; + case QLA84_MGMT_GET_INFO: + mn->options = cpu_to_le16(ACO_REQUEST_INFO); + mn->parameter1 = cpu_to_le32(ql84_mgmt->mgmtp.u.info.type); + mn->parameter2 = cpu_to_le32(ql84_mgmt->mgmtp.u.info.context); + break; + default: + ret = -EIO; + goto exit_mgmt0; + } + + if ((len == ql84_mgmt->len) && + ql84_mgmt->cmd != QLA84_MGMT_CHNG_CONFIG) { + mgmt_b = dma_alloc_coherent(&ha->hw->pdev->dev, len, + &mgmt_dma, GFP_KERNEL); + if (mgmt_b == NULL) { + DEBUG2(printk(KERN_ERR "%s: dma alloc mgmt_b " + "failed%lu\n", __func__, ha->host_no)); + ret = -ENOMEM; + goto exit_mgmt0; + } + mn->total_byte_cnt = cpu_to_le32(ql84_mgmt->len); + mn->dseg_count = cpu_to_le16(1); + mn->dseg_address[0] = cpu_to_le32(LSD(mgmt_dma)); + mn->dseg_address[1] = cpu_to_le32(MSD(mgmt_dma)); + mn->dseg_length = cpu_to_le32(len); + + if (ql84_mgmt->cmd == QLA84_MGMT_WRITE_MEM) { + memcpy(mgmt_b, ql84_mgmt->payload, len); + } + } + + ret = qla2x00_issue_iocb(ha, mn, mn_dma, 0); + if ((ret != QLA_SUCCESS) || (ql84_mgmt->cmd == QLA84_MGMT_WRITE_MEM) + || (ql84_mgmt->cmd == QLA84_MGMT_CHNG_CONFIG)) { + if (ret != QLA_SUCCESS) + DEBUG2(printk(KERN_ERR "%s(%lu): failed\n", + __func__, ha->host_no)); + } else if ((ql84_mgmt->cmd == QLA84_MGMT_READ_MEM) || + (ql84_mgmt->cmd == QLA84_MGMT_GET_INFO)) { + } + + if (mgmt_b) + dma_free_coherent(&ha->hw->pdev->dev, len, mgmt_b, mgmt_dma); + +exit_mgmt0: + dma_pool_free(ha->hw->s_dma_pool, mn, mn_dma); + return ret; +} diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h index 1263d97..afa9561 100644 --- a/drivers/scsi/qla2xxx/qla_def.h +++ b/drivers/scsi/qla2xxx/qla_def.h @@ -31,6 +31,7 @@ #include #include #include +#include #define QLA2XXX_DRIVER_NAME "qla2xxx" @@ -228,6 +229,27 @@ struct srb_logio { uint16_t flags; }; +struct srb_bsg_ctx { +#define SRB_ELS_CMD_RPT 3 +#define SRB_ELS_CMD_HST 4 +#define SRB_CT_CMD 5 + uint16_t type; +}; + +struct srb_bsg { + struct srb_bsg_ctx ctx; + struct fc_bsg_job *bsg_job; +}; + +struct msg_echo_lb { + dma_addr_t send_dma; + dma_addr_t rcv_dma; + uint16_t req_sg_cnt; + uint16_t rsp_sg_cnt; + uint16_t options; + uint32_t transfer_size; +}; + /* * ISP I/O Register Set structure definitions. */ @@ -522,6 +544,8 @@ typedef struct { #define MBA_DISCARD_RND_FRAME 0x8048 /* discard RND frame due to error. */ #define MBA_REJECTED_FCP_CMD 0x8049 /* rejected FCP_CMD. */ +/* ISP mailbox loopback echo diagnostic error code */ +#define MBS_LB_RESET 0x17 /* * Firmware options 1, 2, 3. */ @@ -2230,6 +2254,13 @@ struct req_que { int max_q_depth; }; +/* Place holder for FW buffer parameters */ +struct qlfc_fw { + void *fw_buf; + dma_addr_t fw_dma; + uint32_t len; +}; + /* * Qlogic host adapter specific data structure. */ @@ -2594,6 +2625,7 @@ struct qla_hw_data { struct qla_statistics qla_stats; struct isp_operations *isp_ops; struct workqueue_struct *wq; + struct qlfc_fw fw_buf; }; /* @@ -2766,4 +2798,127 @@ typedef struct scsi_qla_host { #define CMD_SP(Cmnd) ((Cmnd)->SCp.ptr) +/* + * BSG Vendor specific commands + */ + +#define QL_VND_LOOPBACK 0x01 +#define QLA84_RESET 0x02 +#define QLA84_UPDATE_FW 0x03 +#define QLA84_MGMT_CMD 0x04 + +/* BSG definations for interpreting CommandSent field */ +#define INT_DEF_LB_LOOPBACK_CMD 0 +#define INT_DEF_LB_ECHO_CMD 1 + +/* BSG Vendor specific definations */ +typedef struct _A84_RESET { + uint16_t Flags; + uint16_t Reserved; +#define A84_RESET_FLAG_ENABLE_DIAG_FW 1 +} __attribute__((packed)) A84_RESET, *PA84_RESET; + +#define A84_ISSUE_WRITE_TYPE_CMD 0 +#define A84_ISSUE_READ_TYPE_CMD 1 +#define A84_CLEANUP_CMD 2 +#define A84_ISSUE_RESET_OP_FW 3 +#define A84_ISSUE_RESET_DIAG_FW 4 +#define A84_ISSUE_UPDATE_OPFW_CMD 5 +#define A84_ISSUE_UPDATE_DIAGFW_CMD 6 + +struct qla84_mgmt_param { + union { + struct { + uint32_t start_addr; + } mem; /* for QLA84_MGMT_READ/WRITE_MEM */ + struct { + uint32_t id; +#define QLA84_MGMT_CONFIG_ID_UIF 1 +#define QLA84_MGMT_CONFIG_ID_FCOE_COS 2 +#define QLA84_MGMT_CONFIG_ID_PAUSE 3 +#define QLA84_MGMT_CONFIG_ID_TIMEOUTS 4 + + uint32_t param0; + uint32_t param1; + } config; /* for QLA84_MGMT_CHNG_CONFIG */ + + struct { + uint32_t type; +#define QLA84_MGMT_INFO_CONFIG_LOG_DATA 1 /* Get Config Log Data */ +#define QLA84_MGMT_INFO_LOG_DATA 2 /* Get Log Data */ +#define QLA84_MGMT_INFO_PORT_STAT 3 /* Get Port Statistics */ +#define QLA84_MGMT_INFO_LIF_STAT 4 /* Get LIF Statistics */ +#define QLA84_MGMT_INFO_ASIC_STAT 5 /* Get ASIC Statistics */ +#define QLA84_MGMT_INFO_CONFIG_PARAMS 6 /* Get Config Parameters */ +#define QLA84_MGMT_INFO_PANIC_LOG 7 /* Get Panic Log */ + + uint32_t context; +/* +* context definitions for QLA84_MGMT_INFO_CONFIG_LOG_DATA +*/ +#define IC_LOG_DATA_LOG_ID_DEBUG_LOG 0 +#define IC_LOG_DATA_LOG_ID_LEARN_LOG 1 +#define IC_LOG_DATA_LOG_ID_FC_ACL_INGRESS_LOG 2 +#define IC_LOG_DATA_LOG_ID_FC_ACL_EGRESS_LOG 3 +#define IC_LOG_DATA_LOG_ID_ETHERNET_ACL_INGRESS_LOG 4 +#define IC_LOG_DATA_LOG_ID_ETHERNET_ACL_EGRESS_LOG 5 +#define IC_LOG_DATA_LOG_ID_MESSAGE_TRANSMIT_LOG 6 +#define IC_LOG_DATA_LOG_ID_MESSAGE_RECEIVE_LOG 7 +#define IC_LOG_DATA_LOG_ID_LINK_EVENT_LOG 8 +#define IC_LOG_DATA_LOG_ID_DCX_LOG 9 + +/* +* context definitions for QLA84_MGMT_INFO_PORT_STAT +*/ +#define IC_PORT_STATISTICS_PORT_NUMBER_ETHERNET_PORT0 0 +#define IC_PORT_STATISTICS_PORT_NUMBER_ETHERNET_PORT1 1 +#define IC_PORT_STATISTICS_PORT_NUMBER_NSL_PORT0 2 +#define IC_PORT_STATISTICS_PORT_NUMBER_NSL_PORT1 3 +#define IC_PORT_STATISTICS_PORT_NUMBER_FC_PORT0 4 +#define IC_PORT_STATISTICS_PORT_NUMBER_FC_PORT1 5 + + +/* +* context definitions for QLA84_MGMT_INFO_LIF_STAT +*/ +#define IC_LIF_STATISTICS_LIF_NUMBER_ETHERNET_PORT0 0 +#define IC_LIF_STATISTICS_LIF_NUMBER_ETHERNET_PORT1 1 +#define IC_LIF_STATISTICS_LIF_NUMBER_FC_PORT0 2 +#define IC_LIF_STATISTICS_LIF_NUMBER_FC_PORT1 3 +#define IC_LIF_STATISTICS_LIF_NUMBER_CPU 6 + + } info; /* for QLA84_MGMT_GET_INFO */ + } u; +}; + +struct qla84_msg_mgmt { + uint16_t cmd; +#define QLA84_MGMT_READ_MEM 0x00 +#define QLA84_MGMT_WRITE_MEM 0x01 +#define QLA84_MGMT_CHNG_CONFIG 0x02 +#define QLA84_MGMT_GET_INFO 0x03 + uint16_t rsrvd; + struct qla84_mgmt_param mgmtp;/* parameters for cmd */ + uint32_t len; /* bytes in payload following this struct */ + uint8_t payload[0]; /* payload for cmd */ +}; + +struct msg_update_fw { + /* + * diag_fw = 0 operational fw + * otherwise diagnostic fw + * offset, len, fw_len are present to overcome the current limitation + * of 128Kb xfer size. The fw is sent in smaller chunks. Each chunk + * specifies the byte "offset" where it fits in the fw buffer. The + * number of bytes in each chunk is specified in "len". "fw_len" + * is the total size of fw. The first chunk should start at offset = 0. + * When offset+len == fw_len, the fw is written to the HBA. + */ + uint32_t diag_fw; + uint32_t offset;/* start offset */ + uint32_t len; /* num bytes in cur xfer */ + uint32_t fw_len; /* size of fw in bytes */ + uint8_t fw_bytes[0]; +}; + #endif diff --git a/drivers/scsi/qla2xxx/qla_fw.h b/drivers/scsi/qla2xxx/qla_fw.h index 66a8da5..cebf4f1 100644 --- a/drivers/scsi/qla2xxx/qla_fw.h +++ b/drivers/scsi/qla2xxx/qla_fw.h @@ -627,6 +627,39 @@ struct els_entry_24xx { uint32_t rx_len; /* Data segment 1 length. */ }; +struct els_sts_entry_24xx { + uint8_t entry_type; /* Entry type. */ + uint8_t entry_count; /* Entry count. */ + uint8_t sys_define; /* System Defined. */ + uint8_t entry_status; /* Entry Status. */ + + uint32_t handle; /* System handle. */ + + uint16_t comp_status; + + uint16_t nport_handle; /* N_PORT handle. */ + + uint16_t reserved_1; + + uint8_t vp_index; + uint8_t sof_type; + + uint32_t rx_xchg_address; /* Receive exchange address. */ + uint16_t reserved_2; + + uint8_t opcode; + uint8_t reserved_3; + + uint8_t port_id[3]; + uint8_t reserved_4; + + uint16_t reserved_5; + + uint16_t control_flags; /* Control flags. */ + uint32_t total_byte_count; + uint32_t error_subcode_1; + uint32_t error_subcode_2; +}; /* * ISP queue - Mailbox Command entry structure definition. */ diff --git a/drivers/scsi/qla2xxx/qla_gbl.h b/drivers/scsi/qla2xxx/qla_gbl.h index f61fb8d..b42e704 100644 --- a/drivers/scsi/qla2xxx/qla_gbl.h +++ b/drivers/scsi/qla2xxx/qla_gbl.h @@ -60,6 +60,8 @@ extern int qla2x00_async_login_done(struct scsi_qla_host *, fc_port_t *, extern int qla2x00_async_logout_done(struct scsi_qla_host *, fc_port_t *, uint16_t *); +extern fc_port_t * +qla2x00_alloc_fcport(scsi_qla_host_t *, gfp_t ); /* * Global Data in qla_os.c source file. */ @@ -154,6 +156,7 @@ int qla2x00_marker(struct scsi_qla_host *, struct req_que *, struct rsp_que *, int __qla2x00_marker(struct scsi_qla_host *, struct req_que *, struct rsp_que *, uint16_t, uint16_t, uint8_t); extern int qla2x00_start_sp(srb_t *); +extern void qla2x00_ctx_sp_free(srb_t *); /* * Global Function Prototypes in qla_mbx.c source file. @@ -426,6 +429,8 @@ extern void qla2x00_free_sysfs_attr(scsi_qla_host_t *); extern void qla2x00_init_host_attr(scsi_qla_host_t *); extern void qla2x00_alloc_sysfs_attr(scsi_qla_host_t *); extern void qla2x00_free_sysfs_attr(scsi_qla_host_t *); +extern int qla2x00_loopback_test(scsi_qla_host_t *, struct msg_echo_lb *, uint16_t *); +extern int qla2x00_echo_test(scsi_qla_host_t *, struct msg_echo_lb *, uint16_t *); /* * Global Function Prototypes in qla_dfs.c source file. diff --git a/drivers/scsi/qla2xxx/qla_init.c b/drivers/scsi/qla2xxx/qla_init.c index 3f8e849..1128c8d 100644 --- a/drivers/scsi/qla2xxx/qla_init.c +++ b/drivers/scsi/qla2xxx/qla_init.c @@ -62,7 +62,7 @@ qla2x00_ctx_sp_timeout(unsigned long __data) ctx->free(sp); } -static void +void qla2x00_ctx_sp_free(srb_t *sp) { struct srb_ctx *ctx = sp->ctx; @@ -338,6 +338,16 @@ qla2x00_initialize_adapter(scsi_qla_host_t *vha) rval = qla2x00_init_rings(vha); ha->flags.chip_reset_done = 1; + if (rval == QLA_SUCCESS && IS_QLA84XX(ha)) { + /* Issue verify 84xx FW IOCB to complete 84xx initialization */ + rval = qla84xx_init_chip(vha); + if (rval != QLA_SUCCESS) { + qla_printk(KERN_ERR, ha, + "Unable to initialize ISP84XX.\n"); + qla84xx_put_chip(vha); + } + } + return (rval); } @@ -2216,7 +2226,7 @@ qla2x00_rport_del(void *data) * * Returns a pointer to the allocated fcport, or NULL, if none available. */ -static fc_port_t * +fc_port_t * qla2x00_alloc_fcport(scsi_qla_host_t *vha, gfp_t flags) { fc_port_t *fcport; diff --git a/drivers/scsi/qla2xxx/qla_iocb.c b/drivers/scsi/qla2xxx/qla_iocb.c index c5ccac0..8299a98 100644 --- a/drivers/scsi/qla2xxx/qla_iocb.c +++ b/drivers/scsi/qla2xxx/qla_iocb.c @@ -1025,6 +1025,119 @@ qla2x00_logout_iocb(srb_t *sp, struct mbx_entry *mbx) /* Implicit: mbx->mbx10 = 0. */ } +static void +qla24xx_els_iocb(srb_t *sp, struct els_entry_24xx *els_iocb) +{ + struct fc_bsg_job *bsg_job = ((struct srb_bsg*)sp->ctx)->bsg_job; + + els_iocb->entry_type = ELS_IOCB_TYPE; + els_iocb->entry_count = 1; + els_iocb->sys_define = 0; + els_iocb->entry_status = 0; + els_iocb->handle = sp->handle; + els_iocb->nport_handle = cpu_to_le16(sp->fcport->loop_id); + els_iocb->tx_dsd_count = __constant_cpu_to_le16(bsg_job->request_payload.sg_cnt); + els_iocb->vp_index = sp->fcport->vp_idx; + els_iocb->sof_type = EST_SOFI3; + els_iocb->rx_dsd_count = __constant_cpu_to_le16(bsg_job->reply_payload.sg_cnt); + + els_iocb->opcode =(((struct srb_bsg*)sp->ctx)->ctx.type == SRB_ELS_CMD_RPT) ? + bsg_job->request->rqst_data.r_els.els_code : bsg_job->request->rqst_data.h_els.command_code; + els_iocb->port_id[0] = sp->fcport->d_id.b.al_pa; + els_iocb->port_id[1] = sp->fcport->d_id.b.area; + els_iocb->port_id[2] = sp->fcport->d_id.b.domain; + els_iocb->control_flags = 0; + els_iocb->rx_byte_count = + cpu_to_le32(bsg_job->reply_payload.payload_len); + els_iocb->tx_byte_count = + cpu_to_le32(bsg_job->request_payload.payload_len); + + els_iocb->tx_address[0] = cpu_to_le32(LSD(sg_dma_address + (bsg_job->request_payload.sg_list))); + els_iocb->tx_address[1] = cpu_to_le32(MSD(sg_dma_address + (bsg_job->request_payload.sg_list))); + els_iocb->tx_len = cpu_to_le32(sg_dma_len + (bsg_job->request_payload.sg_list)); + + els_iocb->rx_address[0] = cpu_to_le32(LSD(sg_dma_address + (bsg_job->reply_payload.sg_list))); + els_iocb->rx_address[1] = cpu_to_le32(MSD(sg_dma_address + (bsg_job->reply_payload.sg_list))); + els_iocb->rx_len = cpu_to_le32(sg_dma_len + (bsg_job->reply_payload.sg_list)); +} + +static void +qla24xx_ct_iocb(srb_t *sp, struct ct_entry_24xx *ct_iocb) +{ + uint16_t avail_dsds; + uint32_t *cur_dsd; + struct scatterlist *sg; + int index; + uint16_t tot_dsds; + scsi_qla_host_t *vha = sp->fcport->vha; + struct fc_bsg_job *bsg_job = ((struct srb_bsg*)sp->ctx)->bsg_job; + int loop_iterartion = 0; + int cont_iocb_prsnt = 0; + int entry_count = 1; + + ct_iocb->entry_type = CT_IOCB_TYPE; + ct_iocb->entry_status = 0; + ct_iocb->sys_define = 0; + ct_iocb->handle = sp->handle; + + ct_iocb->nport_handle = cpu_to_le16(sp->fcport->loop_id); + ct_iocb->vp_index = sp->fcport->vp_idx; + ct_iocb->comp_status = __constant_cpu_to_le16(0); + + ct_iocb->cmd_dsd_count = + __constant_cpu_to_le16(bsg_job->request_payload.sg_cnt); + ct_iocb->timeout = 0; + ct_iocb->rsp_dsd_count = + __constant_cpu_to_le16(bsg_job->reply_payload.sg_cnt); + ct_iocb->rsp_byte_count = + cpu_to_le32(bsg_job->reply_payload.payload_len); + ct_iocb->cmd_byte_count = + cpu_to_le32(bsg_job->request_payload.payload_len); + ct_iocb->dseg_0_address[0] = cpu_to_le32(LSD(sg_dma_address + (bsg_job->request_payload.sg_list))); + ct_iocb->dseg_0_address[1] = cpu_to_le32(MSD(sg_dma_address + (bsg_job->request_payload.sg_list))); + ct_iocb->dseg_0_len = cpu_to_le32(sg_dma_len + (bsg_job->request_payload.sg_list)); + + avail_dsds = 1; + cur_dsd = (uint32_t *)ct_iocb->dseg_1_address; + index = 0; + tot_dsds = bsg_job->reply_payload.sg_cnt; + + for_each_sg(bsg_job->reply_payload.sg_list, sg, tot_dsds, index) { + dma_addr_t sle_dma; + cont_a64_entry_t *cont_pkt; + + /* Allocate additional continuation packets? */ + if (avail_dsds == 0) { + /* + * Five DSDs are available in the Cont. + * Type 1 IOCB. + */ + cont_pkt = qla2x00_prep_cont_type1_iocb(vha); + cur_dsd = (uint32_t *) cont_pkt->dseg_0_address; + avail_dsds = 5; + cont_iocb_prsnt = 1; + entry_count++; + } + + sle_dma = sg_dma_address(sg); + *cur_dsd++ = cpu_to_le32(LSD(sle_dma)); + *cur_dsd++ = cpu_to_le32(MSD(sle_dma)); + *cur_dsd++ = cpu_to_le32(sg_dma_len(sg)); + loop_iterartion++; + avail_dsds--; + } + ct_iocb->entry_count = entry_count; +} + int qla2x00_start_sp(srb_t *sp) { @@ -1052,6 +1165,13 @@ qla2x00_start_sp(srb_t *sp) qla24xx_logout_iocb(sp, pkt): qla2x00_logout_iocb(sp, pkt); break; + case SRB_ELS_CMD_RPT: + case SRB_ELS_CMD_HST: + qla24xx_els_iocb(sp, pkt); + break; + case SRB_CT_CMD: + qla24xx_ct_iocb(sp, pkt); + break; default: break; } diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c index 5613456..09ba26b 100644 --- a/drivers/scsi/qla2xxx/qla_isr.c +++ b/drivers/scsi/qla2xxx/qla_isr.c @@ -8,6 +8,7 @@ #include #include +#include static void qla2x00_mbx_completion(scsi_qla_host_t *, uint16_t); static void qla2x00_process_completed_request(struct scsi_qla_host *, @@ -881,7 +882,9 @@ qla2x00_get_sp_from_handle(scsi_qla_host_t *vha, const char *func, index); return NULL; } + req->outstanding_cmds[index] = NULL; + done: return sp; } @@ -982,6 +985,100 @@ done_post_logio_done_work: } static void +qla24xx_els_ct_entry(scsi_qla_host_t *vha, struct req_que *req, + struct sts_entry_24xx *pkt, int iocb_type) +{ + const char func[] = "ELS_CT_IOCB"; + const char *type; + struct qla_hw_data *ha = vha->hw; + srb_t *sp; + struct srb_bsg *sp_bsg; + struct fc_bsg_job *bsg_job; + uint16_t comp_status; + uint32_t fw_status[3]; + uint8_t* fw_sts_ptr; + + sp = qla2x00_get_sp_from_handle(vha, func, req, pkt); + if (!sp) + return; + sp_bsg = (struct srb_bsg*)sp->ctx; + bsg_job = sp_bsg->bsg_job; + + type = NULL; + switch (sp_bsg->ctx.type) { + case SRB_ELS_CMD_RPT: + case SRB_ELS_CMD_HST: + type = "els"; + break; + case SRB_CT_CMD: + type = "ct pass-through"; + break; + default: + qla_printk(KERN_WARNING, ha, + "%s: Unrecognized SRB: (%p) type=%d.\n", func, sp, + sp_bsg->ctx.type); + return; + } + + comp_status = fw_status[0] = le16_to_cpu(pkt->comp_status); + fw_status[1] = le16_to_cpu(((struct els_sts_entry_24xx*)pkt)->error_subcode_1); + fw_status[2] = le16_to_cpu(((struct els_sts_entry_24xx*)pkt)->error_subcode_2); + + /* return FC_CTELS_STATUS_OK and leave the decoding of the ELS/CT + * fc payload to the caller + */ + bsg_job->reply->reply_data.ctels_reply.status = FC_CTELS_STATUS_OK; + bsg_job->reply_len = sizeof(struct fc_bsg_reply) + sizeof(fw_status); + + if (comp_status != CS_COMPLETE) { + if (comp_status == CS_DATA_UNDERRUN) { + bsg_job->reply->result = DID_OK << 16; + bsg_job->reply->reply_payload_rcv_len = + le16_to_cpu(((struct els_sts_entry_24xx*)pkt)->total_byte_count); + + DEBUG2(qla_printk(KERN_WARNING, ha, + "scsi(%ld:0x%x): ELS-CT pass-through-%s error comp_status-status=0x%x " + "error subcode 1=0x%x error subcode 2=0x%x total_byte = 0x%x.\n", + vha->host_no, sp->handle, type, comp_status, fw_status[1], fw_status[2], + le16_to_cpu(((struct els_sts_entry_24xx*)pkt)->total_byte_count))); + fw_sts_ptr = ((uint8_t*)bsg_job->req->sense) + sizeof(struct fc_bsg_reply); + memcpy( fw_sts_ptr, fw_status, sizeof(fw_status)); + } + else { + DEBUG2(qla_printk(KERN_WARNING, ha, + "scsi(%ld:0x%x): ELS-CT pass-through-%s error comp_status-status=0x%x " + "error subcode 1=0x%x error subcode 2=0x%x.\n", + vha->host_no, sp->handle, type, comp_status, + le16_to_cpu(((struct els_sts_entry_24xx*)pkt)->error_subcode_1), + le16_to_cpu(((struct els_sts_entry_24xx*)pkt)->error_subcode_2))); + bsg_job->reply->result = DID_ERROR << 16; + bsg_job->reply->reply_payload_rcv_len = 0; + fw_sts_ptr = ((uint8_t*)bsg_job->req->sense) + sizeof(struct fc_bsg_reply); + memcpy( fw_sts_ptr, fw_status, sizeof(fw_status)); + } + DEBUG2(qla2x00_dump_buffer((uint8_t *)pkt, sizeof(*pkt))); + } + else { + bsg_job->reply->result = DID_OK << 16;; + bsg_job->reply->reply_payload_rcv_len = bsg_job->reply_payload.payload_len; + bsg_job->reply_len = 0; + } + + dma_unmap_sg(&ha->pdev->dev, + bsg_job->request_payload.sg_list, + bsg_job->request_payload.sg_cnt, DMA_TO_DEVICE); + dma_unmap_sg(&ha->pdev->dev, + bsg_job->reply_payload.sg_list, + bsg_job->reply_payload.sg_cnt, DMA_FROM_DEVICE); + if ((sp_bsg->ctx.type == SRB_ELS_CMD_HST) || + (sp_bsg->ctx.type == SRB_CT_CMD)) + kfree(sp->fcport); + kfree(sp->ctx); + mempool_free(sp, ha->srb_mempool); + bsg_job->job_done(bsg_job); +} + +static void qla24xx_logio_entry(scsi_qla_host_t *vha, struct req_que *req, struct logio_entry_24xx *logio) { @@ -1749,6 +1846,13 @@ void qla24xx_process_response_queue(struct scsi_qla_host *vha, qla24xx_logio_entry(vha, rsp->req, (struct logio_entry_24xx *)pkt); break; + case CT_IOCB_TYPE: + qla24xx_els_ct_entry(vha, rsp->req, pkt, CT_IOCB_TYPE); + clear_bit(MBX_INTERRUPT, &vha->hw->mbx_cmd_flags); + break; + case ELS_IOCB_TYPE: + qla24xx_els_ct_entry(vha, rsp->req, pkt, ELS_IOCB_TYPE); + break; default: /* Type Not Supported. */ DEBUG4(printk(KERN_WARNING @@ -2046,7 +2150,6 @@ qla24xx_msix_default(int irq, void *dev_id) set_bit(MBX_INTERRUPT, &ha->mbx_cmd_flags); complete(&ha->mbx_intr_comp); } - return IRQ_HANDLED; } diff --git a/drivers/scsi/qla2xxx/qla_mbx.c b/drivers/scsi/qla2xxx/qla_mbx.c index 056e4d4..6e53bdb 100644 --- a/drivers/scsi/qla2xxx/qla_mbx.c +++ b/drivers/scsi/qla2xxx/qla_mbx.c @@ -3636,6 +3636,157 @@ qla2x00_read_ram_word(scsi_qla_host_t *vha, uint32_t risc_addr, uint32_t *data) } int +qla2x00_loopback_test(scsi_qla_host_t *vha, struct msg_echo_lb *mreq, uint16_t *mresp) +{ + int rval; + mbx_cmd_t mc; + mbx_cmd_t *mcp = &mc; + uint32_t iter_cnt = 0x1; + + DEBUG11(printk("scsi(%ld): entered.\n", vha->host_no)); + + memset(mcp->mb, 0 , sizeof(mcp->mb)); + mcp->mb[0] = MBC_DIAGNOSTIC_LOOP_BACK; + mcp->mb[1] = mreq->options | BIT_6; // BIT_6 specifies 64 bit addressing + + /* transfer count */ + mcp->mb[10] = LSW(mreq->transfer_size); + mcp->mb[11] = MSW(mreq->transfer_size); + + /* send data address */ + mcp->mb[14] = LSW(mreq->send_dma); + mcp->mb[15] = MSW(mreq->send_dma); + mcp->mb[20] = LSW(MSD(mreq->send_dma)); + mcp->mb[21] = MSW(MSD(mreq->send_dma)); + + /* recieve data address */ + mcp->mb[16] = LSW(mreq->rcv_dma); + mcp->mb[17] = MSW(mreq->rcv_dma); + mcp->mb[6] = LSW(MSD(mreq->rcv_dma)); + mcp->mb[7] = MSW(MSD(mreq->rcv_dma)); + + /* Iteration count */ + mcp->mb[18] = LSW(iter_cnt); + mcp->mb[19] = MSW(iter_cnt); + + mcp->out_mb = MBX_21|MBX_20|MBX_19|MBX_18|MBX_17|MBX_16|MBX_15| + MBX_14|MBX_13|MBX_12|MBX_11|MBX_10|MBX_7|MBX_6|MBX_1|MBX_0; + if (IS_QLA81XX(vha->hw)) + mcp->out_mb |= MBX_2; + mcp->in_mb = MBX_19|MBX_18|MBX_3|MBX_2|MBX_1|MBX_0; + + mcp->buf_size = mreq->transfer_size; + mcp->tov = MBX_TOV_SECONDS; + mcp->flags = MBX_DMA_OUT|MBX_DMA_IN|IOCTL_CMD; + + rval = qla2x00_mailbox_command(vha, mcp); + + if (rval != QLA_SUCCESS) { + DEBUG2(printk(KERN_WARNING + "(%ld): failed=%x mb[0]=0x%x " + "mb[1]=0x%x mb[2]=0x%x mb[3]=0x%x mb[18]=0x%x mb[19]=0x%x. \n", vha->host_no, rval, + mcp->mb[0], mcp->mb[1], mcp->mb[2], mcp->mb[3], mcp->mb[18], mcp->mb[19])); + } else { + DEBUG2(printk(KERN_WARNING + "scsi(%ld): done.\n", vha->host_no)); + } + + /* Copy mailbox information */ + memcpy( mresp, mcp->mb, 64); + mresp[3] = mcp->mb[18]; + mresp[4] = mcp->mb[19]; + return rval; +} + +int +qla2x00_echo_test(scsi_qla_host_t *vha, struct msg_echo_lb *mreq, uint16_t *mresp) +{ + int rval; + mbx_cmd_t mc; + mbx_cmd_t *mcp = &mc; + struct qla_hw_data *ha = vha->hw; + + DEBUG11(printk("scsi(%ld): entered.\n", vha->host_no)); + + memset(mcp->mb, 0 , sizeof(mcp->mb)); + mcp->mb[0] = MBC_DIAGNOSTIC_ECHO; + mcp->mb[1] = mreq->options | BIT_6; /* BIT_6 specifies 64bit address */ + if (IS_QLA81XX(ha)) + mcp->mb[1] |= BIT_15; + mcp->mb[2] = IS_QLA81XX(ha) ? vha->fcoe_fcf_idx : 0; + mcp->mb[16] = LSW(mreq->rcv_dma); + mcp->mb[17] = MSW(mreq->rcv_dma); + mcp->mb[6] = LSW(MSD(mreq->rcv_dma)); + mcp->mb[7] = MSW(MSD(mreq->rcv_dma)); + + mcp->mb[10] = LSW(mreq->transfer_size); + + mcp->mb[14] = LSW(mreq->send_dma); + mcp->mb[15] = MSW(mreq->send_dma); + mcp->mb[20] = LSW(MSD(mreq->send_dma)); + mcp->mb[21] = MSW(MSD(mreq->send_dma)); + + mcp->out_mb = MBX_21|MBX_20|MBX_17|MBX_16|MBX_15| + MBX_14|MBX_10|MBX_7|MBX_6|MBX_1|MBX_0; + if (IS_QLA81XX(ha)) + mcp->out_mb |= MBX_2; + + mcp->in_mb = MBX_0; + if (IS_QLA24XX_TYPE(ha) || IS_QLA25XX(ha) || IS_QLA81XX(ha)) + mcp->in_mb |= MBX_1; + if (IS_QLA81XX(ha)) + mcp->in_mb |= MBX_3; + + mcp->tov = MBX_TOV_SECONDS; + mcp->flags = MBX_DMA_OUT|MBX_DMA_IN|IOCTL_CMD; + mcp->buf_size = mreq->transfer_size; + + rval = qla2x00_mailbox_command(vha, mcp); + + if (rval != QLA_SUCCESS) { + DEBUG2(printk(KERN_WARNING + "(%ld): failed=%x mb[0]=0x%x mb[1]=0x%x.\n", + vha->host_no, rval, mcp->mb[0], mcp->mb[1])); + } else { + DEBUG2(printk(KERN_WARNING + "scsi(%ld): done.\n", vha->host_no)); + } + + /* Copy mailbox information */ + memcpy( mresp, mcp->mb, 32); + return rval; +} +int +qla84xx_reset_chip(scsi_qla_host_t *ha, uint16_t enable_diagnostic, + uint16_t *cmd_status) +{ + int rval; + mbx_cmd_t mc; + mbx_cmd_t *mcp = &mc; + + DEBUG16(printk("%s(%ld): enable_diag=%d entered.\n", __func__, + ha->host_no, enable_diagnostic)); + + mcp->mb[0] = MBC_ISP84XX_RESET; + mcp->mb[1] = enable_diagnostic; + mcp->out_mb = MBX_1|MBX_0; + mcp->in_mb = MBX_1|MBX_0; + mcp->tov = MBX_TOV_SECONDS; + mcp->flags = MBX_DMA_OUT|MBX_DMA_IN|IOCTL_CMD; + rval = qla2x00_mailbox_command(ha, mcp); + + /* Return mailbox statuses. */ + *cmd_status = mcp->mb[0]; + if (rval != QLA_SUCCESS) + DEBUG16(printk("%s(%ld): failed=%x.\n", __func__, ha->host_no, + rval)); + else + DEBUG16(printk("%s(%ld): done.\n", __func__, ha->host_no)); + + return rval; +} + +int qla2x00_write_ram_word(scsi_qla_host_t *vha, uint32_t risc_addr, uint32_t data) { int rval; diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c index fe34b7f..15ddfc4 100644 --- a/drivers/scsi/qla2xxx/qla_os.c +++ b/drivers/scsi/qla2xxx/qla_os.c @@ -1969,6 +1969,7 @@ qla2x00_probe_one(struct pci_dev *pdev, const struct pci_device_id *id) host->max_channel = MAX_BUSES - 1; host->max_lun = MAX_LUNS; host->transportt = qla2xxx_transport_template; + sht->vendor_id = (SCSI_NL_VID_TYPE_PCI | PCI_VENDOR_ID_QLOGIC); /* Set up the irqs */ ret = qla2x00_request_irqs(ha, rsp); -- cgit v0.10.2 From f8ac60855ebfa239319a4a9945799995feda3b06 Mon Sep 17 00:00:00 2001 From: Giridhar Malavali Date: Tue, 12 Jan 2010 13:02:48 -0800 Subject: [SCSI] qla2xxx: Remove firmware hint for 81xx parts. Firmware is loaded from flash for these ISP types. Signed-off-by: Giridhar Malavali Signed-off-by: James Bottomley diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c index 15ddfc4..ddc1d3d 100644 --- a/drivers/scsi/qla2xxx/qla_os.c +++ b/drivers/scsi/qla2xxx/qla_os.c @@ -3545,4 +3545,3 @@ MODULE_FIRMWARE(FW_FILE_ISP2300); MODULE_FIRMWARE(FW_FILE_ISP2322); MODULE_FIRMWARE(FW_FILE_ISP24XX); MODULE_FIRMWARE(FW_FILE_ISP25XX); -MODULE_FIRMWARE(FW_FILE_ISP81XX); -- cgit v0.10.2 From e8a392444d338d3ce77d157bcbe4975828536efa Mon Sep 17 00:00:00 2001 From: Giridhar Malavali Date: Tue, 12 Jan 2010 13:02:49 -0800 Subject: [SCSI] qla2xxx: Update version number to 8.03.02-k0. Signed-off-by: Giridhar Malavali Signed-off-by: James Bottomley diff --git a/drivers/scsi/qla2xxx/qla_version.h b/drivers/scsi/qla2xxx/qla_version.h index ed36279..082394b 100644 --- a/drivers/scsi/qla2xxx/qla_version.h +++ b/drivers/scsi/qla2xxx/qla_version.h @@ -7,9 +7,9 @@ /* * Driver version */ -#define QLA2XXX_VERSION "8.03.01-k10" +#define QLA2XXX_VERSION "8.03.02-k0" #define QLA_DRIVER_MAJOR_VER 8 #define QLA_DRIVER_MINOR_VER 3 -#define QLA_DRIVER_PATCH_VER 1 +#define QLA_DRIVER_PATCH_VER 2 #define QLA_DRIVER_BETA_VER 0 -- cgit v0.10.2 From 8605c46c171f02859fb3a32655fa226ce446c6ec Mon Sep 17 00:00:00 2001 From: George Kadianakis Date: Sun, 17 Jan 2010 21:19:31 +0200 Subject: [SCSI] lpfc: restore MSI-X/MSI support A Gentoo bug report [1] showed that as of 2.6.31 lpfc only uses INTx interrupts. This patch restores lpfc's ability to support MSI-X/MSI interrupts that the "Addition of SLI4 Interface - Base Support" patch [2] broke. It reestablishes MSI-X as the default interrupt method and in case MSI-X is not supported lpfc_sli{4,}_enable_intr fallbacks to MSI and then to INTx. [1]: http://bugs.gentoo.org/show_bug.cgi?id=296319 [2]: commit da0436e915a5c17ee79e72c1bf978a4ebb1cbf4d [James Smart: Background: Nothing Broke. This was intended. We had originally enabled MSI-X by default, but in qualification within the last 12 months, we encountered a major catch-22: There were at least 4 platforms, from 2 major OEMs, that : - Say they support MSI-X - platform routines work and act as if they do. - We enable it, generate a test interrupt to check they really do deliver it, and it works. - But shortly after attachment, the system hangs or loses interrupts, resulting in a bad system behavior. Given the distro's picking up the 2.6.32 kernel, we had to stick with a default of MSI-X off, with user-enabled MSI-X as these platforms couldn't get fixed. However, we're also now encountering platforms that require MSI-X and never INTx, so we must change. It's desired also for also for performance reasons. So - now (2.6.33) is the right time to re-enable MSI-X by default. ] [jejb: fix up comment on default values] Signed-off-by: George Kadianakis Acked-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/lpfc/lpfc_attr.c b/drivers/scsi/lpfc/lpfc_attr.c index 91542f7..a8908e7 100644 --- a/drivers/scsi/lpfc/lpfc_attr.c +++ b/drivers/scsi/lpfc/lpfc_attr.c @@ -3114,12 +3114,12 @@ LPFC_ATTR_RW(poll_tmo, 10, 1, 255, /* # lpfc_use_msi: Use MSI (Message Signaled Interrupts) in systems that # support this feature -# 0 = MSI disabled (default) +# 0 = MSI disabled # 1 = MSI enabled -# 2 = MSI-X enabled -# Value range is [0,2]. Default value is 0. +# 2 = MSI-X enabled (default) +# Value range is [0,2]. Default value is 2. */ -LPFC_ATTR_R(use_msi, 0, 0, 2, "Use Message Signaled Interrupts (1) or " +LPFC_ATTR_R(use_msi, 2, 0, 2, "Use Message Signaled Interrupts (1) or " "MSI-X (2), if possible"); /* -- cgit v0.10.2 From d57d64080ddc0ff13fcffc898b6251074a482ba1 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 13:34:38 +0900 Subject: sh: Prevent 64-bit pgprot clobbering across ioremap implementations. Presently 'flags' gets passed around a lot between the various ioremap helpers and implementations, which is only 32-bits. In the X2TLB case we use 64-bit pgprots which presently results in the upper 32bits being chopped off (which handily include our read/write/exec permissions). As such, we convert everything internally to using pgprot_t directly and simply convert over with pgprot_val() where needed. With this in place, transparent fixmap utilization for early ioremap works as expected. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/board-sh7785lcr.c b/arch/sh/boards/board-sh7785lcr.c index 511de38..fe7e686 100644 --- a/arch/sh/boards/board-sh7785lcr.c +++ b/arch/sh/boards/board-sh7785lcr.c @@ -21,6 +21,7 @@ #include #include #include +#include #include #include #include @@ -332,15 +333,14 @@ static void __init sh7785lcr_setup(char **cmdline_p) pm_power_off = sh7785lcr_power_off; /* sm501 DRAM configuration */ - sm501_reg = ioremap_fixed(SM107_REG_ADDR, SM501_DRAM_CONTROL, - PAGE_KERNEL); + sm501_reg = ioremap_nocache(SM107_REG_ADDR, SM501_DRAM_CONTROL); if (!sm501_reg) { printk(KERN_ERR "%s: ioremap error.\n", __func__); return; } writel(0x000307c2, sm501_reg + SM501_DRAM_CONTROL); - iounmap_fixed(sm501_reg); + iounmap(sm501_reg); } /* Return the board specific boot mode pin configuration */ diff --git a/arch/sh/boards/mach-landisk/setup.c b/arch/sh/boards/mach-landisk/setup.c index db22ea2..5981635 100644 --- a/arch/sh/boards/mach-landisk/setup.c +++ b/arch/sh/boards/mach-landisk/setup.c @@ -63,7 +63,7 @@ static int __init landisk_devices_setup(void) /* open I/O area window */ paddrbase = virt_to_phys((void *)PA_AREA5_IO); prot = PAGE_KERNEL_PCC(1, _PAGE_PCC_IO16); - cf_ide_base = p3_ioremap(paddrbase, PAGE_SIZE, prot.pgprot); + cf_ide_base = p3_ioremap(paddrbase, PAGE_SIZE, prot); if (!cf_ide_base) { printk("allocate_cf_area : can't open CF I/O window!\n"); return -ENOMEM; diff --git a/arch/sh/boards/mach-lboxre2/setup.c b/arch/sh/boards/mach-lboxre2/setup.c index 2b0b581..408dd5d 100644 --- a/arch/sh/boards/mach-lboxre2/setup.c +++ b/arch/sh/boards/mach-lboxre2/setup.c @@ -57,7 +57,7 @@ static int __init lboxre2_devices_setup(void) paddrbase = virt_to_phys((void*)PA_AREA5_IO); psize = PAGE_SIZE; prot = PAGE_KERNEL_PCC( 1 , _PAGE_PCC_IO16); - cf0_io_base = (u32)p3_ioremap(paddrbase, psize, prot.pgprot); + cf0_io_base = (u32)p3_ioremap(paddrbase, psize, prot); if (!cf0_io_base) { printk(KERN_ERR "%s : can't open CF I/O window!\n" , __func__ ); return -ENOMEM; diff --git a/arch/sh/boards/mach-sh03/setup.c b/arch/sh/boards/mach-sh03/setup.c index 74cfb4b..f14ba0f 100644 --- a/arch/sh/boards/mach-sh03/setup.c +++ b/arch/sh/boards/mach-sh03/setup.c @@ -82,7 +82,7 @@ static int __init sh03_devices_setup(void) /* open I/O area window */ paddrbase = virt_to_phys((void *)PA_AREA5_IO); prot = PAGE_KERNEL_PCC(1, _PAGE_PCC_IO16); - cf_ide_base = p3_ioremap(paddrbase, PAGE_SIZE, prot.pgprot); + cf_ide_base = p3_ioremap(paddrbase, PAGE_SIZE, prot); if (!cf_ide_base) { printk("allocate_cf_area : can't open CF I/O window!\n"); return -ENOMEM; diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index 6a0dd8c..13696df 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -235,7 +235,7 @@ unsigned long long poke_real_address_q(unsigned long long addr, */ #ifdef CONFIG_MMU void __iomem *__ioremap_caller(unsigned long offset, unsigned long size, - unsigned long flags, void *caller); + pgprot_t prot, void *caller); void __iounmap(void __iomem *addr); #ifdef CONFIG_IOREMAP_FIXED @@ -254,13 +254,13 @@ static inline int iounmap_fixed(void __iomem *addr) { return -EINVAL; } #endif static inline void __iomem * -__ioremap(unsigned long offset, unsigned long size, unsigned long flags) +__ioremap(unsigned long offset, unsigned long size, pgprot_t prot) { - return __ioremap_caller(offset, size, flags, __builtin_return_address(0)); + return __ioremap_caller(offset, size, prot, __builtin_return_address(0)); } static inline void __iomem * -__ioremap_29bit(unsigned long offset, unsigned long size, unsigned long flags) +__ioremap_29bit(unsigned long offset, unsigned long size, pgprot_t prot) { #ifdef CONFIG_29BIT unsigned long last_addr = offset + size - 1; @@ -272,7 +272,7 @@ __ioremap_29bit(unsigned long offset, unsigned long size, unsigned long flags) * mapping must be done by the PMB or by using page tables. */ if (likely(PXSEG(offset) < P3SEG && PXSEG(last_addr) < P3SEG)) { - if (unlikely(flags & _PAGE_CACHABLE)) + if (unlikely(pgprot_val(prot) & _PAGE_CACHABLE)) return (void __iomem *)P1SEGADDR(offset); return (void __iomem *)P2SEGADDR(offset); @@ -287,7 +287,7 @@ __ioremap_29bit(unsigned long offset, unsigned long size, unsigned long flags) } static inline void __iomem * -__ioremap_mode(unsigned long offset, unsigned long size, unsigned long flags) +__ioremap_mode(unsigned long offset, unsigned long size, pgprot_t prot) { void __iomem *ret; @@ -295,30 +295,39 @@ __ioremap_mode(unsigned long offset, unsigned long size, unsigned long flags) if (ret) return ret; - ret = __ioremap_29bit(offset, size, flags); + ret = __ioremap_29bit(offset, size, prot); if (ret) return ret; - return __ioremap(offset, size, flags); + return __ioremap(offset, size, prot); } #else -#define __ioremap(offset, size, flags) ((void __iomem *)(offset)) -#define __ioremap_mode(offset, size, flags) ((void __iomem *)(offset)) +#define __ioremap(offset, size, prot) ((void __iomem *)(offset)) +#define __ioremap_mode(offset, size, prot) ((void __iomem *)(offset)) #define __iounmap(addr) do { } while (0) #endif /* CONFIG_MMU */ -#define ioremap(offset, size) \ - __ioremap_mode((offset), (size), 0) -#define ioremap_nocache(offset, size) \ - __ioremap_mode((offset), (size), 0) -#define ioremap_cache(offset, size) \ - __ioremap_mode((offset), (size), _PAGE_CACHABLE) -#define p3_ioremap(offset, size, flags) \ - __ioremap((offset), (size), (flags)) -#define ioremap_prot(offset, size, flags) \ - __ioremap_mode((offset), (size), (flags)) -#define iounmap(addr) \ - __iounmap((addr)) +static inline void __iomem * +ioremap(unsigned long offset, unsigned long size) +{ + return __ioremap_mode(offset, size, PAGE_KERNEL_NOCACHE); +} + +static inline void __iomem * +ioremap_cache(unsigned long offset, unsigned long size) +{ + return __ioremap_mode(offset, size, PAGE_KERNEL); +} + +static inline void __iomem * +ioremap_prot(resource_size_t offset, unsigned long size, unsigned long flags) +{ + return __ioremap_mode(offset, size, __pgprot(flags)); +} + +#define ioremap_nocache ioremap +#define p3_ioremap __ioremap +#define iounmap __iounmap #define maybebadio(port) \ printk(KERN_ERR "bad PC-like io %s:%u for port 0x%lx at 0x%08x\n", \ diff --git a/arch/sh/mm/ioremap.c b/arch/sh/mm/ioremap.c index a130b22..85b420d 100644 --- a/arch/sh/mm/ioremap.c +++ b/arch/sh/mm/ioremap.c @@ -35,11 +35,10 @@ */ void __iomem * __init_refok __ioremap_caller(unsigned long phys_addr, unsigned long size, - unsigned long flags, void *caller) + pgprot_t pgprot, void *caller) { struct vm_struct *area; unsigned long offset, last_addr, addr, orig_addr; - pgprot_t pgprot; /* Don't allow wraparound or zero size */ last_addr = phys_addr + size - 1; @@ -69,7 +68,7 @@ __ioremap_caller(unsigned long phys_addr, unsigned long size, * If we can't yet use the regular approach, go the fixmap route. */ if (!mem_init_done) - return ioremap_fixed(phys_addr, size, __pgprot(flags)); + return ioremap_fixed(phys_addr, size, pgprot); /* * Ok, go for it.. @@ -91,8 +90,9 @@ __ioremap_caller(unsigned long phys_addr, unsigned long size, * PMB entries are all pre-faulted. */ if (unlikely(phys_addr >= P1SEG)) { - unsigned long mapped = pmb_remap(addr, phys_addr, size, flags); + unsigned long mapped; + mapped = pmb_remap(addr, phys_addr, size, pgprot_val(pgprot)); if (likely(mapped)) { addr += mapped; phys_addr += mapped; @@ -101,7 +101,6 @@ __ioremap_caller(unsigned long phys_addr, unsigned long size, } #endif - pgprot = __pgprot(pgprot_val(PAGE_KERNEL_NOCACHE) | flags); if (likely(size)) if (ioremap_page_range(addr, addr + size, phys_addr, pgprot)) { vunmap((void *)orig_addr); -- cgit v0.10.2 From acf2c9685fb8295cb62a623d7358a1cfde8b07ea Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 13:49:19 +0900 Subject: sh: Kill off duplicate address alignment in ioremap_fixed(). This is already taken care of in the top-level ioremap, and now that no one should be calling ioremap_fixed() directly we can simply throw the mapping displacement in as an additional argument. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index 13696df..7026981 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -239,12 +239,14 @@ void __iomem *__ioremap_caller(unsigned long offset, unsigned long size, void __iounmap(void __iomem *addr); #ifdef CONFIG_IOREMAP_FIXED -extern void __iomem *ioremap_fixed(resource_size_t, unsigned long, pgprot_t); +extern void __iomem *ioremap_fixed(resource_size_t, unsigned long, + unsigned long, pgprot_t); extern int iounmap_fixed(void __iomem *); extern void ioremap_fixed_init(void); #else static inline void __iomem * -ioremap_fixed(resource_size t phys_addr, unsigned long size, pgprot_t prot) +ioremap_fixed(resource_size t phys_addr, unsigned long offset, + unsigned long size, pgprot_t prot) { BUG(); } diff --git a/arch/sh/mm/ioremap.c b/arch/sh/mm/ioremap.c index 85b420d..bb03308 100644 --- a/arch/sh/mm/ioremap.c +++ b/arch/sh/mm/ioremap.c @@ -68,7 +68,7 @@ __ioremap_caller(unsigned long phys_addr, unsigned long size, * If we can't yet use the regular approach, go the fixmap route. */ if (!mem_init_done) - return ioremap_fixed(phys_addr, size, pgprot); + return ioremap_fixed(phys_addr, offset, size, pgprot); /* * Ok, go for it.. diff --git a/arch/sh/mm/ioremap_fixed.c b/arch/sh/mm/ioremap_fixed.c index 425f6c6..551b513 100644 --- a/arch/sh/mm/ioremap_fixed.c +++ b/arch/sh/mm/ioremap_fixed.c @@ -45,12 +45,11 @@ void __init ioremap_fixed_init(void) } void __init __iomem * -ioremap_fixed(resource_size_t phys_addr, unsigned long size, pgprot_t prot) +ioremap_fixed(resource_size_t phys_addr, unsigned long offset, + unsigned long size, pgprot_t prot) { enum fixed_addresses idx0, idx; - resource_size_t last_addr; struct ioremap_map *map; - unsigned long offset; unsigned int nrpages; int i, slot; @@ -67,18 +66,6 @@ ioremap_fixed(resource_size_t phys_addr, unsigned long size, pgprot_t prot) if (slot < 0) return NULL; - /* Don't allow wraparound or zero size */ - last_addr = phys_addr + size - 1; - if (!size || last_addr < phys_addr) - return NULL; - - /* - * Fixmap mappings have to be page-aligned - */ - offset = phys_addr & ~PAGE_MASK; - phys_addr &= PAGE_MASK; - size = PAGE_ALIGN(last_addr + 1) - phys_addr; - /* * Mappings have to fit in the FIX_IOREMAP area. */ @@ -111,7 +98,6 @@ int iounmap_fixed(void __iomem *addr) unsigned long offset; unsigned int nrpages; int i, slot; - pgprot_t prot; slot = -1; for (i = 0; i < FIX_N_IOREMAPS; i++) { @@ -133,11 +119,9 @@ int iounmap_fixed(void __iomem *addr) offset = virt_addr & ~PAGE_MASK; nrpages = PAGE_ALIGN(offset + map->size - 1) >> PAGE_SHIFT; - pgprot_val(prot) = _PAGE_WIRED; - idx = FIX_IOREMAP_BEGIN + slot + nrpages; while (nrpages > 0) { - __clear_fixmap(idx, prot); + __clear_fixmap(idx, __pgprot(_PAGE_WIRED)); --idx; --nrpages; } -- cgit v0.10.2 From f33609344acfdde1c1acedf6de3efe6b80af93a6 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 13:55:41 +0900 Subject: sh: Convert p3_ioremap() users to ioremap_prot(). This kills off the ancient p3_ioremap(), converting over to the more generic ioremap_prot() instead. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-landisk/setup.c b/arch/sh/boards/mach-landisk/setup.c index 5981635..2d09d4d 100644 --- a/arch/sh/boards/mach-landisk/setup.c +++ b/arch/sh/boards/mach-landisk/setup.c @@ -63,7 +63,7 @@ static int __init landisk_devices_setup(void) /* open I/O area window */ paddrbase = virt_to_phys((void *)PA_AREA5_IO); prot = PAGE_KERNEL_PCC(1, _PAGE_PCC_IO16); - cf_ide_base = p3_ioremap(paddrbase, PAGE_SIZE, prot); + cf_ide_base = ioremap_prot(paddrbase, PAGE_SIZE, pgprot_val(prot)); if (!cf_ide_base) { printk("allocate_cf_area : can't open CF I/O window!\n"); return -ENOMEM; diff --git a/arch/sh/boards/mach-lboxre2/setup.c b/arch/sh/boards/mach-lboxre2/setup.c index 408dd5d..79b4e0d7 100644 --- a/arch/sh/boards/mach-lboxre2/setup.c +++ b/arch/sh/boards/mach-lboxre2/setup.c @@ -56,8 +56,8 @@ static int __init lboxre2_devices_setup(void) /* open I/O area window */ paddrbase = virt_to_phys((void*)PA_AREA5_IO); psize = PAGE_SIZE; - prot = PAGE_KERNEL_PCC( 1 , _PAGE_PCC_IO16); - cf0_io_base = (u32)p3_ioremap(paddrbase, psize, prot); + prot = PAGE_KERNEL_PCC(1, _PAGE_PCC_IO16); + cf0_io_base = (u32)ioremap_prot(paddrbase, psize, pgprot_val(prot)); if (!cf0_io_base) { printk(KERN_ERR "%s : can't open CF I/O window!\n" , __func__ ); return -ENOMEM; diff --git a/arch/sh/boards/mach-sh03/setup.c b/arch/sh/boards/mach-sh03/setup.c index f14ba0f..af4a0c0 100644 --- a/arch/sh/boards/mach-sh03/setup.c +++ b/arch/sh/boards/mach-sh03/setup.c @@ -82,7 +82,7 @@ static int __init sh03_devices_setup(void) /* open I/O area window */ paddrbase = virt_to_phys((void *)PA_AREA5_IO); prot = PAGE_KERNEL_PCC(1, _PAGE_PCC_IO16); - cf_ide_base = p3_ioremap(paddrbase, PAGE_SIZE, prot); + cf_ide_base = ioremap_prot(paddrbase, PAGE_SIZE, pgprot_val(prot)); if (!cf_ide_base) { printk("allocate_cf_area : can't open CF I/O window!\n"); return -ENOMEM; diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index 7026981..c32a556 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -328,7 +328,6 @@ ioremap_prot(resource_size_t offset, unsigned long size, unsigned long flags) } #define ioremap_nocache ioremap -#define p3_ioremap __ioremap #define iounmap __iounmap #define maybebadio(port) \ -- cgit v0.10.2 From 6d63e73d520b690e4378cef3003eb5f01f7d128c Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 14:00:14 +0900 Subject: sh: Limit ioremap_prot() to 32bit pgprot parts. Presently ioremap_prot() uses an unsigned long to pass the pgprot value around. This results in the upper half of the pgprot being chomped when using 64-bit pgprots on a 32-bit ABI (X2TLB and SH-5). As the only users of ioremap_prot() are presently legacy parts, this doesn't cause too much of an issue. In the future when the interface is converted to use pgprot_t directly this can be re-enabled for the other parts, too. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 5ac0115..8d0eabb 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -13,7 +13,6 @@ config SUPERH select HAVE_LMB select HAVE_OPROFILE select HAVE_GENERIC_DMA_COHERENT - select HAVE_IOREMAP_PROT if MMU select HAVE_ARCH_TRACEHOOK select HAVE_DMA_API_DEBUG select HAVE_DMA_ATTRS @@ -36,6 +35,7 @@ config SUPERH32 def_bool ARCH = "sh" select HAVE_KPROBES select HAVE_KRETPROBES + select HAVE_IOREMAP_PROT if MMU && !X2TLB select HAVE_FUNCTION_TRACER select HAVE_FTRACE_MCOUNT_RECORD select HAVE_DYNAMIC_FTRACE diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index c32a556..4bad72a 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -321,11 +321,13 @@ ioremap_cache(unsigned long offset, unsigned long size) return __ioremap_mode(offset, size, PAGE_KERNEL); } +#ifdef CONFIG_HAVE_IOREMAP_PROT static inline void __iomem * ioremap_prot(resource_size_t offset, unsigned long size, unsigned long flags) { return __ioremap_mode(offset, size, __pgprot(flags)); } +#endif #define ioremap_nocache ioremap #define iounmap __iounmap -- cgit v0.10.2 From 046581f9623b53f551a93864bb74e15ad2514f0c Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 14:23:39 +0900 Subject: sh: Provide a dummy _PAGE_WIRED flag for non-X2TLB parts. This provides a dummy value for legacy parts which permits the entry wiring to be open-coded. The compiler takes care of optimizing the entry wiring away in these cases. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/pgtable_32.h b/arch/sh/include/asm/pgtable_32.h index c573d45..e172d69 100644 --- a/arch/sh/include/asm/pgtable_32.h +++ b/arch/sh/include/asm/pgtable_32.h @@ -143,12 +143,14 @@ static inline unsigned long copy_ptea_attributes(unsigned long x) # elif defined(CONFIG_HUGETLB_PAGE_SIZE_64MB) # define _PAGE_SZHUGE (_PAGE_EXT_ESZ2 | _PAGE_EXT_ESZ3) # endif +# define _PAGE_WIRED (_PAGE_EXT(_PAGE_EXT_WIRED)) #else # if defined(CONFIG_HUGETLB_PAGE_SIZE_64K) # define _PAGE_SZHUGE (_PAGE_SZ1) # elif defined(CONFIG_HUGETLB_PAGE_SIZE_1MB) # define _PAGE_SZHUGE (_PAGE_SZ0 | _PAGE_SZ1) # endif +# define _PAGE_WIRED (0) #endif /* @@ -166,8 +168,6 @@ static inline unsigned long copy_ptea_attributes(unsigned long x) (PTE_MASK | _PAGE_ACCESSED | _PAGE_CACHABLE | \ _PAGE_DIRTY | _PAGE_SPECIAL) -#define _PAGE_WIRED (_PAGE_EXT(_PAGE_EXT_WIRED)) - #ifndef __ASSEMBLY__ #if defined(CONFIG_X2TLB) /* SH-X2 TLB */ -- cgit v0.10.2 From bb29c677b366fdf4f6522cd82228a32567aa98c7 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 15:20:35 +0900 Subject: sh: Split out MMUCR.URB based entry wiring in to shared helper. Presently this is duplicated between tlb-sh4 and tlb-pteaex. Split the helpers out in to a generic tlb-urb that can be used by any parts equipped with MMUCR.URB. At the same time, move the SH-5 code out-of-line, as we require single global state for DTLB entry wiring. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/tlb.h b/arch/sh/include/asm/tlb.h index dfc8fcd..75abb38 100644 --- a/arch/sh/include/asm/tlb.h +++ b/arch/sh/include/asm/tlb.h @@ -98,49 +98,9 @@ tlb_end_vma(struct mmu_gather *tlb, struct vm_area_struct *vma) #define tlb_migrate_finish(mm) do { } while (0) -#ifdef CONFIG_CPU_SH4 +#if defined(CONFIG_CPU_SH4) || defined(CONFIG_SUPERH64) extern void tlb_wire_entry(struct vm_area_struct *, unsigned long, pte_t); extern void tlb_unwire_entry(void); -#elif defined(CONFIG_SUPERH64) -static int dtlb_entry; -static unsigned long long dtlb_entries[64]; - -static inline void tlb_wire_entry(struct vm_area_struct *vma, - unsigned long addr, pte_t pte) -{ - unsigned long long entry; - unsigned long paddr, flags; - - BUG_ON(dtlb_entry == 64); - - local_irq_save(flags); - - entry = sh64_get_wired_dtlb_entry(); - dtlb_entries[dtlb_entry++] = entry; - - paddr = pte_val(pte) & _PAGE_FLAGS_HARDWARE_MASK; - paddr &= ~PAGE_MASK; - - sh64_setup_tlb_slot(entry, addr, get_asid(), paddr); - - local_irq_restore(flags); -} - -static inline void tlb_unwire_entry(void) -{ - unsigned long long entry; - unsigned long flags; - - BUG_ON(!dtlb_entry); - - local_irq_save(flags); - entry = dtlb_entries[dtlb_entry--]; - - sh64_teardown_tlb_slot(entry); - sh64_put_wired_dtlb_entry(entry); - - local_irq_restore(flags); -} #else static inline void tlb_wire_entry(struct vm_area_struct *vma , unsigned long addr, pte_t pte) @@ -152,7 +112,7 @@ static inline void tlb_unwire_entry(void) { BUG(); } -#endif /* CONFIG_CPU_SH4 */ +#endif #else /* CONFIG_MMU */ diff --git a/arch/sh/mm/Makefile b/arch/sh/mm/Makefile index 0027cde..de714cb 100644 --- a/arch/sh/mm/Makefile +++ b/arch/sh/mm/Makefile @@ -26,9 +26,9 @@ endif ifdef CONFIG_MMU tlb-$(CONFIG_CPU_SH3) := tlb-sh3.o -tlb-$(CONFIG_CPU_SH4) := tlb-sh4.o +tlb-$(CONFIG_CPU_SH4) := tlb-sh4.o tlb-urb.o tlb-$(CONFIG_CPU_SH5) := tlb-sh5.o -tlb-$(CONFIG_CPU_HAS_PTEAEX) := tlb-pteaex.o +tlb-$(CONFIG_CPU_HAS_PTEAEX) := tlb-pteaex.o tlb-urb.o obj-y += $(tlb-y) endif diff --git a/arch/sh/mm/tlb-pteaex.c b/arch/sh/mm/tlb-pteaex.c index 88c8bb0..409b7c2 100644 --- a/arch/sh/mm/tlb-pteaex.c +++ b/arch/sh/mm/tlb-pteaex.c @@ -76,69 +76,3 @@ void __uses_jump_to_uncached local_flush_tlb_one(unsigned long asid, __raw_writel(asid, MMU_UTLB_ADDRESS_ARRAY2 | MMU_PAGE_ASSOC_BIT); back_to_cached(); } - -/* - * Load the entry for 'addr' into the TLB and wire the entry. - */ -void tlb_wire_entry(struct vm_area_struct *vma, unsigned long addr, pte_t pte) -{ - unsigned long status, flags; - int urb; - - local_irq_save(flags); - - /* Load the entry into the TLB */ - __update_tlb(vma, addr, pte); - - /* ... and wire it up. */ - status = ctrl_inl(MMUCR); - urb = (status & MMUCR_URB) >> MMUCR_URB_SHIFT; - status &= ~MMUCR_URB; - - /* - * Make sure we're not trying to wire the last TLB entry slot. - */ - BUG_ON(!--urb); - - urb = urb % MMUCR_URB_NENTRIES; - - status |= (urb << MMUCR_URB_SHIFT); - ctrl_outl(status, MMUCR); - ctrl_barrier(); - - local_irq_restore(flags); -} - -/* - * Unwire the last wired TLB entry. - * - * It should also be noted that it is not possible to wire and unwire - * TLB entries in an arbitrary order. If you wire TLB entry N, followed - * by entry N+1, you must unwire entry N+1 first, then entry N. In this - * respect, it works like a stack or LIFO queue. - */ -void tlb_unwire_entry(void) -{ - unsigned long status, flags; - int urb; - - local_irq_save(flags); - - status = ctrl_inl(MMUCR); - urb = (status & MMUCR_URB) >> MMUCR_URB_SHIFT; - status &= ~MMUCR_URB; - - /* - * Make sure we're not trying to unwire a TLB entry when none - * have been wired. - */ - BUG_ON(urb++ == MMUCR_URB_NENTRIES); - - urb = urb % MMUCR_URB_NENTRIES; - - status |= (urb << MMUCR_URB_SHIFT); - ctrl_outl(status, MMUCR); - ctrl_barrier(); - - local_irq_restore(flags); -} diff --git a/arch/sh/mm/tlb-sh4.c b/arch/sh/mm/tlb-sh4.c index 4c62347..8cf550e 100644 --- a/arch/sh/mm/tlb-sh4.c +++ b/arch/sh/mm/tlb-sh4.c @@ -81,69 +81,3 @@ void __uses_jump_to_uncached local_flush_tlb_one(unsigned long asid, ctrl_outl(data, addr); back_to_cached(); } - -/* - * Load the entry for 'addr' into the TLB and wire the entry. - */ -void tlb_wire_entry(struct vm_area_struct *vma, unsigned long addr, pte_t pte) -{ - unsigned long status, flags; - int urb; - - local_irq_save(flags); - - /* Load the entry into the TLB */ - __update_tlb(vma, addr, pte); - - /* ... and wire it up. */ - status = ctrl_inl(MMUCR); - urb = (status & MMUCR_URB) >> MMUCR_URB_SHIFT; - status &= ~MMUCR_URB; - - /* - * Make sure we're not trying to wire the last TLB entry slot. - */ - BUG_ON(!--urb); - - urb = urb % MMUCR_URB_NENTRIES; - - status |= (urb << MMUCR_URB_SHIFT); - ctrl_outl(status, MMUCR); - ctrl_barrier(); - - local_irq_restore(flags); -} - -/* - * Unwire the last wired TLB entry. - * - * It should also be noted that it is not possible to wire and unwire - * TLB entries in an arbitrary order. If you wire TLB entry N, followed - * by entry N+1, you must unwire entry N+1 first, then entry N. In this - * respect, it works like a stack or LIFO queue. - */ -void tlb_unwire_entry(void) -{ - unsigned long status, flags; - int urb; - - local_irq_save(flags); - - status = ctrl_inl(MMUCR); - urb = (status & MMUCR_URB) >> MMUCR_URB_SHIFT; - status &= ~MMUCR_URB; - - /* - * Make sure we're not trying to unwire a TLB entry when none - * have been wired. - */ - BUG_ON(urb++ == MMUCR_URB_NENTRIES); - - urb = urb % MMUCR_URB_NENTRIES; - - status |= (urb << MMUCR_URB_SHIFT); - ctrl_outl(status, MMUCR); - ctrl_barrier(); - - local_irq_restore(flags); -} diff --git a/arch/sh/mm/tlb-sh5.c b/arch/sh/mm/tlb-sh5.c index fdb64e4..f27dbe1 100644 --- a/arch/sh/mm/tlb-sh5.c +++ b/arch/sh/mm/tlb-sh5.c @@ -143,3 +143,42 @@ void sh64_setup_tlb_slot(unsigned long long config_addr, unsigned long eaddr, */ void sh64_teardown_tlb_slot(unsigned long long config_addr) __attribute__ ((alias("__flush_tlb_slot"))); + +static int dtlb_entry; +static unsigned long long dtlb_entries[64]; + +void tlb_wire_entry(struct vm_area_struct *vma, unsigned long addr, pte_t pte) +{ + unsigned long long entry; + unsigned long paddr, flags; + + BUG_ON(dtlb_entry == ARRAY_SIZE(dtlb_entries)); + + local_irq_save(flags); + + entry = sh64_get_wired_dtlb_entry(); + dtlb_entries[dtlb_entry++] = entry; + + paddr = pte_val(pte) & _PAGE_FLAGS_HARDWARE_MASK; + paddr &= ~PAGE_MASK; + + sh64_setup_tlb_slot(entry, addr, get_asid(), paddr); + + local_irq_restore(flags); +} + +void tlb_unwire_entry(void) +{ + unsigned long long entry; + unsigned long flags; + + BUG_ON(!dtlb_entry); + + local_irq_save(flags); + entry = dtlb_entries[dtlb_entry--]; + + sh64_teardown_tlb_slot(entry); + sh64_put_wired_dtlb_entry(entry); + + local_irq_restore(flags); +} diff --git a/arch/sh/mm/tlb-urb.c b/arch/sh/mm/tlb-urb.c new file mode 100644 index 0000000..bb5b909 --- /dev/null +++ b/arch/sh/mm/tlb-urb.c @@ -0,0 +1,81 @@ +/* + * arch/sh/mm/tlb-urb.c + * + * TLB entry wiring helpers for URB-equipped parts. + * + * Copyright (C) 2010 Matt Fleming + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ +#include +#include +#include +#include + +/* + * Load the entry for 'addr' into the TLB and wire the entry. + */ +void tlb_wire_entry(struct vm_area_struct *vma, unsigned long addr, pte_t pte) +{ + unsigned long status, flags; + int urb; + + local_irq_save(flags); + + /* Load the entry into the TLB */ + __update_tlb(vma, addr, pte); + + /* ... and wire it up. */ + status = __raw_readl(MMUCR); + urb = (status & MMUCR_URB) >> MMUCR_URB_SHIFT; + status &= ~MMUCR_URB; + + /* + * Make sure we're not trying to wire the last TLB entry slot. + */ + BUG_ON(!--urb); + + urb = urb % MMUCR_URB_NENTRIES; + + status |= (urb << MMUCR_URB_SHIFT); + __raw_writel(status, MMUCR); + ctrl_barrier(); + + local_irq_restore(flags); +} + +/* + * Unwire the last wired TLB entry. + * + * It should also be noted that it is not possible to wire and unwire + * TLB entries in an arbitrary order. If you wire TLB entry N, followed + * by entry N+1, you must unwire entry N+1 first, then entry N. In this + * respect, it works like a stack or LIFO queue. + */ +void tlb_unwire_entry(void) +{ + unsigned long status, flags; + int urb; + + local_irq_save(flags); + + status = __raw_readl(MMUCR); + urb = (status & MMUCR_URB) >> MMUCR_URB_SHIFT; + status &= ~MMUCR_URB; + + /* + * Make sure we're not trying to unwire a TLB entry when none + * have been wired. + */ + BUG_ON(urb++ == MMUCR_URB_NENTRIES); + + urb = urb % MMUCR_URB_NENTRIES; + + status |= (urb << MMUCR_URB_SHIFT); + __raw_writel(status, MMUCR); + ctrl_barrier(); + + local_irq_restore(flags); +} -- cgit v0.10.2 From cb6d04468d16de5a6161167ec7e76a43be540a80 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 15:22:52 +0900 Subject: sh: Kill off now bogus fixmap/page wiring documentation. The plans for _PAGE_WIRED were detailed in a comment with the fixmap code, but as it's now all taken care of, we no longer have any reason for keeping it around, particularly since it's no longer accurate. Kill it off. Signed-off-by: Paul Mundt diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c index fe578a2..32ebd15 100644 --- a/arch/sh/mm/init.c +++ b/arch/sh/mm/init.c @@ -98,21 +98,6 @@ static void clear_pte_phys(unsigned long addr, pgprot_t prot) local_flush_tlb_one(get_asid(), addr); } -/* - * As a performance optimization, other platforms preserve the fixmap mapping - * across a context switch, we don't presently do this, but this could be done - * in a similar fashion as to the wired TLB interface that sh64 uses (by way - * of the memory mapped UTLB configuration) -- this unfortunately forces us to - * give up a TLB entry for each mapping we want to preserve. While this may be - * viable for a small number of fixmaps, it's not particularly useful for - * everything and needs to be carefully evaluated. (ie, we may want this for - * the vsyscall page). - * - * XXX: Perhaps add a _PAGE_WIRED flag or something similar that we can pass - * in at __set_fixmap() time to determine the appropriate behavior to follow. - * - * -- PFM. - */ void __set_fixmap(enum fixed_addresses idx, unsigned long phys, pgprot_t prot) { unsigned long address = __fix_to_virt(idx); -- cgit v0.10.2 From 3ef2932b8c1fc89408ef1fd4b1e1c2caabc7f07d Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 15:40:03 +0900 Subject: sh64: Fix up the build for the thread_xstate changes. This updates the sh64 processor info with the sh32 changes in order to tie in to the generic task_xstate management code. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/processor.h b/arch/sh/include/asm/processor.h index a522e5d..87a8d1e 100644 --- a/arch/sh/include/asm/processor.h +++ b/arch/sh/include/asm/processor.h @@ -98,9 +98,15 @@ extern struct sh_cpuinfo cpu_data[]; /* Forward decl */ struct seq_operations; +struct task_struct; extern struct pt_regs fake_swapper_regs; +/* arch/sh/kernel/process.c */ +extern unsigned int xstate_size; +extern void free_thread_xstate(struct task_struct *); +extern struct kmem_cache *task_xstate_cachep; + /* arch/sh/mm/init.c */ extern unsigned int mem_init_done; diff --git a/arch/sh/include/asm/processor_32.h b/arch/sh/include/asm/processor_32.h index 5fd8312..488f0a9 100644 --- a/arch/sh/include/asm/processor_32.h +++ b/arch/sh/include/asm/processor_32.h @@ -96,10 +96,6 @@ union thread_xstate { struct sh_fpu_soft_struct softfpu; }; -extern unsigned int xstate_size; -extern void free_thread_xstate(struct task_struct *); -extern struct kmem_cache *task_xstate_cachep; - struct thread_struct { /* Saved registers when thread is descheduled */ unsigned long sp; diff --git a/arch/sh/include/asm/processor_64.h b/arch/sh/include/asm/processor_64.h index 5727d31..7b1560f 100644 --- a/arch/sh/include/asm/processor_64.h +++ b/arch/sh/include/asm/processor_64.h @@ -87,20 +87,21 @@ struct sh_fpu_hard_struct { /* long status; * software status information */ }; -#if 0 /* Dummy fpu emulator */ struct sh_fpu_soft_struct { - unsigned long long fp_regs[32]; + unsigned long fp_regs[64]; unsigned int fpscr; unsigned char lookahead; unsigned long entry_pc; }; -#endif -union sh_fpu_union { - struct sh_fpu_hard_struct hard; - /* 'hard' itself only produces 32 bit alignment, yet we need - to access it using 64 bit load/store as well. */ +union thread_xstate { + struct sh_fpu_hard_struct hardfpu; + struct sh_fpu_soft_struct softfpu; + /* + * The structure definitions only produce 32 bit alignment, yet we need + * to access them using 64 bit load/store as well. + */ unsigned long long alignment_dummy; }; @@ -122,7 +123,7 @@ struct thread_struct { /* Hardware debugging registers may come here */ /* floating point info */ - union sh_fpu_union fpu; + union thread_xstate *xstate; }; #define INIT_MMAP \ @@ -137,7 +138,6 @@ struct thread_struct { .trap_no = 0, \ .error_code = 0, \ .address = 0, \ - .fpu = { { { 0, } }, } \ } /* diff --git a/arch/sh/kernel/cpu/sh5/fpu.c b/arch/sh/kernel/cpu/sh5/fpu.c index 4648cce..92df285 100644 --- a/arch/sh/kernel/cpu/sh5/fpu.c +++ b/arch/sh/kernel/cpu/sh5/fpu.c @@ -27,8 +27,8 @@ #define sNAN64 0xFFFFFFFFFFFFFFFFULL #define sNAN32 0xFFFFFFFFUL -static union sh_fpu_union init_fpuregs = { - .hard = { +static union thread_xstate init_fpuregs = { + .hardfpu = { .fp_regs = { [0 ... 63] = sNAN32 }, .fpscr = FPSCR_INIT } @@ -72,7 +72,7 @@ void save_fpu(struct task_struct *tsk) "fgetscr fr63\n\t" "fst.s %0, (32*8), fr63\n\t" : /* no output */ - : "r" (&tsk->thread.fpu.hard) + : "r" (&tsk->thread.xstate->hardfpu) : "memory"); } @@ -121,7 +121,7 @@ fpload(struct sh_fpu_hard_struct *fpregs) void fpinit(struct sh_fpu_hard_struct *fpregs) { - *fpregs = init_fpuregs.hard; + *fpregs = init_fpuregs.hardfpu; } asmlinkage void @@ -157,10 +157,10 @@ do_fpu_state_restore(unsigned long ex, struct pt_regs *regs) last_task_used_math = current; if (used_math()) { - fpload(¤t->thread.fpu.hard); + fpload(¤t->thread.xstate->hardfpu); } else { /* First time FPU user. */ - fpload(&init_fpuregs.hard); + fpload(&init_fpuregs.hardfpu); set_used_math(); } disable_fpu(); diff --git a/arch/sh/kernel/process_64.c b/arch/sh/kernel/process_64.c index 31f80c6..c9554a7 100644 --- a/arch/sh/kernel/process_64.c +++ b/arch/sh/kernel/process_64.c @@ -410,7 +410,7 @@ int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu) regs->sr |= SR_FD; } - memcpy(fpu, &tsk->thread.fpu.hard, sizeof(*fpu)); + memcpy(fpu, &tsk->thread.xstate->hardfpu, sizeof(*fpu)); } return fpvalid; diff --git a/arch/sh/kernel/ptrace_64.c b/arch/sh/kernel/ptrace_64.c index 873ebdc..67fbcee 100644 --- a/arch/sh/kernel/ptrace_64.c +++ b/arch/sh/kernel/ptrace_64.c @@ -88,7 +88,7 @@ get_fpu_long(struct task_struct *task, unsigned long addr) regs->sr |= SR_FD; } - tmp = ((long *)&task->thread.fpu)[addr / sizeof(unsigned long)]; + tmp = ((long *)task->thread.xstate)[addr / sizeof(unsigned long)]; return tmp; } @@ -114,7 +114,7 @@ put_fpu_long(struct task_struct *task, unsigned long addr, unsigned long data) regs = (struct pt_regs*)((unsigned char *)task + THREAD_SIZE) - 1; if (!tsk_used_math(task)) { - fpinit(&task->thread.fpu.hard); + fpinit(&task->thread.xstate->hardfpu); set_stopped_child_used_math(task); } else if (last_task_used_math == task) { enable_fpu(); @@ -124,7 +124,7 @@ put_fpu_long(struct task_struct *task, unsigned long addr, unsigned long data) regs->sr |= SR_FD; } - ((long *)&task->thread.fpu)[addr / sizeof(unsigned long)] = data; + ((long *)task->thread.xstate)[addr / sizeof(unsigned long)] = data; return 0; } @@ -222,7 +222,7 @@ int fpregs_get(struct task_struct *target, return ret; return user_regset_copyout(&pos, &count, &kbuf, &ubuf, - &target->thread.fpu.hard, 0, -1); + &target->thread.xstate->hardfpu, 0, -1); } static int fpregs_set(struct task_struct *target, @@ -239,7 +239,7 @@ static int fpregs_set(struct task_struct *target, set_stopped_child_used_math(target); return user_regset_copyin(&pos, &count, &kbuf, &ubuf, - &target->thread.fpu.hard, 0, -1); + &target->thread.xstate->hardfpu, 0, -1); } static int fpregs_active(struct task_struct *target, diff --git a/arch/sh/kernel/signal_64.c b/arch/sh/kernel/signal_64.c index ce76dbd..4733bfc 100644 --- a/arch/sh/kernel/signal_64.c +++ b/arch/sh/kernel/signal_64.c @@ -295,7 +295,7 @@ restore_sigcontext_fpu(struct pt_regs *regs, struct sigcontext __user *sc) regs->sr |= SR_FD; } - err |= __copy_from_user(¤t->thread.fpu.hard, &sc->sc_fpregs[0], + err |= __copy_from_user(¤t->thread.xstate->hardfpu, &sc->sc_fpregs[0], (sizeof(long long) * 32) + (sizeof(int) * 1)); return err; @@ -320,7 +320,7 @@ setup_sigcontext_fpu(struct pt_regs *regs, struct sigcontext __user *sc) regs->sr |= SR_FD; } - err |= __copy_to_user(&sc->sc_fpregs[0], ¤t->thread.fpu.hard, + err |= __copy_to_user(&sc->sc_fpregs[0], ¤t->thread.xstate->hardfpu, (sizeof(long long) * 32) + (sizeof(int) * 1)); clear_used_math(); diff --git a/arch/sh/kernel/traps_64.c b/arch/sh/kernel/traps_64.c index d86f531..e3f92eb 100644 --- a/arch/sh/kernel/traps_64.c +++ b/arch/sh/kernel/traps_64.c @@ -611,19 +611,19 @@ static int misaligned_fpu_load(struct pt_regs *regs, switch (width_shift) { case 2: - current->thread.fpu.hard.fp_regs[destreg] = buflo; + current->thread.xstate->hardfpu.fp_regs[destreg] = buflo; break; case 3: if (do_paired_load) { - current->thread.fpu.hard.fp_regs[destreg] = buflo; - current->thread.fpu.hard.fp_regs[destreg+1] = bufhi; + current->thread.xstate->hardfpu.fp_regs[destreg] = buflo; + current->thread.xstate->hardfpu.fp_regs[destreg+1] = bufhi; } else { #if defined(CONFIG_CPU_LITTLE_ENDIAN) - current->thread.fpu.hard.fp_regs[destreg] = bufhi; - current->thread.fpu.hard.fp_regs[destreg+1] = buflo; + current->thread.xstate->hardfpu.fp_regs[destreg] = bufhi; + current->thread.xstate->hardfpu.fp_regs[destreg+1] = buflo; #else - current->thread.fpu.hard.fp_regs[destreg] = buflo; - current->thread.fpu.hard.fp_regs[destreg+1] = bufhi; + current->thread.xstate->hardfpu.fp_regs[destreg] = buflo; + current->thread.xstate->hardfpu.fp_regs[destreg+1] = bufhi; #endif } break; @@ -681,19 +681,19 @@ static int misaligned_fpu_store(struct pt_regs *regs, switch (width_shift) { case 2: - buflo = current->thread.fpu.hard.fp_regs[srcreg]; + buflo = current->thread.xstate->hardfpu.fp_regs[srcreg]; break; case 3: if (do_paired_load) { - buflo = current->thread.fpu.hard.fp_regs[srcreg]; - bufhi = current->thread.fpu.hard.fp_regs[srcreg+1]; + buflo = current->thread.xstate->hardfpu.fp_regs[srcreg]; + bufhi = current->thread.xstate->hardfpu.fp_regs[srcreg+1]; } else { #if defined(CONFIG_CPU_LITTLE_ENDIAN) - bufhi = current->thread.fpu.hard.fp_regs[srcreg]; - buflo = current->thread.fpu.hard.fp_regs[srcreg+1]; + bufhi = current->thread.xstate->hardfpu.fp_regs[srcreg]; + buflo = current->thread.xstate->hardfpu.fp_regs[srcreg+1]; #else - buflo = current->thread.fpu.hard.fp_regs[srcreg]; - bufhi = current->thread.fpu.hard.fp_regs[srcreg+1]; + buflo = current->thread.xstate->hardfpu.fp_regs[srcreg]; + bufhi = current->thread.xstate->hardfpu.fp_regs[srcreg+1]; #endif } break; -- cgit v0.10.2 From 88ea1a445a84fcfbedb810c01e84d6711352bd82 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 15:41:50 +0900 Subject: sh64: Fix up PC casting in unaligned fixup notifier with 32bit ABI. Presently the build bails with the following: CC arch/sh/mm/alignment.o cc1: warnings being treated as errors arch/sh/mm/alignment.c: In function 'unaligned_fixups_notify': arch/sh/mm/alignment.c:69: warning: cast to pointer from integer of different size arch/sh/mm/alignment.c:74: warning: cast to pointer from integer of different size make[2]: *** [arch/sh/mm/alignment.o] Error 1 This is due to the fact that regs->pc is always 64-bit, while the pointer size depends on the ABI. Wrapping through instruction_pointer() takes care of the appropriate casting for both configurations. Signed-off-by: Paul Mundt diff --git a/arch/sh/mm/alignment.c b/arch/sh/mm/alignment.c index e615151..2da8088 100644 --- a/arch/sh/mm/alignment.c +++ b/arch/sh/mm/alignment.c @@ -69,12 +69,12 @@ void unaligned_fixups_notify(struct task_struct *tsk, insn_size_t insn, pr_notice("Fixing up unaligned userspace access " "in \"%s\" pid=%d pc=0x%p ins=0x%04hx\n", tsk->comm, task_pid_nr(tsk), - (void *)regs->pc, insn); + (void *)instruction_pointer(regs), insn); else if (se_kernmode_warn && printk_ratelimit()) pr_notice("Fixing up unaligned kernel access " "in \"%s\" pid=%d pc=0x%p ins=0x%04hx\n", tsk->comm, task_pid_nr(tsk), - (void *)regs->pc, insn); + (void *)instruction_pointer(regs), insn); } static const char *se_usermode_action[] = { -- cgit v0.10.2 From d6db8888c8957fbdcd611e1321a6f6a0d6fb7e15 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 15:55:27 +0900 Subject: sh64: Use the shared FPU state restorer. This kills off the sh64-specific state restorer and switches over to the generic one. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/cpu/sh5/entry.S b/arch/sh/kernel/cpu/sh5/entry.S index 8f13f73..36d031c 100644 --- a/arch/sh/kernel/cpu/sh5/entry.S +++ b/arch/sh/kernel/cpu/sh5/entry.S @@ -1124,7 +1124,7 @@ fpu_error_or_IRQA: pta its_IRQ, tr0 beqi/l r4, EVENT_INTERRUPT, tr0 #ifdef CONFIG_SH_FPU - movi do_fpu_state_restore, r6 + movi fpu_state_restore_trap_handler, r6 #else movi do_exception_error, r6 #endif @@ -1135,7 +1135,7 @@ fpu_error_or_IRQB: pta its_IRQ, tr0 beqi/l r4, EVENT_INTERRUPT, tr0 #ifdef CONFIG_SH_FPU - movi do_fpu_state_restore, r6 + movi fpu_state_restore_trap_handler, r6 #else movi do_exception_error, r6 #endif diff --git a/arch/sh/kernel/cpu/sh5/fpu.c b/arch/sh/kernel/cpu/sh5/fpu.c index 92df285..4b3bb35 100644 --- a/arch/sh/kernel/cpu/sh5/fpu.c +++ b/arch/sh/kernel/cpu/sh5/fpu.c @@ -15,24 +15,6 @@ #include #include #include -#include -#include -#include - -/* - * Initially load the FPU with signalling NANS. This bit pattern - * has the property that no matter whether considered as single or as - * double precision, it still represents a signalling NAN. - */ -#define sNAN64 0xFFFFFFFFFFFFFFFFULL -#define sNAN32 0xFFFFFFFFUL - -static union thread_xstate init_fpuregs = { - .hardfpu = { - .fp_regs = { [0 ... 63] = sNAN32 }, - .fpscr = FPSCR_INIT - } -}; void save_fpu(struct task_struct *tsk) { @@ -76,8 +58,7 @@ void save_fpu(struct task_struct *tsk) : "memory"); } -static inline void -fpload(struct sh_fpu_hard_struct *fpregs) +void restore_fpu(struct task_struct *tsk) { asm volatile("fld.p %0, (0*8), fp0\n\t" "fld.p %0, (1*8), fp2\n\t" @@ -116,16 +97,11 @@ fpload(struct sh_fpu_hard_struct *fpregs) "fld.p %0, (31*8), fp62\n\t" : /* no output */ - : "r" (fpregs) ); -} - -void fpinit(struct sh_fpu_hard_struct *fpregs) -{ - *fpregs = init_fpuregs.hardfpu; + : "r" (&tsk->thread.xstate->hardfpu) + : "memory"); } -asmlinkage void -do_fpu_error(unsigned long ex, struct pt_regs *regs) +asmlinkage void do_fpu_error(unsigned long ex, struct pt_regs *regs) { struct task_struct *tsk = current; @@ -133,35 +109,6 @@ do_fpu_error(unsigned long ex, struct pt_regs *regs) tsk->thread.trap_no = 11; tsk->thread.error_code = 0; - force_sig(SIGFPE, tsk); -} - - -asmlinkage void -do_fpu_state_restore(unsigned long ex, struct pt_regs *regs) -{ - void die(const char *str, struct pt_regs *regs, long err); - - if (! user_mode(regs)) - die("FPU used in kernel", regs, ex); - regs->sr &= ~SR_FD; - - if (last_task_used_math == current) - return; - - enable_fpu(); - if (last_task_used_math != NULL) - /* Other processes fpu state, save away */ - save_fpu(last_task_used_math); - - last_task_used_math = current; - if (used_math()) { - fpload(¤t->thread.xstate->hardfpu); - } else { - /* First time FPU user. */ - fpload(&init_fpuregs.hardfpu); - set_used_math(); - } - disable_fpu(); + force_sig(SIGFPE, tsk); } diff --git a/arch/sh/kernel/ptrace_64.c b/arch/sh/kernel/ptrace_64.c index 67fbcee..23fbd92 100644 --- a/arch/sh/kernel/ptrace_64.c +++ b/arch/sh/kernel/ptrace_64.c @@ -114,8 +114,7 @@ put_fpu_long(struct task_struct *task, unsigned long addr, unsigned long data) regs = (struct pt_regs*)((unsigned char *)task + THREAD_SIZE) - 1; if (!tsk_used_math(task)) { - fpinit(&task->thread.xstate->hardfpu); - set_stopped_child_used_math(task); + init_fpu(task); } else if (last_task_used_math == task) { enable_fpu(); save_fpu(task); -- cgit v0.10.2 From a4ae2b2b18d1766768987dc5de42dfa3c2a6d9f7 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 15:58:27 +0900 Subject: sh64: Fixup build breakage from breakpoint handler rename. The breakpoint handler was renamed on sh32, but sh64 was overlooked in the conversion. Fix it up now. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/cpu/sh5/entry.S b/arch/sh/kernel/cpu/sh5/entry.S index 36d031c..6b80295 100644 --- a/arch/sh/kernel/cpu/sh5/entry.S +++ b/arch/sh/kernel/cpu/sh5/entry.S @@ -187,7 +187,7 @@ trap_jtable: .rept 6 .long do_exception_error /* 0x880 - 0x920 */ .endr - .long do_software_break_point /* 0x940 */ + .long breakpoint_trap_handler /* 0x940 */ .long do_exception_error /* 0x960 */ .long do_single_step /* 0x980 */ diff --git a/arch/sh/kernel/ptrace_64.c b/arch/sh/kernel/ptrace_64.c index 23fbd92..2f6140e 100644 --- a/arch/sh/kernel/ptrace_64.c +++ b/arch/sh/kernel/ptrace_64.c @@ -478,9 +478,10 @@ asmlinkage void do_single_step(unsigned long long vec, struct pt_regs *regs) } /* Called with interrupts disabled */ -asmlinkage void do_software_break_point(unsigned long long vec, - struct pt_regs *regs) +BUILD_TRAP_HANDLER(breakpoint) { + TRAP_HANDLER_DECL; + /* We need to forward step the PC, to counteract the backstep done in signal.c. */ local_irq_enable(); -- cgit v0.10.2 From 43a1839cb1e0e0ed08b8ace0adb3716865fd0c4c Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 19:37:14 +0900 Subject: sh: SH7786 clock framework rewrite. This rewrites the SH7786 clock framework support completely. It's reworked to provide all of the DIV4 and MSTP function clocks. This brings it in line with the current clock framework code and lets us drop SH7786 from the list of CPUs that require legacy CPG handling. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 8d0eabb..352879c 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -551,8 +551,7 @@ config SH_PCLK_FREQ CPU_SUBTYPE_SH7203 || \ CPU_SUBTYPE_SH7206 || \ CPU_SUBTYPE_SH7263 || \ - CPU_SUBTYPE_MXG || \ - CPU_SUBTYPE_SH7786 + CPU_SUBTYPE_MXG default "60000000" if CPU_SUBTYPE_SH7751 || CPU_SUBTYPE_SH7751R default "66000000" if CPU_SUBTYPE_SH4_202 default "50000000" @@ -566,7 +565,8 @@ config SH_CLK_CPG config SH_CLK_CPG_LEGACY depends on SH_CLK_CPG - def_bool y if !CPU_SUBTYPE_SH7785 && !ARCH_SHMOBILE + def_bool y if !CPU_SUBTYPE_SH7785 && !ARCH_SHMOBILE && \ + !CPU_SUBTYPE_SH7786 config SH_CLK_MD int "CPU Mode Pin Setting" diff --git a/arch/sh/kernel/cpu/sh4a/clock-sh7786.c b/arch/sh/kernel/cpu/sh4a/clock-sh7786.c index a0e8869..494c636 100644 --- a/arch/sh/kernel/cpu/sh4a/clock-sh7786.c +++ b/arch/sh/kernel/cpu/sh4a/clock-sh7786.c @@ -3,11 +3,7 @@ * * SH7786 support for the clock framework * - * Copyright (C) 2008, 2009 Renesas Solutions Corp. - * Kuninori Morimoto - * - * Based on SH7785 - * Copyright (C) 2007 Paul Mundt + * Copyright (C) 2010 Paul Mundt * * This file is subject to the terms and conditions of the GNU General Public * License. See the file "COPYING" in the main directory of this archive @@ -15,127 +11,123 @@ */ #include #include +#include +#include #include #include -#include - -static int ifc_divisors[] = { 1, 2, 4, 1 }; -static int sfc_divisors[] = { 1, 1, 4, 1 }; -static int bfc_divisors[] = { 1, 1, 1, 1, 1, 12, 16, 1, - 24, 32, 1, 1, 1, 1, 1, 1 }; -static int mfc_divisors[] = { 1, 1, 4, 1 }; -static int pfc_divisors[] = { 1, 1, 1, 1, 1, 1, 16, 1, - 24, 32, 1, 48, 1, 1, 1, 1 }; - -static void master_clk_init(struct clk *clk) -{ - clk->rate *= pfc_divisors[ctrl_inl(FRQMR1) & 0x000f]; -} -static struct clk_ops sh7786_master_clk_ops = { - .init = master_clk_init, +/* + * Default rate for the root input clock, reset this with clk_set_rate() + * from the platform code. + */ +static struct clk extal_clk = { + .name = "extal", + .id = -1, + .rate = 33333333, }; -static unsigned long module_clk_recalc(struct clk *clk) +static unsigned long pll_recalc(struct clk *clk) { - int idx = (ctrl_inl(FRQMR1) & 0x000f); - return clk->parent->rate / pfc_divisors[idx]; -} + int multiplier; -static struct clk_ops sh7786_module_clk_ops = { - .recalc = module_clk_recalc, -}; + /* + * Clock modes 0, 1, and 2 use an x64 multiplier against PLL1, + * while modes 3, 4, and 5 use an x32. + */ + multiplier = (sh_mv.mv_mode_pins() & 0xf) < 3 ? 64 : 32; -static unsigned long bus_clk_recalc(struct clk *clk) -{ - int idx = ((ctrl_inl(FRQMR1) >> 16) & 0x000f); - return clk->parent->rate / bfc_divisors[idx]; + return clk->parent->rate * multiplier; } -static struct clk_ops sh7786_bus_clk_ops = { - .recalc = bus_clk_recalc, +static struct clk_ops pll_clk_ops = { + .recalc = pll_recalc, }; -static unsigned long cpu_clk_recalc(struct clk *clk) -{ - int idx = ((ctrl_inl(FRQMR1) >> 28) & 0x0003); - return clk->parent->rate / ifc_divisors[idx]; -} - -static struct clk_ops sh7786_cpu_clk_ops = { - .recalc = cpu_clk_recalc, +static struct clk pll_clk = { + .name = "pll_clk", + .id = -1, + .ops = &pll_clk_ops, + .parent = &extal_clk, + .flags = CLK_ENABLE_ON_INIT, }; -static struct clk_ops *sh7786_clk_ops[] = { - &sh7786_master_clk_ops, - &sh7786_module_clk_ops, - &sh7786_bus_clk_ops, - &sh7786_cpu_clk_ops, +static struct clk *clks[] = { + &extal_clk, + &pll_clk, }; -void __init arch_init_clk_ops(struct clk_ops **ops, int idx) -{ - if (idx < ARRAY_SIZE(sh7786_clk_ops)) - *ops = sh7786_clk_ops[idx]; -} - -static unsigned long shyway_clk_recalc(struct clk *clk) -{ - int idx = ((ctrl_inl(FRQMR1) >> 20) & 0x0003); - return clk->parent->rate / sfc_divisors[idx]; -} +static unsigned int div2[] = { 1, 2, 4, 6, 8, 12, 16, 18, + 24, 32, 36, 48 }; -static struct clk_ops sh7786_shyway_clk_ops = { - .recalc = shyway_clk_recalc, +static struct clk_div_mult_table div4_table = { + .divisors = div2, + .nr_divisors = ARRAY_SIZE(div2), }; -static struct clk sh7786_shyway_clk = { - .name = "shyway_clk", - .flags = CLK_ENABLE_ON_INIT, - .ops = &sh7786_shyway_clk_ops, -}; +enum { DIV4_I, DIV4_SH, DIV4_B, DIV4_DDR, DIV4_DU, DIV4_P, DIV4_NR }; -static unsigned long ddr_clk_recalc(struct clk *clk) -{ - int idx = ((ctrl_inl(FRQMR1) >> 12) & 0x0003); - return clk->parent->rate / mfc_divisors[idx]; -} +#define DIV4(_str, _bit, _mask, _flags) \ + SH_CLK_DIV4(_str, &pll_clk, FRQMR1, _bit, _mask, _flags) -static struct clk_ops sh7786_ddr_clk_ops = { - .recalc = ddr_clk_recalc, +struct clk div4_clks[DIV4_NR] = { + [DIV4_P] = DIV4("peripheral_clk", 0, 0x0b40, 0), + [DIV4_DU] = DIV4("du_clk", 4, 0x0010, 0), + [DIV4_DDR] = DIV4("ddr_clk", 12, 0x0002, CLK_ENABLE_ON_INIT), + [DIV4_B] = DIV4("bus_clk", 16, 0x0360, CLK_ENABLE_ON_INIT), + [DIV4_SH] = DIV4("shyway_clk", 20, 0x0002, CLK_ENABLE_ON_INIT), + [DIV4_I] = DIV4("cpu_clk", 28, 0x0006, CLK_ENABLE_ON_INIT), }; -static struct clk sh7786_ddr_clk = { - .name = "ddr_clk", - .flags = CLK_ENABLE_ON_INIT, - .ops = &sh7786_ddr_clk_ops, -}; - -/* - * Additional SH7786-specific on-chip clocks that aren't already part of the - * clock framework - */ -static struct clk *sh7786_onchip_clocks[] = { - &sh7786_shyway_clk, - &sh7786_ddr_clk, +#define MSTPCR0 0xffc40030 +#define MSTPCR1 0xffc40034 + +static struct clk mstp_clks[] = { + /* MSTPCR0 */ + SH_CLK_MSTP32("scif_fck", 5, &div4_clks[DIV4_P], MSTPCR0, 29, 0), + SH_CLK_MSTP32("scif_fck", 4, &div4_clks[DIV4_P], MSTPCR0, 28, 0), + SH_CLK_MSTP32("scif_fck", 3, &div4_clks[DIV4_P], MSTPCR0, 27, 0), + SH_CLK_MSTP32("scif_fck", 2, &div4_clks[DIV4_P], MSTPCR0, 26, 0), + SH_CLK_MSTP32("scif_fck", 1, &div4_clks[DIV4_P], MSTPCR0, 25, 0), + SH_CLK_MSTP32("scif_fck", 0, &div4_clks[DIV4_P], MSTPCR0, 24, 0), + SH_CLK_MSTP32("ssi_fck", 3, &div4_clks[DIV4_P], MSTPCR0, 23, 0), + SH_CLK_MSTP32("ssi_fck", 2, &div4_clks[DIV4_P], MSTPCR0, 22, 0), + SH_CLK_MSTP32("ssi_fck", 1, &div4_clks[DIV4_P], MSTPCR0, 21, 0), + SH_CLK_MSTP32("ssi_fck", 0, &div4_clks[DIV4_P], MSTPCR0, 20, 0), + SH_CLK_MSTP32("hac_fck", 1, &div4_clks[DIV4_P], MSTPCR0, 17, 0), + SH_CLK_MSTP32("hac_fck", 0, &div4_clks[DIV4_P], MSTPCR0, 16, 0), + SH_CLK_MSTP32("i2c_fck", 1, &div4_clks[DIV4_P], MSTPCR0, 15, 0), + SH_CLK_MSTP32("i2c_fck", 0, &div4_clks[DIV4_P], MSTPCR0, 14, 0), + SH_CLK_MSTP32("tmu9_11_fck", -1, &div4_clks[DIV4_P], MSTPCR0, 11, 0), + SH_CLK_MSTP32("tmu678_fck", -1, &div4_clks[DIV4_P], MSTPCR0, 10, 0), + SH_CLK_MSTP32("tmu345_fck", -1, &div4_clks[DIV4_P], MSTPCR0, 9, 0), + SH_CLK_MSTP32("tmu012_fck", -1, &div4_clks[DIV4_P], MSTPCR0, 8, 0), + SH_CLK_MSTP32("sdif_fck", 1, &div4_clks[DIV4_P], MSTPCR0, 5, 0), + SH_CLK_MSTP32("sdif_fck", 0, &div4_clks[DIV4_P], MSTPCR0, 4, 0), + SH_CLK_MSTP32("hspi_fck", -1, &div4_clks[DIV4_P], MSTPCR0, 2, 0), + + /* MSTPCR1 */ + SH_CLK_MSTP32("usb_fck", -1, NULL, MSTPCR1, 12, 0), + SH_CLK_MSTP32("pcie_fck", 2, NULL, MSTPCR1, 10, 0), + SH_CLK_MSTP32("pcie_fck", 1, NULL, MSTPCR1, 9, 0), + SH_CLK_MSTP32("pcie_fck", 0, NULL, MSTPCR1, 8, 0), + SH_CLK_MSTP32("dmac_11_6_fck", -1, NULL, MSTPCR1, 5, 0), + SH_CLK_MSTP32("dmac_5_0_fck", -1, NULL, MSTPCR1, 4, 0), + SH_CLK_MSTP32("du_fck", -1, NULL, MSTPCR1, 3, 0), + SH_CLK_MSTP32("ether_fck", -1, NULL, MSTPCR1, 2, 0), }; int __init arch_clk_init(void) { - struct clk *clk; int i, ret = 0; - cpg_clk_init(); - - clk = clk_get(NULL, "master_clk"); - for (i = 0; i < ARRAY_SIZE(sh7786_onchip_clocks); i++) { - struct clk *clkp = sh7786_onchip_clocks[i]; - - clkp->parent = clk; - ret |= clk_register(clkp); - } + for (i = 0; i < ARRAY_SIZE(clks); i++) + ret |= clk_register(clks[i]); - clk_put(clk); + if (!ret) + ret = sh_clk_div4_register(div4_clks, ARRAY_SIZE(div4_clks), + &div4_table); + if (!ret) + ret = sh_clk_mstp32_register(mstp_clks, ARRAY_SIZE(mstp_clks)); return ret; } -- cgit v0.10.2 From c8098218271d19e7123d8b9137b9a4b87e5ffec0 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 19:38:36 +0900 Subject: sh: mach-sdk7786: Detect/configure/propagate EXTAL. This uses the mode pins exposed through the FPGA to work out whether we're driven from EXTAL or not and does the appropriate setup and propagation through the clock framework. This will also -EINVAL out for anyone adding in their own oscillators, forcing proper configuration with the clock framework instead of proceeding on with bogus clock values. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-sdk7786/setup.c b/arch/sh/boards/mach-sdk7786/setup.c index c38c6cc..0e4b1c3 100644 --- a/arch/sh/boards/mach-sdk7786/setup.c +++ b/arch/sh/boards/mach-sdk7786/setup.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include #include @@ -235,6 +236,27 @@ static int sdk7786_mode_pins(void) return pin_states; } +static int sdk7786_clk_init(void) +{ + struct clk *clk; + int ret; + + /* + * Only handle the EXTAL case, anyone interfacing a crystal + * resonator will need to provide their own input clock. + */ + if (test_mode_pin(MODE_PIN9)) + return -EINVAL; + + clk = clk_get(NULL, "extal"); + if (!clk || IS_ERR(clk)) + return PTR_ERR(clk); + ret = clk_set_rate(clk, 33333333); + clk_put(clk); + + return ret; +} + /* Initialize the board */ static void __init sdk7786_setup(char **cmdline_p) { @@ -248,5 +270,6 @@ static struct sh_machine_vector mv_sdk7786 __initmv = { .mv_name = "SDK7786", .mv_setup = sdk7786_setup, .mv_mode_pins = sdk7786_mode_pins, + .mv_clk_init = sdk7786_clk_init, .mv_init_irq = init_sdk7786_IRQ, }; -- cgit v0.10.2 From dea3cf1c3935047a5b0d21b21d26718e09135f97 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 20:09:28 +0900 Subject: sh: urquell: Handle EXTAL configuration here, too. urquell happens to use the same mode pins and EXTAL configuration as SDK7786, so just copy it over. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/board-urquell.c b/arch/sh/boards/board-urquell.c index 268f09e..8aeac64 100644 --- a/arch/sh/boards/board-urquell.c +++ b/arch/sh/boards/board-urquell.c @@ -2,7 +2,7 @@ * Renesas Technology Corp. SH7786 Urquell Support. * * Copyright (C) 2008 Kuninori Morimoto - * Copyright (C) 2009 Paul Mundt + * Copyright (C) 2009, 2010 Paul Mundt * * Based on board-sh7785lcr.c * Copyright (C) 2008 Yoshihiro Shimoda @@ -19,6 +19,7 @@ #include #include #include +#include #include #include #include @@ -175,6 +176,27 @@ static int urquell_mode_pins(void) return __raw_readw(UBOARDREG(MDSWMR)); } +static int urquell_clk_init(void) +{ + struct clk *clk; + int ret; + + /* + * Only handle the EXTAL case, anyone interfacing a crystal + * resonator will need to provide their own input clock. + */ + if (test_mode_pin(MODE_PIN9)) + return -EINVAL; + + clk = clk_get(NULL, "extal"); + if (!clk || IS_ERR(clk)) + return PTR_ERR(clk); + ret = clk_set_rate(clk, 33333333); + clk_put(clk); + + return ret; +} + /* Initialize the board */ static void __init urquell_setup(char **cmdline_p) { @@ -191,4 +213,5 @@ static struct sh_machine_vector mv_urquell __initmv = { .mv_setup = urquell_setup, .mv_init_irq = urquell_init_irq, .mv_mode_pins = urquell_mode_pins, + .mv_clk_init = urquell_clk_init, }; -- cgit v0.10.2 From 14965f16b4bb8f3e51b09c1d8f61b8e98f9d12db Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 19 Jan 2010 20:10:33 +0900 Subject: sh: Fix up sdk7780 and urquell builds. These two got broken in the heartbeat private data conversion, fix them up. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/board-urquell.c b/arch/sh/boards/board-urquell.c index 8aeac64..a9bd6e3 100644 --- a/arch/sh/boards/board-urquell.c +++ b/arch/sh/boards/board-urquell.c @@ -60,7 +60,7 @@ static struct resource heartbeat_resource = { static struct platform_device heartbeat_device = { .name = "heartbeat", .id = -1, - .num_resources = 1 + .num_resources = 1, .resource = &heartbeat_resource, }; diff --git a/arch/sh/boards/mach-sdk7780/setup.c b/arch/sh/boards/mach-sdk7780/setup.c index 1d04657..b887373 100644 --- a/arch/sh/boards/mach-sdk7780/setup.c +++ b/arch/sh/boards/mach-sdk7780/setup.c @@ -30,7 +30,7 @@ static struct resource heartbeat_resource = { static struct platform_device heartbeat_device = { .name = "heartbeat", .id = -1, - .num_resources = 1 + .num_resources = 1, .resource = &heartbeat_resource, }; -- cgit v0.10.2 From 31c3af503eb75488aafb7a3d292b9e00962f2cee Mon Sep 17 00:00:00 2001 From: Guennadi Liakhovetski Date: Tue, 19 Jan 2010 11:14:31 +0000 Subject: sh: support SIU sourcing from external clock on sh7722 Implement .set_rate() for all SH "div4 clocks," .enable(), .disable(), and .set_parent() for those, that support them. This allows, among other uses, reparenting of SIU clocks to the external source, and enabling and disabling of the IrDA clock on sh7722. Signed-off-by: Guennadi Liakhovetski Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/clock.h b/arch/sh/include/asm/clock.h index 9fe7d7f..501d0b0 100644 --- a/arch/sh/include/asm/clock.h +++ b/arch/sh/include/asm/clock.h @@ -148,6 +148,10 @@ int sh_clk_mstp32_register(struct clk *clks, int nr); int sh_clk_div4_register(struct clk *clks, int nr, struct clk_div_mult_table *table); +int sh_clk_div4_enable_register(struct clk *clks, int nr, + struct clk_div_mult_table *table); +int sh_clk_div4_reparent_register(struct clk *clks, int nr, + struct clk_div_mult_table *table); #define SH_CLK_DIV6(_name, _parent, _reg, _flags) \ { \ diff --git a/arch/sh/kernel/cpu/clock-cpg.c b/arch/sh/kernel/cpu/clock-cpg.c index 6dfe2cc..2827abb 100644 --- a/arch/sh/kernel/cpu/clock-cpg.c +++ b/arch/sh/kernel/cpu/clock-cpg.c @@ -160,13 +160,81 @@ static unsigned long sh_clk_div4_recalc(struct clk *clk) return clk->freq_table[idx].frequency; } +static int sh_clk_div4_set_parent(struct clk *clk, struct clk *parent) +{ + struct clk_div_mult_table *table = clk->priv; + u32 value; + int ret; + + if (!strcmp("pll_clk", parent->name)) + value = __raw_readl(clk->enable_reg) & ~(1 << 7); + else + value = __raw_readl(clk->enable_reg) | (1 << 7); + + ret = clk_reparent(clk, parent); + if (ret < 0) + return ret; + + __raw_writel(value, clk->enable_reg); + + /* Rebiuld the frequency table */ + clk_rate_table_build(clk, clk->freq_table, table->nr_divisors, + table, &clk->arch_flags); + + return 0; +} + +static int sh_clk_div4_set_rate(struct clk *clk, unsigned long rate, int algo_id) +{ + unsigned long value; + int idx = clk_rate_table_find(clk, clk->freq_table, rate); + if (idx < 0) + return idx; + + value = __raw_readl(clk->enable_reg); + value &= ~0xf; + value |= idx; + __raw_writel(value, clk->enable_reg); + + return 0; +} + +static int sh_clk_div4_enable(struct clk *clk) +{ + __raw_writel(__raw_readl(clk->enable_reg) & ~(1 << 8), clk->enable_reg); + return 0; +} + +static void sh_clk_div4_disable(struct clk *clk) +{ + __raw_writel(__raw_readl(clk->enable_reg) | (1 << 8), clk->enable_reg); +} + static struct clk_ops sh_clk_div4_clk_ops = { .recalc = sh_clk_div4_recalc, + .set_rate = sh_clk_div4_set_rate, .round_rate = sh_clk_div_round_rate, }; -int __init sh_clk_div4_register(struct clk *clks, int nr, - struct clk_div_mult_table *table) +static struct clk_ops sh_clk_div4_enable_clk_ops = { + .recalc = sh_clk_div4_recalc, + .set_rate = sh_clk_div4_set_rate, + .round_rate = sh_clk_div_round_rate, + .enable = sh_clk_div4_enable, + .disable = sh_clk_div4_disable, +}; + +static struct clk_ops sh_clk_div4_reparent_clk_ops = { + .recalc = sh_clk_div4_recalc, + .set_rate = sh_clk_div4_set_rate, + .round_rate = sh_clk_div_round_rate, + .enable = sh_clk_div4_enable, + .disable = sh_clk_div4_disable, + .set_parent = sh_clk_div4_set_parent, +}; + +static int __init sh_clk_div4_register_ops(struct clk *clks, int nr, + struct clk_div_mult_table *table, struct clk_ops *ops) { struct clk *clkp; void *freq_table; @@ -185,7 +253,7 @@ int __init sh_clk_div4_register(struct clk *clks, int nr, for (k = 0; !ret && (k < nr); k++) { clkp = clks + k; - clkp->ops = &sh_clk_div4_clk_ops; + clkp->ops = ops; clkp->id = -1; clkp->priv = table; @@ -198,6 +266,26 @@ int __init sh_clk_div4_register(struct clk *clks, int nr, return ret; } +int __init sh_clk_div4_register(struct clk *clks, int nr, + struct clk_div_mult_table *table) +{ + return sh_clk_div4_register_ops(clks, nr, table, &sh_clk_div4_clk_ops); +} + +int __init sh_clk_div4_enable_register(struct clk *clks, int nr, + struct clk_div_mult_table *table) +{ + return sh_clk_div4_register_ops(clks, nr, table, + &sh_clk_div4_enable_clk_ops); +} + +int __init sh_clk_div4_reparent_register(struct clk *clks, int nr, + struct clk_div_mult_table *table) +{ + return sh_clk_div4_register_ops(clks, nr, table, + &sh_clk_div4_reparent_clk_ops); +} + #ifdef CONFIG_SH_CLK_CPG_LEGACY static struct clk master_clk = { .name = "master_clk", diff --git a/arch/sh/kernel/cpu/sh4a/clock-sh7722.c b/arch/sh/kernel/cpu/sh4a/clock-sh7722.c index ea38b55..860ee2b 100644 --- a/arch/sh/kernel/cpu/sh4a/clock-sh7722.c +++ b/arch/sh/kernel/cpu/sh4a/clock-sh7722.c @@ -117,12 +117,11 @@ static struct clk_div_mult_table div4_table = { .nr_multipliers = ARRAY_SIZE(multipliers), }; -enum { DIV4_I, DIV4_U, DIV4_SH, DIV4_B, DIV4_B3, DIV4_P, - DIV4_SIUA, DIV4_SIUB, DIV4_IRDA, DIV4_NR }; - #define DIV4(_str, _reg, _bit, _mask, _flags) \ SH_CLK_DIV4(_str, &pll_clk, _reg, _bit, _mask, _flags) +enum { DIV4_I, DIV4_U, DIV4_SH, DIV4_B, DIV4_B3, DIV4_P, DIV4_NR }; + struct clk div4_clks[DIV4_NR] = { [DIV4_I] = DIV4("cpu_clk", FRQCR, 20, 0x1fef, CLK_ENABLE_ON_INIT), [DIV4_U] = DIV4("umem_clk", FRQCR, 16, 0x1fff, CLK_ENABLE_ON_INIT), @@ -130,9 +129,19 @@ struct clk div4_clks[DIV4_NR] = { [DIV4_B] = DIV4("bus_clk", FRQCR, 8, 0x1fff, CLK_ENABLE_ON_INIT), [DIV4_B3] = DIV4("b3_clk", FRQCR, 4, 0x1fff, CLK_ENABLE_ON_INIT), [DIV4_P] = DIV4("peripheral_clk", FRQCR, 0, 0x1fff, 0), +}; + +enum { DIV4_IRDA, DIV4_ENABLE_NR }; + +struct clk div4_enable_clks[DIV4_ENABLE_NR] = { + [DIV4_IRDA] = DIV4("irda_clk", IRDACLKCR, 0, 0x1fff, 0), +}; + +enum { DIV4_SIUA, DIV4_SIUB, DIV4_REPARENT_NR }; + +struct clk div4_reparent_clks[DIV4_REPARENT_NR] = { [DIV4_SIUA] = DIV4("siua_clk", SCLKACR, 0, 0x1fff, 0), [DIV4_SIUB] = DIV4("siub_clk", SCLKBCR, 0, 0x1fff, 0), - [DIV4_IRDA] = DIV4("irda_clk", IRDACLKCR, 0, 0x1fff, 0), }; struct clk div6_clks[] = { @@ -189,6 +198,14 @@ int __init arch_clk_init(void) ret = sh_clk_div4_register(div4_clks, DIV4_NR, &div4_table); if (!ret) + ret = sh_clk_div4_enable_register(div4_enable_clks, + DIV4_ENABLE_NR, &div4_table); + + if (!ret) + ret = sh_clk_div4_reparent_register(div4_reparent_clks, + DIV4_REPARENT_NR, &div4_table); + + if (!ret) ret = sh_clk_div6_register(div6_clks, ARRAY_SIZE(div6_clks)); if (!ret) -- cgit v0.10.2 From 70911b861bfbba1851cbb3497635030f34703c7a Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 02:45:05 +0900 Subject: sh: Shut up noisy IOREMAP_FIXED=n build. The ioremap_fixed() stub neglected to provide a return value, resulting in a fairly noisy build. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index 4bad72a..e4f563f 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -245,10 +245,11 @@ extern int iounmap_fixed(void __iomem *); extern void ioremap_fixed_init(void); #else static inline void __iomem * -ioremap_fixed(resource_size t phys_addr, unsigned long offset, +ioremap_fixed(resource_size_t phys_addr, unsigned long offset, unsigned long size, pgprot_t prot) { BUG(); + return NULL; } static inline void ioremap_fixed_init(void) { } -- cgit v0.10.2 From 50dd3145a58b5cef03a95c1b98765bcc847a72d0 Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 19 Jan 2010 13:52:28 +0000 Subject: sh: update PFC to allow any enum in MARK lists This patch updates the PFC code with some clarifying comments together with a functional change. The change allows function type of GPIO to select any type of enum in their MARK lists. Without this patch only function type of enums are allowed in MARK lists. Signed-off-by: Magnus Damm Signed-off-by: Paul Mundt diff --git a/drivers/sh/pfc.c b/drivers/sh/pfc.c index 082604e..cf0303ac 100644 --- a/drivers/sh/pfc.c +++ b/drivers/sh/pfc.c @@ -337,12 +337,39 @@ static int pinmux_config_gpio(struct pinmux_info *gpioc, unsigned gpio, if (!enum_id) break; + /* first check if this is a function enum */ in_range = enum_in_range(enum_id, &gpioc->function); - if (!in_range && range) { - in_range = enum_in_range(enum_id, range); - - if (in_range && enum_id == range->force) - continue; + if (!in_range) { + /* not a function enum */ + if (range) { + /* + * other range exists, so this pin is + * a regular GPIO pin that now is being + * bound to a specific direction. + * + * for this case we only allow function enums + * and the enums that match the other range. + */ + in_range = enum_in_range(enum_id, range); + + /* + * special case pass through for fixed + * input-only or output-only pins without + * function enum register association. + */ + if (in_range && enum_id == range->force) + continue; + } else { + /* + * no other range exists, so this pin + * must then be of the function type. + * + * allow function type pins to select + * any combination of function/in/out + * in their MARK lists. + */ + in_range = 1; + } } if (!in_range) -- cgit v0.10.2 From bdc27300f5718626a3817e6478e339f6cca6b994 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 03:38:56 +0900 Subject: sh: Handle SH-4 FPU variants with broken CVR values. Usually we can look to the CVR to work out whether we have an FPU or not. Unfortunately not all parts comply with this, so just set the flag manually for all SH-4 parts and clear it on the only SH-4 that doesn't have one (SH4-501). Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/cpu/sh4/probe.c b/arch/sh/kernel/cpu/sh4/probe.c index 21371f8..cc02b31 100644 --- a/arch/sh/kernel/cpu/sh4/probe.c +++ b/arch/sh/kernel/cpu/sh4/probe.c @@ -71,11 +71,11 @@ int __init detect_cpu_and_cache_system(void) boot_cpu_data.dcache.ways = 4; } else { /* And some SH-4 defaults.. */ - boot_cpu_data.flags |= CPU_HAS_PTEA; + boot_cpu_data.flags |= CPU_HAS_PTEA | CPU_HAS_FPU; boot_cpu_data.family = CPU_FAMILY_SH4; } - /* FPU detection works for everyone */ + /* FPU detection works for almost everyone */ if ((cvr & 0x20000000)) boot_cpu_data.flags |= CPU_HAS_FPU; @@ -161,6 +161,7 @@ int __init detect_cpu_and_cache_system(void) break; case 0x700: boot_cpu_data.type = CPU_SH4_501; + boot_cpu_data.flags &= ~CPU_HAS_FPU; boot_cpu_data.icache.ways = 2; boot_cpu_data.dcache.ways = 2; break; -- cgit v0.10.2 From e2bcd936eb95d0019ca5e05f9fdd27e770ddded1 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 20 Jan 2010 10:44:58 +1100 Subject: xfs: directory names are unsigned Convert the struct xfs_name to use unsigned chars for the name strings to match both what is stored on disk (__uint8_t) and what the VFS expects (unsigned char). Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/xfs_types.h b/fs/xfs/xfs_types.h index d725428..b099045 100644 --- a/fs/xfs/xfs_types.h +++ b/fs/xfs/xfs_types.h @@ -151,8 +151,8 @@ typedef enum { } xfs_btnum_t; struct xfs_name { - const char *name; - int len; + const unsigned char *name; + int len; }; #endif /* __XFS_TYPES_H__ */ -- cgit v0.10.2 From 046ea753130fc51d885835458bf8c1d84765b9ac Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 20 Jan 2010 10:47:08 +1100 Subject: xfs: convert DM ops to use unsigned char names dmops uses a signed char for it's namespace event. To be consistent with the rest of the code, convert them to unsigned char for the namespace string. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index e62fd1c..f4d1441 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -78,7 +78,8 @@ typedef int (*xfs_send_destroy_t)(struct xfs_inode *, dm_right_t); typedef int (*xfs_send_namesp_t)(dm_eventtype_t, struct xfs_mount *, struct xfs_inode *, dm_right_t, struct xfs_inode *, dm_right_t, - const char *, const char *, mode_t, int, int); + const unsigned char *, const unsigned char *, + mode_t, int, int); typedef int (*xfs_send_mount_t)(struct xfs_mount *, dm_right_t, char *, char *); typedef void (*xfs_send_unmount_t)(struct xfs_mount *, struct xfs_inode *, diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 6f26875..9f7c001 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -2199,7 +2199,8 @@ xfs_symlink( if (DM_EVENT_ENABLED(dp, DM_EVENT_SYMLINK)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_SYMLINK, dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, - link_name->name, target_path, 0, 0, 0); + link_name->name, + (unsigned char *)target_path, 0, 0, 0); if (error) return error; } @@ -2395,7 +2396,8 @@ std_return: dp, DM_RIGHT_NULL, error ? NULL : ip, DM_RIGHT_NULL, link_name->name, - target_path, 0, error, 0); + (unsigned char *)target_path, + 0, error, 0); } if (!error) -- cgit v0.10.2 From 2bc754213d40d67c39ddd58cf240f2b948e1951e Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 20 Jan 2010 10:47:17 +1100 Subject: xfs: convert dirnameops to unsigned char names To be consistent across the codebase, convert the dirnameops to pass the directory names by unsigned char strings. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index c0c8869..0ca556b 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -1534,8 +1534,8 @@ xfs_da_hashname(const __uint8_t *name, int namelen) enum xfs_dacmp xfs_da_compname( struct xfs_da_args *args, - const char *name, - int len) + const unsigned char *name, + int len) { return (args->namelen == len && memcmp(args->name, name, len) == 0) ? XFS_CMP_EXACT : XFS_CMP_DIFFERENT; diff --git a/fs/xfs/xfs_da_btree.h b/fs/xfs/xfs_da_btree.h index 30cd08f..fe9f5a8 100644 --- a/fs/xfs/xfs_da_btree.h +++ b/fs/xfs/xfs_da_btree.h @@ -209,7 +209,8 @@ typedef struct xfs_da_state { */ struct xfs_nameops { xfs_dahash_t (*hashname)(struct xfs_name *); - enum xfs_dacmp (*compname)(struct xfs_da_args *, const char *, int); + enum xfs_dacmp (*compname)(struct xfs_da_args *, + const unsigned char *, int); }; @@ -260,7 +261,7 @@ int xfs_da_shrink_inode(xfs_da_args_t *args, xfs_dablk_t dead_blkno, uint xfs_da_hashname(const __uint8_t *name_string, int name_length); enum xfs_dacmp xfs_da_compname(struct xfs_da_args *args, - const char *name, int len); + const unsigned char *name, int len); xfs_da_state_t *xfs_da_state_alloc(void); diff --git a/fs/xfs/xfs_dir2.c b/fs/xfs/xfs_dir2.c index 93634a7..c21c527 100644 --- a/fs/xfs/xfs_dir2.c +++ b/fs/xfs/xfs_dir2.c @@ -66,8 +66,8 @@ xfs_ascii_ci_hashname( STATIC enum xfs_dacmp xfs_ascii_ci_compname( struct xfs_da_args *args, - const char *name, - int len) + const unsigned char *name, + int len) { enum xfs_dacmp result; int i; -- cgit v0.10.2 From a3380ae39fa321282c407ba5e1835e14b64853d9 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 20 Jan 2010 10:47:25 +1100 Subject: xfs: make xfs_dir_cilookup_result use unsigned char For consistency with the result of the code. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/xfs_dir2.c b/fs/xfs/xfs_dir2.c index c21c527..3a8c6ba 100644 --- a/fs/xfs/xfs_dir2.c +++ b/fs/xfs/xfs_dir2.c @@ -247,7 +247,7 @@ xfs_dir_createname( int xfs_dir_cilookup_result( struct xfs_da_args *args, - const char *name, + const unsigned char *name, int len) { if (args->cmpresult == XFS_CMP_DIFFERENT) diff --git a/fs/xfs/xfs_dir2.h b/fs/xfs/xfs_dir2.h index 1d9ef96..74a3b10 100644 --- a/fs/xfs/xfs_dir2.h +++ b/fs/xfs/xfs_dir2.h @@ -100,7 +100,7 @@ extern int xfs_dir2_isleaf(struct xfs_trans *tp, struct xfs_inode *dp, extern int xfs_dir2_shrink_inode(struct xfs_da_args *args, xfs_dir2_db_t db, struct xfs_dabuf *bp); -extern int xfs_dir_cilookup_result(struct xfs_da_args *args, const char *name, - int len); +extern int xfs_dir_cilookup_result(struct xfs_da_args *args, + const unsigned char *name, int len); #endif /* __XFS_DIR2_H__ */ -- cgit v0.10.2 From b9c48649577dfc4a8c263c106d518effa24ea54b Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 20 Jan 2010 10:47:39 +1100 Subject: xfs: xfs_buf_iomove() doesn't care about signedness xfs_buf_iomove() uses xfs_caddr_t as it's parameter types, but it doesn't care about the signedness of the variables as it is just copying the data. Change the prototype to use void * so that we don't get sign warnings at call sites. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index 158fad4..efd745b 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -1443,7 +1443,7 @@ xfs_buf_iomove( xfs_buf_t *bp, /* buffer to process */ size_t boff, /* starting buffer offset */ size_t bsize, /* length to copy */ - caddr_t data, /* data address */ + void *data, /* data address */ xfs_buf_rw_t mode) /* read/write/zero flag */ { size_t bend, cpoff, csize; diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h index 9a29d18..4f2ad66 100644 --- a/fs/xfs/linux-2.6/xfs_buf.h +++ b/fs/xfs/linux-2.6/xfs_buf.h @@ -243,7 +243,7 @@ extern void xfs_buf_ioend(xfs_buf_t *, int); extern void xfs_buf_ioerror(xfs_buf_t *, int); extern int xfs_buf_iorequest(xfs_buf_t *); extern int xfs_buf_iowait(xfs_buf_t *); -extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, xfs_caddr_t, +extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, void *, xfs_buf_rw_t); static inline int xfs_buf_iostrategy(xfs_buf_t *bp) -- cgit v0.10.2 From a9273ca5c6814f393e18ed66645f817b2b71e9ad Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 20 Jan 2010 10:47:48 +1100 Subject: xfs: convert attr to use unsigned names MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit To be consistent with the directory code, the attr code should use unsigned names. Convert the names from the vfs at the highest level to unsigned, and ænsure they are consistenly used as unsigned down to disk. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/linux-2.6/xfs_acl.c b/fs/xfs/linux-2.6/xfs_acl.c index 883ca5a..bf85bbe 100644 --- a/fs/xfs/linux-2.6/xfs_acl.c +++ b/fs/xfs/linux-2.6/xfs_acl.c @@ -106,7 +106,7 @@ xfs_get_acl(struct inode *inode, int type) struct posix_acl *acl; struct xfs_acl *xfs_acl; int len = sizeof(struct xfs_acl); - char *ea_name; + unsigned char *ea_name; int error; acl = get_cached_acl(inode, type); @@ -133,7 +133,8 @@ xfs_get_acl(struct inode *inode, int type) if (!xfs_acl) return ERR_PTR(-ENOMEM); - error = -xfs_attr_get(ip, ea_name, (char *)xfs_acl, &len, ATTR_ROOT); + error = -xfs_attr_get(ip, ea_name, (unsigned char *)xfs_acl, + &len, ATTR_ROOT); if (error) { /* * If the attribute doesn't exist make sure we have a negative @@ -162,7 +163,7 @@ STATIC int xfs_set_acl(struct inode *inode, int type, struct posix_acl *acl) { struct xfs_inode *ip = XFS_I(inode); - char *ea_name; + unsigned char *ea_name; int error; if (S_ISLNK(inode->i_mode)) @@ -194,7 +195,7 @@ xfs_set_acl(struct inode *inode, int type, struct posix_acl *acl) (sizeof(struct xfs_acl_entry) * (XFS_ACL_MAX_ENTRIES - acl->a_count)); - error = -xfs_attr_set(ip, ea_name, (char *)xfs_acl, + error = -xfs_attr_set(ip, ea_name, (unsigned char *)xfs_acl, len, ATTR_ROOT); kfree(xfs_acl); @@ -262,7 +263,7 @@ xfs_set_mode(struct inode *inode, mode_t mode) } static int -xfs_acl_exists(struct inode *inode, char *name) +xfs_acl_exists(struct inode *inode, unsigned char *name) { int len = sizeof(struct xfs_acl); diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c index a034cf6..3906e85 100644 --- a/fs/xfs/linux-2.6/xfs_ioctl.c +++ b/fs/xfs/linux-2.6/xfs_ioctl.c @@ -447,12 +447,12 @@ xfs_attrlist_by_handle( int xfs_attrmulti_attr_get( struct inode *inode, - char *name, - char __user *ubuf, + unsigned char *name, + unsigned char __user *ubuf, __uint32_t *len, __uint32_t flags) { - char *kbuf; + unsigned char *kbuf; int error = EFAULT; if (*len > XATTR_SIZE_MAX) @@ -476,12 +476,12 @@ xfs_attrmulti_attr_get( int xfs_attrmulti_attr_set( struct inode *inode, - char *name, - const char __user *ubuf, + unsigned char *name, + const unsigned char __user *ubuf, __uint32_t len, __uint32_t flags) { - char *kbuf; + unsigned char *kbuf; int error = EFAULT; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) @@ -501,7 +501,7 @@ xfs_attrmulti_attr_set( int xfs_attrmulti_attr_remove( struct inode *inode, - char *name, + unsigned char *name, __uint32_t flags) { if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) @@ -519,7 +519,7 @@ xfs_attrmulti_by_handle( xfs_fsop_attrmulti_handlereq_t am_hreq; struct dentry *dentry; unsigned int i, size; - char *attr_name; + unsigned char *attr_name; if (!capable(CAP_SYS_ADMIN)) return -XFS_ERROR(EPERM); @@ -547,7 +547,7 @@ xfs_attrmulti_by_handle( error = 0; for (i = 0; i < am_hreq.opcount; i++) { - ops[i].am_error = strncpy_from_user(attr_name, + ops[i].am_error = strncpy_from_user((char *)attr_name, ops[i].am_attrname, MAXNAMELEN); if (ops[i].am_error == 0 || ops[i].am_error == MAXNAMELEN) error = -ERANGE; diff --git a/fs/xfs/linux-2.6/xfs_ioctl.h b/fs/xfs/linux-2.6/xfs_ioctl.h index 7bd7c6a..d56173b 100644 --- a/fs/xfs/linux-2.6/xfs_ioctl.h +++ b/fs/xfs/linux-2.6/xfs_ioctl.h @@ -45,23 +45,23 @@ xfs_readlink_by_handle( extern int xfs_attrmulti_attr_get( struct inode *inode, - char *name, - char __user *ubuf, + unsigned char *name, + unsigned char __user *ubuf, __uint32_t *len, __uint32_t flags); extern int - xfs_attrmulti_attr_set( +xfs_attrmulti_attr_set( struct inode *inode, - char *name, - const char __user *ubuf, + unsigned char *name, + const unsigned char __user *ubuf, __uint32_t len, __uint32_t flags); extern int xfs_attrmulti_attr_remove( struct inode *inode, - char *name, + unsigned char *name, __uint32_t flags); extern struct dentry * diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.c b/fs/xfs/linux-2.6/xfs_ioctl32.c index be1527b..0bf6d61 100644 --- a/fs/xfs/linux-2.6/xfs_ioctl32.c +++ b/fs/xfs/linux-2.6/xfs_ioctl32.c @@ -411,7 +411,7 @@ xfs_compat_attrmulti_by_handle( compat_xfs_fsop_attrmulti_handlereq_t am_hreq; struct dentry *dentry; unsigned int i, size; - char *attr_name; + unsigned char *attr_name; if (!capable(CAP_SYS_ADMIN)) return -XFS_ERROR(EPERM); @@ -440,7 +440,7 @@ xfs_compat_attrmulti_by_handle( error = 0; for (i = 0; i < am_hreq.opcount; i++) { - ops[i].am_error = strncpy_from_user(attr_name, + ops[i].am_error = strncpy_from_user((char *)attr_name, compat_ptr(ops[i].am_attrname), MAXNAMELEN); if (ops[i].am_error == 0 || ops[i].am_error == MAXNAMELEN) diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 2259460..e8566bb 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -140,10 +140,10 @@ xfs_init_security( struct xfs_inode *ip = XFS_I(inode); size_t length; void *value; - char *name; + unsigned char *name; int error; - error = security_inode_init_security(inode, dir, &name, + error = security_inode_init_security(inode, dir, (char **)&name, &value, &length); if (error) { if (error == -EOPNOTSUPP) diff --git a/fs/xfs/linux-2.6/xfs_xattr.c b/fs/xfs/linux-2.6/xfs_xattr.c index 0b18788..fa01b9d 100644 --- a/fs/xfs/linux-2.6/xfs_xattr.c +++ b/fs/xfs/linux-2.6/xfs_xattr.c @@ -45,7 +45,7 @@ xfs_xattr_get(struct dentry *dentry, const char *name, value = NULL; } - error = -xfs_attr_get(ip, name, value, &asize, xflags); + error = -xfs_attr_get(ip, (unsigned char *)name, value, &asize, xflags); if (error) return error; return asize; @@ -67,8 +67,9 @@ xfs_xattr_set(struct dentry *dentry, const char *name, const void *value, xflags |= ATTR_REPLACE; if (!value) - return -xfs_attr_remove(ip, name, xflags); - return -xfs_attr_set(ip, name, (void *)value, size, xflags); + return -xfs_attr_remove(ip, (unsigned char *)name, xflags); + return -xfs_attr_set(ip, (unsigned char *)name, + (void *)value, size, xflags); } static struct xattr_handler xfs_xattr_user_handler = { @@ -124,8 +125,13 @@ static const char *xfs_xattr_prefix(int flags) } static int -xfs_xattr_put_listent(struct xfs_attr_list_context *context, int flags, - char *name, int namelen, int valuelen, char *value) +xfs_xattr_put_listent( + struct xfs_attr_list_context *context, + int flags, + unsigned char *name, + int namelen, + int valuelen, + unsigned char *value) { unsigned int prefix_len = xfs_xattr_prefix_len(flags); char *offset; @@ -148,7 +154,7 @@ xfs_xattr_put_listent(struct xfs_attr_list_context *context, int flags, offset = (char *)context->alist + context->count; strncpy(offset, xfs_xattr_prefix(flags), prefix_len); offset += prefix_len; - strncpy(offset, name, namelen); /* real name */ + strncpy(offset, (char *)name, namelen); /* real name */ offset += namelen; *offset = '\0'; context->count += prefix_len + namelen + 1; @@ -156,8 +162,13 @@ xfs_xattr_put_listent(struct xfs_attr_list_context *context, int flags, } static int -xfs_xattr_put_listent_sizes(struct xfs_attr_list_context *context, int flags, - char *name, int namelen, int valuelen, char *value) +xfs_xattr_put_listent_sizes( + struct xfs_attr_list_context *context, + int flags, + unsigned char *name, + int namelen, + int valuelen, + unsigned char *value) { context->count += xfs_xattr_prefix_len(flags) + namelen + 1; return 0; diff --git a/fs/xfs/xfs_acl.h b/fs/xfs/xfs_acl.h index 00fd357c..d13eeba 100644 --- a/fs/xfs/xfs_acl.h +++ b/fs/xfs/xfs_acl.h @@ -36,8 +36,8 @@ struct xfs_acl { }; /* On-disk XFS extended attribute names */ -#define SGI_ACL_FILE "SGI_ACL_FILE" -#define SGI_ACL_DEFAULT "SGI_ACL_DEFAULT" +#define SGI_ACL_FILE (unsigned char *)"SGI_ACL_FILE" +#define SGI_ACL_DEFAULT (unsigned char *)"SGI_ACL_DEFAULT" #define SGI_ACL_FILE_SIZE (sizeof(SGI_ACL_FILE)-1) #define SGI_ACL_DEFAULT_SIZE (sizeof(SGI_ACL_DEFAULT)-1) diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index 9d11eba..f7b426a 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -93,12 +93,12 @@ STATIC int xfs_attr_rmtval_remove(xfs_da_args_t *args); STATIC int xfs_attr_name_to_xname( struct xfs_name *xname, - const char *aname) + const unsigned char *aname) { if (!aname) return EINVAL; xname->name = aname; - xname->len = strlen(aname); + xname->len = strlen((char *)aname); if (xname->len >= MAXNAMELEN) return EFAULT; /* match IRIX behaviour */ @@ -124,7 +124,7 @@ STATIC int xfs_attr_get_int( struct xfs_inode *ip, struct xfs_name *name, - char *value, + unsigned char *value, int *valuelenp, int flags) { @@ -171,8 +171,8 @@ xfs_attr_get_int( int xfs_attr_get( xfs_inode_t *ip, - const char *name, - char *value, + const unsigned char *name, + unsigned char *value, int *valuelenp, int flags) { @@ -235,8 +235,12 @@ xfs_attr_calc_size( } STATIC int -xfs_attr_set_int(xfs_inode_t *dp, struct xfs_name *name, - char *value, int valuelen, int flags) +xfs_attr_set_int( + struct xfs_inode *dp, + struct xfs_name *name, + unsigned char *value, + int valuelen, + int flags) { xfs_da_args_t args; xfs_fsblock_t firstblock; @@ -452,8 +456,8 @@ out: int xfs_attr_set( xfs_inode_t *dp, - const char *name, - char *value, + const unsigned char *name, + unsigned char *value, int valuelen, int flags) { @@ -600,7 +604,7 @@ out: int xfs_attr_remove( xfs_inode_t *dp, - const char *name, + const unsigned char *name, int flags) { int error; @@ -669,9 +673,13 @@ xfs_attr_list_int(xfs_attr_list_context_t *context) */ /*ARGSUSED*/ STATIC int -xfs_attr_put_listent(xfs_attr_list_context_t *context, int flags, - char *name, int namelen, - int valuelen, char *value) +xfs_attr_put_listent( + xfs_attr_list_context_t *context, + int flags, + unsigned char *name, + int namelen, + int valuelen, + unsigned char *value) { struct attrlist *alist = (struct attrlist *)context->alist; attrlist_ent_t *aep; @@ -1980,7 +1988,7 @@ xfs_attr_rmtval_get(xfs_da_args_t *args) xfs_bmbt_irec_t map[ATTR_RMTVALUE_MAPSIZE]; xfs_mount_t *mp; xfs_daddr_t dblkno; - xfs_caddr_t dst; + void *dst; xfs_buf_t *bp; int nmap, error, tmp, valuelen, blkcnt, i; xfs_dablk_t lblkno; @@ -2039,7 +2047,7 @@ xfs_attr_rmtval_set(xfs_da_args_t *args) xfs_inode_t *dp; xfs_bmbt_irec_t map; xfs_daddr_t dblkno; - xfs_caddr_t src; + void *src; xfs_buf_t *bp; xfs_dablk_t lblkno; int blkcnt, valuelen, nmap, error, tmp, committed; diff --git a/fs/xfs/xfs_attr.h b/fs/xfs/xfs_attr.h index 9c3a243..e920d68 100644 --- a/fs/xfs/xfs_attr.h +++ b/fs/xfs/xfs_attr.h @@ -113,7 +113,7 @@ typedef struct attrlist_cursor_kern { typedef int (*put_listent_func_t)(struct xfs_attr_list_context *, int, - char *, int, int, char *); + unsigned char *, int, int, unsigned char *); typedef struct xfs_attr_list_context { struct xfs_inode *dp; /* inode */ diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index baf41b5..52519a2 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -521,11 +521,11 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args) sfe = &sf->list[0]; for (i = 0; i < sf->hdr.count; i++) { - nargs.name = (char *)sfe->nameval; + nargs.name = sfe->nameval; nargs.namelen = sfe->namelen; - nargs.value = (char *)&sfe->nameval[nargs.namelen]; + nargs.value = &sfe->nameval[nargs.namelen]; nargs.valuelen = sfe->valuelen; - nargs.hashval = xfs_da_hashname((char *)sfe->nameval, + nargs.hashval = xfs_da_hashname(sfe->nameval, sfe->namelen); nargs.flags = XFS_ATTR_NSP_ONDISK_TO_ARGS(sfe->flags); error = xfs_attr_leaf_lookup_int(bp, &nargs); /* set a->index */ @@ -612,10 +612,10 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context) for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) { error = context->put_listent(context, sfe->flags, - (char *)sfe->nameval, + sfe->nameval, (int)sfe->namelen, (int)sfe->valuelen, - (char*)&sfe->nameval[sfe->namelen]); + &sfe->nameval[sfe->namelen]); /* * Either search callback finished early or @@ -659,8 +659,8 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context) } sbp->entno = i; - sbp->hash = xfs_da_hashname((char *)sfe->nameval, sfe->namelen); - sbp->name = (char *)sfe->nameval; + sbp->hash = xfs_da_hashname(sfe->nameval, sfe->namelen); + sbp->name = sfe->nameval; sbp->namelen = sfe->namelen; /* These are bytes, and both on-disk, don't endian-flip */ sbp->valuelen = sfe->valuelen; @@ -818,9 +818,9 @@ xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff) continue; ASSERT(entry->flags & XFS_ATTR_LOCAL); name_loc = xfs_attr_leaf_name_local(leaf, i); - nargs.name = (char *)name_loc->nameval; + nargs.name = name_loc->nameval; nargs.namelen = name_loc->namelen; - nargs.value = (char *)&name_loc->nameval[nargs.namelen]; + nargs.value = &name_loc->nameval[nargs.namelen]; nargs.valuelen = be16_to_cpu(name_loc->valuelen); nargs.hashval = be32_to_cpu(entry->hashval); nargs.flags = XFS_ATTR_NSP_ONDISK_TO_ARGS(entry->flags); @@ -2370,10 +2370,10 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context) retval = context->put_listent(context, entry->flags, - (char *)name_loc->nameval, + name_loc->nameval, (int)name_loc->namelen, be16_to_cpu(name_loc->valuelen), - (char *)&name_loc->nameval[name_loc->namelen]); + &name_loc->nameval[name_loc->namelen]); if (retval) return retval; } else { @@ -2397,15 +2397,15 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context) return retval; retval = context->put_listent(context, entry->flags, - (char *)name_rmt->name, + name_rmt->name, (int)name_rmt->namelen, valuelen, - (char*)args.value); + args.value); kmem_free(args.value); } else { retval = context->put_listent(context, entry->flags, - (char *)name_rmt->name, + name_rmt->name, (int)name_rmt->namelen, valuelen, NULL); diff --git a/fs/xfs/xfs_attr_sf.h b/fs/xfs/xfs_attr_sf.h index 76ab7b0..919756e 100644 --- a/fs/xfs/xfs_attr_sf.h +++ b/fs/xfs/xfs_attr_sf.h @@ -52,7 +52,7 @@ typedef struct xfs_attr_sf_sort { __uint8_t valuelen; /* length of value */ __uint8_t flags; /* flags bits (see xfs_attr_leaf.h) */ xfs_dahash_t hash; /* this entry's hash value */ - char *name; /* name value, pointer into buffer */ + unsigned char *name; /* name value, pointer into buffer */ } xfs_attr_sf_sort_t; #define XFS_ATTR_SF_ENTSIZE_BYNAME(nlen,vlen) /* space name/value uses */ \ diff --git a/fs/xfs/xfs_vnodeops.h b/fs/xfs/xfs_vnodeops.h index 167a467..774f407 100644 --- a/fs/xfs/xfs_vnodeops.h +++ b/fs/xfs/xfs_vnodeops.h @@ -43,11 +43,11 @@ int xfs_change_file_space(struct xfs_inode *ip, int cmd, int xfs_rename(struct xfs_inode *src_dp, struct xfs_name *src_name, struct xfs_inode *src_ip, struct xfs_inode *target_dp, struct xfs_name *target_name, struct xfs_inode *target_ip); -int xfs_attr_get(struct xfs_inode *ip, const char *name, char *value, - int *valuelenp, int flags); -int xfs_attr_set(struct xfs_inode *dp, const char *name, char *value, - int valuelen, int flags); -int xfs_attr_remove(struct xfs_inode *dp, const char *name, int flags); +int xfs_attr_get(struct xfs_inode *ip, const unsigned char *name, + unsigned char *value, int *valuelenp, int flags); +int xfs_attr_set(struct xfs_inode *dp, const unsigned char *name, + unsigned char *value, int valuelen, int flags); +int xfs_attr_remove(struct xfs_inode *dp, const unsigned char *name, int flags); int xfs_attr_list(struct xfs_inode *dp, char *buffer, int bufsize, int flags, struct attrlist_cursor_kern *cursor); ssize_t xfs_read(struct xfs_inode *ip, struct kiocb *iocb, -- cgit v0.10.2 From 4a24cb71407dc25035d75dd3d118e0e55679e217 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 20 Jan 2010 10:48:05 +1100 Subject: xfs: clean up sign warnings in dir2 code We are now consistently using unsigned char strings for names so fix up the remaining warnings in the dir2 code to complete the cleanup. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/xfs_dir2.c b/fs/xfs/xfs_dir2.c index 3a8c6ba..42520f0 100644 --- a/fs/xfs/xfs_dir2.c +++ b/fs/xfs/xfs_dir2.c @@ -44,7 +44,7 @@ #include "xfs_vnodeops.h" #include "xfs_trace.h" -struct xfs_name xfs_name_dotdot = {"..", 2}; +struct xfs_name xfs_name_dotdot = { (unsigned char *)"..", 2}; /* * ASCII case-insensitive (ie. A-Z) support for directories that was diff --git a/fs/xfs/xfs_dir2_block.c b/fs/xfs/xfs_dir2_block.c index ddc4ecc..779a267 100644 --- a/fs/xfs/xfs_dir2_block.c +++ b/fs/xfs/xfs_dir2_block.c @@ -57,8 +57,8 @@ static xfs_dahash_t xfs_dir_hash_dot, xfs_dir_hash_dotdot; void xfs_dir_startup(void) { - xfs_dir_hash_dot = xfs_da_hashname(".", 1); - xfs_dir_hash_dotdot = xfs_da_hashname("..", 2); + xfs_dir_hash_dot = xfs_da_hashname((unsigned char *)".", 1); + xfs_dir_hash_dotdot = xfs_da_hashname((unsigned char *)"..", 2); } /* @@ -513,8 +513,9 @@ xfs_dir2_block_getdents( /* * If it didn't fit, set the final offset to here & return. */ - if (filldir(dirent, dep->name, dep->namelen, cook & 0x7fffffff, - be64_to_cpu(dep->inumber), DT_UNKNOWN)) { + if (filldir(dirent, (char *)dep->name, dep->namelen, + cook & 0x7fffffff, be64_to_cpu(dep->inumber), + DT_UNKNOWN)) { *offset = cook & 0x7fffffff; xfs_da_brelse(NULL, bp); return 0; diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c index 29f484c..e2d8985 100644 --- a/fs/xfs/xfs_dir2_leaf.c +++ b/fs/xfs/xfs_dir2_leaf.c @@ -1081,7 +1081,7 @@ xfs_dir2_leaf_getdents( dep = (xfs_dir2_data_entry_t *)ptr; length = xfs_dir2_data_entsize(dep->namelen); - if (filldir(dirent, dep->name, dep->namelen, + if (filldir(dirent, (char *)dep->name, dep->namelen, xfs_dir2_byte_to_dataptr(mp, curoff) & 0x7fffffff, be64_to_cpu(dep->inumber), DT_UNKNOWN)) break; diff --git a/fs/xfs/xfs_dir2_sf.c b/fs/xfs/xfs_dir2_sf.c index 9d4f17a..c1a5945 100644 --- a/fs/xfs/xfs_dir2_sf.c +++ b/fs/xfs/xfs_dir2_sf.c @@ -782,7 +782,7 @@ xfs_dir2_sf_getdents( } ino = xfs_dir2_sf_get_inumber(sfp, xfs_dir2_sf_inumberp(sfep)); - if (filldir(dirent, sfep->name, sfep->namelen, + if (filldir(dirent, (char *)sfep->name, sfep->namelen, off & 0x7fffffff, ino, DT_UNKNOWN)) { *offset = off & 0x7fffffff; return 0; -- cgit v0.10.2 From 58c75cfb51393a52b45262394c1fa81514b4d9bd Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 20 Jan 2010 10:49:18 +1100 Subject: xfs: make compile warn about char sign mismatches again The -fno-unsigned-char directive has no effect anymore as the XFs build is clean. However, the kernel build hides pointer sign differences so turn that back on so that we can clean up all the mismatches prior to a userspace code resync. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index 56641fe..1926701 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -16,7 +16,7 @@ # Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA # -EXTRA_CFLAGS += -I$(src) -I$(src)/linux-2.6 -funsigned-char +EXTRA_CFLAGS += -I$(src) -I$(src)/linux-2.6 -Wpointer-sign XFS_LINUX := linux-2.6 -- cgit v0.10.2 From f0a0eaa8da08ebc6519cacd731df05bbb4ca47ce Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 20 Jan 2010 10:50:06 +1100 Subject: xfs: suppress spurious uninitialised var warning in xfs_bmapi() Initialise the xfs_bmalloca_t structure to zero to avoid uninitialised variable warnings. This is done by zeroing the arg structure rather than using the uninitialised_var() trick so we know for certain that the structure is correctly initialised as xfs_bmapi is a very complex function and it is difficult to prove warnings are spurious. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 7c6d9ac..1869fb9 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4471,7 +4471,7 @@ xfs_bmapi( xfs_fsblock_t abno; /* allocated block number */ xfs_extlen_t alen; /* allocated extent length */ xfs_fileoff_t aoff; /* allocated file offset */ - xfs_bmalloca_t bma; /* args for xfs_bmap_alloc */ + xfs_bmalloca_t bma = { 0 }; /* args for xfs_bmap_alloc */ xfs_btree_cur_t *cur; /* bmap btree cursor */ xfs_fileoff_t end; /* end of mapped file region */ int eof; /* we've hit the end of extents */ -- cgit v0.10.2 From 587aa0feb74ffe3239b5e26ff5d017ba9f5daec9 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Wed, 20 Jan 2010 12:04:53 +1100 Subject: xfs: rearrange xfs_mod_sb() to avoid array subscript warning gcc warns of an array subscript out of bounds in xfs_mod_sb(). The code is written in such a way that if the array subscript is out of bounds, then it will assert fail. Rearrange the code to avoid the bounds check warning. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 0df5045..d95bd18 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1631,15 +1631,14 @@ xfs_mod_sb(xfs_trans_t *tp, __int64_t fields) xfs_sb_to_disk(XFS_BUF_TO_SBP(bp), &mp->m_sb, fields); /* find modified range */ + f = (xfs_sb_field_t)xfs_highbit64((__uint64_t)fields); + ASSERT((1LL << f) & XFS_SB_MOD_BITS); + last = xfs_sb_info[f + 1].offset - 1; f = (xfs_sb_field_t)xfs_lowbit64((__uint64_t)fields); ASSERT((1LL << f) & XFS_SB_MOD_BITS); first = xfs_sb_info[f].offset; - f = (xfs_sb_field_t)xfs_highbit64((__uint64_t)fields); - ASSERT((1LL << f) & XFS_SB_MOD_BITS); - last = xfs_sb_info[f + 1].offset - 1; - xfs_trans_log_buf(tp, bp, first, last); } -- cgit v0.10.2 From efd590d57a5edddabaaf4cfaf532c0f674193b81 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 15:08:36 +0900 Subject: sh: mach-sdk7786: FPGA updates. This does a bit of refactoring of the FPGA management code. The primary FPGA initialization is moved out to its own file in preparation for implementing some of the more complex capabilities, a complete set of register definitions is provided, and all of the existing users in the board code are moved over to use the new interface instead of setting up overlapping mappings. This also corrects the FPGA size, which previously was chomped off at the SDIF control register. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-sdk7786/Makefile b/arch/sh/boards/mach-sdk7786/Makefile index f663768..50c8065 100644 --- a/arch/sh/boards/mach-sdk7786/Makefile +++ b/arch/sh/boards/mach-sdk7786/Makefile @@ -1 +1 @@ -obj-y := setup.o +obj-y := setup.o fpga.o diff --git a/arch/sh/boards/mach-sdk7786/fpga.c b/arch/sh/boards/mach-sdk7786/fpga.c new file mode 100644 index 0000000..99f903c --- /dev/null +++ b/arch/sh/boards/mach-sdk7786/fpga.c @@ -0,0 +1,37 @@ +/* + * SDK7786 FPGA Support. + * + * Copyright (C) 2010 Paul Mundt + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ +#include +#include +#include +#include + +#define FPGA_REGS_BASE 0x07fff800 +#define FPGA_REGS_SIZE 0x490 + +void __iomem *sdk7786_fpga_base; + +void __init sdk7786_fpga_init(void) +{ + u16 version, date; + + sdk7786_fpga_base = ioremap_nocache(FPGA_REGS_BASE, FPGA_REGS_SIZE); + if (unlikely(!sdk7786_fpga_base)) { + panic("FPGA remapping failed.\n"); + return; + } + + version = fpga_read_reg(FPGAVR); + date = fpga_read_reg(FPGADR); + + pr_info("\tFPGA version:\t%d.%d (built on %d/%d/%d)\n", + bcd2bin(version >> 8) & 0xf, bcd2bin(version & 0xf), + ((date >> 12) & 0xf) + 2000, + (date >> 8) & 0xf, bcd2bin(date & 0xff)); +} diff --git a/arch/sh/boards/mach-sdk7786/setup.c b/arch/sh/boards/mach-sdk7786/setup.c index 0e4b1c3..8dbbdea 100644 --- a/arch/sh/boards/mach-sdk7786/setup.c +++ b/arch/sh/boards/mach-sdk7786/setup.c @@ -18,6 +18,7 @@ #include #include #include +#include static struct resource heartbeat_resource = { .start = 0x07fff8b0, @@ -103,27 +104,17 @@ static struct platform_device *sh7786_devices[] __initdata = { &smbus_pcie_device, }; -#define SBCR_REGS_BASE 0x07fff990 - -#define SCBR_I2CMEN (1 << 0) /* FPGA I2C master enable */ -#define SCBR_I2CCEN (1 << 1) /* CPU I2C master enable */ - static int sdk7786_i2c_setup(void) { - void __iomem *sbcr; unsigned int tmp; - sbcr = ioremap_nocache(SBCR_REGS_BASE, SZ_16); - /* * Hand over I2C control to the FPGA. */ - tmp = ioread16(sbcr); + tmp = fpga_read_reg(SBCR); tmp &= ~SCBR_I2CCEN; tmp |= SCBR_I2CMEN; - iowrite16(tmp, sbcr); - - iounmap(sbcr); + fpga_write_reg(tmp, SBCR); return i2c_register_board_info(0, sdk7786_i2c_devices, ARRAY_SIZE(sdk7786_i2c_devices)); @@ -141,43 +132,6 @@ static int __init sdk7786_devices_setup(void) } __initcall(sdk7786_devices_setup); -#define FPGA_REGS_BASE 0x07fff800 -#define FPGA_REGS_SIZE 1152 - -#define INTASR 0x010 -#define INTAMR 0x020 -#define INTBSR 0x090 -#define INTBMR 0x0a0 -#define INTMSR 0x130 - -#define IASELR1 0x210 -#define IASELR2 0x220 -#define IASELR3 0x230 -#define IASELR4 0x240 -#define IASELR5 0x250 -#define IASELR6 0x260 -#define IASELR7 0x270 -#define IASELR8 0x280 -#define IASELR9 0x290 -#define IASELR10 0x2a0 -#define IASELR11 0x2b0 -#define IASELR12 0x2c0 -#define IASELR13 0x2d0 -#define IASELR14 0x2e0 -#define IASELR15 0x2f0 - -static void __iomem *fpga_regs; - -static u16 fpga_read_reg(unsigned int reg) -{ - return __raw_readw(fpga_regs + reg); -} - -static void fpga_write_reg(u16 val, unsigned int reg) -{ - __raw_writew(val, fpga_regs + reg); -} - enum { ATA_IRQ_BIT = 1, SPI_BUSY_BIT = 2, @@ -197,12 +151,6 @@ static void __init init_sdk7786_IRQ(void) { unsigned int tmp; - fpga_regs = ioremap_nocache(FPGA_REGS_BASE, FPGA_REGS_SIZE); - if (!fpga_regs) { - printk(KERN_ERR "Couldn't map FPGA registers\n"); - return; - } - /* Enable priority encoding for all IRLs */ fpga_write_reg(fpga_read_reg(INTMSR) | 0x0303, INTMSR); @@ -219,21 +167,9 @@ static void __init init_sdk7786_IRQ(void) plat_irq_setup_pins(IRQ_MODE_IRL3210_MASK); } -#define MODSWR_REGS 0x07fff830 - static int sdk7786_mode_pins(void) { - void __iomem *modswr; - int pin_states; - - modswr = ioremap_nocache(MODSWR_REGS, SZ_16); - if (!modswr) - return -ENXIO; - - pin_states = ioread16(modswr); - iounmap(modswr); - - return pin_states; + return fpga_read_reg(MODSWR); } static int sdk7786_clk_init(void) @@ -260,7 +196,11 @@ static int sdk7786_clk_init(void) /* Initialize the board */ static void __init sdk7786_setup(char **cmdline_p) { - printk(KERN_INFO "Renesas Technology Corp. SDK7786 support.\n"); + pr_info("Renesas Technology Europe SDK7786 support:\n"); + + sdk7786_fpga_init(); + + pr_info("\tPCB revision:\t%d\n", fpga_read_reg(PCBRR) & 0xf); } /* diff --git a/arch/sh/include/mach-sdk7786/mach/fpga.h b/arch/sh/include/mach-sdk7786/mach/fpga.h new file mode 100644 index 0000000..a85d985 --- /dev/null +++ b/arch/sh/include/mach-sdk7786/mach/fpga.h @@ -0,0 +1,112 @@ +#ifndef __MACH_SDK7786_FPGA_H +#define __MACH_SDK7786_FPGA_H + +#include +#include +#include + +#define SRSTR 0x000 +#define INTASR 0x010 +#define INTAMR 0x020 +#define MODSWR 0x030 +#define INTTESTR 0x040 +#define SYSSR 0x050 +#define NRGPR 0x060 +#define NMISR 0x070 + +#define NMIMR 0x080 +#define NMIMR_MAN_NMIM BIT(0) /* Manual NMI mask */ +#define NMIMR_AUX_NMIM BIT(1) /* Auxiliary NMI mask */ + +#define INTBSR 0x090 +#define INTBMR 0x0a0 +#define USRLEDR 0x0b0 +#define MAPSWR 0x0c0 +#define FPGAVR 0x0d0 +#define FPGADR 0x0e0 +#define PCBRR 0x0f0 +#define RSR 0x100 +#define EXTASR 0x110 +#define SPCAR 0x120 +#define INTMSR 0x130 +#define PCIECR 0x140 +#define FAER 0x150 +#define USRGPIR 0x160 +/* 0x170 reserved */ +#define LCLASR 0x180 + +#define SBCR 0x190 +#define SCBR_I2CMEN BIT(0) /* FPGA I2C master enable */ +#define SCBR_I2CCEN BIT(1) /* CPU I2C master enable */ + +#define PWRCR 0x1a0 +#define SPCBR 0x1b0 +#define SPICR 0x1c0 +#define SPIDR 0x1d0 +#define I2CCR 0x1e0 +#define I2CDR 0x1f0 +#define FPGACR 0x200 +#define IASELR1 0x210 +#define IASELR2 0x220 +#define IASELR3 0x230 +#define IASELR4 0x240 +#define IASELR5 0x250 +#define IASELR6 0x260 +#define IASELR7 0x270 +#define IASELR8 0x280 +#define IASELR9 0x290 +#define IASELR10 0x2a0 +#define IASELR11 0x2b0 +#define IASELR12 0x2c0 +#define IASELR13 0x2d0 +#define IASELR14 0x2e0 +#define IASELR15 0x2f0 +/* 0x300 reserved */ +#define IBSELR1 0x310 +#define IBSELR2 0x320 +#define IBSELR3 0x330 +#define IBSELR4 0x340 +#define IBSELR5 0x350 +#define IBSELR6 0x360 +#define IBSELR7 0x370 +#define IBSELR8 0x380 +#define IBSELR9 0x390 +#define IBSELR10 0x3a0 +#define IBSELR11 0x3b0 +#define IBSELR12 0x3c0 +#define IBSELR13 0x3d0 +#define IBSELR14 0x3e0 +#define IBSELR15 0x3f0 +#define USRACR 0x400 +#define BEEPR 0x410 +#define USRLCDR 0x420 +#define SMBCR 0x430 +#define SMBDR 0x440 +#define USBCR 0x450 +#define AMSR 0x460 +#define ACCR 0x470 +#define SDIFCR 0x480 + +/* arch/sh/boards/mach-sdk7786/fpga.c */ +extern void __iomem *sdk7786_fpga_base; +extern void sdk7786_fpga_init(void); + +#define SDK7786_FPGA_REGADDR(reg) (sdk7786_fpga_base + (reg)) + +/* + * A convenience wrapper from register offset to internal I2C address, + * when the FPGA is in I2C slave mode. + */ +#define SDK7786_FPGA_I2CADDR(reg) ((reg) >> 3) + +static inline u16 fpga_read_reg(unsigned int reg) +{ + return ioread16(sdk7786_fpga_base + reg); +} + +static inline void fpga_write_reg(u16 val, unsigned int reg) +{ + iowrite16(val, sdk7786_fpga_base + reg); +} + +#endif /* __MACH_SDK7786_FPGA_H */ -- cgit v0.10.2 From 580d9e00fdfb85e65c5097dcd739c6efcdbadc96 Mon Sep 17 00:00:00 2001 From: Motohiro KOSAKI Date: Mon, 18 Jan 2010 21:35:05 -0500 Subject: kprobetrace, doc: Shell needs single quote to use $ character Shell interprets $val as shell variable, thus we need quote if we use the echo command. Signed-off-by: KOSAKI Motohiro Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE LKML-Reference: <20100119023505.31880.17367.stgit@localhost6.localdomain6> Signed-off-by: Ingo Molnar diff --git a/Documentation/trace/kprobetrace.txt b/Documentation/trace/kprobetrace.txt index f30978e..ab57f02 100644 --- a/Documentation/trace/kprobetrace.txt +++ b/Documentation/trace/kprobetrace.txt @@ -79,7 +79,7 @@ Usage examples To add a probe as a new event, write a new definition to kprobe_events as below. - echo p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack) > /sys/kernel/debug/tracing/kprobe_events + echo 'p:myprobe do_sys_open dfd=%ax filename=%dx flags=%cx mode=+4($stack)' > /sys/kernel/debug/tracing/kprobe_events This sets a kprobe on the top of do_sys_open() function with recording 1st to 4th arguments as "myprobe" event. Note, which register/stack entry is @@ -88,7 +88,7 @@ the ABI, please try to use probe subcommand of perf-tools (you can find it under tools/perf/). As this example shows, users can choose more familiar names for each arguments. - echo r:myretprobe do_sys_open $retval >> /sys/kernel/debug/tracing/kprobe_events + echo 'r:myretprobe do_sys_open $retval' >> /sys/kernel/debug/tracing/kprobe_events This sets a kretprobe on the return point of do_sys_open() function with recording return value as "myretprobe" event. -- cgit v0.10.2 From df3ab708b787a2b35de5101452bd51d4a8ae0ded Mon Sep 17 00:00:00 2001 From: Motohiro KOSAKI Date: Mon, 18 Jan 2010 21:35:12 -0500 Subject: kprobetrace, doc: Add the explanation to remove probe points Latest kprobetrace can remove probe points selectively, thus the documentation should be updated too. Signed-off-by: KOSAKI Motohiro Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE LKML-Reference: <20100119023512.31880.35535.stgit@localhost6.localdomain6> Signed-off-by: Ingo Molnar diff --git a/Documentation/trace/kprobetrace.txt b/Documentation/trace/kprobetrace.txt index ab57f02..a9100b2 100644 --- a/Documentation/trace/kprobetrace.txt +++ b/Documentation/trace/kprobetrace.txt @@ -24,6 +24,7 @@ Synopsis of kprobe_events ------------------------- p[:[GRP/]EVENT] SYMBOL[+offs]|MEMADDR [FETCHARGS] : Set a probe r[:[GRP/]EVENT] SYMBOL[+0] [FETCHARGS] : Set a return probe + -:[GRP/]EVENT : Clear a probe GRP : Group name. If omitted, use "kprobes" for it. EVENT : Event name. If omitted, the event name is generated @@ -122,6 +123,12 @@ REC->dfd, REC->filename, REC->flags, REC->mode This clears all probe points. + Or, + + echo -:myprobe >> kprobe_events + + This clears probe points selectively. + Right after definition, each event is disabled by default. For tracing these events, you need to enable it. -- cgit v0.10.2 From 5f240718b475dcddee86b138b624f24c681dbb29 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 15:23:54 +0900 Subject: sh: mach-sdk7786: Split out FPGA IRQ controller setup. This moves out the FPGA IRQ controller setup code to its own file, in preparation for switching off of IRL mode and having it provide its own irq_chip. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-sdk7786/Makefile b/arch/sh/boards/mach-sdk7786/Makefile index 50c8065..a29f19e 100644 --- a/arch/sh/boards/mach-sdk7786/Makefile +++ b/arch/sh/boards/mach-sdk7786/Makefile @@ -1 +1 @@ -obj-y := setup.o fpga.o +obj-y := setup.o fpga.o irq.o diff --git a/arch/sh/boards/mach-sdk7786/irq.c b/arch/sh/boards/mach-sdk7786/irq.c new file mode 100644 index 0000000..46943a0 --- /dev/null +++ b/arch/sh/boards/mach-sdk7786/irq.c @@ -0,0 +1,48 @@ +/* + * SDK7786 FPGA IRQ Controller Support. + * + * Copyright (C) 2010 Matt Fleming + * Copyright (C) 2010 Paul Mundt + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ +#include +#include +#include + +enum { + ATA_IRQ_BIT = 1, + SPI_BUSY_BIT = 2, + LIRQ5_BIT = 3, + LIRQ6_BIT = 4, + LIRQ7_BIT = 5, + LIRQ8_BIT = 6, + KEY_IRQ_BIT = 7, + PEN_IRQ_BIT = 8, + ETH_IRQ_BIT = 9, + RTC_ALARM_BIT = 10, + CRYSTAL_FAIL_BIT = 12, + ETH_PME_BIT = 14, +}; + +void __init sdk7786_init_irq(void) +{ + unsigned int tmp; + + /* Enable priority encoding for all IRLs */ + fpga_write_reg(fpga_read_reg(INTMSR) | 0x0303, INTMSR); + + /* Clear FPGA interrupt status registers */ + fpga_write_reg(0x0000, INTASR); + fpga_write_reg(0x0000, INTBSR); + + /* Unmask FPGA interrupts */ + tmp = fpga_read_reg(INTAMR); + tmp &= ~(1 << ETH_IRQ_BIT); + fpga_write_reg(tmp, INTAMR); + + plat_irq_setup_pins(IRQ_MODE_IRL7654_MASK); + plat_irq_setup_pins(IRQ_MODE_IRL3210_MASK); +} diff --git a/arch/sh/boards/mach-sdk7786/setup.c b/arch/sh/boards/mach-sdk7786/setup.c index 8dbbdea..3c3c9a3 100644 --- a/arch/sh/boards/mach-sdk7786/setup.c +++ b/arch/sh/boards/mach-sdk7786/setup.c @@ -15,10 +15,11 @@ #include #include #include +#include +#include #include #include #include -#include static struct resource heartbeat_resource = { .start = 0x07fff8b0, @@ -132,41 +133,6 @@ static int __init sdk7786_devices_setup(void) } __initcall(sdk7786_devices_setup); -enum { - ATA_IRQ_BIT = 1, - SPI_BUSY_BIT = 2, - LIRQ5_BIT = 3, - LIRQ6_BIT = 4, - LIRQ7_BIT = 5, - LIRQ8_BIT = 6, - KEY_IRQ_BIT = 7, - PEN_IRQ_BIT = 8, - ETH_IRQ_BIT = 9, - RTC_ALARM_BIT = 10, - CRYSTAL_FAIL_BIT = 12, - ETH_PME_BIT = 14, -}; - -static void __init init_sdk7786_IRQ(void) -{ - unsigned int tmp; - - /* Enable priority encoding for all IRLs */ - fpga_write_reg(fpga_read_reg(INTMSR) | 0x0303, INTMSR); - - /* Clear FPGA interrupt status registers */ - fpga_write_reg(0x0000, INTASR); - fpga_write_reg(0x0000, INTBSR); - - /* Unmask FPGA interrupts */ - tmp = fpga_read_reg(INTAMR); - tmp &= ~(1 << ETH_IRQ_BIT); - fpga_write_reg(tmp, INTAMR); - - plat_irq_setup_pins(IRQ_MODE_IRL7654_MASK); - plat_irq_setup_pins(IRQ_MODE_IRL3210_MASK); -} - static int sdk7786_mode_pins(void) { return fpga_read_reg(MODSWR); @@ -211,5 +177,5 @@ static struct sh_machine_vector mv_sdk7786 __initmv = { .mv_setup = sdk7786_setup, .mv_mode_pins = sdk7786_mode_pins, .mv_clk_init = sdk7786_clk_init, - .mv_init_irq = init_sdk7786_IRQ, + .mv_init_irq = sdk7786_init_irq, }; diff --git a/arch/sh/include/mach-sdk7786/mach/irq.h b/arch/sh/include/mach-sdk7786/mach/irq.h new file mode 100644 index 0000000..0f58463 --- /dev/null +++ b/arch/sh/include/mach-sdk7786/mach/irq.h @@ -0,0 +1,7 @@ +#ifndef __MACH_SDK7786_IRQ_H +#define __MACH_SDK7786_IRQ_H + +/* arch/sh/boards/mach-sdk7786/irq.c */ +extern void sdk7786_init_irq(void); + +#endif /* __MACH_SDK7786_IRQ_H */ -- cgit v0.10.2 From 2efa53b269ec1e9289a108e1506f53f6f1de440b Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 16:40:48 +0900 Subject: sh: Make 29/32-bit mode check helper generally available. Presently __in_29bit_mode() is only defined for the PMB case, but it's also easily derived from the CONFIG_29BIT and CONFIG_32BIT && CONFIG_PMB=n cases. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/addrspace.h b/arch/sh/include/asm/addrspace.h index 268efd6..446b383 100644 --- a/arch/sh/include/asm/addrspace.h +++ b/arch/sh/include/asm/addrspace.h @@ -65,11 +65,5 @@ #define P3_ADDR_MAX P4SEG #endif -#ifndef __ASSEMBLY__ -#ifdef CONFIG_PMB -extern int __in_29bit_mode(void); -#endif /* CONFIG_PMB */ -#endif /* __ASSEMBLY__ */ - #endif /* __KERNEL__ */ #endif /* __ASM_SH_ADDRSPACE_H */ diff --git a/arch/sh/include/asm/mmu.h b/arch/sh/include/asm/mmu.h index e5e8f48..ca7d91e 100644 --- a/arch/sh/include/asm/mmu.h +++ b/arch/sh/include/asm/mmu.h @@ -31,6 +31,7 @@ #ifndef __ASSEMBLY__ #include +#include /* Default "unsigned long" context */ typedef unsigned long mm_context_id_t[NR_CPUS]; @@ -72,6 +73,7 @@ long pmb_remap(unsigned long virt, unsigned long phys, unsigned long size, unsigned long flags); void pmb_unmap(unsigned long addr); int pmb_init(void); +bool __in_29bit_mode(void); #else static inline long pmb_remap(unsigned long virt, unsigned long phys, unsigned long size, unsigned long flags) @@ -87,8 +89,14 @@ static inline int pmb_init(void) { return -ENODEV; } -#endif /* CONFIG_PMB */ +#ifdef CONFIG_29BIT +#define __in_29bit_mode() (1) +#else +#define __in_29bit_mode() (0) +#endif + +#endif /* CONFIG_PMB */ #endif /* __ASSEMBLY__ */ #endif /* __MMU_H */ diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c index 32ebd15..491d9d5 100644 --- a/arch/sh/mm/init.c +++ b/arch/sh/mm/init.c @@ -338,10 +338,3 @@ EXPORT_SYMBOL_GPL(memory_add_physaddr_to_nid); #endif #endif /* CONFIG_MEMORY_HOTPLUG */ - -#ifdef CONFIG_PMB -int __in_29bit_mode(void) -{ - return !(ctrl_inl(PMB_PASCR) & PASCR_SE); -} -#endif /* CONFIG_PMB */ diff --git a/arch/sh/mm/pmb.c b/arch/sh/mm/pmb.c index b796b6c..d318fa6 100644 --- a/arch/sh/mm/pmb.c +++ b/arch/sh/mm/pmb.c @@ -436,6 +436,11 @@ int __uses_jump_to_uncached pmb_init(void) return 0; } +bool __in_29bit_mode(void) +{ + return (__raw_readl(PMB_PASCR) & PASCR_SE) == 0; +} + static int pmb_seq_show(struct seq_file *file, void *iter) { int i; -- cgit v0.10.2 From fbb82b03653cdb7fd1863b911e7540011259d2ce Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 16:42:52 +0900 Subject: sh: machine_ops based reboot support. This provides a machine_ops-based reboot interface loosely cloned from x86, and converts the native sh32 and sh64 cases over to it. Necessary both for tying in SMP support and also enabling platforms like SDK7786 to add support for their microcontroller-based power managers. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/reboot.h b/arch/sh/include/asm/reboot.h new file mode 100644 index 0000000..b3da0c6 --- /dev/null +++ b/arch/sh/include/asm/reboot.h @@ -0,0 +1,21 @@ +#ifndef __ASM_SH_REBOOT_H +#define __ASM_SH_REBOOT_H + +#include + +struct pt_regs; + +struct machine_ops { + void (*restart)(char *cmd); + void (*halt)(void); + void (*power_off)(void); + void (*shutdown)(void); + void (*crash_shutdown)(struct pt_regs *); +}; + +extern struct machine_ops machine_ops; + +/* arch/sh/kernel/machine_kexec.c */ +void native_machine_crash_shutdown(struct pt_regs *regs); + +#endif /* __ASM_SH_REBOOT_H */ diff --git a/arch/sh/include/asm/system.h b/arch/sh/include/asm/system.h index 62e4fc1..de2fc39 100644 --- a/arch/sh/include/asm/system.h +++ b/arch/sh/include/asm/system.h @@ -143,6 +143,7 @@ extern struct dentry *sh_debugfs_root; void per_cpu_trap_init(void); void default_idle(void); void cpu_idle_wait(void); +void stop_this_cpu(void *); #ifdef CONFIG_SUPERH32 #define BUILD_TRAP_HANDLER(name) \ diff --git a/arch/sh/include/asm/system_32.h b/arch/sh/include/asm/system_32.h index 06814f5..34bd2ba 100644 --- a/arch/sh/include/asm/system_32.h +++ b/arch/sh/include/asm/system_32.h @@ -2,6 +2,7 @@ #define __ASM_SH_SYSTEM_32_H #include +#include #ifdef CONFIG_SH_DSP @@ -216,6 +217,17 @@ static inline reg_size_t register_align(void *val) int handle_unaligned_access(insn_size_t instruction, struct pt_regs *regs, struct mem_access *ma, int); +static inline void trigger_address_error(void) +{ + if (__in_29bit_mode()) + __asm__ __volatile__ ( + "ldc %0, sr\n\t" + "mov.l @%1, %0" + : + : "r" (0x10000000), "r" (0x80000001) + ); +} + asmlinkage void do_address_error(struct pt_regs *regs, unsigned long writeaccess, unsigned long address); diff --git a/arch/sh/include/asm/system_64.h b/arch/sh/include/asm/system_64.h index ab1dd91..604ba78 100644 --- a/arch/sh/include/asm/system_64.h +++ b/arch/sh/include/asm/system_64.h @@ -48,6 +48,13 @@ static inline reg_size_t register_align(void *val) return (unsigned long long)(signed long long)(signed long)val; } +extern void phys_stext(void); + +static inline void trigger_address_error(void) +{ + phys_stext(); +} + #define SR_BL_LL 0x0000000010000000LL static inline void set_bl_bit(void) diff --git a/arch/sh/kernel/Makefile b/arch/sh/kernel/Makefile index 56704a6..02fd3ae 100644 --- a/arch/sh/kernel/Makefile +++ b/arch/sh/kernel/Makefile @@ -14,7 +14,8 @@ CFLAGS_REMOVE_return_address.o = -pg obj-y := debugtraps.o dma-nommu.o dumpstack.o \ idle.o io.o io_generic.o irq.o \ irq_$(BITS).o machvec.o nmi_debug.o process.o \ - process_$(BITS).o ptrace_$(BITS).o return_address.o \ + process_$(BITS).o ptrace_$(BITS).o \ + reboot.o return_address.o \ setup.o signal_$(BITS).o sys_sh.o sys_sh$(BITS).o \ syscalls_$(BITS).o time.o topology.o traps.o \ traps_$(BITS).o unwinder.o diff --git a/arch/sh/kernel/idle.c b/arch/sh/kernel/idle.c index 6b3d706..0fd7b41 100644 --- a/arch/sh/kernel/idle.c +++ b/arch/sh/kernel/idle.c @@ -20,10 +20,9 @@ #include #include -static int hlt_counter; void (*pm_idle)(void) = NULL; -void (*pm_power_off)(void); -EXPORT_SYMBOL(pm_power_off); + +static int hlt_counter; static int __init nohlt_setup(char *__unused) { @@ -131,6 +130,15 @@ static void do_nothing(void *unused) { } +void stop_this_cpu(void *unused) +{ + local_irq_disable(); + cpu_clear(smp_processor_id(), cpu_online_map); + + for (;;) + cpu_sleep(); +} + /* * cpu_idle_wait - Used to ensure that all the CPUs discard old value of * pm_idle and update to new pm_idle value. Required while changing pm_idle diff --git a/arch/sh/kernel/machine_kexec.c b/arch/sh/kernel/machine_kexec.c index f52d8ed..7672141 100644 --- a/arch/sh/kernel/machine_kexec.c +++ b/arch/sh/kernel/machine_kexec.c @@ -22,6 +22,7 @@ #include #include #include +#include typedef void (*relocate_new_kernel_t)(unsigned long indirection_page, unsigned long reboot_code_buffer, @@ -31,12 +32,9 @@ extern const unsigned char relocate_new_kernel[]; extern const unsigned int relocate_new_kernel_size; extern void *vbr_base; -void machine_shutdown(void) -{ -} - -void machine_crash_shutdown(struct pt_regs *regs) +void native_machine_crash_shutdown(struct pt_regs *regs) { + /* Nothing to do for UP, but definitely broken for SMP.. */ } /* diff --git a/arch/sh/kernel/process_32.c b/arch/sh/kernel/process_32.c index 856010f..b6f43f0 100644 --- a/arch/sh/kernel/process_32.c +++ b/arch/sh/kernel/process_32.c @@ -16,63 +16,15 @@ #include #include #include -#include #include -#include -#include -#include -#include #include #include -#include #include #include #include -#include #include #include #include -#include - -#ifdef CONFIG_32BIT -static void watchdog_trigger_immediate(void) -{ - sh_wdt_write_cnt(0xFF); - sh_wdt_write_csr(0xC2); -} - -void machine_restart(char * __unused) -{ - local_irq_disable(); - - /* Use watchdog timer to trigger reset */ - watchdog_trigger_immediate(); - - while (1) - cpu_sleep(); -} -#else -void machine_restart(char * __unused) -{ - /* SR.BL=1 and invoke address error to let CPU reset (manual reset) */ - asm volatile("ldc %0, sr\n\t" - "mov.l @%1, %0" : : "r" (0x10000000), "r" (0x80000001)); -} -#endif - -void machine_halt(void) -{ - local_irq_disable(); - - while (1) - cpu_sleep(); -} - -void machine_power_off(void) -{ - if (pm_power_off) - pm_power_off(); -} void show_regs(struct pt_regs * regs) { diff --git a/arch/sh/kernel/process_64.c b/arch/sh/kernel/process_64.c index c9554a7..6ae5987 100644 --- a/arch/sh/kernel/process_64.c +++ b/arch/sh/kernel/process_64.c @@ -32,30 +32,7 @@ struct task_struct *last_task_used_math = NULL; -void machine_restart(char * __unused) -{ - extern void phys_stext(void); - - phys_stext(); -} - -void machine_halt(void) -{ - for (;;); -} - -void machine_power_off(void) -{ - __asm__ __volatile__ ( - "sleep\n\t" - "synci\n\t" - "nop;nop;nop;nop\n\t" - ); - - panic("Unexpected wakeup!\n"); -} - -void show_regs(struct pt_regs * regs) +void show_regs(struct pt_regs *regs) { unsigned long long ah, al, bh, bl, ch, cl; diff --git a/arch/sh/kernel/reboot.c b/arch/sh/kernel/reboot.c new file mode 100644 index 0000000..b1fca66 --- /dev/null +++ b/arch/sh/kernel/reboot.c @@ -0,0 +1,98 @@ +#include +#include +#include +#include +#include +#ifdef CONFIG_SUPERH32 +#include +#endif +#include +#include +#include + +void (*pm_power_off)(void); +EXPORT_SYMBOL(pm_power_off); + +#ifdef CONFIG_SUPERH32 +static void watchdog_trigger_immediate(void) +{ + sh_wdt_write_cnt(0xFF); + sh_wdt_write_csr(0xC2); +} +#endif + +static void native_machine_restart(char * __unused) +{ + local_irq_disable(); + + /* Address error with SR.BL=1 first. */ + trigger_address_error(); + +#ifdef CONFIG_SUPERH32 + /* If that fails or is unsupported, go for the watchdog next. */ + watchdog_trigger_immediate(); +#endif + + /* + * Give up and sleep. + */ + while (1) + cpu_sleep(); +} + +static void native_machine_shutdown(void) +{ + smp_send_stop(); +} + +static void native_machine_power_off(void) +{ + if (pm_power_off) + pm_power_off(); +} + +static void native_machine_halt(void) +{ + /* stop other cpus */ + machine_shutdown(); + + /* stop this cpu */ + stop_this_cpu(NULL); +} + +struct machine_ops machine_ops = { + .power_off = native_machine_power_off, + .shutdown = native_machine_shutdown, + .restart = native_machine_restart, + .halt = native_machine_halt, +#ifdef CONFIG_KEXEC + .crash_shutdown = native_machine_crash_shutdown, +#endif +}; + +void machine_power_off(void) +{ + machine_ops.power_off(); +} + +void machine_shutdown(void) +{ + machine_ops.shutdown(); +} + +void machine_restart(char *cmd) +{ + machine_ops.restart(cmd); +} + +void machine_halt(void) +{ + machine_ops.halt(); +} + +#ifdef CONFIG_KEXEC +void machine_crash_shutdown(struct pt_regs *regs) +{ + machine_ops.crash_shutdown(regs); +} +#endif diff --git a/arch/sh/kernel/smp.c b/arch/sh/kernel/smp.c index 983e079..e124cf7 100644 --- a/arch/sh/kernel/smp.c +++ b/arch/sh/kernel/smp.c @@ -161,15 +161,6 @@ void smp_send_reschedule(int cpu) plat_send_ipi(cpu, SMP_MSG_RESCHEDULE); } -static void stop_this_cpu(void *unused) -{ - cpu_clear(smp_processor_id(), cpu_online_map); - local_irq_disable(); - - for (;;) - cpu_relax(); -} - void smp_send_stop(void) { smp_call_function(stop_this_cpu, 0, 0); -- cgit v0.10.2 From b51989b8afe9409ee68c67ce2a5de4390693bd2b Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 16:53:11 +0900 Subject: sh: mach-sdk7786: reset controller reboot support. This wires up the machine_ops reboot call to use the system reset controller. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-sdk7786/setup.c b/arch/sh/boards/mach-sdk7786/setup.c index 3c3c9a3..f094ea2 100644 --- a/arch/sh/boards/mach-sdk7786/setup.c +++ b/arch/sh/boards/mach-sdk7786/setup.c @@ -20,6 +20,7 @@ #include #include #include +#include static struct resource heartbeat_resource = { .start = 0x07fff8b0, @@ -159,6 +160,11 @@ static int sdk7786_clk_init(void) return ret; } +static void sdk7786_restart(char *cmd) +{ + fpga_write_reg(0xa5a5, SRSTR); +} + /* Initialize the board */ static void __init sdk7786_setup(char **cmdline_p) { @@ -167,6 +173,8 @@ static void __init sdk7786_setup(char **cmdline_p) sdk7786_fpga_init(); pr_info("\tPCB revision:\t%d\n", fpga_read_reg(PCBRR) & 0xf); + + machine_ops.restart = sdk7786_restart; } /* -- cgit v0.10.2 From e8d433f335d44028d41af231ef5c52fd8a9b280b Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 18 Jan 2010 15:59:19 -0200 Subject: perf archive: Add documentation MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This also makes it appear on the 'perf --help' output, i.e. util/generate-cmdlist.sh now takes it into account. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263837559-24168-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Documentation/perf-archive.txt b/tools/perf/Documentation/perf-archive.txt new file mode 100644 index 0000000..fae174d --- /dev/null +++ b/tools/perf/Documentation/perf-archive.txt @@ -0,0 +1,22 @@ +perf-archive(1) +=============== + +NAME +---- +perf-archive - Create archive with object files with build-ids found in perf.data file + +SYNOPSIS +-------- +[verse] +'perf archive' [file] + +DESCRIPTION +----------- +This command runs runs perf-buildid-list --with-hits, and collects the files +with the buildids found so that analisys of perf.data contents can be possible +on another machine. + + +SEE ALSO +-------- +linkperf:perf-record[1], linkperf:perf-buildid-list[1], linkperf:perf-report[1] diff --git a/tools/perf/command-list.txt b/tools/perf/command-list.txt index f73d1d9..cf6444d 100644 --- a/tools/perf/command-list.txt +++ b/tools/perf/command-list.txt @@ -3,7 +3,7 @@ # command name category [deprecated] [common] # perf-annotate mainporcelain common -perf-archive mainporcelain +perf-archive mainporcelain common perf-bench mainporcelain common perf-buildid-list mainporcelain common perf-diff mainporcelain common -- cgit v0.10.2 From d5526d8cb8e5aa3349c1ff4e409ad9b4cdac380c Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 18 Jan 2010 18:21:42 -0200 Subject: perf archive: Fix installation steps in the Makefile MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fix these warning: acme@parisc:~/git/linux-2.6-tip$ make -C tools/perf/ install make: Entering directory `/home/acme/git/linux-2.6-tip/tools/perf' Makefile:833: warning: overriding commands for target `perf-archive' Makefile:822: warning: ignoring old commands for target `perf-archive' Reported-by: Ingo Molnar Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263846102-24841-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index d739552..ddbeeee 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -311,7 +311,6 @@ PROGRAMS += perf # List built-in command $C whose implementation cmd_$C() is not in # builtin-$C.o but is linked in as part of some other command. # -BUILT_INS += perf-archive # what 'all' will build and 'install' will install, in perfexecdir ALL_PROGRAMS = $(PROGRAMS) $(SCRIPTS) @@ -1004,6 +1003,7 @@ install: all $(INSTALL) perf$X '$(DESTDIR_SQ)$(bindir_SQ)' $(INSTALL) -d -m 755 '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/Perf-Trace-Util/lib/Perf/Trace' $(INSTALL) -d -m 755 '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/bin' + $(INSTALL) perf-archive -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)' $(INSTALL) scripts/perl/Perf-Trace-Util/lib/Perf/Trace/* -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/Perf-Trace-Util/lib/Perf/Trace' $(INSTALL) scripts/perl/*.pl -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl' $(INSTALL) scripts/perl/bin/* -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/bin' -- cgit v0.10.2 From f162f87ad6e98e8bfb2362955da46bed7b2514be Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 19 Jan 2010 10:36:13 -0200 Subject: perf symbols: Set dso->kernel when handling the fake vmlinux MMAP event MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Because it may be possible that there was no buildid section, where we would set this to 1. Found while analysing a perf.data file collected on an ARM machine where an explicitely specified vmlinux was being disregarded. Reported-by: Jamie Iles Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263904574-30732-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index dc13cad..bbaee61 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -373,6 +373,7 @@ int event__process_mmap(event_t *self, struct perf_session *session) if (kernel == NULL) goto out_problem; + kernel->kernel = 1; if (__map_groups__create_kernel_maps(&session->kmaps, session->vmlinux_maps, kernel) < 0) -- cgit v0.10.2 From dc8d6ab2b61a2d92b5d7438565ccd20b29724cb2 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 19 Jan 2010 10:36:14 -0200 Subject: perf symbols: Use only --vmlinux if specified MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Found while analysing a perf.data file collected on an ARM machine where an explicitely specified vmlinux was being disregarded. Reported-by: Jamie Iles Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1263904574-30732-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index a4e7459..b6ab23d 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1572,7 +1572,7 @@ static int dso__load_vmlinux(struct dso *self, struct map *map, return -1; dso__set_loaded(self, map->type); - err = dso__load_sym(self, map, session, self->long_name, fd, filter, 1, 0); + err = dso__load_sym(self, map, session, vmlinux, fd, filter, 1, 0); close(fd); return err; @@ -1584,6 +1584,26 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, int err; const char *kallsyms_filename = NULL; char *kallsyms_allocated_filename = NULL; + /* + * Step 1: if the user specified a vmlinux filename, use it and only + * it, reporting errors to the user if it cannot be used. + * + * For instance, try to analyse an ARM perf.data file _without_ a + * build-id, or if the user specifies the wrong path to the right + * vmlinux file, obviously we can't fallback to another vmlinux (a + * x86_86 one, on the machine where analysis is being performed, say), + * or worse, /proc/kallsyms. + * + * If the specified file _has_ a build-id and there is a build-id + * section in the perf.data file, we will still do the expected + * validation in dso__load_vmlinux and will bail out if they don't + * match. + */ + if (symbol_conf.vmlinux_name != NULL) { + err = dso__load_vmlinux(self, map, session, + symbol_conf.vmlinux_name, filter); + goto out_try_fixup; + } if (vmlinux_path != NULL) { int i; @@ -1618,46 +1638,41 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, goto do_kallsyms; } } - + /* + * Now look if we have it on the build-id cache in + * $HOME/.debug/[kernel.kallsyms]. + */ build_id__sprintf(self->build_id, sizeof(self->build_id), sbuild_id); if (asprintf(&kallsyms_allocated_filename, "%s/.debug/[kernel.kallsyms]/%s", - getenv("HOME"), sbuild_id) != -1) { - if (access(kallsyms_filename, F_OK)) { - kallsyms_filename = kallsyms_allocated_filename; - goto do_kallsyms; - } + getenv("HOME"), sbuild_id) == -1) + return -1; + + if (access(kallsyms_filename, F_OK)) { free(kallsyms_allocated_filename); - kallsyms_allocated_filename = NULL; + return -1; } - goto do_vmlinux; - } - - if (self->long_name[0] == '[') { + kallsyms_filename = kallsyms_allocated_filename; + } else { + /* + * Last resort, if we don't have a build-id and couldn't find + * any vmlinux file, try the running kernel kallsyms table. + */ kallsyms_filename = "/proc/kallsyms"; - goto do_kallsyms; } -do_vmlinux: - err = dso__load_vmlinux(self, map, session, self->long_name, filter); - if (err <= 0) { - if (self->has_build_id) - return -1; - - pr_info("The file %s cannot be used, " - "trying to use /proc/kallsyms...", self->long_name); do_kallsyms: - err = dso__load_kallsyms(self, kallsyms_filename, map, session, filter); - if (err > 0 && kallsyms_filename == NULL) - dso__set_long_name(self, strdup("[kernel.kallsyms]")); - free(kallsyms_allocated_filename); - } + err = dso__load_kallsyms(self, kallsyms_filename, map, session, filter); + free(kallsyms_allocated_filename); +out_try_fixup: if (err > 0) { out_fixup: + if (kallsyms_filename == NULL) + dso__set_long_name(self, strdup("[kernel.kallsyms]")); map__fixup_start(map); map__fixup_end(map); } -- cgit v0.10.2 From 016ddd9b5b315fe52ac6ba3167521f50b3868d34 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Mon, 18 Jan 2010 17:05:40 +0100 Subject: x86: Put 'nopat' in kernel-parameters Add missing 'nopat' boot option into Documentation/kernel-parameters.txt. Signed-off-by: Jiri Kosina LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 736d456..5bf45c0 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -1729,6 +1729,9 @@ and is between 256 and 4096 characters. It is defined in the file nomfgpt [X86-32] Disable Multi-Function General Purpose Timer usage (for AMD Geode machines). + nopat [X86] Disable PAT (page attribute table extension of + pagetables) support. + norandmaps Don't use address space randomization. Equivalent to echo 0 > /proc/sys/kernel/randomize_va_space -- cgit v0.10.2 From 920efaabcbd34e6b8dc05c5b777df3e936af5812 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 18:10:30 +0900 Subject: sh: Correct iounmap fixmap teardown. iounmap_fixed() had a couple of bugs in it that caused it to effectively fail at life. The total number of pages to unmap factored in the mapping offset and aligned up to the next page boundary, which doesn't match the ioremap_fixed() behaviour. When ioremap_fixed() pegs a slot, the address in the mapping data already contains the offset displacement, and the size is recorded verbatim given that we're only interested in total number of pages required. As such, we need to calculate the total number from the original size in the unmap path as well. At the same time, there was also an off-by-1 problem in the fixmap index calculation which has also been corrected. Previously subsequent remaps of an identical fixmap index would trigger the pte_ERROR() in set_pte_phys(): arch/sh/mm/init.c:77: bad pte 8053ffb0(0000781003fff506). arch/sh/mm/init.c:77: bad pte 8053ffb0(0000781003fff506). arch/sh/mm/init.c:77: bad pte 8053ffb0(0000781003fff506). arch/sh/mm/init.c:77: bad pte 8053ffb0(0000781003fff506). arch/sh/mm/init.c:77: bad pte 8053ffb0(0000781003fff506). arch/sh/mm/init.c:77: bad pte 8053ffb0(0000781003fff506). With this patch in place, the iounmap-driven fixmap teardown actually does what it's supposed to do. Signed-off-by: Paul Mundt diff --git a/arch/sh/mm/ioremap_fixed.c b/arch/sh/mm/ioremap_fixed.c index 551b513..0b78b1e 100644 --- a/arch/sh/mm/ioremap_fixed.c +++ b/arch/sh/mm/ioremap_fixed.c @@ -93,9 +93,7 @@ ioremap_fixed(resource_size_t phys_addr, unsigned long offset, int iounmap_fixed(void __iomem *addr) { enum fixed_addresses idx; - unsigned long virt_addr; struct ioremap_map *map; - unsigned long offset; unsigned int nrpages; int i, slot; @@ -114,12 +112,9 @@ int iounmap_fixed(void __iomem *addr) if (slot < 0) return -EINVAL; - virt_addr = (unsigned long)addr; + nrpages = map->size >> PAGE_SHIFT; - offset = virt_addr & ~PAGE_MASK; - nrpages = PAGE_ALIGN(offset + map->size - 1) >> PAGE_SHIFT; - - idx = FIX_IOREMAP_BEGIN + slot + nrpages; + idx = FIX_IOREMAP_BEGIN + slot + nrpages - 1; while (nrpages > 0) { __clear_fixmap(idx, __pgprot(_PAGE_WIRED)); --idx; -- cgit v0.10.2 From d9116d07f60383cef134c43a0ba15ec4375310fc Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 18:25:19 +0900 Subject: sh: mach-sdk7786: Probe system FPGA area mapping. This implements dynamic probing for the system FPGA. The system reset controller contains a fixed magic read word in order to identify the FPGA. This just utilizes a simple loop that scans across all of the fixed physical areas (area 0 through area 6) to locate the FPGA. The FPGA also contains register information detailing the area mappings and chip select settings for all of the other blocks, so this needs to be done before we can set up anything else. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-sdk7786/fpga.c b/arch/sh/boards/mach-sdk7786/fpga.c index 99f903c..3e4ec66 100644 --- a/arch/sh/boards/mach-sdk7786/fpga.c +++ b/arch/sh/boards/mach-sdk7786/fpga.c @@ -11,9 +11,44 @@ #include #include #include +#include -#define FPGA_REGS_BASE 0x07fff800 -#define FPGA_REGS_SIZE 0x490 +#define FPGA_REGS_OFFSET 0x03fff800 +#define FPGA_REGS_SIZE 0x490 + +/* + * The FPGA can be mapped in any of the generally available areas, + * so we attempt to scan for it using the fixed SRSTR read magic. + * + * Once the FPGA is located, the rest of the mapping data for the other + * components can be determined dynamically from its section mapping + * registers. + */ +static void __iomem *sdk7786_fpga_probe(void) +{ + unsigned long area; + void __iomem *base; + + /* + * Iterate over all of the areas where the FPGA could be mapped. + * The possible range is anywhere from area 0 through 6, area 7 + * is reserved. + */ + for (area = PA_AREA0; area < PA_AREA7; area += SZ_64M) { + base = ioremap_nocache(area + FPGA_REGS_OFFSET, FPGA_REGS_SIZE); + if (!base) { + /* Failed to remap this area, move along. */ + continue; + } + + if (ioread16(base + SRSTR) == SRSTR_MAGIC) + return base; /* Found it! */ + + iounmap(base); + } + + return NULL; +} void __iomem *sdk7786_fpga_base; @@ -21,9 +56,9 @@ void __init sdk7786_fpga_init(void) { u16 version, date; - sdk7786_fpga_base = ioremap_nocache(FPGA_REGS_BASE, FPGA_REGS_SIZE); + sdk7786_fpga_base = sdk7786_fpga_probe(); if (unlikely(!sdk7786_fpga_base)) { - panic("FPGA remapping failed.\n"); + panic("FPGA detection failed.\n"); return; } diff --git a/arch/sh/include/cpu-sh4/cpu/addrspace.h b/arch/sh/include/cpu-sh4/cpu/addrspace.h index a3fa733..d51da25 100644 --- a/arch/sh/include/cpu-sh4/cpu/addrspace.h +++ b/arch/sh/include/cpu-sh4/cpu/addrspace.h @@ -28,6 +28,15 @@ #define P4SEG_TLB_DATA 0xf7000000 #define P4SEG_REG_BASE 0xff000000 +#define PA_AREA0 0x00000000 +#define PA_AREA1 0x04000000 +#define PA_AREA2 0x08000000 +#define PA_AREA3 0x0c000000 +#define PA_AREA4 0x10000000 +#define PA_AREA5 0x14000000 +#define PA_AREA6 0x18000000 +#define PA_AREA7 0x1c000000 + #define PA_AREA5_IO 0xb4000000 /* Area 5 IO Memory */ #define PA_AREA6_IO 0xb8000000 /* Area 6 IO Memory */ diff --git a/arch/sh/include/mach-sdk7786/mach/fpga.h b/arch/sh/include/mach-sdk7786/mach/fpga.h index a85d985..2120d67 100644 --- a/arch/sh/include/mach-sdk7786/mach/fpga.h +++ b/arch/sh/include/mach-sdk7786/mach/fpga.h @@ -6,6 +6,8 @@ #include #define SRSTR 0x000 +#define SRSTR_MAGIC 0x1971 /* Fixed magical read value */ + #define INTASR 0x010 #define INTAMR 0x020 #define MODSWR 0x030 -- cgit v0.10.2 From 35f99c0da17beb5004f06ba86a87d82c1a6467be Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 18:48:17 +0900 Subject: sh: pretty print virtual memory map on boot. This cribs the pretty printing from arch/x86/mm/init_32.c to dump the virtual memory layout on boot. This is primarily intended as a debugging aid, given that the newer CPUs have full control over their address space and as such have little to nothing in common with the legacy layout. Signed-off-by: Paul Mundt diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c index 491d9d5..f3e23ad 100644 --- a/arch/sh/mm/init.c +++ b/arch/sh/mm/init.c @@ -260,6 +260,9 @@ void __init mem_init(void) memset(empty_zero_page, 0, PAGE_SIZE); __flush_wback_region(empty_zero_page, PAGE_SIZE); + /* Initialize the vDSO */ + vsyscall_init(); + codesize = (unsigned long) &_etext - (unsigned long) &_text; datasize = (unsigned long) &_edata - (unsigned long) &_etext; initsize = (unsigned long) &__init_end - (unsigned long) &__init_begin; @@ -272,8 +275,39 @@ void __init mem_init(void) datasize >> 10, initsize >> 10); - /* Initialize the vDSO */ - vsyscall_init(); + printk(KERN_INFO "virtual kernel memory layout:\n" + " fixmap : 0x%08lx - 0x%08lx (%4ld kB)\n" +#ifdef CONFIG_HIGHMEM + " pkmap : 0x%08lx - 0x%08lx (%4ld kB)\n" +#endif + " vmalloc : 0x%08lx - 0x%08lx (%4ld MB)\n" + " lowmem : 0x%08lx - 0x%08lx (%4ld MB)\n" + " .init : 0x%08lx - 0x%08lx (%4ld kB)\n" + " .data : 0x%08lx - 0x%08lx (%4ld kB)\n" + " .text : 0x%08lx - 0x%08lx (%4ld kB)\n", + FIXADDR_START, FIXADDR_TOP, + (FIXADDR_TOP - FIXADDR_START) >> 10, + +#ifdef CONFIG_HIGHMEM + PKMAP_BASE, PKMAP_BASE+LAST_PKMAP*PAGE_SIZE, + (LAST_PKMAP*PAGE_SIZE) >> 10, +#endif + + (unsigned long)VMALLOC_START, VMALLOC_END, + (VMALLOC_END - VMALLOC_START) >> 20, + + (unsigned long)memory_start, (unsigned long)high_memory, + ((unsigned long)high_memory - (unsigned long)memory_start) >> 20, + + (unsigned long)&__init_begin, (unsigned long)&__init_end, + ((unsigned long)&__init_end - + (unsigned long)&__init_begin) >> 10, + + (unsigned long)&_etext, (unsigned long)&_edata, + ((unsigned long)&_edata - (unsigned long)&_etext) >> 10, + + (unsigned long)&_text, (unsigned long)&_etext, + ((unsigned long)&_etext - (unsigned long)&_text) >> 10); mem_init_done = 1; } -- cgit v0.10.2 From 358934a60d2180dcd1ed20691dbb66d4fb977ab2 Mon Sep 17 00:00:00 2001 From: Sandeep Paulraj Date: Wed, 16 Dec 2009 22:02:18 +0000 Subject: spi: Add SPI master driver for DaVinci/DA8xx This patch adds support for a SPI master driver for the DaVinci series of SOCs Signed-off-by: Sandeep Paulraj Signed-off-by: Mark A. Greer Signed-off-by: Philby John Signed-off-by: Sudhakar Rajashekhara Signed-off-by: Kevin Hilman Signed-off-by: Grant Likely diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig index f55eb01..82dabbc 100644 --- a/drivers/spi/Kconfig +++ b/drivers/spi/Kconfig @@ -100,6 +100,13 @@ config SPI_BUTTERFLY inexpensive battery powered microcontroller evaluation board. This same cable can be used to flash new firmware. +config SPI_DAVINCI + tristate "SPI controller driver for DaVinci/DA8xx SoC's" + depends on SPI_MASTER && ARCH_DAVINCI + select SPI_BITBANG + help + SPI master controller for DaVinci and DA8xx SPI modules. + config SPI_GPIO tristate "GPIO-based bitbanging SPI Master" depends on GENERIC_GPIO diff --git a/drivers/spi/Makefile b/drivers/spi/Makefile index f3d2810..84b490f 100644 --- a/drivers/spi/Makefile +++ b/drivers/spi/Makefile @@ -16,6 +16,7 @@ obj-$(CONFIG_SPI_BFIN) += spi_bfin5xx.o obj-$(CONFIG_SPI_BITBANG) += spi_bitbang.o obj-$(CONFIG_SPI_AU1550) += au1550_spi.o obj-$(CONFIG_SPI_BUTTERFLY) += spi_butterfly.o +obj-$(CONFIG_SPI_DAVINCI) += davinci_spi.o obj-$(CONFIG_SPI_DESIGNWARE) += dw_spi.o obj-$(CONFIG_SPI_DW_PCI) += dw_spi_pci.o obj-$(CONFIG_SPI_GPIO) += spi_gpio.o diff --git a/drivers/spi/davinci_spi.c b/drivers/spi/davinci_spi.c new file mode 100644 index 0000000..225ab60 --- /dev/null +++ b/drivers/spi/davinci_spi.c @@ -0,0 +1,1255 @@ +/* + * Copyright (C) 2009 Texas Instruments. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +#define SPI_NO_RESOURCE ((resource_size_t)-1) + +#define SPI_MAX_CHIPSELECT 2 + +#define CS_DEFAULT 0xFF + +#define SPI_BUFSIZ (SMP_CACHE_BYTES + 1) +#define DAVINCI_DMA_DATA_TYPE_S8 0x01 +#define DAVINCI_DMA_DATA_TYPE_S16 0x02 +#define DAVINCI_DMA_DATA_TYPE_S32 0x04 + +#define SPIFMT_PHASE_MASK BIT(16) +#define SPIFMT_POLARITY_MASK BIT(17) +#define SPIFMT_DISTIMER_MASK BIT(18) +#define SPIFMT_SHIFTDIR_MASK BIT(20) +#define SPIFMT_WAITENA_MASK BIT(21) +#define SPIFMT_PARITYENA_MASK BIT(22) +#define SPIFMT_ODD_PARITY_MASK BIT(23) +#define SPIFMT_WDELAY_MASK 0x3f000000u +#define SPIFMT_WDELAY_SHIFT 24 +#define SPIFMT_CHARLEN_MASK 0x0000001Fu + +/* SPIGCR1 */ +#define SPIGCR1_SPIENA_MASK 0x01000000u + +/* SPIPC0 */ +#define SPIPC0_DIFUN_MASK BIT(11) /* MISO */ +#define SPIPC0_DOFUN_MASK BIT(10) /* MOSI */ +#define SPIPC0_CLKFUN_MASK BIT(9) /* CLK */ +#define SPIPC0_SPIENA_MASK BIT(8) /* nREADY */ +#define SPIPC0_EN1FUN_MASK BIT(1) +#define SPIPC0_EN0FUN_MASK BIT(0) + +#define SPIINT_MASKALL 0x0101035F +#define SPI_INTLVL_1 0x000001FFu +#define SPI_INTLVL_0 0x00000000u + +/* SPIDAT1 */ +#define SPIDAT1_CSHOLD_SHIFT 28 +#define SPIDAT1_CSNR_SHIFT 16 +#define SPIGCR1_CLKMOD_MASK BIT(1) +#define SPIGCR1_MASTER_MASK BIT(0) +#define SPIGCR1_LOOPBACK_MASK BIT(16) + +/* SPIBUF */ +#define SPIBUF_TXFULL_MASK BIT(29) +#define SPIBUF_RXEMPTY_MASK BIT(31) + +/* Error Masks */ +#define SPIFLG_DLEN_ERR_MASK BIT(0) +#define SPIFLG_TIMEOUT_MASK BIT(1) +#define SPIFLG_PARERR_MASK BIT(2) +#define SPIFLG_DESYNC_MASK BIT(3) +#define SPIFLG_BITERR_MASK BIT(4) +#define SPIFLG_OVRRUN_MASK BIT(6) +#define SPIFLG_RX_INTR_MASK BIT(8) +#define SPIFLG_TX_INTR_MASK BIT(9) +#define SPIFLG_BUF_INIT_ACTIVE_MASK BIT(24) +#define SPIFLG_MASK (SPIFLG_DLEN_ERR_MASK \ + | SPIFLG_TIMEOUT_MASK | SPIFLG_PARERR_MASK \ + | SPIFLG_DESYNC_MASK | SPIFLG_BITERR_MASK \ + | SPIFLG_OVRRUN_MASK | SPIFLG_RX_INTR_MASK \ + | SPIFLG_TX_INTR_MASK \ + | SPIFLG_BUF_INIT_ACTIVE_MASK) + +#define SPIINT_DLEN_ERR_INTR BIT(0) +#define SPIINT_TIMEOUT_INTR BIT(1) +#define SPIINT_PARERR_INTR BIT(2) +#define SPIINT_DESYNC_INTR BIT(3) +#define SPIINT_BITERR_INTR BIT(4) +#define SPIINT_OVRRUN_INTR BIT(6) +#define SPIINT_RX_INTR BIT(8) +#define SPIINT_TX_INTR BIT(9) +#define SPIINT_DMA_REQ_EN BIT(16) +#define SPIINT_ENABLE_HIGHZ BIT(24) + +#define SPI_T2CDELAY_SHIFT 16 +#define SPI_C2TDELAY_SHIFT 24 + +/* SPI Controller registers */ +#define SPIGCR0 0x00 +#define SPIGCR1 0x04 +#define SPIINT 0x08 +#define SPILVL 0x0c +#define SPIFLG 0x10 +#define SPIPC0 0x14 +#define SPIPC1 0x18 +#define SPIPC2 0x1c +#define SPIPC3 0x20 +#define SPIPC4 0x24 +#define SPIPC5 0x28 +#define SPIPC6 0x2c +#define SPIPC7 0x30 +#define SPIPC8 0x34 +#define SPIDAT0 0x38 +#define SPIDAT1 0x3c +#define SPIBUF 0x40 +#define SPIEMU 0x44 +#define SPIDELAY 0x48 +#define SPIDEF 0x4c +#define SPIFMT0 0x50 +#define SPIFMT1 0x54 +#define SPIFMT2 0x58 +#define SPIFMT3 0x5c +#define TGINTVEC0 0x60 +#define TGINTVEC1 0x64 + +struct davinci_spi_slave { + u32 cmd_to_write; + u32 clk_ctrl_to_write; + u32 bytes_per_word; + u8 active_cs; +}; + +/* We have 2 DMA channels per CS, one for RX and one for TX */ +struct davinci_spi_dma { + int dma_tx_channel; + int dma_rx_channel; + int dma_tx_sync_dev; + int dma_rx_sync_dev; + enum dma_event_q eventq; + + struct completion dma_tx_completion; + struct completion dma_rx_completion; +}; + +/* SPI Controller driver's private data. */ +struct davinci_spi { + struct spi_bitbang bitbang; + struct clk *clk; + + u8 version; + resource_size_t pbase; + void __iomem *base; + size_t region_size; + u32 irq; + struct completion done; + + const void *tx; + void *rx; + u8 *tmp_buf; + int count; + struct davinci_spi_dma *dma_channels; + struct davinci_spi_platform_data *pdata; + + void (*get_rx)(u32 rx_data, struct davinci_spi *); + u32 (*get_tx)(struct davinci_spi *); + + struct davinci_spi_slave slave[SPI_MAX_CHIPSELECT]; +}; + +static unsigned use_dma; + +static void davinci_spi_rx_buf_u8(u32 data, struct davinci_spi *davinci_spi) +{ + u8 *rx = davinci_spi->rx; + + *rx++ = (u8)data; + davinci_spi->rx = rx; +} + +static void davinci_spi_rx_buf_u16(u32 data, struct davinci_spi *davinci_spi) +{ + u16 *rx = davinci_spi->rx; + + *rx++ = (u16)data; + davinci_spi->rx = rx; +} + +static u32 davinci_spi_tx_buf_u8(struct davinci_spi *davinci_spi) +{ + u32 data; + const u8 *tx = davinci_spi->tx; + + data = *tx++; + davinci_spi->tx = tx; + return data; +} + +static u32 davinci_spi_tx_buf_u16(struct davinci_spi *davinci_spi) +{ + u32 data; + const u16 *tx = davinci_spi->tx; + + data = *tx++; + davinci_spi->tx = tx; + return data; +} + +static inline void set_io_bits(void __iomem *addr, u32 bits) +{ + u32 v = ioread32(addr); + + v |= bits; + iowrite32(v, addr); +} + +static inline void clear_io_bits(void __iomem *addr, u32 bits) +{ + u32 v = ioread32(addr); + + v &= ~bits; + iowrite32(v, addr); +} + +static inline void set_fmt_bits(void __iomem *addr, u32 bits, int cs_num) +{ + set_io_bits(addr + SPIFMT0 + (0x4 * cs_num), bits); +} + +static inline void clear_fmt_bits(void __iomem *addr, u32 bits, int cs_num) +{ + clear_io_bits(addr + SPIFMT0 + (0x4 * cs_num), bits); +} + +static void davinci_spi_set_dma_req(const struct spi_device *spi, int enable) +{ + struct davinci_spi *davinci_spi = spi_master_get_devdata(spi->master); + + if (enable) + set_io_bits(davinci_spi->base + SPIINT, SPIINT_DMA_REQ_EN); + else + clear_io_bits(davinci_spi->base + SPIINT, SPIINT_DMA_REQ_EN); +} + +/* + * Interface to control the chip select signal + */ +static void davinci_spi_chipselect(struct spi_device *spi, int value) +{ + struct davinci_spi *davinci_spi; + struct davinci_spi_platform_data *pdata; + u32 data1_reg_val = 0; + + davinci_spi = spi_master_get_devdata(spi->master); + pdata = davinci_spi->pdata; + + /* + * Board specific chip select logic decides the polarity and cs + * line for the controller + */ + if (value == BITBANG_CS_INACTIVE) { + set_io_bits(davinci_spi->base + SPIDEF, CS_DEFAULT); + + data1_reg_val |= CS_DEFAULT << SPIDAT1_CSNR_SHIFT; + iowrite32(data1_reg_val, davinci_spi->base + SPIDAT1); + + while ((ioread32(davinci_spi->base + SPIBUF) + & SPIBUF_RXEMPTY_MASK) == 0) + cpu_relax(); + } +} + +/** + * davinci_spi_setup_transfer - This functions will determine transfer method + * @spi: spi device on which data transfer to be done + * @t: spi transfer in which transfer info is filled + * + * This function determines data transfer method (8/16/32 bit transfer). + * It will also set the SPI Clock Control register according to + * SPI slave device freq. + */ +static int davinci_spi_setup_transfer(struct spi_device *spi, + struct spi_transfer *t) +{ + + struct davinci_spi *davinci_spi; + struct davinci_spi_platform_data *pdata; + u8 bits_per_word = 0; + u32 hz = 0, prescale; + + davinci_spi = spi_master_get_devdata(spi->master); + pdata = davinci_spi->pdata; + + if (t) { + bits_per_word = t->bits_per_word; + hz = t->speed_hz; + } + + /* if bits_per_word is not set then set it default */ + if (!bits_per_word) + bits_per_word = spi->bits_per_word; + + /* + * Assign function pointer to appropriate transfer method + * 8bit, 16bit or 32bit transfer + */ + if (bits_per_word <= 8 && bits_per_word >= 2) { + davinci_spi->get_rx = davinci_spi_rx_buf_u8; + davinci_spi->get_tx = davinci_spi_tx_buf_u8; + davinci_spi->slave[spi->chip_select].bytes_per_word = 1; + } else if (bits_per_word <= 16 && bits_per_word >= 2) { + davinci_spi->get_rx = davinci_spi_rx_buf_u16; + davinci_spi->get_tx = davinci_spi_tx_buf_u16; + davinci_spi->slave[spi->chip_select].bytes_per_word = 2; + } else + return -EINVAL; + + if (!hz) + hz = spi->max_speed_hz; + + clear_fmt_bits(davinci_spi->base, SPIFMT_CHARLEN_MASK, + spi->chip_select); + set_fmt_bits(davinci_spi->base, bits_per_word & 0x1f, + spi->chip_select); + + prescale = ((clk_get_rate(davinci_spi->clk) / hz) - 1) & 0xff; + + clear_fmt_bits(davinci_spi->base, 0x0000ff00, spi->chip_select); + set_fmt_bits(davinci_spi->base, prescale << 8, spi->chip_select); + + return 0; +} + +static void davinci_spi_dma_rx_callback(unsigned lch, u16 ch_status, void *data) +{ + struct spi_device *spi = (struct spi_device *)data; + struct davinci_spi *davinci_spi; + struct davinci_spi_dma *davinci_spi_dma; + struct davinci_spi_platform_data *pdata; + + davinci_spi = spi_master_get_devdata(spi->master); + davinci_spi_dma = &(davinci_spi->dma_channels[spi->chip_select]); + pdata = davinci_spi->pdata; + + if (ch_status == DMA_COMPLETE) + edma_stop(davinci_spi_dma->dma_rx_channel); + else + edma_clean_channel(davinci_spi_dma->dma_rx_channel); + + complete(&davinci_spi_dma->dma_rx_completion); + /* We must disable the DMA RX request */ + davinci_spi_set_dma_req(spi, 0); +} + +static void davinci_spi_dma_tx_callback(unsigned lch, u16 ch_status, void *data) +{ + struct spi_device *spi = (struct spi_device *)data; + struct davinci_spi *davinci_spi; + struct davinci_spi_dma *davinci_spi_dma; + struct davinci_spi_platform_data *pdata; + + davinci_spi = spi_master_get_devdata(spi->master); + davinci_spi_dma = &(davinci_spi->dma_channels[spi->chip_select]); + pdata = davinci_spi->pdata; + + if (ch_status == DMA_COMPLETE) + edma_stop(davinci_spi_dma->dma_tx_channel); + else + edma_clean_channel(davinci_spi_dma->dma_tx_channel); + + complete(&davinci_spi_dma->dma_tx_completion); + /* We must disable the DMA TX request */ + davinci_spi_set_dma_req(spi, 0); +} + +static int davinci_spi_request_dma(struct spi_device *spi) +{ + struct davinci_spi *davinci_spi; + struct davinci_spi_dma *davinci_spi_dma; + struct davinci_spi_platform_data *pdata; + struct device *sdev; + int r; + + davinci_spi = spi_master_get_devdata(spi->master); + davinci_spi_dma = &davinci_spi->dma_channels[spi->chip_select]; + pdata = davinci_spi->pdata; + sdev = davinci_spi->bitbang.master->dev.parent; + + r = edma_alloc_channel(davinci_spi_dma->dma_rx_sync_dev, + davinci_spi_dma_rx_callback, spi, + davinci_spi_dma->eventq); + if (r < 0) { + dev_dbg(sdev, "Unable to request DMA channel for SPI RX\n"); + return -EAGAIN; + } + davinci_spi_dma->dma_rx_channel = r; + r = edma_alloc_channel(davinci_spi_dma->dma_tx_sync_dev, + davinci_spi_dma_tx_callback, spi, + davinci_spi_dma->eventq); + if (r < 0) { + edma_free_channel(davinci_spi_dma->dma_rx_channel); + davinci_spi_dma->dma_rx_channel = -1; + dev_dbg(sdev, "Unable to request DMA channel for SPI TX\n"); + return -EAGAIN; + } + davinci_spi_dma->dma_tx_channel = r; + + return 0; +} + +/** + * davinci_spi_setup - This functions will set default transfer method + * @spi: spi device on which data transfer to be done + * + * This functions sets the default transfer method. + */ + +static int davinci_spi_setup(struct spi_device *spi) +{ + int retval; + struct davinci_spi *davinci_spi; + struct davinci_spi_dma *davinci_spi_dma; + struct device *sdev; + + davinci_spi = spi_master_get_devdata(spi->master); + sdev = davinci_spi->bitbang.master->dev.parent; + + /* if bits per word length is zero then set it default 8 */ + if (!spi->bits_per_word) + spi->bits_per_word = 8; + + davinci_spi->slave[spi->chip_select].cmd_to_write = 0; + + if (use_dma && davinci_spi->dma_channels) { + davinci_spi_dma = &davinci_spi->dma_channels[spi->chip_select]; + + if ((davinci_spi_dma->dma_rx_channel == -1) + || (davinci_spi_dma->dma_tx_channel == -1)) { + retval = davinci_spi_request_dma(spi); + if (retval < 0) + return retval; + } + } + + /* + * SPI in DaVinci and DA8xx operate between + * 600 KHz and 50 MHz + */ + if (spi->max_speed_hz < 600000 || spi->max_speed_hz > 50000000) { + dev_dbg(sdev, "Operating frequency is not in acceptable " + "range\n"); + return -EINVAL; + } + + /* + * Set up SPIFMTn register, unique to this chipselect. + * + * NOTE: we could do all of these with one write. Also, some + * of the "version 2" features are found in chips that don't + * support all of them... + */ + if (spi->mode & SPI_LSB_FIRST) + set_fmt_bits(davinci_spi->base, SPIFMT_SHIFTDIR_MASK, + spi->chip_select); + else + clear_fmt_bits(davinci_spi->base, SPIFMT_SHIFTDIR_MASK, + spi->chip_select); + + if (spi->mode & SPI_CPOL) + set_fmt_bits(davinci_spi->base, SPIFMT_POLARITY_MASK, + spi->chip_select); + else + clear_fmt_bits(davinci_spi->base, SPIFMT_POLARITY_MASK, + spi->chip_select); + + if (!(spi->mode & SPI_CPHA)) + set_fmt_bits(davinci_spi->base, SPIFMT_PHASE_MASK, + spi->chip_select); + else + clear_fmt_bits(davinci_spi->base, SPIFMT_PHASE_MASK, + spi->chip_select); + + /* + * Version 1 hardware supports two basic SPI modes: + * - Standard SPI mode uses 4 pins, with chipselect + * - 3 pin SPI is a 4 pin variant without CS (SPI_NO_CS) + * (distinct from SPI_3WIRE, with just one data wire; + * or similar variants without MOSI or without MISO) + * + * Version 2 hardware supports an optional handshaking signal, + * so it can support two more modes: + * - 5 pin SPI variant is standard SPI plus SPI_READY + * - 4 pin with enable is (SPI_READY | SPI_NO_CS) + */ + + if (davinci_spi->version == SPI_VERSION_2) { + clear_fmt_bits(davinci_spi->base, SPIFMT_WDELAY_MASK, + spi->chip_select); + set_fmt_bits(davinci_spi->base, + (davinci_spi->pdata->wdelay + << SPIFMT_WDELAY_SHIFT) + & SPIFMT_WDELAY_MASK, + spi->chip_select); + + if (davinci_spi->pdata->odd_parity) + set_fmt_bits(davinci_spi->base, + SPIFMT_ODD_PARITY_MASK, + spi->chip_select); + else + clear_fmt_bits(davinci_spi->base, + SPIFMT_ODD_PARITY_MASK, + spi->chip_select); + + if (davinci_spi->pdata->parity_enable) + set_fmt_bits(davinci_spi->base, + SPIFMT_PARITYENA_MASK, + spi->chip_select); + else + clear_fmt_bits(davinci_spi->base, + SPIFMT_PARITYENA_MASK, + spi->chip_select); + + if (davinci_spi->pdata->wait_enable) + set_fmt_bits(davinci_spi->base, + SPIFMT_WAITENA_MASK, + spi->chip_select); + else + clear_fmt_bits(davinci_spi->base, + SPIFMT_WAITENA_MASK, + spi->chip_select); + + if (davinci_spi->pdata->timer_disable) + set_fmt_bits(davinci_spi->base, + SPIFMT_DISTIMER_MASK, + spi->chip_select); + else + clear_fmt_bits(davinci_spi->base, + SPIFMT_DISTIMER_MASK, + spi->chip_select); + } + + retval = davinci_spi_setup_transfer(spi, NULL); + + return retval; +} + +static void davinci_spi_cleanup(struct spi_device *spi) +{ + struct davinci_spi *davinci_spi = spi_master_get_devdata(spi->master); + struct davinci_spi_dma *davinci_spi_dma; + + davinci_spi_dma = &davinci_spi->dma_channels[spi->chip_select]; + + if (use_dma && davinci_spi->dma_channels) { + davinci_spi_dma = &davinci_spi->dma_channels[spi->chip_select]; + + if ((davinci_spi_dma->dma_rx_channel != -1) + && (davinci_spi_dma->dma_tx_channel != -1)) { + edma_free_channel(davinci_spi_dma->dma_tx_channel); + edma_free_channel(davinci_spi_dma->dma_rx_channel); + } + } +} + +static int davinci_spi_bufs_prep(struct spi_device *spi, + struct davinci_spi *davinci_spi) +{ + int op_mode = 0; + + /* + * REVISIT unless devices disagree about SPI_LOOP or + * SPI_READY (SPI_NO_CS only allows one device!), this + * should not need to be done before each message... + * optimize for both flags staying cleared. + */ + + op_mode = SPIPC0_DIFUN_MASK + | SPIPC0_DOFUN_MASK + | SPIPC0_CLKFUN_MASK; + if (!(spi->mode & SPI_NO_CS)) + op_mode |= 1 << spi->chip_select; + if (spi->mode & SPI_READY) + op_mode |= SPIPC0_SPIENA_MASK; + + iowrite32(op_mode, davinci_spi->base + SPIPC0); + + if (spi->mode & SPI_LOOP) + set_io_bits(davinci_spi->base + SPIGCR1, + SPIGCR1_LOOPBACK_MASK); + else + clear_io_bits(davinci_spi->base + SPIGCR1, + SPIGCR1_LOOPBACK_MASK); + + return 0; +} + +static int davinci_spi_check_error(struct davinci_spi *davinci_spi, + int int_status) +{ + struct device *sdev = davinci_spi->bitbang.master->dev.parent; + + if (int_status & SPIFLG_TIMEOUT_MASK) { + dev_dbg(sdev, "SPI Time-out Error\n"); + return -ETIMEDOUT; + } + if (int_status & SPIFLG_DESYNC_MASK) { + dev_dbg(sdev, "SPI Desynchronization Error\n"); + return -EIO; + } + if (int_status & SPIFLG_BITERR_MASK) { + dev_dbg(sdev, "SPI Bit error\n"); + return -EIO; + } + + if (davinci_spi->version == SPI_VERSION_2) { + if (int_status & SPIFLG_DLEN_ERR_MASK) { + dev_dbg(sdev, "SPI Data Length Error\n"); + return -EIO; + } + if (int_status & SPIFLG_PARERR_MASK) { + dev_dbg(sdev, "SPI Parity Error\n"); + return -EIO; + } + if (int_status & SPIFLG_OVRRUN_MASK) { + dev_dbg(sdev, "SPI Data Overrun error\n"); + return -EIO; + } + if (int_status & SPIFLG_TX_INTR_MASK) { + dev_dbg(sdev, "SPI TX intr bit set\n"); + return -EIO; + } + if (int_status & SPIFLG_BUF_INIT_ACTIVE_MASK) { + dev_dbg(sdev, "SPI Buffer Init Active\n"); + return -EBUSY; + } + } + + return 0; +} + +/** + * davinci_spi_bufs - functions which will handle transfer data + * @spi: spi device on which data transfer to be done + * @t: spi transfer in which transfer info is filled + * + * This function will put data to be transferred into data register + * of SPI controller and then wait until the completion will be marked + * by the IRQ Handler. + */ +static int davinci_spi_bufs_pio(struct spi_device *spi, struct spi_transfer *t) +{ + struct davinci_spi *davinci_spi; + int int_status, count, ret; + u8 conv, tmp; + u32 tx_data, data1_reg_val; + u32 buf_val, flg_val; + struct davinci_spi_platform_data *pdata; + + davinci_spi = spi_master_get_devdata(spi->master); + pdata = davinci_spi->pdata; + + davinci_spi->tx = t->tx_buf; + davinci_spi->rx = t->rx_buf; + + /* convert len to words based on bits_per_word */ + conv = davinci_spi->slave[spi->chip_select].bytes_per_word; + davinci_spi->count = t->len / conv; + + INIT_COMPLETION(davinci_spi->done); + + ret = davinci_spi_bufs_prep(spi, davinci_spi); + if (ret) + return ret; + + /* Enable SPI */ + set_io_bits(davinci_spi->base + SPIGCR1, SPIGCR1_SPIENA_MASK); + + iowrite32(0 | (pdata->c2tdelay << SPI_C2TDELAY_SHIFT) | + (pdata->t2cdelay << SPI_T2CDELAY_SHIFT), + davinci_spi->base + SPIDELAY); + + count = davinci_spi->count; + data1_reg_val = pdata->cs_hold << SPIDAT1_CSHOLD_SHIFT; + tmp = ~(0x1 << spi->chip_select); + + clear_io_bits(davinci_spi->base + SPIDEF, ~tmp); + + data1_reg_val |= tmp << SPIDAT1_CSNR_SHIFT; + + while ((ioread32(davinci_spi->base + SPIBUF) + & SPIBUF_RXEMPTY_MASK) == 0) + cpu_relax(); + + /* Determine the command to execute READ or WRITE */ + if (t->tx_buf) { + clear_io_bits(davinci_spi->base + SPIINT, SPIINT_MASKALL); + + while (1) { + tx_data = davinci_spi->get_tx(davinci_spi); + + data1_reg_val &= ~(0xFFFF); + data1_reg_val |= (0xFFFF & tx_data); + + buf_val = ioread32(davinci_spi->base + SPIBUF); + if ((buf_val & SPIBUF_TXFULL_MASK) == 0) { + iowrite32(data1_reg_val, + davinci_spi->base + SPIDAT1); + + count--; + } + while (ioread32(davinci_spi->base + SPIBUF) + & SPIBUF_RXEMPTY_MASK) + cpu_relax(); + + /* getting the returned byte */ + if (t->rx_buf) { + buf_val = ioread32(davinci_spi->base + SPIBUF); + davinci_spi->get_rx(buf_val, davinci_spi); + } + if (count <= 0) + break; + } + } else { + if (pdata->poll_mode) { + while (1) { + /* keeps the serial clock going */ + if ((ioread32(davinci_spi->base + SPIBUF) + & SPIBUF_TXFULL_MASK) == 0) + iowrite32(data1_reg_val, + davinci_spi->base + SPIDAT1); + + while (ioread32(davinci_spi->base + SPIBUF) & + SPIBUF_RXEMPTY_MASK) + cpu_relax(); + + flg_val = ioread32(davinci_spi->base + SPIFLG); + buf_val = ioread32(davinci_spi->base + SPIBUF); + + davinci_spi->get_rx(buf_val, davinci_spi); + + count--; + if (count <= 0) + break; + } + } else { /* Receive in Interrupt mode */ + int i; + + for (i = 0; i < davinci_spi->count; i++) { + set_io_bits(davinci_spi->base + SPIINT, + SPIINT_BITERR_INTR + | SPIINT_OVRRUN_INTR + | SPIINT_RX_INTR); + + iowrite32(data1_reg_val, + davinci_spi->base + SPIDAT1); + + while (ioread32(davinci_spi->base + SPIINT) & + SPIINT_RX_INTR) + cpu_relax(); + } + iowrite32((data1_reg_val & 0x0ffcffff), + davinci_spi->base + SPIDAT1); + } + } + + /* + * Check for bit error, desync error,parity error,timeout error and + * receive overflow errors + */ + int_status = ioread32(davinci_spi->base + SPIFLG); + + ret = davinci_spi_check_error(davinci_spi, int_status); + if (ret != 0) + return ret; + + /* SPI Framework maintains the count only in bytes so convert back */ + davinci_spi->count *= conv; + + return t->len; +} + +#define DAVINCI_DMA_DATA_TYPE_S8 0x01 +#define DAVINCI_DMA_DATA_TYPE_S16 0x02 +#define DAVINCI_DMA_DATA_TYPE_S32 0x04 + +static int davinci_spi_bufs_dma(struct spi_device *spi, struct spi_transfer *t) +{ + struct davinci_spi *davinci_spi; + int int_status = 0; + int count, temp_count; + u8 conv = 1; + u8 tmp; + u32 data1_reg_val; + struct davinci_spi_dma *davinci_spi_dma; + int word_len, data_type, ret; + unsigned long tx_reg, rx_reg; + struct davinci_spi_platform_data *pdata; + struct device *sdev; + + davinci_spi = spi_master_get_devdata(spi->master); + pdata = davinci_spi->pdata; + sdev = davinci_spi->bitbang.master->dev.parent; + + davinci_spi_dma = &davinci_spi->dma_channels[spi->chip_select]; + + tx_reg = (unsigned long)davinci_spi->pbase + SPIDAT1; + rx_reg = (unsigned long)davinci_spi->pbase + SPIBUF; + + davinci_spi->tx = t->tx_buf; + davinci_spi->rx = t->rx_buf; + + /* convert len to words based on bits_per_word */ + conv = davinci_spi->slave[spi->chip_select].bytes_per_word; + davinci_spi->count = t->len / conv; + + INIT_COMPLETION(davinci_spi->done); + + init_completion(&davinci_spi_dma->dma_rx_completion); + init_completion(&davinci_spi_dma->dma_tx_completion); + + word_len = conv * 8; + + if (word_len <= 8) + data_type = DAVINCI_DMA_DATA_TYPE_S8; + else if (word_len <= 16) + data_type = DAVINCI_DMA_DATA_TYPE_S16; + else if (word_len <= 32) + data_type = DAVINCI_DMA_DATA_TYPE_S32; + else + return -EINVAL; + + ret = davinci_spi_bufs_prep(spi, davinci_spi); + if (ret) + return ret; + + /* Put delay val if required */ + iowrite32(0 | (pdata->c2tdelay << SPI_C2TDELAY_SHIFT) | + (pdata->t2cdelay << SPI_T2CDELAY_SHIFT), + davinci_spi->base + SPIDELAY); + + count = davinci_spi->count; /* the number of elements */ + data1_reg_val = pdata->cs_hold << SPIDAT1_CSHOLD_SHIFT; + + /* CS default = 0xFF */ + tmp = ~(0x1 << spi->chip_select); + + clear_io_bits(davinci_spi->base + SPIDEF, ~tmp); + + data1_reg_val |= tmp << SPIDAT1_CSNR_SHIFT; + + /* disable all interrupts for dma transfers */ + clear_io_bits(davinci_spi->base + SPIINT, SPIINT_MASKALL); + /* Disable SPI to write configuration bits in SPIDAT */ + clear_io_bits(davinci_spi->base + SPIGCR1, SPIGCR1_SPIENA_MASK); + iowrite32(data1_reg_val, davinci_spi->base + SPIDAT1); + /* Enable SPI */ + set_io_bits(davinci_spi->base + SPIGCR1, SPIGCR1_SPIENA_MASK); + + while ((ioread32(davinci_spi->base + SPIBUF) + & SPIBUF_RXEMPTY_MASK) == 0) + cpu_relax(); + + + if (t->tx_buf) { + t->tx_dma = dma_map_single(&spi->dev, (void *)t->tx_buf, count, + DMA_TO_DEVICE); + if (dma_mapping_error(&spi->dev, t->tx_dma)) { + dev_dbg(sdev, "Unable to DMA map a %d bytes" + " TX buffer\n", count); + return -ENOMEM; + } + temp_count = count; + } else { + /* We need TX clocking for RX transaction */ + t->tx_dma = dma_map_single(&spi->dev, + (void *)davinci_spi->tmp_buf, count + 1, + DMA_TO_DEVICE); + if (dma_mapping_error(&spi->dev, t->tx_dma)) { + dev_dbg(sdev, "Unable to DMA map a %d bytes" + " TX tmp buffer\n", count); + return -ENOMEM; + } + temp_count = count + 1; + } + + edma_set_transfer_params(davinci_spi_dma->dma_tx_channel, + data_type, temp_count, 1, 0, ASYNC); + edma_set_dest(davinci_spi_dma->dma_tx_channel, tx_reg, INCR, W8BIT); + edma_set_src(davinci_spi_dma->dma_tx_channel, t->tx_dma, INCR, W8BIT); + edma_set_src_index(davinci_spi_dma->dma_tx_channel, data_type, 0); + edma_set_dest_index(davinci_spi_dma->dma_tx_channel, 0, 0); + + if (t->rx_buf) { + /* initiate transaction */ + iowrite32(data1_reg_val, davinci_spi->base + SPIDAT1); + + t->rx_dma = dma_map_single(&spi->dev, (void *)t->rx_buf, count, + DMA_FROM_DEVICE); + if (dma_mapping_error(&spi->dev, t->rx_dma)) { + dev_dbg(sdev, "Couldn't DMA map a %d bytes RX buffer\n", + count); + if (t->tx_buf != NULL) + dma_unmap_single(NULL, t->tx_dma, + count, DMA_TO_DEVICE); + return -ENOMEM; + } + edma_set_transfer_params(davinci_spi_dma->dma_rx_channel, + data_type, count, 1, 0, ASYNC); + edma_set_src(davinci_spi_dma->dma_rx_channel, + rx_reg, INCR, W8BIT); + edma_set_dest(davinci_spi_dma->dma_rx_channel, + t->rx_dma, INCR, W8BIT); + edma_set_src_index(davinci_spi_dma->dma_rx_channel, 0, 0); + edma_set_dest_index(davinci_spi_dma->dma_rx_channel, + data_type, 0); + } + + if ((t->tx_buf) || (t->rx_buf)) + edma_start(davinci_spi_dma->dma_tx_channel); + + if (t->rx_buf) + edma_start(davinci_spi_dma->dma_rx_channel); + + if ((t->rx_buf) || (t->tx_buf)) + davinci_spi_set_dma_req(spi, 1); + + if (t->tx_buf) + wait_for_completion_interruptible( + &davinci_spi_dma->dma_tx_completion); + + if (t->rx_buf) + wait_for_completion_interruptible( + &davinci_spi_dma->dma_rx_completion); + + dma_unmap_single(NULL, t->tx_dma, temp_count, DMA_TO_DEVICE); + + if (t->rx_buf) + dma_unmap_single(NULL, t->rx_dma, count, DMA_FROM_DEVICE); + + /* + * Check for bit error, desync error,parity error,timeout error and + * receive overflow errors + */ + int_status = ioread32(davinci_spi->base + SPIFLG); + + ret = davinci_spi_check_error(davinci_spi, int_status); + if (ret != 0) + return ret; + + /* SPI Framework maintains the count only in bytes so convert back */ + davinci_spi->count *= conv; + + return t->len; +} + +/** + * davinci_spi_irq - IRQ handler for DaVinci SPI + * @irq: IRQ number for this SPI Master + * @context_data: structure for SPI Master controller davinci_spi + */ +static irqreturn_t davinci_spi_irq(s32 irq, void *context_data) +{ + struct davinci_spi *davinci_spi = context_data; + u32 int_status, rx_data = 0; + irqreturn_t ret = IRQ_NONE; + + int_status = ioread32(davinci_spi->base + SPIFLG); + + while ((int_status & SPIFLG_RX_INTR_MASK)) { + if (likely(int_status & SPIFLG_RX_INTR_MASK)) { + ret = IRQ_HANDLED; + + rx_data = ioread32(davinci_spi->base + SPIBUF); + davinci_spi->get_rx(rx_data, davinci_spi); + + /* Disable Receive Interrupt */ + iowrite32(~(SPIINT_RX_INTR | SPIINT_TX_INTR), + davinci_spi->base + SPIINT); + } else + (void)davinci_spi_check_error(davinci_spi, int_status); + + int_status = ioread32(davinci_spi->base + SPIFLG); + } + + return ret; +} + +/** + * davinci_spi_probe - probe function for SPI Master Controller + * @pdev: platform_device structure which contains plateform specific data + */ +static int davinci_spi_probe(struct platform_device *pdev) +{ + struct spi_master *master; + struct davinci_spi *davinci_spi; + struct davinci_spi_platform_data *pdata; + struct resource *r, *mem; + resource_size_t dma_rx_chan = SPI_NO_RESOURCE; + resource_size_t dma_tx_chan = SPI_NO_RESOURCE; + resource_size_t dma_eventq = SPI_NO_RESOURCE; + int i = 0, ret = 0; + + pdata = pdev->dev.platform_data; + if (pdata == NULL) { + ret = -ENODEV; + goto err; + } + + master = spi_alloc_master(&pdev->dev, sizeof(struct davinci_spi)); + if (master == NULL) { + ret = -ENOMEM; + goto err; + } + + dev_set_drvdata(&pdev->dev, master); + + davinci_spi = spi_master_get_devdata(master); + if (davinci_spi == NULL) { + ret = -ENOENT; + goto free_master; + } + + r = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (r == NULL) { + ret = -ENOENT; + goto free_master; + } + + davinci_spi->pbase = r->start; + davinci_spi->region_size = resource_size(r); + davinci_spi->pdata = pdata; + + mem = request_mem_region(r->start, davinci_spi->region_size, + pdev->name); + if (mem == NULL) { + ret = -EBUSY; + goto free_master; + } + + davinci_spi->base = (struct davinci_spi_reg __iomem *) + ioremap(r->start, davinci_spi->region_size); + if (davinci_spi->base == NULL) { + ret = -ENOMEM; + goto release_region; + } + + davinci_spi->irq = platform_get_irq(pdev, 0); + if (davinci_spi->irq <= 0) { + ret = -EINVAL; + goto unmap_io; + } + + ret = request_irq(davinci_spi->irq, davinci_spi_irq, IRQF_DISABLED, + dev_name(&pdev->dev), davinci_spi); + if (ret) + goto unmap_io; + + /* Allocate tmp_buf for tx_buf */ + davinci_spi->tmp_buf = kzalloc(SPI_BUFSIZ, GFP_KERNEL); + if (davinci_spi->tmp_buf == NULL) { + ret = -ENOMEM; + goto irq_free; + } + + davinci_spi->bitbang.master = spi_master_get(master); + if (davinci_spi->bitbang.master == NULL) { + ret = -ENODEV; + goto free_tmp_buf; + } + + davinci_spi->clk = clk_get(&pdev->dev, NULL); + if (IS_ERR(davinci_spi->clk)) { + ret = -ENODEV; + goto put_master; + } + clk_enable(davinci_spi->clk); + + + master->bus_num = pdev->id; + master->num_chipselect = pdata->num_chipselect; + master->setup = davinci_spi_setup; + master->cleanup = davinci_spi_cleanup; + + davinci_spi->bitbang.chipselect = davinci_spi_chipselect; + davinci_spi->bitbang.setup_transfer = davinci_spi_setup_transfer; + + davinci_spi->version = pdata->version; + use_dma = pdata->use_dma; + + davinci_spi->bitbang.flags = SPI_NO_CS | SPI_LSB_FIRST | SPI_LOOP; + if (davinci_spi->version == SPI_VERSION_2) + davinci_spi->bitbang.flags |= SPI_READY; + + if (use_dma) { + r = platform_get_resource(pdev, IORESOURCE_DMA, 0); + if (r) + dma_rx_chan = r->start; + r = platform_get_resource(pdev, IORESOURCE_DMA, 1); + if (r) + dma_tx_chan = r->start; + r = platform_get_resource(pdev, IORESOURCE_DMA, 2); + if (r) + dma_eventq = r->start; + } + + if (!use_dma || + dma_rx_chan == SPI_NO_RESOURCE || + dma_tx_chan == SPI_NO_RESOURCE || + dma_eventq == SPI_NO_RESOURCE) { + davinci_spi->bitbang.txrx_bufs = davinci_spi_bufs_pio; + use_dma = 0; + } else { + davinci_spi->bitbang.txrx_bufs = davinci_spi_bufs_dma; + davinci_spi->dma_channels = kzalloc(master->num_chipselect + * sizeof(struct davinci_spi_dma), GFP_KERNEL); + if (davinci_spi->dma_channels == NULL) { + ret = -ENOMEM; + goto free_clk; + } + + for (i = 0; i < master->num_chipselect; i++) { + davinci_spi->dma_channels[i].dma_rx_channel = -1; + davinci_spi->dma_channels[i].dma_rx_sync_dev = + dma_rx_chan; + davinci_spi->dma_channels[i].dma_tx_channel = -1; + davinci_spi->dma_channels[i].dma_tx_sync_dev = + dma_tx_chan; + davinci_spi->dma_channels[i].eventq = dma_eventq; + } + dev_info(&pdev->dev, "DaVinci SPI driver in EDMA mode\n" + "Using RX channel = %d , TX channel = %d and " + "event queue = %d", dma_rx_chan, dma_tx_chan, + dma_eventq); + } + + davinci_spi->get_rx = davinci_spi_rx_buf_u8; + davinci_spi->get_tx = davinci_spi_tx_buf_u8; + + init_completion(&davinci_spi->done); + + /* Reset In/OUT SPI module */ + iowrite32(0, davinci_spi->base + SPIGCR0); + udelay(100); + iowrite32(1, davinci_spi->base + SPIGCR0); + + /* Clock internal */ + if (davinci_spi->pdata->clk_internal) + set_io_bits(davinci_spi->base + SPIGCR1, + SPIGCR1_CLKMOD_MASK); + else + clear_io_bits(davinci_spi->base + SPIGCR1, + SPIGCR1_CLKMOD_MASK); + + /* master mode default */ + set_io_bits(davinci_spi->base + SPIGCR1, SPIGCR1_MASTER_MASK); + + if (davinci_spi->pdata->intr_level) + iowrite32(SPI_INTLVL_1, davinci_spi->base + SPILVL); + else + iowrite32(SPI_INTLVL_0, davinci_spi->base + SPILVL); + + ret = spi_bitbang_start(&davinci_spi->bitbang); + if (ret) + goto free_clk; + + dev_info(&pdev->dev, "Controller at 0x%p \n", davinci_spi->base); + + if (!pdata->poll_mode) + dev_info(&pdev->dev, "Operating in interrupt mode" + " using IRQ %d\n", davinci_spi->irq); + + return ret; + +free_clk: + clk_disable(davinci_spi->clk); + clk_put(davinci_spi->clk); +put_master: + spi_master_put(master); +free_tmp_buf: + kfree(davinci_spi->tmp_buf); +irq_free: + free_irq(davinci_spi->irq, davinci_spi); +unmap_io: + iounmap(davinci_spi->base); +release_region: + release_mem_region(davinci_spi->pbase, davinci_spi->region_size); +free_master: + kfree(master); +err: + return ret; +} + +/** + * davinci_spi_remove - remove function for SPI Master Controller + * @pdev: platform_device structure which contains plateform specific data + * + * This function will do the reverse action of davinci_spi_probe function + * It will free the IRQ and SPI controller's memory region. + * It will also call spi_bitbang_stop to destroy the work queue which was + * created by spi_bitbang_start. + */ +static int __exit davinci_spi_remove(struct platform_device *pdev) +{ + struct davinci_spi *davinci_spi; + struct spi_master *master; + + master = dev_get_drvdata(&pdev->dev); + davinci_spi = spi_master_get_devdata(master); + + spi_bitbang_stop(&davinci_spi->bitbang); + + clk_disable(davinci_spi->clk); + clk_put(davinci_spi->clk); + spi_master_put(master); + kfree(davinci_spi->tmp_buf); + free_irq(davinci_spi->irq, davinci_spi); + iounmap(davinci_spi->base); + release_mem_region(davinci_spi->pbase, davinci_spi->region_size); + + return 0; +} + +static struct platform_driver davinci_spi_driver = { + .driver.name = "spi_davinci", + .remove = __exit_p(davinci_spi_remove), +}; + +static int __init davinci_spi_init(void) +{ + return platform_driver_probe(&davinci_spi_driver, davinci_spi_probe); +} +module_init(davinci_spi_init); + +static void __exit davinci_spi_exit(void) +{ + platform_driver_unregister(&davinci_spi_driver); +} +module_exit(davinci_spi_exit); + +MODULE_DESCRIPTION("TI DaVinci SPI Master Controller Driver"); +MODULE_LICENSE("GPL"); -- cgit v0.10.2 From 68ea2d82c3671d2eccb600e6871fcbec1cac7fca Mon Sep 17 00:00:00 2001 From: "Robert P. J. Day" Date: Wed, 20 Jan 2010 13:49:44 -0700 Subject: spi: Fix reversed args to time_before() in Freescale stmp driver. Signed-off-by: Robert P. J. Day Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_stmp.c b/drivers/spi/spi_stmp.c index 2552bb3..fadff76 100644 --- a/drivers/spi/spi_stmp.c +++ b/drivers/spi/spi_stmp.c @@ -76,7 +76,7 @@ struct stmp_spi { break; \ } \ cpu_relax(); \ - } while (time_before(end_jiffies, jiffies)); \ + } while (time_before(jiffies, end_jiffies)); \ succeeded; \ }) -- cgit v0.10.2 From 631e61b7ca12ef14c834f99f8948e410c539f585 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?M=C3=A1rton=20N=C3=A9meth?= Date: Wed, 20 Jan 2010 13:49:44 -0700 Subject: spi: make Open Firmware device id constant MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The match_table field of the struct of_device_id is constant in so it is worth to make the initialization data constant. The semantic match that finds this kind of pattern is as follows: (http://coccinelle.lip6.fr/) // @r@ disable decl_init,const_decl_init; identifier I1, I2, x; @@ struct I1 { ... const struct I2 *x; ... }; @s@ identifier r.I1, y; identifier r.x, E; @@ struct I1 y = { .x = E, }; @c@ identifier r.I2; identifier s.E; @@ const struct I2 E[] = ... ; @depends on !c@ identifier r.I2; identifier s.E; @@ + const struct I2 E[] = ...; // Signed-off-by: Márton Németh Cc: Julia Lawall Cc: cocci@diku.dk Signed-off-by: Grant Likely diff --git a/drivers/spi/mpc52xx_psc_spi.c b/drivers/spi/mpc52xx_psc_spi.c index f50c81d..0474786 100644 --- a/drivers/spi/mpc52xx_psc_spi.c +++ b/drivers/spi/mpc52xx_psc_spi.c @@ -503,7 +503,7 @@ static int __exit mpc52xx_psc_spi_of_remove(struct of_device *op) return mpc52xx_psc_spi_do_remove(&op->dev); } -static struct of_device_id mpc52xx_psc_spi_of_match[] = { +static const struct of_device_id mpc52xx_psc_spi_of_match[] = { { .compatible = "fsl,mpc5200-psc-spi", }, { .compatible = "mpc5200-psc-spi", }, /* old */ {} diff --git a/drivers/spi/mpc52xx_spi.c b/drivers/spi/mpc52xx_spi.c index 45bfe64..6eab465 100644 --- a/drivers/spi/mpc52xx_spi.c +++ b/drivers/spi/mpc52xx_spi.c @@ -550,7 +550,7 @@ static int __devexit mpc52xx_spi_remove(struct of_device *op) return 0; } -static struct of_device_id mpc52xx_spi_match[] __devinitdata = { +static const struct of_device_id mpc52xx_spi_match[] __devinitconst = { { .compatible = "fsl,mpc5200-spi", }, {} }; diff --git a/drivers/spi/spi_ppc4xx.c b/drivers/spi/spi_ppc4xx.c index 140a18d..6d8d402 100644 --- a/drivers/spi/spi_ppc4xx.c +++ b/drivers/spi/spi_ppc4xx.c @@ -578,7 +578,7 @@ static int __exit spi_ppc4xx_of_remove(struct of_device *op) return 0; } -static struct of_device_id spi_ppc4xx_of_match[] = { +static const struct of_device_id spi_ppc4xx_of_match[] = { { .compatible = "ibm,ppc4xx-spi", }, {}, }; diff --git a/drivers/spi/xilinx_spi_of.c b/drivers/spi/xilinx_spi_of.c index 71dc3ad..ed34a8d 100644 --- a/drivers/spi/xilinx_spi_of.c +++ b/drivers/spi/xilinx_spi_of.c @@ -99,7 +99,7 @@ static int __exit xilinx_spi_of_remove(struct of_device *op) return xilinx_spi_remove(op); } -static struct of_device_id xilinx_spi_of_match[] = { +static const struct of_device_id xilinx_spi_of_match[] = { { .compatible = "xlnx,xps-spi-2.00.a", }, { .compatible = "xlnx,xps-spi-2.00.b", }, {} -- cgit v0.10.2 From 34b8c66173666025020e3a6f8d4a5c238b19cde5 Mon Sep 17 00:00:00 2001 From: Steven King Date: Wed, 20 Jan 2010 13:49:44 -0700 Subject: spi: Add Freescale/Motorola Coldfire QSPI driver Add support for the QSPI controller found some on Freescale/Motorola Coldfire MCUs. Full duplex, active high cs, spi modes 0-3 and word sizes 8-16 bits are supported. The hardware drives the MISO, MOSI and SCLK lines, but the chip selects are managed via GPIO and must be configured by the board code. The QSPI controller has an 80 byte buffer which allows us to transfer up to 16 words at a time. For transfers longer than 16 words, we split the buffer in half so we can update in one half while the controller is operating on the other half. Interrupt latencies then ultimately limits our sustained thru-put to something less than half the maximum speed supported by the part. Signed-off-by: Steven King Signed-off-by: Grant Likely diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig index 82dabbc..28becdd 100644 --- a/drivers/spi/Kconfig +++ b/drivers/spi/Kconfig @@ -100,6 +100,16 @@ config SPI_BUTTERFLY inexpensive battery powered microcontroller evaluation board. This same cable can be used to flash new firmware. +config SPI_COLDFIRE_QSPI + tristate "Freescale Coldfire QSPI controller" + depends on (M520x || M523x || M5249 || M527x || M528x || M532x) + help + This enables support for the Coldfire QSPI controller in master + mode. + + This driver can also be built as a module. If so, the module + will be called coldfire_qspi. + config SPI_DAVINCI tristate "SPI controller driver for DaVinci/DA8xx SoC's" depends on SPI_MASTER && ARCH_DAVINCI diff --git a/drivers/spi/Makefile b/drivers/spi/Makefile index 84b490f..d3a65ab 100644 --- a/drivers/spi/Makefile +++ b/drivers/spi/Makefile @@ -16,6 +16,7 @@ obj-$(CONFIG_SPI_BFIN) += spi_bfin5xx.o obj-$(CONFIG_SPI_BITBANG) += spi_bitbang.o obj-$(CONFIG_SPI_AU1550) += au1550_spi.o obj-$(CONFIG_SPI_BUTTERFLY) += spi_butterfly.o +obj-$(CONFIG_SPI_COLDFIRE_QSPI) += coldfire_qspi.o obj-$(CONFIG_SPI_DAVINCI) += davinci_spi.o obj-$(CONFIG_SPI_DESIGNWARE) += dw_spi.o obj-$(CONFIG_SPI_DW_PCI) += dw_spi_pci.o diff --git a/drivers/spi/coldfire_qspi.c b/drivers/spi/coldfire_qspi.c new file mode 100644 index 0000000..59be3ef --- /dev/null +++ b/drivers/spi/coldfire_qspi.c @@ -0,0 +1,640 @@ +/* + * Freescale/Motorola Coldfire Queued SPI driver + * + * Copyright 2010 Steven King + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA + * +*/ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +#define DRIVER_NAME "mcfqspi" + +#define MCFQSPI_BUSCLK (MCF_BUSCLK / 2) + +#define MCFQSPI_QMR 0x00 +#define MCFQSPI_QMR_MSTR 0x8000 +#define MCFQSPI_QMR_CPOL 0x0200 +#define MCFQSPI_QMR_CPHA 0x0100 +#define MCFQSPI_QDLYR 0x04 +#define MCFQSPI_QDLYR_SPE 0x8000 +#define MCFQSPI_QWR 0x08 +#define MCFQSPI_QWR_HALT 0x8000 +#define MCFQSPI_QWR_WREN 0x4000 +#define MCFQSPI_QWR_CSIV 0x1000 +#define MCFQSPI_QIR 0x0C +#define MCFQSPI_QIR_WCEFB 0x8000 +#define MCFQSPI_QIR_ABRTB 0x4000 +#define MCFQSPI_QIR_ABRTL 0x1000 +#define MCFQSPI_QIR_WCEFE 0x0800 +#define MCFQSPI_QIR_ABRTE 0x0400 +#define MCFQSPI_QIR_SPIFE 0x0100 +#define MCFQSPI_QIR_WCEF 0x0008 +#define MCFQSPI_QIR_ABRT 0x0004 +#define MCFQSPI_QIR_SPIF 0x0001 +#define MCFQSPI_QAR 0x010 +#define MCFQSPI_QAR_TXBUF 0x00 +#define MCFQSPI_QAR_RXBUF 0x10 +#define MCFQSPI_QAR_CMDBUF 0x20 +#define MCFQSPI_QDR 0x014 +#define MCFQSPI_QCR 0x014 +#define MCFQSPI_QCR_CONT 0x8000 +#define MCFQSPI_QCR_BITSE 0x4000 +#define MCFQSPI_QCR_DT 0x2000 + +struct mcfqspi { + void __iomem *iobase; + int irq; + struct clk *clk; + struct mcfqspi_cs_control *cs_control; + + wait_queue_head_t waitq; + + struct work_struct work; + struct workqueue_struct *workq; + spinlock_t lock; + struct list_head msgq; +}; + +static void mcfqspi_wr_qmr(struct mcfqspi *mcfqspi, u16 val) +{ + writew(val, mcfqspi->iobase + MCFQSPI_QMR); +} + +static void mcfqspi_wr_qdlyr(struct mcfqspi *mcfqspi, u16 val) +{ + writew(val, mcfqspi->iobase + MCFQSPI_QDLYR); +} + +static u16 mcfqspi_rd_qdlyr(struct mcfqspi *mcfqspi) +{ + return readw(mcfqspi->iobase + MCFQSPI_QDLYR); +} + +static void mcfqspi_wr_qwr(struct mcfqspi *mcfqspi, u16 val) +{ + writew(val, mcfqspi->iobase + MCFQSPI_QWR); +} + +static void mcfqspi_wr_qir(struct mcfqspi *mcfqspi, u16 val) +{ + writew(val, mcfqspi->iobase + MCFQSPI_QIR); +} + +static void mcfqspi_wr_qar(struct mcfqspi *mcfqspi, u16 val) +{ + writew(val, mcfqspi->iobase + MCFQSPI_QAR); +} + +static void mcfqspi_wr_qdr(struct mcfqspi *mcfqspi, u16 val) +{ + writew(val, mcfqspi->iobase + MCFQSPI_QDR); +} + +static u16 mcfqspi_rd_qdr(struct mcfqspi *mcfqspi) +{ + return readw(mcfqspi->iobase + MCFQSPI_QDR); +} + +static void mcfqspi_cs_select(struct mcfqspi *mcfqspi, u8 chip_select, + bool cs_high) +{ + mcfqspi->cs_control->select(mcfqspi->cs_control, chip_select, cs_high); +} + +static void mcfqspi_cs_deselect(struct mcfqspi *mcfqspi, u8 chip_select, + bool cs_high) +{ + mcfqspi->cs_control->deselect(mcfqspi->cs_control, chip_select, cs_high); +} + +static int mcfqspi_cs_setup(struct mcfqspi *mcfqspi) +{ + return (mcfqspi->cs_control && mcfqspi->cs_control->setup) ? + mcfqspi->cs_control->setup(mcfqspi->cs_control) : 0; +} + +static void mcfqspi_cs_teardown(struct mcfqspi *mcfqspi) +{ + if (mcfqspi->cs_control && mcfqspi->cs_control->teardown) + mcfqspi->cs_control->teardown(mcfqspi->cs_control); +} + +static u8 mcfqspi_qmr_baud(u32 speed_hz) +{ + return clamp((MCFQSPI_BUSCLK + speed_hz - 1) / speed_hz, 2u, 255u); +} + +static bool mcfqspi_qdlyr_spe(struct mcfqspi *mcfqspi) +{ + return mcfqspi_rd_qdlyr(mcfqspi) & MCFQSPI_QDLYR_SPE; +} + +static irqreturn_t mcfqspi_irq_handler(int this_irq, void *dev_id) +{ + struct mcfqspi *mcfqspi = dev_id; + + /* clear interrupt */ + mcfqspi_wr_qir(mcfqspi, MCFQSPI_QIR_SPIFE | MCFQSPI_QIR_SPIF); + wake_up(&mcfqspi->waitq); + + return IRQ_HANDLED; +} + +static void mcfqspi_transfer_msg8(struct mcfqspi *mcfqspi, unsigned count, + const u8 *txbuf, u8 *rxbuf) +{ + unsigned i, n, offset = 0; + + n = min(count, 16u); + + mcfqspi_wr_qar(mcfqspi, MCFQSPI_QAR_CMDBUF); + for (i = 0; i < n; ++i) + mcfqspi_wr_qdr(mcfqspi, MCFQSPI_QCR_BITSE); + + mcfqspi_wr_qar(mcfqspi, MCFQSPI_QAR_TXBUF); + if (txbuf) + for (i = 0; i < n; ++i) + mcfqspi_wr_qdr(mcfqspi, *txbuf++); + else + for (i = 0; i < count; ++i) + mcfqspi_wr_qdr(mcfqspi, 0); + + count -= n; + if (count) { + u16 qwr = 0xf08; + mcfqspi_wr_qwr(mcfqspi, 0x700); + mcfqspi_wr_qdlyr(mcfqspi, MCFQSPI_QDLYR_SPE); + + do { + wait_event(mcfqspi->waitq, !mcfqspi_qdlyr_spe(mcfqspi)); + mcfqspi_wr_qwr(mcfqspi, qwr); + mcfqspi_wr_qdlyr(mcfqspi, MCFQSPI_QDLYR_SPE); + if (rxbuf) { + mcfqspi_wr_qar(mcfqspi, + MCFQSPI_QAR_RXBUF + offset); + for (i = 0; i < 8; ++i) + *rxbuf++ = mcfqspi_rd_qdr(mcfqspi); + } + n = min(count, 8u); + if (txbuf) { + mcfqspi_wr_qar(mcfqspi, + MCFQSPI_QAR_TXBUF + offset); + for (i = 0; i < n; ++i) + mcfqspi_wr_qdr(mcfqspi, *txbuf++); + } + qwr = (offset ? 0x808 : 0) + ((n - 1) << 8); + offset ^= 8; + count -= n; + } while (count); + wait_event(mcfqspi->waitq, !mcfqspi_qdlyr_spe(mcfqspi)); + mcfqspi_wr_qwr(mcfqspi, qwr); + mcfqspi_wr_qdlyr(mcfqspi, MCFQSPI_QDLYR_SPE); + if (rxbuf) { + mcfqspi_wr_qar(mcfqspi, MCFQSPI_QAR_RXBUF + offset); + for (i = 0; i < 8; ++i) + *rxbuf++ = mcfqspi_rd_qdr(mcfqspi); + offset ^= 8; + } + } else { + mcfqspi_wr_qwr(mcfqspi, (n - 1) << 8); + mcfqspi_wr_qdlyr(mcfqspi, MCFQSPI_QDLYR_SPE); + } + wait_event(mcfqspi->waitq, !mcfqspi_qdlyr_spe(mcfqspi)); + if (rxbuf) { + mcfqspi_wr_qar(mcfqspi, MCFQSPI_QAR_RXBUF + offset); + for (i = 0; i < n; ++i) + *rxbuf++ = mcfqspi_rd_qdr(mcfqspi); + } +} + +static void mcfqspi_transfer_msg16(struct mcfqspi *mcfqspi, unsigned count, + const u16 *txbuf, u16 *rxbuf) +{ + unsigned i, n, offset = 0; + + n = min(count, 16u); + + mcfqspi_wr_qar(mcfqspi, MCFQSPI_QAR_CMDBUF); + for (i = 0; i < n; ++i) + mcfqspi_wr_qdr(mcfqspi, MCFQSPI_QCR_BITSE); + + mcfqspi_wr_qar(mcfqspi, MCFQSPI_QAR_TXBUF); + if (txbuf) + for (i = 0; i < n; ++i) + mcfqspi_wr_qdr(mcfqspi, *txbuf++); + else + for (i = 0; i < count; ++i) + mcfqspi_wr_qdr(mcfqspi, 0); + + count -= n; + if (count) { + u16 qwr = 0xf08; + mcfqspi_wr_qwr(mcfqspi, 0x700); + mcfqspi_wr_qdlyr(mcfqspi, MCFQSPI_QDLYR_SPE); + + do { + wait_event(mcfqspi->waitq, !mcfqspi_qdlyr_spe(mcfqspi)); + mcfqspi_wr_qwr(mcfqspi, qwr); + mcfqspi_wr_qdlyr(mcfqspi, MCFQSPI_QDLYR_SPE); + if (rxbuf) { + mcfqspi_wr_qar(mcfqspi, + MCFQSPI_QAR_RXBUF + offset); + for (i = 0; i < 8; ++i) + *rxbuf++ = mcfqspi_rd_qdr(mcfqspi); + } + n = min(count, 8u); + if (txbuf) { + mcfqspi_wr_qar(mcfqspi, + MCFQSPI_QAR_TXBUF + offset); + for (i = 0; i < n; ++i) + mcfqspi_wr_qdr(mcfqspi, *txbuf++); + } + qwr = (offset ? 0x808 : 0x000) + ((n - 1) << 8); + offset ^= 8; + count -= n; + } while (count); + wait_event(mcfqspi->waitq, !mcfqspi_qdlyr_spe(mcfqspi)); + mcfqspi_wr_qwr(mcfqspi, qwr); + mcfqspi_wr_qdlyr(mcfqspi, MCFQSPI_QDLYR_SPE); + if (rxbuf) { + mcfqspi_wr_qar(mcfqspi, MCFQSPI_QAR_RXBUF + offset); + for (i = 0; i < 8; ++i) + *rxbuf++ = mcfqspi_rd_qdr(mcfqspi); + offset ^= 8; + } + } else { + mcfqspi_wr_qwr(mcfqspi, (n - 1) << 8); + mcfqspi_wr_qdlyr(mcfqspi, MCFQSPI_QDLYR_SPE); + } + wait_event(mcfqspi->waitq, !mcfqspi_qdlyr_spe(mcfqspi)); + if (rxbuf) { + mcfqspi_wr_qar(mcfqspi, MCFQSPI_QAR_RXBUF + offset); + for (i = 0; i < n; ++i) + *rxbuf++ = mcfqspi_rd_qdr(mcfqspi); + } +} + +static void mcfqspi_work(struct work_struct *work) +{ + struct mcfqspi *mcfqspi = container_of(work, struct mcfqspi, work); + unsigned long flags; + + spin_lock_irqsave(&mcfqspi->lock, flags); + while (!list_empty(&mcfqspi->msgq)) { + struct spi_message *msg; + struct spi_device *spi; + struct spi_transfer *xfer; + int status = 0; + + msg = container_of(mcfqspi->msgq.next, struct spi_message, + queue); + + list_del_init(&mcfqspi->msgq); + spin_unlock_irqrestore(&mcfqspi->lock, flags); + + spi = msg->spi; + + list_for_each_entry(xfer, &msg->transfers, transfer_list) { + bool cs_high = spi->mode & SPI_CS_HIGH; + u16 qmr = MCFQSPI_QMR_MSTR; + + if (xfer->bits_per_word) + qmr |= xfer->bits_per_word << 10; + else + qmr |= spi->bits_per_word << 10; + if (spi->mode & SPI_CPHA) + qmr |= MCFQSPI_QMR_CPHA; + if (spi->mode & SPI_CPOL) + qmr |= MCFQSPI_QMR_CPOL; + if (xfer->speed_hz) + qmr |= mcfqspi_qmr_baud(xfer->speed_hz); + else + qmr |= mcfqspi_qmr_baud(spi->max_speed_hz); + mcfqspi_wr_qmr(mcfqspi, qmr); + + mcfqspi_cs_select(mcfqspi, spi->chip_select, cs_high); + + mcfqspi_wr_qir(mcfqspi, MCFQSPI_QIR_SPIFE); + if ((xfer->bits_per_word ? xfer->bits_per_word : + spi->bits_per_word) == 8) + mcfqspi_transfer_msg8(mcfqspi, xfer->len, + xfer->tx_buf, + xfer->rx_buf); + else + mcfqspi_transfer_msg16(mcfqspi, xfer->len / 2, + xfer->tx_buf, + xfer->rx_buf); + mcfqspi_wr_qir(mcfqspi, 0); + + if (xfer->delay_usecs) + udelay(xfer->delay_usecs); + if (xfer->cs_change) { + if (!list_is_last(&xfer->transfer_list, + &msg->transfers)) + mcfqspi_cs_deselect(mcfqspi, + spi->chip_select, + cs_high); + } else { + if (list_is_last(&xfer->transfer_list, + &msg->transfers)) + mcfqspi_cs_deselect(mcfqspi, + spi->chip_select, + cs_high); + } + msg->actual_length += xfer->len; + } + msg->status = status; + msg->complete(msg->context); + + spin_lock_irqsave(&mcfqspi->lock, flags); + } + spin_unlock_irqrestore(&mcfqspi->lock, flags); +} + +static int mcfqspi_transfer(struct spi_device *spi, struct spi_message *msg) +{ + struct mcfqspi *mcfqspi; + struct spi_transfer *xfer; + unsigned long flags; + + mcfqspi = spi_master_get_devdata(spi->master); + + list_for_each_entry(xfer, &msg->transfers, transfer_list) { + if (xfer->bits_per_word && ((xfer->bits_per_word < 8) + || (xfer->bits_per_word > 16))) { + dev_dbg(&spi->dev, + "%d bits per word is not supported\n", + xfer->bits_per_word); + goto fail; + } + if (xfer->speed_hz) { + u32 real_speed = MCFQSPI_BUSCLK / + mcfqspi_qmr_baud(xfer->speed_hz); + if (real_speed != xfer->speed_hz) + dev_dbg(&spi->dev, + "using speed %d instead of %d\n", + real_speed, xfer->speed_hz); + } + } + msg->status = -EINPROGRESS; + msg->actual_length = 0; + + spin_lock_irqsave(&mcfqspi->lock, flags); + list_add_tail(&msg->queue, &mcfqspi->msgq); + queue_work(mcfqspi->workq, &mcfqspi->work); + spin_unlock_irqrestore(&mcfqspi->lock, flags); + + return 0; +fail: + msg->status = -EINVAL; + return -EINVAL; +} + +static int mcfqspi_setup(struct spi_device *spi) +{ + if ((spi->bits_per_word < 8) || (spi->bits_per_word > 16)) { + dev_dbg(&spi->dev, "%d bits per word is not supported\n", + spi->bits_per_word); + return -EINVAL; + } + if (spi->chip_select >= spi->master->num_chipselect) { + dev_dbg(&spi->dev, "%d chip select is out of range\n", + spi->chip_select); + return -EINVAL; + } + + mcfqspi_cs_deselect(spi_master_get_devdata(spi->master), + spi->chip_select, spi->mode & SPI_CS_HIGH); + + dev_dbg(&spi->dev, + "bits per word %d, chip select %d, speed %d KHz\n", + spi->bits_per_word, spi->chip_select, + (MCFQSPI_BUSCLK / mcfqspi_qmr_baud(spi->max_speed_hz)) + / 1000); + + return 0; +} + +static int __devinit mcfqspi_probe(struct platform_device *pdev) +{ + struct spi_master *master; + struct mcfqspi *mcfqspi; + struct resource *res; + struct mcfqspi_platform_data *pdata; + int status; + + master = spi_alloc_master(&pdev->dev, sizeof(*mcfqspi)); + if (master == NULL) { + dev_dbg(&pdev->dev, "spi_alloc_master failed\n"); + return -ENOMEM; + } + + mcfqspi = spi_master_get_devdata(master); + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!res) { + dev_dbg(&pdev->dev, "platform_get_resource failed\n"); + status = -ENXIO; + goto fail0; + } + + if (!request_mem_region(res->start, resource_size(res), pdev->name)) { + dev_dbg(&pdev->dev, "request_mem_region failed\n"); + status = -EBUSY; + goto fail0; + } + + mcfqspi->iobase = ioremap(res->start, resource_size(res)); + if (!mcfqspi->iobase) { + dev_dbg(&pdev->dev, "ioremap failed\n"); + status = -ENOMEM; + goto fail1; + } + + mcfqspi->irq = platform_get_irq(pdev, 0); + if (mcfqspi->irq < 0) { + dev_dbg(&pdev->dev, "platform_get_irq failed\n"); + status = -ENXIO; + goto fail2; + } + + status = request_irq(mcfqspi->irq, mcfqspi_irq_handler, IRQF_DISABLED, + pdev->name, mcfqspi); + if (status) { + dev_dbg(&pdev->dev, "request_irq failed\n"); + goto fail2; + } + + mcfqspi->clk = clk_get(&pdev->dev, "qspi_clk"); + if (IS_ERR(mcfqspi->clk)) { + dev_dbg(&pdev->dev, "clk_get failed\n"); + status = PTR_ERR(mcfqspi->clk); + goto fail3; + } + clk_enable(mcfqspi->clk); + + mcfqspi->workq = create_singlethread_workqueue(dev_name(master->dev.parent)); + if (!mcfqspi->workq) { + dev_dbg(&pdev->dev, "create_workqueue failed\n"); + status = -ENOMEM; + goto fail4; + } + INIT_WORK(&mcfqspi->work, mcfqspi_work); + spin_lock_init(&mcfqspi->lock); + INIT_LIST_HEAD(&mcfqspi->msgq); + init_waitqueue_head(&mcfqspi->waitq); + + pdata = pdev->dev.platform_data; + if (!pdata) { + dev_dbg(&pdev->dev, "platform data is missing\n"); + goto fail5; + } + master->bus_num = pdata->bus_num; + master->num_chipselect = pdata->num_chipselect; + + mcfqspi->cs_control = pdata->cs_control; + status = mcfqspi_cs_setup(mcfqspi); + if (status) { + dev_dbg(&pdev->dev, "error initializing cs_control\n"); + goto fail5; + } + + master->mode_bits = SPI_CS_HIGH | SPI_CPOL | SPI_CPHA; + master->setup = mcfqspi_setup; + master->transfer = mcfqspi_transfer; + + platform_set_drvdata(pdev, master); + + status = spi_register_master(master); + if (status) { + dev_dbg(&pdev->dev, "spi_register_master failed\n"); + goto fail6; + } + dev_info(&pdev->dev, "Coldfire QSPI bus driver\n"); + + return 0; + +fail6: + mcfqspi_cs_teardown(mcfqspi); +fail5: + destroy_workqueue(mcfqspi->workq); +fail4: + clk_disable(mcfqspi->clk); + clk_put(mcfqspi->clk); +fail3: + free_irq(mcfqspi->irq, mcfqspi); +fail2: + iounmap(mcfqspi->iobase); +fail1: + release_mem_region(res->start, resource_size(res)); +fail0: + spi_master_put(master); + + dev_dbg(&pdev->dev, "Coldfire QSPI probe failed\n"); + + return status; +} + +static int __devexit mcfqspi_remove(struct platform_device *pdev) +{ + struct spi_master *master = platform_get_drvdata(pdev); + struct mcfqspi *mcfqspi = spi_master_get_devdata(master); + struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + + /* disable the hardware (set the baud rate to 0) */ + mcfqspi_wr_qmr(mcfqspi, MCFQSPI_QMR_MSTR); + + platform_set_drvdata(pdev, NULL); + mcfqspi_cs_teardown(mcfqspi); + destroy_workqueue(mcfqspi->workq); + clk_disable(mcfqspi->clk); + clk_put(mcfqspi->clk); + free_irq(mcfqspi->irq, mcfqspi); + iounmap(mcfqspi->iobase); + release_mem_region(res->start, resource_size(res)); + spi_unregister_master(master); + spi_master_put(master); + + return 0; +} + +#ifdef CONFIG_PM + +static int mcfqspi_suspend(struct device *dev) +{ + struct mcfqspi *mcfqspi = platform_get_drvdata(to_platform_device(dev)); + + clk_disable(mcfqspi->clk); + + return 0; +} + +static int mcfqspi_resume(struct device *dev) +{ + struct mcfqspi *mcfqspi = platform_get_drvdata(to_platform_device(dev)); + + clk_enable(mcfqspi->clk); + + return 0; +} + +static struct dev_pm_ops mcfqspi_dev_pm_ops = { + .suspend = mcfqspi_suspend, + .resume = mcfqspi_resume, +}; + +#define MCFQSPI_DEV_PM_OPS (&mcfqspi_dev_pm_ops) +#else +#define MCFQSPI_DEV_PM_OPS NULL +#endif + +static struct platform_driver mcfqspi_driver = { + .driver.name = DRIVER_NAME, + .driver.owner = THIS_MODULE, + .driver.pm = MCFQSPI_DEV_PM_OPS, + .remove = __devexit_p(mcfqspi_remove), +}; + +static int __init mcfqspi_init(void) +{ + return platform_driver_probe(&mcfqspi_driver, mcfqspi_probe); +} +module_init(mcfqspi_init); + +static void __exit mcfqspi_exit(void) +{ + platform_driver_unregister(&mcfqspi_driver); +} +module_exit(mcfqspi_exit); + +MODULE_AUTHOR("Steven King "); +MODULE_DESCRIPTION("Coldfire QSPI Controller Driver"); +MODULE_LICENSE("GPL"); +MODULE_ALIAS("platform:" DRIVER_NAME); -- cgit v0.10.2 From e9a172f074ba85de144e63b0786c7c5c5ba93c3a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Uwe=20Kleine-K=C3=B6nig?= Date: Wed, 20 Jan 2010 13:49:44 -0700 Subject: spi/mpc8xxx: don't check platform_get_irq's return value against zero MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit platform_get_irq returns -ENXIO on failure, so !irq was probably always true. Make irq a signed variable and compare irq <= 0. Note that a return value of zero is still handled as error even though this could mean irq0. This is a followup to 305b3228f9ff4d59f49e6d34a7034d44ee8ce2f0 that changed the return value of platform_get_irq from 0 to -ENXIO on error. Signed-off-by: Uwe Kleine-König Acked-by: Anton Vorontsov Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_mpc8xxx.c b/drivers/spi/spi_mpc8xxx.c index 1fb2a6e..08065fb 100644 --- a/drivers/spi/spi_mpc8xxx.c +++ b/drivers/spi/spi_mpc8xxx.c @@ -1328,7 +1328,7 @@ static struct of_platform_driver of_mpc8xxx_spi_driver = { static int __devinit plat_mpc8xxx_spi_probe(struct platform_device *pdev) { struct resource *mem; - unsigned int irq; + int irq; struct spi_master *master; if (!pdev->dev.platform_data) @@ -1339,7 +1339,7 @@ static int __devinit plat_mpc8xxx_spi_probe(struct platform_device *pdev) return -EINVAL; irq = platform_get_irq(pdev, 0); - if (!irq) + if (irq <= 0) return -EINVAL; master = mpc8xxx_spi_probe(&pdev->dev, mem, irq); -- cgit v0.10.2 From ad7de729c60380a48844f885f37451158169c50d Mon Sep 17 00:00:00 2001 From: Jassi Brar Date: Wed, 20 Jan 2010 13:49:44 -0700 Subject: spi/s3c64xx: Rename s3c64xx_spi_cntrlr_info Rename 'struct s3c64xx_spi_cntrlr_info' to lesser wordy 'struct s3c64xx_spi_info' Signed-off-by: Jassi Brar Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_s3c64xx.c b/drivers/spi/spi_s3c64xx.c index 88a456d..ad93d5d 100644 --- a/drivers/spi/spi_s3c64xx.c +++ b/drivers/spi/spi_s3c64xx.c @@ -160,7 +160,7 @@ struct s3c64xx_spi_driver_data { struct platform_device *pdev; struct spi_master *master; struct workqueue_struct *workqueue; - struct s3c64xx_spi_cntrlr_info *cntrlr_info; + struct s3c64xx_spi_info *cntrlr_info; struct spi_device *tgl_spi; struct work_struct work; struct list_head queue; @@ -180,7 +180,7 @@ static struct s3c2410_dma_client s3c64xx_spi_dma_client = { static void flush_fifo(struct s3c64xx_spi_driver_data *sdd) { - struct s3c64xx_spi_cntrlr_info *sci = sdd->cntrlr_info; + struct s3c64xx_spi_info *sci = sdd->cntrlr_info; void __iomem *regs = sdd->regs; unsigned long loops; u32 val; @@ -225,7 +225,7 @@ static void enable_datapath(struct s3c64xx_spi_driver_data *sdd, struct spi_device *spi, struct spi_transfer *xfer, int dma_mode) { - struct s3c64xx_spi_cntrlr_info *sci = sdd->cntrlr_info; + struct s3c64xx_spi_info *sci = sdd->cntrlr_info; void __iomem *regs = sdd->regs; u32 modecfg, chcfg; @@ -310,7 +310,7 @@ static inline void enable_cs(struct s3c64xx_spi_driver_data *sdd, static int wait_for_xfer(struct s3c64xx_spi_driver_data *sdd, struct spi_transfer *xfer, int dma_mode) { - struct s3c64xx_spi_cntrlr_info *sci = sdd->cntrlr_info; + struct s3c64xx_spi_info *sci = sdd->cntrlr_info; void __iomem *regs = sdd->regs; unsigned long val; int ms; @@ -389,7 +389,7 @@ static inline void disable_cs(struct s3c64xx_spi_driver_data *sdd, static void s3c64xx_spi_config(struct s3c64xx_spi_driver_data *sdd) { - struct s3c64xx_spi_cntrlr_info *sci = sdd->cntrlr_info; + struct s3c64xx_spi_info *sci = sdd->cntrlr_info; void __iomem *regs = sdd->regs; u32 val; @@ -558,7 +558,7 @@ static void s3c64xx_spi_unmap_mssg(struct s3c64xx_spi_driver_data *sdd, static void handle_msg(struct s3c64xx_spi_driver_data *sdd, struct spi_message *msg) { - struct s3c64xx_spi_cntrlr_info *sci = sdd->cntrlr_info; + struct s3c64xx_spi_info *sci = sdd->cntrlr_info; struct spi_device *spi = msg->spi; struct s3c64xx_spi_csinfo *cs = spi->controller_data; struct spi_transfer *xfer; @@ -786,7 +786,7 @@ static int s3c64xx_spi_setup(struct spi_device *spi) { struct s3c64xx_spi_csinfo *cs = spi->controller_data; struct s3c64xx_spi_driver_data *sdd; - struct s3c64xx_spi_cntrlr_info *sci; + struct s3c64xx_spi_info *sci; struct spi_message *msg; u32 psr, speed; unsigned long flags; @@ -867,7 +867,7 @@ setup_exit: static void s3c64xx_spi_hwinit(struct s3c64xx_spi_driver_data *sdd, int channel) { - struct s3c64xx_spi_cntrlr_info *sci = sdd->cntrlr_info; + struct s3c64xx_spi_info *sci = sdd->cntrlr_info; void __iomem *regs = sdd->regs; unsigned int val; @@ -902,7 +902,7 @@ static int __init s3c64xx_spi_probe(struct platform_device *pdev) { struct resource *mem_res, *dmatx_res, *dmarx_res; struct s3c64xx_spi_driver_data *sdd; - struct s3c64xx_spi_cntrlr_info *sci; + struct s3c64xx_spi_info *sci; struct spi_master *master; int ret; @@ -1078,7 +1078,7 @@ static int s3c64xx_spi_remove(struct platform_device *pdev) { struct spi_master *master = spi_master_get(platform_get_drvdata(pdev)); struct s3c64xx_spi_driver_data *sdd = spi_master_get_devdata(master); - struct s3c64xx_spi_cntrlr_info *sci = sdd->cntrlr_info; + struct s3c64xx_spi_info *sci = sdd->cntrlr_info; struct resource *mem_res; unsigned long flags; @@ -1118,7 +1118,7 @@ static int s3c64xx_spi_suspend(struct platform_device *pdev, pm_message_t state) { struct spi_master *master = spi_master_get(platform_get_drvdata(pdev)); struct s3c64xx_spi_driver_data *sdd = spi_master_get_devdata(master); - struct s3c64xx_spi_cntrlr_info *sci = sdd->cntrlr_info; + struct s3c64xx_spi_info *sci = sdd->cntrlr_info; struct s3c64xx_spi_csinfo *cs; unsigned long flags; @@ -1144,7 +1144,7 @@ static int s3c64xx_spi_resume(struct platform_device *pdev) { struct spi_master *master = spi_master_get(platform_get_drvdata(pdev)); struct s3c64xx_spi_driver_data *sdd = spi_master_get_devdata(master); - struct s3c64xx_spi_cntrlr_info *sci = sdd->cntrlr_info; + struct s3c64xx_spi_info *sci = sdd->cntrlr_info; unsigned long flags; sci->cfg_gpio(pdev); -- cgit v0.10.2 From ee64a37732c23ab2bcef5fe7785fd237a7e38951 Mon Sep 17 00:00:00 2001 From: Jassi Brar Date: Wed, 20 Jan 2010 13:49:44 -0700 Subject: spi/s3c64xx: Differentiate ip and rate clock The instance of SPI clock for controller and that used for generating signals ought to be independently handled. Signed-off-by: Jassi Brar Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_s3c64xx.c b/drivers/spi/spi_s3c64xx.c index ad93d5d..6d03d8f 100644 --- a/drivers/spi/spi_s3c64xx.c +++ b/drivers/spi/spi_s3c64xx.c @@ -1000,10 +1000,7 @@ static int __init s3c64xx_spi_probe(struct platform_device *pdev) goto err4; } - if (sci->src_clk_nr == S3C64XX_SPI_SRCCLK_PCLK) - sci->src_clk = sdd->clk; - else - sci->src_clk = clk_get(&pdev->dev, sci->src_clk_name); + sci->src_clk = clk_get(&pdev->dev, sci->src_clk_name); if (IS_ERR(sci->src_clk)) { dev_err(&pdev->dev, "Unable to acquire clock '%s'\n", sci->src_clk_name); @@ -1011,7 +1008,7 @@ static int __init s3c64xx_spi_probe(struct platform_device *pdev) goto err5; } - if (sci->src_clk != sdd->clk && clk_enable(sci->src_clk)) { + if (clk_enable(sci->src_clk)) { dev_err(&pdev->dev, "Couldn't enable clock '%s'\n", sci->src_clk_name); ret = -EBUSY; @@ -1053,11 +1050,9 @@ static int __init s3c64xx_spi_probe(struct platform_device *pdev) err8: destroy_workqueue(sdd->workqueue); err7: - if (sci->src_clk != sdd->clk) - clk_disable(sci->src_clk); + clk_disable(sci->src_clk); err6: - if (sci->src_clk != sdd->clk) - clk_put(sci->src_clk); + clk_put(sci->src_clk); err5: clk_disable(sdd->clk); err4: @@ -1093,11 +1088,8 @@ static int s3c64xx_spi_remove(struct platform_device *pdev) destroy_workqueue(sdd->workqueue); - if (sci->src_clk != sdd->clk) - clk_disable(sci->src_clk); - - if (sci->src_clk != sdd->clk) - clk_put(sci->src_clk); + clk_disable(sci->src_clk); + clk_put(sci->src_clk); clk_disable(sdd->clk); clk_put(sdd->clk); @@ -1130,9 +1122,7 @@ static int s3c64xx_spi_suspend(struct platform_device *pdev, pm_message_t state) msleep(10); /* Disable the clock */ - if (sci->src_clk != sdd->clk) - clk_disable(sci->src_clk); - + clk_disable(sci->src_clk); clk_disable(sdd->clk); sdd->cur_speed = 0; /* Output Clock is stopped */ @@ -1150,9 +1140,7 @@ static int s3c64xx_spi_resume(struct platform_device *pdev) sci->cfg_gpio(pdev); /* Enable the clock */ - if (sci->src_clk != sdd->clk) - clk_enable(sci->src_clk); - + clk_enable(sci->src_clk); clk_enable(sdd->clk); s3c64xx_spi_hwinit(sdd, pdev->id); -- cgit v0.10.2 From b0d5d6e55340348b0de75eb691b93d7e60dba879 Mon Sep 17 00:00:00 2001 From: Jassi Brar Date: Wed, 20 Jan 2010 13:49:44 -0700 Subject: spi/s3c64xx: Move src_clk to local driver data The pointer to SPI rate source clock had better be the member of driver local data structure rather than platform specific. Also, remove definitions of variable 'sci' that are rendered useless as a consequence. Signed-off-by: Jassi Brar Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_s3c64xx.c b/drivers/spi/spi_s3c64xx.c index 6d03d8f..3acf381 100644 --- a/drivers/spi/spi_s3c64xx.c +++ b/drivers/spi/spi_s3c64xx.c @@ -137,6 +137,7 @@ /** * struct s3c64xx_spi_driver_data - Runtime info holder for SPI driver. * @clk: Pointer to the spi clock. + * @src_clk: Pointer to the clock used to generate SPI signals. * @master: Pointer to the SPI Protocol master. * @workqueue: Work queue for the SPI xfer requests. * @cntrlr_info: Platform specific data for the controller this driver manages. @@ -157,6 +158,7 @@ struct s3c64xx_spi_driver_data { void __iomem *regs; struct clk *clk; + struct clk *src_clk; struct platform_device *pdev; struct spi_master *master; struct workqueue_struct *workqueue; @@ -389,7 +391,6 @@ static inline void disable_cs(struct s3c64xx_spi_driver_data *sdd, static void s3c64xx_spi_config(struct s3c64xx_spi_driver_data *sdd) { - struct s3c64xx_spi_info *sci = sdd->cntrlr_info; void __iomem *regs = sdd->regs; u32 val; @@ -435,7 +436,7 @@ static void s3c64xx_spi_config(struct s3c64xx_spi_driver_data *sdd) /* Configure Clock */ val = readl(regs + S3C64XX_SPI_CLK_CFG); val &= ~S3C64XX_SPI_PSR_MASK; - val |= ((clk_get_rate(sci->src_clk) / sdd->cur_speed / 2 - 1) + val |= ((clk_get_rate(sdd->src_clk) / sdd->cur_speed / 2 - 1) & S3C64XX_SPI_PSR_MASK); writel(val, regs + S3C64XX_SPI_CLK_CFG); @@ -831,17 +832,17 @@ static int s3c64xx_spi_setup(struct spi_device *spi) } /* Check if we can provide the requested rate */ - speed = clk_get_rate(sci->src_clk) / 2 / (0 + 1); /* Max possible */ + speed = clk_get_rate(sdd->src_clk) / 2 / (0 + 1); /* Max possible */ if (spi->max_speed_hz > speed) spi->max_speed_hz = speed; - psr = clk_get_rate(sci->src_clk) / 2 / spi->max_speed_hz - 1; + psr = clk_get_rate(sdd->src_clk) / 2 / spi->max_speed_hz - 1; psr &= S3C64XX_SPI_PSR_MASK; if (psr == S3C64XX_SPI_PSR_MASK) psr--; - speed = clk_get_rate(sci->src_clk) / 2 / (psr + 1); + speed = clk_get_rate(sdd->src_clk) / 2 / (psr + 1); if (spi->max_speed_hz < speed) { if (psr+1 < S3C64XX_SPI_PSR_MASK) { psr++; @@ -851,7 +852,7 @@ static int s3c64xx_spi_setup(struct spi_device *spi) } } - speed = clk_get_rate(sci->src_clk) / 2 / (psr + 1); + speed = clk_get_rate(sdd->src_clk) / 2 / (psr + 1); if (spi->max_speed_hz >= speed) spi->max_speed_hz = speed; else @@ -1000,15 +1001,15 @@ static int __init s3c64xx_spi_probe(struct platform_device *pdev) goto err4; } - sci->src_clk = clk_get(&pdev->dev, sci->src_clk_name); - if (IS_ERR(sci->src_clk)) { + sdd->src_clk = clk_get(&pdev->dev, sci->src_clk_name); + if (IS_ERR(sdd->src_clk)) { dev_err(&pdev->dev, "Unable to acquire clock '%s'\n", sci->src_clk_name); - ret = PTR_ERR(sci->src_clk); + ret = PTR_ERR(sdd->src_clk); goto err5; } - if (clk_enable(sci->src_clk)) { + if (clk_enable(sdd->src_clk)) { dev_err(&pdev->dev, "Couldn't enable clock '%s'\n", sci->src_clk_name); ret = -EBUSY; @@ -1050,9 +1051,9 @@ static int __init s3c64xx_spi_probe(struct platform_device *pdev) err8: destroy_workqueue(sdd->workqueue); err7: - clk_disable(sci->src_clk); + clk_disable(sdd->src_clk); err6: - clk_put(sci->src_clk); + clk_put(sdd->src_clk); err5: clk_disable(sdd->clk); err4: @@ -1073,7 +1074,6 @@ static int s3c64xx_spi_remove(struct platform_device *pdev) { struct spi_master *master = spi_master_get(platform_get_drvdata(pdev)); struct s3c64xx_spi_driver_data *sdd = spi_master_get_devdata(master); - struct s3c64xx_spi_info *sci = sdd->cntrlr_info; struct resource *mem_res; unsigned long flags; @@ -1088,8 +1088,8 @@ static int s3c64xx_spi_remove(struct platform_device *pdev) destroy_workqueue(sdd->workqueue); - clk_disable(sci->src_clk); - clk_put(sci->src_clk); + clk_disable(sdd->src_clk); + clk_put(sdd->src_clk); clk_disable(sdd->clk); clk_put(sdd->clk); @@ -1110,8 +1110,6 @@ static int s3c64xx_spi_suspend(struct platform_device *pdev, pm_message_t state) { struct spi_master *master = spi_master_get(platform_get_drvdata(pdev)); struct s3c64xx_spi_driver_data *sdd = spi_master_get_devdata(master); - struct s3c64xx_spi_info *sci = sdd->cntrlr_info; - struct s3c64xx_spi_csinfo *cs; unsigned long flags; spin_lock_irqsave(&sdd->lock, flags); @@ -1122,7 +1120,7 @@ static int s3c64xx_spi_suspend(struct platform_device *pdev, pm_message_t state) msleep(10); /* Disable the clock */ - clk_disable(sci->src_clk); + clk_disable(sdd->src_clk); clk_disable(sdd->clk); sdd->cur_speed = 0; /* Output Clock is stopped */ @@ -1140,7 +1138,7 @@ static int s3c64xx_spi_resume(struct platform_device *pdev) sci->cfg_gpio(pdev); /* Enable the clock */ - clk_enable(sci->src_clk); + clk_enable(sdd->src_clk); clk_enable(sdd->clk); s3c64xx_spi_hwinit(sdd, pdev->id); -- cgit v0.10.2 From ef6c680dc5a182a79b09567168d6713f46c85784 Mon Sep 17 00:00:00 2001 From: Jassi Brar Date: Wed, 20 Jan 2010 13:49:44 -0700 Subject: spi/s3c64xx: Check before mem-region release Add precautionary check before releasing memory region. Signed-off-by: Jassi Brar Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_s3c64xx.c b/drivers/spi/spi_s3c64xx.c index 3acf381..0e883f9 100644 --- a/drivers/spi/spi_s3c64xx.c +++ b/drivers/spi/spi_s3c64xx.c @@ -1097,7 +1097,8 @@ static int s3c64xx_spi_remove(struct platform_device *pdev) iounmap((void *) sdd->regs); mem_res = platform_get_resource(pdev, IORESOURCE_MEM, 0); - release_mem_region(mem_res->start, resource_size(mem_res)); + if (mem_res != NULL) + release_mem_region(mem_res->start, resource_size(mem_res)); platform_set_drvdata(pdev, NULL); spi_master_put(master); -- cgit v0.10.2 From e6b873c9666015484a01373a4eebc6cfa82e670d Mon Sep 17 00:00:00 2001 From: Jassi Brar Date: Wed, 20 Jan 2010 13:49:45 -0700 Subject: spi/s3c64xx: Include moved header Header for platform specific stuff has been rename to include the SoC type. Include the new header instead. Signed-off-by: Jassi Brar Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_s3c64xx.c b/drivers/spi/spi_s3c64xx.c index 0e883f9..9fa0b99 100644 --- a/drivers/spi/spi_s3c64xx.c +++ b/drivers/spi/spi_s3c64xx.c @@ -28,7 +28,7 @@ #include #include -#include +#include /* Registers and bit-fields */ -- cgit v0.10.2 From fa0fcde66ac3360678360104b24492015e7b852b Mon Sep 17 00:00:00 2001 From: Jassi Brar Date: Wed, 20 Jan 2010 13:49:45 -0700 Subject: spi/s3c64xx: Add new parameter to cs callback Since most of the chip-selects are simply going to be like gpio_set_value, it would do good to have the same callback type so that it could simply be made to point at gpio_set_value. Signed-off-by: Jassi Brar Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_s3c64xx.c b/drivers/spi/spi_s3c64xx.c index 9fa0b99..32db695 100644 --- a/drivers/spi/spi_s3c64xx.c +++ b/drivers/spi/spi_s3c64xx.c @@ -300,13 +300,14 @@ static inline void enable_cs(struct s3c64xx_spi_driver_data *sdd, if (sdd->tgl_spi != spi) { /* if last mssg on diff device */ /* Deselect the last toggled device */ cs = sdd->tgl_spi->controller_data; - cs->set_level(spi->mode & SPI_CS_HIGH ? 0 : 1); + cs->set_level(cs->line, + spi->mode & SPI_CS_HIGH ? 0 : 1); } sdd->tgl_spi = NULL; } cs = spi->controller_data; - cs->set_level(spi->mode & SPI_CS_HIGH ? 1 : 0); + cs->set_level(cs->line, spi->mode & SPI_CS_HIGH ? 1 : 0); } static int wait_for_xfer(struct s3c64xx_spi_driver_data *sdd, @@ -386,7 +387,7 @@ static inline void disable_cs(struct s3c64xx_spi_driver_data *sdd, if (sdd->tgl_spi == spi) sdd->tgl_spi = NULL; - cs->set_level(spi->mode & SPI_CS_HIGH ? 0 : 1); + cs->set_level(cs->line, spi->mode & SPI_CS_HIGH ? 0 : 1); } static void s3c64xx_spi_config(struct s3c64xx_spi_driver_data *sdd) -- cgit v0.10.2 From b490e3704ccee12deb295f96029d68e0daf02feb Mon Sep 17 00:00:00 2001 From: Feng Tang Date: Wed, 20 Jan 2010 13:49:45 -0700 Subject: spi/dw_spi: bug fix in wait_till_not_busy() Make the driver wait at least for 1 jiffie before issuing the warning, no matter what HZ is set to Signed-off-by: Feng Tang Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi.c b/drivers/spi/dw_spi.c index 31620fa..521d680 100644 --- a/drivers/spi/dw_spi.c +++ b/drivers/spi/dw_spi.c @@ -161,7 +161,7 @@ static inline void mrst_spi_debugfs_remove(struct dw_spi *dws) static void wait_till_not_busy(struct dw_spi *dws) { - unsigned long end = jiffies + usecs_to_jiffies(1000); + unsigned long end = jiffies + 1 + usecs_to_jiffies(1000); while (time_before(jiffies, end)) { if (!(dw_readw(dws, sr) & SR_BUSY)) -- cgit v0.10.2 From 51f921c1eb1124fb99ab0728c19e8e14c82c81be Mon Sep 17 00:00:00 2001 From: Feng Tang Date: Wed, 20 Jan 2010 13:49:45 -0700 Subject: spi/dw_spi: add a missed dw_spi_remove_host() in exit sequence Signed-off-by: Feng Tang Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi_pci.c b/drivers/spi/dw_spi_pci.c index 34ba691..7980f14 100644 --- a/drivers/spi/dw_spi_pci.c +++ b/drivers/spi/dw_spi_pci.c @@ -98,6 +98,7 @@ static void __devexit spi_pci_remove(struct pci_dev *pdev) struct dw_spi_pci *dwpci = pci_get_drvdata(pdev); pci_set_drvdata(pdev, NULL); + dw_spi_remove_host(&dwpci->dws); iounmap(dwpci->dws.regs); pci_release_region(pdev, 0); kfree(dwpci); -- cgit v0.10.2 From 552e450929a7298cc8834fd2824a60b2e914f70e Mon Sep 17 00:00:00 2001 From: Feng Tang Date: Wed, 20 Jan 2010 13:49:45 -0700 Subject: spi/dw_spi: refine the IRQ mode working flow Now dw_spi core fully supports 3 transfer modes: pure polling, DMA and IRQ mode. IRQ mode will use the FIFO half empty as the IRQ trigger, so each interface driver need set the fifo_len, so that core driver can handle it properly Signed-off-by: Feng Tang Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi.c b/drivers/spi/dw_spi.c index 521d680..1bb709b 100644 --- a/drivers/spi/dw_spi.c +++ b/drivers/spi/dw_spi.c @@ -358,6 +358,8 @@ static void transfer_complete(struct dw_spi *dws) static irqreturn_t interrupt_transfer(struct dw_spi *dws) { u16 irq_status, irq_mask = 0x3f; + u32 int_level = dws->fifo_len / 2; + u32 left; irq_status = dw_readw(dws, isr) & irq_mask; /* Error handling */ @@ -369,22 +371,23 @@ static irqreturn_t interrupt_transfer(struct dw_spi *dws) return IRQ_HANDLED; } - /* INT comes from tx */ - if (dws->tx && (irq_status & SPI_INT_TXEI)) { - while (dws->tx < dws->tx_end) + if (irq_status & SPI_INT_TXEI) { + spi_mask_intr(dws, SPI_INT_TXEI); + + left = (dws->tx_end - dws->tx) / dws->n_bytes; + left = (left > int_level) ? int_level : left; + + while (left--) dws->write(dws); + dws->read(dws); - if (dws->tx == dws->tx_end) { - spi_mask_intr(dws, SPI_INT_TXEI); + /* Re-enable the IRQ if there is still data left to tx */ + if (dws->tx_end > dws->tx) + spi_umask_intr(dws, SPI_INT_TXEI); + else transfer_complete(dws); - } } - /* INT comes from rx */ - if (dws->rx && (irq_status & SPI_INT_RXFI)) { - if (dws->read(dws)) - transfer_complete(dws); - } return IRQ_HANDLED; } @@ -428,6 +431,7 @@ static void pump_transfers(unsigned long data) u8 bits = 0; u8 imask = 0; u8 cs_change = 0; + u16 txint_level = 0; u16 clk_div = 0; u32 speed = 0; u32 cr0 = 0; @@ -438,6 +442,9 @@ static void pump_transfers(unsigned long data) chip = dws->cur_chip; spi = message->spi; + if (unlikely(!chip->clk_div)) + chip->clk_div = dws->max_freq / chip->speed_hz; + if (message->state == ERROR_STATE) { message->status = -EIO; goto early_exit; @@ -492,7 +499,7 @@ static void pump_transfers(unsigned long data) /* clk_div doesn't support odd number */ clk_div = dws->max_freq / speed; - clk_div = (clk_div >> 1) << 1; + clk_div = (clk_div + 1) & 0xfffe; chip->speed_hz = speed; chip->clk_div = clk_div; @@ -535,11 +542,16 @@ static void pump_transfers(unsigned long data) /* Check if current transfer is a DMA transaction */ dws->dma_mapped = map_dma_buffers(dws); + /* + * Interrupt mode + * we only need set the TXEI IRQ, as TX/RX always happen syncronizely + */ if (!dws->dma_mapped && !chip->poll_mode) { - if (dws->rx) - imask |= SPI_INT_RXFI; - if (dws->tx) - imask |= SPI_INT_TXEI; + int templen = dws->len / dws->n_bytes; + txint_level = dws->fifo_len / 2; + txint_level = (templen > txint_level) ? txint_level : templen; + + imask |= SPI_INT_TXEI; dws->transfer_handler = interrupt_transfer; } @@ -549,21 +561,23 @@ static void pump_transfers(unsigned long data) * 2. clk_div is changed * 3. control value changes */ - if (dw_readw(dws, ctrl0) != cr0 || cs_change || clk_div) { + if (dw_readw(dws, ctrl0) != cr0 || cs_change || clk_div || imask) { spi_enable_chip(dws, 0); if (dw_readw(dws, ctrl0) != cr0) dw_writew(dws, ctrl0, cr0); + spi_set_clk(dws, clk_div ? clk_div : chip->clk_div); + spi_chip_sel(dws, spi->chip_select); + /* Set the interrupt mask, for poll mode just diable all int */ spi_mask_intr(dws, 0xff); - if (!chip->poll_mode) + if (imask) spi_umask_intr(dws, imask); + if (txint_level) + dw_writew(dws, txfltr, txint_level); - spi_set_clk(dws, clk_div ? clk_div : chip->clk_div); - spi_chip_sel(dws, spi->chip_select); spi_enable_chip(dws, 1); - if (cs_change) dws->prev_chip = chip; } @@ -712,11 +726,11 @@ static int dw_spi_setup(struct spi_device *spi) } chip->bits_per_word = spi->bits_per_word; + if (!spi->max_speed_hz) { + dev_err(&spi->dev, "No max speed HZ parameter\n"); + return -EINVAL; + } chip->speed_hz = spi->max_speed_hz; - if (chip->speed_hz) - chip->clk_div = 25000000 / chip->speed_hz; - else - chip->clk_div = 8; /* default value */ chip->tmode = 0; /* Tx & Rx */ /* Default SPI mode is SCPOL = 0, SCPH = 0 */ diff --git a/drivers/spi/dw_spi_pci.c b/drivers/spi/dw_spi_pci.c index 7980f14..1f0735f 100644 --- a/drivers/spi/dw_spi_pci.c +++ b/drivers/spi/dw_spi_pci.c @@ -73,6 +73,7 @@ static int __devinit spi_pci_probe(struct pci_dev *pdev, dws->num_cs = 4; dws->max_freq = 25000000; /* for Moorestwon */ dws->irq = pdev->irq; + dws->fifo_len = 40; /* FIFO has 40 words buffer */ ret = dw_spi_add_host(dws); if (ret) diff --git a/include/linux/spi/dw_spi.h b/include/linux/spi/dw_spi.h index 51b3e77..1a127a3 100644 --- a/include/linux/spi/dw_spi.h +++ b/include/linux/spi/dw_spi.h @@ -90,6 +90,7 @@ struct dw_spi { unsigned long paddr; u32 iolen; int irq; + u32 fifo_len; /* depth of the FIFO buffer */ u32 max_freq; /* max bus freq supported */ u16 bus_num; -- cgit v0.10.2 From ac48eee064f743a198cb38e5a6a62c3f1082ebc1 Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Wed, 20 Jan 2010 13:49:45 -0700 Subject: spi: update MSIOF includes Update the MSIOF driver to remove the architecture speficic spi header file and add err.h. This makes the driver compile on non-SH architectures. Signed-off-by: Magnus Damm Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_sh_msiof.c b/drivers/spi/spi_sh_msiof.c index 51e5e1d..3b5b1be 100644 --- a/drivers/spi/spi_sh_msiof.c +++ b/drivers/spi/spi_sh_msiof.c @@ -20,12 +20,12 @@ #include #include #include +#include #include #include #include -#include #include struct sh_msiof_spi_priv { -- cgit v0.10.2 From f4d4ecfe788b4141d8c90cfc3ac2831f620f5c1b Mon Sep 17 00:00:00 2001 From: Alberto Panizzo Date: Wed, 20 Jan 2010 13:49:45 -0700 Subject: spi/spi_imx: add device information by switching pr_debug() to dev_dbg() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Useful when debugging multiple spi channels. Signed-off-by: Alberto Panizzo Acked-by: Uwe Kleine-König Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_imx.c b/drivers/spi/spi_imx.c index 1893f1e..0ddbbe4 100644 --- a/drivers/spi/spi_imx.c +++ b/drivers/spi/spi_imx.c @@ -469,7 +469,7 @@ static int spi_imx_setup(struct spi_device *spi) struct spi_imx_data *spi_imx = spi_master_get_devdata(spi->master); int gpio = spi_imx->chipselect[spi->chip_select]; - pr_debug("%s: mode %d, %u bpw, %d hz\n", __func__, + dev_dbg(&spi->dev, "%s: mode %d, %u bpw, %d hz\n", __func__, spi->mode, spi->bits_per_word, spi->max_speed_hz); if (gpio >= 0) -- cgit v0.10.2 From 9778214990af88ec6720bd771d7fc0fa1b140b02 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 20 Jan 2010 13:49:45 -0700 Subject: spi: xilinx_spi: Fix up I/O routine wrapping bogosity. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit xilinx_spi presently makes some fairly questionable assumptions about I/O routines, and attempts to assign ioread32/iowrite32 and friends directly to its own internal function pointers. On many platforms these I/O routines are macros or wrappers and not actual functions on their own, resulting in things like: ERROR: "ioread32be" [drivers/spi/xilinx_spi.ko] undefined! ERROR: "iowrite32be" [drivers/spi/xilinx_spi.ko] undefined! ERROR: "iowrite32" [drivers/spi/xilinx_spi.ko] undefined! ERROR: "ioread32" [drivers/spi/xilinx_spi.ko] undefined! If xilinx_spi wants to do this sort of casting, it needs to provide its own wrappers for these, or change how it does accesses completely. I've opted for the first approach, and the attached silly patch does that. If someone with the hardware available wants to give the second option a try that's ok too. In any event, the current code is broken for at least: arm, avr32, blackfin, microblaze, mn10300, and sh. Signed-off-by: Paul Mundt Acked-by: Richard Röjfors Signed-off-by: Grant Likely diff --git a/drivers/spi/xilinx_spi.c b/drivers/spi/xilinx_spi.c index 9f38637..1b47363 100644 --- a/drivers/spi/xilinx_spi.c +++ b/drivers/spi/xilinx_spi.c @@ -93,6 +93,26 @@ struct xilinx_spi { void (*rx_fn) (struct xilinx_spi *); }; +static void xspi_write32(u32 val, void __iomem *addr) +{ + iowrite32(val, addr); +} + +static unsigned int xspi_read32(void __iomem *addr) +{ + return ioread32(addr); +} + +static void xspi_write32_be(u32 val, void __iomem *addr) +{ + iowrite32be(val, addr); +} + +static unsigned int xspi_read32_be(void __iomem *addr) +{ + return ioread32be(addr); +} + static void xspi_tx8(struct xilinx_spi *xspi) { xspi->write_fn(*xspi->tx_ptr, xspi->regs + XSPI_TXD_OFFSET); @@ -374,11 +394,11 @@ struct spi_master *xilinx_spi_init(struct device *dev, struct resource *mem, xspi->mem = *mem; xspi->irq = irq; if (pdata->little_endian) { - xspi->read_fn = ioread32; - xspi->write_fn = iowrite32; + xspi->read_fn = xspi_read32; + xspi->write_fn = xspi_write32; } else { - xspi->read_fn = ioread32be; - xspi->write_fn = iowrite32be; + xspi->read_fn = xspi_read32_be; + xspi->write_fn = xspi_write32_be; } xspi->bits_per_word = pdata->bits_per_word; if (xspi->bits_per_word == 8) { -- cgit v0.10.2 From 99147b5c4167612a987860b661b9f8b79e66b81f Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Wed, 20 Jan 2010 14:03:39 -0700 Subject: spi/dw_spi: fix __init/__devinit section mismatch Section mismatch in reference from the function dw_spi_add_host() to the function init_queue() Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi.c b/drivers/spi/dw_spi.c index 1bb709b..f713af8 100644 --- a/drivers/spi/dw_spi.c +++ b/drivers/spi/dw_spi.c @@ -749,7 +749,7 @@ static void dw_spi_cleanup(struct spi_device *spi) kfree(chip); } -static int __init init_queue(struct dw_spi *dws) +static int __devinit init_queue(struct dw_spi *dws) { INIT_LIST_HEAD(&dws->queue); spin_lock_init(&dws->lock); -- cgit v0.10.2 From 77c2019fc1b4495ce483ef2af9cb12bae23e353b Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 21 Jan 2010 14:19:41 +0900 Subject: sh: initial PMB mapping iteration by helper macro. All of the cached/uncached mapping setup is duplicated for each size, and also misses out on the 16MB case. Rather than duplicating the same iter code for that we just consolidate it in to a helper macro that builds an iter for each size. The 16MB case is then trivially bolted on at the end. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/head_32.S b/arch/sh/kernel/head_32.S index 8ee31a0..ee954d1 100644 --- a/arch/sh/kernel/head_32.S +++ b/arch/sh/kernel/head_32.S @@ -86,6 +86,62 @@ ENTRY(_stext) #endif #if defined(CONFIG_PMB) && !defined(CONFIG_PMB_LEGACY) + +#define __PMB_ITER_BY_SIZE(size) \ +.L##size: \ + mov #(size >> 4), r6; \ + shll16 r6; \ + shll8 r6; \ + \ + cmp/hi r5, r6; \ + bt 9999f; \ + \ + mov #(PMB_SZ_##size##M >> 2), r9; \ + shll2 r9; \ + \ + /* \ + * Cached mapping \ + */ \ + mov #PMB_C, r8; \ + or r0, r8; \ + or r9, r8; \ + mov.l r8, @r1; \ + mov.l r2, @r3; \ + \ + /* Increment to the next PMB_DATA entry */ \ + add r4, r1; \ + /* Increment to the next PMB_ADDR entry */ \ + add r4, r3; \ + /* Increment number of PMB entries */ \ + add #1, r10; \ + \ + /* \ + * Uncached mapping \ + */ \ + mov #(PMB_UB >> 8), r8; \ + shll8 r8; \ + \ + or r0, r8; \ + or r9, r8; \ + mov.l r8, @r1; \ + mov r2, r8; \ + add r7, r8; \ + mov.l r8, @r3; \ + \ + /* Increment to the next PMB_DATA entry */ \ + add r4, r1; \ + /* Increment to the next PMB_ADDR entry */ \ + add r4, r3; \ + /* Increment number of PMB entries */ \ + add #1, r10; \ + \ + sub r6, r5; \ + add r6, r0; \ + add r6, r2; \ + \ + bra .L##size; \ +9999: + /* * Reconfigure the initial PMB mappings setup by the hardware. * @@ -142,154 +198,11 @@ ENTRY(_stext) * r9 = scratch register * r10 = number of PMB entries we've setup */ -.L512: - mov #(512 >> 4), r6 - shll16 r6 - shll8 r6 - - cmp/hi r5, r6 - bt .L128 - - mov #(PMB_SZ_512M >> 2), r9 - shll2 r9 - - /* - * Cached mapping - */ - mov #PMB_C, r8 - or r0, r8 - or r9, r8 - mov.l r8, @r1 - mov.l r2, @r3 - - add r4, r1 /* Increment to the next PMB_DATA entry */ - add r4, r3 /* Increment to the next PMB_ADDR entry */ - - add #1, r10 /* Increment number of PMB entries */ - - /* - * Uncached mapping - */ - mov #(PMB_UB >> 8), r8 - shll8 r8 - - or r0, r8 - or r9, r8 - mov.l r8, @r1 - mov r2, r8 - add r7, r8 - mov.l r8, @r3 - - add r4, r1 /* Increment to the next PMB_DATA entry */ - add r4, r3 /* Increment to the next PMB_ADDR entry */ - - add #1, r10 /* Increment number of PMB entries */ - - sub r6, r5 - add r6, r0 - add r6, r2 - - bra .L512 - -.L128: - mov #(128 >> 4), r6 - shll16 r6 - shll8 r6 - - cmp/hi r5, r6 - bt .L64 - - mov #(PMB_SZ_128M >> 2), r9 - shll2 r9 - - /* - * Cached mapping - */ - mov #PMB_C, r8 - or r0, r8 - or r9, r8 - mov.l r8, @r1 - mov.l r2, @r3 - - add r4, r1 /* Increment to the next PMB_DATA entry */ - add r4, r3 /* Increment to the next PMB_ADDR entry */ - - add #1, r10 /* Increment number of PMB entries */ - - /* - * Uncached mapping - */ - mov #(PMB_UB >> 8), r8 - shll8 r8 - - or r0, r8 - or r9, r8 - mov.l r8, @r1 - mov r2, r8 - add r7, r8 - mov.l r8, @r3 - - add r4, r1 /* Increment to the next PMB_DATA entry */ - add r4, r3 /* Increment to the next PMB_ADDR entry */ - - add #1, r10 /* Increment number of PMB entries */ - - sub r6, r5 - add r6, r0 - add r6, r2 - - bra .L128 - -.L64: - mov #(64 >> 4), r6 - shll16 r6 - shll8 r6 - - cmp/hi r5, r6 - bt .Ldone - - mov #(PMB_SZ_64M >> 2), r9 - shll2 r9 - - /* - * Cached mapping - */ - mov #PMB_C, r8 - or r0, r8 - or r9, r8 - mov.l r8, @r1 - mov.l r2, @r3 - - add r4, r1 /* Increment to the next PMB_DATA entry */ - add r4, r3 /* Increment to the next PMB_ADDR entry */ - - add #1, r10 /* Increment number of PMB entries */ - - /* - * Uncached mapping - */ - mov #(PMB_UB >> 8), r8 - shll8 r8 - - or r0, r8 - or r9, r8 - mov.l r8, @r1 - mov r2, r8 - add r7, r8 - mov.l r8, @r3 - - add r4, r1 /* Increment to the next PMB_DATA entry */ - add r4, r3 /* Increment to the next PMB_ADDR entry */ - - add #1, r10 /* Increment number of PMB entries */ - - sub r6, r5 - add r6, r0 - add r6, r2 - - bra .L64 + __PMB_ITER_BY_SIZE(512) + __PMB_ITER_BY_SIZE(128) + __PMB_ITER_BY_SIZE(64) + __PMB_ITER_BY_SIZE(16) -.Ldone: /* Update cached_to_uncached */ mov.l .Lcached_to_uncached, r0 mov.l r7, @r0 -- cgit v0.10.2 From 2023b843d7b62d99e7356b872e99abc7d6667e49 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 21 Jan 2010 15:42:58 +0900 Subject: sh: Rework P2 to only include kernel text. This effectively neutralizes P2 by getting rid of P1 identity mapping for all available memory and instead only establishes a single unbuffered PMB entry (16MB -- the smallest available) that covers the kernel. As using segmentation for abusing caching attributes in drivers is no longer supported (and there are no drivers that can be enabled in 32-bit mode that do this), this provides us with all of the uncached access needs by the kernel itself. Drivers and their ilk need to specify their caching attributes when remapping through page tables, as usual. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/head_32.S b/arch/sh/kernel/head_32.S index ee954d1..e933201 100644 --- a/arch/sh/kernel/head_32.S +++ b/arch/sh/kernel/head_32.S @@ -86,7 +86,96 @@ ENTRY(_stext) #endif #if defined(CONFIG_PMB) && !defined(CONFIG_PMB_LEGACY) +/* + * Reconfigure the initial PMB mappings setup by the hardware. + * + * When we boot in 32-bit MMU mode there are 2 PMB entries already + * setup for us. + * + * Entry VPN PPN V SZ C UB WT + * --------------------------------------------------------------- + * 0 0x80000000 0x00000000 1 512MB 1 0 1 + * 1 0xA0000000 0x00000000 1 512MB 0 0 0 + * + * But we reprogram them here because we want complete control over + * our address space and the initial mappings may not map PAGE_OFFSET + * to __MEMORY_START (or even map all of our RAM). + * + * Once we've setup cached and uncached mappings we clear the rest of the + * PMB entries. This clearing also deals with the fact that PMB entries + * can persist across reboots. The PMB could have been left in any state + * when the reboot occurred, so to be safe we clear all entries and start + * with with a clean slate. + * + * The uncached mapping is constructed using the smallest possible + * mapping with a single unbufferable page. Only the kernel text needs to + * be covered via the uncached mapping so that certain functions can be + * run uncached. + * + * Drivers and the like that have previously abused the 1:1 identity + * mapping are unsupported in 32-bit mode and must specify their caching + * preference when page tables are constructed. + * + * This frees up the P2 space for more nefarious purposes. + * + * Register utilization is as follows: + * + * r0 = PMB_DATA data field + * r1 = PMB_DATA address field + * r2 = PMB_ADDR data field + * r3 = PMB_ADDR address field + * r4 = PMB_E_SHIFT + * r5 = remaining amount of RAM to map + * r6 = PMB mapping size we're trying to use + * r7 = cached_to_uncached + * r8 = scratch register + * r9 = scratch register + * r10 = number of PMB entries we've setup + */ + + mov.l .LMMUCR, r1 /* Flush the TLB */ + mov.l @r1, r0 + or #MMUCR_TI, r0 + mov.l r0, @r1 + + mov.l .LMEMORY_SIZE, r5 + mov r5, r7 + + mov #PMB_E_SHIFT, r0 + mov #0x1, r4 + shld r0, r4 + mov.l .LFIRST_DATA_ENTRY, r0 + mov.l .LPMB_DATA, r1 + mov.l .LFIRST_ADDR_ENTRY, r2 + mov.l .LPMB_ADDR, r3 + + mov #0, r10 + + /* + * Uncached mapping + */ + mov #(PMB_SZ_16M >> 2), r9 + shll2 r9 + + mov #(PMB_UB >> 8), r8 + shll8 r8 + + or r0, r8 + or r9, r8 + mov.l r8, @r1 + mov r2, r8 + add r7, r8 + mov.l r8, @r3 + + add r4, r1 + add r4, r3 + add #1, r10 + +/* + * Iterate over all of the available sizes from largest to + * smallest for constructing the cached mapping. + */ #define __PMB_ITER_BY_SIZE(size) \ .L##size: \ mov #(size >> 4), r6; \ @@ -115,26 +204,6 @@ ENTRY(_stext) /* Increment number of PMB entries */ \ add #1, r10; \ \ - /* \ - * Uncached mapping \ - */ \ - mov #(PMB_UB >> 8), r8; \ - shll8 r8; \ - \ - or r0, r8; \ - or r9, r8; \ - mov.l r8, @r1; \ - mov r2, r8; \ - add r7, r8; \ - mov.l r8, @r3; \ - \ - /* Increment to the next PMB_DATA entry */ \ - add r4, r1; \ - /* Increment to the next PMB_ADDR entry */ \ - add r4, r3; \ - /* Increment number of PMB entries */ \ - add #1, r10; \ - \ sub r6, r5; \ add r6, r0; \ add r6, r2; \ @@ -142,68 +211,14 @@ ENTRY(_stext) bra .L##size; \ 9999: - /* - * Reconfigure the initial PMB mappings setup by the hardware. - * - * When we boot in 32-bit MMU mode there are 2 PMB entries already - * setup for us. - * - * Entry VPN PPN V SZ C UB WT - * --------------------------------------------------------------- - * 0 0x80000000 0x00000000 1 512MB 1 0 1 - * 1 0xA0000000 0x00000000 1 512MB 0 0 0 - * - * But we reprogram them here because we want complete control over - * our address space and the initial mappings may not map PAGE_OFFSET - * to __MEMORY_START (or even map all of our RAM). - * - * Once we've setup cached and uncached mappings for all of RAM we - * clear the rest of the PMB entries. - * - * This clearing also deals with the fact that PMB entries can persist - * across reboots. The PMB could have been left in any state when the - * reboot occurred, so to be safe we clear all entries and start with - * with a clean slate. - */ - - mov.l .LMMUCR, r1 /* Flush the TLB */ - mov.l @r1, r0 - or #MMUCR_TI, r0 - mov.l r0, @r1 - - mov.l .LMEMORY_SIZE, r5 - mov r5, r7 - - mov #PMB_E_SHIFT, r0 - mov #0x1, r4 - shld r0, r4 - - mov.l .LFIRST_DATA_ENTRY, r0 - mov.l .LPMB_DATA, r1 - mov.l .LFIRST_ADDR_ENTRY, r2 - mov.l .LPMB_ADDR, r3 - - mov #0, r10 - - /* - * r0 = PMB_DATA data field - * r1 = PMB_DATA address field - * r2 = PMB_ADDR data field - * r3 = PMB_ADDR address field - * r4 = PMB_E_SHIFT - * r5 = remaining amount of RAM to map - * r6 = PMB mapping size we're trying to use - * r7 = cached_to_uncached - * r8 = scratch register - * r9 = scratch register - * r10 = number of PMB entries we've setup - */ __PMB_ITER_BY_SIZE(512) __PMB_ITER_BY_SIZE(128) __PMB_ITER_BY_SIZE(64) __PMB_ITER_BY_SIZE(16) - /* Update cached_to_uncached */ + /* + * Now that we can access it, update cached_to_uncached. + */ mov.l .Lcached_to_uncached, r0 mov.l r7, @r0 -- cgit v0.10.2 From 3125ee72dca25fc2157dcddd07e2d740db921fc4 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 21 Jan 2010 15:54:31 +0900 Subject: sh: Track the uncached mapping size. This provides a variable for tracking the uncached mapping size, and uses it for pretty printing the uncached lowmem range. Beyond this, we'll also be building on top of this for figuring out from where the remainder of P2 becomes usable when constructing unrelated mappings. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/system.h b/arch/sh/include/asm/system.h index de2fc39..6442f17 100644 --- a/arch/sh/include/asm/system.h +++ b/arch/sh/include/asm/system.h @@ -137,6 +137,7 @@ extern unsigned int instruction_size(unsigned int insn); #endif extern unsigned long cached_to_uncached; +extern unsigned long uncached_size; extern struct dentry *sh_debugfs_root; diff --git a/arch/sh/kernel/head_32.S b/arch/sh/kernel/head_32.S index e933201..83f2b84 100644 --- a/arch/sh/kernel/head_32.S +++ b/arch/sh/kernel/head_32.S @@ -217,11 +217,18 @@ ENTRY(_stext) __PMB_ITER_BY_SIZE(16) /* - * Now that we can access it, update cached_to_uncached. + * Now that we can access it, update cached_to_uncached and + * uncached_size. */ mov.l .Lcached_to_uncached, r0 mov.l r7, @r0 + mov.l .Luncached_size, r0 + mov #1, r7 + shll16 r7 + shll8 r7 + mov.l r7, @r0 + /* * Clear the remaining PMB entries. * @@ -300,5 +307,6 @@ ENTRY(stack_start) .LFIRST_DATA_ENTRY: .long __MEMORY_START | PMB_V .LMMUCR: .long MMUCR .Lcached_to_uncached: .long cached_to_uncached +.Luncached_size: .long uncached_size .LMEMORY_SIZE: .long __MEMORY_SIZE #endif diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c index f3e23ad..a28ff63 100644 --- a/arch/sh/mm/init.c +++ b/arch/sh/mm/init.c @@ -27,15 +27,17 @@ pgd_t swapper_pg_dir[PTRS_PER_PGD]; #ifdef CONFIG_SUPERH32 /* - * Handle trivial transitions between cached and uncached - * segments, making use of the 1:1 mapping relationship in - * 512MB lowmem. - * * This is the offset of the uncached section from its cached alias. - * Default value only valid in 29 bit mode, in 32bit mode will be - * overridden in pmb_init. + * + * Legacy platforms handle trivial transitions between cached and + * uncached segments by making use of the 1:1 mapping relationship in + * 512MB lowmem, others via a special uncached mapping. + * + * Default value only valid in 29 bit mode, in 32bit mode this will be + * updated by the early PMB initialization code. */ unsigned long cached_to_uncached = P2SEG - P1SEG; +unsigned long uncached_size = 0x20000000; #endif #ifdef CONFIG_MMU @@ -281,7 +283,8 @@ void __init mem_init(void) " pkmap : 0x%08lx - 0x%08lx (%4ld kB)\n" #endif " vmalloc : 0x%08lx - 0x%08lx (%4ld MB)\n" - " lowmem : 0x%08lx - 0x%08lx (%4ld MB)\n" + " lowmem : 0x%08lx - 0x%08lx (%4ld MB) (cached)\n" + " : 0x%08lx - 0x%08lx (%4ld MB) (uncached)\n" " .init : 0x%08lx - 0x%08lx (%4ld kB)\n" " .data : 0x%08lx - 0x%08lx (%4ld kB)\n" " .text : 0x%08lx - 0x%08lx (%4ld kB)\n", @@ -299,6 +302,10 @@ void __init mem_init(void) (unsigned long)memory_start, (unsigned long)high_memory, ((unsigned long)high_memory - (unsigned long)memory_start) >> 20, + (unsigned long)memory_start + cached_to_uncached, + (unsigned long)memory_start + cached_to_uncached + uncached_size, + uncached_size >> 20, + (unsigned long)&__init_begin, (unsigned long)&__init_end, ((unsigned long)&__init_end - (unsigned long)&__init_begin) >> 10, -- cgit v0.10.2 From 2dc2f8e0c46864e2a3722c84eaa96513d4cf8b2f Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 21 Jan 2010 16:05:25 +0900 Subject: sh: Kill off the special uncached section and fixmap. Now that cached_to_uncached works as advertized in 32-bit mode and we're never going to be able to map < 16MB anyways, there's no need for the special uncached section. Kill it off. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/fixmap.h b/arch/sh/include/asm/fixmap.h index 38a1de8..17b22b6c 100644 --- a/arch/sh/include/asm/fixmap.h +++ b/arch/sh/include/asm/fixmap.h @@ -55,11 +55,12 @@ enum fixed_addresses { #define FIX_N_COLOURS 8 FIX_CMAP_BEGIN, FIX_CMAP_END = FIX_CMAP_BEGIN + (FIX_N_COLOURS * NR_CPUS) - 1, - FIX_UNCACHED, + #ifdef CONFIG_HIGHMEM FIX_KMAP_BEGIN, /* reserved pte's for temporary kernel mappings */ FIX_KMAP_END = FIX_KMAP_BEGIN+(KM_TYPE_NR*NR_CPUS)-1, #endif + /* * FIX_IOREMAP entries are useful for mapping physical address * space before ioremap() is useable, e.g. really early in boot @@ -68,6 +69,7 @@ enum fixed_addresses { #define FIX_N_IOREMAPS 32 FIX_IOREMAP_BEGIN, FIX_IOREMAP_END = FIX_IOREMAP_BEGIN + FIX_N_IOREMAPS, + __end_of_fixed_addresses }; diff --git a/arch/sh/include/asm/system_32.h b/arch/sh/include/asm/system_32.h index 34bd2ba..51296b3 100644 --- a/arch/sh/include/asm/system_32.h +++ b/arch/sh/include/asm/system_32.h @@ -145,9 +145,6 @@ do { \ __restore_dsp(prev); \ } while (0) -#define __uses_jump_to_uncached \ - noinline __attribute__ ((__section__ (".uncached.text"))) - /* * Jump to uncached area. * When handling TLB or caches, we need to do it from an uncached area. diff --git a/arch/sh/include/asm/system_64.h b/arch/sh/include/asm/system_64.h index 604ba78..3391bb6 100644 --- a/arch/sh/include/asm/system_64.h +++ b/arch/sh/include/asm/system_64.h @@ -33,8 +33,6 @@ do { \ &next->thread); \ } while (0) -#define __uses_jump_to_uncached - #define jump_to_uncached() do { } while (0) #define back_to_cached() do { } while (0) diff --git a/arch/sh/kernel/cpu/init.c b/arch/sh/kernel/cpu/init.c index a5bb055..05a7d2a 100644 --- a/arch/sh/kernel/cpu/init.c +++ b/arch/sh/kernel/cpu/init.c @@ -98,7 +98,7 @@ static void __init expmask_init(void) #endif /* 2nd-level cache init */ -void __uses_jump_to_uncached __attribute__ ((weak)) l2_cache_init(void) +void __attribute__ ((weak)) l2_cache_init(void) { } @@ -106,7 +106,7 @@ void __uses_jump_to_uncached __attribute__ ((weak)) l2_cache_init(void) * Generic first-level cache init */ #ifdef CONFIG_SUPERH32 -static void __uses_jump_to_uncached cache_init(void) +static void cache_init(void) { unsigned long ccr, flags; diff --git a/arch/sh/kernel/cpu/sh3/probe.c b/arch/sh/kernel/cpu/sh3/probe.c index f9c7df6..c2db657 100644 --- a/arch/sh/kernel/cpu/sh3/probe.c +++ b/arch/sh/kernel/cpu/sh3/probe.c @@ -16,7 +16,7 @@ #include #include -int __uses_jump_to_uncached detect_cpu_and_cache_system(void) +int detect_cpu_and_cache_system(void) { unsigned long addr0, addr1, data0, data1, data2, data3; diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7723.c b/arch/sh/kernel/cpu/sh4a/setup-sh7723.c index 772b926..9d42625 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7723.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7723.c @@ -592,7 +592,8 @@ void __init plat_early_device_setup(void) #define RAMCR_CACHE_L2FC 0x0002 #define RAMCR_CACHE_L2E 0x0001 #define L2_CACHE_ENABLE (RAMCR_CACHE_L2E|RAMCR_CACHE_L2FC) -void __uses_jump_to_uncached l2_cache_init(void) + +void l2_cache_init(void) { /* Enable L2 cache */ ctrl_outl(L2_CACHE_ENABLE, RAMCR); diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7724.c b/arch/sh/kernel/cpu/sh4a/setup-sh7724.c index d32f96c..578e9f8 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7724.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7724.c @@ -714,7 +714,8 @@ void __init plat_early_device_setup(void) #define RAMCR_CACHE_L2FC 0x0002 #define RAMCR_CACHE_L2E 0x0001 #define L2_CACHE_ENABLE (RAMCR_CACHE_L2E|RAMCR_CACHE_L2FC) -void __uses_jump_to_uncached l2_cache_init(void) + +void l2_cache_init(void) { /* Enable L2 cache */ ctrl_outl(L2_CACHE_ENABLE, RAMCR); diff --git a/arch/sh/kernel/vmlinux.lds.S b/arch/sh/kernel/vmlinux.lds.S index 9e5a587..93e9b25 100644 --- a/arch/sh/kernel/vmlinux.lds.S +++ b/arch/sh/kernel/vmlinux.lds.S @@ -53,18 +53,6 @@ SECTIONS NOTES RO_DATA(PAGE_SIZE) - - /* - * Code which must be executed uncached and the associated data - */ - . = ALIGN(PAGE_SIZE); - .uncached : AT(ADDR(.uncached) - LOAD_OFFSET) { - __uncached_start = .; - *(.uncached.text) - *(.uncached.data) - __uncached_end = .; - } - RW_DATA_SECTION(L1_CACHE_BYTES, PAGE_SIZE, THREAD_SIZE) _edata = .; /* End of data section */ diff --git a/arch/sh/mm/cache-debugfs.c b/arch/sh/mm/cache-debugfs.c index 5ba067b..48ce82e 100644 --- a/arch/sh/mm/cache-debugfs.c +++ b/arch/sh/mm/cache-debugfs.c @@ -22,8 +22,7 @@ enum cache_type { CACHE_TYPE_UNIFIED, }; -static int __uses_jump_to_uncached cache_seq_show(struct seq_file *file, - void *iter) +static int cache_seq_show(struct seq_file *file, void *iter) { unsigned int cache_type = (unsigned int)file->private; struct cache_info *cache; diff --git a/arch/sh/mm/cache-sh4.c b/arch/sh/mm/cache-sh4.c index a2301da..87115b3 100644 --- a/arch/sh/mm/cache-sh4.c +++ b/arch/sh/mm/cache-sh4.c @@ -36,7 +36,7 @@ static void __flush_cache_one(unsigned long addr, unsigned long phys, * Called from kernel/module.c:sys_init_module and routine for a.out format, * signal handler code and kprobes code */ -static void __uses_jump_to_uncached sh4_flush_icache_range(void *args) +static void sh4_flush_icache_range(void *args) { struct flusher_data *data = args; unsigned long start, end; @@ -124,7 +124,7 @@ static void sh4_flush_dcache_page(void *arg) } /* TODO: Selective icache invalidation through IC address array.. */ -static void __uses_jump_to_uncached flush_icache_all(void) +static void flush_icache_all(void) { unsigned long flags, ccr; diff --git a/arch/sh/mm/cache-sh7705.c b/arch/sh/mm/cache-sh7705.c index f527fb7..870293e 100644 --- a/arch/sh/mm/cache-sh7705.c +++ b/arch/sh/mm/cache-sh7705.c @@ -78,7 +78,7 @@ static void sh7705_flush_icache_range(void *args) /* * Writeback&Invalidate the D-cache of the page */ -static void __uses_jump_to_uncached __flush_dcache_page(unsigned long phys) +static void __flush_dcache_page(unsigned long phys) { unsigned long ways, waysize, addrstart; unsigned long flags; @@ -144,7 +144,7 @@ static void sh7705_flush_dcache_page(void *arg) __flush_dcache_page(__pa(page_address(page))); } -static void __uses_jump_to_uncached sh7705_flush_cache_all(void *args) +static void sh7705_flush_cache_all(void *args) { unsigned long flags; diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c index a28ff63..dffa6c7 100644 --- a/arch/sh/mm/init.c +++ b/arch/sh/mm/init.c @@ -211,9 +211,6 @@ void __init paging_init(void) } free_area_init_nodes(max_zone_pfns); - - /* Set up the uncached fixmap */ - set_fixmap_nocache(FIX_UNCACHED, __pa(&__uncached_start)); } /* diff --git a/arch/sh/mm/pmb.c b/arch/sh/mm/pmb.c index d318fa6..3d5eece 100644 --- a/arch/sh/mm/pmb.c +++ b/arch/sh/mm/pmb.c @@ -127,14 +127,14 @@ static void __set_pmb_entry(unsigned long vpn, unsigned long ppn, ctrl_outl(ppn | flags | PMB_V, mk_pmb_data(pos)); } -static void __uses_jump_to_uncached set_pmb_entry(struct pmb_entry *pmbe) +static void set_pmb_entry(struct pmb_entry *pmbe) { jump_to_uncached(); __set_pmb_entry(pmbe->vpn, pmbe->ppn, pmbe->flags, pmbe->entry); back_to_cached(); } -static void __uses_jump_to_uncached clear_pmb_entry(struct pmb_entry *pmbe) +static void clear_pmb_entry(struct pmb_entry *pmbe) { unsigned int entry = pmbe->entry; unsigned long addr; @@ -364,7 +364,7 @@ static inline int pmb_apply_legacy_mappings(void) } #endif -int __uses_jump_to_uncached pmb_init(void) +int pmb_init(void) { int i; unsigned long addr, data; diff --git a/arch/sh/mm/tlb-pteaex.c b/arch/sh/mm/tlb-pteaex.c index 409b7c2..32dc674 100644 --- a/arch/sh/mm/tlb-pteaex.c +++ b/arch/sh/mm/tlb-pteaex.c @@ -68,8 +68,7 @@ void __update_tlb(struct vm_area_struct *vma, unsigned long address, pte_t pte) * in extended mode, the legacy 8-bit ASID field in address array 1 has * undefined behaviour. */ -void __uses_jump_to_uncached local_flush_tlb_one(unsigned long asid, - unsigned long page) +void local_flush_tlb_one(unsigned long asid, unsigned long page) { jump_to_uncached(); __raw_writel(page, MMU_UTLB_ADDRESS_ARRAY | MMU_PAGE_ASSOC_BIT); diff --git a/arch/sh/mm/tlb-sh4.c b/arch/sh/mm/tlb-sh4.c index 8cf550e..624c1da 100644 --- a/arch/sh/mm/tlb-sh4.c +++ b/arch/sh/mm/tlb-sh4.c @@ -64,8 +64,7 @@ void __update_tlb(struct vm_area_struct *vma, unsigned long address, pte_t pte) local_irq_restore(flags); } -void __uses_jump_to_uncached local_flush_tlb_one(unsigned long asid, - unsigned long page) +void local_flush_tlb_one(unsigned long asid, unsigned long page) { unsigned long addr, data; -- cgit v0.10.2 From a077e91690fb32a1453423b2cf1df3492fd30c3a Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 21 Jan 2010 16:16:36 +0900 Subject: sh: Conserve fixmap slots when IOREMAP_FIXED=n. Presently the IOREMAP_FIXED fixmaps are always defined, even if the platform isn't capable of supporting it. Since we already have an ifdef for it, ifdef the entries, too. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/fixmap.h b/arch/sh/include/asm/fixmap.h index 17b22b6c..6e7cea4 100644 --- a/arch/sh/include/asm/fixmap.h +++ b/arch/sh/include/asm/fixmap.h @@ -61,6 +61,7 @@ enum fixed_addresses { FIX_KMAP_END = FIX_KMAP_BEGIN+(KM_TYPE_NR*NR_CPUS)-1, #endif +#ifdef CONFIG_IOREMAP_FIXED /* * FIX_IOREMAP entries are useful for mapping physical address * space before ioremap() is useable, e.g. really early in boot @@ -69,6 +70,7 @@ enum fixed_addresses { #define FIX_N_IOREMAPS 32 FIX_IOREMAP_BEGIN, FIX_IOREMAP_END = FIX_IOREMAP_BEGIN + FIX_N_IOREMAPS, +#endif __end_of_fixed_addresses }; -- cgit v0.10.2 From ef12a141306c90336a3a10d40213ecd98624d274 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 20 Jan 2010 15:28:45 -0200 Subject: perf buildid-cache: Add new command to manage build-id cache MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit For now it just has operations to examine a given file, find its build-id and add or remove it to/from the cache. Useful, for instance, when adding binaries sent together with a perf.data file, so that we can add them to the cache and have the tools find it when resolving symbols. It'll also manage the size of the cache like 'ccache' does. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1264008525-29025-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Documentation/perf-buildid-cache.txt b/tools/perf/Documentation/perf-buildid-cache.txt new file mode 100644 index 0000000..88bc3b5 --- /dev/null +++ b/tools/perf/Documentation/perf-buildid-cache.txt @@ -0,0 +1,33 @@ +perf-buildid-cache(1) +===================== + +NAME +---- +perf-buildid-cache - Manage build-id cache. + +SYNOPSIS +-------- +[verse] +'perf buildid-list ' + +DESCRIPTION +----------- +This command manages the build-id cache. It can add and remove files to the +cache. In the future it should as well purge older entries, set upper limits +for the space used by the cache, etc. + +OPTIONS +------- +-a:: +--add=:: + Add specified file to the cache. +-r:: +--remove=:: + Remove specified file to the cache. +-v:: +--verbose:: + Be more verbose. + +SEE ALSO +-------- +linkperf:perf-record[1], linkperf:perf-report[1] diff --git a/tools/perf/Makefile b/tools/perf/Makefile index ddbeeee..9b173e6 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -445,6 +445,7 @@ BUILTIN_OBJS += builtin-diff.o BUILTIN_OBJS += builtin-help.o BUILTIN_OBJS += builtin-sched.o BUILTIN_OBJS += builtin-buildid-list.o +BUILTIN_OBJS += builtin-buildid-cache.o BUILTIN_OBJS += builtin-list.o BUILTIN_OBJS += builtin-record.o BUILTIN_OBJS += builtin-report.o diff --git a/tools/perf/builtin-buildid-cache.c b/tools/perf/builtin-buildid-cache.c new file mode 100644 index 0000000..30a05f5 --- /dev/null +++ b/tools/perf/builtin-buildid-cache.c @@ -0,0 +1,133 @@ +/* + * builtin-buildid-cache.c + * + * Builtin buildid-cache command: Manages build-id cache + * + * Copyright (C) 2010, Red Hat Inc. + * Copyright (C) 2010, Arnaldo Carvalho de Melo + */ +#include "builtin.h" +#include "perf.h" +#include "util/cache.h" +#include "util/debug.h" +#include "util/header.h" +#include "util/parse-options.h" +#include "util/strlist.h" +#include "util/symbol.h" + +static char const *add_name_list_str, *remove_name_list_str; + +static const char * const buildid_cache_usage[] = { + "perf buildid-cache []", + NULL +}; + +static const struct option buildid_cache_options[] = { + OPT_STRING('a', "add", &add_name_list_str, + "file list", "file(s) to add"), + OPT_STRING('r', "remove", &remove_name_list_str, "file list", + "file(s) to remove"), + OPT_BOOLEAN('v', "verbose", &verbose, "be more verbose"), + OPT_END() +}; + +static int build_id_cache__add_file(const char *filename, const char *debugdir) +{ + char sbuild_id[BUILD_ID_SIZE * 2 + 1]; + u8 build_id[BUILD_ID_SIZE]; + int err; + + if (filename__read_build_id(filename, &build_id, sizeof(build_id)) < 0) { + pr_debug("Couldn't read a build-id in %s\n", filename); + return -1; + } + + build_id__sprintf(build_id, sizeof(build_id), sbuild_id); + err = build_id_cache__add_s(sbuild_id, debugdir, filename, false); + if (verbose) + pr_info("Adding %s %s: %s\n", sbuild_id, filename, + err ? "FAIL" : "Ok"); + return err; +} + +static int build_id_cache__remove_file(const char *filename __used, + const char *debugdir __used) +{ + u8 build_id[BUILD_ID_SIZE]; + char sbuild_id[BUILD_ID_SIZE * 2 + 1]; + + int err; + + if (filename__read_build_id(filename, &build_id, sizeof(build_id)) < 0) { + pr_debug("Couldn't read a build-id in %s\n", filename); + return -1; + } + + build_id__sprintf(build_id, sizeof(build_id), sbuild_id); + err = build_id_cache__remove_s(sbuild_id, debugdir); + if (verbose) + pr_info("Removing %s %s: %s\n", sbuild_id, filename, + err ? "FAIL" : "Ok"); + + return err; +} + +static int __cmd_buildid_cache(void) +{ + struct strlist *list; + struct str_node *pos; + char debugdir[PATH_MAX]; + + snprintf(debugdir, sizeof(debugdir), "%s/%s", getenv("HOME"), + DEBUG_CACHE_DIR); + + if (add_name_list_str) { + list = strlist__new(true, add_name_list_str); + if (list) { + strlist__for_each(pos, list) + if (build_id_cache__add_file(pos->s, debugdir)) { + if (errno == EEXIST) { + pr_debug("%s already in the cache\n", + pos->s); + continue; + } + pr_warning("Couldn't add %s: %s\n", + pos->s, strerror(errno)); + } + + strlist__delete(list); + } + } + + if (remove_name_list_str) { + list = strlist__new(true, remove_name_list_str); + if (list) { + strlist__for_each(pos, list) + if (build_id_cache__remove_file(pos->s, debugdir)) { + if (errno == ENOENT) { + pr_debug("%s wasn't in the cache\n", + pos->s); + continue; + } + pr_warning("Couldn't remove %s: %s\n", + pos->s, strerror(errno)); + } + + strlist__delete(list); + } + } + + return 0; +} + +int cmd_buildid_cache(int argc, const char **argv, const char *prefix __used) +{ + argc = parse_options(argc, argv, buildid_cache_options, + buildid_cache_usage, 0); + + if (symbol__init() < 0) + return -1; + + setup_pager(); + return __cmd_buildid_cache(); +} diff --git a/tools/perf/builtin.h b/tools/perf/builtin.h index 18035b1..dee97cf 100644 --- a/tools/perf/builtin.h +++ b/tools/perf/builtin.h @@ -16,6 +16,7 @@ extern int check_pager_config(const char *cmd); extern int cmd_annotate(int argc, const char **argv, const char *prefix); extern int cmd_bench(int argc, const char **argv, const char *prefix); +extern int cmd_buildid_cache(int argc, const char **argv, const char *prefix); extern int cmd_buildid_list(int argc, const char **argv, const char *prefix); extern int cmd_diff(int argc, const char **argv, const char *prefix); extern int cmd_help(int argc, const char **argv, const char *prefix); diff --git a/tools/perf/command-list.txt b/tools/perf/command-list.txt index cf6444d..9afcff2 100644 --- a/tools/perf/command-list.txt +++ b/tools/perf/command-list.txt @@ -5,6 +5,7 @@ perf-annotate mainporcelain common perf-archive mainporcelain common perf-bench mainporcelain common +perf-buildid-cache mainporcelain common perf-buildid-list mainporcelain common perf-diff mainporcelain common perf-list mainporcelain common diff --git a/tools/perf/perf.c b/tools/perf/perf.c index fc89005..05c861c 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -285,6 +285,7 @@ static void handle_internal_command(int argc, const char **argv) { const char *cmd = argv[0]; static struct cmd_struct commands[] = { + { "buildid-cache", cmd_buildid_cache, 0 }, { "buildid-list", cmd_buildid_list, 0 }, { "diff", cmd_diff, 0 }, { "help", cmd_help, 0 }, diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 1b65fed..2bb2bdb 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -231,32 +231,29 @@ static int dsos__write_buildid_table(int fd) return err; } -static int dso__cache_build_id(struct dso *self, const char *debugdir) +int build_id_cache__add_s(const char *sbuild_id, const char *debugdir, + const char *name, bool is_kallsyms) { const size_t size = PATH_MAX; char *filename = malloc(size), - *linkname = malloc(size), *targetname, *sbuild_id; + *linkname = malloc(size), *targetname; int len, err = -1; - bool is_kallsyms = self->kernel && self->long_name[0] != '/'; if (filename == NULL || linkname == NULL) goto out_free; len = snprintf(filename, size, "%s%s%s", - debugdir, is_kallsyms ? "/" : "", self->long_name); + debugdir, is_kallsyms ? "/" : "", name); if (mkdir_p(filename, 0755)) goto out_free; - len += snprintf(filename + len, sizeof(filename) - len, "/"); - sbuild_id = filename + len; - build_id__sprintf(self->build_id, sizeof(self->build_id), sbuild_id); + snprintf(filename + len, sizeof(filename) - len, "/%s", sbuild_id); if (access(filename, F_OK)) { if (is_kallsyms) { if (copyfile("/proc/kallsyms", filename)) goto out_free; - } else if (link(self->long_name, filename) && - copyfile(self->long_name, filename)) + } else if (link(name, filename) && copyfile(name, filename)) goto out_free; } @@ -278,6 +275,63 @@ out_free: return err; } +static int build_id_cache__add_b(const u8 *build_id, size_t build_id_size, + const char *name, const char *debugdir, + bool is_kallsyms) +{ + char sbuild_id[BUILD_ID_SIZE * 2 + 1]; + + build_id__sprintf(build_id, build_id_size, sbuild_id); + + return build_id_cache__add_s(sbuild_id, debugdir, name, is_kallsyms); +} + +int build_id_cache__remove_s(const char *sbuild_id, const char *debugdir) +{ + const size_t size = PATH_MAX; + char *filename = malloc(size), + *linkname = malloc(size); + int err = -1; + + if (filename == NULL || linkname == NULL) + goto out_free; + + snprintf(linkname, size, "%s/.build-id/%.2s/%s", + debugdir, sbuild_id, sbuild_id + 2); + + if (access(linkname, F_OK)) + goto out_free; + + if (readlink(linkname, filename, size) < 0) + goto out_free; + + if (unlink(linkname)) + goto out_free; + + /* + * Since the link is relative, we must make it absolute: + */ + snprintf(linkname, size, "%s/.build-id/%.2s/%s", + debugdir, sbuild_id, filename); + + if (unlink(linkname)) + goto out_free; + + err = 0; +out_free: + free(filename); + free(linkname); + return err; +} + +static int dso__cache_build_id(struct dso *self, const char *debugdir) +{ + bool is_kallsyms = self->kernel && self->long_name[0] != '/'; + + return build_id_cache__add_b(self->build_id, sizeof(self->build_id), + self->long_name, debugdir, is_kallsyms); +} + static int __dsos__cache_build_ids(struct list_head *head, const char *debugdir) { struct dso *pos; diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h index ccc8540..82a6af7 100644 --- a/tools/perf/util/header.h +++ b/tools/perf/util/header.h @@ -5,6 +5,7 @@ #include #include #include "types.h" +#include "event.h" #include @@ -84,4 +85,8 @@ int perf_header__process_sections(struct perf_header *self, int fd, struct perf_header *ph, int feat, int fd)); +int build_id_cache__add_s(const char *sbuild_id, const char *debugdir, + const char *name, bool is_kallsyms); +int build_id_cache__remove_s(const char *sbuild_id, const char *debugdir); + #endif /* __PERF_HEADER_H */ diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index b6ab23d..6f30fe1 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -345,10 +345,10 @@ void dso__sort_by_name(struct dso *self, enum map_type type) &self->symbols[type]); } -int build_id__sprintf(u8 *self, int len, char *bf) +int build_id__sprintf(const u8 *self, int len, char *bf) { char *bid = bf; - u8 *raw = self; + const u8 *raw = self; int i; for (i = 0; i < len; ++i) { diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 525085fd..ffe0b0f 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -144,7 +144,7 @@ struct symbol *dso__find_symbol_by_name(struct dso *self, enum map_type type, int filename__read_build_id(const char *filename, void *bf, size_t size); int sysfs__read_build_id(const char *filename, void *bf, size_t size); bool dsos__read_build_ids(void); -int build_id__sprintf(u8 *self, int len, char *bf); +int build_id__sprintf(const u8 *self, int len, char *bf); int kallsyms__parse(const char *filename, void *arg, int (*process_symbol)(void *arg, const char *name, char type, u64 start)); -- cgit v0.10.2 From 1e3c88bdeb1260edc341e45c9fb8efd182a5c511 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 17:00:43 +0100 Subject: sched: Move load balance code into sched_fair.c Straight fwd code movement. Since non of the load-balance abstractions are used anymore, do away with them and simplify the code some. In preparation move the code around. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 64298a5..13a2acf 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1805,6 +1805,51 @@ static inline void double_unlock_balance(struct rq *this_rq, struct rq *busiest) raw_spin_unlock(&busiest->lock); lock_set_subclass(&this_rq->lock.dep_map, 0, _RET_IP_); } + +/* + * double_rq_lock - safely lock two runqueues + * + * Note this does not disable interrupts like task_rq_lock, + * you need to do so manually before calling. + */ +static void double_rq_lock(struct rq *rq1, struct rq *rq2) + __acquires(rq1->lock) + __acquires(rq2->lock) +{ + BUG_ON(!irqs_disabled()); + if (rq1 == rq2) { + raw_spin_lock(&rq1->lock); + __acquire(rq2->lock); /* Fake it out ;) */ + } else { + if (rq1 < rq2) { + raw_spin_lock(&rq1->lock); + raw_spin_lock_nested(&rq2->lock, SINGLE_DEPTH_NESTING); + } else { + raw_spin_lock(&rq2->lock); + raw_spin_lock_nested(&rq1->lock, SINGLE_DEPTH_NESTING); + } + } + update_rq_clock(rq1); + update_rq_clock(rq2); +} + +/* + * double_rq_unlock - safely unlock two runqueues + * + * Note this does not restore interrupts like task_rq_unlock, + * you need to do so manually after calling. + */ +static void double_rq_unlock(struct rq *rq1, struct rq *rq2) + __releases(rq1->lock) + __releases(rq2->lock) +{ + raw_spin_unlock(&rq1->lock); + if (rq1 != rq2) + raw_spin_unlock(&rq2->lock); + else + __release(rq2->lock); +} + #endif #ifdef CONFIG_FAIR_GROUP_SCHED @@ -1834,18 +1879,14 @@ static inline void __set_task_cpu(struct task_struct *p, unsigned int cpu) #endif } -#include "sched_stats.h" -#include "sched_idletask.c" -#include "sched_fair.c" -#include "sched_rt.c" -#ifdef CONFIG_SCHED_DEBUG -# include "sched_debug.c" -#endif +static const struct sched_class rt_sched_class; #define sched_class_highest (&rt_sched_class) #define for_each_class(class) \ for (class = sched_class_highest; class; class = class->next) +#include "sched_stats.h" + static void inc_nr_running(struct rq *rq) { rq->nr_running++; @@ -1912,6 +1953,37 @@ static void dequeue_task(struct rq *rq, struct task_struct *p, int sleep) } /* + * activate_task - move a task to the runqueue. + */ +static void activate_task(struct rq *rq, struct task_struct *p, int wakeup) +{ + if (task_contributes_to_load(p)) + rq->nr_uninterruptible--; + + enqueue_task(rq, p, wakeup); + inc_nr_running(rq); +} + +/* + * deactivate_task - remove a task from the runqueue. + */ +static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep) +{ + if (task_contributes_to_load(p)) + rq->nr_uninterruptible++; + + dequeue_task(rq, p, sleep); + dec_nr_running(rq); +} + +#include "sched_idletask.c" +#include "sched_fair.c" +#include "sched_rt.c" +#ifdef CONFIG_SCHED_DEBUG +# include "sched_debug.c" +#endif + +/* * __normal_prio - return the priority that is based on the static prio */ static inline int __normal_prio(struct task_struct *p) @@ -1957,30 +2029,6 @@ static int effective_prio(struct task_struct *p) return p->prio; } -/* - * activate_task - move a task to the runqueue. - */ -static void activate_task(struct rq *rq, struct task_struct *p, int wakeup) -{ - if (task_contributes_to_load(p)) - rq->nr_uninterruptible--; - - enqueue_task(rq, p, wakeup); - inc_nr_running(rq); -} - -/* - * deactivate_task - remove a task from the runqueue. - */ -static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep) -{ - if (task_contributes_to_load(p)) - rq->nr_uninterruptible++; - - dequeue_task(rq, p, sleep); - dec_nr_running(rq); -} - /** * task_curr - is this task currently executing on a CPU? * @p: the task in question. @@ -3088,50 +3136,6 @@ static void update_cpu_load(struct rq *this_rq) #ifdef CONFIG_SMP /* - * double_rq_lock - safely lock two runqueues - * - * Note this does not disable interrupts like task_rq_lock, - * you need to do so manually before calling. - */ -static void double_rq_lock(struct rq *rq1, struct rq *rq2) - __acquires(rq1->lock) - __acquires(rq2->lock) -{ - BUG_ON(!irqs_disabled()); - if (rq1 == rq2) { - raw_spin_lock(&rq1->lock); - __acquire(rq2->lock); /* Fake it out ;) */ - } else { - if (rq1 < rq2) { - raw_spin_lock(&rq1->lock); - raw_spin_lock_nested(&rq2->lock, SINGLE_DEPTH_NESTING); - } else { - raw_spin_lock(&rq2->lock); - raw_spin_lock_nested(&rq1->lock, SINGLE_DEPTH_NESTING); - } - } - update_rq_clock(rq1); - update_rq_clock(rq2); -} - -/* - * double_rq_unlock - safely unlock two runqueues - * - * Note this does not restore interrupts like task_rq_unlock, - * you need to do so manually after calling. - */ -static void double_rq_unlock(struct rq *rq1, struct rq *rq2) - __releases(rq1->lock) - __releases(rq2->lock) -{ - raw_spin_unlock(&rq1->lock); - if (rq1 != rq2) - raw_spin_unlock(&rq2->lock); - else - __release(rq2->lock); -} - -/* * sched_exec - execve() is a valuable balancing opportunity, because at * this point the task has the smallest effective memory and cache footprint. */ @@ -3179,1771 +3183,6 @@ again: task_rq_unlock(rq, &flags); } -/* - * pull_task - move a task from a remote runqueue to the local runqueue. - * Both runqueues must be locked. - */ -static void pull_task(struct rq *src_rq, struct task_struct *p, - struct rq *this_rq, int this_cpu) -{ - deactivate_task(src_rq, p, 0); - set_task_cpu(p, this_cpu); - activate_task(this_rq, p, 0); - check_preempt_curr(this_rq, p, 0); -} - -/* - * can_migrate_task - may task p from runqueue rq be migrated to this_cpu? - */ -static -int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned) -{ - int tsk_cache_hot = 0; - /* - * We do not migrate tasks that are: - * 1) running (obviously), or - * 2) cannot be migrated to this CPU due to cpus_allowed, or - * 3) are cache-hot on their current CPU. - */ - if (!cpumask_test_cpu(this_cpu, &p->cpus_allowed)) { - schedstat_inc(p, se.nr_failed_migrations_affine); - return 0; - } - *all_pinned = 0; - - if (task_running(rq, p)) { - schedstat_inc(p, se.nr_failed_migrations_running); - return 0; - } - - /* - * Aggressive migration if: - * 1) task is cache cold, or - * 2) too many balance attempts have failed. - */ - - tsk_cache_hot = task_hot(p, rq->clock, sd); - if (!tsk_cache_hot || - sd->nr_balance_failed > sd->cache_nice_tries) { -#ifdef CONFIG_SCHEDSTATS - if (tsk_cache_hot) { - schedstat_inc(sd, lb_hot_gained[idle]); - schedstat_inc(p, se.nr_forced_migrations); - } -#endif - return 1; - } - - if (tsk_cache_hot) { - schedstat_inc(p, se.nr_failed_migrations_hot); - return 0; - } - return 1; -} - -static unsigned long -balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct rq_iterator *iterator) -{ - int loops = 0, pulled = 0, pinned = 0; - struct task_struct *p; - long rem_load_move = max_load_move; - - if (max_load_move == 0) - goto out; - - pinned = 1; - - /* - * Start the load-balancing iterator: - */ - p = iterator->start(iterator->arg); -next: - if (!p || loops++ > sysctl_sched_nr_migrate) - goto out; - - if ((p->se.load.weight >> 1) > rem_load_move || - !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { - p = iterator->next(iterator->arg); - goto next; - } - - pull_task(busiest, p, this_rq, this_cpu); - pulled++; - rem_load_move -= p->se.load.weight; - -#ifdef CONFIG_PREEMPT - /* - * NEWIDLE balancing is a source of latency, so preemptible kernels - * will stop after the first task is pulled to minimize the critical - * section. - */ - if (idle == CPU_NEWLY_IDLE) - goto out; -#endif - - /* - * We only want to steal up to the prescribed amount of weighted load. - */ - if (rem_load_move > 0) { - if (p->prio < *this_best_prio) - *this_best_prio = p->prio; - p = iterator->next(iterator->arg); - goto next; - } -out: - /* - * Right now, this is one of only two places pull_task() is called, - * so we can safely collect pull_task() stats here rather than - * inside pull_task(). - */ - schedstat_add(sd, lb_gained[idle], pulled); - - if (all_pinned) - *all_pinned = pinned; - - return max_load_move - rem_load_move; -} - -/* - * move_tasks tries to move up to max_load_move weighted load from busiest to - * this_rq, as part of a balancing operation within domain "sd". - * Returns 1 if successful and 0 otherwise. - * - * Called with both runqueues locked. - */ -static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned) -{ - const struct sched_class *class = sched_class_highest; - unsigned long total_load_moved = 0; - int this_best_prio = this_rq->curr->prio; - - do { - total_load_moved += - class->load_balance(this_rq, this_cpu, busiest, - max_load_move - total_load_moved, - sd, idle, all_pinned, &this_best_prio); - class = class->next; - -#ifdef CONFIG_PREEMPT - /* - * NEWIDLE balancing is a source of latency, so preemptible - * kernels will stop after the first task is pulled to minimize - * the critical section. - */ - if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) - break; -#endif - } while (class && max_load_move > total_load_moved); - - return total_load_moved > 0; -} - -static int -iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle, - struct rq_iterator *iterator) -{ - struct task_struct *p = iterator->start(iterator->arg); - int pinned = 0; - - while (p) { - if (can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { - pull_task(busiest, p, this_rq, this_cpu); - /* - * Right now, this is only the second place pull_task() - * is called, so we can safely collect pull_task() - * stats here rather than inside pull_task(). - */ - schedstat_inc(sd, lb_gained[idle]); - - return 1; - } - p = iterator->next(iterator->arg); - } - - return 0; -} - -/* - * move_one_task tries to move exactly one task from busiest to this_rq, as - * part of active balancing operations within "domain". - * Returns 1 if successful and 0 otherwise. - * - * Called with both runqueues locked. - */ -static int move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - const struct sched_class *class; - - for_each_class(class) { - if (class->move_one_task(this_rq, this_cpu, busiest, sd, idle)) - return 1; - } - - return 0; -} -/********** Helpers for find_busiest_group ************************/ -/* - * sd_lb_stats - Structure to store the statistics of a sched_domain - * during load balancing. - */ -struct sd_lb_stats { - struct sched_group *busiest; /* Busiest group in this sd */ - struct sched_group *this; /* Local group in this sd */ - unsigned long total_load; /* Total load of all groups in sd */ - unsigned long total_pwr; /* Total power of all groups in sd */ - unsigned long avg_load; /* Average load across all groups in sd */ - - /** Statistics of this group */ - unsigned long this_load; - unsigned long this_load_per_task; - unsigned long this_nr_running; - - /* Statistics of the busiest group */ - unsigned long max_load; - unsigned long busiest_load_per_task; - unsigned long busiest_nr_running; - - int group_imb; /* Is there imbalance in this sd */ -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) - int power_savings_balance; /* Is powersave balance needed for this sd */ - struct sched_group *group_min; /* Least loaded group in sd */ - struct sched_group *group_leader; /* Group which relieves group_min */ - unsigned long min_load_per_task; /* load_per_task in group_min */ - unsigned long leader_nr_running; /* Nr running of group_leader */ - unsigned long min_nr_running; /* Nr running of group_min */ -#endif -}; - -/* - * sg_lb_stats - stats of a sched_group required for load_balancing - */ -struct sg_lb_stats { - unsigned long avg_load; /*Avg load across the CPUs of the group */ - unsigned long group_load; /* Total load over the CPUs of the group */ - unsigned long sum_nr_running; /* Nr tasks running in the group */ - unsigned long sum_weighted_load; /* Weighted load of group's tasks */ - unsigned long group_capacity; - int group_imb; /* Is there an imbalance in the group ? */ -}; - -/** - * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. - * @group: The group whose first cpu is to be returned. - */ -static inline unsigned int group_first_cpu(struct sched_group *group) -{ - return cpumask_first(sched_group_cpus(group)); -} - -/** - * get_sd_load_idx - Obtain the load index for a given sched domain. - * @sd: The sched_domain whose load_idx is to be obtained. - * @idle: The Idle status of the CPU for whose sd load_icx is obtained. - */ -static inline int get_sd_load_idx(struct sched_domain *sd, - enum cpu_idle_type idle) -{ - int load_idx; - - switch (idle) { - case CPU_NOT_IDLE: - load_idx = sd->busy_idx; - break; - - case CPU_NEWLY_IDLE: - load_idx = sd->newidle_idx; - break; - default: - load_idx = sd->idle_idx; - break; - } - - return load_idx; -} - - -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) -/** - * init_sd_power_savings_stats - Initialize power savings statistics for - * the given sched_domain, during load balancing. - * - * @sd: Sched domain whose power-savings statistics are to be initialized. - * @sds: Variable containing the statistics for sd. - * @idle: Idle status of the CPU at which we're performing load-balancing. - */ -static inline void init_sd_power_savings_stats(struct sched_domain *sd, - struct sd_lb_stats *sds, enum cpu_idle_type idle) -{ - /* - * Busy processors will not participate in power savings - * balance. - */ - if (idle == CPU_NOT_IDLE || !(sd->flags & SD_POWERSAVINGS_BALANCE)) - sds->power_savings_balance = 0; - else { - sds->power_savings_balance = 1; - sds->min_nr_running = ULONG_MAX; - sds->leader_nr_running = 0; - } -} - -/** - * update_sd_power_savings_stats - Update the power saving stats for a - * sched_domain while performing load balancing. - * - * @group: sched_group belonging to the sched_domain under consideration. - * @sds: Variable containing the statistics of the sched_domain - * @local_group: Does group contain the CPU for which we're performing - * load balancing ? - * @sgs: Variable containing the statistics of the group. - */ -static inline void update_sd_power_savings_stats(struct sched_group *group, - struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) -{ - - if (!sds->power_savings_balance) - return; - - /* - * If the local group is idle or completely loaded - * no need to do power savings balance at this domain - */ - if (local_group && (sds->this_nr_running >= sgs->group_capacity || - !sds->this_nr_running)) - sds->power_savings_balance = 0; - - /* - * If a group is already running at full capacity or idle, - * don't include that group in power savings calculations - */ - if (!sds->power_savings_balance || - sgs->sum_nr_running >= sgs->group_capacity || - !sgs->sum_nr_running) - return; - - /* - * Calculate the group which has the least non-idle load. - * This is the group from where we need to pick up the load - * for saving power - */ - if ((sgs->sum_nr_running < sds->min_nr_running) || - (sgs->sum_nr_running == sds->min_nr_running && - group_first_cpu(group) > group_first_cpu(sds->group_min))) { - sds->group_min = group; - sds->min_nr_running = sgs->sum_nr_running; - sds->min_load_per_task = sgs->sum_weighted_load / - sgs->sum_nr_running; - } - - /* - * Calculate the group which is almost near its - * capacity but still has some space to pick up some load - * from other group and save more power - */ - if (sgs->sum_nr_running + 1 > sgs->group_capacity) - return; - - if (sgs->sum_nr_running > sds->leader_nr_running || - (sgs->sum_nr_running == sds->leader_nr_running && - group_first_cpu(group) < group_first_cpu(sds->group_leader))) { - sds->group_leader = group; - sds->leader_nr_running = sgs->sum_nr_running; - } -} - -/** - * check_power_save_busiest_group - see if there is potential for some power-savings balance - * @sds: Variable containing the statistics of the sched_domain - * under consideration. - * @this_cpu: Cpu at which we're currently performing load-balancing. - * @imbalance: Variable to store the imbalance. - * - * Description: - * Check if we have potential to perform some power-savings balance. - * If yes, set the busiest group to be the least loaded group in the - * sched_domain, so that it's CPUs can be put to idle. - * - * Returns 1 if there is potential to perform power-savings balance. - * Else returns 0. - */ -static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, - int this_cpu, unsigned long *imbalance) -{ - if (!sds->power_savings_balance) - return 0; - - if (sds->this != sds->group_leader || - sds->group_leader == sds->group_min) - return 0; - - *imbalance = sds->min_load_per_task; - sds->busiest = sds->group_min; - - return 1; - -} -#else /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ -static inline void init_sd_power_savings_stats(struct sched_domain *sd, - struct sd_lb_stats *sds, enum cpu_idle_type idle) -{ - return; -} - -static inline void update_sd_power_savings_stats(struct sched_group *group, - struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) -{ - return; -} - -static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, - int this_cpu, unsigned long *imbalance) -{ - return 0; -} -#endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ - - -unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu) -{ - return SCHED_LOAD_SCALE; -} - -unsigned long __weak arch_scale_freq_power(struct sched_domain *sd, int cpu) -{ - return default_scale_freq_power(sd, cpu); -} - -unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu) -{ - unsigned long weight = cpumask_weight(sched_domain_span(sd)); - unsigned long smt_gain = sd->smt_gain; - - smt_gain /= weight; - - return smt_gain; -} - -unsigned long __weak arch_scale_smt_power(struct sched_domain *sd, int cpu) -{ - return default_scale_smt_power(sd, cpu); -} - -unsigned long scale_rt_power(int cpu) -{ - struct rq *rq = cpu_rq(cpu); - u64 total, available; - - sched_avg_update(rq); - - total = sched_avg_period() + (rq->clock - rq->age_stamp); - available = total - rq->rt_avg; - - if (unlikely((s64)total < SCHED_LOAD_SCALE)) - total = SCHED_LOAD_SCALE; - - total >>= SCHED_LOAD_SHIFT; - - return div_u64(available, total); -} - -static void update_cpu_power(struct sched_domain *sd, int cpu) -{ - unsigned long weight = cpumask_weight(sched_domain_span(sd)); - unsigned long power = SCHED_LOAD_SCALE; - struct sched_group *sdg = sd->groups; - - if (sched_feat(ARCH_POWER)) - power *= arch_scale_freq_power(sd, cpu); - else - power *= default_scale_freq_power(sd, cpu); - - power >>= SCHED_LOAD_SHIFT; - - if ((sd->flags & SD_SHARE_CPUPOWER) && weight > 1) { - if (sched_feat(ARCH_POWER)) - power *= arch_scale_smt_power(sd, cpu); - else - power *= default_scale_smt_power(sd, cpu); - - power >>= SCHED_LOAD_SHIFT; - } - - power *= scale_rt_power(cpu); - power >>= SCHED_LOAD_SHIFT; - - if (!power) - power = 1; - - sdg->cpu_power = power; -} - -static void update_group_power(struct sched_domain *sd, int cpu) -{ - struct sched_domain *child = sd->child; - struct sched_group *group, *sdg = sd->groups; - unsigned long power; - - if (!child) { - update_cpu_power(sd, cpu); - return; - } - - power = 0; - - group = child->groups; - do { - power += group->cpu_power; - group = group->next; - } while (group != child->groups); - - sdg->cpu_power = power; -} - -/** - * update_sg_lb_stats - Update sched_group's statistics for load balancing. - * @sd: The sched_domain whose statistics are to be updated. - * @group: sched_group whose statistics are to be updated. - * @this_cpu: Cpu for which load balance is currently performed. - * @idle: Idle status of this_cpu - * @load_idx: Load index of sched_domain of this_cpu for load calc. - * @sd_idle: Idle status of the sched_domain containing group. - * @local_group: Does group contain this_cpu. - * @cpus: Set of cpus considered for load balancing. - * @balance: Should we balance. - * @sgs: variable to hold the statistics for this group. - */ -static inline void update_sg_lb_stats(struct sched_domain *sd, - struct sched_group *group, int this_cpu, - enum cpu_idle_type idle, int load_idx, int *sd_idle, - int local_group, const struct cpumask *cpus, - int *balance, struct sg_lb_stats *sgs) -{ - unsigned long load, max_cpu_load, min_cpu_load; - int i; - unsigned int balance_cpu = -1, first_idle_cpu = 0; - unsigned long sum_avg_load_per_task; - unsigned long avg_load_per_task; - - if (local_group) { - balance_cpu = group_first_cpu(group); - if (balance_cpu == this_cpu) - update_group_power(sd, this_cpu); - } - - /* Tally up the load of all CPUs in the group */ - sum_avg_load_per_task = avg_load_per_task = 0; - max_cpu_load = 0; - min_cpu_load = ~0UL; - - for_each_cpu_and(i, sched_group_cpus(group), cpus) { - struct rq *rq = cpu_rq(i); - - if (*sd_idle && rq->nr_running) - *sd_idle = 0; - - /* Bias balancing toward cpus of our domain */ - if (local_group) { - if (idle_cpu(i) && !first_idle_cpu) { - first_idle_cpu = 1; - balance_cpu = i; - } - - load = target_load(i, load_idx); - } else { - load = source_load(i, load_idx); - if (load > max_cpu_load) - max_cpu_load = load; - if (min_cpu_load > load) - min_cpu_load = load; - } - - sgs->group_load += load; - sgs->sum_nr_running += rq->nr_running; - sgs->sum_weighted_load += weighted_cpuload(i); - - sum_avg_load_per_task += cpu_avg_load_per_task(i); - } - - /* - * First idle cpu or the first cpu(busiest) in this sched group - * is eligible for doing load balancing at this and above - * domains. In the newly idle case, we will allow all the cpu's - * to do the newly idle load balance. - */ - if (idle != CPU_NEWLY_IDLE && local_group && - balance_cpu != this_cpu && balance) { - *balance = 0; - return; - } - - /* Adjust by relative CPU power of the group */ - sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; - - - /* - * Consider the group unbalanced when the imbalance is larger - * than the average weight of two tasks. - * - * APZ: with cgroup the avg task weight can vary wildly and - * might not be a suitable number - should we keep a - * normalized nr_running number somewhere that negates - * the hierarchy? - */ - avg_load_per_task = (sum_avg_load_per_task * SCHED_LOAD_SCALE) / - group->cpu_power; - - if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) - sgs->group_imb = 1; - - sgs->group_capacity = - DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); -} - -/** - * update_sd_lb_stats - Update sched_group's statistics for load balancing. - * @sd: sched_domain whose statistics are to be updated. - * @this_cpu: Cpu for which load balance is currently performed. - * @idle: Idle status of this_cpu - * @sd_idle: Idle status of the sched_domain containing group. - * @cpus: Set of cpus considered for load balancing. - * @balance: Should we balance. - * @sds: variable to hold the statistics for this sched_domain. - */ -static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, - enum cpu_idle_type idle, int *sd_idle, - const struct cpumask *cpus, int *balance, - struct sd_lb_stats *sds) -{ - struct sched_domain *child = sd->child; - struct sched_group *group = sd->groups; - struct sg_lb_stats sgs; - int load_idx, prefer_sibling = 0; - - if (child && child->flags & SD_PREFER_SIBLING) - prefer_sibling = 1; - - init_sd_power_savings_stats(sd, sds, idle); - load_idx = get_sd_load_idx(sd, idle); - - do { - int local_group; - - local_group = cpumask_test_cpu(this_cpu, - sched_group_cpus(group)); - memset(&sgs, 0, sizeof(sgs)); - update_sg_lb_stats(sd, group, this_cpu, idle, load_idx, sd_idle, - local_group, cpus, balance, &sgs); - - if (local_group && balance && !(*balance)) - return; - - sds->total_load += sgs.group_load; - sds->total_pwr += group->cpu_power; - - /* - * In case the child domain prefers tasks go to siblings - * first, lower the group capacity to one so that we'll try - * and move all the excess tasks away. - */ - if (prefer_sibling) - sgs.group_capacity = min(sgs.group_capacity, 1UL); - - if (local_group) { - sds->this_load = sgs.avg_load; - sds->this = group; - sds->this_nr_running = sgs.sum_nr_running; - sds->this_load_per_task = sgs.sum_weighted_load; - } else if (sgs.avg_load > sds->max_load && - (sgs.sum_nr_running > sgs.group_capacity || - sgs.group_imb)) { - sds->max_load = sgs.avg_load; - sds->busiest = group; - sds->busiest_nr_running = sgs.sum_nr_running; - sds->busiest_load_per_task = sgs.sum_weighted_load; - sds->group_imb = sgs.group_imb; - } - - update_sd_power_savings_stats(group, sds, local_group, &sgs); - group = group->next; - } while (group != sd->groups); -} - -/** - * fix_small_imbalance - Calculate the minor imbalance that exists - * amongst the groups of a sched_domain, during - * load balancing. - * @sds: Statistics of the sched_domain whose imbalance is to be calculated. - * @this_cpu: The cpu at whose sched_domain we're performing load-balance. - * @imbalance: Variable to store the imbalance. - */ -static inline void fix_small_imbalance(struct sd_lb_stats *sds, - int this_cpu, unsigned long *imbalance) -{ - unsigned long tmp, pwr_now = 0, pwr_move = 0; - unsigned int imbn = 2; - - if (sds->this_nr_running) { - sds->this_load_per_task /= sds->this_nr_running; - if (sds->busiest_load_per_task > - sds->this_load_per_task) - imbn = 1; - } else - sds->this_load_per_task = - cpu_avg_load_per_task(this_cpu); - - if (sds->max_load - sds->this_load + sds->busiest_load_per_task >= - sds->busiest_load_per_task * imbn) { - *imbalance = sds->busiest_load_per_task; - return; - } - - /* - * OK, we don't have enough imbalance to justify moving tasks, - * however we may be able to increase total CPU power used by - * moving them. - */ - - pwr_now += sds->busiest->cpu_power * - min(sds->busiest_load_per_task, sds->max_load); - pwr_now += sds->this->cpu_power * - min(sds->this_load_per_task, sds->this_load); - pwr_now /= SCHED_LOAD_SCALE; - - /* Amount of load we'd subtract */ - tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / - sds->busiest->cpu_power; - if (sds->max_load > tmp) - pwr_move += sds->busiest->cpu_power * - min(sds->busiest_load_per_task, sds->max_load - tmp); - - /* Amount of load we'd add */ - if (sds->max_load * sds->busiest->cpu_power < - sds->busiest_load_per_task * SCHED_LOAD_SCALE) - tmp = (sds->max_load * sds->busiest->cpu_power) / - sds->this->cpu_power; - else - tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / - sds->this->cpu_power; - pwr_move += sds->this->cpu_power * - min(sds->this_load_per_task, sds->this_load + tmp); - pwr_move /= SCHED_LOAD_SCALE; - - /* Move if we gain throughput */ - if (pwr_move > pwr_now) - *imbalance = sds->busiest_load_per_task; -} - -/** - * calculate_imbalance - Calculate the amount of imbalance present within the - * groups of a given sched_domain during load balance. - * @sds: statistics of the sched_domain whose imbalance is to be calculated. - * @this_cpu: Cpu for which currently load balance is being performed. - * @imbalance: The variable to store the imbalance. - */ -static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, - unsigned long *imbalance) -{ - unsigned long max_pull; - /* - * In the presence of smp nice balancing, certain scenarios can have - * max load less than avg load(as we skip the groups at or below - * its cpu_power, while calculating max_load..) - */ - if (sds->max_load < sds->avg_load) { - *imbalance = 0; - return fix_small_imbalance(sds, this_cpu, imbalance); - } - - /* Don't want to pull so many tasks that a group would go idle */ - max_pull = min(sds->max_load - sds->avg_load, - sds->max_load - sds->busiest_load_per_task); - - /* How much load to actually move to equalise the imbalance */ - *imbalance = min(max_pull * sds->busiest->cpu_power, - (sds->avg_load - sds->this_load) * sds->this->cpu_power) - / SCHED_LOAD_SCALE; - - /* - * if *imbalance is less than the average load per runnable task - * there is no gaurantee that any tasks will be moved so we'll have - * a think about bumping its value to force at least one task to be - * moved - */ - if (*imbalance < sds->busiest_load_per_task) - return fix_small_imbalance(sds, this_cpu, imbalance); - -} -/******* find_busiest_group() helpers end here *********************/ - -/** - * find_busiest_group - Returns the busiest group within the sched_domain - * if there is an imbalance. If there isn't an imbalance, and - * the user has opted for power-savings, it returns a group whose - * CPUs can be put to idle by rebalancing those tasks elsewhere, if - * such a group exists. - * - * Also calculates the amount of weighted load which should be moved - * to restore balance. - * - * @sd: The sched_domain whose busiest group is to be returned. - * @this_cpu: The cpu for which load balancing is currently being performed. - * @imbalance: Variable which stores amount of weighted load which should - * be moved to restore balance/put a group to idle. - * @idle: The idle status of this_cpu. - * @sd_idle: The idleness of sd - * @cpus: The set of CPUs under consideration for load-balancing. - * @balance: Pointer to a variable indicating if this_cpu - * is the appropriate cpu to perform load balancing at this_level. - * - * Returns: - the busiest group if imbalance exists. - * - If no imbalance and user has opted for power-savings balance, - * return the least loaded group whose CPUs can be - * put to idle by rebalancing its tasks onto our group. - */ -static struct sched_group * -find_busiest_group(struct sched_domain *sd, int this_cpu, - unsigned long *imbalance, enum cpu_idle_type idle, - int *sd_idle, const struct cpumask *cpus, int *balance) -{ - struct sd_lb_stats sds; - - memset(&sds, 0, sizeof(sds)); - - /* - * Compute the various statistics relavent for load balancing at - * this level. - */ - update_sd_lb_stats(sd, this_cpu, idle, sd_idle, cpus, - balance, &sds); - - /* Cases where imbalance does not exist from POV of this_cpu */ - /* 1) this_cpu is not the appropriate cpu to perform load balancing - * at this level. - * 2) There is no busy sibling group to pull from. - * 3) This group is the busiest group. - * 4) This group is more busy than the avg busieness at this - * sched_domain. - * 5) The imbalance is within the specified limit. - * 6) Any rebalance would lead to ping-pong - */ - if (balance && !(*balance)) - goto ret; - - if (!sds.busiest || sds.busiest_nr_running == 0) - goto out_balanced; - - if (sds.this_load >= sds.max_load) - goto out_balanced; - - sds.avg_load = (SCHED_LOAD_SCALE * sds.total_load) / sds.total_pwr; - - if (sds.this_load >= sds.avg_load) - goto out_balanced; - - if (100 * sds.max_load <= sd->imbalance_pct * sds.this_load) - goto out_balanced; - - sds.busiest_load_per_task /= sds.busiest_nr_running; - if (sds.group_imb) - sds.busiest_load_per_task = - min(sds.busiest_load_per_task, sds.avg_load); - - /* - * We're trying to get all the cpus to the average_load, so we don't - * want to push ourselves above the average load, nor do we wish to - * reduce the max loaded cpu below the average load, as either of these - * actions would just result in more rebalancing later, and ping-pong - * tasks around. Thus we look for the minimum possible imbalance. - * Negative imbalances (*we* are more loaded than anyone else) will - * be counted as no imbalance for these purposes -- we can't fix that - * by pulling tasks to us. Be careful of negative numbers as they'll - * appear as very large values with unsigned longs. - */ - if (sds.max_load <= sds.busiest_load_per_task) - goto out_balanced; - - /* Looks like there is an imbalance. Compute it */ - calculate_imbalance(&sds, this_cpu, imbalance); - return sds.busiest; - -out_balanced: - /* - * There is no obvious imbalance. But check if we can do some balancing - * to save power. - */ - if (check_power_save_busiest_group(&sds, this_cpu, imbalance)) - return sds.busiest; -ret: - *imbalance = 0; - return NULL; -} - -/* - * find_busiest_queue - find the busiest runqueue among the cpus in group. - */ -static struct rq * -find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, - unsigned long imbalance, const struct cpumask *cpus) -{ - struct rq *busiest = NULL, *rq; - unsigned long max_load = 0; - int i; - - for_each_cpu(i, sched_group_cpus(group)) { - unsigned long power = power_of(i); - unsigned long capacity = DIV_ROUND_CLOSEST(power, SCHED_LOAD_SCALE); - unsigned long wl; - - if (!cpumask_test_cpu(i, cpus)) - continue; - - rq = cpu_rq(i); - wl = weighted_cpuload(i) * SCHED_LOAD_SCALE; - wl /= power; - - if (capacity && rq->nr_running == 1 && wl > imbalance) - continue; - - if (wl > max_load) { - max_load = wl; - busiest = rq; - } - } - - return busiest; -} - -/* - * Max backoff if we encounter pinned tasks. Pretty arbitrary value, but - * so long as it is large enough. - */ -#define MAX_PINNED_INTERVAL 512 - -/* Working cpumask for load_balance and load_balance_newidle. */ -static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); - -/* - * Check this_cpu to ensure it is balanced within domain. Attempt to move - * tasks if there is an imbalance. - */ -static int load_balance(int this_cpu, struct rq *this_rq, - struct sched_domain *sd, enum cpu_idle_type idle, - int *balance) -{ - int ld_moved, all_pinned = 0, active_balance = 0, sd_idle = 0; - struct sched_group *group; - unsigned long imbalance; - struct rq *busiest; - unsigned long flags; - struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); - - cpumask_copy(cpus, cpu_active_mask); - - /* - * When power savings policy is enabled for the parent domain, idle - * sibling can pick up load irrespective of busy siblings. In this case, - * let the state of idle sibling percolate up as CPU_IDLE, instead of - * portraying it as CPU_NOT_IDLE. - */ - if (idle != CPU_NOT_IDLE && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - sd_idle = 1; - - schedstat_inc(sd, lb_count[idle]); - -redo: - update_shares(sd); - group = find_busiest_group(sd, this_cpu, &imbalance, idle, &sd_idle, - cpus, balance); - - if (*balance == 0) - goto out_balanced; - - if (!group) { - schedstat_inc(sd, lb_nobusyg[idle]); - goto out_balanced; - } - - busiest = find_busiest_queue(group, idle, imbalance, cpus); - if (!busiest) { - schedstat_inc(sd, lb_nobusyq[idle]); - goto out_balanced; - } - - BUG_ON(busiest == this_rq); - - schedstat_add(sd, lb_imbalance[idle], imbalance); - - ld_moved = 0; - if (busiest->nr_running > 1) { - /* - * Attempt to move tasks. If find_busiest_group has found - * an imbalance but busiest->nr_running <= 1, the group is - * still unbalanced. ld_moved simply stays zero, so it is - * correctly treated as an imbalance. - */ - local_irq_save(flags); - double_rq_lock(this_rq, busiest); - ld_moved = move_tasks(this_rq, this_cpu, busiest, - imbalance, sd, idle, &all_pinned); - double_rq_unlock(this_rq, busiest); - local_irq_restore(flags); - - /* - * some other cpu did the load balance for us. - */ - if (ld_moved && this_cpu != smp_processor_id()) - resched_cpu(this_cpu); - - /* All tasks on this runqueue were pinned by CPU affinity */ - if (unlikely(all_pinned)) { - cpumask_clear_cpu(cpu_of(busiest), cpus); - if (!cpumask_empty(cpus)) - goto redo; - goto out_balanced; - } - } - - if (!ld_moved) { - schedstat_inc(sd, lb_failed[idle]); - sd->nr_balance_failed++; - - if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) { - - raw_spin_lock_irqsave(&busiest->lock, flags); - - /* don't kick the migration_thread, if the curr - * task on busiest cpu can't be moved to this_cpu - */ - if (!cpumask_test_cpu(this_cpu, - &busiest->curr->cpus_allowed)) { - raw_spin_unlock_irqrestore(&busiest->lock, - flags); - all_pinned = 1; - goto out_one_pinned; - } - - if (!busiest->active_balance) { - busiest->active_balance = 1; - busiest->push_cpu = this_cpu; - active_balance = 1; - } - raw_spin_unlock_irqrestore(&busiest->lock, flags); - if (active_balance) - wake_up_process(busiest->migration_thread); - - /* - * We've kicked active balancing, reset the failure - * counter. - */ - sd->nr_balance_failed = sd->cache_nice_tries+1; - } - } else - sd->nr_balance_failed = 0; - - if (likely(!active_balance)) { - /* We were unbalanced, so reset the balancing interval */ - sd->balance_interval = sd->min_interval; - } else { - /* - * If we've begun active balancing, start to back off. This - * case may not be covered by the all_pinned logic if there - * is only 1 task on the busy runqueue (because we don't call - * move_tasks). - */ - if (sd->balance_interval < sd->max_interval) - sd->balance_interval *= 2; - } - - if (!ld_moved && !sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - ld_moved = -1; - - goto out; - -out_balanced: - schedstat_inc(sd, lb_balanced[idle]); - - sd->nr_balance_failed = 0; - -out_one_pinned: - /* tune up the balancing interval */ - if ((all_pinned && sd->balance_interval < MAX_PINNED_INTERVAL) || - (sd->balance_interval < sd->max_interval)) - sd->balance_interval *= 2; - - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - ld_moved = -1; - else - ld_moved = 0; -out: - if (ld_moved) - update_shares(sd); - return ld_moved; -} - -/* - * Check this_cpu to ensure it is balanced within domain. Attempt to move - * tasks if there is an imbalance. - * - * Called from schedule when this_rq is about to become idle (CPU_NEWLY_IDLE). - * this_rq is locked. - */ -static int -load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd) -{ - struct sched_group *group; - struct rq *busiest = NULL; - unsigned long imbalance; - int ld_moved = 0; - int sd_idle = 0; - int all_pinned = 0; - struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); - - cpumask_copy(cpus, cpu_active_mask); - - /* - * When power savings policy is enabled for the parent domain, idle - * sibling can pick up load irrespective of busy siblings. In this case, - * let the state of idle sibling percolate up as IDLE, instead of - * portraying it as CPU_NOT_IDLE. - */ - if (sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - sd_idle = 1; - - schedstat_inc(sd, lb_count[CPU_NEWLY_IDLE]); -redo: - update_shares_locked(this_rq, sd); - group = find_busiest_group(sd, this_cpu, &imbalance, CPU_NEWLY_IDLE, - &sd_idle, cpus, NULL); - if (!group) { - schedstat_inc(sd, lb_nobusyg[CPU_NEWLY_IDLE]); - goto out_balanced; - } - - busiest = find_busiest_queue(group, CPU_NEWLY_IDLE, imbalance, cpus); - if (!busiest) { - schedstat_inc(sd, lb_nobusyq[CPU_NEWLY_IDLE]); - goto out_balanced; - } - - BUG_ON(busiest == this_rq); - - schedstat_add(sd, lb_imbalance[CPU_NEWLY_IDLE], imbalance); - - ld_moved = 0; - if (busiest->nr_running > 1) { - /* Attempt to move tasks */ - double_lock_balance(this_rq, busiest); - /* this_rq->clock is already updated */ - update_rq_clock(busiest); - ld_moved = move_tasks(this_rq, this_cpu, busiest, - imbalance, sd, CPU_NEWLY_IDLE, - &all_pinned); - double_unlock_balance(this_rq, busiest); - - if (unlikely(all_pinned)) { - cpumask_clear_cpu(cpu_of(busiest), cpus); - if (!cpumask_empty(cpus)) - goto redo; - } - } - - if (!ld_moved) { - int active_balance = 0; - - schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]); - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - - if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) - return -1; - - if (sd->nr_balance_failed++ < 2) - return -1; - - /* - * The only task running in a non-idle cpu can be moved to this - * cpu in an attempt to completely freeup the other CPU - * package. The same method used to move task in load_balance() - * have been extended for load_balance_newidle() to speedup - * consolidation at sched_mc=POWERSAVINGS_BALANCE_WAKEUP (2) - * - * The package power saving logic comes from - * find_busiest_group(). If there are no imbalance, then - * f_b_g() will return NULL. However when sched_mc={1,2} then - * f_b_g() will select a group from which a running task may be - * pulled to this cpu in order to make the other package idle. - * If there is no opportunity to make a package idle and if - * there are no imbalance, then f_b_g() will return NULL and no - * action will be taken in load_balance_newidle(). - * - * Under normal task pull operation due to imbalance, there - * will be more than one task in the source run queue and - * move_tasks() will succeed. ld_moved will be true and this - * active balance code will not be triggered. - */ - - /* Lock busiest in correct order while this_rq is held */ - double_lock_balance(this_rq, busiest); - - /* - * don't kick the migration_thread, if the curr - * task on busiest cpu can't be moved to this_cpu - */ - if (!cpumask_test_cpu(this_cpu, &busiest->curr->cpus_allowed)) { - double_unlock_balance(this_rq, busiest); - all_pinned = 1; - return ld_moved; - } - - if (!busiest->active_balance) { - busiest->active_balance = 1; - busiest->push_cpu = this_cpu; - active_balance = 1; - } - - double_unlock_balance(this_rq, busiest); - /* - * Should not call ttwu while holding a rq->lock - */ - raw_spin_unlock(&this_rq->lock); - if (active_balance) - wake_up_process(busiest->migration_thread); - raw_spin_lock(&this_rq->lock); - - } else - sd->nr_balance_failed = 0; - - update_shares_locked(this_rq, sd); - return ld_moved; - -out_balanced: - schedstat_inc(sd, lb_balanced[CPU_NEWLY_IDLE]); - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - sd->nr_balance_failed = 0; - - return 0; -} - -/* - * idle_balance is called by schedule() if this_cpu is about to become - * idle. Attempts to pull tasks from other CPUs. - */ -static void idle_balance(int this_cpu, struct rq *this_rq) -{ - struct sched_domain *sd; - int pulled_task = 0; - unsigned long next_balance = jiffies + HZ; - - this_rq->idle_stamp = this_rq->clock; - - if (this_rq->avg_idle < sysctl_sched_migration_cost) - return; - - for_each_domain(this_cpu, sd) { - unsigned long interval; - - if (!(sd->flags & SD_LOAD_BALANCE)) - continue; - - if (sd->flags & SD_BALANCE_NEWIDLE) - /* If we've pulled tasks over stop searching: */ - pulled_task = load_balance_newidle(this_cpu, this_rq, - sd); - - interval = msecs_to_jiffies(sd->balance_interval); - if (time_after(next_balance, sd->last_balance + interval)) - next_balance = sd->last_balance + interval; - if (pulled_task) { - this_rq->idle_stamp = 0; - break; - } - } - if (pulled_task || time_after(jiffies, this_rq->next_balance)) { - /* - * We are going idle. next_balance may be set based on - * a busy processor. So reset next_balance. - */ - this_rq->next_balance = next_balance; - } -} - -/* - * active_load_balance is run by migration threads. It pushes running tasks - * off the busiest CPU onto idle CPUs. It requires at least 1 task to be - * running on each physical CPU where possible, and avoids physical / - * logical imbalances. - * - * Called with busiest_rq locked. - */ -static void active_load_balance(struct rq *busiest_rq, int busiest_cpu) -{ - int target_cpu = busiest_rq->push_cpu; - struct sched_domain *sd; - struct rq *target_rq; - - /* Is there any task to move? */ - if (busiest_rq->nr_running <= 1) - return; - - target_rq = cpu_rq(target_cpu); - - /* - * This condition is "impossible", if it occurs - * we need to fix it. Originally reported by - * Bjorn Helgaas on a 128-cpu setup. - */ - BUG_ON(busiest_rq == target_rq); - - /* move a task from busiest_rq to target_rq */ - double_lock_balance(busiest_rq, target_rq); - update_rq_clock(busiest_rq); - update_rq_clock(target_rq); - - /* Search for an sd spanning us and the target CPU. */ - for_each_domain(target_cpu, sd) { - if ((sd->flags & SD_LOAD_BALANCE) && - cpumask_test_cpu(busiest_cpu, sched_domain_span(sd))) - break; - } - - if (likely(sd)) { - schedstat_inc(sd, alb_count); - - if (move_one_task(target_rq, target_cpu, busiest_rq, - sd, CPU_IDLE)) - schedstat_inc(sd, alb_pushed); - else - schedstat_inc(sd, alb_failed); - } - double_unlock_balance(busiest_rq, target_rq); -} - -#ifdef CONFIG_NO_HZ -static struct { - atomic_t load_balancer; - cpumask_var_t cpu_mask; - cpumask_var_t ilb_grp_nohz_mask; -} nohz ____cacheline_aligned = { - .load_balancer = ATOMIC_INIT(-1), -}; - -int get_nohz_load_balancer(void) -{ - return atomic_read(&nohz.load_balancer); -} - -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) -/** - * lowest_flag_domain - Return lowest sched_domain containing flag. - * @cpu: The cpu whose lowest level of sched domain is to - * be returned. - * @flag: The flag to check for the lowest sched_domain - * for the given cpu. - * - * Returns the lowest sched_domain of a cpu which contains the given flag. - */ -static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) -{ - struct sched_domain *sd; - - for_each_domain(cpu, sd) - if (sd && (sd->flags & flag)) - break; - - return sd; -} - -/** - * for_each_flag_domain - Iterates over sched_domains containing the flag. - * @cpu: The cpu whose domains we're iterating over. - * @sd: variable holding the value of the power_savings_sd - * for cpu. - * @flag: The flag to filter the sched_domains to be iterated. - * - * Iterates over all the scheduler domains for a given cpu that has the 'flag' - * set, starting from the lowest sched_domain to the highest. - */ -#define for_each_flag_domain(cpu, sd, flag) \ - for (sd = lowest_flag_domain(cpu, flag); \ - (sd && (sd->flags & flag)); sd = sd->parent) - -/** - * is_semi_idle_group - Checks if the given sched_group is semi-idle. - * @ilb_group: group to be checked for semi-idleness - * - * Returns: 1 if the group is semi-idle. 0 otherwise. - * - * We define a sched_group to be semi idle if it has atleast one idle-CPU - * and atleast one non-idle CPU. This helper function checks if the given - * sched_group is semi-idle or not. - */ -static inline int is_semi_idle_group(struct sched_group *ilb_group) -{ - cpumask_and(nohz.ilb_grp_nohz_mask, nohz.cpu_mask, - sched_group_cpus(ilb_group)); - - /* - * A sched_group is semi-idle when it has atleast one busy cpu - * and atleast one idle cpu. - */ - if (cpumask_empty(nohz.ilb_grp_nohz_mask)) - return 0; - - if (cpumask_equal(nohz.ilb_grp_nohz_mask, sched_group_cpus(ilb_group))) - return 0; - - return 1; -} -/** - * find_new_ilb - Finds the optimum idle load balancer for nomination. - * @cpu: The cpu which is nominating a new idle_load_balancer. - * - * Returns: Returns the id of the idle load balancer if it exists, - * Else, returns >= nr_cpu_ids. - * - * This algorithm picks the idle load balancer such that it belongs to a - * semi-idle powersavings sched_domain. The idea is to try and avoid - * completely idle packages/cores just for the purpose of idle load balancing - * when there are other idle cpu's which are better suited for that job. - */ -static int find_new_ilb(int cpu) -{ - struct sched_domain *sd; - struct sched_group *ilb_group; - - /* - * Have idle load balancer selection from semi-idle packages only - * when power-aware load balancing is enabled - */ - if (!(sched_smt_power_savings || sched_mc_power_savings)) - goto out_done; - - /* - * Optimize for the case when we have no idle CPUs or only one - * idle CPU. Don't walk the sched_domain hierarchy in such cases - */ - if (cpumask_weight(nohz.cpu_mask) < 2) - goto out_done; - - for_each_flag_domain(cpu, sd, SD_POWERSAVINGS_BALANCE) { - ilb_group = sd->groups; - - do { - if (is_semi_idle_group(ilb_group)) - return cpumask_first(nohz.ilb_grp_nohz_mask); - - ilb_group = ilb_group->next; - - } while (ilb_group != sd->groups); - } - -out_done: - return cpumask_first(nohz.cpu_mask); -} -#else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ -static inline int find_new_ilb(int call_cpu) -{ - return cpumask_first(nohz.cpu_mask); -} -#endif - -/* - * This routine will try to nominate the ilb (idle load balancing) - * owner among the cpus whose ticks are stopped. ilb owner will do the idle - * load balancing on behalf of all those cpus. If all the cpus in the system - * go into this tickless mode, then there will be no ilb owner (as there is - * no need for one) and all the cpus will sleep till the next wakeup event - * arrives... - * - * For the ilb owner, tick is not stopped. And this tick will be used - * for idle load balancing. ilb owner will still be part of - * nohz.cpu_mask.. - * - * While stopping the tick, this cpu will become the ilb owner if there - * is no other owner. And will be the owner till that cpu becomes busy - * or if all cpus in the system stop their ticks at which point - * there is no need for ilb owner. - * - * When the ilb owner becomes busy, it nominates another owner, during the - * next busy scheduler_tick() - */ -int select_nohz_load_balancer(int stop_tick) -{ - int cpu = smp_processor_id(); - - if (stop_tick) { - cpu_rq(cpu)->in_nohz_recently = 1; - - if (!cpu_active(cpu)) { - if (atomic_read(&nohz.load_balancer) != cpu) - return 0; - - /* - * If we are going offline and still the leader, - * give up! - */ - if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) - BUG(); - - return 0; - } - - cpumask_set_cpu(cpu, nohz.cpu_mask); - - /* time for ilb owner also to sleep */ - if (cpumask_weight(nohz.cpu_mask) == num_active_cpus()) { - if (atomic_read(&nohz.load_balancer) == cpu) - atomic_set(&nohz.load_balancer, -1); - return 0; - } - - if (atomic_read(&nohz.load_balancer) == -1) { - /* make me the ilb owner */ - if (atomic_cmpxchg(&nohz.load_balancer, -1, cpu) == -1) - return 1; - } else if (atomic_read(&nohz.load_balancer) == cpu) { - int new_ilb; - - if (!(sched_smt_power_savings || - sched_mc_power_savings)) - return 1; - /* - * Check to see if there is a more power-efficient - * ilb. - */ - new_ilb = find_new_ilb(cpu); - if (new_ilb < nr_cpu_ids && new_ilb != cpu) { - atomic_set(&nohz.load_balancer, -1); - resched_cpu(new_ilb); - return 0; - } - return 1; - } - } else { - if (!cpumask_test_cpu(cpu, nohz.cpu_mask)) - return 0; - - cpumask_clear_cpu(cpu, nohz.cpu_mask); - - if (atomic_read(&nohz.load_balancer) == cpu) - if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) - BUG(); - } - return 0; -} -#endif - -static DEFINE_SPINLOCK(balancing); - -/* - * It checks each scheduling domain to see if it is due to be balanced, - * and initiates a balancing operation if so. - * - * Balancing parameters are set up in arch_init_sched_domains. - */ -static void rebalance_domains(int cpu, enum cpu_idle_type idle) -{ - int balance = 1; - struct rq *rq = cpu_rq(cpu); - unsigned long interval; - struct sched_domain *sd; - /* Earliest time when we have to do rebalance again */ - unsigned long next_balance = jiffies + 60*HZ; - int update_next_balance = 0; - int need_serialize; - - for_each_domain(cpu, sd) { - if (!(sd->flags & SD_LOAD_BALANCE)) - continue; - - interval = sd->balance_interval; - if (idle != CPU_IDLE) - interval *= sd->busy_factor; - - /* scale ms to jiffies */ - interval = msecs_to_jiffies(interval); - if (unlikely(!interval)) - interval = 1; - if (interval > HZ*NR_CPUS/10) - interval = HZ*NR_CPUS/10; - - need_serialize = sd->flags & SD_SERIALIZE; - - if (need_serialize) { - if (!spin_trylock(&balancing)) - goto out; - } - - if (time_after_eq(jiffies, sd->last_balance + interval)) { - if (load_balance(cpu, rq, sd, idle, &balance)) { - /* - * We've pulled tasks over so either we're no - * longer idle, or one of our SMT siblings is - * not idle. - */ - idle = CPU_NOT_IDLE; - } - sd->last_balance = jiffies; - } - if (need_serialize) - spin_unlock(&balancing); -out: - if (time_after(next_balance, sd->last_balance + interval)) { - next_balance = sd->last_balance + interval; - update_next_balance = 1; - } - - /* - * Stop the load balance at this level. There is another - * CPU in our sched group which is doing load balancing more - * actively. - */ - if (!balance) - break; - } - - /* - * next_balance will be updated only when there is a need. - * When the cpu is attached to null domain for ex, it will not be - * updated. - */ - if (likely(update_next_balance)) - rq->next_balance = next_balance; -} - -/* - * run_rebalance_domains is triggered when needed from the scheduler tick. - * In CONFIG_NO_HZ case, the idle load balance owner will do the - * rebalancing for all the cpus for whom scheduler ticks are stopped. - */ -static void run_rebalance_domains(struct softirq_action *h) -{ - int this_cpu = smp_processor_id(); - struct rq *this_rq = cpu_rq(this_cpu); - enum cpu_idle_type idle = this_rq->idle_at_tick ? - CPU_IDLE : CPU_NOT_IDLE; - - rebalance_domains(this_cpu, idle); - -#ifdef CONFIG_NO_HZ - /* - * If this cpu is the owner for idle load balancing, then do the - * balancing on behalf of the other idle cpus whose ticks are - * stopped. - */ - if (this_rq->idle_at_tick && - atomic_read(&nohz.load_balancer) == this_cpu) { - struct rq *rq; - int balance_cpu; - - for_each_cpu(balance_cpu, nohz.cpu_mask) { - if (balance_cpu == this_cpu) - continue; - - /* - * If this cpu gets work to do, stop the load balancing - * work being done for other cpus. Next load - * balancing owner will pick it up. - */ - if (need_resched()) - break; - - rebalance_domains(balance_cpu, CPU_IDLE); - - rq = cpu_rq(balance_cpu); - if (time_after(this_rq->next_balance, rq->next_balance)) - this_rq->next_balance = rq->next_balance; - } - } -#endif -} - -static inline int on_null_domain(int cpu) -{ - return !rcu_dereference(cpu_rq(cpu)->sd); -} - -/* - * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing. - * - * In case of CONFIG_NO_HZ, this is the place where we nominate a new - * idle load balancing owner or decide to stop the periodic load balancing, - * if the whole system is idle. - */ -static inline void trigger_load_balance(struct rq *rq, int cpu) -{ -#ifdef CONFIG_NO_HZ - /* - * If we were in the nohz mode recently and busy at the current - * scheduler tick, then check if we need to nominate new idle - * load balancer. - */ - if (rq->in_nohz_recently && !rq->idle_at_tick) { - rq->in_nohz_recently = 0; - - if (atomic_read(&nohz.load_balancer) == cpu) { - cpumask_clear_cpu(cpu, nohz.cpu_mask); - atomic_set(&nohz.load_balancer, -1); - } - - if (atomic_read(&nohz.load_balancer) == -1) { - int ilb = find_new_ilb(cpu); - - if (ilb < nr_cpu_ids) - resched_cpu(ilb); - } - } - - /* - * If this cpu is idle and doing idle load balancing for all the - * cpus with ticks stopped, is it time for that to stop? - */ - if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) == cpu && - cpumask_weight(nohz.cpu_mask) == num_online_cpus()) { - resched_cpu(cpu); - return; - } - - /* - * If this cpu is idle and the idle load balancing is done by - * someone else, then no need raise the SCHED_SOFTIRQ - */ - if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) != cpu && - cpumask_test_cpu(cpu, nohz.cpu_mask)) - return; -#endif - /* Don't need to rebalance while attached to NULL domain */ - if (time_after_eq(jiffies, rq->next_balance) && - likely(!on_null_domain(cpu))) - raise_softirq(SCHED_SOFTIRQ); -} - -#else /* CONFIG_SMP */ - -/* - * on UP we do not need to balance between CPUs: - */ -static inline void idle_balance(int cpu, struct rq *rq) -{ -} - #endif DEFINE_PER_CPU(struct kernel_stat, kstat); diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 7177860..5116b81 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1952,6 +1952,1762 @@ move_one_task_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, return 0; } +/* + * pull_task - move a task from a remote runqueue to the local runqueue. + * Both runqueues must be locked. + */ +static void pull_task(struct rq *src_rq, struct task_struct *p, + struct rq *this_rq, int this_cpu) +{ + deactivate_task(src_rq, p, 0); + set_task_cpu(p, this_cpu); + activate_task(this_rq, p, 0); + check_preempt_curr(this_rq, p, 0); +} + +/* + * can_migrate_task - may task p from runqueue rq be migrated to this_cpu? + */ +static +int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, + struct sched_domain *sd, enum cpu_idle_type idle, + int *all_pinned) +{ + int tsk_cache_hot = 0; + /* + * We do not migrate tasks that are: + * 1) running (obviously), or + * 2) cannot be migrated to this CPU due to cpus_allowed, or + * 3) are cache-hot on their current CPU. + */ + if (!cpumask_test_cpu(this_cpu, &p->cpus_allowed)) { + schedstat_inc(p, se.nr_failed_migrations_affine); + return 0; + } + *all_pinned = 0; + + if (task_running(rq, p)) { + schedstat_inc(p, se.nr_failed_migrations_running); + return 0; + } + + /* + * Aggressive migration if: + * 1) task is cache cold, or + * 2) too many balance attempts have failed. + */ + + tsk_cache_hot = task_hot(p, rq->clock, sd); + if (!tsk_cache_hot || + sd->nr_balance_failed > sd->cache_nice_tries) { +#ifdef CONFIG_SCHEDSTATS + if (tsk_cache_hot) { + schedstat_inc(sd, lb_hot_gained[idle]); + schedstat_inc(p, se.nr_forced_migrations); + } +#endif + return 1; + } + + if (tsk_cache_hot) { + schedstat_inc(p, se.nr_failed_migrations_hot); + return 0; + } + return 1; +} + +static unsigned long +balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, struct sched_domain *sd, + enum cpu_idle_type idle, int *all_pinned, + int *this_best_prio, struct rq_iterator *iterator) +{ + int loops = 0, pulled = 0, pinned = 0; + struct task_struct *p; + long rem_load_move = max_load_move; + + if (max_load_move == 0) + goto out; + + pinned = 1; + + /* + * Start the load-balancing iterator: + */ + p = iterator->start(iterator->arg); +next: + if (!p || loops++ > sysctl_sched_nr_migrate) + goto out; + + if ((p->se.load.weight >> 1) > rem_load_move || + !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { + p = iterator->next(iterator->arg); + goto next; + } + + pull_task(busiest, p, this_rq, this_cpu); + pulled++; + rem_load_move -= p->se.load.weight; + +#ifdef CONFIG_PREEMPT + /* + * NEWIDLE balancing is a source of latency, so preemptible kernels + * will stop after the first task is pulled to minimize the critical + * section. + */ + if (idle == CPU_NEWLY_IDLE) + goto out; +#endif + + /* + * We only want to steal up to the prescribed amount of weighted load. + */ + if (rem_load_move > 0) { + if (p->prio < *this_best_prio) + *this_best_prio = p->prio; + p = iterator->next(iterator->arg); + goto next; + } +out: + /* + * Right now, this is one of only two places pull_task() is called, + * so we can safely collect pull_task() stats here rather than + * inside pull_task(). + */ + schedstat_add(sd, lb_gained[idle], pulled); + + if (all_pinned) + *all_pinned = pinned; + + return max_load_move - rem_load_move; +} + +/* + * move_tasks tries to move up to max_load_move weighted load from busiest to + * this_rq, as part of a balancing operation within domain "sd". + * Returns 1 if successful and 0 otherwise. + * + * Called with both runqueues locked. + */ +static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, + struct sched_domain *sd, enum cpu_idle_type idle, + int *all_pinned) +{ + const struct sched_class *class = sched_class_highest; + unsigned long total_load_moved = 0; + int this_best_prio = this_rq->curr->prio; + + do { + total_load_moved += + class->load_balance(this_rq, this_cpu, busiest, + max_load_move - total_load_moved, + sd, idle, all_pinned, &this_best_prio); + class = class->next; + +#ifdef CONFIG_PREEMPT + /* + * NEWIDLE balancing is a source of latency, so preemptible + * kernels will stop after the first task is pulled to minimize + * the critical section. + */ + if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) + break; +#endif + } while (class && max_load_move > total_load_moved); + + return total_load_moved > 0; +} + +static int +iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle, + struct rq_iterator *iterator) +{ + struct task_struct *p = iterator->start(iterator->arg); + int pinned = 0; + + while (p) { + if (can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { + pull_task(busiest, p, this_rq, this_cpu); + /* + * Right now, this is only the second place pull_task() + * is called, so we can safely collect pull_task() + * stats here rather than inside pull_task(). + */ + schedstat_inc(sd, lb_gained[idle]); + + return 1; + } + p = iterator->next(iterator->arg); + } + + return 0; +} + +/* + * move_one_task tries to move exactly one task from busiest to this_rq, as + * part of active balancing operations within "domain". + * Returns 1 if successful and 0 otherwise. + * + * Called with both runqueues locked. + */ +static int move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle) +{ + const struct sched_class *class; + + for_each_class(class) { + if (class->move_one_task(this_rq, this_cpu, busiest, sd, idle)) + return 1; + } + + return 0; +} +/********** Helpers for find_busiest_group ************************/ +/* + * sd_lb_stats - Structure to store the statistics of a sched_domain + * during load balancing. + */ +struct sd_lb_stats { + struct sched_group *busiest; /* Busiest group in this sd */ + struct sched_group *this; /* Local group in this sd */ + unsigned long total_load; /* Total load of all groups in sd */ + unsigned long total_pwr; /* Total power of all groups in sd */ + unsigned long avg_load; /* Average load across all groups in sd */ + + /** Statistics of this group */ + unsigned long this_load; + unsigned long this_load_per_task; + unsigned long this_nr_running; + + /* Statistics of the busiest group */ + unsigned long max_load; + unsigned long busiest_load_per_task; + unsigned long busiest_nr_running; + + int group_imb; /* Is there imbalance in this sd */ +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) + int power_savings_balance; /* Is powersave balance needed for this sd */ + struct sched_group *group_min; /* Least loaded group in sd */ + struct sched_group *group_leader; /* Group which relieves group_min */ + unsigned long min_load_per_task; /* load_per_task in group_min */ + unsigned long leader_nr_running; /* Nr running of group_leader */ + unsigned long min_nr_running; /* Nr running of group_min */ +#endif +}; + +/* + * sg_lb_stats - stats of a sched_group required for load_balancing + */ +struct sg_lb_stats { + unsigned long avg_load; /*Avg load across the CPUs of the group */ + unsigned long group_load; /* Total load over the CPUs of the group */ + unsigned long sum_nr_running; /* Nr tasks running in the group */ + unsigned long sum_weighted_load; /* Weighted load of group's tasks */ + unsigned long group_capacity; + int group_imb; /* Is there an imbalance in the group ? */ +}; + +/** + * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. + * @group: The group whose first cpu is to be returned. + */ +static inline unsigned int group_first_cpu(struct sched_group *group) +{ + return cpumask_first(sched_group_cpus(group)); +} + +/** + * get_sd_load_idx - Obtain the load index for a given sched domain. + * @sd: The sched_domain whose load_idx is to be obtained. + * @idle: The Idle status of the CPU for whose sd load_icx is obtained. + */ +static inline int get_sd_load_idx(struct sched_domain *sd, + enum cpu_idle_type idle) +{ + int load_idx; + + switch (idle) { + case CPU_NOT_IDLE: + load_idx = sd->busy_idx; + break; + + case CPU_NEWLY_IDLE: + load_idx = sd->newidle_idx; + break; + default: + load_idx = sd->idle_idx; + break; + } + + return load_idx; +} + + +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) +/** + * init_sd_power_savings_stats - Initialize power savings statistics for + * the given sched_domain, during load balancing. + * + * @sd: Sched domain whose power-savings statistics are to be initialized. + * @sds: Variable containing the statistics for sd. + * @idle: Idle status of the CPU at which we're performing load-balancing. + */ +static inline void init_sd_power_savings_stats(struct sched_domain *sd, + struct sd_lb_stats *sds, enum cpu_idle_type idle) +{ + /* + * Busy processors will not participate in power savings + * balance. + */ + if (idle == CPU_NOT_IDLE || !(sd->flags & SD_POWERSAVINGS_BALANCE)) + sds->power_savings_balance = 0; + else { + sds->power_savings_balance = 1; + sds->min_nr_running = ULONG_MAX; + sds->leader_nr_running = 0; + } +} + +/** + * update_sd_power_savings_stats - Update the power saving stats for a + * sched_domain while performing load balancing. + * + * @group: sched_group belonging to the sched_domain under consideration. + * @sds: Variable containing the statistics of the sched_domain + * @local_group: Does group contain the CPU for which we're performing + * load balancing ? + * @sgs: Variable containing the statistics of the group. + */ +static inline void update_sd_power_savings_stats(struct sched_group *group, + struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) +{ + + if (!sds->power_savings_balance) + return; + + /* + * If the local group is idle or completely loaded + * no need to do power savings balance at this domain + */ + if (local_group && (sds->this_nr_running >= sgs->group_capacity || + !sds->this_nr_running)) + sds->power_savings_balance = 0; + + /* + * If a group is already running at full capacity or idle, + * don't include that group in power savings calculations + */ + if (!sds->power_savings_balance || + sgs->sum_nr_running >= sgs->group_capacity || + !sgs->sum_nr_running) + return; + + /* + * Calculate the group which has the least non-idle load. + * This is the group from where we need to pick up the load + * for saving power + */ + if ((sgs->sum_nr_running < sds->min_nr_running) || + (sgs->sum_nr_running == sds->min_nr_running && + group_first_cpu(group) > group_first_cpu(sds->group_min))) { + sds->group_min = group; + sds->min_nr_running = sgs->sum_nr_running; + sds->min_load_per_task = sgs->sum_weighted_load / + sgs->sum_nr_running; + } + + /* + * Calculate the group which is almost near its + * capacity but still has some space to pick up some load + * from other group and save more power + */ + if (sgs->sum_nr_running + 1 > sgs->group_capacity) + return; + + if (sgs->sum_nr_running > sds->leader_nr_running || + (sgs->sum_nr_running == sds->leader_nr_running && + group_first_cpu(group) < group_first_cpu(sds->group_leader))) { + sds->group_leader = group; + sds->leader_nr_running = sgs->sum_nr_running; + } +} + +/** + * check_power_save_busiest_group - see if there is potential for some power-savings balance + * @sds: Variable containing the statistics of the sched_domain + * under consideration. + * @this_cpu: Cpu at which we're currently performing load-balancing. + * @imbalance: Variable to store the imbalance. + * + * Description: + * Check if we have potential to perform some power-savings balance. + * If yes, set the busiest group to be the least loaded group in the + * sched_domain, so that it's CPUs can be put to idle. + * + * Returns 1 if there is potential to perform power-savings balance. + * Else returns 0. + */ +static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + if (!sds->power_savings_balance) + return 0; + + if (sds->this != sds->group_leader || + sds->group_leader == sds->group_min) + return 0; + + *imbalance = sds->min_load_per_task; + sds->busiest = sds->group_min; + + return 1; + +} +#else /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ +static inline void init_sd_power_savings_stats(struct sched_domain *sd, + struct sd_lb_stats *sds, enum cpu_idle_type idle) +{ + return; +} + +static inline void update_sd_power_savings_stats(struct sched_group *group, + struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) +{ + return; +} + +static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + return 0; +} +#endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ + + +unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu) +{ + return SCHED_LOAD_SCALE; +} + +unsigned long __weak arch_scale_freq_power(struct sched_domain *sd, int cpu) +{ + return default_scale_freq_power(sd, cpu); +} + +unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu) +{ + unsigned long weight = cpumask_weight(sched_domain_span(sd)); + unsigned long smt_gain = sd->smt_gain; + + smt_gain /= weight; + + return smt_gain; +} + +unsigned long __weak arch_scale_smt_power(struct sched_domain *sd, int cpu) +{ + return default_scale_smt_power(sd, cpu); +} + +unsigned long scale_rt_power(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + u64 total, available; + + sched_avg_update(rq); + + total = sched_avg_period() + (rq->clock - rq->age_stamp); + available = total - rq->rt_avg; + + if (unlikely((s64)total < SCHED_LOAD_SCALE)) + total = SCHED_LOAD_SCALE; + + total >>= SCHED_LOAD_SHIFT; + + return div_u64(available, total); +} + +static void update_cpu_power(struct sched_domain *sd, int cpu) +{ + unsigned long weight = cpumask_weight(sched_domain_span(sd)); + unsigned long power = SCHED_LOAD_SCALE; + struct sched_group *sdg = sd->groups; + + if (sched_feat(ARCH_POWER)) + power *= arch_scale_freq_power(sd, cpu); + else + power *= default_scale_freq_power(sd, cpu); + + power >>= SCHED_LOAD_SHIFT; + + if ((sd->flags & SD_SHARE_CPUPOWER) && weight > 1) { + if (sched_feat(ARCH_POWER)) + power *= arch_scale_smt_power(sd, cpu); + else + power *= default_scale_smt_power(sd, cpu); + + power >>= SCHED_LOAD_SHIFT; + } + + power *= scale_rt_power(cpu); + power >>= SCHED_LOAD_SHIFT; + + if (!power) + power = 1; + + sdg->cpu_power = power; +} + +static void update_group_power(struct sched_domain *sd, int cpu) +{ + struct sched_domain *child = sd->child; + struct sched_group *group, *sdg = sd->groups; + unsigned long power; + + if (!child) { + update_cpu_power(sd, cpu); + return; + } + + power = 0; + + group = child->groups; + do { + power += group->cpu_power; + group = group->next; + } while (group != child->groups); + + sdg->cpu_power = power; +} + +/** + * update_sg_lb_stats - Update sched_group's statistics for load balancing. + * @sd: The sched_domain whose statistics are to be updated. + * @group: sched_group whose statistics are to be updated. + * @this_cpu: Cpu for which load balance is currently performed. + * @idle: Idle status of this_cpu + * @load_idx: Load index of sched_domain of this_cpu for load calc. + * @sd_idle: Idle status of the sched_domain containing group. + * @local_group: Does group contain this_cpu. + * @cpus: Set of cpus considered for load balancing. + * @balance: Should we balance. + * @sgs: variable to hold the statistics for this group. + */ +static inline void update_sg_lb_stats(struct sched_domain *sd, + struct sched_group *group, int this_cpu, + enum cpu_idle_type idle, int load_idx, int *sd_idle, + int local_group, const struct cpumask *cpus, + int *balance, struct sg_lb_stats *sgs) +{ + unsigned long load, max_cpu_load, min_cpu_load; + int i; + unsigned int balance_cpu = -1, first_idle_cpu = 0; + unsigned long sum_avg_load_per_task; + unsigned long avg_load_per_task; + + if (local_group) { + balance_cpu = group_first_cpu(group); + if (balance_cpu == this_cpu) + update_group_power(sd, this_cpu); + } + + /* Tally up the load of all CPUs in the group */ + sum_avg_load_per_task = avg_load_per_task = 0; + max_cpu_load = 0; + min_cpu_load = ~0UL; + + for_each_cpu_and(i, sched_group_cpus(group), cpus) { + struct rq *rq = cpu_rq(i); + + if (*sd_idle && rq->nr_running) + *sd_idle = 0; + + /* Bias balancing toward cpus of our domain */ + if (local_group) { + if (idle_cpu(i) && !first_idle_cpu) { + first_idle_cpu = 1; + balance_cpu = i; + } + + load = target_load(i, load_idx); + } else { + load = source_load(i, load_idx); + if (load > max_cpu_load) + max_cpu_load = load; + if (min_cpu_load > load) + min_cpu_load = load; + } + + sgs->group_load += load; + sgs->sum_nr_running += rq->nr_running; + sgs->sum_weighted_load += weighted_cpuload(i); + + sum_avg_load_per_task += cpu_avg_load_per_task(i); + } + + /* + * First idle cpu or the first cpu(busiest) in this sched group + * is eligible for doing load balancing at this and above + * domains. In the newly idle case, we will allow all the cpu's + * to do the newly idle load balance. + */ + if (idle != CPU_NEWLY_IDLE && local_group && + balance_cpu != this_cpu && balance) { + *balance = 0; + return; + } + + /* Adjust by relative CPU power of the group */ + sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; + + + /* + * Consider the group unbalanced when the imbalance is larger + * than the average weight of two tasks. + * + * APZ: with cgroup the avg task weight can vary wildly and + * might not be a suitable number - should we keep a + * normalized nr_running number somewhere that negates + * the hierarchy? + */ + avg_load_per_task = (sum_avg_load_per_task * SCHED_LOAD_SCALE) / + group->cpu_power; + + if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) + sgs->group_imb = 1; + + sgs->group_capacity = + DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); +} + +/** + * update_sd_lb_stats - Update sched_group's statistics for load balancing. + * @sd: sched_domain whose statistics are to be updated. + * @this_cpu: Cpu for which load balance is currently performed. + * @idle: Idle status of this_cpu + * @sd_idle: Idle status of the sched_domain containing group. + * @cpus: Set of cpus considered for load balancing. + * @balance: Should we balance. + * @sds: variable to hold the statistics for this sched_domain. + */ +static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, + enum cpu_idle_type idle, int *sd_idle, + const struct cpumask *cpus, int *balance, + struct sd_lb_stats *sds) +{ + struct sched_domain *child = sd->child; + struct sched_group *group = sd->groups; + struct sg_lb_stats sgs; + int load_idx, prefer_sibling = 0; + + if (child && child->flags & SD_PREFER_SIBLING) + prefer_sibling = 1; + + init_sd_power_savings_stats(sd, sds, idle); + load_idx = get_sd_load_idx(sd, idle); + + do { + int local_group; + + local_group = cpumask_test_cpu(this_cpu, + sched_group_cpus(group)); + memset(&sgs, 0, sizeof(sgs)); + update_sg_lb_stats(sd, group, this_cpu, idle, load_idx, sd_idle, + local_group, cpus, balance, &sgs); + + if (local_group && balance && !(*balance)) + return; + + sds->total_load += sgs.group_load; + sds->total_pwr += group->cpu_power; + + /* + * In case the child domain prefers tasks go to siblings + * first, lower the group capacity to one so that we'll try + * and move all the excess tasks away. + */ + if (prefer_sibling) + sgs.group_capacity = min(sgs.group_capacity, 1UL); + + if (local_group) { + sds->this_load = sgs.avg_load; + sds->this = group; + sds->this_nr_running = sgs.sum_nr_running; + sds->this_load_per_task = sgs.sum_weighted_load; + } else if (sgs.avg_load > sds->max_load && + (sgs.sum_nr_running > sgs.group_capacity || + sgs.group_imb)) { + sds->max_load = sgs.avg_load; + sds->busiest = group; + sds->busiest_nr_running = sgs.sum_nr_running; + sds->busiest_load_per_task = sgs.sum_weighted_load; + sds->group_imb = sgs.group_imb; + } + + update_sd_power_savings_stats(group, sds, local_group, &sgs); + group = group->next; + } while (group != sd->groups); +} + +/** + * fix_small_imbalance - Calculate the minor imbalance that exists + * amongst the groups of a sched_domain, during + * load balancing. + * @sds: Statistics of the sched_domain whose imbalance is to be calculated. + * @this_cpu: The cpu at whose sched_domain we're performing load-balance. + * @imbalance: Variable to store the imbalance. + */ +static inline void fix_small_imbalance(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + unsigned long tmp, pwr_now = 0, pwr_move = 0; + unsigned int imbn = 2; + + if (sds->this_nr_running) { + sds->this_load_per_task /= sds->this_nr_running; + if (sds->busiest_load_per_task > + sds->this_load_per_task) + imbn = 1; + } else + sds->this_load_per_task = + cpu_avg_load_per_task(this_cpu); + + if (sds->max_load - sds->this_load + sds->busiest_load_per_task >= + sds->busiest_load_per_task * imbn) { + *imbalance = sds->busiest_load_per_task; + return; + } + + /* + * OK, we don't have enough imbalance to justify moving tasks, + * however we may be able to increase total CPU power used by + * moving them. + */ + + pwr_now += sds->busiest->cpu_power * + min(sds->busiest_load_per_task, sds->max_load); + pwr_now += sds->this->cpu_power * + min(sds->this_load_per_task, sds->this_load); + pwr_now /= SCHED_LOAD_SCALE; + + /* Amount of load we'd subtract */ + tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / + sds->busiest->cpu_power; + if (sds->max_load > tmp) + pwr_move += sds->busiest->cpu_power * + min(sds->busiest_load_per_task, sds->max_load - tmp); + + /* Amount of load we'd add */ + if (sds->max_load * sds->busiest->cpu_power < + sds->busiest_load_per_task * SCHED_LOAD_SCALE) + tmp = (sds->max_load * sds->busiest->cpu_power) / + sds->this->cpu_power; + else + tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / + sds->this->cpu_power; + pwr_move += sds->this->cpu_power * + min(sds->this_load_per_task, sds->this_load + tmp); + pwr_move /= SCHED_LOAD_SCALE; + + /* Move if we gain throughput */ + if (pwr_move > pwr_now) + *imbalance = sds->busiest_load_per_task; +} + +/** + * calculate_imbalance - Calculate the amount of imbalance present within the + * groups of a given sched_domain during load balance. + * @sds: statistics of the sched_domain whose imbalance is to be calculated. + * @this_cpu: Cpu for which currently load balance is being performed. + * @imbalance: The variable to store the imbalance. + */ +static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, + unsigned long *imbalance) +{ + unsigned long max_pull; + /* + * In the presence of smp nice balancing, certain scenarios can have + * max load less than avg load(as we skip the groups at or below + * its cpu_power, while calculating max_load..) + */ + if (sds->max_load < sds->avg_load) { + *imbalance = 0; + return fix_small_imbalance(sds, this_cpu, imbalance); + } + + /* Don't want to pull so many tasks that a group would go idle */ + max_pull = min(sds->max_load - sds->avg_load, + sds->max_load - sds->busiest_load_per_task); + + /* How much load to actually move to equalise the imbalance */ + *imbalance = min(max_pull * sds->busiest->cpu_power, + (sds->avg_load - sds->this_load) * sds->this->cpu_power) + / SCHED_LOAD_SCALE; + + /* + * if *imbalance is less than the average load per runnable task + * there is no gaurantee that any tasks will be moved so we'll have + * a think about bumping its value to force at least one task to be + * moved + */ + if (*imbalance < sds->busiest_load_per_task) + return fix_small_imbalance(sds, this_cpu, imbalance); + +} +/******* find_busiest_group() helpers end here *********************/ + +/** + * find_busiest_group - Returns the busiest group within the sched_domain + * if there is an imbalance. If there isn't an imbalance, and + * the user has opted for power-savings, it returns a group whose + * CPUs can be put to idle by rebalancing those tasks elsewhere, if + * such a group exists. + * + * Also calculates the amount of weighted load which should be moved + * to restore balance. + * + * @sd: The sched_domain whose busiest group is to be returned. + * @this_cpu: The cpu for which load balancing is currently being performed. + * @imbalance: Variable which stores amount of weighted load which should + * be moved to restore balance/put a group to idle. + * @idle: The idle status of this_cpu. + * @sd_idle: The idleness of sd + * @cpus: The set of CPUs under consideration for load-balancing. + * @balance: Pointer to a variable indicating if this_cpu + * is the appropriate cpu to perform load balancing at this_level. + * + * Returns: - the busiest group if imbalance exists. + * - If no imbalance and user has opted for power-savings balance, + * return the least loaded group whose CPUs can be + * put to idle by rebalancing its tasks onto our group. + */ +static struct sched_group * +find_busiest_group(struct sched_domain *sd, int this_cpu, + unsigned long *imbalance, enum cpu_idle_type idle, + int *sd_idle, const struct cpumask *cpus, int *balance) +{ + struct sd_lb_stats sds; + + memset(&sds, 0, sizeof(sds)); + + /* + * Compute the various statistics relavent for load balancing at + * this level. + */ + update_sd_lb_stats(sd, this_cpu, idle, sd_idle, cpus, + balance, &sds); + + /* Cases where imbalance does not exist from POV of this_cpu */ + /* 1) this_cpu is not the appropriate cpu to perform load balancing + * at this level. + * 2) There is no busy sibling group to pull from. + * 3) This group is the busiest group. + * 4) This group is more busy than the avg busieness at this + * sched_domain. + * 5) The imbalance is within the specified limit. + * 6) Any rebalance would lead to ping-pong + */ + if (balance && !(*balance)) + goto ret; + + if (!sds.busiest || sds.busiest_nr_running == 0) + goto out_balanced; + + if (sds.this_load >= sds.max_load) + goto out_balanced; + + sds.avg_load = (SCHED_LOAD_SCALE * sds.total_load) / sds.total_pwr; + + if (sds.this_load >= sds.avg_load) + goto out_balanced; + + if (100 * sds.max_load <= sd->imbalance_pct * sds.this_load) + goto out_balanced; + + sds.busiest_load_per_task /= sds.busiest_nr_running; + if (sds.group_imb) + sds.busiest_load_per_task = + min(sds.busiest_load_per_task, sds.avg_load); + + /* + * We're trying to get all the cpus to the average_load, so we don't + * want to push ourselves above the average load, nor do we wish to + * reduce the max loaded cpu below the average load, as either of these + * actions would just result in more rebalancing later, and ping-pong + * tasks around. Thus we look for the minimum possible imbalance. + * Negative imbalances (*we* are more loaded than anyone else) will + * be counted as no imbalance for these purposes -- we can't fix that + * by pulling tasks to us. Be careful of negative numbers as they'll + * appear as very large values with unsigned longs. + */ + if (sds.max_load <= sds.busiest_load_per_task) + goto out_balanced; + + /* Looks like there is an imbalance. Compute it */ + calculate_imbalance(&sds, this_cpu, imbalance); + return sds.busiest; + +out_balanced: + /* + * There is no obvious imbalance. But check if we can do some balancing + * to save power. + */ + if (check_power_save_busiest_group(&sds, this_cpu, imbalance)) + return sds.busiest; +ret: + *imbalance = 0; + return NULL; +} + +/* + * find_busiest_queue - find the busiest runqueue among the cpus in group. + */ +static struct rq * +find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, + unsigned long imbalance, const struct cpumask *cpus) +{ + struct rq *busiest = NULL, *rq; + unsigned long max_load = 0; + int i; + + for_each_cpu(i, sched_group_cpus(group)) { + unsigned long power = power_of(i); + unsigned long capacity = DIV_ROUND_CLOSEST(power, SCHED_LOAD_SCALE); + unsigned long wl; + + if (!cpumask_test_cpu(i, cpus)) + continue; + + rq = cpu_rq(i); + wl = weighted_cpuload(i) * SCHED_LOAD_SCALE; + wl /= power; + + if (capacity && rq->nr_running == 1 && wl > imbalance) + continue; + + if (wl > max_load) { + max_load = wl; + busiest = rq; + } + } + + return busiest; +} + +/* + * Max backoff if we encounter pinned tasks. Pretty arbitrary value, but + * so long as it is large enough. + */ +#define MAX_PINNED_INTERVAL 512 + +/* Working cpumask for load_balance and load_balance_newidle. */ +static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); + +/* + * Check this_cpu to ensure it is balanced within domain. Attempt to move + * tasks if there is an imbalance. + */ +static int load_balance(int this_cpu, struct rq *this_rq, + struct sched_domain *sd, enum cpu_idle_type idle, + int *balance) +{ + int ld_moved, all_pinned = 0, active_balance = 0, sd_idle = 0; + struct sched_group *group; + unsigned long imbalance; + struct rq *busiest; + unsigned long flags; + struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); + + cpumask_copy(cpus, cpu_active_mask); + + /* + * When power savings policy is enabled for the parent domain, idle + * sibling can pick up load irrespective of busy siblings. In this case, + * let the state of idle sibling percolate up as CPU_IDLE, instead of + * portraying it as CPU_NOT_IDLE. + */ + if (idle != CPU_NOT_IDLE && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + sd_idle = 1; + + schedstat_inc(sd, lb_count[idle]); + +redo: + update_shares(sd); + group = find_busiest_group(sd, this_cpu, &imbalance, idle, &sd_idle, + cpus, balance); + + if (*balance == 0) + goto out_balanced; + + if (!group) { + schedstat_inc(sd, lb_nobusyg[idle]); + goto out_balanced; + } + + busiest = find_busiest_queue(group, idle, imbalance, cpus); + if (!busiest) { + schedstat_inc(sd, lb_nobusyq[idle]); + goto out_balanced; + } + + BUG_ON(busiest == this_rq); + + schedstat_add(sd, lb_imbalance[idle], imbalance); + + ld_moved = 0; + if (busiest->nr_running > 1) { + /* + * Attempt to move tasks. If find_busiest_group has found + * an imbalance but busiest->nr_running <= 1, the group is + * still unbalanced. ld_moved simply stays zero, so it is + * correctly treated as an imbalance. + */ + local_irq_save(flags); + double_rq_lock(this_rq, busiest); + ld_moved = move_tasks(this_rq, this_cpu, busiest, + imbalance, sd, idle, &all_pinned); + double_rq_unlock(this_rq, busiest); + local_irq_restore(flags); + + /* + * some other cpu did the load balance for us. + */ + if (ld_moved && this_cpu != smp_processor_id()) + resched_cpu(this_cpu); + + /* All tasks on this runqueue were pinned by CPU affinity */ + if (unlikely(all_pinned)) { + cpumask_clear_cpu(cpu_of(busiest), cpus); + if (!cpumask_empty(cpus)) + goto redo; + goto out_balanced; + } + } + + if (!ld_moved) { + schedstat_inc(sd, lb_failed[idle]); + sd->nr_balance_failed++; + + if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) { + + raw_spin_lock_irqsave(&busiest->lock, flags); + + /* don't kick the migration_thread, if the curr + * task on busiest cpu can't be moved to this_cpu + */ + if (!cpumask_test_cpu(this_cpu, + &busiest->curr->cpus_allowed)) { + raw_spin_unlock_irqrestore(&busiest->lock, + flags); + all_pinned = 1; + goto out_one_pinned; + } + + if (!busiest->active_balance) { + busiest->active_balance = 1; + busiest->push_cpu = this_cpu; + active_balance = 1; + } + raw_spin_unlock_irqrestore(&busiest->lock, flags); + if (active_balance) + wake_up_process(busiest->migration_thread); + + /* + * We've kicked active balancing, reset the failure + * counter. + */ + sd->nr_balance_failed = sd->cache_nice_tries+1; + } + } else + sd->nr_balance_failed = 0; + + if (likely(!active_balance)) { + /* We were unbalanced, so reset the balancing interval */ + sd->balance_interval = sd->min_interval; + } else { + /* + * If we've begun active balancing, start to back off. This + * case may not be covered by the all_pinned logic if there + * is only 1 task on the busy runqueue (because we don't call + * move_tasks). + */ + if (sd->balance_interval < sd->max_interval) + sd->balance_interval *= 2; + } + + if (!ld_moved && !sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + ld_moved = -1; + + goto out; + +out_balanced: + schedstat_inc(sd, lb_balanced[idle]); + + sd->nr_balance_failed = 0; + +out_one_pinned: + /* tune up the balancing interval */ + if ((all_pinned && sd->balance_interval < MAX_PINNED_INTERVAL) || + (sd->balance_interval < sd->max_interval)) + sd->balance_interval *= 2; + + if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + ld_moved = -1; + else + ld_moved = 0; +out: + if (ld_moved) + update_shares(sd); + return ld_moved; +} + +/* + * Check this_cpu to ensure it is balanced within domain. Attempt to move + * tasks if there is an imbalance. + * + * Called from schedule when this_rq is about to become idle (CPU_NEWLY_IDLE). + * this_rq is locked. + */ +static int +load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd) +{ + struct sched_group *group; + struct rq *busiest = NULL; + unsigned long imbalance; + int ld_moved = 0; + int sd_idle = 0; + int all_pinned = 0; + struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); + + cpumask_copy(cpus, cpu_active_mask); + + /* + * When power savings policy is enabled for the parent domain, idle + * sibling can pick up load irrespective of busy siblings. In this case, + * let the state of idle sibling percolate up as IDLE, instead of + * portraying it as CPU_NOT_IDLE. + */ + if (sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + sd_idle = 1; + + schedstat_inc(sd, lb_count[CPU_NEWLY_IDLE]); +redo: + update_shares_locked(this_rq, sd); + group = find_busiest_group(sd, this_cpu, &imbalance, CPU_NEWLY_IDLE, + &sd_idle, cpus, NULL); + if (!group) { + schedstat_inc(sd, lb_nobusyg[CPU_NEWLY_IDLE]); + goto out_balanced; + } + + busiest = find_busiest_queue(group, CPU_NEWLY_IDLE, imbalance, cpus); + if (!busiest) { + schedstat_inc(sd, lb_nobusyq[CPU_NEWLY_IDLE]); + goto out_balanced; + } + + BUG_ON(busiest == this_rq); + + schedstat_add(sd, lb_imbalance[CPU_NEWLY_IDLE], imbalance); + + ld_moved = 0; + if (busiest->nr_running > 1) { + /* Attempt to move tasks */ + double_lock_balance(this_rq, busiest); + /* this_rq->clock is already updated */ + update_rq_clock(busiest); + ld_moved = move_tasks(this_rq, this_cpu, busiest, + imbalance, sd, CPU_NEWLY_IDLE, + &all_pinned); + double_unlock_balance(this_rq, busiest); + + if (unlikely(all_pinned)) { + cpumask_clear_cpu(cpu_of(busiest), cpus); + if (!cpumask_empty(cpus)) + goto redo; + } + } + + if (!ld_moved) { + int active_balance = 0; + + schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]); + if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + return -1; + + if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) + return -1; + + if (sd->nr_balance_failed++ < 2) + return -1; + + /* + * The only task running in a non-idle cpu can be moved to this + * cpu in an attempt to completely freeup the other CPU + * package. The same method used to move task in load_balance() + * have been extended for load_balance_newidle() to speedup + * consolidation at sched_mc=POWERSAVINGS_BALANCE_WAKEUP (2) + * + * The package power saving logic comes from + * find_busiest_group(). If there are no imbalance, then + * f_b_g() will return NULL. However when sched_mc={1,2} then + * f_b_g() will select a group from which a running task may be + * pulled to this cpu in order to make the other package idle. + * If there is no opportunity to make a package idle and if + * there are no imbalance, then f_b_g() will return NULL and no + * action will be taken in load_balance_newidle(). + * + * Under normal task pull operation due to imbalance, there + * will be more than one task in the source run queue and + * move_tasks() will succeed. ld_moved will be true and this + * active balance code will not be triggered. + */ + + /* Lock busiest in correct order while this_rq is held */ + double_lock_balance(this_rq, busiest); + + /* + * don't kick the migration_thread, if the curr + * task on busiest cpu can't be moved to this_cpu + */ + if (!cpumask_test_cpu(this_cpu, &busiest->curr->cpus_allowed)) { + double_unlock_balance(this_rq, busiest); + all_pinned = 1; + return ld_moved; + } + + if (!busiest->active_balance) { + busiest->active_balance = 1; + busiest->push_cpu = this_cpu; + active_balance = 1; + } + + double_unlock_balance(this_rq, busiest); + /* + * Should not call ttwu while holding a rq->lock + */ + raw_spin_unlock(&this_rq->lock); + if (active_balance) + wake_up_process(busiest->migration_thread); + raw_spin_lock(&this_rq->lock); + + } else + sd->nr_balance_failed = 0; + + update_shares_locked(this_rq, sd); + return ld_moved; + +out_balanced: + schedstat_inc(sd, lb_balanced[CPU_NEWLY_IDLE]); + if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + return -1; + sd->nr_balance_failed = 0; + + return 0; +} + +/* + * idle_balance is called by schedule() if this_cpu is about to become + * idle. Attempts to pull tasks from other CPUs. + */ +static void idle_balance(int this_cpu, struct rq *this_rq) +{ + struct sched_domain *sd; + int pulled_task = 0; + unsigned long next_balance = jiffies + HZ; + + this_rq->idle_stamp = this_rq->clock; + + if (this_rq->avg_idle < sysctl_sched_migration_cost) + return; + + for_each_domain(this_cpu, sd) { + unsigned long interval; + + if (!(sd->flags & SD_LOAD_BALANCE)) + continue; + + if (sd->flags & SD_BALANCE_NEWIDLE) + /* If we've pulled tasks over stop searching: */ + pulled_task = load_balance_newidle(this_cpu, this_rq, + sd); + + interval = msecs_to_jiffies(sd->balance_interval); + if (time_after(next_balance, sd->last_balance + interval)) + next_balance = sd->last_balance + interval; + if (pulled_task) { + this_rq->idle_stamp = 0; + break; + } + } + if (pulled_task || time_after(jiffies, this_rq->next_balance)) { + /* + * We are going idle. next_balance may be set based on + * a busy processor. So reset next_balance. + */ + this_rq->next_balance = next_balance; + } +} + +/* + * active_load_balance is run by migration threads. It pushes running tasks + * off the busiest CPU onto idle CPUs. It requires at least 1 task to be + * running on each physical CPU where possible, and avoids physical / + * logical imbalances. + * + * Called with busiest_rq locked. + */ +static void active_load_balance(struct rq *busiest_rq, int busiest_cpu) +{ + int target_cpu = busiest_rq->push_cpu; + struct sched_domain *sd; + struct rq *target_rq; + + /* Is there any task to move? */ + if (busiest_rq->nr_running <= 1) + return; + + target_rq = cpu_rq(target_cpu); + + /* + * This condition is "impossible", if it occurs + * we need to fix it. Originally reported by + * Bjorn Helgaas on a 128-cpu setup. + */ + BUG_ON(busiest_rq == target_rq); + + /* move a task from busiest_rq to target_rq */ + double_lock_balance(busiest_rq, target_rq); + update_rq_clock(busiest_rq); + update_rq_clock(target_rq); + + /* Search for an sd spanning us and the target CPU. */ + for_each_domain(target_cpu, sd) { + if ((sd->flags & SD_LOAD_BALANCE) && + cpumask_test_cpu(busiest_cpu, sched_domain_span(sd))) + break; + } + + if (likely(sd)) { + schedstat_inc(sd, alb_count); + + if (move_one_task(target_rq, target_cpu, busiest_rq, + sd, CPU_IDLE)) + schedstat_inc(sd, alb_pushed); + else + schedstat_inc(sd, alb_failed); + } + double_unlock_balance(busiest_rq, target_rq); +} + +#ifdef CONFIG_NO_HZ +static struct { + atomic_t load_balancer; + cpumask_var_t cpu_mask; + cpumask_var_t ilb_grp_nohz_mask; +} nohz ____cacheline_aligned = { + .load_balancer = ATOMIC_INIT(-1), +}; + +int get_nohz_load_balancer(void) +{ + return atomic_read(&nohz.load_balancer); +} + +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) +/** + * lowest_flag_domain - Return lowest sched_domain containing flag. + * @cpu: The cpu whose lowest level of sched domain is to + * be returned. + * @flag: The flag to check for the lowest sched_domain + * for the given cpu. + * + * Returns the lowest sched_domain of a cpu which contains the given flag. + */ +static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) +{ + struct sched_domain *sd; + + for_each_domain(cpu, sd) + if (sd && (sd->flags & flag)) + break; + + return sd; +} + +/** + * for_each_flag_domain - Iterates over sched_domains containing the flag. + * @cpu: The cpu whose domains we're iterating over. + * @sd: variable holding the value of the power_savings_sd + * for cpu. + * @flag: The flag to filter the sched_domains to be iterated. + * + * Iterates over all the scheduler domains for a given cpu that has the 'flag' + * set, starting from the lowest sched_domain to the highest. + */ +#define for_each_flag_domain(cpu, sd, flag) \ + for (sd = lowest_flag_domain(cpu, flag); \ + (sd && (sd->flags & flag)); sd = sd->parent) + +/** + * is_semi_idle_group - Checks if the given sched_group is semi-idle. + * @ilb_group: group to be checked for semi-idleness + * + * Returns: 1 if the group is semi-idle. 0 otherwise. + * + * We define a sched_group to be semi idle if it has atleast one idle-CPU + * and atleast one non-idle CPU. This helper function checks if the given + * sched_group is semi-idle or not. + */ +static inline int is_semi_idle_group(struct sched_group *ilb_group) +{ + cpumask_and(nohz.ilb_grp_nohz_mask, nohz.cpu_mask, + sched_group_cpus(ilb_group)); + + /* + * A sched_group is semi-idle when it has atleast one busy cpu + * and atleast one idle cpu. + */ + if (cpumask_empty(nohz.ilb_grp_nohz_mask)) + return 0; + + if (cpumask_equal(nohz.ilb_grp_nohz_mask, sched_group_cpus(ilb_group))) + return 0; + + return 1; +} +/** + * find_new_ilb - Finds the optimum idle load balancer for nomination. + * @cpu: The cpu which is nominating a new idle_load_balancer. + * + * Returns: Returns the id of the idle load balancer if it exists, + * Else, returns >= nr_cpu_ids. + * + * This algorithm picks the idle load balancer such that it belongs to a + * semi-idle powersavings sched_domain. The idea is to try and avoid + * completely idle packages/cores just for the purpose of idle load balancing + * when there are other idle cpu's which are better suited for that job. + */ +static int find_new_ilb(int cpu) +{ + struct sched_domain *sd; + struct sched_group *ilb_group; + + /* + * Have idle load balancer selection from semi-idle packages only + * when power-aware load balancing is enabled + */ + if (!(sched_smt_power_savings || sched_mc_power_savings)) + goto out_done; + + /* + * Optimize for the case when we have no idle CPUs or only one + * idle CPU. Don't walk the sched_domain hierarchy in such cases + */ + if (cpumask_weight(nohz.cpu_mask) < 2) + goto out_done; + + for_each_flag_domain(cpu, sd, SD_POWERSAVINGS_BALANCE) { + ilb_group = sd->groups; + + do { + if (is_semi_idle_group(ilb_group)) + return cpumask_first(nohz.ilb_grp_nohz_mask); + + ilb_group = ilb_group->next; + + } while (ilb_group != sd->groups); + } + +out_done: + return cpumask_first(nohz.cpu_mask); +} +#else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ +static inline int find_new_ilb(int call_cpu) +{ + return cpumask_first(nohz.cpu_mask); +} +#endif + +/* + * This routine will try to nominate the ilb (idle load balancing) + * owner among the cpus whose ticks are stopped. ilb owner will do the idle + * load balancing on behalf of all those cpus. If all the cpus in the system + * go into this tickless mode, then there will be no ilb owner (as there is + * no need for one) and all the cpus will sleep till the next wakeup event + * arrives... + * + * For the ilb owner, tick is not stopped. And this tick will be used + * for idle load balancing. ilb owner will still be part of + * nohz.cpu_mask.. + * + * While stopping the tick, this cpu will become the ilb owner if there + * is no other owner. And will be the owner till that cpu becomes busy + * or if all cpus in the system stop their ticks at which point + * there is no need for ilb owner. + * + * When the ilb owner becomes busy, it nominates another owner, during the + * next busy scheduler_tick() + */ +int select_nohz_load_balancer(int stop_tick) +{ + int cpu = smp_processor_id(); + + if (stop_tick) { + cpu_rq(cpu)->in_nohz_recently = 1; + + if (!cpu_active(cpu)) { + if (atomic_read(&nohz.load_balancer) != cpu) + return 0; + + /* + * If we are going offline and still the leader, + * give up! + */ + if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) + BUG(); + + return 0; + } + + cpumask_set_cpu(cpu, nohz.cpu_mask); + + /* time for ilb owner also to sleep */ + if (cpumask_weight(nohz.cpu_mask) == num_active_cpus()) { + if (atomic_read(&nohz.load_balancer) == cpu) + atomic_set(&nohz.load_balancer, -1); + return 0; + } + + if (atomic_read(&nohz.load_balancer) == -1) { + /* make me the ilb owner */ + if (atomic_cmpxchg(&nohz.load_balancer, -1, cpu) == -1) + return 1; + } else if (atomic_read(&nohz.load_balancer) == cpu) { + int new_ilb; + + if (!(sched_smt_power_savings || + sched_mc_power_savings)) + return 1; + /* + * Check to see if there is a more power-efficient + * ilb. + */ + new_ilb = find_new_ilb(cpu); + if (new_ilb < nr_cpu_ids && new_ilb != cpu) { + atomic_set(&nohz.load_balancer, -1); + resched_cpu(new_ilb); + return 0; + } + return 1; + } + } else { + if (!cpumask_test_cpu(cpu, nohz.cpu_mask)) + return 0; + + cpumask_clear_cpu(cpu, nohz.cpu_mask); + + if (atomic_read(&nohz.load_balancer) == cpu) + if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) + BUG(); + } + return 0; +} +#endif + +static DEFINE_SPINLOCK(balancing); + +/* + * It checks each scheduling domain to see if it is due to be balanced, + * and initiates a balancing operation if so. + * + * Balancing parameters are set up in arch_init_sched_domains. + */ +static void rebalance_domains(int cpu, enum cpu_idle_type idle) +{ + int balance = 1; + struct rq *rq = cpu_rq(cpu); + unsigned long interval; + struct sched_domain *sd; + /* Earliest time when we have to do rebalance again */ + unsigned long next_balance = jiffies + 60*HZ; + int update_next_balance = 0; + int need_serialize; + + for_each_domain(cpu, sd) { + if (!(sd->flags & SD_LOAD_BALANCE)) + continue; + + interval = sd->balance_interval; + if (idle != CPU_IDLE) + interval *= sd->busy_factor; + + /* scale ms to jiffies */ + interval = msecs_to_jiffies(interval); + if (unlikely(!interval)) + interval = 1; + if (interval > HZ*NR_CPUS/10) + interval = HZ*NR_CPUS/10; + + need_serialize = sd->flags & SD_SERIALIZE; + + if (need_serialize) { + if (!spin_trylock(&balancing)) + goto out; + } + + if (time_after_eq(jiffies, sd->last_balance + interval)) { + if (load_balance(cpu, rq, sd, idle, &balance)) { + /* + * We've pulled tasks over so either we're no + * longer idle, or one of our SMT siblings is + * not idle. + */ + idle = CPU_NOT_IDLE; + } + sd->last_balance = jiffies; + } + if (need_serialize) + spin_unlock(&balancing); +out: + if (time_after(next_balance, sd->last_balance + interval)) { + next_balance = sd->last_balance + interval; + update_next_balance = 1; + } + + /* + * Stop the load balance at this level. There is another + * CPU in our sched group which is doing load balancing more + * actively. + */ + if (!balance) + break; + } + + /* + * next_balance will be updated only when there is a need. + * When the cpu is attached to null domain for ex, it will not be + * updated. + */ + if (likely(update_next_balance)) + rq->next_balance = next_balance; +} + +/* + * run_rebalance_domains is triggered when needed from the scheduler tick. + * In CONFIG_NO_HZ case, the idle load balance owner will do the + * rebalancing for all the cpus for whom scheduler ticks are stopped. + */ +static void run_rebalance_domains(struct softirq_action *h) +{ + int this_cpu = smp_processor_id(); + struct rq *this_rq = cpu_rq(this_cpu); + enum cpu_idle_type idle = this_rq->idle_at_tick ? + CPU_IDLE : CPU_NOT_IDLE; + + rebalance_domains(this_cpu, idle); + +#ifdef CONFIG_NO_HZ + /* + * If this cpu is the owner for idle load balancing, then do the + * balancing on behalf of the other idle cpus whose ticks are + * stopped. + */ + if (this_rq->idle_at_tick && + atomic_read(&nohz.load_balancer) == this_cpu) { + struct rq *rq; + int balance_cpu; + + for_each_cpu(balance_cpu, nohz.cpu_mask) { + if (balance_cpu == this_cpu) + continue; + + /* + * If this cpu gets work to do, stop the load balancing + * work being done for other cpus. Next load + * balancing owner will pick it up. + */ + if (need_resched()) + break; + + rebalance_domains(balance_cpu, CPU_IDLE); + + rq = cpu_rq(balance_cpu); + if (time_after(this_rq->next_balance, rq->next_balance)) + this_rq->next_balance = rq->next_balance; + } + } +#endif +} + +static inline int on_null_domain(int cpu) +{ + return !rcu_dereference(cpu_rq(cpu)->sd); +} + +/* + * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing. + * + * In case of CONFIG_NO_HZ, this is the place where we nominate a new + * idle load balancing owner or decide to stop the periodic load balancing, + * if the whole system is idle. + */ +static inline void trigger_load_balance(struct rq *rq, int cpu) +{ +#ifdef CONFIG_NO_HZ + /* + * If we were in the nohz mode recently and busy at the current + * scheduler tick, then check if we need to nominate new idle + * load balancer. + */ + if (rq->in_nohz_recently && !rq->idle_at_tick) { + rq->in_nohz_recently = 0; + + if (atomic_read(&nohz.load_balancer) == cpu) { + cpumask_clear_cpu(cpu, nohz.cpu_mask); + atomic_set(&nohz.load_balancer, -1); + } + + if (atomic_read(&nohz.load_balancer) == -1) { + int ilb = find_new_ilb(cpu); + + if (ilb < nr_cpu_ids) + resched_cpu(ilb); + } + } + + /* + * If this cpu is idle and doing idle load balancing for all the + * cpus with ticks stopped, is it time for that to stop? + */ + if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) == cpu && + cpumask_weight(nohz.cpu_mask) == num_online_cpus()) { + resched_cpu(cpu); + return; + } + + /* + * If this cpu is idle and the idle load balancing is done by + * someone else, then no need raise the SCHED_SOFTIRQ + */ + if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) != cpu && + cpumask_test_cpu(cpu, nohz.cpu_mask)) + return; +#endif + /* Don't need to rebalance while attached to NULL domain */ + if (time_after_eq(jiffies, rq->next_balance) && + likely(!on_null_domain(cpu))) + raise_softirq(SCHED_SOFTIRQ); +} + static void rq_online_fair(struct rq *rq) { update_sysctl(); @@ -1962,6 +3718,15 @@ static void rq_offline_fair(struct rq *rq) update_sysctl(); } +#else /* CONFIG_SMP */ + +/* + * on UP we do not need to balance between CPUs: + */ +static inline void idle_balance(int cpu, struct rq *rq) +{ +} + #endif /* CONFIG_SMP */ /* -- cgit v0.10.2 From 3d45fd804a95055ecab5b3eed81f5ab2dbb047a2 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 17:12:46 +0100 Subject: sched: Remove the sched_class load_balance methods Take out the sched_class methods for load-balancing. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/include/linux/sched.h b/include/linux/sched.h index f2f842d..50d685c 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1087,14 +1087,6 @@ struct sched_class { #ifdef CONFIG_SMP int (*select_task_rq)(struct task_struct *p, int sd_flag, int flags); - unsigned long (*load_balance) (struct rq *this_rq, int this_cpu, - struct rq *busiest, unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio); - - int (*move_one_task) (struct rq *this_rq, int this_cpu, - struct rq *busiest, struct sched_domain *sd, - enum cpu_idle_type idle); void (*pre_schedule) (struct rq *this_rq, struct task_struct *task); void (*post_schedule) (struct rq *this_rq); void (*task_waking) (struct rq *this_rq, struct task_struct *task); diff --git a/kernel/sched.c b/kernel/sched.c index 13a2acf..c0be079 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1390,32 +1390,6 @@ static const u32 prio_to_wmult[40] = { /* 15 */ 119304647, 148102320, 186737708, 238609294, 286331153, }; -static void activate_task(struct rq *rq, struct task_struct *p, int wakeup); - -/* - * runqueue iterator, to support SMP load-balancing between different - * scheduling classes, without having to expose their internal data - * structures to the load-balancing proper: - */ -struct rq_iterator { - void *arg; - struct task_struct *(*start)(void *); - struct task_struct *(*next)(void *); -}; - -#ifdef CONFIG_SMP -static unsigned long -balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct rq_iterator *iterator); - -static int -iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle, - struct rq_iterator *iterator); -#endif - /* Time spent by the tasks of the cpu accounting group executing in ... */ enum cpuacct_stat_index { CPUACCT_STAT_USER, /* ... user mode */ diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 5116b81..faf9a2f 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1851,6 +1851,24 @@ static struct task_struct *load_balance_next_fair(void *arg) return __load_balance_iterator(cfs_rq, cfs_rq->balance_iterator); } +/* + * runqueue iterator, to support SMP load-balancing between different + * scheduling classes, without having to expose their internal data + * structures to the load-balancing proper: + */ +struct rq_iterator { + void *arg; + struct task_struct *(*start)(void *); + struct task_struct *(*next)(void *); +}; + +static unsigned long +balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, struct sched_domain *sd, + enum cpu_idle_type idle, int *all_pinned, + int *this_best_prio, struct rq_iterator *iterator); + + static unsigned long __load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, @@ -1929,8 +1947,20 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, #endif static int -move_one_task_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) +iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle, + struct rq_iterator *iterator); + +/* + * move_one_task tries to move exactly one task from busiest to this_rq, as + * part of active balancing operations within "domain". + * Returns 1 if successful and 0 otherwise. + * + * Called with both runqueues locked. + */ +static int +move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle) { struct cfs_rq *busy_cfs_rq; struct rq_iterator cfs_rq_iterator; @@ -2094,16 +2124,15 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned) { - const struct sched_class *class = sched_class_highest; - unsigned long total_load_moved = 0; + unsigned long total_load_moved = 0, load_moved; int this_best_prio = this_rq->curr->prio; do { - total_load_moved += - class->load_balance(this_rq, this_cpu, busiest, + load_moved = load_balance_fair(this_rq, this_cpu, busiest, max_load_move - total_load_moved, sd, idle, all_pinned, &this_best_prio); - class = class->next; + + total_load_moved += load_moved; #ifdef CONFIG_PREEMPT /* @@ -2114,7 +2143,7 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) break; #endif - } while (class && max_load_move > total_load_moved); + } while (load_moved && max_load_move > total_load_moved); return total_load_moved > 0; } @@ -2145,25 +2174,6 @@ iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, return 0; } -/* - * move_one_task tries to move exactly one task from busiest to this_rq, as - * part of active balancing operations within "domain". - * Returns 1 if successful and 0 otherwise. - * - * Called with both runqueues locked. - */ -static int move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - const struct sched_class *class; - - for_each_class(class) { - if (class->move_one_task(this_rq, this_cpu, busiest, sd, idle)) - return 1; - } - - return 0; -} /********** Helpers for find_busiest_group ************************/ /* * sd_lb_stats - Structure to store the statistics of a sched_domain @@ -3873,8 +3883,6 @@ static const struct sched_class fair_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_fair, - .load_balance = load_balance_fair, - .move_one_task = move_one_task_fair, .rq_online = rq_online_fair, .rq_offline = rq_offline_fair, diff --git a/kernel/sched_idletask.c b/kernel/sched_idletask.c index 01332bf..a8a6d8a 100644 --- a/kernel/sched_idletask.c +++ b/kernel/sched_idletask.c @@ -44,24 +44,6 @@ static void put_prev_task_idle(struct rq *rq, struct task_struct *prev) { } -#ifdef CONFIG_SMP -static unsigned long -load_balance_idle(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio) -{ - return 0; -} - -static int -move_one_task_idle(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - return 0; -} -#endif - static void task_tick_idle(struct rq *rq, struct task_struct *curr, int queued) { } @@ -119,9 +101,6 @@ static const struct sched_class idle_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_idle, - - .load_balance = load_balance_idle, - .move_one_task = move_one_task_idle, #endif .set_curr_task = set_curr_task_idle, diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 072b3fc..502bb61 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -1481,24 +1481,6 @@ static void task_woken_rt(struct rq *rq, struct task_struct *p) push_rt_tasks(rq); } -static unsigned long -load_balance_rt(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio) -{ - /* don't touch RT tasks */ - return 0; -} - -static int -move_one_task_rt(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - /* don't touch RT tasks */ - return 0; -} - static void set_cpus_allowed_rt(struct task_struct *p, const struct cpumask *new_mask) { @@ -1746,8 +1728,6 @@ static const struct sched_class rt_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_rt, - .load_balance = load_balance_rt, - .move_one_task = move_one_task_rt, .set_cpus_allowed = set_cpus_allowed_rt, .rq_online = rq_online_rt, .rq_offline = rq_offline_rt, -- cgit v0.10.2 From ee00e66ffff250fb0d3a789e5565462f71c7c9a7 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 17:25:20 +0100 Subject: sched: Remove rq_iterator usage from load_balance_fair Since we only ever iterate the fair class, do away with this abstraction. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index faf9a2f..709deb3 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1866,26 +1866,9 @@ static unsigned long balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct rq_iterator *iterator); + int *this_best_prio, struct cfs_rq *busiest_cfs_rq); -static unsigned long -__load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, int *this_best_prio, - struct cfs_rq *cfs_rq) -{ - struct rq_iterator cfs_rq_iterator; - - cfs_rq_iterator.start = load_balance_start_fair; - cfs_rq_iterator.next = load_balance_next_fair; - cfs_rq_iterator.arg = cfs_rq; - - return balance_tasks(this_rq, this_cpu, busiest, - max_load_move, sd, idle, all_pinned, - this_best_prio, &cfs_rq_iterator); -} - #ifdef CONFIG_FAIR_GROUP_SCHED static unsigned long load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, @@ -1915,9 +1898,9 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, rem_load = (u64)rem_load_move * busiest_weight; rem_load = div_u64(rem_load, busiest_h_load + 1); - moved_load = __load_balance_fair(this_rq, this_cpu, busiest, + moved_load = balance_tasks(this_rq, this_cpu, busiest, rem_load, sd, idle, all_pinned, this_best_prio, - tg->cfs_rq[busiest_cpu]); + busiest_cfs_rq); if (!moved_load) continue; @@ -1940,7 +1923,7 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned, int *this_best_prio) { - return __load_balance_fair(this_rq, this_cpu, busiest, + return balance_tasks(this_rq, this_cpu, busiest, max_load_move, sd, idle, all_pinned, this_best_prio, &busiest->cfs); } @@ -2050,53 +2033,48 @@ static unsigned long balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct rq_iterator *iterator) + int *this_best_prio, struct cfs_rq *busiest_cfs_rq) { int loops = 0, pulled = 0, pinned = 0; - struct task_struct *p; long rem_load_move = max_load_move; + struct task_struct *p, *n; if (max_load_move == 0) goto out; pinned = 1; - /* - * Start the load-balancing iterator: - */ - p = iterator->start(iterator->arg); -next: - if (!p || loops++ > sysctl_sched_nr_migrate) - goto out; + list_for_each_entry_safe(p, n, &busiest_cfs_rq->tasks, se.group_node) { + if (loops++ > sysctl_sched_nr_migrate) + break; - if ((p->se.load.weight >> 1) > rem_load_move || - !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { - p = iterator->next(iterator->arg); - goto next; - } + if ((p->se.load.weight >> 1) > rem_load_move || + !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) + continue; - pull_task(busiest, p, this_rq, this_cpu); - pulled++; - rem_load_move -= p->se.load.weight; + pull_task(busiest, p, this_rq, this_cpu); + pulled++; + rem_load_move -= p->se.load.weight; #ifdef CONFIG_PREEMPT - /* - * NEWIDLE balancing is a source of latency, so preemptible kernels - * will stop after the first task is pulled to minimize the critical - * section. - */ - if (idle == CPU_NEWLY_IDLE) - goto out; + /* + * NEWIDLE balancing is a source of latency, so preemptible + * kernels will stop after the first task is pulled to minimize + * the critical section. + */ + if (idle == CPU_NEWLY_IDLE) + break; #endif - /* - * We only want to steal up to the prescribed amount of weighted load. - */ - if (rem_load_move > 0) { + /* + * We only want to steal up to the prescribed amount of + * weighted load. + */ + if (rem_load_move <= 0) + break; + if (p->prio < *this_best_prio) *this_best_prio = p->prio; - p = iterator->next(iterator->arg); - goto next; } out: /* -- cgit v0.10.2 From 897c395f4c94ae19302f92393a0b8304e414ee06 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 17:45:42 +0100 Subject: sched: Remove rq_iterator from move_one_task Again, since we only iterate the fair class, remove the abstraction. Since this is the last user of the rq_iterator, remove all that too. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 709deb3..e48e459 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1814,54 +1814,6 @@ static void put_prev_task_fair(struct rq *rq, struct task_struct *prev) * Fair scheduling class load-balancing methods: */ -/* - * Load-balancing iterator. Note: while the runqueue stays locked - * during the whole iteration, the current task might be - * dequeued so the iterator has to be dequeue-safe. Here we - * achieve that by always pre-iterating before returning - * the current task: - */ -static struct task_struct * -__load_balance_iterator(struct cfs_rq *cfs_rq, struct list_head *next) -{ - struct task_struct *p = NULL; - struct sched_entity *se; - - if (next == &cfs_rq->tasks) - return NULL; - - se = list_entry(next, struct sched_entity, group_node); - p = task_of(se); - cfs_rq->balance_iterator = next->next; - - return p; -} - -static struct task_struct *load_balance_start_fair(void *arg) -{ - struct cfs_rq *cfs_rq = arg; - - return __load_balance_iterator(cfs_rq, cfs_rq->tasks.next); -} - -static struct task_struct *load_balance_next_fair(void *arg) -{ - struct cfs_rq *cfs_rq = arg; - - return __load_balance_iterator(cfs_rq, cfs_rq->balance_iterator); -} - -/* - * runqueue iterator, to support SMP load-balancing between different - * scheduling classes, without having to expose their internal data - * structures to the load-balancing proper: - */ -struct rq_iterator { - void *arg; - struct task_struct *(*start)(void *); - struct task_struct *(*next)(void *); -}; - static unsigned long balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, @@ -1929,42 +1881,6 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, } #endif -static int -iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle, - struct rq_iterator *iterator); - -/* - * move_one_task tries to move exactly one task from busiest to this_rq, as - * part of active balancing operations within "domain". - * Returns 1 if successful and 0 otherwise. - * - * Called with both runqueues locked. - */ -static int -move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - struct cfs_rq *busy_cfs_rq; - struct rq_iterator cfs_rq_iterator; - - cfs_rq_iterator.start = load_balance_start_fair; - cfs_rq_iterator.next = load_balance_next_fair; - - for_each_leaf_cfs_rq(busiest, busy_cfs_rq) { - /* - * pass busy_cfs_rq argument into - * load_balance_[start|next]_fair iterators - */ - cfs_rq_iterator.arg = busy_cfs_rq; - if (iter_move_one_task(this_rq, this_cpu, busiest, sd, idle, - &cfs_rq_iterator)) - return 1; - } - - return 0; -} - /* * pull_task - move a task from a remote runqueue to the local runqueue. * Both runqueues must be locked. @@ -2029,6 +1945,42 @@ int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, return 1; } +/* + * move_one_task tries to move exactly one task from busiest to this_rq, as + * part of active balancing operations within "domain". + * Returns 1 if successful and 0 otherwise. + * + * Called with both runqueues locked. + */ +static int +move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle) +{ + struct task_struct *p, *n; + struct cfs_rq *cfs_rq; + int pinned = 0; + + for_each_leaf_cfs_rq(busiest, cfs_rq) { + list_for_each_entry_safe(p, n, &cfs_rq->tasks, se.group_node) { + + if (!can_migrate_task(p, busiest, this_cpu, + sd, idle, &pinned)) + continue; + + pull_task(busiest, p, this_rq, this_cpu); + /* + * Right now, this is only the second place pull_task() + * is called, so we can safely collect pull_task() + * stats here rather than inside pull_task(). + */ + schedstat_inc(sd, lb_gained[idle]); + return 1; + } + } + + return 0; +} + static unsigned long balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, @@ -2126,32 +2078,6 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, return total_load_moved > 0; } -static int -iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle, - struct rq_iterator *iterator) -{ - struct task_struct *p = iterator->start(iterator->arg); - int pinned = 0; - - while (p) { - if (can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { - pull_task(busiest, p, this_rq, this_cpu); - /* - * Right now, this is only the second place pull_task() - * is called, so we can safely collect pull_task() - * stats here rather than inside pull_task(). - */ - schedstat_inc(sd, lb_gained[idle]); - - return 1; - } - p = iterator->next(iterator->arg); - } - - return 0; -} - /********** Helpers for find_busiest_group ************************/ /* * sd_lb_stats - Structure to store the statistics of a sched_domain -- cgit v0.10.2 From 230059de77a4e0f6afba98073e73bc9fd471506e Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 17:47:12 +0100 Subject: sched: Remove from fwd decls Move code around to get rid of fwd declarations. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index e48e459..93fccba 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1814,73 +1814,6 @@ static void put_prev_task_fair(struct rq *rq, struct task_struct *prev) * Fair scheduling class load-balancing methods: */ -static unsigned long -balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct cfs_rq *busiest_cfs_rq); - - -#ifdef CONFIG_FAIR_GROUP_SCHED -static unsigned long -load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio) -{ - long rem_load_move = max_load_move; - int busiest_cpu = cpu_of(busiest); - struct task_group *tg; - - rcu_read_lock(); - update_h_load(busiest_cpu); - - list_for_each_entry_rcu(tg, &task_groups, list) { - struct cfs_rq *busiest_cfs_rq = tg->cfs_rq[busiest_cpu]; - unsigned long busiest_h_load = busiest_cfs_rq->h_load; - unsigned long busiest_weight = busiest_cfs_rq->load.weight; - u64 rem_load, moved_load; - - /* - * empty group - */ - if (!busiest_cfs_rq->task_weight) - continue; - - rem_load = (u64)rem_load_move * busiest_weight; - rem_load = div_u64(rem_load, busiest_h_load + 1); - - moved_load = balance_tasks(this_rq, this_cpu, busiest, - rem_load, sd, idle, all_pinned, this_best_prio, - busiest_cfs_rq); - - if (!moved_load) - continue; - - moved_load *= busiest_h_load; - moved_load = div_u64(moved_load, busiest_weight + 1); - - rem_load_move -= moved_load; - if (rem_load_move < 0) - break; - } - rcu_read_unlock(); - - return max_load_move - rem_load_move; -} -#else -static unsigned long -load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio) -{ - return balance_tasks(this_rq, this_cpu, busiest, - max_load_move, sd, idle, all_pinned, - this_best_prio, &busiest->cfs); -} -#endif - /* * pull_task - move a task from a remote runqueue to the local runqueue. * Both runqueues must be locked. @@ -2042,6 +1975,66 @@ out: return max_load_move - rem_load_move; } +#ifdef CONFIG_FAIR_GROUP_SCHED +static unsigned long +load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, + struct sched_domain *sd, enum cpu_idle_type idle, + int *all_pinned, int *this_best_prio) +{ + long rem_load_move = max_load_move; + int busiest_cpu = cpu_of(busiest); + struct task_group *tg; + + rcu_read_lock(); + update_h_load(busiest_cpu); + + list_for_each_entry_rcu(tg, &task_groups, list) { + struct cfs_rq *busiest_cfs_rq = tg->cfs_rq[busiest_cpu]; + unsigned long busiest_h_load = busiest_cfs_rq->h_load; + unsigned long busiest_weight = busiest_cfs_rq->load.weight; + u64 rem_load, moved_load; + + /* + * empty group + */ + if (!busiest_cfs_rq->task_weight) + continue; + + rem_load = (u64)rem_load_move * busiest_weight; + rem_load = div_u64(rem_load, busiest_h_load + 1); + + moved_load = balance_tasks(this_rq, this_cpu, busiest, + rem_load, sd, idle, all_pinned, this_best_prio, + busiest_cfs_rq); + + if (!moved_load) + continue; + + moved_load *= busiest_h_load; + moved_load = div_u64(moved_load, busiest_weight + 1); + + rem_load_move -= moved_load; + if (rem_load_move < 0) + break; + } + rcu_read_unlock(); + + return max_load_move - rem_load_move; +} +#else +static unsigned long +load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, + struct sched_domain *sd, enum cpu_idle_type idle, + int *all_pinned, int *this_best_prio) +{ + return balance_tasks(this_rq, this_cpu, busiest, + max_load_move, sd, idle, all_pinned, + this_best_prio, &busiest->cfs); +} +#endif + /* * move_tasks tries to move up to max_load_move weighted load from busiest to * this_rq, as part of a balancing operation within domain "sd". -- cgit v0.10.2 From baa8c1102f0cd86e69c1497d61d2ee177e663663 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 17 Dec 2009 18:10:09 +0100 Subject: sched: Add a lock break for PREEMPT=y Since load-balancing can hold rq->locks for quite a long while, allow breaking out early when there is lock contention. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 93fccba..65d0820 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2065,6 +2065,10 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, */ if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) break; + + if (raw_spin_is_contended(&this_rq->lock) || + raw_spin_is_contended(&busiest->lock)) + break; #endif } while (load_moved && max_load_move > total_load_moved); -- cgit v0.10.2 From 1af3ed3ddf27499c3f57662c4c29871e2b95e5f9 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 23 Dec 2009 15:10:31 +0100 Subject: sched: Unify load_balance{,_newidle}() load_balance() and load_balance_newidle() look remarkably similar, one key point they differ in is the condition on when to active balance. So split out that logic into a separate function. One side effect is that previously load_balance_newidle() used to fail and return -1 under these conditions, whereas now it doesn't. I've not yet fully figured out the whole -1 return case for either load_balance{,_newidle}(). Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 65d0820..1040832 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2816,6 +2816,39 @@ find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, /* Working cpumask for load_balance and load_balance_newidle. */ static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); +static int need_active_balance(struct sched_domain *sd, int sd_idle, int idle) +{ + if (idle == CPU_NEWLY_IDLE) { + /* + * The only task running in a non-idle cpu can be moved to this + * cpu in an attempt to completely freeup the other CPU + * package. + * + * The package power saving logic comes from + * find_busiest_group(). If there are no imbalance, then + * f_b_g() will return NULL. However when sched_mc={1,2} then + * f_b_g() will select a group from which a running task may be + * pulled to this cpu in order to make the other package idle. + * If there is no opportunity to make a package idle and if + * there are no imbalance, then f_b_g() will return NULL and no + * action will be taken in load_balance_newidle(). + * + * Under normal task pull operation due to imbalance, there + * will be more than one task in the source run queue and + * move_tasks() will succeed. ld_moved will be true and this + * active balance code will not be triggered. + */ + if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + return 0; + + if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) + return 0; + } + + return unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2); +} + /* * Check this_cpu to ensure it is balanced within domain. Attempt to move * tasks if there is an imbalance. @@ -2902,8 +2935,7 @@ redo: schedstat_inc(sd, lb_failed[idle]); sd->nr_balance_failed++; - if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) { - + if (need_active_balance(sd, sd_idle, idle)) { raw_spin_lock_irqsave(&busiest->lock, flags); /* don't kick the migration_thread, if the curr @@ -3049,66 +3081,37 @@ redo: int active_balance = 0; schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]); - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - - if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) - return -1; + sd->nr_balance_failed++; - if (sd->nr_balance_failed++ < 2) - return -1; + if (need_active_balance(sd, sd_idle, CPU_NEWLY_IDLE)) { + double_lock_balance(this_rq, busiest); - /* - * The only task running in a non-idle cpu can be moved to this - * cpu in an attempt to completely freeup the other CPU - * package. The same method used to move task in load_balance() - * have been extended for load_balance_newidle() to speedup - * consolidation at sched_mc=POWERSAVINGS_BALANCE_WAKEUP (2) - * - * The package power saving logic comes from - * find_busiest_group(). If there are no imbalance, then - * f_b_g() will return NULL. However when sched_mc={1,2} then - * f_b_g() will select a group from which a running task may be - * pulled to this cpu in order to make the other package idle. - * If there is no opportunity to make a package idle and if - * there are no imbalance, then f_b_g() will return NULL and no - * action will be taken in load_balance_newidle(). - * - * Under normal task pull operation due to imbalance, there - * will be more than one task in the source run queue and - * move_tasks() will succeed. ld_moved will be true and this - * active balance code will not be triggered. - */ + /* + * don't kick the migration_thread, if the curr + * task on busiest cpu can't be moved to this_cpu + */ + if (!cpumask_test_cpu(this_cpu, + &busiest->curr->cpus_allowed)) { + double_unlock_balance(this_rq, busiest); + all_pinned = 1; + return ld_moved; + } - /* Lock busiest in correct order while this_rq is held */ - double_lock_balance(this_rq, busiest); + if (!busiest->active_balance) { + busiest->active_balance = 1; + busiest->push_cpu = this_cpu; + active_balance = 1; + } - /* - * don't kick the migration_thread, if the curr - * task on busiest cpu can't be moved to this_cpu - */ - if (!cpumask_test_cpu(this_cpu, &busiest->curr->cpus_allowed)) { double_unlock_balance(this_rq, busiest); - all_pinned = 1; - return ld_moved; - } - - if (!busiest->active_balance) { - busiest->active_balance = 1; - busiest->push_cpu = this_cpu; - active_balance = 1; + /* + * Should not call ttwu while holding a rq->lock + */ + raw_spin_unlock(&this_rq->lock); + if (active_balance) + wake_up_process(busiest->migration_thread); + raw_spin_lock(&this_rq->lock); } - - double_unlock_balance(this_rq, busiest); - /* - * Should not call ttwu while holding a rq->lock - */ - raw_spin_unlock(&this_rq->lock); - if (active_balance) - wake_up_process(busiest->migration_thread); - raw_spin_lock(&this_rq->lock); - } else sd->nr_balance_failed = 0; -- cgit v0.10.2 From f492e12ef050e02bf0185b6b57874992591b9be1 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 23 Dec 2009 15:29:42 +0100 Subject: sched: Remove load_balance_newidle() The two functions: load_balance{,_newidle}() are very similar, with the following differences: - rq->lock usage - sb->balance_interval updates - *balance check So remove the load_balance_newidle() call with load_balance(.idle = CPU_NEWLY_IDLE), explicitly unlock the rq->lock before calling (would be done by double_lock_balance() anyway), and ignore the other differences for now. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 1040832..de5ab12 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -3010,125 +3010,6 @@ out: } /* - * Check this_cpu to ensure it is balanced within domain. Attempt to move - * tasks if there is an imbalance. - * - * Called from schedule when this_rq is about to become idle (CPU_NEWLY_IDLE). - * this_rq is locked. - */ -static int -load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd) -{ - struct sched_group *group; - struct rq *busiest = NULL; - unsigned long imbalance; - int ld_moved = 0; - int sd_idle = 0; - int all_pinned = 0; - struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); - - cpumask_copy(cpus, cpu_active_mask); - - /* - * When power savings policy is enabled for the parent domain, idle - * sibling can pick up load irrespective of busy siblings. In this case, - * let the state of idle sibling percolate up as IDLE, instead of - * portraying it as CPU_NOT_IDLE. - */ - if (sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - sd_idle = 1; - - schedstat_inc(sd, lb_count[CPU_NEWLY_IDLE]); -redo: - update_shares_locked(this_rq, sd); - group = find_busiest_group(sd, this_cpu, &imbalance, CPU_NEWLY_IDLE, - &sd_idle, cpus, NULL); - if (!group) { - schedstat_inc(sd, lb_nobusyg[CPU_NEWLY_IDLE]); - goto out_balanced; - } - - busiest = find_busiest_queue(group, CPU_NEWLY_IDLE, imbalance, cpus); - if (!busiest) { - schedstat_inc(sd, lb_nobusyq[CPU_NEWLY_IDLE]); - goto out_balanced; - } - - BUG_ON(busiest == this_rq); - - schedstat_add(sd, lb_imbalance[CPU_NEWLY_IDLE], imbalance); - - ld_moved = 0; - if (busiest->nr_running > 1) { - /* Attempt to move tasks */ - double_lock_balance(this_rq, busiest); - /* this_rq->clock is already updated */ - update_rq_clock(busiest); - ld_moved = move_tasks(this_rq, this_cpu, busiest, - imbalance, sd, CPU_NEWLY_IDLE, - &all_pinned); - double_unlock_balance(this_rq, busiest); - - if (unlikely(all_pinned)) { - cpumask_clear_cpu(cpu_of(busiest), cpus); - if (!cpumask_empty(cpus)) - goto redo; - } - } - - if (!ld_moved) { - int active_balance = 0; - - schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]); - sd->nr_balance_failed++; - - if (need_active_balance(sd, sd_idle, CPU_NEWLY_IDLE)) { - double_lock_balance(this_rq, busiest); - - /* - * don't kick the migration_thread, if the curr - * task on busiest cpu can't be moved to this_cpu - */ - if (!cpumask_test_cpu(this_cpu, - &busiest->curr->cpus_allowed)) { - double_unlock_balance(this_rq, busiest); - all_pinned = 1; - return ld_moved; - } - - if (!busiest->active_balance) { - busiest->active_balance = 1; - busiest->push_cpu = this_cpu; - active_balance = 1; - } - - double_unlock_balance(this_rq, busiest); - /* - * Should not call ttwu while holding a rq->lock - */ - raw_spin_unlock(&this_rq->lock); - if (active_balance) - wake_up_process(busiest->migration_thread); - raw_spin_lock(&this_rq->lock); - } - } else - sd->nr_balance_failed = 0; - - update_shares_locked(this_rq, sd); - return ld_moved; - -out_balanced: - schedstat_inc(sd, lb_balanced[CPU_NEWLY_IDLE]); - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - sd->nr_balance_failed = 0; - - return 0; -} - -/* * idle_balance is called by schedule() if this_cpu is about to become * idle. Attempts to pull tasks from other CPUs. */ @@ -3143,16 +3024,23 @@ static void idle_balance(int this_cpu, struct rq *this_rq) if (this_rq->avg_idle < sysctl_sched_migration_cost) return; + /* + * Drop the rq->lock, but keep IRQ/preempt disabled. + */ + raw_spin_unlock(&this_rq->lock); + for_each_domain(this_cpu, sd) { unsigned long interval; + int balance = 1; if (!(sd->flags & SD_LOAD_BALANCE)) continue; - if (sd->flags & SD_BALANCE_NEWIDLE) + if (sd->flags & SD_BALANCE_NEWIDLE) { /* If we've pulled tasks over stop searching: */ - pulled_task = load_balance_newidle(this_cpu, this_rq, - sd); + pulled_task = load_balance(this_cpu, this_rq, + sd, CPU_NEWLY_IDLE, &balance); + } interval = msecs_to_jiffies(sd->balance_interval); if (time_after(next_balance, sd->last_balance + interval)) @@ -3162,6 +3050,9 @@ static void idle_balance(int this_cpu, struct rq *this_rq) break; } } + + raw_spin_lock(&this_rq->lock); + if (pulled_task || time_after(jiffies, this_rq->next_balance)) { /* * We are going idle. next_balance may be set based on -- cgit v0.10.2 From 8f190fb3f7a405682666d3723f6ec370b5afe4da Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 24 Dec 2009 14:18:21 +0100 Subject: sched: Assume *balance is valid Since all load_balance() callers will have !NULL balance parameters we can now assume so and remove a few checks. Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index de5ab12..0b482f5 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2465,7 +2465,7 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, * to do the newly idle load balance. */ if (idle != CPU_NEWLY_IDLE && local_group && - balance_cpu != this_cpu && balance) { + balance_cpu != this_cpu) { *balance = 0; return; } @@ -2528,7 +2528,7 @@ static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, update_sg_lb_stats(sd, group, this_cpu, idle, load_idx, sd_idle, local_group, cpus, balance, &sgs); - if (local_group && balance && !(*balance)) + if (local_group && !(*balance)) return; sds->total_load += sgs.group_load; @@ -2720,7 +2720,7 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, * 5) The imbalance is within the specified limit. * 6) Any rebalance would lead to ping-pong */ - if (balance && !(*balance)) + if (!(*balance)) goto ret; if (!sds.busiest || sds.busiest_nr_running == 0) -- cgit v0.10.2 From 871e35bc9733f273eaf5ceb69bbd0423b58e5285 Mon Sep 17 00:00:00 2001 From: Gautham R Shenoy Date: Wed, 20 Jan 2010 14:02:44 -0600 Subject: sched: Fix the place where group powers are updated We want to update the sched_group_powers when balance_cpu == this_cpu. Currently the group powers are updated only if the balance_cpu is the first CPU in the local group. But balance_cpu = this_cpu could also be the first idle cpu in the group. Hence fix the place where the group powers are updated. Signed-off-by: Gautham R Shenoy Signed-off-by: Joel Schopp Signed-off-by: Peter Zijlstra LKML-Reference: <1264017764.5717.127.camel@jschopp-laptop> Signed-off-by: Ingo Molnar diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 0b482f5..22231cc 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -2418,11 +2418,8 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, unsigned long sum_avg_load_per_task; unsigned long avg_load_per_task; - if (local_group) { + if (local_group) balance_cpu = group_first_cpu(group); - if (balance_cpu == this_cpu) - update_group_power(sd, this_cpu); - } /* Tally up the load of all CPUs in the group */ sum_avg_load_per_task = avg_load_per_task = 0; @@ -2470,6 +2467,8 @@ static inline void update_sg_lb_stats(struct sched_domain *sd, return; } + update_group_power(sd, this_cpu); + /* Adjust by relative CPU power of the group */ sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; -- cgit v0.10.2 From 7c9414385ebfdd87cc542d4e7e3bb0dbb2d3ce25 Mon Sep 17 00:00:00 2001 From: Dhaval Giani Date: Wed, 20 Jan 2010 13:26:18 +0100 Subject: sched: Remove USER_SCHED Remove the USER_SCHED feature. It has been scheduled to be removed in 2.6.34 as per http://marc.info/?l=linux-kernel&m=125728479022976&w=2 Signed-off-by: Dhaval Giani Signed-off-by: Peter Zijlstra LKML-Reference: <1263990378.24844.3.camel@localhost> Signed-off-by: Ingo Molnar diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt index 870d190..04a3fc3 100644 --- a/Documentation/feature-removal-schedule.txt +++ b/Documentation/feature-removal-schedule.txt @@ -6,21 +6,6 @@ be removed from this file. --------------------------- -What: USER_SCHED -When: 2.6.34 - -Why: USER_SCHED was implemented as a proof of concept for group scheduling. - The effect of USER_SCHED can already be achieved from userspace with - the help of libcgroup. The removal of USER_SCHED will also simplify - the scheduler code with the removal of one major ifdef. There are also - issues USER_SCHED has with USER_NS. A decision was taken not to fix - those and instead remove USER_SCHED. Also new group scheduling - features will not be implemented for USER_SCHED. - -Who: Dhaval Giani - ---------------------------- - What: PRISM54 When: 2.6.34 diff --git a/include/linux/sched.h b/include/linux/sched.h index 50d685c..8b07973 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -731,14 +731,6 @@ struct user_struct { uid_t uid; struct user_namespace *user_ns; -#ifdef CONFIG_USER_SCHED - struct task_group *tg; -#ifdef CONFIG_SYSFS - struct kobject kobj; - struct delayed_work work; -#endif -#endif - #ifdef CONFIG_PERF_EVENTS atomic_long_t locked_vm; #endif @@ -2502,13 +2494,9 @@ extern long sched_getaffinity(pid_t pid, struct cpumask *mask); extern void normalize_rt_tasks(void); -#ifdef CONFIG_GROUP_SCHED +#ifdef CONFIG_CGROUP_SCHED extern struct task_group init_task_group; -#ifdef CONFIG_USER_SCHED -extern struct task_group root_task_group; -extern void set_tg_uid(struct user_struct *user); -#endif extern struct task_group *sched_create_group(struct task_group *parent); extern void sched_destroy_group(struct task_group *tg); diff --git a/init/Kconfig b/init/Kconfig index a23da9f..e9fa300 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -435,57 +435,6 @@ config LOG_BUF_SHIFT config HAVE_UNSTABLE_SCHED_CLOCK bool -config GROUP_SCHED - bool "Group CPU scheduler" - depends on EXPERIMENTAL - default n - help - This feature lets CPU scheduler recognize task groups and control CPU - bandwidth allocation to such task groups. - In order to create a group from arbitrary set of processes, use - CONFIG_CGROUPS. (See Control Group support.) - -config FAIR_GROUP_SCHED - bool "Group scheduling for SCHED_OTHER" - depends on GROUP_SCHED - default GROUP_SCHED - -config RT_GROUP_SCHED - bool "Group scheduling for SCHED_RR/FIFO" - depends on EXPERIMENTAL - depends on GROUP_SCHED - default n - help - This feature lets you explicitly allocate real CPU bandwidth - to users or control groups (depending on the "Basis for grouping tasks" - setting below. If enabled, it will also make it impossible to - schedule realtime tasks for non-root users until you allocate - realtime bandwidth for them. - See Documentation/scheduler/sched-rt-group.txt for more information. - -choice - depends on GROUP_SCHED - prompt "Basis for grouping tasks" - default USER_SCHED - -config USER_SCHED - bool "user id" - help - This option will choose userid as the basis for grouping - tasks, thus providing equal CPU bandwidth to each user. - -config CGROUP_SCHED - bool "Control groups" - depends on CGROUPS - help - This option allows you to create arbitrary task groups - using the "cgroup" pseudo filesystem and control - the cpu bandwidth allocated to each such task group. - Refer to Documentation/cgroups/cgroups.txt for more - information on "cgroup" pseudo filesystem. - -endchoice - menuconfig CGROUPS boolean "Control Group support" help @@ -606,6 +555,36 @@ config CGROUP_MEM_RES_CTLR_SWAP Now, memory usage of swap_cgroup is 2 bytes per entry. If swap page size is 4096bytes, 512k per 1Gbytes of swap. +menuconfig CGROUP_SCHED + bool "Group CPU scheduler" + depends on EXPERIMENTAL && CGROUPS + default n + help + This feature lets CPU scheduler recognize task groups and control CPU + bandwidth allocation to such task groups. It uses cgroups to group + tasks. + +if CGROUP_SCHED +config FAIR_GROUP_SCHED + bool "Group scheduling for SCHED_OTHER" + depends on CGROUP_SCHED + default CGROUP_SCHED + +config RT_GROUP_SCHED + bool "Group scheduling for SCHED_RR/FIFO" + depends on EXPERIMENTAL + depends on CGROUP_SCHED + default n + help + This feature lets you explicitly allocate real CPU bandwidth + to users or control groups (depending on the "Basis for grouping tasks" + setting below. If enabled, it will also make it impossible to + schedule realtime tasks for non-root users until you allocate + realtime bandwidth for them. + See Documentation/scheduler/sched-rt-group.txt for more information. + +endif #CGROUP_SCHED + endif # CGROUPS config MM_OWNER diff --git a/kernel/ksysfs.c b/kernel/ksysfs.c index 3feaf5a..6b1ccc3 100644 --- a/kernel/ksysfs.c +++ b/kernel/ksysfs.c @@ -197,16 +197,8 @@ static int __init ksysfs_init(void) goto group_exit; } - /* create the /sys/kernel/uids/ directory */ - error = uids_sysfs_init(); - if (error) - goto notes_exit; - return 0; -notes_exit: - if (notes_size > 0) - sysfs_remove_bin_file(kernel_kobj, ¬es_attr); group_exit: sysfs_remove_group(kernel_kobj, &kernel_attr_group); kset_exit: diff --git a/kernel/sched.c b/kernel/sched.c index c0be079..41e76d3 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -233,7 +233,7 @@ static void destroy_rt_bandwidth(struct rt_bandwidth *rt_b) */ static DEFINE_MUTEX(sched_domains_mutex); -#ifdef CONFIG_GROUP_SCHED +#ifdef CONFIG_CGROUP_SCHED #include @@ -243,13 +243,7 @@ static LIST_HEAD(task_groups); /* task group related information */ struct task_group { -#ifdef CONFIG_CGROUP_SCHED struct cgroup_subsys_state css; -#endif - -#ifdef CONFIG_USER_SCHED - uid_t uid; -#endif #ifdef CONFIG_FAIR_GROUP_SCHED /* schedulable entities of this group on each cpu */ @@ -274,35 +268,7 @@ struct task_group { struct list_head children; }; -#ifdef CONFIG_USER_SCHED - -/* Helper function to pass uid information to create_sched_user() */ -void set_tg_uid(struct user_struct *user) -{ - user->tg->uid = user->uid; -} - -/* - * Root task group. - * Every UID task group (including init_task_group aka UID-0) will - * be a child to this group. - */ -struct task_group root_task_group; - -#ifdef CONFIG_FAIR_GROUP_SCHED -/* Default task group's sched entity on each cpu */ -static DEFINE_PER_CPU(struct sched_entity, init_sched_entity); -/* Default task group's cfs_rq on each cpu */ -static DEFINE_PER_CPU_SHARED_ALIGNED(struct cfs_rq, init_tg_cfs_rq); -#endif /* CONFIG_FAIR_GROUP_SCHED */ - -#ifdef CONFIG_RT_GROUP_SCHED -static DEFINE_PER_CPU(struct sched_rt_entity, init_sched_rt_entity); -static DEFINE_PER_CPU_SHARED_ALIGNED(struct rt_rq, init_rt_rq_var); -#endif /* CONFIG_RT_GROUP_SCHED */ -#else /* !CONFIG_USER_SCHED */ #define root_task_group init_task_group -#endif /* CONFIG_USER_SCHED */ /* task_group_lock serializes add/remove of task groups and also changes to * a task group's cpu shares. @@ -318,11 +284,7 @@ static int root_task_group_empty(void) } #endif -#ifdef CONFIG_USER_SCHED -# define INIT_TASK_GROUP_LOAD (2*NICE_0_LOAD) -#else /* !CONFIG_USER_SCHED */ # define INIT_TASK_GROUP_LOAD NICE_0_LOAD -#endif /* CONFIG_USER_SCHED */ /* * A weight of 0 or 1 can cause arithmetics problems. @@ -348,11 +310,7 @@ static inline struct task_group *task_group(struct task_struct *p) { struct task_group *tg; -#ifdef CONFIG_USER_SCHED - rcu_read_lock(); - tg = __task_cred(p)->user->tg; - rcu_read_unlock(); -#elif defined(CONFIG_CGROUP_SCHED) +#ifdef CONFIG_CGROUP_SCHED tg = container_of(task_subsys_state(p, cpu_cgroup_subsys_id), struct task_group, css); #else @@ -383,7 +341,7 @@ static inline struct task_group *task_group(struct task_struct *p) return NULL; } -#endif /* CONFIG_GROUP_SCHED */ +#endif /* CONFIG_CGROUP_SCHED */ /* CFS-related fields in a runqueue */ struct cfs_rq { @@ -7678,9 +7636,6 @@ void __init sched_init(void) #ifdef CONFIG_RT_GROUP_SCHED alloc_size += 2 * nr_cpu_ids * sizeof(void **); #endif -#ifdef CONFIG_USER_SCHED - alloc_size *= 2; -#endif #ifdef CONFIG_CPUMASK_OFFSTACK alloc_size += num_possible_cpus() * cpumask_size(); #endif @@ -7694,13 +7649,6 @@ void __init sched_init(void) init_task_group.cfs_rq = (struct cfs_rq **)ptr; ptr += nr_cpu_ids * sizeof(void **); -#ifdef CONFIG_USER_SCHED - root_task_group.se = (struct sched_entity **)ptr; - ptr += nr_cpu_ids * sizeof(void **); - - root_task_group.cfs_rq = (struct cfs_rq **)ptr; - ptr += nr_cpu_ids * sizeof(void **); -#endif /* CONFIG_USER_SCHED */ #endif /* CONFIG_FAIR_GROUP_SCHED */ #ifdef CONFIG_RT_GROUP_SCHED init_task_group.rt_se = (struct sched_rt_entity **)ptr; @@ -7709,13 +7657,6 @@ void __init sched_init(void) init_task_group.rt_rq = (struct rt_rq **)ptr; ptr += nr_cpu_ids * sizeof(void **); -#ifdef CONFIG_USER_SCHED - root_task_group.rt_se = (struct sched_rt_entity **)ptr; - ptr += nr_cpu_ids * sizeof(void **); - - root_task_group.rt_rq = (struct rt_rq **)ptr; - ptr += nr_cpu_ids * sizeof(void **); -#endif /* CONFIG_USER_SCHED */ #endif /* CONFIG_RT_GROUP_SCHED */ #ifdef CONFIG_CPUMASK_OFFSTACK for_each_possible_cpu(i) { @@ -7735,22 +7676,13 @@ void __init sched_init(void) #ifdef CONFIG_RT_GROUP_SCHED init_rt_bandwidth(&init_task_group.rt_bandwidth, global_rt_period(), global_rt_runtime()); -#ifdef CONFIG_USER_SCHED - init_rt_bandwidth(&root_task_group.rt_bandwidth, - global_rt_period(), RUNTIME_INF); -#endif /* CONFIG_USER_SCHED */ #endif /* CONFIG_RT_GROUP_SCHED */ -#ifdef CONFIG_GROUP_SCHED +#ifdef CONFIG_CGROUP_SCHED list_add(&init_task_group.list, &task_groups); INIT_LIST_HEAD(&init_task_group.children); -#ifdef CONFIG_USER_SCHED - INIT_LIST_HEAD(&root_task_group.children); - init_task_group.parent = &root_task_group; - list_add(&init_task_group.siblings, &root_task_group.children); -#endif /* CONFIG_USER_SCHED */ -#endif /* CONFIG_GROUP_SCHED */ +#endif /* CONFIG_CGROUP_SCHED */ #if defined CONFIG_FAIR_GROUP_SCHED && defined CONFIG_SMP update_shares_data = __alloc_percpu(nr_cpu_ids * sizeof(unsigned long), @@ -7790,25 +7722,6 @@ void __init sched_init(void) * directly in rq->cfs (i.e init_task_group->se[] = NULL). */ init_tg_cfs_entry(&init_task_group, &rq->cfs, NULL, i, 1, NULL); -#elif defined CONFIG_USER_SCHED - root_task_group.shares = NICE_0_LOAD; - init_tg_cfs_entry(&root_task_group, &rq->cfs, NULL, i, 0, NULL); - /* - * In case of task-groups formed thr' the user id of tasks, - * init_task_group represents tasks belonging to root user. - * Hence it forms a sibling of all subsequent groups formed. - * In this case, init_task_group gets only a fraction of overall - * system cpu resource, based on the weight assigned to root - * user's cpu share (INIT_TASK_GROUP_LOAD). This is accomplished - * by letting tasks of init_task_group sit in a separate cfs_rq - * (init_tg_cfs_rq) and having one entity represent this group of - * tasks in rq->cfs (i.e init_task_group->se[] != NULL). - */ - init_tg_cfs_entry(&init_task_group, - &per_cpu(init_tg_cfs_rq, i), - &per_cpu(init_sched_entity, i), i, 1, - root_task_group.se[i]); - #endif #endif /* CONFIG_FAIR_GROUP_SCHED */ @@ -7817,12 +7730,6 @@ void __init sched_init(void) INIT_LIST_HEAD(&rq->leaf_rt_rq_list); #ifdef CONFIG_CGROUP_SCHED init_tg_rt_entry(&init_task_group, &rq->rt, NULL, i, 1, NULL); -#elif defined CONFIG_USER_SCHED - init_tg_rt_entry(&root_task_group, &rq->rt, NULL, i, 0, NULL); - init_tg_rt_entry(&init_task_group, - &per_cpu(init_rt_rq_var, i), - &per_cpu(init_sched_rt_entity, i), i, 1, - root_task_group.rt_se[i]); #endif #endif @@ -8218,7 +8125,7 @@ static inline void unregister_rt_sched_group(struct task_group *tg, int cpu) } #endif /* CONFIG_RT_GROUP_SCHED */ -#ifdef CONFIG_GROUP_SCHED +#ifdef CONFIG_CGROUP_SCHED static void free_sched_group(struct task_group *tg) { free_fair_sched_group(tg); @@ -8327,7 +8234,7 @@ void sched_move_task(struct task_struct *tsk) task_rq_unlock(rq, &flags); } -#endif /* CONFIG_GROUP_SCHED */ +#endif /* CONFIG_CGROUP_SCHED */ #ifdef CONFIG_FAIR_GROUP_SCHED static void __set_se_shares(struct sched_entity *se, unsigned long shares) @@ -8469,13 +8376,6 @@ static int tg_schedulable(struct task_group *tg, void *data) runtime = d->rt_runtime; } -#ifdef CONFIG_USER_SCHED - if (tg == &root_task_group) { - period = global_rt_period(); - runtime = global_rt_runtime(); - } -#endif - /* * Cannot have more runtime than the period. */ diff --git a/kernel/sys.c b/kernel/sys.c index 26a6b73..f75bf09 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -569,11 +569,6 @@ static int set_user(struct cred *new) if (!new_user) return -EAGAIN; - if (!task_can_switch_user(new_user, current)) { - free_uid(new_user); - return -EINVAL; - } - if (atomic_read(&new_user->processes) >= current->signal->rlim[RLIMIT_NPROC].rlim_cur && new_user != INIT_USER) { diff --git a/kernel/user.c b/kernel/user.c index 46d0165..766467b 100644 --- a/kernel/user.c +++ b/kernel/user.c @@ -56,9 +56,6 @@ struct user_struct root_user = { .sigpending = ATOMIC_INIT(0), .locked_shm = 0, .user_ns = &init_user_ns, -#ifdef CONFIG_USER_SCHED - .tg = &init_task_group, -#endif }; /* @@ -75,268 +72,6 @@ static void uid_hash_remove(struct user_struct *up) put_user_ns(up->user_ns); } -#ifdef CONFIG_USER_SCHED - -static void sched_destroy_user(struct user_struct *up) -{ - sched_destroy_group(up->tg); -} - -static int sched_create_user(struct user_struct *up) -{ - int rc = 0; - - up->tg = sched_create_group(&root_task_group); - if (IS_ERR(up->tg)) - rc = -ENOMEM; - - set_tg_uid(up); - - return rc; -} - -#else /* CONFIG_USER_SCHED */ - -static void sched_destroy_user(struct user_struct *up) { } -static int sched_create_user(struct user_struct *up) { return 0; } - -#endif /* CONFIG_USER_SCHED */ - -#if defined(CONFIG_USER_SCHED) && defined(CONFIG_SYSFS) - -static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) -{ - struct user_struct *user; - struct hlist_node *h; - - hlist_for_each_entry(user, h, hashent, uidhash_node) { - if (user->uid == uid) { - /* possibly resurrect an "almost deleted" object */ - if (atomic_inc_return(&user->__count) == 1) - cancel_delayed_work(&user->work); - return user; - } - } - - return NULL; -} - -static struct kset *uids_kset; /* represents the /sys/kernel/uids/ directory */ -static DEFINE_MUTEX(uids_mutex); - -static inline void uids_mutex_lock(void) -{ - mutex_lock(&uids_mutex); -} - -static inline void uids_mutex_unlock(void) -{ - mutex_unlock(&uids_mutex); -} - -/* uid directory attributes */ -#ifdef CONFIG_FAIR_GROUP_SCHED -static ssize_t cpu_shares_show(struct kobject *kobj, - struct kobj_attribute *attr, - char *buf) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - - return sprintf(buf, "%lu\n", sched_group_shares(up->tg)); -} - -static ssize_t cpu_shares_store(struct kobject *kobj, - struct kobj_attribute *attr, - const char *buf, size_t size) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - unsigned long shares; - int rc; - - sscanf(buf, "%lu", &shares); - - rc = sched_group_set_shares(up->tg, shares); - - return (rc ? rc : size); -} - -static struct kobj_attribute cpu_share_attr = - __ATTR(cpu_share, 0644, cpu_shares_show, cpu_shares_store); -#endif - -#ifdef CONFIG_RT_GROUP_SCHED -static ssize_t cpu_rt_runtime_show(struct kobject *kobj, - struct kobj_attribute *attr, - char *buf) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - - return sprintf(buf, "%ld\n", sched_group_rt_runtime(up->tg)); -} - -static ssize_t cpu_rt_runtime_store(struct kobject *kobj, - struct kobj_attribute *attr, - const char *buf, size_t size) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - unsigned long rt_runtime; - int rc; - - sscanf(buf, "%ld", &rt_runtime); - - rc = sched_group_set_rt_runtime(up->tg, rt_runtime); - - return (rc ? rc : size); -} - -static struct kobj_attribute cpu_rt_runtime_attr = - __ATTR(cpu_rt_runtime, 0644, cpu_rt_runtime_show, cpu_rt_runtime_store); - -static ssize_t cpu_rt_period_show(struct kobject *kobj, - struct kobj_attribute *attr, - char *buf) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - - return sprintf(buf, "%lu\n", sched_group_rt_period(up->tg)); -} - -static ssize_t cpu_rt_period_store(struct kobject *kobj, - struct kobj_attribute *attr, - const char *buf, size_t size) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - unsigned long rt_period; - int rc; - - sscanf(buf, "%lu", &rt_period); - - rc = sched_group_set_rt_period(up->tg, rt_period); - - return (rc ? rc : size); -} - -static struct kobj_attribute cpu_rt_period_attr = - __ATTR(cpu_rt_period, 0644, cpu_rt_period_show, cpu_rt_period_store); -#endif - -/* default attributes per uid directory */ -static struct attribute *uids_attributes[] = { -#ifdef CONFIG_FAIR_GROUP_SCHED - &cpu_share_attr.attr, -#endif -#ifdef CONFIG_RT_GROUP_SCHED - &cpu_rt_runtime_attr.attr, - &cpu_rt_period_attr.attr, -#endif - NULL -}; - -/* the lifetime of user_struct is not managed by the core (now) */ -static void uids_release(struct kobject *kobj) -{ - return; -} - -static struct kobj_type uids_ktype = { - .sysfs_ops = &kobj_sysfs_ops, - .default_attrs = uids_attributes, - .release = uids_release, -}; - -/* - * Create /sys/kernel/uids//cpu_share file for this user - * We do not create this file for users in a user namespace (until - * sysfs tagging is implemented). - * - * See Documentation/scheduler/sched-design-CFS.txt for ramifications. - */ -static int uids_user_create(struct user_struct *up) -{ - struct kobject *kobj = &up->kobj; - int error; - - memset(kobj, 0, sizeof(struct kobject)); - if (up->user_ns != &init_user_ns) - return 0; - kobj->kset = uids_kset; - error = kobject_init_and_add(kobj, &uids_ktype, NULL, "%d", up->uid); - if (error) { - kobject_put(kobj); - goto done; - } - - kobject_uevent(kobj, KOBJ_ADD); -done: - return error; -} - -/* create these entries in sysfs: - * "/sys/kernel/uids" directory - * "/sys/kernel/uids/0" directory (for root user) - * "/sys/kernel/uids/0/cpu_share" file (for root user) - */ -int __init uids_sysfs_init(void) -{ - uids_kset = kset_create_and_add("uids", NULL, kernel_kobj); - if (!uids_kset) - return -ENOMEM; - - return uids_user_create(&root_user); -} - -/* delayed work function to remove sysfs directory for a user and free up - * corresponding structures. - */ -static void cleanup_user_struct(struct work_struct *w) -{ - struct user_struct *up = container_of(w, struct user_struct, work.work); - unsigned long flags; - int remove_user = 0; - - /* Make uid_hash_remove() + sysfs_remove_file() + kobject_del() - * atomic. - */ - uids_mutex_lock(); - - spin_lock_irqsave(&uidhash_lock, flags); - if (atomic_read(&up->__count) == 0) { - uid_hash_remove(up); - remove_user = 1; - } - spin_unlock_irqrestore(&uidhash_lock, flags); - - if (!remove_user) - goto done; - - if (up->user_ns == &init_user_ns) { - kobject_uevent(&up->kobj, KOBJ_REMOVE); - kobject_del(&up->kobj); - kobject_put(&up->kobj); - } - - sched_destroy_user(up); - key_put(up->uid_keyring); - key_put(up->session_keyring); - kmem_cache_free(uid_cachep, up); - -done: - uids_mutex_unlock(); -} - -/* IRQs are disabled and uidhash_lock is held upon function entry. - * IRQ state (as stored in flags) is restored and uidhash_lock released - * upon function exit. - */ -static void free_user(struct user_struct *up, unsigned long flags) -{ - INIT_DELAYED_WORK(&up->work, cleanup_user_struct); - schedule_delayed_work(&up->work, msecs_to_jiffies(1000)); - spin_unlock_irqrestore(&uidhash_lock, flags); -} - -#else /* CONFIG_USER_SCHED && CONFIG_SYSFS */ - static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) { struct user_struct *user; @@ -352,11 +87,6 @@ static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) return NULL; } -int uids_sysfs_init(void) { return 0; } -static inline int uids_user_create(struct user_struct *up) { return 0; } -static inline void uids_mutex_lock(void) { } -static inline void uids_mutex_unlock(void) { } - /* IRQs are disabled and uidhash_lock is held upon function entry. * IRQ state (as stored in flags) is restored and uidhash_lock released * upon function exit. @@ -365,32 +95,11 @@ static void free_user(struct user_struct *up, unsigned long flags) { uid_hash_remove(up); spin_unlock_irqrestore(&uidhash_lock, flags); - sched_destroy_user(up); key_put(up->uid_keyring); key_put(up->session_keyring); kmem_cache_free(uid_cachep, up); } -#endif - -#if defined(CONFIG_RT_GROUP_SCHED) && defined(CONFIG_USER_SCHED) -/* - * We need to check if a setuid can take place. This function should be called - * before successfully completing the setuid. - */ -int task_can_switch_user(struct user_struct *up, struct task_struct *tsk) -{ - - return sched_rt_can_attach(up->tg, tsk); - -} -#else -int task_can_switch_user(struct user_struct *up, struct task_struct *tsk) -{ - return 1; -} -#endif - /* * Locate the user_struct for the passed UID. If found, take a ref on it. The * caller must undo that ref with free_uid(). @@ -431,8 +140,6 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid) /* Make uid_hash_find() + uids_user_create() + uid_hash_insert() * atomic. */ - uids_mutex_lock(); - spin_lock_irq(&uidhash_lock); up = uid_hash_find(uid, hashent); spin_unlock_irq(&uidhash_lock); @@ -445,14 +152,8 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid) new->uid = uid; atomic_set(&new->__count, 1); - if (sched_create_user(new) < 0) - goto out_free_user; - new->user_ns = get_user_ns(ns); - if (uids_user_create(new)) - goto out_destoy_sched; - /* * Before adding this, check whether we raced * on adding the same user already.. @@ -475,17 +176,11 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid) spin_unlock_irq(&uidhash_lock); } - uids_mutex_unlock(); - return up; -out_destoy_sched: - sched_destroy_user(new); put_user_ns(new->user_ns); -out_free_user: kmem_cache_free(uid_cachep, new); out_unlock: - uids_mutex_unlock(); return NULL; } -- cgit v0.10.2 From c587b6fa05106606053fc5e8e344f07cd34ace23 Mon Sep 17 00:00:00 2001 From: Feng Tang Date: Thu, 21 Jan 2010 10:41:10 +0800 Subject: spi/dw_spi: add a FIFO depth detection FIFO depth is configurable for each implementation of DW core, so add a depth detection for those interface drivers who don't set the fifo_len explicitly Signed-off-by: Feng Tang Acked-by: Jean-Hugues Deschenes Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi.c b/drivers/spi/dw_spi.c index f713af8..d948ef4 100644 --- a/drivers/spi/dw_spi.c +++ b/drivers/spi/dw_spi.c @@ -831,6 +831,22 @@ static void spi_hw_init(struct dw_spi *dws) spi_mask_intr(dws, 0xff); spi_enable_chip(dws, 1); flush(dws); + + /* + * Try to detect the FIFO depth if not set by interface driver, + * the depth could be from 2 to 256 from HW spec + */ + if (!dws->fifo_len) { + u32 fifo; + for (fifo = 2; fifo <= 257; fifo++) { + dw_writew(dws, txfltr, fifo); + if (fifo != dw_readw(dws, txfltr)) + break; + } + + dws->fifo_len = (fifo == 257) ? 0 : fifo; + dw_writew(dws, txfltr, 0); + } } int __devinit dw_spi_add_host(struct dw_spi *dws) -- cgit v0.10.2 From 83fe518a839e317480e50a138ef4acd73510d7ce Mon Sep 17 00:00:00 2001 From: George Shore Date: Thu, 21 Jan 2010 11:40:48 +0000 Subject: spi/dw_spi: enable platform specific chipselect. The driver core allows for a platform-specific chipselect assert/deassert function, however the chipselect function in the core doesn't take advantage of this fact. This enables the use of a custom function, should it be defined. Signed-off-by: George Shore Signed-off-by: Grant Likely diff --git a/include/linux/spi/dw_spi.h b/include/linux/spi/dw_spi.h index 1a127a3..cc813f9 100644 --- a/include/linux/spi/dw_spi.h +++ b/include/linux/spi/dw_spi.h @@ -172,6 +172,10 @@ static inline void spi_chip_sel(struct dw_spi *dws, u16 cs) { if (cs > dws->num_cs) return; + + if (dws->cs_control) + dws->cs_control(1); + dw_writel(dws, ser, 1 << cs); } -- cgit v0.10.2 From 20a588fcc862df79d8fcafbc41950e3ae93dea09 Mon Sep 17 00:00:00 2001 From: George Shore Date: Thu, 21 Jan 2010 11:40:49 +0000 Subject: spi/dw_spi: add return value to empty mrst_spi_debugfs_init() As per the function signature. Signed-off-by: George Shore Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi.c b/drivers/spi/dw_spi.c index d948ef4..cf945a4 100644 --- a/drivers/spi/dw_spi.c +++ b/drivers/spi/dw_spi.c @@ -152,6 +152,7 @@ static void mrst_spi_debugfs_remove(struct dw_spi *dws) #else static inline int mrst_spi_debugfs_init(struct dw_spi *dws) { + return 0; } static inline void mrst_spi_debugfs_remove(struct dw_spi *dws) -- cgit v0.10.2 From 426c0093d8da4d7b6b0e62cda917b1bae26db4c2 Mon Sep 17 00:00:00 2001 From: George Shore Date: Thu, 21 Jan 2010 11:40:50 +0000 Subject: spi/dw_spi: fixed a spelling typo in a warning message. Signed-off-by: George Shore Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi.c b/drivers/spi/dw_spi.c index cf945a4..d0a080a 100644 --- a/drivers/spi/dw_spi.c +++ b/drivers/spi/dw_spi.c @@ -169,7 +169,7 @@ static void wait_till_not_busy(struct dw_spi *dws) return; } dev_err(&dws->master->dev, - "DW SPI: Stutus keeps busy for 1000us after a read/write!\n"); + "DW SPI: Status keeps busy for 1000us after a read/write!\n"); } static void flush(struct dw_spi *dws) -- cgit v0.10.2 From f4aec798ae5a837a1f062e295f9a5f1b00962589 Mon Sep 17 00:00:00 2001 From: George Shore Date: Thu, 21 Jan 2010 11:40:51 +0000 Subject: spi/dw_spi: remove conditional from 'poll_transfer'. The 'poll_transfer' function employs a conditional to test whether the transmit buffer is valid; in doing so, on a receive operation no data is clocked out, thus no data is clocked in and ultimately errors appear. This removes the conditional as the transmit function will be set to a null writer when the transmit buffer is invalid, allowing the driver to clock 0x00 out to the device to receive data from the device. Signed-off-by: George Shore Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi.c b/drivers/spi/dw_spi.c index d0a080a..3853df5 100644 --- a/drivers/spi/dw_spi.c +++ b/drivers/spi/dw_spi.c @@ -408,12 +408,9 @@ static irqreturn_t dw_spi_irq(int irq, void *dev_id) /* Must be called inside pump_transfers() */ static void poll_transfer(struct dw_spi *dws) { - if (dws->tx) { - while (dws->write(dws)) - dws->read(dws); - } + while (dws->write(dws)) + dws->read(dws); - dws->read(dws); transfer_complete(dws); } -- cgit v0.10.2 From 052dc7c45d8f685fb3720a08331ba3e91e87937e Mon Sep 17 00:00:00 2001 From: George Shore Date: Thu, 21 Jan 2010 11:40:52 +0000 Subject: spi/dw_spi: conditional transfer mode changes This allows the switching between transfer modes between 'transmit only', 'receive only' and 'transmit and receive' modes. Due to the design of the SPI block, changing transfer modes requires that the block be disabled; in doing so the chipselect line is inherently deasserted and (usually) the attached device discards its state. Consequentially, switching modes requires that a platform-specific chipselect function has been defined so that the chipselect is not dropped during the change. Signed-off-by: George Shore Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi.c b/drivers/spi/dw_spi.c index 3853df5..e434320 100644 --- a/drivers/spi/dw_spi.c +++ b/drivers/spi/dw_spi.c @@ -537,6 +537,22 @@ static void pump_transfers(unsigned long data) } message->state = RUNNING_STATE; + /* + * Adjust transfer mode if necessary. Requires platform dependent + * chipselect mechanism. + */ + if (dws->cs_control) { + if (dws->rx && dws->tx) + chip->tmode = 0x00; + else if (dws->rx) + chip->tmode = 0x02; + else + chip->tmode = 0x01; + + cr0 &= ~(0x3 << SPI_MODE_OFFSET); + cr0 |= (chip->tmode << SPI_TMOD_OFFSET); + } + /* Check if current transfer is a DMA transaction */ dws->dma_mapped = map_dma_buffers(dws); -- cgit v0.10.2 From 8bcb4a88c5834c6a0fc7140edea32186664fe360 Mon Sep 17 00:00:00 2001 From: Feng Tang Date: Thu, 21 Jan 2010 07:25:38 -0700 Subject: spi/dw_spi: fix missing export of dw_spi_remove_host So that interface drivers could be built as modules Signed-off-by: Feng Tang Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi.c b/drivers/spi/dw_spi.c index e434320..8ed38f1 100644 --- a/drivers/spi/dw_spi.c +++ b/drivers/spi/dw_spi.c @@ -957,6 +957,7 @@ void __devexit dw_spi_remove_host(struct dw_spi *dws) /* Disconnect from the SPI framework */ spi_unregister_master(dws->master); } +EXPORT_SYMBOL(dw_spi_remove_host); int dw_spi_suspend_host(struct dw_spi *dws) { -- cgit v0.10.2 From f7b6fd6d1d3833529f1626c761ba7e338586d35e Mon Sep 17 00:00:00 2001 From: Jean-Hugues Deschenes Date: Thu, 21 Jan 2010 07:46:42 -0700 Subject: Memory-mapped dw_spi driver Adds a memory-mapped I/O dw_spi platform device. Signed-off-by: Jean-Hugues Deschenes Signed-off-by: Grant Likely diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig index 28becdd..201746a 100644 --- a/drivers/spi/Kconfig +++ b/drivers/spi/Kconfig @@ -334,6 +334,10 @@ config SPI_DW_PCI tristate "PCI interface driver for DW SPI core" depends on SPI_DESIGNWARE && PCI +config SPI_DW_MMIO + tristate "Memory-mapped io interface driver for DW SPI core" + depends on SPI_DESIGNWARE + # # There are lots of SPI device types, with sensors and memory # being probably the most widely used ones. diff --git a/drivers/spi/Makefile b/drivers/spi/Makefile index d3a65ab..d7d0f89 100644 --- a/drivers/spi/Makefile +++ b/drivers/spi/Makefile @@ -20,6 +20,7 @@ obj-$(CONFIG_SPI_COLDFIRE_QSPI) += coldfire_qspi.o obj-$(CONFIG_SPI_DAVINCI) += davinci_spi.o obj-$(CONFIG_SPI_DESIGNWARE) += dw_spi.o obj-$(CONFIG_SPI_DW_PCI) += dw_spi_pci.o +obj-$(CONFIG_SPI_DW_MMIO) += dw_spi_mmio.o obj-$(CONFIG_SPI_GPIO) += spi_gpio.o obj-$(CONFIG_SPI_IMX) += spi_imx.o obj-$(CONFIG_SPI_LM70_LLP) += spi_lm70llp.o diff --git a/drivers/spi/dw_spi_mmio.c b/drivers/spi/dw_spi_mmio.c new file mode 100644 index 0000000..26c4b49 --- /dev/null +++ b/drivers/spi/dw_spi_mmio.c @@ -0,0 +1,148 @@ +/* + * dw_spi_mmio.c - Memory-mapped interface driver for DW SPI Core + * + * Copyright (c) 2010, Octasic semiconductor. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + */ + +#include +#include +#include +#include +#include + +#define DRIVER_NAME "dw_spi_mmio" + +struct dw_spi_mmio { + struct dw_spi dws; + struct clk *clk; +}; + +static int __devinit dw_spi_mmio_probe(struct platform_device *pdev) +{ + struct dw_spi_mmio *dwsmmio; + struct dw_spi *dws; + struct resource *mem, *ioarea; + int ret; + + dwsmmio = kzalloc(sizeof(struct dw_spi_mmio), GFP_KERNEL); + if (!dwsmmio) { + ret = -ENOMEM; + goto err_end; + } + + dws = &dwsmmio->dws; + + /* Get basic io resource and map it */ + mem = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!mem) { + dev_err(&pdev->dev, "no mem resource?\n"); + ret = -EINVAL; + goto err_kfree; + } + + ioarea = request_mem_region(mem->start, resource_size(mem), + pdev->name); + if (!ioarea) { + dev_err(&pdev->dev, "SPI region already claimed\n"); + ret = -EBUSY; + goto err_kfree; + } + + dws->regs = ioremap_nocache(mem->start, resource_size(mem)); + if (!dws->regs) { + dev_err(&pdev->dev, "SPI region already mapped\n"); + ret = -ENOMEM; + goto err_release_reg; + } + + dws->irq = platform_get_irq(pdev, 0); + if (dws->irq < 0) { + dev_err(&pdev->dev, "no irq resource?\n"); + ret = dws->irq; /* -ENXIO */ + goto err_unmap; + } + + dwsmmio->clk = clk_get(&pdev->dev, NULL); + if (!dwsmmio->clk) { + ret = -ENODEV; + goto err_irq; + } + clk_enable(dwsmmio->clk); + + dws->parent_dev = &pdev->dev; + dws->bus_num = 0; + dws->num_cs = 4; + dws->max_freq = clk_get_rate(dwsmmio->clk); + + ret = dw_spi_add_host(dws); + if (ret) + goto err_clk; + + platform_set_drvdata(pdev, dwsmmio); + return 0; + +err_clk: + clk_disable(dwsmmio->clk); + clk_put(dwsmmio->clk); + dwsmmio->clk = NULL; +err_irq: + free_irq(dws->irq, dws); +err_unmap: + iounmap(dws->regs); +err_release_reg: + release_mem_region(mem->start, resource_size(mem)); +err_kfree: + kfree(dwsmmio); +err_end: + return ret; +} + +static int __devexit dw_spi_mmio_remove(struct platform_device *pdev) +{ + struct dw_spi_mmio *dwsmmio = platform_get_drvdata(pdev); + struct resource *mem; + + platform_set_drvdata(pdev, NULL); + + clk_disable(dwsmmio->clk); + clk_put(dwsmmio->clk); + dwsmmio->clk = NULL; + + free_irq(dwsmmio->dws.irq, &dwsmmio->dws); + dw_spi_remove_host(&dwsmmio->dws); + iounmap(dwsmmio->dws.regs); + kfree(dwsmmio); + + mem = platform_get_resource(pdev, IORESOURCE_MEM, 0); + release_mem_region(mem->start, resource_size(mem)); + return 0; +} + +static struct platform_driver dw_spi_mmio_driver = { + .remove = __devexit_p(dw_spi_mmio_remove), + .driver = { + .name = DRIVER_NAME, + .owner = THIS_MODULE, + }, +}; + +static int __init dw_spi_mmio_init(void) +{ + return platform_driver_probe(&dw_spi_mmio_driver, dw_spi_mmio_probe); +} + +static void __exit dw_spi_mmio_exit(void) +{ + platform_driver_unregister(&dw_spi_mmio_driver); +} + +module_init(dw_spi_mmio_init); +module_exit(dw_spi_mmio_exit); + +MODULE_AUTHOR("Jean-Hugues Deschenes "); +MODULE_DESCRIPTION("Memory-mapped I/O interface driver for DW SPI Core"); +MODULE_LICENSE("GPL v2"); -- cgit v0.10.2 From 0a4c1d7d446d3ed6179f907541d180e49b56d4f4 Mon Sep 17 00:00:00 2001 From: Jean-Hugues Deschenes Date: Thu, 21 Jan 2010 09:55:42 -0700 Subject: spi/dw_spi: mmio code style fixups Minor code style cleanups following comments by Wolfram Sang Signed-off-by: Jean-Hugues Deschenes Signed-off-by: Grant Likely diff --git a/drivers/spi/dw_spi_mmio.c b/drivers/spi/dw_spi_mmio.c index 26c4b49..e35b45a 100644 --- a/drivers/spi/dw_spi_mmio.c +++ b/drivers/spi/dw_spi_mmio.c @@ -17,8 +17,8 @@ #define DRIVER_NAME "dw_spi_mmio" struct dw_spi_mmio { - struct dw_spi dws; - struct clk *clk; + struct dw_spi dws; + struct clk *clk; }; static int __devinit dw_spi_mmio_probe(struct platform_device *pdev) @@ -134,13 +134,12 @@ static int __init dw_spi_mmio_init(void) { return platform_driver_probe(&dw_spi_mmio_driver, dw_spi_mmio_probe); } +module_init(dw_spi_mmio_init); static void __exit dw_spi_mmio_exit(void) { platform_driver_unregister(&dw_spi_mmio_driver); } - -module_init(dw_spi_mmio_init); module_exit(dw_spi_mmio_exit); MODULE_AUTHOR("Jean-Hugues Deschenes "); -- cgit v0.10.2 From 8ca8d15ade201b7723fa386eadcce2044463ff56 Mon Sep 17 00:00:00 2001 From: Jean-Hugues Deschenes Date: Thu, 21 Jan 2010 09:55:54 -0700 Subject: spi/dw_spi: Allow dw_spi.c to be a module Signed-off-by: Jean-Hugues Deschenes Signed-off-by: Grant Likely diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig index 201746a..1d1e396 100644 --- a/drivers/spi/Kconfig +++ b/drivers/spi/Kconfig @@ -325,7 +325,7 @@ config SPI_NUC900 # config SPI_DESIGNWARE - bool "DesignWare SPI controller core support" + tristate "DesignWare SPI controller core support" depends on SPI_MASTER help general driver for SPI controller core from DesignWare -- cgit v0.10.2 From 512dd1abd9539a474f2792eeaf6783c59ad7778a Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 13 Jan 2010 22:05:48 +0000 Subject: xfs: kill XFS_QMOPT_ASYNC The option is unused and one of the few remaining users of xfs_bawrite, so let's get rid of it. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder diff --git a/fs/xfs/quota/xfs_dquot.c b/fs/xfs/quota/xfs_dquot.c index d7c7eea..a447493 100644 --- a/fs/xfs/quota/xfs_dquot.c +++ b/fs/xfs/quota/xfs_dquot.c @@ -1253,8 +1253,6 @@ xfs_qm_dqflush( if (flags & XFS_QMOPT_DELWRI) { xfs_bdwrite(mp, bp); - } else if (flags & XFS_QMOPT_ASYNC) { - error = xfs_bawrite(mp, bp); } else { error = xfs_bwrite(mp, bp); } diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h index 91bfd60..21d11d9 100644 --- a/fs/xfs/xfs_quota.h +++ b/fs/xfs/xfs_quota.h @@ -226,7 +226,6 @@ typedef struct xfs_qoff_logformat { * flags for dqflush and dqflush_all. */ #define XFS_QMOPT_SYNC 0x1000000 -#define XFS_QMOPT_ASYNC 0x2000000 #define XFS_QMOPT_DELWRI 0x4000000 /* -- cgit v0.10.2 From 4d1f88d75b00c4d23f4c51305ab5b779a86ef74e Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 13 Jan 2010 22:05:49 +0000 Subject: xfs: clean up error handling in xfs_trans_dqresv Move the error code selection after the goto label and fold the xfs_quota_error helper into it. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder diff --git a/fs/xfs/quota/xfs_trans_dquot.c b/fs/xfs/quota/xfs_trans_dquot.c index 97ac964..b9db6f7 100644 --- a/fs/xfs/quota/xfs_trans_dquot.c +++ b/fs/xfs/quota/xfs_trans_dquot.c @@ -589,14 +589,6 @@ xfs_trans_unreserve_and_mod_dquots( } } -STATIC int -xfs_quota_error(uint flags) -{ - if (flags & XFS_QMOPT_ENOSPC) - return ENOSPC; - return EDQUOT; -} - /* * This reserves disk blocks and inodes against a dquot. * Flags indicate if the dquot is to be locked here and also @@ -612,7 +604,6 @@ xfs_trans_dqresv( long ninos, uint flags) { - int error; xfs_qcnt_t hardlimit; xfs_qcnt_t softlimit; time_t timer; @@ -649,7 +640,6 @@ xfs_trans_dqresv( warnlimit = XFS_QI_RTBWARNLIMIT(dqp->q_mount); resbcountp = &dqp->q_res_rtbcount; } - error = 0; if ((flags & XFS_QMOPT_FORCE_RES) == 0 && dqp->q_core.d_id && @@ -667,19 +657,13 @@ xfs_trans_dqresv( * nblks. */ if (hardlimit > 0ULL && - (hardlimit <= nblks + *resbcountp)) { - error = xfs_quota_error(flags); + hardlimit <= nblks + *resbcountp) goto error_return; - } - if (softlimit > 0ULL && - (softlimit <= nblks + *resbcountp)) { - if ((timer != 0 && get_seconds() > timer) || - (warns != 0 && warns >= warnlimit)) { - error = xfs_quota_error(flags); - goto error_return; - } - } + softlimit <= nblks + *resbcountp && + ((timer != 0 && get_seconds() > timer) || + (warns != 0 && warns >= warnlimit))) + goto error_return; } if (ninos > 0) { count = be64_to_cpu(dqp->q_core.d_icount); @@ -692,16 +676,13 @@ xfs_trans_dqresv( softlimit = be64_to_cpu(dqp->q_core.d_ino_softlimit); if (!softlimit) softlimit = q->qi_isoftlimit; - if (hardlimit > 0ULL && count >= hardlimit) { - error = xfs_quota_error(flags); + + if (hardlimit > 0ULL && count >= hardlimit) + goto error_return; + if (softlimit > 0ULL && count >= softlimit && + ((timer != 0 && get_seconds() > timer) || + (warns != 0 && warns >= warnlimit))) goto error_return; - } else if (softlimit > 0ULL && count >= softlimit) { - if ((timer != 0 && get_seconds() > timer) || - (warns != 0 && warns >= warnlimit)) { - error = xfs_quota_error(flags); - goto error_return; - } - } } } @@ -736,9 +717,14 @@ xfs_trans_dqresv( ASSERT(dqp->q_res_rtbcount >= be64_to_cpu(dqp->q_core.d_rtbcount)); ASSERT(dqp->q_res_icount >= be64_to_cpu(dqp->q_core.d_icount)); + xfs_dqunlock(dqp); + return 0; + error_return: xfs_dqunlock(dqp); - return error; + if (flags & XFS_QMOPT_ENOSPC) + return ENOSPC; + return EDQUOT; } -- cgit v0.10.2 From a210c1aa7f6c90b729cc3a72d03e789b13cb6c47 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 17 Jan 2010 22:36:19 +0000 Subject: xfs: implement quota warnings via netlink Wire up quota_send_warning to send quota warnings over netlink. This is used by various desktops to show user quota warnings. Tested by running the quota_nld daemon while running the xfstest quota tests and observing the warnings. I'll see how I can get a more formal testcase for it written. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder diff --git a/fs/xfs/quota/xfs_trans_dquot.c b/fs/xfs/quota/xfs_trans_dquot.c index b9db6f7..c3ab75c 100644 --- a/fs/xfs/quota/xfs_trans_dquot.c +++ b/fs/xfs/quota/xfs_trans_dquot.c @@ -589,6 +589,20 @@ xfs_trans_unreserve_and_mod_dquots( } } +STATIC void +xfs_quota_warn( + struct xfs_mount *mp, + struct xfs_dquot *dqp, + int type) +{ + /* no warnings for project quotas - we just return ENOSPC later */ + if (dqp->dq_flags & XFS_DQ_PROJ) + return; + quota_send_warning((dqp->dq_flags & XFS_DQ_USER) ? USRQUOTA : GRPQUOTA, + be32_to_cpu(dqp->q_core.d_id), mp->m_super->s_dev, + type); +} + /* * This reserves disk blocks and inodes against a dquot. * Flags indicate if the dquot is to be locked here and also @@ -657,13 +671,21 @@ xfs_trans_dqresv( * nblks. */ if (hardlimit > 0ULL && - hardlimit <= nblks + *resbcountp) + hardlimit <= nblks + *resbcountp) { + xfs_quota_warn(mp, dqp, QUOTA_NL_BHARDWARN); goto error_return; + } if (softlimit > 0ULL && - softlimit <= nblks + *resbcountp && - ((timer != 0 && get_seconds() > timer) || - (warns != 0 && warns >= warnlimit))) - goto error_return; + softlimit <= nblks + *resbcountp) { + if ((timer != 0 && get_seconds() > timer) || + (warns != 0 && warns >= warnlimit)) { + xfs_quota_warn(mp, dqp, + QUOTA_NL_BSOFTLONGWARN); + goto error_return; + } + + xfs_quota_warn(mp, dqp, QUOTA_NL_BSOFTWARN); + } } if (ninos > 0) { count = be64_to_cpu(dqp->q_core.d_icount); @@ -677,12 +699,19 @@ xfs_trans_dqresv( if (!softlimit) softlimit = q->qi_isoftlimit; - if (hardlimit > 0ULL && count >= hardlimit) - goto error_return; - if (softlimit > 0ULL && count >= softlimit && - ((timer != 0 && get_seconds() > timer) || - (warns != 0 && warns >= warnlimit))) + if (hardlimit > 0ULL && count >= hardlimit) { + xfs_quota_warn(mp, dqp, QUOTA_NL_IHARDWARN); goto error_return; + } + if (softlimit > 0ULL && count >= softlimit) { + if ((timer != 0 && get_seconds() > timer) || + (warns != 0 && warns >= warnlimit)) { + xfs_quota_warn(mp, dqp, + QUOTA_NL_ISOFTLONGWARN); + goto error_return; + } + xfs_quota_warn(mp, dqp, QUOTA_NL_ISOFTWARN); + } } } -- cgit v0.10.2 From 0cadda1c5f194f98a05d252ff4385d86d2ed0862 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 19 Jan 2010 09:56:44 +0000 Subject: xfs: remove duplicate buffer flags Currently we define aliases for the buffer flags in various namespaces, which only adds confusion. Remove all but the XBF_ flags to clean this up a bit. Note that we still abuse XFS_B_ASYNC/XBF_ASYNC for some non-buffer uses, but I'll clean that up later. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index efd745b..730eff1 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -1169,7 +1169,7 @@ xfs_bioerror_relse( XFS_BUF_STALE(bp); XFS_BUF_CLR_IODONE_FUNC(bp); XFS_BUF_CLR_BDSTRAT_FUNC(bp); - if (!(fl & XFS_B_ASYNC)) { + if (!(fl & XBF_ASYNC)) { /* * Mark b_error and B_ERROR _both_. * Lot's of chunkcache code assumes that. diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h index 4f2ad66..ea8c198 100644 --- a/fs/xfs/linux-2.6/xfs_buf.h +++ b/fs/xfs/linux-2.6/xfs_buf.h @@ -275,33 +275,19 @@ extern void xfs_buf_terminate(void); ({ char __b[BDEVNAME_SIZE]; bdevname((target)->bt_bdev, __b); __b; }) -#define XFS_B_ASYNC XBF_ASYNC -#define XFS_B_DELWRI XBF_DELWRI -#define XFS_B_READ XBF_READ -#define XFS_B_WRITE XBF_WRITE -#define XFS_B_STALE XBF_STALE - -#define XFS_BUF_TRYLOCK XBF_TRYLOCK -#define XFS_INCORE_TRYLOCK XBF_TRYLOCK -#define XFS_BUF_LOCK XBF_LOCK -#define XFS_BUF_MAPPED XBF_MAPPED - -#define BUF_BUSY XBF_DONT_BLOCK - #define XFS_BUF_BFLAGS(bp) ((bp)->b_flags) #define XFS_BUF_ZEROFLAGS(bp) ((bp)->b_flags &= \ ~(XBF_READ|XBF_WRITE|XBF_ASYNC|XBF_DELWRI|XBF_ORDERED)) -#define XFS_BUF_STALE(bp) ((bp)->b_flags |= XFS_B_STALE) -#define XFS_BUF_UNSTALE(bp) ((bp)->b_flags &= ~XFS_B_STALE) -#define XFS_BUF_ISSTALE(bp) ((bp)->b_flags & XFS_B_STALE) +#define XFS_BUF_STALE(bp) ((bp)->b_flags |= XBF_STALE) +#define XFS_BUF_UNSTALE(bp) ((bp)->b_flags &= ~XBF_STALE) +#define XFS_BUF_ISSTALE(bp) ((bp)->b_flags & XBF_STALE) #define XFS_BUF_SUPER_STALE(bp) do { \ XFS_BUF_STALE(bp); \ xfs_buf_delwri_dequeue(bp); \ XFS_BUF_DONE(bp); \ } while (0) -#define XFS_BUF_MANAGE XBF_FS_MANAGED #define XFS_BUF_UNMANAGE(bp) ((bp)->b_flags &= ~XBF_FS_MANAGED) #define XFS_BUF_DELAYWRITE(bp) ((bp)->b_flags |= XBF_DELWRI) @@ -390,7 +376,7 @@ static inline void xfs_buf_relse(xfs_buf_t *bp) #define xfs_biomove(bp, off, len, data, rw) \ xfs_buf_iomove((bp), (off), (len), (data), \ - ((rw) == XFS_B_WRITE) ? XBRW_WRITE : XBRW_READ) + ((rw) == XBF_WRITE) ? XBRW_WRITE : XBRW_READ) #define xfs_biozero(bp, off, len) \ xfs_buf_iomove((bp), (off), (len), NULL, XBRW_ZERO) diff --git a/fs/xfs/linux-2.6/xfs_fs_subr.c b/fs/xfs/linux-2.6/xfs_fs_subr.c index 7501b85..b6918d7 100644 --- a/fs/xfs/linux-2.6/xfs_fs_subr.c +++ b/fs/xfs/linux-2.6/xfs_fs_subr.c @@ -79,7 +79,7 @@ xfs_flush_pages( xfs_iflags_clear(ip, XFS_ITRUNCATED); ret = -filemap_fdatawrite(mapping); } - if (flags & XFS_B_ASYNC) + if (flags & XBF_ASYNC) return ret; ret2 = xfs_wait_on_pages(ip, first, last); if (!ret) diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c index b58f841..58c24be 100644 --- a/fs/xfs/linux-2.6/xfs_sync.c +++ b/fs/xfs/linux-2.6/xfs_sync.c @@ -234,7 +234,7 @@ xfs_sync_inode_data( } error = xfs_flush_pages(ip, 0, -1, (flags & SYNC_WAIT) ? - 0 : XFS_B_ASYNC, FI_NONE); + 0 : XBF_ASYNC, FI_NONE); xfs_iunlock(ip, XFS_IOLOCK_SHARED); out_wait: @@ -370,7 +370,7 @@ xfs_sync_fsdata( if (flags & SYNC_TRYLOCK) { ASSERT(!(flags & SYNC_WAIT)); - bp = xfs_getsb(mp, XFS_BUF_TRYLOCK); + bp = xfs_getsb(mp, XBF_TRYLOCK); if (!bp) goto out; diff --git a/fs/xfs/quota/xfs_dquot.c b/fs/xfs/quota/xfs_dquot.c index a447493..5756392 100644 --- a/fs/xfs/quota/xfs_dquot.c +++ b/fs/xfs/quota/xfs_dquot.c @@ -1527,8 +1527,7 @@ xfs_qm_dqflock_pushbuf_wait( * the flush lock when the I/O completes. */ bp = xfs_incore(dqp->q_mount->m_ddev_targp, dqp->q_blkno, - XFS_QI_DQCHUNKLEN(dqp->q_mount), - XFS_INCORE_TRYLOCK); + XFS_QI_DQCHUNKLEN(dqp->q_mount), XBF_TRYLOCK); if (bp != NULL) { if (XFS_BUF_ISDELAYWRITE(bp)) { int error; diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c index d0d4a9a..37929d1 100644 --- a/fs/xfs/quota/xfs_dquot_item.c +++ b/fs/xfs/quota/xfs_dquot_item.c @@ -237,8 +237,7 @@ xfs_qm_dquot_logitem_pushbuf( } mp = dqp->q_mount; bp = xfs_incore(mp->m_ddev_targp, qip->qli_format.qlf_blkno, - XFS_QI_DQCHUNKLEN(mp), - XFS_INCORE_TRYLOCK); + XFS_QI_DQCHUNKLEN(mp), XBF_TRYLOCK); if (bp != NULL) { if (XFS_BUF_ISDELAYWRITE(bp)) { dopush = ((qip->qli_item.li_flags & XFS_LI_IN_AIL) && diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index 8aa181d..a27aeb7 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -2180,7 +2180,7 @@ xfs_alloc_read_agf( ASSERT(agno != NULLAGNUMBER); error = xfs_read_agf(mp, tp, agno, - (flags & XFS_ALLOC_FLAG_TRYLOCK) ? XFS_BUF_TRYLOCK : 0, + (flags & XFS_ALLOC_FLAG_TRYLOCK) ? XBF_TRYLOCK : 0, bpp); if (error) return error; diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index f7b426a..b9c196a 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -2015,15 +2015,14 @@ xfs_attr_rmtval_get(xfs_da_args_t *args) dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock); blkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount); error = xfs_read_buf(mp, mp->m_ddev_targp, dblkno, - blkcnt, - XFS_BUF_LOCK | XBF_DONT_BLOCK, + blkcnt, XBF_LOCK | XBF_DONT_BLOCK, &bp); if (error) return(error); tmp = (valuelen < XFS_BUF_SIZE(bp)) ? valuelen : XFS_BUF_SIZE(bp); - xfs_biomove(bp, 0, tmp, dst, XFS_B_READ); + xfs_biomove(bp, 0, tmp, dst, XBF_READ); xfs_buf_relse(bp); dst += tmp; valuelen -= tmp; @@ -2149,13 +2148,13 @@ xfs_attr_rmtval_set(xfs_da_args_t *args) blkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount); bp = xfs_buf_get(mp->m_ddev_targp, dblkno, blkcnt, - XFS_BUF_LOCK | XBF_DONT_BLOCK); + XBF_LOCK | XBF_DONT_BLOCK); ASSERT(bp); ASSERT(!XFS_BUF_GETERROR(bp)); tmp = (valuelen < XFS_BUF_SIZE(bp)) ? valuelen : XFS_BUF_SIZE(bp); - xfs_biomove(bp, 0, tmp, src, XFS_B_WRITE); + xfs_biomove(bp, 0, tmp, src, XBF_WRITE); if (tmp < XFS_BUF_SIZE(bp)) xfs_biozero(bp, tmp, XFS_BUF_SIZE(bp) - tmp); if ((error = xfs_bwrite(mp, bp))) {/* GROT: NOTE: synchronous write */ @@ -2216,8 +2215,7 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args) /* * If the "remote" value is in the cache, remove it. */ - bp = xfs_incore(mp->m_ddev_targp, dblkno, blkcnt, - XFS_INCORE_TRYLOCK); + bp = xfs_incore(mp->m_ddev_targp, dblkno, blkcnt, XBF_TRYLOCK); if (bp) { XFS_BUF_STALE(bp); XFS_BUF_UNDELAYWRITE(bp); diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index 52519a2..a90ce74 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -2950,7 +2950,7 @@ xfs_attr_leaf_freextent(xfs_trans_t **trans, xfs_inode_t *dp, map.br_blockcount); bp = xfs_trans_get_buf(*trans, dp->i_mount->m_ddev_targp, - dblkno, dblkcnt, XFS_BUF_LOCK); + dblkno, dblkcnt, XBF_LOCK); xfs_trans_binval(*trans, bp); /* * Roll to next transaction. diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c index 36a0992..96be4b0 100644 --- a/fs/xfs/xfs_btree.c +++ b/fs/xfs/xfs_btree.c @@ -977,7 +977,7 @@ xfs_btree_get_buf_block( xfs_daddr_t d; /* need to sort out how callers deal with failures first */ - ASSERT(!(flags & XFS_BUF_TRYLOCK)); + ASSERT(!(flags & XBF_TRYLOCK)); d = xfs_btree_ptr_to_daddr(cur, ptr); *bpp = xfs_trans_get_buf(cur->bc_tp, mp->m_ddev_targp, d, @@ -1008,7 +1008,7 @@ xfs_btree_read_buf_block( int error; /* need to sort out how callers deal with failures first */ - ASSERT(!(flags & XFS_BUF_TRYLOCK)); + ASSERT(!(flags & XBF_TRYLOCK)); d = xfs_btree_ptr_to_daddr(cur, ptr); error = xfs_trans_read_buf(mp, cur->bc_tp, mp->m_ddev_targp, d, diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index 52c9d00..9d884c1 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -205,7 +205,7 @@ xfs_ialloc_inode_init( d = XFS_AGB_TO_DADDR(mp, agno, agbno + (j * blks_per_cluster)); fbuf = xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize * blks_per_cluster, - XFS_BUF_LOCK); + XBF_LOCK); ASSERT(fbuf); ASSERT(!XFS_BUF_GETERROR(fbuf)); diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 0317b00..bbb3bee 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -151,7 +151,7 @@ xfs_imap_to_bp( "an error %d on %s. Returning error.", error, mp->m_fsname); } else { - ASSERT(buf_flags & XFS_BUF_TRYLOCK); + ASSERT(buf_flags & XBF_TRYLOCK); } return error; } @@ -239,7 +239,7 @@ xfs_inotobp( if (error) return error; - error = xfs_imap_to_bp(mp, tp, &imap, &bp, XFS_BUF_LOCK, imap_flags); + error = xfs_imap_to_bp(mp, tp, &imap, &bp, XBF_LOCK, imap_flags); if (error) return error; @@ -285,7 +285,7 @@ xfs_itobp( return error; if (!bp) { - ASSERT(buf_flags & XFS_BUF_TRYLOCK); + ASSERT(buf_flags & XBF_TRYLOCK); ASSERT(tp == NULL); *bpp = NULL; return EAGAIN; @@ -807,7 +807,7 @@ xfs_iread( * Get pointers to the on-disk inode and the buffer containing it. */ error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &bp, - XFS_BUF_LOCK, iget_flags); + XBF_LOCK, iget_flags); if (error) return error; dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset); @@ -1751,7 +1751,7 @@ xfs_iunlink( * Here we put the head pointer into our next pointer, * and then we fall through to point the head at us. */ - error = xfs_itobp(mp, tp, ip, &dip, &ibp, XFS_BUF_LOCK); + error = xfs_itobp(mp, tp, ip, &dip, &ibp, XBF_LOCK); if (error) return error; @@ -1833,7 +1833,7 @@ xfs_iunlink_remove( * of dealing with the buffer when there is no need to * change it. */ - error = xfs_itobp(mp, tp, ip, &dip, &ibp, XFS_BUF_LOCK); + error = xfs_itobp(mp, tp, ip, &dip, &ibp, XBF_LOCK); if (error) { cmn_err(CE_WARN, "xfs_iunlink_remove: xfs_itobp() returned an error %d on %s. Returning error.", @@ -1895,7 +1895,7 @@ xfs_iunlink_remove( * Now last_ibp points to the buffer previous to us on * the unlinked list. Pull us from the list. */ - error = xfs_itobp(mp, tp, ip, &dip, &ibp, XFS_BUF_LOCK); + error = xfs_itobp(mp, tp, ip, &dip, &ibp, XBF_LOCK); if (error) { cmn_err(CE_WARN, "xfs_iunlink_remove: xfs_itobp() returned an error %d on %s. Returning error.", @@ -2040,7 +2040,7 @@ xfs_ifree_cluster( bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, blkno, mp->m_bsize * blks_per_cluster, - XFS_BUF_LOCK); + XBF_LOCK); pre_flushed = 0; lip = XFS_BUF_FSPRIVATE(bp, xfs_log_item_t *); @@ -2151,7 +2151,7 @@ xfs_ifree( xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - error = xfs_itobp(ip->i_mount, tp, ip, &dip, &ibp, XFS_BUF_LOCK); + error = xfs_itobp(ip->i_mount, tp, ip, &dip, &ibp, XBF_LOCK); if (error) return error; @@ -2952,7 +2952,7 @@ xfs_iflush( * Get the buffer containing the on-disk inode. */ error = xfs_itobp(mp, NULL, ip, &dip, &bp, - noblock ? XFS_BUF_TRYLOCK : XFS_BUF_LOCK); + noblock ? XBF_TRYLOCK : XBF_LOCK); if (error || !bp) { xfs_ifunlock(ip); return error; diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index f38855d..6194fb5 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -785,7 +785,7 @@ xfs_inode_item_pushbuf( mp = ip->i_mount; bp = xfs_incore(mp->m_ddev_targp, iip->ili_format.ilf_blkno, - iip->ili_format.ilf_len, XFS_INCORE_TRYLOCK); + iip->ili_format.ilf_len, XBF_TRYLOCK); if (bp != NULL) { if (XFS_BUF_ISDELAYWRITE(bp)) { diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 65f1f13..97148f0 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -2184,9 +2184,9 @@ xlog_recover_do_buffer_trans( } mp = log->l_mp; - buf_flags = XFS_BUF_LOCK; + buf_flags = XBF_LOCK; if (!(flags & XFS_BLI_INODE_BUF)) - buf_flags |= XFS_BUF_MAPPED; + buf_flags |= XBF_MAPPED; bp = xfs_buf_read(mp->m_ddev_targp, blkno, len, buf_flags); if (XFS_BUF_ISERROR(bp)) { @@ -2288,7 +2288,7 @@ xlog_recover_do_inode_trans( } bp = xfs_buf_read(mp->m_ddev_targp, in_f->ilf_blkno, in_f->ilf_len, - XFS_BUF_LOCK); + XBF_LOCK); if (XFS_BUF_ISERROR(bp)) { xfs_ioerror_alert("xlog_recover_do..(read#2)", mp, bp, in_f->ilf_blkno); @@ -3146,7 +3146,7 @@ xlog_recover_process_one_iunlink( /* * Get the on disk inode to find the next inode in the bucket. */ - error = xfs_itobp(mp, NULL, ip, &dip, &ibp, XFS_BUF_LOCK); + error = xfs_itobp(mp, NULL, ip, &dip, &ibp, XBF_LOCK); if (error) goto fail_iput; diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index d95bd18..bb01540 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -665,7 +665,7 @@ xfs_readsb(xfs_mount_t *mp, int flags) * access to the superblock. */ sector_size = xfs_getsize_buftarg(mp->m_ddev_targp); - extra_flags = XFS_BUF_LOCK | XFS_BUF_MANAGE | XFS_BUF_MAPPED; + extra_flags = XBF_LOCK | XBF_FS_MANAGED | XBF_MAPPED; bp = xfs_buf_read(mp->m_ddev_targp, XFS_SB_DADDR, BTOBB(sector_size), extra_flags); @@ -1969,7 +1969,7 @@ xfs_getsb( ASSERT(mp->m_sb_bp != NULL); bp = mp->m_sb_bp; - if (flags & XFS_BUF_TRYLOCK) { + if (flags & XBF_TRYLOCK) { if (!XFS_BUF_CPSEMA(bp)) { return NULL; } diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index 4913062..5ffd544 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -75,13 +75,14 @@ xfs_trans_get_buf(xfs_trans_t *tp, xfs_buf_log_item_t *bip; if (flags == 0) - flags = XFS_BUF_LOCK | XFS_BUF_MAPPED; + flags = XBF_LOCK | XBF_MAPPED; /* * Default to a normal get_buf() call if the tp is NULL. */ if (tp == NULL) - return xfs_buf_get(target_dev, blkno, len, flags | BUF_BUSY); + return xfs_buf_get(target_dev, blkno, len, + flags | XBF_DONT_BLOCK); /* * If we find the buffer in the cache with this transaction @@ -117,14 +118,14 @@ xfs_trans_get_buf(xfs_trans_t *tp, } /* - * We always specify the BUF_BUSY flag within a transaction so - * that get_buf does not try to push out a delayed write buffer + * We always specify the XBF_DONT_BLOCK flag within a transaction + * so that get_buf does not try to push out a delayed write buffer * which might cause another transaction to take place (if the * buffer was delayed alloc). Such recursive transactions can * easily deadlock with our current transaction as well as cause * us to run out of stack space. */ - bp = xfs_buf_get(target_dev, blkno, len, flags | BUF_BUSY); + bp = xfs_buf_get(target_dev, blkno, len, flags | XBF_DONT_BLOCK); if (bp == NULL) { return NULL; } @@ -290,15 +291,15 @@ xfs_trans_read_buf( int error; if (flags == 0) - flags = XFS_BUF_LOCK | XFS_BUF_MAPPED; + flags = XBF_LOCK | XBF_MAPPED; /* * Default to a normal get_buf() call if the tp is NULL. */ if (tp == NULL) { - bp = xfs_buf_read(target, blkno, len, flags | BUF_BUSY); + bp = xfs_buf_read(target, blkno, len, flags | XBF_DONT_BLOCK); if (!bp) - return (flags & XFS_BUF_TRYLOCK) ? + return (flags & XBF_TRYLOCK) ? EAGAIN : XFS_ERROR(ENOMEM); if (XFS_BUF_GETERROR(bp) != 0) { @@ -385,14 +386,14 @@ xfs_trans_read_buf( } /* - * We always specify the BUF_BUSY flag within a transaction so - * that get_buf does not try to push out a delayed write buffer + * We always specify the XBF_DONT_BLOCK flag within a transaction + * so that get_buf does not try to push out a delayed write buffer * which might cause another transaction to take place (if the * buffer was delayed alloc). Such recursive transactions can * easily deadlock with our current transaction as well as cause * us to run out of stack space. */ - bp = xfs_buf_read(target, blkno, len, flags | BUF_BUSY); + bp = xfs_buf_read(target, blkno, len, flags | XBF_DONT_BLOCK); if (bp == NULL) { *bpp = NULL; return 0; @@ -472,8 +473,8 @@ shutdown_abort: if (XFS_BUF_ISSTALE(bp) && XFS_BUF_ISDELAYWRITE(bp)) cmn_err(CE_NOTE, "about to pop assert, bp == 0x%p", bp); #endif - ASSERT((XFS_BUF_BFLAGS(bp) & (XFS_B_STALE|XFS_B_DELWRI)) != - (XFS_B_STALE|XFS_B_DELWRI)); + ASSERT((XFS_BUF_BFLAGS(bp) & (XBF_STALE|XBF_DELWRI)) != + (XBF_STALE|XBF_DELWRI)); trace_xfs_trans_read_buf_shut(bp, _RET_IP_); xfs_buf_relse(bp); diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 9f7c001..4da96cd 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -256,7 +256,7 @@ xfs_setattr( iattr->ia_size > ip->i_d.di_size) { code = xfs_flush_pages(ip, ip->i_d.di_size, iattr->ia_size, - XFS_B_ASYNC, FI_NONE); + XBF_ASYNC, FI_NONE); } /* wait for all I/O to complete */ @@ -1096,7 +1096,7 @@ xfs_release( */ truncated = xfs_iflags_test_and_clear(ip, XFS_ITRUNCATED); if (truncated && VN_DIRTY(VFS_I(ip)) && ip->i_delayed_blks > 0) - xfs_flush_pages(ip, 0, -1, XFS_B_ASYNC, FI_NONE); + xfs_flush_pages(ip, 0, -1, XBF_ASYNC, FI_NONE); } if (ip->i_d.di_nlink != 0) { -- cgit v0.10.2 From 4139b3b337cffd106744386c842b89dc86e31d4b Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 19 Jan 2010 09:56:45 +0000 Subject: xfs: kill XLOG_VEC_SET_TYPE This macro only obsfucates the log item type assignments, so kill it. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c index 37929d1..116580d 100644 --- a/fs/xfs/quota/xfs_dquot_item.c +++ b/fs/xfs/quota/xfs_dquot_item.c @@ -74,11 +74,11 @@ xfs_qm_dquot_logitem_format( logvec->i_addr = (xfs_caddr_t)&logitem->qli_format; logvec->i_len = sizeof(xfs_dq_logformat_t); - XLOG_VEC_SET_TYPE(logvec, XLOG_REG_TYPE_QFORMAT); + logvec->i_type = XLOG_REG_TYPE_QFORMAT; logvec++; logvec->i_addr = (xfs_caddr_t)&logitem->qli_dquot->q_core; logvec->i_len = sizeof(xfs_disk_dquot_t); - XLOG_VEC_SET_TYPE(logvec, XLOG_REG_TYPE_DQUOT); + logvec->i_type = XLOG_REG_TYPE_DQUOT; ASSERT(2 == logitem->qli_item.li_desc->lid_size); logitem->qli_format.qlf_size = 2; @@ -466,7 +466,7 @@ xfs_qm_qoff_logitem_format(xfs_qoff_logitem_t *qf, log_vector->i_addr = (xfs_caddr_t)&(qf->qql_format); log_vector->i_len = sizeof(xfs_qoff_logitem_t); - XLOG_VEC_SET_TYPE(log_vector, XLOG_REG_TYPE_QUOTAOFF); + log_vector->i_type = XLOG_REG_TYPE_QUOTAOFF; qf->qql_format.qf_size = 1; } diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index a30f7e9..e0a1158 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -250,7 +250,7 @@ xfs_buf_item_format( ((bip->bli_format.blf_map_size - 1) * sizeof(uint))); vecp->i_addr = (xfs_caddr_t)&bip->bli_format; vecp->i_len = base_size; - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_BFORMAT); + vecp->i_type = XLOG_REG_TYPE_BFORMAT; vecp++; nvecs = 1; @@ -297,14 +297,14 @@ xfs_buf_item_format( buffer_offset = first_bit * XFS_BLI_CHUNK; vecp->i_addr = xfs_buf_offset(bp, buffer_offset); vecp->i_len = nbits * XFS_BLI_CHUNK; - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_BCHUNK); + vecp->i_type = XLOG_REG_TYPE_BCHUNK; nvecs++; break; } else if (next_bit != last_bit + 1) { buffer_offset = first_bit * XFS_BLI_CHUNK; vecp->i_addr = xfs_buf_offset(bp, buffer_offset); vecp->i_len = nbits * XFS_BLI_CHUNK; - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_BCHUNK); + vecp->i_type = XLOG_REG_TYPE_BCHUNK; nvecs++; vecp++; first_bit = next_bit; @@ -316,7 +316,7 @@ xfs_buf_item_format( buffer_offset = first_bit * XFS_BLI_CHUNK; vecp->i_addr = xfs_buf_offset(bp, buffer_offset); vecp->i_len = nbits * XFS_BLI_CHUNK; - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_BCHUNK); + vecp->i_type = XLOG_REG_TYPE_BCHUNK; /* You would think we need to bump the nvecs here too, but we do not * this number is used by recovery, and it gets confused by the boundary * split here diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c index 05a4bdd..6f35ed1 100644 --- a/fs/xfs/xfs_extfree_item.c +++ b/fs/xfs/xfs_extfree_item.c @@ -82,7 +82,7 @@ xfs_efi_item_format(xfs_efi_log_item_t *efip, log_vector->i_addr = (xfs_caddr_t)&(efip->efi_format); log_vector->i_len = size; - XLOG_VEC_SET_TYPE(log_vector, XLOG_REG_TYPE_EFI_FORMAT); + log_vector->i_type = XLOG_REG_TYPE_EFI_FORMAT; ASSERT(size >= sizeof(xfs_efi_log_format_t)); } @@ -406,7 +406,7 @@ xfs_efd_item_format(xfs_efd_log_item_t *efdp, log_vector->i_addr = (xfs_caddr_t)&(efdp->efd_format); log_vector->i_len = size; - XLOG_VEC_SET_TYPE(log_vector, XLOG_REG_TYPE_EFD_FORMAT); + log_vector->i_type = XLOG_REG_TYPE_EFD_FORMAT; ASSERT(size >= sizeof(xfs_efd_log_format_t)); } diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index 6194fb5..da4cac6 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -228,7 +228,7 @@ xfs_inode_item_format( vecp->i_addr = (xfs_caddr_t)&iip->ili_format; vecp->i_len = sizeof(xfs_inode_log_format_t); - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_IFORMAT); + vecp->i_type = XLOG_REG_TYPE_IFORMAT; vecp++; nvecs = 1; @@ -279,7 +279,7 @@ xfs_inode_item_format( vecp->i_addr = (xfs_caddr_t)&ip->i_d; vecp->i_len = sizeof(struct xfs_icdinode); - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_ICORE); + vecp->i_type = XLOG_REG_TYPE_ICORE; vecp++; nvecs++; iip->ili_format.ilf_fields |= XFS_ILOG_CORE; @@ -336,7 +336,7 @@ xfs_inode_item_format( vecp->i_addr = (char *)(ip->i_df.if_u1.if_extents); vecp->i_len = ip->i_df.if_bytes; - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_IEXT); + vecp->i_type = XLOG_REG_TYPE_IEXT; } else #endif { @@ -355,7 +355,7 @@ xfs_inode_item_format( vecp->i_addr = (xfs_caddr_t)ext_buffer; vecp->i_len = xfs_iextents_copy(ip, ext_buffer, XFS_DATA_FORK); - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_IEXT); + vecp->i_type = XLOG_REG_TYPE_IEXT; } ASSERT(vecp->i_len <= ip->i_df.if_bytes); iip->ili_format.ilf_dsize = vecp->i_len; @@ -373,7 +373,7 @@ xfs_inode_item_format( ASSERT(ip->i_df.if_broot != NULL); vecp->i_addr = (xfs_caddr_t)ip->i_df.if_broot; vecp->i_len = ip->i_df.if_broot_bytes; - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_IBROOT); + vecp->i_type = XLOG_REG_TYPE_IBROOT; vecp++; nvecs++; iip->ili_format.ilf_dsize = ip->i_df.if_broot_bytes; @@ -399,7 +399,7 @@ xfs_inode_item_format( ASSERT((ip->i_df.if_real_bytes == 0) || (ip->i_df.if_real_bytes == data_bytes)); vecp->i_len = (int)data_bytes; - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_ILOCAL); + vecp->i_type = XLOG_REG_TYPE_ILOCAL; vecp++; nvecs++; iip->ili_format.ilf_dsize = (unsigned)data_bytes; @@ -477,7 +477,7 @@ xfs_inode_item_format( vecp->i_len = xfs_iextents_copy(ip, ext_buffer, XFS_ATTR_FORK); #endif - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_IATTR_EXT); + vecp->i_type = XLOG_REG_TYPE_IATTR_EXT; iip->ili_format.ilf_asize = vecp->i_len; vecp++; nvecs++; @@ -492,7 +492,7 @@ xfs_inode_item_format( ASSERT(ip->i_afp->if_broot != NULL); vecp->i_addr = (xfs_caddr_t)ip->i_afp->if_broot; vecp->i_len = ip->i_afp->if_broot_bytes; - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_IATTR_BROOT); + vecp->i_type = XLOG_REG_TYPE_IATTR_BROOT; vecp++; nvecs++; iip->ili_format.ilf_asize = ip->i_afp->if_broot_bytes; @@ -516,7 +516,7 @@ xfs_inode_item_format( ASSERT((ip->i_afp->if_real_bytes == 0) || (ip->i_afp->if_real_bytes == data_bytes)); vecp->i_len = (int)data_bytes; - XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_IATTR_LOCAL); + vecp->i_type = XLOG_REG_TYPE_IATTR_LOCAL; vecp++; nvecs++; iip->ili_format.ilf_asize = (unsigned)data_bytes; diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 0d17516..20118dd 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -617,7 +617,7 @@ xfs_log_unmount_write(xfs_mount_t *mp) if (! (XLOG_FORCED_SHUTDOWN(log))) { reg[0].i_addr = (void*)&magic; reg[0].i_len = sizeof(magic); - XLOG_VEC_SET_TYPE(®[0], XLOG_REG_TYPE_UNMOUNT); + reg[0].i_type = XLOG_REG_TYPE_UNMOUNT; error = xfs_log_reserve(mp, 600, 1, &tic, XFS_LOG, 0, XLOG_UNMOUNT_REC_TYPE); @@ -1236,7 +1236,7 @@ xlog_commit_record(xfs_mount_t *mp, reg[0].i_addr = NULL; reg[0].i_len = 0; - XLOG_VEC_SET_TYPE(®[0], XLOG_REG_TYPE_COMMIT); + reg[0].i_type = XLOG_REG_TYPE_COMMIT; ASSERT_ALWAYS(iclog); if ((error = xlog_write(mp, reg, 1, ticket, commitlsnp, diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h index d0c9baa..811ccf4 100644 --- a/fs/xfs/xfs_log.h +++ b/fs/xfs/xfs_log.h @@ -110,10 +110,8 @@ static inline xfs_lsn_t _lsn_cmp(xfs_lsn_t lsn1, xfs_lsn_t lsn2) #define XLOG_REG_TYPE_TRANSHDR 19 #define XLOG_REG_TYPE_MAX 19 -#define XLOG_VEC_SET_TYPE(vecp, t) ((vecp)->i_type = (t)) - typedef struct xfs_log_iovec { - xfs_caddr_t i_addr; /* beginning address of region */ + xfs_caddr_t i_addr; /* beginning address of region */ int i_len; /* length in bytes of region */ uint i_type; /* type of region */ } xfs_log_iovec_t; diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index 237badc..7dbe3c3 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -1121,7 +1121,7 @@ xfs_trans_fill_vecs( tp->t_header.th_num_items = nitems; log_vector->i_addr = (xfs_caddr_t)&tp->t_header; log_vector->i_len = sizeof(xfs_trans_header_t); - XLOG_VEC_SET_TYPE(log_vector, XLOG_REG_TYPE_TRANSHDR); + log_vector->i_type = XLOG_REG_TYPE_TRANSHDR; } -- cgit v0.10.2 From a14a348bff2f99471a28e5928eb6801224c053d8 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 19 Jan 2010 09:56:46 +0000 Subject: xfs: cleanup up xfs_log_force calling conventions Remove the XFS_LOG_FORCE argument which was always set, and the XFS_LOG_URGE define, which was never used. Split xfs_log_force into a two helpers - xfs_log_force which forces the whole log, and xfs_log_force_lsn which forces up to the specified LSN. The underlying implementations already were entirely separate, as were the users. Also re-indent the new _xfs_log_force/_xfs_log_force which previously had a weird coding style. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c index 58c24be..c9b863e 100644 --- a/fs/xfs/linux-2.6/xfs_sync.c +++ b/fs/xfs/linux-2.6/xfs_sync.c @@ -296,10 +296,7 @@ xfs_sync_data( if (error) return XFS_ERROR(error); - xfs_log_force(mp, 0, - (flags & SYNC_WAIT) ? - XFS_LOG_FORCE | XFS_LOG_SYNC : - XFS_LOG_FORCE); + xfs_log_force(mp, (flags & SYNC_WAIT) ? XFS_LOG_SYNC : 0); return 0; } @@ -325,10 +322,6 @@ xfs_commit_dummy_trans( struct xfs_inode *ip = mp->m_rootip; struct xfs_trans *tp; int error; - int log_flags = XFS_LOG_FORCE; - - if (flags & SYNC_WAIT) - log_flags |= XFS_LOG_SYNC; /* * Put a dummy transaction in the log to tell recovery @@ -350,7 +343,7 @@ xfs_commit_dummy_trans( xfs_iunlock(ip, XFS_ILOCK_EXCL); /* the log force ensures this transaction is pushed to disk */ - xfs_log_force(mp, 0, log_flags); + xfs_log_force(mp, (flags & SYNC_WAIT) ? XFS_LOG_SYNC : 0); return error; } @@ -390,7 +383,7 @@ xfs_sync_fsdata( * become pinned in between there and here. */ if (XFS_BUF_ISPINNED(bp)) - xfs_log_force(mp, 0, XFS_LOG_FORCE); + xfs_log_force(mp, 0); } @@ -575,7 +568,7 @@ xfs_flush_inodes( igrab(inode); xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_inodes_work, &completion); wait_for_completion(&completion); - xfs_log_force(ip->i_mount, (xfs_lsn_t)0, XFS_LOG_FORCE|XFS_LOG_SYNC); + xfs_log_force(ip->i_mount, XFS_LOG_SYNC); } /* @@ -591,7 +584,7 @@ xfs_sync_worker( int error; if (!(mp->m_flags & XFS_MOUNT_RDONLY)) { - xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE); + xfs_log_force(mp, 0); xfs_reclaim_inodes(mp, XFS_IFLUSH_DELWRI_ELSE_ASYNC); /* dgc: errors ignored here */ error = xfs_qm_sync(mp, SYNC_TRYLOCK); diff --git a/fs/xfs/quota/xfs_dquot.c b/fs/xfs/quota/xfs_dquot.c index 5756392..f9baeed 100644 --- a/fs/xfs/quota/xfs_dquot.c +++ b/fs/xfs/quota/xfs_dquot.c @@ -1248,7 +1248,7 @@ xfs_qm_dqflush( */ if (XFS_BUF_ISPINNED(bp)) { trace_xfs_dqflush_force(dqp); - xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE); + xfs_log_force(mp, 0); } if (flags & XFS_QMOPT_DELWRI) { @@ -1531,11 +1531,9 @@ xfs_qm_dqflock_pushbuf_wait( if (bp != NULL) { if (XFS_BUF_ISDELAYWRITE(bp)) { int error; - if (XFS_BUF_ISPINNED(bp)) { - xfs_log_force(dqp->q_mount, - (xfs_lsn_t)0, - XFS_LOG_FORCE); - } + + if (XFS_BUF_ISPINNED(bp)) + xfs_log_force(dqp->q_mount, 0); error = xfs_bawrite(dqp->q_mount, bp); if (error) xfs_fs_cmn_err(CE_WARN, dqp->q_mount, diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c index 116580d..1b56437 100644 --- a/fs/xfs/quota/xfs_dquot_item.c +++ b/fs/xfs/quota/xfs_dquot_item.c @@ -190,7 +190,7 @@ xfs_qm_dqunpin_wait( /* * Give the log a push so we don't wait here too long. */ - xfs_log_force(dqp->q_mount, (xfs_lsn_t)0, XFS_LOG_FORCE); + xfs_log_force(dqp->q_mount, 0); wait_event(dqp->q_pinwait, (atomic_read(&dqp->q_pincount) == 0)); } @@ -245,10 +245,9 @@ xfs_qm_dquot_logitem_pushbuf( qip->qli_pushbuf_flag = 0; xfs_dqunlock(dqp); - if (XFS_BUF_ISPINNED(bp)) { - xfs_log_force(mp, (xfs_lsn_t)0, - XFS_LOG_FORCE); - } + if (XFS_BUF_ISPINNED(bp)) + xfs_log_force(mp, 0); + if (dopush) { int error; #ifdef XFSRACEDEBUG diff --git a/fs/xfs/quota/xfs_qm_syscalls.c b/fs/xfs/quota/xfs_qm_syscalls.c index 873e07e..5d0ee8d 100644 --- a/fs/xfs/quota/xfs_qm_syscalls.c +++ b/fs/xfs/quota/xfs_qm_syscalls.c @@ -1192,9 +1192,9 @@ xfs_qm_internalqcheck( if (! XFS_IS_QUOTA_ON(mp)) return XFS_ERROR(ESRCH); - xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE | XFS_LOG_SYNC); + xfs_log_force(mp, XFS_LOG_SYNC); XFS_bflush(mp->m_ddev_targp); - xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE | XFS_LOG_SYNC); + xfs_log_force(mp, XFS_LOG_SYNC); XFS_bflush(mp->m_ddev_targp); mutex_lock(&qcheck_lock); diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index a27aeb7..94cddbf 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -2601,5 +2601,5 @@ xfs_alloc_search_busy(xfs_trans_t *tp, * transaction that freed the block */ if (lsn) - xfs_log_force(tp->t_mountp, lsn, XFS_LOG_FORCE|XFS_LOG_SYNC); + xfs_log_force_lsn(tp->t_mountp, lsn, XFS_LOG_SYNC); } diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index bbb3bee..d0d1b5a 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2484,8 +2484,11 @@ __xfs_iunpin_wait( return; /* Give the log a push to start the unpinning I/O */ - xfs_log_force(ip->i_mount, (iip && iip->ili_last_lsn) ? - iip->ili_last_lsn : 0, XFS_LOG_FORCE); + if (iip && iip->ili_last_lsn) + xfs_log_force_lsn(ip->i_mount, iip->ili_last_lsn, 0); + else + xfs_log_force(ip->i_mount, 0); + if (wait) wait_event(ip->i_ipin_wait, (atomic_read(&ip->i_pincount) == 0)); } @@ -2970,7 +2973,7 @@ xfs_iflush( * get stuck waiting in the write for too long. */ if (XFS_BUF_ISPINNED(bp)) - xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE); + xfs_log_force(mp, 0); /* * inode clustering: diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index da4cac6..48ec1c0 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -804,10 +804,9 @@ xfs_inode_item_pushbuf( trace_xfs_inode_item_push(bp, _RET_IP_); - if (XFS_BUF_ISPINNED(bp)) { - xfs_log_force(mp, (xfs_lsn_t)0, - XFS_LOG_FORCE); - } + if (XFS_BUF_ISPINNED(bp)) + xfs_log_force(mp, 0); + if (dopush) { int error; error = xfs_bawrite(mp, bp); diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 20118dd..4f16be4 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -79,11 +79,6 @@ STATIC int xlog_state_release_iclog(xlog_t *log, STATIC void xlog_state_switch_iclogs(xlog_t *log, xlog_in_core_t *iclog, int eventual_size); -STATIC int xlog_state_sync(xlog_t *log, - xfs_lsn_t lsn, - uint flags, - int *log_flushed); -STATIC int xlog_state_sync_all(xlog_t *log, uint flags, int *log_flushed); STATIC void xlog_state_want_sync(xlog_t *log, xlog_in_core_t *iclog); /* local functions to manipulate grant head */ @@ -296,65 +291,6 @@ xfs_log_done(xfs_mount_t *mp, return lsn; } /* xfs_log_done */ - -/* - * Force the in-core log to disk. If flags == XFS_LOG_SYNC, - * the force is done synchronously. - * - * Asynchronous forces are implemented by setting the WANT_SYNC - * bit in the appropriate in-core log and then returning. - * - * Synchronous forces are implemented with a signal variable. All callers - * to force a given lsn to disk will wait on a the sv attached to the - * specific in-core log. When given in-core log finally completes its - * write to disk, that thread will wake up all threads waiting on the - * sv. - */ -int -_xfs_log_force( - xfs_mount_t *mp, - xfs_lsn_t lsn, - uint flags, - int *log_flushed) -{ - xlog_t *log = mp->m_log; - int dummy; - - if (!log_flushed) - log_flushed = &dummy; - - ASSERT(flags & XFS_LOG_FORCE); - - XFS_STATS_INC(xs_log_force); - - if (log->l_flags & XLOG_IO_ERROR) - return XFS_ERROR(EIO); - if (lsn == 0) - return xlog_state_sync_all(log, flags, log_flushed); - else - return xlog_state_sync(log, lsn, flags, log_flushed); -} /* _xfs_log_force */ - -/* - * Wrapper for _xfs_log_force(), to be used when caller doesn't care - * about errors or whether the log was flushed or not. This is the normal - * interface to use when trying to unpin items or move the log forward. - */ -void -xfs_log_force( - xfs_mount_t *mp, - xfs_lsn_t lsn, - uint flags) -{ - int error; - error = _xfs_log_force(mp, lsn, flags, NULL); - if (error) { - xfs_fs_cmn_err(CE_WARN, mp, "xfs_log_force: " - "error %d returned.", error); - } -} - - /* * Attaches a new iclog I/O completion callback routine during * transaction commit. If the log is in error state, a non-zero @@ -601,7 +537,7 @@ xfs_log_unmount_write(xfs_mount_t *mp) if (mp->m_flags & XFS_MOUNT_RDONLY) return 0; - error = _xfs_log_force(mp, 0, XFS_LOG_FORCE|XFS_LOG_SYNC, NULL); + error = _xfs_log_force(mp, XFS_LOG_SYNC, NULL); ASSERT(error || !(XLOG_FORCED_SHUTDOWN(log))); #ifdef DEBUG @@ -2853,7 +2789,6 @@ xlog_state_switch_iclogs(xlog_t *log, log->l_iclog = iclog->ic_next; } /* xlog_state_switch_iclogs */ - /* * Write out all data in the in-core log as of this exact moment in time. * @@ -2881,11 +2816,17 @@ xlog_state_switch_iclogs(xlog_t *log, * b) when we return from flushing out this iclog, it is still * not in the active nor dirty state. */ -STATIC int -xlog_state_sync_all(xlog_t *log, uint flags, int *log_flushed) +int +_xfs_log_force( + struct xfs_mount *mp, + uint flags, + int *log_flushed) { - xlog_in_core_t *iclog; - xfs_lsn_t lsn; + struct log *log = mp->m_log; + struct xlog_in_core *iclog; + xfs_lsn_t lsn; + + XFS_STATS_INC(xs_log_force); spin_lock(&log->l_icloglock); @@ -2931,7 +2872,9 @@ xlog_state_sync_all(xlog_t *log, uint flags, int *log_flushed) if (xlog_state_release_iclog(log, iclog)) return XFS_ERROR(EIO); - *log_flushed = 1; + + if (log_flushed) + *log_flushed = 1; spin_lock(&log->l_icloglock); if (be64_to_cpu(iclog->ic_header.h_lsn) == lsn && iclog->ic_state != XLOG_STATE_DIRTY) @@ -2975,19 +2918,37 @@ maybe_sleep: */ if (iclog->ic_state & XLOG_STATE_IOERROR) return XFS_ERROR(EIO); - *log_flushed = 1; - + if (log_flushed) + *log_flushed = 1; } else { no_sleep: spin_unlock(&log->l_icloglock); } return 0; -} /* xlog_state_sync_all */ +} +/* + * Wrapper for _xfs_log_force(), to be used when caller doesn't care + * about errors or whether the log was flushed or not. This is the normal + * interface to use when trying to unpin items or move the log forward. + */ +void +xfs_log_force( + xfs_mount_t *mp, + uint flags) +{ + int error; + + error = _xfs_log_force(mp, flags, NULL); + if (error) { + xfs_fs_cmn_err(CE_WARN, mp, "xfs_log_force: " + "error %d returned.", error); + } +} /* - * Used by code which implements synchronous log forces. + * Force the in-core log to disk for a specific LSN. * * Find in-core log with lsn. * If it is in the DIRTY state, just return. @@ -2995,109 +2956,142 @@ no_sleep: * state and go to sleep or return. * If it is in any other state, go to sleep or return. * - * If filesystem activity goes to zero, the iclog will get flushed only by - * bdflush(). + * Synchronous forces are implemented with a signal variable. All callers + * to force a given lsn to disk will wait on a the sv attached to the + * specific in-core log. When given in-core log finally completes its + * write to disk, that thread will wake up all threads waiting on the + * sv. */ -STATIC int -xlog_state_sync(xlog_t *log, - xfs_lsn_t lsn, - uint flags, - int *log_flushed) +int +_xfs_log_force_lsn( + struct xfs_mount *mp, + xfs_lsn_t lsn, + uint flags, + int *log_flushed) { - xlog_in_core_t *iclog; - int already_slept = 0; - -try_again: - spin_lock(&log->l_icloglock); - iclog = log->l_iclog; + struct log *log = mp->m_log; + struct xlog_in_core *iclog; + int already_slept = 0; - if (iclog->ic_state & XLOG_STATE_IOERROR) { - spin_unlock(&log->l_icloglock); - return XFS_ERROR(EIO); - } + ASSERT(lsn != 0); - do { - if (be64_to_cpu(iclog->ic_header.h_lsn) != lsn) { - iclog = iclog->ic_next; - continue; - } + XFS_STATS_INC(xs_log_force); - if (iclog->ic_state == XLOG_STATE_DIRTY) { +try_again: + spin_lock(&log->l_icloglock); + iclog = log->l_iclog; + if (iclog->ic_state & XLOG_STATE_IOERROR) { spin_unlock(&log->l_icloglock); - return 0; + return XFS_ERROR(EIO); } - if (iclog->ic_state == XLOG_STATE_ACTIVE) { - /* - * We sleep here if we haven't already slept (e.g. - * this is the first time we've looked at the correct - * iclog buf) and the buffer before us is going to - * be sync'ed. The reason for this is that if we - * are doing sync transactions here, by waiting for - * the previous I/O to complete, we can allow a few - * more transactions into this iclog before we close - * it down. - * - * Otherwise, we mark the buffer WANT_SYNC, and bump - * up the refcnt so we can release the log (which drops - * the ref count). The state switch keeps new transaction - * commits from using this buffer. When the current commits - * finish writing into the buffer, the refcount will drop to - * zero and the buffer will go out then. - */ - if (!already_slept && - (iclog->ic_prev->ic_state & (XLOG_STATE_WANT_SYNC | - XLOG_STATE_SYNCING))) { - ASSERT(!(iclog->ic_state & XLOG_STATE_IOERROR)); - XFS_STATS_INC(xs_log_force_sleep); - sv_wait(&iclog->ic_prev->ic_write_wait, PSWP, - &log->l_icloglock, s); - *log_flushed = 1; - already_slept = 1; - goto try_again; - } else { + do { + if (be64_to_cpu(iclog->ic_header.h_lsn) != lsn) { + iclog = iclog->ic_next; + continue; + } + + if (iclog->ic_state == XLOG_STATE_DIRTY) { + spin_unlock(&log->l_icloglock); + return 0; + } + + if (iclog->ic_state == XLOG_STATE_ACTIVE) { + /* + * We sleep here if we haven't already slept (e.g. + * this is the first time we've looked at the correct + * iclog buf) and the buffer before us is going to + * be sync'ed. The reason for this is that if we + * are doing sync transactions here, by waiting for + * the previous I/O to complete, we can allow a few + * more transactions into this iclog before we close + * it down. + * + * Otherwise, we mark the buffer WANT_SYNC, and bump + * up the refcnt so we can release the log (which + * drops the ref count). The state switch keeps new + * transaction commits from using this buffer. When + * the current commits finish writing into the buffer, + * the refcount will drop to zero and the buffer will + * go out then. + */ + if (!already_slept && + (iclog->ic_prev->ic_state & + (XLOG_STATE_WANT_SYNC | XLOG_STATE_SYNCING))) { + ASSERT(!(iclog->ic_state & XLOG_STATE_IOERROR)); + + XFS_STATS_INC(xs_log_force_sleep); + + sv_wait(&iclog->ic_prev->ic_write_wait, + PSWP, &log->l_icloglock, s); + if (log_flushed) + *log_flushed = 1; + already_slept = 1; + goto try_again; + } atomic_inc(&iclog->ic_refcnt); xlog_state_switch_iclogs(log, iclog, 0); spin_unlock(&log->l_icloglock); if (xlog_state_release_iclog(log, iclog)) return XFS_ERROR(EIO); - *log_flushed = 1; + if (log_flushed) + *log_flushed = 1; spin_lock(&log->l_icloglock); } - } - if ((flags & XFS_LOG_SYNC) && /* sleep */ - !(iclog->ic_state & (XLOG_STATE_ACTIVE | XLOG_STATE_DIRTY))) { + if ((flags & XFS_LOG_SYNC) && /* sleep */ + !(iclog->ic_state & + (XLOG_STATE_ACTIVE | XLOG_STATE_DIRTY))) { + /* + * Don't wait on completion if we know that we've + * gotten a log write error. + */ + if (iclog->ic_state & XLOG_STATE_IOERROR) { + spin_unlock(&log->l_icloglock); + return XFS_ERROR(EIO); + } + XFS_STATS_INC(xs_log_force_sleep); + sv_wait(&iclog->ic_force_wait, PSWP, &log->l_icloglock, s); + /* + * No need to grab the log lock here since we're + * only deciding whether or not to return EIO + * and the memory read should be atomic. + */ + if (iclog->ic_state & XLOG_STATE_IOERROR) + return XFS_ERROR(EIO); - /* - * Don't wait on completion if we know that we've - * gotten a log write error. - */ - if (iclog->ic_state & XLOG_STATE_IOERROR) { + if (log_flushed) + *log_flushed = 1; + } else { /* just return */ spin_unlock(&log->l_icloglock); - return XFS_ERROR(EIO); } - XFS_STATS_INC(xs_log_force_sleep); - sv_wait(&iclog->ic_force_wait, PSWP, &log->l_icloglock, s); - /* - * No need to grab the log lock here since we're - * only deciding whether or not to return EIO - * and the memory read should be atomic. - */ - if (iclog->ic_state & XLOG_STATE_IOERROR) - return XFS_ERROR(EIO); - *log_flushed = 1; - } else { /* just return */ - spin_unlock(&log->l_icloglock); - } - return 0; - } while (iclog != log->l_iclog); + return 0; + } while (iclog != log->l_iclog); - spin_unlock(&log->l_icloglock); - return 0; -} /* xlog_state_sync */ + spin_unlock(&log->l_icloglock); + return 0; +} + +/* + * Wrapper for _xfs_log_force_lsn(), to be used when caller doesn't care + * about errors or whether the log was flushed or not. This is the normal + * interface to use when trying to unpin items or move the log forward. + */ +void +xfs_log_force_lsn( + xfs_mount_t *mp, + xfs_lsn_t lsn, + uint flags) +{ + int error; + error = _xfs_log_force_lsn(mp, lsn, flags, NULL); + if (error) { + xfs_fs_cmn_err(CE_WARN, mp, "xfs_log_force: " + "error %d returned.", error); + } +} /* * Called when we want to mark the current iclog as being ready to sync to @@ -3462,7 +3456,6 @@ xfs_log_force_umount( xlog_ticket_t *tic; xlog_t *log; int retval; - int dummy; log = mp->m_log; @@ -3536,13 +3529,14 @@ xfs_log_force_umount( } spin_unlock(&log->l_grant_lock); - if (! (log->l_iclog->ic_state & XLOG_STATE_IOERROR)) { + if (!(log->l_iclog->ic_state & XLOG_STATE_IOERROR)) { ASSERT(!logerror); /* * Force the incore logs to disk before shutting the * log down completely. */ - xlog_state_sync_all(log, XFS_LOG_FORCE|XFS_LOG_SYNC, &dummy); + _xfs_log_force(mp, XFS_LOG_SYNC, NULL); + spin_lock(&log->l_icloglock); retval = xlog_state_ioerror(log); spin_unlock(&log->l_icloglock); diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h index 811ccf4..7074be9 100644 --- a/fs/xfs/xfs_log.h +++ b/fs/xfs/xfs_log.h @@ -70,14 +70,8 @@ static inline xfs_lsn_t _lsn_cmp(xfs_lsn_t lsn1, xfs_lsn_t lsn2) * Flags to xfs_log_force() * * XFS_LOG_SYNC: Synchronous force in-core log to disk - * XFS_LOG_FORCE: Start in-core log write now. - * XFS_LOG_URGE: Start write within some window of time. - * - * Note: Either XFS_LOG_FORCE or XFS_LOG_URGE must be set. */ #define XFS_LOG_SYNC 0x1 -#define XFS_LOG_FORCE 0x2 -#define XFS_LOG_URGE 0x4 #endif /* __KERNEL__ */ @@ -138,12 +132,17 @@ xfs_lsn_t xfs_log_done(struct xfs_mount *mp, void **iclog, uint flags); int _xfs_log_force(struct xfs_mount *mp, - xfs_lsn_t lsn, uint flags, int *log_forced); void xfs_log_force(struct xfs_mount *mp, - xfs_lsn_t lsn, uint flags); +int _xfs_log_force_lsn(struct xfs_mount *mp, + xfs_lsn_t lsn, + uint flags, + int *log_forced); +void xfs_log_force_lsn(struct xfs_mount *mp, + xfs_lsn_t lsn, + uint flags); int xfs_log_mount(struct xfs_mount *mp, struct xfs_buftarg *log_target, xfs_daddr_t start_block, diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 97148f0..22e6efd 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -3913,8 +3913,7 @@ xlog_recover_finish( * case the unlink transactions would have problems * pushing the EFIs out of the way. */ - xfs_log_force(log->l_mp, (xfs_lsn_t)0, - (XFS_LOG_FORCE | XFS_LOG_SYNC)); + xfs_log_force(log->l_mp, XFS_LOG_SYNC); xlog_recover_process_iunlinks(log); diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index bb01540..7f81ed7 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1455,7 +1455,7 @@ xfs_unmountfs( * push out the iclog we will never get that unlocked. hence we * need to force the log first. */ - xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE | XFS_LOG_SYNC); + xfs_log_force(mp, XFS_LOG_SYNC); xfs_reclaim_inodes(mp, XFS_IFLUSH_ASYNC); xfs_qm_unmount(mp); @@ -1465,7 +1465,7 @@ xfs_unmountfs( * that nothing is pinned. This is important because bflush() * will skip pinned buffers. */ - xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE | XFS_LOG_SYNC); + xfs_log_force(mp, XFS_LOG_SYNC); xfs_binval(mp->m_ddev_targp); if (mp->m_rtdev_targp) { diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index 7dbe3c3..be942d4 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -981,9 +981,8 @@ shut_us_down: */ if (sync) { if (!error) { - error = _xfs_log_force(mp, commit_lsn, - XFS_LOG_FORCE | XFS_LOG_SYNC, - log_flushed); + error = _xfs_log_force_lsn(mp, commit_lsn, + XFS_LOG_SYNC, log_flushed); } XFS_STATS_INC(xs_trans_sync); } else { diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c index 063dfbd..d7b1af8 100644 --- a/fs/xfs/xfs_trans_ail.c +++ b/fs/xfs/xfs_trans_ail.c @@ -371,7 +371,7 @@ xfsaild_push( * move forward in the AIL. */ XFS_STATS_INC(xs_push_ail_flush); - xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE); + xfs_log_force(mp, 0); } if (!count) { diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 4da96cd..fd108b7 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -631,9 +631,8 @@ xfs_fsync( xfs_iunlock(ip, XFS_ILOCK_SHARED); if (xfs_ipincount(ip)) { - error = _xfs_log_force(ip->i_mount, (xfs_lsn_t)0, - XFS_LOG_FORCE | XFS_LOG_SYNC, - &log_flushed); + error = _xfs_log_force(ip->i_mount, XFS_LOG_SYNC, + &log_flushed); } else { /* * If the inode is not pinned and nothing has changed -- cgit v0.10.2 From bdfb04301fa5fdd95f219539a9a5b9663b1e5fc2 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 20 Jan 2010 21:55:30 +0000 Subject: xfs: replace KM_LARGE with explicit vmalloc use We use the KM_LARGE flag to make kmem_alloc and friends use vmalloc if necessary. As we only need this for a few boot/mount time allocations just switch to explicit vmalloc calls there. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/kmem.c b/fs/xfs/linux-2.6/kmem.c index 2d3f90a..bc74055 100644 --- a/fs/xfs/linux-2.6/kmem.c +++ b/fs/xfs/linux-2.6/kmem.c @@ -16,7 +16,6 @@ * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ #include -#include #include #include #include @@ -24,8 +23,25 @@ #include "time.h" #include "kmem.h" -#define MAX_VMALLOCS 6 -#define MAX_SLAB_SIZE 0x20000 +/* + * Greedy allocation. May fail and may return vmalloced memory. + * + * Must be freed using kmem_free_large. + */ +void * +kmem_zalloc_greedy(size_t *size, size_t minsize, size_t maxsize) +{ + void *ptr; + size_t kmsize = maxsize; + + while (!(ptr = kmem_zalloc_large(kmsize))) { + if ((kmsize >>= 1) <= minsize) + kmsize = minsize; + } + if (ptr) + *size = kmsize; + return ptr; +} void * kmem_alloc(size_t size, unsigned int __nocast flags) @@ -34,19 +50,8 @@ kmem_alloc(size_t size, unsigned int __nocast flags) gfp_t lflags = kmem_flags_convert(flags); void *ptr; -#ifdef DEBUG - if (unlikely(!(flags & KM_LARGE) && (size > PAGE_SIZE))) { - printk(KERN_WARNING "Large %s attempt, size=%ld\n", - __func__, (long)size); - dump_stack(); - } -#endif - do { - if (size < MAX_SLAB_SIZE || retries > MAX_VMALLOCS) - ptr = kmalloc(size, lflags); - else - ptr = __vmalloc(size, lflags, PAGE_KERNEL); + ptr = kmalloc(size, lflags); if (ptr || (flags & (KM_MAYFAIL|KM_NOSLEEP))) return ptr; if (!(++retries % 100)) @@ -68,27 +73,6 @@ kmem_zalloc(size_t size, unsigned int __nocast flags) return ptr; } -void * -kmem_zalloc_greedy(size_t *size, size_t minsize, size_t maxsize, - unsigned int __nocast flags) -{ - void *ptr; - size_t kmsize = maxsize; - unsigned int kmflags = (flags & ~KM_SLEEP) | KM_NOSLEEP; - - while (!(ptr = kmem_zalloc(kmsize, kmflags))) { - if ((kmsize <= minsize) && (flags & KM_NOSLEEP)) - break; - if ((kmsize >>= 1) <= minsize) { - kmsize = minsize; - kmflags = flags; - } - } - if (ptr) - *size = kmsize; - return ptr; -} - void kmem_free(const void *ptr) { diff --git a/fs/xfs/linux-2.6/kmem.h b/fs/xfs/linux-2.6/kmem.h index 179cbd6..f7c8f7a 100644 --- a/fs/xfs/linux-2.6/kmem.h +++ b/fs/xfs/linux-2.6/kmem.h @@ -21,6 +21,7 @@ #include #include #include +#include /* * General memory allocation interfaces @@ -30,7 +31,6 @@ #define KM_NOSLEEP 0x0002u #define KM_NOFS 0x0004u #define KM_MAYFAIL 0x0008u -#define KM_LARGE 0x0010u /* * We use a special process flag to avoid recursive callbacks into @@ -42,7 +42,7 @@ kmem_flags_convert(unsigned int __nocast flags) { gfp_t lflags; - BUG_ON(flags & ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL|KM_LARGE)); + BUG_ON(flags & ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL)); if (flags & KM_NOSLEEP) { lflags = GFP_ATOMIC | __GFP_NOWARN; @@ -56,10 +56,25 @@ kmem_flags_convert(unsigned int __nocast flags) extern void *kmem_alloc(size_t, unsigned int __nocast); extern void *kmem_zalloc(size_t, unsigned int __nocast); -extern void *kmem_zalloc_greedy(size_t *, size_t, size_t, unsigned int __nocast); extern void *kmem_realloc(const void *, size_t, size_t, unsigned int __nocast); extern void kmem_free(const void *); +static inline void *kmem_zalloc_large(size_t size) +{ + void *ptr; + + ptr = vmalloc(size); + if (ptr) + memset(ptr, 0, size); + return ptr; +} +static inline void kmem_free_large(void *ptr) +{ + vfree(ptr); +} + +extern void *kmem_zalloc_greedy(size_t *, size_t, size_t); + /* * Zone interfaces */ diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index 730eff1..44e20e5 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -1525,8 +1525,8 @@ xfs_alloc_bufhash( btp->bt_hashshift = external ? 3 : 8; /* 8 or 256 buckets */ btp->bt_hashmask = (1 << btp->bt_hashshift) - 1; - btp->bt_hash = kmem_zalloc((1 << btp->bt_hashshift) * - sizeof(xfs_bufhash_t), KM_SLEEP | KM_LARGE); + btp->bt_hash = kmem_zalloc_large((1 << btp->bt_hashshift) * + sizeof(xfs_bufhash_t)); for (i = 0; i < (1 << btp->bt_hashshift); i++) { spin_lock_init(&btp->bt_hash[i].bh_lock); INIT_LIST_HEAD(&btp->bt_hash[i].bh_list); @@ -1537,7 +1537,7 @@ STATIC void xfs_free_bufhash( xfs_buftarg_t *btp) { - kmem_free(btp->bt_hash); + kmem_free_large(btp->bt_hash); btp->bt_hash = NULL; } diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c index 9e627a8..11cfd82 100644 --- a/fs/xfs/quota/xfs_qm.c +++ b/fs/xfs/quota/xfs_qm.c @@ -118,9 +118,14 @@ xfs_Gqm_init(void) */ udqhash = kmem_zalloc_greedy(&hsize, XFS_QM_HASHSIZE_LOW * sizeof(xfs_dqhash_t), - XFS_QM_HASHSIZE_HIGH * sizeof(xfs_dqhash_t), - KM_SLEEP | KM_MAYFAIL | KM_LARGE); - gdqhash = kmem_zalloc(hsize, KM_SLEEP | KM_LARGE); + XFS_QM_HASHSIZE_HIGH * sizeof(xfs_dqhash_t)); + if (!udqhash) + goto out; + + gdqhash = kmem_zalloc_large(hsize); + if (!udqhash) + goto out_free_udqhash; + hsize /= sizeof(xfs_dqhash_t); ndquot = hsize << 8; @@ -170,6 +175,11 @@ xfs_Gqm_init(void) mutex_init(&qcheck_lock); #endif return xqm; + + out_free_udqhash: + kmem_free_large(udqhash); + out: + return NULL; } /* @@ -189,8 +199,8 @@ xfs_qm_destroy( xfs_qm_list_destroy(&(xqm->qm_usr_dqhtable[i])); xfs_qm_list_destroy(&(xqm->qm_grp_dqhtable[i])); } - kmem_free(xqm->qm_usr_dqhtable); - kmem_free(xqm->qm_grp_dqhtable); + kmem_free_large(xqm->qm_usr_dqhtable); + kmem_free_large(xqm->qm_grp_dqhtable); xqm->qm_usr_dqhtable = NULL; xqm->qm_grp_dqhtable = NULL; xqm->qm_dqhashmask = 0; @@ -219,8 +229,12 @@ xfs_qm_hold_quotafs_ref( */ mutex_lock(&xfs_Gqm_lock); - if (xfs_Gqm == NULL) + if (!xfs_Gqm) { xfs_Gqm = xfs_Gqm_init(); + if (!xfs_Gqm) + return ENOMEM; + } + /* * We can keep a list of all filesystems with quotas mounted for * debugging and statistical purposes, but ... diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index 940307a..3af0231 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -408,8 +408,10 @@ xfs_bulkstat( (XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog); nimask = ~(nicluster - 1); nbcluster = nicluster >> mp->m_sb.sb_inopblog; - irbuf = kmem_zalloc_greedy(&irbsize, PAGE_SIZE, PAGE_SIZE * 4, - KM_SLEEP | KM_MAYFAIL | KM_LARGE); + irbuf = kmem_zalloc_greedy(&irbsize, PAGE_SIZE, PAGE_SIZE * 4); + if (!irbuf) + return ENOMEM; + nirbuf = irbsize / sizeof(*irbuf); /* @@ -727,7 +729,7 @@ xfs_bulkstat( /* * Done, we're either out of filesystem or space to put the data. */ - kmem_free(irbuf); + kmem_free_large(irbuf); *ubcountp = ubelem; /* * Found some inodes, return them now and return the error next time. -- cgit v0.10.2 From 9b00f30762fe9f914eb6e03057a616ed63a4e8ca Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 21 Jan 2010 11:17:20 +0000 Subject: xfs: quota limit statvfs available blocks A "df" run on an NFS client of an exported XFS file system reports the wrong information for "available" blocks. When a block quota is enforced, the amount reported as free is limited by the quota, but the amount reported available is not (and should be). Reported-by: Guk-Bong, Kwon Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/quota/xfs_qm_bhv.c b/fs/xfs/quota/xfs_qm_bhv.c index a534663..97b410c 100644 --- a/fs/xfs/quota/xfs_qm_bhv.c +++ b/fs/xfs/quota/xfs_qm_bhv.c @@ -59,7 +59,7 @@ xfs_fill_statvfs_from_dquot( be64_to_cpu(dp->d_blk_hardlimit); if (limit && statp->f_blocks > limit) { statp->f_blocks = limit; - statp->f_bfree = + statp->f_bfree = statp->f_bavail = (statp->f_blocks > be64_to_cpu(dp->d_bcount)) ? (statp->f_blocks - be64_to_cpu(dp->d_bcount)) : 0; } -- cgit v0.10.2 From 212b3c8b8ab94d983c2e0ee1821f17dd5b4e0859 Mon Sep 17 00:00:00 2001 From: Jean-Hugues Deschenes Date: Fri, 22 Jan 2010 10:08:31 -0700 Subject: spi/dw_spi: Fix dw_spi_mmio to depend on HAVE_CLK dw_spi_mmio is dependent on the clock framework. This marks it as such in Kconfig. Signed-off-by: Jean-Hugues Deschenes Signed-off-by: Grant Likely diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig index 1d1e396..0fee95c 100644 --- a/drivers/spi/Kconfig +++ b/drivers/spi/Kconfig @@ -336,7 +336,7 @@ config SPI_DW_PCI config SPI_DW_MMIO tristate "Memory-mapped io interface driver for DW SPI core" - depends on SPI_DESIGNWARE + depends on SPI_DESIGNWARE && HAVE_CLK # # There are lots of SPI device types, with sensors and memory -- cgit v0.10.2 From ea87bb7853168434f4a82426dd1ea8421f9e604d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 20 Jan 2010 20:58:57 +0000 Subject: sched: Extend enqueue_task to allow head queueing The ability of enqueueing a task to the head of a SCHED_FIFO priority list is required to fix some violations of POSIX scheduling policy. Extend the related functions with a "head" argument. Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra Tested-by: Carsten Emde Tested-by: Mathias Weber LKML-Reference: <20100120171629.734886007@linutronix.de> diff --git a/include/linux/sched.h b/include/linux/sched.h index 8b07973..b35c0c7 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1067,7 +1067,8 @@ struct sched_domain; struct sched_class { const struct sched_class *next; - void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup); + void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup, + bool head); void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep); void (*yield_task) (struct rq *rq); diff --git a/kernel/sched.c b/kernel/sched.c index 41e76d3..f47560f 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1856,13 +1856,14 @@ static void update_avg(u64 *avg, u64 sample) *avg += diff >> 3; } -static void enqueue_task(struct rq *rq, struct task_struct *p, int wakeup) +static void +enqueue_task(struct rq *rq, struct task_struct *p, int wakeup, bool head) { if (wakeup) p->se.start_runtime = p->se.sum_exec_runtime; sched_info_queued(p); - p->sched_class->enqueue_task(rq, p, wakeup); + p->sched_class->enqueue_task(rq, p, wakeup, head); p->se.on_rq = 1; } @@ -1892,7 +1893,7 @@ static void activate_task(struct rq *rq, struct task_struct *p, int wakeup) if (task_contributes_to_load(p)) rq->nr_uninterruptible--; - enqueue_task(rq, p, wakeup); + enqueue_task(rq, p, wakeup, false); inc_nr_running(rq); } @@ -4236,7 +4237,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio) if (running) p->sched_class->set_curr_task(rq); if (on_rq) { - enqueue_task(rq, p, 0); + enqueue_task(rq, p, 0, false); check_class_changed(rq, p, prev_class, oldprio, running); } @@ -4280,7 +4281,7 @@ void set_user_nice(struct task_struct *p, long nice) delta = p->prio - old_prio; if (on_rq) { - enqueue_task(rq, p, 0); + enqueue_task(rq, p, 0, false); /* * If the task increased its priority or is running and * lowered its priority, then reschedule its CPU: @@ -8230,7 +8231,7 @@ void sched_move_task(struct task_struct *tsk) if (unlikely(running)) tsk->sched_class->set_curr_task(rq); if (on_rq) - enqueue_task(rq, tsk, 0); + enqueue_task(rq, tsk, 0, false); task_rq_unlock(rq, &flags); } diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 22231cc..0e7a7af 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1053,7 +1053,8 @@ static inline void hrtick_update(struct rq *rq) * increased. Here we update the fair scheduling stats and * then put the task into the rbtree: */ -static void enqueue_task_fair(struct rq *rq, struct task_struct *p, int wakeup) +static void +enqueue_task_fair(struct rq *rq, struct task_struct *p, int wakeup, bool head) { struct cfs_rq *cfs_rq; struct sched_entity *se = &p->se; diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 502bb61..38076da 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -878,7 +878,8 @@ static void dequeue_rt_entity(struct sched_rt_entity *rt_se) /* * Adding/removing a task to/from a priority array: */ -static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup) +static void +enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup, bool head) { struct sched_rt_entity *rt_se = &p->rt; -- cgit v0.10.2 From 37dad3fce97f01e5149d69de0833d8452c0e862e Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 20 Jan 2010 20:59:01 +0000 Subject: sched: Implement head queueing for sched_rt The ability of enqueueing a task to the head of a SCHED_FIFO priority list is required to fix some violations of POSIX scheduling policy. Implement the functionality in sched_rt. Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra Tested-by: Carsten Emde Tested-by: Mathias Weber LKML-Reference: <20100120171629.772169931@linutronix.de> diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 38076da..ca49ceb 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -194,7 +194,7 @@ static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se) return rt_se->my_q; } -static void enqueue_rt_entity(struct sched_rt_entity *rt_se); +static void enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head); static void dequeue_rt_entity(struct sched_rt_entity *rt_se); static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) @@ -204,7 +204,7 @@ static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) if (rt_rq->rt_nr_running) { if (rt_se && !on_rt_rq(rt_se)) - enqueue_rt_entity(rt_se); + enqueue_rt_entity(rt_se, false); if (rt_rq->highest_prio.curr < curr->prio) resched_task(curr); } @@ -803,7 +803,7 @@ void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) dec_rt_group(rt_se, rt_rq); } -static void __enqueue_rt_entity(struct sched_rt_entity *rt_se) +static void __enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head) { struct rt_rq *rt_rq = rt_rq_of_se(rt_se); struct rt_prio_array *array = &rt_rq->active; @@ -819,7 +819,10 @@ static void __enqueue_rt_entity(struct sched_rt_entity *rt_se) if (group_rq && (rt_rq_throttled(group_rq) || !group_rq->rt_nr_running)) return; - list_add_tail(&rt_se->run_list, queue); + if (head) + list_add(&rt_se->run_list, queue); + else + list_add_tail(&rt_se->run_list, queue); __set_bit(rt_se_prio(rt_se), array->bitmap); inc_rt_tasks(rt_se, rt_rq); @@ -856,11 +859,11 @@ static void dequeue_rt_stack(struct sched_rt_entity *rt_se) } } -static void enqueue_rt_entity(struct sched_rt_entity *rt_se) +static void enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head) { dequeue_rt_stack(rt_se); for_each_sched_rt_entity(rt_se) - __enqueue_rt_entity(rt_se); + __enqueue_rt_entity(rt_se, head); } static void dequeue_rt_entity(struct sched_rt_entity *rt_se) @@ -871,7 +874,7 @@ static void dequeue_rt_entity(struct sched_rt_entity *rt_se) struct rt_rq *rt_rq = group_rt_rq(rt_se); if (rt_rq && rt_rq->rt_nr_running) - __enqueue_rt_entity(rt_se); + __enqueue_rt_entity(rt_se, false); } } @@ -886,7 +889,7 @@ enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup, bool head) if (wakeup) rt_se->timeout = 0; - enqueue_rt_entity(rt_se); + enqueue_rt_entity(rt_se, head); if (!task_current(rq, p) && p->rt.nr_cpus_allowed > 1) enqueue_pushable_task(rq, p); -- cgit v0.10.2 From 60db48cacb9b253d5607a5ff206112a59cd09e34 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 20 Jan 2010 20:59:06 +0000 Subject: sched: Queue a deboosted task to the head of the RT prio queue rtmutex_set_prio() is used to implement priority inheritance for futexes. When a task is deboosted it gets enqueued at the tail of its RT priority list. This is violating the POSIX scheduling semantics: rt priority list X contains two runnable tasks A and B task A runs with priority X and holds mutex M task C preempts A and is blocked on mutex M -> task A is boosted to priority of task C (Y) task A unlocks the mutex M and deboosts itself -> A is dequeued from rt priority list Y -> A is enqueued to the tail of rt priority list X task C schedules away task B runs This is wrong as task A did not schedule away and therefor violates the POSIX scheduling semantics. Enqueue the task to the head of the priority list instead. Reported-by: Mathias Weber Reported-by: Carsten Emde Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra Tested-by: Carsten Emde Tested-by: Mathias Weber LKML-Reference: <20100120171629.809074113@linutronix.de> diff --git a/kernel/sched.c b/kernel/sched.c index f47560f..a56ead4 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4237,7 +4237,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio) if (running) p->sched_class->set_curr_task(rq); if (on_rq) { - enqueue_task(rq, p, 0, false); + enqueue_task(rq, p, 0, oldprio < prio); check_class_changed(rq, p, prev_class, oldprio, running); } -- cgit v0.10.2 From 0406ad336c066190770cbf350b552d608e43ed09 Mon Sep 17 00:00:00 2001 From: Alex Chiang Date: Wed, 20 Jan 2010 00:06:30 -0700 Subject: ACPI: processor: add kernel command line support for early _PDC eval Allow platforms not listed in DMI table to opt-in and evaluate _PDC early. Signed-off-by: Alex Chiang Signed-off-by: Len Brown diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 736d456..826b6e1 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -199,6 +199,10 @@ and is between 256 and 4096 characters. It is defined in the file acpi_display_output=video See above. + acpi_early_pdc_eval [HW,ACPI] Evaluate processor _PDC methods + early. Needed on some platforms to properly + initialize the EC. + acpi_irq_balance [HW,ACPI] ACPI will balance active IRQs default in APIC mode diff --git a/drivers/acpi/processor_pdc.c b/drivers/acpi/processor_pdc.c index 7247819..3bbafe9 100644 --- a/drivers/acpi/processor_pdc.c +++ b/drivers/acpi/processor_pdc.c @@ -151,6 +151,13 @@ static int set_early_pdc_optin(const struct dmi_system_id *id) return 0; } +static int param_early_pdc_optin(char *s) +{ + early_pdc_optin = 1; + return 1; +} +__setup("acpi_early_pdc_eval", param_early_pdc_optin); + static struct dmi_system_id __cpuinitdata early_pdc_optin_table[] = { { set_early_pdc_optin, "HP Envy", { -- cgit v0.10.2 From a4932299d03a1c20e58e4cc40a66fb0a048fb3a7 Mon Sep 17 00:00:00 2001 From: Alex Chiang Date: Wed, 20 Jan 2010 00:06:35 -0700 Subject: ACPI: processor: only evaluate _PDC once per processor If we evaluate _PDC in the early path, we do not want to evaluate it again when the processor driver is loaded. Cc: Venkatesh Pallipadi Signed-off-by: Alex Chiang Signed-off-by: Len Brown diff --git a/drivers/acpi/processor_pdc.c b/drivers/acpi/processor_pdc.c index 3bbafe9..e306ba9 100644 --- a/drivers/acpi/processor_pdc.c +++ b/drivers/acpi/processor_pdc.c @@ -125,6 +125,8 @@ acpi_processor_eval_pdc(acpi_handle handle, struct acpi_object_list *pdc_in) return status; } +static int early_pdc_done; + void acpi_processor_set_pdc(acpi_handle handle) { struct acpi_object_list *obj_list; @@ -132,6 +134,9 @@ void acpi_processor_set_pdc(acpi_handle handle) if (arch_has_acpi_pdc() == false) return; + if (early_pdc_done) + return; + obj_list = acpi_processor_alloc_pdc(); if (!obj_list) return; @@ -199,4 +204,6 @@ void __init acpi_early_processor_set_pdc(void) acpi_walk_namespace(ACPI_TYPE_PROCESSOR, ACPI_ROOT_OBJECT, ACPI_UINT32_MAX, early_init_pdc, NULL, NULL, NULL); + + early_pdc_done = 1; } -- cgit v0.10.2 From a7b480e7f30b3813353ec009f10f2ac7a6669f3b Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 22 Jan 2010 16:01:03 +0100 Subject: x86, lib: Add wbinvd smp helpers Add wbinvd_on_cpu and wbinvd_on_all_cpus stubs for executing wbinvd on a particular CPU. [ hpa: renamed lib/smp.c to lib/cache-smp.c ] [ hpa: wbinvd_on_all_cpus() returns int, but wbinvd() returns void. Thus, the former cannot be a macro for the latter, replace with an inline function. ] Signed-off-by: Borislav Petkov LKML-Reference: <1264172467-25155-2-git-send-email-bp@amd64.org> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h index 1e79678..4cfc908 100644 --- a/arch/x86/include/asm/smp.h +++ b/arch/x86/include/asm/smp.h @@ -135,6 +135,8 @@ int native_cpu_disable(void); void native_cpu_die(unsigned int cpu); void native_play_dead(void); void play_dead_common(void); +void wbinvd_on_cpu(int cpu); +int wbinvd_on_all_cpus(void); void native_send_call_func_ipi(const struct cpumask *mask); void native_send_call_func_single_ipi(int cpu); @@ -147,6 +149,13 @@ static inline int num_booting_cpus(void) { return cpumask_weight(cpu_callout_mask); } +#else /* !CONFIG_SMP */ +#define wbinvd_on_cpu(cpu) wbinvd() +static inline int wbinvd_on_all_cpus(void) +{ + wbinvd(); + return 0; +} #endif /* CONFIG_SMP */ extern unsigned disabled_cpus __cpuinitdata; diff --git a/arch/x86/lib/Makefile b/arch/x86/lib/Makefile index cffd754f..d85e0e4 100644 --- a/arch/x86/lib/Makefile +++ b/arch/x86/lib/Makefile @@ -14,7 +14,7 @@ $(obj)/inat.o: $(obj)/inat-tables.c clean-files := inat-tables.c -obj-$(CONFIG_SMP) += msr-smp.o +obj-$(CONFIG_SMP) += msr-smp.o cache-smp.o lib-y := delay.o lib-y += thunk_$(BITS).o diff --git a/arch/x86/lib/cache-smp.c b/arch/x86/lib/cache-smp.c new file mode 100644 index 0000000..a3c6688 --- /dev/null +++ b/arch/x86/lib/cache-smp.c @@ -0,0 +1,19 @@ +#include +#include + +static void __wbinvd(void *dummy) +{ + wbinvd(); +} + +void wbinvd_on_cpu(int cpu) +{ + smp_call_function_single(cpu, __wbinvd, NULL, 1); +} +EXPORT_SYMBOL(wbinvd_on_cpu); + +int wbinvd_on_all_cpus(void) +{ + return on_each_cpu(__wbinvd, NULL, 1); +} +EXPORT_SYMBOL(wbinvd_on_all_cpus); -- cgit v0.10.2 From 48a719c238bcbb72d6da79de9c5b3b93ab472107 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 22 Jan 2010 16:01:04 +0100 Subject: intel-agp: Switch to wbinvd_on_all_cpus Simplify if-statement while at it. [ hpa: we need to #include ] Cc: Dave Jones Cc: David Airlie Signed-off-by: Borislav Petkov LKML-Reference: <1264172467-25155-3-git-send-email-bp@amd64.org> Signed-off-by: H. Peter Anvin diff --git a/drivers/char/agp/intel-agp.c b/drivers/char/agp/intel-agp.c index 3999a5f..8a713f1 100644 --- a/drivers/char/agp/intel-agp.c +++ b/drivers/char/agp/intel-agp.c @@ -8,6 +8,7 @@ #include #include #include +#include #include "agp.h" /* @@ -815,12 +816,6 @@ static void intel_i830_setup_flush(void) intel_i830_fini_flush(); } -static void -do_wbinvd(void *null) -{ - wbinvd(); -} - /* The chipset_flush interface needs to get data that has already been * flushed out of the CPU all the way out to main memory, because the GPU * doesn't snoop those buffers. @@ -837,12 +832,10 @@ static void intel_i830_chipset_flush(struct agp_bridge_data *bridge) memset(pg, 0, 1024); - if (cpu_has_clflush) { + if (cpu_has_clflush) clflush_cache_range(pg, 1024); - } else { - if (on_each_cpu(do_wbinvd, NULL, 1) != 0) - printk(KERN_ERR "Timed out waiting for cache flush.\n"); - } + else if (wbinvd_on_all_cpus() != 0) + printk(KERN_ERR "Timed out waiting for cache flush.\n"); } /* The intel i830 automatically initializes the agp aperture during POST. -- cgit v0.10.2 From dcf39daf3d6d97f8741e82f0b9fb7554704ed2d1 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 22 Jan 2010 16:01:05 +0100 Subject: x86, cacheinfo: Fix disabling of L3 cache indices * Correct the masks used for writing the cache index disable indices. * Do not turn off L3 scrubber - it is not necessary. * Make sure wbinvd is executed on the same node where the L3 is. * Check for out-of-bounds values written to the registers. * Make show_cache_disable hex values unambiguous * Check for Erratum #388 Signed-off-by: Borislav Petkov LKML-Reference: <1264172467-25155-4-git-send-email-bp@amd64.org> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index fc6c8ef..08c91ab 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -18,6 +18,7 @@ #include #include #include +#include #define LVL_1_INST 1 #define LVL_1_DATA 2 @@ -299,8 +300,10 @@ amd_check_l3_disable(int index, struct _cpuid4_info_regs *this_leaf) if (boot_cpu_data.x86 == 0x11) return; - /* see erratum #382 */ - if ((boot_cpu_data.x86 == 0x10) && (boot_cpu_data.x86_model < 0x8)) + /* see errata #382 and #388 */ + if ((boot_cpu_data.x86 == 0x10) && + ((boot_cpu_data.x86_model < 0x9) || + (boot_cpu_data.x86_mask < 0x1))) return; this_leaf->can_disable = 1; @@ -726,12 +729,12 @@ static ssize_t show_cache_disable(struct _cpuid4_info *this_leaf, char *buf, return -EINVAL; pci_read_config_dword(dev, 0x1BC + index * 4, ®); - return sprintf(buf, "%x\n", reg); + return sprintf(buf, "0x%08x\n", reg); } #define SHOW_CACHE_DISABLE(index) \ static ssize_t \ -show_cache_disable_##index(struct _cpuid4_info *this_leaf, char *buf) \ +show_cache_disable_##index(struct _cpuid4_info *this_leaf, char *buf) \ { \ return show_cache_disable(this_leaf, buf, index); \ } @@ -745,7 +748,9 @@ static ssize_t store_cache_disable(struct _cpuid4_info *this_leaf, int node = cpu_to_node(cpu); struct pci_dev *dev = node_to_k8_nb_misc(node); unsigned long val = 0; - unsigned int scrubber = 0; + +#define SUBCACHE_MASK (3UL << 20) +#define SUBCACHE_INDEX 0xfff if (!this_leaf->can_disable) return -EINVAL; @@ -759,21 +764,24 @@ static ssize_t store_cache_disable(struct _cpuid4_info *this_leaf, if (strict_strtoul(buf, 10, &val) < 0) return -EINVAL; - val |= 0xc0000000; - - pci_read_config_dword(dev, 0x58, &scrubber); - scrubber &= ~0x1f000000; - pci_write_config_dword(dev, 0x58, scrubber); + /* do not allow writes outside of allowed bits */ + if (val & ~(SUBCACHE_MASK | SUBCACHE_INDEX)) + return -EINVAL; - pci_write_config_dword(dev, 0x1BC + index * 4, val & ~0x40000000); - wbinvd(); + val |= BIT(30); pci_write_config_dword(dev, 0x1BC + index * 4, val); + /* + * We need to WBINVD on a core on the node containing the L3 cache which + * indices we disable therefore a simple wbinvd() is not sufficient. + */ + wbinvd_on_cpu(cpu); + pci_write_config_dword(dev, 0x1BC + index * 4, val | BIT(31)); return count; } #define STORE_CACHE_DISABLE(index) \ static ssize_t \ -store_cache_disable_##index(struct _cpuid4_info *this_leaf, \ +store_cache_disable_##index(struct _cpuid4_info *this_leaf, \ const char *buf, size_t count) \ { \ return store_cache_disable(this_leaf, buf, count, index); \ -- cgit v0.10.2 From 897de50e08937663912c86fb12ad7f708af2386c Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 22 Jan 2010 16:01:06 +0100 Subject: x86, cacheinfo: Add cache index disable sysfs attrs only to L3 caches The cache_disable_[01] attribute in /sys/devices/system/cpu/cpu?/cache/index[0-3]/ is enabled on all cache levels although only L3 supports it. Add it only to the cache level that actually supports it. Signed-off-by: Borislav Petkov LKML-Reference: <1264172467-25155-5-git-send-email-bp@amd64.org> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 08c91ab..3976ce9 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -814,16 +814,24 @@ static struct _cache_attr cache_disable_0 = __ATTR(cache_disable_0, 0644, static struct _cache_attr cache_disable_1 = __ATTR(cache_disable_1, 0644, show_cache_disable_1, store_cache_disable_1); +#define DEFAULT_SYSFS_CACHE_ATTRS \ + &type.attr, \ + &level.attr, \ + &coherency_line_size.attr, \ + &physical_line_partition.attr, \ + &ways_of_associativity.attr, \ + &number_of_sets.attr, \ + &size.attr, \ + &shared_cpu_map.attr, \ + &shared_cpu_list.attr + static struct attribute *default_attrs[] = { - &type.attr, - &level.attr, - &coherency_line_size.attr, - &physical_line_partition.attr, - &ways_of_associativity.attr, - &number_of_sets.attr, - &size.attr, - &shared_cpu_map.attr, - &shared_cpu_list.attr, + DEFAULT_SYSFS_CACHE_ATTRS, + NULL +}; + +static struct attribute *default_l3_attrs[] = { + DEFAULT_SYSFS_CACHE_ATTRS, &cache_disable_0.attr, &cache_disable_1.attr, NULL @@ -916,6 +924,7 @@ static int __cpuinit cache_add_dev(struct sys_device * sys_dev) unsigned int cpu = sys_dev->id; unsigned long i, j; struct _index_kobject *this_object; + struct _cpuid4_info *this_leaf; int retval; retval = cpuid4_cache_sysfs_init(cpu); @@ -934,6 +943,14 @@ static int __cpuinit cache_add_dev(struct sys_device * sys_dev) this_object = INDEX_KOBJECT_PTR(cpu, i); this_object->cpu = cpu; this_object->index = i; + + this_leaf = CPUID4_INFO_IDX(cpu, i); + + if (this_leaf->can_disable) + ktype_cache.default_attrs = default_l3_attrs; + else + ktype_cache.default_attrs = default_attrs; + retval = kobject_init_and_add(&(this_object->kobj), &ktype_cache, per_cpu(ici_cache_kobject, cpu), -- cgit v0.10.2 From 048a8774ca43488d78605031f11cc206d7a2682a Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Fri, 22 Jan 2010 16:01:07 +0100 Subject: x86, cacheinfo: Calculate L3 indices We need to know the valid L3 indices interval when disabling them over /sysfs. Do that when the core is brought online and add boundary checks to the sysfs .store attribute. Signed-off-by: Borislav Petkov LKML-Reference: <1264172467-25155-6-git-send-email-bp@amd64.org> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 3976ce9..589b705 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -151,7 +151,8 @@ struct _cpuid4_info { union _cpuid4_leaf_ebx ebx; union _cpuid4_leaf_ecx ecx; unsigned long size; - unsigned long can_disable; + bool can_disable; + unsigned int l3_indices; DECLARE_BITMAP(shared_cpu_map, NR_CPUS); }; @@ -161,7 +162,8 @@ struct _cpuid4_info_regs { union _cpuid4_leaf_ebx ebx; union _cpuid4_leaf_ecx ecx; unsigned long size; - unsigned long can_disable; + bool can_disable; + unsigned int l3_indices; }; unsigned short num_cache_leaves; @@ -291,6 +293,29 @@ amd_cpuid4(int leaf, union _cpuid4_leaf_eax *eax, (ebx->split.ways_of_associativity + 1) - 1; } +static unsigned int __cpuinit amd_calc_l3_indices(void) +{ + /* + * We're called over smp_call_function_single() and therefore + * are on the correct cpu. + */ + int cpu = smp_processor_id(); + int node = cpu_to_node(cpu); + struct pci_dev *dev = node_to_k8_nb_misc(node); + unsigned int sc0, sc1, sc2, sc3; + u32 val; + + pci_read_config_dword(dev, 0x1C4, &val); + + /* calculate subcache sizes */ + sc0 = !(val & BIT(0)); + sc1 = !(val & BIT(4)); + sc2 = !(val & BIT(8)) + !(val & BIT(9)); + sc3 = !(val & BIT(12)) + !(val & BIT(13)); + + return (max(max(max(sc0, sc1), sc2), sc3) << 10) - 1; +} + static void __cpuinit amd_check_l3_disable(int index, struct _cpuid4_info_regs *this_leaf) { @@ -306,7 +331,8 @@ amd_check_l3_disable(int index, struct _cpuid4_info_regs *this_leaf) (boot_cpu_data.x86_mask < 0x1))) return; - this_leaf->can_disable = 1; + this_leaf->can_disable = true; + this_leaf->l3_indices = amd_calc_l3_indices(); } static int @@ -765,7 +791,8 @@ static ssize_t store_cache_disable(struct _cpuid4_info *this_leaf, return -EINVAL; /* do not allow writes outside of allowed bits */ - if (val & ~(SUBCACHE_MASK | SUBCACHE_INDEX)) + if ((val & ~(SUBCACHE_MASK | SUBCACHE_INDEX)) || + ((val & SUBCACHE_INDEX) > this_leaf->l3_indices)) return -EINVAL; val |= BIT(30); -- cgit v0.10.2 From 3970dd8c5169505f0cc5e4c3e2fde7bdd9bbad3e Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Sat, 2 Jan 2010 23:19:45 +0100 Subject: pcmcia: do not lock socket driver module on card insert Do not lock the socket driver module on card insert, as the PCMCIA core can handle a socket module removal, at least if we add a call to socket_remove() on pccardd()'s shutdown. Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/cs.c b/drivers/pcmcia/cs.c index f0630a6..137a5db 100644 --- a/drivers/pcmcia/cs.c +++ b/drivers/pcmcia/cs.c @@ -407,7 +407,7 @@ static void socket_shutdown(struct pcmcia_socket *s) "*** DANGER *** unable to remove socket power\n"); } - cs_socket_put(s); + s->state &= ~SOCKET_INUSE; } static int socket_setup(struct pcmcia_socket *skt, int initial_delay) @@ -496,8 +496,8 @@ static int socket_insert(struct pcmcia_socket *skt) dev_dbg(&skt->dev, "insert\n"); - if (!cs_socket_get(skt)) - return -ENODEV; + WARN_ON(skt->state & SOCKET_INUSE); + skt->state |= SOCKET_INUSE; ret = socket_setup(skt, setup_delay); if (ret == 0) { @@ -697,6 +697,13 @@ static int pccardd(void *__skt) /* make sure we are running before we exit */ set_current_state(TASK_RUNNING); + /* shut down socket, if a device is still present */ + if (skt->state & SOCKET_PRESENT) { + mutex_lock(&skt->skt_mutex); + socket_remove(skt); + mutex_unlock(&skt->skt_mutex); + } + /* remove from the device core */ pccard_sysfs_remove_socket(&skt->dev); device_unregister(&skt->dev); diff --git a/drivers/pcmcia/cs_internal.h b/drivers/pcmcia/cs_internal.h index 3bc02d5..9a3bbad 100644 --- a/drivers/pcmcia/cs_internal.h +++ b/drivers/pcmcia/cs_internal.h @@ -87,26 +87,6 @@ struct pccard_resource_ops { #define SOCKET_CARDBUS 0x8000 #define SOCKET_CARDBUS_CONFIG 0x10000 -static inline int cs_socket_get(struct pcmcia_socket *skt) -{ - int ret; - - WARN_ON(skt->state & SOCKET_INUSE); - - ret = try_module_get(skt->owner); - if (ret) - skt->state |= SOCKET_INUSE; - return ret; -} - -static inline void cs_socket_put(struct pcmcia_socket *skt) -{ - if (skt->state & SOCKET_INUSE) { - skt->state &= ~SOCKET_INUSE; - module_put(skt->owner); - } -} - /* * Stuff internal to module "pcmcia_core": -- cgit v0.10.2 From 385ee871092a524869c71a8180888aadcd6ca36d Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Wed, 6 Jan 2010 11:23:58 +0100 Subject: pcmcia: remove useless indirection As release_resoure_db() used to be called only from one place, and it's a two-line function, remove it. Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/cs.c b/drivers/pcmcia/cs.c index 137a5db..43c90f6 100644 --- a/drivers/pcmcia/cs.c +++ b/drivers/pcmcia/cs.c @@ -283,7 +283,8 @@ void pcmcia_unregister_socket(struct pcmcia_socket *socket) up_write(&pcmcia_socket_list_rwsem); /* wait for sysfs to drop all references */ - release_resource_db(socket); + if (socket->resource_ops->exit) + socket->resource_ops->exit(socket); wait_for_completion(&socket->socket_released); } /* pcmcia_unregister_socket */ EXPORT_SYMBOL(pcmcia_unregister_socket); diff --git a/drivers/pcmcia/cs_internal.h b/drivers/pcmcia/cs_internal.h index 9a3bbad..7f86d09 100644 --- a/drivers/pcmcia/cs_internal.h +++ b/drivers/pcmcia/cs_internal.h @@ -95,9 +95,6 @@ struct pccard_resource_ops { /* cistpl.c */ int verify_cis_cache(struct pcmcia_socket *s); -/* rsrc_mgr.c */ -void release_resource_db(struct pcmcia_socket *s); - /* socket_sysfs.c */ extern int pccard_sysfs_add_socket(struct device *dev); extern void pccard_sysfs_remove_socket(struct device *dev); diff --git a/drivers/pcmcia/rsrc_mgr.c b/drivers/pcmcia/rsrc_mgr.c index 52db172..66c7800 100644 --- a/drivers/pcmcia/rsrc_mgr.c +++ b/drivers/pcmcia/rsrc_mgr.c @@ -58,12 +58,6 @@ struct resource *pcmcia_find_mem_region(u_long base, u_long num, u_long align, } EXPORT_SYMBOL(pcmcia_find_mem_region); -void release_resource_db(struct pcmcia_socket *s) -{ - if (s->resource_ops->exit) - s->resource_ops->exit(s); -} - static int static_init(struct pcmcia_socket *s) { -- cgit v0.10.2 From f9c316f4a2d32e4d03497ecb24e1d2309361a5b8 Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Wed, 6 Jan 2010 11:32:22 +0100 Subject: pcmcia: remove some rsrc_mgr indirections Remove rsrc_mgr indirections only used by pcmcia_resource.c Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/cs_internal.h b/drivers/pcmcia/cs_internal.h index 7f86d09..ad05e3b 100644 --- a/drivers/pcmcia/cs_internal.h +++ b/drivers/pcmcia/cs_internal.h @@ -168,14 +168,6 @@ int pccard_get_tuple_data(struct pcmcia_socket *s, tuple_t *tuple); /* rsrc_mgr.c */ int pcmcia_validate_mem(struct pcmcia_socket *s); -struct resource *pcmcia_find_io_region(unsigned long base, - int num, - unsigned long align, - struct pcmcia_socket *s); -int pcmcia_adjust_io_region(struct resource *res, - unsigned long r_start, - unsigned long r_end, - struct pcmcia_socket *s); struct resource *pcmcia_find_mem_region(u_long base, u_long num, u_long align, diff --git a/drivers/pcmcia/pcmcia_resource.c b/drivers/pcmcia/pcmcia_resource.c index d5db956..880b0b6 100644 --- a/drivers/pcmcia/pcmcia_resource.c +++ b/drivers/pcmcia/pcmcia_resource.c @@ -43,6 +43,23 @@ module_param(io_speed, int, 0444); static u8 pcmcia_used_irq[NR_IRQS]; #endif +static int pcmcia_adjust_io_region(struct resource *res, unsigned long start, + unsigned long end, struct pcmcia_socket *s) +{ + if (s->resource_ops->adjust_io_region) + return s->resource_ops->adjust_io_region(res, start, end, s); + return -ENOMEM; +} + +static struct resource *pcmcia_find_io_region(unsigned long base, int num, + unsigned long align, + struct pcmcia_socket *s) +{ + if (s->resource_ops->find_io) + return s->resource_ops->find_io(base, num, align, s); + return NULL; +} + /** alloc_io_space * diff --git a/drivers/pcmcia/rsrc_mgr.c b/drivers/pcmcia/rsrc_mgr.c index 66c7800..81540c4 100644 --- a/drivers/pcmcia/rsrc_mgr.c +++ b/drivers/pcmcia/rsrc_mgr.c @@ -31,24 +31,6 @@ int pcmcia_validate_mem(struct pcmcia_socket *s) } EXPORT_SYMBOL(pcmcia_validate_mem); -int pcmcia_adjust_io_region(struct resource *res, unsigned long r_start, - unsigned long r_end, struct pcmcia_socket *s) -{ - if (s->resource_ops->adjust_io_region) - return s->resource_ops->adjust_io_region(res, r_start, r_end, s); - return -ENOMEM; -} -EXPORT_SYMBOL(pcmcia_adjust_io_region); - -struct resource *pcmcia_find_io_region(unsigned long base, int num, - unsigned long align, struct pcmcia_socket *s) -{ - if (s->resource_ops->find_io) - return s->resource_ops->find_io(base, num, align, s); - return NULL; -} -EXPORT_SYMBOL(pcmcia_find_io_region); - struct resource *pcmcia_find_mem_region(u_long base, u_long num, u_long align, int low, struct pcmcia_socket *s) { -- cgit v0.10.2 From a7eb169dc7292979d78f2d2f1655026ae3a9ff5f Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Wed, 6 Jan 2010 12:18:13 +0100 Subject: pcmcia: m32r uses static socket resources m32r_cfc sets the socket capabilities to SS_CAP_STATIC_MAP and also sets io_offset != 0. This means no calls to &pccard_nonstatic_ops went through. Therfore, replace it with &pccard_static_ops which is exactly for this case. CC: Mamoru Sakugawa CC: Hirokazu Takata Tested-by: Wolfram Sang Signed-off-by: Dominik Brodowski diff --git a/drivers/pcmcia/Kconfig b/drivers/pcmcia/Kconfig index 9f3adbd..7e9fd38 100644 --- a/drivers/pcmcia/Kconfig +++ b/drivers/pcmcia/Kconfig @@ -238,14 +238,12 @@ config PCMCIA_PROBE config M32R_PCC bool "M32R PCMCIA I/F" depends on M32R && CHIP_M32700 && PCMCIA - select PCCARD_NONSTATIC help Say Y here to use the M32R PCMCIA controller. config M32R_CFC bool "M32R CF I/F Controller" depends on M32R && (PLAT_USRV || PLAT_M32700UT || PLAT_MAPPI2 || PLAT_MAPPI3 || PLAT_OPSPUT) - select PCCARD_NONSTATIC help Say Y here to use the M32R CompactFlash controller. diff --git a/drivers/pcmcia/m32r_cfc.c b/drivers/pcmcia/m32r_cfc.c index 26a621c..0ece2cd 100644 --- a/drivers/pcmcia/m32r_cfc.c +++ b/drivers/pcmcia/m32r_cfc.c @@ -764,7 +764,7 @@ static int __init init_m32r_pcc(void) for (i = 0 ; i < pcc_sockets ; i++) { socket[i].socket.dev.parent = &pcc_device.dev; socket[i].socket.ops = &pcc_operations; - socket[i].socket.resource_ops = &pccard_nonstatic_ops; + socket[i].socket.resource_ops = &pccard_static_ops; socket[i].socket.owner = THIS_MODULE; socket[i].number = i; ret = pcmcia_register_socket(&socket[i].socket); -- cgit v0.10.2 From 9df5f74194871ebd0e51ef5ad2eca5084acaaaba Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Mon, 25 Jan 2010 11:42:20 -0600 Subject: mm: add coherence API for DMA to vmalloc/vmap areas On Virtually Indexed architectures (which don't do automatic alias resolution in their caches), we have to flush via the correct virtual address to prepare pages for DMA. On some architectures (like arm) we cannot prevent the CPU from doing data movein along the alias (and thus giving stale read data), so we not only have to introduce a flush API to push dirty cache lines out, but also an invalidate API to kill inconsistent cache lines that may have moved in before DMA changed the data Signed-off-by: James Bottomley diff --git a/Documentation/cachetlb.txt b/Documentation/cachetlb.txt index da42ab4..b231414 100644 --- a/Documentation/cachetlb.txt +++ b/Documentation/cachetlb.txt @@ -377,3 +377,27 @@ maps this page at its virtual address. All the functionality of flush_icache_page can be implemented in flush_dcache_page and update_mmu_cache. In 2.7 the hope is to remove this interface completely. + +The final category of APIs is for I/O to deliberately aliased address +ranges inside the kernel. Such aliases are set up by use of the +vmap/vmalloc API. Since kernel I/O goes via physical pages, the I/O +subsystem assumes that the user mapping and kernel offset mapping are +the only aliases. This isn't true for vmap aliases, so anything in +the kernel trying to do I/O to vmap areas must manually manage +coherency. It must do this by flushing the vmap range before doing +I/O and invalidating it after the I/O returns. + + void flush_kernel_vmap_range(void *vaddr, int size) + flushes the kernel cache for a given virtual address range in + the vmap area. This is to make sure that any data the kernel + modified in the vmap range is made visible to the physical + page. The design is to make this area safe to perform I/O on. + Note that this API does *not* also flush the offset map alias + of the area. + + void invalidate_kernel_vmap_range(void *vaddr, int size) invalidates + the cache for a given virtual address range in the vmap area + which prevents the processor from making the cache stale by + speculatively reading data while the I/O was occurring to the + physical pages. This is only necessary for data reads into the + vmap area. diff --git a/include/linux/highmem.h b/include/linux/highmem.h index 211ff44..adfe101 100644 --- a/include/linux/highmem.h +++ b/include/linux/highmem.h @@ -17,6 +17,12 @@ static inline void flush_anon_page(struct vm_area_struct *vma, struct page *page static inline void flush_kernel_dcache_page(struct page *page) { } +static inline void flush_kernel_vmap_range(void *vaddr, int size) +{ +} +static inline void invalidate_kernel_vmap_range(void *vaddr, int size) +{ +} #endif #include -- cgit v0.10.2 From cd9ec30da58bcd8ab154eba9eb54d16c67e7ef3b Mon Sep 17 00:00:00 2001 From: Johnathon Harris Date: Thu, 21 Jan 2010 14:36:52 +0000 Subject: HID: add support for Ortek WKB-2000 This patch adds a new USB HID driver for the Ortek WKB-2000, working around an incorrect LogicalMaximum value in the USB resource descriptor. Tracked by http://bugzilla.kernel.org/show_bug.cgi?id=14787 Bug originally reported by Ubuntu users: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/405390 Signed-off-by: Johnathon Harris Tested-by: Daniel J Blueman Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index 38e9692..139668d 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -218,6 +218,13 @@ config HID_NTRIG ---help--- Support for N-Trig touch screen. +config HID_ORTEK + tristate "Ortek" if EMBEDDED + depends on USB_HID + default !EMBEDDED + ---help--- + Support for Ortek WKB-2000 wireless keyboard + mouse trackpad. + config HID_PANTHERLORD tristate "Pantherlord support" if EMBEDDED depends on USB_HID diff --git a/drivers/hid/Makefile b/drivers/hid/Makefile index 15541c4..b62d4b3 100644 --- a/drivers/hid/Makefile +++ b/drivers/hid/Makefile @@ -38,6 +38,7 @@ obj-$(CONFIG_HID_LOGITECH) += hid-logitech.o obj-$(CONFIG_HID_MICROSOFT) += hid-microsoft.o obj-$(CONFIG_HID_MONTEREY) += hid-monterey.o obj-$(CONFIG_HID_NTRIG) += hid-ntrig.o +obj-$(CONFIG_HID_ORTEK) += hid-ortek.o obj-$(CONFIG_HID_QUANTA) += hid-quanta.o obj-$(CONFIG_HID_PANTHERLORD) += hid-pl.o obj-$(CONFIG_HID_PETALYNX) += hid-petalynx.o diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index 116a346..7fe5098 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1337,6 +1337,7 @@ static const struct hid_device_id hid_blacklist[] = { { HID_USB_DEVICE(USB_VENDOR_ID_MICROSOFT, USB_DEVICE_ID_WIRELESS_OPTICAL_DESKTOP_3_0) }, { HID_USB_DEVICE(USB_VENDOR_ID_MONTEREY, USB_DEVICE_ID_GENIUS_KB29E) }, { HID_USB_DEVICE(USB_VENDOR_ID_NTRIG, USB_DEVICE_ID_NTRIG_TOUCH_SCREEN) }, + { HID_USB_DEVICE(USB_VENDOR_ID_ORTEK, USB_DEVICE_ID_ORTEK_WKB2000) }, { HID_USB_DEVICE(USB_VENDOR_ID_PETALYNX, USB_DEVICE_ID_PETALYNX_MAXTER_REMOTE) }, { HID_USB_DEVICE(USB_VENDOR_ID_QUANTA, USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH) }, { HID_USB_DEVICE(USB_VENDOR_ID_SAMSUNG, USB_DEVICE_ID_SAMSUNG_IR_REMOTE) }, diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index 064c09a..6c38359 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -363,6 +363,9 @@ #define USB_VENDOR_ID_ONTRAK 0x0a07 #define USB_DEVICE_ID_ONTRAK_ADU100 0x0064 +#define USB_VENDOR_ID_ORTEK 0x05a4 +#define USB_DEVICE_ID_ORTEK_WKB2000 0x2000 + #define USB_VENDOR_ID_PANJIT 0x134c #define USB_VENDOR_ID_PANTHERLORD 0x0810 diff --git a/drivers/hid/hid-ortek.c b/drivers/hid/hid-ortek.c new file mode 100644 index 0000000..aa9a960 --- /dev/null +++ b/drivers/hid/hid-ortek.c @@ -0,0 +1,56 @@ +/* + * HID driver for Ortek WKB-2000 (wireless keyboard + mouse trackpad). + * Fixes LogicalMaximum error in USB report description, see + * http://bugzilla.kernel.org/show_bug.cgi?id=14787 + * + * Copyright (c) 2010 Johnathon Harris + */ + +/* + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include +#include +#include + +#include "hid-ids.h" + +static void ortek_report_fixup(struct hid_device *hdev, __u8 *rdesc, + unsigned int rsize) +{ + if (rsize >= 56 && rdesc[54] == 0x25 && rdesc[55] == 0x01) { + dev_info(&hdev->dev, "Fixing up Ortek WKB-2000 " + "report descriptor.\n"); + rdesc[55] = 0x92; + } +} + +static const struct hid_device_id ortek_devices[] = { + { HID_USB_DEVICE(USB_VENDOR_ID_ORTEK, USB_DEVICE_ID_ORTEK_WKB2000) }, + { } +}; +MODULE_DEVICE_TABLE(hid, ortek_devices); + +static struct hid_driver ortek_driver = { + .name = "ortek", + .id_table = ortek_devices, + .report_fixup = ortek_report_fixup +}; + +static int __init ortek_init(void) +{ + return hid_register_driver(&ortek_driver); +} + +static void __exit ortek_exit(void) +{ + hid_unregister_driver(&ortek_driver); +} + +module_init(ortek_init); +module_exit(ortek_exit); +MODULE_LICENSE("GPL"); -- cgit v0.10.2 From 9d56dd3b083a3bec56e9da35ce07baca81030b03 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 26 Jan 2010 12:58:40 +0900 Subject: sh: Mass ctrl_in/outX to __raw_read/writeX conversion. The old ctrl in/out routines are non-portable and unsuitable for cross-platform use. While drivers/sh has already been sanitized, there is still quite a lot of code that is not. This converts the arch/sh/ bits over, which permits us to flag the routines as deprecated whilst still building with -Werror for the architecture code, and to ensure that future users are not added. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/board-magicpanelr2.c b/arch/sh/boards/board-magicpanelr2.c index 99ffc5f..efba450 100644 --- a/arch/sh/boards/board-magicpanelr2.c +++ b/arch/sh/boards/board-magicpanelr2.c @@ -23,7 +23,7 @@ #include #include -#define LAN9115_READY (ctrl_inl(0xA8000084UL) & 0x00000001UL) +#define LAN9115_READY (__raw_readl(0xA8000084UL) & 0x00000001UL) /* Prefer cmdline over RedBoot */ static const char *probes[] = { "cmdlinepart", "RedBoot", NULL }; @@ -60,33 +60,33 @@ static void __init setup_chip_select(void) { /* CS2: LAN (0x08000000 - 0x0bffffff) */ /* no idle cycles, normal space, 8 bit data bus */ - ctrl_outl(0x36db0400, CS2BCR); + __raw_writel(0x36db0400, CS2BCR); /* (SW:1.5 WR:3 HW:1.5), ext. wait */ - ctrl_outl(0x000003c0, CS2WCR); + __raw_writel(0x000003c0, CS2WCR); /* CS4: CAN1 (0xb0000000 - 0xb3ffffff) */ /* no idle cycles, normal space, 8 bit data bus */ - ctrl_outl(0x00000200, CS4BCR); + __raw_writel(0x00000200, CS4BCR); /* (SW:1.5 WR:3 HW:1.5), ext. wait */ - ctrl_outl(0x00100981, CS4WCR); + __raw_writel(0x00100981, CS4WCR); /* CS5a: CAN2 (0xb4000000 - 0xb5ffffff) */ /* no idle cycles, normal space, 8 bit data bus */ - ctrl_outl(0x00000200, CS5ABCR); + __raw_writel(0x00000200, CS5ABCR); /* (SW:1.5 WR:3 HW:1.5), ext. wait */ - ctrl_outl(0x00100981, CS5AWCR); + __raw_writel(0x00100981, CS5AWCR); /* CS5b: CAN3 (0xb6000000 - 0xb7ffffff) */ /* no idle cycles, normal space, 8 bit data bus */ - ctrl_outl(0x00000200, CS5BBCR); + __raw_writel(0x00000200, CS5BBCR); /* (SW:1.5 WR:3 HW:1.5), ext. wait */ - ctrl_outl(0x00100981, CS5BWCR); + __raw_writel(0x00100981, CS5BWCR); /* CS6a: Rotary (0xb8000000 - 0xb9ffffff) */ /* no idle cycles, normal space, 8 bit data bus */ - ctrl_outl(0x00000200, CS6ABCR); + __raw_writel(0x00000200, CS6ABCR); /* (SW:1.5 WR:3 HW:1.5), no ext. wait */ - ctrl_outl(0x001009C1, CS6AWCR); + __raw_writel(0x001009C1, CS6AWCR); } static void __init setup_port_multiplexing(void) @@ -94,71 +94,71 @@ static void __init setup_port_multiplexing(void) /* A7 GPO(LED8); A6 GPO(LED7); A5 GPO(LED6); A4 GPO(LED5); * A3 GPO(LED4); A2 GPO(LED3); A1 GPO(LED2); A0 GPO(LED1); */ - ctrl_outw(0x5555, PORT_PACR); /* 01 01 01 01 01 01 01 01 */ + __raw_writew(0x5555, PORT_PACR); /* 01 01 01 01 01 01 01 01 */ /* B7 GPO(RST4); B6 GPO(RST3); B5 GPO(RST2); B4 GPO(RST1); * B3 GPO(PB3); B2 GPO(PB2); B1 GPO(PB1); B0 GPO(PB0); */ - ctrl_outw(0x5555, PORT_PBCR); /* 01 01 01 01 01 01 01 01 */ + __raw_writew(0x5555, PORT_PBCR); /* 01 01 01 01 01 01 01 01 */ /* C7 GPO(PC7); C6 GPO(PC6); C5 GPO(PC5); C4 GPO(PC4); * C3 LCD_DATA3; C2 LCD_DATA2; C1 LCD_DATA1; C0 LCD_DATA0; */ - ctrl_outw(0x5500, PORT_PCCR); /* 01 01 01 01 00 00 00 00 */ + __raw_writew(0x5500, PORT_PCCR); /* 01 01 01 01 00 00 00 00 */ /* D7 GPO(PD7); D6 GPO(PD6); D5 GPO(PD5); D4 GPO(PD4); * D3 GPO(PD3); D2 GPO(PD2); D1 GPO(PD1); D0 GPO(PD0); */ - ctrl_outw(0x5555, PORT_PDCR); /* 01 01 01 01 01 01 01 01 */ + __raw_writew(0x5555, PORT_PDCR); /* 01 01 01 01 01 01 01 01 */ /* E7 (x); E6 GPI(nu); E5 GPI(nu); E4 LCD_M_DISP; * E3 LCD_CL1; E2 LCD_CL2; E1 LCD_DON; E0 LCD_FLM; */ - ctrl_outw(0x3C00, PORT_PECR); /* 00 11 11 00 00 00 00 00 */ + __raw_writew(0x3C00, PORT_PECR); /* 00 11 11 00 00 00 00 00 */ /* F7 (x); F6 DA1(VLCD); F5 DA0(nc); F4 AN3; * F3 AN2(MID_AD); F2 AN1(EARTH_AD); F1 AN0(TEMP); F0 GPI+(nc); */ - ctrl_outw(0x0002, PORT_PFCR); /* 00 00 00 00 00 00 00 10 */ + __raw_writew(0x0002, PORT_PFCR); /* 00 00 00 00 00 00 00 10 */ /* G7 (x); G6 IRQ5(TOUCH_BUSY); G5 IRQ4(TOUCH_IRQ); G4 GPI(KEY2); * G3 GPI(KEY1); G2 GPO(LED11); G1 GPO(LED10); G0 GPO(LED9); */ - ctrl_outw(0x03D5, PORT_PGCR); /* 00 00 00 11 11 01 01 01 */ + __raw_writew(0x03D5, PORT_PGCR); /* 00 00 00 11 11 01 01 01 */ /* H7 (x); H6 /RAS(BRAS); H5 /CAS(BCAS); H4 CKE(BCKE); * H3 GPO(EARTH_OFF); H2 GPO(EARTH_TEST); H1 USB2_PWR; H0 USB1_PWR; */ - ctrl_outw(0x0050, PORT_PHCR); /* 00 00 00 00 01 01 00 00 */ + __raw_writew(0x0050, PORT_PHCR); /* 00 00 00 00 01 01 00 00 */ /* J7 (x); J6 AUDCK; J5 ASEBRKAK; J4 AUDATA3; * J3 AUDATA2; J2 AUDATA1; J1 AUDATA0; J0 AUDSYNC; */ - ctrl_outw(0x0000, PORT_PJCR); /* 00 00 00 00 00 00 00 00 */ + __raw_writew(0x0000, PORT_PJCR); /* 00 00 00 00 00 00 00 00 */ /* K7 (x); K6 (x); K5 (x); K4 (x); * K3 PINT7(/PWR2); K2 PINT6(/PWR1); K1 PINT5(nu); K0 PINT4(FLASH_READY) */ - ctrl_outw(0x00FF, PORT_PKCR); /* 00 00 00 00 11 11 11 11 */ + __raw_writew(0x00FF, PORT_PKCR); /* 00 00 00 00 11 11 11 11 */ /* L7 TRST; L6 TMS; L5 TDO; L4 TDI; * L3 TCK; L2 (x); L1 (x); L0 (x); */ - ctrl_outw(0x0000, PORT_PLCR); /* 00 00 00 00 00 00 00 00 */ + __raw_writew(0x0000, PORT_PLCR); /* 00 00 00 00 00 00 00 00 */ /* M7 GPO(CURRENT_SINK); M6 GPO(PWR_SWITCH); M5 GPO(LAN_SPEED); * M4 GPO(LAN_RESET); M3 GPO(BUZZER); M2 GPO(LCD_BL); * M1 CS5B(CAN3_CS); M0 GPI+(nc); */ - ctrl_outw(0x5552, PORT_PMCR); /* 01 01 01 01 01 01 00 10 */ + __raw_writew(0x5552, PORT_PMCR); /* 01 01 01 01 01 01 00 10 */ /* CURRENT_SINK=off, PWR_SWITCH=off, LAN_SPEED=100MBit, * LAN_RESET=off, BUZZER=off, LCD_BL=off */ #if CONFIG_SH_MAGIC_PANEL_R2_VERSION == 2 - ctrl_outb(0x30, PORT_PMDR); + __raw_writeb(0x30, PORT_PMDR); #elif CONFIG_SH_MAGIC_PANEL_R2_VERSION == 3 - ctrl_outb(0xF0, PORT_PMDR); + __raw_writeb(0xF0, PORT_PMDR); #else #error Unknown revision of PLATFORM_MP_R2 #endif @@ -167,8 +167,8 @@ static void __init setup_port_multiplexing(void) * P4 GPO(nu); P3 IRQ3(LAN_IRQ); P2 IRQ2(CAN3_IRQ); * P1 IRQ1(CAN2_IRQ); P0 IRQ0(CAN1_IRQ) */ - ctrl_outw(0x0100, PORT_PPCR); /* 00 00 00 01 00 00 00 00 */ - ctrl_outb(0x10, PORT_PPDR); + __raw_writew(0x0100, PORT_PPCR); /* 00 00 00 01 00 00 00 00 */ + __raw_writeb(0x10, PORT_PPDR); /* R7 A25; R6 A24; R5 A23; R4 A22; * R3 A21; R2 A20; R1 A19; R0 A0; @@ -185,22 +185,22 @@ static void __init setup_port_multiplexing(void) /* S7 (x); S6 (x); S5 (x); S4 GPO(EEPROM_CS2); * S3 GPO(EEPROM_CS1); S2 SIOF0_TXD; S1 SIOF0_RXD; S0 SIOF0_SCK; */ - ctrl_outw(0x0140, PORT_PSCR); /* 00 00 00 01 01 00 00 00 */ + __raw_writew(0x0140, PORT_PSCR); /* 00 00 00 01 01 00 00 00 */ /* T7 (x); T6 (x); T5 (x); T4 COM1_CTS; * T3 COM1_RTS; T2 COM1_TXD; T1 COM1_RXD; T0 GPO(WDOG) */ - ctrl_outw(0x0001, PORT_PTCR); /* 00 00 00 00 00 00 00 01 */ + __raw_writew(0x0001, PORT_PTCR); /* 00 00 00 00 00 00 00 01 */ /* U7 (x); U6 (x); U5 (x); U4 GPI+(/AC_FAULT); * U3 GPO(TOUCH_CS); U2 TOUCH_TXD; U1 TOUCH_RXD; U0 TOUCH_SCK; */ - ctrl_outw(0x0240, PORT_PUCR); /* 00 00 00 10 01 00 00 00 */ + __raw_writew(0x0240, PORT_PUCR); /* 00 00 00 10 01 00 00 00 */ /* V7 (x); V6 (x); V5 (x); V4 GPO(MID2); * V3 GPO(MID1); V2 CARD_TxD; V1 CARD_RxD; V0 GPI+(/BAT_FAULT); */ - ctrl_outw(0x0142, PORT_PVCR); /* 00 00 00 01 01 00 00 10 */ + __raw_writew(0x0142, PORT_PVCR); /* 00 00 00 01 01 00 00 10 */ } static void __init mpr2_setup(char **cmdline_p) @@ -209,24 +209,24 @@ static void __init mpr2_setup(char **cmdline_p) * /PCC_CD1, /PCC_CD2, PCC_BVD1, PCC_BVD2, * /IOIS16, IRQ4, IRQ5, USB1d_SUSPEND */ - ctrl_outw(0xAABC, PORT_PSELA); + __raw_writew(0xAABC, PORT_PSELA); /* set Pin Select Register B: * /SCIF0_RTS, /SCIF0_CTS, LCD_VCPWC, * LCD_VEPWC, IIC_SDA, IIC_SCL, Reserved */ - ctrl_outw(0x3C00, PORT_PSELB); + __raw_writew(0x3C00, PORT_PSELB); /* set Pin Select Register C: * SIOF1_SCK, SIOF1_RxD, SCIF1_RxD, SCIF1_TxD, Reserved */ - ctrl_outw(0x0000, PORT_PSELC); + __raw_writew(0x0000, PORT_PSELC); /* set Pin Select Register D: Reserved, SIOF1_TxD, Reserved, SIOF1_MCLK, * Reserved, SIOF1_SYNC, Reserved, SCIF1_SCK, Reserved */ - ctrl_outw(0x0000, PORT_PSELD); + __raw_writew(0x0000, PORT_PSELD); /* set USB TxRx Control: Reserved, DRV, Reserved, USB_TRANS, USB_SEL */ - ctrl_outw(0x0101, PORT_UTRCTL); + __raw_writew(0x0101, PORT_UTRCTL); /* set USB Clock Control: USSCS, USSTB, Reserved (HighByte always A5) */ - ctrl_outw(0xA5C0, PORT_UCLKCR_W); + __raw_writew(0xA5C0, PORT_UCLKCR_W); setup_chip_select(); diff --git a/arch/sh/boards/board-polaris.c b/arch/sh/boards/board-polaris.c index 5bc1269..5948663 100644 --- a/arch/sh/boards/board-polaris.c +++ b/arch/sh/boards/board-polaris.c @@ -89,15 +89,15 @@ static int __init polaris_initialise(void) printk(KERN_INFO "Configuring Polaris external bus\n"); /* Configure area 5 with 2 wait states */ - wcr = ctrl_inw(WCR2); + wcr = __raw_readw(WCR2); wcr &= (~AREA5_WAIT_CTRL); wcr |= (WAIT_STATES_10 << 10); - ctrl_outw(wcr, WCR2); + __raw_writew(wcr, WCR2); /* Configure area 5 for 32-bit access */ - bcr_mask = ctrl_inw(BCR2); + bcr_mask = __raw_readw(BCR2); bcr_mask |= 1 << 10; - ctrl_outw(bcr_mask, BCR2); + __raw_writew(bcr_mask, BCR2); return platform_add_devices(polaris_devices, ARRAY_SIZE(polaris_devices)); @@ -128,13 +128,13 @@ static struct ipr_desc ipr_irq_desc = { static void __init init_polaris_irq(void) { /* Disable all interrupts */ - ctrl_outw(0, BCR_ILCRA); - ctrl_outw(0, BCR_ILCRB); - ctrl_outw(0, BCR_ILCRC); - ctrl_outw(0, BCR_ILCRD); - ctrl_outw(0, BCR_ILCRE); - ctrl_outw(0, BCR_ILCRF); - ctrl_outw(0, BCR_ILCRG); + __raw_writew(0, BCR_ILCRA); + __raw_writew(0, BCR_ILCRB); + __raw_writew(0, BCR_ILCRC); + __raw_writew(0, BCR_ILCRD); + __raw_writew(0, BCR_ILCRE); + __raw_writew(0, BCR_ILCRF); + __raw_writew(0, BCR_ILCRG); register_ipr_controller(&ipr_irq_desc); } diff --git a/arch/sh/boards/board-shmin.c b/arch/sh/boards/board-shmin.c index b1dcbbc..325bed5 100644 --- a/arch/sh/boards/board-shmin.c +++ b/arch/sh/boards/board-shmin.c @@ -17,8 +17,8 @@ static void __init init_shmin_irq(void) { - ctrl_outw(0x2a00, PFC_PHCR); // IRQ0-3=IRQ - ctrl_outw(0x0aaa, INTC_ICR1); // IRQ0-3=IRQ-mode,Low-active. + __raw_writew(0x2a00, PFC_PHCR); // IRQ0-3=IRQ + __raw_writew(0x0aaa, INTC_ICR1); // IRQ0-3=IRQ-mode,Low-active. plat_irq_setup_pins(IRQ_MODE_IRQ); } diff --git a/arch/sh/boards/mach-ap325rxa/setup.c b/arch/sh/boards/mach-ap325rxa/setup.c index 1f5fa5c..27277cb 100644 --- a/arch/sh/boards/mach-ap325rxa/setup.c +++ b/arch/sh/boards/mach-ap325rxa/setup.c @@ -159,21 +159,21 @@ static void ap320_wvga_power_on(void *board_data) msleep(100); /* ASD AP-320/325 LCD ON */ - ctrl_outw(FPGA_LCDREG_VAL, FPGA_LCDREG); + __raw_writew(FPGA_LCDREG_VAL, FPGA_LCDREG); /* backlight */ gpio_set_value(GPIO_PTS3, 0); - ctrl_outw(0x100, FPGA_BKLREG); + __raw_writew(0x100, FPGA_BKLREG); } static void ap320_wvga_power_off(void *board_data) { /* backlight */ - ctrl_outw(0, FPGA_BKLREG); + __raw_writew(0, FPGA_BKLREG); gpio_set_value(GPIO_PTS3, 1); /* ASD AP-320/325 LCD OFF */ - ctrl_outw(0, FPGA_LCDREG); + __raw_writew(0, FPGA_LCDREG); } static struct sh_mobile_lcdc_info lcdc_info = { @@ -595,7 +595,7 @@ static int __init ap325rxa_devices_setup(void) gpio_request(GPIO_PTZ4, NULL); gpio_direction_output(GPIO_PTZ4, 0); /* SADDR */ - ctrl_outw(ctrl_inw(PORT_MSELCRB) & ~0x0001, PORT_MSELCRB); + __raw_writew(__raw_readw(PORT_MSELCRB) & ~0x0001, PORT_MSELCRB); /* FLCTL */ gpio_request(GPIO_FN_FCE, NULL); @@ -613,9 +613,9 @@ static int __init ap325rxa_devices_setup(void) gpio_request(GPIO_FN_FWE, NULL); gpio_request(GPIO_FN_FRB, NULL); - ctrl_outw(0, PORT_HIZCRC); - ctrl_outw(0xFFFF, PORT_DRVCRA); - ctrl_outw(0xFFFF, PORT_DRVCRB); + __raw_writew(0, PORT_HIZCRC); + __raw_writew(0xFFFF, PORT_DRVCRA); + __raw_writew(0xFFFF, PORT_DRVCRB); platform_resource_setup_memory(&ceu_device, "ceu", 4 << 20); diff --git a/arch/sh/boards/mach-cayman/irq.c b/arch/sh/boards/mach-cayman/irq.c index 33f7708..1394b07 100644 --- a/arch/sh/boards/mach-cayman/irq.c +++ b/arch/sh/boards/mach-cayman/irq.c @@ -66,9 +66,9 @@ static void enable_cayman_irq(unsigned int irq) reg = EPLD_MASK_BASE + ((irq / 8) << 2); bit = 1<<(irq % 8); local_irq_save(flags); - mask = ctrl_inl(reg); + mask = __raw_readl(reg); mask |= bit; - ctrl_outl(mask, reg); + __raw_writel(mask, reg); local_irq_restore(flags); } @@ -83,9 +83,9 @@ void disable_cayman_irq(unsigned int irq) reg = EPLD_MASK_BASE + ((irq / 8) << 2); bit = 1<<(irq % 8); local_irq_save(flags); - mask = ctrl_inl(reg); + mask = __raw_readl(reg); mask &= ~bit; - ctrl_outl(mask, reg); + __raw_writel(mask, reg); local_irq_restore(flags); } @@ -109,8 +109,8 @@ int cayman_irq_demux(int evt) unsigned long status; int i; - status = ctrl_inl(EPLD_STATUS_BASE) & - ctrl_inl(EPLD_MASK_BASE) & 0xff; + status = __raw_readl(EPLD_STATUS_BASE) & + __raw_readl(EPLD_MASK_BASE) & 0xff; if (status == 0) { irq = -1; } else { @@ -126,8 +126,8 @@ int cayman_irq_demux(int evt) unsigned long status; int i; - status = ctrl_inl(EPLD_STATUS_BASE + 3 * sizeof(u32)) & - ctrl_inl(EPLD_MASK_BASE + 3 * sizeof(u32)) & 0xff; + status = __raw_readl(EPLD_STATUS_BASE + 3 * sizeof(u32)) & + __raw_readl(EPLD_MASK_BASE + 3 * sizeof(u32)) & 0xff; if (status == 0) { irq = -1; } else { diff --git a/arch/sh/boards/mach-dreamcast/rtc.c b/arch/sh/boards/mach-dreamcast/rtc.c index a743368..061d657 100644 --- a/arch/sh/boards/mach-dreamcast/rtc.c +++ b/arch/sh/boards/mach-dreamcast/rtc.c @@ -35,11 +35,11 @@ static void aica_rtc_gettimeofday(struct timespec *ts) unsigned long val1, val2; do { - val1 = ((ctrl_inl(AICA_RTC_SECS_H) & 0xffff) << 16) | - (ctrl_inl(AICA_RTC_SECS_L) & 0xffff); + val1 = ((__raw_readl(AICA_RTC_SECS_H) & 0xffff) << 16) | + (__raw_readl(AICA_RTC_SECS_L) & 0xffff); - val2 = ((ctrl_inl(AICA_RTC_SECS_H) & 0xffff) << 16) | - (ctrl_inl(AICA_RTC_SECS_L) & 0xffff); + val2 = ((__raw_readl(AICA_RTC_SECS_H) & 0xffff) << 16) | + (__raw_readl(AICA_RTC_SECS_L) & 0xffff); } while (val1 != val2); ts->tv_sec = val1 - TWENTY_YEARS; @@ -60,14 +60,14 @@ static int aica_rtc_settimeofday(const time_t secs) unsigned long adj = secs + TWENTY_YEARS; do { - ctrl_outl((adj & 0xffff0000) >> 16, AICA_RTC_SECS_H); - ctrl_outl((adj & 0xffff), AICA_RTC_SECS_L); + __raw_writel((adj & 0xffff0000) >> 16, AICA_RTC_SECS_H); + __raw_writel((adj & 0xffff), AICA_RTC_SECS_L); - val1 = ((ctrl_inl(AICA_RTC_SECS_H) & 0xffff) << 16) | - (ctrl_inl(AICA_RTC_SECS_L) & 0xffff); + val1 = ((__raw_readl(AICA_RTC_SECS_H) & 0xffff) << 16) | + (__raw_readl(AICA_RTC_SECS_L) & 0xffff); - val2 = ((ctrl_inl(AICA_RTC_SECS_H) & 0xffff) << 16) | - (ctrl_inl(AICA_RTC_SECS_L) & 0xffff); + val2 = ((__raw_readl(AICA_RTC_SECS_H) & 0xffff) << 16) | + (__raw_readl(AICA_RTC_SECS_L) & 0xffff); } while (val1 != val2); return 0; diff --git a/arch/sh/boards/mach-ecovec24/setup.c b/arch/sh/boards/mach-ecovec24/setup.c index a49cce1..1135c3b 100644 --- a/arch/sh/boards/mach-ecovec24/setup.c +++ b/arch/sh/boards/mach-ecovec24/setup.c @@ -696,13 +696,13 @@ static struct platform_device camera_devices[] = { #define FCLKBCR 0xa415000c static void fsimck_init(struct clk *clk) { - u32 status = ctrl_inl(clk->enable_reg); + u32 status = __raw_readl(clk->enable_reg); /* use external clock */ status &= ~0x000000ff; status |= 0x00000080; - ctrl_outl(status, clk->enable_reg); + __raw_writel(status, clk->enable_reg); } static struct clk_ops fsimck_clk_ops = { @@ -853,7 +853,7 @@ static int __init arch_setup(void) gpio_direction_output(GPIO_PTG1, 0); gpio_direction_output(GPIO_PTG2, 0); gpio_direction_output(GPIO_PTG3, 0); - ctrl_outw((ctrl_inw(PORT_HIZA) & ~(0x1 << 1)) , PORT_HIZA); + __raw_writew((__raw_readw(PORT_HIZA) & ~(0x1 << 1)) , PORT_HIZA); /* enable SH-Eth */ gpio_request(GPIO_PTA1, NULL); @@ -873,16 +873,16 @@ static int __init arch_setup(void) gpio_request(GPIO_FN_LNKSTA, NULL); /* enable USB */ - ctrl_outw(0x0000, 0xA4D80000); - ctrl_outw(0x0000, 0xA4D90000); + __raw_writew(0x0000, 0xA4D80000); + __raw_writew(0x0000, 0xA4D90000); gpio_request(GPIO_PTB3, NULL); gpio_request(GPIO_PTB4, NULL); gpio_request(GPIO_PTB5, NULL); gpio_direction_input(GPIO_PTB3); gpio_direction_output(GPIO_PTB4, 0); gpio_direction_output(GPIO_PTB5, 0); - ctrl_outw(0x0600, 0xa40501d4); - ctrl_outw(0x0600, 0xa4050192); + __raw_writew(0x0600, 0xa40501d4); + __raw_writew(0x0600, 0xa4050192); if (gpio_get_value(GPIO_PTB3)) { printk(KERN_INFO "USB1 function is selected\n"); @@ -923,7 +923,7 @@ static int __init arch_setup(void) gpio_request(GPIO_FN_LCDVSYN, NULL); gpio_request(GPIO_FN_LCDDON, NULL); gpio_request(GPIO_FN_LCDLCLK, NULL); - ctrl_outw((ctrl_inw(PORT_HIZA) & ~0x0001), PORT_HIZA); + __raw_writew((__raw_readw(PORT_HIZA) & ~0x0001), PORT_HIZA); gpio_request(GPIO_PTE6, NULL); gpio_request(GPIO_PTU1, NULL); @@ -935,7 +935,7 @@ static int __init arch_setup(void) gpio_direction_output(GPIO_PTA2, 0); /* I/O buffer drive ability is high */ - ctrl_outw((ctrl_inw(IODRIVEA) & ~0x00c0) | 0x0080 , IODRIVEA); + __raw_writew((__raw_readw(IODRIVEA) & ~0x00c0) | 0x0080 , IODRIVEA); if (gpio_get_value(GPIO_PTE6)) { /* DVI */ @@ -1067,7 +1067,7 @@ static int __init arch_setup(void) gpio_direction_output(GPIO_PTB7, 0); /* I/O buffer drive ability is high for SDHI1 */ - ctrl_outw((ctrl_inw(IODRIVEA) & ~0x3000) | 0x2000 , IODRIVEA); + __raw_writew((__raw_readw(IODRIVEA) & ~0x3000) | 0x2000 , IODRIVEA); #else /* enable MSIOF0 on CN11 (needs DS2.4 set to OFF) */ gpio_request(GPIO_FN_MSIOF0_TXD, NULL); diff --git a/arch/sh/boards/mach-highlander/irq-r7780mp.c b/arch/sh/boards/mach-highlander/irq-r7780mp.c index 83c28bc..9893fd3 100644 --- a/arch/sh/boards/mach-highlander/irq-r7780mp.c +++ b/arch/sh/boards/mach-highlander/irq-r7780mp.c @@ -64,7 +64,7 @@ static DECLARE_INTC_DESC(intc_desc, "r7780mp", vectors, unsigned char * __init highlander_plat_irq_setup(void) { - if ((ctrl_inw(0xa4000700) & 0xf000) == 0x2000) { + if ((__raw_readw(0xa4000700) & 0xf000) == 0x2000) { printk(KERN_INFO "Using r7780mp interrupt controller.\n"); register_intc_controller(&intc_desc); return irl2irq; diff --git a/arch/sh/boards/mach-highlander/irq-r7780rp.c b/arch/sh/boards/mach-highlander/irq-r7780rp.c index b721e86..0805b21 100644 --- a/arch/sh/boards/mach-highlander/irq-r7780rp.c +++ b/arch/sh/boards/mach-highlander/irq-r7780rp.c @@ -57,7 +57,7 @@ static DECLARE_INTC_DESC(intc_desc, "r7780rp", vectors, unsigned char * __init highlander_plat_irq_setup(void) { - if (ctrl_inw(0xa5000600)) { + if (__raw_readw(0xa5000600)) { printk(KERN_INFO "Using r7780rp interrupt controller.\n"); register_intc_controller(&intc_desc); return irl2irq; diff --git a/arch/sh/boards/mach-highlander/irq-r7785rp.c b/arch/sh/boards/mach-highlander/irq-r7785rp.c index 3811b06..558b248 100644 --- a/arch/sh/boards/mach-highlander/irq-r7785rp.c +++ b/arch/sh/boards/mach-highlander/irq-r7785rp.c @@ -66,20 +66,20 @@ static DECLARE_INTC_DESC(intc_desc, "r7785rp", vectors, unsigned char * __init highlander_plat_irq_setup(void) { - if ((ctrl_inw(0xa4000158) & 0xf000) != 0x1000) + if ((__raw_readw(0xa4000158) & 0xf000) != 0x1000) return NULL; printk(KERN_INFO "Using r7785rp interrupt controller.\n"); - ctrl_outw(0x0000, PA_IRLSSR1); /* FPGA IRLSSR1(CF_CD clear) */ + __raw_writew(0x0000, PA_IRLSSR1); /* FPGA IRLSSR1(CF_CD clear) */ /* Setup the FPGA IRL */ - ctrl_outw(0x0000, PA_IRLPRA); /* FPGA IRLA */ - ctrl_outw(0xe598, PA_IRLPRB); /* FPGA IRLB */ - ctrl_outw(0x7060, PA_IRLPRC); /* FPGA IRLC */ - ctrl_outw(0x0000, PA_IRLPRD); /* FPGA IRLD */ - ctrl_outw(0x4321, PA_IRLPRE); /* FPGA IRLE */ - ctrl_outw(0xdcba, PA_IRLPRF); /* FPGA IRLF */ + __raw_writew(0x0000, PA_IRLPRA); /* FPGA IRLA */ + __raw_writew(0xe598, PA_IRLPRB); /* FPGA IRLB */ + __raw_writew(0x7060, PA_IRLPRC); /* FPGA IRLC */ + __raw_writew(0x0000, PA_IRLPRD); /* FPGA IRLD */ + __raw_writew(0x4321, PA_IRLPRE); /* FPGA IRLE */ + __raw_writew(0xdcba, PA_IRLPRF); /* FPGA IRLF */ register_intc_controller(&intc_desc); return irl2irq; diff --git a/arch/sh/boards/mach-highlander/psw.c b/arch/sh/boards/mach-highlander/psw.c index 37b1a2e..5227863 100644 --- a/arch/sh/boards/mach-highlander/psw.c +++ b/arch/sh/boards/mach-highlander/psw.c @@ -24,7 +24,7 @@ static irqreturn_t psw_irq_handler(int irq, void *arg) unsigned int l, mask; int ret = 0; - l = ctrl_inw(PA_DBSW); + l = __raw_readw(PA_DBSW); /* Nothing to do if there's no state change */ if (psw->state) { @@ -45,7 +45,7 @@ static irqreturn_t psw_irq_handler(int irq, void *arg) out: /* Clear the switch IRQs */ l |= (0x7 << 12); - ctrl_outw(l, PA_DBSW); + __raw_writew(l, PA_DBSW); return IRQ_RETVAL(ret); } diff --git a/arch/sh/boards/mach-highlander/setup.c b/arch/sh/boards/mach-highlander/setup.c index f663c14..affd667 100644 --- a/arch/sh/boards/mach-highlander/setup.c +++ b/arch/sh/boards/mach-highlander/setup.c @@ -311,13 +311,13 @@ device_initcall(r7780rp_devices_setup); */ static int ivdr_clk_enable(struct clk *clk) { - ctrl_outw(ctrl_inw(PA_IVDRCTL) | (1 << IVDR_CK_ON), PA_IVDRCTL); + __raw_writew(__raw_readw(PA_IVDRCTL) | (1 << IVDR_CK_ON), PA_IVDRCTL); return 0; } static void ivdr_clk_disable(struct clk *clk) { - ctrl_outw(ctrl_inw(PA_IVDRCTL) & ~(1 << IVDR_CK_ON), PA_IVDRCTL); + __raw_writew(__raw_readw(PA_IVDRCTL) & ~(1 << IVDR_CK_ON), PA_IVDRCTL); } static struct clk_ops ivdr_clk_ops = { @@ -337,7 +337,7 @@ static struct clk *r7780rp_clocks[] = { static void r7780rp_power_off(void) { if (mach_is_r7780mp() || mach_is_r7785rp()) - ctrl_outw(0x0001, PA_POFF); + __raw_writew(0x0001, PA_POFF); } /* @@ -345,7 +345,7 @@ static void r7780rp_power_off(void) */ static void __init highlander_setup(char **cmdline_p) { - u16 ver = ctrl_inw(PA_VERREG); + u16 ver = __raw_readw(PA_VERREG); int i; printk(KERN_INFO "Renesas Solutions Highlander %s support.\n", @@ -370,12 +370,12 @@ static void __init highlander_setup(char **cmdline_p) clk_enable(clk); } - ctrl_outw(0x0000, PA_OBLED); /* Clear LED. */ + __raw_writew(0x0000, PA_OBLED); /* Clear LED. */ if (mach_is_r7780rp()) - ctrl_outw(0x0001, PA_SDPOW); /* SD Power ON */ + __raw_writew(0x0001, PA_SDPOW); /* SD Power ON */ - ctrl_outw(ctrl_inw(PA_IVDRCTL) | 0x01, PA_IVDRCTL); /* Si13112 */ + __raw_writew(__raw_readw(PA_IVDRCTL) | 0x01, PA_IVDRCTL); /* Si13112 */ pm_power_off = r7780rp_power_off; } diff --git a/arch/sh/boards/mach-hp6xx/hp6xx_apm.c b/arch/sh/boards/mach-hp6xx/hp6xx_apm.c index e85212f..b49535c 100644 --- a/arch/sh/boards/mach-hp6xx/hp6xx_apm.c +++ b/arch/sh/boards/mach-hp6xx/hp6xx_apm.c @@ -53,7 +53,7 @@ static void hp6x0_apm_get_power_status(struct apm_power_info *info) info->ac_line_status = (battery > HP680_BATTERY_AC_ON) ? APM_AC_ONLINE : APM_AC_OFFLINE; - pgdr = ctrl_inb(PGDR); + pgdr = __raw_readb(PGDR); if (pgdr & PGDR_MAIN_BATTERY_OUT) { info->battery_status = APM_BATTERY_STATUS_NOT_PRESENT; info->battery_flag = 0x80; diff --git a/arch/sh/boards/mach-hp6xx/pm.c b/arch/sh/boards/mach-hp6xx/pm.c index d936c1a..4499a37 100644 --- a/arch/sh/boards/mach-hp6xx/pm.c +++ b/arch/sh/boards/mach-hp6xx/pm.c @@ -53,17 +53,17 @@ static void pm_enter(void) sh_wdt_write_cnt(0); /* disable PLL1 */ - frqcr = ctrl_inw(FRQCR); + frqcr = __raw_readw(FRQCR); frqcr &= ~(FRQCR_PLLEN | FRQCR_PSTBY); - ctrl_outw(frqcr, FRQCR); + __raw_writew(frqcr, FRQCR); /* enable standby */ - stbcr = ctrl_inb(STBCR); - ctrl_outb(stbcr | STBCR_STBY | STBCR_MSTP2, STBCR); + stbcr = __raw_readb(STBCR); + __raw_writeb(stbcr | STBCR_STBY | STBCR_MSTP2, STBCR); /* set self-refresh */ - mcr = ctrl_inw(MCR); - ctrl_outw(mcr & ~MCR_RFSH, MCR); + mcr = __raw_readw(MCR); + __raw_writew(mcr & ~MCR_RFSH, MCR); /* set interrupt handler */ asm volatile("stc vbr, %0" : "=r" (vbr_old)); @@ -73,8 +73,8 @@ static void pm_enter(void) &wakeup_start, &wakeup_end - &wakeup_start); asm volatile("ldc %0, vbr" : : "r" (vbr_new)); - ctrl_outw(0, RTCNT); - ctrl_outw(mcr | MCR_RFSH | MCR_RMODE, MCR); + __raw_writew(0, RTCNT); + __raw_writew(mcr | MCR_RFSH | MCR_RMODE, MCR); cpu_sleep(); @@ -83,14 +83,14 @@ static void pm_enter(void) free_page(vbr_new); /* enable PLL1 */ - frqcr = ctrl_inw(FRQCR); + frqcr = __raw_readw(FRQCR); frqcr |= FRQCR_PSTBY; - ctrl_outw(frqcr, FRQCR); + __raw_writew(frqcr, FRQCR); udelay(50); frqcr |= FRQCR_PLLEN; - ctrl_outw(frqcr, FRQCR); + __raw_writew(frqcr, FRQCR); - ctrl_outb(stbcr, STBCR); + __raw_writeb(stbcr, STBCR); clear_bl_bit(); } @@ -115,21 +115,21 @@ static int hp6x0_pm_enter(suspend_state_t state) outw(hd64461_stbcr, HD64461_STBCR); #endif - ctrl_outb(0x1f, DACR); + __raw_writeb(0x1f, DACR); - stbcr = ctrl_inb(STBCR); - ctrl_outb(0x01, STBCR); + stbcr = __raw_readb(STBCR); + __raw_writeb(0x01, STBCR); - stbcr2 = ctrl_inb(STBCR2); - ctrl_outb(0x7f , STBCR2); + stbcr2 = __raw_readb(STBCR2); + __raw_writeb(0x7f , STBCR2); outw(0xf07f, HD64461_SCPUCR); pm_enter(); outw(0, HD64461_SCPUCR); - ctrl_outb(stbcr, STBCR); - ctrl_outb(stbcr2, STBCR2); + __raw_writeb(stbcr, STBCR); + __raw_writeb(stbcr2, STBCR2); #ifdef CONFIG_HD64461_ENABLER hd64461_stbcr = inw(HD64461_STBCR); diff --git a/arch/sh/boards/mach-hp6xx/setup.c b/arch/sh/boards/mach-hp6xx/setup.c index e6dd5e9..8c9add5 100644 --- a/arch/sh/boards/mach-hp6xx/setup.c +++ b/arch/sh/boards/mach-hp6xx/setup.c @@ -149,19 +149,19 @@ static void __init hp6xx_setup(char **cmdline_p) sh_dac_output(0, DAC_SPEAKER_VOLUME); sh_dac_disable(DAC_SPEAKER_VOLUME); - v8 = ctrl_inb(DACR); + v8 = __raw_readb(DACR); v8 &= ~DACR_DAE; - ctrl_outb(v8,DACR); + __raw_writeb(v8,DACR); - v8 = ctrl_inb(SCPDR); + v8 = __raw_readb(SCPDR); v8 |= SCPDR_TS_SCAN_X | SCPDR_TS_SCAN_Y; v8 &= ~SCPDR_TS_SCAN_ENABLE; - ctrl_outb(v8, SCPDR); + __raw_writeb(v8, SCPDR); - v = ctrl_inw(SCPCR); + v = __raw_readw(SCPCR); v &= ~SCPCR_TS_MASK; v |= SCPCR_TS_ENABLE; - ctrl_outw(v, SCPCR); + __raw_writew(v, SCPCR); } device_initcall(hp6xx_devices_setup); diff --git a/arch/sh/boards/mach-kfr2r09/setup.c b/arch/sh/boards/mach-kfr2r09/setup.c index 5d7b5d9..ca9e82d 100644 --- a/arch/sh/boards/mach-kfr2r09/setup.c +++ b/arch/sh/boards/mach-kfr2r09/setup.c @@ -282,7 +282,7 @@ static int camera_power(struct device *dev, int mode) * use 1.8 V for VccQ_VIO * use 2.85V for VccQ_SR */ - ctrl_outw((ctrl_inw(DRVCRB) & ~0x0003) | 0x0001, DRVCRB); + __raw_writew((__raw_readw(DRVCRB) & ~0x0003) | 0x0001, DRVCRB); /* reset clear */ ret = gpio_request(GPIO_PTB4, NULL); @@ -492,13 +492,13 @@ static int kfr2r09_usb0_gadget_setup(void) if (kfr2r09_usb0_gadget_i2c_setup() != 0) return -ENODEV; /* unable to configure using i2c */ - ctrl_outw((ctrl_inw(PORT_MSELCRB) & ~0xc000) | 0x8000, PORT_MSELCRB); + __raw_writew((__raw_readw(PORT_MSELCRB) & ~0xc000) | 0x8000, PORT_MSELCRB); gpio_request(GPIO_FN_PDSTATUS, NULL); /* R-standby disables USB clock */ gpio_request(GPIO_PTV6, NULL); /* USBCLK_ON */ gpio_direction_output(GPIO_PTV6, 1); /* USBCLK_ON = H */ msleep(20); /* wait 20ms to let the clock settle */ clk_enable(clk_get(NULL, "usb0")); - ctrl_outw(0x0600, 0xa40501d4); + __raw_writew(0x0600, 0xa40501d4); return 0; } @@ -526,12 +526,12 @@ static int __init kfr2r09_devices_setup(void) gpio_direction_output(GPIO_PTG3, 1); /* HPON_ON = H */ /* setup NOR flash at CS0 */ - ctrl_outl(0x36db0400, BSC_CS0BCR); - ctrl_outl(0x00000500, BSC_CS0WCR); + __raw_writel(0x36db0400, BSC_CS0BCR); + __raw_writel(0x00000500, BSC_CS0WCR); /* setup NAND flash at CS4 */ - ctrl_outl(0x36db0400, BSC_CS4BCR); - ctrl_outl(0x00000500, BSC_CS4WCR); + __raw_writel(0x36db0400, BSC_CS4BCR); + __raw_writel(0x00000500, BSC_CS4WCR); /* setup KEYSC pins */ gpio_request(GPIO_FN_KEYOUT0, NULL); diff --git a/arch/sh/boards/mach-landisk/gio.c b/arch/sh/boards/mach-landisk/gio.c index 5280131..01e6abb 100644 --- a/arch/sh/boards/mach-landisk/gio.c +++ b/arch/sh/boards/mach-landisk/gio.c @@ -76,39 +76,39 @@ static long gio_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) break; case GIODRV_IOCSGIODATA1: /* write byte */ - ctrl_outb((unsigned char)(0x0ff & data), addr); + __raw_writeb((unsigned char)(0x0ff & data), addr); break; case GIODRV_IOCSGIODATA2: /* write word */ if (addr & 0x01) { return -EFAULT; } - ctrl_outw((unsigned short int)(0x0ffff & data), addr); + __raw_writew((unsigned short int)(0x0ffff & data), addr); break; case GIODRV_IOCSGIODATA4: /* write long */ if (addr & 0x03) { return -EFAULT; } - ctrl_outl(data, addr); + __raw_writel(data, addr); break; case GIODRV_IOCGGIODATA1: /* read byte */ - data = ctrl_inb(addr); + data = __raw_readb(addr); break; case GIODRV_IOCGGIODATA2: /* read word */ if (addr & 0x01) { return -EFAULT; } - data = ctrl_inw(addr); + data = __raw_readw(addr); break; case GIODRV_IOCGGIODATA4: /* read long */ if (addr & 0x03) { return -EFAULT; } - data = ctrl_inl(addr); + data = __raw_readl(addr); break; default: return -EFAULT; diff --git a/arch/sh/boards/mach-landisk/irq.c b/arch/sh/boards/mach-landisk/irq.c index 7b284cd..96f38a4 100644 --- a/arch/sh/boards/mach-landisk/irq.c +++ b/arch/sh/boards/mach-landisk/irq.c @@ -22,14 +22,14 @@ static void disable_landisk_irq(unsigned int irq) { unsigned char mask = 0xff ^ (0x01 << (irq - 5)); - ctrl_outb(ctrl_inb(PA_IMASK) & mask, PA_IMASK); + __raw_writeb(__raw_readb(PA_IMASK) & mask, PA_IMASK); } static void enable_landisk_irq(unsigned int irq) { unsigned char value = (0x01 << (irq - 5)); - ctrl_outb(ctrl_inb(PA_IMASK) | value, PA_IMASK); + __raw_writeb(__raw_readb(PA_IMASK) | value, PA_IMASK); } static struct irq_chip landisk_irq_chip __read_mostly = { @@ -52,5 +52,5 @@ void __init init_landisk_IRQ(void) handle_level_irq, "level"); enable_landisk_irq(i); } - ctrl_outb(0x00, PA_PWRINT_CLR); + __raw_writeb(0x00, PA_PWRINT_CLR); } diff --git a/arch/sh/boards/mach-landisk/psw.c b/arch/sh/boards/mach-landisk/psw.c index e6b0efa..bef8352 100644 --- a/arch/sh/boards/mach-landisk/psw.c +++ b/arch/sh/boards/mach-landisk/psw.c @@ -25,7 +25,7 @@ static irqreturn_t psw_irq_handler(int irq, void *arg) unsigned int sw_value; int ret = 0; - sw_value = (0x0ff & (~ctrl_inb(PA_STATUS))); + sw_value = (0x0ff & (~__raw_readb(PA_STATUS))); /* Nothing to do if there's no state change */ if (psw->state) { @@ -42,7 +42,7 @@ static irqreturn_t psw_irq_handler(int irq, void *arg) out: /* Clear the switch IRQs */ - ctrl_outb(0x00, PA_PWRINT_CLR); + __raw_writeb(0x00, PA_PWRINT_CLR); return IRQ_RETVAL(ret); } diff --git a/arch/sh/boards/mach-landisk/setup.c b/arch/sh/boards/mach-landisk/setup.c index 2d09d4d..50337acc 100644 --- a/arch/sh/boards/mach-landisk/setup.c +++ b/arch/sh/boards/mach-landisk/setup.c @@ -25,7 +25,7 @@ void init_landisk_IRQ(void); static void landisk_power_off(void) { - ctrl_outb(0x01, PA_SHUTDOWN); + __raw_writeb(0x01, PA_SHUTDOWN); } static struct resource cf_ide_resources[3]; @@ -88,7 +88,7 @@ __initcall(landisk_devices_setup); static void __init landisk_setup(char **cmdline_p) { /* LED ON */ - ctrl_outb(ctrl_inb(PA_LED) | 0x03, PA_LED); + __raw_writeb(__raw_readb(PA_LED) | 0x03, PA_LED); printk(KERN_INFO "I-O DATA DEVICE, INC. \"LANDISK Series\" support.\n"); pm_power_off = landisk_power_off; diff --git a/arch/sh/boards/mach-microdev/io.c b/arch/sh/boards/mach-microdev/io.c index 52dd748..2960c65 100644 --- a/arch/sh/boards/mach-microdev/io.c +++ b/arch/sh/boards/mach-microdev/io.c @@ -141,10 +141,10 @@ static inline void delay(void) #if defined(CONFIG_PCI) /* System board present, just make a dummy SRAM access. (CS0 will be mapped to PCI memory, probably good to avoid it.) */ - ctrl_inw(0xa6800000); + __raw_readw(0xa6800000); #else /* CS0 will be mapped to flash, ROM etc so safe to access it. */ - ctrl_inw(0xa0000000); + __raw_readw(0xa0000000); #endif } diff --git a/arch/sh/boards/mach-microdev/irq.c b/arch/sh/boards/mach-microdev/irq.c index b551963..a26d166 100644 --- a/arch/sh/boards/mach-microdev/irq.c +++ b/arch/sh/boards/mach-microdev/irq.c @@ -88,7 +88,7 @@ static void disable_microdev_irq(unsigned int irq) fpgaIrq = fpgaIrqTable[irq].fpgaIrq; /* disable interrupts on the FPGA INTC register */ - ctrl_outl(MICRODEV_FPGA_INTC_MASK(fpgaIrq), MICRODEV_FPGA_INTDSB_REG); + __raw_writel(MICRODEV_FPGA_INTC_MASK(fpgaIrq), MICRODEV_FPGA_INTDSB_REG); } static void enable_microdev_irq(unsigned int irq) @@ -107,13 +107,13 @@ static void enable_microdev_irq(unsigned int irq) priorityReg = MICRODEV_FPGA_INTPRI_REG(fpgaIrq); /* set priority for the interrupt */ - priorities = ctrl_inl(priorityReg); + priorities = __raw_readl(priorityReg); priorities &= ~MICRODEV_FPGA_INTPRI_MASK(fpgaIrq); priorities |= MICRODEV_FPGA_INTPRI_LEVEL(fpgaIrq, pri); - ctrl_outl(priorities, priorityReg); + __raw_writel(priorities, priorityReg); /* enable interrupts on the FPGA INTC register */ - ctrl_outl(MICRODEV_FPGA_INTC_MASK(fpgaIrq), MICRODEV_FPGA_INTENB_REG); + __raw_writel(MICRODEV_FPGA_INTC_MASK(fpgaIrq), MICRODEV_FPGA_INTENB_REG); } /* This function sets the desired irq handler to be a MicroDev type */ @@ -134,7 +134,7 @@ extern void __init init_microdev_irq(void) int i; /* disable interrupts on the FPGA INTC register */ - ctrl_outl(~0ul, MICRODEV_FPGA_INTDSB_REG); + __raw_writel(~0ul, MICRODEV_FPGA_INTDSB_REG); for (i = 0; i < NUM_EXTERNAL_IRQS; i++) make_microdev_irq(i); diff --git a/arch/sh/boards/mach-migor/setup.c b/arch/sh/boards/mach-migor/setup.c index 507c77b..ed2eeeb 100644 --- a/arch/sh/boards/mach-migor/setup.c +++ b/arch/sh/boards/mach-migor/setup.c @@ -516,8 +516,8 @@ static int __init migor_devices_setup(void) /* SMC91C111 - Enable IRQ0, Setup CS4 for 16-bit fast access */ gpio_request(GPIO_FN_IRQ0, NULL); - ctrl_outl(0x00003400, BSC_CS4BCR); - ctrl_outl(0x00110080, BSC_CS4WCR); + __raw_writel(0x00003400, BSC_CS4BCR); + __raw_writel(0x00110080, BSC_CS4WCR); /* KEYSC */ gpio_request(GPIO_FN_KEYOUT0, NULL); @@ -533,7 +533,7 @@ static int __init migor_devices_setup(void) /* NAND Flash */ gpio_request(GPIO_FN_CS6A_CE2B, NULL); - ctrl_outl((ctrl_inl(BSC_CS6ABCR) & ~0x0600) | 0x0200, BSC_CS6ABCR); + __raw_writel((__raw_readl(BSC_CS6ABCR) & ~0x0600) | 0x0200, BSC_CS6ABCR); gpio_request(GPIO_PTA1, NULL); gpio_direction_input(GPIO_PTA1); @@ -627,7 +627,7 @@ static int __init migor_devices_setup(void) #else gpio_direction_output(GPIO_PTT0, 1); #endif - ctrl_outw(ctrl_inw(PORT_MSELCRB) | 0x2000, PORT_MSELCRB); /* D15->D8 */ + __raw_writew(__raw_readw(PORT_MSELCRB) | 0x2000, PORT_MSELCRB); /* D15->D8 */ platform_resource_setup_memory(&migor_ceu_device, "ceu", 4 << 20); diff --git a/arch/sh/boards/mach-r2d/irq.c b/arch/sh/boards/mach-r2d/irq.c index 78d7b27..574f009 100644 --- a/arch/sh/boards/mach-r2d/irq.c +++ b/arch/sh/boards/mach-r2d/irq.c @@ -129,7 +129,7 @@ void __init init_rts7751r2d_IRQ(void) { struct intc_desc *d; - switch (ctrl_inw(PA_VERREG) & 0xf0) { + switch (__raw_readw(PA_VERREG) & 0xf0) { #ifdef CONFIG_RTS7751R2D_PLUS case 0x10: printk(KERN_INFO "Using R2D-PLUS interrupt controller.\n"); @@ -147,7 +147,7 @@ void __init init_rts7751r2d_IRQ(void) #endif default: printk(KERN_INFO "Unknown R2D interrupt controller 0x%04x\n", - ctrl_inw(PA_VERREG)); + __raw_readw(PA_VERREG)); return; } diff --git a/arch/sh/boards/mach-r2d/setup.c b/arch/sh/boards/mach-r2d/setup.c index a625ecb..b84df6a 100644 --- a/arch/sh/boards/mach-r2d/setup.c +++ b/arch/sh/boards/mach-r2d/setup.c @@ -70,7 +70,7 @@ static struct spi_board_info spi_bus[] = { static void r2d_chip_select(struct sh_spi_info *spi, int cs, int state) { BUG_ON(cs != 0); /* Single Epson RTC-9701JE attached on CS0 */ - ctrl_outw(state == BITBANG_CS_ACTIVE, PA_RTCCE); + __raw_writew(state == BITBANG_CS_ACTIVE, PA_RTCCE); } static struct sh_spi_info spi_info = { @@ -262,7 +262,7 @@ __initcall(rts7751r2d_devices_setup); static void rts7751r2d_power_off(void) { - ctrl_outw(0x0001, PA_POWOFF); + __raw_writew(0x0001, PA_POWOFF); } /* @@ -271,14 +271,14 @@ static void rts7751r2d_power_off(void) static void __init rts7751r2d_setup(char **cmdline_p) { void __iomem *sm501_reg; - u16 ver = ctrl_inw(PA_VERREG); + u16 ver = __raw_readw(PA_VERREG); printk(KERN_INFO "Renesas Technology Sales RTS7751R2D support.\n"); printk(KERN_INFO "FPGA version:%d (revision:%d)\n", (ver >> 4) & 0xf, ver & 0xf); - ctrl_outw(0x0000, PA_OUTPORT); + __raw_writew(0x0000, PA_OUTPORT); pm_power_off = rts7751r2d_power_off; /* sm501 dram configuration: diff --git a/arch/sh/boards/mach-rsk/devices-rsk7203.c b/arch/sh/boards/mach-rsk/devices-rsk7203.c index c37617e..4fa08ba 100644 --- a/arch/sh/boards/mach-rsk/devices-rsk7203.c +++ b/arch/sh/boards/mach-rsk/devices-rsk7203.c @@ -96,7 +96,7 @@ static int __init rsk7203_devices_setup(void) gpio_request(GPIO_FN_RXD0, NULL); /* Setup LAN9118: CS1 in 16-bit Big Endian Mode, IRQ0 at Port B */ - ctrl_outl(0x36db0400, 0xfffc0008); /* CS1BCR */ + __raw_writel(0x36db0400, 0xfffc0008); /* CS1BCR */ gpio_request(GPIO_FN_IRQ0_PB, NULL); return platform_add_devices(rsk7203_devices, diff --git a/arch/sh/boards/mach-sdk7780/irq.c b/arch/sh/boards/mach-sdk7780/irq.c index 8555581..e5f7564 100644 --- a/arch/sh/boards/mach-sdk7780/irq.c +++ b/arch/sh/boards/mach-sdk7780/irq.c @@ -37,9 +37,9 @@ void __init init_sdk7780_IRQ(void) { printk(KERN_INFO "Using SDK7780 interrupt controller.\n"); - ctrl_outw(0xFFFF, FPGA_IRQ0MR); + __raw_writew(0xFFFF, FPGA_IRQ0MR); /* Setup IRL 0-3 */ - ctrl_outw(0x0003, FPGA_IMSR); + __raw_writew(0x0003, FPGA_IMSR); plat_irq_setup_pins(IRQ_MODE_IRL3210); register_intc_controller(&fpga_intc_desc); diff --git a/arch/sh/boards/mach-sdk7780/setup.c b/arch/sh/boards/mach-sdk7780/setup.c index b887373..4da38db 100644 --- a/arch/sh/boards/mach-sdk7780/setup.c +++ b/arch/sh/boards/mach-sdk7780/setup.c @@ -74,8 +74,8 @@ device_initcall(sdk7780_devices_setup); static void __init sdk7780_setup(char **cmdline_p) { - u16 ver = ctrl_inw(FPGA_FPVERR); - u16 dateStamp = ctrl_inw(FPGA_FPDATER); + u16 ver = __raw_readw(FPGA_FPVERR); + u16 dateStamp = __raw_readw(FPGA_FPDATER); printk(KERN_INFO "Renesas Technology Europe SDK7780 support.\n"); printk(KERN_INFO "Board version: %d (revision %d), " @@ -85,7 +85,7 @@ static void __init sdk7780_setup(char **cmdline_p) dateStamp); /* Setup pin mux'ing for PCIC */ - ctrl_outw(0x0000, GPIO_PECR); + __raw_writew(0x0000, GPIO_PECR); } /* diff --git a/arch/sh/boards/mach-se/7206/io.c b/arch/sh/boards/mach-se/7206/io.c index 1804556..adadc77 100644 --- a/arch/sh/boards/mach-se/7206/io.c +++ b/arch/sh/boards/mach-se/7206/io.c @@ -16,7 +16,7 @@ static inline void delay(void) { - ctrl_inw(0x20000000); /* P2 ROM Area */ + __raw_readw(0x20000000); /* P2 ROM Area */ } /* MS7750 requires special versions of in*, out* routines, since diff --git a/arch/sh/boards/mach-se/7206/irq.c b/arch/sh/boards/mach-se/7206/irq.c index aef7f05..79be4bc 100644 --- a/arch/sh/boards/mach-se/7206/irq.c +++ b/arch/sh/boards/mach-se/7206/irq.c @@ -32,12 +32,12 @@ static void disable_se7206_irq(unsigned int irq) unsigned short msk0,msk1; /* Set the priority in IPR to 0 */ - val = ctrl_inw(INTC_IPR01); + val = __raw_readw(INTC_IPR01); val &= mask; - ctrl_outw(val, INTC_IPR01); + __raw_writew(val, INTC_IPR01); /* FPGA mask set */ - msk0 = ctrl_inw(INTMSK0); - msk1 = ctrl_inw(INTMSK1); + msk0 = __raw_readw(INTMSK0); + msk1 = __raw_readw(INTMSK1); switch (irq) { case IRQ0_IRQ: @@ -51,8 +51,8 @@ static void disable_se7206_irq(unsigned int irq) msk1 |= 0x00ff; break; } - ctrl_outw(msk0, INTMSK0); - ctrl_outw(msk1, INTMSK1); + __raw_writew(msk0, INTMSK0); + __raw_writew(msk1, INTMSK1); } static void enable_se7206_irq(unsigned int irq) @@ -62,13 +62,13 @@ static void enable_se7206_irq(unsigned int irq) unsigned short msk0,msk1; /* Set priority in IPR back to original value */ - val = ctrl_inw(INTC_IPR01); + val = __raw_readw(INTC_IPR01); val |= value; - ctrl_outw(val, INTC_IPR01); + __raw_writew(val, INTC_IPR01); /* FPGA mask reset */ - msk0 = ctrl_inw(INTMSK0); - msk1 = ctrl_inw(INTMSK1); + msk0 = __raw_readw(INTMSK0); + msk1 = __raw_readw(INTMSK1); switch (irq) { case IRQ0_IRQ: @@ -82,8 +82,8 @@ static void enable_se7206_irq(unsigned int irq) msk1 &= ~0x00ff; break; } - ctrl_outw(msk0, INTMSK0); - ctrl_outw(msk1, INTMSK1); + __raw_writew(msk0, INTMSK0); + __raw_writew(msk1, INTMSK1); } static void eoi_se7206_irq(unsigned int irq) @@ -93,8 +93,8 @@ static void eoi_se7206_irq(unsigned int irq) if (!(irq_desc[irq].status & (IRQ_DISABLED|IRQ_INPROGRESS))) enable_se7206_irq(irq); /* FPGA isr clear */ - sts0 = ctrl_inw(INTSTS0); - sts1 = ctrl_inw(INTSTS1); + sts0 = __raw_readw(INTSTS0); + sts1 = __raw_readw(INTSTS1); switch (irq) { case IRQ0_IRQ: @@ -108,8 +108,8 @@ static void eoi_se7206_irq(unsigned int irq) sts1 &= ~0x00ff; break; } - ctrl_outw(sts0, INTSTS0); - ctrl_outw(sts1, INTSTS1); + __raw_writew(sts0, INTSTS0); + __raw_writew(sts1, INTSTS1); } static struct irq_chip se7206_irq_chip __read_mostly = { @@ -136,11 +136,11 @@ void __init init_se7206_IRQ(void) make_se7206_irq(IRQ0_IRQ); /* SMC91C111 */ make_se7206_irq(IRQ1_IRQ); /* ATA */ make_se7206_irq(IRQ3_IRQ); /* SLOT / PCM */ - ctrl_outw(inw(INTC_ICR1) | 0x000b ,INTC_ICR1 ) ; /* ICR1 */ + __raw_writew(inw(INTC_ICR1) | 0x000b ,INTC_ICR1 ) ; /* ICR1 */ /* FPGA System register setup*/ - ctrl_outw(0x0000,INTSTS0); /* Clear INTSTS0 */ - ctrl_outw(0x0000,INTSTS1); /* Clear INTSTS1 */ + __raw_writew(0x0000,INTSTS0); /* Clear INTSTS0 */ + __raw_writew(0x0000,INTSTS1); /* Clear INTSTS1 */ /* IRQ0=LAN, IRQ1=ATA, IRQ3=SLT,PCM */ - ctrl_outw(0x0001,INTSEL); + __raw_writew(0x0001,INTSEL); } diff --git a/arch/sh/boards/mach-se/7343/irq.c b/arch/sh/boards/mach-se/7343/irq.c index c60fd13..d4305c2 100644 --- a/arch/sh/boards/mach-se/7343/irq.c +++ b/arch/sh/boards/mach-se/7343/irq.c @@ -21,13 +21,13 @@ unsigned int se7343_fpga_irq[SE7343_FPGA_IRQ_NR] = { 0, }; static void disable_se7343_irq(unsigned int irq) { unsigned int bit = (unsigned int)get_irq_chip_data(irq); - ctrl_outw(ctrl_inw(PA_CPLD_IMSK) | 1 << bit, PA_CPLD_IMSK); + __raw_writew(__raw_readw(PA_CPLD_IMSK) | 1 << bit, PA_CPLD_IMSK); } static void enable_se7343_irq(unsigned int irq) { unsigned int bit = (unsigned int)get_irq_chip_data(irq); - ctrl_outw(ctrl_inw(PA_CPLD_IMSK) & ~(1 << bit), PA_CPLD_IMSK); + __raw_writew(__raw_readw(PA_CPLD_IMSK) & ~(1 << bit), PA_CPLD_IMSK); } static struct irq_chip se7343_irq_chip __read_mostly = { @@ -39,7 +39,7 @@ static struct irq_chip se7343_irq_chip __read_mostly = { static void se7343_irq_demux(unsigned int irq, struct irq_desc *desc) { - unsigned short intv = ctrl_inw(PA_CPLD_ST); + unsigned short intv = __raw_readw(PA_CPLD_ST); unsigned int ext_irq = 0; intv &= (1 << SE7343_FPGA_IRQ_NR) - 1; @@ -59,8 +59,8 @@ void __init init_7343se_IRQ(void) { int i, irq; - ctrl_outw(0, PA_CPLD_IMSK); /* disable all irqs */ - ctrl_outw(0x2000, 0xb03fffec); /* mrshpc irq enable */ + __raw_writew(0, PA_CPLD_IMSK); /* disable all irqs */ + __raw_writew(0x2000, 0xb03fffec); /* mrshpc irq enable */ for (i = 0; i < SE7343_FPGA_IRQ_NR; i++) { irq = create_irq(); diff --git a/arch/sh/boards/mach-se/7343/setup.c b/arch/sh/boards/mach-se/7343/setup.c index 3412bb2..d2370af 100644 --- a/arch/sh/boards/mach-se/7343/setup.c +++ b/arch/sh/boards/mach-se/7343/setup.c @@ -161,10 +161,10 @@ device_initcall(sh7343se_devices_setup); */ static void __init sh7343se_setup(char **cmdline_p) { - ctrl_outw(0xf900, FPGA_OUT); /* FPGA */ + __raw_writew(0xf900, FPGA_OUT); /* FPGA */ - ctrl_outw(0x0002, PORT_PECR); /* PORT E 1 = IRQ5 */ - ctrl_outw(0x0020, PORT_PSELD); + __raw_writew(0x0002, PORT_PECR); /* PORT E 1 = IRQ5 */ + __raw_writew(0x0020, PORT_PSELD); printk(KERN_INFO "MS7343CP01 Setup...done\n"); } diff --git a/arch/sh/boards/mach-se/770x/irq.c b/arch/sh/boards/mach-se/770x/irq.c index ec1fea5..1028c17 100644 --- a/arch/sh/boards/mach-se/770x/irq.c +++ b/arch/sh/boards/mach-se/770x/irq.c @@ -96,13 +96,13 @@ static struct ipr_desc ipr_irq_desc = { void __init init_se_IRQ(void) { /* Disable all interrupts */ - ctrl_outw(0, BCR_ILCRA); - ctrl_outw(0, BCR_ILCRB); - ctrl_outw(0, BCR_ILCRC); - ctrl_outw(0, BCR_ILCRD); - ctrl_outw(0, BCR_ILCRE); - ctrl_outw(0, BCR_ILCRF); - ctrl_outw(0, BCR_ILCRG); + __raw_writew(0, BCR_ILCRA); + __raw_writew(0, BCR_ILCRB); + __raw_writew(0, BCR_ILCRC); + __raw_writew(0, BCR_ILCRD); + __raw_writew(0, BCR_ILCRE); + __raw_writew(0, BCR_ILCRF); + __raw_writew(0, BCR_ILCRG); register_ipr_controller(&ipr_irq_desc); } diff --git a/arch/sh/boards/mach-se/7721/irq.c b/arch/sh/boards/mach-se/7721/irq.c index b417acc..d85022e 100644 --- a/arch/sh/boards/mach-se/7721/irq.c +++ b/arch/sh/boards/mach-se/7721/irq.c @@ -38,7 +38,7 @@ static DECLARE_INTC_DESC(intc_desc, "SE7721", vectors, void __init init_se7721_IRQ(void) { /* PPCR */ - ctrl_outw(ctrl_inw(0xa4050118) & ~0x00ff, 0xa4050118); + __raw_writew(__raw_readw(0xa4050118) & ~0x00ff, 0xa4050118); register_intc_controller(&intc_desc); intc_set_priority(MRSHPC_IRQ0, 0xf - MRSHPC_IRQ0); diff --git a/arch/sh/boards/mach-se/7721/setup.c b/arch/sh/boards/mach-se/7721/setup.c index 460da53..7416ad7 100644 --- a/arch/sh/boards/mach-se/7721/setup.c +++ b/arch/sh/boards/mach-se/7721/setup.c @@ -80,10 +80,10 @@ device_initcall(se7721_devices_setup); static void __init se7721_setup(char **cmdline_p) { /* for USB */ - ctrl_outw(0x0000, 0xA405010C); /* PGCR */ - ctrl_outw(0x0000, 0xA405010E); /* PHCR */ - ctrl_outw(0x00AA, 0xA4050118); /* PPCR */ - ctrl_outw(0x0000, 0xA4050124); /* PSELA */ + __raw_writew(0x0000, 0xA405010C); /* PGCR */ + __raw_writew(0x0000, 0xA405010E); /* PHCR */ + __raw_writew(0x00AA, 0xA4050118); /* PPCR */ + __raw_writew(0x0000, 0xA4050124); /* PSELA */ } /* diff --git a/arch/sh/boards/mach-se/7722/irq.c b/arch/sh/boards/mach-se/7722/irq.c index b221b68..61605db 100644 --- a/arch/sh/boards/mach-se/7722/irq.c +++ b/arch/sh/boards/mach-se/7722/irq.c @@ -21,13 +21,13 @@ unsigned int se7722_fpga_irq[SE7722_FPGA_IRQ_NR] = { 0, }; static void disable_se7722_irq(unsigned int irq) { unsigned int bit = (unsigned int)get_irq_chip_data(irq); - ctrl_outw(ctrl_inw(IRQ01_MASK) | 1 << bit, IRQ01_MASK); + __raw_writew(__raw_readw(IRQ01_MASK) | 1 << bit, IRQ01_MASK); } static void enable_se7722_irq(unsigned int irq) { unsigned int bit = (unsigned int)get_irq_chip_data(irq); - ctrl_outw(ctrl_inw(IRQ01_MASK) & ~(1 << bit), IRQ01_MASK); + __raw_writew(__raw_readw(IRQ01_MASK) & ~(1 << bit), IRQ01_MASK); } static struct irq_chip se7722_irq_chip __read_mostly = { @@ -39,7 +39,7 @@ static struct irq_chip se7722_irq_chip __read_mostly = { static void se7722_irq_demux(unsigned int irq, struct irq_desc *desc) { - unsigned short intv = ctrl_inw(IRQ01_STS); + unsigned short intv = __raw_readw(IRQ01_STS); unsigned int ext_irq = 0; intv &= (1 << SE7722_FPGA_IRQ_NR) - 1; @@ -59,8 +59,8 @@ void __init init_se7722_IRQ(void) { int i, irq; - ctrl_outw(0, IRQ01_MASK); /* disable all irqs */ - ctrl_outw(0x2000, 0xb03fffec); /* mrshpc irq enable */ + __raw_writew(0, IRQ01_MASK); /* disable all irqs */ + __raw_writew(0x2000, 0xb03fffec); /* mrshpc irq enable */ for (i = 0; i < SE7722_FPGA_IRQ_NR; i++) { irq = create_irq(); diff --git a/arch/sh/boards/mach-se/7722/setup.c b/arch/sh/boards/mach-se/7722/setup.c index 9367541..80a4e57 100644 --- a/arch/sh/boards/mach-se/7722/setup.c +++ b/arch/sh/boards/mach-se/7722/setup.c @@ -156,32 +156,32 @@ device_initcall(se7722_devices_setup); static void __init se7722_setup(char **cmdline_p) { - ctrl_outw(0x010D, FPGA_OUT); /* FPGA */ + __raw_writew(0x010D, FPGA_OUT); /* FPGA */ - ctrl_outw(0x0000, PORT_PECR); /* PORT E 1 = IRQ5 ,E 0 = BS */ - ctrl_outw(0x1000, PORT_PJCR); /* PORT J 1 = IRQ1,J 0 =IRQ0 */ + __raw_writew(0x0000, PORT_PECR); /* PORT E 1 = IRQ5 ,E 0 = BS */ + __raw_writew(0x1000, PORT_PJCR); /* PORT J 1 = IRQ1,J 0 =IRQ0 */ /* LCDC I/O */ - ctrl_outw(0x0020, PORT_PSELD); + __raw_writew(0x0020, PORT_PSELD); /* SIOF1*/ - ctrl_outw(0x0003, PORT_PSELB); - ctrl_outw(0xe000, PORT_PSELC); - ctrl_outw(0x0000, PORT_PKCR); + __raw_writew(0x0003, PORT_PSELB); + __raw_writew(0xe000, PORT_PSELC); + __raw_writew(0x0000, PORT_PKCR); /* LCDC */ - ctrl_outw(0x4020, PORT_PHCR); - ctrl_outw(0x0000, PORT_PLCR); - ctrl_outw(0x0000, PORT_PMCR); - ctrl_outw(0x0002, PORT_PRCR); - ctrl_outw(0x0000, PORT_PXCR); /* LCDC,CS6A */ + __raw_writew(0x4020, PORT_PHCR); + __raw_writew(0x0000, PORT_PLCR); + __raw_writew(0x0000, PORT_PMCR); + __raw_writew(0x0002, PORT_PRCR); + __raw_writew(0x0000, PORT_PXCR); /* LCDC,CS6A */ /* KEYSC */ - ctrl_outw(0x0A10, PORT_PSELA); /* BS,SHHID2 */ - ctrl_outw(0x0000, PORT_PYCR); - ctrl_outw(0x0000, PORT_PZCR); - ctrl_outw(ctrl_inw(PORT_HIZCRA) & ~0x4000, PORT_HIZCRA); - ctrl_outw(ctrl_inw(PORT_HIZCRC) & ~0xc000, PORT_HIZCRC); + __raw_writew(0x0A10, PORT_PSELA); /* BS,SHHID2 */ + __raw_writew(0x0000, PORT_PYCR); + __raw_writew(0x0000, PORT_PZCR); + __raw_writew(__raw_readw(PORT_HIZCRA) & ~0x4000, PORT_HIZCRA); + __raw_writew(__raw_readw(PORT_HIZCRC) & ~0xc000, PORT_HIZCRC); } /* diff --git a/arch/sh/boards/mach-se/7724/irq.c b/arch/sh/boards/mach-se/7724/irq.c index f76cf3b..e5e021a 100644 --- a/arch/sh/boards/mach-se/7724/irq.c +++ b/arch/sh/boards/mach-se/7724/irq.c @@ -72,14 +72,14 @@ static void disable_se7724_irq(unsigned int irq) { struct fpga_irq set = get_fpga_irq(fpga2irq(irq)); unsigned int bit = irq - set.base; - ctrl_outw(ctrl_inw(set.mraddr) | 0x0001 << bit, set.mraddr); + __raw_writew(__raw_readw(set.mraddr) | 0x0001 << bit, set.mraddr); } static void enable_se7724_irq(unsigned int irq) { struct fpga_irq set = get_fpga_irq(fpga2irq(irq)); unsigned int bit = irq - set.base; - ctrl_outw(ctrl_inw(set.mraddr) & ~(0x0001 << bit), set.mraddr); + __raw_writew(__raw_readw(set.mraddr) & ~(0x0001 << bit), set.mraddr); } static struct irq_chip se7724_irq_chip __read_mostly = { @@ -92,7 +92,7 @@ static struct irq_chip se7724_irq_chip __read_mostly = { static void se7724_irq_demux(unsigned int irq, struct irq_desc *desc) { struct fpga_irq set = get_fpga_irq(irq); - unsigned short intv = ctrl_inw(set.sraddr); + unsigned short intv = __raw_readw(set.sraddr); struct irq_desc *ext_desc; unsigned int ext_irq = set.base; @@ -115,13 +115,13 @@ void __init init_se7724_IRQ(void) { int i; - ctrl_outw(0xffff, IRQ0_MR); /* mask all */ - ctrl_outw(0xffff, IRQ1_MR); /* mask all */ - ctrl_outw(0xffff, IRQ2_MR); /* mask all */ - ctrl_outw(0x0000, IRQ0_SR); /* clear irq */ - ctrl_outw(0x0000, IRQ1_SR); /* clear irq */ - ctrl_outw(0x0000, IRQ2_SR); /* clear irq */ - ctrl_outw(0x002a, IRQ_MODE); /* set irq type */ + __raw_writew(0xffff, IRQ0_MR); /* mask all */ + __raw_writew(0xffff, IRQ1_MR); /* mask all */ + __raw_writew(0xffff, IRQ2_MR); /* mask all */ + __raw_writew(0x0000, IRQ0_SR); /* clear irq */ + __raw_writew(0x0000, IRQ1_SR); /* clear irq */ + __raw_writew(0x0000, IRQ2_SR); /* clear irq */ + __raw_writew(0x002a, IRQ_MODE); /* set irq type */ for (i = 0; i < SE7724_FPGA_IRQ_NR; i++) set_irq_chip_and_handler_name(SE7724_FPGA_IRQ_BASE + i, diff --git a/arch/sh/boards/mach-se/7724/setup.c b/arch/sh/boards/mach-se/7724/setup.c index cbfba78..242dc84 100644 --- a/arch/sh/boards/mach-se/7724/setup.c +++ b/arch/sh/boards/mach-se/7724/setup.c @@ -256,12 +256,12 @@ static struct platform_device ceu1_device = { #define FCLKACR 0xa4150008 static void fsimck_init(struct clk *clk) { - u32 status = ctrl_inl(clk->enable_reg); + u32 status = __raw_readl(clk->enable_reg); /* use external clock */ status &= ~0x000000ff; status |= 0x00000080; - ctrl_outl(status, clk->enable_reg); + __raw_writel(status, clk->enable_reg); } static struct clk_ops fsimck_clk_ops = { @@ -522,7 +522,7 @@ static int __init sh_eth_is_eeprom_ready(void) int t = 10000; while (t--) { - if (!ctrl_inw(EEPROM_STAT)) + if (!__raw_readw(EEPROM_STAT)) return 1; cpu_relax(); } @@ -542,13 +542,13 @@ static void __init sh_eth_init(void) /* read MAC addr from EEPROM */ for (i = 0 ; i < 3 ; i++) { - ctrl_outw(0x0, EEPROM_OP); /* read */ - ctrl_outw(i*2, EEPROM_ADR); - ctrl_outw(0x1, EEPROM_STRT); + __raw_writew(0x0, EEPROM_OP); /* read */ + __raw_writew(i*2, EEPROM_ADR); + __raw_writew(0x1, EEPROM_STRT); if (!sh_eth_is_eeprom_ready()) return; - mac = ctrl_inw(EEPROM_DATA); + mac = __raw_readw(EEPROM_DATA); sh_eth_plat.mac_addr[i << 1] = mac & 0xff; sh_eth_plat.mac_addr[(i << 1) + 1] = mac >> 8; } @@ -585,7 +585,7 @@ arch_initcall(arch_setup); static int __init devices_setup(void) { - u16 sw = ctrl_inw(SW4140); /* select camera, monitor */ + u16 sw = __raw_readw(SW4140); /* select camera, monitor */ struct clk *fsia_clk; /* register board specific self-refresh code */ @@ -595,7 +595,7 @@ static int __init devices_setup(void) &ms7724se_sdram_leave_start, &ms7724se_sdram_leave_end); /* Reset Release */ - ctrl_outw(ctrl_inw(FPGA_OUT) & + __raw_writew(__raw_readw(FPGA_OUT) & ~((1 << 1) | /* LAN */ (1 << 6) | /* VIDEO DAC */ (1 << 7) | /* AK4643 */ @@ -604,7 +604,7 @@ static int __init devices_setup(void) FPGA_OUT); /* turn on USB clocks, use external clock */ - ctrl_outw((ctrl_inw(PORT_MSELCRB) & ~0xc000) | 0x8000, PORT_MSELCRB); + __raw_writew((__raw_readw(PORT_MSELCRB) & ~0xc000) | 0x8000, PORT_MSELCRB); #ifdef CONFIG_PM /* Let LED9 show STATUS2 */ @@ -633,10 +633,10 @@ static int __init devices_setup(void) #endif /* enable USB0 port */ - ctrl_outw(0x0600, 0xa40501d4); + __raw_writew(0x0600, 0xa40501d4); /* enable USB1 port */ - ctrl_outw(0x0600, 0xa4050192); + __raw_writew(0x0600, 0xa4050192); /* enable IRQ 0,1,2 */ gpio_request(GPIO_FN_INTC_IRQ0, NULL); @@ -684,7 +684,7 @@ static int __init devices_setup(void) gpio_request(GPIO_FN_LCDVCPWC, NULL); gpio_request(GPIO_FN_LCDRD, NULL); gpio_request(GPIO_FN_LCDLCLK, NULL); - ctrl_outw((ctrl_inw(PORT_HIZA) & ~0x0001), PORT_HIZA); + __raw_writew((__raw_readw(PORT_HIZA) & ~0x0001), PORT_HIZA); /* enable CEU0 */ gpio_request(GPIO_FN_VIO0_D15, NULL); diff --git a/arch/sh/boards/mach-se/7780/irq.c b/arch/sh/boards/mach-se/7780/irq.c index 121744c..d5c9edc 100644 --- a/arch/sh/boards/mach-se/7780/irq.c +++ b/arch/sh/boards/mach-se/7780/irq.c @@ -24,30 +24,30 @@ void __init init_se7780_IRQ(void) { /* enable all interrupt at FPGA */ - ctrl_outw(0, FPGA_INTMSK1); + __raw_writew(0, FPGA_INTMSK1); /* mask SM501 interrupt */ - ctrl_outw((ctrl_inw(FPGA_INTMSK1) | 0x0002), FPGA_INTMSK1); + __raw_writew((__raw_readw(FPGA_INTMSK1) | 0x0002), FPGA_INTMSK1); /* enable all interrupt at FPGA */ - ctrl_outw(0, FPGA_INTMSK2); + __raw_writew(0, FPGA_INTMSK2); /* set FPGA INTSEL register */ /* FPGA + 0x06 */ - ctrl_outw( ((IRQPIN_SM501 << IRQPOS_SM501) | + __raw_writew( ((IRQPIN_SM501 << IRQPOS_SM501) | (IRQPIN_SMC91CX << IRQPOS_SMC91CX)), FPGA_INTSEL1); /* FPGA + 0x08 */ - ctrl_outw(((IRQPIN_EXTINT4 << IRQPOS_EXTINT4) | + __raw_writew(((IRQPIN_EXTINT4 << IRQPOS_EXTINT4) | (IRQPIN_EXTINT3 << IRQPOS_EXTINT3) | (IRQPIN_EXTINT2 << IRQPOS_EXTINT2) | (IRQPIN_EXTINT1 << IRQPOS_EXTINT1)), FPGA_INTSEL2); /* FPGA + 0x0A */ - ctrl_outw((IRQPIN_PCCPW << IRQPOS_PCCPW), FPGA_INTSEL3); + __raw_writew((IRQPIN_PCCPW << IRQPOS_PCCPW), FPGA_INTSEL3); plat_irq_setup_pins(IRQ_MODE_IRQ); /* install handlers for IRQ0-7 */ /* ICR1: detect low level(for 2ndcut) */ - ctrl_outl(0xAAAA0000, INTC_ICR1); + __raw_writel(0xAAAA0000, INTC_ICR1); /* * FPGA PCISEL register initialize @@ -63,6 +63,6 @@ void __init init_se7780_IRQ(void) * INTD || INTD | INTC | -- | INTA * ------------------------------------- */ - ctrl_outw(0x0013, FPGA_PCI_INTSEL1); - ctrl_outw(0xE402, FPGA_PCI_INTSEL2); + __raw_writew(0x0013, FPGA_PCI_INTSEL1); + __raw_writew(0xE402, FPGA_PCI_INTSEL2); } diff --git a/arch/sh/boards/mach-se/7780/setup.c b/arch/sh/boards/mach-se/7780/setup.c index f7bfb3f..6f7c207 100644 --- a/arch/sh/boards/mach-se/7780/setup.c +++ b/arch/sh/boards/mach-se/7780/setup.c @@ -75,14 +75,14 @@ device_initcall(se7780_devices_setup); static void __init se7780_setup(char **cmdline_p) { /* "SH-Linux" on LED Display */ - ctrl_outw( 'S' , PA_LED_DISP + (DISP_SEL0_ADDR << 1) ); - ctrl_outw( 'H' , PA_LED_DISP + (DISP_SEL1_ADDR << 1) ); - ctrl_outw( '-' , PA_LED_DISP + (DISP_SEL2_ADDR << 1) ); - ctrl_outw( 'L' , PA_LED_DISP + (DISP_SEL3_ADDR << 1) ); - ctrl_outw( 'i' , PA_LED_DISP + (DISP_SEL4_ADDR << 1) ); - ctrl_outw( 'n' , PA_LED_DISP + (DISP_SEL5_ADDR << 1) ); - ctrl_outw( 'u' , PA_LED_DISP + (DISP_SEL6_ADDR << 1) ); - ctrl_outw( 'x' , PA_LED_DISP + (DISP_SEL7_ADDR << 1) ); + __raw_writew( 'S' , PA_LED_DISP + (DISP_SEL0_ADDR << 1) ); + __raw_writew( 'H' , PA_LED_DISP + (DISP_SEL1_ADDR << 1) ); + __raw_writew( '-' , PA_LED_DISP + (DISP_SEL2_ADDR << 1) ); + __raw_writew( 'L' , PA_LED_DISP + (DISP_SEL3_ADDR << 1) ); + __raw_writew( 'i' , PA_LED_DISP + (DISP_SEL4_ADDR << 1) ); + __raw_writew( 'n' , PA_LED_DISP + (DISP_SEL5_ADDR << 1) ); + __raw_writew( 'u' , PA_LED_DISP + (DISP_SEL6_ADDR << 1) ); + __raw_writew( 'x' , PA_LED_DISP + (DISP_SEL7_ADDR << 1) ); printk(KERN_INFO "Hitachi UL Solutions Engine 7780SE03 support.\n"); @@ -93,15 +93,15 @@ static void __init se7780_setup(char **cmdline_p) * REQ2/GNT2 -> Serial ATA * REQ3/GNT3 -> PCI slot */ - ctrl_outw(0x0213, FPGA_REQSEL); + __raw_writew(0x0213, FPGA_REQSEL); /* GPIO setting */ - ctrl_outw(0x0000, GPIO_PECR); - ctrl_outw(ctrl_inw(GPIO_PHCR)&0xfff3, GPIO_PHCR); - ctrl_outw(0x0c00, GPIO_PMSELR); + __raw_writew(0x0000, GPIO_PECR); + __raw_writew(__raw_readw(GPIO_PHCR)&0xfff3, GPIO_PHCR); + __raw_writew(0x0c00, GPIO_PMSELR); /* iVDR Power ON */ - ctrl_outw(0x0001, FPGA_IVDRPW); + __raw_writew(0x0001, FPGA_IVDRPW); } /* diff --git a/arch/sh/boards/mach-sh03/rtc.c b/arch/sh/boards/mach-sh03/rtc.c index a8b9f84..1b20099 100644 --- a/arch/sh/boards/mach-sh03/rtc.c +++ b/arch/sh/boards/mach-sh03/rtc.c @@ -44,15 +44,15 @@ unsigned long get_cmos_time(void) spin_lock(&sh03_rtc_lock); again: do { - sec = (ctrl_inb(RTC_SEC1) & 0xf) + (ctrl_inb(RTC_SEC10) & 0x7) * 10; - min = (ctrl_inb(RTC_MIN1) & 0xf) + (ctrl_inb(RTC_MIN10) & 0xf) * 10; - hour = (ctrl_inb(RTC_HOU1) & 0xf) + (ctrl_inb(RTC_HOU10) & 0xf) * 10; - day = (ctrl_inb(RTC_DAY1) & 0xf) + (ctrl_inb(RTC_DAY10) & 0xf) * 10; - mon = (ctrl_inb(RTC_MON1) & 0xf) + (ctrl_inb(RTC_MON10) & 0xf) * 10; - year = (ctrl_inb(RTC_YEA1) & 0xf) + (ctrl_inb(RTC_YEA10) & 0xf) * 10 - + (ctrl_inb(RTC_YEA100 ) & 0xf) * 100 - + (ctrl_inb(RTC_YEA1000) & 0xf) * 1000; - } while (sec != (ctrl_inb(RTC_SEC1) & 0xf) + (ctrl_inb(RTC_SEC10) & 0x7) * 10); + sec = (__raw_readb(RTC_SEC1) & 0xf) + (__raw_readb(RTC_SEC10) & 0x7) * 10; + min = (__raw_readb(RTC_MIN1) & 0xf) + (__raw_readb(RTC_MIN10) & 0xf) * 10; + hour = (__raw_readb(RTC_HOU1) & 0xf) + (__raw_readb(RTC_HOU10) & 0xf) * 10; + day = (__raw_readb(RTC_DAY1) & 0xf) + (__raw_readb(RTC_DAY10) & 0xf) * 10; + mon = (__raw_readb(RTC_MON1) & 0xf) + (__raw_readb(RTC_MON10) & 0xf) * 10; + year = (__raw_readb(RTC_YEA1) & 0xf) + (__raw_readb(RTC_YEA10) & 0xf) * 10 + + (__raw_readb(RTC_YEA100 ) & 0xf) * 100 + + (__raw_readb(RTC_YEA1000) & 0xf) * 1000; + } while (sec != (__raw_readb(RTC_SEC1) & 0xf) + (__raw_readb(RTC_SEC10) & 0x7) * 10); if (year == 0 || mon < 1 || mon > 12 || day > 31 || day < 1 || hour > 23 || min > 59 || sec > 59) { printk(KERN_ERR @@ -60,16 +60,16 @@ unsigned long get_cmos_time(void) printk("year=%d, mon=%d, day=%d, hour=%d, min=%d, sec=%d\n", year, mon, day, hour, min, sec); - ctrl_outb(0, RTC_SEC1); ctrl_outb(0, RTC_SEC10); - ctrl_outb(0, RTC_MIN1); ctrl_outb(0, RTC_MIN10); - ctrl_outb(0, RTC_HOU1); ctrl_outb(0, RTC_HOU10); - ctrl_outb(6, RTC_WEE1); - ctrl_outb(1, RTC_DAY1); ctrl_outb(0, RTC_DAY10); - ctrl_outb(1, RTC_MON1); ctrl_outb(0, RTC_MON10); - ctrl_outb(0, RTC_YEA1); ctrl_outb(0, RTC_YEA10); - ctrl_outb(0, RTC_YEA100); - ctrl_outb(2, RTC_YEA1000); - ctrl_outb(0, RTC_CTL); + __raw_writeb(0, RTC_SEC1); __raw_writeb(0, RTC_SEC10); + __raw_writeb(0, RTC_MIN1); __raw_writeb(0, RTC_MIN10); + __raw_writeb(0, RTC_HOU1); __raw_writeb(0, RTC_HOU10); + __raw_writeb(6, RTC_WEE1); + __raw_writeb(1, RTC_DAY1); __raw_writeb(0, RTC_DAY10); + __raw_writeb(1, RTC_MON1); __raw_writeb(0, RTC_MON10); + __raw_writeb(0, RTC_YEA1); __raw_writeb(0, RTC_YEA10); + __raw_writeb(0, RTC_YEA100); + __raw_writeb(2, RTC_YEA1000); + __raw_writeb(0, RTC_CTL); goto again; } @@ -93,9 +93,9 @@ static int set_rtc_mmss(unsigned long nowtime) /* gets recalled with irq locally disabled */ spin_lock(&sh03_rtc_lock); for (i = 0 ; i < 1000000 ; i++) /* may take up to 1 second... */ - if (!(ctrl_inb(RTC_CTL) & RTC_BUSY)) + if (!(__raw_readb(RTC_CTL) & RTC_BUSY)) break; - cmos_minutes = (ctrl_inb(RTC_MIN1) & 0xf) + (ctrl_inb(RTC_MIN10) & 0xf) * 10; + cmos_minutes = (__raw_readb(RTC_MIN1) & 0xf) + (__raw_readb(RTC_MIN10) & 0xf) * 10; real_seconds = nowtime % 60; real_minutes = nowtime / 60; if (((abs(real_minutes - cmos_minutes) + 15)/30) & 1) @@ -103,10 +103,10 @@ static int set_rtc_mmss(unsigned long nowtime) real_minutes %= 60; if (abs(real_minutes - cmos_minutes) < 30) { - ctrl_outb(real_seconds % 10, RTC_SEC1); - ctrl_outb(real_seconds / 10, RTC_SEC10); - ctrl_outb(real_minutes % 10, RTC_MIN1); - ctrl_outb(real_minutes / 10, RTC_MIN10); + __raw_writeb(real_seconds % 10, RTC_SEC1); + __raw_writeb(real_seconds / 10, RTC_SEC10); + __raw_writeb(real_minutes % 10, RTC_MIN1); + __raw_writeb(real_minutes / 10, RTC_MIN10); } else { printk(KERN_WARNING "set_rtc_mmss: can't update from %d to %d\n", diff --git a/arch/sh/boards/mach-sh7763rdp/irq.c b/arch/sh/boards/mach-sh7763rdp/irq.c index d8ebfa7..add698c 100644 --- a/arch/sh/boards/mach-sh7763rdp/irq.c +++ b/arch/sh/boards/mach-sh7763rdp/irq.c @@ -28,18 +28,18 @@ void __init init_sh7763rdp_IRQ(void) { /* GPIO enabled */ - ctrl_outl(1 << 25, INTC_INT2MSKCR); + __raw_writel(1 << 25, INTC_INT2MSKCR); /* enable GPIO interrupts */ - ctrl_outl((ctrl_inl(INTC_INT2PRI7) & 0xFF00FFFF) | 0x000F0000, + __raw_writel((__raw_readl(INTC_INT2PRI7) & 0xFF00FFFF) | 0x000F0000, INTC_INT2PRI7); /* USBH enabled */ - ctrl_outl(1 << 17, INTC_INT2MSKCR1); + __raw_writel(1 << 17, INTC_INT2MSKCR1); /* GETHER enabled */ - ctrl_outl(1 << 16, INTC_INT2MSKCR1); + __raw_writel(1 << 16, INTC_INT2MSKCR1); /* DMAC enabled */ - ctrl_outl(1 << 8, INTC_INT2MSKCR); + __raw_writel(1 << 8, INTC_INT2MSKCR); } diff --git a/arch/sh/boards/mach-sh7763rdp/setup.c b/arch/sh/boards/mach-sh7763rdp/setup.c index 390534a..f64a691 100644 --- a/arch/sh/boards/mach-sh7763rdp/setup.c +++ b/arch/sh/boards/mach-sh7763rdp/setup.c @@ -158,50 +158,50 @@ device_initcall(sh7763rdp_devices_setup); static void __init sh7763rdp_setup(char **cmdline_p) { /* Board version check */ - if (ctrl_inw(CPLD_BOARD_ID_ERV_REG) == 0xECB1) + if (__raw_readw(CPLD_BOARD_ID_ERV_REG) == 0xECB1) printk(KERN_INFO "RTE Standard Configuration\n"); else printk(KERN_INFO "RTA Standard Configuration\n"); /* USB pin select bits (clear bit 5-2 to 0) */ - ctrl_outw((ctrl_inw(PORT_PSEL2) & 0xFFC3), PORT_PSEL2); + __raw_writew((__raw_readw(PORT_PSEL2) & 0xFFC3), PORT_PSEL2); /* USBH setup port I controls to other (clear bits 4-9 to 0) */ - ctrl_outw(ctrl_inw(PORT_PICR) & 0xFC0F, PORT_PICR); + __raw_writew(__raw_readw(PORT_PICR) & 0xFC0F, PORT_PICR); /* Select USB Host controller */ - ctrl_outw(0x00, USB_USBHSC); + __raw_writew(0x00, USB_USBHSC); /* For LCD */ /* set PTJ7-1, bits 15-2 of PJCR to 0 */ - ctrl_outw(ctrl_inw(PORT_PJCR) & 0x0003, PORT_PJCR); + __raw_writew(__raw_readw(PORT_PJCR) & 0x0003, PORT_PJCR); /* set PTI5, bits 11-10 of PICR to 0 */ - ctrl_outw(ctrl_inw(PORT_PICR) & 0xF3FF, PORT_PICR); - ctrl_outw(0, PORT_PKCR); - ctrl_outw(0, PORT_PLCR); + __raw_writew(__raw_readw(PORT_PICR) & 0xF3FF, PORT_PICR); + __raw_writew(0, PORT_PKCR); + __raw_writew(0, PORT_PLCR); /* set PSEL2 bits 14-8, 5-4, of PSEL2 to 0 */ - ctrl_outw((ctrl_inw(PORT_PSEL2) & 0x00C0), PORT_PSEL2); + __raw_writew((__raw_readw(PORT_PSEL2) & 0x00C0), PORT_PSEL2); /* set PSEL3 bits 14-12, 6-4, 2-0 of PSEL3 to 0 */ - ctrl_outw((ctrl_inw(PORT_PSEL3) & 0x0700), PORT_PSEL3); + __raw_writew((__raw_readw(PORT_PSEL3) & 0x0700), PORT_PSEL3); /* For HAC */ /* bit3-0 0100:HAC & SSI1 enable */ - ctrl_outw((ctrl_inw(PORT_PSEL1) & 0xFFF0) | 0x0004, PORT_PSEL1); + __raw_writew((__raw_readw(PORT_PSEL1) & 0xFFF0) | 0x0004, PORT_PSEL1); /* bit14 1:SSI_HAC_CLK enable */ - ctrl_outw(ctrl_inw(PORT_PSEL4) | 0x4000, PORT_PSEL4); + __raw_writew(__raw_readw(PORT_PSEL4) | 0x4000, PORT_PSEL4); /* SH-Ether */ - ctrl_outw((ctrl_inw(PORT_PSEL1) & ~0xff00) | 0x2400, PORT_PSEL1); - ctrl_outw(0x0, PORT_PFCR); - ctrl_outw(0x0, PORT_PFCR); - ctrl_outw(0x0, PORT_PFCR); + __raw_writew((__raw_readw(PORT_PSEL1) & ~0xff00) | 0x2400, PORT_PSEL1); + __raw_writew(0x0, PORT_PFCR); + __raw_writew(0x0, PORT_PFCR); + __raw_writew(0x0, PORT_PFCR); /* MMC */ /*selects SCIF and MMC other functions */ - ctrl_outw(0x0001, PORT_PSEL0); + __raw_writew(0x0001, PORT_PSEL0); /* MMC clock operates */ - ctrl_outl(ctrl_inl(MSTPCR1) & ~0x8, MSTPCR1); - ctrl_outw(ctrl_inw(PORT_PACR) & ~0x3000, PORT_PACR); - ctrl_outw(ctrl_inw(PORT_PCCR) & ~0xCFC3, PORT_PCCR); + __raw_writel(__raw_readl(MSTPCR1) & ~0x8, MSTPCR1); + __raw_writew(__raw_readw(PORT_PACR) & ~0x3000, PORT_PACR); + __raw_writew(__raw_readw(PORT_PCCR) & ~0xCFC3, PORT_PCCR); } static struct sh_machine_vector mv_sh7763rdp __initmv = { diff --git a/arch/sh/boards/mach-snapgear/setup.c b/arch/sh/boards/mach-snapgear/setup.c index a3277a2..331745d 100644 --- a/arch/sh/boards/mach-snapgear/setup.c +++ b/arch/sh/boards/mach-snapgear/setup.c @@ -30,7 +30,7 @@ static irqreturn_t eraseconfig_interrupt(int irq, void *dev_id) { - (void)ctrl_inb(0xb8000000); /* dummy read */ + (void)__raw_readb(0xb8000000); /* dummy read */ printk("SnapGear: erase switch interrupt!\n"); diff --git a/arch/sh/boards/mach-systemh/irq.c b/arch/sh/boards/mach-systemh/irq.c index 986a0e7..523aea5 100644 --- a/arch/sh/boards/mach-systemh/irq.c +++ b/arch/sh/boards/mach-systemh/irq.c @@ -41,13 +41,13 @@ static void disable_systemh_irq(unsigned int irq) unsigned long val, mask = 0x01 << 1; /* Clear the "irq"th bit in the mask and set it in the request */ - val = ctrl_inl((unsigned long)systemh_irq_mask_register); + val = __raw_readl((unsigned long)systemh_irq_mask_register); val &= ~mask; - ctrl_outl(val, (unsigned long)systemh_irq_mask_register); + __raw_writel(val, (unsigned long)systemh_irq_mask_register); - val = ctrl_inl((unsigned long)systemh_irq_request_register); + val = __raw_readl((unsigned long)systemh_irq_request_register); val |= mask; - ctrl_outl(val, (unsigned long)systemh_irq_request_register); + __raw_writel(val, (unsigned long)systemh_irq_request_register); } } @@ -57,9 +57,9 @@ static void enable_systemh_irq(unsigned int irq) unsigned long val, mask = 0x01 << 1; /* Set "irq"th bit in the mask register */ - val = ctrl_inl((unsigned long)systemh_irq_mask_register); + val = __raw_readl((unsigned long)systemh_irq_mask_register); val |= mask; - ctrl_outl(val, (unsigned long)systemh_irq_mask_register); + __raw_writel(val, (unsigned long)systemh_irq_mask_register); } } diff --git a/arch/sh/boards/mach-titan/io.c b/arch/sh/boards/mach-titan/io.c index 0130e98..29754c5 100644 --- a/arch/sh/boards/mach-titan/io.c +++ b/arch/sh/boards/mach-titan/io.c @@ -16,8 +16,8 @@ static inline unsigned int port2adr(unsigned int port) u8 titan_inb(unsigned long port) { if (PXSEG(port)) - return ctrl_inb(port); - return ctrl_inw(port2adr(port)) & 0xff; + return __raw_readb(port); + return __raw_readw(port2adr(port)) & 0xff; } u8 titan_inb_p(unsigned long port) @@ -25,9 +25,9 @@ u8 titan_inb_p(unsigned long port) u8 v; if (PXSEG(port)) - v = ctrl_inb(port); + v = __raw_readb(port); else - v = ctrl_inw(port2adr(port)) & 0xff; + v = __raw_readw(port2adr(port)) & 0xff; ctrl_delay(); return v; } @@ -35,9 +35,9 @@ u8 titan_inb_p(unsigned long port) u16 titan_inw(unsigned long port) { if (PXSEG(port)) - return ctrl_inw(port); + return __raw_readw(port); else if (port >= 0x2000) - return ctrl_inw(port2adr(port)); + return __raw_readw(port2adr(port)); else maybebadio(port); return 0; @@ -46,9 +46,9 @@ u16 titan_inw(unsigned long port) u32 titan_inl(unsigned long port) { if (PXSEG(port)) - return ctrl_inl(port); + return __raw_readl(port); else if (port >= 0x2000) - return ctrl_inw(port2adr(port)); + return __raw_readw(port2adr(port)); else maybebadio(port); return 0; @@ -57,26 +57,26 @@ u32 titan_inl(unsigned long port) void titan_outb(u8 value, unsigned long port) { if (PXSEG(port)) - ctrl_outb(value, port); + __raw_writeb(value, port); else - ctrl_outw(value, port2adr(port)); + __raw_writew(value, port2adr(port)); } void titan_outb_p(u8 value, unsigned long port) { if (PXSEG(port)) - ctrl_outb(value, port); + __raw_writeb(value, port); else - ctrl_outw(value, port2adr(port)); + __raw_writew(value, port2adr(port)); ctrl_delay(); } void titan_outw(u16 value, unsigned long port) { if (PXSEG(port)) - ctrl_outw(value, port); + __raw_writew(value, port); else if (port >= 0x2000) - ctrl_outw(value, port2adr(port)); + __raw_writew(value, port2adr(port)); else maybebadio(port); } @@ -84,7 +84,7 @@ void titan_outw(u16 value, unsigned long port) void titan_outl(u32 value, unsigned long port) { if (PXSEG(port)) - ctrl_outl(value, port); + __raw_writel(value, port); else maybebadio(port); } diff --git a/arch/sh/boards/mach-x3proto/ilsel.c b/arch/sh/boards/mach-x3proto/ilsel.c index b5c673c..5c98427 100644 --- a/arch/sh/boards/mach-x3proto/ilsel.c +++ b/arch/sh/boards/mach-x3proto/ilsel.c @@ -70,10 +70,10 @@ static void __ilsel_enable(ilsel_source_t set, unsigned int bit) pr_debug("%s: bit#%d: addr - 0x%08lx (shift %d, set %d)\n", __func__, bit, addr, shift, set); - tmp = ctrl_inw(addr); + tmp = __raw_readw(addr); tmp &= ~(0xf << shift); tmp |= set << shift; - ctrl_outw(tmp, addr); + __raw_writew(tmp, addr); } /** @@ -142,9 +142,9 @@ void ilsel_disable(unsigned int irq) addr = mk_ilsel_addr(irq); - tmp = ctrl_inw(addr); + tmp = __raw_readw(addr); tmp &= ~(0xf << mk_ilsel_shift(irq)); - ctrl_outw(tmp, addr); + __raw_writew(tmp, addr); clear_bit(irq, &ilsel_level_map); } diff --git a/arch/sh/boards/mach-x3proto/setup.c b/arch/sh/boards/mach-x3proto/setup.c index efe4cb9..e284592 100644 --- a/arch/sh/boards/mach-x3proto/setup.c +++ b/arch/sh/boards/mach-x3proto/setup.c @@ -149,7 +149,7 @@ static void __init x3proto_init_irq(void) plat_irq_setup_pins(IRQ_MODE_IRL3210); /* Set ICR0.LVLMODE */ - ctrl_outl(ctrl_inl(0xfe410000) | (1 << 21), 0xfe410000); + __raw_writel(__raw_readl(0xfe410000) | (1 << 21), 0xfe410000); } static struct sh_machine_vector mv_x3proto __initmv = { diff --git a/arch/sh/cchips/hd6446x/hd64461.c b/arch/sh/cchips/hd6446x/hd64461.c index 50aa0c1..cbfedfb 100644 --- a/arch/sh/cchips/hd6446x/hd64461.c +++ b/arch/sh/cchips/hd6446x/hd64461.c @@ -55,7 +55,7 @@ static struct irq_chip hd64461_irq_chip = { static void hd64461_irq_demux(unsigned int irq, struct irq_desc *desc) { - unsigned short intv = ctrl_inw(HD64461_NIRR); + unsigned short intv = __raw_readw(HD64461_NIRR); struct irq_desc *ext_desc; unsigned int ext_irq = HD64461_IRQBASE; diff --git a/arch/sh/drivers/dma/dma-pvr2.c b/arch/sh/drivers/dma/dma-pvr2.c index 391cbe1..3cee58e 100644 --- a/arch/sh/drivers/dma/dma-pvr2.c +++ b/arch/sh/drivers/dma/dma-pvr2.c @@ -40,10 +40,10 @@ static irqreturn_t pvr2_dma_interrupt(int irq, void *dev_id) static int pvr2_request_dma(struct dma_channel *chan) { - if (ctrl_inl(PVR2_DMA_MODE) != 0) + if (__raw_readl(PVR2_DMA_MODE) != 0) return -EBUSY; - ctrl_outl(0, PVR2_DMA_LMMODE0); + __raw_writel(0, PVR2_DMA_LMMODE0); return 0; } @@ -60,9 +60,9 @@ static int pvr2_xfer_dma(struct dma_channel *chan) xfer_complete = 0; - ctrl_outl(chan->dar, PVR2_DMA_ADDR); - ctrl_outl(chan->count, PVR2_DMA_COUNT); - ctrl_outl(chan->mode & DMA_MODE_MASK, PVR2_DMA_MODE); + __raw_writel(chan->dar, PVR2_DMA_ADDR); + __raw_writel(chan->count, PVR2_DMA_COUNT); + __raw_writel(chan->mode & DMA_MODE_MASK, PVR2_DMA_MODE); return 0; } diff --git a/arch/sh/drivers/dma/dma-sh.c b/arch/sh/drivers/dma/dma-sh.c index 37fb5b8..565cc1a 100644 --- a/arch/sh/drivers/dma/dma-sh.c +++ b/arch/sh/drivers/dma/dma-sh.c @@ -54,7 +54,7 @@ static inline unsigned int get_dmte_irq(unsigned int chan) */ static inline unsigned int calc_xmit_shift(struct dma_channel *chan) { - u32 chcr = ctrl_inl(dma_base_addr[chan->chan] + CHCR); + u32 chcr = __raw_readl(dma_base_addr[chan->chan] + CHCR); return ts_shift[(chcr & CHCR_TS_MASK)>>CHCR_TS_SHIFT]; } @@ -70,13 +70,13 @@ static irqreturn_t dma_tei(int irq, void *dev_id) struct dma_channel *chan = dev_id; u32 chcr; - chcr = ctrl_inl(dma_base_addr[chan->chan] + CHCR); + chcr = __raw_readl(dma_base_addr[chan->chan] + CHCR); if (!(chcr & CHCR_TE)) return IRQ_NONE; chcr &= ~(CHCR_IE | CHCR_DE); - ctrl_outl(chcr, (dma_base_addr[chan->chan] + CHCR)); + __raw_writel(chcr, (dma_base_addr[chan->chan] + CHCR)); wake_up(&chan->wait_queue); @@ -115,7 +115,7 @@ sh_dmac_configure_channel(struct dma_channel *chan, unsigned long chcr) chan->flags &= ~DMA_TEI_CAPABLE; } - ctrl_outl(chcr, (dma_base_addr[chan->chan] + CHCR)); + __raw_writel(chcr, (dma_base_addr[chan->chan] + CHCR)); chan->flags |= DMA_CONFIGURED; return 0; @@ -126,13 +126,13 @@ static void sh_dmac_enable_dma(struct dma_channel *chan) int irq; u32 chcr; - chcr = ctrl_inl(dma_base_addr[chan->chan] + CHCR); + chcr = __raw_readl(dma_base_addr[chan->chan] + CHCR); chcr |= CHCR_DE; if (chan->flags & DMA_TEI_CAPABLE) chcr |= CHCR_IE; - ctrl_outl(chcr, (dma_base_addr[chan->chan] + CHCR)); + __raw_writel(chcr, (dma_base_addr[chan->chan] + CHCR)); if (chan->flags & DMA_TEI_CAPABLE) { irq = get_dmte_irq(chan->chan); @@ -150,9 +150,9 @@ static void sh_dmac_disable_dma(struct dma_channel *chan) disable_irq(irq); } - chcr = ctrl_inl(dma_base_addr[chan->chan] + CHCR); + chcr = __raw_readl(dma_base_addr[chan->chan] + CHCR); chcr &= ~(CHCR_DE | CHCR_TE | CHCR_IE); - ctrl_outl(chcr, (dma_base_addr[chan->chan] + CHCR)); + __raw_writel(chcr, (dma_base_addr[chan->chan] + CHCR)); } static int sh_dmac_xfer_dma(struct dma_channel *chan) @@ -183,12 +183,12 @@ static int sh_dmac_xfer_dma(struct dma_channel *chan) */ if (chan->sar || (mach_is_dreamcast() && chan->chan == PVR2_CASCADE_CHAN)) - ctrl_outl(chan->sar, (dma_base_addr[chan->chan]+SAR)); + __raw_writel(chan->sar, (dma_base_addr[chan->chan]+SAR)); if (chan->dar || (mach_is_dreamcast() && chan->chan == PVR2_CASCADE_CHAN)) - ctrl_outl(chan->dar, (dma_base_addr[chan->chan] + DAR)); + __raw_writel(chan->dar, (dma_base_addr[chan->chan] + DAR)); - ctrl_outl(chan->count >> calc_xmit_shift(chan), + __raw_writel(chan->count >> calc_xmit_shift(chan), (dma_base_addr[chan->chan] + TCR)); sh_dmac_enable_dma(chan); @@ -198,10 +198,10 @@ static int sh_dmac_xfer_dma(struct dma_channel *chan) static int sh_dmac_get_dma_residue(struct dma_channel *chan) { - if (!(ctrl_inl(dma_base_addr[chan->chan] + CHCR) & CHCR_DE)) + if (!(__raw_readl(dma_base_addr[chan->chan] + CHCR) & CHCR_DE)) return 0; - return ctrl_inl(dma_base_addr[chan->chan] + TCR) + return __raw_readl(dma_base_addr[chan->chan] + TCR) << calc_xmit_shift(chan); } diff --git a/arch/sh/drivers/dma/dmabrg.c b/arch/sh/drivers/dma/dmabrg.c index 5e22689..72622e3 100644 --- a/arch/sh/drivers/dma/dmabrg.c +++ b/arch/sh/drivers/dma/dmabrg.c @@ -86,8 +86,8 @@ static irqreturn_t dmabrg_irq(int irq, void *data) unsigned long dcr; unsigned int i; - dcr = ctrl_inl(DMABRGCR); - ctrl_outl(dcr & ~0x00ff0003, DMABRGCR); /* ack all */ + dcr = __raw_readl(DMABRGCR); + __raw_writel(dcr & ~0x00ff0003, DMABRGCR); /* ack all */ dcr &= dcr >> 8; /* ignore masked */ /* USB stuff, get it out of the way first */ @@ -109,17 +109,17 @@ static irqreturn_t dmabrg_irq(int irq, void *data) static void dmabrg_disable_irq(unsigned int dmairq) { unsigned long dcr; - dcr = ctrl_inl(DMABRGCR); + dcr = __raw_readl(DMABRGCR); dcr &= ~(1 << ((dmairq > 1) ? dmairq + 22 : dmairq + 8)); - ctrl_outl(dcr, DMABRGCR); + __raw_writel(dcr, DMABRGCR); } static void dmabrg_enable_irq(unsigned int dmairq) { unsigned long dcr; - dcr = ctrl_inl(DMABRGCR); + dcr = __raw_readl(DMABRGCR); dcr |= (1 << ((dmairq > 1) ? dmairq + 22 : dmairq + 8)); - ctrl_outl(dcr, DMABRGCR); + __raw_writel(dcr, DMABRGCR); } int dmabrg_request_irq(unsigned int dmairq, void(*handler)(void*), @@ -165,13 +165,13 @@ static int __init dmabrg_init(void) printk(KERN_INFO "DMABRG: DMAC ch0 not reserved!\n"); #endif - ctrl_outl(0, DMABRGCR); - ctrl_outl(0, DMACHCR0); - ctrl_outl(0x94000000, DMARSRA); /* enable DMABRG in DMAC 0 */ + __raw_writel(0, DMABRGCR); + __raw_writel(0, DMACHCR0); + __raw_writel(0x94000000, DMARSRA); /* enable DMABRG in DMAC 0 */ /* enable DMABRG mode, enable the DMAC */ - or = ctrl_inl(DMAOR); - ctrl_outl(or | DMAOR_BRG | DMAOR_DMEN, DMAOR); + or = __raw_readl(DMAOR); + __raw_writel(or | DMAOR_BRG | DMAOR_DMEN, DMAOR); ret = request_irq(DMABRGI0, dmabrg_irq, IRQF_DISABLED, "DMABRG USB address error", NULL); diff --git a/arch/sh/drivers/pci/fixups-rts7751r2d.c b/arch/sh/drivers/pci/fixups-rts7751r2d.c index 7898f14..e248516 100644 --- a/arch/sh/drivers/pci/fixups-rts7751r2d.c +++ b/arch/sh/drivers/pci/fixups-rts7751r2d.c @@ -43,7 +43,7 @@ int pci_fixup_pcic(struct pci_channel *chan) { unsigned long bcr1, mcr; - bcr1 = ctrl_inl(SH7751_BCR1); + bcr1 = __raw_readl(SH7751_BCR1); bcr1 |= 0x40080000; /* Enable Bit 19 BREQEN, set PCIC to slave */ pci_write_reg(chan, bcr1, SH4_PCIBCR1); @@ -54,7 +54,7 @@ int pci_fixup_pcic(struct pci_channel *chan) pci_write_reg(chan, 0xfb900047, SH7751_PCICONF1); pci_write_reg(chan, 0xab000001, SH7751_PCICONF4); - mcr = ctrl_inl(SH7751_MCR); + mcr = __raw_readl(SH7751_MCR); mcr = (mcr & PCIMCR_MRSET_OFF) & PCIMCR_RFSH_OFF; pci_write_reg(chan, mcr, SH4_PCIMCR); diff --git a/arch/sh/drivers/pci/pci-sh4.h b/arch/sh/drivers/pci/pci-sh4.h index 3d5296c..4744a7d 100644 --- a/arch/sh/drivers/pci/pci-sh4.h +++ b/arch/sh/drivers/pci/pci-sh4.h @@ -167,13 +167,13 @@ struct sh4_pci_address_map { static inline void pci_write_reg(struct pci_channel *chan, unsigned long val, unsigned long reg) { - ctrl_outl(val, chan->reg_base + reg); + __raw_writel(val, chan->reg_base + reg); } static inline unsigned long pci_read_reg(struct pci_channel *chan, unsigned long reg) { - return ctrl_inl(chan->reg_base + reg); + return __raw_readl(chan->reg_base + reg); } #endif /* __PCI_SH4_H */ diff --git a/arch/sh/drivers/pci/pci-sh5.h b/arch/sh/drivers/pci/pci-sh5.h index f277628..3f01dec 100644 --- a/arch/sh/drivers/pci/pci-sh5.h +++ b/arch/sh/drivers/pci/pci-sh5.h @@ -86,14 +86,14 @@ extern unsigned long pcicr_virt; /* #define PCISH5_VCR_REG(x) ( SH5PCI_VCR_BASE (PCISH5_VCR_##x)) */ /* Write I/O functions */ -#define SH5PCI_WRITE(reg,val) ctrl_outl((u32)(val),PCISH5_ICR_REG(reg)) -#define SH5PCI_WRITE_SHORT(reg,val) ctrl_outw((u16)(val),PCISH5_ICR_REG(reg)) -#define SH5PCI_WRITE_BYTE(reg,val) ctrl_outb((u8)(val),PCISH5_ICR_REG(reg)) +#define SH5PCI_WRITE(reg,val) __raw_writel((u32)(val),PCISH5_ICR_REG(reg)) +#define SH5PCI_WRITE_SHORT(reg,val) __raw_writew((u16)(val),PCISH5_ICR_REG(reg)) +#define SH5PCI_WRITE_BYTE(reg,val) __raw_writeb((u8)(val),PCISH5_ICR_REG(reg)) /* Read I/O functions */ -#define SH5PCI_READ(reg) ctrl_inl(PCISH5_ICR_REG(reg)) -#define SH5PCI_READ_SHORT(reg) ctrl_inw(PCISH5_ICR_REG(reg)) -#define SH5PCI_READ_BYTE(reg) ctrl_inb(PCISH5_ICR_REG(reg)) +#define SH5PCI_READ(reg) __raw_readl(PCISH5_ICR_REG(reg)) +#define SH5PCI_READ_SHORT(reg) __raw_readw(PCISH5_ICR_REG(reg)) +#define SH5PCI_READ_BYTE(reg) __raw_readb(PCISH5_ICR_REG(reg)) /* Set PCI config bits */ #define SET_CONFIG_BITS(bus,devfn,where) ((((bus) << 16) | ((devfn) << 8) | ((where) & ~3)) | 0x80000000) diff --git a/arch/sh/drivers/pci/pci-sh7751.c b/arch/sh/drivers/pci/pci-sh7751.c index 70c1999..2455cf3 100644 --- a/arch/sh/drivers/pci/pci-sh7751.c +++ b/arch/sh/drivers/pci/pci-sh7751.c @@ -97,9 +97,9 @@ static int __init sh7751_pci_init(void) return ret; /* Set the BCR's to enable PCI access */ - reg = ctrl_inl(SH7751_BCR1); + reg = __raw_readl(SH7751_BCR1); reg |= 0x80000; - ctrl_outl(reg, SH7751_BCR1); + __raw_writel(reg, SH7751_BCR1); /* Turn the clocks back on (not done in reset)*/ pci_write_reg(chan, 0, SH4_PCICLKR); @@ -159,13 +159,13 @@ static int __init sh7751_pci_init(void) return -1; /* configure the wait control registers */ - word = ctrl_inl(SH7751_WCR1); + word = __raw_readl(SH7751_WCR1); pci_write_reg(chan, word, SH4_PCIWCR1); - word = ctrl_inl(SH7751_WCR2); + word = __raw_readl(SH7751_WCR2); pci_write_reg(chan, word, SH4_PCIWCR2); - word = ctrl_inl(SH7751_WCR3); + word = __raw_readl(SH7751_WCR3); pci_write_reg(chan, word, SH4_PCIWCR3); - word = ctrl_inl(SH7751_MCR); + word = __raw_readl(SH7751_MCR); pci_write_reg(chan, word, SH4_PCIMCR); /* NOTE: I'm ignoring the PCI error IRQs for now.. diff --git a/arch/sh/drivers/superhyway/ops-sh4-202.c b/arch/sh/drivers/superhyway/ops-sh4-202.c index 3b14bf8..6da62e9 100644 --- a/arch/sh/drivers/superhyway/ops-sh4-202.c +++ b/arch/sh/drivers/superhyway/ops-sh4-202.c @@ -134,8 +134,8 @@ static int sh4202_read_vcr(unsigned long base, struct superhyway_vcr_info *vcr) * * Do not trust the documentation, for it is evil. */ - vcrh = ctrl_inl(base); - vcrl = ctrl_inl(base + sizeof(u32)); + vcrh = __raw_readl(base); + vcrl = __raw_readl(base + sizeof(u32)); tmp = ((u64)vcrh << 32) | vcrl; memcpy(vcr, &tmp, sizeof(u64)); @@ -147,8 +147,8 @@ static int sh4202_write_vcr(unsigned long base, struct superhyway_vcr_info vcr) { u64 tmp = *(u64 *)&vcr; - ctrl_outl((tmp >> 32) & 0xffffffff, base); - ctrl_outl(tmp & 0xffffffff, base + sizeof(u32)); + __raw_writel((tmp >> 32) & 0xffffffff, base); + __raw_writel(tmp & 0xffffffff, base + sizeof(u32)); return 0; } diff --git a/arch/sh/include/asm/dma-sh.h b/arch/sh/include/asm/dma-sh.h index 78eed3e..47de109 100644 --- a/arch/sh/include/asm/dma-sh.h +++ b/arch/sh/include/asm/dma-sh.h @@ -20,14 +20,14 @@ defined(CONFIG_CPU_SUBTYPE_SH7780) || \ defined(CONFIG_CPU_SUBTYPE_SH7785) #define dmaor_read_reg(n) \ - (n ? ctrl_inw(SH_DMAC_BASE1 + DMAOR) \ - : ctrl_inw(SH_DMAC_BASE0 + DMAOR)) + (n ? __raw_readw(SH_DMAC_BASE1 + DMAOR) \ + : __raw_readw(SH_DMAC_BASE0 + DMAOR)) #define dmaor_write_reg(n, data) \ - (n ? ctrl_outw(data, SH_DMAC_BASE1 + DMAOR) \ - : ctrl_outw(data, SH_DMAC_BASE0 + DMAOR)) + (n ? __raw_writew(data, SH_DMAC_BASE1 + DMAOR) \ + : __raw_writew(data, SH_DMAC_BASE0 + DMAOR)) #else /* Other CPU */ -#define dmaor_read_reg(n) ctrl_inw(SH_DMAC_BASE0 + DMAOR) -#define dmaor_write_reg(n, data) ctrl_outw(data, SH_DMAC_BASE0 + DMAOR) +#define dmaor_read_reg(n) __raw_readw(SH_DMAC_BASE0 + DMAOR) +#define dmaor_write_reg(n, data) __raw_writew(data, SH_DMAC_BASE0 + DMAOR) #endif static int dmte_irq_map[] __maybe_unused = { diff --git a/arch/sh/include/asm/mmu_context.h b/arch/sh/include/asm/mmu_context.h index 41080b1..384c747 100644 --- a/arch/sh/include/asm/mmu_context.h +++ b/arch/sh/include/asm/mmu_context.h @@ -158,7 +158,7 @@ static inline void enable_mmu(void) unsigned int cpu = smp_processor_id(); /* Enable MMU */ - ctrl_outl(MMU_CONTROL_INIT, MMUCR); + __raw_writel(MMU_CONTROL_INIT, MMUCR); ctrl_barrier(); if (asid_cache(cpu) == NO_CONTEXT) @@ -171,9 +171,9 @@ static inline void disable_mmu(void) { unsigned long cr; - cr = ctrl_inl(MMUCR); + cr = __raw_readl(MMUCR); cr &= ~MMU_CONTROL_INIT; - ctrl_outl(cr, MMUCR); + __raw_writel(cr, MMUCR); ctrl_barrier(); } diff --git a/arch/sh/include/asm/mmu_context_32.h b/arch/sh/include/asm/mmu_context_32.h index 8ef800c..10e2e17 100644 --- a/arch/sh/include/asm/mmu_context_32.h +++ b/arch/sh/include/asm/mmu_context_32.h @@ -49,11 +49,11 @@ static inline unsigned long get_asid(void) /* MMU_TTB is used for optimizing the fault handling. */ static inline void set_TTB(pgd_t *pgd) { - ctrl_outl((unsigned long)pgd, MMU_TTB); + __raw_writel((unsigned long)pgd, MMU_TTB); } static inline pgd_t *get_TTB(void) { - return (pgd_t *)ctrl_inl(MMU_TTB); + return (pgd_t *)__raw_readl(MMU_TTB); } #endif /* __ASM_SH_MMU_CONTEXT_32_H */ diff --git a/arch/sh/include/asm/watchdog.h b/arch/sh/include/asm/watchdog.h index 19dfff5..85a7aca 100644 --- a/arch/sh/include/asm/watchdog.h +++ b/arch/sh/include/asm/watchdog.h @@ -70,7 +70,7 @@ */ static inline __u32 sh_wdt_read_cnt(void) { - return ctrl_inl(WTCNT_R); + return __raw_readl(WTCNT_R); } /** @@ -82,7 +82,7 @@ static inline __u32 sh_wdt_read_cnt(void) */ static inline void sh_wdt_write_cnt(__u32 val) { - ctrl_outl((WTCNT_HIGH << 24) | (__u32)val, WTCNT); + __raw_writel((WTCNT_HIGH << 24) | (__u32)val, WTCNT); } /** @@ -94,7 +94,7 @@ static inline void sh_wdt_write_cnt(__u32 val) */ static inline void sh_wdt_write_bst(__u32 val) { - ctrl_outl((WTBST_HIGH << 24) | (__u32)val, WTBST); + __raw_writel((WTBST_HIGH << 24) | (__u32)val, WTBST); } /** * sh_wdt_read_csr - Read from Control/Status Register @@ -103,7 +103,7 @@ static inline void sh_wdt_write_bst(__u32 val) */ static inline __u32 sh_wdt_read_csr(void) { - return ctrl_inl(WTCSR_R); + return __raw_readl(WTCSR_R); } /** @@ -115,7 +115,7 @@ static inline __u32 sh_wdt_read_csr(void) */ static inline void sh_wdt_write_csr(__u32 val) { - ctrl_outl((WTCSR_HIGH << 24) | (__u32)val, WTCSR); + __raw_writel((WTCSR_HIGH << 24) | (__u32)val, WTCSR); } #else /** @@ -124,7 +124,7 @@ static inline void sh_wdt_write_csr(__u32 val) */ static inline __u8 sh_wdt_read_cnt(void) { - return ctrl_inb(WTCNT_R); + return __raw_readb(WTCNT_R); } /** @@ -136,7 +136,7 @@ static inline __u8 sh_wdt_read_cnt(void) */ static inline void sh_wdt_write_cnt(__u8 val) { - ctrl_outw((WTCNT_HIGH << 8) | (__u16)val, WTCNT); + __raw_writew((WTCNT_HIGH << 8) | (__u16)val, WTCNT); } /** @@ -146,7 +146,7 @@ static inline void sh_wdt_write_cnt(__u8 val) */ static inline __u8 sh_wdt_read_csr(void) { - return ctrl_inb(WTCSR_R); + return __raw_readb(WTCSR_R); } /** @@ -158,7 +158,7 @@ static inline __u8 sh_wdt_read_csr(void) */ static inline void sh_wdt_write_csr(__u8 val) { - ctrl_outw((WTCSR_HIGH << 8) | (__u16)val, WTCSR); + __raw_writew((WTCSR_HIGH << 8) | (__u16)val, WTCSR); } #endif /* CONFIG_CPU_SUBTYPE_SH7785 || CONFIG_CPU_SUBTYPE_SH7780 */ #endif /* __KERNEL__ */ diff --git a/arch/sh/include/cpu-sh2/cpu/watchdog.h b/arch/sh/include/cpu-sh2/cpu/watchdog.h index 393161c..1eab8aa 100644 --- a/arch/sh/include/cpu-sh2/cpu/watchdog.h +++ b/arch/sh/include/cpu-sh2/cpu/watchdog.h @@ -44,7 +44,7 @@ static inline __u8 sh_wdt_read_rstcsr(void) /* * Same read/write brain-damage as for WTCNT here.. */ - return ctrl_inb(RSTCSR_R); + return __raw_readb(RSTCSR_R); } /** @@ -62,7 +62,7 @@ static inline void sh_wdt_write_rstcsr(__u8 val) * we can't presently touch the WOVF bit, since the upper byte * has to be swapped for this. So just leave it alone.. */ - ctrl_outw((WTCNT_HIGH << 8) | (__u16)val, RSTCSR); + __raw_writeb((WTCNT_HIGH << 8) | (__u16)val, RSTCSR); } #endif /* __ASM_CPU_SH2_WATCHDOG_H */ diff --git a/arch/sh/include/cpu-sh3/cpu/dac.h b/arch/sh/include/cpu-sh3/cpu/dac.h index 05fda83..98f1d15 100644 --- a/arch/sh/include/cpu-sh3/cpu/dac.h +++ b/arch/sh/include/cpu-sh3/cpu/dac.h @@ -17,25 +17,25 @@ static __inline__ void sh_dac_enable(int channel) { unsigned char v; - v = ctrl_inb(DACR); + v = __raw_readb(DACR); if(channel) v |= DACR_DAOE1; else v |= DACR_DAOE0; - ctrl_outb(v,DACR); + __raw_writeb(v,DACR); } static __inline__ void sh_dac_disable(int channel) { unsigned char v; - v = ctrl_inb(DACR); + v = __raw_readb(DACR); if(channel) v &= ~DACR_DAOE1; else v &= ~DACR_DAOE0; - ctrl_outb(v,DACR); + __raw_writeb(v,DACR); } static __inline__ void sh_dac_output(u8 value, int channel) { - if(channel) ctrl_outb(value,DADR1); - else ctrl_outb(value,DADR0); + if(channel) __raw_writeb(value,DADR1); + else __raw_writeb(value,DADR0); } #endif /* __ASM_CPU_SH3_DAC_H */ diff --git a/arch/sh/include/mach-common/mach/magicpanelr2.h b/arch/sh/include/mach-common/mach/magicpanelr2.h index c644a77..183a2f7 100644 --- a/arch/sh/include/mach-common/mach/magicpanelr2.h +++ b/arch/sh/include/mach-common/mach/magicpanelr2.h @@ -19,12 +19,12 @@ #include -#define SETBITS_OUTB(mask, reg) ctrl_outb(ctrl_inb(reg) | mask, reg) -#define SETBITS_OUTW(mask, reg) ctrl_outw(ctrl_inw(reg) | mask, reg) -#define SETBITS_OUTL(mask, reg) ctrl_outl(ctrl_inl(reg) | mask, reg) -#define CLRBITS_OUTB(mask, reg) ctrl_outb(ctrl_inb(reg) & ~mask, reg) -#define CLRBITS_OUTW(mask, reg) ctrl_outw(ctrl_inw(reg) & ~mask, reg) -#define CLRBITS_OUTL(mask, reg) ctrl_outl(ctrl_inl(reg) & ~mask, reg) +#define SETBITS_OUTB(mask, reg) __raw_writeb(__raw_readb(reg) | mask, reg) +#define SETBITS_OUTW(mask, reg) __raw_writew(__raw_readw(reg) | mask, reg) +#define SETBITS_OUTL(mask, reg) __raw_writel(__raw_readl(reg) | mask, reg) +#define CLRBITS_OUTB(mask, reg) __raw_writeb(__raw_readb(reg) & ~mask, reg) +#define CLRBITS_OUTW(mask, reg) __raw_writew(__raw_readw(reg) & ~mask, reg) +#define CLRBITS_OUTL(mask, reg) __raw_writel(__raw_readl(reg) & ~mask, reg) #define PA_LED PORT_PADR /* LED */ diff --git a/arch/sh/kernel/cpu/adc.c b/arch/sh/kernel/cpu/adc.c index da3d687..d307571 100644 --- a/arch/sh/kernel/cpu/adc.c +++ b/arch/sh/kernel/cpu/adc.c @@ -18,19 +18,19 @@ int adc_single(unsigned int channel) off = (channel & 0x03) << 2; - csr = ctrl_inb(ADCSR); + csr = __raw_readb(ADCSR); csr = channel | ADCSR_ADST | ADCSR_CKS; - ctrl_outb(csr, ADCSR); + __raw_writeb(csr, ADCSR); do { - csr = ctrl_inb(ADCSR); + csr = __raw_readb(ADCSR); } while ((csr & ADCSR_ADF) == 0); csr &= ~(ADCSR_ADF | ADCSR_ADST); - ctrl_outb(csr, ADCSR); + __raw_writeb(csr, ADCSR); - return (((ctrl_inb(ADDRAH + off) << 8) | - ctrl_inb(ADDRAL + off)) >> 6); + return (((__raw_readb(ADDRAH + off) << 8) | + __raw_readb(ADDRAL + off)) >> 6); } EXPORT_SYMBOL(adc_single); diff --git a/arch/sh/kernel/cpu/init.c b/arch/sh/kernel/cpu/init.c index 05a7d2a..6311b0b 100644 --- a/arch/sh/kernel/cpu/init.c +++ b/arch/sh/kernel/cpu/init.c @@ -61,10 +61,10 @@ onchip_setup(dsp); static void __init speculative_execution_init(void) { /* Clear RABD */ - ctrl_outl(ctrl_inl(CPUOPM) & ~CPUOPM_RABD, CPUOPM); + __raw_writel(__raw_readl(CPUOPM) & ~CPUOPM_RABD, CPUOPM); /* Flush the update */ - (void)ctrl_inl(CPUOPM); + (void)__raw_readl(CPUOPM); ctrl_barrier(); } #else @@ -111,7 +111,7 @@ static void cache_init(void) unsigned long ccr, flags; jump_to_uncached(); - ccr = ctrl_inl(CCR); + ccr = __raw_readl(CCR); /* * At this point we don't know whether the cache is enabled or not - a @@ -155,7 +155,7 @@ static void cache_init(void) for (addr = addrstart; addr < addrstart + waysize; addr += current_cpu_data.dcache.linesz) - ctrl_outl(0, addr); + __raw_writel(0, addr); addrstart += current_cpu_data.dcache.way_incr; } while (--ways); @@ -188,7 +188,7 @@ static void cache_init(void) l2_cache_init(); - ctrl_outl(flags, CCR); + __raw_writel(flags, CCR); back_to_cached(); } #else diff --git a/arch/sh/kernel/cpu/irq/intc-sh5.c b/arch/sh/kernel/cpu/irq/intc-sh5.c index 06e7e29..96a2395 100644 --- a/arch/sh/kernel/cpu/irq/intc-sh5.c +++ b/arch/sh/kernel/cpu/irq/intc-sh5.c @@ -123,7 +123,7 @@ static void enable_intc_irq(unsigned int irq) bitmask = 1 << (irq - 32); } - ctrl_outl(bitmask, reg); + __raw_writel(bitmask, reg); } static void disable_intc_irq(unsigned int irq) @@ -139,7 +139,7 @@ static void disable_intc_irq(unsigned int irq) bitmask = 1 << (irq - 32); } - ctrl_outl(bitmask, reg); + __raw_writel(bitmask, reg); } static void mask_and_ack_intc(unsigned int irq) @@ -170,11 +170,11 @@ void __init plat_irq_setup(void) /* Disable all interrupts and set all priorities to 0 to avoid trouble */ - ctrl_outl(-1, INTC_INTDSB_0); - ctrl_outl(-1, INTC_INTDSB_1); + __raw_writel(-1, INTC_INTDSB_0); + __raw_writel(-1, INTC_INTDSB_1); for (reg = INTC_INTPRI_0, i = 0; i < INTC_INTPRI_PREGS; i++, reg += 8) - ctrl_outl( NO_PRIORITY, reg); + __raw_writel( NO_PRIORITY, reg); #ifdef CONFIG_SH_CAYMAN @@ -199,7 +199,7 @@ void __init plat_irq_setup(void) reg = INTC_ICR_SET; i = IRQ_IRL0; } - ctrl_outl(INTC_ICR_IRLM, reg); + __raw_writel(INTC_ICR_IRLM, reg); /* Set interrupt priorities according to platform description */ for (data = 0, reg = INTC_INTPRI_0; i < NR_INTC_IRQS; i++) { @@ -207,7 +207,7 @@ void __init plat_irq_setup(void) ((i % INTC_INTPRI_PPREG) * 4); if ((i % INTC_INTPRI_PPREG) == (INTC_INTPRI_PPREG - 1)) { /* Upon the 7th, set Priority Register */ - ctrl_outl(data, reg); + __raw_writel(data, reg); data = 0; reg += 8; } diff --git a/arch/sh/kernel/cpu/sh2/clock-sh7619.c b/arch/sh/kernel/cpu/sh2/clock-sh7619.c index 4fe8631..0c9f24d 100644 --- a/arch/sh/kernel/cpu/sh2/clock-sh7619.c +++ b/arch/sh/kernel/cpu/sh2/clock-sh7619.c @@ -31,7 +31,7 @@ static const int pfc_divisors[] = {1,2,0,4}; static void master_clk_init(struct clk *clk) { - clk->rate *= PLL2 * pll1rate[(ctrl_inw(FREQCR) >> 8) & 7]; + clk->rate *= PLL2 * pll1rate[(__raw_readw(FREQCR) >> 8) & 7]; } static struct clk_ops sh7619_master_clk_ops = { @@ -40,7 +40,7 @@ static struct clk_ops sh7619_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FREQCR) & 0x0007); + int idx = (__raw_readw(FREQCR) & 0x0007); return clk->parent->rate / pfc_divisors[idx]; } @@ -50,7 +50,7 @@ static struct clk_ops sh7619_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - return clk->parent->rate / pll1rate[(ctrl_inw(FREQCR) >> 8) & 7]; + return clk->parent->rate / pll1rate[(__raw_readw(FREQCR) >> 8) & 7]; } static struct clk_ops sh7619_bus_clk_ops = { diff --git a/arch/sh/kernel/cpu/sh2a/clock-sh7201.c b/arch/sh/kernel/cpu/sh2a/clock-sh7201.c index 7814c76..b26264d 100644 --- a/arch/sh/kernel/cpu/sh2a/clock-sh7201.c +++ b/arch/sh/kernel/cpu/sh2a/clock-sh7201.c @@ -34,7 +34,7 @@ static const int pfc_divisors[]={1,2,3,4,6,8,12}; static void master_clk_init(struct clk *clk) { - return 10000000 * PLL2 * pll1rate[(ctrl_inw(FREQCR) >> 8) & 0x0007]; + return 10000000 * PLL2 * pll1rate[(__raw_readw(FREQCR) >> 8) & 0x0007]; } static struct clk_ops sh7201_master_clk_ops = { @@ -43,7 +43,7 @@ static struct clk_ops sh7201_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FREQCR) & 0x0007); + int idx = (__raw_readw(FREQCR) & 0x0007); return clk->parent->rate / pfc_divisors[idx]; } @@ -53,7 +53,7 @@ static struct clk_ops sh7201_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FREQCR) & 0x0007); + int idx = (__raw_readw(FREQCR) & 0x0007); return clk->parent->rate / pfc_divisors[idx]; } @@ -63,7 +63,7 @@ static struct clk_ops sh7201_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inw(FREQCR) >> 4) & 0x0007); + int idx = ((__raw_readw(FREQCR) >> 4) & 0x0007); return clk->parent->rate / ifc_divisors[idx]; } diff --git a/arch/sh/kernel/cpu/sh2a/clock-sh7203.c b/arch/sh/kernel/cpu/sh2a/clock-sh7203.c index 9409869..7e75d8f 100644 --- a/arch/sh/kernel/cpu/sh2a/clock-sh7203.c +++ b/arch/sh/kernel/cpu/sh2a/clock-sh7203.c @@ -39,7 +39,7 @@ static const int pfc_divisors[]={1,2,3,4,6,8,12}; static void master_clk_init(struct clk *clk) { - clk->rate *= pll1rate[(ctrl_inw(FREQCR) >> 8) & 0x0003] * PLL2 ; + clk->rate *= pll1rate[(__raw_readw(FREQCR) >> 8) & 0x0003] * PLL2 ; } static struct clk_ops sh7203_master_clk_ops = { @@ -48,7 +48,7 @@ static struct clk_ops sh7203_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FREQCR) & 0x0007); + int idx = (__raw_readw(FREQCR) & 0x0007); return clk->parent->rate / pfc_divisors[idx]; } @@ -58,7 +58,7 @@ static struct clk_ops sh7203_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FREQCR) & 0x0007); + int idx = (__raw_readw(FREQCR) & 0x0007); return clk->parent->rate / pfc_divisors[idx-2]; } diff --git a/arch/sh/kernel/cpu/sh2a/clock-sh7206.c b/arch/sh/kernel/cpu/sh2a/clock-sh7206.c index c2268bd..b27a5e2 100644 --- a/arch/sh/kernel/cpu/sh2a/clock-sh7206.c +++ b/arch/sh/kernel/cpu/sh2a/clock-sh7206.c @@ -34,7 +34,7 @@ static const int pfc_divisors[]={1,2,3,4,6,8,12}; static void master_clk_init(struct clk *clk) { - clk->rate *= PLL2 * pll1rate[(ctrl_inw(FREQCR) >> 8) & 0x0007]; + clk->rate *= PLL2 * pll1rate[(__raw_readw(FREQCR) >> 8) & 0x0007]; } static struct clk_ops sh7206_master_clk_ops = { @@ -43,7 +43,7 @@ static struct clk_ops sh7206_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FREQCR) & 0x0007); + int idx = (__raw_readw(FREQCR) & 0x0007); return clk->parent->rate / pfc_divisors[idx]; } @@ -53,7 +53,7 @@ static struct clk_ops sh7206_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - return clk->parent->rate / pll1rate[(ctrl_inw(FREQCR) >> 8) & 0x0007]; + return clk->parent->rate / pll1rate[(__raw_readw(FREQCR) >> 8) & 0x0007]; } static struct clk_ops sh7206_bus_clk_ops = { @@ -62,7 +62,7 @@ static struct clk_ops sh7206_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FREQCR) & 0x0007); + int idx = (__raw_readw(FREQCR) & 0x0007); return clk->parent->rate / ifc_divisors[idx]; } diff --git a/arch/sh/kernel/cpu/sh3/clock-sh3.c b/arch/sh/kernel/cpu/sh3/clock-sh3.c index 27b8738..b78384a 100644 --- a/arch/sh/kernel/cpu/sh3/clock-sh3.c +++ b/arch/sh/kernel/cpu/sh3/clock-sh3.c @@ -28,7 +28,7 @@ static int pfc_divisors[] = { 1, 2, 3, 4, 6, 1, 1, 1 }; static void master_clk_init(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x2000) >> 11) | (frqcr & 0x0003); clk->rate *= pfc_divisors[idx]; @@ -40,7 +40,7 @@ static struct clk_ops sh3_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x2000) >> 11) | (frqcr & 0x0003); return clk->parent->rate / pfc_divisors[idx]; @@ -52,7 +52,7 @@ static struct clk_ops sh3_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x8000) >> 13) | ((frqcr & 0x0030) >> 4); return clk->parent->rate / stc_multipliers[idx]; @@ -64,7 +64,7 @@ static struct clk_ops sh3_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x4000) >> 12) | ((frqcr & 0x000c) >> 2); return clk->parent->rate / ifc_divisors[idx]; diff --git a/arch/sh/kernel/cpu/sh3/clock-sh7705.c b/arch/sh/kernel/cpu/sh3/clock-sh7705.c index 0ca8f2c..0ecea14 100644 --- a/arch/sh/kernel/cpu/sh3/clock-sh7705.c +++ b/arch/sh/kernel/cpu/sh3/clock-sh7705.c @@ -32,7 +32,7 @@ static int pfc_divisors[] = { 1, 2, 3, 4, 6, 1, 1, 1 }; static void master_clk_init(struct clk *clk) { - clk->rate *= pfc_divisors[ctrl_inw(FRQCR) & 0x0003]; + clk->rate *= pfc_divisors[__raw_readw(FRQCR) & 0x0003]; } static struct clk_ops sh7705_master_clk_ops = { @@ -41,7 +41,7 @@ static struct clk_ops sh7705_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = ctrl_inw(FRQCR) & 0x0003; + int idx = __raw_readw(FRQCR) & 0x0003; return clk->parent->rate / pfc_divisors[idx]; } @@ -51,7 +51,7 @@ static struct clk_ops sh7705_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) & 0x0300) >> 8; + int idx = (__raw_readw(FRQCR) & 0x0300) >> 8; return clk->parent->rate / stc_multipliers[idx]; } @@ -61,7 +61,7 @@ static struct clk_ops sh7705_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) & 0x0030) >> 4; + int idx = (__raw_readw(FRQCR) & 0x0030) >> 4; return clk->parent->rate / ifc_divisors[idx]; } diff --git a/arch/sh/kernel/cpu/sh3/clock-sh7706.c b/arch/sh/kernel/cpu/sh3/clock-sh7706.c index 4bf7887..6f9ff8b 100644 --- a/arch/sh/kernel/cpu/sh3/clock-sh7706.c +++ b/arch/sh/kernel/cpu/sh3/clock-sh7706.c @@ -24,7 +24,7 @@ static int pfc_divisors[] = { 1, 2, 4, 1, 3, 6, 1, 1 }; static void master_clk_init(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x2000) >> 11) | (frqcr & 0x0003); clk->rate *= pfc_divisors[idx]; @@ -36,7 +36,7 @@ static struct clk_ops sh7706_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x2000) >> 11) | (frqcr & 0x0003); return clk->parent->rate / pfc_divisors[idx]; @@ -48,7 +48,7 @@ static struct clk_ops sh7706_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x8000) >> 13) | ((frqcr & 0x0030) >> 4); return clk->parent->rate / stc_multipliers[idx]; @@ -60,7 +60,7 @@ static struct clk_ops sh7706_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x4000) >> 12) | ((frqcr & 0x000c) >> 2); return clk->parent->rate / ifc_divisors[idx]; diff --git a/arch/sh/kernel/cpu/sh3/clock-sh7709.c b/arch/sh/kernel/cpu/sh3/clock-sh7709.c index e874950..f302ba0 100644 --- a/arch/sh/kernel/cpu/sh3/clock-sh7709.c +++ b/arch/sh/kernel/cpu/sh3/clock-sh7709.c @@ -24,7 +24,7 @@ static int pfc_divisors[] = { 1, 2, 4, 1, 3, 6, 1, 1 }; static void master_clk_init(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x2000) >> 11) | (frqcr & 0x0003); clk->rate *= pfc_divisors[idx]; @@ -36,7 +36,7 @@ static struct clk_ops sh7709_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x2000) >> 11) | (frqcr & 0x0003); return clk->parent->rate / pfc_divisors[idx]; @@ -48,7 +48,7 @@ static struct clk_ops sh7709_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = (frqcr & 0x0080) ? ((frqcr & 0x8000) >> 13) | ((frqcr & 0x0030) >> 4) : 1; @@ -61,7 +61,7 @@ static struct clk_ops sh7709_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = ((frqcr & 0x4000) >> 12) | ((frqcr & 0x000c) >> 2); return clk->parent->rate / ifc_divisors[idx]; diff --git a/arch/sh/kernel/cpu/sh3/clock-sh7710.c b/arch/sh/kernel/cpu/sh3/clock-sh7710.c index 030a58b..29a87d8 100644 --- a/arch/sh/kernel/cpu/sh3/clock-sh7710.c +++ b/arch/sh/kernel/cpu/sh3/clock-sh7710.c @@ -26,7 +26,7 @@ static int md_table[] = { 1, 2, 3, 4, 6, 8, 12 }; static void master_clk_init(struct clk *clk) { - clk->rate *= md_table[ctrl_inw(FRQCR) & 0x0007]; + clk->rate *= md_table[__raw_readw(FRQCR) & 0x0007]; } static struct clk_ops sh7710_master_clk_ops = { @@ -35,7 +35,7 @@ static struct clk_ops sh7710_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) & 0x0007); + int idx = (__raw_readw(FRQCR) & 0x0007); return clk->parent->rate / md_table[idx]; } @@ -45,7 +45,7 @@ static struct clk_ops sh7710_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) & 0x0700) >> 8; + int idx = (__raw_readw(FRQCR) & 0x0700) >> 8; return clk->parent->rate / md_table[idx]; } @@ -55,7 +55,7 @@ static struct clk_ops sh7710_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) & 0x0070) >> 4; + int idx = (__raw_readw(FRQCR) & 0x0070) >> 4; return clk->parent->rate / md_table[idx]; } diff --git a/arch/sh/kernel/cpu/sh3/clock-sh7712.c b/arch/sh/kernel/cpu/sh3/clock-sh7712.c index 6428ee6..b0d0c52 100644 --- a/arch/sh/kernel/cpu/sh3/clock-sh7712.c +++ b/arch/sh/kernel/cpu/sh3/clock-sh7712.c @@ -23,7 +23,7 @@ static int divisors[] = { 1, 2, 3, 4, 6 }; static void master_clk_init(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = (frqcr & 0x0300) >> 8; clk->rate *= multipliers[idx]; @@ -35,7 +35,7 @@ static struct clk_ops sh7712_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = frqcr & 0x0007; return clk->parent->rate / divisors[idx]; @@ -47,7 +47,7 @@ static struct clk_ops sh7712_module_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int frqcr = ctrl_inw(FRQCR); + int frqcr = __raw_readw(FRQCR); int idx = (frqcr & 0x0030) >> 4; return clk->parent->rate / divisors[idx]; diff --git a/arch/sh/kernel/cpu/sh3/probe.c b/arch/sh/kernel/cpu/sh3/probe.c index c2db657..295ec4c 100644 --- a/arch/sh/kernel/cpu/sh3/probe.c +++ b/arch/sh/kernel/cpu/sh3/probe.c @@ -30,23 +30,23 @@ int detect_cpu_and_cache_system(void) addr1 = CACHE_OC_ADDRESS_ARRAY + (1 << 12); /* First, write back & invalidate */ - data0 = ctrl_inl(addr0); - ctrl_outl(data0&~(SH_CACHE_VALID|SH_CACHE_UPDATED), addr0); - data1 = ctrl_inl(addr1); - ctrl_outl(data1&~(SH_CACHE_VALID|SH_CACHE_UPDATED), addr1); + data0 = __raw_readl(addr0); + __raw_writel(data0&~(SH_CACHE_VALID|SH_CACHE_UPDATED), addr0); + data1 = __raw_readl(addr1); + __raw_writel(data1&~(SH_CACHE_VALID|SH_CACHE_UPDATED), addr1); /* Next, check if there's shadow or not */ - data0 = ctrl_inl(addr0); + data0 = __raw_readl(addr0); data0 ^= SH_CACHE_VALID; - ctrl_outl(data0, addr0); - data1 = ctrl_inl(addr1); + __raw_writel(data0, addr0); + data1 = __raw_readl(addr1); data2 = data1 ^ SH_CACHE_VALID; - ctrl_outl(data2, addr1); - data3 = ctrl_inl(addr0); + __raw_writel(data2, addr1); + data3 = __raw_readl(addr0); /* Lastly, invaliate them. */ - ctrl_outl(data0&~SH_CACHE_VALID, addr0); - ctrl_outl(data2&~SH_CACHE_VALID, addr1); + __raw_writel(data0&~SH_CACHE_VALID, addr0); + __raw_writel(data2&~SH_CACHE_VALID, addr1); back_to_cached(); @@ -94,9 +94,9 @@ int detect_cpu_and_cache_system(void) boot_cpu_data.dcache.way_incr = (1 << 13); boot_cpu_data.dcache.entry_mask = 0x1ff0; boot_cpu_data.dcache.sets = 512; - ctrl_outl(CCR_CACHE_32KB, CCR3_REG); + __raw_writel(CCR_CACHE_32KB, CCR3_REG); #else - ctrl_outl(CCR_CACHE_16KB, CCR3_REG); + __raw_writel(CCR_CACHE_16KB, CCR3_REG); #endif #endif } diff --git a/arch/sh/kernel/cpu/sh3/setup-sh3.c b/arch/sh/kernel/cpu/sh3/setup-sh3.c index c988468..53be70b 100644 --- a/arch/sh/kernel/cpu/sh3/setup-sh3.c +++ b/arch/sh/kernel/cpu/sh3/setup-sh3.c @@ -58,7 +58,7 @@ static DECLARE_INTC_DESC_ACK(intc_desc_irq45, "sh3-irq45", void __init plat_irq_setup_pins(int mode) { if (mode == IRQ_MODE_IRQ) { - ctrl_outw(ctrl_inw(INTC_ICR1) & ~INTC_ICR1_IRQLVL, INTC_ICR1); + __raw_writew(__raw_readw(INTC_ICR1) & ~INTC_ICR1_IRQLVL, INTC_ICR1); register_intc_controller(&intc_desc_irq0123); return; } diff --git a/arch/sh/kernel/cpu/sh4/clock-sh4-202.c b/arch/sh/kernel/cpu/sh4/clock-sh4-202.c index 21421e3..6b80850 100644 --- a/arch/sh/kernel/cpu/sh4/clock-sh4-202.c +++ b/arch/sh/kernel/cpu/sh4/clock-sh4-202.c @@ -23,7 +23,7 @@ static int frqcr3_values[] = { 0, 1, 2, 3, 4, 5, 6 }; static unsigned long emi_clk_recalc(struct clk *clk) { - int idx = ctrl_inl(CPG2_FRQCR3) & 0x0007; + int idx = __raw_readl(CPG2_FRQCR3) & 0x0007; return clk->parent->rate / frqcr3_divisors[idx]; } @@ -52,7 +52,7 @@ static struct clk sh4202_emi_clk = { static unsigned long femi_clk_recalc(struct clk *clk) { - int idx = (ctrl_inl(CPG2_FRQCR3) >> 3) & 0x0007; + int idx = (__raw_readl(CPG2_FRQCR3) >> 3) & 0x0007; return clk->parent->rate / frqcr3_divisors[idx]; } @@ -92,7 +92,7 @@ static void shoc_clk_init(struct clk *clk) static unsigned long shoc_clk_recalc(struct clk *clk) { - int idx = (ctrl_inl(CPG2_FRQCR3) >> 6) & 0x0007; + int idx = (__raw_readl(CPG2_FRQCR3) >> 6) & 0x0007; return clk->parent->rate / frqcr3_divisors[idx]; } @@ -122,10 +122,10 @@ static int shoc_clk_set_rate(struct clk *clk, unsigned long rate, int algo_id) tmp = frqcr3_lookup(clk, rate); - frqcr3 = ctrl_inl(CPG2_FRQCR3); + frqcr3 = __raw_readl(CPG2_FRQCR3); frqcr3 &= ~(0x0007 << 6); frqcr3 |= tmp << 6; - ctrl_outl(frqcr3, CPG2_FRQCR3); + __raw_writel(frqcr3, CPG2_FRQCR3); clk->rate = clk->parent->rate / frqcr3_divisors[tmp]; diff --git a/arch/sh/kernel/cpu/sh4/clock-sh4.c b/arch/sh/kernel/cpu/sh4/clock-sh4.c index 73294d9..5add75c 100644 --- a/arch/sh/kernel/cpu/sh4/clock-sh4.c +++ b/arch/sh/kernel/cpu/sh4/clock-sh4.c @@ -28,7 +28,7 @@ static int pfc_divisors[] = { 2, 3, 4, 6, 8, 2, 2, 2 }; static void master_clk_init(struct clk *clk) { - clk->rate *= pfc_divisors[ctrl_inw(FRQCR) & 0x0007]; + clk->rate *= pfc_divisors[__raw_readw(FRQCR) & 0x0007]; } static struct clk_ops sh4_master_clk_ops = { @@ -37,7 +37,7 @@ static struct clk_ops sh4_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) & 0x0007); + int idx = (__raw_readw(FRQCR) & 0x0007); return clk->parent->rate / pfc_divisors[idx]; } @@ -47,7 +47,7 @@ static struct clk_ops sh4_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) >> 3) & 0x0007; + int idx = (__raw_readw(FRQCR) >> 3) & 0x0007; return clk->parent->rate / bfc_divisors[idx]; } @@ -57,7 +57,7 @@ static struct clk_ops sh4_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(FRQCR) >> 6) & 0x0007; + int idx = (__raw_readw(FRQCR) >> 6) & 0x0007; return clk->parent->rate / ifc_divisors[idx]; } diff --git a/arch/sh/kernel/cpu/sh4/probe.c b/arch/sh/kernel/cpu/sh4/probe.c index cc02b31..822977a 100644 --- a/arch/sh/kernel/cpu/sh4/probe.c +++ b/arch/sh/kernel/cpu/sh4/probe.c @@ -28,9 +28,9 @@ int __init detect_cpu_and_cache_system(void) [9] = (1 << 16) }; - pvr = (ctrl_inl(CCN_PVR) >> 8) & 0xffffff; - prr = (ctrl_inl(CCN_PRR) >> 4) & 0xff; - cvr = (ctrl_inl(CCN_CVR)); + pvr = (__raw_readl(CCN_PVR) >> 8) & 0xffffff; + prr = (__raw_readl(CCN_PRR) >> 4) & 0xff; + cvr = (__raw_readl(CCN_CVR)); /* * Setup some sane SH-4 defaults for the icache diff --git a/arch/sh/kernel/cpu/sh4/setup-sh4-202.c b/arch/sh/kernel/cpu/sh4/setup-sh4-202.c index 4b73371..b9b7e10 100644 --- a/arch/sh/kernel/cpu/sh4/setup-sh4-202.c +++ b/arch/sh/kernel/cpu/sh4/setup-sh4-202.c @@ -198,7 +198,7 @@ void __init plat_irq_setup_pins(int mode) { switch (mode) { case IRQ_MODE_IRQ: /* individual interrupt mode for IRL3-0 */ - ctrl_outw(ctrl_inw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR); + __raw_writew(__raw_readw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR); register_intc_controller(&intc_desc_irlm); break; default: diff --git a/arch/sh/kernel/cpu/sh4/setup-sh7750.c b/arch/sh/kernel/cpu/sh4/setup-sh7750.c index b2a9df1..ffd79e5 100644 --- a/arch/sh/kernel/cpu/sh4/setup-sh7750.c +++ b/arch/sh/kernel/cpu/sh4/setup-sh7750.c @@ -442,7 +442,7 @@ void __init plat_irq_setup_pins(int mode) switch (mode) { case IRQ_MODE_IRQ: /* individual interrupt mode for IRL3-0 */ - ctrl_outw(ctrl_inw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR); + __raw_writew(__raw_readw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR); register_intc_controller(&intc_desc_irlm); break; default: diff --git a/arch/sh/kernel/cpu/sh4/setup-sh7760.c b/arch/sh/kernel/cpu/sh4/setup-sh7760.c index 5b74cc0..a16eb36 100644 --- a/arch/sh/kernel/cpu/sh4/setup-sh7760.c +++ b/arch/sh/kernel/cpu/sh4/setup-sh7760.c @@ -319,7 +319,7 @@ void __init plat_irq_setup_pins(int mode) { switch (mode) { case IRQ_MODE_IRQ: - ctrl_outw(ctrl_inw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR); + __raw_writew(__raw_readw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR); register_intc_controller(&intc_desc_irq); break; default: diff --git a/arch/sh/kernel/cpu/sh4/sq.c b/arch/sh/kernel/cpu/sh4/sq.c index 8a8a993..97aea9d 100644 --- a/arch/sh/kernel/cpu/sh4/sq.c +++ b/arch/sh/kernel/cpu/sh4/sq.c @@ -43,9 +43,9 @@ static unsigned long *sq_bitmap; #define store_queue_barrier() \ do { \ - (void)ctrl_inl(P4SEG_STORE_QUE); \ - ctrl_outl(0, P4SEG_STORE_QUE + 0); \ - ctrl_outl(0, P4SEG_STORE_QUE + 8); \ + (void)__raw_readl(P4SEG_STORE_QUE); \ + __raw_writel(0, P4SEG_STORE_QUE + 0); \ + __raw_writel(0, P4SEG_STORE_QUE + 8); \ } while (0); /** @@ -123,8 +123,8 @@ static int __sq_remap(struct sq_mapping *map, unsigned long flags) * straightforward, as we can just load up each queue's QACR with * the physical address appropriately masked. */ - ctrl_outl(((map->addr >> 26) << 2) & 0x1c, SQ_QACR0); - ctrl_outl(((map->addr >> 26) << 2) & 0x1c, SQ_QACR1); + __raw_writel(((map->addr >> 26) << 2) & 0x1c, SQ_QACR0); + __raw_writel(((map->addr >> 26) << 2) & 0x1c, SQ_QACR1); #endif return 0; diff --git a/arch/sh/kernel/cpu/sh4a/clock-sh7757.c b/arch/sh/kernel/cpu/sh4a/clock-sh7757.c index ddc235c..86aae60 100644 --- a/arch/sh/kernel/cpu/sh4a/clock-sh7757.c +++ b/arch/sh/kernel/cpu/sh4a/clock-sh7757.c @@ -35,7 +35,7 @@ static struct clk_ops sh7757_master_clk_ops = { static void module_clk_recalc(struct clk *clk) { - int idx = ctrl_inl(FRQCR) & 0x0000000f; + int idx = __raw_readl(FRQCR) & 0x0000000f; clk->rate = clk->parent->rate / p1fc_divisors[idx]; } @@ -45,7 +45,7 @@ static struct clk_ops sh7757_module_clk_ops = { static void bus_clk_recalc(struct clk *clk) { - int idx = (ctrl_inl(FRQCR) >> 8) & 0x0000000f; + int idx = (__raw_readl(FRQCR) >> 8) & 0x0000000f; clk->rate = clk->parent->rate / bfc_divisors[idx]; } @@ -55,7 +55,7 @@ static struct clk_ops sh7757_bus_clk_ops = { static void cpu_clk_recalc(struct clk *clk) { - int idx = (ctrl_inl(FRQCR) >> 20) & 0x0000000f; + int idx = (__raw_readl(FRQCR) >> 20) & 0x0000000f; clk->rate = clk->parent->rate / ifc_divisors[idx]; } @@ -78,7 +78,7 @@ void __init arch_init_clk_ops(struct clk_ops **ops, int idx) static void shyway_clk_recalc(struct clk *clk) { - int idx = (ctrl_inl(FRQCR) >> 12) & 0x0000000f; + int idx = (__raw_readl(FRQCR) >> 12) & 0x0000000f; clk->rate = clk->parent->rate / sfc_divisors[idx]; } diff --git a/arch/sh/kernel/cpu/sh4a/clock-sh7763.c b/arch/sh/kernel/cpu/sh4a/clock-sh7763.c index 370cd47..9f40116 100644 --- a/arch/sh/kernel/cpu/sh4a/clock-sh7763.c +++ b/arch/sh/kernel/cpu/sh4a/clock-sh7763.c @@ -22,7 +22,7 @@ static int cfc_divisors[] = { 1, 1, 4, 1, 1, 1, 1, 1 }; static void master_clk_init(struct clk *clk) { - clk->rate *= p0fc_divisors[(ctrl_inl(FRQCR) >> 4) & 0x07]; + clk->rate *= p0fc_divisors[(__raw_readl(FRQCR) >> 4) & 0x07]; } static struct clk_ops sh7763_master_clk_ops = { @@ -31,7 +31,7 @@ static struct clk_ops sh7763_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> 4) & 0x07); + int idx = ((__raw_readl(FRQCR) >> 4) & 0x07); return clk->parent->rate / p0fc_divisors[idx]; } @@ -41,7 +41,7 @@ static struct clk_ops sh7763_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> 16) & 0x07); + int idx = ((__raw_readl(FRQCR) >> 16) & 0x07); return clk->parent->rate / bfc_divisors[idx]; } @@ -68,7 +68,7 @@ void __init arch_init_clk_ops(struct clk_ops **ops, int idx) static unsigned long shyway_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> 20) & 0x07); + int idx = ((__raw_readl(FRQCR) >> 20) & 0x07); return clk->parent->rate / cfc_divisors[idx]; } diff --git a/arch/sh/kernel/cpu/sh4a/clock-sh7770.c b/arch/sh/kernel/cpu/sh4a/clock-sh7770.c index e0b8967..9e33543 100644 --- a/arch/sh/kernel/cpu/sh4a/clock-sh7770.c +++ b/arch/sh/kernel/cpu/sh4a/clock-sh7770.c @@ -21,7 +21,7 @@ static int pfc_divisors[] = { 1, 8, 1,10,12,16, 1, 1 }; static void master_clk_init(struct clk *clk) { - clk->rate *= pfc_divisors[(ctrl_inl(FRQCR) >> 28) & 0x000f]; + clk->rate *= pfc_divisors[(__raw_readl(FRQCR) >> 28) & 0x000f]; } static struct clk_ops sh7770_master_clk_ops = { @@ -30,7 +30,7 @@ static struct clk_ops sh7770_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> 28) & 0x000f); + int idx = ((__raw_readl(FRQCR) >> 28) & 0x000f); return clk->parent->rate / pfc_divisors[idx]; } @@ -40,7 +40,7 @@ static struct clk_ops sh7770_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = (ctrl_inl(FRQCR) & 0x000f); + int idx = (__raw_readl(FRQCR) & 0x000f); return clk->parent->rate / bfc_divisors[idx]; } @@ -50,7 +50,7 @@ static struct clk_ops sh7770_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> 24) & 0x000f); + int idx = ((__raw_readl(FRQCR) >> 24) & 0x000f); return clk->parent->rate / ifc_divisors[idx]; } diff --git a/arch/sh/kernel/cpu/sh4a/clock-sh7780.c b/arch/sh/kernel/cpu/sh4a/clock-sh7780.c index a249d82..150963a 100644 --- a/arch/sh/kernel/cpu/sh4a/clock-sh7780.c +++ b/arch/sh/kernel/cpu/sh4a/clock-sh7780.c @@ -22,7 +22,7 @@ static int cfc_divisors[] = { 1, 1, 4, 1, 6, 1, 1, 1 }; static void master_clk_init(struct clk *clk) { - clk->rate *= pfc_divisors[ctrl_inl(FRQCR) & 0x0003]; + clk->rate *= pfc_divisors[__raw_readl(FRQCR) & 0x0003]; } static struct clk_ops sh7780_master_clk_ops = { @@ -31,7 +31,7 @@ static struct clk_ops sh7780_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = (ctrl_inl(FRQCR) & 0x0003); + int idx = (__raw_readl(FRQCR) & 0x0003); return clk->parent->rate / pfc_divisors[idx]; } @@ -41,7 +41,7 @@ static struct clk_ops sh7780_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> 16) & 0x0007); + int idx = ((__raw_readl(FRQCR) >> 16) & 0x0007); return clk->parent->rate / bfc_divisors[idx]; } @@ -51,7 +51,7 @@ static struct clk_ops sh7780_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> 24) & 0x0001); + int idx = ((__raw_readl(FRQCR) >> 24) & 0x0001); return clk->parent->rate / ifc_divisors[idx]; } @@ -74,7 +74,7 @@ void __init arch_init_clk_ops(struct clk_ops **ops, int idx) static unsigned long shyway_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> 20) & 0x0007); + int idx = ((__raw_readl(FRQCR) >> 20) & 0x0007); return clk->parent->rate / cfc_divisors[idx]; } diff --git a/arch/sh/kernel/cpu/sh4a/clock-shx3.c b/arch/sh/kernel/cpu/sh4a/clock-shx3.c index 23c27d3..e75c57b 100644 --- a/arch/sh/kernel/cpu/sh4a/clock-shx3.c +++ b/arch/sh/kernel/cpu/sh4a/clock-shx3.c @@ -33,7 +33,7 @@ static int cfc_divisors[] = { 1, 1, 4, 6 }; static void master_clk_init(struct clk *clk) { - clk->rate *= pfc_divisors[(ctrl_inl(FRQCR) >> PFC_POS) & PFC_MSK]; + clk->rate *= pfc_divisors[(__raw_readl(FRQCR) >> PFC_POS) & PFC_MSK]; } static struct clk_ops shx3_master_clk_ops = { @@ -42,7 +42,7 @@ static struct clk_ops shx3_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> PFC_POS) & PFC_MSK); + int idx = ((__raw_readl(FRQCR) >> PFC_POS) & PFC_MSK); return clk->parent->rate / pfc_divisors[idx]; } @@ -52,7 +52,7 @@ static struct clk_ops shx3_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> BFC_POS) & BFC_MSK); + int idx = ((__raw_readl(FRQCR) >> BFC_POS) & BFC_MSK); return clk->parent->rate / bfc_divisors[idx]; } @@ -62,7 +62,7 @@ static struct clk_ops shx3_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> IFC_POS) & IFC_MSK); + int idx = ((__raw_readl(FRQCR) >> IFC_POS) & IFC_MSK); return clk->parent->rate / ifc_divisors[idx]; } @@ -85,7 +85,7 @@ void __init arch_init_clk_ops(struct clk_ops **ops, int idx) static unsigned long shyway_clk_recalc(struct clk *clk) { - int idx = ((ctrl_inl(FRQCR) >> CFC_POS) & CFC_MSK); + int idx = ((__raw_readl(FRQCR) >> CFC_POS) & CFC_MSK); return clk->parent->rate / cfc_divisors[idx]; } diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7723.c b/arch/sh/kernel/cpu/sh4a/setup-sh7723.c index 9d42625..9e8620e 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7723.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7723.c @@ -596,7 +596,7 @@ void __init plat_early_device_setup(void) void l2_cache_init(void) { /* Enable L2 cache */ - ctrl_outl(L2_CACHE_ENABLE, RAMCR); + __raw_writel(L2_CACHE_ENABLE, RAMCR); } enum { diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7724.c b/arch/sh/kernel/cpu/sh4a/setup-sh7724.c index 578e9f8..aa0f6e9 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7724.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7724.c @@ -718,7 +718,7 @@ void __init plat_early_device_setup(void) void l2_cache_init(void) { /* Enable L2 cache */ - ctrl_outl(L2_CACHE_ENABLE, RAMCR); + __raw_writel(L2_CACHE_ENABLE, RAMCR); } enum { diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7757.c b/arch/sh/kernel/cpu/sh4a/setup-sh7757.c index 37e32ef..e75edf5 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7757.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7757.c @@ -487,17 +487,17 @@ static DECLARE_INTC_DESC(intc_desc_irl4567, "sh7757-irl4567", vectors_irl4567, void __init plat_irq_setup(void) { /* disable IRQ3-0 + IRQ7-4 */ - ctrl_outl(0xff000000, INTC_INTMSK0); + __raw_writel(0xff000000, INTC_INTMSK0); /* disable IRL3-0 + IRL7-4 */ - ctrl_outl(0xc0000000, INTC_INTMSK1); - ctrl_outl(0xfffefffe, INTC_INTMSK2); + __raw_writel(0xc0000000, INTC_INTMSK1); + __raw_writel(0xfffefffe, INTC_INTMSK2); /* select IRL mode for IRL3-0 + IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) & ~0x00c00000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) & ~0x00c00000, INTC_ICR0); /* disable holding function, ie enable "SH-4 Mode" */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00200000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00200000, INTC_ICR0); register_intc_controller(&intc_desc); } @@ -507,32 +507,32 @@ void __init plat_irq_setup_pins(int mode) switch (mode) { case IRQ_MODE_IRQ7654: /* select IRQ mode for IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00400000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00400000, INTC_ICR0); register_intc_controller(&intc_desc_irq4567); break; case IRQ_MODE_IRQ3210: /* select IRQ mode for IRL3-0 */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00800000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00800000, INTC_ICR0); register_intc_controller(&intc_desc_irq0123); break; case IRQ_MODE_IRL7654: /* enable IRL7-4 but don't provide any masking */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); - ctrl_outl(0x0000fffe, INTC_INTMSKCLR2); + __raw_writel(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x0000fffe, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL3210: /* enable IRL0-3 but don't provide any masking */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); - ctrl_outl(0xfffe0000, INTC_INTMSKCLR2); + __raw_writel(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0xfffe0000, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL7654_MASK: /* enable IRL7-4 and mask using cpu intc controller */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x40000000, INTC_INTMSKCLR1); register_intc_controller(&intc_desc_irl4567); break; case IRQ_MODE_IRL3210_MASK: /* enable IRL0-3 and mask using cpu intc controller */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0x80000000, INTC_INTMSKCLR1); register_intc_controller(&intc_desc_irl0123); break; default: diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7763.c b/arch/sh/kernel/cpu/sh4a/setup-sh7763.c index 6aba26f..7f6b0a5 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7763.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7763.c @@ -538,11 +538,11 @@ static DECLARE_INTC_DESC(intc_irl3210_desc, "sh7763-irl3210", irl_vectors, void __init plat_irq_setup(void) { /* disable IRQ7-0 */ - ctrl_outl(0xff000000, INTC_INTMSK0); + __raw_writel(0xff000000, INTC_INTMSK0); /* disable IRL3-0 + IRL7-4 */ - ctrl_outl(0xc0000000, INTC_INTMSK1); - ctrl_outl(0xfffefffe, INTC_INTMSK2); + __raw_writel(0xc0000000, INTC_INTMSK1); + __raw_writel(0xfffefffe, INTC_INTMSK2); register_intc_controller(&intc_desc); } @@ -552,27 +552,27 @@ void __init plat_irq_setup_pins(int mode) switch (mode) { case IRQ_MODE_IRQ: /* select IRQ mode for IRL3-0 + IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00c00000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00c00000, INTC_ICR0); register_intc_controller(&intc_irq_desc); break; case IRQ_MODE_IRL7654: /* enable IRL7-4 but don't provide any masking */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); - ctrl_outl(0x0000fffe, INTC_INTMSKCLR2); + __raw_writel(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x0000fffe, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL3210: /* enable IRL0-3 but don't provide any masking */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); - ctrl_outl(0xfffe0000, INTC_INTMSKCLR2); + __raw_writel(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0xfffe0000, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL7654_MASK: /* enable IRL7-4 and mask using cpu intc controller */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x40000000, INTC_INTMSKCLR1); register_intc_controller(&intc_irl7654_desc); break; case IRQ_MODE_IRL3210_MASK: /* enable IRL0-3 and mask using cpu intc controller */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0x80000000, INTC_INTMSKCLR1); register_intc_controller(&intc_irl3210_desc); break; default: diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7770.c b/arch/sh/kernel/cpu/sh4a/setup-sh7770.c index c1643bc..86d681e 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7770.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7770.c @@ -694,17 +694,17 @@ static DECLARE_INTC_DESC(intc_irl3210_desc, "sh7780-irl3210", irl_vectors, void __init plat_irq_setup(void) { /* disable IRQ7-0 */ - ctrl_outl(0xff000000, INTC_INTMSK0); + __raw_writel(0xff000000, INTC_INTMSK0); /* disable IRL3-0 + IRL7-4 */ - ctrl_outl(0xc0000000, INTC_INTMSK1); - ctrl_outl(0xfffefffe, INTC_INTMSK2); + __raw_writel(0xc0000000, INTC_INTMSK1); + __raw_writel(0xfffefffe, INTC_INTMSK2); /* select IRL mode for IRL3-0 + IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) & ~0x00c00000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) & ~0x00c00000, INTC_ICR0); /* disable holding function, ie enable "SH-4 Mode" */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00200000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00200000, INTC_ICR0); register_intc_controller(&intc_desc); } @@ -714,27 +714,27 @@ void __init plat_irq_setup_pins(int mode) switch (mode) { case IRQ_MODE_IRQ: /* select IRQ mode for IRL3-0 + IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00c00000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00c00000, INTC_ICR0); register_intc_controller(&intc_irq_desc); break; case IRQ_MODE_IRL7654: /* enable IRL7-4 but don't provide any masking */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); - ctrl_outl(0x0000fffe, INTC_INTMSKCLR2); + __raw_writel(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x0000fffe, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL3210: /* enable IRL0-3 but don't provide any masking */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); - ctrl_outl(0xfffe0000, INTC_INTMSKCLR2); + __raw_writel(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0xfffe0000, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL7654_MASK: /* enable IRL7-4 and mask using cpu intc controller */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x40000000, INTC_INTMSKCLR1); register_intc_controller(&intc_irl7654_desc); break; case IRQ_MODE_IRL3210_MASK: /* enable IRL0-3 and mask using cpu intc controller */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0x80000000, INTC_INTMSKCLR1); register_intc_controller(&intc_irl3210_desc); break; default: diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7780.c b/arch/sh/kernel/cpu/sh4a/setup-sh7780.c index c310558..f8f2161 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7780.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7780.c @@ -461,17 +461,17 @@ static DECLARE_INTC_DESC(intc_irl3210_desc, "sh7780-irl3210", irl_vectors, void __init plat_irq_setup(void) { /* disable IRQ7-0 */ - ctrl_outl(0xff000000, INTC_INTMSK0); + __raw_writel(0xff000000, INTC_INTMSK0); /* disable IRL3-0 + IRL7-4 */ - ctrl_outl(0xc0000000, INTC_INTMSK1); - ctrl_outl(0xfffefffe, INTC_INTMSK2); + __raw_writel(0xc0000000, INTC_INTMSK1); + __raw_writel(0xfffefffe, INTC_INTMSK2); /* select IRL mode for IRL3-0 + IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) & ~0x00c00000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) & ~0x00c00000, INTC_ICR0); /* disable holding function, ie enable "SH-4 Mode" */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00200000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00200000, INTC_ICR0); register_intc_controller(&intc_desc); } @@ -481,27 +481,27 @@ void __init plat_irq_setup_pins(int mode) switch (mode) { case IRQ_MODE_IRQ: /* select IRQ mode for IRL3-0 + IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00c00000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00c00000, INTC_ICR0); register_intc_controller(&intc_irq_desc); break; case IRQ_MODE_IRL7654: /* enable IRL7-4 but don't provide any masking */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); - ctrl_outl(0x0000fffe, INTC_INTMSKCLR2); + __raw_writel(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x0000fffe, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL3210: /* enable IRL0-3 but don't provide any masking */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); - ctrl_outl(0xfffe0000, INTC_INTMSKCLR2); + __raw_writel(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0xfffe0000, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL7654_MASK: /* enable IRL7-4 and mask using cpu intc controller */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x40000000, INTC_INTMSKCLR1); register_intc_controller(&intc_irl7654_desc); break; case IRQ_MODE_IRL3210_MASK: /* enable IRL0-3 and mask using cpu intc controller */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0x80000000, INTC_INTMSKCLR1); register_intc_controller(&intc_irl3210_desc); break; default: diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7785.c b/arch/sh/kernel/cpu/sh4a/setup-sh7785.c index f685b9b..23448d8 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7785.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7785.c @@ -541,17 +541,17 @@ static DECLARE_INTC_DESC(intc_desc_irl4567, "sh7785-irl4567", vectors_irl4567, void __init plat_irq_setup(void) { /* disable IRQ3-0 + IRQ7-4 */ - ctrl_outl(0xff000000, INTC_INTMSK0); + __raw_writel(0xff000000, INTC_INTMSK0); /* disable IRL3-0 + IRL7-4 */ - ctrl_outl(0xc0000000, INTC_INTMSK1); - ctrl_outl(0xfffefffe, INTC_INTMSK2); + __raw_writel(0xc0000000, INTC_INTMSK1); + __raw_writel(0xfffefffe, INTC_INTMSK2); /* select IRL mode for IRL3-0 + IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) & ~0x00c00000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) & ~0x00c00000, INTC_ICR0); /* disable holding function, ie enable "SH-4 Mode" */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00200000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00200000, INTC_ICR0); register_intc_controller(&intc_desc); } @@ -561,32 +561,32 @@ void __init plat_irq_setup_pins(int mode) switch (mode) { case IRQ_MODE_IRQ7654: /* select IRQ mode for IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00400000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00400000, INTC_ICR0); register_intc_controller(&intc_desc_irq4567); break; case IRQ_MODE_IRQ3210: /* select IRQ mode for IRL3-0 */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00800000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00800000, INTC_ICR0); register_intc_controller(&intc_desc_irq0123); break; case IRQ_MODE_IRL7654: /* enable IRL7-4 but don't provide any masking */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); - ctrl_outl(0x0000fffe, INTC_INTMSKCLR2); + __raw_writel(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x0000fffe, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL3210: /* enable IRL0-3 but don't provide any masking */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); - ctrl_outl(0xfffe0000, INTC_INTMSKCLR2); + __raw_writel(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0xfffe0000, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL7654_MASK: /* enable IRL7-4 and mask using cpu intc controller */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x40000000, INTC_INTMSKCLR1); register_intc_controller(&intc_desc_irl4567); break; case IRQ_MODE_IRL3210_MASK: /* enable IRL0-3 and mask using cpu intc controller */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0x80000000, INTC_INTMSKCLR1); register_intc_controller(&intc_desc_irl0123); break; default: diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7786.c b/arch/sh/kernel/cpu/sh4a/setup-sh7786.c index 7167348..7e58532 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7786.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7786.c @@ -867,14 +867,14 @@ static DECLARE_INTC_DESC(intc_desc_irl4567, "sh7786-irl4567", vectors_irl4567, void __init plat_irq_setup(void) { /* disable IRQ3-0 + IRQ7-4 */ - ctrl_outl(0xff000000, INTC_INTMSK0); + __raw_writel(0xff000000, INTC_INTMSK0); /* disable IRL3-0 + IRL7-4 */ - ctrl_outl(0xc0000000, INTC_INTMSK1); - ctrl_outl(0xfffefffe, INTC_INTMSK2); + __raw_writel(0xc0000000, INTC_INTMSK1); + __raw_writel(0xfffefffe, INTC_INTMSK2); /* select IRL mode for IRL3-0 + IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) & ~0x00c00000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) & ~0x00c00000, INTC_ICR0); register_intc_controller(&intc_desc); } @@ -884,32 +884,32 @@ void __init plat_irq_setup_pins(int mode) switch (mode) { case IRQ_MODE_IRQ7654: /* select IRQ mode for IRL7-4 */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00400000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00400000, INTC_ICR0); register_intc_controller(&intc_desc_irq4567); break; case IRQ_MODE_IRQ3210: /* select IRQ mode for IRL3-0 */ - ctrl_outl(ctrl_inl(INTC_ICR0) | 0x00800000, INTC_ICR0); + __raw_writel(__raw_readl(INTC_ICR0) | 0x00800000, INTC_ICR0); register_intc_controller(&intc_desc_irq0123); break; case IRQ_MODE_IRL7654: /* enable IRL7-4 but don't provide any masking */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); - ctrl_outl(0x0000fffe, INTC_INTMSKCLR2); + __raw_writel(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x0000fffe, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL3210: /* enable IRL0-3 but don't provide any masking */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); - ctrl_outl(0xfffe0000, INTC_INTMSKCLR2); + __raw_writel(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0xfffe0000, INTC_INTMSKCLR2); break; case IRQ_MODE_IRL7654_MASK: /* enable IRL7-4 and mask using cpu intc controller */ - ctrl_outl(0x40000000, INTC_INTMSKCLR1); + __raw_writel(0x40000000, INTC_INTMSKCLR1); register_intc_controller(&intc_desc_irl4567); break; case IRQ_MODE_IRL3210_MASK: /* enable IRL0-3 and mask using cpu intc controller */ - ctrl_outl(0x80000000, INTC_INTMSKCLR1); + __raw_writel(0x80000000, INTC_INTMSKCLR1); register_intc_controller(&intc_desc_irl0123); break; default: diff --git a/arch/sh/kernel/cpu/sh5/clock-sh5.c b/arch/sh/kernel/cpu/sh5/clock-sh5.c index 7f864eb..9cfc19b 100644 --- a/arch/sh/kernel/cpu/sh5/clock-sh5.c +++ b/arch/sh/kernel/cpu/sh5/clock-sh5.c @@ -24,7 +24,7 @@ static unsigned long cprc_base; static void master_clk_init(struct clk *clk) { - int idx = (ctrl_inl(cprc_base + 0x00) >> 6) & 0x0007; + int idx = (__raw_readl(cprc_base + 0x00) >> 6) & 0x0007; clk->rate *= ifc_table[idx]; } @@ -34,7 +34,7 @@ static struct clk_ops sh5_master_clk_ops = { static unsigned long module_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(cprc_base) >> 12) & 0x0007; + int idx = (__raw_readw(cprc_base) >> 12) & 0x0007; return clk->parent->rate / ifc_table[idx]; } @@ -44,7 +44,7 @@ static struct clk_ops sh5_module_clk_ops = { static unsigned long bus_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(cprc_base) >> 3) & 0x0007; + int idx = (__raw_readw(cprc_base) >> 3) & 0x0007; return clk->parent->rate / ifc_table[idx]; } @@ -54,7 +54,7 @@ static struct clk_ops sh5_bus_clk_ops = { static unsigned long cpu_clk_recalc(struct clk *clk) { - int idx = (ctrl_inw(cprc_base) & 0x0007); + int idx = (__raw_readw(cprc_base) & 0x0007); return clk->parent->rate / ifc_table[idx]; } diff --git a/arch/sh/kernel/io_trapped.c b/arch/sh/kernel/io_trapped.c index 69be603..16b83cc 100644 --- a/arch/sh/kernel/io_trapped.c +++ b/arch/sh/kernel/io_trapped.c @@ -184,31 +184,31 @@ static unsigned long long copy_word(unsigned long src_addr, int src_len, switch (src_len) { case 1: - tmp = ctrl_inb(src_addr); + tmp = __raw_readb(src_addr); break; case 2: - tmp = ctrl_inw(src_addr); + tmp = __raw_readw(src_addr); break; case 4: - tmp = ctrl_inl(src_addr); + tmp = __raw_readl(src_addr); break; case 8: - tmp = ctrl_inq(src_addr); + tmp = __raw_readq(src_addr); break; } switch (dst_len) { case 1: - ctrl_outb(tmp, dst_addr); + __raw_writeb(tmp, dst_addr); break; case 2: - ctrl_outw(tmp, dst_addr); + __raw_writew(tmp, dst_addr); break; case 4: - ctrl_outl(tmp, dst_addr); + __raw_writel(tmp, dst_addr); break; case 8: - ctrl_outq(tmp, dst_addr); + __raw_writeq(tmp, dst_addr); break; } diff --git a/arch/sh/kernel/process_32.c b/arch/sh/kernel/process_32.c index b6f43f0..3cb88f1 100644 --- a/arch/sh/kernel/process_32.c +++ b/arch/sh/kernel/process_32.c @@ -41,7 +41,7 @@ void show_regs(struct pt_regs * regs) printk("PC : %08lx SP : %08lx SR : %08lx ", regs->pc, regs->regs[15], regs->sr); #ifdef CONFIG_MMU - printk("TEA : %08x\n", ctrl_inl(MMU_TEA)); + printk("TEA : %08x\n", __raw_readl(MMU_TEA)); #else printk("\n"); #endif diff --git a/arch/sh/kernel/signal_32.c b/arch/sh/kernel/signal_32.c index 6a7cce7..579cd2c 100644 --- a/arch/sh/kernel/signal_32.c +++ b/arch/sh/kernel/signal_32.c @@ -528,7 +528,7 @@ handle_syscall_restart(unsigned long save_r0, struct pt_regs *regs, /* fallthrough */ case -ERESTARTNOINTR: regs->regs[0] = save_r0; - regs->pc -= instruction_size(ctrl_inw(regs->pc - 4)); + regs->pc -= instruction_size(__raw_readw(regs->pc - 4)); break; } } @@ -626,9 +626,9 @@ no_signal: regs->regs[0] == -ERESTARTSYS || regs->regs[0] == -ERESTARTNOINTR) { regs->regs[0] = save_r0; - regs->pc -= instruction_size(ctrl_inw(regs->pc - 4)); + regs->pc -= instruction_size(__raw_readw(regs->pc - 4)); } else if (regs->regs[0] == -ERESTART_RESTARTBLOCK) { - regs->pc -= instruction_size(ctrl_inw(regs->pc - 4)); + regs->pc -= instruction_size(__raw_readw(regs->pc - 4)); regs->regs[3] = __NR_restart_syscall; } } diff --git a/arch/sh/kernel/traps.c b/arch/sh/kernel/traps.c index 7b03633..0830c2a 100644 --- a/arch/sh/kernel/traps.c +++ b/arch/sh/kernel/traps.c @@ -58,7 +58,7 @@ BUILD_TRAP_HANDLER(debug) TRAP_HANDLER_DECL; /* Rewind */ - regs->pc -= instruction_size(ctrl_inw(regs->pc - 4)); + regs->pc -= instruction_size(__raw_readw(regs->pc - 4)); if (notify_die(DIE_TRAP, "debug trap", regs, 0, vec & 0xff, SIGTRAP) == NOTIFY_STOP) @@ -75,7 +75,7 @@ BUILD_TRAP_HANDLER(bug) TRAP_HANDLER_DECL; /* Rewind */ - regs->pc -= instruction_size(ctrl_inw(regs->pc - 4)); + regs->pc -= instruction_size(__raw_readw(regs->pc - 4)); if (notify_die(DIE_TRAP, "bug trap", regs, 0, TRAPA_BUG_OPCODE & 0xff, SIGTRAP) == NOTIFY_STOP) diff --git a/arch/sh/mm/cache-debugfs.c b/arch/sh/mm/cache-debugfs.c index 48ce82e..690ed01 100644 --- a/arch/sh/mm/cache-debugfs.c +++ b/arch/sh/mm/cache-debugfs.c @@ -36,7 +36,7 @@ static int cache_seq_show(struct seq_file *file, void *iter) */ jump_to_uncached(); - ccr = ctrl_inl(CCR); + ccr = __raw_readl(CCR); if ((ccr & CCR_CACHE_ENABLE) == 0) { back_to_cached(); @@ -89,7 +89,7 @@ static int cache_seq_show(struct seq_file *file, void *iter) for (addr = addrstart, line = 0; addr < addrstart + waysize; addr += cache->linesz, line++) { - unsigned long data = ctrl_inl(addr); + unsigned long data = __raw_readl(addr); /* Check the V bit, ignore invalid cachelines */ if ((data & 1) == 0) diff --git a/arch/sh/mm/cache-sh2.c b/arch/sh/mm/cache-sh2.c index 699a71f..defcf71 100644 --- a/arch/sh/mm/cache-sh2.c +++ b/arch/sh/mm/cache-sh2.c @@ -28,10 +28,10 @@ static void sh2__flush_wback_region(void *start, int size) unsigned long addr = CACHE_OC_ADDRESS_ARRAY | (v & 0x00000ff0); int way; for (way = 0; way < 4; way++) { - unsigned long data = ctrl_inl(addr | (way << 12)); + unsigned long data = __raw_readl(addr | (way << 12)); if ((data & CACHE_PHYSADDR_MASK) == (v & CACHE_PHYSADDR_MASK)) { data &= ~SH_CACHE_UPDATED; - ctrl_outl(data, addr | (way << 12)); + __raw_writel(data, addr | (way << 12)); } } } @@ -47,7 +47,7 @@ static void sh2__flush_purge_region(void *start, int size) & ~(L1_CACHE_BYTES-1); for (v = begin; v < end; v+=L1_CACHE_BYTES) - ctrl_outl((v & CACHE_PHYSADDR_MASK), + __raw_writel((v & CACHE_PHYSADDR_MASK), CACHE_OC_ADDRESS_ARRAY | (v & 0x00000ff0) | 0x00000008); } @@ -63,9 +63,9 @@ static void sh2__flush_invalidate_region(void *start, int size) local_irq_save(flags); jump_to_uncached(); - ccr = ctrl_inl(CCR); + ccr = __raw_readl(CCR); ccr |= CCR_CACHE_INVALIDATE; - ctrl_outl(ccr, CCR); + __raw_writel(ccr, CCR); back_to_cached(); local_irq_restore(flags); @@ -78,7 +78,7 @@ static void sh2__flush_invalidate_region(void *start, int size) & ~(L1_CACHE_BYTES-1); for (v = begin; v < end; v+=L1_CACHE_BYTES) - ctrl_outl((v & CACHE_PHYSADDR_MASK), + __raw_writel((v & CACHE_PHYSADDR_MASK), CACHE_OC_ADDRESS_ARRAY | (v & 0x00000ff0) | 0x00000008); #endif } diff --git a/arch/sh/mm/cache-sh2a.c b/arch/sh/mm/cache-sh2a.c index 975899d..1f51225 100644 --- a/arch/sh/mm/cache-sh2a.c +++ b/arch/sh/mm/cache-sh2a.c @@ -32,10 +32,10 @@ static void sh2a__flush_wback_region(void *start, int size) unsigned long addr = CACHE_OC_ADDRESS_ARRAY | (v & 0x000007f0); int way; for (way = 0; way < 4; way++) { - unsigned long data = ctrl_inl(addr | (way << 11)); + unsigned long data = __raw_readl(addr | (way << 11)); if ((data & CACHE_PHYSADDR_MASK) == (v & CACHE_PHYSADDR_MASK)) { data &= ~SH_CACHE_UPDATED; - ctrl_outl(data, addr | (way << 11)); + __raw_writel(data, addr | (way << 11)); } } } @@ -58,7 +58,7 @@ static void sh2a__flush_purge_region(void *start, int size) jump_to_uncached(); for (v = begin; v < end; v+=L1_CACHE_BYTES) { - ctrl_outl((v & CACHE_PHYSADDR_MASK), + __raw_writel((v & CACHE_PHYSADDR_MASK), CACHE_OC_ADDRESS_ARRAY | (v & 0x000007f0) | 0x00000008); } back_to_cached(); @@ -78,17 +78,17 @@ static void sh2a__flush_invalidate_region(void *start, int size) jump_to_uncached(); #ifdef CONFIG_CACHE_WRITEBACK - ctrl_outl(ctrl_inl(CCR) | CCR_OCACHE_INVALIDATE, CCR); + __raw_writel(__raw_readl(CCR) | CCR_OCACHE_INVALIDATE, CCR); /* I-cache invalidate */ for (v = begin; v < end; v+=L1_CACHE_BYTES) { - ctrl_outl((v & CACHE_PHYSADDR_MASK), + __raw_writel((v & CACHE_PHYSADDR_MASK), CACHE_IC_ADDRESS_ARRAY | (v & 0x000007f0) | 0x00000008); } #else for (v = begin; v < end; v+=L1_CACHE_BYTES) { - ctrl_outl((v & CACHE_PHYSADDR_MASK), + __raw_writel((v & CACHE_PHYSADDR_MASK), CACHE_IC_ADDRESS_ARRAY | (v & 0x000007f0) | 0x00000008); - ctrl_outl((v & CACHE_PHYSADDR_MASK), + __raw_writel((v & CACHE_PHYSADDR_MASK), CACHE_OC_ADDRESS_ARRAY | (v & 0x000007f0) | 0x00000008); } #endif @@ -115,14 +115,14 @@ static void sh2a_flush_icache_range(void *args) int way; /* O-Cache writeback */ for (way = 0; way < 4; way++) { - unsigned long data = ctrl_inl(CACHE_OC_ADDRESS_ARRAY | addr | (way << 11)); + unsigned long data = __raw_readl(CACHE_OC_ADDRESS_ARRAY | addr | (way << 11)); if ((data & CACHE_PHYSADDR_MASK) == (v & CACHE_PHYSADDR_MASK)) { data &= ~SH_CACHE_UPDATED; - ctrl_outl(data, CACHE_OC_ADDRESS_ARRAY | addr | (way << 11)); + __raw_writel(data, CACHE_OC_ADDRESS_ARRAY | addr | (way << 11)); } } /* I-Cache invalidate */ - ctrl_outl(addr, + __raw_writel(addr, CACHE_IC_ADDRESS_ARRAY | addr | 0x00000008); } diff --git a/arch/sh/mm/cache-sh3.c b/arch/sh/mm/cache-sh3.c index faef80c..e37523f 100644 --- a/arch/sh/mm/cache-sh3.c +++ b/arch/sh/mm/cache-sh3.c @@ -50,12 +50,12 @@ static void sh3__flush_wback_region(void *start, int size) p = __pa(v); addr = addrstart | (v & current_cpu_data.dcache.entry_mask); local_irq_save(flags); - data = ctrl_inl(addr); + data = __raw_readl(addr); if ((data & CACHE_PHYSADDR_MASK) == (p & CACHE_PHYSADDR_MASK)) { data &= ~SH_CACHE_UPDATED; - ctrl_outl(data, addr); + __raw_writel(data, addr); local_irq_restore(flags); break; } @@ -86,7 +86,7 @@ static void sh3__flush_purge_region(void *start, int size) data = (v & 0xfffffc00); /* _Virtual_ address, ~U, ~V */ addr = CACHE_OC_ADDRESS_ARRAY | (v & current_cpu_data.dcache.entry_mask) | SH_CACHE_ASSOC; - ctrl_outl(data, addr); + __raw_writel(data, addr); } } diff --git a/arch/sh/mm/cache-sh4.c b/arch/sh/mm/cache-sh4.c index 87115b3..2cfae81 100644 --- a/arch/sh/mm/cache-sh4.c +++ b/arch/sh/mm/cache-sh4.c @@ -132,9 +132,9 @@ static void flush_icache_all(void) jump_to_uncached(); /* Flush I-cache */ - ccr = ctrl_inl(CCR); + ccr = __raw_readl(CCR); ccr |= CCR_CACHE_ICI; - ctrl_outl(ccr, CCR); + __raw_writel(ccr, CCR); /* * back_to_cached() will take care of the barrier for us, don't add @@ -377,9 +377,9 @@ extern void __weak sh4__flush_region_init(void); void __init sh4_cache_init(void) { printk("PVR=%08x CVR=%08x PRR=%08x\n", - ctrl_inl(CCN_PVR), - ctrl_inl(CCN_CVR), - ctrl_inl(CCN_PRR)); + __raw_readl(CCN_PVR), + __raw_readl(CCN_CVR), + __raw_readl(CCN_PRR)); local_flush_icache_range = sh4_flush_icache_range; local_flush_dcache_page = sh4_flush_dcache_page; diff --git a/arch/sh/mm/cache-sh7705.c b/arch/sh/mm/cache-sh7705.c index 870293e..f498da1 100644 --- a/arch/sh/mm/cache-sh7705.c +++ b/arch/sh/mm/cache-sh7705.c @@ -48,10 +48,10 @@ static inline void cache_wback_all(void) unsigned long data; int v = SH_CACHE_UPDATED | SH_CACHE_VALID; - data = ctrl_inl(addr); + data = __raw_readl(addr); if ((data & v) == v) - ctrl_outl(data & ~v, addr); + __raw_writel(data & ~v, addr); } @@ -115,10 +115,10 @@ static void __flush_dcache_page(unsigned long phys) addr += current_cpu_data.dcache.linesz) { unsigned long data; - data = ctrl_inl(addr) & (0x1ffffC00 | SH_CACHE_VALID); + data = __raw_readl(addr) & (0x1ffffC00 | SH_CACHE_VALID); if (data == phys) { data &= ~(SH_CACHE_VALID | SH_CACHE_UPDATED); - ctrl_outl(data, addr); + __raw_writel(data, addr); } } diff --git a/arch/sh/mm/pmb.c b/arch/sh/mm/pmb.c index 3d5eece..3c9bf5b 100644 --- a/arch/sh/mm/pmb.c +++ b/arch/sh/mm/pmb.c @@ -112,7 +112,7 @@ static void pmb_free(struct pmb_entry *pmbe) static void __set_pmb_entry(unsigned long vpn, unsigned long ppn, unsigned long flags, int pos) { - ctrl_outl(vpn | PMB_V, mk_pmb_addr(pos)); + __raw_writel(vpn | PMB_V, mk_pmb_addr(pos)); #ifdef CONFIG_CACHE_WRITETHROUGH /* @@ -124,7 +124,7 @@ static void __set_pmb_entry(unsigned long vpn, unsigned long ppn, flags |= PMB_WT; #endif - ctrl_outl(ppn | flags | PMB_V, mk_pmb_data(pos)); + __raw_writel(ppn | flags | PMB_V, mk_pmb_data(pos)); } static void set_pmb_entry(struct pmb_entry *pmbe) @@ -146,10 +146,10 @@ static void clear_pmb_entry(struct pmb_entry *pmbe) /* Clear V-bit */ addr = mk_pmb_addr(entry); - ctrl_outl(ctrl_inl(addr) & ~PMB_V, addr); + __raw_writel(__raw_readl(addr) & ~PMB_V, addr); addr = mk_pmb_data(entry); - ctrl_outl(ctrl_inl(addr) & ~PMB_V, addr); + __raw_writel(__raw_readl(addr) & ~PMB_V, addr); back_to_cached(); } @@ -395,7 +395,7 @@ int pmb_init(void) unsigned long vpn, ppn, flags; addr = PMB_DATA + (i << PMB_E_SHIFT); - data = ctrl_inl(addr); + data = __raw_readl(addr); if (!(data & PMB_V)) continue; @@ -408,7 +408,7 @@ int pmb_init(void) data &= ~(PMB_C | PMB_WT); #endif } - ctrl_outl(data, addr); + __raw_writel(data, addr); ppn = data & PMB_PFN_MASK; @@ -416,7 +416,7 @@ int pmb_init(void) flags |= data & PMB_SZ_MASK; addr = PMB_ADDR + (i << PMB_E_SHIFT); - data = ctrl_inl(addr); + data = __raw_readl(addr); vpn = data & PMB_PFN_MASK; @@ -424,12 +424,12 @@ int pmb_init(void) WARN_ON(IS_ERR(pmbe)); } - ctrl_outl(0, PMB_IRMCR); + __raw_writel(0, PMB_IRMCR); /* Flush out the TLB */ - i = ctrl_inl(MMUCR); + i = __raw_readl(MMUCR); i |= MMUCR_TI; - ctrl_outl(i, MMUCR); + __raw_writel(i, MMUCR); back_to_cached(); @@ -454,8 +454,8 @@ static int pmb_seq_show(struct seq_file *file, void *iter) unsigned int size; char *sz_str = NULL; - addr = ctrl_inl(mk_pmb_addr(i)); - data = ctrl_inl(mk_pmb_data(i)); + addr = __raw_readl(mk_pmb_addr(i)); + data = __raw_readl(mk_pmb_data(i)); size = data & PMB_SZ_MASK; sz_str = (size == PMB_SZ_16M) ? " 16MB": diff --git a/arch/sh/mm/tlb-sh3.c b/arch/sh/mm/tlb-sh3.c index ace8e6d..4f5f7cb 100644 --- a/arch/sh/mm/tlb-sh3.c +++ b/arch/sh/mm/tlb-sh3.c @@ -41,14 +41,14 @@ void __update_tlb(struct vm_area_struct *vma, unsigned long address, pte_t pte) /* Set PTEH register */ vpn = (address & MMU_VPN_MASK) | get_asid(); - ctrl_outl(vpn, MMU_PTEH); + __raw_writel(vpn, MMU_PTEH); pteval = pte_val(pte); /* Set PTEL register */ pteval &= _PAGE_FLAGS_HARDWARE_MASK; /* drop software flags */ /* conveniently, we want all the software flags to be 0 anyway */ - ctrl_outl(pteval, MMU_PTEL); + __raw_writel(pteval, MMU_PTEL); /* Load the TLB */ asm volatile("ldtlb": /* no output */ : /* no input */ : "memory"); @@ -75,5 +75,5 @@ void local_flush_tlb_one(unsigned long asid, unsigned long page) } for (i = 0; i < ways; i++) - ctrl_outl(data, addr + (i << 8)); + __raw_writel(data, addr + (i << 8)); } diff --git a/arch/sh/mm/tlb-sh4.c b/arch/sh/mm/tlb-sh4.c index 624c1da..ccac77f 100644 --- a/arch/sh/mm/tlb-sh4.c +++ b/arch/sh/mm/tlb-sh4.c @@ -29,7 +29,7 @@ void __update_tlb(struct vm_area_struct *vma, unsigned long address, pte_t pte) /* Set PTEH register */ vpn = (address & MMU_VPN_MASK) | get_asid(); - ctrl_outl(vpn, MMU_PTEH); + __raw_writel(vpn, MMU_PTEH); pteval = pte.pte_low; @@ -41,13 +41,13 @@ void __update_tlb(struct vm_area_struct *vma, unsigned long address, pte_t pte) * the protection bits (with the exception of the compat-mode SZ * and PR bits, which are cleared) being written out in PTEL. */ - ctrl_outl(pte.pte_high, MMU_PTEA); + __raw_writel(pte.pte_high, MMU_PTEA); #else if (cpu_data->flags & CPU_HAS_PTEA) { /* The last 3 bits and the first one of pteval contains * the PTEA timing control and space attribute bits */ - ctrl_outl(copy_ptea_attributes(pteval), MMU_PTEA); + __raw_writel(copy_ptea_attributes(pteval), MMU_PTEA); } #endif @@ -57,7 +57,7 @@ void __update_tlb(struct vm_area_struct *vma, unsigned long address, pte_t pte) pteval |= _PAGE_WT; #endif /* conveniently, we want all the software flags to be 0 anyway */ - ctrl_outl(pteval, MMU_PTEL); + __raw_writel(pteval, MMU_PTEL); /* Load the TLB */ asm volatile("ldtlb": /* no output */ : /* no input */ : "memory"); @@ -77,6 +77,6 @@ void local_flush_tlb_one(unsigned long asid, unsigned long page) addr = MMU_UTLB_ADDRESS_ARRAY | MMU_PAGE_ASSOC_BIT; data = page | asid; /* VALID bit is off */ jump_to_uncached(); - ctrl_outl(data, addr); + __raw_writel(data, addr); back_to_cached(); } diff --git a/arch/sh/mm/tlbflush_32.c b/arch/sh/mm/tlbflush_32.c index 6f45c1f..004bb3f 100644 --- a/arch/sh/mm/tlbflush_32.c +++ b/arch/sh/mm/tlbflush_32.c @@ -132,9 +132,9 @@ void local_flush_tlb_all(void) * It's same position, bit #2. */ local_irq_save(flags); - status = ctrl_inl(MMUCR); + status = __raw_readl(MMUCR); status |= 0x04; - ctrl_outl(status, MMUCR); + __raw_writel(status, MMUCR); ctrl_barrier(); local_irq_restore(flags); } -- cgit v0.10.2 From 485773f3e401fca31c112c1ff24797e42ff87afd Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 26 Jan 2010 13:02:10 +0900 Subject: sh: flag ctrl_in/outX as __deprecated. These routines are unsuitable for cross-platform use and no new code should be using them, flag them as deprecated in order to give drivers sufficient time to migrate over. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index e4f563f..98c62fa 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -80,16 +80,51 @@ #define writel(v,a) ({ __raw_writel((v),(a)); mb(); }) #define writeq(v,a) ({ __raw_writeq((v),(a)); mb(); }) -/* SuperH on-chip I/O functions */ -#define ctrl_inb __raw_readb -#define ctrl_inw __raw_readw -#define ctrl_inl __raw_readl -#define ctrl_inq __raw_readq - -#define ctrl_outb __raw_writeb -#define ctrl_outw __raw_writew -#define ctrl_outl __raw_writel -#define ctrl_outq __raw_writeq +/* + * Legacy SuperH on-chip I/O functions + * + * These are all deprecated, all new (and especially cross-platform) code + * should be using the __raw_xxx() routines directly. + */ +static inline u8 __deprecated ctrl_inb(unsigned long addr) +{ + return __raw_readb(addr); +} + +static inline u16 __deprecated ctrl_inw(unsigned long addr) +{ + return __raw_readw(addr); +} + +static inline u32 __deprecated ctrl_inl(unsigned long addr) +{ + return __raw_readl(addr); +} + +static inline u64 __deprecated ctrl_inq(unsigned long addr) +{ + return __raw_readq(addr); +} + +static inline void __deprecated ctrl_outb(u8 v, unsigned long addr) +{ + __raw_writeb(v, addr); +} + +static inline void __deprecated ctrl_outw(u16 v, unsigned long addr) +{ + __raw_writew(v, addr); +} + +static inline void __deprecated ctrl_outl(u32 v, unsigned long addr) +{ + __raw_writel(v, addr); +} + +static inline void __deprecated ctrl_outq(u64 v, unsigned long addr) +{ + __raw_writeq(v, addr); +} extern unsigned long generic_io_base; -- cgit v0.10.2 From cbe132a8bdcff0f9afd9060948fb50597c7400b8 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 26 Jan 2010 15:08:49 +1100 Subject: xfs: don't hold onto reserved blocks on remount,ro If we hold onto reserved blocks when doing a remount,ro we end up writing the blocks used count to disk that includes the reserved blocks. Reserved blocks are not actually used, so this results in the values in the superblock being incorrect. Hence if we run xfs_check or xfs_repair -n while the filesystem is mounted remount,ro we end up with an inconsistent filesystem being reported. Also, running xfs_copy on the remount,ro filesystem will result in an inconsistent image being generated. To fix this, unreserve the blocks when doing the remount,ro, and reserved them again on remount,rw. This way a remount,ro filesystem will appear consistent on disk to all utilities. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index 9f2e398..e9c2145 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -1318,6 +1318,8 @@ xfs_fs_remount( /* ro -> rw */ if ((mp->m_flags & XFS_MOUNT_RDONLY) && !(*flags & MS_RDONLY)) { + __uint64_t resblks; + mp->m_flags &= ~XFS_MOUNT_RDONLY; if (mp->m_flags & XFS_MOUNT_BARRIER) xfs_mountfs_check_barriers(mp); @@ -1335,11 +1337,37 @@ xfs_fs_remount( } mp->m_update_flags = 0; } + + /* + * Fill out the reserve pool if it is empty. Use the stashed + * value if it is non-zero, otherwise go with the default. + */ + if (mp->m_resblks_save) { + resblks = mp->m_resblks_save; + mp->m_resblks_save = 0; + } else { + resblks = mp->m_sb.sb_dblocks; + do_div(resblks, 20); + resblks = min_t(__uint64_t, resblks, 1024); + } + xfs_reserve_blocks(mp, &resblks, NULL); } /* rw -> ro */ if (!(mp->m_flags & XFS_MOUNT_RDONLY) && (*flags & MS_RDONLY)) { + /* + * After we have synced the data but before we sync the + * metadata, we need to free up the reserve block pool so that + * the used block count in the superblock on disk is correct at + * the end of the remount. Stash the current reserve pool size + * so that if we get remounted rw, we can return it to the same + * size. + */ + __uint64_t resblks = 0; + xfs_quiesce_data(mp); + mp->m_resblks_save = mp->m_resblks; + xfs_reserve_blocks(mp, &resblks, NULL); xfs_quiesce_attr(mp); mp->m_flags |= XFS_MOUNT_RDONLY; } diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index f4d1441..02d45f2 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -225,6 +225,7 @@ typedef struct xfs_mount { __uint64_t m_maxioffset; /* maximum inode offset */ __uint64_t m_resblks; /* total reserved blocks */ __uint64_t m_resblks_avail;/* available reserved blocks */ + __uint64_t m_resblks_save; /* reserved blks @ remount,ro */ int m_dalign; /* stripe unit */ int m_swidth; /* stripe width */ int m_sinoalign; /* stripe unit inode alignment */ -- cgit v0.10.2 From 388f1f0c346b533b06d8bc792f7204ebc3e4b7da Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 26 Jan 2010 15:10:15 +1100 Subject: xfs: turn off sign warnings Because they cause warnings in static inline functions conditionally compiled into XFS from the VFS (e.g. fsnotify). Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index 1926701..5c5a366 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -16,7 +16,7 @@ # Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA # -EXTRA_CFLAGS += -I$(src) -I$(src)/linux-2.6 -Wpointer-sign +EXTRA_CFLAGS += -I$(src) -I$(src)/linux-2.6 XFS_LINUX := linux-2.6 -- cgit v0.10.2 From d5db0f97fbbeff11c88dec1aaf1536a975afbaeb Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Fri, 5 Feb 2010 22:59:53 +0000 Subject: xfs: more reserved blocks fixups This mangles the reserved blocks counts a little more. 1) add a helper function for the default reserved count 2) add helper functions to save/restore counts on ro/rw 3) save/restore reserved blocks on freeze/thaw 4) disallow changing reserved count while readonly V2: changed field name to match Dave's changes Signed-off-by: Eric Sandeen Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c index 3906e85..4ea1ee1 100644 --- a/fs/xfs/linux-2.6/xfs_ioctl.c +++ b/fs/xfs/linux-2.6/xfs_ioctl.c @@ -1431,6 +1431,9 @@ xfs_file_ioctl( if (!capable(CAP_SYS_ADMIN)) return -EPERM; + if (mp->m_flags & XFS_MOUNT_RDONLY) + return -XFS_ERROR(EROFS); + if (copy_from_user(&inout, arg, sizeof(inout))) return -XFS_ERROR(EFAULT); diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index e9c2145..6ce828e 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -1256,6 +1256,29 @@ xfs_fs_statfs( return 0; } +STATIC void +xfs_save_resvblks(struct xfs_mount *mp) +{ + __uint64_t resblks = 0; + + mp->m_resblks_save = mp->m_resblks; + xfs_reserve_blocks(mp, &resblks, NULL); +} + +STATIC void +xfs_restore_resvblks(struct xfs_mount *mp) +{ + __uint64_t resblks; + + if (mp->m_resblks_save) { + resblks = mp->m_resblks_save; + mp->m_resblks_save = 0; + } else + resblks = xfs_default_resblks(mp); + + xfs_reserve_blocks(mp, &resblks, NULL); +} + STATIC int xfs_fs_remount( struct super_block *sb, @@ -1318,8 +1341,6 @@ xfs_fs_remount( /* ro -> rw */ if ((mp->m_flags & XFS_MOUNT_RDONLY) && !(*flags & MS_RDONLY)) { - __uint64_t resblks; - mp->m_flags &= ~XFS_MOUNT_RDONLY; if (mp->m_flags & XFS_MOUNT_BARRIER) xfs_mountfs_check_barriers(mp); @@ -1342,15 +1363,7 @@ xfs_fs_remount( * Fill out the reserve pool if it is empty. Use the stashed * value if it is non-zero, otherwise go with the default. */ - if (mp->m_resblks_save) { - resblks = mp->m_resblks_save; - mp->m_resblks_save = 0; - } else { - resblks = mp->m_sb.sb_dblocks; - do_div(resblks, 20); - resblks = min_t(__uint64_t, resblks, 1024); - } - xfs_reserve_blocks(mp, &resblks, NULL); + xfs_restore_resvblks(mp); } /* rw -> ro */ @@ -1363,11 +1376,9 @@ xfs_fs_remount( * so that if we get remounted rw, we can return it to the same * size. */ - __uint64_t resblks = 0; xfs_quiesce_data(mp); - mp->m_resblks_save = mp->m_resblks; - xfs_reserve_blocks(mp, &resblks, NULL); + xfs_save_resvblks(mp); xfs_quiesce_attr(mp); mp->m_flags |= XFS_MOUNT_RDONLY; } @@ -1386,11 +1397,22 @@ xfs_fs_freeze( { struct xfs_mount *mp = XFS_M(sb); + xfs_save_resvblks(mp); xfs_quiesce_attr(mp); return -xfs_fs_log_dummy(mp); } STATIC int +xfs_fs_unfreeze( + struct super_block *sb) +{ + struct xfs_mount *mp = XFS_M(sb); + + xfs_restore_resvblks(mp); + return 0; +} + +STATIC int xfs_fs_show_options( struct seq_file *m, struct vfsmount *mnt) @@ -1612,6 +1634,7 @@ static const struct super_operations xfs_super_operations = { .put_super = xfs_fs_put_super, .sync_fs = xfs_fs_sync_fs, .freeze_fs = xfs_fs_freeze, + .unfreeze_fs = xfs_fs_unfreeze, .statfs = xfs_fs_statfs, .remount_fs = xfs_fs_remount, .show_options = xfs_fs_show_options, diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 7f81ed7..5061149 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1091,6 +1091,22 @@ xfs_mount_reset_sbqflags( return xfs_trans_commit(tp, 0); } +__uint64_t +xfs_default_resblks(xfs_mount_t *mp) +{ + __uint64_t resblks; + + /* + * We default to 5% or 1024 fsbs of space reserved, whichever is smaller. + * This may drive us straight to ENOSPC on mount, but that implies + * we were already there on the last unmount. Warn if this occurs. + */ + resblks = mp->m_sb.sb_dblocks; + do_div(resblks, 20); + resblks = min_t(__uint64_t, resblks, 1024); + return resblks; +} + /* * This function does the following on an initial mount of a file system: * - reads the superblock from disk and init the mount struct @@ -1401,18 +1417,14 @@ xfs_mountfs( * when at ENOSPC. This is needed for operations like create with * attr, unwritten extent conversion at ENOSPC, etc. Data allocations * are not allowed to use this reserved space. - * - * We default to 5% or 1024 fsbs of space reserved, whichever is smaller. - * This may drive us straight to ENOSPC on mount, but that implies - * we were already there on the last unmount. Warn if this occurs. */ - resblks = mp->m_sb.sb_dblocks; - do_div(resblks, 20); - resblks = min_t(__uint64_t, resblks, 1024); - error = xfs_reserve_blocks(mp, &resblks, NULL); - if (error) - cmn_err(CE_WARN, "XFS: Unable to allocate reserve blocks. " - "Continuing without a reserve pool."); + if (!(mp->m_flags & XFS_MOUNT_RDONLY)) { + resblks = xfs_default_resblks(mp); + error = xfs_reserve_blocks(mp, &resblks, NULL); + if (error) + cmn_err(CE_WARN, "XFS: Unable to allocate reserve " + "blocks. Continuing without a reserve pool."); + } return 0; diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 02d45f2..70504fc 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -421,6 +421,7 @@ typedef struct xfs_mod_sb { } xfs_mod_sb_t; extern int xfs_log_sbcount(xfs_mount_t *, uint); +extern __uint64_t xfs_default_resblks(xfs_mount_t *mp); extern int xfs_mountfs(xfs_mount_t *mp); extern void xfs_unmountfs(xfs_mount_t *); -- cgit v0.10.2 From 777df5afdb26c71634edd60582be620ff94e87a0 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sat, 6 Feb 2010 12:37:26 +1100 Subject: xfs: Make inode reclaim states explicit A.K.A.: don't rely on xfs_iflush() return value in reclaim We have gradually been moving checks out of the reclaim code because they are duplicated in xfs_iflush(). We've had a history of problems in this area, and many of them stem from the overloading of the return values from xfs_iflush() and interaction with inode flush locking to determine if the inode is safe to reclaim. With the desire to move to delayed write flushing of inodes and non-blocking inode tree reclaim walks, the overloading of the return value of xfs_iflush makes it very difficult to determine the correct thing to do next. This patch explicitly re-adds the checks to the inode reclaim code, removing the reliance on the return value of xfs_iflush() to determine what to do next. It also means that we can clearly document all the inode states that reclaim must handle and hence we can easily see that we handled all the necessary cases. This also removes the need for the xfs_inode_clean() check in xfs_iflush() as all callers now check this first (safely). Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c index c9b863e..525260c 100644 --- a/fs/xfs/linux-2.6/xfs_sync.c +++ b/fs/xfs/linux-2.6/xfs_sync.c @@ -706,12 +706,43 @@ __xfs_inode_clear_reclaim_tag( XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG); } +/* + * Inodes in different states need to be treated differently, and the return + * value of xfs_iflush is not sufficient to get this right. The following table + * lists the inode states and the reclaim actions necessary for non-blocking + * reclaim: + * + * + * inode state iflush ret required action + * --------------- ---------- --------------- + * bad - reclaim + * shutdown EIO unpin and reclaim + * clean, unpinned 0 reclaim + * stale, unpinned 0 reclaim + * clean, pinned(*) 0 unpin and reclaim + * stale, pinned 0 unpin and reclaim + * dirty, async 0 block on flush lock, reclaim + * dirty, sync flush 0 block on flush lock, reclaim + * + * (*) dgc: I don't think the clean, pinned state is possible but it gets + * handled anyway given the order of checks implemented. + * + * Hence the order of actions after gaining the locks should be: + * bad => reclaim + * shutdown => unpin and reclaim + * pinned => unpin + * stale => reclaim + * clean => reclaim + * dirty => flush, wait and reclaim + */ STATIC int xfs_reclaim_inode( struct xfs_inode *ip, struct xfs_perag *pag, int sync_mode) { + int error; + /* * The radix tree lock here protects a thread in xfs_iget from racing * with us starting reclaim on the inode. Once we have the @@ -729,30 +760,42 @@ xfs_reclaim_inode( spin_unlock(&ip->i_flags_lock); write_unlock(&pag->pag_ici_lock); - /* - * If the inode is still dirty, then flush it out. If the inode - * is not in the AIL, then it will be OK to flush it delwri as - * long as xfs_iflush() does not keep any references to the inode. - * We leave that decision up to xfs_iflush() since it has the - * knowledge of whether it's OK to simply do a delwri flush of - * the inode or whether we need to wait until the inode is - * pulled from the AIL. - * We get the flush lock regardless, though, just to make sure - * we don't free it while it is being flushed. - */ xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_iflock(ip); - /* - * In the case of a forced shutdown we rely on xfs_iflush() to - * wait for the inode to be unpinned before returning an error. - */ - if (!is_bad_inode(VFS_I(ip)) && xfs_iflush(ip, sync_mode) == 0) { - /* synchronize with xfs_iflush_done */ - xfs_iflock(ip); - xfs_ifunlock(ip); + if (is_bad_inode(VFS_I(ip))) + goto reclaim; + if (XFS_FORCED_SHUTDOWN(ip->i_mount)) { + xfs_iunpin_wait(ip); + goto reclaim; + } + if (xfs_ipincount(ip)) + xfs_iunpin_wait(ip); + if (xfs_iflags_test(ip, XFS_ISTALE)) + goto reclaim; + if (xfs_inode_clean(ip)) + goto reclaim; + + /* Now we have an inode that needs flushing */ + error = xfs_iflush(ip, sync_mode); + if (!error) { + switch(sync_mode) { + case XFS_IFLUSH_DELWRI_ELSE_ASYNC: + case XFS_IFLUSH_DELWRI: + case XFS_IFLUSH_ASYNC: + case XFS_IFLUSH_DELWRI_ELSE_SYNC: + case XFS_IFLUSH_SYNC: + /* IO issued, synchronise with IO completion */ + xfs_iflock(ip); + break; + default: + ASSERT(0); + break; + } } +reclaim: + xfs_ifunlock(ip); xfs_iunlock(ip, XFS_ILOCK_EXCL); xfs_ireclaim(ip); return 0; diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index d0d1b5a..8d0666d 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2493,7 +2493,7 @@ __xfs_iunpin_wait( wait_event(ip->i_ipin_wait, (atomic_read(&ip->i_pincount) == 0)); } -static inline void +void xfs_iunpin_wait( xfs_inode_t *ip) { @@ -2849,15 +2849,6 @@ xfs_iflush( mp = ip->i_mount; /* - * If the inode isn't dirty, then just release the inode flush lock and - * do nothing. - */ - if (xfs_inode_clean(ip)) { - xfs_ifunlock(ip); - return 0; - } - - /* * We can't flush the inode until it is unpinned, so wait for it if we * are allowed to block. We know noone new can pin it, because we are * holding the inode lock shared and you need to hold it exclusively to diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index ec1f28c..8b618ea 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -483,6 +483,7 @@ int xfs_iunlink(struct xfs_trans *, xfs_inode_t *); void xfs_iext_realloc(xfs_inode_t *, int, int); void xfs_ipin(xfs_inode_t *); void xfs_iunpin(xfs_inode_t *); +void xfs_iunpin_wait(xfs_inode_t *); int xfs_iflush(xfs_inode_t *, uint); void xfs_ichgtime(xfs_inode_t *, int); void xfs_lock_inodes(xfs_inode_t **, int, uint); -- cgit v0.10.2 From c854363e80b49dd04a4de18ebc379eb8c8806674 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sat, 6 Feb 2010 12:39:36 +1100 Subject: xfs: Use delayed write for inodes rather than async V2 We currently do background inode flush asynchronously, resulting in inodes being written in whatever order the background writeback issues them. Not only that, there are also blocking and non-blocking asynchronous inode flushes, depending on where the flush comes from. This patch completely removes asynchronous inode writeback. It removes all the strange writeback modes and replaces them with either a synchronous flush or a non-blocking delayed write flush. That is, inode flushes will only issue IO directly if they are synchronous, and background flushing may do nothing if the operation would block (e.g. on a pinned inode or buffer lock). Delayed write flushes will now result in the inode buffer sitting in the delwri queue of the buffer cache to be flushed by either an AIL push or by the xfsbufd timing out the buffer. This will allow accumulation of dirty inode buffers in memory and allow optimisation of inode cluster writeback at the xfsbufd level where we have much greater queue depths than the block layer elevators. We will also get adjacent inode cluster buffer IO merging for free when a later patch in the series allows sorting of the delayed write buffers before dispatch. This effectively means that any inode that is written back by background writeback will be seen as flush locked during AIL pushing, and will result in the buffers being pushed from there. This writeback path is currently non-optimal, but the next patch in the series will fix that problem. A side effect of this delayed write mechanism is that background inode reclaim will no longer directly flush inodes, nor can it wait on the flush lock. The result is that inode reclaim must leave the inode in the reclaimable state until it is clean. Hence attempts to reclaim a dirty inode in the background will simply skip the inode until it is clean and this allows other mechanisms (i.e. xfsbufd) to do more optimal writeback of the dirty buffers. As a result, the inode reclaim code has been rewritten so that it no longer relies on the ambiguous return values of xfs_iflush() to determine whether it is safe to reclaim an inode. Portions of this patch are derived from patches by Christoph Hellwig. Version 2: - cleanup reclaim code as suggested by Christoph - log background reclaim inode flush errors - just pass sync flags to xfs_iflush Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index 6ce828e..3b5b46b 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -1064,7 +1064,7 @@ xfs_fs_write_inode( xfs_ilock(ip, XFS_ILOCK_SHARED); xfs_iflock(ip); - error = xfs_iflush(ip, XFS_IFLUSH_SYNC); + error = xfs_iflush(ip, SYNC_WAIT); } else { error = EAGAIN; if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) @@ -1072,7 +1072,7 @@ xfs_fs_write_inode( if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip)) goto out_unlock; - error = xfs_iflush(ip, XFS_IFLUSH_ASYNC_NOBLOCK); + error = xfs_iflush(ip, 0); } out_unlock: diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c index 525260c..a9f6d20 100644 --- a/fs/xfs/linux-2.6/xfs_sync.c +++ b/fs/xfs/linux-2.6/xfs_sync.c @@ -270,8 +270,7 @@ xfs_sync_inode_attr( goto out_unlock; } - error = xfs_iflush(ip, (flags & SYNC_WAIT) ? - XFS_IFLUSH_SYNC : XFS_IFLUSH_DELWRI); + error = xfs_iflush(ip, flags); out_unlock: xfs_iunlock(ip, XFS_ILOCK_SHARED); @@ -460,16 +459,18 @@ xfs_quiesce_fs( { int count = 0, pincount; + xfs_reclaim_inodes(mp, 0); xfs_flush_buftarg(mp->m_ddev_targp, 0); - xfs_reclaim_inodes(mp, XFS_IFLUSH_DELWRI_ELSE_ASYNC); /* * This loop must run at least twice. The first instance of the loop * will flush most meta data but that will generate more meta data * (typically directory updates). Which then must be flushed and - * logged before we can write the unmount record. + * logged before we can write the unmount record. We also so sync + * reclaim of inodes to catch any that the above delwri flush skipped. */ do { + xfs_reclaim_inodes(mp, SYNC_WAIT); xfs_sync_attr(mp, SYNC_WAIT); pincount = xfs_flush_buftarg(mp->m_ddev_targp, 1); if (!pincount) { @@ -585,7 +586,7 @@ xfs_sync_worker( if (!(mp->m_flags & XFS_MOUNT_RDONLY)) { xfs_log_force(mp, 0); - xfs_reclaim_inodes(mp, XFS_IFLUSH_DELWRI_ELSE_ASYNC); + xfs_reclaim_inodes(mp, 0); /* dgc: errors ignored here */ error = xfs_qm_sync(mp, SYNC_TRYLOCK); error = xfs_sync_fsdata(mp, SYNC_TRYLOCK); @@ -719,21 +720,42 @@ __xfs_inode_clear_reclaim_tag( * shutdown EIO unpin and reclaim * clean, unpinned 0 reclaim * stale, unpinned 0 reclaim - * clean, pinned(*) 0 unpin and reclaim - * stale, pinned 0 unpin and reclaim - * dirty, async 0 block on flush lock, reclaim - * dirty, sync flush 0 block on flush lock, reclaim + * clean, pinned(*) 0 requeue + * stale, pinned EAGAIN requeue + * dirty, delwri ok 0 requeue + * dirty, delwri blocked EAGAIN requeue + * dirty, sync flush 0 reclaim * * (*) dgc: I don't think the clean, pinned state is possible but it gets * handled anyway given the order of checks implemented. * + * As can be seen from the table, the return value of xfs_iflush() is not + * sufficient to correctly decide the reclaim action here. The checks in + * xfs_iflush() might look like duplicates, but they are not. + * + * Also, because we get the flush lock first, we know that any inode that has + * been flushed delwri has had the flush completed by the time we check that + * the inode is clean. The clean inode check needs to be done before flushing + * the inode delwri otherwise we would loop forever requeuing clean inodes as + * we cannot tell apart a successful delwri flush and a clean inode from the + * return value of xfs_iflush(). + * + * Note that because the inode is flushed delayed write by background + * writeback, the flush lock may already be held here and waiting on it can + * result in very long latencies. Hence for sync reclaims, where we wait on the + * flush lock, the caller should push out delayed write inodes first before + * trying to reclaim them to minimise the amount of time spent waiting. For + * background relaim, we just requeue the inode for the next pass. + * * Hence the order of actions after gaining the locks should be: * bad => reclaim * shutdown => unpin and reclaim - * pinned => unpin + * pinned, delwri => requeue + * pinned, sync => unpin * stale => reclaim * clean => reclaim - * dirty => flush, wait and reclaim + * dirty, delwri => flush and requeue + * dirty, sync => flush, wait and reclaim */ STATIC int xfs_reclaim_inode( @@ -741,7 +763,7 @@ xfs_reclaim_inode( struct xfs_perag *pag, int sync_mode) { - int error; + int error = 0; /* * The radix tree lock here protects a thread in xfs_iget from racing @@ -761,7 +783,11 @@ xfs_reclaim_inode( write_unlock(&pag->pag_ici_lock); xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_iflock(ip); + if (!xfs_iflock_nowait(ip)) { + if (!(sync_mode & SYNC_WAIT)) + goto out; + xfs_iflock(ip); + } if (is_bad_inode(VFS_I(ip))) goto reclaim; @@ -769,8 +795,13 @@ xfs_reclaim_inode( xfs_iunpin_wait(ip); goto reclaim; } - if (xfs_ipincount(ip)) + if (xfs_ipincount(ip)) { + if (!(sync_mode & SYNC_WAIT)) { + xfs_ifunlock(ip); + goto out; + } xfs_iunpin_wait(ip); + } if (xfs_iflags_test(ip, XFS_ISTALE)) goto reclaim; if (xfs_inode_clean(ip)) @@ -778,27 +809,43 @@ xfs_reclaim_inode( /* Now we have an inode that needs flushing */ error = xfs_iflush(ip, sync_mode); - if (!error) { - switch(sync_mode) { - case XFS_IFLUSH_DELWRI_ELSE_ASYNC: - case XFS_IFLUSH_DELWRI: - case XFS_IFLUSH_ASYNC: - case XFS_IFLUSH_DELWRI_ELSE_SYNC: - case XFS_IFLUSH_SYNC: - /* IO issued, synchronise with IO completion */ - xfs_iflock(ip); - break; - default: - ASSERT(0); - break; - } + if (sync_mode & SYNC_WAIT) { + xfs_iflock(ip); + goto reclaim; } + /* + * When we have to flush an inode but don't have SYNC_WAIT set, we + * flush the inode out using a delwri buffer and wait for the next + * call into reclaim to find it in a clean state instead of waiting for + * it now. We also don't return errors here - if the error is transient + * then the next reclaim pass will flush the inode, and if the error + * is permanent then the next sync reclaim will relcaim the inode and + * pass on the error. + */ + if (error && !XFS_FORCED_SHUTDOWN(ip->i_mount)) { + xfs_fs_cmn_err(CE_WARN, ip->i_mount, + "inode 0x%llx background reclaim flush failed with %d", + (long long)ip->i_ino, error); + } +out: + xfs_iflags_clear(ip, XFS_IRECLAIM); + xfs_iunlock(ip, XFS_ILOCK_EXCL); + /* + * We could return EAGAIN here to make reclaim rescan the inode tree in + * a short while. However, this just burns CPU time scanning the tree + * waiting for IO to complete and xfssyncd never goes back to the idle + * state. Instead, return 0 to let the next scheduled background reclaim + * attempt to reclaim the inode again. + */ + return 0; + reclaim: xfs_ifunlock(ip); xfs_iunlock(ip, XFS_ILOCK_EXCL); xfs_ireclaim(ip); - return 0; + return error; + } int diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 8d0666d..fa31360 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2835,8 +2835,6 @@ xfs_iflush( xfs_dinode_t *dip; xfs_mount_t *mp; int error; - int noblock = (flags == XFS_IFLUSH_ASYNC_NOBLOCK); - enum { INT_DELWRI = (1 << 0), INT_ASYNC = (1 << 1) }; XFS_STATS_INC(xs_iflush_count); @@ -2859,7 +2857,7 @@ xfs_iflush( * in the same cluster are dirty, they will probably write the inode * out for us if they occur after the log force completes. */ - if (noblock && xfs_ipincount(ip)) { + if (!(flags & SYNC_WAIT) && xfs_ipincount(ip)) { xfs_iunpin_nowait(ip); xfs_ifunlock(ip); return EAGAIN; @@ -2893,60 +2891,10 @@ xfs_iflush( } /* - * Decide how buffer will be flushed out. This is done before - * the call to xfs_iflush_int because this field is zeroed by it. - */ - if (iip != NULL && iip->ili_format.ilf_fields != 0) { - /* - * Flush out the inode buffer according to the directions - * of the caller. In the cases where the caller has given - * us a choice choose the non-delwri case. This is because - * the inode is in the AIL and we need to get it out soon. - */ - switch (flags) { - case XFS_IFLUSH_SYNC: - case XFS_IFLUSH_DELWRI_ELSE_SYNC: - flags = 0; - break; - case XFS_IFLUSH_ASYNC_NOBLOCK: - case XFS_IFLUSH_ASYNC: - case XFS_IFLUSH_DELWRI_ELSE_ASYNC: - flags = INT_ASYNC; - break; - case XFS_IFLUSH_DELWRI: - flags = INT_DELWRI; - break; - default: - ASSERT(0); - flags = 0; - break; - } - } else { - switch (flags) { - case XFS_IFLUSH_DELWRI_ELSE_SYNC: - case XFS_IFLUSH_DELWRI_ELSE_ASYNC: - case XFS_IFLUSH_DELWRI: - flags = INT_DELWRI; - break; - case XFS_IFLUSH_ASYNC_NOBLOCK: - case XFS_IFLUSH_ASYNC: - flags = INT_ASYNC; - break; - case XFS_IFLUSH_SYNC: - flags = 0; - break; - default: - ASSERT(0); - flags = 0; - break; - } - } - - /* * Get the buffer containing the on-disk inode. */ error = xfs_itobp(mp, NULL, ip, &dip, &bp, - noblock ? XBF_TRYLOCK : XBF_LOCK); + (flags & SYNC_WAIT) ? XBF_LOCK : XBF_TRYLOCK); if (error || !bp) { xfs_ifunlock(ip); return error; @@ -2974,13 +2922,10 @@ xfs_iflush( if (error) goto cluster_corrupt_out; - if (flags & INT_DELWRI) { - xfs_bdwrite(mp, bp); - } else if (flags & INT_ASYNC) { - error = xfs_bawrite(mp, bp); - } else { + if (flags & SYNC_WAIT) error = xfs_bwrite(mp, bp); - } + else + xfs_bdwrite(mp, bp); return error; corrupt_out: @@ -3015,16 +2960,6 @@ xfs_iflush_int( iip = ip->i_itemp; mp = ip->i_mount; - - /* - * If the inode isn't dirty, then just release the inode - * flush lock and do nothing. - */ - if (xfs_inode_clean(ip)) { - xfs_ifunlock(ip); - return 0; - } - /* set *dip = inode's place in the buffer */ dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset); diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index 8b618ea..6c912b0 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -420,16 +420,6 @@ static inline void xfs_ifunlock(xfs_inode_t *ip) #define XFS_ILOCK_DEP(flags) (((flags) & XFS_ILOCK_DEP_MASK) >> XFS_ILOCK_SHIFT) /* - * Flags for xfs_iflush() - */ -#define XFS_IFLUSH_DELWRI_ELSE_SYNC 1 -#define XFS_IFLUSH_DELWRI_ELSE_ASYNC 2 -#define XFS_IFLUSH_SYNC 3 -#define XFS_IFLUSH_ASYNC 4 -#define XFS_IFLUSH_DELWRI 5 -#define XFS_IFLUSH_ASYNC_NOBLOCK 6 - -/* * Flags for xfs_itruncate_start(). */ #define XFS_ITRUNC_DEFINITE 0x1 diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index 48ec1c0..207553e 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -866,10 +866,14 @@ xfs_inode_item_push( iip->ili_format.ilf_fields != 0); /* - * Write out the inode. The completion routine ('iflush_done') will - * pull it from the AIL, mark it clean, unlock the flush lock. + * Push the inode to it's backing buffer. This will not remove the + * inode from the AIL - a further push will be required to trigger a + * buffer push. However, this allows all the dirty inodes to be pushed + * to the buffer before it is pushed to disk. THe buffer IO completion + * will pull th einode from the AIL, mark it clean and unlock the flush + * lock. */ - (void) xfs_iflush(ip, XFS_IFLUSH_ASYNC); + (void) xfs_iflush(ip, 0); xfs_iunlock(ip, XFS_ILOCK_SHARED); return; diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 5061149..6afaaeb 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1468,7 +1468,18 @@ xfs_unmountfs( * need to force the log first. */ xfs_log_force(mp, XFS_LOG_SYNC); - xfs_reclaim_inodes(mp, XFS_IFLUSH_ASYNC); + + /* + * Do a delwri reclaim pass first so that as many dirty inodes are + * queued up for IO as possible. Then flush the buffers before making + * a synchronous path to catch all the remaining inodes are reclaimed. + * This makes the reclaim process as quick as possible by avoiding + * synchronous writeout and blocking on inodes already in the delwri + * state as much as possible. + */ + xfs_reclaim_inodes(mp, 0); + XFS_bflush(mp->m_ddev_targp); + xfs_reclaim_inodes(mp, SYNC_WAIT); xfs_qm_unmount(mp); -- cgit v0.10.2 From d808f617ad00a413585b806de340feda5ad9a2da Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 2 Feb 2010 10:13:42 +1100 Subject: xfs: Don't issue buffer IO direct from AIL push V2 All buffers logged into the AIL are marked as delayed write. When the AIL needs to push the buffer out, it issues an async write of the buffer. This means that IO patterns are dependent on the order of buffers in the AIL. Instead of flushing the buffer, promote the buffer in the delayed write list so that the next time the xfsbufd is run the buffer will be flushed by the xfsbufd. Return the state to the xfsaild that the buffer was promoted so that the xfsaild knows that it needs to cause the xfsbufd to run to flush the buffers that were promoted. Using the xfsbufd for issuing the IO allows us to dispatch all buffer IO from the one queue. This means that we can make much more enlightened decisions on what order to flush buffers to disk as we don't have multiple places issuing IO. Optimisations to xfsbufd will be in a future patch. Version 2 - kill XFS_ITEM_FLUSHING as it is now unused. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index 44e20e5..b306265 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -1778,6 +1778,35 @@ xfs_buf_delwri_dequeue( trace_xfs_buf_delwri_dequeue(bp, _RET_IP_); } +/* + * If a delwri buffer needs to be pushed before it has aged out, then promote + * it to the head of the delwri queue so that it will be flushed on the next + * xfsbufd run. We do this by resetting the queuetime of the buffer to be older + * than the age currently needed to flush the buffer. Hence the next time the + * xfsbufd sees it is guaranteed to be considered old enough to flush. + */ +void +xfs_buf_delwri_promote( + struct xfs_buf *bp) +{ + struct xfs_buftarg *btp = bp->b_target; + long age = xfs_buf_age_centisecs * msecs_to_jiffies(10) + 1; + + ASSERT(bp->b_flags & XBF_DELWRI); + ASSERT(bp->b_flags & _XBF_DELWRI_Q); + + /* + * Check the buffer age before locking the delayed write queue as we + * don't need to promote buffers that are already past the flush age. + */ + if (bp->b_queuetime < jiffies - age) + return; + bp->b_queuetime = jiffies - age; + spin_lock(&btp->bt_delwrite_lock); + list_move(&bp->b_list, &btp->bt_delwrite_queue); + spin_unlock(&btp->bt_delwrite_lock); +} + STATIC void xfs_buf_runall_queues( struct workqueue_struct *queue) diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h index ea8c198..be45e8c 100644 --- a/fs/xfs/linux-2.6/xfs_buf.h +++ b/fs/xfs/linux-2.6/xfs_buf.h @@ -266,6 +266,7 @@ extern int xfs_buf_ispin(xfs_buf_t *); /* Delayed Write Buffer Routines */ extern void xfs_buf_delwri_dequeue(xfs_buf_t *); +extern void xfs_buf_delwri_promote(xfs_buf_t *); /* Buffer Daemon Setup Routines */ extern int xfs_buf_init(void); @@ -395,6 +396,7 @@ extern void xfs_free_buftarg(struct xfs_mount *, struct xfs_buftarg *); extern void xfs_wait_buftarg(xfs_buftarg_t *); extern int xfs_setsize_buftarg(xfs_buftarg_t *, unsigned int, unsigned int); extern int xfs_flush_buftarg(xfs_buftarg_t *, int); + #ifdef CONFIG_KDB_MODULES extern struct list_head *xfs_get_buftarg_list(void); #endif diff --git a/fs/xfs/linux-2.6/xfs_trace.h b/fs/xfs/linux-2.6/xfs_trace.h index 1bb09e7..a4574dc 100644 --- a/fs/xfs/linux-2.6/xfs_trace.h +++ b/fs/xfs/linux-2.6/xfs_trace.h @@ -483,6 +483,7 @@ DEFINE_BUF_ITEM_EVENT(xfs_buf_item_unlock); DEFINE_BUF_ITEM_EVENT(xfs_buf_item_unlock_stale); DEFINE_BUF_ITEM_EVENT(xfs_buf_item_committed); DEFINE_BUF_ITEM_EVENT(xfs_buf_item_push); +DEFINE_BUF_ITEM_EVENT(xfs_buf_item_pushbuf); DEFINE_BUF_ITEM_EVENT(xfs_trans_get_buf); DEFINE_BUF_ITEM_EVENT(xfs_trans_get_buf_recur); DEFINE_BUF_ITEM_EVENT(xfs_trans_getsb); diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c index 1b56437..dda0fb0 100644 --- a/fs/xfs/quota/xfs_dquot_item.c +++ b/fs/xfs/quota/xfs_dquot_item.c @@ -212,66 +212,31 @@ xfs_qm_dquot_logitem_pushbuf( xfs_dquot_t *dqp; xfs_mount_t *mp; xfs_buf_t *bp; - uint dopush; dqp = qip->qli_dquot; ASSERT(XFS_DQ_IS_LOCKED(dqp)); /* - * The qli_pushbuf_flag keeps others from - * trying to duplicate our effort. - */ - ASSERT(qip->qli_pushbuf_flag != 0); - ASSERT(qip->qli_push_owner == current_pid()); - - /* * If flushlock isn't locked anymore, chances are that the * inode flush completed and the inode was taken off the AIL. * So, just get out. */ if (completion_done(&dqp->q_flush) || ((qip->qli_item.li_flags & XFS_LI_IN_AIL) == 0)) { - qip->qli_pushbuf_flag = 0; xfs_dqunlock(dqp); return; } mp = dqp->q_mount; bp = xfs_incore(mp->m_ddev_targp, qip->qli_format.qlf_blkno, XFS_QI_DQCHUNKLEN(mp), XBF_TRYLOCK); - if (bp != NULL) { - if (XFS_BUF_ISDELAYWRITE(bp)) { - dopush = ((qip->qli_item.li_flags & XFS_LI_IN_AIL) && - !completion_done(&dqp->q_flush)); - qip->qli_pushbuf_flag = 0; - xfs_dqunlock(dqp); - - if (XFS_BUF_ISPINNED(bp)) - xfs_log_force(mp, 0); - - if (dopush) { - int error; -#ifdef XFSRACEDEBUG - delay_for_intr(); - delay(300); -#endif - error = xfs_bawrite(mp, bp); - if (error) - xfs_fs_cmn_err(CE_WARN, mp, - "xfs_qm_dquot_logitem_pushbuf: pushbuf error %d on qip %p, bp %p", - error, qip, bp); - } else { - xfs_buf_relse(bp); - } - } else { - qip->qli_pushbuf_flag = 0; - xfs_dqunlock(dqp); - xfs_buf_relse(bp); - } + xfs_dqunlock(dqp); + if (!bp) return; - } + if (XFS_BUF_ISDELAYWRITE(bp)) + xfs_buf_delwri_promote(bp); + xfs_buf_relse(bp); + return; - qip->qli_pushbuf_flag = 0; - xfs_dqunlock(dqp); } /* @@ -289,50 +254,24 @@ xfs_qm_dquot_logitem_trylock( xfs_dq_logitem_t *qip) { xfs_dquot_t *dqp; - uint retval; dqp = qip->qli_dquot; if (atomic_read(&dqp->q_pincount) > 0) - return (XFS_ITEM_PINNED); + return XFS_ITEM_PINNED; if (! xfs_qm_dqlock_nowait(dqp)) - return (XFS_ITEM_LOCKED); + return XFS_ITEM_LOCKED; - retval = XFS_ITEM_SUCCESS; if (!xfs_dqflock_nowait(dqp)) { /* - * The dquot is already being flushed. It may have been - * flushed delayed write, however, and we don't want to - * get stuck waiting for that to complete. So, we want to check - * to see if we can lock the dquot's buffer without sleeping. - * If we can and it is marked for delayed write, then we - * hold it and send it out from the push routine. We don't - * want to do that now since we might sleep in the device - * strategy routine. We also don't want to grab the buffer lock - * here because we'd like not to call into the buffer cache - * while holding the AIL lock. - * Make sure to only return PUSHBUF if we set pushbuf_flag - * ourselves. If someone else is doing it then we don't - * want to go to the push routine and duplicate their efforts. + * dquot has already been flushed to the backing buffer, + * leave it locked, pushbuf routine will unlock it. */ - if (qip->qli_pushbuf_flag == 0) { - qip->qli_pushbuf_flag = 1; - ASSERT(qip->qli_format.qlf_blkno == dqp->q_blkno); -#ifdef DEBUG - qip->qli_push_owner = current_pid(); -#endif - /* - * The dquot is left locked. - */ - retval = XFS_ITEM_PUSHBUF; - } else { - retval = XFS_ITEM_FLUSHING; - xfs_dqunlock_nonotify(dqp); - } + return XFS_ITEM_PUSHBUF; } ASSERT(qip->qli_item.li_flags & XFS_LI_IN_AIL); - return (retval); + return XFS_ITEM_SUCCESS; } diff --git a/fs/xfs/quota/xfs_dquot_item.h b/fs/xfs/quota/xfs_dquot_item.h index 5a63253..5acae2a 100644 --- a/fs/xfs/quota/xfs_dquot_item.h +++ b/fs/xfs/quota/xfs_dquot_item.h @@ -27,10 +27,6 @@ typedef struct xfs_dq_logitem { xfs_log_item_t qli_item; /* common portion */ struct xfs_dquot *qli_dquot; /* dquot ptr */ xfs_lsn_t qli_flush_lsn; /* lsn at last flush */ - unsigned short qli_pushbuf_flag; /* 1 bit used in push_ail */ -#ifdef DEBUG - uint64_t qli_push_owner; -#endif xfs_dq_logformat_t qli_format; /* logged structure */ } xfs_dq_logitem_t; diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index e0a1158..f3c49e6 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -467,8 +467,10 @@ xfs_buf_item_unpin_remove( /* * This is called to attempt to lock the buffer associated with this * buf log item. Don't sleep on the buffer lock. If we can't get - * the lock right away, return 0. If we can get the lock, pull the - * buffer from the free list, mark it busy, and return 1. + * the lock right away, return 0. If we can get the lock, take a + * reference to the buffer. If this is a delayed write buffer that + * needs AIL help to be written back, invoke the pushbuf routine + * rather than the normal success path. */ STATIC uint xfs_buf_item_trylock( @@ -477,24 +479,18 @@ xfs_buf_item_trylock( xfs_buf_t *bp; bp = bip->bli_buf; - - if (XFS_BUF_ISPINNED(bp)) { + if (XFS_BUF_ISPINNED(bp)) return XFS_ITEM_PINNED; - } - - if (!XFS_BUF_CPSEMA(bp)) { + if (!XFS_BUF_CPSEMA(bp)) return XFS_ITEM_LOCKED; - } - /* - * Remove the buffer from the free list. Only do this - * if it's on the free list. Private buffers like the - * superblock buffer are not. - */ + /* take a reference to the buffer. */ XFS_BUF_HOLD(bp); ASSERT(!(bip->bli_flags & XFS_BLI_STALE)); trace_xfs_buf_item_trylock(bip); + if (XFS_BUF_ISDELAYWRITE(bp)) + return XFS_ITEM_PUSHBUF; return XFS_ITEM_SUCCESS; } @@ -626,11 +622,9 @@ xfs_buf_item_committed( } /* - * This is called to asynchronously write the buffer associated with this - * buf log item out to disk. The buffer will already have been locked by - * a successful call to xfs_buf_item_trylock(). If the buffer still has - * B_DELWRI set, then get it going out to disk with a call to bawrite(). - * If not, then just release the buffer. + * The buffer is locked, but is not a delayed write buffer. This happens + * if we race with IO completion and hence we don't want to try to write it + * again. Just release the buffer. */ STATIC void xfs_buf_item_push( @@ -642,17 +636,29 @@ xfs_buf_item_push( trace_xfs_buf_item_push(bip); bp = bip->bli_buf; + ASSERT(!XFS_BUF_ISDELAYWRITE(bp)); + xfs_buf_relse(bp); +} - if (XFS_BUF_ISDELAYWRITE(bp)) { - int error; - error = xfs_bawrite(bip->bli_item.li_mountp, bp); - if (error) - xfs_fs_cmn_err(CE_WARN, bip->bli_item.li_mountp, - "xfs_buf_item_push: pushbuf error %d on bip %p, bp %p", - error, bip, bp); - } else { - xfs_buf_relse(bp); - } +/* + * The buffer is locked and is a delayed write buffer. Promote the buffer + * in the delayed write queue as the caller knows that they must invoke + * the xfsbufd to get this buffer written. We have to unlock the buffer + * to allow the xfsbufd to write it, too. + */ +STATIC void +xfs_buf_item_pushbuf( + xfs_buf_log_item_t *bip) +{ + xfs_buf_t *bp; + + ASSERT(!(bip->bli_flags & XFS_BLI_STALE)); + trace_xfs_buf_item_pushbuf(bip); + + bp = bip->bli_buf; + ASSERT(XFS_BUF_ISDELAYWRITE(bp)); + xfs_buf_delwri_promote(bp); + xfs_buf_relse(bp); } /* ARGSUSED */ @@ -677,7 +683,7 @@ static struct xfs_item_ops xfs_buf_item_ops = { .iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t)) xfs_buf_item_committed, .iop_push = (void(*)(xfs_log_item_t*))xfs_buf_item_push, - .iop_pushbuf = NULL, + .iop_pushbuf = (void(*)(xfs_log_item_t*))xfs_buf_item_pushbuf, .iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t)) xfs_buf_item_committing }; diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index 207553e..d4dc063 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -602,33 +602,20 @@ xfs_inode_item_trylock( if (!xfs_iflock_nowait(ip)) { /* - * If someone else isn't already trying to push the inode - * buffer, we get to do it. + * inode has already been flushed to the backing buffer, + * leave it locked in shared mode, pushbuf routine will + * unlock it. */ - if (iip->ili_pushbuf_flag == 0) { - iip->ili_pushbuf_flag = 1; -#ifdef DEBUG - iip->ili_push_owner = current_pid(); -#endif - /* - * Inode is left locked in shared mode. - * Pushbuf routine gets to unlock it. - */ - return XFS_ITEM_PUSHBUF; - } else { - /* - * We hold the AIL lock, so we must specify the - * NONOTIFY flag so that we won't double trip. - */ - xfs_iunlock(ip, XFS_ILOCK_SHARED|XFS_IUNLOCK_NONOTIFY); - return XFS_ITEM_FLUSHING; - } - /* NOTREACHED */ + return XFS_ITEM_PUSHBUF; } /* Stale items should force out the iclog */ if (ip->i_flags & XFS_ISTALE) { xfs_ifunlock(ip); + /* + * we hold the AIL lock - notify the unlock routine of this + * so it doesn't try to get the lock again. + */ xfs_iunlock(ip, XFS_ILOCK_SHARED|XFS_IUNLOCK_NONOTIFY); return XFS_ITEM_PINNED; } @@ -746,11 +733,8 @@ xfs_inode_item_committed( * This gets called by xfs_trans_push_ail(), when IOP_TRYLOCK * failed to get the inode flush lock but did get the inode locked SHARED. * Here we're trying to see if the inode buffer is incore, and if so whether it's - * marked delayed write. If that's the case, we'll initiate a bawrite on that - * buffer to expedite the process. - * - * We aren't holding the AIL lock (or the flush lock) when this gets called, - * so it is inherently race-y. + * marked delayed write. If that's the case, we'll promote it and that will + * allow the caller to write the buffer by triggering the xfsbufd to run. */ STATIC void xfs_inode_item_pushbuf( @@ -759,26 +743,16 @@ xfs_inode_item_pushbuf( xfs_inode_t *ip; xfs_mount_t *mp; xfs_buf_t *bp; - uint dopush; ip = iip->ili_inode; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED)); /* - * The ili_pushbuf_flag keeps others from - * trying to duplicate our effort. - */ - ASSERT(iip->ili_pushbuf_flag != 0); - ASSERT(iip->ili_push_owner == current_pid()); - - /* * If a flush is not in progress anymore, chances are that the * inode was taken off the AIL. So, just get out. */ if (completion_done(&ip->i_flush) || ((iip->ili_item.li_flags & XFS_LI_IN_AIL) == 0)) { - iip->ili_pushbuf_flag = 0; xfs_iunlock(ip, XFS_ILOCK_SHARED); return; } @@ -787,53 +761,12 @@ xfs_inode_item_pushbuf( bp = xfs_incore(mp->m_ddev_targp, iip->ili_format.ilf_blkno, iip->ili_format.ilf_len, XBF_TRYLOCK); - if (bp != NULL) { - if (XFS_BUF_ISDELAYWRITE(bp)) { - /* - * We were racing with iflush because we don't hold - * the AIL lock or the flush lock. However, at this point, - * we have the buffer, and we know that it's dirty. - * So, it's possible that iflush raced with us, and - * this item is already taken off the AIL. - * If not, we can flush it async. - */ - dopush = ((iip->ili_item.li_flags & XFS_LI_IN_AIL) && - !completion_done(&ip->i_flush)); - iip->ili_pushbuf_flag = 0; - xfs_iunlock(ip, XFS_ILOCK_SHARED); - - trace_xfs_inode_item_push(bp, _RET_IP_); - - if (XFS_BUF_ISPINNED(bp)) - xfs_log_force(mp, 0); - - if (dopush) { - int error; - error = xfs_bawrite(mp, bp); - if (error) - xfs_fs_cmn_err(CE_WARN, mp, - "xfs_inode_item_pushbuf: pushbuf error %d on iip %p, bp %p", - error, iip, bp); - } else { - xfs_buf_relse(bp); - } - } else { - iip->ili_pushbuf_flag = 0; - xfs_iunlock(ip, XFS_ILOCK_SHARED); - xfs_buf_relse(bp); - } - return; - } - /* - * We have to be careful about resetting pushbuf flag too early (above). - * Even though in theory we can do it as soon as we have the buflock, - * we don't want others to be doing work needlessly. They'll come to - * this function thinking that pushing the buffer is their - * responsibility only to find that the buffer is still locked by - * another doing the same thing - */ - iip->ili_pushbuf_flag = 0; xfs_iunlock(ip, XFS_ILOCK_SHARED); + if (!bp) + return; + if (XFS_BUF_ISDELAYWRITE(bp)) + xfs_buf_delwri_promote(bp); + xfs_buf_relse(bp); return; } @@ -937,7 +870,6 @@ xfs_inode_item_init( /* We have zeroed memory. No need ... iip->ili_extents_buf = NULL; - iip->ili_pushbuf_flag = 0; */ iip->ili_format.ilf_type = XFS_LI_INODE; diff --git a/fs/xfs/xfs_inode_item.h b/fs/xfs/xfs_inode_item.h index cc8df1a..9a46795 100644 --- a/fs/xfs/xfs_inode_item.h +++ b/fs/xfs/xfs_inode_item.h @@ -144,12 +144,6 @@ typedef struct xfs_inode_log_item { data exts */ struct xfs_bmbt_rec *ili_aextents_buf; /* array of logged attr exts */ - unsigned int ili_pushbuf_flag; /* one bit used in push_ail */ - -#ifdef DEBUG - uint64_t ili_push_owner; /* one who sets pushbuf_flag - above gets to push the buf */ -#endif #ifdef XFS_TRANS_DEBUG int ili_root_size; char *ili_orig_root; diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index ca64f33..c93e3a1 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -861,8 +861,7 @@ typedef struct xfs_item_ops { #define XFS_ITEM_SUCCESS 0 #define XFS_ITEM_PINNED 1 #define XFS_ITEM_LOCKED 2 -#define XFS_ITEM_FLUSHING 3 -#define XFS_ITEM_PUSHBUF 4 +#define XFS_ITEM_PUSHBUF 3 /* * This structure is used to maintain a list of block ranges that have been diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c index d7b1af8..e799824 100644 --- a/fs/xfs/xfs_trans_ail.c +++ b/fs/xfs/xfs_trans_ail.c @@ -253,6 +253,7 @@ xfsaild_push( int flush_log, count, stuck; xfs_mount_t *mp = ailp->xa_mount; struct xfs_ail_cursor *cur = &ailp->xa_cursors; + int push_xfsbufd = 0; spin_lock(&ailp->xa_lock); xfs_trans_ail_cursor_init(ailp, cur); @@ -308,6 +309,7 @@ xfsaild_push( XFS_STATS_INC(xs_push_ail_pushbuf); IOP_PUSHBUF(lip); last_pushed_lsn = lsn; + push_xfsbufd = 1; break; case XFS_ITEM_PINNED: @@ -322,12 +324,6 @@ xfsaild_push( stuck++; break; - case XFS_ITEM_FLUSHING: - XFS_STATS_INC(xs_push_ail_flushing); - last_pushed_lsn = lsn; - stuck++; - break; - default: ASSERT(0); break; @@ -374,6 +370,11 @@ xfsaild_push( xfs_log_force(mp, 0); } + if (push_xfsbufd) { + /* we've got delayed write buffers to flush */ + wake_up_process(mp->m_ddev_targp->bt_task); + } + if (!count) { /* We're past our target or empty, so idle */ last_pushed_lsn = 0; -- cgit v0.10.2 From 089716aa1480b7197bcd678b8477774c379a2768 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 26 Jan 2010 15:13:25 +1100 Subject: xfs: Sort delayed write buffers before dispatch MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Currently when the xfsbufd writes delayed write buffers, it pushes them to disk in the order they come off the delayed write list. If there are lots of buffers Ñ•pread widely over the disk, this results in overwhelming the elevator sort queues in the block layer and we end up losing the posibility of merging adjacent buffers to minimise the number of IOs. Use the new generic list_sort function to sort the delwri dispatch queue before issue to ensure that the buffers are pushed in the most friendly order possible to the lower layers. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index b306265..4556a4c 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -33,6 +33,7 @@ #include #include #include +#include #include "xfs_sb.h" #include "xfs_inum.h" @@ -1877,14 +1878,42 @@ xfs_buf_delwri_split( } +/* + * Compare function is more complex than it needs to be because + * the return value is only 32 bits and we are doing comparisons + * on 64 bit values + */ +static int +xfs_buf_cmp( + void *priv, + struct list_head *a, + struct list_head *b) +{ + struct xfs_buf *ap = container_of(a, struct xfs_buf, b_list); + struct xfs_buf *bp = container_of(b, struct xfs_buf, b_list); + xfs_daddr_t diff; + + diff = ap->b_bn - bp->b_bn; + if (diff < 0) + return -1; + if (diff > 0) + return 1; + return 0; +} + +void +xfs_buf_delwri_sort( + xfs_buftarg_t *target, + struct list_head *list) +{ + list_sort(NULL, list, xfs_buf_cmp); +} + STATIC int xfsbufd( void *data) { - struct list_head tmp; - xfs_buftarg_t *target = (xfs_buftarg_t *)data; - int count; - xfs_buf_t *bp; + xfs_buftarg_t *target = (xfs_buftarg_t *)data; current->flags |= PF_MEMALLOC; @@ -1893,6 +1922,8 @@ xfsbufd( do { long age = xfs_buf_age_centisecs * msecs_to_jiffies(10); long tout = xfs_buf_timer_centisecs * msecs_to_jiffies(10); + int count = 0; + struct list_head tmp; if (unlikely(freezing(current))) { set_bit(XBT_FORCE_SLEEP, &target->bt_flags); @@ -1907,11 +1938,10 @@ xfsbufd( schedule_timeout_interruptible(tout); xfs_buf_delwri_split(target, &tmp, age); - count = 0; + list_sort(NULL, &tmp, xfs_buf_cmp); while (!list_empty(&tmp)) { - bp = list_entry(tmp.next, xfs_buf_t, b_list); - ASSERT(target == bp->b_target); - + struct xfs_buf *bp; + bp = list_first_entry(&tmp, struct xfs_buf, b_list); list_del_init(&bp->b_list); xfs_buf_iostrategy(bp); count++; @@ -1937,42 +1967,45 @@ xfs_flush_buftarg( xfs_buftarg_t *target, int wait) { - struct list_head tmp; - xfs_buf_t *bp, *n; + xfs_buf_t *bp; int pincount = 0; + LIST_HEAD(tmp_list); + LIST_HEAD(wait_list); xfs_buf_runall_queues(xfsconvertd_workqueue); xfs_buf_runall_queues(xfsdatad_workqueue); xfs_buf_runall_queues(xfslogd_workqueue); set_bit(XBT_FORCE_FLUSH, &target->bt_flags); - pincount = xfs_buf_delwri_split(target, &tmp, 0); + pincount = xfs_buf_delwri_split(target, &tmp_list, 0); /* - * Dropped the delayed write list lock, now walk the temporary list + * Dropped the delayed write list lock, now walk the temporary list. + * All I/O is issued async and then if we need to wait for completion + * we do that after issuing all the IO. */ - list_for_each_entry_safe(bp, n, &tmp, b_list) { + list_sort(NULL, &tmp_list, xfs_buf_cmp); + while (!list_empty(&tmp_list)) { + bp = list_first_entry(&tmp_list, struct xfs_buf, b_list); ASSERT(target == bp->b_target); - if (wait) + list_del_init(&bp->b_list); + if (wait) { bp->b_flags &= ~XBF_ASYNC; - else - list_del_init(&bp->b_list); - + list_add(&bp->b_list, &wait_list); + } xfs_buf_iostrategy(bp); } - if (wait) + if (wait) { + /* Expedite and wait for IO to complete. */ blk_run_address_space(target->bt_mapping); + while (!list_empty(&wait_list)) { + bp = list_first_entry(&wait_list, struct xfs_buf, b_list); - /* - * Remaining list items must be flushed before returning - */ - while (!list_empty(&tmp)) { - bp = list_entry(tmp.next, xfs_buf_t, b_list); - - list_del_init(&bp->b_list); - xfs_iowait(bp); - xfs_buf_relse(bp); + list_del_init(&bp->b_list); + xfs_iowait(bp); + xfs_buf_relse(bp); + } } return pincount; -- cgit v0.10.2 From 7d6a7bde52e449f21a0e86a7a4955b4e08a49d69 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Tue, 26 Jan 2010 15:13:41 +1100 Subject: xfs: Use delay write promotion for dquot flushing xfs_qm_dqflock_pushbuf_wait() does a very similar trick to item pushing used to do to flush out delayed write dquot buffers. Change it to use the new promotion method rather than an async flush. Also, xfs_qm_dqflock_pushbuf_wait() can return without the flush lock held, yet the callers make the assumption that after this call the flush lock is held. Always return with the flush lock held. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/quota/xfs_dquot.c b/fs/xfs/quota/xfs_dquot.c index f9baeed..1620a56 100644 --- a/fs/xfs/quota/xfs_dquot.c +++ b/fs/xfs/quota/xfs_dquot.c @@ -1528,21 +1528,16 @@ xfs_qm_dqflock_pushbuf_wait( */ bp = xfs_incore(dqp->q_mount->m_ddev_targp, dqp->q_blkno, XFS_QI_DQCHUNKLEN(dqp->q_mount), XBF_TRYLOCK); - if (bp != NULL) { - if (XFS_BUF_ISDELAYWRITE(bp)) { - int error; - - if (XFS_BUF_ISPINNED(bp)) - xfs_log_force(dqp->q_mount, 0); - error = xfs_bawrite(dqp->q_mount, bp); - if (error) - xfs_fs_cmn_err(CE_WARN, dqp->q_mount, - "xfs_qm_dqflock_pushbuf_wait: " - "pushbuf error %d on dqp %p, bp %p", - error, dqp, bp); - } else { - xfs_buf_relse(bp); - } + if (!bp) + goto out_lock; + + if (XFS_BUF_ISDELAYWRITE(bp)) { + if (XFS_BUF_ISPINNED(bp)) + xfs_log_force(dqp->q_mount, 0); + xfs_buf_delwri_promote(bp); + wake_up_process(bp->b_target->bt_task); } + xfs_buf_relse(bp); +out_lock: xfs_dqflock(dqp); } -- cgit v0.10.2 From abd50713944c8ea9e0af5b7bffa0aacae21cc91a Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 26 Jan 2010 18:50:16 +0100 Subject: perf: Reimplement frequency driven sampling There was a bug in the old period code that caused intel_pmu_enable_all() or native_write_msr_safe() to show up quite high in the profiles. In staring at that code it made my head hurt, so I rewrote it in a hopefully simpler fashion. Its now fully symetric between tick and overflow driven adjustments and uses less data to boot. The only complication is that it basically wants to do a u128 division. The code approximates that in a rather simple truncate until it fits fashion, taking care to balance the terms while truncating. This version does not generate that sampling artefact. Signed-off-by: Peter Zijlstra LKML-Reference: Cc: Signed-off-by: Ingo Molnar diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index c6f812e..72b2615 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -498,9 +498,8 @@ struct hw_perf_event { atomic64_t period_left; u64 interrupts; - u64 freq_count; - u64 freq_interrupts; - u64 freq_stamp; + u64 freq_time_stamp; + u64 freq_count_stamp; #endif }; diff --git a/kernel/perf_event.c b/kernel/perf_event.c index edc46b9..251fb95 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1423,14 +1423,83 @@ void perf_event_task_sched_in(struct task_struct *task) static void perf_log_throttle(struct perf_event *event, int enable); -static void perf_adjust_period(struct perf_event *event, u64 events) +static u64 perf_calculate_period(struct perf_event *event, u64 nsec, u64 count) +{ + u64 frequency = event->attr.sample_freq; + u64 sec = NSEC_PER_SEC; + u64 divisor, dividend; + + int count_fls, nsec_fls, frequency_fls, sec_fls; + + count_fls = fls64(count); + nsec_fls = fls64(nsec); + frequency_fls = fls64(frequency); + sec_fls = 30; + + /* + * We got @count in @nsec, with a target of sample_freq HZ + * the target period becomes: + * + * @count * 10^9 + * period = ------------------- + * @nsec * sample_freq + * + */ + + /* + * Reduce accuracy by one bit such that @a and @b converge + * to a similar magnitude. + */ +#define REDUCE_FLS(a, b) \ +do { \ + if (a##_fls > b##_fls) { \ + a >>= 1; \ + a##_fls--; \ + } else { \ + b >>= 1; \ + b##_fls--; \ + } \ +} while (0) + + /* + * Reduce accuracy until either term fits in a u64, then proceed with + * the other, so that finally we can do a u64/u64 division. + */ + while (count_fls + sec_fls > 64 && nsec_fls + frequency_fls > 64) { + REDUCE_FLS(nsec, frequency); + REDUCE_FLS(sec, count); + } + + if (count_fls + sec_fls > 64) { + divisor = nsec * frequency; + + while (count_fls + sec_fls > 64) { + REDUCE_FLS(count, sec); + divisor >>= 1; + } + + dividend = count * sec; + } else { + dividend = count * sec; + + while (nsec_fls + frequency_fls > 64) { + REDUCE_FLS(nsec, frequency); + dividend >>= 1; + } + + divisor = nsec * frequency; + } + + return div64_u64(dividend, divisor); +} + +static void perf_adjust_period(struct perf_event *event, u64 nsec, u64 count) { struct hw_perf_event *hwc = &event->hw; u64 period, sample_period; s64 delta; - events *= hwc->sample_period; - period = div64_u64(events, event->attr.sample_freq); + period = perf_calculate_period(event, nsec, count); delta = (s64)(period - hwc->sample_period); delta = (delta + 7) / 8; /* low pass filter */ @@ -1441,13 +1510,22 @@ static void perf_adjust_period(struct perf_event *event, u64 events) sample_period = 1; hwc->sample_period = sample_period; + + if (atomic64_read(&hwc->period_left) > 8*sample_period) { + perf_disable(); + event->pmu->disable(event); + atomic64_set(&hwc->period_left, 0); + event->pmu->enable(event); + perf_enable(); + } } static void perf_ctx_adjust_freq(struct perf_event_context *ctx) { struct perf_event *event; struct hw_perf_event *hwc; - u64 interrupts, freq; + u64 interrupts, now; + s64 delta; raw_spin_lock(&ctx->lock); list_for_each_entry_rcu(event, &ctx->event_list, event_entry) { @@ -1468,44 +1546,18 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx) if (interrupts == MAX_INTERRUPTS) { perf_log_throttle(event, 1); event->pmu->unthrottle(event); - interrupts = 2*sysctl_perf_event_sample_rate/HZ; } if (!event->attr.freq || !event->attr.sample_freq) continue; - /* - * if the specified freq < HZ then we need to skip ticks - */ - if (event->attr.sample_freq < HZ) { - freq = event->attr.sample_freq; - - hwc->freq_count += freq; - hwc->freq_interrupts += interrupts; - - if (hwc->freq_count < HZ) - continue; - - interrupts = hwc->freq_interrupts; - hwc->freq_interrupts = 0; - hwc->freq_count -= HZ; - } else - freq = HZ; - - perf_adjust_period(event, freq * interrupts); + event->pmu->read(event); + now = atomic64_read(&event->count); + delta = now - hwc->freq_count_stamp; + hwc->freq_count_stamp = now; - /* - * In order to avoid being stalled by an (accidental) huge - * sample period, force reset the sample period if we didn't - * get any events in this freq period. - */ - if (!interrupts) { - perf_disable(); - event->pmu->disable(event); - atomic64_set(&hwc->period_left, 0); - event->pmu->enable(event); - perf_enable(); - } + if (delta > 0) + perf_adjust_period(event, TICK_NSEC, delta); } raw_spin_unlock(&ctx->lock); } @@ -3768,12 +3820,12 @@ static int __perf_event_overflow(struct perf_event *event, int nmi, if (event->attr.freq) { u64 now = perf_clock(); - s64 delta = now - hwc->freq_stamp; + s64 delta = now - hwc->freq_time_stamp; - hwc->freq_stamp = now; + hwc->freq_time_stamp = now; - if (delta > 0 && delta < TICK_NSEC) - perf_adjust_period(event, NSEC_PER_SEC / (int)delta); + if (delta > 0 && delta < 2*TICK_NSEC) + perf_adjust_period(event, delta, hwc->last_period); } /* -- cgit v0.10.2 From 24bfef0f924b4ac4312614422a4982b5f4d9a4c7 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 21 Jan 2010 13:04:43 -0200 Subject: perf top: Fix sample counting MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Broken since "5b2bb75 perf top: Support userspace symbols too". Reported-by: Mike Galbraith Tested-by: Mike Galbraith Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1264086284-1431-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 7a8a77e..8b04988 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -934,8 +934,11 @@ static void event__process_sample(const event_t *self, struct addr_location al; u8 origin = self->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; + ++samples; + switch (origin) { case PERF_RECORD_MISC_USER: + ++userspace_samples; if (hide_user_symbols) return; break; @@ -960,9 +963,6 @@ static void event__process_sample(const event_t *self, if (list_empty(&syme->node) || !syme->node.next) __list_insert_active_sym(syme); pthread_mutex_unlock(&active_symbols_lock); - if (origin == PERF_RECORD_MISC_USER) - ++userspace_samples; - ++samples; } } -- cgit v0.10.2 From 0f35cd4cea08a8893e3e2ea03cbdb65f5d2b0e7a Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 21 Jan 2010 13:04:44 -0200 Subject: perf top: Handle PERF_RECORD_{FORK,EXIT} events MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As noticed by Mike, symbols in new tasks were not being processed as we weren't processing these events. Reported-by: Mike Galbraith Tested-by: Mike Galbraith Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1264086284-1431-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 8b04988..2227b84 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -975,6 +975,10 @@ static int event__process(event_t *event, struct perf_session *session) case PERF_RECORD_MMAP: event__process_mmap(event, session); break; + case PERF_RECORD_FORK: + case PERF_RECORD_EXIT: + event__process_task(event, session); + break; default: break; } -- cgit v0.10.2 From e1c7c6a40c8037478742ce134190c1a955853bfb Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 22 Jan 2010 14:35:01 -0200 Subject: perf symbols: Fix inverted logic for showing kallsyms as the source of symbols MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Only if we parsed /proc/kallsyms (or a copy found in the buildid cache) we should set the dso long name to "[kernel.kallsyms]". Reported-by: Mike Galbraith Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1264178102-4203-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 6f30fe1..1270cf8 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1671,7 +1671,7 @@ do_kallsyms: out_try_fixup: if (err > 0) { out_fixup: - if (kallsyms_filename == NULL) + if (kallsyms_filename != NULL) dso__set_long_name(self, strdup("[kernel.kallsyms]")); map__fixup_start(map); map__fixup_end(map); -- cgit v0.10.2 From 19fc2dedff448120a7aeaa3c136689c6b71777c6 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 22 Jan 2010 14:35:02 -0200 Subject: perf symbols: Use the right variable to check for kallsyms in the cache MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Probably this wasn't noticed when testing this on my parisc machine because I must have copied manually to its cache the vmlinux file used in the x86_64 machine, now that I tried looking on a x86-32 machine with a fresh cache, kernel symbols weren't being resolved even with the right kallsyms copy on its cache, duh. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1264178102-4203-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 1270cf8..f1f609d 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1650,12 +1650,12 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, getenv("HOME"), sbuild_id) == -1) return -1; + kallsyms_filename = kallsyms_allocated_filename; + if (access(kallsyms_filename, F_OK)) { free(kallsyms_allocated_filename); return -1; } - - kallsyms_filename = kallsyms_allocated_filename; } else { /* * Last resort, if we don't have a build-id and couldn't find -- cgit v0.10.2 From 408f0d18ba6b9bb447f807f621b2c9663c5cf638 Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Fri, 22 Jan 2010 22:45:29 +0900 Subject: perf trace: Add -i option for choosing input file perf trace lacks -i option for choosing input file. This patch adds it to perf trace. Signed-off-by: Hitoshi Mitake Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker LKML-Reference: <1264167929-6741-1-git-send-email-mitake@dcl.info.waseda.ac.jp> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 8e9cbfe..0b65779 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -515,6 +515,8 @@ static const struct option options[] = { parse_scriptname), OPT_STRING('g', "gen-script", &generate_script_lang, "lang", "generate perf-trace.xx script in specified language"), + OPT_STRING('i', "input", &input_name, "file", + "input file name"), OPT_END() }; -- cgit v0.10.2 From 08b36c4a02b5a9db609fc6d93b9c41d7fa75713c Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 27 Jan 2010 21:56:57 +0900 Subject: sh: Optimize runtime disabling of trapped I/O. Presently trapped I/O is only registered if it's not explicitly disabled for the platforms that select it openly. From the fault path this runs through an address lookup before figuring out that nothing matches and falls back through the error path, but we can forego the lookup completely by testing if it's been explicitly disabled. This provides a measurable speedup for things like qemu that rely on runtime disabling. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/io_trapped.c b/arch/sh/kernel/io_trapped.c index 16b83cc..4a8bb4e 100644 --- a/arch/sh/kernel/io_trapped.c +++ b/arch/sh/kernel/io_trapped.c @@ -271,6 +271,8 @@ int handle_trapped_io(struct pt_regs *regs, unsigned long address) insn_size_t instruction; int tmp; + if (trapped_io_disable) + return 0; if (!lookup_tiop(address)) return 0; -- cgit v0.10.2 From 660e2acad81c19b404f7d7d06e57a6d5e6ce7426 Mon Sep 17 00:00:00 2001 From: Chris Smith Date: Wed, 27 Jan 2010 22:03:11 +0900 Subject: sh: kmemleak support. Enables support for kmemleak on sh. Signed-off-by: Chris Smith Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/vmlinux.lds.S b/arch/sh/kernel/vmlinux.lds.S index 93e9b25..f0bc6b8 100644 --- a/arch/sh/kernel/vmlinux.lds.S +++ b/arch/sh/kernel/vmlinux.lds.S @@ -50,12 +50,12 @@ SECTIONS } = 0x0009 EXCEPTION_TABLE(16) - NOTES + + _sdata = .; RO_DATA(PAGE_SIZE) RW_DATA_SECTION(L1_CACHE_BYTES, PAGE_SIZE, THREAD_SIZE) - - _edata = .; /* End of data section */ + _edata = .; DWARF_EH_FRAME diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 25c3ed5..d62e3cd 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -355,7 +355,7 @@ config SLUB_STATS config DEBUG_KMEMLEAK bool "Kernel memory leak detector" depends on DEBUG_KERNEL && EXPERIMENTAL && !MEMORY_HOTPLUG && \ - (X86 || ARM || PPC || S390) + (X86 || ARM || PPC || S390 || SUPERH) select DEBUG_FS if SYSFS select STACKTRACE if STACKTRACE_SUPPORT -- cgit v0.10.2 From 7013109f52ff38ad9ea38952cedc144c8e8b9636 Mon Sep 17 00:00:00 2001 From: Giuseppe CAVALLARO Date: Mon, 25 Jan 2010 16:12:07 +0100 Subject: sh: fixed cmpxchg gRB version This patch fixes a bug within the cmpxchg GRB version. A problem was notices while running some tests to stress the priority inheritance, for example pi_stress (http://rt.wiki.kernel.org/index.php/PI_Mutex_Test). Also, without this patch, after applying the latest work to consolidate atomic_cmpxchg() definitions (commit: 8c0b8139c87cfe8b95c6e763b4ca3190aa9b1ad0) the Kernel doesn't boot at all. Signed-off-by: Giuseppe Cavallaro Signed-off-by: Stuart Menefy Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/cmpxchg-grb.h b/arch/sh/include/asm/cmpxchg-grb.h index e2681ab..4676bf5 100644 --- a/arch/sh/include/asm/cmpxchg-grb.h +++ b/arch/sh/include/asm/cmpxchg-grb.h @@ -57,11 +57,10 @@ static inline unsigned long __cmpxchg_u32(volatile int *m, unsigned long old, " mov.l @%1, %0 \n\t" /* load old value */ " cmp/eq %0, %2 \n\t" " bf 1f \n\t" /* if not equal */ - " mov.l %2, @%1 \n\t" /* store new value */ + " mov.l %3, @%1 \n\t" /* store new value */ "1: mov r1, r15 \n\t" /* LOGOUT */ - : "=&r" (retval), - "+r" (m) - : "r" (new) + : "=&r" (retval) + : "r" (m), "r" (old), "r" (new) : "memory" , "r0", "r1", "t"); return retval; -- cgit v0.10.2 From 2854e72b58dad82f5248b30a45dda0df49e5fe05 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Wed, 27 Jan 2010 17:32:22 +0100 Subject: x86: Use helpers for rlimits Make sure compiler won't do weird things with limits. Fetching them twice may return 2 different values after writable limits are implemented. We can either use rlimit helpers added in 3e10e716abf3c71bdb5d86b8f507f9e72236c9cd or ACCESS_ONCE if not applicable; this patch uses the helpers. Signed-off-by: Jiri Slaby LKML-Reference: <1264609942-24621-1-git-send-email-jslaby@suse.cz> Cc: Ingo Molnar Cc: Thomas Gleixner Signed-off-by: H. Peter Anvin diff --git a/arch/x86/ia32/ia32_aout.c b/arch/x86/ia32/ia32_aout.c index 2a4d073..0647478 100644 --- a/arch/x86/ia32/ia32_aout.c +++ b/arch/x86/ia32/ia32_aout.c @@ -297,7 +297,7 @@ static int load_aout_binary(struct linux_binprm *bprm, struct pt_regs *regs) * size limits imposed on them by creating programs with large * arrays in the data or bss. */ - rlim = current->signal->rlim[RLIMIT_DATA].rlim_cur; + rlim = rlimit(RLIMIT_DATA); if (rlim >= RLIM_INFINITY) rlim = ~0; if (ex.a_data + ex.a_bss > rlim) diff --git a/arch/x86/mm/mmap.c b/arch/x86/mm/mmap.c index c8191de..1dab519 100644 --- a/arch/x86/mm/mmap.c +++ b/arch/x86/mm/mmap.c @@ -71,7 +71,7 @@ static int mmap_is_legacy(void) if (current->personality & ADDR_COMPAT_LAYOUT) return 1; - if (current->signal->rlim[RLIMIT_STACK].rlim_cur == RLIM_INFINITY) + if (rlimit(RLIMIT_STACK) == RLIM_INFINITY) return 1; return sysctl_legacy_va_layout; @@ -96,7 +96,7 @@ static unsigned long mmap_rnd(void) static unsigned long mmap_base(void) { - unsigned long gap = current->signal->rlim[RLIMIT_STACK].rlim_cur; + unsigned long gap = rlimit(RLIMIT_STACK); if (gap < MIN_GAP) gap = MIN_GAP; -- cgit v0.10.2 From f655f5e956eaf4fcda983f3b3d37c76371556c27 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 28 Jan 2010 15:08:21 +0900 Subject: sh: mach-titan: Kill off unused PIO port mangling. Nothing is using this, kill it off. Fixing up access sizes can be done with trapped I/O for anyone wanting to make use of this for devices that need it, everything else is already pure MMIO. Signed-off-by: Paul Mundt diff --git a/arch/sh/Makefile b/arch/sh/Makefile index 30c80fd..588579a 100644 --- a/arch/sh/Makefile +++ b/arch/sh/Makefile @@ -149,7 +149,6 @@ machdir-$(CONFIG_SH_X3PROTO) += mach-x3proto machdir-$(CONFIG_SH_SH7763RDP) += mach-sh7763rdp machdir-$(CONFIG_SH_SH4202_MICRODEV) += mach-microdev machdir-$(CONFIG_SH_LANDISK) += mach-landisk -machdir-$(CONFIG_SH_TITAN) += mach-titan machdir-$(CONFIG_SH_LBOX_RE2) += mach-lboxre2 machdir-$(CONFIG_SH_CAYMAN) += mach-cayman machdir-$(CONFIG_SH_RSK) += mach-rsk diff --git a/arch/sh/boards/Makefile b/arch/sh/boards/Makefile index ce0f263..4f90f9b 100644 --- a/arch/sh/boards/Makefile +++ b/arch/sh/boards/Makefile @@ -8,3 +8,4 @@ obj-$(CONFIG_SH_SHMIN) += board-shmin.o obj-$(CONFIG_SH_EDOSK7760) += board-edosk7760.o obj-$(CONFIG_SH_ESPT) += board-espt.o obj-$(CONFIG_SH_POLARIS) += board-polaris.o +obj-$(CONFIG_SH_TITAN) += board-titan.o diff --git a/arch/sh/boards/board-titan.c b/arch/sh/boards/board-titan.c new file mode 100644 index 0000000..94c36c7b --- /dev/null +++ b/arch/sh/boards/board-titan.c @@ -0,0 +1,24 @@ +/* + * arch/sh/boards/titan/setup.c - Setup for Titan + * + * Copyright (C) 2006 Jamie Lenehan + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + */ +#include +#include +#include +#include + +static void __init init_titan_irq(void) +{ + /* enable individual interrupt mode for externals */ + plat_irq_setup_pins(IRQ_MODE_IRQ); +} + +static struct sh_machine_vector mv_titan __initmv = { + .mv_name = "Titan", + .mv_init_irq = init_titan_irq, +}; diff --git a/arch/sh/boards/mach-titan/Makefile b/arch/sh/boards/mach-titan/Makefile deleted file mode 100644 index 08d7537..0000000 --- a/arch/sh/boards/mach-titan/Makefile +++ /dev/null @@ -1,5 +0,0 @@ -# -# Makefile for the Nimble Microsystems TITAN specific parts of the kernel -# - -obj-y := setup.o io.o diff --git a/arch/sh/boards/mach-titan/io.c b/arch/sh/boards/mach-titan/io.c deleted file mode 100644 index 29754c5..0000000 --- a/arch/sh/boards/mach-titan/io.c +++ /dev/null @@ -1,108 +0,0 @@ -/* - * I/O routines for Titan - */ -#include -#include -#include -#include -#include - -static inline unsigned int port2adr(unsigned int port) -{ - maybebadio((unsigned long)port); - return port; -} - -u8 titan_inb(unsigned long port) -{ - if (PXSEG(port)) - return __raw_readb(port); - return __raw_readw(port2adr(port)) & 0xff; -} - -u8 titan_inb_p(unsigned long port) -{ - u8 v; - - if (PXSEG(port)) - v = __raw_readb(port); - else - v = __raw_readw(port2adr(port)) & 0xff; - ctrl_delay(); - return v; -} - -u16 titan_inw(unsigned long port) -{ - if (PXSEG(port)) - return __raw_readw(port); - else if (port >= 0x2000) - return __raw_readw(port2adr(port)); - else - maybebadio(port); - return 0; -} - -u32 titan_inl(unsigned long port) -{ - if (PXSEG(port)) - return __raw_readl(port); - else if (port >= 0x2000) - return __raw_readw(port2adr(port)); - else - maybebadio(port); - return 0; -} - -void titan_outb(u8 value, unsigned long port) -{ - if (PXSEG(port)) - __raw_writeb(value, port); - else - __raw_writew(value, port2adr(port)); -} - -void titan_outb_p(u8 value, unsigned long port) -{ - if (PXSEG(port)) - __raw_writeb(value, port); - else - __raw_writew(value, port2adr(port)); - ctrl_delay(); -} - -void titan_outw(u16 value, unsigned long port) -{ - if (PXSEG(port)) - __raw_writew(value, port); - else if (port >= 0x2000) - __raw_writew(value, port2adr(port)); - else - maybebadio(port); -} - -void titan_outl(u32 value, unsigned long port) -{ - if (PXSEG(port)) - __raw_writel(value, port); - else - maybebadio(port); -} - -void titan_insl(unsigned long port, void *dst, unsigned long count) -{ - maybebadio(port); -} - -void titan_outsl(unsigned long port, const void *src, unsigned long count) -{ - maybebadio(port); -} - -void __iomem *titan_ioport_map(unsigned long port, unsigned int size) -{ - if (PXSEG(port)) - return (void __iomem *)port; - - return (void __iomem *)port2adr(port); -} diff --git a/arch/sh/boards/mach-titan/setup.c b/arch/sh/boards/mach-titan/setup.c deleted file mode 100644 index 81e7e0f..0000000 --- a/arch/sh/boards/mach-titan/setup.c +++ /dev/null @@ -1,44 +0,0 @@ -/* - * arch/sh/boards/titan/setup.c - Setup for Titan - * - * Copyright (C) 2006 Jamie Lenehan - * - * This file is subject to the terms and conditions of the GNU General Public - * License. See the file "COPYING" in the main directory of this archive - * for more details. - */ -#include -#include -#include -#include - -static void __init init_titan_irq(void) -{ - /* enable individual interrupt mode for externals */ - plat_irq_setup_pins(IRQ_MODE_IRQ); -} - -static struct sh_machine_vector mv_titan __initmv = { - .mv_name = "Titan", - - .mv_inb = titan_inb, - .mv_inw = titan_inw, - .mv_inl = titan_inl, - .mv_outb = titan_outb, - .mv_outw = titan_outw, - .mv_outl = titan_outl, - - .mv_inb_p = titan_inb_p, - .mv_inw_p = titan_inw, - .mv_inl_p = titan_inl, - .mv_outb_p = titan_outb_p, - .mv_outw_p = titan_outw, - .mv_outl_p = titan_outl, - - .mv_insl = titan_insl, - .mv_outsl = titan_outsl, - - .mv_ioport_map = titan_ioport_map, - - .mv_init_irq = init_titan_irq, -}; -- cgit v0.10.2 From 621266bdf4b78fa92f7eabd7688b86f0dbd61d64 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 28 Jan 2010 15:55:37 +0900 Subject: sh: Quiet noisy coherent DMA unmapping on R2D. IRQs are re-enabled at a later stage when doing the unmapping on R2D via the sm501 USB coherent DMA, resulting in the irqs_disabled() check producing considerable noise for this configuration. Just kill off the check, which was blindly copied from x86 anyways. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/dma-mapping.h b/arch/sh/include/asm/dma-mapping.h index 87ced13..bea3337 100644 --- a/arch/sh/include/asm/dma-mapping.h +++ b/arch/sh/include/asm/dma-mapping.h @@ -89,8 +89,6 @@ static inline void dma_free_coherent(struct device *dev, size_t size, { struct dma_map_ops *ops = get_dma_ops(dev); - WARN_ON(irqs_disabled()); /* for portability */ - if (dma_release_from_coherent(dev, get_order(size), vaddr)) return; -- cgit v0.10.2 From 849593591ce3b5f9352966c9d410284dc04d03b6 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 28 Jan 2010 18:15:05 +0900 Subject: sh: Fix up read-only variable assignment in pcibios_align_resource(). arch/sh/drivers/pci/pci.c:167: error: assignment of read-only location '*res' Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/pci.c b/arch/sh/drivers/pci/pci.c index c481df6..191075e 100644 --- a/arch/sh/drivers/pci/pci.c +++ b/arch/sh/drivers/pci/pci.c @@ -162,10 +162,8 @@ void pcibios_align_resource(void *data, struct resource *res, /* * Put everything into 0x00-0xff region modulo 0x400. */ - if (start & 0x300) { + if (start & 0x300) start = (start + 0x3ff) & ~0x3ff; - res->start = start; - } } else if (res->flags & IORESOURCE_MEM) { if (start < PCIBIOS_MIN_MEM + chan->mem_resource->start) start = PCIBIOS_MIN_MEM + chan->mem_resource->start; -- cgit v0.10.2 From d627a2ebd1a303aa4ac1962c553d1e8738735d01 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 28 Jan 2010 18:17:29 +0900 Subject: sh: Fix up the ioremap_fixed() build for nommu. arch/sh/kernel/setup.c:455: error: implicit declaration of function 'ioremap_fixed_init' Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h index 98c62fa..bd5fafa 100644 --- a/arch/sh/include/asm/io.h +++ b/arch/sh/include/asm/io.h @@ -273,24 +273,6 @@ void __iomem *__ioremap_caller(unsigned long offset, unsigned long size, pgprot_t prot, void *caller); void __iounmap(void __iomem *addr); -#ifdef CONFIG_IOREMAP_FIXED -extern void __iomem *ioremap_fixed(resource_size_t, unsigned long, - unsigned long, pgprot_t); -extern int iounmap_fixed(void __iomem *); -extern void ioremap_fixed_init(void); -#else -static inline void __iomem * -ioremap_fixed(resource_size_t phys_addr, unsigned long offset, - unsigned long size, pgprot_t prot) -{ - BUG(); - return NULL; -} - -static inline void ioremap_fixed_init(void) { } -static inline int iounmap_fixed(void __iomem *addr) { return -EINVAL; } -#endif - static inline void __iomem * __ioremap(unsigned long offset, unsigned long size, pgprot_t prot) { @@ -365,6 +347,24 @@ ioremap_prot(resource_size_t offset, unsigned long size, unsigned long flags) } #endif +#ifdef CONFIG_IOREMAP_FIXED +extern void __iomem *ioremap_fixed(resource_size_t, unsigned long, + unsigned long, pgprot_t); +extern int iounmap_fixed(void __iomem *); +extern void ioremap_fixed_init(void); +#else +static inline void __iomem * +ioremap_fixed(resource_size_t phys_addr, unsigned long offset, + unsigned long size, pgprot_t prot) +{ + BUG(); + return NULL; +} + +static inline void ioremap_fixed_init(void) { } +static inline int iounmap_fixed(void __iomem *addr) { return -EINVAL; } +#endif + #define ioremap_nocache ioremap #define iounmap __iounmap -- cgit v0.10.2 From abec86a80247ab3a40d2afc9c8e8c26efcab2391 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 28 Jan 2010 18:26:37 +0900 Subject: sh: Kill off superfluous sh_bios references from compressed loader. The VBR value needs to be initialized in order for any of the bios trapping to work, and as we don't do that until slightly later in the boot process, this has no real chance of working. The only reason for using this in the past was due to the fact that early serial initialization wasn't possible, but now that early sh-sci is supported by everyone, there's no longer any reason to support the sh_bios hacks in the first place. Given that, we just kill off the references completely. Signed-off-by: Paul Mundt diff --git a/arch/sh/boot/compressed/Makefile b/arch/sh/boot/compressed/Makefile index 6b95a2a..5d660b9 100644 --- a/arch/sh/boot/compressed/Makefile +++ b/arch/sh/boot/compressed/Makefile @@ -11,10 +11,6 @@ targets := vmlinux vmlinux.bin vmlinux.bin.gz \ OBJECTS = $(obj)/head_$(BITS).o $(obj)/misc.o $(obj)/cache.o -ifdef CONFIG_SH_STANDARD_BIOS -OBJECTS += $(obj)/../../kernel/sh_bios.o -endif - # # IMAGE_OFFSET is the load offset of the compression loader # diff --git a/arch/sh/boot/compressed/misc.c b/arch/sh/boot/compressed/misc.c index 4b319e9..d509a50 100644 --- a/arch/sh/boot/compressed/misc.c +++ b/arch/sh/boot/compressed/misc.c @@ -14,7 +14,6 @@ #include #include #include -#include /* * gzip declarations @@ -66,29 +65,11 @@ static unsigned long free_mem_end_ptr; #include "../../../../lib/decompress_unlzo.c" #endif -#ifdef CONFIG_SH_STANDARD_BIOS -size_t strlen(const char *s) -{ - int i = 0; - - while (*s++) - i++; - return i; -} - -int puts(const char *s) -{ - int len = strlen(s); - sh_bios_console_write(s, len); - return len; -} -#else int puts(const char *s) { /* This should be updated to use the sh-sci routines */ return 0; } -#endif void* memset(void* s, int c, size_t n) { -- cgit v0.10.2 From 339ce1a4dc2ca26444c4f65c31b71a5056f3bb0b Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Mon, 18 Jan 2010 16:47:07 +1100 Subject: perf: Fix inconsistency between IP and callchain sampling When running perf across all cpus with backtracing (-a -g), sometimes we get samples without associated backtraces: 23.44% init [kernel] [k] restore 11.46% init eeba0c [k] 0x00000000eeba0c 6.77% swapper [kernel] [k] .perf_ctx_adjust_freq 5.73% init [kernel] [k] .__trace_hcall_entry 4.69% perf libc-2.9.so [.] 0x0000000006bb8c | |--11.11%-- 0xfffa941bbbc It turns out the backtrace code has a check for the idle task and the IP sampling does not. This creates problems when profiling an interrupt heavy workload (in my case 10Gbit ethernet) since we get no backtraces for interrupts received while idle (ie most of the workload). Right now x86 and sh check that current is not NULL, which should never happen so remove that too. Idle task's exclusion must be performed from the core code, on top of perf_event_attr:exclude_idle. Signed-off-by: Anton Blanchard Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Ingo Molnar Cc: Benjamin Herrenschmidt Cc: Paul Mundt LKML-Reference: <20100118054707.GT12666@kryten> Signed-off-by: Frederic Weisbecker diff --git a/arch/powerpc/kernel/perf_callchain.c b/arch/powerpc/kernel/perf_callchain.c index a3c11ca..95ad9da 100644 --- a/arch/powerpc/kernel/perf_callchain.c +++ b/arch/powerpc/kernel/perf_callchain.c @@ -495,9 +495,6 @@ struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) entry->nr = 0; - if (current->pid == 0) /* idle task? */ - return entry; - if (!user_mode(regs)) { perf_callchain_kernel(regs, entry); if (current->mm) diff --git a/arch/sh/kernel/perf_callchain.c b/arch/sh/kernel/perf_callchain.c index 24ea837..a9dd3ab 100644 --- a/arch/sh/kernel/perf_callchain.c +++ b/arch/sh/kernel/perf_callchain.c @@ -68,9 +68,6 @@ perf_do_callchain(struct pt_regs *regs, struct perf_callchain_entry *entry) is_user = user_mode(regs); - if (!current || current->pid == 0) - return; - if (is_user && current->state != TASK_RUNNING) return; diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index b1bb8c5..ed1998b 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -2425,9 +2425,6 @@ perf_do_callchain(struct pt_regs *regs, struct perf_callchain_entry *entry) is_user = user_mode(regs); - if (!current || current->pid == 0) - return; - if (is_user && current->state != TASK_RUNNING) return; -- cgit v0.10.2 From 1f43cfb9474d1c4f22598b6e3213ec035be6dd56 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 28 Jan 2010 13:47:25 -0700 Subject: of: merge machine_is_compatible() Merge common code between PowerPC and Microblaze Signed-off-by: Grant Likely Tested-by: Wolfram Sang Acked-by: Benjamin Herrenschmidt diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 5505bcf..16a001c 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -318,24 +318,6 @@ void __init early_init_devtree(void *params) pr_debug(" <- early_init_devtree()\n"); } -/** - * Indicates whether the root node has a given value in its - * compatible property. - */ -int machine_is_compatible(const char *compat) -{ - struct device_node *root; - int rc = 0; - - root = of_find_node_by_path("/"); - if (root) { - rc = of_device_is_compatible(root, compat); - of_node_put(root); - } - return rc; -} -EXPORT_SYMBOL(machine_is_compatible); - /******* * * New implementation of the OF "find" APIs, return a refcounted diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 877fad9..caeea9a5 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -754,25 +754,6 @@ void __init early_init_devtree(void *params) DBG(" <- early_init_devtree()\n"); } - -/** - * Indicates whether the root node has a given value in its - * compatible property. - */ -int machine_is_compatible(const char *compat) -{ - struct device_node *root; - int rc = 0; - - root = of_find_node_by_path("/"); - if (root) { - rc = of_device_is_compatible(root, compat); - of_node_put(root); - } - return rc; -} -EXPORT_SYMBOL(machine_is_compatible); - /******* * * New implementation of the OF "find" APIs, return a refcounted diff --git a/drivers/of/base.c b/drivers/of/base.c index ec56739..dba995b 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -144,6 +144,27 @@ int of_device_is_compatible(const struct device_node *device, EXPORT_SYMBOL(of_device_is_compatible); /** + * machine_is_compatible - Test root of device tree for a given compatible value + * @compat: compatible string to look for in root node's compatible property. + * + * Returns true if the root node has the given value in its + * compatible property. + */ +int machine_is_compatible(const char *compat) +{ + struct device_node *root; + int rc = 0; + + root = of_find_node_by_path("/"); + if (root) { + rc = of_device_is_compatible(root, compat); + of_node_put(root); + } + return rc; +} +EXPORT_SYMBOL(machine_is_compatible); + +/** * of_device_is_available - check if a device is available for use * * @device: Node to check for availability -- cgit v0.10.2 From 923f7e30b480438f1e86e01e5cde814248b59a39 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 28 Jan 2010 13:52:53 -0700 Subject: of: Merge of_node_get() and of_node_put() Merge common code between PowerPC and MicroBlaze Signed-off-by: Grant Likely Tested-by: Wolfram Sang Acked-by: Benjamin Herrenschmidt diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 16a001c..46407e6 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -350,80 +350,6 @@ struct device_node *of_find_node_by_phandle(phandle handle) } EXPORT_SYMBOL(of_find_node_by_phandle); -/** - * of_node_get - Increment refcount of a node - * @node: Node to inc refcount, NULL is supported to - * simplify writing of callers - * - * Returns node. - */ -struct device_node *of_node_get(struct device_node *node) -{ - if (node) - kref_get(&node->kref); - return node; -} -EXPORT_SYMBOL(of_node_get); - -static inline struct device_node *kref_to_device_node(struct kref *kref) -{ - return container_of(kref, struct device_node, kref); -} - -/** - * of_node_release - release a dynamically allocated node - * @kref: kref element of the node to be released - * - * In of_node_put() this function is passed to kref_put() - * as the destructor. - */ -static void of_node_release(struct kref *kref) -{ - struct device_node *node = kref_to_device_node(kref); - struct property *prop = node->properties; - - /* We should never be releasing nodes that haven't been detached. */ - if (!of_node_check_flag(node, OF_DETACHED)) { - printk(KERN_INFO "WARNING: Bad of_node_put() on %s\n", - node->full_name); - dump_stack(); - kref_init(&node->kref); - return; - } - - if (!of_node_check_flag(node, OF_DYNAMIC)) - return; - - while (prop) { - struct property *next = prop->next; - kfree(prop->name); - kfree(prop->value); - kfree(prop); - prop = next; - - if (!prop) { - prop = node->deadprops; - node->deadprops = NULL; - } - } - kfree(node->full_name); - kfree(node->data); - kfree(node); -} - -/** - * of_node_put - Decrement refcount of a node - * @node: Node to dec refcount, NULL is supported to - * simplify writing of callers - * - */ -void of_node_put(struct device_node *node) -{ - if (node) - kref_put(&node->kref, of_node_release); -} -EXPORT_SYMBOL(of_node_put); - /* * Plug a device node into the tree and global list. */ diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index caeea9a5..deccd91 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -817,79 +817,6 @@ struct device_node *of_find_next_cache_node(struct device_node *np) return NULL; } -/** - * of_node_get - Increment refcount of a node - * @node: Node to inc refcount, NULL is supported to - * simplify writing of callers - * - * Returns node. - */ -struct device_node *of_node_get(struct device_node *node) -{ - if (node) - kref_get(&node->kref); - return node; -} -EXPORT_SYMBOL(of_node_get); - -static inline struct device_node * kref_to_device_node(struct kref *kref) -{ - return container_of(kref, struct device_node, kref); -} - -/** - * of_node_release - release a dynamically allocated node - * @kref: kref element of the node to be released - * - * In of_node_put() this function is passed to kref_put() - * as the destructor. - */ -static void of_node_release(struct kref *kref) -{ - struct device_node *node = kref_to_device_node(kref); - struct property *prop = node->properties; - - /* We should never be releasing nodes that haven't been detached. */ - if (!of_node_check_flag(node, OF_DETACHED)) { - printk("WARNING: Bad of_node_put() on %s\n", node->full_name); - dump_stack(); - kref_init(&node->kref); - return; - } - - if (!of_node_check_flag(node, OF_DYNAMIC)) - return; - - while (prop) { - struct property *next = prop->next; - kfree(prop->name); - kfree(prop->value); - kfree(prop); - prop = next; - - if (!prop) { - prop = node->deadprops; - node->deadprops = NULL; - } - } - kfree(node->full_name); - kfree(node->data); - kfree(node); -} - -/** - * of_node_put - Decrement refcount of a node - * @node: Node to dec refcount, NULL is supported to - * simplify writing of callers - * - */ -void of_node_put(struct device_node *node) -{ - if (node) - kref_put(&node->kref, of_node_release); -} -EXPORT_SYMBOL(of_node_put); - /* * Plug a device node into the tree and global list. */ diff --git a/drivers/of/base.c b/drivers/of/base.c index dba995b..cf89ee6 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -60,6 +60,81 @@ int of_n_size_cells(struct device_node *np) } EXPORT_SYMBOL(of_n_size_cells); +#if !defined(CONFIG_SPARC) /* SPARC doesn't do ref counting (yet) */ +/** + * of_node_get - Increment refcount of a node + * @node: Node to inc refcount, NULL is supported to + * simplify writing of callers + * + * Returns node. + */ +struct device_node *of_node_get(struct device_node *node) +{ + if (node) + kref_get(&node->kref); + return node; +} +EXPORT_SYMBOL(of_node_get); + +static inline struct device_node *kref_to_device_node(struct kref *kref) +{ + return container_of(kref, struct device_node, kref); +} + +/** + * of_node_release - release a dynamically allocated node + * @kref: kref element of the node to be released + * + * In of_node_put() this function is passed to kref_put() + * as the destructor. + */ +static void of_node_release(struct kref *kref) +{ + struct device_node *node = kref_to_device_node(kref); + struct property *prop = node->properties; + + /* We should never be releasing nodes that haven't been detached. */ + if (!of_node_check_flag(node, OF_DETACHED)) { + pr_err("ERROR: Bad of_node_put() on %s\n", node->full_name); + dump_stack(); + kref_init(&node->kref); + return; + } + + if (!of_node_check_flag(node, OF_DYNAMIC)) + return; + + while (prop) { + struct property *next = prop->next; + kfree(prop->name); + kfree(prop->value); + kfree(prop); + prop = next; + + if (!prop) { + prop = node->deadprops; + node->deadprops = NULL; + } + } + kfree(node->full_name); + kfree(node->data); + kfree(node); +} + +/** + * of_node_put - Decrement refcount of a node + * @node: Node to dec refcount, NULL is supported to + * simplify writing of callers + * + */ +void of_node_put(struct device_node *node) +{ + if (node) + kref_put(&node->kref, of_node_release); +} +EXPORT_SYMBOL(of_node_put); +#endif /* !CONFIG_SPARC */ + struct property *of_find_property(const struct device_node *np, const char *name, int *lenp) -- cgit v0.10.2 From 6016a363f6b56b46b24655bcfc0499b715851cf3 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 28 Jan 2010 14:06:53 -0700 Subject: of: unify phandle name in struct device_node In struct device_node, the phandle is named 'linux_phandle' for PowerPC and MicroBlaze, and 'node' for SPARC. There is no good reason for the difference, it is just an artifact of the code diverging over a couple of years. This patch renames both to simply .phandle. Note: the .node also existed in PowerPC/MicroBlaze, but the only user seems to be arch/powerpc/platforms/powermac/pfunc_core.c. It doesn't look like the assignment between .linux_phandle and .node is significantly different enough to warrant the separate code paths unless ibm,phandle properties actually appear in Apple device trees. I think it is safe to eliminate the old .node property and use phandle everywhere. Signed-off-by: Grant Likely Acked-by: David S. Miller Tested-by: Wolfram Sang Acked-by: Benjamin Herrenschmidt diff --git a/arch/microblaze/kernel/of_platform.c b/arch/microblaze/kernel/of_platform.c index acf4574..1c6d684 100644 --- a/arch/microblaze/kernel/of_platform.c +++ b/arch/microblaze/kernel/of_platform.c @@ -185,7 +185,7 @@ EXPORT_SYMBOL(of_find_device_by_node); static int of_dev_phandle_match(struct device *dev, void *data) { phandle *ph = data; - return to_of_device(dev)->node->linux_phandle == *ph; + return to_of_device(dev)->node->phandle == *ph; } struct of_device *of_find_device_by_phandle(phandle ph) diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 46407e6..6eff83a 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -342,7 +342,7 @@ struct device_node *of_find_node_by_phandle(phandle handle) read_lock(&devtree_lock); for (np = allnodes; np != NULL; np = np->allnext) - if (np->linux_phandle == handle) + if (np->phandle == handle) break; of_node_get(np); read_unlock(&devtree_lock); diff --git a/arch/powerpc/kernel/of_platform.c b/arch/powerpc/kernel/of_platform.c index 1a4fc0d..666d08d 100644 --- a/arch/powerpc/kernel/of_platform.c +++ b/arch/powerpc/kernel/of_platform.c @@ -214,7 +214,7 @@ EXPORT_SYMBOL(of_find_device_by_node); static int of_dev_phandle_match(struct device *dev, void *data) { phandle *ph = data; - return to_of_device(dev)->node->linux_phandle == *ph; + return to_of_device(dev)->node->phandle == *ph; } struct of_device *of_find_device_by_phandle(phandle ph) diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index deccd91..1ed2ec2 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -778,7 +778,7 @@ struct device_node *of_find_node_by_phandle(phandle handle) read_lock(&devtree_lock); for (np = allnodes; np != 0; np = np->allnext) - if (np->linux_phandle == handle) + if (np->phandle == handle) break; of_node_get(np); read_unlock(&devtree_lock); @@ -907,9 +907,9 @@ static int of_finish_dynamic_node(struct device_node *node) if (machine_is(powermac)) return -ENODEV; - /* fix up new node's linux_phandle field */ + /* fix up new node's phandle field */ if ((ibm_phandle = of_get_property(node, "ibm,phandle", NULL))) - node->linux_phandle = *ibm_phandle; + node->phandle = *ibm_phandle; out: of_node_put(parent); diff --git a/arch/powerpc/platforms/cell/spu_manage.c b/arch/powerpc/platforms/cell/spu_manage.c index 4c506c1..891f18e 100644 --- a/arch/powerpc/platforms/cell/spu_manage.c +++ b/arch/powerpc/platforms/cell/spu_manage.c @@ -457,7 +457,7 @@ neighbour_spu(int cbe, struct device_node *target, struct device_node *avoid) continue; vic_handles = of_get_property(spu_dn, "vicinity", &lenp); for (i=0; i < (lenp / sizeof(phandle)); i++) { - if (vic_handles[i] == target->linux_phandle) + if (vic_handles[i] == target->phandle) return spu; } } @@ -499,7 +499,7 @@ static void init_affinity_node(int cbe) if (strcmp(name, "spe") == 0) { spu = devnode_spu(cbe, vic_dn); - avoid_ph = last_spu_dn->linux_phandle; + avoid_ph = last_spu_dn->phandle; } else { /* * "mic-tm" and "bif0" nodes do not have @@ -514,7 +514,7 @@ static void init_affinity_node(int cbe) last_spu->has_mem_affinity = 1; spu->has_mem_affinity = 1; } - avoid_ph = vic_dn->linux_phandle; + avoid_ph = vic_dn->phandle; } list_add_tail(&spu->aff_list, &last_spu->aff_list); diff --git a/arch/powerpc/platforms/powermac/pfunc_core.c b/arch/powerpc/platforms/powermac/pfunc_core.c index 96d5ce5..ede49e7 100644 --- a/arch/powerpc/platforms/powermac/pfunc_core.c +++ b/arch/powerpc/platforms/powermac/pfunc_core.c @@ -842,7 +842,7 @@ struct pmf_function *__pmf_find_function(struct device_node *target, list_for_each_entry(func, &dev->functions, link) { if (name && strcmp(name, func->name)) continue; - if (func->phandle && target->node != func->phandle) + if (func->phandle && target->phandle != func->phandle) continue; if ((func->flags & flags) == 0) continue; diff --git a/arch/sparc/kernel/devices.c b/arch/sparc/kernel/devices.c index b171ae8..b062de9 100644 --- a/arch/sparc/kernel/devices.c +++ b/arch/sparc/kernel/devices.c @@ -59,7 +59,7 @@ static int __cpu_find_by(int (*compare)(int, int, void *), void *compare_arg, cur_inst = 0; for_each_node_by_type(dp, "cpu") { - int err = check_cpu_node(dp->node, &cur_inst, + int err = check_cpu_node(dp->phandle, &cur_inst, compare, compare_arg, prom_node, mid); if (!err) { diff --git a/arch/sparc/kernel/of_device_32.c b/arch/sparc/kernel/of_device_32.c index 4c26eb5..09138d4 100644 --- a/arch/sparc/kernel/of_device_32.c +++ b/arch/sparc/kernel/of_device_32.c @@ -433,7 +433,7 @@ build_resources: if (!parent) dev_set_name(&op->dev, "root"); else - dev_set_name(&op->dev, "%08x", dp->node); + dev_set_name(&op->dev, "%08x", dp->phandle); if (of_device_register(op)) { printk("%s: Could not register of device.\n", diff --git a/arch/sparc/kernel/of_device_64.c b/arch/sparc/kernel/of_device_64.c index 881947e..036f18a 100644 --- a/arch/sparc/kernel/of_device_64.c +++ b/arch/sparc/kernel/of_device_64.c @@ -666,7 +666,7 @@ static struct of_device * __init scan_one_device(struct device_node *dp, if (!parent) dev_set_name(&op->dev, "root"); else - dev_set_name(&op->dev, "%08x", dp->node); + dev_set_name(&op->dev, "%08x", dp->phandle); if (of_device_register(op)) { printk("%s: Could not register of device.\n", diff --git a/arch/sparc/kernel/prom_common.c b/arch/sparc/kernel/prom_common.c index d80a65d..5832e13 100644 --- a/arch/sparc/kernel/prom_common.c +++ b/arch/sparc/kernel/prom_common.c @@ -42,7 +42,7 @@ struct device_node *of_find_node_by_phandle(phandle handle) struct device_node *np; for (np = allnodes; np; np = np->allnext) - if (np->node == handle) + if (np->phandle == handle) break; return np; @@ -89,7 +89,7 @@ int of_set_property(struct device_node *dp, const char *name, void *val, int len void *old_val = prop->value; int ret; - ret = prom_setprop(dp->node, name, val, len); + ret = prom_setprop(dp->phandle, name, val, len); err = -EINVAL; if (ret >= 0) { @@ -236,7 +236,7 @@ static struct device_node * __init prom_create_node(phandle node, dp->name = get_one_property(node, "name"); dp->type = get_one_property(node, "device_type"); - dp->node = node; + dp->phandle = node; dp->properties = build_prop_list(node); @@ -313,7 +313,7 @@ void __init prom_build_devicetree(void) nextp = &allnodes->allnext; allnodes->child = prom_build_tree(allnodes, - prom_getchild(allnodes->node), + prom_getchild(allnodes->phandle), &nextp); of_console_init(); diff --git a/arch/sparc/kernel/smp_64.c b/arch/sparc/kernel/smp_64.c index aa36223..eb14844 100644 --- a/arch/sparc/kernel/smp_64.c +++ b/arch/sparc/kernel/smp_64.c @@ -370,7 +370,7 @@ static int __cpuinit smp_boot_one_cpu(unsigned int cpu) } else { struct device_node *dp = of_find_node_by_cpuid(cpu); - prom_startcpu(dp->node, entry, cookie); + prom_startcpu(dp->phandle, entry, cookie); } for (timeout = 0; timeout < 50000; timeout++) { diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 616a476..7f88611 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -310,12 +310,11 @@ unsigned long __init unflatten_dt_node(unsigned long mem, __alignof__(struct property)); if (allnextpp) { if (strcmp(pname, "linux,phandle") == 0) { - np->node = *((u32 *)*p); - if (np->linux_phandle == 0) - np->linux_phandle = np->node; + if (np->phandle == 0) + np->phandle = *((u32 *)*p); } if (strcmp(pname, "ibm,phandle") == 0) - np->linux_phandle = *((u32 *)*p); + np->phandle = *((u32 *)*p); pp->name = pname; pp->length = sz; pp->value = (void *)*p; diff --git a/drivers/sbus/char/openprom.c b/drivers/sbus/char/openprom.c index 75ac19b..fc2f676 100644 --- a/drivers/sbus/char/openprom.c +++ b/drivers/sbus/char/openprom.c @@ -233,7 +233,7 @@ static int opromnext(void __user *argp, unsigned int cmd, struct device_node *dp ph = 0; if (dp) - ph = dp->node; + ph = dp->phandle; data->current_node = dp; *((int *) op->oprom_array) = ph; @@ -256,7 +256,7 @@ static int oprompci2node(void __user *argp, struct device_node *dp, struct openp dp = pci_device_to_OF_node(pdev); data->current_node = dp; - *((int *)op->oprom_array) = dp->node; + *((int *)op->oprom_array) = dp->phandle; op->oprom_size = sizeof(int); err = copyout(argp, op, bufsize + sizeof(int)); @@ -273,7 +273,7 @@ static int oprompath2node(void __user *argp, struct device_node *dp, struct open dp = of_find_node_by_path(op->oprom_array); if (dp) - ph = dp->node; + ph = dp->phandle; data->current_node = dp; *((int *)op->oprom_array) = ph; op->oprom_size = sizeof(int); @@ -540,7 +540,7 @@ static int opiocgetnext(unsigned int cmd, void __user *argp) } } if (dp) - nd = dp->node; + nd = dp->phandle; if (copy_to_user(argp, &nd, sizeof(phandle))) return -EFAULT; @@ -570,7 +570,7 @@ static int openprom_bsd_ioctl(struct inode * inode, struct file * file, case OPIOCGETOPTNODE: BUILD_BUG_ON(sizeof(phandle) != sizeof(int)); - if (copy_to_user(argp, &options_node->node, sizeof(phandle))) + if (copy_to_user(argp, &options_node->phandle, sizeof(phandle))) return -EFAULT; return 0; diff --git a/drivers/video/aty/atyfb_base.c b/drivers/video/aty/atyfb_base.c index 913b4a4..bb20987 100644 --- a/drivers/video/aty/atyfb_base.c +++ b/drivers/video/aty/atyfb_base.c @@ -3104,7 +3104,7 @@ static int __devinit atyfb_setup_sparc(struct pci_dev *pdev, } dp = pci_device_to_OF_node(pdev); - if (node == dp->node) { + if (node == dp->phandle) { struct fb_var_screeninfo *var = &default_var; unsigned int N, P, Q, M, T, R; u32 v_total, h_total; diff --git a/include/linux/of.h b/include/linux/of.h index d4c014a..dbabf86 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -39,10 +39,7 @@ struct of_irq_controller; struct device_node { const char *name; const char *type; - phandle node; -#if !defined(CONFIG_SPARC) - phandle linux_phandle; -#endif + phandle phandle; char *full_name; struct property *properties; diff --git a/sound/aoa/fabrics/layout.c b/sound/aoa/fabrics/layout.c index 586965f..7a437da 100644 --- a/sound/aoa/fabrics/layout.c +++ b/sound/aoa/fabrics/layout.c @@ -768,7 +768,7 @@ static int check_codec(struct aoa_codec *codec, "required property %s not present\n", propname); return -ENODEV; } - if (*ref != codec->node->linux_phandle) { + if (*ref != codec->node->phandle) { printk(KERN_INFO "snd-aoa-fabric-layout: " "%s doesn't match!\n", propname); return -ENODEV; -- cgit v0.10.2 From ea2c68a08fedb5053ba312d661e47df9f4d72411 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 13 Jan 2010 19:38:30 +0800 Subject: tracing: Simplify test for function_graph tracing start point In the function graph tracer, a calling function is to be traced only when it is enabled through the set_graph_function file, or when it is nested in an enabled function. Current code uses TSK_TRACE_FL_GRAPH to test whether it is nested or not. Looking at the code, we can get this: (trace->depth > 0) <==> (TSK_TRACE_FL_GRAPH is set) trace->depth is more explicit to tell that it is nested. So we use trace->depth directly and simplify the code. No functionality is changed. TSK_TRACE_FL_GRAPH is not removed yet, it is left for future usage. Signed-off-by: Lai Jiangshan Cc: Ingo Molnar Cc: Steven Rostedt LKML-Reference: <4B4DB0B6.7040607@cn.fujitsu.com> Signed-off-by: Frederic Weisbecker diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 4df6a77..ce077fb 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -504,7 +504,7 @@ static inline int ftrace_graph_addr(unsigned long addr) { int i; - if (!ftrace_graph_count || test_tsk_trace_graph(current)) + if (!ftrace_graph_count) return 1; for (i = 0; i < ftrace_graph_count; i++) { diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index f225229..616b135 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -215,7 +215,8 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) if (!ftrace_trace_task(current)) return 0; - if (!ftrace_graph_addr(trace->func)) + /* trace it when it is-nested-in or is a function enabled. */ + if (!(trace->depth || ftrace_graph_addr(trace->func))) return 0; local_irq_save(flags); @@ -228,9 +229,6 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) } else { ret = 0; } - /* Only do the atomic if it is not already set */ - if (!test_tsk_trace_graph(current)) - set_tsk_trace_graph(current); atomic_dec(&data->disabled); local_irq_restore(flags); @@ -278,8 +276,6 @@ void trace_graph_return(struct ftrace_graph_ret *trace) pc = preempt_count(); __trace_graph_return(tr, trace, flags, pc); } - if (!trace->depth) - clear_tsk_trace_graph(current); atomic_dec(&data->disabled); local_irq_restore(flags); } -- cgit v0.10.2 From 430ad5a600a83956749307b13257c464c3826b55 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Thu, 28 Jan 2010 09:32:29 +0800 Subject: perf: Factorize trace events raw sample buffer operations Introduce ftrace_perf_buf_prepare() and ftrace_perf_buf_submit() to gather the common code that operates on raw events sampling buffer. This cleans up redundant code between regular trace events, syscall events and kprobe events. Changelog v1->v2: - Rename function name as per Masami and Frederic's suggestion - Add __kprobes for ftrace_perf_buf_prepare() and make ftrace_perf_buf_submit() inline as per Masami's suggestion - Export ftrace_perf_buf_prepare since modules will use it Signed-off-by: Xiao Guangrong Acked-by: Masami Hiramatsu Cc: Ingo Molnar Cc: Steven Rostedt Cc: Paul Mackerras Cc: Jason Baron Cc: Peter Zijlstra LKML-Reference: <4B60E92D.9000808@cn.fujitsu.com> Signed-off-by: Frederic Weisbecker diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h index 0a09e75..cd95919 100644 --- a/include/linux/ftrace_event.h +++ b/include/linux/ftrace_event.h @@ -5,6 +5,7 @@ #include #include #include +#include struct trace_array; struct tracer; @@ -138,9 +139,6 @@ struct ftrace_event_call { #define FTRACE_MAX_PROFILE_SIZE 2048 -extern char *perf_trace_buf; -extern char *perf_trace_buf_nmi; - #define MAX_FILTER_PRED 32 #define MAX_FILTER_STR_VAL 256 /* Should handle KSYM_SYMBOL_LEN */ @@ -195,6 +193,20 @@ extern void ftrace_profile_disable(int event_id); extern int ftrace_profile_set_filter(struct perf_event *event, int event_id, char *filter_str); extern void ftrace_profile_free_filter(struct perf_event *event); +extern void * +ftrace_perf_buf_prepare(int size, unsigned short type, int *rctxp, + unsigned long *irq_flags); + +static inline void +ftrace_perf_buf_submit(void *raw_data, int size, int rctx, u64 addr, + u64 count, unsigned long irq_flags) +{ + struct trace_entry *entry = raw_data; + + perf_tp_event(entry->type, addr, count, raw_data, size); + perf_swevent_put_recursion_context(rctx); + local_irq_restore(irq_flags); +} #endif #endif /* _LINUX_FTRACE_EVENT_H */ diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h index 4a46a60..f2c09e4 100644 --- a/include/trace/ftrace.h +++ b/include/trace/ftrace.h @@ -850,22 +850,12 @@ ftrace_profile_templ_##call(struct ftrace_event_call *event_call, \ proto) \ { \ struct ftrace_data_offsets_##call __maybe_unused __data_offsets;\ - extern int perf_swevent_get_recursion_context(void); \ - extern void perf_swevent_put_recursion_context(int rctx); \ - extern void perf_tp_event(int, u64, u64, void *, int); \ struct ftrace_raw_##call *entry; \ u64 __addr = 0, __count = 1; \ unsigned long irq_flags; \ - struct trace_entry *ent; \ int __entry_size; \ int __data_size; \ - char *trace_buf; \ - char *raw_data; \ - int __cpu; \ int rctx; \ - int pc; \ - \ - pc = preempt_count(); \ \ __data_size = ftrace_get_offsets_##call(&__data_offsets, args); \ __entry_size = ALIGN(__data_size + sizeof(*entry) + sizeof(u32),\ @@ -875,42 +865,16 @@ ftrace_profile_templ_##call(struct ftrace_event_call *event_call, \ if (WARN_ONCE(__entry_size > FTRACE_MAX_PROFILE_SIZE, \ "profile buffer not large enough")) \ return; \ - \ - local_irq_save(irq_flags); \ - \ - rctx = perf_swevent_get_recursion_context(); \ - if (rctx < 0) \ - goto end_recursion; \ - \ - __cpu = smp_processor_id(); \ - \ - if (in_nmi()) \ - trace_buf = rcu_dereference(perf_trace_buf_nmi); \ - else \ - trace_buf = rcu_dereference(perf_trace_buf); \ - \ - if (!trace_buf) \ - goto end; \ - \ - raw_data = per_cpu_ptr(trace_buf, __cpu); \ - \ - *(u64 *)(&raw_data[__entry_size - sizeof(u64)]) = 0ULL; \ - entry = (struct ftrace_raw_##call *)raw_data; \ - ent = &entry->ent; \ - tracing_generic_entry_update(ent, irq_flags, pc); \ - ent->type = event_call->id; \ - \ + entry = (struct ftrace_raw_##call *)ftrace_perf_buf_prepare( \ + __entry_size, event_call->id, &rctx, &irq_flags); \ + if (!entry) \ + return; \ tstruct \ \ { assign; } \ \ - perf_tp_event(event_call->id, __addr, __count, entry, \ - __entry_size); \ - \ -end: \ - perf_swevent_put_recursion_context(rctx); \ -end_recursion: \ - local_irq_restore(irq_flags); \ + ftrace_perf_buf_submit(entry, __entry_size, rctx, __addr, \ + __count, irq_flags); \ } #undef DEFINE_EVENT diff --git a/kernel/trace/trace_event_profile.c b/kernel/trace/trace_event_profile.c index 9e25573..f0d6930 100644 --- a/kernel/trace/trace_event_profile.c +++ b/kernel/trace/trace_event_profile.c @@ -6,14 +6,12 @@ */ #include +#include #include "trace.h" -char *perf_trace_buf; -EXPORT_SYMBOL_GPL(perf_trace_buf); - -char *perf_trace_buf_nmi; -EXPORT_SYMBOL_GPL(perf_trace_buf_nmi); +static char *perf_trace_buf; +static char *perf_trace_buf_nmi; typedef typeof(char [FTRACE_MAX_PROFILE_SIZE]) perf_trace_t ; @@ -120,3 +118,47 @@ void ftrace_profile_disable(int event_id) } mutex_unlock(&event_mutex); } + +__kprobes void *ftrace_perf_buf_prepare(int size, unsigned short type, + int *rctxp, unsigned long *irq_flags) +{ + struct trace_entry *entry; + char *trace_buf, *raw_data; + int pc, cpu; + + pc = preempt_count(); + + /* Protect the per cpu buffer, begin the rcu read side */ + local_irq_save(*irq_flags); + + *rctxp = perf_swevent_get_recursion_context(); + if (*rctxp < 0) + goto err_recursion; + + cpu = smp_processor_id(); + + if (in_nmi()) + trace_buf = rcu_dereference(perf_trace_buf_nmi); + else + trace_buf = rcu_dereference(perf_trace_buf); + + if (!trace_buf) + goto err; + + raw_data = per_cpu_ptr(trace_buf, cpu); + + /* zero the dead bytes from align to not leak stack to user */ + *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; + + entry = (struct trace_entry *)raw_data; + tracing_generic_entry_update(entry, *irq_flags, pc); + entry->type = type; + + return raw_data; +err: + perf_swevent_put_recursion_context(*rctxp); +err_recursion: + local_irq_restore(*irq_flags); + return NULL; +} +EXPORT_SYMBOL_GPL(ftrace_perf_buf_prepare); diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index d6266ca..2e28ee3 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -1243,14 +1243,10 @@ static __kprobes int kprobe_profile_func(struct kprobe *kp, struct trace_probe *tp = container_of(kp, struct trace_probe, rp.kp); struct ftrace_event_call *call = &tp->call; struct kprobe_trace_entry *entry; - struct trace_entry *ent; - int size, __size, i, pc, __cpu; + int size, __size, i; unsigned long irq_flags; - char *trace_buf; - char *raw_data; int rctx; - pc = preempt_count(); __size = SIZEOF_KPROBE_TRACE_ENTRY(tp->nr_args); size = ALIGN(__size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); @@ -1258,45 +1254,16 @@ static __kprobes int kprobe_profile_func(struct kprobe *kp, "profile buffer not large enough")) return 0; - /* - * Protect the non nmi buffer - * This also protects the rcu read side - */ - local_irq_save(irq_flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - __cpu = smp_processor_id(); - - if (in_nmi()) - trace_buf = rcu_dereference(perf_trace_buf_nmi); - else - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr(trace_buf, __cpu); - - /* Zero dead bytes from alignment to avoid buffer leak to userspace */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; - entry = (struct kprobe_trace_entry *)raw_data; - ent = &entry->ent; + entry = ftrace_perf_buf_prepare(size, call->id, &rctx, &irq_flags); + if (!entry) + return 0; - tracing_generic_entry_update(ent, irq_flags, pc); - ent->type = call->id; entry->nargs = tp->nr_args; entry->ip = (unsigned long)kp->addr; for (i = 0; i < tp->nr_args; i++) entry->args[i] = call_fetch(&tp->args[i].fetch, regs); - perf_tp_event(call->id, entry->ip, 1, entry, size); -end: - perf_swevent_put_recursion_context(rctx); -end_recursion: - local_irq_restore(irq_flags); + ftrace_perf_buf_submit(entry, size, rctx, entry->ip, 1, irq_flags); return 0; } @@ -1308,14 +1275,10 @@ static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, struct trace_probe *tp = container_of(ri->rp, struct trace_probe, rp); struct ftrace_event_call *call = &tp->call; struct kretprobe_trace_entry *entry; - struct trace_entry *ent; - int size, __size, i, pc, __cpu; + int size, __size, i; unsigned long irq_flags; - char *trace_buf; - char *raw_data; int rctx; - pc = preempt_count(); __size = SIZEOF_KRETPROBE_TRACE_ENTRY(tp->nr_args); size = ALIGN(__size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); @@ -1323,46 +1286,17 @@ static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, "profile buffer not large enough")) return 0; - /* - * Protect the non nmi buffer - * This also protects the rcu read side - */ - local_irq_save(irq_flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - __cpu = smp_processor_id(); - - if (in_nmi()) - trace_buf = rcu_dereference(perf_trace_buf_nmi); - else - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr(trace_buf, __cpu); - - /* Zero dead bytes from alignment to avoid buffer leak to userspace */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; - entry = (struct kretprobe_trace_entry *)raw_data; - ent = &entry->ent; + entry = ftrace_perf_buf_prepare(size, call->id, &rctx, &irq_flags); + if (!entry) + return 0; - tracing_generic_entry_update(ent, irq_flags, pc); - ent->type = call->id; entry->nargs = tp->nr_args; entry->func = (unsigned long)tp->rp.kp.addr; entry->ret_ip = (unsigned long)ri->ret_addr; for (i = 0; i < tp->nr_args; i++) entry->args[i] = call_fetch(&tp->args[i].fetch, regs); - perf_tp_event(call->id, entry->ret_ip, 1, entry, size); -end: - perf_swevent_put_recursion_context(rctx); -end_recursion: - local_irq_restore(irq_flags); + ftrace_perf_buf_submit(entry, size, rctx, entry->ret_ip, 1, irq_flags); return 0; } diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index f694f66..4e332b9 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -433,12 +433,9 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) struct syscall_metadata *sys_data; struct syscall_trace_enter *rec; unsigned long flags; - char *trace_buf; - char *raw_data; int syscall_nr; int rctx; int size; - int cpu; syscall_nr = syscall_get_nr(current, regs); if (!test_bit(syscall_nr, enabled_prof_enter_syscalls)) @@ -457,37 +454,15 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) "profile buffer not large enough")) return; - /* Protect the per cpu buffer, begin the rcu read side */ - local_irq_save(flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - cpu = smp_processor_id(); - - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr(trace_buf, cpu); - - /* zero the dead bytes from align to not leak stack to user */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; + rec = (struct syscall_trace_enter *)ftrace_perf_buf_prepare(size, + sys_data->enter_event->id, &rctx, &flags); + if (!rec) + return; - rec = (struct syscall_trace_enter *) raw_data; - tracing_generic_entry_update(&rec->ent, 0, 0); - rec->ent.type = sys_data->enter_event->id; rec->nr = syscall_nr; syscall_get_arguments(current, regs, 0, sys_data->nb_args, (unsigned long *)&rec->args); - perf_tp_event(sys_data->enter_event->id, 0, 1, rec, size); - -end: - perf_swevent_put_recursion_context(rctx); -end_recursion: - local_irq_restore(flags); + ftrace_perf_buf_submit(rec, size, rctx, 0, 1, flags); } int prof_sysenter_enable(struct ftrace_event_call *call) @@ -531,11 +506,8 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) struct syscall_trace_exit *rec; unsigned long flags; int syscall_nr; - char *trace_buf; - char *raw_data; int rctx; int size; - int cpu; syscall_nr = syscall_get_nr(current, regs); if (!test_bit(syscall_nr, enabled_prof_exit_syscalls)) @@ -557,38 +529,15 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) "exit event has grown above profile buffer size")) return; - /* Protect the per cpu buffer, begin the rcu read side */ - local_irq_save(flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - cpu = smp_processor_id(); - - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr(trace_buf, cpu); - - /* zero the dead bytes from align to not leak stack to user */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; - - rec = (struct syscall_trace_exit *)raw_data; + rec = (struct syscall_trace_exit *)ftrace_perf_buf_prepare(size, + sys_data->exit_event->id, &rctx, &flags); + if (!rec) + return; - tracing_generic_entry_update(&rec->ent, 0, 0); - rec->ent.type = sys_data->exit_event->id; rec->nr = syscall_nr; rec->ret = syscall_get_return_value(current, regs); - perf_tp_event(sys_data->exit_event->id, 0, 1, rec, size); - -end: - perf_swevent_put_recursion_context(rctx); -end_recursion: - local_irq_restore(flags); + ftrace_perf_buf_submit(rec, size, rctx, 0, 1, flags); } int prof_sysexit_enable(struct ftrace_event_call *call) -- cgit v0.10.2 From 1e12a4a7a3a78bc9c3aaf3486dde3b8ab1cdf465 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Thu, 28 Jan 2010 09:34:27 +0800 Subject: tracing/kprobe: Cleanup unused return value of tracing functions The return values of the kprobe's tracing functions are meaningless, lets remove these. Signed-off-by: Xiao Guangrong Acked-by: Masami Hiramatsu Cc: Steven Rostedt Cc: Ingo Molnar Cc: Paul Mackerras Cc: Jason Baron Cc: Peter Zijlstra LKML-Reference: <4B60E9A3.2040505@cn.fujitsu.com> [fweisbec@gmail: whitespace fixes, drop useless void returns in end of functions] Signed-off-by: Frederic Weisbecker diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 2e28ee3..6178abf 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -942,7 +942,7 @@ static const struct file_operations kprobe_profile_ops = { }; /* Kprobe handler */ -static __kprobes int kprobe_trace_func(struct kprobe *kp, struct pt_regs *regs) +static __kprobes void kprobe_trace_func(struct kprobe *kp, struct pt_regs *regs) { struct trace_probe *tp = container_of(kp, struct trace_probe, rp.kp); struct kprobe_trace_entry *entry; @@ -962,7 +962,7 @@ static __kprobes int kprobe_trace_func(struct kprobe *kp, struct pt_regs *regs) event = trace_current_buffer_lock_reserve(&buffer, call->id, size, irq_flags, pc); if (!event) - return 0; + return; entry = ring_buffer_event_data(event); entry->nargs = tp->nr_args; @@ -972,11 +972,10 @@ static __kprobes int kprobe_trace_func(struct kprobe *kp, struct pt_regs *regs) if (!filter_current_check_discard(buffer, call, entry, event)) trace_nowake_buffer_unlock_commit(buffer, event, irq_flags, pc); - return 0; } /* Kretprobe handler */ -static __kprobes int kretprobe_trace_func(struct kretprobe_instance *ri, +static __kprobes void kretprobe_trace_func(struct kretprobe_instance *ri, struct pt_regs *regs) { struct trace_probe *tp = container_of(ri->rp, struct trace_probe, rp); @@ -995,7 +994,7 @@ static __kprobes int kretprobe_trace_func(struct kretprobe_instance *ri, event = trace_current_buffer_lock_reserve(&buffer, call->id, size, irq_flags, pc); if (!event) - return 0; + return; entry = ring_buffer_event_data(event); entry->nargs = tp->nr_args; @@ -1006,8 +1005,6 @@ static __kprobes int kretprobe_trace_func(struct kretprobe_instance *ri, if (!filter_current_check_discard(buffer, call, entry, event)) trace_nowake_buffer_unlock_commit(buffer, event, irq_flags, pc); - - return 0; } /* Event entry printers */ @@ -1237,7 +1234,7 @@ static int kretprobe_event_show_format(struct ftrace_event_call *call, #ifdef CONFIG_PERF_EVENTS /* Kprobe profile handler */ -static __kprobes int kprobe_profile_func(struct kprobe *kp, +static __kprobes void kprobe_profile_func(struct kprobe *kp, struct pt_regs *regs) { struct trace_probe *tp = container_of(kp, struct trace_probe, rp.kp); @@ -1252,11 +1249,11 @@ static __kprobes int kprobe_profile_func(struct kprobe *kp, size -= sizeof(u32); if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE, "profile buffer not large enough")) - return 0; + return; entry = ftrace_perf_buf_prepare(size, call->id, &rctx, &irq_flags); if (!entry) - return 0; + return; entry->nargs = tp->nr_args; entry->ip = (unsigned long)kp->addr; @@ -1264,12 +1261,10 @@ static __kprobes int kprobe_profile_func(struct kprobe *kp, entry->args[i] = call_fetch(&tp->args[i].fetch, regs); ftrace_perf_buf_submit(entry, size, rctx, entry->ip, 1, irq_flags); - - return 0; } /* Kretprobe profile handler */ -static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, +static __kprobes void kretprobe_profile_func(struct kretprobe_instance *ri, struct pt_regs *regs) { struct trace_probe *tp = container_of(ri->rp, struct trace_probe, rp); @@ -1284,11 +1279,11 @@ static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, size -= sizeof(u32); if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE, "profile buffer not large enough")) - return 0; + return; entry = ftrace_perf_buf_prepare(size, call->id, &rctx, &irq_flags); if (!entry) - return 0; + return; entry->nargs = tp->nr_args; entry->func = (unsigned long)tp->rp.kp.addr; @@ -1297,8 +1292,6 @@ static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, entry->args[i] = call_fetch(&tp->args[i].fetch, regs); ftrace_perf_buf_submit(entry, size, rctx, entry->ret_ip, 1, irq_flags); - - return 0; } static int probe_profile_enable(struct ftrace_event_call *call) -- cgit v0.10.2 From 40f9249a73f6c251adea492b1c3d19d39e2a9bda Mon Sep 17 00:00:00 2001 From: "K.Prasad" Date: Thu, 28 Jan 2010 16:44:01 +0530 Subject: x86/debug: Clear reserved bits of DR6 in do_debug() Clear the reserved bits from the stored copy of debug status register (DR6). This will help easy bitwise operations such as quick testing of a debug event origin. Signed-off-by: K.Prasad Cc: Roland McGrath Cc: Jan Kiszka Cc: Alan Stern Cc: Ingo Molnar LKML-Reference: <20100128111401.GB13935@in.ibm.com> Signed-off-by: Frederic Weisbecker diff --git a/arch/x86/include/asm/debugreg.h b/arch/x86/include/asm/debugreg.h index 8240f76..b81002f 100644 --- a/arch/x86/include/asm/debugreg.h +++ b/arch/x86/include/asm/debugreg.h @@ -14,6 +14,9 @@ which debugging register was responsible for the trap. The other bits are either reserved or not of interest to us. */ +/* Define reserved bits in DR6 which are always set to 1 */ +#define DR6_RESERVED (0xFFFF0FF0) + #define DR_TRAP0 (0x1) /* db0 */ #define DR_TRAP1 (0x2) /* db1 */ #define DR_TRAP2 (0x4) /* db2 */ diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index 3339917..1168e44 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -534,6 +534,9 @@ dotraplinkage void __kprobes do_debug(struct pt_regs *regs, long error_code) get_debugreg(dr6, 6); + /* Filter out all the reserved bits which are preset to 1 */ + dr6 &= ~DR6_RESERVED; + /* Catch kmemcheck conditions first of all! */ if ((dr6 & DR_STEP) && kmemcheck_trap(regs)) return; -- cgit v0.10.2 From e0e53db6133c32964fd17f20b17073a402f07ed3 Mon Sep 17 00:00:00 2001 From: "K.Prasad" Date: Thu, 28 Jan 2010 16:44:15 +0530 Subject: x86/hw-breakpoints: Optimize return code from notifier chain in hw_breakpoint_handler Processing of debug exceptions in do_debug() can stop if it originated from a hw-breakpoint exception by returning NOTIFY_STOP in most cases. But for certain cases such as: a) user-space breakpoints with pending SIGTRAP signal delivery (as in the case of ptrace induced breakpoints). b) exceptions due to other causes than breakpoints We will continue to process the exception by returning NOTIFY_DONE. Signed-off-by: K.Prasad Cc: Ingo Molnar Cc: Roland McGrath Cc: Alan Stern Cc: Jan Kiszka LKML-Reference: <20100128111415.GC13935@in.ibm.com> Signed-off-by: Frederic Weisbecker diff --git a/arch/x86/kernel/hw_breakpoint.c b/arch/x86/kernel/hw_breakpoint.c index 05d5fec..ae90b47 100644 --- a/arch/x86/kernel/hw_breakpoint.c +++ b/arch/x86/kernel/hw_breakpoint.c @@ -502,8 +502,6 @@ static int __kprobes hw_breakpoint_handler(struct die_args *args) rcu_read_lock(); bp = per_cpu(bp_per_reg[i], cpu); - if (bp) - rc = NOTIFY_DONE; /* * Reset the 'i'th TRAP bit in dr6 to denote completion of * exception handling @@ -522,7 +520,13 @@ static int __kprobes hw_breakpoint_handler(struct die_args *args) rcu_read_unlock(); } - if (dr6 & (~DR_TRAP_BITS)) + /* + * Further processing in do_debug() is needed for a) user-space + * breakpoints (to generate signals) and b) when the system has + * taken exception due to multiple causes + */ + if ((current->thread.debugreg6 & DR_TRAP_BITS) || + (dr6 & (~DR_TRAP_BITS))) rc = NOTIFY_DONE; set_debugreg(dr7, 7); -- cgit v0.10.2 From 9762528f37ddc7071509dddb10e7b4b3b957fd01 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 29 Jan 2010 16:14:29 +0900 Subject: sh: Kill off deprecated fixed PCI memory window accessors. This kills off the deprected fixed memory range accessors for the cases of non-translatable ioremapping. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/pci.h b/arch/sh/include/asm/pci.h index 67f3999..f362d8a 100644 --- a/arch/sh/include/asm/pci.h +++ b/arch/sh/include/asm/pci.h @@ -99,20 +99,6 @@ static inline void pci_dma_burst_advice(struct pci_dev *pdev, } #endif -#ifdef CONFIG_SUPERH32 -/* - * If we're on an SH7751 or SH7780 PCI controller, PCI memory is mapped - * at the end of the address space in a special non-translatable area. - */ -#define PCI_MEM_FIXED_START 0xfd000000 -#define PCI_MEM_FIXED_END (PCI_MEM_FIXED_START + 0x01000000) - -#define is_pci_memory_fixed_range(s, e) \ - ((s) >= PCI_MEM_FIXED_START && (e) < PCI_MEM_FIXED_END) -#else -#define is_pci_memory_fixed_range(s, e) (0) -#endif - /* Board-specific fixup routines. */ int pcibios_map_platform_irq(struct pci_dev *dev, u8 slot, u8 pin); diff --git a/arch/sh/mm/ioremap.c b/arch/sh/mm/ioremap.c index bb03308..94583c5 100644 --- a/arch/sh/mm/ioremap.c +++ b/arch/sh/mm/ioremap.c @@ -46,18 +46,6 @@ __ioremap_caller(unsigned long phys_addr, unsigned long size, return NULL; /* - * If we're in the fixed PCI memory range, mapping through page - * tables is not only pointless, but also fundamentally broken. - * Just return the physical address instead. - * - * For boards that map a small PCI memory aperture somewhere in - * P1/P2 space, ioremap() will already do the right thing, - * and we'll never get this far. - */ - if (is_pci_memory_fixed_range(phys_addr, size)) - return (void __iomem *)phys_addr; - - /* * Mappings have to be page-aligned */ offset = phys_addr & ~PAGE_MASK; @@ -125,9 +113,6 @@ static inline int iomapping_nontranslatable(unsigned long offset) return 1; #endif - if (is_pci_memory_fixed_range(offset, 0)) - return 1; - return 0; } -- cgit v0.10.2 From 1da53e023029c067ba1277a33038c65d6e4c99b3 Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Mon, 18 Jan 2010 10:58:01 +0200 Subject: perf_events, x86: Improve x86 event scheduling This patch improves event scheduling by maximizing the use of PMU registers regardless of the order in which events are created in a group. The algorithm takes into account the list of counter constraints for each event. It assigns events to counters from the most constrained, i.e., works on only one counter, to the least constrained, i.e., works on any counter. Intel Fixed counter events and the BTS special event are also handled via this algorithm which is designed to be fairly generic. The patch also updates the validation of an event to use the scheduling algorithm. This will cause early failure in perf_event_open(). The 2nd version of this patch follows the model used by PPC, by running the scheduling algorithm and the actual assignment separately. Actual assignment takes place in hw_perf_enable() whereas scheduling is implemented in hw_perf_group_sched_in() and x86_pmu_enable(). Signed-off-by: Stephane Eranian [ fixup whitespace and style nits as well as adding is_x86_event() ] Signed-off-by: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <4b5430c6.0f975e0a.1bf9.ffff85fe@mx.google.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h index 8d9f854..dbc0826 100644 --- a/arch/x86/include/asm/perf_event.h +++ b/arch/x86/include/asm/perf_event.h @@ -26,7 +26,14 @@ /* * Includes eventsel and unit mask as well: */ -#define ARCH_PERFMON_EVENT_MASK 0xffff + + +#define INTEL_ARCH_EVTSEL_MASK 0x000000FFULL +#define INTEL_ARCH_UNIT_MASK 0x0000FF00ULL +#define INTEL_ARCH_EDGE_MASK 0x00040000ULL +#define INTEL_ARCH_INV_MASK 0x00800000ULL +#define INTEL_ARCH_CNT_MASK 0xFF000000ULL +#define INTEL_ARCH_EVENT_MASK (INTEL_ARCH_UNIT_MASK|INTEL_ARCH_EVTSEL_MASK) /* * filter mask to validate fixed counter events. @@ -37,7 +44,12 @@ * The other filters are supported by fixed counters. * The any-thread option is supported starting with v3. */ -#define ARCH_PERFMON_EVENT_FILTER_MASK 0xff840000 +#define INTEL_ARCH_FIXED_MASK \ + (INTEL_ARCH_CNT_MASK| \ + INTEL_ARCH_INV_MASK| \ + INTEL_ARCH_EDGE_MASK|\ + INTEL_ARCH_UNIT_MASK|\ + INTEL_ARCH_EVENT_MASK) #define ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL 0x3c #define ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK (0x00 << 8) diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index ed1998b..995ac4a 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -7,6 +7,7 @@ * Copyright (C) 2009 Advanced Micro Devices, Inc., Robert Richter * Copyright (C) 2008-2009 Red Hat, Inc., Peter Zijlstra * Copyright (C) 2009 Intel Corporation, + * Copyright (C) 2009 Google, Inc., Stephane Eranian * * For licencing details see kernel-base/COPYING */ @@ -68,26 +69,37 @@ struct debug_store { u64 pebs_event_reset[MAX_PEBS_EVENTS]; }; +#define BITS_TO_U64(nr) DIV_ROUND_UP(nr, BITS_PER_BYTE * sizeof(u64)) + +struct event_constraint { + u64 idxmsk[BITS_TO_U64(X86_PMC_IDX_MAX)]; + int code; + int cmask; +}; + struct cpu_hw_events { - struct perf_event *events[X86_PMC_IDX_MAX]; - unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; + struct perf_event *events[X86_PMC_IDX_MAX]; /* in counter order */ unsigned long active_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; unsigned long interrupts; int enabled; struct debug_store *ds; -}; -struct event_constraint { - unsigned long idxmsk[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; - int code; + int n_events; + int n_added; + int assign[X86_PMC_IDX_MAX]; /* event to counter assignment */ + struct perf_event *event_list[X86_PMC_IDX_MAX]; /* in enabled order */ }; -#define EVENT_CONSTRAINT(c, m) { .code = (c), .idxmsk[0] = (m) } -#define EVENT_CONSTRAINT_END { .code = 0, .idxmsk[0] = 0 } +#define EVENT_CONSTRAINT(c, n, m) { \ + .code = (c), \ + .cmask = (m), \ + .idxmsk[0] = (n) } -#define for_each_event_constraint(e, c) \ - for ((e) = (c); (e)->idxmsk[0]; (e)++) +#define EVENT_CONSTRAINT_END \ + { .code = 0, .cmask = 0, .idxmsk[0] = 0 } +#define for_each_event_constraint(e, c) \ + for ((e) = (c); (e)->cmask; (e)++) /* * struct x86_pmu - generic x86 pmu @@ -114,8 +126,9 @@ struct x86_pmu { u64 intel_ctrl; void (*enable_bts)(u64 config); void (*disable_bts)(void); - int (*get_event_idx)(struct cpu_hw_events *cpuc, - struct hw_perf_event *hwc); + void (*get_event_constraints)(struct cpu_hw_events *cpuc, struct perf_event *event, u64 *idxmsk); + void (*put_event_constraints)(struct cpu_hw_events *cpuc, struct perf_event *event); + const struct event_constraint *event_constraints; }; static struct x86_pmu x86_pmu __read_mostly; @@ -124,7 +137,8 @@ static DEFINE_PER_CPU(struct cpu_hw_events, cpu_hw_events) = { .enabled = 1, }; -static const struct event_constraint *event_constraints; +static int x86_perf_event_set_period(struct perf_event *event, + struct hw_perf_event *hwc, int idx); /* * Not sure about some of these @@ -171,14 +185,14 @@ static u64 p6_pmu_raw_event(u64 hw_event) return hw_event & P6_EVNTSEL_MASK; } -static const struct event_constraint intel_p6_event_constraints[] = +static struct event_constraint intel_p6_event_constraints[] = { - EVENT_CONSTRAINT(0xc1, 0x1), /* FLOPS */ - EVENT_CONSTRAINT(0x10, 0x1), /* FP_COMP_OPS_EXE */ - EVENT_CONSTRAINT(0x11, 0x1), /* FP_ASSIST */ - EVENT_CONSTRAINT(0x12, 0x2), /* MUL */ - EVENT_CONSTRAINT(0x13, 0x2), /* DIV */ - EVENT_CONSTRAINT(0x14, 0x1), /* CYCLES_DIV_BUSY */ + EVENT_CONSTRAINT(0xc1, 0x1, INTEL_ARCH_EVENT_MASK), /* FLOPS */ + EVENT_CONSTRAINT(0x10, 0x1, INTEL_ARCH_EVENT_MASK), /* FP_COMP_OPS_EXE */ + EVENT_CONSTRAINT(0x11, 0x1, INTEL_ARCH_EVENT_MASK), /* FP_ASSIST */ + EVENT_CONSTRAINT(0x12, 0x2, INTEL_ARCH_EVENT_MASK), /* MUL */ + EVENT_CONSTRAINT(0x13, 0x2, INTEL_ARCH_EVENT_MASK), /* DIV */ + EVENT_CONSTRAINT(0x14, 0x1, INTEL_ARCH_EVENT_MASK), /* CYCLES_DIV_BUSY */ EVENT_CONSTRAINT_END }; @@ -196,32 +210,43 @@ static const u64 intel_perfmon_event_map[] = [PERF_COUNT_HW_BUS_CYCLES] = 0x013c, }; -static const struct event_constraint intel_core_event_constraints[] = -{ - EVENT_CONSTRAINT(0x10, 0x1), /* FP_COMP_OPS_EXE */ - EVENT_CONSTRAINT(0x11, 0x2), /* FP_ASSIST */ - EVENT_CONSTRAINT(0x12, 0x2), /* MUL */ - EVENT_CONSTRAINT(0x13, 0x2), /* DIV */ - EVENT_CONSTRAINT(0x14, 0x1), /* CYCLES_DIV_BUSY */ - EVENT_CONSTRAINT(0x18, 0x1), /* IDLE_DURING_DIV */ - EVENT_CONSTRAINT(0x19, 0x2), /* DELAYED_BYPASS */ - EVENT_CONSTRAINT(0xa1, 0x1), /* RS_UOPS_DISPATCH_CYCLES */ - EVENT_CONSTRAINT(0xcb, 0x1), /* MEM_LOAD_RETIRED */ +static struct event_constraint intel_core_event_constraints[] = +{ + EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32)), INTEL_ARCH_FIXED_MASK), /* INSTRUCTIONS_RETIRED */ + EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33)), INTEL_ARCH_FIXED_MASK), /* UNHALTED_CORE_CYCLES */ + EVENT_CONSTRAINT(0x10, 0x1, INTEL_ARCH_EVENT_MASK), /* FP_COMP_OPS_EXE */ + EVENT_CONSTRAINT(0x11, 0x2, INTEL_ARCH_EVENT_MASK), /* FP_ASSIST */ + EVENT_CONSTRAINT(0x12, 0x2, INTEL_ARCH_EVENT_MASK), /* MUL */ + EVENT_CONSTRAINT(0x13, 0x2, INTEL_ARCH_EVENT_MASK), /* DIV */ + EVENT_CONSTRAINT(0x14, 0x1, INTEL_ARCH_EVENT_MASK), /* CYCLES_DIV_BUSY */ + EVENT_CONSTRAINT(0x18, 0x1, INTEL_ARCH_EVENT_MASK), /* IDLE_DURING_DIV */ + EVENT_CONSTRAINT(0x19, 0x2, INTEL_ARCH_EVENT_MASK), /* DELAYED_BYPASS */ + EVENT_CONSTRAINT(0xa1, 0x1, INTEL_ARCH_EVENT_MASK), /* RS_UOPS_DISPATCH_CYCLES */ + EVENT_CONSTRAINT(0xcb, 0x1, INTEL_ARCH_EVENT_MASK), /* MEM_LOAD_RETIRED */ EVENT_CONSTRAINT_END }; -static const struct event_constraint intel_nehalem_event_constraints[] = -{ - EVENT_CONSTRAINT(0x40, 0x3), /* L1D_CACHE_LD */ - EVENT_CONSTRAINT(0x41, 0x3), /* L1D_CACHE_ST */ - EVENT_CONSTRAINT(0x42, 0x3), /* L1D_CACHE_LOCK */ - EVENT_CONSTRAINT(0x43, 0x3), /* L1D_ALL_REF */ - EVENT_CONSTRAINT(0x4e, 0x3), /* L1D_PREFETCH */ - EVENT_CONSTRAINT(0x4c, 0x3), /* LOAD_HIT_PRE */ - EVENT_CONSTRAINT(0x51, 0x3), /* L1D */ - EVENT_CONSTRAINT(0x52, 0x3), /* L1D_CACHE_PREFETCH_LOCK_FB_HIT */ - EVENT_CONSTRAINT(0x53, 0x3), /* L1D_CACHE_LOCK_FB_HIT */ - EVENT_CONSTRAINT(0xc5, 0x3), /* CACHE_LOCK_CYCLES */ +static struct event_constraint intel_nehalem_event_constraints[] = +{ + EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32)), INTEL_ARCH_FIXED_MASK), /* INSTRUCTIONS_RETIRED */ + EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33)), INTEL_ARCH_FIXED_MASK), /* UNHALTED_CORE_CYCLES */ + EVENT_CONSTRAINT(0x40, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_CACHE_LD */ + EVENT_CONSTRAINT(0x41, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_CACHE_ST */ + EVENT_CONSTRAINT(0x42, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_CACHE_LOCK */ + EVENT_CONSTRAINT(0x43, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_ALL_REF */ + EVENT_CONSTRAINT(0x4e, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_PREFETCH */ + EVENT_CONSTRAINT(0x4c, 0x3, INTEL_ARCH_EVENT_MASK), /* LOAD_HIT_PRE */ + EVENT_CONSTRAINT(0x51, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D */ + EVENT_CONSTRAINT(0x52, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_CACHE_PREFETCH_LOCK_FB_HIT */ + EVENT_CONSTRAINT(0x53, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_CACHE_LOCK_FB_HIT */ + EVENT_CONSTRAINT(0xc5, 0x3, INTEL_ARCH_EVENT_MASK), /* CACHE_LOCK_CYCLES */ + EVENT_CONSTRAINT_END +}; + +static struct event_constraint intel_gen_event_constraints[] = +{ + EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32)), INTEL_ARCH_FIXED_MASK), /* INSTRUCTIONS_RETIRED */ + EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33)), INTEL_ARCH_FIXED_MASK), /* UNHALTED_CORE_CYCLES */ EVENT_CONSTRAINT_END }; @@ -527,11 +552,11 @@ static u64 intel_pmu_raw_event(u64 hw_event) #define CORE_EVNTSEL_REG_MASK 0xFF000000ULL #define CORE_EVNTSEL_MASK \ - (CORE_EVNTSEL_EVENT_MASK | \ - CORE_EVNTSEL_UNIT_MASK | \ - CORE_EVNTSEL_EDGE_MASK | \ - CORE_EVNTSEL_INV_MASK | \ - CORE_EVNTSEL_REG_MASK) + (INTEL_ARCH_EVTSEL_MASK | \ + INTEL_ARCH_UNIT_MASK | \ + INTEL_ARCH_EDGE_MASK | \ + INTEL_ARCH_INV_MASK | \ + INTEL_ARCH_CNT_MASK) return hw_event & CORE_EVNTSEL_MASK; } @@ -1120,9 +1145,15 @@ static void amd_pmu_disable_all(void) void hw_perf_disable(void) { + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + if (!x86_pmu_initialized()) return; - return x86_pmu.disable_all(); + + if (cpuc->enabled) + cpuc->n_added = 0; + + x86_pmu.disable_all(); } static void p6_pmu_enable_all(void) @@ -1189,10 +1220,237 @@ static void amd_pmu_enable_all(void) } } +static const struct pmu pmu; + +static inline int is_x86_event(struct perf_event *event) +{ + return event->pmu == &pmu; +} + +static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) +{ + int i, j , w, num; + int weight, wmax; + unsigned long *c; + u64 constraints[X86_PMC_IDX_MAX][BITS_TO_LONGS(X86_PMC_IDX_MAX)]; + unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; + struct hw_perf_event *hwc; + + bitmap_zero(used_mask, X86_PMC_IDX_MAX); + + for (i = 0; i < n; i++) { + x86_pmu.get_event_constraints(cpuc, + cpuc->event_list[i], + constraints[i]); + } + + /* + * weight = number of possible counters + * + * 1 = most constrained, only works on one counter + * wmax = least constrained, works on any counter + * + * assign events to counters starting with most + * constrained events. + */ + wmax = x86_pmu.num_events; + + /* + * when fixed event counters are present, + * wmax is incremented by 1 to account + * for one more choice + */ + if (x86_pmu.num_events_fixed) + wmax++; + + num = n; + for (w = 1; num && w <= wmax; w++) { + /* for each event */ + for (i = 0; i < n; i++) { + c = (unsigned long *)constraints[i]; + hwc = &cpuc->event_list[i]->hw; + + weight = bitmap_weight(c, X86_PMC_IDX_MAX); + if (weight != w) + continue; + + /* + * try to reuse previous assignment + * + * This is possible despite the fact that + * events or events order may have changed. + * + * What matters is the level of constraints + * of an event and this is constant for now. + * + * This is possible also because we always + * scan from most to least constrained. Thus, + * if a counter can be reused, it means no, + * more constrained events, needed it. And + * next events will either compete for it + * (which cannot be solved anyway) or they + * have fewer constraints, and they can use + * another counter. + */ + j = hwc->idx; + if (j != -1 && !test_bit(j, used_mask)) + goto skip; + + for_each_bit(j, c, X86_PMC_IDX_MAX) { + if (!test_bit(j, used_mask)) + break; + } + + if (j == X86_PMC_IDX_MAX) + break; +skip: + set_bit(j, used_mask); + +#if 0 + pr_debug("CPU%d config=0x%llx idx=%d assign=%c\n", + smp_processor_id(), + hwc->config, + j, + assign ? 'y' : 'n'); +#endif + + if (assign) + assign[i] = j; + num--; + } + } + /* + * scheduling failed or is just a simulation, + * free resources if necessary + */ + if (!assign || num) { + for (i = 0; i < n; i++) { + if (x86_pmu.put_event_constraints) + x86_pmu.put_event_constraints(cpuc, cpuc->event_list[i]); + } + } + return num ? -ENOSPC : 0; +} + +/* + * dogrp: true if must collect siblings events (group) + * returns total number of events and error code + */ +static int collect_events(struct cpu_hw_events *cpuc, struct perf_event *leader, bool dogrp) +{ + struct perf_event *event; + int n, max_count; + + max_count = x86_pmu.num_events + x86_pmu.num_events_fixed; + + /* current number of events already accepted */ + n = cpuc->n_events; + + if (is_x86_event(leader)) { + if (n >= max_count) + return -ENOSPC; + cpuc->event_list[n] = leader; + n++; + } + if (!dogrp) + return n; + + list_for_each_entry(event, &leader->sibling_list, group_entry) { + if (!is_x86_event(event) || + event->state == PERF_EVENT_STATE_OFF) + continue; + + if (n >= max_count) + return -ENOSPC; + + cpuc->event_list[n] = event; + n++; + } + return n; +} + + +static inline void x86_assign_hw_event(struct perf_event *event, + struct hw_perf_event *hwc, int idx) +{ + hwc->idx = idx; + + if (hwc->idx == X86_PMC_IDX_FIXED_BTS) { + hwc->config_base = 0; + hwc->event_base = 0; + } else if (hwc->idx >= X86_PMC_IDX_FIXED) { + hwc->config_base = MSR_ARCH_PERFMON_FIXED_CTR_CTRL; + /* + * We set it so that event_base + idx in wrmsr/rdmsr maps to + * MSR_ARCH_PERFMON_FIXED_CTR0 ... CTR2: + */ + hwc->event_base = + MSR_ARCH_PERFMON_FIXED_CTR0 - X86_PMC_IDX_FIXED; + } else { + hwc->config_base = x86_pmu.eventsel; + hwc->event_base = x86_pmu.perfctr; + } +} + void hw_perf_enable(void) { + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct perf_event *event; + struct hw_perf_event *hwc; + int i; + if (!x86_pmu_initialized()) return; + if (cpuc->n_added) { + /* + * apply assignment obtained either from + * hw_perf_group_sched_in() or x86_pmu_enable() + * + * step1: save events moving to new counters + * step2: reprogram moved events into new counters + */ + for (i = 0; i < cpuc->n_events; i++) { + + event = cpuc->event_list[i]; + hwc = &event->hw; + + if (hwc->idx == -1 || hwc->idx == cpuc->assign[i]) + continue; + + x86_pmu.disable(hwc, hwc->idx); + + clear_bit(hwc->idx, cpuc->active_mask); + barrier(); + cpuc->events[hwc->idx] = NULL; + + x86_perf_event_update(event, hwc, hwc->idx); + + hwc->idx = -1; + } + + for (i = 0; i < cpuc->n_events; i++) { + + event = cpuc->event_list[i]; + hwc = &event->hw; + + if (hwc->idx == -1) { + x86_assign_hw_event(event, hwc, cpuc->assign[i]); + x86_perf_event_set_period(event, hwc, hwc->idx); + } + /* + * need to mark as active because x86_pmu_disable() + * clear active_mask and eventsp[] yet it preserves + * idx + */ + set_bit(hwc->idx, cpuc->active_mask); + cpuc->events[hwc->idx] = event; + + x86_pmu.enable(hwc, hwc->idx); + perf_event_update_userpage(event); + } + cpuc->n_added = 0; + perf_events_lapic_init(); + } x86_pmu.enable_all(); } @@ -1391,148 +1649,43 @@ static void amd_pmu_enable_event(struct hw_perf_event *hwc, int idx) x86_pmu_enable_event(hwc, idx); } -static int fixed_mode_idx(struct hw_perf_event *hwc) -{ - unsigned int hw_event; - - hw_event = hwc->config & ARCH_PERFMON_EVENT_MASK; - - if (unlikely((hw_event == - x86_pmu.event_map(PERF_COUNT_HW_BRANCH_INSTRUCTIONS)) && - (hwc->sample_period == 1))) - return X86_PMC_IDX_FIXED_BTS; - - if (!x86_pmu.num_events_fixed) - return -1; - - /* - * fixed counters do not take all possible filters - */ - if (hwc->config & ARCH_PERFMON_EVENT_FILTER_MASK) - return -1; - - if (unlikely(hw_event == x86_pmu.event_map(PERF_COUNT_HW_INSTRUCTIONS))) - return X86_PMC_IDX_FIXED_INSTRUCTIONS; - if (unlikely(hw_event == x86_pmu.event_map(PERF_COUNT_HW_CPU_CYCLES))) - return X86_PMC_IDX_FIXED_CPU_CYCLES; - if (unlikely(hw_event == x86_pmu.event_map(PERF_COUNT_HW_BUS_CYCLES))) - return X86_PMC_IDX_FIXED_BUS_CYCLES; - - return -1; -} - -/* - * generic counter allocator: get next free counter - */ -static int -gen_get_event_idx(struct cpu_hw_events *cpuc, struct hw_perf_event *hwc) -{ - int idx; - - idx = find_first_zero_bit(cpuc->used_mask, x86_pmu.num_events); - return idx == x86_pmu.num_events ? -1 : idx; -} - /* - * intel-specific counter allocator: check event constraints - */ -static int -intel_get_event_idx(struct cpu_hw_events *cpuc, struct hw_perf_event *hwc) -{ - const struct event_constraint *event_constraint; - int i, code; - - if (!event_constraints) - goto skip; - - code = hwc->config & CORE_EVNTSEL_EVENT_MASK; - - for_each_event_constraint(event_constraint, event_constraints) { - if (code == event_constraint->code) { - for_each_bit(i, event_constraint->idxmsk, X86_PMC_IDX_MAX) { - if (!test_and_set_bit(i, cpuc->used_mask)) - return i; - } - return -1; - } - } -skip: - return gen_get_event_idx(cpuc, hwc); -} - -static int -x86_schedule_event(struct cpu_hw_events *cpuc, struct hw_perf_event *hwc) -{ - int idx; - - idx = fixed_mode_idx(hwc); - if (idx == X86_PMC_IDX_FIXED_BTS) { - /* BTS is already occupied. */ - if (test_and_set_bit(idx, cpuc->used_mask)) - return -EAGAIN; - - hwc->config_base = 0; - hwc->event_base = 0; - hwc->idx = idx; - } else if (idx >= 0) { - /* - * Try to get the fixed event, if that is already taken - * then try to get a generic event: - */ - if (test_and_set_bit(idx, cpuc->used_mask)) - goto try_generic; - - hwc->config_base = MSR_ARCH_PERFMON_FIXED_CTR_CTRL; - /* - * We set it so that event_base + idx in wrmsr/rdmsr maps to - * MSR_ARCH_PERFMON_FIXED_CTR0 ... CTR2: - */ - hwc->event_base = - MSR_ARCH_PERFMON_FIXED_CTR0 - X86_PMC_IDX_FIXED; - hwc->idx = idx; - } else { - idx = hwc->idx; - /* Try to get the previous generic event again */ - if (idx == -1 || test_and_set_bit(idx, cpuc->used_mask)) { -try_generic: - idx = x86_pmu.get_event_idx(cpuc, hwc); - if (idx == -1) - return -EAGAIN; - - set_bit(idx, cpuc->used_mask); - hwc->idx = idx; - } - hwc->config_base = x86_pmu.eventsel; - hwc->event_base = x86_pmu.perfctr; - } - - return idx; -} - -/* - * Find a PMC slot for the freshly enabled / scheduled in event: + * activate a single event + * + * The event is added to the group of enabled events + * but only if it can be scehduled with existing events. + * + * Called with PMU disabled. If successful and return value 1, + * then guaranteed to call perf_enable() and hw_perf_enable() */ static int x86_pmu_enable(struct perf_event *event) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - struct hw_perf_event *hwc = &event->hw; - int idx; + struct hw_perf_event *hwc; + int assign[X86_PMC_IDX_MAX]; + int n, n0, ret; - idx = x86_schedule_event(cpuc, hwc); - if (idx < 0) - return idx; + hwc = &event->hw; - perf_events_lapic_init(); + n0 = cpuc->n_events; + n = collect_events(cpuc, event, false); + if (n < 0) + return n; - x86_pmu.disable(hwc, idx); - - cpuc->events[idx] = event; - set_bit(idx, cpuc->active_mask); + ret = x86_schedule_events(cpuc, n, assign); + if (ret) + return ret; + /* + * copy new assignment, now we know it is possible + * will be used by hw_perf_enable() + */ + memcpy(cpuc->assign, assign, n*sizeof(int)); - x86_perf_event_set_period(event, hwc, idx); - x86_pmu.enable(hwc, idx); + cpuc->n_events = n; + cpuc->n_added = n - n0; - perf_event_update_userpage(event); + if (hwc->idx != -1) + x86_perf_event_set_period(event, hwc, hwc->idx); return 0; } @@ -1576,7 +1729,7 @@ void perf_event_print_debug(void) pr_info("CPU#%d: overflow: %016llx\n", cpu, overflow); pr_info("CPU#%d: fixed: %016llx\n", cpu, fixed); } - pr_info("CPU#%d: used: %016llx\n", cpu, *(u64 *)cpuc->used_mask); + pr_info("CPU#%d: active: %016llx\n", cpu, *(u64 *)cpuc->active_mask); for (idx = 0; idx < x86_pmu.num_events; idx++) { rdmsrl(x86_pmu.eventsel + idx, pmc_ctrl); @@ -1664,7 +1817,7 @@ static void x86_pmu_disable(struct perf_event *event) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct hw_perf_event *hwc = &event->hw; - int idx = hwc->idx; + int i, idx = hwc->idx; /* * Must be done before we disable, otherwise the nmi handler @@ -1690,8 +1843,19 @@ static void x86_pmu_disable(struct perf_event *event) intel_pmu_drain_bts_buffer(cpuc); cpuc->events[idx] = NULL; - clear_bit(idx, cpuc->used_mask); + for (i = 0; i < cpuc->n_events; i++) { + if (event == cpuc->event_list[i]) { + + if (x86_pmu.put_event_constraints) + x86_pmu.put_event_constraints(cpuc, event); + + while (++i < cpuc->n_events) + cpuc->event_list[i-1] = cpuc->event_list[i]; + + --cpuc->n_events; + } + } perf_event_update_userpage(event); } @@ -1962,6 +2126,176 @@ perf_event_nmi_handler(struct notifier_block *self, return NOTIFY_STOP; } +static struct event_constraint bts_constraint = { + .code = 0, + .cmask = 0, + .idxmsk[0] = 1ULL << X86_PMC_IDX_FIXED_BTS +}; + +static int intel_special_constraints(struct perf_event *event, + u64 *idxmsk) +{ + unsigned int hw_event; + + hw_event = event->hw.config & INTEL_ARCH_EVENT_MASK; + + if (unlikely((hw_event == + x86_pmu.event_map(PERF_COUNT_HW_BRANCH_INSTRUCTIONS)) && + (event->hw.sample_period == 1))) { + + bitmap_copy((unsigned long *)idxmsk, + (unsigned long *)bts_constraint.idxmsk, + X86_PMC_IDX_MAX); + return 1; + } + return 0; +} + +static void intel_get_event_constraints(struct cpu_hw_events *cpuc, + struct perf_event *event, + u64 *idxmsk) +{ + const struct event_constraint *c; + + /* + * cleanup bitmask + */ + bitmap_zero((unsigned long *)idxmsk, X86_PMC_IDX_MAX); + + if (intel_special_constraints(event, idxmsk)) + return; + + if (x86_pmu.event_constraints) { + for_each_event_constraint(c, x86_pmu.event_constraints) { + if ((event->hw.config & c->cmask) == c->code) { + + bitmap_copy((unsigned long *)idxmsk, + (unsigned long *)c->idxmsk, + X86_PMC_IDX_MAX); + return; + } + } + } + /* no constraints, means supports all generic counters */ + bitmap_fill((unsigned long *)idxmsk, x86_pmu.num_events); +} + +static void amd_get_event_constraints(struct cpu_hw_events *cpuc, + struct perf_event *event, + u64 *idxmsk) +{ +} + +static int x86_event_sched_in(struct perf_event *event, + struct perf_cpu_context *cpuctx, int cpu) +{ + int ret = 0; + + event->state = PERF_EVENT_STATE_ACTIVE; + event->oncpu = cpu; + event->tstamp_running += event->ctx->time - event->tstamp_stopped; + + if (!is_x86_event(event)) + ret = event->pmu->enable(event); + + if (!ret && !is_software_event(event)) + cpuctx->active_oncpu++; + + if (!ret && event->attr.exclusive) + cpuctx->exclusive = 1; + + return ret; +} + +static void x86_event_sched_out(struct perf_event *event, + struct perf_cpu_context *cpuctx, int cpu) +{ + event->state = PERF_EVENT_STATE_INACTIVE; + event->oncpu = -1; + + if (!is_x86_event(event)) + event->pmu->disable(event); + + event->tstamp_running -= event->ctx->time - event->tstamp_stopped; + + if (!is_software_event(event)) + cpuctx->active_oncpu--; + + if (event->attr.exclusive || !cpuctx->active_oncpu) + cpuctx->exclusive = 0; +} + +/* + * Called to enable a whole group of events. + * Returns 1 if the group was enabled, or -EAGAIN if it could not be. + * Assumes the caller has disabled interrupts and has + * frozen the PMU with hw_perf_save_disable. + * + * called with PMU disabled. If successful and return value 1, + * then guaranteed to call perf_enable() and hw_perf_enable() + */ +int hw_perf_group_sched_in(struct perf_event *leader, + struct perf_cpu_context *cpuctx, + struct perf_event_context *ctx, int cpu) +{ + struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu); + struct perf_event *sub; + int assign[X86_PMC_IDX_MAX]; + int n0, n1, ret; + + /* n0 = total number of events */ + n0 = collect_events(cpuc, leader, true); + if (n0 < 0) + return n0; + + ret = x86_schedule_events(cpuc, n0, assign); + if (ret) + return ret; + + ret = x86_event_sched_in(leader, cpuctx, cpu); + if (ret) + return ret; + + n1 = 1; + list_for_each_entry(sub, &leader->sibling_list, group_entry) { + if (sub->state != PERF_EVENT_STATE_OFF) { + ret = x86_event_sched_in(sub, cpuctx, cpu); + if (ret) + goto undo; + ++n1; + } + } + /* + * copy new assignment, now we know it is possible + * will be used by hw_perf_enable() + */ + memcpy(cpuc->assign, assign, n0*sizeof(int)); + + cpuc->n_events = n0; + cpuc->n_added = n1; + ctx->nr_active += n1; + + /* + * 1 means successful and events are active + * This is not quite true because we defer + * actual activation until hw_perf_enable() but + * this way we* ensure caller won't try to enable + * individual events + */ + return 1; +undo: + x86_event_sched_out(leader, cpuctx, cpu); + n0 = 1; + list_for_each_entry(sub, &leader->sibling_list, group_entry) { + if (sub->state == PERF_EVENT_STATE_ACTIVE) { + x86_event_sched_out(sub, cpuctx, cpu); + if (++n0 == n1) + break; + } + } + return ret; +} + static __read_mostly struct notifier_block perf_event_nmi_notifier = { .notifier_call = perf_event_nmi_handler, .next = NULL, @@ -1993,7 +2327,8 @@ static __initconst struct x86_pmu p6_pmu = { */ .event_bits = 32, .event_mask = (1ULL << 32) - 1, - .get_event_idx = intel_get_event_idx, + .get_event_constraints = intel_get_event_constraints, + .event_constraints = intel_p6_event_constraints }; static __initconst struct x86_pmu intel_pmu = { @@ -2017,7 +2352,7 @@ static __initconst struct x86_pmu intel_pmu = { .max_period = (1ULL << 31) - 1, .enable_bts = intel_pmu_enable_bts, .disable_bts = intel_pmu_disable_bts, - .get_event_idx = intel_get_event_idx, + .get_event_constraints = intel_get_event_constraints }; static __initconst struct x86_pmu amd_pmu = { @@ -2038,7 +2373,7 @@ static __initconst struct x86_pmu amd_pmu = { .apic = 1, /* use highest bit to detect overflow */ .max_period = (1ULL << 47) - 1, - .get_event_idx = gen_get_event_idx, + .get_event_constraints = amd_get_event_constraints }; static __init int p6_pmu_init(void) @@ -2051,12 +2386,9 @@ static __init int p6_pmu_init(void) case 7: case 8: case 11: /* Pentium III */ - event_constraints = intel_p6_event_constraints; - break; case 9: case 13: /* Pentium M */ - event_constraints = intel_p6_event_constraints; break; default: pr_cont("unsupported p6 CPU model %d ", @@ -2121,23 +2453,29 @@ static __init int intel_pmu_init(void) memcpy(hw_cache_event_ids, core2_hw_cache_event_ids, sizeof(hw_cache_event_ids)); + x86_pmu.event_constraints = intel_core_event_constraints; pr_cont("Core2 events, "); - event_constraints = intel_core_event_constraints; break; - default: case 26: memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids, sizeof(hw_cache_event_ids)); - event_constraints = intel_nehalem_event_constraints; + x86_pmu.event_constraints = intel_nehalem_event_constraints; pr_cont("Nehalem/Corei7 events, "); break; case 28: memcpy(hw_cache_event_ids, atom_hw_cache_event_ids, sizeof(hw_cache_event_ids)); + x86_pmu.event_constraints = intel_gen_event_constraints; pr_cont("Atom events, "); break; + default: + /* + * default constraints for v2 and up + */ + x86_pmu.event_constraints = intel_gen_event_constraints; + pr_cont("generic architected perfmon, "); } return 0; } @@ -2234,36 +2572,43 @@ static const struct pmu pmu = { .unthrottle = x86_pmu_unthrottle, }; -static int -validate_event(struct cpu_hw_events *cpuc, struct perf_event *event) -{ - struct hw_perf_event fake_event = event->hw; - - if (event->pmu && event->pmu != &pmu) - return 0; - - return x86_schedule_event(cpuc, &fake_event) >= 0; -} - +/* + * validate a single event group + * + * validation include: + * - check events are compatible which each other + * - events do not compete for the same counter + * - number of events <= number of counters + * + * validation ensures the group can be loaded onto the + * PMU if it was the only group available. + */ static int validate_group(struct perf_event *event) { - struct perf_event *sibling, *leader = event->group_leader; - struct cpu_hw_events fake_pmu; + struct perf_event *leader = event->group_leader; + struct cpu_hw_events fake_cpuc; + int n; - memset(&fake_pmu, 0, sizeof(fake_pmu)); + memset(&fake_cpuc, 0, sizeof(fake_cpuc)); - if (!validate_event(&fake_pmu, leader)) + /* + * the event is not yet connected with its + * siblings therefore we must first collect + * existing siblings, then add the new event + * before we can simulate the scheduling + */ + n = collect_events(&fake_cpuc, leader, true); + if (n < 0) return -ENOSPC; - list_for_each_entry(sibling, &leader->sibling_list, group_entry) { - if (!validate_event(&fake_pmu, sibling)) - return -ENOSPC; - } - - if (!validate_event(&fake_pmu, event)) + fake_cpuc.n_events = n; + n = collect_events(&fake_cpuc, event, false); + if (n < 0) return -ENOSPC; - return 0; + fake_cpuc.n_events = n; + + return x86_schedule_events(&fake_cpuc, n, NULL); } const struct pmu *hw_perf_event_init(struct perf_event *event) -- cgit v0.10.2 From 8113070d6639d2245c6c79afb8df42cedab30540 Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Thu, 21 Jan 2010 17:39:01 +0200 Subject: perf_events: Add fast-path to the rescheduling code Implement correct fastpath scheduling, i.e., reuse previous assignment. Signed-off-by: Stephane Eranian [ split from larger patch] Signed-off-by: Peter Zijlstra LKML-Reference: <4b588464.1818d00a.4456.383b@mx.google.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 995ac4a..0bd23d0 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1245,6 +1245,46 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) } /* + * fastpath, try to reuse previous register + */ + for (i = 0, num = n; i < n; i++, num--) { + hwc = &cpuc->event_list[i]->hw; + c = (unsigned long *)constraints[i]; + + /* never assigned */ + if (hwc->idx == -1) + break; + + /* constraint still honored */ + if (!test_bit(hwc->idx, c)) + break; + + /* not already used */ + if (test_bit(hwc->idx, used_mask)) + break; + +#if 0 + pr_debug("CPU%d fast config=0x%llx idx=%d assign=%c\n", + smp_processor_id(), + hwc->config, + hwc->idx, + assign ? 'y' : 'n'); +#endif + + set_bit(hwc->idx, used_mask); + if (assign) + assign[i] = hwc->idx; + } + if (!num) + goto done; + + /* + * begin slow path + */ + + bitmap_zero(used_mask, X86_PMC_IDX_MAX); + + /* * weight = number of possible counters * * 1 = most constrained, only works on one counter @@ -1263,10 +1303,9 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) if (x86_pmu.num_events_fixed) wmax++; - num = n; - for (w = 1; num && w <= wmax; w++) { + for (w = 1, num = n; num && w <= wmax; w++) { /* for each event */ - for (i = 0; i < n; i++) { + for (i = 0; num && i < n; i++) { c = (unsigned long *)constraints[i]; hwc = &cpuc->event_list[i]->hw; @@ -1274,28 +1313,6 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) if (weight != w) continue; - /* - * try to reuse previous assignment - * - * This is possible despite the fact that - * events or events order may have changed. - * - * What matters is the level of constraints - * of an event and this is constant for now. - * - * This is possible also because we always - * scan from most to least constrained. Thus, - * if a counter can be reused, it means no, - * more constrained events, needed it. And - * next events will either compete for it - * (which cannot be solved anyway) or they - * have fewer constraints, and they can use - * another counter. - */ - j = hwc->idx; - if (j != -1 && !test_bit(j, used_mask)) - goto skip; - for_each_bit(j, c, X86_PMC_IDX_MAX) { if (!test_bit(j, used_mask)) break; @@ -1303,22 +1320,23 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) if (j == X86_PMC_IDX_MAX) break; -skip: - set_bit(j, used_mask); #if 0 - pr_debug("CPU%d config=0x%llx idx=%d assign=%c\n", + pr_debug("CPU%d slow config=0x%llx idx=%d assign=%c\n", smp_processor_id(), hwc->config, j, assign ? 'y' : 'n'); #endif + set_bit(j, used_mask); + if (assign) assign[i] = j; num--; } } +done: /* * scheduling failed or is just a simulation, * free resources if necessary @@ -1357,7 +1375,7 @@ static int collect_events(struct cpu_hw_events *cpuc, struct perf_event *leader, list_for_each_entry(event, &leader->sibling_list, group_entry) { if (!is_x86_event(event) || - event->state == PERF_EVENT_STATE_OFF) + event->state <= PERF_EVENT_STATE_OFF) continue; if (n >= max_count) @@ -2184,6 +2202,8 @@ static void amd_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event, u64 *idxmsk) { + /* no constraints, means supports all generic counters */ + bitmap_fill((unsigned long *)idxmsk, x86_pmu.num_events); } static int x86_event_sched_in(struct perf_event *event, @@ -2258,7 +2278,7 @@ int hw_perf_group_sched_in(struct perf_event *leader, n1 = 1; list_for_each_entry(sub, &leader->sibling_list, group_entry) { - if (sub->state != PERF_EVENT_STATE_OFF) { + if (sub->state > PERF_EVENT_STATE_OFF) { ret = x86_event_sched_in(sub, cpuctx, cpu); if (ret) goto undo; @@ -2613,12 +2633,23 @@ static int validate_group(struct perf_event *event) const struct pmu *hw_perf_event_init(struct perf_event *event) { + const struct pmu *tmp; int err; err = __hw_perf_event_init(event); if (!err) { + /* + * we temporarily connect event to its pmu + * such that validate_group() can classify + * it as an x86 event using is_x86_event() + */ + tmp = event->pmu; + event->pmu = &pmu; + if (event->group_leader != event) err = validate_group(event); + + event->pmu = tmp; } if (err) { if (event->destroy) -- cgit v0.10.2 From 502568d563bcc37ac505a83341c0c95b88c015a8 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Jan 2010 14:35:46 +0100 Subject: perf_event: x86: Allocate the fake_cpuc GCC was complaining the stack usage was too large, so allocate the structure. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: <20100122155535.411197266@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 0bd23d0..7bd359a5 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -2606,10 +2606,13 @@ static const struct pmu pmu = { static int validate_group(struct perf_event *event) { struct perf_event *leader = event->group_leader; - struct cpu_hw_events fake_cpuc; - int n; + struct cpu_hw_events *fake_cpuc; + int ret, n; - memset(&fake_cpuc, 0, sizeof(fake_cpuc)); + ret = -ENOMEM; + fake_cpuc = kmalloc(sizeof(*fake_cpuc), GFP_KERNEL | __GFP_ZERO); + if (!fake_cpuc) + goto out; /* * the event is not yet connected with its @@ -2617,18 +2620,24 @@ static int validate_group(struct perf_event *event) * existing siblings, then add the new event * before we can simulate the scheduling */ - n = collect_events(&fake_cpuc, leader, true); + ret = -ENOSPC; + n = collect_events(fake_cpuc, leader, true); if (n < 0) - return -ENOSPC; + goto out_free; - fake_cpuc.n_events = n; - n = collect_events(&fake_cpuc, event, false); + fake_cpuc->n_events = n; + n = collect_events(fake_cpuc, event, false); if (n < 0) - return -ENOSPC; + goto out_free; - fake_cpuc.n_events = n; + fake_cpuc->n_events = n; - return x86_schedule_events(&fake_cpuc, n, NULL); + ret = x86_schedule_events(fake_cpuc, n, NULL); + +out_free: + kfree(fake_cpuc); +out: + return ret; } const struct pmu *hw_perf_event_init(struct perf_event *event) -- cgit v0.10.2 From 81269a085669b5130058a0275aa7ba9f94abd1fa Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Jan 2010 14:55:22 +0100 Subject: perf_event: x86: Fixup constraints typing issue Constraints gets defined an u64 but in long quantities and then cast to long. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: <20100122155535.504916780@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 7bd359a5..7e181a5 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1232,7 +1232,7 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) int i, j , w, num; int weight, wmax; unsigned long *c; - u64 constraints[X86_PMC_IDX_MAX][BITS_TO_LONGS(X86_PMC_IDX_MAX)]; + unsigned long constraints[X86_PMC_IDX_MAX][BITS_TO_LONGS(X86_PMC_IDX_MAX)]; unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; struct hw_perf_event *hwc; @@ -1249,7 +1249,7 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) */ for (i = 0, num = n; i < n; i++, num--) { hwc = &cpuc->event_list[i]->hw; - c = (unsigned long *)constraints[i]; + c = constraints[i]; /* never assigned */ if (hwc->idx == -1) @@ -1306,7 +1306,7 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) for (w = 1, num = n; num && w <= wmax; w++) { /* for each event */ for (i = 0; num && i < n; i++) { - c = (unsigned long *)constraints[i]; + c = constraints[i]; hwc = &cpuc->event_list[i]->hw; weight = bitmap_weight(c, X86_PMC_IDX_MAX); -- cgit v0.10.2 From c91e0f5da81c6f3a611a1bd6d0cca6717c90fdab Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Jan 2010 15:25:59 +0100 Subject: perf_event: x86: Clean up some of the u64/long bitmask casting We need this to be u64 for direct assigment, but the bitmask functions all work on unsigned long, leading to cast heaven, solve this by using a union. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: <20100122155535.595961269@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 7e181a5..921bbf7 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -69,10 +69,11 @@ struct debug_store { u64 pebs_event_reset[MAX_PEBS_EVENTS]; }; -#define BITS_TO_U64(nr) DIV_ROUND_UP(nr, BITS_PER_BYTE * sizeof(u64)) - struct event_constraint { - u64 idxmsk[BITS_TO_U64(X86_PMC_IDX_MAX)]; + union { + unsigned long idxmsk[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; + u64 idxmsk64[1]; + }; int code; int cmask; }; @@ -90,13 +91,14 @@ struct cpu_hw_events { struct perf_event *event_list[X86_PMC_IDX_MAX]; /* in enabled order */ }; -#define EVENT_CONSTRAINT(c, n, m) { \ - .code = (c), \ - .cmask = (m), \ - .idxmsk[0] = (n) } +#define EVENT_CONSTRAINT(c, n, m) { \ + { .idxmsk64[0] = (n) }, \ + .code = (c), \ + .cmask = (m), \ +} #define EVENT_CONSTRAINT_END \ - { .code = 0, .cmask = 0, .idxmsk[0] = 0 } + EVENT_CONSTRAINT(0, 0, 0) #define for_each_event_constraint(e, c) \ for ((e) = (c); (e)->cmask; (e)++) @@ -126,8 +128,11 @@ struct x86_pmu { u64 intel_ctrl; void (*enable_bts)(u64 config); void (*disable_bts)(void); - void (*get_event_constraints)(struct cpu_hw_events *cpuc, struct perf_event *event, u64 *idxmsk); - void (*put_event_constraints)(struct cpu_hw_events *cpuc, struct perf_event *event); + void (*get_event_constraints)(struct cpu_hw_events *cpuc, + struct perf_event *event, + unsigned long *idxmsk); + void (*put_event_constraints)(struct cpu_hw_events *cpuc, + struct perf_event *event); const struct event_constraint *event_constraints; }; @@ -2144,14 +2149,11 @@ perf_event_nmi_handler(struct notifier_block *self, return NOTIFY_STOP; } -static struct event_constraint bts_constraint = { - .code = 0, - .cmask = 0, - .idxmsk[0] = 1ULL << X86_PMC_IDX_FIXED_BTS -}; +static struct event_constraint bts_constraint = + EVENT_CONSTRAINT(0, 1ULL << X86_PMC_IDX_FIXED_BTS, 0); static int intel_special_constraints(struct perf_event *event, - u64 *idxmsk) + unsigned long *idxmsk) { unsigned int hw_event; @@ -2171,14 +2173,14 @@ static int intel_special_constraints(struct perf_event *event, static void intel_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event, - u64 *idxmsk) + unsigned long *idxmsk) { const struct event_constraint *c; /* * cleanup bitmask */ - bitmap_zero((unsigned long *)idxmsk, X86_PMC_IDX_MAX); + bitmap_zero(idxmsk, X86_PMC_IDX_MAX); if (intel_special_constraints(event, idxmsk)) return; @@ -2186,10 +2188,7 @@ static void intel_get_event_constraints(struct cpu_hw_events *cpuc, if (x86_pmu.event_constraints) { for_each_event_constraint(c, x86_pmu.event_constraints) { if ((event->hw.config & c->cmask) == c->code) { - - bitmap_copy((unsigned long *)idxmsk, - (unsigned long *)c->idxmsk, - X86_PMC_IDX_MAX); + bitmap_copy(idxmsk, c->idxmsk, X86_PMC_IDX_MAX); return; } } @@ -2200,10 +2199,10 @@ static void intel_get_event_constraints(struct cpu_hw_events *cpuc, static void amd_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event, - u64 *idxmsk) + unsigned long *idxmsk) { /* no constraints, means supports all generic counters */ - bitmap_fill((unsigned long *)idxmsk, x86_pmu.num_events); + bitmap_fill(idxmsk, x86_pmu.num_events); } static int x86_event_sched_in(struct perf_event *event, -- cgit v0.10.2 From 8433be1184e4f22c37d4b8ed36cde529a47882f4 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Jan 2010 15:38:26 +0100 Subject: perf_event: x86: Reduce some overly long lines with some MACROs Introduce INTEL_EVENT_CONSTRAINT and FIXED_EVENT_CONSTRAINT to reduce some line length and typing work. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: <20100122155535.688730371@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 921bbf7..4d1ed10 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -97,6 +97,12 @@ struct cpu_hw_events { .cmask = (m), \ } +#define INTEL_EVENT_CONSTRAINT(c, n) \ + EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK) + +#define FIXED_EVENT_CONSTRAINT(c, n) \ + EVENT_CONSTRAINT(c, n, INTEL_ARCH_FIXED_MASK) + #define EVENT_CONSTRAINT_END \ EVENT_CONSTRAINT(0, 0, 0) @@ -192,12 +198,12 @@ static u64 p6_pmu_raw_event(u64 hw_event) static struct event_constraint intel_p6_event_constraints[] = { - EVENT_CONSTRAINT(0xc1, 0x1, INTEL_ARCH_EVENT_MASK), /* FLOPS */ - EVENT_CONSTRAINT(0x10, 0x1, INTEL_ARCH_EVENT_MASK), /* FP_COMP_OPS_EXE */ - EVENT_CONSTRAINT(0x11, 0x1, INTEL_ARCH_EVENT_MASK), /* FP_ASSIST */ - EVENT_CONSTRAINT(0x12, 0x2, INTEL_ARCH_EVENT_MASK), /* MUL */ - EVENT_CONSTRAINT(0x13, 0x2, INTEL_ARCH_EVENT_MASK), /* DIV */ - EVENT_CONSTRAINT(0x14, 0x1, INTEL_ARCH_EVENT_MASK), /* CYCLES_DIV_BUSY */ + INTEL_EVENT_CONSTRAINT(0xc1, 0x1), /* FLOPS */ + INTEL_EVENT_CONSTRAINT(0x10, 0x1), /* FP_COMP_OPS_EXE */ + INTEL_EVENT_CONSTRAINT(0x11, 0x1), /* FP_ASSIST */ + INTEL_EVENT_CONSTRAINT(0x12, 0x2), /* MUL */ + INTEL_EVENT_CONSTRAINT(0x13, 0x2), /* DIV */ + INTEL_EVENT_CONSTRAINT(0x14, 0x1), /* CYCLES_DIV_BUSY */ EVENT_CONSTRAINT_END }; @@ -217,41 +223,41 @@ static const u64 intel_perfmon_event_map[] = static struct event_constraint intel_core_event_constraints[] = { - EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32)), INTEL_ARCH_FIXED_MASK), /* INSTRUCTIONS_RETIRED */ - EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33)), INTEL_ARCH_FIXED_MASK), /* UNHALTED_CORE_CYCLES */ - EVENT_CONSTRAINT(0x10, 0x1, INTEL_ARCH_EVENT_MASK), /* FP_COMP_OPS_EXE */ - EVENT_CONSTRAINT(0x11, 0x2, INTEL_ARCH_EVENT_MASK), /* FP_ASSIST */ - EVENT_CONSTRAINT(0x12, 0x2, INTEL_ARCH_EVENT_MASK), /* MUL */ - EVENT_CONSTRAINT(0x13, 0x2, INTEL_ARCH_EVENT_MASK), /* DIV */ - EVENT_CONSTRAINT(0x14, 0x1, INTEL_ARCH_EVENT_MASK), /* CYCLES_DIV_BUSY */ - EVENT_CONSTRAINT(0x18, 0x1, INTEL_ARCH_EVENT_MASK), /* IDLE_DURING_DIV */ - EVENT_CONSTRAINT(0x19, 0x2, INTEL_ARCH_EVENT_MASK), /* DELAYED_BYPASS */ - EVENT_CONSTRAINT(0xa1, 0x1, INTEL_ARCH_EVENT_MASK), /* RS_UOPS_DISPATCH_CYCLES */ - EVENT_CONSTRAINT(0xcb, 0x1, INTEL_ARCH_EVENT_MASK), /* MEM_LOAD_RETIRED */ + FIXED_EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ + FIXED_EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ + INTEL_EVENT_CONSTRAINT(0x10, 0x1), /* FP_COMP_OPS_EXE */ + INTEL_EVENT_CONSTRAINT(0x11, 0x2), /* FP_ASSIST */ + INTEL_EVENT_CONSTRAINT(0x12, 0x2), /* MUL */ + INTEL_EVENT_CONSTRAINT(0x13, 0x2), /* DIV */ + INTEL_EVENT_CONSTRAINT(0x14, 0x1), /* CYCLES_DIV_BUSY */ + INTEL_EVENT_CONSTRAINT(0x18, 0x1), /* IDLE_DURING_DIV */ + INTEL_EVENT_CONSTRAINT(0x19, 0x2), /* DELAYED_BYPASS */ + INTEL_EVENT_CONSTRAINT(0xa1, 0x1), /* RS_UOPS_DISPATCH_CYCLES */ + INTEL_EVENT_CONSTRAINT(0xcb, 0x1), /* MEM_LOAD_RETIRED */ EVENT_CONSTRAINT_END }; static struct event_constraint intel_nehalem_event_constraints[] = { - EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32)), INTEL_ARCH_FIXED_MASK), /* INSTRUCTIONS_RETIRED */ - EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33)), INTEL_ARCH_FIXED_MASK), /* UNHALTED_CORE_CYCLES */ - EVENT_CONSTRAINT(0x40, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_CACHE_LD */ - EVENT_CONSTRAINT(0x41, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_CACHE_ST */ - EVENT_CONSTRAINT(0x42, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_CACHE_LOCK */ - EVENT_CONSTRAINT(0x43, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_ALL_REF */ - EVENT_CONSTRAINT(0x4e, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_PREFETCH */ - EVENT_CONSTRAINT(0x4c, 0x3, INTEL_ARCH_EVENT_MASK), /* LOAD_HIT_PRE */ - EVENT_CONSTRAINT(0x51, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D */ - EVENT_CONSTRAINT(0x52, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_CACHE_PREFETCH_LOCK_FB_HIT */ - EVENT_CONSTRAINT(0x53, 0x3, INTEL_ARCH_EVENT_MASK), /* L1D_CACHE_LOCK_FB_HIT */ - EVENT_CONSTRAINT(0xc5, 0x3, INTEL_ARCH_EVENT_MASK), /* CACHE_LOCK_CYCLES */ + FIXED_EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ + FIXED_EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ + INTEL_EVENT_CONSTRAINT(0x40, 0x3), /* L1D_CACHE_LD */ + INTEL_EVENT_CONSTRAINT(0x41, 0x3), /* L1D_CACHE_ST */ + INTEL_EVENT_CONSTRAINT(0x42, 0x3), /* L1D_CACHE_LOCK */ + INTEL_EVENT_CONSTRAINT(0x43, 0x3), /* L1D_ALL_REF */ + INTEL_EVENT_CONSTRAINT(0x4e, 0x3), /* L1D_PREFETCH */ + INTEL_EVENT_CONSTRAINT(0x4c, 0x3), /* LOAD_HIT_PRE */ + INTEL_EVENT_CONSTRAINT(0x51, 0x3), /* L1D */ + INTEL_EVENT_CONSTRAINT(0x52, 0x3), /* L1D_CACHE_PREFETCH_LOCK_FB_HIT */ + INTEL_EVENT_CONSTRAINT(0x53, 0x3), /* L1D_CACHE_LOCK_FB_HIT */ + INTEL_EVENT_CONSTRAINT(0xc5, 0x3), /* CACHE_LOCK_CYCLES */ EVENT_CONSTRAINT_END }; static struct event_constraint intel_gen_event_constraints[] = { - EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32)), INTEL_ARCH_FIXED_MASK), /* INSTRUCTIONS_RETIRED */ - EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33)), INTEL_ARCH_FIXED_MASK), /* UNHALTED_CORE_CYCLES */ + FIXED_EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ + FIXED_EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ EVENT_CONSTRAINT_END }; -- cgit v0.10.2 From 9f41699ed067fa695faff8e2e9981b2550abec62 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Jan 2010 15:59:29 +0100 Subject: bitops: Provide compile time HWEIGHT{8,16,32,64} Provide compile time versions of hweight. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian Cc: Linus Torvalds Cc: Andrew Morton Cc: Thomas Gleixner LKML-Reference: <20100122155535.797688466@chello.nl> [ Remove some whitespace damage while we are at it ] Signed-off-by: Ingo Molnar diff --git a/include/linux/bitops.h b/include/linux/bitops.h index c05a29c..ba0fd1e 100644 --- a/include/linux/bitops.h +++ b/include/linux/bitops.h @@ -25,7 +25,7 @@ static __inline__ int get_bitmask_order(unsigned int count) { int order; - + order = fls(count); return order; /* We could be slightly more clever with -1 here... */ } @@ -33,7 +33,7 @@ static __inline__ int get_bitmask_order(unsigned int count) static __inline__ int get_count_order(unsigned int count) { int order; - + order = fls(count) - 1; if (count & (count - 1)) order++; @@ -45,6 +45,20 @@ static inline unsigned long hweight_long(unsigned long w) return sizeof(w) == 4 ? hweight32(w) : hweight64(w); } +#define HWEIGHT8(w) \ + ( (!!((w) & (1ULL << 0))) + \ + (!!((w) & (1ULL << 1))) + \ + (!!((w) & (1ULL << 2))) + \ + (!!((w) & (1ULL << 3))) + \ + (!!((w) & (1ULL << 4))) + \ + (!!((w) & (1ULL << 5))) + \ + (!!((w) & (1ULL << 6))) + \ + (!!((w) & (1ULL << 7))) ) + +#define HWEIGHT16(w) (HWEIGHT8(w) + HWEIGHT8(w >> 8)) +#define HWEIGHT32(w) (HWEIGHT16(w) + HWEIGHT16(w >> 16)) +#define HWEIGHT64(w) (HWEIGHT32(w) + HWEIGHT32(w >> 32)) + /** * rol32 - rotate a 32-bit value left * @word: value to rotate -- cgit v0.10.2 From 63b146490befc027a7e0923e333269e68b20d380 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Jan 2010 16:32:17 +0100 Subject: perf_event: x86: Optimize the constraint searching bits Instead of copying bitmasks around, pass pointers to the constraint structure. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: <20100122155535.887853503@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 4d1ed10..092ad56 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -134,12 +134,14 @@ struct x86_pmu { u64 intel_ctrl; void (*enable_bts)(u64 config); void (*disable_bts)(void); - void (*get_event_constraints)(struct cpu_hw_events *cpuc, - struct perf_event *event, - unsigned long *idxmsk); + + struct event_constraint * + (*get_event_constraints)(struct cpu_hw_events *cpuc, + struct perf_event *event); + void (*put_event_constraints)(struct cpu_hw_events *cpuc, struct perf_event *event); - const struct event_constraint *event_constraints; + struct event_constraint *event_constraints; }; static struct x86_pmu x86_pmu __read_mostly; @@ -1242,17 +1244,15 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) { int i, j , w, num; int weight, wmax; - unsigned long *c; - unsigned long constraints[X86_PMC_IDX_MAX][BITS_TO_LONGS(X86_PMC_IDX_MAX)]; + struct event_constraint *c, *constraints[X86_PMC_IDX_MAX]; unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; struct hw_perf_event *hwc; bitmap_zero(used_mask, X86_PMC_IDX_MAX); for (i = 0; i < n; i++) { - x86_pmu.get_event_constraints(cpuc, - cpuc->event_list[i], - constraints[i]); + constraints[i] = + x86_pmu.get_event_constraints(cpuc, cpuc->event_list[i]); } /* @@ -1267,7 +1267,7 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) break; /* constraint still honored */ - if (!test_bit(hwc->idx, c)) + if (!test_bit(hwc->idx, c->idxmsk)) break; /* not already used */ @@ -1320,11 +1320,11 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) c = constraints[i]; hwc = &cpuc->event_list[i]->hw; - weight = bitmap_weight(c, X86_PMC_IDX_MAX); + weight = bitmap_weight(c->idxmsk, X86_PMC_IDX_MAX); if (weight != w) continue; - for_each_bit(j, c, X86_PMC_IDX_MAX) { + for_each_bit(j, c->idxmsk, X86_PMC_IDX_MAX) { if (!test_bit(j, used_mask)) break; } @@ -2155,11 +2155,13 @@ perf_event_nmi_handler(struct notifier_block *self, return NOTIFY_STOP; } +static struct event_constraint unconstrained; + static struct event_constraint bts_constraint = EVENT_CONSTRAINT(0, 1ULL << X86_PMC_IDX_FIXED_BTS, 0); -static int intel_special_constraints(struct perf_event *event, - unsigned long *idxmsk) +static struct event_constraint * +intel_special_constraints(struct perf_event *event) { unsigned int hw_event; @@ -2169,46 +2171,34 @@ static int intel_special_constraints(struct perf_event *event, x86_pmu.event_map(PERF_COUNT_HW_BRANCH_INSTRUCTIONS)) && (event->hw.sample_period == 1))) { - bitmap_copy((unsigned long *)idxmsk, - (unsigned long *)bts_constraint.idxmsk, - X86_PMC_IDX_MAX); - return 1; + return &bts_constraint; } - return 0; + return NULL; } -static void intel_get_event_constraints(struct cpu_hw_events *cpuc, - struct perf_event *event, - unsigned long *idxmsk) +static struct event_constraint * +intel_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event) { - const struct event_constraint *c; + struct event_constraint *c; - /* - * cleanup bitmask - */ - bitmap_zero(idxmsk, X86_PMC_IDX_MAX); - - if (intel_special_constraints(event, idxmsk)) - return; + c = intel_special_constraints(event); + if (c) + return c; if (x86_pmu.event_constraints) { for_each_event_constraint(c, x86_pmu.event_constraints) { - if ((event->hw.config & c->cmask) == c->code) { - bitmap_copy(idxmsk, c->idxmsk, X86_PMC_IDX_MAX); - return; - } + if ((event->hw.config & c->cmask) == c->code) + return c; } } - /* no constraints, means supports all generic counters */ - bitmap_fill((unsigned long *)idxmsk, x86_pmu.num_events); + + return &unconstrained; } -static void amd_get_event_constraints(struct cpu_hw_events *cpuc, - struct perf_event *event, - unsigned long *idxmsk) +static struct event_constraint * +amd_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event) { - /* no constraints, means supports all generic counters */ - bitmap_fill(idxmsk, x86_pmu.num_events); + return &unconstrained; } static int x86_event_sched_in(struct perf_event *event, @@ -2576,6 +2566,9 @@ void __init init_hw_perf_events(void) perf_events_lapic_init(); register_die_notifier(&perf_event_nmi_notifier); + unconstrained = (struct event_constraint) + EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_events) - 1, 0); + pr_info("... version: %d\n", x86_pmu.version); pr_info("... bit width: %d\n", x86_pmu.event_bits); pr_info("... generic registers: %d\n", x86_pmu.num_events); -- cgit v0.10.2 From 272d30be622c9c6cbd514b1211ff359292001baa Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Jan 2010 16:32:17 +0100 Subject: perf_event: x86: Optimize constraint weight computation Add a weight member to the constraint structure and avoid recomputing the weight at runtime. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: <20100122155535.963944926@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 092ad56..2c22ce4 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -23,6 +23,7 @@ #include #include #include +#include #include #include @@ -76,6 +77,7 @@ struct event_constraint { }; int code; int cmask; + int weight; }; struct cpu_hw_events { @@ -95,6 +97,7 @@ struct cpu_hw_events { { .idxmsk64[0] = (n) }, \ .code = (c), \ .cmask = (m), \ + .weight = HWEIGHT64((u64)(n)), \ } #define INTEL_EVENT_CONSTRAINT(c, n) \ @@ -1242,8 +1245,7 @@ static inline int is_x86_event(struct perf_event *event) static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) { - int i, j , w, num; - int weight, wmax; + int i, j, w, num, wmax; struct event_constraint *c, *constraints[X86_PMC_IDX_MAX]; unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; struct hw_perf_event *hwc; @@ -1320,8 +1322,7 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) c = constraints[i]; hwc = &cpuc->event_list[i]->hw; - weight = bitmap_weight(c->idxmsk, X86_PMC_IDX_MAX); - if (weight != w) + if (c->weight != w) continue; for_each_bit(j, c->idxmsk, X86_PMC_IDX_MAX) { -- cgit v0.10.2 From c933c1a603d5bf700ddce79216c1be0ec3bc0e6c Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Jan 2010 16:40:12 +0100 Subject: perf_event: x86: Optimize the fast path a little more Remove num from the fast path and save a few ops. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: <20100122155536.056430539@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 2c22ce4..33c889f 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1245,9 +1245,9 @@ static inline int is_x86_event(struct perf_event *event) static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) { - int i, j, w, num, wmax; struct event_constraint *c, *constraints[X86_PMC_IDX_MAX]; unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; + int i, j, w, wmax, num = 0; struct hw_perf_event *hwc; bitmap_zero(used_mask, X86_PMC_IDX_MAX); @@ -1260,7 +1260,7 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) /* * fastpath, try to reuse previous register */ - for (i = 0, num = n; i < n; i++, num--) { + for (i = 0; i < n; i++) { hwc = &cpuc->event_list[i]->hw; c = constraints[i]; @@ -1288,7 +1288,7 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) if (assign) assign[i] = hwc->idx; } - if (!num) + if (i == n) goto done; /* -- cgit v0.10.2 From 6c9687abeb24d5b7aae7db5be070c2139ad29e29 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 25 Jan 2010 11:57:25 +0100 Subject: perf_event: x86: Optimize x86_pmu_disable() x86_pmu_disable() removes the event from the cpuc->event_list[], however since an event can only be on that list once, stop looking after we found it. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 33c889f..66de282 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1884,6 +1884,7 @@ static void x86_pmu_disable(struct perf_event *event) cpuc->event_list[i-1] = cpuc->event_list[i]; --cpuc->n_events; + break; } } perf_event_update_userpage(event); -- cgit v0.10.2 From 184f412c3341cd24fbd26604634a5800b83dbdc3 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 27 Jan 2010 08:39:39 +0100 Subject: perf, x86: Clean up event constraints code a bit - Remove stray debug code - Improve ugly macros a bit - Remove some whitespace damage - (Also fix up some accumulated damage in perf_event.h) Signed-off-by: Ingo Molnar Cc: Stephane Eranian Cc: Peter Zijlstra LKML-Reference: diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 66de282..fdbe248 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -93,24 +93,19 @@ struct cpu_hw_events { struct perf_event *event_list[X86_PMC_IDX_MAX]; /* in enabled order */ }; -#define EVENT_CONSTRAINT(c, n, m) { \ +#define EVENT_CONSTRAINT(c, n, m) { \ { .idxmsk64[0] = (n) }, \ .code = (c), \ .cmask = (m), \ .weight = HWEIGHT64((u64)(n)), \ } -#define INTEL_EVENT_CONSTRAINT(c, n) \ - EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK) +#define INTEL_EVENT_CONSTRAINT(c, n) EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK) +#define FIXED_EVENT_CONSTRAINT(c, n) EVENT_CONSTRAINT(c, n, INTEL_ARCH_FIXED_MASK) -#define FIXED_EVENT_CONSTRAINT(c, n) \ - EVENT_CONSTRAINT(c, n, INTEL_ARCH_FIXED_MASK) +#define EVENT_CONSTRAINT_END EVENT_CONSTRAINT(0, 0, 0) -#define EVENT_CONSTRAINT_END \ - EVENT_CONSTRAINT(0, 0, 0) - -#define for_each_event_constraint(e, c) \ - for ((e) = (c); (e)->cmask; (e)++) +#define for_each_event_constraint(e, c) for ((e) = (c); (e)->cmask; (e)++) /* * struct x86_pmu - generic x86 pmu @@ -1276,14 +1271,6 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) if (test_bit(hwc->idx, used_mask)) break; -#if 0 - pr_debug("CPU%d fast config=0x%llx idx=%d assign=%c\n", - smp_processor_id(), - hwc->config, - hwc->idx, - assign ? 'y' : 'n'); -#endif - set_bit(hwc->idx, used_mask); if (assign) assign[i] = hwc->idx; @@ -1333,14 +1320,6 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) if (j == X86_PMC_IDX_MAX) break; -#if 0 - pr_debug("CPU%d slow config=0x%llx idx=%d assign=%c\n", - smp_processor_id(), - hwc->config, - j, - assign ? 'y' : 'n'); -#endif - set_bit(j, used_mask); if (assign) @@ -2596,9 +2575,9 @@ static const struct pmu pmu = { * validate a single event group * * validation include: - * - check events are compatible which each other - * - events do not compete for the same counter - * - number of events <= number of counters + * - check events are compatible which each other + * - events do not compete for the same counter + * - number of events <= number of counters * * validation ensures the group can be loaded onto the * PMU if it was the only group available. diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 72b2615..953c177 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -290,7 +290,7 @@ struct perf_event_mmap_page { }; #define PERF_RECORD_MISC_CPUMODE_MASK (3 << 0) -#define PERF_RECORD_MISC_CPUMODE_UNKNOWN (0 << 0) +#define PERF_RECORD_MISC_CPUMODE_UNKNOWN (0 << 0) #define PERF_RECORD_MISC_KERNEL (1 << 0) #define PERF_RECORD_MISC_USER (2 << 0) #define PERF_RECORD_MISC_HYPERVISOR (3 << 0) @@ -356,8 +356,8 @@ enum perf_event_type { * u64 stream_id; * }; */ - PERF_RECORD_THROTTLE = 5, - PERF_RECORD_UNTHROTTLE = 6, + PERF_RECORD_THROTTLE = 5, + PERF_RECORD_UNTHROTTLE = 6, /* * struct { @@ -371,10 +371,10 @@ enum perf_event_type { /* * struct { - * struct perf_event_header header; - * u32 pid, tid; + * struct perf_event_header header; + * u32 pid, tid; * - * struct read_format values; + * struct read_format values; * }; */ PERF_RECORD_READ = 8, @@ -412,7 +412,7 @@ enum perf_event_type { * char data[size];}&& PERF_SAMPLE_RAW * }; */ - PERF_RECORD_SAMPLE = 9, + PERF_RECORD_SAMPLE = 9, PERF_RECORD_MAX, /* non-ABI */ }; @@ -752,8 +752,7 @@ extern int perf_max_events; extern const struct pmu *hw_perf_event_init(struct perf_event *event); extern void perf_event_task_sched_in(struct task_struct *task); -extern void perf_event_task_sched_out(struct task_struct *task, - struct task_struct *next); +extern void perf_event_task_sched_out(struct task_struct *task, struct task_struct *next); extern void perf_event_task_tick(struct task_struct *task); extern int perf_event_init_task(struct task_struct *child); extern void perf_event_exit_task(struct task_struct *child); @@ -853,8 +852,7 @@ extern int sysctl_perf_event_mlock; extern int sysctl_perf_event_sample_rate; extern void perf_event_init(void); -extern void perf_tp_event(int event_id, u64 addr, u64 count, - void *record, int entry_size); +extern void perf_tp_event(int event_id, u64 addr, u64 count, void *record, int entry_size); extern void perf_bp_event(struct perf_event *event, void *data); #ifndef perf_misc_flags @@ -895,13 +893,13 @@ static inline void perf_sw_event(u32 event_id, u64 nr, int nmi, struct pt_regs *regs, u64 addr) { } static inline void -perf_bp_event(struct perf_event *event, void *data) { } +perf_bp_event(struct perf_event *event, void *data) { } static inline void perf_event_mmap(struct vm_area_struct *vma) { } static inline void perf_event_comm(struct task_struct *tsk) { } static inline void perf_event_fork(struct task_struct *tsk) { } static inline void perf_event_init(void) { } -static inline int perf_swevent_get_recursion_context(void) { return -1; } +static inline int perf_swevent_get_recursion_context(void) { return -1; } static inline void perf_swevent_put_recursion_context(int rctx) { } static inline void perf_event_enable(struct perf_event *event) { } static inline void perf_event_disable(struct perf_event *event) { } -- cgit v0.10.2 From 2e8418736dff9c6fdadb2f87dcc2087cebf32167 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 25 Jan 2010 15:58:43 +0100 Subject: perf_event: x86: Deduplicate the disable code Share the meat of the x86_pmu_disable() code with hw_perf_enable(). Also remove the barrier() from that code, since I could not convince myself we actually need it. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index fdbe248..07fa0c2 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1401,6 +1401,8 @@ static inline void x86_assign_hw_event(struct perf_event *event, } } +static void __x86_pmu_disable(struct perf_event *event, struct cpu_hw_events *cpuc); + void hw_perf_enable(void) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); @@ -1426,13 +1428,7 @@ void hw_perf_enable(void) if (hwc->idx == -1 || hwc->idx == cpuc->assign[i]) continue; - x86_pmu.disable(hwc, hwc->idx); - - clear_bit(hwc->idx, cpuc->active_mask); - barrier(); - cpuc->events[hwc->idx] = NULL; - - x86_perf_event_update(event, hwc, hwc->idx); + __x86_pmu_disable(event, cpuc); hwc->idx = -1; } @@ -1822,11 +1818,10 @@ static void intel_pmu_drain_bts_buffer(struct cpu_hw_events *cpuc) event->pending_kill = POLL_IN; } -static void x86_pmu_disable(struct perf_event *event) +static void __x86_pmu_disable(struct perf_event *event, struct cpu_hw_events *cpuc) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct hw_perf_event *hwc = &event->hw; - int i, idx = hwc->idx; + int idx = hwc->idx; /* * Must be done before we disable, otherwise the nmi handler @@ -1836,12 +1831,6 @@ static void x86_pmu_disable(struct perf_event *event) x86_pmu.disable(hwc, idx); /* - * Make sure the cleared pointer becomes visible before we - * (potentially) free the event: - */ - barrier(); - - /* * Drain the remaining delta count out of a event * that we are disabling: */ @@ -1852,6 +1841,14 @@ static void x86_pmu_disable(struct perf_event *event) intel_pmu_drain_bts_buffer(cpuc); cpuc->events[idx] = NULL; +} + +static void x86_pmu_disable(struct perf_event *event) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + int i; + + __x86_pmu_disable(event, cpuc); for (i = 0; i < cpuc->n_events; i++) { if (event == cpuc->event_list[i]) { -- cgit v0.10.2 From ed8777fc132e589d48a0ba854fdbb5d8203b58e5 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 27 Jan 2010 23:07:46 +0100 Subject: perf_events, x86: Fix event constraint masks Since constraints are specified on the event number, not number and unit mask shorten the constraint masks so that we'll actually match something. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: <20100127221121.967610372@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h index dbc0826..ff5ede1 100644 --- a/arch/x86/include/asm/perf_event.h +++ b/arch/x86/include/asm/perf_event.h @@ -49,7 +49,7 @@ INTEL_ARCH_INV_MASK| \ INTEL_ARCH_EDGE_MASK|\ INTEL_ARCH_UNIT_MASK|\ - INTEL_ARCH_EVENT_MASK) + INTEL_ARCH_EVTSEL_MASK) #define ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL 0x3c #define ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK (0x00 << 8) diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 07fa0c2..951213a 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -100,12 +100,17 @@ struct cpu_hw_events { .weight = HWEIGHT64((u64)(n)), \ } -#define INTEL_EVENT_CONSTRAINT(c, n) EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK) -#define FIXED_EVENT_CONSTRAINT(c, n) EVENT_CONSTRAINT(c, n, INTEL_ARCH_FIXED_MASK) +#define INTEL_EVENT_CONSTRAINT(c, n) \ + EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVTSEL_MASK) -#define EVENT_CONSTRAINT_END EVENT_CONSTRAINT(0, 0, 0) +#define FIXED_EVENT_CONSTRAINT(c, n) \ + EVENT_CONSTRAINT(c, n, INTEL_ARCH_FIXED_MASK) -#define for_each_event_constraint(e, c) for ((e) = (c); (e)->cmask; (e)++) +#define EVENT_CONSTRAINT_END \ + EVENT_CONSTRAINT(0, 0, 0) + +#define for_each_event_constraint(e, c) \ + for ((e) = (c); (e)->cmask; (e)++) /* * struct x86_pmu - generic x86 pmu -- cgit v0.10.2 From 1a6e21f791fe85b40a9ddbafe999ab8ccffc3f78 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 27 Jan 2010 23:07:47 +0100 Subject: perf_events, x86: Clean up hw_perf_*_all() implementation Put the recursion avoidance code in the generic hook instead of replicating it in each implementation. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: <20100127221122.057507285@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 951213a..cf10839 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1099,15 +1099,8 @@ static int __hw_perf_event_init(struct perf_event *event) static void p6_pmu_disable_all(void) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); u64 val; - if (!cpuc->enabled) - return; - - cpuc->enabled = 0; - barrier(); - /* p6 only has one enable register */ rdmsrl(MSR_P6_EVNTSEL0, val); val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE; @@ -1118,12 +1111,6 @@ static void intel_pmu_disable_all(void) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - if (!cpuc->enabled) - return; - - cpuc->enabled = 0; - barrier(); - wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0); if (test_bit(X86_PMC_IDX_FIXED_BTS, cpuc->active_mask)) @@ -1135,17 +1122,6 @@ static void amd_pmu_disable_all(void) struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); int idx; - if (!cpuc->enabled) - return; - - cpuc->enabled = 0; - /* - * ensure we write the disable before we start disabling the - * events proper, so that amd_pmu_enable_event() does the - * right thing. - */ - barrier(); - for (idx = 0; idx < x86_pmu.num_events; idx++) { u64 val; @@ -1166,23 +1142,20 @@ void hw_perf_disable(void) if (!x86_pmu_initialized()) return; - if (cpuc->enabled) - cpuc->n_added = 0; + if (!cpuc->enabled) + return; + + cpuc->n_added = 0; + cpuc->enabled = 0; + barrier(); x86_pmu.disable_all(); } static void p6_pmu_enable_all(void) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); unsigned long val; - if (cpuc->enabled) - return; - - cpuc->enabled = 1; - barrier(); - /* p6 only has one enable register */ rdmsrl(MSR_P6_EVNTSEL0, val); val |= ARCH_PERFMON_EVENTSEL0_ENABLE; @@ -1193,12 +1166,6 @@ static void intel_pmu_enable_all(void) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - if (cpuc->enabled) - return; - - cpuc->enabled = 1; - barrier(); - wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, x86_pmu.intel_ctrl); if (test_bit(X86_PMC_IDX_FIXED_BTS, cpuc->active_mask)) { @@ -1217,12 +1184,6 @@ static void amd_pmu_enable_all(void) struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); int idx; - if (cpuc->enabled) - return; - - cpuc->enabled = 1; - barrier(); - for (idx = 0; idx < x86_pmu.num_events; idx++) { struct perf_event *event = cpuc->events[idx]; u64 val; @@ -1417,6 +1378,10 @@ void hw_perf_enable(void) if (!x86_pmu_initialized()) return; + + if (cpuc->enabled) + return; + if (cpuc->n_added) { /* * apply assignment obtained either from @@ -1461,6 +1426,10 @@ void hw_perf_enable(void) cpuc->n_added = 0; perf_events_lapic_init(); } + + cpuc->enabled = 1; + barrier(); + x86_pmu.enable_all(); } -- cgit v0.10.2 From 452a339a976e7f782c786eb3f73080401e2fa3a6 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 27 Jan 2010 23:07:48 +0100 Subject: perf_events, x86: Implement Intel Westmere support The new Intel documentation includes Westmere arch specific event maps that are significantly different from the Nehalem ones. Add support for this generation. Found the CPUID model numbers on wikipedia. Also ammend some Nehalem constraints, spotted those when looking for the differences between Nehalem and Westmere. Signed-off-by: Peter Zijlstra Cc: Arjan van de Ven Cc: "H. Peter Anvin" Cc: Stephane Eranian LKML-Reference: <20100127221122.151865645@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index cf10839..3fac0bf 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -244,18 +244,26 @@ static struct event_constraint intel_core_event_constraints[] = static struct event_constraint intel_nehalem_event_constraints[] = { - FIXED_EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ - FIXED_EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ + FIXED_EVENT_CONSTRAINT(0xc0, (0xf|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ + FIXED_EVENT_CONSTRAINT(0x3c, (0xf|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ INTEL_EVENT_CONSTRAINT(0x40, 0x3), /* L1D_CACHE_LD */ INTEL_EVENT_CONSTRAINT(0x41, 0x3), /* L1D_CACHE_ST */ INTEL_EVENT_CONSTRAINT(0x42, 0x3), /* L1D_CACHE_LOCK */ INTEL_EVENT_CONSTRAINT(0x43, 0x3), /* L1D_ALL_REF */ + INTEL_EVENT_CONSTRAINT(0x48, 0x3), /* L1D_PEND_MISS */ INTEL_EVENT_CONSTRAINT(0x4e, 0x3), /* L1D_PREFETCH */ - INTEL_EVENT_CONSTRAINT(0x4c, 0x3), /* LOAD_HIT_PRE */ INTEL_EVENT_CONSTRAINT(0x51, 0x3), /* L1D */ - INTEL_EVENT_CONSTRAINT(0x52, 0x3), /* L1D_CACHE_PREFETCH_LOCK_FB_HIT */ - INTEL_EVENT_CONSTRAINT(0x53, 0x3), /* L1D_CACHE_LOCK_FB_HIT */ - INTEL_EVENT_CONSTRAINT(0xc5, 0x3), /* CACHE_LOCK_CYCLES */ + INTEL_EVENT_CONSTRAINT(0x63, 0x3), /* CACHE_LOCK_CYCLES */ + EVENT_CONSTRAINT_END +}; + +static struct event_constraint intel_westmere_event_constraints[] = +{ + FIXED_EVENT_CONSTRAINT(0xc0, (0xf|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ + FIXED_EVENT_CONSTRAINT(0x3c, (0xf|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ + INTEL_EVENT_CONSTRAINT(0x51, 0x3), /* L1D */ + INTEL_EVENT_CONSTRAINT(0x60, 0x1), /* OFFCORE_REQUESTS_OUTSTANDING */ + INTEL_EVENT_CONSTRAINT(0x63, 0x3), /* CACHE_LOCK_CYCLES */ EVENT_CONSTRAINT_END }; @@ -286,6 +294,97 @@ static u64 __read_mostly hw_cache_event_ids [PERF_COUNT_HW_CACHE_OP_MAX] [PERF_COUNT_HW_CACHE_RESULT_MAX]; +static __initconst u64 westmere_hw_cache_event_ids + [PERF_COUNT_HW_CACHE_MAX] + [PERF_COUNT_HW_CACHE_OP_MAX] + [PERF_COUNT_HW_CACHE_RESULT_MAX] = +{ + [ C(L1D) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x010b, /* MEM_INST_RETIRED.LOADS */ + [ C(RESULT_MISS) ] = 0x0151, /* L1D.REPL */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x020b, /* MEM_INST_RETURED.STORES */ + [ C(RESULT_MISS) ] = 0x0251, /* L1D.M_REPL */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x014e, /* L1D_PREFETCH.REQUESTS */ + [ C(RESULT_MISS) ] = 0x024e, /* L1D_PREFETCH.MISS */ + }, + }, + [ C(L1I ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0380, /* L1I.READS */ + [ C(RESULT_MISS) ] = 0x0280, /* L1I.MISSES */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x0, + [ C(RESULT_MISS) ] = 0x0, + }, + }, + [ C(LL ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0324, /* L2_RQSTS.LOADS */ + [ C(RESULT_MISS) ] = 0x0224, /* L2_RQSTS.LD_MISS */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x0c24, /* L2_RQSTS.RFOS */ + [ C(RESULT_MISS) ] = 0x0824, /* L2_RQSTS.RFO_MISS */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x4f2e, /* LLC Reference */ + [ C(RESULT_MISS) ] = 0x412e, /* LLC Misses */ + }, + }, + [ C(DTLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x010b, /* MEM_INST_RETIRED.LOADS */ + [ C(RESULT_MISS) ] = 0x0108, /* DTLB_LOAD_MISSES.ANY */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x020b, /* MEM_INST_RETURED.STORES */ + [ C(RESULT_MISS) ] = 0x010c, /* MEM_STORE_RETIRED.DTLB_MISS */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x0, + [ C(RESULT_MISS) ] = 0x0, + }, + }, + [ C(ITLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x01c0, /* INST_RETIRED.ANY_P */ + [ C(RESULT_MISS) ] = 0x0185, /* ITLB_MISSES.ANY */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, + [ C(BPU ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x00c4, /* BR_INST_RETIRED.ALL_BRANCHES */ + [ C(RESULT_MISS) ] = 0x03e8, /* BPU_CLEARS.ANY */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, +}; + static __initconst u64 nehalem_hw_cache_event_ids [PERF_COUNT_HW_CACHE_MAX] [PERF_COUNT_HW_CACHE_OP_MAX] @@ -2423,7 +2522,9 @@ static __init int intel_pmu_init(void) x86_pmu.event_constraints = intel_core_event_constraints; pr_cont("Core2 events, "); break; - case 26: + + case 26: /* 45 nm nehalem, "Bloomfield" */ + case 30: /* 45 nm nehalem, "Lynnfield" */ memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids, sizeof(hw_cache_event_ids)); @@ -2437,6 +2538,15 @@ static __init int intel_pmu_init(void) x86_pmu.event_constraints = intel_gen_event_constraints; pr_cont("Atom events, "); break; + + case 37: /* 32 nm nehalem, "Clarkdale" */ + case 44: /* 32 nm nehalem, "Gulftown" */ + memcpy(hw_cache_event_ids, westmere_hw_cache_event_ids, + sizeof(hw_cache_event_ids)); + + x86_pmu.event_constraints = intel_westmere_event_constraints; + pr_cont("Westmere events, "); + break; default: /* * default constraints for v2 and up -- cgit v0.10.2 From 18c01f8abff51e4910cc5ffb4b710e8c6eea60c9 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 27 Jan 2010 23:07:49 +0100 Subject: perf_events, x86: Remove spurious counter reset from x86_pmu_enable() At enable time the counter might still have a ->idx pointing to a previously occupied location that might now be taken by another event. Resetting the counter at that location with data from this event will destroy the other counter's count. Signed-off-by: Peter Zijlstra Cc: Stephane Eranian LKML-Reference: <20100127221122.261477183@chello.nl> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 3fac0bf..518eb3e 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1762,9 +1762,6 @@ static int x86_pmu_enable(struct perf_event *event) cpuc->n_events = n; cpuc->n_added = n - n0; - if (hwc->idx != -1) - x86_perf_event_set_period(event, hwc, hwc->idx); - return 0; } -- cgit v0.10.2 From 75c9f3284a7ff957829f44baace82406a6354ceb Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 29 Jan 2010 09:04:26 +0100 Subject: perf_events: Fix sample_period transfer on inherit One problem with frequency driven counters is that we cannot predict the rate at which they trigger, therefore we have to start them at period=1, this causes a ramp up effect. However, if we fail to propagate the stable state on fork each new child will have to ramp up again. This can lead to significant artifacts in sample data. Signed-off-by: Peter Zijlstra Cc: eranian@google.com Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <1264752266.4283.2121.camel@laptop> Signed-off-by: Ingo Molnar diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 251fb95..53dc2a3 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -5002,8 +5002,15 @@ inherit_event(struct perf_event *parent_event, else child_event->state = PERF_EVENT_STATE_OFF; - if (parent_event->attr.freq) - child_event->hw.sample_period = parent_event->hw.sample_period; + if (parent_event->attr.freq) { + u64 sample_period = parent_event->hw.sample_period; + struct hw_perf_event *hwc = &child_event->hw; + + hwc->sample_period = sample_period; + hwc->last_period = sample_period; + + atomic64_set(&hwc->period_left, sample_period); + } child_event->overflow_handler = parent_event->overflow_handler; -- cgit v0.10.2 From 72b8fa1730207274f6818b47b891ce5dff79287e Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 27 Jan 2010 21:05:49 -0200 Subject: perf top: Exit if specified --vmlinux can't be used MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As we do lazy loading of symtabs we only will know if the specified vmlinux file is invalid when we actually have a hit in kernel space and then try to load it. So if we get kernel hits and there are _no_ symbols in the DSO backing the kernel map, bail out. Reported-by: Mike Galbraith Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1264633557-17597-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 2227b84..78f9c45 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -951,9 +951,31 @@ static void event__process_sample(const event_t *self, } if (event__preprocess_sample(self, session, &al, symbol_filter) < 0 || - al.sym == NULL || al.filtered) + al.filtered) return; + if (al.sym == NULL) { + /* + * As we do lazy loading of symtabs we only will know if the + * specified vmlinux file is invalid when we actually have a + * hit in kernel space and then try to load it. So if we get + * here and there are _no_ symbols in the DSO backing the + * kernel map, bail out. + * + * We may never get here, for instance, if we use -K/ + * --hide-kernel-symbols, even if the user specifies an + * invalid --vmlinux ;-) + */ + if (al.map == session->vmlinux_maps[MAP__FUNCTION] && + RB_EMPTY_ROOT(&al.map->dso->symbols[MAP__FUNCTION])) { + pr_err("The %s file can't be used\n", + symbol_conf.vmlinux_name); + exit(1); + } + + return; + } + syme = symbol__priv(al.sym); if (!syme->skip) { syme->count[counter]++; -- cgit v0.10.2 From a19afe46412452fef89cc623873a8931b3685944 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 27 Jan 2010 21:05:50 -0200 Subject: perf symbols: Factor out dso__load_vmlinux_path() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit So that we can call it directly from regression tests, and also to reduce the size of dso__load_kernel_sym(), making it more clear. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1264633557-17597-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index f1f609d..26ec603 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1578,6 +1578,27 @@ static int dso__load_vmlinux(struct dso *self, struct map *map, return err; } +int dso__load_vmlinux_path(struct dso *self, struct map *map, + struct perf_session *session, symbol_filter_t filter) +{ + int i, err = 0; + + pr_debug("Looking at the vmlinux_path (%d entries long)\n", + vmlinux_path__nr_entries); + + for (i = 0; i < vmlinux_path__nr_entries; ++i) { + err = dso__load_vmlinux(self, map, session, vmlinux_path[i], + filter); + if (err > 0) { + pr_debug("Using %s for symbols\n", vmlinux_path[i]); + dso__set_long_name(self, strdup(vmlinux_path[i])); + break; + } + } + + return err; +} + static int dso__load_kernel_sym(struct dso *self, struct map *map, struct perf_session *session, symbol_filter_t filter) { @@ -1606,20 +1627,9 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, } if (vmlinux_path != NULL) { - int i; - pr_debug("Looking at the vmlinux_path (%d entries long)\n", - vmlinux_path__nr_entries); - for (i = 0; i < vmlinux_path__nr_entries; ++i) { - err = dso__load_vmlinux(self, map, session, - vmlinux_path[i], filter); - if (err > 0) { - pr_debug("Using %s for symbols\n", - vmlinux_path[i]); - dso__set_long_name(self, - strdup(vmlinux_path[i])); - goto out_fixup; - } - } + err = dso__load_vmlinux_path(self, map, session, filter); + if (err > 0) + goto out_fixup; } /* diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index ffe0b0f..a94997a 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -129,6 +129,8 @@ struct perf_session; int dso__load(struct dso *self, struct map *map, struct perf_session *session, symbol_filter_t filter); +int dso__load_vmlinux_path(struct dso *self, struct map *map, + struct perf_session *session, symbol_filter_t filter); void dsos__fprintf(FILE *fp); size_t dsos__fprintf_buildid(FILE *fp, bool with_hits); -- cgit v0.10.2 From fd1d908c543fbdfae82839d24b0872c542fceedc Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 27 Jan 2010 21:05:51 -0200 Subject: perf symbols: Split helpers used when creating kernel dso object MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit To make it clear and allow for direct usage by, for instance, regression test suites. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1264633557-17597-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 26ec603..f9049d1 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1762,24 +1762,38 @@ size_t dsos__fprintf_buildid(FILE *fp, bool with_hits) __dsos__fprintf_buildid(&dsos__user, fp, with_hits)); } +struct dso *dso__new_kernel(const char *name) +{ + struct dso *self = dso__new(name ?: "[kernel.kallsyms]"); + + if (self != NULL) { + self->short_name = "[kernel]"; + self->kernel = 1; + } + + return self; +} + +void dso__read_running_kernel_build_id(struct dso *self) +{ + if (sysfs__read_build_id("/sys/kernel/notes", self->build_id, + sizeof(self->build_id)) == 0) + self->has_build_id = true; +} + static struct dso *dsos__create_kernel(const char *vmlinux) { - struct dso *kernel = dso__new(vmlinux ?: "[kernel.kallsyms]"); + struct dso *kernel = dso__new_kernel(vmlinux); if (kernel == NULL) return NULL; - kernel->short_name = "[kernel]"; - kernel->kernel = 1; - vdso = dso__new("[vdso]"); if (vdso == NULL) goto out_delete_kernel_dso; dso__set_loaded(vdso, MAP__FUNCTION); - if (sysfs__read_build_id("/sys/kernel/notes", kernel->build_id, - sizeof(kernel->build_id)) == 0) - kernel->has_build_id = true; + dso__read_running_kernel_build_id(kernel); dsos__add(&dsos__kernel, kernel); dsos__add(&dsos__user, vdso); diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index a94997a..1243027 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -109,6 +109,7 @@ struct dso { }; struct dso *dso__new(const char *name); +struct dso *dso__new_kernel(const char *name); void dso__delete(struct dso *self); bool dso__loaded(const struct dso *self, enum map_type type); @@ -139,6 +140,7 @@ size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp); char dso__symtab_origin(const struct dso *self); void dso__set_long_name(struct dso *self, char *name); void dso__set_build_id(struct dso *self, void *build_id); +void dso__read_running_kernel_build_id(struct dso *self); struct symbol *dso__find_symbol(struct dso *self, enum map_type type, u64 addr); struct symbol *dso__find_symbol_by_name(struct dso *self, enum map_type type, const char *name); -- cgit v0.10.2 From 64abebf731df87e6f4ae7d9ffc340bdf0c033e44 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 27 Jan 2010 21:05:52 -0200 Subject: perf session: Create kernel maps in the constructor MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Removing one extra step needed in the tools that need this, fixing a bug in 'perf probe' where this was not being done. Signed-off-by: Arnaldo Carvalho de Melo Cc: Masami Hiramatsu Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1264633557-17597-4-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 7323d9d..38b8ca9 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -491,11 +491,6 @@ static int __cmd_kmem(void) if (!perf_session__has_traces(session, "kmem record")) goto out_delete; - if (perf_session__create_kernel_maps(session) < 0) { - pr_err("Problems creating kernel maps\n"); - return -1; - } - setup_pager(); err = perf_session__process_events(session, &event_ops); if (err != 0) diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 7bb9ca1..9034522 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -477,11 +477,6 @@ static int __cmd_record(int argc, const char **argv) return -1; } - if (perf_session__create_kernel_maps(session) < 0) { - pr_err("Problems creating kernel maps\n"); - return -1; - } - if (!file_new) { err = perf_header__read(&session->header, output); if (err < 0) diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 78f9c45..1fc018e 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -1191,11 +1191,6 @@ static int __cmd_top(void) if (session == NULL) return -ENOMEM; - if (perf_session__create_kernel_maps(session) < 0) { - pr_err("Problems creating kernel maps\n"); - return -1; - } - if (target_pid != -1) event__synthesize_thread(target_pid, event__process, session); else diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 1951e33..8e7c189 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -70,8 +70,17 @@ struct perf_session *perf_session__new(const char *filename, int mode, bool forc self->unknown_events = 0; map_groups__init(&self->kmaps); - if (mode == O_RDONLY && perf_session__open(self, force) < 0) - goto out_delete; + if (mode == O_RDONLY) { + if (perf_session__open(self, force) < 0) + goto out_delete; + } else if (mode == O_WRONLY) { + /* + * In O_RDONLY mode this will be performed when reading the + * kernel MMAP event, in event__process_mmap(). + */ + if (perf_session__create_kernel_maps(self) < 0) + goto out_delete; + } self->sample_type = perf_header__sample_type(&self->header); out: -- cgit v0.10.2 From 4c574159d03f4d8a136a7adff2d0b1d82cadcb18 Mon Sep 17 00:00:00 2001 From: Thiago Farina Date: Wed, 27 Jan 2010 21:05:55 -0200 Subject: tools/perf/perf.c: Clean up trivial style issues Checked with: ./../scripts/checkpatch.pl --terse --file perf.c perf.c: 51: ERROR: open brace '{' following function declarations go on the next line perf.c: 73: ERROR: "foo*** bar" should be "foo ***bar" perf.c:112: ERROR: space prohibited before that close parenthesis ')' perf.c:127: ERROR: space prohibited before that close parenthesis ')' perf.c:171: ERROR: "foo** bar" should be "foo **bar" perf.c:213: ERROR: "(foo*)" should be "(foo *)" perf.c:216: ERROR: "(foo*)" should be "(foo *)" perf.c:217: ERROR: space required before that '*' (ctx:OxV) perf.c:452: ERROR: do not initialise statics to 0 or NULL perf.c:453: ERROR: do not initialise statics to 0 or NULL Signed-off-by: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker Cc: Masami Hiramatsu LKML-Reference: <1264633557-17597-7-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/perf.c b/tools/perf/perf.c index 05c861c..109b89b 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -48,7 +48,8 @@ int check_pager_config(const char *cmd) return c.val; } -static void commit_pager_choice(void) { +static void commit_pager_choice(void) +{ switch (use_pager) { case 0: setenv("PERF_PAGER", "cat", 1); @@ -70,7 +71,7 @@ static void set_debugfs_path(void) "tracing/events"); } -static int handle_options(const char*** argv, int* argc, int* envchanged) +static int handle_options(const char ***argv, int *argc, int *envchanged) { int handled = 0; @@ -109,7 +110,7 @@ static int handle_options(const char*** argv, int* argc, int* envchanged) *envchanged = 1; } else if (!strcmp(cmd, "--perf-dir")) { if (*argc < 2) { - fprintf(stderr, "No directory given for --perf-dir.\n" ); + fprintf(stderr, "No directory given for --perf-dir.\n"); usage(perf_usage_string); } setenv(PERF_DIR_ENVIRONMENT, (*argv)[1], 1); @@ -124,7 +125,7 @@ static int handle_options(const char*** argv, int* argc, int* envchanged) *envchanged = 1; } else if (!strcmp(cmd, "--work-tree")) { if (*argc < 2) { - fprintf(stderr, "No directory given for --work-tree.\n" ); + fprintf(stderr, "No directory given for --work-tree.\n"); usage(perf_usage_string); } setenv(PERF_WORK_TREE_ENVIRONMENT, (*argv)[1], 1); @@ -168,7 +169,7 @@ static int handle_alias(int *argcp, const char ***argv) { int envchanged = 0, ret = 0, saved_errno = errno; int count, option_count; - const char** new_argv; + const char **new_argv; const char *alias_command; char *alias_string; @@ -210,11 +211,11 @@ static int handle_alias(int *argcp, const char ***argv) if (!strcmp(alias_command, new_argv[0])) die("recursive alias: %s", alias_command); - new_argv = realloc(new_argv, sizeof(char*) * + new_argv = realloc(new_argv, sizeof(char *) * (count + *argcp + 1)); /* insert after command name */ - memcpy(new_argv + count, *argv + 1, sizeof(char*) * *argcp); - new_argv[count+*argcp] = NULL; + memcpy(new_argv + count, *argv + 1, sizeof(char *) * *argcp); + new_argv[count + *argcp] = NULL; *argv = new_argv; *argcp += count - 1; @@ -450,8 +451,8 @@ int main(int argc, const char **argv) setup_path(); while (1) { - static int done_help = 0; - static int was_alias = 0; + static int done_help; + static int was_alias; was_alias = run_argv(&argc, &argv); if (errno != ENOENT) -- cgit v0.10.2 From 6a1b751fb89b61ef7240f2e3ed65a2e2776e7cfd Mon Sep 17 00:00:00 2001 From: John Kacur Date: Wed, 27 Jan 2010 21:05:54 -0200 Subject: perf: Ignore perf-archive temp file Tell git to ignore perf-archive. Signed-off-by: John Kacur Signed-off-by: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <1264633557-17597-6-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/.gitignore b/tools/perf/.gitignore index 124760b..e1d60d7 100644 --- a/tools/perf/.gitignore +++ b/tools/perf/.gitignore @@ -14,6 +14,7 @@ perf*.html common-cmds.h perf.data perf.data.old +perf-archive tags TAGS cscope* -- cgit v0.10.2 From a45635dfb08a1fa2cf77bf1f2c4074961ce2e625 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 29 Jan 2010 22:19:04 +0900 Subject: sh: Reworked SH7780 PCI initialization. This consolidates the PCI initialization code for all of the pci-sh7780 users, and sets up the memory window dynamically as opposed to using hardcoded window positions. A number of bugs were fixed at the same time, including the PIO handling and master abort timeout settings being incorrect. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/fixups-r7780rp.c b/arch/sh/drivers/pci/fixups-r7780rp.c index 15ca65c..08b2d86 100644 --- a/arch/sh/drivers/pci/fixups-r7780rp.c +++ b/arch/sh/drivers/pci/fixups-r7780rp.c @@ -22,15 +22,3 @@ int __init pcibios_map_platform_irq(struct pci_dev *pdev, u8 slot, u8 pin) { return irq_tab[slot]; } - -int pci_fixup_pcic(struct pci_channel *chan) -{ - pci_write_reg(chan, 0x000043ff, SH4_PCIINTM); - pci_write_reg(chan, 0x00000000, SH7780_PCIIBAR); - pci_write_reg(chan, 0x08000000, SH7780_PCICSCR0); - pci_write_reg(chan, 0x0000001b, SH7780_PCICSAR0); - pci_write_reg(chan, 0xfd000000, SH7780_PCICSCR1); - pci_write_reg(chan, 0x0000000f, SH7780_PCICSAR1); - - return 0; -} diff --git a/arch/sh/drivers/pci/fixups-sdk7780.c b/arch/sh/drivers/pci/fixups-sdk7780.c index 250b0ed..0930f98 100644 --- a/arch/sh/drivers/pci/fixups-sdk7780.c +++ b/arch/sh/drivers/pci/fixups-sdk7780.c @@ -31,22 +31,3 @@ int __init pcibios_map_platform_irq(struct pci_dev *pdev, u8 slot, u8 pin) { return sdk7780_irq_tab[pin-1][slot]; } -int pci_fixup_pcic(struct pci_channel *chan) -{ - /* Enable all interrupts, so we know what to fix */ - pci_write_reg(chan, 0x0000C3FF, SH7780_PCIIMR); - - /* Set up standard PCI config registers */ - pci_write_reg(chan, 0x08000000, SH7780_PCIMBAR0); /* PCI */ - pci_write_reg(chan, 0x08000000, SH4_PCILAR0); /* SHwy */ - pci_write_reg(chan, 0x07F00001, SH4_PCILSR0); /* size 128M w/ MBAR */ - - pci_write_reg(chan, 0x00000000, SH7780_PCIMBAR1); - pci_write_reg(chan, 0x00000000, SH4_PCILAR1); - pci_write_reg(chan, 0x00000000, SH4_PCILSR1); - - pci_write_reg(chan, 0xAB000801, SH7780_PCIIBAR); - pci_write_reg(chan, 0xA5000C01, SH4_PCICR); - - return 0; -} diff --git a/arch/sh/drivers/pci/pci-sh4.h b/arch/sh/drivers/pci/pci-sh4.h index 4744a7d..43dddd8 100644 --- a/arch/sh/drivers/pci/pci-sh4.h +++ b/arch/sh/drivers/pci/pci-sh4.h @@ -49,6 +49,17 @@ #define SH4_PCIINT_MWPD 0x00000002 /* Master Write PERR Detect */ #define SH4_PCIINT_MRPD 0x00000001 /* Master Read PERR Detect */ #define SH4_PCIINTM 0x118 /* PCI Interrupt Mask */ + #define SH4_PCIINTM_TTADIM BIT(14) /* Target-target abort interrupt */ + #define SH4_PCIINTM_TMTOIM BIT(9) /* Target retry timeout */ + #define SH4_PCIINTM_MDEIM BIT(8) /* Master function disable error */ + #define SH4_PCIINTM_APEDIM BIT(7) /* Address parity error detection */ + #define SH4_PCIINTM_SDIM BIT(6) /* SERR detection */ + #define SH4_PCIINTM_DPEITWM BIT(5) /* Data parity error for target write */ + #define SH4_PCIINTM_PEDITRM BIT(4) /* PERR detection for target read */ + #define SH4_PCIINTM_TADIMM BIT(3) /* Target abort for master */ + #define SH4_PCIINTM_MADIMM BIT(2) /* Master abort for master */ + #define SH4_PCIINTM_MWPDIM BIT(1) /* Master write data parity error */ + #define SH4_PCIINTM_MRDPEIM BIT(0) /* Master read data parity error */ #define SH4_PCIALR 0x11C /* Error Address Register */ #define SH4_PCICLR 0x120 /* Error Command/Data */ #define SH4_PCICLR_MPIO 0x80000000 @@ -61,7 +72,7 @@ #define SH4_PCIAINT 0x130 /* Arbiter Interrupt Register */ #define SH4_PCIAINT_MBKN 0x00002000 /* Master Broken Interrupt */ #define SH4_PCIAINT_TBTO 0x00001000 /* Target Bus Time Out */ - #define SH4_PCIAINT_MBTO 0x00001000 /* Master Bus Time Out */ + #define SH4_PCIAINT_MBTO 0x00000800 /* Master Bus Time Out */ #define SH4_PCIAINT_TABT 0x00000008 /* Target Abort */ #define SH4_PCIAINT_MABT 0x00000004 /* Master Abort */ #define SH4_PCIAINT_RDPE 0x00000002 /* Read Data Parity Error */ diff --git a/arch/sh/drivers/pci/pci-sh7780.c b/arch/sh/drivers/pci/pci-sh7780.c index 323b92d..019e1af 100644 --- a/arch/sh/drivers/pci/pci-sh7780.c +++ b/arch/sh/drivers/pci/pci-sh7780.c @@ -1,7 +1,7 @@ /* * Low-Level PCI Support for the SH7780 * - * Copyright (C) 2005 - 2009 Paul Mundt + * Copyright (C) 2005 - 2010 Paul Mundt * * This file is subject to the terms and conditions of the GNU General Public * License. See the file "COPYING" in the main directory of this archive @@ -14,11 +14,13 @@ #include #include #include "pci-sh4.h" +#include +#include static struct resource sh7785_io_resource = { .name = "SH7785_IO", - .start = SH7780_PCI_IO_BASE, - .end = SH7780_PCI_IO_BASE + SH7780_PCI_IO_SIZE - 1, + .start = 0x1000, + .end = SH7780_PCI_IO_SIZE - 1, .flags = IORESOURCE_IO }; @@ -38,25 +40,14 @@ static struct pci_channel sh7780_pci_controller = { .io_map_base = SH7780_PCI_IO_BASE, }; -static struct sh4_pci_address_map sh7780_pci_map = { - .window0 = { -#if defined(CONFIG_32BIT) - .base = SH7780_32BIT_DDR_BASE_ADDR, - .size = 0x40000000, -#else - .base = SH7780_CS0_BASE_ADDR, - .size = 0x20000000, -#endif - }, -}; - static int __init sh7780_pci_init(void) { struct pci_channel *chan = &sh7780_pci_controller; + phys_addr_t memphys; + size_t memsize; unsigned int id; - const char *type = NULL; + const char *type; int ret; - u32 word; printk(KERN_NOTICE "PCI: Starting intialization.\n"); @@ -65,17 +56,24 @@ static int __init sh7780_pci_init(void) /* Enable CPU access to the PCIC registers. */ __raw_writel(PCIECR_ENBL, PCIECR); - id = __raw_readw(chan->reg_base + SH7780_PCIVID); - if (id != SH7780_VENDOR_ID) { + /* Reset */ + __raw_writel(SH4_PCICR_PREFIX | SH4_PCICR_PRST, + chan->reg_base + SH4_PCICR); + + /* Wait for it to come back up.. */ + mdelay(100); + + id = __raw_readw(chan->reg_base + PCI_VENDOR_ID); + if (id != PCI_VENDOR_ID_RENESAS) { printk(KERN_ERR "PCI: Unknown vendor ID 0x%04x.\n", id); return -ENODEV; } - id = __raw_readw(chan->reg_base + SH7780_PCIDID); - type = (id == SH7763_DEVICE_ID) ? "SH7763" : - (id == SH7780_DEVICE_ID) ? "SH7780" : - (id == SH7781_DEVICE_ID) ? "SH7781" : - (id == SH7785_DEVICE_ID) ? "SH7785" : + id = __raw_readw(chan->reg_base + PCI_DEVICE_ID); + type = (id == PCI_DEVICE_ID_RENESAS_SH7763) ? "SH7763" : + (id == PCI_DEVICE_ID_RENESAS_SH7780) ? "SH7780" : + (id == PCI_DEVICE_ID_RENESAS_SH7781) ? "SH7781" : + (id == PCI_DEVICE_ID_RENESAS_SH7785) ? "SH7785" : NULL; if (unlikely(!type)) { printk(KERN_ERR "PCI: Found an unsupported Renesas host " @@ -85,59 +83,78 @@ static int __init sh7780_pci_init(void) printk(KERN_NOTICE "PCI: Found a Renesas %s host " "controller, revision %d.\n", type, - __raw_readb(chan->reg_base + SH7780_PCIRID)); + __raw_readb(chan->reg_base + PCI_REVISION_ID)); if ((ret = sh4_pci_check_direct(chan)) != 0) return ret; /* - * Set the class and sub-class codes. + * Now throw it in to register initialization mode and + * start the real work. */ - __raw_writeb(PCI_CLASS_BRIDGE_HOST >> 8, - chan->reg_base + SH7780_PCIBCC); - __raw_writeb(PCI_CLASS_BRIDGE_HOST & 0xff, - chan->reg_base + SH7780_PCISUB); + __raw_writel(SH4_PCICR_PREFIX, chan->reg_base + SH4_PCICR); + + memphys = __pa(memory_start); + memsize = memory_end - memory_start; /* * Set IO and Mem windows to local address * Make PCI and local address the same for easy 1 to 1 mapping */ - pci_write_reg(chan, sh7780_pci_map.window0.size - 0xfffff, SH4_PCILSR0); - /* Set the values on window 0 PCI config registers */ - pci_write_reg(chan, sh7780_pci_map.window0.base, SH4_PCILAR0); - pci_write_reg(chan, sh7780_pci_map.window0.base, SH7780_PCIMBAR0); - - pci_write_reg(chan, 0x0000380f, SH4_PCIAINTM); - - /* Set up standard PCI config registers */ - __raw_writew(0xFB00, chan->reg_base + SH7780_PCISTATUS); - __raw_writew(0x0047, chan->reg_base + SH7780_PCICMD); - __raw_writew(0x1912, chan->reg_base + SH7780_PCISVID); - __raw_writew(0x0001, chan->reg_base + SH7780_PCISID); - - __raw_writeb(0x00, chan->reg_base + SH7780_PCIPIF); - - /* Apply any last-minute PCIC fixups */ - pci_fixup_pcic(chan); - - pci_write_reg(chan, 0xfd000000, SH7780_PCIMBR0); - pci_write_reg(chan, 0x00fc0000, SH7780_PCIMBMR0); - -#ifdef CONFIG_32BIT - pci_write_reg(chan, 0xc0000000, SH7780_PCIMBR2); - pci_write_reg(chan, 0x20000000 - SH7780_PCI_IO_SIZE, SH7780_PCIMBMR2); -#endif - - /* Set IOBR for windows containing area specified in pci.h */ - pci_write_reg(chan, chan->io_resource->start & ~(SH7780_PCI_IO_SIZE-1), - SH7780_PCIIOBR); - pci_write_reg(chan, ((SH7780_PCI_IO_SIZE-1) & (7<<18)), - SH7780_PCIIOBMR); - - /* SH7780 init done, set central function init complete */ - /* use round robin mode to stop a device starving/overruning */ - word = SH4_PCICR_PREFIX | SH4_PCICR_CFIN | SH4_PCICR_FTO; - pci_write_reg(chan, word, SH4_PCICR); + __raw_writel(0, chan->reg_base + PCI_BASE_ADDRESS_0); + + __raw_writel(memphys, chan->reg_base + SH4_PCILAR0); + __raw_writel((memsize - 1) << 9 | 1, + chan->reg_base + SH4_PCILSR0); + + /* Clear out PCI arbiter IRQs */ + __raw_writel(0, chan->reg_base + SH4_PCIAINT); + + /* Unmask all of the arbiter IRQs. */ + __raw_writel(SH4_PCIAINT_MBKN | SH4_PCIAINT_TBTO | SH4_PCIAINT_MBTO | \ + SH4_PCIAINT_TABT | SH4_PCIAINT_MABT | SH4_PCIAINT_RDPE | \ + SH4_PCIAINT_WDPE, chan->reg_base + SH4_PCIAINTM); + + /* Clear all error conditions */ + __raw_writew(PCI_STATUS_DETECTED_PARITY | \ + PCI_STATUS_SIG_SYSTEM_ERROR | \ + PCI_STATUS_REC_MASTER_ABORT | \ + PCI_STATUS_REC_TARGET_ABORT | \ + PCI_STATUS_SIG_TARGET_ABORT | \ + PCI_STATUS_PARITY, chan->reg_base + PCI_STATUS); + + __raw_writew(PCI_COMMAND_SERR | PCI_COMMAND_WAIT | \ + PCI_COMMAND_PARITY | PCI_COMMAND_MASTER | \ + PCI_COMMAND_MEMORY, chan->reg_base + PCI_COMMAND); + + /* Unmask all of the PCI IRQs */ + __raw_writel(SH4_PCIINTM_TTADIM | SH4_PCIINTM_TMTOIM | \ + SH4_PCIINTM_MDEIM | SH4_PCIINTM_APEDIM | \ + SH4_PCIINTM_SDIM | SH4_PCIINTM_DPEITWM | \ + SH4_PCIINTM_PEDITRM | SH4_PCIINTM_TADIMM | \ + SH4_PCIINTM_MADIMM | SH4_PCIINTM_MWPDIM | \ + SH4_PCIINTM_MRDPEIM, chan->reg_base + SH4_PCIINTM); + + /* + * Disable the cache snoop controller for non-coherent DMA. + */ + __raw_writel(0, chan->reg_base + SH7780_PCICSCR0); + __raw_writel(0, chan->reg_base + SH7780_PCICSAR0); + __raw_writel(0, chan->reg_base + SH7780_PCICSCR1); + __raw_writel(0, chan->reg_base + SH7780_PCICSAR1); + + __raw_writel(0xfd000000, chan->reg_base + SH7780_PCIMBR0); + __raw_writel(0x00fc0000, chan->reg_base + SH7780_PCIMBMR0); + + __raw_writel(0, chan->reg_base + SH7780_PCIIOBR); + __raw_writel(0, chan->reg_base + SH7780_PCIIOBMR); + + /* + * Initialization mode complete, release the control register and + * enable round robin mode to stop device overruns/starvation. + */ + __raw_writel(SH4_PCICR_PREFIX | SH4_PCICR_CFIN | SH4_PCICR_FTO, + chan->reg_base + SH4_PCICR); register_pci_controller(chan); diff --git a/arch/sh/drivers/pci/pci-sh7780.h b/arch/sh/drivers/pci/pci-sh7780.h index 4a52478..dee069c 100644 --- a/arch/sh/drivers/pci/pci-sh7780.h +++ b/arch/sh/drivers/pci/pci-sh7780.h @@ -12,12 +12,11 @@ #ifndef _PCI_SH7780_H_ #define _PCI_SH7780_H_ -/* Platform Specific Values */ -#define SH7780_VENDOR_ID 0x1912 -#define SH7781_DEVICE_ID 0x0001 -#define SH7780_DEVICE_ID 0x0002 -#define SH7763_DEVICE_ID 0x0004 -#define SH7785_DEVICE_ID 0x0007 +#define PCI_VENDOR_ID_RENESAS 0x1912 +#define PCI_DEVICE_ID_RENESAS_SH7781 0x0001 +#define PCI_DEVICE_ID_RENESAS_SH7780 0x0002 +#define PCI_DEVICE_ID_RENESAS_SH7763 0x0004 +#define PCI_DEVICE_ID_RENESAS_SH7785 0x0007 /* SH7780 Control Registers */ #define PCIECR 0xFE000008 @@ -36,35 +35,6 @@ #define SH7780_PCIREG_BASE 0xFE040000 /* PCI regs base address */ /* SH7780 PCI Config Registers */ -#define SH7780_PCIVID 0x000 /* Vendor ID */ -#define SH7780_PCIDID 0x002 /* Device ID */ -#define SH7780_PCICMD 0x004 /* Command */ -#define SH7780_PCISTATUS 0x006 /* Status */ -#define SH7780_PCIRID 0x008 /* Revision ID */ -#define SH7780_PCIPIF 0x009 /* Program Interface */ -#define SH7780_PCISUB 0x00a /* Sub class code */ -#define SH7780_PCIBCC 0x00b /* Base class code */ -#define SH7780_PCICLS 0x00c /* Cache line size */ -#define SH7780_PCILTM 0x00d /* latency timer */ -#define SH7780_PCIHDR 0x00e /* Header type */ -#define SH7780_PCIBIST 0x00f /* BIST */ -#define SH7780_PCIIBAR 0x010 /* IO Base address */ -#define SH7780_PCIMBAR0 0x014 /* Memory base address0 */ -#define SH7780_PCIMBAR1 0x018 /* Memory base address1 */ -#define SH7780_PCISVID 0x02c /* Sub system vendor ID */ -#define SH7780_PCISID 0x02e /* Sub system ID */ -#define SH7780_PCICP 0x034 -#define SH7780_PCIINTLINE 0x03c /* Interrupt line */ -#define SH7780_PCIINTPIN 0x03d /* Interrupt pin */ -#define SH7780_PCIMINGNT 0x03e /* Minumum grand */ -#define SH7780_PCIMAXLAT 0x03f /* Maxmum latency */ -#define SH7780_PCICID 0x040 -#define SH7780_PCINIP 0x041 -#define SH7780_PCIPMC 0x042 -#define SH7780_PCIPMCSR 0x044 -#define SH7780_PCIPMCSR_BSE 0x046 -#define SH7780_PCICDD 0x047 - #define SH7780_PCIIR 0x114 /* PCI Interrupt Register */ #define SH7780_PCIIMR 0x118 /* PCI Interrupt Mask Register */ #define SH7780_PCIAIR 0x11C /* Error Address Register */ @@ -78,6 +48,8 @@ #define SH7780_PCIMBR0 0x1E0 #define SH7780_PCIMBMR0 0x1E4 +#define SH7780_PCIMBR1 0x1E8 +#define SH7780_PCIMBMR1 0x1EC #define SH7780_PCIMBR2 0x1F0 #define SH7780_PCIMBMR2 0x1F4 #define SH7780_PCIIOBR 0x1F8 @@ -87,16 +59,4 @@ #define SH7780_PCICSAR0 0x218 /* Cache Snoop1 Addr. Register */ #define SH7780_PCICSAR1 0x21C /* Cache Snoop2 Addr. Register */ -/* General Memory Config Addresses */ -#define SH7780_CS0_BASE_ADDR 0x0 -#define SH7780_MEM_REGION_SIZE 0x04000000 -#define SH7780_CS1_BASE_ADDR (SH7780_CS0_BASE_ADDR + SH7780_MEM_REGION_SIZE) -#define SH7780_CS2_BASE_ADDR (SH7780_CS1_BASE_ADDR + SH7780_MEM_REGION_SIZE) -#define SH7780_CS3_BASE_ADDR (SH7780_CS2_BASE_ADDR + SH7780_MEM_REGION_SIZE) -#define SH7780_CS4_BASE_ADDR (SH7780_CS3_BASE_ADDR + SH7780_MEM_REGION_SIZE) -#define SH7780_CS5_BASE_ADDR (SH7780_CS4_BASE_ADDR + SH7780_MEM_REGION_SIZE) -#define SH7780_CS6_BASE_ADDR (SH7780_CS5_BASE_ADDR + SH7780_MEM_REGION_SIZE) - -#define SH7780_32BIT_DDR_BASE_ADDR 0x40000000 - #endif /* _PCI_SH7780_H_ */ -- cgit v0.10.2 From ac8ab54a8e41a5ed0ee2161d45b6dc855490989f Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 29 Jan 2010 22:22:27 +0900 Subject: sh: Bail out early on PCI resource conflicts. Presently we just call in to request_resource() for the ioport and iomem resources without checking for errors. This has already hidden a couple of bugs, so add some error handling in for good measure. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/pci.c b/arch/sh/drivers/pci/pci.c index 191075e..82e59bc 100644 --- a/arch/sh/drivers/pci/pci.c +++ b/arch/sh/drivers/pci/pci.c @@ -53,8 +53,12 @@ static DEFINE_MUTEX(pci_scan_mutex); void __devinit register_pci_controller(struct pci_channel *hose) { - request_resource(&iomem_resource, hose->mem_resource); - request_resource(&ioport_resource, hose->io_resource); + if (request_resource(&iomem_resource, hose->mem_resource) < 0) + goto out; + if (request_resource(&ioport_resource, hose->io_resource) < 0) { + release_resource(hose->mem_resource); + goto out; + } *hose_tail = hose; hose_tail = &hose->next; @@ -76,6 +80,9 @@ void __devinit register_pci_controller(struct pci_channel *hose) pcibios_scanbus(hose); mutex_unlock(&pci_scan_mutex); } + +out: + printk(KERN_WARNING "Skipping PCI bus scan due to resource conflict\n"); } static int __init pcibios_init(void) @@ -319,20 +326,9 @@ void __iomem *pci_iomap(struct pci_dev *dev, int bar, unsigned long maxlen) if (flags & IORESOURCE_IO) return ioport_map_pci(dev, start, len); - - /* - * Presently the IORESOURCE_MEM case is a bit special, most - * SH7751 style PCI controllers have PCI memory at a fixed - * location in the address space where no remapping is desired. - * With the IORESOURCE_MEM case more care has to be taken - * to inhibit page table mapping for legacy cores, but this is - * punted off to __ioremap(). - * -- PFM. - */ if (flags & IORESOURCE_MEM) { if (flags & IORESOURCE_CACHEABLE) return ioremap(start, len); - return ioremap_nocache(start, len); } -- cgit v0.10.2 From 320e68da59353fe6ad51b81f6865c4b674ad66ea Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 29 Jan 2010 22:38:13 +0900 Subject: sh: support PCI domains. Newer SH parts are now commonly shipping with multiple controllers, so we wire up PCI domain support to deal with them. Shamelessly cloned from the MIPS implementation. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 352879c..bae5383 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -826,11 +826,15 @@ config MAPLE config PCI bool "PCI support" depends on SYS_SUPPORTS_PCI + select PCI_DOMAINS help Find out whether you have a PCI motherboard. PCI is the name of a bus system, i.e. the way the CPU talks to the other stuff inside your box. If you have PCI, say Y, otherwise N. +config PCI_DOMAINS + bool + source "drivers/pci/pcie/Kconfig" source "drivers/pci/Kconfig" diff --git a/arch/sh/drivers/pci/pci.c b/arch/sh/drivers/pci/pci.c index 82e59bc..45a15ca 100644 --- a/arch/sh/drivers/pci/pci.c +++ b/arch/sh/drivers/pci/pci.c @@ -33,15 +33,22 @@ static int pci_initialized; static void __devinit pcibios_scanbus(struct pci_channel *hose) { static int next_busno; + static int need_domain_info; struct pci_bus *bus; bus = pci_scan_bus(next_busno, hose->pci_ops, hose); + hose->bus = bus; + + need_domain_info = need_domain_info || hose->index; + hose->need_domain_info = need_domain_info; if (bus) { next_busno = bus->subordinate + 1; /* Don't allow 8-bit bus number overflow inside the hose - reserve some space for bridges. */ - if (next_busno > 224) + if (next_busno > 224) { next_busno = 0; + need_domain_info = 1; + } pci_bus_size_bridges(bus); pci_bus_assign_resources(bus); @@ -307,9 +314,15 @@ static void __iomem *ioport_map_pci(struct pci_dev *dev, { struct pci_channel *chan = dev->sysdata; - if (!chan->io_map_base) + if (unlikely(!chan->io_map_base)) { chan->io_map_base = generic_io_base; + if (pci_domains_supported) + panic("To avoid data corruption io_map_base MUST be " + "set with multiple PCI domains."); + } + + return (void __iomem *)(chan->io_map_base + port); } diff --git a/arch/sh/include/asm/pci.h b/arch/sh/include/asm/pci.h index f362d8a..d124a00 100644 --- a/arch/sh/include/asm/pci.h +++ b/arch/sh/include/asm/pci.h @@ -15,6 +15,7 @@ */ struct pci_channel { struct pci_channel *next; + struct pci_bus *bus; struct pci_ops *pci_ops; struct resource *io_resource; @@ -24,8 +25,10 @@ struct pci_channel { unsigned long mem_offset; unsigned long reg_base; - unsigned long io_map_base; + + unsigned int index; + unsigned int need_domain_info; }; extern void register_pci_controller(struct pci_channel *hose); @@ -108,6 +111,14 @@ extern void pcibios_resource_to_bus(struct pci_dev *dev, extern void pcibios_bus_to_resource(struct pci_dev *dev, struct resource *res, struct pci_bus_region *region); +#define pci_domain_nr(bus) ((struct pci_channel *)(bus)->sysdata)->index + +static inline int pci_proc_domain(struct pci_bus *bus) +{ + struct pci_channel *hose = bus->sysdata; + return hose->need_domain_info; +} + /* Chances are this interrupt is wired PC-style ... */ static inline int pci_get_legacy_ide_irq(struct pci_dev *dev, int channel) { -- cgit v0.10.2 From 396c56a9c69ebb0baf9171a6365ac9fda322728d Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Sat, 30 Jan 2010 01:41:21 +0900 Subject: sh: Kill off broken type 1 PCI config access checks. The host controllers only support type 1, so there's not much else to test for. Some of the older controllers also supported type 2 accesses, but we've never supported those, and likely never will. Beyond that, the P1SEG test is meaningless for 32-bit mode, so rather than refactoring it, just kill the type 1 test off completely. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/ops-sh4.c b/arch/sh/drivers/pci/ops-sh4.c index 78bebeb..e55e81a 100644 --- a/arch/sh/drivers/pci/ops-sh4.c +++ b/arch/sh/drivers/pci/ops-sh4.c @@ -102,34 +102,6 @@ struct pci_ops sh4_pci_ops = { .write = sh4_pci_write, }; -/* - * Not really related to pci_ops, but it's common and not worth shoving - * somewhere else for now.. - */ -int __init sh4_pci_check_direct(struct pci_channel *chan) -{ - /* - * Check if configuration works. - */ - unsigned int tmp = pci_read_reg(chan, SH4_PCIPAR); - - pci_write_reg(chan, P1SEG, SH4_PCIPAR); - - if (pci_read_reg(chan, SH4_PCIPAR) == P1SEG) { - pci_write_reg(chan, tmp, SH4_PCIPAR); - printk(KERN_INFO "PCI: Using configuration type 1\n"); - request_region(chan->reg_base + SH4_PCIPAR, 8, - "PCI conf1"); - return 0; - } - - pci_write_reg(chan, tmp, SH4_PCIPAR); - - printk(KERN_ERR "PCI: %s failed\n", __func__); - - return -EINVAL; -} - int __attribute__((weak)) pci_fixup_pcic(struct pci_channel *chan) { /* Nothing to do. */ diff --git a/arch/sh/drivers/pci/pci-sh4.h b/arch/sh/drivers/pci/pci-sh4.h index 43dddd8..cbf763b 100644 --- a/arch/sh/drivers/pci/pci-sh4.h +++ b/arch/sh/drivers/pci/pci-sh4.h @@ -162,7 +162,6 @@ /* arch/sh/kernel/drivers/pci/ops-sh4.c */ extern struct pci_ops sh4_pci_ops; -int sh4_pci_check_direct(struct pci_channel *chan); int pci_fixup_pcic(struct pci_channel *chan); struct sh4_pci_address_space { diff --git a/arch/sh/drivers/pci/pci-sh7751.c b/arch/sh/drivers/pci/pci-sh7751.c index 2455cf3..02306dd 100644 --- a/arch/sh/drivers/pci/pci-sh7751.c +++ b/arch/sh/drivers/pci/pci-sh7751.c @@ -79,7 +79,6 @@ static int __init sh7751_pci_init(void) struct pci_channel *chan = &sh7751_pci_controller; unsigned int id; u32 word, reg; - int ret; printk(KERN_NOTICE "PCI: Starting intialization.\n"); @@ -93,9 +92,6 @@ static int __init sh7751_pci_init(void) return -ENODEV; } - if ((ret = sh4_pci_check_direct(chan)) != 0) - return ret; - /* Set the BCR's to enable PCI access */ reg = __raw_readl(SH7751_BCR1); reg |= 0x80000; diff --git a/arch/sh/drivers/pci/pci-sh7780.c b/arch/sh/drivers/pci/pci-sh7780.c index 019e1af..40531cd 100644 --- a/arch/sh/drivers/pci/pci-sh7780.c +++ b/arch/sh/drivers/pci/pci-sh7780.c @@ -47,7 +47,6 @@ static int __init sh7780_pci_init(void) size_t memsize; unsigned int id; const char *type; - int ret; printk(KERN_NOTICE "PCI: Starting intialization.\n"); @@ -85,9 +84,6 @@ static int __init sh7780_pci_init(void) "controller, revision %d.\n", type, __raw_readb(chan->reg_base + PCI_REVISION_ID)); - if ((ret = sh4_pci_check_direct(chan)) != 0) - return ret; - /* * Now throw it in to register initialization mode and * start the real work. -- cgit v0.10.2 From 97b19778ee5dd0484493f12f6a9cf9391cf0aefb Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Sat, 30 Jan 2010 11:04:38 +0900 Subject: sh: mach-se: Fix up irq_desc reference. The irq_desc needs to be accessed with irq_to_desc(), this fixes up a build error with irq_desc being undefined. Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-se/7206/irq.c b/arch/sh/boards/mach-se/7206/irq.c index 79be4bc..8d82175 100644 --- a/arch/sh/boards/mach-se/7206/irq.c +++ b/arch/sh/boards/mach-se/7206/irq.c @@ -89,8 +89,9 @@ static void enable_se7206_irq(unsigned int irq) static void eoi_se7206_irq(unsigned int irq) { unsigned short sts0,sts1; + struct irq_desc *desc = irq_to_desc(irq); - if (!(irq_desc[irq].status & (IRQ_DISABLED|IRQ_INPROGRESS))) + if (!(desc->status & (IRQ_DISABLED|IRQ_INPROGRESS))) enable_se7206_irq(irq); /* FPGA isr clear */ sts0 = __raw_readw(INTSTS0); -- cgit v0.10.2 From 31a090ae0693af189fc7a8ad16cece4878acad33 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Sat, 30 Jan 2010 11:29:56 +0900 Subject: sh: Fix up asm/hw_breakpoint.h header check. Presently headers_check complains about linux/kdebug.h being unexported, so just bump the __KERNEL__ ifdef up, as per the x86 change. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/hw_breakpoint.h b/arch/sh/include/asm/hw_breakpoint.h index 7295d62..965dd78 100644 --- a/arch/sh/include/asm/hw_breakpoint.h +++ b/arch/sh/include/asm/hw_breakpoint.h @@ -1,12 +1,12 @@ #ifndef __ASM_SH_HW_BREAKPOINT_H #define __ASM_SH_HW_BREAKPOINT_H -#include -#include - #ifdef __KERNEL__ #define __ARCH_HW_BREAKPOINT_H +#include +#include + struct arch_hw_breakpoint { char *name; /* Contains name of the symbol to set bkpt */ unsigned long address; -- cgit v0.10.2 From a8e6f734ce9a79d44ebb296f2a341f435227b34e Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Sat, 30 Jan 2010 20:55:41 +0900 Subject: Revert "perf record: Intercept all events" This reverts commit f5a2c3dce03621b55f84496f58adc2d1a87ca16f. This patch is required for making "perf lock rec" work. The commit f5a2c3dce0 changes write_event() of builtin-record.c . And changed write_event() sometimes doesn't stop with perf lock rec. Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: [ that commit also causes perf record to not be Ctrl-C-able, and it's concetually wrong to parse the data at record time (unconditionally - even when not needed), as we eventually want to be able to do zero-copy recording, at least for non-archive recordings. ] Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 9034522..eea5691 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -113,24 +113,16 @@ static void write_output(void *buf, size_t size) static void write_event(event_t *buf, size_t size) { - size_t processed_size = buf->header.size; - event_t *ev = buf; - - do { - /* - * Add it to the list of DSOs, so that when we finish this - * record session we can pick the available build-ids. - */ - if (ev->header.type == PERF_RECORD_MMAP) { - struct list_head *head = &dsos__user; - if (ev->header.misc == 1) - head = &dsos__kernel; - __dsos__findnew(head, ev->mmap.filename); - } - - ev = ((void *)ev) + ev->header.size; - processed_size += ev->header.size; - } while (processed_size < size); + /* + * Add it to the list of DSOs, so that when we finish this + * record session we can pick the available build-ids. + */ + if (buf->header.type == PERF_RECORD_MMAP) { + struct list_head *head = &dsos__user; + if (buf->mmap.header.misc == 1) + head = &dsos__kernel; + __dsos__findnew(head, buf->mmap.filename); + } write_output(buf, size); } -- cgit v0.10.2 From 86d8d29634de4464d568e7c335c0da6cba64e8ab Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Sat, 30 Jan 2010 20:43:23 +0900 Subject: perf tools: Add __data_loc support This patch is required to test the next patch for perf lock. At 064739bc4b3d7f424b2f25547e6611bcf0132415 , support for the modifier "__data_loc" of format is added. But, when I wanted to parse format of lock_acquired (or some event else), raw_field_ptr() did not returned correct pointer. So I modified raw_field_ptr() like this patch. Then raw_field_ptr() works well. Signed-off-by: Hitoshi Mitake Acked-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Tom Zanussi Cc: Steven Rostedt LKML-Reference: <1264851813-8413-2-git-send-email-mitake@dcl.info.waseda.ac.jp> [ v3: fixed minor stylistic detail ] Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/trace-event-parse.c b/tools/perf/util/trace-event-parse.c index c5c32be..c4b3cb8 100644 --- a/tools/perf/util/trace-event-parse.c +++ b/tools/perf/util/trace-event-parse.c @@ -1925,6 +1925,15 @@ void *raw_field_ptr(struct event *event, const char *name, void *data) if (!field) return NULL; + if (field->flags & FIELD_IS_STRING) { + int offset; + + offset = *(int *)(data + field->offset); + offset &= 0xffff; + + return data + offset; + } + return data + field->offset; } -- cgit v0.10.2 From 18e97e06b5fb2d7f6cf272ca07d26d8247db8723 Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Sat, 30 Jan 2010 20:43:24 +0900 Subject: perf: Add util/include/linuxhash.h to include hash.h of kernel linux/hash.h, hash header of kernel, is also useful for perf. util/include/linuxhash.h includes linux/hash.h, so we can use hash facilities (e.g. hash_long()) in perf now. Signed-off-by: Hitoshi Mitake Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker LKML-Reference: <1264851813-8413-3-git-send-email-mitake@dcl.info.waseda.ac.jp> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 9b173e6..b2bce1f 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -333,6 +333,7 @@ LIB_FILE=libperf.a LIB_H += ../../include/linux/perf_event.h LIB_H += ../../include/linux/rbtree.h LIB_H += ../../include/linux/list.h +LIB_H += ../../include/linux/hash.h LIB_H += ../../include/linux/stringify.h LIB_H += util/include/linux/bitmap.h LIB_H += util/include/linux/bitops.h diff --git a/tools/perf/util/include/linux/hash.h b/tools/perf/util/include/linux/hash.h new file mode 100644 index 0000000..201f573 --- /dev/null +++ b/tools/perf/util/include/linux/hash.h @@ -0,0 +1,5 @@ +#include "../../../../include/linux/hash.h" + +#ifndef PERF_HASH_H +#define PERF_HASH_H +#endif -- cgit v0.10.2 From c965be10ca3cb0bdd04016c852764afaf8e647c8 Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Sat, 30 Jan 2010 20:43:32 +0900 Subject: perf lock: Enhance information of lock trace events Add wait time and lock identification details. Signed-off-by: Hitoshi Mitake Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker LKML-Reference: <1264851813-8413-11-git-send-email-mitake@dcl.info.waseda.ac.jp> [ removed the file/line bits as we can do that better via IPs ] Signed-off-by: Ingo Molnar diff --git a/include/trace/events/lock.h b/include/trace/events/lock.h index a870ba1..5c1dcfc 100644 --- a/include/trace/events/lock.h +++ b/include/trace/events/lock.h @@ -20,14 +20,17 @@ TRACE_EVENT(lock_acquire, TP_STRUCT__entry( __field(unsigned int, flags) __string(name, lock->name) + __field(void *, lockdep_addr) ), TP_fast_assign( __entry->flags = (trylock ? 1 : 0) | (read ? 2 : 0); __assign_str(name, lock->name); + __entry->lockdep_addr = lock; ), - TP_printk("%s%s%s", (__entry->flags & 1) ? "try " : "", + TP_printk("%p %s%s%s", __entry->lockdep_addr, + (__entry->flags & 1) ? "try " : "", (__entry->flags & 2) ? "read " : "", __get_str(name)) ); @@ -40,13 +43,16 @@ TRACE_EVENT(lock_release, TP_STRUCT__entry( __string(name, lock->name) + __field(void *, lockdep_addr) ), TP_fast_assign( __assign_str(name, lock->name); + __entry->lockdep_addr = lock; ), - TP_printk("%s", __get_str(name)) + TP_printk("%p %s", + __entry->lockdep_addr, __get_str(name)) ); #ifdef CONFIG_LOCK_STAT @@ -59,13 +65,16 @@ TRACE_EVENT(lock_contended, TP_STRUCT__entry( __string(name, lock->name) + __field(void *, lockdep_addr) ), TP_fast_assign( __assign_str(name, lock->name); + __entry->lockdep_addr = lock; ), - TP_printk("%s", __get_str(name)) + TP_printk("%p %s", + __entry->lockdep_addr, __get_str(name)) ); TRACE_EVENT(lock_acquired, @@ -75,16 +84,18 @@ TRACE_EVENT(lock_acquired, TP_STRUCT__entry( __string(name, lock->name) - __field(unsigned long, wait_usec) - __field(unsigned long, wait_nsec_rem) + __field(s64, wait_nsec) + __field(void *, lockdep_addr) ), + TP_fast_assign( __assign_str(name, lock->name); - __entry->wait_nsec_rem = do_div(waittime, NSEC_PER_USEC); - __entry->wait_usec = (unsigned long) waittime; + __entry->wait_nsec = waittime; + __entry->lockdep_addr = lock; ), - TP_printk("%s (%lu.%03lu us)", __get_str(name), __entry->wait_usec, - __entry->wait_nsec_rem) + TP_printk("%p %s (%llu ns)", __entry->lockdep_addr, + __get_str(name), + __entry->wait_nsec) ); #endif -- cgit v0.10.2 From 9b5e350c7a46a471d5b452836dbafe9aeaeca435 Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Sat, 30 Jan 2010 20:43:33 +0900 Subject: perf lock: Introduce new tool "perf lock", for analyzing lock statistics Adding new subcommand "perf lock" to perf. I have a lot of remaining ToDos, but for now perf lock can already provide minimal functionality for analyzing lock statistics. Signed-off-by: Hitoshi Mitake Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker LKML-Reference: <1264851813-8413-12-git-send-email-mitake@dcl.info.waseda.ac.jp> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index b2bce1f..4296930 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -456,6 +456,7 @@ BUILTIN_OBJS += builtin-top.o BUILTIN_OBJS += builtin-trace.o BUILTIN_OBJS += builtin-probe.o BUILTIN_OBJS += builtin-kmem.o +BUILTIN_OBJS += builtin-lock.o PERFLIBS = $(LIB_FILE) diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c new file mode 100644 index 0000000..2b5f887 --- /dev/null +++ b/tools/perf/builtin-lock.c @@ -0,0 +1,724 @@ +#include "builtin.h" +#include "perf.h" + +#include "util/util.h" +#include "util/cache.h" +#include "util/symbol.h" +#include "util/thread.h" +#include "util/header.h" + +#include "util/parse-options.h" +#include "util/trace-event.h" + +#include "util/debug.h" +#include "util/session.h" + +#include +#include +#include +#include +#include +#include + +#include +#include + +/* based on kernel/lockdep.c */ +#define LOCKHASH_BITS 12 +#define LOCKHASH_SIZE (1UL << LOCKHASH_BITS) + +static struct list_head lockhash_table[LOCKHASH_SIZE]; + +#define __lockhashfn(key) hash_long((unsigned long)key, LOCKHASH_BITS) +#define lockhashentry(key) (lockhash_table + __lockhashfn((key))) + +#define LOCK_STATE_UNLOCKED 0 /* initial state */ +#define LOCK_STATE_LOCKED 1 + +struct lock_stat { + struct list_head hash_entry; + struct rb_node rb; /* used for sorting */ + + /* FIXME: raw_field_value() returns unsigned long long, + * so address of lockdep_map should be dealed as 64bit. + * Is there more better solution? */ + void *addr; /* address of lockdep_map, used as ID */ + char *name; /* for strcpy(), we cannot use const */ + char *file; + unsigned int line; + + int state; + u64 prev_event_time; /* timestamp of previous event */ + + unsigned int nr_acquired; + unsigned int nr_acquire; + unsigned int nr_contended; + unsigned int nr_release; + + /* these times are in nano sec. */ + u64 wait_time_total; + u64 wait_time_min; + u64 wait_time_max; +}; + +/* build simple key function one is bigger than two */ +#define SINGLE_KEY(member) \ + static int lock_stat_key_ ## member(struct lock_stat *one, \ + struct lock_stat *two) \ + { \ + return one->member > two->member; \ + } + +SINGLE_KEY(nr_acquired) +SINGLE_KEY(nr_contended) +SINGLE_KEY(wait_time_total) +SINGLE_KEY(wait_time_min) +SINGLE_KEY(wait_time_max) + +struct lock_key { + /* + * name: the value for specify by user + * this should be simpler than raw name of member + * e.g. nr_acquired -> acquired, wait_time_total -> wait_total + */ + const char *name; + int (*key)(struct lock_stat*, struct lock_stat*); +}; + +static const char *sort_key = "acquired"; +static int (*compare)(struct lock_stat *, struct lock_stat *); + +#define DEF_KEY_LOCK(name, fn_suffix) \ + { #name, lock_stat_key_ ## fn_suffix } +struct lock_key keys[] = { + DEF_KEY_LOCK(acquired, nr_acquired), + DEF_KEY_LOCK(contended, nr_contended), + DEF_KEY_LOCK(wait_total, wait_time_total), + DEF_KEY_LOCK(wait_min, wait_time_min), + DEF_KEY_LOCK(wait_max, wait_time_max), + + /* extra comparisons much complicated should be here */ + + { NULL, NULL } +}; + +static void select_key(void) +{ + int i; + + for (i = 0; keys[i].name; i++) { + if (!strcmp(keys[i].name, sort_key)) { + compare = keys[i].key; + return; + } + } + + die("Unknown compare key:%s\n", sort_key); +} + +static struct rb_root result; /* place to store sorted data */ + +static void insert_to_result(struct lock_stat *st, + int (*bigger)(struct lock_stat *, + struct lock_stat *)) +{ + struct rb_node **rb = &result.rb_node; + struct rb_node *parent = NULL; + struct lock_stat *p; + + while (*rb) { + p = container_of(*rb, struct lock_stat, rb); + parent = *rb; + + if (bigger(st, p)) + rb = &(*rb)->rb_left; + else + rb = &(*rb)->rb_right; + } + + rb_link_node(&st->rb, parent, rb); + rb_insert_color(&st->rb, &result); +} + +/* returns left most element of result, and erase it */ +static struct lock_stat *pop_from_result(void) +{ + struct rb_node *node = result.rb_node; + + if (!node) + return NULL; + + while (node->rb_left) + node = node->rb_left; + + rb_erase(node, &result); + return container_of(node, struct lock_stat, rb); +} + +static struct lock_stat *lock_stat_findnew(void *addr, const char *name, + const char *file, unsigned int line) +{ + struct list_head *entry = lockhashentry(addr); + struct lock_stat *ret, *new; + + list_for_each_entry(ret, entry, hash_entry) { + if (ret->addr == addr) + return ret; + } + + new = zalloc(sizeof(struct lock_stat)); + if (!new) + goto alloc_failed; + + new->addr = addr; + new->name = zalloc(sizeof(char) * strlen(name) + 1); + if (!new->name) + goto alloc_failed; + strcpy(new->name, name); + new->file = zalloc(sizeof(char) * strlen(file) + 1); + if (!new->file) + goto alloc_failed; + strcpy(new->file, file); + new->line = line; + + /* LOCK_STATE_UNLOCKED == 0 isn't guaranteed forever */ + new->state = LOCK_STATE_UNLOCKED; + new->wait_time_min = ULLONG_MAX; + + list_add(&new->hash_entry, entry); + return new; + +alloc_failed: + die("memory allocation failed\n"); +} + +static char const *input_name = "perf.data"; + +static int profile_cpu = -1; + +struct raw_event_sample { + u32 size; + char data[0]; +}; + +struct trace_acquire_event { + void *addr; + const char *name; + const char *file; + unsigned int line; +}; + +struct trace_acquired_event { + void *addr; + const char *name; + const char *file; + unsigned int line; +}; + +struct trace_contended_event { + void *addr; + const char *name; + const char *file; + unsigned int line; +}; + +struct trace_release_event { + void *addr; + const char *name; + const char *file; + unsigned int line; +}; + +struct trace_lock_handler { + void (*acquire_event)(struct trace_acquire_event *, + struct event *, + int cpu, + u64 timestamp, + struct thread *thread); + + void (*acquired_event)(struct trace_acquired_event *, + struct event *, + int cpu, + u64 timestamp, + struct thread *thread); + + void (*contended_event)(struct trace_contended_event *, + struct event *, + int cpu, + u64 timestamp, + struct thread *thread); + + void (*release_event)(struct trace_release_event *, + struct event *, + int cpu, + u64 timestamp, + struct thread *thread); +}; + +static void prof_lock_acquire_event(struct trace_acquire_event *acquire_event, + struct event *__event __used, + int cpu __used, + u64 timestamp, + struct thread *thread __used) +{ + struct lock_stat *st; + + st = lock_stat_findnew(acquire_event->addr, acquire_event->name, + acquire_event->file, acquire_event->line); + + switch (st->state) { + case LOCK_STATE_UNLOCKED: + break; + case LOCK_STATE_LOCKED: + break; + default: + BUG_ON(1); + break; + } + + st->prev_event_time = timestamp; +} + +static void prof_lock_acquired_event(struct trace_acquired_event *acquired_event, + struct event *__event __used, + int cpu __used, + u64 timestamp, + struct thread *thread __used) +{ + struct lock_stat *st; + + st = lock_stat_findnew(acquired_event->addr, acquired_event->name, + acquired_event->file, acquired_event->line); + + switch (st->state) { + case LOCK_STATE_UNLOCKED: + st->state = LOCK_STATE_LOCKED; + st->nr_acquired++; + break; + case LOCK_STATE_LOCKED: + break; + default: + BUG_ON(1); + break; + } + + st->prev_event_time = timestamp; +} + +static void prof_lock_contended_event(struct trace_contended_event *contended_event, + struct event *__event __used, + int cpu __used, + u64 timestamp, + struct thread *thread __used) +{ + struct lock_stat *st; + + st = lock_stat_findnew(contended_event->addr, contended_event->name, + contended_event->file, contended_event->line); + + switch (st->state) { + case LOCK_STATE_UNLOCKED: + break; + case LOCK_STATE_LOCKED: + st->nr_contended++; + break; + default: + BUG_ON(1); + break; + } + + st->prev_event_time = timestamp; +} + +static void prof_lock_release_event(struct trace_release_event *release_event, + struct event *__event __used, + int cpu __used, + u64 timestamp, + struct thread *thread __used) +{ + struct lock_stat *st; + u64 hold_time; + + st = lock_stat_findnew(release_event->addr, release_event->name, + release_event->file, release_event->line); + + switch (st->state) { + case LOCK_STATE_UNLOCKED: + break; + case LOCK_STATE_LOCKED: + st->state = LOCK_STATE_UNLOCKED; + hold_time = timestamp - st->prev_event_time; + + if (timestamp < st->prev_event_time) { + /* terribly, this can happen... */ + goto end; + } + + if (st->wait_time_min > hold_time) + st->wait_time_min = hold_time; + if (st->wait_time_max < hold_time) + st->wait_time_max = hold_time; + st->wait_time_total += hold_time; + + st->nr_release++; + break; + default: + BUG_ON(1); + break; + } + +end: + st->prev_event_time = timestamp; +} + +/* lock oriented handlers */ +/* TODO: handlers for CPU oriented, thread oriented */ +static struct trace_lock_handler prof_lock_ops = { + .acquire_event = prof_lock_acquire_event, + .acquired_event = prof_lock_acquired_event, + .contended_event = prof_lock_contended_event, + .release_event = prof_lock_release_event, +}; + +static struct trace_lock_handler *trace_handler; + +static void +process_lock_acquire_event(void *data, + struct event *event __used, + int cpu __used, + u64 timestamp __used, + struct thread *thread __used) +{ + struct trace_acquire_event acquire_event; + u64 tmp; /* this is required for casting... */ + + tmp = raw_field_value(event, "lockdep_addr", data); + memcpy(&acquire_event.addr, &tmp, sizeof(void *)); + acquire_event.name = (char *)raw_field_ptr(event, "name", data); + acquire_event.file = (char *)raw_field_ptr(event, "file", data); + acquire_event.line = + (unsigned int)raw_field_value(event, "line", data); + + if (trace_handler->acquire_event) { + trace_handler->acquire_event(&acquire_event, + event, cpu, timestamp, thread); + } +} + +static void +process_lock_acquired_event(void *data, + struct event *event __used, + int cpu __used, + u64 timestamp __used, + struct thread *thread __used) +{ + struct trace_acquired_event acquired_event; + u64 tmp; /* this is required for casting... */ + + tmp = raw_field_value(event, "lockdep_addr", data); + memcpy(&acquired_event.addr, &tmp, sizeof(void *)); + acquired_event.name = (char *)raw_field_ptr(event, "name", data); + acquired_event.file = (char *)raw_field_ptr(event, "file", data); + acquired_event.line = + (unsigned int)raw_field_value(event, "line", data); + + if (trace_handler->acquire_event) { + trace_handler->acquired_event(&acquired_event, + event, cpu, timestamp, thread); + } +} + +static void +process_lock_contended_event(void *data, + struct event *event __used, + int cpu __used, + u64 timestamp __used, + struct thread *thread __used) +{ + struct trace_contended_event contended_event; + u64 tmp; /* this is required for casting... */ + + tmp = raw_field_value(event, "lockdep_addr", data); + memcpy(&contended_event.addr, &tmp, sizeof(void *)); + contended_event.name = (char *)raw_field_ptr(event, "name", data); + contended_event.file = (char *)raw_field_ptr(event, "file", data); + contended_event.line = + (unsigned int)raw_field_value(event, "line", data); + + if (trace_handler->acquire_event) { + trace_handler->contended_event(&contended_event, + event, cpu, timestamp, thread); + } +} + +static void +process_lock_release_event(void *data, + struct event *event __used, + int cpu __used, + u64 timestamp __used, + struct thread *thread __used) +{ + struct trace_release_event release_event; + u64 tmp; /* this is required for casting... */ + + tmp = raw_field_value(event, "lockdep_addr", data); + memcpy(&release_event.addr, &tmp, sizeof(void *)); + release_event.name = (char *)raw_field_ptr(event, "name", data); + release_event.file = (char *)raw_field_ptr(event, "file", data); + release_event.line = + (unsigned int)raw_field_value(event, "line", data); + + if (trace_handler->acquire_event) { + trace_handler->release_event(&release_event, + event, cpu, timestamp, thread); + } +} + +static void +process_raw_event(void *data, int cpu, + u64 timestamp, struct thread *thread) +{ + struct event *event; + int type; + + type = trace_parse_common_type(data); + event = trace_find_event(type); + + if (!strcmp(event->name, "lock_acquire")) + process_lock_acquire_event(data, event, cpu, timestamp, thread); + if (!strcmp(event->name, "lock_acquired")) + process_lock_acquired_event(data, event, cpu, timestamp, thread); + if (!strcmp(event->name, "lock_contended")) + process_lock_contended_event(data, event, cpu, timestamp, thread); + if (!strcmp(event->name, "lock_release")) + process_lock_release_event(data, event, cpu, timestamp, thread); +} + +static int process_sample_event(event_t *event, struct perf_session *session) +{ + struct thread *thread; + struct sample_data data; + + bzero(&data, sizeof(struct sample_data)); + event__parse_sample(event, session->sample_type, &data); + thread = perf_session__findnew(session, data.pid); + + /* + * FIXME: this causes warn on 32bit environment + * because of (void *)data.ip (type of data.ip is u64) + */ +/* dump_printf("(IP, %d): %d/%d: %p period: %llu\n", */ +/* event->header.misc, */ +/* data.pid, data.tid, (void *)data.ip, data.period); */ + + if (thread == NULL) { + pr_debug("problem processing %d event, skipping it.\n", + event->header.type); + return -1; + } + + dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); + + if (profile_cpu != -1 && profile_cpu != (int) data.cpu) + return 0; + + process_raw_event(data.raw_data, data.cpu, data.time, thread); + + return 0; +} + +/* TODO: various way to print, coloring, nano or milli sec */ +static void print_result(void) +{ + struct lock_stat *st; + char cut_name[20]; + + printf("%18s ", "ID"); + printf("%20s ", "Name"); + printf("%10s ", "acquired"); + printf("%10s ", "contended"); + + printf("%15s ", "total wait (ns)"); + printf("%15s ", "max wait (ns)"); + printf("%15s ", "min wait (ns)"); + + printf("\n\n"); + + while ((st = pop_from_result())) { + bzero(cut_name, 20); + + printf("%p ", st->addr); + + if (strlen(st->name) < 16) { + /* output raw name */ + printf("%20s ", st->name); + } else { + strncpy(cut_name, st->name, 16); + cut_name[16] = '.'; + cut_name[17] = '.'; + cut_name[18] = '.'; + cut_name[19] = '\0'; + /* cut off name for saving output style */ + printf("%20s ", cut_name); + } + + printf("%10u ", st->nr_acquired); + printf("%10u ", st->nr_contended); + + printf("%15llu ", st->wait_time_total); + printf("%15llu ", st->wait_time_max); + printf("%15llu ", st->wait_time_min == ULLONG_MAX ? + 0 : st->wait_time_min); + printf("\n"); + } +} + +static void dump_map(void) +{ + unsigned int i; + struct lock_stat *st; + + for (i = 0; i < LOCKHASH_SIZE; i++) { + list_for_each_entry(st, &lockhash_table[i], hash_entry) { + printf("%p: %s (src: %s, line: %u)\n", + st->addr, st->name, st->file, st->line); + } + } +} + +static struct perf_event_ops eops = { + .sample = process_sample_event, + .comm = event__process_comm, +}; + +static struct perf_session *session; + +static int read_events(void) +{ + session = perf_session__new(input_name, O_RDONLY, 0); + if (!session) + die("Initializing perf session failed\n"); + + return perf_session__process_events(session, &eops); +} + +static void sort_result(void) +{ + unsigned int i; + struct lock_stat *st; + + for (i = 0; i < LOCKHASH_SIZE; i++) { + list_for_each_entry(st, &lockhash_table[i], hash_entry) { + insert_to_result(st, compare); + } + } +} + +static void __cmd_prof(void) +{ + setup_pager(); + select_key(); + read_events(); + sort_result(); + print_result(); +} + +static const char * const prof_usage[] = { + "perf sched prof []", + NULL +}; + +static const struct option prof_options[] = { + OPT_STRING('k', "key", &sort_key, "acquired", + "key for sorting"), + /* TODO: type */ + OPT_END() +}; + +static const char * const lock_usage[] = { + "perf lock [] {record|trace|prof}", + NULL +}; + +static const struct option lock_options[] = { + OPT_STRING('i', "input", &input_name, "file", + "input file name"), + OPT_BOOLEAN('v', "verbose", &verbose, + "be more verbose (show symbol address, etc)"), + OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, + "dump raw trace in ASCII"), + OPT_END() +}; + +static const char *record_args[] = { + "record", + "-a", + "-R", + "-M", + "-f", + "-m", "1024", + "-c", "1", + "-e", "lock:lock_acquire:r", + "-e", "lock:lock_acquired:r", + "-e", "lock:lock_contended:r", + "-e", "lock:lock_release:r", +}; + +static int __cmd_record(int argc, const char **argv) +{ + unsigned int rec_argc, i, j; + const char **rec_argv; + + rec_argc = ARRAY_SIZE(record_args) + argc - 1; + rec_argv = calloc(rec_argc + 1, sizeof(char *)); + + for (i = 0; i < ARRAY_SIZE(record_args); i++) + rec_argv[i] = strdup(record_args[i]); + + for (j = 1; j < (unsigned int)argc; j++, i++) + rec_argv[i] = argv[j]; + + BUG_ON(i != rec_argc); + + return cmd_record(i, rec_argv, NULL); +} + +int cmd_lock(int argc, const char **argv, const char *prefix __used) +{ + unsigned int i; + + symbol__init(); + for (i = 0; i < LOCKHASH_SIZE; i++) + INIT_LIST_HEAD(lockhash_table + i); + + argc = parse_options(argc, argv, lock_options, lock_usage, + PARSE_OPT_STOP_AT_NON_OPTION); + if (!argc) + usage_with_options(lock_usage, lock_options); + + if (!strncmp(argv[0], "rec", 3)) { + return __cmd_record(argc, argv); + } else if (!strncmp(argv[0], "prof", 4)) { + trace_handler = &prof_lock_ops; + if (argc) { + argc = parse_options(argc, argv, + prof_options, prof_usage, 0); + if (argc) + usage_with_options(prof_usage, prof_options); + } + __cmd_prof(); + } else if (!strcmp(argv[0], "trace")) { + /* Aliased to 'perf trace' */ + return cmd_trace(argc, argv, prefix); + } else if (!strcmp(argv[0], "map")) { + /* recycling prof_lock_ops */ + trace_handler = &prof_lock_ops; + setup_pager(); + read_events(); + dump_map(); + } else { + usage_with_options(lock_usage, lock_options); + } + + return 0; +} diff --git a/tools/perf/builtin.h b/tools/perf/builtin.h index dee97cf..10fe49e 100644 --- a/tools/perf/builtin.h +++ b/tools/perf/builtin.h @@ -31,5 +31,6 @@ extern int cmd_trace(int argc, const char **argv, const char *prefix); extern int cmd_version(int argc, const char **argv, const char *prefix); extern int cmd_probe(int argc, const char **argv, const char *prefix); extern int cmd_kmem(int argc, const char **argv, const char *prefix); +extern int cmd_lock(int argc, const char **argv, const char *prefix); #endif diff --git a/tools/perf/perf.c b/tools/perf/perf.c index 109b89b..57cb107 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -303,6 +303,7 @@ static void handle_internal_command(int argc, const char **argv) { "sched", cmd_sched, 0 }, { "probe", cmd_probe, 0 }, { "kmem", cmd_kmem, 0 }, + { "lock", cmd_lock, 0 }, }; unsigned int i; static const char ext[] = STRIP_EXTENSION; -- cgit v0.10.2 From 59f411b62c9282891274e721fea29026b0eda3cc Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sun, 31 Jan 2010 08:27:58 +0100 Subject: perf lock: Clean up various details Fix up a few small stylistic details: - use consistent vertical spacing/alignment - remove line80 artifacts - group some global variables better - remove dead code Plus rename 'prof' to 'report' to make it more in line with other tools, and remove the line/file keying as we really want to use IPs like the other tools do. Signed-off-by: Ingo Molnar Cc: Hitoshi Mitake Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker LKML-Reference: <1264851813-8413-12-git-send-email-mitake@dcl.info.waseda.ac.jp> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c index 2b5f887..fb9ab2a 100644 --- a/tools/perf/builtin-lock.c +++ b/tools/perf/builtin-lock.c @@ -32,37 +32,37 @@ static struct list_head lockhash_table[LOCKHASH_SIZE]; #define __lockhashfn(key) hash_long((unsigned long)key, LOCKHASH_BITS) #define lockhashentry(key) (lockhash_table + __lockhashfn((key))) -#define LOCK_STATE_UNLOCKED 0 /* initial state */ -#define LOCK_STATE_LOCKED 1 +#define LOCK_STATE_UNLOCKED 0 /* initial state */ +#define LOCK_STATE_LOCKED 1 struct lock_stat { - struct list_head hash_entry; - struct rb_node rb; /* used for sorting */ + struct list_head hash_entry; + struct rb_node rb; /* used for sorting */ - /* FIXME: raw_field_value() returns unsigned long long, + /* + * FIXME: raw_field_value() returns unsigned long long, * so address of lockdep_map should be dealed as 64bit. - * Is there more better solution? */ - void *addr; /* address of lockdep_map, used as ID */ - char *name; /* for strcpy(), we cannot use const */ - char *file; - unsigned int line; + * Is there more better solution? + */ + void *addr; /* address of lockdep_map, used as ID */ + char *name; /* for strcpy(), we cannot use const */ - int state; - u64 prev_event_time; /* timestamp of previous event */ + int state; + u64 prev_event_time; /* timestamp of previous event */ - unsigned int nr_acquired; - unsigned int nr_acquire; - unsigned int nr_contended; - unsigned int nr_release; + unsigned int nr_acquired; + unsigned int nr_acquire; + unsigned int nr_contended; + unsigned int nr_release; /* these times are in nano sec. */ - u64 wait_time_total; - u64 wait_time_min; - u64 wait_time_max; + u64 wait_time_total; + u64 wait_time_min; + u64 wait_time_max; }; /* build simple key function one is bigger than two */ -#define SINGLE_KEY(member) \ +#define SINGLE_KEY(member) \ static int lock_stat_key_ ## member(struct lock_stat *one, \ struct lock_stat *two) \ { \ @@ -81,12 +81,15 @@ struct lock_key { * this should be simpler than raw name of member * e.g. nr_acquired -> acquired, wait_time_total -> wait_total */ - const char *name; - int (*key)(struct lock_stat*, struct lock_stat*); + const char *name; + int (*key)(struct lock_stat*, struct lock_stat*); }; -static const char *sort_key = "acquired"; -static int (*compare)(struct lock_stat *, struct lock_stat *); +static const char *sort_key = "acquired"; + +static int (*compare)(struct lock_stat *, struct lock_stat *); + +static struct rb_root result; /* place to store sorted data */ #define DEF_KEY_LOCK(name, fn_suffix) \ { #name, lock_stat_key_ ## fn_suffix } @@ -116,11 +119,8 @@ static void select_key(void) die("Unknown compare key:%s\n", sort_key); } -static struct rb_root result; /* place to store sorted data */ - static void insert_to_result(struct lock_stat *st, - int (*bigger)(struct lock_stat *, - struct lock_stat *)) + int (*bigger)(struct lock_stat *, struct lock_stat *)) { struct rb_node **rb = &result.rb_node; struct rb_node *parent = NULL; @@ -155,8 +155,7 @@ static struct lock_stat *pop_from_result(void) return container_of(node, struct lock_stat, rb); } -static struct lock_stat *lock_stat_findnew(void *addr, const char *name, - const char *file, unsigned int line) +static struct lock_stat *lock_stat_findnew(void *addr, const char *name) { struct list_head *entry = lockhashentry(addr); struct lock_stat *ret, *new; @@ -175,11 +174,6 @@ static struct lock_stat *lock_stat_findnew(void *addr, const char *name, if (!new->name) goto alloc_failed; strcpy(new->name, name); - new->file = zalloc(sizeof(char) * strlen(file) + 1); - if (!new->file) - goto alloc_failed; - strcpy(new->file, file); - new->line = line; /* LOCK_STATE_UNLOCKED == 0 isn't guaranteed forever */ new->state = LOCK_STATE_UNLOCKED; @@ -197,36 +191,28 @@ static char const *input_name = "perf.data"; static int profile_cpu = -1; struct raw_event_sample { - u32 size; - char data[0]; + u32 size; + char data[0]; }; struct trace_acquire_event { - void *addr; - const char *name; - const char *file; - unsigned int line; + void *addr; + const char *name; }; struct trace_acquired_event { - void *addr; - const char *name; - const char *file; - unsigned int line; + void *addr; + const char *name; }; struct trace_contended_event { - void *addr; - const char *name; - const char *file; - unsigned int line; + void *addr; + const char *name; }; struct trace_release_event { - void *addr; - const char *name; - const char *file; - unsigned int line; + void *addr; + const char *name; }; struct trace_lock_handler { @@ -255,7 +241,8 @@ struct trace_lock_handler { struct thread *thread); }; -static void prof_lock_acquire_event(struct trace_acquire_event *acquire_event, +static void +report_lock_acquire_event(struct trace_acquire_event *acquire_event, struct event *__event __used, int cpu __used, u64 timestamp, @@ -263,8 +250,7 @@ static void prof_lock_acquire_event(struct trace_acquire_event *acquire_event, { struct lock_stat *st; - st = lock_stat_findnew(acquire_event->addr, acquire_event->name, - acquire_event->file, acquire_event->line); + st = lock_stat_findnew(acquire_event->addr, acquire_event->name); switch (st->state) { case LOCK_STATE_UNLOCKED: @@ -279,7 +265,8 @@ static void prof_lock_acquire_event(struct trace_acquire_event *acquire_event, st->prev_event_time = timestamp; } -static void prof_lock_acquired_event(struct trace_acquired_event *acquired_event, +static void +report_lock_acquired_event(struct trace_acquired_event *acquired_event, struct event *__event __used, int cpu __used, u64 timestamp, @@ -287,8 +274,7 @@ static void prof_lock_acquired_event(struct trace_acquired_event *acquired_event { struct lock_stat *st; - st = lock_stat_findnew(acquired_event->addr, acquired_event->name, - acquired_event->file, acquired_event->line); + st = lock_stat_findnew(acquired_event->addr, acquired_event->name); switch (st->state) { case LOCK_STATE_UNLOCKED: @@ -305,7 +291,8 @@ static void prof_lock_acquired_event(struct trace_acquired_event *acquired_event st->prev_event_time = timestamp; } -static void prof_lock_contended_event(struct trace_contended_event *contended_event, +static void +report_lock_contended_event(struct trace_contended_event *contended_event, struct event *__event __used, int cpu __used, u64 timestamp, @@ -313,8 +300,7 @@ static void prof_lock_contended_event(struct trace_contended_event *contended_ev { struct lock_stat *st; - st = lock_stat_findnew(contended_event->addr, contended_event->name, - contended_event->file, contended_event->line); + st = lock_stat_findnew(contended_event->addr, contended_event->name); switch (st->state) { case LOCK_STATE_UNLOCKED: @@ -330,7 +316,8 @@ static void prof_lock_contended_event(struct trace_contended_event *contended_ev st->prev_event_time = timestamp; } -static void prof_lock_release_event(struct trace_release_event *release_event, +static void +report_lock_release_event(struct trace_release_event *release_event, struct event *__event __used, int cpu __used, u64 timestamp, @@ -339,8 +326,7 @@ static void prof_lock_release_event(struct trace_release_event *release_event, struct lock_stat *st; u64 hold_time; - st = lock_stat_findnew(release_event->addr, release_event->name, - release_event->file, release_event->line); + st = lock_stat_findnew(release_event->addr, release_event->name); switch (st->state) { case LOCK_STATE_UNLOCKED: @@ -373,11 +359,11 @@ end: /* lock oriented handlers */ /* TODO: handlers for CPU oriented, thread oriented */ -static struct trace_lock_handler prof_lock_ops = { - .acquire_event = prof_lock_acquire_event, - .acquired_event = prof_lock_acquired_event, - .contended_event = prof_lock_contended_event, - .release_event = prof_lock_release_event, +static struct trace_lock_handler report_lock_ops = { + .acquire_event = report_lock_acquire_event, + .acquired_event = report_lock_acquired_event, + .contended_event = report_lock_contended_event, + .release_event = report_lock_release_event, }; static struct trace_lock_handler *trace_handler; @@ -395,14 +381,9 @@ process_lock_acquire_event(void *data, tmp = raw_field_value(event, "lockdep_addr", data); memcpy(&acquire_event.addr, &tmp, sizeof(void *)); acquire_event.name = (char *)raw_field_ptr(event, "name", data); - acquire_event.file = (char *)raw_field_ptr(event, "file", data); - acquire_event.line = - (unsigned int)raw_field_value(event, "line", data); - if (trace_handler->acquire_event) { - trace_handler->acquire_event(&acquire_event, - event, cpu, timestamp, thread); - } + if (trace_handler->acquire_event) + trace_handler->acquire_event(&acquire_event, event, cpu, timestamp, thread); } static void @@ -418,14 +399,9 @@ process_lock_acquired_event(void *data, tmp = raw_field_value(event, "lockdep_addr", data); memcpy(&acquired_event.addr, &tmp, sizeof(void *)); acquired_event.name = (char *)raw_field_ptr(event, "name", data); - acquired_event.file = (char *)raw_field_ptr(event, "file", data); - acquired_event.line = - (unsigned int)raw_field_value(event, "line", data); - if (trace_handler->acquire_event) { - trace_handler->acquired_event(&acquired_event, - event, cpu, timestamp, thread); - } + if (trace_handler->acquire_event) + trace_handler->acquired_event(&acquired_event, event, cpu, timestamp, thread); } static void @@ -441,14 +417,9 @@ process_lock_contended_event(void *data, tmp = raw_field_value(event, "lockdep_addr", data); memcpy(&contended_event.addr, &tmp, sizeof(void *)); contended_event.name = (char *)raw_field_ptr(event, "name", data); - contended_event.file = (char *)raw_field_ptr(event, "file", data); - contended_event.line = - (unsigned int)raw_field_value(event, "line", data); - if (trace_handler->acquire_event) { - trace_handler->contended_event(&contended_event, - event, cpu, timestamp, thread); - } + if (trace_handler->acquire_event) + trace_handler->contended_event(&contended_event, event, cpu, timestamp, thread); } static void @@ -464,14 +435,9 @@ process_lock_release_event(void *data, tmp = raw_field_value(event, "lockdep_addr", data); memcpy(&release_event.addr, &tmp, sizeof(void *)); release_event.name = (char *)raw_field_ptr(event, "name", data); - release_event.file = (char *)raw_field_ptr(event, "file", data); - release_event.line = - (unsigned int)raw_field_value(event, "line", data); - if (trace_handler->acquire_event) { - trace_handler->release_event(&release_event, - event, cpu, timestamp, thread); - } + if (trace_handler->acquire_event) + trace_handler->release_event(&release_event, event, cpu, timestamp, thread); } static void @@ -503,14 +469,6 @@ static int process_sample_event(event_t *event, struct perf_session *session) event__parse_sample(event, session->sample_type, &data); thread = perf_session__findnew(session, data.pid); - /* - * FIXME: this causes warn on 32bit environment - * because of (void *)data.ip (type of data.ip is u64) - */ -/* dump_printf("(IP, %d): %d/%d: %p period: %llu\n", */ -/* event->header.misc, */ -/* data.pid, data.tid, (void *)data.ip, data.period); */ - if (thread == NULL) { pr_debug("problem processing %d event, skipping it.\n", event->header.type); @@ -580,15 +538,14 @@ static void dump_map(void) for (i = 0; i < LOCKHASH_SIZE; i++) { list_for_each_entry(st, &lockhash_table[i], hash_entry) { - printf("%p: %s (src: %s, line: %u)\n", - st->addr, st->name, st->file, st->line); + printf("%p: %s\n", st->addr, st->name); } } } static struct perf_event_ops eops = { - .sample = process_sample_event, - .comm = event__process_comm, + .sample = process_sample_event, + .comm = event__process_comm, }; static struct perf_session *session; @@ -614,7 +571,7 @@ static void sort_result(void) } } -static void __cmd_prof(void) +static void __cmd_report(void) { setup_pager(); select_key(); @@ -623,12 +580,12 @@ static void __cmd_prof(void) print_result(); } -static const char * const prof_usage[] = { - "perf sched prof []", +static const char * const report_usage[] = { + "perf lock report []", NULL }; -static const struct option prof_options[] = { +static const struct option report_options[] = { OPT_STRING('k', "key", &sort_key, "acquired", "key for sorting"), /* TODO: type */ @@ -636,17 +593,14 @@ static const struct option prof_options[] = { }; static const char * const lock_usage[] = { - "perf lock [] {record|trace|prof}", + "perf lock [] {record|trace|report}", NULL }; static const struct option lock_options[] = { - OPT_STRING('i', "input", &input_name, "file", - "input file name"), - OPT_BOOLEAN('v', "verbose", &verbose, - "be more verbose (show symbol address, etc)"), - OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, - "dump raw trace in ASCII"), + OPT_STRING('i', "input", &input_name, "file", "input file name"), + OPT_BOOLEAN('v', "verbose", &verbose, "be more verbose (show symbol address, etc)"), + OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, "dump raw trace in ASCII"), OPT_END() }; @@ -698,21 +652,21 @@ int cmd_lock(int argc, const char **argv, const char *prefix __used) if (!strncmp(argv[0], "rec", 3)) { return __cmd_record(argc, argv); - } else if (!strncmp(argv[0], "prof", 4)) { - trace_handler = &prof_lock_ops; + } else if (!strncmp(argv[0], "report", 6)) { + trace_handler = &report_lock_ops; if (argc) { argc = parse_options(argc, argv, - prof_options, prof_usage, 0); + report_options, report_usage, 0); if (argc) - usage_with_options(prof_usage, prof_options); + usage_with_options(report_usage, report_options); } - __cmd_prof(); + __cmd_report(); } else if (!strcmp(argv[0], "trace")) { /* Aliased to 'perf trace' */ return cmd_trace(argc, argv, prefix); } else if (!strcmp(argv[0], "map")) { - /* recycling prof_lock_ops */ - trace_handler = &prof_lock_ops; + /* recycling report_lock_ops */ + trace_handler = &report_lock_ops; setup_pager(); read_events(); dump_map(); -- cgit v0.10.2 From aee4467b5ce5047401efb4175b1360ec1734affc Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 1 Feb 2010 11:33:22 +0900 Subject: sh: Fix up large system memory handling for SH7780 PCI. For systems that have more than 512MB we need to set up an additional mapping, this fixes up the rounding to the next power of two and splits out the mapping accordingly between the two local bus mapping windows. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/pci-sh7780.c b/arch/sh/drivers/pci/pci-sh7780.c index 40531cd..8405c8f 100644 --- a/arch/sh/drivers/pci/pci-sh7780.c +++ b/arch/sh/drivers/pci/pci-sh7780.c @@ -13,6 +13,7 @@ #include #include #include +#include #include "pci-sh4.h" #include #include @@ -59,7 +60,11 @@ static int __init sh7780_pci_init(void) __raw_writel(SH4_PCICR_PREFIX | SH4_PCICR_PRST, chan->reg_base + SH4_PCICR); - /* Wait for it to come back up.. */ + /* + * Wait for it to come back up. The spec says to allow for up to + * 1 second after toggling the reset pin, but in practice 100ms + * is more than enough. + */ mdelay(100); id = __raw_readw(chan->reg_base + PCI_VENDOR_ID); @@ -90,17 +95,34 @@ static int __init sh7780_pci_init(void) */ __raw_writel(SH4_PCICR_PREFIX, chan->reg_base + SH4_PCICR); + __raw_writel(0, chan->reg_base + PCI_BASE_ADDRESS_0); + memphys = __pa(memory_start); - memsize = memory_end - memory_start; + memsize = roundup_pow_of_two(memory_end - memory_start); /* - * Set IO and Mem windows to local address - * Make PCI and local address the same for easy 1 to 1 mapping + * If there's more than 512MB of memory, we need to roll over to + * LAR1/LSR1. */ - __raw_writel(0, chan->reg_base + PCI_BASE_ADDRESS_0); + if (memsize > SZ_512M) { + __raw_writel(memphys + SZ_512M, chan->reg_base + SH4_PCILAR1); + __raw_writel((((memsize - SZ_512M) - SZ_1M) & 0x1ff00000) | 1, + chan->reg_base + SH4_PCILSR1); + memsize = SZ_512M; + } else { + /* + * Otherwise just zero it out and disable it. + */ + __raw_writel(0, chan->reg_base + SH4_PCILAR1); + __raw_writel(0, chan->reg_base + SH4_PCILSR1); + } + /* + * LAR0/LSR0 covers up to the first 512MB, which is enough to + * cover all of lowmem on most platforms. + */ __raw_writel(memphys, chan->reg_base + SH4_PCILAR0); - __raw_writel((memsize - 1) << 9 | 1, + __raw_writel(((memsize - SZ_1M) & 0x1ff00000) | 1, chan->reg_base + SH4_PCILSR0); /* Clear out PCI arbiter IRQs */ -- cgit v0.10.2 From d2f6650a950dadd20667a04a9dc785f240d43695 Mon Sep 17 00:00:00 2001 From: Thomas Renninger Date: Fri, 29 Jan 2010 17:48:51 +0100 Subject: ACPI: Add NULL pointer check in acpi_bus_start If acpi_bus_add does not return a device and it's passed to acpi_bus_start, bad things will happen: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 IP: [] acpi_bus_start+0x14/0x24 ... [] acpiphp_bus_add+0xba/0x130 [acpiphp] [] enable_device+0x132/0x2ff [acpiphp] [] acpiphp_enable_slot+0xb8/0x130 [acpiphp] [] handle_hotplug_event_func+0x87/0x190 [acpiphp] Next patch would make this NULL pointer check obsolete, but better having one more than one missing... Signed-off-by: Thomas Renninger Acked-by: Bjorn Helgaas CC: stable@kernel.org Signed-off-by: Len Brown diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c index ff9f622..8044583 100644 --- a/drivers/acpi/scan.c +++ b/drivers/acpi/scan.c @@ -1357,6 +1357,9 @@ int acpi_bus_start(struct acpi_device *device) { struct acpi_bus_ops ops; + if (!device) + return -EINVAL; + memset(&ops, 0, sizeof(ops)); ops.acpi_op_start = 1; -- cgit v0.10.2 From 7779688fc3d1ceddad84846a7b0affbe8e78ec6e Mon Sep 17 00:00:00 2001 From: Thomas Renninger Date: Fri, 29 Jan 2010 17:48:52 +0100 Subject: ACPI: acpi_bus_{scan,bus,add}: return -ENODEV if no device was found Callers (acpi_memhotplug.c, dock.c and others) check for the return value of acpi_bus_add() and assume a valid device was returned in case zero was returned. Thus return -ENODEV if no device was found in acpi_bus_scan and propagate this through acpi_bus_add and acpi_bus_start. Also remove a confusing comment in acpiphp_glue.c, acpi_bus_scan will and cannot invoke if acpi_bus_add returns no valid device. Signed-off-by: Thomas Renninger Acked-by: Bjorn Helgaas Signed-off-by: Len Brown diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c index 8044583..3e00967 100644 --- a/drivers/acpi/scan.c +++ b/drivers/acpi/scan.c @@ -1336,9 +1336,25 @@ static int acpi_bus_scan(acpi_handle handle, struct acpi_bus_ops *ops, if (child) *child = device; - return 0; + + if (device) + return 0; + else + return -ENODEV; } +/* + * acpi_bus_add and acpi_bus_start + * + * scan a given ACPI tree and (probably recently hot-plugged) + * create and add or starts found devices. + * + * If no devices were found -ENODEV is returned which does not + * mean that this is a real error, there just have been no suitable + * ACPI objects in the table trunk from which the kernel could create + * a device and add/start an appropriate driver. + */ + int acpi_bus_add(struct acpi_device **child, struct acpi_device *parent, acpi_handle handle, int type) @@ -1348,8 +1364,7 @@ acpi_bus_add(struct acpi_device **child, memset(&ops, 0, sizeof(ops)); ops.acpi_op_add = 1; - acpi_bus_scan(handle, &ops, child); - return 0; + return acpi_bus_scan(handle, &ops, child); } EXPORT_SYMBOL(acpi_bus_add); @@ -1363,8 +1378,7 @@ int acpi_bus_start(struct acpi_device *device) memset(&ops, 0, sizeof(ops)); ops.acpi_op_start = 1; - acpi_bus_scan(device->handle, &ops, NULL); - return 0; + return acpi_bus_scan(device->handle, &ops, NULL); } EXPORT_SYMBOL(acpi_bus_start); diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c index 8e952fd..cb2fd01 100644 --- a/drivers/pci/hotplug/acpiphp_glue.c +++ b/drivers/pci/hotplug/acpiphp_glue.c @@ -720,12 +720,6 @@ static int acpiphp_bus_add(struct acpiphp_func *func) -ret_val); goto acpiphp_bus_add_out; } - /* - * try to start anyway. We could have failed to add - * simply because this bus had previously been added - * on another add. Don't bother with the return value - * we just keep going. - */ ret_val = acpi_bus_start(device); acpiphp_bus_add_out: -- cgit v0.10.2 From 85b59f5bb24aeca1a987cbb206e228bf630c8327 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 1 Feb 2010 13:01:42 +0900 Subject: sh: Enable PCI66 support for SH7780 host controller. This adds some helper glue for scanning the bus and determining if all of the devices are 66MHz capable or not before flipping on 66MHz mode. This isn't quite to spec, but it's fairly consistent with what other embedded controllers end up having to do. Scanning code cribbed from the MIPS txx9 PCI code. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/Makefile b/arch/sh/drivers/pci/Makefile index 2c458b6..4a59e68 100644 --- a/arch/sh/drivers/pci/Makefile +++ b/arch/sh/drivers/pci/Makefile @@ -1,7 +1,7 @@ # # Makefile for the PCI specific kernel interface routines under Linux. # -obj-y += pci.o +obj-y += common.o pci.o obj-$(CONFIG_CPU_SUBTYPE_SH7751) += pci-sh7751.o ops-sh4.o obj-$(CONFIG_CPU_SUBTYPE_SH7751R) += pci-sh7751.o ops-sh4.o diff --git a/arch/sh/drivers/pci/common.c b/arch/sh/drivers/pci/common.c new file mode 100644 index 0000000..f67c946 --- /dev/null +++ b/arch/sh/drivers/pci/common.c @@ -0,0 +1,64 @@ +#include +#include + +static int __init +early_read_config_word(struct pci_channel *hose, + int top_bus, int bus, int devfn, int offset, u16 *value) +{ + struct pci_dev fake_dev; + struct pci_bus fake_bus; + + fake_dev.bus = &fake_bus; + fake_dev.sysdata = hose; + fake_dev.devfn = devfn; + fake_bus.number = bus; + fake_bus.sysdata = hose; + fake_bus.ops = hose->pci_ops; + + if (bus != top_bus) + /* Fake a parent bus structure. */ + fake_bus.parent = &fake_bus; + else + fake_bus.parent = NULL; + + return pci_read_config_word(&fake_dev, offset, value); +} + +int __init pci_is_66mhz_capable(struct pci_channel *hose, + int top_bus, int current_bus) +{ + u32 pci_devfn; + unsigned short vid; + int cap66 = -1; + u16 stat; + + printk(KERN_INFO "PCI: Checking 66MHz capabilities...\n"); + + for (pci_devfn = 0; pci_devfn < 0xff; pci_devfn++) { + if (PCI_FUNC(pci_devfn)) + continue; + if (early_read_config_word(hose, top_bus, current_bus, + pci_devfn, PCI_VENDOR_ID, &vid) != + PCIBIOS_SUCCESSFUL) + continue; + if (vid == 0xffff) + continue; + + /* check 66MHz capability */ + if (cap66 < 0) + cap66 = 1; + if (cap66) { + early_read_config_word(hose, top_bus, current_bus, + pci_devfn, PCI_STATUS, &stat); + if (!(stat & PCI_STATUS_66MHZ)) { + printk(KERN_DEBUG + "PCI: %02x:%02x not 66MHz capable.\n", + current_bus, pci_devfn); + cap66 = 0; + break; + } + } + } + + return cap66 > 0; +} diff --git a/arch/sh/drivers/pci/pci-sh7780.c b/arch/sh/drivers/pci/pci-sh7780.c index 8405c8f..b68f45b 100644 --- a/arch/sh/drivers/pci/pci-sh7780.c +++ b/arch/sh/drivers/pci/pci-sh7780.c @@ -41,6 +41,29 @@ static struct pci_channel sh7780_pci_controller = { .io_map_base = SH7780_PCI_IO_BASE, }; +static void __init sh7780_pci66_init(struct pci_channel *hose) +{ + unsigned int tmp; + + if (!pci_is_66mhz_capable(hose, 0, 0)) + return; + + /* Enable register access */ + tmp = __raw_readl(hose->reg_base + SH4_PCICR); + tmp |= SH4_PCICR_PREFIX; + __raw_writel(tmp, hose->reg_base + SH4_PCICR); + + /* Enable 66MHz operation */ + tmp = __raw_readw(hose->reg_base + PCI_STATUS); + tmp |= PCI_STATUS_66MHZ; + __raw_writew(tmp, hose->reg_base + PCI_STATUS); + + /* Done */ + tmp = __raw_readl(hose->reg_base + SH4_PCICR); + tmp |= SH4_PCICR_PREFIX | SH4_PCICR_CFIN; + __raw_writel(tmp, hose->reg_base + SH4_PCICR); +} + static int __init sh7780_pci_init(void) { struct pci_channel *chan = &sh7780_pci_controller; @@ -176,6 +199,12 @@ static int __init sh7780_pci_init(void) register_pci_controller(chan); + sh7780_pci66_init(chan); + + printk(KERN_NOTICE "PCI: Running at %dMHz.\n", + (__raw_readw(chan->reg_base + PCI_STATUS) & PCI_STATUS_66MHZ) ? + 66 : 33); + return 0; } arch_initcall(sh7780_pci_init); diff --git a/arch/sh/drivers/pci/pci.c b/arch/sh/drivers/pci/pci.c index 45a15ca..63b11fd 100644 --- a/arch/sh/drivers/pci/pci.c +++ b/arch/sh/drivers/pci/pci.c @@ -88,6 +88,8 @@ void __devinit register_pci_controller(struct pci_channel *hose) mutex_unlock(&pci_scan_mutex); } + return; + out: printk(KERN_WARNING "Skipping PCI bus scan due to resource conflict\n"); } diff --git a/arch/sh/include/asm/pci.h b/arch/sh/include/asm/pci.h index d124a00..5849d43 100644 --- a/arch/sh/include/asm/pci.h +++ b/arch/sh/include/asm/pci.h @@ -32,6 +32,8 @@ struct pci_channel { }; extern void register_pci_controller(struct pci_channel *hose); +extern int pci_is_66mhz_capable(struct pci_channel *hose, + int top_bus, int current_bus); extern unsigned long PCIBIOS_MIN_IO, PCIBIOS_MIN_MEM; -- cgit v0.10.2 From bcf39352eb9e9026f7a1028d4bce3707b65f104b Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 1 Feb 2010 13:11:25 +0900 Subject: sh: Handle PCI controller resource conflicts. register_pci_controller() can fail, but presently is a void function. Change this over to an int so that we can bail early before continuing on with post-registration initialization (such as throwing the controller in to 66MHz mode in the case of the SH7780 host controller). Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/pci-dreamcast.c b/arch/sh/drivers/pci/pci-dreamcast.c index 210f9d4..bd5a1e5 100644 --- a/arch/sh/drivers/pci/pci-dreamcast.c +++ b/arch/sh/drivers/pci/pci-dreamcast.c @@ -95,8 +95,6 @@ static int __init gapspci_init(void) outl(0x00002001, GAPSPCI_BBA_CONFIG+0x10); outl(0x01000000, GAPSPCI_BBA_CONFIG+0x14); - register_pci_controller(&dreamcast_pci_controller); - - return 0; + return register_pci_controller(&dreamcast_pci_controller); } arch_initcall(gapspci_init); diff --git a/arch/sh/drivers/pci/pci-sh5.c b/arch/sh/drivers/pci/pci-sh5.c index 873ed2b..bce73fa 100644 --- a/arch/sh/drivers/pci/pci-sh5.c +++ b/arch/sh/drivers/pci/pci-sh5.c @@ -216,8 +216,6 @@ static int __init sh5pci_init(void) sh5_mem_resource.start = memStart; sh5_mem_resource.end = memStart + memSize; - register_pci_controller(&sh5pci_controller); - - return 0; + return register_pci_controller(&sh5pci_controller); } arch_initcall(sh5pci_init); diff --git a/arch/sh/drivers/pci/pci-sh7751.c b/arch/sh/drivers/pci/pci-sh7751.c index 02306dd..6ad5beb 100644 --- a/arch/sh/drivers/pci/pci-sh7751.c +++ b/arch/sh/drivers/pci/pci-sh7751.c @@ -176,8 +176,6 @@ static int __init sh7751_pci_init(void) word = SH4_PCICR_PREFIX | SH4_PCICR_CFIN | SH4_PCICR_ARBM; pci_write_reg(chan, word, SH4_PCICR); - register_pci_controller(chan); - - return 0; + return register_pci_controller(chan); } arch_initcall(sh7751_pci_init); diff --git a/arch/sh/drivers/pci/pci-sh7780.c b/arch/sh/drivers/pci/pci-sh7780.c index b68f45b..0e0ddd6 100644 --- a/arch/sh/drivers/pci/pci-sh7780.c +++ b/arch/sh/drivers/pci/pci-sh7780.c @@ -71,6 +71,7 @@ static int __init sh7780_pci_init(void) size_t memsize; unsigned int id; const char *type; + int ret; printk(KERN_NOTICE "PCI: Starting intialization.\n"); @@ -197,7 +198,9 @@ static int __init sh7780_pci_init(void) __raw_writel(SH4_PCICR_PREFIX | SH4_PCICR_CFIN | SH4_PCICR_FTO, chan->reg_base + SH4_PCICR); - register_pci_controller(chan); + ret = register_pci_controller(chan); + if (unlikely(ret)) + return ret; sh7780_pci66_init(chan); diff --git a/arch/sh/drivers/pci/pci.c b/arch/sh/drivers/pci/pci.c index 63b11fd..488331c 100644 --- a/arch/sh/drivers/pci/pci.c +++ b/arch/sh/drivers/pci/pci.c @@ -58,7 +58,7 @@ static void __devinit pcibios_scanbus(struct pci_channel *hose) static DEFINE_MUTEX(pci_scan_mutex); -void __devinit register_pci_controller(struct pci_channel *hose) +int __devinit register_pci_controller(struct pci_channel *hose) { if (request_resource(&iomem_resource, hose->mem_resource) < 0) goto out; @@ -88,10 +88,11 @@ void __devinit register_pci_controller(struct pci_channel *hose) mutex_unlock(&pci_scan_mutex); } - return; + return 0; out: printk(KERN_WARNING "Skipping PCI bus scan due to resource conflict\n"); + return -1; } static int __init pcibios_init(void) diff --git a/arch/sh/drivers/pci/pcie-sh7786.c b/arch/sh/drivers/pci/pcie-sh7786.c index ac37ee8..feac1fe 100644 --- a/arch/sh/drivers/pci/pcie-sh7786.c +++ b/arch/sh/drivers/pci/pcie-sh7786.c @@ -296,9 +296,7 @@ static int __devinit sh7786_pcie_init_hw(struct sh7786_pcie_port *port) if (unlikely(ret < 0)) return ret; - register_pci_controller(port->hose); - - return 0; + return register_pci_controller(port->hose); } static struct sh7786_pcie_hwops sh7786_65nm_pcie_hwops __initdata = { diff --git a/arch/sh/include/asm/pci.h b/arch/sh/include/asm/pci.h index 5849d43..bbd10cf 100644 --- a/arch/sh/include/asm/pci.h +++ b/arch/sh/include/asm/pci.h @@ -31,7 +31,7 @@ struct pci_channel { unsigned int need_domain_info; }; -extern void register_pci_controller(struct pci_channel *hose); +extern int register_pci_controller(struct pci_channel *hose); extern int pci_is_66mhz_capable(struct pci_channel *hose, int top_bus, int current_bus); -- cgit v0.10.2 From ef407beefbd9928792ccc93857e408e0057bc17b Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 1 Feb 2010 16:39:46 +0900 Subject: sh: Hook up ERR/PERR/SERR detection for SH7780 PCI host controllers. These were never handled before, so implement some common infrastructure to support them, then make use of that in the SH7780-specific code. In practice there is little here that can not be generalized for SH4 parts, which will be an incremental change as the 7780/7751 code is gradually unified. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/common.c b/arch/sh/drivers/pci/common.c index f67c946..25aec00 100644 --- a/arch/sh/drivers/pci/common.c +++ b/arch/sh/drivers/pci/common.c @@ -1,4 +1,6 @@ #include +#include +#include #include static int __init @@ -62,3 +64,80 @@ int __init pci_is_66mhz_capable(struct pci_channel *hose, return cap66 > 0; } + +static void pcibios_enable_err(unsigned long __data) +{ + struct pci_channel *hose = (struct pci_channel *)__data; + + del_timer(&hose->err_timer); + printk(KERN_DEBUG "PCI: re-enabling error IRQ.\n"); + enable_irq(hose->err_irq); +} + +static void pcibios_enable_serr(unsigned long __data) +{ + struct pci_channel *hose = (struct pci_channel *)__data; + + del_timer(&hose->serr_timer); + printk(KERN_DEBUG "PCI: re-enabling system error IRQ.\n"); + enable_irq(hose->serr_irq); +} + +void pcibios_enable_timers(struct pci_channel *hose) +{ + if (hose->err_irq) { + init_timer(&hose->err_timer); + hose->err_timer.data = (unsigned long)hose; + hose->err_timer.function = pcibios_enable_err; + } + + if (hose->serr_irq) { + init_timer(&hose->serr_timer); + hose->serr_timer.data = (unsigned long)hose; + hose->serr_timer.function = pcibios_enable_serr; + } +} + +/* + * A simple handler for the regular PCI status errors, called from IRQ + * context. + */ +unsigned int pcibios_handle_status_errors(unsigned long addr, + unsigned int status, + struct pci_channel *hose) +{ + unsigned int cmd = 0; + + if (status & PCI_STATUS_REC_MASTER_ABORT) { + printk(KERN_DEBUG "PCI: master abort, pc=0x%08lx\n", addr); + cmd |= PCI_STATUS_REC_MASTER_ABORT; + } + + if (status & PCI_STATUS_REC_TARGET_ABORT) { + printk(KERN_DEBUG "PCI: target abort: "); + pcibios_report_status(PCI_STATUS_REC_TARGET_ABORT | + PCI_STATUS_SIG_TARGET_ABORT | + PCI_STATUS_REC_MASTER_ABORT, 1); + printk("\n"); + + cmd |= PCI_STATUS_REC_TARGET_ABORT; + } + + if (status & (PCI_STATUS_PARITY | PCI_STATUS_DETECTED_PARITY)) { + printk(KERN_DEBUG "PCI: parity error detected: "); + pcibios_report_status(PCI_STATUS_PARITY | + PCI_STATUS_DETECTED_PARITY, 1); + printk("\n"); + + cmd |= PCI_STATUS_PARITY | PCI_STATUS_DETECTED_PARITY; + + /* Now back off of the IRQ for awhile */ + if (hose->err_irq) { + disable_irq(hose->err_irq); + hose->err_timer.expires = jiffies + HZ; + add_timer(&hose->err_timer); + } + } + + return cmd; +} diff --git a/arch/sh/drivers/pci/ops-sh4.c b/arch/sh/drivers/pci/ops-sh4.c index e55e81a..0b81999 100644 --- a/arch/sh/drivers/pci/ops-sh4.c +++ b/arch/sh/drivers/pci/ops-sh4.c @@ -16,7 +16,7 @@ * Direct access to PCI hardware... */ #define CONFIG_CMD(bus, devfn, where) \ - (P1SEG | (bus->number << 16) | (devfn << 8) | (where & ~3)) + (0x80000000 | (bus->number << 16) | (devfn << 8) | (where & ~3)) static DEFINE_SPINLOCK(sh4_pci_lock); diff --git a/arch/sh/drivers/pci/pci-sh7780.c b/arch/sh/drivers/pci/pci-sh7780.c index 0e0ddd6..8637331 100644 --- a/arch/sh/drivers/pci/pci-sh7780.c +++ b/arch/sh/drivers/pci/pci-sh7780.c @@ -11,6 +11,9 @@ #include #include #include +#include +#include +#include #include #include #include @@ -39,8 +42,165 @@ static struct pci_channel sh7780_pci_controller = { .io_resource = &sh7785_io_resource, .io_offset = 0x00000000, .io_map_base = SH7780_PCI_IO_BASE, + .serr_irq = evt2irq(0xa00), + .err_irq = evt2irq(0xaa0), }; +struct pci_errors { + unsigned int mask; + const char *str; +} pci_arbiter_errors[] = { + { SH4_PCIAINT_MBKN, "master broken" }, + { SH4_PCIAINT_TBTO, "target bus time out" }, + { SH4_PCIAINT_MBTO, "master bus time out" }, + { SH4_PCIAINT_TABT, "target abort" }, + { SH4_PCIAINT_MABT, "master abort" }, + { SH4_PCIAINT_RDPE, "read data parity error" }, + { SH4_PCIAINT_WDPE, "write data parity error" }, +}, pci_interrupt_errors[] = { + { SH4_PCIINT_MLCK, "master lock error" }, + { SH4_PCIINT_TABT, "target-target abort" }, + { SH4_PCIINT_TRET, "target retry time out" }, + { SH4_PCIINT_MFDE, "master function disable erorr" }, + { SH4_PCIINT_PRTY, "address parity error" }, + { SH4_PCIINT_SERR, "SERR" }, + { SH4_PCIINT_TWDP, "data parity error for target write" }, + { SH4_PCIINT_TRDP, "PERR detected for target read" }, + { SH4_PCIINT_MTABT, "target abort for master" }, + { SH4_PCIINT_MMABT, "master abort for master" }, + { SH4_PCIINT_MWPD, "master write data parity error" }, + { SH4_PCIINT_MRPD, "master read data parity error" }, +}; + +static irqreturn_t sh7780_pci_err_irq(int irq, void *dev_id) +{ + struct pci_channel *hose = dev_id; + unsigned long addr; + unsigned int status; + unsigned int cmd; + int i; + + addr = __raw_readl(hose->reg_base + SH4_PCIALR); + + /* + * Handle status errors. + */ + status = __raw_readw(hose->reg_base + PCI_STATUS); + if (status & (PCI_STATUS_PARITY | + PCI_STATUS_DETECTED_PARITY | + PCI_STATUS_SIG_TARGET_ABORT | + PCI_STATUS_REC_TARGET_ABORT | + PCI_STATUS_REC_MASTER_ABORT)) { + cmd = pcibios_handle_status_errors(addr, status, hose); + if (likely(cmd)) + __raw_writew(cmd, hose->reg_base + PCI_STATUS); + } + + /* + * Handle arbiter errors. + */ + status = __raw_readl(hose->reg_base + SH4_PCIAINT); + for (i = cmd = 0; i < ARRAY_SIZE(pci_arbiter_errors); i++) { + if (status & pci_arbiter_errors[i].mask) { + printk(KERN_DEBUG "PCI: %s, addr=%08lx\n", + pci_arbiter_errors[i].str, addr); + cmd |= pci_arbiter_errors[i].mask; + } + } + __raw_writel(cmd, hose->reg_base + SH4_PCIAINT); + + /* + * Handle the remaining PCI errors. + */ + status = __raw_readl(hose->reg_base + SH4_PCIINT); + for (i = cmd = 0; i < ARRAY_SIZE(pci_interrupt_errors); i++) { + if (status & pci_interrupt_errors[i].mask) { + printk(KERN_DEBUG "PCI: %s, addr=%08lx\n", + pci_interrupt_errors[i].str, addr); + cmd |= pci_interrupt_errors[i].mask; + } + } + __raw_writel(cmd, hose->reg_base + SH4_PCIINT); + + return IRQ_HANDLED; +} + +static irqreturn_t sh7780_pci_serr_irq(int irq, void *dev_id) +{ + struct pci_channel *hose = dev_id; + + printk(KERN_DEBUG "PCI: system error received: "); + pcibios_report_status(PCI_STATUS_SIG_SYSTEM_ERROR, 1); + printk("\n"); + + /* Deassert SERR */ + __raw_writel(SH4_PCIINTM_SDIM, hose->reg_base + SH4_PCIINTM); + + /* Back off the IRQ for awhile */ + disable_irq(irq); + hose->serr_timer.expires = jiffies + HZ; + add_timer(&hose->serr_timer); + + return IRQ_HANDLED; +} + +static int __init sh7780_pci_setup_irqs(struct pci_channel *hose) +{ + int ret; + + /* Clear out PCI arbiter IRQs */ + __raw_writel(0, hose->reg_base + SH4_PCIAINT); + + /* Clear all error conditions */ + __raw_writew(PCI_STATUS_DETECTED_PARITY | \ + PCI_STATUS_SIG_SYSTEM_ERROR | \ + PCI_STATUS_REC_MASTER_ABORT | \ + PCI_STATUS_REC_TARGET_ABORT | \ + PCI_STATUS_SIG_TARGET_ABORT | \ + PCI_STATUS_PARITY, hose->reg_base + PCI_STATUS); + + ret = request_irq(hose->serr_irq, sh7780_pci_serr_irq, IRQF_DISABLED, + "PCI SERR interrupt", hose); + if (unlikely(ret)) { + printk(KERN_ERR "PCI: Failed hooking SERR IRQ\n"); + return ret; + } + + /* + * The PCI ERR IRQ needs to be IRQF_SHARED since all of the power + * down IRQ vectors are routed through the ERR IRQ vector. We + * only request_irq() once as there is only a single masking + * source for multiple events. + */ + ret = request_irq(hose->err_irq, sh7780_pci_err_irq, IRQF_SHARED, + "PCI ERR interrupt", hose); + if (unlikely(ret)) { + free_irq(hose->serr_irq, hose); + return ret; + } + + /* Unmask all of the arbiter IRQs. */ + __raw_writel(SH4_PCIAINT_MBKN | SH4_PCIAINT_TBTO | SH4_PCIAINT_MBTO | \ + SH4_PCIAINT_TABT | SH4_PCIAINT_MABT | SH4_PCIAINT_RDPE | \ + SH4_PCIAINT_WDPE, hose->reg_base + SH4_PCIAINTM); + + /* Unmask all of the PCI IRQs */ + __raw_writel(SH4_PCIINTM_TTADIM | SH4_PCIINTM_TMTOIM | \ + SH4_PCIINTM_MDEIM | SH4_PCIINTM_APEDIM | \ + SH4_PCIINTM_SDIM | SH4_PCIINTM_DPEITWM | \ + SH4_PCIINTM_PEDITRM | SH4_PCIINTM_TADIMM | \ + SH4_PCIINTM_MADIMM | SH4_PCIINTM_MWPDIM | \ + SH4_PCIINTM_MRDPEIM, hose->reg_base + SH4_PCIINTM); + + return ret; +} + +static inline void __init sh7780_pci_teardown_irqs(struct pci_channel *hose) +{ + free_irq(hose->err_irq, hose); + free_irq(hose->serr_irq, hose); +} + static void __init sh7780_pci66_init(struct pci_channel *hose) { unsigned int tmp; @@ -149,33 +309,12 @@ static int __init sh7780_pci_init(void) __raw_writel(((memsize - SZ_1M) & 0x1ff00000) | 1, chan->reg_base + SH4_PCILSR0); - /* Clear out PCI arbiter IRQs */ - __raw_writel(0, chan->reg_base + SH4_PCIAINT); - - /* Unmask all of the arbiter IRQs. */ - __raw_writel(SH4_PCIAINT_MBKN | SH4_PCIAINT_TBTO | SH4_PCIAINT_MBTO | \ - SH4_PCIAINT_TABT | SH4_PCIAINT_MABT | SH4_PCIAINT_RDPE | \ - SH4_PCIAINT_WDPE, chan->reg_base + SH4_PCIAINTM); - - /* Clear all error conditions */ - __raw_writew(PCI_STATUS_DETECTED_PARITY | \ - PCI_STATUS_SIG_SYSTEM_ERROR | \ - PCI_STATUS_REC_MASTER_ABORT | \ - PCI_STATUS_REC_TARGET_ABORT | \ - PCI_STATUS_SIG_TARGET_ABORT | \ - PCI_STATUS_PARITY, chan->reg_base + PCI_STATUS); - - __raw_writew(PCI_COMMAND_SERR | PCI_COMMAND_WAIT | \ - PCI_COMMAND_PARITY | PCI_COMMAND_MASTER | \ - PCI_COMMAND_MEMORY, chan->reg_base + PCI_COMMAND); - - /* Unmask all of the PCI IRQs */ - __raw_writel(SH4_PCIINTM_TTADIM | SH4_PCIINTM_TMTOIM | \ - SH4_PCIINTM_MDEIM | SH4_PCIINTM_APEDIM | \ - SH4_PCIINTM_SDIM | SH4_PCIINTM_DPEITWM | \ - SH4_PCIINTM_PEDITRM | SH4_PCIINTM_TADIMM | \ - SH4_PCIINTM_MADIMM | SH4_PCIINTM_MWPDIM | \ - SH4_PCIINTM_MRDPEIM, chan->reg_base + SH4_PCIINTM); + /* + * Hook up the ERR and SERR IRQs. + */ + ret = sh7780_pci_setup_irqs(chan); + if (unlikely(ret)) + return ret; /* * Disable the cache snoop controller for non-coherent DMA. @@ -191,6 +330,10 @@ static int __init sh7780_pci_init(void) __raw_writel(0, chan->reg_base + SH7780_PCIIOBR); __raw_writel(0, chan->reg_base + SH7780_PCIIOBMR); + __raw_writew(PCI_COMMAND_SERR | PCI_COMMAND_WAIT | \ + PCI_COMMAND_PARITY | PCI_COMMAND_MASTER | \ + PCI_COMMAND_MEMORY, chan->reg_base + PCI_COMMAND); + /* * Initialization mode complete, release the control register and * enable round robin mode to stop device overruns/starvation. @@ -200,7 +343,7 @@ static int __init sh7780_pci_init(void) ret = register_pci_controller(chan); if (unlikely(ret)) - return ret; + goto err; sh7780_pci66_init(chan); @@ -209,5 +352,9 @@ static int __init sh7780_pci_init(void) 66 : 33); return 0; + +err: + sh7780_pci_teardown_irqs(chan); + return ret; } arch_initcall(sh7780_pci_init); diff --git a/arch/sh/drivers/pci/pci.c b/arch/sh/drivers/pci/pci.c index 488331c..8e42dfb 100644 --- a/arch/sh/drivers/pci/pci.c +++ b/arch/sh/drivers/pci/pci.c @@ -79,6 +79,11 @@ int __devinit register_pci_controller(struct pci_channel *hose) } /* + * Setup the ERR/PERR and SERR timers, if available. + */ + pcibios_enable_timers(hose); + + /* * Scan the bus if it is register after the PCI subsystem * initialization. */ @@ -289,6 +294,52 @@ char * __devinit pcibios_setup(char *str) return str; } +/* + * We can't use pci_find_device() here since we are + * called from interrupt context. + */ +static void pcibios_bus_report_status(struct pci_bus *bus, + unsigned int status_mask, int warn) +{ + struct pci_dev *dev; + + list_for_each_entry(dev, &bus->devices, bus_list) { + u16 status; + + /* + * ignore host bridge - we handle + * that separately + */ + if (dev->bus->number == 0 && dev->devfn == 0) + continue; + + pci_read_config_word(dev, PCI_STATUS, &status); + if (status == 0xffff) + continue; + + if ((status & status_mask) == 0) + continue; + + /* clear the status errors */ + pci_write_config_word(dev, PCI_STATUS, status & status_mask); + + if (warn) + printk("(%s: %04X) ", pci_name(dev), status); + } + + list_for_each_entry(dev, &bus->devices, bus_list) + if (dev->subordinate) + pcibios_bus_report_status(dev->subordinate, status_mask, warn); +} + +void pcibios_report_status(unsigned int status_mask, int warn) +{ + struct pci_channel *hose; + + for (hose = hose_head; hose; hose = hose->next) + pcibios_bus_report_status(hose->bus, status_mask, warn); +} + int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma, enum pci_mmap_state mmap_state, int write_combine) { diff --git a/arch/sh/include/asm/pci.h b/arch/sh/include/asm/pci.h index bbd10cf..1de83f2 100644 --- a/arch/sh/include/asm/pci.h +++ b/arch/sh/include/asm/pci.h @@ -29,9 +29,20 @@ struct pci_channel { unsigned int index; unsigned int need_domain_info; + + /* Optional error handling */ + struct timer_list err_timer, serr_timer; + unsigned int err_irq, serr_irq; }; +/* arch/sh/drivers/pci/pci.c */ extern int register_pci_controller(struct pci_channel *hose); +extern void pcibios_report_status(unsigned int status_mask, int warn); + +/* arch/sh/drivers/pci/common.c */ +extern void pcibios_enable_timers(struct pci_channel *hose); +extern unsigned int pcibios_handle_status_errors(unsigned long addr, + unsigned int status, struct pci_channel *hose); extern int pci_is_66mhz_capable(struct pci_channel *hose, int top_bus, int current_bus); -- cgit v0.10.2 From b6c58b1d987a5795086c5c2babd8c7367d2fdb8c Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 1 Feb 2010 20:01:50 +0900 Subject: sh: Improved multi-resource handling for SH7780 PCI. The SH7780 PCI controller supports 3 different ranges of PCI memory in addition to its PCI I/O window. In the case of 29-bit mode, only 2 memory windows are supported, while in 32-bit mode all 3 are visible. This attempts to make the resource handling completely dynamic and to permit platforms to map in as many apertures as they can handle. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/fixups-dreamcast.c b/arch/sh/drivers/pci/fixups-dreamcast.c index ed7f489..942ef4f 100644 --- a/arch/sh/drivers/pci/fixups-dreamcast.c +++ b/arch/sh/drivers/pci/fixups-dreamcast.c @@ -39,7 +39,7 @@ static void __init gapspci_fixup_resources(struct pci_dev *dev) /* * We also assume that dev->devfn == 0 */ - dev->resource[1].start = p->io_resource->start + 0x100; + dev->resource[1].start = p->resources[0].start + 0x100; dev->resource[1].end = dev->resource[1].start + 0x200 - 1; /* diff --git a/arch/sh/drivers/pci/fixups-se7751.c b/arch/sh/drivers/pci/fixups-se7751.c index 475fa9f..a4c7d3a 100644 --- a/arch/sh/drivers/pci/fixups-se7751.c +++ b/arch/sh/drivers/pci/fixups-se7751.c @@ -97,12 +97,12 @@ int pci_fixup_pcic(struct pci_channel *chan) * meaning all calls go straight through... use BUG_ON to * catch erroneous assumption. */ - BUG_ON(chan->mem_resource->start != SH7751_PCI_MEMORY_BASE); + BUG_ON(chan->resources[1].start != SH7751_PCI_MEMORY_BASE); - PCIC_WRITE(SH7751_PCIMBR, chan->mem_resource->start); + PCIC_WRITE(SH7751_PCIMBR, chan->resources[1].start); /* Set IOBR for window containing area specified in pci.h */ - PCIC_WRITE(SH7751_PCIIOBR, (chan->io_resource->start & SH7751_PCIIOBR_MASK)); + PCIC_WRITE(SH7751_PCIIOBR, (chan->resources[0].start & SH7751_PCIIOBR_MASK)); /* All done, may as well say so... */ printk("SH7751 PCI: Finished initialization of the PCI controller\n"); diff --git a/arch/sh/drivers/pci/pci-dreamcast.c b/arch/sh/drivers/pci/pci-dreamcast.c index bd5a1e5..6336941 100644 --- a/arch/sh/drivers/pci/pci-dreamcast.c +++ b/arch/sh/drivers/pci/pci-dreamcast.c @@ -25,25 +25,25 @@ #include #include -static struct resource gapspci_io_resource = { - .name = "GAPSPCI IO", - .start = GAPSPCI_BBA_CONFIG, - .end = GAPSPCI_BBA_CONFIG + GAPSPCI_BBA_CONFIG_SIZE - 1, - .flags = IORESOURCE_IO, -}; - -static struct resource gapspci_mem_resource = { - .name = "GAPSPCI mem", - .start = GAPSPCI_DMA_BASE, - .end = GAPSPCI_DMA_BASE + GAPSPCI_DMA_SIZE - 1, - .flags = IORESOURCE_MEM, +static struct resource gapspci_resources[] = { + { + .name = "GAPSPCI IO", + .start = GAPSPCI_BBA_CONFIG, + .end = GAPSPCI_BBA_CONFIG + GAPSPCI_BBA_CONFIG_SIZE - 1, + .flags = IORESOURCE_IO, + }, { + .name = "GAPSPCI mem", + .start = GAPSPCI_DMA_BASE, + .end = GAPSPCI_DMA_BASE + GAPSPCI_DMA_SIZE - 1, + .flags = IORESOURCE_MEM, + }, }; static struct pci_channel dreamcast_pci_controller = { .pci_ops = &gapspci_pci_ops, - .io_resource = &gapspci_io_resource, + .resources = gapspci_resources, + .nr_resources = ARRAY_SIZE(gapspci_resources), .io_offset = 0x00000000, - .mem_resource = &gapspci_mem_resource, .mem_offset = 0x00000000, }; diff --git a/arch/sh/drivers/pci/pci-sh5.c b/arch/sh/drivers/pci/pci-sh5.c index bce73fa..0bf296c 100644 --- a/arch/sh/drivers/pci/pci-sh5.c +++ b/arch/sh/drivers/pci/pci-sh5.c @@ -89,14 +89,13 @@ static irqreturn_t pcish5_serr_irq(int irq, void *dev_id) return IRQ_NONE; } -static struct resource sh5_io_resource = { /* place holder */ }; -static struct resource sh5_mem_resource = { /* place holder */ }; +static struct resource sh5_pci_resources[2]; static struct pci_channel sh5pci_controller = { .pci_ops = &sh5_pci_ops, - .mem_resource = &sh5_mem_resource, + .resources = sh5_pci_resources, + .nr_resources = ARRAY_SIZE(sh5_pci_resources), .mem_offset = 0x00000000, - .io_resource = &sh5_io_resource, .io_offset = 0x00000000, }; @@ -210,11 +209,11 @@ static int __init sh5pci_init(void) SH5PCI_WRITE(AINTM, ~0); SH5PCI_WRITE(PINTM, ~0); - sh5_io_resource.start = PCI_IO_AREA; - sh5_io_resource.end = PCI_IO_AREA + 0x10000; + sh5_pci_resources[0].start = PCI_IO_AREA; + sh5_pci_resources[0].end = PCI_IO_AREA + 0x10000; - sh5_mem_resource.start = memStart; - sh5_mem_resource.end = memStart + memSize; + sh5_pci_resources[1].start = memStart; + sh5_pci_resources[1].end = memStart + memSize; return register_pci_controller(&sh5pci_controller); } diff --git a/arch/sh/drivers/pci/pci-sh7751.c b/arch/sh/drivers/pci/pci-sh7751.c index 6ad5beb..17811e5 100644 --- a/arch/sh/drivers/pci/pci-sh7751.c +++ b/arch/sh/drivers/pci/pci-sh7751.c @@ -44,25 +44,25 @@ static int __init __area_sdram_check(struct pci_channel *chan, return 1; } -static struct resource sh7751_io_resource = { - .name = "SH7751_IO", - .start = SH7751_PCI_IO_BASE, - .end = SH7751_PCI_IO_BASE + SH7751_PCI_IO_SIZE - 1, - .flags = IORESOURCE_IO -}; - -static struct resource sh7751_mem_resource = { - .name = "SH7751_mem", - .start = SH7751_PCI_MEMORY_BASE, - .end = SH7751_PCI_MEMORY_BASE + SH7751_PCI_MEM_SIZE - 1, - .flags = IORESOURCE_MEM +static struct resource sh7751_pci_resources[] = { + { + .name = "SH7751_IO", + .start = SH7751_PCI_IO_BASE, + .end = SH7751_PCI_IO_BASE + SH7751_PCI_IO_SIZE - 1, + .flags = IORESOURCE_IO + }, { + .name = "SH7751_mem", + .start = SH7751_PCI_MEMORY_BASE, + .end = SH7751_PCI_MEMORY_BASE + SH7751_PCI_MEM_SIZE - 1, + .flags = IORESOURCE_MEM + }, }; static struct pci_channel sh7751_pci_controller = { .pci_ops = &sh4_pci_ops, - .mem_resource = &sh7751_mem_resource, + .resources = sh7751_pci_resources, + .nr_resources = ARRAY_SIZE(sh7751_pci_resources), .mem_offset = 0x00000000, - .io_resource = &sh7751_io_resource, .io_offset = 0x00000000, .io_map_base = SH7751_PCI_IO_BASE, }; @@ -128,13 +128,13 @@ static int __init sh7751_pci_init(void) /* Set the local 16MB PCI memory space window to * the lowest PCI mapped address */ - word = chan->mem_resource->start & SH4_PCIMBR_MASK; + word = chan->resources[1].start & SH4_PCIMBR_MASK; pr_debug("PCI: Setting upper bits of Memory window to 0x%x\n", word); pci_write_reg(chan, word , SH4_PCIMBR); /* Make sure the MSB's of IO window are set to access PCI space * correctly */ - word = chan->io_resource->start & SH4_PCIIOBR_MASK; + word = chan->resources[0].start & SH4_PCIIOBR_MASK; pr_debug("PCI: Setting upper bits of IO window to 0x%x\n", word); pci_write_reg(chan, word, SH4_PCIIOBR); diff --git a/arch/sh/drivers/pci/pci-sh7780.c b/arch/sh/drivers/pci/pci-sh7780.c index 8637331..472f67a 100644 --- a/arch/sh/drivers/pci/pci-sh7780.c +++ b/arch/sh/drivers/pci/pci-sh7780.c @@ -21,27 +21,40 @@ #include #include -static struct resource sh7785_io_resource = { - .name = "SH7785_IO", - .start = 0x1000, - .end = SH7780_PCI_IO_SIZE - 1, - .flags = IORESOURCE_IO -}; - -static struct resource sh7785_mem_resource = { - .name = "SH7785_mem", - .start = SH7780_PCI_MEMORY_BASE, - .end = SH7780_PCI_MEMORY_BASE + SH7780_PCI_MEM_SIZE - 1, - .flags = IORESOURCE_MEM +static struct resource sh7785_pci_resources[] = { + { + .name = "SH7785_IO", + .start = 0x1000, + .end = SZ_4M - 1, + .flags = IORESOURCE_IO, + }, { + .name = "PCI MEM 0", + .start = 0xfd000000, + .end = 0xfd000000 + SZ_16M - 1, + .flags = IORESOURCE_MEM, + }, { + .name = "PCI MEM 1", + .start = 0x10000000, + .end = 0x10000000 + SZ_64M - 1, + .flags = IORESOURCE_MEM, + }, { + /* + * 32-bit only resources must be last. + */ + .name = "PCI MEM 2", + .start = 0xc0000000, + .end = 0xc0000000 + SZ_512M - 1, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_32BIT, + }, }; static struct pci_channel sh7780_pci_controller = { .pci_ops = &sh4_pci_ops, - .mem_resource = &sh7785_mem_resource, - .mem_offset = 0x00000000, - .io_resource = &sh7785_io_resource, - .io_offset = 0x00000000, - .io_map_base = SH7780_PCI_IO_BASE, + .resources = sh7785_pci_resources, + .nr_resources = ARRAY_SIZE(sh7785_pci_resources), + .io_offset = 0, + .mem_offset = 0, + .io_map_base = 0xfe200000, .serr_irq = evt2irq(0xa00), .err_irq = evt2irq(0xaa0), }; @@ -231,7 +244,7 @@ static int __init sh7780_pci_init(void) size_t memsize; unsigned int id; const char *type; - int ret; + int ret, i; printk(KERN_NOTICE "PCI: Starting intialization.\n"); @@ -279,8 +292,6 @@ static int __init sh7780_pci_init(void) */ __raw_writel(SH4_PCICR_PREFIX, chan->reg_base + SH4_PCICR); - __raw_writel(0, chan->reg_base + PCI_BASE_ADDRESS_0); - memphys = __pa(memory_start); memsize = roundup_pow_of_two(memory_end - memory_start); @@ -324,9 +335,40 @@ static int __init sh7780_pci_init(void) __raw_writel(0, chan->reg_base + SH7780_PCICSCR1); __raw_writel(0, chan->reg_base + SH7780_PCICSAR1); - __raw_writel(0xfd000000, chan->reg_base + SH7780_PCIMBR0); - __raw_writel(0x00fc0000, chan->reg_base + SH7780_PCIMBMR0); + /* + * Setup the memory BARs + */ + for (i = 0; i < chan->nr_resources; i++) { + struct resource *res = chan->resources + (i + 1); + resource_size_t size; + + if (unlikely(res->flags & IORESOURCE_IO)) + continue; + + /* + * Make sure we're in the right physical addressing mode + * for dealing with the resource. + */ + if ((res->flags & IORESOURCE_MEM_32BIT) && __in_29bit_mode()) { + chan->nr_resources--; + continue; + } + size = resource_size(res); + + /* + * The MBMR mask is calculated in units of 256kB, which + * keeps things pretty simple. + */ + __raw_writel(((roundup_pow_of_two(size) / SZ_256K) - 1) << 18, + chan->reg_base + SH7780_PCIMBMR(i)); + __raw_writel(res->start, chan->reg_base + SH7780_PCIMBR(i)); + } + + /* + * And I/O. + */ + __raw_writel(0, chan->reg_base + PCI_BASE_ADDRESS_0); __raw_writel(0, chan->reg_base + SH7780_PCIIOBR); __raw_writel(0, chan->reg_base + SH7780_PCIIOBMR); diff --git a/arch/sh/drivers/pci/pci-sh7780.h b/arch/sh/drivers/pci/pci-sh7780.h index dee069c..205dcbe 100644 --- a/arch/sh/drivers/pci/pci-sh7780.h +++ b/arch/sh/drivers/pci/pci-sh7780.h @@ -26,12 +26,6 @@ #define SH7780_PCI_CONFIG_BASE 0xFD000000 /* Config space base addr */ #define SH7780_PCI_CONFIG_SIZE 0x01000000 /* Config space size */ -#define SH7780_PCI_MEMORY_BASE 0xFD000000 /* Memory space base addr */ -#define SH7780_PCI_MEM_SIZE 0x01000000 /* Size of Memory window */ - -#define SH7780_PCI_IO_BASE 0xFE200000 /* IO space base address */ -#define SH7780_PCI_IO_SIZE 0x00400000 /* Size of IO window */ - #define SH7780_PCIREG_BASE 0xFE040000 /* PCI regs base address */ /* SH7780 PCI Config Registers */ @@ -46,12 +40,8 @@ #define SH7780_PCIPINT 0x1CC /* Power Mgmnt Int. Register */ #define SH7780_PCIPINTM 0x1D0 /* Power Mgmnt Mask Register */ -#define SH7780_PCIMBR0 0x1E0 -#define SH7780_PCIMBMR0 0x1E4 -#define SH7780_PCIMBR1 0x1E8 -#define SH7780_PCIMBMR1 0x1EC -#define SH7780_PCIMBR2 0x1F0 -#define SH7780_PCIMBMR2 0x1F4 +#define SH7780_PCIMBR(x) (0x1E0 + ((x) * 8)) +#define SH7780_PCIMBMR(x) (0x1E4 + ((x) * 8)) #define SH7780_PCIIOBR 0x1F8 #define SH7780_PCIIOBMR 0x1FC #define SH7780_PCICSCR0 0x210 /* Cache Snoop1 Cnt. Register */ diff --git a/arch/sh/drivers/pci/pci.c b/arch/sh/drivers/pci/pci.c index 8e42dfb..f4a6983 100644 --- a/arch/sh/drivers/pci/pci.c +++ b/arch/sh/drivers/pci/pci.c @@ -60,11 +60,18 @@ static DEFINE_MUTEX(pci_scan_mutex); int __devinit register_pci_controller(struct pci_channel *hose) { - if (request_resource(&iomem_resource, hose->mem_resource) < 0) - goto out; - if (request_resource(&ioport_resource, hose->io_resource) < 0) { - release_resource(hose->mem_resource); - goto out; + int i; + + for (i = 0; i < hose->nr_resources; i++) { + struct resource *res = hose->resources + i; + + if (res->flags & IORESOURCE_IO) { + if (request_resource(&ioport_resource, res) < 0) + goto out; + } else { + if (request_resource(&iomem_resource, res) < 0) + goto out; + } } *hose_tail = hose; @@ -96,6 +103,9 @@ int __devinit register_pci_controller(struct pci_channel *hose) return 0; out: + for (--i; i >= 0; i--) + release_resource(&hose->resources[i]); + printk(KERN_WARNING "Skipping PCI bus scan due to resource conflict\n"); return -1; } @@ -149,11 +159,13 @@ void __devinit pcibios_fixup_bus(struct pci_bus *bus) { struct pci_dev *dev = bus->self; struct list_head *ln; - struct pci_channel *chan = bus->sysdata; + struct pci_channel *hose = bus->sysdata; if (!dev) { - bus->resource[0] = chan->io_resource; - bus->resource[1] = chan->mem_resource; + int i; + + for (i = 0; i < hose->nr_resources; i++) + bus->resource[i] = hose->resources + i; } for (ln = bus->devices.next; ln != &bus->devices; ln = ln->next) { @@ -174,21 +186,18 @@ void pcibios_align_resource(void *data, struct resource *res, resource_size_t size, resource_size_t align) { struct pci_dev *dev = data; - struct pci_channel *chan = dev->sysdata; + struct pci_channel *hose = dev->sysdata; resource_size_t start = res->start; if (res->flags & IORESOURCE_IO) { - if (start < PCIBIOS_MIN_IO + chan->io_resource->start) - start = PCIBIOS_MIN_IO + chan->io_resource->start; + if (start < PCIBIOS_MIN_IO + hose->resources[0].start) + start = PCIBIOS_MIN_IO + hose->resources[0].start; /* * Put everything into 0x00-0xff region modulo 0x400. */ if (start & 0x300) start = (start + 0x3ff) & ~0x3ff; - } else if (res->flags & IORESOURCE_MEM) { - if (start < PCIBIOS_MIN_MEM + chan->mem_resource->start) - start = PCIBIOS_MIN_MEM + chan->mem_resource->start; } res->start = start; diff --git a/arch/sh/include/asm/pci.h b/arch/sh/include/asm/pci.h index 1de83f2..6d762cc 100644 --- a/arch/sh/include/asm/pci.h +++ b/arch/sh/include/asm/pci.h @@ -18,8 +18,9 @@ struct pci_channel { struct pci_bus *bus; struct pci_ops *pci_ops; - struct resource *io_resource; - struct resource *mem_resource; + + struct resource *resources; + unsigned int nr_resources; unsigned long io_offset; unsigned long mem_offset; -- cgit v0.10.2 From 3b9cfc0a99f88c0db7c72363620584a9b40b4543 Mon Sep 17 00:00:00 2001 From: Emese Revfy Date: Sun, 31 Jan 2010 20:16:34 +0100 Subject: x86, mtrr: Constify struct mtrr_ops This is part of the ops structure constification effort started by Arjan van de Ven et al. Benefits of this constification: * prevents modification of data that is shared (referenced) by many other structure instances at runtime * detects/prevents accidental (but not intentional) modification attempts on archs that enforce read-only kernel data at runtime * potentially better optimized code as the compiler can assume that the const data cannot be changed * the compiler/linker move const data into .rodata and therefore exclude them from false sharing Signed-off-by: Emese Revfy LKML-Reference: <4B65D712.3080804@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/mtrr/amd.c b/arch/x86/kernel/cpu/mtrr/amd.c index 33af141..92ba9cd 100644 --- a/arch/x86/kernel/cpu/mtrr/amd.c +++ b/arch/x86/kernel/cpu/mtrr/amd.c @@ -108,7 +108,7 @@ amd_validate_add_page(unsigned long base, unsigned long size, unsigned int type) return 0; } -static struct mtrr_ops amd_mtrr_ops = { +static const struct mtrr_ops amd_mtrr_ops = { .vendor = X86_VENDOR_AMD, .set = amd_set_mtrr, .get = amd_get_mtrr, diff --git a/arch/x86/kernel/cpu/mtrr/centaur.c b/arch/x86/kernel/cpu/mtrr/centaur.c index de89f14..316fe3e 100644 --- a/arch/x86/kernel/cpu/mtrr/centaur.c +++ b/arch/x86/kernel/cpu/mtrr/centaur.c @@ -110,7 +110,7 @@ centaur_validate_add_page(unsigned long base, unsigned long size, unsigned int t return 0; } -static struct mtrr_ops centaur_mtrr_ops = { +static const struct mtrr_ops centaur_mtrr_ops = { .vendor = X86_VENDOR_CENTAUR, .set = centaur_set_mcr, .get = centaur_get_mcr, diff --git a/arch/x86/kernel/cpu/mtrr/cyrix.c b/arch/x86/kernel/cpu/mtrr/cyrix.c index 228d982..68a3343 100644 --- a/arch/x86/kernel/cpu/mtrr/cyrix.c +++ b/arch/x86/kernel/cpu/mtrr/cyrix.c @@ -265,7 +265,7 @@ static void cyrix_set_all(void) post_set(); } -static struct mtrr_ops cyrix_mtrr_ops = { +static const struct mtrr_ops cyrix_mtrr_ops = { .vendor = X86_VENDOR_CYRIX, .set_all = cyrix_set_all, .set = cyrix_set_arr, diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c index 55da0c5..4d75584 100644 --- a/arch/x86/kernel/cpu/mtrr/generic.c +++ b/arch/x86/kernel/cpu/mtrr/generic.c @@ -752,7 +752,7 @@ int positive_have_wrcomb(void) /* * Generic structure... */ -struct mtrr_ops generic_mtrr_ops = { +const struct mtrr_ops generic_mtrr_ops = { .use_intel_if = 1, .set_all = generic_set_all, .get = generic_get_mtrr, diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c index 84e83de..fe4622e 100644 --- a/arch/x86/kernel/cpu/mtrr/main.c +++ b/arch/x86/kernel/cpu/mtrr/main.c @@ -60,14 +60,14 @@ static DEFINE_MUTEX(mtrr_mutex); u64 size_or_mask, size_and_mask; static bool mtrr_aps_delayed_init; -static struct mtrr_ops *mtrr_ops[X86_VENDOR_NUM]; +static const struct mtrr_ops *mtrr_ops[X86_VENDOR_NUM]; -struct mtrr_ops *mtrr_if; +const struct mtrr_ops *mtrr_if; static void set_mtrr(unsigned int reg, unsigned long base, unsigned long size, mtrr_type type); -void set_mtrr_ops(struct mtrr_ops *ops) +void set_mtrr_ops(const struct mtrr_ops *ops) { if (ops->vendor && ops->vendor < X86_VENDOR_NUM) mtrr_ops[ops->vendor] = ops; diff --git a/arch/x86/kernel/cpu/mtrr/mtrr.h b/arch/x86/kernel/cpu/mtrr/mtrr.h index a501dee..df5e41f 100644 --- a/arch/x86/kernel/cpu/mtrr/mtrr.h +++ b/arch/x86/kernel/cpu/mtrr/mtrr.h @@ -32,7 +32,7 @@ extern int generic_get_free_region(unsigned long base, unsigned long size, extern int generic_validate_add_page(unsigned long base, unsigned long size, unsigned int type); -extern struct mtrr_ops generic_mtrr_ops; +extern const struct mtrr_ops generic_mtrr_ops; extern int positive_have_wrcomb(void); @@ -53,10 +53,10 @@ void fill_mtrr_var_range(unsigned int index, u32 base_lo, u32 base_hi, u32 mask_lo, u32 mask_hi); void get_mtrr_state(void); -extern void set_mtrr_ops(struct mtrr_ops *ops); +extern void set_mtrr_ops(const struct mtrr_ops *ops); extern u64 size_or_mask, size_and_mask; -extern struct mtrr_ops *mtrr_if; +extern const struct mtrr_ops *mtrr_if; #define is_cpu(vnd) (mtrr_if && mtrr_if->vendor == X86_VENDOR_##vnd) #define use_intel() (mtrr_if && mtrr_if->use_intel_if == 1) -- cgit v0.10.2 From 20026d92013d7bb3abb295337191def6758fc086 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 4 Feb 2010 09:48:58 +1100 Subject: xfs: kill the unused XFS_QMOPT_* flush flags V2 dquots are never flushed asynchronously. Remove the flag and the async write support from the flush function. Make the default flush a delwri flush to make the inode flush code, which leaves the XFS_QMOPT_SYNC the only flag remaining. Convert that to use SYNC_WAIT instead, just like the inode flush code. V2: - just pass flush flags straight through Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/quota/xfs_dquot.c b/fs/xfs/quota/xfs_dquot.c index 1620a56..5f79dd7 100644 --- a/fs/xfs/quota/xfs_dquot.c +++ b/fs/xfs/quota/xfs_dquot.c @@ -1187,7 +1187,7 @@ xfs_qm_dqflush( * block, nada. */ if (!XFS_DQ_IS_DIRTY(dqp) || - (!(flags & XFS_QMOPT_SYNC) && atomic_read(&dqp->q_pincount) > 0)) { + (!(flags & SYNC_WAIT) && atomic_read(&dqp->q_pincount) > 0)) { xfs_dqfunlock(dqp); return 0; } @@ -1251,18 +1251,17 @@ xfs_qm_dqflush( xfs_log_force(mp, 0); } - if (flags & XFS_QMOPT_DELWRI) { - xfs_bdwrite(mp, bp); - } else { + if (flags & SYNC_WAIT) error = xfs_bwrite(mp, bp); - } + else + xfs_bdwrite(mp, bp); trace_xfs_dqflush_done(dqp); /* * dqp is still locked, but caller is free to unlock it now. */ - return (error); + return error; } @@ -1443,7 +1442,7 @@ xfs_qm_dqpurge( * We don't care about getting disk errors here. We need * to purge this dquot anyway, so we go ahead regardless. */ - error = xfs_qm_dqflush(dqp, XFS_QMOPT_SYNC); + error = xfs_qm_dqflush(dqp, SYNC_WAIT); if (error) xfs_fs_cmn_err(CE_WARN, mp, "xfs_qm_dqpurge: dquot %p flush failed", dqp); diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c index dda0fb0..4e4ee9a 100644 --- a/fs/xfs/quota/xfs_dquot_item.c +++ b/fs/xfs/quota/xfs_dquot_item.c @@ -153,7 +153,7 @@ xfs_qm_dquot_logitem_push( * lock without sleeping, then there must not have been * anyone in the process of flushing the dquot. */ - error = xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI); + error = xfs_qm_dqflush(dqp, 0); if (error) xfs_fs_cmn_err(CE_WARN, dqp->q_mount, "xfs_qm_dquot_logitem_push: push error %d on dqp %p", diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c index 11cfd82..8699e51 100644 --- a/fs/xfs/quota/xfs_qm.c +++ b/fs/xfs/quota/xfs_qm.c @@ -450,7 +450,7 @@ xfs_qm_unmount_quotas( STATIC int xfs_qm_dqflush_all( xfs_mount_t *mp, - int flags) + int sync_mode) { int recl; xfs_dquot_t *dqp; @@ -486,7 +486,7 @@ again: * across a disk write. */ xfs_qm_mplist_unlock(mp); - error = xfs_qm_dqflush(dqp, flags); + error = xfs_qm_dqflush(dqp, sync_mode); xfs_dqunlock(dqp); if (error) return error; @@ -926,13 +926,11 @@ xfs_qm_sync( { int recl, restarts; xfs_dquot_t *dqp; - uint flush_flags; int error; if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp)) return 0; - flush_flags = (flags & SYNC_WAIT) ? XFS_QMOPT_SYNC : XFS_QMOPT_DELWRI; restarts = 0; again: @@ -992,7 +990,7 @@ xfs_qm_sync( * across a disk write */ xfs_qm_mplist_unlock(mp); - error = xfs_qm_dqflush(dqp, flush_flags); + error = xfs_qm_dqflush(dqp, flags); xfs_dqunlock(dqp); if (error && XFS_FORCED_SHUTDOWN(mp)) return 0; /* Need to prevent umount failure */ @@ -1796,7 +1794,7 @@ xfs_qm_quotacheck( * successfully. */ if (!error) - error = xfs_qm_dqflush_all(mp, XFS_QMOPT_DELWRI); + error = xfs_qm_dqflush_all(mp, 0); /* * We can get this error if we couldn't do a dquot allocation inside @@ -2018,7 +2016,7 @@ xfs_qm_shake_freelist( * We flush it delayed write, so don't bother * releasing the mplock. */ - error = xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI); + error = xfs_qm_dqflush(dqp, 0); if (error) { xfs_fs_cmn_err(CE_WARN, dqp->q_mount, "xfs_qm_dqflush_all: dquot %p flush failed", dqp); @@ -2201,7 +2199,7 @@ xfs_qm_dqreclaim_one(void) * We flush it delayed write, so don't bother * releasing the freelist lock. */ - error = xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI); + error = xfs_qm_dqflush(dqp, 0); if (error) { xfs_fs_cmn_err(CE_WARN, dqp->q_mount, "xfs_qm_dqreclaim: dquot %p flush failed", dqp); diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h index 21d11d9..fdcab3f 100644 --- a/fs/xfs/xfs_quota.h +++ b/fs/xfs/xfs_quota.h @@ -223,15 +223,9 @@ typedef struct xfs_qoff_logformat { #define XFS_QMOPT_RES_INOS 0x0800000 /* - * flags for dqflush and dqflush_all. - */ -#define XFS_QMOPT_SYNC 0x1000000 -#define XFS_QMOPT_DELWRI 0x4000000 - -/* * flags for dqalloc. */ -#define XFS_QMOPT_INHERIT 0x8000000 +#define XFS_QMOPT_INHERIT 0x1000000 /* * flags to xfs_trans_mod_dquot. -- cgit v0.10.2 From e8b217e7530c6a073ac69f1c85b922d93fdf5647 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 2 Feb 2010 10:16:26 +1100 Subject: xfs: remove invalid barrier optimization from xfs_fsync We always need to flush the disk write cache and can't skip it just because the no inode attributes have changed. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index fd108b7..43241e2 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -597,7 +597,7 @@ xfs_fsync( { xfs_trans_t *tp; int error = 0; - int log_flushed = 0, changed = 1; + int log_flushed = 0; xfs_itrace_entry(ip); @@ -627,18 +627,10 @@ xfs_fsync( * disk yet, the inode will be still be pinned. If it is, * force the log. */ - xfs_iunlock(ip, XFS_ILOCK_SHARED); - if (xfs_ipincount(ip)) { error = _xfs_log_force(ip->i_mount, XFS_LOG_SYNC, &log_flushed); - } else { - /* - * If the inode is not pinned and nothing has changed - * we don't need to flush the cache. - */ - changed = 0; } } else { /* @@ -673,7 +665,7 @@ xfs_fsync( xfs_iunlock(ip, XFS_ILOCK_EXCL); } - if ((ip->i_mount->m_flags & XFS_MOUNT_BARRIER) && changed) { + if (ip->i_mount->m_flags & XFS_MOUNT_BARRIER) { /* * If the log write didn't issue an ordered tag we need * to flush the disk cache for the data device now. -- cgit v0.10.2 From 61ef2489dbf587258526cfd4ebf4bba3b079f401 Mon Sep 17 00:00:00 2001 From: Wu Fengguang Date: Fri, 22 Jan 2010 16:16:19 +0800 Subject: resources: introduce generic page_is_ram() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit It's based on walk_system_ram_range(), for archs that don't have their own page_is_ram(). The static verions in MIPS and SCORE are also made global. v4: prefer plain 1 instead of PAGE_IS_RAM (H. Peter Anvin) v3: add comment (KAMEZAWA Hiroyuki) "AFAIK, this "System RAM" information has been used for kdump to grab valid memory area and seems good for the kernel itself." v2: add PAGE_IS_RAM macro (Américo Wang) Cc: Chen Liqin Cc: Lennox Wu Cc: Américo Wang Cc: linux-mips@linux-mips.org Cc: Yinghai Lu Acked-by: Ralf Baechle Reviewed-by: KAMEZAWA Hiroyuki Signed-off-by: Wu Fengguang LKML-Reference: <20100122081619.GA6431@localhost> Cc: Andrew Morton Signed-off-by: H. Peter Anvin diff --git a/arch/mips/mm/init.c b/arch/mips/mm/init.c index 15aa190..4d72aab 100644 --- a/arch/mips/mm/init.c +++ b/arch/mips/mm/init.c @@ -294,7 +294,7 @@ void __init fixrange_init(unsigned long start, unsigned long end, } #ifndef CONFIG_NEED_MULTIPLE_NODES -static int __init page_is_ram(unsigned long pagenr) +int page_is_ram(unsigned long pagenr) { int i; diff --git a/arch/score/mm/init.c b/arch/score/mm/init.c index 4e3dcd0..f684a59 100644 --- a/arch/score/mm/init.c +++ b/arch/score/mm/init.c @@ -59,7 +59,7 @@ static unsigned long setup_zero_page(void) } #ifndef CONFIG_NEED_MULTIPLE_NODES -static int __init page_is_ram(unsigned long pagenr) +int page_is_ram(unsigned long pagenr) { if (pagenr >= min_low_pfn && pagenr < max_low_pfn) return 1; diff --git a/include/linux/ioport.h b/include/linux/ioport.h index 83aa812..11ef795 100644 --- a/include/linux/ioport.h +++ b/include/linux/ioport.h @@ -188,5 +188,7 @@ extern int walk_system_ram_range(unsigned long start_pfn, unsigned long nr_pages, void *arg, int (*func)(unsigned long, unsigned long, void *)); +extern int page_is_ram(unsigned long pfn); + #endif /* __ASSEMBLY__ */ #endif /* _LINUX_IOPORT_H */ diff --git a/kernel/resource.c b/kernel/resource.c index fb11a58..b4d637a 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -297,6 +297,19 @@ int walk_system_ram_range(unsigned long start_pfn, unsigned long nr_pages, #endif +static int __is_ram(unsigned long pfn, unsigned long nr_pages, void *arg) +{ + return 1; +} +/* + * This generic page_is_ram() returns true if specified address is + * registered as "System RAM" in iomem_resource list. + */ +int __attribute__((weak)) page_is_ram(unsigned long pfn) +{ + return walk_system_ram_range(pfn, 1, NULL, __is_ram) == 1; +} + /* * Find empty slot in the resource tree given range and alignment. */ -- cgit v0.10.2 From e52730071567ec5b6f57e21d6693b112e01e1d0e Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Tue, 26 Jan 2010 16:31:19 -0800 Subject: Generic page_is_ram: use __weak Use __weak instead of __attribute__((weak)). Cc: Wu Fengguang Signed-off-by: Andrew Morton Signed-off-by: H. Peter Anvin diff --git a/kernel/resource.c b/kernel/resource.c index b4d637a..e68cd74 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -305,7 +305,7 @@ static int __is_ram(unsigned long pfn, unsigned long nr_pages, void *arg) * This generic page_is_ram() returns true if specified address is * registered as "System RAM" in iomem_resource list. */ -int __attribute__((weak)) page_is_ram(unsigned long pfn) +int __weak page_is_ram(unsigned long pfn) { return walk_system_ram_range(pfn, 1, NULL, __is_ram) == 1; } -- cgit v0.10.2 From 53df8fdc15fb646b0219e43c989c2cdab1ab100c Mon Sep 17 00:00:00 2001 From: Wu Fengguang Date: Wed, 27 Jan 2010 11:06:39 +0800 Subject: Move page_is_ram() declaration to mm.h Move page_is_ram() declaration to mm.h, it makes no sense in . Signed-off-by: Wu Fengguang LKML-Reference: <20100127030639.GD8132@localhost> Signed-off-by: H. Peter Anvin diff --git a/include/linux/ioport.h b/include/linux/ioport.h index 11ef795..83aa812 100644 --- a/include/linux/ioport.h +++ b/include/linux/ioport.h @@ -188,7 +188,5 @@ extern int walk_system_ram_range(unsigned long start_pfn, unsigned long nr_pages, void *arg, int (*func)(unsigned long, unsigned long, void *)); -extern int page_is_ram(unsigned long pfn); - #endif /* __ASSEMBLY__ */ #endif /* _LINUX_IOPORT_H */ diff --git a/include/linux/mm.h b/include/linux/mm.h index 24c3956..bad433f 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -265,6 +265,8 @@ static inline int get_page_unless_zero(struct page *page) return atomic_inc_not_zero(&page->_count); } +extern int page_is_ram(unsigned long pfn); + /* Support for virtually mapped pages */ struct page *vmalloc_to_page(const void *addr); unsigned long vmalloc_to_pfn(const void *addr); -- cgit v0.10.2 From 1b5576e69a5fe168c08a159685ac366316ac9bbc Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 22 Jan 2010 11:21:04 +0800 Subject: x86: Remove BIOS data range from e820 In preparation for moving to the generic page_is_ram(), make explicit what we expect to be reserved and not reserved. Tested-by: Wu Fengguang Signed-off-by: Yinghai Lu LKML-Reference: <20100122033004.335813103@intel.com> Cc: Andrew Morton Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c index d17d482..230687b 100644 --- a/arch/x86/kernel/e820.c +++ b/arch/x86/kernel/e820.c @@ -517,11 +517,19 @@ u64 __init e820_remove_range(u64 start, u64 size, unsigned old_type, int checktype) { int i; + u64 end; u64 real_removed_size = 0; if (size > (ULLONG_MAX - start)) size = ULLONG_MAX - start; + end = start + size; + printk(KERN_DEBUG "e820 remove range: %016Lx - %016Lx ", + (unsigned long long) start, + (unsigned long long) end); + e820_print_type(old_type); + printk(KERN_CONT "\n"); + for (i = 0; i < e820.nr_map; i++) { struct e820entry *ei = &e820.map[i]; u64 final_start, final_end; diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index cdb6a8a..f9b1f4e 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -650,6 +650,23 @@ static struct dmi_system_id __initdata bad_bios_dmi_table[] = { {} }; +static void __init trim_bios_range(void) +{ + /* + * A special case is the first 4Kb of memory; + * This is a BIOS owned area, not kernel ram, but generally + * not listed as such in the E820 table. + */ + e820_update_range(0, PAGE_SIZE, E820_RAM, E820_RESERVED); + /* + * special case: Some BIOSen report the PC BIOS + * area (640->1Mb) as ram even though it is not. + * take them out. + */ + e820_remove_range(BIOS_BEGIN, BIOS_END - BIOS_BEGIN, E820_RAM, 1); + sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map); +} + /* * Determine if we were loaded by an EFI loader. If so, then we have also been * passed the efi memmap, systab, etc., so we should use these data structures @@ -813,7 +830,7 @@ void __init setup_arch(char **cmdline_p) insert_resource(&iomem_resource, &data_resource); insert_resource(&iomem_resource, &bss_resource); - + trim_bios_range(); #ifdef CONFIG_X86_32 if (ppro_with_ram_bug()) { e820_update_range(0x70000000ULL, 0x40000ULL, E820_RAM, diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c index 334e63c..30e068d 100644 --- a/arch/x86/mm/ioremap.c +++ b/arch/x86/mm/ioremap.c @@ -29,22 +29,6 @@ int page_is_ram(unsigned long pagenr) resource_size_t addr, end; int i; - /* - * A special case is the first 4Kb of memory; - * This is a BIOS owned area, not kernel ram, but generally - * not listed as such in the E820 table. - */ - if (pagenr == 0) - return 0; - - /* - * Second special case: Some BIOSen report the PC BIOS - * area (640->1Mb) as ram even though it is not. - */ - if (pagenr >= (BIOS_BEGIN >> PAGE_SHIFT) && - pagenr < (BIOS_END >> PAGE_SHIFT)) - return 0; - for (i = 0; i < e820.nr_map; i++) { /* * Not usable memory: -- cgit v0.10.2 From 13ca0fcaa33f6b1984c4111b6ec5df42689fea6f Mon Sep 17 00:00:00 2001 From: Wu Fengguang Date: Fri, 22 Jan 2010 11:21:05 +0800 Subject: x86: Use the generic page_is_ram() The generic resource based page_is_ram() works better with memory hotplug/hotremove. So switch the x86 e820map based code to it. CC: Andi Kleen CC: KAMEZAWA Hiroyuki CC: Yinghai Lu Signed-off-by: Wu Fengguang LKML-Reference: <20100122033004.470767217@intel.com> Cc: Andrew Morton Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/page_types.h b/arch/x86/include/asm/page_types.h index 642fe34..a667f24 100644 --- a/arch/x86/include/asm/page_types.h +++ b/arch/x86/include/asm/page_types.h @@ -40,7 +40,6 @@ #ifndef __ASSEMBLY__ -extern int page_is_ram(unsigned long pagenr); extern int devmem_is_allowed(unsigned long pagenr); extern unsigned long max_low_pfn_mapped; diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c index 30e068d..1bf9e08 100644 --- a/arch/x86/mm/ioremap.c +++ b/arch/x86/mm/ioremap.c @@ -24,27 +24,6 @@ #include "physaddr.h" -int page_is_ram(unsigned long pagenr) -{ - resource_size_t addr, end; - int i; - - for (i = 0; i < e820.nr_map; i++) { - /* - * Not usable memory: - */ - if (e820.map[i].type != E820_RAM) - continue; - addr = (e820.map[i].addr + PAGE_SIZE-1) >> PAGE_SHIFT; - end = (e820.map[i].addr + e820.map[i].size) >> PAGE_SHIFT; - - - if ((pagenr >= addr) && (pagenr < end)) - return 1; - } - return 0; -} - /* * Fix up the linear direct mapping of the kernel to avoid cache attribute * conflicts. -- cgit v0.10.2 From b79c7adf82e8b8a6d6ad1dadf7e687a4a030cf8c Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 2 Feb 2010 13:01:25 +0900 Subject: mtd: trivial sh_flctl changes This patch contains a few changes for the sh_flctl driver: - not sh7723-only driver - get rid of kconfig dependency - use dev_err() instead of printk() - use __devinit and __devexit for probe()/remove() - fix probe() return values Signed-off-by: Magnus Damm Acked-by: Yoshihiro Shimoda Signed-off-by: Paul Mundt diff --git a/drivers/mtd/nand/Kconfig b/drivers/mtd/nand/Kconfig index 677cd53..bb64656 100644 --- a/drivers/mtd/nand/Kconfig +++ b/drivers/mtd/nand/Kconfig @@ -457,10 +457,10 @@ config MTD_NAND_NOMADIK config MTD_NAND_SH_FLCTL tristate "Support for NAND on Renesas SuperH FLCTL" - depends on MTD_NAND && SUPERH && CPU_SUBTYPE_SH7723 + depends on MTD_NAND && SUPERH help Several Renesas SuperH CPU has FLCTL. This option enables support - for NAND Flash using FLCTL. This driver support SH7723. + for NAND Flash using FLCTL. config MTD_NAND_DAVINCI tristate "Support NAND on DaVinci SoC" diff --git a/drivers/mtd/nand/sh_flctl.c b/drivers/mtd/nand/sh_flctl.c index 02bef21..ab068a5 100644 --- a/drivers/mtd/nand/sh_flctl.c +++ b/drivers/mtd/nand/sh_flctl.c @@ -1,10 +1,10 @@ /* * SuperH FLCTL nand controller * - * Copyright © 2008 Renesas Solutions Corp. - * Copyright © 2008 Atom Create Engineering Co., Ltd. + * Copyright (c) 2008 Renesas Solutions Corp. + * Copyright (c) 2008 Atom Create Engineering Co., Ltd. * - * Based on fsl_elbc_nand.c, Copyright © 2006-2007 Freescale Semiconductor + * Based on fsl_elbc_nand.c, Copyright (c) 2006-2007 Freescale Semiconductor * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by @@ -75,6 +75,11 @@ static void start_translation(struct sh_flctl *flctl) writeb(TRSTRT, FLTRCR(flctl)); } +static void timeout_error(struct sh_flctl *flctl, const char *str) +{ + dev_err(&flctl->pdev->dev, "Timeout occured in %s\n", str); +} + static void wait_completion(struct sh_flctl *flctl) { uint32_t timeout = LOOP_TIMEOUT_MAX; @@ -87,7 +92,7 @@ static void wait_completion(struct sh_flctl *flctl) udelay(1); } - printk(KERN_ERR "wait_completion(): Timeout occured \n"); + timeout_error(flctl, __func__); writeb(0x0, FLTRCR(flctl)); } @@ -132,7 +137,7 @@ static void wait_rfifo_ready(struct sh_flctl *flctl) return; udelay(1); } - printk(KERN_ERR "wait_rfifo_ready(): Timeout occured \n"); + timeout_error(flctl, __func__); } static void wait_wfifo_ready(struct sh_flctl *flctl) @@ -146,7 +151,7 @@ static void wait_wfifo_ready(struct sh_flctl *flctl) return; udelay(1); } - printk(KERN_ERR "wait_wfifo_ready(): Timeout occured \n"); + timeout_error(flctl, __func__); } static int wait_recfifo_ready(struct sh_flctl *flctl, int sector_number) @@ -198,7 +203,7 @@ static int wait_recfifo_ready(struct sh_flctl *flctl, int sector_number) writel(0, FL4ECCCR(flctl)); } - printk(KERN_ERR "wait_recfifo_ready(): Timeout occured \n"); + timeout_error(flctl, __func__); return 1; /* timeout */ } @@ -214,7 +219,7 @@ static void wait_wecfifo_ready(struct sh_flctl *flctl) return; udelay(1); } - printk(KERN_ERR "wait_wecfifo_ready(): Timeout occured \n"); + timeout_error(flctl, __func__); } static void read_datareg(struct sh_flctl *flctl, int offset) @@ -769,38 +774,36 @@ static int flctl_chip_init_tail(struct mtd_info *mtd) return 0; } -static int __init flctl_probe(struct platform_device *pdev) +static int __devinit flctl_probe(struct platform_device *pdev) { struct resource *res; struct sh_flctl *flctl; struct mtd_info *flctl_mtd; struct nand_chip *nand; struct sh_flctl_platform_data *pdata; - int ret; + int ret = -ENXIO; pdata = pdev->dev.platform_data; if (pdata == NULL) { - printk(KERN_ERR "sh_flctl platform_data not found.\n"); - return -ENODEV; + dev_err(&pdev->dev, "no platform data defined\n"); + return -EINVAL; } flctl = kzalloc(sizeof(struct sh_flctl), GFP_KERNEL); if (!flctl) { - printk(KERN_ERR "Unable to allocate NAND MTD dev structure.\n"); + dev_err(&pdev->dev, "failed to allocate driver data\n"); return -ENOMEM; } res = platform_get_resource(pdev, IORESOURCE_MEM, 0); if (!res) { - printk(KERN_ERR "%s: resource not found.\n", __func__); - ret = -ENODEV; + dev_err(&pdev->dev, "failed to get I/O memory\n"); goto err; } - flctl->reg = ioremap(res->start, res->end - res->start + 1); + flctl->reg = ioremap(res->start, resource_size(res)); if (flctl->reg == NULL) { - printk(KERN_ERR "%s: ioremap error.\n", __func__); - ret = -ENOMEM; + dev_err(&pdev->dev, "failed to remap I/O memory\n"); goto err; } @@ -808,6 +811,7 @@ static int __init flctl_probe(struct platform_device *pdev) flctl_mtd = &flctl->mtd; nand = &flctl->chip; flctl_mtd->priv = nand; + flctl->pdev = pdev; flctl->hwecc = pdata->has_hwecc; flctl_register_init(flctl, pdata->flcmncr_val); @@ -846,7 +850,7 @@ err: return ret; } -static int __exit flctl_remove(struct platform_device *pdev) +static int __devexit flctl_remove(struct platform_device *pdev) { struct sh_flctl *flctl = platform_get_drvdata(pdev); diff --git a/include/linux/mtd/sh_flctl.h b/include/linux/mtd/sh_flctl.h index e77c1ce..164c9d4 100644 --- a/include/linux/mtd/sh_flctl.h +++ b/include/linux/mtd/sh_flctl.h @@ -96,6 +96,7 @@ struct sh_flctl { struct mtd_info mtd; struct nand_chip chip; + struct platform_device *pdev; void __iomem *reg; uint8_t done_buff[2048 + 64]; /* max size 2048 + 64 */ -- cgit v0.10.2 From 010ab820582d03bcd3648416b5837107e8a9c5f3 Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Wed, 27 Jan 2010 09:17:21 +0000 Subject: mtd: sh_flctl SHBUSSEL and SEL_16BIT support This patch extends the sh_flctl driver with support for 16-bit bus configuration using SEL_16BIT and support for multiplexed pins using SHBUSSEL. Signed-off-by: Magnus Damm Acked-by: Yoshihiro Shimoda Signed-off-by: Paul Mundt diff --git a/drivers/mtd/nand/sh_flctl.c b/drivers/mtd/nand/sh_flctl.c index ab068a5..1842df8 100644 --- a/drivers/mtd/nand/sh_flctl.c +++ b/drivers/mtd/nand/sh_flctl.c @@ -105,6 +105,8 @@ static void set_addr(struct mtd_info *mtd, int column, int page_addr) addr = page_addr; /* ERASE1 */ } else if (page_addr != -1) { /* SEQIN, READ0, etc.. */ + if (flctl->chip.options & NAND_BUSWIDTH_16) + column >>= 1; if (flctl->page_size) { addr = column & 0x0FFF; addr |= (page_addr & 0xff) << 16; @@ -280,7 +282,7 @@ static void write_fiforeg(struct sh_flctl *flctl, int rlen, int offset) static void set_cmd_regs(struct mtd_info *mtd, uint32_t cmd, uint32_t flcmcdr_val) { struct sh_flctl *flctl = mtd_to_flctl(mtd); - uint32_t flcmncr_val = readl(FLCMNCR(flctl)); + uint32_t flcmncr_val = readl(FLCMNCR(flctl)) & ~SEL_16BIT; uint32_t flcmdcr_val, addr_len_bytes = 0; /* Set SNAND bit if page size is 2048byte */ @@ -302,6 +304,8 @@ static void set_cmd_regs(struct mtd_info *mtd, uint32_t cmd, uint32_t flcmcdr_va case NAND_CMD_READOOB: addr_len_bytes = flctl->rw_ADRCNT; flcmdcr_val |= CDSRC_E; + if (flctl->chip.options & NAND_BUSWIDTH_16) + flcmncr_val |= SEL_16BIT; break; case NAND_CMD_SEQIN: /* This case is that cmd is READ0 or READ1 or READ00 */ @@ -310,6 +314,8 @@ static void set_cmd_regs(struct mtd_info *mtd, uint32_t cmd, uint32_t flcmcdr_va case NAND_CMD_PAGEPROG: addr_len_bytes = flctl->rw_ADRCNT; flcmdcr_val |= DOCMD2_E | CDSRC_E | SELRW; + if (flctl->chip.options & NAND_BUSWIDTH_16) + flcmncr_val |= SEL_16BIT; break; case NAND_CMD_READID: flcmncr_val &= ~SNAND_E; @@ -528,6 +534,8 @@ static void flctl_cmdfunc(struct mtd_info *mtd, unsigned int command, set_addr(mtd, 0, page_addr); flctl->read_bytes = mtd->writesize + mtd->oobsize; + if (flctl->chip.options & NAND_BUSWIDTH_16) + column >>= 1; flctl->index += column; goto read_normal_exit; @@ -691,6 +699,18 @@ static uint8_t flctl_read_byte(struct mtd_info *mtd) return data; } +static uint16_t flctl_read_word(struct mtd_info *mtd) +{ + struct sh_flctl *flctl = mtd_to_flctl(mtd); + int index = flctl->index; + uint16_t data; + uint16_t *buf = (uint16_t *)&flctl->done_buff[index]; + + data = *buf; + flctl->index += 2; + return data; +} + static void flctl_read_buf(struct mtd_info *mtd, uint8_t *buf, int len) { int i; @@ -829,6 +849,11 @@ static int __devinit flctl_probe(struct platform_device *pdev) nand->select_chip = flctl_select_chip; nand->cmdfunc = flctl_cmdfunc; + if (pdata->flcmncr_val & SEL_16BIT) { + nand->options |= NAND_BUSWIDTH_16; + nand->read_word = flctl_read_word; + } + ret = nand_scan_ident(flctl_mtd, 1); if (ret) goto err; diff --git a/include/linux/mtd/sh_flctl.h b/include/linux/mtd/sh_flctl.h index 164c9d4..ab77609 100644 --- a/include/linux/mtd/sh_flctl.h +++ b/include/linux/mtd/sh_flctl.h @@ -51,6 +51,8 @@ #define _4ECCCNTEN (0x1 << 24) #define _4ECCEN (0x1 << 23) #define _4ECCCORRECT (0x1 << 22) +#define SHBUSSEL (0x1 << 20) +#define SEL_16BIT (0x1 << 19) #define SNAND_E (0x1 << 18) /* SNAND (0=512 1=2048)*/ #define QTSEL_E (0x1 << 17) #define ENDIAN (0x1 << 16) /* 1 = little endian */ -- cgit v0.10.2 From b4f74767a04e175c028336e06507fcc05f5a8618 Mon Sep 17 00:00:00 2001 From: Guennadi Liakhovetski Date: Tue, 19 Jan 2010 07:31:13 +0000 Subject: sh: add high impedance mode management for SIUA pins on sh7722 This improves power management for the SIUA controller on sh7722. Similar patches might be desired for other SIU-enabled SH platforms. Signed-off-by: Guennadi Liakhovetski Acked-by: Magnus Damm Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/cpu/sh4a/pinmux-sh7722.c b/arch/sh/kernel/cpu/sh4a/pinmux-sh7722.c index cb9d07b..0688a75 100644 --- a/arch/sh/kernel/cpu/sh4a/pinmux-sh7722.c +++ b/arch/sh/kernel/cpu/sh4a/pinmux-sh7722.c @@ -278,6 +278,7 @@ enum { HIZA8_LCDC, HIZA8_HIZ, HIZA7_LCDC, HIZA7_HIZ, HIZA6_LCDC, HIZA6_HIZ, + HIZB4_SIUA, HIZB4_HIZ, HIZB1_VIO, HIZB1_HIZ, HIZB0_VIO, HIZB0_HIZ, HIZC15_IRQ7, HIZC15_HIZ, @@ -546,7 +547,7 @@ static pinmux_enum_t pinmux_data[] = { PINMUX_DATA(VIO_VD2_MARK, PSE3_VIO, MSELB9_VIO2, HIZB0_VIO, FOE_VIO_VD2), PINMUX_DATA(VIO_HD2_MARK, PSE3_VIO, MSELB9_VIO2, - HIZB1_VIO, HIZB1_VIO, FCE_VIO_HD2), + HIZB1_VIO, FCE_VIO_HD2), PINMUX_DATA(VIO_CLK2_MARK, PSE3_VIO, MSELB9_VIO2, HIZB1_VIO, FRB_VIO_CLK2), @@ -658,14 +659,14 @@ static pinmux_enum_t pinmux_data[] = { PINMUX_DATA(SDHICLK_MARK, SDHICLK), /* SIU - Port A */ - PINMUX_DATA(SIUAOLR_MARK, PSC13_SIUAOLR, SIUAOLR_SIOF1_SYNC), - PINMUX_DATA(SIUAOBT_MARK, PSC14_SIUAOBT, SIUAOBT_SIOF1_SCK), - PINMUX_DATA(SIUAISLD_MARK, PSC15_SIUAISLD, SIUAISLD_SIOF1_RXD), - PINMUX_DATA(SIUAILR_MARK, PSC11_SIUAILR, SIUAILR_SIOF1_SS2), - PINMUX_DATA(SIUAIBT_MARK, PSC12_SIUAIBT, SIUAIBT_SIOF1_SS1), - PINMUX_DATA(SIUAOSLD_MARK, PSB0_SIUAOSLD, SIUAOSLD_SIOF1_TXD), - PINMUX_DATA(SIUMCKA_MARK, PSE11_SIUMCKA_SIOF1_MCK, PSB1_SIUMCKA, PTK0), - PINMUX_DATA(SIUFCKA_MARK, PSE11_SIUFCKA, PTK0), + PINMUX_DATA(SIUAOLR_MARK, PSC13_SIUAOLR, HIZB4_SIUA, SIUAOLR_SIOF1_SYNC), + PINMUX_DATA(SIUAOBT_MARK, PSC14_SIUAOBT, HIZB4_SIUA, SIUAOBT_SIOF1_SCK), + PINMUX_DATA(SIUAISLD_MARK, PSC15_SIUAISLD, HIZB4_SIUA, SIUAISLD_SIOF1_RXD), + PINMUX_DATA(SIUAILR_MARK, PSC11_SIUAILR, HIZB4_SIUA, SIUAILR_SIOF1_SS2), + PINMUX_DATA(SIUAIBT_MARK, PSC12_SIUAIBT, HIZB4_SIUA, SIUAIBT_SIOF1_SS1), + PINMUX_DATA(SIUAOSLD_MARK, PSB0_SIUAOSLD, HIZB4_SIUA, SIUAOSLD_SIOF1_TXD), + PINMUX_DATA(SIUMCKA_MARK, PSE11_SIUMCKA_SIOF1_MCK, HIZB4_SIUA, PSB1_SIUMCKA, PTK0), + PINMUX_DATA(SIUFCKA_MARK, PSE11_SIUFCKA, HIZB4_SIUA, PTK0), /* SIU - Port B */ PINMUX_DATA(SIUBOLR_MARK, PSB11_SIUBOLR, SIOSTRB1_SIUBOLR), @@ -1612,7 +1613,7 @@ static struct pinmux_cfg_reg pinmux_config_regs[] = { 0, 0, 0, 0, 0, 0, - 0, 0, + HIZB4_SIUA, HIZB4_HIZ, 0, 0, 0, 0, HIZB1_VIO, HIZB1_HIZ, -- cgit v0.10.2 From 90fdbdb48442a03c72cae5463e6edb64cb3a3a7d Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Sun, 31 Jan 2010 20:53:24 +0900 Subject: sched: Use for_each_bit No change in functionality. Signed-off-by: Akinobu Mita Cc: Peter Zijlstra Cc: Andrew Morton LKML-Reference: <1264938810-4173-1-git-send-email-akinobu.mita@gmail.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched_cpupri.c b/kernel/sched_cpupri.c index 597b330..eeb3506 100644 --- a/kernel/sched_cpupri.c +++ b/kernel/sched_cpupri.c @@ -47,9 +47,7 @@ static int convert_prio(int prio) } #define for_each_cpupri_active(array, idx) \ - for (idx = find_first_bit(array, CPUPRI_NR_PRIORITIES); \ - idx < CPUPRI_NR_PRIORITIES; \ - idx = find_next_bit(array, CPUPRI_NR_PRIORITIES, idx+1)) + for_each_bit(idx, array, CPUPRI_NR_PRIORITIES) /** * cpupri_find - find the best (lowest-pri) CPU in the system -- cgit v0.10.2 From 4a461c85b643258e305eb5a3aced34009db2f818 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 1 Feb 2010 11:13:39 +0100 Subject: sched: Remove unused update_shares_locked() Commit f492e12ef050e02bf0185b6b57874992591b9be1 ("sched: Remove load_balance_newidle()") removed the only user of this function, so remove it too. Reported-by: Stephen Rothwell Signed-off-by: Peter Zijlstra LKML-Reference: <1265019219.24455.128.camel@laptop> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index a56ead4..2386f54 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1633,16 +1633,6 @@ static void update_shares(struct sched_domain *sd) } } -static void update_shares_locked(struct rq *rq, struct sched_domain *sd) -{ - if (root_task_group_empty()) - return; - - raw_spin_unlock(&rq->lock); - update_shares(sd); - raw_spin_lock(&rq->lock); -} - static void update_h_load(long cpu) { if (root_task_group_empty()) @@ -1657,10 +1647,6 @@ static inline void update_shares(struct sched_domain *sd) { } -static inline void update_shares_locked(struct rq *rq, struct sched_domain *sd) -{ -} - #endif #ifdef CONFIG_PREEMPT -- cgit v0.10.2 From 8a349d4b13c41c00564cd79f6fabdec347084758 Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Tue, 2 Feb 2010 07:22:13 +0000 Subject: spi/spi_s3c64xx.c: Fix continuation line formats String constants that are continued on subsequent lines with \ are not good. Signed-off-by: Joe Perches Signed-off-by: Grant Likely diff --git a/drivers/spi/spi_s3c64xx.c b/drivers/spi/spi_s3c64xx.c index 32db695..9736581 100644 --- a/drivers/spi/spi_s3c64xx.c +++ b/drivers/spi/spi_s3c64xx.c @@ -634,8 +634,8 @@ static void handle_msg(struct s3c64xx_spi_driver_data *sdd, S3C64XX_SPI_DEACT(sdd); if (status) { - dev_err(&spi->dev, "I/O Error: \ - rx-%d tx-%d res:rx-%c tx-%c len-%d\n", + dev_err(&spi->dev, "I/O Error: " + "rx-%d tx-%d res:rx-%c tx-%c len-%d\n", xfer->rx_buf ? 1 : 0, xfer->tx_buf ? 1 : 0, (sdd->state & RXBUSY) ? 'f' : 'p', (sdd->state & TXBUSY) ? 'f' : 'p', @@ -1039,11 +1039,10 @@ static int __init s3c64xx_spi_probe(struct platform_device *pdev) goto err8; } - dev_dbg(&pdev->dev, "Samsung SoC SPI Driver loaded for Bus SPI-%d \ - with %d Slaves attached\n", + dev_dbg(&pdev->dev, "Samsung SoC SPI Driver loaded for Bus SPI-%d " + "with %d Slaves attached\n", pdev->id, master->num_chipselect); - dev_dbg(&pdev->dev, "\tIOmem=[0x%x-0x%x]\ - \tDMA=[Rx-%d, Tx-%d]\n", + dev_dbg(&pdev->dev, "\tIOmem=[0x%x-0x%x]\tDMA=[Rx-%d, Tx-%d]\n", mem_res->end, mem_res->start, sdd->rx_dmach, sdd->tx_dmach); -- cgit v0.10.2 From e9867c569970d8afb4b882bafbbe81426bd46333 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 2 Feb 2010 17:35:13 +0900 Subject: sh: Provide create_irq_nr() for dynamic IRQ creation by number. This just reworks the existing create_irq_on_node() in to the new create_irq_nr() which is generally exposed. This permits boards that haven't converted over to sparseirq to try and use their existing ranges, rather than having arbitrary vectors assigned to them. Signed-off-by: Paul Mundt diff --git a/drivers/sh/intc.c b/drivers/sh/intc.c index d5d7f23..7d286ae 100644 --- a/drivers/sh/intc.c +++ b/drivers/sh/intc.c @@ -872,7 +872,7 @@ device_initcall(register_intc_sysdevs); /* * Dynamic IRQ allocation and deallocation */ -static unsigned int create_irq_on_node(unsigned int irq_want, int node) +unsigned int create_irq_nr(unsigned int irq_want, int node) { unsigned int irq = 0, new; unsigned long flags; @@ -881,24 +881,28 @@ static unsigned int create_irq_on_node(unsigned int irq_want, int node) spin_lock_irqsave(&vector_lock, flags); /* - * First try the wanted IRQ, then scan. + * First try the wanted IRQ */ - if (test_and_set_bit(irq_want, intc_irq_map)) { + if (test_and_set_bit(irq_want, intc_irq_map) == 0) { + new = irq_want; + } else { + /* .. then fall back to scanning. */ new = find_first_zero_bit(intc_irq_map, nr_irqs); if (unlikely(new == nr_irqs)) goto out_unlock; - desc = irq_to_desc_alloc_node(new, node); - if (unlikely(!desc)) { - pr_info("can't get irq_desc for %d\n", new); - goto out_unlock; - } - - desc = move_irq_desc(desc, node); __set_bit(new, intc_irq_map); - irq = new; } + desc = irq_to_desc_alloc_node(new, node); + if (unlikely(!desc)) { + pr_info("can't get irq_desc for %d\n", new); + goto out_unlock; + } + + desc = move_irq_desc(desc, node); + irq = new; + out_unlock: spin_unlock_irqrestore(&vector_lock, flags); @@ -913,7 +917,7 @@ int create_irq(void) int nid = cpu_to_node(smp_processor_id()); int irq; - irq = create_irq_on_node(NR_IRQS_LEGACY, nid); + irq = create_irq_nr(NR_IRQS_LEGACY, nid); if (irq == 0) irq = -1; -- cgit v0.10.2 From c7a967273a539a9a38413753c58125e777e2d30d Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 2 Feb 2010 17:38:04 +0900 Subject: sh: mach-se: Convert SH7724 solution engine FPGA to sparseirq. This uses the new create_irq_nr() to build up the FPGA's desired virtual IRQ mapping and permits us to finally flip on sparseirq for this board. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index bae5383..0f023bb 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -731,7 +731,7 @@ config GUSA_RB config SPARSE_IRQ def_bool y depends on SUPERH32 && !SH_DREAMCAST && !SH_HIGHLANDER && \ - !SH_RTS7751R2D && !HD64461 && !SH_7724_SOLUTION_ENGINE + !SH_RTS7751R2D && !HD64461 help This enables support for sparse irqs. This is useful in general as most CPUs have a fairly sparse array of IRQ vectors, which diff --git a/arch/sh/boards/mach-se/7724/irq.c b/arch/sh/boards/mach-se/7724/irq.c index e5e021a..0942be2 100644 --- a/arch/sh/boards/mach-se/7724/irq.c +++ b/arch/sh/boards/mach-se/7724/irq.c @@ -93,18 +93,15 @@ static void se7724_irq_demux(unsigned int irq, struct irq_desc *desc) { struct fpga_irq set = get_fpga_irq(irq); unsigned short intv = __raw_readw(set.sraddr); - struct irq_desc *ext_desc; unsigned int ext_irq = set.base; intv &= set.mask; - while (intv) { - if (intv & 0x0001) { - ext_desc = irq_desc + ext_irq; - handle_level_irq(ext_irq, ext_desc); - } - intv >>= 1; - ext_irq++; + for (; intv; intv >>= 1, ext_irq++) { + if (!(intv & 1)) + continue; + + generic_handle_irq(ext_irq); } } @@ -113,7 +110,7 @@ static void se7724_irq_demux(unsigned int irq, struct irq_desc *desc) */ void __init init_se7724_IRQ(void) { - int i; + int i, nid = cpu_to_node(boot_cpu_data); __raw_writew(0xffff, IRQ0_MR); /* mask all */ __raw_writew(0xffff, IRQ1_MR); /* mask all */ @@ -123,10 +120,29 @@ void __init init_se7724_IRQ(void) __raw_writew(0x0000, IRQ2_SR); /* clear irq */ __raw_writew(0x002a, IRQ_MODE); /* set irq type */ - for (i = 0; i < SE7724_FPGA_IRQ_NR; i++) - set_irq_chip_and_handler_name(SE7724_FPGA_IRQ_BASE + i, + for (i = 0; i < SE7724_FPGA_IRQ_NR; i++) { + int irq, wanted; + + wanted = SE7724_FPGA_IRQ_BASE + i; + + irq = create_irq_nr(wanted, nid); + if (unlikely(irq == 0)) { + pr_err("%s: failed hooking irq %d for FPGA\n", + __func__, wanted); + return; + } + + if (unlikely(irq != wanted)) { + pr_err("%s: got irq %d but wanted %d, bailing.\n", + __func__, irq, wanted); + destroy_irq(irq); + return; + } + + set_irq_chip_and_handler_name(irq, &se7724_irq_chip, handle_level_irq, "level"); + } set_irq_chained_handler(IRQ0_IRQ, se7724_irq_demux); set_irq_type(IRQ0_IRQ, IRQ_TYPE_LEVEL_LOW); -- cgit v0.10.2 From 6eb6f98396f7bd653d8fb15b06364c8c7d70e22c Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 2 Feb 2010 17:48:17 +0900 Subject: sh: hd6446x: Convert to sparseirq. Follows the se7724 change and converts the hd64461 IRQ handling to sparseirq. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 0f023bb..9e565d9 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -731,7 +731,7 @@ config GUSA_RB config SPARSE_IRQ def_bool y depends on SUPERH32 && !SH_DREAMCAST && !SH_HIGHLANDER && \ - !SH_RTS7751R2D && !HD64461 + !SH_RTS7751R2D help This enables support for sparse irqs. This is useful in general as most CPUs have a fairly sparse array of IRQ vectors, which diff --git a/arch/sh/cchips/hd6446x/hd64461.c b/arch/sh/cchips/hd6446x/hd64461.c index cbfedfb..b447091 100644 --- a/arch/sh/cchips/hd6446x/hd64461.c +++ b/arch/sh/cchips/hd6446x/hd64461.c @@ -73,7 +73,7 @@ static void hd64461_irq_demux(unsigned int irq, struct irq_desc *desc) int __init setup_hd64461(void) { - int i; + int i, nid = cpu_to_node(boot_cpu_data); if (!MACH_HD64461) return 0; @@ -90,9 +90,26 @@ int __init setup_hd64461(void) __raw_writew(0xffff, HD64461_NIMR); /* IRQ 80 -> 95 belongs to HD64461 */ - for (i = HD64461_IRQBASE; i < HD64461_IRQBASE + 16; i++) + for (i = HD64461_IRQBASE; i < HD64461_IRQBASE + 16; i++) { + unsigned int irq; + + irq = create_irq_nr(i, nid); + if (unlikely(irq == 0)) { + pr_err("%s: failed hooking irq %d for HD64461\n", + __func__, i); + return -EBUSY; + } + + if (unlikely(irq != i)) { + pr_err("%s: got irq %d but wanted %d, bailing.\n", + __func__, irq, i); + destroy_irq(irq); + return -EINVAL; + } + set_irq_chip_and_handler(i, &hd64461_irq_chip, handle_level_irq); + } set_irq_chained_handler(CONFIG_HD64461_IRQ, hd64461_irq_demux); set_irq_type(CONFIG_HD64461_IRQ, IRQ_TYPE_LEVEL_LOW); -- cgit v0.10.2 From deb9b22b8968fa0166d89c8ad1346e816cf1aec4 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 2 Feb 2010 18:01:55 +0900 Subject: sh: mach-dreamcast: Convert to sparseirq. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 9e565d9..3b3c036 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -730,8 +730,7 @@ config GUSA_RB config SPARSE_IRQ def_bool y - depends on SUPERH32 && !SH_DREAMCAST && !SH_HIGHLANDER && \ - !SH_RTS7751R2D + depends on SUPERH32 && !SH_HIGHLANDER && !SH_RTS7751R2D help This enables support for sparse irqs. This is useful in general as most CPUs have a fairly sparse array of IRQ vectors, which diff --git a/arch/sh/boards/mach-dreamcast/irq.c b/arch/sh/boards/mach-dreamcast/irq.c index f55fc8e..d932667 100644 --- a/arch/sh/boards/mach-dreamcast/irq.c +++ b/arch/sh/boards/mach-dreamcast/irq.c @@ -135,3 +135,30 @@ int systemasic_irq_demux(int irq) /* Not reached */ return irq; } + +void systemasic_irq_init(void) +{ + int i, nid = cpu_to_node(boot_cpu_data); + + /* Assign all virtual IRQs to the System ASIC int. handler */ + for (i = HW_EVENT_IRQ_BASE; i < HW_EVENT_IRQ_MAX; i++) { + unsigned int irq; + + irq = create_irq_nr(i, nid); + if (unlikely(irq == 0)) { + pr_err("%s: failed hooking irq %d for systemasic\n", + __func__, i); + return; + } + + if (unlikely(irq != i)) { + pr_err("%s: got irq %d but wanted %d, bailing.\n", + __func__, irq, i); + destroy_irq(irq); + return; + } + + set_irq_chip_and_handler(i, &systemasic_int, + handle_level_irq); + } +} diff --git a/arch/sh/boards/mach-dreamcast/setup.c b/arch/sh/boards/mach-dreamcast/setup.c index a4b7402..ad1a4db 100644 --- a/arch/sh/boards/mach-dreamcast/setup.c +++ b/arch/sh/boards/mach-dreamcast/setup.c @@ -28,25 +28,8 @@ #include #include -extern struct irq_chip systemasic_int; -extern void aica_time_init(void); -extern int systemasic_irq_demux(int); - static void __init dreamcast_setup(char **cmdline_p) { - int i; - - /* Mask all hardware events */ - /* XXX */ - - /* Acknowledge any previous events */ - /* XXX */ - - /* Assign all virtual IRQs to the System ASIC int. handler */ - for (i = HW_EVENT_IRQ_BASE; i < HW_EVENT_IRQ_MAX; i++) - set_irq_chip_and_handler(i, &systemasic_int, - handle_level_irq); - board_time_init = aica_time_init; } @@ -54,4 +37,5 @@ static struct sh_machine_vector mv_dreamcast __initmv = { .mv_name = "Sega Dreamcast", .mv_setup = dreamcast_setup, .mv_irq_demux = systemasic_irq_demux, + .mv_init_irq = systemasic_irq_init, }; diff --git a/arch/sh/include/mach-dreamcast/mach/sysasic.h b/arch/sh/include/mach-dreamcast/mach/sysasic.h index f334266..58f710e 100644 --- a/arch/sh/include/mach-dreamcast/mach/sysasic.h +++ b/arch/sh/include/mach-dreamcast/mach/sysasic.h @@ -39,5 +39,10 @@ #define HW_EVENT_IRQ_MAX (HW_EVENT_IRQ_BASE + 95) +/* arch/sh/boards/mach-dreamcast/irq.c */ +extern int systemasic_irq_demux(int); +extern void systemasic_irq_init(void); +extern void aica_time_init(void); + #endif /* __ASM_SH_DREAMCAST_SYSASIC_H */ -- cgit v0.10.2 From 4b842c8571240d1173eaf00e114cc5cc160f1722 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 2 Feb 2010 19:15:17 +0900 Subject: sh: enable sparseirq for highlander and r2d. highlander and r2d are the only remaining ones that were blocking sparseirq being turned on by default, but it turns out that they already work fine with it by virtue of register_intc_controller(). As such, we can kill off the dependencies and turn it on by default. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 3b3c036..7354e8d 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -87,6 +87,10 @@ config GENERIC_IRQ_PROBE config IRQ_PER_CPU def_bool y +config SPARSE_IRQ + def_bool y + depends on SUPERH32 + config GENERIC_GPIO def_bool n @@ -728,18 +732,6 @@ config GUSA_RB LLSC, this should be more efficient than the other alternative of disabling interrupts around the atomic sequence. -config SPARSE_IRQ - def_bool y - depends on SUPERH32 && !SH_HIGHLANDER && !SH_RTS7751R2D - help - This enables support for sparse irqs. This is useful in general - as most CPUs have a fairly sparse array of IRQ vectors, which - the irq_desc then maps directly on to. Systems with a high - number of off-chip IRQs will want to treat this as - experimental until they have been independently verified. - - If you don't know what to do here, say N. - endmenu menu "Boot options" -- cgit v0.10.2 From 37feecb0ae8c29b713b957c053b05ad10c42c73b Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Tue, 2 Feb 2010 19:17:42 +0900 Subject: sh: Disable generic IRQ probing. IRQ autoprobing hasn't actually worked for us at all since very early in 2.6, but no one seems to have noticed given that none of the drivers that use it see much testing. yenta_socket is the odd one out, and that depends on PCI IRQs which are fixed on all SH platforms anyways. Consequently, turning off autoprobing fixes up crashes triggered by yenta_socket and at least gets it working again on r7785rp. Signed-off-by: Paul Mundt diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 7354e8d..05cef50 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -81,9 +81,6 @@ config GENERIC_HARDIRQS config GENERIC_HARDIRQS_NO__DO_IRQ def_bool y -config GENERIC_IRQ_PROBE - def_bool y - config IRQ_PER_CPU def_bool y -- cgit v0.10.2 From 88f66ea98d7ae6a8b6a34e38b1b4fa51abc1c9ca Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 6 Jan 2010 18:49:44 -0500 Subject: kconfig: Look in both /bin and /sbin for lsmod in streamline_config.pl Distributions now have lsmod in /bin instead of /sbin. But to handle both cases, we look for it in /sbin /bin /usr/bin and /usr/sbin. If lsmod is not found in any of those paths, it defaults to use just lsmod and hopes that it lies in the path of the user. Tested-by: Xavier Chantry Signed-off-by: Steven Rostedt diff --git a/scripts/kconfig/streamline_config.pl b/scripts/kconfig/streamline_config.pl index 9e66fa8..d7f7db7 100644 --- a/scripts/kconfig/streamline_config.pl +++ b/scripts/kconfig/streamline_config.pl @@ -264,7 +264,20 @@ foreach my $makefile (@makefiles) { my %modules; # see what modules are loaded on this system -open(LIN,"/sbin/lsmod|") || die "Cant lsmod"; +my $lsmod; + +foreach $dir ( ("/sbin", "/bin", "/usr/sbin", "/usr/bin") ) { + if ( -x "$dir/lsmod" ) { + $lsmod = "$dir/lsmod"; + last; + } +} +if (!defined($lsmod)) { + # try just the path + $lsmod = "lsmod"; +} + +open(LIN,"$lsmod|") || die "Can not call lsmod with $lsmod"; while () { next if (/^Module/); # Skip the first line. if (/^(\S+)/) { -- cgit v0.10.2 From 7d39e849912f0c3c8c6fc94be7bf7d120b1ee0ba Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Tue, 2 Feb 2010 20:46:34 +0100 Subject: HID: update copyright Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index 7fe5098..1bbd96d 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -4,7 +4,7 @@ * Copyright (c) 1999 Andreas Gal * Copyright (c) 2000-2005 Vojtech Pavlik * Copyright (c) 2005 Michael Haboustak for Concept2, Inc - * Copyright (c) 2006-2007 Jiri Kosina + * Copyright (c) 2006-2010 Jiri Kosina */ /* diff --git a/drivers/hid/hid-input.c b/drivers/hid/hid-input.c index dad7aae..8430d62 100644 --- a/drivers/hid/hid-input.c +++ b/drivers/hid/hid-input.c @@ -1,6 +1,6 @@ /* * Copyright (c) 2000-2001 Vojtech Pavlik - * Copyright (c) 2006-2007 Jiri Kosina + * Copyright (c) 2006-2010 Jiri Kosina * * HID to Linux Input mapping */ diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c index e72751d..eb7e001 100644 --- a/drivers/hid/usbhid/hid-core.c +++ b/drivers/hid/usbhid/hid-core.c @@ -5,7 +5,7 @@ * Copyright (c) 2000-2005 Vojtech Pavlik * Copyright (c) 2005 Michael Haboustak for Concept2, Inc * Copyright (c) 2007-2008 Oliver Neukum - * Copyright (c) 2006-2009 Jiri Kosina + * Copyright (c) 2006-2010 Jiri Kosina */ /* -- cgit v0.10.2 From 1c3a02c215a5b955b342f29dc1719e1a5771eaf1 Mon Sep 17 00:00:00 2001 From: Alexander Shishkin Date: Tue, 2 Feb 2010 18:43:32 +0200 Subject: HID: add NOGET quirk for Prodige Cordless Combo I happen to own a keyboard identified as 05af:3062 which is labeled as "FlatX Coldless Combo" by "Prodige", which exhibits input problems without NOGET quirk. For some reason, lsusb reports this device as "Jing-Mold Enterprise Co., Ltd", which is not mentioned anywhere on the package. A quick search on the intenet shows that there a other people who have this in their lsusb output, but apparently they don't have the problem I am seeing (or they are not such furious typists as myself). Signed-off-by: Alexander Shishkin Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index 6c38359..4e5adc3 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -383,6 +383,9 @@ #define USB_VENDOR_ID_POWERCOM 0x0d9f #define USB_DEVICE_ID_POWERCOM_UPS 0x0002 +#define USB_VENDOR_ID_PRODIGE 0x05af +#define USB_DEVICE_ID_PRODIGE_CORDLESS 0x3062 + #define USB_VENDOR_ID_SAITEK 0x06a3 #define USB_DEVICE_ID_SAITEK_RUMBLEPAD 0xff17 diff --git a/drivers/hid/usbhid/hid-quirks.c b/drivers/hid/usbhid/hid-quirks.c index 38773dc..fc074c1 100644 --- a/drivers/hid/usbhid/hid-quirks.c +++ b/drivers/hid/usbhid/hid-quirks.c @@ -45,6 +45,7 @@ static const struct hid_blacklist { { USB_VENDOR_ID_PANTHERLORD, USB_DEVICE_ID_PANTHERLORD_TWIN_USB_JOYSTICK, HID_QUIRK_MULTI_INPUT | HID_QUIRK_SKIP_OUTPUT_REPORTS }, { USB_VENDOR_ID_PLAYDOTCOM, USB_DEVICE_ID_PLAYDOTCOM_EMS_USBII, HID_QUIRK_MULTI_INPUT }, + { USB_VENDOR_ID_PRODIGE, USB_DEVICE_ID_PRODIGE_CORDLESS, HID_QUIRK_NOGET }, { USB_VENDOR_ID_ATEN, USB_DEVICE_ID_ATEN_UC100KM, HID_QUIRK_NOGET }, { USB_VENDOR_ID_ATEN, USB_DEVICE_ID_ATEN_CS124U, HID_QUIRK_NOGET }, -- cgit v0.10.2 From 8127f4e883666c9960cfa89cffd36313748f8bab Mon Sep 17 00:00:00 2001 From: Thadeu Lima de Souza Cascardo Date: Tue, 2 Feb 2010 18:09:06 -0200 Subject: HID: use multi input quirk for eTurboTouch touchscreen This device generates ABS_Z and ABS_RX events, while it should be generating ABS_X and ABS_Y instead. Using the MULTI_INPUT quirk solves this issue. Signed-off-by: Thadeu Lima de Souza Cascardo Signed-off-by: Daniel Oliveira Nascimento Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index 4e5adc3..bb5f4fa 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -169,6 +169,9 @@ #define USB_VENDOR_ID_ESSENTIAL_REALITY 0x0d7f #define USB_DEVICE_ID_ESSENTIAL_REALITY_P5 0x0100 +#define USB_VENDOR_ID_ETURBOTOUCH 0x22b9 +#define USB_DEVICE_ID_ETURBOTOUCH 0x0006 + #define USB_VENDOR_ID_EZKEY 0x0518 #define USB_DEVICE_ID_BTC_8193 0x0002 diff --git a/drivers/hid/usbhid/hid-quirks.c b/drivers/hid/usbhid/hid-quirks.c index fc074c1..88a1c69 100644 --- a/drivers/hid/usbhid/hid-quirks.c +++ b/drivers/hid/usbhid/hid-quirks.c @@ -43,6 +43,7 @@ static const struct hid_blacklist { { USB_VENDOR_ID_AFATECH, USB_DEVICE_ID_AFATECH_AF9016, HID_QUIRK_FULLSPEED_INTERVAL }, + { USB_VENDOR_ID_ETURBOTOUCH, USB_DEVICE_ID_ETURBOTOUCH, HID_QUIRK_MULTI_INPUT }, { USB_VENDOR_ID_PANTHERLORD, USB_DEVICE_ID_PANTHERLORD_TWIN_USB_JOYSTICK, HID_QUIRK_MULTI_INPUT | HID_QUIRK_SKIP_OUTPUT_REPORTS }, { USB_VENDOR_ID_PLAYDOTCOM, USB_DEVICE_ID_PLAYDOTCOM_EMS_USBII, HID_QUIRK_MULTI_INPUT }, { USB_VENDOR_ID_PRODIGE, USB_DEVICE_ID_PRODIGE_CORDLESS, HID_QUIRK_NOGET }, -- cgit v0.10.2 From 615f0833aa4c4aa944ceb78895bbffa8bd1884df Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 2 Feb 2010 21:51:27 -0500 Subject: kconfig: Add LSMOD=file to override the lsmod for localmodconfig Doing the following: make LSMOD=file localmodconfig Will make the streamline-config code use the given file instead of lsmod. If the file is an executable, it will execute it, otherwise it will read it as text. make LSMOD=/my/local/path/lsmod localmodconfig The above will execute the lsmod in /my/local/path instead of the lsmods that may be located elsewhere. make LSMOD=embedded_board_lsmod localmodconfig The above will read the "embedded_board_lsmod" as a text file. This is useful if you are doing a cross compile and need to run the config against modules that exist on an embedded device. Note, if the LSMOD= file does is not a path, it will add the path to the object directory. That is, the above example will look for "embedded_board_lsmod" in the directory that the binary will be built in (the O=dir directory). Signed-off-by: Steven Rostedt On branch config/linus diff --git a/scripts/kconfig/Makefile b/scripts/kconfig/Makefile index 006c96f..85b9065 100644 --- a/scripts/kconfig/Makefile +++ b/scripts/kconfig/Makefile @@ -30,9 +30,18 @@ silentoldconfig: $(obj)/conf $(Q)mkdir -p include/generated $< -s $(Kconfig) +# if no path is given, then use src directory to find file +ifdef LSMOD +LSMOD_F = $(shell if [ `basename $(LSMOD)` == $(LSMOD) ]; then \ + echo $(objtree)/$(LSMOD); \ + else \ + echo $(LSMOD); \ + fi) +endif + localmodconfig: $(obj)/streamline_config.pl $(obj)/conf $(Q)mkdir -p include/generated - $(Q)perl $< $(srctree) $(Kconfig) > .tmp.config + $(Q)perl $< $(srctree) $(Kconfig) $(LSMOD_F) > .tmp.config $(Q)if [ -f .config ]; then \ cmp -s .tmp.config .config || \ (mv -f .config .config.old.1; \ @@ -47,7 +56,7 @@ localmodconfig: $(obj)/streamline_config.pl $(obj)/conf localyesconfig: $(obj)/streamline_config.pl $(obj)/conf $(Q)mkdir -p include/generated - $(Q)perl $< $(srctree) $(Kconfig) > .tmp.config + $(Q)perl $< $(srctree) $(Kconfig) $(LSMOD_F) > .tmp.config $(Q)sed -i s/=m/=y/ .tmp.config $(Q)if [ -f .config ]; then \ cmp -s .tmp.config .config || \ diff --git a/scripts/kconfig/streamline_config.pl b/scripts/kconfig/streamline_config.pl index d7f7db7..afbd54a 100644 --- a/scripts/kconfig/streamline_config.pl +++ b/scripts/kconfig/streamline_config.pl @@ -113,6 +113,7 @@ find_config; # Get the build source and top level Kconfig file (passed in) my $ksource = $ARGV[0]; my $kconfig = $ARGV[1]; +my $lsmod_file = $ARGV[2]; my @makefiles = `find $ksource -name Makefile`; my %depends; @@ -263,21 +264,36 @@ foreach my $makefile (@makefiles) { my %modules; -# see what modules are loaded on this system -my $lsmod; - -foreach $dir ( ("/sbin", "/bin", "/usr/sbin", "/usr/bin") ) { - if ( -x "$dir/lsmod" ) { - $lsmod = "$dir/lsmod"; - last; +if (defined($lsmod_file)) { + if ( ! -f $lsmod_file) { + die "$lsmod_file not found"; + } + if ( -x $lsmod_file) { + # the file is executable, run it + open(LIN, "$lsmod_file|"); + } else { + # Just read the contents + open(LIN, "$lsmod_file"); } +} else { + + # see what modules are loaded on this system + my $lsmod; + + foreach $dir ( ("/sbin", "/bin", "/usr/sbin", "/usr/bin") ) { + if ( -x "$dir/lsmod" ) { + $lsmod = "$dir/lsmod"; + last; + } } -if (!defined($lsmod)) { - # try just the path - $lsmod = "lsmod"; + if (!defined($lsmod)) { + # try just the path + $lsmod = "lsmod"; + } + + open(LIN,"$lsmod|") || die "Can not call lsmod with $lsmod"; } -open(LIN,"$lsmod|") || die "Can not call lsmod with $lsmod"; while () { next if (/^Module/); # Skip the first line. if (/^(\S+)/) { -- cgit v0.10.2 From dc942cee2fcb734d6f3ef7966040e4e034b67d5b Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Tue, 22 Dec 2009 21:22:59 +0000 Subject: powerpc/viodasd: Remove VIOD_KERN_ macros for printks Use #define pr_fmt(fmt) "viod: " fmt Remove #define VIOD_KERN_WARNING and VIOD_KERN_INFO Convert printk(VIOD_KERN_ to pr_ Coalesce long format strings Signed-off-by: Joe Perches Acked-by: Stephen Rothwell drivers/block/viodasd.c | 86 +++++++++++++++++++--------------------------- 1 files changed, 36 insertions(+), 50 deletions(-) Signed-off-by: Benjamin Herrenschmidt diff --git a/drivers/block/viodasd.c b/drivers/block/viodasd.c index a8c8b56..1b3def1 100644 --- a/drivers/block/viodasd.c +++ b/drivers/block/viodasd.c @@ -28,6 +28,9 @@ * All disk operations are performed by sending messages back and forth to * the OS/400 partition. */ + +#define pr_fmt(fmt) "viod: " fmt + #include #include #include @@ -63,9 +66,6 @@ MODULE_LICENSE("GPL"); #define VIOD_VERS "1.64" -#define VIOD_KERN_WARNING KERN_WARNING "viod: " -#define VIOD_KERN_INFO KERN_INFO "viod: " - enum { PARTITION_SHIFT = 3, MAX_DISKNO = HVMAXARCHITECTEDVIRTUALDISKS, @@ -156,7 +156,7 @@ static int viodasd_open(struct block_device *bdev, fmode_t mode) ((u64)DEVICE_NO(d) << 48) | ((u64)flags << 32), 0, 0, 0); if (hvrc != 0) { - printk(VIOD_KERN_WARNING "HV open failed %d\n", (int)hvrc); + pr_warning("HV open failed %d\n", (int)hvrc); return -EIO; } @@ -167,9 +167,8 @@ static int viodasd_open(struct block_device *bdev, fmode_t mode) const struct vio_error_entry *err = vio_lookup_rc(viodasd_err_table, we.sub_result); - printk(VIOD_KERN_WARNING - "bad rc opening disk: %d:0x%04x (%s)\n", - (int)we.rc, we.sub_result, err->msg); + pr_warning("bad rc opening disk: %d:0x%04x (%s)\n", + (int)we.rc, we.sub_result, err->msg); return -EIO; } @@ -195,8 +194,7 @@ static int viodasd_release(struct gendisk *disk, fmode_t mode) ((u64)DEVICE_NO(d) << 48) /* | ((u64)flags << 32) */, 0, 0, 0); if (hvrc != 0) - printk(VIOD_KERN_WARNING "HV close call failed %d\n", - (int)hvrc); + pr_warning("HV close call failed %d\n", (int)hvrc); return 0; } @@ -288,8 +286,7 @@ static int send_request(struct request *req) bevent = (struct vioblocklpevent *) vio_get_event_buffer(viomajorsubtype_blockio); if (bevent == NULL) { - printk(VIOD_KERN_WARNING - "error allocating disk event buffer\n"); + pr_warning("error allocating disk event buffer\n"); goto error_ret; } @@ -333,9 +330,8 @@ static int send_request(struct request *req) } if (hvrc != HvLpEvent_Rc_Good) { - printk(VIOD_KERN_WARNING - "error sending disk event to OS/400 (rc %d)\n", - (int)hvrc); + pr_warning("error sending disk event to OS/400 (rc %d)\n", + (int)hvrc); goto error_ret; } spin_unlock_irqrestore(&viodasd_spinlock, flags); @@ -402,7 +398,7 @@ retry: ((u64)dev_no << 48) | ((u64)flags<< 32), 0, 0, 0); if (hvrc != 0) { - printk(VIOD_KERN_WARNING "bad rc on HV open %d\n", (int)hvrc); + pr_warning("bad rc on HV open %d\n", (int)hvrc); return 0; } @@ -416,9 +412,8 @@ retry: goto retry; } if (we.max_disk > (MAX_DISKNO - 1)) { - printk_once(VIOD_KERN_INFO - "Only examining the first %d of %d disks connected\n", - MAX_DISKNO, we.max_disk + 1); + printk_once(KERN_INFO pr_fmt("Only examining the first %d of %d disks connected\n"), + MAX_DISKNO, we.max_disk + 1); } /* Send the close event to OS/400. We DON'T expect a response */ @@ -432,17 +427,15 @@ retry: ((u64)dev_no << 48) | ((u64)flags << 32), 0, 0, 0); if (hvrc != 0) { - printk(VIOD_KERN_WARNING - "bad rc sending event to OS/400 %d\n", (int)hvrc); + pr_warning("bad rc sending event to OS/400 %d\n", (int)hvrc); return 0; } if (d->dev == NULL) { /* this is when we reprobe for new disks */ if (vio_create_viodasd(dev_no) == NULL) { - printk(VIOD_KERN_WARNING - "cannot allocate virtual device for disk %d\n", - dev_no); + pr_warning("cannot allocate virtual device for disk %d\n", + dev_no); return 0; } /* @@ -457,15 +450,13 @@ retry: spin_lock_init(&d->q_lock); q = blk_init_queue(do_viodasd_request, &d->q_lock); if (q == NULL) { - printk(VIOD_KERN_WARNING "cannot allocate queue for disk %d\n", - dev_no); + pr_warning("cannot allocate queue for disk %d\n", dev_no); return 0; } g = alloc_disk(1 << PARTITION_SHIFT); if (g == NULL) { - printk(VIOD_KERN_WARNING - "cannot allocate disk structure for disk %d\n", - dev_no); + pr_warning("cannot allocate disk structure for disk %d\n", + dev_no); blk_cleanup_queue(q); return 0; } @@ -489,13 +480,12 @@ retry: g->driverfs_dev = d->dev; set_capacity(g, d->size >> 9); - printk(VIOD_KERN_INFO "disk %d: %lu sectors (%lu MB) " - "CHS=%d/%d/%d sector size %d%s\n", - dev_no, (unsigned long)(d->size >> 9), - (unsigned long)(d->size >> 20), - (int)d->cylinders, (int)d->tracks, - (int)d->sectors, (int)d->bytes_per_sector, - d->read_only ? " (RO)" : ""); + pr_info("disk %d: %lu sectors (%lu MB) CHS=%d/%d/%d sector size %d%s\n", + dev_no, (unsigned long)(d->size >> 9), + (unsigned long)(d->size >> 20), + (int)d->cylinders, (int)d->tracks, + (int)d->sectors, (int)d->bytes_per_sector, + d->read_only ? " (RO)" : ""); /* register us in the global list */ add_disk(g); @@ -580,8 +570,8 @@ static int viodasd_handle_read_write(struct vioblocklpevent *bevent) if (error) { const struct vio_error_entry *err; err = vio_lookup_rc(viodasd_err_table, bevent->sub_result); - printk(VIOD_KERN_WARNING "read/write error %d:0x%04x (%s)\n", - event->xRc, bevent->sub_result, err->msg); + pr_warning("read/write error %d:0x%04x (%s)\n", + event->xRc, bevent->sub_result, err->msg); num_sect = blk_rq_sectors(req); } qlock = req->q->queue_lock; @@ -606,8 +596,7 @@ static void handle_block_event(struct HvLpEvent *event) return; /* First, we should NEVER get an int here...only acks */ if (hvlpevent_is_int(event)) { - printk(VIOD_KERN_WARNING - "Yikes! got an int in viodasd event handler!\n"); + pr_warning("Yikes! got an int in viodasd event handler!\n"); if (hvlpevent_need_ack(event)) { event->xRc = HvLpEvent_Rc_InvalidSubtype; HvCallEvent_ackLpEvent(event); @@ -650,7 +639,7 @@ static void handle_block_event(struct HvLpEvent *event) break; default: - printk(VIOD_KERN_WARNING "invalid subtype!"); + pr_warning("invalid subtype!"); if (hvlpevent_need_ack(event)) { event->xRc = HvLpEvent_Rc_InvalidSubtype; HvCallEvent_ackLpEvent(event); @@ -739,29 +728,26 @@ static int __init viodasd_init(void) vio_set_hostlp(); if (viopath_hostLp == HvLpIndexInvalid) { - printk(VIOD_KERN_WARNING "invalid hosting partition\n"); + pr_warning("invalid hosting partition\n"); rc = -EIO; goto early_fail; } - printk(VIOD_KERN_INFO "vers " VIOD_VERS ", hosting partition %d\n", - viopath_hostLp); + pr_info("vers " VIOD_VERS ", hosting partition %d\n", viopath_hostLp); /* register the block device */ rc = register_blkdev(VIODASD_MAJOR, VIOD_GENHD_NAME); if (rc) { - printk(VIOD_KERN_WARNING - "Unable to get major number %d for %s\n", - VIODASD_MAJOR, VIOD_GENHD_NAME); + pr_warning("Unable to get major number %d for %s\n", + VIODASD_MAJOR, VIOD_GENHD_NAME); goto early_fail; } /* Actually open the path to the hosting partition */ rc = viopath_open(viopath_hostLp, viomajorsubtype_blockio, VIOMAXREQ + 2); if (rc) { - printk(VIOD_KERN_WARNING - "error opening path to host partition %d\n", - viopath_hostLp); + pr_warning("error opening path to host partition %d\n", + viopath_hostLp); goto unregister_blk; } @@ -770,7 +756,7 @@ static int __init viodasd_init(void) rc = vio_register_driver(&viodasd_driver); if (rc) { - printk(VIOD_KERN_WARNING "vio_register_driver failed\n"); + pr_warning("vio_register_driver failed\n"); goto unset_handler; } -- cgit v0.10.2 From 4ba525d134a882e6994bf8f5bd93fd020376588c Mon Sep 17 00:00:00 2001 From: "Robert P. J. Day" Date: Thu, 31 Dec 2009 10:45:50 +0000 Subject: powerpc: Simplify param.h by including Signed-off-by: Robert P. J. Day Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/include/asm/param.h b/arch/powerpc/include/asm/param.h index 094f63d..965d454 100644 --- a/arch/powerpc/include/asm/param.h +++ b/arch/powerpc/include/asm/param.h @@ -1,22 +1 @@ -#ifndef _ASM_POWERPC_PARAM_H -#define _ASM_POWERPC_PARAM_H - -#ifdef __KERNEL__ -#define HZ CONFIG_HZ /* internal kernel timer frequency */ -#define USER_HZ 100 /* for user interfaces in "ticks" */ -#define CLOCKS_PER_SEC (USER_HZ) /* frequency at which times() counts */ -#endif /* __KERNEL__ */ - -#ifndef HZ -#define HZ 100 -#endif - -#define EXEC_PAGESIZE 4096 - -#ifndef NOGROUP -#define NOGROUP (-1) -#endif - -#define MAXHOSTNAMELEN 64 /* max length of hostname */ - -#endif /* _ASM_POWERPC_PARAM_H */ +#include -- cgit v0.10.2 From 61c03ddbdff2e080702521bc1b757854ca58c46a Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 13 Jan 2010 12:04:11 +0000 Subject: powerpc: Replace per_cpu(, smp_processor_id()) with __get_cpu_var() The cputime code has a few places that do per_cpu(, smp_processor_id()). Replace them with __get_cpu_var(). Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/include/asm/cputime.h b/arch/powerpc/include/asm/cputime.h index fa19f3f..8bdc6a9 100644 --- a/arch/powerpc/include/asm/cputime.h +++ b/arch/powerpc/include/asm/cputime.h @@ -73,10 +73,9 @@ static inline unsigned long cputime_to_jiffies(const cputime_t ct) static inline cputime_t cputime_to_scaled(const cputime_t ct) { if (cpu_has_feature(CPU_FTR_SPURR) && - per_cpu(cputime_last_delta, smp_processor_id())) - return ct * - per_cpu(cputime_scaled_last_delta, smp_processor_id())/ - per_cpu(cputime_last_delta, smp_processor_id()); + __get_cpu_var(cputime_last_delta)) + return ct * __get_cpu_var(cputime_scaled_last_delta) / + __get_cpu_var(cputime_last_delta); return ct; } diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 6c6093d..0a5b108 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -265,8 +265,8 @@ void account_system_vtime(struct task_struct *tsk) account_system_time(tsk, 0, delta, deltascaled); else account_idle_time(delta); - per_cpu(cputime_last_delta, smp_processor_id()) = delta; - per_cpu(cputime_scaled_last_delta, smp_processor_id()) = deltascaled; + __get_cpu_var(cputime_last_delta) = delta; + __get_cpu_var(cputime_scaled_last_delta) = deltascaled; local_irq_restore(flags); } EXPORT_SYMBOL_GPL(account_system_vtime); -- cgit v0.10.2 From 5be3492f972b73051ead7ecbac6fb9efd1e8e0ec Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Tue, 12 Jan 2010 00:50:14 +0000 Subject: powerpc: Mark some variables in the page fault path __read_mostly Using perf to trace L1 dcache misses and dumping data addresses I found a few variables taking a lot of misses. Since they are almost never written, they should go into the __read_mostly section. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c index d069ff8..4588d1e 100644 --- a/arch/powerpc/kernel/traps.c +++ b/arch/powerpc/kernel/traps.c @@ -60,13 +60,13 @@ #endif #if defined(CONFIG_DEBUGGER) || defined(CONFIG_KEXEC) -int (*__debugger)(struct pt_regs *regs); -int (*__debugger_ipi)(struct pt_regs *regs); -int (*__debugger_bpt)(struct pt_regs *regs); -int (*__debugger_sstep)(struct pt_regs *regs); -int (*__debugger_iabr_match)(struct pt_regs *regs); -int (*__debugger_dabr_match)(struct pt_regs *regs); -int (*__debugger_fault_handler)(struct pt_regs *regs); +int (*__debugger)(struct pt_regs *regs) __read_mostly; +int (*__debugger_ipi)(struct pt_regs *regs) __read_mostly; +int (*__debugger_bpt)(struct pt_regs *regs) __read_mostly; +int (*__debugger_sstep)(struct pt_regs *regs) __read_mostly; +int (*__debugger_iabr_match)(struct pt_regs *regs) __read_mostly; +int (*__debugger_dabr_match)(struct pt_regs *regs) __read_mostly; +int (*__debugger_fault_handler)(struct pt_regs *regs) __read_mostly; EXPORT_SYMBOL(__debugger); EXPORT_SYMBOL(__debugger_ipi); diff --git a/arch/powerpc/xmon/xmon.c b/arch/powerpc/xmon/xmon.c index 4e6152c..8bad7d5 100644 --- a/arch/powerpc/xmon/xmon.c +++ b/arch/powerpc/xmon/xmon.c @@ -61,7 +61,7 @@ static int xmon_owner; static int xmon_gate; #endif /* CONFIG_SMP */ -static unsigned long in_xmon = 0; +static unsigned long in_xmon __read_mostly = 0; static unsigned long adrs; static int size = 1; -- cgit v0.10.2 From 7b62922a071aea362e879252d7482e448bd63d9c Mon Sep 17 00:00:00 2001 From: Peter Tyser Date: Fri, 18 Dec 2009 12:50:37 +0000 Subject: powerpc/85xx: Fix SMP when "cpu-release-addr" is in lowmem Recent U-Boot commit 5ccd29c3679b3669b0bde5c501c1aa0f325a7acb caused the "cpu-release-addr" device tree property to contain the physical RAM location that secondary cores were spinning at. Previously, the "cpu-release-addr" property contained a value referencing the boot page translation address range of 0xfffffxxx, which then indirectly accessed RAM. The "cpu-release-addr" is currently ioremapped and the secondary cores kicked. However, due to the recent change in "cpu-release-addr", it sometimes points to a memory location in low memory that cannot be ioremapped. For example on a P2020-based board with 512MB of RAM the following error occurs on bootup: <...> mpic: requesting IPIs ... __ioremap(): phys addr 0x1ffff000 is RAM lr c05df9a0 Unable to handle kernel paging request for data at address 0x00000014 Faulting instruction address: 0xc05df9b0 Oops: Kernel access of bad area, sig: 11 [#1] SMP NR_CPUS=2 P2020 RDB Modules linked in: <... eventual kernel panic> Adding logic to conditionally ioremap or access memory directly resolves the issue. Signed-off-by: Peter Tyser Signed-off-by: Nate Case Reported-by: Dipen Dudhat Tested-by: Dipen Dudhat Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/platforms/85xx/smp.c b/arch/powerpc/platforms/85xx/smp.c index 04160a4..a15f582 100644 --- a/arch/powerpc/platforms/85xx/smp.c +++ b/arch/powerpc/platforms/85xx/smp.c @@ -46,6 +46,7 @@ smp_85xx_kick_cpu(int nr) __iomem u32 *bptr_vaddr; struct device_node *np; int n = 0; + int ioremappable; WARN_ON (nr < 0 || nr >= NR_CPUS); @@ -59,21 +60,37 @@ smp_85xx_kick_cpu(int nr) return; } + /* + * A secondary core could be in a spinloop in the bootpage + * (0xfffff000), somewhere in highmem, or somewhere in lowmem. + * The bootpage and highmem can be accessed via ioremap(), but + * we need to directly access the spinloop if its in lowmem. + */ + ioremappable = *cpu_rel_addr > virt_to_phys(high_memory); + /* Map the spin table */ - bptr_vaddr = ioremap(*cpu_rel_addr, SIZE_BOOT_ENTRY); + if (ioremappable) + bptr_vaddr = ioremap(*cpu_rel_addr, SIZE_BOOT_ENTRY); + else + bptr_vaddr = phys_to_virt(*cpu_rel_addr); local_irq_save(flags); out_be32(bptr_vaddr + BOOT_ENTRY_PIR, nr); out_be32(bptr_vaddr + BOOT_ENTRY_ADDR_LOWER, __pa(__early_start)); + if (!ioremappable) + flush_dcache_range((ulong)bptr_vaddr, + (ulong)(bptr_vaddr + SIZE_BOOT_ENTRY)); + /* Wait a bit for the CPU to ack. */ while ((__secondary_hold_acknowledge != nr) && (++n < 1000)) mdelay(1); local_irq_restore(flags); - iounmap(bptr_vaddr); + if (ioremappable) + iounmap(bptr_vaddr); pr_debug("waited %d msecs for CPU #%d.\n", n, nr); } -- cgit v0.10.2 From b51130685817d8c1b56386f9957405b5be2cdfe0 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Mon, 18 Jan 2010 03:44:57 +0000 Subject: hvc_console: Make the ops pointer const. This is nicer for modern R/O protection. And noone needs it non-const, so constify the callers as well. Signed-off-by: Rusty Russell Signed-off-by: Amit Shah To: Christian Borntraeger Cc: linuxppc-dev@ozlabs.org Signed-off-by: Benjamin Herrenschmidt diff --git a/drivers/char/hvc_beat.c b/drivers/char/hvc_beat.c index 0afc8b8..6913fc3 100644 --- a/drivers/char/hvc_beat.c +++ b/drivers/char/hvc_beat.c @@ -84,7 +84,7 @@ static int hvc_beat_put_chars(uint32_t vtermno, const char *buf, int cnt) return cnt; } -static struct hv_ops hvc_beat_get_put_ops = { +static const struct hv_ops hvc_beat_get_put_ops = { .get_chars = hvc_beat_get_chars, .put_chars = hvc_beat_put_chars, }; diff --git a/drivers/char/hvc_console.c b/drivers/char/hvc_console.c index 416d342..d8dac58 100644 --- a/drivers/char/hvc_console.c +++ b/drivers/char/hvc_console.c @@ -125,7 +125,7 @@ static struct hvc_struct *hvc_get_by_index(int index) * console interfaces but can still be used as a tty device. This has to be * static because kmalloc will not work during early console init. */ -static struct hv_ops *cons_ops[MAX_NR_HVC_CONSOLES]; +static const struct hv_ops *cons_ops[MAX_NR_HVC_CONSOLES]; static uint32_t vtermnos[MAX_NR_HVC_CONSOLES] = {[0 ... MAX_NR_HVC_CONSOLES - 1] = -1}; @@ -247,7 +247,7 @@ static void destroy_hvc_struct(struct kref *kref) * vty adapters do NOT get an hvc_instantiate() callback since they * appear after early console init. */ -int hvc_instantiate(uint32_t vtermno, int index, struct hv_ops *ops) +int hvc_instantiate(uint32_t vtermno, int index, const struct hv_ops *ops) { struct hvc_struct *hp; @@ -749,7 +749,8 @@ static const struct tty_operations hvc_ops = { }; struct hvc_struct __devinit *hvc_alloc(uint32_t vtermno, int data, - struct hv_ops *ops, int outbuf_size) + const struct hv_ops *ops, + int outbuf_size) { struct hvc_struct *hp; int i; diff --git a/drivers/char/hvc_console.h b/drivers/char/hvc_console.h index 10950ca..52ddf4d 100644 --- a/drivers/char/hvc_console.h +++ b/drivers/char/hvc_console.h @@ -55,7 +55,7 @@ struct hvc_struct { int outbuf_size; int n_outbuf; uint32_t vtermno; - struct hv_ops *ops; + const struct hv_ops *ops; int irq_requested; int data; struct winsize ws; @@ -76,11 +76,12 @@ struct hv_ops { }; /* Register a vterm and a slot index for use as a console (console_init) */ -extern int hvc_instantiate(uint32_t vtermno, int index, struct hv_ops *ops); +extern int hvc_instantiate(uint32_t vtermno, int index, + const struct hv_ops *ops); /* register a vterm for hvc tty operation (module_init or hotplug add) */ extern struct hvc_struct * __devinit hvc_alloc(uint32_t vtermno, int data, - struct hv_ops *ops, int outbuf_size); + const struct hv_ops *ops, int outbuf_size); /* remove a vterm from hvc tty operation (module_exit or hotplug remove) */ extern int hvc_remove(struct hvc_struct *hp); diff --git a/drivers/char/hvc_iseries.c b/drivers/char/hvc_iseries.c index 936d05b..fd02426 100644 --- a/drivers/char/hvc_iseries.c +++ b/drivers/char/hvc_iseries.c @@ -197,7 +197,7 @@ done: return sent; } -static struct hv_ops hvc_get_put_ops = { +static const struct hv_ops hvc_get_put_ops = { .get_chars = get_chars, .put_chars = put_chars, .notifier_add = notifier_add_irq, diff --git a/drivers/char/hvc_iucv.c b/drivers/char/hvc_iucv.c index fe62bd0..21681a8 100644 --- a/drivers/char/hvc_iucv.c +++ b/drivers/char/hvc_iucv.c @@ -922,7 +922,7 @@ static int hvc_iucv_pm_restore_thaw(struct device *dev) /* HVC operations */ -static struct hv_ops hvc_iucv_ops = { +static const struct hv_ops hvc_iucv_ops = { .get_chars = hvc_iucv_get_chars, .put_chars = hvc_iucv_put_chars, .notifier_add = hvc_iucv_notifier_add, diff --git a/drivers/char/hvc_rtas.c b/drivers/char/hvc_rtas.c index 88590d0..61c4a61 100644 --- a/drivers/char/hvc_rtas.c +++ b/drivers/char/hvc_rtas.c @@ -71,7 +71,7 @@ static int hvc_rtas_read_console(uint32_t vtermno, char *buf, int count) return i; } -static struct hv_ops hvc_rtas_get_put_ops = { +static const struct hv_ops hvc_rtas_get_put_ops = { .get_chars = hvc_rtas_read_console, .put_chars = hvc_rtas_write_console, }; diff --git a/drivers/char/hvc_udbg.c b/drivers/char/hvc_udbg.c index bd63ba8..b0957e6 100644 --- a/drivers/char/hvc_udbg.c +++ b/drivers/char/hvc_udbg.c @@ -58,7 +58,7 @@ static int hvc_udbg_get(uint32_t vtermno, char *buf, int count) return i; } -static struct hv_ops hvc_udbg_ops = { +static const struct hv_ops hvc_udbg_ops = { .get_chars = hvc_udbg_get, .put_chars = hvc_udbg_put, }; diff --git a/drivers/char/hvc_vio.c b/drivers/char/hvc_vio.c index 10be343..27370e9 100644 --- a/drivers/char/hvc_vio.c +++ b/drivers/char/hvc_vio.c @@ -77,7 +77,7 @@ static int filtered_get_chars(uint32_t vtermno, char *buf, int count) return got; } -static struct hv_ops hvc_get_put_ops = { +static const struct hv_ops hvc_get_put_ops = { .get_chars = filtered_get_chars, .put_chars = hvc_put_chars, .notifier_add = notifier_add_irq, diff --git a/drivers/char/hvc_xen.c b/drivers/char/hvc_xen.c index b1a7163..60446f8 100644 --- a/drivers/char/hvc_xen.c +++ b/drivers/char/hvc_xen.c @@ -122,7 +122,7 @@ static int read_console(uint32_t vtermno, char *buf, int len) return recv; } -static struct hv_ops hvc_ops = { +static const struct hv_ops hvc_ops = { .get_chars = read_console, .put_chars = write_console, .notifier_add = notifier_add_irq, -- cgit v0.10.2 From 119ea10947cc1402abbf9d6200815b0606536906 Mon Sep 17 00:00:00 2001 From: Amit Shah Date: Mon, 18 Jan 2010 03:44:58 +0000 Subject: hvc_console: Remove __devinit annotation from hvc_alloc Virtio consoles can be hotplugged, so hvc_alloc gets called from multiple sites: from the initial probe() routine as well as later on from workqueue handlers which aren't __devinit code. So, drop the __devinit annotation for hvc_alloc. Signed-off-by: Amit Shah Cc: linuxppc-dev@ozlabs.org Signed-off-by: Benjamin Herrenschmidt diff --git a/drivers/char/hvc_console.c b/drivers/char/hvc_console.c index d8dac58..4c3b59b 100644 --- a/drivers/char/hvc_console.c +++ b/drivers/char/hvc_console.c @@ -748,9 +748,9 @@ static const struct tty_operations hvc_ops = { .chars_in_buffer = hvc_chars_in_buffer, }; -struct hvc_struct __devinit *hvc_alloc(uint32_t vtermno, int data, - const struct hv_ops *ops, - int outbuf_size) +struct hvc_struct *hvc_alloc(uint32_t vtermno, int data, + const struct hv_ops *ops, + int outbuf_size) { struct hvc_struct *hp; int i; diff --git a/drivers/char/hvc_console.h b/drivers/char/hvc_console.h index 52ddf4d..54381eba 100644 --- a/drivers/char/hvc_console.h +++ b/drivers/char/hvc_console.h @@ -80,8 +80,8 @@ extern int hvc_instantiate(uint32_t vtermno, int index, const struct hv_ops *ops); /* register a vterm for hvc tty operation (module_init or hotplug add) */ -extern struct hvc_struct * __devinit hvc_alloc(uint32_t vtermno, int data, - const struct hv_ops *ops, int outbuf_size); +extern struct hvc_struct * hvc_alloc(uint32_t vtermno, int data, + const struct hv_ops *ops, int outbuf_size); /* remove a vterm from hvc tty operation (module_exit or hotplug remove) */ extern int hvc_remove(struct hvc_struct *hp); -- cgit v0.10.2 From 33a470f6d5e1879c26f16f6b34dc09f82d44f6e9 Mon Sep 17 00:00:00 2001 From: Jean Delvare Date: Sun, 31 Jan 2010 04:00:30 +0000 Subject: macintosh/therm_adt746x: Fix sysfs attributes lifetime Looking at drivers/macintosh/therm_adt746x.c, the sysfs files are created in thermostat_init() and removed in thermostat_exit(), which are the driver's init and exit functions. These files are backed-up by a per-device structure, so it looks like the wrong thing to do: the sysfs files have a lifetime longer than the data structure that is backing it up. I think that sysfs files creation should be moved to the end of probe_thermostat() and sysfs files removal should be moved to the beginning of remove_thermostat(). Signed-off-by: Jean Delvare Tested-by: Christian Kujau Cc: Benjamin Herrenschmidt Cc: Colin Leroy Cc: stable@kernel.org Signed-off-by: Benjamin Herrenschmidt diff --git a/drivers/macintosh/therm_adt746x.c b/drivers/macintosh/therm_adt746x.c index 5ff47ba..58809b0 100644 --- a/drivers/macintosh/therm_adt746x.c +++ b/drivers/macintosh/therm_adt746x.c @@ -90,6 +90,8 @@ static struct task_struct *thread_therm = NULL; static void write_both_fan_speed(struct thermostat *th, int speed); static void write_fan_speed(struct thermostat *th, int speed, int fan); +static void thermostat_create_files(void); +static void thermostat_remove_files(void); static int write_reg(struct thermostat* th, int reg, u8 data) @@ -161,6 +163,8 @@ remove_thermostat(struct i2c_client *client) struct thermostat *th = i2c_get_clientdata(client); int i; + thermostat_remove_files(); + if (thread_therm != NULL) { kthread_stop(thread_therm); } @@ -449,6 +453,8 @@ static int probe_thermostat(struct i2c_client *client, return -ENOMEM; } + thermostat_create_files(); + return 0; } @@ -566,7 +572,6 @@ thermostat_init(void) struct device_node* np; const u32 *prop; int i = 0, offset = 0; - int err; np = of_find_node_by_name(NULL, "fan"); if (!np) @@ -633,6 +638,17 @@ thermostat_init(void) return -ENODEV; } +#ifndef CONFIG_I2C_POWERMAC + request_module("i2c-powermac"); +#endif + + return i2c_add_driver(&thermostat_driver); +} + +static void thermostat_create_files(void) +{ + int err; + err = device_create_file(&of_dev->dev, &dev_attr_sensor1_temperature); err |= device_create_file(&of_dev->dev, &dev_attr_sensor2_temperature); err |= device_create_file(&of_dev->dev, &dev_attr_sensor1_limit); @@ -647,16 +663,9 @@ thermostat_init(void) if (err) printk(KERN_WARNING "Failed to create tempertaure attribute file(s).\n"); - -#ifndef CONFIG_I2C_POWERMAC - request_module("i2c-powermac"); -#endif - - return i2c_add_driver(&thermostat_driver); } -static void __exit -thermostat_exit(void) +static void thermostat_remove_files(void) { if (of_dev) { device_remove_file(&of_dev->dev, &dev_attr_sensor1_temperature); @@ -673,9 +682,14 @@ thermostat_exit(void) device_remove_file(&of_dev->dev, &dev_attr_sensor2_fan_speed); - of_device_unregister(of_dev); } +} + +static void __exit +thermostat_exit(void) +{ i2c_del_driver(&thermostat_driver); + of_device_unregister(of_dev); } module_init(thermostat_init); -- cgit v0.10.2 From 98ceb75c7c14eada76b0aa9f03a635a735cee3cb Mon Sep 17 00:00:00 2001 From: Jean Delvare Date: Sun, 31 Jan 2010 04:03:23 +0000 Subject: macintosh/hwmon/ams: Fix device removal sequence Some code that is in ams_exit() (the module exit code) should instead be called when the device (not module) is removed. It probably doesn't make much of a difference in the PMU case, but in the I2C case it does matter. I make no guarantee that my fix isn't racy, I'm not familiar enough with the ams driver code to tell for sure. Signed-off-by: Jean Delvare Tested-by: Christian Kujau Cc: Benjamin Herrenschmidt Cc: Stelian Pop Cc: Michael Hanselmann Cc: stable@kernel.org Signed-off-by: Benjamin Herrenschmidt diff --git a/drivers/hwmon/ams/ams-core.c b/drivers/hwmon/ams/ams-core.c index 6c9ace1..2ad62c3 100644 --- a/drivers/hwmon/ams/ams-core.c +++ b/drivers/hwmon/ams/ams-core.c @@ -213,7 +213,7 @@ int __init ams_init(void) return -ENODEV; } -void ams_exit(void) +void ams_sensor_detach(void) { /* Remove input device */ ams_input_exit(); @@ -221,9 +221,6 @@ void ams_exit(void) /* Remove attributes */ device_remove_file(&ams_info.of_dev->dev, &dev_attr_current); - /* Shut down implementation */ - ams_info.exit(); - /* Flush interrupt worker * * We do this after ams_info.exit(), because an interrupt might @@ -239,6 +236,12 @@ void ams_exit(void) pmf_unregister_irq_client(&ams_freefall_client); } +static void __exit ams_exit(void) +{ + /* Shut down implementation */ + ams_info.exit(); +} + MODULE_AUTHOR("Stelian Pop, Michael Hanselmann"); MODULE_DESCRIPTION("Apple Motion Sensor driver"); MODULE_LICENSE("GPL"); diff --git a/drivers/hwmon/ams/ams-i2c.c b/drivers/hwmon/ams/ams-i2c.c index 2cbf8a6..abeecd2 100644 --- a/drivers/hwmon/ams/ams-i2c.c +++ b/drivers/hwmon/ams/ams-i2c.c @@ -238,6 +238,8 @@ static int ams_i2c_probe(struct i2c_client *client, static int ams_i2c_remove(struct i2c_client *client) { if (ams_info.has_device) { + ams_sensor_detach(); + /* Disable interrupts */ ams_i2c_set_irq(AMS_IRQ_ALL, 0); diff --git a/drivers/hwmon/ams/ams-pmu.c b/drivers/hwmon/ams/ams-pmu.c index fb18b3d..4f61b3e 100644 --- a/drivers/hwmon/ams/ams-pmu.c +++ b/drivers/hwmon/ams/ams-pmu.c @@ -133,6 +133,8 @@ static void ams_pmu_get_xyz(s8 *x, s8 *y, s8 *z) static void ams_pmu_exit(void) { + ams_sensor_detach(); + /* Disable interrupts */ ams_pmu_set_irq(AMS_IRQ_ALL, 0); diff --git a/drivers/hwmon/ams/ams.h b/drivers/hwmon/ams/ams.h index 5ed387b..b28d7e2 100644 --- a/drivers/hwmon/ams/ams.h +++ b/drivers/hwmon/ams/ams.h @@ -61,6 +61,7 @@ extern struct ams ams_info; extern void ams_sensors(s8 *x, s8 *y, s8 *z); extern int ams_sensor_attach(void); +extern void ams_sensor_detach(void); extern int ams_pmu_init(struct device_node *np); extern int ams_i2c_init(struct device_node *np); -- cgit v0.10.2 From 859aefc5afc6c00dae630680e5470f7a2c27c4a0 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 31 Jan 2010 01:14:03 +0000 Subject: powerpc: Increase NR_IRQS Kconfig maximum to 32768 With dynamic irq descriptors the overhead of a large NR_IRQS is much lower than it used to be. With more MSI-X capable adapters and drivers exploiting multiple vectors we may as well allow the user to increase it beyond the current maximum of 512. 32768 seems large enough that we'd never have to bump it again (although I bet my prediction is horribly wrong). It boot tests OK and the vmlinux footprint increase is only around 500kB due to: struct irq_map_entry irq_map[NR_IRQS]; We format /proc/interrupts correctly with the previous changes: CPU0 CPU1 CPU2 CPU3 CPU4 CPU5 286: 0 0 0 0 0 0 516: 0 0 0 0 0 0 16689: 1833 0 0 0 0 0 17157: 0 0 0 0 0 0 17158: 319 0 0 0 0 0 25092: 0 0 0 0 0 0 Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index ba3948c..bf15e7b 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -58,7 +58,7 @@ config IRQ_PER_CPU config NR_IRQS int "Number of virtual interrupt numbers" - range 32 512 + range 32 32768 default "512" help This defines the number of virtual interrupt numbers the kernel -- cgit v0.10.2 From 24551f64d47af9539a7f324343bffeea09d9dcfa Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Tue, 12 Jan 2010 21:25:24 +0000 Subject: lmb: Add lmb_free() We can free memory allocated with lmb_alloc() by removing it from the list of reserved LMBs. Rework lmb_remove() to allow that possibility and add lmb_free() which exploits it. BenH: Removed some useless parenthesis Signed-off-by: Michael Ellerman Signed-off-by: Benjamin Herrenschmidt diff --git a/include/linux/lmb.h b/include/linux/lmb.h index ef82b8f..f3d1433 100644 --- a/include/linux/lmb.h +++ b/include/linux/lmb.h @@ -42,6 +42,7 @@ extern void __init lmb_init(void); extern void __init lmb_analyze(void); extern long lmb_add(u64 base, u64 size); extern long lmb_remove(u64 base, u64 size); +extern long __init lmb_free(u64 base, u64 size); extern long __init lmb_reserve(u64 base, u64 size); extern u64 __init lmb_alloc_nid(u64 size, u64 align, int nid, u64 (*nid_range)(u64, u64, int *)); diff --git a/lib/lmb.c b/lib/lmb.c index 9cee171..b1fc526 100644 --- a/lib/lmb.c +++ b/lib/lmb.c @@ -205,9 +205,8 @@ long lmb_add(u64 base, u64 size) } -long lmb_remove(u64 base, u64 size) +static long __lmb_remove(struct lmb_region *rgn, u64 base, u64 size) { - struct lmb_region *rgn = &(lmb.memory); u64 rgnbegin, rgnend; u64 end = base + size; int i; @@ -254,6 +253,16 @@ long lmb_remove(u64 base, u64 size) return lmb_add_region(rgn, end, rgnend - end); } +long lmb_remove(u64 base, u64 size) +{ + return __lmb_remove(&lmb.memory, base, size); +} + +long __init lmb_free(u64 base, u64 size) +{ + return __lmb_remove(&lmb.reserved, base, size); +} + long __init lmb_reserve(u64 base, u64 size) { struct lmb_region *_rgn = &lmb.reserved; -- cgit v0.10.2 From 6c504d423147770d7bd73aab3228726379700786 Mon Sep 17 00:00:00 2001 From: Thadeu Lima de Souza Cascardo Date: Sun, 17 Jan 2010 11:23:14 +0000 Subject: powerpc: Fix typo s/leve/level/ in TLB code Signed-off-by: Thadeu Lima de Souza Cascardo Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/mm/tlb_low_64e.S b/arch/powerpc/mm/tlb_low_64e.S index f288279..8b04c54 100644 --- a/arch/powerpc/mm/tlb_low_64e.S +++ b/arch/powerpc/mm/tlb_low_64e.S @@ -1,5 +1,5 @@ /* - * Low leve TLB miss handlers for Book3E + * Low level TLB miss handlers for Book3E * * Copyright (C) 2008-2009 * Ben. Herrenschmidt (benh@kernel.crashing.org), IBM Corp. -- cgit v0.10.2 From 28bb9ee13aa0ee4c57dc3568f539cc84920b43aa Mon Sep 17 00:00:00 2001 From: "jschopp@austin.ibm.com" Date: Mon, 1 Feb 2010 12:50:48 +0000 Subject: powerpc: Add static fields to ibm,client-architecture call This patch adds 2 fields to the ibm_architecture_vec array. The first of these fields indicates the number of cores which Linux can boot. It does not account for SMT, so it may result in cpus assigned to Linux which cannot be booted. A second patch follows that dynamically updates this for SMT. The second field just indicates that our OS is Linux, and not another OS. The system may or may not use this hint to performance tune settings for Linux. Signed-off-by: Joel Schopp Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c index bafac2e..59d5bd1 100644 --- a/arch/powerpc/kernel/prom_init.c +++ b/arch/powerpc/kernel/prom_init.c @@ -654,6 +654,9 @@ static void __init early_cmdline_parse(void) #define OV5_CMO 0x00 #endif +/* Option Vector 6: IBM PAPR hints */ +#define OV6_LINUX 0x02 /* Linux is our OS */ + /* * The architecture vector has an array of PVR mask/value pairs, * followed by # option vectors - 1, followed by the option vectors. @@ -665,7 +668,7 @@ static unsigned char ibm_architecture_vec[] = { W(0xffffffff), W(0x0f000003), /* all 2.06-compliant */ W(0xffffffff), W(0x0f000002), /* all 2.05-compliant */ W(0xfffffffe), W(0x0f000001), /* all 2.04-compliant and earlier */ - 5 - 1, /* 5 option vectors */ + 6 - 1, /* 6 option vectors */ /* option vector 1: processor architectures supported */ 3 - 2, /* length */ @@ -697,12 +700,24 @@ static unsigned char ibm_architecture_vec[] = { 0, /* don't halt */ /* option vector 5: PAPR/OF options */ - 5 - 2, /* length */ + 13 - 2, /* length */ 0, /* don't ignore, don't halt */ OV5_LPAR | OV5_SPLPAR | OV5_LARGE_PAGES | OV5_DRCONF_MEMORY | OV5_DONATE_DEDICATE_CPU | OV5_MSI, 0, OV5_CMO, + 0, + 0, + 0, + 0, + W(NR_CPUS), /* number of cores supported*/ + + /* option vector 6: IBM PAPR hints */ + 4 - 2, /* length */ + 0, + 0, + OV6_LINUX, + }; /* Old method - ELF header with PT_NOTE sections */ -- cgit v0.10.2 From 9ad62ec4f752c82b39aa5927f23d894b46ae10b9 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 3 Feb 2010 16:46:20 +0900 Subject: sh: Fix up early PCI PERR/SERR IRQ handling. This adds support for handling early PERR/SERR triggering in between controller registration and the initial bus scan. Buggy cards end up asserting these as soon as the M66EN scan is undertaken, resulting in an early crash. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/common.c b/arch/sh/drivers/pci/common.c index 25aec00..dbf1381 100644 --- a/arch/sh/drivers/pci/common.c +++ b/arch/sh/drivers/pci/common.c @@ -3,29 +3,48 @@ #include #include -static int __init -early_read_config_word(struct pci_channel *hose, - int top_bus, int bus, int devfn, int offset, u16 *value) +/* + * These functions are used early on before PCI scanning is done + * and all of the pci_dev and pci_bus structures have been created. + */ +static struct pci_dev *fake_pci_dev(struct pci_channel *hose, + int top_bus, int busnr, int devfn) { - struct pci_dev fake_dev; - struct pci_bus fake_bus; + static struct pci_dev dev; + static struct pci_bus bus; - fake_dev.bus = &fake_bus; - fake_dev.sysdata = hose; - fake_dev.devfn = devfn; - fake_bus.number = bus; - fake_bus.sysdata = hose; - fake_bus.ops = hose->pci_ops; + dev.bus = &bus; + dev.sysdata = hose; + dev.devfn = devfn; + bus.number = busnr; + bus.sysdata = hose; + bus.ops = hose->pci_ops; - if (bus != top_bus) + if(busnr != top_bus) /* Fake a parent bus structure. */ - fake_bus.parent = &fake_bus; + bus.parent = &bus; else - fake_bus.parent = NULL; + bus.parent = NULL; - return pci_read_config_word(&fake_dev, offset, value); + return &dev; } +#define EARLY_PCI_OP(rw, size, type) \ +int __init early_##rw##_config_##size(struct pci_channel *hose, \ + int top_bus, int bus, int devfn, int offset, type value) \ +{ \ + return pci_##rw##_config_##size( \ + fake_pci_dev(hose, top_bus, bus, devfn), \ + offset, value); \ +} + +EARLY_PCI_OP(read, byte, u8 *) +EARLY_PCI_OP(read, word, u16 *) +EARLY_PCI_OP(read, dword, u32 *) +EARLY_PCI_OP(write, byte, u8) +EARLY_PCI_OP(write, word, u16) +EARLY_PCI_OP(write, dword, u32) + int __init pci_is_66mhz_capable(struct pci_channel *hose, int top_bus, int current_bus) { @@ -133,7 +152,7 @@ unsigned int pcibios_handle_status_errors(unsigned long addr, /* Now back off of the IRQ for awhile */ if (hose->err_irq) { - disable_irq(hose->err_irq); + disable_irq_nosync(hose->err_irq); hose->err_timer.expires = jiffies + HZ; add_timer(&hose->err_timer); } diff --git a/arch/sh/drivers/pci/pci-sh7780.c b/arch/sh/drivers/pci/pci-sh7780.c index 472f67a..1e147f4 100644 --- a/arch/sh/drivers/pci/pci-sh7780.c +++ b/arch/sh/drivers/pci/pci-sh7780.c @@ -150,7 +150,7 @@ static irqreturn_t sh7780_pci_serr_irq(int irq, void *dev_id) __raw_writel(SH4_PCIINTM_SDIM, hose->reg_base + SH4_PCIINTM); /* Back off the IRQ for awhile */ - disable_irq(irq); + disable_irq_nosync(irq); hose->serr_timer.expires = jiffies + HZ; add_timer(&hose->serr_timer); diff --git a/arch/sh/drivers/pci/pci.c b/arch/sh/drivers/pci/pci.c index f4a6983..41d8f01 100644 --- a/arch/sh/drivers/pci/pci.c +++ b/arch/sh/drivers/pci/pci.c @@ -204,7 +204,7 @@ void pcibios_align_resource(void *data, struct resource *res, } void pcibios_resource_to_bus(struct pci_dev *dev, struct pci_bus_region *region, - struct resource *res) + struct resource *res) { struct pci_channel *hose = dev->sysdata; unsigned long offset = 0; @@ -218,9 +218,8 @@ void pcibios_resource_to_bus(struct pci_dev *dev, struct pci_bus_region *region, region->end = res->end - offset; } -void __devinit -pcibios_bus_to_resource(struct pci_dev *dev, struct resource *res, - struct pci_bus_region *region) +void pcibios_bus_to_resource(struct pci_dev *dev, struct resource *res, + struct pci_bus_region *region) { struct pci_channel *hose = dev->sysdata; unsigned long offset = 0; @@ -303,12 +302,41 @@ char * __devinit pcibios_setup(char *str) return str; } +static void __init +pcibios_bus_report_status_early(struct pci_channel *hose, + int top_bus, int current_bus, + unsigned int status_mask, int warn) +{ + unsigned int pci_devfn; + u16 status; + int ret; + + for (pci_devfn = 0; pci_devfn < 0xff; pci_devfn++) { + if (PCI_FUNC(pci_devfn)) + continue; + ret = early_read_config_word(hose, top_bus, current_bus, + pci_devfn, PCI_STATUS, &status); + if (ret != PCIBIOS_SUCCESSFUL) + continue; + if (status == 0xffff) + continue; + + early_write_config_word(hose, top_bus, current_bus, + pci_devfn, PCI_STATUS, + status & status_mask); + if (warn) + printk("(%02x:%02x: %04X) ", current_bus, + pci_devfn, status); + } +} + /* * We can't use pci_find_device() here since we are * called from interrupt context. */ -static void pcibios_bus_report_status(struct pci_bus *bus, - unsigned int status_mask, int warn) +static void __init_refok +pcibios_bus_report_status(struct pci_bus *bus, unsigned int status_mask, + int warn) { struct pci_dev *dev; @@ -341,12 +369,17 @@ static void pcibios_bus_report_status(struct pci_bus *bus, pcibios_bus_report_status(dev->subordinate, status_mask, warn); } -void pcibios_report_status(unsigned int status_mask, int warn) +void __init_refok pcibios_report_status(unsigned int status_mask, int warn) { struct pci_channel *hose; - for (hose = hose_head; hose; hose = hose->next) - pcibios_bus_report_status(hose->bus, status_mask, warn); + for (hose = hose_head; hose; hose = hose->next) { + if (unlikely(!hose->bus)) + pcibios_bus_report_status_early(hose, hose_head->index, + hose->index, status_mask, warn); + else + pcibios_bus_report_status(hose->bus, status_mask, warn); + } } int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma, diff --git a/arch/sh/include/asm/pci.h b/arch/sh/include/asm/pci.h index 6d762cc..1042f7f 100644 --- a/arch/sh/include/asm/pci.h +++ b/arch/sh/include/asm/pci.h @@ -41,6 +41,18 @@ extern int register_pci_controller(struct pci_channel *hose); extern void pcibios_report_status(unsigned int status_mask, int warn); /* arch/sh/drivers/pci/common.c */ +extern int early_read_config_byte(struct pci_channel *hose, int top_bus, + int bus, int devfn, int offset, u8 *value); +extern int early_read_config_word(struct pci_channel *hose, int top_bus, + int bus, int devfn, int offset, u16 *value); +extern int early_read_config_dword(struct pci_channel *hose, int top_bus, + int bus, int devfn, int offset, u32 *value); +extern int early_write_config_byte(struct pci_channel *hose, int top_bus, + int bus, int devfn, int offset, u8 value); +extern int early_write_config_word(struct pci_channel *hose, int top_bus, + int bus, int devfn, int offset, u16 value); +extern int early_write_config_dword(struct pci_channel *hose, int top_bus, + int bus, int devfn, int offset, u32 value); extern void pcibios_enable_timers(struct pci_channel *hose); extern unsigned int pcibios_handle_status_errors(unsigned long addr, unsigned int status, struct pci_channel *hose); -- cgit v0.10.2 From b8f46c5a34fa64fd456295388d18f50ae69d9f37 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Wed, 3 Feb 2010 11:53:14 +0800 Subject: perf tools: Use O_LARGEFILE to open perf data file Open perf data file with O_LARGEFILE flag since its size is easily larger that 2G. For example: # rm -rf perf.data # ./perf kmem record sleep 300 [ perf record: Woken up 0 times to write data ] [ perf record: Captured and wrote 3142.147 MB perf.data (~137282513 samples) ] # ll -h perf.data -rw------- 1 root root 3.1G ..... Signed-off-by: Xiao Guangrong Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul Mackerras Cc: Peter Zijlstra LKML-Reference: <4B68F32A.9040203@cn.fujitsu.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index eea5691..949167e 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -5,6 +5,9 @@ * (or a CPU, or a PID) into the perf.data output file - for * later analysis via perf report. */ +#define _LARGEFILE64_SOURCE +#define _FILE_OFFSET_BITS 64 + #include "builtin.h" #include "perf.h" @@ -451,7 +454,7 @@ static int __cmd_record(int argc, const char **argv) append_file = 0; } - flags = O_CREAT|O_RDWR; + flags = O_CREAT|O_RDWR|O_LARGEFILE; if (append_file) file_new = 0; else diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 2bb2bdb..ed3efd7 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -1,3 +1,6 @@ +#define _LARGEFILE64_SOURCE +#define _FILE_OFFSET_BITS 64 + #include #include #include @@ -382,7 +385,7 @@ static int perf_header__adds_write(struct perf_header *self, int fd) sec_size = sizeof(*feat_sec) * nr_sections; sec_start = self->data_offset + self->data_size; - lseek(fd, sec_start + sec_size, SEEK_SET); + lseek64(fd, sec_start + sec_size, SEEK_SET); if (perf_header__has_feat(self, HEADER_TRACE_INFO)) { struct perf_file_section *trace_sec; @@ -390,9 +393,9 @@ static int perf_header__adds_write(struct perf_header *self, int fd) trace_sec = &feat_sec[idx++]; /* Write trace info */ - trace_sec->offset = lseek(fd, 0, SEEK_CUR); + trace_sec->offset = lseek64(fd, 0, SEEK_CUR); read_tracing_data(fd, attrs, nr_counters); - trace_sec->size = lseek(fd, 0, SEEK_CUR) - trace_sec->offset; + trace_sec->size = lseek64(fd, 0, SEEK_CUR) - trace_sec->offset; } @@ -402,17 +405,18 @@ static int perf_header__adds_write(struct perf_header *self, int fd) buildid_sec = &feat_sec[idx++]; /* Write build-ids */ - buildid_sec->offset = lseek(fd, 0, SEEK_CUR); + buildid_sec->offset = lseek64(fd, 0, SEEK_CUR); err = dsos__write_buildid_table(fd); if (err < 0) { pr_debug("failed to write buildid table\n"); goto out_free; } - buildid_sec->size = lseek(fd, 0, SEEK_CUR) - buildid_sec->offset; + buildid_sec->size = lseek64(fd, 0, SEEK_CUR) - + buildid_sec->offset; dsos__cache_build_ids(); } - lseek(fd, sec_start, SEEK_SET); + lseek64(fd, sec_start, SEEK_SET); err = do_write(fd, feat_sec, sec_size); if (err < 0) pr_debug("failed to write feature section\n"); @@ -506,7 +510,7 @@ int perf_header__write(struct perf_header *self, int fd, bool at_exit) pr_debug("failed to write perf header\n"); return err; } - lseek(fd, self->data_offset + self->data_size, SEEK_SET); + lseek64(fd, self->data_offset + self->data_size, SEEK_SET); self->frozen = 1; return 0; @@ -560,7 +564,7 @@ int perf_header__process_sections(struct perf_header *self, int fd, sec_size = sizeof(*feat_sec) * nr_sections; - lseek(fd, self->data_offset + self->data_size, SEEK_SET); + lseek64(fd, self->data_offset + self->data_size, SEEK_SET); if (perf_header__getbuffer64(self, fd, feat_sec, sec_size)) goto out_free; @@ -634,7 +638,7 @@ static int perf_file_section__process(struct perf_file_section *self, struct perf_header *ph, int feat, int fd) { - if (lseek(fd, self->offset, SEEK_SET) < 0) { + if (lseek64(fd, self->offset, SEEK_SET) < 0) { pr_debug("Failed to lseek to %Ld offset for feature %d, " "continuing...\n", self->offset, feat); return 0; diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 8e7c189..cf91d09 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1,3 +1,6 @@ +#define _LARGEFILE64_SOURCE +#define _FILE_OFFSET_BITS 64 + #include #include @@ -12,7 +15,7 @@ static int perf_session__open(struct perf_session *self, bool force) { struct stat input_stat; - self->fd = open(self->filename, O_RDONLY); + self->fd = open(self->filename, O_RDONLY|O_LARGEFILE); if (self->fd < 0) { pr_err("failed to open file: %s", self->filename); if (!strcmp(self->filename, "perf.data")) diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c index 1744422..ca3c26d 100644 --- a/tools/perf/util/trace-event-read.c +++ b/tools/perf/util/trace-event-read.c @@ -83,7 +83,7 @@ static char *read_string(void) char *str = NULL; int size = 0; int i; - int r; + s64 r; for (;;) { r = read(input_fd, buf, BUFSIZ); @@ -117,7 +117,7 @@ static char *read_string(void) i++; /* move the file descriptor to the end of the string */ - r = lseek(input_fd, -(r - i), SEEK_CUR); + r = lseek64(input_fd, -(r - i), SEEK_CUR); if (r < 0) die("lseek"); -- cgit v0.10.2 From f54405db66fbec11679241daefd16fd8291a5762 Mon Sep 17 00:00:00 2001 From: Alex Neblett Date: Tue, 2 Feb 2010 21:16:03 -0800 Subject: HID: add support for Pixart Imaging Optical Touch Screen Added support for the Pixart Imaging Inc. Optical Touch Screen found in the MSI AE2220 and other new all in one computers to the Quanta Optical Touch dual-touch panel driver found in the latest git clone git://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git. Signed-off-by: Alex Neblett Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index 2dd9b28..282f4a1 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1343,6 +1343,7 @@ static const struct hid_device_id hid_blacklist[] = { { HID_USB_DEVICE(USB_VENDOR_ID_ORTEK, USB_DEVICE_ID_ORTEK_WKB2000) }, { HID_USB_DEVICE(USB_VENDOR_ID_PETALYNX, USB_DEVICE_ID_PETALYNX_MAXTER_REMOTE) }, { HID_USB_DEVICE(USB_VENDOR_ID_QUANTA, USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH) }, + { HID_USB_DEVICE(USB_VENDOR_ID_QUANTA, USB_DEVICE_ID_PIXART_IMAGING_INC_OPTICAL_TOUCH_SCREEN) }, { HID_USB_DEVICE(USB_VENDOR_ID_SAMSUNG, USB_DEVICE_ID_SAMSUNG_IR_REMOTE) }, { HID_USB_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_PS3_CONTROLLER) }, { HID_USB_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_VAIO_VGX_MOUSE) }, diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index 39ff98a..b29d9da 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -400,6 +400,7 @@ #define USB_VENDOR_ID_QUANTA 0x0408 #define USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH 0x3000 +#define USB_DEVICE_ID_PIXART_IMAGING_INC_OPTICAL_TOUCH_SCREEN 0x3001 #define USB_VENDOR_ID_SAMSUNG 0x0419 #define USB_DEVICE_ID_SAMSUNG_IR_REMOTE 0x0001 diff --git a/drivers/hid/hid-quanta.c b/drivers/hid/hid-quanta.c index 244d61c..9ac4928 100644 --- a/drivers/hid/hid-quanta.c +++ b/drivers/hid/hid-quanta.c @@ -224,6 +224,8 @@ static void quanta_remove(struct hid_device *hdev) static const struct hid_device_id quanta_devices[] = { { HID_USB_DEVICE(USB_VENDOR_ID_QUANTA, USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH) }, + { HID_USB_DEVICE(USB_VENDOR_ID_QUANTA, + USB_DEVICE_ID_PIXART_IMAGING_INC_OPTICAL_TOUCH_SCREEN) }, { } }; MODULE_DEVICE_TABLE(hid, quanta_devices); -- cgit v0.10.2 From d4bfa033ed84e0ae446eff445d107ffd5ee78df3 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Fri, 29 Jan 2010 15:03:36 +0100 Subject: HID: make raw reports possible for both feature and output reports In commit 2da31939a42 ("Bluetooth: Implement raw output support for HIDP layer"), support for Bluetooth hid_output_raw_report was added, but it pushes the data to the intr socket instead of the ctrl one. This has been fixed by 6bf8268f9a91f1 ("Bluetooth: Use the control channel for raw HID reports") Still, it is necessary to distinguish whether the report in question should be either FEATURE or OUTPUT. For this, we have to extend the generic HID API, so that hid_output_raw_report() callback provides means to specify this value so that it can be passed down to lower level hardware drivers (currently Bluetooth and USB). Based on original patch by Bastien Nocera Acked-by: Marcel Holtmann Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hidraw.c b/drivers/hid/hidraw.c index cdd1369..d044767 100644 --- a/drivers/hid/hidraw.c +++ b/drivers/hid/hidraw.c @@ -134,7 +134,7 @@ static ssize_t hidraw_write(struct file *file, const char __user *buffer, size_t goto out; } - ret = dev->hid_output_raw_report(dev, buf, count); + ret = dev->hid_output_raw_report(dev, buf, count, HID_OUTPUT_REPORT); out: kfree(buf); return ret; diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c index e2997a8..caa16c0 100644 --- a/drivers/hid/usbhid/hid-core.c +++ b/drivers/hid/usbhid/hid-core.c @@ -774,7 +774,8 @@ static int hid_alloc_buffers(struct usb_device *dev, struct hid_device *hid) return 0; } -static int usbhid_output_raw_report(struct hid_device *hid, __u8 *buf, size_t count) +static int usbhid_output_raw_report(struct hid_device *hid, __u8 *buf, size_t count, + unsigned char report_type) { struct usbhid_device *usbhid = hid->driver_data; struct usb_device *dev = hid_to_usb_dev(hid); @@ -785,7 +786,7 @@ static int usbhid_output_raw_report(struct hid_device *hid, __u8 *buf, size_t co ret = usb_control_msg(dev, usb_sndctrlpipe(dev, 0), HID_REQ_SET_REPORT, USB_DIR_OUT | USB_TYPE_CLASS | USB_RECIP_INTERFACE, - ((HID_OUTPUT_REPORT + 1) << 8) | *buf, + ((report_type + 1) << 8) | *buf, interface->desc.bInterfaceNumber, buf + 1, count - 1, USB_CTRL_SET_TIMEOUT); diff --git a/include/linux/hid.h b/include/linux/hid.h index 8709365..3661a62 100644 --- a/include/linux/hid.h +++ b/include/linux/hid.h @@ -501,7 +501,7 @@ struct hid_device { /* device report descriptor */ void (*hiddev_report_event) (struct hid_device *, struct hid_report *); /* handler for raw output data, used by hidraw */ - int (*hid_output_raw_report) (struct hid_device *, __u8 *, size_t); + int (*hid_output_raw_report) (struct hid_device *, __u8 *, size_t, unsigned char); /* debugging support via debugfs */ unsigned short debug; diff --git a/net/bluetooth/hidp/core.c b/net/bluetooth/hidp/core.c index 6cf526d..37ba153 100644 --- a/net/bluetooth/hidp/core.c +++ b/net/bluetooth/hidp/core.c @@ -313,10 +313,21 @@ static int hidp_send_report(struct hidp_session *session, struct hid_report *rep return hidp_queue_report(session, buf, rsize); } -static int hidp_output_raw_report(struct hid_device *hid, unsigned char *data, size_t count) +static int hidp_output_raw_report(struct hid_device *hid, unsigned char *data, size_t count, + unsigned char report_type) { - if (hidp_send_ctrl_message(hid->driver_data, - HIDP_TRANS_SET_REPORT | HIDP_DATA_RTYPE_FEATURE, + switch (report_type) { + case HID_FEATURE_REPORT: + report_type = HIDP_TRANS_SET_REPORT | HIDP_DATA_RTYPE_FEATURE; + break; + case HID_OUTPUT_REPORT: + report_type = HIDP_TRANS_DATA | HIDP_DATA_RTYPE_OUPUT; + break; + default: + return -EINVAL; + } + + if (hidp_send_ctrl_message(hid->driver_data, report_type, data, count)) return -ENOMEM; return count; -- cgit v0.10.2 From 46a709b900bfcf43244cd19cf3245c77484ec733 Mon Sep 17 00:00:00 2001 From: Bastien Nocera Date: Wed, 20 Jan 2010 12:00:53 +0000 Subject: HID: Implement Wacom quirk in the kernel The hid-wacom driver required user-space to poke at the tablet to make it send data about the cursor location. This patch makes it do the same thing but in the kernel. Signed-off-by: Bastien Nocera Acked-by: Marcel Holtmann Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-wacom.c b/drivers/hid/hid-wacom.c index 12dcda5..b8778db 100644 --- a/drivers/hid/hid-wacom.c +++ b/drivers/hid/hid-wacom.c @@ -156,7 +156,9 @@ static int wacom_probe(struct hid_device *hdev, struct hid_input *hidinput; struct input_dev *input; struct wacom_data *wdata; + char rep_data[2]; int ret; + int limit; wdata = kzalloc(sizeof(*wdata), GFP_KERNEL); if (wdata == NULL) { @@ -166,6 +168,7 @@ static int wacom_probe(struct hid_device *hdev, hid_set_drvdata(hdev, wdata); + /* Parse the HID report now */ ret = hid_parse(hdev); if (ret) { dev_err(&hdev->dev, "parse failed\n"); @@ -178,6 +181,30 @@ static int wacom_probe(struct hid_device *hdev, goto err_free; } + /* Set Wacom mode2 */ + rep_data[0] = 0x03; rep_data[1] = 0x00; + limit = 3; + do { + ret = hdev->hid_output_raw_report(hdev, rep_data, 2, + HID_FEATURE_REPORT); + } while (ret < 0 && limit-- > 0); + if (ret < 0) { + dev_err(&hdev->dev, "failed to poke device #1, %d\n", ret); + goto err_free; + } + + /* 0x06 - high reporting speed, 0x05 - low speed */ + rep_data[0] = 0x06; rep_data[1] = 0x00; + limit = 3; + do { + ret = hdev->hid_output_raw_report(hdev, rep_data, 2, + HID_FEATURE_REPORT); + } while (ret < 0 && limit-- > 0); + if (ret < 0) { + dev_err(&hdev->dev, "failed to poke device #2, %d\n", ret); + goto err_free; + } + hidinput = list_entry(hdev->inputs.next, struct hid_input, list); input = hidinput->input; -- cgit v0.10.2 From f9ce7c283c16538955d5d094101889792bcde109 Mon Sep 17 00:00:00 2001 From: Bastien Nocera Date: Wed, 20 Jan 2010 12:01:53 +0000 Subject: HID: Enable Sixaxis controller over Bluetooth Now that hid_output_raw_report works, port the PS3 Sixaxis Bluetooth quirk from user-space, into kernel-space. Signed-off-by: Bastien Nocera Acked-by: Marcel Holtmann Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index eabe5f8..f7f80e1 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1340,6 +1340,7 @@ static const struct hid_device_id hid_blacklist[] = { { HID_USB_DEVICE(USB_VENDOR_ID_PETALYNX, USB_DEVICE_ID_PETALYNX_MAXTER_REMOTE) }, { HID_USB_DEVICE(USB_VENDOR_ID_SAMSUNG, USB_DEVICE_ID_SAMSUNG_IR_REMOTE) }, { HID_USB_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_PS3_CONTROLLER) }, + { HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_PS3_CONTROLLER) }, { HID_USB_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_VAIO_VGX_MOUSE) }, { HID_USB_DEVICE(USB_VENDOR_ID_SUNPLUS, USB_DEVICE_ID_SUNPLUS_WDESKTOP) }, { HID_USB_DEVICE(USB_VENDOR_ID_THRUSTMASTER, 0xb300) }, diff --git a/drivers/hid/hid-sony.c b/drivers/hid/hid-sony.c index 4e84502..6ced140 100644 --- a/drivers/hid/hid-sony.c +++ b/drivers/hid/hid-sony.c @@ -48,7 +48,7 @@ static void sony_report_fixup(struct hid_device *hdev, __u8 *rdesc, * to "operational". Without this, the ps3 controller will not report any * events. */ -static int sony_set_operational(struct hid_device *hdev) +static int sony_set_operational_usb(struct hid_device *hdev) { struct usb_interface *intf = to_usb_interface(hdev->dev.parent); struct usb_device *dev = interface_to_usbdev(intf); @@ -73,6 +73,12 @@ static int sony_set_operational(struct hid_device *hdev) return ret; } +static int sony_set_operational_bt(struct hid_device *hdev) +{ + unsigned char buf[] = { 0x53, 0xf4, 0x42, 0x03, 0x00, 0x00 }; + return hdev->hid_output_raw_report(hdev, buf, sizeof(buf), HID_FEATURE_REPORT); +} + static int sony_probe(struct hid_device *hdev, const struct hid_device_id *id) { int ret; @@ -101,7 +107,17 @@ static int sony_probe(struct hid_device *hdev, const struct hid_device_id *id) goto err_free; } - ret = sony_set_operational(hdev); + switch (hdev->bus) { + case BUS_USB: + ret = sony_set_operational_usb(hdev); + break; + case BUS_BLUETOOTH: + ret = sony_set_operational_bt(hdev); + break; + default: + ret = 0; + } + if (ret < 0) goto err_stop; @@ -121,6 +137,7 @@ static void sony_remove(struct hid_device *hdev) static const struct hid_device_id sony_devices[] = { { HID_USB_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_PS3_CONTROLLER) }, + { HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_PS3_CONTROLLER) }, { HID_USB_DEVICE(USB_VENDOR_ID_SONY, USB_DEVICE_ID_SONY_VAIO_VGX_MOUSE), .driver_data = VAIO_RDESC_CONSTANT }, { } -- cgit v0.10.2 From 342f31e84eec9002b75f6fcdec6bd932ac77a390 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Wed, 3 Feb 2010 15:52:31 +0100 Subject: HID: make Wacom modesetting failures non-fatal With Wacom tablet mode-setting moved from userspace into kernel, we don't have to consider failures of device queries through the _raw callback as hard failure, as the driver can safely continue anyway. This is consistent with the current USB driver in wacom_sys.c Reported-by: Ping Cheng Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-wacom.c b/drivers/hid/hid-wacom.c index b8778db..8d3b46f 100644 --- a/drivers/hid/hid-wacom.c +++ b/drivers/hid/hid-wacom.c @@ -181,6 +181,11 @@ static int wacom_probe(struct hid_device *hdev, goto err_free; } + /* + * Note that if the raw queries fail, it's not a hard failure and it + * is safe to continue + */ + /* Set Wacom mode2 */ rep_data[0] = 0x03; rep_data[1] = 0x00; limit = 3; @@ -188,10 +193,8 @@ static int wacom_probe(struct hid_device *hdev, ret = hdev->hid_output_raw_report(hdev, rep_data, 2, HID_FEATURE_REPORT); } while (ret < 0 && limit-- > 0); - if (ret < 0) { - dev_err(&hdev->dev, "failed to poke device #1, %d\n", ret); - goto err_free; - } + if (ret < 0) + dev_warn(&hdev->dev, "failed to poke device #1, %d\n", ret); /* 0x06 - high reporting speed, 0x05 - low speed */ rep_data[0] = 0x06; rep_data[1] = 0x00; @@ -200,10 +203,8 @@ static int wacom_probe(struct hid_device *hdev, ret = hdev->hid_output_raw_report(hdev, rep_data, 2, HID_FEATURE_REPORT); } while (ret < 0 && limit-- > 0); - if (ret < 0) { - dev_err(&hdev->dev, "failed to poke device #2, %d\n", ret); - goto err_free; - } + if (ret < 0) + dev_warn(&hdev->dev, "failed to poke device #2, %d\n", ret); hidinput = list_entry(hdev->inputs.next, struct hid_input, list); input = hidinput->input; -- cgit v0.10.2 From 2dbf209d7a7ab94266b936bd2da6a4026c279992 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Wed, 3 Feb 2010 16:11:12 +0100 Subject: HID: make full-fledged hid-bus drivers properly selectable For historical reasons, we don't have most of the in-tree drivers residing on hid-bus properly selectable in kernel configuration unless CONFIG_EMBEDDED is set. This has been introduced on Linus' request from 14 Oct === As to the Kconfig options - do they really add so much space that you need to ask for the quirks? You didn't use to. Can you make the questions depend on EMBEDDED, or at least on the HID_COMPAT thing or whatever? === This still makes perfect sense for small and tiny drivers, which just fix report descriptors, fix up HID->input mappings that slightly violates HUT standard, send one extra packet to the device that is needed before it becomes functional, etc. Since then, we have been gathering more and more HID-bus drivers, which are full-fledged drivers. For these, the size argument becomes more valid. Plus the devices are much more special than "just violates HID specification in this one or two tiny unimportant points". Therefore I am marking such drivers as properly selectable no matter the setting of CONFIG_EMBEDDED, while keeping all the small and tiny ones compiled by default. Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index 139668d..617ed7d 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -212,9 +212,8 @@ config HID_MONTEREY Support for Monterey Genius KB29E. config HID_NTRIG - tristate "NTrig" if EMBEDDED + tristate "NTrig" depends on USB_HID - default !EMBEDDED ---help--- Support for N-Trig touch screen. @@ -338,9 +337,8 @@ config THRUSTMASTER_FF Rumble Force or Force Feedback Wheel. config HID_WACOM - tristate "Wacom Bluetooth devices support" if EMBEDDED + tristate "Wacom Bluetooth devices support" depends on BT_HIDP - default !EMBEDDED ---help--- Support for Wacom Graphire Bluetooth tablet. -- cgit v0.10.2 From 3cebbb81c7e75321e25cc586d07a25a3d98278fc Mon Sep 17 00:00:00 2001 From: Michal Marek Date: Wed, 3 Feb 2010 17:20:14 +0100 Subject: kconfig: Simplify LSMOD= handling Signed-off-by: Michal Marek LKML-Reference: <20100203162014.GA10956@sepie.suse.cz> Signed-off-by: Steven Rostedt diff --git a/scripts/kconfig/Makefile b/scripts/kconfig/Makefile index 85b9065..186c466 100644 --- a/scripts/kconfig/Makefile +++ b/scripts/kconfig/Makefile @@ -32,11 +32,10 @@ silentoldconfig: $(obj)/conf # if no path is given, then use src directory to find file ifdef LSMOD -LSMOD_F = $(shell if [ `basename $(LSMOD)` == $(LSMOD) ]; then \ - echo $(objtree)/$(LSMOD); \ - else \ - echo $(LSMOD); \ - fi) +LSMOD_F := $(LSMOD) +ifeq ($(findstring /,$(LSMOD)),) + LSMOD_F := $(objtree)/$(LSMOD) +endif endif localmodconfig: $(obj)/streamline_config.pl $(obj)/conf -- cgit v0.10.2 From 07fec73625dc0db6f9aed68019918208a2ca53f5 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 9 Feb 2010 11:43:49 +1100 Subject: xfs: log changed inodes instead of writing them synchronously When an inode has already be flushed delayed write, xfs_inode_clean() returns true and hence xfs_fs_write_inode() can return on a synchronous inode write without having written the inode. Currently these sycnhronous writes only come sync(1), unmount, a sycnhronous NFS export and cachefiles so should be relatively rare and out of common performance paths. Realistically, a synchronous inode write is not necessary here; we can avoid writing the inode by logging any non-transactional changes that are pending. This needs to be done with synchronous transactions, but it avoids seeking between the log and inode clusters as we do now. We don't force the log if the inode is pinned, though, so this differs from the fsync case. For normal sys_sync and unmount behaviour this is fine because we do a synchronous log force in xfs_sync_data which is called from the ->sync_fs code. It does however break the NFS synchronous export guarantees for now, but work is under way to fix this at a higher level or for the higher level to provide an additional flag in the writeback control to tell us that a log force is needed. Portions of this patch are based on work from Dave Chinner. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Reviewed-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index 3b5b46b..25ea240 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -1021,12 +1021,45 @@ xfs_fs_dirty_inode( XFS_I(inode)->i_update_core = 1; } -/* - * Attempt to flush the inode, this will actually fail - * if the inode is pinned, but we dirty the inode again - * at the point when it is unpinned after a log write, - * since this is when the inode itself becomes flushable. - */ +STATIC int +xfs_log_inode( + struct xfs_inode *ip) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_trans *tp; + int error; + + xfs_iunlock(ip, XFS_ILOCK_SHARED); + tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS); + error = xfs_trans_reserve(tp, 0, XFS_FSYNC_TS_LOG_RES(mp), 0, 0, 0); + + if (error) { + xfs_trans_cancel(tp, 0); + /* we need to return with the lock hold shared */ + xfs_ilock(ip, XFS_ILOCK_SHARED); + return error; + } + + xfs_ilock(ip, XFS_ILOCK_EXCL); + + /* + * Note - it's possible that we might have pushed ourselves out of the + * way during trans_reserve which would flush the inode. But there's + * no guarantee that the inode buffer has actually gone out yet (it's + * delwri). Plus the buffer could be pinned anyway if it's part of + * an inode in another recent transaction. So we play it safe and + * fire off the transaction anyway. + */ + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); + xfs_trans_ihold(tp, ip); + xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); + xfs_trans_set_sync(tp); + error = xfs_trans_commit(tp, 0); + xfs_ilock_demote(ip, XFS_ILOCK_EXCL); + + return error; +} + STATIC int xfs_fs_write_inode( struct inode *inode, @@ -1034,7 +1067,7 @@ xfs_fs_write_inode( { struct xfs_inode *ip = XFS_I(inode); struct xfs_mount *mp = ip->i_mount; - int error = 0; + int error = EAGAIN; xfs_itrace_entry(ip); @@ -1045,35 +1078,55 @@ xfs_fs_write_inode( error = xfs_wait_on_pages(ip, 0, -1); if (error) goto out; - } - - /* - * Bypass inodes which have already been cleaned by - * the inode flush clustering code inside xfs_iflush - */ - if (xfs_inode_clean(ip)) - goto out; - /* - * We make this non-blocking if the inode is contended, return - * EAGAIN to indicate to the caller that they did not succeed. - * This prevents the flush path from blocking on inodes inside - * another operation right now, they get caught later by xfs_sync. - */ - if (sync) { + /* + * Make sure the inode has hit stable storage. By using the + * log and the fsync transactions we reduce the IOs we have + * to do here from two (log and inode) to just the log. + * + * Note: We still need to do a delwri write of the inode after + * this to flush it to the backing buffer so that bulkstat + * works properly if this is the first time the inode has been + * written. Because we hold the ilock atomically over the + * transaction commit and the inode flush we are guaranteed + * that the inode is not pinned when it returns. If the flush + * lock is already held, then the inode has already been + * flushed once and we don't need to flush it again. Hence + * the code will only flush the inode if it isn't already + * being flushed. + */ xfs_ilock(ip, XFS_ILOCK_SHARED); - xfs_iflock(ip); - - error = xfs_iflush(ip, SYNC_WAIT); + if (ip->i_update_core) { + error = xfs_log_inode(ip); + if (error) + goto out_unlock; + } } else { - error = EAGAIN; + /* + * We make this non-blocking if the inode is contended, return + * EAGAIN to indicate to the caller that they did not succeed. + * This prevents the flush path from blocking on inodes inside + * another operation right now, they get caught later by xfs_sync. + */ if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) goto out; - if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip)) - goto out_unlock; + } + + if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip)) + goto out_unlock; - error = xfs_iflush(ip, 0); + /* + * Now we have the flush lock and the inode is not pinned, we can check + * if the inode is really clean as we know that there are no pending + * transaction completions, it is not waiting on the delayed write + * queue and there is no IO in progress. + */ + if (xfs_inode_clean(ip)) { + xfs_ifunlock(ip); + error = 0; + goto out_unlock; } + error = xfs_iflush(ip, 0); out_unlock: xfs_iunlock(ip, XFS_ILOCK_SHARED); -- cgit v0.10.2 From 5322892d867e186c6b4c5fff5c99ea4863696a60 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 4 Feb 2010 10:09:14 +1100 Subject: xfs: kill xfs_bawrite There are no more users of this function left in the XFS code now that we've switched everything to delayed write flushing. Remove it. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index 4556a4c..d50df3a 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -1078,25 +1078,6 @@ xfs_bwrite( return error; } -int -xfs_bawrite( - void *mp, - struct xfs_buf *bp) -{ - trace_xfs_buf_bawrite(bp, _RET_IP_); - - ASSERT(bp->b_bn != XFS_BUF_DADDR_NULL); - - xfs_buf_delwri_dequeue(bp); - - bp->b_flags &= ~(XBF_READ | XBF_DELWRI | XBF_READ_AHEAD); - bp->b_flags |= (XBF_WRITE | XBF_ASYNC | _XBF_RUN_QUEUES); - - bp->b_mount = mp; - bp->b_strat = xfs_bdstrat_cb; - return xfs_bdstrat_cb(bp); -} - void xfs_bdwrite( void *mp, diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h index be45e8c..386e736 100644 --- a/fs/xfs/linux-2.6/xfs_buf.h +++ b/fs/xfs/linux-2.6/xfs_buf.h @@ -233,7 +233,6 @@ extern void xfs_buf_unlock(xfs_buf_t *); /* Buffer Read and Write Routines */ extern int xfs_bwrite(struct xfs_mount *mp, struct xfs_buf *bp); -extern int xfs_bawrite(void *mp, xfs_buf_t *bp); extern void xfs_bdwrite(void *mp, xfs_buf_t *bp); extern void xfsbdstrat(struct xfs_mount *, struct xfs_buf *); -- cgit v0.10.2 From e054f1647162d7098a9ff619405a72bd7c417213 Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Thu, 4 Feb 2010 11:39:13 +1100 Subject: crypto: geode-aes - Fix cip/blk confusion a crypto_cipher cip member was set where a crypto_cipher blk members should have been. Signed-off-by: Roel Kluin Signed-off-by: Herbert Xu diff --git a/drivers/crypto/geode-aes.c b/drivers/crypto/geode-aes.c index 03e71b1..c7a5a43 100644 --- a/drivers/crypto/geode-aes.c +++ b/drivers/crypto/geode-aes.c @@ -141,7 +141,7 @@ static int geode_setkey_cip(struct crypto_tfm *tfm, const u8 *key, ret = crypto_cipher_setkey(op->fallback.cip, key, len); if (ret) { tfm->crt_flags &= ~CRYPTO_TFM_RES_MASK; - tfm->crt_flags |= (op->fallback.blk->base.crt_flags & CRYPTO_TFM_RES_MASK); + tfm->crt_flags |= (op->fallback.cip->base.crt_flags & CRYPTO_TFM_RES_MASK); } return ret; } -- cgit v0.10.2 From 5a1436beec5744029f3ac90b6fe71a698dcd6155 Mon Sep 17 00:00:00 2001 From: Steffen Klassert Date: Thu, 4 Feb 2010 11:40:17 +1100 Subject: crypto: pcrypt - call the complete function on error This fixes three forgotten calls to the complete function in the error case. Signed-off-by: Steffen Klassert Signed-off-by: Herbert Xu diff --git a/crypto/pcrypt.c b/crypto/pcrypt.c index b9527d0..8020124 100644 --- a/crypto/pcrypt.c +++ b/crypto/pcrypt.c @@ -114,7 +114,7 @@ static void pcrypt_aead_enc(struct padata_priv *padata) padata->info = crypto_aead_encrypt(req); - if (padata->info) + if (padata->info == -EINPROGRESS) return; padata_do_serial(padata); @@ -158,7 +158,7 @@ static void pcrypt_aead_dec(struct padata_priv *padata) padata->info = crypto_aead_decrypt(req); - if (padata->info) + if (padata->info == -EINPROGRESS) return; padata_do_serial(padata); @@ -202,7 +202,7 @@ static void pcrypt_aead_givenc(struct padata_priv *padata) padata->info = crypto_aead_givencrypt(req); - if (padata->info) + if (padata->info == -EINPROGRESS) return; padata_do_serial(padata); -- cgit v0.10.2 From efec959f63de850fbd2442189f7dfc9c38efe251 Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Thu, 4 Feb 2010 14:33:54 +1100 Subject: powerpc/pseries: Pass more accurate number of supported cores to firmware Updated variant of a patch by Joel Schopp. The field containing the number of supported cores which we pass to firmware via the ibm,client-architecture call was set by a previous patch statically as high as is possible (NR_CPUS). However, that value isn't quite right for a system that supports multiple threads per core, thus permitting the firmware to assign more cores to a Linux partition than it can really cope with. This patch improves it by using the device-tree to determine the number of threads supported by the processors in order to adjust the value passed to firmware. Signed-off-by: Joel Schopp Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c index 59d5bd1..5f306c4 100644 --- a/arch/powerpc/kernel/prom_init.c +++ b/arch/powerpc/kernel/prom_init.c @@ -710,7 +710,12 @@ static unsigned char ibm_architecture_vec[] = { 0, 0, 0, - W(NR_CPUS), /* number of cores supported*/ + /* WARNING: The offset of the "number of cores" field below + * must match by the macro below. Update the definition if + * the structure layout changes. + */ +#define IBM_ARCH_VEC_NRCORES_OFFSET 100 + W(NR_CPUS), /* number of cores supported */ /* option vector 6: IBM PAPR hints */ 4 - 2, /* length */ @@ -807,13 +812,70 @@ static struct fake_elf { } }; +static int __init prom_count_smt_threads(void) +{ + phandle node; + char type[64]; + unsigned int plen; + + /* Pick up th first CPU node we can find */ + for (node = 0; prom_next_node(&node); ) { + type[0] = 0; + prom_getprop(node, "device_type", type, sizeof(type)); + + if (strcmp(type, RELOC("cpu"))) + continue; + /* + * There is an entry for each smt thread, each entry being + * 4 bytes long. All cpus should have the same number of + * smt threads, so return after finding the first. + */ + plen = prom_getproplen(node, "ibm,ppc-interrupt-server#s"); + if (plen == PROM_ERROR) + break; + plen >>= 2; + prom_debug("Found 0x%x smt threads per core\n", (unsigned long)plen); + + /* Sanity check */ + if (plen < 1 || plen > 64) { + prom_printf("Threads per core 0x%x out of bounds, assuming 1\n", + (unsigned long)plen); + return 1; + } + return plen; + } + prom_debug("No threads found, assuming 1 per core\n"); + + return 1; + +} + + static void __init prom_send_capabilities(void) { ihandle elfloader, root; prom_arg_t ret; + u32 *cores; root = call_prom("open", 1, 1, ADDR("/")); if (root != 0) { + /* We need to tell the FW about the number of cores we support. + * + * To do that, we count the number of threads on the first core + * (we assume this is the same for all cores) and use it to + * divide NR_CPUS. + */ + cores = (u32 *)PTRRELOC(&ibm_architecture_vec[IBM_ARCH_VEC_NRCORES_OFFSET]); + if (*cores != NR_CPUS) { + prom_printf("WARNING ! " + "ibm_architecture_vec structure inconsistent: 0x%x !\n", + *cores); + } else { + *cores = NR_CPUS / prom_count_smt_threads(); + prom_printf("Max number of cores passed to firmware: 0x%x\n", + (unsigned long)*cores); + } + /* try calling the ibm,client-architecture-support method */ prom_printf("Calling ibm,client-architecture-support..."); if (call_prom_ret("call-method", 3, 2, &ret, -- cgit v0.10.2 From f266d7f5f89652a68e21e9882c44ee9104ad8d61 Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Wed, 3 Feb 2010 21:21:32 +0200 Subject: x86_64: Print modules like i386 does Print modules list during kernel BUG. Signed-off-by: Alexey Dobriyan Cc: Arjan van de Ven Cc: Linus Torvalds Cc: Andrew Morton LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index 0ad9597..907a90e 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -291,6 +291,7 @@ void show_registers(struct pt_regs *regs) sp = regs->sp; printk("CPU %d ", cpu); + print_modules(); __show_regs(regs, 1); printk("Process %s (pid: %d, threadinfo %p, task %p)\n", cur->comm, cur->pid, task_thread_info(cur), cur); -- cgit v0.10.2 From 9de89fe7c577847877ae00ea1aa6315559b10243 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 3 Feb 2010 16:52:00 -0200 Subject: perf symbols: Remove perf_session usage in symbols layer MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit I noticed while writing the first test in 'perf regtest' that to just test the symbol handling routines one needs to create a perf session, that is a layer centered on a perf.data file, events, etc, so I untied these layers. This reduces the complexity for the users as the number of parameters to most of the symbols and session APIs now was reduced while not adding more state to all the map instances by only having data that is needed to split the kernel (kallsyms and ELF symtab sections) maps and do vmlinux relocation on the main kernel map. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1265223128-11786-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 5d5dc6b..924a951 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -369,7 +369,7 @@ static void __print_result(struct rb_root *root, struct perf_session *session, if (is_caller) { addr = data->call_site; if (!raw_ip) - sym = map_groups__find_function(&session->kmaps, session, addr, NULL); + sym = map_groups__find_function(&session->kmaps, addr, NULL); } else addr = data->ptr; diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index 34f2acb..4fa73ec 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -122,8 +122,7 @@ static int opt_del_probe_event(const struct option *opt __used, static void evaluate_probe_point(struct probe_point *pp) { struct symbol *sym; - sym = map__find_symbol_by_name(session.kmap, pp->function, - session.psession, NULL); + sym = map__find_symbol_by_name(session.kmap, pp->function, NULL); if (!sym) die("Kernel symbol \'%s\' not found - probe not added.", pp->function); @@ -132,7 +131,7 @@ static void evaluate_probe_point(struct probe_point *pp) #ifndef NO_LIBDWARF static int open_vmlinux(void) { - if (map__load(session.kmap, session.psession, NULL) < 0) { + if (map__load(session.kmap, NULL) < 0) { pr_debug("Failed to load kernel map.\n"); return -EINVAL; } diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index bbaee61..c3831f6 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -374,9 +374,7 @@ int event__process_mmap(event_t *self, struct perf_session *session) goto out_problem; kernel->kernel = 1; - if (__map_groups__create_kernel_maps(&session->kmaps, - session->vmlinux_maps, - kernel) < 0) + if (__perf_session__create_kernel_maps(session, kernel) < 0) goto out_problem; session->vmlinux_maps[MAP__FUNCTION]->start = self->mmap.start; @@ -476,7 +474,7 @@ void thread__find_addr_location(struct thread *self, { thread__find_addr_map(self, session, cpumode, type, addr, al); if (al->map != NULL) - al->sym = map__find_symbol(al->map, session, al->addr, filter); + al->sym = map__find_symbol(al->map, al->addr, filter); else al->sym = NULL; } diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index c4d55a0..36ff0bf 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -104,8 +104,7 @@ void map__fixup_end(struct map *self) #define DSO__DELETED "(deleted)" -int map__load(struct map *self, struct perf_session *session, - symbol_filter_t filter) +int map__load(struct map *self, symbol_filter_t filter) { const char *name = self->dso->long_name; int nr; @@ -113,7 +112,7 @@ int map__load(struct map *self, struct perf_session *session, if (dso__loaded(self->dso, self->type)) return 0; - nr = dso__load(self->dso, self, session, filter); + nr = dso__load(self->dso, self, filter); if (nr < 0) { if (self->dso->has_build_id) { char sbuild_id[BUILD_ID_SIZE * 2 + 1]; @@ -144,24 +143,29 @@ int map__load(struct map *self, struct perf_session *session, return -1; } + /* + * Only applies to the kernel, as its symtabs aren't relative like the + * module ones. + */ + if (self->dso->kernel) + map__reloc_vmlinux(self); return 0; } -struct symbol *map__find_symbol(struct map *self, struct perf_session *session, - u64 addr, symbol_filter_t filter) +struct symbol *map__find_symbol(struct map *self, u64 addr, + symbol_filter_t filter) { - if (map__load(self, session, filter) < 0) + if (map__load(self, filter) < 0) return NULL; return dso__find_symbol(self->dso, self->type, addr); } struct symbol *map__find_symbol_by_name(struct map *self, const char *name, - struct perf_session *session, symbol_filter_t filter) { - if (map__load(self, session, filter) < 0) + if (map__load(self, filter) < 0) return NULL; if (!dso__sorted_by_name(self->dso, self->type)) diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h index 72f0b6a..de04839 100644 --- a/tools/perf/util/map.h +++ b/tools/perf/util/map.h @@ -14,6 +14,8 @@ enum map_type { #define MAP__NR_TYPES (MAP__VARIABLE + 1) struct dso; +struct ref_reloc_sym; +struct map_groups; struct map { union { @@ -29,6 +31,16 @@ struct map { struct dso *dso; }; +struct kmap { + struct ref_reloc_sym *ref_reloc_sym; + struct map_groups *kmaps; +}; + +static inline struct kmap *map__kmap(struct map *self) +{ + return (struct kmap *)(self + 1); +} + static inline u64 map__map_ip(struct map *map, u64 ip) { return ip - map->start + map->pgoff; @@ -58,16 +70,14 @@ struct map *map__clone(struct map *self); int map__overlap(struct map *l, struct map *r); size_t map__fprintf(struct map *self, FILE *fp); -struct perf_session; - -int map__load(struct map *self, struct perf_session *session, - symbol_filter_t filter); -struct symbol *map__find_symbol(struct map *self, struct perf_session *session, +int map__load(struct map *self, symbol_filter_t filter); +struct symbol *map__find_symbol(struct map *self, u64 addr, symbol_filter_t filter); struct symbol *map__find_symbol_by_name(struct map *self, const char *name, - struct perf_session *session, symbol_filter_t filter); void map__fixup_start(struct map *self); void map__fixup_end(struct map *self); +void map__reloc_vmlinux(struct map *self); + #endif /* __PERF_MAP_H */ diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index cf91d09..aa8a031 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -53,6 +53,11 @@ out_close: return -1; } +static inline int perf_session__create_kernel_maps(struct perf_session *self) +{ + return map_groups__create_kernel_maps(&self->kmaps, self->vmlinux_maps); +} + struct perf_session *perf_session__new(const char *filename, int mode, bool force) { size_t len = filename ? strlen(filename) + 1 : 0; @@ -507,6 +512,7 @@ int perf_session__set_kallsyms_ref_reloc_sym(struct perf_session *self, u64 addr) { char *bracket; + enum map_type i; self->ref_reloc_sym.name = strdup(symbol_name); if (self->ref_reloc_sym.name == NULL) @@ -517,6 +523,12 @@ int perf_session__set_kallsyms_ref_reloc_sym(struct perf_session *self, *bracket = '\0'; self->ref_reloc_sym.addr = addr; + + for (i = 0; i < MAP__NR_TYPES; ++i) { + struct kmap *kmap = map__kmap(self->vmlinux_maps[i]); + kmap->ref_reloc_sym = &self->ref_reloc_sym; + } + return 0; } @@ -530,20 +542,21 @@ static u64 map__reloc_unmap_ip(struct map *map, u64 ip) return ip - (s64)map->pgoff; } -void perf_session__reloc_vmlinux_maps(struct perf_session *self, - u64 unrelocated_addr) +void map__reloc_vmlinux(struct map *self) { - enum map_type type; - s64 reloc = unrelocated_addr - self->ref_reloc_sym.addr; + struct kmap *kmap = map__kmap(self); + s64 reloc; - if (!reloc) + if (!kmap->ref_reloc_sym || !kmap->ref_reloc_sym->unrelocated_addr) return; - for (type = 0; type < MAP__NR_TYPES; ++type) { - struct map *map = self->vmlinux_maps[type]; + reloc = (kmap->ref_reloc_sym->unrelocated_addr - + kmap->ref_reloc_sym->addr); - map->map_ip = map__reloc_map_ip; - map->unmap_ip = map__reloc_unmap_ip; - map->pgoff = reloc; - } + if (!reloc) + return; + + self->map_ip = map__reloc_map_ip; + self->unmap_ip = map__reloc_unmap_ip; + self->pgoff = reloc; } diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 36d1a80..752d75a 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -3,13 +3,13 @@ #include "event.h" #include "header.h" +#include "symbol.h" #include "thread.h" #include #include "../../../include/linux/perf_event.h" struct ip_callchain; struct thread; -struct symbol; struct perf_session { struct perf_header header; @@ -24,10 +24,7 @@ struct perf_session { unsigned long unknown_events; struct rb_root hists; u64 sample_type; - struct { - const char *name; - u64 addr; - } ref_reloc_sym; + struct ref_reloc_sym ref_reloc_sym; int fd; int cwdlen; char *cwd; @@ -69,9 +66,20 @@ int perf_header__read_build_ids(struct perf_header *self, int input, int perf_session__set_kallsyms_ref_reloc_sym(struct perf_session *self, const char *symbol_name, u64 addr); -void perf_session__reloc_vmlinux_maps(struct perf_session *self, - u64 unrelocated_addr); void mem_bswap_64(void *src, int byte_size); +static inline int __perf_session__create_kernel_maps(struct perf_session *self, + struct dso *kernel) +{ + return __map_groups__create_kernel_maps(&self->kmaps, + self->vmlinux_maps, kernel); +} + +static inline struct map * + perf_session__new_module_map(struct perf_session *self, + u64 start, const char *filename) +{ + return map_groups__new_module(&self->kmaps, start, filename); +} #endif /* __PERF_SESSION_H */ diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index f9049d1..6138742 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1,6 +1,5 @@ #include "util.h" #include "../perf.h" -#include "session.h" #include "sort.h" #include "string.h" #include "symbol.h" @@ -34,7 +33,7 @@ enum dso_origin { static void dsos__add(struct list_head *head, struct dso *dso); static struct map *map__new2(u64 start, struct dso *dso, enum map_type type); static int dso__load_kernel_sym(struct dso *self, struct map *map, - struct perf_session *session, symbol_filter_t filter); + symbol_filter_t filter); static int vmlinux_path__nr_entries; static char **vmlinux_path; @@ -480,8 +479,9 @@ static int dso__load_all_kallsyms(struct dso *self, const char *filename, * the original ELF section names vmlinux have. */ static int dso__split_kallsyms(struct dso *self, struct map *map, - struct perf_session *session, symbol_filter_t filter) + symbol_filter_t filter) { + struct map_groups *kmaps = map__kmap(map)->kmaps; struct map *curr_map = map; struct symbol *pos; int count = 0; @@ -503,7 +503,7 @@ static int dso__split_kallsyms(struct dso *self, struct map *map, *module++ = '\0'; if (strcmp(curr_map->dso->short_name, module)) { - curr_map = map_groups__find_by_name(&session->kmaps, map->type, module); + curr_map = map_groups__find_by_name(kmaps, map->type, module); if (curr_map == NULL) { pr_debug("/proc/{kallsyms,modules} " "inconsistency while looking " @@ -538,7 +538,7 @@ static int dso__split_kallsyms(struct dso *self, struct map *map, } curr_map->map_ip = curr_map->unmap_ip = identity__map_ip; - map_groups__insert(&session->kmaps, curr_map); + map_groups__insert(kmaps, curr_map); ++kernel_range; } @@ -557,9 +557,8 @@ discard_symbol: rb_erase(&pos->rb_node, root); return count; } - -static int dso__load_kallsyms(struct dso *self, const char *filename, struct map *map, - struct perf_session *session, symbol_filter_t filter) +int dso__load_kallsyms(struct dso *self, const char *filename, + struct map *map, symbol_filter_t filter) { if (dso__load_all_kallsyms(self, filename, map) < 0) return -1; @@ -567,7 +566,7 @@ static int dso__load_kallsyms(struct dso *self, const char *filename, struct map symbols__fixup_end(&self->symbols[map->type]); self->origin = DSO__ORIG_KERNEL; - return dso__split_kallsyms(self, map, session, filter); + return dso__split_kallsyms(self, map, filter); } static int dso__load_perf_map(struct dso *self, struct map *map, @@ -893,10 +892,10 @@ static bool elf_sec__is_a(GElf_Shdr *self, Elf_Data *secstrs, enum map_type type } } -static int dso__load_sym(struct dso *self, struct map *map, - struct perf_session *session, const char *name, int fd, - symbol_filter_t filter, int kernel, int kmodule) +static int dso__load_sym(struct dso *self, struct map *map, const char *name, + int fd, symbol_filter_t filter, int kmodule) { + struct kmap *kmap = self->kernel ? map__kmap(map) : NULL; struct map *curr_map = map; struct dso *curr_dso = self; size_t dso_name_len = strlen(self->short_name); @@ -953,7 +952,7 @@ static int dso__load_sym(struct dso *self, struct map *map, nr_syms = shdr.sh_size / shdr.sh_entsize; memset(&sym, 0, sizeof(sym)); - if (!kernel) { + if (!self->kernel) { self->adjust_symbols = (ehdr.e_type == ET_EXEC || elf_section_by_name(elf, &ehdr, &shdr, ".gnu.prelink_undo", @@ -967,9 +966,9 @@ static int dso__load_sym(struct dso *self, struct map *map, int is_label = elf_sym__is_label(&sym); const char *section_name; - if (kernel && session->ref_reloc_sym.name != NULL && - strcmp(elf_name, session->ref_reloc_sym.name) == 0) - perf_session__reloc_vmlinux_maps(session, sym.st_value); + if (kmap && kmap->ref_reloc_sym && kmap->ref_reloc_sym->name && + strcmp(elf_name, kmap->ref_reloc_sym->name) == 0) + kmap->ref_reloc_sym->unrelocated_addr = sym.st_value; if (!is_label && !elf_sym__is_a(&sym, map->type)) continue; @@ -985,7 +984,7 @@ static int dso__load_sym(struct dso *self, struct map *map, section_name = elf_sec__name(&shdr, secstrs); - if (kernel || kmodule) { + if (self->kernel || kmodule) { char dso_name[PATH_MAX]; if (strcmp(section_name, @@ -1001,7 +1000,7 @@ static int dso__load_sym(struct dso *self, struct map *map, snprintf(dso_name, sizeof(dso_name), "%s%s", self->short_name, section_name); - curr_map = map_groups__find_by_name(&session->kmaps, map->type, dso_name); + curr_map = map_groups__find_by_name(kmap->kmaps, map->type, dso_name); if (curr_map == NULL) { u64 start = sym.st_value; @@ -1020,7 +1019,7 @@ static int dso__load_sym(struct dso *self, struct map *map, curr_map->map_ip = identity__map_ip; curr_map->unmap_ip = identity__map_ip; curr_dso->origin = DSO__ORIG_KERNEL; - map_groups__insert(&session->kmaps, curr_map); + map_groups__insert(kmap->kmaps, curr_map); dsos__add(&dsos__kernel, curr_dso); } else curr_dso = curr_map->dso; @@ -1236,8 +1235,7 @@ char dso__symtab_origin(const struct dso *self) return origin[self->origin]; } -int dso__load(struct dso *self, struct map *map, struct perf_session *session, - symbol_filter_t filter) +int dso__load(struct dso *self, struct map *map, symbol_filter_t filter) { int size = PATH_MAX; char *name; @@ -1249,7 +1247,7 @@ int dso__load(struct dso *self, struct map *map, struct perf_session *session, dso__set_loaded(self, map->type); if (self->kernel) - return dso__load_kernel_sym(self, map, session, filter); + return dso__load_kernel_sym(self, map, filter); name = malloc(size); if (!name) @@ -1320,7 +1318,7 @@ open_file: fd = open(name, O_RDONLY); } while (fd < 0); - ret = dso__load_sym(self, map, NULL, name, fd, filter, 0, 0); + ret = dso__load_sym(self, map, name, fd, filter, 0); close(fd); /* @@ -1376,7 +1374,7 @@ static int dso__kernel_module_get_build_id(struct dso *self) return 0; } -static int perf_session__set_modules_path_dir(struct perf_session *self, char *dirname) +static int map_groups__set_modules_path_dir(struct map_groups *self, char *dirname) { struct dirent *dent; DIR *dir = opendir(dirname); @@ -1396,7 +1394,7 @@ static int perf_session__set_modules_path_dir(struct perf_session *self, char *d snprintf(path, sizeof(path), "%s/%s", dirname, dent->d_name); - if (perf_session__set_modules_path_dir(self, path) < 0) + if (map_groups__set_modules_path_dir(self, path) < 0) goto failure; } else { char *dot = strrchr(dent->d_name, '.'), @@ -1410,7 +1408,7 @@ static int perf_session__set_modules_path_dir(struct perf_session *self, char *d (int)(dot - dent->d_name), dent->d_name); strxfrchar(dso_name, '-', '_'); - map = map_groups__find_by_name(&self->kmaps, MAP__FUNCTION, dso_name); + map = map_groups__find_by_name(self, MAP__FUNCTION, dso_name); if (map == NULL) continue; @@ -1431,7 +1429,7 @@ failure: return -1; } -static int perf_session__set_modules_path(struct perf_session *self) +static int map_groups__set_modules_path(struct map_groups *self) { struct utsname uts; char modules_path[PATH_MAX]; @@ -1442,7 +1440,7 @@ static int perf_session__set_modules_path(struct perf_session *self) snprintf(modules_path, sizeof(modules_path), "/lib/modules/%s/kernel", uts.release); - return perf_session__set_modules_path_dir(self, modules_path); + return map_groups__set_modules_path_dir(self, modules_path); } /* @@ -1452,8 +1450,8 @@ static int perf_session__set_modules_path(struct perf_session *self) */ static struct map *map__new2(u64 start, struct dso *dso, enum map_type type) { - struct map *self = malloc(sizeof(*self)); - + struct map *self = zalloc(sizeof(*self) + + (dso->kernel ? sizeof(struct kmap) : 0)); if (self != NULL) { /* * ->end will be filled after we load all the symbols @@ -1464,8 +1462,8 @@ static struct map *map__new2(u64 start, struct dso *dso, enum map_type type) return self; } -struct map *perf_session__new_module_map(struct perf_session *self, u64 start, - const char *filename) +struct map *map_groups__new_module(struct map_groups *self, u64 start, + const char *filename) { struct map *map; struct dso *dso = __dsos__findnew(&dsos__kernel, filename); @@ -1478,11 +1476,11 @@ struct map *perf_session__new_module_map(struct perf_session *self, u64 start, return NULL; dso->origin = DSO__ORIG_KMODULE; - map_groups__insert(&self->kmaps, map); + map_groups__insert(self, map); return map; } -static int perf_session__create_module_maps(struct perf_session *self) +static int map_groups__create_modules(struct map_groups *self) { char *line = NULL; size_t n; @@ -1520,7 +1518,7 @@ static int perf_session__create_module_maps(struct perf_session *self) *sep = '\0'; snprintf(name, sizeof(name), "[%s]", line); - map = perf_session__new_module_map(self, start, name); + map = map_groups__new_module(self, start, name); if (map == NULL) goto out_delete_line; dso__kernel_module_get_build_id(map->dso); @@ -1529,7 +1527,7 @@ static int perf_session__create_module_maps(struct perf_session *self) free(line); fclose(file); - return perf_session__set_modules_path(self); + return map_groups__set_modules_path(self); out_delete_line: free(line); @@ -1538,7 +1536,6 @@ out_failure: } static int dso__load_vmlinux(struct dso *self, struct map *map, - struct perf_session *session, const char *vmlinux, symbol_filter_t filter) { int err = -1, fd; @@ -1572,14 +1569,14 @@ static int dso__load_vmlinux(struct dso *self, struct map *map, return -1; dso__set_loaded(self, map->type); - err = dso__load_sym(self, map, session, vmlinux, fd, filter, 1, 0); + err = dso__load_sym(self, map, vmlinux, fd, filter, 0); close(fd); return err; } int dso__load_vmlinux_path(struct dso *self, struct map *map, - struct perf_session *session, symbol_filter_t filter) + symbol_filter_t filter) { int i, err = 0; @@ -1587,8 +1584,7 @@ int dso__load_vmlinux_path(struct dso *self, struct map *map, vmlinux_path__nr_entries); for (i = 0; i < vmlinux_path__nr_entries; ++i) { - err = dso__load_vmlinux(self, map, session, vmlinux_path[i], - filter); + err = dso__load_vmlinux(self, map, vmlinux_path[i], filter); if (err > 0) { pr_debug("Using %s for symbols\n", vmlinux_path[i]); dso__set_long_name(self, strdup(vmlinux_path[i])); @@ -1600,7 +1596,7 @@ int dso__load_vmlinux_path(struct dso *self, struct map *map, } static int dso__load_kernel_sym(struct dso *self, struct map *map, - struct perf_session *session, symbol_filter_t filter) + symbol_filter_t filter) { int err; const char *kallsyms_filename = NULL; @@ -1621,13 +1617,13 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, * match. */ if (symbol_conf.vmlinux_name != NULL) { - err = dso__load_vmlinux(self, map, session, + err = dso__load_vmlinux(self, map, symbol_conf.vmlinux_name, filter); goto out_try_fixup; } if (vmlinux_path != NULL) { - err = dso__load_vmlinux_path(self, map, session, filter); + err = dso__load_vmlinux_path(self, map, filter); if (err > 0) goto out_fixup; } @@ -1675,7 +1671,7 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, } do_kallsyms: - err = dso__load_kallsyms(self, kallsyms_filename, map, session, filter); + err = dso__load_kallsyms(self, kallsyms_filename, map, filter); free(kallsyms_allocated_filename); out_try_fixup: @@ -1812,30 +1808,23 @@ int __map_groups__create_kernel_maps(struct map_groups *self, enum map_type type; for (type = 0; type < MAP__NR_TYPES; ++type) { + struct kmap *kmap; + vmlinux_maps[type] = map__new2(0, kernel, type); if (vmlinux_maps[type] == NULL) return -1; vmlinux_maps[type]->map_ip = vmlinux_maps[type]->unmap_ip = identity__map_ip; + + kmap = map__kmap(vmlinux_maps[type]); + kmap->kmaps = self; map_groups__insert(self, vmlinux_maps[type]); } return 0; } -static int map_groups__create_kernel_maps(struct map_groups *self, - struct map *vmlinux_maps[MAP__NR_TYPES], - const char *vmlinux) -{ - struct dso *kernel = dsos__create_kernel(vmlinux); - - if (kernel == NULL) - return -1; - - return __map_groups__create_kernel_maps(self, vmlinux_maps, kernel); -} - static void vmlinux_path__exit(void) { while (--vmlinux_path__nr_entries >= 0) { @@ -1941,19 +1930,22 @@ out_free_comm_list: return -1; } -int perf_session__create_kernel_maps(struct perf_session *self) +int map_groups__create_kernel_maps(struct map_groups *self, + struct map *vmlinux_maps[MAP__NR_TYPES]) { - if (map_groups__create_kernel_maps(&self->kmaps, self->vmlinux_maps, - symbol_conf.vmlinux_name) < 0) + struct dso *kernel = dsos__create_kernel(symbol_conf.vmlinux_name); + + if (kernel == NULL) + return -1; + + if (__map_groups__create_kernel_maps(self, vmlinux_maps, kernel) < 0) return -1; - if (symbol_conf.use_modules && - perf_session__create_module_maps(self) < 0) - pr_debug("Failed to load list of modules for session %s, " - "continuing...\n", self->filename); + if (symbol_conf.use_modules && map_groups__create_modules(self) < 0) + return -1; /* * Now that we have all the maps created, just set the ->end of them: */ - map_groups__fixup_end(&self->kmaps); + map_groups__fixup_end(self); return 0; } diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 1243027..e6a59e5 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -80,6 +80,12 @@ static inline void *symbol__priv(struct symbol *self) return ((void *)self) - symbol_conf.priv_size; } +struct ref_reloc_sym { + const char *name; + u64 addr; + u64 unrelocated_addr; +}; + struct addr_location { struct thread *thread; struct map *map; @@ -126,12 +132,11 @@ static inline struct dso *dsos__findnew(const char *name) return __dsos__findnew(&dsos__user, name); } -struct perf_session; - -int dso__load(struct dso *self, struct map *map, struct perf_session *session, - symbol_filter_t filter); +int dso__load(struct dso *self, struct map *map, symbol_filter_t filter); int dso__load_vmlinux_path(struct dso *self, struct map *map, - struct perf_session *session, symbol_filter_t filter); + symbol_filter_t filter); +int dso__load_kallsyms(struct dso *self, const char *filename, struct map *map, + symbol_filter_t filter); void dsos__fprintf(FILE *fp); size_t dsos__fprintf_buildid(FILE *fp, bool with_hits); @@ -156,9 +161,5 @@ int kallsyms__parse(const char *filename, void *arg, int symbol__init(void); bool symbol_type__is_a(char symbol_type, enum map_type map_type); -int perf_session__create_kernel_maps(struct perf_session *self); - -struct map *perf_session__new_module_map(struct perf_session *self, u64 start, - const char *filename); extern struct dso *vdso; #endif /* __PERF_SYMBOL */ diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 4a08dcf..634b7f7 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -282,14 +282,13 @@ size_t perf_session__fprintf(struct perf_session *self, FILE *fp) } struct symbol *map_groups__find_symbol(struct map_groups *self, - struct perf_session *session, enum map_type type, u64 addr, symbol_filter_t filter) { struct map *map = map_groups__find(self, type, addr); if (map != NULL) - return map__find_symbol(map, session, map->map_ip(map, addr), filter); + return map__find_symbol(map, map->map_ip(map, addr), filter); return NULL; } diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index e35653c..56f317b 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -59,15 +59,14 @@ void thread__find_addr_location(struct thread *self, struct addr_location *al, symbol_filter_t filter); struct symbol *map_groups__find_symbol(struct map_groups *self, - struct perf_session *session, enum map_type type, u64 addr, symbol_filter_t filter); -static inline struct symbol * -map_groups__find_function(struct map_groups *self, struct perf_session *session, - u64 addr, symbol_filter_t filter) +static inline struct symbol *map_groups__find_function(struct map_groups *self, + u64 addr, + symbol_filter_t filter) { - return map_groups__find_symbol(self, session, MAP__FUNCTION, addr, filter); + return map_groups__find_symbol(self, MAP__FUNCTION, addr, filter); } struct map *map_groups__find_by_name(struct map_groups *self, @@ -76,4 +75,9 @@ struct map *map_groups__find_by_name(struct map_groups *self, int __map_groups__create_kernel_maps(struct map_groups *self, struct map *vmlinux_maps[MAP__NR_TYPES], struct dso *kernel); +int map_groups__create_kernel_maps(struct map_groups *self, + struct map *vmlinux_maps[MAP__NR_TYPES]); + +struct map *map_groups__new_module(struct map_groups *self, u64 start, + const char *filename); #endif /* __PERF_THREAD_H */ -- cgit v0.10.2 From 6275ce2d5f44ae4f8575c24724525cbb2a3a141b Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 3 Feb 2010 16:52:01 -0200 Subject: perf symbols: Fixup vsyscall maps MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit While debugging a problem reported by Pekka Enberg by printing the IP and all the maps for a thread when we don't find a map for an IP I noticed that dso__load_sym needs to fixup these extra maps it creates to hold symbols in different ELF sections than the main kernel one. Now we're back showing things like: [root@doppio linux-2.6-tip]# perf report | grep vsyscall 0.02% mutt [kernel.kallsyms].vsyscall_fn [.] vread_hpet 0.01% named [kernel.kallsyms].vsyscall_fn [.] vread_hpet 0.01% NetworkManager [kernel.kallsyms].vsyscall_fn [.] vread_hpet 0.01% gconfd-2 [kernel.kallsyms].vsyscall_0 [.] vgettimeofday 0.01% hald-addon-rfki [kernel.kallsyms].vsyscall_fn [.] vread_hpet 0.00% dbus-daemon [kernel.kallsyms].vsyscall_fn [.] vread_hpet [root@doppio linux-2.6-tip]# Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Pekka Enberg Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1265223128-11786-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 6138742..051d71b 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1011,7 +1011,7 @@ static int dso__load_sym(struct dso *self, struct map *map, const char *name, if (curr_dso == NULL) goto out_elf_end; curr_map = map__new2(start, curr_dso, - MAP__FUNCTION); + map->type); if (curr_map == NULL) { dso__delete(curr_dso); goto out_elf_end; @@ -1021,6 +1021,7 @@ static int dso__load_sym(struct dso *self, struct map *map, const char *name, curr_dso->origin = DSO__ORIG_KERNEL; map_groups__insert(kmap->kmaps, curr_map); dsos__add(&dsos__kernel, curr_dso); + dso__set_loaded(curr_dso, map->type); } else curr_dso = curr_map->dso; @@ -1058,8 +1059,16 @@ new_symbol: /* * For misannotated, zeroed, ASM function sizes. */ - if (nr > 0) + if (nr > 0) { symbols__fixup_end(&self->symbols[map->type]); + if (kmap) { + /* + * We need to fixup this here too because we create new + * maps here, for things like vsyscall sections. + */ + __map_groups__fixup_end(kmap->kmaps, map->type); + } + } err = nr; out_elf_end: elf_end(elf); -- cgit v0.10.2 From 8d92c02ab07602786eaa6d4e5b519395730b3fd3 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 3 Feb 2010 16:52:02 -0200 Subject: perf symbols: Ditch vdso global variable MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We can check using strcmp, most DSOs don't start with '[' so the test is cheap enough and we had to test it there anyway since when reading perf.data files we weren't calling the routine that created this global variable and thus weren't setting it as "loaded", which was causing a bogus: Failed to open [vdso], continuing without symbols Message as the first line of 'perf report'. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1265223128-11786-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 36ff0bf..f6626cc 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -68,8 +68,13 @@ struct map *map__new(struct mmap_event *event, enum map_type type, map__init(self, type, event->start, event->start + event->len, event->pgoff, dso); - if (self->dso == vdso || anon) + if (anon) { +set_identity: self->map_ip = self->unmap_ip = identity__map_ip; + } else if (strcmp(filename, "[vdso]") == 0) { + dso__set_loaded(dso, self->type); + goto set_identity; + } } return self; out_delete: diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 051d71b..e752837 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -53,11 +53,6 @@ bool dso__sorted_by_name(const struct dso *self, enum map_type type) return self->sorted_by_name & (1 << type); } -static void dso__set_loaded(struct dso *self, enum map_type type) -{ - self->loaded |= (1 << type); -} - static void dso__set_sorted_by_name(struct dso *self, enum map_type type) { self->sorted_by_name |= (1 << type); @@ -1697,7 +1692,6 @@ out_fixup: LIST_HEAD(dsos__user); LIST_HEAD(dsos__kernel); -struct dso *vdso; static void dsos__add(struct list_head *head, struct dso *dso) { @@ -1790,24 +1784,12 @@ static struct dso *dsos__create_kernel(const char *vmlinux) { struct dso *kernel = dso__new_kernel(vmlinux); - if (kernel == NULL) - return NULL; - - vdso = dso__new("[vdso]"); - if (vdso == NULL) - goto out_delete_kernel_dso; - dso__set_loaded(vdso, MAP__FUNCTION); - - dso__read_running_kernel_build_id(kernel); - - dsos__add(&dsos__kernel, kernel); - dsos__add(&dsos__user, vdso); + if (kernel != NULL) { + dso__read_running_kernel_build_id(kernel); + dsos__add(&dsos__kernel, kernel); + } return kernel; - -out_delete_kernel_dso: - dso__delete(kernel); - return NULL; } int __map_groups__create_kernel_maps(struct map_groups *self, diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index e6a59e5..e90568a 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -121,6 +121,11 @@ void dso__delete(struct dso *self); bool dso__loaded(const struct dso *self, enum map_type type); bool dso__sorted_by_name(const struct dso *self, enum map_type type); +static inline void dso__set_loaded(struct dso *self, enum map_type type) +{ + self->loaded |= (1 << type); +} + void dso__sort_by_name(struct dso *self, enum map_type type); extern struct list_head dsos__user, dsos__kernel; @@ -161,5 +166,4 @@ int kallsyms__parse(const char *filename, void *arg, int symbol__init(void); bool symbol_type__is_a(char symbol_type, enum map_type map_type); -extern struct dso *vdso; #endif /* __PERF_SYMBOL */ -- cgit v0.10.2 From 8ad94c6052649a8e32120b464eefa0ffd8f2f04f Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 3 Feb 2010 16:52:03 -0200 Subject: perf probe: Don't use a perf_session instance just to resolve symbols MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit With the recent modifications done to untie the session and symbol layers, 'perf probe' now can use just the symbols layer. Signed-off-by: Arnaldo Carvalho de Melo Acked-by: Masami Hiramatsu Cc: Frédéric Weisbecker Cc: Masami Hiramatsu Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index 4fa73ec..ad47bd4 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -41,7 +41,6 @@ #include "util/debugfs.h" #include "util/symbol.h" #include "util/thread.h" -#include "util/session.h" #include "util/parse-options.h" #include "util/parse-events.h" /* For debugfs_path */ #include "util/probe-finder.h" @@ -59,8 +58,8 @@ static struct { int nr_probe; struct probe_point probes[MAX_PROBES]; struct strlist *dellist; - struct perf_session *psession; - struct map *kmap; + struct map_groups kmap_groups; + struct map *kmaps[MAP__NR_TYPES]; struct line_range line_range; } session; @@ -122,7 +121,8 @@ static int opt_del_probe_event(const struct option *opt __used, static void evaluate_probe_point(struct probe_point *pp) { struct symbol *sym; - sym = map__find_symbol_by_name(session.kmap, pp->function, NULL); + sym = map__find_symbol_by_name(session.kmaps[MAP__FUNCTION], + pp->function, NULL); if (!sym) die("Kernel symbol \'%s\' not found - probe not added.", pp->function); @@ -131,12 +131,13 @@ static void evaluate_probe_point(struct probe_point *pp) #ifndef NO_LIBDWARF static int open_vmlinux(void) { - if (map__load(session.kmap, NULL) < 0) { + if (map__load(session.kmaps[MAP__FUNCTION], NULL) < 0) { pr_debug("Failed to load kernel map.\n"); return -EINVAL; } - pr_debug("Try to open %s\n", session.kmap->dso->long_name); - return open(session.kmap->dso->long_name, O_RDONLY); + pr_debug("Try to open %s\n", + session.kmaps[MAP__FUNCTION]->dso->long_name); + return open(session.kmaps[MAP__FUNCTION]->dso->long_name, O_RDONLY); } static int opt_show_lines(const struct option *opt __used, @@ -212,12 +213,11 @@ static void init_vmlinux(void) pr_debug("Use vmlinux: %s\n", symbol_conf.vmlinux_name); if (symbol__init() < 0) die("Failed to init symbol map."); - session.psession = perf_session__new(NULL, O_WRONLY, false); - if (session.psession == NULL) - die("Failed to init perf_session."); - session.kmap = session.psession->vmlinux_maps[MAP__FUNCTION]; - if (!session.kmap) - die("Could not find kernel map.\n"); + + map_groups__init(&session.kmap_groups); + if (map_groups__create_kernel_maps(&session.kmap_groups, + session.kmaps) < 0) + die("Failed to create kernel maps."); } int cmd_probe(int argc, const char **argv, const char *prefix __used) -- cgit v0.10.2 From 7b2567c1f57c059de29d3f2ca03aca84473865c8 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 3 Feb 2010 16:52:04 -0200 Subject: perf build-id: Move the routine to find DSOs with hits to the lib MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Because 'perf record' will have to find the build-ids in after we stop recording, so as to reduce even more the impact in the workload while we do the measurement. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1265223128-11786-5-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 4296930..3a5fb36 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -357,6 +357,7 @@ LIB_H += util/include/asm/uaccess.h LIB_H += perf.h LIB_H += util/cache.h LIB_H += util/callchain.h +LIB_H += util/build-id.h LIB_H += util/debug.h LIB_H += util/debugfs.h LIB_H += util/event.h @@ -390,6 +391,7 @@ LIB_H += util/probe-event.h LIB_OBJS += util/abspath.o LIB_OBJS += util/alias.o +LIB_OBJS += util/build-id.o LIB_OBJS += util/config.o LIB_OBJS += util/ctype.o LIB_OBJS += util/debugfs.o diff --git a/tools/perf/builtin-buildid-list.c b/tools/perf/builtin-buildid-list.c index 431f204..d0675c0 100644 --- a/tools/perf/builtin-buildid-list.c +++ b/tools/perf/builtin-buildid-list.c @@ -8,6 +8,7 @@ */ #include "builtin.h" #include "perf.h" +#include "util/build-id.h" #include "util/cache.h" #include "util/debug.h" #include "util/parse-options.h" @@ -33,34 +34,6 @@ static const struct option options[] = { OPT_END() }; -static int build_id_list__process_event(event_t *event, - struct perf_session *session) -{ - struct addr_location al; - u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; - struct thread *thread = perf_session__findnew(session, event->ip.pid); - - if (thread == NULL) { - pr_err("problem processing %d event, skipping it.\n", - event->header.type); - return -1; - } - - thread__find_addr_map(thread, session, cpumode, MAP__FUNCTION, - event->ip.ip, &al); - - if (al.map != NULL) - al.map->dso->hit = 1; - - return 0; -} - -static struct perf_event_ops build_id_list__event_ops = { - .sample = build_id_list__process_event, - .mmap = event__process_mmap, - .fork = event__process_task, -}; - static int __cmd_buildid_list(void) { int err = -1; @@ -71,7 +44,7 @@ static int __cmd_buildid_list(void) return -1; if (with_hits) - perf_session__process_events(session, &build_id_list__event_ops); + perf_session__process_events(session, &build_id__mark_dso_hit_ops); dsos__fprintf_buildid(stdout, with_hits); diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c new file mode 100644 index 0000000..04904b3 --- /dev/null +++ b/tools/perf/util/build-id.c @@ -0,0 +1,39 @@ +/* + * build-id.c + * + * build-id support + * + * Copyright (C) 2009, 2010 Red Hat Inc. + * Copyright (C) 2009, 2010 Arnaldo Carvalho de Melo + */ +#include "build-id.h" +#include "event.h" +#include "symbol.h" +#include + +static int build_id__mark_dso_hit(event_t *event, struct perf_session *session) +{ + struct addr_location al; + u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; + struct thread *thread = perf_session__findnew(session, event->ip.pid); + + if (thread == NULL) { + pr_err("problem processing %d event, skipping it.\n", + event->header.type); + return -1; + } + + thread__find_addr_map(thread, session, cpumode, MAP__FUNCTION, + event->ip.ip, &al); + + if (al.map != NULL) + al.map->dso->hit = 1; + + return 0; +} + +struct perf_event_ops build_id__mark_dso_hit_ops = { + .sample = build_id__mark_dso_hit, + .mmap = event__process_mmap, + .fork = event__process_task, +}; diff --git a/tools/perf/util/build-id.h b/tools/perf/util/build-id.h new file mode 100644 index 0000000..1d981d6 --- /dev/null +++ b/tools/perf/util/build-id.h @@ -0,0 +1,8 @@ +#ifndef PERF_BUILD_ID_H_ +#define PERF_BUILD_ID_H_ 1 + +#include "session.h" + +extern struct perf_event_ops build_id__mark_dso_hit_ops; + +#endif -- cgit v0.10.2 From 6122e4e4f5d0913e319ef8a4dc60a47afe4abc0a Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 3 Feb 2010 16:52:05 -0200 Subject: perf record: Stop intercepting events, use postprocessing to get build-ids MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We want to stream events as fast as possible to perf.data, and also in the future we want to have splice working, when no interception will be possible. Using build_id__mark_dso_hit_ops to create the list of DSOs that back MMAPs we also optimize disk usage in the build-id cache by only caching DSOs that had hits. Suggested-by: Peter Zijlstra Signed-off-by: Arnaldo Carvalho de Melo Cc: Xiao Guangrong Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Paul Mackerras LKML-Reference: <1265223128-11786-6-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 949167e..706f001 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -12,6 +12,7 @@ #include "perf.h" +#include "util/build-id.h" #include "util/util.h" #include "util/parse-options.h" #include "util/parse-events.h" @@ -65,6 +66,7 @@ static int nr_poll = 0; static int nr_cpu = 0; static int file_new = 1; +static off_t post_processing_offset; static struct perf_session *session; @@ -114,26 +116,10 @@ static void write_output(void *buf, size_t size) } } -static void write_event(event_t *buf, size_t size) -{ - /* - * Add it to the list of DSOs, so that when we finish this - * record session we can pick the available build-ids. - */ - if (buf->header.type == PERF_RECORD_MMAP) { - struct list_head *head = &dsos__user; - if (buf->mmap.header.misc == 1) - head = &dsos__kernel; - __dsos__findnew(head, buf->mmap.filename); - } - - write_output(buf, size); -} - static int process_synthesized_event(event_t *event, struct perf_session *self __used) { - write_event(event, event->header.size); + write_output(event, event->header.size); return 0; } @@ -185,14 +171,14 @@ static void mmap_read(struct mmap_data *md) size = md->mask + 1 - (old & md->mask); old += size; - write_event(buf, size); + write_output(buf, size); } buf = &data[old & md->mask]; size = head - old; old += size; - write_event(buf, size); + write_output(buf, size); md->prev = old; mmap_write_tail(md, old); @@ -402,10 +388,21 @@ static void open_counters(int cpu, pid_t pid) nr_cpu++; } +static int process_buildids(void) +{ + u64 size = lseek(output, 0, SEEK_CUR); + + session->fd = output; + return __perf_session__process_events(session, post_processing_offset, + size - post_processing_offset, + size, &build_id__mark_dso_hit_ops); +} + static void atexit_header(void) { session->header.data_size += bytes_written; + process_buildids(); perf_header__write(&session->header, output, true); } @@ -558,6 +555,8 @@ static int __cmd_record(int argc, const char **argv) return err; } + post_processing_offset = lseek(output, 0, SEEK_CUR); + err = event__synthesize_kernel_mmap(process_synthesized_event, session, "_text"); if (err < 0) { diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index ed3efd7..d5facd5 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -205,8 +205,11 @@ static int __dsos__write_buildid_table(struct list_head *head, u16 misc, int fd) dsos__for_each_with_build_id(pos, head) { int err; struct build_id_event b; - size_t len = pos->long_name_len + 1; + size_t len; + if (!pos->hit) + continue; + len = pos->long_name_len + 1; len = ALIGN(len, NAME_ALIGN); memset(&b, 0, sizeof(b)); memcpy(&b.build_id, pos->build_id, sizeof(pos->build_id)); @@ -371,7 +374,7 @@ static int perf_header__adds_write(struct perf_header *self, int fd) u64 sec_start; int idx = 0, err; - if (dsos__read_build_ids()) + if (dsos__read_build_ids(true)) perf_header__set_feat(self, HEADER_BUILD_ID); nr_sections = bitmap_weight(self->adds_features, HEADER_FEAT_BITS); diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index aa8a031..74cbc64 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -385,8 +385,9 @@ static struct thread *perf_session__register_idle_thread(struct perf_session *se return thread; } -int perf_session__process_events(struct perf_session *self, - struct perf_event_ops *ops) +int __perf_session__process_events(struct perf_session *self, + u64 data_offset, u64 data_size, + u64 file_size, struct perf_event_ops *ops) { int err, mmap_prot, mmap_flags; u64 head, shift; @@ -396,32 +397,11 @@ int perf_session__process_events(struct perf_session *self, uint32_t size; char *buf; - if (perf_session__register_idle_thread(self) == NULL) - return -ENOMEM; - perf_event_ops__fill_defaults(ops); page_size = sysconf(_SC_PAGESIZE); - head = self->header.data_offset; - - if (!symbol_conf.full_paths) { - char bf[PATH_MAX]; - - if (getcwd(bf, sizeof(bf)) == NULL) { - err = -errno; -out_getcwd_err: - pr_err("failed to get the current directory\n"); - goto out_err; - } - self->cwd = strdup(bf); - if (self->cwd == NULL) { - err = -ENOMEM; - goto out_getcwd_err; - } - self->cwdlen = strlen(self->cwd); - } - + head = data_offset; shift = page_size * (head / page_size); offset += shift; head -= shift; @@ -486,10 +466,10 @@ more: head += size; - if (offset + head >= self->header.data_offset + self->header.data_size) + if (offset + head >= data_offset + data_size) goto done; - if (offset + head < self->size) + if (offset + head < file_size) goto more; done: err = 0; @@ -497,6 +477,38 @@ out_err: return err; } +int perf_session__process_events(struct perf_session *self, + struct perf_event_ops *ops) +{ + int err; + + if (perf_session__register_idle_thread(self) == NULL) + return -ENOMEM; + + if (!symbol_conf.full_paths) { + char bf[PATH_MAX]; + + if (getcwd(bf, sizeof(bf)) == NULL) { + err = -errno; +out_getcwd_err: + pr_err("failed to get the current directory\n"); + goto out_err; + } + self->cwd = strdup(bf); + if (self->cwd == NULL) { + err = -ENOMEM; + goto out_getcwd_err; + } + self->cwdlen = strlen(self->cwd); + } + + err = __perf_session__process_events(self, self->header.data_offset, + self->header.data_size, + self->size, ops); +out_err: + return err; +} + bool perf_session__has_traces(struct perf_session *self, const char *msg) { if (!(self->sample_type & PERF_SAMPLE_RAW)) { diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 752d75a..31950fc 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -50,6 +50,9 @@ void perf_session__delete(struct perf_session *self); void perf_event_header__bswap(struct perf_event_header *self); +int __perf_session__process_events(struct perf_session *self, + u64 data_offset, u64 data_size, u64 size, + struct perf_event_ops *ops); int perf_session__process_events(struct perf_session *self, struct perf_event_ops *event_ops); diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index e752837..bfb0554 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1076,25 +1076,28 @@ static bool dso__build_id_equal(const struct dso *self, u8 *build_id) return memcmp(self->build_id, build_id, sizeof(self->build_id)) == 0; } -static bool __dsos__read_build_ids(struct list_head *head) +static bool __dsos__read_build_ids(struct list_head *head, bool with_hits) { bool have_build_id = false; struct dso *pos; - list_for_each_entry(pos, head, node) + list_for_each_entry(pos, head, node) { + if (with_hits && !pos->hit) + continue; if (filename__read_build_id(pos->long_name, pos->build_id, sizeof(pos->build_id)) > 0) { have_build_id = true; pos->has_build_id = true; } + } return have_build_id; } -bool dsos__read_build_ids(void) +bool dsos__read_build_ids(bool with_hits) { - bool kbuildids = __dsos__read_build_ids(&dsos__kernel), - ubuildids = __dsos__read_build_ids(&dsos__user); + bool kbuildids = __dsos__read_build_ids(&dsos__kernel, with_hits), + ubuildids = __dsos__read_build_ids(&dsos__user, with_hits); return kbuildids || ubuildids; } diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index e90568a..1b4192e 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -157,7 +157,7 @@ struct symbol *dso__find_symbol_by_name(struct dso *self, enum map_type type, int filename__read_build_id(const char *filename, void *bf, size_t size); int sysfs__read_build_id(const char *filename, void *bf, size_t size); -bool dsos__read_build_ids(void); +bool dsos__read_build_ids(bool with_hits); int build_id__sprintf(const u8 *self, int len, char *bf); int kallsyms__parse(const char *filename, void *arg, int (*process_symbol)(void *arg, const char *name, -- cgit v0.10.2 From 29a9f66d703cb9464e24084e09edab5683e1b6b8 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 3 Feb 2010 16:52:06 -0200 Subject: perf tools: Adjust some verbosity levels MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Not to pollute too much 'perf annotate' debugging sessions. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1265223128-11786-7-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 73c202e..4fc3899 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -97,9 +97,7 @@ static void hist_hit(struct hist_entry *he, u64 ip) sym_size = sym->end - sym->start; offset = ip - sym->start; - if (verbose) - fprintf(stderr, "%s: ip=%Lx\n", __func__, - he->map->unmap_ip(he->map, ip)); + pr_debug3("%s: ip=%#Lx\n", __func__, he->map->unmap_ip(he->map, ip)); if (offset >= sym_size) return; @@ -108,12 +106,8 @@ static void hist_hit(struct hist_entry *he, u64 ip) h->sum++; h->ip[offset]++; - if (verbose >= 3) - printf("%p %s: count++ [ip: %p, %08Lx] => %Ld\n", - (void *)(unsigned long)he->sym->start, - he->sym->name, - (void *)(unsigned long)ip, ip - he->sym->start, - h->ip[offset]); + pr_debug3("%#Lx %s: count++ [ip: %#Lx, %#Lx] => %Ld\n", he->sym->start, + he->sym->name, ip, ip - he->sym->start, h->ip[offset]); } static int perf_session__add_hist_entry(struct perf_session *self, @@ -136,14 +130,14 @@ static int process_sample_event(event_t *event, struct perf_session *session) event->ip.pid, event->ip.ip); if (event__preprocess_sample(event, session, &al, symbol_filter) < 0) { - fprintf(stderr, "problem processing %d event, skipping it.\n", - event->header.type); + pr_warning("problem processing %d event, skipping it.\n", + event->header.type); return -1; } if (!al.filtered && perf_session__add_hist_entry(session, &al, 1)) { - fprintf(stderr, "problem incrementing symbol count, " - "skipping event\n"); + pr_warning("problem incrementing symbol count, " + "skipping event\n"); return -1; } @@ -378,11 +372,9 @@ static void annotate_sym(struct hist_entry *he) if (!filename) return; - if (verbose) - fprintf(stderr, "%s: filename=%s, sym=%s, start=%Lx, end=%Lx\n", - __func__, filename, sym->name, - map->unmap_ip(map, sym->start), - map->unmap_ip(map, sym->end)); + pr_debug("%s: filename=%s, sym=%s, start=%#Lx, end=%#Lx\n", __func__, + filename, sym->name, map->unmap_ip(map, sym->start), + map->unmap_ip(map, sym->end)); if (full_paths) d_filename = filename; @@ -542,9 +534,8 @@ int cmd_annotate(int argc, const char **argv, const char *prefix __used) setup_pager(); if (field_sep && *field_sep == '.') { - fputs("'.' is the only non valid --field-separator argument\n", - stderr); - exit(129); + pr_err("'.' is the only non valid --field-separator argument\n"); + return -1; } return __cmd_annotate(); diff --git a/tools/perf/util/include/linux/kernel.h b/tools/perf/util/include/linux/kernel.h index 21c0274..f261165 100644 --- a/tools/perf/util/include/linux/kernel.h +++ b/tools/perf/util/include/linux/kernel.h @@ -101,5 +101,6 @@ simple_strtoul(const char *nptr, char **endptr, int base) eprintf(n, pr_fmt(fmt), ##__VA_ARGS__) #define pr_debug2(fmt, ...) pr_debugN(2, pr_fmt(fmt), ##__VA_ARGS__) #define pr_debug3(fmt, ...) pr_debugN(3, pr_fmt(fmt), ##__VA_ARGS__) +#define pr_debug4(fmt, ...) pr_debugN(4, pr_fmt(fmt), ##__VA_ARGS__) #endif diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index bfb0554..a60ba2b 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -137,7 +137,7 @@ static struct symbol *symbol__new(u64 start, u64 len, const char *name) self->start = start; self->end = len ? start + len - 1 : start; - pr_debug3("%s: %s %#Lx-%#Lx\n", __func__, name, start, self->end); + pr_debug4("%s: %s %#Lx-%#Lx\n", __func__, name, start, self->end); memcpy(self->name, name, namelen); @@ -1024,9 +1024,10 @@ static int dso__load_sym(struct dso *self, struct map *map, const char *name, } if (curr_dso->adjust_symbols) { - pr_debug2("adjusting symbol: st_value: %Lx sh_addr: " - "%Lx sh_offset: %Lx\n", (u64)sym.st_value, - (u64)shdr.sh_addr, (u64)shdr.sh_offset); + pr_debug4("%s: adjusting symbol: st_value: %#Lx " + "sh_addr: %#Lx sh_offset: %#Lx\n", __func__, + (u64)sym.st_value, (u64)shdr.sh_addr, + (u64)shdr.sh_offset); sym.st_value -= shdr.sh_addr - shdr.sh_offset; } /* -- cgit v0.10.2 From 7a2b6209863626cf8362e5ff4653491558f91e67 Mon Sep 17 00:00:00 2001 From: Kirill Smelkov Date: Wed, 3 Feb 2010 16:52:07 -0200 Subject: perf annotate: Fix it for non-prelinked *.so The problem was we were incorrectly calculating objdump addresses for sym->start and sym->end, look: For simple ET_DYN type DSO (*.so) with one function, objdump -dS output is something like this: 000004ac : int my_strlen(const char *s) 4ac: 55 push %ebp 4ad: 89 e5 mov %esp,%ebp 4af: 83 ec 10 sub $0x10,%esp { i.e. we have relative-to-dso-mapping IPs (=RIP) there. For ET_EXEC type and probably for prelinked libs as well (sorry can't test - I don't use prelink) objdump outputs absolute IPs, e.g. 08048604 : extern "C" int zz_strlen(const char *s) 8048604: 55 push %ebp 8048605: 89 e5 mov %esp,%ebp 8048607: 83 ec 10 sub $0x10,%esp { So, if sym->start is always relative to dso mapping(*), we'll have to unmap it for ET_EXEC like cases, and leave as is for ET_DYN cases. (*) and it is - we've explicitely made it relative. Look for adjust_symbols handling in dso__load_sym() Previously we were always unmapping sym->start and for ET_DYN dsos resulting addresses were wrong, and so objdump output was empty. The end result was that perf annotate output for symbols from non-prelinked *.so had always 0.00% percents only, which is wrong. To fix it, let's introduce a helper for converting rip to objdump address, and also let's document what map_ip() and unmap_ip() do -- I had to study sources for several hours to understand it. Signed-off-by: Kirill Smelkov Signed-off-by: Arnaldo Carvalho de Melo Cc: Mike Galbraith LKML-Reference: <1265223128-11786-8-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 4fc3899..28ea4e0 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -189,7 +189,7 @@ static int parse_line(FILE *file, struct hist_entry *he, u64 len) line_ip = -1; } - start = he->map->unmap_ip(he->map, sym->start); + start = map__rip_2objdump(he->map, sym->start); if (line_ip != -1) { const char *path = NULL; @@ -397,7 +397,8 @@ static void annotate_sym(struct hist_entry *he) dso, dso->long_name, sym, sym->name); sprintf(command, "objdump --start-address=0x%016Lx --stop-address=0x%016Lx -dS %s|grep -v %s", - map->unmap_ip(map, sym->start), map->unmap_ip(map, sym->end), + map__rip_2objdump(map, sym->start), + map__rip_2objdump(map, sym->end), filename, filename); if (verbose >= 3) diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index f6626cc..af5805f 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -210,3 +210,15 @@ size_t map__fprintf(struct map *self, FILE *fp) return fprintf(fp, " %Lx-%Lx %Lx %s\n", self->start, self->end, self->pgoff, self->dso->name); } + +/* + * objdump wants/reports absolute IPs for ET_EXEC, and RIPs for ET_DYN. + * map->dso->adjust_symbols==1 for ET_EXEC-like cases. + */ +u64 map__rip_2objdump(struct map *map, u64 rip) +{ + u64 addr = map->dso->adjust_symbols ? + map->unmap_ip(map, rip) : /* RIP -> IP */ + rip; + return addr; +} diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h index de04839..9cee9c7 100644 --- a/tools/perf/util/map.h +++ b/tools/perf/util/map.h @@ -26,8 +26,12 @@ struct map { u64 end; enum map_type type; u64 pgoff; + + /* ip -> dso rip */ u64 (*map_ip)(struct map *, u64); + /* dso rip -> ip */ u64 (*unmap_ip)(struct map *, u64); + struct dso *dso; }; @@ -56,6 +60,11 @@ static inline u64 identity__map_ip(struct map *map __used, u64 ip) return ip; } + +/* rip -> addr suitable for passing to `objdump --start-address=` */ +u64 map__rip_2objdump(struct map *map, u64 rip); + + struct symbol; struct mmap_event; -- cgit v0.10.2 From 6cff0e8dbaa4d5d822a814e5028683d7e71c3291 Mon Sep 17 00:00:00 2001 From: Kirill Smelkov Date: Wed, 3 Feb 2010 16:52:08 -0200 Subject: perf top: Teach it to autolocate vmlinux By relying on logic in dso__load_kernel_sym(), we can automatically load vmlinux. The only thing which needs to be adjusted, is how --sym-annotate option is handled - now we can't rely on vmlinux been loaded until full successful pass of dso__load_vmlinux(), but that's not the case if we'll do sym_filter_entry setup in symbol_filter(). So move this step right after event__process_sample() where we know the whole dso__load_kernel_sym() pass is done. By the way, though conceptually similar `perf top` still can't annotate userspace - see next patches with fixes. Signed-off-by: Kirill Smelkov Signed-off-by: Arnaldo Carvalho de Melo Cc: Mike Galbraith LKML-Reference: <1265223128-11786-9-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/Documentation/perf-top.txt b/tools/perf/Documentation/perf-top.txt index 4a7d558..785b9fc 100644 --- a/tools/perf/Documentation/perf-top.txt +++ b/tools/perf/Documentation/perf-top.txt @@ -74,7 +74,7 @@ OPTIONS -s :: --sym-annotate=:: - Annotate this symbol. Requires -k option. + Annotate this symbol. -v:: --verbose:: diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 1fc018e..83c09c8 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -94,6 +94,7 @@ struct source_line { static char *sym_filter = NULL; struct sym_entry *sym_filter_entry = NULL; +struct sym_entry *sym_filter_entry_sched = NULL; static int sym_pcnt_filter = 5; static int sym_counter = 0; static int display_weighted = -1; @@ -695,11 +696,9 @@ static void print_mapped_keys(void) fprintf(stdout, "\t[f] profile display filter (count). \t(%d)\n", count_filter); - if (symbol_conf.vmlinux_name) { - fprintf(stdout, "\t[F] annotate display filter (percent). \t(%d%%)\n", sym_pcnt_filter); - fprintf(stdout, "\t[s] annotate symbol. \t(%s)\n", name?: "NULL"); - fprintf(stdout, "\t[S] stop annotation.\n"); - } + fprintf(stdout, "\t[F] annotate display filter (percent). \t(%d%%)\n", sym_pcnt_filter); + fprintf(stdout, "\t[s] annotate symbol. \t(%s)\n", name?: "NULL"); + fprintf(stdout, "\t[S] stop annotation.\n"); if (nr_counters > 1) fprintf(stdout, "\t[w] toggle display weighted/count[E]r. \t(%d)\n", display_weighted ? 1 : 0); @@ -725,14 +724,13 @@ static int key_mapped(int c) case 'Q': case 'K': case 'U': + case 'F': + case 's': + case 'S': return 1; case 'E': case 'w': return nr_counters > 1 ? 1 : 0; - case 'F': - case 's': - case 'S': - return symbol_conf.vmlinux_name ? 1 : 0; default: break; } @@ -910,8 +908,12 @@ static int symbol_filter(struct map *map, struct symbol *sym) syme = symbol__priv(sym); syme->map = map; syme->src = NULL; - if (!sym_filter_entry && sym_filter && !strcmp(name, sym_filter)) - sym_filter_entry = syme; + + if (!sym_filter_entry && sym_filter && !strcmp(name, sym_filter)) { + /* schedule initial sym_filter_entry setup */ + sym_filter_entry_sched = syme; + sym_filter = NULL; + } for (i = 0; skip_symbols[i]; i++) { if (!strcmp(skip_symbols[i], name)) { @@ -976,6 +978,13 @@ static void event__process_sample(const event_t *self, return; } + /* let's see, whether we need to install initial sym_filter_entry */ + if (sym_filter_entry_sched) { + sym_filter_entry = sym_filter_entry_sched; + sym_filter_entry_sched = NULL; + parse_source(sym_filter_entry); + } + syme = symbol__priv(al.sym); if (!syme->skip) { syme->count[counter]++; @@ -1270,7 +1279,7 @@ static const struct option options[] = { OPT_BOOLEAN('i', "inherit", &inherit, "child tasks inherit counters"), OPT_STRING('s', "sym-annotate", &sym_filter, "symbol name", - "symbol to annotate - requires -k option"), + "symbol to annotate"), OPT_BOOLEAN('z', "zero", &zero, "zero history across updates"), OPT_INTEGER('F', "freq", &freq, @@ -1306,16 +1315,14 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) symbol_conf.priv_size = (sizeof(struct sym_entry) + (nr_counters + 1) * sizeof(unsigned long)); - if (symbol_conf.vmlinux_name == NULL) - symbol_conf.try_vmlinux_path = true; + + symbol_conf.try_vmlinux_path = (symbol_conf.vmlinux_name == NULL); if (symbol__init() < 0) return -1; if (delay_secs < 1) delay_secs = 1; - parse_source(sym_filter_entry); - /* * User specified count overrides default frequency. */ -- cgit v0.10.2 From 57d818895f9d294ab9080e5a662675fdee943ff1 Mon Sep 17 00:00:00 2001 From: Mike Galbraith Date: Thu, 4 Feb 2010 07:31:46 +0100 Subject: perf annotate: Fix perf top module symbol annotation Signed-off-by: Mike Galbraith Cc: Kirill Smelkov Cc: Arnaldo Carvalho de Melo Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Frederic Weisbecker LKML-Reference: <1265265106.6364.5.camel@marge.simson.net> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 83c09c8..e4156bc 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -204,8 +204,8 @@ static void parse_source(struct sym_entry *syme) sprintf(command, "objdump --start-address=0x%016Lx " "--stop-address=0x%016Lx -dS %s", - map->unmap_ip(map, sym->start), - map->unmap_ip(map, sym->end), path); + map__rip_2objdump(map, sym->start), + map__rip_2objdump(map, sym->end), path); file = popen(command, "r"); if (!file) -- cgit v0.10.2 From 615d0ebbc782b67296e3226c293f520f93f93515 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 2 Feb 2010 16:49:04 -0500 Subject: kprobes: Disable booster when CONFIG_PREEMPT=y Disable kprobe booster when CONFIG_PREEMPT=y at this time, because it can't ensure that all kernel threads preempted on kprobe's boosted slot run out from the slot even using freeze_processes(). The booster on preemptive kernel will be resumed if synchronize_tasks() or something like that is introduced. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Ananth N Mavinakayanahalli Cc: Frederic Weisbecker Cc: Jim Keniston Cc: Mathieu Desnoyers Cc: Steven Rostedt LKML-Reference: <20100202214904.4694.24330.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/arch/ia64/kernel/kprobes.c b/arch/ia64/kernel/kprobes.c index 9adac44..7026b29 100644 --- a/arch/ia64/kernel/kprobes.c +++ b/arch/ia64/kernel/kprobes.c @@ -870,7 +870,7 @@ static int __kprobes pre_kprobes_handler(struct die_args *args) return 1; ss_probe: -#if !defined(CONFIG_PREEMPT) || defined(CONFIG_FREEZER) +#if !defined(CONFIG_PREEMPT) if (p->ainsn.inst_flag == INST_FLAG_BOOSTABLE && !p->post_handler) { /* Boost up -- we can execute copied instructions directly */ ia64_psr(regs)->ri = p->ainsn.slot; diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c index 5b8c750..9453815 100644 --- a/arch/x86/kernel/kprobes.c +++ b/arch/x86/kernel/kprobes.c @@ -429,7 +429,7 @@ void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri, static void __kprobes setup_singlestep(struct kprobe *p, struct pt_regs *regs, struct kprobe_ctlblk *kcb) { -#if !defined(CONFIG_PREEMPT) || defined(CONFIG_FREEZER) +#if !defined(CONFIG_PREEMPT) if (p->ainsn.boostable == 1 && !p->post_handler) { /* Boost up -- we can execute copied instructions directly */ reset_current_kprobe(); diff --git a/kernel/kprobes.c b/kernel/kprobes.c index b7df302..9907a03 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -124,30 +124,6 @@ static LIST_HEAD(kprobe_insn_pages); static int kprobe_garbage_slots; static int collect_garbage_slots(void); -static int __kprobes check_safety(void) -{ - int ret = 0; -#if defined(CONFIG_PREEMPT) && defined(CONFIG_FREEZER) - ret = freeze_processes(); - if (ret == 0) { - struct task_struct *p, *q; - do_each_thread(p, q) { - if (p != current && p->state == TASK_RUNNING && - p->pid != 0) { - printk("Check failed: %s is running\n",p->comm); - ret = -1; - goto loop_end; - } - } while_each_thread(p, q); - } -loop_end: - thaw_processes(); -#else - synchronize_sched(); -#endif - return ret; -} - /** * __get_insn_slot() - Find a slot on an executable page for an instruction. * We allocate an executable page if there's no room on existing ones. @@ -235,9 +211,8 @@ static int __kprobes collect_garbage_slots(void) { struct kprobe_insn_page *kip, *next; - /* Ensure no-one is preepmted on the garbages */ - if (check_safety()) - return -EAGAIN; + /* Ensure no-one is interrupted on the garbages */ + synchronize_sched(); list_for_each_entry_safe(kip, next, &kprobe_insn_pages, list) { int i; -- cgit v0.10.2 From 2cfa19780d61740f65790c5bae363b759d7c96fa Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 2 Feb 2010 16:49:11 -0500 Subject: ftrace/alternatives: Introducing *_text_reserved functions Introducing *_text_reserved functions for checking the text address range is partially reserved or not. This patch provides checking routines for x86 smp alternatives and dynamic ftrace. Since both functions modify fixed pieces of kernel text, they should reserve and protect those from other dynamic text modifier, like kprobes. This will also be extended when introducing other subsystems which modify fixed pieces of kernel text. Dynamic text modifiers should avoid those. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Steven Rostedt Cc: przemyslaw@pawelczyk.it Cc: Frederic Weisbecker Cc: Ananth N Mavinakayanahalli Cc: Jim Keniston Cc: Mathieu Desnoyers Cc: Jason Baron LKML-Reference: <20100202214911.4694.16587.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/include/asm/alternative.h b/arch/x86/include/asm/alternative.h index 69b74a7..ac80b7d 100644 --- a/arch/x86/include/asm/alternative.h +++ b/arch/x86/include/asm/alternative.h @@ -65,12 +65,17 @@ extern void alternatives_smp_module_add(struct module *mod, char *name, void *text, void *text_end); extern void alternatives_smp_module_del(struct module *mod); extern void alternatives_smp_switch(int smp); +extern int alternatives_text_reserved(void *start, void *end); #else static inline void alternatives_smp_module_add(struct module *mod, char *name, void *locks, void *locks_end, void *text, void *text_end) {} static inline void alternatives_smp_module_del(struct module *mod) {} static inline void alternatives_smp_switch(int smp) {} +static inline int alternatives_text_reserved(void *start, void *end) +{ + return 0; +} #endif /* CONFIG_SMP */ /* alternative assembly primitive: */ diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index de7353c..3c13284 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -390,6 +390,22 @@ void alternatives_smp_switch(int smp) mutex_unlock(&smp_alt); } +/* Return 1 if the address range is reserved for smp-alternatives */ +int alternatives_text_reserved(void *start, void *end) +{ + struct smp_alt_module *mod; + u8 **ptr; + + list_for_each_entry(mod, &smp_alt_modules, next) { + if (mod->text > end || mod->text_end < start) + continue; + for (ptr = mod->locks; ptr < mod->locks_end; ptr++) + if (start <= *ptr && end >= *ptr) + return 1; + } + + return 0; +} #endif #ifdef CONFIG_PARAVIRT diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 0b4f97d..9d127ef 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -134,6 +134,8 @@ extern void unregister_ftrace_function_probe_func(char *glob, struct ftrace_probe_ops *ops); extern void unregister_ftrace_function_probe_all(char *glob); +extern int ftrace_text_reserved(void *start, void *end); + enum { FTRACE_FL_FREE = (1 << 0), FTRACE_FL_FAILED = (1 << 1), @@ -250,6 +252,10 @@ static inline int unregister_ftrace_command(char *cmd_name) { return -EINVAL; } +static inline int ftrace_text_reserved(void *start, void *end) +{ + return 0; +} #endif /* CONFIG_DYNAMIC_FTRACE */ /* totally disable ftrace - can not re-enable after this */ diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 1e6640f..3d90661 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1025,6 +1025,21 @@ static void ftrace_bug(int failed, unsigned long ip) } +/* Return 1 if the address range is reserved for ftrace */ +int ftrace_text_reserved(void *start, void *end) +{ + struct dyn_ftrace *rec; + struct ftrace_page *pg; + + do_for_each_ftrace_rec(pg, rec) { + if (rec->ip <= (unsigned long)end && + rec->ip + MCOUNT_INSN_SIZE > (unsigned long)start) + return 1; + } while_for_each_ftrace_rec(); + return 0; +} + + static int __ftrace_replace_code(struct dyn_ftrace *rec, int enable) { -- cgit v0.10.2 From 4554dbcb85a4ed2abaa2b6fa15649b796699ec89 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 2 Feb 2010 16:49:18 -0500 Subject: kprobes: Check probe address is reserved Check whether the address of new probe is already reserved by ftrace or alternatives (on x86) when registering new probe. If reserved, it returns an error and not register the probe. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Steven Rostedt Cc: przemyslaw@pawelczyk.it Cc: Frederic Weisbecker Cc: Ananth N Mavinakayanahalli Cc: Jim Keniston Cc: Mathieu Desnoyers Cc: Jason Baron LKML-Reference: <20100202214918.4694.94179.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c index 9453815..5de9f4a 100644 --- a/arch/x86/kernel/kprobes.c +++ b/arch/x86/kernel/kprobes.c @@ -337,6 +337,9 @@ static void __kprobes arch_copy_kprobe(struct kprobe *p) int __kprobes arch_prepare_kprobe(struct kprobe *p) { + if (alternatives_text_reserved(p->addr, p->addr)) + return -EINVAL; + if (!can_probe((unsigned long)p->addr)) return -EILSEQ; /* insn: must be on special executable page on x86. */ diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 9907a03..c3340e8 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -44,6 +44,7 @@ #include #include #include +#include #include #include @@ -703,7 +704,8 @@ int __kprobes register_kprobe(struct kprobe *p) preempt_disable(); if (!kernel_text_address((unsigned long) p->addr) || - in_kprobes_functions((unsigned long) p->addr)) { + in_kprobes_functions((unsigned long) p->addr) || + ftrace_text_reserved(p->addr, p->addr)) { preempt_enable(); return -EINVAL; } -- cgit v0.10.2 From f24bb999d2b9f2950e5cac5b69bffedf73c24ea4 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 2 Feb 2010 16:49:25 -0500 Subject: ftrace: Remove record freezing Remove record freezing. Because kprobes never puts probe on ftrace's mcount call anymore, it doesn't need ftrace to check whether kprobes on it. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Steven Rostedt Cc: przemyslaw@pawelczyk.it Cc: Frederic Weisbecker LKML-Reference: <20100202214925.4694.73469.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 9d127ef..eb054ae 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -143,7 +143,6 @@ enum { FTRACE_FL_ENABLED = (1 << 3), FTRACE_FL_NOTRACE = (1 << 4), FTRACE_FL_CONVERTED = (1 << 5), - FTRACE_FL_FROZEN = (1 << 6), }; struct dyn_ftrace { diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 3d90661..1904797 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -22,7 +22,6 @@ #include #include #include -#include #include #include #include @@ -898,36 +897,6 @@ static struct dyn_ftrace *ftrace_free_records; } \ } -#ifdef CONFIG_KPROBES - -static int frozen_record_count; - -static inline void freeze_record(struct dyn_ftrace *rec) -{ - if (!(rec->flags & FTRACE_FL_FROZEN)) { - rec->flags |= FTRACE_FL_FROZEN; - frozen_record_count++; - } -} - -static inline void unfreeze_record(struct dyn_ftrace *rec) -{ - if (rec->flags & FTRACE_FL_FROZEN) { - rec->flags &= ~FTRACE_FL_FROZEN; - frozen_record_count--; - } -} - -static inline int record_frozen(struct dyn_ftrace *rec) -{ - return rec->flags & FTRACE_FL_FROZEN; -} -#else -# define freeze_record(rec) ({ 0; }) -# define unfreeze_record(rec) ({ 0; }) -# define record_frozen(rec) ({ 0; }) -#endif /* CONFIG_KPROBES */ - static void ftrace_free_rec(struct dyn_ftrace *rec) { rec->freelist = ftrace_free_records; @@ -1091,14 +1060,6 @@ static void ftrace_replace_code(int enable) !(rec->flags & FTRACE_FL_CONVERTED)) continue; - /* ignore updates to this record's mcount site */ - if (get_kprobe((void *)rec->ip)) { - freeze_record(rec); - continue; - } else { - unfreeze_record(rec); - } - failed = __ftrace_replace_code(rec, enable); if (failed) { rec->flags |= FTRACE_FL_FAILED; -- cgit v0.10.2 From 74b7eb5885415ed41d012f432398d1b697115b5f Mon Sep 17 00:00:00 2001 From: Yong Zhang Date: Fri, 29 Jan 2010 14:57:52 +0800 Subject: sched: Change usage of rt_rq->rt_se to rt_rq->tg->rt_se[cpu] This is the first step to remove rt_rq member rt_se because it have the same meaning with tg->rt_se[cpu]. And the latter style is also used by the fair scheduling class. Signed-off-by: Yong Zhang Cc: Rusty Russell Signed-off-by: Peter Zijlstra LKML-Reference: <2674af741001282257r28c97a92o9f90cf16fe8d3d84@mail.gmail.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index ca49ceb..bf3e38f 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -199,8 +199,11 @@ static void dequeue_rt_entity(struct sched_rt_entity *rt_se); static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) { + int this_cpu = smp_processor_id(); struct task_struct *curr = rq_of_rt_rq(rt_rq)->curr; - struct sched_rt_entity *rt_se = rt_rq->rt_se; + struct sched_rt_entity *rt_se; + + rt_se = rt_rq->tg->rt_se[this_cpu]; if (rt_rq->rt_nr_running) { if (rt_se && !on_rt_rq(rt_se)) @@ -212,7 +215,10 @@ static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) static void sched_rt_rq_dequeue(struct rt_rq *rt_rq) { - struct sched_rt_entity *rt_se = rt_rq->rt_se; + int this_cpu = smp_processor_id(); + struct sched_rt_entity *rt_se; + + rt_se = rt_rq->tg->rt_se[this_cpu]; if (rt_se && on_rt_rq(rt_se)) dequeue_rt_entity(rt_se); -- cgit v0.10.2 From 23577256953c870de9b724c3a2611ce7be6a1e4e Mon Sep 17 00:00:00 2001 From: Yong Zhang Date: Fri, 29 Jan 2010 14:58:47 +0800 Subject: sched: Remove member rt_se from struct rt_rq It's a duplicate of tg->rt_se[cpu] and the only usage is sched_rt_rq_dequeue() and sched_rt_rq_enqueue(). After the first patch to those two function. rt_se can be removed. Signed-off-by: Yong Zhang Cc: Rusty Russell Signed-off-by: Peter Zijlstra LKML-Reference: <2674af741001282258q38781619u653ca4a7dd267347@mail.gmail.com> Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 2386f54..6cee227 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -436,7 +436,6 @@ struct rt_rq { struct rq *rq; struct list_head leaf_rt_rq_list; struct task_group *tg; - struct sched_rt_entity *rt_se; #endif }; @@ -7592,7 +7591,6 @@ static void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq, tg->rt_rq[cpu] = rt_rq; init_rt_rq(rt_rq, rq); rt_rq->tg = tg; - rt_rq->rt_se = rt_se; rt_rq->rt_runtime = tg->rt_bandwidth.rt_runtime; if (add) list_add(&rt_rq->leaf_rt_rq_list, &rq->leaf_rt_rq_list); -- cgit v0.10.2 From 9717e6cd3db22eade7dbae0fc9235c66325a7132 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 28 Jan 2010 13:57:44 +0100 Subject: perf_events: Optimize perf_event_task_tick() Pretty much all of the calls do perf_disable/perf_enable cycles, pull that out to cut back on hardware programming. Signed-off-by: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 40f8b07..087025f 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1573,12 +1573,8 @@ static void rotate_ctx(struct perf_event_context *ctx) raw_spin_lock(&ctx->lock); /* Rotate the first entry last of non-pinned groups */ - perf_disable(); - list_rotate_left(&ctx->flexible_groups); - perf_enable(); - raw_spin_unlock(&ctx->lock); } @@ -1593,6 +1589,8 @@ void perf_event_task_tick(struct task_struct *curr) cpuctx = &__get_cpu_var(perf_cpu_context); ctx = curr->perf_event_ctxp; + perf_disable(); + perf_ctx_adjust_freq(&cpuctx->ctx); if (ctx) perf_ctx_adjust_freq(ctx); @@ -1608,6 +1606,8 @@ void perf_event_task_tick(struct task_struct *curr) cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE); if (ctx) task_ctx_sched_in(curr, EVENT_FLEXIBLE); + + perf_enable(); } static int event_enable_on_exec(struct perf_event *event, -- cgit v0.10.2 From 8c48e444191de0ff84e85d41180d7bc3e74f14ef Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 29 Jan 2010 13:25:31 +0100 Subject: perf_events, x86: Implement intel core solo/duo support Implement Intel Core Solo/Duo, aka. Intel Architectural Performance Monitoring Version 1. Signed-off-by: Peter Zijlstra Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: Arjan van de Ven LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 1846ead..5b91992 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -228,6 +228,17 @@ static const u64 intel_perfmon_event_map[] = static struct event_constraint intel_core_event_constraints[] = { + INTEL_EVENT_CONSTRAINT(0x11, 0x2), /* FP_ASSIST */ + INTEL_EVENT_CONSTRAINT(0x12, 0x2), /* MUL */ + INTEL_EVENT_CONSTRAINT(0x13, 0x2), /* DIV */ + INTEL_EVENT_CONSTRAINT(0x14, 0x1), /* CYCLES_DIV_BUSY */ + INTEL_EVENT_CONSTRAINT(0x19, 0x2), /* DELAYED_BYPASS */ + INTEL_EVENT_CONSTRAINT(0xc1, 0x1), /* FP_COMP_INSTR_RET */ + EVENT_CONSTRAINT_END +}; + +static struct event_constraint intel_core2_event_constraints[] = +{ FIXED_EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ FIXED_EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ INTEL_EVENT_CONSTRAINT(0x10, 0x1), /* FP_COMP_OPS_EXE */ @@ -1216,7 +1227,7 @@ static void intel_pmu_disable_all(void) intel_pmu_disable_bts(); } -static void amd_pmu_disable_all(void) +static void x86_pmu_disable_all(void) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); int idx; @@ -1226,11 +1237,11 @@ static void amd_pmu_disable_all(void) if (!test_bit(idx, cpuc->active_mask)) continue; - rdmsrl(MSR_K7_EVNTSEL0 + idx, val); + rdmsrl(x86_pmu.eventsel + idx, val); if (!(val & ARCH_PERFMON_EVENTSEL0_ENABLE)) continue; val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE; - wrmsrl(MSR_K7_EVNTSEL0 + idx, val); + wrmsrl(x86_pmu.eventsel + idx, val); } } @@ -1278,7 +1289,7 @@ static void intel_pmu_enable_all(void) } } -static void amd_pmu_enable_all(void) +static void x86_pmu_enable_all(void) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); int idx; @@ -1292,7 +1303,7 @@ static void amd_pmu_enable_all(void) val = event->hw.config; val |= ARCH_PERFMON_EVENTSEL0_ENABLE; - wrmsrl(MSR_K7_EVNTSEL0 + idx, val); + wrmsrl(x86_pmu.eventsel + idx, val); } } @@ -1546,7 +1557,7 @@ static inline void intel_pmu_ack_status(u64 ack) wrmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, ack); } -static inline void x86_pmu_enable_event(struct hw_perf_event *hwc, int idx) +static inline void __x86_pmu_enable_event(struct hw_perf_event *hwc, int idx) { (void)checking_wrmsrl(hwc->config_base + idx, hwc->config | ARCH_PERFMON_EVENTSEL0_ENABLE); @@ -1598,12 +1609,6 @@ intel_pmu_disable_event(struct hw_perf_event *hwc, int idx) x86_pmu_disable_event(hwc, idx); } -static inline void -amd_pmu_disable_event(struct hw_perf_event *hwc, int idx) -{ - x86_pmu_disable_event(hwc, idx); -} - static DEFINE_PER_CPU(u64 [X86_PMC_IDX_MAX], pmc_prev_left); /* @@ -1723,15 +1728,14 @@ static void intel_pmu_enable_event(struct hw_perf_event *hwc, int idx) return; } - x86_pmu_enable_event(hwc, idx); + __x86_pmu_enable_event(hwc, idx); } -static void amd_pmu_enable_event(struct hw_perf_event *hwc, int idx) +static void x86_pmu_enable_event(struct hw_perf_event *hwc, int idx) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - if (cpuc->enabled) - x86_pmu_enable_event(hwc, idx); + __x86_pmu_enable_event(hwc, idx); } /* @@ -1988,50 +1992,6 @@ static void intel_pmu_reset(void) local_irq_restore(flags); } -static int p6_pmu_handle_irq(struct pt_regs *regs) -{ - struct perf_sample_data data; - struct cpu_hw_events *cpuc; - struct perf_event *event; - struct hw_perf_event *hwc; - int idx, handled = 0; - u64 val; - - data.addr = 0; - data.raw = NULL; - - cpuc = &__get_cpu_var(cpu_hw_events); - - for (idx = 0; idx < x86_pmu.num_events; idx++) { - if (!test_bit(idx, cpuc->active_mask)) - continue; - - event = cpuc->events[idx]; - hwc = &event->hw; - - val = x86_perf_event_update(event, hwc, idx); - if (val & (1ULL << (x86_pmu.event_bits - 1))) - continue; - - /* - * event overflow - */ - handled = 1; - data.period = event->hw.last_period; - - if (!x86_perf_event_set_period(event, hwc, idx)) - continue; - - if (perf_event_overflow(event, 1, &data, regs)) - p6_pmu_disable_event(hwc, idx); - } - - if (handled) - inc_irq_stat(apic_perf_irqs); - - return handled; -} - /* * This handler is triggered by the local APIC, so the APIC IRQ handling * rules apply: @@ -2098,7 +2058,7 @@ again: return 1; } -static int amd_pmu_handle_irq(struct pt_regs *regs) +static int x86_pmu_handle_irq(struct pt_regs *regs) { struct perf_sample_data data; struct cpu_hw_events *cpuc; @@ -2133,7 +2093,7 @@ static int amd_pmu_handle_irq(struct pt_regs *regs) continue; if (perf_event_overflow(event, 1, &data, regs)) - amd_pmu_disable_event(hwc, idx); + x86_pmu.disable(hwc, idx); } if (handled) @@ -2374,7 +2334,7 @@ static __read_mostly struct notifier_block perf_event_nmi_notifier = { static __initconst struct x86_pmu p6_pmu = { .name = "p6", - .handle_irq = p6_pmu_handle_irq, + .handle_irq = x86_pmu_handle_irq, .disable_all = p6_pmu_disable_all, .enable_all = p6_pmu_enable_all, .enable = p6_pmu_enable_event, @@ -2401,6 +2361,29 @@ static __initconst struct x86_pmu p6_pmu = { .event_constraints = intel_p6_event_constraints }; +static __initconst struct x86_pmu core_pmu = { + .name = "core", + .handle_irq = x86_pmu_handle_irq, + .disable_all = x86_pmu_disable_all, + .enable_all = x86_pmu_enable_all, + .enable = x86_pmu_enable_event, + .disable = x86_pmu_disable_event, + .eventsel = MSR_ARCH_PERFMON_EVENTSEL0, + .perfctr = MSR_ARCH_PERFMON_PERFCTR0, + .event_map = intel_pmu_event_map, + .raw_event = intel_pmu_raw_event, + .max_events = ARRAY_SIZE(intel_perfmon_event_map), + .apic = 1, + /* + * Intel PMCs cannot be accessed sanely above 32 bit width, + * so we install an artificial 1<<31 period regardless of + * the generic event period: + */ + .max_period = (1ULL << 31) - 1, + .get_event_constraints = intel_get_event_constraints, + .event_constraints = intel_core_event_constraints, +}; + static __initconst struct x86_pmu intel_pmu = { .name = "Intel", .handle_irq = intel_pmu_handle_irq, @@ -2427,11 +2410,11 @@ static __initconst struct x86_pmu intel_pmu = { static __initconst struct x86_pmu amd_pmu = { .name = "AMD", - .handle_irq = amd_pmu_handle_irq, - .disable_all = amd_pmu_disable_all, - .enable_all = amd_pmu_enable_all, - .enable = amd_pmu_enable_event, - .disable = amd_pmu_disable_event, + .handle_irq = x86_pmu_handle_irq, + .disable_all = x86_pmu_disable_all, + .enable_all = x86_pmu_enable_all, + .enable = x86_pmu_enable_event, + .disable = x86_pmu_disable_event, .eventsel = MSR_K7_EVNTSEL0, .perfctr = MSR_K7_PERFCTR0, .event_map = amd_pmu_event_map, @@ -2498,9 +2481,10 @@ static __init int intel_pmu_init(void) version = eax.split.version_id; if (version < 2) - return -ENODEV; + x86_pmu = core_pmu; + else + x86_pmu = intel_pmu; - x86_pmu = intel_pmu; x86_pmu.version = version; x86_pmu.num_events = eax.split.num_events; x86_pmu.event_bits = eax.split.bit_width; @@ -2510,12 +2494,17 @@ static __init int intel_pmu_init(void) * Quirk: v2 perfmon does not report fixed-purpose events, so * assume at least 3 events: */ - x86_pmu.num_events_fixed = max((int)edx.split.num_events_fixed, 3); + if (version > 1) + x86_pmu.num_events_fixed = max((int)edx.split.num_events_fixed, 3); /* * Install the hw-cache-events table: */ switch (boot_cpu_data.x86_model) { + case 14: /* 65 nm core solo/duo, "Yonah" */ + pr_cont("Core events, "); + break; + case 15: /* original 65 nm celeron/pentium/core2/xeon, "Merom"/"Conroe" */ case 22: /* single-core 65 nm celeron/core2solo "Merom-L"/"Conroe-L" */ case 23: /* current 45 nm celeron/core2/xeon "Penryn"/"Wolfdale" */ @@ -2523,7 +2512,7 @@ static __init int intel_pmu_init(void) memcpy(hw_cache_event_ids, core2_hw_cache_event_ids, sizeof(hw_cache_event_ids)); - x86_pmu.event_constraints = intel_core_event_constraints; + x86_pmu.event_constraints = intel_core2_event_constraints; pr_cont("Core2 events, "); break; -- cgit v0.10.2 From fce877e3a429940a986e085a41e8b57f2d922e36 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 29 Jan 2010 13:25:12 +0100 Subject: bitops: Ensure the compile time HWEIGHT is only used for such Avoid accidental misuse by failing to compile things Suggested-by: Andrew Morton Signed-off-by: Peter Zijlstra Cc: Linus Torvalds LKML-Reference: Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 5b91992..96cfc1a 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -93,13 +93,16 @@ struct cpu_hw_events { struct perf_event *event_list[X86_PMC_IDX_MAX]; /* in enabled order */ }; -#define EVENT_CONSTRAINT(c, n, m) { \ +#define __EVENT_CONSTRAINT(c, n, m, w) {\ { .idxmsk64[0] = (n) }, \ .code = (c), \ .cmask = (m), \ - .weight = HWEIGHT64((u64)(n)), \ + .weight = (w), \ } +#define EVENT_CONSTRAINT(c, n, m) \ + __EVENT_CONSTRAINT(c, n, m, HWEIGHT(n)) + #define INTEL_EVENT_CONSTRAINT(c, n) \ EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVTSEL_MASK) @@ -2622,7 +2625,8 @@ void __init init_hw_perf_events(void) register_die_notifier(&perf_event_nmi_notifier); unconstrained = (struct event_constraint) - EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_events) - 1, 0); + __EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_events) - 1, + 0, x86_pmu.num_events); pr_info("... version: %d\n", x86_pmu.version); pr_info("... bit width: %d\n", x86_pmu.event_bits); diff --git a/include/linux/bitops.h b/include/linux/bitops.h index ba0fd1e..25b8b2f 100644 --- a/include/linux/bitops.h +++ b/include/linux/bitops.h @@ -45,19 +45,30 @@ static inline unsigned long hweight_long(unsigned long w) return sizeof(w) == 4 ? hweight32(w) : hweight64(w); } -#define HWEIGHT8(w) \ - ( (!!((w) & (1ULL << 0))) + \ - (!!((w) & (1ULL << 1))) + \ - (!!((w) & (1ULL << 2))) + \ - (!!((w) & (1ULL << 3))) + \ - (!!((w) & (1ULL << 4))) + \ - (!!((w) & (1ULL << 5))) + \ - (!!((w) & (1ULL << 6))) + \ +/* + * Clearly slow versions of the hweightN() functions, their benefit is + * of course compile time evaluation of constant arguments. + */ +#define HWEIGHT8(w) \ + ( BUILD_BUG_ON_ZERO(!__builtin_constant_p(w)) + \ + (!!((w) & (1ULL << 0))) + \ + (!!((w) & (1ULL << 1))) + \ + (!!((w) & (1ULL << 2))) + \ + (!!((w) & (1ULL << 3))) + \ + (!!((w) & (1ULL << 4))) + \ + (!!((w) & (1ULL << 5))) + \ + (!!((w) & (1ULL << 6))) + \ (!!((w) & (1ULL << 7))) ) -#define HWEIGHT16(w) (HWEIGHT8(w) + HWEIGHT8(w >> 8)) -#define HWEIGHT32(w) (HWEIGHT16(w) + HWEIGHT16(w >> 16)) -#define HWEIGHT64(w) (HWEIGHT32(w) + HWEIGHT32(w >> 32)) +#define HWEIGHT16(w) (HWEIGHT8(w) + HWEIGHT8((w) >> 8)) +#define HWEIGHT32(w) (HWEIGHT16(w) + HWEIGHT16((w) >> 16)) +#define HWEIGHT64(w) (HWEIGHT32(w) + HWEIGHT32((w) >> 32)) + +/* + * Type invariant version that simply casts things to the + * largest type. + */ +#define HWEIGHT(w) HWEIGHT64((u64)(w)) /** * rol32 - rotate a 32-bit value left -- cgit v0.10.2 From 447a194b393f32699607fd99617a40abd6a95114 Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Mon, 1 Feb 2010 14:50:01 +0200 Subject: perf_events, x86: Fix bug in hw_perf_enable() We cannot assume that because hwc->idx == assign[i], we can avoid reprogramming the counter in hw_perf_enable(). The event may have been scheduled out and another event may have been programmed into this counter. Thus, we need a more robust way of verifying if the counter still contains config/data related to an event. This patch adds a generation number to each counter on each cpu. Using this mechanism we can verify reliabilty whether the content of a counter corresponds to an event. Signed-off-by: Stephane Eranian Signed-off-by: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <4b66dc67.0b38560a.1635.ffffae18@mx.google.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 96cfc1a..a920f17 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -90,6 +90,7 @@ struct cpu_hw_events { int n_events; int n_added; int assign[X86_PMC_IDX_MAX]; /* event to counter assignment */ + u64 tags[X86_PMC_IDX_MAX]; struct perf_event *event_list[X86_PMC_IDX_MAX]; /* in enabled order */ }; @@ -1142,6 +1143,8 @@ static int __hw_perf_event_init(struct perf_event *event) hwc->config = ARCH_PERFMON_EVENTSEL_INT; hwc->idx = -1; + hwc->last_cpu = -1; + hwc->last_tag = ~0ULL; /* * Count user and OS events unless requested not to. @@ -1457,11 +1460,14 @@ static int collect_events(struct cpu_hw_events *cpuc, struct perf_event *leader, return n; } - static inline void x86_assign_hw_event(struct perf_event *event, - struct hw_perf_event *hwc, int idx) + struct cpu_hw_events *cpuc, int i) { - hwc->idx = idx; + struct hw_perf_event *hwc = &event->hw; + + hwc->idx = cpuc->assign[i]; + hwc->last_cpu = smp_processor_id(); + hwc->last_tag = ++cpuc->tags[i]; if (hwc->idx == X86_PMC_IDX_FIXED_BTS) { hwc->config_base = 0; @@ -1480,6 +1486,15 @@ static inline void x86_assign_hw_event(struct perf_event *event, } } +static inline int match_prev_assignment(struct hw_perf_event *hwc, + struct cpu_hw_events *cpuc, + int i) +{ + return hwc->idx == cpuc->assign[i] && + hwc->last_cpu == smp_processor_id() && + hwc->last_tag == cpuc->tags[i]; +} + static void __x86_pmu_disable(struct perf_event *event, struct cpu_hw_events *cpuc); void hw_perf_enable(void) @@ -1508,7 +1523,14 @@ void hw_perf_enable(void) event = cpuc->event_list[i]; hwc = &event->hw; - if (hwc->idx == -1 || hwc->idx == cpuc->assign[i]) + /* + * we can avoid reprogramming counter if: + * - assigned same counter as last time + * - running on same CPU as last time + * - no other event has used the counter since + */ + if (hwc->idx == -1 || + match_prev_assignment(hwc, cpuc, i)) continue; __x86_pmu_disable(event, cpuc); @@ -1522,12 +1544,12 @@ void hw_perf_enable(void) hwc = &event->hw; if (hwc->idx == -1) { - x86_assign_hw_event(event, hwc, cpuc->assign[i]); + x86_assign_hw_event(event, cpuc, i); x86_perf_event_set_period(event, hwc, hwc->idx); } /* * need to mark as active because x86_pmu_disable() - * clear active_mask and eventsp[] yet it preserves + * clear active_mask and events[] yet it preserves * idx */ set_bit(hwc->idx, cpuc->active_mask); diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 556b0f4..071a7db 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -478,9 +478,11 @@ struct hw_perf_event { union { struct { /* hardware */ u64 config; + u64 last_tag; unsigned long config_base; unsigned long event_base; int idx; + int last_cpu; }; struct { /* software */ s64 remaining; -- cgit v0.10.2 From 34d2819f20782feb60f9434470ecfb200875fd41 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Thu, 4 Feb 2010 09:51:28 +0100 Subject: x86, mtrr: Remove unused mtrr/state.c The last reference to the helpers in went away with 9a6b344ea967efa0bb5ca4cb5405f840652b66c4 leaving unused code. Remove it. Signed-off-by: Borislav Petkov LKML-Reference: <20100204085128.GA513@liondog.tnic> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/cpu/mtrr/Makefile b/arch/x86/kernel/cpu/mtrr/Makefile index f4361b5..ad9e5ed 100644 --- a/arch/x86/kernel/cpu/mtrr/Makefile +++ b/arch/x86/kernel/cpu/mtrr/Makefile @@ -1,3 +1,3 @@ -obj-y := main.o if.o generic.o state.o cleanup.o +obj-y := main.o if.o generic.o cleanup.o obj-$(CONFIG_X86_32) += amd.o cyrix.o centaur.o diff --git a/arch/x86/kernel/cpu/mtrr/state.c b/arch/x86/kernel/cpu/mtrr/state.c deleted file mode 100644 index dfc80b4..0000000 --- a/arch/x86/kernel/cpu/mtrr/state.c +++ /dev/null @@ -1,94 +0,0 @@ -#include -#include -#include - -#include -#include -#include -#include - -#include "mtrr.h" - -/* Put the processor into a state where MTRRs can be safely set */ -void set_mtrr_prepare_save(struct set_mtrr_context *ctxt) -{ - unsigned int cr0; - - /* Disable interrupts locally */ - local_irq_save(ctxt->flags); - - if (use_intel() || is_cpu(CYRIX)) { - - /* Save value of CR4 and clear Page Global Enable (bit 7) */ - if (cpu_has_pge) { - ctxt->cr4val = read_cr4(); - write_cr4(ctxt->cr4val & ~X86_CR4_PGE); - } - - /* - * Disable and flush caches. Note that wbinvd flushes the TLBs - * as a side-effect - */ - cr0 = read_cr0() | X86_CR0_CD; - wbinvd(); - write_cr0(cr0); - wbinvd(); - - if (use_intel()) { - /* Save MTRR state */ - rdmsr(MSR_MTRRdefType, ctxt->deftype_lo, ctxt->deftype_hi); - } else { - /* - * Cyrix ARRs - - * everything else were excluded at the top - */ - ctxt->ccr3 = getCx86(CX86_CCR3); - } - } -} - -void set_mtrr_cache_disable(struct set_mtrr_context *ctxt) -{ - if (use_intel()) { - /* Disable MTRRs, and set the default type to uncached */ - mtrr_wrmsr(MSR_MTRRdefType, ctxt->deftype_lo & 0xf300UL, - ctxt->deftype_hi); - } else { - if (is_cpu(CYRIX)) { - /* Cyrix ARRs - everything else were excluded at the top */ - setCx86(CX86_CCR3, (ctxt->ccr3 & 0x0f) | 0x10); - } - } -} - -/* Restore the processor after a set_mtrr_prepare */ -void set_mtrr_done(struct set_mtrr_context *ctxt) -{ - if (use_intel() || is_cpu(CYRIX)) { - - /* Flush caches and TLBs */ - wbinvd(); - - /* Restore MTRRdefType */ - if (use_intel()) { - /* Intel (P6) standard MTRRs */ - mtrr_wrmsr(MSR_MTRRdefType, ctxt->deftype_lo, - ctxt->deftype_hi); - } else { - /* - * Cyrix ARRs - - * everything else was excluded at the top - */ - setCx86(CX86_CCR3, ctxt->ccr3); - } - - /* Enable caches */ - write_cr0(read_cr0() & 0xbfffffff); - - /* Restore value of CR4 */ - if (cpu_has_pge) - write_cr4(ctxt->cr4val); - } - /* Re-enable interrupts locally (if enabled previously) */ - local_irq_restore(ctxt->flags); -} -- cgit v0.10.2 From f887f3019e56389a73617f4e70f512e82cc89adb Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Thu, 4 Feb 2010 16:46:42 +0800 Subject: perf tools: Clean up O_LARGEFILE et al usage Setting _FILE_OFFSET_BITS and using O_LARGEFILE, lseek64, etc, is redundant. Thanks H. Peter Anvin for pointing it out. So, this patch removes O_LARGEFILE, lseek64, etc. Suggested-by: "H. Peter Anvin" Signed-off-by: Xiao Guangrong Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Paul Mackerras Cc: Peter Zijlstra LKML-Reference: <4B6A8972.3070605@cn.fujitsu.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 706f001..3ad599b 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -5,7 +5,6 @@ * (or a CPU, or a PID) into the perf.data output file - for * later analysis via perf report. */ -#define _LARGEFILE64_SOURCE #define _FILE_OFFSET_BITS 64 #include "builtin.h" @@ -451,7 +450,7 @@ static int __cmd_record(int argc, const char **argv) append_file = 0; } - flags = O_CREAT|O_RDWR|O_LARGEFILE; + flags = O_CREAT|O_RDWR; if (append_file) file_new = 0; else diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index d5facd5..6c9aa16 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -1,4 +1,3 @@ -#define _LARGEFILE64_SOURCE #define _FILE_OFFSET_BITS 64 #include @@ -388,7 +387,7 @@ static int perf_header__adds_write(struct perf_header *self, int fd) sec_size = sizeof(*feat_sec) * nr_sections; sec_start = self->data_offset + self->data_size; - lseek64(fd, sec_start + sec_size, SEEK_SET); + lseek(fd, sec_start + sec_size, SEEK_SET); if (perf_header__has_feat(self, HEADER_TRACE_INFO)) { struct perf_file_section *trace_sec; @@ -396,9 +395,9 @@ static int perf_header__adds_write(struct perf_header *self, int fd) trace_sec = &feat_sec[idx++]; /* Write trace info */ - trace_sec->offset = lseek64(fd, 0, SEEK_CUR); + trace_sec->offset = lseek(fd, 0, SEEK_CUR); read_tracing_data(fd, attrs, nr_counters); - trace_sec->size = lseek64(fd, 0, SEEK_CUR) - trace_sec->offset; + trace_sec->size = lseek(fd, 0, SEEK_CUR) - trace_sec->offset; } @@ -408,18 +407,18 @@ static int perf_header__adds_write(struct perf_header *self, int fd) buildid_sec = &feat_sec[idx++]; /* Write build-ids */ - buildid_sec->offset = lseek64(fd, 0, SEEK_CUR); + buildid_sec->offset = lseek(fd, 0, SEEK_CUR); err = dsos__write_buildid_table(fd); if (err < 0) { pr_debug("failed to write buildid table\n"); goto out_free; } - buildid_sec->size = lseek64(fd, 0, SEEK_CUR) - - buildid_sec->offset; + buildid_sec->size = lseek(fd, 0, SEEK_CUR) - + buildid_sec->offset; dsos__cache_build_ids(); } - lseek64(fd, sec_start, SEEK_SET); + lseek(fd, sec_start, SEEK_SET); err = do_write(fd, feat_sec, sec_size); if (err < 0) pr_debug("failed to write feature section\n"); @@ -513,7 +512,7 @@ int perf_header__write(struct perf_header *self, int fd, bool at_exit) pr_debug("failed to write perf header\n"); return err; } - lseek64(fd, self->data_offset + self->data_size, SEEK_SET); + lseek(fd, self->data_offset + self->data_size, SEEK_SET); self->frozen = 1; return 0; @@ -567,7 +566,7 @@ int perf_header__process_sections(struct perf_header *self, int fd, sec_size = sizeof(*feat_sec) * nr_sections; - lseek64(fd, self->data_offset + self->data_size, SEEK_SET); + lseek(fd, self->data_offset + self->data_size, SEEK_SET); if (perf_header__getbuffer64(self, fd, feat_sec, sec_size)) goto out_free; @@ -641,7 +640,7 @@ static int perf_file_section__process(struct perf_file_section *self, struct perf_header *ph, int feat, int fd) { - if (lseek64(fd, self->offset, SEEK_SET) < 0) { + if (lseek(fd, self->offset, SEEK_SET) == (off_t)-1) { pr_debug("Failed to lseek to %Ld offset for feature %d, " "continuing...\n", self->offset, feat); return 0; diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 74cbc64..0de7258 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1,4 +1,3 @@ -#define _LARGEFILE64_SOURCE #define _FILE_OFFSET_BITS 64 #include @@ -15,7 +14,7 @@ static int perf_session__open(struct perf_session *self, bool force) { struct stat input_stat; - self->fd = open(self->filename, O_RDONLY|O_LARGEFILE); + self->fd = open(self->filename, O_RDONLY); if (self->fd < 0) { pr_err("failed to open file: %s", self->filename); if (!strcmp(self->filename, "perf.data")) diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c index ca3c26d..7cd1193 100644 --- a/tools/perf/util/trace-event-read.c +++ b/tools/perf/util/trace-event-read.c @@ -18,7 +18,7 @@ * * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ */ -#define _LARGEFILE64_SOURCE +#define _FILE_OFFSET_BITS 64 #include #include @@ -83,7 +83,7 @@ static char *read_string(void) char *str = NULL; int size = 0; int i; - s64 r; + off_t r; for (;;) { r = read(input_fd, buf, BUFSIZ); @@ -117,8 +117,8 @@ static char *read_string(void) i++; /* move the file descriptor to the end of the string */ - r = lseek64(input_fd, -(r - i), SEEK_CUR); - if (r < 0) + r = lseek(input_fd, -(r - i), SEEK_CUR); + if (r == (off_t)-1) die("lseek"); if (str) { @@ -282,8 +282,8 @@ static void update_cpu_data_index(int cpu) static void get_next_page(int cpu) { - off64_t save_seek; - off64_t ret; + off_t save_seek; + off_t ret; if (!cpu_data[cpu].page) return; @@ -298,17 +298,17 @@ static void get_next_page(int cpu) update_cpu_data_index(cpu); /* other parts of the code may expect the pointer to not move */ - save_seek = lseek64(input_fd, 0, SEEK_CUR); + save_seek = lseek(input_fd, 0, SEEK_CUR); - ret = lseek64(input_fd, cpu_data[cpu].offset, SEEK_SET); - if (ret < 0) + ret = lseek(input_fd, cpu_data[cpu].offset, SEEK_SET); + if (ret == (off_t)-1) die("failed to lseek"); ret = read(input_fd, cpu_data[cpu].page, page_size); if (ret < 0) die("failed to read page"); /* reset the file pointer back */ - lseek64(input_fd, save_seek, SEEK_SET); + lseek(input_fd, save_seek, SEEK_SET); return; } -- cgit v0.10.2 From 2161db969313cb94ffd9377a525fb75c3fee9eeb Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 4 Feb 2010 10:22:01 +0100 Subject: perf tools: Fix session init on non-modular kernels MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit perf top and perf record refuses to initialize on non-modular kernels: refuse to initialize: $ perf top -v map_groups__set_modules_path_dir: cannot open /lib/modules/2.6.33-rc6-tip-00586-g398dde3-dirty/ Cc: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1265223128-11786-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index a60ba2b..6882e9f 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1937,7 +1937,7 @@ int map_groups__create_kernel_maps(struct map_groups *self, return -1; if (symbol_conf.use_modules && map_groups__create_modules(self) < 0) - return -1; + return 0; /* * Now that we have all the maps created, just set the ->end of them: */ -- cgit v0.10.2 From c2c3489c5b0fdb8fbf0f5e9424905c2994ab5660 Mon Sep 17 00:00:00 2001 From: Thadeu Lima de Souza Cascardo Date: Thu, 4 Feb 2010 13:36:24 -0200 Subject: HID: use multi input quirk for TouchPack touchscreen This device generates ABS_Z and ABS_RX events, while it should be generating ABS_X and ABS_Y instead. Using the MULTI_INPUT quirk solves this issue. Signed-off-by: Thadeu Lima de Souza Cascardo Signed-off-by: Daniel Oliveira Nascimento [jkosina@suse.cz: fixed blacklist ordering while resolving conflict] [jkosina@suse.cz: fixed typo to make it compile] Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index b29d9da..282bfe8 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -424,6 +424,9 @@ #define USB_VENDOR_ID_THRUSTMASTER 0x044f +#define USB_VENDOR_ID_TOUCHPACK 0x1bfd +#define USB_DEVICE_ID_TOUCHPACK_RTS 0x1688 + #define USB_VENDOR_ID_TOPMAX 0x0663 #define USB_DEVICE_ID_TOPMAX_COBRAPAD 0x0103 diff --git a/drivers/hid/usbhid/hid-quirks.c b/drivers/hid/usbhid/hid-quirks.c index 88a1c69..7844280 100644 --- a/drivers/hid/usbhid/hid-quirks.c +++ b/drivers/hid/usbhid/hid-quirks.c @@ -46,7 +46,7 @@ static const struct hid_blacklist { { USB_VENDOR_ID_ETURBOTOUCH, USB_DEVICE_ID_ETURBOTOUCH, HID_QUIRK_MULTI_INPUT }, { USB_VENDOR_ID_PANTHERLORD, USB_DEVICE_ID_PANTHERLORD_TWIN_USB_JOYSTICK, HID_QUIRK_MULTI_INPUT | HID_QUIRK_SKIP_OUTPUT_REPORTS }, { USB_VENDOR_ID_PLAYDOTCOM, USB_DEVICE_ID_PLAYDOTCOM_EMS_USBII, HID_QUIRK_MULTI_INPUT }, - { USB_VENDOR_ID_PRODIGE, USB_DEVICE_ID_PRODIGE_CORDLESS, HID_QUIRK_NOGET }, + { USB_VENDOR_ID_TOUCHPACK, USB_DEVICE_ID_TOUCHPACK_RTS, HID_QUIRK_MULTI_INPUT }, { USB_VENDOR_ID_ATEN, USB_DEVICE_ID_ATEN_UC100KM, HID_QUIRK_NOGET }, { USB_VENDOR_ID_ATEN, USB_DEVICE_ID_ATEN_CS124U, HID_QUIRK_NOGET }, @@ -59,6 +59,7 @@ static const struct hid_blacklist { { USB_VENDOR_ID_CH, USB_DEVICE_ID_CH_3AXIS_5BUTTON_STICK, HID_QUIRK_NOGET }, { USB_VENDOR_ID_DMI, USB_DEVICE_ID_DMI_ENC, HID_QUIRK_NOGET }, { USB_VENDOR_ID_ELO, USB_DEVICE_ID_ELO_TS2700, HID_QUIRK_NOGET }, + { USB_VENDOR_ID_PRODIGE, USB_DEVICE_ID_PRODIGE_CORDLESS, HID_QUIRK_NOGET }, { USB_VENDOR_ID_SUN, USB_DEVICE_ID_RARITAN_KVM_DONGLE, HID_QUIRK_NOGET }, { USB_VENDOR_ID_TURBOX, USB_DEVICE_ID_TURBOX_KEYBOARD, HID_QUIRK_NOGET }, { USB_VENDOR_ID_UCLOGIC, USB_DEVICE_ID_UCLOGIC_TABLET_PF1209, HID_QUIRK_MULTI_INPUT }, -- cgit v0.10.2 From 3b0be1a4f2f7d8280574aa6e5eac2dd3dd57e2b7 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 5 Feb 2010 16:11:25 +0900 Subject: sh: Fix an off-by-1 in SH7780 PCIC memory resource mapping. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/pci-sh7780.c b/arch/sh/drivers/pci/pci-sh7780.c index 1e147f4..ffdcbf1 100644 --- a/arch/sh/drivers/pci/pci-sh7780.c +++ b/arch/sh/drivers/pci/pci-sh7780.c @@ -23,7 +23,7 @@ static struct resource sh7785_pci_resources[] = { { - .name = "SH7785_IO", + .name = "PCI IO", .start = 0x1000, .end = SZ_4M - 1, .flags = IORESOURCE_IO, @@ -338,8 +338,8 @@ static int __init sh7780_pci_init(void) /* * Setup the memory BARs */ - for (i = 0; i < chan->nr_resources; i++) { - struct resource *res = chan->resources + (i + 1); + for (i = 1; i < chan->nr_resources; i++) { + struct resource *res = chan->resources + i; resource_size_t size; if (unlikely(res->flags & IORESOURCE_IO)) @@ -361,8 +361,8 @@ static int __init sh7780_pci_init(void) * keeps things pretty simple. */ __raw_writel(((roundup_pow_of_two(size) / SZ_256K) - 1) << 18, - chan->reg_base + SH7780_PCIMBMR(i)); - __raw_writel(res->start, chan->reg_base + SH7780_PCIMBR(i)); + chan->reg_base + SH7780_PCIMBMR(i - 1)); + __raw_writel(res->start, chan->reg_base + SH7780_PCIMBR(i - 1)); } /* -- cgit v0.10.2 From 5ecaafdbf44b1ba400b746c60c401d54c7ee0863 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Fri, 5 Feb 2010 01:24:34 -0500 Subject: kprobes: Add mcount to the kprobes blacklist Since mcount function can be called from everywhere, it should be blacklisted. Moreover, the "mcount" symbol is a special symbol name. So, it is better to put it in the generic blacklist. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Ananth N Mavinakayanahalli Cc: Steven Rostedt LKML-Reference: <20100205062433.3745.36726.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/kernel/kprobes.c b/kernel/kprobes.c index c3340e8..ccec774 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -94,6 +94,7 @@ static struct kprobe_blackpoint kprobe_blacklist[] = { {"native_get_debugreg",}, {"irq_entries_start",}, {"common_interrupt",}, + {"mcount",}, /* mcount can be called from everywhere */ {NULL} /* Terminator */ }; -- cgit v0.10.2 From b06ede84dd1473dec7c6af03a41c8d04d2fee437 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 5 Feb 2010 17:45:14 +0900 Subject: sh: Fix up hp6xx build. With the sparseirq conversion there was a stray irq_desc reference left over, this tidies it up and brings the demuxer in line with what the solution engine boards are doing. Signed-off-by: Paul Mundt diff --git a/arch/sh/cchips/hd6446x/hd64461.c b/arch/sh/cchips/hd6446x/hd64461.c index b447091..bcb31ae 100644 --- a/arch/sh/cchips/hd6446x/hd64461.c +++ b/arch/sh/cchips/hd6446x/hd64461.c @@ -56,18 +56,15 @@ static struct irq_chip hd64461_irq_chip = { static void hd64461_irq_demux(unsigned int irq, struct irq_desc *desc) { unsigned short intv = __raw_readw(HD64461_NIRR); - struct irq_desc *ext_desc; unsigned int ext_irq = HD64461_IRQBASE; intv &= (1 << HD64461_IRQ_NUM) - 1; - while (intv) { - if (intv & 1) { - ext_desc = irq_desc + ext_irq; - handle_level_irq(ext_irq, ext_desc); - } - intv >>= 1; - ext_irq++; + for (; intv; intv >>= 1, ext_irq++) { + if (!(intv & 1)) + continue; + + generic_handle_irq(ext_irq); } } -- cgit v0.10.2 From ef7cc35b0ee03431731186320b18e5da585341ff Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Mon, 25 Jan 2010 11:42:21 -0600 Subject: parisc: add mm API for DMA to vmalloc/vmap areas We already have an API to flush a kernel page along an alias address, so use it. The TLB purge prevents the CPU from doing speculative moveins on the flushed address, so we don't need to implement and invalidate. Acked-by: Kyle McMartin Signed-off-by: James Bottomley diff --git a/arch/parisc/include/asm/cacheflush.h b/arch/parisc/include/asm/cacheflush.h index 7a73b61..4772777 100644 --- a/arch/parisc/include/asm/cacheflush.h +++ b/arch/parisc/include/asm/cacheflush.h @@ -38,6 +38,18 @@ void flush_cache_mm(struct mm_struct *mm); #define flush_kernel_dcache_range(start,size) \ flush_kernel_dcache_range_asm((start), (start)+(size)); +/* vmap range flushes and invalidates. Architecturally, we don't need + * the invalidate, because the CPU should refuse to speculate once an + * area has been flushed, so invalidate is left empty */ +static inline void flush_kernel_vmap_range(void *vaddr, int size) +{ + unsigned long start = (unsigned long)vaddr; + + flush_kernel_dcache_range_asm(start, start + size); +} +static inline void invalidate_kernel_vmap_range(void *vaddr, int size) +{ +} #define flush_cache_vmap(start, end) flush_cache_all() #define flush_cache_vunmap(start, end) flush_cache_all() -- cgit v0.10.2 From 252a9afff76097667429b583e8b5b170b47665a4 Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Mon, 25 Jan 2010 11:42:22 -0600 Subject: arm: add mm API for DMA to vmalloc/vmap areas ARM cannot prevent cache movein, so this patch implements both the flush and invalidate pieces of the API. Signed-off-by: James Bottomley diff --git a/arch/arm/include/asm/cacheflush.h b/arch/arm/include/asm/cacheflush.h index 730aefc..4ae503c 100644 --- a/arch/arm/include/asm/cacheflush.h +++ b/arch/arm/include/asm/cacheflush.h @@ -432,6 +432,16 @@ static inline void __flush_icache_all(void) : "r" (0)); #endif } +static inline void flush_kernel_vmap_range(void *addr, int size) +{ + if ((cache_is_vivt() || cache_is_vipt_aliasing())) + __cpuc_flush_dcache_area(addr, (size_t)size); +} +static inline void invalidate_kernel_vmap_range(void *addr, int size) +{ + if ((cache_is_vivt() || cache_is_vipt_aliasing())) + __cpuc_flush_dcache_area(addr, (size_t)size); +} #define ARCH_HAS_FLUSH_ANON_PAGE static inline void flush_anon_page(struct vm_area_struct *vma, -- cgit v0.10.2 From c9334f6067dbe0380141fc75b122e0a533878838 Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Mon, 25 Jan 2010 11:42:23 -0600 Subject: sh: add mm API for DMA to vmalloc/vmap areas Signed-off-by: James Bottomley diff --git a/arch/sh/include/asm/cacheflush.h b/arch/sh/include/asm/cacheflush.h index dda96eb..da3ebec 100644 --- a/arch/sh/include/asm/cacheflush.h +++ b/arch/sh/include/asm/cacheflush.h @@ -63,6 +63,14 @@ static inline void flush_anon_page(struct vm_area_struct *vma, if (boot_cpu_data.dcache.n_aliases && PageAnon(page)) __flush_anon_page(page, vmaddr); } +static inline void flush_kernel_vmap_range(void *addr, int size) +{ + __flush_wback_region(addr, size); +} +static inline void invalidate_kernel_vmap_range(void *addr, int size) +{ + __flush_invalidate_region(addr, size); +} #define ARCH_HAS_FLUSH_KERNEL_DCACHE_PAGE static inline void flush_kernel_dcache_page(struct page *page) -- cgit v0.10.2 From 73c77e2ccc14413c232c3e0b3aa43a0c4b72ec70 Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Mon, 25 Jan 2010 11:42:24 -0600 Subject: xfs: fix xfs to work with Virtually Indexed architectures xfs_buf.c includes what is essentially a hand rolled version of blk_rq_map_kern(). In order to work properly with the vmalloc buffers that xfs uses, this hand rolled routine must also implement the flushing API for vmap/vmalloc areas. [style updates from hch@lst.de] Acked-by: Christoph Hellwig Signed-off-by: James Bottomley diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index 77b8be8..6f3ebb6 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -76,6 +76,27 @@ struct workqueue_struct *xfsconvertd_workqueue; #define xfs_buf_deallocate(bp) \ kmem_zone_free(xfs_buf_zone, (bp)); +static inline int +xfs_buf_is_vmapped( + struct xfs_buf *bp) +{ + /* + * Return true if the buffer is vmapped. + * + * The XBF_MAPPED flag is set if the buffer should be mapped, but the + * code is clever enough to know it doesn't have to map a single page, + * so the check has to be both for XBF_MAPPED and bp->b_page_count > 1. + */ + return (bp->b_flags & XBF_MAPPED) && bp->b_page_count > 1; +} + +static inline int +xfs_buf_vmap_len( + struct xfs_buf *bp) +{ + return (bp->b_page_count * PAGE_SIZE) - bp->b_offset; +} + /* * Page Region interfaces. * @@ -314,7 +335,7 @@ xfs_buf_free( if (bp->b_flags & (_XBF_PAGE_CACHE|_XBF_PAGES)) { uint i; - if ((bp->b_flags & XBF_MAPPED) && (bp->b_page_count > 1)) + if (xfs_buf_is_vmapped(bp)) free_address(bp->b_addr - bp->b_offset); for (i = 0; i < bp->b_page_count; i++) { @@ -1107,6 +1128,9 @@ xfs_buf_bio_end_io( xfs_buf_ioerror(bp, -error); + if (!error && xfs_buf_is_vmapped(bp) && (bp->b_flags & XBF_READ)) + invalidate_kernel_vmap_range(bp->b_addr, xfs_buf_vmap_len(bp)); + do { struct page *page = bvec->bv_page; @@ -1216,6 +1240,10 @@ next_chunk: submit_io: if (likely(bio->bi_size)) { + if (xfs_buf_is_vmapped(bp)) { + flush_kernel_vmap_range(bp->b_addr, + xfs_buf_vmap_len(bp)); + } submit_bio(rw, bio); if (size) goto next_chunk; -- cgit v0.10.2 From 5c64c7019e571a726f4aa9c1896402c15391a8ed Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 5 Feb 2010 09:37:03 -0500 Subject: x86-32: Move XQUAD definitions to numaq.h The XQUAD stuff is part of the NUMAQ architecture, so move it there. Signed-off-by: Brian Gerst LKML-Reference: <1265380629-3212-2-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/io_32.h b/arch/x86/include/asm/io_32.h index a299900..e16b9db 100644 --- a/arch/x86/include/asm/io_32.h +++ b/arch/x86/include/asm/io_32.h @@ -37,9 +37,6 @@ * - Arnaldo Carvalho de Melo */ -#define XQUAD_PORTIO_BASE 0xfe400000 -#define XQUAD_PORTIO_QUAD 0x40000 /* 256k per quad. */ - #ifdef __KERNEL__ #include diff --git a/arch/x86/include/asm/numaq.h b/arch/x86/include/asm/numaq.h index 9f0a5f5..13370b9 100644 --- a/arch/x86/include/asm/numaq.h +++ b/arch/x86/include/asm/numaq.h @@ -33,6 +33,10 @@ extern int get_memcfg_numaq(void); extern void *xquad_portio; +#define XQUAD_PORTIO_BASE 0xfe400000 +#define XQUAD_PORTIO_QUAD 0x40000 /* 256k per quad. */ +#define XQUAD_PORT_ADDR(port, quad) (xquad_portio + (XQUAD_PORTIO_QUAD*quad) + port) + /* * SYS_CFG_DATA_PRIV_ADDR, struct eachquadmem, and struct sys_cfg_data are the */ diff --git a/arch/x86/pci/numaq_32.c b/arch/x86/pci/numaq_32.c index 8eb295e..8884a1c 100644 --- a/arch/x86/pci/numaq_32.c +++ b/arch/x86/pci/numaq_32.c @@ -8,9 +8,7 @@ #include #include #include - -#define XQUAD_PORTIO_BASE 0xfe400000 -#define XQUAD_PORTIO_QUAD 0x40000 /* 256k per quad. */ +#include #define BUS2QUAD(global) (mp_bus_id_to_node[global]) @@ -18,8 +16,6 @@ #define QUADLOCAL2BUS(quad,local) (quad_local_to_mp_bus_id[quad][local]) -#define XQUAD_PORT_ADDR(port, quad) (xquad_portio + (XQUAD_PORTIO_QUAD*quad) + port) - #define PCI_CONF1_MQ_ADDRESS(bus, devfn, reg) \ (0x80000000 | (BUS2LOCAL(bus) << 16) | (devfn << 8) | (reg & ~3)) -- cgit v0.10.2 From bd2984e96452855d148ebce76f696dcecbc96340 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 5 Feb 2010 09:37:04 -0500 Subject: x86-32: Remove _local variants of in/out from io_32.h These were leftover from the numaq support that was removed in commit 1fba38703d0ce8a5ff0fad9df3eccc6b55cf2cfb. Signed-off-by: Brian Gerst LKML-Reference: <1265380629-3212-3-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/io_32.h b/arch/x86/include/asm/io_32.h index e16b9db..72a6a4a 100644 --- a/arch/x86/include/asm/io_32.h +++ b/arch/x86/include/asm/io_32.h @@ -120,47 +120,21 @@ static inline void slow_down_io(void) #endif -#define __BUILDIO(bwl, bw, type) \ -static inline void out##bwl(unsigned type value, int port) \ -{ \ - out##bwl##_local(value, port); \ -} \ - \ -static inline unsigned type in##bwl(int port) \ -{ \ - return in##bwl##_local(port); \ -} - #define BUILDIO(bwl, bw, type) \ -static inline void out##bwl##_local(unsigned type value, int port) \ +static inline void out##bwl(unsigned type value, int port) \ { \ - asm volatile("out" #bwl " %" #bw "0, %w1" \ + asm volatile("out" #bwl " %" #bw "0, %w1" \ : : "a"(value), "Nd"(port)); \ } \ \ -static inline unsigned type in##bwl##_local(int port) \ +static inline unsigned type in##bwl(int port) \ { \ unsigned type value; \ - asm volatile("in" #bwl " %w1, %" #bw "0" \ + asm volatile("in" #bwl " %w1, %" #bw "0" \ : "=a"(value) : "Nd"(port)); \ return value; \ } \ \ -static inline void out##bwl##_local_p(unsigned type value, int port) \ -{ \ - out##bwl##_local(value, port); \ - slow_down_io(); \ -} \ - \ -static inline unsigned type in##bwl##_local_p(int port) \ -{ \ - unsigned type value = in##bwl##_local(port); \ - slow_down_io(); \ - return value; \ -} \ - \ -__BUILDIO(bwl, bw, type) \ - \ static inline void out##bwl##_p(unsigned type value, int port) \ { \ out##bwl(value, port); \ -- cgit v0.10.2 From 2e16fc7728a77755b5b2dc6b27dde62cd97b9ea5 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 5 Feb 2010 09:37:05 -0500 Subject: x86-64: Reorganize io_64.h Make it more similar to io_32.h. No real code changes. Signed-off-by: Brian Gerst LKML-Reference: <1265380629-3212-4-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/io_64.h b/arch/x86/include/asm/io_64.h index 2440678..040bf74 100644 --- a/arch/x86/include/asm/io_64.h +++ b/arch/x86/include/asm/io_64.h @@ -35,6 +35,54 @@ * - Arnaldo Carvalho de Melo */ +#ifdef __KERNEL__ + +#include + +#include + +/* + * Convert a virtual cached pointer to an uncached pointer + */ +#define xlate_dev_kmem_ptr(p) p + +void memset_io(volatile void __iomem *a, int b, size_t c); + +void __memcpy_fromio(void *, unsigned long, unsigned); +static inline void memcpy_fromio(void *to, const volatile void __iomem *from, + unsigned len) +{ + __memcpy_fromio(to, (unsigned long)from, len); +} + +void __memcpy_toio(unsigned long, const void *, unsigned); +static inline void memcpy_toio(volatile void __iomem *to, const void *from, + unsigned len) +{ + __memcpy_toio((unsigned long)to, from, len); +} + +/* + * ISA space is 'always mapped' on a typical x86 system, no need to + * explicitly ioremap() it. The fact that the ISA IO space is mapped + * to PAGE_OFFSET is pure coincidence - it does not mean ISA values + * are physical addresses. The following constant pointer can be + * used as the IO-area pointer (it can be iounmapped as well, so the + * analogy with PCI is quite large): + */ +#define __ISA_IO_base ((char __iomem *)(PAGE_OFFSET)) + +/* + * Cache management + * + * This needed for two cases + * 1. Out of order aware processors + * 2. Accidentally out of order processors (PPro errata #51) + */ +#define flush_write_buffers() do { } while (0) + +#endif /* __KERNEL__ */ + extern void native_io_delay(void); extern int io_delay_type; @@ -53,6 +101,7 @@ static inline void slow_down_io(void) native_io_delay(); #endif } + #endif /* @@ -136,46 +185,5 @@ __OUTS(b) __OUTS(w) __OUTS(l) -#if defined(__KERNEL__) && defined(__x86_64__) - -#include - -#include - -void __memcpy_fromio(void *, unsigned long, unsigned); -void __memcpy_toio(unsigned long, const void *, unsigned); - -static inline void memcpy_fromio(void *to, const volatile void __iomem *from, - unsigned len) -{ - __memcpy_fromio(to, (unsigned long)from, len); -} - -static inline void memcpy_toio(volatile void __iomem *to, const void *from, - unsigned len) -{ - __memcpy_toio((unsigned long)to, from, len); -} - -void memset_io(volatile void __iomem *a, int b, size_t c); - -/* - * ISA space is 'always mapped' on a typical x86 system, no need to - * explicitly ioremap() it. The fact that the ISA IO space is mapped - * to PAGE_OFFSET is pure coincidence - it does not mean ISA values - * are physical addresses. The following constant pointer can be - * used as the IO-area pointer (it can be iounmapped as well, so the - * analogy with PCI is quite large): - */ -#define __ISA_IO_base ((char __iomem *)(PAGE_OFFSET)) - -#define flush_write_buffers() - -/* - * Convert a virtual cached pointer to an uncached pointer - */ -#define xlate_dev_kmem_ptr(p) p - -#endif /* __KERNEL__ */ #endif /* _ASM_X86_IO_64_H */ -- cgit v0.10.2 From 2b4df4d4f7de1a834d252c7da3197fce634cbf0e Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 5 Feb 2010 09:37:06 -0500 Subject: x86-64: Use BUILDIO in io_64.h Copied from io_32.h. Signed-off-by: Brian Gerst LKML-Reference: <1265380629-3212-5-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/io_64.h b/arch/x86/include/asm/io_64.h index 040bf74..4a94aef 100644 --- a/arch/x86/include/asm/io_64.h +++ b/arch/x86/include/asm/io_64.h @@ -104,86 +104,48 @@ static inline void slow_down_io(void) #endif -/* - * Talk about misusing macros.. - */ -#define __OUT1(s, x) \ -static inline void out##s(unsigned x value, unsigned short port) { - -#define __OUT2(s, s1, s2) \ -asm volatile ("out" #s " %" s1 "0,%" s2 "1" - -#ifndef REALLY_SLOW_IO -#define REALLY_SLOW_IO -#define UNSET_REALLY_SLOW_IO -#endif - -#define __OUT(s, s1, x) \ - __OUT1(s, x) __OUT2(s, s1, "w") : : "a" (value), "Nd" (port)); \ - } \ - __OUT1(s##_p, x) __OUT2(s, s1, "w") : : "a" (value), "Nd" (port)); \ +#define BUILDIO(bwl, bw, type) \ +static inline void out##bwl(unsigned type value, int port) \ +{ \ + asm volatile("out" #bwl " %" #bw "0, %w1" \ + : : "a"(value), "Nd"(port)); \ +} \ + \ +static inline unsigned type in##bwl(int port) \ +{ \ + unsigned type value; \ + asm volatile("in" #bwl " %w1, %" #bw "0" \ + : "=a"(value) : "Nd"(port)); \ + return value; \ +} \ + \ +static inline void out##bwl##_p(unsigned type value, int port) \ +{ \ + out##bwl(value, port); \ slow_down_io(); \ -} - -#define __IN1(s) \ -static inline RETURN_TYPE in##s(unsigned short port) \ +} \ + \ +static inline unsigned type in##bwl##_p(int port) \ { \ - RETURN_TYPE _v; - -#define __IN2(s, s1, s2) \ - asm volatile ("in" #s " %" s2 "1,%" s1 "0" - -#define __IN(s, s1, i...) \ - __IN1(s) __IN2(s, s1, "w") : "=a" (_v) : "Nd" (port), ##i); \ - return _v; \ - } \ - __IN1(s##_p) __IN2(s, s1, "w") : "=a" (_v) : "Nd" (port), ##i); \ - slow_down_io(); \ - return _v; } - -#ifdef UNSET_REALLY_SLOW_IO -#undef REALLY_SLOW_IO -#endif - -#define __INS(s) \ -static inline void ins##s(unsigned short port, void *addr, \ - unsigned long count) \ + unsigned type value = in##bwl(port); \ + slow_down_io(); \ + return value; \ +} \ + \ +static inline void outs##bwl(int port, const void *addr, unsigned long count) \ { \ - asm volatile ("rep ; ins" #s \ - : "=D" (addr), "=c" (count) \ - : "d" (port), "0" (addr), "1" (count)); \ -} - -#define __OUTS(s) \ -static inline void outs##s(unsigned short port, const void *addr, \ - unsigned long count) \ + asm volatile("rep; outs" #bwl \ + : "+S"(addr), "+c"(count) : "d"(port)); \ +} \ + \ +static inline void ins##bwl(int port, void *addr, unsigned long count) \ { \ - asm volatile ("rep ; outs" #s \ - : "=S" (addr), "=c" (count) \ - : "d" (port), "0" (addr), "1" (count)); \ + asm volatile("rep; ins" #bwl \ + : "+D"(addr), "+c"(count) : "d"(port)); \ } -#define RETURN_TYPE unsigned char -__IN(b, "") -#undef RETURN_TYPE -#define RETURN_TYPE unsigned short -__IN(w, "") -#undef RETURN_TYPE -#define RETURN_TYPE unsigned int -__IN(l, "") -#undef RETURN_TYPE - -__OUT(b, "b", char) -__OUT(w, "w", short) -__OUT(l, , int) - -__INS(b) -__INS(w) -__INS(l) - -__OUTS(b) -__OUTS(w) -__OUTS(l) - +BUILDIO(b, b, char) +BUILDIO(w, w, short) +BUILDIO(l, , int) #endif /* _ASM_X86_IO_64_H */ -- cgit v0.10.2 From 6175ddf06b6172046a329e3abfd9c901a43efd2e Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 5 Feb 2010 09:37:07 -0500 Subject: x86: Clean up mem*io functions. Iomem has no special significance on x86. Use the standard mem* functions instead of trying to call other versions. Some fixups are needed to match the function prototypes. Signed-off-by: Brian Gerst LKML-Reference: <1265380629-3212-6-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/boot/compressed/misc.c b/arch/x86/boot/compressed/misc.c index 3b22fe8..88042e8 100644 --- a/arch/x86/boot/compressed/misc.c +++ b/arch/x86/boot/compressed/misc.c @@ -19,11 +19,6 @@ #define _ASM_X86_DESC_H 1 #endif -#ifdef CONFIG_X86_64 -#define _LINUX_STRING_H_ 1 -#define __LINUX_BITMAP_H 1 -#endif - #include #include #include @@ -131,8 +126,8 @@ static void error(char *m); static struct boot_params *real_mode; /* Pointer to real-mode data */ static int quiet; -static void *memset(void *s, int c, unsigned n); -void *memcpy(void *dest, const void *src, unsigned n); +void *memset(void *s, int c, size_t n); +void *memcpy(void *dest, const void *src, size_t n); static void __putstr(int, const char *); #define putstr(__x) __putstr(0, __x) @@ -223,7 +218,7 @@ static void __putstr(int error, const char *s) outb(0xff & (pos >> 1), vidport+1); } -static void *memset(void *s, int c, unsigned n) +void *memset(void *s, int c, size_t n) { int i; char *ss = s; @@ -233,7 +228,7 @@ static void *memset(void *s, int c, unsigned n) return s; } -void *memcpy(void *dest, const void *src, unsigned n) +void *memcpy(void *dest, const void *src, size_t n) { int i; const char *s = src; diff --git a/arch/x86/include/asm/io_32.h b/arch/x86/include/asm/io_32.h index 72a6a4a..685e332 100644 --- a/arch/x86/include/asm/io_32.h +++ b/arch/x86/include/asm/io_32.h @@ -49,21 +49,21 @@ #define xlate_dev_kmem_ptr(p) p static inline void -memset_io(volatile void __iomem *addr, unsigned char val, int count) +memset_io(volatile void __iomem *addr, unsigned char val, size_t count) { memset((void __force *)addr, val, count); } static inline void -memcpy_fromio(void *dst, const volatile void __iomem *src, int count) +memcpy_fromio(void *dst, const volatile void __iomem *src, size_t count) { - __memcpy(dst, (const void __force *)src, count); + memcpy(dst, (const void __force *)src, count); } static inline void -memcpy_toio(volatile void __iomem *dst, const void *src, int count) +memcpy_toio(volatile void __iomem *dst, const void *src, size_t count) { - __memcpy((void __force *)dst, src, count); + memcpy((void __force *)dst, src, count); } /* diff --git a/arch/x86/include/asm/io_64.h b/arch/x86/include/asm/io_64.h index 4a94aef..1305525 100644 --- a/arch/x86/include/asm/io_64.h +++ b/arch/x86/include/asm/io_64.h @@ -1,6 +1,8 @@ #ifndef _ASM_X86_IO_64_H #define _ASM_X86_IO_64_H +#include +#include /* * This file contains the definitions for the x86 IO instructions @@ -46,20 +48,22 @@ */ #define xlate_dev_kmem_ptr(p) p -void memset_io(volatile void __iomem *a, int b, size_t c); +static inline void +memset_io(volatile void __iomem *addr, unsigned char val, size_t count) +{ + memset((void __force *)addr, val, count); +} -void __memcpy_fromio(void *, unsigned long, unsigned); -static inline void memcpy_fromio(void *to, const volatile void __iomem *from, - unsigned len) +static inline void +memcpy_fromio(void *dst, const volatile void __iomem *src, size_t count) { - __memcpy_fromio(to, (unsigned long)from, len); + memcpy(dst, (const void __force *)src, count); } -void __memcpy_toio(unsigned long, const void *, unsigned); -static inline void memcpy_toio(volatile void __iomem *to, const void *from, - unsigned len) +static inline void +memcpy_toio(volatile void __iomem *dst, const void *src, size_t count) { - __memcpy_toio((unsigned long)to, from, len); + memcpy((void __force *)dst, src, count); } /* diff --git a/arch/x86/lib/Makefile b/arch/x86/lib/Makefile index cffd754f..fff1427 100644 --- a/arch/x86/lib/Makefile +++ b/arch/x86/lib/Makefile @@ -34,7 +34,7 @@ ifneq ($(CONFIG_X86_CMPXCHG64),y) endif lib-$(CONFIG_X86_USE_3DNOW) += mmx_32.o else - obj-y += io_64.o iomap_copy_64.o + obj-y += iomap_copy_64.o lib-y += csum-partial_64.o csum-copy_64.o csum-wrappers_64.o lib-y += thunk_64.o clear_page_64.o copy_page_64.o lib-y += memmove_64.o memset_64.o diff --git a/arch/x86/lib/io_64.c b/arch/x86/lib/io_64.c deleted file mode 100644 index 3f1eb59..0000000 --- a/arch/x86/lib/io_64.c +++ /dev/null @@ -1,25 +0,0 @@ -#include -#include -#include - -void __memcpy_toio(unsigned long dst, const void *src, unsigned len) -{ - __inline_memcpy((void *)dst, src, len); -} -EXPORT_SYMBOL(__memcpy_toio); - -void __memcpy_fromio(void *dst, unsigned long src, unsigned len) -{ - __inline_memcpy(dst, (const void *)src, len); -} -EXPORT_SYMBOL(__memcpy_fromio); - -void memset_io(volatile void __iomem *a, int b, size_t c) -{ - /* - * TODO: memset can mangle the IO patterns quite a bit. - * perhaps it would be better to use a dumb one: - */ - memset((void *)a, b, c); -} -EXPORT_SYMBOL(memset_io); -- cgit v0.10.2 From 910bf6ad0be3e1efbda0e9d358794937b52c9860 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 5 Feb 2010 09:37:08 -0500 Subject: x86: Simplify flush_write_buffers() Always make it an inline instead of using a macro for the no-op case. Signed-off-by: Brian Gerst LKML-Reference: <1265380629-3212-7-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/io_32.h b/arch/x86/include/asm/io_32.h index 685e332..e8177f3 100644 --- a/arch/x86/include/asm/io_32.h +++ b/arch/x86/include/asm/io_32.h @@ -84,18 +84,12 @@ memcpy_toio(volatile void __iomem *dst, const void *src, size_t count) * 2. Accidentally out of order processors (PPro errata #51) */ -#if defined(CONFIG_X86_OOSTORE) || defined(CONFIG_X86_PPRO_FENCE) - static inline void flush_write_buffers(void) { +#if defined(CONFIG_X86_OOSTORE) || defined(CONFIG_X86_PPRO_FENCE) asm volatile("lock; addl $0,0(%%esp)": : :"memory"); -} - -#else - -#define flush_write_buffers() do { } while (0) - #endif +} #endif /* __KERNEL__ */ diff --git a/arch/x86/include/asm/io_64.h b/arch/x86/include/asm/io_64.h index 1305525..6964a1c 100644 --- a/arch/x86/include/asm/io_64.h +++ b/arch/x86/include/asm/io_64.h @@ -83,7 +83,13 @@ memcpy_toio(volatile void __iomem *dst, const void *src, size_t count) * 1. Out of order aware processors * 2. Accidentally out of order processors (PPro errata #51) */ -#define flush_write_buffers() do { } while (0) + +static inline void flush_write_buffers(void) +{ +#if defined(CONFIG_X86_OOSTORE) || defined(CONFIG_X86_PPRO_FENCE) + asm volatile("lock; addl $0,0(%%esp)": : :"memory"); +#endif +} #endif /* __KERNEL__ */ -- cgit v0.10.2 From 1c5b9069e12e20d2fe883076ae0bf73966492108 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Fri, 5 Feb 2010 09:37:09 -0500 Subject: x86: Merge io.h io_32.h and io_64.h are now identical. Merge them into io.h. Signed-off-by: Brian Gerst LKML-Reference: <1265380629-3212-8-git-send-email-brgerst@gmail.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h index 7373932..a1dcfa3 100644 --- a/arch/x86/include/asm/io.h +++ b/arch/x86/include/asm/io.h @@ -1,8 +1,42 @@ #ifndef _ASM_X86_IO_H #define _ASM_X86_IO_H +/* + * This file contains the definitions for the x86 IO instructions + * inb/inw/inl/outb/outw/outl and the "string versions" of the same + * (insb/insw/insl/outsb/outsw/outsl). You can also use "pausing" + * versions of the single-IO instructions (inb_p/inw_p/..). + * + * This file is not meant to be obfuscating: it's just complicated + * to (a) handle it all in a way that makes gcc able to optimize it + * as well as possible and (b) trying to avoid writing the same thing + * over and over again with slight variations and possibly making a + * mistake somewhere. + */ + +/* + * Thanks to James van Artsdalen for a better timing-fix than + * the two short jumps: using outb's to a nonexistent port seems + * to guarantee better timings even on fast machines. + * + * On the other hand, I'd like to be sure of a non-existent port: + * I feel a bit unsafe about using 0x80 (should be safe, though) + * + * Linus + */ + + /* + * Bit simplified and optimized by Jan Hubicka + * Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999. + * + * isa_memset_io, isa_memcpy_fromio, isa_memcpy_toio added, + * isa_read[wl] and isa_write[wl] fixed + * - Arnaldo Carvalho de Melo + */ + #define ARCH_HAS_IOREMAP_WC +#include #include #include #include @@ -173,11 +207,126 @@ static inline void __iomem *ioremap(resource_size_t offset, unsigned long size) extern void iounmap(volatile void __iomem *addr); -#ifdef CONFIG_X86_32 -# include "io_32.h" +#ifdef __KERNEL__ + +#include + +#include + +/* + * Convert a virtual cached pointer to an uncached pointer + */ +#define xlate_dev_kmem_ptr(p) p + +static inline void +memset_io(volatile void __iomem *addr, unsigned char val, size_t count) +{ + memset((void __force *)addr, val, count); +} + +static inline void +memcpy_fromio(void *dst, const volatile void __iomem *src, size_t count) +{ + memcpy(dst, (const void __force *)src, count); +} + +static inline void +memcpy_toio(volatile void __iomem *dst, const void *src, size_t count) +{ + memcpy((void __force *)dst, src, count); +} + +/* + * ISA space is 'always mapped' on a typical x86 system, no need to + * explicitly ioremap() it. The fact that the ISA IO space is mapped + * to PAGE_OFFSET is pure coincidence - it does not mean ISA values + * are physical addresses. The following constant pointer can be + * used as the IO-area pointer (it can be iounmapped as well, so the + * analogy with PCI is quite large): + */ +#define __ISA_IO_base ((char __iomem *)(PAGE_OFFSET)) + +/* + * Cache management + * + * This needed for two cases + * 1. Out of order aware processors + * 2. Accidentally out of order processors (PPro errata #51) + */ + +static inline void flush_write_buffers(void) +{ +#if defined(CONFIG_X86_OOSTORE) || defined(CONFIG_X86_PPRO_FENCE) + asm volatile("lock; addl $0,0(%%esp)": : :"memory"); +#endif +} + +#endif /* __KERNEL__ */ + +extern void native_io_delay(void); + +extern int io_delay_type; +extern void io_delay_init(void); + +#if defined(CONFIG_PARAVIRT) +#include #else -# include "io_64.h" + +static inline void slow_down_io(void) +{ + native_io_delay(); +#ifdef REALLY_SLOW_IO + native_io_delay(); + native_io_delay(); + native_io_delay(); #endif +} + +#endif + +#define BUILDIO(bwl, bw, type) \ +static inline void out##bwl(unsigned type value, int port) \ +{ \ + asm volatile("out" #bwl " %" #bw "0, %w1" \ + : : "a"(value), "Nd"(port)); \ +} \ + \ +static inline unsigned type in##bwl(int port) \ +{ \ + unsigned type value; \ + asm volatile("in" #bwl " %w1, %" #bw "0" \ + : "=a"(value) : "Nd"(port)); \ + return value; \ +} \ + \ +static inline void out##bwl##_p(unsigned type value, int port) \ +{ \ + out##bwl(value, port); \ + slow_down_io(); \ +} \ + \ +static inline unsigned type in##bwl##_p(int port) \ +{ \ + unsigned type value = in##bwl(port); \ + slow_down_io(); \ + return value; \ +} \ + \ +static inline void outs##bwl(int port, const void *addr, unsigned long count) \ +{ \ + asm volatile("rep; outs" #bwl \ + : "+S"(addr), "+c"(count) : "d"(port)); \ +} \ + \ +static inline void ins##bwl(int port, void *addr, unsigned long count) \ +{ \ + asm volatile("rep; ins" #bwl \ + : "+D"(addr), "+c"(count) : "d"(port)); \ +} + +BUILDIO(b, b, char) +BUILDIO(w, w, short) +BUILDIO(l, , int) extern void *xlate_dev_mem_ptr(unsigned long phys); extern void unxlate_dev_mem_ptr(unsigned long phys, void *addr); diff --git a/arch/x86/include/asm/io_32.h b/arch/x86/include/asm/io_32.h deleted file mode 100644 index e8177f3..0000000 --- a/arch/x86/include/asm/io_32.h +++ /dev/null @@ -1,161 +0,0 @@ -#ifndef _ASM_X86_IO_32_H -#define _ASM_X86_IO_32_H - -#include -#include - -/* - * This file contains the definitions for the x86 IO instructions - * inb/inw/inl/outb/outw/outl and the "string versions" of the same - * (insb/insw/insl/outsb/outsw/outsl). You can also use "pausing" - * versions of the single-IO instructions (inb_p/inw_p/..). - * - * This file is not meant to be obfuscating: it's just complicated - * to (a) handle it all in a way that makes gcc able to optimize it - * as well as possible and (b) trying to avoid writing the same thing - * over and over again with slight variations and possibly making a - * mistake somewhere. - */ - -/* - * Thanks to James van Artsdalen for a better timing-fix than - * the two short jumps: using outb's to a nonexistent port seems - * to guarantee better timings even on fast machines. - * - * On the other hand, I'd like to be sure of a non-existent port: - * I feel a bit unsafe about using 0x80 (should be safe, though) - * - * Linus - */ - - /* - * Bit simplified and optimized by Jan Hubicka - * Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999. - * - * isa_memset_io, isa_memcpy_fromio, isa_memcpy_toio added, - * isa_read[wl] and isa_write[wl] fixed - * - Arnaldo Carvalho de Melo - */ - -#ifdef __KERNEL__ - -#include - -#include - -/* - * Convert a virtual cached pointer to an uncached pointer - */ -#define xlate_dev_kmem_ptr(p) p - -static inline void -memset_io(volatile void __iomem *addr, unsigned char val, size_t count) -{ - memset((void __force *)addr, val, count); -} - -static inline void -memcpy_fromio(void *dst, const volatile void __iomem *src, size_t count) -{ - memcpy(dst, (const void __force *)src, count); -} - -static inline void -memcpy_toio(volatile void __iomem *dst, const void *src, size_t count) -{ - memcpy((void __force *)dst, src, count); -} - -/* - * ISA space is 'always mapped' on a typical x86 system, no need to - * explicitly ioremap() it. The fact that the ISA IO space is mapped - * to PAGE_OFFSET is pure coincidence - it does not mean ISA values - * are physical addresses. The following constant pointer can be - * used as the IO-area pointer (it can be iounmapped as well, so the - * analogy with PCI is quite large): - */ -#define __ISA_IO_base ((char __iomem *)(PAGE_OFFSET)) - -/* - * Cache management - * - * This needed for two cases - * 1. Out of order aware processors - * 2. Accidentally out of order processors (PPro errata #51) - */ - -static inline void flush_write_buffers(void) -{ -#if defined(CONFIG_X86_OOSTORE) || defined(CONFIG_X86_PPRO_FENCE) - asm volatile("lock; addl $0,0(%%esp)": : :"memory"); -#endif -} - -#endif /* __KERNEL__ */ - -extern void native_io_delay(void); - -extern int io_delay_type; -extern void io_delay_init(void); - -#if defined(CONFIG_PARAVIRT) -#include -#else - -static inline void slow_down_io(void) -{ - native_io_delay(); -#ifdef REALLY_SLOW_IO - native_io_delay(); - native_io_delay(); - native_io_delay(); -#endif -} - -#endif - -#define BUILDIO(bwl, bw, type) \ -static inline void out##bwl(unsigned type value, int port) \ -{ \ - asm volatile("out" #bwl " %" #bw "0, %w1" \ - : : "a"(value), "Nd"(port)); \ -} \ - \ -static inline unsigned type in##bwl(int port) \ -{ \ - unsigned type value; \ - asm volatile("in" #bwl " %w1, %" #bw "0" \ - : "=a"(value) : "Nd"(port)); \ - return value; \ -} \ - \ -static inline void out##bwl##_p(unsigned type value, int port) \ -{ \ - out##bwl(value, port); \ - slow_down_io(); \ -} \ - \ -static inline unsigned type in##bwl##_p(int port) \ -{ \ - unsigned type value = in##bwl(port); \ - slow_down_io(); \ - return value; \ -} \ - \ -static inline void outs##bwl(int port, const void *addr, unsigned long count) \ -{ \ - asm volatile("rep; outs" #bwl \ - : "+S"(addr), "+c"(count) : "d"(port)); \ -} \ - \ -static inline void ins##bwl(int port, void *addr, unsigned long count) \ -{ \ - asm volatile("rep; ins" #bwl \ - : "+D"(addr), "+c"(count) : "d"(port)); \ -} - -BUILDIO(b, b, char) -BUILDIO(w, w, short) -BUILDIO(l, , int) - -#endif /* _ASM_X86_IO_32_H */ diff --git a/arch/x86/include/asm/io_64.h b/arch/x86/include/asm/io_64.h deleted file mode 100644 index 6964a1c..0000000 --- a/arch/x86/include/asm/io_64.h +++ /dev/null @@ -1,161 +0,0 @@ -#ifndef _ASM_X86_IO_64_H -#define _ASM_X86_IO_64_H - -#include -#include - -/* - * This file contains the definitions for the x86 IO instructions - * inb/inw/inl/outb/outw/outl and the "string versions" of the same - * (insb/insw/insl/outsb/outsw/outsl). You can also use "pausing" - * versions of the single-IO instructions (inb_p/inw_p/..). - * - * This file is not meant to be obfuscating: it's just complicated - * to (a) handle it all in a way that makes gcc able to optimize it - * as well as possible and (b) trying to avoid writing the same thing - * over and over again with slight variations and possibly making a - * mistake somewhere. - */ - -/* - * Thanks to James van Artsdalen for a better timing-fix than - * the two short jumps: using outb's to a nonexistent port seems - * to guarantee better timings even on fast machines. - * - * On the other hand, I'd like to be sure of a non-existent port: - * I feel a bit unsafe about using 0x80 (should be safe, though) - * - * Linus - */ - - /* - * Bit simplified and optimized by Jan Hubicka - * Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999. - * - * isa_memset_io, isa_memcpy_fromio, isa_memcpy_toio added, - * isa_read[wl] and isa_write[wl] fixed - * - Arnaldo Carvalho de Melo - */ - -#ifdef __KERNEL__ - -#include - -#include - -/* - * Convert a virtual cached pointer to an uncached pointer - */ -#define xlate_dev_kmem_ptr(p) p - -static inline void -memset_io(volatile void __iomem *addr, unsigned char val, size_t count) -{ - memset((void __force *)addr, val, count); -} - -static inline void -memcpy_fromio(void *dst, const volatile void __iomem *src, size_t count) -{ - memcpy(dst, (const void __force *)src, count); -} - -static inline void -memcpy_toio(volatile void __iomem *dst, const void *src, size_t count) -{ - memcpy((void __force *)dst, src, count); -} - -/* - * ISA space is 'always mapped' on a typical x86 system, no need to - * explicitly ioremap() it. The fact that the ISA IO space is mapped - * to PAGE_OFFSET is pure coincidence - it does not mean ISA values - * are physical addresses. The following constant pointer can be - * used as the IO-area pointer (it can be iounmapped as well, so the - * analogy with PCI is quite large): - */ -#define __ISA_IO_base ((char __iomem *)(PAGE_OFFSET)) - -/* - * Cache management - * - * This needed for two cases - * 1. Out of order aware processors - * 2. Accidentally out of order processors (PPro errata #51) - */ - -static inline void flush_write_buffers(void) -{ -#if defined(CONFIG_X86_OOSTORE) || defined(CONFIG_X86_PPRO_FENCE) - asm volatile("lock; addl $0,0(%%esp)": : :"memory"); -#endif -} - -#endif /* __KERNEL__ */ - -extern void native_io_delay(void); - -extern int io_delay_type; -extern void io_delay_init(void); - -#if defined(CONFIG_PARAVIRT) -#include -#else - -static inline void slow_down_io(void) -{ - native_io_delay(); -#ifdef REALLY_SLOW_IO - native_io_delay(); - native_io_delay(); - native_io_delay(); -#endif -} - -#endif - -#define BUILDIO(bwl, bw, type) \ -static inline void out##bwl(unsigned type value, int port) \ -{ \ - asm volatile("out" #bwl " %" #bw "0, %w1" \ - : : "a"(value), "Nd"(port)); \ -} \ - \ -static inline unsigned type in##bwl(int port) \ -{ \ - unsigned type value; \ - asm volatile("in" #bwl " %w1, %" #bw "0" \ - : "=a"(value) : "Nd"(port)); \ - return value; \ -} \ - \ -static inline void out##bwl##_p(unsigned type value, int port) \ -{ \ - out##bwl(value, port); \ - slow_down_io(); \ -} \ - \ -static inline unsigned type in##bwl##_p(int port) \ -{ \ - unsigned type value = in##bwl(port); \ - slow_down_io(); \ - return value; \ -} \ - \ -static inline void outs##bwl(int port, const void *addr, unsigned long count) \ -{ \ - asm volatile("rep; outs" #bwl \ - : "+S"(addr), "+c"(count) : "d"(port)); \ -} \ - \ -static inline void ins##bwl(int port, void *addr, unsigned long count) \ -{ \ - asm volatile("rep; ins" #bwl \ - : "+D"(addr), "+c"(count) : "d"(port)); \ -} - -BUILDIO(b, b, char) -BUILDIO(w, w, short) -BUILDIO(l, , int) - -#endif /* _ASM_X86_IO_64_H */ -- cgit v0.10.2 From 95a8b6efc5d07103583f706c8a5889437d537939 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 2 Feb 2010 14:38:13 -0800 Subject: pci: Update pci_set_vga_state() to call arch functions Update pci_set_vga_state to call arch dependent functions to enable Legacy VGA I/O transactions to be redirected to correct target. [akpm@linux-foundation.org: make pci_register_set_vga_state() __init] Signed-off-by: Mike Travis LKML-Reference: <201002022238.o12McE1J018723@imap1.linux-foundation.org> Cc: Thomas Gleixner Cc: Robin Holt Cc: Jack Steiner Cc: Ingo Molnar Cc: Jesse Barnes Cc: David Airlie Signed-off-by: Andrew Morton Signed-off-by: H. Peter Anvin diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 315fea4..ac2a576 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -2615,6 +2615,23 @@ int pci_resource_bar(struct pci_dev *dev, int resno, enum pci_bar_type *type) return 0; } +/* Some architectures require additional programming to enable VGA */ +static arch_set_vga_state_t arch_set_vga_state; + +void __init pci_register_set_vga_state(arch_set_vga_state_t func) +{ + arch_set_vga_state = func; /* NULL disables */ +} + +static int pci_set_vga_state_arch(struct pci_dev *dev, bool decode, + unsigned int command_bits, bool change_bridge) +{ + if (arch_set_vga_state) + return arch_set_vga_state(dev, decode, command_bits, + change_bridge); + return 0; +} + /** * pci_set_vga_state - set VGA decode state on device and parents if requested * @dev: the PCI device @@ -2628,9 +2645,15 @@ int pci_set_vga_state(struct pci_dev *dev, bool decode, struct pci_bus *bus; struct pci_dev *bridge; u16 cmd; + int rc; WARN_ON(command_bits & ~(PCI_COMMAND_IO|PCI_COMMAND_MEMORY)); + /* ARCH specific VGA enables */ + rc = pci_set_vga_state_arch(dev, decode, command_bits, change_bridge); + if (rc) + return rc; + pci_read_config_word(dev, PCI_COMMAND, &cmd); if (decode == true) cmd |= command_bits; @@ -2845,6 +2868,7 @@ EXPORT_SYMBOL(pcim_pin_device); EXPORT_SYMBOL(pci_disable_device); EXPORT_SYMBOL(pci_find_capability); EXPORT_SYMBOL(pci_bus_find_capability); +EXPORT_SYMBOL(pci_register_set_vga_state); EXPORT_SYMBOL(pci_release_regions); EXPORT_SYMBOL(pci_request_regions); EXPORT_SYMBOL(pci_request_regions_exclusive); @@ -2877,4 +2901,3 @@ EXPORT_SYMBOL(pci_target_state); EXPORT_SYMBOL(pci_prepare_to_sleep); EXPORT_SYMBOL(pci_back_from_sleep); EXPORT_SYMBOL_GPL(pci_set_pcie_reset_state); - diff --git a/include/linux/pci.h b/include/linux/pci.h index 174e539..5da166e 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -955,6 +955,11 @@ static inline int pci_proc_domain(struct pci_bus *bus) } #endif /* CONFIG_PCI_DOMAINS */ +/* some architectures require additional setup to direct VGA traffic */ +typedef int (*arch_set_vga_state_t)(struct pci_dev *pdev, bool decode, + unsigned int command_bits, bool change_bridge); +extern void pci_register_set_vga_state(arch_set_vga_state_t func); + #else /* CONFIG_PCI is not enabled */ /* -- cgit v0.10.2 From 841582ea9e29a8f757c30c5377ce649586ba793a Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 2 Feb 2010 14:38:14 -0800 Subject: x86, uv: Update UV arch to target Legacy VGA I/O correctly. Add function to direct Legacy VGA I/O traffic to correct I/O Hub. Signed-off-by: Mike Travis LKML-Reference: <201002022238.o12McEbi018727@imap1.linux-foundation.org> Cc: Thomas Gleixner Cc: Robin Holt Cc: Jack Steiner Cc: Ingo Molnar Cc: Jesse Barnes Cc: David Airlie Signed-off-by: Andrew Morton Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/uv/bios.h b/arch/x86/include/asm/uv/bios.h index 2751f30..1634275 100644 --- a/arch/x86/include/asm/uv/bios.h +++ b/arch/x86/include/asm/uv/bios.h @@ -36,7 +36,8 @@ enum uv_bios_cmd { UV_BIOS_WATCHLIST_ALLOC, UV_BIOS_WATCHLIST_FREE, UV_BIOS_MEMPROTECT, - UV_BIOS_GET_PARTITION_ADDR + UV_BIOS_GET_PARTITION_ADDR, + UV_BIOS_SET_LEGACY_VGA_TARGET }; /* @@ -96,6 +97,7 @@ extern int uv_bios_mq_watchlist_alloc(unsigned long, unsigned int, extern int uv_bios_mq_watchlist_free(int, int); extern s64 uv_bios_change_memprotect(u64, u64, enum uv_memprotect); extern s64 uv_bios_reserved_page_pa(u64, u64 *, u64 *, u64 *); +extern int uv_bios_set_legacy_vga_target(bool decode, int domain, int bus); extern void uv_bios_init(void); diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index 21db3cb..6ef2899 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -20,6 +20,7 @@ #include #include #include +#include #include #include @@ -34,6 +35,8 @@ DEFINE_PER_CPU(int, x2apic_extra_bits); +#define PR_DEVEL(fmt, args...) pr_devel("%s: " fmt, __func__, args) + static enum uv_system_type uv_system_type; static u64 gru_start_paddr, gru_end_paddr; int uv_min_hub_revision_id; @@ -553,6 +556,30 @@ late_initcall(uv_init_heartbeat); #endif /* !CONFIG_HOTPLUG_CPU */ +/* Direct Legacy VGA I/O traffic to designated IOH */ +int uv_set_vga_state(struct pci_dev *pdev, bool decode, + unsigned int command_bits, bool change_bridge) +{ + int domain, bus, rc; + + PR_DEVEL("devfn %x decode %d cmd %x chg_brdg %d\n", + pdev->devfn, decode, command_bits, change_bridge); + + if (!change_bridge) + return 0; + + if ((command_bits & PCI_COMMAND_IO) == 0) + return 0; + + domain = pci_domain_nr(pdev->bus); + bus = pdev->bus->number; + + rc = uv_bios_set_legacy_vga_target(decode, domain, bus); + PR_DEVEL("vga decode %d %x:%x, rc: %d\n", decode, domain, bus, rc); + + return rc; +} + /* * Called on each cpu to initialize the per_cpu UV data area. * FIXME: hotplug not supported yet @@ -691,4 +718,7 @@ void __init uv_system_init(void) uv_cpu_init(); uv_scir_register_cpu_notifier(); proc_mkdir("sgi_uv", NULL); + + /* register Legacy VGA I/O redirection handler */ + pci_register_set_vga_state(uv_set_vga_state); } diff --git a/arch/x86/kernel/bios_uv.c b/arch/x86/kernel/bios_uv.c index b0206a2..575127a 100644 --- a/arch/x86/kernel/bios_uv.c +++ b/arch/x86/kernel/bios_uv.c @@ -154,6 +154,25 @@ s64 uv_bios_freq_base(u64 clock_type, u64 *ticks_per_second) } EXPORT_SYMBOL_GPL(uv_bios_freq_base); +/* + * uv_bios_set_legacy_vga_target - Set Legacy VGA I/O Target + * @decode: true to enable target, false to disable target + * @domain: PCI domain number + * @bus: PCI bus number + * + * Returns: + * 0: Success + * -EINVAL: Invalid domain or bus number + * -ENOSYS: Capability not available + * -EBUSY: Legacy VGA I/O cannot be retargeted at this time + */ +int uv_bios_set_legacy_vga_target(bool decode, int domain, int bus) +{ + return uv_bios_call(UV_BIOS_SET_LEGACY_VGA_TARGET, + (u64)decode, (u64)domain, (u64)bus, 0, 0); +} +EXPORT_SYMBOL_GPL(uv_bios_set_legacy_vga_target); + #ifdef CONFIG_EFI void uv_bios_init(void) -- cgit v0.10.2 From 773a38dbdad03474c5ee235f7d9bf9f51c9e3c2b Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 2 Feb 2010 14:38:15 -0800 Subject: vgaarb: Fix VGA arbiter to accept PCI domains other than 0 Update the VGA Arbiter to accept PCI Domains other than 0. Signed-off-by: Mike Travis LKML-Reference: <201002022238.o12McFe8018730@imap1.linux-foundation.org> Cc: Thomas Gleixner Cc: Robin Holt Cc: Jack Steiner Cc: Ingo Molnar Cc: Jesse Barnes Cc: David Airlie Signed-off-by: Andrew Morton Signed-off-by: H. Peter Anvin diff --git a/drivers/gpu/vga/vgaarb.c b/drivers/gpu/vga/vgaarb.c index 1ac0c93..ba7fa9e 100644 --- a/drivers/gpu/vga/vgaarb.c +++ b/drivers/gpu/vga/vgaarb.c @@ -954,6 +954,7 @@ static ssize_t vga_arb_write(struct file *file, const char __user * buf, } } else if (strncmp(curr_pos, "target ", 7) == 0) { + struct pci_bus *pbus; unsigned int domain, bus, devfn; struct vga_device *vgadev; @@ -969,18 +970,31 @@ static ssize_t vga_arb_write(struct file *file, const char __user * buf, ret_val = -EPROTO; goto done; } - - pdev = pci_get_bus_and_slot(bus, devfn); + pr_devel("vgaarb: %s ==> %x:%x:%x.%x\n", curr_pos, + domain, bus, PCI_SLOT(devfn), PCI_FUNC(devfn)); + + pbus = pci_find_bus(domain, bus); + pr_devel("vgaarb: pbus %p\n", pbus); + if (pbus == NULL) { + pr_err("vgaarb: invalid PCI domain and/or bus address %x:%x\n", + domain, bus); + ret_val = -ENODEV; + goto done; + } + pdev = pci_get_slot(pbus, devfn); + pr_devel("vgaarb: pdev %p\n", pdev); if (!pdev) { - pr_info("vgaarb: invalid PCI address!\n"); + pr_err("vgaarb: invalid PCI address %x:%x\n", + bus, devfn); ret_val = -ENODEV; goto done; } } vgadev = vgadev_find(pdev); + pr_devel("vgaarb: vgadev %p\n", vgadev); if (vgadev == NULL) { - pr_info("vgaarb: this pci device is not a vga device\n"); + pr_err("vgaarb: this pci device is not a vga device\n"); pci_dev_put(pdev); ret_val = -ENODEV; goto done; @@ -998,7 +1012,8 @@ static ssize_t vga_arb_write(struct file *file, const char __user * buf, } } if (i == MAX_USER_CARDS) { - pr_err("vgaarb: maximum user cards number reached!\n"); + pr_err("vgaarb: maximum user cards (%d) number reached!\n", + MAX_USER_CARDS); pci_dev_put(pdev); /* XXX: which value to return? */ ret_val = -ENOMEM; -- cgit v0.10.2 From 36028f3383872eefb558a4aae4c12ec2b5fa640f Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Tue, 2 Feb 2010 17:45:01 -0800 Subject: vgaarb: Add user selectability of the number of GPUS in a system Update the VGA Arbiter to allow the user to select the number of GPU's supported in a system. v2: simplify setting of MAX_USER_CARDS, revert back to original default of 16 Signed-off-by: Mike Travis LKML-Reference: <4B68D51D.6090401@sgi.com> Cc: Thomas Gleixner Cc: Robin Holt Cc: Jack Steiner Cc: Ingo Molnar Cc: Jesse Barnes Cc: David Airlie Signed-off-by: Andrew Morton Signed-off-by: H. Peter Anvin diff --git a/drivers/gpu/vga/Kconfig b/drivers/gpu/vga/Kconfig index 790e675..0920492 100644 --- a/drivers/gpu/vga/Kconfig +++ b/drivers/gpu/vga/Kconfig @@ -8,3 +8,11 @@ config VGA_ARB are accessed at same time they need some kind of coordination. Please see Documentation/vgaarbiter.txt for more details. Select this to enable VGA arbiter. + +config VGA_ARB_MAX_GPUS + int "Maximum number of GPUs" + default 16 + depends on VGA_ARB + help + Reserves space in the kernel to maintain resource locking for + multiple GPUS. The overhead for each GPU is very small. diff --git a/drivers/gpu/vga/vgaarb.c b/drivers/gpu/vga/vgaarb.c index ba7fa9e..325a533 100644 --- a/drivers/gpu/vga/vgaarb.c +++ b/drivers/gpu/vga/vgaarb.c @@ -688,7 +688,7 @@ EXPORT_SYMBOL(vga_client_register); * the arbiter. */ -#define MAX_USER_CARDS 16 +#define MAX_USER_CARDS CONFIG_VGA_ARB_MAX_GPUS #define PCI_INVALID_CARD ((struct pci_dev *)-1UL) /* -- cgit v0.10.2 From ee11b90b12eb1ec25e1044bac861e90bfd19ec9e Mon Sep 17 00:00:00 2001 From: Kirill Smelkov Date: Sun, 7 Feb 2010 11:46:15 -0200 Subject: perf top: Fix annotate for userspace First, for programs and prelinked libraries, annotate code was fooled by objdump output IPs (src->eip in the code) being wrongly converted to absolute IPs. In such case there were no conversion needed, but in src->eip = strtoull(src->line, NULL, 16); src->eip = map->unmap_ip(map, src->eip); // = eip + map->start - map->pgoff we were reading absolute address from objdump (e.g. 8048604) and then almost doubling it, because eip & map->start are approximately close for small programs. Needless to say, that later, in record_precise_ip() there was no matching with real runtime IPs. And second, like with `perf annotate` the problem with non-prelinked *.so was that we were doing rip -> objdump address conversion wrong. Also, because unlike `perf annotate`, `perf top` code does annotation based on absolute IPs for performance reasons(*), new helper for mapping objdump addresse to IP is introduced. (*) we get samples info in absolute IPs, and since we do lots of hit-testing on absolute IPs at runtime in record_precise_ip(), it's better to convert objdump addresses to IPs once and do no conversion at runtime. I also had to fix how objdump output is parsed (with hardcoded 8/16 characters format, which was inappropriate for ET_DYN dsos with small addresses like '4ac') Also note, that not all objdump output lines has associtated IPs, e.g. look at source lines here: 000004ac : extern "C" int my_strlen(const char *s) 4ac: 55 push %ebp 4ad: 89 e5 mov %esp,%ebp 4af: 83 ec 10 sub $0x10,%esp { int len = 0; 4b2: c7 45 fc 00 00 00 00 movl $0x0,-0x4(%ebp) 4b9: eb 08 jmp 4c3 while (*s) { ++len; 4bb: 83 45 fc 01 addl $0x1,-0x4(%ebp) ++s; 4bf: 83 45 08 01 addl $0x1,0x8(%ebp) So we mark them with eip=0, and ignore such lines in annotate lookup code. Signed-off-by: Kirill Smelkov [ Note: one hunk of this patch was applied by Mike in 57d8188 ] Signed-off-by: Arnaldo Carvalho de Melo Cc: Mike Galbraith LKML-Reference: <1265550376-12665-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index e4156bc..befa57e 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -216,7 +216,7 @@ static void parse_source(struct sym_entry *syme) while (!feof(file)) { struct source_line *src; size_t dummy = 0; - char *c; + char *c, *sep; src = malloc(sizeof(struct source_line)); assert(src != NULL); @@ -235,14 +235,11 @@ static void parse_source(struct sym_entry *syme) *source->lines_tail = src; source->lines_tail = &src->next; - if (strlen(src->line)>8 && src->line[8] == ':') { - src->eip = strtoull(src->line, NULL, 16); - src->eip = map->unmap_ip(map, src->eip); - } - if (strlen(src->line)>8 && src->line[16] == ':') { - src->eip = strtoull(src->line, NULL, 16); - src->eip = map->unmap_ip(map, src->eip); - } + src->eip = strtoull(src->line, &sep, 16); + if (*sep == ':') + src->eip = map__objdump_2ip(map, src->eip); + else /* this line has no ip info (e.g. source line) */ + src->eip = 0; } pclose(file); out_assign: @@ -277,6 +274,9 @@ static void record_precise_ip(struct sym_entry *syme, int counter, u64 ip) goto out_unlock; for (line = syme->src->lines; line; line = line->next) { + /* skip lines without IP info */ + if (line->eip == 0) + continue; if (line->eip == ip) { line->count[counter]++; break; diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index af5805f..138e3cb 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -222,3 +222,11 @@ u64 map__rip_2objdump(struct map *map, u64 rip) rip; return addr; } + +u64 map__objdump_2ip(struct map *map, u64 addr) +{ + u64 ip = map->dso->adjust_symbols ? + addr : + map->unmap_ip(map, addr); /* RIP -> IP */ + return ip; +} diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h index 9cee9c7..86f77cb 100644 --- a/tools/perf/util/map.h +++ b/tools/perf/util/map.h @@ -61,9 +61,9 @@ static inline u64 identity__map_ip(struct map *map __used, u64 ip) } -/* rip -> addr suitable for passing to `objdump --start-address=` */ +/* rip/ip <-> addr suitable for passing to `objdump --start-address=` */ u64 map__rip_2objdump(struct map *map, u64 rip); - +u64 map__objdump_2ip(struct map *map, u64 addr); struct symbol; struct mmap_event; -- cgit v0.10.2 From 5f485364365f00853e5249cb3ae31f876936b552 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 7 Feb 2010 11:46:16 -0200 Subject: perf top: Use address pattern in lookup_sym_source MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Because we may have aliases, like __GI___strcoll_l in /lib64/libc-2.10.2.so that appears in objdump as: $ objdump --start-address=0x0000003715a86420 \ --stop-address=0x0000003715a872dc -dS /lib64/libc-2.10.2.so 0000003715a86420 <__strcoll_l>: 3715a86420: 55 push %rbp 3715a86421: 48 89 e5 mov %rsp,%rbp 3715a86424: 41 57 push %r15 [root@doppio linux-2.6-tip]# So look for the address exactly at the start of the line instead so that annotation can work for in these cases. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Kirill Smelkov Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1265550376-12665-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index befa57e..c72ab50 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -202,10 +202,9 @@ static void parse_source(struct sym_entry *syme) len = sym->end - sym->start; sprintf(command, - "objdump --start-address=0x%016Lx " - "--stop-address=0x%016Lx -dS %s", - map__rip_2objdump(map, sym->start), - map__rip_2objdump(map, sym->end), path); + "objdump --start-address=%#0*Lx --stop-address=%#0*Lx -dS %s", + BITS_PER_LONG / 4, map__rip_2objdump(map, sym->start), + BITS_PER_LONG / 4, map__rip_2objdump(map, sym->end), path); file = popen(command, "r"); if (!file) @@ -292,13 +291,15 @@ static void lookup_sym_source(struct sym_entry *syme) { struct symbol *symbol = sym_entry__symbol(syme); struct source_line *line; - char pattern[PATH_MAX]; + const size_t pattern_len = BITS_PER_LONG / 4 + 2; + char pattern[pattern_len + 1]; - sprintf(pattern, "<%s>:", symbol->name); + sprintf(pattern, "%0*Lx <", BITS_PER_LONG / 4, + map__rip_2objdump(syme->map, symbol->start)); pthread_mutex_lock(&syme->src->lock); for (line = syme->src->lines; line; line = line->next) { - if (strstr(line->line, pattern)) { + if (memcmp(line->line, pattern, pattern_len) == 0) { syme->src->source = line; break; } -- cgit v0.10.2 From 3235dc3f22378f35ce77eba0d0f62db2d9c4844e Mon Sep 17 00:00:00 2001 From: Frans Pop Date: Sat, 6 Feb 2010 18:47:17 +0100 Subject: x86: Remove trailing spaces in messages Signed-off-by: Frans Pop Cc: Avi Kivity Cc: x86@kernel.org LKML-Reference: <1265478443-31072-10-git-send-email-elendil@planet.nl> [ Left out the KVM bits. ] Signed-off-by: Ingo Molnar diff --git a/arch/x86/boot/mkcpustr.c b/arch/x86/boot/mkcpustr.c index 8ef60f2..919257f 100644 --- a/arch/x86/boot/mkcpustr.c +++ b/arch/x86/boot/mkcpustr.c @@ -22,7 +22,7 @@ int main(void) int i, j; const char *str; - printf("static const char x86_cap_strs[] = \n"); + printf("static const char x86_cap_strs[] =\n"); for (i = 0; i < NCAPINTS; i++) { for (j = 0; j < 32; j++) { diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index e80f291..71c4443 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -587,7 +587,7 @@ calibrate_by_pmtimer(long deltapm, long *delta, long *deltatsc) res = (((u64)(*deltatsc)) * pm_100ms); do_div(res, deltapm); apic_printk(APIC_VERBOSE, "TSC delta adjusted to " - "PM-Timer: %lu (%ld) \n", + "PM-Timer: %lu (%ld)\n", (unsigned long)res, *deltatsc); *deltatsc = (long)res; } diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 53243ca..6bdd2c7 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1647,7 +1647,7 @@ __apicdebuginit(void) print_IO_APIC(void) printk(KERN_DEBUG ".... IRQ redirection table:\n"); printk(KERN_DEBUG " NR Dst Mask Trig IRR Pol" - " Stat Dmod Deli Vect: \n"); + " Stat Dmod Deli Vect:\n"); for (i = 0; i <= reg_01.bits.entries; i++) { struct IO_APIC_route_entry entry; diff --git a/arch/x86/kernel/apic/numaq_32.c b/arch/x86/kernel/apic/numaq_32.c index 98c4665..47dd856 100644 --- a/arch/x86/kernel/apic/numaq_32.c +++ b/arch/x86/kernel/apic/numaq_32.c @@ -225,7 +225,7 @@ static void __init smp_read_mpc_oem(struct mpc_table *mpc) mpc_record = 0; printk(KERN_INFO - "Found an OEM MPC table at %8p - parsing it ... \n", oemtable); + "Found an OEM MPC table at %8p - parsing it...\n", oemtable); if (memcmp(oemtable->signature, MPC_OEM_SIGNATURE, 4)) { printk(KERN_WARNING diff --git a/arch/x86/kernel/apm_32.c b/arch/x86/kernel/apm_32.c index b5b6b23..031aa88 100644 --- a/arch/x86/kernel/apm_32.c +++ b/arch/x86/kernel/apm_32.c @@ -1992,8 +1992,8 @@ static int __init apm_is_horked_d850md(const struct dmi_system_id *d) apm_info.disabled = 1; printk(KERN_INFO "%s machine detected. " "Disabling APM.\n", d->ident); - printk(KERN_INFO "This bug is fixed in bios P15 which is available for \n"); - printk(KERN_INFO "download from support.intel.com \n"); + printk(KERN_INFO "This bug is fixed in bios P15 which is available for\n"); + printk(KERN_INFO "download from support.intel.com\n"); } return 0; } diff --git a/arch/x86/kernel/efi.c b/arch/x86/kernel/efi.c index cdcfb12..c2fa9b8 100644 --- a/arch/x86/kernel/efi.c +++ b/arch/x86/kernel/efi.c @@ -362,7 +362,7 @@ void __init efi_init(void) printk(KERN_ERR PFX "Could not map the firmware vendor!\n"); early_iounmap(tmp, 2); - printk(KERN_INFO "EFI v%u.%.02u by %s \n", + printk(KERN_INFO "EFI v%u.%.02u by %s\n", efi.systab->hdr.revision >> 16, efi.systab->hdr.revision & 0xffff, vendor); diff --git a/arch/x86/kernel/microcode_intel.c b/arch/x86/kernel/microcode_intel.c index ebd193e..85a343e 100644 --- a/arch/x86/kernel/microcode_intel.c +++ b/arch/x86/kernel/microcode_intel.c @@ -328,7 +328,7 @@ static int apply_microcode(int cpu) cpu_num, mc_intel->hdr.rev); return -1; } - pr_info("CPU%d updated to revision 0x%x, date = %04x-%02x-%02x \n", + pr_info("CPU%d updated to revision 0x%x, date = %04x-%02x-%02x\n", cpu_num, val[1], mc_intel->hdr.date & 0xffff, mc_intel->hdr.date >> 24, diff --git a/arch/x86/kernel/uv_sysfs.c b/arch/x86/kernel/uv_sysfs.c index 36afb98..309c70f 100644 --- a/arch/x86/kernel/uv_sysfs.c +++ b/arch/x86/kernel/uv_sysfs.c @@ -54,19 +54,19 @@ static int __init sgi_uv_sysfs_init(void) if (!sgi_uv_kobj) sgi_uv_kobj = kobject_create_and_add("sgi_uv", firmware_kobj); if (!sgi_uv_kobj) { - printk(KERN_WARNING "kobject_create_and_add sgi_uv failed \n"); + printk(KERN_WARNING "kobject_create_and_add sgi_uv failed\n"); return -EINVAL; } ret = sysfs_create_file(sgi_uv_kobj, &partition_id_attr.attr); if (ret) { - printk(KERN_WARNING "sysfs_create_file partition_id failed \n"); + printk(KERN_WARNING "sysfs_create_file partition_id failed\n"); return ret; } ret = sysfs_create_file(sgi_uv_kobj, &coherence_id_attr.attr); if (ret) { - printk(KERN_WARNING "sysfs_create_file coherence_id failed \n"); + printk(KERN_WARNING "sysfs_create_file coherence_id failed\n"); return ret; } diff --git a/arch/x86/tools/test_get_len.c b/arch/x86/tools/test_get_len.c index bee8d6a..13403fc 100644 --- a/arch/x86/tools/test_get_len.c +++ b/arch/x86/tools/test_get_len.c @@ -43,7 +43,7 @@ static int x86_64; static void usage(void) { fprintf(stderr, "Usage: objdump -d a.out | awk -f distill.awk |" - " %s [-y|-n] [-v] \n", prog); + " %s [-y|-n] [-v]\n", prog); fprintf(stderr, "\t-y 64bit mode\n"); fprintf(stderr, "\t-n 32bit mode\n"); fprintf(stderr, "\t-v verbose mode\n"); @@ -69,7 +69,7 @@ static void dump_field(FILE *fp, const char *name, const char *indent, static void dump_insn(FILE *fp, struct insn *insn) { - fprintf(fp, "Instruction = { \n"); + fprintf(fp, "Instruction = {\n"); dump_field(fp, "prefixes", "\t", &insn->prefixes); dump_field(fp, "rex_prefix", "\t", &insn->rex_prefix); dump_field(fp, "vex_prefix", "\t", &insn->vex_prefix); -- cgit v0.10.2 From 076dc4a65a6d99a16979e2c7917e669fb8c91ee5 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Fri, 5 Feb 2010 12:16:47 -0500 Subject: x86/alternatives: Fix build warning Fixes these warnings: arch/x86/kernel/alternative.c: In function 'alternatives_text_reserved': arch/x86/kernel/alternative.c:402: warning: comparison of distinct pointer types lacks a cast arch/x86/kernel/alternative.c:402: warning: comparison of distinct pointer types lacks a cast arch/x86/kernel/alternative.c:405: warning: comparison of distinct pointer types lacks a cast arch/x86/kernel/alternative.c:405: warning: comparison of distinct pointer types lacks a cast Caused by: 2cfa197: ftrace/alternatives: Introducing *_text_reserved functions Changes in v2: - Use local variables to compare, instead of type casts. Reported-by: Ingo Molnar Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE LKML-Reference: <20100205171647.15750.37221.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index 3c13284..e63b80e 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -395,12 +395,14 @@ int alternatives_text_reserved(void *start, void *end) { struct smp_alt_module *mod; u8 **ptr; + u8 *text_start = start; + u8 *text_end = end; list_for_each_entry(mod, &smp_alt_modules, next) { - if (mod->text > end || mod->text_end < start) + if (mod->text > text_end || mod->text_end < text_start) continue; for (ptr = mod->locks; ptr < mod->locks_end; ptr++) - if (start <= *ptr && end >= *ptr) + if (text_start <= *ptr && text_end >= *ptr) return 1; } -- cgit v0.10.2 From fc4618575f79eea062cdc51715040e40cd35b71c Mon Sep 17 00:00:00 2001 From: Guennadi Liakhovetski Date: Tue, 19 Jan 2010 07:24:55 +0000 Subject: sh: prepare the DMA driver for slave functionality Slave DMA functionality uses scatter-gather arrays for data transfers, whereas memcpy just uses a single data buffer. This patch converts the current memcpy implementation in shdma.c to use scatter-gather, making it just a special case with one SG-element. This allows us to isolate descriptor list manipulations and locking into one function, thus reducing error chances. Signed-off-by: Guennadi Liakhovetski Acked-by: Dan Williams Signed-off-by: Paul Mundt diff --git a/drivers/dma/shdma.c b/drivers/dma/shdma.c index d10cc89..427c3ef 100644 --- a/drivers/dma/shdma.c +++ b/drivers/dma/shdma.c @@ -53,12 +53,12 @@ static void sh_dmae_chan_ld_cleanup(struct sh_dmae_chan *sh_chan, bool all); #define SH_DMAC_CHAN_BASE(id) (dma_base_addr[id]) static void sh_dmae_writel(struct sh_dmae_chan *sh_dc, u32 data, u32 reg) { - ctrl_outl(data, (SH_DMAC_CHAN_BASE(sh_dc->id) + reg)); + ctrl_outl(data, SH_DMAC_CHAN_BASE(sh_dc->id) + reg); } static u32 sh_dmae_readl(struct sh_dmae_chan *sh_dc, u32 reg) { - return ctrl_inl((SH_DMAC_CHAN_BASE(sh_dc->id) + reg)); + return ctrl_inl(SH_DMAC_CHAN_BASE(sh_dc->id) + reg); } static void dmae_init(struct sh_dmae_chan *sh_chan) @@ -95,14 +95,14 @@ static int sh_dmae_rst(int id) return 0; } -static int dmae_is_busy(struct sh_dmae_chan *sh_chan) +static bool dmae_is_busy(struct sh_dmae_chan *sh_chan) { u32 chcr = sh_dmae_readl(sh_chan, CHCR); - if (chcr & CHCR_DE) { - if (!(chcr & CHCR_TE)) - return -EBUSY; /* working */ - } - return 0; /* waiting */ + + if ((chcr & (CHCR_DE | CHCR_TE)) == CHCR_DE) + return true; /* working */ + + return false; /* waiting */ } static inline unsigned int calc_xmit_shift(struct sh_dmae_chan *sh_chan) @@ -136,10 +136,9 @@ static void dmae_halt(struct sh_dmae_chan *sh_chan) static int dmae_set_chcr(struct sh_dmae_chan *sh_chan, u32 val) { - int ret = dmae_is_busy(sh_chan); /* When DMA was working, can not set data to CHCR */ - if (ret) - return ret; + if (dmae_is_busy(sh_chan)) + return -EBUSY; sh_dmae_writel(sh_chan, val, CHCR); return 0; @@ -153,9 +152,9 @@ static int dmae_set_dmars(struct sh_dmae_chan *sh_chan, u16 val) { u32 addr; int shift = 0; - int ret = dmae_is_busy(sh_chan); - if (ret) - return ret; + + if (dmae_is_busy(sh_chan)) + return -EBUSY; if (sh_chan->id & DMARS_CHAN_MSK) shift = DMARS_SHIFT; @@ -301,23 +300,95 @@ static void sh_dmae_free_chan_resources(struct dma_chan *chan) kfree(desc); } -static struct dma_async_tx_descriptor *sh_dmae_prep_memcpy( - struct dma_chan *chan, dma_addr_t dma_dest, dma_addr_t dma_src, - size_t len, unsigned long flags) +/* + * sh_dmae_add_desc - get, set up and return one transfer descriptor + * @sh_chan: DMA channel + * @flags: DMA transfer flags + * @dest: destination DMA address, incremented when direction equals + * DMA_FROM_DEVICE or DMA_BIDIRECTIONAL + * @src: source DMA address, incremented when direction equals + * DMA_TO_DEVICE or DMA_BIDIRECTIONAL + * @len: DMA transfer length + * @first: if NULL, set to the current descriptor and cookie set to -EBUSY + * @direction: needed for slave DMA to decide which address to keep constant, + * equals DMA_BIDIRECTIONAL for MEMCPY + * Returns 0 or an error + * Locks: called with desc_lock held + */ +static struct sh_desc *sh_dmae_add_desc(struct sh_dmae_chan *sh_chan, + unsigned long flags, dma_addr_t *dest, dma_addr_t *src, size_t *len, + struct sh_desc **first, enum dma_data_direction direction) { - struct sh_dmae_chan *sh_chan; - struct sh_desc *first = NULL, *prev = NULL, *new; + struct sh_desc *new; size_t copy_size; - LIST_HEAD(tx_list); - int chunks = (len + SH_DMA_TCR_MAX) / (SH_DMA_TCR_MAX + 1); - if (!chan) + if (!*len) return NULL; - if (!len) + /* Allocate the link descriptor from the free list */ + new = sh_dmae_get_desc(sh_chan); + if (!new) { + dev_err(sh_chan->dev, "No free link descriptor available\n"); return NULL; + } - sh_chan = to_sh_chan(chan); + copy_size = min(*len, (size_t)SH_DMA_TCR_MAX + 1); + + new->hw.sar = *src; + new->hw.dar = *dest; + new->hw.tcr = copy_size; + + if (!*first) { + /* First desc */ + new->async_tx.cookie = -EBUSY; + *first = new; + } else { + /* Other desc - invisible to the user */ + new->async_tx.cookie = -EINVAL; + } + + dev_dbg(sh_chan->dev, "chaining (%u/%u)@%x -> %x with %p, cookie %d\n", + copy_size, *len, *src, *dest, &new->async_tx, + new->async_tx.cookie); + + new->mark = DESC_PREPARED; + new->async_tx.flags = flags; + + *len -= copy_size; + if (direction == DMA_BIDIRECTIONAL || direction == DMA_TO_DEVICE) + *src += copy_size; + if (direction == DMA_BIDIRECTIONAL || direction == DMA_FROM_DEVICE) + *dest += copy_size; + + return new; +} + +/* + * sh_dmae_prep_sg - prepare transfer descriptors from an SG list + * + * Common routine for public (MEMCPY) and slave DMA. The MEMCPY case is also + * converted to scatter-gather to guarantee consistent locking and a correct + * list manipulation. For slave DMA direction carries the usual meaning, and, + * logically, the SG list is RAM and the addr variable contains slave address, + * e.g., the FIFO I/O register. For MEMCPY direction equals DMA_BIDIRECTIONAL + * and the SG list contains only one element and points at the source buffer. + */ +static struct dma_async_tx_descriptor *sh_dmae_prep_sg(struct sh_dmae_chan *sh_chan, + struct scatterlist *sgl, unsigned int sg_len, dma_addr_t *addr, + enum dma_data_direction direction, unsigned long flags) +{ + struct scatterlist *sg; + struct sh_desc *first = NULL, *new = NULL /* compiler... */; + LIST_HEAD(tx_list); + int chunks = 0; + int i; + + if (!sg_len) + return NULL; + + for_each_sg(sgl, sg, sg_len, i) + chunks += (sg_dma_len(sg) + SH_DMA_TCR_MAX) / + (SH_DMA_TCR_MAX + 1); /* Have to lock the whole loop to protect against concurrent release */ spin_lock_bh(&sh_chan->desc_lock); @@ -333,49 +404,32 @@ static struct dma_async_tx_descriptor *sh_dmae_prep_memcpy( * only during this function, then they are immediately spliced * back onto the free list in form of a chain */ - do { - /* Allocate the link descriptor from the free list */ - new = sh_dmae_get_desc(sh_chan); - if (!new) { - dev_err(sh_chan->dev, - "No free memory for link descriptor\n"); - list_for_each_entry(new, &tx_list, node) - new->mark = DESC_IDLE; - list_splice(&tx_list, &sh_chan->ld_free); - spin_unlock_bh(&sh_chan->desc_lock); - return NULL; - } - - copy_size = min(len, (size_t)SH_DMA_TCR_MAX + 1); - - new->hw.sar = dma_src; - new->hw.dar = dma_dest; - new->hw.tcr = copy_size; - if (!first) { - /* First desc */ - new->async_tx.cookie = -EBUSY; - first = new; - } else { - /* Other desc - invisible to the user */ - new->async_tx.cookie = -EINVAL; - } - - dev_dbg(sh_chan->dev, - "chaining %u of %u with %p, dst %x, cookie %d\n", - copy_size, len, &new->async_tx, dma_dest, - new->async_tx.cookie); - - new->mark = DESC_PREPARED; - new->async_tx.flags = flags; - new->chunks = chunks--; - - prev = new; - len -= copy_size; - dma_src += copy_size; - dma_dest += copy_size; - /* Insert the link descriptor to the LD ring */ - list_add_tail(&new->node, &tx_list); - } while (len); + for_each_sg(sgl, sg, sg_len, i) { + dma_addr_t sg_addr = sg_dma_address(sg); + size_t len = sg_dma_len(sg); + + if (!len) + goto err_get_desc; + + do { + dev_dbg(sh_chan->dev, "Add SG #%d@%p[%d], dma %llx\n", + i, sg, len, (unsigned long long)sg_addr); + + if (direction == DMA_FROM_DEVICE) + new = sh_dmae_add_desc(sh_chan, flags, + &sg_addr, addr, &len, &first, + direction); + else + new = sh_dmae_add_desc(sh_chan, flags, + addr, &sg_addr, &len, &first, + direction); + if (!new) + goto err_get_desc; + + new->chunks = chunks--; + list_add_tail(&new->node, &tx_list); + } while (len); + } if (new != first) new->async_tx.cookie = -ENOSPC; @@ -386,6 +440,37 @@ static struct dma_async_tx_descriptor *sh_dmae_prep_memcpy( spin_unlock_bh(&sh_chan->desc_lock); return &first->async_tx; + +err_get_desc: + list_for_each_entry(new, &tx_list, node) + new->mark = DESC_IDLE; + list_splice(&tx_list, &sh_chan->ld_free); + + spin_unlock_bh(&sh_chan->desc_lock); + + return NULL; +} + +static struct dma_async_tx_descriptor *sh_dmae_prep_memcpy( + struct dma_chan *chan, dma_addr_t dma_dest, dma_addr_t dma_src, + size_t len, unsigned long flags) +{ + struct sh_dmae_chan *sh_chan; + struct scatterlist sg; + + if (!chan || !len) + return NULL; + + sh_chan = to_sh_chan(chan); + + sg_init_table(&sg, 1); + sg_set_page(&sg, pfn_to_page(PFN_DOWN(dma_src)), len, + offset_in_page(dma_src)); + sg_dma_address(&sg) = dma_src; + sg_dma_len(&sg) = len; + + return sh_dmae_prep_sg(sh_chan, &sg, 1, &dma_dest, DMA_BIDIRECTIONAL, + flags); } static dma_async_tx_callback __ld_cleanup(struct sh_dmae_chan *sh_chan, bool all) @@ -559,7 +644,7 @@ static irqreturn_t sh_dmae_err(int irq, void *data) /* IRQ Multi */ if (shdev->pdata.mode & SHDMA_MIX_IRQ) { - int cnt = 0; + int __maybe_unused cnt = 0; switch (irq) { #if defined(DMTE6_IRQ) && defined(DMAE1_IRQ) case DMTE6_IRQ: -- cgit v0.10.2 From 623b4ac4bf9e767991c66e29b47dd4b19458fb42 Mon Sep 17 00:00:00 2001 From: Guennadi Liakhovetski Date: Wed, 3 Feb 2010 14:44:12 +0000 Subject: sh: fix Transfer Size calculation in both DMA drivers Both the original arch/sh/drivers/dma/dma-sh.c and the new SH dmaengine drivers do not take into account bits 3:2 of the Transfer Size field in the CHCR register, besides, bit-field defines set bit 2, but the mask only passes bits 1:0 through. TS_16BLK and TS_32BLK macros are bogus too. This patch fixes all these issues for sh7722 and sh7724, other CPUs stay unchanged and might need to be fixed too. Signed-off-by: Guennadi Liakhovetski Acked-by: Dan Williams Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/dma/dma-sh.c b/arch/sh/drivers/dma/dma-sh.c index 37fb5b8..31830cb 100644 --- a/arch/sh/drivers/dma/dma-sh.c +++ b/arch/sh/drivers/dma/dma-sh.c @@ -52,11 +52,14 @@ static inline unsigned int get_dmte_irq(unsigned int chan) * * iterations to complete the transfer. */ +static unsigned int ts_shift[] = TS_SHIFT; static inline unsigned int calc_xmit_shift(struct dma_channel *chan) { u32 chcr = ctrl_inl(dma_base_addr[chan->chan] + CHCR); + int cnt = ((chcr & CHCR_TS_LOW_MASK) >> CHCR_TS_LOW_SHIFT) | + ((chcr & CHCR_TS_HIGH_MASK) >> CHCR_TS_HIGH_SHIFT); - return ts_shift[(chcr & CHCR_TS_MASK)>>CHCR_TS_SHIFT]; + return ts_shift[cnt]; } /* diff --git a/arch/sh/include/asm/dma-sh.h b/arch/sh/include/asm/dma-sh.h index 78eed3e..01d2fc7 100644 --- a/arch/sh/include/asm/dma-sh.h +++ b/arch/sh/include/asm/dma-sh.h @@ -83,7 +83,7 @@ static int dmte_irq_map[] __maybe_unused = { * Define the default configuration for dual address memory-memory transfer. * The 0x400 value represents auto-request, external->external. */ -#define RS_DUAL (DM_INC | SM_INC | 0x400 | TS_32) +#define RS_DUAL (DM_INC | SM_INC | 0x400 | TS_INDEX2VAL(XMIT_SZ_32BIT)) /* DMA base address */ static u32 dma_base_addr[] __maybe_unused = { diff --git a/arch/sh/include/cpu-sh3/cpu/dma.h b/arch/sh/include/cpu-sh3/cpu/dma.h index 0ea15f3..207811a 100644 --- a/arch/sh/include/cpu-sh3/cpu/dma.h +++ b/arch/sh/include/cpu-sh3/cpu/dma.h @@ -20,8 +20,10 @@ #define TS_32 0x00000010 #define TS_128 0x00000018 -#define CHCR_TS_MASK 0x18 -#define CHCR_TS_SHIFT 3 +#define CHCR_TS_LOW_MASK 0x18 +#define CHCR_TS_LOW_SHIFT 3 +#define CHCR_TS_HIGH_MASK 0 +#define CHCR_TS_HIGH_SHIFT 0 #define DMAOR_INIT DMAOR_DME @@ -36,11 +38,13 @@ enum { XMIT_SZ_128BIT, }; -static unsigned int ts_shift[] __maybe_unused = { - [XMIT_SZ_8BIT] = 0, - [XMIT_SZ_16BIT] = 1, - [XMIT_SZ_32BIT] = 2, - [XMIT_SZ_128BIT] = 4, -}; +#define TS_SHIFT { \ + [XMIT_SZ_8BIT] = 0, \ + [XMIT_SZ_16BIT] = 1, \ + [XMIT_SZ_32BIT] = 2, \ + [XMIT_SZ_128BIT] = 4, \ +} + +#define TS_INDEX2VAL(i) (((i) & 3) << CHCR_TS_LOW_SHIFT) #endif /* __ASM_CPU_SH3_DMA_H */ diff --git a/arch/sh/include/cpu-sh4/cpu/dma-sh4a.h b/arch/sh/include/cpu-sh4/cpu/dma-sh4a.h index c4ed660..cc1cf3e 100644 --- a/arch/sh/include/cpu-sh4/cpu/dma-sh4a.h +++ b/arch/sh/include/cpu-sh4/cpu/dma-sh4a.h @@ -2,13 +2,26 @@ #define __ASM_SH_CPU_SH4_DMA_SH7780_H #if defined(CONFIG_CPU_SUBTYPE_SH7343) || \ - defined(CONFIG_CPU_SUBTYPE_SH7722) || \ defined(CONFIG_CPU_SUBTYPE_SH7730) #define DMTE0_IRQ 48 #define DMTE4_IRQ 76 #define DMAE0_IRQ 78 /* DMA Error IRQ*/ #define SH_DMAC_BASE0 0xFE008020 #define SH_DMARS_BASE 0xFE009000 +#define CHCR_TS_LOW_MASK 0x00000018 +#define CHCR_TS_LOW_SHIFT 3 +#define CHCR_TS_HIGH_MASK 0 +#define CHCR_TS_HIGH_SHIFT 0 +#elif defined(CONFIG_CPU_SUBTYPE_SH7722) +#define DMTE0_IRQ 48 +#define DMTE4_IRQ 76 +#define DMAE0_IRQ 78 /* DMA Error IRQ*/ +#define SH_DMAC_BASE0 0xFE008020 +#define SH_DMARS_BASE 0xFE009000 +#define CHCR_TS_LOW_MASK 0x00000018 +#define CHCR_TS_LOW_SHIFT 3 +#define CHCR_TS_HIGH_MASK 0x00300000 +#define CHCR_TS_HIGH_SHIFT 20 #elif defined(CONFIG_CPU_SUBTYPE_SH7763) || \ defined(CONFIG_CPU_SUBTYPE_SH7764) #define DMTE0_IRQ 34 @@ -16,8 +29,11 @@ #define DMAE0_IRQ 38 #define SH_DMAC_BASE0 0xFF608020 #define SH_DMARS_BASE 0xFF609000 -#elif defined(CONFIG_CPU_SUBTYPE_SH7723) || \ - defined(CONFIG_CPU_SUBTYPE_SH7724) +#define CHCR_TS_LOW_MASK 0x00000018 +#define CHCR_TS_LOW_SHIFT 3 +#define CHCR_TS_HIGH_MASK 0 +#define CHCR_TS_HIGH_SHIFT 0 +#elif defined(CONFIG_CPU_SUBTYPE_SH7723) #define DMTE0_IRQ 48 /* DMAC0A*/ #define DMTE4_IRQ 76 /* DMAC0B */ #define DMTE6_IRQ 40 @@ -30,6 +46,27 @@ #define SH_DMAC_BASE0 0xFE008020 #define SH_DMAC_BASE1 0xFDC08020 #define SH_DMARS_BASE 0xFDC09000 +#define CHCR_TS_LOW_MASK 0x00000018 +#define CHCR_TS_LOW_SHIFT 3 +#define CHCR_TS_HIGH_MASK 0 +#define CHCR_TS_HIGH_SHIFT 0 +#elif defined(CONFIG_CPU_SUBTYPE_SH7724) +#define DMTE0_IRQ 48 /* DMAC0A*/ +#define DMTE4_IRQ 76 /* DMAC0B */ +#define DMTE6_IRQ 40 +#define DMTE8_IRQ 42 /* DMAC1A */ +#define DMTE9_IRQ 43 +#define DMTE10_IRQ 72 /* DMAC1B */ +#define DMTE11_IRQ 73 +#define DMAE0_IRQ 78 /* DMA Error IRQ*/ +#define DMAE1_IRQ 74 /* DMA Error IRQ*/ +#define SH_DMAC_BASE0 0xFE008020 +#define SH_DMAC_BASE1 0xFDC08020 +#define SH_DMARS_BASE 0xFDC09000 +#define CHCR_TS_LOW_MASK 0x00000018 +#define CHCR_TS_LOW_SHIFT 3 +#define CHCR_TS_HIGH_MASK 0x00600000 +#define CHCR_TS_HIGH_SHIFT 21 #elif defined(CONFIG_CPU_SUBTYPE_SH7780) #define DMTE0_IRQ 34 #define DMTE4_IRQ 44 @@ -42,6 +79,10 @@ #define SH_DMAC_BASE0 0xFC808020 #define SH_DMAC_BASE1 0xFC818020 #define SH_DMARS_BASE 0xFC809000 +#define CHCR_TS_LOW_MASK 0x00000018 +#define CHCR_TS_LOW_SHIFT 3 +#define CHCR_TS_HIGH_MASK 0 +#define CHCR_TS_HIGH_SHIFT 0 #else /* SH7785 */ #define DMTE0_IRQ 33 #define DMTE4_IRQ 37 @@ -55,17 +96,16 @@ #define SH_DMAC_BASE0 0xFC808020 #define SH_DMAC_BASE1 0xFCC08020 #define SH_DMARS_BASE 0xFC809000 +#define CHCR_TS_LOW_MASK 0x00000018 +#define CHCR_TS_LOW_SHIFT 3 +#define CHCR_TS_HIGH_MASK 0 +#define CHCR_TS_HIGH_SHIFT 0 #endif -#define REQ_HE 0x000000C0 -#define REQ_H 0x00000080 -#define REQ_LE 0x00000040 -#define TM_BURST 0x0000020 -#define TS_8 0x00000000 -#define TS_16 0x00000008 -#define TS_32 0x00000010 -#define TS_16BLK 0x00000018 -#define TS_32BLK 0x00100000 +#define REQ_HE 0x000000C0 +#define REQ_H 0x00000080 +#define REQ_LE 0x00000040 +#define TM_BURST 0x00000020 /* * The SuperH DMAC supports a number of transmit sizes, we list them here, @@ -74,22 +114,31 @@ * Defaults to a 64-bit transfer size. */ enum { - XMIT_SZ_8BIT, - XMIT_SZ_16BIT, - XMIT_SZ_32BIT, - XMIT_SZ_128BIT, - XMIT_SZ_256BIT, + XMIT_SZ_8BIT = 0, + XMIT_SZ_16BIT = 1, + XMIT_SZ_32BIT = 2, + XMIT_SZ_64BIT = 7, + XMIT_SZ_128BIT = 3, + XMIT_SZ_256BIT = 4, + XMIT_SZ_128BIT_BLK = 0xb, + XMIT_SZ_256BIT_BLK = 0xc, }; /* * The DMA count is defined as the number of bytes to transfer. */ -static unsigned int ts_shift[] __maybe_unused = { - [XMIT_SZ_8BIT] = 0, - [XMIT_SZ_16BIT] = 1, - [XMIT_SZ_32BIT] = 2, - [XMIT_SZ_128BIT] = 4, - [XMIT_SZ_256BIT] = 5, -}; +#define TS_SHIFT { \ + [XMIT_SZ_8BIT] = 0, \ + [XMIT_SZ_16BIT] = 1, \ + [XMIT_SZ_32BIT] = 2, \ + [XMIT_SZ_64BIT] = 3, \ + [XMIT_SZ_128BIT] = 4, \ + [XMIT_SZ_256BIT] = 5, \ + [XMIT_SZ_128BIT_BLK] = 4, \ + [XMIT_SZ_256BIT_BLK] = 5, \ +} + +#define TS_INDEX2VAL(i) ((((i) & 3) << CHCR_TS_LOW_SHIFT) | \ + ((((i) >> 2) & 3) << CHCR_TS_HIGH_SHIFT)) #endif /* __ASM_SH_CPU_SH4_DMA_SH7780_H */ diff --git a/arch/sh/include/cpu-sh4/cpu/dma.h b/arch/sh/include/cpu-sh4/cpu/dma.h index bcb3024..114a369 100644 --- a/arch/sh/include/cpu-sh4/cpu/dma.h +++ b/arch/sh/include/cpu-sh4/cpu/dma.h @@ -6,8 +6,6 @@ #ifdef CONFIG_CPU_SH4A #define DMAOR_INIT (DMAOR_DME) -#define CHCR_TS_MASK 0x18 -#define CHCR_TS_SHIFT 3 #include #else /* CONFIG_CPU_SH4A */ @@ -29,8 +27,10 @@ #define TS_32 0x00000030 #define TS_64 0x00000000 -#define CHCR_TS_MASK 0x70 -#define CHCR_TS_SHIFT 4 +#define CHCR_TS_LOW_MASK 0x70 +#define CHCR_TS_LOW_SHIFT 4 +#define CHCR_TS_HIGH_MASK 0 +#define CHCR_TS_HIGH_SHIFT 0 #define DMAOR_COD 0x00000008 @@ -41,23 +41,26 @@ * Defaults to a 64-bit transfer size. */ enum { - XMIT_SZ_64BIT, - XMIT_SZ_8BIT, - XMIT_SZ_16BIT, - XMIT_SZ_32BIT, - XMIT_SZ_256BIT, + XMIT_SZ_8BIT = 1, + XMIT_SZ_16BIT = 2, + XMIT_SZ_32BIT = 3, + XMIT_SZ_64BIT = 0, + XMIT_SZ_256BIT = 4, }; /* * The DMA count is defined as the number of bytes to transfer. */ -static unsigned int ts_shift[] __maybe_unused = { - [XMIT_SZ_64BIT] = 3, - [XMIT_SZ_8BIT] = 0, - [XMIT_SZ_16BIT] = 1, - [XMIT_SZ_32BIT] = 2, - [XMIT_SZ_256BIT] = 5, -}; +#define TS_SHIFT { \ + [XMIT_SZ_8BIT] = 0, \ + [XMIT_SZ_16BIT] = 1, \ + [XMIT_SZ_32BIT] = 2, \ + [XMIT_SZ_64BIT] = 3, \ + [XMIT_SZ_256BIT] = 5, \ +} + +#define TS_INDEX2VAL(i) (((i) & 7) << CHCR_TS_LOW_SHIFT) + #endif #endif /* __ASM_CPU_SH4_DMA_H */ diff --git a/drivers/dma/shdma.c b/drivers/dma/shdma.c index 427c3ef..3e1037c 100644 --- a/drivers/dma/shdma.c +++ b/drivers/dma/shdma.c @@ -105,10 +105,14 @@ static bool dmae_is_busy(struct sh_dmae_chan *sh_chan) return false; /* waiting */ } +static unsigned int ts_shift[] = TS_SHIFT; static inline unsigned int calc_xmit_shift(struct sh_dmae_chan *sh_chan) { u32 chcr = sh_dmae_readl(sh_chan, CHCR); - return ts_shift[(chcr & CHCR_TS_MASK) >> CHCR_TS_SHIFT]; + int cnt = ((chcr & CHCR_TS_LOW_MASK) >> CHCR_TS_LOW_SHIFT) | + ((chcr & CHCR_TS_HIGH_MASK) >> CHCR_TS_HIGH_SHIFT); + + return ts_shift[cnt]; } static void dmae_set_reg(struct sh_dmae_chan *sh_chan, struct sh_dmae_regs *hw) -- cgit v0.10.2 From cfefe99795251d76d92e8457f4152f532a961ec5 Mon Sep 17 00:00:00 2001 From: Guennadi Liakhovetski Date: Wed, 3 Feb 2010 14:46:41 +0000 Subject: sh: implement DMA_SLAVE capability in SH dmaengine driver Tested to work with a SIU ASoC driver on sh7722 (migor). Signed-off-by: Guennadi Liakhovetski Acked-by: Dan Williams Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/dma-sh.h b/arch/sh/include/asm/dma-sh.h index 01d2fc7..c8d8ce7 100644 --- a/arch/sh/include/asm/dma-sh.h +++ b/arch/sh/include/asm/dma-sh.h @@ -64,8 +64,10 @@ static int dmte_irq_map[] __maybe_unused = { #define ACK_L 0x00010000 #define DM_INC 0x00004000 #define DM_DEC 0x00008000 +#define DM_FIX 0x0000c000 #define SM_INC 0x00001000 #define SM_DEC 0x00002000 +#define SM_FIX 0x00003000 #define RS_IN 0x00000200 #define RS_OUT 0x00000300 #define TS_BLK 0x00000040 @@ -123,10 +125,47 @@ static u32 dma_base_addr[] __maybe_unused = { */ #define SHDMA_MIX_IRQ (1 << 1) #define SHDMA_DMAOR1 (1 << 2) -#define SHDMA_DMAE1 (1 << 3) +#define SHDMA_DMAE1 (1 << 3) + +enum sh_dmae_slave_chan_id { + SHDMA_SLAVE_SCIF0_TX, + SHDMA_SLAVE_SCIF0_RX, + SHDMA_SLAVE_SCIF1_TX, + SHDMA_SLAVE_SCIF1_RX, + SHDMA_SLAVE_SCIF2_TX, + SHDMA_SLAVE_SCIF2_RX, + SHDMA_SLAVE_SCIF3_TX, + SHDMA_SLAVE_SCIF3_RX, + SHDMA_SLAVE_SCIF4_TX, + SHDMA_SLAVE_SCIF4_RX, + SHDMA_SLAVE_SCIF5_TX, + SHDMA_SLAVE_SCIF5_RX, + SHDMA_SLAVE_SIUA_TX, + SHDMA_SLAVE_SIUA_RX, + SHDMA_SLAVE_SIUB_TX, + SHDMA_SLAVE_SIUB_RX, + SHDMA_SLAVE_NUMBER, /* Must stay last */ +}; + +struct sh_dmae_slave_config { + enum sh_dmae_slave_chan_id slave_id; + dma_addr_t addr; + u32 chcr; + char mid_rid; +}; struct sh_dmae_pdata { unsigned int mode; + struct sh_dmae_slave_config *config; + int config_num; +}; + +struct device; + +struct sh_dmae_slave { + enum sh_dmae_slave_chan_id slave_id; /* Set by the platform */ + struct device *dma_dev; /* Set by the platform */ + struct sh_dmae_slave_config *config; /* Set by the driver */ }; #endif /* __DMA_SH_H */ diff --git a/arch/sh/include/cpu-sh4/cpu/dma-sh4a.h b/arch/sh/include/cpu-sh4/cpu/dma-sh4a.h index cc1cf3e..e734ea4 100644 --- a/arch/sh/include/cpu-sh4/cpu/dma-sh4a.h +++ b/arch/sh/include/cpu-sh4/cpu/dma-sh4a.h @@ -7,7 +7,7 @@ #define DMTE4_IRQ 76 #define DMAE0_IRQ 78 /* DMA Error IRQ*/ #define SH_DMAC_BASE0 0xFE008020 -#define SH_DMARS_BASE 0xFE009000 +#define SH_DMARS_BASE0 0xFE009000 #define CHCR_TS_LOW_MASK 0x00000018 #define CHCR_TS_LOW_SHIFT 3 #define CHCR_TS_HIGH_MASK 0 @@ -17,7 +17,7 @@ #define DMTE4_IRQ 76 #define DMAE0_IRQ 78 /* DMA Error IRQ*/ #define SH_DMAC_BASE0 0xFE008020 -#define SH_DMARS_BASE 0xFE009000 +#define SH_DMARS_BASE0 0xFE009000 #define CHCR_TS_LOW_MASK 0x00000018 #define CHCR_TS_LOW_SHIFT 3 #define CHCR_TS_HIGH_MASK 0x00300000 @@ -28,7 +28,7 @@ #define DMTE4_IRQ 44 #define DMAE0_IRQ 38 #define SH_DMAC_BASE0 0xFF608020 -#define SH_DMARS_BASE 0xFF609000 +#define SH_DMARS_BASE0 0xFF609000 #define CHCR_TS_LOW_MASK 0x00000018 #define CHCR_TS_LOW_SHIFT 3 #define CHCR_TS_HIGH_MASK 0 @@ -45,7 +45,7 @@ #define DMAE1_IRQ 74 /* DMA Error IRQ*/ #define SH_DMAC_BASE0 0xFE008020 #define SH_DMAC_BASE1 0xFDC08020 -#define SH_DMARS_BASE 0xFDC09000 +#define SH_DMARS_BASE0 0xFDC09000 #define CHCR_TS_LOW_MASK 0x00000018 #define CHCR_TS_LOW_SHIFT 3 #define CHCR_TS_HIGH_MASK 0 @@ -62,7 +62,8 @@ #define DMAE1_IRQ 74 /* DMA Error IRQ*/ #define SH_DMAC_BASE0 0xFE008020 #define SH_DMAC_BASE1 0xFDC08020 -#define SH_DMARS_BASE 0xFDC09000 +#define SH_DMARS_BASE0 0xFE009000 +#define SH_DMARS_BASE1 0xFDC09000 #define CHCR_TS_LOW_MASK 0x00000018 #define CHCR_TS_LOW_SHIFT 3 #define CHCR_TS_HIGH_MASK 0x00600000 @@ -78,7 +79,7 @@ #define DMAE0_IRQ 38 /* DMA Error IRQ */ #define SH_DMAC_BASE0 0xFC808020 #define SH_DMAC_BASE1 0xFC818020 -#define SH_DMARS_BASE 0xFC809000 +#define SH_DMARS_BASE0 0xFC809000 #define CHCR_TS_LOW_MASK 0x00000018 #define CHCR_TS_LOW_SHIFT 3 #define CHCR_TS_HIGH_MASK 0 @@ -95,7 +96,7 @@ #define DMAE1_IRQ 58 /* DMA Error IRQ1 */ #define SH_DMAC_BASE0 0xFC808020 #define SH_DMAC_BASE1 0xFCC08020 -#define SH_DMARS_BASE 0xFC809000 +#define SH_DMARS_BASE0 0xFC809000 #define CHCR_TS_LOW_MASK 0x00000018 #define CHCR_TS_LOW_SHIFT 3 #define CHCR_TS_HIGH_MASK 0 diff --git a/drivers/dma/shdma.c b/drivers/dma/shdma.c index 3e1037c..b75ce8b 100644 --- a/drivers/dma/shdma.c +++ b/drivers/dma/shdma.c @@ -48,6 +48,9 @@ enum sh_dmae_desc_status { */ #define RS_DEFAULT (RS_DUAL) +/* A bitmask with bits enough for enum sh_dmae_slave_chan_id */ +static unsigned long sh_dmae_slave_used[BITS_TO_LONGS(SHDMA_SLAVE_NUMBER)]; + static void sh_dmae_chan_ld_cleanup(struct sh_dmae_chan *sh_chan, bool all); #define SH_DMAC_CHAN_BASE(id) (dma_base_addr[id]) @@ -61,12 +64,6 @@ static u32 sh_dmae_readl(struct sh_dmae_chan *sh_dc, u32 reg) return ctrl_inl(SH_DMAC_CHAN_BASE(sh_dc->id) + reg); } -static void dmae_init(struct sh_dmae_chan *sh_chan) -{ - u32 chcr = RS_DEFAULT; /* default is DUAL mode */ - sh_dmae_writel(sh_chan, chcr, CHCR); -} - /* * Reset DMA controller * @@ -106,9 +103,8 @@ static bool dmae_is_busy(struct sh_dmae_chan *sh_chan) } static unsigned int ts_shift[] = TS_SHIFT; -static inline unsigned int calc_xmit_shift(struct sh_dmae_chan *sh_chan) +static inline unsigned int calc_xmit_shift(u32 chcr) { - u32 chcr = sh_dmae_readl(sh_chan, CHCR); int cnt = ((chcr & CHCR_TS_LOW_MASK) >> CHCR_TS_LOW_SHIFT) | ((chcr & CHCR_TS_HIGH_MASK) >> CHCR_TS_HIGH_SHIFT); @@ -119,7 +115,7 @@ static void dmae_set_reg(struct sh_dmae_chan *sh_chan, struct sh_dmae_regs *hw) { sh_dmae_writel(sh_chan, hw->sar, SAR); sh_dmae_writel(sh_chan, hw->dar, DAR); - sh_dmae_writel(sh_chan, hw->tcr >> calc_xmit_shift(sh_chan), TCR); + sh_dmae_writel(sh_chan, hw->tcr >> sh_chan->xmit_shift, TCR); } static void dmae_start(struct sh_dmae_chan *sh_chan) @@ -127,7 +123,7 @@ static void dmae_start(struct sh_dmae_chan *sh_chan) u32 chcr = sh_dmae_readl(sh_chan, CHCR); chcr |= CHCR_DE | CHCR_IE; - sh_dmae_writel(sh_chan, chcr, CHCR); + sh_dmae_writel(sh_chan, chcr & ~CHCR_TE, CHCR); } static void dmae_halt(struct sh_dmae_chan *sh_chan) @@ -138,20 +134,27 @@ static void dmae_halt(struct sh_dmae_chan *sh_chan) sh_dmae_writel(sh_chan, chcr, CHCR); } +static void dmae_init(struct sh_dmae_chan *sh_chan) +{ + u32 chcr = RS_DEFAULT; /* default is DUAL mode */ + sh_chan->xmit_shift = calc_xmit_shift(chcr); + sh_dmae_writel(sh_chan, chcr, CHCR); +} + static int dmae_set_chcr(struct sh_dmae_chan *sh_chan, u32 val) { /* When DMA was working, can not set data to CHCR */ if (dmae_is_busy(sh_chan)) return -EBUSY; + sh_chan->xmit_shift = calc_xmit_shift(val); sh_dmae_writel(sh_chan, val, CHCR); + return 0; } -#define DMARS1_ADDR 0x04 -#define DMARS2_ADDR 0x08 -#define DMARS_SHIFT 8 -#define DMARS_CHAN_MSK 0x01 +#define DMARS_SHIFT 8 +#define DMARS_CHAN_MSK 0x01 static int dmae_set_dmars(struct sh_dmae_chan *sh_chan, u16 val) { u32 addr; @@ -163,29 +166,18 @@ static int dmae_set_dmars(struct sh_dmae_chan *sh_chan, u16 val) if (sh_chan->id & DMARS_CHAN_MSK) shift = DMARS_SHIFT; - switch (sh_chan->id) { - /* DMARS0 */ - case 0: - case 1: - addr = SH_DMARS_BASE; - break; - /* DMARS1 */ - case 2: - case 3: - addr = (SH_DMARS_BASE + DMARS1_ADDR); - break; - /* DMARS2 */ - case 4: - case 5: - addr = (SH_DMARS_BASE + DMARS2_ADDR); - break; - default: + if (sh_chan->id < 6) + /* DMA0RS0 - DMA0RS2 */ + addr = SH_DMARS_BASE0 + (sh_chan->id / 2) * 4; +#ifdef SH_DMARS_BASE1 + else if (sh_chan->id < 12) + /* DMA1RS0 - DMA1RS2 */ + addr = SH_DMARS_BASE1 + ((sh_chan->id - 6) / 2) * 4; +#endif + else return -EINVAL; - } - ctrl_outw((val << shift) | - (ctrl_inw(addr) & (shift ? 0xFF00 : 0x00FF)), - addr); + ctrl_outw((val << shift) | (ctrl_inw(addr) & (0xFF00 >> shift)), addr); return 0; } @@ -253,10 +245,53 @@ static struct sh_desc *sh_dmae_get_desc(struct sh_dmae_chan *sh_chan) return NULL; } +static struct sh_dmae_slave_config *sh_dmae_find_slave( + struct sh_dmae_chan *sh_chan, enum sh_dmae_slave_chan_id slave_id) +{ + struct dma_device *dma_dev = sh_chan->common.device; + struct sh_dmae_device *shdev = container_of(dma_dev, + struct sh_dmae_device, common); + struct sh_dmae_pdata *pdata = &shdev->pdata; + int i; + + if ((unsigned)slave_id >= SHDMA_SLAVE_NUMBER) + return NULL; + + for (i = 0; i < pdata->config_num; i++) + if (pdata->config[i].slave_id == slave_id) + return pdata->config + i; + + return NULL; +} + static int sh_dmae_alloc_chan_resources(struct dma_chan *chan) { struct sh_dmae_chan *sh_chan = to_sh_chan(chan); struct sh_desc *desc; + struct sh_dmae_slave *param = chan->private; + + /* + * This relies on the guarantee from dmaengine that alloc_chan_resources + * never runs concurrently with itself or free_chan_resources. + */ + if (param) { + struct sh_dmae_slave_config *cfg; + + cfg = sh_dmae_find_slave(sh_chan, param->slave_id); + if (!cfg) + return -EINVAL; + + if (test_and_set_bit(param->slave_id, sh_dmae_slave_used)) + return -EBUSY; + + param->config = cfg; + + dmae_set_dmars(sh_chan, cfg->mid_rid); + dmae_set_chcr(sh_chan, cfg->chcr); + } else { + if ((sh_dmae_readl(sh_chan, CHCR) & 0x700) != 0x400) + dmae_set_chcr(sh_chan, RS_DEFAULT); + } spin_lock_bh(&sh_chan->desc_lock); while (sh_chan->descs_allocated < NR_DESCS_PER_CHANNEL) { @@ -289,10 +324,18 @@ static void sh_dmae_free_chan_resources(struct dma_chan *chan) struct sh_desc *desc, *_desc; LIST_HEAD(list); + dmae_halt(sh_chan); + /* Prepared and not submitted descriptors can still be on the queue */ if (!list_empty(&sh_chan->ld_queue)) sh_dmae_chan_ld_cleanup(sh_chan, true); + if (chan->private) { + /* The caller is holding dma_list_mutex */ + struct sh_dmae_slave *param = chan->private; + clear_bit(param->slave_id, sh_dmae_slave_used); + } + spin_lock_bh(&sh_chan->desc_lock); list_splice_init(&sh_chan->ld_free, &list); @@ -304,7 +347,7 @@ static void sh_dmae_free_chan_resources(struct dma_chan *chan) kfree(desc); } -/* +/** * sh_dmae_add_desc - get, set up and return one transfer descriptor * @sh_chan: DMA channel * @flags: DMA transfer flags @@ -351,12 +394,14 @@ static struct sh_desc *sh_dmae_add_desc(struct sh_dmae_chan *sh_chan, new->async_tx.cookie = -EINVAL; } - dev_dbg(sh_chan->dev, "chaining (%u/%u)@%x -> %x with %p, cookie %d\n", + dev_dbg(sh_chan->dev, + "chaining (%u/%u)@%x -> %x with %p, cookie %d, shift %d\n", copy_size, *len, *src, *dest, &new->async_tx, - new->async_tx.cookie); + new->async_tx.cookie, sh_chan->xmit_shift); new->mark = DESC_PREPARED; new->async_tx.flags = flags; + new->direction = direction; *len -= copy_size; if (direction == DMA_BIDIRECTIONAL || direction == DMA_TO_DEVICE) @@ -465,6 +510,8 @@ static struct dma_async_tx_descriptor *sh_dmae_prep_memcpy( if (!chan || !len) return NULL; + chan->private = NULL; + sh_chan = to_sh_chan(chan); sg_init_table(&sg, 1); @@ -477,6 +524,44 @@ static struct dma_async_tx_descriptor *sh_dmae_prep_memcpy( flags); } +static struct dma_async_tx_descriptor *sh_dmae_prep_slave_sg( + struct dma_chan *chan, struct scatterlist *sgl, unsigned int sg_len, + enum dma_data_direction direction, unsigned long flags) +{ + struct sh_dmae_slave *param; + struct sh_dmae_chan *sh_chan; + + if (!chan) + return NULL; + + sh_chan = to_sh_chan(chan); + param = chan->private; + + /* Someone calling slave DMA on a public channel? */ + if (!param || !sg_len) { + dev_warn(sh_chan->dev, "%s: bad parameter: %p, %d, %d\n", + __func__, param, sg_len, param ? param->slave_id : -1); + return NULL; + } + + /* + * if (param != NULL), this is a successfully requested slave channel, + * therefore param->config != NULL too. + */ + return sh_dmae_prep_sg(sh_chan, sgl, sg_len, ¶m->config->addr, + direction, flags); +} + +static void sh_dmae_terminate_all(struct dma_chan *chan) +{ + struct sh_dmae_chan *sh_chan = to_sh_chan(chan); + + if (!chan) + return; + + sh_dmae_chan_ld_cleanup(sh_chan, true); +} + static dma_async_tx_callback __ld_cleanup(struct sh_dmae_chan *sh_chan, bool all) { struct sh_desc *desc, *_desc; @@ -508,7 +593,11 @@ static dma_async_tx_callback __ld_cleanup(struct sh_dmae_chan *sh_chan, bool all cookie = tx->cookie; if (desc->mark == DESC_COMPLETED && desc->chunks == 1) { - BUG_ON(sh_chan->completed_cookie != desc->cookie - 1); + if (sh_chan->completed_cookie != desc->cookie - 1) + dev_dbg(sh_chan->dev, + "Completing cookie %d, expected %d\n", + desc->cookie, + sh_chan->completed_cookie + 1); sh_chan->completed_cookie = desc->cookie; } @@ -581,7 +670,7 @@ static void sh_chan_xfer_ld_queue(struct sh_dmae_chan *sh_chan) return; } - /* Find the first un-transfer desciptor */ + /* Find the first not transferred desciptor */ list_for_each_entry(sd, &sh_chan->ld_queue, node) if (sd->mark == DESC_SUBMITTED) { /* Get the ld start address from ld_queue */ @@ -685,11 +774,14 @@ static void dmae_do_tasklet(unsigned long data) struct sh_dmae_chan *sh_chan = (struct sh_dmae_chan *)data; struct sh_desc *desc; u32 sar_buf = sh_dmae_readl(sh_chan, SAR); + u32 dar_buf = sh_dmae_readl(sh_chan, DAR); spin_lock(&sh_chan->desc_lock); list_for_each_entry(desc, &sh_chan->ld_queue, node) { - if ((desc->hw.sar + desc->hw.tcr) == sar_buf && - desc->mark == DESC_SUBMITTED) { + if (desc->mark == DESC_SUBMITTED && + ((desc->direction == DMA_FROM_DEVICE && + (desc->hw.dar + desc->hw.tcr) == dar_buf) || + (desc->hw.sar + desc->hw.tcr) == sar_buf)) { dev_dbg(sh_chan->dev, "done #%d@%p dst %u\n", desc->async_tx.cookie, &desc->async_tx, desc->hw.dar); @@ -762,7 +854,7 @@ static int __devinit sh_dmae_chan_probe(struct sh_dmae_device *shdev, int id) } snprintf(new_sh_chan->dev_id, sizeof(new_sh_chan->dev_id), - "sh-dmae%d", new_sh_chan->id); + "sh-dmae%d", new_sh_chan->id); /* set up channel irq */ err = request_irq(irq, &sh_dmae_interrupt, irqflags, @@ -773,11 +865,6 @@ static int __devinit sh_dmae_chan_probe(struct sh_dmae_device *shdev, int id) goto err_no_irq; } - /* CHCR register control function */ - new_sh_chan->set_chcr = dmae_set_chcr; - /* DMARS register control function */ - new_sh_chan->set_dmars = dmae_set_dmars; - shdev->chan[id] = new_sh_chan; return 0; @@ -848,12 +935,19 @@ static int __init sh_dmae_probe(struct platform_device *pdev) INIT_LIST_HEAD(&shdev->common.channels); dma_cap_set(DMA_MEMCPY, shdev->common.cap_mask); + dma_cap_set(DMA_SLAVE, shdev->common.cap_mask); + shdev->common.device_alloc_chan_resources = sh_dmae_alloc_chan_resources; shdev->common.device_free_chan_resources = sh_dmae_free_chan_resources; shdev->common.device_prep_dma_memcpy = sh_dmae_prep_memcpy; shdev->common.device_is_tx_complete = sh_dmae_is_complete; shdev->common.device_issue_pending = sh_dmae_memcpy_issue_pending; + + /* Compulsory for DMA_SLAVE fields */ + shdev->common.device_prep_slave_sg = sh_dmae_prep_slave_sg; + shdev->common.device_terminate_all = sh_dmae_terminate_all; + shdev->common.dev = &pdev->dev; /* Default transfer size of 32 bytes requires 32-byte alignment */ shdev->common.copy_align = 5; diff --git a/drivers/dma/shdma.h b/drivers/dma/shdma.h index 108f1cf..7e227f3 100644 --- a/drivers/dma/shdma.h +++ b/drivers/dma/shdma.h @@ -29,6 +29,7 @@ struct sh_desc { struct sh_dmae_regs hw; struct list_head node; struct dma_async_tx_descriptor async_tx; + enum dma_data_direction direction; dma_cookie_t cookie; int chunks; int mark; @@ -45,13 +46,9 @@ struct sh_dmae_chan { struct device *dev; /* Channel device */ struct tasklet_struct tasklet; /* Tasklet */ int descs_allocated; /* desc count */ + int xmit_shift; /* log_2(bytes_per_xfer) */ int id; /* Raw id of this channel */ char dev_id[16]; /* unique name per DMAC of channel */ - - /* Set chcr */ - int (*set_chcr)(struct sh_dmae_chan *sh_chan, u32 regs); - /* Set DMA resource */ - int (*set_dmars)(struct sh_dmae_chan *sh_chan, u16 res); }; struct sh_dmae_device { -- cgit v0.10.2 From 858918b77b29d0e9ce7f524d1b57d602d85f5d64 Mon Sep 17 00:00:00 2001 From: Matt Fleming Date: Sun, 7 Feb 2010 12:40:36 +0000 Subject: sh: Optimise FDE/CIE lookup by using red-black trees Now that the DWARF unwinder is being used to provide perf callstacks unwinding speed is an issue. It is no longer being used in exceptional circumstances where we don't care about runtime performance, e.g. when panicing, so it makes sense improve performance is possible. With this patch I saw a 42% improvement in unwind time when calling return_address(1). Greater improvements will be seen as the number of levels unwound increases as each unwind is now cheaper. Note that insertion time has doubled but that's just the price we pay for keeping the trees balanced. However, this is a one-time cost for kernel boot/module load and so the improvements in lookup time dominate the extra time we spend keeping the trees balanced. Signed-off-by: Matt Fleming Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/dwarf.h b/arch/sh/include/asm/dwarf.h index bdccbbf..d62abd1 100644 --- a/arch/sh/include/asm/dwarf.h +++ b/arch/sh/include/asm/dwarf.h @@ -243,16 +243,13 @@ struct dwarf_cie { unsigned long cie_pointer; - struct list_head link; - unsigned long flags; #define DWARF_CIE_Z_AUGMENTATION (1 << 0) - /* - * 'mod' will be non-NULL if this CIE came from a module's - * .eh_frame section. - */ - struct module *mod; + /* linked-list entry if this CIE is from a module */ + struct list_head link; + + struct rb_node node; }; /** @@ -266,13 +263,11 @@ struct dwarf_fde { unsigned long address_range; unsigned char *instructions; unsigned char *end; + + /* linked-list entry if this FDE is from a module */ struct list_head link; - /* - * 'mod' will be non-NULL if this FDE came from a module's - * .eh_frame section. - */ - struct module *mod; + struct rb_node node; }; /** diff --git a/arch/sh/include/asm/module.h b/arch/sh/include/asm/module.h index 068bf16..b7927de 100644 --- a/arch/sh/include/asm/module.h +++ b/arch/sh/include/asm/module.h @@ -1,7 +1,22 @@ #ifndef _ASM_SH_MODULE_H #define _ASM_SH_MODULE_H -#include +struct mod_arch_specific { +#ifdef CONFIG_DWARF_UNWINDER + struct list_head fde_list; + struct list_head cie_list; +#endif +}; + +#ifdef CONFIG_64BIT +#define Elf_Shdr Elf64_Shdr +#define Elf_Sym Elf64_Sym +#define Elf_Ehdr Elf64_Ehdr +#else +#define Elf_Shdr Elf32_Shdr +#define Elf_Sym Elf32_Sym +#define Elf_Ehdr Elf32_Ehdr +#endif #ifdef CONFIG_CPU_LITTLE_ENDIAN # ifdef CONFIG_CPU_SH2 diff --git a/arch/sh/kernel/dwarf.c b/arch/sh/kernel/dwarf.c index e511680..bd1c497 100644 --- a/arch/sh/kernel/dwarf.c +++ b/arch/sh/kernel/dwarf.c @@ -39,10 +39,10 @@ static mempool_t *dwarf_frame_pool; static struct kmem_cache *dwarf_reg_cachep; static mempool_t *dwarf_reg_pool; -static LIST_HEAD(dwarf_cie_list); +static struct rb_root cie_root; static DEFINE_SPINLOCK(dwarf_cie_lock); -static LIST_HEAD(dwarf_fde_list); +static struct rb_root fde_root; static DEFINE_SPINLOCK(dwarf_fde_lock); static struct dwarf_cie *cached_cie; @@ -301,7 +301,8 @@ static inline int dwarf_entry_len(char *addr, unsigned long *len) */ static struct dwarf_cie *dwarf_lookup_cie(unsigned long cie_ptr) { - struct dwarf_cie *cie; + struct rb_node **rb_node = &cie_root.rb_node; + struct dwarf_cie *cie = NULL; unsigned long flags; spin_lock_irqsave(&dwarf_cie_lock, flags); @@ -315,16 +316,24 @@ static struct dwarf_cie *dwarf_lookup_cie(unsigned long cie_ptr) goto out; } - list_for_each_entry(cie, &dwarf_cie_list, link) { - if (cie->cie_pointer == cie_ptr) { - cached_cie = cie; - break; + while (*rb_node) { + struct dwarf_cie *cie_tmp; + + cie_tmp = rb_entry(*rb_node, struct dwarf_cie, node); + BUG_ON(!cie_tmp); + + if (cie_ptr == cie_tmp->cie_pointer) { + cie = cie_tmp; + cached_cie = cie_tmp; + goto out; + } else { + if (cie_ptr < cie_tmp->cie_pointer) + rb_node = &(*rb_node)->rb_left; + else + rb_node = &(*rb_node)->rb_right; } } - /* Couldn't find the entry in the list. */ - if (&cie->link == &dwarf_cie_list) - cie = NULL; out: spin_unlock_irqrestore(&dwarf_cie_lock, flags); return cie; @@ -336,25 +345,34 @@ out: */ struct dwarf_fde *dwarf_lookup_fde(unsigned long pc) { - struct dwarf_fde *fde; + struct rb_node **rb_node = &fde_root.rb_node; + struct dwarf_fde *fde = NULL; unsigned long flags; spin_lock_irqsave(&dwarf_fde_lock, flags); - list_for_each_entry(fde, &dwarf_fde_list, link) { - unsigned long start, end; + while (*rb_node) { + struct dwarf_fde *fde_tmp; + unsigned long tmp_start, tmp_end; - start = fde->initial_location; - end = fde->initial_location + fde->address_range; + fde_tmp = rb_entry(*rb_node, struct dwarf_fde, node); + BUG_ON(!fde_tmp); - if (pc >= start && pc < end) - break; - } + tmp_start = fde_tmp->initial_location; + tmp_end = fde_tmp->initial_location + fde_tmp->address_range; - /* Couldn't find the entry in the list. */ - if (&fde->link == &dwarf_fde_list) - fde = NULL; + if (pc < tmp_start) { + rb_node = &(*rb_node)->rb_left; + } else { + if (pc < tmp_end) { + fde = fde_tmp; + goto out; + } else + rb_node = &(*rb_node)->rb_right; + } + } +out: spin_unlock_irqrestore(&dwarf_fde_lock, flags); return fde; @@ -552,8 +570,8 @@ extern void ret_from_irq(void); * on the callstack. Each of the lower (older) stack frames are * linked via the "prev" member. */ -struct dwarf_frame * dwarf_unwind_stack(unsigned long pc, - struct dwarf_frame *prev) +struct dwarf_frame *dwarf_unwind_stack(unsigned long pc, + struct dwarf_frame *prev) { struct dwarf_frame *frame; struct dwarf_cie *cie; @@ -708,6 +726,8 @@ bail: static int dwarf_parse_cie(void *entry, void *p, unsigned long len, unsigned char *end, struct module *mod) { + struct rb_node **rb_node = &cie_root.rb_node; + struct rb_node *parent; struct dwarf_cie *cie; unsigned long flags; int count; @@ -802,11 +822,30 @@ static int dwarf_parse_cie(void *entry, void *p, unsigned long len, cie->initial_instructions = p; cie->instructions_end = end; - cie->mod = mod; - /* Add to list */ spin_lock_irqsave(&dwarf_cie_lock, flags); - list_add_tail(&cie->link, &dwarf_cie_list); + + while (*rb_node) { + struct dwarf_cie *cie_tmp; + + cie_tmp = rb_entry(*rb_node, struct dwarf_cie, node); + + parent = *rb_node; + + if (cie->cie_pointer < cie_tmp->cie_pointer) + rb_node = &parent->rb_left; + else if (cie->cie_pointer >= cie_tmp->cie_pointer) + rb_node = &parent->rb_right; + else + WARN_ON(1); + } + + rb_link_node(&cie->node, parent, rb_node); + rb_insert_color(&cie->node, &cie_root); + + if (mod != NULL) + list_add_tail(&cie->link, &mod->arch.cie_list); + spin_unlock_irqrestore(&dwarf_cie_lock, flags); return 0; @@ -816,6 +855,8 @@ static int dwarf_parse_fde(void *entry, u32 entry_type, void *start, unsigned long len, unsigned char *end, struct module *mod) { + struct rb_node **rb_node = &fde_root.rb_node; + struct rb_node *parent; struct dwarf_fde *fde; struct dwarf_cie *cie; unsigned long flags; @@ -863,11 +904,38 @@ static int dwarf_parse_fde(void *entry, u32 entry_type, fde->instructions = p; fde->end = end; - fde->mod = mod; - /* Add to list. */ spin_lock_irqsave(&dwarf_fde_lock, flags); - list_add_tail(&fde->link, &dwarf_fde_list); + + while (*rb_node) { + struct dwarf_fde *fde_tmp; + unsigned long tmp_start, tmp_end; + unsigned long start, end; + + fde_tmp = rb_entry(*rb_node, struct dwarf_fde, node); + + start = fde->initial_location; + end = fde->initial_location + fde->address_range; + + tmp_start = fde_tmp->initial_location; + tmp_end = fde_tmp->initial_location + fde_tmp->address_range; + + parent = *rb_node; + + if (start < tmp_start) + rb_node = &parent->rb_left; + else if (start >= tmp_end) + rb_node = &parent->rb_right; + else + WARN_ON(1); + } + + rb_link_node(&fde->node, parent, rb_node); + rb_insert_color(&fde->node, &fde_root); + + if (mod != NULL) + list_add_tail(&fde->link, &mod->arch.fde_list); + spin_unlock_irqrestore(&dwarf_fde_lock, flags); return 0; @@ -912,19 +980,29 @@ static struct unwinder dwarf_unwinder = { static void dwarf_unwinder_cleanup(void) { - struct dwarf_cie *cie, *cie_tmp; - struct dwarf_fde *fde, *fde_tmp; + struct rb_node **fde_rb_node = &fde_root.rb_node; + struct rb_node **cie_rb_node = &cie_root.rb_node; /* * Deallocate all the memory allocated for the DWARF unwinder. * Traverse all the FDE/CIE lists and remove and free all the * memory associated with those data structures. */ - list_for_each_entry_safe(cie, cie_tmp, &dwarf_cie_list, link) - kfree(cie); + while (*fde_rb_node) { + struct dwarf_fde *fde; - list_for_each_entry_safe(fde, fde_tmp, &dwarf_fde_list, link) + fde = rb_entry(*fde_rb_node, struct dwarf_fde, node); + rb_erase(*fde_rb_node, &fde_root); kfree(fde); + } + + while (*cie_rb_node) { + struct dwarf_cie *cie; + + cie = rb_entry(*cie_rb_node, struct dwarf_cie, node); + rb_erase(*cie_rb_node, &cie_root); + kfree(cie); + } kmem_cache_destroy(dwarf_reg_cachep); kmem_cache_destroy(dwarf_frame_cachep); @@ -1024,6 +1102,8 @@ int module_dwarf_finalize(const Elf_Ehdr *hdr, const Elf_Shdr *sechdrs, /* Did we find the .eh_frame section? */ if (i != hdr->e_shnum) { + INIT_LIST_HEAD(&me->arch.cie_list); + INIT_LIST_HEAD(&me->arch.fde_list); err = dwarf_parse_section((char *)start, (char *)end, me); if (err) { printk(KERN_WARNING "%s: failed to parse DWARF info\n", @@ -1044,38 +1124,26 @@ int module_dwarf_finalize(const Elf_Ehdr *hdr, const Elf_Shdr *sechdrs, */ void module_dwarf_cleanup(struct module *mod) { - struct dwarf_fde *fde; - struct dwarf_cie *cie; + struct dwarf_fde *fde, *ftmp; + struct dwarf_cie *cie, *ctmp; unsigned long flags; spin_lock_irqsave(&dwarf_cie_lock, flags); -again_cie: - list_for_each_entry(cie, &dwarf_cie_list, link) { - if (cie->mod == mod) - break; - } - - if (&cie->link != &dwarf_cie_list) { + list_for_each_entry_safe(cie, ctmp, &mod->arch.cie_list, link) { list_del(&cie->link); + rb_erase(&cie->node, &cie_root); kfree(cie); - goto again_cie; } spin_unlock_irqrestore(&dwarf_cie_lock, flags); spin_lock_irqsave(&dwarf_fde_lock, flags); -again_fde: - list_for_each_entry(fde, &dwarf_fde_list, link) { - if (fde->mod == mod) - break; - } - - if (&fde->link != &dwarf_fde_list) { + list_for_each_entry_safe(fde, ftmp, &mod->arch.fde_list, link) { list_del(&fde->link); + rb_erase(&fde->node, &fde_root); kfree(fde); - goto again_fde; } spin_unlock_irqrestore(&dwarf_fde_lock, flags); @@ -1094,8 +1162,6 @@ again_fde: static int __init dwarf_unwinder_init(void) { int err; - INIT_LIST_HEAD(&dwarf_cie_list); - INIT_LIST_HEAD(&dwarf_fde_list); dwarf_frame_cachep = kmem_cache_create("dwarf_frames", sizeof(struct dwarf_frame), 0, -- cgit v0.10.2 From 9e9622d1a064705181bea0600ea9eacd95adab7f Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Mon, 8 Feb 2010 11:47:44 +0900 Subject: serial: sh-sci: remove SCIF code in sci_rxd_in() Since sci_rxd_in() is used by SCI only, clean up the header file by killing off code dealing with SCIF ports and their register definitions. Also introduce a default sci_rxd_in() function which can be shared by all SCIF-only processors. Signed-off-by: Magnus Damm Signed-off-by: Paul Mundt diff --git a/drivers/serial/sh-sci.h b/drivers/serial/sh-sci.h index 0efcded..f7d2589 100644 --- a/drivers/serial/sh-sci.h +++ b/drivers/serial/sh-sci.h @@ -518,34 +518,6 @@ static inline int sci_rxd_in(struct uart_port *port) { if (port->mapbase == 0xfffffe80) return __raw_readb(SCPDR)&0x01 ? 1 : 0; /* SCI */ - if (port->mapbase == 0xa4000150) - return __raw_readb(SCPDR)&0x10 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xa4000140) - return __raw_readb(SCPDR)&0x04 ? 1 : 0; /* IRDA */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7705) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == SCIF0) - return __raw_readb(SCPDR)&0x04 ? 1 : 0; /* IRDA */ - if (port->mapbase == SCIF2) - return __raw_readb(SCPDR)&0x10 ? 1 : 0; /* SCIF */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7710) || defined(CONFIG_CPU_SUBTYPE_SH7712) -static inline int sci_rxd_in(struct uart_port *port) -{ - return sci_in(port,SCxSR)&0x0010 ? 1 : 0; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7720) || \ - defined(CONFIG_CPU_SUBTYPE_SH7721) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xa4430000) - return sci_in(port, SCxSR) & 0x0003 ? 1 : 0; - else if (port->mapbase == 0xa4438000) - return sci_in(port, SCxSR) & 0x0003 ? 1 : 0; return 1; } #elif defined(CONFIG_CPU_SUBTYPE_SH7750) || \ @@ -558,207 +530,17 @@ static inline int sci_rxd_in(struct uart_port *port) { if (port->mapbase == 0xffe00000) return __raw_readb(SCSPTR1)&0x01 ? 1 : 0; /* SCI */ - if (port->mapbase == 0xffe80000) - return __raw_readw(SCSPTR2)&0x0001 ? 1 : 0; /* SCIF */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH4_202) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xffe80000) - return __raw_readw(SCSPTR2)&0x0001 ? 1 : 0; /* SCIF */ return 1; } -#elif defined(CONFIG_CPU_SUBTYPE_SH7757) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xfe4b0000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; - if (port->mapbase == 0xfe4c0000) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; - if (port->mapbase == 0xfe4d0000) - return __raw_readw(SCSPTR2) & 0x0001 ? 1 : 0; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7760) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xfe600000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xfe610000) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xfe620000) - return __raw_readw(SCSPTR2) & 0x0001 ? 1 : 0; /* SCIF */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7343) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xffe00000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffe10000) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffe20000) - return __raw_readw(SCSPTR2) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffe30000) - return __raw_readw(SCSPTR3) & 0x0001 ? 1 : 0; /* SCIF */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7366) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xffe00000) - return __raw_readb(SCPDR0) & 0x0001 ? 1 : 0; /* SCIF0 */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7722) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xffe00000) - return __raw_readb(PSDR) & 0x02 ? 1 : 0; /* SCIF0 */ - if (port->mapbase == 0xffe10000) - return __raw_readb(PADR) & 0x40 ? 1 : 0; /* SCIF1 */ - if (port->mapbase == 0xffe20000) - return __raw_readb(PWDR) & 0x04 ? 1 : 0; /* SCIF2 */ - - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7723) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xffe00000) - return __raw_readb(SCSPTR0) & 0x0008 ? 1 : 0; /* SCIF0 */ - if (port->mapbase == 0xffe10000) - return __raw_readb(SCSPTR1) & 0x0020 ? 1 : 0; /* SCIF1 */ - if (port->mapbase == 0xffe20000) - return __raw_readb(SCSPTR2) & 0x0001 ? 1 : 0; /* SCIF2 */ - if (port->mapbase == 0xa4e30000) - return __raw_readb(SCSPTR3) & 0x0001 ? 1 : 0; /* SCIF3 */ - if (port->mapbase == 0xa4e40000) - return __raw_readb(SCSPTR4) & 0x0001 ? 1 : 0; /* SCIF4 */ - if (port->mapbase == 0xa4e50000) - return __raw_readb(SCSPTR5) & 0x0008 ? 1 : 0; /* SCIF5 */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7724) -# define SCFSR 0x0010 -# define SCASSR 0x0014 -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->type == PORT_SCIF) - return __raw_readw((port->mapbase + SCFSR)) & SCIF_BRK ? 1 : 0; - if (port->type == PORT_SCIFA) - return __raw_readw((port->mapbase + SCASSR)) & SCIF_BRK ? 1 : 0; - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH5_101) || defined(CONFIG_CPU_SUBTYPE_SH5_103) -static inline int sci_rxd_in(struct uart_port *port) -{ - return sci_in(port, SCSPTR)&0x0001 ? 1 : 0; /* SCIF */ -} #elif defined(__H8300H__) || defined(__H8300S__) static inline int sci_rxd_in(struct uart_port *port) { int ch = (port->mapbase - SMR0) >> 3; return (H8300_SCI_DR(ch) & h8300_sci_pins[ch].rx) ? 1 : 0; } -#elif defined(CONFIG_CPU_SUBTYPE_SH7763) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xffe00000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffe08000) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffe10000) - return __raw_readw(SCSPTR2) & 0x0001 ? 1 : 0; /* SCIF/IRDA */ - - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7770) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xff923000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xff924000) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xff925000) - return __raw_readw(SCSPTR2) & 0x0001 ? 1 : 0; /* SCIF */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7780) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xffe00000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffe10000) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; /* SCIF */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7785) || \ - defined(CONFIG_CPU_SUBTYPE_SH7786) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xffea0000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffeb0000) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffec0000) - return __raw_readw(SCSPTR2) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffed0000) - return __raw_readw(SCSPTR3) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffee0000) - return __raw_readw(SCSPTR4) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffef0000) - return __raw_readw(SCSPTR5) & 0x0001 ? 1 : 0; /* SCIF */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7201) || \ - defined(CONFIG_CPU_SUBTYPE_SH7203) || \ - defined(CONFIG_CPU_SUBTYPE_SH7206) || \ - defined(CONFIG_CPU_SUBTYPE_SH7263) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xfffe8000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xfffe8800) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xfffe9000) - return __raw_readw(SCSPTR2) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xfffe9800) - return __raw_readw(SCSPTR3) & 0x0001 ? 1 : 0; /* SCIF */ -#if defined(CONFIG_CPU_SUBTYPE_SH7201) - if (port->mapbase == 0xfffeA000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xfffeA800) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xfffeB000) - return __raw_readw(SCSPTR2) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xfffeB800) - return __raw_readw(SCSPTR3) & 0x0001 ? 1 : 0; /* SCIF */ -#endif - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SH7619) -static inline int sci_rxd_in(struct uart_port *port) -{ - if (port->mapbase == 0xf8400000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xf8410000) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xf8420000) - return __raw_readw(SCSPTR2) & 0x0001 ? 1 : 0; /* SCIF */ - return 1; -} -#elif defined(CONFIG_CPU_SUBTYPE_SHX3) +#else /* default case for non-SCI processors */ static inline int sci_rxd_in(struct uart_port *port) { - if (port->mapbase == 0xffc30000) - return __raw_readw(SCSPTR0) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffc40000) - return __raw_readw(SCSPTR1) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffc50000) - return __raw_readw(SCSPTR2) & 0x0001 ? 1 : 0; /* SCIF */ - if (port->mapbase == 0xffc60000) - return __raw_readw(SCSPTR3) & 0x0001 ? 1 : 0; /* SCIF */ return 1; } #endif -- cgit v0.10.2 From 7561f2dd393bd0c6397e6b2a6b021cdb827a2eb1 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 8 Feb 2010 16:36:56 +0900 Subject: sh: Fix up SH7786 PCI resource definitions. This adds in some of the missing memory resources for channels 1/2 and gets the code building again for the recent changes. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/pcie-sh7786.c b/arch/sh/drivers/pci/pcie-sh7786.c index feac1fe..95d095f 100644 --- a/arch/sh/drivers/pci/pcie-sh7786.c +++ b/arch/sh/drivers/pci/pcie-sh7786.c @@ -1,7 +1,7 @@ /* * Low-Level PCI Express Support for the SH7786 * - * Copyright (C) 2009 Paul Mundt + * Copyright (C) 2009 - 2010 Paul Mundt * * This file is subject to the terms and conditions of the GNU General Public * License. See the file "COPYING" in the main directory of this archive @@ -30,60 +30,84 @@ static struct sh7786_pcie_hwops { int (*port_init_hw)(struct sh7786_pcie_port *port); } *sh7786_pcie_hwops; -static struct resource sh7786_pci_32bit_mem_resources[] = { +static struct resource sh7786_pci0_resources[] = { { - .name = "pci0_mem", - .start = SH4A_PCIMEM_BASEA, - .end = SH4A_PCIMEM_BASEA + SZ_64M - 1, - .flags = IORESOURCE_MEM, + .name = "PCIe0 IO", + .start = 0xfd000000, + .end = 0xfd000000 + SZ_8M - 1, + .flags = IORESOURCE_IO, }, { - .name = "pci1_mem", - .start = SH4A_PCIMEM_BASEA1, - .end = SH4A_PCIMEM_BASEA1 + SZ_64M - 1, - .flags = IORESOURCE_MEM, + .name = "PCIe0 MEM 0", + .start = 0xc0000000, + .end = 0xc0000000 + SZ_512M - 1, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_32BIT, }, { - .name = "pci2_mem", - .start = SH4A_PCIMEM_BASEA2, - .end = SH4A_PCIMEM_BASEA2 + SZ_64M - 1, + .name = "PCIe0 MEM 1", + .start = 0x10000000, + .end = 0x10000000 + SZ_64M - 1, .flags = IORESOURCE_MEM, + }, { + .name = "PCIe0 MEM 2", + .start = 0xfe100000, + .end = 0xfe100000 + SZ_1M - 1, }, }; -static struct resource sh7786_pci_29bit_mem_resource = { - .start = SH4A_PCIMEM_BASE, - .end = SH4A_PCIMEM_BASE + SZ_64M - 1, - .flags = IORESOURCE_MEM, +static struct resource sh7786_pci1_resources[] = { + { + .name = "PCIe1 IO", + .start = 0xfd800000, + .end = 0xfd800000 + SZ_8M - 1, + .flags = IORESOURCE_IO, + }, { + .name = "PCIe1 MEM 0", + .start = 0xa0000000, + .end = 0xa0000000 + SZ_512M - 1, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_32BIT, + }, { + .name = "PCIe1 MEM 1", + .start = 0x30000000, + .end = 0x30000000 + SZ_256M - 1, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_32BIT, + }, { + .name = "PCIe1 MEM 2", + .start = 0xfe300000, + .end = 0xfe300000 + SZ_1M - 1, + }, }; -static struct resource sh7786_pci_io_resources[] = { +static struct resource sh7786_pci2_resources[] = { { - .name = "pci0_io", - .start = SH4A_PCIIO_BASE, - .end = SH4A_PCIIO_BASE + SZ_8M - 1, - .flags = IORESOURCE_IO, + .name = "PCIe2 IO", + .start = 0xfc800000, + .end = 0xfc800000 + SZ_4M - 1, }, { - .name = "pci1_io", - .start = SH4A_PCIIO_BASE1, - .end = SH4A_PCIIO_BASE1 + SZ_8M - 1, - .flags = IORESOURCE_IO, + .name = "PCIe2 MEM 0", + .start = 0x80000000, + .end = 0x80000000 + SZ_512M - 1, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_32BIT, }, { - .name = "pci2_io", - .start = SH4A_PCIIO_BASE2, - .end = SH4A_PCIIO_BASE2 + SZ_4M - 1, - .flags = IORESOURCE_IO, + .name = "PCIe2 MEM 1", + .start = 0x20000000, + .end = 0x20000000 + SZ_256M - 1, + .flags = IORESOURCE_MEM | IORESOURCE_MEM_32BIT, + }, { + .name = "PCIe2 MEM 2", + .start = 0xfcd00000, + .end = 0xfcd00000 + SZ_1M - 1, }, }; extern struct pci_ops sh7786_pci_ops; -#define DEFINE_CONTROLLER(start, idx) \ -{ \ - .pci_ops = &sh7786_pci_ops, \ - .reg_base = start, \ - /* mem_resource filled in at probe time */ \ - .mem_offset = 0, \ - .io_resource = &sh7786_pci_io_resources[idx], \ - .io_offset = 0, \ +#define DEFINE_CONTROLLER(start, idx) \ +{ \ + .pci_ops = &sh7786_pci_ops, \ + .resources = sh7786_pci##idx##_resources, \ + .nr_resources = ARRAY_SIZE(sh7786_pci##idx##_resources), \ + .reg_base = start, \ + .mem_offset = 0, \ + .io_offset = 0, \ } static struct pci_channel sh7786_pci_channels[] = { @@ -330,17 +354,7 @@ static int __init sh7786_pcie_init(void) port->index = i; port->hose = sh7786_pci_channels + i; - port->hose->io_map_base = port->hose->io_resource->start; - - /* - * Check if we are booting in 29 or 32-bit mode - * - * 32-bit mode provides each controller with its own - * memory window, while 29-bit mode uses a shared one. - */ - port->hose->mem_resource = test_mode_pin(MODE_PIN10) ? - &sh7786_pci_32bit_mem_resources[i] : - &sh7786_pci_29bit_mem_resource; + port->hose->io_map_base = port->hose->resources[0].start; ret |= sh7786_pcie_hwops->port_init_hw(port); } diff --git a/arch/sh/drivers/pci/pcie-sh7786.h b/arch/sh/drivers/pci/pcie-sh7786.h index c655290..6666ea2 100644 --- a/arch/sh/drivers/pci/pcie-sh7786.h +++ b/arch/sh/drivers/pci/pcie-sh7786.h @@ -30,47 +30,9 @@ * for other(Max Payload Size=4096B,PCIIO_SIZE=8M) */ -/* PCI0-0: PCI I/O space */ -#define SH4A_PCIIO_BASE 0xFD000000 /* PCI I/O for controller 0 */ -#define SH4A_PCIIO_BASE1 0xFD800000 /* PCI I/O for controller 1 (Rev1.14)*/ -#define SH4A_PCIIO_BASE2 0xFC800000 /* PCI I/O for controller 2 (Rev1.171)*/ - -#define SH4A_PCIIO_SIZE64 0x00010000 /* PLX allows only 64K */ -#define SH4A_PCIIO_SIZE 0x00800000 /* 8M */ -#define SH4A_PCIIO_SIZE2 0x00400000 /* 4M (Rev1.171)*/ - -/* PCI0-1: PCI memory space 29-bit address */ -#define SH4A_PCIMEM_BASE 0x10000000 -#define SH4A_PCIMEM_SIZE 0x04000000 /* 64M */ - -/* PCI0-2: PCI memory space 32-bit address */ -#define SH4A_PCIMEM_BASEA 0xC0000000 /* for controller 0 */ -#define SH4A_PCIMEM_BASEA1 0xA0000000 /* for controller 1 (Rev1.14)*/ -#define SH4A_PCIMEM_BASEA2 0x80000000 /* for controller 2 (Rev1.171)*/ -#define SH4A_PCIMEM_SIZEA 0x20000000 /* 512M */ - /* PCI0: PCI memory target transfer 32-bit address translation value(Rev1.11T)*/ #define SH4A_PCIBMSTR_TRANSLATION 0x20000000 -#define SH4A_PCI_DEVICE_ID 0x0002 -#define SH4A_PCI_VENDOR_ID 0x1912 - -// PCI compatible 000-03f -#define PCI_CMD 0x004 -#define PCI_RID 0x008 -#define PCI_IBAR 0x010 -#define PCI_MBAR0 0x014 -#define PCI_MBAR1 0x018 - -/* PCI power management/MSI/capablity 040-0ff */ -/* PCIE extended 100-fff */ - -/* SH7786 device identification */ // Rev1.171 -#define SH4A_PVR (0xFF000030) -#define SH4A_PVR_SHX3 (0x10400000) -#define SH4A_PRR (0xFF000044) -#define SH4A_PRR_SH7786 (0x00000400) // Rev1.171 - /* SPVCR0 */ #define SH4A_PCIEVCR0 (0x000000) /* R - 0x0000 0000 32 */ #define BITS_TOP_MB (24) -- cgit v0.10.2 From 50200df462023b187d80a99a52f5f2cfe3c86c26 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Tue, 2 Feb 2010 14:46:13 -0800 Subject: kernel/sched.c: Suppress unused var warning On UP: kernel/sched.c: In function 'wake_up_new_task': kernel/sched.c:2631: warning: unused variable 'cpu' Signed-off-by: Andrew Morton Acked-by: Peter Zijlstra Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index 3a8fb30..e3199df 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2647,7 +2647,7 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) { unsigned long flags; struct rq *rq; - int cpu = get_cpu(); + int cpu __maybe_unused = get_cpu(); #ifdef CONFIG_SMP /* -- cgit v0.10.2 From 0c9cf2efd74dbc90354e2ccc7dbd6bad68ec6c4d Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Tue, 2 Feb 2010 14:46:10 -0800 Subject: percpu_counter: Make __percpu_counter_add an inline function on UP Even though batch isn't used on UP, we may want to pass one in to keep the SMP and UP code paths similar. Convert __percpu_counter_add to an inline function so we wont get variable unused warnings if we do. Signed-off-by: Anton Blanchard Cc: KOSAKI Motohiro Cc: Peter Zijlstra Cc: Martin Schwidefsky Cc: "Luck, Tony" Cc: Balbir Singh Signed-off-by: Andrew Morton Signed-off-by: Ingo Molnar diff --git a/include/linux/percpu_counter.h b/include/linux/percpu_counter.h index a7684a5..794662b 100644 --- a/include/linux/percpu_counter.h +++ b/include/linux/percpu_counter.h @@ -98,9 +98,6 @@ static inline void percpu_counter_set(struct percpu_counter *fbc, s64 amount) fbc->count = amount; } -#define __percpu_counter_add(fbc, amount, batch) \ - percpu_counter_add(fbc, amount) - static inline void percpu_counter_add(struct percpu_counter *fbc, s64 amount) { @@ -109,6 +106,12 @@ percpu_counter_add(struct percpu_counter *fbc, s64 amount) preempt_enable(); } +static inline void +__percpu_counter_add(struct percpu_counter *fbc, s64 amount, s32 batch) +{ + percpu_counter_add(fbc, amount); +} + static inline s64 percpu_counter_read(struct percpu_counter *fbc) { return fbc->count; -- cgit v0.10.2 From fa535a77bd3fa32b9215ba375d6a202fe73e1dd6 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Tue, 2 Feb 2010 14:46:13 -0800 Subject: sched: cpuacct: Use bigger percpu counter batch values for stats counters When CONFIG_VIRT_CPU_ACCOUNTING and CONFIG_CGROUP_CPUACCT are enabled we can call cpuacct_update_stats with values much larger than percpu_counter_batch. This means the call to percpu_counter_add will always add to the global count which is protected by a spinlock and we end up with a global spinlock in the scheduler. Based on an idea by KOSAKI Motohiro, this patch scales the batch value by cputime_one_jiffy such that we have the same batch limit as we would if CONFIG_VIRT_CPU_ACCOUNTING was disabled. His patch did this once at boot but that initialisation happened too early on PowerPC (before time_init) and it was never updated at runtime as a result of a hotplug cpu add/remove. This patch instead scales percpu_counter_batch by cputime_one_jiffy at runtime, which keeps the batch correct even after cpu hotplug operations. We cap it at INT_MAX in case of overflow. For architectures that do not support CONFIG_VIRT_CPU_ACCOUNTING, cputime_one_jiffy is the constant 1 and gcc is smart enough to optimise min(s32 percpu_counter_batch, INT_MAX) to just percpu_counter_batch at least on x86 and PowerPC. So there is no need to add an #ifdef. On a 64 thread PowerPC box with CONFIG_VIRT_CPU_ACCOUNTING and CONFIG_CGROUP_CPUACCT enabled, a context switch microbenchmark is 234x faster and almost matches a CONFIG_CGROUP_CPUACCT disabled kernel: CONFIG_CGROUP_CPUACCT disabled: 16906698 ctx switches/sec CONFIG_CGROUP_CPUACCT enabled: 61720 ctx switches/sec CONFIG_CGROUP_CPUACCT + patch: 16663217 ctx switches/sec Tested with: wget http://ozlabs.org/~anton/junkcode/context_switch.c make context_switch for i in `seq 0 63`; do taskset -c $i ./context_switch & done vmstat 1 Signed-off-by: Anton Blanchard Reviewed-by: KOSAKI Motohiro Acked-by: Balbir Singh Tested-by: Balbir Singh Cc: Peter Zijlstra Cc: Martin Schwidefsky Cc: "Luck, Tony" Signed-off-by: Andrew Morton Signed-off-by: Ingo Molnar diff --git a/kernel/sched.c b/kernel/sched.c index f96be93..bae6fcf 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -8998,12 +8998,30 @@ static void cpuacct_charge(struct task_struct *tsk, u64 cputime) } /* + * When CONFIG_VIRT_CPU_ACCOUNTING is enabled one jiffy can be very large + * in cputime_t units. As a result, cpuacct_update_stats calls + * percpu_counter_add with values large enough to always overflow the + * per cpu batch limit causing bad SMP scalability. + * + * To fix this we scale percpu_counter_batch by cputime_one_jiffy so we + * batch the same amount of time with CONFIG_VIRT_CPU_ACCOUNTING disabled + * and enabled. We cap it at INT_MAX which is the largest allowed batch value. + */ +#ifdef CONFIG_SMP +#define CPUACCT_BATCH \ + min_t(long, percpu_counter_batch * cputime_one_jiffy, INT_MAX) +#else +#define CPUACCT_BATCH 0 +#endif + +/* * Charge the system/user time to the task's accounting group. */ static void cpuacct_update_stats(struct task_struct *tsk, enum cpuacct_stat_index idx, cputime_t val) { struct cpuacct *ca; + int batch = CPUACCT_BATCH; if (unlikely(!cpuacct_subsys.active)) return; @@ -9012,7 +9030,7 @@ static void cpuacct_update_stats(struct task_struct *tsk, ca = task_ca(tsk); do { - percpu_counter_add(&ca->cpustat[idx], val); + __percpu_counter_add(&ca->cpustat[idx], val, batch); ca = ca->parent; } while (ca); rcu_read_unlock(); -- cgit v0.10.2 From f7e7ee36757f68778700cde1aaed89e1d23e59fd Mon Sep 17 00:00:00 2001 From: "austin_zhang@linux.intel.com" Date: Fri, 5 Feb 2010 09:02:42 -0800 Subject: perf record: Fix existing process callgraph symbol When 'perf record -g' a existing process, even with debuginfo packages, still cannnot get symbol from 'perf report'. try: perf record -g -p `pidof xxx` -f perf report 68.26% :1181 b74870f2 [.] 0x000000b74870f2 | |--32.09%-- 0xb73b5b44 | 0xb7487102 | 0xb748a4e2 | 0xb748633d | 0xb73b41cd | 0xb73b4467 | 0xb747d531 The reason is: for existing process, in __cmd_record(), the pid is 0 rather than the existing process id. Signed-off-by: Austin Zhang Acked-by: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <4710.10.255.24.35.1265389362.squirrel@linux.intel.com> Signed-off-by: Ingo Molnar diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 3ad599b..771533c 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -570,7 +570,7 @@ static int __cmd_record(int argc, const char **argv) } if (!system_wide && profile_cpu == -1) - event__synthesize_thread(pid, process_synthesized_event, + event__synthesize_thread(target_pid, process_synthesized_event, session); else event__synthesize_threads(process_synthesized_event, session); -- cgit v0.10.2 From 0f88009d5cfae890bd3466126d1622ad4f16b798 Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Mon, 18 Jan 2010 10:14:51 -0600 Subject: [SCSI] scsi_transport_sas: add support for transport layer retries (TLR) The mpt2sas driver wants to use transport layer retries (TLR) so the simplest thing to do seems to be to add the enabling flags and checks to the SAS transport class, since they're a SAS specific protocol feature. Signed-off-by: James Bottomley diff --git a/drivers/scsi/scsi_sas_internal.h b/drivers/scsi/scsi_sas_internal.h index 998cb5b..6266a5d 100644 --- a/drivers/scsi/scsi_sas_internal.h +++ b/drivers/scsi/scsi_sas_internal.h @@ -5,7 +5,7 @@ #define SAS_PHY_ATTRS 17 #define SAS_PORT_ATTRS 1 #define SAS_RPORT_ATTRS 7 -#define SAS_END_DEV_ATTRS 3 +#define SAS_END_DEV_ATTRS 5 #define SAS_EXPANDER_ATTRS 7 struct sas_internal { diff --git a/drivers/scsi/scsi_transport_sas.c b/drivers/scsi/scsi_transport_sas.c index f27e52d..927e99c 100644 --- a/drivers/scsi/scsi_transport_sas.c +++ b/drivers/scsi/scsi_transport_sas.c @@ -155,6 +155,17 @@ static struct { sas_bitfield_name_search(linkspeed, sas_linkspeed_names) sas_bitfield_name_set(linkspeed, sas_linkspeed_names) +static struct sas_end_device *sas_sdev_to_rdev(struct scsi_device *sdev) +{ + struct sas_rphy *rphy = target_to_rphy(sdev->sdev_target); + struct sas_end_device *rdev; + + BUG_ON(rphy->identify.device_type != SAS_END_DEVICE); + + rdev = rphy_to_end_device(rphy); + return rdev; +} + static void sas_smp_request(struct request_queue *q, struct Scsi_Host *shost, struct sas_rphy *rphy) { @@ -358,6 +369,85 @@ void sas_remove_host(struct Scsi_Host *shost) } EXPORT_SYMBOL(sas_remove_host); +/** + * sas_tlr_supported - checking TLR bit in vpd 0x90 + * @sdev: scsi device struct + * + * Check Transport Layer Retries are supported or not. + * If vpd page 0x90 is present, TRL is supported. + * + */ +unsigned int +sas_tlr_supported(struct scsi_device *sdev) +{ + const int vpd_len = 32; + struct sas_end_device *rdev = sas_sdev_to_rdev(sdev); + char *buffer = kzalloc(vpd_len, GFP_KERNEL); + int ret = 0; + + if (scsi_get_vpd_page(sdev, 0x90, buffer, vpd_len)) + goto out; + + /* + * Magic numbers: the VPD Protocol page (0x90) + * has a 4 byte header and then one entry per device port + * the TLR bit is at offset 8 on each port entry + * if we take the first port, that's at total offset 12 + */ + ret = buffer[12] & 0x01; + + out: + kfree(buffer); + rdev->tlr_supported = ret; + return ret; + +} +EXPORT_SYMBOL_GPL(sas_tlr_supported); + +/** + * sas_disable_tlr - setting TLR flags + * @sdev: scsi device struct + * + * Seting tlr_enabled flag to 0. + * + */ +void +sas_disable_tlr(struct scsi_device *sdev) +{ + struct sas_end_device *rdev = sas_sdev_to_rdev(sdev); + + rdev->tlr_enabled = 0; +} +EXPORT_SYMBOL_GPL(sas_disable_tlr); + +/** + * sas_enable_tlr - setting TLR flags + * @sdev: scsi device struct + * + * Seting tlr_enabled flag 1. + * + */ +void sas_enable_tlr(struct scsi_device *sdev) +{ + unsigned int tlr_supported = 0; + tlr_supported = sas_tlr_supported(sdev); + + if (tlr_supported) { + struct sas_end_device *rdev = sas_sdev_to_rdev(sdev); + + rdev->tlr_enabled = 1; + } + + return; +} +EXPORT_SYMBOL_GPL(sas_enable_tlr); + +unsigned int sas_is_tlr_enabled(struct scsi_device *sdev) +{ + struct sas_end_device *rdev = sas_sdev_to_rdev(sdev); + return rdev->tlr_enabled; +} +EXPORT_SYMBOL_GPL(sas_is_tlr_enabled); /* * SAS Phy attributes @@ -1146,15 +1236,10 @@ sas_rphy_simple_attr(identify.phy_identifier, phy_identifier, "%d\n", u8); int sas_read_port_mode_page(struct scsi_device *sdev) { char *buffer = kzalloc(BUF_SIZE, GFP_KERNEL), *msdata; - struct sas_rphy *rphy = target_to_rphy(sdev->sdev_target); - struct sas_end_device *rdev; + struct sas_end_device *rdev = sas_sdev_to_rdev(sdev); struct scsi_mode_data mode_data; int res, error; - BUG_ON(rphy->identify.device_type != SAS_END_DEVICE); - - rdev = rphy_to_end_device(rphy); - if (!buffer) return -ENOMEM; @@ -1207,6 +1292,10 @@ sas_end_dev_simple_attr(I_T_nexus_loss_timeout, I_T_nexus_loss_timeout, "%d\n", int); sas_end_dev_simple_attr(initiator_response_timeout, initiator_response_timeout, "%d\n", int); +sas_end_dev_simple_attr(tlr_supported, tlr_supported, + "%d\n", int); +sas_end_dev_simple_attr(tlr_enabled, tlr_enabled, + "%d\n", int); static DECLARE_TRANSPORT_CLASS(sas_expander_class, "sas_expander", NULL, NULL, NULL); @@ -1733,6 +1822,8 @@ sas_attach_transport(struct sas_function_template *ft) SETUP_END_DEV_ATTRIBUTE(end_dev_ready_led_meaning); SETUP_END_DEV_ATTRIBUTE(end_dev_I_T_nexus_loss_timeout); SETUP_END_DEV_ATTRIBUTE(end_dev_initiator_response_timeout); + SETUP_END_DEV_ATTRIBUTE(end_dev_tlr_supported); + SETUP_END_DEV_ATTRIBUTE(end_dev_tlr_enabled); i->end_dev_attrs[count] = NULL; count = 0; diff --git a/include/scsi/scsi_transport_sas.h b/include/scsi/scsi_transport_sas.h index 61ad359..ffeebc3 100644 --- a/include/scsi/scsi_transport_sas.h +++ b/include/scsi/scsi_transport_sas.h @@ -107,6 +107,8 @@ struct sas_end_device { struct sas_rphy rphy; /* flags */ unsigned ready_led_meaning:1; + unsigned tlr_supported:1; + unsigned tlr_enabled:1; /* parameters */ u16 I_T_nexus_loss_timeout; u16 initiator_response_timeout; @@ -181,6 +183,11 @@ extern int sas_phy_add(struct sas_phy *); extern void sas_phy_delete(struct sas_phy *); extern int scsi_is_sas_phy(const struct device *); +unsigned int sas_tlr_supported(struct scsi_device *); +unsigned int sas_is_tlr_enabled(struct scsi_device *); +void sas_disable_tlr(struct scsi_device *); +void sas_enable_tlr(struct scsi_device *); + extern struct sas_rphy *sas_end_device_alloc(struct sas_port *); extern struct sas_rphy *sas_expander_alloc(struct sas_port *, enum sas_device_type); void sas_rphy_free(struct sas_rphy *); -- cgit v0.10.2 From d7384b28afb2bf2b7be835ddc8c852bdc5e0ce1c Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 18:50:06 +0530 Subject: [SCSI] mpt2sas: Delete volume before HBA detach. The driver hangs when doing `rmmod mpt2sas` if there are any IR volumes present.The hang is due the scsi midlayer trying to access the IR volumes after the driver releases controller resources. Perhaps when scsi_remove_host is called,the scsi mid layer is sending some request. This doesn't occur for bare drives becuase the driver is already reporting those drives deleted prior to calling mpt2sas_base_detach. To solve this issue, we need to delete the volumes as well. Signed-off-by: Kashyap Desai Reviewed-by: Eric Moore Signed-off-by: James Bottomley diff --git a/drivers/scsi/mpt2sas/mpt2sas_scsih.c b/drivers/scsi/mpt2sas/mpt2sas_scsih.c index efabea1..cd55176 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_scsih.c +++ b/drivers/scsi/mpt2sas/mpt2sas_scsih.c @@ -5998,6 +5998,8 @@ _scsih_remove(struct pci_dev *pdev) struct _sas_port *mpt2sas_port; struct _sas_device *sas_device; struct _sas_node *expander_sibling; + struct _raid_device *raid_device, *next; + struct MPT2SAS_TARGET *sas_target_priv_data; struct workqueue_struct *wq; unsigned long flags; @@ -6011,6 +6013,21 @@ _scsih_remove(struct pci_dev *pdev) if (wq) destroy_workqueue(wq); + /* release all the volumes */ + list_for_each_entry_safe(raid_device, next, &ioc->raid_device_list, + list) { + if (raid_device->starget) { + sas_target_priv_data = + raid_device->starget->hostdata; + sas_target_priv_data->deleted = 1; + scsi_remove_target(&raid_device->starget->dev); + } + printk(MPT2SAS_INFO_FMT "removing handle(0x%04x), wwid" + "(0x%016llx)\n", ioc->name, raid_device->handle, + (unsigned long long) raid_device->wwid); + _scsih_raid_device_remove(ioc, raid_device); + } + /* free ports attached to the sas_host */ retry_again: list_for_each_entry(mpt2sas_port, -- cgit v0.10.2 From b2ff36ba19dec02b05d3e26e36923b71ab64f34a Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 18:51:05 +0530 Subject: [SCSI] mpt2sas: Return -ENODATA on IOCTL timeout The driver was modified to return -ENODATA when there is a timeout via ioctl path. Signed-off-by: Kashyap Desai Reviewed-by: Eric Moore Signed-off-by: James Bottomley diff --git a/drivers/scsi/mpt2sas/mpt2sas_ctl.c b/drivers/scsi/mpt2sas/mpt2sas_ctl.c index 84a124f..198ecdd 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_ctl.c +++ b/drivers/scsi/mpt2sas/mpt2sas_ctl.c @@ -891,6 +891,7 @@ _ctl_do_mpt_command(struct MPT2SAS_ADAPTER *ioc, issue_host_reset: if (issue_reset) { + ret = -ENODATA; if ((mpi_request->Function == MPI2_FUNCTION_SCSI_IO_REQUEST || mpi_request->Function == MPI2_FUNCTION_RAID_SCSI_IO_PASSTHROUGH)) { -- cgit v0.10.2 From 6846e75cdfa76c6609ac0381a6739c9f4af8fd3b Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 18:51:45 +0530 Subject: [SCSI] mpt2sas: User resource_size_t instead of unsigned long Use resource_size_t to define the type resource for the system interface register set. The existing implementation was using "unsigned long" which would be 32 bit in 32 bit OS. If 32 bit OS is using 64 bit physical address space for the system interface register set, we need to shift to using resource_size_t which takes care of physical address space. Signed-off-by: Kashyap Desai Reviewed-by: Eric Moore Signed-off-by: James Bottomley diff --git a/drivers/scsi/mpt2sas/mpt2sas_base.c b/drivers/scsi/mpt2sas/mpt2sas_base.c index 89d0240..bbd4f71 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_base.c +++ b/drivers/scsi/mpt2sas/mpt2sas_base.c @@ -1222,6 +1222,8 @@ mpt2sas_base_map_resources(struct MPT2SAS_ADAPTER *ioc) u32 memap_sz; u32 pio_sz; int i, r = 0; + u64 pio_chip = 0; + u64 chip_phys = 0; dinitprintk(ioc, printk(MPT2SAS_DEBUG_FMT "%s\n", ioc->name, __func__)); @@ -1255,12 +1257,13 @@ mpt2sas_base_map_resources(struct MPT2SAS_ADAPTER *ioc) if (pci_resource_flags(pdev, i) & PCI_BASE_ADDRESS_SPACE_IO) { if (pio_sz) continue; - ioc->pio_chip = pci_resource_start(pdev, i); + pio_chip = (u64)pci_resource_start(pdev, i); pio_sz = pci_resource_len(pdev, i); } else { if (memap_sz) continue; ioc->chip_phys = pci_resource_start(pdev, i); + chip_phys = (u64)ioc->chip_phys; memap_sz = pci_resource_len(pdev, i); ioc->chip = ioremap(ioc->chip_phys, memap_sz); if (ioc->chip == NULL) { @@ -1280,10 +1283,10 @@ mpt2sas_base_map_resources(struct MPT2SAS_ADAPTER *ioc) printk(MPT2SAS_INFO_FMT "%s: IRQ %d\n", ioc->name, ((ioc->msix_enable) ? "PCI-MSI-X enabled" : "IO-APIC enabled"), ioc->pci_irq); - printk(MPT2SAS_INFO_FMT "iomem(0x%lx), mapped(0x%p), size(%d)\n", - ioc->name, ioc->chip_phys, ioc->chip, memap_sz); - printk(MPT2SAS_INFO_FMT "ioport(0x%lx), size(%d)\n", - ioc->name, ioc->pio_chip, pio_sz); + printk(MPT2SAS_INFO_FMT "iomem(0x%016llx), mapped(0x%p), size(%d)\n", + ioc->name, (unsigned long long)chip_phys, ioc->chip, memap_sz); + printk(MPT2SAS_INFO_FMT "ioport(0x%016llx), size(%d)\n", + ioc->name, (unsigned long long)pio_chip, pio_sz); return 0; diff --git a/drivers/scsi/mpt2sas/mpt2sas_base.h b/drivers/scsi/mpt2sas/mpt2sas_base.h index bb4f146..12fa18b 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_base.h +++ b/drivers/scsi/mpt2sas/mpt2sas_base.h @@ -464,7 +464,6 @@ typedef void (*MPT_ADD_SGE)(void *paddr, u32 flags_length, dma_addr_t dma_addr); * @pdev: pci pdev object * @chip: memory mapped register space * @chip_phys: physical addrss prior to mapping - * @pio_chip: I/O mapped register space * @logging_level: see mpt2sas_debug.h * @fwfault_debug: debuging FW timeouts * @ir_firmware: IR firmware present @@ -587,8 +586,7 @@ struct MPT2SAS_ADAPTER { char tmp_string[MPT_STRING_LENGTH]; struct pci_dev *pdev; Mpi2SystemInterfaceRegs_t __iomem *chip; - unsigned long chip_phys; - unsigned long pio_chip; + resource_size_t chip_phys; int logging_level; int fwfault_debug; u8 ir_firmware; -- cgit v0.10.2 From e75b9b6d9d81e35aed5ef2692285c4948e173c5c Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 18:52:39 +0530 Subject: [SCSI] mpt2sas: Set ioc->fwfault_debug to the cmd line option mpt2sas_fwfault_debug. (1) change the formentioned string from logging_level to fwfault_debug (2) set ioc->fwfault_debug to the command line option mpt2sas_fwfault_debug setting at driver load time. Signed-off-by: Kashyap Desai Reviewed-by: Eric Moore Signed-off-by: James Bottomley diff --git a/drivers/scsi/mpt2sas/mpt2sas_base.c b/drivers/scsi/mpt2sas/mpt2sas_base.c index bbd4f71..88e6eeb 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_base.c +++ b/drivers/scsi/mpt2sas/mpt2sas_base.c @@ -107,8 +107,7 @@ _scsih_set_fwfault_debug(const char *val, struct kernel_param *kp) if (ret) return ret; - printk(KERN_INFO "setting logging_level(0x%08x)\n", - mpt2sas_fwfault_debug); + printk(KERN_INFO "setting fwfault_debug(%d)\n", mpt2sas_fwfault_debug); list_for_each_entry(ioc, &mpt2sas_ioc_list, list) ioc->fwfault_debug = mpt2sas_fwfault_debug; return 0; @@ -3576,6 +3575,8 @@ mpt2sas_base_attach(struct MPT2SAS_ADAPTER *ioc) init_waitqueue_head(&ioc->reset_wq); + ioc->fwfault_debug = mpt2sas_fwfault_debug; + /* base internal command bits */ mutex_init(&ioc->base_cmds.mutex); ioc->base_cmds.reply = kzalloc(ioc->reply_sz, GFP_KERNEL); -- cgit v0.10.2 From 22c88425e03e8476176375ab0457c88ff3a5b68c Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 18:53:04 +0530 Subject: [SCSI] mpt2sas: Use compat_ptr to setup the pointer compatibility. On ppc64, an 32bit application was failing due to data buffers not being copied properly from user to kernel memory. The problem due to improper conversion of 32 to 64 bit pointers. The fix is to use compat_ptr to setup the pointer compatibility in the routine _ctl_compat_mpt_command. Signed-off-by: Kashyap Desai Reviewed-by: Eric Moore Signed-off-by: James Bottomley diff --git a/drivers/scsi/mpt2sas/mpt2sas_ctl.c b/drivers/scsi/mpt2sas/mpt2sas_ctl.c index 198ecdd..fa9bf83 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_ctl.c +++ b/drivers/scsi/mpt2sas/mpt2sas_ctl.c @@ -2203,14 +2203,10 @@ _ctl_compat_mpt_command(struct file *file, unsigned cmd, unsigned long arg) karg.data_out_size = karg32.data_out_size; karg.max_sense_bytes = karg32.max_sense_bytes; karg.data_sge_offset = karg32.data_sge_offset; - memcpy(&karg.reply_frame_buf_ptr, &karg32.reply_frame_buf_ptr, - sizeof(uint32_t)); - memcpy(&karg.data_in_buf_ptr, &karg32.data_in_buf_ptr, - sizeof(uint32_t)); - memcpy(&karg.data_out_buf_ptr, &karg32.data_out_buf_ptr, - sizeof(uint32_t)); - memcpy(&karg.sense_data_ptr, &karg32.sense_data_ptr, - sizeof(uint32_t)); + karg.reply_frame_buf_ptr = compat_ptr(karg32.reply_frame_buf_ptr); + karg.data_in_buf_ptr = compat_ptr(karg32.data_in_buf_ptr); + karg.data_out_buf_ptr = compat_ptr(karg32.data_out_buf_ptr); + karg.sense_data_ptr = compat_ptr(karg32.sense_data_ptr); state = (file->f_flags & O_NONBLOCK) ? NON_BLOCKING : BLOCKING; return _ctl_do_mpt_command(ioc, karg, &uarg->mf, state); } -- cgit v0.10.2 From f7c95ef02b564d9984c0655c9659791b1dd5d7ad Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 18:54:42 +0530 Subject: [SCSI] mpt2sas: Added raid transport support Adding support for raid transport layer. This will provide sysfs attributes containing raid level, state, and resync rate. MPT2SAS module will select RAID_ATTRS. Signed-off-by: Kashyap Desai Reviewed-by: Eric Moore Signed-off-by: James Bottomley diff --git a/drivers/scsi/mpt2sas/Kconfig b/drivers/scsi/mpt2sas/Kconfig index 70c4c24..ba8e128 100644 --- a/drivers/scsi/mpt2sas/Kconfig +++ b/drivers/scsi/mpt2sas/Kconfig @@ -44,6 +44,7 @@ config SCSI_MPT2SAS tristate "LSI MPT Fusion SAS 2.0 Device Driver" depends on PCI && SCSI select SCSI_SAS_ATTRS + select RAID_ATTRS ---help--- This driver supports PCI-Express SAS 6Gb/s Host Adapters. diff --git a/drivers/scsi/mpt2sas/mpt2sas_base.h b/drivers/scsi/mpt2sas/mpt2sas_base.h index 12fa18b..014318f 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_base.h +++ b/drivers/scsi/mpt2sas/mpt2sas_base.h @@ -323,6 +323,7 @@ struct _sas_device { * @device_info: bitfield provides detailed info about the hidden components * @num_pds: number of hidden raid components * @responding: used in _scsih_raid_device_mark_responding + * @percent_complete: resync percent complete */ struct _raid_device { struct list_head list; @@ -336,6 +337,7 @@ struct _raid_device { u32 device_info; u8 num_pds; u8 responding; + u8 percent_complete; }; /** diff --git a/drivers/scsi/mpt2sas/mpt2sas_scsih.c b/drivers/scsi/mpt2sas/mpt2sas_scsih.c index cd55176..ca984cb 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_scsih.c +++ b/drivers/scsi/mpt2sas/mpt2sas_scsih.c @@ -52,6 +52,7 @@ #include #include #include +#include #include "mpt2sas_base.h" @@ -133,6 +134,9 @@ struct fw_event_work { void *event_data; }; +/* raid transport support */ +static struct raid_template *mpt2sas_raid_template; + /** * struct _scsi_io_transfer - scsi io transfer * @handle: sas device handle (assigned by firmware) @@ -1419,6 +1423,140 @@ _scsih_display_sata_capabilities(struct MPT2SAS_ADAPTER *ioc, } /** + * _scsih_is_raid - return boolean indicating device is raid volume + * @dev the device struct object + */ +static int +_scsih_is_raid(struct device *dev) +{ + struct scsi_device *sdev = to_scsi_device(dev); + + return (sdev->channel == RAID_CHANNEL) ? 1 : 0; +} + +/** + * _scsih_get_resync - get raid volume resync percent complete + * @dev the device struct object + */ +static void +_scsih_get_resync(struct device *dev) +{ + struct scsi_device *sdev = to_scsi_device(dev); + struct MPT2SAS_ADAPTER *ioc = shost_priv(sdev->host); + static struct _raid_device *raid_device; + unsigned long flags; + Mpi2RaidVolPage0_t vol_pg0; + Mpi2ConfigReply_t mpi_reply; + u32 volume_status_flags; + u8 percent_complete = 0; + + spin_lock_irqsave(&ioc->raid_device_lock, flags); + raid_device = _scsih_raid_device_find_by_id(ioc, sdev->id, + sdev->channel); + spin_unlock_irqrestore(&ioc->raid_device_lock, flags); + + if (!raid_device) + goto out; + + if (mpt2sas_config_get_raid_volume_pg0(ioc, &mpi_reply, &vol_pg0, + MPI2_RAID_VOLUME_PGAD_FORM_HANDLE, raid_device->handle, + sizeof(Mpi2RaidVolPage0_t))) { + printk(MPT2SAS_ERR_FMT "failure at %s:%d/%s()!\n", + ioc->name, __FILE__, __LINE__, __func__); + goto out; + } + + volume_status_flags = le32_to_cpu(vol_pg0.VolumeStatusFlags); + if (volume_status_flags & MPI2_RAIDVOL0_STATUS_FLAG_RESYNC_IN_PROGRESS) + percent_complete = raid_device->percent_complete; + out: + raid_set_resync(mpt2sas_raid_template, dev, percent_complete); +} + +/** + * _scsih_get_state - get raid volume level + * @dev the device struct object + */ +static void +_scsih_get_state(struct device *dev) +{ + struct scsi_device *sdev = to_scsi_device(dev); + struct MPT2SAS_ADAPTER *ioc = shost_priv(sdev->host); + static struct _raid_device *raid_device; + unsigned long flags; + Mpi2RaidVolPage0_t vol_pg0; + Mpi2ConfigReply_t mpi_reply; + u32 volstate; + enum raid_state state = RAID_STATE_UNKNOWN; + + spin_lock_irqsave(&ioc->raid_device_lock, flags); + raid_device = _scsih_raid_device_find_by_id(ioc, sdev->id, + sdev->channel); + spin_unlock_irqrestore(&ioc->raid_device_lock, flags); + + if (!raid_device) + goto out; + + if (mpt2sas_config_get_raid_volume_pg0(ioc, &mpi_reply, &vol_pg0, + MPI2_RAID_VOLUME_PGAD_FORM_HANDLE, raid_device->handle, + sizeof(Mpi2RaidVolPage0_t))) { + printk(MPT2SAS_ERR_FMT "failure at %s:%d/%s()!\n", + ioc->name, __FILE__, __LINE__, __func__); + goto out; + } + + volstate = le32_to_cpu(vol_pg0.VolumeStatusFlags); + if (volstate & MPI2_RAIDVOL0_STATUS_FLAG_RESYNC_IN_PROGRESS) { + state = RAID_STATE_RESYNCING; + goto out; + } + + switch (vol_pg0.VolumeState) { + case MPI2_RAID_VOL_STATE_OPTIMAL: + case MPI2_RAID_VOL_STATE_ONLINE: + state = RAID_STATE_ACTIVE; + break; + case MPI2_RAID_VOL_STATE_DEGRADED: + state = RAID_STATE_DEGRADED; + break; + case MPI2_RAID_VOL_STATE_FAILED: + case MPI2_RAID_VOL_STATE_MISSING: + state = RAID_STATE_OFFLINE; + break; + } + out: + raid_set_state(mpt2sas_raid_template, dev, state); +} + +/** + * _scsih_set_level - set raid level + * @sdev: scsi device struct + * @raid_device: raid_device object + */ +static void +_scsih_set_level(struct scsi_device *sdev, struct _raid_device *raid_device) +{ + enum raid_level level = RAID_LEVEL_UNKNOWN; + + switch (raid_device->volume_type) { + case MPI2_RAID_VOL_TYPE_RAID0: + level = RAID_LEVEL_0; + break; + case MPI2_RAID_VOL_TYPE_RAID10: + level = RAID_LEVEL_10; + break; + case MPI2_RAID_VOL_TYPE_RAID1E: + level = RAID_LEVEL_1E; + break; + case MPI2_RAID_VOL_TYPE_RAID1: + level = RAID_LEVEL_1; + break; + } + + raid_set_level(mpt2sas_raid_template, &sdev->sdev_gendev, level); +} + +/** * _scsih_get_volume_capabilities - volume capabilities * @ioc: per adapter object * @sas_device: the raid_device object @@ -1574,6 +1712,8 @@ _scsih_slave_configure(struct scsi_device *sdev) (unsigned long long)raid_device->wwid, raid_device->num_pds, ds); _scsih_change_queue_depth(sdev, qdepth, SCSI_QDEPTH_DEFAULT); + /* raid transport support */ + _scsih_set_level(sdev, raid_device); return 0; } @@ -5170,11 +5310,33 @@ static void _scsih_sas_ir_operation_status_event(struct MPT2SAS_ADAPTER *ioc, struct fw_event_work *fw_event) { + Mpi2EventDataIrOperationStatus_t *event_data = fw_event->event_data; + static struct _raid_device *raid_device; + unsigned long flags; + u16 handle; + #ifdef CONFIG_SCSI_MPT2SAS_LOGGING if (ioc->logging_level & MPT_DEBUG_EVENT_WORK_TASK) _scsih_sas_ir_operation_status_event_debug(ioc, - fw_event->event_data); + event_data); #endif + + /* code added for raid transport support */ + if (event_data->RAIDOperation == MPI2_EVENT_IR_RAIDOP_RESYNC) { + + handle = le16_to_cpu(event_data->VolDevHandle); + + spin_lock_irqsave(&ioc->raid_device_lock, flags); + raid_device = _scsih_raid_device_find_by_handle(ioc, handle); + spin_unlock_irqrestore(&ioc->raid_device_lock, flags); + + if (!raid_device) + return; + + if (event_data->RAIDOperation == MPI2_EVENT_IR_RAIDOP_RESYNC) + raid_device->percent_complete = + event_data->PercentComplete; + } } /** @@ -6390,6 +6552,13 @@ static struct pci_driver scsih_driver = { #endif }; +/* raid transport support */ +static struct raid_function_template mpt2sas_raid_functions = { + .cookie = &scsih_driver_template, + .is_raid = _scsih_is_raid, + .get_resync = _scsih_get_resync, + .get_state = _scsih_get_state, +}; /** * _scsih_init - main entry point for this driver. @@ -6409,6 +6578,12 @@ _scsih_init(void) sas_attach_transport(&mpt2sas_transport_functions); if (!mpt2sas_transport_template) return -ENODEV; + /* raid transport support */ + mpt2sas_raid_template = raid_class_attach(&mpt2sas_raid_functions); + if (!mpt2sas_raid_template) { + sas_release_transport(mpt2sas_transport_template); + return -ENODEV; + } mpt2sas_base_initialize_callback_handler(); @@ -6443,8 +6618,11 @@ _scsih_init(void) mpt2sas_ctl_init(); error = pci_register_driver(&scsih_driver); - if (error) + if (error) { + /* raid transport support */ + raid_class_release(mpt2sas_raid_template); sas_release_transport(mpt2sas_transport_template); + } return error; } @@ -6462,7 +6640,8 @@ _scsih_exit(void) pci_unregister_driver(&scsih_driver); - sas_release_transport(mpt2sas_transport_template); + mpt2sas_ctl_exit(); + mpt2sas_base_release_callback_handler(scsi_io_cb_idx); mpt2sas_base_release_callback_handler(tm_cb_idx); mpt2sas_base_release_callback_handler(base_cb_idx); @@ -6474,7 +6653,10 @@ _scsih_exit(void) mpt2sas_base_release_callback_handler(tm_tr_cb_idx); mpt2sas_base_release_callback_handler(tm_sas_control_cb_idx); - mpt2sas_ctl_exit(); + /* raid transport support */ + raid_class_release(mpt2sas_raid_template); + sas_release_transport(mpt2sas_transport_template); + } module_init(_scsih_init); diff --git a/include/linux/raid_class.h b/include/linux/raid_class.h index 6b537f1..31e1ff6 100644 --- a/include/linux/raid_class.h +++ b/include/linux/raid_class.h @@ -32,6 +32,7 @@ enum raid_level { RAID_LEVEL_0, RAID_LEVEL_1, RAID_LEVEL_10, + RAID_LEVEL_1E, RAID_LEVEL_3, RAID_LEVEL_4, RAID_LEVEL_5, -- cgit v0.10.2 From 50d5c60634673a79f8d88564e10345b50fca7378 Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 18:55:26 +0530 Subject: [SCSI] mpt2sas: Added phy_enable and set_phy_speed sysfs callback support. Added new callbacks phy_enable and set_phy_speed in the mpt2sas_transport_functions template. This will allow end user to enable/disable phys and change links rates using the SysFS interface. Current implementation only supports direct attached phys, but we could in the future add support for expander based phys. A new subroutine mpt2sas_config_set_sas_iounit_pg1 was added; this wrapper function used to send request to controller firmware to modify the phys and link rates. A new subroutine _transport_find_local_phy was added; a function for easly obtaining the local phy object for direct attached. Example to disable a phy echo 0 > /sys/class/phy3:0/enable Example to enable the same phy echo 1 > /sys/class/phy3:0/enable Example to change the link rate to 1.5 #echo "1.5 Gbit" > /sys/class/phy3:0/maximum_linkrate #cat /sys/class/phy3:0/negotiated_linkrate 1.5 Gbit Example to change the link rate to 3.0 #echo "3.0 Gbit" > /sys/class/phy3:0/maximum_linkrate #cat /sys/class/phy3:0/negotiated_linkrate 3.0 Gbit Example to change the link rate to 6.0 #echo "6.0 Gbit" > /sys/class/phy3:0/maximum_linkrate #cat /sys/class/phy3:0/negotiated_linkrate 6.0 Gbit Signed-off-by: Kashyap Desai Reviewed-by: Eric Moore Signed-off-by: James Bottomley diff --git a/drivers/scsi/mpt2sas/mpt2sas_base.h b/drivers/scsi/mpt2sas/mpt2sas_base.h index 014318f..93c067f 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_base.h +++ b/drivers/scsi/mpt2sas/mpt2sas_base.h @@ -853,6 +853,8 @@ int mpt2sas_config_set_iounit_pg1(struct MPT2SAS_ADAPTER *ioc, Mpi2ConfigReply_t *mpi_reply, Mpi2IOUnitPage1_t *config_page); int mpt2sas_config_get_sas_iounit_pg1(struct MPT2SAS_ADAPTER *ioc, Mpi2ConfigReply_t *mpi_reply, Mpi2SasIOUnitPage1_t *config_page, u16 sz); +int mpt2sas_config_set_sas_iounit_pg1(struct MPT2SAS_ADAPTER *ioc, + Mpi2ConfigReply_t *mpi_reply, Mpi2SasIOUnitPage1_t *config_page, u16 sz); int mpt2sas_config_get_ioc_pg8(struct MPT2SAS_ADAPTER *ioc, Mpi2ConfigReply_t *mpi_reply, Mpi2IOCPage8_t *config_page); int mpt2sas_config_get_expander_pg0(struct MPT2SAS_ADAPTER *ioc, Mpi2ConfigReply_t diff --git a/drivers/scsi/mpt2sas/mpt2sas_config.c b/drivers/scsi/mpt2sas/mpt2sas_config.c index 594a389..411c27d 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_config.c +++ b/drivers/scsi/mpt2sas/mpt2sas_config.c @@ -324,7 +324,9 @@ _config_request(struct MPT2SAS_ADAPTER *ioc, Mpi2ConfigRequest_t if (r != 0) goto out; if (mpi_request->Action == - MPI2_CONFIG_ACTION_PAGE_WRITE_CURRENT) { + MPI2_CONFIG_ACTION_PAGE_WRITE_CURRENT || + mpi_request->Action == + MPI2_CONFIG_ACTION_PAGE_WRITE_NVRAM) { ioc->base_add_sg_single(&mpi_request->PageBufferSGE, MPT2_CONFIG_COMMON_WRITE_SGLFLAGS | mem.sz, mem.page_dma); @@ -882,7 +884,7 @@ mpt2sas_config_get_sas_iounit_pg0(struct MPT2SAS_ADAPTER *ioc, Mpi2ConfigReply_t } /** - * mpt2sas_config_get_sas_iounit_pg1 - obtain sas iounit page 0 + * mpt2sas_config_get_sas_iounit_pg1 - obtain sas iounit page 1 * @ioc: per adapter object * @mpi_reply: reply mf payload returned from firmware * @config_page: contents of the config page @@ -907,7 +909,7 @@ mpt2sas_config_get_sas_iounit_pg1(struct MPT2SAS_ADAPTER *ioc, Mpi2ConfigReply_t mpi_request.Header.PageType = MPI2_CONFIG_PAGETYPE_EXTENDED; mpi_request.ExtPageType = MPI2_CONFIG_EXTPAGETYPE_SAS_IO_UNIT; mpi_request.Header.PageNumber = 1; - mpi_request.Header.PageVersion = MPI2_SASIOUNITPAGE0_PAGEVERSION; + mpi_request.Header.PageVersion = MPI2_SASIOUNITPAGE1_PAGEVERSION; mpt2sas_base_build_zero_len_sge(ioc, &mpi_request.PageBufferSGE); r = _config_request(ioc, &mpi_request, mpi_reply, MPT2_CONFIG_PAGE_DEFAULT_TIMEOUT, NULL, 0); @@ -922,6 +924,49 @@ mpt2sas_config_get_sas_iounit_pg1(struct MPT2SAS_ADAPTER *ioc, Mpi2ConfigReply_t } /** + * mpt2sas_config_set_sas_iounit_pg1 - send sas iounit page 1 + * @ioc: per adapter object + * @mpi_reply: reply mf payload returned from firmware + * @config_page: contents of the config page + * @sz: size of buffer passed in config_page + * Context: sleep. + * + * Calling function should call config_get_number_hba_phys prior to + * this function, so enough memory is allocated for config_page. + * + * Returns 0 for success, non-zero for failure. + */ +int +mpt2sas_config_set_sas_iounit_pg1(struct MPT2SAS_ADAPTER *ioc, Mpi2ConfigReply_t + *mpi_reply, Mpi2SasIOUnitPage1_t *config_page, u16 sz) +{ + Mpi2ConfigRequest_t mpi_request; + int r; + + memset(&mpi_request, 0, sizeof(Mpi2ConfigRequest_t)); + mpi_request.Function = MPI2_FUNCTION_CONFIG; + mpi_request.Action = MPI2_CONFIG_ACTION_PAGE_HEADER; + mpi_request.Header.PageType = MPI2_CONFIG_PAGETYPE_EXTENDED; + mpi_request.ExtPageType = MPI2_CONFIG_EXTPAGETYPE_SAS_IO_UNIT; + mpi_request.Header.PageNumber = 1; + mpi_request.Header.PageVersion = MPI2_SASIOUNITPAGE1_PAGEVERSION; + mpt2sas_base_build_zero_len_sge(ioc, &mpi_request.PageBufferSGE); + r = _config_request(ioc, &mpi_request, mpi_reply, + MPT2_CONFIG_PAGE_DEFAULT_TIMEOUT, NULL, 0); + if (r) + goto out; + + mpi_request.Action = MPI2_CONFIG_ACTION_PAGE_WRITE_CURRENT; + _config_request(ioc, &mpi_request, mpi_reply, + MPT2_CONFIG_PAGE_DEFAULT_TIMEOUT, config_page, sz); + mpi_request.Action = MPI2_CONFIG_ACTION_PAGE_WRITE_NVRAM; + r = _config_request(ioc, &mpi_request, mpi_reply, + MPT2_CONFIG_PAGE_DEFAULT_TIMEOUT, config_page, sz); + out: + return r; +} + +/** * mpt2sas_config_get_expander_pg0 - obtain expander page 0 * @ioc: per adapter object * @mpi_reply: reply mf payload returned from firmware diff --git a/drivers/scsi/mpt2sas/mpt2sas_transport.c b/drivers/scsi/mpt2sas/mpt2sas_transport.c index 3a82872..789f9ee 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_transport.c +++ b/drivers/scsi/mpt2sas/mpt2sas_transport.c @@ -855,6 +855,17 @@ rphy_to_ioc(struct sas_rphy *rphy) return shost_priv(shost); } +static struct _sas_phy * +_transport_find_local_phy(struct MPT2SAS_ADAPTER *ioc, struct sas_phy *phy) +{ + int i; + + for (i = 0; i < ioc->sas_hba.num_phys; i++) + if (ioc->sas_hba.phy[i].phy == phy) + return(&ioc->sas_hba.phy[i]); + return NULL; +} + /** * _transport_get_linkerrors - * @phy: The sas phy object @@ -870,14 +881,8 @@ _transport_get_linkerrors(struct sas_phy *phy) struct _sas_phy *mpt2sas_phy; Mpi2ConfigReply_t mpi_reply; Mpi2SasPhyPage1_t phy_pg1; - int i; - for (i = 0, mpt2sas_phy = NULL; i < ioc->sas_hba.num_phys && - !mpt2sas_phy; i++) { - if (ioc->sas_hba.phy[i].phy != phy) - continue; - mpt2sas_phy = &ioc->sas_hba.phy[i]; - } + mpt2sas_phy = _transport_find_local_phy(ioc, phy); if (!mpt2sas_phy) /* this phy not on sas_host */ return -EINVAL; @@ -971,14 +976,8 @@ _transport_phy_reset(struct sas_phy *phy, int hard_reset) struct _sas_phy *mpt2sas_phy; Mpi2SasIoUnitControlReply_t mpi_reply; Mpi2SasIoUnitControlRequest_t mpi_request; - int i; - for (i = 0, mpt2sas_phy = NULL; i < ioc->sas_hba.num_phys && - !mpt2sas_phy; i++) { - if (ioc->sas_hba.phy[i].phy != phy) - continue; - mpt2sas_phy = &ioc->sas_hba.phy[i]; - } + mpt2sas_phy = _transport_find_local_phy(ioc, phy); if (!mpt2sas_phy) /* this phy not on sas_host */ return -EINVAL; @@ -1006,6 +1005,173 @@ _transport_phy_reset(struct sas_phy *phy, int hard_reset) } /** + * _transport_phy_enable - enable/disable phys + * @phy: The sas phy object + * @enable: enable phy when true + * + * Only support sas_host direct attached phys. + * Returns 0 for success, non-zero for failure. + */ +static int +_transport_phy_enable(struct sas_phy *phy, int enable) +{ + struct MPT2SAS_ADAPTER *ioc = phy_to_ioc(phy); + struct _sas_phy *mpt2sas_phy; + Mpi2SasIOUnitPage1_t *sas_iounit_pg1 = NULL; + Mpi2ConfigReply_t mpi_reply; + u16 ioc_status; + u16 sz; + int rc = 0; + + mpt2sas_phy = _transport_find_local_phy(ioc, phy); + + if (!mpt2sas_phy) /* this phy not on sas_host */ + return -EINVAL; + + /* sas_iounit page 1 */ + sz = offsetof(Mpi2SasIOUnitPage1_t, PhyData) + (ioc->sas_hba.num_phys * + sizeof(Mpi2SasIOUnit1PhyData_t)); + sas_iounit_pg1 = kzalloc(sz, GFP_KERNEL); + if (!sas_iounit_pg1) { + printk(MPT2SAS_ERR_FMT "failure at %s:%d/%s()!\n", + ioc->name, __FILE__, __LINE__, __func__); + rc = -ENOMEM; + goto out; + } + if ((mpt2sas_config_get_sas_iounit_pg1(ioc, &mpi_reply, + sas_iounit_pg1, sz))) { + printk(MPT2SAS_ERR_FMT "failure at %s:%d/%s()!\n", + ioc->name, __FILE__, __LINE__, __func__); + rc = -ENXIO; + goto out; + } + ioc_status = le16_to_cpu(mpi_reply.IOCStatus) & + MPI2_IOCSTATUS_MASK; + if (ioc_status != MPI2_IOCSTATUS_SUCCESS) { + printk(MPT2SAS_ERR_FMT "failure at %s:%d/%s()!\n", + ioc->name, __FILE__, __LINE__, __func__); + rc = -EIO; + goto out; + } + + if (enable) + sas_iounit_pg1->PhyData[mpt2sas_phy->phy_id].PhyFlags + &= ~MPI2_SASIOUNIT1_PHYFLAGS_PHY_DISABLE; + else + sas_iounit_pg1->PhyData[mpt2sas_phy->phy_id].PhyFlags + |= MPI2_SASIOUNIT1_PHYFLAGS_PHY_DISABLE; + + mpt2sas_config_set_sas_iounit_pg1(ioc, &mpi_reply, sas_iounit_pg1, sz); + + out: + kfree(sas_iounit_pg1); + return rc; +} + +/** + * _transport_phy_speed - set phy min/max link rates + * @phy: The sas phy object + * @rates: rates defined in sas_phy_linkrates + * + * Only support sas_host direct attached phys. + * Returns 0 for success, non-zero for failure. + */ +static int +_transport_phy_speed(struct sas_phy *phy, struct sas_phy_linkrates *rates) +{ + struct MPT2SAS_ADAPTER *ioc = phy_to_ioc(phy); + struct _sas_phy *mpt2sas_phy; + Mpi2SasIOUnitPage1_t *sas_iounit_pg1 = NULL; + Mpi2SasPhyPage0_t phy_pg0; + Mpi2ConfigReply_t mpi_reply; + u16 ioc_status; + u16 sz; + int i; + int rc = 0; + + mpt2sas_phy = _transport_find_local_phy(ioc, phy); + + if (!mpt2sas_phy) /* this phy not on sas_host */ + return -EINVAL; + + if (!rates->minimum_linkrate) + rates->minimum_linkrate = phy->minimum_linkrate; + else if (rates->minimum_linkrate < phy->minimum_linkrate_hw) + rates->minimum_linkrate = phy->minimum_linkrate_hw; + + if (!rates->maximum_linkrate) + rates->maximum_linkrate = phy->maximum_linkrate; + else if (rates->maximum_linkrate > phy->maximum_linkrate_hw) + rates->maximum_linkrate = phy->maximum_linkrate_hw; + + /* sas_iounit page 1 */ + sz = offsetof(Mpi2SasIOUnitPage1_t, PhyData) + (ioc->sas_hba.num_phys * + sizeof(Mpi2SasIOUnit1PhyData_t)); + sas_iounit_pg1 = kzalloc(sz, GFP_KERNEL); + if (!sas_iounit_pg1) { + printk(MPT2SAS_ERR_FMT "failure at %s:%d/%s()!\n", + ioc->name, __FILE__, __LINE__, __func__); + rc = -ENOMEM; + goto out; + } + if ((mpt2sas_config_get_sas_iounit_pg1(ioc, &mpi_reply, + sas_iounit_pg1, sz))) { + printk(MPT2SAS_ERR_FMT "failure at %s:%d/%s()!\n", + ioc->name, __FILE__, __LINE__, __func__); + rc = -ENXIO; + goto out; + } + ioc_status = le16_to_cpu(mpi_reply.IOCStatus) & + MPI2_IOCSTATUS_MASK; + if (ioc_status != MPI2_IOCSTATUS_SUCCESS) { + printk(MPT2SAS_ERR_FMT "failure at %s:%d/%s()!\n", + ioc->name, __FILE__, __LINE__, __func__); + rc = -EIO; + goto out; + } + + for (i = 0; i < ioc->sas_hba.num_phys; i++) { + if (mpt2sas_phy->phy_id != i) { + sas_iounit_pg1->PhyData[i].MaxMinLinkRate = + (ioc->sas_hba.phy[i].phy->minimum_linkrate + + (ioc->sas_hba.phy[i].phy->maximum_linkrate << 4)); + } else { + sas_iounit_pg1->PhyData[i].MaxMinLinkRate = + (rates->minimum_linkrate + + (rates->maximum_linkrate << 4)); + } + } + + if (mpt2sas_config_set_sas_iounit_pg1(ioc, &mpi_reply, sas_iounit_pg1, + sz)) { + printk(MPT2SAS_ERR_FMT "failure at %s:%d/%s()!\n", + ioc->name, __FILE__, __LINE__, __func__); + rc = -ENXIO; + goto out; + } + + /* link reset */ + _transport_phy_reset(phy, 0); + + /* read phy page 0, then update the rates in the sas transport phy */ + if (!mpt2sas_config_get_phy_pg0(ioc, &mpi_reply, &phy_pg0, + mpt2sas_phy->phy_id)) { + phy->minimum_linkrate = _transport_convert_phy_link_rate( + phy_pg0.ProgrammedLinkRate & MPI2_SAS_PRATE_MIN_RATE_MASK); + phy->maximum_linkrate = _transport_convert_phy_link_rate( + phy_pg0.ProgrammedLinkRate >> 4); + phy->negotiated_linkrate = _transport_convert_phy_link_rate( + phy_pg0.NegotiatedLinkRate & + MPI2_SAS_NEG_LINK_RATE_MASK_PHYSICAL); + } + + out: + kfree(sas_iounit_pg1); + return rc; +} + + +/** * _transport_smp_handler - transport portal for smp passthru * @shost: shost object * @rphy: sas transport rphy object @@ -1207,6 +1373,8 @@ struct sas_function_template mpt2sas_transport_functions = { .get_enclosure_identifier = _transport_get_enclosure_identifier, .get_bay_identifier = _transport_get_bay_identifier, .phy_reset = _transport_phy_reset, + .phy_enable = _transport_phy_enable, + .set_phy_speed = _transport_phy_speed, .smp_handler = _transport_smp_handler, }; -- cgit v0.10.2 From f4af3c14113d1b0d98d5a5e717b8aa1f484065b6 Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 18:55:54 +0530 Subject: [SCSI] mpt2sas: Update driver header to latest MPI Spec. Update header to latest MPI SPEC revision. Signed-off-by: Kashyap Desai Reviewed-by: Eric Moore Signed-off-by: James Bottomley diff --git a/drivers/scsi/mpt2sas/mpi/mpi2.h b/drivers/scsi/mpt2sas/mpi/mpi2.h index 9141681..9958d84 100644 --- a/drivers/scsi/mpt2sas/mpi/mpi2.h +++ b/drivers/scsi/mpt2sas/mpi/mpi2.h @@ -8,7 +8,7 @@ * scatter/gather formats. * Creation Date: June 21, 2006 * - * mpi2.h Version: 02.00.13 + * mpi2.h Version: 02.00.14 * * Version History * --------------- @@ -53,6 +53,10 @@ * bytes reserved. * Added RAID Accelerator functionality. * 07-30-09 02.00.13 Bumped MPI2_HEADER_VERSION_UNIT. + * 10-28-09 02.00.14 Bumped MPI2_HEADER_VERSION_UNIT. + * Added MSI-x index mask and shift for Reply Post Host + * Index register. + * Added function code for Host Based Discovery Action. * -------------------------------------------------------------------------- */ @@ -78,7 +82,7 @@ #define MPI2_VERSION_02_00 (0x0200) /* versioning for this MPI header set */ -#define MPI2_HEADER_VERSION_UNIT (0x0D) +#define MPI2_HEADER_VERSION_UNIT (0x0E) #define MPI2_HEADER_VERSION_DEV (0x00) #define MPI2_HEADER_VERSION_UNIT_MASK (0xFF00) #define MPI2_HEADER_VERSION_UNIT_SHIFT (8) @@ -232,9 +236,12 @@ typedef volatile struct _MPI2_SYSTEM_INTERFACE_REGS #define MPI2_REPLY_FREE_HOST_INDEX_OFFSET (0x00000048) /* - * Offset for the Reply Descriptor Post Queue + * Defines for the Reply Descriptor Post Queue */ #define MPI2_REPLY_POST_HOST_INDEX_OFFSET (0x0000006C) +#define MPI2_REPLY_POST_HOST_INDEX_MASK (0x00FFFFFF) +#define MPI2_RPHI_MSIX_INDEX_MASK (0xFF000000) +#define MPI2_RPHI_MSIX_INDEX_SHIFT (24) /* * Defines for the HCBSize and address @@ -497,12 +504,13 @@ typedef union _MPI2_REPLY_DESCRIPTORS_UNION #define MPI2_FUNCTION_TARGET_CMD_BUF_BASE_POST (0x24) /* Target Command Buffer Post Base */ #define MPI2_FUNCTION_TARGET_CMD_BUF_LIST_POST (0x25) /* Target Command Buffer Post List */ #define MPI2_FUNCTION_RAID_ACCELERATOR (0x2C) /* RAID Accelerator*/ +/* Host Based Discovery Action */ +#define MPI2_FUNCTION_HOST_BASED_DISCOVERY_ACTION (0x2F) /* Doorbell functions */ #define MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET (0x40) -/* #define MPI2_FUNCTION_IO_UNIT_RESET (0x41) */ #define MPI2_FUNCTION_HANDSHAKE (0x42) diff --git a/drivers/scsi/mpt2sas/mpi/mpi2_cnfg.h b/drivers/scsi/mpt2sas/mpi/mpi2_cnfg.h index 1611c57..cf0ac9f 100644 --- a/drivers/scsi/mpt2sas/mpi/mpi2_cnfg.h +++ b/drivers/scsi/mpt2sas/mpi/mpi2_cnfg.h @@ -6,7 +6,7 @@ * Title: MPI Configuration messages and pages * Creation Date: November 10, 2006 * - * mpi2_cnfg.h Version: 02.00.12 + * mpi2_cnfg.h Version: 02.00.13 * * Version History * --------------- @@ -107,6 +107,8 @@ * to SAS Device Page 0 Flags field. * Added PhyInfo defines for power condition. * Added Ethernet configuration pages. + * 10-28-09 02.00.13 Added MPI2_IOUNITPAGE1_ENABLE_HOST_BASED_DISCOVERY. + * Added SAS PHY Page 4 structure and defines. * -------------------------------------------------------------------------- */ @@ -712,6 +714,7 @@ typedef struct _MPI2_CONFIG_PAGE_IO_UNIT_1 #define MPI2_IOUNITPAGE1_PAGEVERSION (0x04) /* IO Unit Page 1 Flags defines */ +#define MPI2_IOUNITPAGE1_ENABLE_HOST_BASED_DISCOVERY (0x00000800) #define MPI2_IOUNITPAGE1_MASK_SATA_WRITE_CACHE (0x00000600) #define MPI2_IOUNITPAGE1_ENABLE_SATA_WRITE_CACHE (0x00000000) #define MPI2_IOUNITPAGE1_DISABLE_SATA_WRITE_CACHE (0x00000200) @@ -2291,6 +2294,26 @@ typedef struct _MPI2_CONFIG_PAGE_SAS_PHY_3 { #define MPI2_SASPHY3_PAGEVERSION (0x00) +/* SAS PHY Page 4 */ + +typedef struct _MPI2_CONFIG_PAGE_SAS_PHY_4 { + MPI2_CONFIG_EXTENDED_PAGE_HEADER Header; /* 0x00 */ + U16 Reserved1; /* 0x08 */ + U8 Reserved2; /* 0x0A */ + U8 Flags; /* 0x0B */ + U8 InitialFrame[28]; /* 0x0C */ +} MPI2_CONFIG_PAGE_SAS_PHY_4, MPI2_POINTER PTR_MPI2_CONFIG_PAGE_SAS_PHY_4, + Mpi2SasPhyPage4_t, MPI2_POINTER pMpi2SasPhyPage4_t; + +#define MPI2_SASPHY4_PAGEVERSION (0x00) + +/* values for the Flags field */ +#define MPI2_SASPHY4_FLAGS_FRAME_VALID (0x02) +#define MPI2_SASPHY4_FLAGS_SATA_FRAME (0x01) + + + + /**************************************************************************** * SAS Port Config Pages ****************************************************************************/ diff --git a/drivers/scsi/mpt2sas/mpi/mpi2_history.txt b/drivers/scsi/mpt2sas/mpi/mpi2_history.txt index 65fcaa3..c4adf76 100644 --- a/drivers/scsi/mpt2sas/mpi/mpi2_history.txt +++ b/drivers/scsi/mpt2sas/mpi/mpi2_history.txt @@ -5,23 +5,24 @@ Copyright (c) 2000-2009 LSI Corporation. --------------------------------------- - Header Set Release Version: 02.00.12 - Header Set Release Date: 05-06-09 + Header Set Release Version: 02.00.14 + Header Set Release Date: 10-28-09 --------------------------------------- Filename Current version Prior version ---------- --------------- ------------- - mpi2.h 02.00.12 02.00.11 - mpi2_cnfg.h 02.00.11 02.00.10 - mpi2_init.h 02.00.07 02.00.06 - mpi2_ioc.h 02.00.11 02.00.10 - mpi2_raid.h 02.00.03 02.00.03 - mpi2_sas.h 02.00.02 02.00.02 + mpi2.h 02.00.14 02.00.13 + mpi2_cnfg.h 02.00.13 02.00.12 + mpi2_init.h 02.00.08 02.00.07 + mpi2_ioc.h 02.00.13 02.00.12 + mpi2_raid.h 02.00.04 02.00.04 + mpi2_sas.h 02.00.03 02.00.02 mpi2_targ.h 02.00.03 02.00.03 - mpi2_tool.h 02.00.03 02.00.02 + mpi2_tool.h 02.00.04 02.00.04 mpi2_type.h 02.00.00 02.00.00 - mpi2_ra.h 02.00.00 - mpi2_history.txt 02.00.11 02.00.12 + mpi2_ra.h 02.00.00 02.00.00 + mpi2_hbd.h 02.00.00 + mpi2_history.txt 02.00.14 02.00.13 * Date Version Description @@ -65,6 +66,11 @@ mpi2.h * MPI2_SCSI_IO_SUCCESS_REPLY_DESCRIPTOR and made those * bytes reserved. * Added RAID Accelerator functionality. + * 07-30-09 02.00.13 Bumped MPI2_HEADER_VERSION_UNIT. + * 10-28-09 02.00.14 Bumped MPI2_HEADER_VERSION_UNIT. + * Added MSI-x index mask and shift for Reply Post Host + * Index register. + * Added function code for Host Based Discovery Action. * -------------------------------------------------------------------------- mpi2_cnfg.h @@ -155,6 +161,15 @@ mpi2_cnfg.h * Added expander reduced functionality data to SAS * Expander Page 0. * Added SAS PHY Page 2 and SAS PHY Page 3. + * 07-30-09 02.00.12 Added IO Unit Page 7. + * Added new device ids. + * Added SAS IO Unit Page 5. + * Added partial and slumber power management capable flags + * to SAS Device Page 0 Flags field. + * Added PhyInfo defines for power condition. + * Added Ethernet configuration pages. + * 10-28-09 02.00.13 Added MPI2_IOUNITPAGE1_ENABLE_HOST_BASED_DISCOVERY. + * Added SAS PHY Page 4 structure and defines. * -------------------------------------------------------------------------- mpi2_init.h @@ -172,6 +187,10 @@ mpi2_init.h * Query Asynchronous Event. * Defined two new bits in the SlotStatus field of the SCSI * Enclosure Processor Request and Reply. + * 10-28-09 02.00.08 Added defines for decoding the ResponseInfo bytes for + * both SCSI IO Error Reply and SCSI Task Management Reply. + * Added ResponseInfo field to MPI2_SCSI_TASK_MANAGE_REPLY. + * Added MPI2_SCSITASKMGMT_RSP_TM_OVERLAPPED_TAG define. * -------------------------------------------------------------------------- mpi2_ioc.h @@ -246,6 +265,20 @@ mpi2_ioc.h * Added two new reason codes for SAS Device Status Change * Event. * Added new event: SAS PHY Counter. + * 07-30-09 02.00.12 Added GPIO Interrupt event define and structure. + * Added MPI2_IOCFACTS_CAPABILITY_EXTENDED_BUFFER define. + * Added new product id family for 2208. + * 10-28-09 02.00.13 Added HostMSIxVectors field to MPI2_IOC_INIT_REQUEST. + * Added MaxMSIxVectors field to MPI2_IOC_FACTS_REPLY. + * Added MinDevHandle field to MPI2_IOC_FACTS_REPLY. + * Added MPI2_IOCFACTS_CAPABILITY_HOST_BASED_DISCOVERY. + * Added MPI2_EVENT_HOST_BASED_DISCOVERY_PHY define. + * Added MPI2_EVENT_SAS_TOPO_ES_NO_EXPANDER define. + * Added Host Based Discovery Phy Event data. + * Added defines for ProductID Product field + * (MPI2_FW_HEADER_PID_). + * Modified values for SAS ProductID Family + * (MPI2_FW_HEADER_PID_FAMILY_). * -------------------------------------------------------------------------- mpi2_raid.h @@ -256,6 +289,8 @@ mpi2_raid.h * 05-21-08 02.00.03 Added MPI2_RAID_VOL_CREATION_NUM_PHYSDISKS so that * the PhysDisk array in MPI2_RAID_VOLUME_CREATION_STRUCT * can be sized by the build environment. + * 07-30-09 02.00.04 Added proper define for the Use Default Settings bit of + * VolumeCreationFlags and marked the old one as obsolete. * -------------------------------------------------------------------------- mpi2_sas.h @@ -264,6 +299,8 @@ mpi2_sas.h * Control Request. * 10-02-08 02.00.02 Added Set IOC Parameter Operation to SAS IO Unit Control * Request. + * 10-28-09 02.00.03 Changed the type of SGL in MPI2_SATA_PASSTHROUGH_REQUEST + * to MPI2_SGE_IO_UNION since it supports chained SGLs. * -------------------------------------------------------------------------- mpi2_targ.h @@ -283,6 +320,10 @@ mpi2_tool.h * structures and defines. * 02-29-08 02.00.02 Modified various names to make them 32-character unique. * 05-06-09 02.00.03 Added ISTWI Read Write Tool and Diagnostic CLI Tool. + * 07-30-09 02.00.04 Added ExtendedType field to DiagnosticBufferPost request + * and reply messages. + * Added MPI2_DIAG_BUF_TYPE_EXTENDED. + * Incremented MPI2_DIAG_BUF_TYPE_COUNT. * -------------------------------------------------------------------------- mpi2_type.h @@ -293,20 +334,26 @@ mpi2_ra.h * 05-06-09 02.00.00 Initial version. * -------------------------------------------------------------------------- +mpi2_hbd.h + * 10-28-09 02.00.00 Initial version. + * -------------------------------------------------------------------------- + + mpi2_history.txt Parts list history -Filename 02.00.12 ----------- -------- -mpi2.h 02.00.12 -mpi2_cnfg.h 02.00.11 -mpi2_init.h 02.00.07 -mpi2_ioc.h 02.00.11 -mpi2_raid.h 02.00.03 -mpi2_sas.h 02.00.02 -mpi2_targ.h 02.00.03 -mpi2_tool.h 02.00.03 -mpi2_type.h 02.00.00 -mpi2_ra.h 02.00.00 +Filename 02.00.14 02.00.13 02.00.12 +---------- -------- -------- -------- +mpi2.h 02.00.14 02.00.13 02.00.12 +mpi2_cnfg.h 02.00.13 02.00.12 02.00.11 +mpi2_init.h 02.00.08 02.00.07 02.00.07 +mpi2_ioc.h 02.00.13 02.00.12 02.00.11 +mpi2_raid.h 02.00.04 02.00.04 02.00.03 +mpi2_sas.h 02.00.03 02.00.02 02.00.02 +mpi2_targ.h 02.00.03 02.00.03 02.00.03 +mpi2_tool.h 02.00.04 02.00.04 02.00.03 +mpi2_type.h 02.00.00 02.00.00 02.00.00 +mpi2_ra.h 02.00.00 02.00.00 02.00.00 +mpi2_hbd.h 02.00.00 Filename 02.00.11 02.00.10 02.00.09 02.00.08 02.00.07 02.00.06 ---------- -------- -------- -------- -------- -------- -------- diff --git a/drivers/scsi/mpt2sas/mpi/mpi2_init.h b/drivers/scsi/mpt2sas/mpi/mpi2_init.h index 563e56d..6541945 100644 --- a/drivers/scsi/mpt2sas/mpi/mpi2_init.h +++ b/drivers/scsi/mpt2sas/mpi/mpi2_init.h @@ -6,7 +6,7 @@ * Title: MPI SCSI initiator mode messages and structures * Creation Date: June 23, 2006 * - * mpi2_init.h Version: 02.00.07 + * mpi2_init.h Version: 02.00.08 * * Version History * --------------- @@ -27,6 +27,10 @@ * Query Asynchronous Event. * Defined two new bits in the SlotStatus field of the SCSI * Enclosure Processor Request and Reply. + * 10-28-09 02.00.08 Added defines for decoding the ResponseInfo bytes for + * both SCSI IO Error Reply and SCSI Task Management Reply. + * Added ResponseInfo field to MPI2_SCSI_TASK_MANAGE_REPLY. + * Added MPI2_SCSITASKMGMT_RSP_TM_OVERLAPPED_TAG define. * -------------------------------------------------------------------------- */ @@ -254,6 +258,11 @@ typedef struct _MPI2_SCSI_IO_REPLY #define MPI2_SCSI_STATE_AUTOSENSE_FAILED (0x02) #define MPI2_SCSI_STATE_AUTOSENSE_VALID (0x01) +/* masks and shifts for the ResponseInfo field */ + +#define MPI2_SCSI_RI_MASK_REASONCODE (0x000000FF) +#define MPI2_SCSI_RI_SHIFT_REASONCODE (0) + #define MPI2_SCSI_TASKTAG_UNKNOWN (0xFFFF) @@ -327,6 +336,7 @@ typedef struct _MPI2_SCSI_TASK_MANAGE_REPLY U16 IOCStatus; /* 0x0E */ U32 IOCLogInfo; /* 0x10 */ U32 TerminationCount; /* 0x14 */ + U32 ResponseInfo; /* 0x18 */ } MPI2_SCSI_TASK_MANAGE_REPLY, MPI2_POINTER PTR_MPI2_SCSI_TASK_MANAGE_REPLY, Mpi2SCSITaskManagementReply_t, MPI2_POINTER pMpi2SCSIManagementReply_t; @@ -339,8 +349,20 @@ typedef struct _MPI2_SCSI_TASK_MANAGE_REPLY #define MPI2_SCSITASKMGMT_RSP_TM_FAILED (0x05) #define MPI2_SCSITASKMGMT_RSP_TM_SUCCEEDED (0x08) #define MPI2_SCSITASKMGMT_RSP_TM_INVALID_LUN (0x09) +#define MPI2_SCSITASKMGMT_RSP_TM_OVERLAPPED_TAG (0x0A) #define MPI2_SCSITASKMGMT_RSP_IO_QUEUED_ON_IOC (0x80) +/* masks and shifts for the ResponseInfo field */ + +#define MPI2_SCSITASKMGMT_RI_MASK_REASONCODE (0x000000FF) +#define MPI2_SCSITASKMGMT_RI_SHIFT_REASONCODE (0) +#define MPI2_SCSITASKMGMT_RI_MASK_ARI2 (0x0000FF00) +#define MPI2_SCSITASKMGMT_RI_SHIFT_ARI2 (8) +#define MPI2_SCSITASKMGMT_RI_MASK_ARI1 (0x00FF0000) +#define MPI2_SCSITASKMGMT_RI_SHIFT_ARI1 (16) +#define MPI2_SCSITASKMGMT_RI_MASK_ARI0 (0xFF000000) +#define MPI2_SCSITASKMGMT_RI_SHIFT_ARI0 (24) + /**************************************************************************** * SCSI Enclosure Processor messages diff --git a/drivers/scsi/mpt2sas/mpi/mpi2_ioc.h b/drivers/scsi/mpt2sas/mpi/mpi2_ioc.h index ea51ce8..7549384 100644 --- a/drivers/scsi/mpt2sas/mpi/mpi2_ioc.h +++ b/drivers/scsi/mpt2sas/mpi/mpi2_ioc.h @@ -6,7 +6,7 @@ * Title: MPI IOC, Port, Event, FW Download, and FW Upload messages * Creation Date: October 11, 2006 * - * mpi2_ioc.h Version: 02.00.12 + * mpi2_ioc.h Version: 02.00.13 * * Version History * --------------- @@ -87,6 +87,17 @@ * 07-30-09 02.00.12 Added GPIO Interrupt event define and structure. * Added MPI2_IOCFACTS_CAPABILITY_EXTENDED_BUFFER define. * Added new product id family for 2208. + * 10-28-09 02.00.13 Added HostMSIxVectors field to MPI2_IOC_INIT_REQUEST. + * Added MaxMSIxVectors field to MPI2_IOC_FACTS_REPLY. + * Added MinDevHandle field to MPI2_IOC_FACTS_REPLY. + * Added MPI2_IOCFACTS_CAPABILITY_HOST_BASED_DISCOVERY. + * Added MPI2_EVENT_HOST_BASED_DISCOVERY_PHY define. + * Added MPI2_EVENT_SAS_TOPO_ES_NO_EXPANDER define. + * Added Host Based Discovery Phy Event data. + * Added defines for ProductID Product field + * (MPI2_FW_HEADER_PID_). + * Modified values for SAS ProductID Family + * (MPI2_FW_HEADER_PID_FAMILY_). * -------------------------------------------------------------------------- */ @@ -119,8 +130,10 @@ typedef struct _MPI2_IOC_INIT_REQUEST U16 MsgVersion; /* 0x0C */ U16 HeaderVersion; /* 0x0E */ U32 Reserved5; /* 0x10 */ - U32 Reserved6; /* 0x14 */ - U16 Reserved7; /* 0x18 */ + U16 Reserved6; /* 0x14 */ + U8 Reserved7; /* 0x16 */ + U8 HostMSIxVectors; /* 0x17 */ + U16 Reserved8; /* 0x18 */ U16 SystemRequestFrameSize; /* 0x1A */ U16 ReplyDescriptorPostQueueDepth; /* 0x1C */ U16 ReplyFreeQueueDepth; /* 0x1E */ @@ -215,7 +228,7 @@ typedef struct _MPI2_IOC_FACTS_REPLY U8 MaxChainDepth; /* 0x14 */ U8 WhoInit; /* 0x15 */ U8 NumberOfPorts; /* 0x16 */ - U8 Reserved2; /* 0x17 */ + U8 MaxMSIxVectors; /* 0x17 */ U16 RequestCredit; /* 0x18 */ U16 ProductID; /* 0x1A */ U32 IOCCapabilities; /* 0x1C */ @@ -233,7 +246,8 @@ typedef struct _MPI2_IOC_FACTS_REPLY U8 MaxVolumes; /* 0x37 */ U16 MaxDevHandle; /* 0x38 */ U16 MaxPersistentEntries; /* 0x3A */ - U32 Reserved4; /* 0x3C */ + U16 MinDevHandle; /* 0x3C */ + U16 Reserved4; /* 0x3E */ } MPI2_IOC_FACTS_REPLY, MPI2_POINTER PTR_MPI2_IOC_FACTS_REPLY, Mpi2IOCFactsReply_t, MPI2_POINTER pMpi2IOCFactsReply_t; @@ -269,6 +283,7 @@ typedef struct _MPI2_IOC_FACTS_REPLY /* ProductID field uses MPI2_FW_HEADER_PID_ */ /* IOCCapabilities */ +#define MPI2_IOCFACTS_CAPABILITY_HOST_BASED_DISCOVERY (0x00010000) #define MPI2_IOCFACTS_CAPABILITY_MSI_X_INDEX (0x00008000) #define MPI2_IOCFACTS_CAPABILITY_RAID_ACCELERATOR (0x00004000) #define MPI2_IOCFACTS_CAPABILITY_EVENT_REPLAY (0x00002000) @@ -453,6 +468,7 @@ typedef struct _MPI2_EVENT_NOTIFICATION_REPLY #define MPI2_EVENT_LOG_ENTRY_ADDED (0x0021) #define MPI2_EVENT_SAS_PHY_COUNTER (0x0022) #define MPI2_EVENT_GPIO_INTERRUPT (0x0023) +#define MPI2_EVENT_HOST_BASED_DISCOVERY_PHY (0x0024) /* Log Entry Added Event data */ @@ -793,6 +809,7 @@ typedef struct _MPI2_EVENT_DATA_SAS_TOPOLOGY_CHANGE_LIST MPI2_POINTER pMpi2EventDataSasTopologyChangeList_t; /* values for the ExpStatus field */ +#define MPI2_EVENT_SAS_TOPO_ES_NO_EXPANDER (0x00) #define MPI2_EVENT_SAS_TOPO_ES_ADDED (0x01) #define MPI2_EVENT_SAS_TOPO_ES_NOT_RESPONDING (0x02) #define MPI2_EVENT_SAS_TOPO_ES_RESPONDING (0x03) @@ -878,6 +895,44 @@ typedef struct _MPI2_EVENT_DATA_SAS_PHY_COUNTER { * */ +/* Host Based Discovery Phy Event data */ + +typedef struct _MPI2_EVENT_HBD_PHY_SAS { + U8 Flags; /* 0x00 */ + U8 NegotiatedLinkRate; /* 0x01 */ + U8 PhyNum; /* 0x02 */ + U8 PhysicalPort; /* 0x03 */ + U32 Reserved1; /* 0x04 */ + U8 InitialFrame[28]; /* 0x08 */ +} MPI2_EVENT_HBD_PHY_SAS, MPI2_POINTER PTR_MPI2_EVENT_HBD_PHY_SAS, + Mpi2EventHbdPhySas_t, MPI2_POINTER pMpi2EventHbdPhySas_t; + +/* values for the Flags field */ +#define MPI2_EVENT_HBD_SAS_FLAGS_FRAME_VALID (0x02) +#define MPI2_EVENT_HBD_SAS_FLAGS_SATA_FRAME (0x01) + +/* use MPI2_SAS_NEG_LINK_RATE_ defines from mpi2_cnfg.h for + * the NegotiatedLinkRate field */ + +typedef union _MPI2_EVENT_HBD_DESCRIPTOR { + MPI2_EVENT_HBD_PHY_SAS Sas; +} MPI2_EVENT_HBD_DESCRIPTOR, MPI2_POINTER PTR_MPI2_EVENT_HBD_DESCRIPTOR, + Mpi2EventHbdDescriptor_t, MPI2_POINTER pMpi2EventHbdDescriptor_t; + +typedef struct _MPI2_EVENT_DATA_HBD_PHY { + U8 DescriptorType; /* 0x00 */ + U8 Reserved1; /* 0x01 */ + U16 Reserved2; /* 0x02 */ + U32 Reserved3; /* 0x04 */ + MPI2_EVENT_HBD_DESCRIPTOR Descriptor; /* 0x08 */ +} MPI2_EVENT_DATA_HBD_PHY, MPI2_POINTER PTR_MPI2_EVENT_DATA_HBD_PHY, + Mpi2EventDataHbdPhy_t, MPI2_POINTER pMpi2EventDataMpi2EventDataHbdPhy_t; + +/* values for the DescriptorType field */ +#define MPI2_EVENT_HBD_DT_SAS (0x01) + + + /**************************************************************************** * EventAck message ****************************************************************************/ @@ -1126,13 +1181,17 @@ typedef struct _MPI2_FW_IMAGE_HEADER #define MPI2_FW_HEADER_PID_TYPE_MASK (0xF000) #define MPI2_FW_HEADER_PID_TYPE_SAS (0x2000) -#define MPI2_FW_HEADER_PID_PROD_MASK (0x0F00) -#define MPI2_FW_HEADER_PID_PROD_A (0x0000) +#define MPI2_FW_HEADER_PID_PROD_MASK (0x0F00) +#define MPI2_FW_HEADER_PID_PROD_A (0x0000) +#define MPI2_FW_HEADER_PID_PROD_MASK (0x0F00) +#define MPI2_FW_HEADER_PID_PROD_TARGET_INITIATOR_SCSI (0x0200) +#define MPI2_FW_HEADER_PID_PROD_IR_SCSI (0x0700) + #define MPI2_FW_HEADER_PID_FAMILY_MASK (0x00FF) /* SAS */ -#define MPI2_FW_HEADER_PID_FAMILY_2108_SAS (0x0010) -#define MPI2_FW_HEADER_PID_FAMILY_2208_SAS (0x0011) +#define MPI2_FW_HEADER_PID_FAMILY_2108_SAS (0x0013) +#define MPI2_FW_HEADER_PID_FAMILY_2208_SAS (0x0014) /* use MPI2_IOCFACTS_PROTOCOL_ defines for ProtocolFlags field */ diff --git a/drivers/scsi/mpt2sas/mpi/mpi2_sas.h b/drivers/scsi/mpt2sas/mpi/mpi2_sas.h index 8a42b13..2d8aeed 100644 --- a/drivers/scsi/mpt2sas/mpi/mpi2_sas.h +++ b/drivers/scsi/mpt2sas/mpi/mpi2_sas.h @@ -6,7 +6,7 @@ * Title: MPI Serial Attached SCSI structures and definitions * Creation Date: February 9, 2007 * - * mpi2.h Version: 02.00.02 + * mpi2.h Version: 02.00.03 * * Version History * --------------- @@ -18,6 +18,8 @@ * Control Request. * 10-02-08 02.00.02 Added Set IOC Parameter Operation to SAS IO Unit Control * Request. + * 10-28-09 02.00.03 Changed the type of SGL in MPI2_SATA_PASSTHROUGH_REQUEST + * to MPI2_SGE_IO_UNION since it supports chained SGLs. * -------------------------------------------------------------------------- */ @@ -160,7 +162,7 @@ typedef struct _MPI2_SATA_PASSTHROUGH_REQUEST U32 Reserved4; /* 0x14 */ U32 DataLength; /* 0x18 */ U8 CommandFIS[20]; /* 0x1C */ - MPI2_SIMPLE_SGE_UNION SGL; /* 0x20 */ + MPI2_SGE_IO_UNION SGL; /* 0x20 */ } MPI2_SATA_PASSTHROUGH_REQUEST, MPI2_POINTER PTR_MPI2_SATA_PASSTHROUGH_REQUEST, Mpi2SataPassthroughRequest_t, MPI2_POINTER pMpi2SataPassthroughRequest_t; -- cgit v0.10.2 From 84f0b04a0e3b279a0b0a851b93eb403a626ca4b8 Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 18:56:28 +0530 Subject: [SCSI] mpt2sas: Enable TLR for SSP TAPE drives (Added SAS Transport APIs) If TLR is supported for end device, MPT2SAS driver will enable the TLR bit in the SCSI_IO for every request. If there is a response with MPI2_SCSITASKMGMT_RSP_INVALID_FRAME, the driver will turn off the TLR logic. [jejb: updated to new transport class TLR API] Signed-off-by: Kashyap Desai Reviewed-by: Eric Moore Signed-off-by: James Bottomley diff --git a/drivers/scsi/mpt2sas/mpt2sas_scsih.c b/drivers/scsi/mpt2sas/mpt2sas_scsih.c index ca984cb..344a22a 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_scsih.c +++ b/drivers/scsi/mpt2sas/mpt2sas_scsih.c @@ -1309,7 +1309,6 @@ _scsih_slave_alloc(struct scsi_device *sdev) struct MPT2SAS_DEVICE *sas_device_priv_data; struct scsi_target *starget; struct _raid_device *raid_device; - struct _sas_device *sas_device; unsigned long flags; sas_device_priv_data = kzalloc(sizeof(struct scsi_device), GFP_KERNEL); @@ -1336,21 +1335,8 @@ _scsih_slave_alloc(struct scsi_device *sdev) if (raid_device) raid_device->sdev = sdev; /* raid is single lun */ spin_unlock_irqrestore(&ioc->raid_device_lock, flags); - } else { - /* set TLR bit for SSP devices */ - if (!(ioc->facts.IOCCapabilities & - MPI2_IOCFACTS_CAPABILITY_TLR)) - goto out; - spin_lock_irqsave(&ioc->sas_device_lock, flags); - sas_device = mpt2sas_scsih_sas_device_find_by_sas_address(ioc, - sas_device_priv_data->sas_target->sas_address); - spin_unlock_irqrestore(&ioc->sas_device_lock, flags); - if (sas_device && sas_device->device_info & - MPI2_SAS_DEVICE_INFO_SSP_TARGET) - sas_device_priv_data->flags |= MPT_DEVICE_TLR_ON; } - out: return 0; } @@ -1617,6 +1603,32 @@ _scsih_get_volume_capabilities(struct MPT2SAS_ADAPTER *ioc, } /** + * _scsih_enable_tlr - setting TLR flags + * @ioc: per adapter object + * @sdev: scsi device struct + * + * Enabling Transaction Layer Retries for tape devices when + * vpd page 0x90 is present + * + */ +static void +_scsih_enable_tlr(struct MPT2SAS_ADAPTER *ioc, struct scsi_device *sdev) +{ + /* only for TAPE */ + if (sdev->type != TYPE_TAPE) + return; + + if (!(ioc->facts.IOCCapabilities & MPI2_IOCFACTS_CAPABILITY_TLR)) + return; + + sas_enable_tlr(sdev); + sdev_printk(KERN_INFO, sdev, "TLR %s\n", + sas_is_tlr_enabled(sdev) ? "Enabled" : "Disabled"); + return; + +} + +/** * _scsih_slave_configure - device configure routine. * @sdev: scsi device struct * @@ -1761,8 +1773,10 @@ _scsih_slave_configure(struct scsi_device *sdev) _scsih_change_queue_depth(sdev, qdepth, SCSI_QDEPTH_DEFAULT); - if (ssp_target) + if (ssp_target) { sas_read_port_mode_page(sdev); + _scsih_enable_tlr(ioc, sdev); + } return 0; } @@ -3049,7 +3063,7 @@ _scsih_qcmd(struct scsi_cmnd *scmd, void (*done)(struct scsi_cmnd *)) } else mpi_control |= MPI2_SCSIIO_CONTROL_SIMPLEQ; - if ((sas_device_priv_data->flags & MPT_DEVICE_TLR_ON)) + if (sas_is_tlr_enabled(scmd->device)) mpi_control |= MPI2_SCSIIO_CONTROL_TLR_ON; smid = mpt2sas_base_get_smid_scsiio(ioc, ioc->scsi_io_cb_idx, scmd); @@ -3438,10 +3452,11 @@ _scsih_io_done(struct MPT2SAS_ADAPTER *ioc, u16 smid, u8 msix_index, u32 reply) le32_to_cpu(mpi_reply->ResponseInfo) & 0xFF; if (!sas_device_priv_data->tlr_snoop_check) { sas_device_priv_data->tlr_snoop_check++; - if ((sas_device_priv_data->flags & MPT_DEVICE_TLR_ON) && - response_code == MPI2_SCSITASKMGMT_RSP_INVALID_FRAME) - sas_device_priv_data->flags &= - ~MPT_DEVICE_TLR_ON; + if (sas_is_tlr_enabled(scmd->device) && + response_code == MPI2_SCSITASKMGMT_RSP_INVALID_FRAME) { + sas_disable_tlr(scmd->device); + sdev_printk(KERN_INFO, scmd->device, "TLR disabled\n"); + } } xfer_cnt = le32_to_cpu(mpi_reply->TransferCount); -- cgit v0.10.2 From e4b8972c6172950ce226afdf7a921098f7672d5c Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Wed, 16 Dec 2009 18:57:02 +0530 Subject: [SCSI] mpt2sas: Bump version 04.100.01.00 Version upgraded to 04.100.01.00. Signed-off-by: Kashyap Desai Reviewed-by: Eric Moore Signed-off-by: James Bottomley diff --git a/drivers/scsi/mpt2sas/mpt2sas_base.h b/drivers/scsi/mpt2sas/mpt2sas_base.h index 93c067f..e18b054 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_base.h +++ b/drivers/scsi/mpt2sas/mpt2sas_base.h @@ -69,10 +69,10 @@ #define MPT2SAS_DRIVER_NAME "mpt2sas" #define MPT2SAS_AUTHOR "LSI Corporation " #define MPT2SAS_DESCRIPTION "LSI MPT Fusion SAS 2.0 Device Driver" -#define MPT2SAS_DRIVER_VERSION "03.100.03.00" -#define MPT2SAS_MAJOR_VERSION 03 +#define MPT2SAS_DRIVER_VERSION "04.100.01.00" +#define MPT2SAS_MAJOR_VERSION 04 #define MPT2SAS_MINOR_VERSION 100 -#define MPT2SAS_BUILD_VERSION 03 +#define MPT2SAS_BUILD_VERSION 01 #define MPT2SAS_RELEASE_VERSION 00 /* -- cgit v0.10.2 From 48bd86cf213a5780346b603b703c4b83978611a8 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Thu, 7 Jan 2010 01:50:19 +0530 Subject: [SCSI] be2iscsi: Fixing Bug for multiple SGEs The patch fixes a but where the sg_next is not assigned and hence the first sge was being resused wrongly Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 9fc4446..d611a13 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -1711,7 +1711,8 @@ hwi_write_sgl(struct iscsi_wrb *pwrb, struct scatterlist *sg, io_task->bhs_pa.u.a32.address_hi); l_sg = sg; - for (index = 0; (index < num_sg) && (index < 2); index++, sg_next(sg)) { + for (index = 0; (index < num_sg) && (index < 2); index++, + sg = sg_next(sg)) { if (index == 0) { sg_len = sg_dma_len(sg); addr = (u64) sg_dma_address(sg); @@ -1755,7 +1756,7 @@ hwi_write_sgl(struct iscsi_wrb *pwrb, struct scatterlist *sg, psgl++; psgl++; offset = 0; - for (index = 0; index < num_sg; index++, sg_next(sg), psgl++) { + for (index = 0; index < num_sg; index++, sg = sg_next(sg), psgl++) { sg_len = sg_dma_len(sg); addr = (u64) sg_dma_address(sg); AMAP_SET_BITS(struct amap_iscsi_sge, addr_lo, psgl, -- cgit v0.10.2 From aa3590329fdf789902e3c1383589bad12a74fb09 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Thu, 7 Jan 2010 01:51:04 +0530 Subject: [SCSI] be2iscsi: Fixing the number of SGE's The number of SGE's supported is fixed to what the chip expects. Also, the max sectors set to tested values Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.h b/drivers/scsi/be2iscsi/be_main.h index 4cde8f6..13ff02c 100644 --- a/drivers/scsi/be2iscsi/be_main.h +++ b/drivers/scsi/be2iscsi/be_main.h @@ -56,10 +56,10 @@ #define BE2_DEFPDU_DATA_SZ 8192 #define MAX_CPUS 31 -#define BEISCSI_SGLIST_ELEMENTS BE2_SGE +#define BEISCSI_SGLIST_ELEMENTS 30 #define BEISCSI_CMD_PER_LUN 128 /* scsi_host->cmd_per_lun */ -#define BEISCSI_MAX_SECTORS 2048 /* scsi_host->max_sectors */ +#define BEISCSI_MAX_SECTORS 256 /* scsi_host->max_sectors */ #define BEISCSI_MAX_CMD_LEN 16 /* scsi_host->max_cmd_len */ #define BEISCSI_NUM_MAX_LUN 256 /* scsi_host->max_lun */ -- cgit v0.10.2 From a3babda5b3b60b5d28dbab127b1ecc8ba707b6e7 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Thu, 7 Jan 2010 01:51:44 +0530 Subject: [SCSI] be2iscsi: Fix to allow driver to load when the FW allows more cids This fix allows the driver to load when the FW allows more cids than than the driver supports. The driver will limit the number of cid to what it can support. There was no reason to fail the driver load,so, correcting that Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_mgmt.c b/drivers/scsi/be2iscsi/be_mgmt.c index 18f8041..b0bb54d 100644 --- a/drivers/scsi/be2iscsi/be_mgmt.c +++ b/drivers/scsi/be2iscsi/be_mgmt.c @@ -49,12 +49,12 @@ unsigned char mgmt_get_fw_config(struct be_ctrl_info *ctrl, phba->fw_config.iscsi_cid_count = pfw_cfg->ulp[0].sq_count; if (phba->fw_config.iscsi_cid_count > (BE2_MAX_SESSIONS / 2)) { - status = 1; - shost_printk(KERN_WARNING, phba->shost, - "FW reported MAX CXNS as %d \t" - "Max Supported = %d. Failing to load \n", - phba->fw_config.iscsi_cid_count, - BE2_MAX_SESSIONS); + SE_DEBUG(DBG_LVL_8, + "FW reported MAX CXNS as %d \t" + "Max Supported = %d.\n", + phba->fw_config.iscsi_cid_count, + BE2_MAX_SESSIONS); + phba->fw_config.iscsi_cid_count = BE2_MAX_SESSIONS / 2; } } else { shost_printk(KERN_WARNING, phba->shost, -- cgit v0.10.2 From d44a6d2bbffd9e5c87f0a78cbe9c089e21289162 Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Sun, 17 Jan 2010 16:15:57 +0100 Subject: [SCSI] lpfc: unify two if branches with the same code in lpfc_decode_firmware_rev() Regardless of the flag state, the branches execute the same code Signed-off-by: Roel Kluin Acked-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/lpfc/lpfc_ct.c b/drivers/scsi/lpfc/lpfc_ct.c index 0ebcd9b..bf7bf62 100644 --- a/drivers/scsi/lpfc/lpfc_ct.c +++ b/drivers/scsi/lpfc/lpfc_ct.c @@ -1843,12 +1843,7 @@ lpfc_decode_firmware_rev(struct lpfc_hba *phba, char *fwrevision, int flag) c = (rev & 0x0000ff00) >> 8; b4 = (rev & 0x000000ff); - if (flag) - sprintf(fwrevision, "%d.%d%d%c%d ", b1, - b2, b3, c, b4); - else - sprintf(fwrevision, "%d.%d%d%c%d ", b1, - b2, b3, c, b4); + sprintf(fwrevision, "%d.%d%d%c%d", b1, b2, b3, c, b4); } return; } -- cgit v0.10.2 From fa4698fcf59c3bd01c171e5e558bae9e8eb396f1 Mon Sep 17 00:00:00 2001 From: "Martin K. Petersen" Date: Mon, 18 Jan 2010 18:43:18 -0500 Subject: [SCSI] sd: Combine DIF/DIX error handling DIF and DIX errors are handled identically at this point. Collapse the switch cases into one and let scsi_io_completion print result and sense data. Signed-off-by: Martin K. Petersen Signed-off-by: James Bottomley diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c index c5e9a99..908d400 100644 --- a/drivers/scsi/sd.c +++ b/drivers/scsi/sd.c @@ -1196,19 +1196,10 @@ static int sd_done(struct scsi_cmnd *SCpnt) SCpnt->result = 0; memset(SCpnt->sense_buffer, 0, SCSI_SENSE_BUFFERSIZE); break; - case ABORTED_COMMAND: - if (sshdr.asc == 0x10) { /* DIF: Disk detected corruption */ - scsi_print_result(SCpnt); - scsi_print_sense("sd", SCpnt); + case ABORTED_COMMAND: /* DIF: Target detected corruption */ + case ILLEGAL_REQUEST: /* DIX: Host detected corruption */ + if (sshdr.asc == 0x10) good_bytes = sd_completed_bytes(SCpnt); - } - break; - case ILLEGAL_REQUEST: - if (sshdr.asc == 0x10) { /* DIX: HBA detected corruption */ - scsi_print_result(SCpnt); - scsi_print_sense("sd", SCpnt); - good_bytes = sd_completed_bytes(SCpnt); - } break; default: break; -- cgit v0.10.2 From d0e2ddff7c4b493acff50a9000564b67cbe7d676 Mon Sep 17 00:00:00 2001 From: Dmitry Torokhov Date: Tue, 19 Jan 2010 10:24:40 -0800 Subject: [SCSI] vmw_pvscsi: fix signedness warning with newer compilers pvscsi_setup_msix() expects 'irq' argument to be an int but is being passed unsigned int. Unsigned int is more proper type for IRQ number so let's use it. This shuts off a compile warning with recent compilers. Signed-off-by: Dmitry Torokhov Acked-by: Alok Kataria Signed-off-by: James Bottomley diff --git a/drivers/scsi/vmw_pvscsi.c b/drivers/scsi/vmw_pvscsi.c index d2604c8..e4ac582 100644 --- a/drivers/scsi/vmw_pvscsi.c +++ b/drivers/scsi/vmw_pvscsi.c @@ -1069,7 +1069,8 @@ static void pvscsi_free_sgls(const struct pvscsi_adapter *adapter) free_pages((unsigned long)ctx->sgl, get_order(SGL_SIZE)); } -static int pvscsi_setup_msix(const struct pvscsi_adapter *adapter, int *irq) +static int pvscsi_setup_msix(const struct pvscsi_adapter *adapter, + unsigned int *irq) { struct msix_entry entry = { 0, PVSCSI_VECTOR_COMPLETION }; int ret; -- cgit v0.10.2 From 32951dd8cd70e525eb523c9e2b62d79d9157fae9 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Sat, 23 Jan 2010 05:34:24 +0530 Subject: [SCSI] be2iscsi: Remove Ring mode from driver Ring mode is not used. This patch removes the code. Signed-off-by: Mike Christie Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index d611a13..522e7d5 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -40,7 +40,6 @@ static unsigned int be_iopoll_budget = 10; static unsigned int be_max_phys_size = 64; static unsigned int enable_msix = 1; -static unsigned int ring_mode; MODULE_DEVICE_TABLE(pci, beiscsi_pci_id_table); MODULE_DESCRIPTION(DRV_DESC " " BUILD_STR); @@ -671,9 +670,7 @@ static void free_wrb_handle(struct beiscsi_hba *phba, struct hwi_wrb_context *pwrb_context, struct wrb_handle *pwrb_handle) { - if (!ring_mode) - pwrb_context->pwrb_handle_base[pwrb_context->free_index] = - pwrb_handle; + pwrb_context->pwrb_handle_base[pwrb_context->free_index] = pwrb_handle; pwrb_context->wrb_handles_available++; if (pwrb_context->free_index == (phba->params.wrbs_per_cxn - 1)) pwrb_context->free_index = 0; @@ -865,7 +862,6 @@ hwi_complete_drvr_msgs(struct beiscsi_conn *beiscsi_conn, { struct hwi_wrb_context *pwrb_context; struct wrb_handle *pwrb_handle = NULL; - struct sgl_handle *psgl_handle = NULL; struct hwi_controller *phwi_ctrlr; struct iscsi_task *task; struct beiscsi_io_task *io_task; @@ -873,23 +869,14 @@ hwi_complete_drvr_msgs(struct beiscsi_conn *beiscsi_conn, struct iscsi_session *session = conn->session; phwi_ctrlr = phba->phwi_ctrlr; - if (ring_mode) { - psgl_handle = phba->sgl_hndl_array[((psol-> - dw[offsetof(struct amap_sol_cqe_ring, icd_index) / - 32] & SOL_ICD_INDEX_MASK) >> 6)]; - pwrb_context = &phwi_ctrlr->wrb_context[psgl_handle->cid]; - task = psgl_handle->task; - pwrb_handle = NULL; - } else { - pwrb_context = &phwi_ctrlr->wrb_context[((psol-> + pwrb_context = &phwi_ctrlr->wrb_context[((psol-> dw[offsetof(struct amap_sol_cqe, cid) / 32] & SOL_CID_MASK) >> 6) - phba->fw_config.iscsi_cid_start]; - pwrb_handle = pwrb_context->pwrb_handle_basestd[((psol-> + pwrb_handle = pwrb_context->pwrb_handle_basestd[((psol-> dw[offsetof(struct amap_sol_cqe, wrb_index) / 32] & SOL_WRB_INDEX_MASK) >> 16)]; - task = pwrb_handle->pio_handle; - } + task = pwrb_handle->pio_handle; io_task = task->dd_data; spin_lock(&phba->mgmt_sgl_lock); @@ -930,32 +917,23 @@ static void hwi_complete_cmd(struct beiscsi_conn *beiscsi_conn, struct iscsi_wrb *pwrb = NULL; struct hwi_controller *phwi_ctrlr; struct iscsi_task *task; - struct sgl_handle *psgl_handle = NULL; unsigned int type; struct iscsi_conn *conn = beiscsi_conn->conn; struct iscsi_session *session = conn->session; phwi_ctrlr = phba->phwi_ctrlr; - if (ring_mode) { - psgl_handle = phba->sgl_hndl_array[((psol-> - dw[offsetof(struct amap_sol_cqe_ring, icd_index) / - 32] & SOL_ICD_INDEX_MASK) >> 6)]; - task = psgl_handle->task; - type = psgl_handle->type; - } else { - pwrb_context = &phwi_ctrlr-> - wrb_context[((psol->dw[offsetof + pwrb_context = &phwi_ctrlr->wrb_context[((psol->dw[offsetof (struct amap_sol_cqe, cid) / 32] & SOL_CID_MASK) >> 6) - phba->fw_config.iscsi_cid_start]; - pwrb_handle = pwrb_context->pwrb_handle_basestd[((psol-> + pwrb_handle = pwrb_context->pwrb_handle_basestd[((psol-> dw[offsetof(struct amap_sol_cqe, wrb_index) / 32] & SOL_WRB_INDEX_MASK) >> 16)]; - task = pwrb_handle->pio_handle; - pwrb = pwrb_handle->pwrb; - type = (pwrb->dw[offsetof(struct amap_iscsi_wrb, type) / 32] & - WRB_TYPE_MASK) >> 28; - } + task = pwrb_handle->pio_handle; + pwrb = pwrb_handle->pwrb; + type = (pwrb->dw[offsetof(struct amap_iscsi_wrb, type) / 32] & + WRB_TYPE_MASK) >> 28; + spin_lock_bh(&session->lock); switch (type) { case HWH_TYPE_IO: @@ -986,15 +964,7 @@ static void hwi_complete_cmd(struct beiscsi_conn *beiscsi_conn, break; default: - if (ring_mode) - shost_printk(KERN_WARNING, phba->shost, - "In hwi_complete_cmd, unknown type = %d" - "icd_index 0x%x CID 0x%x\n", type, - ((psol->dw[offsetof(struct amap_sol_cqe_ring, - icd_index) / 32] & SOL_ICD_INDEX_MASK) >> 6), - psgl_handle->cid); - else - shost_printk(KERN_WARNING, phba->shost, + shost_printk(KERN_WARNING, phba->shost, "In hwi_complete_cmd, unknown type = %d" "wrb_index 0x%x CID 0x%x\n", type, ((psol->dw[offsetof(struct amap_iscsi_wrb, @@ -1483,7 +1453,6 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) unsigned int num_processed = 0; unsigned int tot_nump = 0; struct beiscsi_conn *beiscsi_conn; - struct sgl_handle *psgl_handle = NULL; struct beiscsi_endpoint *beiscsi_ep; struct iscsi_endpoint *ep; struct beiscsi_hba *phba; @@ -1496,18 +1465,11 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) CQE_VALID_MASK) { be_dws_le_to_cpu(sol, sizeof(struct sol_cqe)); - if (ring_mode) { - psgl_handle = phba->sgl_hndl_array[((sol-> - dw[offsetof(struct amap_sol_cqe_ring, - icd_index) / 32] & SOL_ICD_INDEX_MASK) - >> 6)]; - ep = phba->ep_array[psgl_handle->cid]; - } else { - ep = phba->ep_array[(u32) ((sol-> + ep = phba->ep_array[(u32) ((sol-> dw[offsetof(struct amap_sol_cqe, cid) / 32] & SOL_CID_MASK) >> 6) - phba->fw_config.iscsi_cid_start]; - } + beiscsi_ep = ep->dd_data; beiscsi_conn = beiscsi_ep->conn; @@ -1553,21 +1515,13 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) case CMD_CXN_KILLED_ITT_INVALID: case CMD_CXN_KILLED_SEQ_OUTOFORDER: case CMD_CXN_KILLED_INVALID_DATASN_RCVD: - if (ring_mode) { - SE_DEBUG(DBG_LVL_1, - "CQ Error notification for cmd.. " - "code %d cid 0x%x\n", - sol->dw[offsetof(struct amap_sol_cqe, code) / - 32] & CQE_CODE_MASK, psgl_handle->cid); - } else { - SE_DEBUG(DBG_LVL_1, + SE_DEBUG(DBG_LVL_1, "CQ Error notification for cmd.. " "code %d cid 0x%x\n", sol->dw[offsetof(struct amap_sol_cqe, code) / 32] & CQE_CODE_MASK, (sol->dw[offsetof(struct amap_sol_cqe, cid) / 32] & SOL_CID_MASK)); - } break; case UNSOL_DATA_DIGEST_ERROR_NOTIFY: SE_DEBUG(DBG_LVL_1, @@ -1589,37 +1543,23 @@ static unsigned int beiscsi_process_cq(struct be_eq_obj *pbe_eq) case CXN_KILLED_OVER_RUN_RESIDUAL: case CXN_KILLED_UNDER_RUN_RESIDUAL: case CXN_KILLED_CMND_DATA_NOT_ON_SAME_CONN: - if (ring_mode) { - SE_DEBUG(DBG_LVL_1, "CQ Error %d, reset CID " - "0x%x...\n", - sol->dw[offsetof(struct amap_sol_cqe, code) / - 32] & CQE_CODE_MASK, psgl_handle->cid); - } else { - SE_DEBUG(DBG_LVL_1, "CQ Error %d, reset CID " + SE_DEBUG(DBG_LVL_1, "CQ Error %d, reset CID " "0x%x...\n", sol->dw[offsetof(struct amap_sol_cqe, code) / 32] & CQE_CODE_MASK, (sol->dw[offsetof(struct amap_sol_cqe, cid) / 32] & CQE_CID_MASK)); - } iscsi_conn_failure(beiscsi_conn->conn, ISCSI_ERR_CONN_FAILED); break; case CXN_KILLED_RST_SENT: case CXN_KILLED_RST_RCVD: - if (ring_mode) { - SE_DEBUG(DBG_LVL_1, "CQ Error %d, reset" - "received/sent on CID 0x%x...\n", - sol->dw[offsetof(struct amap_sol_cqe, code) / - 32] & CQE_CODE_MASK, psgl_handle->cid); - } else { - SE_DEBUG(DBG_LVL_1, "CQ Error %d, reset" + SE_DEBUG(DBG_LVL_1, "CQ Error %d, reset" "received/sent on CID 0x%x...\n", sol->dw[offsetof(struct amap_sol_cqe, code) / 32] & CQE_CODE_MASK, (sol->dw[offsetof(struct amap_sol_cqe, cid) / 32] & CQE_CID_MASK)); - } iscsi_conn_failure(beiscsi_conn->conn, ISCSI_ERR_CONN_FAILED); break; @@ -2848,11 +2788,6 @@ static int hwi_init_port(struct beiscsi_hba *phba) goto error; } - if (phba->fw_config.iscsi_features == 0x1) - ring_mode = 1; - else - ring_mode = 0; - status = beiscsi_create_cqs(phba, phwi_context); if (status != 0) { shost_printk(KERN_ERR, phba->shost, "CQ not created\n"); @@ -2980,17 +2915,6 @@ static int beiscsi_init_sgl_handle(struct beiscsi_hba *phba) phba->io_sgl_hndl_avbl = 0; phba->eh_sgl_hndl_avbl = 0; - if (ring_mode) { - phba->sgl_hndl_array = kzalloc(sizeof(struct sgl_handle *) * - phba->params.icds_per_ctrl, - GFP_KERNEL); - if (!phba->sgl_hndl_array) { - shost_printk(KERN_ERR, phba->shost, - "Mem Alloc Failed. Failing to load\n"); - return -ENOMEM; - } - } - mem_descr_sglh = phba->init_mem; mem_descr_sglh += HWI_MEM_SGLH; if (1 == mem_descr_sglh->num_elements) { @@ -2998,8 +2922,6 @@ static int beiscsi_init_sgl_handle(struct beiscsi_hba *phba) phba->params.ios_per_ctrl, GFP_KERNEL); if (!phba->io_sgl_hndl_base) { - if (ring_mode) - kfree(phba->sgl_hndl_array); shost_printk(KERN_ERR, phba->shost, "Mem Alloc Failed. Failing to load\n"); return -ENOMEM; @@ -3184,8 +3106,6 @@ static int beiscsi_init_port(struct beiscsi_hba *phba) if (hba_setup_cid_tbls(phba)) { shost_printk(KERN_ERR, phba->shost, "Failed in hba_setup_cid_tbls\n"); - if (ring_mode) - kfree(phba->sgl_hndl_array); kfree(phba->io_sgl_hndl_base); kfree(phba->eh_sgl_hndl_base); goto do_cleanup_ctrlr; @@ -3242,8 +3162,6 @@ static void beiscsi_clean_port(struct beiscsi_hba *phba) hwi_purge_eq(phba); hwi_cleanup(phba); - if (ring_mode) - kfree(phba->sgl_hndl_array); kfree(phba->io_sgl_hndl_base); kfree(phba->eh_sgl_hndl_base); kfree(phba->cid_array); @@ -3329,8 +3247,7 @@ beiscsi_offload_connection(struct beiscsi_conn *beiscsi_conn, be_dws_le_to_cpu(pwrb, sizeof(struct iscsi_target_context_update_wrb)); doorbell |= beiscsi_conn->beiscsi_conn_cid & DB_WRB_POST_CID_MASK; - if (!ring_mode) - doorbell |= (pwrb_handle->wrb_index & DB_DEF_PDU_WRB_INDEX_MASK) + doorbell |= (pwrb_handle->wrb_index & DB_DEF_PDU_WRB_INDEX_MASK) << DB_DEF_PDU_WRB_INDEX_SHIFT; doorbell |= 1 << DB_DEF_PDU_NUM_POSTED_SHIFT; @@ -3417,15 +3334,7 @@ static int beiscsi_alloc_pdu(struct iscsi_task *task, uint8_t opcode) itt = (itt_t) cpu_to_be32(((unsigned int)io_task->pwrb_handle-> wrb_index << 16) | (unsigned int) (io_task->psgl_handle->sgl_index)); - if (ring_mode) { - phba->sgl_hndl_array[io_task->psgl_handle->sgl_index - - phba->fw_config.iscsi_icd_start] = - io_task->psgl_handle; - io_task->psgl_handle->task = task; - io_task->psgl_handle->cid = beiscsi_conn->beiscsi_conn_cid - - phba->fw_config.iscsi_cid_start; - } else - io_task->pwrb_handle->pio_handle = task; + io_task->pwrb_handle->pio_handle = task; io_task->cmd_bhs->iscsi_hdr.itt = itt; return 0; @@ -3511,18 +3420,12 @@ static int beiscsi_iotask(struct iscsi_task *task, struct scatterlist *sg, ISCSI_OPCODE_SCSI_DATA_OUT); AMAP_SET_BITS(struct amap_pdu_data_out, final_bit, &io_task->cmd_bhs->iscsi_data_pdu, 1); - if (ring_mode) - io_task->psgl_handle->type = INI_WR_CMD; - else - AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, - INI_WR_CMD); + AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, + INI_WR_CMD); AMAP_SET_BITS(struct amap_iscsi_wrb, dsp, pwrb, 1); } else { - if (ring_mode) - io_task->psgl_handle->type = INI_RD_CMD; - else - AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, - INI_RD_CMD); + AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, + INI_RD_CMD); AMAP_SET_BITS(struct amap_iscsi_wrb, dsp, pwrb, 0); } memcpy(&io_task->cmd_bhs->iscsi_data_pdu. @@ -3547,8 +3450,7 @@ static int beiscsi_iotask(struct iscsi_task *task, struct scatterlist *sg, be_dws_le_to_cpu(pwrb, sizeof(struct iscsi_wrb)); doorbell |= beiscsi_conn->beiscsi_conn_cid & DB_WRB_POST_CID_MASK; - if (!ring_mode) - doorbell |= (io_task->pwrb_handle->wrb_index & + doorbell |= (io_task->pwrb_handle->wrb_index & DB_DEF_PDU_WRB_INDEX_MASK) << DB_DEF_PDU_WRB_INDEX_SHIFT; doorbell |= 1 << DB_DEF_PDU_NUM_POSTED_SHIFT; @@ -3583,21 +3485,15 @@ static int beiscsi_mtask(struct iscsi_task *task) switch (task->hdr->opcode & ISCSI_OPCODE_MASK) { case ISCSI_OP_LOGIN: - if (ring_mode) - io_task->psgl_handle->type = TGT_DM_CMD; - else - AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, - TGT_DM_CMD); + AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, + TGT_DM_CMD); AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 0); AMAP_SET_BITS(struct amap_iscsi_wrb, cmdsn_itt, pwrb, 1); hwi_write_buffer(pwrb, task); break; case ISCSI_OP_NOOP_OUT: - if (ring_mode) - io_task->psgl_handle->type = INI_RD_CMD; - else - AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, - INI_RD_CMD); + AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, + INI_RD_CMD); if (task->hdr->ttt == ISCSI_RESERVED_TAG) AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 0); else @@ -3605,11 +3501,8 @@ static int beiscsi_mtask(struct iscsi_task *task) hwi_write_buffer(pwrb, task); break; case ISCSI_OP_TEXT: - if (ring_mode) - io_task->psgl_handle->type = INI_WR_CMD; - else - AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, - INI_WR_CMD); + AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, + INI_WR_CMD); AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 0); AMAP_SET_BITS(struct amap_iscsi_wrb, dsp, pwrb, 1); hwi_write_buffer(pwrb, task); @@ -3642,19 +3535,13 @@ static int beiscsi_mtask(struct iscsi_task *task) phba->ctrl.mcc_numtag[tag]); free_mcc_tag(&phba->ctrl, tag); } - if (ring_mode) - io_task->psgl_handle->type = INI_TMF_CMD; - else - AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, - INI_TMF_CMD); + AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, + INI_TMF_CMD); AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 0); hwi_write_buffer(pwrb, task); break; case ISCSI_OP_LOGOUT: AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 0); - if (ring_mode) - io_task->psgl_handle->type = HWH_TYPE_LOGOUT; - else AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, HWH_TYPE_LOGOUT); hwi_write_buffer(pwrb, task); @@ -3673,8 +3560,7 @@ static int beiscsi_mtask(struct iscsi_task *task) be_dws_le_to_cpu(pwrb, sizeof(struct iscsi_wrb)); doorbell |= cid & DB_WRB_POST_CID_MASK; - if (!ring_mode) - doorbell |= (io_task->pwrb_handle->wrb_index & + doorbell |= (io_task->pwrb_handle->wrb_index & DB_DEF_PDU_WRB_INDEX_MASK) << DB_DEF_PDU_WRB_INDEX_SHIFT; doorbell |= 1 << DB_DEF_PDU_NUM_POSTED_SHIFT; iowrite32(doorbell, phba->db_va + DB_TXULP0_OFFSET); @@ -3999,7 +3885,6 @@ static int __init beiscsi_module_init(void) "beiscsi pci driver.\n"); goto unregister_iscsi_transport; } - ring_mode = 0; return 0; unregister_iscsi_transport: -- cgit v0.10.2 From d2eeb1ac35a7146cbd4b1f6bf09bb6384ccdfb0d Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Sat, 23 Jan 2010 05:35:15 +0530 Subject: [SCSI] be2iscsi: changing copyright to 2010 This patch replaces 2009 with 2010 in copyright statement Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be.h b/drivers/scsi/be2iscsi/be.h index 3861cf4..0063740 100644 --- a/drivers/scsi/be2iscsi/be.h +++ b/drivers/scsi/be2iscsi/be.h @@ -1,5 +1,5 @@ /** - * Copyright (C) 2005 - 2009 ServerEngines + * Copyright (C) 2005 - 2010 ServerEngines * All rights reserved. * * This program is free software; you can redistribute it and/or diff --git a/drivers/scsi/be2iscsi/be_cmds.c b/drivers/scsi/be2iscsi/be_cmds.c index 44e69bd..6709857 100644 --- a/drivers/scsi/be2iscsi/be_cmds.c +++ b/drivers/scsi/be2iscsi/be_cmds.c @@ -1,5 +1,5 @@ /** - * Copyright (C) 2005 - 2009 ServerEngines + * Copyright (C) 2005 - 2010 ServerEngines * All rights reserved. * * This program is free software; you can redistribute it and/or diff --git a/drivers/scsi/be2iscsi/be_cmds.h b/drivers/scsi/be2iscsi/be_cmds.h index 69dddfa..49fcc78 100644 --- a/drivers/scsi/be2iscsi/be_cmds.h +++ b/drivers/scsi/be2iscsi/be_cmds.h @@ -1,5 +1,5 @@ /** - * Copyright (C) 2005 - 2009 ServerEngines + * Copyright (C) 2005 - 2010 ServerEngines * All rights reserved. * * This program is free software; you can redistribute it and/or diff --git a/drivers/scsi/be2iscsi/be_iscsi.c b/drivers/scsi/be2iscsi/be_iscsi.c index 95694d3..3847a58 100644 --- a/drivers/scsi/be2iscsi/be_iscsi.c +++ b/drivers/scsi/be2iscsi/be_iscsi.c @@ -1,5 +1,5 @@ /** - * Copyright (C) 2005 - 2009 ServerEngines + * Copyright (C) 2005 - 2010 ServerEngines * All rights reserved. * * This program is free software; you can redistribute it and/or diff --git a/drivers/scsi/be2iscsi/be_iscsi.h b/drivers/scsi/be2iscsi/be_iscsi.h index f92ffc5..1f512c2 100644 --- a/drivers/scsi/be2iscsi/be_iscsi.h +++ b/drivers/scsi/be2iscsi/be_iscsi.h @@ -1,5 +1,5 @@ /** - * Copyright (C) 2005 - 2009 ServerEngines + * Copyright (C) 2005 - 2010 ServerEngines * All rights reserved. * * This program is free software; you can redistribute it and/or diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 522e7d5..2f0027c 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -1,5 +1,5 @@ /** - * Copyright (C) 2005 - 2009 ServerEngines + * Copyright (C) 2005 - 2010 ServerEngines * All rights reserved. * * This program is free software; you can redistribute it and/or diff --git a/drivers/scsi/be2iscsi/be_main.h b/drivers/scsi/be2iscsi/be_main.h index 13ff02c..8cbe908 100644 --- a/drivers/scsi/be2iscsi/be_main.h +++ b/drivers/scsi/be2iscsi/be_main.h @@ -1,5 +1,5 @@ /** - * Copyright (C) 2005 - 2009 ServerEngines + * Copyright (C) 2005 - 2010 ServerEngines * All rights reserved. * * This program is free software; you can redistribute it and/or diff --git a/drivers/scsi/be2iscsi/be_mgmt.c b/drivers/scsi/be2iscsi/be_mgmt.c index b0bb54d..39106da 100644 --- a/drivers/scsi/be2iscsi/be_mgmt.c +++ b/drivers/scsi/be2iscsi/be_mgmt.c @@ -1,5 +1,5 @@ /** - * Copyright (C) 2005 - 2009 ServerEngines + * Copyright (C) 2005 - 2010 ServerEngines * All rights reserved. * * This program is free software; you can redistribute it and/or diff --git a/drivers/scsi/be2iscsi/be_mgmt.h b/drivers/scsi/be2iscsi/be_mgmt.h index f873a66..ecead6a 100644 --- a/drivers/scsi/be2iscsi/be_mgmt.h +++ b/drivers/scsi/be2iscsi/be_mgmt.h @@ -1,5 +1,5 @@ /** - * Copyright (C) 2005 - 2009 ServerEngines + * Copyright (C) 2005 - 2010 ServerEngines * All rights reserved. * * This program is free software; you can redistribute it and/or -- cgit v0.10.2 From 230dceb4316da9c8e05c82b64f2527aee95da2ff Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Sat, 23 Jan 2010 05:36:10 +0530 Subject: [SCSI] be2iscsi: Fix for first_burst This patch fixes the first_burst being modified instead of max_burst Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_iscsi.c b/drivers/scsi/be2iscsi/be_iscsi.c index 3847a58..d5712bc 100644 --- a/drivers/scsi/be2iscsi/be_iscsi.c +++ b/drivers/scsi/be2iscsi/be_iscsi.c @@ -274,8 +274,8 @@ int beiscsi_set_param(struct iscsi_cls_conn *cls_conn, conn->max_recv_dlength = 65536; break; case ISCSI_PARAM_MAX_BURST: - if (session->first_burst > 262144) - session->first_burst = 262144; + if (session->max_burst > 262144) + session->max_burst = 262144; break; default: return 0; -- cgit v0.10.2 From 5dc1c416b39531e3bcf0651ca4123e64228dd553 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Sat, 23 Jan 2010 05:36:52 +0530 Subject: [SCSI] be2iscsi: Proper checking of state This patch adds proper checking of value in for hba state. We would be adding more states later on Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_iscsi.c b/drivers/scsi/be2iscsi/be_iscsi.c index d5712bc..29a3aaf 100644 --- a/drivers/scsi/be2iscsi/be_iscsi.c +++ b/drivers/scsi/be2iscsi/be_iscsi.c @@ -561,7 +561,7 @@ beiscsi_ep_connect(struct Scsi_Host *shost, struct sockaddr *dst_addr, return ERR_PTR(ret); } - if (phba->state) { + if (phba->state != BE_ADAPTER_UP) { ret = -EBUSY; SE_DEBUG(DBG_LVL_1, "The Adapter state is Not UP \n"); return ERR_PTR(ret); -- cgit v0.10.2 From f55a24f2c24b63f8b4329eb28b4aec02a4bcf5d0 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Sat, 23 Jan 2010 05:37:40 +0530 Subject: [SCSI] be2iscsi: correcting the return This patch fixes an issue where return was not called properly. Thanks to Mike Christie for spotting this Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 2f0027c..6b700ad 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -3873,7 +3873,7 @@ static int __init beiscsi_module_init(void) SE_DEBUG(DBG_LVL_1, "beiscsi_module_init - Unable to register beiscsi" "transport.\n"); - ret = -ENOMEM; + return -ENOMEM; } SE_DEBUG(DBG_LVL_8, "In beiscsi_module_init, tt=%p \n", &beiscsi_iscsi_transport); -- cgit v0.10.2 From caf818f118b1ea60226a884072f1c3a9d5714766 Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Sat, 23 Jan 2010 05:38:18 +0530 Subject: [SCSI] be2iscsi: Ensure clean reuse of wrb This patch ensures that wrb is cleanly resued for io path and is memset to zero for non io path Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 6b700ad..85c0e80 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -1663,11 +1663,7 @@ hwi_write_sgl(struct iscsi_wrb *pwrb, struct scatterlist *sg, AMAP_SET_BITS(struct amap_iscsi_wrb, sge0_len, pwrb, sg_len); sge_len = sg_len; - AMAP_SET_BITS(struct amap_iscsi_wrb, sge0_last, pwrb, - 1); } else { - AMAP_SET_BITS(struct amap_iscsi_wrb, sge0_last, pwrb, - 0); AMAP_SET_BITS(struct amap_iscsi_wrb, sge1_r2t_offset, pwrb, sge_len); sg_len = sg_dma_len(sg); @@ -1690,8 +1686,22 @@ hwi_write_sgl(struct iscsi_wrb *pwrb, struct scatterlist *sg, AMAP_SET_BITS(struct amap_iscsi_sge, addr_lo, psgl, io_task->bhs_pa.u.a32.address_lo); - if (num_sg == 2) - AMAP_SET_BITS(struct amap_iscsi_wrb, sge1_last, pwrb, 1); + if (num_sg == 1) { + AMAP_SET_BITS(struct amap_iscsi_wrb, sge0_last, pwrb, + 1); + AMAP_SET_BITS(struct amap_iscsi_wrb, sge1_last, pwrb, + 0); + } else if (num_sg == 2) { + AMAP_SET_BITS(struct amap_iscsi_wrb, sge0_last, pwrb, + 0); + AMAP_SET_BITS(struct amap_iscsi_wrb, sge1_last, pwrb, + 1); + } else { + AMAP_SET_BITS(struct amap_iscsi_wrb, sge0_last, pwrb, + 0); + AMAP_SET_BITS(struct amap_iscsi_wrb, sge1_last, pwrb, + 0); + } sg = l_sg; psgl++; psgl++; @@ -3476,6 +3486,7 @@ static int beiscsi_mtask(struct iscsi_task *task) cid = beiscsi_conn->beiscsi_conn_cid; pwrb = io_task->pwrb_handle->pwrb; + memset(pwrb, 0, sizeof(*pwrb)); AMAP_SET_BITS(struct amap_iscsi_wrb, cmdsn_itt, pwrb, be32_to_cpu(task->cmdsn)); AMAP_SET_BITS(struct amap_iscsi_wrb, wrb_idx, pwrb, -- cgit v0.10.2 From b30c6dab1d87cc638a7413dd0e778d8c9a5232eb Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Sat, 23 Jan 2010 05:38:56 +0530 Subject: [SCSI] be2iscsi: changing the chip opcode for TEXT This patch corrects the chipopcode for text and chooses correct paramters for that command Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 85c0e80..932cb8a 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -3513,9 +3513,8 @@ static int beiscsi_mtask(struct iscsi_task *task) break; case ISCSI_OP_TEXT: AMAP_SET_BITS(struct amap_iscsi_wrb, type, pwrb, - INI_WR_CMD); + TGT_DM_CMD); AMAP_SET_BITS(struct amap_iscsi_wrb, dmsg, pwrb, 0); - AMAP_SET_BITS(struct amap_iscsi_wrb, dsp, pwrb, 1); hwi_write_buffer(pwrb, task); break; case ISCSI_OP_SCSI_TMFUNC: -- cgit v0.10.2 From 7c56533cf4fb072abc39e850e7ef4fb6166cc83b Mon Sep 17 00:00:00 2001 From: Jayamohan Kallickal Date: Sat, 23 Jan 2010 05:39:37 +0530 Subject: [SCSI] be2iscsi: correction in the claculation for num_cxn_wrb This patch correct the math done for num_cxn_wrb Signed-off-by: Jayamohan Kallickal Signed-off-by: James Bottomley diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 932cb8a..687d535 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -2042,10 +2042,9 @@ static void beiscsi_init_wrb_handle(struct beiscsi_hba *phba) } idx = 0; pwrb = mem_descr_wrb->mem_array[idx].virtual_address; - num_cxn_wrb = - ((mem_descr_wrb->mem_array[idx].size) / (sizeof(struct iscsi_wrb)) * - phba->params.wrbs_per_cxn); - + num_cxn_wrb = (mem_descr_wrb->mem_array[idx].size) / + ((sizeof(struct iscsi_wrb) * + phba->params.wrbs_per_cxn)); for (index = 0; index < phba->params.cxns_per_ctrl; index += 2) { pwrb_context = &phwi_ctrlr->wrb_context[index]; if (num_cxn_wrb) { @@ -2058,9 +2057,9 @@ static void beiscsi_init_wrb_handle(struct beiscsi_hba *phba) } else { idx++; pwrb = mem_descr_wrb->mem_array[idx].virtual_address; - num_cxn_wrb = ((mem_descr_wrb->mem_array[idx].size) / - (sizeof(struct iscsi_wrb)) * - phba->params.wrbs_per_cxn); + num_cxn_wrb = (mem_descr_wrb->mem_array[idx].size) / + ((sizeof(struct iscsi_wrb) * + phba->params.wrbs_per_cxn)); for (j = 0; j < phba->params.wrbs_per_cxn; j++) { pwrb_handle = pwrb_context->pwrb_handle_base[j]; pwrb_handle->pwrb = pwrb; -- cgit v0.10.2 From 2cec802980727f1daa46d8c31b411e083d49d7a2 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Sat, 23 Jan 2010 18:40:29 +0000 Subject: [SCSI] qla1280: Drop host_lock while requesting firmware request_firmware() may sleep and it appears to be safe to release the spinlock here. Signed-off-by: Ben Hutchings Cc: stable@kernel.org Signed-off-by: James Bottomley diff --git a/drivers/scsi/qla1280.c b/drivers/scsi/qla1280.c index 8371d91..49ac414 100644 --- a/drivers/scsi/qla1280.c +++ b/drivers/scsi/qla1280.c @@ -1640,8 +1640,10 @@ qla1280_load_firmware_pio(struct scsi_qla_host *ha) uint16_t mb[MAILBOX_REGISTER_COUNT], i; int err; + spin_unlock_irq(ha->host->host_lock); err = request_firmware(&fw, ql1280_board_tbl[ha->devnum].fwname, &ha->pdev->dev); + spin_lock_irq(ha->host->host_lock); if (err) { printk(KERN_ERR "Failed to load image \"%s\" err %d\n", ql1280_board_tbl[ha->devnum].fwname, err); @@ -1699,8 +1701,10 @@ qla1280_load_firmware_dma(struct scsi_qla_host *ha) return -ENOMEM; #endif + spin_unlock_irq(ha->host->host_lock); err = request_firmware(&fw, ql1280_board_tbl[ha->devnum].fwname, &ha->pdev->dev); + spin_lock_irq(ha->host->host_lock); if (err) { printk(KERN_ERR "Failed to load image \"%s\" err %d\n", ql1280_board_tbl[ha->devnum].fwname, err); -- cgit v0.10.2 From 341af10239c4c87192bf762f53c7bcb1f3a1e767 Mon Sep 17 00:00:00 2001 From: James Smart Date: Tue, 26 Jan 2010 23:07:37 -0500 Subject: [SCSI] lpfc 8.3.8: BugFixes: SLI relates changes Fix hardware/SLI relates issues: - Handle XB bit so that ELS XRIs are not prematurely released. - Handle XB bit so that FCP XRIs are not prematurely released. - Define new security SLI Commands. - Remove unused security SLI commands - Skip receive data size parameter check on received FLOGI. - Added LPFC_USE_FCPWQIDX flag to iocb to force SLI layer to submit abort WQE on same WQ as the command WQE. Signed-off-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/lpfc/lpfc_crtn.h b/drivers/scsi/lpfc/lpfc_crtn.h index 650494d..14ee7d4 100644 --- a/drivers/scsi/lpfc/lpfc_crtn.h +++ b/drivers/scsi/lpfc/lpfc_crtn.h @@ -99,7 +99,7 @@ int lpfc_disc_state_machine(struct lpfc_vport *, struct lpfc_nodelist *, void *, void lpfc_do_scr_ns_plogi(struct lpfc_hba *, struct lpfc_vport *); int lpfc_check_sparm(struct lpfc_vport *, struct lpfc_nodelist *, - struct serv_parm *, uint32_t); + struct serv_parm *, uint32_t, int); int lpfc_els_abort(struct lpfc_hba *, struct lpfc_nodelist *); void lpfc_more_plogi(struct lpfc_vport *); void lpfc_more_adisc(struct lpfc_vport *); diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c index 2cc3968..32e3c8d 100644 --- a/drivers/scsi/lpfc/lpfc_els.c +++ b/drivers/scsi/lpfc/lpfc_els.c @@ -4385,7 +4385,7 @@ lpfc_els_rcv_flogi(struct lpfc_vport *vport, struct lpfc_iocbq *cmdiocb, did = Fabric_DID; - if ((lpfc_check_sparm(vport, ndlp, sp, CLASS3))) { + if ((lpfc_check_sparm(vport, ndlp, sp, CLASS3, 1))) { /* For a FLOGI we accept, then if our portname is greater * then the remote portname we initiate Nport login. */ diff --git a/drivers/scsi/lpfc/lpfc_hw.h b/drivers/scsi/lpfc/lpfc_hw.h index c9faa1d..44c2587 100644 --- a/drivers/scsi/lpfc/lpfc_hw.h +++ b/drivers/scsi/lpfc/lpfc_hw.h @@ -1465,17 +1465,13 @@ typedef struct { /* FireFly BIU registers */ #define CMD_IOCB_LOGENTRY_CN 0x94 #define CMD_IOCB_LOGENTRY_ASYNC_CN 0x96 -/* Unhandled Data Security SLI Commands */ -#define DSSCMD_IWRITE64_CR 0xD8 -#define DSSCMD_IWRITE64_CX 0xD9 -#define DSSCMD_IREAD64_CR 0xDA -#define DSSCMD_IREAD64_CX 0xDB -#define DSSCMD_INVALIDATE_DEK 0xDC -#define DSSCMD_SET_KEK 0xDD -#define DSSCMD_GET_KEK_ID 0xDE -#define DSSCMD_GEN_XFER 0xDF - -#define CMD_MAX_IOCB_CMD 0xE6 +/* Data Security SLI Commands */ +#define DSSCMD_IWRITE64_CR 0xF8 +#define DSSCMD_IWRITE64_CX 0xF9 +#define DSSCMD_IREAD64_CR 0xFA +#define DSSCMD_IREAD64_CX 0xFB + +#define CMD_MAX_IOCB_CMD 0xFB #define CMD_IOCB_MASK 0xff #define MAX_MSG_DATA 28 /* max msg data in CMD_ADAPTER_MSG diff --git a/drivers/scsi/lpfc/lpfc_nportdisc.c b/drivers/scsi/lpfc/lpfc_nportdisc.c index 2ed6af1..293234a 100644 --- a/drivers/scsi/lpfc/lpfc_nportdisc.c +++ b/drivers/scsi/lpfc/lpfc_nportdisc.c @@ -62,7 +62,7 @@ lpfc_check_adisc(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, int lpfc_check_sparm(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, - struct serv_parm * sp, uint32_t class) + struct serv_parm *sp, uint32_t class, int flogi) { volatile struct serv_parm *hsp = &vport->fc_sparam; uint16_t hsp_value, ssp_value = 0; @@ -75,49 +75,56 @@ lpfc_check_sparm(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, * correcting the byte values. */ if (sp->cls1.classValid) { - hsp_value = (hsp->cls1.rcvDataSizeMsb << 8) | - hsp->cls1.rcvDataSizeLsb; - ssp_value = (sp->cls1.rcvDataSizeMsb << 8) | - sp->cls1.rcvDataSizeLsb; - if (!ssp_value) - goto bad_service_param; - if (ssp_value > hsp_value) { - sp->cls1.rcvDataSizeLsb = hsp->cls1.rcvDataSizeLsb; - sp->cls1.rcvDataSizeMsb = hsp->cls1.rcvDataSizeMsb; + if (!flogi) { + hsp_value = ((hsp->cls1.rcvDataSizeMsb << 8) | + hsp->cls1.rcvDataSizeLsb); + ssp_value = ((sp->cls1.rcvDataSizeMsb << 8) | + sp->cls1.rcvDataSizeLsb); + if (!ssp_value) + goto bad_service_param; + if (ssp_value > hsp_value) { + sp->cls1.rcvDataSizeLsb = + hsp->cls1.rcvDataSizeLsb; + sp->cls1.rcvDataSizeMsb = + hsp->cls1.rcvDataSizeMsb; + } } - } else if (class == CLASS1) { + } else if (class == CLASS1) goto bad_service_param; - } - if (sp->cls2.classValid) { - hsp_value = (hsp->cls2.rcvDataSizeMsb << 8) | - hsp->cls2.rcvDataSizeLsb; - ssp_value = (sp->cls2.rcvDataSizeMsb << 8) | - sp->cls2.rcvDataSizeLsb; - if (!ssp_value) - goto bad_service_param; - if (ssp_value > hsp_value) { - sp->cls2.rcvDataSizeLsb = hsp->cls2.rcvDataSizeLsb; - sp->cls2.rcvDataSizeMsb = hsp->cls2.rcvDataSizeMsb; + if (!flogi) { + hsp_value = ((hsp->cls2.rcvDataSizeMsb << 8) | + hsp->cls2.rcvDataSizeLsb); + ssp_value = ((sp->cls2.rcvDataSizeMsb << 8) | + sp->cls2.rcvDataSizeLsb); + if (!ssp_value) + goto bad_service_param; + if (ssp_value > hsp_value) { + sp->cls2.rcvDataSizeLsb = + hsp->cls2.rcvDataSizeLsb; + sp->cls2.rcvDataSizeMsb = + hsp->cls2.rcvDataSizeMsb; + } } - } else if (class == CLASS2) { + } else if (class == CLASS2) goto bad_service_param; - } - if (sp->cls3.classValid) { - hsp_value = (hsp->cls3.rcvDataSizeMsb << 8) | - hsp->cls3.rcvDataSizeLsb; - ssp_value = (sp->cls3.rcvDataSizeMsb << 8) | - sp->cls3.rcvDataSizeLsb; - if (!ssp_value) - goto bad_service_param; - if (ssp_value > hsp_value) { - sp->cls3.rcvDataSizeLsb = hsp->cls3.rcvDataSizeLsb; - sp->cls3.rcvDataSizeMsb = hsp->cls3.rcvDataSizeMsb; + if (!flogi) { + hsp_value = ((hsp->cls3.rcvDataSizeMsb << 8) | + hsp->cls3.rcvDataSizeLsb); + ssp_value = ((sp->cls3.rcvDataSizeMsb << 8) | + sp->cls3.rcvDataSizeLsb); + if (!ssp_value) + goto bad_service_param; + if (ssp_value > hsp_value) { + sp->cls3.rcvDataSizeLsb = + hsp->cls3.rcvDataSizeLsb; + sp->cls3.rcvDataSizeMsb = + hsp->cls3.rcvDataSizeMsb; + } } - } else if (class == CLASS3) { + } else if (class == CLASS3) goto bad_service_param; - } /* * Preserve the upper four bits of the MSB from the PLOGI response. @@ -295,7 +302,7 @@ lpfc_rcv_plogi(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, NULL); return 0; } - if ((lpfc_check_sparm(vport, ndlp, sp, CLASS3) == 0)) { + if ((lpfc_check_sparm(vport, ndlp, sp, CLASS3, 0) == 0)) { /* Reject this request because invalid parameters */ stat.un.b.lsRjtRsnCode = LSRJT_UNABLE_TPC; stat.un.b.lsRjtRsnCodeExp = LSEXP_SPARM_OPTIONS; @@ -831,7 +838,7 @@ lpfc_cmpl_plogi_plogi_issue(struct lpfc_vport *vport, "0142 PLOGI RSP: Invalid WWN.\n"); goto out; } - if (!lpfc_check_sparm(vport, ndlp, sp, CLASS3)) + if (!lpfc_check_sparm(vport, ndlp, sp, CLASS3, 0)) goto out; /* PLOGI chkparm OK */ lpfc_printf_vlog(vport, KERN_INFO, LOG_ELS, diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c index a246410..d5cc6b8 100644 --- a/drivers/scsi/lpfc/lpfc_scsi.c +++ b/drivers/scsi/lpfc/lpfc_scsi.c @@ -626,6 +626,7 @@ lpfc_sli4_fcp_xri_aborted(struct lpfc_hba *phba, &phba->sli4_hba.lpfc_abts_scsi_buf_list, list) { if (psb->cur_iocbq.sli4_xritag == xri) { list_del(&psb->list); + psb->exch_busy = 0; psb->status = IOSTAT_SUCCESS; spin_unlock_irqrestore( &phba->sli4_hba.abts_scsi_buf_list_lock, @@ -688,11 +689,12 @@ lpfc_sli4_repost_scsi_sgl_list(struct lpfc_hba *phba) list); if (status) { /* Put this back on the abort scsi list */ - psb->status = IOSTAT_LOCAL_REJECT; - psb->result = IOERR_ABORT_REQUESTED; + psb->exch_busy = 1; rc++; - } else + } else { + psb->exch_busy = 0; psb->status = IOSTAT_SUCCESS; + } /* Put it back into the SCSI buffer list */ lpfc_release_scsi_buf_s4(phba, psb); } @@ -839,11 +841,12 @@ lpfc_new_scsi_buf_s4(struct lpfc_vport *vport, int num_to_alloc) psb->cur_iocbq.sli4_xritag); if (status) { /* Put this back on the abort scsi list */ - psb->status = IOSTAT_LOCAL_REJECT; - psb->result = IOERR_ABORT_REQUESTED; + psb->exch_busy = 1; rc++; - } else + } else { + psb->exch_busy = 0; psb->status = IOSTAT_SUCCESS; + } /* Put it back into the SCSI buffer list */ lpfc_release_scsi_buf_s4(phba, psb); break; @@ -857,11 +860,12 @@ lpfc_new_scsi_buf_s4(struct lpfc_vport *vport, int num_to_alloc) list); if (status) { /* Put this back on the abort scsi list */ - psb->status = IOSTAT_LOCAL_REJECT; - psb->result = IOERR_ABORT_REQUESTED; + psb->exch_busy = 1; rc++; - } else + } else { + psb->exch_busy = 0; psb->status = IOSTAT_SUCCESS; + } /* Put it back into the SCSI buffer list */ lpfc_release_scsi_buf_s4(phba, psb); } @@ -951,8 +955,7 @@ lpfc_release_scsi_buf_s4(struct lpfc_hba *phba, struct lpfc_scsi_buf *psb) { unsigned long iflag = 0; - if (psb->status == IOSTAT_LOCAL_REJECT - && psb->result == IOERR_ABORT_REQUESTED) { + if (psb->exch_busy) { spin_lock_irqsave(&phba->sli4_hba.abts_scsi_buf_list_lock, iflag); psb->pCmd = NULL; @@ -2221,6 +2224,9 @@ lpfc_scsi_cmd_iocb_cmpl(struct lpfc_hba *phba, struct lpfc_iocbq *pIocbIn, lpfc_cmd->result = pIocbOut->iocb.un.ulpWord[4]; lpfc_cmd->status = pIocbOut->iocb.ulpStatus; + /* pick up SLI4 exhange busy status from HBA */ + lpfc_cmd->exch_busy = pIocbOut->iocb_flag & LPFC_EXCHANGE_BUSY; + if (pnode && NLP_CHK_NODE_ACT(pnode)) atomic_dec(&pnode->cmd_pending); @@ -2990,6 +2996,7 @@ lpfc_abort_handler(struct scsi_cmnd *cmnd) /* ABTS WQE must go to the same WQ as the WQE to be aborted */ abtsiocb->fcp_wqidx = iocb->fcp_wqidx; + abtsiocb->iocb_flag |= LPFC_USE_FCPWQIDX; if (lpfc_is_link_up(phba)) icmd->ulpCommand = CMD_ABORT_XRI_CN; diff --git a/drivers/scsi/lpfc/lpfc_scsi.h b/drivers/scsi/lpfc/lpfc_scsi.h index 65dfc8b..5932273 100644 --- a/drivers/scsi/lpfc/lpfc_scsi.h +++ b/drivers/scsi/lpfc/lpfc_scsi.h @@ -118,6 +118,7 @@ struct lpfc_scsi_buf { uint32_t timeout; + uint16_t exch_busy; /* SLI4 hba reported XB on complete WCQE */ uint16_t status; /* From IOCB Word 7- ulpStatus */ uint32_t result; /* From IOCB Word 4. */ diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c index 589549b..dc7c5c1 100644 --- a/drivers/scsi/lpfc/lpfc_sli.c +++ b/drivers/scsi/lpfc/lpfc_sli.c @@ -580,10 +580,7 @@ __lpfc_sli_release_iocbq_s4(struct lpfc_hba *phba, struct lpfc_iocbq *iocbq) else sglq = __lpfc_clear_active_sglq(phba, iocbq->sli4_xritag); if (sglq) { - if (iocbq->iocb_flag & LPFC_DRIVER_ABORTED - && ((iocbq->iocb.ulpStatus == IOSTAT_LOCAL_REJECT) - && (iocbq->iocb.un.ulpWord[4] - == IOERR_ABORT_REQUESTED))) { + if (iocbq->iocb_flag & LPFC_EXCHANGE_BUSY) { spin_lock_irqsave(&phba->sli4_hba.abts_sgl_list_lock, iflag); list_add(&sglq->list, @@ -764,10 +761,6 @@ lpfc_sli_iocb_cmd_type(uint8_t iocb_cmnd) case DSSCMD_IWRITE64_CX: case DSSCMD_IREAD64_CR: case DSSCMD_IREAD64_CX: - case DSSCMD_INVALIDATE_DEK: - case DSSCMD_SET_KEK: - case DSSCMD_GET_KEK_ID: - case DSSCMD_GEN_XFER: type = LPFC_SOL_IOCB; break; case CMD_ABORT_XRI_CN: @@ -2228,9 +2221,15 @@ lpfc_sli_process_sol_iocb(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, * All other are passed to the completion callback. */ if (pring->ringno == LPFC_ELS_RING) { - if (cmdiocbp->iocb_flag & LPFC_DRIVER_ABORTED) { + if ((phba->sli_rev < LPFC_SLI_REV4) && + (cmdiocbp->iocb_flag & + LPFC_DRIVER_ABORTED)) { + spin_lock_irqsave(&phba->hbalock, + iflag); cmdiocbp->iocb_flag &= ~LPFC_DRIVER_ABORTED; + spin_unlock_irqrestore(&phba->hbalock, + iflag); saveq->iocb.ulpStatus = IOSTAT_LOCAL_REJECT; saveq->iocb.un.ulpWord[4] = @@ -2240,7 +2239,47 @@ lpfc_sli_process_sol_iocb(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, * of DMAing payload, so don't free data * buffer till after a hbeat. */ + spin_lock_irqsave(&phba->hbalock, + iflag); saveq->iocb_flag |= LPFC_DELAY_MEM_FREE; + spin_unlock_irqrestore(&phba->hbalock, + iflag); + } + if ((phba->sli_rev == LPFC_SLI_REV4) && + (saveq->iocb_flag & LPFC_EXCHANGE_BUSY)) { + /* Set cmdiocb flag for the exchange + * busy so sgl (xri) will not be + * released until the abort xri is + * received from hba, clear the + * LPFC_DRIVER_ABORTED bit in case + * it was driver initiated abort. + */ + spin_lock_irqsave(&phba->hbalock, + iflag); + cmdiocbp->iocb_flag &= + ~LPFC_DRIVER_ABORTED; + cmdiocbp->iocb_flag |= + LPFC_EXCHANGE_BUSY; + spin_unlock_irqrestore(&phba->hbalock, + iflag); + cmdiocbp->iocb.ulpStatus = + IOSTAT_LOCAL_REJECT; + cmdiocbp->iocb.un.ulpWord[4] = + IOERR_ABORT_REQUESTED; + /* + * For SLI4, irsiocb contains NO_XRI + * in sli_xritag, it shall not affect + * releasing sgl (xri) process. + */ + saveq->iocb.ulpStatus = + IOSTAT_LOCAL_REJECT; + saveq->iocb.un.ulpWord[4] = + IOERR_SLI_ABORTED; + spin_lock_irqsave(&phba->hbalock, + iflag); + saveq->iocb_flag |= LPFC_DELAY_MEM_FREE; + spin_unlock_irqrestore(&phba->hbalock, + iflag); } } (cmdiocbp->iocb_cmpl) (phba, cmdiocbp, saveq); @@ -5987,12 +6026,10 @@ lpfc_sli4_iocb2wqe(struct lpfc_hba *phba, struct lpfc_iocbq *iocbq, else bf_set(abort_cmd_ia, &wqe->abort_cmd, 0); bf_set(abort_cmd_criteria, &wqe->abort_cmd, T_XRI_TAG); - abort_tag = iocbq->iocb.un.acxri.abortIoTag; wqe->words[5] = 0; bf_set(lpfc_wqe_gen_ct, &wqe->generic, ((iocbq->iocb.ulpCt_h << 1) | iocbq->iocb.ulpCt_l)); abort_tag = iocbq->iocb.un.acxri.abortIoTag; - wqe->generic.abort_tag = abort_tag; /* * The abort handler will send us CMD_ABORT_XRI_CN or * CMD_CLOSE_XRI_CN and the fw only accepts CMD_ABORT_XRI_CX @@ -6121,15 +6158,15 @@ __lpfc_sli_issue_iocb_s4(struct lpfc_hba *phba, uint32_t ring_number, if (lpfc_sli4_iocb2wqe(phba, piocb, &wqe)) return IOCB_ERROR; - if (piocb->iocb_flag & LPFC_IO_FCP) { + if ((piocb->iocb_flag & LPFC_IO_FCP) || + (piocb->iocb_flag & LPFC_USE_FCPWQIDX)) { /* * For FCP command IOCB, get a new WQ index to distribute * WQE across the WQsr. On the other hand, for abort IOCB, * it carries the same WQ index to the original command * IOCB. */ - if ((piocb->iocb.ulpCommand != CMD_ABORT_XRI_CN) && - (piocb->iocb.ulpCommand != CMD_CLOSE_XRI_CN)) + if (piocb->iocb_flag & LPFC_IO_FCP) piocb->fcp_wqidx = lpfc_sli4_scmd_to_wqidx_distr(phba); if (lpfc_sli4_wq_put(phba->sli4_hba.fcp_wq[piocb->fcp_wqidx], &wqe)) @@ -7004,7 +7041,14 @@ lpfc_sli_abort_els_cmpl(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, abort_iocb->iocb.ulpContext != abort_context || (abort_iocb->iocb_flag & LPFC_DRIVER_ABORTED) == 0) spin_unlock_irq(&phba->hbalock); - else { + else if (phba->sli_rev < LPFC_SLI_REV4) { + /* + * leave the SLI4 aborted command on the txcmplq + * list and the command complete WCQE's XB bit + * will tell whether the SGL (XRI) can be released + * immediately or to the aborted SGL list for the + * following abort XRI from the HBA. + */ list_del_init(&abort_iocb->list); pring->txcmplq_cnt--; spin_unlock_irq(&phba->hbalock); @@ -7013,11 +7057,13 @@ lpfc_sli_abort_els_cmpl(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, * payload, so don't free data buffer till after * a hbeat. */ + spin_lock_irq(&phba->hbalock); abort_iocb->iocb_flag |= LPFC_DELAY_MEM_FREE; - abort_iocb->iocb_flag &= ~LPFC_DRIVER_ABORTED; + spin_unlock_irq(&phba->hbalock); + abort_iocb->iocb.ulpStatus = IOSTAT_LOCAL_REJECT; - abort_iocb->iocb.un.ulpWord[4] = IOERR_SLI_ABORTED; + abort_iocb->iocb.un.ulpWord[4] = IOERR_ABORT_REQUESTED; (abort_iocb->iocb_cmpl)(phba, abort_iocb, abort_iocb); } } @@ -7106,7 +7152,7 @@ lpfc_sli_issue_abort_iotag(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, return 0; /* This signals the response to set the correct status - * before calling the completion handler. + * before calling the completion handler */ cmdiocb->iocb_flag |= LPFC_DRIVER_ABORTED; @@ -7124,6 +7170,8 @@ lpfc_sli_issue_abort_iotag(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, /* ABTS WQE must go to the same WQ as the WQE to be aborted */ abtsiocbp->fcp_wqidx = cmdiocb->fcp_wqidx; + if (cmdiocb->iocb_flag & LPFC_IO_FCP) + abtsiocbp->iocb_flag |= LPFC_USE_FCPWQIDX; if (phba->link_state >= LPFC_LINK_UP) iabt->ulpCommand = CMD_ABORT_XRI_CN; @@ -7330,6 +7378,8 @@ lpfc_sli_abort_iocb(struct lpfc_vport *vport, struct lpfc_sli_ring *pring, /* ABTS WQE must go to the same WQ as the WQE to be aborted */ abtsiocb->fcp_wqidx = iocbq->fcp_wqidx; + if (iocbq->iocb_flag & LPFC_IO_FCP) + abtsiocb->iocb_flag |= LPFC_USE_FCPWQIDX; if (lpfc_is_link_up(phba)) abtsiocb->iocb.ulpCommand = CMD_ABORT_XRI_CN; @@ -8359,11 +8409,24 @@ void lpfc_sli4_els_xri_abort_event_proc(struct lpfc_hba *phba) } } +/** + * lpfc_sli4_iocb_param_transfer - Transfer pIocbOut and cmpl status to pIocbIn + * @phba: pointer to lpfc hba data structure + * @pIocbIn: pointer to the rspiocbq + * @pIocbOut: pointer to the cmdiocbq + * @wcqe: pointer to the complete wcqe + * + * This routine transfers the fields of a command iocbq to a response iocbq + * by copying all the IOCB fields from command iocbq and transferring the + * completion status information from the complete wcqe. + **/ static void -lpfc_sli4_iocb_param_transfer(struct lpfc_iocbq *pIocbIn, +lpfc_sli4_iocb_param_transfer(struct lpfc_hba *phba, + struct lpfc_iocbq *pIocbIn, struct lpfc_iocbq *pIocbOut, struct lpfc_wcqe_complete *wcqe) { + unsigned long iflags; size_t offset = offsetof(struct lpfc_iocbq, iocb); memcpy((char *)pIocbIn + offset, (char *)pIocbOut + offset, @@ -8379,6 +8442,13 @@ lpfc_sli4_iocb_param_transfer(struct lpfc_iocbq *pIocbIn, pIocbIn->iocb.un.ulpWord[4] = wcqe->parameter; else pIocbIn->iocb.un.ulpWord[4] = wcqe->parameter; + + /* Pick up HBA exchange busy condition */ + if (bf_get(lpfc_wcqe_c_xb, wcqe)) { + spin_lock_irqsave(&phba->hbalock, iflags); + pIocbIn->iocb_flag |= LPFC_EXCHANGE_BUSY; + spin_unlock_irqrestore(&phba->hbalock, iflags); + } } /** @@ -8419,7 +8489,7 @@ lpfc_sli4_els_wcqe_to_rspiocbq(struct lpfc_hba *phba, } /* Fake the irspiocbq and copy necessary response information */ - lpfc_sli4_iocb_param_transfer(irspiocbq, cmdiocbq, wcqe); + lpfc_sli4_iocb_param_transfer(phba, irspiocbq, cmdiocbq, wcqe); return irspiocbq; } @@ -8976,7 +9046,7 @@ lpfc_sli4_fp_handle_fcp_wcqe(struct lpfc_hba *phba, } /* Fake the irspiocb and copy necessary response information */ - lpfc_sli4_iocb_param_transfer(&irspiocbq, cmdiocbq, wcqe); + lpfc_sli4_iocb_param_transfer(phba, &irspiocbq, cmdiocbq, wcqe); /* Pass the cmd_iocb and the rsp state to the upper layer */ (cmdiocbq->iocb_cmpl)(phba, cmdiocbq, &irspiocbq); diff --git a/drivers/scsi/lpfc/lpfc_sli.h b/drivers/scsi/lpfc/lpfc_sli.h index ba38de3..dfcf543 100644 --- a/drivers/scsi/lpfc/lpfc_sli.h +++ b/drivers/scsi/lpfc/lpfc_sli.h @@ -53,17 +53,19 @@ struct lpfc_iocbq { IOCB_t iocb; /* IOCB cmd */ uint8_t retry; /* retry counter for IOCB cmd - if needed */ - uint8_t iocb_flag; + uint16_t iocb_flag; #define LPFC_IO_LIBDFC 1 /* libdfc iocb */ #define LPFC_IO_WAKE 2 /* High Priority Queue signal flag */ #define LPFC_IO_FCP 4 /* FCP command -- iocbq in scsi_buf */ #define LPFC_DRIVER_ABORTED 8 /* driver aborted this request */ #define LPFC_IO_FABRIC 0x10 /* Iocb send using fabric scheduler */ #define LPFC_DELAY_MEM_FREE 0x20 /* Defer free'ing of FC data */ -#define LPFC_FIP_ELS_ID_MASK 0xc0 /* ELS_ID range 0-3 */ -#define LPFC_FIP_ELS_ID_SHIFT 6 +#define LPFC_EXCHANGE_BUSY 0x40 /* SLI4 hba reported XB in response */ +#define LPFC_USE_FCPWQIDX 0x80 /* Submit to specified FCPWQ index */ + +#define LPFC_FIP_ELS_ID_MASK 0xc000 /* ELS_ID range 0-3, non-shifted mask */ +#define LPFC_FIP_ELS_ID_SHIFT 14 - uint8_t abort_count; uint8_t rsvd2; uint32_t drvrTimeout; /* driver timeout in seconds */ uint32_t fcp_wqidx; /* index to FCP work queue */ -- cgit v0.10.2 From 695a814e18561c52456acf5051fac0ea4b8111da Mon Sep 17 00:00:00 2001 From: James Smart Date: Tue, 26 Jan 2010 23:08:03 -0500 Subject: [SCSI] lpfc 8.3.8: BugFixes: Discovery relates changes Discovery relates changes: - Separated VPI_REGISTERED state of physical port into VFI_REGISTERED and VPI_REGISTERED state so that driver can unregister physical port VPI independent of VFI. - Add code to unregister, re-init and re-register physical port VPI when physical port NportID change. - Add code to unregister and re-register VPI of a vport when its Nport ID change. - Add code in FDISC completion path to re-start FLOGI discovery when a FDISC complete with LOGIN_REQUIRED reason code. - Fix a memory leak in lpfc_init_vpi_cmpl - Add code to start a timer for vport to retry FDISC when CVL is received by a vport or physical port. If all Nports receive CVLs, then all timers are cancelled and a logical link level discovery will be started after one second. - Flush ELS commands after killing all delayed ELS commands. Signed-off-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/lpfc/lpfc.h b/drivers/scsi/lpfc/lpfc.h index 1cc23a6..197dc3e 100644 --- a/drivers/scsi/lpfc/lpfc.h +++ b/drivers/scsi/lpfc/lpfc.h @@ -315,6 +315,9 @@ struct lpfc_vport { #define FC_VPORT_NEEDS_REG_VPI 0x80000 /* Needs to have its vpi registered */ #define FC_RSCN_DEFERRED 0x100000 /* A deferred RSCN being processed */ #define FC_VPORT_NEEDS_INIT_VPI 0x200000 /* Need to INIT_VPI before FDISC */ +#define FC_VPORT_CVL_RCVD 0x400000 /* VLink failed due to CVL */ +#define FC_VFI_REGISTERED 0x800000 /* VFI is registered */ +#define FC_FDISC_COMPLETED 0x1000000/* FDISC completed */ uint32_t ct_flags; #define FC_CT_RFF_ID 0x1 /* RFF_ID accepted by switch */ diff --git a/drivers/scsi/lpfc/lpfc_crtn.h b/drivers/scsi/lpfc/lpfc_crtn.h index 14ee7d4..b9e8cd5 100644 --- a/drivers/scsi/lpfc/lpfc_crtn.h +++ b/drivers/scsi/lpfc/lpfc_crtn.h @@ -44,6 +44,8 @@ int lpfc_reg_rpi(struct lpfc_hba *, uint16_t, uint32_t, uint8_t *, void lpfc_unreg_login(struct lpfc_hba *, uint16_t, uint32_t, LPFC_MBOXQ_t *); void lpfc_unreg_did(struct lpfc_hba *, uint16_t, uint32_t, LPFC_MBOXQ_t *); void lpfc_reg_vpi(struct lpfc_vport *, LPFC_MBOXQ_t *); +void lpfc_register_new_vport(struct lpfc_hba *, struct lpfc_vport *, + struct lpfc_nodelist *); void lpfc_unreg_vpi(struct lpfc_hba *, uint16_t, LPFC_MBOXQ_t *); void lpfc_init_link(struct lpfc_hba *, LPFC_MBOXQ_t *, uint32_t, uint32_t); void lpfc_request_features(struct lpfc_hba *, struct lpfcMboxq *); @@ -52,10 +54,13 @@ struct lpfc_vport *lpfc_find_vport_by_did(struct lpfc_hba *, uint32_t); void lpfc_cleanup_rcv_buffers(struct lpfc_vport *); void lpfc_rcv_seq_check_edtov(struct lpfc_vport *); void lpfc_cleanup_rpis(struct lpfc_vport *, int); +void lpfc_cleanup_pending_mbox(struct lpfc_vport *); int lpfc_linkdown(struct lpfc_hba *); void lpfc_linkdown_port(struct lpfc_vport *); void lpfc_port_link_failure(struct lpfc_vport *); void lpfc_mbx_cmpl_read_la(struct lpfc_hba *, LPFC_MBOXQ_t *); +void lpfc_init_vpi_cmpl(struct lpfc_hba *, LPFC_MBOXQ_t *); +void lpfc_retry_pport_discovery(struct lpfc_hba *); void lpfc_mbx_cmpl_reg_login(struct lpfc_hba *, LPFC_MBOXQ_t *); void lpfc_mbx_cmpl_dflt_rpi(struct lpfc_hba *, LPFC_MBOXQ_t *); diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c index 32e3c8d..2c62349 100644 --- a/drivers/scsi/lpfc/lpfc_els.c +++ b/drivers/scsi/lpfc/lpfc_els.c @@ -50,9 +50,6 @@ static int lpfc_issue_els_fdisc(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, uint8_t retry); static int lpfc_issue_fabric_iocb(struct lpfc_hba *phba, struct lpfc_iocbq *iocb); -static void lpfc_register_new_vport(struct lpfc_hba *phba, - struct lpfc_vport *vport, - struct lpfc_nodelist *ndlp); static int lpfc_max_els_tries = 3; @@ -604,10 +601,13 @@ lpfc_cmpl_els_flogi_fabric(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, } else { ndlp->nlp_type |= NLP_FABRIC; lpfc_nlp_set_state(vport, ndlp, NLP_STE_UNMAPPED_NODE); - if (vport->vpi_state & LPFC_VPI_REGISTERED) { + if ((!(vport->fc_flag & FC_VPORT_NEEDS_REG_VPI)) && + (vport->vpi_state & LPFC_VPI_REGISTERED)) { lpfc_start_fdiscs(phba); lpfc_do_scr_ns_plogi(phba, vport); - } else + } else if (vport->fc_flag & FC_VFI_REGISTERED) + lpfc_register_new_vport(phba, vport, ndlp); + else lpfc_issue_reg_vfi(vport); } return 0; @@ -804,6 +804,9 @@ lpfc_cmpl_els_flogi(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, irsp->ulpTimeout); goto flogifail; } + spin_lock_irq(shost->host_lock); + vport->fc_flag &= ~FC_VPORT_CVL_RCVD; + spin_unlock_irq(shost->host_lock); /* * The FLogI succeeded. Sync the data for the CPU before @@ -2720,7 +2723,7 @@ lpfc_els_retry(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, if (did == FDMI_DID) retry = 1; - if ((cmd == ELS_CMD_FLOGI) && + if (((cmd == ELS_CMD_FLOGI) || (cmd == ELS_CMD_FDISC)) && (phba->fc_topology != TOPOLOGY_LOOP) && !lpfc_error_lost_link(irsp)) { /* FLOGI retry policy */ @@ -5915,6 +5918,7 @@ lpfc_cmpl_reg_new_vport(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) struct Scsi_Host *shost = lpfc_shost_from_vport(vport); struct lpfc_nodelist *ndlp = (struct lpfc_nodelist *) pmb->context2; MAILBOX_t *mb = &pmb->u.mb; + int rc; spin_lock_irq(shost->host_lock); vport->fc_flag &= ~FC_VPORT_NEEDS_REG_VPI; @@ -5936,6 +5940,26 @@ lpfc_cmpl_reg_new_vport(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) spin_unlock_irq(shost->host_lock); lpfc_can_disctmo(vport); break; + /* If reg_vpi fail with invalid VPI status, re-init VPI */ + case 0x20: + spin_lock_irq(shost->host_lock); + vport->fc_flag |= FC_VPORT_NEEDS_REG_VPI; + spin_unlock_irq(shost->host_lock); + lpfc_init_vpi(phba, pmb, vport->vpi); + pmb->vport = vport; + pmb->mbox_cmpl = lpfc_init_vpi_cmpl; + rc = lpfc_sli_issue_mbox(phba, pmb, + MBX_NOWAIT); + if (rc == MBX_NOT_FINISHED) { + lpfc_printf_vlog(vport, + KERN_ERR, LOG_MBOX, + "2732 Failed to issue INIT_VPI" + " mailbox command\n"); + } else { + lpfc_nlp_put(ndlp); + return; + } + default: /* Try to recover from this error */ lpfc_mbx_unreg_vpi(vport); @@ -5949,13 +5973,17 @@ lpfc_cmpl_reg_new_vport(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) break; } } else { + spin_lock_irq(shost->host_lock); vport->vpi_state |= LPFC_VPI_REGISTERED; - if (vport == phba->pport) + spin_unlock_irq(shost->host_lock); + if (vport == phba->pport) { if (phba->sli_rev < LPFC_SLI_REV4) lpfc_issue_fabric_reglogin(vport); - else - lpfc_issue_reg_vfi(vport); - else + else { + lpfc_start_fdiscs(phba); + lpfc_do_scr_ns_plogi(phba, vport); + } + } else lpfc_do_scr_ns_plogi(phba, vport); } @@ -5977,7 +6005,7 @@ lpfc_cmpl_reg_new_vport(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) * This routine registers the @vport as a new virtual port with a HBA. * It is done through a registering vpi mailbox command. **/ -static void +void lpfc_register_new_vport(struct lpfc_hba *phba, struct lpfc_vport *vport, struct lpfc_nodelist *ndlp) { @@ -6018,6 +6046,78 @@ mbox_err_exit: } /** + * lpfc_retry_pport_discovery - Start timer to retry FLOGI. + * @phba: pointer to lpfc hba data structure. + * + * This routine abort all pending discovery commands and + * start a timer to retry FLOGI for the physical port + * discovery. + **/ +void +lpfc_retry_pport_discovery(struct lpfc_hba *phba) +{ + struct lpfc_vport **vports; + struct lpfc_nodelist *ndlp; + struct Scsi_Host *shost; + int i; + uint32_t link_state; + + /* Treat this failure as linkdown for all vports */ + link_state = phba->link_state; + lpfc_linkdown(phba); + phba->link_state = link_state; + + vports = lpfc_create_vport_work_array(phba); + + if (vports) { + for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++) { + ndlp = lpfc_findnode_did(vports[i], Fabric_DID); + if (ndlp) + lpfc_cancel_retry_delay_tmo(vports[i], ndlp); + lpfc_els_flush_cmd(vports[i]); + } + lpfc_destroy_vport_work_array(phba, vports); + } + + /* If fabric require FLOGI, then re-instantiate physical login */ + ndlp = lpfc_findnode_did(phba->pport, Fabric_DID); + if (!ndlp) + return; + + + shost = lpfc_shost_from_vport(phba->pport); + mod_timer(&ndlp->nlp_delayfunc, jiffies + HZ); + spin_lock_irq(shost->host_lock); + ndlp->nlp_flag |= NLP_DELAY_TMO; + spin_unlock_irq(shost->host_lock); + ndlp->nlp_last_elscmd = ELS_CMD_FLOGI; + phba->pport->port_state = LPFC_FLOGI; + return; +} + +/** + * lpfc_fabric_login_reqd - Check if FLOGI required. + * @phba: pointer to lpfc hba data structure. + * @cmdiocb: pointer to FDISC command iocb. + * @rspiocb: pointer to FDISC response iocb. + * + * This routine checks if a FLOGI is reguired for FDISC + * to succeed. + **/ +static int +lpfc_fabric_login_reqd(struct lpfc_hba *phba, + struct lpfc_iocbq *cmdiocb, + struct lpfc_iocbq *rspiocb) +{ + + if ((rspiocb->iocb.ulpStatus != IOSTAT_FABRIC_RJT) || + (rspiocb->iocb.un.ulpWord[4] != RJT_LOGIN_REQUIRED)) + return 0; + else + return 1; +} + +/** * lpfc_cmpl_els_fdisc - Completion function for fdisc iocb command * @phba: pointer to lpfc hba data structure. * @cmdiocb: pointer to lpfc command iocb data structure. @@ -6066,6 +6166,12 @@ lpfc_cmpl_els_fdisc(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, irsp->ulpStatus, irsp->un.ulpWord[4], vport->fc_prevDID); if (irsp->ulpStatus) { + + if (lpfc_fabric_login_reqd(phba, cmdiocb, rspiocb)) { + lpfc_retry_pport_discovery(phba); + goto out; + } + /* Check for retry */ if (lpfc_els_retry(phba, cmdiocb, rspiocb)) goto out; @@ -6076,6 +6182,7 @@ lpfc_cmpl_els_fdisc(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, goto fdisc_failed; } spin_lock_irq(shost->host_lock); + vport->fc_flag &= ~FC_VPORT_CVL_RCVD; vport->fc_flag |= FC_FABRIC; if (vport->phba->fc_topology == TOPOLOGY_LOOP) vport->fc_flag |= FC_PUBLIC_LOOP; diff --git a/drivers/scsi/lpfc/lpfc_hbadisc.c b/drivers/scsi/lpfc/lpfc_hbadisc.c index 2445e39..7143d71 100644 --- a/drivers/scsi/lpfc/lpfc_hbadisc.c +++ b/drivers/scsi/lpfc/lpfc_hbadisc.c @@ -706,6 +706,8 @@ lpfc_cleanup_rpis(struct lpfc_vport *vport, int remove) void lpfc_port_link_failure(struct lpfc_vport *vport) { + lpfc_vport_set_state(vport, FC_VPORT_LINKDOWN); + /* Cleanup any outstanding received buffers */ lpfc_cleanup_rcv_buffers(vport); @@ -1695,10 +1697,11 @@ out: * * This function handles completion of init vpi mailbox command. */ -static void +void lpfc_init_vpi_cmpl(struct lpfc_hba *phba, LPFC_MBOXQ_t *mboxq) { struct lpfc_vport *vport = mboxq->vport; + struct lpfc_nodelist *ndlp; if (mboxq->u.mb.mbxStatus) { lpfc_printf_vlog(vport, KERN_ERR, LOG_MBOX, @@ -1712,6 +1715,20 @@ lpfc_init_vpi_cmpl(struct lpfc_hba *phba, LPFC_MBOXQ_t *mboxq) vport->fc_flag &= ~FC_VPORT_NEEDS_INIT_VPI; spin_unlock_irq(&phba->hbalock); + /* If this port is physical port or FDISC is done, do reg_vpi */ + if ((phba->pport == vport) || (vport->port_state == LPFC_FDISC)) { + ndlp = lpfc_findnode_did(vport, Fabric_DID); + if (!ndlp) + lpfc_printf_vlog(vport, KERN_ERR, + LOG_DISCOVERY, + "2731 Cannot find fabric " + "controller node\n"); + else + lpfc_register_new_vport(phba, vport, ndlp); + mempool_free(mboxq, phba->mbox_mem_pool); + return; + } + if (phba->link_flag & LS_NPIV_FAB_SUPPORTED) lpfc_initial_fdisc(vport); else { @@ -1719,6 +1736,7 @@ lpfc_init_vpi_cmpl(struct lpfc_hba *phba, LPFC_MBOXQ_t *mboxq) lpfc_printf_vlog(vport, KERN_ERR, LOG_ELS, "2606 No NPIV Fabric support\n"); } + mempool_free(mboxq, phba->mbox_mem_pool); return; } @@ -1814,6 +1832,9 @@ lpfc_mbx_cmpl_reg_vfi(struct lpfc_hba *phba, LPFC_MBOXQ_t *mboxq) } /* The VPI is implicitly registered when the VFI is registered */ vport->vpi_state |= LPFC_VPI_REGISTERED; + vport->fc_flag |= FC_VFI_REGISTERED; + + vport->fc_flag &= ~FC_VPORT_NEEDS_REG_VPI; if (vport->port_state == LPFC_FABRIC_CFG_LINK) { lpfc_start_fdiscs(phba); @@ -2333,6 +2354,7 @@ lpfc_mbx_cmpl_reg_vpi(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) } vport->vpi_state |= LPFC_VPI_REGISTERED; + vport->fc_flag &= ~FC_VPORT_NEEDS_REG_VPI; vport->num_disc_nodes = 0; /* go thru NPR list and issue ELS PLOGIs */ if (vport->fc_npr_cnt) @@ -4462,6 +4484,7 @@ lpfc_unregister_unused_fcf(struct lpfc_hba *phba) int rc; struct lpfc_vport **vports; int i; + struct lpfc_nodelist *ndlp; spin_lock_irq(&phba->hbalock); /* @@ -4489,6 +4512,10 @@ lpfc_unregister_unused_fcf(struct lpfc_hba *phba) if (vports && (phba->sli3_options & LPFC_SLI3_NPIV_ENABLED)) for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++) { + /* Stop FLOGI/FDISC retries */ + ndlp = lpfc_findnode_did(vports[i], Fabric_DID); + if (ndlp) + lpfc_cancel_retry_delay_tmo(vports[i], ndlp); lpfc_mbx_unreg_vpi(vports[i]); spin_lock_irq(&phba->hbalock); vports[i]->fc_flag |= FC_VPORT_NEEDS_INIT_VPI; @@ -4497,6 +4524,9 @@ lpfc_unregister_unused_fcf(struct lpfc_hba *phba) } lpfc_destroy_vport_work_array(phba, vports); + /* Cleanup any outstanding ELS commands */ + lpfc_els_flush_all_cmd(phba); + /* Unregister VFI */ mbox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); if (!mbox) { @@ -4521,6 +4551,10 @@ lpfc_unregister_unused_fcf(struct lpfc_hba *phba) return; } + spin_lock_irq(&phba->hbalock); + phba->pport->fc_flag &= ~FC_VFI_REGISTERED; + spin_unlock_irq(&phba->hbalock); + /* Unregister FCF */ mbox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); if (!mbox) { diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c index b8eb1b6..4d20c41 100644 --- a/drivers/scsi/lpfc/lpfc_init.c +++ b/drivers/scsi/lpfc/lpfc_init.c @@ -2246,6 +2246,9 @@ lpfc_offline_prep(struct lpfc_hba * phba) if (vports[i]->load_flag & FC_UNLOADING) continue; vports[i]->vpi_state &= ~LPFC_VPI_REGISTERED; + vports[i]->fc_flag |= FC_VPORT_NEEDS_REG_VPI; + vports[i]->fc_flag &= ~FC_VFI_REGISTERED; + shost = lpfc_shost_from_vport(vports[i]); list_for_each_entry_safe(ndlp, next_ndlp, &vports[i]->fc_nodes, @@ -3007,6 +3010,9 @@ lpfc_sli4_async_fcoe_evt(struct lpfc_hba *phba, struct lpfc_nodelist *ndlp; struct Scsi_Host *shost; uint32_t link_state; + int active_vlink_present; + struct lpfc_vport **vports; + int i; phba->fc_eventTag = acqe_fcoe->event_tag; phba->fcoe_eventtag = acqe_fcoe->event_tag; @@ -3074,14 +3080,46 @@ lpfc_sli4_async_fcoe_evt(struct lpfc_hba *phba, if (!ndlp) break; shost = lpfc_shost_from_vport(vport); + if (phba->pport->port_state <= LPFC_FLOGI) + break; + /* If virtual link is not yet instantiated ignore CVL */ + if (vport->port_state <= LPFC_FDISC) + break; + lpfc_linkdown_port(vport); - if (vport->port_type != LPFC_NPIV_PORT) { + lpfc_cleanup_pending_mbox(vport); + spin_lock_irq(shost->host_lock); + vport->fc_flag |= FC_VPORT_CVL_RCVD; + spin_unlock_irq(shost->host_lock); + active_vlink_present = 0; + + vports = lpfc_create_vport_work_array(phba); + if (vports) { + for (i = 0; i <= phba->max_vports && vports[i] != NULL; + i++) { + if ((!(vports[i]->fc_flag & + FC_VPORT_CVL_RCVD)) && + (vports[i]->port_state > LPFC_FDISC)) { + active_vlink_present = 1; + break; + } + } + lpfc_destroy_vport_work_array(phba, vports); + } + + if (active_vlink_present) { + /* + * If there are other active VLinks present, + * re-instantiate the Vlink using FDISC. + */ mod_timer(&ndlp->nlp_delayfunc, jiffies + HZ); spin_lock_irq(shost->host_lock); ndlp->nlp_flag |= NLP_DELAY_TMO; spin_unlock_irq(shost->host_lock); - ndlp->nlp_last_elscmd = ELS_CMD_FLOGI; - vport->port_state = LPFC_FLOGI; + ndlp->nlp_last_elscmd = ELS_CMD_FDISC; + vport->port_state = LPFC_FDISC; + } else { + lpfc_retry_pport_discovery(phba); } break; default: diff --git a/drivers/scsi/lpfc/lpfc_nportdisc.c b/drivers/scsi/lpfc/lpfc_nportdisc.c index 293234a..d20ae6b 100644 --- a/drivers/scsi/lpfc/lpfc_nportdisc.c +++ b/drivers/scsi/lpfc/lpfc_nportdisc.c @@ -254,7 +254,7 @@ lpfc_rcv_plogi(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, int rc; memset(&stat, 0, sizeof (struct ls_rjt)); - if (vport->port_state <= LPFC_FLOGI) { + if (vport->port_state <= LPFC_FDISC) { /* Before responding to PLOGI, check for pt2pt mode. * If we are pt2pt, with an outstanding FLOGI, abort * the FLOGI and resend it first. diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c index dc7c5c1..8d666d9 100644 --- a/drivers/scsi/lpfc/lpfc_sli.c +++ b/drivers/scsi/lpfc/lpfc_sli.c @@ -1710,6 +1710,7 @@ lpfc_sli_def_mbox_cmpl(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) struct lpfc_dmabuf *mp; uint16_t rpi, vpi; int rc; + struct lpfc_vport *vport = pmb->vport; mp = (struct lpfc_dmabuf *) (pmb->context1); @@ -1738,6 +1739,18 @@ lpfc_sli_def_mbox_cmpl(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb) return; } + /* Unreg VPI, if the REG_VPI succeed after VLink failure */ + if ((pmb->u.mb.mbxCommand == MBX_REG_VPI) && + !(phba->pport->load_flag & FC_UNLOADING) && + !pmb->u.mb.mbxStatus) { + lpfc_unreg_vpi(phba, pmb->u.mb.un.varRegVpi.vpi, pmb); + pmb->vport = vport; + pmb->mbox_cmpl = lpfc_sli_def_mbox_cmpl; + rc = lpfc_sli_issue_mbox(phba, pmb, MBX_NOWAIT); + if (rc != MBX_NOT_FINISHED) + return; + } + if (bf_get(lpfc_mqe_command, &pmb->u.mqe) == MBX_SLI4_CONFIG) lpfc_sli4_mbox_cmd_free(phba, pmb); else @@ -8440,8 +8453,10 @@ lpfc_sli4_iocb_param_transfer(struct lpfc_hba *phba, wcqe->total_data_placed; else pIocbIn->iocb.un.ulpWord[4] = wcqe->parameter; - else + else { pIocbIn->iocb.un.ulpWord[4] = wcqe->parameter; + pIocbIn->iocb.un.genreq64.bdl.bdeSize = wcqe->total_data_placed; + } /* Pick up HBA exchange busy condition */ if (bf_get(lpfc_wcqe_c_xb, wcqe)) { @@ -12139,3 +12154,48 @@ out: kfree(rgn23_data); return; } + +/** + * lpfc_cleanup_pending_mbox - Free up vport discovery mailbox commands. + * @vport: pointer to vport data structure. + * + * This function iterate through the mailboxq and clean up all REG_LOGIN + * and REG_VPI mailbox commands associated with the vport. This function + * is called when driver want to restart discovery of the vport due to + * a Clear Virtual Link event. + **/ +void +lpfc_cleanup_pending_mbox(struct lpfc_vport *vport) +{ + struct lpfc_hba *phba = vport->phba; + LPFC_MBOXQ_t *mb, *nextmb; + struct lpfc_dmabuf *mp; + + spin_lock_irq(&phba->hbalock); + list_for_each_entry_safe(mb, nextmb, &phba->sli.mboxq, list) { + if (mb->vport != vport) + continue; + + if ((mb->u.mb.mbxCommand != MBX_REG_LOGIN64) && + (mb->u.mb.mbxCommand != MBX_REG_VPI)) + continue; + + if (mb->u.mb.mbxCommand == MBX_REG_LOGIN64) { + mp = (struct lpfc_dmabuf *) (mb->context1); + if (mp) { + __lpfc_mbuf_free(phba, mp->virt, mp->phys); + kfree(mp); + } + } + list_del(&mb->list); + mempool_free(mb, phba->mbox_mem_pool); + } + mb = phba->sli.mbox_active; + if (mb && (mb->vport == vport)) { + if ((mb->u.mb.mbxCommand == MBX_REG_LOGIN64) || + (mb->u.mb.mbxCommand == MBX_REG_VPI)) + mb->mbox_cmpl = lpfc_sli_def_mbox_cmpl; + } + spin_unlock_irq(&phba->hbalock); +} + diff --git a/drivers/scsi/lpfc/lpfc_vport.c b/drivers/scsi/lpfc/lpfc_vport.c index e3c7fa6..281ff03 100644 --- a/drivers/scsi/lpfc/lpfc_vport.c +++ b/drivers/scsi/lpfc/lpfc_vport.c @@ -389,7 +389,7 @@ lpfc_vport_create(struct fc_vport *fc_vport, bool disable) * by the port. */ if ((phba->sli_rev == LPFC_SLI_REV4) && - (pport->vpi_state & LPFC_VPI_REGISTERED)) { + (pport->fc_flag & FC_VFI_REGISTERED)) { rc = lpfc_sli4_init_vpi(phba, vpi); if (rc) { lpfc_printf_log(phba, KERN_ERR, LOG_VPORT, -- cgit v0.10.2 From 65467b6bdffd3efde111444663bc9de35b59b22a Mon Sep 17 00:00:00 2001 From: James Smart Date: Tue, 26 Jan 2010 23:08:29 -0500 Subject: [SCSI] lpfc 8.3.8: Add code to display logical link speed Display Logical Link Speed when supported and is non-zero. Signed-off-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/lpfc/lpfc_hw4.h b/drivers/scsi/lpfc/lpfc_hw4.h index 8a2a1c5..c2bec6e 100644 --- a/drivers/scsi/lpfc/lpfc_hw4.h +++ b/drivers/scsi/lpfc/lpfc_hw4.h @@ -1959,6 +1959,9 @@ struct lpfc_acqe_link { #define LPFC_ASYNC_LINK_FAULT_NONE 0x0 #define LPFC_ASYNC_LINK_FAULT_LOCAL 0x1 #define LPFC_ASYNC_LINK_FAULT_REMOTE 0x2 +#define lpfc_acqe_qos_link_speed_SHIFT 16 +#define lpfc_acqe_qos_link_speed_MASK 0x0000FFFF +#define lpfc_acqe_qos_link_speed_WORD word1 uint32_t event_tag; uint32_t trailer; }; diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c index 4d20c41..e80e95d 100644 --- a/drivers/scsi/lpfc/lpfc_init.c +++ b/drivers/scsi/lpfc/lpfc_init.c @@ -2981,6 +2981,8 @@ lpfc_sli4_async_link_evt(struct lpfc_hba *phba, bf_get(lpfc_acqe_link_physical, acqe_link); phba->sli4_hba.link_state.fault = bf_get(lpfc_acqe_link_fault, acqe_link); + phba->sli4_hba.link_state.logical_speed = + bf_get(lpfc_acqe_qos_link_speed, acqe_link); /* Invoke the lpfc_handle_latt mailbox command callback function */ lpfc_mbx_cmpl_read_la(phba, pmb); diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c index d5cc6b8..8f4b90a 100644 --- a/drivers/scsi/lpfc/lpfc_scsi.c +++ b/drivers/scsi/lpfc/lpfc_scsi.c @@ -2701,6 +2701,13 @@ lpfc_info(struct Scsi_Host *host) " port %s", phba->Port); } + len = strlen(lpfcinfobuf); + if (phba->sli4_hba.link_state.logical_speed) { + snprintf(lpfcinfobuf + len, + 384-len, + " Logical Link Speed: %d Mbps", + phba->sli4_hba.link_state.logical_speed * 10); + } } return lpfcinfobuf; } diff --git a/drivers/scsi/lpfc/lpfc_sli4.h b/drivers/scsi/lpfc/lpfc_sli4.h index 44e5f57..2e5e405 100644 --- a/drivers/scsi/lpfc/lpfc_sli4.h +++ b/drivers/scsi/lpfc/lpfc_sli4.h @@ -126,6 +126,7 @@ struct lpfc_sli4_link { uint8_t status; uint8_t physical; uint8_t fault; + uint16_t logical_speed; }; struct lpfc_fcf { -- cgit v0.10.2 From 4fede78f7552479c4bb3bab221133ec5244e4154 Mon Sep 17 00:00:00 2001 From: James Smart Date: Tue, 26 Jan 2010 23:08:55 -0500 Subject: [SCSI] lpfc 8.3.8: (BSG1) Update BSG infrastructure Update BSG infrastructure to handle new vendor specific BSG commands. Signed-off-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/lpfc/lpfc.h b/drivers/scsi/lpfc/lpfc.h index 197dc3e..a40a45b 100644 --- a/drivers/scsi/lpfc/lpfc.h +++ b/drivers/scsi/lpfc/lpfc.h @@ -1,7 +1,7 @@ /******************************************************************* * This file is part of the Emulex Linux Device Driver for * * Fibre Channel Host Bus Adapters. * - * Copyright (C) 2004-2009 Emulex. All rights reserved. * + * Copyright (C) 2004-2010 Emulex. All rights reserved. * * EMULEX and SLI are trademarks of Emulex. * * www.emulex.com * * Portions Copyright (C) 2004-2005 Christoph Hellwig * @@ -451,6 +451,8 @@ struct unsol_rcv_ct_ctx { uint32_t ctxt_id; uint32_t SID; uint32_t oxid; + uint32_t flags; +#define UNSOL_VALID 0x00000001 }; struct lpfc_hba { @@ -793,7 +795,7 @@ struct lpfc_hba { uint16_t vlan_id; struct list_head fcf_conn_rec_list; - struct mutex ct_event_mutex; /* synchronize access to ct_ev_waiters */ + spinlock_t ct_ev_lock; /* synchronize access to ct_ev_waiters */ struct list_head ct_ev_waiters; struct unsol_rcv_ct_ctx ct_ctx[64]; uint32_t ctx_idx; diff --git a/drivers/scsi/lpfc/lpfc_bsg.c b/drivers/scsi/lpfc/lpfc_bsg.c index a5d9048..dfb1f73 100644 --- a/drivers/scsi/lpfc/lpfc_bsg.c +++ b/drivers/scsi/lpfc/lpfc_bsg.c @@ -1,7 +1,7 @@ /******************************************************************* * This file is part of the Emulex Linux Device Driver for * * Fibre Channel Host Bus Adapters. * - * Copyright (C) 2009 Emulex. All rights reserved. * + * Copyright (C) 2009-2010 Emulex. All rights reserved. * * EMULEX and SLI are trademarks of Emulex. * * www.emulex.com * * * @@ -33,6 +33,7 @@ #include "lpfc_sli.h" #include "lpfc_sli4.h" #include "lpfc_nl.h" +#include "lpfc_bsg.h" #include "lpfc_disc.h" #include "lpfc_scsi.h" #include "lpfc.h" @@ -476,7 +477,7 @@ enum ELX_LOOPBACK_CMD { * This function is called when an unsolicited CT command is received. It * forwards the event to any processes registerd to receive CT events. */ -void +int lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, struct lpfc_iocbq *piocbq) { @@ -496,6 +497,7 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, struct lpfc_dmabuf *bdeBuf2 = piocbq->context3; struct lpfc_hbq_entry *hbqe; struct lpfc_sli_ct_request *ct_req; + unsigned long flags; INIT_LIST_HEAD(&head); list_add_tail(&head, &piocbq->list); @@ -519,7 +521,7 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, if (!(phba->sli3_options & LPFC_SLI3_HBQ_ENABLED)) lpfc_sli_ringpostbuf_put(phba, pring, dmabuf); - mutex_lock(&phba->ct_event_mutex); + spin_lock_irqsave(&phba->ct_ev_lock, flags); list_for_each_entry(evt, &phba->ct_ev_waiters, node) { if (evt->req_id != evt_req_id) continue; @@ -535,7 +537,7 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, break; } - mutex_unlock(&phba->ct_event_mutex); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); if (phba->sli3_options & LPFC_SLI3_HBQ_ENABLED) { /* take accumulated byte count from the last iocbq */ @@ -556,9 +558,9 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, "CT event data, size %d\n", evt_dat->len); kfree(evt_dat); - mutex_lock(&phba->ct_event_mutex); + spin_lock_irqsave(&phba->ct_ev_lock, flags); lpfc_ct_event_unref(evt); - mutex_unlock(&phba->ct_event_mutex); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); goto error_ct_unsol_exit; } @@ -601,9 +603,11 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, iocbq); kfree(evt_dat->data); kfree(evt_dat); - mutex_lock(&phba->ct_event_mutex); + spin_lock_irqsave(&phba->ct_ev_lock, + flags); lpfc_ct_event_unref(evt); - mutex_unlock(&phba->ct_event_mutex); + spin_unlock_irqrestore( + &phba->ct_ev_lock, flags); goto error_ct_unsol_exit; } memcpy((char *)(evt_dat->data) + offset, @@ -638,7 +642,7 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, } } - mutex_lock(&phba->ct_event_mutex); + spin_lock_irqsave(&phba->ct_ev_lock, flags); if (phba->sli_rev == LPFC_SLI_REV4) { evt_dat->immed_dat = phba->ctx_idx; phba->ctx_idx = (phba->ctx_idx + 1) % 64; @@ -656,13 +660,13 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, if (evt_req_id == SLI_CT_ELX_LOOPBACK) break; } - mutex_unlock(&phba->ct_event_mutex); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); error_ct_unsol_exit: if (!list_empty(&head)) list_del(&head); - return; + return 1; } /** @@ -676,6 +680,7 @@ lpfc_bsg_set_event(struct fc_bsg_job *job) struct lpfc_hba *phba = vport->phba; struct set_ct_event *event_req; struct lpfc_ct_event *evt; + unsigned long flags; int rc = 0; if (job->request_len < @@ -689,7 +694,7 @@ lpfc_bsg_set_event(struct fc_bsg_job *job) event_req = (struct set_ct_event *) job->request->rqst_data.h_vendor.vendor_cmd; - mutex_lock(&phba->ct_event_mutex); + spin_lock_irqsave(&phba->ct_ev_lock, flags); list_for_each_entry(evt, &phba->ct_ev_waiters, node) { if (evt->reg_id == event_req->ev_reg_id) { lpfc_ct_event_ref(evt); @@ -697,7 +702,7 @@ lpfc_bsg_set_event(struct fc_bsg_job *job) break; } } - mutex_unlock(&phba->ct_event_mutex); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); if (&evt->node == &phba->ct_ev_waiters) { /* no event waiting struct yet - first call */ @@ -710,19 +715,19 @@ lpfc_bsg_set_event(struct fc_bsg_job *job) return -ENOMEM; } - mutex_lock(&phba->ct_event_mutex); + spin_lock_irqsave(&phba->ct_ev_lock, flags); list_add(&evt->node, &phba->ct_ev_waiters); lpfc_ct_event_ref(evt); - mutex_unlock(&phba->ct_event_mutex); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); } evt->waiting = 1; if (wait_event_interruptible(evt->wq, !list_empty(&evt->events_to_see))) { - mutex_lock(&phba->ct_event_mutex); + spin_lock_irqsave(&phba->ct_ev_lock, flags); lpfc_ct_event_unref(evt); /* release ref */ lpfc_ct_event_unref(evt); /* delete */ - mutex_unlock(&phba->ct_event_mutex); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); rc = -EINTR; goto set_event_out; } @@ -730,10 +735,10 @@ lpfc_bsg_set_event(struct fc_bsg_job *job) evt->wait_time_stamp = jiffies; evt->waiting = 0; - mutex_lock(&phba->ct_event_mutex); + spin_lock_irqsave(&phba->ct_ev_lock, flags); list_move(evt->events_to_see.prev, &evt->events_to_get); lpfc_ct_event_unref(evt); /* release ref */ - mutex_unlock(&phba->ct_event_mutex); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); set_event_out: /* set_event carries no reply payload */ @@ -759,6 +764,7 @@ lpfc_bsg_get_event(struct fc_bsg_job *job) struct get_ct_event_reply *event_reply; struct lpfc_ct_event *evt; struct event_data *evt_dat = NULL; + unsigned long flags; int rc = 0; if (job->request_len < @@ -775,7 +781,7 @@ lpfc_bsg_get_event(struct fc_bsg_job *job) event_reply = (struct get_ct_event_reply *) job->reply->reply_data.vendor_reply.vendor_rsp; - mutex_lock(&phba->ct_event_mutex); + spin_lock_irqsave(&phba->ct_ev_lock, flags); list_for_each_entry(evt, &phba->ct_ev_waiters, node) { if (evt->reg_id == event_req->ev_reg_id) { if (list_empty(&evt->events_to_get)) @@ -788,7 +794,7 @@ lpfc_bsg_get_event(struct fc_bsg_job *job) break; } } - mutex_unlock(&phba->ct_event_mutex); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); if (!evt_dat) { job->reply->reply_payload_rcv_len = 0; @@ -818,9 +824,9 @@ lpfc_bsg_get_event(struct fc_bsg_job *job) if (evt_dat) kfree(evt_dat->data); kfree(evt_dat); - mutex_lock(&phba->ct_event_mutex); + spin_lock_irqsave(&phba->ct_ev_lock, flags); lpfc_ct_event_unref(evt); - mutex_unlock(&phba->ct_event_mutex); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); error_get_event_exit: /* make error code available to userspace */ diff --git a/drivers/scsi/lpfc/lpfc_crtn.h b/drivers/scsi/lpfc/lpfc_crtn.h index b9e8cd5..107899d 100644 --- a/drivers/scsi/lpfc/lpfc_crtn.h +++ b/drivers/scsi/lpfc/lpfc_crtn.h @@ -1,7 +1,7 @@ /******************************************************************* * This file is part of the Emulex Linux Device Driver for * * Fibre Channel Host Bus Adapters. * - * Copyright (C) 2004-2008 Emulex. All rights reserved. * + * Copyright (C) 2004-2010 Emulex. All rights reserved. * * EMULEX and SLI are trademarks of Emulex. * * www.emulex.com * * * @@ -383,5 +383,5 @@ struct lpfc_vport *lpfc_find_vport_by_vpid(struct lpfc_hba *, uint16_t); /* functions to support SGIOv4/bsg interface */ int lpfc_bsg_request(struct fc_bsg_job *); int lpfc_bsg_timeout(struct fc_bsg_job *); -void lpfc_bsg_ct_unsol_event(struct lpfc_hba *, struct lpfc_sli_ring *, +int lpfc_bsg_ct_unsol_event(struct lpfc_hba *, struct lpfc_sli_ring *, struct lpfc_iocbq *); diff --git a/drivers/scsi/lpfc/lpfc_ct.c b/drivers/scsi/lpfc/lpfc_ct.c index bf7bf62..c7e9219 100644 --- a/drivers/scsi/lpfc/lpfc_ct.c +++ b/drivers/scsi/lpfc/lpfc_ct.c @@ -1,7 +1,7 @@ /******************************************************************* * This file is part of the Emulex Linux Device Driver for * * Fibre Channel Host Bus Adapters. * - * Copyright (C) 2004-2009 Emulex. All rights reserved. * + * Copyright (C) 2004-2010 Emulex. All rights reserved. * * EMULEX and SLI are trademarks of Emulex. * * www.emulex.com * * * @@ -97,7 +97,8 @@ lpfc_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, struct list_head head; struct lpfc_dmabuf *bdeBuf; - lpfc_bsg_ct_unsol_event(phba, pring, piocbq); + if (lpfc_bsg_ct_unsol_event(phba, pring, piocbq) == 0) + return; if (unlikely(icmd->ulpStatus == IOSTAT_NEED_BUFFER)) { lpfc_sli_hbqbuf_add_hbqs(phba, LPFC_ELS_HBQ); @@ -181,7 +182,8 @@ lpfc_sli4_ct_abort_unsol_event(struct lpfc_hba *phba, uint32_t size; /* Forward abort event to any process registered to receive ct event */ - lpfc_bsg_ct_unsol_event(phba, pring, piocbq); + if (lpfc_bsg_ct_unsol_event(phba, pring, piocbq) == 0) + return; /* If there is no BDE associated with IOCB, there is nothing to do */ if (icmd->ulpBdeCount == 0) diff --git a/drivers/scsi/lpfc/lpfc_hw.h b/drivers/scsi/lpfc/lpfc_hw.h index 44c2587..89ff7c0 100644 --- a/drivers/scsi/lpfc/lpfc_hw.h +++ b/drivers/scsi/lpfc/lpfc_hw.h @@ -1,7 +1,7 @@ /******************************************************************* * This file is part of the Emulex Linux Device Driver for * * Fibre Channel Host Bus Adapters. * - * Copyright (C) 2004-2009 Emulex. All rights reserved. * + * Copyright (C) 2004-2010 Emulex. All rights reserved. * * EMULEX and SLI are trademarks of Emulex. * * www.emulex.com * * * @@ -1346,6 +1346,9 @@ typedef struct { /* FireFly BIU registers */ #define MBX_HEARTBEAT 0x31 #define MBX_WRITE_VPARMS 0x32 #define MBX_ASYNCEVT_ENABLE 0x33 +#define MBX_READ_EVENT_LOG_STATUS 0x37 +#define MBX_READ_EVENT_LOG 0x38 +#define MBX_WRITE_EVENT_LOG 0x39 #define MBX_PORT_CAPABILITIES 0x3B #define MBX_PORT_IOV_CONTROL 0x3C diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c index e80e95d..52fc758 100644 --- a/drivers/scsi/lpfc/lpfc_init.c +++ b/drivers/scsi/lpfc/lpfc_init.c @@ -1,7 +1,7 @@ /******************************************************************* * This file is part of the Emulex Linux Device Driver for * * Fibre Channel Host Bus Adapters. * - * Copyright (C) 2004-2009 Emulex. All rights reserved. * + * Copyright (C) 2004-2010 Emulex. All rights reserved. * * EMULEX and SLI are trademarks of Emulex. * * www.emulex.com * * Portions Copyright (C) 2004-2005 Christoph Hellwig * @@ -4331,7 +4331,7 @@ lpfc_hba_alloc(struct pci_dev *pdev) return NULL; } - mutex_init(&phba->ct_event_mutex); + spin_lock_init(&phba->ct_ev_lock); INIT_LIST_HEAD(&phba->ct_ev_waiters); return phba; diff --git a/drivers/scsi/lpfc/lpfc_nl.h b/drivers/scsi/lpfc/lpfc_nl.h index d655ed3..f3cfbe2 100644 --- a/drivers/scsi/lpfc/lpfc_nl.h +++ b/drivers/scsi/lpfc/lpfc_nl.h @@ -1,7 +1,7 @@ /******************************************************************* * This file is part of the Emulex Linux Device Driver for * * Fibre Channel Host Bus Adapters. * - * Copyright (C) 2008 Emulex. All rights reserved. * + * Copyright (C) 2010 Emulex. All rights reserved. * * EMULEX and SLI are trademarks of Emulex. * * www.emulex.com * * * @@ -177,23 +177,3 @@ struct temp_event { uint32_t data; }; -/* bsg definitions */ -#define LPFC_BSG_VENDOR_SET_CT_EVENT 1 -#define LPFC_BSG_VENDOR_GET_CT_EVENT 2 - -struct set_ct_event { - uint32_t command; - uint32_t ev_req_id; - uint32_t ev_reg_id; -}; - -struct get_ct_event { - uint32_t command; - uint32_t ev_reg_id; - uint32_t ev_req_id; -}; - -struct get_ct_event_reply { - uint32_t immed_data; - uint32_t type; -}; -- cgit v0.10.2 From c79c1292df87fa9c63383ca551fa719c0c2fda7c Mon Sep 17 00:00:00 2001 From: James Smart Date: Tue, 26 Jan 2010 23:09:22 -0500 Subject: [SCSI] lpfc 8.3.8: (BSG2) Create lpfc_bsg.h Create lpfc_bsg.h - structures, etc used by bsg-related routines. Signed-off-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/lpfc/lpfc_bsg.h b/drivers/scsi/lpfc/lpfc_bsg.h new file mode 100644 index 0000000..6c8f87e --- /dev/null +++ b/drivers/scsi/lpfc/lpfc_bsg.h @@ -0,0 +1,98 @@ +/******************************************************************* + * This file is part of the Emulex Linux Device Driver for * + * Fibre Channel Host Bus Adapters. * + * Copyright (C) 2010 Emulex. All rights reserved. * + * EMULEX and SLI are trademarks of Emulex. * + * www.emulex.com * + * * + * This program is free software; you can redistribute it and/or * + * modify it under the terms of version 2 of the GNU General * + * Public License as published by the Free Software Foundation. * + * This program is distributed in the hope that it will be useful. * + * ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND * + * WARRANTIES, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, * + * FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE * + * DISCLAIMED, EXCEPT TO THE EXTENT THAT SUCH DISCLAIMERS ARE HELD * + * TO BE LEGALLY INVALID. See the GNU General Public License for * + * more details, a copy of which can be found in the file COPYING * + * included with this package. * + *******************************************************************/ +/* bsg definitions + * No pointers to user data are allowed, all application buffers and sizes will + * derived through the bsg interface. + * + * These are the vendor unique structures passed in using the bsg + * FC_BSG_HST_VENDOR message code type. + */ +#define LPFC_BSG_VENDOR_SET_CT_EVENT 1 +#define LPFC_BSG_VENDOR_GET_CT_EVENT 2 +#define LPFC_BSG_VENDOR_SEND_MGMT_RESP 3 +#define LPFC_BSG_VENDOR_DIAG_MODE 4 +#define LPFC_BSG_VENDOR_DIAG_TEST 5 +#define LPFC_BSG_VENDOR_GET_MGMT_REV 6 +#define LPFC_BSG_VENDOR_MBOX 7 + +struct set_ct_event { + uint32_t command; + uint32_t type_mask; + uint32_t ev_req_id; + uint32_t ev_reg_id; +}; + +struct get_ct_event { + uint32_t command; + uint32_t ev_reg_id; + uint32_t ev_req_id; +}; + +struct get_ct_event_reply { + uint32_t immed_data; + uint32_t type; +}; + +struct send_mgmt_resp { + uint32_t command; + uint32_t tag; +}; + + +#define INTERNAL_LOOP_BACK 0x1 /* adapter short cuts the loop internally */ +#define EXTERNAL_LOOP_BACK 0x2 /* requires an external loopback plug */ + +struct diag_mode_set { + uint32_t command; + uint32_t type; + uint32_t timeout; +}; + +struct diag_mode_test { + uint32_t command; +}; + +#define LPFC_WWNN_TYPE 0 +#define LPFC_WWPN_TYPE 1 + +struct get_mgmt_rev { + uint32_t command; +}; + +#define MANAGEMENT_MAJOR_REV 1 +#define MANAGEMENT_MINOR_REV 0 + +/* the MgmtRevInfo structure */ +struct MgmtRevInfo { + uint32_t a_Major; + uint32_t a_Minor; +}; + +struct get_mgmt_rev_reply { + struct MgmtRevInfo info; +}; + +struct dfc_mbox_req { + uint32_t command; + uint32_t inExtWLen; + uint32_t outExtWLen; + uint8_t mbOffset; +}; + -- cgit v0.10.2 From 4cc0e56e977f12e6f400cbab3df7cf1e11d6f58a Mon Sep 17 00:00:00 2001 From: James Smart Date: Tue, 26 Jan 2010 23:09:48 -0500 Subject: [SCSI] lpfc 8.3.8: (BSG3) Modify BSG commands to operate asynchronously Modify the following BSG commands to operate asynchronously. - FC_BSG_RPT_ELS - FC_BSG_RPT_CT - LPFC_BSG_VENDOR_GET_CT_EVENT - LPFC_BSG_VENDOR_SET_CT_EVENT Signed-off-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/lpfc/lpfc_bsg.c b/drivers/scsi/lpfc/lpfc_bsg.c index dfb1f73..a7e8921 100644 --- a/drivers/scsi/lpfc/lpfc_bsg.c +++ b/drivers/scsi/lpfc/lpfc_bsg.c @@ -42,14 +42,169 @@ #include "lpfc_vport.h" #include "lpfc_version.h" +struct lpfc_bsg_event { + struct list_head node; + struct kref kref; + wait_queue_head_t wq; + + /* Event type and waiter identifiers */ + uint32_t type_mask; + uint32_t req_id; + uint32_t reg_id; + + /* next two flags are here for the auto-delete logic */ + unsigned long wait_time_stamp; + int waiting; + + /* seen and not seen events */ + struct list_head events_to_get; + struct list_head events_to_see; + + /* job waiting for this event to finish */ + struct fc_bsg_job *set_job; +}; + +struct lpfc_bsg_iocb { + struct lpfc_iocbq *cmdiocbq; + struct lpfc_iocbq *rspiocbq; + struct lpfc_dmabuf *bmp; + struct lpfc_nodelist *ndlp; + + /* job waiting for this iocb to finish */ + struct fc_bsg_job *set_job; +}; + +#define TYPE_EVT 1 +#define TYPE_IOCB 2 +struct bsg_job_data { + uint32_t type; + union { + struct lpfc_bsg_event *evt; + struct lpfc_bsg_iocb iocb; + } context_un; +}; + +struct event_data { + struct list_head node; + uint32_t type; + uint32_t immed_dat; + void *data; + uint32_t len; +}; + +#define SLI_CT_ELX_LOOPBACK 0x10 + +enum ELX_LOOPBACK_CMD { + ELX_LOOPBACK_XRI_SETUP, + ELX_LOOPBACK_DATA, +}; + +struct lpfc_dmabufext { + struct lpfc_dmabuf dma; + uint32_t size; + uint32_t flag; +}; + +/** + * lpfc_bsg_send_mgmt_cmd_cmp - lpfc_bsg_send_mgmt_cmd's completion handler + * @phba: Pointer to HBA context object. + * @cmdiocbq: Pointer to command iocb. + * @rspiocbq: Pointer to response iocb. + * + * This function is the completion handler for iocbs issued using + * lpfc_bsg_send_mgmt_cmd function. This function is called by the + * ring event handler function without any lock held. This function + * can be called from both worker thread context and interrupt + * context. This function also can be called from another thread which + * cleans up the SLI layer objects. + * This function copies the contents of the response iocb to the + * response iocb memory object provided by the caller of + * lpfc_sli_issue_iocb_wait and then wakes up the thread which + * sleeps for the iocb completion. + **/ +static void +lpfc_bsg_send_mgmt_cmd_cmp(struct lpfc_hba *phba, + struct lpfc_iocbq *cmdiocbq, + struct lpfc_iocbq *rspiocbq) +{ + unsigned long iflags; + struct bsg_job_data *dd_data; + struct fc_bsg_job *job; + IOCB_t *rsp; + struct lpfc_dmabuf *bmp; + struct lpfc_nodelist *ndlp; + struct lpfc_bsg_iocb *iocb; + unsigned long flags; + int rc = 0; + + spin_lock_irqsave(&phba->ct_ev_lock, flags); + dd_data = cmdiocbq->context1; + if (!dd_data) { + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + return; + } + + iocb = &dd_data->context_un.iocb; + job = iocb->set_job; + job->dd_data = NULL; /* so timeout handler does not reply */ + + spin_lock_irqsave(&phba->hbalock, iflags); + cmdiocbq->iocb_flag |= LPFC_IO_WAKE; + if (cmdiocbq->context2 && rspiocbq) + memcpy(&((struct lpfc_iocbq *)cmdiocbq->context2)->iocb, + &rspiocbq->iocb, sizeof(IOCB_t)); + spin_unlock_irqrestore(&phba->hbalock, iflags); + + bmp = iocb->bmp; + rspiocbq = iocb->rspiocbq; + rsp = &rspiocbq->iocb; + ndlp = iocb->ndlp; + + pci_unmap_sg(phba->pcidev, job->request_payload.sg_list, + job->request_payload.sg_cnt, DMA_TO_DEVICE); + pci_unmap_sg(phba->pcidev, job->reply_payload.sg_list, + job->reply_payload.sg_cnt, DMA_FROM_DEVICE); + + if (rsp->ulpStatus) { + if (rsp->ulpStatus == IOSTAT_LOCAL_REJECT) { + switch (rsp->un.ulpWord[4] & 0xff) { + case IOERR_SEQUENCE_TIMEOUT: + rc = -ETIMEDOUT; + break; + case IOERR_INVALID_RPI: + rc = -EFAULT; + break; + default: + rc = -EACCES; + break; + } + } else + rc = -EACCES; + } else + job->reply->reply_payload_rcv_len = + rsp->un.genreq64.bdl.bdeSize; + + lpfc_mbuf_free(phba, bmp->virt, bmp->phys); + lpfc_sli_release_iocbq(phba, rspiocbq); + lpfc_sli_release_iocbq(phba, cmdiocbq); + lpfc_nlp_put(ndlp); + kfree(bmp); + kfree(dd_data); + /* make error code available to userspace */ + job->reply->result = rc; + /* complete the job back to userspace */ + job->job_done(job); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + return; +} + /** - * lpfc_bsg_rport_ct - send a CT command from a bsg request + * lpfc_bsg_send_mgmt_cmd - send a CT command from a bsg request * @job: fc_bsg_job to handle */ static int -lpfc_bsg_rport_ct(struct fc_bsg_job *job) +lpfc_bsg_send_mgmt_cmd(struct fc_bsg_job *job) { - struct Scsi_Host *shost = job->shost; struct lpfc_vport *vport = (struct lpfc_vport *)job->shost->hostdata; struct lpfc_hba *phba = vport->phba; struct lpfc_rport_data *rdata = job->rport->dd_data; @@ -66,57 +221,60 @@ lpfc_bsg_rport_ct(struct fc_bsg_job *job) struct scatterlist *sgel = NULL; int numbde; dma_addr_t busaddr; + struct bsg_job_data *dd_data; + uint32_t creg_val; int rc = 0; /* in case no data is transferred */ job->reply->reply_payload_rcv_len = 0; + /* allocate our bsg tracking structure */ + dd_data = kmalloc(sizeof(struct bsg_job_data), GFP_KERNEL); + if (!dd_data) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2733 Failed allocation of dd_data\n"); + rc = -ENOMEM; + goto no_dd_data; + } + if (!lpfc_nlp_get(ndlp)) { - job->reply->result = -ENODEV; - return 0; + rc = -ENODEV; + goto no_ndlp; + } + + bmp = kmalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL); + if (!bmp) { + rc = -ENOMEM; + goto free_ndlp; } if (ndlp->nlp_flag & NLP_ELS_SND_MASK) { rc = -ENODEV; - goto free_ndlp_exit; + goto free_bmp; } - spin_lock_irq(shost->host_lock); cmdiocbq = lpfc_sli_get_iocbq(phba); if (!cmdiocbq) { rc = -ENOMEM; - spin_unlock_irq(shost->host_lock); - goto free_ndlp_exit; + goto free_bmp; } - cmd = &cmdiocbq->iocb; + cmd = &cmdiocbq->iocb; rspiocbq = lpfc_sli_get_iocbq(phba); if (!rspiocbq) { rc = -ENOMEM; goto free_cmdiocbq; } - spin_unlock_irq(shost->host_lock); rsp = &rspiocbq->iocb; - - bmp = kmalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL); - if (!bmp) { - rc = -ENOMEM; - spin_lock_irq(shost->host_lock); - goto free_rspiocbq; - } - - spin_lock_irq(shost->host_lock); bmp->virt = lpfc_mbuf_alloc(phba, 0, &bmp->phys); if (!bmp->virt) { rc = -ENOMEM; - goto free_bmp; + goto free_rspiocbq; } - spin_unlock_irq(shost->host_lock); INIT_LIST_HEAD(&bmp->list); bpl = (struct ulp_bde64 *) bmp->virt; - request_nseg = pci_map_sg(phba->pcidev, job->request_payload.sg_list, job->request_payload.sg_cnt, DMA_TO_DEVICE); for_each_sg(job->request_payload.sg_list, sgel, request_nseg, numbde) { @@ -158,72 +316,146 @@ lpfc_bsg_rport_ct(struct fc_bsg_job *job) cmd->ulpContext = ndlp->nlp_rpi; cmd->ulpOwner = OWN_CHIP; cmdiocbq->vport = phba->pport; - cmdiocbq->context1 = NULL; - cmdiocbq->context2 = NULL; + cmdiocbq->context3 = bmp; cmdiocbq->iocb_flag |= LPFC_IO_LIBDFC; - timeout = phba->fc_ratov * 2; - job->dd_data = cmdiocbq; - - rc = lpfc_sli_issue_iocb_wait(phba, LPFC_ELS_RING, cmdiocbq, rspiocbq, - timeout + LPFC_DRVR_TIMEOUT); - - if (rc != IOCB_TIMEDOUT) { - pci_unmap_sg(phba->pcidev, job->request_payload.sg_list, - job->request_payload.sg_cnt, DMA_TO_DEVICE); - pci_unmap_sg(phba->pcidev, job->reply_payload.sg_list, - job->reply_payload.sg_cnt, DMA_FROM_DEVICE); + cmd->ulpTimeout = timeout; + + cmdiocbq->iocb_cmpl = lpfc_bsg_send_mgmt_cmd_cmp; + cmdiocbq->context1 = dd_data; + cmdiocbq->context2 = rspiocbq; + dd_data->type = TYPE_IOCB; + dd_data->context_un.iocb.cmdiocbq = cmdiocbq; + dd_data->context_un.iocb.rspiocbq = rspiocbq; + dd_data->context_un.iocb.set_job = job; + dd_data->context_un.iocb.bmp = bmp; + dd_data->context_un.iocb.ndlp = ndlp; + + if (phba->cfg_poll & DISABLE_FCP_RING_INT) { + creg_val = readl(phba->HCregaddr); + creg_val |= (HC_R0INT_ENA << LPFC_FCP_RING); + writel(creg_val, phba->HCregaddr); + readl(phba->HCregaddr); /* flush */ } - if (rc == IOCB_TIMEDOUT) { - lpfc_sli_release_iocbq(phba, rspiocbq); - rc = -EACCES; - goto free_ndlp_exit; - } + rc = lpfc_sli_issue_iocb(phba, LPFC_ELS_RING, cmdiocbq, 0); - if (rc != IOCB_SUCCESS) { - rc = -EACCES; - goto free_outdmp; - } + if (rc == IOCB_SUCCESS) + return 0; /* done for now */ - if (rsp->ulpStatus) { - if (rsp->ulpStatus == IOSTAT_LOCAL_REJECT) { - switch (rsp->un.ulpWord[4] & 0xff) { - case IOERR_SEQUENCE_TIMEOUT: - rc = -ETIMEDOUT; - break; - case IOERR_INVALID_RPI: - rc = -EFAULT; - break; - default: - rc = -EACCES; - break; - } - goto free_outdmp; - } - } else - job->reply->reply_payload_rcv_len = - rsp->un.genreq64.bdl.bdeSize; + /* iocb failed so cleanup */ + pci_unmap_sg(phba->pcidev, job->request_payload.sg_list, + job->request_payload.sg_cnt, DMA_TO_DEVICE); + pci_unmap_sg(phba->pcidev, job->reply_payload.sg_list, + job->reply_payload.sg_cnt, DMA_FROM_DEVICE); -free_outdmp: - spin_lock_irq(shost->host_lock); lpfc_mbuf_free(phba, bmp->virt, bmp->phys); -free_bmp: - kfree(bmp); + free_rspiocbq: lpfc_sli_release_iocbq(phba, rspiocbq); free_cmdiocbq: lpfc_sli_release_iocbq(phba, cmdiocbq); - spin_unlock_irq(shost->host_lock); -free_ndlp_exit: +free_bmp: + kfree(bmp); +free_ndlp: lpfc_nlp_put(ndlp); +no_ndlp: + kfree(dd_data); +no_dd_data: + /* make error code available to userspace */ + job->reply->result = rc; + job->dd_data = NULL; + return rc; +} + +/** + * lpfc_bsg_rport_els_cmp - lpfc_bsg_rport_els's completion handler + * @phba: Pointer to HBA context object. + * @cmdiocbq: Pointer to command iocb. + * @rspiocbq: Pointer to response iocb. + * + * This function is the completion handler for iocbs issued using + * lpfc_bsg_rport_els_cmp function. This function is called by the + * ring event handler function without any lock held. This function + * can be called from both worker thread context and interrupt + * context. This function also can be called from other thread which + * cleans up the SLI layer objects. + * This function copy the contents of the response iocb to the + * response iocb memory object provided by the caller of + * lpfc_sli_issue_iocb_wait and then wakes up the thread which + * sleeps for the iocb completion. + **/ +static void +lpfc_bsg_rport_els_cmp(struct lpfc_hba *phba, + struct lpfc_iocbq *cmdiocbq, + struct lpfc_iocbq *rspiocbq) +{ + struct bsg_job_data *dd_data; + struct fc_bsg_job *job; + IOCB_t *rsp; + struct lpfc_nodelist *ndlp; + struct lpfc_dmabuf *pbuflist = NULL; + struct fc_bsg_ctels_reply *els_reply; + uint8_t *rjt_data; + unsigned long flags; + int rc = 0; + + spin_lock_irqsave(&phba->ct_ev_lock, flags); + dd_data = cmdiocbq->context1; + /* normal completion and timeout crossed paths, already done */ + if (!dd_data) { + spin_unlock_irqrestore(&phba->hbalock, flags); + return; + } + + cmdiocbq->iocb_flag |= LPFC_IO_WAKE; + if (cmdiocbq->context2 && rspiocbq) + memcpy(&((struct lpfc_iocbq *)cmdiocbq->context2)->iocb, + &rspiocbq->iocb, sizeof(IOCB_t)); + + job = dd_data->context_un.iocb.set_job; + cmdiocbq = dd_data->context_un.iocb.cmdiocbq; + rspiocbq = dd_data->context_un.iocb.rspiocbq; + rsp = &rspiocbq->iocb; + ndlp = dd_data->context_un.iocb.ndlp; + + pci_unmap_sg(phba->pcidev, job->request_payload.sg_list, + job->request_payload.sg_cnt, DMA_TO_DEVICE); + pci_unmap_sg(phba->pcidev, job->reply_payload.sg_list, + job->reply_payload.sg_cnt, DMA_FROM_DEVICE); + + if (job->reply->result == -EAGAIN) + rc = -EAGAIN; + else if (rsp->ulpStatus == IOSTAT_SUCCESS) + job->reply->reply_payload_rcv_len = + rsp->un.elsreq64.bdl.bdeSize; + else if (rsp->ulpStatus == IOSTAT_LS_RJT) { + job->reply->reply_payload_rcv_len = + sizeof(struct fc_bsg_ctels_reply); + /* LS_RJT data returned in word 4 */ + rjt_data = (uint8_t *)&rsp->un.ulpWord[4]; + els_reply = &job->reply->reply_data.ctels_reply; + els_reply->status = FC_CTELS_STATUS_REJECT; + els_reply->rjt_data.action = rjt_data[3]; + els_reply->rjt_data.reason_code = rjt_data[2]; + els_reply->rjt_data.reason_explanation = rjt_data[1]; + els_reply->rjt_data.vendor_unique = rjt_data[0]; + } else + rc = -EIO; + pbuflist = (struct lpfc_dmabuf *) cmdiocbq->context3; + lpfc_mbuf_free(phba, pbuflist->virt, pbuflist->phys); + lpfc_sli_release_iocbq(phba, rspiocbq); + lpfc_sli_release_iocbq(phba, cmdiocbq); + lpfc_nlp_put(ndlp); + kfree(dd_data); /* make error code available to userspace */ job->reply->result = rc; + job->dd_data = NULL; /* complete the job back to userspace */ job->job_done(job); - - return 0; + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + return; } /** @@ -237,7 +469,6 @@ lpfc_bsg_rport_els(struct fc_bsg_job *job) struct lpfc_hba *phba = vport->phba; struct lpfc_rport_data *rdata = job->rport->dd_data; struct lpfc_nodelist *ndlp = rdata->pnode; - uint32_t elscmd; uint32_t cmdsize; uint32_t rspsize; @@ -249,20 +480,30 @@ lpfc_bsg_rport_els(struct fc_bsg_job *job) struct lpfc_dmabuf *prsp; struct lpfc_dmabuf *pbuflist = NULL; struct ulp_bde64 *bpl; - int iocb_status; int request_nseg; int reply_nseg; struct scatterlist *sgel = NULL; int numbde; dma_addr_t busaddr; + struct bsg_job_data *dd_data; + uint32_t creg_val; int rc = 0; /* in case no data is transferred */ job->reply->reply_payload_rcv_len = 0; + /* allocate our bsg tracking structure */ + dd_data = kmalloc(sizeof(struct bsg_job_data), GFP_KERNEL); + if (!dd_data) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2735 Failed allocation of dd_data\n"); + rc = -ENOMEM; + goto no_dd_data; + } + if (!lpfc_nlp_get(ndlp)) { rc = -ENODEV; - goto out; + goto free_dd_data; } elscmd = job->request->rqst_data.r_els.els_code; @@ -272,24 +513,24 @@ lpfc_bsg_rport_els(struct fc_bsg_job *job) if (!rspiocbq) { lpfc_nlp_put(ndlp); rc = -ENOMEM; - goto out; + goto free_dd_data; } rsp = &rspiocbq->iocb; rpi = ndlp->nlp_rpi; - cmdiocbq = lpfc_prep_els_iocb(phba->pport, 1, cmdsize, 0, ndlp, + cmdiocbq = lpfc_prep_els_iocb(vport, 1, cmdsize, 0, ndlp, ndlp->nlp_DID, elscmd); - if (!cmdiocbq) { - lpfc_sli_release_iocbq(phba, rspiocbq); - return -EIO; + rc = -EIO; + goto free_rspiocbq; } - job->dd_data = cmdiocbq; + /* prep els iocb set context1 to the ndlp, context2 to the command + * dmabuf, context3 holds the data dmabuf + */ pcmd = (struct lpfc_dmabuf *) cmdiocbq->context2; prsp = (struct lpfc_dmabuf *) pcmd->list.next; - lpfc_mbuf_free(phba, pcmd->virt, pcmd->phys); kfree(pcmd); lpfc_mbuf_free(phba, prsp->virt, prsp->phys); @@ -301,7 +542,6 @@ lpfc_bsg_rport_els(struct fc_bsg_job *job) request_nseg = pci_map_sg(phba->pcidev, job->request_payload.sg_list, job->request_payload.sg_cnt, DMA_TO_DEVICE); - for_each_sg(job->request_payload.sg_list, sgel, request_nseg, numbde) { busaddr = sg_dma_address(sgel); bpl->tus.f.bdeFlags = BUFF_TYPE_BDE_64; @@ -323,7 +563,6 @@ lpfc_bsg_rport_els(struct fc_bsg_job *job) bpl->addrHigh = cpu_to_le32(putPaddrHigh(busaddr)); bpl++; } - cmdiocbq->iocb.un.elsreq64.bdl.bdeSize = (request_nseg + reply_nseg) * sizeof(struct ulp_bde64); cmdiocbq->iocb.ulpContext = rpi; @@ -331,102 +570,54 @@ lpfc_bsg_rport_els(struct fc_bsg_job *job) cmdiocbq->context1 = NULL; cmdiocbq->context2 = NULL; - iocb_status = lpfc_sli_issue_iocb_wait(phba, LPFC_ELS_RING, cmdiocbq, - rspiocbq, (phba->fc_ratov * 2) - + LPFC_DRVR_TIMEOUT); - - /* release the new ndlp once the iocb completes */ + cmdiocbq->iocb_cmpl = lpfc_bsg_rport_els_cmp; + cmdiocbq->context1 = dd_data; + cmdiocbq->context2 = rspiocbq; + dd_data->type = TYPE_IOCB; + dd_data->context_un.iocb.cmdiocbq = cmdiocbq; + dd_data->context_un.iocb.rspiocbq = rspiocbq; + dd_data->context_un.iocb.set_job = job; + dd_data->context_un.iocb.bmp = NULL;; + dd_data->context_un.iocb.ndlp = ndlp; + + if (phba->cfg_poll & DISABLE_FCP_RING_INT) { + creg_val = readl(phba->HCregaddr); + creg_val |= (HC_R0INT_ENA << LPFC_FCP_RING); + writel(creg_val, phba->HCregaddr); + readl(phba->HCregaddr); /* flush */ + } + rc = lpfc_sli_issue_iocb(phba, LPFC_ELS_RING, cmdiocbq, 0); lpfc_nlp_put(ndlp); - if (iocb_status != IOCB_TIMEDOUT) { - pci_unmap_sg(phba->pcidev, job->request_payload.sg_list, - job->request_payload.sg_cnt, DMA_TO_DEVICE); - pci_unmap_sg(phba->pcidev, job->reply_payload.sg_list, - job->reply_payload.sg_cnt, DMA_FROM_DEVICE); - } - - if (iocb_status == IOCB_SUCCESS) { - if (rsp->ulpStatus == IOSTAT_SUCCESS) { - job->reply->reply_payload_rcv_len = - rsp->un.elsreq64.bdl.bdeSize; - rc = 0; - } else if (rsp->ulpStatus == IOSTAT_LS_RJT) { - struct fc_bsg_ctels_reply *els_reply; - /* LS_RJT data returned in word 4 */ - uint8_t *rjt_data = (uint8_t *)&rsp->un.ulpWord[4]; - - els_reply = &job->reply->reply_data.ctels_reply; - job->reply->result = 0; - els_reply->status = FC_CTELS_STATUS_REJECT; - els_reply->rjt_data.action = rjt_data[0]; - els_reply->rjt_data.reason_code = rjt_data[1]; - els_reply->rjt_data.reason_explanation = rjt_data[2]; - els_reply->rjt_data.vendor_unique = rjt_data[3]; - } else - rc = -EIO; - } else - rc = -EIO; - - if (iocb_status != IOCB_TIMEDOUT) - lpfc_els_free_iocb(phba, cmdiocbq); - - lpfc_sli_release_iocbq(phba, rspiocbq); - -out: - /* make error code available to userspace */ - job->reply->result = rc; - /* complete the job back to userspace */ - job->job_done(job); - - return 0; -} - -struct lpfc_ct_event { - struct list_head node; - int ref; - wait_queue_head_t wq; - - /* Event type and waiter identifiers */ - uint32_t type_mask; - uint32_t req_id; - uint32_t reg_id; + if (rc == IOCB_SUCCESS) + return 0; /* done for now */ - /* next two flags are here for the auto-delete logic */ - unsigned long wait_time_stamp; - int waiting; + pci_unmap_sg(phba->pcidev, job->request_payload.sg_list, + job->request_payload.sg_cnt, DMA_TO_DEVICE); + pci_unmap_sg(phba->pcidev, job->reply_payload.sg_list, + job->reply_payload.sg_cnt, DMA_FROM_DEVICE); - /* seen and not seen events */ - struct list_head events_to_get; - struct list_head events_to_see; -}; + lpfc_mbuf_free(phba, pbuflist->virt, pbuflist->phys); -struct event_data { - struct list_head node; - uint32_t type; - uint32_t immed_dat; - void *data; - uint32_t len; -}; + lpfc_sli_release_iocbq(phba, cmdiocbq); -static struct lpfc_ct_event * -lpfc_ct_event_new(int ev_reg_id, uint32_t ev_req_id) -{ - struct lpfc_ct_event *evt = kzalloc(sizeof(*evt), GFP_KERNEL); - if (!evt) - return NULL; +free_rspiocbq: + lpfc_sli_release_iocbq(phba, rspiocbq); - INIT_LIST_HEAD(&evt->events_to_get); - INIT_LIST_HEAD(&evt->events_to_see); - evt->req_id = ev_req_id; - evt->reg_id = ev_reg_id; - evt->wait_time_stamp = jiffies; - init_waitqueue_head(&evt->wq); +free_dd_data: + kfree(dd_data); - return evt; +no_dd_data: + /* make error code available to userspace */ + job->reply->result = rc; + job->dd_data = NULL; + return rc; } static void -lpfc_ct_event_free(struct lpfc_ct_event *evt) +lpfc_bsg_event_free(struct kref *kref) { + struct lpfc_bsg_event *evt = container_of(kref, struct lpfc_bsg_event, + kref); struct event_data *ed; list_del(&evt->node); @@ -449,24 +640,62 @@ lpfc_ct_event_free(struct lpfc_ct_event *evt) } static inline void -lpfc_ct_event_ref(struct lpfc_ct_event *evt) +lpfc_bsg_event_ref(struct lpfc_bsg_event *evt) { - evt->ref++; + kref_get(&evt->kref); } static inline void -lpfc_ct_event_unref(struct lpfc_ct_event *evt) +lpfc_bsg_event_unref(struct lpfc_bsg_event *evt) { - if (--evt->ref < 0) - lpfc_ct_event_free(evt); + kref_put(&evt->kref, lpfc_bsg_event_free); } -#define SLI_CT_ELX_LOOPBACK 0x10 +static struct lpfc_bsg_event * +lpfc_bsg_event_new(uint32_t ev_mask, int ev_reg_id, uint32_t ev_req_id) +{ + struct lpfc_bsg_event *evt = kzalloc(sizeof(*evt), GFP_KERNEL); -enum ELX_LOOPBACK_CMD { - ELX_LOOPBACK_XRI_SETUP, - ELX_LOOPBACK_DATA, -}; + if (!evt) + return NULL; + + INIT_LIST_HEAD(&evt->events_to_get); + INIT_LIST_HEAD(&evt->events_to_see); + evt->type_mask = ev_mask; + evt->req_id = ev_req_id; + evt->reg_id = ev_reg_id; + evt->wait_time_stamp = jiffies; + init_waitqueue_head(&evt->wq); + kref_init(&evt->kref); + return evt; +} + +static int +dfc_cmd_data_free(struct lpfc_hba *phba, struct lpfc_dmabufext *mlist) +{ + struct lpfc_dmabufext *mlast; + struct pci_dev *pcidev; + struct list_head head, *curr, *next; + + if ((!mlist) || (!lpfc_is_link_up(phba) && + (phba->link_flag & LS_LOOPBACK_MODE))) { + return 0; + } + + pcidev = phba->pcidev; + list_add_tail(&head, &mlist->dma.list); + + list_for_each_safe(curr, next, &head) { + mlast = list_entry(curr, struct lpfc_dmabufext , dma.list); + if (mlast->dma.virt) + dma_free_coherent(&pcidev->dev, + mlast->size, + mlast->dma.virt, + mlast->dma.phys); + kfree(mlast); + } + return 0; +} /** * lpfc_bsg_ct_unsol_event - process an unsolicited CT command @@ -475,7 +704,7 @@ enum ELX_LOOPBACK_CMD { * @piocbq: * * This function is called when an unsolicited CT command is received. It - * forwards the event to any processes registerd to receive CT events. + * forwards the event to any processes registered to receive CT events. */ int lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, @@ -485,7 +714,7 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, uint32_t cmd; uint32_t len; struct lpfc_dmabuf *dmabuf = NULL; - struct lpfc_ct_event *evt; + struct lpfc_bsg_event *evt; struct event_data *evt_dat = NULL; struct lpfc_iocbq *iocbq; size_t offset = 0; @@ -497,7 +726,9 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, struct lpfc_dmabuf *bdeBuf2 = piocbq->context3; struct lpfc_hbq_entry *hbqe; struct lpfc_sli_ct_request *ct_req; + struct fc_bsg_job *job = NULL; unsigned long flags; + int size = 0; INIT_LIST_HEAD(&head); list_add_tail(&head, &piocbq->list); @@ -506,6 +737,10 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, piocbq->iocb.un.cont64[0].tus.f.bdeSize == 0) goto error_ct_unsol_exit; + if (phba->link_state == LPFC_HBA_ERROR || + (!(phba->sli.sli_flag & LPFC_SLI_ACTIVE))) + goto error_ct_unsol_exit; + if (phba->sli3_options & LPFC_SLI3_HBQ_ENABLED) dmabuf = bdeBuf1; else { @@ -513,7 +748,8 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, piocbq->iocb.un.cont64[0].addrLow); dmabuf = lpfc_sli_ringpostbuf_get(phba, pring, dma_addr); } - + if (dmabuf == NULL) + goto error_ct_unsol_exit; ct_req = (struct lpfc_sli_ct_request *)dmabuf->virt; evt_req_id = ct_req->FsType; cmd = ct_req->CommandResponse.bits.CmdRsp; @@ -523,22 +759,22 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, spin_lock_irqsave(&phba->ct_ev_lock, flags); list_for_each_entry(evt, &phba->ct_ev_waiters, node) { - if (evt->req_id != evt_req_id) + if (!(evt->type_mask & FC_REG_CT_EVENT) || + evt->req_id != evt_req_id) continue; - lpfc_ct_event_ref(evt); - + lpfc_bsg_event_ref(evt); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); evt_dat = kzalloc(sizeof(*evt_dat), GFP_KERNEL); - if (!evt_dat) { - lpfc_ct_event_unref(evt); + if (evt_dat == NULL) { + spin_lock_irqsave(&phba->ct_ev_lock, flags); + lpfc_bsg_event_unref(evt); lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, "2614 Memory allocation failed for " "CT event\n"); break; } - spin_unlock_irqrestore(&phba->ct_ev_lock, flags); - if (phba->sli3_options & LPFC_SLI3_HBQ_ENABLED) { /* take accumulated byte count from the last iocbq */ iocbq = list_entry(head.prev, typeof(*iocbq), list); @@ -552,25 +788,25 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, } evt_dat->data = kzalloc(evt_dat->len, GFP_KERNEL); - if (!evt_dat->data) { + if (evt_dat->data == NULL) { lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, "2615 Memory allocation failed for " "CT event data, size %d\n", evt_dat->len); kfree(evt_dat); spin_lock_irqsave(&phba->ct_ev_lock, flags); - lpfc_ct_event_unref(evt); + lpfc_bsg_event_unref(evt); spin_unlock_irqrestore(&phba->ct_ev_lock, flags); goto error_ct_unsol_exit; } list_for_each_entry(iocbq, &head, list) { + size = 0; if (phba->sli3_options & LPFC_SLI3_HBQ_ENABLED) { bdeBuf1 = iocbq->context2; bdeBuf2 = iocbq->context3; } for (i = 0; i < iocbq->iocb.ulpBdeCount; i++) { - int size = 0; if (phba->sli3_options & LPFC_SLI3_HBQ_ENABLED) { if (i == 0) { @@ -605,7 +841,7 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, kfree(evt_dat); spin_lock_irqsave(&phba->ct_ev_lock, flags); - lpfc_ct_event_unref(evt); + lpfc_bsg_event_unref(evt); spin_unlock_irqrestore( &phba->ct_ev_lock, flags); goto error_ct_unsol_exit; @@ -620,15 +856,24 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, dmabuf); } else { switch (cmd) { + case ELX_LOOPBACK_DATA: + dfc_cmd_data_free(phba, + (struct lpfc_dmabufext *) + dmabuf); + break; case ELX_LOOPBACK_XRI_SETUP: - if (!(phba->sli3_options & - LPFC_SLI3_HBQ_ENABLED)) + if ((phba->sli_rev == + LPFC_SLI_REV2) || + (phba->sli3_options & + LPFC_SLI3_HBQ_ENABLED + )) { + lpfc_in_buf_free(phba, + dmabuf); + } else { lpfc_post_buffer(phba, pring, 1); - else - lpfc_in_buf_free(phba, - dmabuf); + } break; default: if (!(phba->sli3_options & @@ -655,49 +900,79 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, evt_dat->type = FC_REG_CT_EVENT; list_add(&evt_dat->node, &evt->events_to_see); - wake_up_interruptible(&evt->wq); - lpfc_ct_event_unref(evt); - if (evt_req_id == SLI_CT_ELX_LOOPBACK) + if (evt_req_id == SLI_CT_ELX_LOOPBACK) { + wake_up_interruptible(&evt->wq); + lpfc_bsg_event_unref(evt); break; + } + + list_move(evt->events_to_see.prev, &evt->events_to_get); + lpfc_bsg_event_unref(evt); + + job = evt->set_job; + evt->set_job = NULL; + if (job) { + job->reply->reply_payload_rcv_len = size; + /* make error code available to userspace */ + job->reply->result = 0; + job->dd_data = NULL; + /* complete the job back to userspace */ + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + job->job_done(job); + spin_lock_irqsave(&phba->ct_ev_lock, flags); + } } spin_unlock_irqrestore(&phba->ct_ev_lock, flags); error_ct_unsol_exit: if (!list_empty(&head)) list_del(&head); - + if (evt_req_id == SLI_CT_ELX_LOOPBACK) + return 0; return 1; } /** - * lpfc_bsg_set_event - process a SET_EVENT bsg vendor command + * lpfc_bsg_hba_set_event - process a SET_EVENT bsg vendor command * @job: SET_EVENT fc_bsg_job */ static int -lpfc_bsg_set_event(struct fc_bsg_job *job) +lpfc_bsg_hba_set_event(struct fc_bsg_job *job) { struct lpfc_vport *vport = (struct lpfc_vport *)job->shost->hostdata; struct lpfc_hba *phba = vport->phba; struct set_ct_event *event_req; - struct lpfc_ct_event *evt; - unsigned long flags; + struct lpfc_bsg_event *evt; int rc = 0; + struct bsg_job_data *dd_data = NULL; + uint32_t ev_mask; + unsigned long flags; if (job->request_len < sizeof(struct fc_bsg_request) + sizeof(struct set_ct_event)) { lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, "2612 Received SET_CT_EVENT below minimum " "size\n"); - return -EINVAL; + rc = -EINVAL; + goto job_error; + } + + dd_data = kmalloc(sizeof(struct bsg_job_data), GFP_KERNEL); + if (dd_data == NULL) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2734 Failed allocation of dd_data\n"); + rc = -ENOMEM; + goto job_error; } event_req = (struct set_ct_event *) job->request->rqst_data.h_vendor.vendor_cmd; - + ev_mask = ((uint32_t)(unsigned long)event_req->type_mask & + FC_REG_EVENT_MASK); spin_lock_irqsave(&phba->ct_ev_lock, flags); list_for_each_entry(evt, &phba->ct_ev_waiters, node) { if (evt->reg_id == event_req->ev_reg_id) { - lpfc_ct_event_ref(evt); + lpfc_bsg_event_ref(evt); evt->wait_time_stamp = jiffies; break; } @@ -706,73 +981,63 @@ lpfc_bsg_set_event(struct fc_bsg_job *job) if (&evt->node == &phba->ct_ev_waiters) { /* no event waiting struct yet - first call */ - evt = lpfc_ct_event_new(event_req->ev_reg_id, + evt = lpfc_bsg_event_new(ev_mask, event_req->ev_reg_id, event_req->ev_req_id); if (!evt) { lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, "2617 Failed allocation of event " "waiter\n"); - return -ENOMEM; + rc = -ENOMEM; + goto job_error; } spin_lock_irqsave(&phba->ct_ev_lock, flags); list_add(&evt->node, &phba->ct_ev_waiters); - lpfc_ct_event_ref(evt); + lpfc_bsg_event_ref(evt); + evt->wait_time_stamp = jiffies; spin_unlock_irqrestore(&phba->ct_ev_lock, flags); } - evt->waiting = 1; - if (wait_event_interruptible(evt->wq, - !list_empty(&evt->events_to_see))) { - spin_lock_irqsave(&phba->ct_ev_lock, flags); - lpfc_ct_event_unref(evt); /* release ref */ - lpfc_ct_event_unref(evt); /* delete */ - spin_unlock_irqrestore(&phba->ct_ev_lock, flags); - rc = -EINTR; - goto set_event_out; - } - - evt->wait_time_stamp = jiffies; - evt->waiting = 0; - spin_lock_irqsave(&phba->ct_ev_lock, flags); - list_move(evt->events_to_see.prev, &evt->events_to_get); - lpfc_ct_event_unref(evt); /* release ref */ + evt->waiting = 1; + dd_data->type = TYPE_EVT; + dd_data->context_un.evt = evt; + evt->set_job = job; /* for unsolicited command */ + job->dd_data = dd_data; /* for fc transport timeout callback*/ spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + return 0; /* call job done later */ -set_event_out: - /* set_event carries no reply payload */ - job->reply->reply_payload_rcv_len = 0; - /* make error code available to userspace */ - job->reply->result = rc; - /* complete the job back to userspace */ - job->job_done(job); +job_error: + if (dd_data != NULL) + kfree(dd_data); - return 0; + job->dd_data = NULL; + return rc; } /** - * lpfc_bsg_get_event - process a GET_EVENT bsg vendor command + * lpfc_bsg_hba_get_event - process a GET_EVENT bsg vendor command * @job: GET_EVENT fc_bsg_job */ static int -lpfc_bsg_get_event(struct fc_bsg_job *job) +lpfc_bsg_hba_get_event(struct fc_bsg_job *job) { struct lpfc_vport *vport = (struct lpfc_vport *)job->shost->hostdata; struct lpfc_hba *phba = vport->phba; struct get_ct_event *event_req; struct get_ct_event_reply *event_reply; - struct lpfc_ct_event *evt; + struct lpfc_bsg_event *evt; struct event_data *evt_dat = NULL; unsigned long flags; - int rc = 0; + uint32_t rc = 0; if (job->request_len < sizeof(struct fc_bsg_request) + sizeof(struct get_ct_event)) { lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, "2613 Received GET_CT_EVENT request below " "minimum size\n"); - return -EINVAL; + rc = -EINVAL; + goto job_error; } event_req = (struct get_ct_event *) @@ -780,13 +1045,12 @@ lpfc_bsg_get_event(struct fc_bsg_job *job) event_reply = (struct get_ct_event_reply *) job->reply->reply_data.vendor_reply.vendor_rsp; - spin_lock_irqsave(&phba->ct_ev_lock, flags); list_for_each_entry(evt, &phba->ct_ev_waiters, node) { if (evt->reg_id == event_req->ev_reg_id) { if (list_empty(&evt->events_to_get)) break; - lpfc_ct_event_ref(evt); + lpfc_bsg_event_ref(evt); evt->wait_time_stamp = jiffies; evt_dat = list_entry(evt->events_to_get.prev, struct event_data, node); @@ -796,44 +1060,49 @@ lpfc_bsg_get_event(struct fc_bsg_job *job) } spin_unlock_irqrestore(&phba->ct_ev_lock, flags); - if (!evt_dat) { + /* The app may continue to ask for event data until it gets + * an error indicating that there isn't anymore + */ + if (evt_dat == NULL) { job->reply->reply_payload_rcv_len = 0; rc = -ENOENT; - goto error_get_event_exit; + goto job_error; } - if (evt_dat->len > job->reply_payload.payload_len) { - evt_dat->len = job->reply_payload.payload_len; - lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, - "2618 Truncated event data at %d " - "bytes\n", - job->reply_payload.payload_len); + if (evt_dat->len > job->request_payload.payload_len) { + evt_dat->len = job->request_payload.payload_len; + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2618 Truncated event data at %d " + "bytes\n", + job->request_payload.payload_len); } + event_reply->type = evt_dat->type; event_reply->immed_data = evt_dat->immed_dat; - if (evt_dat->len > 0) job->reply->reply_payload_rcv_len = - sg_copy_from_buffer(job->reply_payload.sg_list, - job->reply_payload.sg_cnt, + sg_copy_from_buffer(job->request_payload.sg_list, + job->request_payload.sg_cnt, evt_dat->data, evt_dat->len); else job->reply->reply_payload_rcv_len = 0; - rc = 0; - if (evt_dat) + if (evt_dat) { kfree(evt_dat->data); - kfree(evt_dat); + kfree(evt_dat); + } + spin_lock_irqsave(&phba->ct_ev_lock, flags); - lpfc_ct_event_unref(evt); + lpfc_bsg_event_unref(evt); spin_unlock_irqrestore(&phba->ct_ev_lock, flags); - -error_get_event_exit: - /* make error code available to userspace */ - job->reply->result = rc; - /* complete the job back to userspace */ + job->dd_data = NULL; + job->reply->result = 0; job->job_done(job); + return 0; +job_error: + job->dd_data = NULL; + job->reply->result = rc; return rc; } @@ -845,19 +1114,25 @@ static int lpfc_bsg_hst_vendor(struct fc_bsg_job *job) { int command = job->request->rqst_data.h_vendor.vendor_cmd[0]; + int rc; switch (command) { case LPFC_BSG_VENDOR_SET_CT_EVENT: - return lpfc_bsg_set_event(job); + rc = lpfc_bsg_hba_set_event(job); break; case LPFC_BSG_VENDOR_GET_CT_EVENT: - return lpfc_bsg_get_event(job); + rc = lpfc_bsg_hba_get_event(job); break; - default: - return -EINVAL; + rc = -EINVAL; + job->reply->reply_payload_rcv_len = 0; + /* make error code available to userspace */ + job->reply->result = rc; + break; } + + return rc; } /** @@ -868,10 +1143,9 @@ int lpfc_bsg_request(struct fc_bsg_job *job) { uint32_t msgcode; - int rc = -EINVAL; + int rc; msgcode = job->request->msgcode; - switch (msgcode) { case FC_BSG_HST_VENDOR: rc = lpfc_bsg_hst_vendor(job); @@ -880,9 +1154,13 @@ lpfc_bsg_request(struct fc_bsg_job *job) rc = lpfc_bsg_rport_els(job); break; case FC_BSG_RPT_CT: - rc = lpfc_bsg_rport_ct(job); + rc = lpfc_bsg_send_mgmt_cmd(job); break; default: + rc = -EINVAL; + job->reply->reply_payload_rcv_len = 0; + /* make error code available to userspace */ + job->reply->result = rc; break; } @@ -901,11 +1179,54 @@ lpfc_bsg_timeout(struct fc_bsg_job *job) { struct lpfc_vport *vport = (struct lpfc_vport *)job->shost->hostdata; struct lpfc_hba *phba = vport->phba; - struct lpfc_iocbq *cmdiocb = (struct lpfc_iocbq *)job->dd_data; + struct lpfc_iocbq *cmdiocb; + struct lpfc_bsg_event *evt; + struct lpfc_bsg_iocb *iocb; struct lpfc_sli_ring *pring = &phba->sli.ring[LPFC_ELS_RING]; + struct bsg_job_data *dd_data; + unsigned long flags; - if (cmdiocb) + spin_lock_irqsave(&phba->ct_ev_lock, flags); + dd_data = (struct bsg_job_data *)job->dd_data; + /* timeout and completion crossed paths if no dd_data */ + if (!dd_data) { + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + return 0; + } + + switch (dd_data->type) { + case TYPE_IOCB: + iocb = &dd_data->context_un.iocb; + cmdiocb = iocb->cmdiocbq; + /* hint to completion handler that the job timed out */ + job->reply->result = -EAGAIN; + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + /* this will call our completion handler */ + spin_lock_irq(&phba->hbalock); lpfc_sli_issue_abort_iotag(phba, pring, cmdiocb); + spin_unlock_irq(&phba->hbalock); + break; + case TYPE_EVT: + evt = dd_data->context_un.evt; + /* this event has no job anymore */ + evt->set_job = NULL; + job->dd_data = NULL; + job->reply->reply_payload_rcv_len = 0; + /* Return -EAGAIN which is our way of signallying the + * app to retry. + */ + job->reply->result = -EAGAIN; + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + job->job_done(job); + break; + default: + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + break; + } + /* scsi transport fc fc_bsg_job_timeout expects a zero return code, + * otherwise an error message will be displayed on the console + * so always return success (zero) + */ return 0; } -- cgit v0.10.2 From 3b5dd52aaffd291edea9f939ed46a960b240bb45 Mon Sep 17 00:00:00 2001 From: James Smart Date: Tue, 26 Jan 2010 23:10:15 -0500 Subject: [SCSI] lpfc 8.3.8: (BSG4) Add new vendor specific BSG Commands Add the following new vendor specific BSG commands. - Add LPFC_BSG_VENDOR_GET_MGMT_REV command - Add LPFC_BSG_VENDOR_MBOX command - Add LPFC_BSG_VENDOR_DIAG_MODE command - Add LPFC_BSG_VENDOR_DIAG_TEST command Signed-off-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/lpfc/lpfc_bsg.c b/drivers/scsi/lpfc/lpfc_bsg.c index a7e8921..64ab075 100644 --- a/drivers/scsi/lpfc/lpfc_bsg.c +++ b/drivers/scsi/lpfc/lpfc_bsg.c @@ -74,13 +74,23 @@ struct lpfc_bsg_iocb { struct fc_bsg_job *set_job; }; +struct lpfc_bsg_mbox { + LPFC_MBOXQ_t *pmboxq; + MAILBOX_t *mb; + + /* job waiting for this mbox command to finish */ + struct fc_bsg_job *set_job; +}; + #define TYPE_EVT 1 #define TYPE_IOCB 2 +#define TYPE_MBOX 3 struct bsg_job_data { uint32_t type; union { struct lpfc_bsg_event *evt; struct lpfc_bsg_iocb iocb; + struct lpfc_bsg_mbox mbox; } context_un; }; @@ -92,6 +102,7 @@ struct event_data { uint32_t len; }; +#define BUF_SZ_4K 4096 #define SLI_CT_ELX_LOOPBACK 0x10 enum ELX_LOOPBACK_CMD { @@ -99,6 +110,9 @@ enum ELX_LOOPBACK_CMD { ELX_LOOPBACK_DATA, }; +#define ELX_LOOPBACK_HEADER_SZ \ + (size_t)(&((struct lpfc_sli_ct_request *)NULL)->un) + struct lpfc_dmabufext { struct lpfc_dmabuf dma; uint32_t size; @@ -201,7 +215,7 @@ lpfc_bsg_send_mgmt_cmd_cmp(struct lpfc_hba *phba, /** * lpfc_bsg_send_mgmt_cmd - send a CT command from a bsg request * @job: fc_bsg_job to handle - */ + **/ static int lpfc_bsg_send_mgmt_cmd(struct fc_bsg_job *job) { @@ -380,7 +394,7 @@ no_dd_data: * can be called from both worker thread context and interrupt * context. This function also can be called from other thread which * cleans up the SLI layer objects. - * This function copy the contents of the response iocb to the + * This function copies the contents of the response iocb to the * response iocb memory object provided by the caller of * lpfc_sli_issue_iocb_wait and then wakes up the thread which * sleeps for the iocb completion. @@ -461,7 +475,7 @@ lpfc_bsg_rport_els_cmp(struct lpfc_hba *phba, /** * lpfc_bsg_rport_els - send an ELS command from a bsg request * @job: fc_bsg_job to handle - */ + **/ static int lpfc_bsg_rport_els(struct fc_bsg_job *job) { @@ -527,8 +541,8 @@ lpfc_bsg_rport_els(struct fc_bsg_job *job) } /* prep els iocb set context1 to the ndlp, context2 to the command - * dmabuf, context3 holds the data dmabuf - */ + * dmabuf, context3 holds the data dmabuf + */ pcmd = (struct lpfc_dmabuf *) cmdiocbq->context2; prsp = (struct lpfc_dmabuf *) pcmd->list.next; lpfc_mbuf_free(phba, pcmd->virt, pcmd->phys); @@ -613,6 +627,14 @@ no_dd_data: return rc; } +/** + * lpfc_bsg_event_free - frees an allocated event structure + * @kref: Pointer to a kref. + * + * Called from kref_put. Back cast the kref into an event structure address. + * Free any events to get, delete associated nodes, free any events to see, + * free any data then free the event itself. + **/ static void lpfc_bsg_event_free(struct kref *kref) { @@ -639,18 +661,32 @@ lpfc_bsg_event_free(struct kref *kref) kfree(evt); } +/** + * lpfc_bsg_event_ref - increments the kref for an event + * @evt: Pointer to an event structure. + **/ static inline void lpfc_bsg_event_ref(struct lpfc_bsg_event *evt) { kref_get(&evt->kref); } +/** + * lpfc_bsg_event_unref - Uses kref_put to free an event structure + * @evt: Pointer to an event structure. + **/ static inline void lpfc_bsg_event_unref(struct lpfc_bsg_event *evt) { kref_put(&evt->kref, lpfc_bsg_event_free); } +/** + * lpfc_bsg_event_new - allocate and initialize a event structure + * @ev_mask: Mask of events. + * @ev_reg_id: Event reg id. + * @ev_req_id: Event request id. + **/ static struct lpfc_bsg_event * lpfc_bsg_event_new(uint32_t ev_mask, int ev_reg_id, uint32_t ev_req_id) { @@ -670,8 +706,13 @@ lpfc_bsg_event_new(uint32_t ev_mask, int ev_reg_id, uint32_t ev_req_id) return evt; } +/** + * diag_cmd_data_free - Frees an lpfc dma buffer extension + * @phba: Pointer to HBA context object. + * @mlist: Pointer to an lpfc dma buffer extension. + **/ static int -dfc_cmd_data_free(struct lpfc_hba *phba, struct lpfc_dmabufext *mlist) +diag_cmd_data_free(struct lpfc_hba *phba, struct lpfc_dmabufext *mlist) { struct lpfc_dmabufext *mlast; struct pci_dev *pcidev; @@ -705,7 +746,7 @@ dfc_cmd_data_free(struct lpfc_hba *phba, struct lpfc_dmabufext *mlist) * * This function is called when an unsolicited CT command is received. It * forwards the event to any processes registered to receive CT events. - */ + **/ int lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, struct lpfc_iocbq *piocbq) @@ -857,7 +898,7 @@ lpfc_bsg_ct_unsol_event(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, } else { switch (cmd) { case ELX_LOOPBACK_DATA: - dfc_cmd_data_free(phba, + diag_cmd_data_free(phba, (struct lpfc_dmabufext *) dmabuf); break; @@ -935,7 +976,7 @@ error_ct_unsol_exit: /** * lpfc_bsg_hba_set_event - process a SET_EVENT bsg vendor command * @job: SET_EVENT fc_bsg_job - */ + **/ static int lpfc_bsg_hba_set_event(struct fc_bsg_job *job) { @@ -1018,7 +1059,7 @@ job_error: /** * lpfc_bsg_hba_get_event - process a GET_EVENT bsg vendor command * @job: GET_EVENT fc_bsg_job - */ + **/ static int lpfc_bsg_hba_get_event(struct fc_bsg_job *job) { @@ -1107,91 +1148,1595 @@ job_error: } /** - * lpfc_bsg_hst_vendor - process a vendor-specific fc_bsg_job - * @job: fc_bsg_job to handle - */ + * lpfc_issue_ct_rsp_cmp - lpfc_issue_ct_rsp's completion handler + * @phba: Pointer to HBA context object. + * @cmdiocbq: Pointer to command iocb. + * @rspiocbq: Pointer to response iocb. + * + * This function is the completion handler for iocbs issued using + * lpfc_issue_ct_rsp_cmp function. This function is called by the + * ring event handler function without any lock held. This function + * can be called from both worker thread context and interrupt + * context. This function also can be called from other thread which + * cleans up the SLI layer objects. + * This function copy the contents of the response iocb to the + * response iocb memory object provided by the caller of + * lpfc_sli_issue_iocb_wait and then wakes up the thread which + * sleeps for the iocb completion. + **/ +static void +lpfc_issue_ct_rsp_cmp(struct lpfc_hba *phba, + struct lpfc_iocbq *cmdiocbq, + struct lpfc_iocbq *rspiocbq) +{ + struct bsg_job_data *dd_data; + struct fc_bsg_job *job; + IOCB_t *rsp; + struct lpfc_dmabuf *bmp; + struct lpfc_nodelist *ndlp; + unsigned long flags; + int rc = 0; + + spin_lock_irqsave(&phba->ct_ev_lock, flags); + dd_data = cmdiocbq->context1; + /* normal completion and timeout crossed paths, already done */ + if (!dd_data) { + spin_unlock_irqrestore(&phba->hbalock, flags); + return; + } + + job = dd_data->context_un.iocb.set_job; + bmp = dd_data->context_un.iocb.bmp; + rsp = &rspiocbq->iocb; + ndlp = dd_data->context_un.iocb.ndlp; + + pci_unmap_sg(phba->pcidev, job->request_payload.sg_list, + job->request_payload.sg_cnt, DMA_TO_DEVICE); + + if (rsp->ulpStatus) { + if (rsp->ulpStatus == IOSTAT_LOCAL_REJECT) { + switch (rsp->un.ulpWord[4] & 0xff) { + case IOERR_SEQUENCE_TIMEOUT: + rc = -ETIMEDOUT; + break; + case IOERR_INVALID_RPI: + rc = -EFAULT; + break; + default: + rc = -EACCES; + break; + } + } else + rc = -EACCES; + } else + job->reply->reply_payload_rcv_len = + rsp->un.genreq64.bdl.bdeSize; + + lpfc_mbuf_free(phba, bmp->virt, bmp->phys); + lpfc_sli_release_iocbq(phba, cmdiocbq); + lpfc_nlp_put(ndlp); + kfree(bmp); + kfree(dd_data); + /* make error code available to userspace */ + job->reply->result = rc; + job->dd_data = NULL; + /* complete the job back to userspace */ + job->job_done(job); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + return; +} + +/** + * lpfc_issue_ct_rsp - issue a ct response + * @phba: Pointer to HBA context object. + * @job: Pointer to the job object. + * @tag: tag index value into the ports context exchange array. + * @bmp: Pointer to a dma buffer descriptor. + * @num_entry: Number of enties in the bde. + **/ static int -lpfc_bsg_hst_vendor(struct fc_bsg_job *job) +lpfc_issue_ct_rsp(struct lpfc_hba *phba, struct fc_bsg_job *job, uint32_t tag, + struct lpfc_dmabuf *bmp, int num_entry) { - int command = job->request->rqst_data.h_vendor.vendor_cmd[0]; - int rc; + IOCB_t *icmd; + struct lpfc_iocbq *ctiocb = NULL; + int rc = 0; + struct lpfc_nodelist *ndlp = NULL; + struct bsg_job_data *dd_data; + uint32_t creg_val; - switch (command) { - case LPFC_BSG_VENDOR_SET_CT_EVENT: - rc = lpfc_bsg_hba_set_event(job); - break; + /* allocate our bsg tracking structure */ + dd_data = kmalloc(sizeof(struct bsg_job_data), GFP_KERNEL); + if (!dd_data) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2736 Failed allocation of dd_data\n"); + rc = -ENOMEM; + goto no_dd_data; + } - case LPFC_BSG_VENDOR_GET_CT_EVENT: - rc = lpfc_bsg_hba_get_event(job); - break; - default: - rc = -EINVAL; - job->reply->reply_payload_rcv_len = 0; - /* make error code available to userspace */ - job->reply->result = rc; - break; + /* Allocate buffer for command iocb */ + ctiocb = lpfc_sli_get_iocbq(phba); + if (!ctiocb) { + rc = ENOMEM; + goto no_ctiocb; + } + + icmd = &ctiocb->iocb; + icmd->un.xseq64.bdl.ulpIoTag32 = 0; + icmd->un.xseq64.bdl.addrHigh = putPaddrHigh(bmp->phys); + icmd->un.xseq64.bdl.addrLow = putPaddrLow(bmp->phys); + icmd->un.xseq64.bdl.bdeFlags = BUFF_TYPE_BLP_64; + icmd->un.xseq64.bdl.bdeSize = (num_entry * sizeof(struct ulp_bde64)); + icmd->un.xseq64.w5.hcsw.Fctl = (LS | LA); + icmd->un.xseq64.w5.hcsw.Dfctl = 0; + icmd->un.xseq64.w5.hcsw.Rctl = FC_RCTL_DD_SOL_CTL; + icmd->un.xseq64.w5.hcsw.Type = FC_TYPE_CT; + + /* Fill in rest of iocb */ + icmd->ulpCommand = CMD_XMIT_SEQUENCE64_CX; + icmd->ulpBdeCount = 1; + icmd->ulpLe = 1; + icmd->ulpClass = CLASS3; + if (phba->sli_rev == LPFC_SLI_REV4) { + /* Do not issue unsol response if oxid not marked as valid */ + if (!(phba->ct_ctx[tag].flags & UNSOL_VALID)) { + rc = IOCB_ERROR; + goto issue_ct_rsp_exit; + } + icmd->ulpContext = phba->ct_ctx[tag].oxid; + ndlp = lpfc_findnode_did(phba->pport, phba->ct_ctx[tag].SID); + if (!ndlp) { + lpfc_printf_log(phba, KERN_WARNING, LOG_ELS, + "2721 ndlp null for oxid %x SID %x\n", + icmd->ulpContext, + phba->ct_ctx[tag].SID); + rc = IOCB_ERROR; + goto issue_ct_rsp_exit; + } + icmd->un.ulpWord[3] = ndlp->nlp_rpi; + /* The exchange is done, mark the entry as invalid */ + phba->ct_ctx[tag].flags &= ~UNSOL_VALID; + } else + icmd->ulpContext = (ushort) tag; + + icmd->ulpTimeout = phba->fc_ratov * 2; + + /* Xmit CT response on exchange */ + lpfc_printf_log(phba, KERN_INFO, LOG_ELS, + "2722 Xmit CT response on exchange x%x Data: x%x x%x\n", + icmd->ulpContext, icmd->ulpIoTag, phba->link_state); + + ctiocb->iocb_cmpl = NULL; + ctiocb->iocb_flag |= LPFC_IO_LIBDFC; + ctiocb->vport = phba->pport; + ctiocb->context3 = bmp; + + ctiocb->iocb_cmpl = lpfc_issue_ct_rsp_cmp; + ctiocb->context1 = dd_data; + ctiocb->context2 = NULL; + dd_data->type = TYPE_IOCB; + dd_data->context_un.iocb.cmdiocbq = ctiocb; + dd_data->context_un.iocb.rspiocbq = NULL; + dd_data->context_un.iocb.set_job = job; + dd_data->context_un.iocb.bmp = bmp; + dd_data->context_un.iocb.ndlp = ndlp; + + if (phba->cfg_poll & DISABLE_FCP_RING_INT) { + creg_val = readl(phba->HCregaddr); + creg_val |= (HC_R0INT_ENA << LPFC_FCP_RING); + writel(creg_val, phba->HCregaddr); + readl(phba->HCregaddr); /* flush */ } + rc = lpfc_sli_issue_iocb(phba, LPFC_ELS_RING, ctiocb, 0); + + if (rc == IOCB_SUCCESS) + return 0; /* done for now */ + +issue_ct_rsp_exit: + lpfc_sli_release_iocbq(phba, ctiocb); +no_ctiocb: + kfree(dd_data); +no_dd_data: return rc; } /** - * lpfc_bsg_request - handle a bsg request from the FC transport - * @job: fc_bsg_job to handle - */ -int -lpfc_bsg_request(struct fc_bsg_job *job) + * lpfc_bsg_send_mgmt_rsp - process a SEND_MGMT_RESP bsg vendor command + * @job: SEND_MGMT_RESP fc_bsg_job + **/ +static int +lpfc_bsg_send_mgmt_rsp(struct fc_bsg_job *job) { - uint32_t msgcode; - int rc; + struct lpfc_vport *vport = (struct lpfc_vport *)job->shost->hostdata; + struct lpfc_hba *phba = vport->phba; + struct send_mgmt_resp *mgmt_resp = (struct send_mgmt_resp *) + job->request->rqst_data.h_vendor.vendor_cmd; + struct ulp_bde64 *bpl; + struct lpfc_dmabuf *bmp = NULL; + struct scatterlist *sgel = NULL; + int request_nseg; + int numbde; + dma_addr_t busaddr; + uint32_t tag = mgmt_resp->tag; + unsigned long reqbfrcnt = + (unsigned long)job->request_payload.payload_len; + int rc = 0; - msgcode = job->request->msgcode; - switch (msgcode) { - case FC_BSG_HST_VENDOR: - rc = lpfc_bsg_hst_vendor(job); - break; - case FC_BSG_RPT_ELS: - rc = lpfc_bsg_rport_els(job); - break; - case FC_BSG_RPT_CT: - rc = lpfc_bsg_send_mgmt_cmd(job); - break; - default: - rc = -EINVAL; - job->reply->reply_payload_rcv_len = 0; - /* make error code available to userspace */ - job->reply->result = rc; - break; + /* in case no data is transferred */ + job->reply->reply_payload_rcv_len = 0; + + if (!reqbfrcnt || (reqbfrcnt > (80 * BUF_SZ_4K))) { + rc = -ERANGE; + goto send_mgmt_rsp_exit; + } + + bmp = kmalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL); + if (!bmp) { + rc = -ENOMEM; + goto send_mgmt_rsp_exit; + } + + bmp->virt = lpfc_mbuf_alloc(phba, 0, &bmp->phys); + if (!bmp->virt) { + rc = -ENOMEM; + goto send_mgmt_rsp_free_bmp; + } + + INIT_LIST_HEAD(&bmp->list); + bpl = (struct ulp_bde64 *) bmp->virt; + request_nseg = pci_map_sg(phba->pcidev, job->request_payload.sg_list, + job->request_payload.sg_cnt, DMA_TO_DEVICE); + for_each_sg(job->request_payload.sg_list, sgel, request_nseg, numbde) { + busaddr = sg_dma_address(sgel); + bpl->tus.f.bdeFlags = BUFF_TYPE_BDE_64; + bpl->tus.f.bdeSize = sg_dma_len(sgel); + bpl->tus.w = cpu_to_le32(bpl->tus.w); + bpl->addrLow = cpu_to_le32(putPaddrLow(busaddr)); + bpl->addrHigh = cpu_to_le32(putPaddrHigh(busaddr)); + bpl++; } + rc = lpfc_issue_ct_rsp(phba, job, tag, bmp, request_nseg); + + if (rc == IOCB_SUCCESS) + return 0; /* done for now */ + + /* TBD need to handle a timeout */ + pci_unmap_sg(phba->pcidev, job->request_payload.sg_list, + job->request_payload.sg_cnt, DMA_TO_DEVICE); + rc = -EACCES; + lpfc_mbuf_free(phba, bmp->virt, bmp->phys); + +send_mgmt_rsp_free_bmp: + kfree(bmp); +send_mgmt_rsp_exit: + /* make error code available to userspace */ + job->reply->result = rc; + job->dd_data = NULL; return rc; } /** - * lpfc_bsg_timeout - handle timeout of a bsg request from the FC transport - * @job: fc_bsg_job that has timed out + * lpfc_bsg_diag_mode - process a LPFC_BSG_VENDOR_DIAG_MODE bsg vendor command + * @job: LPFC_BSG_VENDOR_DIAG_MODE * - * This function just aborts the job's IOCB. The aborted IOCB will return to - * the waiting function which will handle passing the error back to userspace + * This function is responsible for placing a port into diagnostic loopback + * mode in order to perform a diagnostic loopback test. + * All new scsi requests are blocked, a small delay is used to allow the + * scsi requests to complete then the link is brought down. If the link is + * is placed in loopback mode then scsi requests are again allowed + * so the scsi mid-layer doesn't give up on the port. + * All of this is done in-line. */ -int -lpfc_bsg_timeout(struct fc_bsg_job *job) +static int +lpfc_bsg_diag_mode(struct fc_bsg_job *job) { + struct Scsi_Host *shost = job->shost; struct lpfc_vport *vport = (struct lpfc_vport *)job->shost->hostdata; struct lpfc_hba *phba = vport->phba; - struct lpfc_iocbq *cmdiocb; - struct lpfc_bsg_event *evt; - struct lpfc_bsg_iocb *iocb; - struct lpfc_sli_ring *pring = &phba->sli.ring[LPFC_ELS_RING]; - struct bsg_job_data *dd_data; - unsigned long flags; + struct diag_mode_set *loopback_mode; + struct lpfc_sli *psli = &phba->sli; + struct lpfc_sli_ring *pring = &psli->ring[LPFC_FCP_RING]; + uint32_t link_flags; + uint32_t timeout; + struct lpfc_vport **vports; + LPFC_MBOXQ_t *pmboxq; + int mbxstatus; + int i = 0; + int rc = 0; - spin_lock_irqsave(&phba->ct_ev_lock, flags); - dd_data = (struct bsg_job_data *)job->dd_data; - /* timeout and completion crossed paths if no dd_data */ - if (!dd_data) { - spin_unlock_irqrestore(&phba->ct_ev_lock, flags); - return 0; + /* no data to return just the return code */ + job->reply->reply_payload_rcv_len = 0; + + if (job->request_len < + sizeof(struct fc_bsg_request) + sizeof(struct diag_mode_set)) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2738 Received DIAG MODE request below minimum " + "size\n"); + rc = -EINVAL; + goto job_error; + } + + loopback_mode = (struct diag_mode_set *) + job->request->rqst_data.h_vendor.vendor_cmd; + link_flags = loopback_mode->type; + timeout = loopback_mode->timeout; + + if ((phba->link_state == LPFC_HBA_ERROR) || + (psli->sli_flag & LPFC_BLOCK_MGMT_IO) || + (!(psli->sli_flag & LPFC_SLI_ACTIVE))) { + rc = -EACCES; + goto job_error; + } + + pmboxq = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); + if (!pmboxq) { + rc = -ENOMEM; + goto job_error; + } + + vports = lpfc_create_vport_work_array(phba); + if (vports) { + for (i = 0; i <= phba->max_vpi && vports[i] != NULL; i++) { + shost = lpfc_shost_from_vport(vports[i]); + scsi_block_requests(shost); + } + + lpfc_destroy_vport_work_array(phba, vports); + } else { + shost = lpfc_shost_from_vport(phba->pport); + scsi_block_requests(shost); + } + + while (pring->txcmplq_cnt) { + if (i++ > 500) /* wait up to 5 seconds */ + break; + + msleep(10); + } + + memset((void *)pmboxq, 0, sizeof(LPFC_MBOXQ_t)); + pmboxq->u.mb.mbxCommand = MBX_DOWN_LINK; + pmboxq->u.mb.mbxOwner = OWN_HOST; + + mbxstatus = lpfc_sli_issue_mbox_wait(phba, pmboxq, LPFC_MBOX_TMO); + + if ((mbxstatus == MBX_SUCCESS) && (pmboxq->u.mb.mbxStatus == 0)) { + /* wait for link down before proceeding */ + i = 0; + while (phba->link_state != LPFC_LINK_DOWN) { + if (i++ > timeout) { + rc = -ETIMEDOUT; + goto loopback_mode_exit; + } + + msleep(10); + } + + memset((void *)pmboxq, 0, sizeof(LPFC_MBOXQ_t)); + if (link_flags == INTERNAL_LOOP_BACK) + pmboxq->u.mb.un.varInitLnk.link_flags = FLAGS_LOCAL_LB; + else + pmboxq->u.mb.un.varInitLnk.link_flags = + FLAGS_TOPOLOGY_MODE_LOOP; + + pmboxq->u.mb.mbxCommand = MBX_INIT_LINK; + pmboxq->u.mb.mbxOwner = OWN_HOST; + + mbxstatus = lpfc_sli_issue_mbox_wait(phba, pmboxq, + LPFC_MBOX_TMO); + + if ((mbxstatus != MBX_SUCCESS) || (pmboxq->u.mb.mbxStatus)) + rc = -ENODEV; + else { + phba->link_flag |= LS_LOOPBACK_MODE; + /* wait for the link attention interrupt */ + msleep(100); + + i = 0; + while (phba->link_state != LPFC_HBA_READY) { + if (i++ > timeout) { + rc = -ETIMEDOUT; + break; + } + + msleep(10); + } + } + + } else + rc = -ENODEV; + +loopback_mode_exit: + vports = lpfc_create_vport_work_array(phba); + if (vports) { + for (i = 0; i <= phba->max_vpi && vports[i] != NULL; i++) { + shost = lpfc_shost_from_vport(vports[i]); + scsi_unblock_requests(shost); + } + lpfc_destroy_vport_work_array(phba, vports); + } else { + shost = lpfc_shost_from_vport(phba->pport); + scsi_unblock_requests(shost); + } + + /* + * Let SLI layer release mboxq if mbox command completed after timeout. + */ + if (mbxstatus != MBX_TIMEOUT) + mempool_free(pmboxq, phba->mbox_mem_pool); + +job_error: + /* make error code available to userspace */ + job->reply->result = rc; + /* complete the job back to userspace if no error */ + if (rc == 0) + job->job_done(job); + return rc; +} + +/** + * lpfcdiag_loop_self_reg - obtains a remote port login id + * @phba: Pointer to HBA context object + * @rpi: Pointer to a remote port login id + * + * This function obtains a remote port login id so the diag loopback test + * can send and receive its own unsolicited CT command. + **/ +static int lpfcdiag_loop_self_reg(struct lpfc_hba *phba, uint16_t * rpi) +{ + LPFC_MBOXQ_t *mbox; + struct lpfc_dmabuf *dmabuff; + int status; + + mbox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); + if (!mbox) + return ENOMEM; + + status = lpfc_reg_rpi(phba, 0, phba->pport->fc_myDID, + (uint8_t *)&phba->pport->fc_sparam, mbox, 0); + if (status) { + mempool_free(mbox, phba->mbox_mem_pool); + return ENOMEM; + } + + dmabuff = (struct lpfc_dmabuf *) mbox->context1; + mbox->context1 = NULL; + status = lpfc_sli_issue_mbox_wait(phba, mbox, LPFC_MBOX_TMO); + + if ((status != MBX_SUCCESS) || (mbox->u.mb.mbxStatus)) { + lpfc_mbuf_free(phba, dmabuff->virt, dmabuff->phys); + kfree(dmabuff); + if (status != MBX_TIMEOUT) + mempool_free(mbox, phba->mbox_mem_pool); + return ENODEV; + } + + *rpi = mbox->u.mb.un.varWords[0]; + + lpfc_mbuf_free(phba, dmabuff->virt, dmabuff->phys); + kfree(dmabuff); + mempool_free(mbox, phba->mbox_mem_pool); + return 0; +} + +/** + * lpfcdiag_loop_self_unreg - unregs from the rpi + * @phba: Pointer to HBA context object + * @rpi: Remote port login id + * + * This function unregisters the rpi obtained in lpfcdiag_loop_self_reg + **/ +static int lpfcdiag_loop_self_unreg(struct lpfc_hba *phba, uint16_t rpi) +{ + LPFC_MBOXQ_t *mbox; + int status; + + /* Allocate mboxq structure */ + mbox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); + if (mbox == NULL) + return ENOMEM; + + lpfc_unreg_login(phba, 0, rpi, mbox); + status = lpfc_sli_issue_mbox_wait(phba, mbox, LPFC_MBOX_TMO); + + if ((status != MBX_SUCCESS) || (mbox->u.mb.mbxStatus)) { + if (status != MBX_TIMEOUT) + mempool_free(mbox, phba->mbox_mem_pool); + return EIO; + } + + mempool_free(mbox, phba->mbox_mem_pool); + return 0; +} + +/** + * lpfcdiag_loop_get_xri - obtains the transmit and receive ids + * @phba: Pointer to HBA context object + * @rpi: Remote port login id + * @txxri: Pointer to transmit exchange id + * @rxxri: Pointer to response exchabge id + * + * This function obtains the transmit and receive ids required to send + * an unsolicited ct command with a payload. A special lpfc FsType and CmdRsp + * flags are used to the unsolicted response handler is able to process + * the ct command sent on the same port. + **/ +static int lpfcdiag_loop_get_xri(struct lpfc_hba *phba, uint16_t rpi, + uint16_t *txxri, uint16_t * rxxri) +{ + struct lpfc_bsg_event *evt; + struct lpfc_iocbq *cmdiocbq, *rspiocbq; + IOCB_t *cmd, *rsp; + struct lpfc_dmabuf *dmabuf; + struct ulp_bde64 *bpl = NULL; + struct lpfc_sli_ct_request *ctreq = NULL; + int ret_val = 0; + unsigned long flags; + + *txxri = 0; + *rxxri = 0; + evt = lpfc_bsg_event_new(FC_REG_CT_EVENT, current->pid, + SLI_CT_ELX_LOOPBACK); + if (!evt) + return ENOMEM; + + spin_lock_irqsave(&phba->ct_ev_lock, flags); + list_add(&evt->node, &phba->ct_ev_waiters); + lpfc_bsg_event_ref(evt); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + + cmdiocbq = lpfc_sli_get_iocbq(phba); + rspiocbq = lpfc_sli_get_iocbq(phba); + + dmabuf = kmalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL); + if (dmabuf) { + dmabuf->virt = lpfc_mbuf_alloc(phba, 0, &dmabuf->phys); + INIT_LIST_HEAD(&dmabuf->list); + bpl = (struct ulp_bde64 *) dmabuf->virt; + memset(bpl, 0, sizeof(*bpl)); + ctreq = (struct lpfc_sli_ct_request *)(bpl + 1); + bpl->addrHigh = + le32_to_cpu(putPaddrHigh(dmabuf->phys + sizeof(*bpl))); + bpl->addrLow = + le32_to_cpu(putPaddrLow(dmabuf->phys + sizeof(*bpl))); + bpl->tus.f.bdeFlags = 0; + bpl->tus.f.bdeSize = ELX_LOOPBACK_HEADER_SZ; + bpl->tus.w = le32_to_cpu(bpl->tus.w); + } + + if (cmdiocbq == NULL || rspiocbq == NULL || + dmabuf == NULL || bpl == NULL || ctreq == NULL) { + ret_val = ENOMEM; + goto err_get_xri_exit; + } + + cmd = &cmdiocbq->iocb; + rsp = &rspiocbq->iocb; + + memset(ctreq, 0, ELX_LOOPBACK_HEADER_SZ); + + ctreq->RevisionId.bits.Revision = SLI_CT_REVISION; + ctreq->RevisionId.bits.InId = 0; + ctreq->FsType = SLI_CT_ELX_LOOPBACK; + ctreq->FsSubType = 0; + ctreq->CommandResponse.bits.CmdRsp = ELX_LOOPBACK_XRI_SETUP; + ctreq->CommandResponse.bits.Size = 0; + + + cmd->un.xseq64.bdl.addrHigh = putPaddrHigh(dmabuf->phys); + cmd->un.xseq64.bdl.addrLow = putPaddrLow(dmabuf->phys); + cmd->un.xseq64.bdl.bdeFlags = BUFF_TYPE_BLP_64; + cmd->un.xseq64.bdl.bdeSize = sizeof(*bpl); + + cmd->un.xseq64.w5.hcsw.Fctl = LA; + cmd->un.xseq64.w5.hcsw.Dfctl = 0; + cmd->un.xseq64.w5.hcsw.Rctl = FC_RCTL_DD_UNSOL_CTL; + cmd->un.xseq64.w5.hcsw.Type = FC_TYPE_CT; + + cmd->ulpCommand = CMD_XMIT_SEQUENCE64_CR; + cmd->ulpBdeCount = 1; + cmd->ulpLe = 1; + cmd->ulpClass = CLASS3; + cmd->ulpContext = rpi; + + cmdiocbq->iocb_flag |= LPFC_IO_LIBDFC; + cmdiocbq->vport = phba->pport; + + ret_val = lpfc_sli_issue_iocb_wait(phba, LPFC_ELS_RING, cmdiocbq, + rspiocbq, + (phba->fc_ratov * 2) + + LPFC_DRVR_TIMEOUT); + if (ret_val) + goto err_get_xri_exit; + + *txxri = rsp->ulpContext; + + evt->waiting = 1; + evt->wait_time_stamp = jiffies; + ret_val = wait_event_interruptible_timeout( + evt->wq, !list_empty(&evt->events_to_see), + ((phba->fc_ratov * 2) + LPFC_DRVR_TIMEOUT) * HZ); + if (list_empty(&evt->events_to_see)) + ret_val = (ret_val) ? EINTR : ETIMEDOUT; + else { + ret_val = IOCB_SUCCESS; + spin_lock_irqsave(&phba->ct_ev_lock, flags); + list_move(evt->events_to_see.prev, &evt->events_to_get); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + *rxxri = (list_entry(evt->events_to_get.prev, + typeof(struct event_data), + node))->immed_dat; + } + evt->waiting = 0; + +err_get_xri_exit: + spin_lock_irqsave(&phba->ct_ev_lock, flags); + lpfc_bsg_event_unref(evt); /* release ref */ + lpfc_bsg_event_unref(evt); /* delete */ + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + + if (dmabuf) { + if (dmabuf->virt) + lpfc_mbuf_free(phba, dmabuf->virt, dmabuf->phys); + kfree(dmabuf); + } + + if (cmdiocbq && (ret_val != IOCB_TIMEDOUT)) + lpfc_sli_release_iocbq(phba, cmdiocbq); + if (rspiocbq) + lpfc_sli_release_iocbq(phba, rspiocbq); + return ret_val; +} + +/** + * diag_cmd_data_alloc - fills in a bde struct with dma buffers + * @phba: Pointer to HBA context object + * @bpl: Pointer to 64 bit bde structure + * @size: Number of bytes to process + * @nocopydata: Flag to copy user data into the allocated buffer + * + * This function allocates page size buffers and populates an lpfc_dmabufext. + * If allowed the user data pointed to with indataptr is copied into the kernel + * memory. The chained list of page size buffers is returned. + **/ +static struct lpfc_dmabufext * +diag_cmd_data_alloc(struct lpfc_hba *phba, + struct ulp_bde64 *bpl, uint32_t size, + int nocopydata) +{ + struct lpfc_dmabufext *mlist = NULL; + struct lpfc_dmabufext *dmp; + int cnt, offset = 0, i = 0; + struct pci_dev *pcidev; + + pcidev = phba->pcidev; + + while (size) { + /* We get chunks of 4K */ + if (size > BUF_SZ_4K) + cnt = BUF_SZ_4K; + else + cnt = size; + + /* allocate struct lpfc_dmabufext buffer header */ + dmp = kmalloc(sizeof(struct lpfc_dmabufext), GFP_KERNEL); + if (!dmp) + goto out; + + INIT_LIST_HEAD(&dmp->dma.list); + + /* Queue it to a linked list */ + if (mlist) + list_add_tail(&dmp->dma.list, &mlist->dma.list); + else + mlist = dmp; + + /* allocate buffer */ + dmp->dma.virt = dma_alloc_coherent(&pcidev->dev, + cnt, + &(dmp->dma.phys), + GFP_KERNEL); + + if (!dmp->dma.virt) + goto out; + + dmp->size = cnt; + + if (nocopydata) { + bpl->tus.f.bdeFlags = 0; + pci_dma_sync_single_for_device(phba->pcidev, + dmp->dma.phys, LPFC_BPL_SIZE, PCI_DMA_TODEVICE); + + } else { + memset((uint8_t *)dmp->dma.virt, 0, cnt); + bpl->tus.f.bdeFlags = BUFF_TYPE_BDE_64I; + } + + /* build buffer ptr list for IOCB */ + bpl->addrLow = le32_to_cpu(putPaddrLow(dmp->dma.phys)); + bpl->addrHigh = le32_to_cpu(putPaddrHigh(dmp->dma.phys)); + bpl->tus.f.bdeSize = (ushort) cnt; + bpl->tus.w = le32_to_cpu(bpl->tus.w); + bpl++; + + i++; + offset += cnt; + size -= cnt; + } + + mlist->flag = i; + return mlist; +out: + diag_cmd_data_free(phba, mlist); + return NULL; +} + +/** + * lpfcdiag_loop_post_rxbufs - post the receive buffers for an unsol CT cmd + * @phba: Pointer to HBA context object + * @rxxri: Receive exchange id + * @len: Number of data bytes + * + * This function allocates and posts a data buffer of sufficient size to recieve + * an unsolicted CT command. + **/ +static int lpfcdiag_loop_post_rxbufs(struct lpfc_hba *phba, uint16_t rxxri, + size_t len) +{ + struct lpfc_sli *psli = &phba->sli; + struct lpfc_sli_ring *pring = &psli->ring[LPFC_ELS_RING]; + struct lpfc_iocbq *cmdiocbq; + IOCB_t *cmd = NULL; + struct list_head head, *curr, *next; + struct lpfc_dmabuf *rxbmp; + struct lpfc_dmabuf *dmp; + struct lpfc_dmabuf *mp[2] = {NULL, NULL}; + struct ulp_bde64 *rxbpl = NULL; + uint32_t num_bde; + struct lpfc_dmabufext *rxbuffer = NULL; + int ret_val = 0; + int i = 0; + + cmdiocbq = lpfc_sli_get_iocbq(phba); + rxbmp = kmalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL); + if (rxbmp != NULL) { + rxbmp->virt = lpfc_mbuf_alloc(phba, 0, &rxbmp->phys); + INIT_LIST_HEAD(&rxbmp->list); + rxbpl = (struct ulp_bde64 *) rxbmp->virt; + rxbuffer = diag_cmd_data_alloc(phba, rxbpl, len, 0); + } + + if (!cmdiocbq || !rxbmp || !rxbpl || !rxbuffer) { + ret_val = ENOMEM; + goto err_post_rxbufs_exit; + } + + /* Queue buffers for the receive exchange */ + num_bde = (uint32_t)rxbuffer->flag; + dmp = &rxbuffer->dma; + + cmd = &cmdiocbq->iocb; + i = 0; + + INIT_LIST_HEAD(&head); + list_add_tail(&head, &dmp->list); + list_for_each_safe(curr, next, &head) { + mp[i] = list_entry(curr, struct lpfc_dmabuf, list); + list_del(curr); + + if (phba->sli3_options & LPFC_SLI3_HBQ_ENABLED) { + mp[i]->buffer_tag = lpfc_sli_get_buffer_tag(phba); + cmd->un.quexri64cx.buff.bde.addrHigh = + putPaddrHigh(mp[i]->phys); + cmd->un.quexri64cx.buff.bde.addrLow = + putPaddrLow(mp[i]->phys); + cmd->un.quexri64cx.buff.bde.tus.f.bdeSize = + ((struct lpfc_dmabufext *)mp[i])->size; + cmd->un.quexri64cx.buff.buffer_tag = mp[i]->buffer_tag; + cmd->ulpCommand = CMD_QUE_XRI64_CX; + cmd->ulpPU = 0; + cmd->ulpLe = 1; + cmd->ulpBdeCount = 1; + cmd->unsli3.que_xri64cx_ext_words.ebde_count = 0; + + } else { + cmd->un.cont64[i].addrHigh = putPaddrHigh(mp[i]->phys); + cmd->un.cont64[i].addrLow = putPaddrLow(mp[i]->phys); + cmd->un.cont64[i].tus.f.bdeSize = + ((struct lpfc_dmabufext *)mp[i])->size; + cmd->ulpBdeCount = ++i; + + if ((--num_bde > 0) && (i < 2)) + continue; + + cmd->ulpCommand = CMD_QUE_XRI_BUF64_CX; + cmd->ulpLe = 1; + } + + cmd->ulpClass = CLASS3; + cmd->ulpContext = rxxri; + + ret_val = lpfc_sli_issue_iocb(phba, LPFC_ELS_RING, cmdiocbq, 0); + + if (ret_val == IOCB_ERROR) { + diag_cmd_data_free(phba, + (struct lpfc_dmabufext *)mp[0]); + if (mp[1]) + diag_cmd_data_free(phba, + (struct lpfc_dmabufext *)mp[1]); + dmp = list_entry(next, struct lpfc_dmabuf, list); + ret_val = EIO; + goto err_post_rxbufs_exit; + } + + lpfc_sli_ringpostbuf_put(phba, pring, mp[0]); + if (mp[1]) { + lpfc_sli_ringpostbuf_put(phba, pring, mp[1]); + mp[1] = NULL; + } + + /* The iocb was freed by lpfc_sli_issue_iocb */ + cmdiocbq = lpfc_sli_get_iocbq(phba); + if (!cmdiocbq) { + dmp = list_entry(next, struct lpfc_dmabuf, list); + ret_val = EIO; + goto err_post_rxbufs_exit; + } + + cmd = &cmdiocbq->iocb; + i = 0; + } + list_del(&head); + +err_post_rxbufs_exit: + + if (rxbmp) { + if (rxbmp->virt) + lpfc_mbuf_free(phba, rxbmp->virt, rxbmp->phys); + kfree(rxbmp); + } + + if (cmdiocbq) + lpfc_sli_release_iocbq(phba, cmdiocbq); + return ret_val; +} + +/** + * lpfc_bsg_diag_test - with a port in loopback issues a Ct cmd to itself + * @job: LPFC_BSG_VENDOR_DIAG_TEST fc_bsg_job + * + * This function receives a user data buffer to be transmitted and received on + * the same port, the link must be up and in loopback mode prior + * to being called. + * 1. A kernel buffer is allocated to copy the user data into. + * 2. The port registers with "itself". + * 3. The transmit and receive exchange ids are obtained. + * 4. The receive exchange id is posted. + * 5. A new els loopback event is created. + * 6. The command and response iocbs are allocated. + * 7. The cmd iocb FsType is set to elx loopback and the CmdRsp to looppback. + * + * This function is meant to be called n times while the port is in loopback + * so it is the apps responsibility to issue a reset to take the port out + * of loopback mode. + **/ +static int +lpfc_bsg_diag_test(struct fc_bsg_job *job) +{ + struct lpfc_vport *vport = (struct lpfc_vport *)job->shost->hostdata; + struct lpfc_hba *phba = vport->phba; + struct diag_mode_test *diag_mode; + struct lpfc_bsg_event *evt; + struct event_data *evdat; + struct lpfc_sli *psli = &phba->sli; + uint32_t size; + uint32_t full_size; + size_t segment_len = 0, segment_offset = 0, current_offset = 0; + uint16_t rpi; + struct lpfc_iocbq *cmdiocbq, *rspiocbq; + IOCB_t *cmd, *rsp; + struct lpfc_sli_ct_request *ctreq; + struct lpfc_dmabuf *txbmp; + struct ulp_bde64 *txbpl = NULL; + struct lpfc_dmabufext *txbuffer = NULL; + struct list_head head; + struct lpfc_dmabuf *curr; + uint16_t txxri, rxxri; + uint32_t num_bde; + uint8_t *ptr = NULL, *rx_databuf = NULL; + int rc = 0; + unsigned long flags; + void *dataout = NULL; + uint32_t total_mem; + + /* in case no data is returned return just the return code */ + job->reply->reply_payload_rcv_len = 0; + + if (job->request_len < + sizeof(struct fc_bsg_request) + sizeof(struct diag_mode_test)) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2739 Received DIAG TEST request below minimum " + "size\n"); + rc = -EINVAL; + goto loopback_test_exit; + } + + if (job->request_payload.payload_len != + job->reply_payload.payload_len) { + rc = -EINVAL; + goto loopback_test_exit; + } + + diag_mode = (struct diag_mode_test *) + job->request->rqst_data.h_vendor.vendor_cmd; + + if ((phba->link_state == LPFC_HBA_ERROR) || + (psli->sli_flag & LPFC_BLOCK_MGMT_IO) || + (!(psli->sli_flag & LPFC_SLI_ACTIVE))) { + rc = -EACCES; + goto loopback_test_exit; + } + + if (!lpfc_is_link_up(phba) || !(phba->link_flag & LS_LOOPBACK_MODE)) { + rc = -EACCES; + goto loopback_test_exit; + } + + size = job->request_payload.payload_len; + full_size = size + ELX_LOOPBACK_HEADER_SZ; /* plus the header */ + + if ((size == 0) || (size > 80 * BUF_SZ_4K)) { + rc = -ERANGE; + goto loopback_test_exit; + } + + if (size >= BUF_SZ_4K) { + /* + * Allocate memory for ioctl data. If buffer is bigger than 64k, + * then we allocate 64k and re-use that buffer over and over to + * xfer the whole block. This is because Linux kernel has a + * problem allocating more than 120k of kernel space memory. Saw + * problem with GET_FCPTARGETMAPPING... + */ + if (size <= (64 * 1024)) + total_mem = size; + else + total_mem = 64 * 1024; + } else + /* Allocate memory for ioctl data */ + total_mem = BUF_SZ_4K; + + dataout = kmalloc(total_mem, GFP_KERNEL); + if (dataout == NULL) { + rc = -ENOMEM; + goto loopback_test_exit; + } + + ptr = dataout; + ptr += ELX_LOOPBACK_HEADER_SZ; + sg_copy_to_buffer(job->request_payload.sg_list, + job->request_payload.sg_cnt, + ptr, size); + + rc = lpfcdiag_loop_self_reg(phba, &rpi); + if (rc) { + rc = -ENOMEM; + goto loopback_test_exit; + } + + rc = lpfcdiag_loop_get_xri(phba, rpi, &txxri, &rxxri); + if (rc) { + lpfcdiag_loop_self_unreg(phba, rpi); + rc = -ENOMEM; + goto loopback_test_exit; + } + + rc = lpfcdiag_loop_post_rxbufs(phba, rxxri, full_size); + if (rc) { + lpfcdiag_loop_self_unreg(phba, rpi); + rc = -ENOMEM; + goto loopback_test_exit; + } + + evt = lpfc_bsg_event_new(FC_REG_CT_EVENT, current->pid, + SLI_CT_ELX_LOOPBACK); + if (!evt) { + lpfcdiag_loop_self_unreg(phba, rpi); + rc = -ENOMEM; + goto loopback_test_exit; + } + + spin_lock_irqsave(&phba->ct_ev_lock, flags); + list_add(&evt->node, &phba->ct_ev_waiters); + lpfc_bsg_event_ref(evt); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + + cmdiocbq = lpfc_sli_get_iocbq(phba); + rspiocbq = lpfc_sli_get_iocbq(phba); + txbmp = kmalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL); + + if (txbmp) { + txbmp->virt = lpfc_mbuf_alloc(phba, 0, &txbmp->phys); + INIT_LIST_HEAD(&txbmp->list); + txbpl = (struct ulp_bde64 *) txbmp->virt; + if (txbpl) + txbuffer = diag_cmd_data_alloc(phba, + txbpl, full_size, 0); + } + + if (!cmdiocbq || !rspiocbq || !txbmp || !txbpl || !txbuffer) { + rc = -ENOMEM; + goto err_loopback_test_exit; + } + + cmd = &cmdiocbq->iocb; + rsp = &rspiocbq->iocb; + + INIT_LIST_HEAD(&head); + list_add_tail(&head, &txbuffer->dma.list); + list_for_each_entry(curr, &head, list) { + segment_len = ((struct lpfc_dmabufext *)curr)->size; + if (current_offset == 0) { + ctreq = curr->virt; + memset(ctreq, 0, ELX_LOOPBACK_HEADER_SZ); + ctreq->RevisionId.bits.Revision = SLI_CT_REVISION; + ctreq->RevisionId.bits.InId = 0; + ctreq->FsType = SLI_CT_ELX_LOOPBACK; + ctreq->FsSubType = 0; + ctreq->CommandResponse.bits.CmdRsp = ELX_LOOPBACK_DATA; + ctreq->CommandResponse.bits.Size = size; + segment_offset = ELX_LOOPBACK_HEADER_SZ; + } else + segment_offset = 0; + + BUG_ON(segment_offset >= segment_len); + memcpy(curr->virt + segment_offset, + ptr + current_offset, + segment_len - segment_offset); + + current_offset += segment_len - segment_offset; + BUG_ON(current_offset > size); + } + list_del(&head); + + /* Build the XMIT_SEQUENCE iocb */ + + num_bde = (uint32_t)txbuffer->flag; + + cmd->un.xseq64.bdl.addrHigh = putPaddrHigh(txbmp->phys); + cmd->un.xseq64.bdl.addrLow = putPaddrLow(txbmp->phys); + cmd->un.xseq64.bdl.bdeFlags = BUFF_TYPE_BLP_64; + cmd->un.xseq64.bdl.bdeSize = (num_bde * sizeof(struct ulp_bde64)); + + cmd->un.xseq64.w5.hcsw.Fctl = (LS | LA); + cmd->un.xseq64.w5.hcsw.Dfctl = 0; + cmd->un.xseq64.w5.hcsw.Rctl = FC_RCTL_DD_UNSOL_CTL; + cmd->un.xseq64.w5.hcsw.Type = FC_TYPE_CT; + + cmd->ulpCommand = CMD_XMIT_SEQUENCE64_CX; + cmd->ulpBdeCount = 1; + cmd->ulpLe = 1; + cmd->ulpClass = CLASS3; + cmd->ulpContext = txxri; + + cmdiocbq->iocb_flag |= LPFC_IO_LIBDFC; + cmdiocbq->vport = phba->pport; + + rc = lpfc_sli_issue_iocb_wait(phba, LPFC_ELS_RING, cmdiocbq, rspiocbq, + (phba->fc_ratov * 2) + LPFC_DRVR_TIMEOUT); + + if ((rc != IOCB_SUCCESS) || (rsp->ulpStatus != IOCB_SUCCESS)) { + rc = -EIO; + goto err_loopback_test_exit; + } + + evt->waiting = 1; + rc = wait_event_interruptible_timeout( + evt->wq, !list_empty(&evt->events_to_see), + ((phba->fc_ratov * 2) + LPFC_DRVR_TIMEOUT) * HZ); + evt->waiting = 0; + if (list_empty(&evt->events_to_see)) + rc = (rc) ? -EINTR : -ETIMEDOUT; + else { + spin_lock_irqsave(&phba->ct_ev_lock, flags); + list_move(evt->events_to_see.prev, &evt->events_to_get); + evdat = list_entry(evt->events_to_get.prev, + typeof(*evdat), node); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + rx_databuf = evdat->data; + if (evdat->len != full_size) { + lpfc_printf_log(phba, KERN_ERR, LOG_LIBDFC, + "1603 Loopback test did not receive expected " + "data length. actual length 0x%x expected " + "length 0x%x\n", + evdat->len, full_size); + rc = -EIO; + } else if (rx_databuf == NULL) + rc = -EIO; + else { + rc = IOCB_SUCCESS; + /* skip over elx loopback header */ + rx_databuf += ELX_LOOPBACK_HEADER_SZ; + job->reply->reply_payload_rcv_len = + sg_copy_from_buffer(job->reply_payload.sg_list, + job->reply_payload.sg_cnt, + rx_databuf, size); + job->reply->reply_payload_rcv_len = size; + } + } + +err_loopback_test_exit: + lpfcdiag_loop_self_unreg(phba, rpi); + + spin_lock_irqsave(&phba->ct_ev_lock, flags); + lpfc_bsg_event_unref(evt); /* release ref */ + lpfc_bsg_event_unref(evt); /* delete */ + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + + if (cmdiocbq != NULL) + lpfc_sli_release_iocbq(phba, cmdiocbq); + + if (rspiocbq != NULL) + lpfc_sli_release_iocbq(phba, rspiocbq); + + if (txbmp != NULL) { + if (txbpl != NULL) { + if (txbuffer != NULL) + diag_cmd_data_free(phba, txbuffer); + lpfc_mbuf_free(phba, txbmp->virt, txbmp->phys); + } + kfree(txbmp); + } + +loopback_test_exit: + kfree(dataout); + /* make error code available to userspace */ + job->reply->result = rc; + job->dd_data = NULL; + /* complete the job back to userspace if no error */ + if (rc == 0) + job->job_done(job); + return rc; +} + +/** + * lpfc_bsg_get_dfc_rev - process a GET_DFC_REV bsg vendor command + * @job: GET_DFC_REV fc_bsg_job + **/ +static int +lpfc_bsg_get_dfc_rev(struct fc_bsg_job *job) +{ + struct lpfc_vport *vport = (struct lpfc_vport *)job->shost->hostdata; + struct lpfc_hba *phba = vport->phba; + struct get_mgmt_rev *event_req; + struct get_mgmt_rev_reply *event_reply; + int rc = 0; + + if (job->request_len < + sizeof(struct fc_bsg_request) + sizeof(struct get_mgmt_rev)) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2740 Received GET_DFC_REV request below " + "minimum size\n"); + rc = -EINVAL; + goto job_error; + } + + event_req = (struct get_mgmt_rev *) + job->request->rqst_data.h_vendor.vendor_cmd; + + event_reply = (struct get_mgmt_rev_reply *) + job->reply->reply_data.vendor_reply.vendor_rsp; + + if (job->reply_len < + sizeof(struct fc_bsg_request) + sizeof(struct get_mgmt_rev_reply)) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2741 Received GET_DFC_REV reply below " + "minimum size\n"); + rc = -EINVAL; + goto job_error; + } + + event_reply->info.a_Major = MANAGEMENT_MAJOR_REV; + event_reply->info.a_Minor = MANAGEMENT_MINOR_REV; +job_error: + job->reply->result = rc; + if (rc == 0) + job->job_done(job); + return rc; +} + +/** + * lpfc_bsg_wake_mbox_wait - lpfc_bsg_issue_mbox mbox completion handler + * @phba: Pointer to HBA context object. + * @pmboxq: Pointer to mailbox command. + * + * This is completion handler function for mailbox commands issued from + * lpfc_bsg_issue_mbox function. This function is called by the + * mailbox event handler function with no lock held. This function + * will wake up thread waiting on the wait queue pointed by context1 + * of the mailbox. + **/ +void +lpfc_bsg_wake_mbox_wait(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmboxq) +{ + struct bsg_job_data *dd_data; + MAILBOX_t *pmb; + MAILBOX_t *mb; + struct fc_bsg_job *job; + uint32_t size; + unsigned long flags; + + spin_lock_irqsave(&phba->ct_ev_lock, flags); + dd_data = pmboxq->context1; + if (!dd_data) { + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + return; + } + + pmb = &dd_data->context_un.mbox.pmboxq->u.mb; + mb = dd_data->context_un.mbox.mb; + job = dd_data->context_un.mbox.set_job; + memcpy(mb, pmb, sizeof(*pmb)); + size = job->request_payload.payload_len; + job->reply->reply_payload_rcv_len = + sg_copy_from_buffer(job->reply_payload.sg_list, + job->reply_payload.sg_cnt, + mb, size); + job->reply->result = 0; + dd_data->context_un.mbox.set_job = NULL; + job->dd_data = NULL; + job->job_done(job); + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + mempool_free(dd_data->context_un.mbox.pmboxq, phba->mbox_mem_pool); + kfree(mb); + kfree(dd_data); + return; +} + +/** + * lpfc_bsg_check_cmd_access - test for a supported mailbox command + * @phba: Pointer to HBA context object. + * @mb: Pointer to a mailbox object. + * @vport: Pointer to a vport object. + * + * Some commands require the port to be offline, some may not be called from + * the application. + **/ +static int lpfc_bsg_check_cmd_access(struct lpfc_hba *phba, + MAILBOX_t *mb, struct lpfc_vport *vport) +{ + /* return negative error values for bsg job */ + switch (mb->mbxCommand) { + /* Offline only */ + case MBX_INIT_LINK: + case MBX_DOWN_LINK: + case MBX_CONFIG_LINK: + case MBX_CONFIG_RING: + case MBX_RESET_RING: + case MBX_UNREG_LOGIN: + case MBX_CLEAR_LA: + case MBX_DUMP_CONTEXT: + case MBX_RUN_DIAGS: + case MBX_RESTART: + case MBX_SET_MASK: + if (!(vport->fc_flag & FC_OFFLINE_MODE)) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2743 Command 0x%x is illegal in on-line " + "state\n", + mb->mbxCommand); + return -EPERM; + } + case MBX_WRITE_NV: + case MBX_WRITE_VPARMS: + case MBX_LOAD_SM: + case MBX_READ_NV: + case MBX_READ_CONFIG: + case MBX_READ_RCONFIG: + case MBX_READ_STATUS: + case MBX_READ_XRI: + case MBX_READ_REV: + case MBX_READ_LNK_STAT: + case MBX_DUMP_MEMORY: + case MBX_DOWN_LOAD: + case MBX_UPDATE_CFG: + case MBX_KILL_BOARD: + case MBX_LOAD_AREA: + case MBX_LOAD_EXP_ROM: + case MBX_BEACON: + case MBX_DEL_LD_ENTRY: + case MBX_SET_DEBUG: + case MBX_WRITE_WWN: + case MBX_SLI4_CONFIG: + case MBX_READ_EVENT_LOG_STATUS: + case MBX_WRITE_EVENT_LOG: + case MBX_PORT_CAPABILITIES: + case MBX_PORT_IOV_CONTROL: + break; + case MBX_SET_VARIABLE: + case MBX_RUN_BIU_DIAG64: + case MBX_READ_EVENT_LOG: + case MBX_READ_SPARM64: + case MBX_READ_LA: + case MBX_READ_LA64: + case MBX_REG_LOGIN: + case MBX_REG_LOGIN64: + case MBX_CONFIG_PORT: + case MBX_RUN_BIU_DIAG: + default: + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2742 Unknown Command 0x%x\n", + mb->mbxCommand); + return -EPERM; + } + + return 0; /* ok */ +} + +/** + * lpfc_bsg_issue_mbox - issues a mailbox command on behalf of an app + * @phba: Pointer to HBA context object. + * @mb: Pointer to a mailbox object. + * @vport: Pointer to a vport object. + * + * Allocate a tracking object, mailbox command memory, get a mailbox + * from the mailbox pool, copy the caller mailbox command. + * + * If offline and the sli is active we need to poll for the command (port is + * being reset) and com-plete the job, otherwise issue the mailbox command and + * let our completion handler finish the command. + **/ +static uint32_t +lpfc_bsg_issue_mbox(struct lpfc_hba *phba, struct fc_bsg_job *job, + struct lpfc_vport *vport) +{ + LPFC_MBOXQ_t *pmboxq; + MAILBOX_t *pmb; + MAILBOX_t *mb; + struct bsg_job_data *dd_data; + uint32_t size; + int rc = 0; + + /* allocate our bsg tracking structure */ + dd_data = kmalloc(sizeof(struct bsg_job_data), GFP_KERNEL); + if (!dd_data) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2727 Failed allocation of dd_data\n"); + return -ENOMEM; + } + + mb = kzalloc(PAGE_SIZE, GFP_KERNEL); + if (!mb) { + kfree(dd_data); + return -ENOMEM; + } + + pmboxq = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); + if (!pmboxq) { + kfree(dd_data); + kfree(mb); + return -ENOMEM; + } + + size = job->request_payload.payload_len; + job->reply->reply_payload_rcv_len = + sg_copy_to_buffer(job->request_payload.sg_list, + job->request_payload.sg_cnt, + mb, size); + + rc = lpfc_bsg_check_cmd_access(phba, mb, vport); + if (rc != 0) { + kfree(dd_data); + kfree(mb); + mempool_free(pmboxq, phba->mbox_mem_pool); + return rc; /* must be negative */ + } + + memset(pmboxq, 0, sizeof(LPFC_MBOXQ_t)); + pmb = &pmboxq->u.mb; + memcpy(pmb, mb, sizeof(*pmb)); + pmb->mbxOwner = OWN_HOST; + pmboxq->context1 = NULL; + pmboxq->vport = vport; + + if ((vport->fc_flag & FC_OFFLINE_MODE) || + (!(phba->sli.sli_flag & LPFC_SLI_ACTIVE))) { + rc = lpfc_sli_issue_mbox(phba, pmboxq, MBX_POLL); + if (rc != MBX_SUCCESS) { + if (rc != MBX_TIMEOUT) { + kfree(dd_data); + kfree(mb); + mempool_free(pmboxq, phba->mbox_mem_pool); + } + return (rc == MBX_TIMEOUT) ? -ETIME : -ENODEV; + } + + memcpy(mb, pmb, sizeof(*pmb)); + job->reply->reply_payload_rcv_len = + sg_copy_from_buffer(job->reply_payload.sg_list, + job->reply_payload.sg_cnt, + mb, size); + kfree(dd_data); + kfree(mb); + mempool_free(pmboxq, phba->mbox_mem_pool); + /* not waiting mbox already done */ + return 0; + } + + /* setup wake call as IOCB callback */ + pmboxq->mbox_cmpl = lpfc_bsg_wake_mbox_wait; + /* setup context field to pass wait_queue pointer to wake function */ + pmboxq->context1 = dd_data; + dd_data->type = TYPE_MBOX; + dd_data->context_un.mbox.pmboxq = pmboxq; + dd_data->context_un.mbox.mb = mb; + dd_data->context_un.mbox.set_job = job; + job->dd_data = dd_data; + rc = lpfc_sli_issue_mbox(phba, pmboxq, MBX_NOWAIT); + if ((rc != MBX_SUCCESS) && (rc != MBX_BUSY)) { + kfree(dd_data); + kfree(mb); + mempool_free(pmboxq, phba->mbox_mem_pool); + return -EIO; + } + + return 1; +} + +/** + * lpfc_bsg_mbox_cmd - process an fc bsg LPFC_BSG_VENDOR_MBOX command + * @job: MBOX fc_bsg_job for LPFC_BSG_VENDOR_MBOX. + **/ +static int +lpfc_bsg_mbox_cmd(struct fc_bsg_job *job) +{ + struct lpfc_vport *vport = (struct lpfc_vport *)job->shost->hostdata; + struct lpfc_hba *phba = vport->phba; + int rc = 0; + + /* in case no data is transferred */ + job->reply->reply_payload_rcv_len = 0; + if (job->request_len < + sizeof(struct fc_bsg_request) + sizeof(struct dfc_mbox_req)) { + lpfc_printf_log(phba, KERN_WARNING, LOG_LIBDFC, + "2737 Received MBOX_REQ request below " + "minimum size\n"); + rc = -EINVAL; + goto job_error; + } + + if (job->request_payload.payload_len != PAGE_SIZE) { + rc = -EINVAL; + goto job_error; + } + + if (phba->sli.sli_flag & LPFC_BLOCK_MGMT_IO) { + rc = -EAGAIN; + goto job_error; + } + + rc = lpfc_bsg_issue_mbox(phba, job, vport); + +job_error: + if (rc == 0) { + /* job done */ + job->reply->result = 0; + job->dd_data = NULL; + job->job_done(job); + } else if (rc == 1) + /* job submitted, will complete later*/ + rc = 0; /* return zero, no error */ + else { + /* some error occurred */ + job->reply->result = rc; + job->dd_data = NULL; + } + + return rc; +} + +/** + * lpfc_bsg_hst_vendor - process a vendor-specific fc_bsg_job + * @job: fc_bsg_job to handle + **/ +static int +lpfc_bsg_hst_vendor(struct fc_bsg_job *job) +{ + int command = job->request->rqst_data.h_vendor.vendor_cmd[0]; + int rc; + + switch (command) { + case LPFC_BSG_VENDOR_SET_CT_EVENT: + rc = lpfc_bsg_hba_set_event(job); + break; + case LPFC_BSG_VENDOR_GET_CT_EVENT: + rc = lpfc_bsg_hba_get_event(job); + break; + case LPFC_BSG_VENDOR_SEND_MGMT_RESP: + rc = lpfc_bsg_send_mgmt_rsp(job); + break; + case LPFC_BSG_VENDOR_DIAG_MODE: + rc = lpfc_bsg_diag_mode(job); + break; + case LPFC_BSG_VENDOR_DIAG_TEST: + rc = lpfc_bsg_diag_test(job); + break; + case LPFC_BSG_VENDOR_GET_MGMT_REV: + rc = lpfc_bsg_get_dfc_rev(job); + break; + case LPFC_BSG_VENDOR_MBOX: + rc = lpfc_bsg_mbox_cmd(job); + break; + default: + rc = -EINVAL; + job->reply->reply_payload_rcv_len = 0; + /* make error code available to userspace */ + job->reply->result = rc; + break; + } + + return rc; +} + +/** + * lpfc_bsg_request - handle a bsg request from the FC transport + * @job: fc_bsg_job to handle + **/ +int +lpfc_bsg_request(struct fc_bsg_job *job) +{ + uint32_t msgcode; + int rc; + + msgcode = job->request->msgcode; + switch (msgcode) { + case FC_BSG_HST_VENDOR: + rc = lpfc_bsg_hst_vendor(job); + break; + case FC_BSG_RPT_ELS: + rc = lpfc_bsg_rport_els(job); + break; + case FC_BSG_RPT_CT: + rc = lpfc_bsg_send_mgmt_cmd(job); + break; + default: + rc = -EINVAL; + job->reply->reply_payload_rcv_len = 0; + /* make error code available to userspace */ + job->reply->result = rc; + break; + } + + return rc; +} + +/** + * lpfc_bsg_timeout - handle timeout of a bsg request from the FC transport + * @job: fc_bsg_job that has timed out + * + * This function just aborts the job's IOCB. The aborted IOCB will return to + * the waiting function which will handle passing the error back to userspace + **/ +int +lpfc_bsg_timeout(struct fc_bsg_job *job) +{ + struct lpfc_vport *vport = (struct lpfc_vport *)job->shost->hostdata; + struct lpfc_hba *phba = vport->phba; + struct lpfc_iocbq *cmdiocb; + struct lpfc_bsg_event *evt; + struct lpfc_bsg_iocb *iocb; + struct lpfc_bsg_mbox *mbox; + struct lpfc_sli_ring *pring = &phba->sli.ring[LPFC_ELS_RING]; + struct bsg_job_data *dd_data; + unsigned long flags; + + spin_lock_irqsave(&phba->ct_ev_lock, flags); + dd_data = (struct bsg_job_data *)job->dd_data; + /* timeout and completion crossed paths if no dd_data */ + if (!dd_data) { + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + return 0; } switch (dd_data->type) { @@ -1219,6 +2764,16 @@ lpfc_bsg_timeout(struct fc_bsg_job *job) spin_unlock_irqrestore(&phba->ct_ev_lock, flags); job->job_done(job); break; + case TYPE_MBOX: + mbox = &dd_data->context_un.mbox; + /* this mbox has no job anymore */ + mbox->set_job = NULL; + job->dd_data = NULL; + job->reply->reply_payload_rcv_len = 0; + job->reply->result = -EAGAIN; + spin_unlock_irqrestore(&phba->ct_ev_lock, flags); + job->job_done(job); + break; default: spin_unlock_irqrestore(&phba->ct_ev_lock, flags); break; -- cgit v0.10.2 From 7c0616b85ba23d0467bf0b5a6f288f5296eb05ff Mon Sep 17 00:00:00 2001 From: James Smart Date: Tue, 26 Jan 2010 23:10:40 -0500 Subject: [SCSI] lpfc 8.3.8: Update Driver version to 8.3.8 Update Driver version to 8.3.8 Signed-off-by: James Smart Signed-off-by: James Bottomley diff --git a/drivers/scsi/lpfc/lpfc_version.h b/drivers/scsi/lpfc/lpfc_version.h index 792f722..e550d02 100644 --- a/drivers/scsi/lpfc/lpfc_version.h +++ b/drivers/scsi/lpfc/lpfc_version.h @@ -1,7 +1,7 @@ /******************************************************************* * This file is part of the Emulex Linux Device Driver for * * Fibre Channel Host Bus Adapters. * - * Copyright (C) 2004-2009 Emulex. All rights reserved. * + * Copyright (C) 2004-2010 Emulex. All rights reserved. * * EMULEX and SLI are trademarks of Emulex. * * www.emulex.com * * * @@ -18,7 +18,7 @@ * included with this package. * *******************************************************************/ -#define LPFC_DRIVER_VERSION "8.3.7" +#define LPFC_DRIVER_VERSION "8.3.8" #define LPFC_DRIVER_NAME "lpfc" #define LPFC_SP_DRIVER_HANDLER_NAME "lpfc:sp" #define LPFC_FP_DRIVER_HANDLER_NAME "lpfc:fp" -- cgit v0.10.2 From 25ef231de2678690198d95dca949c3caa67a76de Mon Sep 17 00:00:00 2001 From: Will Schmidt Date: Wed, 25 Nov 2009 06:12:09 +0000 Subject: powerpc/pseries: Hypervisor call tracepoints hcall_stats touchup The tb_total and purr_total values reported via the hcall_stats code should be cumulative, rather than being replaced by the latest delta tb or purr value. Tested-by: Will Schmidt Signed-off-by: Will Schmidt Acked-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/platforms/pseries/hvCall_inst.c b/arch/powerpc/platforms/pseries/hvCall_inst.c index 2f58c71..1fefae7 100644 --- a/arch/powerpc/platforms/pseries/hvCall_inst.c +++ b/arch/powerpc/platforms/pseries/hvCall_inst.c @@ -124,8 +124,8 @@ static void probe_hcall_exit(unsigned long opcode, unsigned long retval, h = &__get_cpu_var(hcall_stats)[opcode / 4]; h->num_calls++; - h->tb_total = mftb() - h->tb_start; - h->purr_total = mfspr(SPRN_PURR) - h->purr_start; + h->tb_total += mftb() - h->tb_start; + h->purr_total += mfspr(SPRN_PURR) - h->purr_start; put_cpu_var(hcall_stats); } -- cgit v0.10.2 From 5a2ad98e920dd8ac4cfbc0c1556e930ba2d65427 Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Sun, 31 Jan 2010 10:02:03 +0000 Subject: arch/powerpc: Fix continuation line formats String constants that are continued on subsequent lines with \ are not good. Signed-off-by: Joe Perches Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/nvram_64.c b/arch/powerpc/kernel/nvram_64.c index ad461e7..9cf197f 100644 --- a/arch/powerpc/kernel/nvram_64.c +++ b/arch/powerpc/kernel/nvram_64.c @@ -338,8 +338,8 @@ static int __init nvram_create_os_partition(void) rc = nvram_write_header(new_part); if (rc <= 0) { - printk(KERN_ERR "nvram_create_os_partition: nvram_write_header \ - failed (%d)\n", rc); + printk(KERN_ERR "nvram_create_os_partition: nvram_write_header " + "failed (%d)\n", rc); return rc; } @@ -349,7 +349,7 @@ static int __init nvram_create_os_partition(void) rc = ppc_md.nvram_write((char *)&seq_init, sizeof(seq_init), &tmp_index); if (rc <= 0) { printk(KERN_ERR "nvram_create_os_partition: nvram_write " - "failed (%d)\n", rc); + "failed (%d)\n", rc); return rc; } diff --git a/arch/powerpc/platforms/pseries/hotplug-cpu.c b/arch/powerpc/platforms/pseries/hotplug-cpu.c index 6ea4698..a70de10 100644 --- a/arch/powerpc/platforms/pseries/hotplug-cpu.c +++ b/arch/powerpc/platforms/pseries/hotplug-cpu.c @@ -397,12 +397,12 @@ static int parse_cede_parameters(void) CEDE_LATENCY_PARAM_MAX_LENGTH); if (call_status != 0) - printk(KERN_INFO "CEDE_LATENCY: \ - %s %s Error calling get-system-parameter(0x%x)\n", + printk(KERN_INFO "CEDE_LATENCY: " + "%s %s Error calling get-system-parameter(0x%x)\n", __FILE__, __func__, call_status); else - printk(KERN_INFO "CEDE_LATENCY: \ - get-system-parameter successful.\n"); + printk(KERN_INFO "CEDE_LATENCY: " + "get-system-parameter successful.\n"); return call_status; } diff --git a/arch/powerpc/platforms/pseries/smp.c b/arch/powerpc/platforms/pseries/smp.c index b488663..4e7f89a 100644 --- a/arch/powerpc/platforms/pseries/smp.c +++ b/arch/powerpc/platforms/pseries/smp.c @@ -144,8 +144,8 @@ static void __devinit smp_pSeries_kick_cpu(int nr) hcpuid = get_hard_smp_processor_id(nr); rc = plpar_hcall_norets(H_PROD, hcpuid); if (rc != H_SUCCESS) - printk(KERN_ERR "Error: Prod to wake up processor %d\ - Ret= %ld\n", nr, rc); + printk(KERN_ERR "Error: Prod to wake up processor %d " + "Ret= %ld\n", nr, rc); } } -- cgit v0.10.2 From 20a8ab97372e453f58a7b49e0d5e0f832406cffe Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 7 Feb 2010 13:52:05 +0000 Subject: powerpc/pseries: Quieten cede latency printk The cede latency stuff is relatively new and we don't need to complain about it not working on older firmware. Signed-off-by: Anton Blanchard Acked-by: Balbir Singh Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/platforms/pseries/hotplug-cpu.c b/arch/powerpc/platforms/pseries/hotplug-cpu.c index a70de10..d1b124e 100644 --- a/arch/powerpc/platforms/pseries/hotplug-cpu.c +++ b/arch/powerpc/platforms/pseries/hotplug-cpu.c @@ -387,24 +387,12 @@ static char cede_parameters[CEDE_LATENCY_PARAM_MAX_LENGTH]; static int parse_cede_parameters(void) { - int call_status; - memset(cede_parameters, 0, CEDE_LATENCY_PARAM_MAX_LENGTH); - call_status = rtas_call(rtas_token("ibm,get-system-parameter"), 3, 1, - NULL, - CEDE_LATENCY_TOKEN, - __pa(cede_parameters), - CEDE_LATENCY_PARAM_MAX_LENGTH); - - if (call_status != 0) - printk(KERN_INFO "CEDE_LATENCY: " - "%s %s Error calling get-system-parameter(0x%x)\n", - __FILE__, __func__, call_status); - else - printk(KERN_INFO "CEDE_LATENCY: " - "get-system-parameter successful.\n"); - - return call_status; + return rtas_call(rtas_token("ibm,get-system-parameter"), 3, 1, + NULL, + CEDE_LATENCY_TOKEN, + __pa(cede_parameters), + CEDE_LATENCY_PARAM_MAX_LENGTH); } static int __init pseries_cpu_hotplug_init(void) -- cgit v0.10.2 From b926a88a7b77bc27c0e7f67166f34b854fe21ccc Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 7 Feb 2010 12:26:28 +0000 Subject: powerpc: Reduce differences between pseries and ppc64 defconfigs The pseries and ppc64 defconfigs have drifted apart over the years. Reduce some of the differences while still keeping the idea that the ppc64 defconfig is cross platform but enables fewer features than pseries, eg NR_CPUS is lower. Also enable a number of common adapters as modules. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/configs/ppc64_defconfig b/arch/powerpc/configs/ppc64_defconfig index 7b3804a..b5b2599 100644 --- a/arch/powerpc/configs/ppc64_defconfig +++ b/arch/powerpc/configs/ppc64_defconfig @@ -137,8 +137,9 @@ CONFIG_TRACEPOINTS=y CONFIG_MARKERS=y CONFIG_OPROFILE=y CONFIG_HAVE_OPROFILE=y -# CONFIG_KPROBES is not set +CONFIG_KPROBES=y CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y +CONFIG_KRETPROBES=y CONFIG_HAVE_IOREMAP_PROT=y CONFIG_HAVE_KPROBES=y CONFIG_HAVE_KRETPROBES=y @@ -191,6 +192,7 @@ CONFIG_SCANLOG=m CONFIG_LPARCFG=y CONFIG_PPC_SMLPAR=y CONFIG_CMM=y +CONFIG_DTL=y CONFIG_PPC_ISERIES=y # @@ -328,9 +330,10 @@ CONFIG_ARCH_ENABLE_MEMORY_HOTREMOVE=y CONFIG_KEXEC=y # CONFIG_PHYP_DUMP is not set CONFIG_IRQ_ALL_CPUS=y -# CONFIG_NUMA is not set +CONFIG_NUMA=y +CONFIG_NODES_SHIFT=8 +CONFIG_MAX_ACTIVE_REGIONS=256 CONFIG_ARCH_SELECT_MEMORY_MODEL=y -CONFIG_ARCH_FLATMEM_ENABLE=y CONFIG_ARCH_SPARSEMEM_ENABLE=y CONFIG_ARCH_SPARSEMEM_DEFAULT=y CONFIG_ARCH_POPULATES_NODE_MAP=y @@ -339,6 +342,7 @@ CONFIG_SELECT_MEMORY_MODEL=y # CONFIG_DISCONTIGMEM_MANUAL is not set CONFIG_SPARSEMEM_MANUAL=y CONFIG_SPARSEMEM=y +CONFIG_NEED_MULTIPLE_NODES=y CONFIG_HAVE_MEMORY_PRESENT=y CONFIG_SPARSEMEM_EXTREME=y CONFIG_SPARSEMEM_VMEMMAP_ENABLE=y @@ -354,11 +358,12 @@ CONFIG_PHYS_ADDR_T_64BIT=y CONFIG_ZONE_DMA_FLAG=1 CONFIG_BOUNCE=y CONFIG_UNEVICTABLE_LRU=y +CONFIG_NODES_SPAN_OTHER_NODES=y CONFIG_ARCH_MEMORY_PROBE=y CONFIG_PPC_HAS_HASH_64K=y # CONFIG_PPC_64K_PAGES is not set CONFIG_FORCE_MAX_ZONEORDER=13 -# CONFIG_SCHED_SMT is not set +CONFIG_SCHED_SMT=y CONFIG_PROC_DEVICETREE=y # CONFIG_CMDLINE_BOOL is not set CONFIG_EXTRA_TARGETS="" @@ -790,12 +795,12 @@ CONFIG_SCSI_IPR=y CONFIG_SCSI_IPR_TRACE=y CONFIG_SCSI_IPR_DUMP=y # CONFIG_SCSI_QLOGIC_1280 is not set -# CONFIG_SCSI_QLA_FC is not set +CONFIG_SCSI_QLA_FC=m # CONFIG_SCSI_QLA_ISCSI is not set CONFIG_SCSI_LPFC=m # CONFIG_SCSI_DC395x is not set # CONFIG_SCSI_DC390T is not set -CONFIG_SCSI_DEBUG=m +# CONFIG_SCSI_DEBUG is not set # CONFIG_SCSI_SRP is not set # CONFIG_SCSI_LOWLEVEL_PCMCIA is not set # CONFIG_SCSI_DH is not set @@ -867,9 +872,8 @@ CONFIG_MD_AUTODETECT=y CONFIG_MD_LINEAR=y CONFIG_MD_RAID0=y CONFIG_MD_RAID1=y -CONFIG_MD_RAID10=y -CONFIG_MD_RAID456=y -CONFIG_MD_RAID5_RESHAPE=y +CONFIG_MD_RAID10=m +CONFIG_MD_RAID456=m CONFIG_MD_MULTIPATH=m CONFIG_MD_FAULTY=m CONFIG_BLK_DEV_DM=y @@ -984,7 +988,7 @@ CONFIG_ACENIC=m CONFIG_ACENIC_OMIT_TIGON_I=y # CONFIG_DL2K is not set CONFIG_E1000=y -# CONFIG_E1000E is not set +CONFIG_E1000E=m # CONFIG_IP1000 is not set # CONFIG_IGB is not set # CONFIG_NS83820 is not set @@ -1006,19 +1010,19 @@ CONFIG_GELIC_WIRELESS=y # CONFIG_ATL1E is not set # CONFIG_JME is not set CONFIG_NETDEV_10000=y -# CONFIG_CHELSIO_T1 is not set -# CONFIG_CHELSIO_T3 is not set +CONFIG_CHELSIO_T1=m +CONFIG_CHELSIO_T3=m CONFIG_EHEA=m # CONFIG_ENIC is not set -# CONFIG_IXGBE is not set +CONFIG_IXGBE=m CONFIG_IXGB=m -# CONFIG_S2IO is not set -# CONFIG_MYRI10GE is not set -# CONFIG_NETXEN_NIC is not set +CONFIG_S2IO=m +CONFIG_MYRI10GE=m +CONFIG_NETXEN_NIC=m # CONFIG_NIU is not set CONFIG_PASEMI_MAC=y -# CONFIG_MLX4_EN is not set -# CONFIG_MLX4_CORE is not set +CONFIG_MLX4_EN=m +CONFIG_MLX4_CORE=m # CONFIG_TEHUTI is not set # CONFIG_BNX2X is not set # CONFIG_QLGE is not set @@ -1169,7 +1173,7 @@ CONFIG_SERIAL_TXX9=y CONFIG_HAS_TXX9_SERIAL=y CONFIG_SERIAL_TXX9_NR_UARTS=6 CONFIG_SERIAL_TXX9_CONSOLE=y -# CONFIG_SERIAL_JSM is not set +CONFIG_SERIAL_JSM=m # CONFIG_SERIAL_OF_PLATFORM is not set CONFIG_UNIX98_PTYS=y CONFIG_LEGACY_PTYS=y @@ -1586,7 +1590,7 @@ CONFIG_USB_DEVICEFS=y CONFIG_USB_DEVICE_CLASS=y # CONFIG_USB_DYNAMIC_MINORS is not set # CONFIG_USB_OTG is not set -# CONFIG_USB_MON is not set +CONFIG_USB_MON=m # CONFIG_USB_WUSB is not set # CONFIG_USB_WUSB_CBAF is not set @@ -1686,21 +1690,22 @@ CONFIG_USB_APPLEDISPLAY=m # CONFIG_NEW_LEDS is not set # CONFIG_ACCESSIBILITY is not set CONFIG_INFINIBAND=m -# CONFIG_INFINIBAND_USER_MAD is not set -# CONFIG_INFINIBAND_USER_ACCESS is not set +CONFIG_INFINIBAND_USER_MAD=m +CONFIG_INFINIBAND_USER_ACCESS=m +CONFIG_INFINIBAND_USER_MEM=y CONFIG_INFINIBAND_ADDR_TRANS=y CONFIG_INFINIBAND_MTHCA=m CONFIG_INFINIBAND_MTHCA_DEBUG=y -# CONFIG_INFINIBAND_IPATH is not set +CONFIG_INFINIBAND_IPATH=m CONFIG_INFINIBAND_EHCA=m # CONFIG_INFINIBAND_AMSO1100 is not set -# CONFIG_MLX4_INFINIBAND is not set +CONFIG_MLX4_INFINIBAND=m # CONFIG_INFINIBAND_NES is not set CONFIG_INFINIBAND_IPOIB=m -# CONFIG_INFINIBAND_IPOIB_CM is not set +CONFIG_INFINIBAND_IPOIB_CM=y CONFIG_INFINIBAND_IPOIB_DEBUG=y # CONFIG_INFINIBAND_IPOIB_DEBUG_DATA is not set -# CONFIG_INFINIBAND_SRP is not set +CONFIG_INFINIBAND_SRP=m CONFIG_INFINIBAND_ISER=m CONFIG_EDAC=y @@ -1798,7 +1803,7 @@ CONFIG_REISERFS_FS=y CONFIG_REISERFS_FS_XATTR=y CONFIG_REISERFS_FS_POSIX_ACL=y CONFIG_REISERFS_FS_SECURITY=y -CONFIG_JFS_FS=y +CONFIG_JFS_FS=m CONFIG_JFS_POSIX_ACL=y CONFIG_JFS_SECURITY=y # CONFIG_JFS_DEBUG is not set @@ -1811,14 +1816,22 @@ CONFIG_XFS_POSIX_ACL=y # CONFIG_XFS_RT is not set # CONFIG_XFS_DEBUG is not set # CONFIG_GFS2_FS is not set -# CONFIG_OCFS2_FS is not set +CONFIG_OCFS2_FS=m +CONFIG_OCFS2_FS_O2CB=m +CONFIG_OCFS2_FS_STATS=y +CONFIG_OCFS2_DEBUG_MASKLOG=y +# CONFIG_OCFS2_DEBUG_FS is not set +# CONFIG_OCFS2_COMPAT_JBD is not set +CONFIG_BTRFS_FS=m +CONFIG_BTRFS_FS_POSIX_ACL=y +CONFIG_NILFS2_FS=m CONFIG_DNOTIFY=y CONFIG_INOTIFY=y CONFIG_INOTIFY_USER=y # CONFIG_QUOTA is not set # CONFIG_AUTOFS_FS is not set CONFIG_AUTOFS4_FS=m -# CONFIG_FUSE_FS is not set +CONFIG_FUSE_FS=m # # CD-ROM/DVD Filesystems @@ -1851,7 +1864,7 @@ CONFIG_TMPFS=y # CONFIG_TMPFS_POSIX_ACL is not set CONFIG_HUGETLBFS=y CONFIG_HUGETLB_PAGE=y -# CONFIG_CONFIGFS_FS is not set +CONFIG_CONFIGFS_FS=m # # Miscellaneous filesystems @@ -2075,7 +2088,7 @@ CONFIG_XMON=y CONFIG_XMON_DISASSEMBLY=y CONFIG_DEBUGGER=y CONFIG_IRQSTACKS=y -# CONFIG_VIRQ_DEBUG is not set +CONFIG_VIRQ_DEBUG=y CONFIG_BOOTX_TEXT=y # CONFIG_PPC_EARLY_DEBUG is not set diff --git a/arch/powerpc/configs/pseries_defconfig b/arch/powerpc/configs/pseries_defconfig index ca9ff9a..41de3dd 100644 --- a/arch/powerpc/configs/pseries_defconfig +++ b/arch/powerpc/configs/pseries_defconfig @@ -159,7 +159,7 @@ CONFIG_MODULE_SRCVERSION_ALL=y CONFIG_KMOD=y CONFIG_STOP_MACHINE=y CONFIG_BLOCK=y -# CONFIG_BLK_DEV_IO_TRACE is not set +CONFIG_BLK_DEV_IO_TRACE=y CONFIG_BLK_DEV_BSG=y # CONFIG_BLK_DEV_INTEGRITY is not set CONFIG_BLOCK_COMPAT=y @@ -191,6 +191,7 @@ CONFIG_SCANLOG=m CONFIG_LPARCFG=y CONFIG_PPC_SMLPAR=y CONFIG_CMM=y +CONFIG_DTL=y # CONFIG_PPC_ISERIES is not set # CONFIG_PPC_PMAC is not set # CONFIG_PPC_MAPLE is not set @@ -255,7 +256,8 @@ CONFIG_KEXEC=y # CONFIG_PHYP_DUMP is not set CONFIG_IRQ_ALL_CPUS=y CONFIG_NUMA=y -CONFIG_NODES_SHIFT=4 +CONFIG_NODES_SHIFT=8 +CONFIG_MAX_ACTIVE_REGIONS=256 CONFIG_ARCH_SELECT_MEMORY_MODEL=y CONFIG_ARCH_SPARSEMEM_ENABLE=y CONFIG_ARCH_SPARSEMEM_DEFAULT=y @@ -270,7 +272,9 @@ CONFIG_HAVE_MEMORY_PRESENT=y CONFIG_SPARSEMEM_EXTREME=y CONFIG_SPARSEMEM_VMEMMAP_ENABLE=y CONFIG_SPARSEMEM_VMEMMAP=y -# CONFIG_MEMORY_HOTPLUG is not set +CONFIG_MEMORY_HOTPLUG=y +CONFIG_MEMORY_HOTPLUG_SPARSE=y +CONFIG_MEMORY_HOTREMOVE=y CONFIG_PAGEFLAGS_EXTENDED=y CONFIG_SPLIT_PTLOCK_CPUS=4 CONFIG_MIGRATION=y @@ -705,7 +709,7 @@ CONFIG_MD_LINEAR=y CONFIG_MD_RAID0=y CONFIG_MD_RAID1=y CONFIG_MD_RAID10=m -# CONFIG_MD_RAID456 is not set +CONFIG_MD_RAID456=m CONFIG_MD_MULTIPATH=m CONFIG_MD_FAULTY=m CONFIG_BLK_DEV_DM=y @@ -800,7 +804,7 @@ CONFIG_ACENIC=m CONFIG_ACENIC_OMIT_TIGON_I=y # CONFIG_DL2K is not set CONFIG_E1000=y -# CONFIG_E1000E is not set +CONFIG_E1000E=m # CONFIG_IP1000 is not set # CONFIG_IGB is not set # CONFIG_NS83820 is not set @@ -818,18 +822,18 @@ CONFIG_TIGON3=y # CONFIG_ATL1E is not set # CONFIG_JME is not set CONFIG_NETDEV_10000=y -# CONFIG_CHELSIO_T1 is not set -# CONFIG_CHELSIO_T3 is not set +CONFIG_CHELSIO_T1=m +CONFIG_CHELSIO_T3=m CONFIG_EHEA=y # CONFIG_ENIC is not set -# CONFIG_IXGBE is not set +CONFIG_IXGBE=m CONFIG_IXGB=m CONFIG_S2IO=m -# CONFIG_MYRI10GE is not set -# CONFIG_NETXEN_NIC is not set +CONFIG_MYRI10GE=m +CONFIG_NETXEN_NIC=m # CONFIG_NIU is not set -# CONFIG_MLX4_EN is not set -# CONFIG_MLX4_CORE is not set +CONFIG_MLX4_EN=m +CONFIG_MLX4_CORE=m # CONFIG_TEHUTI is not set # CONFIG_BNX2X is not set # CONFIG_QLGE is not set @@ -894,7 +898,7 @@ CONFIG_INPUT_MOUSEDEV=y CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024 CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768 # CONFIG_INPUT_JOYDEV is not set -# CONFIG_INPUT_EVDEV is not set +CONFIG_INPUT_EVDEV=m # CONFIG_INPUT_EVBUG is not set # @@ -1271,7 +1275,7 @@ CONFIG_USB_DEVICEFS=y CONFIG_USB_DEVICE_CLASS=y # CONFIG_USB_DYNAMIC_MINORS is not set # CONFIG_USB_OTG is not set -CONFIG_USB_MON=y +CONFIG_USB_MON=m # CONFIG_USB_WUSB is not set # CONFIG_USB_WUSB_CBAF is not set @@ -1311,7 +1315,7 @@ CONFIG_USB_OHCI_LITTLE_ENDIAN=y # # may also be needed; see USB_STORAGE Help for more information # -CONFIG_USB_STORAGE=y +CONFIG_USB_STORAGE=m # CONFIG_USB_STORAGE_DEBUG is not set # CONFIG_USB_STORAGE_DATAFAB is not set # CONFIG_USB_STORAGE_FREECOM is not set @@ -1322,7 +1326,7 @@ CONFIG_USB_STORAGE=y # CONFIG_USB_STORAGE_SDDR55 is not set # CONFIG_USB_STORAGE_JUMPSHOT is not set # CONFIG_USB_STORAGE_ALAUDA is not set -CONFIG_USB_STORAGE_ONETOUCH=y +# CONFIG_USB_STORAGE_ONETOUCH is not set # CONFIG_USB_STORAGE_KARMA is not set # CONFIG_USB_STORAGE_CYPRESS_ATACB is not set # CONFIG_USB_LIBUSUAL is not set @@ -1377,17 +1381,17 @@ CONFIG_INFINIBAND_USER_MEM=y CONFIG_INFINIBAND_ADDR_TRANS=y CONFIG_INFINIBAND_MTHCA=m CONFIG_INFINIBAND_MTHCA_DEBUG=y -# CONFIG_INFINIBAND_IPATH is not set +CONFIG_INFINIBAND_IPATH=m CONFIG_INFINIBAND_EHCA=m # CONFIG_INFINIBAND_AMSO1100 is not set -# CONFIG_MLX4_INFINIBAND is not set +CONFIG_MLX4_INFINIBAND=m # CONFIG_INFINIBAND_NES is not set CONFIG_INFINIBAND_IPOIB=m -# CONFIG_INFINIBAND_IPOIB_CM is not set +CONFIG_INFINIBAND_IPOIB_CM=y CONFIG_INFINIBAND_IPOIB_DEBUG=y # CONFIG_INFINIBAND_IPOIB_DEBUG_DATA is not set CONFIG_INFINIBAND_SRP=m -# CONFIG_INFINIBAND_ISER is not set +CONFIG_INFINIBAND_ISER=m # CONFIG_EDAC is not set # CONFIG_RTC_CLASS is not set # CONFIG_DMADEVICES is not set @@ -1443,6 +1447,9 @@ CONFIG_OCFS2_FS_STATS=y CONFIG_OCFS2_DEBUG_MASKLOG=y # CONFIG_OCFS2_DEBUG_FS is not set # CONFIG_OCFS2_COMPAT_JBD is not set +CONFIG_BTRFS_FS=m +CONFIG_BTRFS_FS_POSIX_ACL=y +CONFIG_NILFS2_FS=m CONFIG_DNOTIFY=y CONFIG_INOTIFY=y CONFIG_INOTIFY_USER=y @@ -1455,8 +1462,8 @@ CONFIG_FUSE_FS=m # CD-ROM/DVD Filesystems # CONFIG_ISO9660_FS=y -CONFIG_JOLIET=y -CONFIG_ZISOFS=y +# CONFIG_JOLIET is not set +# CONFIG_ZISOFS is not set CONFIG_UDF_FS=m CONFIG_UDF_NLS=y @@ -1508,14 +1515,14 @@ CONFIG_NFS_FS=y CONFIG_NFS_V3=y CONFIG_NFS_V3_ACL=y CONFIG_NFS_V4=y -CONFIG_NFSD=y +CONFIG_NFSD=m CONFIG_NFSD_V2_ACL=y CONFIG_NFSD_V3=y CONFIG_NFSD_V3_ACL=y CONFIG_NFSD_V4=y CONFIG_LOCKD=y CONFIG_LOCKD_V4=y -CONFIG_EXPORTFS=y +CONFIG_EXPORTFS=m CONFIG_NFS_ACL_SUPPORT=y CONFIG_NFS_COMMON=y CONFIG_SUNRPC=y @@ -1681,12 +1688,12 @@ CONFIG_DYNAMIC_PRINTK_DEBUG=y CONFIG_HAVE_ARCH_KGDB=y # CONFIG_KGDB is not set CONFIG_DEBUG_STACKOVERFLOW=y -# CONFIG_DEBUG_STACK_USAGE is not set +CONFIG_DEBUG_STACK_USAGE=y # CONFIG_DEBUG_PAGEALLOC is not set # CONFIG_HCALL_STATS is not set -# CONFIG_CODE_PATCHING_SELFTEST is not set -# CONFIG_FTR_FIXUP_SELFTEST is not set -# CONFIG_MSI_BITMAP_SELFTEST is not set +CONFIG_CODE_PATCHING_SELFTEST=y +CONFIG_FTR_FIXUP_SELFTEST=y +CONFIG_MSI_BITMAP_SELFTEST=y CONFIG_XMON=y CONFIG_XMON_DEFAULT=y CONFIG_XMON_DISASSEMBLY=y -- cgit v0.10.2 From 66fcb1059deeae072c1bf1536a8e2613028bce3e Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 7 Feb 2010 14:44:16 +0000 Subject: powerpc: Add last sysfs file and dump of ftrace buffer to oops printout Add printout of last accessed sysfs file, added to x86 in ae87221d3ce49d9de1e43756da834fd0bf05a2ad (sysfs: crash debugging) Also add the notify_die hook that allows us to print out the ftrace buffer on oops. This is useful in conjunction with ftrace function_graph: Oops: Kernel access of bad area, sig: 11 [#1] SMP NR_CPUS=128 NUMA pSeries last sysfs file: /sys/class/net/tunl0/type Dumping ftrace buffer: ... 0) | .sysrq_handle_crash() { 0) 0.476 us | .hash_page(); 0) 0.488 us | .xmon_fault_handler(); 0) | .bad_page_fault() { 0) | .search_exception_tables() { 0) 0.590 us | .search_module_extables(); 0) 2.546 us | } 0) | .printk() { 0) | .vprintk() { 0) 0.488 us | ._raw_spin_lock(); 0) 0.572 us | .emit_log_char(); Showing the function graph of a sysrq-c crash. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c index 4588d1e..0a320db 100644 --- a/arch/powerpc/kernel/traps.c +++ b/arch/powerpc/kernel/traps.c @@ -146,6 +146,11 @@ int die(const char *str, struct pt_regs *regs, long err) #endif printk("%s\n", ppc_md.name ? ppc_md.name : ""); + sysfs_printk_last_file(); + if (notify_die(DIE_OOPS, str, regs, err, 255, + SIGSEGV) == NOTIFY_STOP) + return 1; + print_modules(); show_regs(regs); } else { -- cgit v0.10.2 From 7317ac87119dfb6fac36af4bcc64a25b62b5898d Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 7 Feb 2010 12:30:12 +0000 Subject: powerpc: Convert mmu context allocator from idr to ida We can use the much more lightweight ida allocator since we don't need the pointer storage idr provides. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/mm/mmu_context_hash64.c b/arch/powerpc/mm/mmu_context_hash64.c index b910d37..51622da 100644 --- a/arch/powerpc/mm/mmu_context_hash64.c +++ b/arch/powerpc/mm/mmu_context_hash64.c @@ -23,7 +23,7 @@ #include static DEFINE_SPINLOCK(mmu_context_lock); -static DEFINE_IDR(mmu_context_idr); +static DEFINE_IDA(mmu_context_ida); /* * The proto-VSID space has 2^35 - 1 segments available for user mappings. @@ -39,11 +39,11 @@ int __init_new_context(void) int err; again: - if (!idr_pre_get(&mmu_context_idr, GFP_KERNEL)) + if (!ida_pre_get(&mmu_context_ida, GFP_KERNEL)) return -ENOMEM; spin_lock(&mmu_context_lock); - err = idr_get_new_above(&mmu_context_idr, NULL, 1, &index); + err = ida_get_new_above(&mmu_context_ida, 1, &index); spin_unlock(&mmu_context_lock); if (err == -EAGAIN) @@ -53,7 +53,7 @@ again: if (index > MAX_CONTEXT) { spin_lock(&mmu_context_lock); - idr_remove(&mmu_context_idr, index); + ida_remove(&mmu_context_ida, index); spin_unlock(&mmu_context_lock); return -ENOMEM; } @@ -85,7 +85,7 @@ int init_new_context(struct task_struct *tsk, struct mm_struct *mm) void __destroy_context(int context_id) { spin_lock(&mmu_context_lock); - idr_remove(&mmu_context_idr, context_id); + ida_remove(&mmu_context_ida, context_id); spin_unlock(&mmu_context_lock); } EXPORT_SYMBOL_GPL(__destroy_context); -- cgit v0.10.2 From a13672fb30b46eaf9fa50357e1e863de00631401 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 7 Feb 2010 13:23:30 +0000 Subject: powerpc: Reformat SD_NODE_INIT to match x86 Clean up SD_NODE_INITS so we can easily compare it to x86. Similar to the work in 47734f89be0614b5acbd6a532390f9c72f019648 (sched: Clean up topology.h) Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/include/asm/topology.h b/arch/powerpc/include/asm/topology.h index bbf8970..8eaec31 100644 --- a/arch/powerpc/include/asm/topology.h +++ b/arch/powerpc/include/asm/topology.h @@ -38,27 +38,33 @@ static inline int pcibus_to_node(struct pci_bus *bus) cpumask_of_node(pcibus_to_node(bus))) /* sched_domains SD_NODE_INIT for PPC64 machines */ -#define SD_NODE_INIT (struct sched_domain) { \ - .parent = NULL, \ - .child = NULL, \ - .groups = NULL, \ - .min_interval = 8, \ - .max_interval = 32, \ - .busy_factor = 32, \ - .imbalance_pct = 125, \ - .cache_nice_tries = 1, \ - .busy_idx = 3, \ - .idle_idx = 1, \ - .newidle_idx = 0, \ - .wake_idx = 0, \ - .flags = SD_LOAD_BALANCE \ - | SD_BALANCE_EXEC \ - | SD_BALANCE_FORK \ - | SD_BALANCE_NEWIDLE \ - | SD_SERIALIZE, \ - .last_balance = jiffies, \ - .balance_interval = 1, \ - .nr_balance_failed = 0, \ +#define SD_NODE_INIT (struct sched_domain) { \ + .min_interval = 8, \ + .max_interval = 32, \ + .busy_factor = 32, \ + .imbalance_pct = 125, \ + .cache_nice_tries = 1, \ + .busy_idx = 3, \ + .idle_idx = 1, \ + .newidle_idx = 0, \ + .wake_idx = 0, \ + .forkexec_idx = 0, \ + \ + .flags = 1*SD_LOAD_BALANCE \ + | 1*SD_BALANCE_NEWIDLE \ + | 1*SD_BALANCE_EXEC \ + | 1*SD_BALANCE_FORK \ + | 0*SD_BALANCE_WAKE \ + | 0*SD_WAKE_AFFINE \ + | 0*SD_PREFER_LOCAL \ + | 0*SD_SHARE_CPUPOWER \ + | 0*SD_POWERSAVINGS_BALANCE \ + | 0*SD_SHARE_PKG_RESOURCES \ + | 1*SD_SERIALIZE \ + | 0*SD_PREFER_SIBLING \ + , \ + .last_balance = jiffies, \ + .balance_interval = 1, \ } extern void __init dump_numa_cpu_topology(void); -- cgit v0.10.2 From 0b9612c210f6bf6efb32ab84b8f34018eaacb08d Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 7 Feb 2010 13:07:36 +0000 Subject: powerpc: Make powerpc_firmware_features __read_mostly We use firmware_has_feature quite a lot these days, so it's worth putting powerpc_firmware_features into __read_mostly. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/firmware.c b/arch/powerpc/kernel/firmware.c index 1679a70..6b1f427 100644 --- a/arch/powerpc/kernel/firmware.c +++ b/arch/powerpc/kernel/firmware.c @@ -17,5 +17,5 @@ #include -unsigned long powerpc_firmware_features; +unsigned long powerpc_firmware_features __read_mostly; EXPORT_SYMBOL_GPL(powerpc_firmware_features); -- cgit v0.10.2 From 8354be9c104db8e92bb079e29540fbafb7f09b5c Mon Sep 17 00:00:00 2001 From: Frans Pop Date: Sat, 6 Feb 2010 07:47:20 +0000 Subject: powerpc: Remove trailing space in messages Signed-off-by: Frans Pop Cc: linuxppc-dev@ozlabs.org Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/lparcfg.c b/arch/powerpc/kernel/lparcfg.c index 79a00bb..d09d1c6 100644 --- a/arch/powerpc/kernel/lparcfg.c +++ b/arch/powerpc/kernel/lparcfg.c @@ -359,7 +359,7 @@ static void parse_system_parameter_string(struct seq_file *m) unsigned char *local_buffer = kmalloc(SPLPAR_MAXLENGTH, GFP_KERNEL); if (!local_buffer) { - printk(KERN_ERR "%s %s kmalloc failure at line %d \n", + printk(KERN_ERR "%s %s kmalloc failure at line %d\n", __FILE__, __func__, __LINE__); return; } @@ -383,13 +383,13 @@ static void parse_system_parameter_string(struct seq_file *m) int idx, w_idx; char *workbuffer = kzalloc(SPLPAR_MAXLENGTH, GFP_KERNEL); if (!workbuffer) { - printk(KERN_ERR "%s %s kmalloc failure at line %d \n", + printk(KERN_ERR "%s %s kmalloc failure at line %d\n", __FILE__, __func__, __LINE__); kfree(local_buffer); return; } #ifdef LPARCFG_DEBUG - printk(KERN_INFO "success calling get-system-parameter \n"); + printk(KERN_INFO "success calling get-system-parameter\n"); #endif splpar_strlen = local_buffer[0] * 256 + local_buffer[1]; local_buffer += 2; /* step over strlen value */ @@ -440,7 +440,7 @@ static int lparcfg_count_active_processors(void) while ((cpus_dn = of_find_node_by_type(cpus_dn, "cpu"))) { #ifdef LPARCFG_DEBUG - printk(KERN_ERR "cpus_dn %p \n", cpus_dn); + printk(KERN_ERR "cpus_dn %p\n", cpus_dn); #endif count++; } @@ -725,7 +725,7 @@ static int lparcfg_data(struct seq_file *m, void *v) const unsigned int *lp_index_ptr; unsigned int lp_index = 0; - seq_printf(m, "%s %s \n", MODULE_NAME, MODULE_VERS); + seq_printf(m, "%s %s\n", MODULE_NAME, MODULE_VERS); rootdn = of_find_node_by_path("/"); if (rootdn) { diff --git a/arch/powerpc/kernel/pci_of_scan.c b/arch/powerpc/kernel/pci_of_scan.c index 4aa1740..cd11d5c 100644 --- a/arch/powerpc/kernel/pci_of_scan.c +++ b/arch/powerpc/kernel/pci_of_scan.c @@ -304,7 +304,7 @@ static void __devinit __of_scan_bus(struct device_node *node, int reglen, devfn; struct pci_dev *dev; - pr_debug("of_scan_bus(%s) bus no %d... \n", + pr_debug("of_scan_bus(%s) bus no %d...\n", node->full_name, bus->number); /* Scan direct children */ diff --git a/arch/powerpc/platforms/85xx/stx_gp3.c b/arch/powerpc/platforms/85xx/stx_gp3.c index f559918..bc33d18 100644 --- a/arch/powerpc/platforms/85xx/stx_gp3.c +++ b/arch/powerpc/platforms/85xx/stx_gp3.c @@ -134,7 +134,7 @@ static void stx_gp3_show_cpuinfo(struct seq_file *m) pvid = mfspr(SPRN_PVR); svid = mfspr(SPRN_SVR); - seq_printf(m, "Vendor\t\t: RPC Electronics STx \n"); + seq_printf(m, "Vendor\t\t: RPC Electronics STx\n"); seq_printf(m, "PVR\t\t: 0x%x\n", pvid); seq_printf(m, "SVR\t\t: 0x%x\n", svid); diff --git a/arch/powerpc/platforms/iseries/proc.c b/arch/powerpc/platforms/iseries/proc.c index 91f4c6c..0676368 100644 --- a/arch/powerpc/platforms/iseries/proc.c +++ b/arch/powerpc/platforms/iseries/proc.c @@ -85,7 +85,7 @@ static int proc_titantod_show(struct seq_file *m, void *v) seq_printf(m, " titan elapsed = %lu uSec\n", titan_usec); seq_printf(m, " tb elapsed = %lu ticks\n", tb_ticks); - seq_printf(m, " titan jiffies = %lu.%04lu \n", titan_jiffies, + seq_printf(m, " titan jiffies = %lu.%04lu\n", titan_jiffies, titan_jiff_rem_usec); seq_printf(m, " tb jiffies = %lu.%04lu\n", tb_jiffies, tb_jiff_rem_usec); diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index a6cd339..b086341 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -256,7 +256,7 @@ static unsigned long iSeries_process_mainstore_vpd(struct MemoryBlock *mb_array, mem_blocks = iSeries_process_Condor_mainstore_vpd(mb_array, max_entries); - printk("Mainstore_VPD: numMemoryBlocks = %ld \n", mem_blocks); + printk("Mainstore_VPD: numMemoryBlocks = %ld\n", mem_blocks); for (i = 0; i < mem_blocks; ++i) { printk("Mainstore_VPD: block %3ld logical chunks %016lx - %016lx\n" " abs chunks %016lx - %016lx\n", diff --git a/arch/powerpc/platforms/powermac/bootx_init.c b/arch/powerpc/platforms/powermac/bootx_init.c index 9dd789a..84d7fd9 100644 --- a/arch/powerpc/platforms/powermac/bootx_init.c +++ b/arch/powerpc/platforms/powermac/bootx_init.c @@ -539,7 +539,7 @@ void __init bootx_init(unsigned long r3, unsigned long r4) if (model && (strcmp(model, "iMac,1") == 0 || strcmp(model, "PowerMac1,1") == 0)) { - bootx_printf("iMac,1 detected, shutting down USB \n"); + bootx_printf("iMac,1 detected, shutting down USB\n"); out_le32((unsigned __iomem *)0x80880008, 1); /* XXX */ } } @@ -554,7 +554,7 @@ void __init bootx_init(unsigned long r3, unsigned long r4) } else space = bi->totalParamsSize; - bootx_printf("Total space used by parameters & ramdisk: 0x%x \n", space); + bootx_printf("Total space used by parameters & ramdisk: 0x%x\n", space); /* New BootX will have flushed all TLBs and enters kernel with * MMU switched OFF, so this should not be useful anymore. diff --git a/arch/powerpc/platforms/pseries/eeh_driver.c b/arch/powerpc/platforms/pseries/eeh_driver.c index ef8e454..0c252c3 100644 --- a/arch/powerpc/platforms/pseries/eeh_driver.c +++ b/arch/powerpc/platforms/pseries/eeh_driver.c @@ -478,9 +478,9 @@ excess_failures: * due to actual, failed cards. */ printk(KERN_ERR - "EEH: PCI device at location=%s driver=%s pci addr=%s \n" + "EEH: PCI device at location=%s driver=%s pci addr=%s\n" "has failed %d times in the last hour " - "and has been permanently disabled. \n" + "and has been permanently disabled.\n" "Please try reseating this device or replacing it.\n", location, drv_str, pci_str, frozen_pdn->eeh_freeze_count); goto perm_error; @@ -488,7 +488,7 @@ excess_failures: hard_fail: printk(KERN_ERR "EEH: Unable to recover from failure of PCI device " - "at location=%s driver=%s pci addr=%s \n" + "at location=%s driver=%s pci addr=%s\n" "Please try reseating this device or replacing it.\n", location, drv_str, pci_str); diff --git a/arch/powerpc/platforms/pseries/pci_dlpar.c b/arch/powerpc/platforms/pseries/pci_dlpar.c index b6fa3e4..4b7a062 100644 --- a/arch/powerpc/platforms/pseries/pci_dlpar.c +++ b/arch/powerpc/platforms/pseries/pci_dlpar.c @@ -165,7 +165,7 @@ int remove_phb_dynamic(struct pci_controller *phb) struct resource *res; int rc, i; - pr_debug("PCI: Removing PHB %04x:%02x... \n", + pr_debug("PCI: Removing PHB %04x:%02x...\n", pci_domain_nr(b), b->number); /* We cannot to remove a root bus that has children */ diff --git a/arch/powerpc/platforms/pseries/phyp_dump.c b/arch/powerpc/platforms/pseries/phyp_dump.c index 15eb610..225a50a 100644 --- a/arch/powerpc/platforms/pseries/phyp_dump.c +++ b/arch/powerpc/platforms/pseries/phyp_dump.c @@ -150,7 +150,7 @@ static void print_dump_header(const struct phyp_dump_header *ph) printk(KERN_INFO "Max auto time= %d\n", ph->maxtime_to_auto); /*set cpu state and hpte states as well scratch pad area */ - printk(KERN_INFO " CPU AREA \n"); + printk(KERN_INFO " CPU AREA\n"); printk(KERN_INFO "cpu dump_flags =%d\n", ph->cpu_data.dump_flags); printk(KERN_INFO "cpu source_type =%d\n", ph->cpu_data.source_type); printk(KERN_INFO "cpu error_flags =%d\n", ph->cpu_data.error_flags); @@ -161,7 +161,7 @@ static void print_dump_header(const struct phyp_dump_header *ph) printk(KERN_INFO "cpu length_copied =%llx\n", ph->cpu_data.length_copied); - printk(KERN_INFO " HPTE AREA \n"); + printk(KERN_INFO " HPTE AREA\n"); printk(KERN_INFO "HPTE dump_flags =%d\n", ph->hpte_data.dump_flags); printk(KERN_INFO "HPTE source_type =%d\n", ph->hpte_data.source_type); printk(KERN_INFO "HPTE error_flags =%d\n", ph->hpte_data.error_flags); @@ -172,7 +172,7 @@ static void print_dump_header(const struct phyp_dump_header *ph) printk(KERN_INFO "HPTE length_copied =%llx\n", ph->hpte_data.length_copied); - printk(KERN_INFO " SRSD AREA \n"); + printk(KERN_INFO " SRSD AREA\n"); printk(KERN_INFO "SRSD dump_flags =%d\n", ph->kernel_data.dump_flags); printk(KERN_INFO "SRSD source_type =%d\n", ph->kernel_data.source_type); printk(KERN_INFO "SRSD error_flags =%d\n", ph->kernel_data.error_flags); diff --git a/arch/powerpc/platforms/pseries/xics.c b/arch/powerpc/platforms/pseries/xics.c index 1ee66db..e9a52ec 100644 --- a/arch/powerpc/platforms/pseries/xics.c +++ b/arch/powerpc/platforms/pseries/xics.c @@ -127,7 +127,7 @@ static inline unsigned int lpar_xirr_info_get(void) lpar_rc = plpar_xirr(&return_value); if (lpar_rc != H_SUCCESS) - panic(" bad return code xirr - rc = %lx \n", lpar_rc); + panic(" bad return code xirr - rc = %lx\n", lpar_rc); return (unsigned int)return_value; } diff --git a/arch/powerpc/sysdev/mpic.c b/arch/powerpc/sysdev/mpic.c index 470dc6c..314f0fa 100644 --- a/arch/powerpc/sysdev/mpic.c +++ b/arch/powerpc/sysdev/mpic.c @@ -1575,7 +1575,7 @@ void mpic_request_ipis(void) int i; BUG_ON(mpic == NULL); - printk(KERN_INFO "mpic: requesting IPIs ... \n"); + printk(KERN_INFO "mpic: requesting IPIs...\n"); for (i = 0; i < 4; i++) { unsigned int vipi = irq_create_mapping(mpic->irqhost, diff --git a/arch/powerpc/sysdev/qe_lib/qe_ic.c b/arch/powerpc/sysdev/qe_lib/qe_ic.c index 2acc928..5f0b036 100644 --- a/arch/powerpc/sysdev/qe_lib/qe_ic.c +++ b/arch/powerpc/sysdev/qe_lib/qe_ic.c @@ -256,7 +256,7 @@ static int qe_ic_host_map(struct irq_host *h, unsigned int virq, struct irq_chip *chip; if (qe_ic_info[hw].mask == 0) { - printk(KERN_ERR "Can't map reserved IRQ \n"); + printk(KERN_ERR "Can't map reserved IRQ\n"); return -EINVAL; } /* Default chip */ diff --git a/arch/powerpc/sysdev/qe_lib/qe_io.c b/arch/powerpc/sysdev/qe_lib/qe_io.c index 7c87460..77e4934 100644 --- a/arch/powerpc/sysdev/qe_lib/qe_io.c +++ b/arch/powerpc/sysdev/qe_lib/qe_io.c @@ -157,13 +157,13 @@ int par_io_of_config(struct device_node *np) const unsigned int *pio_map; if (par_io == NULL) { - printk(KERN_ERR "par_io not initialized \n"); + printk(KERN_ERR "par_io not initialized\n"); return -1; } ph = of_get_property(np, "pio-handle", NULL); if (ph == NULL) { - printk(KERN_ERR "pio-handle not available \n"); + printk(KERN_ERR "pio-handle not available\n"); return -1; } @@ -171,12 +171,12 @@ int par_io_of_config(struct device_node *np) pio_map = of_get_property(pio, "pio-map", &pio_map_len); if (pio_map == NULL) { - printk(KERN_ERR "pio-map is not set! \n"); + printk(KERN_ERR "pio-map is not set!\n"); return -1; } pio_map_len /= sizeof(unsigned int); if ((pio_map_len % 6) != 0) { - printk(KERN_ERR "pio-map format wrong! \n"); + printk(KERN_ERR "pio-map format wrong!\n"); return -1; } diff --git a/drivers/macintosh/therm_adt746x.c b/drivers/macintosh/therm_adt746x.c index 58809b0..c42eeb4 100644 --- a/drivers/macintosh/therm_adt746x.c +++ b/drivers/macintosh/therm_adt746x.c @@ -316,7 +316,7 @@ static void update_fans_speed (struct thermostat *th) if (verbose) printk(KERN_DEBUG "adt746x: Setting fans speed to %d " - "(limit exceeded by %d on %s) \n", + "(limit exceeded by %d on %s)\n", new_speed, var, sensor_location[fan_number+1]); write_both_fan_speed(th, new_speed); diff --git a/drivers/ps3/ps3av.c b/drivers/ps3/ps3av.c index e82d8c9..95a689b 100644 --- a/drivers/ps3/ps3av.c +++ b/drivers/ps3/ps3av.c @@ -532,7 +532,7 @@ static void ps3av_set_videomode_packet(u32 id) res = ps3av_cmd_avb_param(&avb_param, len); if (res == PS3AV_STATUS_NO_SYNC_HEAD) printk(KERN_WARNING - "%s: Command failed. Please try your request again. \n", + "%s: Command failed. Please try your request again.\n", __func__); else if (res) dev_dbg(&ps3av->dev->core, "ps3av_cmd_avb_param failed\n"); -- cgit v0.10.2 From 44c9f3cc1a07503b653b571eac5792436da39a30 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 7 Feb 2010 19:37:29 +0000 Subject: powerpc: Clear MSR_RI during RTAS calls RTAS should never cause an exception but if it does (for example accessing outside our RMO) then we might go a long way through the kernel before oopsing. If we unset MSR_RI we should at least stop things on exception exit. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S index bdcb557..07109d8 100644 --- a/arch/powerpc/kernel/entry_64.S +++ b/arch/powerpc/kernel/entry_64.S @@ -791,9 +791,8 @@ _GLOBAL(enter_rtas) li r9,1 rldicr r9,r9,MSR_SF_LG,(63-MSR_SF_LG) - ori r9,r9,MSR_IR|MSR_DR|MSR_FE0|MSR_FE1|MSR_FP + ori r9,r9,MSR_IR|MSR_DR|MSR_FE0|MSR_FE1|MSR_FP|MSR_RI andc r6,r0,r9 - ori r6,r6,MSR_RI sync /* disable interrupts so SRR0/1 */ mtmsrd r0 /* don't get trashed */ -- cgit v0.10.2 From b919ee827e048826786fd7e889a2b04f63382fe6 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 7 Feb 2010 19:26:29 +0000 Subject: powerpc: Only print clockevent settings once The clockevent multiplier and shift is useful information, but we only need to print it once. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 0a5b108..ed1c0f5 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -935,8 +935,8 @@ static void register_decrementer_clockevent(int cpu) *dec = decrementer_clockevent; dec->cpumask = cpumask_of(cpu); - printk(KERN_DEBUG "clockevent: %s mult[%x] shift[%d] cpu[%d]\n", - dec->name, dec->mult, dec->shift, cpu); + printk_once(KERN_DEBUG "clockevent: %s mult[%x] shift[%d] cpu[%d]\n", + dec->name, dec->mult, dec->shift, cpu); clockevents_register_device(dec); } -- cgit v0.10.2 From 577cd7584cf5199f1ea22cca0ad1fa129a98effa Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 9 Feb 2010 04:24:46 +0000 Subject: sh: extend INTC with struct intc_hw_desc This patch updates the INTC code by moving all vectors, groups and registers from struct intc_desc to struct intc_hw_desc. The idea is that INTC tables should go from using the macro(s) DECLARE_INTC_DESC..() only to using struct intc_desc with name and hw initialized using the macro INTC_HW_DESC(). This move makes it easy to initialize an extended struct intc_desc in the future. Signed-off-by: Magnus Damm Signed-off-by: Paul Mundt diff --git a/drivers/sh/intc.c b/drivers/sh/intc.c index d5d7f23..5ec0fff 100644 --- a/drivers/sh/intc.c +++ b/drivers/sh/intc.c @@ -400,11 +400,11 @@ static unsigned int __init intc_get_reg(struct intc_desc_int *d, static intc_enum __init intc_grp_id(struct intc_desc *desc, intc_enum enum_id) { - struct intc_group *g = desc->groups; + struct intc_group *g = desc->hw.groups; unsigned int i, j; - for (i = 0; g && enum_id && i < desc->nr_groups; i++) { - g = desc->groups + i; + for (i = 0; g && enum_id && i < desc->hw.nr_groups; i++) { + g = desc->hw.groups + i; for (j = 0; g->enum_ids[j]; j++) { if (g->enum_ids[j] != enum_id) @@ -421,12 +421,12 @@ static unsigned int __init intc_mask_data(struct intc_desc *desc, struct intc_desc_int *d, intc_enum enum_id, int do_grps) { - struct intc_mask_reg *mr = desc->mask_regs; + struct intc_mask_reg *mr = desc->hw.mask_regs; unsigned int i, j, fn, mode; unsigned long reg_e, reg_d; - for (i = 0; mr && enum_id && i < desc->nr_mask_regs; i++) { - mr = desc->mask_regs + i; + for (i = 0; mr && enum_id && i < desc->hw.nr_mask_regs; i++) { + mr = desc->hw.mask_regs + i; for (j = 0; j < ARRAY_SIZE(mr->enum_ids); j++) { if (mr->enum_ids[j] != enum_id) @@ -469,12 +469,12 @@ static unsigned int __init intc_prio_data(struct intc_desc *desc, struct intc_desc_int *d, intc_enum enum_id, int do_grps) { - struct intc_prio_reg *pr = desc->prio_regs; + struct intc_prio_reg *pr = desc->hw.prio_regs; unsigned int i, j, fn, mode, bit; unsigned long reg_e, reg_d; - for (i = 0; pr && enum_id && i < desc->nr_prio_regs; i++) { - pr = desc->prio_regs + i; + for (i = 0; pr && enum_id && i < desc->hw.nr_prio_regs; i++) { + pr = desc->hw.prio_regs + i; for (j = 0; j < ARRAY_SIZE(pr->enum_ids); j++) { if (pr->enum_ids[j] != enum_id) @@ -517,12 +517,12 @@ static unsigned int __init intc_ack_data(struct intc_desc *desc, struct intc_desc_int *d, intc_enum enum_id) { - struct intc_mask_reg *mr = desc->ack_regs; + struct intc_mask_reg *mr = desc->hw.ack_regs; unsigned int i, j, fn, mode; unsigned long reg_e, reg_d; - for (i = 0; mr && enum_id && i < desc->nr_ack_regs; i++) { - mr = desc->ack_regs + i; + for (i = 0; mr && enum_id && i < desc->hw.nr_ack_regs; i++) { + mr = desc->hw.ack_regs + i; for (j = 0; j < ARRAY_SIZE(mr->enum_ids); j++) { if (mr->enum_ids[j] != enum_id) @@ -549,11 +549,11 @@ static unsigned int __init intc_sense_data(struct intc_desc *desc, struct intc_desc_int *d, intc_enum enum_id) { - struct intc_sense_reg *sr = desc->sense_regs; + struct intc_sense_reg *sr = desc->hw.sense_regs; unsigned int i, j, fn, bit; - for (i = 0; sr && enum_id && i < desc->nr_sense_regs; i++) { - sr = desc->sense_regs + i; + for (i = 0; sr && enum_id && i < desc->hw.nr_sense_regs; i++) { + sr = desc->hw.sense_regs + i; for (j = 0; j < ARRAY_SIZE(sr->enum_ids); j++) { if (sr->enum_ids[j] != enum_id) @@ -656,7 +656,7 @@ static void __init intc_register_irq(struct intc_desc *desc, /* irq should be disabled by default */ d->chip.mask(irq); - if (desc->ack_regs) + if (desc->hw.ack_regs) ack_handle[irq] = intc_ack_data(desc, d, enum_id); } @@ -684,6 +684,7 @@ static void intc_redirect_irq(unsigned int irq, struct irq_desc *desc) void __init register_intc_controller(struct intc_desc *desc) { unsigned int i, k, smp; + struct intc_hw_desc *hw = &desc->hw; struct intc_desc_int *d; d = kzalloc(sizeof(*d), GFP_NOWAIT); @@ -691,10 +692,10 @@ void __init register_intc_controller(struct intc_desc *desc) INIT_LIST_HEAD(&d->list); list_add(&d->list, &intc_list); - d->nr_reg = desc->mask_regs ? desc->nr_mask_regs * 2 : 0; - d->nr_reg += desc->prio_regs ? desc->nr_prio_regs * 2 : 0; - d->nr_reg += desc->sense_regs ? desc->nr_sense_regs : 0; - d->nr_reg += desc->ack_regs ? desc->nr_ack_regs : 0; + d->nr_reg = hw->mask_regs ? hw->nr_mask_regs * 2 : 0; + d->nr_reg += hw->prio_regs ? hw->nr_prio_regs * 2 : 0; + d->nr_reg += hw->sense_regs ? hw->nr_sense_regs : 0; + d->nr_reg += hw->ack_regs ? hw->nr_ack_regs : 0; d->reg = kzalloc(d->nr_reg * sizeof(*d->reg), GFP_NOWAIT); #ifdef CONFIG_SMP @@ -702,30 +703,31 @@ void __init register_intc_controller(struct intc_desc *desc) #endif k = 0; - if (desc->mask_regs) { - for (i = 0; i < desc->nr_mask_regs; i++) { - smp = IS_SMP(desc->mask_regs[i]); - k += save_reg(d, k, desc->mask_regs[i].set_reg, smp); - k += save_reg(d, k, desc->mask_regs[i].clr_reg, smp); + if (hw->mask_regs) { + for (i = 0; i < hw->nr_mask_regs; i++) { + smp = IS_SMP(hw->mask_regs[i]); + k += save_reg(d, k, hw->mask_regs[i].set_reg, smp); + k += save_reg(d, k, hw->mask_regs[i].clr_reg, smp); } } - if (desc->prio_regs) { - d->prio = kzalloc(desc->nr_vectors * sizeof(*d->prio), GFP_NOWAIT); + if (hw->prio_regs) { + d->prio = kzalloc(hw->nr_vectors * sizeof(*d->prio), + GFP_NOWAIT); - for (i = 0; i < desc->nr_prio_regs; i++) { - smp = IS_SMP(desc->prio_regs[i]); - k += save_reg(d, k, desc->prio_regs[i].set_reg, smp); - k += save_reg(d, k, desc->prio_regs[i].clr_reg, smp); + for (i = 0; i < hw->nr_prio_regs; i++) { + smp = IS_SMP(hw->prio_regs[i]); + k += save_reg(d, k, hw->prio_regs[i].set_reg, smp); + k += save_reg(d, k, hw->prio_regs[i].clr_reg, smp); } } - if (desc->sense_regs) { - d->sense = kzalloc(desc->nr_vectors * sizeof(*d->sense), GFP_NOWAIT); + if (hw->sense_regs) { + d->sense = kzalloc(hw->nr_vectors * sizeof(*d->sense), + GFP_NOWAIT); - for (i = 0; i < desc->nr_sense_regs; i++) { - k += save_reg(d, k, desc->sense_regs[i].reg, 0); - } + for (i = 0; i < hw->nr_sense_regs; i++) + k += save_reg(d, k, hw->sense_regs[i].reg, 0); } d->chip.name = desc->name; @@ -738,9 +740,9 @@ void __init register_intc_controller(struct intc_desc *desc) d->chip.set_type = intc_set_sense; d->chip.set_wake = intc_set_wake; - if (desc->ack_regs) { - for (i = 0; i < desc->nr_ack_regs; i++) - k += save_reg(d, k, desc->ack_regs[i].set_reg, 0); + if (hw->ack_regs) { + for (i = 0; i < hw->nr_ack_regs; i++) + k += save_reg(d, k, hw->ack_regs[i].set_reg, 0); d->chip.mask_ack = intc_mask_ack; } @@ -748,8 +750,8 @@ void __init register_intc_controller(struct intc_desc *desc) BUG_ON(k > 256); /* _INTC_ADDR_E() and _INTC_ADDR_D() are 8 bits */ /* register the vectors one by one */ - for (i = 0; i < desc->nr_vectors; i++) { - struct intc_vect *vect = desc->vectors + i; + for (i = 0; i < hw->nr_vectors; i++) { + struct intc_vect *vect = hw->vectors + i; unsigned int irq = evt2irq(vect->vect); struct irq_desc *irq_desc; @@ -764,8 +766,8 @@ void __init register_intc_controller(struct intc_desc *desc) intc_register_irq(desc, d, vect->enum_id, irq); - for (k = i + 1; k < desc->nr_vectors; k++) { - struct intc_vect *vect2 = desc->vectors + k; + for (k = i + 1; k < hw->nr_vectors; k++) { + struct intc_vect *vect2 = hw->vectors + k; unsigned int irq2 = evt2irq(vect2->vect); if (vect->enum_id != vect2->enum_id) diff --git a/include/linux/sh_intc.h b/include/linux/sh_intc.h index 4ef246f..7b37526 100644 --- a/include/linux/sh_intc.h +++ b/include/linux/sh_intc.h @@ -45,7 +45,7 @@ struct intc_sense_reg { #define INTC_SMP(stride, nr) #endif -struct intc_desc { +struct intc_hw_desc { struct intc_vect *vectors; unsigned int nr_vectors; struct intc_group *groups; @@ -56,29 +56,38 @@ struct intc_desc { unsigned int nr_prio_regs; struct intc_sense_reg *sense_regs; unsigned int nr_sense_regs; - char *name; struct intc_mask_reg *ack_regs; unsigned int nr_ack_regs; }; #define _INTC_ARRAY(a) a, sizeof(a)/sizeof(*a) +#define INTC_HW_DESC(vectors, groups, mask_regs, \ + prio_regs, sense_regs, ack_regs) \ +{ \ + _INTC_ARRAY(vectors), _INTC_ARRAY(groups), \ + _INTC_ARRAY(mask_regs), _INTC_ARRAY(prio_regs), \ + _INTC_ARRAY(sense_regs), _INTC_ARRAY(ack_regs), \ +} + +struct intc_desc { + char *name; + struct intc_hw_desc hw; +}; + #define DECLARE_INTC_DESC(symbol, chipname, vectors, groups, \ mask_regs, prio_regs, sense_regs) \ struct intc_desc symbol __initdata = { \ - _INTC_ARRAY(vectors), _INTC_ARRAY(groups), \ - _INTC_ARRAY(mask_regs), _INTC_ARRAY(prio_regs), \ - _INTC_ARRAY(sense_regs), \ - chipname, \ + .name = chipname, \ + .hw = INTC_HW_DESC(vectors, groups, mask_regs, \ + prio_regs, sense_regs, NULL), \ } #define DECLARE_INTC_DESC_ACK(symbol, chipname, vectors, groups, \ mask_regs, prio_regs, sense_regs, ack_regs) \ struct intc_desc symbol __initdata = { \ - _INTC_ARRAY(vectors), _INTC_ARRAY(groups), \ - _INTC_ARRAY(mask_regs), _INTC_ARRAY(prio_regs), \ - _INTC_ARRAY(sense_regs), \ - chipname, \ - _INTC_ARRAY(ack_regs), \ + .name = chipname, \ + .hw = INTC_HW_DESC(vectors, groups, mask_regs, \ + prio_regs, sense_regs, ack_regs), \ } void __init register_intc_controller(struct intc_desc *desc); -- cgit v0.10.2 From d519095344fda705c9840a579acf6aa6205c37cc Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 9 Feb 2010 04:29:22 +0000 Subject: sh: extend INTC with force_enable Extend the shared INTC code with force_enable support to allow keeping mask bits statically enabled. Needed by upcoming INTC SDHI patches that mux together a bunch of vectors to a single linux interrupt which is masked by a priority register, but needs individual mask bits constantly enabled. Signed-off-by: Magnus Damm Signed-off-by: Paul Mundt diff --git a/drivers/sh/intc.c b/drivers/sh/intc.c index 5ec0fff..ccee189 100644 --- a/drivers/sh/intc.c +++ b/drivers/sh/intc.c @@ -259,6 +259,43 @@ static void intc_disable(unsigned int irq) } } +static void (*intc_enable_noprio_fns[])(unsigned long addr, + unsigned long handle, + void (*fn)(unsigned long, + unsigned long, + unsigned long), + unsigned int irq) = { + [MODE_ENABLE_REG] = intc_mode_field, + [MODE_MASK_REG] = intc_mode_zero, + [MODE_DUAL_REG] = intc_mode_field, + [MODE_PRIO_REG] = intc_mode_field, + [MODE_PCLR_REG] = intc_mode_field, +}; + +static void intc_enable_disable(struct intc_desc_int *d, + unsigned long handle, int do_enable) +{ + unsigned long addr; + unsigned int cpu; + void (*fn)(unsigned long, unsigned long, + void (*)(unsigned long, unsigned long, unsigned long), + unsigned int); + + if (do_enable) { + for (cpu = 0; cpu < SMP_NR(d, _INTC_ADDR_E(handle)); cpu++) { + addr = INTC_REG(d, _INTC_ADDR_E(handle), cpu); + fn = intc_enable_noprio_fns[_INTC_MODE(handle)]; + fn(addr, handle, intc_reg_fns[_INTC_FN(handle)], 0); + } + } else { + for (cpu = 0; cpu < SMP_NR(d, _INTC_ADDR_D(handle)); cpu++) { + addr = INTC_REG(d, _INTC_ADDR_D(handle), cpu); + fn = intc_disable_fns[_INTC_MODE(handle)]; + fn(addr, handle, intc_reg_fns[_INTC_FN(handle)], 0); + } + } +} + static int intc_set_wake(unsigned int irq, unsigned int on) { return 0; /* allow wakeup, but setup hardware in intc_suspend() */ @@ -417,19 +454,21 @@ static intc_enum __init intc_grp_id(struct intc_desc *desc, return 0; } -static unsigned int __init intc_mask_data(struct intc_desc *desc, - struct intc_desc_int *d, - intc_enum enum_id, int do_grps) +static unsigned int __init _intc_mask_data(struct intc_desc *desc, + struct intc_desc_int *d, + intc_enum enum_id, + unsigned int *reg_idx, + unsigned int *fld_idx) { struct intc_mask_reg *mr = desc->hw.mask_regs; - unsigned int i, j, fn, mode; + unsigned int fn, mode; unsigned long reg_e, reg_d; - for (i = 0; mr && enum_id && i < desc->hw.nr_mask_regs; i++) { - mr = desc->hw.mask_regs + i; + while (mr && enum_id && *reg_idx < desc->hw.nr_mask_regs) { + mr = desc->hw.mask_regs + *reg_idx; - for (j = 0; j < ARRAY_SIZE(mr->enum_ids); j++) { - if (mr->enum_ids[j] != enum_id) + for (; *fld_idx < ARRAY_SIZE(mr->enum_ids); (*fld_idx)++) { + if (mr->enum_ids[*fld_idx] != enum_id) continue; if (mr->set_reg && mr->clr_reg) { @@ -455,29 +494,49 @@ static unsigned int __init intc_mask_data(struct intc_desc *desc, intc_get_reg(d, reg_e), intc_get_reg(d, reg_d), 1, - (mr->reg_width - 1) - j); + (mr->reg_width - 1) - *fld_idx); } + + *fld_idx = 0; + (*reg_idx)++; } + return 0; +} + +static unsigned int __init intc_mask_data(struct intc_desc *desc, + struct intc_desc_int *d, + intc_enum enum_id, int do_grps) +{ + unsigned int i = 0; + unsigned int j = 0; + unsigned int ret; + + ret = _intc_mask_data(desc, d, enum_id, &i, &j); + if (ret) + return ret; + if (do_grps) return intc_mask_data(desc, d, intc_grp_id(desc, enum_id), 0); return 0; } -static unsigned int __init intc_prio_data(struct intc_desc *desc, - struct intc_desc_int *d, - intc_enum enum_id, int do_grps) +static unsigned int __init _intc_prio_data(struct intc_desc *desc, + struct intc_desc_int *d, + intc_enum enum_id, + unsigned int *reg_idx, + unsigned int *fld_idx) { struct intc_prio_reg *pr = desc->hw.prio_regs; - unsigned int i, j, fn, mode, bit; + unsigned int fn, n, mode, bit; unsigned long reg_e, reg_d; - for (i = 0; pr && enum_id && i < desc->hw.nr_prio_regs; i++) { - pr = desc->hw.prio_regs + i; + while (pr && enum_id && *reg_idx < desc->hw.nr_prio_regs) { + pr = desc->hw.prio_regs + *reg_idx; - for (j = 0; j < ARRAY_SIZE(pr->enum_ids); j++) { - if (pr->enum_ids[j] != enum_id) + for (; *fld_idx < ARRAY_SIZE(pr->enum_ids); (*fld_idx)++) { + if (pr->enum_ids[*fld_idx] != enum_id) continue; if (pr->set_reg && pr->clr_reg) { @@ -495,24 +554,69 @@ static unsigned int __init intc_prio_data(struct intc_desc *desc, } fn += (pr->reg_width >> 3) - 1; + n = *fld_idx + 1; - BUG_ON((j + 1) * pr->field_width > pr->reg_width); + BUG_ON(n * pr->field_width > pr->reg_width); - bit = pr->reg_width - ((j + 1) * pr->field_width); + bit = pr->reg_width - (n * pr->field_width); return _INTC_MK(fn, mode, intc_get_reg(d, reg_e), intc_get_reg(d, reg_d), pr->field_width, bit); } + + *fld_idx = 0; + (*reg_idx)++; } + return 0; +} + +static unsigned int __init intc_prio_data(struct intc_desc *desc, + struct intc_desc_int *d, + intc_enum enum_id, int do_grps) +{ + unsigned int i = 0; + unsigned int j = 0; + unsigned int ret; + + ret = _intc_prio_data(desc, d, enum_id, &i, &j); + if (ret) + return ret; + if (do_grps) return intc_prio_data(desc, d, intc_grp_id(desc, enum_id), 0); return 0; } +static void __init intc_enable_disable_enum(struct intc_desc *desc, + struct intc_desc_int *d, + intc_enum enum_id, int enable) +{ + unsigned int i, j, data; + + /* go through and enable/disable all mask bits */ + i = j = 0; + do { + data = _intc_mask_data(desc, d, enum_id, &i, &j); + if (data) + intc_enable_disable(d, data, enable); + j++; + } while (data); + + /* go through and enable/disable all priority fields */ + i = j = 0; + do { + data = _intc_prio_data(desc, d, enum_id, &i, &j); + if (data) + intc_enable_disable(d, data, enable); + + j++; + } while (data); +} + static unsigned int __init intc_ack_data(struct intc_desc *desc, struct intc_desc_int *d, intc_enum enum_id) @@ -747,6 +851,11 @@ void __init register_intc_controller(struct intc_desc *desc) d->chip.mask_ack = intc_mask_ack; } + + /* disable bits matching force_enable before registering irqs */ + if (desc->force_enable) + intc_enable_disable_enum(desc, d, desc->force_enable, 0); + BUG_ON(k > 256); /* _INTC_ADDR_E() and _INTC_ADDR_D() are 8 bits */ /* register the vectors one by one */ @@ -792,6 +901,10 @@ void __init register_intc_controller(struct intc_desc *desc) set_irq_data(irq2, (void *)irq); } } + + /* enable bits matching force_enable after registering irqs */ + if (desc->force_enable) + intc_enable_disable_enum(desc, d, desc->force_enable, 1); } static int intc_suspend(struct sys_device *dev, pm_message_t state) diff --git a/include/linux/sh_intc.h b/include/linux/sh_intc.h index 7b37526..66b4b0c 100644 --- a/include/linux/sh_intc.h +++ b/include/linux/sh_intc.h @@ -71,6 +71,7 @@ struct intc_hw_desc { struct intc_desc { char *name; + intc_enum force_enable; struct intc_hw_desc hw; }; -- cgit v0.10.2 From 8d9adabac3a3a3742c7a47da9e456108a0fd2efb Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 9 Feb 2010 06:49:48 +0000 Subject: sh: sh7722/Migo-R SDHI vector merge Merge the SDHI vectors in the sh7722 INTC table and update the SDHI platform data for Migo-R. Signed-off-by: Magnus Damm Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-migor/setup.c b/arch/sh/boards/mach-migor/setup.c index ed2eeeb..d1e51a9 100644 --- a/arch/sh/boards/mach-migor/setup.c +++ b/arch/sh/boards/mach-migor/setup.c @@ -397,7 +397,7 @@ static struct resource sdhi_cn9_resources[] = { .flags = IORESOURCE_MEM, }, [1] = { - .start = 101, + .start = 100, .flags = IORESOURCE_IRQ, }, }; diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7722.c b/arch/sh/kernel/cpu/sh4a/setup-sh7722.c index b5335b5..0949135 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7722.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7722.c @@ -446,6 +446,7 @@ void __init plat_early_device_setup(void) enum { UNUSED=0, + ENABLED, /* interrupt sources */ IRQ0, IRQ1, IRQ2, IRQ3, IRQ4, IRQ5, IRQ6, IRQ7, @@ -461,7 +462,6 @@ enum { SCIF0, SCIF1, SCIF2, SIOF0, SIOF1, SIO, FLCTL_FLSTEI, FLCTL_FLENDI, FLCTL_FLTREQ0I, FLCTL_FLTREQ1I, I2C_ALI, I2C_TACKI, I2C_WAITI, I2C_DTEI, - SDHI0, SDHI1, SDHI2, SDHI3, CMT, TSIF, SIU, TWODG, TMU0, TMU1, TMU2, IRDA, JPU, LCDC, @@ -494,8 +494,8 @@ static struct intc_vect vectors[] __initdata = { INTC_VECT(FLCTL_FLTREQ0I, 0xdc0), INTC_VECT(FLCTL_FLTREQ1I, 0xde0), INTC_VECT(I2C_ALI, 0xe00), INTC_VECT(I2C_TACKI, 0xe20), INTC_VECT(I2C_WAITI, 0xe40), INTC_VECT(I2C_DTEI, 0xe60), - INTC_VECT(SDHI0, 0xe80), INTC_VECT(SDHI1, 0xea0), - INTC_VECT(SDHI2, 0xec0), INTC_VECT(SDHI3, 0xee0), + INTC_VECT(SDHI, 0xe80), INTC_VECT(SDHI, 0xea0), + INTC_VECT(SDHI, 0xec0), INTC_VECT(SDHI, 0xee0), INTC_VECT(CMT, 0xf00), INTC_VECT(TSIF, 0xf20), INTC_VECT(SIU, 0xf80), INTC_VECT(TWODG, 0xfa0), INTC_VECT(TMU0, 0x400), INTC_VECT(TMU1, 0x420), @@ -513,7 +513,6 @@ static struct intc_group groups[] __initdata = { INTC_GROUP(FLCTL, FLCTL_FLSTEI, FLCTL_FLENDI, FLCTL_FLTREQ0I, FLCTL_FLTREQ1I), INTC_GROUP(I2C, I2C_ALI, I2C_TACKI, I2C_WAITI, I2C_DTEI), - INTC_GROUP(SDHI, SDHI0, SDHI1, SDHI2, SDHI3), }; static struct intc_mask_reg mask_registers[] __initdata = { @@ -535,7 +534,7 @@ static struct intc_mask_reg mask_registers[] __initdata = { { I2C_DTEI, I2C_WAITI, I2C_TACKI, I2C_ALI, FLCTL_FLTREQ1I, FLCTL_FLTREQ0I, FLCTL_FLENDI, FLCTL_FLSTEI } }, { 0xa40800a0, 0xa40800e0, 8, /* IMR8 / IMCR8 */ - { SDHI3, SDHI2, SDHI1, SDHI0, 0, 0, TWODG, SIU } }, + { 0, 0, ENABLED, ENABLED, 0, 0, TWODG, SIU } }, { 0xa40800a4, 0xa40800e4, 8, /* IMR9 / IMCR9 */ { 0, 0, 0, CMT, 0, USB_USBI1, USB_USBI0, } }, { 0xa40800a8, 0xa40800e8, 8, /* IMR10 / IMCR10 */ @@ -573,9 +572,12 @@ static struct intc_mask_reg ack_registers[] __initdata = { { IRQ0, IRQ1, IRQ2, IRQ3, IRQ4, IRQ5, IRQ6, IRQ7 } }, }; -static DECLARE_INTC_DESC_ACK(intc_desc, "sh7722", vectors, groups, - mask_registers, prio_registers, sense_registers, - ack_registers); +static struct intc_desc intc_desc __initdata = { + .name = "sh7722", + .force_enable = ENABLED, + .hw = INTC_HW_DESC(vectors, groups, mask_registers, + prio_registers, sense_registers, ack_registers), +}; void __init plat_irq_setup(void) { -- cgit v0.10.2 From e3e80046e0ce2a595e607a62b650a9b4efea6558 Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 9 Feb 2010 06:49:56 +0000 Subject: sh: sh7723/AP325 SDHI vector merge Merge the SDHI vectors in the sh7723 INTC table and update the SDHI platform data for AP325. Signed-off-by: Magnus Damm Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-ap325rxa/setup.c b/arch/sh/boards/mach-ap325rxa/setup.c index 27277cb..b598069 100644 --- a/arch/sh/boards/mach-ap325rxa/setup.c +++ b/arch/sh/boards/mach-ap325rxa/setup.c @@ -420,7 +420,7 @@ static struct resource sdhi0_cn3_resources[] = { .flags = IORESOURCE_MEM, }, [1] = { - .start = 101, + .start = 100, .flags = IORESOURCE_IRQ, }, }; @@ -443,7 +443,7 @@ static struct resource sdhi1_cn7_resources[] = { .flags = IORESOURCE_MEM, }, [1] = { - .start = 24, + .start = 23, .flags = IORESOURCE_IRQ, }, }; diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7723.c b/arch/sh/kernel/cpu/sh4a/setup-sh7723.c index 9e8620e..7b90ebc 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7723.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7723.c @@ -601,6 +601,7 @@ void l2_cache_init(void) enum { UNUSED=0, + ENABLED, /* interrupt sources */ IRQ0, IRQ1, IRQ2, IRQ3, IRQ4, IRQ5, IRQ6, IRQ7, @@ -623,7 +624,6 @@ enum { SCIFA_SCIFA1, FLCTL_FLSTEI,FLCTL_FLTENDI,FLCTL_FLTREQ0I,FLCTL_FLTREQ1I, I2C_ALI,I2C_TACKI,I2C_WAITI,I2C_DTEI, - SDHI0_SDHII0,SDHI0_SDHII1,SDHI0_SDHII2, CMT_CMTI, TSIF_TSIFI, SIU_SIUI, @@ -631,7 +631,6 @@ enum { TMU0_TUNI0, TMU0_TUNI1, TMU0_TUNI2, IRDA_IRDAI, ATAPI_ATAPII, - SDHI1_SDHII0,SDHI1_SDHII1,SDHI1_SDHII2, VEU2H1_VEU2HI, LCDC_LCDCI, TMU1_TUNI0,TMU1_TUNI1,TMU1_TUNI2, @@ -702,9 +701,9 @@ static struct intc_vect vectors[] __initdata = { INTC_VECT(I2C_WAITI,0xE40), INTC_VECT(I2C_DTEI,0xE60), - INTC_VECT(SDHI0_SDHII0,0xE80), - INTC_VECT(SDHI0_SDHII1,0xEA0), - INTC_VECT(SDHI0_SDHII2,0xEC0), + INTC_VECT(SDHI0, 0xE80), + INTC_VECT(SDHI0, 0xEA0), + INTC_VECT(SDHI0, 0xEC0), INTC_VECT(CMT_CMTI,0xF00), INTC_VECT(TSIF_TSIFI,0xF20), @@ -718,9 +717,9 @@ static struct intc_vect vectors[] __initdata = { INTC_VECT(IRDA_IRDAI,0x480), INTC_VECT(ATAPI_ATAPII,0x4A0), - INTC_VECT(SDHI1_SDHII0,0x4E0), - INTC_VECT(SDHI1_SDHII1,0x500), - INTC_VECT(SDHI1_SDHII2,0x520), + INTC_VECT(SDHI1, 0x4E0), + INTC_VECT(SDHI1, 0x500), + INTC_VECT(SDHI1, 0x520), INTC_VECT(VEU2H1_VEU2HI,0x560), INTC_VECT(LCDC_LCDCI,0x580), @@ -739,15 +738,14 @@ static struct intc_group groups[] __initdata = { INTC_GROUP(FLCTL,FLCTL_FLSTEI,FLCTL_FLTENDI,FLCTL_FLTREQ0I,FLCTL_FLTREQ1I), INTC_GROUP(I2C,I2C_ALI,I2C_TACKI,I2C_WAITI,I2C_DTEI), INTC_GROUP(_2DG, _2DG_TRI,_2DG_INI,_2DG_CEI), - INTC_GROUP(SDHI1, SDHI1_SDHII0,SDHI1_SDHII1,SDHI1_SDHII2), INTC_GROUP(RTC, RTC_ATI,RTC_PRI,RTC_CUI), INTC_GROUP(DMAC1B, DMAC1B_DEI4,DMAC1B_DEI5,DMAC1B_DADERR), - INTC_GROUP(SDHI0,SDHI0_SDHII0,SDHI0_SDHII1,SDHI0_SDHII2), }; static struct intc_mask_reg mask_registers[] __initdata = { { 0xa4080080, 0xa40800c0, 8, /* IMR0 / IMCR0 */ - { 0, TMU1_TUNI2,TMU1_TUNI1,TMU1_TUNI0,0,SDHI1_SDHII2,SDHI1_SDHII1,SDHI1_SDHII0} }, + { 0, TMU1_TUNI2, TMU1_TUNI1, TMU1_TUNI0, + 0, 0, ENABLED, ENABLED } }, { 0xa4080084, 0xa40800c4, 8, /* IMR1 / IMCR1 */ { VIO_VOUI, VIO_VEU2HI,VIO_BEUI,VIO_CEUI,DMAC0A_DEI3,DMAC0A_DEI2,DMAC0A_DEI1,DMAC0A_DEI0 } }, { 0xa4080088, 0xa40800c8, 8, /* IMR2 / IMCR2 */ @@ -764,7 +762,8 @@ static struct intc_mask_reg mask_registers[] __initdata = { { I2C_DTEI, I2C_WAITI, I2C_TACKI, I2C_ALI, FLCTL_FLTREQ1I, FLCTL_FLTREQ0I, FLCTL_FLTENDI, FLCTL_FLSTEI } }, { 0xa40800a0, 0xa40800e0, 8, /* IMR8 / IMCR8 */ - { 0,SDHI0_SDHII2,SDHI0_SDHII1,SDHI0_SDHII0,0,0,SCIFA_SCIFA2,SIU_SIUI } }, + { 0, 0, ENABLED, ENABLED, + 0, 0, SCIFA_SCIFA2, SIU_SIUI } }, { 0xa40800a4, 0xa40800e4, 8, /* IMR9 / IMCR9 */ { 0, 0, 0, CMT_CMTI, 0, 0, USB_USI0,0 } }, { 0xa40800a8, 0xa40800e8, 8, /* IMR10 / IMCR10 */ @@ -804,9 +803,12 @@ static struct intc_mask_reg ack_registers[] __initdata = { { IRQ0, IRQ1, IRQ2, IRQ3, IRQ4, IRQ5, IRQ6, IRQ7 } }, }; -static DECLARE_INTC_DESC_ACK(intc_desc, "sh7723", vectors, groups, - mask_registers, prio_registers, sense_registers, - ack_registers); +static struct intc_desc intc_desc __initdata = { + .name = "sh7723", + .force_enable = ENABLED, + .hw = INTC_HW_DESC(vectors, groups, mask_registers, + prio_registers, sense_registers, ack_registers), +}; void __init plat_irq_setup(void) { -- cgit v0.10.2 From 3844eadcfd2ba975110e3ca8479efa8c093129ce Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 9 Feb 2010 06:50:04 +0000 Subject: sh: sh7724/Ecovec24/KFR2R09/MS7724SE SDHI vector merge Merge the SDHI vectors in the sh7724 INTC table and update the SDHI platform data for Ecovec24, KFR2R09 and MS7724SE. Signed-off-by: Magnus Damm Signed-off-by: Paul Mundt diff --git a/arch/sh/boards/mach-ecovec24/setup.c b/arch/sh/boards/mach-ecovec24/setup.c index 1135c3b..8e3b05f 100644 --- a/arch/sh/boards/mach-ecovec24/setup.c +++ b/arch/sh/boards/mach-ecovec24/setup.c @@ -453,7 +453,7 @@ static struct resource sdhi0_resources[] = { .flags = IORESOURCE_MEM, }, [1] = { - .start = 101, + .start = 100, .flags = IORESOURCE_IRQ, }, }; @@ -489,7 +489,7 @@ static struct resource sdhi1_resources[] = { .flags = IORESOURCE_MEM, }, [1] = { - .start = 24, + .start = 23, .flags = IORESOURCE_IRQ, }, }; diff --git a/arch/sh/boards/mach-kfr2r09/setup.c b/arch/sh/boards/mach-kfr2r09/setup.c index ca9e82d..b2cd0ed 100644 --- a/arch/sh/boards/mach-kfr2r09/setup.c +++ b/arch/sh/boards/mach-kfr2r09/setup.c @@ -351,7 +351,7 @@ static struct resource kfr2r09_sh_sdhi0_resources[] = { .flags = IORESOURCE_MEM, }, [1] = { - .start = 101, + .start = 100, .flags = IORESOURCE_IRQ, }, }; diff --git a/arch/sh/boards/mach-se/7724/setup.c b/arch/sh/boards/mach-se/7724/setup.c index c7dbbec..50e4d15 100644 --- a/arch/sh/boards/mach-se/7724/setup.c +++ b/arch/sh/boards/mach-se/7724/setup.c @@ -451,7 +451,7 @@ static struct resource sdhi0_cn7_resources[] = { .flags = IORESOURCE_MEM, }, [1] = { - .start = 101, + .start = 100, .flags = IORESOURCE_IRQ, }, }; @@ -474,7 +474,7 @@ static struct resource sdhi1_cn8_resources[] = { .flags = IORESOURCE_MEM, }, [1] = { - .start = 24, + .start = 23, .flags = IORESOURCE_IRQ, }, }; diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7724.c b/arch/sh/kernel/cpu/sh4a/setup-sh7724.c index aa0f6e9..c0815a4 100644 --- a/arch/sh/kernel/cpu/sh4a/setup-sh7724.c +++ b/arch/sh/kernel/cpu/sh4a/setup-sh7724.c @@ -723,6 +723,7 @@ void l2_cache_init(void) enum { UNUSED = 0, + ENABLED, /* interrupt sources */ IRQ0, IRQ1, IRQ2, IRQ3, IRQ4, IRQ5, IRQ6, IRQ7, @@ -751,14 +752,12 @@ enum { ETHI, I2C1_ALI, I2C1_TACKI, I2C1_WAITI, I2C1_DTEI, I2C0_ALI, I2C0_TACKI, I2C0_WAITI, I2C0_DTEI, - SDHI0_SDHII0, SDHI0_SDHII1, SDHI0_SDHII2, SDHI0_SDHII3, CMT, TSIF, FSI, SCIFA5, TMU0_TUNI0, TMU0_TUNI1, TMU0_TUNI2, IRDA, - SDHI1_SDHII0, SDHI1_SDHII1, SDHI1_SDHII2, JPU, _2DDMAC, MMC_MMC2I, MMC_MMC3I, @@ -840,10 +839,10 @@ static struct intc_vect vectors[] __initdata = { INTC_VECT(I2C0_WAITI, 0xE40), INTC_VECT(I2C0_DTEI, 0xE60), - INTC_VECT(SDHI0_SDHII0, 0xE80), - INTC_VECT(SDHI0_SDHII1, 0xEA0), - INTC_VECT(SDHI0_SDHII2, 0xEC0), - INTC_VECT(SDHI0_SDHII3, 0xEE0), + INTC_VECT(SDHI0, 0xE80), + INTC_VECT(SDHI0, 0xEA0), + INTC_VECT(SDHI0, 0xEC0), + INTC_VECT(SDHI0, 0xEE0), INTC_VECT(CMT, 0xF00), INTC_VECT(TSIF, 0xF20), @@ -856,9 +855,9 @@ static struct intc_vect vectors[] __initdata = { INTC_VECT(IRDA, 0x480), - INTC_VECT(SDHI1_SDHII0, 0x4E0), - INTC_VECT(SDHI1_SDHII1, 0x500), - INTC_VECT(SDHI1_SDHII2, 0x520), + INTC_VECT(SDHI1, 0x4E0), + INTC_VECT(SDHI1, 0x500), + INTC_VECT(SDHI1, 0x520), INTC_VECT(JPU, 0x560), INTC_VECT(_2DDMAC, 0x4A0), @@ -884,8 +883,6 @@ static struct intc_group groups[] __initdata = { INTC_GROUP(DMAC0B, DMAC0B_DEI4, DMAC0B_DEI5, DMAC0B_DADERR), INTC_GROUP(I2C0, I2C0_ALI, I2C0_TACKI, I2C0_WAITI, I2C0_DTEI), INTC_GROUP(I2C1, I2C1_ALI, I2C1_TACKI, I2C1_WAITI, I2C1_DTEI), - INTC_GROUP(SDHI0, SDHI0_SDHII0, SDHI0_SDHII1, SDHI0_SDHII2, SDHI0_SDHII3), - INTC_GROUP(SDHI1, SDHI1_SDHII0, SDHI1_SDHII1, SDHI1_SDHII2), INTC_GROUP(SPU, SPU_SPUI0, SPU_SPUI1), INTC_GROUP(MMCIF, MMC_MMC2I, MMC_MMC3I), }; @@ -893,7 +890,7 @@ static struct intc_group groups[] __initdata = { static struct intc_mask_reg mask_registers[] __initdata = { { 0xa4080080, 0xa40800c0, 8, /* IMR0 / IMCR0 */ { 0, TMU1_TUNI2, TMU1_TUNI1, TMU1_TUNI0, - 0, SDHI1_SDHII2, SDHI1_SDHII1, SDHI1_SDHII0 } }, + 0, 0, ENABLED, ENABLED } }, { 0xa4080084, 0xa40800c4, 8, /* IMR1 / IMCR1 */ { VIO_VOU, VIO_VEU1, VIO_BEU0, VIO_CEU0, DMAC0A_DEI3, DMAC0A_DEI2, DMAC0A_DEI1, DMAC0A_DEI0 } }, @@ -915,7 +912,7 @@ static struct intc_mask_reg mask_registers[] __initdata = { { I2C0_DTEI, I2C0_WAITI, I2C0_TACKI, I2C0_ALI, I2C1_DTEI, I2C1_WAITI, I2C1_TACKI, I2C1_ALI } }, { 0xa40800a0, 0xa40800e0, 8, /* IMR8 / IMCR8 */ - { SDHI0_SDHII3, SDHI0_SDHII2, SDHI0_SDHII1, SDHI0_SDHII0, + { 0, 0, ENABLED, ENABLED, 0, 0, SCIFA5, FSI } }, { 0xa40800a4, 0xa40800e4, 8, /* IMR9 / IMCR9 */ { 0, 0, 0, CMT, 0, USB1, USB0, 0 } }, @@ -962,9 +959,12 @@ static struct intc_mask_reg ack_registers[] __initdata = { { IRQ0, IRQ1, IRQ2, IRQ3, IRQ4, IRQ5, IRQ6, IRQ7 } }, }; -static DECLARE_INTC_DESC_ACK(intc_desc, "sh7724", vectors, groups, - mask_registers, prio_registers, sense_registers, - ack_registers); +static struct intc_desc intc_desc __initdata = { + .name = "sh7724", + .force_enable = ENABLED, + .hw = INTC_HW_DESC(vectors, groups, mask_registers, + prio_registers, sense_registers, ack_registers), +}; void __init plat_irq_setup(void) { -- cgit v0.10.2 From e6f077592d1de2f6a4fc760e7b5d6f20b37d3a27 Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 9 Feb 2010 07:17:20 +0000 Subject: sh: fix INTC to use set_irq_chained_handler() for redirects This patch updates the shared INTC code to use set_irq_chained_handler() for intc_redirect_irq(). With this in place request_irq() on a merged irq which has been redirected will now return -EINVAL instead of 0 together with a crash. This thanks to the protection of the IRQ_NOREQUEST flag set for chained interrupt handlers. Signed-off-by: Magnus Damm Signed-off-by: Paul Mundt diff --git a/drivers/sh/intc.c b/drivers/sh/intc.c index 3a68739..66594eb 100644 --- a/drivers/sh/intc.c +++ b/drivers/sh/intc.c @@ -896,8 +896,8 @@ void __init register_intc_controller(struct intc_desc *desc) vect2->enum_id = 0; /* redirect this interrupts to the first one */ - set_irq_chip_and_handler_name(irq2, &d->chip, - intc_redirect_irq, "redirect"); + set_irq_chained_handler(irq2, intc_redirect_irq); + set_irq_chip(irq2, &d->chip); set_irq_data(irq2, (void *)irq); } } -- cgit v0.10.2 From 801cd56e3e2c2b727399d2c50c50139b2d7c98e8 Mon Sep 17 00:00:00 2001 From: Magnus Damm Date: Tue, 9 Feb 2010 08:22:30 +0000 Subject: sh: break out enable/reparent div4 clocks on sh7723 Break out sh7723 div4 clocks for SIU and IRDA as reparent / enable clocks. Similar to the SIU clock patch for sh7722 by Guennadi. Signed-off-by: Magnus Damm Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/cpu/sh4a/clock-sh7723.c b/arch/sh/kernel/cpu/sh4a/clock-sh7723.c index 20a31c2..76731b6 100644 --- a/arch/sh/kernel/cpu/sh4a/clock-sh7723.c +++ b/arch/sh/kernel/cpu/sh4a/clock-sh7723.c @@ -117,8 +117,7 @@ static struct clk_div_mult_table div4_table = { .nr_multipliers = ARRAY_SIZE(multipliers), }; -enum { DIV4_I, DIV4_U, DIV4_SH, DIV4_B, DIV4_B3, DIV4_P, - DIV4_SIUA, DIV4_SIUB, DIV4_IRDA, DIV4_NR }; +enum { DIV4_I, DIV4_U, DIV4_SH, DIV4_B, DIV4_B3, DIV4_P, DIV4_NR }; #define DIV4(_str, _reg, _bit, _mask, _flags) \ SH_CLK_DIV4(_str, &pll_clk, _reg, _bit, _mask, _flags) @@ -130,11 +129,20 @@ struct clk div4_clks[DIV4_NR] = { [DIV4_B] = DIV4("bus_clk", FRQCR, 8, 0x0dbf, CLK_ENABLE_ON_INIT), [DIV4_B3] = DIV4("b3_clk", FRQCR, 4, 0x0db4, CLK_ENABLE_ON_INIT), [DIV4_P] = DIV4("peripheral_clk", FRQCR, 0, 0x0dbf, 0), - [DIV4_SIUA] = DIV4("siua_clk", SCLKACR, 0, 0x0dbf, 0), - [DIV4_SIUB] = DIV4("siub_clk", SCLKBCR, 0, 0x0dbf, 0), +}; + +enum { DIV4_IRDA, DIV4_ENABLE_NR }; + +struct clk div4_enable_clks[DIV4_ENABLE_NR] = { [DIV4_IRDA] = DIV4("irda_clk", IRDACLKCR, 0, 0x0dbf, 0), }; +enum { DIV4_SIUA, DIV4_SIUB, DIV4_REPARENT_NR }; + +struct clk div4_reparent_clks[DIV4_REPARENT_NR] = { + [DIV4_SIUA] = DIV4("siua_clk", SCLKACR, 0, 0x0dbf, 0), + [DIV4_SIUB] = DIV4("siub_clk", SCLKBCR, 0, 0x0dbf, 0), +}; struct clk div6_clks[] = { SH_CLK_DIV6("video_clk", &pll_clk, VCLKCR, 0), }; @@ -216,6 +224,14 @@ int __init arch_clk_init(void) ret = sh_clk_div4_register(div4_clks, DIV4_NR, &div4_table); if (!ret) + ret = sh_clk_div4_enable_register(div4_enable_clks, + DIV4_ENABLE_NR, &div4_table); + + if (!ret) + ret = sh_clk_div4_reparent_register(div4_reparent_clks, + DIV4_REPARENT_NR, &div4_table); + + if (!ret) ret = sh_clk_div6_register(div6_clks, ARRAY_SIZE(div6_clks)); if (!ret) -- cgit v0.10.2 From 0690535d6bcec5eb28573824df5c8a49ec85b696 Mon Sep 17 00:00:00 2001 From: Leo P White Date: Mon, 8 Feb 2010 13:02:05 +0000 Subject: HID: add mapping for "AL Network Chat" usage Adding a mapping for the 'AL Network Chat' usage from the 'Consumer' usage page (USB HID Usage Tables v1.11). This usage is used by some keyboards for a multimedia key. Signed-off-by: Leo P White Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-input.c b/drivers/hid/hid-input.c index 8430d62..a01a1b0 100644 --- a/drivers/hid/hid-input.c +++ b/drivers/hid/hid-input.c @@ -405,6 +405,7 @@ static void hidinput_configure_usage(struct hid_input *hidinput, struct hid_fiel case 0x192: map_key_clear(KEY_CALC); break; case 0x194: map_key_clear(KEY_FILE); break; case 0x196: map_key_clear(KEY_WWW); break; + case 0x199: map_key_clear(KEY_CHAT); break; case 0x19c: map_key_clear(KEY_LOGOFF); break; case 0x19e: map_key_clear(KEY_COFFEE); break; case 0x1a6: map_key_clear(KEY_HELP); break; -- cgit v0.10.2 From eabe5c90580a065aed8ce6a5ba53eb443d317fae Mon Sep 17 00:00:00 2001 From: Bastien Nocera Date: Tue, 9 Feb 2010 11:43:19 +0100 Subject: HID: fix typo in error message Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-sony.c b/drivers/hid/hid-sony.c index 4e84502..96d723e 100644 --- a/drivers/hid/hid-sony.c +++ b/drivers/hid/hid-sony.c @@ -81,7 +81,7 @@ static int sony_probe(struct hid_device *hdev, const struct hid_device_id *id) sc = kzalloc(sizeof(*sc), GFP_KERNEL); if (sc == NULL) { - dev_err(&hdev->dev, "can't alloc apple descriptor\n"); + dev_err(&hdev->dev, "can't alloc sony descriptor\n"); return -ENOMEM; } -- cgit v0.10.2 From 301ba0457f1ed853fc08e57785f8c87fe7e49c68 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Tue, 9 Feb 2010 15:07:40 +1100 Subject: kthread, sched: Remove reference to kthread_create_on_cpu kthread_create_on_cpu doesn't exist so update a comment in kthread.c to reflect this. Signed-off-by: Anton Blanchard Acked-by: Rusty Russell Cc: Peter Zijlstra LKML-Reference: <20100209040740.GB3702@kryten> Signed-off-by: Ingo Molnar diff --git a/kernel/kthread.c b/kernel/kthread.c index fbb6222..82ed0ea 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -101,7 +101,7 @@ static void create_kthread(struct kthread_create_info *create) * * Description: This helper function creates and names a kernel * thread. The thread will be stopped: use wake_up_process() to start - * it. See also kthread_run(), kthread_create_on_cpu(). + * it. See also kthread_run(). * * When woken, the thread will run @threadfn() with @data as its * argument. @threadfn() can either call do_exit() directly if it is a -- cgit v0.10.2 From 580537140568caddbc8a727d4c2f238d38707f68 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Thu, 28 Jan 2010 22:18:56 -0700 Subject: microblaze: remove early_init_dt_scan_cpus() and phyp_dump_*() Microblaze only has one CPU, it isn't SMP at all. early_init_dt_scan_cpus() is effectively just a no-op, so remove it. Microblaze doesn't support hypervisor assisted dump either, so the phyp stuff can also go. Signed-off-by: Grant Likely Acked-by: Michal Simek diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 6eff83a..cd158ef 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -47,67 +47,6 @@ struct device_node *of_chosen; #define early_init_dt_scan_drconf_memory(node) 0 -static int __init early_init_dt_scan_cpus(unsigned long node, - const char *uname, int depth, - void *data) -{ - static int logical_cpuid; - char *type = of_get_flat_dt_prop(node, "device_type", NULL); - const u32 *intserv; - int i, nthreads; - int found = 0; - - /* We are scanning "cpu" nodes only */ - if (type == NULL || strcmp(type, "cpu") != 0) - return 0; - - /* Get physical cpuid */ - intserv = of_get_flat_dt_prop(node, "reg", NULL); - nthreads = 1; - - /* - * Now see if any of these threads match our boot cpu. - * NOTE: This must match the parsing done in smp_setup_cpu_maps. - */ - for (i = 0; i < nthreads; i++) { - /* - * version 2 of the kexec param format adds the phys cpuid of - * booted proc. - */ - if (initial_boot_params && initial_boot_params->version >= 2) { - if (intserv[i] == - initial_boot_params->boot_cpuid_phys) { - found = 1; - break; - } - } else { - /* - * Check if it's the boot-cpu, set it's hw index now, - * unfortunately this format did not support booting - * off secondary threads. - */ - if (of_get_flat_dt_prop(node, - "linux,boot-cpu", NULL) != NULL) { - found = 1; - break; - } - } - -#ifdef CONFIG_SMP - /* logical cpu id is always 0 on UP kernels */ - logical_cpuid++; -#endif - } - - if (found) { - pr_debug("boot cpu: logical %d physical %d\n", logical_cpuid, - intserv[i]); - boot_cpuid = logical_cpuid; - } - - return 0; -} - void __init early_init_dt_scan_chosen_arch(unsigned long node) { /* No Microblaze specific code here */ @@ -163,86 +102,6 @@ static int __init early_init_dt_scan_memory(unsigned long node, return 0; } -#ifdef CONFIG_PHYP_DUMP -/** - * phyp_dump_calculate_reserve_size() - reserve variable boot area 5% or arg - * - * Function to find the largest size we need to reserve - * during early boot process. - * - * It either looks for boot param and returns that OR - * returns larger of 256 or 5% rounded down to multiples of 256MB. - * - */ -static inline unsigned long phyp_dump_calculate_reserve_size(void) -{ - unsigned long tmp; - - if (phyp_dump_info->reserve_bootvar) - return phyp_dump_info->reserve_bootvar; - - /* divide by 20 to get 5% of value */ - tmp = lmb_end_of_DRAM(); - do_div(tmp, 20); - - /* round it down in multiples of 256 */ - tmp = tmp & ~0x0FFFFFFFUL; - - return (tmp > PHYP_DUMP_RMR_END ? tmp : PHYP_DUMP_RMR_END); -} - -/** - * phyp_dump_reserve_mem() - reserve all not-yet-dumped mmemory - * - * This routine may reserve memory regions in the kernel only - * if the system is supported and a dump was taken in last - * boot instance or if the hardware is supported and the - * scratch area needs to be setup. In other instances it returns - * without reserving anything. The memory in case of dump being - * active is freed when the dump is collected (by userland tools). - */ -static void __init phyp_dump_reserve_mem(void) -{ - unsigned long base, size; - unsigned long variable_reserve_size; - - if (!phyp_dump_info->phyp_dump_configured) { - printk(KERN_ERR "Phyp-dump not supported on this hardware\n"); - return; - } - - if (!phyp_dump_info->phyp_dump_at_boot) { - printk(KERN_INFO "Phyp-dump disabled at boot time\n"); - return; - } - - variable_reserve_size = phyp_dump_calculate_reserve_size(); - - if (phyp_dump_info->phyp_dump_is_active) { - /* Reserve *everything* above RMR.Area freed by userland tools*/ - base = variable_reserve_size; - size = lmb_end_of_DRAM() - base; - - /* XXX crashed_ram_end is wrong, since it may be beyond - * the memory_limit, it will need to be adjusted. */ - lmb_reserve(base, size); - - phyp_dump_info->init_reserve_start = base; - phyp_dump_info->init_reserve_size = size; - } else { - size = phyp_dump_info->cpu_state_size + - phyp_dump_info->hpte_region_size + - variable_reserve_size; - base = lmb_end_of_DRAM() - size; - lmb_reserve(base, size); - phyp_dump_info->init_reserve_start = base; - phyp_dump_info->init_reserve_size = size; - } -} -#else -static inline void __init phyp_dump_reserve_mem(void) {} -#endif /* CONFIG_PHYP_DUMP && CONFIG_PPC_RTAS */ - #ifdef CONFIG_EARLY_PRINTK /* MS this is Microblaze specifig function */ static int __init early_init_dt_scan_serial(unsigned long node, @@ -284,11 +143,6 @@ void __init early_init_devtree(void *params) /* Setup flat device-tree pointer */ initial_boot_params = params; -#ifdef CONFIG_PHYP_DUMP - /* scan tree to see if dump occured during last boot */ - of_scan_flat_dt(early_init_dt_scan_phyp_dump, NULL); -#endif - /* Retrieve various informations from the /chosen node of the * device-tree, including the platform type, initrd location and * size, TCE reserve, and more ... @@ -308,13 +162,6 @@ void __init early_init_devtree(void *params) pr_debug("Phys. mem: %lx\n", (unsigned long) lmb_phys_mem_size()); - pr_debug("Scanning CPUs ...\n"); - - /* Retreive CPU related informations from the flat tree - * (altivec support, boot CPU ID, ...) - */ - of_scan_flat_dt(early_init_dt_scan_cpus, NULL); - pr_debug(" <- early_init_devtree()\n"); } -- cgit v0.10.2 From fcdeb7fedf89f4bbc2e11959794968080cd8426e Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Fri, 29 Jan 2010 05:04:33 -0700 Subject: of: merge of_attach_node() & of_detach_node() Merge common code between PowerPC and Microblaze Signed-off-by: Grant Likely Tested-by: Wolfram Sang Acked-by: Benjamin Herrenschmidt diff --git a/arch/microblaze/include/asm/prom.h b/arch/microblaze/include/asm/prom.h index 07d1063..6c6b386 100644 --- a/arch/microblaze/include/asm/prom.h +++ b/arch/microblaze/include/asm/prom.h @@ -39,10 +39,6 @@ extern struct device_node *of_chosen; extern rwlock_t devtree_lock; /* temporary while merging */ -/* For updating the device tree at runtime */ -extern void of_attach_node(struct device_node *); -extern void of_detach_node(struct device_node *); - /* Other Prototypes */ extern int early_uartlite_console(void); diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index cd158ef..8171282 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -197,65 +197,6 @@ struct device_node *of_find_node_by_phandle(phandle handle) } EXPORT_SYMBOL(of_find_node_by_phandle); -/* - * Plug a device node into the tree and global list. - */ -void of_attach_node(struct device_node *np) -{ - unsigned long flags; - - write_lock_irqsave(&devtree_lock, flags); - np->sibling = np->parent->child; - np->allnext = allnodes; - np->parent->child = np; - allnodes = np; - write_unlock_irqrestore(&devtree_lock, flags); -} - -/* - * "Unplug" a node from the device tree. The caller must hold - * a reference to the node. The memory associated with the node - * is not freed until its refcount goes to zero. - */ -void of_detach_node(struct device_node *np) -{ - struct device_node *parent; - unsigned long flags; - - write_lock_irqsave(&devtree_lock, flags); - - parent = np->parent; - if (!parent) - goto out_unlock; - - if (allnodes == np) - allnodes = np->allnext; - else { - struct device_node *prev; - for (prev = allnodes; - prev->allnext != np; - prev = prev->allnext) - ; - prev->allnext = np->allnext; - } - - if (parent->child == np) - parent->child = np->sibling; - else { - struct device_node *prevsib; - for (prevsib = np->parent->child; - prevsib->sibling != np; - prevsib = prevsib->sibling) - ; - prevsib->sibling = np->sibling; - } - - of_node_set_flag(np, OF_DETACHED); - -out_unlock: - write_unlock_irqrestore(&devtree_lock, flags); -} - #if defined(CONFIG_DEBUG_FS) && defined(DEBUG) static struct debugfs_blob_wrapper flat_dt_blob; diff --git a/arch/powerpc/include/asm/prom.h b/arch/powerpc/include/asm/prom.h index 2ab9cbd..f384db8 100644 --- a/arch/powerpc/include/asm/prom.h +++ b/arch/powerpc/include/asm/prom.h @@ -34,10 +34,6 @@ extern struct device_node *of_chosen; #define HAVE_ARCH_DEVTREE_FIXUPS -/* For updating the device tree at runtime */ -extern void of_attach_node(struct device_node *); -extern void of_detach_node(struct device_node *); - #ifdef CONFIG_PPC32 /* * PCI <-> OF matching functions diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 1ed2ec2..f954c71 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -817,65 +817,6 @@ struct device_node *of_find_next_cache_node(struct device_node *np) return NULL; } -/* - * Plug a device node into the tree and global list. - */ -void of_attach_node(struct device_node *np) -{ - unsigned long flags; - - write_lock_irqsave(&devtree_lock, flags); - np->sibling = np->parent->child; - np->allnext = allnodes; - np->parent->child = np; - allnodes = np; - write_unlock_irqrestore(&devtree_lock, flags); -} - -/* - * "Unplug" a node from the device tree. The caller must hold - * a reference to the node. The memory associated with the node - * is not freed until its refcount goes to zero. - */ -void of_detach_node(struct device_node *np) -{ - struct device_node *parent; - unsigned long flags; - - write_lock_irqsave(&devtree_lock, flags); - - parent = np->parent; - if (!parent) - goto out_unlock; - - if (allnodes == np) - allnodes = np->allnext; - else { - struct device_node *prev; - for (prev = allnodes; - prev->allnext != np; - prev = prev->allnext) - ; - prev->allnext = np->allnext; - } - - if (parent->child == np) - parent->child = np->sibling; - else { - struct device_node *prevsib; - for (prevsib = np->parent->child; - prevsib->sibling != np; - prevsib = prevsib->sibling) - ; - prevsib->sibling = np->sibling; - } - - of_node_set_flag(np, OF_DETACHED); - -out_unlock: - write_unlock_irqrestore(&devtree_lock, flags); -} - #ifdef CONFIG_PPC_PSERIES /* * Fix up the uninitialized fields in a new device node: diff --git a/drivers/of/Kconfig b/drivers/of/Kconfig index 462825e..7cecc8f 100644 --- a/drivers/of/Kconfig +++ b/drivers/of/Kconfig @@ -2,6 +2,10 @@ config OF_FLATTREE bool depends on OF +config OF_DYNAMIC + def_bool y + depends on OF && PPC_OF + config OF_DEVICE def_bool y depends on OF && (SPARC || PPC_OF || MICROBLAZE) diff --git a/drivers/of/base.c b/drivers/of/base.c index cf89ee6..2ce58be 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -870,3 +870,74 @@ int prom_update_property(struct device_node *np, return 0; } + +#if defined(CONFIG_OF_DYNAMIC) +/* + * Support for dynamic device trees. + * + * On some platforms, the device tree can be manipulated at runtime. + * The routines in this section support adding, removing and changing + * device tree nodes. + */ + +/** + * of_attach_node - Plug a device node into the tree and global list. + */ +void of_attach_node(struct device_node *np) +{ + unsigned long flags; + + write_lock_irqsave(&devtree_lock, flags); + np->sibling = np->parent->child; + np->allnext = allnodes; + np->parent->child = np; + allnodes = np; + write_unlock_irqrestore(&devtree_lock, flags); +} + +/** + * of_detach_node - "Unplug" a node from the device tree. + * + * The caller must hold a reference to the node. The memory associated with + * the node is not freed until its refcount goes to zero. + */ +void of_detach_node(struct device_node *np) +{ + struct device_node *parent; + unsigned long flags; + + write_lock_irqsave(&devtree_lock, flags); + + parent = np->parent; + if (!parent) + goto out_unlock; + + if (allnodes == np) + allnodes = np->allnext; + else { + struct device_node *prev; + for (prev = allnodes; + prev->allnext != np; + prev = prev->allnext) + ; + prev->allnext = np->allnext; + } + + if (parent->child == np) + parent->child = np->sibling; + else { + struct device_node *prevsib; + for (prevsib = np->parent->child; + prevsib->sibling != np; + prevsib = prevsib->sibling) + ; + prevsib->sibling = np->sibling; + } + + of_node_set_flag(np, OF_DETACHED); + +out_unlock: + write_unlock_irqrestore(&devtree_lock, flags); +} +#endif /* defined(CONFIG_OF_DYNAMIC) */ + diff --git a/include/linux/of.h b/include/linux/of.h index dbabf86..3cc0d7a 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -184,4 +184,10 @@ extern int of_parse_phandles_with_args(struct device_node *np, const char *list_name, const char *cells_name, int index, struct device_node **out_node, const void **out_args); +#if defined(CONFIG_OF_DYNAMIC) +/* For updating the device tree at runtime */ +extern void of_attach_node(struct device_node *); +extern void of_detach_node(struct device_node *); +#endif + #endif /* _LINUX_OF_H */ -- cgit v0.10.2 From 89751a7cb70a20f0d604dd7c4be29dd7b0011718 Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Mon, 1 Feb 2010 21:34:11 -0700 Subject: of: merge of_find_node_by_phandle Merge common function between powerpc, sparc and microblaze. Code is identical for powerpc and microblaze, but adds a lock (and release) of the devtree_lock on sparc. Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 8171282..f7bd0ee 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -176,27 +176,6 @@ void __init early_init_devtree(void *params) * *******/ -/** - * of_find_node_by_phandle - Find a node given a phandle - * @handle: phandle of the node to find - * - * Returns a node pointer with refcount incremented, use - * of_node_put() on it when done. - */ -struct device_node *of_find_node_by_phandle(phandle handle) -{ - struct device_node *np; - - read_lock(&devtree_lock); - for (np = allnodes; np != NULL; np = np->allnext) - if (np->phandle == handle) - break; - of_node_get(np); - read_unlock(&devtree_lock); - return np; -} -EXPORT_SYMBOL(of_find_node_by_phandle); - #if defined(CONFIG_DEBUG_FS) && defined(DEBUG) static struct debugfs_blob_wrapper flat_dt_blob; diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index f954c71..28be19a 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -766,27 +766,6 @@ void __init early_init_devtree(void *params) *******/ /** - * of_find_node_by_phandle - Find a node given a phandle - * @handle: phandle of the node to find - * - * Returns a node pointer with refcount incremented, use - * of_node_put() on it when done. - */ -struct device_node *of_find_node_by_phandle(phandle handle) -{ - struct device_node *np; - - read_lock(&devtree_lock); - for (np = allnodes; np != 0; np = np->allnext) - if (np->phandle == handle) - break; - of_node_get(np); - read_unlock(&devtree_lock); - return np; -} -EXPORT_SYMBOL(of_find_node_by_phandle); - -/** * of_find_next_cache_node - Find a node's subsidiary cache * @np: node of type "cpu" or "cache" * diff --git a/arch/sparc/kernel/prom_common.c b/arch/sparc/kernel/prom_common.c index 5832e13..57ac9e2 100644 --- a/arch/sparc/kernel/prom_common.c +++ b/arch/sparc/kernel/prom_common.c @@ -37,18 +37,6 @@ EXPORT_SYMBOL(of_console_path); char *of_console_options; EXPORT_SYMBOL(of_console_options); -struct device_node *of_find_node_by_phandle(phandle handle) -{ - struct device_node *np; - - for (np = allnodes; np; np = np->allnext) - if (np->phandle == handle) - break; - - return np; -} -EXPORT_SYMBOL(of_find_node_by_phandle); - int of_getintprop_default(struct device_node *np, const char *name, int def) { struct property *prop; diff --git a/drivers/of/base.c b/drivers/of/base.c index 2ce58be..785e9cc 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -615,6 +615,27 @@ int of_modalias_node(struct device_node *node, char *modalias, int len) EXPORT_SYMBOL_GPL(of_modalias_node); /** + * of_find_node_by_phandle - Find a node given a phandle + * @handle: phandle of the node to find + * + * Returns a node pointer with refcount incremented, use + * of_node_put() on it when done. + */ +struct device_node *of_find_node_by_phandle(phandle handle) +{ + struct device_node *np; + + read_lock(&devtree_lock); + for (np = allnodes; np; np = np->allnext) + if (np->phandle == handle) + break; + of_node_get(np); + read_unlock(&devtree_lock); + return np; +} +EXPORT_SYMBOL(of_find_node_by_phandle); + +/** * of_parse_phandle - Resolve a phandle property to a device_node pointer * @np: Pointer to device node holding phandle property * @phandle_name: Name of property holding a phandle value -- cgit v0.10.2 From 71a157e8edca55198e808f8561dd49017a54ee34 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 1 Feb 2010 21:34:14 -0700 Subject: of: add 'of_' prefix to machine_is_compatible() machine is compatible is an OF-specific call. It should have the of_ prefix to protect the global namespace. Signed-off-by: Grant Likely Acked-by: Michal Simek diff --git a/arch/powerpc/kernel/pci_64.c b/arch/powerpc/kernel/pci_64.c index ccf56ac..d43fc65 100644 --- a/arch/powerpc/kernel/pci_64.c +++ b/arch/powerpc/kernel/pci_64.c @@ -224,7 +224,7 @@ long sys_pciconfig_iobase(long which, unsigned long in_bus, * G5 machines... So when something asks for bus 0 io base * (bus 0 is HT root), we return the AGP one instead. */ - if (in_bus == 0 && machine_is_compatible("MacRISC4")) { + if (in_bus == 0 && of_machine_is_compatible("MacRISC4")) { struct device_node *agp; agp = of_find_compatible_node(NULL, NULL, "u3-agp"); diff --git a/arch/powerpc/platforms/85xx/xes_mpc85xx.c b/arch/powerpc/platforms/85xx/xes_mpc85xx.c index 1b42605..0125604 100644 --- a/arch/powerpc/platforms/85xx/xes_mpc85xx.c +++ b/arch/powerpc/platforms/85xx/xes_mpc85xx.c @@ -80,8 +80,8 @@ static void xes_mpc85xx_configure_l2(void __iomem *l2_base) printk(KERN_INFO "xes_mpc85xx: Enabling L2 as cache\n"); ctl = MPC85xx_L2CTL_L2E | MPC85xx_L2CTL_L2I; - if (machine_is_compatible("MPC8540") || - machine_is_compatible("MPC8560")) + if (of_machine_is_compatible("MPC8540") || + of_machine_is_compatible("MPC8560")) /* * Assume L2 SRAM is used fully for cache, so set * L2BLKSZ (bits 4:5) to match L2SIZ (bits 2:3). diff --git a/arch/powerpc/platforms/cell/cbe_powerbutton.c b/arch/powerpc/platforms/cell/cbe_powerbutton.c index dcddaa5..f75a4da 100644 --- a/arch/powerpc/platforms/cell/cbe_powerbutton.c +++ b/arch/powerpc/platforms/cell/cbe_powerbutton.c @@ -48,7 +48,7 @@ static int __init cbe_powerbutton_init(void) int ret = 0; struct input_dev *dev; - if (!machine_is_compatible("IBM,CBPLUS-1.0")) { + if (!of_machine_is_compatible("IBM,CBPLUS-1.0")) { printk(KERN_ERR "%s: Not a cell blade.\n", __func__); ret = -ENODEV; goto out; diff --git a/arch/powerpc/platforms/cell/ras.c b/arch/powerpc/platforms/cell/ras.c index 5e0a191..608fd2b 100644 --- a/arch/powerpc/platforms/cell/ras.c +++ b/arch/powerpc/platforms/cell/ras.c @@ -255,7 +255,7 @@ static int __init cbe_sysreset_init(void) { struct cbe_pmd_regs __iomem *regs; - sysreset_hack = machine_is_compatible("IBM,CBPLUS-1.0"); + sysreset_hack = of_machine_is_compatible("IBM,CBPLUS-1.0"); if (!sysreset_hack) return 0; diff --git a/arch/powerpc/platforms/pasemi/cpufreq.c b/arch/powerpc/platforms/pasemi/cpufreq.c index be2527a..d35e052 100644 --- a/arch/powerpc/platforms/pasemi/cpufreq.c +++ b/arch/powerpc/platforms/pasemi/cpufreq.c @@ -304,8 +304,8 @@ static struct cpufreq_driver pas_cpufreq_driver = { static int __init pas_cpufreq_init(void) { - if (!machine_is_compatible("PA6T-1682M") && - !machine_is_compatible("pasemi,pwrficient")) + if (!of_machine_is_compatible("PA6T-1682M") && + !of_machine_is_compatible("pasemi,pwrficient")) return -ENODEV; return cpufreq_register_driver(&pas_cpufreq_driver); diff --git a/arch/powerpc/platforms/powermac/cpufreq_32.c b/arch/powerpc/platforms/powermac/cpufreq_32.c index 08d94e4..d4f127d 100644 --- a/arch/powerpc/platforms/powermac/cpufreq_32.c +++ b/arch/powerpc/platforms/powermac/cpufreq_32.c @@ -657,31 +657,31 @@ static int __init pmac_cpufreq_setup(void) cur_freq = (*value) / 1000; /* Check for 7447A based MacRISC3 */ - if (machine_is_compatible("MacRISC3") && + if (of_machine_is_compatible("MacRISC3") && of_get_property(cpunode, "dynamic-power-step", NULL) && PVR_VER(mfspr(SPRN_PVR)) == 0x8003) { pmac_cpufreq_init_7447A(cpunode); /* Check for other MacRISC3 machines */ - } else if (machine_is_compatible("PowerBook3,4") || - machine_is_compatible("PowerBook3,5") || - machine_is_compatible("MacRISC3")) { + } else if (of_machine_is_compatible("PowerBook3,4") || + of_machine_is_compatible("PowerBook3,5") || + of_machine_is_compatible("MacRISC3")) { pmac_cpufreq_init_MacRISC3(cpunode); /* Else check for iBook2 500/600 */ - } else if (machine_is_compatible("PowerBook4,1")) { + } else if (of_machine_is_compatible("PowerBook4,1")) { hi_freq = cur_freq; low_freq = 400000; set_speed_proc = pmu_set_cpu_speed; is_pmu_based = 1; } /* Else check for TiPb 550 */ - else if (machine_is_compatible("PowerBook3,3") && cur_freq == 550000) { + else if (of_machine_is_compatible("PowerBook3,3") && cur_freq == 550000) { hi_freq = cur_freq; low_freq = 500000; set_speed_proc = pmu_set_cpu_speed; is_pmu_based = 1; } /* Else check for TiPb 400 & 500 */ - else if (machine_is_compatible("PowerBook3,2")) { + else if (of_machine_is_compatible("PowerBook3,2")) { /* We only know about the 400 MHz and the 500Mhz model * they both have 300 MHz as low frequency */ diff --git a/arch/powerpc/platforms/powermac/cpufreq_64.c b/arch/powerpc/platforms/powermac/cpufreq_64.c index 708c751..3ed288e 100644 --- a/arch/powerpc/platforms/powermac/cpufreq_64.c +++ b/arch/powerpc/platforms/powermac/cpufreq_64.c @@ -398,11 +398,11 @@ static int __init g5_neo2_cpufreq_init(struct device_node *cpus) int rc = -ENODEV; /* Check supported platforms */ - if (machine_is_compatible("PowerMac8,1") || - machine_is_compatible("PowerMac8,2") || - machine_is_compatible("PowerMac9,1")) + if (of_machine_is_compatible("PowerMac8,1") || + of_machine_is_compatible("PowerMac8,2") || + of_machine_is_compatible("PowerMac9,1")) use_volts_smu = 1; - else if (machine_is_compatible("PowerMac11,2")) + else if (of_machine_is_compatible("PowerMac11,2")) use_volts_vdnap = 1; else return -ENODEV; @@ -729,9 +729,9 @@ static int __init g5_cpufreq_init(void) return -ENODEV; } - if (machine_is_compatible("PowerMac7,2") || - machine_is_compatible("PowerMac7,3") || - machine_is_compatible("RackMac3,1")) + if (of_machine_is_compatible("PowerMac7,2") || + of_machine_is_compatible("PowerMac7,3") || + of_machine_is_compatible("RackMac3,1")) rc = g5_pm72_cpufreq_init(cpus); #ifdef CONFIG_PMAC_SMU else diff --git a/arch/powerpc/platforms/powermac/feature.c b/arch/powerpc/platforms/powermac/feature.c index fbc9bbd..33e815f 100644 --- a/arch/powerpc/platforms/powermac/feature.c +++ b/arch/powerpc/platforms/powermac/feature.c @@ -2426,7 +2426,7 @@ static int __init probe_motherboard(void) } } for(i=0; imode_info.connector_table = radeon_connector_table; if (rdev->mode_info.connector_table == CT_NONE) { #ifdef CONFIG_PPC_PMAC - if (machine_is_compatible("PowerBook3,3")) { + if (of_machine_is_compatible("PowerBook3,3")) { /* powerbook with VGA */ rdev->mode_info.connector_table = CT_POWERBOOK_VGA; - } else if (machine_is_compatible("PowerBook3,4") || - machine_is_compatible("PowerBook3,5")) { + } else if (of_machine_is_compatible("PowerBook3,4") || + of_machine_is_compatible("PowerBook3,5")) { /* powerbook with internal tmds */ rdev->mode_info.connector_table = CT_POWERBOOK_INTERNAL; - } else if (machine_is_compatible("PowerBook5,1") || - machine_is_compatible("PowerBook5,2") || - machine_is_compatible("PowerBook5,3") || - machine_is_compatible("PowerBook5,4") || - machine_is_compatible("PowerBook5,5")) { + } else if (of_machine_is_compatible("PowerBook5,1") || + of_machine_is_compatible("PowerBook5,2") || + of_machine_is_compatible("PowerBook5,3") || + of_machine_is_compatible("PowerBook5,4") || + of_machine_is_compatible("PowerBook5,5")) { /* powerbook with external single link tmds (sil164) */ rdev->mode_info.connector_table = CT_POWERBOOK_EXTERNAL; - } else if (machine_is_compatible("PowerBook5,6")) { + } else if (of_machine_is_compatible("PowerBook5,6")) { /* powerbook with external dual or single link tmds */ rdev->mode_info.connector_table = CT_POWERBOOK_EXTERNAL; - } else if (machine_is_compatible("PowerBook5,7") || - machine_is_compatible("PowerBook5,8") || - machine_is_compatible("PowerBook5,9")) { + } else if (of_machine_is_compatible("PowerBook5,7") || + of_machine_is_compatible("PowerBook5,8") || + of_machine_is_compatible("PowerBook5,9")) { /* PowerBook6,2 ? */ /* powerbook with external dual link tmds (sil1178?) */ rdev->mode_info.connector_table = CT_POWERBOOK_EXTERNAL; - } else if (machine_is_compatible("PowerBook4,1") || - machine_is_compatible("PowerBook4,2") || - machine_is_compatible("PowerBook4,3") || - machine_is_compatible("PowerBook6,3") || - machine_is_compatible("PowerBook6,5") || - machine_is_compatible("PowerBook6,7")) { + } else if (of_machine_is_compatible("PowerBook4,1") || + of_machine_is_compatible("PowerBook4,2") || + of_machine_is_compatible("PowerBook4,3") || + of_machine_is_compatible("PowerBook6,3") || + of_machine_is_compatible("PowerBook6,5") || + of_machine_is_compatible("PowerBook6,7")) { /* ibook */ rdev->mode_info.connector_table = CT_IBOOK; - } else if (machine_is_compatible("PowerMac4,4")) { + } else if (of_machine_is_compatible("PowerMac4,4")) { /* emac */ rdev->mode_info.connector_table = CT_EMAC; - } else if (machine_is_compatible("PowerMac10,1")) { + } else if (of_machine_is_compatible("PowerMac10,1")) { /* mini with internal tmds */ rdev->mode_info.connector_table = CT_MINI_INTERNAL; - } else if (machine_is_compatible("PowerMac10,2")) { + } else if (of_machine_is_compatible("PowerMac10,2")) { /* mini with external tmds */ rdev->mode_info.connector_table = CT_MINI_EXTERNAL; - } else if (machine_is_compatible("PowerMac12,1")) { + } else if (of_machine_is_compatible("PowerMac12,1")) { /* PowerMac8,1 ? */ /* imac g5 isight */ rdev->mode_info.connector_table = CT_IMAC_G5_ISIGHT; diff --git a/drivers/macintosh/adb.c b/drivers/macintosh/adb.c index 23741ce..d840a10 100644 --- a/drivers/macintosh/adb.c +++ b/drivers/macintosh/adb.c @@ -322,8 +322,8 @@ static int __init adb_init(void) adb_controller = NULL; } else { #ifdef CONFIG_PPC - if (machine_is_compatible("AAPL,PowerBook1998") || - machine_is_compatible("PowerBook1,1")) + if (of_machine_is_compatible("AAPL,PowerBook1998") || + of_machine_is_compatible("PowerBook1,1")) sleepy_trackpad = 1; #endif /* CONFIG_PPC */ diff --git a/drivers/macintosh/therm_pm72.c b/drivers/macintosh/therm_pm72.c index 454bc50..5738d8b 100644 --- a/drivers/macintosh/therm_pm72.c +++ b/drivers/macintosh/therm_pm72.c @@ -1899,7 +1899,7 @@ static int create_control_loops(void) */ if (rackmac) cpu_pid_type = CPU_PID_TYPE_RACKMAC; - else if (machine_is_compatible("PowerMac7,3") + else if (of_machine_is_compatible("PowerMac7,3") && (cpu_count > 1) && fcu_fans[CPUA_PUMP_RPM_INDEX].id != FCU_FAN_ABSENT_ID && fcu_fans[CPUB_PUMP_RPM_INDEX].id != FCU_FAN_ABSENT_ID) { @@ -2234,10 +2234,10 @@ static int __init therm_pm72_init(void) { struct device_node *np; - rackmac = machine_is_compatible("RackMac3,1"); + rackmac = of_machine_is_compatible("RackMac3,1"); - if (!machine_is_compatible("PowerMac7,2") && - !machine_is_compatible("PowerMac7,3") && + if (!of_machine_is_compatible("PowerMac7,2") && + !of_machine_is_compatible("PowerMac7,3") && !rackmac) return -ENODEV; diff --git a/drivers/macintosh/therm_windtunnel.c b/drivers/macintosh/therm_windtunnel.c index ba48fd7..7fb8b4d 100644 --- a/drivers/macintosh/therm_windtunnel.c +++ b/drivers/macintosh/therm_windtunnel.c @@ -490,7 +490,7 @@ g4fan_init( void ) info = of_get_property(np, "thermal-info", NULL); of_node_put(np); - if( !info || !machine_is_compatible("PowerMac3,6") ) + if( !info || !of_machine_is_compatible("PowerMac3,6") ) return -ENODEV; if( info->id != 3 ) { diff --git a/drivers/macintosh/via-pmu-backlight.c b/drivers/macintosh/via-pmu-backlight.c index a348bb0..4f3c447 100644 --- a/drivers/macintosh/via-pmu-backlight.c +++ b/drivers/macintosh/via-pmu-backlight.c @@ -150,13 +150,13 @@ void __init pmu_backlight_init() /* Special case for the old PowerBook since I can't test on it */ autosave = - machine_is_compatible("AAPL,3400/2400") || - machine_is_compatible("AAPL,3500"); + of_machine_is_compatible("AAPL,3400/2400") || + of_machine_is_compatible("AAPL,3500"); if (!autosave && !pmac_has_backlight_type("pmu") && - !machine_is_compatible("AAPL,PowerBook1998") && - !machine_is_compatible("PowerBook1,1")) + !of_machine_is_compatible("AAPL,PowerBook1998") && + !of_machine_is_compatible("PowerBook1,1")) return; snprintf(name, sizeof(name), "pmubl"); diff --git a/drivers/macintosh/via-pmu.c b/drivers/macintosh/via-pmu.c index db379c3..4276484 100644 --- a/drivers/macintosh/via-pmu.c +++ b/drivers/macintosh/via-pmu.c @@ -463,8 +463,8 @@ static int __init via_pmu_dev_init(void) #endif #ifdef CONFIG_PPC32 - if (machine_is_compatible("AAPL,3400/2400") || - machine_is_compatible("AAPL,3500")) { + if (of_machine_is_compatible("AAPL,3400/2400") || + of_machine_is_compatible("AAPL,3500")) { int mb = pmac_call_feature(PMAC_FTR_GET_MB_INFO, NULL, PMAC_MB_INFO_MODEL, 0); pmu_battery_count = 1; @@ -472,8 +472,8 @@ static int __init via_pmu_dev_init(void) pmu_batteries[0].flags |= PMU_BATT_TYPE_COMET; else pmu_batteries[0].flags |= PMU_BATT_TYPE_HOOPER; - } else if (machine_is_compatible("AAPL,PowerBook1998") || - machine_is_compatible("PowerBook1,1")) { + } else if (of_machine_is_compatible("AAPL,PowerBook1998") || + of_machine_is_compatible("PowerBook1,1")) { pmu_battery_count = 2; pmu_batteries[0].flags |= PMU_BATT_TYPE_SMART; pmu_batteries[1].flags |= PMU_BATT_TYPE_SMART; diff --git a/drivers/macintosh/windfarm_core.c b/drivers/macintosh/windfarm_core.c index 075b4d9..437f55c 100644 --- a/drivers/macintosh/windfarm_core.c +++ b/drivers/macintosh/windfarm_core.c @@ -468,9 +468,9 @@ static int __init windfarm_core_init(void) DBG("wf: core loaded\n"); /* Don't register on old machines that use therm_pm72 for now */ - if (machine_is_compatible("PowerMac7,2") || - machine_is_compatible("PowerMac7,3") || - machine_is_compatible("RackMac3,1")) + if (of_machine_is_compatible("PowerMac7,2") || + of_machine_is_compatible("PowerMac7,3") || + of_machine_is_compatible("RackMac3,1")) return -ENODEV; platform_device_register(&wf_platform_device); return 0; diff --git a/drivers/macintosh/windfarm_cpufreq_clamp.c b/drivers/macintosh/windfarm_cpufreq_clamp.c index 900aade..1a77a7c 100644 --- a/drivers/macintosh/windfarm_cpufreq_clamp.c +++ b/drivers/macintosh/windfarm_cpufreq_clamp.c @@ -76,9 +76,9 @@ static int __init wf_cpufreq_clamp_init(void) struct wf_control *clamp; /* Don't register on old machines that use therm_pm72 for now */ - if (machine_is_compatible("PowerMac7,2") || - machine_is_compatible("PowerMac7,3") || - machine_is_compatible("RackMac3,1")) + if (of_machine_is_compatible("PowerMac7,2") || + of_machine_is_compatible("PowerMac7,3") || + of_machine_is_compatible("RackMac3,1")) return -ENODEV; clamp = kmalloc(sizeof(struct wf_control), GFP_KERNEL); diff --git a/drivers/macintosh/windfarm_lm75_sensor.c b/drivers/macintosh/windfarm_lm75_sensor.c index ed6426a..d8257d3 100644 --- a/drivers/macintosh/windfarm_lm75_sensor.c +++ b/drivers/macintosh/windfarm_lm75_sensor.c @@ -239,9 +239,9 @@ static struct i2c_driver wf_lm75_driver = { static int __init wf_lm75_sensor_init(void) { /* Don't register on old machines that use therm_pm72 for now */ - if (machine_is_compatible("PowerMac7,2") || - machine_is_compatible("PowerMac7,3") || - machine_is_compatible("RackMac3,1")) + if (of_machine_is_compatible("PowerMac7,2") || + of_machine_is_compatible("PowerMac7,3") || + of_machine_is_compatible("RackMac3,1")) return -ENODEV; return i2c_add_driver(&wf_lm75_driver); } diff --git a/drivers/macintosh/windfarm_max6690_sensor.c b/drivers/macintosh/windfarm_max6690_sensor.c index a67b349..b486eb9 100644 --- a/drivers/macintosh/windfarm_max6690_sensor.c +++ b/drivers/macintosh/windfarm_max6690_sensor.c @@ -188,9 +188,9 @@ static struct i2c_driver wf_max6690_driver = { static int __init wf_max6690_sensor_init(void) { /* Don't register on old machines that use therm_pm72 for now */ - if (machine_is_compatible("PowerMac7,2") || - machine_is_compatible("PowerMac7,3") || - machine_is_compatible("RackMac3,1")) + if (of_machine_is_compatible("PowerMac7,2") || + of_machine_is_compatible("PowerMac7,3") || + of_machine_is_compatible("RackMac3,1")) return -ENODEV; return i2c_add_driver(&wf_max6690_driver); } diff --git a/drivers/macintosh/windfarm_pm112.c b/drivers/macintosh/windfarm_pm112.c index 73d695d..e0ee807 100644 --- a/drivers/macintosh/windfarm_pm112.c +++ b/drivers/macintosh/windfarm_pm112.c @@ -676,7 +676,7 @@ static int __init wf_pm112_init(void) { struct device_node *cpu; - if (!machine_is_compatible("PowerMac11,2")) + if (!of_machine_is_compatible("PowerMac11,2")) return -ENODEV; /* Count the number of CPU cores */ diff --git a/drivers/macintosh/windfarm_pm121.c b/drivers/macintosh/windfarm_pm121.c index 66ec4fb1..947d4af 100644 --- a/drivers/macintosh/windfarm_pm121.c +++ b/drivers/macintosh/windfarm_pm121.c @@ -1008,7 +1008,7 @@ static int __init pm121_init(void) { int rc = -ENODEV; - if (machine_is_compatible("PowerMac12,1")) + if (of_machine_is_compatible("PowerMac12,1")) rc = pm121_init_pm(); if (rc == 0) { diff --git a/drivers/macintosh/windfarm_pm81.c b/drivers/macintosh/windfarm_pm81.c index abbe206..565d5b2 100644 --- a/drivers/macintosh/windfarm_pm81.c +++ b/drivers/macintosh/windfarm_pm81.c @@ -779,8 +779,8 @@ static int __init wf_smu_init(void) { int rc = -ENODEV; - if (machine_is_compatible("PowerMac8,1") || - machine_is_compatible("PowerMac8,2")) + if (of_machine_is_compatible("PowerMac8,1") || + of_machine_is_compatible("PowerMac8,2")) rc = wf_init_pm(); if (rc == 0) { diff --git a/drivers/macintosh/windfarm_pm91.c b/drivers/macintosh/windfarm_pm91.c index 764c525..bea9916 100644 --- a/drivers/macintosh/windfarm_pm91.c +++ b/drivers/macintosh/windfarm_pm91.c @@ -711,7 +711,7 @@ static int __init wf_smu_init(void) { int rc = -ENODEV; - if (machine_is_compatible("PowerMac9,1")) + if (of_machine_is_compatible("PowerMac9,1")) rc = wf_init_pm(); if (rc == 0) { diff --git a/drivers/macintosh/windfarm_smu_sensors.c b/drivers/macintosh/windfarm_smu_sensors.c index 9c567b9..3c19350 100644 --- a/drivers/macintosh/windfarm_smu_sensors.c +++ b/drivers/macintosh/windfarm_smu_sensors.c @@ -363,9 +363,9 @@ smu_cpu_power_create(struct wf_sensor *volts, struct wf_sensor *amps) * I yet have to figure out what's up with 8,2 and will have to * adjust for later, unless we can 100% trust the SDB partition... */ - if ((machine_is_compatible("PowerMac8,1") || - machine_is_compatible("PowerMac8,2") || - machine_is_compatible("PowerMac9,1")) && + if ((of_machine_is_compatible("PowerMac8,1") || + of_machine_is_compatible("PowerMac8,2") || + of_machine_is_compatible("PowerMac9,1")) && cpuvcp_version >= 2) { pow->quadratic = 1; DBG("windfarm: CPU Power using quadratic transform\n"); diff --git a/drivers/net/mace.c b/drivers/net/mace.c index d9fbad3..43aea91 100644 --- a/drivers/net/mace.c +++ b/drivers/net/mace.c @@ -206,7 +206,7 @@ static int __devinit mace_probe(struct macio_dev *mdev, const struct of_device_i mp->port_aaui = port_aaui; else { /* Apple Network Server uses the AAUI port */ - if (machine_is_compatible("AAPL,ShinerESB")) + if (of_machine_is_compatible("AAPL,ShinerESB")) mp->port_aaui = 1; else { #ifdef CONFIG_MACE_AAUI_PORT diff --git a/drivers/of/base.c b/drivers/of/base.c index 785e9cc..6bc8740 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -219,13 +219,13 @@ int of_device_is_compatible(const struct device_node *device, EXPORT_SYMBOL(of_device_is_compatible); /** - * machine_is_compatible - Test root of device tree for a given compatible value + * of_machine_is_compatible - Test root of device tree for a given compatible value * @compat: compatible string to look for in root node's compatible property. * * Returns true if the root node has the given value in its * compatible property. */ -int machine_is_compatible(const char *compat) +int of_machine_is_compatible(const char *compat) { struct device_node *root; int rc = 0; @@ -237,7 +237,7 @@ int machine_is_compatible(const char *compat) } return rc; } -EXPORT_SYMBOL(machine_is_compatible); +EXPORT_SYMBOL(of_machine_is_compatible); /** * of_device_is_available - check if a device is available for use diff --git a/drivers/serial/pmac_zilog.c b/drivers/serial/pmac_zilog.c index 683e66f..3e2ae48 100644 --- a/drivers/serial/pmac_zilog.c +++ b/drivers/serial/pmac_zilog.c @@ -2031,9 +2031,9 @@ static int __init pmz_console_setup(struct console *co, char *options) /* * XServe's default to 57600 bps */ - if (machine_is_compatible("RackMac1,1") - || machine_is_compatible("RackMac1,2") - || machine_is_compatible("MacRISC4")) + if (of_machine_is_compatible("RackMac1,1") + || of_machine_is_compatible("RackMac1,2") + || of_machine_is_compatible("MacRISC4")) baud = 57600; /* diff --git a/drivers/video/aty/aty128fb.c b/drivers/video/aty/aty128fb.c index e4e4d43..9ee67d6 100644 --- a/drivers/video/aty/aty128fb.c +++ b/drivers/video/aty/aty128fb.c @@ -1931,22 +1931,22 @@ static int __devinit aty128_init(struct pci_dev *pdev, const struct pci_device_i * PowerMac2,2 summer 2000 iMacs * PowerMac4,1 january 2001 iMacs "flower power" */ - if (machine_is_compatible("PowerMac2,1") || - machine_is_compatible("PowerMac2,2") || - machine_is_compatible("PowerMac4,1")) + if (of_machine_is_compatible("PowerMac2,1") || + of_machine_is_compatible("PowerMac2,2") || + of_machine_is_compatible("PowerMac4,1")) default_vmode = VMODE_1024_768_75; /* iBook SE */ - if (machine_is_compatible("PowerBook2,2")) + if (of_machine_is_compatible("PowerBook2,2")) default_vmode = VMODE_800_600_60; /* PowerBook Firewire (Pismo), iBook Dual USB */ - if (machine_is_compatible("PowerBook3,1") || - machine_is_compatible("PowerBook4,1")) + if (of_machine_is_compatible("PowerBook3,1") || + of_machine_is_compatible("PowerBook4,1")) default_vmode = VMODE_1024_768_60; /* PowerBook Titanium */ - if (machine_is_compatible("PowerBook3,2")) + if (of_machine_is_compatible("PowerBook3,2")) default_vmode = VMODE_1152_768_60; if (default_cmode > 16) diff --git a/drivers/video/aty/atyfb_base.c b/drivers/video/aty/atyfb_base.c index 5f1b580..e45ab8d 100644 --- a/drivers/video/aty/atyfb_base.c +++ b/drivers/video/aty/atyfb_base.c @@ -2439,7 +2439,7 @@ static int __devinit aty_init(struct fb_info *info) * The Apple iBook1 uses non-standard memory frequencies. * We detect it and set the frequency manually. */ - if (machine_is_compatible("PowerBook2,1")) { + if (of_machine_is_compatible("PowerBook2,1")) { par->pll_limits.mclk = 70; par->pll_limits.xclk = 53; } @@ -2659,7 +2659,7 @@ static int __devinit aty_init(struct fb_info *info) FBINFO_HWACCEL_YPAN; #ifdef CONFIG_PMAC_BACKLIGHT - if (M64_HAS(G3_PB_1_1) && machine_is_compatible("PowerBook1,1")) { + if (M64_HAS(G3_PB_1_1) && of_machine_is_compatible("PowerBook1,1")) { /* * these bits let the 101 powerbook * wake up from sleep -- paulus @@ -2690,9 +2690,9 @@ static int __devinit aty_init(struct fb_info *info) if (M64_HAS(G3_PB_1024x768)) /* G3 PowerBook with 1024x768 LCD */ default_vmode = VMODE_1024_768_60; - else if (machine_is_compatible("iMac")) + else if (of_machine_is_compatible("iMac")) default_vmode = VMODE_1024_768_75; - else if (machine_is_compatible("PowerBook2,1")) + else if (of_machine_is_compatible("PowerBook2,1")) /* iBook with 800x600 LCD */ default_vmode = VMODE_800_600_60; else diff --git a/drivers/video/aty/radeon_backlight.c b/drivers/video/aty/radeon_backlight.c index 1a056ad..fa1198c 100644 --- a/drivers/video/aty/radeon_backlight.c +++ b/drivers/video/aty/radeon_backlight.c @@ -175,9 +175,9 @@ void radeonfb_bl_init(struct radeonfb_info *rinfo) #ifdef CONFIG_PMAC_BACKLIGHT pdata->negative = pdata->negative || - machine_is_compatible("PowerBook4,3") || - machine_is_compatible("PowerBook6,3") || - machine_is_compatible("PowerBook6,5"); + of_machine_is_compatible("PowerBook4,3") || + of_machine_is_compatible("PowerBook6,3") || + of_machine_is_compatible("PowerBook6,5"); #endif rinfo->info->bl_dev = bd; diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index 8118d45..fbf2961 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -85,7 +85,7 @@ extern int early_init_dt_scan_root(unsigned long node, const char *uname, extern void finish_device_tree(void); extern void unflatten_device_tree(void); extern void early_init_devtree(void *); -extern int machine_is_compatible(const char *compat); +extern int of_machine_is_compatible(const char *compat); extern void print_properties(struct device_node *node); extern int prom_n_intr_cells(struct device_node* np); extern void prom_get_irq_senses(unsigned char *senses, int off, int max); diff --git a/sound/ppc/awacs.c b/sound/ppc/awacs.c index 2e15646..b366793 100644 --- a/sound/ppc/awacs.c +++ b/sound/ppc/awacs.c @@ -751,8 +751,8 @@ static void snd_pmac_awacs_suspend(struct snd_pmac *chip) static void snd_pmac_awacs_resume(struct snd_pmac *chip) { - if (machine_is_compatible("PowerBook3,1") - || machine_is_compatible("PowerBook3,2")) { + if (of_machine_is_compatible("PowerBook3,1") + || of_machine_is_compatible("PowerBook3,2")) { msleep(100); snd_pmac_awacs_write_reg(chip, 1, chip->awacs_reg[1] & ~MASK_PAROUT); @@ -780,16 +780,16 @@ static void snd_pmac_awacs_resume(struct snd_pmac *chip) } #endif /* CONFIG_PM */ -#define IS_PM7500 (machine_is_compatible("AAPL,7500") \ - || machine_is_compatible("AAPL,8500") \ - || machine_is_compatible("AAPL,9500")) -#define IS_PM5500 (machine_is_compatible("AAPL,e411")) -#define IS_BEIGE (machine_is_compatible("AAPL,Gossamer")) -#define IS_IMAC1 (machine_is_compatible("PowerMac2,1")) -#define IS_IMAC2 (machine_is_compatible("PowerMac2,2") \ - || machine_is_compatible("PowerMac4,1")) -#define IS_G4AGP (machine_is_compatible("PowerMac3,1")) -#define IS_LOMBARD (machine_is_compatible("PowerBook1,1")) +#define IS_PM7500 (of_machine_is_compatible("AAPL,7500") \ + || of_machine_is_compatible("AAPL,8500") \ + || of_machine_is_compatible("AAPL,9500")) +#define IS_PM5500 (of_machine_is_compatible("AAPL,e411")) +#define IS_BEIGE (of_machine_is_compatible("AAPL,Gossamer")) +#define IS_IMAC1 (of_machine_is_compatible("PowerMac2,1")) +#define IS_IMAC2 (of_machine_is_compatible("PowerMac2,2") \ + || of_machine_is_compatible("PowerMac4,1")) +#define IS_G4AGP (of_machine_is_compatible("PowerMac3,1")) +#define IS_LOMBARD (of_machine_is_compatible("PowerBook1,1")) static int imac1, imac2; diff --git a/sound/ppc/burgundy.c b/sound/ppc/burgundy.c index 0accfe4..1f72e1c 100644 --- a/sound/ppc/burgundy.c +++ b/sound/ppc/burgundy.c @@ -582,7 +582,7 @@ static int snd_pmac_burgundy_detect_headphone(struct snd_pmac *chip) static void snd_pmac_burgundy_update_automute(struct snd_pmac *chip, int do_notify) { if (chip->auto_mute) { - int imac = machine_is_compatible("iMac"); + int imac = of_machine_is_compatible("iMac"); int reg, oreg; reg = oreg = snd_pmac_burgundy_rcb(chip, MASK_ADDR_BURGUNDY_MORE_OUTPUTENABLES); @@ -620,7 +620,7 @@ static void snd_pmac_burgundy_update_automute(struct snd_pmac *chip, int do_noti */ int __devinit snd_pmac_burgundy_init(struct snd_pmac *chip) { - int imac = machine_is_compatible("iMac"); + int imac = of_machine_is_compatible("iMac"); int i, err; /* Checks to see the chip is alive and kicking */ diff --git a/sound/ppc/pmac.c b/sound/ppc/pmac.c index 7bc492e..8508117 100644 --- a/sound/ppc/pmac.c +++ b/sound/ppc/pmac.c @@ -922,11 +922,11 @@ static void __devinit detect_byte_swap(struct snd_pmac *chip) } /* it seems the Pismo & iBook can't byte-swap in hardware. */ - if (machine_is_compatible("PowerBook3,1") || - machine_is_compatible("PowerBook2,1")) + if (of_machine_is_compatible("PowerBook3,1") || + of_machine_is_compatible("PowerBook2,1")) chip->can_byte_swap = 0 ; - if (machine_is_compatible("PowerBook2,1")) + if (of_machine_is_compatible("PowerBook2,1")) chip->can_duplex = 0; } @@ -959,11 +959,11 @@ static int __devinit snd_pmac_detect(struct snd_pmac *chip) chip->control_mask = MASK_IEPC | MASK_IEE | 0x11; /* default */ /* check machine type */ - if (machine_is_compatible("AAPL,3400/2400") - || machine_is_compatible("AAPL,3500")) + if (of_machine_is_compatible("AAPL,3400/2400") + || of_machine_is_compatible("AAPL,3500")) chip->is_pbook_3400 = 1; - else if (machine_is_compatible("PowerBook1,1") - || machine_is_compatible("AAPL,PowerBook1998")) + else if (of_machine_is_compatible("PowerBook1,1") + || of_machine_is_compatible("AAPL,PowerBook1998")) chip->is_pbook_G3 = 1; chip->node = of_find_node_by_name(NULL, "awacs"); sound = of_node_get(chip->node); @@ -1033,8 +1033,8 @@ static int __devinit snd_pmac_detect(struct snd_pmac *chip) } if (of_device_is_compatible(sound, "tumbler")) { chip->model = PMAC_TUMBLER; - chip->can_capture = machine_is_compatible("PowerMac4,2") - || machine_is_compatible("PowerBook4,1"); + chip->can_capture = of_machine_is_compatible("PowerMac4,2") + || of_machine_is_compatible("PowerBook4,1"); chip->can_duplex = 0; // chip->can_byte_swap = 0; /* FIXME: check this */ chip->num_freqs = ARRAY_SIZE(tumbler_freqs); diff --git a/sound/soc/fsl/efika-audio-fabric.c b/sound/soc/fsl/efika-audio-fabric.c index 3326e2a..1a5b8e0 100644 --- a/sound/soc/fsl/efika-audio-fabric.c +++ b/sound/soc/fsl/efika-audio-fabric.c @@ -55,7 +55,7 @@ static __init int efika_fabric_init(void) struct platform_device *pdev; int rc; - if (!machine_is_compatible("bplan,efika")) + if (!of_machine_is_compatible("bplan,efika")) return -ENODEV; card.platform = &mpc5200_audio_dma_platform; diff --git a/sound/soc/fsl/pcm030-audio-fabric.c b/sound/soc/fsl/pcm030-audio-fabric.c index b928ef7..6644cba 100644 --- a/sound/soc/fsl/pcm030-audio-fabric.c +++ b/sound/soc/fsl/pcm030-audio-fabric.c @@ -55,7 +55,7 @@ static __init int pcm030_fabric_init(void) struct platform_device *pdev; int rc; - if (!machine_is_compatible("phytec,pcm030")) + if (!of_machine_is_compatible("phytec,pcm030")) return -ENODEV; card.platform = &mpc5200_audio_dma_platform; -- cgit v0.10.2 From 51975db0b7333cf389b64b5040c2a910341d241a Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Mon, 1 Feb 2010 21:34:14 -0700 Subject: of/flattree: merge early_init_dt_scan_memory() common code Merge common code between PowerPC and Microblaze architectures. Signed-off-by: Grant Likely Acked-by: Michal Simek diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index f7bd0ee..459c32e 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -45,61 +45,14 @@ /* export that to outside world */ struct device_node *of_chosen; -#define early_init_dt_scan_drconf_memory(node) 0 - void __init early_init_dt_scan_chosen_arch(unsigned long node) { /* No Microblaze specific code here */ } -static int __init early_init_dt_scan_memory(unsigned long node, - const char *uname, int depth, void *data) +void __init early_init_dt_add_memory_arch(u64 base, u64 size) { - char *type = of_get_flat_dt_prop(node, "device_type", NULL); - __be32 *reg, *endp; - unsigned long l; - - /* Look for the ibm,dynamic-reconfiguration-memory node */ -/* if (depth == 1 && - strcmp(uname, "ibm,dynamic-reconfiguration-memory") == 0) - return early_init_dt_scan_drconf_memory(node); -*/ - /* We are scanning "memory" nodes only */ - if (type == NULL) { - /* - * The longtrail doesn't have a device_type on the - * /memory node, so look for the node called /memory@0. - */ - if (depth != 1 || strcmp(uname, "memory@0") != 0) - return 0; - } else if (strcmp(type, "memory") != 0) - return 0; - - reg = (__be32 *)of_get_flat_dt_prop(node, "linux,usable-memory", &l); - if (reg == NULL) - reg = (__be32 *)of_get_flat_dt_prop(node, "reg", &l); - if (reg == NULL) - return 0; - - endp = reg + (l / sizeof(__be32)); - - pr_debug("memory scan node %s, reg size %ld, data: %x %x %x %x,\n", - uname, l, reg[0], reg[1], reg[2], reg[3]); - - while ((endp - reg) >= (dt_root_addr_cells + dt_root_size_cells)) { - u64 base, size; - - base = dt_mem_next_cell(dt_root_addr_cells, ®); - size = dt_mem_next_cell(dt_root_size_cells, ®); - - if (size == 0) - continue; - pr_debug(" - %llx , %llx\n", (unsigned long long)base, - (unsigned long long)size); - - lmb_add(base, size); - } - return 0; + lmb_add(base, size); } #ifdef CONFIG_EARLY_PRINTK diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 28be19a..e0f368f 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -483,64 +483,31 @@ static int __init early_init_dt_scan_drconf_memory(unsigned long node) #define early_init_dt_scan_drconf_memory(node) 0 #endif /* CONFIG_PPC_PSERIES */ -static int __init early_init_dt_scan_memory(unsigned long node, - const char *uname, int depth, void *data) +static int __init early_init_dt_scan_memory_ppc(unsigned long node, + const char *uname, + int depth, void *data) { - char *type = of_get_flat_dt_prop(node, "device_type", NULL); - __be32 *reg, *endp; - unsigned long l; - - /* Look for the ibm,dynamic-reconfiguration-memory node */ if (depth == 1 && strcmp(uname, "ibm,dynamic-reconfiguration-memory") == 0) return early_init_dt_scan_drconf_memory(node); + + return early_init_dt_scan_memory(node, uname, depth, data); +} - /* We are scanning "memory" nodes only */ - if (type == NULL) { - /* - * The longtrail doesn't have a device_type on the - * /memory node, so look for the node called /memory@0. - */ - if (depth != 1 || strcmp(uname, "memory@0") != 0) - return 0; - } else if (strcmp(type, "memory") != 0) - return 0; - - reg = of_get_flat_dt_prop(node, "linux,usable-memory", &l); - if (reg == NULL) - reg = of_get_flat_dt_prop(node, "reg", &l); - if (reg == NULL) - return 0; - - endp = reg + (l / sizeof(__be32)); - - DBG("memory scan node %s, reg size %ld, data: %x %x %x %x,\n", - uname, l, reg[0], reg[1], reg[2], reg[3]); - - while ((endp - reg) >= (dt_root_addr_cells + dt_root_size_cells)) { - u64 base, size; - - base = dt_mem_next_cell(dt_root_addr_cells, ®); - size = dt_mem_next_cell(dt_root_size_cells, ®); - - if (size == 0) - continue; - DBG(" - %llx , %llx\n", (unsigned long long)base, - (unsigned long long)size); -#ifdef CONFIG_PPC64 - if (iommu_is_off) { - if (base >= 0x80000000ul) - continue; - if ((base + size) > 0x80000000ul) - size = 0x80000000ul - base; - } +void __init early_init_dt_add_memory_arch(u64 base, u64 size) +{ +#if defined(CONFIG_PPC64) + if (iommu_is_off) { + if (base >= 0x80000000ul) + return; + if ((base + size) > 0x80000000ul) + size = 0x80000000ul - base; + } #endif - lmb_add(base, size); - memstart_addr = min((u64)memstart_addr, base); - } + lmb_add(base, size); - return 0; + memstart_addr = min((u64)memstart_addr, base); } static void __init early_reserve_mem(void) @@ -706,7 +673,7 @@ void __init early_init_devtree(void *params) /* Scan memory nodes and rebuild LMBs */ lmb_init(); of_scan_flat_dt(early_init_dt_scan_root, NULL); - of_scan_flat_dt(early_init_dt_scan_memory, NULL); + of_scan_flat_dt(early_init_dt_scan_memory_ppc, NULL); /* Save command line for /proc/cmdline and then parse parameters */ strlcpy(boot_command_line, cmd_line, COMMAND_LINE_SIZE); diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 7f88611..f84152d 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -15,6 +15,7 @@ #include #include + #ifdef CONFIG_PPC #include #endif /* CONFIG_PPC */ @@ -443,6 +444,55 @@ u64 __init dt_mem_next_cell(int s, u32 **cellp) return of_read_number(p, s); } +/** + * early_init_dt_scan_memory - Look for an parse memory nodes + */ +int __init early_init_dt_scan_memory(unsigned long node, const char *uname, + int depth, void *data) +{ + char *type = of_get_flat_dt_prop(node, "device_type", NULL); + __be32 *reg, *endp; + unsigned long l; + + /* We are scanning "memory" nodes only */ + if (type == NULL) { + /* + * The longtrail doesn't have a device_type on the + * /memory node, so look for the node called /memory@0. + */ + if (depth != 1 || strcmp(uname, "memory@0") != 0) + return 0; + } else if (strcmp(type, "memory") != 0) + return 0; + + reg = of_get_flat_dt_prop(node, "linux,usable-memory", &l); + if (reg == NULL) + reg = of_get_flat_dt_prop(node, "reg", &l); + if (reg == NULL) + return 0; + + endp = reg + (l / sizeof(__be32)); + + pr_debug("memory scan node %s, reg size %ld, data: %x %x %x %x,\n", + uname, l, reg[0], reg[1], reg[2], reg[3]); + + while ((endp - reg) >= (dt_root_addr_cells + dt_root_size_cells)) { + u64 base, size; + + base = dt_mem_next_cell(dt_root_addr_cells, ®); + size = dt_mem_next_cell(dt_root_size_cells, ®); + + if (size == 0) + continue; + pr_debug(" - %llx , %llx\n", (unsigned long long)base, + (unsigned long long)size); + + early_init_dt_add_memory_arch(base, size); + } + + return 0; +} + int __init early_init_dt_scan_chosen(unsigned long node, const char *uname, int depth, void *data) { diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index fbf2961..bf26bd5 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -75,6 +75,9 @@ extern void early_init_dt_scan_chosen_arch(unsigned long node); extern int early_init_dt_scan_chosen(unsigned long node, const char *uname, int depth, void *data); extern void early_init_dt_check_for_initrd(unsigned long node); +extern int early_init_dt_scan_memory(unsigned long node, const char *uname, + int depth, void *data); +extern void early_init_dt_add_memory_arch(u64 base, u64 size); extern u64 dt_mem_next_cell(int s, u32 **cellp); /* Early flat tree scan hooks */ -- cgit v0.10.2 From a9f2f63a671d5e91ed89ade408d87f1692a373de Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Mon, 1 Feb 2010 21:34:14 -0700 Subject: of: include linux/proc_fs.h We use a few procfs-specific functions (eg, proc_device_tree_*) which aren't covered by the current includes. This causes the following build error on arm: drivers/of/base.c: In function 'prom_add_property': drivers/of/base.c:861: error: implicit declaration of function 'proc_device_tree_add_prop' drivers/of/base.c: In function 'prom_remove_property': drivers/of/base.c:902: error: implicit declaration of function 'proc_device_tree_remove_prop' drivers/of/base.c: In function 'prom_update_property': drivers/of/base.c:946: error: implicit declaration of function 'proc_device_tree_update_prop' Add proc_fs.h for these prototypes. Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely diff --git a/drivers/of/base.c b/drivers/of/base.c index 6bc8740..524645a 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -20,6 +20,7 @@ #include #include #include +#include struct device_node *allnodes; -- cgit v0.10.2 From 8cfb3343f70bcf9403218df120ecf345f06dd585 Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Mon, 1 Feb 2010 21:34:14 -0700 Subject: of: make set_node_proc_entry private to proc_devtree.c We only need set_node_proc_entry in proc_devtree.c, so move it there. This fixes the !HAVE_ARCH_DEVTREE_FIXUPS build, as we can't make make the definition in linux/of.h conditional on this #define (definitions in asm/prom.h can't be exposed to linux/of.h, due to the enforced #include ordering). Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely diff --git a/fs/proc/proc_devtree.c b/fs/proc/proc_devtree.c index 123257b..2309bf1 100644 --- a/fs/proc/proc_devtree.c +++ b/fs/proc/proc_devtree.c @@ -14,12 +14,13 @@ #include #include "internal.h" -#ifndef HAVE_ARCH_DEVTREE_FIXUPS static inline void set_node_proc_entry(struct device_node *np, struct proc_dir_entry *de) { -} +#ifdef HAVE_ARCH_DEVTREE_FIXUPS + np->pde = de; #endif +} static struct proc_dir_entry *proc_device_tree; diff --git a/include/linux/of.h b/include/linux/of.h index 3cc0d7a..fd47c81 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -73,12 +73,6 @@ static inline void of_node_set_flag(struct device_node *n, unsigned long flag) set_bit(flag, &n->_flags); } -static inline void -set_node_proc_entry(struct device_node *dn, struct proc_dir_entry *de) -{ - dn->pde = de; -} - extern struct device_node *of_find_all_nodes(struct device_node *prev); #if defined(CONFIG_SPARC) -- cgit v0.10.2 From 50ab2fe147e22c8786552cda1791a61ae81b84d2 Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Mon, 1 Feb 2010 21:34:14 -0700 Subject: proc_devtree: include linux/of.h Currenly, proc_devtree.c depends on asm/prom.h to include linux/of.h, to provide some device-tree definitions (eg, struct property). Instead, include linux/of.h directly. We still need asm/prom.h for HAVE_ARCH_DEVTREE_FIXUPS. Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely diff --git a/fs/proc/proc_devtree.c b/fs/proc/proc_devtree.c index 2309bf1..0ec4511 100644 --- a/fs/proc/proc_devtree.c +++ b/fs/proc/proc_devtree.c @@ -10,6 +10,7 @@ #include #include #include +#include #include #include #include "internal.h" -- cgit v0.10.2 From 1406bc2f57787797d1f6a3675c019a7093769275 Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Sat, 30 Jan 2010 01:31:21 -0700 Subject: of/flattree: use callback to setup initrd from /chosen At present, the fdt code sets the kernel-wide initrd_start and initrd_end variables when parsing /chosen. On ARM, we only set these once the bootmem has been reserved. This change adds an arch hook to setup the initrd from the device tree: void early_init_dt_setup_initrd_arch(unsigned long start, unsigned long end); The arch-specific code can then setup the initrd however it likes. Compiled on powerpc, with CONFIG_BLK_DEV_INITRD=y and =n. Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 459c32e..050b799 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -118,6 +118,16 @@ void __init early_init_devtree(void *params) pr_debug(" <- early_init_devtree()\n"); } +#ifdef CONFIG_BLK_DEV_INITRD +void __init early_init_dt_setup_initrd_arch(unsigned long start, + unsigned long end) +{ + initrd_start = (unsigned long)__va(start); + initrd_end = (unsigned long)__va(end); + initrd_below_start_ok = 1; +} +#endif + /******* * * New implementation of the OF "find" APIs, return a refcounted diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index e0f368f..40fce1c 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -510,6 +510,16 @@ void __init early_init_dt_add_memory_arch(u64 base, u64 size) memstart_addr = min((u64)memstart_addr, base); } +#ifdef CONFIG_BLK_DEV_INITRD +void __init early_init_dt_setup_initrd_arch(unsigned long start, + unsigned long end) +{ + initrd_start = (unsigned long)__va(start); + initrd_end = (unsigned long)__va(end); + initrd_below_start_ok = 1; +} +#endif + static void __init early_reserve_mem(void) { u64 base, size; diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index f84152d..9290ca5 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -384,28 +384,23 @@ unsigned long __init unflatten_dt_node(unsigned long mem, */ void __init early_init_dt_check_for_initrd(unsigned long node) { - unsigned long len; + unsigned long start, end, len; u32 *prop; pr_debug("Looking for initrd properties... "); prop = of_get_flat_dt_prop(node, "linux,initrd-start", &len); - if (prop) { - initrd_start = (unsigned long) - __va(of_read_ulong(prop, len/4)); - - prop = of_get_flat_dt_prop(node, "linux,initrd-end", &len); - if (prop) { - initrd_end = (unsigned long) - __va(of_read_ulong(prop, len/4)); - initrd_below_start_ok = 1; - } else { - initrd_start = 0; - } - } + if (!prop) + return; + start = of_read_ulong(prop, len/4); + + prop = of_get_flat_dt_prop(node, "linux,initrd-end", &len); + if (!prop) + return; + end = of_read_ulong(prop, len/4); - pr_debug("initrd_start=0x%lx initrd_end=0x%lx\n", - initrd_start, initrd_end); + early_init_dt_setup_initrd_arch(start, end); + pr_debug("initrd_start=0x%lx initrd_end=0x%lx\n", start, end); } #else inline void early_init_dt_check_for_initrd(unsigned long node) diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index bf26bd5..f32f0fc 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -80,6 +80,16 @@ extern int early_init_dt_scan_memory(unsigned long node, const char *uname, extern void early_init_dt_add_memory_arch(u64 base, u64 size); extern u64 dt_mem_next_cell(int s, u32 **cellp); +/* + * If BLK_DEV_INITRD, the fdt early init code will call this function, + * to be provided by the arch code. start and end are specified as + * physical addresses. + */ +#ifdef CONFIG_BLK_DEV_INITRD +extern void early_init_dt_setup_initrd_arch(unsigned long start, + unsigned long end); +#endif + /* Early flat tree scan hooks */ extern int early_init_dt_scan_root(unsigned long node, const char *uname, int depth, void *data); -- cgit v0.10.2 From 36b9d3070d653af5807cef74ff129721d9047107 Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Mon, 1 Feb 2010 21:34:14 -0700 Subject: of/flattree: use OF_ROOT_NODE_{SIZE,ADDR}_CELLS DEFAULT for fdt parsing At present we're using hard-coded values for defaults when parsing the FDT. This change uses the #defines instead. Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 9290ca5..56fbd6e 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -420,11 +420,11 @@ int __init early_init_dt_scan_root(unsigned long node, const char *uname, return 0; prop = of_get_flat_dt_prop(node, "#size-cells", NULL); - dt_root_size_cells = (prop == NULL) ? 1 : *prop; + dt_root_size_cells = prop ? *prop : OF_ROOT_NODE_SIZE_CELLS_DEFAULT; pr_debug("dt_root_size_cells = %x\n", dt_root_size_cells); prop = of_get_flat_dt_prop(node, "#address-cells", NULL); - dt_root_addr_cells = (prop == NULL) ? 2 : *prop; + dt_root_addr_cells = prop ? *prop : OF_ROOT_NODE_ADDR_CELLS_DEFAULT; pr_debug("dt_root_addr_cells = %x\n", dt_root_addr_cells); /* break now */ -- cgit v0.10.2 From 2e89e685a8fd0e8334de967739d11e2e28c1a4dd Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Sat, 30 Jan 2010 01:41:49 -0700 Subject: of: use __be32 for cell value accessors Currently, we're using u32 for cell values, and hence assuming host-endian device trees. As we'd like to support little-endian platforms, use a __be32 for cell values, and convert in the cell accessors. Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 56fbd6e..968a86a 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -431,9 +431,9 @@ int __init early_init_dt_scan_root(unsigned long node, const char *uname, return 1; } -u64 __init dt_mem_next_cell(int s, u32 **cellp) +u64 __init dt_mem_next_cell(int s, __be32 **cellp) { - u32 *p = *cellp; + __be32 *p = *cellp; *cellp = p + s; return of_read_number(p, s); diff --git a/include/linux/of.h b/include/linux/of.h index fd47c81..fa5571b 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -20,6 +20,8 @@ #include #include +#include + typedef u32 phandle; typedef u32 ihandle; @@ -95,16 +97,16 @@ extern void of_node_put(struct device_node *node); */ /* Helper to read a big number; size is in cells (not bytes) */ -static inline u64 of_read_number(const u32 *cell, int size) +static inline u64 of_read_number(const __be32 *cell, int size) { u64 r = 0; while (size--) - r = (r << 32) | *(cell++); + r = (r << 32) | be32_to_cpu(*(cell++)); return r; } /* Like of_read_number, but we want an unsigned long result */ -static inline unsigned long of_read_ulong(const u32 *cell, int size) +static inline unsigned long of_read_ulong(const __be32 *cell, int size) { /* toss away upper bits if unsigned long is smaller than u64 */ return of_read_number(cell, size); diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index f32f0fc..6a35d91 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -78,7 +78,7 @@ extern void early_init_dt_check_for_initrd(unsigned long node); extern int early_init_dt_scan_memory(unsigned long node, const char *uname, int depth, void *data); extern void early_init_dt_add_memory_arch(u64 base, u64 size); -extern u64 dt_mem_next_cell(int s, u32 **cellp); +extern u64 dt_mem_next_cell(int s, __be32 **cellp); /* * If BLK_DEV_INITRD, the fdt early init code will call this function, -- cgit v0.10.2 From 337148812f97368a8ec4a69f1691e4c5ce3af494 Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Sat, 30 Jan 2010 01:45:26 -0700 Subject: of: assume big-endian properties, adding conversions where necessary Properties in the device tree are specified as big-endian. At present, the only platforms to support device trees are also big-endian, so we've been acessing the properties as raw values. We'd like to add device tree support to little-endian platforms too, so add endian conversion to the sites where we access property values in the common of code. Compiled on powerpc (ppc44x_defconfig & ppc64_defconfig) and arm (fdt support only for now). Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely diff --git a/drivers/of/base.c b/drivers/of/base.c index 524645a..873479a 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -38,7 +38,7 @@ int of_n_addr_cells(struct device_node *np) np = np->parent; ip = of_get_property(np, "#address-cells", NULL); if (ip) - return *ip; + return be32_to_cpup(ip); } while (np->parent); /* No #address-cells property for the root node */ return OF_ROOT_NODE_ADDR_CELLS_DEFAULT; @@ -54,7 +54,7 @@ int of_n_size_cells(struct device_node *np) np = np->parent; ip = of_get_property(np, "#size-cells", NULL); if (ip) - return *ip; + return be32_to_cpup(ip); } while (np->parent); /* No #size-cells property for the root node */ return OF_ROOT_NODE_SIZE_CELLS_DEFAULT; @@ -696,8 +696,8 @@ int of_parse_phandles_with_args(struct device_node *np, const char *list_name, const void **out_args) { int ret = -EINVAL; - const u32 *list; - const u32 *list_end; + const __be32 *list; + const __be32 *list_end; int size; int cur_index = 0; struct device_node *node = NULL; @@ -711,7 +711,7 @@ int of_parse_phandles_with_args(struct device_node *np, const char *list_name, list_end = list + size / sizeof(*list); while (list < list_end) { - const u32 *cells; + const __be32 *cells; const phandle *phandle; phandle = list++; @@ -735,7 +735,7 @@ int of_parse_phandles_with_args(struct device_node *np, const char *list_name, goto err1; } - list += *cells; + list += be32_to_cpup(cells); if (list > list_end) { pr_debug("%s: insufficient arguments length\n", np->full_name); diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 968a86a..5c5f03e 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -51,7 +51,7 @@ int __init of_scan_flat_dt(int (*it)(unsigned long node, int depth = -1; do { - u32 tag = *((u32 *)p); + u32 tag = be32_to_cpup((__be32 *)p); char *pathp; p += 4; @@ -64,7 +64,7 @@ int __init of_scan_flat_dt(int (*it)(unsigned long node, if (tag == OF_DT_END) break; if (tag == OF_DT_PROP) { - u32 sz = *((u32 *)p); + u32 sz = be32_to_cpup((__be32 *)p); p += 8; if (initial_boot_params->version < 0x10) p = _ALIGN(p, sz >= 8 ? 8 : 4); @@ -103,9 +103,9 @@ unsigned long __init of_get_flat_dt_root(void) unsigned long p = ((unsigned long)initial_boot_params) + initial_boot_params->off_dt_struct; - while (*((u32 *)p) == OF_DT_NOP) + while (be32_to_cpup((__be32 *)p) == OF_DT_NOP) p += 4; - BUG_ON(*((u32 *)p) != OF_DT_BEGIN_NODE); + BUG_ON(be32_to_cpup((__be32 *)p) != OF_DT_BEGIN_NODE); p += 4; return _ALIGN(p + strlen((char *)p) + 1, 4); } @@ -122,7 +122,7 @@ void *__init of_get_flat_dt_prop(unsigned long node, const char *name, unsigned long p = node; do { - u32 tag = *((u32 *)p); + u32 tag = be32_to_cpup((__be32 *)p); u32 sz, noff; const char *nstr; @@ -132,8 +132,8 @@ void *__init of_get_flat_dt_prop(unsigned long node, const char *name, if (tag != OF_DT_PROP) return NULL; - sz = *((u32 *)p); - noff = *((u32 *)(p + 4)); + sz = be32_to_cpup((__be32 *)p); + noff = be32_to_cpup((__be32 *)(p + 4)); p += 8; if (initial_boot_params->version < 0x10) p = _ALIGN(p, sz >= 8 ? 8 : 4); @@ -210,7 +210,7 @@ unsigned long __init unflatten_dt_node(unsigned long mem, int has_name = 0; int new_format = 0; - tag = *((u32 *)(*p)); + tag = be32_to_cpup((__be32 *)(*p)); if (tag != OF_DT_BEGIN_NODE) { pr_err("Weird tag at start of node: %x\n", tag); return mem; @@ -285,7 +285,7 @@ unsigned long __init unflatten_dt_node(unsigned long mem, u32 sz, noff; char *pname; - tag = *((u32 *)(*p)); + tag = be32_to_cpup((__be32 *)(*p)); if (tag == OF_DT_NOP) { *p += 4; continue; @@ -293,8 +293,8 @@ unsigned long __init unflatten_dt_node(unsigned long mem, if (tag != OF_DT_PROP) break; *p += 4; - sz = *((u32 *)(*p)); - noff = *((u32 *)((*p) + 4)); + sz = be32_to_cpup((__be32 *)(*p)); + noff = be32_to_cpup((__be32 *)((*p) + 4)); *p += 8; if (initial_boot_params->version < 0x10) *p = _ALIGN(*p, sz >= 8 ? 8 : 4); @@ -367,7 +367,7 @@ unsigned long __init unflatten_dt_node(unsigned long mem, } while (tag == OF_DT_BEGIN_NODE) { mem = unflatten_dt_node(mem, p, np, allnextpp, fpsize); - tag = *((u32 *)(*p)); + tag = be32_to_cpup((__be32 *)(*p)); } if (tag != OF_DT_END_NODE) { pr_err("Weird tag at end of node: %x\n", tag); @@ -385,7 +385,7 @@ unsigned long __init unflatten_dt_node(unsigned long mem, void __init early_init_dt_check_for_initrd(unsigned long node) { unsigned long start, end, len; - u32 *prop; + __be32 *prop; pr_debug("Looking for initrd properties... "); @@ -414,17 +414,22 @@ inline void early_init_dt_check_for_initrd(unsigned long node) int __init early_init_dt_scan_root(unsigned long node, const char *uname, int depth, void *data) { - u32 *prop; + __be32 *prop; if (depth != 0) return 0; + dt_root_size_cells = OF_ROOT_NODE_SIZE_CELLS_DEFAULT; + dt_root_addr_cells = OF_ROOT_NODE_ADDR_CELLS_DEFAULT; + prop = of_get_flat_dt_prop(node, "#size-cells", NULL); - dt_root_size_cells = prop ? *prop : OF_ROOT_NODE_SIZE_CELLS_DEFAULT; + if (prop) + dt_root_size_cells = be32_to_cpup(prop); pr_debug("dt_root_size_cells = %x\n", dt_root_size_cells); prop = of_get_flat_dt_prop(node, "#address-cells", NULL); - dt_root_addr_cells = prop ? *prop : OF_ROOT_NODE_ADDR_CELLS_DEFAULT; + if (prop) + dt_root_addr_cells = be32_to_cpup(prop); pr_debug("dt_root_addr_cells = %x\n", dt_root_addr_cells); /* break now */ @@ -549,7 +554,7 @@ void __init unflatten_device_tree(void) mem = lmb_alloc(size + 4, __alignof__(struct device_node)); mem = (unsigned long) __va(mem); - ((u32 *)mem)[size / 4] = 0xdeadbeef; + ((__be32 *)mem)[size / 4] = cpu_to_be32(0xdeadbeef); pr_debug(" unflattening %lx...\n", mem); @@ -557,11 +562,11 @@ void __init unflatten_device_tree(void) start = ((unsigned long)initial_boot_params) + initial_boot_params->off_dt_struct; unflatten_dt_node(mem, &start, NULL, &allnextp, 0); - if (*((u32 *)start) != OF_DT_END) + if (be32_to_cpup((__be32 *)start) != OF_DT_END) pr_warning("Weird tag at end of tree: %08x\n", *((u32 *)start)); - if (((u32 *)mem)[size / 4] != 0xdeadbeef) + if (be32_to_cpu(((__be32 *)mem)[size / 4]) != 0xdeadbeef) pr_warning("End of tree marker overwritten: %08x\n", - ((u32 *)mem)[size / 4]); + be32_to_cpu(((__be32 *)mem)[size / 4])); *allnextp = NULL; /* Get pointer to OF "/chosen" node for use everywhere */ diff --git a/drivers/of/gpio.c b/drivers/of/gpio.c index 6eea601..24c3606 100644 --- a/drivers/of/gpio.c +++ b/drivers/of/gpio.c @@ -36,7 +36,7 @@ int of_get_gpio_flags(struct device_node *np, int index, struct of_gpio_chip *of_gc = NULL; int size; const void *gpio_spec; - const u32 *gpio_cells; + const __be32 *gpio_cells; ret = of_parse_phandles_with_args(np, "gpios", "#gpio-cells", index, &gc, &gpio_spec); @@ -55,7 +55,7 @@ int of_get_gpio_flags(struct device_node *np, int index, gpio_cells = of_get_property(gc, "#gpio-cells", &size); if (!gpio_cells || size != sizeof(*gpio_cells) || - *gpio_cells != of_gc->gpio_cells) { + be32_to_cpup(gpio_cells) != of_gc->gpio_cells) { pr_debug("%s: wrong #gpio-cells for %s\n", np->full_name, gc->full_name); ret = -EINVAL; @@ -127,7 +127,8 @@ EXPORT_SYMBOL(of_gpio_count); int of_gpio_simple_xlate(struct of_gpio_chip *of_gc, struct device_node *np, const void *gpio_spec, enum of_gpio_flags *flags) { - const u32 *gpio = gpio_spec; + const __be32 *gpio = gpio_spec; + const u32 n = be32_to_cpup(gpio); /* * We're discouraging gpio_cells < 2, since that way you'll have to @@ -140,13 +141,13 @@ int of_gpio_simple_xlate(struct of_gpio_chip *of_gc, struct device_node *np, return -EINVAL; } - if (*gpio > of_gc->gc.ngpio) + if (n > of_gc->gc.ngpio) return -EINVAL; if (flags) - *flags = gpio[1]; + *flags = be32_to_cpu(gpio[1]); - return *gpio; + return n; } EXPORT_SYMBOL(of_gpio_simple_xlate); diff --git a/drivers/of/of_i2c.c b/drivers/of/of_i2c.c index fa65a2b..a3a708e 100644 --- a/drivers/of/of_i2c.c +++ b/drivers/of/of_i2c.c @@ -25,7 +25,7 @@ void of_register_i2c_devices(struct i2c_adapter *adap, for_each_child_of_node(adap_node, node) { struct i2c_board_info info = {}; struct dev_archdata dev_ad = {}; - const u32 *addr; + const __be32 *addr; int len; if (of_modalias_node(node, info.type, sizeof(info.type)) < 0) @@ -40,7 +40,7 @@ void of_register_i2c_devices(struct i2c_adapter *adap, info.irq = irq_of_parse_and_map(node, 0); - info.addr = *addr; + info.addr = be32_to_cpup(addr); dev_archdata_set_node(&dev_ad, node); info.archdata = &dev_ad; diff --git a/drivers/of/of_mdio.c b/drivers/of/of_mdio.c index 4b22ba5..18ecae4 100644 --- a/drivers/of/of_mdio.c +++ b/drivers/of/of_mdio.c @@ -51,7 +51,7 @@ int of_mdiobus_register(struct mii_bus *mdio, struct device_node *np) /* Loop over the child nodes and register a phy_device for each one */ for_each_child_of_node(np, child) { - const u32 *addr; + const __be32 *addr; int len; /* A PHY must have a reg property in the range [0-31] */ @@ -68,7 +68,7 @@ int of_mdiobus_register(struct mii_bus *mdio, struct device_node *np) mdio->irq[*addr] = PHY_POLL; } - phy = get_phy_device(mdio, *addr); + phy = get_phy_device(mdio, be32_to_cpup(addr)); if (!phy) { dev_err(&mdio->dev, "error probing PHY at address %i\n", *addr); @@ -160,7 +160,7 @@ struct phy_device *of_phy_connect_fixed_link(struct net_device *dev, struct device_node *net_np; char bus_id[MII_BUS_ID_SIZE + 3]; struct phy_device *phy; - const u32 *phy_id; + const __be32 *phy_id; int sz; if (!dev->dev.parent) @@ -174,7 +174,7 @@ struct phy_device *of_phy_connect_fixed_link(struct net_device *dev, if (!phy_id || sz < sizeof(*phy_id)) return NULL; - sprintf(bus_id, PHY_ID_FMT, "0", phy_id[0]); + sprintf(bus_id, PHY_ID_FMT, "0", be32_to_cpu(phy_id[0])); phy = phy_connect(dev, bus_id, hndlr, 0, iface); return IS_ERR(phy) ? NULL : phy; diff --git a/drivers/of/of_spi.c b/drivers/of/of_spi.c index bed0ed6..f65f48b 100644 --- a/drivers/of/of_spi.c +++ b/drivers/of/of_spi.c @@ -23,7 +23,7 @@ void of_register_spi_devices(struct spi_master *master, struct device_node *np) { struct spi_device *spi; struct device_node *nc; - const u32 *prop; + const __be32 *prop; int rc; int len; @@ -54,7 +54,7 @@ void of_register_spi_devices(struct spi_master *master, struct device_node *np) spi_dev_put(spi); continue; } - spi->chip_select = *prop; + spi->chip_select = be32_to_cpup(prop); /* Mode (clock phase/polarity/etc.) */ if (of_find_property(nc, "spi-cpha", NULL)) @@ -72,7 +72,7 @@ void of_register_spi_devices(struct spi_master *master, struct device_node *np) spi_dev_put(spi); continue; } - spi->max_speed_hz = *prop; + spi->max_speed_hz = be32_to_cpup(prop); /* IRQ */ spi->irq = irq_of_parse_and_map(nc, 0); -- cgit v0.10.2 From 087f79c48c090a2c0cd9ee45231d63290d2036d2 Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Sat, 30 Jan 2010 04:14:19 -0700 Subject: of/flattree: endian-convert members of boot_param_header The boot_param_header has big-endian fields, so change the types to __be32, and perform endian conversion when we access them. Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 40fce1c..43c78d7 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -98,7 +98,7 @@ static void __init move_device_tree(void) DBG("-> move_device_tree\n"); start = __pa(initial_boot_params); - size = initial_boot_params->totalsize; + size = be32_to_cpu(initial_boot_params->totalsize); if ((memory_limit && (start + size) > memory_limit) || overlaps_crashkernel(start, size)) { diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 5c5f03e..18d282f 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -28,7 +28,7 @@ struct boot_param_header *initial_boot_params; char *find_flat_dt_string(u32 offset) { return ((char *)initial_boot_params) + - initial_boot_params->off_dt_strings + offset; + be32_to_cpu(initial_boot_params->off_dt_strings) + offset; } /** @@ -46,7 +46,7 @@ int __init of_scan_flat_dt(int (*it)(unsigned long node, void *data) { unsigned long p = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; + be32_to_cpu(initial_boot_params->off_dt_struct); int rc = 0; int depth = -1; @@ -66,7 +66,7 @@ int __init of_scan_flat_dt(int (*it)(unsigned long node, if (tag == OF_DT_PROP) { u32 sz = be32_to_cpup((__be32 *)p); p += 8; - if (initial_boot_params->version < 0x10) + if (be32_to_cpu(initial_boot_params->version) < 0x10) p = _ALIGN(p, sz >= 8 ? 8 : 4); p += sz; p = _ALIGN(p, 4); @@ -101,7 +101,7 @@ int __init of_scan_flat_dt(int (*it)(unsigned long node, unsigned long __init of_get_flat_dt_root(void) { unsigned long p = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; + be32_to_cpu(initial_boot_params->off_dt_struct); while (be32_to_cpup((__be32 *)p) == OF_DT_NOP) p += 4; @@ -135,7 +135,7 @@ void *__init of_get_flat_dt_prop(unsigned long node, const char *name, sz = be32_to_cpup((__be32 *)p); noff = be32_to_cpup((__be32 *)(p + 4)); p += 8; - if (initial_boot_params->version < 0x10) + if (be32_to_cpu(initial_boot_params->version) < 0x10) p = _ALIGN(p, sz >= 8 ? 8 : 4); nstr = find_flat_dt_string(noff); @@ -296,7 +296,7 @@ unsigned long __init unflatten_dt_node(unsigned long mem, sz = be32_to_cpup((__be32 *)(*p)); noff = be32_to_cpup((__be32 *)((*p) + 4)); *p += 8; - if (initial_boot_params->version < 0x10) + if (be32_to_cpu(initial_boot_params->version) < 0x10) *p = _ALIGN(*p, sz >= 8 ? 8 : 4); pname = find_flat_dt_string(noff); @@ -544,7 +544,7 @@ void __init unflatten_device_tree(void) /* First pass, scan for size */ start = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; + be32_to_cpu(initial_boot_params->off_dt_struct); size = unflatten_dt_node(0, &start, NULL, NULL, 0); size = (size | 3) + 1; @@ -560,7 +560,7 @@ void __init unflatten_device_tree(void) /* Second pass, do actual unflattening */ start = ((unsigned long)initial_boot_params) + - initial_boot_params->off_dt_struct; + be32_to_cpu(initial_boot_params->off_dt_struct); unflatten_dt_node(mem, &start, NULL, &allnextp, 0); if (be32_to_cpup((__be32 *)start) != OF_DT_END) pr_warning("Weird tag at end of tree: %08x\n", *((u32 *)start)); diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index 6a35d91..0007187 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -42,19 +42,19 @@ * ends when size is 0 */ struct boot_param_header { - u32 magic; /* magic word OF_DT_HEADER */ - u32 totalsize; /* total size of DT block */ - u32 off_dt_struct; /* offset to structure */ - u32 off_dt_strings; /* offset to strings */ - u32 off_mem_rsvmap; /* offset to memory reserve map */ - u32 version; /* format version */ - u32 last_comp_version; /* last compatible version */ + __be32 magic; /* magic word OF_DT_HEADER */ + __be32 totalsize; /* total size of DT block */ + __be32 off_dt_struct; /* offset to structure */ + __be32 off_dt_strings; /* offset to strings */ + __be32 off_mem_rsvmap; /* offset to memory reserve map */ + __be32 version; /* format version */ + __be32 last_comp_version; /* last compatible version */ /* version 2 fields below */ - u32 boot_cpuid_phys; /* Physical CPU id we're booting on */ + __be32 boot_cpuid_phys; /* Physical CPU id we're booting on */ /* version 3 fields below */ - u32 dt_strings_size; /* size of the DT strings block */ + __be32 dt_strings_size; /* size of the DT strings block */ /* version 17 fields below */ - u32 dt_struct_size; /* size of the DT structure block */ + __be32 dt_struct_size; /* size of the DT structure block */ }; /* TBD: Temporary export of fdt globals - remove when code fully merged */ -- cgit v0.10.2 From 04b954a673dd02f585a2769c4945a43880faa989 Mon Sep 17 00:00:00 2001 From: David Gibson Date: Mon, 1 Feb 2010 21:34:15 -0700 Subject: of/flattree: Make the kernel accept ePAPR style phandle information Currently when processing flattened device trees, the kernel expects the phandle in a property called "linux,phandle". The ePAPR spec - not being Linux specific - instead requires phandles to be encoded in a property named simply "phandle". This patch makes the kernel accept either form when unflattening the device tree. Signed-off-by: David Gibson Signed-off-by: Grant Likely diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index 18d282f..b51f797 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -310,10 +310,19 @@ unsigned long __init unflatten_dt_node(unsigned long mem, pp = unflatten_dt_alloc(&mem, sizeof(struct property), __alignof__(struct property)); if (allnextpp) { - if (strcmp(pname, "linux,phandle") == 0) { + /* We accept flattened tree phandles either in + * ePAPR-style "phandle" properties, or the + * legacy "linux,phandle" properties. If both + * appear and have different values, things + * will get weird. Don't do that. */ + if ((strcmp(pname, "phandle") == 0) || + (strcmp(pname, "linux,phandle") == 0)) { if (np->phandle == 0) np->phandle = *((u32 *)*p); } + /* And we process the "ibm,phandle" property + * used in pSeries dynamic device tree + * stuff */ if (strcmp(pname, "ibm,phandle") == 0) np->phandle = *((u32 *)*p); pp->name = pname; -- cgit v0.10.2 From ede55c9d78101fef0d8e620940a5163f14b02f29 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 27 Jan 2010 11:25:54 -0500 Subject: tracing: Add correct/incorrect to sort keys for branch annotation output The branch annotation is a bit difficult to see the worst offenders because it only sorts by percentage: correct incorrect % Function File Line ------- --------- - -------- ---- ---- 0 163 100 qdisc_restart sch_generic.c 179 0 163 100 pfifo_fast_dequeue sch_generic.c 447 0 4 100 pskb_trim_rcsum skbuff.h 1689 0 4 100 llc_rcv llc_input.c 170 0 18 100 psmouse_interrupt psmouse-base.c 304 0 3 100 atkbd_interrupt atkbd.c 389 0 5 100 usb_alloc_dev usb.c 437 0 11 100 vsscanf vsprintf.c 1897 0 2 100 IS_ERR err.h 34 0 23 100 __rmqueue_fallback page_alloc.c 865 0 4 100 probe_wakeup_sched_switch trace_sched_wakeup.c 142 0 3 100 move_masked_irq migration.c 11 Adding the incorrect and correct values as sort keys makes this file a bit more informative: correct incorrect % Function File Line ------- --------- - -------- ---- ---- 0 366541 100 audit_syscall_entry auditsc.c 1637 0 366538 100 audit_syscall_exit auditsc.c 1685 0 115839 100 sched_info_switch sched_stats.h 269 0 74567 100 sched_info_queued sched_stats.h 222 0 66578 100 sched_info_dequeued sched_stats.h 177 0 15113 100 trace_workqueue_insertion workqueue.h 38 0 15107 100 trace_workqueue_execution workqueue.h 45 0 3622 100 syscall_trace_leave ptrace.c 1772 0 2750 100 sched_move_task sched.c 10100 0 2750 100 sched_move_task sched.c 10110 0 1815 100 pre_schedule_rt sched_rt.c 1462 0 837 100 audit_alloc auditsc.c 879 0 814 100 tcp_mss_split_point tcp_output.c 1302 Signed-off-by: Steven Rostedt diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 4a194f0..b9bc4d4 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -307,8 +307,23 @@ static int annotated_branch_stat_cmp(void *p1, void *p2) return -1; if (percent_a > percent_b) return 1; - else - return 0; + + if (a->incorrect < b->incorrect) + return -1; + if (a->incorrect > b->incorrect) + return 1; + + /* + * Since the above shows worse (incorrect) cases + * first, we continue that by showing best (correct) + * cases last. + */ + if (a->correct > b->correct) + return -1; + if (a->correct < b->correct) + return 1; + + return 0; } static struct tracer_stat annotated_branch_stats = { -- cgit v0.10.2 From 7578a4c625a5cc32812946338a4549f3090be113 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 10 Feb 2010 16:00:58 +0900 Subject: sh: Fix up multi-resource mapping for SH7786 PCIe. This reworks some of the SH7786 PCIe initialization code to dynamically setup and size the various resource windows, as opposed to the original code that simply wired in a couple of them statically. At the same time, we tidy up the initialization code a bit, kill off some read-only register twiddling that was gleaned from the bus analyzer, and also propagate the physical slot/channel mapping. Signed-off-by: Paul Mundt diff --git a/arch/sh/drivers/pci/pcie-sh7786.c b/arch/sh/drivers/pci/pcie-sh7786.c index 95d095f..ae91a2d 100644 --- a/arch/sh/drivers/pci/pcie-sh7786.c +++ b/arch/sh/drivers/pci/pcie-sh7786.c @@ -204,7 +204,9 @@ static int pcie_init(struct sh7786_pcie_port *port) { struct pci_channel *chan = port->hose; unsigned int data; - int ret; + phys_addr_t memphys; + size_t memsize; + int ret, i; /* Begin initialization */ pci_write_reg(chan, 0, SH4A_PCIETCTLR); @@ -227,15 +229,24 @@ static int pcie_init(struct sh7786_pcie_port *port) data |= PCI_CAP_ID_EXP; pci_write_reg(chan, data, SH4A_PCIEEXPCAP0); - /* Enable x4 link width and extended sync. */ + /* Enable data link layer active state reporting */ + pci_write_reg(chan, PCI_EXP_LNKCAP_DLLLARC, SH4A_PCIEEXPCAP3); + + /* Enable extended sync and ASPM L0s support */ data = pci_read_reg(chan, SH4A_PCIEEXPCAP4); - data &= ~(PCI_EXP_LNKSTA_NLW << 16); - data |= (1 << 22) | PCI_EXP_LNKCTL_ES; + data &= ~PCI_EXP_LNKCTL_ASPMC; + data |= PCI_EXP_LNKCTL_ES | 1; pci_write_reg(chan, data, SH4A_PCIEEXPCAP4); + /* Write out the physical slot number */ + data = pci_read_reg(chan, SH4A_PCIEEXPCAP5); + data &= ~PCI_EXP_SLTCAP_PSN; + data |= (port->index + 1) << 19; + pci_write_reg(chan, data, SH4A_PCIEEXPCAP5); + /* Set the completion timer timeout to the maximum 32ms. */ data = pci_read_reg(chan, SH4A_PCIETLCTLR); - data &= ~0xffff; + data &= ~0x3f00; data |= 0x32 << 8; pci_write_reg(chan, data, SH4A_PCIETLCTLR); @@ -248,6 +259,33 @@ static int pcie_init(struct sh7786_pcie_port *port) data |= (0xff << 16); pci_write_reg(chan, data, SH4A_PCIEMACCTLR); + memphys = __pa(memory_start); + memsize = roundup_pow_of_two(memory_end - memory_start); + + /* + * If there's more than 512MB of memory, we need to roll over to + * LAR1/LAMR1. + */ + if (memsize > SZ_512M) { + __raw_writel(memphys + SZ_512M, chan->reg_base + SH4A_PCIELAR1); + __raw_writel(((memsize - SZ_512M) - SZ_256) | 1, + chan->reg_base + SH4A_PCIELAMR1); + memsize = SZ_512M; + } else { + /* + * Otherwise just zero it out and disable it. + */ + __raw_writel(0, chan->reg_base + SH4A_PCIELAR1); + __raw_writel(0, chan->reg_base + SH4A_PCIELAMR1); + } + + /* + * LAR0/LAMR0 covers up to the first 512MB, which is enough to + * cover all of lowmem on most platforms. + */ + __raw_writel(memphys, chan->reg_base + SH4A_PCIELAR0); + __raw_writel((memsize - SZ_256) | 1, chan->reg_base + SH4A_PCIELAMR0); + /* Finish initialization */ data = pci_read_reg(chan, SH4A_PCIETCTLR); data |= 0x1; @@ -267,10 +305,14 @@ static int pcie_init(struct sh7786_pcie_port *port) if (unlikely(ret != 0)) return -ENODEV; - pci_write_reg(chan, 0x00100007, SH4A_PCIEPCICONF1); + data = pci_read_reg(chan, SH4A_PCIEPCICONF1); + data &= ~(PCI_STATUS_DEVSEL_MASK << 16); + data |= PCI_COMMAND_IO | PCI_COMMAND_MEMORY | PCI_COMMAND_MASTER | + (PCI_STATUS_CAP_LIST | PCI_STATUS_DEVSEL_FAST) << 16; + pci_write_reg(chan, data, SH4A_PCIEPCICONF1); + pci_write_reg(chan, 0x80888000, SH4A_PCIETXVC0DCTLR); pci_write_reg(chan, 0x00222000, SH4A_PCIERXVC0DCTLR); - pci_write_reg(chan, 0x000050A0, SH4A_PCIEEXPCAP2); wmb(); @@ -278,15 +320,32 @@ static int pcie_init(struct sh7786_pcie_port *port) printk(KERN_NOTICE "PCI: PCIe#%d link width %d\n", port->index, (data >> 20) & 0x3f); - pci_write_reg(chan, 0x007c0000, SH4A_PCIEPAMR0); - pci_write_reg(chan, 0x00000000, SH4A_PCIEPARH0); - pci_write_reg(chan, 0x00000000, SH4A_PCIEPARL0); - pci_write_reg(chan, 0x80000100, SH4A_PCIEPTCTLR0); - pci_write_reg(chan, 0x03fc0000, SH4A_PCIEPAMR2); - pci_write_reg(chan, 0x00000000, SH4A_PCIEPARH2); - pci_write_reg(chan, 0x00000000, SH4A_PCIEPARL2); - pci_write_reg(chan, 0x80000000, SH4A_PCIEPTCTLR2); + for (i = 0; i < chan->nr_resources; i++) { + struct resource *res = chan->resources + i; + resource_size_t size; + u32 enable_mask; + + pci_write_reg(chan, 0x00000000, SH4A_PCIEPTCTLR(i)); + + size = resource_size(res); + + /* + * The PAMR mask is calculated in units of 256kB, which + * keeps things pretty simple. + */ + __raw_writel(((roundup_pow_of_two(size) / SZ_256K) - 1) << 18, + chan->reg_base + SH4A_PCIEPAMR(i)); + + pci_write_reg(chan, 0x00000000, SH4A_PCIEPARH(i)); + pci_write_reg(chan, 0x00000000, SH4A_PCIEPARL(i)); + + enable_mask = MASK_PARE; + if (res->flags & IORESOURCE_IO) + enable_mask |= MASK_SPC; + + pci_write_reg(chan, enable_mask, SH4A_PCIEPTCTLR(i)); + } return 0; } diff --git a/arch/sh/drivers/pci/pcie-sh7786.h b/arch/sh/drivers/pci/pcie-sh7786.h index 6666ea2..90a6992 100644 --- a/arch/sh/drivers/pci/pcie-sh7786.h +++ b/arch/sh/drivers/pci/pcie-sh7786.h @@ -312,23 +312,23 @@ #define SH4A_PCIECSAR5 (0x0202B4) /* R/W R/W 0x0000 0000 32 */ #define SH4A_PCIESTCTLR5 (0x0202B8) /* R/W R/W 0x0000 0000 32 */ -/* PCIEPARL0 */ -#define SH4A_PCIEPARL0 (0x020400) /* R/W R/W 0x0000 0000 32 */ +/* PCIEPARL */ +#define SH4A_PCIEPARL(x) (0x020400 + ((x) * 0x20)) /* R/W R/W 0x0000 0000 32 */ #define BITS_PAL (18) #define MASK_PAL (0x3fff< Date: Wed, 10 Feb 2010 16:06:42 +0900 Subject: sh: update sdk7786 defconfig. This plugs in USB and PCI and other bits for SDK7786. Signed-off-by: Paul Mundt diff --git a/arch/sh/configs/sdk7786_defconfig b/arch/sh/configs/sdk7786_defconfig index 1721d2c..9b331ea 100644 --- a/arch/sh/configs/sdk7786_defconfig +++ b/arch/sh/configs/sdk7786_defconfig @@ -1,7 +1,7 @@ # # Automatically generated make config: don't edit -# Linux kernel version: 2.6.33-rc3 -# Thu Jan 14 20:37:44 2010 +# Linux kernel version: 2.6.33-rc7 +# Tue Feb 9 15:27:06 2010 # CONFIG_SUPERH=y CONFIG_SUPERH32=y @@ -13,8 +13,8 @@ CONFIG_GENERIC_FIND_NEXT_BIT=y CONFIG_GENERIC_HWEIGHT=y CONFIG_GENERIC_HARDIRQS=y CONFIG_GENERIC_HARDIRQS_NO__DO_IRQ=y -CONFIG_GENERIC_IRQ_PROBE=y CONFIG_IRQ_PER_CPU=y +CONFIG_SPARSE_IRQ=y # CONFIG_GENERIC_GPIO is not set CONFIG_GENERIC_TIME=y CONFIG_GENERIC_CLOCKEVENTS=y @@ -23,6 +23,7 @@ CONFIG_ARCH_HIBERNATION_POSSIBLE=y CONFIG_SYS_SUPPORTS_HUGETLBFS=y CONFIG_SYS_SUPPORTS_SMP=y CONFIG_SYS_SUPPORTS_NUMA=y +CONFIG_SYS_SUPPORTS_PCI=y CONFIG_SYS_SUPPORTS_TMU=y CONFIG_STACKTRACE_SUPPORT=y CONFIG_LOCKDEP_SUPPORT=y @@ -134,6 +135,7 @@ CONFIG_EVENT_PROFILE=y # CONFIG_PERF_COUNTERS is not set # CONFIG_DEBUG_PERF_USE_VMALLOC is not set CONFIG_VM_EVENT_COUNTERS=y +CONFIG_PCI_QUIRKS=y # CONFIG_COMPAT_BRK is not set CONFIG_SLAB=y # CONFIG_SLUB is not set @@ -143,7 +145,6 @@ CONFIG_TRACEPOINTS=y # CONFIG_OPROFILE is not set CONFIG_HAVE_OPROFILE=y # CONFIG_KPROBES is not set -CONFIG_HAVE_IOREMAP_PROT=y CONFIG_HAVE_KPROBES=y CONFIG_HAVE_KRETPROBES=y CONFIG_HAVE_ARCH_TRACEHOOK=y @@ -267,7 +268,7 @@ CONFIG_MMU=y CONFIG_PAGE_OFFSET=0x80000000 CONFIG_FORCE_MAX_ZONEORDER=11 CONFIG_MEMORY_START=0x60000000 -CONFIG_MEMORY_SIZE=0x08000000 +CONFIG_MEMORY_SIZE=0x20000000 # CONFIG_29BIT is not set CONFIG_32BIT=y CONFIG_PMB=y @@ -284,6 +285,7 @@ CONFIG_ARCH_SELECT_MEMORY_MODEL=y CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y CONFIG_ARCH_ENABLE_MEMORY_HOTREMOVE=y CONFIG_ARCH_MEMORY_PROBE=y +CONFIG_IOREMAP_FIXED=y CONFIG_PAGE_SIZE_4KB=y # CONFIG_PAGE_SIZE_8KB is not set # CONFIG_PAGE_SIZE_16KB is not set @@ -341,9 +343,7 @@ CONFIG_SH_SDK7786=y # Timer and clock configuration # CONFIG_SH_TIMER_TMU=y -CONFIG_SH_PCLK_FREQ=50000000 CONFIG_SH_CLK_CPG=y -CONFIG_SH_CLK_CPG_LEGACY=y CONFIG_TICK_ONESHOT=y CONFIG_NO_HZ=y CONFIG_HIGH_RES_TIMERS=y @@ -401,7 +401,6 @@ CONFIG_SECCOMP=y # CONFIG_PREEMPT_VOLUNTARY is not set CONFIG_PREEMPT=y CONFIG_GUSA=y -CONFIG_SPARSE_IRQ=y # # Boot options @@ -411,13 +410,26 @@ CONFIG_BOOT_LINK_OFFSET=0x00800000 CONFIG_ENTRY_OFFSET=0x00001000 CONFIG_CMDLINE_OVERWRITE=y # CONFIG_CMDLINE_EXTEND is not set -CONFIG_CMDLINE="console=ttySC1,115200 earlyprintk=sh-sci.1,115200 root=/dev/nfs ip=dhcp" +CONFIG_CMDLINE="console=ttySC1,115200 earlyprintk=sh-sci.1,115200 root=/dev/sda1 nmi_debug=state,debounce rootdelay=10" # # Bus options # +CONFIG_PCI=y +CONFIG_PCI_DOMAINS=y +CONFIG_PCIEPORTBUS=y +CONFIG_PCIEAER=y +# CONFIG_PCIE_ECRC is not set +CONFIG_PCIEAER_INJECT=y +CONFIG_PCIEASPM=y +CONFIG_PCIEASPM_DEBUG=y # CONFIG_ARCH_SUPPORTS_MSI is not set +# CONFIG_PCI_LEGACY is not set +CONFIG_PCI_DEBUG=y +# CONFIG_PCI_STUB is not set +# CONFIG_PCI_IOV is not set # CONFIG_PCCARD is not set +# CONFIG_HOTPLUG_PCI is not set # # Executable file formats @@ -430,8 +442,14 @@ CONFIG_BINFMT_MISC=y # # Power management options (EXPERIMENTAL) # -# CONFIG_PM is not set -# CONFIG_CPU_IDLE is not set +CONFIG_PM=y +CONFIG_PM_DEBUG=y +CONFIG_PM_VERBOSE=y +# CONFIG_HIBERNATION is not set +CONFIG_PM_RUNTIME=y +CONFIG_CPU_IDLE=y +CONFIG_CPU_IDLE_GOV_LADDER=y +CONFIG_CPU_IDLE_GOV_MENU=y CONFIG_NET=y # @@ -538,6 +556,9 @@ CONFIG_PREVENT_FIRMWARE_BUILD=y # CONFIG_MTD is not set # CONFIG_PARPORT is not set CONFIG_BLK_DEV=y +# CONFIG_BLK_CPQ_CISS_DA is not set +# CONFIG_BLK_DEV_DAC960 is not set +# CONFIG_BLK_DEV_UMEM is not set # CONFIG_BLK_DEV_COW_COMMON is not set # CONFIG_BLK_DEV_LOOP is not set @@ -545,6 +566,7 @@ CONFIG_BLK_DEV=y # DRBD disabled because PROC_FS, INET or CONNECTOR not selected # # CONFIG_BLK_DEV_NBD is not set +# CONFIG_BLK_DEV_SX8 is not set # CONFIG_BLK_DEV_UB is not set CONFIG_BLK_DEV_RAM=y CONFIG_BLK_DEV_RAM_COUNT=16 @@ -555,8 +577,12 @@ CONFIG_BLK_DEV_RAM_SIZE=4096 # CONFIG_BLK_DEV_HD is not set CONFIG_MISC_DEVICES=y # CONFIG_AD525X_DPOT is not set +# CONFIG_PHANTOM is not set +# CONFIG_SGI_IOC4 is not set +# CONFIG_TIFM_CORE is not set # CONFIG_ICS932S401 is not set # CONFIG_ENCLOSURE_SERVICES is not set +# CONFIG_HP_ILO is not set # CONFIG_ISL29003 is not set # CONFIG_DS1682 is not set # CONFIG_TI_DAC7512 is not set @@ -570,6 +596,7 @@ CONFIG_MISC_DEVICES=y # CONFIG_EEPROM_LEGACY is not set # CONFIG_EEPROM_MAX6875 is not set # CONFIG_EEPROM_93CX6 is not set +# CONFIG_CB710_CORE is not set CONFIG_HAVE_IDE=y # CONFIG_IDE is not set @@ -608,19 +635,131 @@ CONFIG_SCSI_WAIT_SCAN=m # CONFIG_SCSI_SRP_ATTRS is not set CONFIG_SCSI_LOWLEVEL=y # CONFIG_ISCSI_TCP is not set +# CONFIG_SCSI_BNX2_ISCSI is not set +# CONFIG_BE2ISCSI is not set +# CONFIG_BLK_DEV_3W_XXXX_RAID is not set +# CONFIG_SCSI_HPSA is not set +# CONFIG_SCSI_3W_9XXX is not set +# CONFIG_SCSI_3W_SAS is not set +# CONFIG_SCSI_ACARD is not set +# CONFIG_SCSI_AACRAID is not set +# CONFIG_SCSI_AIC7XXX is not set +# CONFIG_SCSI_AIC7XXX_OLD is not set +# CONFIG_SCSI_AIC79XX is not set +# CONFIG_SCSI_AIC94XX is not set +# CONFIG_SCSI_MVSAS is not set +# CONFIG_SCSI_ARCMSR is not set +# CONFIG_MEGARAID_NEWGEN is not set +# CONFIG_MEGARAID_LEGACY is not set +# CONFIG_MEGARAID_SAS is not set +# CONFIG_SCSI_MPT2SAS is not set +# CONFIG_SCSI_HPTIOP is not set # CONFIG_LIBFC is not set # CONFIG_LIBFCOE is not set +# CONFIG_FCOE is not set +# CONFIG_SCSI_DMX3191D is not set +# CONFIG_SCSI_FUTURE_DOMAIN is not set +# CONFIG_SCSI_IPS is not set +# CONFIG_SCSI_INITIO is not set +# CONFIG_SCSI_INIA100 is not set +# CONFIG_SCSI_STEX is not set +# CONFIG_SCSI_SYM53C8XX_2 is not set +# CONFIG_SCSI_IPR is not set +# CONFIG_SCSI_QLOGIC_1280 is not set +# CONFIG_SCSI_QLA_FC is not set +# CONFIG_SCSI_QLA_ISCSI is not set +# CONFIG_SCSI_LPFC is not set +# CONFIG_SCSI_DC395x is not set +# CONFIG_SCSI_DC390T is not set +# CONFIG_SCSI_NSP32 is not set # CONFIG_SCSI_DEBUG is not set +# CONFIG_SCSI_PMCRAID is not set +# CONFIG_SCSI_PM8001 is not set +# CONFIG_SCSI_SRP is not set +# CONFIG_SCSI_BFA_FC is not set # CONFIG_SCSI_DH is not set # CONFIG_SCSI_OSD_INITIATOR is not set CONFIG_ATA=y # CONFIG_ATA_NONSTANDARD is not set CONFIG_ATA_VERBOSE_ERROR=y CONFIG_SATA_PMP=y +# CONFIG_SATA_AHCI is not set +CONFIG_SATA_SIL24=y CONFIG_ATA_SFF=y +# CONFIG_SATA_SVW is not set +# CONFIG_ATA_PIIX is not set # CONFIG_SATA_MV is not set +# CONFIG_SATA_NV is not set +# CONFIG_PDC_ADMA is not set +# CONFIG_SATA_QSTOR is not set +# CONFIG_SATA_PROMISE is not set +# CONFIG_SATA_SX4 is not set +# CONFIG_SATA_SIL is not set +# CONFIG_SATA_SIS is not set +# CONFIG_SATA_ULI is not set +# CONFIG_SATA_VIA is not set +# CONFIG_SATA_VITESSE is not set +# CONFIG_SATA_INIC162X is not set +# CONFIG_PATA_ALI is not set +# CONFIG_PATA_AMD is not set +# CONFIG_PATA_ARTOP is not set +# CONFIG_PATA_ATP867X is not set +# CONFIG_PATA_ATIIXP is not set +# CONFIG_PATA_CMD640_PCI is not set +# CONFIG_PATA_CMD64X is not set +# CONFIG_PATA_CS5520 is not set +# CONFIG_PATA_CS5530 is not set +# CONFIG_PATA_CYPRESS is not set +# CONFIG_PATA_EFAR is not set +# CONFIG_ATA_GENERIC is not set +# CONFIG_PATA_HPT366 is not set +# CONFIG_PATA_HPT37X is not set +# CONFIG_PATA_HPT3X2N is not set +# CONFIG_PATA_HPT3X3 is not set +# CONFIG_PATA_IT821X is not set +# CONFIG_PATA_IT8213 is not set +# CONFIG_PATA_JMICRON is not set +# CONFIG_PATA_TRIFLEX is not set +# CONFIG_PATA_MARVELL is not set +# CONFIG_PATA_MPIIX is not set +# CONFIG_PATA_OLDPIIX is not set +# CONFIG_PATA_NETCELL is not set +# CONFIG_PATA_NINJA32 is not set +# CONFIG_PATA_NS87410 is not set +# CONFIG_PATA_NS87415 is not set +# CONFIG_PATA_OPTI is not set +# CONFIG_PATA_OPTIDMA is not set +# CONFIG_PATA_PDC2027X is not set +# CONFIG_PATA_PDC_OLD is not set +# CONFIG_PATA_RADISYS is not set +# CONFIG_PATA_RDC is not set +# CONFIG_PATA_RZ1000 is not set +# CONFIG_PATA_SC1200 is not set +# CONFIG_PATA_SERVERWORKS is not set +# CONFIG_PATA_SIL680 is not set +# CONFIG_PATA_SIS is not set +# CONFIG_PATA_TOSHIBA is not set +# CONFIG_PATA_VIA is not set +# CONFIG_PATA_WINBOND is not set CONFIG_PATA_PLATFORM=y +# CONFIG_PATA_SCH is not set # CONFIG_MD is not set +# CONFIG_FUSION is not set + +# +# IEEE 1394 (FireWire) support +# + +# +# You can enable one or both FireWire driver stacks. +# + +# +# The newer stack is recommended. +# +# CONFIG_FIREWIRE is not set +# CONFIG_IEEE1394 is not set +# CONFIG_I2O is not set CONFIG_NETDEVICES=y # CONFIG_DUMMY is not set # CONFIG_BONDING is not set @@ -628,6 +767,7 @@ CONFIG_NETDEVICES=y # CONFIG_EQUALIZER is not set # CONFIG_TUN is not set # CONFIG_VETH is not set +# CONFIG_ARCNET is not set CONFIG_PHYLIB=y # @@ -652,12 +792,18 @@ CONFIG_NET_ETHERNET=y CONFIG_MII=y # CONFIG_AX88796 is not set # CONFIG_STNIC is not set +# CONFIG_HAPPYMEAL is not set +# CONFIG_SUNGEM is not set +# CONFIG_CASSINI is not set +# CONFIG_NET_VENDOR_3COM is not set CONFIG_SMC91X=y # CONFIG_ENC28J60 is not set # CONFIG_ETHOC is not set # CONFIG_SMC911X is not set CONFIG_SMSC911X=y # CONFIG_DNET is not set +# CONFIG_NET_TULIP is not set +# CONFIG_HP100 is not set # CONFIG_IBM_NEW_EMAC_ZMII is not set # CONFIG_IBM_NEW_EMAC_RGMII is not set # CONFIG_IBM_NEW_EMAC_TAH is not set @@ -665,13 +811,18 @@ CONFIG_SMSC911X=y # CONFIG_IBM_NEW_EMAC_NO_FLOW_CTRL is not set # CONFIG_IBM_NEW_EMAC_MAL_CLR_ICINTSTAT is not set # CONFIG_IBM_NEW_EMAC_MAL_COMMON_ERR is not set +# CONFIG_NET_PCI is not set # CONFIG_B44 is not set # CONFIG_KS8842 is not set # CONFIG_KS8851 is not set # CONFIG_KS8851_MLL is not set +# CONFIG_ATL2 is not set # CONFIG_NETDEV_1000 is not set # CONFIG_NETDEV_10000 is not set +# CONFIG_TR is not set CONFIG_WLAN=y +# CONFIG_ATMEL is not set +# CONFIG_PRISM54 is not set # CONFIG_USB_ZD1201 is not set # CONFIG_HOSTAP is not set @@ -688,31 +839,94 @@ CONFIG_WLAN=y # CONFIG_USB_RTL8150 is not set # CONFIG_USB_USBNET is not set # CONFIG_WAN is not set +# CONFIG_FDDI is not set +# CONFIG_HIPPI is not set # CONFIG_PPP is not set # CONFIG_SLIP is not set +# CONFIG_NET_FC is not set # CONFIG_NETCONSOLE is not set # CONFIG_NETPOLL is not set # CONFIG_NET_POLL_CONTROLLER is not set +# CONFIG_VMXNET3 is not set # CONFIG_ISDN is not set # CONFIG_PHONE is not set # # Input device support # -# CONFIG_INPUT is not set +CONFIG_INPUT=y +# CONFIG_INPUT_FF_MEMLESS is not set +# CONFIG_INPUT_POLLDEV is not set +# CONFIG_INPUT_SPARSEKMAP is not set + +# +# Userland interfaces +# +CONFIG_INPUT_MOUSEDEV=y +CONFIG_INPUT_MOUSEDEV_PSAUX=y +CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024 +CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768 +# CONFIG_INPUT_JOYDEV is not set +# CONFIG_INPUT_EVDEV is not set +# CONFIG_INPUT_EVBUG is not set + +# +# Input Device Drivers +# +CONFIG_INPUT_KEYBOARD=y +# CONFIG_KEYBOARD_ADP5588 is not set +CONFIG_KEYBOARD_ATKBD=y +# CONFIG_QT2160 is not set +# CONFIG_KEYBOARD_LKKBD is not set +# CONFIG_KEYBOARD_MAX7359 is not set +# CONFIG_KEYBOARD_NEWTON is not set +# CONFIG_KEYBOARD_OPENCORES is not set +# CONFIG_KEYBOARD_STOWAWAY is not set +# CONFIG_KEYBOARD_SUNKBD is not set +# CONFIG_KEYBOARD_SH_KEYSC is not set +# CONFIG_KEYBOARD_XTKBD is not set +CONFIG_INPUT_MOUSE=y +CONFIG_MOUSE_PS2=y +CONFIG_MOUSE_PS2_ALPS=y +CONFIG_MOUSE_PS2_LOGIPS2PP=y +CONFIG_MOUSE_PS2_SYNAPTICS=y +CONFIG_MOUSE_PS2_TRACKPOINT=y +# CONFIG_MOUSE_PS2_ELANTECH is not set +# CONFIG_MOUSE_PS2_SENTELIC is not set +# CONFIG_MOUSE_PS2_TOUCHKIT is not set +# CONFIG_MOUSE_SERIAL is not set +# CONFIG_MOUSE_APPLETOUCH is not set +# CONFIG_MOUSE_BCM5974 is not set +# CONFIG_MOUSE_VSXXXAA is not set +# CONFIG_MOUSE_SYNAPTICS_I2C is not set +# CONFIG_INPUT_JOYSTICK is not set +# CONFIG_INPUT_TABLET is not set +# CONFIG_INPUT_TOUCHSCREEN is not set +# CONFIG_INPUT_MISC is not set # # Hardware I/O ports # -# CONFIG_SERIO is not set +CONFIG_SERIO=y +CONFIG_SERIO_I8042=y +CONFIG_SERIO_SERPORT=y +# CONFIG_SERIO_PCIPS2 is not set +CONFIG_SERIO_LIBPS2=y +# CONFIG_SERIO_RAW is not set +# CONFIG_SERIO_ALTERA_PS2 is not set # CONFIG_GAMEPORT is not set # # Character devices # -# CONFIG_VT is not set +CONFIG_VT=y +CONFIG_CONSOLE_TRANSLATIONS=y +CONFIG_VT_CONSOLE=y +CONFIG_HW_CONSOLE=y +# CONFIG_VT_HW_CONSOLE_BINDING is not set CONFIG_DEVKMEM=y # CONFIG_SERIAL_NONSTANDARD is not set +# CONFIG_NOZOMI is not set # # Serial drivers @@ -728,14 +942,17 @@ CONFIG_SERIAL_SH_SCI_NR_UARTS=6 CONFIG_SERIAL_SH_SCI_CONSOLE=y CONFIG_SERIAL_CORE=y CONFIG_SERIAL_CORE_CONSOLE=y +# CONFIG_SERIAL_JSM is not set CONFIG_UNIX98_PTYS=y # CONFIG_DEVPTS_MULTIPLE_INSTANCES is not set # CONFIG_LEGACY_PTYS is not set # CONFIG_IPMI_HANDLER is not set # CONFIG_HW_RANDOM is not set # CONFIG_R3964 is not set +# CONFIG_APPLICOM is not set # CONFIG_RAW_DRIVER is not set # CONFIG_TCG_TPM is not set +CONFIG_DEVPORT=y CONFIG_I2C=y CONFIG_I2C_BOARDINFO=y # CONFIG_I2C_COMPAT is not set @@ -747,11 +964,28 @@ CONFIG_I2C_HELPER_AUTO=y # # +# PC SMBus host controller drivers +# +# CONFIG_I2C_ALI1535 is not set +# CONFIG_I2C_ALI1563 is not set +# CONFIG_I2C_ALI15X3 is not set +# CONFIG_I2C_AMD756 is not set +# CONFIG_I2C_AMD8111 is not set +# CONFIG_I2C_I801 is not set +# CONFIG_I2C_ISCH is not set +# CONFIG_I2C_PIIX4 is not set +# CONFIG_I2C_NFORCE2 is not set +# CONFIG_I2C_SIS5595 is not set +# CONFIG_I2C_SIS630 is not set +# CONFIG_I2C_SIS96X is not set +# CONFIG_I2C_VIA is not set +# CONFIG_I2C_VIAPRO is not set + +# # I2C system bus drivers (mostly embedded / system-on-chip) # # CONFIG_I2C_DESIGNWARE is not set # CONFIG_I2C_OCORES is not set -CONFIG_I2C_SDK7786=y # CONFIG_I2C_SH_MOBILE is not set # CONFIG_I2C_SIMTEC is not set @@ -772,10 +1006,10 @@ CONFIG_I2C_SDK7786=y # Miscellaneous I2C Chip support # # CONFIG_SENSORS_TSL2550 is not set -CONFIG_I2C_DEBUG_CORE=y -CONFIG_I2C_DEBUG_ALGO=y -CONFIG_I2C_DEBUG_BUS=y -CONFIG_I2C_DEBUG_CHIP=y +# CONFIG_I2C_DEBUG_CORE is not set +# CONFIG_I2C_DEBUG_ALGO is not set +# CONFIG_I2C_DEBUG_BUS is not set +# CONFIG_I2C_DEBUG_CHIP is not set CONFIG_SPI=y # CONFIG_SPI_DEBUG is not set CONFIG_SPI_MASTER=y @@ -810,9 +1044,16 @@ CONFIG_WATCHDOG=y # Watchdog Device Drivers # # CONFIG_SOFT_WATCHDOG is not set +# CONFIG_ALIM7101_WDT is not set # CONFIG_SH_WDT is not set # +# PCI-based Watchdog Cards +# +# CONFIG_PCIPCWATCHDOG is not set +# CONFIG_WDTPCI is not set + +# # USB-based Watchdog Cards # # CONFIG_USBPCWATCHDOG is not set @@ -849,6 +1090,8 @@ CONFIG_SSB_POSSIBLE=y # # Graphics support # +CONFIG_VGA_ARB=y +# CONFIG_DRM is not set # CONFIG_VGASTATE is not set CONFIG_VIDEO_OUTPUT_CONTROL=m # CONFIG_FB is not set @@ -858,11 +1101,56 @@ CONFIG_VIDEO_OUTPUT_CONTROL=m # Display device support # # CONFIG_DISPLAY_SUPPORT is not set + +# +# Console display driver support +# +CONFIG_DUMMY_CONSOLE=y # CONFIG_SOUND is not set +CONFIG_HID_SUPPORT=y +CONFIG_HID=y +# CONFIG_HIDRAW is not set + +# +# USB Input Devices +# +CONFIG_USB_HID=y +# CONFIG_HID_PID is not set +# CONFIG_USB_HIDDEV is not set + +# +# Special HID drivers +# +# CONFIG_HID_A4TECH is not set +# CONFIG_HID_APPLE is not set +# CONFIG_HID_BELKIN is not set +# CONFIG_HID_CHERRY is not set +# CONFIG_HID_CHICONY is not set +# CONFIG_HID_CYPRESS is not set +# CONFIG_HID_DRAGONRISE is not set +# CONFIG_HID_EZKEY is not set +# CONFIG_HID_KYE is not set +# CONFIG_HID_GYRATION is not set +# CONFIG_HID_TWINHAN is not set +# CONFIG_HID_KENSINGTON is not set +# CONFIG_HID_LOGITECH is not set +# CONFIG_HID_MICROSOFT is not set +# CONFIG_HID_MONTEREY is not set +# CONFIG_HID_NTRIG is not set +# CONFIG_HID_PANTHERLORD is not set +# CONFIG_HID_PETALYNX is not set +# CONFIG_HID_SAMSUNG is not set +# CONFIG_HID_SONY is not set +# CONFIG_HID_SUNPLUS is not set +# CONFIG_HID_GREENASIA is not set +# CONFIG_HID_SMARTJOYPLUS is not set +# CONFIG_HID_TOPSEED is not set +# CONFIG_HID_THRUSTMASTER is not set +# CONFIG_HID_ZEROPLUS is not set CONFIG_USB_SUPPORT=y CONFIG_USB_ARCH_HAS_HCD=y CONFIG_USB_ARCH_HAS_OHCI=y -# CONFIG_USB_ARCH_HAS_EHCI is not set +CONFIG_USB_ARCH_HAS_EHCI=y CONFIG_USB=y # CONFIG_USB_DEBUG is not set # CONFIG_USB_ANNOUNCE_NEW_DEVICES is not set @@ -873,6 +1161,7 @@ CONFIG_USB=y # CONFIG_USB_DEVICEFS is not set CONFIG_USB_DEVICE_CLASS=y # CONFIG_USB_DYNAMIC_MINORS is not set +# CONFIG_USB_SUSPEND is not set # CONFIG_USB_OTG is not set # CONFIG_USB_OTG_WHITELIST is not set # CONFIG_USB_OTG_BLACKLIST_HUB is not set @@ -884,13 +1173,20 @@ CONFIG_USB_MON=y # USB Host Controller Drivers # # CONFIG_USB_C67X00_HCD is not set +# CONFIG_USB_XHCI_HCD is not set +# CONFIG_USB_EHCI_HCD is not set # CONFIG_USB_OXU210HP_HCD is not set # CONFIG_USB_ISP116X_HCD is not set # CONFIG_USB_ISP1760_HCD is not set # CONFIG_USB_ISP1362_HCD is not set -# CONFIG_USB_OHCI_HCD is not set +CONFIG_USB_OHCI_HCD=y +# CONFIG_USB_OHCI_BIG_ENDIAN_DESC is not set +# CONFIG_USB_OHCI_BIG_ENDIAN_MMIO is not set +CONFIG_USB_OHCI_LITTLE_ENDIAN=y +# CONFIG_USB_UHCI_HCD is not set # CONFIG_USB_SL811_HCD is not set -CONFIG_USB_R8A66597_HCD=m +# CONFIG_USB_R8A66597_HCD is not set +# CONFIG_USB_WHCI_HCD is not set # CONFIG_USB_HWA_HCD is not set # CONFIG_USB_GADGET_MUSB_HDRC is not set @@ -909,7 +1205,19 @@ CONFIG_USB_R8A66597_HCD=m # # also be needed; see USB_STORAGE Help for more info # -# CONFIG_USB_STORAGE is not set +CONFIG_USB_STORAGE=y +# CONFIG_USB_STORAGE_DEBUG is not set +# CONFIG_USB_STORAGE_DATAFAB is not set +# CONFIG_USB_STORAGE_FREECOM is not set +# CONFIG_USB_STORAGE_ISD200 is not set +# CONFIG_USB_STORAGE_USBAT is not set +# CONFIG_USB_STORAGE_SDDR09 is not set +# CONFIG_USB_STORAGE_SDDR55 is not set +# CONFIG_USB_STORAGE_JUMPSHOT is not set +# CONFIG_USB_STORAGE_ALAUDA is not set +# CONFIG_USB_STORAGE_ONETOUCH is not set +# CONFIG_USB_STORAGE_KARMA is not set +# CONFIG_USB_STORAGE_CYPRESS_ATACB is not set # CONFIG_USB_LIBUSUAL is not set # @@ -989,10 +1297,12 @@ CONFIG_USB_GADGET_DUALSPEED=y # OTG and related infrastructure # # CONFIG_NOP_USB_XCEIV is not set +# CONFIG_UWB is not set # CONFIG_MMC is not set # CONFIG_MEMSTICK is not set # CONFIG_NEW_LEDS is not set # CONFIG_ACCESSIBILITY is not set +# CONFIG_INFINIBAND is not set CONFIG_RTC_LIB=y CONFIG_RTC_CLASS=y CONFIG_RTC_HCTOSYS=y @@ -1063,10 +1373,13 @@ CONFIG_RTC_DRV_SH=y # CONFIG_DMADEVICES is not set # CONFIG_AUXDISPLAY is not set CONFIG_UIO=m +# CONFIG_UIO_CIF is not set # CONFIG_UIO_PDRV is not set # CONFIG_UIO_PDRV_GENIRQ is not set # CONFIG_UIO_SMX is not set +# CONFIG_UIO_AEC is not set # CONFIG_UIO_SERCOS3 is not set +# CONFIG_UIO_PCI_GENERIC is not set # # TI VLYNQ @@ -1418,6 +1731,7 @@ CONFIG_CRYPTO=y # # CONFIG_CRYPTO_ANSI_CPRNG is not set CONFIG_CRYPTO_HW=y +# CONFIG_CRYPTO_DEV_HIFN_795X is not set CONFIG_BINARY_PRINTF=y # -- cgit v0.10.2 From b32758c7216f337044ceb6dcaa754b8eda95a59f Mon Sep 17 00:00:00 2001 From: Stephane Chatty Date: Wed, 10 Feb 2010 12:09:17 +0100 Subject: HID: fixed bug in single-touch emulation on the stantum panel Fixed stupid copy-paste bug in touchscreen emulation for the Stantum multitouch panel: a flag was reset just before being tested. Signed-off-by: Stephane Chatty Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-stantum.c b/drivers/hid/hid-stantum.c index add965d..1b7b165 100644 --- a/drivers/hid/hid-stantum.c +++ b/drivers/hid/hid-stantum.c @@ -147,7 +147,6 @@ static void stantum_filter_event(struct stantum_data *sd, input_mt_sync(input); sd->valid = false; - sd->first = false; /* touchscreen emulation */ if (sd->first) { @@ -158,6 +157,7 @@ static void stantum_filter_event(struct stantum_data *sd, input_event(input, EV_ABS, ABS_X, sd->x); input_event(input, EV_ABS, ABS_Y, sd->y); } + sd->first = false; } -- cgit v0.10.2 From 580363db92572cccbe6226bf83321e50a9ea50ea Mon Sep 17 00:00:00 2001 From: Stephane Chatty Date: Sat, 6 Feb 2010 15:20:03 +0100 Subject: HID: add pressure support for the Stantum multitouch panel Added pressure handling for Stantum multitouch panels Signed-off-by: Stephane Chatty Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-stantum.c b/drivers/hid/hid-stantum.c index 1b7b165..04b0b98 100644 --- a/drivers/hid/hid-stantum.c +++ b/drivers/hid/hid-stantum.c @@ -66,7 +66,6 @@ static int stantum_input_mapping(struct hid_device *hdev, struct hid_input *hi, case HID_DG_DEVICEINDEX: case HID_DG_CONTACTCOUNT: case HID_DG_CONTACTMAX: - case HID_DG_TIPPRESSURE: return -1; case HID_DG_TIPSWITCH: @@ -84,6 +83,11 @@ static int stantum_input_mapping(struct hid_device *hdev, struct hid_input *hi, input_set_abs_params(hi->input, ABS_MT_ORIENTATION, 1, 1, 0, 0); return 1; + case HID_DG_TIPPRESSURE: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_PRESSURE); + return 1; + case HID_DG_CONTACTID: hid_map_usage(hi, usage, bit, max, EV_ABS, ABS_MT_TRACKING_ID); @@ -140,10 +144,7 @@ static void stantum_filter_event(struct stantum_data *sd, input_event(input, EV_ABS, ABS_MT_TOUCH_MAJOR, wide ? sd->w : sd->h); input_event(input, EV_ABS, ABS_MT_TOUCH_MINOR, wide ? sd->h : sd->w); -#if 0 - /* MT_PRESSURE does not exist yet */ input_event(input, EV_ABS, ABS_MT_PRESSURE, sd->z); -#endif input_mt_sync(input); sd->valid = false; -- cgit v0.10.2 From 28ef35eb33d7e291bf0e1fe1137915a7e38805ff Mon Sep 17 00:00:00 2001 From: Stefan Roese Date: Tue, 9 Feb 2010 23:08:17 +0000 Subject: powerpc/44x: Fix L2-cache support for 460GT Also set L2C_CFG_RDBW on 460GT platforms and not only on 460EX. Signed-off-by: Stefan Roese Cc: Josh Boyer Signed-off-by: Josh Boyer diff --git a/arch/powerpc/sysdev/ppc4xx_soc.c b/arch/powerpc/sysdev/ppc4xx_soc.c index 5b32adc..5c01435 100644 --- a/arch/powerpc/sysdev/ppc4xx_soc.c +++ b/arch/powerpc/sysdev/ppc4xx_soc.c @@ -174,7 +174,8 @@ static int __init ppc4xx_l2c_probe(void) | L2C_CFG_CPIM | L2C_CFG_TPIM | L2C_CFG_LIM | L2C_CFG_SMCM; /* Check for 460EX/GT special handling */ - if (of_device_is_compatible(np, "ibm,l2-cache-460ex")) + if (of_device_is_compatible(np, "ibm,l2-cache-460ex") || + of_device_is_compatible(np, "ibm,l2-cache-460gt")) r |= L2C_CFG_RDBW; mtdcr(dcrbase_l2c + DCRN_L2C0_CFG, r); -- cgit v0.10.2 From 036f290d8952fdcafa9d6c9e2d94efbcb793d123 Mon Sep 17 00:00:00 2001 From: Stefan Roese Date: Tue, 26 Jan 2010 05:56:30 +0000 Subject: powerpc/44x: Add MTD support (NOR FLASH) to Katmai dts This patch adds NOR FLASH MTD support to the Katmai (440SPe) dts file. For this the OPB ranges address is mapped differently (base 0x00000000 -> 0xe0000000). This results in the address being identical to the lower 32bit of its physical address. This is needed for the MTD mapping to work correctly, since U-Boot will insert the physical addresses of the EBC chip selects into the EBC ranges property. This is the way its done in most other 4xx dts files as well. Additionally with a small whitespace cleanup. Signed-off-by: Stefan Roese Cc: Josh Boyer Signed-off-by: Josh Boyer diff --git a/arch/powerpc/boot/dts/katmai.dts b/arch/powerpc/boot/dts/katmai.dts index 8f345de9..8cf2c0c 100644 --- a/arch/powerpc/boot/dts/katmai.dts +++ b/arch/powerpc/boot/dts/katmai.dts @@ -156,7 +156,7 @@ compatible = "ibm,opb-440spe", "ibm,opb-440gp", "ibm,opb"; #address-cells = <1>; #size-cells = <1>; - ranges = <0x00000000 0x00000004 0xe0000000 0x20000000>; + ranges = <0xe0000000 0x00000004 0xe0000000 0x20000000>; clock-frequency = <0>; /* Filled in by zImage */ EBC0: ebc { @@ -165,14 +165,47 @@ #address-cells = <2>; #size-cells = <1>; clock-frequency = <0>; /* Filled in by zImage */ + /* ranges property is supplied by U-Boot */ interrupts = <0x5 0x1>; interrupt-parent = <&UIC1>; + + nor_flash@0,0 { + compatible = "cfi-flash"; + bank-width = <2>; + reg = <0x00000000 0x00000000 0x01000000>; + #address-cells = <1>; + #size-cells = <1>; + partition@0 { + label = "kernel"; + reg = <0x00000000 0x001e0000>; + }; + partition@1e0000 { + label = "dtb"; + reg = <0x001e0000 0x00020000>; + }; + partition@200000 { + label = "root"; + reg = <0x00200000 0x00200000>; + }; + partition@400000 { + label = "user"; + reg = <0x00400000 0x00b60000>; + }; + partition@f60000 { + label = "env"; + reg = <0x00f60000 0x00040000>; + }; + partition@fa0000 { + label = "u-boot"; + reg = <0x00fa0000 0x00060000>; + }; + }; }; - UART0: serial@10000200 { + UART0: serial@f0000200 { device_type = "serial"; compatible = "ns16550"; - reg = <0x10000200 0x00000008>; + reg = <0xf0000200 0x00000008>; virtual-reg = <0xa0000200>; clock-frequency = <0>; /* Filled in by zImage */ current-speed = <115200>; @@ -180,10 +213,10 @@ interrupts = <0x0 0x4>; }; - UART1: serial@10000300 { + UART1: serial@f0000300 { device_type = "serial"; compatible = "ns16550"; - reg = <0x10000300 0x00000008>; + reg = <0xf0000300 0x00000008>; virtual-reg = <0xa0000300>; clock-frequency = <0>; current-speed = <0>; @@ -192,10 +225,10 @@ }; - UART2: serial@10000600 { + UART2: serial@f0000600 { device_type = "serial"; compatible = "ns16550"; - reg = <0x10000600 0x00000008>; + reg = <0xf0000600 0x00000008>; virtual-reg = <0xa0000600>; clock-frequency = <0>; current-speed = <0>; @@ -203,27 +236,27 @@ interrupts = <0x5 0x4>; }; - IIC0: i2c@10000400 { + IIC0: i2c@f0000400 { compatible = "ibm,iic-440spe", "ibm,iic-440gp", "ibm,iic"; - reg = <0x10000400 0x00000014>; + reg = <0xf0000400 0x00000014>; interrupt-parent = <&UIC0>; interrupts = <0x2 0x4>; }; - IIC1: i2c@10000500 { + IIC1: i2c@f0000500 { compatible = "ibm,iic-440spe", "ibm,iic-440gp", "ibm,iic"; - reg = <0x10000500 0x00000014>; + reg = <0xf0000500 0x00000014>; interrupt-parent = <&UIC0>; interrupts = <0x3 0x4>; }; - EMAC0: ethernet@10000800 { + EMAC0: ethernet@f0000800 { linux,network-index = <0x0>; device_type = "network"; compatible = "ibm,emac-440spe", "ibm,emac4"; interrupt-parent = <&UIC1>; interrupts = <0x1c 0x4 0x1d 0x4>; - reg = <0x10000800 0x00000074>; + reg = <0xf0000800 0x00000074>; local-mac-address = [000000000000]; mal-device = <&MAL0>; mal-tx-channel = <0>; @@ -248,11 +281,11 @@ primary; large-inbound-windows; enable-msi-hole; - reg = <0x0000000c 0x0ec00000 0x00000008 /* Config space access */ - 0x00000000 0x00000000 0x00000000 /* no IACK cycles */ - 0x0000000c 0x0ed00000 0x00000004 /* Special cycles */ - 0x0000000c 0x0ec80000 0x00000100 /* Internal registers */ - 0x0000000c 0x0ec80100 0x000000fc>; /* Internal messaging registers */ + reg = <0x0000000c 0x0ec00000 0x00000008 /* Config space access */ + 0x00000000 0x00000000 0x00000000 /* no IACK cycles */ + 0x0000000c 0x0ed00000 0x00000004 /* Special cycles */ + 0x0000000c 0x0ec80000 0x00000100 /* Internal registers */ + 0x0000000c 0x0ec80100 0x000000fc>; /* Internal messaging registers */ /* Outbound ranges, one memory and one IO, * later cannot be changed @@ -453,6 +486,6 @@ }; chosen { - linux,stdout-path = "/plb/opb/serial@10000200"; + linux,stdout-path = "/plb/opb/serial@f0000200"; }; }; -- cgit v0.10.2 From 6f57518cfa097b170a8f500d6d70651e81589e30 Mon Sep 17 00:00:00 2001 From: Stefan Roese Date: Tue, 9 Feb 2010 23:08:34 +0000 Subject: powerpc/44x: Update Arches dts Sync Arches dts with latest Canyonlands version: - Add 16k FIFO size to supported EMAC nodes - Add next-level-cache property - Add Crypto device node Signed-off-by: Stefan Roese Cc: Josh Boyer Signed-off-by: Josh Boyer diff --git a/arch/powerpc/boot/dts/arches.dts b/arch/powerpc/boot/dts/arches.dts index 414ef8b..30f4120 100644 --- a/arch/powerpc/boot/dts/arches.dts +++ b/arch/powerpc/boot/dts/arches.dts @@ -60,6 +60,7 @@ d-cache-size = <32768>; dcr-controller; dcr-access-method = "native"; + next-level-cache = <&L2C0>; }; }; @@ -146,6 +147,13 @@ dcr-reg = <0x010 0x002>; }; + CRYPTO: crypto@180000 { + compatible = "amcc,ppc460gt-crypto", "amcc,ppc4xx-crypto"; + reg = <4 0x00180000 0x80400>; + interrupt-parent = <&UIC0>; + interrupts = <0x1d 0x4>; + }; + MAL0: mcmal { compatible = "ibm,mcmal-460gt", "ibm,mcmal2"; dcr-reg = <0x180 0x062>; @@ -274,6 +282,7 @@ max-frame-size = <9000>; rx-fifo-size = <4096>; tx-fifo-size = <2048>; + rx-fifo-size-gige = <16384>; phy-mode = "sgmii"; phy-map = <0xffffffff>; gpcs-address = <0x0000000a>; @@ -302,6 +311,7 @@ max-frame-size = <9000>; rx-fifo-size = <4096>; tx-fifo-size = <2048>; + rx-fifo-size-gige = <16384>; phy-mode = "sgmii"; phy-map = <0x00000000>; gpcs-address = <0x0000000b>; @@ -331,6 +341,8 @@ max-frame-size = <9000>; rx-fifo-size = <4096>; tx-fifo-size = <2048>; + rx-fifo-size-gige = <16384>; + tx-fifo-size-gige = <16384>; /* emac2&3 only */ phy-mode = "sgmii"; phy-map = <0x00000001>; gpcs-address = <0x0000000C>; -- cgit v0.10.2 From 5a6543e8da6e338fef3cd42cc79ac85f6f53a8e7 Mon Sep 17 00:00:00 2001 From: Stefan Roese Date: Tue, 9 Feb 2010 23:08:28 +0000 Subject: powerpc/44x: Update Glacier dts Sync Glacier dts with latest Canyonlands version: - Add l2 cache support - Add NDFC support - Add RTC support - Add AD7414 hwmon support - Change EMAC compatible node from emac4 to emac4sync and correct the register size - Add support for ISA holes on 4xx PCI/X/E (as done in Benjamin Herrenschmidt's patch for Canyonlands) - Add Crypto device node Signed-off-by: Stefan Roese Cc: Josh Boyer Signed-off-by: Josh Boyer diff --git a/arch/powerpc/boot/dts/glacier.dts b/arch/powerpc/boot/dts/glacier.dts index f6f6189..d62a4fb 100644 --- a/arch/powerpc/boot/dts/glacier.dts +++ b/arch/powerpc/boot/dts/glacier.dts @@ -1,7 +1,7 @@ /* * Device Tree Source for AMCC Glacier (460GT) * - * Copyright 2008 DENX Software Engineering, Stefan Roese + * Copyright 2008-2010 DENX Software Engineering, Stefan Roese * * This file is licensed under the terms of the GNU General Public * License version 2. This program is licensed "as is" without @@ -42,6 +42,7 @@ d-cache-size = <32768>; dcr-controller; dcr-access-method = "native"; + next-level-cache = <&L2C0>; }; }; @@ -106,6 +107,16 @@ dcr-reg = <0x00c 0x002>; }; + L2C0: l2c { + compatible = "ibm,l2-cache-460gt", "ibm,l2-cache"; + dcr-reg = <0x020 0x008 /* Internal SRAM DCR's */ + 0x030 0x008>; /* L2 cache DCR's */ + cache-line-size = <32>; /* 32 bytes */ + cache-size = <262144>; /* L2, 256K */ + interrupt-parent = <&UIC1>; + interrupts = <11 1>; + }; + plb { compatible = "ibm,plb-460gt", "ibm,plb4"; #address-cells = <2>; @@ -118,6 +129,13 @@ dcr-reg = <0x010 0x002>; }; + CRYPTO: crypto@180000 { + compatible = "amcc,ppc460gt-crypto", "amcc,ppc4xx-crypto"; + reg = <4 0x00180000 0x80400>; + interrupt-parent = <&UIC0>; + interrupts = <0x1d 0x4>; + }; + MAL0: mcmal { compatible = "ibm,mcmal-460gt", "ibm,mcmal2"; dcr-reg = <0x180 0x062>; @@ -186,6 +204,29 @@ reg = <0x03fa0000 0x00060000>; }; }; + + ndfc@3,0 { + compatible = "ibm,ndfc"; + reg = <0x00000003 0x00000000 0x00002000>; + ccr = <0x00001000>; + bank-settings = <0x80002222>; + #address-cells = <1>; + #size-cells = <1>; + + nand { + #address-cells = <1>; + #size-cells = <1>; + + partition@0 { + label = "u-boot"; + reg = <0x00000000 0x00100000>; + }; + partition@100000 { + label = "user"; + reg = <0x00000000 0x03f00000>; + }; + }; + }; }; UART0: serial@ef600300 { @@ -237,6 +278,20 @@ reg = <0xef600700 0x00000014>; interrupt-parent = <&UIC0>; interrupts = <0x2 0x4>; + #address-cells = <1>; + #size-cells = <0>; + rtc@68 { + compatible = "stm,m41t80"; + reg = <0x68>; + interrupt-parent = <&UIC2>; + interrupts = <0x19 0x8>; + }; + sttm@48 { + compatible = "ad,ad7414"; + reg = <0x48>; + interrupt-parent = <&UIC1>; + interrupts = <0x14 0x8>; + }; }; IIC1: i2c@ef600800 { @@ -275,7 +330,7 @@ EMAC0: ethernet@ef600e00 { device_type = "network"; - compatible = "ibm,emac-460gt", "ibm,emac4"; + compatible = "ibm,emac-460gt", "ibm,emac4sync"; interrupt-parent = <&EMAC0>; interrupts = <0x0 0x1>; #interrupt-cells = <1>; @@ -283,7 +338,7 @@ #size-cells = <0>; interrupt-map = ; - reg = <0xef600e00 0x00000074>; + reg = <0xef600e00 0x000000c4>; local-mac-address = [000000000000]; /* Filled in by U-Boot */ mal-device = <&MAL0>; mal-tx-channel = <0>; @@ -305,7 +360,7 @@ EMAC1: ethernet@ef600f00 { device_type = "network"; - compatible = "ibm,emac-460gt", "ibm,emac4"; + compatible = "ibm,emac-460gt", "ibm,emac4sync"; interrupt-parent = <&EMAC1>; interrupts = <0x0 0x1>; #interrupt-cells = <1>; @@ -313,7 +368,7 @@ #size-cells = <0>; interrupt-map = ; - reg = <0xef600f00 0x00000074>; + reg = <0xef600f00 0x000000c4>; local-mac-address = [000000000000]; /* Filled in by U-Boot */ mal-device = <&MAL0>; mal-tx-channel = <1>; @@ -336,7 +391,7 @@ EMAC2: ethernet@ef601100 { device_type = "network"; - compatible = "ibm,emac-460gt", "ibm,emac4"; + compatible = "ibm,emac-460gt", "ibm,emac4sync"; interrupt-parent = <&EMAC2>; interrupts = <0x0 0x1>; #interrupt-cells = <1>; @@ -344,7 +399,7 @@ #size-cells = <0>; interrupt-map = ; - reg = <0xef601100 0x00000074>; + reg = <0xef601100 0x000000c4>; local-mac-address = [000000000000]; /* Filled in by U-Boot */ mal-device = <&MAL0>; mal-tx-channel = <2>; @@ -366,7 +421,7 @@ EMAC3: ethernet@ef601200 { device_type = "network"; - compatible = "ibm,emac-460gt", "ibm,emac4"; + compatible = "ibm,emac-460gt", "ibm,emac4sync"; interrupt-parent = <&EMAC3>; interrupts = <0x0 0x1>; #interrupt-cells = <1>; @@ -374,7 +429,7 @@ #size-cells = <0>; interrupt-map = ; - reg = <0xef601200 0x00000074>; + reg = <0xef601200 0x000000c4>; local-mac-address = [000000000000]; /* Filled in by U-Boot */ mal-device = <&MAL0>; mal-tx-channel = <3>; @@ -414,6 +469,7 @@ * later cannot be changed */ ranges = <0x02000000 0x00000000 0x80000000 0x0000000d 0x80000000 0x00000000 0x80000000 + 0x02000000 0x00000000 0x00000000 0x0000000c 0x0ee00000 0x00000000 0x00100000 0x01000000 0x00000000 0x00000000 0x0000000c 0x08000000 0x00000000 0x00010000>; /* Inbound 2GB range starting at 0 */ @@ -444,6 +500,7 @@ * later cannot be changed */ ranges = <0x02000000 0x00000000 0x80000000 0x0000000e 0x00000000 0x00000000 0x80000000 + 0x02000000 0x00000000 0x00000000 0x0000000f 0x00000000 0x00000000 0x00100000 0x01000000 0x00000000 0x00000000 0x0000000f 0x80000000 0x00000000 0x00010000>; /* Inbound 2GB range starting at 0 */ @@ -485,6 +542,7 @@ * later cannot be changed */ ranges = <0x02000000 0x00000000 0x80000000 0x0000000e 0x80000000 0x00000000 0x80000000 + 0x02000000 0x00000000 0x00000000 0x0000000f 0x00100000 0x00000000 0x00100000 0x01000000 0x00000000 0x00000000 0x0000000f 0x80010000 0x00000000 0x00010000>; /* Inbound 2GB range starting at 0 */ -- cgit v0.10.2 From 573bff5ab9437b4e59ba9be34a7b08bca5ff054a Mon Sep 17 00:00:00 2001 From: Stefan Roese Date: Tue, 26 Jan 2010 05:53:20 +0000 Subject: powerpc/44x: Add MTD support to katmai defconfig Signed-off-by: Stefan Roese Cc: Josh Boyer Signed-off-by: Josh Boyer diff --git a/arch/powerpc/configs/44x/katmai_defconfig b/arch/powerpc/configs/44x/katmai_defconfig index dec901f..af244e1 100644 --- a/arch/powerpc/configs/44x/katmai_defconfig +++ b/arch/powerpc/configs/44x/katmai_defconfig @@ -1,7 +1,7 @@ # # Automatically generated make config: don't edit -# Linux kernel version: 2.6.33-rc1 -# Mon Jan 4 14:55:34 2010 +# Linux kernel version: 2.6.33-rc5 +# Tue Jan 26 14:40:58 2010 # # CONFIG_PPC64 is not set @@ -106,6 +106,7 @@ CONFIG_INITRAMFS_SOURCE="" CONFIG_RD_GZIP=y # CONFIG_RD_BZIP2 is not set # CONFIG_RD_LZMA is not set +# CONFIG_RD_LZO is not set # CONFIG_CC_OPTIMIZE_FOR_SIZE is not set CONFIG_SYSCTL=y CONFIG_ANON_INODES=y @@ -442,7 +443,90 @@ CONFIG_EXTRA_FIRMWARE="" # CONFIG_SYS_HYPERVISOR is not set CONFIG_CONNECTOR=y CONFIG_PROC_EVENTS=y -# CONFIG_MTD is not set +CONFIG_MTD=y +# CONFIG_MTD_DEBUG is not set +# CONFIG_MTD_TESTS is not set +# CONFIG_MTD_CONCAT is not set +CONFIG_MTD_PARTITIONS=y +# CONFIG_MTD_REDBOOT_PARTS is not set +CONFIG_MTD_CMDLINE_PARTS=y +CONFIG_MTD_OF_PARTS=y +# CONFIG_MTD_AR7_PARTS is not set + +# +# User Modules And Translation Layers +# +CONFIG_MTD_CHAR=y +CONFIG_MTD_BLKDEVS=y +CONFIG_MTD_BLOCK=y +# CONFIG_FTL is not set +# CONFIG_NFTL is not set +# CONFIG_INFTL is not set +# CONFIG_RFD_FTL is not set +# CONFIG_SSFDC is not set +# CONFIG_MTD_OOPS is not set + +# +# RAM/ROM/Flash chip drivers +# +CONFIG_MTD_CFI=y +# CONFIG_MTD_JEDECPROBE is not set +CONFIG_MTD_GEN_PROBE=y +# CONFIG_MTD_CFI_ADV_OPTIONS is not set +CONFIG_MTD_MAP_BANK_WIDTH_1=y +CONFIG_MTD_MAP_BANK_WIDTH_2=y +CONFIG_MTD_MAP_BANK_WIDTH_4=y +# CONFIG_MTD_MAP_BANK_WIDTH_8 is not set +# CONFIG_MTD_MAP_BANK_WIDTH_16 is not set +# CONFIG_MTD_MAP_BANK_WIDTH_32 is not set +CONFIG_MTD_CFI_I1=y +CONFIG_MTD_CFI_I2=y +# CONFIG_MTD_CFI_I4 is not set +# CONFIG_MTD_CFI_I8 is not set +# CONFIG_MTD_CFI_INTELEXT is not set +CONFIG_MTD_CFI_AMDSTD=y +# CONFIG_MTD_CFI_STAA is not set +CONFIG_MTD_CFI_UTIL=y +# CONFIG_MTD_RAM is not set +# CONFIG_MTD_ROM is not set +# CONFIG_MTD_ABSENT is not set + +# +# Mapping drivers for chip access +# +# CONFIG_MTD_COMPLEX_MAPPINGS is not set +# CONFIG_MTD_PHYSMAP is not set +CONFIG_MTD_PHYSMAP_OF=y +# CONFIG_MTD_INTEL_VR_NOR is not set +# CONFIG_MTD_PLATRAM is not set + +# +# Self-contained MTD device drivers +# +# CONFIG_MTD_PMC551 is not set +# CONFIG_MTD_SLRAM is not set +# CONFIG_MTD_PHRAM is not set +# CONFIG_MTD_MTDRAM is not set +# CONFIG_MTD_BLOCK2MTD is not set + +# +# Disk-On-Chip Device Drivers +# +# CONFIG_MTD_DOC2000 is not set +# CONFIG_MTD_DOC2001 is not set +# CONFIG_MTD_DOC2001PLUS is not set +# CONFIG_MTD_NAND is not set +# CONFIG_MTD_ONENAND is not set + +# +# LPDDR flash memory drivers +# +# CONFIG_MTD_LPDDR is not set + +# +# UBI - Unsorted block images +# +# CONFIG_MTD_UBI is not set CONFIG_OF_DEVICE=y # CONFIG_PARPORT is not set CONFIG_BLK_DEV=y @@ -500,7 +584,7 @@ CONFIG_HAVE_IDE=y # # -# See the help texts for more information. +# The newer stack is recommended. # # CONFIG_FIREWIRE is not set # CONFIG_IEEE1394 is not set @@ -763,7 +847,6 @@ CONFIG_EXT2_FS=y # CONFIG_EXT2_FS_XIP is not set # CONFIG_EXT3_FS is not set # CONFIG_EXT4_FS is not set -CONFIG_EXT4_USE_FOR_EXT23=y # CONFIG_REISERFS_FS is not set # CONFIG_JFS_FS is not set # CONFIG_FS_POSIX_ACL is not set @@ -820,6 +903,7 @@ CONFIG_MISC_FILESYSTEMS=y # CONFIG_BEFS_FS is not set # CONFIG_BFS_FS is not set # CONFIG_EFS_FS is not set +# CONFIG_JFFS2_FS is not set CONFIG_CRAMFS=y # CONFIG_SQUASHFS is not set # CONFIG_VXFS_FS is not set -- cgit v0.10.2 From 77f720b71d88a3cbf574c113566a31c93099f97d Mon Sep 17 00:00:00 2001 From: Stephane Chatty Date: Sat, 6 Feb 2010 15:17:13 +0100 Subject: HID: Support for MosArt multitouch panel Added support for MosArt dual-touch panels, present in the Asus T91MT notebook. Signed-off-by: Stephane Chatty Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index 617ed7d..2b93d7f 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -204,6 +204,12 @@ config HID_MICROSOFT ---help--- Support for Microsoft devices that are not fully compliant with HID standard. +config HID_MOSART + tristate "MosArt" + depends on USB_HID + ---help--- + Support for MosArt dual-touch panels. + config HID_MONTEREY tristate "Monterey" if EMBEDDED depends on USB_HID diff --git a/drivers/hid/Makefile b/drivers/hid/Makefile index b62d4b3..0305169 100644 --- a/drivers/hid/Makefile +++ b/drivers/hid/Makefile @@ -37,6 +37,7 @@ obj-$(CONFIG_HID_KYE) += hid-kye.o obj-$(CONFIG_HID_LOGITECH) += hid-logitech.o obj-$(CONFIG_HID_MICROSOFT) += hid-microsoft.o obj-$(CONFIG_HID_MONTEREY) += hid-monterey.o +obj-$(CONFIG_HID_MOSART) += hid-mosart.o obj-$(CONFIG_HID_NTRIG) += hid-ntrig.o obj-$(CONFIG_HID_ORTEK) += hid-ortek.o obj-$(CONFIG_HID_QUANTA) += hid-quanta.o diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index 282f4a1..da97195 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1550,8 +1550,9 @@ static const struct hid_device_id hid_ignore_list[] = { { HID_USB_DEVICE(USB_VENDOR_ID_AIPTEK, USB_DEVICE_ID_AIPTEK_24) }, { HID_USB_DEVICE(USB_VENDOR_ID_AIRCABLE, USB_DEVICE_ID_AIRCABLE1) }, { HID_USB_DEVICE(USB_VENDOR_ID_ALCOR, USB_DEVICE_ID_ALCOR_USBRS232) }, - { HID_USB_DEVICE(USB_VENDOR_ID_ASUS, USB_DEVICE_ID_ASUS_LCM)}, - { HID_USB_DEVICE(USB_VENDOR_ID_ASUS, USB_DEVICE_ID_ASUS_LCM2)}, + { HID_USB_DEVICE(USB_VENDOR_ID_ASUS, USB_DEVICE_ID_ASUS_T91MT)}, + { HID_USB_DEVICE(USB_VENDOR_ID_ASUSTEK, USB_DEVICE_ID_ASUSTEK_LCM)}, + { HID_USB_DEVICE(USB_VENDOR_ID_ASUSTEK, USB_DEVICE_ID_ASUSTEK_LCM2)}, { HID_USB_DEVICE(USB_VENDOR_ID_AVERMEDIA, USB_DEVICE_ID_AVER_FM_MR800) }, { HID_USB_DEVICE(USB_VENDOR_ID_BERKSHIRE, USB_DEVICE_ID_BERKSHIRE_PCWD) }, { HID_USB_DEVICE(USB_VENDOR_ID_CIDC, 0x0103) }, diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index 282bfe8..ab370f1 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -99,9 +99,12 @@ #define USB_DEVICE_ID_APPLE_ATV_IRCONTROL 0x8241 #define USB_DEVICE_ID_APPLE_IRCONTROL4 0x8242 -#define USB_VENDOR_ID_ASUS 0x0b05 -#define USB_DEVICE_ID_ASUS_LCM 0x1726 -#define USB_DEVICE_ID_ASUS_LCM2 0x175b +#define USB_VENDOR_ID_ASUS 0x0486 +#define USB_DEVICE_ID_ASUS_T91MT 0x0185 + +#define USB_VENDOR_ID_ASUSTEK 0x0b05 +#define USB_DEVICE_ID_ASUSTEK_LCM 0x1726 +#define USB_DEVICE_ID_ASUSTEK_LCM2 0x175b #define USB_VENDOR_ID_ATEN 0x0557 #define USB_DEVICE_ID_ATEN_UC100KM 0x2004 diff --git a/drivers/hid/hid-mosart.c b/drivers/hid/hid-mosart.c new file mode 100644 index 0000000..f6ec4f1 --- /dev/null +++ b/drivers/hid/hid-mosart.c @@ -0,0 +1,274 @@ +/* + * HID driver for the multitouch panel on the ASUS EeePC T91MT + * + * Copyright (c) 2009-2010 Stephane Chatty + * Copyright (c) 2010 Teemu Tuominen + * + */ + +/* + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include +#include +#include +#include +#include "usbhid/usbhid.h" + +MODULE_VERSION("1.00"); +MODULE_AUTHOR("Stephane Chatty "); +MODULE_DESCRIPTION("MosArt dual-touch panel"); +MODULE_LICENSE("GPL"); + +#include "hid-ids.h" + +struct mosart_data { + __u16 x, y; + __u8 id; + bool valid; /* valid finger data, or just placeholder? */ + bool first; /* is this the first finger in this frame? */ + bool activity_now; /* at least one active finger in this frame? */ + bool activity; /* at least one active finger previously? */ +}; + +static int mosart_input_mapping(struct hid_device *hdev, struct hid_input *hi, + struct hid_field *field, struct hid_usage *usage, + unsigned long **bit, int *max) +{ + switch (usage->hid & HID_USAGE_PAGE) { + + case HID_UP_GENDESK: + switch (usage->hid) { + case HID_GD_X: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_POSITION_X); + /* touchscreen emulation */ + input_set_abs_params(hi->input, ABS_X, + field->logical_minimum, + field->logical_maximum, 0, 0); + return 1; + case HID_GD_Y: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_POSITION_Y); + /* touchscreen emulation */ + input_set_abs_params(hi->input, ABS_Y, + field->logical_minimum, + field->logical_maximum, 0, 0); + return 1; + } + return 0; + + case HID_UP_DIGITIZER: + switch (usage->hid) { + case HID_DG_CONFIDENCE: + case HID_DG_TIPSWITCH: + case HID_DG_INPUTMODE: + case HID_DG_DEVICEINDEX: + case HID_DG_CONTACTCOUNT: + case HID_DG_CONTACTMAX: + case HID_DG_TIPPRESSURE: + case HID_DG_WIDTH: + case HID_DG_HEIGHT: + return -1; + case HID_DG_INRANGE: + /* touchscreen emulation */ + hid_map_usage(hi, usage, bit, max, EV_KEY, BTN_TOUCH); + return 1; + + case HID_DG_CONTACTID: + hid_map_usage(hi, usage, bit, max, + EV_ABS, ABS_MT_TRACKING_ID); + return 1; + + } + return 0; + + case 0xff000000: + /* ignore HID features */ + return -1; + } + + return 0; +} + +static int mosart_input_mapped(struct hid_device *hdev, struct hid_input *hi, + struct hid_field *field, struct hid_usage *usage, + unsigned long **bit, int *max) +{ + if (usage->type == EV_KEY || usage->type == EV_ABS) + clear_bit(usage->code, *bit); + + return 0; +} + +/* + * this function is called when a whole finger has been parsed, + * so that it can decide what to send to the input layer. + */ +static void mosart_filter_event(struct mosart_data *td, struct input_dev *input) +{ + td->first = !td->first; /* touchscreen emulation */ + + if (!td->valid) { + /* + * touchscreen emulation: if no finger in this frame is valid + * and there previously was finger activity, this is a release + */ + if (!td->first && !td->activity_now && td->activity) { + input_event(input, EV_KEY, BTN_TOUCH, 0); + td->activity = false; + } + return; + } + + input_event(input, EV_ABS, ABS_MT_TRACKING_ID, td->id); + input_event(input, EV_ABS, ABS_MT_POSITION_X, td->x); + input_event(input, EV_ABS, ABS_MT_POSITION_Y, td->y); + + input_mt_sync(input); + td->valid = false; + + /* touchscreen emulation: if first active finger in this frame... */ + if (!td->activity_now) { + /* if there was no previous activity, emit touch event */ + if (!td->activity) { + input_event(input, EV_KEY, BTN_TOUCH, 1); + td->activity = true; + } + td->activity_now = true; + /* and in any case this is our preferred finger */ + input_event(input, EV_ABS, ABS_X, td->x); + input_event(input, EV_ABS, ABS_Y, td->y); + } +} + + +static int mosart_event(struct hid_device *hid, struct hid_field *field, + struct hid_usage *usage, __s32 value) +{ + struct mosart_data *td = hid_get_drvdata(hid); + + if (hid->claimed & HID_CLAIMED_INPUT) { + struct input_dev *input = field->hidinput->input; + switch (usage->hid) { + case HID_DG_INRANGE: + td->valid = !!value; + break; + case HID_GD_X: + td->x = value; + break; + case HID_GD_Y: + td->y = value; + mosart_filter_event(td, input); + break; + case HID_DG_CONTACTID: + td->id = value; + break; + case HID_DG_CONTACTCOUNT: + /* touch emulation: this is the last field in a frame */ + td->first = false; + td->activity_now = false; + break; + case HID_DG_CONFIDENCE: + case HID_DG_TIPSWITCH: + /* avoid interference from generic hidinput handling */ + break; + + default: + /* fallback to the generic hidinput handling */ + return 0; + } + } + + /* we have handled the hidinput part, now remains hiddev */ + if (hid->claimed & HID_CLAIMED_HIDDEV && hid->hiddev_hid_event) + hid->hiddev_hid_event(hid, field, usage, value); + + return 1; +} + +static int mosart_probe(struct hid_device *hdev, const struct hid_device_id *id) +{ + int ret; + struct mosart_data *td; + + + td = kmalloc(sizeof(struct mosart_data), GFP_KERNEL); + if (!td) { + dev_err(&hdev->dev, "cannot allocate MosArt data\n"); + return -ENOMEM; + } + td->valid = false; + td->activity = false; + td->activity_now = false; + td->first = false; + hid_set_drvdata(hdev, td); + + /* currently, it's better to have one evdev device only */ +#if 0 + hdev->quirks |= HID_QUIRK_MULTI_INPUT; +#endif + + ret = hid_parse(hdev); + if (ret == 0) + ret = hid_hw_start(hdev, HID_CONNECT_DEFAULT); + + if (ret == 0) { + struct hid_report_enum *re = hdev->report_enum + + HID_FEATURE_REPORT; + struct hid_report *r = re->report_id_hash[7]; + + r->field[0]->value[0] = 0x02; + usbhid_submit_report(hdev, r, USB_DIR_OUT); + } else + kfree(td); + + return ret; +} + +static void mosart_remove(struct hid_device *hdev) +{ + hid_hw_stop(hdev); + kfree(hid_get_drvdata(hdev)); + hid_set_drvdata(hdev, NULL); +} + +static const struct hid_device_id mosart_devices[] = { + { HID_USB_DEVICE(USB_VENDOR_ID_ASUS, USB_DEVICE_ID_ASUS_T91MT) }, + { } +}; +MODULE_DEVICE_TABLE(hid, mosart_devices); + +static const struct hid_usage_id mosart_grabbed_usages[] = { + { HID_ANY_ID, HID_ANY_ID, HID_ANY_ID }, + { HID_ANY_ID - 1, HID_ANY_ID - 1, HID_ANY_ID - 1} +}; + +static struct hid_driver mosart_driver = { + .name = "mosart", + .id_table = mosart_devices, + .probe = mosart_probe, + .remove = mosart_remove, + .input_mapping = mosart_input_mapping, + .input_mapped = mosart_input_mapped, + .usage_table = mosart_grabbed_usages, + .event = mosart_event, +}; + +static int __init mosart_init(void) +{ + return hid_register_driver(&mosart_driver); +} + +static void __exit mosart_exit(void) +{ + hid_unregister_driver(&mosart_driver); +} + +module_init(mosart_init); +module_exit(mosart_exit); + -- cgit v0.10.2 From 90a006abf8015c8cab893555244d8fc673b24839 Mon Sep 17 00:00:00 2001 From: Michael Poole Date: Sun, 24 Jan 2010 22:32:29 -0500 Subject: HID: Export hid_register_report The Apple Magic Mouse (and probably other devices) publish reports that are not called out in their HID report descriptors -- they only send them when enabled through other writes to the device. This allows a driver to handle these unlisted reports. Signed-off-by: Michael Poole Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index f7f80e1..66a91eb 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -51,7 +51,7 @@ EXPORT_SYMBOL_GPL(hid_debug); * Register a new report for a device. */ -static struct hid_report *hid_register_report(struct hid_device *device, unsigned type, unsigned id) +struct hid_report *hid_register_report(struct hid_device *device, unsigned type, unsigned id) { struct hid_report_enum *report_enum = device->report_enum + type; struct hid_report *report; @@ -75,6 +75,7 @@ static struct hid_report *hid_register_report(struct hid_device *device, unsigne return report; } +EXPORT_SYMBOL_GPL(hid_register_report); /* * Register a new field for this report. diff --git a/include/linux/hid.h b/include/linux/hid.h index 3661a62..456838c 100644 --- a/include/linux/hid.h +++ b/include/linux/hid.h @@ -690,6 +690,7 @@ int hid_input_report(struct hid_device *, int type, u8 *, int, int); int hidinput_find_field(struct hid_device *hid, unsigned int type, unsigned int code, struct hid_field **field); void hid_output_report(struct hid_report *report, __u8 *data); struct hid_device *hid_allocate_device(void); +struct hid_report *hid_register_report(struct hid_device *device, unsigned type, unsigned id); int hid_parse_report(struct hid_device *hid, __u8 *start, unsigned size); int hid_check_keys_pressed(struct hid_device *hid); int hid_connect(struct hid_device *hid, unsigned int connect_mask); -- cgit v0.10.2 From 128537cea464d919febeaea2000e256749f317eb Mon Sep 17 00:00:00 2001 From: Michael Poole Date: Sat, 6 Feb 2010 12:24:36 -0500 Subject: HID: add a device driver for the Apple Magic Mouse. The Magic Mouse requires that a driver send an unlock Report(Feature) command, similar to the Wacom wireless tablet and Sixaxis controller quirks. This turns on an Input Report that isn't published in the input Report descriptor that contains touch data (and usually overrides the normal motion and click Report). Because the mouse has only one switch and no scroll wheel, the driver (under control of parameters) emulates a middle button and scroll wheel. User space could also ignore and/or re-synthesize those events based on the reported events. Some user-space tools to talk to the mouse directly (that is, when it is not associated with the host's HIDP stack) are at http://github.com/entrope/linux-magicmouse Signed-off-by: Michael Poole Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index 24d90ea..ba14ec8 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -183,6 +183,16 @@ config LOGIRUMBLEPAD2_FF Say Y here if you want to enable force feedback support for Logitech Rumblepad 2 devices. +config HID_MAGICMOUSE + tristate "Apple" if EMBEDDED + depends on BT_HIDP + default !EMBEDDED + ---help--- + Support for the Apple Magic Mouse. + + Say Y here if you want support for the multi-touch features of the + Apple Wireless "Magic" Mouse. + config HID_MICROSOFT tristate "Microsoft" if EMBEDDED depends on USB_HID diff --git a/drivers/hid/Makefile b/drivers/hid/Makefile index 0de2dff..45d81e9 100644 --- a/drivers/hid/Makefile +++ b/drivers/hid/Makefile @@ -31,6 +31,7 @@ obj-$(CONFIG_HID_GYRATION) += hid-gyration.o obj-$(CONFIG_HID_KENSINGTON) += hid-kensington.o obj-$(CONFIG_HID_KYE) += hid-kye.o obj-$(CONFIG_HID_LOGITECH) += hid-logitech.o +obj-$(CONFIG_HID_MAGICMOUSE) += hid-magicmouse.o obj-$(CONFIG_HID_MICROSOFT) += hid-microsoft.o obj-$(CONFIG_HID_MONTEREY) += hid-monterey.o obj-$(CONFIG_HID_NTRIG) += hid-ntrig.o diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c index 66a91eb..f038422 100644 --- a/drivers/hid/hid-core.c +++ b/drivers/hid/hid-core.c @@ -1254,6 +1254,7 @@ static const struct hid_device_id hid_blacklist[] = { { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_ATV_IRCONTROL) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_IRCONTROL4) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_MIGHTYMOUSE) }, + { HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_MAGICMOUSE) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_FOUNTAIN_ANSI) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_FOUNTAIN_ISO) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_GEYSER_ANSI) }, diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index 010368e..11e5218 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -56,6 +56,7 @@ #define USB_VENDOR_ID_APPLE 0x05ac #define USB_DEVICE_ID_APPLE_MIGHTYMOUSE 0x0304 +#define USB_DEVICE_ID_APPLE_MAGICMOUSE 0x030d #define USB_DEVICE_ID_APPLE_FOUNTAIN_ANSI 0x020e #define USB_DEVICE_ID_APPLE_FOUNTAIN_ISO 0x020f #define USB_DEVICE_ID_APPLE_GEYSER_ANSI 0x0214 diff --git a/drivers/hid/hid-magicmouse.c b/drivers/hid/hid-magicmouse.c new file mode 100644 index 0000000..f94b3e4 --- /dev/null +++ b/drivers/hid/hid-magicmouse.c @@ -0,0 +1,469 @@ +/* + * Apple "Magic" Wireless Mouse driver + * + * Copyright (c) 2010 Michael Poole + */ + +/* + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include +#include +#include +#include + +#include "hid-ids.h" + +static bool emulate_3button = 1; +module_param(emulate_3button, bool, 0644); +MODULE_PARM_DESC(emulate_3button, "Emulate a middle button"); + +static int middle_button_start = -350; +static int middle_button_stop = +350; + +static bool emulate_scroll_wheel = 1; +module_param(emulate_scroll_wheel, bool, 0644); +MODULE_PARM_DESC(emulate_scroll_wheel, "Emulate a scroll wheel"); + +static bool report_touches = 1; +module_param(report_touches, bool, 0644); +MODULE_PARM_DESC(report_touches, "Emit touch records (otherwise, only use them for emulation)"); + +static bool report_undeciphered = 0; +module_param(report_undeciphered, bool, 0644); +MODULE_PARM_DESC(report_undeciphered, "Report undeciphered multi-touch state field using a MSC_RAW event"); + +#define TOUCH_REPORT_ID 0x29 +/* These definitions are not precise, but they're close enough. (Bits + * 0x03 seem to indicate the aspect ratio of the touch, bits 0x70 seem + * to be some kind of bit mask -- 0x20 may be a near-field reading, + * and 0x40 is actual contact, and 0x10 may be a start/stop or change + * indication.) + */ +#define TOUCH_STATE_MASK 0xf0 +#define TOUCH_STATE_NONE 0x00 +#define TOUCH_STATE_START 0x30 +#define TOUCH_STATE_DRAG 0x40 + +/** + * struct magicmouse_sc - Tracks Magic Mouse-specific data. + * @input: Input device through which we report events. + * @quirks: Currently unused. + * @last_timestamp: Timestamp from most recent (18-bit) touch report + * (units of milliseconds over short windows, but seems to + * increase faster when there are no touches). + * @delta_time: 18-bit difference between the two most recent touch + * reports from the mouse. + * @ntouches: Number of touches in most recent touch report. + * @scroll_accel: Number of consecutive scroll motions. + * @scroll_jiffies: Time of last scroll motion. + * @touches: Most recent data for a touch, indexed by tracking ID. + * @tracking_ids: Mapping of current touch input data to @touches. + */ +struct magicmouse_sc { + struct input_dev *input; + unsigned long quirks; + + int last_timestamp; + int delta_time; + int ntouches; + int scroll_accel; + unsigned long scroll_jiffies; + + struct { + short x; + short y; + short scroll_y; + u8 size; + } touches[16]; + int tracking_ids[16]; +}; + +static int magicmouse_firm_touch(struct magicmouse_sc *msc) +{ + int touch = -1; + int ii; + + /* If there is only one "firm" touch, set touch to its + * tracking ID. + */ + for (ii = 0; ii < msc->ntouches; ii++) { + int idx = msc->tracking_ids[ii]; + if (msc->touches[idx].size < 8) { + /* Ignore this touch. */ + } else if (touch >= 0) { + touch = -1; + break; + } else { + touch = idx; + } + } + + return touch; +} + +static void magicmouse_emit_buttons(struct magicmouse_sc *msc, int state) +{ + int last_state = test_bit(BTN_LEFT, msc->input->key) << 0 | + test_bit(BTN_RIGHT, msc->input->key) << 1 | + test_bit(BTN_MIDDLE, msc->input->key) << 2; + + if (emulate_3button) { + int id; + + /* If some button was pressed before, keep it held + * down. Otherwise, if there's exactly one firm + * touch, use that to override the mouse's guess. + */ + if (state == 0) { + /* The button was released. */ + } else if (last_state != 0) { + state = last_state; + } else if ((id = magicmouse_firm_touch(msc)) >= 0) { + int x = msc->touches[id].x; + if (x < middle_button_start) + state = 1; + else if (x > middle_button_stop) + state = 2; + else + state = 4; + } /* else: we keep the mouse's guess */ + + input_report_key(msc->input, BTN_MIDDLE, state & 4); + } + + input_report_key(msc->input, BTN_LEFT, state & 1); + input_report_key(msc->input, BTN_RIGHT, state & 2); + + if (state != last_state) + msc->scroll_accel = 0; +} + +static void magicmouse_emit_touch(struct magicmouse_sc *msc, int raw_id, u8 *tdata) +{ + struct input_dev *input = msc->input; + __s32 x_y = tdata[0] << 8 | tdata[1] << 16 | tdata[2] << 24; + int misc = tdata[5] | tdata[6] << 8; + int id = (misc >> 6) & 15; + int x = x_y << 12 >> 20; + int y = -(x_y >> 20); + + /* Store tracking ID and other fields. */ + msc->tracking_ids[raw_id] = id; + msc->touches[id].x = x; + msc->touches[id].y = y; + msc->touches[id].size = misc & 63; + + /* If requested, emulate a scroll wheel by detecting small + * vertical touch motions along the middle of the mouse. + */ + if (emulate_scroll_wheel && + middle_button_start < x && x < middle_button_stop) { + static const int accel_profile[] = { + 256, 228, 192, 160, 128, 96, 64, 32, + }; + unsigned long now = jiffies; + int step = msc->touches[id].scroll_y - y; + + /* Reset acceleration after half a second. */ + if (time_after(now, msc->scroll_jiffies + HZ / 2)) + msc->scroll_accel = 0; + + /* Calculate and apply the scroll motion. */ + switch (tdata[7] & TOUCH_STATE_MASK) { + case TOUCH_STATE_START: + msc->touches[id].scroll_y = y; + msc->scroll_accel = min_t(int, msc->scroll_accel + 1, + ARRAY_SIZE(accel_profile) - 1); + break; + case TOUCH_STATE_DRAG: + step = step / accel_profile[msc->scroll_accel]; + if (step != 0) { + msc->touches[id].scroll_y = y; + msc->scroll_jiffies = now; + input_report_rel(input, REL_WHEEL, step); + } + break; + } + } + + /* Generate the input events for this touch. */ + if (report_touches) { + int orientation = (misc >> 10) - 32; + + input_report_abs(input, ABS_MT_TRACKING_ID, id); + input_report_abs(input, ABS_MT_TOUCH_MAJOR, tdata[3]); + input_report_abs(input, ABS_MT_TOUCH_MINOR, tdata[4]); + input_report_abs(input, ABS_MT_ORIENTATION, orientation); + input_report_abs(input, ABS_MT_POSITION_X, x); + input_report_abs(input, ABS_MT_POSITION_Y, y); + + if (report_undeciphered) { + input_event(input, EV_MSC, MSC_RAW, tdata[7]); + } + + input_mt_sync(input); + } +} + +static int magicmouse_raw_event(struct hid_device *hdev, + struct hid_report *report, u8 *data, int size) +{ + struct magicmouse_sc *msc = hid_get_drvdata(hdev); + struct input_dev *input = msc->input; + int x, y, ts, ii, clicks; + + switch (data[0]) { + case 0x10: + if (size != 6) + return 0; + x = (__s16)(data[2] | data[3] << 8); + y = (__s16)(data[4] | data[5] << 8); + clicks = data[1]; + break; + case TOUCH_REPORT_ID: + /* Expect six bytes of prefix, and N*8 bytes of touch data. */ + if (size < 6 || ((size - 6) % 8) != 0) + return 0; + ts = data[3] >> 6 | data[4] << 2 | data[5] << 10; + msc->delta_time = (ts - msc->last_timestamp) & 0x3ffff; + msc->last_timestamp = ts; + msc->ntouches = (size - 6) / 8; + for (ii = 0; ii < msc->ntouches; ii++) + magicmouse_emit_touch(msc, ii, data + ii * 8 + 6); + /* When emulating three-button mode, it is important + * to have the current touch information before + * generating a click event. + */ + x = (signed char)data[1]; + y = (signed char)data[2]; + clicks = data[3]; + break; + case 0x20: /* Theoretically battery status (0-100), but I have + * never seen it -- maybe it is only upon request. + */ + case 0x60: /* Unknown, maybe laser on/off. */ + case 0x61: /* Laser reflection status change. + * data[1]: 0 = spotted, 1 = lost + */ + default: + return 0; + } + + magicmouse_emit_buttons(msc, clicks & 3); + input_report_rel(input, REL_X, x); + input_report_rel(input, REL_Y, y); + input_sync(input); + return 1; +} + +static int magicmouse_input_open(struct input_dev *dev) +{ + struct hid_device *hid = input_get_drvdata(dev); + + return hid->ll_driver->open(hid); +} + +static void magicmouse_input_close(struct input_dev *dev) +{ + struct hid_device *hid = input_get_drvdata(dev); + + hid->ll_driver->close(hid); +} + +static void magicmouse_setup_input(struct input_dev *input, struct hid_device *hdev) +{ + input_set_drvdata(input, hdev); + input->event = hdev->ll_driver->hidinput_input_event; + input->open = magicmouse_input_open; + input->close = magicmouse_input_close; + + input->name = hdev->name; + input->phys = hdev->phys; + input->uniq = hdev->uniq; + input->id.bustype = hdev->bus; + input->id.vendor = hdev->vendor; + input->id.product = hdev->product; + input->id.version = hdev->version; + input->dev.parent = hdev->dev.parent; + + set_bit(EV_KEY, input->evbit); + set_bit(BTN_LEFT, input->keybit); + set_bit(BTN_RIGHT, input->keybit); + if (emulate_3button) + set_bit(BTN_MIDDLE, input->keybit); + set_bit(BTN_TOOL_FINGER, input->keybit); + + set_bit(EV_REL, input->evbit); + set_bit(REL_X, input->relbit); + set_bit(REL_Y, input->relbit); + if (emulate_scroll_wheel) + set_bit(REL_WHEEL, input->relbit); + + if (report_touches) { + set_bit(EV_ABS, input->evbit); + + set_bit(ABS_MT_TRACKING_ID, input->absbit); + input->absmin[ABS_MT_TRACKING_ID] = 0; + input->absmax[ABS_MT_TRACKING_ID] = 15; + input->absfuzz[ABS_MT_TRACKING_ID] = 0; + + set_bit(ABS_MT_TOUCH_MAJOR, input->absbit); + input->absmin[ABS_MT_TOUCH_MAJOR] = 0; + input->absmax[ABS_MT_TOUCH_MAJOR] = 255; + input->absfuzz[ABS_MT_TOUCH_MAJOR] = 4; + + set_bit(ABS_MT_TOUCH_MINOR, input->absbit); + input->absmin[ABS_MT_TOUCH_MINOR] = 0; + input->absmax[ABS_MT_TOUCH_MINOR] = 255; + input->absfuzz[ABS_MT_TOUCH_MINOR] = 4; + + set_bit(ABS_MT_ORIENTATION, input->absbit); + input->absmin[ABS_MT_ORIENTATION] = -32; + input->absmax[ABS_MT_ORIENTATION] = 31; + input->absfuzz[ABS_MT_ORIENTATION] = 1; + + set_bit(ABS_MT_POSITION_X, input->absbit); + input->absmin[ABS_MT_POSITION_X] = -1100; + input->absmax[ABS_MT_POSITION_X] = 1358; + input->absfuzz[ABS_MT_POSITION_X] = 4; + + /* Note: Touch Y position from the device is inverted relative + * to how pointer motion is reported (and relative to how USB + * HID recommends the coordinates work). This driver keeps + * the origin at the same position, and just uses the additive + * inverse of the reported Y. + */ + set_bit(ABS_MT_POSITION_Y, input->absbit); + input->absmin[ABS_MT_POSITION_Y] = -1589; + input->absmax[ABS_MT_POSITION_Y] = 2047; + input->absfuzz[ABS_MT_POSITION_Y] = 4; + } + + if (report_undeciphered) { + set_bit(EV_MSC, input->evbit); + set_bit(MSC_RAW, input->mscbit); + } +} + +static int magicmouse_probe(struct hid_device *hdev, + const struct hid_device_id *id) +{ + __u8 feature_1[] = { 0xd7, 0x01 }; + __u8 feature_2[] = { 0xf8, 0x01, 0x32 }; + struct input_dev *input; + struct magicmouse_sc *msc; + struct hid_report *report; + int ret; + + msc = kzalloc(sizeof(*msc), GFP_KERNEL); + if (msc == NULL) { + dev_err(&hdev->dev, "can't alloc magicmouse descriptor\n"); + return -ENOMEM; + } + + msc->quirks = id->driver_data; + hid_set_drvdata(hdev, msc); + + ret = hid_parse(hdev); + if (ret) { + dev_err(&hdev->dev, "magicmouse hid parse failed\n"); + goto err_free; + } + + ret = hid_hw_start(hdev, HID_CONNECT_DEFAULT); + if (ret) { + dev_err(&hdev->dev, "magicmouse hw start failed\n"); + goto err_free; + } + + report = hid_register_report(hdev, HID_INPUT_REPORT, TOUCH_REPORT_ID); + if (!report) { + dev_err(&hdev->dev, "unable to register touch report\n"); + ret = -ENOMEM; + goto err_free; + } + report->size = 6; + + ret = hdev->hid_output_raw_report(hdev, feature_1, sizeof(feature_1), + HID_FEATURE_REPORT); + if (ret != sizeof(feature_1)) { + dev_err(&hdev->dev, "unable to request touch data (1:%d)\n", + ret); + goto err_free; + } + ret = hdev->hid_output_raw_report(hdev, feature_2, + sizeof(feature_2), HID_FEATURE_REPORT); + if (ret != sizeof(feature_2)) { + dev_err(&hdev->dev, "unable to request touch data (2:%d)\n", + ret); + goto err_free; + } + + input = input_allocate_device(); + if (!input) { + dev_err(&hdev->dev, "can't alloc input device\n"); + ret = -ENOMEM; + goto err_free; + } + magicmouse_setup_input(input, hdev); + + ret = input_register_device(input); + if (ret) { + dev_err(&hdev->dev, "input device registration failed\n"); + goto err_both; + } + msc->input = input; + + return 0; + err_both: + input_free_device(input); + err_free: + kfree(msc); + return ret; +} + +static void magicmouse_remove(struct hid_device *hdev) +{ + hid_hw_stop(hdev); + kfree(hid_get_drvdata(hdev)); +} + +static const struct hid_device_id magic_mice[] = { + { HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_MAGICMOUSE), + .driver_data = 0 }, + { } +}; +MODULE_DEVICE_TABLE(hid, magic_mice); + +static struct hid_driver magicmouse_driver = { + .name = "magicmouse", + .id_table = magic_mice, + .probe = magicmouse_probe, + .remove = magicmouse_remove, + .raw_event = magicmouse_raw_event, +}; + +static int __init magicmouse_init(void) +{ + int ret; + + ret = hid_register_driver(&magicmouse_driver); + if (ret) + printk(KERN_ERR "can't register magicmouse driver\n"); + + return ret; +} + +static void __exit magicmouse_exit(void) +{ + hid_unregister_driver(&magicmouse_driver); +} + +module_init(magicmouse_init); +module_exit(magicmouse_exit); +MODULE_LICENSE("GPL"); -- cgit v0.10.2 From 9f5231472340ebcaf2dec75428b67d5d0d872857 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Wed, 10 Feb 2010 14:59:03 +0100 Subject: HID: fix up Kconfig entry for MagicMouse Make Apple MagicMouse Kconfig entry consistent with other dirvers. Also expand the tristate text a little bit more, so that it doesn't clash with already existing HID_APPLE. Signed-off-by: Jiri Kosina diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index ba14ec8..0e668cc 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -184,11 +184,10 @@ config LOGIRUMBLEPAD2_FF Rumblepad 2 devices. config HID_MAGICMOUSE - tristate "Apple" if EMBEDDED + tristate "Apple MagicMouse multi-touch support" depends on BT_HIDP - default !EMBEDDED ---help--- - Support for the Apple Magic Mouse. + Support for the Apple Magic Mouse multi-touch. Say Y here if you want support for the multi-touch features of the Apple Wireless "Magic" Mouse. -- cgit v0.10.2 From c8a8602b76b6703df1243e31be01cf0e4451e4a6 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Wed, 10 Feb 2010 15:29:02 +0100 Subject: HID: remove MODULE_VERSION from new drivers MODULE_VERSION doesn't make too much sense for drivers merged into main tree, as git is much better tracking revisions than any developer might ever be. Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-3m-pct.c b/drivers/hid/hid-3m-pct.c index 6d11e3d..2370aef 100644 --- a/drivers/hid/hid-3m-pct.c +++ b/drivers/hid/hid-3m-pct.c @@ -17,7 +17,6 @@ #include #include -MODULE_VERSION("0.6"); MODULE_AUTHOR("Stephane Chatty "); MODULE_DESCRIPTION("3M PCT multitouch panels"); MODULE_LICENSE("GPL"); diff --git a/drivers/hid/hid-mosart.c b/drivers/hid/hid-mosart.c index f6ec4f1..c871816 100644 --- a/drivers/hid/hid-mosart.c +++ b/drivers/hid/hid-mosart.c @@ -19,7 +19,6 @@ #include #include "usbhid/usbhid.h" -MODULE_VERSION("1.00"); MODULE_AUTHOR("Stephane Chatty "); MODULE_DESCRIPTION("MosArt dual-touch panel"); MODULE_LICENSE("GPL"); diff --git a/drivers/hid/hid-quanta.c b/drivers/hid/hid-quanta.c index 9ac4928..01dd51c 100644 --- a/drivers/hid/hid-quanta.c +++ b/drivers/hid/hid-quanta.c @@ -16,7 +16,6 @@ #include #include -MODULE_VERSION("1.00"); MODULE_AUTHOR("Stephane Chatty "); MODULE_DESCRIPTION("Quanta dual-touch panel"); MODULE_LICENSE("GPL"); diff --git a/drivers/hid/hid-stantum.c b/drivers/hid/hid-stantum.c index 04b0b98..2e592a0 100644 --- a/drivers/hid/hid-stantum.c +++ b/drivers/hid/hid-stantum.c @@ -16,7 +16,6 @@ #include #include -MODULE_VERSION("0.6"); MODULE_AUTHOR("Stephane Chatty "); MODULE_DESCRIPTION("Stantum HID multitouch panels"); MODULE_LICENSE("GPL"); -- cgit v0.10.2 From 0271f91003d3703675be13b8865618359a6caa1f Mon Sep 17 00:00:00 2001 From: Haicheng Li Date: Thu, 4 Feb 2010 19:06:33 +0800 Subject: x86, acpi: Map hotadded cpu to correct node. When hotadd new cpu to system, if its affinitive node is online, should map the cpu to its own node. Otherwise, let kernel select one online node for the new cpu later. Signed-off-by: Haicheng Li LKML-Reference: <4B6AAA39.6000300@linux.intel.com> Tested-by: Thomas Renninger Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 036d28a..7db15e1 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -49,6 +49,7 @@ EXPORT_SYMBOL(acpi_disabled); #ifdef CONFIG_X86_64 # include +# include #endif /* X86 */ #define BAD_MADT_ENTRY(entry, end) ( \ @@ -482,6 +483,25 @@ int acpi_register_gsi(struct device *dev, u32 gsi, int trigger, int polarity) */ #ifdef CONFIG_ACPI_HOTPLUG_CPU +static void acpi_map_cpu2node(acpi_handle handle, int cpu, int physid) +{ +#ifdef CONFIG_ACPI_NUMA + int nid; + + nid = acpi_get_node(handle); + if (nid == -1 || !node_online(nid)) + return; +#ifdef CONFIG_X86_64 + apicid_to_node[physid] = nid; + numa_set_node(cpu, nid); +#else /* CONFIG_X86_32 */ + apicid_2_node[physid] = nid; + cpu_to_node_map[cpu] = nid; +#endif + +#endif +} + static int __cpuinit _acpi_map_lsapic(acpi_handle handle, int *pcpu) { struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL }; @@ -540,6 +560,7 @@ static int __cpuinit _acpi_map_lsapic(acpi_handle handle, int *pcpu) } cpu = cpumask_first(new_map); + acpi_map_cpu2node(handle, cpu, physid); *pcpu = cpu; retval = 0; -- cgit v0.10.2 From 71b38bd4c1cc4f2b653064357e4efab77dfd711d Mon Sep 17 00:00:00 2001 From: Michael Poole Date: Thu, 11 Feb 2010 00:32:57 -0500 Subject: HID: magicmouse: coding style and probe failure fixes Use proper values to initialize bool configuration variables, tabs rather than spaces, no braces for one-line else clause, __set_bit() when the operation doesn't have to be atomic, input_set_abs_params() rather than writing the fields directly, and call hid_hw_stop() when appropriate to handle failures in the probe. Signed-off-by: Michael Poole Signed-off-by: Jiri Kosina diff --git a/drivers/hid/hid-magicmouse.c b/drivers/hid/hid-magicmouse.c index f94b3e4..4a3a94f 100644 --- a/drivers/hid/hid-magicmouse.c +++ b/drivers/hid/hid-magicmouse.c @@ -18,22 +18,22 @@ #include "hid-ids.h" -static bool emulate_3button = 1; +static bool emulate_3button = true; module_param(emulate_3button, bool, 0644); MODULE_PARM_DESC(emulate_3button, "Emulate a middle button"); static int middle_button_start = -350; static int middle_button_stop = +350; -static bool emulate_scroll_wheel = 1; +static bool emulate_scroll_wheel = true; module_param(emulate_scroll_wheel, bool, 0644); MODULE_PARM_DESC(emulate_scroll_wheel, "Emulate a scroll wheel"); -static bool report_touches = 1; +static bool report_touches = true; module_param(report_touches, bool, 0644); MODULE_PARM_DESC(report_touches, "Emit touch records (otherwise, only use them for emulation)"); -static bool report_undeciphered = 0; +static bool report_undeciphered; module_param(report_undeciphered, bool, 0644); MODULE_PARM_DESC(report_undeciphered, "Report undeciphered multi-touch state field using a MSC_RAW event"); @@ -108,9 +108,9 @@ static int magicmouse_firm_touch(struct magicmouse_sc *msc) static void magicmouse_emit_buttons(struct magicmouse_sc *msc, int state) { - int last_state = test_bit(BTN_LEFT, msc->input->key) << 0 | - test_bit(BTN_RIGHT, msc->input->key) << 1 | - test_bit(BTN_MIDDLE, msc->input->key) << 2; + int last_state = test_bit(BTN_LEFT, msc->input->key) << 0 | + test_bit(BTN_RIGHT, msc->input->key) << 1 | + test_bit(BTN_MIDDLE, msc->input->key) << 2; if (emulate_3button) { int id; @@ -177,7 +177,7 @@ static void magicmouse_emit_touch(struct magicmouse_sc *msc, int raw_id, u8 *tda switch (tdata[7] & TOUCH_STATE_MASK) { case TOUCH_STATE_START: msc->touches[id].scroll_y = y; - msc->scroll_accel = min_t(int, msc->scroll_accel + 1, + msc->scroll_accel = min_t(int, msc->scroll_accel + 1, ARRAY_SIZE(accel_profile) - 1); break; case TOUCH_STATE_DRAG: @@ -193,7 +193,7 @@ static void magicmouse_emit_touch(struct magicmouse_sc *msc, int raw_id, u8 *tda /* Generate the input events for this touch. */ if (report_touches) { - int orientation = (misc >> 10) - 32; + int orientation = (misc >> 10) - 32; input_report_abs(input, ABS_MT_TRACKING_ID, id); input_report_abs(input, ABS_MT_TOUCH_MAJOR, tdata[3]); @@ -202,9 +202,8 @@ static void magicmouse_emit_touch(struct magicmouse_sc *msc, int raw_id, u8 *tda input_report_abs(input, ABS_MT_POSITION_X, x); input_report_abs(input, ABS_MT_POSITION_Y, y); - if (report_undeciphered) { + if (report_undeciphered) input_event(input, EV_MSC, MSC_RAW, tdata[7]); - } input_mt_sync(input); } @@ -291,62 +290,41 @@ static void magicmouse_setup_input(struct input_dev *input, struct hid_device *h input->id.version = hdev->version; input->dev.parent = hdev->dev.parent; - set_bit(EV_KEY, input->evbit); - set_bit(BTN_LEFT, input->keybit); - set_bit(BTN_RIGHT, input->keybit); + __set_bit(EV_KEY, input->evbit); + __set_bit(BTN_LEFT, input->keybit); + __set_bit(BTN_RIGHT, input->keybit); if (emulate_3button) - set_bit(BTN_MIDDLE, input->keybit); - set_bit(BTN_TOOL_FINGER, input->keybit); + __set_bit(BTN_MIDDLE, input->keybit); + __set_bit(BTN_TOOL_FINGER, input->keybit); - set_bit(EV_REL, input->evbit); - set_bit(REL_X, input->relbit); - set_bit(REL_Y, input->relbit); + __set_bit(EV_REL, input->evbit); + __set_bit(REL_X, input->relbit); + __set_bit(REL_Y, input->relbit); if (emulate_scroll_wheel) - set_bit(REL_WHEEL, input->relbit); + __set_bit(REL_WHEEL, input->relbit); if (report_touches) { - set_bit(EV_ABS, input->evbit); - - set_bit(ABS_MT_TRACKING_ID, input->absbit); - input->absmin[ABS_MT_TRACKING_ID] = 0; - input->absmax[ABS_MT_TRACKING_ID] = 15; - input->absfuzz[ABS_MT_TRACKING_ID] = 0; - - set_bit(ABS_MT_TOUCH_MAJOR, input->absbit); - input->absmin[ABS_MT_TOUCH_MAJOR] = 0; - input->absmax[ABS_MT_TOUCH_MAJOR] = 255; - input->absfuzz[ABS_MT_TOUCH_MAJOR] = 4; - - set_bit(ABS_MT_TOUCH_MINOR, input->absbit); - input->absmin[ABS_MT_TOUCH_MINOR] = 0; - input->absmax[ABS_MT_TOUCH_MINOR] = 255; - input->absfuzz[ABS_MT_TOUCH_MINOR] = 4; - - set_bit(ABS_MT_ORIENTATION, input->absbit); - input->absmin[ABS_MT_ORIENTATION] = -32; - input->absmax[ABS_MT_ORIENTATION] = 31; - input->absfuzz[ABS_MT_ORIENTATION] = 1; - - set_bit(ABS_MT_POSITION_X, input->absbit); - input->absmin[ABS_MT_POSITION_X] = -1100; - input->absmax[ABS_MT_POSITION_X] = 1358; - input->absfuzz[ABS_MT_POSITION_X] = 4; - + __set_bit(EV_ABS, input->evbit); + + input_set_abs_params(input, ABS_MT_TRACKING_ID, 0, 15, 0, 0); + input_set_abs_params(input, ABS_MT_TOUCH_MAJOR, 0, 255, 4, 0); + input_set_abs_params(input, ABS_MT_TOUCH_MINOR, 0, 255, 4, 0); + input_set_abs_params(input, ABS_MT_ORIENTATION, -32, 31, 1, 0); + input_set_abs_params(input, ABS_MT_POSITION_X, -1100, 1358, + 4, 0); /* Note: Touch Y position from the device is inverted relative * to how pointer motion is reported (and relative to how USB * HID recommends the coordinates work). This driver keeps * the origin at the same position, and just uses the additive * inverse of the reported Y. */ - set_bit(ABS_MT_POSITION_Y, input->absbit); - input->absmin[ABS_MT_POSITION_Y] = -1589; - input->absmax[ABS_MT_POSITION_Y] = 2047; - input->absfuzz[ABS_MT_POSITION_Y] = 4; + input_set_abs_params(input, ABS_MT_POSITION_Y, -1589, 2047, + 4, 0); } if (report_undeciphered) { - set_bit(EV_MSC, input->evbit); - set_bit(MSC_RAW, input->mscbit); + __set_bit(EV_MSC, input->evbit); + __set_bit(MSC_RAW, input->mscbit); } } @@ -385,7 +363,7 @@ static int magicmouse_probe(struct hid_device *hdev, if (!report) { dev_err(&hdev->dev, "unable to register touch report\n"); ret = -ENOMEM; - goto err_free; + goto err_stop_hw; } report->size = 6; @@ -394,35 +372,37 @@ static int magicmouse_probe(struct hid_device *hdev, if (ret != sizeof(feature_1)) { dev_err(&hdev->dev, "unable to request touch data (1:%d)\n", ret); - goto err_free; + goto err_stop_hw; } ret = hdev->hid_output_raw_report(hdev, feature_2, sizeof(feature_2), HID_FEATURE_REPORT); if (ret != sizeof(feature_2)) { dev_err(&hdev->dev, "unable to request touch data (2:%d)\n", ret); - goto err_free; + goto err_stop_hw; } input = input_allocate_device(); if (!input) { dev_err(&hdev->dev, "can't alloc input device\n"); ret = -ENOMEM; - goto err_free; + goto err_stop_hw; } magicmouse_setup_input(input, hdev); ret = input_register_device(input); if (ret) { dev_err(&hdev->dev, "input device registration failed\n"); - goto err_both; + goto err_input; } msc->input = input; return 0; - err_both: +err_input: input_free_device(input); - err_free: +err_stop_hw: + hid_hw_stop(hdev); +err_free: kfree(msc); return ret; } -- cgit v0.10.2 From c7c6b1fe9f942c1a30585ec2210a09dfff238506 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 10 Feb 2010 15:43:04 +0800 Subject: ftrace: Allow to remove a single function from function graph filter I don't see why we can only clear all functions from the filter. After patching: # echo sys_open > set_graph_function # echo sys_close >> set_graph_function # cat set_graph_function sys_open sys_close # echo '!sys_close' >> set_graph_function # cat set_graph_function sys_open Signed-off-by: Li Zefan LKML-Reference: <4B726388.2000408@cn.fujitsu.com> Signed-off-by: Steven Rostedt diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 7968762..43bec99 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -2426,6 +2426,7 @@ static const struct file_operations ftrace_notrace_fops = { static DEFINE_MUTEX(graph_lock); int ftrace_graph_count; +int ftrace_graph_filter_enabled; unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly; static void * @@ -2448,7 +2449,7 @@ static void *g_start(struct seq_file *m, loff_t *pos) mutex_lock(&graph_lock); /* Nothing, tell g_show to print all functions are enabled */ - if (!ftrace_graph_count && !*pos) + if (!ftrace_graph_filter_enabled && !*pos) return (void *)1; return __g_next(m, pos); @@ -2494,6 +2495,7 @@ ftrace_graph_open(struct inode *inode, struct file *file) mutex_lock(&graph_lock); if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC)) { + ftrace_graph_filter_enabled = 0; ftrace_graph_count = 0; memset(ftrace_graph_funcs, 0, sizeof(ftrace_graph_funcs)); } @@ -2519,7 +2521,7 @@ ftrace_set_func(unsigned long *array, int *idx, char *buffer) struct dyn_ftrace *rec; struct ftrace_page *pg; int search_len; - int found = 0; + int fail = 1; int type, not; char *search; bool exists; @@ -2530,37 +2532,51 @@ ftrace_set_func(unsigned long *array, int *idx, char *buffer) /* decode regex */ type = filter_parse_regex(buffer, strlen(buffer), &search, ¬); - if (not) - return -EINVAL; + if (!not && *idx >= FTRACE_GRAPH_MAX_FUNCS) + return -EBUSY; search_len = strlen(search); mutex_lock(&ftrace_lock); do_for_each_ftrace_rec(pg, rec) { - if (*idx >= FTRACE_GRAPH_MAX_FUNCS) - break; - if (rec->flags & (FTRACE_FL_FAILED | FTRACE_FL_FREE)) continue; if (ftrace_match_record(rec, search, search_len, type)) { - /* ensure it is not already in the array */ + /* if it is in the array */ exists = false; - for (i = 0; i < *idx; i++) + for (i = 0; i < *idx; i++) { if (array[i] == rec->ip) { exists = true; break; } - if (!exists) - array[(*idx)++] = rec->ip; - found = 1; + } + + if (!not) { + fail = 0; + if (!exists) { + array[(*idx)++] = rec->ip; + if (*idx >= FTRACE_GRAPH_MAX_FUNCS) + goto out; + } + } else { + if (exists) { + array[i] = array[--(*idx)]; + array[*idx] = 0; + fail = 0; + } + } } } while_for_each_ftrace_rec(); - +out: mutex_unlock(&ftrace_lock); - return found ? 0 : -EINVAL; + if (fail) + return -EINVAL; + + ftrace_graph_filter_enabled = 1; + return 0; } static ssize_t @@ -2570,16 +2586,11 @@ ftrace_graph_write(struct file *file, const char __user *ubuf, struct trace_parser parser; ssize_t read, ret; - if (!cnt || cnt < 0) + if (!cnt) return 0; mutex_lock(&graph_lock); - if (ftrace_graph_count >= FTRACE_GRAPH_MAX_FUNCS) { - ret = -EBUSY; - goto out_unlock; - } - if (trace_parser_get_init(&parser, FTRACE_BUFF_MAX)) { ret = -ENOMEM; goto out_unlock; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index ce077fb..b477fce 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -497,6 +497,7 @@ trace_print_graph_duration(unsigned long long duration, struct trace_seq *s); #ifdef CONFIG_DYNAMIC_FTRACE /* TODO: make this variable */ #define FTRACE_GRAPH_MAX_FUNCS 32 +extern int ftrace_graph_filter_enabled; extern int ftrace_graph_count; extern unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS]; @@ -504,7 +505,7 @@ static inline int ftrace_graph_addr(unsigned long addr) { int i; - if (!ftrace_graph_count) + if (!ftrace_graph_filter_enabled) return 1; for (i = 0; i < ftrace_graph_count; i++) { -- cgit v0.10.2 From 5b3efd500854d45d305b53c54c97db5970959980 Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Thu, 11 Feb 2010 11:50:59 -0800 Subject: x86, ptrace: regset extensions to support xstate Add the xstate regset support which helps extend the kernel ptrace and the core-dump interfaces to support AVX state etc. This regset interface is designed to support all the future state that gets supported using xsave/xrstor infrastructure. Looking at the memory layout saved by "xsave", one can't say which state is represented in the memory layout. This is because if a particular state is in init state, in the xsave hdr it can be represented by bit '0'. And hence we can't really say by the xsave header wether a state is in init state or the state is not saved in the memory layout. And hence the xsave memory layout available through this regset interface uses SW usable bytes [464..511] to convey what state is represented in the memory layout. First 8 bytes of the sw_usable_bytes[464..467] will be set to OS enabled xstate mask(which is same as the 64bit mask returned by the xgetbv's xCR0). The note NT_X86_XSTATE represents the extended state information in the core file, using the above mentioned memory layout. Signed-off-by: Suresh Siddha LKML-Reference: <20100211195614.802495327@sbs-t61.sc.intel.com> Signed-off-by: Hongjiu Lu Cc: Roland McGrath Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h index ebfb8a9..da29309 100644 --- a/arch/x86/include/asm/i387.h +++ b/arch/x86/include/asm/i387.h @@ -33,8 +33,16 @@ extern void init_thread_xstate(void); extern int dump_fpu(struct pt_regs *, struct user_i387_struct *); extern user_regset_active_fn fpregs_active, xfpregs_active; -extern user_regset_get_fn fpregs_get, xfpregs_get, fpregs_soft_get; -extern user_regset_set_fn fpregs_set, xfpregs_set, fpregs_soft_set; +extern user_regset_get_fn fpregs_get, xfpregs_get, fpregs_soft_get, + xstateregs_get; +extern user_regset_set_fn fpregs_set, xfpregs_set, fpregs_soft_set, + xstateregs_set; + +/* + * xstateregs_active == fpregs_active. Please refer to the comment + * at the definition of fpregs_active. + */ +#define xstateregs_active fpregs_active extern struct _fpx_sw_bytes fx_sw_reserved; #ifdef CONFIG_IA32_EMULATION diff --git a/arch/x86/include/asm/user.h b/arch/x86/include/asm/user.h index 999873b..24532c7 100644 --- a/arch/x86/include/asm/user.h +++ b/arch/x86/include/asm/user.h @@ -1,5 +1,63 @@ +#ifndef _ASM_X86_USER_H +#define _ASM_X86_USER_H + #ifdef CONFIG_X86_32 # include "user_32.h" #else # include "user_64.h" #endif + +#include + +struct user_ymmh_regs { + /* 16 * 16 bytes for each YMMH-reg */ + __u32 ymmh_space[64]; +}; + +struct user_xsave_hdr { + __u64 xstate_bv; + __u64 reserved1[2]; + __u64 reserved2[5]; +}; + +/* + * The structure layout of user_xstateregs, used for exporting the + * extended register state through ptrace and core-dump (NT_X86_XSTATE note) + * interfaces will be same as the memory layout of xsave used by the processor + * (except for the bytes 464..511, which can be used by the software) and hence + * the size of this structure varies depending on the features supported by the + * processor and OS. The size of the structure that users need to use can be + * obtained by doing: + * cpuid_count(0xd, 0, &eax, &ptrace_xstateregs_struct_size, &ecx, &edx); + * i.e., cpuid.(eax=0xd,ecx=0).ebx will be the size that user (debuggers, etc.) + * need to use. + * + * For now, only the first 8 bytes of the software usable bytes[464..471] will + * be used and will be set to OS enabled xstate mask (which is same as the + * 64bit mask returned by the xgetbv's xCR0). Users (analyzing core dump + * remotely, etc.) can use this mask as well as the mask saved in the + * xstate_hdr bytes and interpret what states the processor/OS supports + * and what states are in modified/initialized conditions for the + * particular process/thread. + * + * Also when the user modifies certain state FP/SSE/etc through the + * ptrace interface, they must ensure that the xsave_hdr.xstate_bv + * bytes[512..519] of the memory layout are updated correspondingly. + * i.e., for example when FP state is modified to a non-init state, + * xsave_hdr.xstate_bv's bit 0 must be set to '1', when SSE is modified to + * non-init state, xsave_hdr.xstate_bv's bit 1 must to be set to '1', etc. + */ +#define USER_XSTATE_FX_SW_WORDS 6 +#define USER_XSTATE_XCR0_WORD 0 + +struct user_xstateregs { + struct { + __u64 fpx_space[58]; + __u64 xstate_fx_sw[USER_XSTATE_FX_SW_WORDS]; + } i387; + struct user_xsave_hdr xsave_hdr; + struct user_ymmh_regs ymmh; + /* further processor state extensions go here */ +}; + +#endif /* _ASM_X86_USER_H */ diff --git a/arch/x86/include/asm/xsave.h b/arch/x86/include/asm/xsave.h index 727acc1..ddc04cc 100644 --- a/arch/x86/include/asm/xsave.h +++ b/arch/x86/include/asm/xsave.h @@ -27,9 +27,11 @@ extern unsigned int xstate_size; extern u64 pcntxt_mask; extern struct xsave_struct *init_xstate_buf; +extern u64 xstate_fx_sw_bytes[USER_XSTATE_FX_SW_WORDS]; extern void xsave_cntxt_init(void); extern void xsave_init(void); +extern void update_regset_xstate_info(unsigned int size, u64 xstate_mask); extern int init_fpu(struct task_struct *child); extern int check_for_xstate(struct i387_fxsave_struct __user *buf, void __user *fpstate, diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c index f2f8540..7a8a193 100644 --- a/arch/x86/kernel/i387.c +++ b/arch/x86/kernel/i387.c @@ -164,6 +164,11 @@ int init_fpu(struct task_struct *tsk) return 0; } +/* + * The xstateregs_active() routine is the same as the fpregs_active() routine, + * as the "regset->n" for the xstate regset will be updated based on the feature + * capabilites supported by the xsave. + */ int fpregs_active(struct task_struct *target, const struct user_regset *regset) { return tsk_used_math(target) ? regset->n : 0; @@ -224,6 +229,84 @@ int xfpregs_set(struct task_struct *target, const struct user_regset *regset, return ret; } +int xstateregs_get(struct task_struct *target, const struct user_regset *regset, + unsigned int pos, unsigned int count, + void *kbuf, void __user *ubuf) +{ + int ret; + + if (!cpu_has_xsave) + return -ENODEV; + + ret = init_fpu(target); + if (ret) + return ret; + + /* + * First copy the fxsave bytes 0..463. + */ + ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, + &target->thread.xstate->xsave, 0, + offsetof(struct user_xstateregs, + i387.xstate_fx_sw)); + if (ret) + return ret; + + /* + * Copy the 48bytes defined by software. + */ + ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, + xstate_fx_sw_bytes, + offsetof(struct user_xstateregs, + i387.xstate_fx_sw), + offsetof(struct user_xstateregs, + xsave_hdr)); + if (ret) + return ret; + + /* + * Copy the rest of xstate memory layout. + */ + ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, + &target->thread.xstate->xsave.xsave_hdr, + offsetof(struct user_xstateregs, + xsave_hdr), -1); + return ret; +} + +int xstateregs_set(struct task_struct *target, const struct user_regset *regset, + unsigned int pos, unsigned int count, + const void *kbuf, const void __user *ubuf) +{ + int ret; + struct xsave_hdr_struct *xsave_hdr; + + if (!cpu_has_xsave) + return -ENODEV; + + ret = init_fpu(target); + if (ret) + return ret; + + ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf, + &target->thread.xstate->xsave, 0, -1); + + /* + * mxcsr reserved bits must be masked to zero for security reasons. + */ + target->thread.xstate->fxsave.mxcsr &= mxcsr_feature_mask; + + xsave_hdr = &target->thread.xstate->xsave.xsave_hdr; + + xsave_hdr->xstate_bv &= pcntxt_mask; + /* + * These bits must be zero. + */ + xsave_hdr->reserved1[0] = xsave_hdr->reserved1[1] = 0; + + return ret; +} + #if defined CONFIG_X86_32 || defined CONFIG_IA32_EMULATION /* diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 017d937..16433a5 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -48,6 +48,7 @@ enum x86_regset { REGSET_FP, REGSET_XFP, REGSET_IOPERM64 = REGSET_XFP, + REGSET_XSTATE, REGSET_TLS, REGSET_IOPERM32, }; @@ -1584,7 +1585,7 @@ long compat_arch_ptrace(struct task_struct *child, compat_long_t request, #ifdef CONFIG_X86_64 -static const struct user_regset x86_64_regsets[] = { +static struct user_regset x86_64_regsets[] __read_mostly = { [REGSET_GENERAL] = { .core_note_type = NT_PRSTATUS, .n = sizeof(struct user_regs_struct) / sizeof(long), @@ -1597,6 +1598,12 @@ static const struct user_regset x86_64_regsets[] = { .size = sizeof(long), .align = sizeof(long), .active = xfpregs_active, .get = xfpregs_get, .set = xfpregs_set }, + [REGSET_XSTATE] = { + .core_note_type = NT_X86_XSTATE, + .size = sizeof(u64), .align = sizeof(u64), + .active = xstateregs_active, .get = xstateregs_get, + .set = xstateregs_set + }, [REGSET_IOPERM64] = { .core_note_type = NT_386_IOPERM, .n = IO_BITMAP_LONGS, @@ -1622,7 +1629,7 @@ static const struct user_regset_view user_x86_64_view = { #endif /* CONFIG_X86_64 */ #if defined CONFIG_X86_32 || defined CONFIG_IA32_EMULATION -static const struct user_regset x86_32_regsets[] = { +static struct user_regset x86_32_regsets[] __read_mostly = { [REGSET_GENERAL] = { .core_note_type = NT_PRSTATUS, .n = sizeof(struct user_regs_struct32) / sizeof(u32), @@ -1641,6 +1648,12 @@ static const struct user_regset x86_32_regsets[] = { .size = sizeof(u32), .align = sizeof(u32), .active = xfpregs_active, .get = xfpregs_get, .set = xfpregs_set }, + [REGSET_XSTATE] = { + .core_note_type = NT_X86_XSTATE, + .size = sizeof(u64), .align = sizeof(u64), + .active = xstateregs_active, .get = xstateregs_get, + .set = xstateregs_set + }, [REGSET_TLS] = { .core_note_type = NT_386_TLS, .n = GDT_ENTRY_TLS_ENTRIES, .bias = GDT_ENTRY_TLS_MIN, @@ -1663,6 +1676,23 @@ static const struct user_regset_view user_x86_32_view = { }; #endif +/* + * This represents bytes 464..511 in the memory layout exported through + * the REGSET_XSTATE interface. + */ +u64 xstate_fx_sw_bytes[USER_XSTATE_FX_SW_WORDS]; + +void update_regset_xstate_info(unsigned int size, u64 xstate_mask) +{ +#ifdef CONFIG_X86_64 + x86_64_regsets[REGSET_XSTATE].n = size / sizeof(u64); +#endif +#if defined CONFIG_X86_32 || defined CONFIG_IA32_EMULATION + x86_32_regsets[REGSET_XSTATE].n = size / sizeof(u64); +#endif + xstate_fx_sw_bytes[USER_XSTATE_XCR0_WORD] = xstate_mask; +} + const struct user_regset_view *task_user_regset_view(struct task_struct *task) { #ifdef CONFIG_IA32_EMULATION diff --git a/arch/x86/kernel/xsave.c b/arch/x86/kernel/xsave.c index c5ee17e..782c3a36 100644 --- a/arch/x86/kernel/xsave.c +++ b/arch/x86/kernel/xsave.c @@ -337,6 +337,7 @@ void __ref xsave_cntxt_init(void) cpuid_count(0xd, 0, &eax, &ebx, &ecx, &edx); xstate_size = ebx; + update_regset_xstate_info(xstate_size, pcntxt_mask); prepare_fx_sw_frame(); setup_xstate_init(); diff --git a/include/linux/elf.h b/include/linux/elf.h index 0cc4d55..a8c4af0 100644 --- a/include/linux/elf.h +++ b/include/linux/elf.h @@ -361,6 +361,7 @@ typedef struct elf64_shdr { #define NT_PPC_VSX 0x102 /* PowerPC VSX registers */ #define NT_386_TLS 0x200 /* i386 TLS slots (struct user_desc) */ #define NT_386_IOPERM 0x201 /* x86 io permission bitmap (1=deny) */ +#define NT_X86_XSTATE 0x202 /* x86 extended state using xsave */ #define NT_S390_HIGH_GPRS 0x300 /* s390 upper register halves */ -- cgit v0.10.2 From 2225a122ae26d542bdce523d9d87a4a7ba10e07b Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Thu, 11 Feb 2010 11:51:00 -0800 Subject: ptrace: Add support for generic PTRACE_GETREGSET/PTRACE_SETREGSET Generic support for PTRACE_GETREGSET/PTRACE_SETREGSET commands which export the regsets supported by each architecture using the correponding NT_* types. These NT_* types are already part of the userland ABI, used in representing the architecture specific register sets as different NOTES in an ELF core file. 'addr' parameter for the ptrace system call encode the REGSET type (using the corresppnding NT_* type) and the 'data' parameter points to the struct iovec having the user buffer and the length of that buffer. struct iovec iov = { buf, len}; ret = ptrace(PTRACE_GETREGSET/PTRACE_SETREGSET, pid, NT_XXX_TYPE, &iov); On successful completion, iov.len will be updated by the kernel specifying how much the kernel has written/read to/from the user's iov.buf. x86 extended state registers are primarily exported using this interface. Signed-off-by: Suresh Siddha LKML-Reference: <20100211195614.886724710@sbs-t61.sc.intel.com> Acked-by: Hongjiu Lu Cc: Roland McGrath Signed-off-by: H. Peter Anvin diff --git a/include/linux/elf.h b/include/linux/elf.h index a8c4af0..d8e6e61 100644 --- a/include/linux/elf.h +++ b/include/linux/elf.h @@ -349,7 +349,11 @@ typedef struct elf64_shdr { #define ELF_OSABI ELFOSABI_NONE #endif -/* Notes used in ET_CORE */ +/* + * Notes used in ET_CORE. Architectures export some of the arch register sets + * using the corresponding note types via the PTRACE_GETREGSET and + * PTRACE_SETREGSET requests. + */ #define NT_PRSTATUS 1 #define NT_PRFPREG 2 #define NT_PRPSINFO 3 diff --git a/include/linux/ptrace.h b/include/linux/ptrace.h index 56f2d63..dbfa821 100644 --- a/include/linux/ptrace.h +++ b/include/linux/ptrace.h @@ -27,6 +27,21 @@ #define PTRACE_GETSIGINFO 0x4202 #define PTRACE_SETSIGINFO 0x4203 +/* + * Generic ptrace interface that exports the architecture specific regsets + * using the corresponding NT_* types (which are also used in the core dump). + * + * This interface usage is as follows: + * struct iovec iov = { buf, len}; + * + * ret = ptrace(PTRACE_GETREGSET/PTRACE_SETREGSET, pid, NT_XXX_TYPE, &iov); + * + * On the successful completion, iov.len will be updated by the kernel, + * specifying how much the kernel has written/read to/from the user's iov.buf. + */ +#define PTRACE_GETREGSET 0x4204 +#define PTRACE_SETREGSET 0x4205 + /* options set using PTRACE_SETOPTIONS */ #define PTRACE_O_TRACESYSGOOD 0x00000001 #define PTRACE_O_TRACEFORK 0x00000002 diff --git a/kernel/ptrace.c b/kernel/ptrace.c index 23bd09c..13b4554 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -22,6 +22,7 @@ #include #include #include +#include /* @@ -511,6 +512,47 @@ static int ptrace_resume(struct task_struct *child, long request, long data) return 0; } +#ifdef CONFIG_HAVE_ARCH_TRACEHOOK + +static const struct user_regset * +find_regset(const struct user_regset_view *view, unsigned int type) +{ + const struct user_regset *regset; + int n; + + for (n = 0; n < view->n; ++n) { + regset = view->regsets + n; + if (regset->core_note_type == type) + return regset; + } + + return NULL; +} + +static int ptrace_regset(struct task_struct *task, int req, unsigned int type, + struct iovec *kiov) +{ + const struct user_regset_view *view = task_user_regset_view(task); + const struct user_regset *regset = find_regset(view, type); + int regset_no; + + if (!regset || (kiov->iov_len % regset->size) != 0) + return -EIO; + + regset_no = regset - view->regsets; + kiov->iov_len = min(kiov->iov_len, + (__kernel_size_t) (regset->n * regset->size)); + + if (req == PTRACE_GETREGSET) + return copy_regset_to_user(task, view, regset_no, 0, + kiov->iov_len, kiov->iov_base); + else + return copy_regset_from_user(task, view, regset_no, 0, + kiov->iov_len, kiov->iov_base); +} + +#endif + int ptrace_request(struct task_struct *child, long request, long addr, long data) { @@ -573,6 +615,26 @@ int ptrace_request(struct task_struct *child, long request, return 0; return ptrace_resume(child, request, SIGKILL); +#ifdef CONFIG_HAVE_ARCH_TRACEHOOK + case PTRACE_GETREGSET: + case PTRACE_SETREGSET: + { + struct iovec kiov; + struct iovec __user *uiov = (struct iovec __user *) data; + + if (!access_ok(VERIFY_WRITE, uiov, sizeof(*uiov))) + return -EFAULT; + + if (__get_user(kiov.iov_base, &uiov->iov_base) || + __get_user(kiov.iov_len, &uiov->iov_len)) + return -EFAULT; + + ret = ptrace_regset(child, request, addr, &kiov); + if (!ret) + ret = __put_user(kiov.iov_len, &uiov->iov_len); + break; + } +#endif default: break; } @@ -711,6 +773,32 @@ int compat_ptrace_request(struct task_struct *child, compat_long_t request, else ret = ptrace_setsiginfo(child, &siginfo); break; +#ifdef CONFIG_HAVE_ARCH_TRACEHOOK + case PTRACE_GETREGSET: + case PTRACE_SETREGSET: + { + struct iovec kiov; + struct compat_iovec __user *uiov = + (struct compat_iovec __user *) datap; + compat_uptr_t ptr; + compat_size_t len; + + if (!access_ok(VERIFY_WRITE, uiov, sizeof(*uiov))) + return -EFAULT; + + if (__get_user(ptr, &uiov->iov_base) || + __get_user(len, &uiov->iov_len)) + return -EFAULT; + + kiov.iov_base = compat_ptr(ptr); + kiov.iov_len = len; + + ret = ptrace_regset(child, request, addr, &kiov); + if (!ret) + ret = __put_user(kiov.iov_len, &uiov->iov_len); + break; + } +#endif default: ret = ptrace_request(child, request, addr, data); -- cgit v0.10.2 From b0f3ae03aca0f331b851ae94bc066124e7f104df Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 12 Feb 2010 15:40:00 +0900 Subject: sh: Isolate uncached mapping support. This splits out the uncached mapping support under its own config option, presently only used by 29-bit mode and 32-bit + PMB. This will make it possible to optionally add an uncached mapping on sh64 as well as booting without an uncached mapping for 32-bit. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/ptrace.h b/arch/sh/include/asm/ptrace.h index 201d11e..e879dff 100644 --- a/arch/sh/include/asm/ptrace.h +++ b/arch/sh/include/asm/ptrace.h @@ -102,13 +102,15 @@ struct pt_dspregs { #define PTRACE_GETDSPREGS 55 /* DSP registers */ #define PTRACE_SETDSPREGS 56 -#define PT_TEXT_END_ADDR 240 -#define PT_TEXT_ADDR 244 /* &(struct user)->start_code */ -#define PT_DATA_ADDR 248 /* &(struct user)->start_data */ +#define PT_TEXT_END_ADDR 240 +#define PT_TEXT_ADDR 244 /* &(struct user)->start_code */ +#define PT_DATA_ADDR 248 /* &(struct user)->start_data */ #define PT_TEXT_LEN 252 #ifdef __KERNEL__ #include +#include +#include #define user_mode(regs) (((regs)->sr & 0x40000000)==0) #define instruction_pointer(regs) ((unsigned long)(regs)->pc) @@ -137,9 +139,14 @@ static inline unsigned long profile_pc(struct pt_regs *regs) { unsigned long pc = instruction_pointer(regs); -#ifdef P2SEG - if (pc >= P2SEG && pc < P3SEG) - pc -= 0x20000000; +#ifdef CONFIG_UNCACHED_MAPPING + /* + * If PC points in to the uncached mapping, fix it up and hand + * back the cached equivalent. + */ + if ((pc >= (memory_start + cached_to_uncached)) && + (pc < (memory_start + cached_to_uncached + uncached_size))) + pc -= cached_to_uncached; #endif return pc; diff --git a/arch/sh/include/asm/system.h b/arch/sh/include/asm/system.h index 6442f17..0bd7a17 100644 --- a/arch/sh/include/asm/system.h +++ b/arch/sh/include/asm/system.h @@ -10,7 +10,6 @@ #include #include #include -#include #define AT_VECTOR_SIZE_ARCH 5 /* entries in ARCH_DLINFO */ @@ -114,6 +113,8 @@ static inline unsigned long __cmpxchg(volatile void * ptr, unsigned long old, (unsigned long)_n_, sizeof(*(ptr))); \ }) +struct pt_regs; + extern void die(const char *str, struct pt_regs *regs, long err) __attribute__ ((noreturn)); void free_initmem(void); void free_initrd_mem(unsigned long start, unsigned long end); diff --git a/arch/sh/include/asm/system_64.h b/arch/sh/include/asm/system_64.h index 3391bb6..3633864 100644 --- a/arch/sh/include/asm/system_64.h +++ b/arch/sh/include/asm/system_64.h @@ -18,6 +18,7 @@ /* * switch_to() should switch tasks to task nr n, first */ +struct thread_struct; struct task_struct *sh64_switch_to(struct task_struct *prev, struct thread_struct *prev_thread, struct task_struct *next, diff --git a/arch/sh/mm/Kconfig b/arch/sh/mm/Kconfig index b890752..65cb5b8 100644 --- a/arch/sh/mm/Kconfig +++ b/arch/sh/mm/Kconfig @@ -75,6 +75,7 @@ config MEMORY_SIZE config 29BIT def_bool !32BIT depends on SUPERH32 + select UNCACHED_MAPPING config 32BIT bool @@ -84,6 +85,7 @@ config PMB bool "Support 32-bit physical addressing through PMB" depends on MMU && EXPERIMENTAL && CPU_SH4A && !CPU_SH4AL_DSP select 32BIT + select UNCACHED_MAPPING help If you say Y here, physical addressing will be extended to 32-bits through the SH-4A PMB. If this is not set, legacy @@ -173,6 +175,9 @@ config IOREMAP_FIXED def_bool y depends on X2TLB || SUPERH64 +config UNCACHED_MAPPING + bool + choice prompt "Kernel page size" default PAGE_SIZE_4KB diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c index dffa6c7..58012b6 100644 --- a/arch/sh/mm/init.c +++ b/arch/sh/mm/init.c @@ -21,11 +21,12 @@ #include #include #include +#include DEFINE_PER_CPU(struct mmu_gather, mmu_gathers); pgd_t swapper_pg_dir[PTRS_PER_PGD]; -#ifdef CONFIG_SUPERH32 +#ifdef CONFIG_UNCACHED_MAPPING /* * This is the offset of the uncached section from its cached alias. * @@ -36,8 +37,8 @@ pgd_t swapper_pg_dir[PTRS_PER_PGD]; * Default value only valid in 29 bit mode, in 32bit mode this will be * updated by the early PMB initialization code. */ -unsigned long cached_to_uncached = P2SEG - P1SEG; -unsigned long uncached_size = 0x20000000; +unsigned long cached_to_uncached = 0x20000000; +unsigned long uncached_size = SZ_512M; #endif #ifdef CONFIG_MMU @@ -281,7 +282,9 @@ void __init mem_init(void) #endif " vmalloc : 0x%08lx - 0x%08lx (%4ld MB)\n" " lowmem : 0x%08lx - 0x%08lx (%4ld MB) (cached)\n" +#ifdef CONFIG_UNCACHED_MAPPING " : 0x%08lx - 0x%08lx (%4ld MB) (uncached)\n" +#endif " .init : 0x%08lx - 0x%08lx (%4ld kB)\n" " .data : 0x%08lx - 0x%08lx (%4ld kB)\n" " .text : 0x%08lx - 0x%08lx (%4ld kB)\n", @@ -299,9 +302,11 @@ void __init mem_init(void) (unsigned long)memory_start, (unsigned long)high_memory, ((unsigned long)high_memory - (unsigned long)memory_start) >> 20, +#ifdef CONFIG_UNCACHED_MAPPING (unsigned long)memory_start + cached_to_uncached, (unsigned long)memory_start + cached_to_uncached + uncached_size, uncached_size >> 20, +#endif (unsigned long)&__init_begin, (unsigned long)&__init_end, ((unsigned long)&__init_end - -- cgit v0.10.2 From 19f6b8b44e3f633d5d7d1ed68848b1eb89a1e800 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Fri, 12 Feb 2010 15:41:45 +0900 Subject: sh64: fix up memory offset calculation. The linker script offsets were broken by the recent 29/32-bit integration, so this fixes it up for sh64. Signed-off-by: Paul Mundt diff --git a/arch/sh/include/asm/vmlinux.lds.h b/arch/sh/include/asm/vmlinux.lds.h index 244ec4a..d58ad49 100644 --- a/arch/sh/include/asm/vmlinux.lds.h +++ b/arch/sh/include/asm/vmlinux.lds.h @@ -14,4 +14,12 @@ #define DWARF_EH_FRAME #endif +#ifdef CONFIG_SUPERH64 +#define EXTRA_TEXT \ + *(.text64) \ + *(.text..SHmedia32) +#else +#define EXTRA_TEXT +#endif + #endif /* __ASM_SH_VMLINUX_LDS_H */ diff --git a/arch/sh/kernel/vmlinux.lds.S b/arch/sh/kernel/vmlinux.lds.S index f0bc6b8..dcf4891 100644 --- a/arch/sh/kernel/vmlinux.lds.S +++ b/arch/sh/kernel/vmlinux.lds.S @@ -3,10 +3,12 @@ * Written by Niibe Yutaka and Paul Mundt */ #ifdef CONFIG_SUPERH64 -#define LOAD_OFFSET CONFIG_PAGE_OFFSET +#define LOAD_OFFSET PAGE_OFFSET +#define MEMORY_OFFSET __MEMORY_START OUTPUT_ARCH(sh:sh5) #else #define LOAD_OFFSET 0 +#define MEMORY_OFFSET 0 OUTPUT_ARCH(sh) #endif @@ -14,16 +16,10 @@ OUTPUT_ARCH(sh) #include #include -#if defined(CONFIG_32BIT) && !defined(CONFIG_PMB_LEGACY) -#define MEMORY_OFFSET 0 -#else -#define MEMORY_OFFSET (CONFIG_MEMORY_START & 0x1fffffff) -#endif - ENTRY(_start) SECTIONS { - . = CONFIG_PAGE_OFFSET + MEMORY_OFFSET + CONFIG_ZERO_PAGE_OFFSET; + . = PAGE_OFFSET + MEMORY_OFFSET + CONFIG_ZERO_PAGE_OFFSET; _text = .; /* Text and read-only data */ @@ -34,12 +30,7 @@ SECTIONS .text : AT(ADDR(.text) - LOAD_OFFSET) { HEAD_TEXT TEXT_TEXT - -#ifdef CONFIG_SUPERH64 - *(.text64) - *(.text..SHmedia32) -#endif - + EXTRA_TEXT SCHED_TEXT LOCK_TEXT KPROBES_TEXT -- cgit v0.10.2 From 858155fbcc0cd713f6382c527bb1c3abc0ed6d00 Mon Sep 17 00:00:00 2001 From: Oliver Neukum Date: Fri, 12 Feb 2010 13:02:28 +0100 Subject: HID: usbhid: introduce timeout for stuck ctrl/out URBs Some devices do not react to a control request (seen on APC UPS's) resulting in a slow stream of messages, "generic-usb ... control queue full". Therefore request needs a timeout. Cc: stable@kernel.org Signed-off-by: Oliver Neukum Signed-off-by: David Fries Signed-off-by: Jiri Kosina diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c index 5406074..74bd3ca 100644 --- a/drivers/hid/usbhid/hid-core.c +++ b/drivers/hid/usbhid/hid-core.c @@ -316,6 +316,7 @@ static int hid_submit_out(struct hid_device *hid) err_hid("usb_submit_urb(out) failed"); return -1; } + usbhid->last_out = jiffies; } else { /* * queue work to wake up the device. @@ -377,6 +378,7 @@ static int hid_submit_ctrl(struct hid_device *hid) err_hid("usb_submit_urb(ctrl) failed"); return -1; } + usbhid->last_ctrl = jiffies; } else { /* * queue work to wake up the device. @@ -512,9 +514,20 @@ static void __usbhid_submit_report(struct hid_device *hid, struct hid_report *re usbhid->out[usbhid->outhead].report = report; usbhid->outhead = head; - if (!test_and_set_bit(HID_OUT_RUNNING, &usbhid->iofl)) + if (!test_and_set_bit(HID_OUT_RUNNING, &usbhid->iofl)) { if (hid_submit_out(hid)) clear_bit(HID_OUT_RUNNING, &usbhid->iofl); + } else { + /* + * the queue is known to run + * but an earlier request may be stuck + * we may need to time out + * no race because this is called under + * spinlock + */ + if (time_after(jiffies, usbhid->last_out + HZ * 5)) + usb_unlink_urb(usbhid->urbout); + } return; } @@ -535,9 +548,20 @@ static void __usbhid_submit_report(struct hid_device *hid, struct hid_report *re usbhid->ctrl[usbhid->ctrlhead].dir = dir; usbhid->ctrlhead = head; - if (!test_and_set_bit(HID_CTRL_RUNNING, &usbhid->iofl)) + if (!test_and_set_bit(HID_CTRL_RUNNING, &usbhid->iofl)) { if (hid_submit_ctrl(hid)) clear_bit(HID_CTRL_RUNNING, &usbhid->iofl); + } else { + /* + * the queue is known to run + * but an earlier request may be stuck + * we may need to time out + * no race because this is called under + * spinlock + */ + if (time_after(jiffies, usbhid->last_ctrl + HZ * 5)) + usb_unlink_urb(usbhid->urbctrl); + } } void usbhid_submit_report(struct hid_device *hid, struct hid_report *report, unsigned char dir) diff --git a/drivers/hid/usbhid/usbhid.h b/drivers/hid/usbhid/usbhid.h index 08f505c..ec20400 100644 --- a/drivers/hid/usbhid/usbhid.h +++ b/drivers/hid/usbhid/usbhid.h @@ -80,12 +80,14 @@ struct usbhid_device { unsigned char ctrlhead, ctrltail; /* Control fifo head & tail */ char *ctrlbuf; /* Control buffer */ dma_addr_t ctrlbuf_dma; /* Control buffer dma */ + unsigned long last_ctrl; /* record of last output for timeouts */ struct urb *urbout; /* Output URB */ struct hid_output_fifo out[HID_CONTROL_FIFO_SIZE]; /* Output pipe fifo */ unsigned char outhead, outtail; /* Output pipe fifo head & tail */ char *outbuf; /* Output buffer */ dma_addr_t outbuf_dma; /* Output buffer dma */ + unsigned long last_out; /* record of last output for timeouts */ spinlock_t lock; /* fifo spinlock */ unsigned long iofl; /* I/O flags (CTRL_RUNNING, OUT_RUNNING) */ -- cgit v0.10.2 From c7b66698122796623f1838c0ee6a8e12ec832468 Mon Sep 17 00:00:00 2001 From: Stefan Roese Date: Wed, 10 Feb 2010 03:54:59 +0000 Subject: powerpc/40x: Add support for PPC40x boards with > 512MB SDRAM This patch adds support for boards with more that 512MByte RAM. Currently only 512MB of memory are enabled in the DCCR/ICCR real-mode cache control registers. This patch now enables caching in real-mode for 2GByte. Signed-off-by: Stefan Roese Cc: Benjamin Herrenschmidt Cc: Josh Boyer Signed-off-by: Josh Boyer diff --git a/arch/powerpc/mm/40x_mmu.c b/arch/powerpc/mm/40x_mmu.c index 08dfa8e..65abfcf 100644 --- a/arch/powerpc/mm/40x_mmu.c +++ b/arch/powerpc/mm/40x_mmu.c @@ -84,8 +84,8 @@ void __init MMU_init_hw(void) * vectors and the kernel live in real-mode. */ - mtspr(SPRN_DCCR, 0xF0000000); /* 512 MB of data space at 0x0. */ - mtspr(SPRN_ICCR, 0xF0000000); /* 512 MB of instr. space at 0x0. */ + mtspr(SPRN_DCCR, 0xFFFF0000); /* 2GByte of data space at 0x0. */ + mtspr(SPRN_ICCR, 0xFFFF0000); /* 2GByte of instr. space at 0x0. */ } #define LARGE_PAGE_SIZE_16M (1<<24) -- cgit v0.10.2 From 87185517de81101da5afbc82cefdeed6eeaa38fb Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 3 Feb 2010 19:43:31 +0000 Subject: xfs: only clear the suid bit once in xfs_write file_remove_suid already calls into ->setattr to clear the suid and sgid bits if needed, no need to start a second transaction to do it ourselves. Note that xfs_write_clear_setuid issues a sync transaction while the path through ->setattr doesn't, but that is consistant with the other filesystems. Signed-off-by: Christoph Hellwig Reviewed-by: Alex Elder Signed-off-by: Alex Elder diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c index c80fa00d..eac6f80 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.c +++ b/fs/xfs/linux-2.6/xfs_lrw.c @@ -630,18 +630,9 @@ start: * by root. This keeps people from modifying setuid and * setgid binaries. */ - - if (((xip->i_d.di_mode & S_ISUID) || - ((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) == - (S_ISGID | S_IXGRP))) && - !capable(CAP_FSETID)) { - error = xfs_write_clear_setuid(xip); - if (likely(!error)) - error = -file_remove_suid(file); - if (unlikely(error)) { - goto out_unlock_internal; - } - } + error = -file_remove_suid(file); + if (unlikely(error)) + goto out_unlock_internal; /* We can write back this queue in page reclaim */ current->backing_dev_info = mapping->backing_dev_info; diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c index abb2c45..e336742 100644 --- a/fs/xfs/xfs_rw.c +++ b/fs/xfs/xfs_rw.c @@ -47,48 +47,6 @@ #include "xfs_trace.h" /* - * This is a subroutine for xfs_write() and other writers (xfs_ioctl) - * which clears the setuid and setgid bits when a file is written. - */ -int -xfs_write_clear_setuid( - xfs_inode_t *ip) -{ - xfs_mount_t *mp; - xfs_trans_t *tp; - int error; - - mp = ip->i_mount; - tp = xfs_trans_alloc(mp, XFS_TRANS_WRITEID); - if ((error = xfs_trans_reserve(tp, 0, - XFS_WRITEID_LOG_RES(mp), - 0, 0, 0))) { - xfs_trans_cancel(tp, 0); - return error; - } - xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - xfs_trans_ihold(tp, ip); - ip->i_d.di_mode &= ~S_ISUID; - - /* - * Note that we don't have to worry about mandatory - * file locking being disabled here because we only - * clear the S_ISGID bit if the Group execute bit is - * on, but if it was on then mandatory locking wouldn't - * have been enabled. - */ - if (ip->i_d.di_mode & S_IXGRP) { - ip->i_d.di_mode &= ~S_ISGID; - } - xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - xfs_trans_set_sync(tp); - error = xfs_trans_commit(tp, 0); - xfs_iunlock(ip, XFS_ILOCK_EXCL); - return 0; -} - -/* * Force a shutdown of the filesystem instantly while keeping * the filesystem consistent. We don't do an unmount here; just shutdown * the shop, make sure that absolutely nothing persistent happens to diff --git a/fs/xfs/xfs_rw.h b/fs/xfs/xfs_rw.h index a54c3b7..11c41ec 100644 --- a/fs/xfs/xfs_rw.h +++ b/fs/xfs/xfs_rw.h @@ -39,7 +39,6 @@ xfs_fsb_to_db(struct xfs_inode *ip, xfs_fsblock_t fsb) /* * Prototypes for functions in xfs_rw.c. */ -extern int xfs_write_clear_setuid(struct xfs_inode *ip); extern int xfs_read_buf(struct xfs_mount *mp, xfs_buftarg_t *btp, xfs_daddr_t blkno, int len, uint flags, struct xfs_buf **bpp); -- cgit v0.10.2 From 180040b89ee2aed88c0a0b1fcf7ada9a512b12e3 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 5 Feb 2010 09:57:55 +0000 Subject: xfs: optimize log flushing in xfs_fsync If we have a pinned inode it must have a log item attached to it. Usually that log item will have ili_last_lsn already set, in which case we only need to flush the log up to that LSN instead of doing a full log force. This gives speedups of about 5% in some fsync heavy workloads. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 43241e2..ddd2c5d 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -629,8 +629,14 @@ xfs_fsync( */ xfs_iunlock(ip, XFS_ILOCK_SHARED); if (xfs_ipincount(ip)) { - error = _xfs_log_force(ip->i_mount, XFS_LOG_SYNC, - &log_flushed); + if (ip->i_itemp->ili_last_lsn) { + error = _xfs_log_force_lsn(ip->i_mount, + ip->i_itemp->ili_last_lsn, + XFS_LOG_SYNC, &log_flushed); + } else { + error = _xfs_log_force(ip->i_mount, + XFS_LOG_SYNC, &log_flushed); + } } } else { /* -- cgit v0.10.2 From 232486e1e9f34889424b68ee6270440b554479a2 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 12 Feb 2010 12:03:45 -0800 Subject: sparc64: Tighten checks in kstack_valid(). The kernel stack pointer is invalid if it is not 16-byte aligned. Based upon a report by Meelis Roos Signed-off-by: David S. Miller diff --git a/arch/sparc/kernel/kstack.h b/arch/sparc/kernel/kstack.h index 4248d96..5247283 100644 --- a/arch/sparc/kernel/kstack.h +++ b/arch/sparc/kernel/kstack.h @@ -11,6 +11,10 @@ static inline bool kstack_valid(struct thread_info *tp, unsigned long sp) { unsigned long base = (unsigned long) tp; + /* Stack pointer must be 16-byte aligned. */ + if (sp & (16UL - 1)) + return false; + if (sp >= (base + sizeof(struct thread_info)) && sp <= (base + THREAD_SIZE - sizeof(struct sparc_stackf))) return true; -- cgit v0.10.2 From d67b1b03254c501fef371b0e5916c94a52bfc2c5 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Sat, 6 Feb 2010 08:45:15 +0000 Subject: fs/xfs: Correct NULL test Test the value that was just allocated rather than the previously tested one. A simplified version of the semantic match that finds this problem is as follows: (http://coccinelle.lip6.fr/) // @r@ expression *x; expression e; identifier l; @@ if (x == NULL || ...) { ... when forall return ...; } ... when != goto l; when != x = e when != &x *x == NULL // Signed-off-by: Julia Lawall Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c index 8699e51..417e61e 100644 --- a/fs/xfs/quota/xfs_qm.c +++ b/fs/xfs/quota/xfs_qm.c @@ -123,7 +123,7 @@ xfs_Gqm_init(void) goto out; gdqhash = kmem_zalloc_large(hsize); - if (!udqhash) + if (!gdqhash) goto out_free_udqhash; hsize /= sizeof(xfs_dqhash_t); -- cgit v0.10.2 From fa644298eb24ab05b32acf6cc0f2265b833280e1 Mon Sep 17 00:00:00 2001 From: Anton Vorontsov Date: Tue, 15 Dec 2009 12:58:09 +0000 Subject: powerpc/85xx: Fix oops during MSI driver probe on MPC85xxMDS boards MPC85xx chips report the wrong value in feature reporting register, and that causes the following oops: Unable to handle kernel paging request for data at address 0x00000c00 Faulting instruction address: 0xc0019294 Oops: Kernel access of bad area, sig: 11 [#1] MPC8569 MDS Modules linked in: [...] NIP [c0019294] mpic_set_irq_type+0x2f0/0x368 LR [c0019124] mpic_set_irq_type+0x180/0x368 Call Trace: [ef851d60] [c0019124] mpic_set_irq_type+0x180/0x368 (unreliable) [ef851d90] [c007958c] __irq_set_trigger+0x44/0xd4 [ef851db0] [c007b550] set_irq_type+0x40/0x7c [ef851dc0] [c0004a60] irq_create_of_mapping+0xb4/0x114 [ef851df0] [c0004af0] irq_of_parse_and_map+0x30/0x40 [ef851e20] [c0405678] fsl_of_msi_probe+0x1a0/0x328 [ef851e60] [c02e6438] of_platform_device_probe+0x5c/0x84 [...] This is because mpic_alloc() assigns wrong values to mpic->isu_{size,shift,mask}, and things eventually break when _mpic_irq_read() is trying to use them. This patch fixes the issue by enabling MPIC_BROKEN_FRR_NIRQS quirk. Signed-off-by: Anton Vorontsov Signed-off-by: Kumar Gala diff --git a/arch/powerpc/platforms/85xx/mpc85xx_mds.c b/arch/powerpc/platforms/85xx/mpc85xx_mds.c index 21f61b8..cc29c0f 100644 --- a/arch/powerpc/platforms/85xx/mpc85xx_mds.c +++ b/arch/powerpc/platforms/85xx/mpc85xx_mds.c @@ -338,7 +338,8 @@ static void __init mpc85xx_mds_pic_init(void) } mpic = mpic_alloc(np, r.start, - MPIC_PRIMARY | MPIC_WANTS_RESET | MPIC_BIG_ENDIAN, + MPIC_PRIMARY | MPIC_WANTS_RESET | MPIC_BIG_ENDIAN | + MPIC_BROKEN_FRR_NIRQS, 0, 256, " OpenPIC "); BUG_ON(mpic == NULL); of_node_put(np); -- cgit v0.10.2 From d1d47ec6e62ab08d2ebb925fd9203abfad3adfbf Mon Sep 17 00:00:00 2001 From: Peter Tyser Date: Fri, 18 Dec 2009 16:50:37 -0600 Subject: powerpc/85xx: Fix SMP when "cpu-release-addr" is in lowmem Recent U-Boot commit 5ccd29c3679b3669b0bde5c501c1aa0f325a7acb caused the "cpu-release-addr" device tree property to contain the physical RAM location that secondary cores were spinning at. Previously, the "cpu-release-addr" property contained a value referencing the boot page translation address range of 0xfffffxxx, which then indirectly accessed RAM. The "cpu-release-addr" is currently ioremapped and the secondary cores kicked. However, due to the recent change in "cpu-release-addr", it sometimes points to a memory location in low memory that cannot be ioremapped. For example on a P2020-based board with 512MB of RAM the following error occurs on bootup: <...> mpic: requesting IPIs ... __ioremap(): phys addr 0x1ffff000 is RAM lr c05df9a0 Unable to handle kernel paging request for data at address 0x00000014 Faulting instruction address: 0xc05df9b0 Oops: Kernel access of bad area, sig: 11 [#1] SMP NR_CPUS=2 P2020 RDB Modules linked in: <... eventual kernel panic> Adding logic to conditionally ioremap or access memory directly resolves the issue. Signed-off-by: Peter Tyser Signed-off-by: Nate Case Reported-by: Dipen Dudhat Tested-by: Dipen Dudhat Signed-off-by: Kumar Gala diff --git a/arch/powerpc/platforms/85xx/smp.c b/arch/powerpc/platforms/85xx/smp.c index 04160a4..a15f582 100644 --- a/arch/powerpc/platforms/85xx/smp.c +++ b/arch/powerpc/platforms/85xx/smp.c @@ -46,6 +46,7 @@ smp_85xx_kick_cpu(int nr) __iomem u32 *bptr_vaddr; struct device_node *np; int n = 0; + int ioremappable; WARN_ON (nr < 0 || nr >= NR_CPUS); @@ -59,21 +60,37 @@ smp_85xx_kick_cpu(int nr) return; } + /* + * A secondary core could be in a spinloop in the bootpage + * (0xfffff000), somewhere in highmem, or somewhere in lowmem. + * The bootpage and highmem can be accessed via ioremap(), but + * we need to directly access the spinloop if its in lowmem. + */ + ioremappable = *cpu_rel_addr > virt_to_phys(high_memory); + /* Map the spin table */ - bptr_vaddr = ioremap(*cpu_rel_addr, SIZE_BOOT_ENTRY); + if (ioremappable) + bptr_vaddr = ioremap(*cpu_rel_addr, SIZE_BOOT_ENTRY); + else + bptr_vaddr = phys_to_virt(*cpu_rel_addr); local_irq_save(flags); out_be32(bptr_vaddr + BOOT_ENTRY_PIR, nr); out_be32(bptr_vaddr + BOOT_ENTRY_ADDR_LOWER, __pa(__early_start)); + if (!ioremappable) + flush_dcache_range((ulong)bptr_vaddr, + (ulong)(bptr_vaddr + SIZE_BOOT_ENTRY)); + /* Wait a bit for the CPU to ack. */ while ((__secondary_hold_acknowledge != nr) && (++n < 1000)) mdelay(1); local_irq_restore(flags); - iounmap(bptr_vaddr); + if (ioremappable) + iounmap(bptr_vaddr); pr_debug("waited %d msecs for CPU #%d.\n", n, nr); } -- cgit v0.10.2 From 0d1622d7f526311d87d7da2ee7dd14b73e45d3fc Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Sat, 13 Feb 2010 10:33:12 +0200 Subject: x86-64, rwsem: Avoid store forwarding hazard in __downgrade_write The Intel Architecture Optimization Reference Manual states that a short load that follows a long store to the same object will suffer a store forwading penalty, particularly if the two accesses use different addresses. Trivially, a long load that follows a short store will also suffer a penalty. __downgrade_write() in rwsem incurs both penalties: the increment operation will not be able to reuse a recently-loaded rwsem value, and its result will not be reused by any recently-following rwsem operation. A comment in the code states that this is because 64-bit immediates are special and expensive; but while they are slightly special (only a single instruction allows them), they aren't expensive: a test shows that two loops, one loading a 32-bit immediate and one loading a 64-bit immediate, both take 1.5 cycles per iteration. Fix this by changing __downgrade_write to use the same add instruction on i386 and on x86_64, so that it uses the same operand size as all the other rwsem functions. Signed-off-by: Avi Kivity LKML-Reference: <1266049992-17419-1-git-send-email-avi@redhat.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/rwsem.h b/arch/x86/include/asm/rwsem.h index 10204a2..606ede1 100644 --- a/arch/x86/include/asm/rwsem.h +++ b/arch/x86/include/asm/rwsem.h @@ -232,34 +232,19 @@ static inline void __up_write(struct rw_semaphore *sem) */ static inline void __downgrade_write(struct rw_semaphore *sem) { -#ifdef CONFIG_X86_64 -# if RWSEM_WAITING_BIAS != -0x100000000 -# error "This code assumes RWSEM_WAITING_BIAS == -2^32" -# endif - - /* 64-bit immediates are special and expensive, and not needed here */ - asm volatile("# beginning __downgrade_write\n\t" - LOCK_PREFIX "incl 4(%1)\n\t" - /* transitions 0xZZZZZZZZ00000001 -> 0xYYYYYYYY00000001 */ - " jns 1f\n\t" - " call call_rwsem_downgrade_wake\n" - "1:\n\t" - "# ending __downgrade_write\n" - : "+m" (sem->count) - : "a" (sem) - : "memory", "cc"); -#else asm volatile("# beginning __downgrade_write\n\t" LOCK_PREFIX _ASM_ADD "%2,(%1)\n\t" - /* transitions 0xZZZZ0001 -> 0xYYYY0001 */ + /* + * transitions 0xZZZZ0001 -> 0xYYYY0001 (i386) + * 0xZZZZZZZZ00000001 -> 0xYYYYYYYY00000001 (x86_64) + */ " jns 1f\n\t" " call call_rwsem_downgrade_wake\n" "1:\n\t" "# ending __downgrade_write\n" : "+m" (sem->count) - : "a" (sem), "i" (-RWSEM_WAITING_BIAS) + : "a" (sem), "er" (-RWSEM_WAITING_BIAS) : "memory", "cc"); -#endif } /* -- cgit v0.10.2 From 414bb144efa2d2fe16d104d836d0d6b6e9265788 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Mon, 14 Dec 2009 13:08:41 +0100 Subject: x86, cpu: Print AMD virtualization features in /proc/cpuinfo This patch adds code to cpu initialization path to detect the extended virtualization features of AMD cpus to show them in /proc/cpuinfo. Signed-off-by: Joerg Roedel LKML-Reference: <1260792521-15212-1-git-send-email-joerg.roedel@amd.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index 637e1ec..0cd82d0 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -168,6 +168,10 @@ #define X86_FEATURE_FLEXPRIORITY (8*32+ 2) /* Intel FlexPriority */ #define X86_FEATURE_EPT (8*32+ 3) /* Intel Extended Page Table */ #define X86_FEATURE_VPID (8*32+ 4) /* Intel Virtual Processor ID */ +#define X86_FEATURE_NPT (8*32+5) /* AMD Nested Page Table support */ +#define X86_FEATURE_LBRV (8*32+6) /* AMD LBR Virtualization support */ +#define X86_FEATURE_SVML (8*32+7) /* "svm_lock" AMD SVM locking MSR */ +#define X86_FEATURE_NRIPS (8*32+8) /* "nrip_save" AMD SVM next_rip save */ #if defined(__KERNEL__) && !defined(__ASSEMBLY__) diff --git a/arch/x86/kernel/cpu/addon_cpuid_features.c b/arch/x86/kernel/cpu/addon_cpuid_features.c index 468489b..97ad79c 100644 --- a/arch/x86/kernel/cpu/addon_cpuid_features.c +++ b/arch/x86/kernel/cpu/addon_cpuid_features.c @@ -32,6 +32,10 @@ void __cpuinit init_scattered_cpuid_features(struct cpuinfo_x86 *c) static const struct cpuid_bit __cpuinitconst cpuid_bits[] = { { X86_FEATURE_IDA, CR_EAX, 1, 0x00000006 }, { X86_FEATURE_ARAT, CR_EAX, 2, 0x00000006 }, + { X86_FEATURE_NPT, CR_EDX, 0, 0x8000000a }, + { X86_FEATURE_LBRV, CR_EDX, 1, 0x8000000a }, + { X86_FEATURE_SVML, CR_EDX, 2, 0x8000000a }, + { X86_FEATURE_NRIPS, CR_EDX, 3, 0x8000000a }, { 0, 0, 0, 0 } }; -- cgit v0.10.2 From 21b082ecdd7e6b8a5eba2cc013cae41b24de7f51 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Sun, 14 Feb 2010 07:13:38 -0700 Subject: of: Remove old and misplaced function declarations The following functions don't exist: finish_device_tree() print_properties() prom_n_intr_cells() prom_get_irq_senses() The following functions are in drivers/of/base.c, so the declaration belongs in of.h instead of of_fdt.h of_machine_is_compatible() prom_add_property() prom_remove_property() prom_update_property() Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Acked-by: Michal Simek diff --git a/include/linux/of.h b/include/linux/of.h index fa5571b..5c7b6a6 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -180,6 +180,14 @@ extern int of_parse_phandles_with_args(struct device_node *np, const char *list_name, const char *cells_name, int index, struct device_node **out_node, const void **out_args); +extern int of_machine_is_compatible(const char *compat); + +extern int prom_add_property(struct device_node* np, struct property* prop); +extern int prom_remove_property(struct device_node *np, struct property *prop); +extern int prom_update_property(struct device_node *np, + struct property *newprop, + struct property *oldprop); + #if defined(CONFIG_OF_DYNAMIC) /* For updating the device tree at runtime */ extern void of_attach_node(struct device_node *); diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index 0007187..c9cb8a7 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -95,18 +95,8 @@ extern int early_init_dt_scan_root(unsigned long node, const char *uname, int depth, void *data); /* Other Prototypes */ -extern void finish_device_tree(void); extern void unflatten_device_tree(void); extern void early_init_devtree(void *); -extern int of_machine_is_compatible(const char *compat); -extern void print_properties(struct device_node *node); -extern int prom_n_intr_cells(struct device_node* np); -extern void prom_get_irq_senses(unsigned char *senses, int off, int max); -extern int prom_add_property(struct device_node* np, struct property* prop); -extern int prom_remove_property(struct device_node *np, struct property *prop); -extern int prom_update_property(struct device_node *np, - struct property *newprop, - struct property *oldprop); #endif /* __ASSEMBLY__ */ #endif /* _LINUX_OF_FDT_H */ -- cgit v0.10.2 From 7c540d9e3da38c3d1c15fb8059e4577a84ac0066 Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Sun, 14 Feb 2010 07:13:41 -0700 Subject: proc_devtree: fix THIS_MODULE without module.h Commit e22f628395432b967f2f505858c64450f7835365 introduced a build breakage for ARM devtree work: the THIS_MODULE macro was added, but we don't have module.h This change adds the necessary #include to get THIS_MODULE defined. While we could just replace it with NULL (PROC_FS is a bool, not a tristate), using THIS_MODULE will prevent unexpected breakage if we ever do compile this as a module. Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Acked-by: Michal Simek diff --git a/fs/proc/proc_devtree.c b/fs/proc/proc_devtree.c index 0ec4511..f8650dc 100644 --- a/fs/proc/proc_devtree.c +++ b/fs/proc/proc_devtree.c @@ -11,6 +11,7 @@ #include #include #include +#include #include #include #include "internal.h" -- cgit v0.10.2 From 9dfbf207802c7e8cda9d081a8d750b50633c82d2 Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Sun, 14 Feb 2010 07:13:43 -0700 Subject: of: protect linux/of.h with CONFIG_OF For platforms that have CONFIG_OF optional, we need to make the contents of linux/of.h conditional on CONFIG_OF. Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Acked-by: Michal Simek diff --git a/include/linux/of.h b/include/linux/of.h index 5c7b6a6..48b0ee6 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -22,6 +22,8 @@ #include +#ifdef CONFIG_OF + typedef u32 phandle; typedef u32 ihandle; @@ -194,4 +196,5 @@ extern void of_attach_node(struct device_node *); extern void of_detach_node(struct device_node *); #endif +#endif /* CONFIG_OF */ #endif /* _LINUX_OF_H */ -- cgit v0.10.2 From 4ef7b373df330bc0ff037dc4792d373c9346375f Mon Sep 17 00:00:00 2001 From: Jeremy Kerr Date: Sun, 14 Feb 2010 07:13:47 -0700 Subject: of/flattree: Don't assume HAVE_LMB We don't always have lmb available, so make arches provide an early_init_dt_alloc_memory_arch() to handle the allocation of memory in the fdt code. When we don't have lmb.h included, we need asm/page.h for __va. Signed-off-by: Jeremy Kerr Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Acked-by: Michal Simek diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index 050b799..a7dcaf0 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -55,6 +55,11 @@ void __init early_init_dt_add_memory_arch(u64 base, u64 size) lmb_add(base, size); } +u64 __init early_init_dt_alloc_memory_arch(u64 size, u64 align) +{ + return lmb_alloc(size, align); +} + #ifdef CONFIG_EARLY_PRINTK /* MS this is Microblaze specifig function */ static int __init early_init_dt_scan_serial(unsigned long node, diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 43c78d7..5bbbdb29 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -510,6 +510,11 @@ void __init early_init_dt_add_memory_arch(u64 base, u64 size) memstart_addr = min((u64)memstart_addr, base); } +u64 __init early_init_dt_alloc_memory_arch(u64 size, u64 align) +{ + return lmb_alloc(size, align); +} + #ifdef CONFIG_BLK_DEV_INITRD void __init early_init_dt_setup_initrd_arch(unsigned long start, unsigned long end) diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index b51f797..406757a 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -10,16 +10,18 @@ */ #include -#include #include #include #include - +#include +#include #ifdef CONFIG_PPC #include #endif /* CONFIG_PPC */ +#include + int __initdata dt_root_addr_cells; int __initdata dt_root_size_cells; @@ -560,7 +562,8 @@ void __init unflatten_device_tree(void) pr_debug(" size is %lx, allocating...\n", size); /* Allocate memory for the expanded device tree */ - mem = lmb_alloc(size + 4, __alignof__(struct device_node)); + mem = early_init_dt_alloc_memory_arch(size + 4, + __alignof__(struct device_node)); mem = (unsigned long) __va(mem); ((__be32 *)mem)[size / 4] = cpu_to_be32(0xdeadbeef); diff --git a/include/linux/of_fdt.h b/include/linux/of_fdt.h index c9cb8a7..a1ca92c 100644 --- a/include/linux/of_fdt.h +++ b/include/linux/of_fdt.h @@ -78,6 +78,7 @@ extern void early_init_dt_check_for_initrd(unsigned long node); extern int early_init_dt_scan_memory(unsigned long node, const char *uname, int depth, void *data); extern void early_init_dt_add_memory_arch(u64 base, u64 size); +extern u64 early_init_dt_alloc_memory_arch(u64 size, u64 align); extern u64 dt_mem_next_cell(int s, __be32 **cellp); /* -- cgit v0.10.2 From 7c7b60cb87547b1664a4385c187f029bf514a737 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Sun, 14 Feb 2010 07:13:50 -0700 Subject: of: put default string compare and #a/s-cell values into common header Most architectures don't need to change these. Put them into common code to eliminate some duplication Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Acked-by: Michal Simek diff --git a/arch/microblaze/include/asm/prom.h b/arch/microblaze/include/asm/prom.h index 6c6b386..ddc5c57 100644 --- a/arch/microblaze/include/asm/prom.h +++ b/arch/microblaze/include/asm/prom.h @@ -26,13 +26,6 @@ #include #include -#define OF_ROOT_NODE_ADDR_CELLS_DEFAULT 1 -#define OF_ROOT_NODE_SIZE_CELLS_DEFAULT 1 - -#define of_compat_cmp(s1, s2, l) strncasecmp((s1), (s2), (l)) -#define of_prop_cmp(s1, s2) strcmp((s1), (s2)) -#define of_node_cmp(s1, s2) strcasecmp((s1), (s2)) - extern struct device_node *of_chosen; #define HAVE_ARCH_DEVTREE_FIXUPS diff --git a/arch/powerpc/include/asm/prom.h b/arch/powerpc/include/asm/prom.h index f384db8..4a5070e 100644 --- a/arch/powerpc/include/asm/prom.h +++ b/arch/powerpc/include/asm/prom.h @@ -23,13 +23,6 @@ #include #include -#define OF_ROOT_NODE_ADDR_CELLS_DEFAULT 1 -#define OF_ROOT_NODE_SIZE_CELLS_DEFAULT 1 - -#define of_compat_cmp(s1, s2, l) strcasecmp((s1), (s2)) -#define of_prop_cmp(s1, s2) strcmp((s1), (s2)) -#define of_node_cmp(s1, s2) strcasecmp((s1), (s2)) - extern struct device_node *of_chosen; #define HAVE_ARCH_DEVTREE_FIXUPS diff --git a/include/linux/of.h b/include/linux/of.h index 48b0ee6..5cd2840 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -116,6 +116,19 @@ static inline unsigned long of_read_ulong(const __be32 *cell, int size) #include +/* Default #address and #size cells. Allow arch asm/prom.h to override */ +#if !defined(OF_ROOT_NODE_ADDR_CELLS_DEFAULT) +#define OF_ROOT_NODE_ADDR_CELLS_DEFAULT 1 +#define OF_ROOT_NODE_SIZE_CELLS_DEFAULT 1 +#endif + +/* Default string compare functions, Allow arch asm/prom.h to override */ +#if !defined(of_compat_cmp) +#define of_compat_cmp(s1, s2, l) strncasecmp((s1), (s2), (l)) +#define of_prop_cmp(s1, s2) strcmp((s1), (s2)) +#define of_node_cmp(s1, s2) strcasecmp((s1), (s2)) +#endif + /* flag descriptions */ #define OF_DYNAMIC 1 /* node and properties were allocated via kmalloc */ #define OF_DETACHED 2 /* node has been detached from the device tree */ -- cgit v0.10.2 From 22d5579e66101162fd1119f2e7f4f999ca8b48c7 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Sun, 14 Feb 2010 07:13:52 -0700 Subject: of: remove unused extern reference to devtree_lock Neither the powerpc nor the microblaze code use devtree_lock anymore. Remove the extern reference. Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Acked-by: Michal Simek diff --git a/arch/microblaze/include/asm/prom.h b/arch/microblaze/include/asm/prom.h index ddc5c57..8b1ebd3 100644 --- a/arch/microblaze/include/asm/prom.h +++ b/arch/microblaze/include/asm/prom.h @@ -30,8 +30,6 @@ extern struct device_node *of_chosen; #define HAVE_ARCH_DEVTREE_FIXUPS -extern rwlock_t devtree_lock; /* temporary while merging */ - /* Other Prototypes */ extern int early_uartlite_console(void); diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 5bbbdb29..4869c93 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -67,8 +67,6 @@ int __initdata iommu_force_on; unsigned long tce_alloc_start, tce_alloc_end; #endif -extern rwlock_t devtree_lock; /* temporary while merging */ - /* export that to outside world */ struct device_node *of_chosen; -- cgit v0.10.2 From fc0bdae49d810e4cb32d7b547bc6d4dfb08f9e2e Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Sun, 14 Feb 2010 07:13:55 -0700 Subject: of: move definition of of_chosen into common code. Rather than defining of_chosen in each arch, it can be defined for all in driver/of/base.c Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Acked-by: Michal Simek diff --git a/arch/microblaze/include/asm/prom.h b/arch/microblaze/include/asm/prom.h index 8b1ebd3..aa1a437 100644 --- a/arch/microblaze/include/asm/prom.h +++ b/arch/microblaze/include/asm/prom.h @@ -26,8 +26,6 @@ #include #include -extern struct device_node *of_chosen; - #define HAVE_ARCH_DEVTREE_FIXUPS /* Other Prototypes */ diff --git a/arch/microblaze/kernel/prom.c b/arch/microblaze/kernel/prom.c index a7dcaf0..a15ef6d 100644 --- a/arch/microblaze/kernel/prom.c +++ b/arch/microblaze/kernel/prom.c @@ -42,9 +42,6 @@ #include #include -/* export that to outside world */ -struct device_node *of_chosen; - void __init early_init_dt_scan_chosen_arch(unsigned long node) { /* No Microblaze specific code here */ diff --git a/arch/powerpc/include/asm/prom.h b/arch/powerpc/include/asm/prom.h index 4a5070e..7f9a50aa 100644 --- a/arch/powerpc/include/asm/prom.h +++ b/arch/powerpc/include/asm/prom.h @@ -23,8 +23,6 @@ #include #include -extern struct device_node *of_chosen; - #define HAVE_ARCH_DEVTREE_FIXUPS #ifdef CONFIG_PPC32 diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 4869c93..43238b2 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -67,9 +67,6 @@ int __initdata iommu_force_on; unsigned long tce_alloc_start, tce_alloc_end; #endif -/* export that to outside world */ -struct device_node *of_chosen; - static int __init early_parse_mem(char *p) { if (!p) diff --git a/drivers/of/base.c b/drivers/of/base.c index 873479a..cb96888 100644 --- a/drivers/of/base.c +++ b/drivers/of/base.c @@ -23,6 +23,7 @@ #include struct device_node *allnodes; +struct device_node *of_chosen; /* use when traversing tree through the allnext, child, sibling, * or parent members of struct device_node. diff --git a/include/linux/of.h b/include/linux/of.h index 5cd2840..d34cc5d 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -66,6 +66,7 @@ struct device_node { /* Pointer for first entry in chain of all nodes. */ extern struct device_node *allnodes; +extern struct device_node *of_chosen; static inline int of_node_check_flag(struct device_node *n, unsigned long flag) { -- cgit v0.10.2 From 6f93d0a7c83772997b81c30d6f519a9a5dbab6a9 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Sun, 14 Feb 2010 11:12:04 +0100 Subject: perf_events: Fix FORK events Commit 22e19085 ("Honour event state for aux stream data") introduced a bug where we would drop FORK events. The thing is that we deliver FORK events to the child process' event, which at that time will be PERF_EVENT_STATE_INACTIVE because the child won't be scheduled in (we're in the middle of fork). Solve this twice, change the event state filter to exclude only disabled (STATE_OFF) or worse, and deliver FORK events to the current (parent). Signed-off-by: Peter Zijlstra Cc: Anton Blanchard Cc: Arnaldo Carvalho de Melo LKML-Reference: <1266142324.5273.411.camel@laptop> Signed-off-by: Ingo Molnar diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 2b19297..2ae7409 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -3259,8 +3259,6 @@ static void perf_event_task_output(struct perf_event *event, task_event->event_id.tid = perf_event_tid(event, task); task_event->event_id.ptid = perf_event_tid(event, current); - task_event->event_id.time = perf_clock(); - perf_output_put(&handle, task_event->event_id); perf_output_end(&handle); @@ -3268,7 +3266,7 @@ static void perf_event_task_output(struct perf_event *event, static int perf_event_task_match(struct perf_event *event) { - if (event->state != PERF_EVENT_STATE_ACTIVE) + if (event->state < PERF_EVENT_STATE_INACTIVE) return 0; if (event->cpu != -1 && event->cpu != smp_processor_id()) @@ -3300,7 +3298,7 @@ static void perf_event_task_event(struct perf_task_event *task_event) cpuctx = &get_cpu_var(perf_cpu_context); perf_event_task_ctx(&cpuctx->ctx, task_event); if (!ctx) - ctx = rcu_dereference(task_event->task->perf_event_ctxp); + ctx = rcu_dereference(current->perf_event_ctxp); if (ctx) perf_event_task_ctx(ctx, task_event); put_cpu_var(perf_cpu_context); @@ -3331,6 +3329,7 @@ static void perf_event_task(struct task_struct *task, /* .ppid */ /* .tid */ /* .ptid */ + .time = perf_clock(), }, }; @@ -3380,7 +3379,7 @@ static void perf_event_comm_output(struct perf_event *event, static int perf_event_comm_match(struct perf_event *event) { - if (event->state != PERF_EVENT_STATE_ACTIVE) + if (event->state < PERF_EVENT_STATE_INACTIVE) return 0; if (event->cpu != -1 && event->cpu != smp_processor_id()) @@ -3500,7 +3499,7 @@ static void perf_event_mmap_output(struct perf_event *event, static int perf_event_mmap_match(struct perf_event *event, struct perf_mmap_event *mmap_event) { - if (event->state != PERF_EVENT_STATE_ACTIVE) + if (event->state < PERF_EVENT_STATE_INACTIVE) return 0; if (event->cpu != -1 && event->cpu != smp_processor_id()) -- cgit v0.10.2 From 0d351c3e932c2e155ef5e4c3f5b87223abd4eea6 Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Sun, 14 Feb 2010 14:13:57 -0700 Subject: of/sparc: Remove sparc-local declaration of allnodes and devtree_lock Both allnodes and devtree_lock are defined in common code. The extern declaration should be in the common header too so that the compiler can type check. allnodes is already in of.h, but devtree_lock should be declared there too. This patch removes the SPARC declarations and uses decls in of.h instead. Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Acked-by: Michal Simek Acked-by: David S. Miller diff --git a/arch/sparc/kernel/prom.h b/arch/sparc/kernel/prom.h index 453397f..a8591ef 100644 --- a/arch/sparc/kernel/prom.h +++ b/arch/sparc/kernel/prom.h @@ -4,9 +4,6 @@ #include #include -extern struct device_node *allnodes; /* temporary while merging */ -extern rwlock_t devtree_lock; /* temporary while merging */ - extern void * prom_early_alloc(unsigned long size); extern void irq_trans_init(struct device_node *dp); diff --git a/include/linux/of.h b/include/linux/of.h index d34cc5d..f6d9cbc 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -19,6 +19,7 @@ #include #include #include +#include #include @@ -67,6 +68,7 @@ struct device_node { /* Pointer for first entry in chain of all nodes. */ extern struct device_node *allnodes; extern struct device_node *of_chosen; +extern rwlock_t devtree_lock; static inline int of_node_check_flag(struct device_node *n, unsigned long flag) { -- cgit v0.10.2 From df0edeb59eb559be0bee53452fda2f5cc0ae133f Mon Sep 17 00:00:00 2001 From: Grant Likely Date: Sun, 14 Feb 2010 14:14:00 -0700 Subject: of: remove undefined request_OF_resource & release_OF_resource Neither request_OF_resource or release_OF_resource are defined anywhere. Remove the declarations. Signed-off-by: Grant Likely Acked-by: Benjamin Herrenschmidt Acked-by: Michal Simek diff --git a/arch/microblaze/include/asm/prom.h b/arch/microblaze/include/asm/prom.h index aa1a437..03f45a9 100644 --- a/arch/microblaze/include/asm/prom.h +++ b/arch/microblaze/include/asm/prom.h @@ -31,10 +31,6 @@ /* Other Prototypes */ extern int early_uartlite_console(void); -extern struct resource *request_OF_resource(struct device_node *node, - int index, const char *name_postfix); -extern int release_OF_resource(struct device_node *node, int index); - /* * OF address retreival & translation */ diff --git a/arch/powerpc/include/asm/prom.h b/arch/powerpc/include/asm/prom.h index 7f9a50aa..ddd408a 100644 --- a/arch/powerpc/include/asm/prom.h +++ b/arch/powerpc/include/asm/prom.h @@ -39,11 +39,6 @@ extern struct device_node* pci_device_to_OF_node(struct pci_dev *); extern void pci_create_OF_bus_map(void); #endif -extern struct resource *request_OF_resource(struct device_node* node, - int index, const char* name_postfix); -extern int release_OF_resource(struct device_node* node, int index); - - /* * OF address retreival & translation */ -- cgit v0.10.2 From 8b833c506c05c498d4215e2c260be44225daf6de Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Fri, 5 Feb 2010 01:24:34 -0500 Subject: kprobes: Add mcount to the kprobes blacklist Since mcount function can be called from everywhere, it should be blacklisted. Moreover, the "mcount" symbol is a special symbol name. So, it is better to put it in the generic blacklist. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Ananth N Mavinakayanahalli Cc: Steven Rostedt LKML-Reference: <20100205062433.3745.36726.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar diff --git a/kernel/kprobes.c b/kernel/kprobes.c index b7df302..c4b4343 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -93,6 +93,7 @@ static struct kprobe_blackpoint kprobe_blacklist[] = { {"native_get_debugreg",}, {"irq_entries_start",}, {"common_interrupt",}, + {"mcount",}, /* mcount can be called from everywhere */ {NULL} /* Terminator */ }; -- cgit v0.10.2 From 04c869735541c27dd137c55f35f8a18bb372bbe1 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Mon, 15 Feb 2010 16:10:57 +0900 Subject: sh: Fix up legacy PMB mode offset calculation. The change for fixing up sh64 inadvertently inverted the logic for legacy PMB, fix that back up. Signed-off-by: Paul Mundt diff --git a/arch/sh/kernel/vmlinux.lds.S b/arch/sh/kernel/vmlinux.lds.S index dcf4891..0e66c7b 100644 --- a/arch/sh/kernel/vmlinux.lds.S +++ b/arch/sh/kernel/vmlinux.lds.S @@ -4,11 +4,9 @@ */ #ifdef CONFIG_SUPERH64 #define LOAD_OFFSET PAGE_OFFSET -#define MEMORY_OFFSET __MEMORY_START OUTPUT_ARCH(sh:sh5) #else #define LOAD_OFFSET 0 -#define MEMORY_OFFSET 0 OUTPUT_ARCH(sh) #endif @@ -16,6 +14,13 @@ OUTPUT_ARCH(sh) #include #include +#if defined(CONFIG_29BIT) || defined(CONFIG_SUPERH64) || \ + defined(CONFIG_PMB_LEGACY) + #define MEMORY_OFFSET __MEMORY_START +#else + #define MEMORY_OFFSET 0 +#endif + ENTRY(_start) SECTIONS { -- cgit v0.10.2 From 68fd111e02b979876359c7b471a8bcbca0628b75 Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Mon, 15 Feb 2010 13:43:25 -0800 Subject: x86, numa: Fix numa emulation calculation of big nodes numa=fake=N uses split_nodes_interleave() to partition the system into N fake nodes. Each node size must have be a multiple of FAKE_NODE_MIN_SIZE, otherwise it is possible to get strange alignments. Because of this, the remaining memory from each node when rounded to FAKE_NODE_MIN_SIZE is consolidated into a number of "big nodes" that are bigger than the rest. The calculation of the number of big nodes is incorrect since it is using a logical AND operator when it should be multiplying the rounded-off portion of each node with N. Signed-off-by: David Rientjes LKML-Reference: Signed-off-by: H. Peter Anvin diff --git a/arch/x86/mm/numa_64.c b/arch/x86/mm/numa_64.c index 83bbc70..2ecbe0c 100644 --- a/arch/x86/mm/numa_64.c +++ b/arch/x86/mm/numa_64.c @@ -427,7 +427,7 @@ static int __init split_nodes_interleave(u64 addr, u64 max_addr, * Calculate the number of big nodes that can be allocated as a result * of consolidating the remainder. */ - big = ((size & ~FAKE_NODE_MIN_HASH_MASK) & nr_nodes) / + big = ((size & ~FAKE_NODE_MIN_HASH_MASK) * nr_nodes) / FAKE_NODE_MIN_SIZE; size &= FAKE_NODE_MIN_HASH_MASK; -- cgit v0.10.2 From 8df5bb34defd685fe86f60746bbf3d47d1c6f033 Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Mon, 15 Feb 2010 13:43:30 -0800 Subject: x86, numa: Add fixed node size option for numa emulation numa=fake=N specifies the number of fake nodes, N, to partition the system into and then allocates them by interleaving over physical nodes. This requires knowledge of the system capacity when attempting to allocate nodes of a certain size: either very large nodes to benchmark scalability of code that operates on individual nodes, or very small nodes to find bugs in the VM. This patch introduces numa=fake=[MG] so it is possible to specify the size of each node to allocate. When used, nodes of the size specified will be allocated and interleaved over the set of physical nodes. FAKE_NODE_MIN_SIZE was also moved to the more-appropriate include/asm/numa_64.h. Signed-off-by: David Rientjes LKML-Reference: Signed-off-by: H. Peter Anvin diff --git a/Documentation/x86/x86_64/boot-options.txt b/Documentation/x86/x86_64/boot-options.txt index 29a6ff8..01150c6 100644 --- a/Documentation/x86/x86_64/boot-options.txt +++ b/Documentation/x86/x86_64/boot-options.txt @@ -166,6 +166,10 @@ NUMA numa=noacpi Don't parse the SRAT table for NUMA setup + numa=fake=[MG] + If given as a memory unit, fills all system RAM with nodes of + size interleaved over physical nodes. + numa=fake=CMDLINE If a number, fakes CMDLINE nodes and ignores NUMA setup of the actual machine. Otherwise, system memory is configured diff --git a/arch/x86/include/asm/mmzone_64.h b/arch/x86/include/asm/mmzone_64.h index a29f48c..288b96f 100644 --- a/arch/x86/include/asm/mmzone_64.h +++ b/arch/x86/include/asm/mmzone_64.h @@ -39,11 +39,5 @@ static inline __attribute__((pure)) int phys_to_nid(unsigned long addr) #define node_start_pfn(nid) (NODE_DATA(nid)->node_start_pfn) #define node_end_pfn(nid) (NODE_DATA(nid)->node_start_pfn + \ NODE_DATA(nid)->node_spanned_pages) - -#ifdef CONFIG_NUMA_EMU -#define FAKE_NODE_MIN_SIZE (64 * 1024 * 1024) -#define FAKE_NODE_MIN_HASH_MASK (~(FAKE_NODE_MIN_SIZE - 1UL)) -#endif - #endif #endif /* _ASM_X86_MMZONE_64_H */ diff --git a/arch/x86/include/asm/numa_64.h b/arch/x86/include/asm/numa_64.h index c4ae822..823e070 100644 --- a/arch/x86/include/asm/numa_64.h +++ b/arch/x86/include/asm/numa_64.h @@ -36,6 +36,11 @@ extern void __cpuinit numa_set_node(int cpu, int node); extern void __cpuinit numa_clear_node(int cpu); extern void __cpuinit numa_add_cpu(int cpu); extern void __cpuinit numa_remove_cpu(int cpu); + +#ifdef CONFIG_NUMA_EMU +#define FAKE_NODE_MIN_SIZE ((u64)64 << 20) +#define FAKE_NODE_MIN_HASH_MASK (~(FAKE_NODE_MIN_SIZE - 1UL)) +#endif /* CONFIG_NUMA_EMU */ #else static inline void init_cpu_to_node(void) { } static inline void numa_set_node(int cpu, int node) { } diff --git a/arch/x86/mm/numa_64.c b/arch/x86/mm/numa_64.c index 2ecbe0c..c47c78b 100644 --- a/arch/x86/mm/numa_64.c +++ b/arch/x86/mm/numa_64.c @@ -502,6 +502,102 @@ static int __init split_nodes_interleave(u64 addr, u64 max_addr, } /* + * Returns the end address of a node so that there is at least `size' amount of + * non-reserved memory or `max_addr' is reached. + */ +static u64 __init find_end_of_node(u64 start, u64 max_addr, u64 size) +{ + u64 end = start + size; + + while (end - start - e820_hole_size(start, end) < size) { + end += FAKE_NODE_MIN_SIZE; + if (end > max_addr) { + end = max_addr; + break; + } + } + return end; +} + +/* + * Sets up fake nodes of `size' interleaved over physical nodes ranging from + * `addr' to `max_addr'. The return value is the number of nodes allocated. + */ +static int __init split_nodes_size_interleave(u64 addr, u64 max_addr, u64 size) +{ + nodemask_t physnode_mask = NODE_MASK_NONE; + u64 min_size; + int ret = 0; + int i; + + if (!size) + return -1; + /* + * The limit on emulated nodes is MAX_NUMNODES, so the size per node is + * increased accordingly if the requested size is too small. This + * creates a uniform distribution of node sizes across the entire + * machine (but not necessarily over physical nodes). + */ + min_size = (max_addr - addr - e820_hole_size(addr, max_addr)) / + MAX_NUMNODES; + min_size = max(min_size, FAKE_NODE_MIN_SIZE); + if ((min_size & FAKE_NODE_MIN_HASH_MASK) < min_size) + min_size = (min_size + FAKE_NODE_MIN_SIZE) & + FAKE_NODE_MIN_HASH_MASK; + if (size < min_size) { + pr_err("Fake node size %LuMB too small, increasing to %LuMB\n", + size >> 20, min_size >> 20); + size = min_size; + } + size &= FAKE_NODE_MIN_HASH_MASK; + + for (i = 0; i < MAX_NUMNODES; i++) + if (physnodes[i].start != physnodes[i].end) + node_set(i, physnode_mask); + /* + * Fill physical nodes with fake nodes of size until there is no memory + * left on any of them. + */ + while (nodes_weight(physnode_mask)) { + for_each_node_mask(i, physnode_mask) { + u64 dma32_end = MAX_DMA32_PFN << PAGE_SHIFT; + u64 end; + + end = find_end_of_node(physnodes[i].start, + physnodes[i].end, size); + /* + * If there won't be at least FAKE_NODE_MIN_SIZE of + * non-reserved memory in ZONE_DMA32 for the next node, + * this one must extend to the boundary. + */ + if (end < dma32_end && dma32_end - end - + e820_hole_size(end, dma32_end) < FAKE_NODE_MIN_SIZE) + end = dma32_end; + + /* + * If there won't be enough non-reserved memory for the + * next node, this one must extend to the end of the + * physical node. + */ + if (physnodes[i].end - end - + e820_hole_size(end, physnodes[i].end) < size) + end = physnodes[i].end; + + /* + * Setup the fake node that will be allocated as bootmem + * later. If setup_node_range() returns non-zero, there + * is no more memory available on this physical node. + */ + if (setup_node_range(ret++, &physnodes[i].start, + end - physnodes[i].start, + physnodes[i].end) < 0) + node_clear(i, physnode_mask); + } + } + return ret; +} + +/* * Splits num_nodes nodes up equally starting at node_start. The return value * is the number of nodes split up and addr is adjusted to be at the end of the * last node allocated. @@ -546,14 +642,7 @@ static int __init split_nodes_equally(u64 *addr, u64 max_addr, int node_start, if (i == num_nodes + node_start - 1) end = max_addr; else - while (end - *addr - e820_hole_size(*addr, end) < - size) { - end += FAKE_NODE_MIN_SIZE; - if (end > max_addr) { - end = max_addr; - break; - } - } + end = find_end_of_node(*addr, max_addr, size); if (setup_node_range(i, addr, end - *addr, max_addr) < 0) break; } @@ -589,6 +678,18 @@ static int __init numa_emulation(unsigned long start_pfn, num_phys_nodes = setup_physnodes(addr, max_addr, acpi, k8); /* + * If the numa=fake command-line contains a 'M' or 'G', it represents + * the fixed node size. + */ + if (strchr(cmdline, 'M') || strchr(cmdline, 'G')) { + size = memparse(cmdline, &cmdline); + num_nodes = split_nodes_size_interleave(addr, max_addr, size); + if (num_nodes < 0) + return num_nodes; + goto out; + } + + /* * If the numa=fake command-line is just a single number N, split the * system RAM into N fake nodes. */ -- cgit v0.10.2 From ca2107c9d6cf44fb915402d6f12b9d9ff3925cd7 Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Mon, 15 Feb 2010 13:43:33 -0800 Subject: x86, numa: Remove configurable node size support for numa emulation Now that numa=fake=[MG] is implemented, it is possible to remove configurable node size support. The command-line parsing was already broken (numa=fake=*128, for example, would not work) and since fake nodes are now interleaved over physical nodes, this support is no longer required. Signed-off-by: David Rientjes LKML-Reference: Signed-off-by: H. Peter Anvin diff --git a/Documentation/x86/x86_64/boot-options.txt b/Documentation/x86/x86_64/boot-options.txt index 01150c6..7fbbaf8 100644 --- a/Documentation/x86/x86_64/boot-options.txt +++ b/Documentation/x86/x86_64/boot-options.txt @@ -170,19 +170,9 @@ NUMA If given as a memory unit, fills all system RAM with nodes of size interleaved over physical nodes. - numa=fake=CMDLINE - If a number, fakes CMDLINE nodes and ignores NUMA setup of the - actual machine. Otherwise, system memory is configured - depending on the sizes and coefficients listed. For example: - numa=fake=2*512,1024,4*256,*128 - gives two 512M nodes, a 1024M node, four 256M nodes, and the - rest split into 128M chunks. If the last character of CMDLINE - is a *, the remaining memory is divided up equally among its - coefficient: - numa=fake=2*512,2* - gives two 512M nodes and the rest split into two nodes. - Otherwise, the remaining system RAM is allocated to an - additional node. + numa=fake= + If given as an integer, fills all system RAM with N fake nodes + interleaved over physical nodes. ACPI diff --git a/arch/x86/mm/numa_64.c b/arch/x86/mm/numa_64.c index c47c78b..3307ea8 100644 --- a/arch/x86/mm/numa_64.c +++ b/arch/x86/mm/numa_64.c @@ -598,172 +598,38 @@ static int __init split_nodes_size_interleave(u64 addr, u64 max_addr, u64 size) } /* - * Splits num_nodes nodes up equally starting at node_start. The return value - * is the number of nodes split up and addr is adjusted to be at the end of the - * last node allocated. - */ -static int __init split_nodes_equally(u64 *addr, u64 max_addr, int node_start, - int num_nodes) -{ - unsigned int big; - u64 size; - int i; - - if (num_nodes <= 0) - return -1; - if (num_nodes > MAX_NUMNODES) - num_nodes = MAX_NUMNODES; - size = (max_addr - *addr - e820_hole_size(*addr, max_addr)) / - num_nodes; - /* - * Calculate the number of big nodes that can be allocated as a result - * of consolidating the leftovers. - */ - big = ((size & ~FAKE_NODE_MIN_HASH_MASK) * num_nodes) / - FAKE_NODE_MIN_SIZE; - - /* Round down to nearest FAKE_NODE_MIN_SIZE. */ - size &= FAKE_NODE_MIN_HASH_MASK; - if (!size) { - printk(KERN_ERR "Not enough memory for each node. " - "NUMA emulation disabled.\n"); - return -1; - } - - for (i = node_start; i < num_nodes + node_start; i++) { - u64 end = *addr + size; - - if (i < big) - end += FAKE_NODE_MIN_SIZE; - /* - * The final node can have the remaining system RAM. Other - * nodes receive roughly the same amount of available pages. - */ - if (i == num_nodes + node_start - 1) - end = max_addr; - else - end = find_end_of_node(*addr, max_addr, size); - if (setup_node_range(i, addr, end - *addr, max_addr) < 0) - break; - } - return i - node_start + 1; -} - -/* - * Splits the remaining system RAM into chunks of size. The remaining memory is - * always assigned to a final node and can be asymmetric. Returns the number of - * nodes split. - */ -static int __init split_nodes_by_size(u64 *addr, u64 max_addr, int node_start, - u64 size) -{ - int i = node_start; - size = (size << 20) & FAKE_NODE_MIN_HASH_MASK; - while (!setup_node_range(i++, addr, size, max_addr)) - ; - return i - node_start; -} - -/* * Sets up the system RAM area from start_pfn to last_pfn according to the * numa=fake command-line option. */ static int __init numa_emulation(unsigned long start_pfn, unsigned long last_pfn, int acpi, int k8) { - u64 size, addr = start_pfn << PAGE_SHIFT; + u64 addr = start_pfn << PAGE_SHIFT; u64 max_addr = last_pfn << PAGE_SHIFT; - int num_nodes = 0, num = 0, coeff_flag, coeff = -1, i; int num_phys_nodes; + int num_nodes; + int i; num_phys_nodes = setup_physnodes(addr, max_addr, acpi, k8); /* * If the numa=fake command-line contains a 'M' or 'G', it represents - * the fixed node size. + * the fixed node size. Otherwise, if it is just a single number N, + * split the system RAM into N fake nodes. */ if (strchr(cmdline, 'M') || strchr(cmdline, 'G')) { + u64 size; + size = memparse(cmdline, &cmdline); num_nodes = split_nodes_size_interleave(addr, max_addr, size); - if (num_nodes < 0) - return num_nodes; - goto out; - } + } else { + unsigned long n; - /* - * If the numa=fake command-line is just a single number N, split the - * system RAM into N fake nodes. - */ - if (!strchr(cmdline, '*') && !strchr(cmdline, ',')) { - long n = simple_strtol(cmdline, NULL, 0); - - num_nodes = split_nodes_interleave(addr, max_addr, - num_phys_nodes, n); - if (num_nodes < 0) - return num_nodes; - goto out; + n = simple_strtoul(cmdline, NULL, 0); + num_nodes = split_nodes_interleave(addr, max_addr, num_phys_nodes, n); } - /* Parse the command line. */ - for (coeff_flag = 0; ; cmdline++) { - if (*cmdline && isdigit(*cmdline)) { - num = num * 10 + *cmdline - '0'; - continue; - } - if (*cmdline == '*') { - if (num > 0) - coeff = num; - coeff_flag = 1; - } - if (!*cmdline || *cmdline == ',') { - if (!coeff_flag) - coeff = 1; - /* - * Round down to the nearest FAKE_NODE_MIN_SIZE. - * Command-line coefficients are in megabytes. - */ - size = ((u64)num << 20) & FAKE_NODE_MIN_HASH_MASK; - if (size) - for (i = 0; i < coeff; i++, num_nodes++) - if (setup_node_range(num_nodes, &addr, - size, max_addr) < 0) - goto done; - if (!*cmdline) - break; - coeff_flag = 0; - coeff = -1; - } - num = 0; - } -done: - if (!num_nodes) - return -1; - /* Fill remainder of system RAM, if appropriate. */ - if (addr < max_addr) { - if (coeff_flag && coeff < 0) { - /* Split remaining nodes into num-sized chunks */ - num_nodes += split_nodes_by_size(&addr, max_addr, - num_nodes, num); - goto out; - } - switch (*(cmdline - 1)) { - case '*': - /* Split remaining nodes into coeff chunks */ - if (coeff <= 0) - break; - num_nodes += split_nodes_equally(&addr, max_addr, - num_nodes, coeff); - break; - case ',': - /* Do not allocate remaining system RAM */ - break; - default: - /* Give one final node */ - setup_node_range(num_nodes, &addr, max_addr - addr, - max_addr); - num_nodes++; - } - } -out: + if (num_nodes < 0) + return num_nodes; memnode_shift = compute_hash_shift(nodes, num_nodes, NULL); if (memnode_shift < 0) { memnode_shift = 0; -- cgit v0.10.2 From d9184fa97b6f48d399636e5e2669bc8419f9369e Mon Sep 17 00:00:00 2001 From: Ben Skeggs Date: Tue, 16 Feb 2010 11:14:14 +1000 Subject: drm/nouveau: use mutex for vbios lock Signed-off-by: Ben Skeggs diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c index 2cd0fad..0e9cd1d 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bios.c +++ b/drivers/gpu/drm/nouveau/nouveau_bios.c @@ -5861,13 +5861,12 @@ nouveau_bios_run_init_table(struct drm_device *dev, uint16_t table, struct drm_nouveau_private *dev_priv = dev->dev_private; struct nvbios *bios = &dev_priv->VBIOS; struct init_exec iexec = { true, false }; - unsigned long flags; - spin_lock_irqsave(&bios->lock, flags); + mutex_lock(&bios->lock); bios->display.output = dcbent; parse_init_table(bios, table, &iexec); bios->display.output = NULL; - spin_unlock_irqrestore(&bios->lock, flags); + mutex_unlock(&bios->lock); } static bool NVInitVBIOS(struct drm_device *dev) @@ -5876,7 +5875,7 @@ static bool NVInitVBIOS(struct drm_device *dev) struct nvbios *bios = &dev_priv->VBIOS; memset(bios, 0, sizeof(struct nvbios)); - spin_lock_init(&bios->lock); + mutex_init(&bios->lock); bios->dev = dev; if (!NVShadowVBIOS(dev, bios->data)) diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.h b/drivers/gpu/drm/nouveau/nouveau_bios.h index 68446fd..fd94bd6 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bios.h +++ b/drivers/gpu/drm/nouveau/nouveau_bios.h @@ -205,7 +205,7 @@ struct nvbios { struct drm_device *dev; struct nouveau_bios_info pub; - spinlock_t lock; + struct mutex lock; uint8_t data[NV_PROM_SIZE]; unsigned int length; -- cgit v0.10.2 From bf929efa56ac174bf6d4f54cd6fe811181a51ae5 Mon Sep 17 00:00:00 2001 From: Francisco Jerez Date: Thu, 11 Feb 2010 12:47:40 +0100 Subject: drm/nouveau: Force TV encoder DPMS reinit after resume. Signed-off-by: Francisco Jerez Signed-off-by: Ben Skeggs diff --git a/drivers/gpu/drm/nouveau/nv17_tv.c b/drivers/gpu/drm/nouveau/nv17_tv.c index 58b917c..21ac6e4 100644 --- a/drivers/gpu/drm/nouveau/nv17_tv.c +++ b/drivers/gpu/drm/nouveau/nv17_tv.c @@ -579,6 +579,8 @@ static void nv17_tv_restore(struct drm_encoder *encoder) nouveau_encoder(encoder)->restore.output); nv17_tv_state_load(dev, &to_tv_enc(encoder)->saved_state); + + nouveau_encoder(encoder)->last_dpms = NV_DPMS_CLEARED; } static int nv17_tv_create_resources(struct drm_encoder *encoder, -- cgit v0.10.2 From 942fa3b63eb525aa0512ba28c42e656d8efc6787 Mon Sep 17 00:00:00 2001 From: Alan Cox Date: Mon, 8 Feb 2010 10:03:17 +0000 Subject: x86, mtrr: Kill over the top warn Fixes bugzilla: http://bugzilla.kernel.org/show_bug.cgi?id=12558 Fixes bugzilla: http://bugzilla.kernel.org/show_bug.cgi?id=12317 (and if this really needed to be a warn you'd be responding to the bugs left in bugzilla from it...) Signed-off-by: Alan Cox LKML-Reference: <20100208100239.2568.2940.stgit@localhost.localdomain> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c index 4d75584..163e59e 100644 --- a/arch/x86/kernel/cpu/mtrr/generic.c +++ b/arch/x86/kernel/cpu/mtrr/generic.c @@ -464,7 +464,7 @@ static void generic_get_mtrr(unsigned int reg, unsigned long *base, tmp |= ~((1<<(hi - 1)) - 1); if (tmp != mask_lo) { - WARN_ONCE(1, KERN_INFO "mtrr: your BIOS has set up an incorrect mask, fixing it up.\n"); + printk(KERN_WARNING "mtrr: your BIOS has configured an incorrect mask, fixing it.\n"); mask_lo = tmp; } } -- cgit v0.10.2 From 40331b21f5fdb746e80fc609ef60ef71b5cd47d9 Mon Sep 17 00:00:00 2001 From: Phil Edworthy Date: Mon, 15 Feb 2010 13:57:49 +0000 Subject: video: sh_mobile_lcdcfb: Add wait for vsync. Added FBIO_WAITFORVSYNC ioctl for SH-Mobile devices. Tested on MS7724 and MigoR boards against 2.6.33-rc7. Signed-off-by: Phil Edworthy Signed-off-by: Paul Mundt diff --git a/drivers/video/sh_mobile_lcdcfb.c b/drivers/video/sh_mobile_lcdcfb.c index a69830d..37aeced 100644 --- a/drivers/video/sh_mobile_lcdcfb.c +++ b/drivers/video/sh_mobile_lcdcfb.c @@ -19,6 +19,7 @@ #include #include #include +#include #include