From 1adcaafe7414c5731f758b158aa0525057225deb Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Mon, 17 Aug 2009 13:23:50 -0700 Subject: x86, pat: Allow ISA memory range uncacheable mapping requests Max Vozeler reported: > Bug 13877 - bogl-term broken with CONFIG_X86_PAT=y, works with =n > > strace of bogl-term: > 814 mmap2(NULL, 65536, PROT_READ|PROT_WRITE, MAP_SHARED, 4, 0) > = -1 EAGAIN (Resource temporarily unavailable) > 814 write(2, "bogl: mmaping /dev/fb0: Resource temporarily unavailable\n", > 57) = 57 PAT code maps the ISA memory range as WB in the PAT attribute, so that fixed range MTRR registers define the actual memory type (UC/WC/WT etc). But the upper level is_new_memtype_allowed() API checks are failing, as the request here is for UC and the return tracked type is WB (Tracked type is WB as MTRR type for this legacy range potentially will be different for each 4k page). Fix is_new_memtype_allowed() by always succeeding the ISA address range checks, as the null PAT (WB) and def MTRR fixed range register settings satisfy the memory type needs of the applications that map the ISA address range. Reported-and-Tested-by: Max Vozeler Signed-off-by: Suresh Siddha Signed-off-by: Venkatesh Pallipadi Signed-off-by: H. Peter Anvin diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index 3cc06e3..1674807 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -2,6 +2,7 @@ #define _ASM_X86_PGTABLE_H #include +#include #include @@ -269,10 +270,17 @@ static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot) #define canon_pgprot(p) __pgprot(massage_pgprot(p)) -static inline int is_new_memtype_allowed(unsigned long flags, - unsigned long new_flags) +static inline int is_new_memtype_allowed(u64 paddr, unsigned long size, + unsigned long flags, + unsigned long new_flags) { /* + * PAT type is always WB for ISA. So no need to check. + */ + if (is_ISA_range(paddr, paddr + size - 1)) + return 1; + + /* * Certain new memtypes are not allowed with certain * requested memtype: * - request is uncached, return cannot be write-back diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c index e6718bb..352aa9e 100644 --- a/arch/x86/mm/pat.c +++ b/arch/x86/mm/pat.c @@ -623,7 +623,8 @@ static int reserve_pfn_range(u64 paddr, unsigned long size, pgprot_t *vma_prot, return ret; if (flags != want_flags) { - if (strict_prot || !is_new_memtype_allowed(want_flags, flags)) { + if (strict_prot || + !is_new_memtype_allowed(paddr, size, want_flags, flags)) { free_memtype(paddr, paddr + size); printk(KERN_ERR "%s:%d map pfn expected mapping type %s" " for %Lx-%Lx, got %s\n", -- cgit v0.10.2 From 78b89ecd731798f2fec8cc26ca90739253cec33c Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 18 Aug 2009 16:41:33 +0100 Subject: i386: Fix section mismatches for init code with !HOTPLUG_CPU Commit 0e83815be719d3391bf5ea24b7fe696c07dbd417 changed the section the initial_code variable gets allocated in, in an attempt to address a section conflict warning. This, however created a new section conflict when building without HOTPLUG_CPU. The apparently only (reasonable) way to address this is to always use __REFDATA. Once at it, also fix a second section mismatch when not using HOTPLUG_CPU. Signed-off-by: Jan Beulich Cc: Robert Richter LKML-Reference: <4A8AE7CD020000780001054B@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S index 0d98a01..cc827ac 100644 --- a/arch/x86/kernel/head_32.S +++ b/arch/x86/kernel/head_32.S @@ -261,9 +261,7 @@ page_pde_offset = (__PAGE_OFFSET >> 20); * which will be freed later */ -#ifndef CONFIG_HOTPLUG_CPU -.section .init.text,"ax",@progbits -#endif +__CPUINIT #ifdef CONFIG_SMP ENTRY(startup_32_smp) @@ -602,11 +600,7 @@ ignore_int: #endif iret -#ifndef CONFIG_HOTPLUG_CPU - __CPUINITDATA -#else __REFDATA -#endif .align 4 ENTRY(initial_code) .long i386_start_kernel -- cgit v0.10.2 From 5416c2663517ebd0be0664c4d4ce3df0b116c059 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Mon, 17 Aug 2009 12:25:41 -0700 Subject: x86: make sure load_percpu_segment has no stackprotector load_percpu_segment() is used to set up the per-cpu segment registers, which are also used for -fstack-protector. Make sure that the load_percpu_segment() function doesn't have stackprotector enabled. [ Impact: allow percpu setup before calling stack-protected functions ] Signed-off-by: Jeremy Fitzhardinge diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile index 4e242f9..8b5b9b6 100644 --- a/arch/x86/kernel/cpu/Makefile +++ b/arch/x86/kernel/cpu/Makefile @@ -7,6 +7,10 @@ ifdef CONFIG_FUNCTION_TRACER CFLAGS_REMOVE_common.o = -pg endif +# Make sure load_percpu_segment has no stackprotector +nostackp := $(call cc-option, -fno-stack-protector) +CFLAGS_common.o := $(nostackp) + obj-y := intel_cacheinfo.o addon_cpuid_features.o obj-y += proc.o capflags.o powerflags.o common.o obj-y += vmware.o hypervisor.o -- cgit v0.10.2 From ce2eef33d35cd7b932492b5a81fb0febd2b323cd Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Mon, 17 Aug 2009 12:26:53 -0700 Subject: xen: rearrange things to fix stackprotector Make sure the stack-protector segment registers are properly set up before calling any functions which may have stack-protection compiled into them. [ Impact: prevent Xen early-boot crash when stack-protector is enabled ] Signed-off-by: Jeremy Fitzhardinge diff --git a/arch/x86/xen/Makefile b/arch/x86/xen/Makefile index 3b767d0..a5b9288 100644 --- a/arch/x86/xen/Makefile +++ b/arch/x86/xen/Makefile @@ -5,6 +5,10 @@ CFLAGS_REMOVE_time.o = -pg CFLAGS_REMOVE_irq.o = -pg endif +# Make sure early boot has no stackprotector +nostackp := $(call cc-option, -fno-stack-protector) +CFLAGS_enlighten.o := $(nostackp) + obj-y := enlighten.o setup.o multicalls.o mmu.o irq.o \ time.o xen-asm.o xen-asm_$(BITS).o \ grant-table.o suspend.o diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index f09e8c3..edcf72a 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -925,10 +925,6 @@ asmlinkage void __init xen_start_kernel(void) xen_domain_type = XEN_PV_DOMAIN; - BUG_ON(memcmp(xen_start_info->magic, "xen-3", 5) != 0); - - xen_setup_features(); - /* Install Xen paravirt ops */ pv_info = xen_info; pv_init_ops = xen_init_ops; @@ -937,8 +933,15 @@ asmlinkage void __init xen_start_kernel(void) pv_apic_ops = xen_apic_ops; pv_mmu_ops = xen_mmu_ops; - xen_init_irq_ops(); +#ifdef CONFIG_X86_64 + /* + * Setup percpu state. We only need to do this for 64-bit + * because 32-bit already has %fs set properly. + */ + load_percpu_segment(0); +#endif + xen_init_irq_ops(); xen_init_cpuid_mask(); #ifdef CONFIG_X86_LOCAL_APIC @@ -948,6 +951,8 @@ asmlinkage void __init xen_start_kernel(void) set_xen_basic_apic_ops(); #endif + xen_setup_features(); + if (xen_feature(XENFEAT_mmu_pt_update_preserve_ad)) { pv_mmu_ops.ptep_modify_prot_start = xen_ptep_modify_prot_start; pv_mmu_ops.ptep_modify_prot_commit = xen_ptep_modify_prot_commit; @@ -955,13 +960,6 @@ asmlinkage void __init xen_start_kernel(void) machine_ops = xen_machine_ops; -#ifdef CONFIG_X86_64 - /* - * Setup percpu state. We only need to do this for 64-bit - * because 32-bit already has %fs set properly. - */ - load_percpu_segment(0); -#endif /* * The only reliable way to retain the initial address of the * percpu gdt_page is to remember it here, so we can go and -- cgit v0.10.2 From fc0ce23506d943b9eaa731a051769d0e0605eb03 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 20 Aug 2009 16:14:15 +0100 Subject: x86: add vmlinux.lds to targets in arch/x86/boot/compressed/Makefile The absence of vmlinux.lds here keeps .vmlinux.lds.cmd from being included, which in turn leads to it and all its dependents always getting rebuilt independent of whether they are already up-to-date. Signed-off-by: Jan Beulich LKML-Reference: <4A8D84670200007800010D31@vpn.id2.novell.com> Signed-off-by: H. Peter Anvin diff --git a/arch/x86/boot/compressed/Makefile b/arch/x86/boot/compressed/Makefile index e2ff504..f8ed065 100644 --- a/arch/x86/boot/compressed/Makefile +++ b/arch/x86/boot/compressed/Makefile @@ -4,7 +4,7 @@ # create a compressed vmlinux image from the original vmlinux # -targets := vmlinux vmlinux.bin vmlinux.bin.gz vmlinux.bin.bz2 vmlinux.bin.lzma head_$(BITS).o misc.o piggy.o +targets := vmlinux.lds vmlinux vmlinux.bin vmlinux.bin.gz vmlinux.bin.bz2 vmlinux.bin.lzma head_$(BITS).o misc.o piggy.o KBUILD_CFLAGS := -m$(BITS) -D__KERNEL__ $(LINUX_INCLUDE) -O2 KBUILD_CFLAGS += -fno-strict-aliasing -fPIC -- cgit v0.10.2 From a6a06f7b577f89d0b916c5ccaff67ca5ed444a78 Mon Sep 17 00:00:00 2001 From: Amerigo Wang Date: Fri, 21 Aug 2009 04:34:45 -0400 Subject: x86: Fix an incorrect argument of reserve_bootmem() This line looks suspicious, because if this is true, then the 'flags' parameter of function reserve_bootmem_generic() will be unused when !CONFIG_NUMA. I don't think this is what we want. Signed-off-by: WANG Cong Cc: Yinghai Lu Cc: akpm@linux-foundation.org LKML-Reference: <20090821083709.5098.52505.sendpatchset@localhost.localdomain> Signed-off-by: Ingo Molnar diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index 6176fe8..ea56b8c 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -796,7 +796,7 @@ int __init reserve_bootmem_generic(unsigned long phys, unsigned long len, return ret; #else - reserve_bootmem(phys, len, BOOTMEM_DEFAULT); + reserve_bootmem(phys, len, flags); #endif if (phys+len <= MAX_DMA_PFN*PAGE_SIZE) { -- cgit v0.10.2 From c62e43202e7cf50ca24bce58b255df7bf5de69d0 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 25 Aug 2009 14:50:53 +0100 Subject: x86: Fix build with older binutils and consolidate linker script binutils prior to 2.17 can't deal with the currently possible situation of a new segment following the per-CPU segment, but that new segment being empty - objcopy misplaces the .bss (and perhaps also the .brk) sections outside of any segment. However, the current ordering of sections really just appears to be the effect of cumulative unrelated changes; re-ordering things allows to easily guarantee that the segment following the per-CPU one is non-empty, and at once eliminates the need for the bogus data.init2 segment. Once touching this code, also use the various data section helper macros from include/asm-generic/vmlinux.lds.h. -v2: fix !SMP builds. Signed-off-by: Jan Beulich Cc: LKML-Reference: <4A94085D02000078000119A5@vpn.id2.novell.com> Signed-off-by: Ingo Molnar diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S index 78d185d..9fc1782 100644 --- a/arch/x86/kernel/vmlinux.lds.S +++ b/arch/x86/kernel/vmlinux.lds.S @@ -46,11 +46,10 @@ PHDRS { data PT_LOAD FLAGS(7); /* RWE */ #ifdef CONFIG_X86_64 user PT_LOAD FLAGS(7); /* RWE */ - data.init PT_LOAD FLAGS(7); /* RWE */ #ifdef CONFIG_SMP percpu PT_LOAD FLAGS(7); /* RWE */ #endif - data.init2 PT_LOAD FLAGS(7); /* RWE */ + init PT_LOAD FLAGS(7); /* RWE */ #endif note PT_NOTE FLAGS(0); /* ___ */ } @@ -103,65 +102,43 @@ SECTIONS __stop___ex_table = .; } :text = 0x9090 - RODATA + RO_DATA(PAGE_SIZE) /* Data */ - . = ALIGN(PAGE_SIZE); .data : AT(ADDR(.data) - LOAD_OFFSET) { /* Start of data section */ _sdata = .; - DATA_DATA - CONSTRUCTORS - } :data + + /* init_task */ + INIT_TASK_DATA(THREAD_SIZE) #ifdef CONFIG_X86_32 - /* 32 bit has nosave before _edata */ - . = ALIGN(PAGE_SIZE); - .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) { - __nosave_begin = .; - *(.data.nosave) - . = ALIGN(PAGE_SIZE); - __nosave_end = .; - } + /* 32 bit has nosave before _edata */ + NOSAVE_DATA #endif - . = ALIGN(PAGE_SIZE); - .data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) { - *(.data.page_aligned) + PAGE_ALIGNED_DATA(PAGE_SIZE) *(.data.idt) - } -#ifdef CONFIG_X86_32 - . = ALIGN(32); -#else - . = ALIGN(PAGE_SIZE); - . = ALIGN(CONFIG_X86_L1_CACHE_BYTES); -#endif - .data.cacheline_aligned : - AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) { - *(.data.cacheline_aligned) - } + CACHELINE_ALIGNED_DATA(CONFIG_X86_L1_CACHE_BYTES) - /* rarely changed data like cpu maps */ -#ifdef CONFIG_X86_32 - . = ALIGN(32); -#else - . = ALIGN(CONFIG_X86_INTERNODE_CACHE_BYTES); -#endif - .data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) { - *(.data.read_mostly) + DATA_DATA + CONSTRUCTORS + + /* rarely changed data like cpu maps */ + READ_MOSTLY_DATA(CONFIG_X86_INTERNODE_CACHE_BYTES) /* End of data section */ _edata = .; - } + } :data #ifdef CONFIG_X86_64 #define VSYSCALL_ADDR (-10*1024*1024) -#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data.read_mostly) + \ - SIZEOF(.data.read_mostly) + 4095) & ~(4095)) -#define VSYSCALL_VIRT_ADDR ((ADDR(.data.read_mostly) + \ - SIZEOF(.data.read_mostly) + 4095) & ~(4095)) +#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data) + SIZEOF(.data) + \ + PAGE_SIZE - 1) & ~(PAGE_SIZE - 1)) +#define VSYSCALL_VIRT_ADDR ((ADDR(.data) + SIZEOF(.data) + \ + PAGE_SIZE - 1) & ~(PAGE_SIZE - 1)) #define VLOAD_OFFSET (VSYSCALL_ADDR - VSYSCALL_PHYS_ADDR) #define VLOAD(x) (ADDR(x) - VLOAD_OFFSET) @@ -227,35 +204,29 @@ SECTIONS #endif /* CONFIG_X86_64 */ - /* init_task */ - . = ALIGN(THREAD_SIZE); - .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) { - *(.data.init_task) + /* Init code and data - will be freed after init */ + . = ALIGN(PAGE_SIZE); + .init.begin : AT(ADDR(.init.begin) - LOAD_OFFSET) { + __init_begin = .; /* paired with __init_end */ } -#ifdef CONFIG_X86_64 - :data.init -#endif +#if defined(CONFIG_X86_64) && defined(CONFIG_SMP) /* - * smp_locks might be freed after init - * start/end must be page aligned + * percpu offsets are zero-based on SMP. PERCPU_VADDR() changes the + * output PHDR, so the next output section - .init.text - should + * start another segment - init. */ - . = ALIGN(PAGE_SIZE); - .smp_locks : AT(ADDR(.smp_locks) - LOAD_OFFSET) { - __smp_locks = .; - *(.smp_locks) - __smp_locks_end = .; - . = ALIGN(PAGE_SIZE); - } + PERCPU_VADDR(0, :percpu) +#endif - /* Init code and data - will be freed after init */ - . = ALIGN(PAGE_SIZE); .init.text : AT(ADDR(.init.text) - LOAD_OFFSET) { - __init_begin = .; /* paired with __init_end */ _sinittext = .; INIT_TEXT _einittext = .; } +#ifdef CONFIG_X86_64 + :init +#endif .init.data : AT(ADDR(.init.data) - LOAD_OFFSET) { INIT_DATA @@ -326,17 +297,7 @@ SECTIONS } #endif -#if defined(CONFIG_X86_64) && defined(CONFIG_SMP) - /* - * percpu offsets are zero-based on SMP. PERCPU_VADDR() changes the - * output PHDR, so the next output section - __data_nosave - should - * start another section data.init2. Also, pda should be at the head of - * percpu area. Preallocate it and define the percpu offset symbol - * so that it can be accessed as a percpu variable. - */ - . = ALIGN(PAGE_SIZE); - PERCPU_VADDR(0, :percpu) -#else +#if !defined(CONFIG_X86_64) || !defined(CONFIG_SMP) PERCPU(PAGE_SIZE) #endif @@ -347,15 +308,22 @@ SECTIONS __init_end = .; } + /* + * smp_locks might be freed after init + * start/end must be page aligned + */ + . = ALIGN(PAGE_SIZE); + .smp_locks : AT(ADDR(.smp_locks) - LOAD_OFFSET) { + __smp_locks = .; + *(.smp_locks) + __smp_locks_end = .; + . = ALIGN(PAGE_SIZE); + } + #ifdef CONFIG_X86_64 .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) { - . = ALIGN(PAGE_SIZE); - __nosave_begin = .; - *(.data.nosave) - . = ALIGN(PAGE_SIZE); - __nosave_end = .; - } :data.init2 - /* use another section data.init2, see PERCPU_VADDR() above */ + NOSAVE_DATA + } #endif /* BSS */ -- cgit v0.10.2