diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2016-03-16 21:10:53 (GMT) |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2016-03-16 21:10:53 (GMT) |
commit | 277edbabf6fece057b14fb6db5e3a34e00f42f42 (patch) | |
tree | d33314ae118cf387fa697643d10f1549ba4d6bfe | |
parent | 271ecc5253e2b317d729d366560789cd7f93836c (diff) | |
parent | 0d571b62dd8eb341788599259c3dbc92c0dc8f22 (diff) | |
download | linux-277edbabf6fece057b14fb6db5e3a34e00f42f42.tar.xz |
Merge tag 'pm+acpi-4.6-rc1-1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull power management and ACPI updates from Rafael Wysocki:
"This time the majority of changes go into cpufreq and they are
significant.
First off, the way CPU frequency updates are triggered is different
now. Instead of having to set up and manage a deferrable timer for
each CPU in the system to evaluate and possibly change its frequency
periodically, cpufreq governors set up callbacks to be invoked by the
scheduler on a regular basis (basically on utilization updates). The
"old" governors, "ondemand" and "conservative", still do all of their
work in process context (although that is triggered by the scheduler
now), but intel_pstate does it all in the callback invoked by the
scheduler with no need for any additional asynchronous processing.
Of course, this eliminates the overhead related to the management of
all those timers, but also it allows the cpufreq governor code to be
simplified quite a bit. On top of that, the common code and data
structures used by the "ondemand" and "conservative" governors are
cleaned up and made more straightforward and some long-standing and
quite annoying problems are addressed. In particular, the handling of
governor sysfs attributes is modified and the related locking becomes
more fine grained which allows some concurrency problems to be avoided
(particularly deadlocks with the core cpufreq code).
In principle, the new mechanism for triggering frequency updates
allows utilization information to be passed from the scheduler to
cpufreq. Although the current code doesn't make use of it, in the
works is a new cpufreq governor that will make decisions based on the
scheduler's utilization data. That should allow the scheduler and
cpufreq to work more closely together in the long run.
In addition to the core and governor changes, cpufreq drivers are
updated too. Fixes and optimizations go into intel_pstate, the
cpufreq-dt driver is updated on top of some modification in the
Operating Performance Points (OPP) framework and there are fixes and
other updates in the powernv cpufreq driver.
Apart from the cpufreq updates there is some new ACPICA material,
including a fix for a problem introduced by previous ACPICA updates,
and some less significant changes in the ACPI code, like CPPC code
optimizations, ACPI processor driver cleanups and support for loading
ACPI tables from initrd.
Also updated are the generic power domains framework, the Intel RAPL
power capping driver and the turbostat utility and we have a bunch of
traditional assorted fixes and cleanups.
Specifics:
- Redesign of cpufreq governors and the intel_pstate driver to make
them use callbacks invoked by the scheduler to trigger CPU
frequency evaluation instead of using per-CPU deferrable timers for
that purpose (Rafael Wysocki).
- Reorganization and cleanup of cpufreq governor code to make it more
straightforward and fix some concurrency problems in it (Rafael
Wysocki, Viresh Kumar).
- Cleanup and improvements of locking in the cpufreq core (Viresh
Kumar).
- Assorted cleanups in the cpufreq core (Rafael Wysocki, Viresh
Kumar, Eric Biggers).
- intel_pstate driver updates including fixes, optimizations and a
modification to make it enable enable hardware-coordinated P-state
selection (HWP) by default if supported by the processor (Philippe
Longepe, Srinivas Pandruvada, Rafael Wysocki, Viresh Kumar, Felipe
Franciosi).
- Operating Performance Points (OPP) framework updates to improve its
handling of voltage regulators and device clocks and updates of the
cpufreq-dt driver on top of that (Viresh Kumar, Jon Hunter).
- Updates of the powernv cpufreq driver to fix initialization and
cleanup problems in it and correct its worker thread handling with
respect to CPU offline, new powernv_throttle tracepoint (Shilpasri
Bhat).
- ACPI cpufreq driver optimization and cleanup (Rafael Wysocki).
- ACPICA updates including one fix for a regression introduced by
previos changes in the ACPICA code (Bob Moore, Lv Zheng, David Box,
Colin Ian King).
- Support for installing ACPI tables from initrd (Lv Zheng).
- Optimizations of the ACPI CPPC code (Prashanth Prakash, Ashwin
Chaugule).
- Support for _HID(ACPI0010) devices (ACPI processor containers) and
ACPI processor driver cleanups (Sudeep Holla).
- Support for ACPI-based enumeration of the AMBA bus (Graeme Gregory,
Aleksey Makarov).
- Modification of the ACPI PCI IRQ management code to make it treat
255 in the Interrupt Line register as "not connected" on x86 (as
per the specification) and avoid attempts to use that value as a
valid interrupt vector (Chen Fan).
- ACPI APEI fixes related to resource leaks (Josh Hunt).
- Removal of modularity from a few ACPI drivers (BGRT, GHES,
intel_pmic_crc) that cannot be built as modules in practice (Paul
Gortmaker).
- PNP framework update to make it treat ACPI_RESOURCE_TYPE_SERIAL_BUS
as a valid resource type (Harb Abdulhamid).
- New device ID (future AMD I2C controller) in the ACPI driver for
AMD SoCs (APD) and in the designware I2C driver (Xiangliang Yu).
- Assorted ACPI cleanups (Colin Ian King, Kaiyen Chang, Oleg Drokin).
- cpuidle menu governor optimization to avoid a square root
computation in it (Rasmus Villemoes).
- Fix for potential use-after-free in the generic device properties
framework (Heikki Krogerus).
- Updates of the generic power domains (genpd) framework including
support for multiple power states of a domain, fixes and debugfs
output improvements (Axel Haslam, Jon Hunter, Laurent Pinchart,
Geert Uytterhoeven).
- Intel RAPL power capping driver updates to reduce IPI overhead in
it (Jacob Pan).
- System suspend/hibernation code cleanups (Eric Biggers, Saurabh
Sengar).
- Year 2038 fix for the process freezer (Abhilash Jindal).
- turbostat utility updates including new features (decoding of more
registers and CPUID fields, sub-second intervals support, GFX MHz
and RC6 printout, --out command line option), fixes (syscall jitter
detection and workaround, reductioin of the number of syscalls
made, fixes related to Xeon x200 processors, compiler warning
fixes) and cleanups (Len Brown, Hubert Chrzaniuk, Chen Yu)"
* tag 'pm+acpi-4.6-rc1-1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (182 commits)
tools/power turbostat: bugfix: TDP MSRs print bits fixing
tools/power turbostat: correct output for MSR_NHM_SNB_PKG_CST_CFG_CTL dump
tools/power turbostat: call __cpuid() instead of __get_cpuid()
tools/power turbostat: indicate SMX and SGX support
tools/power turbostat: detect and work around syscall jitter
tools/power turbostat: show GFX%rc6
tools/power turbostat: show GFXMHz
tools/power turbostat: show IRQs per CPU
tools/power turbostat: make fewer systems calls
tools/power turbostat: fix compiler warnings
tools/power turbostat: add --out option for saving output in a file
tools/power turbostat: re-name "%Busy" field to "Busy%"
tools/power turbostat: Intel Xeon x200: fix turbo-ratio decoding
tools/power turbostat: Intel Xeon x200: fix erroneous bclk value
tools/power turbostat: allow sub-sec intervals
ACPI / APEI: ERST: Fixed leaked resources in erst_init
ACPI / APEI: Fix leaked resources
intel_pstate: Do not skip samples partially
intel_pstate: Remove freq calculation from intel_pstate_calc_busy()
intel_pstate: Move intel_pstate_calc_busy() into get_target_pstate_use_performance()
...
101 files changed, 4454 insertions, 2943 deletions
diff --git a/Documentation/cpu-freq/intel-pstate.txt b/Documentation/cpu-freq/intel-pstate.txt index f7b12c0..e6bd1e6 100644 --- a/Documentation/cpu-freq/intel-pstate.txt +++ b/Documentation/cpu-freq/intel-pstate.txt @@ -25,7 +25,7 @@ callback, so cpufreq core can't request a transition to a specific frequency. The driver provides minimum and maximum frequency limits and callbacks to set a policy. The policy in cpufreq sysfs is referred to as the "scaling governor". The cpufreq core can request the driver to operate in any of the two policies: -"performance: and "powersave". The driver decides which frequency to use based +"performance" and "powersave". The driver decides which frequency to use based on the above policy selection considering minimum and maximum frequency limits. The Intel P-State driver falls under the latter category, which implements the diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 5b47acb..def4791 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -193,6 +193,12 @@ bytes respectively. Such letter suffixes can also be entirely omitted. (e.g. thinkpad_acpi, sony_acpi, etc.) instead of the ACPI video.ko driver. + acpi_force_32bit_fadt_addr + force FADT to use 32 bit addresses rather than the + 64 bit X_* addresses. Some firmware have broken 64 + bit addresses for force ACPI ignore these and use + the older legacy 32 bit addresses. + acpica_no_return_repair [HW, ACPI] Disable AML predefined validation mechanism This mechanism can repair the evaluation result to make diff --git a/arch/arm/mach-imx/gpc.c b/arch/arm/mach-imx/gpc.c index cfc696b..fd87205 100644 --- a/arch/arm/mach-imx/gpc.c +++ b/arch/arm/mach-imx/gpc.c @@ -374,8 +374,13 @@ static struct pu_domain imx6q_pu_domain = { .name = "PU", .power_off = imx6q_pm_pu_power_off, .power_on = imx6q_pm_pu_power_on, - .power_off_latency_ns = 25000, - .power_on_latency_ns = 2000000, + .states = { + [0] = { + .power_off_latency_ns = 25000, + .power_on_latency_ns = 2000000, + }, + }, + .state_count = 1, }, }; diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h index 984ab75..2da46ac 100644 --- a/arch/x86/include/asm/msr-index.h +++ b/arch/x86/include/asm/msr-index.h @@ -235,10 +235,10 @@ #define HWP_PACKAGE_LEVEL_REQUEST_BIT (1<<11) /* IA32_HWP_CAPABILITIES */ -#define HWP_HIGHEST_PERF(x) (x & 0xff) -#define HWP_GUARANTEED_PERF(x) ((x & (0xff << 8)) >>8) -#define HWP_MOSTEFFICIENT_PERF(x) ((x & (0xff << 16)) >>16) -#define HWP_LOWEST_PERF(x) ((x & (0xff << 24)) >>24) +#define HWP_HIGHEST_PERF(x) (((x) >> 0) & 0xff) +#define HWP_GUARANTEED_PERF(x) (((x) >> 8) & 0xff) +#define HWP_MOSTEFFICIENT_PERF(x) (((x) >> 16) & 0xff) +#define HWP_LOWEST_PERF(x) (((x) >> 24) & 0xff) /* IA32_HWP_REQUEST */ #define HWP_MIN_PERF(x) (x & 0xff) diff --git a/drivers/acpi/Makefile b/drivers/acpi/Makefile index cb648a4..edeb2d1 100644 --- a/drivers/acpi/Makefile +++ b/drivers/acpi/Makefile @@ -43,6 +43,7 @@ acpi-y += pci_root.o pci_link.o pci_irq.o acpi-y += acpi_lpss.o acpi_apd.o acpi-y += acpi_platform.o acpi-y += acpi_pnp.o +acpi-$(CONFIG_ARM_AMBA) += acpi_amba.o acpi-y += int340x_thermal.o acpi-y += power.o acpi-y += event.o diff --git a/drivers/acpi/acpi_amba.c b/drivers/acpi/acpi_amba.c new file mode 100644 index 0000000..2a61b54 --- /dev/null +++ b/drivers/acpi/acpi_amba.c @@ -0,0 +1,122 @@ + +/* + * ACPI support for platform bus type. + * + * Copyright (C) 2015, Linaro Ltd + * Author: Graeme Gregory <graeme.gregory@linaro.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 <linux/acpi.h> +#include <linux/amba/bus.h> +#include <linux/clkdev.h> +#include <linux/clk-provider.h> +#include <linux/device.h> +#include <linux/err.h> +#include <linux/ioport.h> +#include <linux/kernel.h> +#include <linux/module.h> + +#include "internal.h" + +static const struct acpi_device_id amba_id_list[] = { + {"ARMH0061", 0}, /* PL061 GPIO Device */ + {"", 0}, +}; + +static void amba_register_dummy_clk(void) +{ + static struct clk *amba_dummy_clk; + + /* If clock already registered */ + if (amba_dummy_clk) + return; + + amba_dummy_clk = clk_register_fixed_rate(NULL, "apb_pclk", NULL, + CLK_IS_ROOT, 0); + clk_register_clkdev(amba_dummy_clk, "apb_pclk", NULL); +} + +static int amba_handler_attach(struct acpi_device *adev, + const struct acpi_device_id *id) +{ + struct amba_device *dev; + struct resource_entry *rentry; + struct list_head resource_list; + bool address_found = false; + int irq_no = 0; + int ret; + + /* If the ACPI node already has a physical device attached, skip it. */ + if (adev->physical_node_count) + return 0; + + dev = amba_device_alloc(dev_name(&adev->dev), 0, 0); + if (!dev) { + dev_err(&adev->dev, "%s(): amba_device_alloc() failed\n", + __func__); + return -ENOMEM; + } + + INIT_LIST_HEAD(&resource_list); + ret = acpi_dev_get_resources(adev, &resource_list, NULL, NULL); + if (ret < 0) + goto err_free; + + list_for_each_entry(rentry, &resource_list, node) { + switch (resource_type(rentry->res)) { + case IORESOURCE_MEM: + if (!address_found) { + dev->res = *rentry->res; + address_found = true; + } + break; + case IORESOURCE_IRQ: + if (irq_no < AMBA_NR_IRQS) + dev->irq[irq_no++] = rentry->res->start; + break; + default: + dev_warn(&adev->dev, "Invalid resource\n"); + break; + } + } + + acpi_dev_free_resource_list(&resource_list); + + /* + * If the ACPI node has a parent and that parent has a physical device + * attached to it, that physical device should be the parent of + * the amba device we are about to create. + */ + if (adev->parent) + dev->dev.parent = acpi_get_first_physical_node(adev->parent); + + ACPI_COMPANION_SET(&dev->dev, adev); + + ret = amba_device_add(dev, &iomem_resource); + if (ret) { + dev_err(&adev->dev, "%s(): amba_device_add() failed (%d)\n", + __func__, ret); + goto err_free; + } + + return 1; + +err_free: + amba_device_put(dev); + return ret; +} + +static struct acpi_scan_handler amba_handler = { + .ids = amba_id_list, + .attach = amba_handler_attach, +}; + +void __init acpi_amba_init(void) +{ + amba_register_dummy_clk(); + acpi_scan_add_handler(&amba_handler); +} diff --git a/drivers/acpi/acpi_apd.c b/drivers/acpi/acpi_apd.c index d507cf6..d0aad06 100644 --- a/drivers/acpi/acpi_apd.c +++ b/drivers/acpi/acpi_apd.c @@ -143,6 +143,7 @@ static const struct acpi_device_id acpi_apd_device_ids[] = { /* Generic apd devices */ #ifdef CONFIG_X86_AMD_PLATFORM_DEVICE { "AMD0010", APD_ADDR(cz_i2c_desc) }, + { "AMDI0010", APD_ADDR(cz_i2c_desc) }, { "AMD0020", APD_ADDR(cz_uart_desc) }, { "AMD0030", }, #endif diff --git a/drivers/acpi/acpi_platform.c b/drivers/acpi/acpi_platform.c index b6f7fa3..159f7f1 100644 --- a/drivers/acpi/acpi_platform.c +++ b/drivers/acpi/acpi_platform.c @@ -43,7 +43,6 @@ static const struct acpi_device_id forbidden_id_list[] = { struct platform_device *acpi_create_platform_device(struct acpi_device *adev) { struct platform_device *pdev = NULL; - struct acpi_device *acpi_parent; struct platform_device_info pdevinfo; struct resource_entry *rentry; struct list_head resource_list; @@ -82,22 +81,8 @@ struct platform_device *acpi_create_platform_device(struct acpi_device *adev) * attached to it, that physical device should be the parent of the * platform device we are about to create. */ - pdevinfo.parent = NULL; - acpi_parent = adev->parent; - if (acpi_parent) { - struct acpi_device_physical_node *entry; - struct list_head *list; - - mutex_lock(&acpi_parent->physical_node_lock); - list = &acpi_parent->physical_node_list; - if (!list_empty(list)) { - entry = list_first_entry(list, - struct acpi_device_physical_node, - node); - pdevinfo.parent = entry->dev; - } - mutex_unlock(&acpi_parent->physical_node_lock); - } + pdevinfo.parent = adev->parent ? + acpi_get_first_physical_node(adev->parent) : NULL; pdevinfo.name = dev_name(&adev->dev); pdevinfo.id = -1; pdevinfo.res = resources; diff --git a/drivers/acpi/acpi_processor.c b/drivers/acpi/acpi_processor.c index 6979186..b5e54f2 100644 --- a/drivers/acpi/acpi_processor.c +++ b/drivers/acpi/acpi_processor.c @@ -514,7 +514,24 @@ static struct acpi_scan_handler processor_handler = { }, }; +static int acpi_processor_container_attach(struct acpi_device *dev, + const struct acpi_device_id *id) +{ + return 1; +} + +static const struct acpi_device_id processor_container_ids[] = { + { ACPI_PROCESSOR_CONTAINER_HID, }, + { } +}; + +static struct acpi_scan_handler processor_container_handler = { + .ids = processor_container_ids, + .attach = acpi_processor_container_attach, +}; + void __init acpi_processor_init(void) { acpi_scan_add_handler_with_hotplug(&processor_handler, "processor"); + acpi_scan_add_handler(&processor_container_handler); } diff --git a/drivers/acpi/acpi_video.c b/drivers/acpi/acpi_video.c index a76f8be..4361bc9 100644 --- a/drivers/acpi/acpi_video.c +++ b/drivers/acpi/acpi_video.c @@ -218,13 +218,6 @@ struct acpi_video_device { struct thermal_cooling_device *cooling_dev; }; -static const char device_decode[][30] = { - "motherboard VGA device", - "PCI VGA device", - "AGP VGA device", - "UNKNOWN", -}; - static void acpi_video_device_notify(acpi_handle handle, u32 event, void *data); static void acpi_video_device_rebind(struct acpi_video_bus *video); static void acpi_video_device_bind(struct acpi_video_bus *video, diff --git a/drivers/acpi/acpica/acglobal.h b/drivers/acpi/acpica/acglobal.h index 55c8197..51b073b 100644 --- a/drivers/acpi/acpica/acglobal.h +++ b/drivers/acpi/acpica/acglobal.h @@ -165,7 +165,7 @@ ACPI_GLOBAL(u8, acpi_gbl_next_owner_id_offset); /* Initialization sequencing */ -ACPI_INIT_GLOBAL(u8, acpi_gbl_reg_methods_enabled, FALSE); +ACPI_INIT_GLOBAL(u8, acpi_gbl_namespace_initialized, FALSE); /* Misc */ diff --git a/drivers/acpi/acpica/aclocal.h b/drivers/acpi/acpica/aclocal.h index e4977fa..9562a10 100644 --- a/drivers/acpi/acpica/aclocal.h +++ b/drivers/acpi/acpica/aclocal.h @@ -85,7 +85,7 @@ union acpi_parse_object; #define ACPI_MTX_MEMORY 5 /* Debug memory tracking lists */ #define ACPI_MAX_MUTEX 5 -#define ACPI_NUM_MUTEX ACPI_MAX_MUTEX+1 +#define ACPI_NUM_MUTEX (ACPI_MAX_MUTEX+1) /* Lock structure for reader/writer interfaces */ @@ -103,11 +103,11 @@ struct acpi_rw_lock { #define ACPI_LOCK_HARDWARE 1 #define ACPI_MAX_LOCK 1 -#define ACPI_NUM_LOCK ACPI_MAX_LOCK+1 +#define ACPI_NUM_LOCK (ACPI_MAX_LOCK+1) /* This Thread ID means that the mutex is not in use (unlocked) */ -#define ACPI_MUTEX_NOT_ACQUIRED (acpi_thread_id) 0 +#define ACPI_MUTEX_NOT_ACQUIRED ((acpi_thread_id) 0) /* This Thread ID means an invalid thread ID */ diff --git a/drivers/acpi/acpica/acnamesp.h b/drivers/acpi/acpica/acnamesp.h index 9684ed6..022d69c 100644 --- a/drivers/acpi/acpica/acnamesp.h +++ b/drivers/acpi/acpica/acnamesp.h @@ -88,7 +88,7 @@ */ acpi_status acpi_ns_initialize_objects(void); -acpi_status acpi_ns_initialize_devices(void); +acpi_status acpi_ns_initialize_devices(u32 flags); /* * nsload - Namespace loading diff --git a/drivers/acpi/acpica/acpredef.h b/drivers/acpi/acpica/acpredef.h index 52f6bee..5faeab4 100644 --- a/drivers/acpi/acpica/acpredef.h +++ b/drivers/acpi/acpica/acpredef.h @@ -1125,7 +1125,7 @@ const union acpi_predefined_info acpi_gbl_resource_names[] = { PACKAGE_INFO(0, 0, 0, 0, 0, 0) /* Table terminator */ }; -static const union acpi_predefined_info acpi_gbl_scope_names[] = { +const union acpi_predefined_info acpi_gbl_scope_names[] = { {{"_GPE", 0, 0}}, {{"_PR_", 0, 0}}, {{"_SB_", 0, 0}}, diff --git a/drivers/acpi/acpica/dbcmds.c b/drivers/acpi/acpica/dbcmds.c index 7ec62c4..772178c 100644 --- a/drivers/acpi/acpica/dbcmds.c +++ b/drivers/acpi/acpica/dbcmds.c @@ -348,7 +348,7 @@ void acpi_db_display_table_info(char *table_arg) } else { /* If the pointer is null, the table has been unloaded */ - ACPI_INFO((AE_INFO, "%4.4s - Table has been unloaded", + ACPI_INFO(("%4.4s - Table has been unloaded", table_desc->signature.ascii)); } } diff --git a/drivers/acpi/acpica/dbconvert.c b/drivers/acpi/acpica/dbconvert.c index 9fee88f..68f4e0f4 100644 --- a/drivers/acpi/acpica/dbconvert.c +++ b/drivers/acpi/acpica/dbconvert.c @@ -408,7 +408,7 @@ void acpi_db_dump_pld_buffer(union acpi_object *obj_desc) new_buffer = acpi_db_encode_pld_buffer(pld_info); if (!new_buffer) { - return; + goto exit; } /* The two bit-packed buffers should match */ @@ -479,6 +479,7 @@ void acpi_db_dump_pld_buffer(union acpi_object *obj_desc) pld_info->horizontal_offset); } - ACPI_FREE(pld_info); ACPI_FREE(new_buffer); +exit: + ACPI_FREE(pld_info); } diff --git a/drivers/acpi/acpica/dsmethod.c b/drivers/acpi/acpica/dsmethod.c index 6a72047..1982310 100644 --- a/drivers/acpi/acpica/dsmethod.c +++ b/drivers/acpi/acpica/dsmethod.c @@ -809,8 +809,7 @@ acpi_ds_terminate_control_method(union acpi_operand_object *method_desc, if (method_desc->method. info_flags & ACPI_METHOD_SERIALIZED_PENDING) { if (walk_state) { - ACPI_INFO((AE_INFO, - "Marking method %4.4s as Serialized " + ACPI_INFO(("Marking method %4.4s as Serialized " "because of AE_ALREADY_EXISTS error", walk_state->method_node->name. ascii)); diff --git a/drivers/acpi/acpica/dsobject.c b/drivers/acpi/acpica/dsobject.c index c303e9d..a91de2b 100644 --- a/drivers/acpi/acpica/dsobject.c +++ b/drivers/acpi/acpica/dsobject.c @@ -524,8 +524,7 @@ acpi_ds_build_internal_package_obj(struct acpi_walk_state *walk_state, arg = arg->common.next; } - ACPI_INFO((AE_INFO, - "Actual Package length (%u) is larger than " + ACPI_INFO(("Actual Package length (%u) is larger than " "NumElements field (%u), truncated", i, element_count)); } else if (i < element_count) { diff --git a/drivers/acpi/acpica/evgpeblk.c b/drivers/acpi/acpica/evgpeblk.c index 9275e62..447fa1c 100644 --- a/drivers/acpi/acpica/evgpeblk.c +++ b/drivers/acpi/acpica/evgpeblk.c @@ -499,8 +499,7 @@ acpi_ev_initialize_gpe_block(struct acpi_gpe_xrupt_info *gpe_xrupt_info, } if (gpe_enabled_count) { - ACPI_INFO((AE_INFO, - "Enabled %u GPEs in block %02X to %02X", + ACPI_INFO(("Enabled %u GPEs in block %02X to %02X", gpe_enabled_count, (u32)gpe_block->block_base_number, (u32)(gpe_block->block_base_number + (gpe_block->gpe_count - 1)))); diff --git a/drivers/acpi/acpica/evgpeinit.c b/drivers/acpi/acpica/evgpeinit.c index 9fdd8d0..7dc7547 100644 --- a/drivers/acpi/acpica/evgpeinit.c +++ b/drivers/acpi/acpica/evgpeinit.c @@ -281,7 +281,7 @@ void acpi_ev_update_gpes(acpi_owner_id table_owner_id) } if (walk_info.count) { - ACPI_INFO((AE_INFO, "Enabled %u new GPEs", walk_info.count)); + ACPI_INFO(("Enabled %u new GPEs", walk_info.count)); } (void)acpi_ut_release_mutex(ACPI_MTX_EVENTS); diff --git a/drivers/acpi/acpica/evregion.c b/drivers/acpi/acpica/evregion.c index 47092b4..63924d1 100644 --- a/drivers/acpi/acpica/evregion.c +++ b/drivers/acpi/acpica/evregion.c @@ -600,7 +600,7 @@ acpi_ev_execute_reg_method(union acpi_operand_object *region_obj, u32 function) if (region_obj2->extra.method_REG == NULL || region_obj->region.handler == NULL || - !acpi_gbl_reg_methods_enabled) { + !acpi_gbl_namespace_initialized) { return_ACPI_STATUS(AE_OK); } diff --git a/drivers/acpi/acpica/exconfig.c b/drivers/acpi/acpica/exconfig.c index 011df21..f741613 100644 --- a/drivers/acpi/acpica/exconfig.c +++ b/drivers/acpi/acpica/exconfig.c @@ -252,7 +252,7 @@ acpi_ex_load_table_op(struct acpi_walk_state *walk_state, status = acpi_get_table_by_index(table_index, &table); if (ACPI_SUCCESS(status)) { - ACPI_INFO((AE_INFO, "Dynamic OEM Table Load:")); + ACPI_INFO(("Dynamic OEM Table Load:")); acpi_tb_print_table_header(0, table); } @@ -472,7 +472,7 @@ acpi_ex_load_op(union acpi_operand_object *obj_desc, /* Install the new table into the local data structures */ - ACPI_INFO((AE_INFO, "Dynamic OEM Table Load:")); + ACPI_INFO(("Dynamic OEM Table Load:")); (void)acpi_ut_acquire_mutex(ACPI_MTX_TABLES); status = acpi_tb_install_standard_table(ACPI_PTR_TO_PHYSADDR(table), diff --git a/drivers/acpi/acpica/exoparg3.c b/drivers/acpi/acpica/exoparg3.c index 28eb861..5aa21c4 100644 --- a/drivers/acpi/acpica/exoparg3.c +++ b/drivers/acpi/acpica/exoparg3.c @@ -123,8 +123,10 @@ acpi_status acpi_ex_opcode_3A_0T_0R(struct acpi_walk_state *walk_state) * op is intended for use by disassemblers in order to properly * disassemble control method invocations. The opcode or group of * opcodes should be surrounded by an "if (0)" clause to ensure that - * AML interpreters never see the opcode. + * AML interpreters never see the opcode. Thus, something is + * wrong if an external opcode ever gets here. */ + ACPI_ERROR((AE_INFO, "Executed External Op")); status = AE_OK; goto cleanup; diff --git a/drivers/acpi/acpica/nseval.c b/drivers/acpi/acpica/nseval.c index 65d58be..5d59cfc 100644 --- a/drivers/acpi/acpica/nseval.c +++ b/drivers/acpi/acpica/nseval.c @@ -378,8 +378,7 @@ void acpi_ns_exec_module_code_list(void) acpi_ut_remove_reference(prev); } - ACPI_INFO((AE_INFO, - "Executed %u blocks of module-level executable AML code", + ACPI_INFO(("Executed %u blocks of module-level executable AML code", method_count)); ACPI_FREE(info); diff --git a/drivers/acpi/acpica/nsinit.c b/drivers/acpi/acpica/nsinit.c index bd75d46..d4aa8b6 100644 --- a/drivers/acpi/acpica/nsinit.c +++ b/drivers/acpi/acpica/nsinit.c @@ -46,6 +46,7 @@ #include "acnamesp.h" #include "acdispat.h" #include "acinterp.h" +#include "acevents.h" #define _COMPONENT ACPI_NAMESPACE ACPI_MODULE_NAME("nsinit") @@ -83,6 +84,8 @@ acpi_status acpi_ns_initialize_objects(void) ACPI_FUNCTION_TRACE(ns_initialize_objects); + ACPI_DEBUG_PRINT((ACPI_DB_EXEC, + "[Init] Completing Initialization of ACPI Objects\n")); ACPI_DEBUG_PRINT((ACPI_DB_DISPATCH, "**** Starting initialization of namespace objects ****\n")); ACPI_DEBUG_PRINT_RAW((ACPI_DB_INIT, @@ -133,82 +136,108 @@ acpi_status acpi_ns_initialize_objects(void) * ******************************************************************************/ -acpi_status acpi_ns_initialize_devices(void) +acpi_status acpi_ns_initialize_devices(u32 flags) { - acpi_status status; + acpi_status status = AE_OK; struct acpi_device_walk_info info; ACPI_FUNCTION_TRACE(ns_initialize_devices); - /* Init counters */ + if (!(flags & ACPI_NO_DEVICE_INIT)) { + ACPI_DEBUG_PRINT((ACPI_DB_EXEC, + "[Init] Initializing ACPI Devices\n")); - info.device_count = 0; - info.num_STA = 0; - info.num_INI = 0; + /* Init counters */ - ACPI_DEBUG_PRINT_RAW((ACPI_DB_INIT, - "Initializing Device/Processor/Thermal objects " - "and executing _INI/_STA methods:\n")); + info.device_count = 0; + info.num_STA = 0; + info.num_INI = 0; - /* Tree analysis: find all subtrees that contain _INI methods */ + ACPI_DEBUG_PRINT_RAW((ACPI_DB_INIT, + "Initializing Device/Processor/Thermal objects " + "and executing _INI/_STA methods:\n")); - status = acpi_ns_walk_namespace(ACPI_TYPE_ANY, ACPI_ROOT_OBJECT, - ACPI_UINT32_MAX, FALSE, - acpi_ns_find_ini_methods, NULL, &info, - NULL); - if (ACPI_FAILURE(status)) { - goto error_exit; - } + /* Tree analysis: find all subtrees that contain _INI methods */ + + status = acpi_ns_walk_namespace(ACPI_TYPE_ANY, ACPI_ROOT_OBJECT, + ACPI_UINT32_MAX, FALSE, + acpi_ns_find_ini_methods, NULL, + &info, NULL); + if (ACPI_FAILURE(status)) { + goto error_exit; + } + + /* Allocate the evaluation information block */ - /* Allocate the evaluation information block */ + info.evaluate_info = + ACPI_ALLOCATE_ZEROED(sizeof(struct acpi_evaluate_info)); + if (!info.evaluate_info) { + status = AE_NO_MEMORY; + goto error_exit; + } - info.evaluate_info = - ACPI_ALLOCATE_ZEROED(sizeof(struct acpi_evaluate_info)); - if (!info.evaluate_info) { - status = AE_NO_MEMORY; - goto error_exit; + /* + * Execute the "global" _INI method that may appear at the root. + * This support is provided for Windows compatibility (Vista+) and + * is not part of the ACPI specification. + */ + info.evaluate_info->prefix_node = acpi_gbl_root_node; + info.evaluate_info->relative_pathname = METHOD_NAME__INI; + info.evaluate_info->parameters = NULL; + info.evaluate_info->flags = ACPI_IGNORE_RETURN_VALUE; + + status = acpi_ns_evaluate(info.evaluate_info); + if (ACPI_SUCCESS(status)) { + info.num_INI++; + } } /* - * Execute the "global" _INI method that may appear at the root. This - * support is provided for Windows compatibility (Vista+) and is not - * part of the ACPI specification. + * Run all _REG methods + * + * Note: Any objects accessed by the _REG methods will be automatically + * initialized, even if they contain executable AML (see the call to + * acpi_ns_initialize_objects below). */ - info.evaluate_info->prefix_node = acpi_gbl_root_node; - info.evaluate_info->relative_pathname = METHOD_NAME__INI; - info.evaluate_info->parameters = NULL; - info.evaluate_info->flags = ACPI_IGNORE_RETURN_VALUE; + if (!(flags & ACPI_NO_ADDRESS_SPACE_INIT)) { + ACPI_DEBUG_PRINT((ACPI_DB_EXEC, + "[Init] Executing _REG OpRegion methods\n")); - status = acpi_ns_evaluate(info.evaluate_info); - if (ACPI_SUCCESS(status)) { - info.num_INI++; + status = acpi_ev_initialize_op_regions(); + if (ACPI_FAILURE(status)) { + goto error_exit; + } } - /* Walk namespace to execute all _INIs on present devices */ + if (!(flags & ACPI_NO_DEVICE_INIT)) { - status = acpi_ns_walk_namespace(ACPI_TYPE_ANY, ACPI_ROOT_OBJECT, - ACPI_UINT32_MAX, FALSE, - acpi_ns_init_one_device, NULL, &info, - NULL); + /* Walk namespace to execute all _INIs on present devices */ - /* - * Any _OSI requests should be completed by now. If the BIOS has - * requested any Windows OSI strings, we will always truncate - * I/O addresses to 16 bits -- for Windows compatibility. - */ - if (acpi_gbl_osi_data >= ACPI_OSI_WIN_2000) { - acpi_gbl_truncate_io_addresses = TRUE; - } + status = acpi_ns_walk_namespace(ACPI_TYPE_ANY, ACPI_ROOT_OBJECT, + ACPI_UINT32_MAX, FALSE, + acpi_ns_init_one_device, NULL, + &info, NULL); - ACPI_FREE(info.evaluate_info); - if (ACPI_FAILURE(status)) { - goto error_exit; - } + /* + * Any _OSI requests should be completed by now. If the BIOS has + * requested any Windows OSI strings, we will always truncate + * I/O addresses to 16 bits -- for Windows compatibility. + */ + if (acpi_gbl_osi_data >= ACPI_OSI_WIN_2000) { + acpi_gbl_truncate_io_addresses = TRUE; + } - ACPI_DEBUG_PRINT_RAW((ACPI_DB_INIT, - " Executed %u _INI methods requiring %u _STA executions " - "(examined %u objects)\n", - info.num_INI, info.num_STA, info.device_count)); + ACPI_FREE(info.evaluate_info); + if (ACPI_FAILURE(status)) { + goto error_exit; + } + + ACPI_DEBUG_PRINT_RAW((ACPI_DB_INIT, + " Executed %u _INI methods requiring %u _STA executions " + "(examined %u objects)\n", + info.num_INI, info.num_STA, + info.device_count)); + } return_ACPI_STATUS(status); diff --git a/drivers/acpi/acpica/tbinstal.c b/drivers/acpi/acpica/tbinstal.c index b661a1e..4dc6108 100644 --- a/drivers/acpi/acpica/tbinstal.c +++ b/drivers/acpi/acpica/tbinstal.c @@ -267,8 +267,7 @@ acpi_tb_install_standard_table(acpi_physical_address address, if (!reload && acpi_gbl_disable_ssdt_table_install && ACPI_COMPARE_NAME(&new_table_desc.signature, ACPI_SIG_SSDT)) { - ACPI_INFO((AE_INFO, - "Ignoring installation of %4.4s at %8.8X%8.8X", + ACPI_INFO(("Ignoring installation of %4.4s at %8.8X%8.8X", new_table_desc.signature.ascii, ACPI_FORMAT_UINT64(address))); goto release_and_exit; @@ -432,7 +431,7 @@ finish_override: return; } - ACPI_INFO((AE_INFO, "%4.4s 0x%8.8X%8.8X" + ACPI_INFO(("%4.4s 0x%8.8X%8.8X" " %s table override, new table: 0x%8.8X%8.8X", old_table_desc->signature.ascii, ACPI_FORMAT_UINT64(old_table_desc->address), diff --git a/drivers/acpi/acpica/tbprint.c b/drivers/acpi/acpica/tbprint.c index fd4146d..26d61db 100644 --- a/drivers/acpi/acpica/tbprint.c +++ b/drivers/acpi/acpica/tbprint.c @@ -132,7 +132,7 @@ acpi_tb_print_table_header(acpi_physical_address address, /* FACS only has signature and length fields */ - ACPI_INFO((AE_INFO, "%-4.4s 0x%8.8X%8.8X %06X", + ACPI_INFO(("%-4.4s 0x%8.8X%8.8X %06X", header->signature, ACPI_FORMAT_UINT64(address), header->length)); } else if (ACPI_VALIDATE_RSDP_SIG(header->signature)) { @@ -144,7 +144,7 @@ acpi_tb_print_table_header(acpi_physical_address address, ACPI_OEM_ID_SIZE); acpi_tb_fix_string(local_header.oem_id, ACPI_OEM_ID_SIZE); - ACPI_INFO((AE_INFO, "RSDP 0x%8.8X%8.8X %06X (v%.2d %-6.6s)", + ACPI_INFO(("RSDP 0x%8.8X%8.8X %06X (v%.2d %-6.6s)", ACPI_FORMAT_UINT64(address), (ACPI_CAST_PTR(struct acpi_table_rsdp, header)-> revision > @@ -158,8 +158,7 @@ acpi_tb_print_table_header(acpi_physical_address address, acpi_tb_cleanup_table_header(&local_header, header); - ACPI_INFO((AE_INFO, - "%-4.4s 0x%8.8X%8.8X" + ACPI_INFO(("%-4.4s 0x%8.8X%8.8X" " %06X (v%.2d %-6.6s %-8.8s %08X %-4.4s %08X)", local_header.signature, ACPI_FORMAT_UINT64(address), local_header.length, local_header.revision, diff --git a/drivers/acpi/acpica/tbutils.c b/drivers/acpi/acpica/tbutils.c index 3269bef..9240c76 100644 --- a/drivers/acpi/acpica/tbutils.c +++ b/drivers/acpi/acpica/tbutils.c @@ -174,9 +174,7 @@ struct acpi_table_header *acpi_tb_copy_dsdt(u32 table_index) ACPI_TABLE_ORIGIN_INTERNAL_VIRTUAL, new_table); - ACPI_INFO((AE_INFO, - "Forced DSDT copy: length 0x%05X copied locally, original unmapped", - new_table->length)); + ACPI_INFO(("Forced DSDT copy: length 0x%05X copied locally, original unmapped", new_table->length)); return (new_table); } diff --git a/drivers/acpi/acpica/tbxfload.c b/drivers/acpi/acpica/tbxfload.c index 278666e..3151968 100644 --- a/drivers/acpi/acpica/tbxfload.c +++ b/drivers/acpi/acpica/tbxfload.c @@ -47,6 +47,7 @@ #include "accommon.h" #include "acnamesp.h" #include "actables.h" +#include "acevents.h" #define _COMPONENT ACPI_TABLES ACPI_MODULE_NAME("tbxfload") @@ -68,6 +69,25 @@ acpi_status __init acpi_load_tables(void) ACPI_FUNCTION_TRACE(acpi_load_tables); + /* + * Install the default operation region handlers. These are the + * handlers that are defined by the ACPI specification to be + * "always accessible" -- namely, system_memory, system_IO, and + * PCI_Config. This also means that no _REG methods need to be + * run for these address spaces. We need to have these handlers + * installed before any AML code can be executed, especially any + * module-level code (11/2015). + * Note that we allow OSPMs to install their own region handlers + * between acpi_initialize_subsystem() and acpi_load_tables() to use + * their customized default region handlers. + */ + status = acpi_ev_install_region_handlers(); + if (ACPI_FAILURE(status) && status != AE_ALREADY_EXISTS) { + ACPI_EXCEPTION((AE_INFO, status, + "During Region initialization")); + return_ACPI_STATUS(status); + } + /* Load the namespace from the tables */ status = acpi_tb_load_namespace(); @@ -83,6 +103,20 @@ acpi_status __init acpi_load_tables(void) "While loading namespace from ACPI tables")); } + if (!acpi_gbl_group_module_level_code) { + /* + * Initialize the objects that remain uninitialized. This + * runs the executable AML that may be part of the + * declaration of these objects: + * operation_regions, buffer_fields, Buffers, and Packages. + */ + status = acpi_ns_initialize_objects(); + if (ACPI_FAILURE(status)) { + return_ACPI_STATUS(status); + } + } + + acpi_gbl_namespace_initialized = TRUE; return_ACPI_STATUS(status); } @@ -206,9 +240,7 @@ acpi_status acpi_tb_load_namespace(void) } if (!tables_failed) { - ACPI_INFO((AE_INFO, - "%u ACPI AML tables successfully acquired and loaded\n", - tables_loaded)); + ACPI_INFO(("%u ACPI AML tables successfully acquired and loaded\n", tables_loaded)); } else { ACPI_ERROR((AE_INFO, "%u table load failures, %u successful", @@ -301,7 +333,7 @@ acpi_status acpi_load_table(struct acpi_table_header *table) /* Install the table and load it into the namespace */ - ACPI_INFO((AE_INFO, "Host-directed Dynamic ACPI Table Load:")); + ACPI_INFO(("Host-directed Dynamic ACPI Table Load:")); (void)acpi_ut_acquire_mutex(ACPI_MTX_TABLES); status = acpi_tb_install_standard_table(ACPI_PTR_TO_PHYSADDR(table), diff --git a/drivers/acpi/acpica/utcache.c b/drivers/acpi/acpica/utcache.c index c9a720f..f8e9978 100644 --- a/drivers/acpi/acpica/utcache.c +++ b/drivers/acpi/acpica/utcache.c @@ -245,7 +245,7 @@ void *acpi_os_acquire_object(struct acpi_memory_list *cache) acpi_status status; void *object; - ACPI_FUNCTION_NAME(os_acquire_object); + ACPI_FUNCTION_TRACE(os_acquire_object); if (!cache) { return_PTR(NULL); diff --git a/drivers/acpi/acpica/utnonansi.c b/drivers/acpi/acpica/utnonansi.c index c427a5c..d5c3adf 100644 --- a/drivers/acpi/acpica/utnonansi.c +++ b/drivers/acpi/acpica/utnonansi.c @@ -140,6 +140,67 @@ int acpi_ut_stricmp(char *string1, char *string2) return (c1 - c2); } +#if defined (ACPI_DEBUGGER) || defined (ACPI_APPLICATION) +/******************************************************************************* + * + * FUNCTION: acpi_ut_safe_strcpy, acpi_ut_safe_strcat, acpi_ut_safe_strncat + * + * PARAMETERS: Adds a "DestSize" parameter to each of the standard string + * functions. This is the size of the Destination buffer. + * + * RETURN: TRUE if the operation would overflow the destination buffer. + * + * DESCRIPTION: Safe versions of standard Clib string functions. Ensure that + * the result of the operation will not overflow the output string + * buffer. + * + * NOTE: These functions are typically only helpful for processing + * user input and command lines. For most ACPICA code, the + * required buffer length is precisely calculated before buffer + * allocation, so the use of these functions is unnecessary. + * + ******************************************************************************/ + +u8 acpi_ut_safe_strcpy(char *dest, acpi_size dest_size, char *source) +{ + + if (strlen(source) >= dest_size) { + return (TRUE); + } + + strcpy(dest, source); + return (FALSE); +} + +u8 acpi_ut_safe_strcat(char *dest, acpi_size dest_size, char *source) +{ + + if ((strlen(dest) + strlen(source)) >= dest_size) { + return (TRUE); + } + + strcat(dest, source); + return (FALSE); +} + +u8 +acpi_ut_safe_strncat(char *dest, + acpi_size dest_size, + char *source, acpi_size max_transfer_length) +{ + acpi_size actual_transfer_length; + + actual_transfer_length = ACPI_MIN(max_transfer_length, strlen(source)); + + if ((strlen(dest) + actual_transfer_length) >= dest_size) { + return (TRUE); + } + + strncat(dest, source, max_transfer_length); + return (FALSE); +} +#endif + /******************************************************************************* * * FUNCTION: acpi_ut_strtoul64 @@ -155,7 +216,15 @@ int acpi_ut_stricmp(char *string1, char *string2) * 32-bit or 64-bit conversion, depending on the current mode * of the interpreter. * - * NOTE: Does not support Octal strings, not needed. + * NOTES: acpi_gbl_integer_byte_width should be set to the proper width. + * For the core ACPICA code, this width depends on the DSDT + * version. For iASL, the default byte width is always 8. + * + * Does not support Octal strings, not needed at this time. + * + * There is an earlier version of the function after this one, + * below. It is slightly different than this one, and the two + * may eventually may need to be merged. (01/2016). * ******************************************************************************/ @@ -171,7 +240,7 @@ acpi_status acpi_ut_strtoul64(char *string, u32 base, u64 *ret_integer) u8 sign_of0x = 0; u8 term = 0; - ACPI_FUNCTION_TRACE_STR(ut_stroul64, string); + ACPI_FUNCTION_TRACE_STR(ut_strtoul64, string); switch (base) { case ACPI_ANY_BASE: @@ -318,63 +387,162 @@ error_exit: } } -#if defined (ACPI_DEBUGGER) || defined (ACPI_APPLICATION) +#ifdef _OBSOLETE_FUNCTIONS +/* TBD: use version in ACPICA main code base? */ +/* DONE: 01/2016 */ + /******************************************************************************* * - * FUNCTION: acpi_ut_safe_strcpy, acpi_ut_safe_strcat, acpi_ut_safe_strncat + * FUNCTION: strtoul64 * - * PARAMETERS: Adds a "DestSize" parameter to each of the standard string - * functions. This is the size of the Destination buffer. + * PARAMETERS: string - Null terminated string + * terminater - Where a pointer to the terminating byte + * is returned + * base - Radix of the string * - * RETURN: TRUE if the operation would overflow the destination buffer. + * RETURN: Converted value * - * DESCRIPTION: Safe versions of standard Clib string functions. Ensure that - * the result of the operation will not overflow the output string - * buffer. - * - * NOTE: These functions are typically only helpful for processing - * user input and command lines. For most ACPICA code, the - * required buffer length is precisely calculated before buffer - * allocation, so the use of these functions is unnecessary. + * DESCRIPTION: Convert a string into an unsigned value. * ******************************************************************************/ -u8 acpi_ut_safe_strcpy(char *dest, acpi_size dest_size, char *source) +acpi_status strtoul64(char *string, u32 base, u64 *ret_integer) { + u32 index; + u32 sign; + u64 return_value = 0; + acpi_status status = AE_OK; - if (strlen(source) >= dest_size) { - return (TRUE); + *ret_integer = 0; + + switch (base) { + case 0: + case 8: + case 10: + case 16: + + break; + + default: + /* + * The specified Base parameter is not in the domain of + * this function: + */ + return (AE_BAD_PARAMETER); } - strcpy(dest, source); - return (FALSE); -} + /* Skip over any white space in the buffer: */ -u8 acpi_ut_safe_strcat(char *dest, acpi_size dest_size, char *source) -{ + while (isspace((int)*string) || *string == '\t') { + ++string; + } - if ((strlen(dest) + strlen(source)) >= dest_size) { - return (TRUE); + /* + * The buffer may contain an optional plus or minus sign. + * If it does, then skip over it but remember what is was: + */ + if (*string == '-') { + sign = ACPI_SIGN_NEGATIVE; + ++string; + } else if (*string == '+') { + ++string; + sign = ACPI_SIGN_POSITIVE; + } else { + sign = ACPI_SIGN_POSITIVE; } - strcat(dest, source); - return (FALSE); -} + /* + * If the input parameter Base is zero, then we need to + * determine if it is octal, decimal, or hexadecimal: + */ + if (base == 0) { + if (*string == '0') { + if (tolower((int)*(++string)) == 'x') { + base = 16; + ++string; + } else { + base = 8; + } + } else { + base = 10; + } + } -u8 -acpi_ut_safe_strncat(char *dest, - acpi_size dest_size, - char *source, acpi_size max_transfer_length) -{ - acpi_size actual_transfer_length; + /* + * For octal and hexadecimal bases, skip over the leading + * 0 or 0x, if they are present. + */ + if (base == 8 && *string == '0') { + string++; + } - actual_transfer_length = ACPI_MIN(max_transfer_length, strlen(source)); + if (base == 16 && *string == '0' && tolower((int)*(++string)) == 'x') { + string++; + } - if ((strlen(dest) + actual_transfer_length) >= dest_size) { - return (TRUE); + /* Main loop: convert the string to an unsigned long */ + + while (*string) { + if (isdigit((int)*string)) { + index = ((u8)*string) - '0'; + } else { + index = (u8)toupper((int)*string); + if (isupper((int)index)) { + index = index - 'A' + 10; + } else { + goto error_exit; + } + } + + if (index >= base) { + goto error_exit; + } + + /* Check to see if value is out of range: */ + + if (return_value > ((ACPI_UINT64_MAX - (u64)index) / (u64)base)) { + goto error_exit; + } else { + return_value *= base; + return_value += index; + } + + ++string; } - strncat(dest, source, max_transfer_length); - return (FALSE); + /* If a minus sign was present, then "the conversion is negated": */ + + if (sign == ACPI_SIGN_NEGATIVE) { + return_value = (ACPI_UINT32_MAX - return_value) + 1; + } + + *ret_integer = return_value; + return (status); + +error_exit: + switch (base) { + case 8: + + status = AE_BAD_OCTAL_CONSTANT; + break; + + case 10: + + status = AE_BAD_DECIMAL_CONSTANT; + break; + + case 16: + + status = AE_BAD_HEX_CONSTANT; + break; + + default: + + /* Base validated above */ + + break; + } + + return (status); } #endif diff --git a/drivers/acpi/acpica/uttrack.c b/drivers/acpi/acpica/uttrack.c index c7c2bb8..60c406a 100644 --- a/drivers/acpi/acpica/uttrack.c +++ b/drivers/acpi/acpica/uttrack.c @@ -712,7 +712,7 @@ void acpi_ut_dump_allocations(u32 component, const char *module) /* Print summary */ if (!num_outstanding) { - ACPI_INFO((AE_INFO, "No outstanding allocations")); + ACPI_INFO(("No outstanding allocations")); } else { ACPI_ERROR((AE_INFO, "%u(0x%X) Outstanding allocations", num_outstanding, num_outstanding)); diff --git a/drivers/acpi/acpica/utxferror.c b/drivers/acpi/acpica/utxferror.c index 6fe5959..d9f15cb 100644 --- a/drivers/acpi/acpica/utxferror.c +++ b/drivers/acpi/acpica/utxferror.c @@ -175,8 +175,7 @@ ACPI_EXPORT_SYMBOL(acpi_warning) * TBD: module_name and line_number args are not needed, should be removed. * ******************************************************************************/ -void ACPI_INTERNAL_VAR_XFACE -acpi_info(const char *module_name, u32 line_number, const char *format, ...) +void ACPI_INTERNAL_VAR_XFACE acpi_info(const char *format, ...) { va_list arg_list; diff --git a/drivers/acpi/acpica/utxfinit.c b/drivers/acpi/acpica/utxfinit.c index 721b87c..75b5f27 100644 --- a/drivers/acpi/acpica/utxfinit.c +++ b/drivers/acpi/acpica/utxfinit.c @@ -154,21 +154,6 @@ acpi_status __init acpi_enable_subsystem(u32 flags) */ acpi_gbl_early_initialization = FALSE; - /* - * Install the default operation region handlers. These are the - * handlers that are defined by the ACPI specification to be - * "always accessible" -- namely, system_memory, system_IO, and - * PCI_Config. This also means that no _REG methods need to be - * run for these address spaces. We need to have these handlers - * installed before any AML code can be executed, especially any - * module-level code (11/2015). - */ - status = acpi_ev_install_region_handlers(); - if (ACPI_FAILURE(status)) { - ACPI_EXCEPTION((AE_INFO, status, - "During Region initialization")); - return_ACPI_STATUS(status); - } #if (!ACPI_REDUCED_HARDWARE) /* Enable ACPI mode */ @@ -260,23 +245,6 @@ acpi_status __init acpi_initialize_objects(u32 flags) ACPI_FUNCTION_TRACE(acpi_initialize_objects); - /* - * Run all _REG methods - * - * Note: Any objects accessed by the _REG methods will be automatically - * initialized, even if they contain executable AML (see the call to - * acpi_ns_initialize_objects below). - */ - acpi_gbl_reg_methods_enabled = TRUE; - if (!(flags & ACPI_NO_ADDRESS_SPACE_INIT)) { - ACPI_DEBUG_PRINT((ACPI_DB_EXEC, - "[Init] Executing _REG OpRegion methods\n")); - - status = acpi_ev_initialize_op_regions(); - if (ACPI_FAILURE(status)) { - return_ACPI_STATUS(status); - } - } #ifdef ACPI_EXEC_APP /* * This call implements the "initialization file" option for acpi_exec. @@ -299,32 +267,27 @@ acpi_status __init acpi_initialize_objects(u32 flags) */ if (acpi_gbl_group_module_level_code) { acpi_ns_exec_module_code_list(); - } - - /* - * Initialize the objects that remain uninitialized. This runs the - * executable AML that may be part of the declaration of these objects: - * operation_regions, buffer_fields, Buffers, and Packages. - */ - if (!(flags & ACPI_NO_OBJECT_INIT)) { - ACPI_DEBUG_PRINT((ACPI_DB_EXEC, - "[Init] Completing Initialization of ACPI Objects\n")); - status = acpi_ns_initialize_objects(); - if (ACPI_FAILURE(status)) { - return_ACPI_STATUS(status); + /* + * Initialize the objects that remain uninitialized. This + * runs the executable AML that may be part of the + * declaration of these objects: + * operation_regions, buffer_fields, Buffers, and Packages. + */ + if (!(flags & ACPI_NO_OBJECT_INIT)) { + status = acpi_ns_initialize_objects(); + if (ACPI_FAILURE(status)) { + return_ACPI_STATUS(status); + } } } /* - * Initialize all device objects in the namespace. This runs the device - * _STA and _INI methods. + * Initialize all device/region objects in the namespace. This runs + * the device _STA and _INI methods and region _REG methods. */ - if (!(flags & ACPI_NO_DEVICE_INIT)) { - ACPI_DEBUG_PRINT((ACPI_DB_EXEC, - "[Init] Initializing ACPI Devices\n")); - - status = acpi_ns_initialize_devices(); + if (!(flags & (ACPI_NO_DEVICE_INIT | ACPI_NO_ADDRESS_SPACE_INIT))) { + status = acpi_ns_initialize_devices(flags); if (ACPI_FAILURE(status)) { return_ACPI_STATUS(status); } diff --git a/drivers/acpi/apei/apei-base.c b/drivers/acpi/apei/apei-base.c index a2c8d7a..da370e1 100644 --- a/drivers/acpi/apei/apei-base.c +++ b/drivers/acpi/apei/apei-base.c @@ -536,7 +536,8 @@ int apei_resources_request(struct apei_resources *resources, goto err_unmap_ioport; } - return 0; + goto arch_res_fini; + err_unmap_ioport: list_for_each_entry(res, &resources->ioport, list) { if (res == res_bak) @@ -551,7 +552,8 @@ err_unmap_iomem: release_mem_region(res->start, res->end - res->start); } arch_res_fini: - apei_resources_fini(&arch_res); + if (arch_apei_filter_addr) + apei_resources_fini(&arch_res); nvs_res_fini: apei_resources_fini(&nvs_resources); return rc; diff --git a/drivers/acpi/apei/erst.c b/drivers/acpi/apei/erst.c index 6e6bc10..006c389 100644 --- a/drivers/acpi/apei/erst.c +++ b/drivers/acpi/apei/erst.c @@ -1207,6 +1207,9 @@ static int __init erst_init(void) "Failed to allocate %lld bytes for persistent store error log.\n", erst_erange.size); + /* Cleanup ERST Resources */ + apei_resources_fini(&erst_resources); + return 0; err_release_erange: diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c index 3dd9c46..60746ef 100644 --- a/drivers/acpi/apei/ghes.c +++ b/drivers/acpi/apei/ghes.c @@ -26,7 +26,7 @@ */ #include <linux/kernel.h> -#include <linux/module.h> +#include <linux/moduleparam.h> #include <linux/init.h> #include <linux/acpi.h> #include <linux/io.h> @@ -79,6 +79,11 @@ ((struct acpi_hest_generic_status *) \ ((struct ghes_estatus_node *)(estatus_node) + 1)) +/* + * This driver isn't really modular, however for the time being, + * continuing to use module_param is the easiest way to remain + * compatible with existing boot arg use cases. + */ bool ghes_disable; module_param_named(disable, ghes_disable, bool, 0); @@ -1148,18 +1153,4 @@ err_ioremap_exit: err: return rc; } - -static void __exit ghes_exit(void) -{ - platform_driver_unregister(&ghes_platform_driver); - ghes_estatus_pool_exit(); - ghes_ioremap_exit(); -} - -module_init(ghes_init); -module_exit(ghes_exit); - -MODULE_AUTHOR("Huang Ying"); -MODULE_DESCRIPTION("APEI Generic Hardware Error Source support"); -MODULE_LICENSE("GPL"); -MODULE_ALIAS("platform:GHES"); +device_initcall(ghes_init); diff --git a/drivers/acpi/bgrt.c b/drivers/acpi/bgrt.c index a83e3c6..75f128e 100644 --- a/drivers/acpi/bgrt.c +++ b/drivers/acpi/bgrt.c @@ -1,4 +1,6 @@ /* + * BGRT boot graphic support + * Authors: Matthew Garrett, Josh Triplett <josh@joshtriplett.org> * Copyright 2012 Red Hat, Inc <mjg@redhat.com> * Copyright 2012 Intel Corporation * @@ -8,7 +10,6 @@ */ #include <linux/kernel.h> -#include <linux/module.h> #include <linux/init.h> #include <linux/device.h> #include <linux/sysfs.h> @@ -103,9 +104,4 @@ out_kobject: kobject_put(bgrt_kobj); return ret; } - -module_init(bgrt_init); - -MODULE_AUTHOR("Matthew Garrett, Josh Triplett <josh@joshtriplett.org>"); -MODULE_DESCRIPTION("BGRT boot graphic support"); -MODULE_LICENSE("GPL"); +device_initcall(bgrt_init); diff --git a/drivers/acpi/bus.c b/drivers/acpi/bus.c index 891c42d..0e85678 100644 --- a/drivers/acpi/bus.c +++ b/drivers/acpi/bus.c @@ -479,24 +479,38 @@ static void acpi_device_remove_notify_handler(struct acpi_device *device) Device Matching -------------------------------------------------------------------------- */ -static struct acpi_device *acpi_primary_dev_companion(struct acpi_device *adev, - const struct device *dev) +/** + * acpi_get_first_physical_node - Get first physical node of an ACPI device + * @adev: ACPI device in question + * + * Return: First physical node of ACPI device @adev + */ +struct device *acpi_get_first_physical_node(struct acpi_device *adev) { struct mutex *physical_node_lock = &adev->physical_node_lock; + struct device *phys_dev; mutex_lock(physical_node_lock); if (list_empty(&adev->physical_node_list)) { - adev = NULL; + phys_dev = NULL; } else { const struct acpi_device_physical_node *node; node = list_first_entry(&adev->physical_node_list, struct acpi_device_physical_node, node); - if (node->dev != dev) - adev = NULL; + + phys_dev = node->dev; } mutex_unlock(physical_node_lock); - return adev; + return phys_dev; +} + +static struct acpi_device *acpi_primary_dev_companion(struct acpi_device *adev, + const struct device *dev) +{ + const struct device *phys_dev = acpi_get_first_physical_node(adev); + + return phys_dev && phys_dev == dev ? adev : NULL; } /** diff --git a/drivers/acpi/cppc_acpi.c b/drivers/acpi/cppc_acpi.c index 6730f96..8adac69 100644 --- a/drivers/acpi/cppc_acpi.c +++ b/drivers/acpi/cppc_acpi.c @@ -39,6 +39,7 @@ #include <linux/cpufreq.h> #include <linux/delay.h> +#include <linux/ktime.h> #include <acpi/cppc_acpi.h> /* @@ -63,58 +64,140 @@ static struct mbox_chan *pcc_channel; static void __iomem *pcc_comm_addr; static u64 comm_base_addr; static int pcc_subspace_idx = -1; -static u16 pcc_cmd_delay; static bool pcc_channel_acquired; +static ktime_t deadline; +static unsigned int pcc_mpar, pcc_mrtt; + +/* pcc mapped address + header size + offset within PCC subspace */ +#define GET_PCC_VADDR(offs) (pcc_comm_addr + 0x8 + (offs)) /* * Arbitrary Retries in case the remote processor is slow to respond - * to PCC commands. + * to PCC commands. Keeping it high enough to cover emulators where + * the processors run painfully slow. */ #define NUM_RETRIES 500 +static int check_pcc_chan(void) +{ + int ret = -EIO; + struct acpi_pcct_shared_memory __iomem *generic_comm_base = pcc_comm_addr; + ktime_t next_deadline = ktime_add(ktime_get(), deadline); + + /* Retry in case the remote processor was too slow to catch up. */ + while (!ktime_after(ktime_get(), next_deadline)) { + /* + * Per spec, prior to boot the PCC space wil be initialized by + * platform and should have set the command completion bit when + * PCC can be used by OSPM + */ + if (readw_relaxed(&generic_comm_base->status) & PCC_CMD_COMPLETE) { + ret = 0; + break; + } + /* + * Reducing the bus traffic in case this loop takes longer than + * a few retries. + */ + udelay(3); + } + + return ret; +} + static int send_pcc_cmd(u16 cmd) { - int retries, result = -EIO; - struct acpi_pcct_hw_reduced *pcct_ss = pcc_channel->con_priv; + int ret = -EIO; struct acpi_pcct_shared_memory *generic_comm_base = (struct acpi_pcct_shared_memory *) pcc_comm_addr; - u32 cmd_latency = pcct_ss->latency; + static ktime_t last_cmd_cmpl_time, last_mpar_reset; + static int mpar_count; + unsigned int time_delta; - /* Min time OS should wait before sending next command. */ - udelay(pcc_cmd_delay); + /* + * For CMD_WRITE we know for a fact the caller should have checked + * the channel before writing to PCC space + */ + if (cmd == CMD_READ) { + ret = check_pcc_chan(); + if (ret) + return ret; + } + + /* + * Handle the Minimum Request Turnaround Time(MRTT) + * "The minimum amount of time that OSPM must wait after the completion + * of a command before issuing the next command, in microseconds" + */ + if (pcc_mrtt) { + time_delta = ktime_us_delta(ktime_get(), last_cmd_cmpl_time); + if (pcc_mrtt > time_delta) + udelay(pcc_mrtt - time_delta); + } + + /* + * Handle the non-zero Maximum Periodic Access Rate(MPAR) + * "The maximum number of periodic requests that the subspace channel can + * support, reported in commands per minute. 0 indicates no limitation." + * + * This parameter should be ideally zero or large enough so that it can + * handle maximum number of requests that all the cores in the system can + * collectively generate. If it is not, we will follow the spec and just + * not send the request to the platform after hitting the MPAR limit in + * any 60s window + */ + if (pcc_mpar) { + if (mpar_count == 0) { + time_delta = ktime_ms_delta(ktime_get(), last_mpar_reset); + if (time_delta < 60 * MSEC_PER_SEC) { + pr_debug("PCC cmd not sent due to MPAR limit"); + return -EIO; + } + last_mpar_reset = ktime_get(); + mpar_count = pcc_mpar; + } + mpar_count--; + } /* Write to the shared comm region. */ - writew(cmd, &generic_comm_base->command); + writew_relaxed(cmd, &generic_comm_base->command); /* Flip CMD COMPLETE bit */ - writew(0, &generic_comm_base->status); + writew_relaxed(0, &generic_comm_base->status); /* Ring doorbell */ - result = mbox_send_message(pcc_channel, &cmd); - if (result < 0) { + ret = mbox_send_message(pcc_channel, &cmd); + if (ret < 0) { pr_err("Err sending PCC mbox message. cmd:%d, ret:%d\n", - cmd, result); - return result; + cmd, ret); + return ret; } - /* Wait for a nominal time to let platform process command. */ - udelay(cmd_latency); - - /* Retry in case the remote processor was too slow to catch up. */ - for (retries = NUM_RETRIES; retries > 0; retries--) { - if (readw_relaxed(&generic_comm_base->status) & PCC_CMD_COMPLETE) { - result = 0; - break; - } + /* + * For READs we need to ensure the cmd completed to ensure + * the ensuing read()s can proceed. For WRITEs we dont care + * because the actual write()s are done before coming here + * and the next READ or WRITE will check if the channel + * is busy/free at the entry of this call. + * + * If Minimum Request Turnaround Time is non-zero, we need + * to record the completion time of both READ and WRITE + * command for proper handling of MRTT, so we need to check + * for pcc_mrtt in addition to CMD_READ + */ + if (cmd == CMD_READ || pcc_mrtt) { + ret = check_pcc_chan(); + if (pcc_mrtt) + last_cmd_cmpl_time = ktime_get(); } - mbox_client_txdone(pcc_channel, result); - return result; + mbox_client_txdone(pcc_channel, ret); + return ret; } static void cppc_chan_tx_done(struct mbox_client *cl, void *msg, int ret) { - if (ret) + if (ret < 0) pr_debug("TX did not complete: CMD sent:%x, ret:%d\n", *(u16 *)msg, ret); else @@ -306,6 +389,7 @@ static int register_pcc_channel(int pcc_subspace_idx) { struct acpi_pcct_hw_reduced *cppc_ss; unsigned int len; + u64 usecs_lat; if (pcc_subspace_idx >= 0) { pcc_channel = pcc_mbox_request_channel(&cppc_mbox_cl, @@ -335,7 +419,16 @@ static int register_pcc_channel(int pcc_subspace_idx) */ comm_base_addr = cppc_ss->base_address; len = cppc_ss->length; - pcc_cmd_delay = cppc_ss->min_turnaround_time; + + /* + * cppc_ss->latency is just a Nominal value. In reality + * the remote processor could be much slower to reply. + * So add an arbitrary amount of wait on top of Nominal. + */ + usecs_lat = NUM_RETRIES * cppc_ss->latency; + deadline = ns_to_ktime(usecs_lat * NSEC_PER_USEC); + pcc_mrtt = cppc_ss->min_turnaround_time; + pcc_mpar = cppc_ss->max_access_rate; pcc_comm_addr = acpi_os_ioremap(comm_base_addr, len); if (!pcc_comm_addr) { @@ -546,29 +639,74 @@ void acpi_cppc_processor_exit(struct acpi_processor *pr) } EXPORT_SYMBOL_GPL(acpi_cppc_processor_exit); -static u64 get_phys_addr(struct cpc_reg *reg) -{ - /* PCC communication addr space begins at byte offset 0x8. */ - if (reg->space_id == ACPI_ADR_SPACE_PLATFORM_COMM) - return (u64)comm_base_addr + 0x8 + reg->address; - else - return reg->address; -} +/* + * Since cpc_read and cpc_write are called while holding pcc_lock, it should be + * as fast as possible. We have already mapped the PCC subspace during init, so + * we can directly write to it. + */ -static void cpc_read(struct cpc_reg *reg, u64 *val) +static int cpc_read(struct cpc_reg *reg, u64 *val) { - u64 addr = get_phys_addr(reg); + int ret_val = 0; - acpi_os_read_memory((acpi_physical_address)addr, - val, reg->bit_width); + *val = 0; + if (reg->space_id == ACPI_ADR_SPACE_PLATFORM_COMM) { + void __iomem *vaddr = GET_PCC_VADDR(reg->address); + + switch (reg->bit_width) { + case 8: + *val = readb_relaxed(vaddr); + break; + case 16: + *val = readw_relaxed(vaddr); + break; + case 32: + *val = readl_relaxed(vaddr); + break; + case 64: + *val = readq_relaxed(vaddr); + break; + default: + pr_debug("Error: Cannot read %u bit width from PCC\n", + reg->bit_width); + ret_val = -EFAULT; + } + } else + ret_val = acpi_os_read_memory((acpi_physical_address)reg->address, + val, reg->bit_width); + return ret_val; } -static void cpc_write(struct cpc_reg *reg, u64 val) +static int cpc_write(struct cpc_reg *reg, u64 val) { - u64 addr = get_phys_addr(reg); + int ret_val = 0; - acpi_os_write_memory((acpi_physical_address)addr, - val, reg->bit_width); + if (reg->space_id == ACPI_ADR_SPACE_PLATFORM_COMM) { + void __iomem *vaddr = GET_PCC_VADDR(reg->address); + + switch (reg->bit_width) { + case 8: + writeb_relaxed(val, vaddr); + break; + case 16: + writew_relaxed(val, vaddr); + break; + case 32: + writel_relaxed(val, vaddr); + break; + case 64: + writeq_relaxed(val, vaddr); + break; + default: + pr_debug("Error: Cannot write %u bit width to PCC\n", + reg->bit_width); + ret_val = -EFAULT; + break; + } + } else + ret_val = acpi_os_write_memory((acpi_physical_address)reg->address, + val, reg->bit_width); + return ret_val; } /** @@ -604,7 +742,7 @@ int cppc_get_perf_caps(int cpunum, struct cppc_perf_caps *perf_caps) (ref_perf->cpc_entry.reg.space_id == ACPI_ADR_SPACE_PLATFORM_COMM) || (nom_perf->cpc_entry.reg.space_id == ACPI_ADR_SPACE_PLATFORM_COMM)) { /* Ring doorbell once to update PCC subspace */ - if (send_pcc_cmd(CMD_READ)) { + if (send_pcc_cmd(CMD_READ) < 0) { ret = -EIO; goto out_err; } @@ -662,7 +800,7 @@ int cppc_get_perf_ctrs(int cpunum, struct cppc_perf_fb_ctrs *perf_fb_ctrs) if ((delivered_reg->cpc_entry.reg.space_id == ACPI_ADR_SPACE_PLATFORM_COMM) || (reference_reg->cpc_entry.reg.space_id == ACPI_ADR_SPACE_PLATFORM_COMM)) { /* Ring doorbell once to update PCC subspace */ - if (send_pcc_cmd(CMD_READ)) { + if (send_pcc_cmd(CMD_READ) < 0) { ret = -EIO; goto out_err; } @@ -713,6 +851,13 @@ int cppc_set_perf(int cpu, struct cppc_perf_ctrls *perf_ctrls) spin_lock(&pcc_lock); + /* If this is PCC reg, check if channel is free before writing */ + if (desired_reg->cpc_entry.reg.space_id == ACPI_ADR_SPACE_PLATFORM_COMM) { + ret = check_pcc_chan(); + if (ret) + goto busy_channel; + } + /* * Skip writing MIN/MAX until Linux knows how to come up with * useful values. @@ -722,10 +867,10 @@ int cppc_set_perf(int cpu, struct cppc_perf_ctrls *perf_ctrls) /* Is this a PCC reg ?*/ if (desired_reg->cpc_entry.reg.space_id == ACPI_ADR_SPACE_PLATFORM_COMM) { /* Ring doorbell so Remote can get our perf request. */ - if (send_pcc_cmd(CMD_WRITE)) + if (send_pcc_cmd(CMD_WRITE) < 0) ret = -EIO; } - +busy_channel: spin_unlock(&pcc_lock); return ret; diff --git a/drivers/acpi/ec_sys.c b/drivers/acpi/ec_sys.c index bea8e42..6c7dd7a 100644 --- a/drivers/acpi/ec_sys.c +++ b/drivers/acpi/ec_sys.c @@ -73,6 +73,9 @@ static ssize_t acpi_ec_write_io(struct file *f, const char __user *buf, loff_t init_off = *off; int err = 0; + if (!write_support) + return -EINVAL; + if (*off >= EC_SPACE_SIZE) return 0; if (*off + count >= EC_SPACE_SIZE) { diff --git a/drivers/acpi/fan.c b/drivers/acpi/fan.c index 6322db6..384cfc3 100644 --- a/drivers/acpi/fan.c +++ b/drivers/acpi/fan.c @@ -46,7 +46,7 @@ MODULE_DEVICE_TABLE(acpi, fan_device_ids); #ifdef CONFIG_PM_SLEEP static int acpi_fan_suspend(struct device *dev); static int acpi_fan_resume(struct device *dev); -static struct dev_pm_ops acpi_fan_pm = { +static const struct dev_pm_ops acpi_fan_pm = { .resume = acpi_fan_resume, .freeze = acpi_fan_suspend, .thaw = acpi_fan_resume, diff --git a/drivers/acpi/internal.h b/drivers/acpi/internal.h index 1e6833a..a37508e 100644 --- a/drivers/acpi/internal.h +++ b/drivers/acpi/internal.h @@ -20,6 +20,7 @@ #define PREFIX "ACPI: " +void acpi_initrd_initialize_tables(void); acpi_status acpi_os_initialize1(void); void init_acpi_device_notify(void); int acpi_scan_init(void); @@ -29,6 +30,11 @@ void acpi_processor_init(void); void acpi_platform_init(void); void acpi_pnp_init(void); void acpi_int340x_thermal_init(void); +#ifdef CONFIG_ARM_AMBA +void acpi_amba_init(void); +#else +static inline void acpi_amba_init(void) {} +#endif int acpi_sysfs_init(void); void acpi_container_init(void); void acpi_memory_hotplug_init(void); @@ -106,6 +112,7 @@ bool acpi_device_is_present(struct acpi_device *adev); bool acpi_device_is_battery(struct acpi_device *adev); bool acpi_device_is_first_physical_node(struct acpi_device *adev, const struct device *dev); +struct device *acpi_get_first_physical_node(struct acpi_device *adev); /* -------------------------------------------------------------------------- Device Matching and Notification diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c index 67da6fb..814d5f8 100644 --- a/drivers/acpi/osl.c +++ b/drivers/acpi/osl.c @@ -602,6 +602,14 @@ acpi_os_predefined_override(const struct acpi_predefined_names *init_val, return AE_OK; } +static void acpi_table_taint(struct acpi_table_header *table) +{ + pr_warn(PREFIX + "Override [%4.4s-%8.8s], this is unsafe: tainting kernel\n", + table->signature, table->oem_table_id); + add_taint(TAINT_OVERRIDDEN_ACPI_TABLE, LOCKDEP_NOW_UNRELIABLE); +} + #ifdef CONFIG_ACPI_INITRD_TABLE_OVERRIDE #include <linux/earlycpio.h> #include <linux/memblock.h> @@ -636,6 +644,7 @@ static const char * const table_sigs[] = { #define ACPI_OVERRIDE_TABLES 64 static struct cpio_data __initdata acpi_initrd_files[ACPI_OVERRIDE_TABLES]; +static DECLARE_BITMAP(acpi_initrd_installed, ACPI_OVERRIDE_TABLES); #define MAP_CHUNK_SIZE (NR_FIX_BTMAPS << PAGE_SHIFT) @@ -746,96 +755,125 @@ void __init acpi_initrd_override(void *data, size_t size) } } } -#endif /* CONFIG_ACPI_INITRD_TABLE_OVERRIDE */ -static void acpi_table_taint(struct acpi_table_header *table) +acpi_status +acpi_os_physical_table_override(struct acpi_table_header *existing_table, + acpi_physical_address *address, u32 *length) { - pr_warn(PREFIX - "Override [%4.4s-%8.8s], this is unsafe: tainting kernel\n", - table->signature, table->oem_table_id); - add_taint(TAINT_OVERRIDDEN_ACPI_TABLE, LOCKDEP_NOW_UNRELIABLE); -} + int table_offset = 0; + int table_index = 0; + struct acpi_table_header *table; + u32 table_length; + *length = 0; + *address = 0; + if (!acpi_tables_addr) + return AE_OK; -acpi_status -acpi_os_table_override(struct acpi_table_header * existing_table, - struct acpi_table_header ** new_table) -{ - if (!existing_table || !new_table) - return AE_BAD_PARAMETER; + while (table_offset + ACPI_HEADER_SIZE <= all_tables_size) { + table = acpi_os_map_memory(acpi_tables_addr + table_offset, + ACPI_HEADER_SIZE); + if (table_offset + table->length > all_tables_size) { + acpi_os_unmap_memory(table, ACPI_HEADER_SIZE); + WARN_ON(1); + return AE_OK; + } - *new_table = NULL; + table_length = table->length; -#ifdef CONFIG_ACPI_CUSTOM_DSDT - if (strncmp(existing_table->signature, "DSDT", 4) == 0) - *new_table = (struct acpi_table_header *)AmlCode; -#endif - if (*new_table != NULL) + /* Only override tables matched */ + if (test_bit(table_index, acpi_initrd_installed) || + memcmp(existing_table->signature, table->signature, 4) || + memcmp(table->oem_table_id, existing_table->oem_table_id, + ACPI_OEM_TABLE_ID_SIZE)) { + acpi_os_unmap_memory(table, ACPI_HEADER_SIZE); + goto next_table; + } + + *length = table_length; + *address = acpi_tables_addr + table_offset; acpi_table_taint(existing_table); + acpi_os_unmap_memory(table, ACPI_HEADER_SIZE); + set_bit(table_index, acpi_initrd_installed); + break; + +next_table: + table_offset += table_length; + table_index++; + } return AE_OK; } -acpi_status -acpi_os_physical_table_override(struct acpi_table_header *existing_table, - acpi_physical_address *address, - u32 *table_length) +void __init acpi_initrd_initialize_tables(void) { -#ifndef CONFIG_ACPI_INITRD_TABLE_OVERRIDE - *table_length = 0; - *address = 0; - return AE_OK; -#else int table_offset = 0; + int table_index = 0; + u32 table_length; struct acpi_table_header *table; - *table_length = 0; - *address = 0; - if (!acpi_tables_addr) - return AE_OK; - - do { - if (table_offset + ACPI_HEADER_SIZE > all_tables_size) { - WARN_ON(1); - return AE_OK; - } + return; + while (table_offset + ACPI_HEADER_SIZE <= all_tables_size) { table = acpi_os_map_memory(acpi_tables_addr + table_offset, ACPI_HEADER_SIZE); - if (table_offset + table->length > all_tables_size) { acpi_os_unmap_memory(table, ACPI_HEADER_SIZE); WARN_ON(1); - return AE_OK; + return; } - table_offset += table->length; + table_length = table->length; - if (memcmp(existing_table->signature, table->signature, 4)) { - acpi_os_unmap_memory(table, - ACPI_HEADER_SIZE); - continue; - } - - /* Only override tables with matching oem id */ - if (memcmp(table->oem_table_id, existing_table->oem_table_id, - ACPI_OEM_TABLE_ID_SIZE)) { - acpi_os_unmap_memory(table, - ACPI_HEADER_SIZE); - continue; + /* Skip RSDT/XSDT which should only be used for override */ + if (test_bit(table_index, acpi_initrd_installed) || + ACPI_COMPARE_NAME(table->signature, ACPI_SIG_RSDT) || + ACPI_COMPARE_NAME(table->signature, ACPI_SIG_XSDT)) { + acpi_os_unmap_memory(table, ACPI_HEADER_SIZE); + goto next_table; } - table_offset -= table->length; - *table_length = table->length; + acpi_table_taint(table); acpi_os_unmap_memory(table, ACPI_HEADER_SIZE); - *address = acpi_tables_addr + table_offset; - break; - } while (table_offset + ACPI_HEADER_SIZE < all_tables_size); + acpi_install_table(acpi_tables_addr + table_offset, TRUE); + set_bit(table_index, acpi_initrd_installed); +next_table: + table_offset += table_length; + table_index++; + } +} +#else +acpi_status +acpi_os_physical_table_override(struct acpi_table_header *existing_table, + acpi_physical_address *address, + u32 *table_length) +{ + *table_length = 0; + *address = 0; + return AE_OK; +} + +void __init acpi_initrd_initialize_tables(void) +{ +} +#endif /* CONFIG_ACPI_INITRD_TABLE_OVERRIDE */ - if (*address != 0) +acpi_status +acpi_os_table_override(struct acpi_table_header *existing_table, + struct acpi_table_header **new_table) +{ + if (!existing_table || !new_table) + return AE_BAD_PARAMETER; + + *new_table = NULL; + +#ifdef CONFIG_ACPI_CUSTOM_DSDT + if (strncmp(existing_table->signature, "DSDT", 4) == 0) + *new_table = (struct acpi_table_header *)AmlCode; +#endif + if (*new_table != NULL) acpi_table_taint(existing_table); return AE_OK; -#endif } static irqreturn_t acpi_irq(int irq, void *dev_id) diff --git a/drivers/acpi/pci_irq.c b/drivers/acpi/pci_irq.c index c8e169e..2c45dd3 100644 --- a/drivers/acpi/pci_irq.c +++ b/drivers/acpi/pci_irq.c @@ -33,6 +33,7 @@ #include <linux/pci.h> #include <linux/acpi.h> #include <linux/slab.h> +#include <linux/interrupt.h> #define PREFIX "ACPI: " @@ -387,6 +388,23 @@ static inline int acpi_isa_register_gsi(struct pci_dev *dev) } #endif +static inline bool acpi_pci_irq_valid(struct pci_dev *dev, u8 pin) +{ +#ifdef CONFIG_X86 + /* + * On x86 irq line 0xff means "unknown" or "no connection" + * (PCI 3.0, Section 6.2.4, footnote on page 223). + */ + if (dev->irq == 0xff) { + dev->irq = IRQ_NOTCONNECTED; + dev_warn(&dev->dev, "PCI INT %c: not connected\n", + pin_name(pin)); + return false; + } +#endif + return true; +} + int acpi_pci_irq_enable(struct pci_dev *dev) { struct acpi_prt_entry *entry; @@ -431,11 +449,14 @@ int acpi_pci_irq_enable(struct pci_dev *dev) } else gsi = -1; - /* - * No IRQ known to the ACPI subsystem - maybe the BIOS / - * driver reported one, then use it. Exit in any case. - */ if (gsi < 0) { + /* + * No IRQ known to the ACPI subsystem - maybe the BIOS / + * driver reported one, then use it. Exit in any case. + */ + if (!acpi_pci_irq_valid(dev, pin)) + return 0; + if (acpi_isa_register_gsi(dev)) dev_warn(&dev->dev, "PCI INT %c: no GSI\n", pin_name(pin)); diff --git a/drivers/acpi/pmic/intel_pmic_crc.c b/drivers/acpi/pmic/intel_pmic_crc.c index 42df46a..fcd1852 100644 --- a/drivers/acpi/pmic/intel_pmic_crc.c +++ b/drivers/acpi/pmic/intel_pmic_crc.c @@ -13,7 +13,7 @@ * GNU General Public License for more details. */ -#include <linux/module.h> +#include <linux/init.h> #include <linux/acpi.h> #include <linux/mfd/intel_soc_pmic.h> #include <linux/regmap.h> @@ -205,7 +205,4 @@ static int __init intel_crc_pmic_opregion_driver_init(void) { return platform_driver_register(&intel_crc_pmic_opregion_driver); } -module_init(intel_crc_pmic_opregion_driver_init); - -MODULE_DESCRIPTION("CrystalCove ACPI operation region driver"); -MODULE_LICENSE("GPL"); +device_initcall(intel_crc_pmic_opregion_driver_init); diff --git a/drivers/acpi/processor_driver.c b/drivers/acpi/processor_driver.c index 11154a3..d2fa8cb 100644 --- a/drivers/acpi/processor_driver.c +++ b/drivers/acpi/processor_driver.c @@ -314,7 +314,6 @@ static int __init acpi_processor_driver_init(void) if (result < 0) return result; - acpi_processor_syscore_init(); register_hotcpu_notifier(&acpi_cpu_notifier); acpi_thermal_cpufreq_init(); acpi_processor_ppc_init(); @@ -330,7 +329,6 @@ static void __exit acpi_processor_driver_exit(void) acpi_processor_ppc_exit(); acpi_thermal_cpufreq_exit(); unregister_hotcpu_notifier(&acpi_cpu_notifier); - acpi_processor_syscore_exit(); driver_unregister(&acpi_processor_driver); } diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c index 9ca2b2f..444e374 100644 --- a/drivers/acpi/processor_idle.c +++ b/drivers/acpi/processor_idle.c @@ -23,6 +23,7 @@ * * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ */ +#define pr_fmt(fmt) "ACPI: " fmt #include <linux/module.h> #include <linux/acpi.h> @@ -30,7 +31,6 @@ #include <linux/sched.h> /* need_resched() */ #include <linux/tick.h> #include <linux/cpuidle.h> -#include <linux/syscore_ops.h> #include <acpi/processor.h> /* @@ -43,8 +43,6 @@ #include <asm/apic.h> #endif -#define PREFIX "ACPI: " - #define ACPI_PROCESSOR_CLASS "processor" #define _COMPONENT ACPI_PROCESSOR_COMPONENT ACPI_MODULE_NAME("processor_idle"); @@ -81,9 +79,9 @@ static int set_max_cstate(const struct dmi_system_id *id) if (max_cstate > ACPI_PROCESSOR_MAX_POWER) return 0; - printk(KERN_NOTICE PREFIX "%s detected - limiting to C%ld max_cstate." - " Override with \"processor.max_cstate=%d\"\n", id->ident, - (long)id->driver_data, ACPI_PROCESSOR_MAX_POWER + 1); + pr_notice("%s detected - limiting to C%ld max_cstate." + " Override with \"processor.max_cstate=%d\"\n", id->ident, + (long)id->driver_data, ACPI_PROCESSOR_MAX_POWER + 1); max_cstate = (long)id->driver_data; @@ -194,42 +192,6 @@ static void lapic_timer_state_broadcast(struct acpi_processor *pr, #endif -#ifdef CONFIG_PM_SLEEP -static u32 saved_bm_rld; - -static int acpi_processor_suspend(void) -{ - acpi_read_bit_register(ACPI_BITREG_BUS_MASTER_RLD, &saved_bm_rld); - return 0; -} - -static void acpi_processor_resume(void) -{ - u32 resumed_bm_rld = 0; - - acpi_read_bit_register(ACPI_BITREG_BUS_MASTER_RLD, &resumed_bm_rld); - if (resumed_bm_rld == saved_bm_rld) - return; - - acpi_write_bit_register(ACPI_BITREG_BUS_MASTER_RLD, saved_bm_rld); -} - -static struct syscore_ops acpi_processor_syscore_ops = { - .suspend = acpi_processor_suspend, - .resume = acpi_processor_resume, -}; - -void acpi_processor_syscore_init(void) -{ - register_syscore_ops(&acpi_processor_syscore_ops); -} - -void acpi_processor_syscore_exit(void) -{ - unregister_syscore_ops(&acpi_processor_syscore_ops); -} -#endif /* CONFIG_PM_SLEEP */ - #if defined(CONFIG_X86) static void tsc_check_state(int state) { @@ -351,7 +313,7 @@ static int acpi_processor_get_power_info_cst(struct acpi_processor *pr) /* There must be at least 2 elements */ if (!cst || (cst->type != ACPI_TYPE_PACKAGE) || cst->package.count < 2) { - printk(KERN_ERR PREFIX "not enough elements in _CST\n"); + pr_err("not enough elements in _CST\n"); ret = -EFAULT; goto end; } @@ -360,7 +322,7 @@ static int acpi_processor_get_power_info_cst(struct acpi_processor *pr) /* Validate number of power states. */ if (count < 1 || count != cst->package.count - 1) { - printk(KERN_ERR PREFIX "count given by _CST is not valid\n"); + pr_err("count given by _CST is not valid\n"); ret = -EFAULT; goto end; } @@ -469,11 +431,9 @@ static int acpi_processor_get_power_info_cst(struct acpi_processor *pr) * (From 1 through ACPI_PROCESSOR_MAX_POWER - 1) */ if (current_count >= (ACPI_PROCESSOR_MAX_POWER - 1)) { - printk(KERN_WARNING - "Limiting number of power states to max (%d)\n", - ACPI_PROCESSOR_MAX_POWER); - printk(KERN_WARNING - "Please increase ACPI_PROCESSOR_MAX_POWER if needed.\n"); + pr_warn("Limiting number of power states to max (%d)\n", + ACPI_PROCESSOR_MAX_POWER); + pr_warn("Please increase ACPI_PROCESSOR_MAX_POWER if needed.\n"); break; } } @@ -1097,8 +1057,8 @@ int acpi_processor_power_init(struct acpi_processor *pr) retval = cpuidle_register_driver(&acpi_idle_driver); if (retval) return retval; - printk(KERN_DEBUG "ACPI: %s registered with cpuidle\n", - acpi_idle_driver.name); + pr_debug("%s registered with cpuidle\n", + acpi_idle_driver.name); } dev = kzalloc(sizeof(*dev), GFP_KERNEL); diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c index 407a376..5f28cf7 100644 --- a/drivers/acpi/scan.c +++ b/drivers/acpi/scan.c @@ -1930,6 +1930,7 @@ int __init acpi_scan_init(void) acpi_memory_hotplug_init(); acpi_pnp_init(); acpi_int340x_thermal_init(); + acpi_amba_init(); acpi_scan_add_handler(&generic_device_handler); diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c index 9cb9752..fbfcce3 100644 --- a/drivers/acpi/sleep.c +++ b/drivers/acpi/sleep.c @@ -19,6 +19,7 @@ #include <linux/reboot.h> #include <linux/acpi.h> #include <linux/module.h> +#include <linux/syscore_ops.h> #include <asm/io.h> #include <trace/events/power.h> @@ -677,6 +678,39 @@ static void acpi_sleep_suspend_setup(void) static inline void acpi_sleep_suspend_setup(void) {} #endif /* !CONFIG_SUSPEND */ +#ifdef CONFIG_PM_SLEEP +static u32 saved_bm_rld; + +static int acpi_save_bm_rld(void) +{ + acpi_read_bit_register(ACPI_BITREG_BUS_MASTER_RLD, &saved_bm_rld); + return 0; +} + +static void acpi_restore_bm_rld(void) +{ + u32 resumed_bm_rld = 0; + + acpi_read_bit_register(ACPI_BITREG_BUS_MASTER_RLD, &resumed_bm_rld); + if (resumed_bm_rld == saved_bm_rld) + return; + + acpi_write_bit_register(ACPI_BITREG_BUS_MASTER_RLD, saved_bm_rld); +} + +static struct syscore_ops acpi_sleep_syscore_ops = { + .suspend = acpi_save_bm_rld, + .resume = acpi_restore_bm_rld, +}; + +void acpi_sleep_syscore_init(void) +{ + register_syscore_ops(&acpi_sleep_syscore_ops); +} +#else +static inline void acpi_sleep_syscore_init(void) {} +#endif /* CONFIG_PM_SLEEP */ + #ifdef CONFIG_HIBERNATION static unsigned long s4_hardware_signature; static struct acpi_table_facs *facs; @@ -839,6 +873,7 @@ int __init acpi_sleep_init(void) sleep_states[ACPI_STATE_S0] = 1; + acpi_sleep_syscore_init(); acpi_sleep_suspend_setup(); acpi_sleep_hibernate_setup(); diff --git a/drivers/acpi/tables.c b/drivers/acpi/tables.c index 6c0f079..f49c024 100644 --- a/drivers/acpi/tables.c +++ b/drivers/acpi/tables.c @@ -32,6 +32,7 @@ #include <linux/errno.h> #include <linux/acpi.h> #include <linux/bootmem.h> +#include "internal.h" #define ACPI_MAX_TABLES 128 @@ -456,6 +457,7 @@ int __init acpi_table_init(void) status = acpi_initialize_tables(initial_tables, ACPI_MAX_TABLES, 0); if (ACPI_FAILURE(status)) return -EINVAL; + acpi_initrd_initialize_tables(); check_multiple_madt(); return 0; @@ -484,3 +486,13 @@ static int __init acpi_force_table_verification_setup(char *s) } early_param("acpi_force_table_verification", acpi_force_table_verification_setup); + +static int __init acpi_force_32bit_fadt_addr(char *s) +{ + pr_info("Forcing 32 Bit FADT addresses\n"); + acpi_gbl_use32_bit_fadt_addresses = TRUE; + + return 0; +} + +early_param("acpi_force_32bit_fadt_addr", acpi_force_32bit_fadt_addr); diff --git a/drivers/acpi/utils.c b/drivers/acpi/utils.c index f2f9873..f12a724 100644 --- a/drivers/acpi/utils.c +++ b/drivers/acpi/utils.c @@ -201,10 +201,6 @@ acpi_extract_package(union acpi_object *package, u8 **pointer = NULL; union acpi_object *element = &(package->package.elements[i]); - if (!element) { - return AE_BAD_DATA; - } - switch (element->type) { case ACPI_TYPE_INTEGER: diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c index 301b785..56705b5 100644 --- a/drivers/base/power/domain.c +++ b/drivers/base/power/domain.c @@ -104,6 +104,7 @@ static void genpd_sd_counter_inc(struct generic_pm_domain *genpd) static int genpd_power_on(struct generic_pm_domain *genpd, bool timed) { + unsigned int state_idx = genpd->state_idx; ktime_t time_start; s64 elapsed_ns; int ret; @@ -120,10 +121,10 @@ static int genpd_power_on(struct generic_pm_domain *genpd, bool timed) return ret; elapsed_ns = ktime_to_ns(ktime_sub(ktime_get(), time_start)); - if (elapsed_ns <= genpd->power_on_latency_ns) + if (elapsed_ns <= genpd->states[state_idx].power_on_latency_ns) return ret; - genpd->power_on_latency_ns = elapsed_ns; + genpd->states[state_idx].power_on_latency_ns = elapsed_ns; genpd->max_off_time_changed = true; pr_debug("%s: Power-%s latency exceeded, new value %lld ns\n", genpd->name, "on", elapsed_ns); @@ -133,6 +134,7 @@ static int genpd_power_on(struct generic_pm_domain *genpd, bool timed) static int genpd_power_off(struct generic_pm_domain *genpd, bool timed) { + unsigned int state_idx = genpd->state_idx; ktime_t time_start; s64 elapsed_ns; int ret; @@ -149,10 +151,10 @@ static int genpd_power_off(struct generic_pm_domain *genpd, bool timed) return ret; elapsed_ns = ktime_to_ns(ktime_sub(ktime_get(), time_start)); - if (elapsed_ns <= genpd->power_off_latency_ns) + if (elapsed_ns <= genpd->states[state_idx].power_off_latency_ns) return ret; - genpd->power_off_latency_ns = elapsed_ns; + genpd->states[state_idx].power_off_latency_ns = elapsed_ns; genpd->max_off_time_changed = true; pr_debug("%s: Power-%s latency exceeded, new value %lld ns\n", genpd->name, "off", elapsed_ns); @@ -485,8 +487,13 @@ static int pm_genpd_runtime_resume(struct device *dev) if (timed && runtime_pm) time_start = ktime_get(); - genpd_start_dev(genpd, dev); - genpd_restore_dev(genpd, dev); + ret = genpd_start_dev(genpd, dev); + if (ret) + goto err_poweroff; + + ret = genpd_restore_dev(genpd, dev); + if (ret) + goto err_stop; /* Update resume latency value if the measured time exceeds it. */ if (timed && runtime_pm) { @@ -501,6 +508,17 @@ static int pm_genpd_runtime_resume(struct device *dev) } return 0; + +err_stop: + genpd_stop_dev(genpd, dev); +err_poweroff: + if (!dev->power.irq_safe) { + mutex_lock(&genpd->lock); + genpd_poweroff(genpd, 0); + mutex_unlock(&genpd->lock); + } + + return ret; } static bool pd_ignore_unused; @@ -585,6 +603,8 @@ static void pm_genpd_sync_poweroff(struct generic_pm_domain *genpd, || atomic_read(&genpd->sd_count) > 0) return; + /* Choose the deepest state when suspending */ + genpd->state_idx = genpd->state_count - 1; genpd_power_off(genpd, timed); genpd->status = GPD_STATE_POWER_OFF; @@ -1378,7 +1398,7 @@ int pm_genpd_remove_subdomain(struct generic_pm_domain *genpd, mutex_lock(&subdomain->lock); mutex_lock_nested(&genpd->lock, SINGLE_DEPTH_NESTING); - if (!list_empty(&subdomain->slave_links) || subdomain->device_count) { + if (!list_empty(&subdomain->master_links) || subdomain->device_count) { pr_warn("%s: unable to remove subdomain %s\n", genpd->name, subdomain->name); ret = -EBUSY; @@ -1508,6 +1528,20 @@ void pm_genpd_init(struct generic_pm_domain *genpd, genpd->dev_ops.start = pm_clk_resume; } + if (genpd->state_idx >= GENPD_MAX_NUM_STATES) { + pr_warn("Initial state index out of bounds.\n"); + genpd->state_idx = GENPD_MAX_NUM_STATES - 1; + } + + if (genpd->state_count > GENPD_MAX_NUM_STATES) { + pr_warn("Limiting states to %d\n", GENPD_MAX_NUM_STATES); + genpd->state_count = GENPD_MAX_NUM_STATES; + } + + /* Use only one "off" state if there were no states declared */ + if (genpd->state_count == 0) + genpd->state_count = 1; + mutex_lock(&gpd_list_lock); list_add(&genpd->gpd_list_node, &gpd_list); mutex_unlock(&gpd_list_lock); @@ -1668,6 +1702,9 @@ struct generic_pm_domain *of_genpd_get_from_provider( struct generic_pm_domain *genpd = ERR_PTR(-ENOENT); struct of_genpd_provider *provider; + if (!genpdspec) + return ERR_PTR(-EINVAL); + mutex_lock(&of_genpd_mutex); /* Check if we have such a provider in our array */ @@ -1864,6 +1901,7 @@ static int pm_genpd_summary_one(struct seq_file *s, struct pm_domain_data *pm_data; const char *kobj_path; struct gpd_link *link; + char state[16]; int ret; ret = mutex_lock_interruptible(&genpd->lock); @@ -1872,7 +1910,13 @@ static int pm_genpd_summary_one(struct seq_file *s, if (WARN_ON(genpd->status >= ARRAY_SIZE(status_lookup))) goto exit; - seq_printf(s, "%-30s %-15s ", genpd->name, status_lookup[genpd->status]); + if (genpd->status == GPD_STATE_POWER_OFF) + snprintf(state, sizeof(state), "%s-%u", + status_lookup[genpd->status], genpd->state_idx); + else + snprintf(state, sizeof(state), "%s", + status_lookup[genpd->status]); + seq_printf(s, "%-30s %-15s ", genpd->name, state); /* * Modifications on the list require holding locks on both diff --git a/drivers/base/power/domain_governor.c b/drivers/base/power/domain_governor.c index 1e937ac..00a5436 100644 --- a/drivers/base/power/domain_governor.c +++ b/drivers/base/power/domain_governor.c @@ -98,7 +98,8 @@ static bool default_stop_ok(struct device *dev) * * This routine must be executed under the PM domain's lock. */ -static bool default_power_down_ok(struct dev_pm_domain *pd) +static bool __default_power_down_ok(struct dev_pm_domain *pd, + unsigned int state) { struct generic_pm_domain *genpd = pd_to_genpd(pd); struct gpd_link *link; @@ -106,27 +107,9 @@ static bool default_power_down_ok(struct dev_pm_domain *pd) s64 min_off_time_ns; s64 off_on_time_ns; - if (genpd->max_off_time_changed) { - struct gpd_link *link; - - /* - * We have to invalidate the cached results for the masters, so - * use the observation that default_power_down_ok() is not - * going to be called for any master until this instance - * returns. - */ - list_for_each_entry(link, &genpd->slave_links, slave_node) - link->master->max_off_time_changed = true; - - genpd->max_off_time_changed = false; - genpd->cached_power_down_ok = false; - genpd->max_off_time_ns = -1; - } else { - return genpd->cached_power_down_ok; - } + off_on_time_ns = genpd->states[state].power_off_latency_ns + + genpd->states[state].power_on_latency_ns; - off_on_time_ns = genpd->power_off_latency_ns + - genpd->power_on_latency_ns; min_off_time_ns = -1; /* @@ -186,8 +169,6 @@ static bool default_power_down_ok(struct dev_pm_domain *pd) min_off_time_ns = constraint_ns; } - genpd->cached_power_down_ok = true; - /* * If the computed minimum device off time is negative, there are no * latency constraints, so the domain can spend arbitrary time in the @@ -201,10 +182,45 @@ static bool default_power_down_ok(struct dev_pm_domain *pd) * time and the time needed to turn the domain on is the maximum * theoretical time this domain can spend in the "off" state. */ - genpd->max_off_time_ns = min_off_time_ns - genpd->power_on_latency_ns; + genpd->max_off_time_ns = min_off_time_ns - + genpd->states[state].power_on_latency_ns; return true; } +static bool default_power_down_ok(struct dev_pm_domain *pd) +{ + struct generic_pm_domain *genpd = pd_to_genpd(pd); + struct gpd_link *link; + + if (!genpd->max_off_time_changed) + return genpd->cached_power_down_ok; + + /* + * We have to invalidate the cached results for the masters, so + * use the observation that default_power_down_ok() is not + * going to be called for any master until this instance + * returns. + */ + list_for_each_entry(link, &genpd->slave_links, slave_node) + link->master->max_off_time_changed = true; + + genpd->max_off_time_ns = -1; + genpd->max_off_time_changed = false; + genpd->cached_power_down_ok = true; + genpd->state_idx = genpd->state_count - 1; + + /* Find a state to power down to, starting from the deepest. */ + while (!__default_power_down_ok(pd, genpd->state_idx)) { + if (genpd->state_idx == 0) { + genpd->cached_power_down_ok = false; + break; + } + genpd->state_idx--; + } + + return genpd->cached_power_down_ok; +} + static bool always_on_power_down_ok(struct dev_pm_domain *domain) { return false; diff --git a/drivers/base/power/opp/core.c b/drivers/base/power/opp/core.c index cf351d3..433b600 100644 --- a/drivers/base/power/opp/core.c +++ b/drivers/base/power/opp/core.c @@ -13,50 +13,52 @@ #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt +#include <linux/clk.h> #include <linux/errno.h> #include <linux/err.h> #include <linux/slab.h> #include <linux/device.h> #include <linux/of.h> #include <linux/export.h> +#include <linux/regulator/consumer.h> #include "opp.h" /* - * The root of the list of all devices. All device_opp structures branch off - * from here, with each device_opp containing the list of opp it supports in + * The root of the list of all opp-tables. All opp_table structures branch off + * from here, with each opp_table containing the list of opps it supports in * various states of availability. */ -static LIST_HEAD(dev_opp_list); +static LIST_HEAD(opp_tables); /* Lock to allow exclusive modification to the device and opp lists */ -DEFINE_MUTEX(dev_opp_list_lock); +DEFINE_MUTEX(opp_table_lock); #define opp_rcu_lockdep_assert() \ do { \ RCU_LOCKDEP_WARN(!rcu_read_lock_held() && \ - !lockdep_is_held(&dev_opp_list_lock), \ - "Missing rcu_read_lock() or " \ - "dev_opp_list_lock protection"); \ + !lockdep_is_held(&opp_table_lock), \ + "Missing rcu_read_lock() or " \ + "opp_table_lock protection"); \ } while (0) -static struct device_list_opp *_find_list_dev(const struct device *dev, - struct device_opp *dev_opp) +static struct opp_device *_find_opp_dev(const struct device *dev, + struct opp_table *opp_table) { - struct device_list_opp *list_dev; + struct opp_device *opp_dev; - list_for_each_entry(list_dev, &dev_opp->dev_list, node) - if (list_dev->dev == dev) - return list_dev; + list_for_each_entry(opp_dev, &opp_table->dev_list, node) + if (opp_dev->dev == dev) + return opp_dev; return NULL; } -static struct device_opp *_managed_opp(const struct device_node *np) +static struct opp_table *_managed_opp(const struct device_node *np) { - struct device_opp *dev_opp; + struct opp_table *opp_table; - list_for_each_entry_rcu(dev_opp, &dev_opp_list, node) { - if (dev_opp->np == np) { + list_for_each_entry_rcu(opp_table, &opp_tables, node) { + if (opp_table->np == np) { /* * Multiple devices can point to the same OPP table and * so will have same node-pointer, np. @@ -64,7 +66,7 @@ static struct device_opp *_managed_opp(const struct device_node *np) * But the OPPs will be considered as shared only if the * OPP table contains a "opp-shared" property. */ - return dev_opp->shared_opp ? dev_opp : NULL; + return opp_table->shared_opp ? opp_table : NULL; } } @@ -72,24 +74,24 @@ static struct device_opp *_managed_opp(const struct device_node *np) } /** - * _find_device_opp() - find device_opp struct using device pointer - * @dev: device pointer used to lookup device OPPs + * _find_opp_table() - find opp_table struct using device pointer + * @dev: device pointer used to lookup OPP table * - * Search list of device OPPs for one containing matching device. Does a RCU - * reader operation to grab the pointer needed. + * Search OPP table for one containing matching device. Does a RCU reader + * operation to grab the pointer needed. * - * Return: pointer to 'struct device_opp' if found, otherwise -ENODEV or + * Return: pointer to 'struct opp_table' if found, otherwise -ENODEV or * -EINVAL based on type of error. * * Locking: For readers, this function must be called under rcu_read_lock(). - * device_opp is a RCU protected pointer, which means that device_opp is valid + * opp_table is a RCU protected pointer, which means that opp_table is valid * as long as we are under RCU lock. * - * For Writers, this function must be called with dev_opp_list_lock held. + * For Writers, this function must be called with opp_table_lock held. */ -struct device_opp *_find_device_opp(struct device *dev) +struct opp_table *_find_opp_table(struct device *dev) { - struct device_opp *dev_opp; + struct opp_table *opp_table; opp_rcu_lockdep_assert(); @@ -98,9 +100,9 @@ struct device_opp *_find_device_opp(struct device *dev) return ERR_PTR(-EINVAL); } - list_for_each_entry_rcu(dev_opp, &dev_opp_list, node) - if (_find_list_dev(dev, dev_opp)) - return dev_opp; + list_for_each_entry_rcu(opp_table, &opp_tables, node) + if (_find_opp_dev(dev, opp_table)) + return opp_table; return ERR_PTR(-ENODEV); } @@ -213,16 +215,16 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_is_turbo); */ unsigned long dev_pm_opp_get_max_clock_latency(struct device *dev) { - struct device_opp *dev_opp; + struct opp_table *opp_table; unsigned long clock_latency_ns; rcu_read_lock(); - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp)) + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) clock_latency_ns = 0; else - clock_latency_ns = dev_opp->clock_latency_ns_max; + clock_latency_ns = opp_table->clock_latency_ns_max; rcu_read_unlock(); return clock_latency_ns; @@ -230,6 +232,82 @@ unsigned long dev_pm_opp_get_max_clock_latency(struct device *dev) EXPORT_SYMBOL_GPL(dev_pm_opp_get_max_clock_latency); /** + * dev_pm_opp_get_max_volt_latency() - Get max voltage latency in nanoseconds + * @dev: device for which we do this operation + * + * Return: This function returns the max voltage latency in nanoseconds. + * + * Locking: This function takes rcu_read_lock(). + */ +unsigned long dev_pm_opp_get_max_volt_latency(struct device *dev) +{ + struct opp_table *opp_table; + struct dev_pm_opp *opp; + struct regulator *reg; + unsigned long latency_ns = 0; + unsigned long min_uV = ~0, max_uV = 0; + int ret; + + rcu_read_lock(); + + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) { + rcu_read_unlock(); + return 0; + } + + reg = opp_table->regulator; + if (IS_ERR(reg)) { + /* Regulator may not be required for device */ + if (reg) + dev_err(dev, "%s: Invalid regulator (%ld)\n", __func__, + PTR_ERR(reg)); + rcu_read_unlock(); + return 0; + } + + list_for_each_entry_rcu(opp, &opp_table->opp_list, node) { + if (!opp->available) + continue; + + if (opp->u_volt_min < min_uV) + min_uV = opp->u_volt_min; + if (opp->u_volt_max > max_uV) + max_uV = opp->u_volt_max; + } + + rcu_read_unlock(); + + /* + * The caller needs to ensure that opp_table (and hence the regulator) + * isn't freed, while we are executing this routine. + */ + ret = regulator_set_voltage_time(reg, min_uV, max_uV); + if (ret > 0) + latency_ns = ret * 1000; + + return latency_ns; +} +EXPORT_SYMBOL_GPL(dev_pm_opp_get_max_volt_latency); + +/** + * dev_pm_opp_get_max_transition_latency() - Get max transition latency in + * nanoseconds + * @dev: device for which we do this operation + * + * Return: This function returns the max transition latency, in nanoseconds, to + * switch from one OPP to other. + * + * Locking: This function takes rcu_read_lock(). + */ +unsigned long dev_pm_opp_get_max_transition_latency(struct device *dev) +{ + return dev_pm_opp_get_max_volt_latency(dev) + + dev_pm_opp_get_max_clock_latency(dev); +} +EXPORT_SYMBOL_GPL(dev_pm_opp_get_max_transition_latency); + +/** * dev_pm_opp_get_suspend_opp() - Get suspend opp * @dev: device for which we do this operation * @@ -244,21 +322,21 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_get_max_clock_latency); */ struct dev_pm_opp *dev_pm_opp_get_suspend_opp(struct device *dev) { - struct device_opp *dev_opp; + struct opp_table *opp_table; opp_rcu_lockdep_assert(); - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp) || !dev_opp->suspend_opp || - !dev_opp->suspend_opp->available) + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table) || !opp_table->suspend_opp || + !opp_table->suspend_opp->available) return NULL; - return dev_opp->suspend_opp; + return opp_table->suspend_opp; } EXPORT_SYMBOL_GPL(dev_pm_opp_get_suspend_opp); /** - * dev_pm_opp_get_opp_count() - Get number of opps available in the opp list + * dev_pm_opp_get_opp_count() - Get number of opps available in the opp table * @dev: device for which we do this operation * * Return: This function returns the number of available opps if there are any, @@ -268,21 +346,21 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_get_suspend_opp); */ int dev_pm_opp_get_opp_count(struct device *dev) { - struct device_opp *dev_opp; + struct opp_table *opp_table; struct dev_pm_opp *temp_opp; int count = 0; rcu_read_lock(); - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp)) { - count = PTR_ERR(dev_opp); - dev_err(dev, "%s: device OPP not found (%d)\n", + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) { + count = PTR_ERR(opp_table); + dev_err(dev, "%s: OPP table not found (%d)\n", __func__, count); goto out_unlock; } - list_for_each_entry_rcu(temp_opp, &dev_opp->opp_list, node) { + list_for_each_entry_rcu(temp_opp, &opp_table->opp_list, node) { if (temp_opp->available) count++; } @@ -299,7 +377,7 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_get_opp_count); * @freq: frequency to search for * @available: true/false - match for available opp * - * Return: Searches for exact match in the opp list and returns pointer to the + * Return: Searches for exact match in the opp table and returns pointer to the * matching opp if found, else returns ERR_PTR in case of error and should * be handled using IS_ERR. Error return values can be: * EINVAL: for bad pointer @@ -323,19 +401,20 @@ struct dev_pm_opp *dev_pm_opp_find_freq_exact(struct device *dev, unsigned long freq, bool available) { - struct device_opp *dev_opp; + struct opp_table *opp_table; struct dev_pm_opp *temp_opp, *opp = ERR_PTR(-ERANGE); opp_rcu_lockdep_assert(); - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp)) { - int r = PTR_ERR(dev_opp); - dev_err(dev, "%s: device OPP not found (%d)\n", __func__, r); + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) { + int r = PTR_ERR(opp_table); + + dev_err(dev, "%s: OPP table not found (%d)\n", __func__, r); return ERR_PTR(r); } - list_for_each_entry_rcu(temp_opp, &dev_opp->opp_list, node) { + list_for_each_entry_rcu(temp_opp, &opp_table->opp_list, node) { if (temp_opp->available == available && temp_opp->rate == freq) { opp = temp_opp; @@ -371,7 +450,7 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_find_freq_exact); struct dev_pm_opp *dev_pm_opp_find_freq_ceil(struct device *dev, unsigned long *freq) { - struct device_opp *dev_opp; + struct opp_table *opp_table; struct dev_pm_opp *temp_opp, *opp = ERR_PTR(-ERANGE); opp_rcu_lockdep_assert(); @@ -381,11 +460,11 @@ struct dev_pm_opp *dev_pm_opp_find_freq_ceil(struct device *dev, return ERR_PTR(-EINVAL); } - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp)) - return ERR_CAST(dev_opp); + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) + return ERR_CAST(opp_table); - list_for_each_entry_rcu(temp_opp, &dev_opp->opp_list, node) { + list_for_each_entry_rcu(temp_opp, &opp_table->opp_list, node) { if (temp_opp->available && temp_opp->rate >= *freq) { opp = temp_opp; *freq = opp->rate; @@ -421,7 +500,7 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_find_freq_ceil); struct dev_pm_opp *dev_pm_opp_find_freq_floor(struct device *dev, unsigned long *freq) { - struct device_opp *dev_opp; + struct opp_table *opp_table; struct dev_pm_opp *temp_opp, *opp = ERR_PTR(-ERANGE); opp_rcu_lockdep_assert(); @@ -431,11 +510,11 @@ struct dev_pm_opp *dev_pm_opp_find_freq_floor(struct device *dev, return ERR_PTR(-EINVAL); } - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp)) - return ERR_CAST(dev_opp); + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) + return ERR_CAST(opp_table); - list_for_each_entry_rcu(temp_opp, &dev_opp->opp_list, node) { + list_for_each_entry_rcu(temp_opp, &opp_table->opp_list, node) { if (temp_opp->available) { /* go to the next node, before choosing prev */ if (temp_opp->rate > *freq) @@ -451,130 +530,343 @@ struct dev_pm_opp *dev_pm_opp_find_freq_floor(struct device *dev, } EXPORT_SYMBOL_GPL(dev_pm_opp_find_freq_floor); -/* List-dev Helpers */ -static void _kfree_list_dev_rcu(struct rcu_head *head) +/* + * The caller needs to ensure that opp_table (and hence the clk) isn't freed, + * while clk returned here is used. + */ +static struct clk *_get_opp_clk(struct device *dev) +{ + struct opp_table *opp_table; + struct clk *clk; + + rcu_read_lock(); + + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) { + dev_err(dev, "%s: device opp doesn't exist\n", __func__); + clk = ERR_CAST(opp_table); + goto unlock; + } + + clk = opp_table->clk; + if (IS_ERR(clk)) + dev_err(dev, "%s: No clock available for the device\n", + __func__); + +unlock: + rcu_read_unlock(); + return clk; +} + +static int _set_opp_voltage(struct device *dev, struct regulator *reg, + unsigned long u_volt, unsigned long u_volt_min, + unsigned long u_volt_max) +{ + int ret; + + /* Regulator not available for device */ + if (IS_ERR(reg)) { + dev_dbg(dev, "%s: regulator not available: %ld\n", __func__, + PTR_ERR(reg)); + return 0; + } + + dev_dbg(dev, "%s: voltages (mV): %lu %lu %lu\n", __func__, u_volt_min, + u_volt, u_volt_max); + + ret = regulator_set_voltage_triplet(reg, u_volt_min, u_volt, + u_volt_max); + if (ret) + dev_err(dev, "%s: failed to set voltage (%lu %lu %lu mV): %d\n", + __func__, u_volt_min, u_volt, u_volt_max, ret); + + return ret; +} + +/** + * dev_pm_opp_set_rate() - Configure new OPP based on frequency + * @dev: device for which we do this operation + * @target_freq: frequency to achieve + * + * This configures the power-supplies and clock source to the levels specified + * by the OPP corresponding to the target_freq. + * + * Locking: This function takes rcu_read_lock(). + */ +int dev_pm_opp_set_rate(struct device *dev, unsigned long target_freq) +{ + struct opp_table *opp_table; + struct dev_pm_opp *old_opp, *opp; + struct regulator *reg; + struct clk *clk; + unsigned long freq, old_freq; + unsigned long u_volt, u_volt_min, u_volt_max; + unsigned long ou_volt, ou_volt_min, ou_volt_max; + int ret; + + if (unlikely(!target_freq)) { + dev_err(dev, "%s: Invalid target frequency %lu\n", __func__, + target_freq); + return -EINVAL; + } + + clk = _get_opp_clk(dev); + if (IS_ERR(clk)) + return PTR_ERR(clk); + + freq = clk_round_rate(clk, target_freq); + if ((long)freq <= 0) + freq = target_freq; + + old_freq = clk_get_rate(clk); + + /* Return early if nothing to do */ + if (old_freq == freq) { + dev_dbg(dev, "%s: old/new frequencies (%lu Hz) are same, nothing to do\n", + __func__, freq); + return 0; + } + + rcu_read_lock(); + + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) { + dev_err(dev, "%s: device opp doesn't exist\n", __func__); + rcu_read_unlock(); + return PTR_ERR(opp_table); + } + + old_opp = dev_pm_opp_find_freq_ceil(dev, &old_freq); + if (!IS_ERR(old_opp)) { + ou_volt = old_opp->u_volt; + ou_volt_min = old_opp->u_volt_min; + ou_volt_max = old_opp->u_volt_max; + } else { + dev_err(dev, "%s: failed to find current OPP for freq %lu (%ld)\n", + __func__, old_freq, PTR_ERR(old_opp)); + } + + opp = dev_pm_opp_find_freq_ceil(dev, &freq); + if (IS_ERR(opp)) { + ret = PTR_ERR(opp); + dev_err(dev, "%s: failed to find OPP for freq %lu (%d)\n", + __func__, freq, ret); + rcu_read_unlock(); + return ret; + } + + u_volt = opp->u_volt; + u_volt_min = opp->u_volt_min; + u_volt_max = opp->u_volt_max; + + reg = opp_table->regulator; + + rcu_read_unlock(); + + /* Scaling up? Scale voltage before frequency */ + if (freq > old_freq) { + ret = _set_opp_voltage(dev, reg, u_volt, u_volt_min, + u_volt_max); + if (ret) + goto restore_voltage; + } + + /* Change frequency */ + + dev_dbg(dev, "%s: switching OPP: %lu Hz --> %lu Hz\n", + __func__, old_freq, freq); + + ret = clk_set_rate(clk, freq); + if (ret) { + dev_err(dev, "%s: failed to set clock rate: %d\n", __func__, + ret); + goto restore_voltage; + } + + /* Scaling down? Scale voltage after frequency */ + if (freq < old_freq) { + ret = _set_opp_voltage(dev, reg, u_volt, u_volt_min, + u_volt_max); + if (ret) + goto restore_freq; + } + + return 0; + +restore_freq: + if (clk_set_rate(clk, old_freq)) + dev_err(dev, "%s: failed to restore old-freq (%lu Hz)\n", + __func__, old_freq); +restore_voltage: + /* This shouldn't harm even if the voltages weren't updated earlier */ + if (!IS_ERR(old_opp)) + _set_opp_voltage(dev, reg, ou_volt, ou_volt_min, ou_volt_max); + + return ret; +} +EXPORT_SYMBOL_GPL(dev_pm_opp_set_rate); + +/* OPP-dev Helpers */ +static void _kfree_opp_dev_rcu(struct rcu_head *head) { - struct device_list_opp *list_dev; + struct opp_device *opp_dev; - list_dev = container_of(head, struct device_list_opp, rcu_head); - kfree_rcu(list_dev, rcu_head); + opp_dev = container_of(head, struct opp_device, rcu_head); + kfree_rcu(opp_dev, rcu_head); } -static void _remove_list_dev(struct device_list_opp *list_dev, - struct device_opp *dev_opp) +static void _remove_opp_dev(struct opp_device *opp_dev, + struct opp_table *opp_table) { - opp_debug_unregister(list_dev, dev_opp); - list_del(&list_dev->node); - call_srcu(&dev_opp->srcu_head.srcu, &list_dev->rcu_head, - _kfree_list_dev_rcu); + opp_debug_unregister(opp_dev, opp_table); + list_del(&opp_dev->node); + call_srcu(&opp_table->srcu_head.srcu, &opp_dev->rcu_head, + _kfree_opp_dev_rcu); } -struct device_list_opp *_add_list_dev(const struct device *dev, - struct device_opp *dev_opp) +struct opp_device *_add_opp_dev(const struct device *dev, + struct opp_table *opp_table) { - struct device_list_opp *list_dev; + struct opp_device *opp_dev; int ret; - list_dev = kzalloc(sizeof(*list_dev), GFP_KERNEL); - if (!list_dev) + opp_dev = kzalloc(sizeof(*opp_dev), GFP_KERNEL); + if (!opp_dev) return NULL; - /* Initialize list-dev */ - list_dev->dev = dev; - list_add_rcu(&list_dev->node, &dev_opp->dev_list); + /* Initialize opp-dev */ + opp_dev->dev = dev; + list_add_rcu(&opp_dev->node, &opp_table->dev_list); - /* Create debugfs entries for the dev_opp */ - ret = opp_debug_register(list_dev, dev_opp); + /* Create debugfs entries for the opp_table */ + ret = opp_debug_register(opp_dev, opp_table); if (ret) dev_err(dev, "%s: Failed to register opp debugfs (%d)\n", __func__, ret); - return list_dev; + return opp_dev; } /** - * _add_device_opp() - Find device OPP table or allocate a new one + * _add_opp_table() - Find OPP table or allocate a new one * @dev: device for which we do this operation * * It tries to find an existing table first, if it couldn't find one, it * allocates a new OPP table and returns that. * - * Return: valid device_opp pointer if success, else NULL. + * Return: valid opp_table pointer if success, else NULL. */ -static struct device_opp *_add_device_opp(struct device *dev) +static struct opp_table *_add_opp_table(struct device *dev) { - struct device_opp *dev_opp; - struct device_list_opp *list_dev; + struct opp_table *opp_table; + struct opp_device *opp_dev; + struct device_node *np; + int ret; - /* Check for existing list for 'dev' first */ - dev_opp = _find_device_opp(dev); - if (!IS_ERR(dev_opp)) - return dev_opp; + /* Check for existing table for 'dev' first */ + opp_table = _find_opp_table(dev); + if (!IS_ERR(opp_table)) + return opp_table; /* - * Allocate a new device OPP table. In the infrequent case where a new + * Allocate a new OPP table. In the infrequent case where a new * device is needed to be added, we pay this penalty. */ - dev_opp = kzalloc(sizeof(*dev_opp), GFP_KERNEL); - if (!dev_opp) + opp_table = kzalloc(sizeof(*opp_table), GFP_KERNEL); + if (!opp_table) return NULL; - INIT_LIST_HEAD(&dev_opp->dev_list); + INIT_LIST_HEAD(&opp_table->dev_list); - list_dev = _add_list_dev(dev, dev_opp); - if (!list_dev) { - kfree(dev_opp); + opp_dev = _add_opp_dev(dev, opp_table); + if (!opp_dev) { + kfree(opp_table); return NULL; } - srcu_init_notifier_head(&dev_opp->srcu_head); - INIT_LIST_HEAD(&dev_opp->opp_list); + /* + * Only required for backward compatibility with v1 bindings, but isn't + * harmful for other cases. And so we do it unconditionally. + */ + np = of_node_get(dev->of_node); + if (np) { + u32 val; + + if (!of_property_read_u32(np, "clock-latency", &val)) + opp_table->clock_latency_ns_max = val; + of_property_read_u32(np, "voltage-tolerance", + &opp_table->voltage_tolerance_v1); + of_node_put(np); + } + + /* Set regulator to a non-NULL error value */ + opp_table->regulator = ERR_PTR(-ENXIO); + + /* Find clk for the device */ + opp_table->clk = clk_get(dev, NULL); + if (IS_ERR(opp_table->clk)) { + ret = PTR_ERR(opp_table->clk); + if (ret != -EPROBE_DEFER) + dev_dbg(dev, "%s: Couldn't find clock: %d\n", __func__, + ret); + } - /* Secure the device list modification */ - list_add_rcu(&dev_opp->node, &dev_opp_list); - return dev_opp; + srcu_init_notifier_head(&opp_table->srcu_head); + INIT_LIST_HEAD(&opp_table->opp_list); + + /* Secure the device table modification */ + list_add_rcu(&opp_table->node, &opp_tables); + return opp_table; } /** - * _kfree_device_rcu() - Free device_opp RCU handler + * _kfree_device_rcu() - Free opp_table RCU handler * @head: RCU head */ static void _kfree_device_rcu(struct rcu_head *head) { - struct device_opp *device_opp = container_of(head, struct device_opp, rcu_head); + struct opp_table *opp_table = container_of(head, struct opp_table, + rcu_head); - kfree_rcu(device_opp, rcu_head); + kfree_rcu(opp_table, rcu_head); } /** - * _remove_device_opp() - Removes a device OPP table - * @dev_opp: device OPP table to be removed. + * _remove_opp_table() - Removes a OPP table + * @opp_table: OPP table to be removed. * - * Removes/frees device OPP table it it doesn't contain any OPPs. + * Removes/frees OPP table if it doesn't contain any OPPs. */ -static void _remove_device_opp(struct device_opp *dev_opp) +static void _remove_opp_table(struct opp_table *opp_table) { - struct device_list_opp *list_dev; + struct opp_device *opp_dev; + + if (!list_empty(&opp_table->opp_list)) + return; - if (!list_empty(&dev_opp->opp_list)) + if (opp_table->supported_hw) return; - if (dev_opp->supported_hw) + if (opp_table->prop_name) return; - if (dev_opp->prop_name) + if (!IS_ERR(opp_table->regulator)) return; - list_dev = list_first_entry(&dev_opp->dev_list, struct device_list_opp, - node); + /* Release clk */ + if (!IS_ERR(opp_table->clk)) + clk_put(opp_table->clk); - _remove_list_dev(list_dev, dev_opp); + opp_dev = list_first_entry(&opp_table->dev_list, struct opp_device, + node); + + _remove_opp_dev(opp_dev, opp_table); /* dev_list must be empty now */ - WARN_ON(!list_empty(&dev_opp->dev_list)); + WARN_ON(!list_empty(&opp_table->dev_list)); - list_del_rcu(&dev_opp->node); - call_srcu(&dev_opp->srcu_head.srcu, &dev_opp->rcu_head, + list_del_rcu(&opp_table->node); + call_srcu(&opp_table->srcu_head.srcu, &opp_table->rcu_head, _kfree_device_rcu); } @@ -591,17 +883,17 @@ static void _kfree_opp_rcu(struct rcu_head *head) /** * _opp_remove() - Remove an OPP from a table definition - * @dev_opp: points back to the device_opp struct this opp belongs to + * @opp_table: points back to the opp_table struct this opp belongs to * @opp: pointer to the OPP to remove * @notify: OPP_EVENT_REMOVE notification should be sent or not * - * This function removes an opp definition from the opp list. + * This function removes an opp definition from the opp table. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * It is assumed that the caller holds required mutex for an RCU updater * strategy. */ -static void _opp_remove(struct device_opp *dev_opp, +static void _opp_remove(struct opp_table *opp_table, struct dev_pm_opp *opp, bool notify) { /* @@ -609,22 +901,23 @@ static void _opp_remove(struct device_opp *dev_opp, * frequency/voltage list. */ if (notify) - srcu_notifier_call_chain(&dev_opp->srcu_head, OPP_EVENT_REMOVE, opp); + srcu_notifier_call_chain(&opp_table->srcu_head, + OPP_EVENT_REMOVE, opp); opp_debug_remove_one(opp); list_del_rcu(&opp->node); - call_srcu(&dev_opp->srcu_head.srcu, &opp->rcu_head, _kfree_opp_rcu); + call_srcu(&opp_table->srcu_head.srcu, &opp->rcu_head, _kfree_opp_rcu); - _remove_device_opp(dev_opp); + _remove_opp_table(opp_table); } /** - * dev_pm_opp_remove() - Remove an OPP from OPP list + * dev_pm_opp_remove() - Remove an OPP from OPP table * @dev: device for which we do this operation * @freq: OPP to remove with matching 'freq' * - * This function removes an opp from the opp list. + * This function removes an opp from the opp table. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function internally uses RCU updater strategy with mutex locks * to keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where @@ -633,17 +926,17 @@ static void _opp_remove(struct device_opp *dev_opp, void dev_pm_opp_remove(struct device *dev, unsigned long freq) { struct dev_pm_opp *opp; - struct device_opp *dev_opp; + struct opp_table *opp_table; bool found = false; - /* Hold our list modification lock here */ - mutex_lock(&dev_opp_list_lock); + /* Hold our table modification lock here */ + mutex_lock(&opp_table_lock); - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp)) + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) goto unlock; - list_for_each_entry(opp, &dev_opp->opp_list, node) { + list_for_each_entry(opp, &opp_table->opp_list, node) { if (opp->rate == freq) { found = true; break; @@ -656,14 +949,14 @@ void dev_pm_opp_remove(struct device *dev, unsigned long freq) goto unlock; } - _opp_remove(dev_opp, opp, true); + _opp_remove(opp_table, opp, true); unlock: - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); } EXPORT_SYMBOL_GPL(dev_pm_opp_remove); static struct dev_pm_opp *_allocate_opp(struct device *dev, - struct device_opp **dev_opp) + struct opp_table **opp_table) { struct dev_pm_opp *opp; @@ -674,8 +967,8 @@ static struct dev_pm_opp *_allocate_opp(struct device *dev, INIT_LIST_HEAD(&opp->node); - *dev_opp = _add_device_opp(dev); - if (!*dev_opp) { + *opp_table = _add_opp_table(dev); + if (!*opp_table) { kfree(opp); return NULL; } @@ -683,22 +976,38 @@ static struct dev_pm_opp *_allocate_opp(struct device *dev, return opp; } +static bool _opp_supported_by_regulators(struct dev_pm_opp *opp, + struct opp_table *opp_table) +{ + struct regulator *reg = opp_table->regulator; + + if (!IS_ERR(reg) && + !regulator_is_supported_voltage(reg, opp->u_volt_min, + opp->u_volt_max)) { + pr_warn("%s: OPP minuV: %lu maxuV: %lu, not supported by regulator\n", + __func__, opp->u_volt_min, opp->u_volt_max); + return false; + } + + return true; +} + static int _opp_add(struct device *dev, struct dev_pm_opp *new_opp, - struct device_opp *dev_opp) + struct opp_table *opp_table) { struct dev_pm_opp *opp; - struct list_head *head = &dev_opp->opp_list; + struct list_head *head = &opp_table->opp_list; int ret; /* * Insert new OPP in order of increasing frequency and discard if * already present. * - * Need to use &dev_opp->opp_list in the condition part of the 'for' + * Need to use &opp_table->opp_list in the condition part of the 'for' * loop, don't replace it with head otherwise it will become an infinite * loop. */ - list_for_each_entry_rcu(opp, &dev_opp->opp_list, node) { + list_for_each_entry_rcu(opp, &opp_table->opp_list, node) { if (new_opp->rate > opp->rate) { head = &opp->node; continue; @@ -716,14 +1025,20 @@ static int _opp_add(struct device *dev, struct dev_pm_opp *new_opp, 0 : -EEXIST; } - new_opp->dev_opp = dev_opp; + new_opp->opp_table = opp_table; list_add_rcu(&new_opp->node, head); - ret = opp_debug_create_one(new_opp, dev_opp); + ret = opp_debug_create_one(new_opp, opp_table); if (ret) dev_err(dev, "%s: Failed to register opp to debugfs (%d)\n", __func__, ret); + if (!_opp_supported_by_regulators(new_opp, opp_table)) { + new_opp->available = false; + dev_warn(dev, "%s: OPP not supported by regulators (%lu)\n", + __func__, new_opp->rate); + } + return 0; } @@ -734,14 +1049,14 @@ static int _opp_add(struct device *dev, struct dev_pm_opp *new_opp, * @u_volt: Voltage in uVolts for this OPP * @dynamic: Dynamically added OPPs. * - * This function adds an opp definition to the opp list and returns status. + * This function adds an opp definition to the opp table and returns status. * The opp is made available by default and it can be controlled using * dev_pm_opp_enable/disable functions and may be removed by dev_pm_opp_remove. * * NOTE: "dynamic" parameter impacts OPPs added by the dev_pm_opp_of_add_table * and freed by dev_pm_opp_of_remove_table. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function internally uses RCU updater strategy with mutex locks * to keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where @@ -757,14 +1072,15 @@ static int _opp_add(struct device *dev, struct dev_pm_opp *new_opp, static int _opp_add_v1(struct device *dev, unsigned long freq, long u_volt, bool dynamic) { - struct device_opp *dev_opp; + struct opp_table *opp_table; struct dev_pm_opp *new_opp; + unsigned long tol; int ret; - /* Hold our list modification lock here */ - mutex_lock(&dev_opp_list_lock); + /* Hold our table modification lock here */ + mutex_lock(&opp_table_lock); - new_opp = _allocate_opp(dev, &dev_opp); + new_opp = _allocate_opp(dev, &opp_table); if (!new_opp) { ret = -ENOMEM; goto unlock; @@ -772,33 +1088,36 @@ static int _opp_add_v1(struct device *dev, unsigned long freq, long u_volt, /* populate the opp table */ new_opp->rate = freq; + tol = u_volt * opp_table->voltage_tolerance_v1 / 100; new_opp->u_volt = u_volt; + new_opp->u_volt_min = u_volt - tol; + new_opp->u_volt_max = u_volt + tol; new_opp->available = true; new_opp->dynamic = dynamic; - ret = _opp_add(dev, new_opp, dev_opp); + ret = _opp_add(dev, new_opp, opp_table); if (ret) goto free_opp; - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); /* * Notify the changes in the availability of the operable * frequency/voltage list. */ - srcu_notifier_call_chain(&dev_opp->srcu_head, OPP_EVENT_ADD, new_opp); + srcu_notifier_call_chain(&opp_table->srcu_head, OPP_EVENT_ADD, new_opp); return 0; free_opp: - _opp_remove(dev_opp, new_opp, false); + _opp_remove(opp_table, new_opp, false); unlock: - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); return ret; } /* TODO: Support multiple regulators */ static int opp_parse_supplies(struct dev_pm_opp *opp, struct device *dev, - struct device_opp *dev_opp) + struct opp_table *opp_table) { u32 microvolt[3] = {0}; u32 val; @@ -807,9 +1126,9 @@ static int opp_parse_supplies(struct dev_pm_opp *opp, struct device *dev, char name[NAME_MAX]; /* Search for "opp-microvolt-<name>" */ - if (dev_opp->prop_name) { + if (opp_table->prop_name) { snprintf(name, sizeof(name), "opp-microvolt-%s", - dev_opp->prop_name); + opp_table->prop_name); prop = of_find_property(opp->np, name, NULL); } @@ -844,14 +1163,20 @@ static int opp_parse_supplies(struct dev_pm_opp *opp, struct device *dev, } opp->u_volt = microvolt[0]; - opp->u_volt_min = microvolt[1]; - opp->u_volt_max = microvolt[2]; + + if (count == 1) { + opp->u_volt_min = opp->u_volt; + opp->u_volt_max = opp->u_volt; + } else { + opp->u_volt_min = microvolt[1]; + opp->u_volt_max = microvolt[2]; + } /* Search for "opp-microamp-<name>" */ prop = NULL; - if (dev_opp->prop_name) { + if (opp_table->prop_name) { snprintf(name, sizeof(name), "opp-microamp-%s", - dev_opp->prop_name); + opp_table->prop_name); prop = of_find_property(opp->np, name, NULL); } @@ -878,7 +1203,7 @@ static int opp_parse_supplies(struct dev_pm_opp *opp, struct device *dev, * OPPs, which are available for those versions, based on its 'opp-supported-hw' * property. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function internally uses RCU updater strategy with mutex locks * to keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where @@ -887,44 +1212,44 @@ static int opp_parse_supplies(struct dev_pm_opp *opp, struct device *dev, int dev_pm_opp_set_supported_hw(struct device *dev, const u32 *versions, unsigned int count) { - struct device_opp *dev_opp; + struct opp_table *opp_table; int ret = 0; - /* Hold our list modification lock here */ - mutex_lock(&dev_opp_list_lock); + /* Hold our table modification lock here */ + mutex_lock(&opp_table_lock); - dev_opp = _add_device_opp(dev); - if (!dev_opp) { + opp_table = _add_opp_table(dev); + if (!opp_table) { ret = -ENOMEM; goto unlock; } - /* Make sure there are no concurrent readers while updating dev_opp */ - WARN_ON(!list_empty(&dev_opp->opp_list)); + /* Make sure there are no concurrent readers while updating opp_table */ + WARN_ON(!list_empty(&opp_table->opp_list)); - /* Do we already have a version hierarchy associated with dev_opp? */ - if (dev_opp->supported_hw) { + /* Do we already have a version hierarchy associated with opp_table? */ + if (opp_table->supported_hw) { dev_err(dev, "%s: Already have supported hardware list\n", __func__); ret = -EBUSY; goto err; } - dev_opp->supported_hw = kmemdup(versions, count * sizeof(*versions), + opp_table->supported_hw = kmemdup(versions, count * sizeof(*versions), GFP_KERNEL); - if (!dev_opp->supported_hw) { + if (!opp_table->supported_hw) { ret = -ENOMEM; goto err; } - dev_opp->supported_hw_count = count; - mutex_unlock(&dev_opp_list_lock); + opp_table->supported_hw_count = count; + mutex_unlock(&opp_table_lock); return 0; err: - _remove_device_opp(dev_opp); + _remove_opp_table(opp_table); unlock: - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); return ret; } @@ -932,13 +1257,13 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_set_supported_hw); /** * dev_pm_opp_put_supported_hw() - Releases resources blocked for supported hw - * @dev: Device for which supported-hw has to be set. + * @dev: Device for which supported-hw has to be put. * * This is required only for the V2 bindings, and is called for a matching - * dev_pm_opp_set_supported_hw(). Until this is called, the device_opp structure + * dev_pm_opp_set_supported_hw(). Until this is called, the opp_table structure * will not be freed. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function internally uses RCU updater strategy with mutex locks * to keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where @@ -946,42 +1271,43 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_set_supported_hw); */ void dev_pm_opp_put_supported_hw(struct device *dev) { - struct device_opp *dev_opp; + struct opp_table *opp_table; - /* Hold our list modification lock here */ - mutex_lock(&dev_opp_list_lock); + /* Hold our table modification lock here */ + mutex_lock(&opp_table_lock); - /* Check for existing list for 'dev' first */ - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp)) { - dev_err(dev, "Failed to find dev_opp: %ld\n", PTR_ERR(dev_opp)); + /* Check for existing table for 'dev' first */ + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) { + dev_err(dev, "Failed to find opp_table: %ld\n", + PTR_ERR(opp_table)); goto unlock; } - /* Make sure there are no concurrent readers while updating dev_opp */ - WARN_ON(!list_empty(&dev_opp->opp_list)); + /* Make sure there are no concurrent readers while updating opp_table */ + WARN_ON(!list_empty(&opp_table->opp_list)); - if (!dev_opp->supported_hw) { + if (!opp_table->supported_hw) { dev_err(dev, "%s: Doesn't have supported hardware list\n", __func__); goto unlock; } - kfree(dev_opp->supported_hw); - dev_opp->supported_hw = NULL; - dev_opp->supported_hw_count = 0; + kfree(opp_table->supported_hw); + opp_table->supported_hw = NULL; + opp_table->supported_hw_count = 0; - /* Try freeing device_opp if this was the last blocking resource */ - _remove_device_opp(dev_opp); + /* Try freeing opp_table if this was the last blocking resource */ + _remove_opp_table(opp_table); unlock: - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); } EXPORT_SYMBOL_GPL(dev_pm_opp_put_supported_hw); /** * dev_pm_opp_set_prop_name() - Set prop-extn name - * @dev: Device for which the regulator has to be set. + * @dev: Device for which the prop-name has to be set. * @name: name to postfix to properties. * * This is required only for the V2 bindings, and it enables a platform to @@ -989,7 +1315,7 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_put_supported_hw); * which the extension will apply are opp-microvolt and opp-microamp. OPP core * should postfix the property name with -<name> while looking for them. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function internally uses RCU updater strategy with mutex locks * to keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where @@ -997,42 +1323,42 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_put_supported_hw); */ int dev_pm_opp_set_prop_name(struct device *dev, const char *name) { - struct device_opp *dev_opp; + struct opp_table *opp_table; int ret = 0; - /* Hold our list modification lock here */ - mutex_lock(&dev_opp_list_lock); + /* Hold our table modification lock here */ + mutex_lock(&opp_table_lock); - dev_opp = _add_device_opp(dev); - if (!dev_opp) { + opp_table = _add_opp_table(dev); + if (!opp_table) { ret = -ENOMEM; goto unlock; } - /* Make sure there are no concurrent readers while updating dev_opp */ - WARN_ON(!list_empty(&dev_opp->opp_list)); + /* Make sure there are no concurrent readers while updating opp_table */ + WARN_ON(!list_empty(&opp_table->opp_list)); - /* Do we already have a prop-name associated with dev_opp? */ - if (dev_opp->prop_name) { + /* Do we already have a prop-name associated with opp_table? */ + if (opp_table->prop_name) { dev_err(dev, "%s: Already have prop-name %s\n", __func__, - dev_opp->prop_name); + opp_table->prop_name); ret = -EBUSY; goto err; } - dev_opp->prop_name = kstrdup(name, GFP_KERNEL); - if (!dev_opp->prop_name) { + opp_table->prop_name = kstrdup(name, GFP_KERNEL); + if (!opp_table->prop_name) { ret = -ENOMEM; goto err; } - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); return 0; err: - _remove_device_opp(dev_opp); + _remove_opp_table(opp_table); unlock: - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); return ret; } @@ -1040,13 +1366,13 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_set_prop_name); /** * dev_pm_opp_put_prop_name() - Releases resources blocked for prop-name - * @dev: Device for which the regulator has to be set. + * @dev: Device for which the prop-name has to be put. * * This is required only for the V2 bindings, and is called for a matching - * dev_pm_opp_set_prop_name(). Until this is called, the device_opp structure + * dev_pm_opp_set_prop_name(). Until this is called, the opp_table structure * will not be freed. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function internally uses RCU updater strategy with mutex locks * to keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where @@ -1054,45 +1380,154 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_set_prop_name); */ void dev_pm_opp_put_prop_name(struct device *dev) { - struct device_opp *dev_opp; + struct opp_table *opp_table; - /* Hold our list modification lock here */ - mutex_lock(&dev_opp_list_lock); + /* Hold our table modification lock here */ + mutex_lock(&opp_table_lock); - /* Check for existing list for 'dev' first */ - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp)) { - dev_err(dev, "Failed to find dev_opp: %ld\n", PTR_ERR(dev_opp)); + /* Check for existing table for 'dev' first */ + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) { + dev_err(dev, "Failed to find opp_table: %ld\n", + PTR_ERR(opp_table)); goto unlock; } - /* Make sure there are no concurrent readers while updating dev_opp */ - WARN_ON(!list_empty(&dev_opp->opp_list)); + /* Make sure there are no concurrent readers while updating opp_table */ + WARN_ON(!list_empty(&opp_table->opp_list)); - if (!dev_opp->prop_name) { + if (!opp_table->prop_name) { dev_err(dev, "%s: Doesn't have a prop-name\n", __func__); goto unlock; } - kfree(dev_opp->prop_name); - dev_opp->prop_name = NULL; + kfree(opp_table->prop_name); + opp_table->prop_name = NULL; - /* Try freeing device_opp if this was the last blocking resource */ - _remove_device_opp(dev_opp); + /* Try freeing opp_table if this was the last blocking resource */ + _remove_opp_table(opp_table); unlock: - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); } EXPORT_SYMBOL_GPL(dev_pm_opp_put_prop_name); -static bool _opp_is_supported(struct device *dev, struct device_opp *dev_opp, +/** + * dev_pm_opp_set_regulator() - Set regulator name for the device + * @dev: Device for which regulator name is being set. + * @name: Name of the regulator. + * + * In order to support OPP switching, OPP layer needs to know the name of the + * device's regulator, as the core would be required to switch voltages as well. + * + * This must be called before any OPPs are initialized for the device. + * + * Locking: The internal opp_table and opp structures are RCU protected. + * Hence this function internally uses RCU updater strategy with mutex locks + * to keep the integrity of the internal data structures. Callers should ensure + * that this function is *NOT* called under RCU protection or in contexts where + * mutex cannot be locked. + */ +int dev_pm_opp_set_regulator(struct device *dev, const char *name) +{ + struct opp_table *opp_table; + struct regulator *reg; + int ret; + + mutex_lock(&opp_table_lock); + + opp_table = _add_opp_table(dev); + if (!opp_table) { + ret = -ENOMEM; + goto unlock; + } + + /* This should be called before OPPs are initialized */ + if (WARN_ON(!list_empty(&opp_table->opp_list))) { + ret = -EBUSY; + goto err; + } + + /* Already have a regulator set */ + if (WARN_ON(!IS_ERR(opp_table->regulator))) { + ret = -EBUSY; + goto err; + } + /* Allocate the regulator */ + reg = regulator_get_optional(dev, name); + if (IS_ERR(reg)) { + ret = PTR_ERR(reg); + if (ret != -EPROBE_DEFER) + dev_err(dev, "%s: no regulator (%s) found: %d\n", + __func__, name, ret); + goto err; + } + + opp_table->regulator = reg; + + mutex_unlock(&opp_table_lock); + return 0; + +err: + _remove_opp_table(opp_table); +unlock: + mutex_unlock(&opp_table_lock); + + return ret; +} +EXPORT_SYMBOL_GPL(dev_pm_opp_set_regulator); + +/** + * dev_pm_opp_put_regulator() - Releases resources blocked for regulator + * @dev: Device for which regulator was set. + * + * Locking: The internal opp_table and opp structures are RCU protected. + * Hence this function internally uses RCU updater strategy with mutex locks + * to keep the integrity of the internal data structures. Callers should ensure + * that this function is *NOT* called under RCU protection or in contexts where + * mutex cannot be locked. + */ +void dev_pm_opp_put_regulator(struct device *dev) +{ + struct opp_table *opp_table; + + mutex_lock(&opp_table_lock); + + /* Check for existing table for 'dev' first */ + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) { + dev_err(dev, "Failed to find opp_table: %ld\n", + PTR_ERR(opp_table)); + goto unlock; + } + + if (IS_ERR(opp_table->regulator)) { + dev_err(dev, "%s: Doesn't have regulator set\n", __func__); + goto unlock; + } + + /* Make sure there are no concurrent readers while updating opp_table */ + WARN_ON(!list_empty(&opp_table->opp_list)); + + regulator_put(opp_table->regulator); + opp_table->regulator = ERR_PTR(-ENXIO); + + /* Try freeing opp_table if this was the last blocking resource */ + _remove_opp_table(opp_table); + +unlock: + mutex_unlock(&opp_table_lock); +} +EXPORT_SYMBOL_GPL(dev_pm_opp_put_regulator); + +static bool _opp_is_supported(struct device *dev, struct opp_table *opp_table, struct device_node *np) { - unsigned int count = dev_opp->supported_hw_count; + unsigned int count = opp_table->supported_hw_count; u32 version; int ret; - if (!dev_opp->supported_hw) + if (!opp_table->supported_hw) return true; while (count--) { @@ -1105,7 +1540,7 @@ static bool _opp_is_supported(struct device *dev, struct device_opp *dev_opp, } /* Both of these are bitwise masks of the versions */ - if (!(version & dev_opp->supported_hw[count])) + if (!(version & opp_table->supported_hw[count])) return false; } @@ -1117,11 +1552,11 @@ static bool _opp_is_supported(struct device *dev, struct device_opp *dev_opp, * @dev: device for which we do this operation * @np: device node * - * This function adds an opp definition to the opp list and returns status. The + * This function adds an opp definition to the opp table and returns status. The * opp can be controlled using dev_pm_opp_enable/disable functions and may be * removed by dev_pm_opp_remove. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function internally uses RCU updater strategy with mutex locks * to keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where @@ -1137,16 +1572,16 @@ static bool _opp_is_supported(struct device *dev, struct device_opp *dev_opp, */ static int _opp_add_static_v2(struct device *dev, struct device_node *np) { - struct device_opp *dev_opp; + struct opp_table *opp_table; struct dev_pm_opp *new_opp; u64 rate; u32 val; int ret; - /* Hold our list modification lock here */ - mutex_lock(&dev_opp_list_lock); + /* Hold our table modification lock here */ + mutex_lock(&opp_table_lock); - new_opp = _allocate_opp(dev, &dev_opp); + new_opp = _allocate_opp(dev, &opp_table); if (!new_opp) { ret = -ENOMEM; goto unlock; @@ -1159,7 +1594,7 @@ static int _opp_add_static_v2(struct device *dev, struct device_node *np) } /* Check if the OPP supports hardware's hierarchy of versions or not */ - if (!_opp_is_supported(dev, dev_opp, np)) { + if (!_opp_is_supported(dev, opp_table, np)) { dev_dbg(dev, "OPP not supported by hardware: %llu\n", rate); goto free_opp; } @@ -1179,30 +1614,30 @@ static int _opp_add_static_v2(struct device *dev, struct device_node *np) if (!of_property_read_u32(np, "clock-latency-ns", &val)) new_opp->clock_latency_ns = val; - ret = opp_parse_supplies(new_opp, dev, dev_opp); + ret = opp_parse_supplies(new_opp, dev, opp_table); if (ret) goto free_opp; - ret = _opp_add(dev, new_opp, dev_opp); + ret = _opp_add(dev, new_opp, opp_table); if (ret) goto free_opp; /* OPP to select on device suspend */ if (of_property_read_bool(np, "opp-suspend")) { - if (dev_opp->suspend_opp) { + if (opp_table->suspend_opp) { dev_warn(dev, "%s: Multiple suspend OPPs found (%lu %lu)\n", - __func__, dev_opp->suspend_opp->rate, + __func__, opp_table->suspend_opp->rate, new_opp->rate); } else { new_opp->suspend = true; - dev_opp->suspend_opp = new_opp; + opp_table->suspend_opp = new_opp; } } - if (new_opp->clock_latency_ns > dev_opp->clock_latency_ns_max) - dev_opp->clock_latency_ns_max = new_opp->clock_latency_ns; + if (new_opp->clock_latency_ns > opp_table->clock_latency_ns_max) + opp_table->clock_latency_ns_max = new_opp->clock_latency_ns; - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); pr_debug("%s: turbo:%d rate:%lu uv:%lu uvmin:%lu uvmax:%lu latency:%lu\n", __func__, new_opp->turbo, new_opp->rate, new_opp->u_volt, @@ -1213,13 +1648,13 @@ static int _opp_add_static_v2(struct device *dev, struct device_node *np) * Notify the changes in the availability of the operable * frequency/voltage list. */ - srcu_notifier_call_chain(&dev_opp->srcu_head, OPP_EVENT_ADD, new_opp); + srcu_notifier_call_chain(&opp_table->srcu_head, OPP_EVENT_ADD, new_opp); return 0; free_opp: - _opp_remove(dev_opp, new_opp, false); + _opp_remove(opp_table, new_opp, false); unlock: - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); return ret; } @@ -1229,11 +1664,11 @@ unlock: * @freq: Frequency in Hz for this OPP * @u_volt: Voltage in uVolts for this OPP * - * This function adds an opp definition to the opp list and returns status. + * This function adds an opp definition to the opp table and returns status. * The opp is made available by default and it can be controlled using * dev_pm_opp_enable/disable functions. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function internally uses RCU updater strategy with mutex locks * to keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where @@ -1265,7 +1700,7 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_add); * copy operation, returns 0 if no modification was done OR modification was * successful. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function internally uses RCU updater strategy with mutex locks to * keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where @@ -1274,7 +1709,7 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_add); static int _opp_set_availability(struct device *dev, unsigned long freq, bool availability_req) { - struct device_opp *dev_opp; + struct opp_table *opp_table; struct dev_pm_opp *new_opp, *tmp_opp, *opp = ERR_PTR(-ENODEV); int r = 0; @@ -1283,18 +1718,18 @@ static int _opp_set_availability(struct device *dev, unsigned long freq, if (!new_opp) return -ENOMEM; - mutex_lock(&dev_opp_list_lock); + mutex_lock(&opp_table_lock); - /* Find the device_opp */ - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp)) { - r = PTR_ERR(dev_opp); + /* Find the opp_table */ + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) { + r = PTR_ERR(opp_table); dev_warn(dev, "%s: Device OPP not found (%d)\n", __func__, r); goto unlock; } /* Do we have the frequency? */ - list_for_each_entry(tmp_opp, &dev_opp->opp_list, node) { + list_for_each_entry(tmp_opp, &opp_table->opp_list, node) { if (tmp_opp->rate == freq) { opp = tmp_opp; break; @@ -1315,21 +1750,21 @@ static int _opp_set_availability(struct device *dev, unsigned long freq, new_opp->available = availability_req; list_replace_rcu(&opp->node, &new_opp->node); - mutex_unlock(&dev_opp_list_lock); - call_srcu(&dev_opp->srcu_head.srcu, &opp->rcu_head, _kfree_opp_rcu); + mutex_unlock(&opp_table_lock); + call_srcu(&opp_table->srcu_head.srcu, &opp->rcu_head, _kfree_opp_rcu); /* Notify the change of the OPP availability */ if (availability_req) - srcu_notifier_call_chain(&dev_opp->srcu_head, OPP_EVENT_ENABLE, - new_opp); + srcu_notifier_call_chain(&opp_table->srcu_head, + OPP_EVENT_ENABLE, new_opp); else - srcu_notifier_call_chain(&dev_opp->srcu_head, OPP_EVENT_DISABLE, - new_opp); + srcu_notifier_call_chain(&opp_table->srcu_head, + OPP_EVENT_DISABLE, new_opp); return 0; unlock: - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); kfree(new_opp); return r; } @@ -1343,7 +1778,7 @@ unlock: * corresponding error value. It is meant to be used for users an OPP available * after being temporarily made unavailable with dev_pm_opp_disable. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function indirectly uses RCU and mutex locks to keep the * integrity of the internal data structures. Callers should ensure that * this function is *NOT* called under RCU protection or in contexts where @@ -1369,7 +1804,7 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_enable); * control by users to make this OPP not available until the circumstances are * right to make it available again (with a call to dev_pm_opp_enable). * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function indirectly uses RCU and mutex locks to keep the * integrity of the internal data structures. Callers should ensure that * this function is *NOT* called under RCU protection or in contexts where @@ -1387,26 +1822,26 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_disable); /** * dev_pm_opp_get_notifier() - find notifier_head of the device with opp - * @dev: device pointer used to lookup device OPPs. + * @dev: device pointer used to lookup OPP table. * * Return: pointer to notifier head if found, otherwise -ENODEV or * -EINVAL based on type of error casted as pointer. value must be checked * with IS_ERR to determine valid pointer or error result. * - * Locking: This function must be called under rcu_read_lock(). dev_opp is a RCU - * protected pointer. The reason for the same is that the opp pointer which is - * returned will remain valid for use with opp_get_{voltage, freq} only while + * Locking: This function must be called under rcu_read_lock(). opp_table is a + * RCU protected pointer. The reason for the same is that the opp pointer which + * is returned will remain valid for use with opp_get_{voltage, freq} only while * under the locked area. The pointer returned must be used prior to unlocking * with rcu_read_unlock() to maintain the integrity of the pointer. */ struct srcu_notifier_head *dev_pm_opp_get_notifier(struct device *dev) { - struct device_opp *dev_opp = _find_device_opp(dev); + struct opp_table *opp_table = _find_opp_table(dev); - if (IS_ERR(dev_opp)) - return ERR_CAST(dev_opp); /* matching type */ + if (IS_ERR(opp_table)) + return ERR_CAST(opp_table); /* matching type */ - return &dev_opp->srcu_head; + return &opp_table->srcu_head; } EXPORT_SYMBOL_GPL(dev_pm_opp_get_notifier); @@ -1414,11 +1849,11 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_get_notifier); /** * dev_pm_opp_of_remove_table() - Free OPP table entries created from static DT * entries - * @dev: device pointer used to lookup device OPPs. + * @dev: device pointer used to lookup OPP table. * * Free OPPs created using static entries present in DT. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function indirectly uses RCU updater strategy with mutex locks * to keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where @@ -1426,38 +1861,38 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_get_notifier); */ void dev_pm_opp_of_remove_table(struct device *dev) { - struct device_opp *dev_opp; + struct opp_table *opp_table; struct dev_pm_opp *opp, *tmp; - /* Hold our list modification lock here */ - mutex_lock(&dev_opp_list_lock); + /* Hold our table modification lock here */ + mutex_lock(&opp_table_lock); - /* Check for existing list for 'dev' */ - dev_opp = _find_device_opp(dev); - if (IS_ERR(dev_opp)) { - int error = PTR_ERR(dev_opp); + /* Check for existing table for 'dev' */ + opp_table = _find_opp_table(dev); + if (IS_ERR(opp_table)) { + int error = PTR_ERR(opp_table); if (error != -ENODEV) - WARN(1, "%s: dev_opp: %d\n", + WARN(1, "%s: opp_table: %d\n", IS_ERR_OR_NULL(dev) ? "Invalid device" : dev_name(dev), error); goto unlock; } - /* Find if dev_opp manages a single device */ - if (list_is_singular(&dev_opp->dev_list)) { + /* Find if opp_table manages a single device */ + if (list_is_singular(&opp_table->dev_list)) { /* Free static OPPs */ - list_for_each_entry_safe(opp, tmp, &dev_opp->opp_list, node) { + list_for_each_entry_safe(opp, tmp, &opp_table->opp_list, node) { if (!opp->dynamic) - _opp_remove(dev_opp, opp, true); + _opp_remove(opp_table, opp, true); } } else { - _remove_list_dev(_find_list_dev(dev, dev_opp), dev_opp); + _remove_opp_dev(_find_opp_dev(dev, opp_table), opp_table); } unlock: - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); } EXPORT_SYMBOL_GPL(dev_pm_opp_of_remove_table); @@ -1478,22 +1913,22 @@ struct device_node *_of_get_opp_desc_node(struct device *dev) static int _of_add_opp_table_v2(struct device *dev, struct device_node *opp_np) { struct device_node *np; - struct device_opp *dev_opp; + struct opp_table *opp_table; int ret = 0, count = 0; - mutex_lock(&dev_opp_list_lock); + mutex_lock(&opp_table_lock); - dev_opp = _managed_opp(opp_np); - if (dev_opp) { + opp_table = _managed_opp(opp_np); + if (opp_table) { /* OPPs are already managed */ - if (!_add_list_dev(dev, dev_opp)) + if (!_add_opp_dev(dev, opp_table)) ret = -ENOMEM; - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); return ret; } - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); - /* We have opp-list node now, iterate over it and add OPPs */ + /* We have opp-table node now, iterate over it and add OPPs */ for_each_available_child_of_node(opp_np, np) { count++; @@ -1509,19 +1944,19 @@ static int _of_add_opp_table_v2(struct device *dev, struct device_node *opp_np) if (WARN_ON(!count)) return -ENOENT; - mutex_lock(&dev_opp_list_lock); + mutex_lock(&opp_table_lock); - dev_opp = _find_device_opp(dev); - if (WARN_ON(IS_ERR(dev_opp))) { - ret = PTR_ERR(dev_opp); - mutex_unlock(&dev_opp_list_lock); + opp_table = _find_opp_table(dev); + if (WARN_ON(IS_ERR(opp_table))) { + ret = PTR_ERR(opp_table); + mutex_unlock(&opp_table_lock); goto free_table; } - dev_opp->np = opp_np; - dev_opp->shared_opp = of_property_read_bool(opp_np, "opp-shared"); + opp_table->np = opp_np; + opp_table->shared_opp = of_property_read_bool(opp_np, "opp-shared"); - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); return 0; @@ -1550,7 +1985,7 @@ static int _of_add_opp_table_v1(struct device *dev) */ nr = prop->length / sizeof(u32); if (nr % 2) { - dev_err(dev, "%s: Invalid OPP list\n", __func__); + dev_err(dev, "%s: Invalid OPP table\n", __func__); return -EINVAL; } @@ -1570,11 +2005,11 @@ static int _of_add_opp_table_v1(struct device *dev) /** * dev_pm_opp_of_add_table() - Initialize opp table from device tree - * @dev: device pointer used to lookup device OPPs. + * @dev: device pointer used to lookup OPP table. * * Register the initial OPP table with the OPP library for given device. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Hence this function indirectly uses RCU updater strategy with mutex locks * to keep the integrity of the internal data structures. Callers should ensure * that this function is *NOT* called under RCU protection or in contexts where diff --git a/drivers/base/power/opp/cpu.c b/drivers/base/power/opp/cpu.c index 9f0c155..ba2bdbd 100644 --- a/drivers/base/power/opp/cpu.c +++ b/drivers/base/power/opp/cpu.c @@ -31,7 +31,7 @@ * @table: Cpufreq table returned back to caller * * Generate a cpufreq table for a provided device- this assumes that the - * opp list is already initialized and ready for usage. + * opp table is already initialized and ready for usage. * * This function allocates required memory for the cpufreq table. It is * expected that the caller does the required maintenance such as freeing @@ -44,7 +44,7 @@ * WARNING: It is important for the callers to ensure refreshing their copy of * the table if any of the mentioned functions have been invoked in the interim. * - * Locking: The internal device_opp and opp structures are RCU protected. + * Locking: The internal opp_table and opp structures are RCU protected. * Since we just use the regular accessor functions to access the internal data * structures, we use RCU read lock inside this function. As a result, users of * this function DONOT need to use explicit locks for invoking. @@ -122,15 +122,15 @@ EXPORT_SYMBOL_GPL(dev_pm_opp_free_cpufreq_table); /* Required only for V1 bindings, as v2 can manage it from DT itself */ int dev_pm_opp_set_sharing_cpus(struct device *cpu_dev, cpumask_var_t cpumask) { - struct device_list_opp *list_dev; - struct device_opp *dev_opp; + struct opp_device *opp_dev; + struct opp_table *opp_table; struct device *dev; int cpu, ret = 0; - mutex_lock(&dev_opp_list_lock); + mutex_lock(&opp_table_lock); - dev_opp = _find_device_opp(cpu_dev); - if (IS_ERR(dev_opp)) { + opp_table = _find_opp_table(cpu_dev); + if (IS_ERR(opp_table)) { ret = -EINVAL; goto unlock; } @@ -146,15 +146,15 @@ int dev_pm_opp_set_sharing_cpus(struct device *cpu_dev, cpumask_var_t cpumask) continue; } - list_dev = _add_list_dev(dev, dev_opp); - if (!list_dev) { - dev_err(dev, "%s: failed to add list-dev for cpu%d device\n", + opp_dev = _add_opp_dev(dev, opp_table); + if (!opp_dev) { + dev_err(dev, "%s: failed to add opp-dev for cpu%d device\n", __func__, cpu); continue; } } unlock: - mutex_unlock(&dev_opp_list_lock); + mutex_unlock(&opp_table_lock); return ret; } diff --git a/drivers/base/power/opp/debugfs.c b/drivers/base/power/opp/debugfs.c index ddfe477..ef1ae6b 100644 --- a/drivers/base/power/opp/debugfs.c +++ b/drivers/base/power/opp/debugfs.c @@ -34,9 +34,9 @@ void opp_debug_remove_one(struct dev_pm_opp *opp) debugfs_remove_recursive(opp->dentry); } -int opp_debug_create_one(struct dev_pm_opp *opp, struct device_opp *dev_opp) +int opp_debug_create_one(struct dev_pm_opp *opp, struct opp_table *opp_table) { - struct dentry *pdentry = dev_opp->dentry; + struct dentry *pdentry = opp_table->dentry; struct dentry *d; char name[25]; /* 20 chars for 64 bit value + 5 (opp:\0) */ @@ -83,52 +83,52 @@ int opp_debug_create_one(struct dev_pm_opp *opp, struct device_opp *dev_opp) return 0; } -static int device_opp_debug_create_dir(struct device_list_opp *list_dev, - struct device_opp *dev_opp) +static int opp_list_debug_create_dir(struct opp_device *opp_dev, + struct opp_table *opp_table) { - const struct device *dev = list_dev->dev; + const struct device *dev = opp_dev->dev; struct dentry *d; - opp_set_dev_name(dev, dev_opp->dentry_name); + opp_set_dev_name(dev, opp_table->dentry_name); /* Create device specific directory */ - d = debugfs_create_dir(dev_opp->dentry_name, rootdir); + d = debugfs_create_dir(opp_table->dentry_name, rootdir); if (!d) { dev_err(dev, "%s: Failed to create debugfs dir\n", __func__); return -ENOMEM; } - list_dev->dentry = d; - dev_opp->dentry = d; + opp_dev->dentry = d; + opp_table->dentry = d; return 0; } -static int device_opp_debug_create_link(struct device_list_opp *list_dev, - struct device_opp *dev_opp) +static int opp_list_debug_create_link(struct opp_device *opp_dev, + struct opp_table *opp_table) { - const struct device *dev = list_dev->dev; + const struct device *dev = opp_dev->dev; char name[NAME_MAX]; struct dentry *d; - opp_set_dev_name(list_dev->dev, name); + opp_set_dev_name(opp_dev->dev, name); /* Create device specific directory link */ - d = debugfs_create_symlink(name, rootdir, dev_opp->dentry_name); + d = debugfs_create_symlink(name, rootdir, opp_table->dentry_name); if (!d) { dev_err(dev, "%s: Failed to create link\n", __func__); return -ENOMEM; } - list_dev->dentry = d; + opp_dev->dentry = d; return 0; } /** * opp_debug_register - add a device opp node to the debugfs 'opp' directory - * @list_dev: list-dev pointer for device - * @dev_opp: the device-opp being added + * @opp_dev: opp-dev pointer for device + * @opp_table: the device-opp being added * * Dynamically adds device specific directory in debugfs 'opp' directory. If the * device-opp is shared with other devices, then links will be created for all @@ -136,73 +136,72 @@ static int device_opp_debug_create_link(struct device_list_opp *list_dev, * * Return: 0 on success, otherwise negative error. */ -int opp_debug_register(struct device_list_opp *list_dev, - struct device_opp *dev_opp) +int opp_debug_register(struct opp_device *opp_dev, struct opp_table *opp_table) { if (!rootdir) { pr_debug("%s: Uninitialized rootdir\n", __func__); return -EINVAL; } - if (dev_opp->dentry) - return device_opp_debug_create_link(list_dev, dev_opp); + if (opp_table->dentry) + return opp_list_debug_create_link(opp_dev, opp_table); - return device_opp_debug_create_dir(list_dev, dev_opp); + return opp_list_debug_create_dir(opp_dev, opp_table); } -static void opp_migrate_dentry(struct device_list_opp *list_dev, - struct device_opp *dev_opp) +static void opp_migrate_dentry(struct opp_device *opp_dev, + struct opp_table *opp_table) { - struct device_list_opp *new_dev; + struct opp_device *new_dev; const struct device *dev; struct dentry *dentry; - /* Look for next list-dev */ - list_for_each_entry(new_dev, &dev_opp->dev_list, node) - if (new_dev != list_dev) + /* Look for next opp-dev */ + list_for_each_entry(new_dev, &opp_table->dev_list, node) + if (new_dev != opp_dev) break; /* new_dev is guaranteed to be valid here */ dev = new_dev->dev; debugfs_remove_recursive(new_dev->dentry); - opp_set_dev_name(dev, dev_opp->dentry_name); + opp_set_dev_name(dev, opp_table->dentry_name); - dentry = debugfs_rename(rootdir, list_dev->dentry, rootdir, - dev_opp->dentry_name); + dentry = debugfs_rename(rootdir, opp_dev->dentry, rootdir, + opp_table->dentry_name); if (!dentry) { dev_err(dev, "%s: Failed to rename link from: %s to %s\n", - __func__, dev_name(list_dev->dev), dev_name(dev)); + __func__, dev_name(opp_dev->dev), dev_name(dev)); return; } new_dev->dentry = dentry; - dev_opp->dentry = dentry; + opp_table->dentry = dentry; } /** * opp_debug_unregister - remove a device opp node from debugfs opp directory - * @list_dev: list-dev pointer for device - * @dev_opp: the device-opp being removed + * @opp_dev: opp-dev pointer for device + * @opp_table: the device-opp being removed * * Dynamically removes device specific directory from debugfs 'opp' directory. */ -void opp_debug_unregister(struct device_list_opp *list_dev, - struct device_opp *dev_opp) +void opp_debug_unregister(struct opp_device *opp_dev, + struct opp_table *opp_table) { - if (list_dev->dentry == dev_opp->dentry) { + if (opp_dev->dentry == opp_table->dentry) { /* Move the real dentry object under another device */ - if (!list_is_singular(&dev_opp->dev_list)) { - opp_migrate_dentry(list_dev, dev_opp); + if (!list_is_singular(&opp_table->dev_list)) { + opp_migrate_dentry(opp_dev, opp_table); goto out; } - dev_opp->dentry = NULL; + opp_table->dentry = NULL; } - debugfs_remove_recursive(list_dev->dentry); + debugfs_remove_recursive(opp_dev->dentry); out: - list_dev->dentry = NULL; + opp_dev->dentry = NULL; } static int __init opp_debug_init(void) diff --git a/drivers/base/power/opp/opp.h b/drivers/base/power/opp/opp.h index 690638e..f67f806 100644 --- a/drivers/base/power/opp/opp.h +++ b/drivers/base/power/opp/opp.h @@ -22,13 +22,16 @@ #include <linux/rculist.h> #include <linux/rcupdate.h> +struct clk; +struct regulator; + /* Lock to allow exclusive modification to the device and opp lists */ -extern struct mutex dev_opp_list_lock; +extern struct mutex opp_table_lock; /* * Internal data structure organization with the OPP layer library is as * follows: - * dev_opp_list (root) + * opp_tables (root) * |- device 1 (represents voltage domain 1) * | |- opp 1 (availability, freq, voltage) * | |- opp 2 .. @@ -37,18 +40,18 @@ extern struct mutex dev_opp_list_lock; * |- device 2 (represents the next voltage domain) * ... * `- device m (represents mth voltage domain) - * device 1, 2.. are represented by dev_opp structure while each opp + * device 1, 2.. are represented by opp_table structure while each opp * is represented by the opp structure. */ /** * struct dev_pm_opp - Generic OPP description structure - * @node: opp list node. The nodes are maintained throughout the lifetime + * @node: opp table node. The nodes are maintained throughout the lifetime * of boot. It is expected only an optimal set of OPPs are * added to the library by the SoC framework. - * RCU usage: opp list is traversed with RCU locks. node + * RCU usage: opp table is traversed with RCU locks. node * modification is possible realtime, hence the modifications - * are protected by the dev_opp_list_lock for integrity. + * are protected by the opp_table_lock for integrity. * IMPORTANT: the opp nodes should be maintained in increasing * order. * @available: true/false - marks if this OPP as available or not @@ -62,7 +65,7 @@ extern struct mutex dev_opp_list_lock; * @u_amp: Maximum current drawn by the device in microamperes * @clock_latency_ns: Latency (in nanoseconds) of switching to this OPP's * frequency from any other OPP's frequency. - * @dev_opp: points back to the device_opp struct this opp belongs to + * @opp_table: points back to the opp_table struct this opp belongs to * @rcu_head: RCU callback head used for deferred freeing * @np: OPP's device node. * @dentry: debugfs dentry pointer (per opp) @@ -84,7 +87,7 @@ struct dev_pm_opp { unsigned long u_amp; unsigned long clock_latency_ns; - struct device_opp *dev_opp; + struct opp_table *opp_table; struct rcu_head rcu_head; struct device_node *np; @@ -95,16 +98,16 @@ struct dev_pm_opp { }; /** - * struct device_list_opp - devices managed by 'struct device_opp' + * struct opp_device - devices managed by 'struct opp_table' * @node: list node * @dev: device to which the struct object belongs * @rcu_head: RCU callback head used for deferred freeing * @dentry: debugfs dentry pointer (per device) * - * This is an internal data structure maintaining the list of devices that are - * managed by 'struct device_opp'. + * This is an internal data structure maintaining the devices that are managed + * by 'struct opp_table'. */ -struct device_list_opp { +struct opp_device { struct list_head node; const struct device *dev; struct rcu_head rcu_head; @@ -115,16 +118,16 @@ struct device_list_opp { }; /** - * struct device_opp - Device opp structure - * @node: list node - contains the devices with OPPs that + * struct opp_table - Device opp structure + * @node: table node - contains the devices with OPPs that * have been registered. Nodes once added are not modified in this - * list. - * RCU usage: nodes are not modified in the list of device_opp, - * however addition is possible and is secured by dev_opp_list_lock + * table. + * RCU usage: nodes are not modified in the table of opp_table, + * however addition is possible and is secured by opp_table_lock * @srcu_head: notifier head to notify the OPP availability changes. * @rcu_head: RCU callback head used for deferred freeing * @dev_list: list of devices that share these OPPs - * @opp_list: list of opps + * @opp_list: table of opps * @np: struct device_node pointer for opp's DT node. * @clock_latency_ns_max: Max clock latency in nanoseconds. * @shared_opp: OPP is shared between multiple devices. @@ -132,9 +135,13 @@ struct device_list_opp { * @supported_hw: Array of version number to support. * @supported_hw_count: Number of elements in supported_hw array. * @prop_name: A name to postfix to many DT properties, while parsing them. + * @clk: Device's clock handle + * @regulator: Supply regulator * @dentry: debugfs dentry pointer of the real device directory (not links). * @dentry_name: Name of the real dentry. * + * @voltage_tolerance_v1: In percentage, for v1 bindings only. + * * This is an internal data structure maintaining the link to opps attached to * a device. This structure is not meant to be shared to users as it is * meant for book keeping and private to OPP library. @@ -143,7 +150,7 @@ struct device_list_opp { * need to wait for the grace period of both of them before freeing any * resources. And so we have used kfree_rcu() from within call_srcu() handlers. */ -struct device_opp { +struct opp_table { struct list_head node; struct srcu_notifier_head srcu_head; @@ -153,12 +160,18 @@ struct device_opp { struct device_node *np; unsigned long clock_latency_ns_max; + + /* For backward compatibility with v1 bindings */ + unsigned int voltage_tolerance_v1; + bool shared_opp; struct dev_pm_opp *suspend_opp; unsigned int *supported_hw; unsigned int supported_hw_count; const char *prop_name; + struct clk *clk; + struct regulator *regulator; #ifdef CONFIG_DEBUG_FS struct dentry *dentry; @@ -167,30 +180,27 @@ struct device_opp { }; /* Routines internal to opp core */ -struct device_opp *_find_device_opp(struct device *dev); -struct device_list_opp *_add_list_dev(const struct device *dev, - struct device_opp *dev_opp); +struct opp_table *_find_opp_table(struct device *dev); +struct opp_device *_add_opp_dev(const struct device *dev, struct opp_table *opp_table); struct device_node *_of_get_opp_desc_node(struct device *dev); #ifdef CONFIG_DEBUG_FS void opp_debug_remove_one(struct dev_pm_opp *opp); -int opp_debug_create_one(struct dev_pm_opp *opp, struct device_opp *dev_opp); -int opp_debug_register(struct device_list_opp *list_dev, - struct device_opp *dev_opp); -void opp_debug_unregister(struct device_list_opp *list_dev, - struct device_opp *dev_opp); +int opp_debug_create_one(struct dev_pm_opp *opp, struct opp_table *opp_table); +int opp_debug_register(struct opp_device *opp_dev, struct opp_table *opp_table); +void opp_debug_unregister(struct opp_device *opp_dev, struct opp_table *opp_table); #else static inline void opp_debug_remove_one(struct dev_pm_opp *opp) {} static inline int opp_debug_create_one(struct dev_pm_opp *opp, - struct device_opp *dev_opp) + struct opp_table *opp_table) { return 0; } -static inline int opp_debug_register(struct device_list_opp *list_dev, - struct device_opp *dev_opp) +static inline int opp_debug_register(struct opp_device *opp_dev, + struct opp_table *opp_table) { return 0; } -static inline void opp_debug_unregister(struct device_list_opp *list_dev, - struct device_opp *dev_opp) +static inline void opp_debug_unregister(struct opp_device *opp_dev, + struct opp_table *opp_table) { } #endif /* DEBUG_FS */ diff --git a/drivers/base/power/trace.c b/drivers/base/power/trace.c index a311cfa..a697579 100644 --- a/drivers/base/power/trace.c +++ b/drivers/base/power/trace.c @@ -166,14 +166,14 @@ void generate_pm_trace(const void *tracedata, unsigned int user) } EXPORT_SYMBOL(generate_pm_trace); -extern char __tracedata_start, __tracedata_end; +extern char __tracedata_start[], __tracedata_end[]; static int show_file_hash(unsigned int value) { int match; char *tracedata; match = 0; - for (tracedata = &__tracedata_start ; tracedata < &__tracedata_end ; + for (tracedata = __tracedata_start ; tracedata < __tracedata_end ; tracedata += 2 + sizeof(unsigned long)) { unsigned short lineno = *(unsigned short *)tracedata; const char *file = *(const char **)(tracedata + 2); diff --git a/drivers/base/property.c b/drivers/base/property.c index a163f2c5..76628a7 100644 --- a/drivers/base/property.c +++ b/drivers/base/property.c @@ -218,7 +218,8 @@ bool fwnode_property_present(struct fwnode_handle *fwnode, const char *propname) bool ret; ret = __fwnode_property_present(fwnode, propname); - if (ret == false && fwnode && !IS_ERR_OR_NULL(fwnode->secondary)) + if (ret == false && !IS_ERR_OR_NULL(fwnode) && + !IS_ERR_OR_NULL(fwnode->secondary)) ret = __fwnode_property_present(fwnode->secondary, propname); return ret; } @@ -423,7 +424,8 @@ EXPORT_SYMBOL_GPL(device_property_match_string); int _ret_; \ _ret_ = FWNODE_PROP_READ(_fwnode_, _propname_, _type_, _proptype_, \ _val_, _nval_); \ - if (_ret_ == -EINVAL && _fwnode_ && !IS_ERR_OR_NULL(_fwnode_->secondary)) \ + if (_ret_ == -EINVAL && !IS_ERR_OR_NULL(_fwnode_) && \ + !IS_ERR_OR_NULL(_fwnode_->secondary)) \ _ret_ = FWNODE_PROP_READ(_fwnode_->secondary, _propname_, _type_, \ _proptype_, _val_, _nval_); \ _ret_; \ @@ -593,7 +595,8 @@ int fwnode_property_read_string_array(struct fwnode_handle *fwnode, int ret; ret = __fwnode_property_read_string_array(fwnode, propname, val, nval); - if (ret == -EINVAL && fwnode && !IS_ERR_OR_NULL(fwnode->secondary)) + if (ret == -EINVAL && !IS_ERR_OR_NULL(fwnode) && + !IS_ERR_OR_NULL(fwnode->secondary)) ret = __fwnode_property_read_string_array(fwnode->secondary, propname, val, nval); return ret; @@ -621,7 +624,8 @@ int fwnode_property_read_string(struct fwnode_handle *fwnode, int ret; ret = __fwnode_property_read_string(fwnode, propname, val); - if (ret == -EINVAL && fwnode && !IS_ERR_OR_NULL(fwnode->secondary)) + if (ret == -EINVAL && !IS_ERR_OR_NULL(fwnode) && + !IS_ERR_OR_NULL(fwnode->secondary)) ret = __fwnode_property_read_string(fwnode->secondary, propname, val); return ret; @@ -820,11 +824,16 @@ void device_remove_property_set(struct device *dev) * the pset. If there is no real firmware node (ACPI/DT) primary * will hold the pset. */ - if (!is_pset_node(fwnode)) - fwnode = fwnode->secondary; - if (!IS_ERR(fwnode) && is_pset_node(fwnode)) + if (is_pset_node(fwnode)) { + set_primary_fwnode(dev, NULL); pset_free_set(to_pset_node(fwnode)); - set_secondary_fwnode(dev, NULL); + } else { + fwnode = fwnode->secondary; + if (!IS_ERR(fwnode) && is_pset_node(fwnode)) { + set_secondary_fwnode(dev, NULL); + pset_free_set(to_pset_node(fwnode)); + } + } } EXPORT_SYMBOL_GPL(device_remove_property_set); diff --git a/drivers/cpufreq/Kconfig b/drivers/cpufreq/Kconfig index f935110..a7f4585 100644 --- a/drivers/cpufreq/Kconfig +++ b/drivers/cpufreq/Kconfig @@ -19,6 +19,7 @@ config CPU_FREQ if CPU_FREQ config CPU_FREQ_GOV_COMMON + select IRQ_WORK bool config CPU_FREQ_BOOST_SW diff --git a/drivers/cpufreq/acpi-cpufreq.c b/drivers/cpufreq/acpi-cpufreq.c index 51eef87..59a7b38 100644 --- a/drivers/cpufreq/acpi-cpufreq.c +++ b/drivers/cpufreq/acpi-cpufreq.c @@ -70,6 +70,8 @@ struct acpi_cpufreq_data { unsigned int cpu_feature; unsigned int acpi_perf_cpu; cpumask_var_t freqdomain_cpus; + void (*cpu_freq_write)(struct acpi_pct_register *reg, u32 val); + u32 (*cpu_freq_read)(struct acpi_pct_register *reg); }; /* acpi_perf_data is a pointer to percpu data. */ @@ -243,125 +245,119 @@ static unsigned extract_freq(u32 val, struct acpi_cpufreq_data *data) } } -struct msr_addr { - u32 reg; -}; +u32 cpu_freq_read_intel(struct acpi_pct_register *not_used) +{ + u32 val, dummy; -struct io_addr { - u16 port; - u8 bit_width; -}; + rdmsr(MSR_IA32_PERF_CTL, val, dummy); + return val; +} + +void cpu_freq_write_intel(struct acpi_pct_register *not_used, u32 val) +{ + u32 lo, hi; + + rdmsr(MSR_IA32_PERF_CTL, lo, hi); + lo = (lo & ~INTEL_MSR_RANGE) | (val & INTEL_MSR_RANGE); + wrmsr(MSR_IA32_PERF_CTL, lo, hi); +} + +u32 cpu_freq_read_amd(struct acpi_pct_register *not_used) +{ + u32 val, dummy; + + rdmsr(MSR_AMD_PERF_CTL, val, dummy); + return val; +} + +void cpu_freq_write_amd(struct acpi_pct_register *not_used, u32 val) +{ + wrmsr(MSR_AMD_PERF_CTL, val, 0); +} + +u32 cpu_freq_read_io(struct acpi_pct_register *reg) +{ + u32 val; + + acpi_os_read_port(reg->address, &val, reg->bit_width); + return val; +} + +void cpu_freq_write_io(struct acpi_pct_register *reg, u32 val) +{ + acpi_os_write_port(reg->address, val, reg->bit_width); +} struct drv_cmd { - unsigned int type; - const struct cpumask *mask; - union { - struct msr_addr msr; - struct io_addr io; - } addr; + struct acpi_pct_register *reg; u32 val; + union { + void (*write)(struct acpi_pct_register *reg, u32 val); + u32 (*read)(struct acpi_pct_register *reg); + } func; }; /* Called via smp_call_function_single(), on the target CPU */ static void do_drv_read(void *_cmd) { struct drv_cmd *cmd = _cmd; - u32 h; - switch (cmd->type) { - case SYSTEM_INTEL_MSR_CAPABLE: - case SYSTEM_AMD_MSR_CAPABLE: - rdmsr(cmd->addr.msr.reg, cmd->val, h); - break; - case SYSTEM_IO_CAPABLE: - acpi_os_read_port((acpi_io_address)cmd->addr.io.port, - &cmd->val, - (u32)cmd->addr.io.bit_width); - break; - default: - break; - } + cmd->val = cmd->func.read(cmd->reg); } -/* Called via smp_call_function_many(), on the target CPUs */ -static void do_drv_write(void *_cmd) +static u32 drv_read(struct acpi_cpufreq_data *data, const struct cpumask *mask) { - struct drv_cmd *cmd = _cmd; - u32 lo, hi; + struct acpi_processor_performance *perf = to_perf_data(data); + struct drv_cmd cmd = { + .reg = &perf->control_register, + .func.read = data->cpu_freq_read, + }; + int err; - switch (cmd->type) { - case SYSTEM_INTEL_MSR_CAPABLE: - rdmsr(cmd->addr.msr.reg, lo, hi); - lo = (lo & ~INTEL_MSR_RANGE) | (cmd->val & INTEL_MSR_RANGE); - wrmsr(cmd->addr.msr.reg, lo, hi); - break; - case SYSTEM_AMD_MSR_CAPABLE: - wrmsr(cmd->addr.msr.reg, cmd->val, 0); - break; - case SYSTEM_IO_CAPABLE: - acpi_os_write_port((acpi_io_address)cmd->addr.io.port, - cmd->val, - (u32)cmd->addr.io.bit_width); - break; - default: - break; - } + err = smp_call_function_any(mask, do_drv_read, &cmd, 1); + WARN_ON_ONCE(err); /* smp_call_function_any() was buggy? */ + return cmd.val; } -static void drv_read(struct drv_cmd *cmd) +/* Called via smp_call_function_many(), on the target CPUs */ +static void do_drv_write(void *_cmd) { - int err; - cmd->val = 0; + struct drv_cmd *cmd = _cmd; - err = smp_call_function_any(cmd->mask, do_drv_read, cmd, 1); - WARN_ON_ONCE(err); /* smp_call_function_any() was buggy? */ + cmd->func.write(cmd->reg, cmd->val); } -static void drv_write(struct drv_cmd *cmd) +static void drv_write(struct acpi_cpufreq_data *data, + const struct cpumask *mask, u32 val) { + struct acpi_processor_performance *perf = to_perf_data(data); + struct drv_cmd cmd = { + .reg = &perf->control_register, + .val = val, + .func.write = data->cpu_freq_write, + }; int this_cpu; this_cpu = get_cpu(); - if (cpumask_test_cpu(this_cpu, cmd->mask)) - do_drv_write(cmd); - smp_call_function_many(cmd->mask, do_drv_write, cmd, 1); + if (cpumask_test_cpu(this_cpu, mask)) + do_drv_write(&cmd); + + smp_call_function_many(mask, do_drv_write, &cmd, 1); put_cpu(); } -static u32 -get_cur_val(const struct cpumask *mask, struct acpi_cpufreq_data *data) +static u32 get_cur_val(const struct cpumask *mask, struct acpi_cpufreq_data *data) { - struct acpi_processor_performance *perf; - struct drv_cmd cmd; + u32 val; if (unlikely(cpumask_empty(mask))) return 0; - switch (data->cpu_feature) { - case SYSTEM_INTEL_MSR_CAPABLE: - cmd.type = SYSTEM_INTEL_MSR_CAPABLE; - cmd.addr.msr.reg = MSR_IA32_PERF_CTL; - break; - case SYSTEM_AMD_MSR_CAPABLE: - cmd.type = SYSTEM_AMD_MSR_CAPABLE; - cmd.addr.msr.reg = MSR_AMD_PERF_CTL; - break; - case SYSTEM_IO_CAPABLE: - cmd.type = SYSTEM_IO_CAPABLE; - perf = to_perf_data(data); - cmd.addr.io.port = perf->control_register.address; - cmd.addr.io.bit_width = perf->control_register.bit_width; - break; - default: - return 0; - } - - cmd.mask = mask; - drv_read(&cmd); + val = drv_read(data, mask); - pr_debug("get_cur_val = %u\n", cmd.val); + pr_debug("get_cur_val = %u\n", val); - return cmd.val; + return val; } static unsigned int get_cur_freq_on_cpu(unsigned int cpu) @@ -416,7 +412,7 @@ static int acpi_cpufreq_target(struct cpufreq_policy *policy, { struct acpi_cpufreq_data *data = policy->driver_data; struct acpi_processor_performance *perf; - struct drv_cmd cmd; + const struct cpumask *mask; unsigned int next_perf_state = 0; /* Index into perf table */ int result = 0; @@ -434,42 +430,21 @@ static int acpi_cpufreq_target(struct cpufreq_policy *policy, } else { pr_debug("Already at target state (P%d)\n", next_perf_state); - goto out; + return 0; } } - switch (data->cpu_feature) { - case SYSTEM_INTEL_MSR_CAPABLE: - cmd.type = SYSTEM_INTEL_MSR_CAPABLE; - cmd.addr.msr.reg = MSR_IA32_PERF_CTL; - cmd.val = (u32) perf->states[next_perf_state].control; - break; - case SYSTEM_AMD_MSR_CAPABLE: - cmd.type = SYSTEM_AMD_MSR_CAPABLE; - cmd.addr.msr.reg = MSR_AMD_PERF_CTL; - cmd.val = (u32) perf->states[next_perf_state].control; - break; - case SYSTEM_IO_CAPABLE: - cmd.type = SYSTEM_IO_CAPABLE; - cmd.addr.io.port = perf->control_register.address; - cmd.addr.io.bit_width = perf->control_register.bit_width; - cmd.val = (u32) perf->states[next_perf_state].control; - break; - default: - result = -ENODEV; - goto out; - } - - /* cpufreq holds the hotplug lock, so we are safe from here on */ - if (policy->shared_type != CPUFREQ_SHARED_TYPE_ANY) - cmd.mask = policy->cpus; - else - cmd.mask = cpumask_of(policy->cpu); + /* + * The core won't allow CPUs to go away until the governor has been + * stopped, so we can rely on the stability of policy->cpus. + */ + mask = policy->shared_type == CPUFREQ_SHARED_TYPE_ANY ? + cpumask_of(policy->cpu) : policy->cpus; - drv_write(&cmd); + drv_write(data, mask, perf->states[next_perf_state].control); if (acpi_pstate_strict) { - if (!check_freqs(cmd.mask, data->freq_table[index].frequency, + if (!check_freqs(mask, data->freq_table[index].frequency, data)) { pr_debug("acpi_cpufreq_target failed (%d)\n", policy->cpu); @@ -480,7 +455,6 @@ static int acpi_cpufreq_target(struct cpufreq_policy *policy, if (!result) perf->state = next_perf_state; -out: return result; } @@ -740,15 +714,21 @@ static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy) } pr_debug("SYSTEM IO addr space\n"); data->cpu_feature = SYSTEM_IO_CAPABLE; + data->cpu_freq_read = cpu_freq_read_io; + data->cpu_freq_write = cpu_freq_write_io; break; case ACPI_ADR_SPACE_FIXED_HARDWARE: pr_debug("HARDWARE addr space\n"); if (check_est_cpu(cpu)) { data->cpu_feature = SYSTEM_INTEL_MSR_CAPABLE; + data->cpu_freq_read = cpu_freq_read_intel; + data->cpu_freq_write = cpu_freq_write_intel; break; } if (check_amd_hwpstate_cpu(cpu)) { data->cpu_feature = SYSTEM_AMD_MSR_CAPABLE; + data->cpu_freq_read = cpu_freq_read_amd; + data->cpu_freq_write = cpu_freq_write_amd; break; } result = -ENODEV; diff --git a/drivers/cpufreq/amd_freq_sensitivity.c b/drivers/cpufreq/amd_freq_sensitivity.c index f6b79ab..404360c 100644 --- a/drivers/cpufreq/amd_freq_sensitivity.c +++ b/drivers/cpufreq/amd_freq_sensitivity.c @@ -21,7 +21,7 @@ #include <asm/msr.h> #include <asm/cpufeature.h> -#include "cpufreq_governor.h" +#include "cpufreq_ondemand.h" #define MSR_AMD64_FREQ_SENSITIVITY_ACTUAL 0xc0010080 #define MSR_AMD64_FREQ_SENSITIVITY_REFERENCE 0xc0010081 @@ -45,10 +45,10 @@ static unsigned int amd_powersave_bias_target(struct cpufreq_policy *policy, long d_actual, d_reference; struct msr actual, reference; struct cpu_data_t *data = &per_cpu(cpu_data, policy->cpu); - struct dbs_data *od_data = policy->governor_data; + struct policy_dbs_info *policy_dbs = policy->governor_data; + struct dbs_data *od_data = policy_dbs->dbs_data; struct od_dbs_tuners *od_tuners = od_data->tuners; - struct od_cpu_dbs_info_s *od_info = - od_data->cdata->get_cpu_dbs_info_s(policy->cpu); + struct od_policy_dbs_info *od_info = to_dbs_info(policy_dbs); if (!od_info->freq_table) return freq_next; diff --git a/drivers/cpufreq/cpufreq-dt.c b/drivers/cpufreq/cpufreq-dt.c index 0ca74d0..f951f91 100644 --- a/drivers/cpufreq/cpufreq-dt.c +++ b/drivers/cpufreq/cpufreq-dt.c @@ -31,9 +31,8 @@ struct private_data { struct device *cpu_dev; - struct regulator *cpu_reg; struct thermal_cooling_device *cdev; - unsigned int voltage_tolerance; /* in percentage */ + const char *reg_name; }; static struct freq_attr *cpufreq_dt_attr[] = { @@ -44,175 +43,128 @@ static struct freq_attr *cpufreq_dt_attr[] = { static int set_target(struct cpufreq_policy *policy, unsigned int index) { - struct dev_pm_opp *opp; - struct cpufreq_frequency_table *freq_table = policy->freq_table; - struct clk *cpu_clk = policy->clk; struct private_data *priv = policy->driver_data; - struct device *cpu_dev = priv->cpu_dev; - struct regulator *cpu_reg = priv->cpu_reg; - unsigned long volt = 0, tol = 0; - int volt_old = 0; - unsigned int old_freq, new_freq; - long freq_Hz, freq_exact; - int ret; - - freq_Hz = clk_round_rate(cpu_clk, freq_table[index].frequency * 1000); - if (freq_Hz <= 0) - freq_Hz = freq_table[index].frequency * 1000; - freq_exact = freq_Hz; - new_freq = freq_Hz / 1000; - old_freq = clk_get_rate(cpu_clk) / 1000; + return dev_pm_opp_set_rate(priv->cpu_dev, + policy->freq_table[index].frequency * 1000); +} - if (!IS_ERR(cpu_reg)) { - unsigned long opp_freq; +/* + * An earlier version of opp-v1 bindings used to name the regulator + * "cpu0-supply", we still need to handle that for backwards compatibility. + */ +static const char *find_supply_name(struct device *dev) +{ + struct device_node *np; + struct property *pp; + int cpu = dev->id; + const char *name = NULL; - rcu_read_lock(); - opp = dev_pm_opp_find_freq_ceil(cpu_dev, &freq_Hz); - if (IS_ERR(opp)) { - rcu_read_unlock(); - dev_err(cpu_dev, "failed to find OPP for %ld\n", - freq_Hz); - return PTR_ERR(opp); - } - volt = dev_pm_opp_get_voltage(opp); - opp_freq = dev_pm_opp_get_freq(opp); - rcu_read_unlock(); - tol = volt * priv->voltage_tolerance / 100; - volt_old = regulator_get_voltage(cpu_reg); - dev_dbg(cpu_dev, "Found OPP: %ld kHz, %ld uV\n", - opp_freq / 1000, volt); - } + np = of_node_get(dev->of_node); - dev_dbg(cpu_dev, "%u MHz, %d mV --> %u MHz, %ld mV\n", - old_freq / 1000, (volt_old > 0) ? volt_old / 1000 : -1, - new_freq / 1000, volt ? volt / 1000 : -1); + /* This must be valid for sure */ + if (WARN_ON(!np)) + return NULL; - /* scaling up? scale voltage before frequency */ - if (!IS_ERR(cpu_reg) && new_freq > old_freq) { - ret = regulator_set_voltage_tol(cpu_reg, volt, tol); - if (ret) { - dev_err(cpu_dev, "failed to scale voltage up: %d\n", - ret); - return ret; + /* Try "cpu0" for older DTs */ + if (!cpu) { + pp = of_find_property(np, "cpu0-supply", NULL); + if (pp) { + name = "cpu0"; + goto node_put; } } - ret = clk_set_rate(cpu_clk, freq_exact); - if (ret) { - dev_err(cpu_dev, "failed to set clock rate: %d\n", ret); - if (!IS_ERR(cpu_reg) && volt_old > 0) - regulator_set_voltage_tol(cpu_reg, volt_old, tol); - return ret; + pp = of_find_property(np, "cpu-supply", NULL); + if (pp) { + name = "cpu"; + goto node_put; } - /* scaling down? scale voltage after frequency */ - if (!IS_ERR(cpu_reg) && new_freq < old_freq) { - ret = regulator_set_voltage_tol(cpu_reg, volt, tol); - if (ret) { - dev_err(cpu_dev, "failed to scale voltage down: %d\n", - ret); - clk_set_rate(cpu_clk, old_freq * 1000); - } - } - - return ret; + dev_dbg(dev, "no regulator for cpu%d\n", cpu); +node_put: + of_node_put(np); + return name; } -static int allocate_resources(int cpu, struct device **cdev, - struct regulator **creg, struct clk **cclk) +static int resources_available(void) { struct device *cpu_dev; struct regulator *cpu_reg; struct clk *cpu_clk; int ret = 0; - char *reg_cpu0 = "cpu0", *reg_cpu = "cpu", *reg; + const char *name; - cpu_dev = get_cpu_device(cpu); + cpu_dev = get_cpu_device(0); if (!cpu_dev) { - pr_err("failed to get cpu%d device\n", cpu); + pr_err("failed to get cpu0 device\n"); return -ENODEV; } - /* Try "cpu0" for older DTs */ - if (!cpu) - reg = reg_cpu0; - else - reg = reg_cpu; - -try_again: - cpu_reg = regulator_get_optional(cpu_dev, reg); - ret = PTR_ERR_OR_ZERO(cpu_reg); + cpu_clk = clk_get(cpu_dev, NULL); + ret = PTR_ERR_OR_ZERO(cpu_clk); if (ret) { /* - * If cpu's regulator supply node is present, but regulator is - * not yet registered, we should try defering probe. + * If cpu's clk node is present, but clock is not yet + * registered, we should try defering probe. */ - if (ret == -EPROBE_DEFER) { - dev_dbg(cpu_dev, "cpu%d regulator not ready, retry\n", - cpu); - return ret; - } - - /* Try with "cpu-supply" */ - if (reg == reg_cpu0) { - reg = reg_cpu; - goto try_again; - } + if (ret == -EPROBE_DEFER) + dev_dbg(cpu_dev, "clock not ready, retry\n"); + else + dev_err(cpu_dev, "failed to get clock: %d\n", ret); - dev_dbg(cpu_dev, "no regulator for cpu%d: %d\n", cpu, ret); + return ret; } - cpu_clk = clk_get(cpu_dev, NULL); - ret = PTR_ERR_OR_ZERO(cpu_clk); - if (ret) { - /* put regulator */ - if (!IS_ERR(cpu_reg)) - regulator_put(cpu_reg); + clk_put(cpu_clk); + name = find_supply_name(cpu_dev); + /* Platform doesn't require regulator */ + if (!name) + return 0; + + cpu_reg = regulator_get_optional(cpu_dev, name); + ret = PTR_ERR_OR_ZERO(cpu_reg); + if (ret) { /* - * If cpu's clk node is present, but clock is not yet - * registered, we should try defering probe. + * If cpu's regulator supply node is present, but regulator is + * not yet registered, we should try defering probe. */ if (ret == -EPROBE_DEFER) - dev_dbg(cpu_dev, "cpu%d clock not ready, retry\n", cpu); + dev_dbg(cpu_dev, "cpu0 regulator not ready, retry\n"); else - dev_err(cpu_dev, "failed to get cpu%d clock: %d\n", cpu, - ret); - } else { - *cdev = cpu_dev; - *creg = cpu_reg; - *cclk = cpu_clk; + dev_dbg(cpu_dev, "no regulator for cpu0: %d\n", ret); + + return ret; } - return ret; + regulator_put(cpu_reg); + return 0; } static int cpufreq_init(struct cpufreq_policy *policy) { struct cpufreq_frequency_table *freq_table; - struct device_node *np; struct private_data *priv; struct device *cpu_dev; - struct regulator *cpu_reg; struct clk *cpu_clk; struct dev_pm_opp *suspend_opp; - unsigned long min_uV = ~0, max_uV = 0; unsigned int transition_latency; - bool need_update = false; + bool opp_v1 = false; + const char *name; int ret; - ret = allocate_resources(policy->cpu, &cpu_dev, &cpu_reg, &cpu_clk); - if (ret) { - pr_err("%s: Failed to allocate resources: %d\n", __func__, ret); - return ret; + cpu_dev = get_cpu_device(policy->cpu); + if (!cpu_dev) { + pr_err("failed to get cpu%d device\n", policy->cpu); + return -ENODEV; } - np = of_node_get(cpu_dev->of_node); - if (!np) { - dev_err(cpu_dev, "failed to find cpu%d node\n", policy->cpu); - ret = -ENOENT; - goto out_put_reg_clk; + cpu_clk = clk_get(cpu_dev, NULL); + if (IS_ERR(cpu_clk)) { + ret = PTR_ERR(cpu_clk); + dev_err(cpu_dev, "%s: failed to get clk: %d\n", __func__, ret); + return ret; } /* Get OPP-sharing information from "operating-points-v2" bindings */ @@ -223,9 +175,23 @@ static int cpufreq_init(struct cpufreq_policy *policy) * finding shared-OPPs for backward compatibility. */ if (ret == -ENOENT) - need_update = true; + opp_v1 = true; else - goto out_node_put; + goto out_put_clk; + } + + /* + * OPP layer will be taking care of regulators now, but it needs to know + * the name of the regulator first. + */ + name = find_supply_name(cpu_dev); + if (name) { + ret = dev_pm_opp_set_regulator(cpu_dev, name); + if (ret) { + dev_err(cpu_dev, "Failed to set regulator for cpu%d: %d\n", + policy->cpu, ret); + goto out_put_clk; + } } /* @@ -246,12 +212,12 @@ static int cpufreq_init(struct cpufreq_policy *policy) */ ret = dev_pm_opp_get_opp_count(cpu_dev); if (ret <= 0) { - pr_debug("OPP table is not ready, deferring probe\n"); + dev_dbg(cpu_dev, "OPP table is not ready, deferring probe\n"); ret = -EPROBE_DEFER; goto out_free_opp; } - if (need_update) { + if (opp_v1) { struct cpufreq_dt_platform_data *pd = cpufreq_get_driver_data(); if (!pd || !pd->independent_clocks) @@ -265,10 +231,6 @@ static int cpufreq_init(struct cpufreq_policy *policy) if (ret) dev_err(cpu_dev, "%s: failed to mark OPPs as shared: %d\n", __func__, ret); - - of_property_read_u32(np, "clock-latency", &transition_latency); - } else { - transition_latency = dev_pm_opp_get_max_clock_latency(cpu_dev); } priv = kzalloc(sizeof(*priv), GFP_KERNEL); @@ -277,62 +239,16 @@ static int cpufreq_init(struct cpufreq_policy *policy) goto out_free_opp; } - of_property_read_u32(np, "voltage-tolerance", &priv->voltage_tolerance); - - if (!transition_latency) - transition_latency = CPUFREQ_ETERNAL; - - if (!IS_ERR(cpu_reg)) { - unsigned long opp_freq = 0; - - /* - * Disable any OPPs where the connected regulator isn't able to - * provide the specified voltage and record minimum and maximum - * voltage levels. - */ - while (1) { - struct dev_pm_opp *opp; - unsigned long opp_uV, tol_uV; - - rcu_read_lock(); - opp = dev_pm_opp_find_freq_ceil(cpu_dev, &opp_freq); - if (IS_ERR(opp)) { - rcu_read_unlock(); - break; - } - opp_uV = dev_pm_opp_get_voltage(opp); - rcu_read_unlock(); - - tol_uV = opp_uV * priv->voltage_tolerance / 100; - if (regulator_is_supported_voltage(cpu_reg, - opp_uV - tol_uV, - opp_uV + tol_uV)) { - if (opp_uV < min_uV) - min_uV = opp_uV; - if (opp_uV > max_uV) - max_uV = opp_uV; - } else { - dev_pm_opp_disable(cpu_dev, opp_freq); - } - - opp_freq++; - } - - ret = regulator_set_voltage_time(cpu_reg, min_uV, max_uV); - if (ret > 0) - transition_latency += ret * 1000; - } + priv->reg_name = name; ret = dev_pm_opp_init_cpufreq_table(cpu_dev, &freq_table); if (ret) { - pr_err("failed to init cpufreq table: %d\n", ret); + dev_err(cpu_dev, "failed to init cpufreq table: %d\n", ret); goto out_free_priv; } priv->cpu_dev = cpu_dev; - priv->cpu_reg = cpu_reg; policy->driver_data = priv; - policy->clk = cpu_clk; rcu_read_lock(); @@ -357,9 +273,11 @@ static int cpufreq_init(struct cpufreq_policy *policy) cpufreq_dt_attr[1] = &cpufreq_freq_attr_scaling_boost_freqs; } - policy->cpuinfo.transition_latency = transition_latency; + transition_latency = dev_pm_opp_get_max_transition_latency(cpu_dev); + if (!transition_latency) + transition_latency = CPUFREQ_ETERNAL; - of_node_put(np); + policy->cpuinfo.transition_latency = transition_latency; return 0; @@ -369,12 +287,10 @@ out_free_priv: kfree(priv); out_free_opp: dev_pm_opp_of_cpumask_remove_table(policy->cpus); -out_node_put: - of_node_put(np); -out_put_reg_clk: + if (name) + dev_pm_opp_put_regulator(cpu_dev); +out_put_clk: clk_put(cpu_clk); - if (!IS_ERR(cpu_reg)) - regulator_put(cpu_reg); return ret; } @@ -386,9 +302,10 @@ static int cpufreq_exit(struct cpufreq_policy *policy) cpufreq_cooling_unregister(priv->cdev); dev_pm_opp_free_cpufreq_table(priv->cpu_dev, &policy->freq_table); dev_pm_opp_of_cpumask_remove_table(policy->related_cpus); + if (priv->reg_name) + dev_pm_opp_put_regulator(priv->cpu_dev); + clk_put(policy->clk); - if (!IS_ERR(priv->cpu_reg)) - regulator_put(priv->cpu_reg); kfree(priv); return 0; @@ -441,9 +358,6 @@ static struct cpufreq_driver dt_cpufreq_driver = { static int dt_cpufreq_probe(struct platform_device *pdev) { - struct device *cpu_dev; - struct regulator *cpu_reg; - struct clk *cpu_clk; int ret; /* @@ -453,19 +367,15 @@ static int dt_cpufreq_probe(struct platform_device *pdev) * * FIXME: Is checking this only for CPU0 sufficient ? */ - ret = allocate_resources(0, &cpu_dev, &cpu_reg, &cpu_clk); + ret = resources_available(); if (ret) return ret; - clk_put(cpu_clk); - if (!IS_ERR(cpu_reg)) - regulator_put(cpu_reg); - dt_cpufreq_driver.driver_data = dev_get_platdata(&pdev->dev); ret = cpufreq_register_driver(&dt_cpufreq_driver); if (ret) - dev_err(cpu_dev, "failed register driver: %d\n", ret); + dev_err(&pdev->dev, "failed register driver: %d\n", ret); return ret; } diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index e979ec7..4c78258 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -38,48 +38,10 @@ static inline bool policy_is_inactive(struct cpufreq_policy *policy) return cpumask_empty(policy->cpus); } -static bool suitable_policy(struct cpufreq_policy *policy, bool active) -{ - return active == !policy_is_inactive(policy); -} - -/* Finds Next Acive/Inactive policy */ -static struct cpufreq_policy *next_policy(struct cpufreq_policy *policy, - bool active) -{ - do { - /* No more policies in the list */ - if (list_is_last(&policy->policy_list, &cpufreq_policy_list)) - return NULL; - - policy = list_next_entry(policy, policy_list); - } while (!suitable_policy(policy, active)); - - return policy; -} - -static struct cpufreq_policy *first_policy(bool active) -{ - struct cpufreq_policy *policy; - - /* No policies in the list */ - if (list_empty(&cpufreq_policy_list)) - return NULL; - - policy = list_first_entry(&cpufreq_policy_list, typeof(*policy), - policy_list); - - if (!suitable_policy(policy, active)) - policy = next_policy(policy, active); - - return policy; -} - /* Macros to iterate over CPU policies */ -#define for_each_suitable_policy(__policy, __active) \ - for (__policy = first_policy(__active); \ - __policy; \ - __policy = next_policy(__policy, __active)) +#define for_each_suitable_policy(__policy, __active) \ + list_for_each_entry(__policy, &cpufreq_policy_list, policy_list) \ + if ((__active) == !policy_is_inactive(__policy)) #define for_each_active_policy(__policy) \ for_each_suitable_policy(__policy, true) @@ -102,7 +64,6 @@ static LIST_HEAD(cpufreq_governor_list); static struct cpufreq_driver *cpufreq_driver; static DEFINE_PER_CPU(struct cpufreq_policy *, cpufreq_cpu_data); static DEFINE_RWLOCK(cpufreq_driver_lock); -DEFINE_MUTEX(cpufreq_governor_lock); /* Flag to suspend/resume CPUFreq governors */ static bool cpufreq_suspended; @@ -113,10 +74,8 @@ static inline bool has_target(void) } /* internal prototypes */ -static int __cpufreq_governor(struct cpufreq_policy *policy, - unsigned int event); +static int cpufreq_governor(struct cpufreq_policy *policy, unsigned int event); static unsigned int __cpufreq_get(struct cpufreq_policy *policy); -static void handle_update(struct work_struct *work); /** * Two notifier lists: the "policy" list is involved in the @@ -818,12 +777,7 @@ static ssize_t show(struct kobject *kobj, struct attribute *attr, char *buf) ssize_t ret; down_read(&policy->rwsem); - - if (fattr->show) - ret = fattr->show(policy, buf); - else - ret = -EIO; - + ret = fattr->show(policy, buf); up_read(&policy->rwsem); return ret; @@ -838,18 +792,12 @@ static ssize_t store(struct kobject *kobj, struct attribute *attr, get_online_cpus(); - if (!cpu_online(policy->cpu)) - goto unlock; - - down_write(&policy->rwsem); - - if (fattr->store) + if (cpu_online(policy->cpu)) { + down_write(&policy->rwsem); ret = fattr->store(policy, buf, count); - else - ret = -EIO; + up_write(&policy->rwsem); + } - up_write(&policy->rwsem); -unlock: put_online_cpus(); return ret; @@ -959,6 +907,11 @@ static int cpufreq_add_dev_interface(struct cpufreq_policy *policy) return cpufreq_add_dev_symlink(policy); } +__weak struct cpufreq_governor *cpufreq_default_governor(void) +{ + return NULL; +} + static int cpufreq_init_policy(struct cpufreq_policy *policy) { struct cpufreq_governor *gov = NULL; @@ -968,11 +921,14 @@ static int cpufreq_init_policy(struct cpufreq_policy *policy) /* Update governor of new_policy to the governor used before hotplug */ gov = find_governor(policy->last_governor); - if (gov) + if (gov) { pr_debug("Restoring governor %s for cpu %d\n", policy->governor->name, policy->cpu); - else - gov = CPUFREQ_DEFAULT_GOVERNOR; + } else { + gov = cpufreq_default_governor(); + if (!gov) + return -ENODATA; + } new_policy.governor = gov; @@ -996,36 +952,45 @@ static int cpufreq_add_policy_cpu(struct cpufreq_policy *policy, unsigned int cp if (cpumask_test_cpu(cpu, policy->cpus)) return 0; + down_write(&policy->rwsem); if (has_target()) { - ret = __cpufreq_governor(policy, CPUFREQ_GOV_STOP); + ret = cpufreq_governor(policy, CPUFREQ_GOV_STOP); if (ret) { pr_err("%s: Failed to stop governor\n", __func__); - return ret; + goto unlock; } } - down_write(&policy->rwsem); cpumask_set_cpu(cpu, policy->cpus); - up_write(&policy->rwsem); if (has_target()) { - ret = __cpufreq_governor(policy, CPUFREQ_GOV_START); + ret = cpufreq_governor(policy, CPUFREQ_GOV_START); if (!ret) - ret = __cpufreq_governor(policy, CPUFREQ_GOV_LIMITS); + ret = cpufreq_governor(policy, CPUFREQ_GOV_LIMITS); - if (ret) { + if (ret) pr_err("%s: Failed to start governor\n", __func__); - return ret; - } } - return 0; +unlock: + up_write(&policy->rwsem); + return ret; +} + +static void handle_update(struct work_struct *work) +{ + struct cpufreq_policy *policy = + container_of(work, struct cpufreq_policy, update); + unsigned int cpu = policy->cpu; + pr_debug("handle_update for cpu %u called\n", cpu); + cpufreq_update_policy(cpu); } static struct cpufreq_policy *cpufreq_policy_alloc(unsigned int cpu) { struct device *dev = get_cpu_device(cpu); struct cpufreq_policy *policy; + int ret; if (WARN_ON(!dev)) return NULL; @@ -1043,7 +1008,13 @@ static struct cpufreq_policy *cpufreq_policy_alloc(unsigned int cpu) if (!zalloc_cpumask_var(&policy->real_cpus, GFP_KERNEL)) goto err_free_rcpumask; - kobject_init(&policy->kobj, &ktype_cpufreq); + ret = kobject_init_and_add(&policy->kobj, &ktype_cpufreq, + cpufreq_global_kobject, "policy%u", cpu); + if (ret) { + pr_err("%s: failed to init policy->kobj: %d\n", __func__, ret); + goto err_free_real_cpus; + } + INIT_LIST_HEAD(&policy->policy_list); init_rwsem(&policy->rwsem); spin_lock_init(&policy->transition_lock); @@ -1054,6 +1025,8 @@ static struct cpufreq_policy *cpufreq_policy_alloc(unsigned int cpu) policy->cpu = cpu; return policy; +err_free_real_cpus: + free_cpumask_var(policy->real_cpus); err_free_rcpumask: free_cpumask_var(policy->related_cpus); err_free_cpumask: @@ -1158,16 +1131,6 @@ static int cpufreq_online(unsigned int cpu) cpumask_copy(policy->related_cpus, policy->cpus); /* Remember CPUs present at the policy creation time. */ cpumask_and(policy->real_cpus, policy->cpus, cpu_present_mask); - - /* Name and add the kobject */ - ret = kobject_add(&policy->kobj, cpufreq_global_kobject, - "policy%u", - cpumask_first(policy->related_cpus)); - if (ret) { - pr_err("%s: failed to add policy->kobj: %d\n", __func__, - ret); - goto out_exit_policy; - } } /* @@ -1309,9 +1272,10 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif) return ret; } -static void cpufreq_offline_prepare(unsigned int cpu) +static void cpufreq_offline(unsigned int cpu) { struct cpufreq_policy *policy; + int ret; pr_debug("%s: unregistering CPU %u\n", __func__, cpu); @@ -1321,13 +1285,13 @@ static void cpufreq_offline_prepare(unsigned int cpu) return; } + down_write(&policy->rwsem); if (has_target()) { - int ret = __cpufreq_governor(policy, CPUFREQ_GOV_STOP); + ret = cpufreq_governor(policy, CPUFREQ_GOV_STOP); if (ret) pr_err("%s: Failed to stop governor\n", __func__); } - down_write(&policy->rwsem); cpumask_clear_cpu(cpu, policy->cpus); if (policy_is_inactive(policy)) { @@ -1340,39 +1304,27 @@ static void cpufreq_offline_prepare(unsigned int cpu) /* Nominate new CPU */ policy->cpu = cpumask_any(policy->cpus); } - up_write(&policy->rwsem); /* Start governor again for active policy */ if (!policy_is_inactive(policy)) { if (has_target()) { - int ret = __cpufreq_governor(policy, CPUFREQ_GOV_START); + ret = cpufreq_governor(policy, CPUFREQ_GOV_START); if (!ret) - ret = __cpufreq_governor(policy, CPUFREQ_GOV_LIMITS); + ret = cpufreq_governor(policy, CPUFREQ_GOV_LIMITS); if (ret) pr_err("%s: Failed to start governor\n", __func__); } - } else if (cpufreq_driver->stop_cpu) { - cpufreq_driver->stop_cpu(policy); - } -} -static void cpufreq_offline_finish(unsigned int cpu) -{ - struct cpufreq_policy *policy = per_cpu(cpufreq_cpu_data, cpu); - - if (!policy) { - pr_debug("%s: No cpu_data found\n", __func__); - return; + goto unlock; } - /* Only proceed for inactive policies */ - if (!policy_is_inactive(policy)) - return; + if (cpufreq_driver->stop_cpu) + cpufreq_driver->stop_cpu(policy); /* If cpu is last user of policy, free policy */ if (has_target()) { - int ret = __cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT); + ret = cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT); if (ret) pr_err("%s: Failed to exit governor\n", __func__); } @@ -1386,6 +1338,9 @@ static void cpufreq_offline_finish(unsigned int cpu) cpufreq_driver->exit(policy); policy->freq_table = NULL; } + +unlock: + up_write(&policy->rwsem); } /** @@ -1401,10 +1356,8 @@ static void cpufreq_remove_dev(struct device *dev, struct subsys_interface *sif) if (!policy) return; - if (cpu_online(cpu)) { - cpufreq_offline_prepare(cpu); - cpufreq_offline_finish(cpu); - } + if (cpu_online(cpu)) + cpufreq_offline(cpu); cpumask_clear_cpu(cpu, policy->real_cpus); remove_cpu_dev_symlink(policy, cpu); @@ -1413,15 +1366,6 @@ static void cpufreq_remove_dev(struct device *dev, struct subsys_interface *sif) cpufreq_policy_free(policy, true); } -static void handle_update(struct work_struct *work) -{ - struct cpufreq_policy *policy = - container_of(work, struct cpufreq_policy, update); - unsigned int cpu = policy->cpu; - pr_debug("handle_update for cpu %u called\n", cpu); - cpufreq_update_policy(cpu); -} - /** * cpufreq_out_of_sync - If actual and saved CPU frequency differs, we're * in deep trouble. @@ -1584,6 +1528,7 @@ EXPORT_SYMBOL(cpufreq_generic_suspend); void cpufreq_suspend(void) { struct cpufreq_policy *policy; + int ret; if (!cpufreq_driver) return; @@ -1594,7 +1539,11 @@ void cpufreq_suspend(void) pr_debug("%s: Suspending Governors\n", __func__); for_each_active_policy(policy) { - if (__cpufreq_governor(policy, CPUFREQ_GOV_STOP)) + down_write(&policy->rwsem); + ret = cpufreq_governor(policy, CPUFREQ_GOV_STOP); + up_write(&policy->rwsem); + + if (ret) pr_err("%s: Failed to stop governor for policy: %p\n", __func__, policy); else if (cpufreq_driver->suspend @@ -1616,6 +1565,7 @@ suspend: void cpufreq_resume(void) { struct cpufreq_policy *policy; + int ret; if (!cpufreq_driver) return; @@ -1628,13 +1578,20 @@ void cpufreq_resume(void) pr_debug("%s: Resuming Governors\n", __func__); for_each_active_policy(policy) { - if (cpufreq_driver->resume && cpufreq_driver->resume(policy)) + if (cpufreq_driver->resume && cpufreq_driver->resume(policy)) { pr_err("%s: Failed to resume driver: %p\n", __func__, policy); - else if (__cpufreq_governor(policy, CPUFREQ_GOV_START) - || __cpufreq_governor(policy, CPUFREQ_GOV_LIMITS)) - pr_err("%s: Failed to start governor for policy: %p\n", - __func__, policy); + } else { + down_write(&policy->rwsem); + ret = cpufreq_governor(policy, CPUFREQ_GOV_START); + if (!ret) + cpufreq_governor(policy, CPUFREQ_GOV_LIMITS); + up_write(&policy->rwsem); + + if (ret) + pr_err("%s: Failed to start governor for policy: %p\n", + __func__, policy); + } } /* @@ -1846,7 +1803,8 @@ int __cpufreq_driver_target(struct cpufreq_policy *policy, unsigned int relation) { unsigned int old_target_freq = target_freq; - int retval = -EINVAL; + struct cpufreq_frequency_table *freq_table; + int index, retval; if (cpufreq_disabled()) return -ENODEV; @@ -1873,34 +1831,28 @@ int __cpufreq_driver_target(struct cpufreq_policy *policy, policy->restore_freq = policy->cur; if (cpufreq_driver->target) - retval = cpufreq_driver->target(policy, target_freq, relation); - else if (cpufreq_driver->target_index) { - struct cpufreq_frequency_table *freq_table; - int index; - - freq_table = cpufreq_frequency_get_table(policy->cpu); - if (unlikely(!freq_table)) { - pr_err("%s: Unable to find freq_table\n", __func__); - goto out; - } + return cpufreq_driver->target(policy, target_freq, relation); - retval = cpufreq_frequency_table_target(policy, freq_table, - target_freq, relation, &index); - if (unlikely(retval)) { - pr_err("%s: Unable to find matching freq\n", __func__); - goto out; - } + if (!cpufreq_driver->target_index) + return -EINVAL; - if (freq_table[index].frequency == policy->cur) { - retval = 0; - goto out; - } + freq_table = cpufreq_frequency_get_table(policy->cpu); + if (unlikely(!freq_table)) { + pr_err("%s: Unable to find freq_table\n", __func__); + return -EINVAL; + } - retval = __target_index(policy, freq_table, index); + retval = cpufreq_frequency_table_target(policy, freq_table, target_freq, + relation, &index); + if (unlikely(retval)) { + pr_err("%s: Unable to find matching freq\n", __func__); + return retval; } -out: - return retval; + if (freq_table[index].frequency == policy->cur) + return 0; + + return __target_index(policy, freq_table, index); } EXPORT_SYMBOL_GPL(__cpufreq_driver_target); @@ -1920,20 +1872,14 @@ int cpufreq_driver_target(struct cpufreq_policy *policy, } EXPORT_SYMBOL_GPL(cpufreq_driver_target); -static int __cpufreq_governor(struct cpufreq_policy *policy, - unsigned int event) +__weak struct cpufreq_governor *cpufreq_fallback_governor(void) { - int ret; + return NULL; +} - /* Only must be defined when default governor is known to have latency - restrictions, like e.g. conservative or ondemand. - That this is the case is already ensured in Kconfig - */ -#ifdef CONFIG_CPU_FREQ_GOV_PERFORMANCE - struct cpufreq_governor *gov = &cpufreq_gov_performance; -#else - struct cpufreq_governor *gov = NULL; -#endif +static int cpufreq_governor(struct cpufreq_policy *policy, unsigned int event) +{ + int ret; /* Don't start any governor operations if we are entering suspend */ if (cpufreq_suspended) @@ -1948,12 +1894,14 @@ static int __cpufreq_governor(struct cpufreq_policy *policy, if (policy->governor->max_transition_latency && policy->cpuinfo.transition_latency > policy->governor->max_transition_latency) { - if (!gov) - return -EINVAL; - else { + struct cpufreq_governor *gov = cpufreq_fallback_governor(); + + if (gov) { pr_warn("%s governor failed, too long transition latency of HW, fallback to %s governor\n", policy->governor->name, gov->name); policy->governor = gov; + } else { + return -EINVAL; } } @@ -1963,21 +1911,6 @@ static int __cpufreq_governor(struct cpufreq_policy *policy, pr_debug("%s: for CPU %u, event %u\n", __func__, policy->cpu, event); - mutex_lock(&cpufreq_governor_lock); - if ((policy->governor_enabled && event == CPUFREQ_GOV_START) - || (!policy->governor_enabled - && (event == CPUFREQ_GOV_LIMITS || event == CPUFREQ_GOV_STOP))) { - mutex_unlock(&cpufreq_governor_lock); - return -EBUSY; - } - - if (event == CPUFREQ_GOV_STOP) - policy->governor_enabled = false; - else if (event == CPUFREQ_GOV_START) - policy->governor_enabled = true; - - mutex_unlock(&cpufreq_governor_lock); - ret = policy->governor->governor(policy, event); if (!ret) { @@ -1985,14 +1918,6 @@ static int __cpufreq_governor(struct cpufreq_policy *policy, policy->governor->initialized++; else if (event == CPUFREQ_GOV_POLICY_EXIT) policy->governor->initialized--; - } else { - /* Restore original values */ - mutex_lock(&cpufreq_governor_lock); - if (event == CPUFREQ_GOV_STOP) - policy->governor_enabled = true; - else if (event == CPUFREQ_GOV_START) - policy->governor_enabled = false; - mutex_unlock(&cpufreq_governor_lock); } if (((event == CPUFREQ_GOV_POLICY_INIT) && ret) || @@ -2147,7 +2072,7 @@ static int cpufreq_set_policy(struct cpufreq_policy *policy, old_gov = policy->governor; /* end old governor */ if (old_gov) { - ret = __cpufreq_governor(policy, CPUFREQ_GOV_STOP); + ret = cpufreq_governor(policy, CPUFREQ_GOV_STOP); if (ret) { /* This can happen due to race with other operations */ pr_debug("%s: Failed to Stop Governor: %s (%d)\n", @@ -2155,10 +2080,7 @@ static int cpufreq_set_policy(struct cpufreq_policy *policy, return ret; } - up_write(&policy->rwsem); - ret = __cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT); - down_write(&policy->rwsem); - + ret = cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT); if (ret) { pr_err("%s: Failed to Exit Governor: %s (%d)\n", __func__, old_gov->name, ret); @@ -2168,32 +2090,30 @@ static int cpufreq_set_policy(struct cpufreq_policy *policy, /* start new governor */ policy->governor = new_policy->governor; - ret = __cpufreq_governor(policy, CPUFREQ_GOV_POLICY_INIT); + ret = cpufreq_governor(policy, CPUFREQ_GOV_POLICY_INIT); if (!ret) { - ret = __cpufreq_governor(policy, CPUFREQ_GOV_START); + ret = cpufreq_governor(policy, CPUFREQ_GOV_START); if (!ret) goto out; - up_write(&policy->rwsem); - __cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT); - down_write(&policy->rwsem); + cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT); } /* new governor failed, so re-start old one */ pr_debug("starting governor %s failed\n", policy->governor->name); if (old_gov) { policy->governor = old_gov; - if (__cpufreq_governor(policy, CPUFREQ_GOV_POLICY_INIT)) + if (cpufreq_governor(policy, CPUFREQ_GOV_POLICY_INIT)) policy->governor = NULL; else - __cpufreq_governor(policy, CPUFREQ_GOV_START); + cpufreq_governor(policy, CPUFREQ_GOV_START); } return ret; out: pr_debug("governor: change or update limits\n"); - return __cpufreq_governor(policy, CPUFREQ_GOV_LIMITS); + return cpufreq_governor(policy, CPUFREQ_GOV_LIMITS); } /** @@ -2260,11 +2180,7 @@ static int cpufreq_cpu_callback(struct notifier_block *nfb, break; case CPU_DOWN_PREPARE: - cpufreq_offline_prepare(cpu); - break; - - case CPU_POST_DEAD: - cpufreq_offline_finish(cpu); + cpufreq_offline(cpu); break; case CPU_DOWN_FAILED: @@ -2297,8 +2213,11 @@ static int cpufreq_boost_set_sw(int state) __func__); break; } + + down_write(&policy->rwsem); policy->user_policy.max = policy->max; - __cpufreq_governor(policy, CPUFREQ_GOV_LIMITS); + cpufreq_governor(policy, CPUFREQ_GOV_LIMITS); + up_write(&policy->rwsem); } } @@ -2384,7 +2303,7 @@ EXPORT_SYMBOL_GPL(cpufreq_boost_enabled); * submitted by the CPU Frequency driver. * * Registers a CPU Frequency driver to this core code. This code - * returns zero on success, -EBUSY when another driver got here first + * returns zero on success, -EEXIST when another driver got here first * (and isn't unregistered in the meantime). * */ diff --git a/drivers/cpufreq/cpufreq_conservative.c b/drivers/cpufreq/cpufreq_conservative.c index 606ad74ab..bf4913f 100644 --- a/drivers/cpufreq/cpufreq_conservative.c +++ b/drivers/cpufreq/cpufreq_conservative.c @@ -14,6 +14,22 @@ #include <linux/slab.h> #include "cpufreq_governor.h" +struct cs_policy_dbs_info { + struct policy_dbs_info policy_dbs; + unsigned int down_skip; + unsigned int requested_freq; +}; + +static inline struct cs_policy_dbs_info *to_dbs_info(struct policy_dbs_info *policy_dbs) +{ + return container_of(policy_dbs, struct cs_policy_dbs_info, policy_dbs); +} + +struct cs_dbs_tuners { + unsigned int down_threshold; + unsigned int freq_step; +}; + /* Conservative governor macros */ #define DEF_FREQUENCY_UP_THRESHOLD (80) #define DEF_FREQUENCY_DOWN_THRESHOLD (20) @@ -21,21 +37,6 @@ #define DEF_SAMPLING_DOWN_FACTOR (1) #define MAX_SAMPLING_DOWN_FACTOR (10) -static DEFINE_PER_CPU(struct cs_cpu_dbs_info_s, cs_cpu_dbs_info); - -static int cs_cpufreq_governor_dbs(struct cpufreq_policy *policy, - unsigned int event); - -#ifndef CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE -static -#endif -struct cpufreq_governor cpufreq_gov_conservative = { - .name = "conservative", - .governor = cs_cpufreq_governor_dbs, - .max_transition_latency = TRANSITION_LATENCY_LIMIT, - .owner = THIS_MODULE, -}; - static inline unsigned int get_freq_target(struct cs_dbs_tuners *cs_tuners, struct cpufreq_policy *policy) { @@ -57,27 +58,28 @@ static inline unsigned int get_freq_target(struct cs_dbs_tuners *cs_tuners, * Any frequency increase takes it to the maximum frequency. Frequency reduction * happens at minimum steps of 5% (default) of maximum frequency */ -static void cs_check_cpu(int cpu, unsigned int load) +static unsigned int cs_dbs_timer(struct cpufreq_policy *policy) { - struct cs_cpu_dbs_info_s *dbs_info = &per_cpu(cs_cpu_dbs_info, cpu); - struct cpufreq_policy *policy = dbs_info->cdbs.shared->policy; - struct dbs_data *dbs_data = policy->governor_data; + struct policy_dbs_info *policy_dbs = policy->governor_data; + struct cs_policy_dbs_info *dbs_info = to_dbs_info(policy_dbs); + struct dbs_data *dbs_data = policy_dbs->dbs_data; struct cs_dbs_tuners *cs_tuners = dbs_data->tuners; + unsigned int load = dbs_update(policy); /* * break out if we 'cannot' reduce the speed as the user might * want freq_step to be zero */ if (cs_tuners->freq_step == 0) - return; + goto out; /* Check for frequency increase */ - if (load > cs_tuners->up_threshold) { + if (load > dbs_data->up_threshold) { dbs_info->down_skip = 0; /* if we are already at full speed then break out early */ if (dbs_info->requested_freq == policy->max) - return; + goto out; dbs_info->requested_freq += get_freq_target(cs_tuners, policy); @@ -86,12 +88,12 @@ static void cs_check_cpu(int cpu, unsigned int load) __cpufreq_driver_target(policy, dbs_info->requested_freq, CPUFREQ_RELATION_H); - return; + goto out; } /* if sampling_down_factor is active break out early */ - if (++dbs_info->down_skip < cs_tuners->sampling_down_factor) - return; + if (++dbs_info->down_skip < dbs_data->sampling_down_factor) + goto out; dbs_info->down_skip = 0; /* Check for frequency decrease */ @@ -101,7 +103,7 @@ static void cs_check_cpu(int cpu, unsigned int load) * if we cannot reduce the frequency anymore, break out early */ if (policy->cur == policy->min) - return; + goto out; freq_target = get_freq_target(cs_tuners, policy); if (dbs_info->requested_freq > freq_target) @@ -111,58 +113,25 @@ static void cs_check_cpu(int cpu, unsigned int load) __cpufreq_driver_target(policy, dbs_info->requested_freq, CPUFREQ_RELATION_L); - return; } -} - -static unsigned int cs_dbs_timer(struct cpufreq_policy *policy, bool modify_all) -{ - struct dbs_data *dbs_data = policy->governor_data; - struct cs_dbs_tuners *cs_tuners = dbs_data->tuners; - - if (modify_all) - dbs_check_cpu(dbs_data, policy->cpu); - return delay_for_sampling_rate(cs_tuners->sampling_rate); + out: + return dbs_data->sampling_rate; } static int dbs_cpufreq_notifier(struct notifier_block *nb, unsigned long val, - void *data) -{ - struct cpufreq_freqs *freq = data; - struct cs_cpu_dbs_info_s *dbs_info = - &per_cpu(cs_cpu_dbs_info, freq->cpu); - struct cpufreq_policy *policy = cpufreq_cpu_get_raw(freq->cpu); - - if (!policy) - return 0; - - /* policy isn't governed by conservative governor */ - if (policy->governor != &cpufreq_gov_conservative) - return 0; - - /* - * we only care if our internally tracked freq moves outside the 'valid' - * ranges of frequency available to us otherwise we do not change it - */ - if (dbs_info->requested_freq > policy->max - || dbs_info->requested_freq < policy->min) - dbs_info->requested_freq = freq->new; - - return 0; -} + void *data); static struct notifier_block cs_cpufreq_notifier_block = { .notifier_call = dbs_cpufreq_notifier, }; /************************** sysfs interface ************************/ -static struct common_dbs_data cs_dbs_cdata; +static struct dbs_governor cs_dbs_gov; static ssize_t store_sampling_down_factor(struct dbs_data *dbs_data, const char *buf, size_t count) { - struct cs_dbs_tuners *cs_tuners = dbs_data->tuners; unsigned int input; int ret; ret = sscanf(buf, "%u", &input); @@ -170,22 +139,7 @@ static ssize_t store_sampling_down_factor(struct dbs_data *dbs_data, if (ret != 1 || input > MAX_SAMPLING_DOWN_FACTOR || input < 1) return -EINVAL; - cs_tuners->sampling_down_factor = input; - return count; -} - -static ssize_t store_sampling_rate(struct dbs_data *dbs_data, const char *buf, - size_t count) -{ - struct cs_dbs_tuners *cs_tuners = dbs_data->tuners; - unsigned int input; - int ret; - ret = sscanf(buf, "%u", &input); - - if (ret != 1) - return -EINVAL; - - cs_tuners->sampling_rate = max(input, dbs_data->min_sampling_rate); + dbs_data->sampling_down_factor = input; return count; } @@ -200,7 +154,7 @@ static ssize_t store_up_threshold(struct dbs_data *dbs_data, const char *buf, if (ret != 1 || input > 100 || input <= cs_tuners->down_threshold) return -EINVAL; - cs_tuners->up_threshold = input; + dbs_data->up_threshold = input; return count; } @@ -214,7 +168,7 @@ static ssize_t store_down_threshold(struct dbs_data *dbs_data, const char *buf, /* cannot be lower than 11 otherwise freq will not fall */ if (ret != 1 || input < 11 || input > 100 || - input >= cs_tuners->up_threshold) + input >= dbs_data->up_threshold) return -EINVAL; cs_tuners->down_threshold = input; @@ -224,8 +178,7 @@ static ssize_t store_down_threshold(struct dbs_data *dbs_data, const char *buf, static ssize_t store_ignore_nice_load(struct dbs_data *dbs_data, const char *buf, size_t count) { - struct cs_dbs_tuners *cs_tuners = dbs_data->tuners; - unsigned int input, j; + unsigned int input; int ret; ret = sscanf(buf, "%u", &input); @@ -235,21 +188,14 @@ static ssize_t store_ignore_nice_load(struct dbs_data *dbs_data, if (input > 1) input = 1; - if (input == cs_tuners->ignore_nice_load) /* nothing to do */ + if (input == dbs_data->ignore_nice_load) /* nothing to do */ return count; - cs_tuners->ignore_nice_load = input; + dbs_data->ignore_nice_load = input; /* we need to re-evaluate prev_cpu_idle */ - for_each_online_cpu(j) { - struct cs_cpu_dbs_info_s *dbs_info; - dbs_info = &per_cpu(cs_cpu_dbs_info, j); - dbs_info->cdbs.prev_cpu_idle = get_cpu_idle_time(j, - &dbs_info->cdbs.prev_cpu_wall, 0); - if (cs_tuners->ignore_nice_load) - dbs_info->cdbs.prev_cpu_nice = - kcpustat_cpu(j).cpustat[CPUTIME_NICE]; - } + gov_update_cpu_data(dbs_data); + return count; } @@ -275,55 +221,47 @@ static ssize_t store_freq_step(struct dbs_data *dbs_data, const char *buf, return count; } -show_store_one(cs, sampling_rate); -show_store_one(cs, sampling_down_factor); -show_store_one(cs, up_threshold); -show_store_one(cs, down_threshold); -show_store_one(cs, ignore_nice_load); -show_store_one(cs, freq_step); -declare_show_sampling_rate_min(cs); - -gov_sys_pol_attr_rw(sampling_rate); -gov_sys_pol_attr_rw(sampling_down_factor); -gov_sys_pol_attr_rw(up_threshold); -gov_sys_pol_attr_rw(down_threshold); -gov_sys_pol_attr_rw(ignore_nice_load); -gov_sys_pol_attr_rw(freq_step); -gov_sys_pol_attr_ro(sampling_rate_min); - -static struct attribute *dbs_attributes_gov_sys[] = { - &sampling_rate_min_gov_sys.attr, - &sampling_rate_gov_sys.attr, - &sampling_down_factor_gov_sys.attr, - &up_threshold_gov_sys.attr, - &down_threshold_gov_sys.attr, - &ignore_nice_load_gov_sys.attr, - &freq_step_gov_sys.attr, +gov_show_one_common(sampling_rate); +gov_show_one_common(sampling_down_factor); +gov_show_one_common(up_threshold); +gov_show_one_common(ignore_nice_load); +gov_show_one_common(min_sampling_rate); +gov_show_one(cs, down_threshold); +gov_show_one(cs, freq_step); + +gov_attr_rw(sampling_rate); +gov_attr_rw(sampling_down_factor); +gov_attr_rw(up_threshold); +gov_attr_rw(ignore_nice_load); +gov_attr_ro(min_sampling_rate); +gov_attr_rw(down_threshold); +gov_attr_rw(freq_step); + +static struct attribute *cs_attributes[] = { + &min_sampling_rate.attr, + &sampling_rate.attr, + &sampling_down_factor.attr, + &up_threshold.attr, + &down_threshold.attr, + &ignore_nice_load.attr, + &freq_step.attr, NULL }; -static struct attribute_group cs_attr_group_gov_sys = { - .attrs = dbs_attributes_gov_sys, - .name = "conservative", -}; +/************************** sysfs end ************************/ -static struct attribute *dbs_attributes_gov_pol[] = { - &sampling_rate_min_gov_pol.attr, - &sampling_rate_gov_pol.attr, - &sampling_down_factor_gov_pol.attr, - &up_threshold_gov_pol.attr, - &down_threshold_gov_pol.attr, - &ignore_nice_load_gov_pol.attr, - &freq_step_gov_pol.attr, - NULL -}; +static struct policy_dbs_info *cs_alloc(void) +{ + struct cs_policy_dbs_info *dbs_info; -static struct attribute_group cs_attr_group_gov_pol = { - .attrs = dbs_attributes_gov_pol, - .name = "conservative", -}; + dbs_info = kzalloc(sizeof(*dbs_info), GFP_KERNEL); + return dbs_info ? &dbs_info->policy_dbs : NULL; +} -/************************** sysfs end ************************/ +static void cs_free(struct policy_dbs_info *policy_dbs) +{ + kfree(to_dbs_info(policy_dbs)); +} static int cs_init(struct dbs_data *dbs_data, bool notify) { @@ -335,11 +273,11 @@ static int cs_init(struct dbs_data *dbs_data, bool notify) return -ENOMEM; } - tuners->up_threshold = DEF_FREQUENCY_UP_THRESHOLD; tuners->down_threshold = DEF_FREQUENCY_DOWN_THRESHOLD; - tuners->sampling_down_factor = DEF_SAMPLING_DOWN_FACTOR; - tuners->ignore_nice_load = 0; tuners->freq_step = DEF_FREQUENCY_STEP; + dbs_data->up_threshold = DEF_FREQUENCY_UP_THRESHOLD; + dbs_data->sampling_down_factor = DEF_SAMPLING_DOWN_FACTOR; + dbs_data->ignore_nice_load = 0; dbs_data->tuners = tuners; dbs_data->min_sampling_rate = MIN_SAMPLING_RATE_RATIO * @@ -361,35 +299,66 @@ static void cs_exit(struct dbs_data *dbs_data, bool notify) kfree(dbs_data->tuners); } -define_get_cpu_dbs_routines(cs_cpu_dbs_info); +static void cs_start(struct cpufreq_policy *policy) +{ + struct cs_policy_dbs_info *dbs_info = to_dbs_info(policy->governor_data); + + dbs_info->down_skip = 0; + dbs_info->requested_freq = policy->cur; +} -static struct common_dbs_data cs_dbs_cdata = { - .governor = GOV_CONSERVATIVE, - .attr_group_gov_sys = &cs_attr_group_gov_sys, - .attr_group_gov_pol = &cs_attr_group_gov_pol, - .get_cpu_cdbs = get_cpu_cdbs, - .get_cpu_dbs_info_s = get_cpu_dbs_info_s, +static struct dbs_governor cs_dbs_gov = { + .gov = { + .name = "conservative", + .governor = cpufreq_governor_dbs, + .max_transition_latency = TRANSITION_LATENCY_LIMIT, + .owner = THIS_MODULE, + }, + .kobj_type = { .default_attrs = cs_attributes }, .gov_dbs_timer = cs_dbs_timer, - .gov_check_cpu = cs_check_cpu, + .alloc = cs_alloc, + .free = cs_free, .init = cs_init, .exit = cs_exit, - .mutex = __MUTEX_INITIALIZER(cs_dbs_cdata.mutex), + .start = cs_start, }; -static int cs_cpufreq_governor_dbs(struct cpufreq_policy *policy, - unsigned int event) +#define CPU_FREQ_GOV_CONSERVATIVE (&cs_dbs_gov.gov) + +static int dbs_cpufreq_notifier(struct notifier_block *nb, unsigned long val, + void *data) { - return cpufreq_governor_dbs(policy, &cs_dbs_cdata, event); + struct cpufreq_freqs *freq = data; + struct cpufreq_policy *policy = cpufreq_cpu_get_raw(freq->cpu); + struct cs_policy_dbs_info *dbs_info; + + if (!policy) + return 0; + + /* policy isn't governed by conservative governor */ + if (policy->governor != CPU_FREQ_GOV_CONSERVATIVE) + return 0; + + dbs_info = to_dbs_info(policy->governor_data); + /* + * we only care if our internally tracked freq moves outside the 'valid' + * ranges of frequency available to us otherwise we do not change it + */ + if (dbs_info->requested_freq > policy->max + || dbs_info->requested_freq < policy->min) + dbs_info->requested_freq = freq->new; + + return 0; } static int __init cpufreq_gov_dbs_init(void) { - return cpufreq_register_governor(&cpufreq_gov_conservative); + return cpufreq_register_governor(CPU_FREQ_GOV_CONSERVATIVE); } static void __exit cpufreq_gov_dbs_exit(void) { - cpufreq_unregister_governor(&cpufreq_gov_conservative); + cpufreq_unregister_governor(CPU_FREQ_GOV_CONSERVATIVE); } MODULE_AUTHOR("Alexander Clouter <alex@digriz.org.uk>"); @@ -399,6 +368,11 @@ MODULE_DESCRIPTION("'cpufreq_conservative' - A dynamic cpufreq governor for " MODULE_LICENSE("GPL"); #ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE +struct cpufreq_governor *cpufreq_default_governor(void) +{ + return CPU_FREQ_GOV_CONSERVATIVE; +} + fs_initcall(cpufreq_gov_dbs_init); #else module_init(cpufreq_gov_dbs_init); diff --git a/drivers/cpufreq/cpufreq_governor.c b/drivers/cpufreq/cpufreq_governor.c index e0d1110..1c25ef4 100644 --- a/drivers/cpufreq/cpufreq_governor.c +++ b/drivers/cpufreq/cpufreq_governor.c @@ -18,95 +18,193 @@ #include <linux/export.h> #include <linux/kernel_stat.h> +#include <linux/sched.h> #include <linux/slab.h> #include "cpufreq_governor.h" -static struct attribute_group *get_sysfs_attr(struct dbs_data *dbs_data) -{ - if (have_governor_per_policy()) - return dbs_data->cdata->attr_group_gov_pol; - else - return dbs_data->cdata->attr_group_gov_sys; -} +static DEFINE_PER_CPU(struct cpu_dbs_info, cpu_dbs); + +static DEFINE_MUTEX(gov_dbs_data_mutex); -void dbs_check_cpu(struct dbs_data *dbs_data, int cpu) +/* Common sysfs tunables */ +/** + * store_sampling_rate - update sampling rate effective immediately if needed. + * + * If new rate is smaller than the old, simply updating + * dbs.sampling_rate might not be appropriate. For example, if the + * original sampling_rate was 1 second and the requested new sampling rate is 10 + * ms because the user needs immediate reaction from ondemand governor, but not + * sure if higher frequency will be required or not, then, the governor may + * change the sampling rate too late; up to 1 second later. Thus, if we are + * reducing the sampling rate, we need to make the new value effective + * immediately. + * + * This must be called with dbs_data->mutex held, otherwise traversing + * policy_dbs_list isn't safe. + */ +ssize_t store_sampling_rate(struct dbs_data *dbs_data, const char *buf, + size_t count) { - struct cpu_dbs_info *cdbs = dbs_data->cdata->get_cpu_cdbs(cpu); - struct od_dbs_tuners *od_tuners = dbs_data->tuners; - struct cs_dbs_tuners *cs_tuners = dbs_data->tuners; - struct cpufreq_policy *policy = cdbs->shared->policy; - unsigned int sampling_rate; - unsigned int max_load = 0; - unsigned int ignore_nice; - unsigned int j; + struct policy_dbs_info *policy_dbs; + unsigned int rate; + int ret; + ret = sscanf(buf, "%u", &rate); + if (ret != 1) + return -EINVAL; - if (dbs_data->cdata->governor == GOV_ONDEMAND) { - struct od_cpu_dbs_info_s *od_dbs_info = - dbs_data->cdata->get_cpu_dbs_info_s(cpu); + dbs_data->sampling_rate = max(rate, dbs_data->min_sampling_rate); + /* + * We are operating under dbs_data->mutex and so the list and its + * entries can't be freed concurrently. + */ + list_for_each_entry(policy_dbs, &dbs_data->policy_dbs_list, list) { + mutex_lock(&policy_dbs->timer_mutex); /* - * Sometimes, the ondemand governor uses an additional - * multiplier to give long delays. So apply this multiplier to - * the 'sampling_rate', so as to keep the wake-up-from-idle - * detection logic a bit conservative. + * On 32-bit architectures this may race with the + * sample_delay_ns read in dbs_update_util_handler(), but that + * really doesn't matter. If the read returns a value that's + * too big, the sample will be skipped, but the next invocation + * of dbs_update_util_handler() (when the update has been + * completed) will take a sample. + * + * If this runs in parallel with dbs_work_handler(), we may end + * up overwriting the sample_delay_ns value that it has just + * written, but it will be corrected next time a sample is + * taken, so it shouldn't be significant. */ - sampling_rate = od_tuners->sampling_rate; - sampling_rate *= od_dbs_info->rate_mult; + gov_update_sample_delay(policy_dbs, 0); + mutex_unlock(&policy_dbs->timer_mutex); + } - ignore_nice = od_tuners->ignore_nice_load; - } else { - sampling_rate = cs_tuners->sampling_rate; - ignore_nice = cs_tuners->ignore_nice_load; + return count; +} +EXPORT_SYMBOL_GPL(store_sampling_rate); + +/** + * gov_update_cpu_data - Update CPU load data. + * @dbs_data: Top-level governor data pointer. + * + * Update CPU load data for all CPUs in the domain governed by @dbs_data + * (that may be a single policy or a bunch of them if governor tunables are + * system-wide). + * + * Call under the @dbs_data mutex. + */ +void gov_update_cpu_data(struct dbs_data *dbs_data) +{ + struct policy_dbs_info *policy_dbs; + + list_for_each_entry(policy_dbs, &dbs_data->policy_dbs_list, list) { + unsigned int j; + + for_each_cpu(j, policy_dbs->policy->cpus) { + struct cpu_dbs_info *j_cdbs = &per_cpu(cpu_dbs, j); + + j_cdbs->prev_cpu_idle = get_cpu_idle_time(j, &j_cdbs->prev_cpu_wall, + dbs_data->io_is_busy); + if (dbs_data->ignore_nice_load) + j_cdbs->prev_cpu_nice = kcpustat_cpu(j).cpustat[CPUTIME_NICE]; + } } +} +EXPORT_SYMBOL_GPL(gov_update_cpu_data); + +static inline struct dbs_data *to_dbs_data(struct kobject *kobj) +{ + return container_of(kobj, struct dbs_data, kobj); +} + +static inline struct governor_attr *to_gov_attr(struct attribute *attr) +{ + return container_of(attr, struct governor_attr, attr); +} + +static ssize_t governor_show(struct kobject *kobj, struct attribute *attr, + char *buf) +{ + struct dbs_data *dbs_data = to_dbs_data(kobj); + struct governor_attr *gattr = to_gov_attr(attr); + + return gattr->show(dbs_data, buf); +} + +static ssize_t governor_store(struct kobject *kobj, struct attribute *attr, + const char *buf, size_t count) +{ + struct dbs_data *dbs_data = to_dbs_data(kobj); + struct governor_attr *gattr = to_gov_attr(attr); + int ret = -EBUSY; + + mutex_lock(&dbs_data->mutex); + + if (dbs_data->usage_count) + ret = gattr->store(dbs_data, buf, count); + + mutex_unlock(&dbs_data->mutex); + + return ret; +} + +/* + * Sysfs Ops for accessing governor attributes. + * + * All show/store invocations for governor specific sysfs attributes, will first + * call the below show/store callbacks and the attribute specific callback will + * be called from within it. + */ +static const struct sysfs_ops governor_sysfs_ops = { + .show = governor_show, + .store = governor_store, +}; + +unsigned int dbs_update(struct cpufreq_policy *policy) +{ + struct policy_dbs_info *policy_dbs = policy->governor_data; + struct dbs_data *dbs_data = policy_dbs->dbs_data; + unsigned int ignore_nice = dbs_data->ignore_nice_load; + unsigned int max_load = 0; + unsigned int sampling_rate, io_busy, j; + + /* + * Sometimes governors may use an additional multiplier to increase + * sample delays temporarily. Apply that multiplier to sampling_rate + * so as to keep the wake-up-from-idle detection logic a bit + * conservative. + */ + sampling_rate = dbs_data->sampling_rate * policy_dbs->rate_mult; + /* + * For the purpose of ondemand, waiting for disk IO is an indication + * that you're performance critical, and not that the system is actually + * idle, so do not add the iowait time to the CPU idle time then. + */ + io_busy = dbs_data->io_is_busy; /* Get Absolute Load */ for_each_cpu(j, policy->cpus) { - struct cpu_dbs_info *j_cdbs; + struct cpu_dbs_info *j_cdbs = &per_cpu(cpu_dbs, j); u64 cur_wall_time, cur_idle_time; unsigned int idle_time, wall_time; unsigned int load; - int io_busy = 0; - - j_cdbs = dbs_data->cdata->get_cpu_cdbs(j); - /* - * For the purpose of ondemand, waiting for disk IO is - * an indication that you're performance critical, and - * not that the system is actually idle. So do not add - * the iowait time to the cpu idle time. - */ - if (dbs_data->cdata->governor == GOV_ONDEMAND) - io_busy = od_tuners->io_is_busy; cur_idle_time = get_cpu_idle_time(j, &cur_wall_time, io_busy); - wall_time = (unsigned int) - (cur_wall_time - j_cdbs->prev_cpu_wall); + wall_time = cur_wall_time - j_cdbs->prev_cpu_wall; j_cdbs->prev_cpu_wall = cur_wall_time; - if (cur_idle_time < j_cdbs->prev_cpu_idle) - cur_idle_time = j_cdbs->prev_cpu_idle; - - idle_time = (unsigned int) - (cur_idle_time - j_cdbs->prev_cpu_idle); - j_cdbs->prev_cpu_idle = cur_idle_time; + if (cur_idle_time <= j_cdbs->prev_cpu_idle) { + idle_time = 0; + } else { + idle_time = cur_idle_time - j_cdbs->prev_cpu_idle; + j_cdbs->prev_cpu_idle = cur_idle_time; + } if (ignore_nice) { - u64 cur_nice; - unsigned long cur_nice_jiffies; - - cur_nice = kcpustat_cpu(j).cpustat[CPUTIME_NICE] - - cdbs->prev_cpu_nice; - /* - * Assumption: nice time between sampling periods will - * be less than 2^32 jiffies for 32 bit sys - */ - cur_nice_jiffies = (unsigned long) - cputime64_to_jiffies64(cur_nice); + u64 cur_nice = kcpustat_cpu(j).cpustat[CPUTIME_NICE]; - cdbs->prev_cpu_nice = - kcpustat_cpu(j).cpustat[CPUTIME_NICE]; - idle_time += jiffies_to_usecs(cur_nice_jiffies); + idle_time += cputime_to_usecs(cur_nice - j_cdbs->prev_cpu_nice); + j_cdbs->prev_cpu_nice = cur_nice; } if (unlikely(!wall_time || wall_time < idle_time)) @@ -128,10 +226,10 @@ void dbs_check_cpu(struct dbs_data *dbs_data, int cpu) * dropped down. So we perform the copy only once, upon the * first wake-up from idle.) * - * Detecting this situation is easy: the governor's deferrable - * timer would not have fired during CPU-idle periods. Hence - * an unusually large 'wall_time' (as compared to the sampling - * rate) indicates this scenario. + * Detecting this situation is easy: the governor's utilization + * update handler would not have run during CPU-idle periods. + * Hence, an unusually large 'wall_time' (as compared to the + * sampling rate) indicates this scenario. * * prev_load can be zero in two cases and we must recalculate it * for both cases: @@ -156,222 +254,224 @@ void dbs_check_cpu(struct dbs_data *dbs_data, int cpu) if (load > max_load) max_load = load; } - - dbs_data->cdata->gov_check_cpu(cpu, max_load); + return max_load; } -EXPORT_SYMBOL_GPL(dbs_check_cpu); +EXPORT_SYMBOL_GPL(dbs_update); -void gov_add_timers(struct cpufreq_policy *policy, unsigned int delay) +static void gov_set_update_util(struct policy_dbs_info *policy_dbs, + unsigned int delay_us) { - struct dbs_data *dbs_data = policy->governor_data; - struct cpu_dbs_info *cdbs; + struct cpufreq_policy *policy = policy_dbs->policy; int cpu; + gov_update_sample_delay(policy_dbs, delay_us); + policy_dbs->last_sample_time = 0; + for_each_cpu(cpu, policy->cpus) { - cdbs = dbs_data->cdata->get_cpu_cdbs(cpu); - cdbs->timer.expires = jiffies + delay; - add_timer_on(&cdbs->timer, cpu); + struct cpu_dbs_info *cdbs = &per_cpu(cpu_dbs, cpu); + + cpufreq_set_update_util_data(cpu, &cdbs->update_util); } } -EXPORT_SYMBOL_GPL(gov_add_timers); -static inline void gov_cancel_timers(struct cpufreq_policy *policy) +static inline void gov_clear_update_util(struct cpufreq_policy *policy) { - struct dbs_data *dbs_data = policy->governor_data; - struct cpu_dbs_info *cdbs; int i; - for_each_cpu(i, policy->cpus) { - cdbs = dbs_data->cdata->get_cpu_cdbs(i); - del_timer_sync(&cdbs->timer); - } -} + for_each_cpu(i, policy->cpus) + cpufreq_set_update_util_data(i, NULL); -void gov_cancel_work(struct cpu_common_dbs_info *shared) -{ - /* Tell dbs_timer_handler() to skip queuing up work items. */ - atomic_inc(&shared->skip_work); - /* - * If dbs_timer_handler() is already running, it may not notice the - * incremented skip_work, so wait for it to complete to prevent its work - * item from being queued up after the cancel_work_sync() below. - */ - gov_cancel_timers(shared->policy); - /* - * In case dbs_timer_handler() managed to run and spawn a work item - * before the timers have been canceled, wait for that work item to - * complete and then cancel all of the timers set up by it. If - * dbs_timer_handler() runs again at that point, it will see the - * positive value of skip_work and won't spawn any more work items. - */ - cancel_work_sync(&shared->work); - gov_cancel_timers(shared->policy); - atomic_set(&shared->skip_work, 0); + synchronize_sched(); } -EXPORT_SYMBOL_GPL(gov_cancel_work); -/* Will return if we need to evaluate cpu load again or not */ -static bool need_load_eval(struct cpu_common_dbs_info *shared, - unsigned int sampling_rate) +static void gov_cancel_work(struct cpufreq_policy *policy) { - if (policy_is_shared(shared->policy)) { - ktime_t time_now = ktime_get(); - s64 delta_us = ktime_us_delta(time_now, shared->time_stamp); - - /* Do nothing if we recently have sampled */ - if (delta_us < (s64)(sampling_rate / 2)) - return false; - else - shared->time_stamp = time_now; - } + struct policy_dbs_info *policy_dbs = policy->governor_data; - return true; + gov_clear_update_util(policy_dbs->policy); + irq_work_sync(&policy_dbs->irq_work); + cancel_work_sync(&policy_dbs->work); + atomic_set(&policy_dbs->work_count, 0); + policy_dbs->work_in_progress = false; } static void dbs_work_handler(struct work_struct *work) { - struct cpu_common_dbs_info *shared = container_of(work, struct - cpu_common_dbs_info, work); + struct policy_dbs_info *policy_dbs; struct cpufreq_policy *policy; - struct dbs_data *dbs_data; - unsigned int sampling_rate, delay; - bool eval_load; - - policy = shared->policy; - dbs_data = policy->governor_data; + struct dbs_governor *gov; - /* Kill all timers */ - gov_cancel_timers(policy); + policy_dbs = container_of(work, struct policy_dbs_info, work); + policy = policy_dbs->policy; + gov = dbs_governor_of(policy); - if (dbs_data->cdata->governor == GOV_CONSERVATIVE) { - struct cs_dbs_tuners *cs_tuners = dbs_data->tuners; - - sampling_rate = cs_tuners->sampling_rate; - } else { - struct od_dbs_tuners *od_tuners = dbs_data->tuners; - - sampling_rate = od_tuners->sampling_rate; - } - - eval_load = need_load_eval(shared, sampling_rate); + /* + * Make sure cpufreq_governor_limits() isn't evaluating load or the + * ondemand governor isn't updating the sampling rate in parallel. + */ + mutex_lock(&policy_dbs->timer_mutex); + gov_update_sample_delay(policy_dbs, gov->gov_dbs_timer(policy)); + mutex_unlock(&policy_dbs->timer_mutex); + /* Allow the utilization update handler to queue up more work. */ + atomic_set(&policy_dbs->work_count, 0); /* - * Make sure cpufreq_governor_limits() isn't evaluating load in - * parallel. + * If the update below is reordered with respect to the sample delay + * modification, the utilization update handler may end up using a stale + * sample delay value. */ - mutex_lock(&shared->timer_mutex); - delay = dbs_data->cdata->gov_dbs_timer(policy, eval_load); - mutex_unlock(&shared->timer_mutex); + smp_wmb(); + policy_dbs->work_in_progress = false; +} - atomic_dec(&shared->skip_work); +static void dbs_irq_work(struct irq_work *irq_work) +{ + struct policy_dbs_info *policy_dbs; - gov_add_timers(policy, delay); + policy_dbs = container_of(irq_work, struct policy_dbs_info, irq_work); + schedule_work(&policy_dbs->work); } -static void dbs_timer_handler(unsigned long data) +static void dbs_update_util_handler(struct update_util_data *data, u64 time, + unsigned long util, unsigned long max) { - struct cpu_dbs_info *cdbs = (struct cpu_dbs_info *)data; - struct cpu_common_dbs_info *shared = cdbs->shared; + struct cpu_dbs_info *cdbs = container_of(data, struct cpu_dbs_info, update_util); + struct policy_dbs_info *policy_dbs = cdbs->policy_dbs; + u64 delta_ns, lst; /* - * Timer handler may not be allowed to queue the work at the moment, - * because: - * - Another timer handler has done that - * - We are stopping the governor - * - Or we are updating the sampling rate of the ondemand governor + * The work may not be allowed to be queued up right now. + * Possible reasons: + * - Work has already been queued up or is in progress. + * - It is too early (too little time from the previous sample). */ - if (atomic_inc_return(&shared->skip_work) > 1) - atomic_dec(&shared->skip_work); - else - queue_work(system_wq, &shared->work); -} + if (policy_dbs->work_in_progress) + return; -static void set_sampling_rate(struct dbs_data *dbs_data, - unsigned int sampling_rate) -{ - if (dbs_data->cdata->governor == GOV_CONSERVATIVE) { - struct cs_dbs_tuners *cs_tuners = dbs_data->tuners; - cs_tuners->sampling_rate = sampling_rate; - } else { - struct od_dbs_tuners *od_tuners = dbs_data->tuners; - od_tuners->sampling_rate = sampling_rate; + /* + * If the reads below are reordered before the check above, the value + * of sample_delay_ns used in the computation may be stale. + */ + smp_rmb(); + lst = READ_ONCE(policy_dbs->last_sample_time); + delta_ns = time - lst; + if ((s64)delta_ns < policy_dbs->sample_delay_ns) + return; + + /* + * If the policy is not shared, the irq_work may be queued up right away + * at this point. Otherwise, we need to ensure that only one of the + * CPUs sharing the policy will do that. + */ + if (policy_dbs->is_shared) { + if (!atomic_add_unless(&policy_dbs->work_count, 1, 1)) + return; + + /* + * If another CPU updated last_sample_time in the meantime, we + * shouldn't be here, so clear the work counter and bail out. + */ + if (unlikely(lst != READ_ONCE(policy_dbs->last_sample_time))) { + atomic_set(&policy_dbs->work_count, 0); + return; + } } + + policy_dbs->last_sample_time = time; + policy_dbs->work_in_progress = true; + irq_work_queue(&policy_dbs->irq_work); } -static int alloc_common_dbs_info(struct cpufreq_policy *policy, - struct common_dbs_data *cdata) +static struct policy_dbs_info *alloc_policy_dbs_info(struct cpufreq_policy *policy, + struct dbs_governor *gov) { - struct cpu_common_dbs_info *shared; + struct policy_dbs_info *policy_dbs; int j; - /* Allocate memory for the common information for policy->cpus */ - shared = kzalloc(sizeof(*shared), GFP_KERNEL); - if (!shared) - return -ENOMEM; + /* Allocate memory for per-policy governor data. */ + policy_dbs = gov->alloc(); + if (!policy_dbs) + return NULL; - /* Set shared for all CPUs, online+offline */ - for_each_cpu(j, policy->related_cpus) - cdata->get_cpu_cdbs(j)->shared = shared; + policy_dbs->policy = policy; + mutex_init(&policy_dbs->timer_mutex); + atomic_set(&policy_dbs->work_count, 0); + init_irq_work(&policy_dbs->irq_work, dbs_irq_work); + INIT_WORK(&policy_dbs->work, dbs_work_handler); - mutex_init(&shared->timer_mutex); - atomic_set(&shared->skip_work, 0); - INIT_WORK(&shared->work, dbs_work_handler); - return 0; + /* Set policy_dbs for all CPUs, online+offline */ + for_each_cpu(j, policy->related_cpus) { + struct cpu_dbs_info *j_cdbs = &per_cpu(cpu_dbs, j); + + j_cdbs->policy_dbs = policy_dbs; + j_cdbs->update_util.func = dbs_update_util_handler; + } + return policy_dbs; } -static void free_common_dbs_info(struct cpufreq_policy *policy, - struct common_dbs_data *cdata) +static void free_policy_dbs_info(struct policy_dbs_info *policy_dbs, + struct dbs_governor *gov) { - struct cpu_dbs_info *cdbs = cdata->get_cpu_cdbs(policy->cpu); - struct cpu_common_dbs_info *shared = cdbs->shared; int j; - mutex_destroy(&shared->timer_mutex); + mutex_destroy(&policy_dbs->timer_mutex); - for_each_cpu(j, policy->cpus) - cdata->get_cpu_cdbs(j)->shared = NULL; + for_each_cpu(j, policy_dbs->policy->related_cpus) { + struct cpu_dbs_info *j_cdbs = &per_cpu(cpu_dbs, j); - kfree(shared); + j_cdbs->policy_dbs = NULL; + j_cdbs->update_util.func = NULL; + } + gov->free(policy_dbs); } -static int cpufreq_governor_init(struct cpufreq_policy *policy, - struct dbs_data *dbs_data, - struct common_dbs_data *cdata) +static int cpufreq_governor_init(struct cpufreq_policy *policy) { + struct dbs_governor *gov = dbs_governor_of(policy); + struct dbs_data *dbs_data; + struct policy_dbs_info *policy_dbs; unsigned int latency; - int ret; + int ret = 0; /* State should be equivalent to EXIT */ if (policy->governor_data) return -EBUSY; - if (dbs_data) { - if (WARN_ON(have_governor_per_policy())) - return -EINVAL; + policy_dbs = alloc_policy_dbs_info(policy, gov); + if (!policy_dbs) + return -ENOMEM; - ret = alloc_common_dbs_info(policy, cdata); - if (ret) - return ret; + /* Protect gov->gdbs_data against concurrent updates. */ + mutex_lock(&gov_dbs_data_mutex); + dbs_data = gov->gdbs_data; + if (dbs_data) { + if (WARN_ON(have_governor_per_policy())) { + ret = -EINVAL; + goto free_policy_dbs_info; + } + policy_dbs->dbs_data = dbs_data; + policy->governor_data = policy_dbs; + + mutex_lock(&dbs_data->mutex); dbs_data->usage_count++; - policy->governor_data = dbs_data; - return 0; + list_add(&policy_dbs->list, &dbs_data->policy_dbs_list); + mutex_unlock(&dbs_data->mutex); + goto out; } dbs_data = kzalloc(sizeof(*dbs_data), GFP_KERNEL); - if (!dbs_data) - return -ENOMEM; - - ret = alloc_common_dbs_info(policy, cdata); - if (ret) - goto free_dbs_data; + if (!dbs_data) { + ret = -ENOMEM; + goto free_policy_dbs_info; + } - dbs_data->cdata = cdata; - dbs_data->usage_count = 1; + INIT_LIST_HEAD(&dbs_data->policy_dbs_list); + mutex_init(&dbs_data->mutex); - ret = cdata->init(dbs_data, !policy->governor->initialized); + ret = gov->init(dbs_data, !policy->governor->initialized); if (ret) - goto free_common_dbs_info; + goto free_policy_dbs_info; /* policy latency is in ns. Convert it to us first */ latency = policy->cpuinfo.transition_latency / 1000; @@ -381,216 +481,156 @@ static int cpufreq_governor_init(struct cpufreq_policy *policy, /* Bring kernel and HW constraints together */ dbs_data->min_sampling_rate = max(dbs_data->min_sampling_rate, MIN_LATENCY_MULTIPLIER * latency); - set_sampling_rate(dbs_data, max(dbs_data->min_sampling_rate, - latency * LATENCY_MULTIPLIER)); + dbs_data->sampling_rate = max(dbs_data->min_sampling_rate, + LATENCY_MULTIPLIER * latency); if (!have_governor_per_policy()) - cdata->gdbs_data = dbs_data; + gov->gdbs_data = dbs_data; - policy->governor_data = dbs_data; + policy->governor_data = policy_dbs; - ret = sysfs_create_group(get_governor_parent_kobj(policy), - get_sysfs_attr(dbs_data)); - if (ret) - goto reset_gdbs_data; + policy_dbs->dbs_data = dbs_data; + dbs_data->usage_count = 1; + list_add(&policy_dbs->list, &dbs_data->policy_dbs_list); - return 0; + gov->kobj_type.sysfs_ops = &governor_sysfs_ops; + ret = kobject_init_and_add(&dbs_data->kobj, &gov->kobj_type, + get_governor_parent_kobj(policy), + "%s", gov->gov.name); + if (!ret) + goto out; + + /* Failure, so roll back. */ + pr_err("cpufreq: Governor initialization failed (dbs_data kobject init error %d)\n", ret); -reset_gdbs_data: policy->governor_data = NULL; if (!have_governor_per_policy()) - cdata->gdbs_data = NULL; - cdata->exit(dbs_data, !policy->governor->initialized); -free_common_dbs_info: - free_common_dbs_info(policy, cdata); -free_dbs_data: + gov->gdbs_data = NULL; + gov->exit(dbs_data, !policy->governor->initialized); kfree(dbs_data); + +free_policy_dbs_info: + free_policy_dbs_info(policy_dbs, gov); + +out: + mutex_unlock(&gov_dbs_data_mutex); return ret; } -static int cpufreq_governor_exit(struct cpufreq_policy *policy, - struct dbs_data *dbs_data) +static int cpufreq_governor_exit(struct cpufreq_policy *policy) { - struct common_dbs_data *cdata = dbs_data->cdata; - struct cpu_dbs_info *cdbs = cdata->get_cpu_cdbs(policy->cpu); + struct dbs_governor *gov = dbs_governor_of(policy); + struct policy_dbs_info *policy_dbs = policy->governor_data; + struct dbs_data *dbs_data = policy_dbs->dbs_data; + int count; - /* State should be equivalent to INIT */ - if (!cdbs->shared || cdbs->shared->policy) - return -EBUSY; + /* Protect gov->gdbs_data against concurrent updates. */ + mutex_lock(&gov_dbs_data_mutex); + + mutex_lock(&dbs_data->mutex); + list_del(&policy_dbs->list); + count = --dbs_data->usage_count; + mutex_unlock(&dbs_data->mutex); - if (!--dbs_data->usage_count) { - sysfs_remove_group(get_governor_parent_kobj(policy), - get_sysfs_attr(dbs_data)); + if (!count) { + kobject_put(&dbs_data->kobj); policy->governor_data = NULL; if (!have_governor_per_policy()) - cdata->gdbs_data = NULL; + gov->gdbs_data = NULL; - cdata->exit(dbs_data, policy->governor->initialized == 1); + gov->exit(dbs_data, policy->governor->initialized == 1); + mutex_destroy(&dbs_data->mutex); kfree(dbs_data); } else { policy->governor_data = NULL; } - free_common_dbs_info(policy, cdata); + free_policy_dbs_info(policy_dbs, gov); + + mutex_unlock(&gov_dbs_data_mutex); return 0; } -static int cpufreq_governor_start(struct cpufreq_policy *policy, - struct dbs_data *dbs_data) +static int cpufreq_governor_start(struct cpufreq_policy *policy) { - struct common_dbs_data *cdata = dbs_data->cdata; - unsigned int sampling_rate, ignore_nice, j, cpu = policy->cpu; - struct cpu_dbs_info *cdbs = cdata->get_cpu_cdbs(cpu); - struct cpu_common_dbs_info *shared = cdbs->shared; - int io_busy = 0; + struct dbs_governor *gov = dbs_governor_of(policy); + struct policy_dbs_info *policy_dbs = policy->governor_data; + struct dbs_data *dbs_data = policy_dbs->dbs_data; + unsigned int sampling_rate, ignore_nice, j; + unsigned int io_busy; if (!policy->cur) return -EINVAL; - /* State should be equivalent to INIT */ - if (!shared || shared->policy) - return -EBUSY; + policy_dbs->is_shared = policy_is_shared(policy); + policy_dbs->rate_mult = 1; - if (cdata->governor == GOV_CONSERVATIVE) { - struct cs_dbs_tuners *cs_tuners = dbs_data->tuners; - - sampling_rate = cs_tuners->sampling_rate; - ignore_nice = cs_tuners->ignore_nice_load; - } else { - struct od_dbs_tuners *od_tuners = dbs_data->tuners; - - sampling_rate = od_tuners->sampling_rate; - ignore_nice = od_tuners->ignore_nice_load; - io_busy = od_tuners->io_is_busy; - } - - shared->policy = policy; - shared->time_stamp = ktime_get(); + sampling_rate = dbs_data->sampling_rate; + ignore_nice = dbs_data->ignore_nice_load; + io_busy = dbs_data->io_is_busy; for_each_cpu(j, policy->cpus) { - struct cpu_dbs_info *j_cdbs = cdata->get_cpu_cdbs(j); + struct cpu_dbs_info *j_cdbs = &per_cpu(cpu_dbs, j); unsigned int prev_load; - j_cdbs->prev_cpu_idle = - get_cpu_idle_time(j, &j_cdbs->prev_cpu_wall, io_busy); + j_cdbs->prev_cpu_idle = get_cpu_idle_time(j, &j_cdbs->prev_cpu_wall, io_busy); - prev_load = (unsigned int)(j_cdbs->prev_cpu_wall - - j_cdbs->prev_cpu_idle); - j_cdbs->prev_load = 100 * prev_load / - (unsigned int)j_cdbs->prev_cpu_wall; + prev_load = j_cdbs->prev_cpu_wall - j_cdbs->prev_cpu_idle; + j_cdbs->prev_load = 100 * prev_load / (unsigned int)j_cdbs->prev_cpu_wall; if (ignore_nice) j_cdbs->prev_cpu_nice = kcpustat_cpu(j).cpustat[CPUTIME_NICE]; - - __setup_timer(&j_cdbs->timer, dbs_timer_handler, - (unsigned long)j_cdbs, - TIMER_DEFERRABLE | TIMER_IRQSAFE); } - if (cdata->governor == GOV_CONSERVATIVE) { - struct cs_cpu_dbs_info_s *cs_dbs_info = - cdata->get_cpu_dbs_info_s(cpu); - - cs_dbs_info->down_skip = 0; - cs_dbs_info->requested_freq = policy->cur; - } else { - struct od_ops *od_ops = cdata->gov_ops; - struct od_cpu_dbs_info_s *od_dbs_info = cdata->get_cpu_dbs_info_s(cpu); - - od_dbs_info->rate_mult = 1; - od_dbs_info->sample_type = OD_NORMAL_SAMPLE; - od_ops->powersave_bias_init_cpu(cpu); - } + gov->start(policy); - gov_add_timers(policy, delay_for_sampling_rate(sampling_rate)); + gov_set_update_util(policy_dbs, sampling_rate); return 0; } -static int cpufreq_governor_stop(struct cpufreq_policy *policy, - struct dbs_data *dbs_data) +static int cpufreq_governor_stop(struct cpufreq_policy *policy) { - struct cpu_dbs_info *cdbs = dbs_data->cdata->get_cpu_cdbs(policy->cpu); - struct cpu_common_dbs_info *shared = cdbs->shared; - - /* State should be equivalent to START */ - if (!shared || !shared->policy) - return -EBUSY; - - gov_cancel_work(shared); - shared->policy = NULL; - + gov_cancel_work(policy); return 0; } -static int cpufreq_governor_limits(struct cpufreq_policy *policy, - struct dbs_data *dbs_data) +static int cpufreq_governor_limits(struct cpufreq_policy *policy) { - struct common_dbs_data *cdata = dbs_data->cdata; - unsigned int cpu = policy->cpu; - struct cpu_dbs_info *cdbs = cdata->get_cpu_cdbs(cpu); + struct policy_dbs_info *policy_dbs = policy->governor_data; - /* State should be equivalent to START */ - if (!cdbs->shared || !cdbs->shared->policy) - return -EBUSY; + mutex_lock(&policy_dbs->timer_mutex); + + if (policy->max < policy->cur) + __cpufreq_driver_target(policy, policy->max, CPUFREQ_RELATION_H); + else if (policy->min > policy->cur) + __cpufreq_driver_target(policy, policy->min, CPUFREQ_RELATION_L); + + gov_update_sample_delay(policy_dbs, 0); - mutex_lock(&cdbs->shared->timer_mutex); - if (policy->max < cdbs->shared->policy->cur) - __cpufreq_driver_target(cdbs->shared->policy, policy->max, - CPUFREQ_RELATION_H); - else if (policy->min > cdbs->shared->policy->cur) - __cpufreq_driver_target(cdbs->shared->policy, policy->min, - CPUFREQ_RELATION_L); - dbs_check_cpu(dbs_data, cpu); - mutex_unlock(&cdbs->shared->timer_mutex); + mutex_unlock(&policy_dbs->timer_mutex); return 0; } -int cpufreq_governor_dbs(struct cpufreq_policy *policy, - struct common_dbs_data *cdata, unsigned int event) +int cpufreq_governor_dbs(struct cpufreq_policy *policy, unsigned int event) { - struct dbs_data *dbs_data; - int ret; - - /* Lock governor to block concurrent initialization of governor */ - mutex_lock(&cdata->mutex); - - if (have_governor_per_policy()) - dbs_data = policy->governor_data; - else - dbs_data = cdata->gdbs_data; - - if (!dbs_data && (event != CPUFREQ_GOV_POLICY_INIT)) { - ret = -EINVAL; - goto unlock; - } - - switch (event) { - case CPUFREQ_GOV_POLICY_INIT: - ret = cpufreq_governor_init(policy, dbs_data, cdata); - break; - case CPUFREQ_GOV_POLICY_EXIT: - ret = cpufreq_governor_exit(policy, dbs_data); - break; - case CPUFREQ_GOV_START: - ret = cpufreq_governor_start(policy, dbs_data); - break; - case CPUFREQ_GOV_STOP: - ret = cpufreq_governor_stop(policy, dbs_data); - break; - case CPUFREQ_GOV_LIMITS: - ret = cpufreq_governor_limits(policy, dbs_data); - break; - default: - ret = -EINVAL; + if (event == CPUFREQ_GOV_POLICY_INIT) { + return cpufreq_governor_init(policy); + } else if (policy->governor_data) { + switch (event) { + case CPUFREQ_GOV_POLICY_EXIT: + return cpufreq_governor_exit(policy); + case CPUFREQ_GOV_START: + return cpufreq_governor_start(policy); + case CPUFREQ_GOV_STOP: + return cpufreq_governor_stop(policy); + case CPUFREQ_GOV_LIMITS: + return cpufreq_governor_limits(policy); + } } - -unlock: - mutex_unlock(&cdata->mutex); - - return ret; + return -EINVAL; } EXPORT_SYMBOL_GPL(cpufreq_governor_dbs); diff --git a/drivers/cpufreq/cpufreq_governor.h b/drivers/cpufreq/cpufreq_governor.h index 91e767a..61ff82f 100644 --- a/drivers/cpufreq/cpufreq_governor.h +++ b/drivers/cpufreq/cpufreq_governor.h @@ -18,6 +18,7 @@ #define _CPUFREQ_GOVERNOR_H #include <linux/atomic.h> +#include <linux/irq_work.h> #include <linux/cpufreq.h> #include <linux/kernel_stat.h> #include <linux/module.h> @@ -41,96 +42,68 @@ enum {OD_NORMAL_SAMPLE, OD_SUB_SAMPLE}; /* - * Macro for creating governors sysfs routines - * - * - gov_sys: One governor instance per whole system - * - gov_pol: One governor instance per policy + * Abbreviations: + * dbs: used as a shortform for demand based switching It helps to keep variable + * names smaller, simpler + * cdbs: common dbs + * od_*: On-demand governor + * cs_*: Conservative governor */ -/* Create attributes */ -#define gov_sys_attr_ro(_name) \ -static struct global_attr _name##_gov_sys = \ -__ATTR(_name, 0444, show_##_name##_gov_sys, NULL) - -#define gov_sys_attr_rw(_name) \ -static struct global_attr _name##_gov_sys = \ -__ATTR(_name, 0644, show_##_name##_gov_sys, store_##_name##_gov_sys) - -#define gov_pol_attr_ro(_name) \ -static struct freq_attr _name##_gov_pol = \ -__ATTR(_name, 0444, show_##_name##_gov_pol, NULL) - -#define gov_pol_attr_rw(_name) \ -static struct freq_attr _name##_gov_pol = \ -__ATTR(_name, 0644, show_##_name##_gov_pol, store_##_name##_gov_pol) +/* Governor demand based switching data (per-policy or global). */ +struct dbs_data { + int usage_count; + void *tuners; + unsigned int min_sampling_rate; + unsigned int ignore_nice_load; + unsigned int sampling_rate; + unsigned int sampling_down_factor; + unsigned int up_threshold; + unsigned int io_is_busy; -#define gov_sys_pol_attr_rw(_name) \ - gov_sys_attr_rw(_name); \ - gov_pol_attr_rw(_name) + struct kobject kobj; + struct list_head policy_dbs_list; + /* + * Protect concurrent updates to governor tunables from sysfs, + * policy_dbs_list and usage_count. + */ + struct mutex mutex; +}; -#define gov_sys_pol_attr_ro(_name) \ - gov_sys_attr_ro(_name); \ - gov_pol_attr_ro(_name) +/* Governor's specific attributes */ +struct dbs_data; +struct governor_attr { + struct attribute attr; + ssize_t (*show)(struct dbs_data *dbs_data, char *buf); + ssize_t (*store)(struct dbs_data *dbs_data, const char *buf, + size_t count); +}; -/* Create show/store routines */ -#define show_one(_gov, file_name) \ -static ssize_t show_##file_name##_gov_sys \ -(struct kobject *kobj, struct attribute *attr, char *buf) \ +#define gov_show_one(_gov, file_name) \ +static ssize_t show_##file_name \ +(struct dbs_data *dbs_data, char *buf) \ { \ - struct _gov##_dbs_tuners *tuners = _gov##_dbs_cdata.gdbs_data->tuners; \ - return sprintf(buf, "%u\n", tuners->file_name); \ -} \ - \ -static ssize_t show_##file_name##_gov_pol \ -(struct cpufreq_policy *policy, char *buf) \ -{ \ - struct dbs_data *dbs_data = policy->governor_data; \ struct _gov##_dbs_tuners *tuners = dbs_data->tuners; \ return sprintf(buf, "%u\n", tuners->file_name); \ } -#define store_one(_gov, file_name) \ -static ssize_t store_##file_name##_gov_sys \ -(struct kobject *kobj, struct attribute *attr, const char *buf, size_t count) \ -{ \ - struct dbs_data *dbs_data = _gov##_dbs_cdata.gdbs_data; \ - return store_##file_name(dbs_data, buf, count); \ -} \ - \ -static ssize_t store_##file_name##_gov_pol \ -(struct cpufreq_policy *policy, const char *buf, size_t count) \ +#define gov_show_one_common(file_name) \ +static ssize_t show_##file_name \ +(struct dbs_data *dbs_data, char *buf) \ { \ - struct dbs_data *dbs_data = policy->governor_data; \ - return store_##file_name(dbs_data, buf, count); \ + return sprintf(buf, "%u\n", dbs_data->file_name); \ } -#define show_store_one(_gov, file_name) \ -show_one(_gov, file_name); \ -store_one(_gov, file_name) +#define gov_attr_ro(_name) \ +static struct governor_attr _name = \ +__ATTR(_name, 0444, show_##_name, NULL) -/* create helper routines */ -#define define_get_cpu_dbs_routines(_dbs_info) \ -static struct cpu_dbs_info *get_cpu_cdbs(int cpu) \ -{ \ - return &per_cpu(_dbs_info, cpu).cdbs; \ -} \ - \ -static void *get_cpu_dbs_info_s(int cpu) \ -{ \ - return &per_cpu(_dbs_info, cpu); \ -} - -/* - * Abbreviations: - * dbs: used as a shortform for demand based switching It helps to keep variable - * names smaller, simpler - * cdbs: common dbs - * od_*: On-demand governor - * cs_*: Conservative governor - */ +#define gov_attr_rw(_name) \ +static struct governor_attr _name = \ +__ATTR(_name, 0644, show_##_name, store_##_name) /* Common to all CPUs of a policy */ -struct cpu_common_dbs_info { +struct policy_dbs_info { struct cpufreq_policy *policy; /* * Per policy mutex that serializes load evaluation from limit-change @@ -138,11 +111,27 @@ struct cpu_common_dbs_info { */ struct mutex timer_mutex; - ktime_t time_stamp; - atomic_t skip_work; + u64 last_sample_time; + s64 sample_delay_ns; + atomic_t work_count; + struct irq_work irq_work; struct work_struct work; + /* dbs_data may be shared between multiple policy objects */ + struct dbs_data *dbs_data; + struct list_head list; + /* Multiplier for increasing sample delay temporarily. */ + unsigned int rate_mult; + /* Status indicators */ + bool is_shared; /* This object is used by multiple CPUs */ + bool work_in_progress; /* Work is being queued up or in progress */ }; +static inline void gov_update_sample_delay(struct policy_dbs_info *policy_dbs, + unsigned int delay_us) +{ + policy_dbs->sample_delay_ns = delay_us * NSEC_PER_USEC; +} + /* Per cpu structures */ struct cpu_dbs_info { u64 prev_cpu_idle; @@ -155,54 +144,14 @@ struct cpu_dbs_info { * wake-up from idle. */ unsigned int prev_load; - struct timer_list timer; - struct cpu_common_dbs_info *shared; -}; - -struct od_cpu_dbs_info_s { - struct cpu_dbs_info cdbs; - struct cpufreq_frequency_table *freq_table; - unsigned int freq_lo; - unsigned int freq_lo_jiffies; - unsigned int freq_hi_jiffies; - unsigned int rate_mult; - unsigned int sample_type:1; -}; - -struct cs_cpu_dbs_info_s { - struct cpu_dbs_info cdbs; - unsigned int down_skip; - unsigned int requested_freq; -}; - -/* Per policy Governors sysfs tunables */ -struct od_dbs_tuners { - unsigned int ignore_nice_load; - unsigned int sampling_rate; - unsigned int sampling_down_factor; - unsigned int up_threshold; - unsigned int powersave_bias; - unsigned int io_is_busy; -}; - -struct cs_dbs_tuners { - unsigned int ignore_nice_load; - unsigned int sampling_rate; - unsigned int sampling_down_factor; - unsigned int up_threshold; - unsigned int down_threshold; - unsigned int freq_step; + struct update_util_data update_util; + struct policy_dbs_info *policy_dbs; }; /* Common Governor data across policies */ -struct dbs_data; -struct common_dbs_data { - /* Common across governors */ - #define GOV_ONDEMAND 0 - #define GOV_CONSERVATIVE 1 - int governor; - struct attribute_group *attr_group_gov_sys; /* one governor - system */ - struct attribute_group *attr_group_gov_pol; /* one governor - policy */ +struct dbs_governor { + struct cpufreq_governor gov; + struct kobj_type kobj_type; /* * Common data for platforms that don't set @@ -210,74 +159,32 @@ struct common_dbs_data { */ struct dbs_data *gdbs_data; - struct cpu_dbs_info *(*get_cpu_cdbs)(int cpu); - void *(*get_cpu_dbs_info_s)(int cpu); - unsigned int (*gov_dbs_timer)(struct cpufreq_policy *policy, - bool modify_all); - void (*gov_check_cpu)(int cpu, unsigned int load); + unsigned int (*gov_dbs_timer)(struct cpufreq_policy *policy); + struct policy_dbs_info *(*alloc)(void); + void (*free)(struct policy_dbs_info *policy_dbs); int (*init)(struct dbs_data *dbs_data, bool notify); void (*exit)(struct dbs_data *dbs_data, bool notify); - - /* Governor specific ops, see below */ - void *gov_ops; - - /* - * Protects governor's data (struct dbs_data and struct common_dbs_data) - */ - struct mutex mutex; + void (*start)(struct cpufreq_policy *policy); }; -/* Governor Per policy data */ -struct dbs_data { - struct common_dbs_data *cdata; - unsigned int min_sampling_rate; - int usage_count; - void *tuners; -}; +static inline struct dbs_governor *dbs_governor_of(struct cpufreq_policy *policy) +{ + return container_of(policy->governor, struct dbs_governor, gov); +} -/* Governor specific ops, will be passed to dbs_data->gov_ops */ +/* Governor specific operations */ struct od_ops { - void (*powersave_bias_init_cpu)(int cpu); unsigned int (*powersave_bias_target)(struct cpufreq_policy *policy, unsigned int freq_next, unsigned int relation); - void (*freq_increase)(struct cpufreq_policy *policy, unsigned int freq); }; -static inline int delay_for_sampling_rate(unsigned int sampling_rate) -{ - int delay = usecs_to_jiffies(sampling_rate); - - /* We want all CPUs to do sampling nearly on same jiffy */ - if (num_online_cpus() > 1) - delay -= jiffies % delay; - - return delay; -} - -#define declare_show_sampling_rate_min(_gov) \ -static ssize_t show_sampling_rate_min_gov_sys \ -(struct kobject *kobj, struct attribute *attr, char *buf) \ -{ \ - struct dbs_data *dbs_data = _gov##_dbs_cdata.gdbs_data; \ - return sprintf(buf, "%u\n", dbs_data->min_sampling_rate); \ -} \ - \ -static ssize_t show_sampling_rate_min_gov_pol \ -(struct cpufreq_policy *policy, char *buf) \ -{ \ - struct dbs_data *dbs_data = policy->governor_data; \ - return sprintf(buf, "%u\n", dbs_data->min_sampling_rate); \ -} - -extern struct mutex cpufreq_governor_lock; - -void gov_add_timers(struct cpufreq_policy *policy, unsigned int delay); -void gov_cancel_work(struct cpu_common_dbs_info *shared); -void dbs_check_cpu(struct dbs_data *dbs_data, int cpu); -int cpufreq_governor_dbs(struct cpufreq_policy *policy, - struct common_dbs_data *cdata, unsigned int event); +unsigned int dbs_update(struct cpufreq_policy *policy); +int cpufreq_governor_dbs(struct cpufreq_policy *policy, unsigned int event); void od_register_powersave_bias_handler(unsigned int (*f) (struct cpufreq_policy *, unsigned int, unsigned int), unsigned int powersave_bias); void od_unregister_powersave_bias_handler(void); +ssize_t store_sampling_rate(struct dbs_data *dbs_data, const char *buf, + size_t count); +void gov_update_cpu_data(struct dbs_data *dbs_data); #endif /* _CPUFREQ_GOVERNOR_H */ diff --git a/drivers/cpufreq/cpufreq_ondemand.c b/drivers/cpufreq/cpufreq_ondemand.c index eae5107..acd8027 100644 --- a/drivers/cpufreq/cpufreq_ondemand.c +++ b/drivers/cpufreq/cpufreq_ondemand.c @@ -16,7 +16,8 @@ #include <linux/percpu-defs.h> #include <linux/slab.h> #include <linux/tick.h> -#include "cpufreq_governor.h" + +#include "cpufreq_ondemand.h" /* On-demand governor macros */ #define DEF_FREQUENCY_UP_THRESHOLD (80) @@ -27,24 +28,10 @@ #define MIN_FREQUENCY_UP_THRESHOLD (11) #define MAX_FREQUENCY_UP_THRESHOLD (100) -static DEFINE_PER_CPU(struct od_cpu_dbs_info_s, od_cpu_dbs_info); - static struct od_ops od_ops; -#ifndef CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND -static struct cpufreq_governor cpufreq_gov_ondemand; -#endif - static unsigned int default_powersave_bias; -static void ondemand_powersave_bias_init_cpu(int cpu) -{ - struct od_cpu_dbs_info_s *dbs_info = &per_cpu(od_cpu_dbs_info, cpu); - - dbs_info->freq_table = cpufreq_frequency_get_table(cpu); - dbs_info->freq_lo = 0; -} - /* * Not all CPUs want IO time to be accounted as busy; this depends on how * efficient idling at a higher frequency/voltage is. @@ -70,8 +57,8 @@ static int should_io_be_busy(void) /* * Find right freq to be set now with powersave_bias on. - * Returns the freq_hi to be used right now and will set freq_hi_jiffies, - * freq_lo, and freq_lo_jiffies in percpu area for averaging freqs. + * Returns the freq_hi to be used right now and will set freq_hi_delay_us, + * freq_lo, and freq_lo_delay_us in percpu area for averaging freqs. */ static unsigned int generic_powersave_bias_target(struct cpufreq_policy *policy, unsigned int freq_next, unsigned int relation) @@ -79,15 +66,15 @@ static unsigned int generic_powersave_bias_target(struct cpufreq_policy *policy, unsigned int freq_req, freq_reduc, freq_avg; unsigned int freq_hi, freq_lo; unsigned int index = 0; - unsigned int jiffies_total, jiffies_hi, jiffies_lo; - struct od_cpu_dbs_info_s *dbs_info = &per_cpu(od_cpu_dbs_info, - policy->cpu); - struct dbs_data *dbs_data = policy->governor_data; + unsigned int delay_hi_us; + struct policy_dbs_info *policy_dbs = policy->governor_data; + struct od_policy_dbs_info *dbs_info = to_dbs_info(policy_dbs); + struct dbs_data *dbs_data = policy_dbs->dbs_data; struct od_dbs_tuners *od_tuners = dbs_data->tuners; if (!dbs_info->freq_table) { dbs_info->freq_lo = 0; - dbs_info->freq_lo_jiffies = 0; + dbs_info->freq_lo_delay_us = 0; return freq_next; } @@ -110,31 +97,30 @@ static unsigned int generic_powersave_bias_target(struct cpufreq_policy *policy, /* Find out how long we have to be in hi and lo freqs */ if (freq_hi == freq_lo) { dbs_info->freq_lo = 0; - dbs_info->freq_lo_jiffies = 0; + dbs_info->freq_lo_delay_us = 0; return freq_lo; } - jiffies_total = usecs_to_jiffies(od_tuners->sampling_rate); - jiffies_hi = (freq_avg - freq_lo) * jiffies_total; - jiffies_hi += ((freq_hi - freq_lo) / 2); - jiffies_hi /= (freq_hi - freq_lo); - jiffies_lo = jiffies_total - jiffies_hi; + delay_hi_us = (freq_avg - freq_lo) * dbs_data->sampling_rate; + delay_hi_us += (freq_hi - freq_lo) / 2; + delay_hi_us /= freq_hi - freq_lo; + dbs_info->freq_hi_delay_us = delay_hi_us; dbs_info->freq_lo = freq_lo; - dbs_info->freq_lo_jiffies = jiffies_lo; - dbs_info->freq_hi_jiffies = jiffies_hi; + dbs_info->freq_lo_delay_us = dbs_data->sampling_rate - delay_hi_us; return freq_hi; } -static void ondemand_powersave_bias_init(void) +static void ondemand_powersave_bias_init(struct cpufreq_policy *policy) { - int i; - for_each_online_cpu(i) { - ondemand_powersave_bias_init_cpu(i); - } + struct od_policy_dbs_info *dbs_info = to_dbs_info(policy->governor_data); + + dbs_info->freq_table = cpufreq_frequency_get_table(policy->cpu); + dbs_info->freq_lo = 0; } static void dbs_freq_increase(struct cpufreq_policy *policy, unsigned int freq) { - struct dbs_data *dbs_data = policy->governor_data; + struct policy_dbs_info *policy_dbs = policy->governor_data; + struct dbs_data *dbs_data = policy_dbs->dbs_data; struct od_dbs_tuners *od_tuners = dbs_data->tuners; if (od_tuners->powersave_bias) @@ -152,21 +138,21 @@ static void dbs_freq_increase(struct cpufreq_policy *policy, unsigned int freq) * (default), then we try to increase frequency. Else, we adjust the frequency * proportional to load. */ -static void od_check_cpu(int cpu, unsigned int load) +static void od_update(struct cpufreq_policy *policy) { - struct od_cpu_dbs_info_s *dbs_info = &per_cpu(od_cpu_dbs_info, cpu); - struct cpufreq_policy *policy = dbs_info->cdbs.shared->policy; - struct dbs_data *dbs_data = policy->governor_data; + struct policy_dbs_info *policy_dbs = policy->governor_data; + struct od_policy_dbs_info *dbs_info = to_dbs_info(policy_dbs); + struct dbs_data *dbs_data = policy_dbs->dbs_data; struct od_dbs_tuners *od_tuners = dbs_data->tuners; + unsigned int load = dbs_update(policy); dbs_info->freq_lo = 0; /* Check for frequency increase */ - if (load > od_tuners->up_threshold) { + if (load > dbs_data->up_threshold) { /* If switching to max speed, apply sampling_down_factor */ if (policy->cur < policy->max) - dbs_info->rate_mult = - od_tuners->sampling_down_factor; + policy_dbs->rate_mult = dbs_data->sampling_down_factor; dbs_freq_increase(policy, policy->max); } else { /* Calculate the next frequency proportional to load */ @@ -177,177 +163,70 @@ static void od_check_cpu(int cpu, unsigned int load) freq_next = min_f + load * (max_f - min_f) / 100; /* No longer fully busy, reset rate_mult */ - dbs_info->rate_mult = 1; + policy_dbs->rate_mult = 1; - if (!od_tuners->powersave_bias) { - __cpufreq_driver_target(policy, freq_next, - CPUFREQ_RELATION_C); - return; - } + if (od_tuners->powersave_bias) + freq_next = od_ops.powersave_bias_target(policy, + freq_next, + CPUFREQ_RELATION_L); - freq_next = od_ops.powersave_bias_target(policy, freq_next, - CPUFREQ_RELATION_L); __cpufreq_driver_target(policy, freq_next, CPUFREQ_RELATION_C); } } -static unsigned int od_dbs_timer(struct cpufreq_policy *policy, bool modify_all) +static unsigned int od_dbs_timer(struct cpufreq_policy *policy) { - struct dbs_data *dbs_data = policy->governor_data; - unsigned int cpu = policy->cpu; - struct od_cpu_dbs_info_s *dbs_info = &per_cpu(od_cpu_dbs_info, - cpu); - struct od_dbs_tuners *od_tuners = dbs_data->tuners; - int delay = 0, sample_type = dbs_info->sample_type; - - if (!modify_all) - goto max_delay; + struct policy_dbs_info *policy_dbs = policy->governor_data; + struct dbs_data *dbs_data = policy_dbs->dbs_data; + struct od_policy_dbs_info *dbs_info = to_dbs_info(policy_dbs); + int sample_type = dbs_info->sample_type; /* Common NORMAL_SAMPLE setup */ dbs_info->sample_type = OD_NORMAL_SAMPLE; - if (sample_type == OD_SUB_SAMPLE) { - delay = dbs_info->freq_lo_jiffies; + /* + * OD_SUB_SAMPLE doesn't make sense if sample_delay_ns is 0, so ignore + * it then. + */ + if (sample_type == OD_SUB_SAMPLE && policy_dbs->sample_delay_ns > 0) { __cpufreq_driver_target(policy, dbs_info->freq_lo, CPUFREQ_RELATION_H); - } else { - dbs_check_cpu(dbs_data, cpu); - if (dbs_info->freq_lo) { - /* Setup timer for SUB_SAMPLE */ - dbs_info->sample_type = OD_SUB_SAMPLE; - delay = dbs_info->freq_hi_jiffies; - } + return dbs_info->freq_lo_delay_us; } -max_delay: - if (!delay) - delay = delay_for_sampling_rate(od_tuners->sampling_rate - * dbs_info->rate_mult); - - return delay; -} - -/************************** sysfs interface ************************/ -static struct common_dbs_data od_dbs_cdata; - -/** - * update_sampling_rate - update sampling rate effective immediately if needed. - * @new_rate: new sampling rate - * - * If new rate is smaller than the old, simply updating - * dbs_tuners_int.sampling_rate might not be appropriate. For example, if the - * original sampling_rate was 1 second and the requested new sampling rate is 10 - * ms because the user needs immediate reaction from ondemand governor, but not - * sure if higher frequency will be required or not, then, the governor may - * change the sampling rate too late; up to 1 second later. Thus, if we are - * reducing the sampling rate, we need to make the new value effective - * immediately. - */ -static void update_sampling_rate(struct dbs_data *dbs_data, - unsigned int new_rate) -{ - struct od_dbs_tuners *od_tuners = dbs_data->tuners; - struct cpumask cpumask; - int cpu; - - od_tuners->sampling_rate = new_rate = max(new_rate, - dbs_data->min_sampling_rate); - - /* - * Lock governor so that governor start/stop can't execute in parallel. - */ - mutex_lock(&od_dbs_cdata.mutex); - - cpumask_copy(&cpumask, cpu_online_mask); - - for_each_cpu(cpu, &cpumask) { - struct cpufreq_policy *policy; - struct od_cpu_dbs_info_s *dbs_info; - struct cpu_dbs_info *cdbs; - struct cpu_common_dbs_info *shared; - unsigned long next_sampling, appointed_at; - - dbs_info = &per_cpu(od_cpu_dbs_info, cpu); - cdbs = &dbs_info->cdbs; - shared = cdbs->shared; - - /* - * A valid shared and shared->policy means governor hasn't - * stopped or exited yet. - */ - if (!shared || !shared->policy) - continue; - - policy = shared->policy; - - /* clear all CPUs of this policy */ - cpumask_andnot(&cpumask, &cpumask, policy->cpus); + od_update(policy); - /* - * Update sampling rate for CPUs whose policy is governed by - * dbs_data. In case of governor_per_policy, only a single - * policy will be governed by dbs_data, otherwise there can be - * multiple policies that are governed by the same dbs_data. - */ - if (dbs_data != policy->governor_data) - continue; - - /* - * Checking this for any CPU should be fine, timers for all of - * them are scheduled together. - */ - next_sampling = jiffies + usecs_to_jiffies(new_rate); - appointed_at = dbs_info->cdbs.timer.expires; - - if (time_before(next_sampling, appointed_at)) { - gov_cancel_work(shared); - gov_add_timers(policy, usecs_to_jiffies(new_rate)); - - } + if (dbs_info->freq_lo) { + /* Setup timer for SUB_SAMPLE */ + dbs_info->sample_type = OD_SUB_SAMPLE; + return dbs_info->freq_hi_delay_us; } - mutex_unlock(&od_dbs_cdata.mutex); + return dbs_data->sampling_rate * policy_dbs->rate_mult; } -static ssize_t store_sampling_rate(struct dbs_data *dbs_data, const char *buf, - size_t count) -{ - unsigned int input; - int ret; - ret = sscanf(buf, "%u", &input); - if (ret != 1) - return -EINVAL; - - update_sampling_rate(dbs_data, input); - return count; -} +/************************** sysfs interface ************************/ +static struct dbs_governor od_dbs_gov; static ssize_t store_io_is_busy(struct dbs_data *dbs_data, const char *buf, size_t count) { - struct od_dbs_tuners *od_tuners = dbs_data->tuners; unsigned int input; int ret; - unsigned int j; ret = sscanf(buf, "%u", &input); if (ret != 1) return -EINVAL; - od_tuners->io_is_busy = !!input; + dbs_data->io_is_busy = !!input; /* we need to re-evaluate prev_cpu_idle */ - for_each_online_cpu(j) { - struct od_cpu_dbs_info_s *dbs_info = &per_cpu(od_cpu_dbs_info, - j); - dbs_info->cdbs.prev_cpu_idle = get_cpu_idle_time(j, - &dbs_info->cdbs.prev_cpu_wall, od_tuners->io_is_busy); - } + gov_update_cpu_data(dbs_data); + return count; } static ssize_t store_up_threshold(struct dbs_data *dbs_data, const char *buf, size_t count) { - struct od_dbs_tuners *od_tuners = dbs_data->tuners; unsigned int input; int ret; ret = sscanf(buf, "%u", &input); @@ -357,40 +236,43 @@ static ssize_t store_up_threshold(struct dbs_data *dbs_data, const char *buf, return -EINVAL; } - od_tuners->up_threshold = input; + dbs_data->up_threshold = input; return count; } static ssize_t store_sampling_down_factor(struct dbs_data *dbs_data, const char *buf, size_t count) { - struct od_dbs_tuners *od_tuners = dbs_data->tuners; - unsigned int input, j; + struct policy_dbs_info *policy_dbs; + unsigned int input; int ret; ret = sscanf(buf, "%u", &input); if (ret != 1 || input > MAX_SAMPLING_DOWN_FACTOR || input < 1) return -EINVAL; - od_tuners->sampling_down_factor = input; + + dbs_data->sampling_down_factor = input; /* Reset down sampling multiplier in case it was active */ - for_each_online_cpu(j) { - struct od_cpu_dbs_info_s *dbs_info = &per_cpu(od_cpu_dbs_info, - j); - dbs_info->rate_mult = 1; + list_for_each_entry(policy_dbs, &dbs_data->policy_dbs_list, list) { + /* + * Doing this without locking might lead to using different + * rate_mult values in od_update() and od_dbs_timer(). + */ + mutex_lock(&policy_dbs->timer_mutex); + policy_dbs->rate_mult = 1; + mutex_unlock(&policy_dbs->timer_mutex); } + return count; } static ssize_t store_ignore_nice_load(struct dbs_data *dbs_data, const char *buf, size_t count) { - struct od_dbs_tuners *od_tuners = dbs_data->tuners; unsigned int input; int ret; - unsigned int j; - ret = sscanf(buf, "%u", &input); if (ret != 1) return -EINVAL; @@ -398,22 +280,14 @@ static ssize_t store_ignore_nice_load(struct dbs_data *dbs_data, if (input > 1) input = 1; - if (input == od_tuners->ignore_nice_load) { /* nothing to do */ + if (input == dbs_data->ignore_nice_load) { /* nothing to do */ return count; } - od_tuners->ignore_nice_load = input; + dbs_data->ignore_nice_load = input; /* we need to re-evaluate prev_cpu_idle */ - for_each_online_cpu(j) { - struct od_cpu_dbs_info_s *dbs_info; - dbs_info = &per_cpu(od_cpu_dbs_info, j); - dbs_info->cdbs.prev_cpu_idle = get_cpu_idle_time(j, - &dbs_info->cdbs.prev_cpu_wall, od_tuners->io_is_busy); - if (od_tuners->ignore_nice_load) - dbs_info->cdbs.prev_cpu_nice = - kcpustat_cpu(j).cpustat[CPUTIME_NICE]; + gov_update_cpu_data(dbs_data); - } return count; } @@ -421,6 +295,7 @@ static ssize_t store_powersave_bias(struct dbs_data *dbs_data, const char *buf, size_t count) { struct od_dbs_tuners *od_tuners = dbs_data->tuners; + struct policy_dbs_info *policy_dbs; unsigned int input; int ret; ret = sscanf(buf, "%u", &input); @@ -432,59 +307,54 @@ static ssize_t store_powersave_bias(struct dbs_data *dbs_data, const char *buf, input = 1000; od_tuners->powersave_bias = input; - ondemand_powersave_bias_init(); + + list_for_each_entry(policy_dbs, &dbs_data->policy_dbs_list, list) + ondemand_powersave_bias_init(policy_dbs->policy); + return count; } -show_store_one(od, sampling_rate); -show_store_one(od, io_is_busy); -show_store_one(od, up_threshold); -show_store_one(od, sampling_down_factor); -show_store_one(od, ignore_nice_load); -show_store_one(od, powersave_bias); -declare_show_sampling_rate_min(od); - -gov_sys_pol_attr_rw(sampling_rate); -gov_sys_pol_attr_rw(io_is_busy); -gov_sys_pol_attr_rw(up_threshold); -gov_sys_pol_attr_rw(sampling_down_factor); -gov_sys_pol_attr_rw(ignore_nice_load); -gov_sys_pol_attr_rw(powersave_bias); -gov_sys_pol_attr_ro(sampling_rate_min); - -static struct attribute *dbs_attributes_gov_sys[] = { - &sampling_rate_min_gov_sys.attr, - &sampling_rate_gov_sys.attr, - &up_threshold_gov_sys.attr, - &sampling_down_factor_gov_sys.attr, - &ignore_nice_load_gov_sys.attr, - &powersave_bias_gov_sys.attr, - &io_is_busy_gov_sys.attr, +gov_show_one_common(sampling_rate); +gov_show_one_common(up_threshold); +gov_show_one_common(sampling_down_factor); +gov_show_one_common(ignore_nice_load); +gov_show_one_common(min_sampling_rate); +gov_show_one_common(io_is_busy); +gov_show_one(od, powersave_bias); + +gov_attr_rw(sampling_rate); +gov_attr_rw(io_is_busy); +gov_attr_rw(up_threshold); +gov_attr_rw(sampling_down_factor); +gov_attr_rw(ignore_nice_load); +gov_attr_rw(powersave_bias); +gov_attr_ro(min_sampling_rate); + +static struct attribute *od_attributes[] = { + &min_sampling_rate.attr, + &sampling_rate.attr, + &up_threshold.attr, + &sampling_down_factor.attr, + &ignore_nice_load.attr, + &powersave_bias.attr, + &io_is_busy.attr, NULL }; -static struct attribute_group od_attr_group_gov_sys = { - .attrs = dbs_attributes_gov_sys, - .name = "ondemand", -}; +/************************** sysfs end ************************/ -static struct attribute *dbs_attributes_gov_pol[] = { - &sampling_rate_min_gov_pol.attr, - &sampling_rate_gov_pol.attr, - &up_threshold_gov_pol.attr, - &sampling_down_factor_gov_pol.attr, - &ignore_nice_load_gov_pol.attr, - &powersave_bias_gov_pol.attr, - &io_is_busy_gov_pol.attr, - NULL -}; +static struct policy_dbs_info *od_alloc(void) +{ + struct od_policy_dbs_info *dbs_info; -static struct attribute_group od_attr_group_gov_pol = { - .attrs = dbs_attributes_gov_pol, - .name = "ondemand", -}; + dbs_info = kzalloc(sizeof(*dbs_info), GFP_KERNEL); + return dbs_info ? &dbs_info->policy_dbs : NULL; +} -/************************** sysfs end ************************/ +static void od_free(struct policy_dbs_info *policy_dbs) +{ + kfree(to_dbs_info(policy_dbs)); +} static int od_init(struct dbs_data *dbs_data, bool notify) { @@ -503,7 +373,7 @@ static int od_init(struct dbs_data *dbs_data, bool notify) put_cpu(); if (idle_time != -1ULL) { /* Idle micro accounting is supported. Use finer thresholds */ - tuners->up_threshold = MICRO_FREQUENCY_UP_THRESHOLD; + dbs_data->up_threshold = MICRO_FREQUENCY_UP_THRESHOLD; /* * In nohz/micro accounting case we set the minimum frequency * not depending on HZ, but fixed (very low). The deferred @@ -511,17 +381,17 @@ static int od_init(struct dbs_data *dbs_data, bool notify) */ dbs_data->min_sampling_rate = MICRO_FREQUENCY_MIN_SAMPLE_RATE; } else { - tuners->up_threshold = DEF_FREQUENCY_UP_THRESHOLD; + dbs_data->up_threshold = DEF_FREQUENCY_UP_THRESHOLD; /* For correct statistics, we need 10 ticks for each measure */ dbs_data->min_sampling_rate = MIN_SAMPLING_RATE_RATIO * jiffies_to_usecs(10); } - tuners->sampling_down_factor = DEF_SAMPLING_DOWN_FACTOR; - tuners->ignore_nice_load = 0; + dbs_data->sampling_down_factor = DEF_SAMPLING_DOWN_FACTOR; + dbs_data->ignore_nice_load = 0; tuners->powersave_bias = default_powersave_bias; - tuners->io_is_busy = should_io_be_busy(); + dbs_data->io_is_busy = should_io_be_busy(); dbs_data->tuners = tuners; return 0; @@ -532,33 +402,38 @@ static void od_exit(struct dbs_data *dbs_data, bool notify) kfree(dbs_data->tuners); } -define_get_cpu_dbs_routines(od_cpu_dbs_info); +static void od_start(struct cpufreq_policy *policy) +{ + struct od_policy_dbs_info *dbs_info = to_dbs_info(policy->governor_data); + + dbs_info->sample_type = OD_NORMAL_SAMPLE; + ondemand_powersave_bias_init(policy); +} static struct od_ops od_ops = { - .powersave_bias_init_cpu = ondemand_powersave_bias_init_cpu, .powersave_bias_target = generic_powersave_bias_target, - .freq_increase = dbs_freq_increase, }; -static struct common_dbs_data od_dbs_cdata = { - .governor = GOV_ONDEMAND, - .attr_group_gov_sys = &od_attr_group_gov_sys, - .attr_group_gov_pol = &od_attr_group_gov_pol, - .get_cpu_cdbs = get_cpu_cdbs, - .get_cpu_dbs_info_s = get_cpu_dbs_info_s, +static struct dbs_governor od_dbs_gov = { + .gov = { + .name = "ondemand", + .governor = cpufreq_governor_dbs, + .max_transition_latency = TRANSITION_LATENCY_LIMIT, + .owner = THIS_MODULE, + }, + .kobj_type = { .default_attrs = od_attributes }, .gov_dbs_timer = od_dbs_timer, - .gov_check_cpu = od_check_cpu, - .gov_ops = &od_ops, + .alloc = od_alloc, + .free = od_free, .init = od_init, .exit = od_exit, - .mutex = __MUTEX_INITIALIZER(od_dbs_cdata.mutex), + .start = od_start, }; +#define CPU_FREQ_GOV_ONDEMAND (&od_dbs_gov.gov) + static void od_set_powersave_bias(unsigned int powersave_bias) { - struct cpufreq_policy *policy; - struct dbs_data *dbs_data; - struct od_dbs_tuners *od_tuners; unsigned int cpu; cpumask_t done; @@ -567,22 +442,25 @@ static void od_set_powersave_bias(unsigned int powersave_bias) get_online_cpus(); for_each_online_cpu(cpu) { - struct cpu_common_dbs_info *shared; + struct cpufreq_policy *policy; + struct policy_dbs_info *policy_dbs; + struct dbs_data *dbs_data; + struct od_dbs_tuners *od_tuners; if (cpumask_test_cpu(cpu, &done)) continue; - shared = per_cpu(od_cpu_dbs_info, cpu).cdbs.shared; - if (!shared) + policy = cpufreq_cpu_get_raw(cpu); + if (!policy || policy->governor != CPU_FREQ_GOV_ONDEMAND) continue; - policy = shared->policy; - cpumask_or(&done, &done, policy->cpus); - - if (policy->governor != &cpufreq_gov_ondemand) + policy_dbs = policy->governor_data; + if (!policy_dbs) continue; - dbs_data = policy->governor_data; + cpumask_or(&done, &done, policy->cpus); + + dbs_data = policy_dbs->dbs_data; od_tuners = dbs_data->tuners; od_tuners->powersave_bias = default_powersave_bias; } @@ -605,30 +483,14 @@ void od_unregister_powersave_bias_handler(void) } EXPORT_SYMBOL_GPL(od_unregister_powersave_bias_handler); -static int od_cpufreq_governor_dbs(struct cpufreq_policy *policy, - unsigned int event) -{ - return cpufreq_governor_dbs(policy, &od_dbs_cdata, event); -} - -#ifndef CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND -static -#endif -struct cpufreq_governor cpufreq_gov_ondemand = { - .name = "ondemand", - .governor = od_cpufreq_governor_dbs, - .max_transition_latency = TRANSITION_LATENCY_LIMIT, - .owner = THIS_MODULE, -}; - static int __init cpufreq_gov_dbs_init(void) { - return cpufreq_register_governor(&cpufreq_gov_ondemand); + return cpufreq_register_governor(CPU_FREQ_GOV_ONDEMAND); } static void __exit cpufreq_gov_dbs_exit(void) { - cpufreq_unregister_governor(&cpufreq_gov_ondemand); + cpufreq_unregister_governor(CPU_FREQ_GOV_ONDEMAND); } MODULE_AUTHOR("Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>"); @@ -638,6 +500,11 @@ MODULE_DESCRIPTION("'cpufreq_ondemand' - A dynamic cpufreq governor for " MODULE_LICENSE("GPL"); #ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND +struct cpufreq_governor *cpufreq_default_governor(void) +{ + return CPU_FREQ_GOV_ONDEMAND; +} + fs_initcall(cpufreq_gov_dbs_init); #else module_init(cpufreq_gov_dbs_init); diff --git a/drivers/cpufreq/cpufreq_ondemand.h b/drivers/cpufreq/cpufreq_ondemand.h new file mode 100644 index 0000000..f0121db --- /dev/null +++ b/drivers/cpufreq/cpufreq_ondemand.h @@ -0,0 +1,30 @@ +/* + * Header file for CPUFreq ondemand governor and related code. + * + * Copyright (C) 2016, Intel Corporation + * Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com> + * + * 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 "cpufreq_governor.h" + +struct od_policy_dbs_info { + struct policy_dbs_info policy_dbs; + struct cpufreq_frequency_table *freq_table; + unsigned int freq_lo; + unsigned int freq_lo_delay_us; + unsigned int freq_hi_delay_us; + unsigned int sample_type:1; +}; + +static inline struct od_policy_dbs_info *to_dbs_info(struct policy_dbs_info *policy_dbs) +{ + return container_of(policy_dbs, struct od_policy_dbs_info, policy_dbs); +} + +struct od_dbs_tuners { + unsigned int powersave_bias; +}; diff --git a/drivers/cpufreq/cpufreq_performance.c b/drivers/cpufreq/cpufreq_performance.c index cf117de..af9f4b9 100644 --- a/drivers/cpufreq/cpufreq_performance.c +++ b/drivers/cpufreq/cpufreq_performance.c @@ -33,10 +33,7 @@ static int cpufreq_governor_performance(struct cpufreq_policy *policy, return 0; } -#ifdef CONFIG_CPU_FREQ_GOV_PERFORMANCE_MODULE -static -#endif -struct cpufreq_governor cpufreq_gov_performance = { +static struct cpufreq_governor cpufreq_gov_performance = { .name = "performance", .governor = cpufreq_governor_performance, .owner = THIS_MODULE, @@ -52,6 +49,19 @@ static void __exit cpufreq_gov_performance_exit(void) cpufreq_unregister_governor(&cpufreq_gov_performance); } +#ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE +struct cpufreq_governor *cpufreq_default_governor(void) +{ + return &cpufreq_gov_performance; +} +#endif +#ifndef CONFIG_CPU_FREQ_GOV_PERFORMANCE_MODULE +struct cpufreq_governor *cpufreq_fallback_governor(void) +{ + return &cpufreq_gov_performance; +} +#endif + MODULE_AUTHOR("Dominik Brodowski <linux@brodo.de>"); MODULE_DESCRIPTION("CPUfreq policy governor 'performance'"); MODULE_LICENSE("GPL"); diff --git a/drivers/cpufreq/cpufreq_powersave.c b/drivers/cpufreq/cpufreq_powersave.c index e3b874c..b8b4002 100644 --- a/drivers/cpufreq/cpufreq_powersave.c +++ b/drivers/cpufreq/cpufreq_powersave.c @@ -33,10 +33,7 @@ static int cpufreq_governor_powersave(struct cpufreq_policy *policy, return 0; } -#ifndef CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE -static -#endif -struct cpufreq_governor cpufreq_gov_powersave = { +static struct cpufreq_governor cpufreq_gov_powersave = { .name = "powersave", .governor = cpufreq_governor_powersave, .owner = THIS_MODULE, @@ -57,6 +54,11 @@ MODULE_DESCRIPTION("CPUfreq policy governor 'powersave'"); MODULE_LICENSE("GPL"); #ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE +struct cpufreq_governor *cpufreq_default_governor(void) +{ + return &cpufreq_gov_powersave; +} + fs_initcall(cpufreq_gov_powersave_init); #else module_init(cpufreq_gov_powersave_init); diff --git a/drivers/cpufreq/cpufreq_userspace.c b/drivers/cpufreq/cpufreq_userspace.c index 4dbf1db..4d16f45 100644 --- a/drivers/cpufreq/cpufreq_userspace.c +++ b/drivers/cpufreq/cpufreq_userspace.c @@ -89,10 +89,7 @@ static int cpufreq_governor_userspace(struct cpufreq_policy *policy, return rc; } -#ifndef CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE -static -#endif -struct cpufreq_governor cpufreq_gov_userspace = { +static struct cpufreq_governor cpufreq_gov_userspace = { .name = "userspace", .governor = cpufreq_governor_userspace, .store_setspeed = cpufreq_set, @@ -116,6 +113,11 @@ MODULE_DESCRIPTION("CPUfreq policy governor 'userspace'"); MODULE_LICENSE("GPL"); #ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE +struct cpufreq_governor *cpufreq_default_governor(void) +{ + return &cpufreq_gov_userspace; +} + fs_initcall(cpufreq_gov_userspace_init); #else module_init(cpufreq_gov_userspace_init); diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c index 3a4b39a..cb56074 100644 --- a/drivers/cpufreq/intel_pstate.c +++ b/drivers/cpufreq/intel_pstate.c @@ -71,7 +71,7 @@ struct sample { u64 mperf; u64 tsc; int freq; - ktime_t time; + u64 time; }; struct pstate_data { @@ -103,13 +103,13 @@ struct _pid { struct cpudata { int cpu; - struct timer_list timer; + struct update_util_data update_util; struct pstate_data pstate; struct vid_data vid; struct _pid pid; - ktime_t last_sample_time; + u64 last_sample_time; u64 prev_aperf; u64 prev_mperf; u64 prev_tsc; @@ -120,6 +120,7 @@ struct cpudata { static struct cpudata **all_cpu_data; struct pstate_adjust_policy { int sample_rate_ms; + s64 sample_rate_ns; int deadband; int setpoint; int p_gain_pct; @@ -197,8 +198,8 @@ static struct perf_limits *limits = &powersave_limits; static inline void pid_reset(struct _pid *pid, int setpoint, int busy, int deadband, int integral) { - pid->setpoint = setpoint; - pid->deadband = deadband; + pid->setpoint = int_tofp(setpoint); + pid->deadband = int_tofp(deadband); pid->integral = int_tofp(integral); pid->last_err = int_tofp(setpoint) - int_tofp(busy); } @@ -224,9 +225,9 @@ static signed int pid_calc(struct _pid *pid, int32_t busy) int32_t pterm, dterm, fp_error; int32_t integral_limit; - fp_error = int_tofp(pid->setpoint) - busy; + fp_error = pid->setpoint - busy; - if (abs(fp_error) <= int_tofp(pid->deadband)) + if (abs(fp_error) <= pid->deadband) return 0; pterm = mul_fp(pid->p_gain, fp_error); @@ -286,7 +287,7 @@ static inline void update_turbo_state(void) cpu->pstate.max_pstate == cpu->pstate.turbo_pstate); } -static void intel_pstate_hwp_set(void) +static void intel_pstate_hwp_set(const struct cpumask *cpumask) { int min, hw_min, max, hw_max, cpu, range, adj_range; u64 value, cap; @@ -296,9 +297,7 @@ static void intel_pstate_hwp_set(void) hw_max = HWP_HIGHEST_PERF(cap); range = hw_max - hw_min; - get_online_cpus(); - - for_each_online_cpu(cpu) { + for_each_cpu(cpu, cpumask) { rdmsrl_on_cpu(cpu, MSR_HWP_REQUEST, &value); adj_range = limits->min_perf_pct * range / 100; min = hw_min + adj_range; @@ -317,7 +316,12 @@ static void intel_pstate_hwp_set(void) value |= HWP_MAX_PERF(max); wrmsrl_on_cpu(cpu, MSR_HWP_REQUEST, value); } +} +static void intel_pstate_hwp_set_online_cpus(void) +{ + get_online_cpus(); + intel_pstate_hwp_set(cpu_online_mask); put_online_cpus(); } @@ -439,7 +443,7 @@ static ssize_t store_no_turbo(struct kobject *a, struct attribute *b, limits->no_turbo = clamp_t(int, input, 0, 1); if (hwp_active) - intel_pstate_hwp_set(); + intel_pstate_hwp_set_online_cpus(); return count; } @@ -465,7 +469,7 @@ static ssize_t store_max_perf_pct(struct kobject *a, struct attribute *b, int_tofp(100)); if (hwp_active) - intel_pstate_hwp_set(); + intel_pstate_hwp_set_online_cpus(); return count; } @@ -490,7 +494,7 @@ static ssize_t store_min_perf_pct(struct kobject *a, struct attribute *b, int_tofp(100)); if (hwp_active) - intel_pstate_hwp_set(); + intel_pstate_hwp_set_online_cpus(); return count; } @@ -531,6 +535,9 @@ static void __init intel_pstate_sysfs_expose_params(void) static void intel_pstate_hwp_enable(struct cpudata *cpudata) { + /* First disable HWP notification interrupt as we don't process them */ + wrmsrl_on_cpu(cpudata->cpu, MSR_HWP_INTERRUPT, 0x00); + wrmsrl_on_cpu(cpudata->cpu, MSR_PM_ENABLE, 0x1); } @@ -712,7 +719,7 @@ static void core_set_pstate(struct cpudata *cpudata, int pstate) if (limits->no_turbo && !limits->turbo_disabled) val |= (u64)1 << 32; - wrmsrl_on_cpu(cpudata->cpu, MSR_IA32_PERF_CTL, val); + wrmsrl(MSR_IA32_PERF_CTL, val); } static int knl_get_turbo_pstate(void) @@ -824,11 +831,11 @@ static void intel_pstate_get_min_max(struct cpudata *cpu, int *min, int *max) * policy, or by cpu specific default values determined through * experimentation. */ - max_perf_adj = fp_toint(mul_fp(int_tofp(max_perf), limits->max_perf)); + max_perf_adj = fp_toint(max_perf * limits->max_perf); *max = clamp_t(int, max_perf_adj, cpu->pstate.min_pstate, cpu->pstate.turbo_pstate); - min_perf = fp_toint(mul_fp(int_tofp(max_perf), limits->min_perf)); + min_perf = fp_toint(max_perf * limits->min_perf); *min = clamp_t(int, min_perf, cpu->pstate.min_pstate, max_perf); } @@ -874,16 +881,10 @@ static inline void intel_pstate_calc_busy(struct cpudata *cpu) core_pct = int_tofp(sample->aperf) * int_tofp(100); core_pct = div64_u64(core_pct, int_tofp(sample->mperf)); - sample->freq = fp_toint( - mul_fp(int_tofp( - cpu->pstate.max_pstate_physical * - cpu->pstate.scaling / 100), - core_pct)); - sample->core_pct_busy = (int32_t)core_pct; } -static inline void intel_pstate_sample(struct cpudata *cpu) +static inline bool intel_pstate_sample(struct cpudata *cpu, u64 time) { u64 aperf, mperf; unsigned long flags; @@ -893,14 +894,14 @@ static inline void intel_pstate_sample(struct cpudata *cpu) rdmsrl(MSR_IA32_APERF, aperf); rdmsrl(MSR_IA32_MPERF, mperf); tsc = rdtsc(); - if ((cpu->prev_mperf == mperf) || (cpu->prev_tsc == tsc)) { + if (cpu->prev_mperf == mperf || cpu->prev_tsc == tsc) { local_irq_restore(flags); - return; + return false; } local_irq_restore(flags); cpu->last_sample_time = cpu->sample.time; - cpu->sample.time = ktime_get(); + cpu->sample.time = time; cpu->sample.aperf = aperf; cpu->sample.mperf = mperf; cpu->sample.tsc = tsc; @@ -908,27 +909,16 @@ static inline void intel_pstate_sample(struct cpudata *cpu) cpu->sample.mperf -= cpu->prev_mperf; cpu->sample.tsc -= cpu->prev_tsc; - intel_pstate_calc_busy(cpu); - cpu->prev_aperf = aperf; cpu->prev_mperf = mperf; cpu->prev_tsc = tsc; + return true; } -static inline void intel_hwp_set_sample_time(struct cpudata *cpu) -{ - int delay; - - delay = msecs_to_jiffies(50); - mod_timer_pinned(&cpu->timer, jiffies + delay); -} - -static inline void intel_pstate_set_sample_time(struct cpudata *cpu) +static inline int32_t get_avg_frequency(struct cpudata *cpu) { - int delay; - - delay = msecs_to_jiffies(pid_params.sample_rate_ms); - mod_timer_pinned(&cpu->timer, jiffies + delay); + return div64_u64(cpu->pstate.max_pstate_physical * cpu->sample.aperf * + cpu->pstate.scaling, cpu->sample.mperf); } static inline int32_t get_target_pstate_use_cpu_load(struct cpudata *cpu) @@ -954,7 +944,6 @@ static inline int32_t get_target_pstate_use_cpu_load(struct cpudata *cpu) mperf = cpu->sample.mperf + delta_iowait_mperf; cpu->prev_cummulative_iowait = cummulative_iowait; - /* * The load can be estimated as the ratio of the mperf counter * running at a constant frequency during active periods @@ -970,8 +959,9 @@ static inline int32_t get_target_pstate_use_cpu_load(struct cpudata *cpu) static inline int32_t get_target_pstate_use_performance(struct cpudata *cpu) { int32_t core_busy, max_pstate, current_pstate, sample_ratio; - s64 duration_us; - u32 sample_time; + u64 duration_ns; + + intel_pstate_calc_busy(cpu); /* * core_busy is the ratio of actual performance to max @@ -990,18 +980,16 @@ static inline int32_t get_target_pstate_use_performance(struct cpudata *cpu) core_busy = mul_fp(core_busy, div_fp(max_pstate, current_pstate)); /* - * Since we have a deferred timer, it will not fire unless - * we are in C0. So, determine if the actual elapsed time - * is significantly greater (3x) than our sample interval. If it - * is, then we were idle for a long enough period of time - * to adjust our busyness. + * Since our utilization update callback will not run unless we are + * in C0, check if the actual elapsed time is significantly greater (3x) + * than our sample interval. If it is, then we were idle for a long + * enough period of time to adjust our busyness. */ - sample_time = pid_params.sample_rate_ms * USEC_PER_MSEC; - duration_us = ktime_us_delta(cpu->sample.time, - cpu->last_sample_time); - if (duration_us > sample_time * 3) { - sample_ratio = div_fp(int_tofp(sample_time), - int_tofp(duration_us)); + duration_ns = cpu->sample.time - cpu->last_sample_time; + if ((s64)duration_ns > pid_params.sample_rate_ns * 3 + && cpu->last_sample_time > 0) { + sample_ratio = div_fp(int_tofp(pid_params.sample_rate_ns), + int_tofp(duration_ns)); core_busy = mul_fp(core_busy, sample_ratio); } @@ -1028,26 +1016,21 @@ static inline void intel_pstate_adjust_busy_pstate(struct cpudata *cpu) sample->mperf, sample->aperf, sample->tsc, - sample->freq); + get_avg_frequency(cpu)); } -static void intel_hwp_timer_func(unsigned long __data) +static void intel_pstate_update_util(struct update_util_data *data, u64 time, + unsigned long util, unsigned long max) { - struct cpudata *cpu = (struct cpudata *) __data; + struct cpudata *cpu = container_of(data, struct cpudata, update_util); + u64 delta_ns = time - cpu->sample.time; - intel_pstate_sample(cpu); - intel_hwp_set_sample_time(cpu); -} + if ((s64)delta_ns >= pid_params.sample_rate_ns) { + bool sample_taken = intel_pstate_sample(cpu, time); -static void intel_pstate_timer_func(unsigned long __data) -{ - struct cpudata *cpu = (struct cpudata *) __data; - - intel_pstate_sample(cpu); - - intel_pstate_adjust_busy_pstate(cpu); - - intel_pstate_set_sample_time(cpu); + if (sample_taken && !hwp_active) + intel_pstate_adjust_busy_pstate(cpu); + } } #define ICPU(model, policy) \ @@ -1095,24 +1078,19 @@ static int intel_pstate_init_cpu(unsigned int cpunum) cpu->cpu = cpunum; - if (hwp_active) + if (hwp_active) { intel_pstate_hwp_enable(cpu); + pid_params.sample_rate_ms = 50; + pid_params.sample_rate_ns = 50 * NSEC_PER_MSEC; + } intel_pstate_get_cpu_pstates(cpu); - init_timer_deferrable(&cpu->timer); - cpu->timer.data = (unsigned long)cpu; - cpu->timer.expires = jiffies + HZ/100; - - if (!hwp_active) - cpu->timer.function = intel_pstate_timer_func; - else - cpu->timer.function = intel_hwp_timer_func; - intel_pstate_busy_pid_reset(cpu); - intel_pstate_sample(cpu); + intel_pstate_sample(cpu, 0); - add_timer_on(&cpu->timer, cpunum); + cpu->update_util.func = intel_pstate_update_util; + cpufreq_set_update_util_data(cpunum, &cpu->update_util); pr_debug("intel_pstate: controlling: cpu %d\n", cpunum); @@ -1128,7 +1106,7 @@ static unsigned int intel_pstate_get(unsigned int cpu_num) if (!cpu) return 0; sample = &cpu->sample; - return sample->freq; + return get_avg_frequency(cpu); } static int intel_pstate_set_policy(struct cpufreq_policy *policy) @@ -1141,7 +1119,7 @@ static int intel_pstate_set_policy(struct cpufreq_policy *policy) pr_debug("intel_pstate: set performance\n"); limits = &performance_limits; if (hwp_active) - intel_pstate_hwp_set(); + intel_pstate_hwp_set(policy->cpus); return 0; } @@ -1173,7 +1151,7 @@ static int intel_pstate_set_policy(struct cpufreq_policy *policy) int_tofp(100)); if (hwp_active) - intel_pstate_hwp_set(); + intel_pstate_hwp_set(policy->cpus); return 0; } @@ -1196,7 +1174,9 @@ static void intel_pstate_stop_cpu(struct cpufreq_policy *policy) pr_debug("intel_pstate: CPU %d exiting\n", cpu_num); - del_timer_sync(&all_cpu_data[cpu_num]->timer); + cpufreq_set_update_util_data(cpu_num, NULL); + synchronize_sched(); + if (hwp_active) return; @@ -1260,6 +1240,7 @@ static int intel_pstate_msrs_not_valid(void) static void copy_pid_params(struct pstate_adjust_policy *policy) { pid_params.sample_rate_ms = policy->sample_rate_ms; + pid_params.sample_rate_ns = pid_params.sample_rate_ms * NSEC_PER_MSEC; pid_params.p_gain_pct = policy->p_gain_pct; pid_params.i_gain_pct = policy->i_gain_pct; pid_params.d_gain_pct = policy->d_gain_pct; @@ -1397,6 +1378,11 @@ static inline bool intel_pstate_platform_pwr_mgmt_exists(void) { return false; } static inline bool intel_pstate_has_acpi_ppc(void) { return false; } #endif /* CONFIG_ACPI */ +static const struct x86_cpu_id hwp_support_ids[] __initconst = { + { X86_VENDOR_INTEL, 6, X86_MODEL_ANY, X86_FEATURE_HWP }, + {} +}; + static int __init intel_pstate_init(void) { int cpu, rc = 0; @@ -1406,17 +1392,16 @@ static int __init intel_pstate_init(void) if (no_load) return -ENODEV; + if (x86_match_cpu(hwp_support_ids) && !no_hwp) { + copy_cpu_funcs(&core_params.funcs); + hwp_active++; + goto hwp_cpu_matched; + } + id = x86_match_cpu(intel_pstate_cpu_ids); if (!id) return -ENODEV; - /* - * The Intel pstate driver will be ignored if the platform - * firmware has its own power management modes. - */ - if (intel_pstate_platform_pwr_mgmt_exists()) - return -ENODEV; - cpu_def = (struct cpu_defaults *)id->driver_data; copy_pid_params(&cpu_def->pid_policy); @@ -1425,17 +1410,20 @@ static int __init intel_pstate_init(void) if (intel_pstate_msrs_not_valid()) return -ENODEV; +hwp_cpu_matched: + /* + * The Intel pstate driver will be ignored if the platform + * firmware has its own power management modes. + */ + if (intel_pstate_platform_pwr_mgmt_exists()) + return -ENODEV; + pr_info("Intel P-state driver initializing.\n"); all_cpu_data = vzalloc(sizeof(void *) * num_possible_cpus()); if (!all_cpu_data) return -ENOMEM; - if (static_cpu_has(X86_FEATURE_HWP) && !no_hwp) { - pr_info("intel_pstate: HWP enabled\n"); - hwp_active++; - } - if (!hwp_active && hwp_only) goto out; @@ -1446,12 +1434,16 @@ static int __init intel_pstate_init(void) intel_pstate_debug_expose_params(); intel_pstate_sysfs_expose_params(); + if (hwp_active) + pr_info("intel_pstate: HWP enabled\n"); + return rc; out: get_online_cpus(); for_each_online_cpu(cpu) { if (all_cpu_data[cpu]) { - del_timer_sync(&all_cpu_data[cpu]->timer); + cpufreq_set_update_util_data(cpu, NULL); + synchronize_sched(); kfree(all_cpu_data[cpu]); } } diff --git a/drivers/cpufreq/powernv-cpufreq.c b/drivers/cpufreq/powernv-cpufreq.c index 547890f..50bf120 100644 --- a/drivers/cpufreq/powernv-cpufreq.c +++ b/drivers/cpufreq/powernv-cpufreq.c @@ -28,6 +28,8 @@ #include <linux/of.h> #include <linux/reboot.h> #include <linux/slab.h> +#include <linux/cpu.h> +#include <trace/events/power.h> #include <asm/cputhreads.h> #include <asm/firmware.h> @@ -42,13 +44,24 @@ static struct cpufreq_frequency_table powernv_freqs[POWERNV_MAX_PSTATES+1]; static bool rebooting, throttled, occ_reset; +static unsigned int *core_to_chip_map; + +static const char * const throttle_reason[] = { + "No throttling", + "Power Cap", + "Processor Over Temperature", + "Power Supply Failure", + "Over Current", + "OCC Reset" +}; static struct chip { unsigned int id; bool throttled; + bool restore; + u8 throttle_reason; cpumask_t mask; struct work_struct throttle; - bool restore; } *chips; static int nr_chips; @@ -312,13 +325,14 @@ static inline unsigned int get_nominal_index(void) static void powernv_cpufreq_throttle_check(void *data) { unsigned int cpu = smp_processor_id(); + unsigned int chip_id = core_to_chip_map[cpu_core_index_of_thread(cpu)]; unsigned long pmsr; int pmsr_pmax, i; pmsr = get_pmspr(SPRN_PMSR); for (i = 0; i < nr_chips; i++) - if (chips[i].id == cpu_to_chip_id(cpu)) + if (chips[i].id == chip_id) break; /* Check for Pmax Capping */ @@ -328,17 +342,17 @@ static void powernv_cpufreq_throttle_check(void *data) goto next; chips[i].throttled = true; if (pmsr_pmax < powernv_pstate_info.nominal) - pr_crit("CPU %d on Chip %u has Pmax reduced below nominal frequency (%d < %d)\n", - cpu, chips[i].id, pmsr_pmax, - powernv_pstate_info.nominal); - else - pr_info("CPU %d on Chip %u has Pmax reduced below turbo frequency (%d < %d)\n", - cpu, chips[i].id, pmsr_pmax, - powernv_pstate_info.max); + pr_warn_once("CPU %d on Chip %u has Pmax reduced below nominal frequency (%d < %d)\n", + cpu, chips[i].id, pmsr_pmax, + powernv_pstate_info.nominal); + trace_powernv_throttle(chips[i].id, + throttle_reason[chips[i].throttle_reason], + pmsr_pmax); } else if (chips[i].throttled) { chips[i].throttled = false; - pr_info("CPU %d on Chip %u has Pmax restored to %d\n", cpu, - chips[i].id, pmsr_pmax); + trace_powernv_throttle(chips[i].id, + throttle_reason[chips[i].throttle_reason], + pmsr_pmax); } /* Check if Psafe_mode_active is set in PMSR. */ @@ -356,7 +370,7 @@ next: if (throttled) { pr_info("PMSR = %16lx\n", pmsr); - pr_crit("CPU Frequency could be throttled\n"); + pr_warn("CPU Frequency could be throttled\n"); } } @@ -423,18 +437,19 @@ void powernv_cpufreq_work_fn(struct work_struct *work) { struct chip *chip = container_of(work, struct chip, throttle); unsigned int cpu; - cpumask_var_t mask; + cpumask_t mask; - smp_call_function_any(&chip->mask, + get_online_cpus(); + cpumask_and(&mask, &chip->mask, cpu_online_mask); + smp_call_function_any(&mask, powernv_cpufreq_throttle_check, NULL, 0); if (!chip->restore) - return; + goto out; chip->restore = false; - cpumask_copy(mask, &chip->mask); - for_each_cpu_and(cpu, mask, cpu_online_mask) { - int index, tcpu; + for_each_cpu(cpu, &mask) { + int index; struct cpufreq_policy policy; cpufreq_get_policy(&policy, cpu); @@ -442,20 +457,12 @@ void powernv_cpufreq_work_fn(struct work_struct *work) policy.cur, CPUFREQ_RELATION_C, &index); powernv_cpufreq_target_index(&policy, index); - for_each_cpu(tcpu, policy.cpus) - cpumask_clear_cpu(tcpu, mask); + cpumask_andnot(&mask, &mask, policy.cpus); } +out: + put_online_cpus(); } -static char throttle_reason[][30] = { - "No throttling", - "Power Cap", - "Processor Over Temperature", - "Power Supply Failure", - "Over Current", - "OCC Reset" - }; - static int powernv_cpufreq_occ_msg(struct notifier_block *nb, unsigned long msg_type, void *_msg) { @@ -481,7 +488,7 @@ static int powernv_cpufreq_occ_msg(struct notifier_block *nb, */ if (!throttled) { throttled = true; - pr_crit("CPU frequency is throttled for duration\n"); + pr_warn("CPU frequency is throttled for duration\n"); } break; @@ -505,23 +512,18 @@ static int powernv_cpufreq_occ_msg(struct notifier_block *nb, return 0; } - if (omsg.throttle_status && + for (i = 0; i < nr_chips; i++) + if (chips[i].id == omsg.chip) + break; + + if (omsg.throttle_status >= 0 && omsg.throttle_status <= OCC_MAX_THROTTLE_STATUS) - pr_info("OCC: Chip %u Pmax reduced due to %s\n", - (unsigned int)omsg.chip, - throttle_reason[omsg.throttle_status]); - else if (!omsg.throttle_status) - pr_info("OCC: Chip %u %s\n", (unsigned int)omsg.chip, - throttle_reason[omsg.throttle_status]); - else - return 0; + chips[i].throttle_reason = omsg.throttle_status; - for (i = 0; i < nr_chips; i++) - if (chips[i].id == omsg.chip) { - if (!omsg.throttle_status) - chips[i].restore = true; - schedule_work(&chips[i].throttle); - } + if (!omsg.throttle_status) + chips[i].restore = true; + + schedule_work(&chips[i].throttle); } return 0; } @@ -556,29 +558,54 @@ static int init_chip_info(void) unsigned int chip[256]; unsigned int cpu, i; unsigned int prev_chip_id = UINT_MAX; + cpumask_t cpu_mask; + int ret = -ENOMEM; + + core_to_chip_map = kcalloc(cpu_nr_cores(), sizeof(unsigned int), + GFP_KERNEL); + if (!core_to_chip_map) + goto out; - for_each_possible_cpu(cpu) { + cpumask_copy(&cpu_mask, cpu_possible_mask); + for_each_cpu(cpu, &cpu_mask) { unsigned int id = cpu_to_chip_id(cpu); if (prev_chip_id != id) { prev_chip_id = id; chip[nr_chips++] = id; } + core_to_chip_map[cpu_core_index_of_thread(cpu)] = id; + cpumask_andnot(&cpu_mask, &cpu_mask, cpu_sibling_mask(cpu)); } - chips = kmalloc_array(nr_chips, sizeof(struct chip), GFP_KERNEL); + chips = kcalloc(nr_chips, sizeof(struct chip), GFP_KERNEL); if (!chips) - return -ENOMEM; + goto free_chip_map; for (i = 0; i < nr_chips; i++) { chips[i].id = chip[i]; - chips[i].throttled = false; cpumask_copy(&chips[i].mask, cpumask_of_node(chip[i])); INIT_WORK(&chips[i].throttle, powernv_cpufreq_work_fn); - chips[i].restore = false; } return 0; +free_chip_map: + kfree(core_to_chip_map); +out: + return ret; +} + +static inline void clean_chip_info(void) +{ + kfree(chips); + kfree(core_to_chip_map); +} + +static inline void unregister_all_notifiers(void) +{ + opal_message_notifier_unregister(OPAL_MSG_OCC, + &powernv_cpufreq_opal_nb); + unregister_reboot_notifier(&powernv_cpufreq_reboot_nb); } static int __init powernv_cpufreq_init(void) @@ -591,28 +618,35 @@ static int __init powernv_cpufreq_init(void) /* Discover pstates from device tree and init */ rc = init_powernv_pstates(); - if (rc) { - pr_info("powernv-cpufreq disabled. System does not support PState control\n"); - return rc; - } + if (rc) + goto out; /* Populate chip info */ rc = init_chip_info(); if (rc) - return rc; + goto out; register_reboot_notifier(&powernv_cpufreq_reboot_nb); opal_message_notifier_register(OPAL_MSG_OCC, &powernv_cpufreq_opal_nb); - return cpufreq_register_driver(&powernv_cpufreq_driver); + + rc = cpufreq_register_driver(&powernv_cpufreq_driver); + if (!rc) + return 0; + + pr_info("Failed to register the cpufreq driver (%d)\n", rc); + unregister_all_notifiers(); + clean_chip_info(); +out: + pr_info("Platform driver disabled. System does not support PState control\n"); + return rc; } module_init(powernv_cpufreq_init); static void __exit powernv_cpufreq_exit(void) { - unregister_reboot_notifier(&powernv_cpufreq_reboot_nb); - opal_message_notifier_unregister(OPAL_MSG_OCC, - &powernv_cpufreq_opal_nb); cpufreq_unregister_driver(&powernv_cpufreq_driver); + unregister_all_notifiers(); + clean_chip_info(); } module_exit(powernv_cpufreq_exit); diff --git a/drivers/cpuidle/governors/menu.c b/drivers/cpuidle/governors/menu.c index 0742b32..27fc733 100644 --- a/drivers/cpuidle/governors/menu.c +++ b/drivers/cpuidle/governors/menu.c @@ -199,8 +199,8 @@ static void menu_update(struct cpuidle_driver *drv, struct cpuidle_device *dev); static void get_typical_interval(struct menu_device *data) { int i, divisor; - unsigned int max, thresh; - uint64_t avg, stddev; + unsigned int max, thresh, avg; + uint64_t sum, variance; thresh = UINT_MAX; /* Discard outliers above this value */ @@ -208,52 +208,51 @@ again: /* First calculate the average of past intervals */ max = 0; - avg = 0; + sum = 0; divisor = 0; for (i = 0; i < INTERVALS; i++) { unsigned int value = data->intervals[i]; if (value <= thresh) { - avg += value; + sum += value; divisor++; if (value > max) max = value; } } if (divisor == INTERVALS) - avg >>= INTERVAL_SHIFT; + avg = sum >> INTERVAL_SHIFT; else - do_div(avg, divisor); + avg = div_u64(sum, divisor); - /* Then try to determine standard deviation */ - stddev = 0; + /* Then try to determine variance */ + variance = 0; for (i = 0; i < INTERVALS; i++) { unsigned int value = data->intervals[i]; if (value <= thresh) { - int64_t diff = value - avg; - stddev += diff * diff; + int64_t diff = (int64_t)value - avg; + variance += diff * diff; } } if (divisor == INTERVALS) - stddev >>= INTERVAL_SHIFT; + variance >>= INTERVAL_SHIFT; else - do_div(stddev, divisor); + do_div(variance, divisor); /* - * The typical interval is obtained when standard deviation is small - * or standard deviation is small compared to the average interval. - * - * int_sqrt() formal parameter type is unsigned long. When the - * greatest difference to an outlier exceeds ~65 ms * sqrt(divisor) - * the resulting squared standard deviation exceeds the input domain - * of int_sqrt on platforms where unsigned long is 32 bits in size. - * In such case reject the candidate average. + * The typical interval is obtained when standard deviation is + * small (stddev <= 20 us, variance <= 400 us^2) or standard + * deviation is small compared to the average interval (avg > + * 6*stddev, avg^2 > 36*variance). The average is smaller than + * UINT_MAX aka U32_MAX, so computing its square does not + * overflow a u64. We simply reject this candidate average if + * the standard deviation is greater than 715 s (which is + * rather unlikely). * * Use this result only if there is no timer to wake us up sooner. */ - if (likely(stddev <= ULONG_MAX)) { - stddev = int_sqrt(stddev); - if (((avg > stddev * 6) && (divisor * 4 >= INTERVALS * 3)) - || stddev <= 20) { + if (likely(variance <= U64_MAX/36)) { + if ((((u64)avg*avg > variance*36) && (divisor * 4 >= INTERVALS * 3)) + || variance <= 400) { if (data->next_timer_us > avg) data->predicted_us = avg; return; diff --git a/drivers/i2c/busses/i2c-designware-platdrv.c b/drivers/i2c/busses/i2c-designware-platdrv.c index 438f1b4..d656657 100644 --- a/drivers/i2c/busses/i2c-designware-platdrv.c +++ b/drivers/i2c/busses/i2c-designware-platdrv.c @@ -123,6 +123,7 @@ static const struct acpi_device_id dw_i2c_acpi_match[] = { { "80860F41", 0 }, { "808622C1", 0 }, { "AMD0010", ACCESS_INTR_MASK }, + { "AMDI0010", ACCESS_INTR_MASK }, { "AMDI0510", 0 }, { "APMC0D0F", 0 }, { } diff --git a/drivers/mailbox/pcc.c b/drivers/mailbox/pcc.c index 8f779a1..0ddf638 100644 --- a/drivers/mailbox/pcc.c +++ b/drivers/mailbox/pcc.c @@ -63,6 +63,7 @@ #include <linux/platform_device.h> #include <linux/mailbox_controller.h> #include <linux/mailbox_client.h> +#include <linux/io-64-nonatomic-lo-hi.h> #include "mailbox.h" @@ -70,6 +71,9 @@ static struct mbox_chan *pcc_mbox_channels; +/* Array of cached virtual address for doorbell registers */ +static void __iomem **pcc_doorbell_vaddr; + static struct mbox_controller pcc_mbox_ctrl = {}; /** * get_pcc_channel - Given a PCC subspace idx, get @@ -160,6 +164,66 @@ void pcc_mbox_free_channel(struct mbox_chan *chan) } EXPORT_SYMBOL_GPL(pcc_mbox_free_channel); +/* + * PCC can be used with perf critical drivers such as CPPC + * So it makes sense to locally cache the virtual address and + * use it to read/write to PCC registers such as doorbell register + * + * The below read_register and write_registers are used to read and + * write from perf critical registers such as PCC doorbell register + */ +static int read_register(void __iomem *vaddr, u64 *val, unsigned int bit_width) +{ + int ret_val = 0; + + switch (bit_width) { + case 8: + *val = readb(vaddr); + break; + case 16: + *val = readw(vaddr); + break; + case 32: + *val = readl(vaddr); + break; + case 64: + *val = readq(vaddr); + break; + default: + pr_debug("Error: Cannot read register of %u bit width", + bit_width); + ret_val = -EFAULT; + break; + } + return ret_val; +} + +static int write_register(void __iomem *vaddr, u64 val, unsigned int bit_width) +{ + int ret_val = 0; + + switch (bit_width) { + case 8: + writeb(val, vaddr); + break; + case 16: + writew(val, vaddr); + break; + case 32: + writel(val, vaddr); + break; + case 64: + writeq(val, vaddr); + break; + default: + pr_debug("Error: Cannot write register of %u bit width", + bit_width); + ret_val = -EFAULT; + break; + } + return ret_val; +} + /** * pcc_send_data - Called from Mailbox Controller code. Used * here only to ring the channel doorbell. The PCC client @@ -175,21 +239,39 @@ EXPORT_SYMBOL_GPL(pcc_mbox_free_channel); static int pcc_send_data(struct mbox_chan *chan, void *data) { struct acpi_pcct_hw_reduced *pcct_ss = chan->con_priv; - struct acpi_generic_address doorbell; + struct acpi_generic_address *doorbell; u64 doorbell_preserve; u64 doorbell_val; u64 doorbell_write; + u32 id = chan - pcc_mbox_channels; + int ret = 0; + + if (id >= pcc_mbox_ctrl.num_chans) { + pr_debug("pcc_send_data: Invalid mbox_chan passed\n"); + return -ENOENT; + } - doorbell = pcct_ss->doorbell_register; + doorbell = &pcct_ss->doorbell_register; doorbell_preserve = pcct_ss->preserve_mask; doorbell_write = pcct_ss->write_mask; /* Sync notification from OS to Platform. */ - acpi_read(&doorbell_val, &doorbell); - acpi_write((doorbell_val & doorbell_preserve) | doorbell_write, - &doorbell); - - return 0; + if (pcc_doorbell_vaddr[id]) { + ret = read_register(pcc_doorbell_vaddr[id], &doorbell_val, + doorbell->bit_width); + if (ret) + return ret; + ret = write_register(pcc_doorbell_vaddr[id], + (doorbell_val & doorbell_preserve) | doorbell_write, + doorbell->bit_width); + } else { + ret = acpi_read(&doorbell_val, doorbell); + if (ret) + return ret; + ret = acpi_write((doorbell_val & doorbell_preserve) | doorbell_write, + doorbell); + } + return ret; } static const struct mbox_chan_ops pcc_chan_ops = { @@ -265,14 +347,29 @@ static int __init acpi_pcc_probe(void) return -ENOMEM; } + pcc_doorbell_vaddr = kcalloc(count, sizeof(void *), GFP_KERNEL); + if (!pcc_doorbell_vaddr) { + kfree(pcc_mbox_channels); + return -ENOMEM; + } + /* Point to the first PCC subspace entry */ pcct_entry = (struct acpi_subtable_header *) ( (unsigned long) pcct_tbl + sizeof(struct acpi_table_pcct)); for (i = 0; i < count; i++) { + struct acpi_generic_address *db_reg; + struct acpi_pcct_hw_reduced *pcct_ss; pcc_mbox_channels[i].con_priv = pcct_entry; pcct_entry = (struct acpi_subtable_header *) ((unsigned long) pcct_entry + pcct_entry->length); + + /* If doorbell is in system memory cache the virt address */ + pcct_ss = (struct acpi_pcct_hw_reduced *)pcct_entry; + db_reg = &pcct_ss->doorbell_register; + if (db_reg->space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) + pcc_doorbell_vaddr[i] = acpi_os_ioremap(db_reg->address, + db_reg->bit_width/8); } pcc_mbox_ctrl.num_chans = count; diff --git a/drivers/pnp/pnpacpi/rsparser.c b/drivers/pnp/pnpacpi/rsparser.c index 0579649..4b717c6 100644 --- a/drivers/pnp/pnpacpi/rsparser.c +++ b/drivers/pnp/pnpacpi/rsparser.c @@ -252,6 +252,10 @@ static acpi_status pnpacpi_allocated_resource(struct acpi_resource *res, case ACPI_RESOURCE_TYPE_GENERIC_REGISTER: break; + case ACPI_RESOURCE_TYPE_SERIAL_BUS: + /* serial bus connections (I2C/SPI/UART) are not pnp */ + break; + default: dev_warn(&dev->dev, "unknown resource type %d in _CRS\n", res->type); diff --git a/drivers/powercap/intel_rapl.c b/drivers/powercap/intel_rapl.c index 6c592dc..cdfd01f0 100644 --- a/drivers/powercap/intel_rapl.c +++ b/drivers/powercap/intel_rapl.c @@ -133,6 +133,12 @@ struct rapl_domain_data { unsigned long timestamp; }; +struct msrl_action { + u32 msr_no; + u64 clear_mask; + u64 set_mask; + int err; +}; #define DOMAIN_STATE_INACTIVE BIT(0) #define DOMAIN_STATE_POWER_LIMIT_SET BIT(1) @@ -149,6 +155,7 @@ struct rapl_power_limit { static const char pl1_name[] = "long_term"; static const char pl2_name[] = "short_term"; +struct rapl_package; struct rapl_domain { const char *name; enum rapl_domain_type id; @@ -159,7 +166,7 @@ struct rapl_domain { u64 attr_map; /* track capabilities */ unsigned int state; unsigned int domain_energy_unit; - int package_id; + struct rapl_package *rp; }; #define power_zone_to_rapl_domain(_zone) \ container_of(_zone, struct rapl_domain, power_zone) @@ -184,6 +191,7 @@ struct rapl_package { * notify interrupt enable status. */ struct list_head plist; + int lead_cpu; /* one active cpu per package for access */ }; struct rapl_defaults { @@ -231,10 +239,10 @@ static int rapl_read_data_raw(struct rapl_domain *rd, static int rapl_write_data_raw(struct rapl_domain *rd, enum rapl_primitives prim, unsigned long long value); -static u64 rapl_unit_xlate(struct rapl_domain *rd, int package, +static u64 rapl_unit_xlate(struct rapl_domain *rd, enum unit_type type, u64 value, int to_raw); -static void package_power_limit_irq_save(int package_id); +static void package_power_limit_irq_save(struct rapl_package *rp); static LIST_HEAD(rapl_packages); /* guarded by CPU hotplug lock */ @@ -260,20 +268,6 @@ static struct rapl_package *find_package_by_id(int id) return NULL; } -/* caller to ensure CPU hotplug lock is held */ -static int find_active_cpu_on_package(int package_id) -{ - int i; - - for_each_online_cpu(i) { - if (topology_physical_package_id(i) == package_id) - return i; - } - /* all CPUs on this package are offline */ - - return -ENODEV; -} - /* caller must hold cpu hotplug lock */ static void rapl_cleanup_data(void) { @@ -312,25 +306,19 @@ static int get_max_energy_counter(struct powercap_zone *pcd_dev, u64 *energy) { struct rapl_domain *rd = power_zone_to_rapl_domain(pcd_dev); - *energy = rapl_unit_xlate(rd, 0, ENERGY_UNIT, ENERGY_STATUS_MASK, 0); + *energy = rapl_unit_xlate(rd, ENERGY_UNIT, ENERGY_STATUS_MASK, 0); return 0; } static int release_zone(struct powercap_zone *power_zone) { struct rapl_domain *rd = power_zone_to_rapl_domain(power_zone); - struct rapl_package *rp; + struct rapl_package *rp = rd->rp; /* package zone is the last zone of a package, we can free * memory here since all children has been unregistered. */ if (rd->id == RAPL_DOMAIN_PACKAGE) { - rp = find_package_by_id(rd->package_id); - if (!rp) { - dev_warn(&power_zone->dev, "no package id %s\n", - rd->name); - return -ENODEV; - } kfree(rd); rp->domains = NULL; } @@ -432,11 +420,7 @@ static int set_power_limit(struct powercap_zone *power_zone, int id, get_online_cpus(); rd = power_zone_to_rapl_domain(power_zone); - rp = find_package_by_id(rd->package_id); - if (!rp) { - ret = -ENODEV; - goto set_exit; - } + rp = rd->rp; if (rd->state & DOMAIN_STATE_BIOS_LOCKED) { dev_warn(&power_zone->dev, "%s locked by BIOS, monitoring only\n", @@ -456,7 +440,7 @@ static int set_power_limit(struct powercap_zone *power_zone, int id, ret = -EINVAL; } if (!ret) - package_power_limit_irq_save(rd->package_id); + package_power_limit_irq_save(rp); set_exit: put_online_cpus(); return ret; @@ -655,24 +639,19 @@ static void rapl_init_domains(struct rapl_package *rp) break; } if (mask) { - rd->package_id = rp->id; + rd->rp = rp; rd++; } } } -static u64 rapl_unit_xlate(struct rapl_domain *rd, int package, - enum unit_type type, u64 value, - int to_raw) +static u64 rapl_unit_xlate(struct rapl_domain *rd, enum unit_type type, + u64 value, int to_raw) { u64 units = 1; - struct rapl_package *rp; + struct rapl_package *rp = rd->rp; u64 scale = 1; - rp = find_package_by_id(package); - if (!rp) - return value; - switch (type) { case POWER_UNIT: units = rp->power_unit; @@ -769,10 +748,8 @@ static int rapl_read_data_raw(struct rapl_domain *rd, msr = rd->msrs[rp->id]; if (!msr) return -EINVAL; - /* use physical package id to look up active cpus */ - cpu = find_active_cpu_on_package(rd->package_id); - if (cpu < 0) - return cpu; + + cpu = rd->rp->lead_cpu; /* special-case package domain, which uses a different bit*/ if (prim == FW_LOCK && rd->id == RAPL_DOMAIN_PACKAGE) { @@ -793,42 +770,66 @@ static int rapl_read_data_raw(struct rapl_domain *rd, final = value & rp->mask; final = final >> rp->shift; if (xlate) - *data = rapl_unit_xlate(rd, rd->package_id, rp->unit, final, 0); + *data = rapl_unit_xlate(rd, rp->unit, final, 0); else *data = final; return 0; } + +static int msrl_update_safe(u32 msr_no, u64 clear_mask, u64 set_mask) +{ + int err; + u64 val; + + err = rdmsrl_safe(msr_no, &val); + if (err) + goto out; + + val &= ~clear_mask; + val |= set_mask; + + err = wrmsrl_safe(msr_no, val); + +out: + return err; +} + +static void msrl_update_func(void *info) +{ + struct msrl_action *ma = info; + + ma->err = msrl_update_safe(ma->msr_no, ma->clear_mask, ma->set_mask); +} + /* Similar use of primitive info in the read counterpart */ static int rapl_write_data_raw(struct rapl_domain *rd, enum rapl_primitives prim, unsigned long long value) { - u64 msr_val; - u32 msr; struct rapl_primitive_info *rp = &rpi[prim]; int cpu; + u64 bits; + struct msrl_action ma; + int ret; - cpu = find_active_cpu_on_package(rd->package_id); - if (cpu < 0) - return cpu; - msr = rd->msrs[rp->id]; - if (rdmsrl_safe_on_cpu(cpu, msr, &msr_val)) { - dev_dbg(&rd->power_zone.dev, - "failed to read msr 0x%x on cpu %d\n", msr, cpu); - return -EIO; - } - value = rapl_unit_xlate(rd, rd->package_id, rp->unit, value, 1); - msr_val &= ~rp->mask; - msr_val |= value << rp->shift; - if (wrmsrl_safe_on_cpu(cpu, msr, msr_val)) { - dev_dbg(&rd->power_zone.dev, - "failed to write msr 0x%x on cpu %d\n", msr, cpu); - return -EIO; - } + cpu = rd->rp->lead_cpu; + bits = rapl_unit_xlate(rd, rp->unit, value, 1); + bits |= bits << rp->shift; + memset(&ma, 0, sizeof(ma)); - return 0; + ma.msr_no = rd->msrs[rp->id]; + ma.clear_mask = rp->mask; + ma.set_mask = bits; + + ret = smp_call_function_single(cpu, msrl_update_func, &ma, 1); + if (ret) + WARN_ON_ONCE(ret); + else + ret = ma.err; + + return ret; } /* @@ -893,6 +894,21 @@ static int rapl_check_unit_atom(struct rapl_package *rp, int cpu) return 0; } +static void power_limit_irq_save_cpu(void *info) +{ + u32 l, h = 0; + struct rapl_package *rp = (struct rapl_package *)info; + + /* save the state of PLN irq mask bit before disabling it */ + rdmsr_safe(MSR_IA32_PACKAGE_THERM_INTERRUPT, &l, &h); + if (!(rp->power_limit_irq & PACKAGE_PLN_INT_SAVED)) { + rp->power_limit_irq = l & PACKAGE_THERM_INT_PLN_ENABLE; + rp->power_limit_irq |= PACKAGE_PLN_INT_SAVED; + } + l &= ~PACKAGE_THERM_INT_PLN_ENABLE; + wrmsr_safe(MSR_IA32_PACKAGE_THERM_INTERRUPT, l, h); +} + /* REVISIT: * When package power limit is set artificially low by RAPL, LVT @@ -904,61 +920,40 @@ static int rapl_check_unit_atom(struct rapl_package *rp, int cpu) * to do by adding an atomic notifier. */ -static void package_power_limit_irq_save(int package_id) +static void package_power_limit_irq_save(struct rapl_package *rp) { - u32 l, h = 0; - int cpu; - struct rapl_package *rp; - - rp = find_package_by_id(package_id); - if (!rp) - return; - if (!boot_cpu_has(X86_FEATURE_PTS) || !boot_cpu_has(X86_FEATURE_PLN)) return; - cpu = find_active_cpu_on_package(package_id); - if (cpu < 0) - return; - /* save the state of PLN irq mask bit before disabling it */ - rdmsr_safe_on_cpu(cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT, &l, &h); - if (!(rp->power_limit_irq & PACKAGE_PLN_INT_SAVED)) { - rp->power_limit_irq = l & PACKAGE_THERM_INT_PLN_ENABLE; - rp->power_limit_irq |= PACKAGE_PLN_INT_SAVED; - } - l &= ~PACKAGE_THERM_INT_PLN_ENABLE; - wrmsr_on_cpu(cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT, l, h); + smp_call_function_single(rp->lead_cpu, power_limit_irq_save_cpu, rp, 1); } -/* restore per package power limit interrupt enable state */ -static void package_power_limit_irq_restore(int package_id) +static void power_limit_irq_restore_cpu(void *info) { - u32 l, h; - int cpu; - struct rapl_package *rp; + u32 l, h = 0; + struct rapl_package *rp = (struct rapl_package *)info; - rp = find_package_by_id(package_id); - if (!rp) - return; + rdmsr_safe(MSR_IA32_PACKAGE_THERM_INTERRUPT, &l, &h); - if (!boot_cpu_has(X86_FEATURE_PTS) || !boot_cpu_has(X86_FEATURE_PLN)) - return; + if (rp->power_limit_irq & PACKAGE_THERM_INT_PLN_ENABLE) + l |= PACKAGE_THERM_INT_PLN_ENABLE; + else + l &= ~PACKAGE_THERM_INT_PLN_ENABLE; + + wrmsr_safe(MSR_IA32_PACKAGE_THERM_INTERRUPT, l, h); +} - cpu = find_active_cpu_on_package(package_id); - if (cpu < 0) +/* restore per package power limit interrupt enable state */ +static void package_power_limit_irq_restore(struct rapl_package *rp) +{ + if (!boot_cpu_has(X86_FEATURE_PTS) || !boot_cpu_has(X86_FEATURE_PLN)) return; /* irq enable state not saved, nothing to restore */ if (!(rp->power_limit_irq & PACKAGE_PLN_INT_SAVED)) return; - rdmsr_safe_on_cpu(cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT, &l, &h); - - if (rp->power_limit_irq & PACKAGE_THERM_INT_PLN_ENABLE) - l |= PACKAGE_THERM_INT_PLN_ENABLE; - else - l &= ~PACKAGE_THERM_INT_PLN_ENABLE; - wrmsr_on_cpu(cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT, l, h); + smp_call_function_single(rp->lead_cpu, power_limit_irq_restore_cpu, rp, 1); } static void set_floor_freq_default(struct rapl_domain *rd, bool mode) @@ -1141,7 +1136,7 @@ static int rapl_unregister_powercap(void) * hotplug lock held */ list_for_each_entry(rp, &rapl_packages, plist) { - package_power_limit_irq_restore(rp->id); + package_power_limit_irq_restore(rp); for (rd = rp->domains; rd < rp->domains + rp->nr_domains; rd++) { @@ -1392,7 +1387,8 @@ static int rapl_detect_topology(void) /* add the new package to the list */ new_package->id = phy_package_id; new_package->nr_cpus = 1; - + /* use the first active cpu of the package to access */ + new_package->lead_cpu = i; /* check if the package contains valid domains */ if (rapl_detect_domains(new_package, i) || rapl_defaults->check_unit(new_package, i)) { @@ -1448,6 +1444,8 @@ static int rapl_add_package(int cpu) /* add the new package to the list */ rp->id = phy_package_id; rp->nr_cpus = 1; + rp->lead_cpu = cpu; + /* check if the package contains valid domains */ if (rapl_detect_domains(rp, cpu) || rapl_defaults->check_unit(rp, cpu)) { @@ -1480,6 +1478,7 @@ static int rapl_cpu_callback(struct notifier_block *nfb, unsigned long cpu = (unsigned long)hcpu; int phy_package_id; struct rapl_package *rp; + int lead_cpu; phy_package_id = topology_physical_package_id(cpu); switch (action) { @@ -1500,6 +1499,15 @@ static int rapl_cpu_callback(struct notifier_block *nfb, break; if (--rp->nr_cpus == 0) rapl_remove_package(rp); + else if (cpu == rp->lead_cpu) { + /* choose another active cpu in the package */ + lead_cpu = cpumask_any_but(topology_core_cpumask(cpu), cpu); + if (lead_cpu < nr_cpu_ids) + rp->lead_cpu = lead_cpu; + else /* should never go here */ + pr_err("no active cpu available for package %d\n", + phy_package_id); + } } return NOTIFY_OK; diff --git a/include/acpi/acoutput.h b/include/acpi/acoutput.h index 5bfc619..34f601e 100644 --- a/include/acpi/acoutput.h +++ b/include/acpi/acoutput.h @@ -262,7 +262,7 @@ #define ACPI_GET_FUNCTION_NAME _acpi_function_name /* - * The Name parameter should be the procedure name as a quoted string. + * The Name parameter should be the procedure name as a non-quoted string. * The function name is also used by the function exit macros below. * Note: (const char) is used to be compatible with the debug interfaces * and macros such as __func__. diff --git a/include/acpi/acpixf.h b/include/acpi/acpixf.h index c96621e..1755697 100644 --- a/include/acpi/acpixf.h +++ b/include/acpi/acpixf.h @@ -897,11 +897,9 @@ ACPI_MSG_DEPENDENT_RETURN_VOID(ACPI_PRINTF_LIKE(3) acpi_warning(const char *module_name, u32 line_number, const char *format, ...)) -ACPI_MSG_DEPENDENT_RETURN_VOID(ACPI_PRINTF_LIKE(3) +ACPI_MSG_DEPENDENT_RETURN_VOID(ACPI_PRINTF_LIKE(1) void ACPI_INTERNAL_VAR_XFACE - acpi_info(const char *module_name, - u32 line_number, - const char *format, ...)) + acpi_info(const char *format, ...)) ACPI_MSG_DEPENDENT_RETURN_VOID(ACPI_PRINTF_LIKE(3) void ACPI_INTERNAL_VAR_XFACE acpi_bios_error(const char *module_name, diff --git a/include/acpi/processor.h b/include/acpi/processor.h index 07fb100..6f1805d 100644 --- a/include/acpi/processor.h +++ b/include/acpi/processor.h @@ -9,6 +9,7 @@ #define ACPI_PROCESSOR_CLASS "processor" #define ACPI_PROCESSOR_DEVICE_NAME "Processor" #define ACPI_PROCESSOR_DEVICE_HID "ACPI0007" +#define ACPI_PROCESSOR_CONTAINER_HID "ACPI0010" #define ACPI_PROCESSOR_BUSY_METRIC 10 @@ -394,14 +395,6 @@ static inline int acpi_processor_hotplug(struct acpi_processor *pr) } #endif /* CONFIG_ACPI_PROCESSOR_IDLE */ -#if defined(CONFIG_PM_SLEEP) & defined(CONFIG_ACPI_PROCESSOR_IDLE) -void acpi_processor_syscore_init(void); -void acpi_processor_syscore_exit(void); -#else -static inline void acpi_processor_syscore_init(void) {} -static inline void acpi_processor_syscore_exit(void) {} -#endif - /* in processor_thermal.c */ int acpi_processor_get_limit_info(struct acpi_processor *pr); extern const struct thermal_cooling_device_ops processor_cooling_ops; diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h index 88a4215..718e872 100644 --- a/include/linux/cpufreq.h +++ b/include/linux/cpufreq.h @@ -80,7 +80,6 @@ struct cpufreq_policy { unsigned int last_policy; /* policy before unplug */ struct cpufreq_governor *governor; /* see below */ void *governor_data; - bool governor_enabled; /* governor start/stop flag */ char last_governor[CPUFREQ_NAME_LEN]; /* last governor used */ struct work_struct update; /* if update_policy() needs to be @@ -100,10 +99,6 @@ struct cpufreq_policy { * - Any routine that will write to the policy structure and/or may take away * the policy altogether (eg. CPU hotplug), will hold this lock in write * mode before doing so. - * - * Additional rules: - * - Lock should not be held across - * __cpufreq_governor(data, CPUFREQ_GOV_POLICY_EXIT); */ struct rw_semaphore rwsem; @@ -464,29 +459,8 @@ int __cpufreq_driver_target(struct cpufreq_policy *policy, int cpufreq_register_governor(struct cpufreq_governor *governor); void cpufreq_unregister_governor(struct cpufreq_governor *governor); -/* CPUFREQ DEFAULT GOVERNOR */ -/* - * Performance governor is fallback governor if any other gov failed to auto - * load due latency restrictions - */ -#ifdef CONFIG_CPU_FREQ_GOV_PERFORMANCE -extern struct cpufreq_governor cpufreq_gov_performance; -#endif -#ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE -#define CPUFREQ_DEFAULT_GOVERNOR (&cpufreq_gov_performance) -#elif defined(CONFIG_CPU_FREQ_DEFAULT_GOV_POWERSAVE) -extern struct cpufreq_governor cpufreq_gov_powersave; -#define CPUFREQ_DEFAULT_GOVERNOR (&cpufreq_gov_powersave) -#elif defined(CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE) -extern struct cpufreq_governor cpufreq_gov_userspace; -#define CPUFREQ_DEFAULT_GOVERNOR (&cpufreq_gov_userspace) -#elif defined(CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND) -extern struct cpufreq_governor cpufreq_gov_ondemand; -#define CPUFREQ_DEFAULT_GOVERNOR (&cpufreq_gov_ondemand) -#elif defined(CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE) -extern struct cpufreq_governor cpufreq_gov_conservative; -#define CPUFREQ_DEFAULT_GOVERNOR (&cpufreq_gov_conservative) -#endif +struct cpufreq_governor *cpufreq_default_governor(void); +struct cpufreq_governor *cpufreq_fallback_governor(void); /********************************************************************* * FREQUENCY TABLE HELPERS * @@ -525,16 +499,6 @@ static inline void dev_pm_opp_free_cpufreq_table(struct device *dev, } #endif -static inline bool cpufreq_next_valid(struct cpufreq_frequency_table **pos) -{ - while ((*pos)->frequency != CPUFREQ_TABLE_END) - if ((*pos)->frequency != CPUFREQ_ENTRY_INVALID) - return true; - else - (*pos)++; - return false; -} - /* * cpufreq_for_each_entry - iterate over a cpufreq_frequency_table * @pos: the cpufreq_frequency_table * to use as a loop cursor. @@ -551,8 +515,11 @@ static inline bool cpufreq_next_valid(struct cpufreq_frequency_table **pos) * @table: the cpufreq_frequency_table * to iterate over. */ -#define cpufreq_for_each_valid_entry(pos, table) \ - for (pos = table; cpufreq_next_valid(&pos); pos++) +#define cpufreq_for_each_valid_entry(pos, table) \ + for (pos = table; pos->frequency != CPUFREQ_TABLE_END; pos++) \ + if (pos->frequency == CPUFREQ_ENTRY_INVALID) \ + continue; \ + else int cpufreq_frequency_table_cpuinfo(struct cpufreq_policy *policy, struct cpufreq_frequency_table *table); diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index 0e95fcc..358076e 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -125,6 +125,16 @@ struct irqaction { extern irqreturn_t no_action(int cpl, void *dev_id); +/* + * If a (PCI) device interrupt is not connected we set dev->irq to + * IRQ_NOTCONNECTED. This causes request_irq() to fail with -ENOTCONN, so we + * can distingiush that case from other error returns. + * + * 0x80000000 is guaranteed to be outside the available range of interrupts + * and easy to distinguish from other possible incorrect values. + */ +#define IRQ_NOTCONNECTED (1U << 31) + extern int __must_check request_threaded_irq(unsigned int irq, irq_handler_t handler, irq_handler_t thread_fn, diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h index db21d39..49cd889 100644 --- a/include/linux/pm_domain.h +++ b/include/linux/pm_domain.h @@ -19,6 +19,8 @@ /* Defines used for the flags field in the struct generic_pm_domain */ #define GENPD_FLAG_PM_CLK (1U << 0) /* PM domain uses PM clk */ +#define GENPD_MAX_NUM_STATES 8 /* Number of possible low power states */ + enum gpd_status { GPD_STATE_ACTIVE = 0, /* PM domain is active */ GPD_STATE_POWER_OFF, /* PM domain is off */ @@ -37,6 +39,11 @@ struct gpd_dev_ops { bool (*active_wakeup)(struct device *dev); }; +struct genpd_power_state { + s64 power_off_latency_ns; + s64 power_on_latency_ns; +}; + struct generic_pm_domain { struct dev_pm_domain domain; /* PM domain operations */ struct list_head gpd_list_node; /* Node in the global PM domains list */ @@ -54,9 +61,7 @@ struct generic_pm_domain { unsigned int prepared_count; /* Suspend counter of prepared devices */ bool suspend_power_off; /* Power status before system suspend */ int (*power_off)(struct generic_pm_domain *domain); - s64 power_off_latency_ns; int (*power_on)(struct generic_pm_domain *domain); - s64 power_on_latency_ns; struct gpd_dev_ops dev_ops; s64 max_off_time_ns; /* Maximum allowed "suspended" time. */ bool max_off_time_changed; @@ -66,6 +71,10 @@ struct generic_pm_domain { void (*detach_dev)(struct generic_pm_domain *domain, struct device *dev); unsigned int flags; /* Bit field of configs for genpd */ + struct genpd_power_state states[GENPD_MAX_NUM_STATES]; + unsigned int state_count; /* number of states */ + unsigned int state_idx; /* state that genpd will go to when off */ + }; static inline struct generic_pm_domain *pd_to_genpd(struct dev_pm_domain *pd) diff --git a/include/linux/pm_opp.h b/include/linux/pm_opp.h index 95403d2..cccaf4a 100644 --- a/include/linux/pm_opp.h +++ b/include/linux/pm_opp.h @@ -34,6 +34,8 @@ bool dev_pm_opp_is_turbo(struct dev_pm_opp *opp); int dev_pm_opp_get_opp_count(struct device *dev); unsigned long dev_pm_opp_get_max_clock_latency(struct device *dev); +unsigned long dev_pm_opp_get_max_volt_latency(struct device *dev); +unsigned long dev_pm_opp_get_max_transition_latency(struct device *dev); struct dev_pm_opp *dev_pm_opp_get_suspend_opp(struct device *dev); struct dev_pm_opp *dev_pm_opp_find_freq_exact(struct device *dev, @@ -60,6 +62,9 @@ int dev_pm_opp_set_supported_hw(struct device *dev, const u32 *versions, void dev_pm_opp_put_supported_hw(struct device *dev); int dev_pm_opp_set_prop_name(struct device *dev, const char *name); void dev_pm_opp_put_prop_name(struct device *dev); +int dev_pm_opp_set_regulator(struct device *dev, const char *name); +void dev_pm_opp_put_regulator(struct device *dev); +int dev_pm_opp_set_rate(struct device *dev, unsigned long target_freq); #else static inline unsigned long dev_pm_opp_get_voltage(struct dev_pm_opp *opp) { @@ -86,6 +91,16 @@ static inline unsigned long dev_pm_opp_get_max_clock_latency(struct device *dev) return 0; } +static inline unsigned long dev_pm_opp_get_max_volt_latency(struct device *dev) +{ + return 0; +} + +static inline unsigned long dev_pm_opp_get_max_transition_latency(struct device *dev) +{ + return 0; +} + static inline struct dev_pm_opp *dev_pm_opp_get_suspend_opp(struct device *dev) { return NULL; @@ -151,6 +166,18 @@ static inline int dev_pm_opp_set_prop_name(struct device *dev, const char *name) static inline void dev_pm_opp_put_prop_name(struct device *dev) {} +static inline int dev_pm_opp_set_regulator(struct device *dev, const char *name) +{ + return -EINVAL; +} + +static inline void dev_pm_opp_put_regulator(struct device *dev) {} + +static inline int dev_pm_opp_set_rate(struct device *dev, unsigned long target_freq) +{ + return -EINVAL; +} + #endif /* CONFIG_PM_OPP */ #if defined(CONFIG_PM_OPP) && defined(CONFIG_OF) diff --git a/include/linux/sched.h b/include/linux/sched.h index c617ea1..eb7f2f8 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -3212,4 +3212,13 @@ static inline unsigned long rlimit_max(unsigned int limit) return task_rlimit_max(current, limit); } +#ifdef CONFIG_CPU_FREQ +struct update_util_data { + void (*func)(struct update_util_data *data, + u64 time, unsigned long util, unsigned long max); +}; + +void cpufreq_set_update_util_data(int cpu, struct update_util_data *data); +#endif /* CONFIG_CPU_FREQ */ + #endif diff --git a/include/trace/events/power.h b/include/trace/events/power.h index 284244e..19e5030 100644 --- a/include/trace/events/power.h +++ b/include/trace/events/power.h @@ -38,6 +38,28 @@ DEFINE_EVENT(cpu, cpu_idle, TP_ARGS(state, cpu_id) ); +TRACE_EVENT(powernv_throttle, + + TP_PROTO(int chip_id, const char *reason, int pmax), + + TP_ARGS(chip_id, reason, pmax), + + TP_STRUCT__entry( + __field(int, chip_id) + __string(reason, reason) + __field(int, pmax) + ), + + TP_fast_assign( + __entry->chip_id = chip_id; + __assign_str(reason, reason); + __entry->pmax = pmax; + ), + + TP_printk("Chip %d Pmax %d %s", __entry->chip_id, + __entry->pmax, __get_str(reason)) +); + TRACE_EVENT(pstate_sample, TP_PROTO(u32 core_busy, diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 3ddd229..64731e8 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -1607,6 +1607,9 @@ int request_threaded_irq(unsigned int irq, irq_handler_t handler, struct irq_desc *desc; int retval; + if (irq == IRQ_NOTCONNECTED) + return -ENOTCONN; + /* * Sanity-check: shared interrupts must pass in a real dev-ID, * otherwise we'll have trouble later trying to figure out @@ -1697,9 +1700,13 @@ EXPORT_SYMBOL(request_threaded_irq); int request_any_context_irq(unsigned int irq, irq_handler_t handler, unsigned long flags, const char *name, void *dev_id) { - struct irq_desc *desc = irq_to_desc(irq); + struct irq_desc *desc; int ret; + if (irq == IRQ_NOTCONNECTED) + return -ENOTCONN; + + desc = irq_to_desc(irq); if (!desc) return -EINVAL; diff --git a/kernel/power/process.c b/kernel/power/process.c index 564f786..df058be 100644 --- a/kernel/power/process.c +++ b/kernel/power/process.c @@ -30,13 +30,12 @@ static int try_to_freeze_tasks(bool user_only) unsigned long end_time; unsigned int todo; bool wq_busy = false; - struct timeval start, end; - u64 elapsed_msecs64; + ktime_t start, end, elapsed; unsigned int elapsed_msecs; bool wakeup = false; int sleep_usecs = USEC_PER_MSEC; - do_gettimeofday(&start); + start = ktime_get_boottime(); end_time = jiffies + msecs_to_jiffies(freeze_timeout_msecs); @@ -78,10 +77,9 @@ static int try_to_freeze_tasks(bool user_only) sleep_usecs *= 2; } - do_gettimeofday(&end); - elapsed_msecs64 = timeval_to_ns(&end) - timeval_to_ns(&start); - do_div(elapsed_msecs64, NSEC_PER_MSEC); - elapsed_msecs = elapsed_msecs64; + end = ktime_get_boottime(); + elapsed = ktime_sub(end, start); + elapsed_msecs = ktime_to_ms(elapsed); if (todo) { pr_cont("\n"); diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c index f9fe133..230a772 100644 --- a/kernel/power/suspend.c +++ b/kernel/power/suspend.c @@ -248,7 +248,7 @@ static int suspend_test(int level) { #ifdef CONFIG_PM_DEBUG if (pm_test_level == level) { - printk(KERN_INFO "suspend debug: Waiting for %d second(s).\n", + pr_info("suspend debug: Waiting for %d second(s).\n", pm_test_delay); mdelay(pm_test_delay * 1000); return 1; @@ -320,7 +320,7 @@ static int suspend_enter(suspend_state_t state, bool *wakeup) error = dpm_suspend_late(PMSG_SUSPEND); if (error) { - printk(KERN_ERR "PM: late suspend of devices failed\n"); + pr_err("PM: late suspend of devices failed\n"); goto Platform_finish; } error = platform_suspend_prepare_late(state); @@ -329,7 +329,7 @@ static int suspend_enter(suspend_state_t state, bool *wakeup) error = dpm_suspend_noirq(PMSG_SUSPEND); if (error) { - printk(KERN_ERR "PM: noirq suspend of devices failed\n"); + pr_err("PM: noirq suspend of devices failed\n"); goto Platform_early_resume; } error = platform_suspend_prepare_noirq(state); diff --git a/kernel/sched/Makefile b/kernel/sched/Makefile index 7d4cba2..302d6eb 100644 --- a/kernel/sched/Makefile +++ b/kernel/sched/Makefile @@ -19,3 +19,4 @@ obj-$(CONFIG_SCHED_AUTOGROUP) += auto_group.o obj-$(CONFIG_SCHEDSTATS) += stats.o obj-$(CONFIG_SCHED_DEBUG) += debug.o obj-$(CONFIG_CGROUP_CPUACCT) += cpuacct.o +obj-$(CONFIG_CPU_FREQ) += cpufreq.o diff --git a/kernel/sched/cpufreq.c b/kernel/sched/cpufreq.c new file mode 100644 index 0000000..928c4ba --- /dev/null +++ b/kernel/sched/cpufreq.c @@ -0,0 +1,37 @@ +/* + * Scheduler code and data structures related to cpufreq. + * + * Copyright (C) 2016, Intel Corporation + * Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com> + * + * 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 "sched.h" + +DEFINE_PER_CPU(struct update_util_data *, cpufreq_update_util_data); + +/** + * cpufreq_set_update_util_data - Populate the CPU's update_util_data pointer. + * @cpu: The CPU to set the pointer for. + * @data: New pointer value. + * + * Set and publish the update_util_data pointer for the given CPU. That pointer + * points to a struct update_util_data object containing a callback function + * to call from cpufreq_update_util(). That function will be called from an RCU + * read-side critical section, so it must not sleep. + * + * Callers must use RCU-sched callbacks to free any memory that might be + * accessed via the old update_util_data pointer or invoke synchronize_sched() + * right after this function to avoid use-after-free. + */ +void cpufreq_set_update_util_data(int cpu, struct update_util_data *data) +{ + if (WARN_ON(data && !data->func)) + return; + + rcu_assign_pointer(per_cpu(cpufreq_update_util_data, cpu), data); +} +EXPORT_SYMBOL_GPL(cpufreq_set_update_util_data); diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c index c7a036f..affd97e 100644 --- a/kernel/sched/deadline.c +++ b/kernel/sched/deadline.c @@ -717,6 +717,10 @@ static void update_curr_dl(struct rq *rq) if (!dl_task(curr) || !on_dl_rq(dl_se)) return; + /* Kick cpufreq (see the comment in linux/cpufreq.h). */ + if (cpu_of(rq) == smp_processor_id()) + cpufreq_trigger_update(rq_clock(rq)); + /* * Consumed budget is computed considering the time as * observed by schedulable tasks (excluding time spent diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 3313052..46d64e4 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -2856,7 +2856,8 @@ static inline void update_load_avg(struct sched_entity *se, int update_tg) { struct cfs_rq *cfs_rq = cfs_rq_of(se); u64 now = cfs_rq_clock_task(cfs_rq); - int cpu = cpu_of(rq_of(cfs_rq)); + struct rq *rq = rq_of(cfs_rq); + int cpu = cpu_of(rq); /* * Track task load average for carrying it to new CPU after migrated, and @@ -2868,6 +2869,29 @@ static inline void update_load_avg(struct sched_entity *se, int update_tg) if (update_cfs_rq_load_avg(now, cfs_rq) && update_tg) update_tg_load_avg(cfs_rq, 0); + + if (cpu == smp_processor_id() && &rq->cfs == cfs_rq) { + unsigned long max = rq->cpu_capacity_orig; + + /* + * There are a few boundary cases this might miss but it should + * get called often enough that that should (hopefully) not be + * a real problem -- added to that it only calls on the local + * CPU, so if we enqueue remotely we'll miss an update, but + * the next tick/schedule should update. + * + * It will not get called when we go idle, because the idle + * thread is a different class (!fair), nor will the utilization + * number include things like RT tasks. + * + * As is, the util number is not freq-invariant (we'd have to + * implement arch_scale_freq_capacity() for that). + * + * See cpu_util(). + */ + cpufreq_update_util(rq_clock(rq), + min(cfs_rq->avg.util_avg, max), max); + } } static void attach_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se) diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 5624713..c41ea7a 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -953,6 +953,10 @@ static void update_curr_rt(struct rq *rq) if (curr->sched_class != &rt_sched_class) return; + /* Kick cpufreq (see the comment in linux/cpufreq.h). */ + if (cpu_of(rq) == smp_processor_id()) + cpufreq_trigger_update(rq_clock(rq)); + delta_exec = rq_clock_task(rq) - curr->se.exec_start; if (unlikely((s64)delta_exec <= 0)) return; diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index b2ff5a2..382848a 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -1793,3 +1793,51 @@ static inline u64 irq_time_read(int cpu) } #endif /* CONFIG_64BIT */ #endif /* CONFIG_IRQ_TIME_ACCOUNTING */ + +#ifdef CONFIG_CPU_FREQ +DECLARE_PER_CPU(struct update_util_data *, cpufreq_update_util_data); + +/** + * cpufreq_update_util - Take a note about CPU utilization changes. + * @time: Current time. + * @util: Current utilization. + * @max: Utilization ceiling. + * + * This function is called by the scheduler on every invocation of + * update_load_avg() on the CPU whose utilization is being updated. + * + * It can only be called from RCU-sched read-side critical sections. + */ +static inline void cpufreq_update_util(u64 time, unsigned long util, unsigned long max) +{ + struct update_util_data *data; + + data = rcu_dereference_sched(*this_cpu_ptr(&cpufreq_update_util_data)); + if (data) + data->func(data, time, util, max); +} + +/** + * cpufreq_trigger_update - Trigger CPU performance state evaluation if needed. + * @time: Current time. + * + * The way cpufreq is currently arranged requires it to evaluate the CPU + * performance state (frequency/voltage) on a regular basis to prevent it from + * being stuck in a completely inadequate performance level for too long. + * That is not guaranteed to happen if the updates are only triggered from CFS, + * though, because they may not be coming in if RT or deadline tasks are active + * all the time (or there are RT and DL tasks only). + * + * As a workaround for that issue, this function is called by the RT and DL + * sched classes to trigger extra cpufreq updates to prevent it from stalling, + * but that really is a band-aid. Going forward it should be replaced with + * solutions targeted more specifically at RT and DL tasks. + */ +static inline void cpufreq_trigger_update(u64 time) +{ + cpufreq_update_util(time, ULONG_MAX, 0); +} +#else +static inline void cpufreq_update_util(u64 time, unsigned long util, unsigned long max) {} +static inline void cpufreq_trigger_update(u64 time) {} +#endif /* CONFIG_CPU_FREQ */ diff --git a/kernel/trace/power-traces.c b/kernel/trace/power-traces.c index eb4220a..81b8745 100644 --- a/kernel/trace/power-traces.c +++ b/kernel/trace/power-traces.c @@ -15,4 +15,5 @@ EXPORT_TRACEPOINT_SYMBOL_GPL(suspend_resume); EXPORT_TRACEPOINT_SYMBOL_GPL(cpu_idle); +EXPORT_TRACEPOINT_SYMBOL_GPL(powernv_throttle); diff --git a/tools/power/x86/turbostat/turbostat.8 b/tools/power/x86/turbostat/turbostat.8 index 622db68..89a55d5 100644 --- a/tools/power/x86/turbostat/turbostat.8 +++ b/tools/power/x86/turbostat/turbostat.8 @@ -34,7 +34,10 @@ name as necessary to disambiguate it from others is necessary. Note that option \fB--debug\fP displays additional system configuration information. Invoking this parameter more than once may also enable internal turbostat debug information. .PP -\fB--interval seconds\fP overrides the default 5-second measurement interval. +\fB--interval seconds\fP overrides the default 5.0 second measurement interval. +.PP +\fB--out output_file\fP turbostat output is written to the specified output_file. +The file is truncated if it already exists, and it is created if it does not exist. .PP \fB--help\fP displays usage for the most common parameters. .PP @@ -61,7 +64,7 @@ displays the statistics gathered since it was forked. .nf \fBCPU\fP Linux CPU (logical processor) number. Yes, it is okay that on many systems the CPUs are not listed in numerical order -- for efficiency reasons, turbostat runs in topology order, so HT siblings appear together. \fBAVG_MHz\fP number of cycles executed divided by time elapsed. -\fB%Busy\fP percent of the interval that the CPU retired instructions, aka. % of time in "C0" state. +\fBBusy%\fP percent of the interval that the CPU retired instructions, aka. % of time in "C0" state. \fBBzy_MHz\fP average clock rate while the CPU was busy (in "c0" state). \fBTSC_MHz\fP average MHz that the TSC ran during the entire interval. .fi @@ -83,13 +86,14 @@ Note that multiple CPUs per core indicate support for Intel(R) Hyper-Threading T \fBRAM_%\fP percent of the interval that RAPL throttling was active on DRAM. .fi .PP -.SH EXAMPLE +.SH PERIODIC EXAMPLE Without any parameters, turbostat displays statistics ever 5 seconds. -(override interval with "-i sec" option, or specify a command -for turbostat to fork). +Periodic output goes to stdout, by default, unless --out is used to specify an output file. +The 5-second interval can be changed with th "-i sec" option. +Or a command may be specified as in "FORK EXAMPLE" below. .nf [root@hsw]# ./turbostat - CPU Avg_MHz %Busy Bzy_MHz TSC_MHz + CPU Avg_MHz Busy% Bzy_MHz TSC_MHz - 488 12.51 3898 3498 0 0 0.01 3885 3498 4 3897 99.99 3898 3498 @@ -145,7 +149,7 @@ cpu0: MSR_IA32_THERM_STATUS: 0x88340000 (48 C +/- 1) cpu1: MSR_IA32_THERM_STATUS: 0x88440000 (32 C +/- 1) cpu2: MSR_IA32_THERM_STATUS: 0x88450000 (31 C +/- 1) cpu3: MSR_IA32_THERM_STATUS: 0x88490000 (27 C +/- 1) - Core CPU Avg_MHz %Busy Bzy_MHz TSC_MHz SMI CPU%c1 CPU%c3 CPU%c6 CPU%c7 CoreTmp PkgTmp PkgWatt CorWatt GFXWatt + Core CPU Avg_MHz Busy% Bzy_MHz TSC_MHz SMI CPU%c1 CPU%c3 CPU%c6 CPU%c7 CoreTmp PkgTmp PkgWatt CorWatt GFXWatt - - 493 12.64 3898 3498 0 12.64 0.00 0.00 74.72 47 47 21.62 13.74 0.00 0 0 4 0.11 3894 3498 0 99.89 0.00 0.00 0.00 47 47 21.62 13.74 0.00 0 4 3897 99.98 3898 3498 0 0.02 @@ -171,14 +175,16 @@ The --debug option adds additional columns to the measurement ouput, including C See the field definitions above. .SH FORK EXAMPLE If turbostat is invoked with a command, it will fork that command -and output the statistics gathered when the command exits. +and output the statistics gathered after the command exits. +In this case, turbostat output goes to stderr, by default. +Output can instead be saved to a file using the --out option. eg. Here a cycle soaker is run on 1 CPU (see %c0) for a few seconds until ^C while the other CPUs are mostly idle: .nf root@hsw: turbostat cat /dev/zero > /dev/null ^C - CPU Avg_MHz %Busy Bzy_MHz TSC_MHz + CPU Avg_MHz Busy% Bzy_MHz TSC_MHz - 482 12.51 3854 3498 0 0 0.01 1960 3498 4 0 0.00 2128 3498 @@ -192,12 +198,12 @@ root@hsw: turbostat cat /dev/zero > /dev/null .fi Above the cycle soaker drives cpu5 up its 3.9 GHz turbo limit. -The first row shows the average MHz and %Busy across all the processors in the system. +The first row shows the average MHz and Busy% across all the processors in the system. Note that the Avg_MHz column reflects the total number of cycles executed -divided by the measurement interval. If the %Busy column is 100%, +divided by the measurement interval. If the Busy% column is 100%, then the processor was running at that speed the entire interval. -The Avg_MHz multiplied by the %Busy results in the Bzy_MHz -- +The Avg_MHz multiplied by the Busy% results in the Bzy_MHz -- which is the average frequency while the processor was executing -- not including any non-busy idle time. @@ -233,7 +239,7 @@ in the brand string in /proc/cpuinfo. On a system where the TSC stops in idle, TSC_MHz will drop below the processor's base frequency. -%Busy = MPERF_delta/TSC_delta +Busy% = MPERF_delta/TSC_delta Bzy_MHz = TSC_delta/APERF_delta/MPERF_delta/measurement_interval diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index 3fa94e2..20a257a 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -38,12 +38,15 @@ #include <string.h> #include <ctype.h> #include <sched.h> +#include <time.h> #include <cpuid.h> #include <linux/capability.h> #include <errno.h> char *proc_stat = "/proc/stat"; -unsigned int interval_sec = 5; +FILE *outf; +int *fd_percpu; +struct timespec interval_ts = {5, 0}; unsigned int debug; unsigned int rapl_joules; unsigned int summary_only; @@ -72,6 +75,7 @@ unsigned int extra_msr_offset64; unsigned int extra_delta_offset32; unsigned int extra_delta_offset64; unsigned int aperf_mperf_multiplier = 1; +int do_irq = 1; int do_smi; double bclk; double base_hz; @@ -86,6 +90,10 @@ char *output_buffer, *outp; unsigned int do_rapl; unsigned int do_dts; unsigned int do_ptm; +unsigned int do_gfx_rc6_ms; +unsigned long long gfx_cur_rc6_ms; +unsigned int do_gfx_mhz; +unsigned int gfx_cur_mhz; unsigned int tcc_activation_temp; unsigned int tcc_activation_temp_override; double rapl_power_units, rapl_time_units; @@ -98,6 +106,12 @@ unsigned int crystal_hz; unsigned long long tsc_hz; int base_cpu; double discover_bclk(unsigned int family, unsigned int model); +unsigned int has_hwp; /* IA32_PM_ENABLE, IA32_HWP_CAPABILITIES */ + /* IA32_HWP_REQUEST, IA32_HWP_STATUS */ +unsigned int has_hwp_notify; /* IA32_HWP_INTERRUPT */ +unsigned int has_hwp_activity_window; /* IA32_HWP_REQUEST[bits 41:32] */ +unsigned int has_hwp_epp; /* IA32_HWP_REQUEST[bits 31:24] */ +unsigned int has_hwp_pkg; /* IA32_HWP_REQUEST_PKG */ #define RAPL_PKG (1 << 0) /* 0x610 MSR_PKG_POWER_LIMIT */ @@ -145,6 +159,7 @@ struct thread_data { unsigned long long extra_delta64; unsigned long long extra_msr32; unsigned long long extra_delta32; + unsigned int irq_count; unsigned int smi_count; unsigned int cpu_id; unsigned int flags; @@ -172,6 +187,8 @@ struct pkg_data { unsigned long long pkg_any_core_c0; unsigned long long pkg_any_gfxe_c0; unsigned long long pkg_both_core_gfxe_c0; + unsigned long long gfx_rc6_ms; + unsigned int gfx_mhz; unsigned int package_id; unsigned int energy_pkg; /* MSR_PKG_ENERGY_STATUS */ unsigned int energy_dram; /* MSR_DRAM_ENERGY_STATUS */ @@ -212,6 +229,9 @@ struct topo_params { struct timeval tv_even, tv_odd, tv_delta; +int *irq_column_2_cpu; /* /proc/interrupts column numbers */ +int *irqs_per_cpu; /* indexed by cpu_num */ + void setup_all_buffers(void); int cpu_is_not_present(int cpu) @@ -262,23 +282,34 @@ int cpu_migrate(int cpu) else return 0; } - -int get_msr(int cpu, off_t offset, unsigned long long *msr) +int get_msr_fd(int cpu) { - ssize_t retval; char pathname[32]; int fd; + fd = fd_percpu[cpu]; + + if (fd) + return fd; + sprintf(pathname, "/dev/cpu/%d/msr", cpu); fd = open(pathname, O_RDONLY); if (fd < 0) err(-1, "%s open failed, try chown or chmod +r /dev/cpu/*/msr, or run as root", pathname); - retval = pread(fd, msr, sizeof *msr, offset); - close(fd); + fd_percpu[cpu] = fd; + + return fd; +} + +int get_msr(int cpu, off_t offset, unsigned long long *msr) +{ + ssize_t retval; + + retval = pread(get_msr_fd(cpu), msr, sizeof(*msr), offset); if (retval != sizeof *msr) - err(-1, "%s offset 0x%llx read failed", pathname, (unsigned long long)offset); + err(-1, "msr %d offset 0x%llx read failed", cpu, (unsigned long long)offset); return 0; } @@ -286,8 +317,8 @@ int get_msr(int cpu, off_t offset, unsigned long long *msr) /* * Example Format w/ field column widths: * - * Package Core CPU Avg_MHz Bzy_MHz TSC_MHz SMI %Busy CPU_%c1 CPU_%c3 CPU_%c6 CPU_%c7 CoreTmp PkgTmp Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 PkgWatt CorWatt GFXWatt - * 123456781234567812345678123456781234567812345678123456781234567812345678123456781234567812345678123456781234567812345678123456781234567812345678123456781234567812345678 + * Package Core CPU Avg_MHz Bzy_MHz TSC_MHz IRQ SMI Busy% CPU_%c1 CPU_%c3 CPU_%c6 CPU_%c7 CoreTmp PkgTmp GFXMHz Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 PkgWatt CorWatt GFXWatt + * 12345678123456781234567812345678123456781234567812345678123456781234567812345678123456781234567812345678123456781234567812345678123456781234567812345678123456781234567812345678 */ void print_header(void) @@ -301,7 +332,7 @@ void print_header(void) if (has_aperf) outp += sprintf(outp, " Avg_MHz"); if (has_aperf) - outp += sprintf(outp, " %%Busy"); + outp += sprintf(outp, " Busy%%"); if (has_aperf) outp += sprintf(outp, " Bzy_MHz"); outp += sprintf(outp, " TSC_MHz"); @@ -318,6 +349,8 @@ void print_header(void) if (!debug) goto done; + if (do_irq) + outp += sprintf(outp, " IRQ"); if (do_smi) outp += sprintf(outp, " SMI"); @@ -335,6 +368,12 @@ void print_header(void) if (do_ptm) outp += sprintf(outp, " PkgTmp"); + if (do_gfx_rc6_ms) + outp += sprintf(outp, " GFX%%rc6"); + + if (do_gfx_mhz) + outp += sprintf(outp, " GFXMHz"); + if (do_skl_residency) { outp += sprintf(outp, " Totl%%C0"); outp += sprintf(outp, " Any%%C0"); @@ -409,6 +448,8 @@ int dump_counters(struct thread_data *t, struct core_data *c, extra_msr_offset32, t->extra_msr32); outp += sprintf(outp, "msr0x%x: %016llX\n", extra_msr_offset64, t->extra_msr64); + if (do_irq) + outp += sprintf(outp, "IRQ: %08X\n", t->irq_count); if (do_smi) outp += sprintf(outp, "SMI: %08X\n", t->smi_count); } @@ -504,7 +545,7 @@ int format_counters(struct thread_data *t, struct core_data *c, outp += sprintf(outp, "%8.0f", 1.0 / units * t->aperf / interval_float); - /* %Busy */ + /* Busy% */ if (has_aperf) { if (!skip_c0) outp += sprintf(outp, "%8.2f", 100.0 * t->mperf/t->tsc/tsc_tweak); @@ -542,6 +583,10 @@ int format_counters(struct thread_data *t, struct core_data *c, if (!debug) goto done; + /* IRQ */ + if (do_irq) + outp += sprintf(outp, "%8d", t->irq_count); + /* SMI */ if (do_smi) outp += sprintf(outp, "%8d", t->smi_count); @@ -575,6 +620,14 @@ int format_counters(struct thread_data *t, struct core_data *c, if (do_ptm) outp += sprintf(outp, "%8d", p->pkg_temp_c); + /* GFXrc6 */ + if (do_gfx_rc6_ms) + outp += sprintf(outp, "%8.2f", 100.0 * p->gfx_rc6_ms / 1000.0 / interval_float); + + /* GFXMHz */ + if (do_gfx_mhz) + outp += sprintf(outp, "%8d", p->gfx_mhz); + /* Totl%C0, Any%C0 GFX%C0 CPUGFX% */ if (do_skl_residency) { outp += sprintf(outp, "%8.2f", 100.0 * p->pkg_wtd_core_c0/t->tsc); @@ -645,15 +698,24 @@ done: return 0; } -void flush_stdout() +void flush_output_stdout(void) { - fputs(output_buffer, stdout); - fflush(stdout); + FILE *filep; + + if (outf == stderr) + filep = stdout; + else + filep = outf; + + fputs(output_buffer, filep); + fflush(filep); + outp = output_buffer; } -void flush_stderr() +void flush_output_stderr(void) { - fputs(output_buffer, stderr); + fputs(output_buffer, outf); + fflush(outf); outp = output_buffer; } void format_all_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p) @@ -704,6 +766,9 @@ delta_package(struct pkg_data *new, struct pkg_data *old) old->pc10 = new->pc10 - old->pc10; old->pkg_temp_c = new->pkg_temp_c; + old->gfx_rc6_ms = new->gfx_rc6_ms - old->gfx_rc6_ms; + old->gfx_mhz = new->gfx_mhz; + DELTA_WRAP32(new->energy_pkg, old->energy_pkg); DELTA_WRAP32(new->energy_cores, old->energy_cores); DELTA_WRAP32(new->energy_gfx, old->energy_gfx); @@ -745,9 +810,9 @@ delta_thread(struct thread_data *new, struct thread_data *old, } else { if (!aperf_mperf_unstable) { - fprintf(stderr, "%s: APERF or MPERF went backwards *\n", progname); - fprintf(stderr, "* Frequency results do not cover entire interval *\n"); - fprintf(stderr, "* fix this by running Linux-2.6.30 or later *\n"); + fprintf(outf, "%s: APERF or MPERF went backwards *\n", progname); + fprintf(outf, "* Frequency results do not cover entire interval *\n"); + fprintf(outf, "* fix this by running Linux-2.6.30 or later *\n"); aperf_mperf_unstable = 1; } @@ -782,7 +847,8 @@ delta_thread(struct thread_data *new, struct thread_data *old, } if (old->mperf == 0) { - if (debug > 1) fprintf(stderr, "cpu%d MPERF 0!\n", old->cpu_id); + if (debug > 1) + fprintf(outf, "cpu%d MPERF 0!\n", old->cpu_id); old->mperf = 1; /* divide by 0 protection */ } @@ -797,6 +863,9 @@ delta_thread(struct thread_data *new, struct thread_data *old, old->extra_msr32 = new->extra_msr32; old->extra_msr64 = new->extra_msr64; + if (do_irq) + old->irq_count = new->irq_count - old->irq_count; + if (do_smi) old->smi_count = new->smi_count - old->smi_count; } @@ -826,10 +895,12 @@ void clear_counters(struct thread_data *t, struct core_data *c, struct pkg_data t->mperf = 0; t->c1 = 0; - t->smi_count = 0; t->extra_delta32 = 0; t->extra_delta64 = 0; + t->irq_count = 0; + t->smi_count = 0; + /* tells format_counters to dump all fields from this set */ t->flags = CPU_IS_FIRST_THREAD_IN_CORE | CPU_IS_FIRST_CORE_IN_PACKAGE; @@ -861,6 +932,9 @@ void clear_counters(struct thread_data *t, struct core_data *c, struct pkg_data p->rapl_pkg_perf_status = 0; p->rapl_dram_perf_status = 0; p->pkg_temp_c = 0; + + p->gfx_rc6_ms = 0; + p->gfx_mhz = 0; } int sum_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p) @@ -873,6 +947,9 @@ int sum_counters(struct thread_data *t, struct core_data *c, average.threads.extra_delta32 += t->extra_delta32; average.threads.extra_delta64 += t->extra_delta64; + average.threads.irq_count += t->irq_count; + average.threads.smi_count += t->smi_count; + /* sum per-core values only for 1st thread in core */ if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE)) return 0; @@ -910,6 +987,9 @@ int sum_counters(struct thread_data *t, struct core_data *c, average.packages.energy_cores += p->energy_cores; average.packages.energy_gfx += p->energy_gfx; + average.packages.gfx_rc6_ms = p->gfx_rc6_ms; + average.packages.gfx_mhz = p->gfx_mhz; + average.packages.pkg_temp_c = MAX(average.packages.pkg_temp_c, p->pkg_temp_c); average.packages.rapl_pkg_perf_status += p->rapl_pkg_perf_status; @@ -970,7 +1050,6 @@ static unsigned long long rdtsc(void) return low | ((unsigned long long)high) << 32; } - /* * get_counters(...) * migrate to cpu @@ -980,23 +1059,74 @@ int get_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p) { int cpu = t->cpu_id; unsigned long long msr; + int aperf_mperf_retry_count = 0; if (cpu_migrate(cpu)) { - fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + fprintf(outf, "Could not migrate to CPU %d\n", cpu); return -1; } +retry: t->tsc = rdtsc(); /* we are running on local CPU of interest */ if (has_aperf) { + unsigned long long tsc_before, tsc_between, tsc_after, aperf_time, mperf_time; + + /* + * The TSC, APERF and MPERF must be read together for + * APERF/MPERF and MPERF/TSC to give accurate results. + * + * Unfortunately, APERF and MPERF are read by + * individual system call, so delays may occur + * between them. If the time to read them + * varies by a large amount, we re-read them. + */ + + /* + * This initial dummy APERF read has been seen to + * reduce jitter in the subsequent reads. + */ + + if (get_msr(cpu, MSR_IA32_APERF, &t->aperf)) + return -3; + + t->tsc = rdtsc(); /* re-read close to APERF */ + + tsc_before = t->tsc; + if (get_msr(cpu, MSR_IA32_APERF, &t->aperf)) return -3; + + tsc_between = rdtsc(); + if (get_msr(cpu, MSR_IA32_MPERF, &t->mperf)) return -4; + + tsc_after = rdtsc(); + + aperf_time = tsc_between - tsc_before; + mperf_time = tsc_after - tsc_between; + + /* + * If the system call latency to read APERF and MPERF + * differ by more than 2x, then try again. + */ + if ((aperf_time > (2 * mperf_time)) || (mperf_time > (2 * aperf_time))) { + aperf_mperf_retry_count++; + if (aperf_mperf_retry_count < 5) + goto retry; + else + warnx("cpu%d jitter %lld %lld", + cpu, aperf_time, mperf_time); + } + aperf_mperf_retry_count = 0; + t->aperf = t->aperf * aperf_mperf_multiplier; t->mperf = t->mperf * aperf_mperf_multiplier; } + if (do_irq) + t->irq_count = irqs_per_cpu[cpu]; if (do_smi) { if (get_msr(cpu, MSR_SMI_COUNT, &msr)) return -5; @@ -1124,6 +1254,13 @@ int get_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p) return -17; p->pkg_temp_c = tcc_activation_temp - ((msr >> 16) & 0x7F); } + + if (do_gfx_rc6_ms) + p->gfx_rc6_ms = gfx_cur_rc6_ms; + + if (do_gfx_mhz) + p->gfx_mhz = gfx_cur_mhz; + return 0; } @@ -1175,18 +1312,18 @@ dump_nhm_platform_info(void) get_msr(base_cpu, MSR_PLATFORM_INFO, &msr); - fprintf(stderr, "cpu%d: MSR_PLATFORM_INFO: 0x%08llx\n", base_cpu, msr); + fprintf(outf, "cpu%d: MSR_PLATFORM_INFO: 0x%08llx\n", base_cpu, msr); ratio = (msr >> 40) & 0xFF; - fprintf(stderr, "%d * %.0f = %.0f MHz max efficiency frequency\n", + fprintf(outf, "%d * %.0f = %.0f MHz max efficiency frequency\n", ratio, bclk, ratio * bclk); ratio = (msr >> 8) & 0xFF; - fprintf(stderr, "%d * %.0f = %.0f MHz base frequency\n", + fprintf(outf, "%d * %.0f = %.0f MHz base frequency\n", ratio, bclk, ratio * bclk); get_msr(base_cpu, MSR_IA32_POWER_CTL, &msr); - fprintf(stderr, "cpu%d: MSR_IA32_POWER_CTL: 0x%08llx (C1E auto-promotion: %sabled)\n", + fprintf(outf, "cpu%d: MSR_IA32_POWER_CTL: 0x%08llx (C1E auto-promotion: %sabled)\n", base_cpu, msr, msr & 0x2 ? "EN" : "DIS"); return; @@ -1200,16 +1337,16 @@ dump_hsw_turbo_ratio_limits(void) get_msr(base_cpu, MSR_TURBO_RATIO_LIMIT2, &msr); - fprintf(stderr, "cpu%d: MSR_TURBO_RATIO_LIMIT2: 0x%08llx\n", base_cpu, msr); + fprintf(outf, "cpu%d: MSR_TURBO_RATIO_LIMIT2: 0x%08llx\n", base_cpu, msr); ratio = (msr >> 8) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 18 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 18 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 0) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 17 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 17 active cores\n", ratio, bclk, ratio * bclk); return; } @@ -1222,46 +1359,46 @@ dump_ivt_turbo_ratio_limits(void) get_msr(base_cpu, MSR_TURBO_RATIO_LIMIT1, &msr); - fprintf(stderr, "cpu%d: MSR_TURBO_RATIO_LIMIT1: 0x%08llx\n", base_cpu, msr); + fprintf(outf, "cpu%d: MSR_TURBO_RATIO_LIMIT1: 0x%08llx\n", base_cpu, msr); ratio = (msr >> 56) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 16 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 16 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 48) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 15 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 15 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 40) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 14 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 14 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 32) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 13 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 13 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 24) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 12 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 12 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 16) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 11 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 11 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 8) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 10 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 10 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 0) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 9 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 9 active cores\n", ratio, bclk, ratio * bclk); return; } @@ -1274,46 +1411,46 @@ dump_nhm_turbo_ratio_limits(void) get_msr(base_cpu, MSR_TURBO_RATIO_LIMIT, &msr); - fprintf(stderr, "cpu%d: MSR_TURBO_RATIO_LIMIT: 0x%08llx\n", base_cpu, msr); + fprintf(outf, "cpu%d: MSR_TURBO_RATIO_LIMIT: 0x%08llx\n", base_cpu, msr); ratio = (msr >> 56) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 8 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 8 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 48) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 7 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 7 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 40) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 6 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 6 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 32) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 5 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 5 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 24) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 4 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 4 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 16) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 3 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 3 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 8) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 2 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 2 active cores\n", ratio, bclk, ratio * bclk); ratio = (msr >> 0) & 0xFF; if (ratio) - fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 1 active cores\n", + fprintf(outf, "%d * %.0f = %.0f MHz max turbo 1 active cores\n", ratio, bclk, ratio * bclk); return; } @@ -1321,21 +1458,23 @@ dump_nhm_turbo_ratio_limits(void) static void dump_knl_turbo_ratio_limits(void) { - int cores; - unsigned int ratio; + const unsigned int buckets_no = 7; + unsigned long long msr; - int delta_cores; - int delta_ratio; - int i; + int delta_cores, delta_ratio; + int i, b_nr; + unsigned int cores[buckets_no]; + unsigned int ratio[buckets_no]; get_msr(base_cpu, MSR_NHM_TURBO_RATIO_LIMIT, &msr); - fprintf(stderr, "cpu%d: MSR_NHM_TURBO_RATIO_LIMIT: 0x%08llx\n", + fprintf(outf, "cpu%d: MSR_TURBO_RATIO_LIMIT: 0x%08llx\n", base_cpu, msr); /** * Turbo encoding in KNL is as follows: - * [7:0] -- Base value of number of active cores of bucket 1. + * [0] -- Reserved + * [7:1] -- Base value of number of active cores of bucket 1. * [15:8] -- Base value of freq ratio of bucket 1. * [20:16] -- +ve delta of number of active cores of bucket 2. * i.e. active cores of bucket 2 = @@ -1354,29 +1493,25 @@ dump_knl_turbo_ratio_limits(void) * [60:56]-- +ve delta of number of active cores of bucket 7. * [63:61]-- -ve delta of freq ratio of bucket 7. */ - cores = msr & 0xFF; - ratio = (msr >> 8) && 0xFF; - if (ratio > 0) - fprintf(stderr, - "%d * %.0f = %.0f MHz max turbo %d active cores\n", - ratio, bclk, ratio * bclk, cores); - - for (i = 16; i < 64; i = i + 8) { + + b_nr = 0; + cores[b_nr] = (msr & 0xFF) >> 1; + ratio[b_nr] = (msr >> 8) & 0xFF; + + for (i = 16; i < 64; i += 8) { delta_cores = (msr >> i) & 0x1F; - delta_ratio = (msr >> (i + 5)) && 0x7; - if (!delta_cores || !delta_ratio) - return; - cores = cores + delta_cores; - ratio = ratio - delta_ratio; - - /** -ve ratios will make successive ratio calculations - * negative. Hence return instead of carrying on. - */ - if (ratio > 0) - fprintf(stderr, - "%d * %.0f = %.0f MHz max turbo %d active cores\n", - ratio, bclk, ratio * bclk, cores); + delta_ratio = (msr >> (i + 5)) & 0x7; + + cores[b_nr + 1] = cores[b_nr] + delta_cores; + ratio[b_nr + 1] = ratio[b_nr] - delta_ratio; + b_nr++; } + + for (i = buckets_no - 1; i >= 0; i--) + if (i > 0 ? ratio[i] != ratio[i - 1] : 1) + fprintf(outf, + "%d * %.0f = %.0f MHz max turbo %d active cores\n", + ratio[i], bclk, ratio[i] * bclk, cores[i]); } static void @@ -1389,15 +1524,15 @@ dump_nhm_cst_cfg(void) #define SNB_C1_AUTO_UNDEMOTE (1UL << 27) #define SNB_C3_AUTO_UNDEMOTE (1UL << 28) - fprintf(stderr, "cpu%d: MSR_NHM_SNB_PKG_CST_CFG_CTL: 0x%08llx", base_cpu, msr); + fprintf(outf, "cpu%d: MSR_NHM_SNB_PKG_CST_CFG_CTL: 0x%08llx", base_cpu, msr); - fprintf(stderr, " (%s%s%s%s%slocked: pkg-cstate-limit=%d: %s)\n", + fprintf(outf, " (%s%s%s%s%slocked: pkg-cstate-limit=%d: %s)\n", (msr & SNB_C3_AUTO_UNDEMOTE) ? "UNdemote-C3, " : "", (msr & SNB_C1_AUTO_UNDEMOTE) ? "UNdemote-C1, " : "", (msr & NHM_C3_AUTO_DEMOTE) ? "demote-C3, " : "", (msr & NHM_C1_AUTO_DEMOTE) ? "demote-C1, " : "", (msr & (1 << 15)) ? "" : "UN", - (unsigned int)msr & 7, + (unsigned int)msr & 0xF, pkg_cstate_limit_strings[pkg_cstate_limit]); return; } @@ -1408,48 +1543,59 @@ dump_config_tdp(void) unsigned long long msr; get_msr(base_cpu, MSR_CONFIG_TDP_NOMINAL, &msr); - fprintf(stderr, "cpu%d: MSR_CONFIG_TDP_NOMINAL: 0x%08llx", base_cpu, msr); - fprintf(stderr, " (base_ratio=%d)\n", (unsigned int)msr & 0xEF); + fprintf(outf, "cpu%d: MSR_CONFIG_TDP_NOMINAL: 0x%08llx", base_cpu, msr); + fprintf(outf, " (base_ratio=%d)\n", (unsigned int)msr & 0xFF); get_msr(base_cpu, MSR_CONFIG_TDP_LEVEL_1, &msr); - fprintf(stderr, "cpu%d: MSR_CONFIG_TDP_LEVEL_1: 0x%08llx (", base_cpu, msr); + fprintf(outf, "cpu%d: MSR_CONFIG_TDP_LEVEL_1: 0x%08llx (", base_cpu, msr); if (msr) { - fprintf(stderr, "PKG_MIN_PWR_LVL1=%d ", (unsigned int)(msr >> 48) & 0xEFFF); - fprintf(stderr, "PKG_MAX_PWR_LVL1=%d ", (unsigned int)(msr >> 32) & 0xEFFF); - fprintf(stderr, "LVL1_RATIO=%d ", (unsigned int)(msr >> 16) & 0xEF); - fprintf(stderr, "PKG_TDP_LVL1=%d", (unsigned int)(msr) & 0xEFFF); + fprintf(outf, "PKG_MIN_PWR_LVL1=%d ", (unsigned int)(msr >> 48) & 0x7FFF); + fprintf(outf, "PKG_MAX_PWR_LVL1=%d ", (unsigned int)(msr >> 32) & 0x7FFF); + fprintf(outf, "LVL1_RATIO=%d ", (unsigned int)(msr >> 16) & 0xFF); + fprintf(outf, "PKG_TDP_LVL1=%d", (unsigned int)(msr) & 0x7FFF); } - fprintf(stderr, ")\n"); + fprintf(outf, ")\n"); get_msr(base_cpu, MSR_CONFIG_TDP_LEVEL_2, &msr); - fprintf(stderr, "cpu%d: MSR_CONFIG_TDP_LEVEL_2: 0x%08llx (", base_cpu, msr); + fprintf(outf, "cpu%d: MSR_CONFIG_TDP_LEVEL_2: 0x%08llx (", base_cpu, msr); if (msr) { - fprintf(stderr, "PKG_MIN_PWR_LVL2=%d ", (unsigned int)(msr >> 48) & 0xEFFF); - fprintf(stderr, "PKG_MAX_PWR_LVL2=%d ", (unsigned int)(msr >> 32) & 0xEFFF); - fprintf(stderr, "LVL2_RATIO=%d ", (unsigned int)(msr >> 16) & 0xEF); - fprintf(stderr, "PKG_TDP_LVL2=%d", (unsigned int)(msr) & 0xEFFF); + fprintf(outf, "PKG_MIN_PWR_LVL2=%d ", (unsigned int)(msr >> 48) & 0x7FFF); + fprintf(outf, "PKG_MAX_PWR_LVL2=%d ", (unsigned int)(msr >> 32) & 0x7FFF); + fprintf(outf, "LVL2_RATIO=%d ", (unsigned int)(msr >> 16) & 0xFF); + fprintf(outf, "PKG_TDP_LVL2=%d", (unsigned int)(msr) & 0x7FFF); } - fprintf(stderr, ")\n"); + fprintf(outf, ")\n"); get_msr(base_cpu, MSR_CONFIG_TDP_CONTROL, &msr); - fprintf(stderr, "cpu%d: MSR_CONFIG_TDP_CONTROL: 0x%08llx (", base_cpu, msr); + fprintf(outf, "cpu%d: MSR_CONFIG_TDP_CONTROL: 0x%08llx (", base_cpu, msr); if ((msr) & 0x3) - fprintf(stderr, "TDP_LEVEL=%d ", (unsigned int)(msr) & 0x3); - fprintf(stderr, " lock=%d", (unsigned int)(msr >> 31) & 1); - fprintf(stderr, ")\n"); - + fprintf(outf, "TDP_LEVEL=%d ", (unsigned int)(msr) & 0x3); + fprintf(outf, " lock=%d", (unsigned int)(msr >> 31) & 1); + fprintf(outf, ")\n"); + get_msr(base_cpu, MSR_TURBO_ACTIVATION_RATIO, &msr); - fprintf(stderr, "cpu%d: MSR_TURBO_ACTIVATION_RATIO: 0x%08llx (", base_cpu, msr); - fprintf(stderr, "MAX_NON_TURBO_RATIO=%d", (unsigned int)(msr) & 0x7F); - fprintf(stderr, " lock=%d", (unsigned int)(msr >> 31) & 1); - fprintf(stderr, ")\n"); + fprintf(outf, "cpu%d: MSR_TURBO_ACTIVATION_RATIO: 0x%08llx (", base_cpu, msr); + fprintf(outf, "MAX_NON_TURBO_RATIO=%d", (unsigned int)(msr) & 0xFF); + fprintf(outf, " lock=%d", (unsigned int)(msr >> 31) & 1); + fprintf(outf, ")\n"); +} +void free_fd_percpu(void) +{ + int i; + + for (i = 0; i < topo.max_cpu_num; ++i) { + if (fd_percpu[i] != 0) + close(fd_percpu[i]); + } + + free(fd_percpu); } void free_all_buffers(void) { CPU_FREE(cpu_present_set); cpu_present_set = NULL; - cpu_present_set = 0; + cpu_present_setsize = 0; CPU_FREE(cpu_affinity_set); cpu_affinity_set = NULL; @@ -1474,6 +1620,11 @@ void free_all_buffers(void) free(output_buffer); output_buffer = NULL; outp = NULL; + + free_fd_percpu(); + + free(irq_column_2_cpu); + free(irqs_per_cpu); } /* @@ -1481,7 +1632,7 @@ void free_all_buffers(void) */ FILE *fopen_or_die(const char *path, const char *mode) { - FILE *filep = fopen(path, "r"); + FILE *filep = fopen(path, mode); if (!filep) err(1, "%s: open failed", path); return filep; @@ -1696,6 +1847,136 @@ int mark_cpu_present(int cpu) return 0; } +/* + * snapshot_proc_interrupts() + * + * read and record summary of /proc/interrupts + * + * return 1 if config change requires a restart, else return 0 + */ +int snapshot_proc_interrupts(void) +{ + static FILE *fp; + int column, retval; + + if (fp == NULL) + fp = fopen_or_die("/proc/interrupts", "r"); + else + rewind(fp); + + /* read 1st line of /proc/interrupts to get cpu* name for each column */ + for (column = 0; column < topo.num_cpus; ++column) { + int cpu_number; + + retval = fscanf(fp, " CPU%d", &cpu_number); + if (retval != 1) + break; + + if (cpu_number > topo.max_cpu_num) { + warn("/proc/interrupts: cpu%d: > %d", cpu_number, topo.max_cpu_num); + return 1; + } + + irq_column_2_cpu[column] = cpu_number; + irqs_per_cpu[cpu_number] = 0; + } + + /* read /proc/interrupt count lines and sum up irqs per cpu */ + while (1) { + int column; + char buf[64]; + + retval = fscanf(fp, " %s:", buf); /* flush irq# "N:" */ + if (retval != 1) + break; + + /* read the count per cpu */ + for (column = 0; column < topo.num_cpus; ++column) { + + int cpu_number, irq_count; + + retval = fscanf(fp, " %d", &irq_count); + if (retval != 1) + break; + + cpu_number = irq_column_2_cpu[column]; + irqs_per_cpu[cpu_number] += irq_count; + + } + + while (getc(fp) != '\n') + ; /* flush interrupt description */ + + } + return 0; +} +/* + * snapshot_gfx_rc6_ms() + * + * record snapshot of + * /sys/class/drm/card0/power/rc6_residency_ms + * + * return 1 if config change requires a restart, else return 0 + */ +int snapshot_gfx_rc6_ms(void) +{ + FILE *fp; + int retval; + + fp = fopen_or_die("/sys/class/drm/card0/power/rc6_residency_ms", "r"); + + retval = fscanf(fp, "%lld", &gfx_cur_rc6_ms); + if (retval != 1) + err(1, "GFX rc6"); + + fclose(fp); + + return 0; +} +/* + * snapshot_gfx_mhz() + * + * record snapshot of + * /sys/class/graphics/fb0/device/drm/card0/gt_cur_freq_mhz + * + * return 1 if config change requires a restart, else return 0 + */ +int snapshot_gfx_mhz(void) +{ + static FILE *fp; + int retval; + + if (fp == NULL) + fp = fopen_or_die("/sys/class/graphics/fb0/device/drm/card0/gt_cur_freq_mhz", "r"); + else + rewind(fp); + + retval = fscanf(fp, "%d", &gfx_cur_mhz); + if (retval != 1) + err(1, "GFX MHz"); + + return 0; +} + +/* + * snapshot /proc and /sys files + * + * return 1 if configuration restart needed, else return 0 + */ +int snapshot_proc_sysfs_files(void) +{ + if (snapshot_proc_interrupts()) + return 1; + + if (do_gfx_rc6_ms) + snapshot_gfx_rc6_ms(); + + if (do_gfx_mhz) + snapshot_gfx_mhz(); + + return 0; +} + void turbostat_loop() { int retval; @@ -1704,6 +1985,7 @@ void turbostat_loop() restart: restarted++; + snapshot_proc_sysfs_files(); retval = for_all_cpus(get_counters, EVEN_COUNTERS); if (retval < -1) { exit(retval); @@ -1722,7 +2004,9 @@ restart: re_initialize(); goto restart; } - sleep(interval_sec); + nanosleep(&interval_ts, NULL); + if (snapshot_proc_sysfs_files()) + goto restart; retval = for_all_cpus(get_counters, ODD_COUNTERS); if (retval < -1) { exit(retval); @@ -1735,8 +2019,10 @@ restart: for_all_cpus_2(delta_cpu, ODD_COUNTERS, EVEN_COUNTERS); compute_average(EVEN_COUNTERS); format_all_counters(EVEN_COUNTERS); - flush_stdout(); - sleep(interval_sec); + flush_output_stdout(); + nanosleep(&interval_ts, NULL); + if (snapshot_proc_sysfs_files()) + goto restart; retval = for_all_cpus(get_counters, EVEN_COUNTERS); if (retval < -1) { exit(retval); @@ -1749,7 +2035,7 @@ restart: for_all_cpus_2(delta_cpu, EVEN_COUNTERS, ODD_COUNTERS); compute_average(ODD_COUNTERS); format_all_counters(ODD_COUNTERS); - flush_stdout(); + flush_output_stdout(); } } @@ -1889,6 +2175,7 @@ int has_nhm_turbo_ratio_limit(unsigned int family, unsigned int model) /* Nehalem compatible, but do not include turbo-ratio limit support */ case 0x2E: /* Nehalem-EX Xeon - Beckton */ case 0x2F: /* Westmere-EX Xeon - Eagleton */ + case 0x57: /* PHI - Knights Landing (different MSR definition) */ return 0; default: return 1; @@ -2016,7 +2303,7 @@ int print_epb(struct thread_data *t, struct core_data *c, struct pkg_data *p) return 0; if (cpu_migrate(cpu)) { - fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + fprintf(outf, "Could not migrate to CPU %d\n", cpu); return -1; } @@ -2037,7 +2324,98 @@ int print_epb(struct thread_data *t, struct core_data *c, struct pkg_data *p) epb_string = "custom"; break; } - fprintf(stderr, "cpu%d: MSR_IA32_ENERGY_PERF_BIAS: 0x%08llx (%s)\n", cpu, msr, epb_string); + fprintf(outf, "cpu%d: MSR_IA32_ENERGY_PERF_BIAS: 0x%08llx (%s)\n", cpu, msr, epb_string); + + return 0; +} +/* + * print_hwp() + * Decode the MSR_HWP_CAPABILITIES + */ +int print_hwp(struct thread_data *t, struct core_data *c, struct pkg_data *p) +{ + unsigned long long msr; + int cpu; + + if (!has_hwp) + return 0; + + cpu = t->cpu_id; + + /* MSR_HWP_CAPABILITIES is per-package */ + if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE) || !(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) + return 0; + + if (cpu_migrate(cpu)) { + fprintf(outf, "Could not migrate to CPU %d\n", cpu); + return -1; + } + + if (get_msr(cpu, MSR_PM_ENABLE, &msr)) + return 0; + + fprintf(outf, "cpu%d: MSR_PM_ENABLE: 0x%08llx (%sHWP)\n", + cpu, msr, (msr & (1 << 0)) ? "" : "No-"); + + /* MSR_PM_ENABLE[1] == 1 if HWP is enabled and MSRs visible */ + if ((msr & (1 << 0)) == 0) + return 0; + + if (get_msr(cpu, MSR_HWP_CAPABILITIES, &msr)) + return 0; + + fprintf(outf, "cpu%d: MSR_HWP_CAPABILITIES: 0x%08llx " + "(high 0x%x guar 0x%x eff 0x%x low 0x%x)\n", + cpu, msr, + (unsigned int)HWP_HIGHEST_PERF(msr), + (unsigned int)HWP_GUARANTEED_PERF(msr), + (unsigned int)HWP_MOSTEFFICIENT_PERF(msr), + (unsigned int)HWP_LOWEST_PERF(msr)); + + if (get_msr(cpu, MSR_HWP_REQUEST, &msr)) + return 0; + + fprintf(outf, "cpu%d: MSR_HWP_REQUEST: 0x%08llx " + "(min 0x%x max 0x%x des 0x%x epp 0x%x window 0x%x pkg 0x%x)\n", + cpu, msr, + (unsigned int)(((msr) >> 0) & 0xff), + (unsigned int)(((msr) >> 8) & 0xff), + (unsigned int)(((msr) >> 16) & 0xff), + (unsigned int)(((msr) >> 24) & 0xff), + (unsigned int)(((msr) >> 32) & 0xff3), + (unsigned int)(((msr) >> 42) & 0x1)); + + if (has_hwp_pkg) { + if (get_msr(cpu, MSR_HWP_REQUEST_PKG, &msr)) + return 0; + + fprintf(outf, "cpu%d: MSR_HWP_REQUEST_PKG: 0x%08llx " + "(min 0x%x max 0x%x des 0x%x epp 0x%x window 0x%x)\n", + cpu, msr, + (unsigned int)(((msr) >> 0) & 0xff), + (unsigned int)(((msr) >> 8) & 0xff), + (unsigned int)(((msr) >> 16) & 0xff), + (unsigned int)(((msr) >> 24) & 0xff), + (unsigned int)(((msr) >> 32) & 0xff3)); + } + if (has_hwp_notify) { + if (get_msr(cpu, MSR_HWP_INTERRUPT, &msr)) + return 0; + + fprintf(outf, "cpu%d: MSR_HWP_INTERRUPT: 0x%08llx " + "(%s_Guaranteed_Perf_Change, %s_Excursion_Min)\n", + cpu, msr, + ((msr) & 0x1) ? "EN" : "Dis", + ((msr) & 0x2) ? "EN" : "Dis"); + } + if (get_msr(cpu, MSR_HWP_STATUS, &msr)) + return 0; + + fprintf(outf, "cpu%d: MSR_HWP_STATUS: 0x%08llx " + "(%sGuaranteed_Perf_Change, %sExcursion_Min)\n", + cpu, msr, + ((msr) & 0x1) ? "" : "No-", + ((msr) & 0x2) ? "" : "No-"); return 0; } @@ -2057,14 +2435,14 @@ int print_perf_limit(struct thread_data *t, struct core_data *c, struct pkg_data return 0; if (cpu_migrate(cpu)) { - fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + fprintf(outf, "Could not migrate to CPU %d\n", cpu); return -1; } if (do_core_perf_limit_reasons) { get_msr(cpu, MSR_CORE_PERF_LIMIT_REASONS, &msr); - fprintf(stderr, "cpu%d: MSR_CORE_PERF_LIMIT_REASONS, 0x%08llx", cpu, msr); - fprintf(stderr, " (Active: %s%s%s%s%s%s%s%s%s%s%s%s%s%s)", + fprintf(outf, "cpu%d: MSR_CORE_PERF_LIMIT_REASONS, 0x%08llx", cpu, msr); + fprintf(outf, " (Active: %s%s%s%s%s%s%s%s%s%s%s%s%s%s)", (msr & 1 << 15) ? "bit15, " : "", (msr & 1 << 14) ? "bit14, " : "", (msr & 1 << 13) ? "Transitions, " : "", @@ -2079,7 +2457,7 @@ int print_perf_limit(struct thread_data *t, struct core_data *c, struct pkg_data (msr & 1 << 2) ? "bit2, " : "", (msr & 1 << 1) ? "ThermStatus, " : "", (msr & 1 << 0) ? "PROCHOT, " : ""); - fprintf(stderr, " (Logged: %s%s%s%s%s%s%s%s%s%s%s%s%s%s)\n", + fprintf(outf, " (Logged: %s%s%s%s%s%s%s%s%s%s%s%s%s%s)\n", (msr & 1 << 31) ? "bit31, " : "", (msr & 1 << 30) ? "bit30, " : "", (msr & 1 << 29) ? "Transitions, " : "", @@ -2098,8 +2476,8 @@ int print_perf_limit(struct thread_data *t, struct core_data *c, struct pkg_data } if (do_gfx_perf_limit_reasons) { get_msr(cpu, MSR_GFX_PERF_LIMIT_REASONS, &msr); - fprintf(stderr, "cpu%d: MSR_GFX_PERF_LIMIT_REASONS, 0x%08llx", cpu, msr); - fprintf(stderr, " (Active: %s%s%s%s%s%s%s%s)", + fprintf(outf, "cpu%d: MSR_GFX_PERF_LIMIT_REASONS, 0x%08llx", cpu, msr); + fprintf(outf, " (Active: %s%s%s%s%s%s%s%s)", (msr & 1 << 0) ? "PROCHOT, " : "", (msr & 1 << 1) ? "ThermStatus, " : "", (msr & 1 << 4) ? "Graphics, " : "", @@ -2108,7 +2486,7 @@ int print_perf_limit(struct thread_data *t, struct core_data *c, struct pkg_data (msr & 1 << 9) ? "GFXPwr, " : "", (msr & 1 << 10) ? "PkgPwrL1, " : "", (msr & 1 << 11) ? "PkgPwrL2, " : ""); - fprintf(stderr, " (Logged: %s%s%s%s%s%s%s%s)\n", + fprintf(outf, " (Logged: %s%s%s%s%s%s%s%s)\n", (msr & 1 << 16) ? "PROCHOT, " : "", (msr & 1 << 17) ? "ThermStatus, " : "", (msr & 1 << 20) ? "Graphics, " : "", @@ -2120,15 +2498,15 @@ int print_perf_limit(struct thread_data *t, struct core_data *c, struct pkg_data } if (do_ring_perf_limit_reasons) { get_msr(cpu, MSR_RING_PERF_LIMIT_REASONS, &msr); - fprintf(stderr, "cpu%d: MSR_RING_PERF_LIMIT_REASONS, 0x%08llx", cpu, msr); - fprintf(stderr, " (Active: %s%s%s%s%s%s)", + fprintf(outf, "cpu%d: MSR_RING_PERF_LIMIT_REASONS, 0x%08llx", cpu, msr); + fprintf(outf, " (Active: %s%s%s%s%s%s)", (msr & 1 << 0) ? "PROCHOT, " : "", (msr & 1 << 1) ? "ThermStatus, " : "", (msr & 1 << 6) ? "VR-Therm, " : "", (msr & 1 << 8) ? "Amps, " : "", (msr & 1 << 10) ? "PkgPwrL1, " : "", (msr & 1 << 11) ? "PkgPwrL2, " : ""); - fprintf(stderr, " (Logged: %s%s%s%s%s%s)\n", + fprintf(outf, " (Logged: %s%s%s%s%s%s)\n", (msr & 1 << 16) ? "PROCHOT, " : "", (msr & 1 << 17) ? "ThermStatus, " : "", (msr & 1 << 22) ? "VR-Therm, " : "", @@ -2251,7 +2629,7 @@ void rapl_probe(unsigned int family, unsigned int model) rapl_joule_counter_range = 0xFFFFFFFF * rapl_energy_units / tdp; if (debug) - fprintf(stderr, "RAPL: %.0f sec. Joule Counter Range, at %.0f Watts\n", rapl_joule_counter_range, tdp); + fprintf(outf, "RAPL: %.0f sec. Joule Counter Range, at %.0f Watts\n", rapl_joule_counter_range, tdp); return; } @@ -2293,7 +2671,7 @@ int print_thermal(struct thread_data *t, struct core_data *c, struct pkg_data *p return 0; if (cpu_migrate(cpu)) { - fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + fprintf(outf, "Could not migrate to CPU %d\n", cpu); return -1; } @@ -2302,7 +2680,7 @@ int print_thermal(struct thread_data *t, struct core_data *c, struct pkg_data *p return 0; dts = (msr >> 16) & 0x7F; - fprintf(stderr, "cpu%d: MSR_IA32_PACKAGE_THERM_STATUS: 0x%08llx (%d C)\n", + fprintf(outf, "cpu%d: MSR_IA32_PACKAGE_THERM_STATUS: 0x%08llx (%d C)\n", cpu, msr, tcc_activation_temp - dts); #ifdef THERM_DEBUG @@ -2311,7 +2689,7 @@ int print_thermal(struct thread_data *t, struct core_data *c, struct pkg_data *p dts = (msr >> 16) & 0x7F; dts2 = (msr >> 8) & 0x7F; - fprintf(stderr, "cpu%d: MSR_IA32_PACKAGE_THERM_INTERRUPT: 0x%08llx (%d C, %d C)\n", + fprintf(outf, "cpu%d: MSR_IA32_PACKAGE_THERM_INTERRUPT: 0x%08llx (%d C, %d C)\n", cpu, msr, tcc_activation_temp - dts, tcc_activation_temp - dts2); #endif } @@ -2325,7 +2703,7 @@ int print_thermal(struct thread_data *t, struct core_data *c, struct pkg_data *p dts = (msr >> 16) & 0x7F; resolution = (msr >> 27) & 0xF; - fprintf(stderr, "cpu%d: MSR_IA32_THERM_STATUS: 0x%08llx (%d C +/- %d)\n", + fprintf(outf, "cpu%d: MSR_IA32_THERM_STATUS: 0x%08llx (%d C +/- %d)\n", cpu, msr, tcc_activation_temp - dts, resolution); #ifdef THERM_DEBUG @@ -2334,17 +2712,17 @@ int print_thermal(struct thread_data *t, struct core_data *c, struct pkg_data *p dts = (msr >> 16) & 0x7F; dts2 = (msr >> 8) & 0x7F; - fprintf(stderr, "cpu%d: MSR_IA32_THERM_INTERRUPT: 0x%08llx (%d C, %d C)\n", + fprintf(outf, "cpu%d: MSR_IA32_THERM_INTERRUPT: 0x%08llx (%d C, %d C)\n", cpu, msr, tcc_activation_temp - dts, tcc_activation_temp - dts2); #endif } return 0; } - + void print_power_limit_msr(int cpu, unsigned long long msr, char *label) { - fprintf(stderr, "cpu%d: %s: %sabled (%f Watts, %f sec, clamp %sabled)\n", + fprintf(outf, "cpu%d: %s: %sabled (%f Watts, %f sec, clamp %sabled)\n", cpu, label, ((msr >> 15) & 1) ? "EN" : "DIS", ((msr >> 0) & 0x7FFF) * rapl_power_units, @@ -2368,7 +2746,7 @@ int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p) cpu = t->cpu_id; if (cpu_migrate(cpu)) { - fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + fprintf(outf, "Could not migrate to CPU %d\n", cpu); return -1; } @@ -2376,7 +2754,7 @@ int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p) return -1; if (debug) { - fprintf(stderr, "cpu%d: MSR_RAPL_POWER_UNIT: 0x%08llx " + fprintf(outf, "cpu%d: MSR_RAPL_POWER_UNIT: 0x%08llx " "(%f Watts, %f Joules, %f sec.)\n", cpu, msr, rapl_power_units, rapl_energy_units, rapl_time_units); } @@ -2386,7 +2764,7 @@ int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p) return -5; - fprintf(stderr, "cpu%d: MSR_PKG_POWER_INFO: 0x%08llx (%.0f W TDP, RAPL %.0f - %.0f W, %f sec.)\n", + fprintf(outf, "cpu%d: MSR_PKG_POWER_INFO: 0x%08llx (%.0f W TDP, RAPL %.0f - %.0f W, %f sec.)\n", cpu, msr, ((msr >> 0) & RAPL_POWER_GRANULARITY) * rapl_power_units, ((msr >> 16) & RAPL_POWER_GRANULARITY) * rapl_power_units, @@ -2399,11 +2777,11 @@ int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p) if (get_msr(cpu, MSR_PKG_POWER_LIMIT, &msr)) return -9; - fprintf(stderr, "cpu%d: MSR_PKG_POWER_LIMIT: 0x%08llx (%slocked)\n", + fprintf(outf, "cpu%d: MSR_PKG_POWER_LIMIT: 0x%08llx (%slocked)\n", cpu, msr, (msr >> 63) & 1 ? "": "UN"); print_power_limit_msr(cpu, msr, "PKG Limit #1"); - fprintf(stderr, "cpu%d: PKG Limit #2: %sabled (%f Watts, %f* sec, clamp %sabled)\n", + fprintf(outf, "cpu%d: PKG Limit #2: %sabled (%f Watts, %f* sec, clamp %sabled)\n", cpu, ((msr >> 47) & 1) ? "EN" : "DIS", ((msr >> 32) & 0x7FFF) * rapl_power_units, @@ -2415,7 +2793,7 @@ int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p) if (get_msr(cpu, MSR_DRAM_POWER_INFO, &msr)) return -6; - fprintf(stderr, "cpu%d: MSR_DRAM_POWER_INFO,: 0x%08llx (%.0f W TDP, RAPL %.0f - %.0f W, %f sec.)\n", + fprintf(outf, "cpu%d: MSR_DRAM_POWER_INFO,: 0x%08llx (%.0f W TDP, RAPL %.0f - %.0f W, %f sec.)\n", cpu, msr, ((msr >> 0) & RAPL_POWER_GRANULARITY) * rapl_power_units, ((msr >> 16) & RAPL_POWER_GRANULARITY) * rapl_power_units, @@ -2425,7 +2803,7 @@ int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p) if (do_rapl & RAPL_DRAM) { if (get_msr(cpu, MSR_DRAM_POWER_LIMIT, &msr)) return -9; - fprintf(stderr, "cpu%d: MSR_DRAM_POWER_LIMIT: 0x%08llx (%slocked)\n", + fprintf(outf, "cpu%d: MSR_DRAM_POWER_LIMIT: 0x%08llx (%slocked)\n", cpu, msr, (msr >> 31) & 1 ? "": "UN"); print_power_limit_msr(cpu, msr, "DRAM Limit"); @@ -2435,7 +2813,7 @@ int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p) if (get_msr(cpu, MSR_PP0_POLICY, &msr)) return -7; - fprintf(stderr, "cpu%d: MSR_PP0_POLICY: %lld\n", cpu, msr & 0xF); + fprintf(outf, "cpu%d: MSR_PP0_POLICY: %lld\n", cpu, msr & 0xF); } } if (do_rapl & RAPL_CORES) { @@ -2443,7 +2821,7 @@ int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p) if (get_msr(cpu, MSR_PP0_POWER_LIMIT, &msr)) return -9; - fprintf(stderr, "cpu%d: MSR_PP0_POWER_LIMIT: 0x%08llx (%slocked)\n", + fprintf(outf, "cpu%d: MSR_PP0_POWER_LIMIT: 0x%08llx (%slocked)\n", cpu, msr, (msr >> 31) & 1 ? "": "UN"); print_power_limit_msr(cpu, msr, "Cores Limit"); } @@ -2453,11 +2831,11 @@ int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p) if (get_msr(cpu, MSR_PP1_POLICY, &msr)) return -8; - fprintf(stderr, "cpu%d: MSR_PP1_POLICY: %lld\n", cpu, msr & 0xF); + fprintf(outf, "cpu%d: MSR_PP1_POLICY: %lld\n", cpu, msr & 0xF); if (get_msr(cpu, MSR_PP1_POWER_LIMIT, &msr)) return -9; - fprintf(stderr, "cpu%d: MSR_PP1_POWER_LIMIT: 0x%08llx (%slocked)\n", + fprintf(outf, "cpu%d: MSR_PP1_POWER_LIMIT: 0x%08llx (%slocked)\n", cpu, msr, (msr >> 31) & 1 ? "": "UN"); print_power_limit_msr(cpu, msr, "GFX Limit"); } @@ -2583,23 +2961,23 @@ double slm_bclk(void) double freq; if (get_msr(base_cpu, MSR_FSB_FREQ, &msr)) - fprintf(stderr, "SLM BCLK: unknown\n"); + fprintf(outf, "SLM BCLK: unknown\n"); i = msr & 0xf; if (i >= SLM_BCLK_FREQS) { - fprintf(stderr, "SLM BCLK[%d] invalid\n", i); + fprintf(outf, "SLM BCLK[%d] invalid\n", i); msr = 3; } freq = slm_freq_table[i]; - fprintf(stderr, "SLM BCLK: %.1f Mhz\n", freq); + fprintf(outf, "SLM BCLK: %.1f Mhz\n", freq); return freq; } double discover_bclk(unsigned int family, unsigned int model) { - if (has_snb_msrs(family, model)) + if (has_snb_msrs(family, model) || is_knl(family, model)) return 100.00; else if (is_slm(family, model)) return slm_bclk(); @@ -2635,13 +3013,13 @@ int set_temperature_target(struct thread_data *t, struct core_data *c, struct pk cpu = t->cpu_id; if (cpu_migrate(cpu)) { - fprintf(stderr, "Could not migrate to CPU %d\n", cpu); + fprintf(outf, "Could not migrate to CPU %d\n", cpu); return -1; } if (tcc_activation_temp_override != 0) { tcc_activation_temp = tcc_activation_temp_override; - fprintf(stderr, "cpu%d: Using cmdline TCC Target (%d C)\n", + fprintf(outf, "cpu%d: Using cmdline TCC Target (%d C)\n", cpu, tcc_activation_temp); return 0; } @@ -2656,7 +3034,7 @@ int set_temperature_target(struct thread_data *t, struct core_data *c, struct pk target_c_local = (msr >> 16) & 0xFF; if (debug) - fprintf(stderr, "cpu%d: MSR_IA32_TEMPERATURE_TARGET: 0x%08llx (%d C)\n", + fprintf(outf, "cpu%d: MSR_IA32_TEMPERATURE_TARGET: 0x%08llx (%d C)\n", cpu, msr, target_c_local); if (!target_c_local) @@ -2668,37 +3046,93 @@ int set_temperature_target(struct thread_data *t, struct core_data *c, struct pk guess: tcc_activation_temp = TJMAX_DEFAULT; - fprintf(stderr, "cpu%d: Guessing tjMax %d C, Please use -T to specify\n", + fprintf(outf, "cpu%d: Guessing tjMax %d C, Please use -T to specify\n", cpu, tcc_activation_temp); return 0; } + +void decode_feature_control_msr(void) +{ + unsigned long long msr; + + if (!get_msr(base_cpu, MSR_IA32_FEATURE_CONTROL, &msr)) + fprintf(outf, "cpu%d: MSR_IA32_FEATURE_CONTROL: 0x%08llx (%sLocked %s)\n", + base_cpu, msr, + msr & FEATURE_CONTROL_LOCKED ? "" : "UN-", + msr & (1 << 18) ? "SGX" : ""); +} + +void decode_misc_enable_msr(void) +{ + unsigned long long msr; + + if (!get_msr(base_cpu, MSR_IA32_MISC_ENABLE, &msr)) + fprintf(outf, "cpu%d: MSR_IA32_MISC_ENABLE: 0x%08llx (%s %s %s)\n", + base_cpu, msr, + msr & (1 << 3) ? "TCC" : "", + msr & (1 << 16) ? "EIST" : "", + msr & (1 << 18) ? "MONITOR" : ""); +} + +/* + * Decode MSR_MISC_PWR_MGMT + * + * Decode the bits according to the Nehalem documentation + * bit[0] seems to continue to have same meaning going forward + * bit[1] less so... + */ +void decode_misc_pwr_mgmt_msr(void) +{ + unsigned long long msr; + + if (!do_nhm_platform_info) + return; + + if (!get_msr(base_cpu, MSR_MISC_PWR_MGMT, &msr)) + fprintf(outf, "cpu%d: MSR_MISC_PWR_MGMT: 0x%08llx (%sable-EIST_Coordination %sable-EPB)\n", + base_cpu, msr, + msr & (1 << 0) ? "DIS" : "EN", + msr & (1 << 1) ? "EN" : "DIS"); +} + void process_cpuid() { - unsigned int eax, ebx, ecx, edx, max_level; + unsigned int eax, ebx, ecx, edx, max_level, max_extended_level; unsigned int fms, family, model, stepping; eax = ebx = ecx = edx = 0; - __get_cpuid(0, &max_level, &ebx, &ecx, &edx); + __cpuid(0, max_level, ebx, ecx, edx); if (ebx == 0x756e6547 && edx == 0x49656e69 && ecx == 0x6c65746e) genuine_intel = 1; if (debug) - fprintf(stderr, "CPUID(0): %.4s%.4s%.4s ", + fprintf(outf, "CPUID(0): %.4s%.4s%.4s ", (char *)&ebx, (char *)&edx, (char *)&ecx); - __get_cpuid(1, &fms, &ebx, &ecx, &edx); + __cpuid(1, fms, ebx, ecx, edx); family = (fms >> 8) & 0xf; model = (fms >> 4) & 0xf; stepping = fms & 0xf; if (family == 6 || family == 0xf) model += ((fms >> 16) & 0xf) << 4; - if (debug) - fprintf(stderr, "%d CPUID levels; family:model:stepping 0x%x:%x:%x (%d:%d:%d)\n", + if (debug) { + fprintf(outf, "%d CPUID levels; family:model:stepping 0x%x:%x:%x (%d:%d:%d)\n", max_level, family, model, stepping, family, model, stepping); + fprintf(outf, "CPUID(1): %s %s %s %s %s %s %s %s %s\n", + ecx & (1 << 0) ? "SSE3" : "-", + ecx & (1 << 3) ? "MONITOR" : "-", + ecx & (1 << 6) ? "SMX" : "-", + ecx & (1 << 7) ? "EIST" : "-", + ecx & (1 << 8) ? "TM2" : "-", + edx & (1 << 4) ? "TSC" : "-", + edx & (1 << 5) ? "MSR" : "-", + edx & (1 << 22) ? "ACPI-TM" : "-", + edx & (1 << 29) ? "TM" : "-"); + } if (!(edx & (1 << 5))) errx(1, "CPUID: no MSR"); @@ -2709,15 +3143,15 @@ void process_cpuid() * This check is valid for both Intel and AMD. */ ebx = ecx = edx = 0; - __get_cpuid(0x80000000, &max_level, &ebx, &ecx, &edx); + __cpuid(0x80000000, max_extended_level, ebx, ecx, edx); - if (max_level >= 0x80000007) { + if (max_extended_level >= 0x80000007) { /* * Non-Stop TSC is advertised by CPUID.EAX=0x80000007: EDX.bit8 * this check is valid for both Intel and AMD */ - __get_cpuid(0x80000007, &eax, &ebx, &ecx, &edx); + __cpuid(0x80000007, eax, ebx, ecx, edx); has_invariant_tsc = edx & (1 << 8); } @@ -2726,20 +3160,48 @@ void process_cpuid() * this check is valid for both Intel and AMD */ - __get_cpuid(0x6, &eax, &ebx, &ecx, &edx); + __cpuid(0x6, eax, ebx, ecx, edx); has_aperf = ecx & (1 << 0); do_dts = eax & (1 << 0); do_ptm = eax & (1 << 6); + has_hwp = eax & (1 << 7); + has_hwp_notify = eax & (1 << 8); + has_hwp_activity_window = eax & (1 << 9); + has_hwp_epp = eax & (1 << 10); + has_hwp_pkg = eax & (1 << 11); has_epb = ecx & (1 << 3); if (debug) - fprintf(stderr, "CPUID(6): %sAPERF, %sDTS, %sPTM, %sEPB\n", - has_aperf ? "" : "No ", - do_dts ? "" : "No ", - do_ptm ? "" : "No ", - has_epb ? "" : "No "); + fprintf(outf, "CPUID(6): %sAPERF, %sDTS, %sPTM, %sHWP, " + "%sHWPnotify, %sHWPwindow, %sHWPepp, %sHWPpkg, %sEPB\n", + has_aperf ? "" : "No-", + do_dts ? "" : "No-", + do_ptm ? "" : "No-", + has_hwp ? "" : "No-", + has_hwp_notify ? "" : "No-", + has_hwp_activity_window ? "" : "No-", + has_hwp_epp ? "" : "No-", + has_hwp_pkg ? "" : "No-", + has_epb ? "" : "No-"); + + if (debug) + decode_misc_enable_msr(); + + if (max_level >= 0x7) { + int has_sgx; - if (max_level > 0x15) { + ecx = 0; + + __cpuid_count(0x7, 0, eax, ebx, ecx, edx); + + has_sgx = ebx & (1 << 2); + fprintf(outf, "CPUID(7): %sSGX\n", has_sgx ? "" : "No-"); + + if (has_sgx) + decode_feature_control_msr(); + } + + if (max_level >= 0x15) { unsigned int eax_crystal; unsigned int ebx_tsc; @@ -2747,12 +3209,12 @@ void process_cpuid() * CPUID 15H TSC/Crystal ratio, possibly Crystal Hz */ eax_crystal = ebx_tsc = crystal_hz = edx = 0; - __get_cpuid(0x15, &eax_crystal, &ebx_tsc, &crystal_hz, &edx); + __cpuid(0x15, eax_crystal, ebx_tsc, crystal_hz, edx); if (ebx_tsc != 0) { if (debug && (ebx != 0)) - fprintf(stderr, "CPUID(0x15): eax_crystal: %d ebx_tsc: %d ecx_crystal_hz: %d\n", + fprintf(outf, "CPUID(0x15): eax_crystal: %d ebx_tsc: %d ecx_crystal_hz: %d\n", eax_crystal, ebx_tsc, crystal_hz); if (crystal_hz == 0) @@ -2768,11 +3230,24 @@ void process_cpuid() if (crystal_hz) { tsc_hz = (unsigned long long) crystal_hz * ebx_tsc / eax_crystal; if (debug) - fprintf(stderr, "TSC: %lld MHz (%d Hz * %d / %d / 1000000)\n", + fprintf(outf, "TSC: %lld MHz (%d Hz * %d / %d / 1000000)\n", tsc_hz / 1000000, crystal_hz, ebx_tsc, eax_crystal); } } } + if (max_level >= 0x16) { + unsigned int base_mhz, max_mhz, bus_mhz, edx; + + /* + * CPUID 16H Base MHz, Max MHz, Bus MHz + */ + base_mhz = max_mhz = bus_mhz = edx = 0; + + __cpuid(0x16, base_mhz, max_mhz, bus_mhz, edx); + if (debug) + fprintf(outf, "CPUID(0x16): base_mhz: %d max_mhz: %d bus_mhz: %d\n", + base_mhz, max_mhz, bus_mhz); + } if (has_aperf) aperf_mperf_multiplier = get_aperf_mperf_multiplier(family, model); @@ -2788,6 +3263,9 @@ void process_cpuid() do_slm_cstates = is_slm(family, model); do_knl_cstates = is_knl(family, model); + if (debug) + decode_misc_pwr_mgmt_msr(); + rapl_probe(family, model); perf_limit_reasons_probe(family, model); @@ -2797,12 +3275,16 @@ void process_cpuid() if (has_skl_msrs(family, model)) calculate_tsc_tweak(); + do_gfx_rc6_ms = !access("/sys/class/drm/card0/power/rc6_residency_ms", R_OK); + + do_gfx_mhz = !access("/sys/class/graphics/fb0/device/drm/card0/gt_cur_freq_mhz", R_OK); + return; } void help() { - fprintf(stderr, + fprintf(outf, "Usage: turbostat [OPTIONS][(--interval seconds) | COMMAND ...]\n" "\n" "Turbostat forks the specified COMMAND and prints statistics\n" @@ -2814,6 +3296,7 @@ void help() "--help print this help message\n" "--counter msr print 32-bit counter at address \"msr\"\n" "--Counter msr print 64-bit Counter at address \"msr\"\n" + "--out file create or truncate \"file\" for all output\n" "--msr msr print 32-bit value at address \"msr\"\n" "--MSR msr print 64-bit Value at address \"msr\"\n" "--version print version information\n" @@ -2858,7 +3341,7 @@ void topology_probe() show_cpu = 1; if (debug > 1) - fprintf(stderr, "num_cpus %d max_cpu_num %d\n", topo.num_cpus, topo.max_cpu_num); + fprintf(outf, "num_cpus %d max_cpu_num %d\n", topo.num_cpus, topo.max_cpu_num); cpus = calloc(1, (topo.max_cpu_num + 1) * sizeof(struct cpu_topology)); if (cpus == NULL) @@ -2893,7 +3376,7 @@ void topology_probe() if (cpu_is_not_present(i)) { if (debug > 1) - fprintf(stderr, "cpu%d NOT PRESENT\n", i); + fprintf(outf, "cpu%d NOT PRESENT\n", i); continue; } cpus[i].core_id = get_core_id(i); @@ -2908,26 +3391,26 @@ void topology_probe() if (siblings > max_siblings) max_siblings = siblings; if (debug > 1) - fprintf(stderr, "cpu %d pkg %d core %d\n", + fprintf(outf, "cpu %d pkg %d core %d\n", i, cpus[i].physical_package_id, cpus[i].core_id); } topo.num_cores_per_pkg = max_core_id + 1; if (debug > 1) - fprintf(stderr, "max_core_id %d, sizing for %d cores per package\n", + fprintf(outf, "max_core_id %d, sizing for %d cores per package\n", max_core_id, topo.num_cores_per_pkg); if (debug && !summary_only && topo.num_cores_per_pkg > 1) show_core = 1; topo.num_packages = max_package_id + 1; if (debug > 1) - fprintf(stderr, "max_package_id %d, sizing for %d packages\n", + fprintf(outf, "max_package_id %d, sizing for %d packages\n", max_package_id, topo.num_packages); if (debug && !summary_only && topo.num_packages > 1) show_pkg = 1; topo.num_threads_per_core = max_siblings; if (debug > 1) - fprintf(stderr, "max_siblings %d\n", max_siblings); + fprintf(outf, "max_siblings %d\n", max_siblings); free(cpus); } @@ -3019,10 +3502,27 @@ void allocate_output_buffer() if (outp == NULL) err(-1, "calloc output buffer"); } +void allocate_fd_percpu(void) +{ + fd_percpu = calloc(topo.max_cpu_num, sizeof(int)); + if (fd_percpu == NULL) + err(-1, "calloc fd_percpu"); +} +void allocate_irq_buffers(void) +{ + irq_column_2_cpu = calloc(topo.num_cpus, sizeof(int)); + if (irq_column_2_cpu == NULL) + err(-1, "calloc %d", topo.num_cpus); + irqs_per_cpu = calloc(topo.max_cpu_num, sizeof(int)); + if (irqs_per_cpu == NULL) + err(-1, "calloc %d", topo.max_cpu_num); +} void setup_all_buffers(void) { topology_probe(); + allocate_irq_buffers(); + allocate_fd_percpu(); allocate_counters(&thread_even, &core_even, &package_even); allocate_counters(&thread_odd, &core_odd, &package_odd); allocate_output_buffer(); @@ -3036,7 +3536,7 @@ void set_base_cpu(void) err(-ENODEV, "No valid cpus found"); if (debug > 1) - fprintf(stderr, "base_cpu = %d\n", base_cpu); + fprintf(outf, "base_cpu = %d\n", base_cpu); } void turbostat_init() @@ -3049,6 +3549,9 @@ void turbostat_init() if (debug) + for_all_cpus(print_hwp, ODD_COUNTERS); + + if (debug) for_all_cpus(print_epb, ODD_COUNTERS); if (debug) @@ -3100,9 +3603,10 @@ int fork_it(char **argv) for_all_cpus_2(delta_cpu, ODD_COUNTERS, EVEN_COUNTERS); compute_average(EVEN_COUNTERS); format_all_counters(EVEN_COUNTERS); - flush_stderr(); - fprintf(stderr, "%.6f sec\n", tv_delta.tv_sec + tv_delta.tv_usec/1000000.0); + fprintf(outf, "%.6f sec\n", tv_delta.tv_sec + tv_delta.tv_usec/1000000.0); + + flush_output_stderr(); return status; } @@ -3119,13 +3623,13 @@ int get_and_dump_counters(void) if (status) return status; - flush_stdout(); + flush_output_stdout(); return status; } void print_version() { - fprintf(stderr, "turbostat version 4.8 26-Sep, 2015" + fprintf(outf, "turbostat version 4.11 27 Feb 2016" " - Len Brown <lenb@kernel.org>\n"); } @@ -3143,6 +3647,7 @@ void cmdline(int argc, char **argv) {"Joules", no_argument, 0, 'J'}, {"MSR", required_argument, 0, 'M'}, {"msr", required_argument, 0, 'm'}, + {"out", required_argument, 0, 'o'}, {"Package", no_argument, 0, 'p'}, {"processor", no_argument, 0, 'p'}, {"Summary", no_argument, 0, 'S'}, @@ -3153,7 +3658,7 @@ void cmdline(int argc, char **argv) progname = argv[0]; - while ((opt = getopt_long_only(argc, argv, "+C:c:Ddhi:JM:m:PpST:v", + while ((opt = getopt_long_only(argc, argv, "+C:c:Ddhi:JM:m:o:PpST:v", long_options, &option_index)) != -1) { switch (opt) { case 'C': @@ -3173,7 +3678,18 @@ void cmdline(int argc, char **argv) help(); exit(1); case 'i': - interval_sec = atoi(optarg); + { + double interval = strtod(optarg, NULL); + + if (interval < 0.001) { + fprintf(outf, "interval %f seconds is too small\n", + interval); + exit(2); + } + + interval_ts.tv_sec = interval; + interval_ts.tv_nsec = (interval - interval_ts.tv_sec) * 1000000000; + } break; case 'J': rapl_joules++; @@ -3184,6 +3700,9 @@ void cmdline(int argc, char **argv) case 'm': sscanf(optarg, "%x", &extra_msr_offset32); break; + case 'o': + outf = fopen_or_die(optarg, "w"); + break; case 'P': show_pkg_only++; break; @@ -3206,6 +3725,8 @@ void cmdline(int argc, char **argv) int main(int argc, char **argv) { + outf = stderr; + cmdline(argc, argv); if (debug) |