From 7f7acd193ba8aaa8ed07cfadc335bb17a991fd42 Mon Sep 17 00:00:00 2001 From: Brian Norris Date: Thu, 25 Sep 2025 12:42:14 -0700 Subject: [PATCH 01/19] PM: runtime: Add basic kunit tests for API contracts In exploring the various return codes and failure modes of runtime PM APIs, I found it helpful to verify and codify many of them in unit tests, especially given that even the kerneldoc can be rather complex to reason through, and it also has had subtle errors of its own. Notably, I avoid testing the return codes for pm_runtime_put() and pm_runtime_put_autosuspend(), since code that checks them is probably wrong, and we're considering making them return 'void' altogether. I still test the sync() variants, since those have a bit more meaning to them. Signed-off-by: Brian Norris Signed-off-by: Rafael J. Wysocki --- drivers/base/Kconfig | 6 + drivers/base/power/Makefile | 1 + drivers/base/power/runtime-test.c | 253 ++++++++++++++++++++++++++++++ 3 files changed, 260 insertions(+) create mode 100644 drivers/base/power/runtime-test.c diff --git a/drivers/base/Kconfig b/drivers/base/Kconfig index 064eb52ff7e2..1786d87b29e2 100644 --- a/drivers/base/Kconfig +++ b/drivers/base/Kconfig @@ -167,6 +167,12 @@ config PM_QOS_KUNIT_TEST depends on KUNIT=y default KUNIT_ALL_TESTS +config PM_RUNTIME_KUNIT_TEST + tristate "KUnit Tests for runtime PM" if !KUNIT_ALL_TESTS + depends on KUNIT + depends on PM + default KUNIT_ALL_TESTS + config HMEM_REPORTING bool default n diff --git a/drivers/base/power/Makefile b/drivers/base/power/Makefile index 01f11629d241..2989e42d0161 100644 --- a/drivers/base/power/Makefile +++ b/drivers/base/power/Makefile @@ -4,5 +4,6 @@ obj-$(CONFIG_PM_SLEEP) += main.o wakeup.o wakeup_stats.o obj-$(CONFIG_PM_TRACE_RTC) += trace.o obj-$(CONFIG_HAVE_CLK) += clock_ops.o obj-$(CONFIG_PM_QOS_KUNIT_TEST) += qos-test.o +obj-$(CONFIG_PM_RUNTIME_KUNIT_TEST) += runtime-test.o ccflags-$(CONFIG_DEBUG_DRIVER) := -DDEBUG diff --git a/drivers/base/power/runtime-test.c b/drivers/base/power/runtime-test.c new file mode 100644 index 000000000000..2e966fd96664 --- /dev/null +++ b/drivers/base/power/runtime-test.c @@ -0,0 +1,253 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright 2025 Google, Inc. + */ + +#include +#include +#include +#include + +#define DEVICE_NAME "pm_runtime_test_device" + +static void pm_runtime_depth_test(struct kunit *test) +{ + struct device *dev = kunit_device_register(test, DEVICE_NAME); + + KUNIT_ASSERT_PTR_NE(test, NULL, dev); + + pm_runtime_enable(dev); + + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); + KUNIT_EXPECT_EQ(test, 0, pm_runtime_get_sync(dev)); + KUNIT_EXPECT_TRUE(test, pm_runtime_active(dev)); + KUNIT_EXPECT_EQ(test, 1, pm_runtime_get_sync(dev)); /* "already active" */ + KUNIT_EXPECT_EQ(test, 0, pm_runtime_put_sync(dev)); + KUNIT_EXPECT_EQ(test, 0, pm_runtime_put_sync(dev)); + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); +} + +/* Test pm_runtime_put() and friends when already suspended. */ +static void pm_runtime_already_suspended_test(struct kunit *test) +{ + struct device *dev = kunit_device_register(test, DEVICE_NAME); + + KUNIT_ASSERT_PTR_NE(test, NULL, dev); + + pm_runtime_enable(dev); + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); + + pm_runtime_get_noresume(dev); + KUNIT_EXPECT_EQ(test, 0, pm_runtime_barrier(dev)); /* no wakeup needed */ + pm_runtime_put(dev); + + pm_runtime_get_noresume(dev); + KUNIT_EXPECT_EQ(test, -EAGAIN, pm_runtime_put_sync(dev)); + + KUNIT_EXPECT_EQ(test, 1, pm_runtime_suspend(dev)); + KUNIT_EXPECT_EQ(test, 1, pm_runtime_autosuspend(dev)); + KUNIT_EXPECT_EQ(test, 1, pm_request_autosuspend(dev)); + + pm_runtime_get_noresume(dev); + KUNIT_EXPECT_EQ(test, 1, pm_runtime_put_sync_autosuspend(dev)); + + pm_runtime_get_noresume(dev); + pm_runtime_put_autosuspend(dev); + + /* Grab 2 refcounts */ + pm_runtime_get_noresume(dev); + pm_runtime_get_noresume(dev); + /* The first put() sees usage_count 1 */ + KUNIT_EXPECT_EQ(test, 0, pm_runtime_put_sync_autosuspend(dev)); + /* The second put() sees usage_count 0 but tells us "already suspended". */ + KUNIT_EXPECT_EQ(test, 1, pm_runtime_put_sync_autosuspend(dev)); + + /* Should have remained suspended the whole time. */ + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); +} + +static void pm_runtime_idle_test(struct kunit *test) +{ + struct device *dev = kunit_device_register(test, DEVICE_NAME); + + KUNIT_ASSERT_PTR_NE(test, NULL, dev); + + pm_runtime_enable(dev); + + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); + KUNIT_EXPECT_EQ(test, 0, pm_runtime_get_sync(dev)); + KUNIT_EXPECT_TRUE(test, pm_runtime_active(dev)); + KUNIT_EXPECT_EQ(test, -EAGAIN, pm_runtime_idle(dev)); + KUNIT_EXPECT_TRUE(test, pm_runtime_active(dev)); + pm_runtime_put_noidle(dev); + KUNIT_EXPECT_TRUE(test, pm_runtime_active(dev)); + KUNIT_EXPECT_EQ(test, 0, pm_runtime_idle(dev)); + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); + KUNIT_EXPECT_EQ(test, -EAGAIN, pm_runtime_idle(dev)); + KUNIT_EXPECT_EQ(test, -EAGAIN, pm_request_idle(dev)); +} + +static void pm_runtime_disabled_test(struct kunit *test) +{ + struct device *dev = kunit_device_register(test, DEVICE_NAME); + + KUNIT_ASSERT_PTR_NE(test, NULL, dev); + + /* Never called pm_runtime_enable() */ + KUNIT_EXPECT_FALSE(test, pm_runtime_enabled(dev)); + + /* "disabled" is treated as "active" */ + KUNIT_EXPECT_TRUE(test, pm_runtime_active(dev)); + KUNIT_EXPECT_FALSE(test, pm_runtime_suspended(dev)); + + /* + * Note: these "fail", but they still acquire/release refcounts, so + * keep them balanced. + */ + KUNIT_EXPECT_EQ(test, -EACCES, pm_runtime_get(dev)); + pm_runtime_put(dev); + + KUNIT_EXPECT_EQ(test, -EACCES, pm_runtime_get_sync(dev)); + KUNIT_EXPECT_EQ(test, -EACCES, pm_runtime_put_sync(dev)); + + KUNIT_EXPECT_EQ(test, -EACCES, pm_runtime_get(dev)); + pm_runtime_put_autosuspend(dev); + + KUNIT_EXPECT_EQ(test, -EACCES, pm_runtime_resume_and_get(dev)); + KUNIT_EXPECT_EQ(test, -EACCES, pm_runtime_idle(dev)); + KUNIT_EXPECT_EQ(test, -EACCES, pm_request_idle(dev)); + KUNIT_EXPECT_EQ(test, -EACCES, pm_request_resume(dev)); + KUNIT_EXPECT_EQ(test, -EACCES, pm_request_autosuspend(dev)); + KUNIT_EXPECT_EQ(test, -EACCES, pm_runtime_suspend(dev)); + KUNIT_EXPECT_EQ(test, -EACCES, pm_runtime_resume(dev)); + KUNIT_EXPECT_EQ(test, -EACCES, pm_runtime_autosuspend(dev)); + + /* Still disabled */ + KUNIT_EXPECT_TRUE(test, pm_runtime_active(dev)); + KUNIT_EXPECT_FALSE(test, pm_runtime_enabled(dev)); +} + +static void pm_runtime_error_test(struct kunit *test) +{ + struct device *dev = kunit_device_register(test, DEVICE_NAME); + + KUNIT_ASSERT_PTR_NE(test, NULL, dev); + + pm_runtime_enable(dev); + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); + + /* Fake a .runtime_resume() error */ + dev->power.runtime_error = -EIO; + + /* + * Note: these "fail", but they still acquire/release refcounts, so + * keep them balanced. + */ + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_get(dev)); + pm_runtime_put(dev); + + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_get_sync(dev)); + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_put_sync(dev)); + + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_get(dev)); + pm_runtime_put_autosuspend(dev); + + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_get(dev)); + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_put_sync_autosuspend(dev)); + + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_resume_and_get(dev)); + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_idle(dev)); + KUNIT_EXPECT_EQ(test, -EINVAL, pm_request_idle(dev)); + KUNIT_EXPECT_EQ(test, -EINVAL, pm_request_resume(dev)); + KUNIT_EXPECT_EQ(test, -EINVAL, pm_request_autosuspend(dev)); + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_suspend(dev)); + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_resume(dev)); + KUNIT_EXPECT_EQ(test, -EINVAL, pm_runtime_autosuspend(dev)); + + /* Error is still pending */ + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); + KUNIT_EXPECT_EQ(test, -EIO, dev->power.runtime_error); + /* Clear error */ + KUNIT_EXPECT_EQ(test, 0, pm_runtime_set_suspended(dev)); + KUNIT_EXPECT_EQ(test, 0, dev->power.runtime_error); + /* Still suspended */ + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); + + KUNIT_EXPECT_EQ(test, 0, pm_runtime_get(dev)); + KUNIT_EXPECT_EQ(test, 1, pm_runtime_barrier(dev)); /* resume was pending */ + pm_runtime_put(dev); + pm_runtime_suspend(dev); /* flush the put(), to suspend */ + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); + + KUNIT_EXPECT_EQ(test, 0, pm_runtime_get_sync(dev)); + KUNIT_EXPECT_EQ(test, 0, pm_runtime_put_sync(dev)); + + KUNIT_EXPECT_EQ(test, 0, pm_runtime_get_sync(dev)); + pm_runtime_put_autosuspend(dev); + + KUNIT_EXPECT_EQ(test, 0, pm_runtime_resume_and_get(dev)); + + /* + * The following should all return -EAGAIN (usage is non-zero) or 1 + * (already resumed). + */ + KUNIT_EXPECT_EQ(test, -EAGAIN, pm_runtime_idle(dev)); + KUNIT_EXPECT_EQ(test, -EAGAIN, pm_request_idle(dev)); + KUNIT_EXPECT_EQ(test, 1, pm_request_resume(dev)); + KUNIT_EXPECT_EQ(test, -EAGAIN, pm_request_autosuspend(dev)); + KUNIT_EXPECT_EQ(test, -EAGAIN, pm_runtime_suspend(dev)); + KUNIT_EXPECT_EQ(test, 1, pm_runtime_resume(dev)); + KUNIT_EXPECT_EQ(test, -EAGAIN, pm_runtime_autosuspend(dev)); + + KUNIT_EXPECT_EQ(test, 0, pm_runtime_put_sync(dev)); + + /* Suspended again */ + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); +} + +/* + * Explore a typical probe() sequence in which a device marks itself powered, + * but doesn't hold any runtime PM reference, so it suspends as soon as it goes + * idle. + */ +static void pm_runtime_probe_active_test(struct kunit *test) +{ + struct device *dev = kunit_device_register(test, DEVICE_NAME); + + KUNIT_ASSERT_PTR_NE(test, NULL, dev); + + KUNIT_EXPECT_TRUE(test, pm_runtime_status_suspended(dev)); + + KUNIT_EXPECT_EQ(test, 0, pm_runtime_set_active(dev)); + KUNIT_EXPECT_TRUE(test, pm_runtime_active(dev)); + + pm_runtime_enable(dev); + KUNIT_EXPECT_TRUE(test, pm_runtime_active(dev)); + + /* Nothing to flush. We stay active. */ + KUNIT_EXPECT_EQ(test, 0, pm_runtime_barrier(dev)); + KUNIT_EXPECT_TRUE(test, pm_runtime_active(dev)); + + /* Ask for idle? Now we suspend. */ + KUNIT_EXPECT_EQ(test, 0, pm_runtime_idle(dev)); + KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); +} + +static struct kunit_case pm_runtime_test_cases[] = { + KUNIT_CASE(pm_runtime_depth_test), + KUNIT_CASE(pm_runtime_already_suspended_test), + KUNIT_CASE(pm_runtime_idle_test), + KUNIT_CASE(pm_runtime_disabled_test), + KUNIT_CASE(pm_runtime_error_test), + KUNIT_CASE(pm_runtime_probe_active_test), + {} +}; + +static struct kunit_suite pm_runtime_test_suite = { + .name = "pm_runtime_test_cases", + .test_cases = pm_runtime_test_cases, +}; + +kunit_test_suite(pm_runtime_test_suite); +MODULE_DESCRIPTION("Runtime power management unit test suite"); +MODULE_LICENSE("GPL"); From d0b8651a026125d58b50b464aeb78f2c5956179f Mon Sep 17 00:00:00 2001 From: Brian Norris Date: Thu, 25 Sep 2025 12:42:15 -0700 Subject: [PATCH 02/19] PM: runtime: Make put{,_sync}() return 1 when already suspended The pm_runtime.h docs say pm_runtime_put() and pm_runtime_put_sync() return 1 when already suspended, but this is not true -- they return -EAGAIN. On the other hand, pm_runtime_put_sync_suspend() and pm_runtime_put_sync_autosuspend() *do* return 1. This is an artifact of the fact that the former are built on rpm_idle(), whereas the latter are built on rpm_suspend(). There are precious few pm_runtime_put()/pm_runtime_put_sync() callers that check the return code at all, but most of them only log errors, and usually only for negative error codes. None of them should be treating this as an error, so: * at best, this may fix some case where a driver treats this condition as an error, when it shouldn't; * at worst, this should make no effect; and * somewhere in between, we could potentially clear up non-fatal log messages. Fix the pm_runtime_already_suspended_test() while tweaking the behavior. The test makes a lot more sense when these all return 1 when the device is already suspended: pm_runtime_put_sync(dev); pm_runtime_suspend(dev); pm_runtime_autosuspend(dev); pm_request_autosuspend(dev); pm_runtime_put_sync_autosuspend(dev); Notably, I've avoided testing the return codes for these, since they really should be ignored by callers, and we may make them 'void' altogether: pm_runtime_put(dev); pm_runtime_put_autosuspend(dev); Signed-off-by: Brian Norris Reviewed-by: Dhruva Gole Reviewed-by: Sakari Ailus Signed-off-by: Rafael J. Wysocki --- drivers/base/power/runtime-test.c | 2 +- drivers/base/power/runtime.c | 3 +++ 2 files changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/base/power/runtime-test.c b/drivers/base/power/runtime-test.c index 2e966fd96664..eca9885e807d 100644 --- a/drivers/base/power/runtime-test.c +++ b/drivers/base/power/runtime-test.c @@ -42,7 +42,7 @@ static void pm_runtime_already_suspended_test(struct kunit *test) pm_runtime_put(dev); pm_runtime_get_noresume(dev); - KUNIT_EXPECT_EQ(test, -EAGAIN, pm_runtime_put_sync(dev)); + KUNIT_EXPECT_EQ(test, 1, pm_runtime_put_sync(dev)); KUNIT_EXPECT_EQ(test, 1, pm_runtime_suspend(dev)); KUNIT_EXPECT_EQ(test, 1, pm_runtime_autosuspend(dev)); diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c index 3e84dc4122de..faa68bf9ef3d 100644 --- a/drivers/base/power/runtime.c +++ b/drivers/base/power/runtime.c @@ -498,6 +498,9 @@ static int rpm_idle(struct device *dev, int rpmflags) if (retval < 0) ; /* Conditions are wrong. */ + else if ((rpmflags & RPM_GET_PUT) && retval == 1) + ; /* put() is allowed in RPM_SUSPENDED */ + /* Idle notifications are allowed only in the RPM_ACTIVE state. */ else if (dev->power.runtime_status != RPM_ACTIVE) retval = -EAGAIN; From fed7eaa4f037361fe4f3d4170649d6849a25998d Mon Sep 17 00:00:00 2001 From: Brian Norris Date: Thu, 25 Sep 2025 12:42:16 -0700 Subject: [PATCH 03/19] PM: runtime: Update kerneldoc return codes APIs based on __pm_runtime_idle() (pm_runtime_idle(), pm_request_idle()) do not return 1 when already suspended. They return -EAGAIN. This is already covered in the docs, so the entry for "1" is redundant and conflicting. (pm_runtime_put() and pm_runtime_put_sync() were previously incorrect, but that's fixed in "PM: runtime: pm_runtime_put{,_sync}() returns 1 when already suspended", to ensure consistency with APIs like pm_runtime_put_autosuspend().) RPM_GET_PUT APIs based on __pm_runtime_suspend() do return 1 when already suspended, but the language is a little unclear -- it's not really an "error", so it seems better to list as a clarification before the 0/success case. Additionally, they only actually return 1 when the refcount makes it to 0; if the usage counter is still non-zero, we return 0. pm_runtime_put(), etc., also don't appear at first like they can ever see "-EAGAIN: Runtime PM usage_count non-zero", because in non-racy conditions, pm_runtime_put() would drop its reference count, see it's non-zero, and return early (in __pm_runtime_idle()). However, it's possible to race with another actor that increments the usage_count afterward, since rpm_idle() is protected by a separate lock; in such a case, we may see -EAGAIN. Because this case is only seen in the presence of concurrent actors, it makes sense to clarify that this is when "usage_count **became** non-zero", by way of some racing actor. Lastly, pm_runtime_put_sync_suspend() duplicated some -EAGAIN language. Fix that. Fixes: 271ff96d6066 ("PM: runtime: Document return values of suspend-related API functions") Link: https://lore.kernel.org/linux-pm/aJ5pkEJuixTaybV4@google.com/ Signed-off-by: Brian Norris Reviewed-by: Sakari Ailus Cc: 6.17+ # 6.17+ Signed-off-by: Rafael J. Wysocki --- include/linux/pm_runtime.h | 56 +++++++++++++++++++++----------------- 1 file changed, 31 insertions(+), 25 deletions(-) diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index d88d6b6ccf5b..d1ff76e0e2d0 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -350,13 +350,12 @@ static inline int pm_runtime_force_resume(struct device *dev) { return -ENXIO; } * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero, Runtime PM status change ongoing - * or device not in %RPM_ACTIVE state. + * * -EAGAIN: Runtime PM usage counter non-zero, Runtime PM status change + * ongoing or device not in %RPM_ACTIVE state. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -EINPROGRESS: Suspend already in progress. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. * Other values and conditions for the above values are possible as returned by * Runtime PM idle and suspend callbacks. */ @@ -370,14 +369,15 @@ static inline int pm_runtime_idle(struct device *dev) * @dev: Target device. * * Return: + * * 1: Success; device was already suspended. * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero or Runtime PM status change ongoing. + * * -EAGAIN: Runtime PM usage counter non-zero or Runtime PM status change + * ongoing. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. * Other values and conditions for the above values are possible as returned by * Runtime PM suspend callbacks. */ @@ -396,14 +396,15 @@ static inline int pm_runtime_suspend(struct device *dev) * engaging its "idle check" callback. * * Return: + * * 1: Success; device was already suspended. * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero or Runtime PM status change ongoing. + * * -EAGAIN: Runtime PM usage counter non-zero or Runtime PM status change + * ongoing. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. * Other values and conditions for the above values are possible as returned by * Runtime PM suspend callbacks. */ @@ -433,13 +434,12 @@ static inline int pm_runtime_resume(struct device *dev) * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero, Runtime PM status change ongoing - * or device not in %RPM_ACTIVE state. + * * -EAGAIN: Runtime PM usage counter non-zero, Runtime PM status change + * ongoing or device not in %RPM_ACTIVE state. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -EINPROGRESS: Suspend already in progress. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. */ static inline int pm_request_idle(struct device *dev) { @@ -464,15 +464,16 @@ static inline int pm_request_resume(struct device *dev) * equivalent pm_runtime_autosuspend() for @dev asynchronously. * * Return: + * * 1: Success; device was already suspended. * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero or Runtime PM status change ongoing. + * * -EAGAIN: Runtime PM usage counter non-zero or Runtime PM status change + * ongoing. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -EINPROGRESS: Suspend already in progress. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. */ static inline int pm_request_autosuspend(struct device *dev) { @@ -540,15 +541,16 @@ static inline int pm_runtime_resume_and_get(struct device *dev) * equal to 0, queue up a work item for @dev like in pm_request_idle(). * * Return: + * * 1: Success. Usage counter dropped to zero, but device was already suspended. * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero or Runtime PM status change ongoing. + * * -EAGAIN: Runtime PM usage counter became non-zero or Runtime PM status + * change ongoing. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -EINPROGRESS: Suspend already in progress. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. */ static inline int pm_runtime_put(struct device *dev) { @@ -565,15 +567,16 @@ DEFINE_FREE(pm_runtime_put, struct device *, if (_T) pm_runtime_put(_T)) * equal to 0, queue up a work item for @dev like in pm_request_autosuspend(). * * Return: + * * 1: Success. Usage counter dropped to zero, but device was already suspended. * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero or Runtime PM status change ongoing. + * * -EAGAIN: Runtime PM usage counter became non-zero or Runtime PM status + * change ongoing. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -EINPROGRESS: Suspend already in progress. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. */ static inline int __pm_runtime_put_autosuspend(struct device *dev) { @@ -590,15 +593,16 @@ static inline int __pm_runtime_put_autosuspend(struct device *dev) * in pm_request_autosuspend(). * * Return: + * * 1: Success. Usage counter dropped to zero, but device was already suspended. * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero or Runtime PM status change ongoing. + * * -EAGAIN: Runtime PM usage counter became non-zero or Runtime PM status + * change ongoing. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -EINPROGRESS: Suspend already in progress. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. */ static inline int pm_runtime_put_autosuspend(struct device *dev) { @@ -619,14 +623,15 @@ static inline int pm_runtime_put_autosuspend(struct device *dev) * if it returns an error code. * * Return: + * * 1: Success. Usage counter dropped to zero, but device was already suspended. * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero or Runtime PM status change ongoing. + * * -EAGAIN: Runtime PM usage counter became non-zero or Runtime PM status + * change ongoing. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. * Other values and conditions for the above values are possible as returned by * Runtime PM suspend callbacks. */ @@ -646,15 +651,15 @@ static inline int pm_runtime_put_sync(struct device *dev) * if it returns an error code. * * Return: + * * 1: Success. Usage counter dropped to zero, but device was already suspended. * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero or Runtime PM status change ongoing. - * * -EAGAIN: usage_count non-zero or Runtime PM status change ongoing. + * * -EAGAIN: Runtime PM usage counter became non-zero or Runtime PM status + * change ongoing. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. * Other values and conditions for the above values are possible as returned by * Runtime PM suspend callbacks. */ @@ -677,15 +682,16 @@ static inline int pm_runtime_put_sync_suspend(struct device *dev) * if it returns an error code. * * Return: + * * 1: Success. Usage counter dropped to zero, but device was already suspended. * * 0: Success. * * -EINVAL: Runtime PM error. * * -EACCES: Runtime PM disabled. - * * -EAGAIN: Runtime PM usage_count non-zero or Runtime PM status change ongoing. + * * -EAGAIN: Runtime PM usage counter became non-zero or Runtime PM status + * change ongoing. * * -EBUSY: Runtime PM child_count non-zero. * * -EPERM: Device PM QoS resume latency 0. * * -EINPROGRESS: Suspend already in progress. * * -ENOSYS: CONFIG_PM not enabled. - * * 1: Device already suspended. * Other values and conditions for the above values are possible as returned by * Runtime PM suspend callbacks. */ From 632d31067be2f414c57955efcf29c79290cc749b Mon Sep 17 00:00:00 2001 From: Pin-yen Lin Date: Fri, 26 Sep 2025 18:23:18 +0800 Subject: [PATCH 04/19] PM: sleep: Do not wait on SYNC_STATE_ONLY device links Device links with DL_FLAG_SYNC_STATE_ONLY should not affect system suspend and resume, and functions like device_reorder_to_tail() and device_link_add() don't try to reorder the consumers with that flag. However, dpm_wait_for_consumers() and dpm_wait_for_suppliers() don't check thas flag before triggering dpm_wait(), leading to potential hang during suspend/resume. This can be reproduced on MT8186 Corsola Chromebook with devicetree like: usb-a-connector { compatible = "usb-a-connector"; port { usb_a_con: endpoint { remote-endpoint = <&usb_hs>; }; }; }; usb_host { compatible = "mediatek,mt8186-xhci", "mediatek,mtk-xhci"; port { usb_hs: endpoint { remote-endpoint = <&usb_a_con>; }; }; }; In this case, the two nodes form a cycle and a SYNC_STATE_ONLY devlink between usb_host (supplier) and usb-a-connector (consumer) is created. Address this by exporting device_link_flag_is_sync_state_only() and making dpm_wait_for_consumers() and dpm_wait_for_suppliers() use it when deciding if dpm_wait() should be called. Fixes: 05ef983e0d65a ("driver core: Add device link support for SYNC_STATE_ONLY flag") Signed-off-by: Pin-yen Lin Link: https://patch.msgid.link/20250926102320.4053167-1-treapking@chromium.org [ rjw: Subject and changelog edits ] Signed-off-by: Rafael J. Wysocki --- drivers/base/base.h | 1 + drivers/base/core.c | 2 +- drivers/base/power/main.c | 6 ++++-- 3 files changed, 6 insertions(+), 3 deletions(-) diff --git a/drivers/base/base.h b/drivers/base/base.h index 700aecd22fd3..86fa7fbb3548 100644 --- a/drivers/base/base.h +++ b/drivers/base/base.h @@ -248,6 +248,7 @@ void device_links_driver_cleanup(struct device *dev); void device_links_no_driver(struct device *dev); bool device_links_busy(struct device *dev); void device_links_unbind_consumers(struct device *dev); +bool device_link_flag_is_sync_state_only(u32 flags); void fw_devlink_drivers_done(void); void fw_devlink_probing_done(void); diff --git a/drivers/base/core.c b/drivers/base/core.c index d22d6b23e758..a54ec6df1058 100644 --- a/drivers/base/core.c +++ b/drivers/base/core.c @@ -287,7 +287,7 @@ static bool device_is_ancestor(struct device *dev, struct device *target) #define DL_MARKER_FLAGS (DL_FLAG_INFERRED | \ DL_FLAG_CYCLE | \ DL_FLAG_MANAGED) -static inline bool device_link_flag_is_sync_state_only(u32 flags) +bool device_link_flag_is_sync_state_only(u32 flags) { return (flags & ~DL_MARKER_FLAGS) == DL_FLAG_SYNC_STATE_ONLY; } diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c index e3a7f25bf8d3..c97f8b139dc2 100644 --- a/drivers/base/power/main.c +++ b/drivers/base/power/main.c @@ -278,7 +278,8 @@ static void dpm_wait_for_suppliers(struct device *dev, bool async) * walking. */ dev_for_each_link_to_supplier(link, dev) - if (READ_ONCE(link->status) != DL_STATE_DORMANT) + if (READ_ONCE(link->status) != DL_STATE_DORMANT && + !device_link_flag_is_sync_state_only(link->flags)) dpm_wait(link->supplier, async); device_links_read_unlock(idx); @@ -335,7 +336,8 @@ static void dpm_wait_for_consumers(struct device *dev, bool async) * unregistration). */ dev_for_each_link_to_consumer(link, dev) - if (READ_ONCE(link->status) != DL_STATE_DORMANT) + if (READ_ONCE(link->status) != DL_STATE_DORMANT && + !device_link_flag_is_sync_state_only(link->flags)) dpm_wait(link->consumer, async); device_links_read_unlock(idx); From ebc4ed14a4dbf51307102bb7ffc82ed6c16a37c2 Mon Sep 17 00:00:00 2001 From: Johan Hovold Date: Tue, 9 Sep 2025 09:38:19 +0200 Subject: [PATCH 05/19] cpufreq: mediatek: fix device leak on probe failure Make sure to drop the reference to the cci device taken by of_find_device_by_node() on probe failure (e.g. probe deferral). Fixes: 0daa47325bae ("cpufreq: mediatek: Link CCI device to CPU") Cc: Jia-Wei Chang Cc: Rex-BC Chen Signed-off-by: Johan Hovold Reviewed-by: AngeloGioacchino Del Regno Reviewed-by: Chen-Yu Tsai Signed-off-by: Viresh Kumar --- drivers/cpufreq/mediatek-cpufreq.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/drivers/cpufreq/mediatek-cpufreq.c b/drivers/cpufreq/mediatek-cpufreq.c index 00de1166188a..5d50a231f944 100644 --- a/drivers/cpufreq/mediatek-cpufreq.c +++ b/drivers/cpufreq/mediatek-cpufreq.c @@ -403,9 +403,11 @@ static int mtk_cpu_dvfs_info_init(struct mtk_cpu_dvfs_info *info, int cpu) } info->cpu_clk = clk_get(cpu_dev, "cpu"); - if (IS_ERR(info->cpu_clk)) - return dev_err_probe(cpu_dev, PTR_ERR(info->cpu_clk), - "cpu%d: failed to get cpu clk\n", cpu); + if (IS_ERR(info->cpu_clk)) { + ret = PTR_ERR(info->cpu_clk); + dev_err_probe(cpu_dev, ret, "cpu%d: failed to get cpu clk\n", cpu); + goto out_put_cci_dev; + } info->inter_clk = clk_get(cpu_dev, "intermediate"); if (IS_ERR(info->inter_clk)) { @@ -551,6 +553,10 @@ static int mtk_cpu_dvfs_info_init(struct mtk_cpu_dvfs_info *info, int cpu) out_free_mux_clock: clk_put(info->cpu_clk); +out_put_cci_dev: + if (info->soc_data->ccifreq_supported) + put_device(info->cci_dev); + return ret; } @@ -568,6 +574,8 @@ static void mtk_cpu_dvfs_info_release(struct mtk_cpu_dvfs_info *info) clk_put(info->inter_clk); dev_pm_opp_of_cpumask_remove_table(&info->cpus); dev_pm_opp_unregister_notifier(info->cpu_dev, &info->opp_nb); + if (info->soc_data->ccifreq_supported) + put_device(info->cci_dev); } static int mtk_cpufreq_init(struct cpufreq_policy *policy) From 24287f902095d845c6af9c2c369ba96877f5eb79 Mon Sep 17 00:00:00 2001 From: Thorsten Blum Date: Mon, 15 Sep 2025 15:59:54 +0200 Subject: [PATCH 06/19] rust: cpufreq: streamline find_supply_names Remove local variables from find_supply_names() and use .and_then() with the more concise kernel::kvec![] macro, instead of KVec::with_capacity() followed by .push() and Some(). No functional changes intended. Signed-off-by: Thorsten Blum Signed-off-by: Viresh Kumar --- drivers/cpufreq/rcpufreq_dt.rs | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) diff --git a/drivers/cpufreq/rcpufreq_dt.rs b/drivers/cpufreq/rcpufreq_dt.rs index 7e1fbf9a091f..224d063c7cec 100644 --- a/drivers/cpufreq/rcpufreq_dt.rs +++ b/drivers/cpufreq/rcpufreq_dt.rs @@ -28,15 +28,11 @@ fn find_supply_name_exact(dev: &Device, name: &str) -> Option { /// Finds supply name for the CPU from DT. fn find_supply_names(dev: &Device, cpu: cpu::CpuId) -> Option> { // Try "cpu0" for older DTs, fallback to "cpu". - let name = (cpu.as_u32() == 0) + (cpu.as_u32() == 0) .then(|| find_supply_name_exact(dev, "cpu0")) .flatten() - .or_else(|| find_supply_name_exact(dev, "cpu"))?; - - let mut list = KVec::with_capacity(1, GFP_KERNEL).ok()?; - list.push(name, GFP_KERNEL).ok()?; - - Some(list) + .or_else(|| find_supply_name_exact(dev, "cpu")) + .and_then(|name| kernel::kvec![name].ok()) } /// Represents the cpufreq dt device. From 0b1bb980fd7cae126ee3d59f817068a13e321b07 Mon Sep 17 00:00:00 2001 From: Aaron Kling Date: Thu, 28 Aug 2025 21:48:12 -0500 Subject: [PATCH 07/19] cpufreq: tegra186: Set target frequency for all cpus in policy The original commit set all cores in a cluster to a shared policy, but did not update set_target to apply a frequency change to all cores for the policy. This caused most cores to remain stuck at their boot frequency. Fixes: be4ae8c19492 ("cpufreq: tegra186: Share policy per cluster") Signed-off-by: Aaron Kling Reviewed-by: Mikko Perttunen Signed-off-by: Viresh Kumar --- drivers/cpufreq/tegra186-cpufreq.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/cpufreq/tegra186-cpufreq.c b/drivers/cpufreq/tegra186-cpufreq.c index 4270686fc3e3..a9e1f9bd2a4e 100644 --- a/drivers/cpufreq/tegra186-cpufreq.c +++ b/drivers/cpufreq/tegra186-cpufreq.c @@ -93,10 +93,14 @@ static int tegra186_cpufreq_set_target(struct cpufreq_policy *policy, { struct tegra186_cpufreq_data *data = cpufreq_get_driver_data(); struct cpufreq_frequency_table *tbl = policy->freq_table + index; - unsigned int edvd_offset = data->cpus[policy->cpu].edvd_offset; + unsigned int edvd_offset; u32 edvd_val = tbl->driver_data; + u32 cpu; - writel(edvd_val, data->regs + edvd_offset); + for_each_cpu(cpu, policy->cpus) { + edvd_offset = data->cpus[cpu].edvd_offset; + writel(edvd_val, data->regs + edvd_offset); + } return 0; } From ba6018929165fc914c665f071f8e8cdbac844a49 Mon Sep 17 00:00:00 2001 From: Aaron Kling Date: Thu, 28 Aug 2025 21:48:13 -0500 Subject: [PATCH 08/19] cpufreq: tegra186: Initialize all cores to max frequencies During initialization, the EDVD_COREx_VOLT_FREQ registers for some cores are still at reset values and not reflecting the actual frequency. This causes get calls to fail. Set all cores to their respective max frequency during probe to initialize the registers to working values. Suggested-by: Mikko Perttunen Signed-off-by: Aaron Kling Reviewed-by: Mikko Perttunen Signed-off-by: Viresh Kumar --- drivers/cpufreq/tegra186-cpufreq.c | 27 +++++++++++++++++++++------ 1 file changed, 21 insertions(+), 6 deletions(-) diff --git a/drivers/cpufreq/tegra186-cpufreq.c b/drivers/cpufreq/tegra186-cpufreq.c index a9e1f9bd2a4e..136ab102f636 100644 --- a/drivers/cpufreq/tegra186-cpufreq.c +++ b/drivers/cpufreq/tegra186-cpufreq.c @@ -136,13 +136,14 @@ static struct cpufreq_driver tegra186_cpufreq_driver = { static struct cpufreq_frequency_table *init_vhint_table( struct platform_device *pdev, struct tegra_bpmp *bpmp, - struct tegra186_cpufreq_cluster *cluster, unsigned int cluster_id) + struct tegra186_cpufreq_cluster *cluster, unsigned int cluster_id, + int *num_rates) { struct cpufreq_frequency_table *table; struct mrq_cpu_vhint_request req; struct tegra_bpmp_message msg; struct cpu_vhint_data *data; - int err, i, j, num_rates = 0; + int err, i, j; dma_addr_t phys; void *virt; @@ -172,6 +173,7 @@ static struct cpufreq_frequency_table *init_vhint_table( goto free; } + *num_rates = 0; for (i = data->vfloor; i <= data->vceil; i++) { u16 ndiv = data->ndiv[i]; @@ -182,10 +184,10 @@ static struct cpufreq_frequency_table *init_vhint_table( if (i > 0 && ndiv == data->ndiv[i - 1]) continue; - num_rates++; + (*num_rates)++; } - table = devm_kcalloc(&pdev->dev, num_rates + 1, sizeof(*table), + table = devm_kcalloc(&pdev->dev, *num_rates + 1, sizeof(*table), GFP_KERNEL); if (!table) { table = ERR_PTR(-ENOMEM); @@ -227,7 +229,9 @@ static int tegra186_cpufreq_probe(struct platform_device *pdev) { struct tegra186_cpufreq_data *data; struct tegra_bpmp *bpmp; - unsigned int i = 0, err; + unsigned int i = 0, err, edvd_offset; + int num_rates = 0; + u32 edvd_val, cpu; data = devm_kzalloc(&pdev->dev, struct_size(data, clusters, TEGRA186_NUM_CLUSTERS), @@ -250,10 +254,21 @@ static int tegra186_cpufreq_probe(struct platform_device *pdev) for (i = 0; i < TEGRA186_NUM_CLUSTERS; i++) { struct tegra186_cpufreq_cluster *cluster = &data->clusters[i]; - cluster->table = init_vhint_table(pdev, bpmp, cluster, i); + cluster->table = init_vhint_table(pdev, bpmp, cluster, i, &num_rates); if (IS_ERR(cluster->table)) { err = PTR_ERR(cluster->table); goto put_bpmp; + } else if (!num_rates) { + err = -EINVAL; + goto put_bpmp; + } + + for (cpu = 0; cpu < ARRAY_SIZE(tegra186_cpus); cpu++) { + if (data->cpus[cpu].bpmp_cluster_id == i) { + edvd_val = cluster->table[num_rates - 1].driver_data; + edvd_offset = data->cpus[cpu].edvd_offset; + writel(edvd_val, data->regs + edvd_offset); + } } } From 9a0abc39450a3123fd52533a662fbd37e0d1508c Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 26 Sep 2025 17:47:14 +0200 Subject: [PATCH 09/19] PM: runtime: Add auto-cleanup macros for "resume and get" operations It is generally useful to be able to automatically drop a device's runtime PM usage counter incremented by runtime PM operations that resume a device and bump up its usage counter [1]. To that end, add guard definition macros allowing pm_runtime_put() and pm_runtime_put_autosuspend() to be used for the auto-cleanup in those cases. Simply put, a piece of code like below: pm_runtime_get_sync(dev); ..... pm_runtime_put(dev); return 0; can be transformed with guard() like: guard(pm_runtime_active)(dev); ..... return 0; (see the pm_runtime_put() call is gone). However, it is better to do proper error handling in the majority of cases, so doing something like this instead of the above is recommended: ACQUIRE(pm_runtime_active_try, pm)(dev); if (ACQUIRE_ERR(pm_runtime_active_try, &pm)) return -ENXIO; ..... return 0; In all of the cases in which runtime PM is known to be enabled for the given device or the device can be regarded as operational (and so it can be accessed) with runtime PM disabled, a piece of code like: ret = pm_runtime_resume_and_get(dev); if (ret < 0) return ret; ..... pm_runtime_put(dev); return 0; can be changed as follows: ACQUIRE(pm_runtime_active_try, pm)(dev); ret = ACQUIRE_ERR(pm_runtime_active_try, &pm); if (ret < 0) return ret; ..... return 0; (again, see the pm_runtime_put() call is gone). Still, if the device cannot be accessed unless runtime PM has been enabled for it, the pm_runtime_active_try_enabled guard variant needs to be used, that is (in the context of the example above): ACQUIRE(pm_runtime_active_try_enabled, pm)(dev); ret = ACQUIRE_ERR(pm_runtime_active_try_enabled, &pm); if (ret < 0) return ret; ..... return 0; When the original code calls pm_runtime_put_autosuspend(), use one of the "auto" guard variants, pm_runtime_active_auto/_try/_enabled, so for example, a piece of code like: ret = pm_runtime_resume_and_get(dev); if (ret < 0) return ret; ..... pm_runtime_put_autosuspend(dev); return 0; will become: ACQUIRE(pm_runtime_active_auto_try_enabled, pm)(dev); ret = ACQUIRE_ERR(pm_runtime_active_auto_try_enabled, &pm); if (ret < 0) return ret; ..... return 0; Note that the cases in which the return value of pm_runtime_get_sync() is checked can also be handled with the help of the new guard macros. For example, a piece of code like: ret = pm_runtime_get_sync(dev); if (ret < 0) { pm_runtime_put(dev); return ret; } ..... pm_runtime_put(dev); return 0; can be rewritten as: ACQUIRE(pm_runtime_active_auto_try_enabled, pm)(dev); ret = ACQUIRE_ERR(pm_runtime_active_auto_try_enabled, &pm); if (ret < 0) return ret; ..... return 0; or pm_runtime_get_active_try can be used if transparent handling of disabled runtime PM is desirable. Link: https://lore.kernel.org/linux-pm/878qimv24u.wl-tiwai@suse.de/ [1] Link: https://lore.kernel.org/linux-pm/20250926150613.000073a4@huawei.com/ Signed-off-by: Rafael J. Wysocki Acked-by: Dan Williams Reviewed-by: Takashi Iwai Link: https://patch.msgid.link/2238241.irdbgypaU6@rafael.j.wysocki [ rjw: Fixed leftovers from the previous version in the changelog ] Reviewed-by: Jonathan Cameron Reviewed-by: Dhruva Gole Signed-off-by: Rafael J. Wysocki --- drivers/base/power/runtime.c | 2 ++ include/linux/pm_runtime.h | 44 ++++++++++++++++++++++++++++-------- 2 files changed, 37 insertions(+), 9 deletions(-) diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c index faa68bf9ef3d..0fc825066ede 100644 --- a/drivers/base/power/runtime.c +++ b/drivers/base/power/runtime.c @@ -799,6 +799,8 @@ static int rpm_resume(struct device *dev, int rpmflags) if (dev->power.runtime_status == RPM_ACTIVE && dev->power.last_status == RPM_ACTIVE) retval = 1; + else if (rpmflags & RPM_TRANSPARENT) + goto out; else retval = -EACCES; } diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index d1ff76e0e2d0..e5426bdd0c9f 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -21,6 +21,7 @@ #define RPM_GET_PUT 0x04 /* Increment/decrement the usage_count */ #define RPM_AUTO 0x08 /* Use autosuspend_delay */ +#define RPM_TRANSPARENT 0x10 /* Succeed if runtime PM is disabled */ /* * Use this for defining a set of PM operations to be used in all situations @@ -512,6 +513,19 @@ static inline int pm_runtime_get_sync(struct device *dev) return __pm_runtime_resume(dev, RPM_GET_PUT); } +static inline int pm_runtime_get_active(struct device *dev, int rpmflags) +{ + int ret; + + ret = __pm_runtime_resume(dev, RPM_GET_PUT | rpmflags); + if (ret < 0) { + pm_runtime_put_noidle(dev); + return ret; + } + + return 0; +} + /** * pm_runtime_resume_and_get - Bump up usage counter of a device and resume it. * @dev: Target device. @@ -522,15 +536,7 @@ static inline int pm_runtime_get_sync(struct device *dev) */ static inline int pm_runtime_resume_and_get(struct device *dev) { - int ret; - - ret = __pm_runtime_resume(dev, RPM_GET_PUT); - if (ret < 0) { - pm_runtime_put_noidle(dev); - return ret; - } - - return 0; + return pm_runtime_get_active(dev, 0); } /** @@ -610,6 +616,26 @@ static inline int pm_runtime_put_autosuspend(struct device *dev) return __pm_runtime_put_autosuspend(dev); } +DEFINE_GUARD(pm_runtime_active, struct device *, + pm_runtime_get_sync(_T), pm_runtime_put(_T)); +DEFINE_GUARD(pm_runtime_active_auto, struct device *, + pm_runtime_get_sync(_T), pm_runtime_put_autosuspend(_T)); +/* + * Use the following guards with ACQUIRE()/ACQUIRE_ERR(). + * + * The difference between the "_try" and "_try_enabled" variants is that the + * former do not produce an error when runtime PM is disabled for the given + * device. + */ +DEFINE_GUARD_COND(pm_runtime_active, _try, + pm_runtime_get_active(_T, RPM_TRANSPARENT)) +DEFINE_GUARD_COND(pm_runtime_active, _try_enabled, + pm_runtime_resume_and_get(_T)) +DEFINE_GUARD_COND(pm_runtime_active_auto, _try, + pm_runtime_get_active(_T, RPM_TRANSPARENT)) +DEFINE_GUARD_COND(pm_runtime_active_auto, _try_enabled, + pm_runtime_resume_and_get(_T)) + /** * pm_runtime_put_sync - Drop device usage counter and run "idle check" if 0. * @dev: Target device. From 8ff5aaa7b8c9fb3e66df6f440ee109d00f8d7a1b Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 26 Sep 2025 18:24:05 +0200 Subject: [PATCH 10/19] PCI/sysfs: Use runtime PM guard macro for auto-cleanup Use the newly introduced pm_runtime_active_try guard to simplify the code and add the proper error handling for PM runtime resume errors. Based on an earlier patch from Takashi Iwai [1]. Link: https://patch.msgid.link/20250919163147.4743-3-tiwai@suse.de [1] Acked-by: Bjorn Helgaas Signed-off-by: Rafael J. Wysocki Reviewed-by: Takashi Iwai Reviewed-by: Jonathan Cameron --- drivers/pci/pci-sysfs.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c index 5eea14c1f7f5..c96301026f01 100644 --- a/drivers/pci/pci-sysfs.c +++ b/drivers/pci/pci-sysfs.c @@ -1475,8 +1475,9 @@ static ssize_t reset_method_store(struct device *dev, return count; } - pm_runtime_get_sync(dev); - struct device *pmdev __free(pm_runtime_put) = dev; + ACQUIRE(pm_runtime_active_try, pm)(dev); + if (ACQUIRE_ERR(pm_runtime_active_try, &pm)) + return -ENXIO; if (sysfs_streq(buf, "default")) { pci_init_reset_methods(pdev); From d5e58ce1fb0f13a9a0845851f267ede3551cd9fe Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 26 Sep 2025 18:26:40 +0200 Subject: [PATCH 11/19] PM: runtime: Drop DEFINE_FREE() for pm_runtime_put() The DEFINE_FREE() for pm_runtime_put has been superseded by recently introduced runtime PM auto-cleanup macros and its only user has been converted to using one of the new macros, so drop it. Signed-off-by: Rafael J. Wysocki Reviewed-by: Dhruva Gole Reviewed-by: Takashi Iwai Reviewed-by: Jonathan Cameron --- include/linux/pm_runtime.h | 2 -- 1 file changed, 2 deletions(-) diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index e5426bdd0c9f..edb8aed5ef62 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -563,8 +563,6 @@ static inline int pm_runtime_put(struct device *dev) return __pm_runtime_idle(dev, RPM_GET_PUT | RPM_ASYNC); } -DEFINE_FREE(pm_runtime_put, struct device *, if (_T) pm_runtime_put(_T)) - /** * __pm_runtime_put_autosuspend - Drop device usage counter and queue autosuspend if 0. * @dev: Target device. From f97aef092e199c10a3da96ae79b571edd5362faa Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 26 Sep 2025 12:12:37 +0200 Subject: [PATCH 12/19] cpufreq: Make drivers using CPUFREQ_ETERNAL specify transition latency Commit a755d0e2d41b ("cpufreq: Honour transition_latency over transition_delay_us") caused platforms where cpuinfo.transition_latency is CPUFREQ_ETERNAL to get a very large transition latency whereas previously it had been capped at 10 ms (and later at 2 ms). This led to a user-observable regression between 6.6 and 6.12 as described by Shawn: "The dbs sampling_rate was 10000 us on 6.6 and suddently becomes 6442450 us (4294967295 / 1000 * 1.5) on 6.12 for these platforms because the default transition delay was dropped [...]. It slows down dbs governor's reacting to CPU loading change dramatically. Also, as transition_delay_us is used by schedutil governor as rate_limit_us, it shows a negative impact on device idle power consumption, because the device gets slightly less time in the lowest OPP." Evidently, the expectation of the drivers using CPUFREQ_ETERNAL as cpuinfo.transition_latency was that it would be capped by the core, but they may as well return a default transition latency value instead of CPUFREQ_ETERNAL and the core need not do anything with it. Accordingly, introduce CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS and make all of the drivers in question use it instead of CPUFREQ_ETERNAL. Also update the related Rust binding. Fixes: a755d0e2d41b ("cpufreq: Honour transition_latency over transition_delay_us") Closes: https://lore.kernel.org/linux-pm/20250922125929.453444-1-shawnguo2@yeah.net/ Reported-by: Shawn Guo Reviewed-by: Mario Limonciello (AMD) Reviewed-by: Jie Zhan Acked-by: Viresh Kumar Cc: 6.6+ # 6.6+ Signed-off-by: Rafael J. Wysocki Link: https://patch.msgid.link/2264949.irdbgypaU6@rafael.j.wysocki [ rjw: Fix typo in new symbol name, drop redundant type cast from Rust binding ] Tested-by: Shawn Guo # with cpufreq-dt driver Reviewed-by: Qais Yousef Signed-off-by: Rafael J. Wysocki --- drivers/cpufreq/cpufreq-dt.c | 2 +- drivers/cpufreq/imx6q-cpufreq.c | 2 +- drivers/cpufreq/mediatek-cpufreq-hw.c | 2 +- drivers/cpufreq/rcpufreq_dt.rs | 2 +- drivers/cpufreq/scmi-cpufreq.c | 2 +- drivers/cpufreq/scpi-cpufreq.c | 2 +- drivers/cpufreq/spear-cpufreq.c | 2 +- include/linux/cpufreq.h | 3 +++ rust/kernel/cpufreq.rs | 7 ++++--- 9 files changed, 14 insertions(+), 10 deletions(-) diff --git a/drivers/cpufreq/cpufreq-dt.c b/drivers/cpufreq/cpufreq-dt.c index 506437489b4d..7d5079fd1688 100644 --- a/drivers/cpufreq/cpufreq-dt.c +++ b/drivers/cpufreq/cpufreq-dt.c @@ -104,7 +104,7 @@ static int cpufreq_init(struct cpufreq_policy *policy) transition_latency = dev_pm_opp_get_max_transition_latency(cpu_dev); if (!transition_latency) - transition_latency = CPUFREQ_ETERNAL; + transition_latency = CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS; cpumask_copy(policy->cpus, priv->cpus); policy->driver_data = priv; diff --git a/drivers/cpufreq/imx6q-cpufreq.c b/drivers/cpufreq/imx6q-cpufreq.c index db1c88e9d3f9..e93697d3edfd 100644 --- a/drivers/cpufreq/imx6q-cpufreq.c +++ b/drivers/cpufreq/imx6q-cpufreq.c @@ -442,7 +442,7 @@ static int imx6q_cpufreq_probe(struct platform_device *pdev) } if (of_property_read_u32(np, "clock-latency", &transition_latency)) - transition_latency = CPUFREQ_ETERNAL; + transition_latency = CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS; /* * Calculate the ramp time for max voltage change in the diff --git a/drivers/cpufreq/mediatek-cpufreq-hw.c b/drivers/cpufreq/mediatek-cpufreq-hw.c index fce5aa5ceea0..ae4500ab4891 100644 --- a/drivers/cpufreq/mediatek-cpufreq-hw.c +++ b/drivers/cpufreq/mediatek-cpufreq-hw.c @@ -309,7 +309,7 @@ static int mtk_cpufreq_hw_cpu_init(struct cpufreq_policy *policy) latency = readl_relaxed(data->reg_bases[REG_FREQ_LATENCY]) * 1000; if (!latency) - latency = CPUFREQ_ETERNAL; + latency = CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS; policy->cpuinfo.transition_latency = latency; policy->fast_switch_possible = true; diff --git a/drivers/cpufreq/rcpufreq_dt.rs b/drivers/cpufreq/rcpufreq_dt.rs index 7e1fbf9a091f..3909022e1c74 100644 --- a/drivers/cpufreq/rcpufreq_dt.rs +++ b/drivers/cpufreq/rcpufreq_dt.rs @@ -123,7 +123,7 @@ fn init(policy: &mut cpufreq::Policy) -> Result { let mut transition_latency = opp_table.max_transition_latency_ns() as u32; if transition_latency == 0 { - transition_latency = cpufreq::ETERNAL_LATENCY_NS; + transition_latency = cpufreq::DEFAULT_TRANSITION_LATENCY_NS; } policy diff --git a/drivers/cpufreq/scmi-cpufreq.c b/drivers/cpufreq/scmi-cpufreq.c index 38c165d526d1..d2a110079f5f 100644 --- a/drivers/cpufreq/scmi-cpufreq.c +++ b/drivers/cpufreq/scmi-cpufreq.c @@ -294,7 +294,7 @@ static int scmi_cpufreq_init(struct cpufreq_policy *policy) latency = perf_ops->transition_latency_get(ph, domain); if (!latency) - latency = CPUFREQ_ETERNAL; + latency = CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS; policy->cpuinfo.transition_latency = latency; diff --git a/drivers/cpufreq/scpi-cpufreq.c b/drivers/cpufreq/scpi-cpufreq.c index dcbb0ae7dd47..e530345baddf 100644 --- a/drivers/cpufreq/scpi-cpufreq.c +++ b/drivers/cpufreq/scpi-cpufreq.c @@ -157,7 +157,7 @@ static int scpi_cpufreq_init(struct cpufreq_policy *policy) latency = scpi_ops->get_transition_latency(cpu_dev); if (!latency) - latency = CPUFREQ_ETERNAL; + latency = CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS; policy->cpuinfo.transition_latency = latency; diff --git a/drivers/cpufreq/spear-cpufreq.c b/drivers/cpufreq/spear-cpufreq.c index 707c71090cc3..2a1550e1aa21 100644 --- a/drivers/cpufreq/spear-cpufreq.c +++ b/drivers/cpufreq/spear-cpufreq.c @@ -182,7 +182,7 @@ static int spear_cpufreq_probe(struct platform_device *pdev) if (of_property_read_u32(np, "clock-latency", &spear_cpufreq.transition_latency)) - spear_cpufreq.transition_latency = CPUFREQ_ETERNAL; + spear_cpufreq.transition_latency = CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS; cnt = of_property_count_u32_elems(np, "cpufreq_tbl"); if (cnt <= 0) { diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h index 40966512ea18..bc8c083bc16a 100644 --- a/include/linux/cpufreq.h +++ b/include/linux/cpufreq.h @@ -32,6 +32,9 @@ */ #define CPUFREQ_ETERNAL (-1) + +#define CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS NSEC_PER_MSEC + #define CPUFREQ_NAME_LEN 16 /* Print length for names. Extra 1 space for accommodating '\n' in prints */ #define CPUFREQ_NAME_PLEN (CPUFREQ_NAME_LEN + 1) diff --git a/rust/kernel/cpufreq.rs b/rust/kernel/cpufreq.rs index eea57ba95f24..2ea735700ae7 100644 --- a/rust/kernel/cpufreq.rs +++ b/rust/kernel/cpufreq.rs @@ -39,7 +39,8 @@ const CPUFREQ_NAME_LEN: usize = bindings::CPUFREQ_NAME_LEN as usize; /// Default transition latency value in nanoseconds. -pub const ETERNAL_LATENCY_NS: u32 = bindings::CPUFREQ_ETERNAL as u32; +pub const DEFAULT_TRANSITION_LATENCY_NS: u32 = + bindings::CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS; /// CPU frequency driver flags. pub mod flags { @@ -400,13 +401,13 @@ pub fn to_table(mut self) -> Result { /// The following example demonstrates how to create a CPU frequency table. /// /// ``` -/// use kernel::cpufreq::{ETERNAL_LATENCY_NS, Policy}; +/// use kernel::cpufreq::{DEFAULT_TRANSITION_LATENCY_NS, Policy}; /// /// fn update_policy(policy: &mut Policy) { /// policy /// .set_dvfs_possible_from_any_cpu(true) /// .set_fast_switch_possible(true) -/// .set_transition_latency_ns(ETERNAL_LATENCY_NS); +/// .set_transition_latency_ns(DEFAULT_TRANSITION_LATENCY_NS); /// /// pr_info!("The policy details are: {:?}\n", (policy.cpu(), policy.cur())); /// } From f965d111e68f4a993cc44d487d416e3d954eea11 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 26 Sep 2025 12:19:41 +0200 Subject: [PATCH 13/19] cpufreq: CPPC: Avoid using CPUFREQ_ETERNAL as transition delay If cppc_get_transition_latency() returns CPUFREQ_ETERNAL to indicate a failure to retrieve the transition latency value from the platform firmware, the CPPC cpufreq driver will use that value (converted to microseconds) as the policy transition delay, but it is way too large for any practical use. Address this by making the driver use the cpufreq's default transition latency value (in microseconds) as the transition delay if CPUFREQ_ETERNAL is returned by cppc_get_transition_latency(). Fixes: d4f3388afd48 ("cpufreq / CPPC: Set platform specific transition_delay_us") Cc: 5.19+ # 5.19 Signed-off-by: Rafael J. Wysocki Reviewed-by: Mario Limonciello (AMD) Reviewed-by: Jie Zhan Acked-by: Viresh Kumar Reviewed-by: Qais Yousef --- drivers/cpufreq/cppc_cpufreq.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/drivers/cpufreq/cppc_cpufreq.c b/drivers/cpufreq/cppc_cpufreq.c index 12de0ac7bbaf..b71946937c52 100644 --- a/drivers/cpufreq/cppc_cpufreq.c +++ b/drivers/cpufreq/cppc_cpufreq.c @@ -308,6 +308,16 @@ static int cppc_verify_policy(struct cpufreq_policy_data *policy) return 0; } +static unsigned int __cppc_cpufreq_get_transition_delay_us(unsigned int cpu) +{ + unsigned int transition_latency_ns = cppc_get_transition_latency(cpu); + + if (transition_latency_ns == CPUFREQ_ETERNAL) + return CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS / NSEC_PER_USEC; + + return transition_latency_ns / NSEC_PER_USEC; +} + /* * The PCC subspace describes the rate at which platform can accept commands * on the shared PCC channel (including READs which do not count towards freq @@ -330,12 +340,12 @@ static unsigned int cppc_cpufreq_get_transition_delay_us(unsigned int cpu) return 10000; } } - return cppc_get_transition_latency(cpu) / NSEC_PER_USEC; + return __cppc_cpufreq_get_transition_delay_us(cpu); } #else static unsigned int cppc_cpufreq_get_transition_delay_us(unsigned int cpu) { - return cppc_get_transition_latency(cpu) / NSEC_PER_USEC; + return __cppc_cpufreq_get_transition_delay_us(cpu); } #endif From c28a280bd465690981099cd6e43dfcfa5c28b133 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 26 Sep 2025 12:29:50 +0200 Subject: [PATCH 14/19] ACPI: CPPC: Do not use CPUFREQ_ETERNAL as an error value Instead of using CPUFREQ_ETERNAL for signaling an error condition in cppc_get_transition_latency(), change the return value type of that function to int and make it return a proper negative error code on failures. No intentional functional impact. Reviewed-by: Mario Limonciello (AMD) Reviewed-by: Jie Zhan Signed-off-by: Rafael J. Wysocki Acked-by: Viresh Kumar Reviewed-by: Qais Yousef --- drivers/acpi/cppc_acpi.c | 16 +++++++--------- drivers/cpufreq/amd-pstate.c | 8 ++++---- drivers/cpufreq/cppc_cpufreq.c | 4 ++-- include/acpi/cppc_acpi.h | 6 +++--- 4 files changed, 16 insertions(+), 18 deletions(-) diff --git a/drivers/acpi/cppc_acpi.c b/drivers/acpi/cppc_acpi.c index 6b649031808f..ab4651205e8a 100644 --- a/drivers/acpi/cppc_acpi.c +++ b/drivers/acpi/cppc_acpi.c @@ -1876,7 +1876,7 @@ EXPORT_SYMBOL_GPL(cppc_set_perf); * If desired_reg is in the SystemMemory or SystemIo ACPI address space, * then assume there is no latency. */ -unsigned int cppc_get_transition_latency(int cpu_num) +int cppc_get_transition_latency(int cpu_num) { /* * Expected transition latency is based on the PCCT timing values @@ -1889,31 +1889,29 @@ unsigned int cppc_get_transition_latency(int cpu_num) * completion of a command before issuing the next command, * in microseconds. */ - unsigned int latency_ns = 0; struct cpc_desc *cpc_desc; struct cpc_register_resource *desired_reg; int pcc_ss_id = per_cpu(cpu_pcc_subspace_idx, cpu_num); struct cppc_pcc_data *pcc_ss_data; + int latency_ns = 0; cpc_desc = per_cpu(cpc_desc_ptr, cpu_num); if (!cpc_desc) - return CPUFREQ_ETERNAL; + return -ENODATA; desired_reg = &cpc_desc->cpc_regs[DESIRED_PERF]; if (CPC_IN_SYSTEM_MEMORY(desired_reg) || CPC_IN_SYSTEM_IO(desired_reg)) return 0; - else if (!CPC_IN_PCC(desired_reg)) - return CPUFREQ_ETERNAL; - if (pcc_ss_id < 0) - return CPUFREQ_ETERNAL; + if (!CPC_IN_PCC(desired_reg) || pcc_ss_id < 0) + return -ENODATA; pcc_ss_data = pcc_data[pcc_ss_id]; if (pcc_ss_data->pcc_mpar) latency_ns = 60 * (1000 * 1000 * 1000 / pcc_ss_data->pcc_mpar); - latency_ns = max(latency_ns, pcc_ss_data->pcc_nominal * 1000); - latency_ns = max(latency_ns, pcc_ss_data->pcc_mrtt * 1000); + latency_ns = max_t(int, latency_ns, pcc_ss_data->pcc_nominal * 1000); + latency_ns = max_t(int, latency_ns, pcc_ss_data->pcc_mrtt * 1000); return latency_ns; } diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c index b4c79fde1979..298e92d8cc03 100644 --- a/drivers/cpufreq/amd-pstate.c +++ b/drivers/cpufreq/amd-pstate.c @@ -872,10 +872,10 @@ static void amd_pstate_update_limits(struct cpufreq_policy *policy) */ static u32 amd_pstate_get_transition_delay_us(unsigned int cpu) { - u32 transition_delay_ns; + int transition_delay_ns; transition_delay_ns = cppc_get_transition_latency(cpu); - if (transition_delay_ns == CPUFREQ_ETERNAL) { + if (transition_delay_ns < 0) { if (cpu_feature_enabled(X86_FEATURE_AMD_FAST_CPPC)) return AMD_PSTATE_FAST_CPPC_TRANSITION_DELAY; else @@ -891,10 +891,10 @@ static u32 amd_pstate_get_transition_delay_us(unsigned int cpu) */ static u32 amd_pstate_get_transition_latency(unsigned int cpu) { - u32 transition_latency; + int transition_latency; transition_latency = cppc_get_transition_latency(cpu); - if (transition_latency == CPUFREQ_ETERNAL) + if (transition_latency < 0) return AMD_PSTATE_TRANSITION_LATENCY; return transition_latency; diff --git a/drivers/cpufreq/cppc_cpufreq.c b/drivers/cpufreq/cppc_cpufreq.c index b71946937c52..e23d9abea135 100644 --- a/drivers/cpufreq/cppc_cpufreq.c +++ b/drivers/cpufreq/cppc_cpufreq.c @@ -310,9 +310,9 @@ static int cppc_verify_policy(struct cpufreq_policy_data *policy) static unsigned int __cppc_cpufreq_get_transition_delay_us(unsigned int cpu) { - unsigned int transition_latency_ns = cppc_get_transition_latency(cpu); + int transition_latency_ns = cppc_get_transition_latency(cpu); - if (transition_latency_ns == CPUFREQ_ETERNAL) + if (transition_latency_ns < 0) return CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS / NSEC_PER_USEC; return transition_latency_ns / NSEC_PER_USEC; diff --git a/include/acpi/cppc_acpi.h b/include/acpi/cppc_acpi.h index 20f3d62e7a16..13fa81504844 100644 --- a/include/acpi/cppc_acpi.h +++ b/include/acpi/cppc_acpi.h @@ -160,7 +160,7 @@ extern unsigned int cppc_khz_to_perf(struct cppc_perf_caps *caps, unsigned int f extern bool acpi_cpc_valid(void); extern bool cppc_allow_fast_switch(void); extern int acpi_get_psd_map(unsigned int cpu, struct cppc_cpudata *cpu_data); -extern unsigned int cppc_get_transition_latency(int cpu); +extern int cppc_get_transition_latency(int cpu); extern bool cpc_ffh_supported(void); extern bool cpc_supported_by_cpu(void); extern int cpc_read_ffh(int cpunum, struct cpc_reg *reg, u64 *val); @@ -216,9 +216,9 @@ static inline bool cppc_allow_fast_switch(void) { return false; } -static inline unsigned int cppc_get_transition_latency(int cpu) +static inline int cppc_get_transition_latency(int cpu) { - return CPUFREQ_ETERNAL; + return -ENODATA; } static inline bool cpc_ffh_supported(void) { From 950c6451a5c38d375993c3b9da427e2e69b01c30 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 26 Sep 2025 12:31:47 +0200 Subject: [PATCH 15/19] cpufreq: Drop unused symbol CPUFREQ_ETERNAL Drop CPUFREQ_ETERNAL that has no users any more along with all references to it in the documentation. No functional impact. Reviewed-by: Mario Limonciello (AMD) Reviewed-by: Jie Zhan Signed-off-by: Rafael J. Wysocki Acked-by: Viresh Kumar Reviewed-by: Qais Yousef --- Documentation/admin-guide/pm/cpufreq.rst | 4 ---- Documentation/cpu-freq/cpu-drivers.rst | 3 +-- Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst | 3 +-- Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst | 3 +-- include/linux/cpufreq.h | 5 ----- 5 files changed, 3 insertions(+), 15 deletions(-) diff --git a/Documentation/admin-guide/pm/cpufreq.rst b/Documentation/admin-guide/pm/cpufreq.rst index cacb9f0307dd..738d7b4dc33a 100644 --- a/Documentation/admin-guide/pm/cpufreq.rst +++ b/Documentation/admin-guide/pm/cpufreq.rst @@ -274,10 +274,6 @@ are the following: The time it takes to switch the CPUs belonging to this policy from one P-state to another, in nanoseconds. - If unknown or if known to be so high that the scaling driver does not - work with the `ondemand`_ governor, -1 (:c:macro:`CPUFREQ_ETERNAL`) - will be returned by reads from this attribute. - ``related_cpus`` List of all (online and offline) CPUs belonging to this policy. diff --git a/Documentation/cpu-freq/cpu-drivers.rst b/Documentation/cpu-freq/cpu-drivers.rst index d84ededb66f9..c5635ac3de54 100644 --- a/Documentation/cpu-freq/cpu-drivers.rst +++ b/Documentation/cpu-freq/cpu-drivers.rst @@ -109,8 +109,7 @@ Then, the driver must fill in the following values: +-----------------------------------+--------------------------------------+ |policy->cpuinfo.transition_latency | the time it takes on this CPU to | | | switch between two frequencies in | -| | nanoseconds (if appropriate, else | -| | specify CPUFREQ_ETERNAL) | +| | nanoseconds | +-----------------------------------+--------------------------------------+ |policy->cur | The current operating frequency of | | | this CPU (if appropriate) | diff --git a/Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst b/Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst index 2ca92042767b..8238f4c6e4f5 100644 --- a/Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst +++ b/Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst @@ -112,8 +112,7 @@ CPUfreq核心层注册一个cpufreq_driver结构体。 | | | +-----------------------------------+--------------------------------------+ |policy->cpuinfo.transition_latency | CPU在两个频率之间切换所需的时间，以 | -| | 纳秒为单位（如不适用，设定为 | -| | CPUFREQ_ETERNAL） | +| | 纳秒为单位 | | | | +-----------------------------------+--------------------------------------+ |policy->cur | 该CPU当前的工作频率(如适用) | diff --git a/Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst b/Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst index add3de2d4523..5435c3928d4b 100644 --- a/Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst +++ b/Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst @@ -112,8 +112,7 @@ CPUfreq核心層註冊一個cpufreq_driver結構體。 | | | +-----------------------------------+--------------------------------------+ |policy->cpuinfo.transition_latency | CPU在兩個頻率之間切換所需的時間，以 | -| | 納秒爲單位（如不適用，設定爲 | -| | CPUFREQ_ETERNAL） | +| | 納秒爲單位 | | | | +-----------------------------------+--------------------------------------+ |policy->cur | 該CPU當前的工作頻率(如適用) | diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h index bc8c083bc16a..0465d1e6f72a 100644 --- a/include/linux/cpufreq.h +++ b/include/linux/cpufreq.h @@ -26,13 +26,8 @@ *********************************************************************/ /* * Frequency values here are CPU kHz - * - * Maximum transition latency is in nanoseconds - if it's unknown, - * CPUFREQ_ETERNAL shall be used. */ -#define CPUFREQ_ETERNAL (-1) - #define CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS NSEC_PER_MSEC #define CPUFREQ_NAME_LEN 16 From 384b52ce32110db974d3b61d463af48347eb73fb Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Wed, 1 Oct 2025 19:34:29 +0200 Subject: [PATCH 16/19] PM: runtime: Introduce one more usage counter guard Follow previous commit 9a0abc39450a ("PM: runtime: Add auto-cleanup macros for "resume and get" operations") and define a runtime PM usage counter guard in which pm_runtime_get_noresume() and pm_runtime_put_noidle() will be used for incrementing and decrementing it, respectively. Signed-off-by: Rafael J. Wysocki Reviewed-by: Jonathan Cameron --- include/linux/pm_runtime.h | 3 +++ 1 file changed, 3 insertions(+) diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index edb8aed5ef62..a3f44f6c2da1 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -614,6 +614,9 @@ static inline int pm_runtime_put_autosuspend(struct device *dev) return __pm_runtime_put_autosuspend(dev); } +DEFINE_GUARD(pm_runtime_noresume, struct device *, + pm_runtime_get_noresume(_T), pm_runtime_put_noidle(_T)); + DEFINE_GUARD(pm_runtime_active, struct device *, pm_runtime_get_sync(_T), pm_runtime_put(_T)); DEFINE_GUARD(pm_runtime_active_auto, struct device *, From 92158fae2ed986f44347fc5b9a269830862c1529 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 3 Oct 2025 12:29:31 +0300 Subject: [PATCH 17/19] PM: runtime: Fix error checking for kunit_device_register() The kunit_device_register() function never returns NULL, it returns error pointers. Update the assertions to use KUNIT_ASSERT_NOT_ERR_OR_NULL() instead of checking for NULL. Fixes: 7f7acd193ba8 ("PM: runtime: Add basic kunit tests for API contracts") Signed-off-by: Dan Carpenter Reviewed-by: Brian Norris Signed-off-by: Rafael J. Wysocki --- drivers/base/power/runtime-test.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/base/power/runtime-test.c b/drivers/base/power/runtime-test.c index eca9885e807d..477feca804c7 100644 --- a/drivers/base/power/runtime-test.c +++ b/drivers/base/power/runtime-test.c @@ -14,7 +14,7 @@ static void pm_runtime_depth_test(struct kunit *test) { struct device *dev = kunit_device_register(test, DEVICE_NAME); - KUNIT_ASSERT_PTR_NE(test, NULL, dev); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, dev); pm_runtime_enable(dev); @@ -32,7 +32,7 @@ static void pm_runtime_already_suspended_test(struct kunit *test) { struct device *dev = kunit_device_register(test, DEVICE_NAME); - KUNIT_ASSERT_PTR_NE(test, NULL, dev); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, dev); pm_runtime_enable(dev); KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); @@ -70,7 +70,7 @@ static void pm_runtime_idle_test(struct kunit *test) { struct device *dev = kunit_device_register(test, DEVICE_NAME); - KUNIT_ASSERT_PTR_NE(test, NULL, dev); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, dev); pm_runtime_enable(dev); @@ -91,7 +91,7 @@ static void pm_runtime_disabled_test(struct kunit *test) { struct device *dev = kunit_device_register(test, DEVICE_NAME); - KUNIT_ASSERT_PTR_NE(test, NULL, dev); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, dev); /* Never called pm_runtime_enable() */ KUNIT_EXPECT_FALSE(test, pm_runtime_enabled(dev)); @@ -131,7 +131,7 @@ static void pm_runtime_error_test(struct kunit *test) { struct device *dev = kunit_device_register(test, DEVICE_NAME); - KUNIT_ASSERT_PTR_NE(test, NULL, dev); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, dev); pm_runtime_enable(dev); KUNIT_EXPECT_TRUE(test, pm_runtime_suspended(dev)); @@ -214,7 +214,7 @@ static void pm_runtime_probe_active_test(struct kunit *test) { struct device *dev = kunit_device_register(test, DEVICE_NAME); - KUNIT_ASSERT_PTR_NE(test, NULL, dev); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, dev); KUNIT_EXPECT_TRUE(test, pm_runtime_status_suspended(dev)); From 989ed3cad2fdb7921bf0f9f0f730e1bb065946c2 Mon Sep 17 00:00:00 2001 From: Gopi Krishna Menon Date: Fri, 3 Oct 2025 23:06:00 +0530 Subject: [PATCH 18/19] docs/zh_TW: Fix malformed table Running "make htmldocs" generate the following build error in zh_TW/cpu-freq/cpu-drivers.rst: Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst:109: ERROR: Malformed table. Fix the table formatting to resolve the error. Signed-off-by: Gopi Krishna Menon Signed-off-by: Rafael J. Wysocki --- Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst b/Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst index 5435c3928d4b..7f751a7add56 100644 --- a/Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst +++ b/Documentation/translations/zh_TW/cpu-freq/cpu-drivers.rst @@ -112,7 +112,7 @@ CPUfreq核心層註冊一個cpufreq_driver結構體。 | | | +-----------------------------------+--------------------------------------+ |policy->cpuinfo.transition_latency | CPU在兩個頻率之間切換所需的時間，以 | -| | 納秒爲單位 | +| | 納秒爲單位 | | | | +-----------------------------------+--------------------------------------+ |policy->cur | 該CPU當前的工作頻率(如適用) | From 7e8f305a081e22ce81aab7f7b9ce01437cbd38b3 Mon Sep 17 00:00:00 2001 From: Gopi Krishna Menon Date: Fri, 3 Oct 2025 23:05:59 +0530 Subject: [PATCH 19/19] docs/zh_CN: Fix malformed table Running "make htmldocs" generates the following build error in zh_CN/cpu-freq/cpu-drivers.rst: Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst:109: ERROR: Malformed table. Fix the table formatting to resolve the error. Signed-off-by: Gopi Krishna Menon Signed-off-by: Rafael J. Wysocki --- Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst b/Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst index 8238f4c6e4f5..2d5e84d8e58d 100644 --- a/Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst +++ b/Documentation/translations/zh_CN/cpu-freq/cpu-drivers.rst @@ -112,7 +112,7 @@ CPUfreq核心层注册一个cpufreq_driver结构体。 | | | +-----------------------------------+--------------------------------------+ |policy->cpuinfo.transition_latency | CPU在两个频率之间切换所需的时间，以 | -| | 纳秒为单位 | +| | 纳秒为单位 | | | | +-----------------------------------+--------------------------------------+ |policy->cur | 该CPU当前的工作频率(如适用) |