linux-stable.git/drivers/cpufreq, branch v6.14 Linux kernel stable tree Merge tag 'amd-pstate-v6.14-2025-02-06' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/superm1/linux 2025-02-06T19:39:43+00:00 Rafael J. Wysocki rafael.j.wysocki@intel.com 2025-02-06T19:39:43+00:00 b3cc5afc4d972ca30dc5d25ed5e1056e62b5efa8 Merge amd-pstate driver fixes for 6.14-rc2 from Mario Limonciello: "* Fix some error cleanup paths with mutex use and boost * Fix a ref counting issue * Fix a schedutil issue" * tag 'amd-pstate-v6.14-2025-02-06' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/superm1/linux: cpufreq/amd-pstate: Fix cpufreq_policy ref counting cpufreq/amd-pstate: Fix max_perf updation with schedutil cpufreq/amd-pstate: Remove the goto label in amd_pstate_update_limits cpufreq/amd-pstate: Fix per-policy boost flag incorrect when fail 
Merge amd-pstate driver fixes for 6.14-rc2 from Mario Limonciello:

"* Fix some error cleanup paths with mutex use and boost
 * Fix a ref counting issue
 * Fix a schedutil issue"

* tag 'amd-pstate-v6.14-2025-02-06' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/superm1/linux:
  cpufreq/amd-pstate: Fix cpufreq_policy ref counting
  cpufreq/amd-pstate: Fix max_perf updation with schedutil
  cpufreq/amd-pstate: Remove the goto label in amd_pstate_update_limits
  cpufreq/amd-pstate: Fix per-policy boost flag incorrect when fail
cpufreq/amd-pstate: Fix cpufreq_policy ref counting 2025-02-06T19:19:36+00:00 Dhananjay Ugwekar dhananjay.ugwekar@amd.com 2025-02-05T11:25:20+00:00 3ace20038e19f23fe73259513f1f08d4bf1a3c83 amd_pstate_update_limits() takes a cpufreq_policy reference but doesn't decrement the refcount in one of the exit paths, fix that. Fixes: 45722e777fd9 ("cpufreq: amd-pstate: Optimize amd_pstate_update_limits()") Signed-off-by: Dhananjay Ugwekar <dhananjay.ugwekar@amd.com> Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Link: https://lore.kernel.org/r/20250205112523.201101-10-dhananjay.ugwekar@amd.com Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> 
amd_pstate_update_limits() takes a cpufreq_policy reference but doesn't
decrement the refcount in one of the exit paths, fix that.

Fixes: 45722e777fd9 ("cpufreq: amd-pstate: Optimize amd_pstate_update_limits()")
Signed-off-by: Dhananjay Ugwekar <dhananjay.ugwekar@amd.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20250205112523.201101-10-dhananjay.ugwekar@amd.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
cpufreq: prevent NULL dereference in cpufreq_online() 2025-02-05T20:02:39+00:00 Aboorva Devarajan aboorvad@linux.ibm.com 2025-02-05T18:13:47+00:00 0813fd2e14ca6ecd4e6ba005a9766f08e26020d7 Ensure cpufreq_driver->set_boost is non-NULL before using it in cpufreq_online() to prevent a potential NULL pointer dereference. Reported-by: Gautam Menghani <gautam@linux.ibm.com> Closes: https://lore.kernel.org/all/c9e56c5f54cc33338762c94e9bed7b5a0d5de812.camel@linux.ibm.com/ Fixes: dd016f379ebc ("cpufreq: Introduce a more generic way to set default per-policy boost flag") Suggested-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Aboorva Devarajan <aboorvad@linux.ibm.com> Link: https://patch.msgid.link/20250205181347.2079272-1-aboorvad@linux.ibm.com [ rjw: Minor edits in the subject and changelog ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 
Ensure cpufreq_driver->set_boost is non-NULL before using it in
cpufreq_online() to prevent a potential NULL pointer dereference.

Reported-by: Gautam Menghani <gautam@linux.ibm.com>
Closes: https://lore.kernel.org/all/c9e56c5f54cc33338762c94e9bed7b5a0d5de812.camel@linux.ibm.com/
Fixes: dd016f379ebc ("cpufreq: Introduce a more generic way to set default per-policy boost flag")
Suggested-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Aboorva Devarajan <aboorvad@linux.ibm.com>
Link: https://patch.msgid.link/20250205181347.2079272-1-aboorvad@linux.ibm.com
[ rjw: Minor edits in the subject and changelog ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
cpufreq: airoha: modify CONFIG_OF dependency 2025-02-05T19:59:42+00:00 Arnd Bergmann arnd@arndb.de 2025-02-03T10:54:15+00:00 90508a1bb8f00618fa12cb2ad2276bc783656fc5 Compile-testing without CONFIG_OF leads to a harmless build warning: drivers/cpufreq/airoha-cpufreq.c:109:34: error: 'airoha_cpufreq_match_list' defined but not used [-Werror=unused-const-variable=] 109 | static const struct of_device_id airoha_cpufreq_match_list[] __initconst = { | ^~~~~~~~~~~~~~~~~~~~~~~~~ It would be possible to mark the variable as __maybe_unused to shut up that warning, but a Kconfig dependency seems more appropriate as this still allows build testing in allmodconfig and randconfig builds on all architectures. An earlier commit, b865a8404642 ("cpufreq: airoha: Depends on OF"), tried to fix it incorrectly. ARCH_AIROHA already requires CONFIG_OF, so this change does nothing, and the dependency is still missing for the COMPILE_TEST case. Fix it properly. Fixes: 84cf9e541ccc ("cpufreq: airoha: Add EN7581 CPUFreq SMCCC driver") Fixes: b865a8404642 ("cpufreq: airoha: Depends on OF") Signed-off-by: Arnd Bergmann <arnd@arndb.de> [ Viresh: updated commit log and fixed rebase conflict ] Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Link: https://patch.msgid.link/9d51d2710061dfa7f2568287c6ed125b858b7318.1738580005.git.viresh.kumar@linaro.org Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 
Compile-testing without CONFIG_OF leads to a harmless build warning:

drivers/cpufreq/airoha-cpufreq.c:109:34: error: 'airoha_cpufreq_match_list' defined but not used [-Werror=unused-const-variable=]
  109 | static const struct of_device_id airoha_cpufreq_match_list[] __initconst = {
      |                                  ^~~~~~~~~~~~~~~~~~~~~~~~~

It would be possible to mark the variable as __maybe_unused to shut up
that warning, but a Kconfig dependency seems more appropriate as this still
allows build testing in allmodconfig and randconfig builds on all
architectures.

An earlier commit, b865a8404642 ("cpufreq: airoha: Depends on OF"),
tried to fix it incorrectly. ARCH_AIROHA already requires CONFIG_OF, so
this change does nothing, and the dependency is still missing for the
COMPILE_TEST case.

Fix it properly.

Fixes: 84cf9e541ccc ("cpufreq: airoha: Add EN7581 CPUFreq SMCCC driver")
Fixes: b865a8404642 ("cpufreq: airoha: Depends on OF")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
[ Viresh: updated commit log and fixed rebase conflict ]
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://patch.msgid.link/9d51d2710061dfa7f2568287c6ed125b858b7318.1738580005.git.viresh.kumar@linaro.org
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
cpufreq/amd-pstate: Fix max_perf updation with schedutil 2025-02-05T18:18:27+00:00 Dhananjay Ugwekar dhananjay.ugwekar@amd.com 2025-02-05T11:25:13+00:00 55db9b73c3a77544efc671d5e796d9674772c330 In adjust_perf() callback, we are setting the max_perf to highest_perf, as opposed to the correct limit value i.e. max_limit_perf. Fix that. Fixes: 3f7b835fa4d0 ("cpufreq/amd-pstate: Move limit updating code") Signed-off-by: Dhananjay Ugwekar <dhananjay.ugwekar@amd.com> Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Link: https://lore.kernel.org/r/20250205112523.201101-3-dhananjay.ugwekar@amd.com Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> 
In adjust_perf() callback, we are setting the max_perf to highest_perf,
as opposed to the correct limit value i.e. max_limit_perf. Fix that.

Fixes: 3f7b835fa4d0 ("cpufreq/amd-pstate: Move limit updating code")
Signed-off-by: Dhananjay Ugwekar <dhananjay.ugwekar@amd.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20250205112523.201101-3-dhananjay.ugwekar@amd.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
cpufreq/amd-pstate: Remove the goto label in amd_pstate_update_limits 2025-02-05T18:18:27+00:00 Dhananjay Ugwekar dhananjay.ugwekar@amd.com 2025-02-05T11:25:12+00:00 d364eee14c682b141f4667efc3c65191339d88bd Scope based guard/cleanup macros should not be used together with goto labels. Hence, remove the goto label. Fixes: 6c093d5a5b73 ("cpufreq/amd-pstate: convert mutex use to guard()") Signed-off-by: Dhananjay Ugwekar <dhananjay.ugwekar@amd.com> Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Link: https://lore.kernel.org/r/20250205112523.201101-2-dhananjay.ugwekar@amd.com Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> 
Scope based guard/cleanup macros should not be used together with goto
labels. Hence, remove the goto label.

Fixes: 6c093d5a5b73 ("cpufreq/amd-pstate: convert mutex use to guard()")
Signed-off-by: Dhananjay Ugwekar <dhananjay.ugwekar@amd.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20250205112523.201101-2-dhananjay.ugwekar@amd.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
cpufreq/amd-pstate: Fix per-policy boost flag incorrect when fail 2025-02-03T06:04:23+00:00 Lifeng Zheng zhenglifeng1@huawei.com 2025-01-10T09:19:49+00:00 fa803513ab68ba07369643393f1754b845160030 Commit c8c68c38b56f ("cpufreq: amd-pstate: initialize core precision boost state") sets per-policy boost flag to false when boost fail. However, this boost flag will be set to reverse value in store_local_boost() and cpufreq_boost_trigger_state() in cpufreq.c. This will cause the per-policy boost flag set to true when fail to set boost. Remove the extra assignment in amd_pstate_set_boost() and keep all operations on per-policy boost flag outside of set_boost() to fix this problem. Fixes: c8c68c38b56f ("cpufreq: amd-pstate: initialize core precision boost state") Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com> Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Link: https://lore.kernel.org/r/20250110091949.3610770-1-zhenglifeng1@huawei.com Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> 
Commit c8c68c38b56f ("cpufreq: amd-pstate: initialize core precision
boost state") sets per-policy boost flag to false when boost fail.
However, this boost flag will be set to reverse value in
store_local_boost() and cpufreq_boost_trigger_state() in cpufreq.c. This
will cause the per-policy boost flag set to true when fail to set boost.
Remove the extra assignment in amd_pstate_set_boost() and keep all
operations on per-policy boost flag outside of set_boost() to fix this
problem.

Fixes: c8c68c38b56f ("cpufreq: amd-pstate: initialize core precision boost state")
Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20250110091949.3610770-1-zhenglifeng1@huawei.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
cpufreq: airoha: Depends on OF 2025-01-29T09:56:11+00:00 Viresh Kumar viresh.kumar@linaro.org 2025-01-28T07:51:23+00:00 b865a8404642279e53644fc7288d172afd6a170e The Airoha cpufreq depends on OF and must be marked as such. With the kernel compiled without OF support, we get following warning: drivers/cpufreq/airoha-cpufreq.c:109:34: warning: 'airoha_cpufreq_match_list' defined but not used [-Wunused-const-variable=] 109 | static const struct of_device_id airoha_cpufreq_match_list[] __initconst = { | ^~~~~~~~~~~~~~~~~~~~~~~~~ Reported-by: kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202501251941.0fXlcd1D-lkp@intel.com/ Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Link: https://patch.msgid.link/455e18c947bd9529701a2f1c796f0f934d1354d7.1738050679.git.viresh.kumar@linaro.org Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 
The Airoha cpufreq depends on OF and must be marked as such. With the
kernel compiled without OF support, we get following warning:

drivers/cpufreq/airoha-cpufreq.c:109:34: warning: 'airoha_cpufreq_match_list' defined but not used [-Wunused-const-variable=]
    109 | static const struct of_device_id airoha_cpufreq_match_list[] __initconst = {
        |                                  ^~~~~~~~~~~~~~~~~~~~~~~~~

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202501251941.0fXlcd1D-lkp@intel.com/
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://patch.msgid.link/455e18c947bd9529701a2f1c796f0f934d1354d7.1738050679.git.viresh.kumar@linaro.org
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
cpufreq: ACPI: Remove set_boost in acpi_cpufreq_cpu_init() 2025-01-23T20:06:33+00:00 Lifeng Zheng zhenglifeng1@huawei.com 2025-01-17T10:14:57+00:00 2b16c631832df6cf8782fb1fdc7df8a4f03f4f16 At the end of cpufreq_online() in cpufreq.c, set_boost is executed and the per-policy boost flag is set to mirror the cpufreq_driver boost, so it is not necessary to run set_boost in acpi_cpufreq_cpu_init(). Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Link: https://patch.msgid.link/20250117101457.1530653-5-zhenglifeng1@huawei.com [ rjw: Changelog edits ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 
At the end of cpufreq_online() in cpufreq.c, set_boost is executed and
the per-policy boost flag is set to mirror the cpufreq_driver boost, so
it is not necessary to run set_boost in acpi_cpufreq_cpu_init().

Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://patch.msgid.link/20250117101457.1530653-5-zhenglifeng1@huawei.com
[ rjw: Changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
cpufreq: CPPC: Fix wrong max_freq in policy initialization 2025-01-23T20:06:33+00:00 Lifeng Zheng zhenglifeng1@huawei.com 2025-01-17T10:14:56+00:00 03d8b4e76266e11662c5e544854b737843173e2d In policy initialization, policy->max and policy->cpuinfo.max_freq are always set to the value calculated from caps->nominal_perf. This will cause the frequency stay on base frequency even if the policy is already boosted when a CPU is going online. Fix this by using policy->boost_enabled to determine which value should be set. Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Link: https://patch.msgid.link/20250117101457.1530653-4-zhenglifeng1@huawei.com [ rjw: Changelog edits ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 
In policy initialization, policy->max and policy->cpuinfo.max_freq are
always set to the value calculated from caps->nominal_perf.

This will cause the frequency stay on base frequency even if the policy
is already boosted when a CPU is going online.

Fix this by using policy->boost_enabled to determine which value should
be set.

Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://patch.msgid.link/20250117101457.1530653-4-zhenglifeng1@huawei.com
[ rjw: Changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>