When boost has been disabled the limit for perf should be nominal perf not
the highest perf. Using the latter to do calculations will lead to
incorrect values that are still above nominal.
Fixes: ad4caad58d91 ("cpufreq: amd-pstate: Merge amd_pstate_highest_perf_set() into amd_get_boost_ratio_numerator()")
Reported-by: Peter Jung <ptr1337@cachyos.org>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219348
Reviewed-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Tested-by: Dhananjay Ugwekar <dhananjay.ugwekar@amd.com>
Link: https://lore.kernel.org/r/20241012174519.897-1-mario.limonciello@amd.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Rather than hooking up the PM domains through _opp_attach_genpd() and
manually manage runtime PM for the corresponding virtual devices created by
genpd during attach, let's avoid the boilerplate-code by converting into
dev_pm_domain_attach|detach_list.
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20241002122232.194245-11-ulf.hansson@linaro.org
With the Silicon revision being taken directly from socinfo, there's no
longer any need for reading any SOC register for revision from this driver.
Hence, we do not require any rev_offset for AM62 family of devices.
Signed-off-by: Dhruva Gole <d-gole@ti.com>
Reviewed-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
The AM625 syscon for efuse was being taken earlier from the wkup_conf node
where the entire wkup_conf was marked as "syscon". This is wrong and will
be fixed in the devicetree. However, whenever that does happen will end up
breaking this driver for that device because of the change in efuse offset.
Hence, to avoid breaking any sort of backward compatibility of devicetrees
use a quirk to distinguish and accordingly use 0x0 offset for the new
syscon node.
Suggested-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Dhruva Gole <d-gole@ti.com>
Reviewed-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
While switching the driver mode between active and passive, Collaborative
Processor Performance Control (CPPC) is disabled in
amd_pstate_unregister_driver(). But, it is not enabled back while registering
the new driver (passive or active). This leads to the new driver mode not
working correctly, so enable it back in amd_pstate_register_driver().
Fixes: 3ca7bc818d8c ("cpufreq: amd-pstate: Add guided mode control support via sysfs")
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20241004122303.94283-1-Dhananjay.Ugwekar@amd.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
- Fix CPU device node reference counting in the cpufreq core (Miquel
Sabaté Solà).
- Turn the spinlock used by the intel_pstate driver in hard IRQ
context into a raw one to prevent the driver from crashing when
PREEMPT_RT is enabled (Uwe Kleine-König).
-----BEGIN PGP SIGNATURE-----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=mgVf
-----END PGP SIGNATURE-----
Merge tag 'pm-6.12-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull power management fixes from Rafael Wysocki:
"These fix two cpufreq issues, one in the core and one in the
intel_pstate driver:
- Fix CPU device node reference counting in the cpufreq core (Miquel
Sabaté Solà)
- Turn the spinlock used by the intel_pstate driver in hard IRQ
context into a raw one to prevent the driver from crashing when
PREEMPT_RT is enabled (Uwe Kleine-König)"
* tag 'pm-6.12-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
cpufreq: Avoid a bad reference count on CPU node
cpufreq: intel_pstate: Make hwp_notify_lock a raw spinlock
Since commit 6c8d750f9784 ("cpufreq / cppc: Work around for Hisilicon CPPC
cpufreq"), we introduce a workround for HiSilicon platforms that do not
support performance feedback counters, whereas they can get the actual
frequency from the desired perf register. Later on, FIE is disabled in
that workaround as well.
Now the workround can be handled by the common code. Desired perf would be
read and converted to frequency if feedback counters don't change. FIE
would be disabled if the CPPC regs are in PCC region.
Hence, the workaround is no longer needed and can be safely removed, in an
effort to consolidate the driver procedure.
Signed-off-by: Jie Zhan <zhanjie9@hisilicon.com>
Reviewed-by: Xiongfeng Wang <wangxiongfeng2@huawei.com>
Reviewed-by: Huisong Li <lihuisong@huawei.com>
[ Viresh: Move fie_disabled withing CONFIG option to fix warning ]
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
The CPPC performance feedback counters could be 0 or unchanged when the
target cpu is in a low-power idle state, e.g. power-gated or clock-gated.
When the counters are 0, cppc_cpufreq_get_rate() returns 0 KHz, which makes
cpufreq_online() get a false error and fail to generate a cpufreq policy.
When the counters are unchanged, the existing cppc_perf_from_fbctrs()
returns a cached desired perf, but some platforms may update the real
frequency back to the desired perf reg.
For the above cases in cppc_cpufreq_get_rate(), get the latest desired perf
from the CPPC reg to reflect the frequency because some platforms may
update the actual frequency back there; if failed, use the cached desired
perf.
Fixes: 6a4fec4f6d30 ("cpufreq: cppc: cppc_cpufreq_get_rate() returns zero in all error cases.")
Signed-off-by: Jie Zhan <zhanjie9@hisilicon.com>
Reviewed-by: Zeng Heng <zengheng4@huawei.com>
Reviewed-by: Ionela Voinescu <ionela.voinescu@arm.com>
Reviewed-by: Huisong Li <lihuisong@huawei.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
asm/unaligned.h is always an include of asm-generic/unaligned.h;
might as well move that thing to linux/unaligned.h and include
that - there's nothing arch-specific in that header.
auto-generated by the following:
for i in `git grep -l -w asm/unaligned.h`; do
sed -i -e "s/asm\/unaligned.h/linux\/unaligned.h/" $i
done
for i in `git grep -l -w asm-generic/unaligned.h`; do
sed -i -e "s/asm-generic\/unaligned.h/linux\/unaligned.h/" $i
done
git mv include/asm-generic/unaligned.h include/linux/unaligned.h
git mv tools/include/asm-generic/unaligned.h tools/include/linux/unaligned.h
sed -i -e "/unaligned.h/d" include/asm-generic/Kbuild
sed -i -e "s/__ASM_GENERIC/__LINUX/" include/linux/unaligned.h tools/include/linux/unaligned.h
notify_hwp_interrupt() is called via sysvec_thermal() ->
smp_thermal_vector() -> intel_thermal_interrupt() in hard irq context.
For this reason it must not use a simple spin_lock that sleeps with
PREEMPT_RT enabled. So convert it to a raw spinlock.
Reported-by: xiao sheng wen <atzlinux@sina.com>
Link: https://bugs.debian.org/1076483
Signed-off-by: Uwe Kleine-König <ukleinek@debian.org>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: xiao sheng wen <atzlinux@sina.com>
Link: https://patch.msgid.link/20240919081121.10784-2-ukleinek@debian.org
Cc: All applicable <stable@vger.kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
- Implement the SCHED_DEADLINE server infrastructure - Daniel Bristot de Oliveira's
last major contribution to the kernel:
"SCHED_DEADLINE servers can help fixing starvation issues of low priority
tasks (e.g., SCHED_OTHER) when higher priority tasks monopolize CPU
cycles. Today we have RT Throttling; DEADLINE servers should be able to
replace and improve that."
(Daniel Bristot de Oliveira, Peter Zijlstra, Joel Fernandes,
Youssef Esmat, Huang Shijie)
- Preparatory changes for sched_ext integration:
- Use set_next_task(.first) where required
- Fix up set_next_task() implementations
- Clean up DL server vs. core sched
- Split up put_prev_task_balance()
- Rework pick_next_task()
- Combine the last put_prev_task() and the first set_next_task()
- Rework dl_server
- Add put_prev_task(.next)
(Peter Zijlstra, with a fix by Tejun Heo)
- Complete the EEVDF transition and refine EEVDF scheduling:
- Implement delayed dequeue
- Allow shorter slices to wakeup-preempt
- Use sched_attr::sched_runtime to set request/slice suggestion
- Document the new feature flags
- Remove unused and duplicate-functionality fields
- Simplify & unify pick_next_task_fair()
- Misc debuggability enhancements
(Peter Zijlstra, with fixes/cleanups by Dietmar Eggemann,
Valentin Schneider and Chuyi Zhou)
- Initialize the vruntime of a new task when it is first enqueued,
resulting in significant decrease in latency of newly woken tasks.
(Zhang Qiao)
- Introduce SM_IDLE and an idle re-entry fast-path in __schedule()
(K Prateek Nayak, Peter Zijlstra)
- Clean up and clarify the usage of Clean up usage of rt_task()
(Qais Yousef)
- Preempt SCHED_IDLE entities in strict cgroup hierarchies
(Tianchen Ding)
- Clarify the documentation of time units for deadline scheduler
parameters. (Christian Loehle)
- Remove the HZ_BW chicken-bit feature flag introduced a year ago,
the original change seems to be working fine.
(Phil Auld)
- Misc fixes and cleanups (Chen Yu, Dan Carpenter, Huang Shijie,
Peilin He, Qais Yousefm and Vincent Guittot)
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-----BEGIN PGP SIGNATURE-----
iQJFBAABCgAvFiEEBpT5eoXrXCwVQwEKEnMQ0APhK1gFAmbr8qcRHG1pbmdvQGtl
cm5lbC5vcmcACgkQEnMQ0APhK1gdbw/+Mj3zWfYP+dtUkfgrR2FClPAJoo1/9Dz0
LYD8XgYHu8rEJ0Aq+VbdkgYGUt9utvzUFPIxvWFDcldQl57KwhF4hp9Ir+PqJyYC
NolQ1q8ddo1hnslxnEg6SgHVzQq/4FqMM0nDNUkQETCx6zTyFFeRf+q7o/2c2m5B
uI9dSU1Wrx7XrXm2D3kB8+xP+ZRy+qhbFN5Pfuz96mhelfklylgKMfPzgAiCT/7T
JTbQhQ2HdcCNgiLoSrWsHBDy2UYpouP4zb4jyd+lDQzhSUJrj3u4Xy4vVmuTKq+y
sTgWlgKB+MTuh9UuJ4UYzSnMqg161UlMvtXeH84ABmAqDNGHRPtOKrrlcLtJ3D4x
m1SPhNnsvpjOu2pH0XLIS8al3VUesWND5S+rucHRYSq6Nvhivf4MTvRJlicXXurL
Mt2APnIlhGJuKBNWnmyZovVdtO0ZUUPlaZWfr3rCS4txAVo+HwWhsm3uhtTycQqN
gazsCiuGh6Jds90ZqA/BvdLWG+DY8J0xLlV3ex4pCXuQ/HFrabVWTyThJsULhrZ2
5mTdWIsocPctNMO9/RHMy7vJI7G7ljgHEquWVn5kiGGzXhK6VwVwKAMpfgXGw+YA
yVP6/M7a7g2yEzj69gXkcDa8k/kedMVquJ/G/8YhZM7u7sPqsMjpmaGsqsJRfnpT
ChngAzap+kA=
=TEC6
-----END PGP SIGNATURE-----
Merge tag 'sched-core-2024-09-19' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler updates from Ingo Molnar:
- Implement the SCHED_DEADLINE server infrastructure - Daniel Bristot
de Oliveira's last major contribution to the kernel:
"SCHED_DEADLINE servers can help fixing starvation issues of low
priority tasks (e.g., SCHED_OTHER) when higher priority tasks
monopolize CPU cycles. Today we have RT Throttling; DEADLINE
servers should be able to replace and improve that."
(Daniel Bristot de Oliveira, Peter Zijlstra, Joel Fernandes, Youssef
Esmat, Huang Shijie)
- Preparatory changes for sched_ext integration:
- Use set_next_task(.first) where required
- Fix up set_next_task() implementations
- Clean up DL server vs. core sched
- Split up put_prev_task_balance()
- Rework pick_next_task()
- Combine the last put_prev_task() and the first set_next_task()
- Rework dl_server
- Add put_prev_task(.next)
(Peter Zijlstra, with a fix by Tejun Heo)
- Complete the EEVDF transition and refine EEVDF scheduling:
- Implement delayed dequeue
- Allow shorter slices to wakeup-preempt
- Use sched_attr::sched_runtime to set request/slice suggestion
- Document the new feature flags
- Remove unused and duplicate-functionality fields
- Simplify & unify pick_next_task_fair()
- Misc debuggability enhancements
(Peter Zijlstra, with fixes/cleanups by Dietmar Eggemann, Valentin
Schneider and Chuyi Zhou)
- Initialize the vruntime of a new task when it is first enqueued,
resulting in significant decrease in latency of newly woken tasks
(Zhang Qiao)
- Introduce SM_IDLE and an idle re-entry fast-path in __schedule()
(K Prateek Nayak, Peter Zijlstra)
- Clean up and clarify the usage of Clean up usage of rt_task()
(Qais Yousef)
- Preempt SCHED_IDLE entities in strict cgroup hierarchies
(Tianchen Ding)
- Clarify the documentation of time units for deadline scheduler
parameters (Christian Loehle)
- Remove the HZ_BW chicken-bit feature flag introduced a year ago,
the original change seems to be working fine (Phil Auld)
- Misc fixes and cleanups (Chen Yu, Dan Carpenter, Huang Shijie,
Peilin He, Qais Yousefm and Vincent Guittot)
* tag 'sched-core-2024-09-19' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (64 commits)
sched/cpufreq: Use NSEC_PER_MSEC for deadline task
cpufreq/cppc: Use NSEC_PER_MSEC for deadline task
sched/deadline: Clarify nanoseconds in uapi
sched/deadline: Convert schedtool example to chrt
sched/debug: Fix the runnable tasks output
sched: Fix sched_delayed vs sched_core
kernel/sched: Fix util_est accounting for DELAY_DEQUEUE
kthread: Fix task state in kthread worker if being frozen
sched/pelt: Use rq_clock_task() for hw_pressure
sched/fair: Move effective_cpu_util() and effective_cpu_util() in fair.c
sched/core: Introduce SM_IDLE and an idle re-entry fast-path in __schedule()
sched: Add put_prev_task(.next)
sched: Rework dl_server
sched: Combine the last put_prev_task() and the first set_next_task()
sched: Rework pick_next_task()
sched: Split up put_prev_task_balance()
sched: Clean up DL server vs core sched
sched: Fixup set_next_task() implementations
sched: Use set_next_task(.first) where required
sched/fair: Properly deactivate sched_delayed task upon class change
...
* Move the calculation of the AMD boost numerator outside of
amd-pstate, correcting acpi-cpufreq on systems with preferred cores
* Harden preferred core detection to avoid potential false positives
* Add extra unit test coverage for mode state machine
-----BEGIN PGP SIGNATURE-----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=R4HZ
-----END PGP SIGNATURE-----
Merge tag 'amd-pstate-v6.12-2024-09-11' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/superm1/linux
Merge the second round of amd-pstate changes for 6.12 from Mario
Limonciello:
"* Move the calculation of the AMD boost numerator outside of
amd-pstate, correcting acpi-cpufreq on systems with preferred cores
* Harden preferred core detection to avoid potential false positives
* Add extra unit test coverage for mode state machine"
* tag 'amd-pstate-v6.12-2024-09-11' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/superm1/linux:
cpufreq/amd-pstate-ut: Fix an "Uninitialized variables" issue
cpufreq/amd-pstate-ut: Add test case for mode switches
cpufreq/amd-pstate: Export symbols for changing modes
amd-pstate: Add missing documentation for `amd_pstate_prefcore_ranking`
cpufreq: amd-pstate: Add documentation for `amd_pstate_hw_prefcore`
cpufreq: amd-pstate: Optimize amd_pstate_update_limits()
cpufreq: amd-pstate: Merge amd_pstate_highest_perf_set() into amd_get_boost_ratio_numerator()
x86/amd: Detect preferred cores in amd_get_boost_ratio_numerator()
x86/amd: Move amd_get_highest_perf() out of amd-pstate
ACPI: CPPC: Adjust debug messages in amd_set_max_freq_ratio() to warn
ACPI: CPPC: Drop check for non zero perf ratio
x86/amd: Rename amd_get_highest_perf() to amd_get_boost_ratio_numerator()
ACPI: CPPC: Adjust return code for inline functions in !CONFIG_ACPI_CPPC_LIB
x86/amd: Move amd_get_highest_perf() from amd.c to cppc.c
Using uninitialized value "mode2" when calling "amd_pstate_get_mode_string".
Set "mode2" to "AMD_PSTATE_DISABLE" by default.
Signed-off-by: Qianqiang Liu <qianqiang.liu@163.com>
Link: https://lore.kernel.org/r/20240910233923.46470-1-qianqiang.liu@163.com
Acked-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
There is a state machine in the amd-pstate driver utilized for
switches for all modes. To make sure that cleanup and setup works
properly for each mode add a unit test case that tries all
combinations.
Reviewed-by: Perry Yuan <perry.yuan@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
In order to effectively test all mode switch combinations export
everything necessarily for amd-pstate-ut to trigger a mode switch.
Reviewed-by: Perry Yuan <Perry.Yuan@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Don't take and release the mutex when prefcore isn't present and
avoid initialization of variables that will be initially set
in the function.
Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Reviewed-by: Perry Yuan <perry.yuan@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
The special case in amd_pstate_highest_perf_set() is the value used
for calculating the boost numerator. Merge this into
amd_get_boost_ratio_numerator() and then use that to calculate boost
ratio.
This allows dropping more special casing of the highest perf value.
Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
AMD systems that support preferred cores will use "166" as their
numerator for max frequency calculations instead of "255".
Add a function for detecting preferred cores by looking at the
highest perf value on all cores.
If preferred cores are enabled return 166 and if disabled the
value in the highest perf register. As the function will be called
multiple times, cache the values for the boost numerator and if
preferred cores will be enabled in global variables.
Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
amd_pstate_get_highest_perf() is a helper used to get the highest perf
value on AMD systems. It's used in amd-pstate as part of preferred
core handling, but applicable for acpi-cpufreq as well.
Move it out to cppc handling code as amd_get_highest_perf().
Reviewed-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
The function name is ambiguous because it returns an intermediate value
for calculating maximum frequency rather than the CPPC 'Highest Perf'
register.
Rename the function to clarify its use and allow the function to return
errors. Adjust the consumer in acpi-cpufreq to catch errors.
Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Convert the cppc deadline task attributes to use the available
definitions to make them more readable.
No functional change.
Signed-off-by: Christian Loehle <christian.loehle@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Juri Lelli <juri.lelli@redhat.com>
Acked-by: Rafael J. Wysocki <rafael@kernel.org>
Link: https://lore.kernel.org/r/20240813144348.1180344-4-christian.loehle@arm.com
- Several OF related cleanups in cpufreq drivers (Rob Herring).
- Enable COMPILE_TEST for ARM drivers (Rob Herrring).
- Introduce quirks for syscon failures and use socinfo to get revision
for TI cpufreq driver (Dhruva Gole and Nishanth Menon).
- Minor cleanups in amd-pstate driver (Anastasia Belova and Dhananjay
Ugwekar).
- Minor cleanups for loongson, cpufreq-dt and powernv cpufreq drivers
(Danila Tikhonov, Huacai Chen, and Liu Jing).
-----BEGIN PGP SIGNATURE-----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=BlNP
-----END PGP SIGNATURE-----
Merge tag 'cpufreq-arm-updates-6.12' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/vireshk/pm
Merge ARM cpufreq updates for 6.12 from Viresh Kumar:
"- Several OF related cleanups in cpufreq drivers (Rob Herring).
- Enable COMPILE_TEST for ARM drivers (Rob Herrring).
- Introduce quirks for syscon failures and use socinfo to get revision
for TI cpufreq driver (Dhruva Gole and Nishanth Menon).
- Minor cleanups in amd-pstate driver (Anastasia Belova and Dhananjay
Ugwekar).
- Minor cleanups for loongson, cpufreq-dt and powernv cpufreq drivers
(Danila Tikhonov, Huacai Chen, and Liu Jing)."
* tag 'cpufreq-arm-updates-6.12' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/vireshk/pm:
cpufreq: ti-cpufreq: Use socinfo to get revision in AM62 family
cpufreq: Fix the cacography in powernv-cpufreq.c
cpufreq: ti-cpufreq: Introduce quirks to handle syscon fails appropriately
cpufreq: loongson3: Use raw_smp_processor_id() in do_service_request()
cpufreq: amd-pstate: add check for cpufreq_cpu_get's return value
cpufreq: Add SM7325 to cpufreq-dt-platdev blocklist
cpufreq: Fix warning on unused of_device_id tables for !CONFIG_OF
cpufreq/amd-pstate: Add the missing cpufreq_cpu_put()
cpufreq: Drop CONFIG_ARM and CONFIG_ARM64 dependency on Arm drivers
cpufreq: Enable COMPILE_TEST on Arm drivers
cpufreq: armada-8k: Avoid excessive stack usage
cpufreq: omap: Drop asm includes
cpufreq: qcom: Add explicit io.h include for readl/writel_relaxed
cpufreq: spear: Use of_property_for_each_u32() instead of open coding
cpufreq: Use of_property_present()
* Validate return of any attempt to update EPP limits, which fixes
the masking hardware problems.
-----BEGIN PGP SIGNATURE-----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=nPcu
-----END PGP SIGNATURE-----
Merge tag 'amd-pstate-v6.12-2024-09-04' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/superm1/linux
Merge an amd-pstate driver update for 6.12 from Mario Limonciello:
"amd-pstate development for 6.12:
* Validate return of any attempt to update EPP limits, which fixes
the masking hardware problems."
* tag 'amd-pstate-v6.12-2024-09-04' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/superm1/linux:
cpufreq/amd-pstate: Catch failures for amd_pstate_epp_update_limit()
amd_pstate_set_epp() calls cppc_set_epp_perf() which can fail for
a variety of reasons but this is ignored. Change the return flow
to allow failures.
Reviewed-by: Perry Yuan <perry.yuan@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
In the AM62x, AM62Ax, and AM62Px devices, we already have the revision
info within the k3-socinfo driver. Hence, re-use this information from
there instead of re using the offset for 2 drivers trying to get the
same information ie. revision.
Signed-off-by: Dhruva Gole <d-gole@ti.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Commit b4bc9f9e27ed ("cpufreq: ti-cpufreq: add support for omap34xx
and omap36xx") introduced special handling for OMAP3 class devices
where syscon node may not be present. However, this also creates a bug
where the syscon node is present, however the offset used to read
is beyond the syscon defined range.
Fix this by providing a quirk option that is populated when such
special handling is required. This allows proper failure for all other
platforms when the syscon node and efuse offsets are mismatched.
Fixes: b4bc9f9e27ed ("cpufreq: ti-cpufreq: add support for omap34xx and omap36xx")
Signed-off-by: Nishanth Menon <nm@ti.com>
Tested-by: Dhruva Gole <d-gole@ti.com>
Reviewed-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Use raw_smp_processor_id() instead of plain smp_processor_id() in
do_service_request(), otherwise we may get some errors with the driver
enabled:
BUG: using smp_processor_id() in preemptible [00000000] code: (udev-worker)/208
caller is loongson3_cpufreq_probe+0x5c/0x250 [loongson3_cpufreq]
Reported-by: Xi Ruoyao <xry111@xry111.site>
Tested-by: Binbin Zhou <zhoubinbin@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
cpufreq_cpu_get may return NULL. To avoid NULL-dereference check it
and return in case of error.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Signed-off-by: Anastasia Belova <abelova@astralinux.ru>
Reviewed-by: Perry Yuan <perry.yuan@amd.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
The Qualcomm SM7325 platform uses the qcom-cpufreq-hw driver, so add
it to the cpufreq-dt-platdev driver's blocklist.
Signed-off-by: Danila Tikhonov <danila@jiaxyga.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
!CONFIG_OF builds cause warnings on unused of_device_id tables. This is
due to of_match_node() being a macro rather than static inline function.
Add a __maybe_unused annotation to the of_device_id tables.
Fixes: c7582ec85342 ("cpufreq: Drop CONFIG_ARM and CONFIG_ARM64 dependency on Arm drivers")
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202408090714.wcrqU6Pk-lkp@intel.com/
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Fix the reference counting of cpufreq_policy object in amd_pstate_update()
function by adding the missing cpufreq_cpu_put().
Fixes: e8f555daacd3 ("cpufreq/amd-pstate: fix setting policy current frequency value")
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Reviewed-by: Perry Yuan <perry.yuan@amd.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
The CONFIG_ARM and CONFIG_ARM64 dependency is redundant as all the
drivers have necessary sub-arch dependency and don't depend on the
architecture support.
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
COMPILE_TEST is useful for build testing without requiring a specific
architecture's compiler. Enable it for most of the Arm CPUFreq drivers.
As Kconfig.arm is only included on ARM and ARM64, COMPILE_TEST is only
enabled for those architectures until that is dropped.
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Make intel_pstate use the HWP_HIGHEST_PERF values from
MSR_HWP_CAPABILITIES to set asymmetric CPU capacity information
via the previously introduced arch_set_cpu_capacity() on hybrid
systems without SMT.
Setting asymmetric CPU capacity is generally necessary to allow the
scheduler to compute task sizes in a consistent way across all CPUs
in a system where they differ by capacity. That, in turn, should help
to improve scheduling decisions. It is also necessary for the schedutil
cpufreq governor to operate as expected on hybrid systems where tasks
migrate between CPUs of different capacities.
The underlying observation is that intel_pstate already uses
MSR_HWP_CAPABILITIES to get CPU performance information which is
exposed by it via sysfs and CPU performance scaling is based on it.
Thus using this information for setting asymmetric CPU capacity is
consistent with what the driver has been doing already. Moreover,
HWP_HIGHEST_PERF reflects the maximum capacity of a given CPU including
both the instructions-per-cycle (IPC) factor and the maximum turbo
frequency and the units in which that value is expressed are the same
for all CPUs in the system, so the maximum capacity ratio between two
CPUs can be obtained by computing the ratio of their HWP_HIGHEST_PERF
values. Of course, in principle that capacity ratio need not be
directly applicable at lower frequencies, so using it for providing the
asymmetric CPU capacity information to the scheduler is a rough
approximation, but it is as good as it gets. Also, measurements
indicate that this approximation is not too bad in practice.
If the given system is hybrid and non-SMT, the new code disables ITMT
support in the scheduler (because it may get in the way of asymmetric CPU
capacity code in the scheduler that automatically gets enabled by setting
asymmetric CPU capacity) after initializing all online CPUs and finds
the one with the maximum HWP_HIGHEST_PERF value. Next, it computes the
capacity number for each (online) CPU by dividing the product of its
HWP_HIGHEST_PERF and SCHED_CAPACITY_SCALE by the maximum HWP_HIGHEST_PERF.
When a CPU goes offline, its capacity is reset to SCHED_CAPACITY_SCALE
and if it is the one with the maximum HWP_HIGHEST_PERF value, the
capacity numbers for all of the other online CPUs are recomputed. This
also takes care of a cleanup during driver operation mode changes.
Analogously, when a new CPU goes online, its capacity number is updated
and if its HWP_HIGHEST_PERF value is greater than the current maximum
one, the capacity numbers for all of the other online CPUs are
recomputed.
The case when the driver is notified of a CPU capacity change, either
through the HWP interrupt or through an ACPI notification, is handled
similarly to the CPU online case above, except that if the target CPU
is the current highest-capacity one and its capacity is reduced, the
capacity numbers for all of the other online CPUs need to be recomputed
either.
If the driver's "no_trubo" sysfs attribute is updated, all of the CPU
capacity information is computed from scratch to reflect the new turbo
status.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Tested-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com> # scale invariance
Link: https://patch.msgid.link/1979653.PYKUYFuaPT@rjwysocki.net
[ rjw: Fixed a typo in the changelog ]
[ rjw: Renamed 3 new functions and added a comment ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
commit bff7d13c190a ("cpufreq: amd-pstate: add debug message while
CPPC is supported and disabled by SBIOS") issues a warning on plaforms
where the X86_FEATURE_CPPC is expected to be enabled, but is not due
to it being disabled in the BIOS.
This feature bit corresponds to CPUID 0x80000008.ebx[27] which is a
reserved bit on the Zen1 processors and a reserved bit on Zen2 based
models 0x70-0x7F, and is expected to be cleared on these
platforms. Thus printing the warning message for these models when
X86_FEATURE_CPPC is unavailable is incorrect. Fix this.
Modify some of the comments, and use switch-case for model range
checking for improved readability while at it.
Fixes: bff7d13c190a ("cpufreq: amd-pstate: add debug message while CPPC is supported and disabled by SBIOS")
Cc: Xiaojian Du <xiaojian.du@amd.com>
Reported-by: David Wang <00107082@163.com>
Closes: https://lore.kernel.org/lkml/20240730140111.4491-1-00107082@163.com/
Signed-off-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Acked-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
If a system is using preferred cores the highest perf will be inconsistent
as it can change from system events.
Skip the checks for it.
Fixes: e571a5e2068e ("cpufreq: amd-pstate: Update amd-pstate preferred core ranking dynamically")
Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
After the commit 63edbaa48a57 ("x86/cpu/topology: Add support for the
AMD 0x80000026 leaf"), the topolgy_logical_die_id() function returns
the logical Core Chiplet Die (CCD) ID instead of the logical socket
ID.
Since this is currently used to set MSR_AMD_CPPC_ENABLE, which needs
to be set on any one of the threads of the socket, it is prudent to
use topology_logical_package_id() in place of
topology_logical_die_id().
Fixes: 63edbaa48a57 ("x86/cpu/topology: Add support for the AMD 0x80000026 leaf")
cc: stable@vger.kernel.org # 6.10
Signed-off-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Tested-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Link: https://lore.kernel.org/lkml/20240801124509.3650-1-Dhananjay.Ugwekar@amd.com/
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Smatch complains that "ret" could be uninitialized:
drivers/cpufreq/amd-pstate.c:734 amd_pstate_cpu_boost_update()
error: uninitialized symbol 'ret'.
This seems like it probably is a real issue. Initialize "ret" to zero to
be safe.
Fixes: c8c68c38b56f ("cpufreq: amd-pstate: initialize core precision boost state")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Perry Yuan <perry.yuan@amd.com>
Acked-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Link: https://lore.kernel.org/lkml/7ff53543-6c04-48a0-8d99-7dc010b93b3a@stanley.mountain/T/
Signed-off-by: Gautham R. Shenoy <gautham.shenoy@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Prevent intel_pstate from loading when OOB (Out Of Band) P-states mode is
enabled.
The OOB identifying bits are same as for the prior generation CPUs like
Emerald Rapids servers. Add Granite Rapids and Sierra Forest CPU models to
intel_pstate_cpu_oob_ids[].
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Link: https://patch.msgid.link/20240802184839.1909091-1-srinivas.pandruvada@linux.intel.com
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
In some build configurations (e.g. x86 allmodconfig), 2 cpu_mask
variables exceeds the max stack frame size:
drivers/cpufreq/armada-8k-cpufreq.c:203:1: error: the frame size of 2128 bytes is larger than 2048 bytes
Fix it by making "cpus" static which is fine given that module init is
only called once.
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
The omap driver doesn't actually need asm/smp_plat.h, so drop it.
asm/cpu.h is not needed either as linux/cpu.h is already included.
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Acked-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
The qcom-cpufreq-hw driver is relying on an implicit include of io.h
which doesn't get included on some architectures.
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Use of_property_count_u32_elems() and of_property_for_each_u32() instead
of of_find_property() and open coding the property parsing.
This is part of a larger effort to remove callers of of_find_property()
and similar functions. of_find_property() leaks the DT struct property
and data pointers which is a problem for dynamically allocated nodes
which may be freed.
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
[Viresh: Fixed build failure and initialize / increment 'i']
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Use of_property_present() to test for property presence rather than
of_(find|get)_property(). This is part of a larger effort to remove
callers of of_find_property() and similar functions. of_find_property()
leaks the DT struct property and data pointers which is a problem for
dynamically allocated nodes which may be freed.
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Acked-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
The current LATENCY_MULTIPLIER which has been around for nearly 20 years
causes rate_limit_us to be always in ms range.
On M1 mac mini I get 50 and 56us transition latency, but due to the 1000
multiplier we end up setting rate_limit_us to 50 and 56ms, which gets
capped into 2ms and was 10ms before e13aa799c2a6 ("cpufreq: Change
default transition delay to 2ms")
On Intel I5 system transition latency is 20us but due to the multiplier
we end up with 20ms that again is capped to 2ms.
Given how good modern hardware and how modern workloads require systems
to be more responsive to cater for sudden changes in workload (tasks
sleeping/wakeup/migrating, uclamp causing a sudden boost or cap) and
that 2ms is quarter of the time of 120Hz refresh rate system, drop the
old logic in favour of providing 50% headroom.
rate_limit_us = 1.5 * latency.
I considered not adding any headroom which could mean that we can end up
with infinite back-to-back requests.
I also considered providing a constant headroom (e.g: 100us) assuming
that any h/w or f/w dealing with the request shouldn't require a large
headroom when transition_latency is actually high.
But for both cases I wasn't sure if h/w or f/w can end up being
overwhelmed dealing with the freq requests in a potentially busy system.
So I opted for providing 50% breathing room.
This is expected to impact schedutil only as the other user,
dbs_governor, takes the max(2*tick, transition_delay_us) and the former
was at least 2ms on 1ms TICK, which is equivalent to the max_delay_us
before applying this patch. For systems with TICK of 4ms, this value
would have almost always ended up with 8ms sampling rate.
For systems that report 0 transition latency, we still default to
returning 1ms as transition delay.
This helps in eliminating a source of latency for applying requests as
mentioned in [1]. For example if we have a 1ms tick, most systems will
miss sending an update at tick when updating the util_avg for a task/CPU
(rate_limit_us will be 2ms for most systems).
Link: https://lore.kernel.org/lkml/20240724212255.mfr2ybiv2j2uqek7@airbuntu/ # [1]
Link: https://lore.kernel.org/lkml/20240205022500.2232124-1-qyousef@layalina.io/
Signed-off-by: Qais Yousef <qyousef@layalina.io>
Link: https://patch.msgid.link/20240728192659.58115-1-qyousef@layalina.io
[ rjw: Subject edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
With ARCH=powerpc, make allmodconfig && make W=1 C=1 reports:
WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/cpufreq/ppc-cbe-cpufreq.o
WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/cpufreq/powernv-cpufreq.o
Add the missing invocation of the MODULE_DESCRIPTION() macro to all
files which have a MODULE_LICENSE().
This includes three additional files which, although they did not
produce a warning with the powerpc allmodconfig configuration, may
cause this warning with specific options enabled in the kernel
configuration.
Signed-off-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Acked-by: Michael Ellerman <mpe@ellerman.id.au> (powerpc)
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Link: https://patch.msgid.link/20240722-md-powerpc-drivers-cpufreq-v2-1-bb84d715eb3d@quicinc.com
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
On Intel Emerald Rapids machines, we ship the Energy Performance Preference
(EPP) default for balance_performance as 128. However, during an internal
investigation together with Intel, we have determined that 32 is a more
suitable value. This leads to significant improvements in both performance
and energy:
POV-Ray: 32% faster | 12% less energy
OpenSSL: 12% faster | energy within 1%
Build Linux Kernel: 29% faster | 18% less energy
Therefore, we should move the default EPP for balance_performance to 32.
This is in line with what has already been done for Sapphire Rapids.
Signed-off-by: Pedro Henrique Kopper <pedro.kopper@canonical.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Link: https://patch.msgid.link/Zqu6zjVMoiXwROBI@capivara
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
