[AMD Official Use Only - General]
Hi Oleksandr:
-----Original Message----- From: Oleksandr Natalenko oleksandr@natalenko.name Sent: Monday, October 9, 2023 9:00 PM To: Rafael J . Wysocki rafael.j.wysocki@intel.com; Huang, Ray Ray.Huang@amd.com; Meng, Li (Jassmine) Li.Meng@amd.com Cc: linux-pm@vger.kernel.org; linux-kernel@vger.kernel.org; x86@kernel.org; linux-acpi@vger.kernel.org; Shuah Khan skhan@linuxfoundation.org; linux-kselftest@vger.kernel.org; Fontenot, Nathan Nathan.Fontenot@amd.com; Sharma, Deepak Deepak.Sharma@amd.com; Deucher, Alexander Alexander.Deucher@amd.com; Limonciello, Mario Mario.Limonciello@amd.com; Huang, Shimmer Shimmer.Huang@amd.com; Yuan, Perry Perry.Yuan@amd.com; Du, Xiaojian Xiaojian.Du@amd.com; Viresh Kumar viresh.kumar@linaro.org; Borislav Petkov bp@alien8.de Subject: Re: [PATCH V8 0/7] amd-pstate preferred core
Hello.
On pondělí 9. října 2023 9:23:29 CEST Meng, Li (Jassmine) wrote:
[AMD Official Use Only - General]
Hi Oleksandr:
-----Original Message----- From: Oleksandr Natalenko oleksandr@natalenko.name Sent: Monday, October 9, 2023 2:55 PM To: Rafael J . Wysocki rafael.j.wysocki@intel.com; Huang, Ray Ray.Huang@amd.com; Meng, Li (Jassmine) Li.Meng@amd.com Cc: linux-pm@vger.kernel.org; linux-kernel@vger.kernel.org; x86@kernel.org; linux-acpi@vger.kernel.org; Shuah Khan skhan@linuxfoundation.org; linux-kselftest@vger.kernel.org; Fontenot, Nathan Nathan.Fontenot@amd.com; Sharma, Deepak Deepak.Sharma@amd.com; Deucher, Alexander Alexander.Deucher@amd.com; Limonciello, Mario Mario.Limonciello@amd.com; Huang, Shimmer
Yuan, Perry Perry.Yuan@amd.com; Du, Xiaojian Xiaojian.Du@amd.com; Viresh Kumar viresh.kumar@linaro.org; Borislav Petkov bp@alien8.de; Meng, Li (Jassmine) Li.Meng@amd.com Subject: Re: [PATCH V8 0/7] amd-pstate preferred core
Hello.
On pondělí 9. října 2023 4:49:25 CEST Meng Li wrote:
Hi all:
The core frequency is subjected to the process variation in
semiconductors.
Not all cores are able to reach the maximum frequency respecting the infrastructure limits. Consequently, AMD has redefined the concept of maximum frequency of a part. This means that a fraction of cores can reach maximum frequency. To find the best process scheduling policy for a given scenario, OS needs to know the core ordering informed by the platform through highest performance capability register of the CPPC
interface.
Earlier implementations of amd-pstate preferred core only support a static core ranking and targeted performance. Now it has the ability to dynamically change the preferred core based on the workload and platform conditions and accounting for thermals and
aging.
Amd-pstate driver utilizes the functions and data structures provided by the ITMT architecture to enable the scheduler to favor scheduling on cores which can be get a higher frequency with lower
voltage.
We call it amd-pstate preferred core.
Here sched_set_itmt_core_prio() is called to set priorities and sched_set_itmt_support() is called to enable ITMT feature. Amd-pstate driver uses the highest performance value to indicate the priority of CPU. The higher value has a higher priority.
Amd-pstate driver will provide an initial core ordering at boot time. It relies on the CPPC interface to communicate the core ranking to the operating system and scheduler to make sure that OS is choosing the cores with highest performance firstly for scheduling the
process.
When amd-pstate driver receives a message with the highest performance change, it will update the core ranking.
Changes form V7->V8:
- all:
- pick up Review-By flag added by Mario and Ray.
- cpufreq: amd-pstate:
- use hw_prefcore embeds into cpudata structure.
- delete preferred core init from cpu online/off.
Could you please let me know if this change means a fix for the report I've sent previously? [1]
[Meng, Li (Jassmine)] Yes. I have deleted online handle function of amd pstate driver. It doesn't re-initialize preferred core. This online function will set incorrect des perf value.
Thank you for the confirmation. I've built v6.5.5 with this patchset applied, and now the frequency is as expected after the suspend-resume cycle.
I've also added the following modification to accommodate recent feedback:
commit 1450ac395434c532f995521e1a2497d09ddf106c Author: Oleksandr Natalenko <oleksandr@natalenko.name> Date: Mon Oct 9 11:19:50 2023 +0200 cpufreq/amd-pstate: show prefcore_ranking separately Signed-off-by: Oleksandr Natalenko <oleksandr@natalenko.name> diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c index d3369247c6c9c..86999d861e87b 100644 --- a/drivers/cpufreq/amd-pstate.c +++ b/drivers/cpufreq/amd-pstate.c @@ -954,6 +954,17 @@ static ssize_t show_amd_pstate_highest_perf(struct cpufreq_policy *policy, u32 perf; struct amd_cpudata *cpudata = policy->driver_data; + perf = READ_ONCE(cpudata->highest_perf); + + return sysfs_emit(buf, "%u\n", perf); +} + +static ssize_t show_amd_pstate_prefcore_ranking(struct cpufreq_policy *policy, + char *buf) +{ + u32 perf; + struct amd_cpudata *cpudata = policy->driver_data; + perf = READ_ONCE(cpudata->prefcore_ranking); return sysfs_emit(buf, "%u\n", perf); @@ -1172,6 +1183,7 @@ cpufreq_freq_attr_ro(amd_pstate_max_freq); cpufreq_freq_attr_ro(amd_pstate_lowest_nonlinear_freq); cpufreq_freq_attr_ro(amd_pstate_highest_perf); +cpufreq_freq_attr_ro(amd_pstate_prefcore_ranking); cpufreq_freq_attr_ro(amd_pstate_hw_prefcore); cpufreq_freq_attr_rw(energy_performance_preference); cpufreq_freq_attr_ro(energy_performance_available_preferences); @@ -1182,6 +1194,7 @@ static struct freq_attr *amd_pstate_attr[] = { &amd_pstate_max_freq, &amd_pstate_lowest_nonlinear_freq, &amd_pstate_highest_perf, + &amd_pstate_prefcore_ranking, &amd_pstate_hw_prefcore, NULL, }; @@ -1190,6 +1203,7 @@ static struct freq_attr *amd_pstate_epp_attr[] = { &amd_pstate_max_freq, &amd_pstate_lowest_nonlinear_freq, &amd_pstate_highest_perf, + &amd_pstate_prefcore_ranking, &amd_pstate_hw_prefcore, &energy_performance_preference, &energy_performance_available_preferences,with the following output as a result:
[~]> grep . /sys/devices/system/cpu*/cpufreq/policy*/amd_pstate_highest_perf /sys/devices/system/cpu/cpufreq/policy0/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy1/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy2/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy3/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy4/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy5/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy6/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy7/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy8/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy9/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy10/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy11/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy12/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy13/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy14/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy15/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy16/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy17/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy18/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy19/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy20/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy21/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy22/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy23/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy24/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy25/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy26/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy27/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy28/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy29/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy30/amd_pstate_highest_perf:166 /sys/devices/system/cpu/cpufreq/policy31/amd_pstate_highest_perf:166 [~]> grep . /sys/devices/system/cpu*/cpufreq/policy*/amd_pstate_hw_prefcore /sys/devices/system/cpu/cpufreq/policy0/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy1/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy2/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy3/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy4/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy5/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy6/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy7/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy8/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy9/amd_pstate_hw_prefcore:suppo rted /sys/devices/system/cpu/cpufreq/policy10/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy11/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy12/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy13/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy14/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy15/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy16/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy17/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy18/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy19/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy20/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy21/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy22/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy23/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy24/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy25/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy26/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy27/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy28/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy29/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy30/amd_pstate_hw_prefcore:supp orted /sys/devices/system/cpu/cpufreq/policy31/amd_pstate_hw_prefcore:supp orted [~]> grep . /sys/devices/system/cpu*/cpufreq/policy*/amd_pstate_prefcore_ranking /sys/devices/system/cpu/cpufreq/policy0/amd_pstate_prefcore_ranking:22 6 /sys/devices/system/cpu/cpufreq/policy1/amd_pstate_prefcore_ranking:23 1 /sys/devices/system/cpu/cpufreq/policy2/amd_pstate_prefcore_ranking:21 1 /sys/devices/system/cpu/cpufreq/policy3/amd_pstate_prefcore_ranking:23 6 /sys/devices/system/cpu/cpufreq/policy4/amd_pstate_prefcore_ranking:21 6 /sys/devices/system/cpu/cpufreq/policy5/amd_pstate_prefcore_ranking:23 6 /sys/devices/system/cpu/cpufreq/policy6/amd_pstate_prefcore_ranking:20 6 /sys/devices/system/cpu/cpufreq/policy7/amd_pstate_prefcore_ranking:22 1 /sys/devices/system/cpu/cpufreq/policy8/amd_pstate_prefcore_ranking:19 1 /sys/devices/system/cpu/cpufreq/policy9/amd_pstate_prefcore_ranking:20 1 /sys/devices/system/cpu/cpufreq/policy10/amd_pstate_prefcore_ranking:1 86 /sys/devices/system/cpu/cpufreq/policy11/amd_pstate_prefcore_ranking:1 96 /sys/devices/system/cpu/cpufreq/policy12/amd_pstate_prefcore_ranking:1 71 /sys/devices/system/cpu/cpufreq/policy13/amd_pstate_prefcore_ranking:1 66 /sys/devices/system/cpu/cpufreq/policy14/amd_pstate_prefcore_ranking:1 76 /sys/devices/system/cpu/cpufreq/policy15/amd_pstate_prefcore_ranking:1 81 /sys/devices/system/cpu/cpufreq/policy16/amd_pstate_prefcore_ranking:2 26 /sys/devices/system/cpu/cpufreq/policy17/amd_pstate_prefcore_ranking:2 31 /sys/devices/system/cpu/cpufreq/policy18/amd_pstate_prefcore_ranking:2 11 /sys/devices/system/cpu/cpufreq/policy19/amd_pstate_prefcore_ranking:2 36 /sys/devices/system/cpu/cpufreq/policy20/amd_pstate_prefcore_ranking:2 16 /sys/devices/system/cpu/cpufreq/policy21/amd_pstate_prefcore_ranking:2 36 /sys/devices/system/cpu/cpufreq/policy22/amd_pstate_prefcore_ranking:2 06 /sys/devices/system/cpu/cpufreq/policy23/amd_pstate_prefcore_ranking:2 21 /sys/devices/system/cpu/cpufreq/policy24/amd_pstate_prefcore_ranking:1 91 /sys/devices/system/cpu/cpufreq/policy25/amd_pstate_prefcore_ranking:2 01 /sys/devices/system/cpu/cpufreq/policy26/amd_pstate_prefcore_ranking:1 86 /sys/devices/system/cpu/cpufreq/policy27/amd_pstate_prefcore_ranking:1 96 /sys/devices/system/cpu/cpufreq/policy28/amd_pstate_prefcore_ranking:1 71 /sys/devices/system/cpu/cpufreq/policy29/amd_pstate_prefcore_ranking:1 66 /sys/devices/system/cpu/cpufreq/policy30/amd_pstate_prefcore_ranking:1 76 /sys/devices/system/cpu/cpufreq/policy31/amd_pstate_prefcore_ranking:1 81When I run `dd if=/dev/zero of=/dev/null`, the load lands onto cores 3, 5, 19 or 21, IOW, those that have the highest `amd_pstate_prefcore_ranking` value given `schedutil` is in use.
If all of the above is as expected, please add:
Tested-by: Oleksandr Natalenko oleksandr@natalenko.name
Would you also be able to Cc me on the next iteration of this patchset?
[Meng, Li (Jassmine)] OK.
Thanks.
[Meng, Li (Jassmine)] Thanks a lot. Based on Wyes's suggestion, I also made similar modifications in the next patches. All the log information above is in line with expectations.
Thank you!
[1] https://lore.kernel.org/lkml/5973628.lOV4Wx5bFT@natalenko.name/
Changes form V6->V7:
- x86:
- Modify kconfig about X86_AMD_PSTATE.
- cpufreq: amd-pstate:
- modify incorrect comments about scheduler_work().
- convert highest_perf data type.
- modify preferred core init when cpu init and online.
- acpi: cppc:
- modify link of CPPC highest performance.
- cpufreq:
- modify link of CPPC highest performance changed.
Changes form V5->V6:
- cpufreq: amd-pstate:
- modify the wrong tag order.
- modify warning about hw_prefcore sysfs attribute.
- delete duplicate comments.
- modify the variable name cppc_highest_perf to prefcore_ranking.
- modify judgment conditions for setting highest_perf.
- modify sysfs attribute for CPPC highest perf to pr_debug message.
- Documentation: amd-pstate:
- modify warning: title underline too short.
Changes form V4->V5:
- cpufreq: amd-pstate:
- modify sysfs attribute for CPPC highest perf.
- modify warning about comments
- rebase linux-next
- cpufreq:
- Moidfy warning about function declarations.
- Documentation: amd-pstate:
- align with ``amd-pstat``
Changes form V3->V4:
- Documentation: amd-pstate:
- Modify inappropriate descriptions.
Changes form V2->V3:
- x86:
- Modify kconfig and description.
- cpufreq: amd-pstate:
- Add Co-developed-by tag in commit message.
- cpufreq:
- Modify commit message.
- Documentation: amd-pstate:
- Modify inappropriate descriptions.
Changes form V1->V2:
- acpi: cppc:
- Add reference link.
- cpufreq:
- Moidfy link error.
- cpufreq: amd-pstate:
- Init the priorities of all online CPUs
- Use a single variable to represent the status of preferred core.
- Documentation:
- Default enabled preferred core.
- Documentation: amd-pstate:
- Modify inappropriate descriptions.
- Default enabled preferred core.
- Use a single variable to represent the status of preferred core.
Meng Li (7): x86: Drop CPU_SUP_INTEL from SCHED_MC_PRIO for the expansion. acpi: cppc: Add get the highest performance cppc control cpufreq: amd-pstate: Enable amd-pstate preferred core supporting. cpufreq: Add a notification message that the highest perf has changed cpufreq: amd-pstate: Update amd-pstate preferred core ranking dynamically Documentation: amd-pstate: introduce amd-pstate preferred core Documentation: introduce amd-pstate preferrd core mode kernel
command
line options.../admin-guide/kernel-parameters.txt | 5 + Documentation/admin-guide/pm/amd-pstate.rst | 59 +++++- arch/x86/Kconfig | 5 +- drivers/acpi/cppc_acpi.c | 13 ++ drivers/acpi/processor_driver.c | 6 + drivers/cpufreq/amd-pstate.c | 186 ++++++++++++++++-- drivers/cpufreq/cpufreq.c | 13 ++ include/acpi/cppc_acpi.h | 5 + include/linux/amd-pstate.h | 10 + include/linux/cpufreq.h | 5 + 10 files changed, 285 insertions(+), 22 deletions(-)
-- Oleksandr Natalenko (post-factum)
-- Oleksandr Natalenko (post-factum)