Linux kernel stable tree (mirror) https://git.radix-linux.su/kernel/linux.git/atom?h=v6.1 2022-11-22T18:57:15+00:00 2022-11-22T18:57:15+00:00 Perry Yuan Perry.Yuan@amd.com 2022-11-17T07:35:39+00:00 urn:sha1:202e683df37cdf4c38e06e56ac91cc170ef49058 When the amd_pstate driver is built-in users still need a method to be able enable or disable it depending upon their circumstance. Add support for an early parameter to do this. There is some performance degradation on a number of ASICs in the passive mode. This performance issue was originally discovered in shared memory systems but it has been proven that certain workloads on MSR systems also suffer performance issues. Set the amd-pstate driver as disabled by default to temporarily mitigate the performance problem. 1) with `amd_pstate=disable`, pstate driver will be disabled to load at kernel booting. 2) with `amd_pstate=passive`, pstate driver will be enabled and loaded as non-autonomous working mode supported in the low-level power management firmware. 3) If neither parameter is specified, the driver will be disabled by default to avoid triggering performance regressions in certain ASICs Acked-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com> Tested-by: Wyes Karny <wyes.karny@amd.com> Signed-off-by: Perry Yuan <Perry.Yuan@amd.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 2022-11-22T18:57:15+00:00 Perry Yuan Perry.Yuan@amd.com 2022-11-17T07:35:38+00:00 urn:sha1:456ca88d8a5258fc66edc42a10053ac8473de2b1 Currently when the amd-pstate and acpi_cpufreq are both built into kernel as module driver, amd-pstate will not be loaded by default in this case. Change amd-pstate driver as built-in type, it will resolve the loading sequence problem to allow user to make amd-pstate driver as the default cpufreq scaling driver. Acked-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com> Tested-by: Wyes Karny <wyes.karny@amd.com> Signed-off-by: Perry Yuan <Perry.Yuan@amd.com> Fixes: ec437d71db77 ("cpufreq: amd-pstate: Introduce a new AMD P-State driver to support future processors") Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 2022-11-22T18:57:14+00:00 Wyes Karny wyes.karny@amd.com 2022-11-17T07:35:37+00:00 urn:sha1:919f4557696939625085435ebde09a539de2349c MSR_AMD_PERF_CTL is guaranteed to be 0 on a cold boot. However, on a kexec boot, for instance, it may have a non-zero value (if the cpu was in a non-P0 Pstate). In such cases, the cores with non-P0 Pstates at boot will never be pushed to P0, let alone boost frequencies. Kexec is a common workflow for reboot on Linux and this creates a regression in performance. Fix it by explicitly setting the MSR_AMD_PERF_CTL to 0 during amd_pstate driver init. Cc: All applicable <stable@vger.kernel.org> Acked-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Gautham R. Shenoy <gautham.shenoy@amd.com> Tested-by: Wyes Karny <wyes.karny@amd.com> Signed-off-by: Wyes Karny <wyes.karny@amd.com> Signed-off-by: Perry Yuan <Perry.Yuan@amd.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 2022-10-06T19:53:15+00:00 Linus Torvalds torvalds@linux-foundation.org 2022-10-06T19:53:15+00:00 urn:sha1:dd42d9c3f461a7dc896076691ba42cf97225973e Pull Kselftest updates from Shuah Khan: "Fixes and new tests: - Add an amd-pstate-ut test module, used by kselftest to unit test amd-pstate functionality - Fixes and cleanups to to cpu-hotplug to delete the fault injection test code - Improvements to vm test to use top_srcdir for builds" * tag 'linux-kselftest-next-6.1-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest: docs:kselftest: fix kselftest_module.h path of example module cpufreq: amd-pstate: Add explanation for X86_AMD_PSTATE_UT selftests/cpu-hotplug: Add log info when test success selftests/cpu-hotplug: Reserve one cpu online at least selftests/cpu-hotplug: Delete fault injection related code selftests/cpu-hotplug: Use return instead of exit selftests/cpu-hotplug: Correct log info cpufreq: amd-pstate: modify type in argument 2 for filp_open Documentation: amd-pstate: Add unit test introduction selftests: amd-pstate: Add test trigger for amd-pstate driver cpufreq: amd-pstate: Add test module for amd-pstate driver cpufreq: amd-pstate: Expose struct amd_cpudata selftests/vm: use top_srcdir instead of recomputing relative paths 2022-10-05T17:05:17+00:00 Meng Li li.meng@amd.com 2022-08-17T03:46:27+00:00 urn:sha1:f1375ec1df09fb09fecfdf9418d3784e7ba12469 Expose struct amd_cpudata to AMD P-State unit test module. This data struct will be used on the following AMD P-State unit test (amd-pstate-ut) module. The amd-pstate-ut module can get some AMD infomations by this data struct. For example: highest perf, nominal perf, boost supported etc. Signed-off-by: Meng Li <li.meng@amd.com> Acked-by: Huang Rui <ray.huang@amd.com> Acked-by: Shuah Khan <skhan@linuxfoundation.org> Signed-off-by: Shuah Khan <skhan@linuxfoundation.org> 2022-10-03T20:26:47+00:00 Linus Torvalds torvalds@linux-foundation.org 2022-10-03T20:26:47+00:00 urn:sha1:c79e6fa98ca8628556a01fe277022bda64829fdf Pull power management updates from Rafael Wysocki: "These add support for some new hardware, extend the existing hardware support, fix some issues and clean up code Specifics: - Add isupport for Tiger Lake in no-HWP mode to intel_pstate (Doug Smythies) - Update the AMD P-state driver (Perry Yuan): - Fix wrong lowest perf fetch - Map desired perf into pstate scope for powersave governor - Update pstate frequency transition delay time - Fix initial highest_perf value - Clean up - Move max CPU capacity to sugov_policy in the schedutil cpufreq governor (Lukasz Luba) - Add SM6115 to cpufreq-dt blocklist (Adam Skladowski) - Add support for Tegra239 and minor cleanups (Sumit Gupta, ye xingchen, and Yang Yingliang) - Add freq qos for qcom cpufreq driver and minor cleanups (Xuewen Yan, and Viresh Kumar) - Minor cleanups around functions called at module_init() (Xiu Jianfeng) - Use module_init and add module_exit for bmips driver (Zhang Jianhua) - Add AlderLake-N support to intel_idle (Zhang Rui) - Replace strlcpy() with unused retval with strscpy() in intel_idle (Wolfram Sang) - Remove redundant check from cpuidle_switch_governor() (Yu Liao) - Replace strlcpy() with unused retval with strscpy() in the powernv cpuidle driver (Wolfram Sang) - Drop duplicate word from a comment in the coupled cpuidle driver (Jason Wang) - Make rpm_resume() return -EINPROGRESS if RPM_NOWAIT is passed to it in the flags and the device is about to resume (Rafael Wysocki) - Add extra debugging statement for multiple active IRQs to system wakeup handling code (Mario Limonciello) - Replace strlcpy() with unused retval with strscpy() in the core system suspend support code (Wolfram Sang) - Update the intel_rapl power capping driver: - Use standard Energy Unit for SPR Dram RAPL domain (Zhang Rui). - Add support for RAPTORLAKE_S (Zhang Rui). - Fix UBSAN shift-out-of-bounds issue (Chao Qin) - Handle -EPROBE_DEFER when regulator is not probed on mtk-ci-devfreq.c (AngeloGioacchino Del Regno) - Fix message typo and use dev_err_probe() in rockchip-dfi.c (Christophe JAILLET)" * tag 'pm-6.1-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (29 commits) cpufreq: qcom-cpufreq-hw: Add cpufreq qos for LMh cpufreq: Add __init annotation to module init funcs cpufreq: tegra194: change tegra239_cpufreq_soc to static PM / devfreq: rockchip-dfi: Fix an error message PM / devfreq: mtk-cci: Handle sram regulator probe deferral powercap: intel_rapl: Use standard Energy Unit for SPR Dram RAPL domain PM: runtime: Return -EINPROGRESS from rpm_resume() in the RPM_NOWAIT case intel_idle: Add AlderLake-N support powercap: intel_rapl: fix UBSAN shift-out-of-bounds issue cpufreq: tegra194: Add support for Tegra239 cpufreq: qcom-cpufreq-hw: Fix uninitialized throttled_freq warning cpufreq: intel_pstate: Add Tigerlake support in no-HWP mode powercap: intel_rapl: Add support for RAPTORLAKE_S cpufreq: amd-pstate: Fix initial highest_perf value cpuidle: Remove redundant check in cpuidle_switch_governor() PM: wakeup: Add extra debugging statement for multiple active IRQs cpufreq: tegra194: Remove the unneeded result variable PM: suspend: move from strlcpy() with unused retval to strscpy() intel_idle: move from strlcpy() with unused retval to strscpy() cpuidle: powernv: move from strlcpy() with unused retval to strscpy() ... 2022-09-03T18:17:32+00:00 Perry Yuan Perry.Yuan@amd.com 2022-08-30T01:56:00+00:00 urn:sha1:bedadcfb011fef55273bd686e8893fdd8911dcdb To avoid some new AMD processors use wrong highest perf when amd pstate driver loaded, this fix will query the highest perf from MSR register MSR_AMD_CPPC_CAP1 and cppc_acpi interface firstly, then compare with the highest perf value got by calling amd_get_highest_perf() function. The lower value will be the correct highest perf we need to use. Otherwise the CPU max MHz will be incorrect if the amd_get_highest_perf() did not cover the new process family and model ID. Like this lscpu info, the max frequency is incorrect. Vendor ID: AuthenticAMD Socket(s): 1 Stepping: 2 CPU max MHz: 5410.0000 CPU min MHz: 400.0000 BogoMIPS: 5600.54 Fixes: 3743d55b289c2 (x86, sched: Fix the AMD CPPC maximum performance value on certain AMD Ryzen generations) Acked-by: Huang Rui <ray.huang@amd.com> Signed-off-by: Perry Yuan <Perry.Yuan@amd.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 2022-08-31T19:05:38+00:00 Perry Yuan Perry.Yuan@amd.com 2022-08-14T16:35:47+00:00 urn:sha1:ca08e46d4215e85eb3cef2481255431d98da56da Change the default transition latency to be 20ms that is more reasonable transition delay for AMD processors in non-EPP driver mode. Update transition delay time to 1ms, in the AMD CPU autonomous mode and non-autonomous mode, CPPC firmware will decide frequency at 1ms timescale based on the workload utilization. Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Perry Yuan <Perry.Yuan@amd.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 2022-08-31T19:05:38+00:00 Perry Yuan Perry.Yuan@amd.com 2022-08-14T16:35:46+00:00 urn:sha1:0e9a86386bca646d3b0fc50e5e15eedb1eb620b3 The patch will fix the invalid desired perf value for powersave governor. This issue is found when testing on one AMD EPYC system, the actual des_perf is smaller than the min_perf value, that is invalid value. because the min_perf is the lowest_perf system can support in idle state. Reviewed-by: Huang Rui <ray.huang@amd.com> Signed-off-by: Perry Yuan <Perry.Yuan@amd.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 2022-08-31T19:05:37+00:00 Perry Yuan Perry.Yuan@amd.com 2022-08-14T16:35:45+00:00 urn:sha1:b185c5053c65b7704ead4537e4d4d9b33dc398dc Fix the wrong lowest perf value reading which is used for new des_perf calculation by governor requested, the incorrect min_perf will get incorrect des_perf to be set , that will cause the system frequency changing unexpectedly. Reviewed-by: Huang Rui <ray.huang@amd.com> Acked-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Perry Yuan <Perry.Yuan@amd.com> Signed-off-by: Su Jinzhou <jinzhou.su@amd.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>