[Bug 221473] New: amd-pstate dynamic EPP regression on Ryzen 5 3600

[Bug 221473] New: amd-pstate dynamic EPP regression on Ryzen 5 3600

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221473

            Bug ID: 221473
           Summary: amd-pstate dynamic EPP regression on Ryzen 5 3600
           Product: Power Management
           Version: 2.5
    Kernel Version: 7.1.0-rc2
          Hardware: AMD
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P3
         Component: cpufreq
          Assignee: linux-pm@vger.kernel.org
          Reporter: stuartmeckle@gmail.com
                CC: mario.limonciello@amd.com
        Regression: Yes
           Bisected e30ca6dd5345c5b8ba05f346a8e81105352fe571
         commit-id:

Created attachment 310042
  --> https://bugzilla.kernel.org/attachment.cgi?id=310042&action=edit
dmesg and config

amd-pstate dynamic EPP regression on Ryzen 5 3600

My CPU which normally boosts to 4.0GHz (PBO enabled) was stuck at 1.7GHz under
full load. Same result with CONFIG_X86_AMD_PSTATE_DYNAMIC_EPP set to either 'y'
or 'n'. I bisected to this commit:

commit e30ca6dd5345c5b8ba05f346a8e81105352fe571
Author: Mario Limonciello (AMD) <superm1@kernel.org>

    cpufreq/amd-pstate: Add dynamic energy performance preference


Arch Linux
Linux version 7.1.0-rc2-00099-gadc1e5c6203c (stuart@romano) (gcc (GCC) 16.1.1
20260430, GNU ld (GNU Binutils) 2.46.0) #1 SMP PREEMPT_DYNAMIC Wed May  6
15:30:26 UTC 2026
Gigabyte Technology Co., Ltd. B450M DS3H/B450M DS3H-CF, BIOS F67 10/28/2025
Advanced Micro Devices, Inc. [AMD/ATI] Navi 44 [Radeon RX 9060 XT] (rev c0)
AMD Ryzen 5 3600 6-Core Processor
32GB DDR4 DRAM F4-3000C16-8GVRB


CONFIG_CPU_FREQ_DEFAULT_GOV_SCHEDUTIL=y
CONFIG_CPU_FREQ_GOV_SCHEDUTIL=y
CONFIG_X86_AMD_PSTATE=y
CONFIG_X86_AMD_PSTATE_DEFAULT_MODE=3
CONFIG_X86_AMD_PSTATE_DYNAMIC_EPP=y
CONFIG_X86_AMD_PSTATE_UT=m


==> /sys/devices/system/cpu/cpu0/cpufreq/affected_cpus <==
0
==> /sys/devices/system/cpu/cpu0/cpufreq/amd_pstate_highest_perf <==
166
==> /sys/devices/system/cpu/cpu0/cpufreq/amd_pstate_hw_prefcore <==
enabled
==> /sys/devices/system/cpu/cpu0/cpufreq/amd_pstate_lowest_nonlinear_freq <==
1723944
==> /sys/devices/system/cpu/cpu0/cpufreq/amd_pstate_max_freq <==
4208451
==> /sys/devices/system/cpu/cpu0/cpufreq/amd_pstate_prefcore_ranking <==
171
==> /sys/devices/system/cpu/cpu0/cpufreq/boost <==
1
==> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_avg_freq <==
1723944
==> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq <==
4208451
==> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq <==
557747
==> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency <==
0
==>
/sys/devices/system/cpu/cpu0/cpufreq/energy_performance_available_preferences
<==
default performance balance_performance balance_power power custom
==> /sys/devices/system/cpu/cpu0/cpufreq/energy_performance_preference <==
performance
==> /sys/devices/system/cpu/cpu0/cpufreq/related_cpus <==
0
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors <==
performance powersave
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq <==
1736801
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver <==
amd-pstate-epp
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor <==
powersave
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq <==
4208451
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq <==
1723944
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed <==
<unsupported>

For comparison, here are the same settings on kernel 7.0.3-arch1-2:

==> /sys/devices/system/cpu/cpu0/cpufreq/affected_cpus <==
0
==> /sys/devices/system/cpu/cpu0/cpufreq/amd_pstate_highest_perf <==
166
==> /sys/devices/system/cpu/cpu0/cpufreq/amd_pstate_hw_prefcore <==
enabled
==> /sys/devices/system/cpu/cpu0/cpufreq/amd_pstate_lowest_nonlinear_freq <==
1723944
==> /sys/devices/system/cpu/cpu0/cpufreq/amd_pstate_max_freq <==
4208451
==> /sys/devices/system/cpu/cpu0/cpufreq/amd_pstate_prefcore_ranking <==
171
==> /sys/devices/system/cpu/cpu0/cpufreq/boost <==
1
==> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_avg_freq <==
3656148
==> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq <==
4208451
==> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq <==
557747
==> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency <==
0
==>
/sys/devices/system/cpu/cpu0/cpufreq/energy_performance_available_preferences
<==
default performance balance_performance balance_power power
==> /sys/devices/system/cpu/cpu0/cpufreq/energy_performance_preference <==
balance_performance
==> /sys/devices/system/cpu/cpu0/cpufreq/related_cpus <==
0
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors <==
performance powersave
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq <==
3693566
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver <==
amd-pstate-epp
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor <==
powersave
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq <==
4208451
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq <==
1723944
==> /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed <==
<unsupported>

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are the assignee for the bug.

[Bug 221473] amd-pstate dynamic EPP regression on Ryzen 5 3600

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221473

--- Comment #1 from Mario Limonciello (AMD) (mario.limonciello@amd.com) ---
Can you please share an acpidump as well as the output from the amd-pstate tool
at https://git.kernel.org/pub/scm/linux/kernel/git/superm1/amd-debug-tools.git/

Thanks!

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are the assignee for the bug.

[Bug 221473] amd-pstate dynamic EPP regression on Ryzen 5 3600

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221473

Mario Limonciello (AMD) (mario.limonciello@amd.com) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|linux-pm@vger.kernel.org    |mario.limonciello@amd.com

--- Comment #2 from Mario Limonciello (AMD) (mario.limonciello@amd.com) ---
Created attachment 310056
  --> https://bugzilla.kernel.org/attachment.cgi?id=310056&action=edit
possible patch

Can you see if this patch helps - specifically in dynamic EPP turned off.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are the assignee for the bug.