You are not logged in.
Hello,
I've been having difficulty getting cpupower running again, and I'm not entirely sure when it stopped working.
I'm on linux-lts 6.12.45-1, using a Ryzen 3700X.
I thought that amd-pstate-epp might be the issue, so I tried:
- loading acpi-cpufreq. but I get this: modprobe: ERROR: could not insert 'acpi_cpufreq': No such device
- loading amd-pstate, it loads but didn't fix the issue. Maybe I need to unload the amd-pstate-epp first?
Here's some info:
$ systemctl status cpupower.service --no-pager -l
× cpupower.service - Apply cpupower configuration
Loaded: loaded (/usr/lib/systemd/system/cpupower.service; enabled; preset: disabled)
Active: failed (Result: exit-code) since Sat 2025-09-06 10:52:49 BST; 4min 29s ago
Invocation: c0916234e9fd4e7db827e2691631cb94
Process: 1924981 ExecStart=/usr/lib/systemd/scripts/cpupower (code=exited, status=1/FAILURE)
Main PID: 1924981 (code=exited, status=1/FAILURE)
Mem peak: 1.7M
CPU: 7ms
Sep 06 10:52:49 samurai systemd[1]: Starting Apply cpupower configuration...
Sep 06 10:52:49 samurai systemd[1]: cpupower.service: Main process exited, code=exited, status=1/FAILURE
Sep 06 10:52:49 samurai systemd[1]: cpupower.service: Failed with result 'exit-code'.
Sep 06 10:52:49 samurai systemd[1]: Failed to start Apply cpupower configuration.
$ ls /usr/lib/modules/$(uname -r)/kernel/drivers/cpufreq/
acpi-cpufreq.ko.zst amd-pstate-ut.ko.zst pcc-cpufreq.ko.zst speedstep-lib.ko.zst
amd_freq_sensitivity.ko.zst p4-clockmod.ko.zst powernow-k8.ko.zst
$ cpupower frequency-info
analyzing CPU 1:
driver: amd-pstate-epp
CPUs which run at the same hardware frequency: 1
CPUs which need to have their frequency coordinated by software: 1
energy performance preference: balance_power
hardware limits: 550 MHz - 4.43 GHz
available cpufreq governors: performance powersave
current policy: frequency should be within 550 MHz and 4.43 GHz.
The governor "powersave" may decide which speed to use
within this range.
current CPU frequency: 3.67 GHz (asserted by call to kernel)
boost state support:
Supported: yes
Active: yes
amd-pstate limits:
Highest Performance: 196. Maximum Frequency: 4.43 GHz.
Nominal Performance: 135. Nominal Frequency: 3.60 GHz.
Lowest Non-linear Performance: 66. Lowest Non-linear Frequency: 1.76 GHz.
Lowest Performance: 21. Lowest Frequency: 550 MHz.
Preferred Core Support: 1. Preferred Core Ranking: 196.
Thanks in advance for any help
Last edited by Qwerty-Space (2025-10-02 16:01:10)
Offline
Does running the script directly from the terminal provide any more information?
. /etc/default/cpupower
. /etc/cpupower-service.conf
/usr/lib/systemd/scripts/cpupower
You can use tmpfiles.d(5) instead, that can do everything cpupower can do but without the extra package, service & scripts. But I suppose that is off-topic for this thread, strictly speaking.
Jin, Jîyan, Azadî
Offline
Does running the script directly from the terminal provide any more information?
. /etc/default/cpupower . /etc/cpupower-service.conf /usr/lib/systemd/scripts/cpupower
You can use tmpfiles.d(5) instead, that can do everything cpupower can do but without the extra package, service & scripts. But I suppose that is off-topic for this thread, strictly speaking.
Everything in cpupower-service.conf is commented.
/etc/default/cpupower is not executable apparently. /usr/lib/systemd/scripts/cpupower outputs nothing.
Offline
/etc/default/cpupower is not executable
My suggested command sourced the file rather than executed it. Did you include the space after the dot?
/usr/lib/systemd/scripts/cpupower outputs nothing
Did it effect the desired changes?
Does the journal show anything more about cpupower.service?
journalctl -u cpupower.service
If not we should probably check the entire journal for context.
Jin, Jîyan, Azadî
Offline
Oh I wasn't aware that that sources it. I've tried that again, it seemed to source it, but the service still won't run.
output of journalctl -xeu:
Sep 06 20:34:46 samurai systemd[1]: Starting Apply cpupower configuration...
░░ Subject: A start job for unit cpupower.service has begun execution
░░ Defined-By: systemd
░░ Support: https://lists.freedesktop.org/mailman/listinfo/systemd-devel
░░
░░ A start job for unit cpupower.service has begun execution.
░░
░░ The job identifier is 5613.
Sep 06 20:34:46 samurai systemd[1]: cpupower.service: Main process exited, code=exited, status=1/FAILURE
░░ Subject: Unit process exited
░░ Defined-By: systemd
░░ Support: https://lists.freedesktop.org/mailman/listinfo/systemd-devel
░░
░░ An ExecStart= process belonging to unit cpupower.service has exited.
░░
░░ The process' exit code is 'exited' and its exit status is 1.
Sep 06 20:34:46 samurai systemd[1]: cpupower.service: Failed with result 'exit-code'.
░░ Subject: Unit failed
░░ Defined-By: systemd
░░ Support: https://lists.freedesktop.org/mailman/listinfo/systemd-devel
░░
░░ The unit cpupower.service has entered the 'failed' state with result 'exit-code'.
Sep 06 20:34:46 samurai systemd[1]: Failed to start Apply cpupower configuration.
░░ Subject: A start job for unit cpupower.service has failed
░░ Defined-By: systemd
░░ Support: https://lists.freedesktop.org/mailman/listinfo/systemd-devel
░░
░░ A start job for unit cpupower.service has finished with a failure.
░░
░░ The job identifier is 5613 and the job result is failed.
Last edited by Qwerty-Space (2025-09-06 19:55:34)
Offline
Please stop full-quoting unnecessarily, it makes the thread much harder to follow.
I didn't ask you to use the -x & -e options for the journalctl command because I didn't want you to use those options. Your posted output is cropped because it's run in a pager and has useless spam from the -x option.
Please share a full journal, preferably from as brief a boot as possible.
And once more:
Did it effect the desired changes?
For clarity: when you ran the cpupower script directly from the terminal after sourcing the configuration file did it change the configuration so that it matched what is defined in the configuration file?
FWIW https://bbs.archlinux.org/viewtopic.php?id=278187 suggests this might be a kernel regression. Does the non-LTS kernel behave differently? It might be worth debugging the cpupower script, see seth's suggestions in the linked thread for the method (bash -x).
Jin, Jîyan, Azadî
Offline
Sorry for the delay work has been busy, here's my journalctl -b --no-pager: https://pastebin.com/raw/xFRaCKJm
Here's my journalctl -u cpupower --no-pager:
Sep 26 16:50:01 samurai systemd[1]: Starting Apply cpupower configuration...
Sep 26 16:50:01 samurai systemd[1]: cpupower.service: Main process exited, code=exited, status=1/FAILURE
Sep 26 16:50:01 samurai systemd[1]: cpupower.service: Failed with result 'exit-code'.
Sep 26 16:50:01 samurai systemd[1]: Failed to start Apply cpupower configuration.
Offline
No, it didn't effect the desired changes. cpupower service still won't run.
For clarity: when you ran the cpupower script directly from the terminal after sourcing the configuration file did it change the configuration so that it matched what is defined in the configuration file?
I'm not entirely sure how to verify this.
See below for the config file:
$ cat /etc/default/cpupower
# Define CPUs governor
# valid governors: ondemand, performance, powersave, conservative, userspace.
governor='ondemand'
# Limit frequency range
# Valid suffixes: Hz, kHz (default), MHz, GHz, THz
#min_freq="2.25GHz"
#max_freq="3GHz"
# Specific frequency to be set.
# Requires userspace governor to be available.
# Do not set governor field if you use this one.
#freq=
# Utilizes cores in one processor package/socket first before processes are
# scheduled to other processor packages/sockets.
# See man (1) CPUPOWER-SET for additional details.
#mc_scheduler=
# Utilizes thread siblings of one processor core first before processes are
# scheduled to other cores. See man (1) CPUPOWER-SET for additional details.
#smp_scheduler=
# Sets a register on supported Intel processore which allows software to convey
# its policy for the relative importance of performance versus energy savings to
# the processor. See man (1) CPUPOWER-SET for additional details.
#perf_bias=
# vim:set ts=2 sw=2 ft=sh et:
Last edited by Qwerty-Space (2025-09-26 16:16:03)
Offline
The ondemand governor doesn't exist for amd-pstate. Technically powersave should have roughly equivalent semantics on the pstate drivers. So if this is the reason for running the service you can as well just disable it, since powersave is default.
Offline
Thanks V1del
Offline