Re: Cpufreq drivers not working on T480S

[Jan Beulich <jbeulich@xxxxxxxx>]

On 06.01.2026 02:03, Jason Andryuk wrote:
> On Tue, Dec 30, 2025 at 8:50 PM Milky <milky_way_303030@xxxxxxxxx> wrote:
>> (Re-CC'ing the ML because I forgot by accident. Hopefully the quoted
>> sections provide sufficient context)
>>
>> On Tuesday, December 30th, 2025 at 10:44 AM, Jason Andryuk 
>> <jandryuk@xxxxxxxxx> wrote:
> 
>> As suggested, I added the debug parameters to the dom0 kernel. Before or
>> after `modprobe xen-acpi-processor dyndbg==pmf`, there is no useful
>> debug information that I could find, apart from the
>> `xen_acpi_processor:get_max_acpi_id` message as seen below.
>>
>> ```
>> # sudo dmesg | grep xen.acpi
>> [    2.282851] Kernel command line: placeholder 
>> root=/dev/mapper/qubes_dom0-root ro rd.luks.uuid=<...> 
>> rd.lvm.lv=qubes_dom0/root rd.lvm.lv=qubes_dom0/swap 
>> plymouth.ignore-serial-consoles 6.6.77-1.qubes.fc37.x86_64 x86_64 rhgb 
>> loglevel=9 "dyndbg=module xen_acpi_processor +p" 
>> "xen_acpi_processor.dyndbg=func * +p" rd.qubes.hide_all_usb
>> [    5.224092] xen_acpi_processor: Max ACPI ID: 6
> 
> You successfully turned on dyndbg to get that output, but there is no
> further output.  This makes me think something else is wrong and
> xen-acpi-processor doesn't upload anything.
> 
> The call here 
> https://elixir.bootlin.com/linux/v6.18.2/source/drivers/xen/xen-acpi-processor.c#L557
> to
> https://elixir.bootlin.com/linux/v6.18.2/source/drivers/acpi/processor_perflib.c#L421
> goes into some acpi code.  Maybe there are other messages in dmesg
> around the same time?  Maybe you'd have to turn on more debugging to
> get them.
> 
>> # sudo lsmod | grep xen_acpi
>> <no output>
>>
>> # sudo modprobe xen-acpi-processor dyndbg==pmf
>> modprobe: ERROR: could not insert 'xen_acpi_processor': No such device
>> ```
> 
>>> Maybe also with Xen's command line try cpufreq=xen:no-hwp to disable
>>> HWP and see if the regular ACPI cpufreq driver works better.
>>>
>>> I'm thinking it's something where xen-acpi-processor didn't upload
>>> ACPI CPU data, which means cpufreq isn't running. That may also be
>>> why you see that bogus CPU frequency.
>>
>> After booting with `xen:no-hwp`, I wasn't sure how to check if the
>> regular ACPI cpufreq driver is operational. Is `xenpm` still the
>> correct way to query for CPU info? I've tried the following:
>>
>> ```
>> # sudo xl dmesg | grep -i hwp
>> (XEN) Command line: placeholder cpufreq=xen:no-hwp,verbose loglvl=all 
>> dom0_mem=min:1024M dom0_mem=max:4096M ucode=scan smt=off 
>> gnttab_max_frames=2048 gnttab_max_maptrack_frames=4096 no-real-mode edd=off
>> (XEN) HWP: 1 notify: 1 act-window: 1 energy-perf: 1 pkg-level: 0 peci: 0
>> (XEN) HWP: Hardware Duty Cycling (HDC) supported, enabled
>> (XEN) HWP: HW_FEEDBACK not supported
> 
> no-hwp failed to disable HWP.  But if there is no ACPI CPU data, it
> wouldn't work either.

There isn't any "no-hwp" option that we would recognize, is there? Iirc HWP
isn't enabled by default, so simply not saying "cpufreq=hwp" should disable
the driver? (I already found the original report confusing in this regard,
hence why I preferred to not reply so far. I wonder if there are local
patches in use.)

Jan