[PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[Rong Zhang]

When booting a virtual machine that simulates or passes-through a
relatively new CPUID, one or two warning messages are printed to kmsg:

    amd_pstate: The CPPC feature is supported but currently disabled by the BIOS.
    Please enable it if your BIOS has the CPPC option.
    amd_pstate: the _CPC object is not present in SBIOS or ACPI disabled

Technically the check is not wrong and acts as a safetynet that prevents
the driver from being registered incorrectly, but the warning messages
are noisy and not really helpful, as virtual machines cannot adjust the
CPU frequency from the very beginning. Moreover, mentioning BIOS here
makes the messages even more confusing.

Fix it by bailing out early when running in virtual machines (excluding
Xen Dom0).

Fixes: cb817ec6673b ("cpufreq: amd-pstate: show CPPC debug message if CPPC is not supported")
Signed-off-by: Rong Zhang <i@rong.moe>
---
 drivers/cpufreq/amd-pstate.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c
index a74a4cf99d22..f2115f935318 100644
--- a/drivers/cpufreq/amd-pstate.c
+++ b/drivers/cpufreq/amd-pstate.c
@@ -2228,6 +2228,9 @@ static int __init amd_pstate_init(void)
 	if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD)
 		return -ENODEV;
 
+	if (cpu_feature_enabled(X86_FEATURE_HYPERVISOR) && !xen_initial_domain())
+		return -ENODEV;
+
 	/* show debug message only if CPPC is not supported */
 	if (!amd_cppc_supported())
 		return -EOPNOTSUPP;

---
base-commit: 58717b2a1365d06c8c64b72aa948541b53fe31eb
change-id: 1cc24037-amd-pstate-vm-d6ab4c959bd3

Thanks,
Rong

Re: [PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[Mario Limonciello]

+Jason
+Boris

On 7/15/26 12:41, Rong Zhang wrote:
> When booting a virtual machine that simulates or passes-through a
> relatively new CPUID, one or two warning messages are printed to kmsg:
> 
>      amd_pstate: The CPPC feature is supported but currently disabled by the BIOS.
>      Please enable it if your BIOS has the CPPC option.
>      amd_pstate: the _CPC object is not present in SBIOS or ACPI disabled
> 
> Technically the check is not wrong and acts as a safetynet that prevents
> the driver from being registered incorrectly, but the warning messages
> are noisy and not really helpful, as virtual machines cannot adjust the
> CPU frequency from the very beginning. Moreover, mentioning BIOS here
> makes the messages even more confusing.
> 
> Fix it by bailing out early when running in virtual machines (excluding
> Xen Dom0).

Conceptually this makes sense to me, but I would like some comments from 
others if this is the right way forward.

Jason, Boris,

comments please?

> 
> Fixes: cb817ec6673b ("cpufreq: amd-pstate: show CPPC debug message if CPPC is not supported")
> Signed-off-by: Rong Zhang <i@rong.moe>
> ---
>   drivers/cpufreq/amd-pstate.c | 3 +++
>   1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c
> index a74a4cf99d22..f2115f935318 100644
> --- a/drivers/cpufreq/amd-pstate.c
> +++ b/drivers/cpufreq/amd-pstate.c
> @@ -2228,6 +2228,9 @@ static int __init amd_pstate_init(void)
>   	if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD)
>   		return -ENODEV;
>   
> +	if (cpu_feature_enabled(X86_FEATURE_HYPERVISOR) && !xen_initial_domain())
> +		return -ENODEV;
> +
>   	/* show debug message only if CPPC is not supported */
>   	if (!amd_cppc_supported())
>   		return -EOPNOTSUPP;
> 
> ---
> base-commit: 58717b2a1365d06c8c64b72aa948541b53fe31eb
> change-id: 1cc24037-amd-pstate-vm-d6ab4c959bd3
> 
> Thanks,
> Rong
> 

Re: [PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[Borislav Petkov]

On Wed, Jul 15, 2026 at 12:46:24PM -0500, Mario Limonciello wrote:
> Conceptually this makes sense to me, but I would like some comments from
> others if this is the right way forward.
> 
> Jason, Boris,
> 
> comments please?

We don't care if someone "passes-through a relatively new CPUID" - I am not
sure what that means - but we won't patch the upstream kernel just because it
dumped an error message in the guest.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: [PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[Rong Zhang]

Hi Borislav,

Thanks for reviewing the patch.

On Wed, 2026-07-15 at 11:04 -0700, Borislav Petkov wrote:
> On Wed, Jul 15, 2026 at 12:46:24PM -0500, Mario Limonciello wrote:
> > Conceptually this makes sense to me, but I would like some comments from
> > others if this is the right way forward.
> > 
> > Jason, Boris,
> > 
> > comments please?
> 
> We don't care if someone "passes-through a relatively new CPUID" - I am not
> sure what that means - 
> 

Sorry, typo here. I meant CPU model.

I usually boot a KVM with `qemu-system-x86_64 -accel kvm -cpu host
<other_args...>' to test various patches. These messages is a bit
disturbing to me as they are KERN_WARNING messages.

FYI, my device is Zen 5 based and always triggers both messages.

I can also reproduce both messages with `-cpu EPYC-Genoa' and `-cpu
EPYC-Turin'. It's just due to how amd_cppc_supported() is implemented.

> but we won't patch the upstream kernel just because it
> dumped an error message in the guest.

Well, the reason why I decided to send this is not only about how noisy
it is, but also about the correctness.

The first messages says:

   The CPPC feature is supported but currently disabled by the BIOS.
   Please enable it if your BIOS has the CPPC option.

The CPPC feature is neither supported in a KVM nor disabled by the VM's
BIOS.

IIUC, it's disabled by kvm_cpu_cap_init(CPUID_8000_0008_EBX, ...), so
the upstream kernel both disables it as a hypervisor and warns about it
as a guest. What a paradox!

At least the first message is nonsense in this case and the paradox
should really be fixed. If you prefer preserving the second message
("the _CPC object is not present in SBIOS or ACPI disabled") as an
informative hint anyway, how about moving the virtualization check into
amd_cppc_supported() to only mute the first one? ;-)

Thanks,
Rong

Re: [PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[Jason Andryuk]

On 2026-07-15 15:21, Rong Zhang wrote:
> Hi Borislav,
> 
> Thanks for reviewing the patch.
> 
> On Wed, 2026-07-15 at 11:04 -0700, Borislav Petkov wrote:
>> On Wed, Jul 15, 2026 at 12:46:24PM -0500, Mario Limonciello wrote:
>>> Conceptually this makes sense to me, but I would like some comments from
>>> others if this is the right way forward.
>>>
>>> Jason, Boris,
>>>
>>> comments please?
>>
>> We don't care if someone "passes-through a relatively new CPUID" - I am not
>> sure what that means -
>>
> 
> Sorry, typo here. I meant CPU model.
> 
> I usually boot a KVM with `qemu-system-x86_64 -accel kvm -cpu host
> <other_args...>' to test various patches. These messages is a bit
> disturbing to me as they are KERN_WARNING messages.
> 
> FYI, my device is Zen 5 based and always triggers both messages.
> 
> I can also reproduce both messages with `-cpu EPYC-Genoa' and `-cpu
> EPYC-Turin'. It's just due to how amd_cppc_supported() is implemented.
> 
>> but we won't patch the upstream kernel just because it
>> dumped an error message in the guest.
> 
> Well, the reason why I decided to send this is not only about how noisy
> it is, but also about the correctness.
> 
> The first messages says:
> 
>     The CPPC feature is supported but currently disabled by the BIOS.
>     Please enable it if your BIOS has the CPPC option.
> 
> The CPPC feature is neither supported in a KVM nor disabled by the VM's
> BIOS.
> 
> IIUC, it's disabled by kvm_cpu_cap_init(CPUID_8000_0008_EBX, ...), so
> the upstream kernel both disables it as a hypervisor and warns about it
> as a guest. What a paradox!
> 
> At least the first message is nonsense in this case and the paradox
> should really be fixed. If you prefer preserving the second message
> ("the _CPC object is not present in SBIOS or ACPI disabled") as an
> informative hint anyway, how about moving the virtualization check into
> amd_cppc_supported() to only mute the first one? ;-)

I think changing amd_cppc_supported() to return false for 
X86_FEATURE_HYPERVISOR makes sense and can remove the message.  The 
heuristic there doesn't seem relevant when virtualized.

With Xen, the Xen hypervisor itself controls CPPC.  dom0 
(xen_initial_domain()) doesn't need the amd-pstate driver.

Regards,
Jason

Re: [PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[Mario Limonciello]

On 7/15/26 15:32, Jason Andryuk wrote:
> On 2026-07-15 15:21, Rong Zhang wrote:
>> Hi Borislav,
>>
>> Thanks for reviewing the patch.
>>
>> On Wed, 2026-07-15 at 11:04 -0700, Borislav Petkov wrote:
>>> On Wed, Jul 15, 2026 at 12:46:24PM -0500, Mario Limonciello wrote:
>>>> Conceptually this makes sense to me, but I would like some comments 
>>>> from
>>>> others if this is the right way forward.
>>>>
>>>> Jason, Boris,
>>>>
>>>> comments please?
>>>
>>> We don't care if someone "passes-through a relatively new CPUID" - I 
>>> am not
>>> sure what that means -
>>>
>>
>> Sorry, typo here. I meant CPU model.
>>
>> I usually boot a KVM with `qemu-system-x86_64 -accel kvm -cpu host
>> <other_args...>' to test various patches. These messages is a bit
>> disturbing to me as they are KERN_WARNING messages.
>>
>> FYI, my device is Zen 5 based and always triggers both messages.
>>
>> I can also reproduce both messages with `-cpu EPYC-Genoa' and `-cpu
>> EPYC-Turin'. It's just due to how amd_cppc_supported() is implemented.
>>
>>> but we won't patch the upstream kernel just because it
>>> dumped an error message in the guest.
>>
>> Well, the reason why I decided to send this is not only about how noisy
>> it is, but also about the correctness.
>>
>> The first messages says:
>>
>>     The CPPC feature is supported but currently disabled by the BIOS.
>>     Please enable it if your BIOS has the CPPC option.
>>
>> The CPPC feature is neither supported in a KVM nor disabled by the VM's
>> BIOS.
>>
>> IIUC, it's disabled by kvm_cpu_cap_init(CPUID_8000_0008_EBX, ...), so
>> the upstream kernel both disables it as a hypervisor and warns about it
>> as a guest. What a paradox!
>>
>> At least the first message is nonsense in this case and the paradox
>> should really be fixed. If you prefer preserving the second message
>> ("the _CPC object is not present in SBIOS or ACPI disabled") as an
>> informative hint anyway, how about moving the virtualization check into
>> amd_cppc_supported() to only mute the first one? ;-)
> 
> I think changing amd_cppc_supported() to return false for 
> X86_FEATURE_HYPERVISOR makes sense and can remove the message.  The 
> heuristic there doesn't seem relevant when virtualized.
> 
> With Xen, the Xen hypervisor itself controls CPPC.  dom0 
> (xen_initial_domain()) doesn't need the amd-pstate driver.
> 
> Regards,
> Jason

OK, Rong - please send an updated patch.

Re: [PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[Borislav Petkov]

On Wed, Jul 15, 2026 at 04:32:53PM -0400, Jason Andryuk wrote:
> I think changing amd_cppc_supported() to return false for
> X86_FEATURE_HYPERVISOR makes sense and can remove the message.  The
> heuristic there doesn't seem relevant when virtualized.

I don't maintain that code but if I were, I won't sprinkle if
X86_FEATURE_HYPERVISOR for no good reason. And the guest warning is not good
enough a reason.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

RE: [PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[Michael Kelley]

From: Rong Zhang <i@rong.moe> Sent: Wednesday, July 15, 2026 12:22 PM
> 
> Hi Borislav,
> 
> Thanks for reviewing the patch.
> 
> On Wed, 2026-07-15 at 11:04 -0700, Borislav Petkov wrote:
> > On Wed, Jul 15, 2026 at 12:46:24PM -0500, Mario Limonciello wrote:
> > > Conceptually this makes sense to me, but I would like some comments from
> > > others if this is the right way forward.
> > >
> > > Jason, Boris,
> > >
> > > comments please?
> >
> > We don't care if someone "passes-through a relatively new CPUID" - I am not
> > sure what that means -
> >
> 
> Sorry, typo here. I meant CPU model.
> 
> I usually boot a KVM with `qemu-system-x86_64 -accel kvm -cpu host
> <other_args...>' to test various patches. These messages is a bit
> disturbing to me as they are KERN_WARNING messages.
> 
> FYI, my device is Zen 5 based and always triggers both messages.
> 
> I can also reproduce both messages with `-cpu EPYC-Genoa' and `-cpu
> EPYC-Turin'. It's just due to how amd_cppc_supported() is implemented.
> 
> > but we won't patch the upstream kernel just because it
> > dumped an error message in the guest.
> 
> Well, the reason why I decided to send this is not only about how noisy
> it is, but also about the correctness.
> 
> The first messages says:
> 
>    The CPPC feature is supported but currently disabled by the BIOS.
>    Please enable it if your BIOS has the CPPC option.
> 
> The CPPC feature is neither supported in a KVM nor disabled by the VM's
> BIOS.
> 
> IIUC, it's disabled by kvm_cpu_cap_init(CPUID_8000_0008_EBX, ...), so
> the upstream kernel both disables it as a hypervisor and warns about it
> as a guest. What a paradox!
> 
> At least the first message is nonsense in this case and the paradox
> should really be fixed. If you prefer preserving the second message
> ("the _CPC object is not present in SBIOS or ACPI disabled") as an
> informative hint anyway, how about moving the virtualization check into
> amd_cppc_supported() to only mute the first one? ;-)
> 
> Thanks,
> Rong

FWIW, the same warnings are output when running as a guest on Hyper-V
hosts. Over the years, I've cleaned up various bogus warnings and errors in
the boot logs for Hyper-V guests so they looks clean. I'd love to see these
go away as well.

Just my $.02 ...

Michael

Re: [PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[K Prateek Nayak]

Hello Rong,

On 7/15/2026 11:11 PM, Rong Zhang wrote:
> When booting a virtual machine that simulates or passes-through a
> relatively new CPUID, one or two warning messages are printed to kmsg:
> 
>     amd_pstate: The CPPC feature is supported but currently disabled by the BIOS.
>     Please enable it if your BIOS has the CPPC option.
>     amd_pstate: the _CPC object is not present in SBIOS or ACPI disabled
> 
> Technically the check is not wrong and acts as a safetynet that prevents
> the driver from being registered incorrectly, but the warning messages
> are noisy and not really helpful, as virtual machines cannot adjust the
> CPU frequency from the very beginning. Moreover, mentioning BIOS here
> makes the messages even more confusing.
> 
> Fix it by bailing out early when running in virtual machines (excluding
> Xen Dom0).
> 
> Fixes: cb817ec6673b ("cpufreq: amd-pstate: show CPPC debug message if CPPC is not supported")
> Signed-off-by: Rong Zhang <i@rong.moe>
> ---
>  drivers/cpufreq/amd-pstate.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c
> index a74a4cf99d22..f2115f935318 100644
> --- a/drivers/cpufreq/amd-pstate.c
> +++ b/drivers/cpufreq/amd-pstate.c
> @@ -2228,6 +2228,9 @@ static int __init amd_pstate_init(void)
>  	if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD)
>  		return -ENODEV;
>  
> +	if (cpu_feature_enabled(X86_FEATURE_HYPERVISOR) && !xen_initial_domain())
> +		return -ENODEV;
> +

I think, instead of this, you can just check for X86_FEATURE_HW_PSTATE.

Using check_amd_hwpstate_cpu() from drivers/cpufreq/acpi-cpufreq.c
here should be sufficient to know if the platform supports frequency
scaling or not.

If it does and _CPC object is missing, there is a good chance CPPC is
disabled in BIOS.

KVM doesn't expose X86_FEATURE_HW_PSTATE to guest. I'm not sure if
Xen does.

>  	/* show debug message only if CPPC is not supported */
>  	if (!amd_cppc_supported())
>  		return -EOPNOTSUPP;

-- 
Thanks and Regards,
Prateek

Re: [PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[Rong Zhang]

Hi K Prateek,

Thanks for reviewing this.

On Thu, 2026-07-16 at 23:30 +0530, K Prateek Nayak wrote:
> Hello Rong,
> 
> On 7/15/2026 11:11 PM, Rong Zhang wrote:
> > When booting a virtual machine that simulates or passes-through a
> > relatively new CPUID, one or two warning messages are printed to kmsg:
> > 
> >     amd_pstate: The CPPC feature is supported but currently disabled by the BIOS.
> >     Please enable it if your BIOS has the CPPC option.
> >     amd_pstate: the _CPC object is not present in SBIOS or ACPI disabled
> > 
> > Technically the check is not wrong and acts as a safetynet that prevents
> > the driver from being registered incorrectly, but the warning messages
> > are noisy and not really helpful, as virtual machines cannot adjust the
> > CPU frequency from the very beginning. Moreover, mentioning BIOS here
> > makes the messages even more confusing.
> > 
> > Fix it by bailing out early when running in virtual machines (excluding
> > Xen Dom0).
> > 
> > Fixes: cb817ec6673b ("cpufreq: amd-pstate: show CPPC debug message if CPPC is not supported")
> > Signed-off-by: Rong Zhang <i@rong.moe>
> > ---
> >  drivers/cpufreq/amd-pstate.c | 3 +++
> >  1 file changed, 3 insertions(+)
> > 
> > diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c
> > index a74a4cf99d22..f2115f935318 100644
> > --- a/drivers/cpufreq/amd-pstate.c
> > +++ b/drivers/cpufreq/amd-pstate.c
> > @@ -2228,6 +2228,9 @@ static int __init amd_pstate_init(void)
> >  	if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD)
> >  		return -ENODEV;
> >  
> > +	if (cpu_feature_enabled(X86_FEATURE_HYPERVISOR) && !xen_initial_domain())
> > +		return -ENODEV;
> > +
> 
> I think, instead of this, you can just check for X86_FEATURE_HW_PSTATE.
> 
> Using check_amd_hwpstate_cpu() from drivers/cpufreq/acpi-cpufreq.c
> here should be sufficient to know if the platform supports frequency
> scaling or not.
> 
> If it does and _CPC object is missing, there is a good chance CPPC is
> disabled in BIOS.

Thanks a lot for your suggestion! It makes a lot of sense to me and
sounds promising. I will try the approach in v2.

> 
> KVM doesn't expose X86_FEATURE_HW_PSTATE to guest. I'm not sure if
> Xen does.

As Jason said:

   With Xen, the Xen hypervisor itself controls CPPC.  dom0
   (xen_initial_domain()) doesn't need the amd-pstate driver.

I think we don't need to worry about Xen.

@Michael, if you have the time, could you test if Hyper-V exposes
X86_FEATURE_HW_PSTATE (i.e., 0x80000007/edx bit 7)?

Thanks,
Rong

> 
> >  	/* show debug message only if CPPC is not supported */
> >  	if (!amd_cppc_supported())
> >  		return -EOPNOTSUPP;

RE: [PATCH] cpufreq/amd-pstate: Bail out early in virtual machines and don't warn

[Michael Kelley]

From: Rong Zhang <i@rong.moe> Sent: Thursday, July 16, 2026 11:29 AM
> 
> Hi K Prateek,
> 
> Thanks for reviewing this.
> 

[snip]

> 
> As Jason said:
> 
>    With Xen, the Xen hypervisor itself controls CPPC.  dom0
>    (xen_initial_domain()) doesn't need the amd-pstate driver.
> 
> I think we don't need to worry about Xen.
> 
> @Michael, if you have the time, could you test if Hyper-V exposes
> X86_FEATURE_HW_PSTATE (i.e., 0x80000007/edx bit 7)?

I checked a VM running on Hyper-V on an AMD Zen 2 (EPYC 7452),
and another on a Zen 4 (EPYC 9V74) processor. Neither has
X86_FEATURE_HW_PSTATE set. There are other variations that
could affect what a guest on Hyper-V sees, but these spot tests
are sufficient evidence that this flag does what we want for
Hyper-V guests.

Michael