On 10.12.2025 11:26, Teddy Astie wrote:
> Le 10/12/2025 à 09:50, Jan Beulich a écrit :
>> On 09.12.2025 18:19, Teddy Astie wrote:
>>> --- a/tools/misc/xenpm.c
>>> +++ b/tools/misc/xenpm.c
>>> @@ -30,6 +30,7 @@
>>> #include <inttypes.h>
>>> #include <sys/time.h>
>>>
>>> +#include <xen/asm/msr-index.h>
>>
>> For this to not break non-x86 builds, don't you need to constrain the
>> building
>> of the tool to CONFIG_X86? (I have no clue why it is being built for Arm as
>> well right now, as I don't see how it could provide any value there.)
>
> I don't know what are the plans on that area for ARM, the only thing
> that seems supported right now is "get-cpu-topology".
Anthony and/or the Arm maintainers will then need to decide whether the tool
wants to continue to be kept building for non-x86.
>>> + if ( ret == 2 )
>>> + tjmax = (entries[1].val >> 16) & 0xff;
>>> + else
>>> + {
>>> + /*
>>> + * The CPU doesn't support MSR_IA32_TEMPERATURE_TARGET, we assume
>>> it's 100 which
>>> + * is correct aside a few selected Atom CPUs. Check coretemp
>>> source code for more
>>> + * information.
>>> + */
>>
>> What is "coretemp source code" in xen.git context? (I understand you mean the
>> Linux driver, but that also needs saying then.)
>
> Is "Linux kernel's coretemp.c" better ?
Yes.
Jan
