On 07.05.2025 07:27, Penny, Zheng wrote:
>> -----Original Message-----
>> From: Jan Beulich <jbeul...@suse.com>
>> Sent: Tuesday, April 29, 2025 8:52 PM
>>
>> On 14.04.2025 09:40, Penny Zheng wrote:
>>> @@ -573,6 +576,14 @@ ret_t do_platform_op(
>>> }
>>>
>>> case XEN_PM_CPPC:
>>> + if ( !(xen_processor_pmbits & XEN_PROCESSOR_PM_CPPC) )
>>> + {
>>> + ret = -EOPNOTSUPP;
>>> + break;
>>> + }
>>
>> While at least you no longer use -ENOSYS here, I question this behavior,
>> including
>> that for the pre-existing cases: How is the caller supposed to know whether
>> to
>> invoke this sub-op? Ignoring errors is generally not a good idea, so it
>> would be
>> better if the caller could blindly issue this request, getting back success
>> unless
>> there really was an issue with the data provided.
>>
>
> Understood.
> I will change it to ret = 0. Do you think we shall add warning info here?
No, for precisely ...
> Dom0 will send the CPPC data whenever it could.
... this reason.
Jan
> XEN_PROCESSOR_PM_CPPC
> is not set could largely be users choosing not to. In such case, silently
> getting back success
> shall be enough.
>
>
>> Jan
