On 8/18/26 11:46 AM, Khushit Shah wrote:
>
>> On 14 Aug 2026, at 7:08 PM, Eric Auger <[email protected]> wrote:
>>
>> !-------------------------------------------------------------------|
>> CAUTION: External Email
>>
>> |-------------------------------------------------------------------!
>>
>> Hi Khushit,
>>
>> On 8/5/26 8:47 AM, Khushit Shah wrote:
>>>> On 4 Aug 2026, at 12:15 PM, Markus Armbruster <[email protected]> wrote:
>>>>
>>>> !-------------------------------------------------------------------|
>>>> CAUTION: External Email
>>>>
>>>> |-------------------------------------------------------------------!
>>>>
>>>> Khushit Shah <[email protected]> writes:
>>>>
>>>>> Introduce the 'query-cpu-props-info' QMP command. For CPU model
>>>>> properties it reports the type (boolean or number) and the set of values
>>>>> supported under the active accelerator and host.
>>>>>
>>>>> For numbers the supported values are expressed as a list of inclusive
>>>>> {min, max} ranges; for booleans as the list of allowed true/false
>>>>> values.
>>>>>
>>>>> The command and its return types (CpuPropertyType, CpuPropertyInfo and
>>>>> friends) are defined in qapi/machine.json rather than an Arm-specific
>>>>> schema, since the concept applies to other targets (e.g. x86, riscv)
>>>>> too. Target that do not implement it return an error.
>>>>>
>>>>> The Arm/KVM implementation spins up a scratch "host" vCPU object,
>>>>> walks its QOM properties: SYSREG_ ID-register fields become 'number'
>>>>> properties whose ranges come from arm_field_get_supported_values(),
>>>>> and the advertised feature toggles become 'boolean' properties.
>>>>>
>>>>> Signed-off-by: Khushit Shah <[email protected]>
>>>> This explains what the command does and how it's implemented. It's
>>>> silent on the most important aspect: why do we want the command? What
>>>> is its intended purpose?
>>> Will add something like:
>>>
>>> This helps the management stack to discover which values are supported
>>> for a given CPU property. This is especially important for non-boolean
>>> properties, where knowing that the host has value 'x' does not inherently
>>> reveal the full range of valid values supported on the host.
>> How does x86 handle this at the moment, they don't have the cmd either?
>> How do they currently introspect what are the valid values for a
>> specific host? Is it easier for them?
> x86 is much simpler, all the cpu features properties are backed by cpuid bits
> and assumption is simple, if host value is “true”, “false" is supported, and
> if the host value is “false", “true" is not supported.
OK thanks!
Eric
>
> Warm Regards,
> Khushit
>>
>> Thanks
>>
>> Eric
>>>
>>>