On 1/7/26 11:55 AM, Zhao Liu wrote:
> 
> 
>>>> As background, current Zhaoxin CPUs implement several 
>>>> CPUID.(EAX=0xC0000001,
>>>> ECX=0):EDX feature bits that are not yet defined in the Linux kernel, for
>>>> example SM2/SM2_EN, SM3/SM4 and their enable bits, PARALLAX/PARALLAX_EN,
>>>> TM3/TM3_EN, RNG2/RNG2_EN, PHE2/PHE2_EN, and RSA/RSA_EN.
>>>>
>>>> We previously tried to upstream all these extra feature bits in one
>>>> patch(https://lore.kernel.org/all/[email protected]/),
>>>> but the maintainer rejected it because there was no in-tree code using 
>>>> these
>>>> features yet. So our current plan is to add the CPUID bits together with 
>>>> real
>>>> kernel users step by step.
>>>
>>> I see. I think it's enough to document missing CPUIDs in comment.
>>>
>>
>> Would the following comment be acceptable?
>>
>> /*
>>  * missing: SM2/SM2_EN, CCS/CCS_EN, PARALLAX/PARALLAX_EN,
>>  * TM3/TM3_EN, RNG2/RNG2_EN, PHE2/PHE2_EN, RSA/RSA_EN
>>  */
> 
> Yes, look good to me.
> 
>> Do you think I should also include the lore link in the commit message/cover
>> letter for additional context?
> 
> Yes, mentioning the link in commit message is good. More information is
> helpful.
> 
>>> Could we make Shijidadao-Client as a v2 of Shijidadao-Server, and create an
>>> alias for this v2?
>>>
>>> .alias = "Shijidadao-Client"
>>>
>>> Then we could rename Shijidadao-Server to Shijidadao, and its v2 is for
>>> client.
>>>
>>>> This is also aligned with how QEMU models other vendors' 
>>>> micro-architectures
>>>> where client and server products have slightly different feature sets.
>>>
>>> The main use case for CPU models is to easy migration across mixed CPU
>>> clusters [*]. So, IMO, not all products require a model.
>>
>> For the CPU model naming/versioning, my plan is:
>> The current Shijidadao will be equivalent to the old Shijidadao-Client-v2, 
>> drop
>> the old Shijidadao-Client-v1 according to your advice, Shijidadao-v1 will 
>> have
>> the alias Shijidadao-Client, and Shijidadao-v2 will have the alias
>> Shijidadao-Server.
> 
> Migration should have more use cases for the server. Personally, I feel
> using the server version as the base model might be more convenient?
> Anyway, it's up to you. Overall, these are fine for me.

I hadn't considered that server-side migration is a more common use case, thanks
a lot. See you in v3.

Best wishes,
Ewan.



Reply via email to