Re: [PATCH v3 09/10] x86/fpu/xstate: Introduce XFEATURE_MASK_KERNEL_DYNAMIC xfeature set

[Chao Gao]

On Mon, Mar 10, 2025 at 01:53:09PM +0800, Chao Gao wrote:
>On Sun, Mar 09, 2025 at 10:20:47PM -0700, Chang S. Bae wrote:
>>On 3/9/2025 8:49 PM, Chao Gao wrote:
>>> 
>>> It was suggested by Sean [1].
>>...
>>> [1]: https://lore.kernel.org/kvm/ZTf5wPKXuHBQk0AN@xxxxxxxxxx/
>>
>>But, you're defining a kernel "dynamic" feature while introducing a
>>"guest-only" xfeature concept. Both seem to be mixed together with this
>>patch. Why not call it as a guest-only feature? That's what Sean was
>>suggesting, no?
>
>Yes. I agree that we should call it as a guest-only feature. That's also why I
>included a note in this patch below the "---" to seek feedback on the naming:
>
>	I am tempted to rename XFEATURE_MASK_KERNEL_DYNAMIC to
>	XFEATURE_MASK_GUEST_ONLY. But I am not sure if this was discussed
>	and rejected.
>
>Thanks for confirming that the renaming is necessary.

Hi Chang,

I dug through the history and found a discussion about the naming at:

https://lore.kernel.org/all/893ac578-baaf-4f4f-96ee-e012dfc073a8@xxxxxxxxx/#t

I think I should revise the changelog to call out why 'DYNAMIC' is preferred
over 'GUEST' and reference that discussion.