Re: [PATCH v5 3/7] x86/fpu/xstate: Differentiate default features for host and guest FPUs

[Chao Gao <chao.gao@intel.com>]

On Tue, Apr 29, 2025 at 11:36:40AM +0800, Edgecombe, Rick P wrote:
>On Mon, 2025-04-28 at 20:22 -0700, Chang S. Bae wrote:
>> > 
>> > This patch adds struct vcpu_fpu_config, with new fields user_size,
>> > user_features. Then those fields are used to configure the guest FPU, where
>> > today it just uses fpu_user_cfg.default_features, etc.
>> > 
>> > KVM doesn't refer to any of those fields specifically, but since they are
>> > used
>> > to configure struct fpu_guest they become part of KVM's uABI.
>> 
>> Today, fpu_alloc_guest_fpstate() -> __fpstate_reset() sets
>> vcpu->arch.guest_fpu.fpstate->user_xfeatures using 
>> fpu_user_cfg.default_features.
>> 
>> Are you really saying that switching this to 
>> guest_default_cfg.user_features would constitute a uABI change?
>
>I'm not saying there is a uABI change... I don't see a change in uABI.

Yes. We all agree that this series has no uABI change.

>
>>  Do you 
>> consider fpu_user_cfg.default_features to be part of the uABI or 
>> anything else?
>
>KVM_GET_XSAVE is part of KVM's API. It uses fields configured in struct
>fpu_guest. If fpu_user_cfg.default_features changes value (in the current code)
>it would change KVM's uABI. But I'm starting to suspect we are talking past each
>other.
>
>It should be simple. Two new configuration fields are added in this patch that

Yes. it is a minor issue.

>match the existing concept and values of existing configurations fields. Per
>Sean, there are no plans to have them diverge. So why add them. If anyone feels
>strongly, I won't argue. But I think there is just miscommunication.

Ok. I will drop vcpu_fpu_config.user*.