Re: [PATCH v2 4/6] x86/fpu/xstate: Introduce fpu_guest_cfg for guest FPU configuration

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

On Tue, Mar 04, 2025 at 02:50:48PM -0800, Dave Hansen wrote:
>On 11/26/24 02:17, Chao Gao wrote:
>> From: Yang Weijiang <weijiang.yang@intel.com>
>> 
>> Define new fpu_guest_cfg to hold all guest FPU settings so that it can
>> differ from generic kernel FPU settings, e.g., enabling CET supervisor
>> xstate by default for guest fpstate while it's remained disabled in
>> kernel FPU config.
>> 
>> The kernel dynamic xfeatures are specifically used by guest fpstate now,
>> add the mask for guest fpstate so that guest_perm.__state_perm ==
>> (fpu_kernel_cfg.default_xfeature | XFEATURE_MASK_KERNEL_DYNAMIC). And
>> if guest fpstate is re-allocated to hold user dynamic xfeatures, the
>> resulting permissions are consumed before calculate new guest fpstate.
>
>This kinda restates what the code does, but I don't think it matches the
>code.

Actually, it does (see more below)

>
>> With new guest FPU config added, there're 3 categories of FPU configs in
>> kernel, the usages and key fields are recapped as below.
>
>This changelog is pretty rough. It's got a lot of words but not much
>substance.

Will revise the changelog.

[...]

>
>
>> 
>>   @fpu_guest_cfg.default_size
>>   - size of compacted buffer with 'fpu_guest_cfg.default_features'
>
>This looks like kerneldoc, not changelog. Are you sure you want it _here_?

No, I agree this should be a comment above fpu_guest/kernel/user_cfg.

>
>
>> @@ -829,6 +835,10 @@ void __init fpu__init_system_xstate(unsigned int legacy_size)
>>  	fpu_user_cfg.default_features = fpu_user_cfg.max_features;
>>  	fpu_user_cfg.default_features &= ~XFEATURE_MASK_USER_DYNAMIC;
>>  
>> +	fpu_guest_cfg.max_features = fpu_kernel_cfg.max_features;
>> +	fpu_guest_cfg.default_features = fpu_guest_cfg.max_features;
>> +	fpu_guest_cfg.default_features &= ~XFEATURE_MASK_USER_DYNAMIC;
>
>I thought this was saying above that it was _setting_ dynamic features.
>
>Why _not_ set them by default?

At first glance, I had the same question.

XFEATURE_MASK_*KERNEL*_DYNAMIC is not excluded here, so it is enabled by
default. I believe the confusion arises partly from the order of the
patches. I will reorder the patches as you suggested in patch 5.