Re: [patch V2 15/52] x86/fpu: Use copy_uabi_xstate_to_membuf() in xfpregs_get()

[Borislav Petkov]

On Mon, Jun 14, 2021 at 05:44:23PM +0200, Thomas Gleixner wrote:
> Use the new functionality of copy_uabi_xstate_to_membuf() to retrieve the
> FX state when XSAVE* is in use. This avoids to overwrite the FPU state

					avoids overwriting...

> buffer with fpstate_sanitize_xstate() which is error prone and duplicated
> code.
> 
> Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> ---
> V2: New patch
> ---
>  arch/x86/kernel/fpu/regset.c |   11 ++++++++---
>  1 file changed, 8 insertions(+), 3 deletions(-)
> 
> --- a/arch/x86/kernel/fpu/regset.c
> +++ b/arch/x86/kernel/fpu/regset.c
> @@ -33,13 +33,18 @@ int xfpregs_get(struct task_struct *targ

So AFAICT, this thing is called by PTRACE_GETFPREGS but looking at ltp:

$ git grep PTRACE_GETFPREGS
$

so this is used - if at all used - by some super duper old binaries
somewhere.

manpage says "PTRACE_GETREGS and PTRACE_GETFPREGS are not present on all
architectures." which could explain why. I wonder if we should add some
stupid test cases so that we can at least exercise this...

>  	struct fpu *fpu = &target->thread.fpu;
>  
> -	if (!boot_cpu_has(X86_FEATURE_FXSR))
> +	if (!static_cpu_has(X86_FEATURE_FXSR))

cpu_feature_enabled

>  		return -ENODEV;
>  
>  	fpu__prepare_read(fpu);
> -	fpstate_sanitize_xstate(fpu);
>  
> -	return membuf_write(&to, &fpu->state.fxsave, sizeof(struct fxregs_state));
> +	if (!use_xsave()) {
> +		return membuf_write(&to, &fpu->state.fxsave,
> +				    sizeof(fpu->state.fxsave));
> +	}
> +
> +	copy_uabi_xstate_to_membuf(to, &fpu->state.xsave, XSTATE_COPY_FX);
> +	return 0;

With the above nitpicks addressed:

Reviewed-by: Borislav Petkov <bp@xxxxxxx>

Thx.

-- 
Regards/Gruss,
    Boris.

SUSE Software Solutions Germany GmbH, GF: Felix Imendörffer, HRB 36809, AG Nürnberg