Re: [PATCH v6 1/2] x86/fpu: Allow PKRU to be (once again) written by ptrace.

[Sean Christopherson <seanjc@google.com>]

On Mon, Aug 29, 2022, Kyle Huey wrote:
> @@ -1246,6 +1246,21 @@ static int copy_uabi_to_xstate(struct fpstate *fpstate, const void *kbuf,
>  		}
>  	}
>  
> +	/*
> +	 * Update the user protection key storage. Allow KVM to
> +	 * pass in a NULL pkru pointer if the mask bit is unset
> +	 * for its legacy ABI behavior.
> +	 */
> +	if (pkru)
> +		*pkru = 0;
> +
> +	if (hdr.xfeatures & XFEATURE_MASK_PKRU) {
> +		struct pkru_state *xpkru;
> +
> +		xpkru = __raw_xsave_addr(xsave, XFEATURE_PKRU);
> +		*pkru = xpkru->pkru;
> +	}

What about writing this as:

	if (hdr.xfeatures & XFEATURE_MASK_PKRU) {
		...

		*pkru = xpkru->pkru;
	} else if (pkru) {
		*pkru = 0;
	}

to make it slightly more obvious that @pkru must be non-NULL if the feature flag
is enabled?

Or we could be paranoid, though I'm not sure this is worthwhile.

	if ((hdr.xfeatures & XFEATURE_MASK_PKRU) &&
	    !WARN_ON_ONCE(!pkru)) {
		...

		*pkru = xpkru->pkru;
	} else if (pkru) {
		*pkru = 0;
	}


Otherwise, looks good from a KVM perspective.  Thanks!