Received-SPF: pass (google.com: domain of linux-crypto+bounces-3431-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Message-ID: <73f90bdd2b5767b3004d130031ef042bd28f1698.camel@linux.intel.com>
Subject: Re: [PATCH 2/3] crypto: x86/sha256-avx2 - add missing vzeroupper
From: Tim Chen <tim.c.chen@linux.intel.com>
To: Eric Biggers <ebiggers@kernel.org>, linux-crypto@vger.kernel.org
Cc: x86@kernel.org
Date: Tue, 09 Apr 2024 15:38:16 -0700
In-Reply-To: <20240406002610.37202-3-ebiggers@kernel.org>
References: <20240406002610.37202-1-ebiggers@kernel.org>
	 <20240406002610.37202-3-ebiggers@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.44.4 (3.44.4-2.fc36) 
Precedence: bulk
MIME-Version: 1.0

On Fri, 2024-04-05 at 20:26 -0400, Eric Biggers wrote:
> From: Eric Biggers <ebiggers@google.com>
>=20
> Since sha256_transform_rorx() uses ymm registers, execute vzeroupper
> before returning from it.  This is necessary to avoid reducing the
> performance of SSE code.
>=20
> Fixes: d34a460092d8 ("crypto: sha256 - Optimized sha256 x86_64 routine us=
ing AVX2's RORX instructions")
> Signed-off-by: Eric Biggers <ebiggers@google.com>
> ---
>  arch/x86/crypto/sha256-avx2-asm.S | 1 +
>  1 file changed, 1 insertion(+)
>=20
> diff --git a/arch/x86/crypto/sha256-avx2-asm.S b/arch/x86/crypto/sha256-a=
vx2-asm.S
> index 9918212faf91..0ffb072be956 100644
> --- a/arch/x86/crypto/sha256-avx2-asm.S
> +++ b/arch/x86/crypto/sha256-avx2-asm.S
> @@ -714,10 +714,11 @@ SYM_TYPED_FUNC_START(sha256_transform_rorx)
>  	popq	%r15
>  	popq	%r14
>  	popq	%r13
>  	popq	%r12
>  	popq	%rbx
> +	vzeroupper
>  	RET
>  SYM_FUNC_END(sha256_transform_rorx)
> =20
>  .section	.rodata.cst512.K256, "aM", @progbits, 512
>  .align 64

Acked-by: Tim Chen <tim.c.chen@linux.intel.com>