Received-SPF: pass (google.com: domain of linux-crypto+bounces-4516-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Precedence: bulk
Mime-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset=UTF-8
Date: Thu, 30 May 2024 08:28:01 +0300
Message-Id: <D1MQBJSYUBRS.12KH2S8FUK0XS@kernel.org>
Cc: <linux-kernel@vger.kernel.org>, <lukas@wunner.de>
Subject: Re: [PATCH 1/2] crypto: ecdsa - Use ecc_digits_from_bytes to create
 hash digits array
From: "Jarkko Sakkinen" <jarkko@kernel.org>
To: "Stefan Berger" <stefanb@linux.ibm.com>, <keyrings@vger.kernel.org>,
 <linux-crypto@vger.kernel.org>, <herbert@gondor.apana.org.au>,
 <davem@davemloft.net>
References: <20240529230827.379111-1-stefanb@linux.ibm.com>
 <20240529230827.379111-2-stefanb@linux.ibm.com>
In-Reply-To: <20240529230827.379111-2-stefanb@linux.ibm.com>

On Thu May 30, 2024 at 2:08 AM EEST, Stefan Berger wrote:
> Since ecc_digits_from_bytes will provide zeros when an insufficient numbe=
r
> of bytes are passed in the input byte array, use it to create the hash
> digits directly from the input byte array. This avoids going through an
> intermediate byte array (rawhash) that has the first few bytes filled wit=
h
> zeros.
>
> Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
> ---
>  crypto/ecdsa.c | 17 ++++-------------
>  1 file changed, 4 insertions(+), 13 deletions(-)
>
> diff --git a/crypto/ecdsa.c b/crypto/ecdsa.c
> index 258fffbf623d..fa029f36110b 100644
> --- a/crypto/ecdsa.c
> +++ b/crypto/ecdsa.c
> @@ -142,10 +142,8 @@ static int ecdsa_verify(struct akcipher_request *req=
)
>  	struct ecdsa_signature_ctx sig_ctx =3D {
>  		.curve =3D ctx->curve,
>  	};
> -	u8 rawhash[ECC_MAX_BYTES];
>  	u64 hash[ECC_MAX_DIGITS];
>  	unsigned char *buffer;
> -	ssize_t diff;
>  	int ret;
> =20
>  	if (unlikely(!ctx->pub_key_set))
> @@ -164,18 +162,11 @@ static int ecdsa_verify(struct akcipher_request *re=
q)
>  	if (ret < 0)
>  		goto error;
> =20
> -	/* if the hash is shorter then we will add leading zeros to fit to ndig=
its */
> -	diff =3D bufsize - req->dst_len;
> -	if (diff >=3D 0) {
> -		if (diff)
> -			memset(rawhash, 0, diff);
> -		memcpy(&rawhash[diff], buffer + req->src_len, req->dst_len);
> -	} else if (diff < 0) {
> -		/* given hash is longer, we take the left-most bytes */
> -		memcpy(&rawhash, buffer + req->src_len, bufsize);
> -	}
> +	if (bufsize > req->dst_len)
> +		bufsize =3D req->dst_len;
> =20
> -	ecc_swap_digits((u64 *)rawhash, hash, ctx->curve->g.ndigits);
> +	ecc_digits_from_bytes(buffer + req->src_len, bufsize,
> +			      hash, ctx->curve->g.ndigits);
> =20
>  	ret =3D _ecdsa_verify(ctx, hash, sig_ctx.r, sig_ctx.s);
> =20

Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>

I don't think it'd be even nit-picking to say that the function
called would really need kdoc. I had to spend about 20 minutes
to reacall ecc_digits_from_bytes().

Like something to remind what, how and why... So that you can
recap quickly. Once I got grip of it (for the 2nd time) the
code itself was just fine, no complains on that.

BR, Jarkko