Received-SPF: pass (google.com: domain of linux-crypto-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Saulo Alessandre <saulo.alessandre@gmail.com>
To:     herbert@gondor.apana.org.au
Cc:     ardb@kernel.org, daniele.alessandrelli@linux.intel.com,
        davem@davemloft.net, elena.reshetova@intel.com,
        linux-crypto@vger.kernel.org, linux-kernel@vger.kernel.org,
        mgross@linux.intel.com, prabhjot.khurana@intel.com,
        wangzhou1@hisilicon.com, xuzaibo@huawei.com, yumeng18@huawei.com,
        saulo.alessandre@gmail.com
Subject: [PATCH v7 4/7] crypto: add ecc curve and expose them
Date:   Wed,  3 Feb 2021 15:03:44 -0300
Message-Id: <20210203180344.2418055-1-saulo.alessandre@gmail.com>
In-Reply-To: <20210128050354.GA30874@gondor.apana.org.au>
References: <20210128050354.GA30874@gondor.apana.org.au>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 28/01/2021 02:03, Herbert Xu wrote:
> On Fri, Jan 22, 2021 at 03:09:52PM +0800, Meng Yu wrote:
>> 1. Add ecc curves(P224, P384, P521) for ECDH;
>
> OK I think this is getting unwieldy.
>
> In light of the fact that we already have hardware that supports
> a specific subset of curves, I think perhaps it would be better
> to move the curve ID from the key into the algorithm name instead.
>
I think I understand you, I'm not using ECDH at the moment, but IMHO maybe
we could use enum OID of oid_registry.h as curve ID and eliminate the 
duplicate ECC_CURVE_NIST_{...} from ecdh.h. Or perhaps put another param
in struct ecc_curve with OID, because the name already exists.

> IOW, instead of allocating ecdh, you would allocate ecdh-nist-pXXX.
>
> Any comments?

I recently sent a patch for the ECDSA signature verification, that use
the NIST-P curves to check elf32 binary modules and signatures in
about 450k T-DRE voting machines, in the Brazilian Elections across
the country. I put the other curves because we started testing them
(P256, P384) for speed measurement, but we ended up using P521 in our
production version since 2017 in the 4.9.xxx kernel, and now in 5.4.xxx.

In this patch I'm using akcipher allocate like ecdsa(sha1,sha256,...), 
because the ecdsa algo is generic, and using the curve name and ndigits
inside vli_mmod_fast to discover the curve, but I agree the correct way
would be allocate ecdsa-nist-p521(sha1,...) and have all params for the
curve inside strut ecc_curve, remembering that we have anothers curves 
incoming, like Edwards.

regards,
--
Email: Saulo Alessandre <saulo.alessandre@gmail.com>