Received-SPF: pass (google.com: best guess record for domain of linux-crypto-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 12 Feb 2020 19:57:49 +0100
From:   Corentin Labbe <clabbe.montjoie@gmail.com>
To:     Herbert Xu <herbert@gondor.apana.org.au>
Cc:     Eric Biggers <ebiggers@kernel.org>, davem@davemloft.net,
        linux-crypto@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [BUG] crypto: export() overran state buffer on test vector
Message-ID: <20200212185749.GA9886@Red>
References: <20200206085442.GA5585@Red>
 <20200207065719.GA8284@sol.localdomain>
 <20200207104659.GA10979@Red>
 <20200208085713.ftuqxhatk6iioz7e@gondor.apana.org.au>
 <20200211192118.GA24059@Red>
 <20200212020628.7grnopgwm6shn3hi@gondor.apana.org.au>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200212020628.7grnopgwm6shn3hi@gondor.apana.org.au>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-crypto-owner@vger.kernel.org
Precedence: bulk

On Wed, Feb 12, 2020 at 10:06:28AM +0800, Herbert Xu wrote:
> On Tue, Feb 11, 2020 at 08:21:18PM +0100, Corentin Labbe wrote:
> > 
> > Do you mean that I should abandon ahash as a fallback ?
> 
> Perhaps switching to shash is not as straightforward because of
> SG handling.  But if you're getting problems with statesize
> perhaps at least force a sync ahash algorithm might be a good
> idea.
> 

I just found the problem, it happen with ARM CE which has a bigger state_size (sha256_ce_state) of 4bytes.

Since my driver didnt use statesize at all, I detect this on cra_init() like this:
if (algt->alg.hash.halg.statesize < crypto_ahash_statesize(op->fallback_tfm))
	algt->alg.hash.halg.statesize = crypto_ahash_statesize(op->fallback_tfm);

Regards