linux-kernel - Re: [PATCH] crypto: nx: Fix memcpy() over-reading in nonce

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <202106181648.0C5FA93@keescook>
Date:   Fri, 18 Jun 2021 16:49:34 -0700
From:   Kees Cook <keescook@...omium.org>
To:     Michael Ellerman <mpe@...erman.id.au>
Cc:     Herbert Xu <herbert@...dor.apana.org.au>, stable@...r.kernel.org,
        Breno Leitão <leitao@...ian.org>,
        Nayna Jain <nayna@...ux.ibm.com>,
        Paulo Flabiano Smorigo <pfsmorigo@...il.com>,
        Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Paul Mackerras <paulus@...ba.org>,
        "David S. Miller" <davem@...emloft.net>,
        linux-kernel@...r.kernel.org, linux-crypto@...r.kernel.org,
        linuxppc-dev@...ts.ozlabs.org, linux-hardening@...r.kernel.org
Subject: Re: [PATCH] crypto: nx: Fix memcpy() over-reading in nonce

On Thu, Jun 17, 2021 at 04:08:15PM +1000, Michael Ellerman wrote:
> Kees Cook <keescook@...omium.org> writes:
> > Fix typo in memcpy() where size should be CTR_RFC3686_NONCE_SIZE.
> >
> > Fixes: 030f4e968741 ("crypto: nx - Fix reentrancy bugs")
> > Cc: stable@...r.kernel.org
> > Signed-off-by: Kees Cook <keescook@...omium.org>
> 
> Thanks.
> 
> > ---
> >  drivers/crypto/nx/nx-aes-ctr.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/crypto/nx/nx-aes-ctr.c b/drivers/crypto/nx/nx-aes-ctr.c
> > index 13f518802343..6120e350ff71 100644
> > --- a/drivers/crypto/nx/nx-aes-ctr.c
> > +++ b/drivers/crypto/nx/nx-aes-ctr.c
> > @@ -118,7 +118,7 @@ static int ctr3686_aes_nx_crypt(struct skcipher_request *req)
> >  	struct nx_crypto_ctx *nx_ctx = crypto_skcipher_ctx(tfm);
> >  	u8 iv[16];
> >  
> > -	memcpy(iv, nx_ctx->priv.ctr.nonce, CTR_RFC3686_IV_SIZE);
> > +	memcpy(iv, nx_ctx->priv.ctr.nonce, CTR_RFC3686_NONCE_SIZE);
> >  	memcpy(iv + CTR_RFC3686_NONCE_SIZE, req->iv, CTR_RFC3686_IV_SIZE);
> >  	iv[12] = iv[13] = iv[14] = 0;
> >  	iv[15] = 1;
> 
> Where IV_SIZE is 8 and NONCE_SIZE is 4.
> 
> And iv is 16 bytes, so it's not a buffer overflow.
> 
> But priv.ctr.nonce is 4 bytes, and at the end of the struct, so it reads
> 4 bytes past the end of the nx_crypto_ctx, which is not good.
> 
> But then immediately overwrites whatever it read with req->iv.
> 
> So seems pretty harmless in practice?

Right -- there's no damage done, but future memcpy() FORTIFY work alerts
on this, so I'm going through cleaning all of these up. :)

-- 
Kees Cook