Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 29 Nov 2018 16:28:45 +0800
From:   Herbert Xu <herbert@gondor.apana.org.au>
To:     David CARLIER <devnexen@gmail.com>
Cc:     linux-kernel@vger.kernel.org,
        "David S. Miller" <davem@davemloft.net>,
        linux-crypto@vger.kernel.org
Subject: Re: [PATCH] crypto: use memzero_explicit instead of memset to clear
 contexts.
Message-ID: <20181129082845.r3lakg6ckvnkklp7@gondor.apana.org.au>
References: <CA+XhMqz7ZaGVNpkrRM-t4h2f9UZVQPjLVB5Txco4bKGRm8k=mA@mail.gmail.com>
 <20181129064913.wbudlexjzszpa7dq@gondor.apana.org.au>
 <CA+XhMqzw9pqFw30QD=XpFTAqu+t1v5q24ODR+zynZQJJrS9xZw@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CA+XhMqzw9pqFw30QD=XpFTAqu+t1v5q24ODR+zynZQJJrS9xZw@mail.gmail.com>
User-Agent: NeoMutt/20170113 (1.7.2)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Nov 29, 2018 at 06:59:50AM +0000, David CARLIER wrote:
> Meant regardless how the kernel is compiled (ie optimisation level),
> the contexts are guaranteed to be wiped because of the memory fences
> used.

Please avoid top-posting.

> On Thu, 29 Nov 2018 at 06:49, Herbert Xu <herbert@gondor.apana.org.au> wrote:
> >
> > On Wed, Nov 28, 2018 at 07:36:50PM +0000, David CARLIER wrote:
> > > There might be a little performance drop but to make sure it stands
> > > by it comments, we really wipe the whole context after usage.
> > > ---
> > >  crypto/chacha20poly1305.c | 3 ++-
> > >  crypto/md5.c              | 2 +-
> > >  crypto/rmd128.c           | 3 ++-
> > >  crypto/rmd160.c           | 3 ++-
> > >  crypto/rmd256.c           | 3 ++-
> > >  crypto/rmd320.c           | 3 ++-
> > >  crypto/sha3_generic.c     | 3 ++-
> > >  7 files changed, 13 insertions(+), 7 deletions(-)
> > >
> > > diff --git a/crypto/chacha20poly1305.c b/crypto/chacha20poly1305.c
> > > index 600afa99941f..6e93d998109e 100644
> > > --- a/crypto/chacha20poly1305.c
> > > +++ b/crypto/chacha20poly1305.c
> > > @@ -19,6 +19,7 @@
> > >  #include <linux/init.h>
> > >  #include <linux/kernel.h>
> > >  #include <linux/module.h>
> > > +#include <linux/string.h>
> > >
> > >  #include "internal.h"
> > >
> > > @@ -388,7 +389,7 @@ static int poly_genkey(struct aead_request *req)
> > >         }
> > >
> > >         sg_init_table(creq->src, 1);
> > > -       memset(rctx->key, 0, sizeof(rctx->key));
> > > +       memzero_explicit(rctx->key, sizeof(rctx->key));
> >
> > Please explain the purpose of this patch.  As it stands this
> > makes no sense.

memzero_explicit is really only useful for stack memory.  You need
to explain why it makes sense to use it in this context.  For
example, by providing an example of a miscompile or how it is
even possible for this to happen.

Thanks,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt