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;
MIME-Version: 1.0
References: <20200313110258.94A0668C4E@verein.lst.de> <20200317221743.GD20788@willie-the-truck>
In-Reply-To: <20200317221743.GD20788@willie-the-truck>
From:   Ard Biesheuvel <ardb@kernel.org>
Date:   Wed, 18 Mar 2020 16:13:33 -0400
Message-ID: <CAKv+Gu9_gV0aVwa2QG7jgaR71bTz12vs386R9uPjdQTtm0HcUw@mail.gmail.com>
Subject: Re: [Patch][Fix] crypto: arm{,64} neon: memzero_explicit aes-cbc key
To:     Will Deacon <will@kernel.org>
Cc:     Torsten Duwe <duwe@lst.de>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Herbert Xu <herbert@gondor.apana.org.au>,
        "David S. Miller" <davem@davemloft.net>,
        Russell King <linux@armlinux.org.uk>,
        "open list:HARDWARE RANDOM NUMBER GENERATOR CORE" 
        <linux-crypto@vger.kernel.org>,
        linux-arm-kernel <linux-arm-kernel@lists.infradead.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-crypto-owner@vger.kernel.org
Precedence: bulk

On Tue, 17 Mar 2020 at 18:17, Will Deacon <will@kernel.org> wrote:
>
> [+Ard]
>
> On Fri, Mar 13, 2020 at 12:02:58PM +0100, Torsten Duwe wrote:
> > From: Torsten Duwe <duwe@suse.de>
> >
> > At function exit, do not leave the expanded key in the rk struct
> > which got allocated on the stack.
> >
> > Signed-off-by: Torsten Duwe <duwe@suse.de>
> > ---
> > Another small fix from our FIPS evaluation. I hope you don't mind I merged
> > arm32 and arm64 into one patch -- this is really simple.
> > --- a/arch/arm/crypto/aes-neonbs-glue.c
> > +++ b/arch/arm/crypto/aes-neonbs-glue.c
> > @@ -138,6 +138,7 @@ static int aesbs_cbc_setkey(struct crypto_skcipher *tfm, const u8 *in_key,
> >       kernel_neon_begin();
> >       aesbs_convert_key(ctx->key.rk, rk.key_enc, ctx->key.rounds);
> >       kernel_neon_end();
> > +     memzero_explicit(&rk, sizeof(rk));
> >
> >       return crypto_cipher_setkey(ctx->enc_tfm, in_key, key_len);
> >  }
> > diff --git a/arch/arm64/crypto/aes-neonbs-glue.c b/arch/arm64/crypto/aes-neonbs-glue.c
> > index e3e27349a9fe..c0b980503643 100644
> > --- a/arch/arm64/crypto/aes-neonbs-glue.c
> > +++ b/arch/arm64/crypto/aes-neonbs-glue.c
> > @@ -151,6 +151,7 @@ static int aesbs_cbc_setkey(struct crypto_skcipher *tfm, const u8 *in_key,
> >       kernel_neon_begin();
> >       aesbs_convert_key(ctx->key.rk, rk.key_enc, ctx->key.rounds);
> >       kernel_neon_end();
> > +     memzero_explicit(&rk, sizeof(rk));
> >
> >       return 0;
> >  }
>
> I'm certainly not a crypto person, but this looks sensible to me and I
> couldn't find any other similar stack variable usage under
> arch/arm64/crypto/ at a quick glance.
>
> Acked-by: Will Deacon <will@kernel.org>
>

Acked-by: Ard Biesheuvel <ardb@kernel.org>