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:   Wed, 14 Mar 2018 18:51:39 -0700
From:   Eric Biggers <ebiggers3@gmail.com>
To:     Tycho Andersen <tycho@tycho.ws>
Cc:     David Howells <dhowells@redhat.com>, keyrings@vger.kernel.org,
        linux-security-module@vger.kernel.org,
        linux-kernel@vger.kernel.org, kernel-hardening@lists.openwall.com,
        James Morris <jmorris@namei.org>,
        "Serge E. Hallyn" <serge@hallyn.com>,
        "Jason A . Donenfeld" <Jason@zx2c4.com>
Subject: Re: [PATCH 1/2] big key: get rid of stack array allocation
Message-ID: <20180315015139.GA641@zzz.localdomain>
References: <20180313042907.29598-1-tycho@tycho.ws>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20180313042907.29598-1-tycho@tycho.ws>
User-Agent: Mutt/1.9.4 (2018-02-28)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Mar 12, 2018 at 10:29:06PM -0600, Tycho Andersen wrote:
> We're interested in getting rid of all of the stack allocated arrays in the
> kernel [1]. This patch removes one in keys by switching to malloc/free.
> Note that we use kzalloc, to avoid leaking the nonce. I'm not sure this is
> really necessary, but extra paranoia seems prudent.
> 
> Manually tested using the program from the add_key man page to trigger
> big_key.
> 
> [1]: https://lkml.org/lkml/2018/3/7/621
> 
> Signed-off-by: Tycho Andersen <tycho@tycho.ws>
> CC: David Howells <dhowells@redhat.com>
> CC: James Morris <jmorris@namei.org>
> CC: "Serge E. Hallyn" <serge@hallyn.com>
> CC: Jason A. Donenfeld <Jason@zx2c4.com>
> ---
>  security/keys/big_key.c | 12 +++++++++---
>  1 file changed, 9 insertions(+), 3 deletions(-)
> 
> diff --git a/security/keys/big_key.c b/security/keys/big_key.c
> index fa728f662a6f..70f9f785c59d 100644
> --- a/security/keys/big_key.c
> +++ b/security/keys/big_key.c
> @@ -108,13 +108,18 @@ static int big_key_crypt(enum big_key_op op, struct big_key_buf *buf, size_t dat
>  	 * an .update function, so there's no chance we'll wind up reusing the
>  	 * key to encrypt updated data. Simply put: one key, one encryption.
>  	 */
> -	u8 zero_nonce[crypto_aead_ivsize(big_key_aead)];
> +	u8 *zero_nonce;
> +
> +	zero_nonce = kzalloc(crypto_aead_ivsize(big_key_aead), GFP_KERNEL);
> +	if (!zero_nonce)
> +		return -ENOMEM;
>  
>  	aead_req = aead_request_alloc(big_key_aead, GFP_KERNEL);
> -	if (!aead_req)
> +	if (!aead_req) {
> +		kfree(zero_nonce);
>  		return -ENOMEM;
> +	}
>  
> -	memset(zero_nonce, 0, sizeof(zero_nonce));
>  	aead_request_set_crypt(aead_req, buf->sg, buf->sg, datalen, zero_nonce);
>  	aead_request_set_callback(aead_req, CRYPTO_TFM_REQ_MAY_SLEEP, NULL, NULL);
>  	aead_request_set_ad(aead_req, 0);
> @@ -131,6 +136,7 @@ static int big_key_crypt(enum big_key_op op, struct big_key_buf *buf, size_t dat
>  error:
>  	mutex_unlock(&big_key_aead_lock);
>  	aead_request_free(aead_req);
> +	kzfree(zero_nonce);
>  	return ret;

A dynamic allocation here doesn't make sense -- the algorithm is hard-coded to
AES-GCM, so the IV size is fixed.  You should just include <crypto/gcm.h> and
use GCM_AES_IV_LEN.  As a sanity check you can add
'BUG_ON(crypto_aead_ivsize(big_key_aead) != GCM_AES_IV_LEN' to big_key_init().

kzfree() also doesn't make sense since the nonce is not secret information.

Thanks,

Eric