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 for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date:   Sat, 3 Nov 2018 15:07:55 -0700
From:   Eric Biggers <ebiggers@...nel.org>
To:     Corentin Labbe <clabbe@...libre.com>
Cc:     dan.carpenter@...cle.com, davem@...emloft.net,
        herbert@...dor.apana.org.au, linux-crypto@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] crypto: crypto_user_stat: Zeroize whole structure given
 to user space

On Fri, Oct 19, 2018 at 12:27:25PM +0000, Corentin Labbe wrote:
> For preventing un-initilized data to be given to user-space (and so leak
> potential useful data), the crypto_stat structure must be correctly
> initialised.
> 
> Reported-by: Dan Carpenter <dan.carpenter@...cle.com>
> Fixes: cac5818c25d0 ("crypto: user - Implement a generic crypto statistics")
> Signed-off-by: Corentin Labbe <clabbe@...libre.com>
> ---
>  crypto/crypto_user_stat.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
> 
> diff --git a/crypto/crypto_user_stat.c b/crypto/crypto_user_stat.c
> index 021ad06bbb62..9990c599f648 100644
> --- a/crypto/crypto_user_stat.c
> +++ b/crypto/crypto_user_stat.c
> @@ -37,6 +37,7 @@ static int crypto_report_aead(struct sk_buff *skb, struct crypto_alg *alg)
>  	u64 v64;
>  	u32 v32;
>  
> +	memset(&raead, 0, sizeof(struct crypto_stat));
>  	strncpy(raead.type, "aead", sizeof(raead.type));
>  
>  	v32 = atomic_read(&alg->encrypt_cnt);
> @@ -65,6 +66,7 @@ static int crypto_report_cipher(struct sk_buff *skb, struct crypto_alg *alg)
>  	u64 v64;
>  	u32 v32;
>  
> +	memset(&rcipher, 0, sizeof(struct crypto_stat));
>  	strlcpy(rcipher.type, "cipher", sizeof(rcipher.type));
>  
>  	v32 = atomic_read(&alg->encrypt_cnt);
> @@ -93,6 +95,7 @@ static int crypto_report_comp(struct sk_buff *skb, struct crypto_alg *alg)
>  	u64 v64;
>  	u32 v32;
>  
> +	memset(&rcomp, 0, sizeof(struct crypto_stat));
>  	strlcpy(rcomp.type, "compression", sizeof(rcomp.type));
>  	v32 = atomic_read(&alg->compress_cnt);
>  	rcomp.stat_compress_cnt = v32;
> @@ -120,6 +123,7 @@ static int crypto_report_acomp(struct sk_buff *skb, struct crypto_alg *alg)
>  	u64 v64;
>  	u32 v32;
>  
> +	memset(&racomp, 0, sizeof(struct crypto_stat));
>  	strlcpy(racomp.type, "acomp", sizeof(racomp.type));
>  	v32 = atomic_read(&alg->compress_cnt);
>  	racomp.stat_compress_cnt = v32;
> @@ -147,6 +151,7 @@ static int crypto_report_akcipher(struct sk_buff *skb, struct crypto_alg *alg)
>  	u64 v64;
>  	u32 v32;
>  
> +	memset(&rakcipher, 0, sizeof(struct crypto_stat));
>  	strncpy(rakcipher.type, "akcipher", sizeof(rakcipher.type));
>  	v32 = atomic_read(&alg->encrypt_cnt);
>  	rakcipher.stat_encrypt_cnt = v32;
> @@ -177,6 +182,7 @@ static int crypto_report_kpp(struct sk_buff *skb, struct crypto_alg *alg)
>  	struct crypto_stat rkpp;
>  	u32 v;
>  
> +	memset(&rkpp, 0, sizeof(struct crypto_stat));
>  	strlcpy(rkpp.type, "kpp", sizeof(rkpp.type));
>  
>  	v = atomic_read(&alg->setsecret_cnt);
> @@ -203,6 +209,7 @@ static int crypto_report_ahash(struct sk_buff *skb, struct crypto_alg *alg)
>  	u64 v64;
>  	u32 v32;
>  
> +	memset(&rhash, 0, sizeof(struct crypto_stat));
>  	strncpy(rhash.type, "ahash", sizeof(rhash.type));
>  
>  	v32 = atomic_read(&alg->hash_cnt);
> @@ -227,6 +234,7 @@ static int crypto_report_shash(struct sk_buff *skb, struct crypto_alg *alg)
>  	u64 v64;
>  	u32 v32;
>  
> +	memset(&rhash, 0, sizeof(struct crypto_stat));
>  	strncpy(rhash.type, "shash", sizeof(rhash.type));
>  
>  	v32 = atomic_read(&alg->hash_cnt);
> @@ -251,6 +259,7 @@ static int crypto_report_rng(struct sk_buff *skb, struct crypto_alg *alg)
>  	u64 v64;
>  	u32 v32;
>  
> +	memset(&rrng, 0, sizeof(struct crypto_stat));
>  	strncpy(rrng.type, "rng", sizeof(rrng.type));
>  
>  	v32 = atomic_read(&alg->generate_cnt);
> -- 
> 2.18.1
> 

There are two more info leaks in the file that this doesn't fix.  I sent out a
fixed version here:

	https://patchwork.kernel.org/patch/10666775/

... as part of the series "crypto: crypto_user reporting fixes and cleanups".

- Eric

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ