All of lore.kernel.org
 help / color / mirror / Atom feed
From: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
To: Herbert Xu <herbert@gondor.apana.org.au>
Cc: "David S. Miller" <davem@davemloft.net>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>,
	linux-crypto@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org
Subject: Re: [v4 PATCH] crypto: arm64/sm4-gcm - Fix possible crash in GCM cryption
Date: Fri, 3 Feb 2023 10:44:53 +0800	[thread overview]
Message-ID: <f4df6e4f-9d88-bb63-42bf-e2fa23b9a9bc@linux.alibaba.com> (raw)
In-Reply-To: <Y9t1a2HkqlPT+Zfn@gondor.apana.org.au>

Hi Herbert,

On 2/2/23 4:33 PM, Herbert Xu wrote:
> On Wed, Feb 01, 2023 at 08:31:33PM +0800, Tianjia Zhang wrote:
>>
>> +	sm4_ce_pmull_gcm_crypt(ctx->key.rkey_enc, walk->dst.virt.addr,
>> +			       walk->src.virt.addr, iv, walk->nbytes, ghash,
>> +			       ctx->ghash_table, (const u8 *)&lengths);
> 
> I still think this is error-prone.  When walk->nbytes == 0,
> walk->src and walk->dst are undefined.  Sure you could argue
> that the underlying assembly code won't touch the values, but
> accessing uninitialised memory even if just to throw them away
> is still a bit icky.

You're right, whether used or not, accessing an undefined pointer is
always ugly. This benefited me a lot.

> 
> Anyway, here's my attempt at rewriting the gcm loop:
> 
> ---8<---
> An often overlooked aspect of the skcipher walker API is that an
> error is not just indicated by a non-zero return value, but by the
> fact that walk->nbytes is zero.
> 
> Thus it is an error to call skcipher_walk_done after getting back
> walk->nbytes == 0 from the previous interaction with the walker.
> 
> This is because when walk->nbytes is zero the walker is left in
> an undefined state and any further calls to it may try to free
> uninitialised stack memory.
> 
> The sm4 arm64 ccm code gets this wrong and ends up calling
> skcipher_walk_done even when walk->nbytes is zero.
> 
> This patch rewrites the loop in a form that resembles other callers.
> 
> Reported-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
> Fixes: ae1b83c7d572 ("crypto: arm64/sm4 - add CE implementation for GCM mode")
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Thanks for the fix, this patch works find to me, so

Tested-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>

> 
> diff --git a/arch/arm64/crypto/sm4-ce-gcm-glue.c b/arch/arm64/crypto/sm4-ce-gcm-glue.c
> index c450a2025ca9..73bfb6972d3a 100644
> --- a/arch/arm64/crypto/sm4-ce-gcm-glue.c
> +++ b/arch/arm64/crypto/sm4-ce-gcm-glue.c
> @@ -135,22 +135,23 @@ static void gcm_calculate_auth_mac(struct aead_request *req, u8 ghash[])
>   }
>   
>   static int gcm_crypt(struct aead_request *req, struct skcipher_walk *walk,
> -		     struct sm4_gcm_ctx *ctx, u8 ghash[],
> +		     u8 ghash[], int err,
>   		     void (*sm4_ce_pmull_gcm_crypt)(const u32 *rkey_enc,
>   				u8 *dst, const u8 *src, u8 *iv,
>   				unsigned int nbytes, u8 *ghash,
>   				const u8 *ghash_table, const u8 *lengths))
>   {
> +	struct crypto_aead *aead = crypto_aead_reqtfm(req);
> +	struct sm4_gcm_ctx *ctx = crypto_aead_ctx(aead);
>   	u8 __aligned(8) iv[SM4_BLOCK_SIZE];
>   	be128 __aligned(8) lengths;
> -	int err;
>   
>   	memset(ghash, 0, SM4_BLOCK_SIZE);
>   
>   	lengths.a = cpu_to_be64(req->assoclen * 8);
>   	lengths.b = cpu_to_be64(walk->total * 8);
>   
> -	memcpy(iv, walk->iv, GCM_IV_SIZE);
> +	memcpy(iv, req->iv, GCM_IV_SIZE);
>   	put_unaligned_be32(2, iv + GCM_IV_SIZE);
>   
>   	kernel_neon_begin();
> @@ -158,49 +159,51 @@ static int gcm_crypt(struct aead_request *req, struct skcipher_walk *walk,
>   	if (req->assoclen)
>   		gcm_calculate_auth_mac(req, ghash);
>   
> -	do {
> +	while (walk->nbytes) {
>   		unsigned int tail = walk->nbytes % SM4_BLOCK_SIZE;
>   		const u8 *src = walk->src.virt.addr;
>   		u8 *dst = walk->dst.virt.addr;
>   
>   		if (walk->nbytes == walk->total) {
> -			tail = 0;
> -
>   			sm4_ce_pmull_gcm_crypt(ctx->key.rkey_enc, dst, src, iv,
>   					       walk->nbytes, ghash,
>   					       ctx->ghash_table,
>   					       (const u8 *)&lengths);
> -		} else if (walk->nbytes - tail) {
> -			sm4_ce_pmull_gcm_crypt(ctx->key.rkey_enc, dst, src, iv,
> -					       walk->nbytes - tail, ghash,
> -					       ctx->ghash_table, NULL);
> +
> +			kernel_neon_end();
> +
> +			return skcipher_walk_done(walk, 0);
>   		}
>   
> +		sm4_ce_pmull_gcm_crypt(ctx->key.rkey_enc, dst, src, iv,
> +				       walk->nbytes - tail, ghash,
> +				       ctx->ghash_table, NULL);
> +
>   		kernel_neon_end();
>   
>   		err = skcipher_walk_done(walk, tail);
> -		if (err)
> -			return err;
> -		if (walk->nbytes)
> -			kernel_neon_begin();
> -	} while (walk->nbytes > 0);
>   
> -	return 0;
> +		kernel_neon_begin();
> +	}
> +
> +	sm4_ce_pmull_gcm_crypt(ctx->key.rkey_enc, NULL, NULL, iv,
> +			       walk->nbytes, ghash, ctx->ghash_table,
> +			       (const u8 *)&lengths);
> +
> +	kernel_neon_end();
> +
> +	return err;
>   }
>   
>   static int gcm_encrypt(struct aead_request *req)
>   {
>   	struct crypto_aead *aead = crypto_aead_reqtfm(req);
> -	struct sm4_gcm_ctx *ctx = crypto_aead_ctx(aead);
>   	u8 __aligned(8) ghash[SM4_BLOCK_SIZE];
>   	struct skcipher_walk walk;
>   	int err;
>   
>   	err = skcipher_walk_aead_encrypt(&walk, req, false);
> -	if (err)
> -		return err;
> -
> -	err = gcm_crypt(req, &walk, ctx, ghash, sm4_ce_pmull_gcm_enc);
> +	err = gcm_crypt(req, &walk, ghash, err, sm4_ce_pmull_gcm_enc);
>   	if (err)
>   		return err;
>   
> @@ -215,17 +218,13 @@ static int gcm_decrypt(struct aead_request *req)
>   {
>   	struct crypto_aead *aead = crypto_aead_reqtfm(req);
>   	unsigned int authsize = crypto_aead_authsize(aead);
> -	struct sm4_gcm_ctx *ctx = crypto_aead_ctx(aead);
>   	u8 __aligned(8) ghash[SM4_BLOCK_SIZE];
>   	u8 authtag[SM4_BLOCK_SIZE];
>   	struct skcipher_walk walk;
>   	int err;
>   
>   	err = skcipher_walk_aead_decrypt(&walk, req, false);
> -	if (err)
> -		return err;
> -
> -	err = gcm_crypt(req, &walk, ctx, ghash, sm4_ce_pmull_gcm_dec);
> +	err = gcm_crypt(req, &walk, ghash, err, sm4_ce_pmull_gcm_dec);
>   	if (err)
>   		return err;
>   

WARNING: multiple messages have this Message-ID (diff)
From: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
To: Herbert Xu <herbert@gondor.apana.org.au>
Cc: "David S. Miller" <davem@davemloft.net>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>,
	linux-crypto@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org
Subject: Re: [v4 PATCH] crypto: arm64/sm4-gcm - Fix possible crash in GCM cryption
Date: Fri, 3 Feb 2023 10:44:53 +0800	[thread overview]
Message-ID: <f4df6e4f-9d88-bb63-42bf-e2fa23b9a9bc@linux.alibaba.com> (raw)
In-Reply-To: <Y9t1a2HkqlPT+Zfn@gondor.apana.org.au>

Hi Herbert,

On 2/2/23 4:33 PM, Herbert Xu wrote:
> On Wed, Feb 01, 2023 at 08:31:33PM +0800, Tianjia Zhang wrote:
>>
>> +	sm4_ce_pmull_gcm_crypt(ctx->key.rkey_enc, walk->dst.virt.addr,
>> +			       walk->src.virt.addr, iv, walk->nbytes, ghash,
>> +			       ctx->ghash_table, (const u8 *)&lengths);
> 
> I still think this is error-prone.  When walk->nbytes == 0,
> walk->src and walk->dst are undefined.  Sure you could argue
> that the underlying assembly code won't touch the values, but
> accessing uninitialised memory even if just to throw them away
> is still a bit icky.

You're right, whether used or not, accessing an undefined pointer is
always ugly. This benefited me a lot.

> 
> Anyway, here's my attempt at rewriting the gcm loop:
> 
> ---8<---
> An often overlooked aspect of the skcipher walker API is that an
> error is not just indicated by a non-zero return value, but by the
> fact that walk->nbytes is zero.
> 
> Thus it is an error to call skcipher_walk_done after getting back
> walk->nbytes == 0 from the previous interaction with the walker.
> 
> This is because when walk->nbytes is zero the walker is left in
> an undefined state and any further calls to it may try to free
> uninitialised stack memory.
> 
> The sm4 arm64 ccm code gets this wrong and ends up calling
> skcipher_walk_done even when walk->nbytes is zero.
> 
> This patch rewrites the loop in a form that resembles other callers.
> 
> Reported-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
> Fixes: ae1b83c7d572 ("crypto: arm64/sm4 - add CE implementation for GCM mode")
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Thanks for the fix, this patch works find to me, so

Tested-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>

> 
> diff --git a/arch/arm64/crypto/sm4-ce-gcm-glue.c b/arch/arm64/crypto/sm4-ce-gcm-glue.c
> index c450a2025ca9..73bfb6972d3a 100644
> --- a/arch/arm64/crypto/sm4-ce-gcm-glue.c
> +++ b/arch/arm64/crypto/sm4-ce-gcm-glue.c
> @@ -135,22 +135,23 @@ static void gcm_calculate_auth_mac(struct aead_request *req, u8 ghash[])
>   }
>   
>   static int gcm_crypt(struct aead_request *req, struct skcipher_walk *walk,
> -		     struct sm4_gcm_ctx *ctx, u8 ghash[],
> +		     u8 ghash[], int err,
>   		     void (*sm4_ce_pmull_gcm_crypt)(const u32 *rkey_enc,
>   				u8 *dst, const u8 *src, u8 *iv,
>   				unsigned int nbytes, u8 *ghash,
>   				const u8 *ghash_table, const u8 *lengths))
>   {
> +	struct crypto_aead *aead = crypto_aead_reqtfm(req);
> +	struct sm4_gcm_ctx *ctx = crypto_aead_ctx(aead);
>   	u8 __aligned(8) iv[SM4_BLOCK_SIZE];
>   	be128 __aligned(8) lengths;
> -	int err;
>   
>   	memset(ghash, 0, SM4_BLOCK_SIZE);
>   
>   	lengths.a = cpu_to_be64(req->assoclen * 8);
>   	lengths.b = cpu_to_be64(walk->total * 8);
>   
> -	memcpy(iv, walk->iv, GCM_IV_SIZE);
> +	memcpy(iv, req->iv, GCM_IV_SIZE);
>   	put_unaligned_be32(2, iv + GCM_IV_SIZE);
>   
>   	kernel_neon_begin();
> @@ -158,49 +159,51 @@ static int gcm_crypt(struct aead_request *req, struct skcipher_walk *walk,
>   	if (req->assoclen)
>   		gcm_calculate_auth_mac(req, ghash);
>   
> -	do {
> +	while (walk->nbytes) {
>   		unsigned int tail = walk->nbytes % SM4_BLOCK_SIZE;
>   		const u8 *src = walk->src.virt.addr;
>   		u8 *dst = walk->dst.virt.addr;
>   
>   		if (walk->nbytes == walk->total) {
> -			tail = 0;
> -
>   			sm4_ce_pmull_gcm_crypt(ctx->key.rkey_enc, dst, src, iv,
>   					       walk->nbytes, ghash,
>   					       ctx->ghash_table,
>   					       (const u8 *)&lengths);
> -		} else if (walk->nbytes - tail) {
> -			sm4_ce_pmull_gcm_crypt(ctx->key.rkey_enc, dst, src, iv,
> -					       walk->nbytes - tail, ghash,
> -					       ctx->ghash_table, NULL);
> +
> +			kernel_neon_end();
> +
> +			return skcipher_walk_done(walk, 0);
>   		}
>   
> +		sm4_ce_pmull_gcm_crypt(ctx->key.rkey_enc, dst, src, iv,
> +				       walk->nbytes - tail, ghash,
> +				       ctx->ghash_table, NULL);
> +
>   		kernel_neon_end();
>   
>   		err = skcipher_walk_done(walk, tail);
> -		if (err)
> -			return err;
> -		if (walk->nbytes)
> -			kernel_neon_begin();
> -	} while (walk->nbytes > 0);
>   
> -	return 0;
> +		kernel_neon_begin();
> +	}
> +
> +	sm4_ce_pmull_gcm_crypt(ctx->key.rkey_enc, NULL, NULL, iv,
> +			       walk->nbytes, ghash, ctx->ghash_table,
> +			       (const u8 *)&lengths);
> +
> +	kernel_neon_end();
> +
> +	return err;
>   }
>   
>   static int gcm_encrypt(struct aead_request *req)
>   {
>   	struct crypto_aead *aead = crypto_aead_reqtfm(req);
> -	struct sm4_gcm_ctx *ctx = crypto_aead_ctx(aead);
>   	u8 __aligned(8) ghash[SM4_BLOCK_SIZE];
>   	struct skcipher_walk walk;
>   	int err;
>   
>   	err = skcipher_walk_aead_encrypt(&walk, req, false);
> -	if (err)
> -		return err;
> -
> -	err = gcm_crypt(req, &walk, ctx, ghash, sm4_ce_pmull_gcm_enc);
> +	err = gcm_crypt(req, &walk, ghash, err, sm4_ce_pmull_gcm_enc);
>   	if (err)
>   		return err;
>   
> @@ -215,17 +218,13 @@ static int gcm_decrypt(struct aead_request *req)
>   {
>   	struct crypto_aead *aead = crypto_aead_reqtfm(req);
>   	unsigned int authsize = crypto_aead_authsize(aead);
> -	struct sm4_gcm_ctx *ctx = crypto_aead_ctx(aead);
>   	u8 __aligned(8) ghash[SM4_BLOCK_SIZE];
>   	u8 authtag[SM4_BLOCK_SIZE];
>   	struct skcipher_walk walk;
>   	int err;
>   
>   	err = skcipher_walk_aead_decrypt(&walk, req, false);
> -	if (err)
> -		return err;
> -
> -	err = gcm_crypt(req, &walk, ctx, ghash, sm4_ce_pmull_gcm_dec);
> +	err = gcm_crypt(req, &walk, ghash, err, sm4_ce_pmull_gcm_dec);
>   	if (err)
>   		return err;
>   

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

  reply	other threads:[~2023-02-03  2:45 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-02-01 12:31 [PATCH v3] crypto: arm64/sm4-gcm - Fix possible crash in GCM cryption Tianjia Zhang
2023-02-01 12:31 ` Tianjia Zhang
2023-02-02  8:33 ` [v4 PATCH] " Herbert Xu
2023-02-02  8:33   ` Herbert Xu
2023-02-03  2:44   ` Tianjia Zhang [this message]
2023-02-03  2:44     ` Tianjia Zhang

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=f4df6e4f-9d88-bb63-42bf-e2fa23b9a9bc@linux.alibaba.com \
    --to=tianjia.zhang@linux.alibaba.com \
    --cc=catalin.marinas@arm.com \
    --cc=davem@davemloft.net \
    --cc=herbert@gondor.apana.org.au \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-crypto@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=will@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.