Re: [PATCH 1/3] crypto: Fix the pointer voodoo in unaligned ahash

On Tue, Jan 14, 2014 at 01:35:50PM -0600, Tom Lendacky wrote:

On Tuesday, January 14, 2014 06:33:47 PM Marek Vasut wrote:

quoted

Add documentation for the pointer voodoo that is happening in crypto/ahash.c
in ahash_op_unaligned(). This code is quite confusing, so add a beefy chunk
of documentation.

Moreover, make sure the mangled request is completely restored after finishing
this unaligned operation. This means restoring all of .result, .priv, .base.data
and .base.complete .

Also, remove the crypto_completion_t complete = ... line present in the
ahash_op_unaligned_done() function. This type actually declares a function
pointer, which is very confusing.

Finally, yet very important nonetheless, make sure the req->priv is free()'d
only after the original request is restored in ahash_op_unaligned_done().
The req->priv data must not be free()'d before that in ahash_op_unaligned_finish(),
since we would be accessing previously free()'d data in ahash_op_unaligned_done()
and cause corruption.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: David S. Miller <davem@davemloft.net>
Cc: Fabio Estevam <redacted>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Shawn Guo <redacted>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
---
 crypto/ahash.c | 65 ++++++++++++++++++++++++++++++++++++++++++++++++----------
 1 file changed, 54 insertions(+), 11 deletions(-)

diff --git a/crypto/ahash.c b/crypto/ahash.c
index a92dc38..5ca8ede 100644
--- a/crypto/ahash.c
+++ b/crypto/ahash.c

@@ -29,6 +29,7 @@
 struct ahash_request_priv {
 	crypto_completion_t complete;
 	void *data;
+	void *priv;
 	u8 *result;
 	void *ubuf[] CRYPTO_MINALIGN_ATTR;
 };

@@ -200,23 +201,38 @@ static void ahash_op_unaligned_finish(struct ahash_request *req, int err)
 	if (!err)
 		memcpy(priv->result, req->result,
 		       crypto_ahash_digestsize(crypto_ahash_reqtfm(req)));
-
-	kzfree(priv);

You can't move/remove this kzfree since a synchronous operation will not take
the ahash_op_unaligned_done path.  A synchronous operation will never return
-EINPROGRESS and the effect will be to never free the priv structure.

Marek, did you have a chance to address this?

Thanks,
-- 
Email: Herbert Xu [off-list ref]
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt