diff mbox series

crypto: vmx/xts - reject inputs that are too short

Message ID 20200108050646.29220-1-dja@axtens.net (mailing list archive)
State Accepted
Delegated to: Herbert Xu
Headers show
Series crypto: vmx/xts - reject inputs that are too short | expand

Commit Message

Daniel Axtens Jan. 8, 2020, 5:06 a.m. UTC
When the kernel XTS implementation was extended to deal with ciphertext
stealing in commit 8083b1bf8163 ("crypto: xts - add support for ciphertext
stealing"), a check was added to reject inputs that were too short.

However, in the vmx enablement - commit 239668419349 ("crypto: vmx/xts -
use fallback for ciphertext stealing"), that check wasn't added to the
vmx implementation. This disparity leads to errors like the following:

alg: skcipher: p8_aes_xts encryption unexpectedly succeeded on test vector "random: len=0 klen=64"; expected_error=-22, cfg="random: inplace may_sleep use_finup src_divs=[<flush>66.99%@+10, 33.1%@alignmask+1155]"

Return -EINVAL if asked to operate with a cryptlen smaller than the AES
block size. This brings vmx in line with the generic implementation.

Reported-by: Erhard Furtner <erhard_f@mailbox.org>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=206049
Fixes: 239668419349 ("crypto: vmx/xts - use fallback for ciphertext stealing")
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: stable@vger.kernel.org # v5.4+
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
[dja: commit message]
Signed-off-by: Daniel Axtens <dja@axtens.net>
---
 drivers/crypto/vmx/aes_xts.c | 3 +++
 1 file changed, 3 insertions(+)

Comments

Herbert Xu Jan. 16, 2020, 7:28 a.m. UTC | #1
On Wed, Jan 08, 2020 at 04:06:46PM +1100, Daniel Axtens wrote:
> When the kernel XTS implementation was extended to deal with ciphertext
> stealing in commit 8083b1bf8163 ("crypto: xts - add support for ciphertext
> stealing"), a check was added to reject inputs that were too short.
> 
> However, in the vmx enablement - commit 239668419349 ("crypto: vmx/xts -
> use fallback for ciphertext stealing"), that check wasn't added to the
> vmx implementation. This disparity leads to errors like the following:
> 
> alg: skcipher: p8_aes_xts encryption unexpectedly succeeded on test vector "random: len=0 klen=64"; expected_error=-22, cfg="random: inplace may_sleep use_finup src_divs=[<flush>66.99%@+10, 33.1%@alignmask+1155]"
> 
> Return -EINVAL if asked to operate with a cryptlen smaller than the AES
> block size. This brings vmx in line with the generic implementation.
> 
> Reported-by: Erhard Furtner <erhard_f@mailbox.org>
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=206049
> Fixes: 239668419349 ("crypto: vmx/xts - use fallback for ciphertext stealing")
> Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> Cc: stable@vger.kernel.org # v5.4+
> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
> [dja: commit message]
> Signed-off-by: Daniel Axtens <dja@axtens.net>
> ---
>  drivers/crypto/vmx/aes_xts.c | 3 +++
>  1 file changed, 3 insertions(+)

Patch applied.  Thanks.
diff mbox series

Patch

diff --git a/drivers/crypto/vmx/aes_xts.c b/drivers/crypto/vmx/aes_xts.c
index d59e736882f6..9fee1b1532a4 100644
--- a/drivers/crypto/vmx/aes_xts.c
+++ b/drivers/crypto/vmx/aes_xts.c
@@ -84,6 +84,9 @@  static int p8_aes_xts_crypt(struct skcipher_request *req, int enc)
 	u8 tweak[AES_BLOCK_SIZE];
 	int ret;
 
+	if (req->cryptlen < AES_BLOCK_SIZE)
+		return -EINVAL;
+
 	if (!crypto_simd_usable() || (req->cryptlen % XTS_BLOCK_SIZE) != 0) {
 		struct skcipher_request *subreq = skcipher_request_ctx(req);