From: Eric Biggers <ebigg...@google.com>
commit 12455e320e19e9cc7ad97f4ab89c280fe297387c upstream.
The arm64 NEON bit-sliced implementation of AES-CTR fails the improved
skcipher tests because it sometimes produces the wrong ciphertext. The
bug is that the final keystream block isn't returned from the assembly
code when the number of non-final blocks is zero. This can happen if
the input data ends a few bytes after a page boundary. In this case the
last bytes get "encrypted" by XOR'ing them with uninitialized memory.
Fix the assembly code to return the final keystream block when needed.
Fixes: 88a3f582bea9 ("crypto: arm64/aes - don't use IV buffer to return final
keystream block")
Cc: <sta...@vger.kernel.org> # v4.11+
Reviewed-by: Ard Biesheuvel <ard.biesheu...@linaro.org>
Signed-off-by: Eric Biggers <ebigg...@google.com>
Signed-off-by: Herbert Xu <herb...@gondor.apana.org.au>
---
Please apply to 4.14-stable. This resolves conflicts due to
"crypto: arm64/aes-bs - yield NEON after every block of input"
not being present in 4.14, but that has other dependencies.
Tested using the crypto self-tests from v5.1-rc1 backported to 4.14.
"rfc3686(ctr-aes-neonbs)" now passes the tests.
arch/arm64/crypto/aes-neonbs-core.S | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/arch/arm64/crypto/aes-neonbs-core.S
b/arch/arm64/crypto/aes-neonbs-core.S
index ca0472500433..3b18e3e79531 100644
--- a/arch/arm64/crypto/aes-neonbs-core.S
+++ b/arch/arm64/crypto/aes-neonbs-core.S
@@ -940,7 +940,7 @@ CPU_LE( rev x8, x8 )
8: next_ctr v0
cbnz x4, 99b
-0: st1 {v0.16b}, [x5]
+ st1 {v0.16b}, [x5]
ldp x29, x30, [sp], #16
ret
@@ -948,6 +948,9 @@ CPU_LE( rev x8, x8 )
* If we are handling the tail of the input (x6 != NULL), return the
* final keystream block back to the caller.
*/
+0: cbz x6, 8b
+ st1 {v0.16b}, [x6]
+ b 8b
1: cbz x6, 8b
st1 {v1.16b}, [x6]
b 8b
--
2.21.0
