4.4-stable review patch. If anyone has any objections, please let me know.
------------------
From: Eric Biggers <ebiggers@xxxxxxxxxx>
[It's a minimal fix for a bug that was fixed incidentally by a large
refactoring in v4.8.]
In the CTS template, when the input length is <= one block cipher block
(e.g. <= 16 bytes for AES) pass the correct length to the underlying CBC
transform rather than one block. This matches the upstream behavior and
makes the encryption/decryption operation correctly return -EINVAL when
1 <= nbytes < bsize or succeed when nbytes == 0, rather than crashing.
This was fixed upstream incidentally by a large refactoring,
commit 0605c41cc53c ("crypto: cts - Convert to skcipher"). But
syzkaller easily trips over this when running on older kernels, as it's
easily reachable via AF_ALG. Therefore, this patch makes the minimal
fix for older kernels.
Cc: linux-crypto@xxxxxxxxxxxxxxx
Fixes: 76cb9521795a ("[CRYPTO] cts: Add CTS mode required for Kerberos AES support")
Signed-off-by: Eric Biggers <ebiggers@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
crypto/cts.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
--- a/crypto/cts.c
+++ b/crypto/cts.c
@@ -137,8 +137,8 @@ static int crypto_cts_encrypt(struct blk
lcldesc.info = desc->info;
lcldesc.flags = desc->flags;
- if (tot_blocks == 1) {
- err = crypto_blkcipher_encrypt_iv(&lcldesc, dst, src, bsize);
+ if (tot_blocks <= 1) {
+ err = crypto_blkcipher_encrypt_iv(&lcldesc, dst, src, nbytes);
} else if (nbytes <= bsize * 2) {
err = cts_cbc_encrypt(ctx, desc, dst, src, 0, nbytes);
} else {
@@ -232,8 +232,8 @@ static int crypto_cts_decrypt(struct blk
lcldesc.info = desc->info;
lcldesc.flags = desc->flags;
- if (tot_blocks == 1) {
- err = crypto_blkcipher_decrypt_iv(&lcldesc, dst, src, bsize);
+ if (tot_blocks <= 1) {
+ err = crypto_blkcipher_decrypt_iv(&lcldesc, dst, src, nbytes);
} else if (nbytes <= bsize * 2) {
err = cts_cbc_decrypt(ctx, desc, dst, src, 0, nbytes);
} else {
