This is a note to let you know that I've just added the patch titled
crypto: sha512-mb - initialize pending lengths correctly
to the 4.14-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
crypto-sha512-mb-initialize-pending-lengths-correctly.patch
and it can be found in the queue-4.14 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
>From eff84b379089cd8b4e83599639c1f5f6e34ef7bf Mon Sep 17 00:00:00 2001
From: Eric Biggers <ebiggers@xxxxxxxxxx>
Date: Wed, 24 Jan 2018 00:31:27 -0800
Subject: crypto: sha512-mb - initialize pending lengths correctly
From: Eric Biggers <ebiggers@xxxxxxxxxx>
commit eff84b379089cd8b4e83599639c1f5f6e34ef7bf upstream.
The SHA-512 multibuffer code keeps track of the number of blocks pending
in each lane. The minimum of these values is used to identify the next
lane that will be completed. Unused lanes are set to a large number
(0xFFFFFFFF) so that they don't affect this calculation.
However, it was forgotten to set the lengths to this value in the
initial state, where all lanes are unused. As a result it was possible
for sha512_mb_mgr_get_comp_job_avx2() to select an unused lane, causing
a NULL pointer dereference. Specifically this could happen in the case
where ->update() was passed fewer than SHA512_BLOCK_SIZE bytes of data,
so it then called sha_complete_job() without having actually submitted
any blocks to the multi-buffer code. This hit a NULL pointer
dereference if another task happened to have submitted blocks
concurrently to the same CPU and the flush timer had not yet expired.
Fix this by initializing sha512_mb_mgr->lens correctly.
As usual, this bug was found by syzkaller.
Fixes: 45691e2d9b18 ("crypto: sha512-mb - submit/flush routines for AVX2")
Reported-by: syzbot <syzkaller@xxxxxxxxxxxxxxxx>
Signed-off-by: Eric Biggers <ebiggers@xxxxxxxxxx>
Signed-off-by: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
arch/x86/crypto/sha512-mb/sha512_mb_mgr_init_avx2.c | 10 ++++++----
1 file changed, 6 insertions(+), 4 deletions(-)
--- a/arch/x86/crypto/sha512-mb/sha512_mb_mgr_init_avx2.c
+++ b/arch/x86/crypto/sha512-mb/sha512_mb_mgr_init_avx2.c
@@ -57,10 +57,12 @@ void sha512_mb_mgr_init_avx2(struct sha5
{
unsigned int j;
- state->lens[0] = 0;
- state->lens[1] = 1;
- state->lens[2] = 2;
- state->lens[3] = 3;
+ /* initially all lanes are unused */
+ state->lens[0] = 0xFFFFFFFF00000000;
+ state->lens[1] = 0xFFFFFFFF00000001;
+ state->lens[2] = 0xFFFFFFFF00000002;
+ state->lens[3] = 0xFFFFFFFF00000003;
+
state->unused_lanes = 0xFF03020100;
for (j = 0; j < 4; j++)
state->ldata[j].job_in_lane = NULL;
Patches currently in stable-queue which might be from ebiggers@xxxxxxxxxx are
queue-4.14/pipe-fix-off-by-one-error-when-checking-buffer-limits.patch
queue-4.14/crypto-hash-annotate-algorithms-taking-optional-key.patch
queue-4.14/crypto-cryptd-pass-through-absence-of-setkey.patch
queue-4.14/crypto-hash-prevent-using-keyed-hashes-without-setting-key.patch
queue-4.14/ubifs-free-the-encrypted-symlink-target.patch
queue-4.14/pipe-actually-allow-root-to-exceed-the-pipe-buffer-limits.patch
queue-4.14/kernel-relay.c-revert-kernel-relay.c-fix-potential-memory-leak.patch
queue-4.14/nfs-reject-request-for-id_legacy-key-without-auxdata.patch
queue-4.14/crypto-poly1305-remove-setkey-method.patch
queue-4.14/crypto-sha512-mb-initialize-pending-lengths-correctly.patch
queue-4.14/crypto-hash-introduce-crypto_hash_alg_has_setkey.patch
queue-4.14/crypto-mcryptd-pass-through-absence-of-setkey.patch