[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20241101095955.9786-1-zoo868e@gmail.com>
Date: Fri, 1 Nov 2024 17:59:55 +0800
From: Matt Jan <zoo868e@...il.com>
To: Dave Kleikamp <shaggy@...nel.org>,
jfs-discussion@...ts.sourceforge.net,
linux-kernel@...r.kernel.org,
Shuan Khan <skhan@...uxfoundation.org>
Cc: Matt Jan <zoo868e@...il.com>,
syzbot+9e90a1c5eedb9dc4c6cc@...kaller.appspotmail.com
Subject: [PATCH v4] jfs: UBSAN: shift-out-of-bounds in dbFindBits
Ensure l2nb is less than BUDMIN by performing a sanity check in the caller.
Return -EIO if the check fails.
#syz test
Reported-by: syzbot+9e90a1c5eedb9dc4c6cc@...kaller.appspotmail.com
Signed-off-by: Matt Jan <zoo868e@...il.com>
---
Changes in v4: Thanks to Shaggy for the review. We now perform a sanity check instead of continuing as if nothing is wrong.
Changes in v3: Return the result earlier instead of assert it
Changes in v2: Test if the patch resolve the issue through syzbot and reference the reporter
fs/jfs/jfs_dmap.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/jfs/jfs_dmap.c b/fs/jfs/jfs_dmap.c
index 974ecf5e0d95..89c22a18314f 100644
--- a/fs/jfs/jfs_dmap.c
+++ b/fs/jfs/jfs_dmap.c
@@ -1217,7 +1217,7 @@ dbAllocNear(struct bmap * bmp,
int word, lword, rc;
s8 *leaf;
- if (dp->tree.leafidx != cpu_to_le32(LEAFIND)) {
+ if (dp->tree.leafidx != cpu_to_le32(LEAFIND) || l2nb >= L2DBWORD) {
jfs_error(bmp->db_ipbmap->i_sb, "Corrupt dmap page\n");
return -EIO;
}
@@ -1969,7 +1969,7 @@ dbAllocDmapLev(struct bmap * bmp,
if (dbFindLeaf((dmtree_t *) &dp->tree, l2nb, &leafidx, false))
return -ENOSPC;
- if (leafidx < 0)
+ if (leafidx < 0 || l2nb >= L2DBWORD)
return -EIO;
/* determine the block number within the file system corresponding
--
2.25.1
Powered by blists - more mailing lists