Skip to content

Commit 78193d6

Browse files
committed
Squashfs: check the inode number is not the invalid value of zero
jira LE-1907 cve CVE-2024-26982 Rebuild_History Non-Buildable kernel-5.14.0-427.28.1.el9_4 commit-author Phillip Lougher <phillip@squashfs.org.uk> commit 9253c54 Empty-Commit: Cherry-Pick Conflicts during history rebuild. Will be included in final tarball splat. Ref for failed cherry-pick at: ciq/ciq_backports/kernel-5.14.0-427.28.1.el9_4/9253c54e.failed Syskiller has produced an out of bounds access in fill_meta_index(). That out of bounds access is ultimately caused because the inode has an inode number with the invalid value of zero, which was not checked. The reason this causes the out of bounds access is due to following sequence of events: 1. Fill_meta_index() is called to allocate (via empty_meta_index()) and fill a metadata index. It however suffers a data read error and aborts, invalidating the newly returned empty metadata index. It does this by setting the inode number of the index to zero, which means unused (zero is not a valid inode number). 2. When fill_meta_index() is subsequently called again on another read operation, locate_meta_index() returns the previous index because it matches the inode number of 0. Because this index has been returned it is expected to have been filled, and because it hasn't been, an out of bounds access is performed. This patch adds a sanity check which checks that the inode number is not zero when the inode is created and returns -EINVAL if it is. [phillip@squashfs.org.uk: whitespace fix] Link: https://lkml.kernel.org/r/20240409204723.446925-1-phillip@squashfs.org.uk Link: https://lkml.kernel.org/r/20240408220206.435788-1-phillip@squashfs.org.uk Signed-off-by: Phillip Lougher <phillip@squashfs.org.uk> Reported-by: "Ubisectech Sirius" <bugreport@ubisectech.com> Closes: https://lore.kernel.org/lkml/87f5c007-b8a5-41ae-8b57-431e924c5915.bugreport@ubisectech.com/ Cc: Christian Brauner <brauner@kernel.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> (cherry picked from commit 9253c54) Signed-off-by: Jonathan Maple <jmaple@ciq.com> # Conflicts: # fs/squashfs/inode.c
1 parent 72255bd commit 78193d6

File tree

1 file changed

+70
-0
lines changed

1 file changed

+70
-0
lines changed
Lines changed: 70 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,70 @@
1+
Squashfs: check the inode number is not the invalid value of zero
2+
3+
jira LE-1907
4+
cve CVE-2024-26982
5+
Rebuild_History Non-Buildable kernel-5.14.0-427.28.1.el9_4
6+
commit-author Phillip Lougher <phillip@squashfs.org.uk>
7+
commit 9253c54e01b6505d348afbc02abaa4d9f8a01395
8+
Empty-Commit: Cherry-Pick Conflicts during history rebuild.
9+
Will be included in final tarball splat. Ref for failed cherry-pick at:
10+
ciq/ciq_backports/kernel-5.14.0-427.28.1.el9_4/9253c54e.failed
11+
12+
Syskiller has produced an out of bounds access in fill_meta_index().
13+
14+
That out of bounds access is ultimately caused because the inode
15+
has an inode number with the invalid value of zero, which was not checked.
16+
17+
The reason this causes the out of bounds access is due to following
18+
sequence of events:
19+
20+
1. Fill_meta_index() is called to allocate (via empty_meta_index())
21+
and fill a metadata index. It however suffers a data read error
22+
and aborts, invalidating the newly returned empty metadata index.
23+
It does this by setting the inode number of the index to zero,
24+
which means unused (zero is not a valid inode number).
25+
26+
2. When fill_meta_index() is subsequently called again on another
27+
read operation, locate_meta_index() returns the previous index
28+
because it matches the inode number of 0. Because this index
29+
has been returned it is expected to have been filled, and because
30+
it hasn't been, an out of bounds access is performed.
31+
32+
This patch adds a sanity check which checks that the inode number
33+
is not zero when the inode is created and returns -EINVAL if it is.
34+
35+
[phillip@squashfs.org.uk: whitespace fix]
36+
Link: https://lkml.kernel.org/r/20240409204723.446925-1-phillip@squashfs.org.uk
37+
Link: https://lkml.kernel.org/r/20240408220206.435788-1-phillip@squashfs.org.uk
38+
Signed-off-by: Phillip Lougher <phillip@squashfs.org.uk>
39+
Reported-by: "Ubisectech Sirius" <bugreport@ubisectech.com>
40+
Closes: https://lore.kernel.org/lkml/87f5c007-b8a5-41ae-8b57-431e924c5915.bugreport@ubisectech.com/
41+
Cc: Christian Brauner <brauner@kernel.org>
42+
Cc: <stable@vger.kernel.org>
43+
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
44+
(cherry picked from commit 9253c54e01b6505d348afbc02abaa4d9f8a01395)
45+
Signed-off-by: Jonathan Maple <jmaple@ciq.com>
46+
47+
# Conflicts:
48+
# fs/squashfs/inode.c
49+
diff --cc fs/squashfs/inode.c
50+
index 24463145b351,16bd693d0b3a..000000000000
51+
--- a/fs/squashfs/inode.c
52+
+++ b/fs/squashfs/inode.c
53+
@@@ -58,10 -62,9 +62,16 @@@ static int squashfs_new_inode(struct su
54+
55+
i_uid_write(inode, i_uid);
56+
i_gid_write(inode, i_gid);
57+
++<<<<<<< HEAD
58+
+ inode->i_ino = le32_to_cpu(sqsh_ino->inode_number);
59+
+ inode->i_mtime.tv_sec = le32_to_cpu(sqsh_ino->mtime);
60+
+ inode->i_atime.tv_sec = inode->i_mtime.tv_sec;
61+
+ inode->i_ctime.tv_sec = inode->i_mtime.tv_sec;
62+
++=======
63+
+ inode_set_mtime(inode, le32_to_cpu(sqsh_ino->mtime), 0);
64+
+ inode_set_atime(inode, inode_get_mtime_sec(inode), 0);
65+
+ inode_set_ctime(inode, inode_get_mtime_sec(inode), 0);
66+
++>>>>>>> 9253c54e01b6 (Squashfs: check the inode number is not the invalid value of zero)
67+
inode->i_mode = le16_to_cpu(sqsh_ino->mode);
68+
inode->i_size = 0;
69+
70+
* Unmerged path fs/squashfs/inode.c

0 commit comments

Comments
 (0)