diff mbox series

[SRU,B/X,CVE-2018-10322,v2,4/4] xfs: enhance dinode verifier

Message ID 20200902194139.67480-6-william.gray@canonical.com
State New
Headers show
Series XFS xfs_dinode_verify() DOS | expand

Commit Message

William Breathitt Gray Sept. 2, 2020, 7:41 p.m. UTC
From: Eric Sandeen <sandeen@sandeen.net>

Add several more validations to xfs_dinode_verify:

- For LOCAL data fork formats, di_nextents must be 0.
- For LOCAL attr fork formats, di_anextents must be 0.
- For inodes with no attr fork offset,
  - format must be XFS_DINODE_FMT_EXTENTS if set at all
  - di_anextents must be 0.

Thanks to dchinner for pointing out a couple related checks I had
forgotten to add.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=199377
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

CVE-2018-10322

(backported from commit b42db0860e13067fcc7cbfba3966c9e652668bbc)
[ vilhelmgray: context adjustments ]
Signed-off-by: William Breathitt Gray <william.gray@canonical.com>
---
 fs/xfs/libxfs/xfs_inode_buf.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)
diff mbox series

Patch

diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index 936a72e420b1..a03b9c02d09c 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -349,6 +349,8 @@  xfs_dinode_verify(
 				return false;
 			if (di_size > XFS_DFORK_DSIZE(dip, mp))
 				return false;
+			if (dip->di_nextents)
+				return false;
 			/* fall through */
 		case XFS_DINODE_FMT_EXTENTS:
 		case XFS_DINODE_FMT_BTREE:
@@ -367,12 +369,31 @@  xfs_dinode_verify(
 	if (XFS_DFORK_Q(dip)) {
 		switch (dip->di_aformat) {
 		case XFS_DINODE_FMT_LOCAL:
+			if (dip->di_anextents)
+				return false;
+		/* fall through */
 		case XFS_DINODE_FMT_EXTENTS:
 		case XFS_DINODE_FMT_BTREE:
 			break;
 		default:
 			return false;
 		}
+	} else {
+		/*
+		 * If there is no fork offset, this may be a freshly-made inode
+		 * in a new disk cluster, in which case di_aformat is zeroed.
+		 * Otherwise, such an inode must be in EXTENTS format; this goes
+		 * for freed inodes as well.
+		 */
+		switch (dip->di_aformat) {
+		case 0:
+		case XFS_DINODE_FMT_EXTENTS:
+			break;
+		default:
+			return false;
+		}
+		if (dip->di_anextents)
+			return false;
 	}
 
 	/* only version 3 or greater inodes are extensively verified here */