[PATCH 08/50] libxfs: implement some sanity checking for enormous rgcount

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



From: Darrick J. Wong <djwong@xxxxxxxxxx>

Similar to what we do for suspiciously large sb_agcount values, if
someone tries to get libxfs to load a filesystem with a very large
realtime group count, let's do some basic checks of the rt device to
see if it's really that large.  If the read fails, only load the first
rtgroup and warn the user.

Signed-off-by: "Darrick J. Wong" <djwong@xxxxxxxxxx>
---
 libxfs/init.c |   46 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 46 insertions(+)


diff --git a/libxfs/init.c b/libxfs/init.c
index 6642cd50c00b5f..16291466ac86d3 100644
--- a/libxfs/init.c
+++ b/libxfs/init.c
@@ -654,6 +654,49 @@ xfs_set_low_space_thresholds(
 		mp->m_low_space[i] = dblocks * (i + 1);
 }
 
+/*
+ * libxfs_initialize_rtgroup will allocate a rtgroup structure for each
+ * rtgroup.  If rgcount is corrupted and insanely high, this will OOM the box.
+ * Try to read what would be the last rtgroup superblock.  If that fails, read
+ * the first one and let the user know to check the geometry.
+ */
+static inline bool
+check_many_rtgroups(
+	struct xfs_mount	*mp,
+	struct xfs_sb		*sbp)
+{
+	struct xfs_buf		*bp;
+	xfs_daddr_t		d;
+	int			error;
+
+	if (!mp->m_rtdev->bt_bdev) {
+		fprintf(stderr, _("%s: no rt device, ignoring rgcount %u\n"),
+				progname, sbp->sb_rgcount);
+		if (!xfs_is_debugger(mp))
+			return false;
+
+		sbp->sb_rgcount = 0;
+		return true;
+	}
+
+	d = (xfs_daddr_t)XFS_FSB_TO_BB(mp, mp->m_sb.sb_rblocks);
+	error = libxfs_buf_read(mp->m_rtdev, d - XFS_FSB_TO_BB(mp, 1), 1, 0,
+			&bp, NULL);
+	if (!error) {
+		libxfs_buf_relse(bp);
+		return true;
+	}
+
+	fprintf(stderr, _("%s: read of rtgroup %u failed\n"), progname,
+			sbp->sb_rgcount - 1);
+	if (!xfs_is_debugger(mp))
+		return false;
+
+	fprintf(stderr, _("%s: limiting reads to rtgroup 0\n"), progname);
+	sbp->sb_rgcount = 1;
+	return true;
+}
+
 /*
  * Mount structure initialization, provides a filled-in xfs_mount_t
  * such that the numerous XFS_* macros can be used.  If dev is zero,
@@ -810,6 +853,9 @@ libxfs_mount(
 			libxfs_buf_relse(bp);
 	}
 
+	if (sbp->sb_rgcount > 1000000 && !check_many_rtgroups(mp, sbp))
+		goto out_da;
+
 	error = libxfs_initialize_perag(mp, 0, sbp->sb_agcount,
 			sbp->sb_dblocks, &mp->m_maxagi);
 	if (error) {





[Index of Archives]     [XFS Filesystem Development (older mail)]     [Linux Filesystem Development]     [Linux Audio Users]     [Yosemite Trails]     [Linux Kernel]     [Linux RAID]     [Linux SCSI]


  Powered by Linux