linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] btrfs-progs: Don't return any fs_info that contain NULL tree_root or fs_root.
@ 2014-09-30  2:39 Qu Wenruo
  2014-09-30 11:35 ` David Sterba
  2014-10-06  1:17 ` Qu Wenruo
  0 siblings, 2 replies; 5+ messages in thread
From: Qu Wenruo @ 2014-09-30  2:39 UTC (permalink / raw)
  To: linux-btrfs

[BUG]
btrfsck will segfault if it fails to open the fs tree or tree root.

[REPRODUCER]
Execute btrfsck on a highly damaged btrfs image.
fsfuzz can be used to make a junk btrfs image.

[REASON]
Current open_ctree() in btrfs-progs support OPEN_CTREE_PARTIAL flag to
allow return fs_info even some of the trees is missing.

However it is too loose and even allows fs_info containing no tree to be
returned.

And when it happens, fs_info->fs_root is NULL,
close_ctree(fs_info->fs_root) will cause the access to NULL pointer and
segfault.

[FIX]
This patch will add checks for fs_info->tree_root and fs_info->fs_root
before return fs_info.

Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
---
 disk-io.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/disk-io.c b/disk-io.c
index 26a532e..21a3083 100644
--- a/disk-io.c
+++ b/disk-io.c
@@ -1134,7 +1134,8 @@ static struct btrfs_fs_info *__open_ctree_fd(int fp, const char *path,
 	return fs_info;
 
 out_failed:
-	if (flags & OPEN_CTREE_PARTIAL)
+	if (flags & OPEN_CTREE_PARTIAL &&
+	    fs_info->tree_root && fs_info->fs_root)
 		return fs_info;
 out_chunk:
 	btrfs_release_all_roots(fs_info);
-- 
2.1.1


^ permalink raw reply related	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2014-10-06  1:16 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-09-30  2:39 [PATCH] btrfs-progs: Don't return any fs_info that contain NULL tree_root or fs_root Qu Wenruo
2014-09-30 11:35 ` David Sterba
2014-09-30 15:24   ` Zach Brown
2014-10-06  1:12   ` Qu Wenruo
2014-10-06  1:17 ` Qu Wenruo

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).