[3.4 stable] ext4: protect group inode free counting with group lock
From: Benjamin LaHaise <bcrl@kvack.org>
Date: 2014-02-14 19:20:36
Subsystem:
ext4 file system, filesystems (vfs and infrastructure), the rest · Maintainers:
"Theodore Ts'o", Alexander Viro, Christian Brauner, Linus Torvalds
Hello Ted et al,
Commit 6f2e9f0e7d795214b9cf5a47724a273b705fd113 is missing from the 3.4
stable tree. I came across this during testing, and it's pretty easy to
trigger under heavy load just by writing out 10,000 8MB files to a disk by
way of 8 worker threads. Are there any other significant commits like this
that are missing from the 3.4 stable tree? Please note that Tao's original
patch doesn't apply entirely cleanly; my tested backport is at the bottom
of this email.
-ben
----snip----
[3.4 stable] ext4: protect group inode free counting with group lock
This patch is a straightforward backport of the following commit. During
testing I encounted group descriptor corruption that matches Tao's original
report. My testing consists of writing 10,000 8MB files to disk using 8
worker threads. This workload manages to trigger the free inode count
corruption almost 100% of the time.
commit 6f2e9f0e7d795214b9cf5a47724a273b705fd113
Author: Tao Ma [off-list ref]
Date: Mon May 28 18:20:59 2012 -0400
ext4: protect group inode free counting with group lock
Now when we set the group inode free count, we don't have a proper
group lock so that multiple threads may decrease the inode free
count at the same time. And e2fsck will complain something like:
Free inodes count wrong for group #1 (1, counted=0).
Fix? no
Free inodes count wrong for group #2 (3, counted=0).
Fix? no
Directories count wrong for group #2 (780, counted=779).
Fix? no
Free inodes count wrong for group #3 (2272, counted=2273).
Fix? no
So this patch try to protect it with the ext4_lock_group.
btw, it is found by xfstests test case 269 and the volume is
mkfsed with the parameter
"-O ^resize_inode,^uninit_bg,extent,meta_bg,flex_bg,ext_attr"
and I have run it 100 times and the error in e2fsck doesn't
show up again.
Signed-off-by: Tao Ma [off-list ref]
Signed-off-by: "Theodore Ts'o" [off-list ref]
Signed-off-by: Benjamin LaHaise <bcrl@kvack.org>
diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
index c90ab78..30ac458 100644
--- a/fs/ext4/ialloc.c
+++ b/fs/ext4/ialloc.c@@ -782,6 +782,8 @@ got: ext4_itable_unused_set(sb, gdp, (EXT4_INODES_PER_GROUP(sb) - ino)); up_read(&grp->alloc_sem); + } else { + ext4_lock_group(sb, group); } ext4_free_inodes_set(sb, gdp, ext4_free_inodes_count(sb, gdp) - 1); if (S_ISDIR(mode)) {
@@ -794,8 +796,8 @@ got: } if (EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_GDT_CSUM)) { gdp->bg_checksum = ext4_group_desc_csum(sbi, group, gdp); - ext4_unlock_group(sb, group); } + ext4_unlock_group(sb, group); BUFFER_TRACE(group_desc_bh, "call ext4_handle_dirty_metadata"); err = ext4_handle_dirty_metadata(handle, NULL, group_desc_bh);