ext3: Fix fsync error handling after filesystem abort.
If filesystem was aborted we will return success
due to (sb->s_flags & MS_RDONLY) which is incorrect and
results in data loss.
In order to handle fs abort correctly we have to check
fs state once we discover that it is in MS_RDONLY state
Test case: http://patchwork.ozlabs.org/patch/244297/
Changes from V1:
- fix spelling
- fix smp_rmb()/debug order
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: Jan Kara <jack@suse.cz>
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index 6356665..c47f147 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -174,6 +174,11 @@
if (test_opt (sb, ERRORS_RO)) {
ext3_msg(sb, KERN_CRIT,
"error: remounting filesystem read-only");
+ /*
+ * Make sure updated value of ->s_mount_state will be visible
+ * before ->s_flags update.
+ */
+ smp_wmb();
sb->s_flags |= MS_RDONLY;
}
ext3_commit_super(sb, es, 1);
@@ -291,8 +296,14 @@
ext3_msg(sb, KERN_CRIT,
"error: remounting filesystem read-only");
EXT3_SB(sb)->s_mount_state |= EXT3_ERROR_FS;
- sb->s_flags |= MS_RDONLY;
set_opt(EXT3_SB(sb)->s_mount_opt, ABORT);
+ /*
+ * Make sure updated value of ->s_mount_state will be visible
+ * before ->s_flags update.
+ */
+ smp_wmb();
+ sb->s_flags |= MS_RDONLY;
+
if (EXT3_SB(sb)->s_journal)
journal_abort(EXT3_SB(sb)->s_journal, -EIO);
}