drop mbcache-fix-potential-double-count-when-removing-entry
[ext4-patch-queue.git] / save-error-to-disk-in-ext4_grp_locked_error
blob4cac03b76a1fccb04bdb2bd1e042a8eb181e6a3f
1 ext4: save error to disk in __ext4_grp_locked_error()
3 From: Zhouyi Zhou <zhouzhouyi@gmail.com>
5 In the function __ext4_grp_locked_error(), __save_error_info()
6 is called to save error info in super block block, but does not sync
7 that information to disk to info the subsequence fsck after reboot.
9 This patch writes the error information to disk.  After this patch,
10 I think there is no obvious EXT4 error handle branches which leads to
11 "Remounting filesystem read-only" will leave the disk partition miss
12 the subsequence fsck.
14 Signed-off-by: Zhouyi Zhou <zhouzhouyi@gmail.com>
15 Signed-off-by: Theodore Ts'o <tytso@mit.edu>
16 ---
17  fs/ext4/super.c | 1 +
18  1 file changed, 1 insertion(+)
20 diff --git a/fs/ext4/super.c b/fs/ext4/super.c
21 index 7c46693..7159438 100644
22 --- a/fs/ext4/super.c
23 +++ b/fs/ext4/super.c
24 @@ -742,6 +742,7 @@ __acquires(bitlock)
25         }
27         ext4_unlock_group(sb, grp);
28 +       ext4_commit_super(sb, 1);
29         ext4_handle_error(sb);
30         /*
31          * We only get here in the ERRORS_RO case; relocking the group
32 -- 
33 2.1.4