ext4: fix bug_on in start_this_handle during umount filesystem

We got issue as follows:
------------[ cut here ]------------
kernel BUG at fs/jbd2/transaction.c:389!
invalid opcode: 0000 [#1] PREEMPT SMP KASAN PTI
CPU: 9 PID: 131 Comm: kworker/9:1 Not tainted 5.17.0-862.14.0.6.x86_64-00001-g23f87daf7d74-dirty #197
Workqueue: events flush_stashed_error_work
RIP: 0010:start_this_handle+0x41c/0x1160
RSP: 0018:ffff888106b47c20 EFLAGS: 00010202
RAX: ffffed10251b8400 RBX: ffff888128dc204c RCX: ffffffffb52972ac
RDX: 0000000000000200 RSI: 0000000000000004 RDI: ffff888128dc2050
RBP: 0000000000000039 R08: 0000000000000001 R09: ffffed10251b840a
R10: ffff888128dc204f R11: ffffed10251b8409 R12: ffff888116d78000
R13: 0000000000000000 R14: dffffc0000000000 R15: ffff888128dc2000
FS:  0000000000000000(0000) GS:ffff88839d680000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000001620068 CR3: 0000000376c0e000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 <TASK>
 jbd2__journal_start+0x38a/0x790
 jbd2_journal_start+0x19/0x20
 flush_stashed_error_work+0x110/0x2b3
 process_one_work+0x688/0x1080
 worker_thread+0x8b/0xc50
 kthread+0x26f/0x310
 ret_from_fork+0x22/0x30
 </TASK>
Modules linked in:
---[ end trace 0000000000000000 ]---

Above issue may happen as follows:
      umount            read procfs            error_work
ext4_put_super
  flush_work(&sbi->s_error_work);

                      ext4_mb_seq_groups_show
	                ext4_mb_load_buddy_gfp
			  ext4_mb_init_group
			    ext4_mb_init_cache
	                      ext4_read_block_bitmap_nowait
			        ext4_validate_block_bitmap
				  ext4_error
			            ext4_handle_error
			              schedule_work(&EXT4_SB(sb)->s_error_work);

  ext4_unregister_sysfs(sb);
  jbd2_journal_destroy(sbi->s_journal);
    journal_kill_thread
      journal->j_flags |= JBD2_UNMOUNT;

                                          flush_stashed_error_work
				            jbd2_journal_start
					      start_this_handle
					        BUG_ON(journal->j_flags & JBD2_UNMOUNT);

To solve this issue, we call 'ext4_unregister_sysfs() before flushing
s_error_work in ext4_put_super().

Signed-off-by: Ye Bin <yebin10@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Ritesh Harjani <riteshh@linux.ibm.com>
Link: https://lore.kernel.org/r/20220322012419.725457-1-yebin10@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>

diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 4120a743be521831b2d7315af64dde011fc9e9a3..f2a5e78f93a9809caea5a047782cd5d8518962dd 100644 (file)
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -1199,20 +1199,25 @@ static void ext4_put_super(struct super_block *sb)
        int aborted = 0;
        int i, err;
 
-       ext4_unregister_li_request(sb);
-       ext4_quota_off_umount(sb);
-
-       flush_work(&sbi->s_error_work);
-       destroy_workqueue(sbi->rsv_conversion_wq);
-       ext4_release_orphan_info(sb);
-
        /*
         * Unregister sysfs before destroying jbd2 journal.
         * Since we could still access attr_journal_task attribute via sysfs
         * path which could have sbi->s_journal->j_task as NULL
+        * Unregister sysfs before flush sbi->s_error_work.
+        * Since user may read /proc/fs/ext4/xx/mb_groups during umount, If
+        * read metadata verify failed then will queue error work.
+        * flush_stashed_error_work will call start_this_handle may trigger
+        * BUG_ON.
         */
        ext4_unregister_sysfs(sb);
 
+       ext4_unregister_li_request(sb);
+       ext4_quota_off_umount(sb);
+
+       flush_work(&sbi->s_error_work);
+       destroy_workqueue(sbi->rsv_conversion_wq);
+       ext4_release_orphan_info(sb);
+
        if (sbi->s_journal) {
                aborted = is_journal_aborted(sbi->s_journal);
                err = jbd2_journal_destroy(sbi->s_journal);