Unbalanced unlock with sysrq+SUB

Hi

For some time - when I reboot via 'sysrq + SUB' I get this BUG: message My system is running rawhide (systemd) on ext4 filesystem. systemd-195-7.fc18.x86_64

Zdenek

SysRq : Emergency Remount R/O

=====================================
[ BUG: bad unlock balance detected! ] 3.7.0-rc6-00028-g88e75b6 #101 Not tainted
-------------------------------------
kworker/1:2/79 is trying to release lock (sb_writers) at:
[<ffffffff811b33b4>] mnt_drop_write+0x24/0x30 but there are no more locks to release!

other info that might help us debug this:
4 locks held by kworker/1:2/79:

 #0:  (events){.+.+.+}, at: [<ffffffff81064a79>] process_one_work+0x149/0x750
 #1:  ((work)#3){+.+...}, at: [<ffffffff81064a79>] process_one_work+0x149/0x750
 #2:  (&type->s_umount_key#19){+++++.}, at: [<ffffffff81194140>] 
do_emergency_remount+0x80/0x120
 #3:  (files_lglock){.+.+..}, at: [<ffffffff8119202b>] mark_files_ro+0x2b/0x100

stack backtrace:
Pid: 79, comm: kworker/1:2 Not tainted 3.7.0-rc6-00028-g88e75b6 #101
Call Trace:
 [<ffffffff811b33b4>] ? mnt_drop_write+0x24/0x30
 [<ffffffff810ae52e>] print_unlock_inbalance_bug+0xfe/0x110
 [<ffffffff810b10fe>] lock_release_non_nested+0x21e/0x320
 [<ffffffff8113c9a5>] ? pagevec_lookup_tag+0x25/0x40
 [<ffffffff811b33b4>] ? mnt_drop_write+0x24/0x30
 [<ffffffff810b1297>] lock_release+0x97/0x300
 [<ffffffff811926b1>] __sb_end_write+0x81/0x90
 [<ffffffff811b33b4>] mnt_drop_write+0x24/0x30
 [<ffffffff811b33d2>] mnt_drop_write_file+0x12/0x20
 [<ffffffff811920f7>] mark_files_ro+0xf7/0x100
 [<ffffffff81193f6d>] do_remount_sb+0x13d/0x1b0
 [<ffffffff81194140>] ? do_emergency_remount+0x80/0x120
 [<ffffffff811941dc>] do_emergency_remount+0x11c/0x120
 [<ffffffff81064add>] process_one_work+0x1ad/0x750
 [<ffffffff81064a79>] ? process_one_work+0x149/0x750
 [<ffffffff8106550b>] ? worker_thread+0x21b/0x450
 [<ffffffff81557219>] ? _raw_spin_lock_irq+0x19/0x80
 [<ffffffff811940c0>] ? mount_single+0xe0/0xe0
 [<ffffffff8106544d>] worker_thread+0x15d/0x450
 [<ffffffff810652f0>] ? rescuer_thread+0x230/0x230
 [<ffffffff8106f50b>] kthread+0xdb/0xe0
 [<ffffffff8106f430>] ? kthread_create_on_node+0x140/0x140
 [<ffffffff8155fe1c>] ret_from_fork+0x7c/0xb0
 [<ffffffff8106f430>] ? kthread_create_on_node+0x140/0x140

EXT4-fs (sda2): re-mounted. Opts: (null) EXT4-fs (sda5): re-mounted. Opts: (null) Emergency Remount complete
systemd[1]: Stopping Sendmail Mail Transport Client... systemd[1]: Stopping Sendmail Mail Transport Agent... systemd[1]: Starting Sendmail Mail Transport Agent... systemd[1]: Started Sendmail Mail Transport Agent. systemd[1]: Starting Sendmail Mail Transport Client... systemd[1]: Started Sendmail Mail Transport Client. SysRq : Resetting

Hello,

Thanks for report. Looks like my problem. Attached patch should fix it.

Honza

Jan Kara <[email protected]>
SUSE Labs, CR

From: Jan Kara <[email protected]>
Date: Thu, 22 Nov 2012 23:44:01 +0100
Subject: [PATCH] fs: Fix imbalance in freeze protection in mark_files_ro()

File descriptors (even those for writing) do not hold freeze protection. Thus mark_files_ro() must call __mnt_drop_write() to only drop protection against remount read-only. Calling mnt_drop_write_file() as we do now results in:

[ BUG: bad unlock balance detected! ] 3.7.0-rc6-00028-g88e75b6 #101 Not tainted
-------------------------------------
kworker/1:2/79 is trying to release lock (sb_writers) at:
[<ffffffff811b33b4>] mnt_drop_write+0x24/0x30 but there are no more locks to release!

---
 fs/file_table.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

index a72bf9d..de9e965 100644
@@ -458,8 +458,8 @@ void mark_files_ro(struct super_block *sb)
 		spin_unlock(&f->f_lock);
 		if (file_check_writeable(f) != 0)
 			continue;
+		__mnt_drop_write(f->f_path.mnt);
 		file_release_write(f);
-		mnt_drop_write_file(f);
 	} while_file_list_for_each_entry;
 	lg_global_unlock(&files_lglock);
 }

1.7.1