FYI, we noticed the below changes on

git://git.kernel.org/pub/scm/linux/kernel/git/josef/btrfs-next.git superblock-scaling
commit 2709a56a0a029b769676cf270b1daa046edbd9cb ("inode: convert per-sb inode list to a list_lru")


+----------------------------------------------------+------------+------------+
|                                                    | f545223a1c | 2709a56a0a |
+----------------------------------------------------+------------+------------+
| boot_successes                                     | 20         | 0          |
| boot_failures                                      | 0          | 15         |
| INFO:possible_circular_locking_dependency_detected | 0          | 15         |
| backtrace:kern_mount_data                          | 0          | 15         |
| backtrace:bdev_cache_init                          | 0          | 15         |
| backtrace:vfs_caches_init                          | 0          | 15         |
+----------------------------------------------------+------------+------------+


wfg: rebooting
[    4.857686] 
[    4.857848] ======================================================
[    4.858377] [ INFO: possible circular locking dependency detected ]
[    4.858885] 4.0.0-rc3-00299-g36e71fc #134 Not tainted
[    4.859322] -------------------------------------------------------
[    4.859806] reboot/137 is trying to acquire lock:
[    4.860009]  (&sb->s_type->i_lock_key#2){+.+...}, at: [<c1187878>] bdev_iter_cb+0x18/0xa0
[    4.860009] 
[    4.860009] but task is already holding lock:
[    4.860009]  (&(&lru->node[i].lock)->rlock){+.+...}, at: [<c1129eec>] __list_lru_walk_one+0x1c/0x110
[    4.860009] 
[    4.860009] which lock already depends on the new lock.
[    4.860009] 
[    4.860009] 
[    4.860009] the existing dependency chain (in reverse order) is:
[    4.860009] 
-> #2 (&(&lru->node[i].lock)->rlock){+.+...}:
[    4.860009]        [<c10bbd7e>] lock_acquire+0x6e/0xa0
[    4.860009]        [<c207e693>] _raw_spin_lock+0x23/0x40
[    4.860009]        [<c1129ff9>] list_lru_add+0x19/0x60
[    4.860009]        [<c116bbf7>] dput+0x2c7/0x350
[    4.860009]        [<c1156c64>] __fput+0x104/0x1b0
[    4.860009]        [<c1156d38>] ____fput+0x8/0x10
[    4.860009]        [<c1094769>] task_work_run+0x99/0xc0
[    4.860009]        [<c1037f28>] do_notify_resume+0x58/0x60
[    4.860009]        [<c207f2aa>] work_notifysig+0x35/0x3b
[    4.860009] 
-> #1 (&(&dentry->d_lockref.lock)->rlock){+.+...}:
[    4.860009]        [<c10bbd7e>] lock_acquire+0x6e/0xa0
[    4.860009]        [<c207e693>] _raw_spin_lock+0x23/0x40
[    4.860009]        [<c1169aff>] __d_instantiate+0x1f/0xe0
[    4.860009]        [<c116c8bb>] d_instantiate+0x2b/0x60
[    4.860009]        [<c117ab11>] mount_pseudo+0xd1/0x140
[    4.860009]        [<c1187239>] bd_mount+0x19/0x20
[    4.860009]        [<c1158836>] mount_fs+0x26/0x150
[    4.860009]        [<c11746c3>] vfs_kern_mount+0x43/0x130
[    4.860009]        [<c1177620>] kern_mount_data+0x10/0x30
[    4.860009]        [<c288068a>] bdev_cache_init+0x49/0x69
[    4.860009]        [<c28801e0>] vfs_caches_init+0x8a/0xe5
[    4.860009]        [<c2864ab6>] start_kernel+0x371/0x3a0
[    4.860009]        [<c28642bb>] i386_start_kernel+0x85/0x89
[    4.860009] 
-> #0 (&sb->s_type->i_lock_key#2){+.+...}:
[    4.860009]        [<c10bb34a>] __lock_acquire+0x1a5a/0x1c90
[    4.860009]        [<c10bbd7e>] lock_acquire+0x6e/0xa0
[    4.860009]        [<c207e693>] _raw_spin_lock+0x23/0x40
[    4.860009]        [<c1187878>] bdev_iter_cb+0x18/0xa0
[    4.860009]        [<c1129f2e>] __list_lru_walk_one+0x5e/0x110
[    4.860009]        [<c112a16e>] list_lru_walk_node+0xe/0x10
[    4.860009]        [<c1189338>] iterate_bdevs+0x38/0x50
[    4.860009]        [<c1180b42>] sys_sync+0x52/0x80
[    4.860009]        [<c207f19b>] syscall_after_call+0x0/0x4
[    4.860009] 
[    4.860009] other info that might help us debug this:
[    4.860009] 
[    4.860009] Chain exists of:
  &sb->s_type->i_lock_key#2 --> &(&dentry->d_lockref.lock)->rlock --> &(&lru->node[i].lock)->rlock

[    4.860009]  Possible unsafe locking scenario:
[    4.860009] 
[    4.860009]        CPU0                    CPU1
[    4.860009]        ----                    ----
[    4.860009]   lock(&(&lru->node[i].lock)->rlock);
[    4.860009]                                lock(&(&dentry->d_lockref.lock)->rlock);
[    4.860009]                                lock(&(&lru->node[i].lock)->rlock);
[    4.860009]   lock(&sb->s_type->i_lock_key#2);
[    4.860009] 
[    4.860009]  *** DEADLOCK ***
[    4.860009] 
[    4.860009] 1 lock held by reboot/137:
[    4.860009]  #0:  (&(&lru->node[i].lock)->rlock){+.+...}, at: [<c1129eec>] __list_lru_walk_one+0x1c/0x110
[    4.860009] 
[    4.860009] stack backtrace:
[    4.860009] CPU: 0 PID: 137 Comm: reboot Not tainted 4.0.0-rc3-00299-g36e71fc #134
[    4.860009] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.7.5-20140531_083030-gandalf 04/01/2014
[    4.860009]  00000000 00000000 cd635e44 c206a841 c2d79720 cd635e74 c10b67c0 c25264e1
[    4.860009]  c252647e c2526461 c2526476 c2526461 c2f77e60 cd656000 cd6564b4 cd656000
[    4.860009]  cd65649c cd635ee0 c10bb34a cd65649c cd656000 00000001 00000000 00000000
[    4.860009] Call Trace:
[    4.860009]  [<c206a841>] dump_stack+0x48/0x60
[    4.860009]  [<c10b67c0>] print_circular_bug+0x190/0x280
[    4.860009]  [<c10bb34a>] __lock_acquire+0x1a5a/0x1c90
[    4.860009]  [<c10bbd7e>] lock_acquire+0x6e/0xa0
[    4.860009]  [<c1187878>] ? bdev_iter_cb+0x18/0xa0
[    4.860009]  [<c207e693>] _raw_spin_lock+0x23/0x40
[    4.860009]  [<c1187878>] ? bdev_iter_cb+0x18/0xa0
[    4.860009]  [<c1187878>] bdev_iter_cb+0x18/0xa0
[    4.860009]  [<c1187860>] ? block_llseek+0xf0/0xf0
[    4.860009]  [<c1129f2e>] __list_lru_walk_one+0x5e/0x110
[    4.860009]  [<c1187860>] ? block_llseek+0xf0/0xf0
[    4.860009]  [<c112a16e>] list_lru_walk_node+0xe/0x10
[    4.860009]  [<c1189338>] iterate_bdevs+0x38/0x50
[    4.860009]  [<c11809d0>] ? fdatawait_one_bdev+0x10/0x10
[    4.860009]  [<c1180b42>] sys_sync+0x52/0x80
[    4.860009]  [<c207f19b>] syscall_call+0x7/0x7
[    4.896543] Unregister pv shared memory for cpu 0
[    4.897080] spin_lock-torture: Unscheduled system shutdown detected
[    4.897605] no ifx modem active;



Thanks,
Ying Huang