FYI, we noticed the below changes on

https://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git mmap-clear-setid
commit b953933104b23dc80a166b3c2bdc9c8fe077c8a3 ("fs: clear file privilege bits when mmap writing")


+----------------------------------------------------+------------+------------+
|                                                    | 25364a9e54 | b953933104 |
+----------------------------------------------------+------------+------------+
| boot_successes                                     | 30         | 2          |
| boot_failures                                      | 4          | 14         |
| BUG:kernel_early-boot_hang                         | 2          | 2          |
| BUG:kernel_boot_hang                               | 2          |            |
| INFO:possible_circular_locking_dependency_detected | 0          | 12         |
| backtrace:iterate_dir                              | 0          | 12         |
| backtrace:SyS_getdents                             | 0          | 12         |
| backtrace:vm_mmap_pgoff                            | 0          | 12         |
| backtrace:SyS_mmap_pgoff                           | 0          | 12         |
| backtrace:SyS_mmap                                 | 0          | 12         |
+----------------------------------------------------+------------+------------+


[   75.509649] ======================================================
[   75.511059] [ INFO: possible circular locking dependency detected ]
[   75.512313] 4.4.0-rc3-00025-gb953933 #141 Not tainted
[   75.513316] -------------------------------------------------------
[   75.531487] systemd-journal/308 is trying to acquire lock:
[   75.532574]  (&sb->s_type->i_mutex_key#10){+.+.+.}, at: [<ffffffff811d3ccf>] do_mmap+0x213/0x362
[   75.543301] 
[   75.543301] but task is already holding lock:
[   75.544595]  (&mm->mmap_sem){++++++}, at: [<ffffffff811bf304>] vm_mmap_pgoff+0x5d/0xb3
[   75.558997] 
[   75.558997] which lock already depends on the new lock.
[   75.558997] 
[   75.560864] 
[   75.560864] the existing dependency chain (in reverse order) is:
[   75.562354] 
-> #1 (&mm->mmap_sem){++++++}:
[   75.572016]        [<ffffffff81116121>] lock_acquire+0x10a/0x196
[   75.573530]        [<ffffffff811cb662>] __might_fault+0x6a/0x8d
[   75.583071]        [<ffffffff8120ea82>] filldir+0x9d/0x10b
[   75.584465]        [<ffffffff8121f91d>] dir_emit_dots+0x40/0xc4
[   75.585878]        [<ffffffff81220106>] dcache_readdir+0x26/0x14f
[   75.595370]        [<ffffffff8120e958>] iterate_dir+0x90/0x11d
[   75.596745]        [<ffffffff8120ec14>] SyS_getdents+0x97/0xff
[   75.598029]        [<ffffffff82e146f2>] entry_SYSCALL_64_fastpath+0x12/0x76
[   75.607616] 
-> #0 (&sb->s_type->i_mutex_key#10){+.+.+.}:
[   75.609567]        [<ffffffff81115abb>] __lock_acquire+0xb1e/0xdee
[   75.619128]        [<ffffffff81116121>] lock_acquire+0x10a/0x196
[   75.623536]        [<ffffffff82e10be8>] mutex_lock_nested+0x79/0x3ab
[   75.624818]        [<ffffffff811d3ccf>] do_mmap+0x213/0x362
[   75.626059]        [<ffffffff811bf325>] vm_mmap_pgoff+0x7e/0xb3
[   75.631415]        [<ffffffff811d2644>] SyS_mmap_pgoff+0x192/0x1c0
[   75.632676]        [<ffffffff8105b1ca>] SyS_mmap+0x1b/0x25
[   75.633860]        [<ffffffff82e146f2>] entry_SYSCALL_64_fastpath+0x12/0x76
[   75.639318] 
[   75.639318] other info that might help us debug this:
[   75.639318] 
[   75.641060]  Possible unsafe locking scenario:
[   75.641060] 
[   75.642301]        CPU0                    CPU1
[   75.647399]        ----                    ----
[   75.648578]   lock(&mm->mmap_sem);
[   75.649774]                                lock(&sb->s_type->i_mutex_key#10);
[   75.655655]                                lock(&mm->mmap_sem);
[   75.657042]   lock(&sb->s_type->i_mutex_key#10);
[   75.658431] 
[   75.658431]  *** DEADLOCK ***
[   75.658431] 
[   75.664182] 1 lock held by systemd-journal/308:
[   75.665237]  #0:  (&mm->mmap_sem){++++++}, at: [<ffffffff811bf304>] vm_mmap_pgoff+0x5d/0xb3
[   75.671508] 
[   75.671508] stack backtrace:
[   75.672561] CPU: 0 PID: 308 Comm: systemd-journal Not tainted 4.4.0-rc3-00025-gb953933 #141
[   75.674286] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Debian-1.8.2-1 04/01/2014
[   75.680315]  0000000000000000 ffff880028117c30 ffffffff8173d3a8 ffffffff850dd010
[   75.682200]  ffff880028117c70 ffffffff81113f3a ffff880028110ac0 ffff880028111308
[   75.688211]  ffff8800281112d0 0000000000000001 0000000000000001 ffff880028110ac0
[   75.690224] Call Trace:
[   75.694936]  [<ffffffff8173d3a8>] dump_stack+0x4b/0x63
[   75.696122]  [<ffffffff81113f3a>] print_circular_bug+0x1d2/0x1e0
[   75.697429]  [<ffffffff81115abb>] __lock_acquire+0xb1e/0xdee
[   75.698705]  [<ffffffff81116121>] lock_acquire+0x10a/0x196
[   75.703926]  [<ffffffff81116121>] ? lock_acquire+0x10a/0x196
[   75.705186]  [<ffffffff811d3ccf>] ? do_mmap+0x213/0x362
[   75.706359]  [<ffffffff811d3ccf>] ? do_mmap+0x213/0x362
[   75.711511]  [<ffffffff82e10be8>] mutex_lock_nested+0x79/0x3ab
[   75.712615]  [<ffffffff811d3ccf>] ? do_mmap+0x213/0x362
[   75.713654]  [<ffffffff811d3ccf>] do_mmap+0x213/0x362
[   75.714794]  [<ffffffff811d3ccf>] ? do_mmap+0x213/0x362
[   75.720036]  [<ffffffff811bf325>] vm_mmap_pgoff+0x7e/0xb3
[   75.721236]  [<ffffffff811d2644>] SyS_mmap_pgoff+0x192/0x1c0
[   75.722478]  [<ffffffff8105b1ca>] SyS_mmap+0x1b/0x25
[   75.727533]  [<ffffffff82e146f2>] entry_SYSCALL_64_fastpath+0x12/0x76


Thanks,
lkp