* master - btrfs lockdep splat
@ 2017-01-12 18:12 Mike Galbraith
2017-01-12 18:44 ` Liu Bo
0 siblings, 1 reply; 3+ messages in thread
From: Mike Galbraith @ 2017-01-12 18:12 UTC (permalink / raw)
To: LKML; +Cc: linux-btrfs
Greetings,
I wanted to do some -rt testing, but seems non-rt kernels aren't
lockdep clean with btrfs /, making -rt testing a bit premature.
(hm, 28a235931 Btrfs: fix lockdep warning on deadlock against an inode's log mutex)
[ 876.622587] =============================================
[ 876.622588] [ INFO: possible recursive locking detected ]
[ 876.622589] 4.10.0-master #36 Tainted: G E
[ 876.622590] ---------------------------------------------
[ 876.622591] vi/3364 is trying to acquire lock:
[ 876.622592] (&ei->log_mutex){+.+...}, at: [<ffffffffa03d42ec>] btrfs_log_inode+0x13c/0xbd0 [btrfs]
[ 876.622628] but task is already holding lock:
[ 876.622629] (&ei->log_mutex){+.+...}, at: [<ffffffffa03d42ec>] btrfs_log_inode+0x13c/0xbd0 [btrfs]
[ 876.622641] other info that might help us debug this:
[ 876.622642] Possible unsafe locking scenario:
[ 876.622643] CPU0
[ 876.622644] ----
[ 876.622644] lock(&ei->log_mutex);
[ 876.622648] lock(&ei->log_mutex);
[ 876.622649] *** DEADLOCK ***
[ 876.622650] May be due to missing lock nesting notation
[ 876.622651] 3 locks held by vi/3364:
[ 876.622651] #0: (&sb->s_type->i_mutex_key#11){+.+.+.}, at: [<ffffffffa03a3c84>] btrfs_sync_file+0x154/0x480 [btrfs]
[ 876.622664] #1: (sb_internal){.+.+..}, at: [<ffffffffa038b397>] start_transaction+0x2a7/0x540 [btrfs]
[ 876.622674] #2: (&ei->log_mutex){+.+...}, at: [<ffffffffa03d42ec>] btrfs_log_inode+0x13c/0xbd0 [btrfs]
[ 876.622685] stack backtrace:
[ 876.622687] CPU: 3 PID: 3364 Comm: vi Tainted: G E 4.10.0-master #36
[ 876.622688] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.8.1-0-g4adadbd-20161202_174313-build11a 04/01/2014
[ 876.622689] Call Trace:
[ 876.622698] dump_stack+0x85/0xc9
[ 876.622704] __lock_acquire+0x9f9/0x1550
[ 876.622715] ? __btrfs_release_delayed_node+0x79/0x2d0 [btrfs]
[ 876.622717] lock_acquire+0xbd/0x200
[ 876.622726] ? btrfs_log_inode+0x13c/0xbd0 [btrfs]
[ 876.622732] mutex_lock_nested+0x69/0x660
[ 876.622741] ? btrfs_log_inode+0x13c/0xbd0 [btrfs]
[ 876.622750] ? __btrfs_release_delayed_node+0x79/0x2d0 [btrfs]
[ 876.622759] ? btrfs_commit_inode_delayed_inode+0xeb/0x130 [btrfs]
[ 876.622767] btrfs_log_inode+0x13c/0xbd0 [btrfs]
[ 876.622771] ? __might_sleep+0x4a/0x90
[ 876.622781] ? btrfs_i_callback+0x20/0x20 [btrfs]
[ 876.622791] ? free_extent_buffer+0x4b/0x90 [btrfs]
[ 876.622799] btrfs_log_inode+0x572/0xbd0 [btrfs]
[ 876.622808] btrfs_log_inode_parent+0x26a/0x9b0 [btrfs]
[ 876.622812] ? dget_parent+0x77/0x170
[ 876.622821] btrfs_log_dentry_safe+0x62/0x80 [btrfs]
[ 876.622830] btrfs_sync_file+0x2eb/0x480 [btrfs]
[ 876.622834] vfs_fsync_range+0x3d/0xb0
[ 876.622836] ? trace_hardirqs_on_caller+0xf9/0x1c0
[ 876.622837] do_fsync+0x3d/0x70
[ 876.622839] SyS_fsync+0x10/0x20
[ 876.622840] entry_SYSCALL_64_fastpath+0x1f/0xc2
[ 876.622842] RIP: 0033:0x7f7fbe3da290
[ 876.622843] RSP: 002b:00007ffe2778f0b8 EFLAGS: 00000246 ORIG_RAX: 000000000000004a
[ 876.622844] RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 00007f7fbe3da290
[ 876.622845] RDX: 000000000000103d RSI: 000000000143e5d0 RDI: 0000000000000003
[ 876.622846] RBP: 0000000001285f10 R08: 000000000143e5d0 R09: 0000000000000000
[ 876.622847] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
[ 876.622847] R13: 0000000000002000 R14: 0000000000000001 R15: 00000000012821a0
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: master - btrfs lockdep splat
2017-01-12 18:12 master - btrfs lockdep splat Mike Galbraith
@ 2017-01-12 18:44 ` Liu Bo
2017-01-12 19:37 ` Mike Galbraith
0 siblings, 1 reply; 3+ messages in thread
From: Liu Bo @ 2017-01-12 18:44 UTC (permalink / raw)
To: Mike Galbraith; +Cc: LKML, linux-btrfs
On Thu, Jan 12, 2017 at 07:12:12PM +0100, Mike Galbraith wrote:
> Greetings,
>
> I wanted to do some -rt testing, but seems non-rt kernels aren't
> lockdep clean with btrfs /, making -rt testing a bit premature.
>
> (hm, 28a235931 Btrfs: fix lockdep warning on deadlock against an inode's log mutex)
It's rather a false-positive lockdep warning than a real deadlock, and a
patch[1] has been queued to fix it.
[1]: https://patchwork.kernel.org/patch/9473431/
Thanks,
-liubo
>
> [ 876.622587] =============================================
> [ 876.622588] [ INFO: possible recursive locking detected ]
> [ 876.622589] 4.10.0-master #36 Tainted: G E
> [ 876.622590] ---------------------------------------------
> [ 876.622591] vi/3364 is trying to acquire lock:
> [ 876.622592] (&ei->log_mutex){+.+...}, at: [<ffffffffa03d42ec>] btrfs_log_inode+0x13c/0xbd0 [btrfs]
> [ 876.622628] but task is already holding lock:
> [ 876.622629] (&ei->log_mutex){+.+...}, at: [<ffffffffa03d42ec>] btrfs_log_inode+0x13c/0xbd0 [btrfs]
> [ 876.622641] other info that might help us debug this:
> [ 876.622642] Possible unsafe locking scenario:
> [ 876.622643] CPU0
> [ 876.622644] ----
> [ 876.622644] lock(&ei->log_mutex);
> [ 876.622648] lock(&ei->log_mutex);
> [ 876.622649] *** DEADLOCK ***
> [ 876.622650] May be due to missing lock nesting notation
> [ 876.622651] 3 locks held by vi/3364:
> [ 876.622651] #0: (&sb->s_type->i_mutex_key#11){+.+.+.}, at: [<ffffffffa03a3c84>] btrfs_sync_file+0x154/0x480 [btrfs]
> [ 876.622664] #1: (sb_internal){.+.+..}, at: [<ffffffffa038b397>] start_transaction+0x2a7/0x540 [btrfs]
> [ 876.622674] #2: (&ei->log_mutex){+.+...}, at: [<ffffffffa03d42ec>] btrfs_log_inode+0x13c/0xbd0 [btrfs]
> [ 876.622685] stack backtrace:
> [ 876.622687] CPU: 3 PID: 3364 Comm: vi Tainted: G E 4.10.0-master #36
> [ 876.622688] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.8.1-0-g4adadbd-20161202_174313-build11a 04/01/2014
> [ 876.622689] Call Trace:
> [ 876.622698] dump_stack+0x85/0xc9
> [ 876.622704] __lock_acquire+0x9f9/0x1550
> [ 876.622715] ? __btrfs_release_delayed_node+0x79/0x2d0 [btrfs]
> [ 876.622717] lock_acquire+0xbd/0x200
> [ 876.622726] ? btrfs_log_inode+0x13c/0xbd0 [btrfs]
> [ 876.622732] mutex_lock_nested+0x69/0x660
> [ 876.622741] ? btrfs_log_inode+0x13c/0xbd0 [btrfs]
> [ 876.622750] ? __btrfs_release_delayed_node+0x79/0x2d0 [btrfs]
> [ 876.622759] ? btrfs_commit_inode_delayed_inode+0xeb/0x130 [btrfs]
> [ 876.622767] btrfs_log_inode+0x13c/0xbd0 [btrfs]
> [ 876.622771] ? __might_sleep+0x4a/0x90
> [ 876.622781] ? btrfs_i_callback+0x20/0x20 [btrfs]
> [ 876.622791] ? free_extent_buffer+0x4b/0x90 [btrfs]
> [ 876.622799] btrfs_log_inode+0x572/0xbd0 [btrfs]
> [ 876.622808] btrfs_log_inode_parent+0x26a/0x9b0 [btrfs]
> [ 876.622812] ? dget_parent+0x77/0x170
> [ 876.622821] btrfs_log_dentry_safe+0x62/0x80 [btrfs]
> [ 876.622830] btrfs_sync_file+0x2eb/0x480 [btrfs]
> [ 876.622834] vfs_fsync_range+0x3d/0xb0
> [ 876.622836] ? trace_hardirqs_on_caller+0xf9/0x1c0
> [ 876.622837] do_fsync+0x3d/0x70
> [ 876.622839] SyS_fsync+0x10/0x20
> [ 876.622840] entry_SYSCALL_64_fastpath+0x1f/0xc2
> [ 876.622842] RIP: 0033:0x7f7fbe3da290
> [ 876.622843] RSP: 002b:00007ffe2778f0b8 EFLAGS: 00000246 ORIG_RAX: 000000000000004a
> [ 876.622844] RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 00007f7fbe3da290
> [ 876.622845] RDX: 000000000000103d RSI: 000000000143e5d0 RDI: 0000000000000003
> [ 876.622846] RBP: 0000000001285f10 R08: 000000000143e5d0 R09: 0000000000000000
> [ 876.622847] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
> [ 876.622847] R13: 0000000000002000 R14: 0000000000000001 R15: 00000000012821a0
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: master - btrfs lockdep splat
2017-01-12 18:44 ` Liu Bo
@ 2017-01-12 19:37 ` Mike Galbraith
0 siblings, 0 replies; 3+ messages in thread
From: Mike Galbraith @ 2017-01-12 19:37 UTC (permalink / raw)
To: bo.li.liu; +Cc: LKML, linux-btrfs
On Thu, 2017-01-12 at 10:44 -0800, Liu Bo wrote:
> On Thu, Jan 12, 2017 at 07:12:12PM +0100, Mike Galbraith wrote:
> > Greetings,
> >
> > I wanted to do some -rt testing, but seems non-rt kernels aren't
> > lockdep clean with btrfs /, making -rt testing a bit premature.
> >
> > (hm, 28a235931 Btrfs: fix lockdep warning on deadlock against an inode's log mutex)
>
> It's rather a false-positive lockdep warning than a real deadlock, and a
> patch[1] has been queued to fix it.
(yeah, just stops lockdep from perhaps finding something real)
> [1]: https://patchwork.kernel.org/patch/9473431/
Yup, virtual box seems to be a happy camper now. Thanks.
-Mike
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2017-01-12 19:37 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2017-01-12 18:12 master - btrfs lockdep splat Mike Galbraith
2017-01-12 18:44 ` Liu Bo
2017-01-12 19:37 ` Mike Galbraith
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).