On 11/23/22 4:17 PM, Bart Van Assche wrote:
> Hi,
>
> If I run blktests test block/027 the warning and bug shown below appear.
> It is the first time that I see a complaint like this while running
> blktests. I have not yet tried to root-cause this issue.
>
> root[7931]: run blktests block/027
> [ ... ]
> kernel: ------------[ cut here ]------------
> kernel: ODEBUG: init active (active state 0) object type: work_struct hint: css_release_work_fn+0x0/0x480
> kernel: WARNING: CPU: 38 PID: 498 at lib/debugobjects.c:502 debug_print_object+0xda/0x110
> kernel: Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.0-debian-1.16.0-4 04/01/2014
> kernel: Workqueue: cgwb_release cgwb_release_workfn
> kernel: RIP: 0010:debug_print_object+0xda/0x110
> kernel: Call Trace:
> kernel: <TASK>
> kernel: __debug_object_init+0x217/0x290
> kernel: debug_object_init+0x16/0x20
> kernel: __init_work+0x20/0x30
> kernel: css_release+0x1f/0xb0
> kernel: percpu_ref_put_many.constprop.0+0x150/0x160
> kernel: blkcg_destroy_blkgs+0x20b/0x230
> kernel: blkcg_unpin_online+0x4e/0x90
> kernel: cgwb_release_workfn+0xba/0x210
> kernel: process_one_work+0x57d/0xa80
> kernel: worker_thread+0x90/0x650
> kernel: kthread+0x185/0x1c0
> kernel: ret_from_fork+0x1f/0x30
> kernel: </TASK>
> kernel: irq event stamp: 121765
> kernel: hardirqs last enabled at (121775): [<ffffffff811ae168>] __up_console_sem+0x58/0x60
> kernel: hardirqs last disabled at (121792): [<ffffffff811ae14d>] __up_console_sem+0x3d/0x60
> kernel: softirqs last enabled at (121790): [<ffffffff824004d0>] __do_softirq+0x4d0/0x757
> kernel: softirqs last disabled at (121785): [<ffffffff810e4e61>] __irq_exit_rcu+0xd1/0x140
> kernel: ---[ end trace 0000000000000000 ]---
> kernel: BUG: sleeping function called from invalid context at kernel/workqueue.c:3010
> kernel: in_atomic(): 0, irqs_disabled(): 0, non_block: 0, pid: 498, name: kworker/38:1
> kernel: preempt_count: 0, expected: 0
> kernel: RCU nest depth: 1, expected: 0
> kernel: 3 locks held by kworker/38:1/498:
> kernel: #0: ffff888102c6b538 ((wq_completion)cgwb_release){+.+.}-{0:0}, at: process_one_work+0x479/0xa80
> kernel: #1: ffff88810357fdf0 ((work_completion)(&wb->release_work)){+.+.}-{0:0}, at: process_one_work+0x479/0xa80
> kernel: #2: ffffffff830c52a0 (rcu_read_lock){....}-{1:2}, at: percpu_ref_put_many.constprop.0+0x0/0x160
> kernel: CPU: 38 PID: 498 Comm: kworker/38:1 Tainted: G W E 6.1.0-rc6-dbg #5
> kernel: Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.0-debian-1.16.0-4 04/01/2014
> kernel: Workqueue: cgwb_release cgwb_release_workfn
> kernel: Call Trace:
> kernel: <TASK>
> kernel: show_stack+0x4e/0x53
> kernel: dump_stack_lvl+0x51/0x66
> kernel: dump_stack+0x10/0x12
> kernel: __might_resched.cold+0x173/0x198
> kernel: __might_sleep+0x72/0xe0
> kernel: start_flush_work+0x30/0x560
> kernel: __flush_work+0xf7/0x170
> kernel: __cancel_work_timer+0x22a/0x2c0
> kernel: work_fixup_init+0x20/0x40
> kernel: __debug_object_init+0x23b/0x290
> kernel: debug_object_init+0x16/0x20
> kernel: __init_work+0x20/0x30
> kernel: css_release+0x1f/0xb0
> kernel: percpu_ref_put_many.constprop.0+0x150/0x160
> kernel: blkcg_destroy_blkgs+0x20b/0x230
> kernel: blkcg_unpin_online+0x4e/0x90
> kernel: cgwb_release_workfn+0xba/0x210
> kernel: process_one_work+0x57d/0xa80
> kernel: worker_thread+0x90/0x650
> kernel: kthread+0x185/0x1c0
> kernel: ret_from_fork+0x1f/0x30
> kernel: </TASK>
> kernel: null_blk: disk nullb0 created
> kernel: null_blk: module loaded
There was another blkcg related crash reported that looks similar
to this. Can you try and bisect for-6.2/block and see where you
end up?
--
Jens Axboe
