Re: [PATCH v2 4/5] workqueue: Show all busy workers in stall diagnostics

[Petr Mladek]

On Thu 2026-06-11 07:50:04, Breno Leitao wrote:
> On Fri, Mar 20, 2026 at 03:41:13AM -0700, Breno Leitao wrote:
> > On Wed, Mar 18, 2026 at 04:11:54PM +0100, Petr Mladek wrote:
> > > On Wed 2026-03-18 04:31:08, Breno Leitao wrote:
> > > Otherwise, I like this patch.
> > > 
> > > I still think what might be the reason that there is no worker
> > > in the running state. Let's see if this patch brings some useful info.
> > > 
> > > One more idea. It might be useful to store a timestamp when the last
> > > worker was woken. And then print either the timestamp or delta.
> > > It would help to make sure that kick_pool() was really called
> > > during the reported stall.
> > 
> > Ack, this is the following patch I will deploy in production, let's see
> > how useful it is.
> 
> I got this running in production (backported to 6.16), and we finally got the culprit.
> 
> 	05:42:00  BUG: workqueue lockup - pool cpus=2 node=0 flags=0x0 nice=0 stuck for 115s!
> 		NMI backtrace for cpu 2
> 		CPU: 2 UID: 0 PID: 411 Comm: kworker/u288:2 Tainted: G           O        6.16.1-0_fbk4_0_gb849430a436c #1 NONE
> 		Tainted: [O]=OOT_MODULE
> 		Hardware name: <foo>
> 		Workqueue: efi_rts_wq efi_call_rts
> 		pstate: 23401009 (nzCv daif +PAN -UAO +TCO +DIT +SSBS BTYPE=--)
> 		pc : 0x4052f10900
> 		lr : 0x4052f10e94
> 		sp : ffff800088cefc90
> 		x29: ffff800088cefc90 x28: 0000000048524641 x27: 0000004052b60000
> 		x26: 0000000000010058 x25: 0000004043ba0000 x24: 0000000001280000
> 		x23: 000000405a02807f x22: 0000000000010080 x21: 0000004053ac0097
> 		x20: 000000405a028080 x19: 0000004053ac0098 x18: 0000000000000000
> 		x17: 0000000000000030 x16: 0000004052eb6de0 x15: 0000004042ba0030
> 		x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000001
> 		x11: 0000000001d00d09 x10: 0000004042ba0028 x9 : ffff800088cefc90
> 		x8 : 0000000001d00cd9 x7 : 0000000000000000 x6 : 0000004043ba0000
> 		x5 : 0000004043bb0000 x4 : 0000004053ac0098 x3 : 000000405a028080
> 		x2 : 0000000000000000 x1 : 0000000000000000 x0 : ffffffffffffe1e8
> 		Call trace:
> 		0x4052f10900 (P)
> 		0x4052f10e94
> 		0x4052b00ed0
> 		0x4052b02e38
> 		0x4052b0175c
> 		0x4052b517b4
> 		0x4052a70b84
> 		0x4052cb11d4
> 		__efi_rt_asm_wrapper+0x50/0x78
> 		efi_call_rts+0x178/0x240
> 		process_scheduled_works+0x17c/0x420
> 		worker_thread+0x184/0x4d8
> 		kthread+0xcc/0x1f8
> 		ret_from_fork+0x10/0x20
> 	05:42:30  BUG: workqueue lockup - pool cpus=2 node=0 flags=0x0 nice=0 stuck for 145s!
> 		NMI backtrace for cpu 2
> 		CPU: 2 UID: 0 PID: 411 Comm: kworker/u288:2 Tainted: G           O        6.16.1-0_fbk4_0_gb849430a436c #1 NONE
> 		Tainted: [O]=OOT_MODULE
> 		Hardware name: <foo>
> 		Workqueue: efi_rts_wq efi_call_rts
> 		pstate: 63401009 (nZCv daif +PAN -UAO +TCO +DIT +SSBS BTYPE=--)
> 		pc : 0x4052f11ecc
> 		lr : 0x4052f10b8c
> 		sp : ffff800088cefc30
> 		x29: ffff800088cefc40 x28: 0000000048524641 x27: 0000004052b60000
> 		x26: 0000000000010058 x25: 0000004043fb0000 x24: 0000000001690000
> 		x23: 0000004053ab0040 x22: 0000000000010080 x21: ffff800088cefd00
> 
> 	rinse and repeat..
> 
> Unfortunately I didn't get the other pr_info(), because of console settings,
> but, I can say the following from this issue and previous code:
> 
> 1) in show_cpu_pool_hog, found_running variable is set to false.
> 2) hash_for_each() never found any running task
> 3) The following code was trigger and was very helpful:
> 
>          if (!found_running)
>                 trigger_single_cpu_backtrace(cpu);

Great. So, the extra complexity was worth it. Should I clean it and
send a proper patch? Or would you like to do so?

Also I wonder whether it would make sense to revert the commit
8823eaef45da7f ("workqueue: Show all busy workers in stall
diagnostics"). If I get it correctly then printing all busy workers
was not that helpful. Namely, the sleeping workers should not prevent
progress.

Best Regards,
Petr