* RXE and ib_write_bw
@ 2016-06-15 21:26 Steve Wise
2016-06-15 21:29 ` Steve Wise
0 siblings, 1 reply; 2+ messages in thread
From: Steve Wise @ 2016-06-15 21:26 UTC (permalink / raw)
To: 'Moni Shoua'
Cc: 'Doug Ledford', 'linux-rdma',
'Matan Barak', 'Majd Dibbiny',
'Liran Liss', 'Kamal Heib', 'Amir Vadai',
'Leon Romanovsky'
Hey Moni,
When I run ib_write_bw, I get this logged on one or sometimes both nodes. I think RXE is perhaps doing too much in the interrupt context or in its tasklet?
The test is:
ib_write_bw -a -R -d rxe0 --report_gbits --bidir 10.0.1.14
The dmesg log:
INFO: rcu_sched self-detected stall on CPU
8-...: (1 GPs behind) idle=49f/140000000000001/0 softirq=16690/16690 fqs=6706
INFO: rcu_sched detected stalls on CPUs/tasks:
8-...: (1 GPs behind) idle=49f/140000000000001/0 softirq=16690/16690 fqs=6706
(detected by 13, t=21008 jiffies, g=17642, c=17641, q=4090)
Task dump for CPU 8:
ksoftirqd/8 R running task 0 48 2 0x10000008
ffffffff00000100 ffff88107861ed80 ffffffff81a050f0 0000000000000006
0000000000000100 ffff881078627e68 ffffffff8161abbb 0420804000000000
ffff881078624008 000000010018d17d 0000000a7861ed80 ffffffff81a050f0
Call Trace:
[<ffffffff8161abbb>] ? __do_softirq+0xeb/0x2d8
[<ffffffff81086240>] ? smpboot_create_threads+0x80/0x80
[<ffffffff81614eb0>] ? schedule+0x40/0xb0
[<ffffffff81086240>] ? smpboot_create_threads+0x80/0x80
[<ffffffff810686dd>] ? run_ksoftirqd+0x1d/0x40
[<ffffffff8108634c>] ? smpboot_thread_fn+0x10c/0x160
[<ffffffff8108255c>] ? kthread+0xcc/0xf0
[<ffffffff8108cade>] ? schedule_tail+0x1e/0xc0
[<ffffffff816186cf>] ? ret_from_fork+0x1f/0x40
[<ffffffff81082490>] ? kthread_freezable_should_stop+0x70/0x70
(t=21081 jiffies g=17642 c=17641 q=4090)
Task dump for CPU 8:
ksoftirqd/8 R running task 0 48 2 0x10000008
0000000000000009 ffff88107f403ba8 ffffffff8108956d ffff881078bfa040
0000000000000008 000000000000000f 0000000000000006 0000000000000008
ffffffff81a48240 ffff88107f403bc8 ffffffff8108960f ffff88107f403bc8
Call Trace:
<IRQ> [<ffffffff8108956d>] sched_show_task+0xbd/0x120
[<ffffffff8108960f>] dump_cpu_task+0x3f/0x50
[<ffffffff810c0c3c>] rcu_dump_cpu_stacks+0x7c/0xb0
[<ffffffff810c0f01>] print_cpu_stall+0xf1/0x150
[<ffffffff810c1360>] check_cpu_stall+0x110/0x120
[<ffffffff8111b3b1>] ? trace_buffer_unlock_commit_regs+0x61/0x80
[<ffffffff8112c7f6>] ? trace_event_buffer_commit+0x146/0x1d0
[<ffffffff810682fd>] ? trace_event_raw_event_softirq+0x5d/0xa0
[<ffffffff810c3878>] rcu_check_callbacks+0xe8/0x330
[<ffffffff81068822>] ? __raise_softirq_irqoff+0x72/0x80
[<ffffffff8110aacc>] ? acct_account_cputime+0x1c/0x20
[<ffffffff81092c80>] ? account_system_time+0x90/0x130
[<ffffffff810c6639>] update_process_times+0x39/0x70
[<ffffffff810d7e0a>] tick_sched_handle+0x3a/0x50
[<ffffffff810d8842>] tick_sched_timer+0x52/0xa0
[<ffffffff810c92b7>] __run_hrtimer+0x77/0x1d0
[<ffffffff810d87f0>] ? tick_nohz_handler+0xc0/0xc0
[<ffffffff810682fd>] ? trace_event_raw_event_softirq+0x5d/0xa0
[<ffffffff810cfe5c>] ? ktime_get+0x4c/0xc0
[<ffffffff810c9477>] __hrtimer_run_queues+0x67/0x90
[<ffffffff810c961b>] hrtimer_interrupt+0x9b/0x170
[<ffffffff81044e19>] local_apic_timer_interrupt+0x39/0x60
[<ffffffff8161a9b1>] smp_apic_timer_interrupt+0x41/0x60
[<ffffffff8161908f>] apic_timer_interrupt+0x7f/0x90
<EOI> [<ffffffffa0293662>] ? rxe_requester+0x252/0x780 [ib_rxe]
[<ffffffffa0293643>] ? rxe_requester+0x233/0x780 [ib_rxe]
[<ffffffff8112c7f6>] ? trace_event_buffer_commit+0x146/0x1d0
[<ffffffffa029df89>] rxe_do_task+0x59/0x100 [ib_rxe]
[<ffffffff81068993>] tasklet_action+0xa3/0xc0
[<ffffffff8161abbb>] __do_softirq+0xeb/0x2d8
[<ffffffff81086240>] ? smpboot_create_threads+0x80/0x80
[<ffffffff81614eb0>] ? schedule+0x40/0xb0
[<ffffffff81086240>] ? smpboot_create_threads+0x80/0x80
[<ffffffff810686dd>] run_ksoftirqd+0x1d/0x40
[<ffffffff8108634c>] smpboot_thread_fn+0x10c/0x160
[<ffffffff8108255c>] kthread+0xcc/0xf0
[<ffffffff8108cade>] ? schedule_tail+0x1e/0xc0
[<ffffffff816186cf>] ret_from_fork+0x1f/0x40
[<ffffffff81082490>] ? kthread_freezable_should_stop+0x70/0x70
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 2+ messages in thread* RE: RXE and ib_write_bw
2016-06-15 21:26 RXE and ib_write_bw Steve Wise
@ 2016-06-15 21:29 ` Steve Wise
0 siblings, 0 replies; 2+ messages in thread
From: Steve Wise @ 2016-06-15 21:29 UTC (permalink / raw)
To: 'Moni Shoua'
Cc: 'Doug Ledford', 'linux-rdma',
'Matan Barak', 'Majd Dibbiny',
'Liran Liss', 'Kamal Heib', 'Amir Vadai',
'Leon Romanovsky'
> -----Original Message-----
> From: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org [mailto:linux-rdma-
> owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org] On Behalf Of Steve Wise
> Sent: Wednesday, June 15, 2016 4:27 PM
> To: 'Moni Shoua'
> Cc: 'Doug Ledford'; 'linux-rdma'; 'Matan Barak'; 'Majd Dibbiny'; 'Liran Liss'; 'Kamal
> Heib'; 'Amir Vadai'; 'Leon Romanovsky'
> Subject: RXE and ib_write_bw
>
> Hey Moni,
>
> When I run ib_write_bw, I get this logged on one or sometimes both nodes. I think
> RXE is perhaps doing too much in the interrupt context or in its tasklet?
>
I also see this logged this in the log:
Failed sending ack
Failed sending ack
Failed sending ack
Failed sending ack
> The test is:
>
> ib_write_bw -a -R -d rxe0 --report_gbits --bidir 10.0.1.14
>
> The dmesg log:
>
> INFO: rcu_sched self-detected stall on CPU
> 8-...: (1 GPs behind) idle=49f/140000000000001/0 softirq=16690/16690
> fqs=6706
> INFO: rcu_sched detected stalls on CPUs/tasks:
> 8-...: (1 GPs behind) idle=49f/140000000000001/0 softirq=16690/16690
> fqs=6706
> (detected by 13, t=21008 jiffies, g=17642, c=17641, q=4090)
> Task dump for CPU 8:
> ksoftirqd/8 R running task 0 48 2 0x10000008
> ffffffff00000100 ffff88107861ed80 ffffffff81a050f0 0000000000000006
> 0000000000000100 ffff881078627e68 ffffffff8161abbb 0420804000000000
> ffff881078624008 000000010018d17d 0000000a7861ed80 ffffffff81a050f0
> Call Trace:
> [<ffffffff8161abbb>] ? __do_softirq+0xeb/0x2d8
> [<ffffffff81086240>] ? smpboot_create_threads+0x80/0x80
> [<ffffffff81614eb0>] ? schedule+0x40/0xb0
> [<ffffffff81086240>] ? smpboot_create_threads+0x80/0x80
> [<ffffffff810686dd>] ? run_ksoftirqd+0x1d/0x40
> [<ffffffff8108634c>] ? smpboot_thread_fn+0x10c/0x160
> [<ffffffff8108255c>] ? kthread+0xcc/0xf0
> [<ffffffff8108cade>] ? schedule_tail+0x1e/0xc0
> [<ffffffff816186cf>] ? ret_from_fork+0x1f/0x40
> [<ffffffff81082490>] ? kthread_freezable_should_stop+0x70/0x70
> (t=21081 jiffies g=17642 c=17641 q=4090)
> Task dump for CPU 8:
> ksoftirqd/8 R running task 0 48 2 0x10000008
> 0000000000000009 ffff88107f403ba8 ffffffff8108956d ffff881078bfa040
> 0000000000000008 000000000000000f 0000000000000006 0000000000000008
> ffffffff81a48240 ffff88107f403bc8 ffffffff8108960f ffff88107f403bc8
> Call Trace:
> <IRQ> [<ffffffff8108956d>] sched_show_task+0xbd/0x120
> [<ffffffff8108960f>] dump_cpu_task+0x3f/0x50
> [<ffffffff810c0c3c>] rcu_dump_cpu_stacks+0x7c/0xb0
> [<ffffffff810c0f01>] print_cpu_stall+0xf1/0x150
> [<ffffffff810c1360>] check_cpu_stall+0x110/0x120
> [<ffffffff8111b3b1>] ? trace_buffer_unlock_commit_regs+0x61/0x80
> [<ffffffff8112c7f6>] ? trace_event_buffer_commit+0x146/0x1d0
> [<ffffffff810682fd>] ? trace_event_raw_event_softirq+0x5d/0xa0
> [<ffffffff810c3878>] rcu_check_callbacks+0xe8/0x330
> [<ffffffff81068822>] ? __raise_softirq_irqoff+0x72/0x80
> [<ffffffff8110aacc>] ? acct_account_cputime+0x1c/0x20
> [<ffffffff81092c80>] ? account_system_time+0x90/0x130
> [<ffffffff810c6639>] update_process_times+0x39/0x70
> [<ffffffff810d7e0a>] tick_sched_handle+0x3a/0x50
> [<ffffffff810d8842>] tick_sched_timer+0x52/0xa0
> [<ffffffff810c92b7>] __run_hrtimer+0x77/0x1d0
> [<ffffffff810d87f0>] ? tick_nohz_handler+0xc0/0xc0
> [<ffffffff810682fd>] ? trace_event_raw_event_softirq+0x5d/0xa0
> [<ffffffff810cfe5c>] ? ktime_get+0x4c/0xc0
> [<ffffffff810c9477>] __hrtimer_run_queues+0x67/0x90
> [<ffffffff810c961b>] hrtimer_interrupt+0x9b/0x170
> [<ffffffff81044e19>] local_apic_timer_interrupt+0x39/0x60
> [<ffffffff8161a9b1>] smp_apic_timer_interrupt+0x41/0x60
> [<ffffffff8161908f>] apic_timer_interrupt+0x7f/0x90
> <EOI> [<ffffffffa0293662>] ? rxe_requester+0x252/0x780 [ib_rxe]
> [<ffffffffa0293643>] ? rxe_requester+0x233/0x780 [ib_rxe]
> [<ffffffff8112c7f6>] ? trace_event_buffer_commit+0x146/0x1d0
> [<ffffffffa029df89>] rxe_do_task+0x59/0x100 [ib_rxe]
> [<ffffffff81068993>] tasklet_action+0xa3/0xc0
> [<ffffffff8161abbb>] __do_softirq+0xeb/0x2d8
> [<ffffffff81086240>] ? smpboot_create_threads+0x80/0x80
> [<ffffffff81614eb0>] ? schedule+0x40/0xb0
> [<ffffffff81086240>] ? smpboot_create_threads+0x80/0x80
> [<ffffffff810686dd>] run_ksoftirqd+0x1d/0x40
> [<ffffffff8108634c>] smpboot_thread_fn+0x10c/0x160
> [<ffffffff8108255c>] kthread+0xcc/0xf0
> [<ffffffff8108cade>] ? schedule_tail+0x1e/0xc0
> [<ffffffff816186cf>] ret_from_fork+0x1f/0x40
> [<ffffffff81082490>] ? kthread_freezable_should_stop+0x70/0x70
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2016-06-15 21:29 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-06-15 21:26 RXE and ib_write_bw Steve Wise
2016-06-15 21:29 ` Steve Wise
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox