Messages in this thread | | | Date | Thu, 3 Jul 2008 11:07:53 +0200 | From | "Vegard Nossum" <> | Subject | Re: [BUG] hotplug_cpu vs no_hz |
| |
Hi!
On Thu, Jul 3, 2008 at 8:35 AM, Lai Jiangshan <laijs@cn.fujitsu.com> wrote: > after several seconds ~ several hours, "echo 1 > /sys/devices/system/cpu/cpu1/online" > was blocked, cpu#1 can not be used and the output of dmesg: > > BUG: soft lockup - CPU#1 stuck for 61s! [events/1:9898] > CPU 1: > Modules linked in: > Pid: 9898, comm: events/1 Not tainted 2.6.26-rc8-official-LAI-00089-ge1441b9 #5 > RIP: 0010:[<ffffffff80237612>] [<ffffffff80237612>] __do_softirq+0x4b/0xc7 > RSP: 0018:ffff81006b42ff20 EFLAGS: 00000206 > RAX: ffff81006a9b9fd8 RBX: ffff81006b42ff40 RCX: 0000000000000006 > RDX: 0000000000000042 RSI: ffffffff8022da16 RDI: ffffffff8022da16 > RBP: ffff81006b42fea0 R08: ffff81007f2c9178 R09: ffff81007f2c9140 > R10: ffff8100807cc000 R11: 0000000000000000 R12: ffffffff8020be36 > R13: ffff81006b42fea0 R14: ffffffff807a5100 R15: 0000000000000042 > FS: 0000000000000000(0000) GS:ffff81007fb3ccc0(0000) knlGS:0000000000000000 > CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b > CR2: 00007f4cc97b6000 CR3: 0000000000201000 CR4: 00000000000006a0 > DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 > DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 > > Call Trace: > <IRQ> [<ffffffff8020c38c>] ? call_softirq+0x1c/0x28 > [<ffffffff8020dad6>] ? do_softirq+0x34/0x72 > [<ffffffff80237586>] ? irq_exit+0x3f/0x80 > [<ffffffff8021b128>] ? smp_apic_timer_interrupt+0x8b/0xa7 > [<ffffffff8020be36>] ? apic_timer_interrupt+0x66/0x70 > <EOI> [<ffffffff8022da16>] ? finish_task_switch+0x31/0x82 > [<ffffffff80590e7d>] ? thread_return+0x3d/0x9c > [<ffffffff80241ca1>] ? worker_thread+0xa3/0xe5 > [<ffffffff80244780>] ? autoremove_wake_function+0x0/0x38 > [<ffffffff80241bfe>] ? worker_thread+0x0/0xe5 > [<ffffffff80244645>] ? kthread+0x49/0x78 > [<ffffffff8020c018>] ? child_rip+0xa/0x12 > [<ffffffff802445fc>] ? kthread+0x0/0x78 > [<ffffffff8020c00e>] ? child_rip+0x0/0x12
I believe I have experienced this as well, and I tried to debug it with Peter Zijlstra (Cc added).
I only managed to get this message once. But I made a patch that could help us debug this, see
commit 8d5be7f4e8515af461cbc8f07687ccc81507d508 Date: Wed Jun 25 08:50:10 2008 +0200
softlockup: show irqtrace
from the core/softlockup branch in the -tip tree. (This commit is also present in tip/master.)
In order to see the effect of this, you need CONFIG_TRACE_IRQFLAGS=y. This will tell us when/where irqs were last disabled and will hopefully give a hint of where the block is really occurring. (Unless it is already obvious to you or others; it isn't to me :-))
Thanks,
Vegard
-- "The animistic metaphor of the bug that maliciously sneaked in while the programmer was not looking is intellectually dishonest as it disguises that the error is the programmer's own creation." -- E. W. Dijkstra, EWD1036
| |