On 2014-07-10 23:27, Michael Mattsson wrote:
Hey, I don't get any kernel messages about stuck processes. echo t > /proc/sysrq-trigger gives the following related to fio: fio S 0000000000000004 0 4189 3638 0x00000080 ffff8817a0fd5bf8 0000000000000086 0000000000000000 ffffffff8111f867 ffff8817a0fd5d01 00007fa2f43b6000 ffff8817a0fd5bd8 ffffffff810aec10 ffff88186b867ab8 ffff8817a0fd5fd8 000000000000fbc8 ffff88186b867ab8 Call Trace: [<ffffffff8111f867>] ? unlock_page+0x27/0x30 [<ffffffff810aec10>] ? get_futex_key+0x180/0x2b0 [<ffffffff810ae559>] futex_wait_queue_me+0xb9/0xf0 [<ffffffff810af668>] futex_wait+0x1f8/0x380 [<ffffffff8100988e>] ? __switch_to+0x26e/0x320 [<ffffffff810b0f31>] do_futex+0x121/0xb50 [<ffffffff8109f491>] ? lock_hrtimer_base+0x31/0x60 [<ffffffff810a010f>] ? hrtimer_try_to_cancel+0x3f/0xd0 [<ffffffff810a01c2>] ? hrtimer_cancel+0x22/0x30 [<ffffffff8152a413>] ? do_nanosleep+0x93/0xc0 [<ffffffff810a0294>] ? hrtimer_nanosleep+0xc4/0x180 [<ffffffff810b19db>] sys_futex+0x7b/0x170 [<ffffffff810e1e87>] ? audit_syscall_entry+0x1d7/0x200 [<ffffffff810e1c7e>] ? __audit_syscall_exit+0x25e/0x290 [<ffffffff8100b072>] system_call_fastpath+0x16/0x1b
OK, so nothing unexpected there. But it's all still very weird. Could you try and attach gdb to a fio process that is stuck like this, and generate 'bt' backtraces?
-- Jens Axboe -- To unsubscribe from this list: send the line "unsubscribe fio" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
