Hello so I modified my perf_fuzzer to not try tracepoint event 0x18 (irq_work_exit which is the event causing problems in the other thread).
But I got another lockup, presumbaly with a different trace event. I'll try to narrow down which one caused it. Should we just be disabling the perf/tracepoint interface if not run as root at this point? Vince [243573.968001] ------------[ cut here ]------------ [243573.968001] WARNING: CPU: 0 PID: 4159 at kernel/watchdog.c:245 watchdog_ove) [243573.968001] Watchdog detected hard LOCKUP on cpu 0 [243573.968001] Modules linked in: cpufreq_userspace cpufreq_stats cpufreq_powen [243573.968001] CPU: 0 PID: 4159 Comm: perf_fuzzer Tainted: G W 3.12.2 [243573.968001] Hardware name: AOpen DE7000/nMCP7ALPx-DE R1.06 Oct.19.2012, B2 [243573.968001] 00000000000000f5 ffff88011fc07c48 ffffffff8151aad6 ffff88011fc8 [243573.968001] ffffffff8103cd68 0000000100000000 ffffffff810a101c ffff88011b30 [243573.968001] ffff88011fc07ef8 ffff88011fc07d58 0000000000000000 ffff88011fc8 [243573.968001] Call Trace: [243573.968001] <NMI> [<ffffffff8151aad6>] ? dump_stack+0x41/0x53 [243573.968001] [<ffffffff8103cd68>] ? warn_slowpath_common+0x81/0x9c [243573.968001] [<ffffffff810a101c>] ? watchdog_overflow_callback+0x9b/0xa6 [243573.968001] [<ffffffff8103ce25>] ? warn_slowpath_fmt+0x46/0x48 [243573.968001] [<ffffffff810c9ced>] ? perf_prepare_sample+0x62/0x2f4 [243573.968001] [<ffffffff810a101c>] ? watchdog_overflow_callback+0x9b/0xa6 [243573.968001] [<ffffffff810ca4a2>] ? __perf_event_overflow+0x12e/0x1ab [243573.968001] [<ffffffff810167d3>] ? intel_pmu_lbr_read+0x253/0x349 [243573.968001] [<ffffffff81018c95>] ? intel_pmu_handle_irq+0x2b8/0x34d [243573.968001] [<ffffffff810c79e3>] ? perf_event_context_sched_in+0x70/0x8d [243573.968001] [<ffffffff8151edfc>] ? nmi_handle+0xa8/0x13a [243573.968001] [<ffffffff8151f473>] ? perf_event_nmi_handler+0x24/0x3d [243573.968001] [<ffffffff8151edb2>] ? nmi_handle+0x5e/0x13a [243573.968001] [<ffffffff8151ef35>] ? do_nmi+0xa7/0x2bd [243573.968001] [<ffffffff8151e537>] ? end_repeat_nmi+0x1e/0x2e [243573.968001] [<ffffffff81526160>] ? irq_work_interrupt+0x70/0x70 [243573.968001] [<ffffffff81526160>] ? irq_work_interrupt+0x70/0x70 [243573.968001] [<ffffffff81526160>] ? irq_work_interrupt+0x70/0x70 [243573.968001] <<EOE>> [<ffffffff8106658c>] ? finish_task_switch+0x57/0x98 [243573.968001] [<ffffffff8106657b>] ? finish_task_switch+0x46/0x98 [243573.968001] [<ffffffff8151c9ee>] ? __schedule+0x4f7/0x54b [243573.968001] [<ffffffff8151ab1e>] ? schedule_timeout+0x36/0xe5 [243573.968001] [<ffffffff8106657b>] ? finish_task_switch+0x46/0x98 [243573.968001] [<ffffffff8151ca13>] ? __schedule+0x51c/0x54b [243573.968001] [<ffffffff8151aae8>] ? dump_stack+0x53/0x53 [243573.968001] [<ffffffff8151cb96>] ? __wait_for_common+0xce/0x14a [243573.968001] [<ffffffff810664d6>] ? try_to_wake_up+0x1ab/0x1ab [243573.968001] [<ffffffff810a2fff>] ? T.901+0x1e4/0x1e4 [243573.968001] [<ffffffff810561db>] ? wait_rcu_gp+0x4b/0x54 [243573.968001] [<ffffffff810561e4>] ? wait_rcu_gp+0x54/0x54 [243573.968001] [<ffffffff810be94a>] ? perf_trace_event_unreg+0x2e/0xbd [243573.968001] [<ffffffff810bea07>] ? perf_trace_destroy+0x2e/0x3b [243573.968001] [<ffffffff810c6f9e>] ? __free_event+0x2d/0x52 [243573.968001] [<ffffffff810c8116>] ? perf_event_release_kernel+0x74/0x7b [243573.968001] [<ffffffff810c8358>] ? perf_release+0x10/0x14 [243573.968001] [<ffffffff811133c5>] ? __fput+0xfa/0x1be [243573.968001] [<ffffffff81055ebd>] ? task_work_run+0x88/0xa0 [243573.968001] [<ffffffff81524ea0>] ? int_signal+0x12/0x17 [243573.968001] ---[ end trace 199b386ed4ca91db ]--- [243573.968001] perf samples too long (7436949 > 10000), lowering kernel.perf_e0 [243573.968001] INFO: NMI handler (perf_event_nmi_handler) took too long to runs [243589.024001] perf samples too long (7378854 > 20000), lowering kernel.perf_e0 [243589.844001] perf samples too long (7321212 > 40000), lowering kernel.perf_e0 [243590.692001] perf samples too long (7264020 > 76923), lowering kernel.perf_e0 [243591.540001] perf samples too long (7207276 > 142857), lowering kernel.perf_0 [243592.388001] perf samples too long (7150974 > 250000), lowering kernel.perf_0 [243593.236001] perf samples too long (7095112 > 500000), lowering kernel.perf_0 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
