Re: lockdep warning: inconsistent hardirq-on-W -> in-hardirq-R

[Marcin Slusarz]

On Tue, Oct 14, 2008 at 12:41:41PM +0800, Dave Young wrote:
> Hi,
> 
> kernel built from today's linux-2.6 git, lockdep warnings as following:
> 
> [   52.984024] =================================
> [   52.984030] [ INFO: inconsistent lock state ]
> [   52.984034] 2.6.27-smp #3
> [   52.984037] ---------------------------------
> [   52.984040] inconsistent {hardirq-on-W} -> {in-hardirq-R} usage.
> [   52.984045] pulseaudio/3400 [HC1[1]:SC0[0]:HE0:SE1] takes:
> [   52.984048]  (tasklist_lock){--?-}, at: [<c0142552>] run_posix_cpu_timers+0x52/0x9a0
> [   52.984064] {hardirq-on-W} state was registered at:
> [   52.984068]   [<c01507a0>] mark_held_locks+0x30/0x80
> [   52.984075]   [<c015090a>] trace_hardirqs_on_caller+0x7a/0x140
> [   52.984081]   [<c01509db>] trace_hardirqs_on+0xb/0x10
> [   52.984086]   [<c043a9b2>] _spin_unlock_irq+0x22/0x30
> [   52.984093]   [<c02aea15>] proc_clear_tty+0x45/0x60
> [   52.984099]   [<c013b8a6>] sys_setsid+0x96/0xa0
> [   52.984105]   [<c0103e22>] syscall_call+0x7/0xb
> [   52.984111]   [<ffffffff>] 0xffffffff
> [   52.984124] irq event stamp: 636
> [   52.984127] hardirqs last  enabled at (635): [<c0264294>] trace_hardirqs_on_thunk+0xc/0x10
> [   52.984135] hardirqs last disabled at (636): [<c02642a4>] trace_hardirqs_off_thunk+0xc/0x18
> [   52.984141] softirqs last  enabled at (0): [<c012bcfa>] copy_process+0x26a/0x1160
> [   52.984149] softirqs last disabled at (0): [<00000000>] 0x0
> [   52.984155] 
> [   52.984155] other info that might help us debug this:
> [   52.984160] no locks held by pulseaudio/3400.
> [   52.984162] 
> [   52.984163] stack backtrace:
> [   52.984168] Pid: 3400, comm: pulseaudio Not tainted 2.6.27-smp #3
> [   52.984171] Call Trace:
> [   52.984178]  [<c014ebc6>] print_usage_bug+0x166/0x1a0
> [   52.984183]  [<c01506f6>] mark_lock+0xa96/0xb10
> [   52.984189]  [<c014d83b>] ? trace_hardirqs_off+0xb/0x10
> [   52.984196]  [<c010a53f>] ? native_sched_clock+0x7f/0xb0
> [   52.984202]  [<c0152452>] __lock_acquire+0x552/0x9a0
> [   52.984208]  [<c01507d5>] ? mark_held_locks+0x65/0x80
> [   52.984213]  [<c0152916>] lock_acquire+0x76/0xa0
> [   52.984219]  [<c0142552>] ? run_posix_cpu_timers+0x52/0x9a0
> [   52.984224]  [<c043a826>] _read_lock+0x36/0x60
> [   52.984229]  [<c0142552>] ? run_posix_cpu_timers+0x52/0x9a0
> [   52.984234]  [<c0142552>] run_posix_cpu_timers+0x52/0x9a0
> [   52.984239]  [<c014d83b>] ? trace_hardirqs_off+0xb/0x10
> [   52.984245]  [<c0132b55>] ? raise_softirq+0x65/0x90
> [   52.984251]  [<c012b11c>] ? scheduler_tick+0x13c/0x1b0
> [   52.984256]  [<c01370ef>] update_process_times+0x4f/0x60
> [   52.984262]  [<c014c250>] tick_sched_timer+0x50/0xc0
> [   52.984266]  [<c014c200>] ? tick_sched_timer+0x0/0xc0
> [   52.984271]  [<c014c200>] ? tick_sched_timer+0x0/0xc0
> [   52.984276]  [<c01442d5>] __run_hrtimer+0x75/0x90
> [   52.984282]  [<c0144d1d>] hrtimer_interrupt+0x11d/0x180
> [   52.984287]  [<c0115996>] smp_apic_timer_interrupt+0x56/0x90
> [   52.984293]  [<c02642a4>] ? trace_hardirqs_off_thunk+0xc/0x18
> [   52.984298]  [<c01048d1>] apic_timer_interrupt+0x2d/0x34

Known bug, see: http://lkml.org/lkml/2008/10/13/368

Marcin
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/