[RFC] panic on cascade() of kernel/time/timer.c

[RFC] panic on cascade() of kernel/time/timer.c

[Yisheng Xie]

Hi all,
The kernel version is 4.1.18, and we got a panic at cascade().
from the log it seems cause by BUG_ON of kernel/time/timer.c:1120.
Any ideas about about it?

Any reply will be appreciated.

Thanks,
Yisheng.

-------------------------
[76661.805297s][pid:457,cpu3,surfaceflinger]CPU: 3 PID: 457 Comm: surfaceflinger Not tainted 4.1.18-gffd2c04 #1
[76661.805297s][pid:457,cpu3,surfaceflinger]TGID: 457 Comm: surfaceflinger
[76661.805297s][pid:457,cpu3,surfaceflinger]Hardware name: hi3660 (DT)
[76661.805328s][pid:457,cpu3,surfaceflinger]task: ffffffc0cf056600 ti: ffffffc0cc5ec160 task.ti: ffffffc0cc5ec160
[76661.805358s][pid:457,cpu3,surfaceflinger]PC is at panic+0x84/0x29c
[76661.805389s][pid:457,cpu3,surfaceflinger]LR is at cascade+0xdc/0x108
[76661.805389s][pid:457,cpu3,surfaceflinger]pc : [<ffffffc00120ae90>] lr : [<ffffffc0001118a0>] pstate: 000001c4
[76661.805389s][pid:457,cpu3,surfaceflinger]sp : ffffffc0cc5ef210
[76661.805419s]x29: ffffffc0cc5ef210 x28: 0000000000000000
[76661.805419s]x27: 0000000000000100 x26: ffffffc001e7b000
[76661.805450s]x25: ffffffc0cc5ec000 x24: 0000000000000282
[76661.805480s]x23: ffffffc001a7c000 x22: ffffffc0016547f0
[76661.805480s]x21: ffffffc0cc5ef458 x20: ffffffc001a99000
[76661.805511s]x19: ffffffc001d35000 x18: 000000000000000f
[76661.805541s]x17: 0000000000000045 x16: 00000000000001f7
[76661.805541s]x15: 0000000000000640 x14: 2f65646f635f746e
[76661.805572s]x13: 656e6f706d6f632f x12: 646c6975625f616e
[76661.805603s]x11: 6968635f4f474143 x10: 4948435f746e656e
[76661.805603s]x9 : 0000000000000000 x8 : ffffffc0cc5ef3b8
[76661.805633s]x7 : 0000000000000000 x6 : 000000000000003f
[76661.805633s]x5 : 0000000000000040 x4 : ffffffffffffffe0
[76661.805664s]x3 : 0000000000000030 x2 : 0000000000000008
[76661.805694s]x1 : 0000000000000000 x0 : ffffffc0cc5ef298
[...]
[76661.810882s][pid:457,cpu3,surfaceflinger]
[76661.810882s][pid:457,cpu3,surfaceflinger]Kernel panic - not syncing: BUG!
[76661.810913s][pid:457,cpu3,surfaceflinger]CPU: 3 PID: 457 Comm: surfaceflinger Not tainted 4.1.18-gffd2c04 #1
[76661.810913s][pid:457,cpu3,surfaceflinger]TGID: 457 Comm: surfaceflinger
[76661.810943s][pid:457,cpu3,surfaceflinger]Hardware name: hi3660 (DT)
[76661.810943s][pid:457,cpu3,surfaceflinger]Call trace:
[76661.810974s][pid:457,cpu3,surfaceflinger][<ffffffc00008a5c8>] dump_backtrace+0x0/0x15c
[76661.810974s][pid:457,cpu3,surfaceflinger][<ffffffc00008a744>] show_stack+0x20/0x28
[76661.811004s][pid:457,cpu3,surfaceflinger][<ffffffc00120cf78>] dump_stack+0x84/0xa8
[76661.811004s][pid:457,cpu3,surfaceflinger][<ffffffc00120af60>] panic+0x154/0x29c
[76661.811035s][pid:457,cpu3,surfaceflinger][<ffffffc0001118a0>] cascade+0xdc/0x108
[76661.811035s][pid:457,cpu3,surfaceflinger][<ffffffc000111c2c>] run_timer_softirq+0x2f8/0x350
[76661.811065s][pid:457,cpu3,surfaceflinger][<ffffffc0000a782c>] __do_softirq+0xe4/0x40c
[76661.811065s][pid:457,cpu3,surfaceflinger][<ffffffc0000a7e58>] irq_exit+0xb4/0xe4
[76661.811096s][pid:457,cpu3,surfaceflinger][<ffffffc0000fd45c>] __handle_domain_irq+0xb0/0x110
[76661.811096s][pid:457,cpu3,surfaceflinger][<ffffffc000081570>] gic_handle_irq+0x3c/0x8c

Re: [RFC] panic on cascade() of kernel/time/timer.c

[Thomas Gleixner]

On Wed, 30 Nov 2016, Yisheng Xie wrote:

> Hi all,
> The kernel version is 4.1.18, and we got a panic at cascade().
> from the log it seems cause by BUG_ON of kernel/time/timer.c:1120.
> Any ideas about about it?

Probably a timer which got freed while enqueued or some other wreckage in
that area. Certainly not a core code problem.

Enable CONFIG_DEBUG_OBJECTS* and try to reproduce. The debug facility will
tell you which driver/whatever causes the problem.

Thanks,

	tglx

Re: [RFC] panic on cascade() of kernel/time/timer.c

[Yisheng Xie]

On 2016/11/30 17:18, Thomas Gleixner wrote:
> On Wed, 30 Nov 2016, Yisheng Xie wrote:
> 
>> Hi all,
>> The kernel version is 4.1.18, and we got a panic at cascade().
>> from the log it seems cause by BUG_ON of kernel/time/timer.c:1120.
>> Any ideas about about it?
> 
> Probably a timer which got freed while enqueued or some other wreckage in
> that area. Certainly not a core code problem.
> 
> Enable CONFIG_DEBUG_OBJECTS* and try to reproduce. The debug facility will
> tell you which driver/whatever causes the problem.
> 
Hi tglx,
Thank you for your suggestion, and we will do that test.

Thanks,
Yisheng Xie
> Thanks,
> 
> 	tglx
> 
> 
> .
>