Re: BUG during shutdown - bisected to commit e2912009

[Marc Dionne <marc.c.dionne@gmail.com>]

On Wed, Jan 6, 2010 at 10:07 PM, Marc Dionne <marc.c.dionne@gmail.com> wrote:
> On Wed, Jan 6, 2010 at 9:51 PM, Xiaotian Feng <dfeng@redhat.com> wrote:
>> On 01/07/2010 08:44 AM, Marc Dionne wrote:
>>>
>>> On Wed, Jan 6, 2010 at 4:42 AM, Xiaotian Feng<dfeng@redhat.com>  wrote:
>>>>
>>>> On 01/06/2010 06:58 AM, Marc Dionne wrote:
>>>>>
>>>>> On Tue, Jan 5, 2010 at 5:18 AM, Xiaotian Feng<dfeng@redhat.com>
>>>>>  wrote:
>>>>>>
>>>>>> This is outputed by sound module, but it will not affect clockevents,
>>>>>> could
>>>>>> you please try following patch and let me know the output before BUG_ON
>>>>>> happens? We can gather more information on the BUG_ON. Thank you.
>>>>>>
>>>>>> diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c
>>>>>> index 6f740d9..7c945e8 100644
>>>>>> --- a/kernel/time/clockevents.c
>>>>>> +++ b/kernel/time/clockevents.c
>>>>>> @@ -260,6 +260,9 @@ void clockevents_notify(unsigned long reason, void
>>>>>> *arg)
>>>>>>                list_for_each_entry_safe(dev, tmp,&clockevent_devices,
>>>>>> list)
>>>>>> {
>>>>>>                        if (cpumask_test_cpu(cpu, dev->cpumask)&&
>>>>>>                            cpumask_weight(dev->cpumask) == 1) {
>>>>>> +                               if (dev->mode != CLOCK_EVT_MODE_UNUSED)
>>>>>> +                                       printk("invalid dev %s mode %d
>>>>>> on
>>>>>> cpu %d\n", dev->name,
>>>>>> +                                               dev->mode, cpu);
>>>>>>                                BUG_ON(dev->mode !=
>>>>>> CLOCK_EVT_MODE_UNUSED);
>>>>>>                                list_del(&dev->list);
>>>>>
>>>>> I don't get anything on screen from the printk - is there a trick
>>>>> needed to getting printk output at that stage of shutting down?  I
>>>>> tried inserting an mdelay() before the BUG, which delayed the bug
>>>>> output but still didn't print the invalid dev message.
>>>>
>>>> Did you notice this BUG when you're doing suspend/resume?
>>>>
>>>> Does the BUG still appear if we changed BUG_ON line to BUG_ON(dev->mode
>>>> !=
>>>> CLOCK_EVT_MODE_UNUSED&&  dev->mode != CLOCK_EVT_MODE_SHUTDOWN)?
>>>
>>> I only see the BUG on halt - reboot works normally and suspend
>>> actually freezes and doesn't suspend, but that's perhaps unrelated.
>>>
>>> I managed to get your suggested printk to work by adding KERN_CRIT
>>> (otherwise I got no output), and the offending dev is:
>>>  "hpet", mode 3 (CLOCK_EVT_MODE_ONESHOT?), cpu 4.
>>
>> It looks like kernel is trying to remove broadcast device, could you please
>> try following patch?
>>
>> diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c
>> index 6f740d9..d7395fd 100644
>> --- a/kernel/time/clockevents.c
>> +++ b/kernel/time/clockevents.c
>> @@ -259,7 +259,8 @@ void clockevents_notify(unsigned long reason, void *arg)
>>                cpu = *((int *)arg);
>>                list_for_each_entry_safe(dev, tmp, &clockevent_devices, list)
>> {
>>                        if (cpumask_test_cpu(cpu, dev->cpumask) &&
>> -                           cpumask_weight(dev->cpumask) == 1) {
>> +                           cpumask_weight(dev->cpumask) == 1 &&
>> +                           !tick_is_broadcast_device(dev)) {
>>                                BUG_ON(dev->mode != CLOCK_EVT_MODE_UNUSED);
>>                                list_del(&dev->list);
>>                        }
>
> That works - no problem shutting down with that patch applied.

And after doing a bit more testing, it turns out that applying this
patch also makes suspend/resume work normally again, so it looks like
the hang on suspend was also related to this.

Thanks,
Marc