Re: [BUG 2.6.25-rc3] scheduler/hotplug: some processes aredealocked when cpu is set to offline

[Yi Yang]

On Fri, 2008-03-07 at 16:21 +0530, Gautham R Shenoy wrote:
> On Fri, Mar 07, 2008 at 02:40:49PM +0530, Gautham R Shenoy wrote:
> > 
> > > 
> > > Just to be sure, there were no "bad ->cpu..." messages, yes?
> > 
> > Hopefully should be able to catch them now. If yes, it's a problem in
> > the way we do migration after cpu-hotplug as Yi suggested in an earlier
> > mail.
> > 
> > http://lkml.org/lkml/2008/3/6/437
> > 
> > This mail from akpm says the same thing.
> 
> Yup! There are a quite a few "bad->cpu" messages.
> All of them only for watchdog/1.
> 
> What surprises me is the fact that the first of  task-hung messages come
> after 136 successful cpu-hotplug attempts.
> 
> To answer Andrew's question, migration_call() *should* ideally be the
> first notifier called, because it's registered with the priority 10.
> 
> Unless there's some change that has recently gone in the
> notifier_call_chain subsystem that doesn't honour the registration priority
> anymore, migration_call() will be called first,
> which means the task affinity for watchdog thread is broken and hence
> we shouldn't see the message below.
> 
> I'll probe to see if there's a change in the call order that's causing
> this hang.
I have traced all the cpu callback functions, i ensure migration_call is
called firstly, i have ever set notifier priority of cpu_callback in
kernel/softlockup.c to maximum int and check what will happen, the
result is same.

> 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> INFO: task watchdog/1:10958 can't get CPU for more than 120 seconds.
> bad ->cpu
>        f524ffac 00000046 c011befc f5276aa0 f5276bd8 c443db80 00000001 f6130b80 
>        f524ff94 00000246 c07130f4 03d28000 00000000 f524ff9c c01438ad f524ffac 
>        00000001 c0143a1f 00000000 f524ffd0 c0143a66 00000001 c0119fea 00000000 
> Call Trace:
>  [<c011befc>] ? cpu_clock+0x4e/0x59
>  [<c01438ad>] ? get_timestamp+0x8/0x11
>  [<c0143a1f>] ? watchdog+0x0/0x239
>  [<c0143a66>] watchdog+0x47/0x239
>  [<c0119fea>] ? complete+0x34/0x3e
>  [<c0143a1f>] ? watchdog+0x0/0x239
>  [<c012f797>] kthread+0x3b/0x64
>  [<c012f75c>] ? kthread+0x0/0x64
>  [<c01056eb>] kernel_thread_helper+0x7/0x10
>  =======================
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> 
> 
> >
> > >
> > > Oleg.
> > 
> > -- 
> > Thanks and Regards
> > gautham
> 

--
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/