2.5.8-pre2: preempt: exits with preempt_count 1

2.5.8-pre2: preempt: exits with preempt_count 1

[Duncan Sands]

My system (x86 K6 UP running 2.5.8-pre2 with preemption) on powering down
gave:

...
Power down.
error: halt[411] exited with preempt_count 1

This was after about 24 hours of up time.  What can I do to help
track down this locking problem?

All the best,

Duncan.

Re: 2.5.8-pre2: preempt: exits with preempt_count 1

[Robert Love]

On Wed, 2002-04-10 at 08:53, Duncan Sands wrote:

> error: halt[411] exited with preempt_count 1
> 
> This was after about 24 hours of up time.  What can I do to help
> track down this locking problem?

It is not a big deal.  The issue is that in the system shutdown code,
something does not release a lock but just figures "the system is going
down, what is the point?"  It is probably the BKL ...

For the sake of code readability and having nothing quit with a nonzero
preempt_count, we should explicitly drop the lock, but it is not hurting
anything (now, if you get this message elsewhere, there may be a
problem).

I am trying to find what is the cause but I have not tracked it down yet
...

	Robert Love

Re: 2.5.8-pre2: preempt: exits with preempt_count 1

[Pierre Rousselet]

Robert Love wrote:
> On Wed, 2002-04-10 at 08:53, Duncan Sands wrote:
> 
> 
>>error: halt[411] exited with preempt_count 1
>>
>>This was after about 24 hours of up time.  What can I do to help
>>track down this locking problem?

Duncan Sands wrote:
 > UP x86 K6 system running 2.5.8-pre3 with preemption.
 > Using usb-uhci.  I got the following bug when powering off:

It looks like one problem, caused by some usb device driver not exiting 
cleanly.

Pierre
-- 
------------------------------------------------
  Pierre Rousselet <pierre.rousselet@wanadoo.fr>
------------------------------------------------

Re: 2.5.8-pre2: preempt: exits with preempt_count 1

[Duncan Sands]

On Wednesday 10 April 2002 8:42 pm, Pierre Rousselet wrote:
> Robert Love wrote:
> > On Wed, 2002-04-10 at 08:53, Duncan Sands wrote:
> >>error: halt[411] exited with preempt_count 1
> >>
> >>This was after about 24 hours of up time.  What can I do to help
> >>track down this locking problem?
>
> Duncan Sands wrote:
>  > UP x86 K6 system running 2.5.8-pre3 with preemption.
>  > Using usb-uhci.  I got the following bug when powering off:
>
> It looks like one problem, caused by some usb device driver not exiting
> cleanly.

Probably.  There was no BUG with 2.5.8-pre2 though.

Duncan.

Re: 2.5.8-pre2: preempt: exits with preempt_count 1

[Chris Wright]

* Robert Love (rml@tech9.net) wrote:
> On Wed, 2002-04-10 at 08:53, Duncan Sands wrote:
> 
> > error: halt[411] exited with preempt_count 1
> > 
> > This was after about 24 hours of up time.  What can I do to help
> > track down this locking problem?
> 
> It is not a big deal.  The issue is that in the system shutdown code,
> something does not release a lock but just figures "the system is going
> down, what is the point?"  It is probably the BKL ...
> 
> For the sake of code readability and having nothing quit with a nonzero
> preempt_count, we should explicitly drop the lock, but it is not hurting
> anything (now, if you get this message elsewhere, there may be a
> problem).

ditto with nfs.  doesn't require a system shutdown, just stopping the
nfs server.

error: nfsd[983] exited with preempt_count 1
error: rpciod[994] exited with preempt_count 1
error: lockd[993] exited with preempt_count 1

cheers,
-chris

Re: 2.5.8-pre2: preempt: exits with preempt_count 1

[Bongani]

On Wed, 2002-04-10 at 20:28, Robert Love wrote:
> On Wed, 2002-04-10 at 08:53, Duncan Sands wrote:
> 
> > error: halt[411] exited with preempt_count 1
> > 
> > This was after about 24 hours of up time.  What can I do to help
> > track down this locking problem?
> 
> It is not a big deal.  The issue is that in the system shutdown code,
> something does not release a lock but just figures "the system is going
> down, what is the point?"  It is probably the BKL ...
> 
> For the sake of code readability and having nothing quit with a nonzero
> preempt_count, we should explicitly drop the lock, but it is not hurting
> anything (now, if you get this message elsewhere, there may be a
> problem).
> 
> I am trying to find what is the cause but I have not tracked it down yet

This is also still happening on 2.4.19-pre6 + preempt 


> ...

> 
> 	Robert Love





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

Re: 2.5.8-pre2: preempt: exits with preempt_count 1

[Skip Ford]

Duncan Sands wrote:
> My system (x86 K6 UP running 2.5.8-pre2 with preemption) on powering down
> gave:
> ...
> Power down.
> error: halt[411] exited with preempt_count 1
> 
> This was after about 24 hours of up time.  What can I do to help
> track down this locking problem?

If you fixed the oopsing 2.5.8-pre1 kernel by putting 'lock_kernel()'
in exit.c, then you have to remove that line.

The correct fix was applied later, and if you leave in the call
to 'lock_kernel()' you get the exact message you're reporting.

If line 505 of kernel/exit.c is lock_kernel() in your tree then delete
it, and try again.

-- 
Skip

Re: 2.5.8-pre2: preempt: exits with preempt_count 1

[Duncan Sands]

On Wednesday 10 April 2002 8:40 pm, Skip Ford wrote:
> Duncan Sands wrote:
> > My system (x86 K6 UP running 2.5.8-pre2 with preemption) on powering down
> > gave:
> > ...
> > Power down.
> > error: halt[411] exited with preempt_count 1
> >
> > This was after about 24 hours of up time.  What can I do to help
> > track down this locking problem?
>
> If you fixed the oopsing 2.5.8-pre1 kernel by putting 'lock_kernel()'
> in exit.c, then you have to remove that line.
>
> The correct fix was applied later, and if you leave in the call
> to 'lock_kernel()' you get the exact message you're reporting.
>
> If line 505 of kernel/exit.c is lock_kernel() in your tree then delete
> it, and try again.

No, it wasn't that, but thanks for thinking about it.

Ciao, Duncan.