Re: kern/46402: LWPs created after exit_lwp() is called can hang the process....

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Thu, May 03, 2012 at 10:22:21AM -0600, Greg Oster wrote:
> On Thu,  3 May 2012 15:00:07 +0000 (UTC)
> Manuel Bouyer <bouyer%antioche.eu.org@localhost> wrote:
> 
> > The following reply was made to PR kern/46402; it has been noted by
> > GNATS.
> > 
> > From: Manuel Bouyer <bouyer%antioche.eu.org@localhost>
> > To: gnats-bugs%NetBSD.org@localhost
> > Cc: kern-bug-people%NetBSD.org@localhost, gnats-admin%NetBSD.org@localhost,
> > netbsd-bugs%NetBSD.org@localhost Subject: Re: kern/46402: LWPs created after
> > exit_lwp() is called can hang the process....
> > Date: Thu, 3 May 2012 16:57:09 +0200
> > 
> >  On Wed, May 02, 2012 at 09:30:00PM +0000, oster%NetBSD.org@localhost wrote:
> >  > While running an extended series of aft-tests I noticed that the
> >  > test were occasionally hanging.  It turns out they were hanging at
> >  > the same place, and that the test is repeatable.  Running the
> >  > following:
> >  > 
> >  >  cd /usr/tests/lib/libpthread
> >  >  foreach i (`jot 10000`)
> >  >  ./t_cond cond_timedwait_race 
> >  >  echo $i
> >  >  end
> >  > 
> >  
> >  Sounds familiar: kern/46168
> 
> yes!  That's exactly the same bug....  I ran with a patch very, very
> similar to the one you had in 46168, but I'd end up with a very hard
> kernel hang after some time... (couldn't break into ddb as normal,
> even).
> 
> >  Andrew Doran had some improvements to the patch I proposed here, but
> > it seems he forgot to send me an updated patch (this was a month ago).
> >  I guess we should get the patch in kern/46168 commited then ...
> 
> I don't think it's sufficient, unfortunately... there's still something
> else missing.. :(  

It may be a different bug; see the thread about hard hang on port-amd64.
I've never seen these hard hangs myself, and I've tested different hardware ...

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--