Re: panic: vm_fault: fault on nofault entry

On Mon, 2014-03-10 at 15:10 -0400, Glen Barber wrote:
> On Mon, Mar 10, 2014 at 09:01:12PM +0200, Konstantin Belousov wrote:
> > On Mon, Mar 10, 2014 at 02:05:08PM -0400, Glen Barber wrote:
> > > Unread portion of the kernel message buffer:
> > > Sleeping thread (tid 100702, pid 24712) owns a non-sleepable lock
> > 
> > Would be nice to see the full message before and panic from the console.
> 
> I will include it in the future.
> 
> > From what I see, this is a lock leak, I forgot to unlock the map.
> > It is nice that it is so simple to reproduce the issue in your setup.
> > 
> > Try this update.
> > 
> 
> I will have the machine updated with this patch in the next few minutes.
> 
> Thank you.
> 
> Glen
> 


All 4 machines have been patched and have been grinding away for several
days now.  I'd say this is a good test and we should commit this.

$ for i in 1 2 3 4; do ssh redbuild0${i} uptime; done
 5:47PM  up 1 day, 23:22, 1 user, load averages: 1.36, 1.10, 0.57
 5:47PM  up 1 day, 23:23, 1 user, load averages: 4.33, 3.87, 2.08
 5:47PM  up 3 days, 22:45, 1 user, load averages: 16.87, 12.47, 10.11
 5:47PM  up 9 days, 20:10, 1 user, load averages: 11.58, 12.34, 10.93


sean