Re: 2.6.16-rc5-mm1

[Mike Galbraith <efault@gmx.de>]

On Wed, 2006-03-01 at 02:32 -0800, Andrew Morton wrote:
> Laurent Riffard <laurent.riffard@free.fr> wrote:
> >
> > Le 01.03.2006 01:21, Andrew Morton a écrit :
> >  > Laurent Riffard <laurent.riffard@free.fr> wrote:
> >  > 
> >  >>BUG: unable to handle kernel NULL pointer dereference at virtual address 00000034
> >  > 
> >  > 
> >  > I booted that thing on five machines, four architectures :(
> >  > 
> >  > Could people please test a couple more patchsets, see if we can isolate it?
> >  > 
> >  > http://www.zip.com.au/~akpm/linux/patches/stuff/2.6.16-rc5-mm1.1.gz
> >  > 
> >  > is 2.6.16-rc5-mm1 minus:
> >  > 
> >  > proc-make-proc_numbuf-the-buffer-size-for-holding-a.patch
> >  > tref-implement-task-references.patch
> >  > proc-dont-lock-task_structs-indefinitely.patch
> >  > proc-dont-lock-task_structs-indefinitely-git-nfs-fix.patch
> >  > proc-dont-lock-task_structs-indefinitely-cpuset-fix.patch
> >  > proc-optimize-proc_check_dentry_visible.patch
> > 
> >  Ok, 2.6.16-rc5-mm1.1 works for me:
> >  - I can run java from command line in runlevel 1
> >  - I can launch Mozilla in X
> 
> Useful, thanks.  So the second batch of /proc patches are indeed the problem.
> 
> If you have (even more) time you could test
> http://www.zip.com.au/~akpm/linux/patches/stuff/2.6.16-rc5-mm2-pre1.gz. 
> That's the latest of everything with the problematic sysfs patches reverted
> and Eric's recent /proc fixes.

Seems to work OK here, with debug settings as in Paul Jackson's report.
No java crash, no fuser -n tcp NNNN crash.  Only thing I see so far is
the proc symbolic link to nirvana permissions thingie.  Box is P4 HT.

	-Mike