lkml.org 
[lkml]   [2006]   [Sep]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [BUG] i386 2.6.18 cpu_up: attempt to bring up CPU 4 failed : kernel BUG at mm/slab.c:2698!
From
Date
On Thu, 2006-09-21 at 17:41 -0700, Andrew Morton wrote:
> On Thu, 21 Sep 2006 17:17:31 -0700
> keith mannthey <kmannth@us.ibm.com> wrote:
>
> > I wanted to just give 2.6.18 a spin and I tripped over something I
> > didn't expect.
> >
> >
> > cpu_up: attempt to bring up CPU 4 failed
> > kfree_debugcheck: bad ptr c15f6000h.
> > ------------[ cut here ]------------
> > kernel BUG at mm/slab.c:2698!
> > invalid opcode: 0000 [#1]
> > SMP
> > Modules linked in:
> > CPU: 0
> > EIP: 0060:[<c106ce51>] Not tainted VLI
> > EFLAGS: 00010046 (2.6.18 #1)
> > EIP is at kfree_debugcheck+0x7f/0x90
> > eax: 00000028 ebx: 000015f6 ecx: c1025289 edx: c7653540
> > esi: c15f6000 edi: c15f6000 ebp: c764af38 esp: c764af28
> > ds: 007b es: 007b ss: 0068
> > Process swapper (pid: 1, ti=c764a000 task=c7653540 task.ti=c764a000)
> > Stack: c122c68d c15f6000 c1635000 00000004 c764af5c c106ef93 00000286
> > c76a77d0
> > 00000004 00000001 c1635000 00000004 00000004 c764af6c c10557f6
> > c1274eac
> > c12743dc c764af84 c1207467 00000004 c12734c0 00000004 00000004
> > c764af98
> > Call Trace:
> > [<c106ef93>] kfree+0x24/0x1d8
> > [<c10557f6>] pageset_cpuup_callback+0x40/0x58
> > [<c1207467>] notifier_call_chain+0x20/0x31
> > [<c1031530>] blocking_notifier_call_chain+0x1d/0x2d
> > [<c103f80c>] cpu_up+0xb5/0xcf
> > [<c1000372>] init+0x78/0x296
> > [<c1002005>] kernel_thread_helper+0x5/0xb
>
> I think we have two problems here:
>
> a) CPU4 didn't come up. To diagnose that I think we'll need to ask you
> to into cpu_up(), add debug printks to blocking_notifier_call_chain(),
> work out which entry on that chain returned NOTIFY_BAD, then work out
> why it did so.

That unhappy caller in the chain is cpuup_callback in mm/slab.c. I am
still working out as to why, there is a lot going on if this function.

> b) pageset_cpuup_callback()'s CPU_UP_CANCELED path possibly hasn't been
> tested before. I'd be guessing that we're not zeroing out the
> zone.pageset[] array when the `struct zone' is first allocated, but I
> don't immediately recall where that code lives.

Thanks,
Keith


-
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/

\
 
 \ /
  Last update: 2006-09-22 03:39    [from the cache]
©2003-2011 Jasper Spaans