Re: 2.6.18-mm1

["Aaron Durbin" <adurbin@google.com>]

On 9/26/06, Eric W. Biederman <ebiederm@xmission.com> wrote:
> Andrew Morton <akpm@osdl.org> writes:
>
> > On Tue, 26 Sep 2006 20:04:05 -0600
> > ebiederm@xmission.com (Eric W. Biederman) wrote:
> >
> >> When I apply:
> >> x86_64-mm-insert-ioapics-and-local-apic-into-resource-map
> >>
> >> My e1000 fails to initializes and complains about a bad eeprom checksum.
> >> I haven't tracked this down to root cause yet and I am in the process of
> > building
> >> 2.6.18-mm1 with just that patch reverted to confirm that is the only cause.
> >>
> >> I could not see anything obvious in the patch.  I don't have a clue the patch
> >> could be triggering the problem I'm seeing.
> >>
> >> At a quick visual diff I'm not seeing any other differences in the kernel boot
> >> logs, or in /proc/iomem.
> >
> > This bit looks fishy:
> >
> >  GSI 17 sharing vector 0x4A and IRQ 17
> >  PCI->APIC IRQ transform: 0000:05:0c.0[A] -> IRQ 17
> > +PCI: Cannot allocate resource region 8 of bridge 0000:00:02.0
> > +PCI: Cannot allocate resource region 8 of bridge 0000:01:00.0
> > +PCI: Cannot allocate resource region 8 of bridge 0000:01:00.2
> > +PCI: Cannot allocate resource region 0 of device 0000:01:00.1
> > +PCI: Cannot allocate resource region 0 of device 0000:01:00.3
> > +PCI: Cannot allocate resource region 0 of device 0000:03:04.0
> > +PCI: Cannot allocate resource region 0 of device 0000:03:04.1
> >  PCI-GART: No AMD northbridge found.
> >  PCI: Bridge: 0000:01:00.0
> >    IO window: disabled.
> > -  MEM window: fe000000-fe0fffff
> > +  MEM window: e2000000-e20fffff
> >    PREFETCH window: fd000000-fdffffff
> >  PCI: Bridge: 0000:01:00.2
> >    IO window: 1000-1fff
> > -  MEM window: fe100000-fe1fffff
> > +  MEM window: e2100000-e21fffff
> >    PREFETCH window: disabled.
> >  PCI: Bridge: 0000:00:02.0
> >    IO window: 1000-1fff
> > -  MEM window: fe000000-fe2fffff
> > +  MEM window: e2000000-e22fffff
> >    PREFETCH window: fd000000-fdffffff
> >  PCI: Bridge: 0000:00:06.0
> >    IO window: disabled.
> > @@ -123,17 +131,17 @@
> >  PCI: Bridge: 0000:00:1e.0
> >    IO window: 2000-2fff
> >    MEM window: fb000000-fc0fffff
> > -  PREFETCH window: e2000000-e20fffff
> >
> >
> > Wanna hack into arch/i386/pci/i386.c:pcibios_allocate_bus_resources() and
> > see what is conflicting with what?
>
> Good catch.  Add that to the earlier /proc/iomem output.
> > fe200000-fe200fff : IOAPIC 1
> > fe201000-fe201fff : IOAPIC 2
>
> On that board I have ioapics on pci devices and it appears the way
> the patch is reserving them it doesn't account for ioapics that
> have that property.  I.e.  Those ioapics regions show up in lspci
> on an ioapic pci device and the regions are specified with normal
> base address registers.
>

That patch marks each IOAPIC resource as IORESOURCE_BUSY.  This why
the pci allocation fails. I will post patch to fix this in the
morning.

-Aaron