Re: [2.6.37-rc1 / rc2 Regression] PCI: Kernel Hangs On Boot

[Bjorn Helgaas <bjorn.helgaas@hp.com>]

On Wed, Nov 17, 2010 at 09:10:19PM +0530, David John wrote:
> Commit dc9887dc ("x86/PCI: allocate space from the end of a region, not
> the beginning") causes the kernel to hang on my Dell Inspiron 1545.
> 2.6.36 and booting with pci=nocrs works fine. Reverting this commit also
> works. I am attaching the kernel log and system iomem from the working
> 2.6.36. It hangs at this location consistently:
> 
> pci-stub: invalid id string ""
> ACPI: AC Adapter [AC] (on-line)
> input: Lid Switch as
> /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0C0D:00/input/input0
> 
> < hang >
> 
> Let me know if you need any other info.

Thanks a lot for this report.  I opened this bugzilla:
    https://bugzilla.kernel.org/show_bug.cgi?id=23132
to track it.

Would you mind attaching the dmesg log from 2.6.37-rc2 with
commit dc9887dc reverted?  That should have a little more
information (PCI allocation closer to what fails, and also
more PNP resource information).

The PCI assignments Linux did are only for bridge windows,
and I don't see anything actually *using* those windows, so
I suspect they might conflict with an ACPI/PNP device, which
is what we're poking at when it hangs.

If you happen to have Windows on this box as well, an
Everest report (http://lavalys.com) would be interesting.
If not, no worries.

Thanks again,
  Bjorn