powerpc/pci sysdata batch hangs G5 boot

[Hugh Dickins]

Hi Grant,

I've been unable to boot mmotm on the G5 for a few weeks; and now that
the problem has reached Linus, I've bisected and it converges on your:

commit b5d937de0367d26f65b9af1aef5f2c34c1939be0
    powerpc/pci: Make both ppc32 and ppc64 use sysdata for pci_controller
   
    Currently, ppc32 uses sysdata for the pci_controller pointer, and
    ppc64 uses it to hold the device_node pointer.  This patch moves the
    of_node pointer into (struct pci_bus*)->dev.of_node and
    (struct pci_dev*)->dev.of_node so that sysdata can be converted to always
    use the pci_controller pointer instead.  It also fixes up the
    allocating of pci devices so that the of_node pointer gets assigned
    consistently and increments the ref count.

The last message I see on screen when booting is
mpic: ISU size: 124, shift: 7, mask: 7f

Usually that would be followed by
mpic: Initializing for 124 sources

but with your commit it just hangs; revert that commit,
from mmotm or from Linus's current git, and all is fine.

Config available privately on request.  Noticing that patch is a
PPC OF PCI thing, and that I have CONFIG_PPC_OF_PLATFORM_PCI off,
I did try switching that on, but it made no difference.

Thanks,
Hugh