Booting Imac G5 : success !

[Benjamin Herrenschmidt]

> OK, you're right. The problem is that in the PCI configuration space of
> the SATA root, there's no IRQ line defined. I can see we're in the same
> case for the IDE controller, and I saw this code:
> drivers/ide/ppc/pmac.c:pmac_ide_pci_attach
> 	np = pci_device_to_OF_node(pdev);
> 	...
> 	if (np->n_intrs == 0)
> 		pmif->irq = pdev->irq;
> 	else
> 		pmif->irq = np->intrs[0].line;
> Then I did the same in the sata driver and I got it working. Do you
> think this is an acceptable fix ?

Wait wait, what is the problem exactly ? is the device lacking the
INTERRUPT_PIN in the config space ? or is pci_device_to_OF_node()
returning NULL or is node->n_intrs == 0 ?

In the 2 later cases, that means there is something wrong with the
parsing of the OF interrupt tree in prom.c and that has to be fixed.
Once you have booted with your hack fix, can you send me a tarball
of /proc/device-tree please ?

Ben.