On Mon, 2007-10-01 at 16:48 -0500, Scott Wood wrote: > Benjamin Herrenschmidt wrote: > > On Mon, 2007-10-01 at 16:37 -0500, Scott Wood wrote: > >> Scott Wood wrote: > >> Actually, it doesn't -- it should stop when it sees the > >> interrupt-controller property in the i8259 node, at which point it'll be > >> trying to use the raw PCI IRQ pin number as an i8259 IRQ. This is > >> Unlikely To Work(tm). > > > > It will work in the specific 8259 case I suppose since it gets the > > legacy 1:1 mapping... but it sucks :-) > > The mapping between INTA-D and i8259 numbers isn't generally 1:1, and it > looked as if it'd try using the former... though the code is > sufficiently complicated that I could easily be missing something.
If the whole of_* thing totally fails, pci_read_irq_line should pickup the value in the config space PCI_INTERRUPT_LINE. That will work on things like Pegasos and possibly on this AmigaOne but it sucks. Ben. _______________________________________________ Linuxppc-dev mailing list Linuxppc-dev@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-dev