Subject: Re: port-alpha/21472 and kern/16379 (ohci/USB problems on Alpha)
To: NetBSD/alpha Discussion List <port-alpha@NetBSD.org>
From: Nathan J. Williams <email@example.com>
Date: 09/23/2004 17:50:00
"Greg A. Woods" <firstname.lastname@example.org> writes:
> I find the following in sys/arch/alpha/pci/pci_6600.c:
> revision 1.2
> date: 2000/03/19 02:25:29; author: thorpej; state: Exp; lines: +69 -26
> branches: 1.2.2;
> Add support for mapping the OHCI USB controller interrupt (which is wired toan ISA IRQ because it's in the same package as the PCI-ISA bridge).
> However on the ES40 the device probes as follows:
> ohci0 at pci0 dev 19 function 0: Acer Labs M5237 USB Host Controller (rev. 0x03)
> ohci0: interrupting at isa irq 10
> ohci0: OHCI version 1.0, legacy support
> So is it really on an ISA IRQ when it probes as a PCI device?
Probably, yes. There's no such thing as a "PCI interrupt controller"
at this level; the system can wire up the interrupt pins from the PCI
bus to whatever interrupt controller it likes, and on systems with ISA
bridges (usually just "legacy" keyboard/mouse/serial/parallel/floppy),
it's common to wire them up to the ISA interrupt controller that's
included as part of that package.
> According to the SRM's "show config" the only thing that actually seems
> to be on the ISA bus is the floppy (of the things it knows about):
It's not really "on the ISA bus", just using its interrupt
controller. Not that SRM can safely be considered authoritative for
ISA devices anyway.
> Does this suggest that Nathan's patches for USB on PWS500au to
> pci_550.c still also need to be made to pci_6600.c?
Er, which patch was that?
> Shall I build a PCI_CONFIG_DUMP kernel?
You can get everything you need from userland with "pcictl /dev/pci0
dump -d 1 -f 2", with appropriate values for 0, 1, and 2.