PCCARD unrecogonized by FreeBSD 5.1-RELEASE on IBM Thinkpad 600

[Andrew Heybey]

"M. Warner Losh" <imp at bsdimp.com> writes:

> In message: <20040113185231.22FBC2051D at wolfe.bbn.com>
>             Greg Troxel <gdt at bbn.com> writes:
> : Probing cbb, I get a problem with "unconfigured IRQ".  Turning on
> : various PCIBIOS options (in NetBSD), I found that the IRQ in the PIR
> : was set to -1, which means that the BIOS failed to configure the
> : device properly.   I set a PCIBIOS_FIXUP_INTR option (actually many):
> : 
> : options         PCIBIOS                 # PCI BIOS support
> : options         PCIBIOSVERBOSE          # PCI BIOS verbose info
> : options         PCIBIOS_ADDR_FIXUP      # fixup PCI I/O addresses
> : options         PCIBIOS_BUS_FIXUP       # fixup PCI bus numbering
> : options         PCIBIOS_INTR_FIXUP      # fixup PCI interrupt routing
> : options         PCIBIOS_IRQS_HINT=0x0a00 # PCI interrupts hint. IRQ 9 or 11
> : options         PCIBIOS_INTR_GUESS      # see pcibios(4)
>
> Many of these are rolled up into FreeBSD's code already.  I'm working
> on not needing addr_fixup.  bus_fixup is needed.  The intr stuff is
> being worked on by John Baldwin.  The bus_fixup stuff is needed by a
> few laptops now to get cardbus working...

To chime in with my experience:  I recently installed 5.2-RELEASE on an IBM
Thinkpad 600X (2645-4EU).

I could not get PCI interrupts working correctly using ACPI.  I ran
out of time to fiddle with it and went back to APM and it is working.
(Well, pccard and cardbus cards are working; suspend still wedges the
box).

I did upgrade to the latest BIOS available from IBM (ITET55WW, 1.11).

The cardbus/pccard problems are because ACPI does not route a usable
interrupt to cbb0/cbb1.  In dmesg I get:

pcib0: slot 2 INTA routed to irq 3 via \\_SB_.LNKA

...

cbb0: <TI1450 PCI-CardBus Bridge> mem 0x50103000-0x501-3fff irq 3 at device 2.0 on pci0

cbb1 also gets irq 3.  In W98SE or FreeBSD 4.9 or 5.2 w/ apm, they get
irqs 9 & 11 respectively.

I tried setting a hint, eg (in /boot/device.hints):

hw.acpi.pci.link.0.2.0.irq=9

but it does not work because the call to 
dev/acpica/acpi_pci_link.c:acpi_pci_link_set_irq() dies because its
call to acpi_pci_get_current_irq() dies because its call to
acpi_pci_link_get_irq_resources() fails because NumberOfInterrupts ==
0 (in acpi_pci_link_get_irq_resources()).  I assume that this is
related to Greg's finding the the IRQ in the PIR was set to -1???
(Sorry I am ACPI-ignorant.)

That is as far as I got in trying to figure out what is going on.  Any
suggestions are welcome.  I do not include a full dmesg, but am happy
to provide such or any other information on request.

andrew