Hi,

I've been trying to track down a problem where an interrupt from a
cardbus card is being routed incorrectly.  I'm running 2.6.3 with ACPI
enabled on an IBM Thinkpad T40 which has a TI1250 cardbus controller.
The cardbus controller I'm using is a PCILynx IEEE1394 controller, but
that doesn't seem to be the problem as I get the same behavior with a
OHCI IEEE1394 controller.

The problem is that when I use the second cardbus slot the device gets
assigned IRQ 9, but it generates interrupts on IRQ 5.  After a short
while the kernel disables IRQ 5 as nobody is handling the interrupt.

I've tried to print out dev->irq before calling pci_enable_device() and
at this point it's 5, but after enabling it has been set to 9.

I've attached the output of lspci -vv and dmesg.  Before loading the
driver for the cardbus device lspci reports IRQ 5, after loading the
driver lspci reports IRQ 9. One thing that looks strange to me is that
the cardbus device (07:00.0) has interrupt pin A, but the cardbus
controller that it is attached to (02:00.1) has interrupt pin B.  Or is 
it a problem with the cardbus controller not honoring the new IRQ that's 
assigned to it?

I'm booting with pci=noacpi now, which fixes the problem, but I'm ready
to help track this down, if somebody has a suggestion.  I've also
attached noacpi versions of lspci and dmesg outputs (after loading the 
driver).

Kristian