Strange problem: if_xe only works in GENERIC kernel

[Joerg Wunsch]

As Joerg Wunsch wrote:

> OK, at kernel #11 :), I can now say it's the USB subsystem.  Just
> leaving "device usb" (and also "device uhci") in makes it work.
> 
> So the question appears to be why keeping the USB driver in makes the
> interrupt storm detection work...

Maybe that's the relationship?

camel# dmesg | fgrep 'irq 11'
vgapci0: <VGA-compatible display> mem 0xe0000000-0xe0ffffff,0x70000000-0x703fffff,0x70400000-0x704fffff irq 11 at device 0.0 on pci1
cbb0: <TI1251 PCI-CardBus Bridge> mem 0x50102000-0x50102fff irq 11 at device 2.0 on pci0
cbb1: <TI1251 PCI-CardBus Bridge> mem 0x50101000-0x50101fff irq 11 at device 2.1 on pci0
uhci0: <Intel 82371AB/EB (PIIX4) USB controller> port 0x8400-0x841f irq 11 at device 7.2 on pci0
xe0: <Intel EtherExpress(TM) PRO/100 PC Card Mobile Adapter16> at port 0x100-0x10f iomem 0x20000000-0x20000fff irq 11 function 0 config 1 on pccard1

I guess vgapci0 doesn't really use interrupts, so this leaves cbb0/1
and uhci0 sharing an interrupt.  Apparently, the interrupt storm at
cbb gets detected correctly as long as at least another device
installs an interrupt handler on irq 11.

Does that make any sense as an explanation?

I can live with the current situation (and proceed in setting up that
machine as a firewall, which was my original intention), although I
could also spend another day into debugging that symptom if someone
can get me some directions.
-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)