Well here's what I can see on my screen since networking isn't coming  
up most of the time and I can't type anything without USB:

uhub_explore:usb_new_device failed, error=SET_ADDR_FAILED
uhub1: device problem, disabling port 1
wd1 at atabus1 drive 0: <Maxtor 6Y160M0>
wd1 blah
usb_new_device: set address 2 failed
uhub_explore: usb_new_device failed, error=SET_ADDR_FAILED
uhub0: device problem, disabling port 1
ehci_sync_hc: tsleep() = 35
usb_new_device: set address 2 failed
ehci_sync_hc: tsleep() = 35
uhub_explore: usb_new_device failed, error=SET_ADDR_FAILED
uhub4: device problem, disabling port 4
Searching for RAID components...
...

I can hook up a serial console and try to debug over that, but I'm  
not sure how much I'm going to be able to do while the kernel is  
booting. Would modloading the usb system after boot be a better  
alternative?

I know people want to play the bad hardware card but I ran the Dell  
Diagnostics, MPmemtest, Memtest86+, and dual boot this with windows  
and it works flawlessly in windows.

On Oct 7, 2005, at 9:11 PM, Andreas Wrede wrote:
> On Oct 7, 2005, at 20:30 , Chris Tribo wrote:
>> I just came across a bug report on OpenBSD PR 4455 that describes  
>> the same
>> problem I'm having with EHCI. Except that it's on an nVidia AMD64  
>> board.
>> I'll try his suggestion of running a non MP kernel, but it still  
>> sounds
>> like something is wrong in APIC interrupt handler/PCI Express land.
>
> I filed kern/31457 a couple of days ago. Maybe it is related, since  
> the problem occurs with a dual CPU system and an APCI enabled  
> kernel if a bge sharing an IRQ with a isp interface is accessed?