cvs commit: src/sys/dev/usb ehci.c ehci_pci.c ehcivar.h

[Ian Dowse]

In message <20060129235559.GA29869 at saturn.kn-bremen.de>, Juergen Lock writes:
>Hmm.  Now I just got the following panic with it, pipe->queue.stqh_first
>shouldn't be NULL right?  (Maybe it already got a new command for the
>same umass device while there still is an ehci_intrlist_timeout pending
>for the last command and so the new one interferes?)

Is that with umass devices only, or were there other USB devices
in use? The trace looks familiour so it might be something that has
been fixed in -current. Could you see if there were any related
console messages just before the crash with

	dmesg -a -M vmcore_file -N kernel_file

and if possible recompile your kernel to contain the drivers for
any USB devices you use rather than loading modules? Using modules
causes the stack trace to miss out information from module frames.

Ian