a new (for me) panic: usbtask

[Mikhail T.]

Fresh -current (October 10):

 Fatal trap 12: ...
 fault virtual address = 0x4c
 [...]
 current process = 46 (usbtask)
 [thread 100049]
 Stopped at usb_transfer_complete+0xe7: movl 0x4c(%eax),%eax 

Dual Pentium2 Xeon, 1.5Gb of RAM; two big compiles in parallel where
humming along, when a print request was submitted to the USB printer.

Upon restart, the system started to print. As I went on to restart one
of the compiles, however, it promptly paniced again -- in the same place.

When I did 'call boot(0)', the machine successfully stopped 'vnlru', but
hung waiting for 'bufdaemon' to stop...

	-mi