re: kern/54023: assertion panic during shutdown

[matthew green <mrg%eterna.com.au@localhost>]

not particular thoughts on the assert itself (which is fairly
new -- only added last year), however...

> 	It then locks up hard trying to dump core, and has to be
> 	powered off.

this part is probably related to the fact USB crashed and the
dump device is behind USB.  it may be that USB dumps don't
work at all... i will see about testing this situation.


.mrg.

ps. actually, there is a known issue in usb abort still that
we haven't finished patching, so it may be related to that,
though the race is very hard now, compared to the old ways
these bugs could be triggered.