> Yes, that seems plausible, after reading David Forbes' message. I don't
> think my crash has anything to do with yours. You aren't running usbd by any
> chance are you?

Aye Carumba!! USB in -current was broken for a few days this week. See
http://mail-index.netbsd.org/mlist/current-users/1998/12/12/0001.html
and get a fresh SUP. Uninitialized struct pointers, very bad.

The worst part is, the old gcc flagged them down, but EGCS did not. Brrr.

> Snapshot? Oh goody. I was getting worried that I'd be stuck with my old,
> obsolete (at least it sure feels old considering all the src commits I've
> seen flying by on the lists) October kernel for ship date in January. =)

Hmm, do you have 100 megs of FTP space that could host it when I'm done?
Source is about 68 megs, the tarballs should be another fifty or so.

> BTW, if anyone is interested in robotics, the project I'm doing this for is
> at http://web.mit.edu/6.270/ - I'm an organizer for the MIT 6.270 contest.
> As of Friday, I have NetBSD/arm32 running on a prototype EBSA285-like board,
> except that this board is about the size of an index card and will be used
> to control robots in January...

Cool! You know what I'd really love: a cheap PCI card that has the guts of an
EBSA285 on it with a single right-angle DIMM slot (so the card only takes a
single PCI slot) and 10BT plus fcom0 on the back panel. I'd buy as many as I
could afford, stick them into every empty PCI slot in my CATS and i386 boxes,
hook them all up to the 8-port hub, and boot them diskless off of one of the
other machines and use them as compile engines or something. Eventually it
would be better to have them use PCI bus-mastering to show up as point-to-
point interfaces on the motherboard kernel's ifconfig list. After all, that's
pretty much what the EBSA285 was originally designed to do, as a foundation
for I2O products...

"We don't need no steenkin' kernel threads. I got your SMP right here!"

Todd Whitesel
toddpw @ best.com