Todd Whitesel <toddpw@best.com> writes:
> Well, I was thinking of switching this guy to 1.4.2 also, I don't suppose
> you've pulled up the nextdma.c fix yet? My slab runs a darn sight better
> in spite of this latest crash, and I'd certainly like to do 1.4.2 builds
> (right now my next68k is the only build machine I have that _hasn't_ gone
> back to 1.4.2_ALPHA).
The nextdma.c fix was pulled up last night.

> Hm, can we throw a specific panic in to detect that case, so when it happens
> again we might be able to confirm the theory?

I could, except this is pretty much what the recently applied patch
disabled.  The applied patch also changed the bits it was setting in
the registers slightly, so the restart behavior may be a little
different now.  I need to sit down and think about my model of the dma
chip operation again and see if I can find the holes.

Darrin