> > >But not with a fresh enough -current to expose the problem, I
> > >guess.
> >
> > Or maybe I'm just really lucky?
>
> Is the bug supposed to affect all cards that use the ep driver, or just
> the 3c509? I've got a 3c595 that seems to be working just fine... July 21
> srcs, kernel compiled with -O2.

I saw the problem with a 3c595 with sources a little older (early
July IIRC).  It's the common backend which exposes the problem
(elink3.c).

- H=E5vard