Greg,
        48 MB RAM
        1GB MicroDrive
        3C589 PCMCIA Network

The behavior is the same with or without the NIC and didn't seem to me to correlate
to heave CPU or disk IO.  The symtoms were consistent.   I also though it was odd
that the axises (axi?) were each mirror flipped.   I'll try the generic Kernel as
soon as I get a chance.  If other people are working, the problem must either be with
my configuation or with my hardware.

Thanks,
-Jed

Greg Hughes wrote:

> Could you give me a quick rundown of the hardware components you're using
> with your z50?
>
> >From what I've seen on my own machine the mouse should be pretty smooth in
> all but high-volume interrupt situations (e.g. network or disk IO).
>
> One of the problems with the z50 design is that they attached the
> TrackPoint to the vr4121's debug serial interface unit.  This serial unit
> only has a 1 byte buffer, so sometimes it drops bytes (overruns) if the
> interrupt handler doesn't get around to them in time.
>
> I was trying to get rid of or minimize that problem, but it looks like
> I'll have to move over to a polling design to get 100% reliability from
> the trackpoint.
>
> - Greg
>
> On Wed, 5 Sep 2001, JS wrote:
>
> > > ah! but there *is* trackpoint support on the z50 under NetBSD and X!!!!
> > > follow the link on the hpcmips port page to greg's page w/a compiled
> > > GENERIC kernel that supports the TrackPoint!!!
> >
> > Kind of- try it.   When I boot from the kernel with support (the one with LCD
> > power save, the other wasn't out when I played with it) or try to compile my
> > own, the mouse sort of works.   The mouse moves jerkily in _roughly_ the
> > direction you'd expect if you flipped right with left and up with down.   It's
> > too jerky to be usable for me.   I suppose it's possible that I have bad
> > hardware; let me know what your experience is.   I'll screw around with it some
> > more when work lets up.
> >
> > -Jed
> >
> >
> >