Quoting Izumi Tsutsui <tsutsui@ceres.dti.ne.jp>:

> Does anyone test zs mouse driver on IP12/IP20?

I never have. When I wrote it the grtwo driver may have been incomplete, plus my
Indigo was broken, and in any event we don't have an Xserver for grtwo, making
the mouse a little less useful. The driver undoubtedly does not function
properly, as I had no idea what to do with the actual axis and button values in
terms of what wsmouse really wanted, nor how users of wsmouse are expected to
interpret the data. I guessed that they need to understand the type of mouse. In
short: the driver can safely be assumed broken.

> I'm still working on ews4800mips port and
> its mouse has the similar protocol with sgimips one.
> I just take sgimips/dev/zs_ms.c, but the zsms driver
> seems to pass wrong values to MI wsmouse(4) layer:
> - Packets for delta X and delta Y should be signed

Yes, that makes absolute sense. The IRIX mouse(7) man page says: 'Positive
values indicate movement to the right or upward.'. I don't know if this matches
wsmouse(4).

I do not know, either, if the values need to be tweaked by different function or
if they're okay as-is. I mean: I don't know if the relationship between x1 and
x2 (also y1 and y2) are simply additive. That was just a guess. I guess that
your test program will be indicative of proper behaviour.

> - 1 is DOWN, and 0 is UP for buttons on wsmouse(4) 

Yes, you're right. Could you please update the wsmouse(9) manpage to reflect
this? There's only a little comment in the source.

> - The most left button is LSB on wsmouse(4)

If that is true, would you please add a note to wsmouse(9) as well?

> Is it okay to commit attached patch?

It looks alright to me.

Thanks for fixing this!
Steve

P.S.: In case you're interested, here's a URL for the IRIX mouse(7) man page:
http://techpubs.sgi.com/library/tpl/cgi-bin/getdoc.cgi?coll=0650&db=man&fname=/usr/share/catman/a_man/cat7/mouse.z&srch=mouse