In article <9601168245.AA824525266@ccgate.infoworld.com>,
Brett Glass <Brett_Glass@ccgate.infoworld.com> wrote:
> In any event, how has the zs code changed in -current compared to the
> released 1.1? If it used to handle 19200 baud, there is definitely
> something wrong now.

Well, it changed just as I described: there's now a zsc "bus", to which
zstty, kbd, and ms drivers attach. (the names might not be right, but
that's how it's organized). Perhaps Gordon introcuded a bug that slowed
things down, perhaps the new organization just introduced too much
overhead, and perhaps that can be avoided somehow. I think Gordon will
have to figure it out.

One of these days, when I free up my 3/60 from doing serious work so I
can install NetBSD (not that NetBSD is necesarily unsuitable, just that
I don't want to futz too much with a machine that's important and
working just fine the way it is), I'll add in support for detecting and
enabling the larger FIFO on a Z85230 ESCC, which sould be a drop-in
replacement for the Z8530.  Shouldn't be very difficult.  Then maybe add
support for changing FIFO trigger depth (like the i386 com.c driver) to
reduce interrupt load.  Of course, I said all this back before 1.0, and
I'm STILL waiting for a new machine to replace the 3/60...  sigh :-/