After 3 hours of pounding at dz.c to make it fit neatly into what
we have in the 2.1.100 release - I actually got a kernel that looks
like it talks to the serial port for the console... I got the
configuration stuff set up so that it works as well. Resolved the name
clashes between promcon.c,tc.c and dz.c.
I've yet to pony up the 30 or 40 dollars it would take to have a cable
made - and my friend hasn't come through on sending me the mmj serial
cable that I was hoping for.... The end result of it is that its
a bit hard to actually tell if its working correctly. Tomorrow I plan
on dragging myself into the cable place and forking over the dollars...
Anyway - the questions about io.h were answered in a comment by
Ralf at the top of io.h - that is his macros will not work in inline
functions - the rz.c driver has inline functions for the core
functions. I think that answers the question. Chalk it up to not
enough research before asking the question.
I'm going to pound at tz.c for another couple of hours this evening and
see if I can clean it up even more.
Given enough eyeballs all bugs seem shallow.
On Tue, 1 Sep 1998, Harald Koerfgen wrote:
> Hi Thomas,
> On 01-Sep-98 Thomas Riemer wrote:
> > I've run across a reference to outw and inw - that
> > has a comment indicating it was supposed to replace
> > outw and inw....
> > I'm guessing it was written prior to work done on asm-mips/io.h.
> > Can anyone tell me the state of outw and inw in asm-mips/io.h
> > and perhaps the history of when it was worked on?
> AFAIK the macros in asm-mips/io.h are only needed for MIPS machines with an
> bus. They are intended to help using device drivers intentionally written for
> Intel based boxes.
> As far as I can tell from my private communication with him, Olivier used
> drivers/char/serial.c as a template and the inw outw instructions may be
> The serial device driver for the DC7085 is supposed to work only on MIPS based
> non-ISA machines (unless the Linux/VAX port comes to new life), so it would be
> cleaner to kick them out and work with pointers.
> BTW does anybody know if the VAX processors have memory mapped I/O too?
> Happy hacking.