Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!rutgers!sri-unix!sri-spam!ames!ucbcad!ucbvax!decvax!decwrl!labrea!Lindy!vandys
From: vandys@Lindy.UUCP
Newsgroups: comp.os.minix
Subject: RS-232 on the way
Message-ID: <199@Lindy.STANFORD.EDU>
Date: Thu, 30-Apr-87 13:13:45 EDT
Article-I.D.: Lindy.199
Posted: Thu Apr 30 13:13:45 1987
Date-Received: Sat, 2-May-87 01:28:09 EDT
Organization: Stanford University Information Technology Services
Lines: 29
Keywords: RS-232, MINIX


  I have an RS-232 driver in the works, and will ship it out to all
of you as soon as it's tidied up and well-tested.  Unfortunately, I
am going on vacation for two weeks, so let's call it three weeks.

  A couple notes about RS-232 on MINIX:
1. The keyboard input routines for MINIX have the right idea, but the
  screen output routines have the *wrong* idea.  For fun, try control-S
  'ing your console output while a compile runs in background.  Notice
  how your compile stops.  TTY output is not correctly designed for
  asynchronous I/O.

2. Device-dependent/device-independent routines are not correctly
  seperated.  The most flagrant violation is echo(), which directly
  calls out_char() and flush().  In a later release I will split out
  the stuff into three files: independent, console, and RS-232.

3. do_read() correctly handles the asynchronicity of requests (in the
  book, lines 3792-3795), but do_write() does not.

  Oh, well.  It's always been my experience that TTY drivers are the
hardest thing to get right in a new OS (or even a new port of an OS).
The decision to not support RS-232 on first release, however, let the
TTY driver design stray a bit far; hopefully the next edition can correct
this.

                          Andrew Valencia
                          vandys@lindy.stanford.edu
                          br.ajv@rlg.BITNET