home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!ogicse!das-news.harvard.edu!spdcc!rbraun
- From: rbraun@spdcc.com (Rich Braun)
- Newsgroups: comp.os.linux
- Subject: Serial driver issues
- Message-ID: <1992Dec20.051300.1475@spdcc.com>
- Date: 20 Dec 92 05:13:00 GMT
- Article-I.D.: spdcc.1992Dec20.051300.1475
- Organization: Jupiter Technology Inc., Waltham MA
- Lines: 35
-
- I've been playing around with Linux 0.98p5, trying to get it to communicate
- with two serial printers, an internal modem, and a terminal. At the
- moment, I'm typing this posting through the modem on /dev/ttys1 (with
- the ATS25=50 setting I mentioned in an earlier posting, to deal with
- momentary DTR lossage), and there's an interesting conflict with the
- interactive session on /dev/ttys3.
-
- COM1 and COM3 use the same IRQ vector, and what I see is that output
- sent to ttys3 out at a *very* slow rate (like 30cps), and no input for
- ttys3 is handled until after a byte is fetched on ttys1. (Example,
- I type 'pwd' on the other session, and nothing echoes. I then type
- another character into this posting, a 'p' shows up on the other session,
- and the 'wd' are discarded.)
-
- This may be a problem with misconfigured hardware (COM3 might be
- jumpered wrong) or a bug in Linux' serial driver. Any theories? Has
- anyone else run this kind of test?
-
- Along similar lines, I notice that XON/XOFF flow control does not work
- on this port. Even if I set the stop char to something other than
- ^S, output won't stop when I type it. Haven't tried other ports.
- This means my serial printer won't work as it did under DOS.
-
- Finally, the getty program appears to alter the clocal setting and a
- few others for the serial device. I can't get it to work at all with
- this serial port, since getty sees a hangup signal immediately. I've
- had to make do with the 'doshell' program. Is there any way to make
- getty not do this? Do I need to patch the source? Is this considered
- a bug or a deliberate design?
-
- If there are actually bugs here, as opposed to mere operator confusion,
- I'd like to try fixing them prior to the 1.0 release.
-
- Thanks!
- -rich
-