Return to BSD News archive
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!bunyip.cc.uq.oz.au!munnari.oz.au!news.Hawaii.Edu!ames!hookup!swrinde!cs.utexas.edu!howland.reston.ans.net!europa.eng.gtefsd.com!MathWorks.Com!news2.near.net!bloom-beacon.mit.edu!usenet.elf.com!sundog.tiac.net!sundog.tiac.net!rick From: rick@vox.trystero.com (Richard E. Nickle) Newsgroups: comp.os.386bsd.misc Subject: Re: serial port speed Date: 08 Aug 1994 20:45:04 GMT Organization: The Internet Access Company Lines: 49 Message-ID: <RICK.94Aug8164510@vox.trystero.com> References: <31t958$q2h@sundog.tiac.net> NNTP-Posting-Host: vox.trystero.com In-reply-to: bill@bhhome.ci.net's message of 5 Aug 1994 11:52:08 GMT In article <31t958$q2h@sundog.tiac.net> bill@bhhome.ci.net (Bill Heiser) writes: > I posted this once before but I guess it never made it out. > I am running FreeBSD 1.1.5.1 (RELEASE) on a 486/33 with 8mb. > I am trying to use tip to talk to another machine connected > by null modem ... and if I try to communicate above 19200, I > just get garbage when trying to connect with TIP. Maybe you should use 'cu' instead. TIP seems too difficult to fathom for me. I do: cu -l /dev/tty00 -s 38400 It saves me from having to screw around with all the config files that tip wants, and I'm too lazy to read about. What the hell *IS* the advantage of tip anyway? From what I see it's a bastardized version of cu? Anyone? > Is there something special I need to do to get higher comms rates? > I'd expect to be able to do at least 38400 if not 115200! I talk at 57600 all the time. No problems. > I am trying to connect to a LINUX box on which I have a getty running > at 38400. Aha! Easy solution. Blame Linux. Post a hundred page diatribe to the net about how rotten this loose cannon REALLY IS. Help kill it before it pollutes the world. First off, I'd grab an external modem that can do 38400 and try talking to that with Linux, then try to do the same thing with FreeBSD. Figure out if either end is obviously screwed up somehow. Do you have 16550AFNs? Or those buggy early versions? Or 16450s? I'd consider attacking the problem next by checking your line. I've had serious serial communications problems at high speed with lines that: - were too long (that 50' limit is REAL, and the faster you go, the shorter the cable has to be to be reliable) - didn't have all the flow control lines connected - had intermittant shorts on one end or the other I ended up solving these problems by switching to coaxial ethernet to make long runs between machines. :) :) :) Rick