Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: Notesfiles $Revision: 1.6.2.17 $; site smu.UUCP Path: utzoo!watmath!clyde!cbosgd!ihnp4!inuxc!pur-ee!uiucdcs!smu!pedz From: pedz@smu.UUCP Newsgroups: net.unix-wizards Subject: Autobaud Modems Message-ID: <32300009@smu.UUCP> Date: Fri, 7-Dec-84 03:22:00 EST Article-I.D.: smu.32300009 Posted: Fri Dec 7 03:22:00 1984 Date-Received: Tue, 11-Dec-84 04:19:36 EST Lines: 44 Nf-ID: #N:smu:32300009:000:2086 Nf-From: smu!pedz Dec 7 02:22:00 1984 Some time ago there was a lot of debate about auto-baud modems and different ways to do them. The solutions were all done totally in software. This solution uses hardware to aid the software and the only problems with it is that it can only be used at two speeds. Typically the speeds are 1200 and 300. The concept is quite simple. The ring detect line is not used for anything usefull (at least in the DH driver on a 4.2 system). Most autobaud modems have a speed line availiable on the RS-232 connector. So it is simply a matter of connecting the speed line of the modem to the ring detect line of the DH. The software that needs to be changed is the DH driver to add ioctl's to get the status of the ring detect line and also the carrier detect line (if you do one you might was well do the other). These calls were never implemented in the 4.2 DH driver. Second is to setup a special entry in gettytab which getty will treat special (I choose "m" for modem). With this entry, getty first opens the line (the soft carrier detect option must be off for this to work). The open will block until carrier is detected. At this point, the speed line is set so it is simply a matter of doing the ioctl call to get the status of the line. Care must be taken in this step since there seems to be a significant time delay in the setting of this line. I put in a sleep for 1 second to solve the problem. When the line is sampled, the speed of the line is determined to be 1200 if the line is high and 300 if the line is low. The baud rate is set and normal processing resumes. Obviously this system assumes some things about the hardware being used. It also is based upon a 4.2 system and I am not sure what other system do with the ring detect line. It does however work very very nicely. No more wacking on the break or the @ key and then hitting four or five returns, then repeating until something comes to life. Diffs are offered to anyone interrested. You need to tell me which diffs (the DH driver, getty, or both) you need. Perry convex!smu!pedz pedz@smu (csnet)