marks@aivax.radc.af.mil (David Marks) (04/04/91)
I want to connect 3 modems (actually NIU ports, but they LOOK and ACT like modems) to our 4D/70GT (soon to be a 4D/340VGX). We have the 6 serial port expansion interface, but those ports don't support DTR for detecting presence of a connection; an attribute we require. Ports 1-4 support DTR, but 2 and 3 are dedicated to the Buttons & Dials Box and the SpaceBall. Questions: 1. Can the Buttons & Dials Box and SpaceBall be moved over to other ports (5-10) which don't support DTR? 2. Does the 6 serial port expansion interface for the 3xxVGX support DTR? 3. Does it cost more than $5.00 of parts to support DTR (and the like) on a serial port? (If not, why not provide that support on ALL serial ports, or at least make it an option?!) 4. Is there any reason to put a dumb terminal console on port 1 versus any other port? 5. Can stty be used to configure a port for RTS/CTS hardware handshaking? I've seen the net traffic on how to get it through a program, but I'd like to avoid writing stty2.c ... Thanks Dave Marks Rome Laboratory marks@aivax.radc.af.mil
Dan Karron@UCBVAX.BERKELEY.EDU (04/04/91)
Dave Marks (marks@aivax.radc.af.mil) writes: >I want to connect 3 modems (actually NIU ports, but they LOOK and ACT like >modems) to our 4D/70GT (soon to be a 4D/340VGX). We have the 6 serial port >expansion interface, but those ports don't support DTR for detecting presence >of a connection; an attribute we require. Ports 1-4 support DTR, but 2 and 3 >are dedicated to the Buttons & Dials Box and the SpaceBall. I had a 70G machine, and I maintain that serial io is not right on that class machine. I had to move to a newer machine (85) for all the serial io problems to clear up. My advice is to wait until you get your 340VGX machine. The design intent is that all the serial ports are wired the same, but that is not the case for the 70 machines, despite what all the fe's say. If you are using Ungermann-Bass NIU's, you are in for more grief. I am surprised that the serial expansion ports don't support hardware handshake. You need it for fast (>9600 baud) io. Cheers! dan. | karron@nyu.edu (e-mail alias ) Dan Karron, Research Associate | | Phone: 212 263 5210 Fax: 212 263 7190 New York University Medical Center | | 560 First Avenue Digital Pager <1> (212) 397 9330 | | New York, New York 10016 <2> 10896 <3> <your-number-here> |
vjs@rhyolite.wpd.sgi.com (Vernon Schryver) (04/04/91)
In article <9104031623.AA18586@AIVAX.RADC.AF.MIL>, marks@aivax.radc.af.mil (David Marks) writes: > I want to connect 3 modems (actually NIU ports, but they LOOK and ACT like > modems) to our 4D/70GT (soon to be a 4D/340VGX). We have the 6 serial port > expansion interface, but those ports don't support DTR for detecting presence > of a connection; an attribute we require.... On the contrary, the 6-port serial board supports a full 7-wire interface, including pins 2,3,4,5,7,8, and 9, or RD,TD,RTS,CTS,SG,DCD, and DTR. Please note that an IRIS is a DTE, and as such raises DTR just in case it is prepared to talk. DCD is the signal that standards say that a DCE uses to tell a DTE that the modem (DCE) is interested in passing data, or what might reasonably be called the "presence of a connection." As described in Chapter 10 of "IRIX System Administrator's Guide", doc. no. 007-0603-030, - DCD/DTR hand shaking is available only when using the ttym* or ttyf* device name. - RTS/CTS non-standard but nearly universal "hardware flow control" is available only when using the ttyf* device name. That manual describes the behavior of all 7 signals with the three device names. As described in /usr/sysgen/master.d/cdsio, there is an old error in the cables used to connect the 6-port board to the connectors on the outside of the machine. That error confused DSR and DCD. If you do not require hardware flow control on any of the 6-ports, then the "old" cables can be used. Otherwise, it might be possible to get SGI Field Service to fix the cables as described in the file. Recently shipped 6-port boards for machines with 9-U VME buses should come with fixed cables. Obviously, one could work around the error in the cable between the computer and the DCE. The "integral" serial ports and the 6-port serial board ports act the same, modulo bugs. More to the point, there are no such bugs known concerning DCD,DTR,RTS, or RTS, except for the cabling error mentioned above. There is no programatic way to turn RTS/CTS flow control on or off, except by choosing the correct minor device number, usually by choosing a suitable name in /dev. A "dumb console," as opposed to "the graphics console," must be on port 1. That is where the PROM and a host of other things expect to find it. Vernon Schryver, vjs@sgi.com.