era@ncar.ucar.edu (Ed Arnold) (03/23/91)
In sun-spots v10n7, I wrote with regard to usage of an ALM-1 on a 4/490 running 4.1: |If you do decide to run it on your 4/400, keep in mind that the login |shells on ALM-1 ports will become a security hole, cuz they'll never |receive hangup signals. I subsequently received the following reply from a member of the Sun technical staff: |This is not true. See mti(4S) for how to wire your ALM-1 lines to |handle carrier properly. See ttysoftcar(8) for how to set up the |system to pay attention to carrier. On a 4/490 running 4.1, which has a single ALM-1, we have done the following: 1) We have set the "carrier detection mode" flag of all ALM-1 ttys to something other than "local". (It has been set to "foobar", which seems eminently appropriate in this case. :-)) "ttysoftcar" is executed from our rc script. 2) The ALM-1 is fed by a Gandalf PACX serial port switcher. We have OR-wired its DCD line to the ALM-1's CTS, DSR, and DCD pins, presumably satisfying this paragraph from the WIRING section of the 4.1 MTI(4S) man page: |The Systech requires the CTS modem control signal to |operate. If the device does not supply CTS then RTS should |be jumpered to CTS at the distribution panel (short pins 4 |to 5). Also, the CD (carrier detect) line does not work |properly. When connecting a modem, the modem's CD line |should be wired to DSR, which the software will treat as |carrier detect. (It is interesting to note that the same wired-OR arrangment is used on Sun 3/280s and 4/280s with ALM-1s here, running 4.0.3, on which it works perfectly.) I am interested in receiving e-mail from three classes of people (will summarize of course): 1) Users who have been able to make the ALM-1 work on any 4/400, i.e. it will generate a SIGHUP to a login shell when CTS, DSR, and DCD go away. 2) Users who have tried every combination under the Sun, proving to themselves that it can't be made to work. Thanks in advance, era@ncar.ucar.edu