rwl@umree.UUCP (Wayne Little) (10/19/85)
We are running 4.2 BSD, and recently installed the 4.3 version uucp sporting in/out dialing on the same modem. This version was recently distributed by Tom Truscott & Bob Gray via Rick Adams March 4, 1985, according to the README. We are experiencing a side effect with the acucntrl part of uucp 4.3. The dial out part "turns off" the getty and works fine, but the getty does not get restarted on the line & I can't get it to come back (kill -HUP 1 doesn't work, and the line in /etc/ttys is enabled). The only thing that has reenabled the getty is a reboot (following a crash - we're not that desparate for lines). I can disable and reenable other lines manually (edit /etc/ttys & kill -HUP 1). I also tried another line besides tty07 to make sure it wasnt the line. I also tried physically disconnecting the line to ensure a drop of DTR. We are using a hayes modem; line from L-devices is: ACU tty07 inout 1200 hayesq The dialogue from uucico says that the line is being enabled, but no cigar and no getty thereafter (until a reboot). Any words of advice or enlightenment would be greatly appreciated. (I have another dialing scheme lined up to try if no fix is pending.) Thanks in advance!
rwl@umree.UUCP (Wayne Little) (11/01/85)
In article <106@umree.UUCP>, rwl@umree.UUCP (Wayne Little) writes: > We are experiencing a side effect with the acucntrl part of uucp 4.3. > The dial out part "turns off" the getty and works fine, but the getty does > not get restarted on the line & I can't get it to come back (kill -HUP 1 > doesn't work, and the line in /etc/ttys is enabled). The FIX involved wiring up leads 1-8 & 20 of the RS 232 cable to the modem. [We didn't have any flat 25 wire cable, & I knew the others weren't used much in the RS 232 protocol; we only had 2, 3 & 7 before - dumb, I know (now).] A hypotheis of the problem, supplied by Dave Greunke - {seismo,ihnp4,harpo}!uwvax!astroatc!vid, follows. The key may be the cabling to your modem. We have a 25 conductor flat cable going from our modems to the Able DH-11 multiplexers so all signals are brought thru to the modems (We use Ventel 212+ rackmounts). I have poked thru the acucntrl code a bit (I don't claim to understand it yet) and I know that Sebok pokes into the kernel tty structures (via /dev/kmem) and fiddles with the modem control bits of the the multiplexers. What may be happening is acucntrl may drop DTR from the mux (to reset the modem) and then expects the modem to respond (with what I'm not quite sure). If the modem doesn't respond the mux may hang. A side note is that other people have experienced the getty not being put back on the line following the acucntrl poking, UNTIL someone dialed in on the line, and then all was well. This was not the case with ours. Thanks also to Kurt Haserodt (ihnp4!wucs!kurt) who also suggested running a cable with the DTR lead in it, and to David Herron (cbosgd!ukma!david) for his reply.