nghiem@ut-emx.UUCP (Alex Nghiem) (10/24/88)
In article <7139@dasys1.UUCP>, alexis@dasys1.UUCP (Alexis Rosen) writes: > > Has anyone gotten Hayes-compatible modems working as dial-in devices? My > (possibly flawed) understanding is that 2400 bps models won't work right. > I believe this problem is endemic to all SysV based systems. If you have > a dial-in line working, what modem are you using? > We have got AUX to work with Microcomm 2400c and 9600+ in AT dumb mode. Apple is working on problems with running the modems in AT smart mode. AUX does not work properly with the Everex 2400e, probably the same problem as above. I understand that AUX dialup will work with Hayes modems. We have found that if carrier is lost for any reason, ie someone trips over the phone cord, the line is hit by noise, there is a power failure at the remote end, etc. AUX will not kill the dialup shell process. Anybody can call back and get into the system. Apple.com is working on this problem. I have since learned that this may be a SysV problem and not necessarily just an AUX problem. This article is posted for factual information only. Any misrepresentation, if any, is purely unintentional. Any opinion expressed or implicit in these remarks are solely my own. nghiem@emx.utexas.edu "Why?...Because we LIKE you!!!"
alexis@ccnysci.UUCP (Alexis Rosen) (10/25/88)
In article <7216@ut-emx.UUCP> nghiem@ut-emx.UUCP (Alex Nghiem) writes: >In article <7139@dasys1.UUCP>, alexis@dasys1.UUCP (Alexis Rosen) writes: >> Has anyone gotten Hayes-compatible modems working as dial-in devices? My >> (possibly flawed) understanding is that 2400 bps models won't work right. >> I believe this problem is endemic to all SysV based systems. If you have >> a dial-in line working, what modem are you using? > >We have got AUX to work with Microcomm 2400c and 9600+ in AT dumb mode. >Apple is working on problems with running the modems in AT smart mode. AUX >does not work properly with the Everex 2400e, probably the same problem as >above. I understand that AUX dialup will work with Hayes modems. Please clarify this. Does the Everex fail in the same way as the MicroCom, or all the time? Also, could you briefly discuss the "AT dumb mode"? My hayes modem hacking is limited to getting the damn thing running with my terminal emulator... How do you 'understand' that the Hayes modems work? Have you used them, or heard reports? If the latter, can you tell us who it is that has the Hayes modems running? >We have found that if carrier is lost for any reason, ie someone >trips over the phone cord, the line is hit by noise, there is a power >failure at the remote end, etc. AUX will not kill the dialup shell process. >Anybody can call back and get into the system. Apple.com is working >on this problem. I have since learned that this may be a SysV problem and >not necessarily just an AUX problem. If this is true then you are ahead of the game. On some other System V systems, you have to kill the modem BY HAND to get it to work again. There has been much talk about this in various newsgroups (notably pubnet.something) but no definitive conclusion was reached, other than that Hayes-type modems were risky business for SysV machines. There are other problems. In some cases, if you call in on a 2400 bps modem at 1200, it won't reset the speed. This can cause major problems if your news feed is trying to call at 2400. Note that all this is not rumor. I use a system daily that suffers from the problems I mentioned. It is NOT an A/UX machine, however. If I knew for a fact that A/UX were susceptible or immune wrt these problems, I wouldn't be posting... Does anyone have a non-Hayes-compatible, non-Telebit modem working under A/UX for DIAL-IN use? Sigh... ---- Alexis Rosen alexis@dasys1.UUCP or alexis@ccnysci.UUCP Writing from {allegra,philabs,cmcl2}!phri\ The Big Electric Cat uunet!dasys1!alexis Public UNIX {portal,well,sun}!hoptoad/
wmd@tippy.uucp (10/26/88)
I am currently using three Hayes 2400's somewhat successfully with my A/UX
box. The following problems occurred and I will attempt to describe my
kludge to get them working.
The modem settings:
The modems are set to reset on DTR drop, set DCD according to the
state of the carrier, command mode disabled, and numeric result
codes.
The A/UX box settings:
getty to respawn using the mo_1200 gettydef label. mo_1200 was modified
to try mo_2400 after a break, then mo_300, and then back to mo_1200.
The problem:
If someone calls in connects and hangs up before logging in, getty
dies and is not respawned by init. Modem control does not take effect
until login occurs. That is, a loss of carrier does not result in
a termination of processes and respawn of getty until a successful
login has occurred.
The Solutions attempted:
I tried rewriting getty but was unsuccessfull in getting my implementation
to work. I sent multiple Applelinks to Apple on the problem. They must
not read their Applelink mail in the A/UX group. I offerred to write a
special implementation of getty for them (for free!) if they would give
me the source to their getty under non-disclosure... not a word from them.
The Kludge at it now stands:
There is a cron demon running each minute saying '/etc/telinit q' which
forces init to re-examine the inittab and respawn the getty processes
which have failed to respawn. Users still need to generate a break to
cause getty to attempt the next baud rate. It works, I don't like it,
I never will, but I have seen no alternatives from Apple nor even a response
to my queries. There appears to be a real problem using Hayes modems
with A/UX and someone needs to offer a solution...I thought I'd outline
mine. Hope it helps.
====----====----====----====----====----====----====----====----====----====
Malcolm Duncan Duncan Communications Public Access Usenet System
Internet: tippy!root@newton.physics.purdue.edu (317) 567-2143
US Mail: RR #1 Box 98E, Battle Ground, IN 47920
Disclaimer: Duncan Communications is a private for-profit (ha!) organization.
All opinions expressed here are my own and those sharing them
may send me royalty checks at the beginning of each year. Mail
them to the above address. ($50 per opinion, $10 per resentment)