kevin@msa3b.UUCP (Kevin P. Kleinfelter) (07/15/89)
Does AIX have a problem with Carrier Detect?
I issue an
open ("/dev/tty0", O_RDWR, 0);
on a port with "-clocal".
As expected, if CD is active when this call is issued, the call proceeds.
Also, if CD is inactive, the call does not proceed. However, when CD
goes active, THE CALL STILL DOES NOT PROCEED.
Now I'm not really interested in issuing this call. Somewhere in
"getty" a call like this HAS to be issued. My trouble is that getty
hangs forever if CD is not active when getty is started.
I suspect that it is a problem with the task going to sleep and waiting
on an event and never getting restarted after the event. Another possibility
is that it gets swapped out, and for some reason it never gets swapped back
in.
I called IBM's AIX support center and got absolutely nowhere! They referred
me back to my local S.E. (who is NOT proficient in AIX).
Once again, I turn to the net for help.
Thanks.
P.S. If you're tired of AIX PS/2 questions in "inappropriate" groups,
you might join in the discussion on news.groups.
--
Kevin Kleinfelter @ Management Science America, Inc (404) 239-2347
gatech!nanovx!msa3b!kevinpim@ctisbv.UUCP (Pim Zandbergen) (07/17/89)
kevin@msa3b.UUCP (Kevin P. Kleinfelter) writes: >Does AIX have a problem with Carrier Detect? >I issue an > open ("/dev/tty0", O_RDWR, 0); >on a port with "-clocal". I have found out that "stty clocal" or "stty -clocal" have no effect in AIX. You select modem control *once* when you run "devices", afterwards you cannot change it. Have fun with this: Make a device with (without) modem control type on this device: $ stty clocal ($ stty -clocal) $ stty $ stty <`tty` and watch the difference! -- --------------------+----------------------+----------------------------------- Pim Zandbergen | phone: +31 70 542302 | CTI Software BV pim@ctisbv.UUCP | fax : +31 70 512837 | Laan Copes van Cattenburch 70 ...!uunet!mcvax!hp4nl!ctisbv!pim | 2585 GD The Hague, The Netherlands
karish@forel.stanford.edu (Chuck Karish) (07/17/89)
In article <786@ctisbv.UUCP> pim@ctisbv.UUCP (Pim Zandbergen) wrote: >I have found out that "stty clocal" or "stty -clocal" have no effect in AIX. >You select modem control *once* when you run "devices", afterwards >you cannot change it. All `devices' seems to do with these parameters is to edit the appropriate entry in /etc/ports. It would seem that parameters not offered for change in `devices' could be changed there, but this doesn't work for all of them. In particular, I've noticed that istrip doesn't work (AIX 2.2.1). Chuck Karish {decwrl,hpda}!mindcrf!karish (415) 493-7277 karish@forel.stanford.edu
pim@ctisbv.UUCP (Pim Zandbergen) (07/18/89)
karish@forel.stanford.edu (Chuck Karish) writes: >In article <786@ctisbv.UUCP> I, pim@ctisbv.UUCP (Pim Zandbergen) wrote: >>I have found out that "stty clocal" or "stty -clocal" have no effect in AIX. >>You select modem control *once* when you run "devices", afterwards >>you cannot change it. >All `devices' seems to do with these parameters is to edit the >appropriate entry in /etc/ports. No, there is nothing in /etc/ports that reflects the modem control of a device. I suppose modem control is handled in VRM, and cannot be changed in AIX. -- --------------------+----------------------+----------------------------------- Pim Zandbergen | phone: +31 70 542302 | CTI Software BV pim@ctisbv.UUCP | fax : +31 70 512837 | Laan Copes van Cattenburch 70 ...!uunet!mcvax!hp4nl!ctisbv!pim | 2585 GD The Hague, The Netherlands
kevin@msa3b.UUCP (Kevin P. Kleinfelter) (07/18/89)
In article <3680@portia.Stanford.EDU>, karish@forel.stanford.edu (Chuck Karish) writes: > In article <786@ctisbv.UUCP> pim@ctisbv.UUCP (Pim Zandbergen) wrote: > >I have found out that "stty clocal" or "stty -clocal" have no effect in AIX. > >You select modem control *once* when you run "devices", afterwards > >you cannot change it. > > All `devices' seems to do with these parameters is to edit the > appropriate entry in /etc/ports. It would seem that parameters not Not quite true. It also does a "mknod c /dev/tty0 5 96" for tty0. The minor device number contains a bit mask which determines modem control. In order to get modem control, you must have the correct minor device number AND -clocal. As near as I can tell, when processing tty devices, "devices updates /dev/tty? /etc/ports /etc/inittab I see no reason why these things cannot ALL be done manually if desired. Am I wrong? -- Kevin Kleinfelter @ Management Science America, Inc (404) 239-2347 gatech!nanovx!msa3b!kevin
karish@forel.stanford.edu (Chuck Karish) (07/21/89)
In article <671@msa3b.UUCP> kevin@msa3b.UUCP (Kevin P. Kleinfelter) wrote: >In article <3680@portia.Stanford.EDU>, I wrote: >> All `devices' seems to do with [stty] parameters is to edit the >> appropriate entry in /etc/ports. It would seem that parameters not >Not quite true. It also does a "mknod c /dev/tty0 5 96" for tty0. >The minor device number contains a bit mask which determines modem control. >In order to get modem control, you must have the correct minor device number >AND -clocal. I discovered today that HP does it with the minor device number, too. Both HP and IBM hide the documentation for this feature pretty effectively. The AIX line protocols arent't discussed comprehensibely anywhere; to figure out what's going on, you have to read the system management guide and the VRM device driver reference. >As near as I can tell, when processing tty devices, "devices updates > /dev/tty? > /etc/ports > /etc/inittab >I see no reason why these things cannot ALL be done manually if desired. >Am I wrong? I think you're right. I've been successful changing both disk partition and terminal parameters by hand. The policy of making system administration magic, and hiding the details from the administrator, is annoying. Especially when the tools provided don't do enough. Try to re-arrange your disk partitions using minidisks! I've learned to use the VRM maintenance facility, and to edit /etc/system and /etc/filesystems by hand. Changing simple options by editing /etc/ports works OK. It's also possible to change some stty options that aren't offered by `devices'. Chuck Karish {decwrl,hpda}!mindcrf!karish (415) 493-7277 karish@forel.stanford.edu