Controlling ttys, and how to get rid of them.

[Dave Horsfall]

In article <176 at ucms.ucms.uucp> dave at ucms.UUCP (Dave Settle) writes:
| 
| I'm having intermittent problems with 'login' on System V machines. It's
| happened on a number of machines, ranging from AT&T 3B2's to more
| flavoursome versions of SVR2.
| 
[ Details suppressed ]

This is caused by a process doing a setpgrp() to detach itself from the
controlling TTY, and the kernel not getting it right.  Subsequently, "login"
attempts to disable echo with ioctl(), and fails.

A work-around is a liberal application of "fuser -k /dev/tty??" to kill it.
If that fails, there's a little switch on most machines...

Complain to your vendor, and if possible supply a little program that
demonstrates it.  We had the problem with CCI 6/32 Unix V for a long time,
and it finally got fixed in 2.22.02.

-- 
Dave Horsfall (VK2KFU),  Alcatel-STC Australia,  dave at stcns3.stc.oz
dave%stcns3.stc.oz.AU at uunet.UU.NET,  ...munnari!stcns3.stc.oz.AU!dave
    PCs haven't changed computing history - merely repeated it