[BusyBox] job control turned off, yet again...

[Jeff "Tiny" Warren]

Vernon,

Thanks for the advice.  Your 2nd item in particular helped me out
significantly.  I looked at the stat file for ash and init when I get
the warning and this is what I see..

pid | name | state | ppid | pgrp | session | tty | tpgid
--------------------------------------------------------
34  | ash  | S     | 1    | 34   | 34      |  0  | -1
--------------------------------------------------------
1   | init | S     | 0    | 0    | 0       |  0  | -1

and this is where it gets interesting... I looked at inetd which I have
running for telnet and other daemons.

pid | name | state | ppid | pgrp | session | tty | tpgid
--------------------------------------------------------
33  |inetd | S     | 1    | 33   | 33      | 1088| 33

So inetd successfully has a tty, but what is 1088?

So I took out inetd out of inittab and uncommented out the line:

ttyS0::askfirst:-/bin/ash

and no warning!

When I look at the stat for ash now, under tty it has 1088!
So I am guessing that ttyS0 is 1088 and inetd was grabbing it first. 
Unfortunately I will need to run inetd as well, does anyone know why
this is occurring or a workaround for this?

I am using netkit-base-0.17 unmodified.

Thanks again Vernon,
your suggestions pointed me in the right direction.
Regards,

Jeff Warren

>Vernon wrote:
>
>Some comments from my limited experience for what they are worth:
>
>1 - The kernel command line has "console=ttyS0,115200,mem=32M". I think
>the
>comma before the mem= should be a space. The console= handler might be
>getting confused??
>
>2 - The controlling tty of a process is listed in /proc/<pid>/stat. See
>"man
>proc" for more information. If the TTY field is zero, there is no
>controlling terminal. If it is some other number, I haven't yet figured
>out
>how to track that back to a fd.
>
>Best of luck,
>Vernon