Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 15 Oct 2014 14:51:02 -0400
From:      Lowell Gilbert <freebsd-questions-local@be-well.ilk.org>
To:        freebsd-questions@freebsd.org
Subject:   Re: exiting Xorg locks up 9.2-STABLE system
Message-ID:  <44tx35w4ix.fsf@be-well.ilk.org>
In-Reply-To: <20141015185327.4fc2590d.freebsd@edvax.de> (Polytropon's message of "Wed, 15 Oct 2014 18:53:27 +0200")
References:  <20141015162547.GJ3963@itcom245.staff.itd.umich.edu> <20141015185327.4fc2590d.freebsd@edvax.de>

Next in thread | Previous in thread | Raw E-Mail | Index | Archive | Help
Polytropon <freebsd@edvax.de> writes:

> On Wed, 15 Oct 2014 12:25:47 -0400, William Bulley wrote:
>> According to Polytropon <freebsd@edvax.de> on Wed, 10/15/14 at 12:18:
>> > 
>> > The last line of your ~/.xinitrc file should be:
>> > 
>> > 	exec /usr/local/bin/mwm
>> > 
>> > This makes sure that the process "continues" as mwm, and when
>> > mwm exits, then X also exists.
>> 
>> Thanks.  The last line of my ~/.xinitrc file is:
>> 
>>    /usr/local/bin/mwm 2>&1 /dev/null
>> 
>> and since I start Xorg thusly:
>> 
>>    /usr/local/bin/xinit -- /usr/local/bin/Xorg
>> 
>> I had thought when I exit mwm(1) using the "f.quit_mwm" feature,
>> that then Xorg would exit.  Is that not the case?  Or am I missing
>> your point above?  This configuration has been working for me for
>> about fifteen years.   :-)
>
> Interesting, but it's not what The FreeBSD Handbook suggests
> and what _I_ have been using for about fifteen years. :-)
>
> The "exec" statement makes the xinit process "continue" as
> the window manager process. Its f.quit_mwm function exits
> the mwm program, but that does not have any effect on the
> X session - the session continues without a window manager
> (which is possible). Only if the X process is "connected"
> to the window manager process using "exec", it will exit,
> and so quit the X session entirely.

Without "exec", the windows manager is running as a subprocess of
shell executing the .xinitrc file. When it exits, the shell continues,
runs the shell input to completion, and the shell process terminates,
which also terminates the X server session. 

In other words, either way will work fine as far as this situation is
concerned. Using "exec" avoids having an extra process sticking around,
which is slightly more efficient -- but keeps your .xinitrc from being
able to do any custom cleanup after the window manager exits. I venture
to say that only a vanishing few would ever notice the difference.

> This is what "init" or "startx" help to do: They source
> the .xinitrc file and execute it. Its last line, the "exec"
> statement, creates the "connection" between the X session
> and the program which has control over its life, usually
> the window manager - but it could also be an xterm, and
> when _that_ xterm is closed, the X session terminates.
> The line "exec /usr/local/bin/mwm" will do this.

There are several factual errors in here, although nothing that the
original poster would need to know. It's mostly a matter of not properly
conveying the concept of a child process, or of the fact that a parent
process (in these cases) waits for its child process to complete before
continuing on with its own business.



Want to link to this message? Use this URL: <http://docs.FreeBSD.org/cgi/mid.cgi?44tx35w4ix.fsf>