Subject: Re: remotely exitting a process
To: Frank van der Linden <>
From: Jaromir Dolecek <>
List: tech-kern
Date: 12/04/2005 15:06:37
On Sun, Dec 04, 2005 at 01:18:50PM +0100, Frank van der Linden wrote:
> Jaromir Dolecek wrote:
> >Perhaps we could just use a new process flag for that - it's only
> >necessary to reset exitsig to 0 for exit_group() in sigexit(), introducing
> >a function hook for that seems overkill.
> > 
> >
> Not quite. You need to be able to pass on the exit code, passed as an 
> argument to the exit_group system call. That value needs to be obtained 
> in sigexit() somehow.

Ah, right - let's do the sigexit() hook then. Do I understand correctly
that this way exit_group() would:

1. for all processes in the same thread group, do:
   a) mark thread exit code in emuldata
   b) send it SIGKILL
2. have a emulation-specific hook in sigexit() to reset the exit code to that
   provided in exit_group()
3. don't need any hook in wait() function for parent, since the exit
   would appear as normal process exit from parent POV
Sounds fine.

Jaromir Dolecek <>  
-=- We can walk our road together if our goals are all the same;     -=-
-=- We can run alone and free if we pursue a different aim.          -=-