httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Trawick <traw...@attglobal.net>
Subject Re: clean_child_exit, just_die and exit
Date Thu, 27 Sep 2001 17:09:30 GMT
Jim Jagielski <jim@jaguNET.com> writes:

> Bill Stoddard wrote:
> > 
> > We (and third party modules) call lots of stuff in clean_child_exit that is not
safe to
> > call from a signal handler. Dean has pointed this out many times. We need to eliminate
> > this particular architectural flaw in 2.0.
> > 
> > BTW, Jeff is working on a patch to 1.3 that will help out some though I do not know
all
> > the details.
> > 
> 
> I agree that we should have a "signal" aware version of c_c_e rather
> than re-utilizing what we have. The PR's main point was that one key
> problem was using exit() instead of _exit() and *that's* a new
> one on me, if it's even true. Hell, some default handlers call exit()
> directly...

Does this signal-aware version of c_c_e avoid calling child exit
handlers?  

If the signal-aware version of c_c_e does call child exit handlers,
then do child exit handlers have to know they're running from a signal
handler and avoid any questionable code?

Isn't it better just to avoid running c_c_e in a signal handler
context, as my patch attempts to implement?

Thanks,

Jeff

-- 
Jeff Trawick | trawick@attglobal.net | PGP public key at web site:
       http://www.geocities.com/SiliconValley/Park/9289/
             Born in Roswell... married an alien...

Mime
View raw message