[ILUG] Signal handling

[Brian Foster]

  | From: "Dale Dunlea" <daledunlea at commergy.com>
  | Date: Tue, 7 Dec 2004 13:36:20 -0000
  | 
  |  > man ptrace
  | 
  | Right, I've had a look at that, and I'm using PTRACE_ATTACH.
  | The thing is, when my program tries to attach to the running
  | process, said process seems to halt.

 this is correct.  (_read_ the man page!)
 and yer problem is....?

 as others have pointed out, yer doing something
 quite strange; and as Pádraig is hinting, this
 can be quite hard to make robust.  also, the
 original description was slightly muddled, but
 the ghist seems to be that if some process Þ is
 orphaned, then a (new) process is to be notified
 via SIGCHLD if (when) þ terminates.  to emphasise
 Pádraig's point, that whole _idea_ is filled full
 of some nasty race conditions and security issues.

 methinks you should rethink yer algorithm.
 and a few more details would help.
cheers!
	-blf-

  |  > Can you guarantee the pid will be correct.
  |  > I.E. can you guarantee the child process didn't
  |  > itself die and another process now holds the PID?
  | 
  | Yes.

 interesting.  how?

-- 
Experienced (20+ years) kernel engineer: |  Brian Foster      Montpellier,
 · Unix, ChorusOS, others;               |  blf at utvinternet.ie      FRANCE
 · IDL, automated testing, process, &tc. |    Stop E$$o (ExxonMobile)!
Résumé:  http://www.blf.utvinternet.ie   |        http://www.stopesso.com