On Sunday 18 June 2006 09:51, Dinesh Ahuja wrote:
> As per older handling of SIGCHLD, if we set the
> disposition of SIGHCHLD to be caught, the kernel
> immediately checks if there are aby child processes
> ready to be waited for and , if so, calls SIGCHLD
> behavior.
>
> Following code snipped should give out of stack space
> and should terminate.
> // Signal Handler for SIGCHLD
> static void sig_chld()
> {
> int status;
> pid_t pid;
> signal(SIGCHLD,sig_chld);
>
> pid = wait(&status);
> return;
> }
>
> I fully agree that POSIX signals are reliable and
> hence we need not set signal handler again in signal
> handler. But what should happen if I do it explicitly?
What happens to the registered handler is system dependent.
Use sigaction() for more explicit and portable behavior.
> I was expecting to have a recursive call to signal
> handler and hence out of stack should have occured.
I wouldn't expect that.
> But the aboVe code works fine on Linux Kernel 2.4.20.
> How does POSIX handles if we establish a signal
> handler and there exists a terminated child which
> needs to be waited for.
I believe the signal handler needs to be registered before the
child dies or it won't be called.
--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive: http://mail.nl.linux.org/kernelnewbies/
FAQ: http://kernelnewbies.org/faq/
