Hi all,

i am porting a system, that used to manipulate memory across processes using an interrupt handler - which means that upon return from the interrupt handler the memory change would be finished. I am trying to simulate this using signals on Linux 2.6.2x.

What i would like to know is wether sending a signal (ex. kill( pid , SIGUSR1 ) ) from a process will cause the kernel to perform taskswitching and put the calling process in queue, allowing the signal handler to run if it is ready?


A simplified scenario:
We have a shared memory area attached to both processes "share".

Process1:
...
share->variable = 0;
kill( Process2_pid , SIGUSR1 )
printf("share->variable= %d\n",share->variable);
...

Process2:
SIGUSR1_handler()
{
share->variable = 1;
}

It seems that the printf allways prints a "1" . Which would mean that taskswitching occurs, and that the handler in process2 has finished when process1 returns from kill().

But is this something one can rely on Running Linux 2.6.2x ?


Regards
Paul