I've recently finished Section 10 (Signals) of "Advanced Programming in the Unix Environment" (3rd edition) and I've come across a piece of code I don't entirely understand:
#include "apue.h"
static volatile sig_atomic_t sigflag; /* set nonzero by sig handler */
static sigset_t newmask, oldmask, zeromask;
static void
sig_usr(int signo) /* one signal handler for SIGUSR1 and SIGUSR2 */
{
sigflag = 1;
}
void
TELL_WAIT(void)
{
if (signal(SIGUSR1, sig_usr) == SIG_ERR)
err_sys("signal(SIGUSR1) error");
if (signal(SIGUSR2, sig_usr) == SIG_ERR)
err_sys("signal(SIGUSR2) error");
sigemptyset(&zeromask);
sigemptyset(&newmask);
sigaddset(&newmask, SIGUSR1);
sigaddset(&newmask, SIGUSR2);
/* Block SIGUSR1 and SIGUSR2, and save current signal mask */
if (sigprocmask(SIG_BLOCK, &newmask, &oldmask) < 0)
err_sys("SIG_BLOCK error");
}
void
TELL_PARENT(pid_t pid)
{
kill(pid, SIGUSR2); /* tell parent we're done */
}
void
WAIT_PARENT(void)
{
while (sigflag == 0)
sigsuspend(&zeromask); /* and wait for parent */
sigflag = 0;
/* Reset signal mask to original value */
if (sigprocmask(SIG_SETMASK, &oldmask, NULL) < 0)
err_sys("SIG_SETMASK error");
}
void
TELL_CHILD(pid_t pid)
{
kill(pid, SIGUSR1); /* tell child we're done */
}
void
WAIT_CHILD(void)
{
while (sigflag == 0)
sigsuspend(&zeromask); /* and wait for child */
sigflag = 0;
/* Reset signal mask to original value */
if (sigprocmask(SIG_SETMASK, &oldmask, NULL) < 0)
err_sys("SIG_SETMASK error");
}
The routines above are used (as you certainly know) to synchronize processes using signals. Although I understand every single line on its own, I can't see (understand) the big picture. The code itself is used it the following scenario: to avoid a race condition in our program, after we fork(), we make the child process TELL_PARENT and WAIT_PARENT, and then we do the same to the parent with TELL_CHILD and WAIT_CHILD. My questions are:
1.) How can a child communicate with its parent through a variable while both of them work with their own set (copy) of variables? Is it because the child doesn't modify sigflag directly but through a signal handler (the same goes for the parent)? 2.) Why do we need to block SIGUSR1 and SIGUSR2 and then unblock it with sigprocmask?
A program that uses three of those routines could be (taken from the book):
#include "apue.h"
static void charatatime(char *);
int
main(void)
{
pid_t pid;
TELL_WAIT();
if ((pid = fork()) < 0) {
err_sys("fork error");
} else if (pid == 0) {
WAIT_PARENT(); /* parent goes first */
charatatime("output from child\n");
} else {
charatatime("output from parent\n");
TELL_CHILD(pid);
}
exit(0);
}
static void
charatatime(char *str)
{
char *ptr;
int c;
setbuf(stdout, NULL); /* set unbuffered */
for (ptr = str; (c = *ptr++) != 0; )
putc(c, stdout);
}
Cheers,
1) They are not communicating through "variable" - the sole communication facility used here is
kill
function. We "tell" things by invokingkill
, we "wait" to be told withsigsuspend
.sig_flag
is not shared, it's a local state of each process, and it says whether this particular process has been "told" by the other.2) Were the signals not blocked prior to
fork
, the parent process could send the signal to the child before the child has started waiting for it. That is, the timeline could be like that:kill
But this signal has already been delivered, and so waits indefinitely. Therefore, we must ensure the signal is not delivered to the child process before it starts the waiting loop. To this end, we block it before
fork
, and atomically unblock it and start waiting for it. Atomicity is the key; required invariant cannot be achieved with this operation performed as two independent steps, as the signal could be delivered inbetween.