/* * Copyright (c) 2002-2003, Intel Corporation. All rights reserved. * Created by: rusty.lynch REMOVE-THIS AT intel DOT com * This file is licensed under the GPL license. For the full content * of this license, see the COPYING file at the top level of this * source tree. Test assertion #17 by verifying that select returns -1 with errno set to EINTR if a handler for the SIGTRAP signal is setup with the SA_RESTART flag cleared. * 12/18/02 - Adding in include of sys/time.h per * rodrigc REMOVE-THIS AT attbi DOT com input that it needs * to be included whenever the timeval struct is used. */ #include #include #include #include #include #include #include #include #include #include "posixtest.h" volatile sig_atomic_t wakeup = 1; void handler(int signo) { printf("Caught SIGTRAP\n"); wakeup++; } int main(void) { pid_t pid; struct timeval tv; if ((pid = fork()) == 0) { /* child */ struct sigaction act; act.sa_handler = handler; act.sa_flags = 0; sigemptyset(&act.sa_mask); sigaction(SIGTRAP, &act, 0); while (wakeup == 1) { tv.tv_sec = 3; tv.tv_usec = 0; if (select(0, NULL, NULL, NULL, &tv) == -1 && errno == EINTR) { perror("select"); return PTS_PASS; } } return PTS_FAIL; } else { /* parent */ int s; /* There is a race condition between the parent process sending the SIGTRAP signal, and the child process being inside the 'select' function call. I could not find a pure POSIX method for determining the state of the child process, so I just added a delay so that the test is valid in most conditions. (The problem is that it would be perfectly legal for a POSIX conformant OS to not schedule the child process for a long time.) */ tv.tv_sec = 1; tv.tv_usec = 0; select(0, NULL, NULL, NULL, &tv); kill(pid, SIGTRAP); waitpid(pid, &s, 0); if (WEXITSTATUS(s) == PTS_PASS) { printf("Test PASSED\n"); return PTS_PASS; } } printf("Test FAILED\n"); return PTS_FAIL; }