static int do_child(orte_app_context_t* context,
orte_proc_t *child,
char **environ_copy,
orte_job_t *jobdat, int write_fd,
orte_iof_base_io_conf_t opts)
{
int i, rc;
sigset_t sigs;
char *param, *msg;
/* Setup the pipe to be close-on-exec */
opal_fd_set_cloexec(write_fd);
if (NULL != child) {
/* setup stdout/stderr so that any error messages that we
may print out will get displayed back at orterun.
NOTE: Definitely do this AFTER we check contexts so
that any error message from those two functions doesn't
come out to the user. IF we didn't do it in this order,
THEN a user who gives us a bad executable name or
working directory would get N error messages, where
N=num_procs. This would be very annoying for large
jobs, so instead we set things up so that orterun
always outputs a nice, single message indicating what
happened
*/
if (ORTE_SUCCESS != (i = orte_iof_base_setup_child(&opts,
&environ_copy))) {
ORTE_ERROR_LOG(i);
send_error_show_help(write_fd, 1,
"help-orte-odls-alps.txt",
"iof setup failed",
orte_process_info.nodename, context->app);
/* Does not return */
}
/* now set any child-level controls such as binding */
orte_rtc.set(jobdat, child, &environ_copy, write_fd);
} else if (!ORTE_FLAG_TEST(jobdat, ORTE_JOB_FLAG_FORWARD_OUTPUT)) {
/* tie stdin/out/err/internal to /dev/null */
int fdnull;
for (i=0; i < 3; i++) {
fdnull = open("/dev/null", O_RDONLY, 0);
if (fdnull > i && i != write_fd) {
dup2(fdnull, i);
}
close(fdnull);
}
fdnull = open("/dev/null", O_RDONLY, 0);
if (fdnull > opts.p_internal[1]) {
dup2(fdnull, opts.p_internal[1]);
}
close(fdnull);
}
/* if the user requested it, set the system resource limits */
if (OPAL_SUCCESS != (rc = opal_util_init_sys_limits(&msg))) {
send_error_show_help(write_fd, 1, "help-orte-odls-alps.txt",
"set limit",
orte_process_info.nodename, context->app,
__FILE__, __LINE__, msg);
}
/* ensure we only do this once */
(void) mca_base_var_env_name("opal_set_max_sys_limits", ¶m);
opal_unsetenv(param, &environ_copy);
free(param);
if (ORTE_SUCCESS != close_open_file_descriptors(write_fd, opts)) {
send_error_show_help(write_fd, 1, "help-orte-odls-alps.txt",
"close fds",
orte_process_info.nodename, context->app,
__FILE__, __LINE__);
}
if (context->argv == NULL) {
context->argv = malloc(sizeof(char*)*2);
context->argv[0] = strdup(context->app);
context->argv[1] = NULL;
}
/* Set signal handlers back to the default. Do this close to
the exev() because the event library may (and likely will)
reset them. If we don't do this, the event library may
have left some set that, at least on some OS's, don't get
reset via fork() or exec(). Hence, the launched process
could be unkillable (for example). */
set_handler_alps(SIGTERM);
set_handler_alps(SIGINT);
set_handler_alps(SIGHUP);
set_handler_alps(SIGPIPE);
set_handler_alps(SIGCHLD);
/* Unblock all signals, for many of the same reasons that we
set the default handlers, above. This is noticable on
Linux where the event library blocks SIGTERM, but we don't
want that blocked by the launched process. */
sigprocmask(0, 0, &sigs);
sigprocmask(SIG_UNBLOCK, &sigs, 0);
/* Exec the new executable */
if (10 < opal_output_get_verbosity(orte_odls_base_framework.framework_output)) {
int jout;
opal_output(0, "%s STARTING %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), context->app);
for (jout=0; NULL != context->argv[jout]; jout++) {
opal_output(0, "%s\tARGV[%d]: %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), jout, context->argv[jout]);
}
for (jout=0; NULL != environ_copy[jout]; jout++) {
opal_output(0, "%s\tENVIRON[%d]: %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), jout, environ_copy[jout]);
}
}
execve(context->app, context->argv, environ_copy);
send_error_show_help(write_fd, 1,
"help-orte-odls-alps.txt", "execve error",
orte_process_info.nodename, context->app, strerror(errno));
/* Does not return */
}
static int do_child(orte_odls_spawn_caddy_t *cd, int write_fd)
{
int i;
sigset_t sigs;
/* Setup the pipe to be close-on-exec */
opal_fd_set_cloexec(write_fd);
if (NULL != cd->child) {
/* setup stdout/stderr so that any error messages that we
may print out will get displayed back at orterun.
NOTE: Definitely do this AFTER we check contexts so
that any error message from those two functions doesn't
come out to the user. IF we didn't do it in this order,
THEN a user who gives us a bad executable name or
working directory would get N error messages, where
N=num_procs. This would be very annoying for large
jobs, so instead we set things up so that orterun
always outputs a nice, single message indicating what
happened
*/
if (ORTE_FLAG_TEST(cd->jdata, ORTE_JOB_FLAG_FORWARD_OUTPUT)) {
if (ORTE_SUCCESS != (i = orte_iof_base_setup_child(&cd->opts, &cd->env))) {
ORTE_ERROR_LOG(i);
send_error_show_help(write_fd, 1,
"help-orte-odls-alps.txt",
"iof setup failed",
orte_process_info.nodename, cd->app->app);
/* Does not return */
}
}
/* now set any child-level controls such as binding */
orte_rtc.set(cd->jdata, cd->child, &cd->env, write_fd);
} else if (!ORTE_FLAG_TEST(cd->jdata, ORTE_JOB_FLAG_FORWARD_OUTPUT)) {
/* tie stdin/out/err/internal to /dev/null */
int fdnull;
for (i=0; i < 3; i++) {
fdnull = open("/dev/null", O_RDONLY, 0);
if (fdnull > i && i != write_fd) {
dup2(fdnull, i);
}
close(fdnull);
}
#if OPAL_PMIX_V1
fdnull = open("/dev/null", O_RDONLY, 0);
if (fdnull > cd->opts.p_internal[1]) {
dup2(fdnull, cd->opts.p_internal[1]);
}
close(fdnull);
#endif
}
if (ORTE_SUCCESS != close_open_file_descriptors(write_fd, cd->opts)) {
send_error_show_help(write_fd, 1, "help-orte-odls-alps.txt",
"close fds",
orte_process_info.nodename, cd->app->app,
__FILE__, __LINE__);
}
if (cd->argv == NULL) {
cd->argv = malloc(sizeof(char*)*2);
cd->argv[0] = strdup(cd->app->app);
cd->argv[1] = NULL;
}
/* Set signal handlers back to the default. Do this close to
the exev() because the event library may (and likely will)
reset them. If we don't do this, the event library may
have left some set that, at least on some OS's, don't get
reset via fork() or exec(). Hence, the launched process
could be unkillable (for example). */
set_handler_alps(SIGTERM);
set_handler_alps(SIGINT);
set_handler_alps(SIGHUP);
set_handler_alps(SIGPIPE);
set_handler_alps(SIGCHLD);
/* Unblock all signals, for many of the same reasons that we
set the default handlers, above. This is noticable on
Linux where the event library blocks SIGTERM, but we don't
want that blocked by the launched process. */
sigprocmask(0, 0, &sigs);
sigprocmask(SIG_UNBLOCK, &sigs, 0);
/* take us to the correct wdir */
if (NULL != cd->wdir) {
if (0 != chdir(cd->wdir)) {
send_error_show_help(write_fd, 1,
"help-orterun.txt",
"orterun:wdir-not-found",
"orted",
cd->wdir,
orte_process_info.nodename,
(NULL == cd->child) ? 0 : cd->child->app_rank);
/* Does not return */
}
}
/* Exec the new executable */
if (10 < opal_output_get_verbosity(orte_odls_base_framework.framework_output)) {
int jout;
opal_output(0, "%s STARTING %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), cd->app->app);
for (jout=0; NULL != cd->argv[jout]; jout++) {
opal_output(0, "%s\tARGV[%d]: %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), jout, cd->argv[jout]);
}
for (jout=0; NULL != cd->env[jout]; jout++) {
opal_output(0, "%s\tENVIRON[%d]: %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), jout, cd->env[jout]);
}
}
execve(cd->cmd, cd->argv, cd->env);
send_error_show_help(write_fd, 1,
"help-orte-odls-alps.txt", "execve error",
orte_process_info.nodename, cd->app->app, strerror(errno));
/* Does not return */
}
