/*
* CR Init
*/
int ompi_cr_init(void)
{
int val;
/*
* Register some MCA parameters
*/
mca_base_param_reg_int_name("ompi_cr", "verbose",
"Verbose output for the OMPI Checkpoint/Restart functionality",
false, false,
0,
&val);
if(0 != val) {
ompi_cr_output = opal_output_open(NULL);
opal_output_set_verbosity(ompi_cr_output, val);
} else {
ompi_cr_output = opal_cr_output;
}
/* Typically this is not needed. Individual BTLs will set this as needed */
ompi_cr_continue_like_restart = false;
opal_output_verbose(10, ompi_cr_output,
"ompi_cr: init: ompi_cr_init()");
/* Register the OMPI interlevel coordination callback */
opal_cr_reg_coord_callback(ompi_cr_coord, &prev_coord_callback);
return OMPI_SUCCESS;
}
/*
* CR Init
*/
int orte_cr_init(void)
{
int ret, exit_status = ORTE_SUCCESS;
int val;
/*
* OPAL Frameworks
*/
if (OPAL_SUCCESS != (ret = opal_cr_init() ) ) {
exit_status = ret;
goto cleanup;
}
/*
* Register MCA Parameters
*/
mca_base_param_reg_int_name("orte_cr", "verbose",
"Verbose output for the ORTE Checkpoint/Restart functionality",
false, false,
0,
&val);
/*** RHC: This is going to crash-and-burn when the output conversion is
* completed as opal_output will have no idea what opal_cr_output stream means,
* or even worse, will have assigned it to someone else!
*/
if(0 != val) {
orte_cr_output = opal_output_open(NULL);
opal_output_set_verbosity(orte_cr_output, val);
} else {
orte_cr_output = opal_cr_output;
}
opal_output_verbose(10, orte_cr_output,
"orte_cr: init: orte_cr_init()\n");
/* Init ORTE Entry Point Function */
if( ORTE_SUCCESS != (ret = orte_cr_entry_point_init()) ) {
exit_status = ret;
goto cleanup;
}
/* Register the ORTE interlevel coordination callback */
opal_cr_reg_coord_callback(orte_cr_coord, &prev_coord_callback);
/* Typically this is not needed. Individual BTLs will set this as needed */
orte_cr_continue_like_restart = false;
orte_cr_flush_restart_files = true;
cleanup:
return exit_status;
}
/*
* CR Init
*/
int ompi_cr_init(void)
{
/*
* Register some MCA variables
*/
ompi_cr_verbosity = 0;
(void) mca_base_var_register("ompi", "ompi", "cr", "verbose",
"Verbose output for the OMPI Checkpoint/Restart functionality",
MCA_BASE_VAR_TYPE_INT, NULL, 0, 0,
OPAL_INFO_LVL_9,
MCA_BASE_VAR_SCOPE_READONLY,
&ompi_cr_verbosity);
if(0 != ompi_cr_verbosity) {
ompi_cr_output = opal_output_open(NULL);
opal_output_set_verbosity(ompi_cr_output, ompi_cr_verbosity);
} else {
ompi_cr_output = opal_cr_output;
}
opal_output_verbose(10, ompi_cr_output,
"ompi_cr: init: ompi_cr_init()");
/* Register the OMPI interlevel coordination callback */
opal_cr_reg_coord_callback(ompi_cr_coord, &prev_coord_callback);
#if OPAL_ENABLE_CRDEBUG == 1
/* Check for C/R enabled debugging */
if( MPIR_debug_with_checkpoint ) {
char *uri = NULL;
char *sep = NULL;
char *hostname = NULL;
/* Mark as debuggable with C/R */
MPIR_checkpointable = 1;
/* Set the checkpoint and restart commands */
/* Add the full path to the binary */
asprintf(&MPIR_checkpoint_command,
"%s/ompi-checkpoint --crdebug --hnp-jobid %u",
opal_install_dirs.bindir,
ORTE_PROC_MY_HNP->jobid);
asprintf(&MPIR_restart_command,
"%s/ompi-restart --crdebug ",
opal_install_dirs.bindir);
asprintf(&MPIR_checkpoint_listing_command,
"%s/ompi-checkpoint -l --crdebug ",
opal_install_dirs.bindir);
/* Set contact information for HNP */
uri = strdup(ompi_process_info.my_hnp_uri);
hostname = strchr(uri, ';') + 1;
sep = strchr(hostname, ';');
if (sep) {
*sep = 0;
}
if (strncmp(hostname, "tcp://", 6) == 0) {
hostname += 6;
sep = strchr(hostname, ':');
*sep = 0;
MPIR_controller_hostname = strdup(hostname);
} else {
MPIR_controller_hostname = strdup("localhost");
}
/* Cleanup */
if( NULL != uri ) {
free(uri);
uri = NULL;
}
}
#endif
return OMPI_SUCCESS;
}
int opal_cr_init(void )
{
int ret, exit_status = OPAL_SUCCESS;
opal_cr_coord_callback_fn_t prev_coord_func;
if( ++opal_cr_initalized != 1 ) {
if( opal_cr_initalized < 1 ) {
exit_status = OPAL_ERROR;
goto cleanup;
}
exit_status = OPAL_SUCCESS;
goto cleanup;
}
ret = opal_cr_register ();
if (OPAL_SUCCESS != ret) {
return ret;
}
if(0 != opal_cr_verbose) {
opal_cr_output = opal_output_open(NULL);
opal_output_set_verbosity(opal_cr_output, opal_cr_verbose);
}
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Verbose Level: %d",
opal_cr_verbose);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: FT Enabled: %s",
opal_cr_is_enabled ? "true" : "false");
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Is a tool program: %s",
opal_cr_is_tool ? "true" : "false");
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Debug SIGPIPE: %d (%s)",
opal_cr_verbose, (opal_cr_debug_sigpipe ? "True" : "False"));
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Checkpoint Signal: %d",
opal_cr_entry_point_signal);
#if OPAL_ENABLE_FT_THREAD == 1
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: FT Use thread: %s",
opal_cr_thread_use_if_avail ? "true" : "false");
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: FT thread sleep: check = %d, wait = %d",
opal_cr_thread_sleep_check, opal_cr_thread_sleep_wait);
/* If we have a thread, then attach the SIGPIPE signal handler there since
* it is most likely to be the one that needs it.
*/
if( opal_cr_debug_sigpipe && !opal_cr_thread_use_if_avail ) {
if( SIG_ERR == signal(SIGPIPE, opal_cr_sigpipe_debug_signal_handler) ) {
;
}
}
#else
if( opal_cr_debug_sigpipe ) {
if( SIG_ERR == signal(SIGPIPE, opal_cr_sigpipe_debug_signal_handler) ) {
;
}
}
#endif
#if OPAL_ENABLE_CRDEBUG == 1
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: C/R Debugging Enabled [%s]\n",
(MPIR_debug_with_checkpoint ? "True": "False"));
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Checkpoint Signal (Debug): %d",
opal_cr_debug_signal);
if( SIG_ERR == signal(opal_cr_debug_signal, MPIR_checkpoint_debugger_signal_handler) ) {
opal_output(opal_cr_output,
"opal_cr: init: Failed to register C/R debug signal (%d)",
opal_cr_debug_signal);
}
#endif
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Temp Directory: %s",
opal_cr_pipe_dir);
if( !opal_cr_is_tool ) {
/* Register the OPAL interlevel coordination callback */
opal_cr_reg_coord_callback(opal_cr_coord, &prev_coord_func);
opal_cr_stall_check = false;
opal_cr_currently_stalled = false;
} /* End opal_cr_is_tool = true */
/*
* If fault tolerance was not compiled in then
* we need to make sure that the listener thread is active to tell
* the tools that this is not a checkpointable job.
* We don't need the CRS framework to be initalized.
*/
#if OPAL_ENABLE_FT_CR == 1
/*
* Open the checkpoint / restart service components
*/
if (OPAL_SUCCESS != (ret = mca_base_framework_open(&opal_crs_base_framework, 0))) {
opal_show_help( "help-opal-runtime.txt",
"opal_cr_init:no-crs", true,
"opal_crs_base_open", ret );
exit_status = ret;
goto cleanup;
}
if (OPAL_SUCCESS != (ret = opal_crs_base_select())) {
opal_show_help( "help-opal-runtime.txt",
"opal_cr_init:no-crs", true,
"opal_crs_base_select", ret );
exit_status = ret;
goto cleanup;
}
#endif
#if OPAL_ENABLE_FT_THREAD == 1
if( !opal_cr_is_tool && opal_cr_thread_use_if_avail) {
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: starting the thread\n");
/* JJH: We really do need this line below since it enables
* actual locks for threads. However currently the
* upper layers will deadlock if it is enabled.
* So hack around the problem for now, while working
* on a complete solution. See ticket #2741 for more
* details.
* opal_set_using_threads(true);
*/
/*
* Start the thread
*/
OBJ_CONSTRUCT(&opal_cr_thread, opal_thread_t);
OBJ_CONSTRUCT(&opal_cr_thread_lock, opal_mutex_t);
opal_cr_thread_is_done = false;
opal_cr_thread_is_active = false;
opal_cr_thread_in_library = false;
opal_cr_thread_num_in_library = 0;
opal_cr_thread.t_run = opal_cr_thread_fn;
opal_cr_thread.t_arg = NULL;
opal_thread_start(&opal_cr_thread);
} /* End opal_cr_is_tool = true */
else {
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: *Not* Using C/R thread\n");
}
#endif /* OPAL_ENABLE_FT_THREAD == 1 */
cleanup:
return exit_status;
}
int opal_cr_init(void )
{
int ret, exit_status = OPAL_SUCCESS;
opal_cr_coord_callback_fn_t prev_coord_func;
int val;
if( ++opal_cr_initalized != 1 ) {
if( opal_cr_initalized < 1 ) {
exit_status = OPAL_ERROR;
goto cleanup;
}
exit_status = OPAL_SUCCESS;
goto cleanup;
}
/*
* Some startup MCA parameters
*/
ret = mca_base_param_reg_int_name("opal_cr", "verbose",
"Verbose output level for the runtime OPAL Checkpoint/Restart functionality",
false, false,
0,
&val);
if(0 != val) {
opal_cr_output = opal_output_open(NULL);
} else {
opal_cr_output = -1;
}
opal_output_set_verbosity(opal_cr_output, val);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Verbose Level: %d",
val);
mca_base_param_reg_int_name("ft", "cr_enabled",
"Enable fault tolerance for this program",
false, false,
0, &val);
opal_cr_set_enabled(OPAL_INT_TO_BOOL(val));
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: FT Enabled: %d",
val);
mca_base_param_reg_int_name("opal_cr", "enable_timer",
"Enable Checkpoint timer (Default: Disabled)",
false, false,
0, &val);
opal_cr_timing_enabled = OPAL_INT_TO_BOOL(val);
mca_base_param_reg_int_name("opal_cr", "enable_timer_barrier",
"Enable Checkpoint timer Barrier (Default: Disabled)",
false, false,
0, &val);
if( opal_cr_timing_enabled ) {
opal_cr_timing_barrier_enabled = OPAL_INT_TO_BOOL(val);
} else {
opal_cr_timing_barrier_enabled = false;
}
mca_base_param_reg_int_name("opal_cr", "timer_target_rank",
"Target Rank for the timer (Default: 0)",
false, false,
0, &val);
opal_cr_timing_target_rank = val;
#if OPAL_ENABLE_FT_THREAD == 1
mca_base_param_reg_int_name("opal_cr", "use_thread",
"Use an async thread to checkpoint this program (Default: Disabled)",
false, false,
0, &val);
opal_cr_thread_use_if_avail = OPAL_INT_TO_BOOL(val);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: FT Use thread: %d",
val);
mca_base_param_reg_int_name("opal_cr", "thread_sleep_check",
"Time to sleep between checking for a checkpoint (Default: 0)",
false, false,
0, &val);
opal_cr_thread_sleep_check = val;
mca_base_param_reg_int_name("opal_cr", "thread_sleep_wait",
"Time to sleep waiting for process to exit MPI library (Default: 0)",
false, false,
0, &val);
opal_cr_thread_sleep_wait = val;
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: FT thread sleep: check = %d, wait = %d",
opal_cr_thread_sleep_check, opal_cr_thread_sleep_wait);
#endif
mca_base_param_reg_int_name("opal_cr", "is_tool",
"Is this a tool program, meaning does it require a fully operational OPAL or just enough to exec.",
false, false,
0,
&val);
opal_cr_is_tool = OPAL_INT_TO_BOOL(val);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Is a tool program: %d",
val);
#ifndef __WINDOWS__
mca_base_param_reg_int_name("opal_cr", "signal",
"Checkpoint/Restart signal used to initialize an OPAL Only checkpoint of a program",
false, false,
SIGUSR1,
&opal_cr_entry_point_signal);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Checkpoint Signal: %d",
opal_cr_entry_point_signal);
mca_base_param_reg_int_name("opal_cr", "debug_sigpipe",
"Activate a signal handler for debugging SIGPIPE Errors that can happen on restart. (Default: Disabled)",
false, false,
0, &val);
opal_cr_debug_sigpipe = OPAL_INT_TO_BOOL(val);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Debug SIGPIPE: %d (%s)",
val, (opal_cr_debug_sigpipe ? "True" : "False"));
#if OPAL_ENABLE_FT_THREAD == 1
/* If we have a thread, then attach the SIGPIPE signal handler there since
* it is most likely to be the one that needs it.
*/
if( opal_cr_debug_sigpipe && !opal_cr_thread_use_if_avail ) {
if( SIG_ERR == signal(SIGPIPE, opal_cr_sigpipe_debug_signal_handler) ) {
;
}
}
#else
if( opal_cr_debug_sigpipe ) {
if( SIG_ERR == signal(SIGPIPE, opal_cr_sigpipe_debug_signal_handler) ) {
;
}
}
#endif
#else
opal_cr_is_tool = true; /* no support for CR on Windows yet */
#endif /* __WINDOWS__ */
mca_base_param_reg_string_name("opal_cr", "tmp_dir",
"Temporary directory to place rendezvous files for a checkpoint",
false, false,
"/tmp",
&opal_cr_pipe_dir);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Temp Directory: %s",
opal_cr_pipe_dir);
if( !opal_cr_is_tool ) {
/* Register the OPAL interlevel coordination callback */
opal_cr_reg_coord_callback(opal_cr_coord, &prev_coord_func);
opal_cr_stall_check = false;
opal_cr_currently_stalled = false;
} /* End opal_cr_is_tool = true */
/*
* If fault tolerance was not compiled in then
* we need to make sure that the listener thread is active to tell
* the tools that this is not a checkpointable job.
* We don't need the CRS framework to be initalized.
*/
#if OPAL_ENABLE_FT == 1
/*
* Open the checkpoint / restart service components
*/
if (OPAL_SUCCESS != (ret = opal_crs_base_open())) {
opal_output(opal_cr_output,
"opal_cr: init: opal_crs_base_open Failed to open. (%d)\n", ret);
exit_status = ret;
goto cleanup;
}
if (OPAL_SUCCESS != (ret = opal_crs_base_select())) {
opal_output(opal_cr_output,
"opal_cr: init: opal_crs_base_select Failed. (%d)\n", ret);
exit_status = ret;
goto cleanup;
}
#endif
#if OPAL_ENABLE_FT_THREAD == 1
if( !opal_cr_is_tool && opal_cr_thread_use_if_avail) {
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: starting the thread\n");
opal_set_using_threads(true);
/*
* Start the thread
*/
OBJ_CONSTRUCT(&opal_cr_thread, opal_thread_t);
OBJ_CONSTRUCT(&opal_cr_thread_lock, opal_mutex_t);
opal_cr_thread_is_done = false;
opal_cr_thread_is_active = false;
opal_cr_thread_in_library = false;
opal_cr_thread_num_in_library = 0;
opal_cr_thread.t_run = opal_cr_thread_fn;
opal_cr_thread.t_arg = NULL;
opal_thread_start(&opal_cr_thread);
} /* End opal_cr_is_tool = true */
else {
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: *Not* Using C/R thread\n");
}
#endif /* OPAL_ENABLE_FT_THREAD == 1 */
cleanup:
return exit_status;
}
int opal_cr_init(void )
{
int ret, exit_status = OPAL_SUCCESS;
opal_cr_coord_callback_fn_t prev_coord_func;
int val, t;
if( ++opal_cr_initalized != 1 ) {
if( opal_cr_initalized < 1 ) {
exit_status = OPAL_ERROR;
goto cleanup;
}
exit_status = OPAL_SUCCESS;
goto cleanup;
}
/*
* Some startup MCA parameters
*/
ret = mca_base_param_reg_int_name("opal_cr", "verbose",
"Verbose output level for the runtime OPAL Checkpoint/Restart functionality",
false, false,
0,
&val);
if(0 != val) {
opal_cr_output = opal_output_open(NULL);
} else {
opal_cr_output = -1;
}
opal_output_set_verbosity(opal_cr_output, val);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Verbose Level: %d",
val);
mca_base_param_reg_int_name("ft", "cr_enabled",
"Enable fault tolerance for this program",
false, false,
0, &val);
opal_cr_set_enabled(OPAL_INT_TO_BOOL(val));
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: FT Enabled: %d",
val);
mca_base_param_reg_int_name("opal_cr", "enable_timer",
"Enable Checkpoint timer (Default: Disabled)",
false, false,
0, &val);
opal_cr_timing_enabled = OPAL_INT_TO_BOOL(val);
mca_base_param_reg_int_name("opal_cr", "enable_timer_barrier",
"Enable Checkpoint timer Barrier (Default: Disabled)",
false, false,
0, &val);
if( opal_cr_timing_enabled ) {
opal_cr_timing_barrier_enabled = OPAL_INT_TO_BOOL(val);
} else {
opal_cr_timing_barrier_enabled = false;
}
mca_base_param_reg_int_name("opal_cr", "timer_target_rank",
"Target Rank for the timer (Default: 0)",
false, false,
0, &val);
opal_cr_timing_target_rank = val;
#if OPAL_ENABLE_FT_THREAD == 1
mca_base_param_reg_int_name("opal_cr", "use_thread",
"Use an async thread to checkpoint this program (Default: Disabled)",
false, false,
0, &val);
opal_cr_thread_use_if_avail = OPAL_INT_TO_BOOL(val);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: FT Use thread: %d",
val);
mca_base_param_reg_int_name("opal_cr", "thread_sleep_check",
"Time to sleep between checking for a checkpoint (Default: 0)",
false, false,
0, &val);
opal_cr_thread_sleep_check = val;
mca_base_param_reg_int_name("opal_cr", "thread_sleep_wait",
"Time to sleep waiting for process to exit MPI library (Default: 1000)",
false, false,
1000, &val);
opal_cr_thread_sleep_wait = val;
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: FT thread sleep: check = %d, wait = %d",
opal_cr_thread_sleep_check, opal_cr_thread_sleep_wait);
#endif
mca_base_param_reg_int_name("opal_cr", "is_tool",
"Is this a tool program, meaning does it require a fully operational OPAL or just enough to exec.",
false, false,
0,
&val);
opal_cr_is_tool = OPAL_INT_TO_BOOL(val);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Is a tool program: %d",
val);
#if OPAL_ENABLE_CRDEBUG == 1
mca_base_param_reg_int_name("opal_cr", "enable_crdebug",
"Enable checkpoint/restart debugging",
false, false,
0,
&val);
MPIR_debug_with_checkpoint = OPAL_INT_TO_BOOL(val);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: C/R Debugging Enabled [%s]\n",
(MPIR_debug_with_checkpoint ? "True": "False"));
#endif
#ifndef __WINDOWS__
mca_base_param_reg_int_name("opal_cr", "signal",
"Checkpoint/Restart signal used to initialize an OPAL Only checkpoint of a program",
false, false,
SIGUSR1,
&opal_cr_entry_point_signal);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Checkpoint Signal: %d",
opal_cr_entry_point_signal);
mca_base_param_reg_int_name("opal_cr", "debug_sigpipe",
"Activate a signal handler for debugging SIGPIPE Errors that can happen on restart. (Default: Disabled)",
false, false,
0, &val);
opal_cr_debug_sigpipe = OPAL_INT_TO_BOOL(val);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Debug SIGPIPE: %d (%s)",
val, (opal_cr_debug_sigpipe ? "True" : "False"));
#if OPAL_ENABLE_FT_THREAD == 1
/* If we have a thread, then attach the SIGPIPE signal handler there since
* it is most likely to be the one that needs it.
*/
if( opal_cr_debug_sigpipe && !opal_cr_thread_use_if_avail ) {
if( SIG_ERR == signal(SIGPIPE, opal_cr_sigpipe_debug_signal_handler) ) {
;
}
}
#else
if( opal_cr_debug_sigpipe ) {
if( SIG_ERR == signal(SIGPIPE, opal_cr_sigpipe_debug_signal_handler) ) {
;
}
}
#endif
#else
opal_cr_is_tool = true; /* no support for CR on Windows yet */
#endif /* __WINDOWS__ */
#if OPAL_ENABLE_CRDEBUG == 1
opal_cr_debug_num_free_threads = 3;
opal_cr_debug_free_threads = (opal_thread_t **)malloc(sizeof(opal_thread_t *) * opal_cr_debug_num_free_threads );
for(t = 0; t < opal_cr_debug_num_free_threads; ++t ) {
opal_cr_debug_free_threads[t] = NULL;
}
mca_base_param_reg_int_name("opal_cr", "crdebug_signal",
"Checkpoint/Restart signal used to hold threads when debugging",
false, false,
SIGTSTP,
&opal_cr_debug_signal);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Checkpoint Signal (Debug): %d",
opal_cr_debug_signal);
if( SIG_ERR == signal(opal_cr_debug_signal, MPIR_checkpoint_debugger_signal_handler) ) {
opal_output(opal_cr_output,
"opal_cr: init: Failed to register C/R debug signal (%d)",
opal_cr_debug_signal);
}
#else
/* Silence a compiler warning */
t = 0;
#endif
mca_base_param_reg_string_name("opal_cr", "tmp_dir",
"Temporary directory to place rendezvous files for a checkpoint",
false, false,
opal_tmp_directory(),
&opal_cr_pipe_dir);
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: Temp Directory: %s",
opal_cr_pipe_dir);
if( !opal_cr_is_tool ) {
/* Register the OPAL interlevel coordination callback */
opal_cr_reg_coord_callback(opal_cr_coord, &prev_coord_func);
opal_cr_stall_check = false;
opal_cr_currently_stalled = false;
} /* End opal_cr_is_tool = true */
/*
* If fault tolerance was not compiled in then
* we need to make sure that the listener thread is active to tell
* the tools that this is not a checkpointable job.
* We don't need the CRS framework to be initalized.
*/
#if OPAL_ENABLE_FT_CR == 1
/*
* Open the checkpoint / restart service components
*/
if (OPAL_SUCCESS != (ret = opal_crs_base_open())) {
opal_show_help( "help-opal-runtime.txt",
"opal_cr_init:no-crs", true,
"opal_crs_base_open", ret );
exit_status = ret;
goto cleanup;
}
if (OPAL_SUCCESS != (ret = opal_crs_base_select())) {
opal_show_help( "help-opal-runtime.txt",
"opal_cr_init:no-crs", true,
"opal_crs_base_select", ret );
exit_status = ret;
goto cleanup;
}
#endif
#if OPAL_ENABLE_FT_THREAD == 1
if( !opal_cr_is_tool && opal_cr_thread_use_if_avail) {
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: starting the thread\n");
/* JJH: We really do need this line below since it enables
* actual locks for threads. However currently the
* upper layers will deadlock if it is enabled.
* So hack around the problem for now, while working
* on a complete solution. See ticket #2741 for more
* details.
* opal_set_using_threads(true);
*/
/*
* Start the thread
*/
OBJ_CONSTRUCT(&opal_cr_thread, opal_thread_t);
OBJ_CONSTRUCT(&opal_cr_thread_lock, opal_mutex_t);
opal_cr_thread_is_done = false;
opal_cr_thread_is_active = false;
opal_cr_thread_in_library = false;
opal_cr_thread_num_in_library = 0;
opal_cr_thread.t_run = opal_cr_thread_fn;
opal_cr_thread.t_arg = NULL;
opal_thread_start(&opal_cr_thread);
} /* End opal_cr_is_tool = true */
else {
opal_output_verbose(10, opal_cr_output,
"opal_cr: init: *Not* Using C/R thread\n");
}
#endif /* OPAL_ENABLE_FT_THREAD == 1 */
cleanup:
return exit_status;
}
