Skip to content

Fix application finalize race condition in RML #730

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 22, 2015
Merged

Fix application finalize race condition in RML #730

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
33 changes: 28 additions & 5 deletions orte/mca/rml/base/rml_base_frame.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,6 +26,7 @@

#include "orte/mca/rml/rml.h"
#include "orte/mca/state/state.h"
#include "orte/runtime/orte_wait.h"
#include "orte/util/name_fns.h"

#include "orte/mca/rml/base/base.h"
Expand Down Expand Up @@ -74,14 +75,36 @@ static int orte_rml_base_register(mca_base_register_flag_t flags)
return ORTE_SUCCESS;
}

static int orte_rml_base_close(void)
static void cleanup(int sd, short args, void *cbdata)
{
opal_list_item_t *item;
bool *active = (bool*)cbdata;

while (NULL != (item = opal_list_remove_first(&orte_rml_base.posted_recvs))) {
OBJ_RELEASE(item);
OPAL_LIST_DESTRUCT(&orte_rml_base.posted_recvs);
if (NULL != active) {
*active = false;
}
OBJ_DESTRUCT(&orte_rml_base.posted_recvs);
}

static int orte_rml_base_close(void)
{
bool active;

/* because the RML posted recvs list is in a separate
* async thread for apps, we can't just destruct it here.
* Instead, we push it into that event thread and destruct
* it there */
if (ORTE_PROC_IS_APP) {
opal_event_t ev;
active = true;
opal_event_set(orte_event_base, &ev, -1,
OPAL_EV_WRITE, cleanup, &active);
opal_event_set_priority(&ev, ORTE_ERROR_PRI);
opal_event_active(&ev, OPAL_EV_WRITE, 1);
ORTE_WAIT_FOR_COMPLETION(active);
} else {
/* we can call the destruct directly */
cleanup(0, 0, NULL);
}

OPAL_TIMING_REPORT(orte_rml_base.timing, &tm_rml);
OBJ_DESTRUCT(&orte_rml_base.open_channels);
Expand Down