Skip to content

Enable the PMIx event notification capability #1767

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
Jun 16, 2016
Merged

Enable the PMIx event notification capability #1767

Show file tree
Hide file tree
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
19 changes: 10 additions & 9 deletions ompi/errhandler/errhandler.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,7 @@
* Copyright (c) 2009 Sun Microsystems, Inc. All rights reserved.
* Copyright (c) 2015 Research Organization for Information Science
* and Technology (RIST). All rights reserved.
* Copyright (c) 2015 Intel, Inc. All rights reserved.
* Copyright (c) 2015-2016 Intel, Inc. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
Expand Down Expand Up @@ -42,7 +42,7 @@ opal_pointer_array_t ompi_errhandler_f_to_c_table = {{0}};
/*
* default errhandler id
*/
static int default_errhandler_id = -1;
static size_t default_errhandler_id = SIZE_MAX;

/*
* Class information
Expand Down Expand Up @@ -163,7 +163,7 @@ int ompi_errhandler_finalize(void)

/* JMS Add stuff here checking for unreleased errorhandlers,
similar to communicators, info handles, etc. */
opal_pmix.deregister_errhandler(default_errhandler_id, NULL, NULL);
opal_pmix.deregister_evhandler(default_errhandler_id, NULL, NULL);

/* Remove errhandler F2C table */

Expand Down Expand Up @@ -222,7 +222,7 @@ ompi_errhandler_t *ompi_errhandler_create(ompi_errhandler_type_t object_type,

/* registration callback */
void ompi_errhandler_registration_callback(int status,
int errhandler_ref,
size_t errhandler_ref,
void *cbdata)
{
ompi_errhandler_errtrk_t *errtrk = (ompi_errhandler_errtrk_t*)cbdata;
Expand All @@ -236,14 +236,15 @@ void ompi_errhandler_registration_callback(int status,
* Default errhandler callback
*/
void ompi_errhandler_callback(int status,
opal_list_t *procs,
opal_list_t *info,
opal_pmix_release_cbfunc_t cbfunc,
const opal_process_name_t *source,
opal_list_t *info, opal_list_t *results,
opal_pmix_notification_complete_fn_t cbfunc,
void *cbdata)
{
/* allow the caller to release its data */
/* tell the event chain engine to go no further - we
* will handle this */
if (NULL != cbfunc) {
cbfunc(cbdata);
cbfunc(OMPI_ERR_HANDLERS_COMPLETE, NULL, NULL, NULL, cbdata);
}
/* our default action is to abort */
ompi_mpi_abort(MPI_COMM_WORLD, status);
Expand Down
10 changes: 5 additions & 5 deletions ompi/errhandler/errhandler.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@
* All rights reserved.
* Copyright (c) 2008-2012 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2008-2009 Sun Microsystems, Inc. All rights reserved.
* Copyright (c) 2015 Intel, Inc. All rights reserved.
* Copyright (c) 2015-2016 Intel, Inc. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
Expand Down Expand Up @@ -380,13 +380,13 @@ typedef struct {
} ompi_errhandler_errtrk_t;

OMPI_DECLSPEC void ompi_errhandler_callback(int status,
opal_list_t *procs,
opal_list_t *info,
opal_pmix_release_cbfunc_t cbfunc,
const opal_process_name_t *source,
opal_list_t *info, opal_list_t *results,
opal_pmix_notification_complete_fn_t cbfunc,
void *cbdata);

OMPI_DECLSPEC void ompi_errhandler_registration_callback(int status,
int errhandler_ref,
size_t errhandler_ref,
void *cbdata);
/**
* Check to see if an errhandler is intrinsic.
Expand Down
2 changes: 2 additions & 0 deletions ompi/include/ompi/constants.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@
* University of Stuttgart. All rights reserved.
* Copyright (c) 2004-2005 The Regents of the University of California.
* All rights reserved.
* Copyright (c) 2016 Intel, Inc. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
Expand Down Expand Up @@ -63,6 +64,7 @@ enum {

OMPI_ERR_BUFFER = OPAL_ERR_BUFFER,
OMPI_ERR_SILENT = OPAL_ERR_SILENT,
OMPI_ERR_HANDLERS_COMPLETE = OPAL_ERR_HANDLERS_COMPLETE,

OMPI_ERR_REQUEST = OMPI_ERR_BASE - 1,
OMPI_ERR_RMA_SYNC = OMPI_ERR_BASE - 2,
Expand Down
65 changes: 47 additions & 18 deletions ompi/mca/rte/orte/rte_orte_module.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -95,6 +95,34 @@ void ompi_rte_abort(int error_code, char *fmt, ...)
exit(-1);
}

static size_t handler = SIZE_MAX;
static bool debugger_register_active = true;
static bool debugger_event_active = true;

static void _release_fn(int status,
const opal_process_name_t *source,
opal_list_t *info, opal_list_t *results,
opal_pmix_notification_complete_fn_t cbfunc,
void *cbdata)
{
/* must let the notifier know we are done */
if (NULL != cbfunc) {
cbfunc(ORTE_SUCCESS, NULL, NULL, NULL, cbdata);
}
debugger_event_active = false;
}

static void _register_fn(int status,
size_t evhandler_ref,
void *cbdata)
{
opal_list_t *codes = (opal_list_t*)cbdata;

handler = evhandler_ref;
OPAL_LIST_RELEASE(codes);
debugger_register_active = false;
}

/*
* Wait for a debugger if asked. We support two ways of waiting for
* attaching debuggers -- see big comment in
Expand All @@ -103,7 +131,8 @@ void ompi_rte_abort(int error_code, char *fmt, ...)
void ompi_rte_wait_for_debugger(void)
{
int debugger;
orte_rml_recv_cb_t xfer;
opal_list_t *codes;
opal_value_t *kv;

/* See lengthy comment in orte/tools/orterun/debuggers.c about
orte_in_parallel_debugger */
Expand Down Expand Up @@ -133,23 +162,23 @@ void ompi_rte_wait_for_debugger(void)
#endif
}
} else {
/* only the rank=0 proc waits for either a message from the
* HNP or for the debugger to attach - everyone else will just
* spin in * the grpcomm barrier in ompi_mpi_init until rank=0
* joins them.
*/
if (0 != ORTE_PROC_MY_NAME->vpid) {
return;
}

/* VPID 0 waits for a message from the HNP */
OBJ_CONSTRUCT(&xfer, orte_rml_recv_cb_t);
xfer.active = true;
orte_rml.recv_buffer_nb(OMPI_NAME_WILDCARD,
ORTE_RML_TAG_DEBUGGER_RELEASE,
ORTE_RML_NON_PERSISTENT,
orte_rml_recv_callback, &xfer);
/* let the MPI progress engine run while we wait */
OMPI_WAIT_FOR_COMPLETION(xfer.active);
/* register an event handler for the ORTE_ERR_DEBUGGER_RELEASE event */
codes = OBJ_NEW(opal_list_t);
kv = OBJ_NEW(opal_value_t);
kv->key = strdup("errorcode");
kv->type = OPAL_INT;
kv->data.integer = ORTE_ERR_DEBUGGER_RELEASE;
opal_list_append(codes, &kv->super);

opal_pmix.register_evhandler(codes, NULL, _release_fn, _register_fn, codes);
/* let the MPI progress engine run while we wait for registration to complete */
OMPI_WAIT_FOR_COMPLETION(debugger_register_active);

/* let the MPI progress engine run while we wait for debugger release */
OMPI_WAIT_FOR_COMPLETION(debugger_event_active);

/* deregister the event handler */
opal_pmix.deregister_evhandler(handler, NULL, NULL);
}
}
8 changes: 4 additions & 4 deletions ompi/mpi/c/lookup_name.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,10 +12,10 @@
* All rights reserved.
* Copyright (c) 2013 Los Alamos National Security, LLC. All rights
* reserved.
* Copyright (c) 2015 Intel, Inc. All rights reserved.
* Copyright (c) 2015-2016 Intel, Inc. All rights reserved.
* Copyright (c) 2015 Research Organization for Information Science
* and Technology (RIST). All rights reserved.
* Copyright (c) 2015 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2015 Cisco Systems, Inc. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
Expand Down Expand Up @@ -94,13 +94,13 @@ int MPI_Lookup_name(const char *service_name, MPI_Info info, char *port_name)
rng = OBJ_NEW(opal_value_t);
rng->key = strdup(OPAL_PMIX_RANGE);
rng->type = OPAL_INT;
rng->data.integer = OPAL_PMIX_NAMESPACE; // share only with procs in same nspace
rng->data.integer = OPAL_PMIX_RANGE_NAMESPACE; // share only with procs in same nspace
opal_list_append(&pinfo, &rng->super);
} else if (0 == strcmp(range, "session")) {
rng = OBJ_NEW(opal_value_t);
rng->key = strdup(OPAL_PMIX_RANGE);
rng->type = OPAL_INT;
rng->data.integer = OPAL_PMIX_SESSION; // share only with procs in same session
rng->data.integer = OPAL_PMIX_RANGE_SESSION; // share only with procs in same session
opal_list_append(&pinfo, &rng->super);
} else {
/* unrecognized scope */
Expand Down
8 changes: 4 additions & 4 deletions ompi/mpi/c/publish_name.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,10 +12,10 @@
* All rights reserved.
* Copyright (c) 2013 Los Alamos National Security, LLC. All rights
* reserved.
* Copyright (c) 2015 Intel, Inc. All rights reserved.
* Copyright (c) 2015-2016 Intel, Inc. All rights reserved.
* Copyright (c) 2015 Research Organization for Information Science
* and Technology (RIST). All rights reserved.
* Copyright (c) 2015 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2015 Cisco Systems, Inc. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
Expand Down Expand Up @@ -94,13 +94,13 @@ int MPI_Publish_name(const char *service_name, MPI_Info info,
rng = OBJ_NEW(opal_value_t);
rng->key = strdup(OPAL_PMIX_RANGE);
rng->type = OPAL_INT;
rng->data.integer = OPAL_PMIX_NAMESPACE; // share only with procs in same nspace
rng->data.integer = OPAL_PMIX_RANGE_NAMESPACE; // share only with procs in same nspace
opal_list_append(&values, &rng->super);
} else if (0 == strcmp(range, "session")) {
rng = OBJ_NEW(opal_value_t);
rng->key = strdup(OPAL_PMIX_RANGE);
rng->type = OPAL_INT;
rng->data.integer = OPAL_PMIX_SESSION; // share only with procs in same session
rng->data.integer = OPAL_PMIX_RANGE_SESSION; // share only with procs in same session
opal_list_append(&values, &rng->super);
} else {
/* unrecognized scope */
Expand Down
8 changes: 4 additions & 4 deletions ompi/mpi/c/unpublish_name.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,10 +12,10 @@
* All rights reserved.
* Copyright (c) 2012-2013 Los Alamos National Security, LLC. All rights
* reserved.
* Copyright (c) 2015 Intel, Inc. All rights reserved.
* Copyright (c) 2015-2016 Intel, Inc. All rights reserved.
* Copyright (c) 2015 Research Organization for Information Science
* and Technology (RIST). All rights reserved.
* Copyright (c) 2015 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2015 Cisco Systems, Inc. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
Expand Down Expand Up @@ -96,13 +96,13 @@ int MPI_Unpublish_name(const char *service_name, MPI_Info info,
rng = OBJ_NEW(opal_value_t);
rng->key = strdup(OPAL_PMIX_RANGE);
rng->type = OPAL_INT;
rng->data.integer = OPAL_PMIX_NAMESPACE; // share only with procs in same nspace
rng->data.integer = OPAL_PMIX_RANGE_NAMESPACE; // share only with procs in same nspace
opal_list_append(&pinfo, &rng->super);
} else if (0 == strcmp(range, "session")) {
rng = OBJ_NEW(opal_value_t);
rng->key = strdup(OPAL_PMIX_RANGE);
rng->type = OPAL_INT;
rng->data.integer = OPAL_PMIX_SESSION; // share only with procs in same session
rng->data.integer = OPAL_PMIX_RANGE_SESSION; // share only with procs in same session
opal_list_append(&pinfo, &rng->super);
} else {
/* unrecognized scope */
Expand Down
14 changes: 11 additions & 3 deletions ompi/runtime/ompi_mpi_init.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -377,6 +377,8 @@ int ompi_mpi_init(int argc, char **argv, int requested, int *provided)
char *cmd=NULL, *av=NULL;
ompi_errhandler_errtrk_t errtrk;
volatile bool active;
opal_list_t info;
opal_value_t *kv;
OPAL_TIMING_DECLARE(tm);
OPAL_TIMING_INIT_EXT(&tm, OPAL_TIMING_GET_TIME_OF_DAY);

Expand Down Expand Up @@ -519,10 +521,16 @@ int ompi_mpi_init(int argc, char **argv, int requested, int *provided)
/* Register the default errhandler callback */
errtrk.status = OPAL_ERROR;
errtrk.active = true;
opal_pmix.register_errhandler(NULL, ompi_errhandler_callback,
ompi_errhandler_registration_callback,
(void*)&errtrk);
/* we want to go first */
OBJ_CONSTRUCT(&info, opal_list_t);
kv = OBJ_NEW(opal_value_t);
kv->key = strdup(OPAL_PMIX_EVENT_ORDER_PREPEND);
opal_list_append(&info, &kv->super);
opal_pmix.register_evhandler(NULL, &info, ompi_errhandler_callback,
ompi_errhandler_registration_callback,
(void*)&errtrk);
OMPI_WAIT_FOR_COMPLETION(errtrk.active);
OPAL_LIST_DESTRUCT(&info);
if (OPAL_SUCCESS != errtrk.status) {
error = "Error handler registration";
ret = errtrk.status;
Expand Down
3 changes: 1 addition & 2 deletions opal/Makefile.am
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@
# Copyright (c) 2004-2005 The Regents of the University of California.
# All rights reserved.
# Copyright (c) 2009-2015 Cisco Systems, Inc. All rights reserved.
# Copyright (c) 2015 Intel, Inc. All rights reserved.
# Copyright (c) 2015-2016 Intel, Inc. All rights reserved.
# $COPYRIGHT$
#
# Additional copyrights may follow
Expand Down Expand Up @@ -76,7 +76,6 @@ nobase_opal_HEADERS = $(headers)
endif

include class/Makefile.am
include errhandler/Makefile.am
include memoryhooks/Makefile.am
include runtime/Makefile.am
include threads/Makefile.am
Expand Down
17 changes: 0 additions & 17 deletions opal/errhandler/Makefile.am
View file
Open in desktop

This file was deleted.

34 changes: 0 additions & 34 deletions opal/errhandler/opal_errhandler.c
View file
Open in desktop

This file was deleted.

25 changes: 0 additions & 25 deletions opal/errhandler/opal_errhandler.h
View file
Open in desktop

This file was deleted.

Loading