ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность

Refs trac:3275.

Просмотр исходного кода 
We ran into a case where the OMPI SVN trunk grew a new acceptable MCA
parameter value, but this new value was not accepted on the v1.6
branch (hwloc_base_mem_bind_failure_action -- on the trunk it accepts
the value "silent", but on the older v1.6 branch, it doesn't).  If you
set "hwloc_base_mem_bind_failure_action=silent" in the default MCA
params file and then accidentally ran with the v1.6 branch, every OMPI
executable (including ompi_info) just failed because hwloc_base_open()
would say "hey, 'silent' is not a valid value for
hwloc_base_mem_bind_failure_action!".  Kaboom.

The only problem is that it didn't give you any indication of where
this value was being set.  Quite maddening, from a user perspective.

So we changed the ompi_info handles this case.  If any framework open
function return OMPI_ERR_BAD_PARAM (either because its base MCA params
got a bad value or because one of its component register/open
functions return OMPI_ERR_BAD_PARAM), ompi_info will stop, print out
a warning that it received and error, and then dump out the parameters
that it has received so far in the framework that had a problem.

At a minimum, this will show the user the MCA param that had an error
(it's usually the last one), and ''where it was set from'' (so that
they can go fix it).  

We updated ompi_info to check for O???_ERR_BAD_PARAM from each from
the framework opens.  Also updated the doxygen docs in mca.h for this
O???_BAD_PARAM behavior.  And we noticed that mca.h had MCA_SUCCESS
and MCA_ERR_??? codes.  Why?  I think we used them in exactly one
place in the code base (mca_base_components_open.c).  So we deleted
those and just used the normal OPAL_* codes instead.

While we were doing this, we also cleaned up a little memory
management during ompi_info/orte-info/opal-info finalization.
Valgrind still reports a truckload of memory still in use at ompi_info
termination, but they mostly look to be components not freeing
memory/resources properly (and outside the scope of this fix).

This commit was SVN r27306.

The following Trac tickets were found above:
  Ticket 3275 --> https://svn.open-mpi.org/trac/ompi/ticket/3275
Этот коммит содержится в:
Jeff Squyres 2012-09-11 20:47:24 +00:00
родитель a0ffeb205a
Коммит fb2e543a57
10 изменённых файлов: 520 добавлений и 199 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

7
ompi/mca/btl/openib/btl_openib_mca.c
Просмотреть файл

@ -9,7 +9,7 @@
* University of Stuttgart. All rights reserved.
* Copyright (c) 2004-2005 The Regents of the University of California.
* All rights reserved.
* Copyright (c) 2006-2011 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2006-2012 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2006-2009 Mellanox Technologies. All rights reserved.
* Copyright (c) 2006-2007 Los Alamos National Security, LLC. All rights
* reserved.
@ -190,7 +190,7 @@ int btl_openib_register_mca_params(void)
orte_show_help("help-mpi-btl-openib.txt",
"ibv_fork requested but not supported", true,
orte_process_info.nodename);
return OMPI_ERROR;
return OMPI_ERR_BAD_PARAM;
}
#endif
@ -221,7 +221,7 @@ int btl_openib_register_mca_params(void)
orte_show_help("help-mpi-btl-openib.txt",
"ibv_fork requested but not supported", true,
orte_process_info.nodename);
return OMPI_ERROR;
return OMPI_ERR_BAD_PARAM;
}
free(str);
@ -646,6 +646,7 @@ int btl_openib_register_mca_params(void)
}
mca_btl_openib_component.memalign_threshold = (size_t)ival;
#endif
/* Register any MCA params for the connect pseudo-components */
if (OMPI_SUCCESS == ret) {
ret = ompi_btl_openib_connect_base_register();

227
ompi/tools/ompi_info/components.c
Просмотреть файл

@ -68,9 +68,6 @@
#include "ompi/tools/ompi_info/ompi_info.h"
/*
* Public variables
*/
/*
@ -90,7 +87,7 @@ static bool opened_components = false;
int ompi_info_register_components(opal_pointer_array_t *mca_types,
opal_pointer_array_t *component_map)
{
int i;
int i, rc;
char *env, *str;
char *target, *save, *type;
char **env_save=NULL;
@ -121,8 +118,11 @@ int ompi_info_register_components(opal_pointer_array_t *mca_types,
}
/* Register the MPI layer's MCA parameters */
if (OMPI_SUCCESS != ompi_mpi_register_params()) {
str = "ompi_mpi_Register_params failed";
if (OMPI_SUCCESS != (rc = ompi_mpi_register_params())) {
str = "ompi_mpi_register_params";
if (OMPI_ERR_BAD_PARAM == rc) {
goto breakout;
}
goto error;
}
@ -131,123 +131,205 @@ int ompi_info_register_components(opal_pointer_array_t *mca_types,
map->type = strdup("base");
opal_pointer_array_add(component_map, map);
/* set default error message from here forward */
str = "A component framework failed to open properly.";
/* MPI frameworks */
if (OMPI_SUCCESS != mca_allocator_base_open()) {
if (OMPI_SUCCESS != (rc = mca_allocator_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "allocator open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("allocator");
map->components = &mca_allocator_base_components;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != mca_btl_base_open()) {
if (OMPI_ERR_BAD_PARAM == rc) {
str = "allocator";
goto breakout;
}
if (OMPI_SUCCESS != (rc = mca_btl_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "btl open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("btl");
map->components = &mca_btl_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != mca_coll_base_open()) {
if (OMPI_ERR_BAD_PARAM == rc) {
str = "btl";
goto breakout;
}
if (OMPI_SUCCESS != (rc = mca_coll_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "coll open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("coll");
map->components = &mca_coll_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OMPI_ERR_BAD_PARAM == rc) {
str = "coll";
goto breakout;
}
#if OPAL_ENABLE_FT_CR == 1
if (OMPI_SUCCESS != ompi_crcp_base_open()) {
if (OMPI_SUCCESS != (rc = ompi_crcp_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "crcp open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("crcp");
map->components = &ompi_crcp_base_components_available;
opal_pointer_array_add(component_map, map);
if (OMPI_ERR_BAD_PARAM == rc) {
str = "crcp";
goto breakout;
}
#endif
if (OMPI_SUCCESS != ompi_dpm_base_open()) {
if (OMPI_SUCCESS != (rc = ompi_dpm_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "dpm open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("dpm");
map->components = &ompi_dpm_base_components_available;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != mca_fbtl_base_open()) {
if (OMPI_ERR_BAD_PARAM == rc) {
str = "dpm";
goto breakout;
}
if (OMPI_SUCCESS != (rc = mca_fbtl_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "fbtl open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("fbtl");
map->components = &mca_fbtl_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OMPI_ERR_BAD_PARAM == rc) {
str = "fbtl";
goto breakout;
}
if (OMPI_SUCCESS != mca_fcoll_base_open()) {
if (OMPI_SUCCESS != (rc = mca_fcoll_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "fcoll open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("fcoll");
map->components = &mca_fcoll_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OMPI_ERR_BAD_PARAM == rc) {
str = "fcoll";
goto breakout;
}
if (OMPI_SUCCESS != mca_fs_base_open()) {
if (OMPI_SUCCESS != (rc = mca_fs_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "fs open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("fs");
map->components = &mca_fs_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OMPI_ERR_BAD_PARAM == rc) {
str = "fs";
goto breakout;
}
if (OMPI_SUCCESS != mca_io_base_open()) {
if (OMPI_SUCCESS != (rc = mca_io_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "io open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("io");
map->components = &mca_io_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OMPI_ERR_BAD_PARAM == rc) {
str = "io";
goto breakout;
}
if (OMPI_SUCCESS != mca_mpool_base_open()) {
if (OMPI_SUCCESS != (rc = mca_mpool_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "mpool open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("mpool");
map->components = &mca_mpool_base_components;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != ompi_mtl_base_open()) {
if (OMPI_ERR_BAD_PARAM == rc) {
str = "mpool";
goto breakout;
}
if (OMPI_SUCCESS != (rc = ompi_mtl_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "mtl open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("mtl");
map->components = &ompi_mtl_base_components_opened;
opal_pointer_array_add(component_map, map);
ompi_op_base_open();
if (OMPI_ERR_BAD_PARAM == rc) {
str = "mtl";
goto breakout;
}
if (OMPI_SUCCESS != (rc = ompi_op_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "op open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("op");
map->components = &ompi_op_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != ompi_osc_base_open()) {
if (OMPI_ERR_BAD_PARAM == rc) {
str = "op";
goto breakout;
}
if (OMPI_SUCCESS != (rc = ompi_osc_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "osc open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("osc");
map->components = &ompi_osc_base_open_components;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != mca_pml_base_open()) {
if (OMPI_ERR_BAD_PARAM == rc) {
str = "osc";
goto breakout;
}
if (OMPI_SUCCESS != (rc = mca_pml_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "pml open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("pml");
map->components = &mca_pml_base_components_available;
opal_pointer_array_add(component_map, map);
if (OMPI_ERR_BAD_PARAM == rc) {
str = "pml";
goto breakout;
}
/* No need to call the bml_base_open() because the ob1 pml calls it.
* mca_bml_base_open();
*/
@ -256,49 +338,80 @@ int ompi_info_register_components(opal_pointer_array_t *mca_types,
map->components = &mca_bml_base_components_available;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != ompi_pubsub_base_open()) {
if (OMPI_SUCCESS != (rc = ompi_pubsub_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "pubsub open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("pubsub");
map->components = &ompi_pubsub_base_components_available;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != mca_rcache_base_open()) {
if (OMPI_ERR_BAD_PARAM == rc) {
str = "pubsub";
goto breakout;
}
if (OMPI_SUCCESS != (rc = mca_rcache_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "rcache open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("rcache");
map->components = &mca_rcache_base_components;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != mca_sharedfp_base_open()) {
if (OMPI_ERR_BAD_PARAM == rc) {
str = "rcache";
goto breakout;
}
if (OMPI_SUCCESS != (rc = mca_sharedfp_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "sharedfp open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("sharedfp");
map->components = &mca_sharedfp_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != mca_topo_base_open()) {
if (OMPI_ERR_BAD_PARAM == rc) {
str = "sharedfp";
goto breakout;
}
if (OMPI_SUCCESS != (rc = mca_topo_base_open()) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "topo open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("topo");
map->components = &mca_topo_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OMPI_SUCCESS != mca_vprotocol_base_open(NULL)) {
if (OMPI_ERR_BAD_PARAM == rc) {
str = "topo";
goto breakout;
}
if (OMPI_SUCCESS != (rc = mca_vprotocol_base_open(NULL)) &&
OMPI_ERR_BAD_PARAM != rc) {
str = "vprotocol open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("vprotocol");
map->components = &mca_vprotocol_base_components_available;
opal_pointer_array_add(component_map, map);
if (OMPI_ERR_BAD_PARAM == rc) {
str = "vprotocol";
goto breakout;
}
/* flag that we need to close components */
need_close_components = true;
breakout:
/* Restore the environment to what it was before we started so that
* if users setenv OMPI_MCA_<framework name> to some value, they'll
* see that value when it is shown via --param output.
@ -310,18 +423,17 @@ int ompi_info_register_components(opal_pointer_array_t *mca_types,
}
}
/* All done */
opened_components = true;
return OMPI_SUCCESS;
error:
fprintf(stderr, "%s\n", str);
fprintf(stderr, "ompi_info will likely not display all configuration information\n");
if (need_close_components) {
opened_components = true;
ompi_info_close_components();
if (OPAL_ERR_BAD_PARAM == rc) {
fprintf(stderr, "\nA \"bad parameter\" error was encountered when opening the OMPI %s framework\n", str);
fprintf(stderr, "The output received from that framework includes the following parameters:\n\n");
}
opened_components = true;
return rc;
error:
fprintf(stderr, "ompi_info_register: %s failed\n", str);
fprintf(stderr, "ompi_info will likely not display all configuration information\n");
return OMPI_ERROR;
}
@ -338,12 +450,6 @@ void ompi_info_close_components()
* error?
*/
/* close the OPAL components */
(void) opal_info_close_components();
/* close the ORTE components */
(void) orte_info_close_components();
#if OPAL_ENABLE_FT_CR == 1
(void) ompi_crcp_base_close();
#endif
@ -364,4 +470,7 @@ void ompi_info_close_components()
(void) mca_coll_base_close();
(void) mca_allocator_base_close();
(void) ompi_osc_base_close();
/* close the ORTE components */
(void) orte_info_close_components();
}

24
ompi/tools/ompi_info/ompi_info.c
Просмотреть файл

@ -137,17 +137,31 @@ int main(int argc, char *argv[])
/* Register OPAL's params */
if (OPAL_SUCCESS != (ret = opal_info_register_components(&mca_types, &component_map))) {
exit(ret);
if (OPAL_ERR_BAD_PARAM == ret) {
/* output where the error occurred */
opal_info_err_params(&component_map);
}
exit(1);
}
/* Register ORTE's params */
if (ORTE_SUCCESS != (ret = orte_info_register_components(&mca_types, &component_map))) {
exit(ret);
if (OPAL_ERR_BAD_PARAM == ret) {
/* output what we got */
opal_info_do_params(true, opal_cmd_line_is_taken(ompi_info_cmd_line, "internal"),
&mca_types, NULL);
}
exit(1);
}
/* Register OMPI's params */
if (OMPI_SUCCESS != (ret = ompi_info_register_components(&mca_types, &component_map))) {
exit(ret);
if (OMPI_ERR_BAD_PARAM == ret) {
/* output what we got */
opal_info_do_params(true, opal_cmd_line_is_taken(ompi_info_cmd_line, "internal"),
&mca_types, NULL);
}
exit(1);
}
/* Execute the desired action(s) */
@ -219,6 +233,10 @@ int main(int argc, char *argv[])
OBJ_DESTRUCT(&component_map);
opal_info_finalize();
/* Put our own call to opal_finalize_util() here because we called
it up above (and it refcounts) */
opal_finalize_util();
return 0;
}

6
opal/mca/base/mca_base_components_open.c
Просмотреть файл

@ -9,7 +9,7 @@
* University of Stuttgart. All rights reserved.
* Copyright (c) 2004-2005 The Regents of the University of California.
* All rights reserved.
* Copyright (c) 2008 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2008-2012 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2011-2012 Los Alamos National Security, LLC.
* All rights reserved.
* $COPYRIGHT$
@ -387,7 +387,7 @@ static int open_components(const char *type_name, int output_id,
component->mca_component_name);
} else {
ret = component->mca_register_component_params();
if (MCA_SUCCESS == ret) {
if (OPAL_SUCCESS == ret) {
registered = true;
opal_output_verbose(10, output_id,
"mca: base: components_open: "
@ -427,7 +427,7 @@ static int open_components(const char *type_name, int output_id,
} else {
called_open = true;
ret = component->mca_open_component();
if (MCA_SUCCESS == ret) {
if (OPAL_SUCCESS == ret) {
opened = true;
opal_output_verbose(10, output_id,
"mca: base: components_open: "

4
opal/mca/hwloc/base/hwloc_base_open.c
Просмотреть файл

@ -1,5 +1,5 @@
/*
* Copyright (c) 2011 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2011-2012 Cisco Systems, Inc. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
@ -213,7 +213,7 @@ int opal_hwloc_base_open(void)
/* error - cannot redefine the default ranking policy */
opal_show_help("help-opal-hwloc-base.txt", "redefining-policy", true,
"core", opal_hwloc_base_print_binding(opal_hwloc_binding_policy));
return OPAL_ERR_SILENT;
return OPAL_ERR_BAD_PARAM;
}
OPAL_SET_BINDING_POLICY(opal_hwloc_binding_policy, OPAL_BIND_TO_CORE);
opal_hwloc_binding_policy |= OPAL_BIND_GIVEN;

70
opal/mca/mca.h
Просмотреть файл

@ -9,7 +9,7 @@
* University of Stuttgart. All rights reserved.
* Copyright (c) 2004-2005 The Regents of the University of California.
* All rights reserved.
* Copyright (c) 2008 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2008-2012 Cisco Systems, Inc. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
@ -110,8 +110,7 @@ typedef struct mca_base_module_2_0_0_t mca_base_module_2_0_0_t;
/**
* MCA component open function.
*
* @retval MCA_SUCCESS (or OPAL_SUCCESS or ORTE_SUCCESS or
* OMPI_SUCCESS) This component can be used in the process.
* @retval OPAL_SUCCESS This component can be used in the process.
*
* @retval OPAL_ERR_NOT_AVAILABLE Silently ignore this component for
* the duration of the process (it may even be unloaded from the
@ -122,19 +121,19 @@ typedef struct mca_base_module_2_0_0_t mca_base_module_2_0_0_t;
* be unloaded from the process).
*
* All MCA components can have an "open" function that is invoked once
* per process, when the component is located and loaded. This function
* should register any MCA parameters (using
* mca_base_param_register_int() and mca_base_param_register_string())
* that will be used by the component. Parameter registrations should
* occur in this function because the ompi_info command can be used by
* users to display all available MCA parameters (and their default
* values). However, the ompi_info command \em only invokes this open
* function on all components (i.e., no other component API methods).
* per process, when the component is located and loaded.
*
* This function should avoid registering MCA parameters (use the
* component "register" function for that; i.e.,
* mca_base_register_component_params_2_0_0_fn_t for that). Legacy
* components still register MCA params in their component "open"
* function, but their authors should update them to use the component
* "register" function.
*
* This function can also be used to allocate any resources necessary
* for the component (e.g., heap memory).
*
* This function should return MCA_SUCCESS if it wishes to remain
* This function should return OPAL_SUCCESS if it wishes to remain
* loaded in the process. Any other return value will cause the MCA
* base to unload the component. Although most components do not use
* this mechanism to force themselves to be unloaded (because if they
@ -145,14 +144,14 @@ typedef struct mca_base_module_2_0_0_t mca_base_module_2_0_0_t;
* resources to allocate, and c) can always be used in a process
* (albiet perhaps not selected), it may provide NULL for this
* function. In this cause, the MCA will act as if it called the open
* function and it returned MCA_SUCCESS.
* function and it returned OPAL_SUCCESS.
*/
typedef int (*mca_base_open_component_1_0_0_fn_t)(void);
/**
* MCA component close function.
*
* @retval MCA_SUCCESS The component successfully shut down.
* @retval OPAL_SUCCESS The component successfully shut down.
*
* @retval any_other_value Some error occurred, but is likely to be
* ignored.
@ -168,7 +167,7 @@ typedef int (*mca_base_open_component_1_0_0_fn_t)(void);
*
* If the component has no resources to free, it may provide NULL for
* this function. In this case, the MCA will act as if it called the
* close function and it returned MCA_SUCCESS.
* close function and it returned OPAL_SUCCESS.
*/
typedef int (*mca_base_close_component_1_0_0_fn_t)(void);
@ -194,8 +193,10 @@ typedef int (*mca_base_query_component_2_0_0_fn_t)(mca_base_module_2_0_0_t **mod
/**
* MCA component parameter registration function.
*
* @retval MCA_SUCCESS This component successfully registered its
* @retval OPAL_SUCCESS This component successfully registered its
* parameters and can be used in this process.
* @retval OPAL_ERR_BAD_PARAM Indicates that the register function
* failed because an MCA parameter got an invalid/incorrect value.
*
* @retval anything_else The MCA will ignore this component for the
* duration of the process.
@ -214,18 +215,27 @@ typedef int (*mca_base_query_component_2_0_0_fn_t)(mca_base_module_2_0_0_t **mod
* function invoked on a component; component authors should take care
* that no resources are leaked in this case.
*
* This function should return MCA_SUCCESS if it wishes to remain
* This function should return OPAL_SUCCESS if it wishes to remain
* loaded in the process. Any other return value will cause the MCA
* base to unload the component. Although most components do not use
* this mechanism to force themselves to be unloaded (because if they
* are immediately unloaded, ompi_info will not display them), the
* mechanism is available should the need arise.
*
* Note that if the function returns OPAL_ERR_BAD_PARAM, it is
* possible (likely?) that the component didn't register all of its
* parameters. When this happens, ompi_info (and friends) will stop
* execution and print out all existing registered parameters from the
* entire framework (since ompi_info doesn't track individual
* component register failures). This allows a user to know exactly
* what value is incorrect, and from where it was set (e.g., via an
* MCA params file).
*
* If the component a) has no MCA parameters to register, b) no
* resources to allocate, and c) can always be used in a process
* (albiet perhaps not selected), it may provide NULL for this
* function. In this cause, the MCA will act as if it called the
* registration function and it returned MCA_SUCCESS.
* registration function and it returned OPAL_SUCCESS.
*/
typedef int (*mca_base_register_component_params_2_0_0_fn_t)(void);
@ -334,28 +344,4 @@ typedef struct mca_base_component_data_2_0_0_t mca_base_component_data_2_0_0_t;
#define MCA_BASE_VERSION_2_0_0 MCA_BASE_VERSION_MAJOR, MCA_BASE_VERSION_MINOR, MCA_BASE_VERSION_RELEASE
/**
* MCA return codes.
*/
enum {
MCA_SUCCESS = 0,
/**< Success. */
MCA_ERROR = -1,
/**< General error. */
MCA_ERR_OUT_OF_RESOURCE = -2,
/**< Out of resources; a fatal error. */
MCA_ERR_TEMP_OUT_OF_RESOURCE = -3,
/**< Out of resources; try again later. */
MCA_ERR_BAD_PARAM = -5,
/**< Equivalent to MPI_ERR_ARG error code. */
MCA_ERR_NOT_IMPLEMENTED = -10,
/**< Returned by functions or functionality that has not yet been
implemented */
MCA_ERR_NOT_SUPPORTED = -11,
/**< Returned by functionality that is not supported. */
MCA_MAX_ERROR = -20
/**< Maximum error code. */
};
#endif /* OPAL_MCA_H */

160
opal/runtime/opal_info_support.c
Просмотреть файл

@ -251,7 +251,7 @@ int opal_info_register_components(opal_pointer_array_t *mca_types,
{
opal_info_component_map_t *map;
char *env, *str;
int i;
int i, rc;
char *target, *save, *type;
char **env_save=NULL;
@ -278,52 +278,77 @@ int opal_info_register_components(opal_pointer_array_t *mca_types,
}
/* some components require the event library be active, so activate it */
if (OPAL_SUCCESS != opal_event_base_open()) {
if (OPAL_SUCCESS != (rc = opal_event_base_open())) {
if (OPAL_ERR_BAD_PARAM == rc) {
goto breakout;
}
str = "opal_event_base_open";
goto error;
}
/* Open the DSS */
if (OPAL_SUCCESS != opal_dss_open()) {
str = "Unable to initialize the DSS";
if (OPAL_SUCCESS != (rc = opal_dss_open())) {
if (OPAL_ERR_BAD_PARAM == rc) {
str = "opal_dss";
goto breakout;
}
str = "dss_open";
goto error;
}
/* Register the OPAL layer's MCA parameters */
if (OPAL_SUCCESS != opal_register_params()) {
str = "opal_register_params failed";
if (OPAL_SUCCESS != (rc = opal_register_params())) {
str = "opal_register_params";
if (OPAL_ERR_BAD_PARAM == rc) {
goto breakout;
}
goto error;
}
/* OPAL frameworks */
if (OPAL_SUCCESS != opal_backtrace_base_open()) {
str = "backtrace open failed";
if (OPAL_SUCCESS != (rc = opal_backtrace_base_open()) &&
OPAL_ERR_BAD_PARAM != rc) {
str = "backtrace open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("backtrace");
map->components = &opal_backtrace_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OPAL_ERR_BAD_PARAM == rc) {
str = "backtrace";
goto breakout;
}
#if OPAL_ENABLE_FT_CR == 1
if (OPAL_SUCCESS != opal_compress_base_open()) {
str = "compress open failed";
if (OPAL_SUCCESS != (rc = opal_compress_base_open()) &&
OPAL_ERR_BAD_PARAM != rc) {
str = "compress open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("compress");
map->components = &opal_compress_base_components_available;
opal_pointer_array_add(component_map, map);
if (OPAL_ERR_BAD_PARAM == rc) {
str = "compress";
goto breakout;
}
if (OPAL_SUCCESS != opal_crs_base_open()) {
str = "crs open failed";
if (OPAL_SUCCESS != (rc = opal_crs_base_open()) &&
OPAL_ERR_BAD_PARAM != rc) {
str = "crs open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("crs");
map->components = &opal_crs_base_components_available;
opal_pointer_array_add(component_map, map);
if (OPAL_ERR_BAD_PARAM == rc) {
str = "crs";
goto breakout;
}
#endif
/* the event framework is already open - just get its components */
@ -333,24 +358,34 @@ int opal_info_register_components(opal_pointer_array_t *mca_types,
opal_pointer_array_add(component_map, map);
#if OPAL_HAVE_HWLOC
if (OPAL_SUCCESS != opal_hwloc_base_open()) {
str = "hwloc open failed";
if (OPAL_SUCCESS != (rc = opal_hwloc_base_open()) &&
OPAL_ERR_BAD_PARAM != rc) {
str = "hwloc open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("hwloc");
map->components = &opal_hwloc_base_components;
opal_pointer_array_add(component_map, map);
if (OPAL_ERR_BAD_PARAM == rc) {
str = "hwloc";
goto breakout;
}
#endif
if (OPAL_SUCCESS != opal_if_base_open()) {
str = "if open failed";
if (OPAL_SUCCESS != (rc = opal_if_base_open()) &&
OPAL_ERR_BAD_PARAM != rc) {
str = "if open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("if");
map->components = &opal_if_components;
opal_pointer_array_add(component_map, map);
if (OPAL_ERR_BAD_PARAM == rc) {
str = "if";
goto breakout;
}
/* OPAL's installdirs base open has already been called as part of
* opal_init_util() back in main().
@ -360,51 +395,76 @@ int opal_info_register_components(opal_pointer_array_t *mca_types,
map->components = &opal_installdirs_components;
opal_pointer_array_add(component_map, map);
if (OPAL_SUCCESS != opal_memory_base_open()) {
str = "memory open failed";
if (OPAL_SUCCESS != (rc = opal_memory_base_open()) &&
OPAL_ERR_BAD_PARAM != rc) {
str = "memory open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("memory");
map->components = &opal_memory_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OPAL_ERR_BAD_PARAM == rc) {
str = "memory";
goto breakout;
}
if (OPAL_SUCCESS != opal_memcpy_base_open()) {
str = "memcpy open failed";
if (OPAL_SUCCESS != (rc = opal_memcpy_base_open()) &&
OPAL_ERR_BAD_PARAM != rc) {
str = "memcpy open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("memcpy");
map->components = &opal_memcpy_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OPAL_ERR_BAD_PARAM == rc) {
str = "memcpy";
goto breakout;
}
if (OPAL_SUCCESS != opal_memchecker_base_open()) {
str = "memchecker open failed";
if (OPAL_SUCCESS != (rc = opal_memchecker_base_open()) &&
OPAL_ERR_BAD_PARAM != rc) {
str = "memchecker open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("memchecker");
map->components = &opal_memchecker_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OPAL_SUCCESS != opal_shmem_base_open()) {
str = "shmem open failed";
if (OPAL_ERR_BAD_PARAM == rc) {
str = "memchecker";
goto breakout;
}
if (OPAL_SUCCESS != (rc = opal_shmem_base_open()) &&
OPAL_ERR_BAD_PARAM != rc) {
str = "shmem open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("shmem");
map->components = &opal_shmem_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OPAL_SUCCESS != opal_timer_base_open()) {
str = "timer open failed";
if (OPAL_ERR_BAD_PARAM == rc) {
str = "shmem";
goto breakout;
}
if (OPAL_SUCCESS != (rc = opal_timer_base_open()) &&
OPAL_ERR_BAD_PARAM != rc) {
str = "timer open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("timer");
map->components = &opal_timer_base_components_opened;
opal_pointer_array_add(component_map, map);
if (OPAL_ERR_BAD_PARAM == rc) {
str = "timer";
}
breakout:
/* Restore the environment to what it was before we started so that
* if users setenv OMPI_MCA_<framework name> to some value, they'll
* see that value when it is shown via --param output.
@ -416,13 +476,19 @@ int opal_info_register_components(opal_pointer_array_t *mca_types,
}
}
return OPAL_SUCCESS;
if (OPAL_ERR_BAD_PARAM == rc) {
fprintf(stderr, "\nA \"bad parameter\" error was encountered when opening the OPAL %s framework.\n", str);
fprintf(stderr, "The output received from that framework includes the following parameters:\n\n");
}
return rc;
error:
fprintf(stderr, "opal_info_register: %s\n", str);
fprintf(stderr, "opal_info_register: %s failed\n", str);
return OPAL_ERROR;
}
void opal_info_close_components(void)
{
(void) opal_backtrace_base_close();
@ -440,10 +506,11 @@ void opal_info_close_components(void)
(void) opal_event_base_close();
/* Do not call OPAL's installdirs close; it will be handled in
* opal_finalize_util().
* opal_finalize_util(). Ditto for opal_if_base_close().
*/
}
void opal_info_show_path(const char *type, const char *value)
{
char *pretty, *path;
@ -573,8 +640,8 @@ void opal_info_do_params(bool want_all_in, bool want_internal,
if (want_all_in) {
want_all = true;
} else {
/* See if the special param "all" was givin to --param; that
* superceeds any individual type
/* See if the special param "all" was given to --param; that
* supercedes any individual type
*/
count = opal_cmd_line_get_ninsts(opal_info_cmd_line, p);
for (i = 0; i < count; ++i) {
@ -628,6 +695,27 @@ void opal_info_do_params(bool want_all_in, bool want_internal,
mca_base_param_dump_release(info);
}
void opal_info_err_params(opal_pointer_array_t *component_map)
{
opal_info_component_map_t *map, *mptr;
int i;
opal_list_t *info=NULL;
/* all we want to do is display the LAST entry in the
* component_map array as this is the one that generated the error
*/
for (i=0; i < component_map->size; i++) {
if (NULL == (mptr = (opal_info_component_map_t*)opal_pointer_array_get_item(component_map, i))) {
continue;
}
map = mptr;
}
mca_base_param_dump(&info, true);
opal_info_show_mca_params(info, map->type, opal_info_component_all, true);
fprintf(stderr, "\n");
return;
}
void opal_info_show_mca_params(opal_list_t *info,
const char *type, const char *component,
bool want_internal)
@ -640,7 +728,7 @@ void opal_info_show_mca_params(opal_list_t *info,
mca_base_param_source_t source;
char *src_file;
for (i = opal_list_get_first(info); i != opal_list_get_last(info);
for (i = opal_list_get_first(info); i != opal_list_get_end(info);
i = opal_list_get_next(i)) {
p = (mca_base_param_info_t*) i;

1
opal/runtime/opal_info_support.h
Просмотреть файл

@ -64,6 +64,7 @@ OPAL_DECLSPEC int opal_info_register_components(opal_pointer_array_t *mca_types,
opal_pointer_array_t *component_map);
OPAL_DECLSPEC void opal_info_close_components(void);
OPAL_DECLSPEC void opal_info_err_params(opal_pointer_array_t *component_map);
OPAL_DECLSPEC void opal_info_do_params(bool want_all_in, bool want_internal,
opal_pointer_array_t *mca_type,

215
orte/runtime/orte_info_support.c
Просмотреть файл

@ -87,7 +87,7 @@ int orte_info_register_components(opal_pointer_array_t *mca_types,
{
opal_info_component_map_t *map;
char *env, *str;
int i;
int i, rc;
char *target, *save, *type;
char **env_save=NULL;
@ -113,157 +113,262 @@ int orte_info_register_components(opal_pointer_array_t *mca_types,
free(env);
}
/* Set orte_process_info.proc_type to HNP to force all frameworks to
/* Set orte_process_info.proc_type to HNP to force all frameworks to
* open components
*/
orte_process_info.proc_type = ORTE_PROC_HNP;
/* Register the ORTE layer's MCA parameters */
if (ORTE_SUCCESS != orte_register_params()) {
return ORTE_ERROR;
if (ORTE_SUCCESS != (rc = orte_register_params()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "orte_register_params";
goto error;
}
if (ORTE_SUCCESS != orte_db_base_open()) {
return ORTE_ERROR;
if (ORTE_SUCCESS != (rc = orte_db_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "db_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("db");
map->components = &orte_db_base.available_components;
opal_pointer_array_add(component_map, map);
if (ORTE_SUCCESS != orte_errmgr_base_open()) {
return ORTE_ERROR;
if (ORTE_ERR_BAD_PARAM == rc) {
str = "db";
goto breakout;
}
if (ORTE_SUCCESS != (rc = orte_errmgr_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "errmgr_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("errmgr");
map->components = &orte_errmgr_base_components_available;
opal_pointer_array_add(component_map, map);
if (ORTE_SUCCESS != orte_ess_base_open()) {
return ORTE_ERROR;
if (ORTE_ERR_BAD_PARAM == rc) {
str = "errmgr";
goto breakout;
}
if (ORTE_SUCCESS != (rc = orte_ess_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "ess_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("ess");
map->components = &orte_ess_base_components_available;
opal_pointer_array_add(component_map, map);
if (ORTE_SUCCESS != orte_filem_base_open()) {
return ORTE_ERROR;
if (ORTE_ERR_BAD_PARAM == rc) {
str = "ess";
goto breakout;
}
if (ORTE_SUCCESS != (rc = orte_filem_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "filem_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("filem");
map->components = &orte_filem_base_components_available;
opal_pointer_array_add(component_map, map);
if (ORTE_ERR_BAD_PARAM == rc) {
str = "filem";
goto breakout;
}
if (ORTE_SUCCESS != orte_grpcomm_base_open()) {
return ORTE_ERROR;
if (ORTE_SUCCESS != (rc = orte_grpcomm_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "grpcomm_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("grpcomm");
map->components = &orte_grpcomm_base.components_available;
opal_pointer_array_add(component_map, map);
if (ORTE_SUCCESS != orte_iof_base_open()) {
return ORTE_ERROR;
if (ORTE_ERR_BAD_PARAM == rc) {
str = "grpcomm";
goto breakout;
}
if (ORTE_SUCCESS != (rc = orte_iof_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "iof_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("iof");
map->components = &orte_iof_base.iof_components_opened;
opal_pointer_array_add(component_map, map);
if (ORTE_SUCCESS != orte_odls_base_open()) {
return ORTE_ERROR;
if (ORTE_ERR_BAD_PARAM == rc) {
str = "iof";
goto breakout;
}
if (ORTE_SUCCESS != (rc = orte_odls_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "odls_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("odls");
map->components = &orte_odls_base.available_components;
opal_pointer_array_add(component_map, map);
if (ORTE_SUCCESS != mca_oob_base_open()) {
return ORTE_ERROR;
if (ORTE_ERR_BAD_PARAM == rc) {
str = "odls";
goto breakout;
}
if (ORTE_SUCCESS != (rc = mca_oob_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "oob_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("oob");
map->components = &mca_oob_base_components;
opal_pointer_array_add(component_map, map);
if (ORTE_SUCCESS != orte_plm_base_open()) {
return ORTE_ERROR;
if (ORTE_ERR_BAD_PARAM == rc) {
str = "oob";
goto breakout;
}
if (ORTE_SUCCESS != (rc = orte_plm_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "plm_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("plm");
map->components = &orte_plm_base.available_components;
opal_pointer_array_add(component_map, map);
if (ORTE_ERR_BAD_PARAM == rc) {
str = "plm";
goto breakout;
}
if (ORTE_SUCCESS != orte_ras_base_open()) {
return ORTE_ERROR;
if (ORTE_SUCCESS != (rc = orte_ras_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "ras_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("ras");
map->components = &orte_ras_base.ras_opened;
opal_pointer_array_add(component_map, map);
if (ORTE_SUCCESS != orte_rmaps_base_open()) {
return ORTE_ERROR;
if (ORTE_ERR_BAD_PARAM == rc) {
str = "ras";
goto breakout;
}
if (ORTE_SUCCESS != (rc = orte_rmaps_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "rmaps_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("rmaps");
map->components = &orte_rmaps_base.available_components;
opal_pointer_array_add(component_map, map);
if (ORTE_SUCCESS != orte_routed_base_open()) {
return ORTE_ERROR;
if (ORTE_ERR_BAD_PARAM == rc) {
str = "rmaps";
goto breakout;
}
if (ORTE_SUCCESS != (rc = orte_routed_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "routed_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("routed");
map->components = &orte_routed_base_components;
opal_pointer_array_add(component_map, map);
if (ORTE_SUCCESS != orte_rml_base_open()) {
return ORTE_ERROR;
if (ORTE_ERR_BAD_PARAM == rc) {
str = "routed";
goto breakout;
}
if (ORTE_SUCCESS != (rc = orte_rml_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "rml_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("rml");
map->components = &orte_rml_base_components;
opal_pointer_array_add(component_map, map);
if (ORTE_ERR_BAD_PARAM == rc) {
str = "rml";
goto breakout;
}
#if ORTE_ENABLE_SENSORS
if (ORTE_SUCCESS != orte_sensor_base_open()) {
return ORTE_ERROR;
if (ORTE_SUCCESS != (rc = orte_sensor_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "sensor_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("sensor");
map->components = &mca_sensor_base_components_available;
opal_pointer_array_add(component_map, map);
if (ORTE_ERR_BAD_PARAM == rc) {
str = "sensor";
goto breakout;
}
#endif
#if OPAL_ENABLE_FT_CR == 1
if (ORTE_SUCCESS != orte_snapc_base_open()) {
return ORTE_ERROR;
if (ORTE_SUCCESS != (rc = orte_snapc_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "snapc_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("snapc");
map->components = &orte_snapc_base_components_available;
opal_pointer_array_add(component_map, map);
if (ORTE_ERR_BAD_PARAM == rc) {
str = "snapc";
goto breakout;
}
if (ORTE_SUCCESS != orte_sstore_base_open()) {
return ORTE_ERROR;
if (ORTE_SUCCESS != (rc = orte_sstore_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "sstore_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("sstore");
map->components = &orte_sstore_base_components_available;
opal_pointer_array_add(component_map, map);
if (ORTE_ERR_BAD_PARAM == rc) {
str = "sstore";
goto breakout;
}
#endif
if (ORTE_SUCCESS != orte_state_base_open()) {
return ORTE_ERROR;
if (ORTE_SUCCESS != (rc = orte_state_base_open()) &&
ORTE_ERR_BAD_PARAM != rc) {
str = "state_base_open";
goto error;
}
map = OBJ_NEW(opal_info_component_map_t);
map->type = strdup("state");
map->components = &orte_state_base_components_available;
opal_pointer_array_add(component_map, map);
if (ORTE_ERR_BAD_PARAM == rc) {
str = "state";
goto breakout;
}
breakout:
/* Restore the environment to what it was before we started so that
* if users setenv OMPI_MCA_<framework name> to some value, they'll
* see that value when it is shown via --param output.
@ -274,7 +379,17 @@ int orte_info_register_components(opal_pointer_array_t *mca_types,
putenv(env_save[i]);
}
}
return ORTE_SUCCESS;
if (ORTE_ERR_BAD_PARAM == rc) {
fprintf(stderr, "\nA \"bad parameter\" error was encountered when opening the ORTE %s framework\n", str);
fprintf(stderr, "The output received from that framework includes the following parameters:\n\n");
}
return rc;
error:
fprintf(stderr, "orte_info_register: %s failed\n", str);
return ORTE_ERROR;
}
void orte_info_close_components(void)

5
orte/tools/orte-info/components.c
Просмотреть файл

@ -27,7 +27,7 @@
#include "orte/runtime/runtime.h"
#include "opal/util/argv.h"
#include "opal/runtime/opal_info_support.h"
#include "opal/mca/event/base/base.h"
#include "opal/util/output.h"
#include "opal/mca/base/base.h"
@ -528,6 +528,9 @@ void orte_info_close_components()
}
}
OBJ_DESTRUCT(&component_map);
/* close the OPAL components */
(void) opal_info_close_components();
}
opened_components = false;