Refs trac:1255

This commit repairs the debugger initialization procedure. I am not closing the ticket, however, pending Jeff's review of how it interfaces to the ompi_debugger code he implemented. There were duplicate symbols being created in that code, but not used anywhere. I replaced them with the ORTE-created symbols instead. However, since they aren't used anywhere, I have no way of checking to ensure I didn't break something.

So the ticket can be checked by Jeff when he returns from vacation... :-)

This commit was SVN r18625.

The following Trac tickets were found above:
  Ticket 1255 --> https://svn.open-mpi.org/trac/ompi/ticket/1255

ompi/debuggers/ompi_debuggers.c

@@ -64,9 +64,8 @@
 #include "ompi/datatype/datatype.h"
 #include "ompi/include/mpi.h"
 
-OMPI_DECLSPEC int MPIR_being_debugged = 0;
-OMPI_DECLSPEC volatile int MPIR_debug_gate = 0;
-OMPI_DECLSPEC volatile int MPIR_debug_state = 0;
+#include "orte/util/totalview.h"
+
 #if defined(OMPI_MSGQ_DLL)
 /* This variable is old/deprecated -- the mpimsgq_dll_locations[]
    method is preferred because it's more flexible */
@@ -85,14 +84,6 @@ OMPI_DECLSPEC int MPIR_debug_typedefs_sizeof[] = {
     sizeof(size_t)
 };
     
-struct MPIR_PROCDESC {
-    char *host_name;        /* something that can be passed to inet_addr */
-    char *executable_name;  /* name of binary */
-    int pid;                /* process pid */
-};
-OMPI_DECLSPEC struct MPIR_PROCDESC *MPIR_proctable = NULL;
-OMPI_DECLSPEC int MPIR_proctable_size = 0;
-
 /**
  * There is an issue with the debugger running on different architectures
  * compared with the debugged program. We need to know the sizes of the types

orte/mca/plm/base/plm_base_launch_support.c

@@ -27,6 +27,7 @@
 
 #include "opal/util/argv.h"
 #include "opal/runtime/opal_progress.h"
+#include "opal/class/opal_pointer_array.h"
 
 #include "opal/dss/dss.h"
 #include "orte/util/show_help.h"
@@ -47,14 +48,10 @@
 #include "orte/runtime/runtime.h"
 #include "orte/runtime/orte_locks.h"
 #include "orte/runtime/orte_wait.h"
+
 #include "orte/util/name_fns.h"
-
-
+#include "orte/util/totalview.h"
 #include "orte/util/nidmap.h"
-#include "opal/class/opal_pointer_array.h"
-
-
-#include "orte/tools/orterun/totalview.h"
 
 #include "orte/mca/plm/base/plm_private.h"
 #include "orte/mca/plm/base/base.h"
@@ -217,10 +214,9 @@ int orte_plm_base_launch_apps(orte_jobid_t job)
         return rc;
     }
     
-    /* init the debuggers */
-#if 0
+    /* init any debuggers */
     orte_totalview_init_after_spawn(job);
-#endif
+
     OPAL_OUTPUT_VERBOSE((5, orte_plm_globals.output,
                          "%s plm:base:launch completed for job %s",
                          ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),

orte/tools/orterun/Makefile.am

@@ -39,7 +39,6 @@ endif # OMPI_INSTALL_BINARIES
 orterun_SOURCES = \
         main.c \
         orterun.c \
-        orterun.h \
-        totalview.c \
-        totalview.h 
+        orterun.h
+ 
 orterun_LDADD = $(top_builddir)/orte/libopen-rte.la

orte/tools/orterun/orterun.c

@@ -71,6 +71,7 @@
 #include "orte/util/pre_condition_transports.h"
 #include "orte/util/session_dir.h"
 #include "orte/util/name_fns.h"
+#include "orte/util/totalview.h"
 
 #include "orte/mca/odls/odls.h"
 #include "orte/mca/plm/plm.h"
@@ -90,7 +91,6 @@
 #include "orte/orted/orted.h"
 
 #include "orterun.h"
-#include "totalview.h"
 
 /*
  * Globals
@@ -599,6 +599,9 @@ static void job_completed(int trigpipe, short event, void *arg)
         }
     }
     
+    /* if the debuggers were run, clean up */
+    orte_totalview_finalize();
+    
     /* the job is complete - now setup an event that will
      * trigger when the orteds are gone and tell the orteds that it is
      * okay to finalize and exit, we are done with them.
@@ -1100,7 +1103,7 @@ static int parse_globals(int argc, char* argv[], opal_cmd_line_t *cmd_line)
     /* Do we want a user-level debugger? */
 
     if (orterun_globals.debugger) {
-        orte_run_debugger(orterun_basename, cmd_line, argc, argv);
+        orte_run_debugger(orterun_basename, cmd_line, argc, argv, orterun_globals.num_procs);
     }
 
     /* Allocate and map by node or by slot?  Shortcut for setting an

orte/util/Makefile.am

@@ -36,7 +36,8 @@ headers += \
         util/dash_host/dash_host.h \
         util/comm/comm.h \
         util/nidmap.h \
-        util/show_help.h
+        util/show_help.h \
+        util/totalview.h
 
 libopen_rte_la_SOURCES += \
         util/error_strings.c \
@@ -51,5 +52,6 @@ libopen_rte_la_SOURCES += \
         util/dash_host/dash_host.c \
         util/comm/comm.c \
         util/nidmap.c \
-        util/show_help.c
+        util/show_help.c \
+        util/totalview.c
 

orte/util/show_help.h

@@ -25,8 +25,8 @@
  *
  */
 
-#ifndef _ORTE_OUTPUT_H_
-#define _ORTE_OUTPUT_H_
+#ifndef _ORTE_SHOW_HELP_H_
+#define _ORTE_SHOW_HELP_H_
 
 #include "orte_config.h"
 #include "orte/types.h"

orte/util/totalview.c

@@ -67,20 +67,13 @@
 #include "orte/runtime/runtime.h"
 #include "orte/runtime/orte_globals.h"
 
-#include "orterun.h"
-#include "totalview.h"
+#include "orte/util/totalview.h"
 
 /* +++ begin MPICH/TotalView interface definitions */
 
 #define MPIR_DEBUG_SPAWNED  1
 #define MPIR_DEBUG_ABORTING 2
 
-struct MPIR_PROCDESC {
-    char *host_name;        /* something that can be passed to inet_addr */
-    char *executable_name;  /* name of binary */
-    int pid;                /* process pid */
-};
-
 struct MPIR_PROCDESC *MPIR_proctable = NULL;
 int MPIR_proctable_size = 0;
 int MPIR_being_debugged = 0;
@@ -92,14 +85,6 @@ volatile int MPIR_acquired_pre_main = 0;
 
 /* --- end MPICH/TotalView interface definitions */
 
-/*
- * NOTE: The job description in the registry will likely evolve to use
- * the "jobgrp_t", but this works for now.
- * 
- * An initial skeleton of how to implement this with jobgrp_t is
- * available in SVN as orte/tools/orterun/totalview.c, version 7075.
- */
-
 
 #define DUMP_INT(X) fprintf(stderr, "  %s = %d\n", # X, X);
 
@@ -130,7 +115,7 @@ static void dump(void)
  * new_argv.
  */
 static int process(char *orig_line, char *basename, opal_cmd_line_t *cmd_line,
-                   int argc, char **argv, char ***new_argv) 
+                   int argc, char **argv, char ***new_argv, int num_procs) 
 {
     int i;
     char *line, *full_line = strdup(orig_line);
@@ -212,7 +197,7 @@ static int process(char *orig_line, char *basename, opal_cmd_line_t *cmd_line,
             asprintf(&tmp, "%s%s%s", line, user_argv, line + i + 14);
         } else if (0 == strncmp(line + i, "@np@", 4)) {
             line[i] = '\0';
-            asprintf(&tmp, "%s%d%s", line, orterun_globals.num_procs,
+            asprintf(&tmp, "%s%d%s", line, num_procs,
                      line + i + 4);
             used_num_procs = true;
         } else if (0 == strncmp(line + i, "@single_app@", 12)) {
@@ -276,7 +261,7 @@ static int process(char *orig_line, char *basename, opal_cmd_line_t *cmd_line,
         /* We do not support launching a debugger that requires the
            -np value if the user did not specify -np on the command
            line. */
-        if (used_num_procs && 0 == orterun_globals.num_procs) {
+        if (used_num_procs && 0 == num_procs) {
             orte_show_help("help-orterun.txt", "debugger requires -np",
                            true, (*new_argv)[0], argv[0], user_argv, 
                            (*new_argv)[0]);
@@ -321,7 +306,7 @@ static int process(char *orig_line, char *basename, opal_cmd_line_t *cmd_line,
  * Run a user-level debugger
  */
 void orte_run_debugger(char *basename, opal_cmd_line_t *cmd_line,
-                       int argc, char *argv[])
+                       int argc, char *argv[], int num_procs)
 {
     int i, id;
     char **new_argv = NULL;
@@ -350,7 +335,7 @@ void orte_run_debugger(char *basename, opal_cmd_line_t *cmd_line,
     free(value);
     for (i = 0; NULL != lines[i]; ++i) {
         if (ORTE_SUCCESS == process(lines[i], basename, cmd_line, argc, argv, 
-                                    &new_argv)) {
+                                    &new_argv, num_procs)) {
             break;
         }
     }

orte/util/totalview.h

@@ -25,13 +25,25 @@
 BEGIN_C_DECLS
 
 void orte_run_debugger(char *basename, opal_cmd_line_t *cmd_line,
-                       int argc, char *argv[]) __opal_attribute_noreturn__;
+                       int argc, char *argv[], int num_procs) __opal_attribute_noreturn__;
 void orte_totalview_init_before_spawn(void);
 void orte_totalview_init_after_spawn(orte_jobid_t jobid);
 void orte_totalview_finalize(void);
 
 ORTE_DECLSPEC extern void *MPIR_Breakpoint(void);
 
+struct MPIR_PROCDESC {
+    char *host_name;        /* something that can be passed to inet_addr */
+    char *executable_name;  /* name of binary */
+    int pid;                /* process pid */
+};
+
+ORTE_DECLSPEC extern struct MPIR_PROCDESC *MPIR_proctable;
+ORTE_DECLSPEC extern int MPIR_proctable_size;
+ORTE_DECLSPEC extern int MPIR_being_debugged;
+ORTE_DECLSPEC extern volatile int MPIR_debug_state;
+ORTE_DECLSPEC extern volatile int MPIR_debug_gate;
+
 END_C_DECLS
 
 #endif /* ORTERUN_TOTALVIEW_H */