Merge pull request #3717 from rhc54/cmr20/signal

Shift the signal forwarding code to ess/base

Author: jsquyres

orte/mca/ess/base/Makefile.am

@@ -10,7 +10,7 @@
 # Copyright (c) 2004-2005 The Regents of the University of California.
 #                         All rights reserved.
 # Copyright (c) 2013      Los Alamos National Security, LLC.  All rights reserved.
-# Copyright (c) 2015      Intel, Inc.  All rights reserved.
+# Copyright (c) 2015-2017 Intel, Inc. All rights reserved.
 # $COPYRIGHT$
 #
 # Additional copyrights may follow
@@ -26,10 +26,9 @@ headers += \
 libmca_ess_la_SOURCES += \
         base/ess_base_frame.c \
         base/ess_base_select.c \
-	base/ess_base_get.c \
-	base/ess_base_std_tool.c \
-	base/ess_base_std_app.c \
-	base/ess_base_std_orted.c \
+		base/ess_base_get.c \
+		base/ess_base_std_tool.c \
+		base/ess_base_std_app.c \
+		base/ess_base_std_orted.c \
         base/ess_base_std_prolog.c \
         base/ess_base_fns.c
-

orte/mca/ess/base/base.h

@@ -12,7 +12,7 @@
  * Copyright (c) 2011-2012 Cisco Systems, Inc.  All rights reserved.
  * Copyright (c) 2012      Oak Ridge National Labs.  All rights reserved.
  * Copyright (c) 2013      Los Alamos National Security, LLC.  All rights reserved.
- * Copyright (c) 2013      Intel, Inc. All rights reserved.
+ * Copyright (c) 2013-2017 Intel, Inc. All rights reserved.
  * $COPYRIGHT$
  *
  * Additional copyrights may follow
@@ -52,6 +52,7 @@ ORTE_DECLSPEC extern int orte_ess_base_std_buffering;
 ORTE_DECLSPEC extern int orte_ess_base_num_procs;
 ORTE_DECLSPEC extern char *orte_ess_base_jobid;
 ORTE_DECLSPEC extern char *orte_ess_base_vpid;
+ORTE_DECLSPEC extern opal_list_t orte_ess_base_signals;
 
 /*
  * Internal helper functions used by components
@@ -82,6 +83,13 @@ ORTE_DECLSPEC int orte_ess_env_put(orte_std_cntr_t num_procs,
                                    orte_std_cntr_t num_local_procs,
                                    char ***env);
 
+typedef struct {
+    opal_list_item_t super;
+    char *signame;
+    int signal;
+} orte_ess_base_signal_t;
+OBJ_CLASS_DECLARATION(orte_ess_base_signal_t);
+
 END_C_DECLS
 
 #endif

orte/mca/ess/base/ess_base_frame.c

@@ -11,6 +11,7 @@
  *                         All rights reserved.
  * Copyright (c) 2011 Cisco Systems, Inc.  All rights reserved.
  * Copyright (c) 2012      Oak Ridge National Labs.  All rights reserved.
+ * Copyright (c) 2017      Intel, Inc. All rights reserved.
  * $COPYRIGHT$
  *
  * Additional copyrights may follow
@@ -23,8 +24,10 @@
 #include "orte/constants.h"
 
 #include "orte/mca/mca.h"
+#include "opal/util/argv.h"
 #include "opal/util/output.h"
 #include "opal/mca/base/base.h"
+#include "orte/util/show_help.h"
 
 #include "orte/mca/ess/base/base.h"
 
@@ -46,6 +49,7 @@ int orte_ess_base_std_buffering = -1;
 int orte_ess_base_num_procs = -1;
 char *orte_ess_base_jobid = NULL;
 char *orte_ess_base_vpid = NULL;
+opal_list_t orte_ess_base_signals = {0};
 
 static mca_base_var_enum_value_t stream_buffering_values[] = {
   {-1, "default"},
@@ -55,6 +59,9 @@ static mca_base_var_enum_value_t stream_buffering_values[] = {
   {0, NULL}
 };
 
+static int setup_signals(void);
+static char *forwarded_signals = NULL;
+
 static int orte_ess_base_register(mca_base_register_flag_t flags)
 {
     mca_base_var_enum_t *new_enum;
@@ -96,21 +103,200 @@ static int orte_ess_base_register(mca_base_register_flag_t flags)
                                 MCA_BASE_VAR_SCOPE_READONLY, &orte_ess_base_num_procs);
     mca_base_var_register_synonym(ret, "orte", "orte", "ess", "num_procs", 0);
 
+    forwarded_signals = NULL;
+    ret = mca_base_var_register ("orte", "ess", "base", "forward_signals",
+                                 "Comma-delimited list of additional signals (names or integers) to forward to "
+                                 "application processes [\"none\" => forward nothing]. Signals provided by "
+                                 "default include SIGTSTP, SIGUSR1, SIGUSR2, SIGABRT, SIGALRM, and SIGCONT",
+                                 MCA_BASE_VAR_TYPE_STRING, NULL, 0, 0,
+                                 OPAL_INFO_LVL_4, MCA_BASE_VAR_SCOPE_READONLY,
+                                 &forwarded_signals);
+    mca_base_var_register_synonym(ret, "orte", "ess", "hnp", "forward_signals", 0);
+
+
     return ORTE_SUCCESS;
 }
 
 static int orte_ess_base_close(void)
 {
+    OPAL_LIST_DESTRUCT(&orte_ess_base_signals);
+
     return mca_base_framework_components_close(&orte_ess_base_framework, NULL);
 }
 
 static int orte_ess_base_open(mca_base_open_flag_t flags)
 {
+    int rc;
+
+    OBJ_CONSTRUCT(&orte_ess_base_signals, opal_list_t);
+
+    if (ORTE_PROC_IS_HNP || ORTE_PROC_IS_DAEMON) {
+        if (ORTE_SUCCESS != (rc = setup_signals())) {
+            return rc;
+        }
+    }
     return mca_base_framework_components_open(&orte_ess_base_framework, flags);
 }
 
 MCA_BASE_FRAMEWORK_DECLARE(orte, ess, "ORTE Environmenal System Setup",
                            orte_ess_base_register, orte_ess_base_open, orte_ess_base_close,
                            mca_ess_base_static_components, 0);
 
+/* signal forwarding */
+
+/* setup signal forwarding list */
+struct known_signal {
+    /** signal number */
+    int signal;
+    /** signal name */
+    char *signame;
+    /** can this signal be forwarded */
+    bool can_forward;
+};
+
+static struct known_signal known_signals[] = {
+    {SIGTERM, "SIGTERM", false},
+    {SIGHUP, "SIGHUP", false},
+    {SIGINT, "SIGINT", false},
+    {SIGKILL, "SIGKILL", false},
+#ifdef SIGSYS
+    {SIGSYS, "SIGSYS", true},
+#endif
+#ifdef SIGXCPU
+    {SIGXCPU, "SIGXCPU", true},
+#endif
+    {SIGXFSZ, "SIGXFSZ", true},
+#ifdef SIGVTALRM
+    {SIGVTALRM, "SIGVTALRM", true},
+#endif
+#ifdef SIGPROF
+    {SIGPROF, "SIGPROF", true},
+#endif
+#ifdef SIGINFO
+    {SIGINFO, "SIGINFO", true},
+#endif
+#ifdef SIGPWR
+    {SIGPWR, "SIGPWR", true},
+#endif
+#ifdef SIGURG
+    {SIGURG, "SIGURG", true},
+#endif
+#ifdef SIGUSR1
+    {SIGUSR1, "SIGUSR1", true},
+#endif
+#ifdef SIGUSR2
+    {SIGUSR2, "SIGUSR2", true},
+#endif
+    {0, NULL},
+};
+
+#define ESS_ADDSIGNAL(x, s)                                                 \
+    do {                                                                    \
+        orte_ess_base_signal_t *_sig;                                       \
+        _sig = OBJ_NEW(orte_ess_base_signal_t);                             \
+        _sig->signal = (x);                                                 \
+        _sig->signame = strdup((s));                                        \
+        opal_list_append(&orte_ess_base_signals, &_sig->super);             \
+    } while(0)
+
+static int setup_signals(void)
+{
+    int i, sval, nsigs;
+    char **signals, *tmp;
+    orte_ess_base_signal_t *sig;
+    bool ignore, found;
+
+    /* if they told us "none", then nothing to do */
+    if (NULL != forwarded_signals &&
+        0 == strcmp(forwarded_signals, "none")) {
+        return ORTE_SUCCESS;
+    }
 
+    /* we know that some signals are (nearly) always defined, regardless
+     * of environment, so add them here */
+    nsigs = sizeof(known_signals) / sizeof(struct known_signal);
+    for (i=0; i < nsigs; i++) {
+        if (known_signals[i].can_forward) {
+            ESS_ADDSIGNAL(known_signals[i].signal, known_signals[i].signame);
+        }
+    }
+
+    /* see if they asked for anything beyond those - note that they may
+     * have asked for some we already cover, and so we ignore any duplicates */
+    if (NULL != forwarded_signals) {
+        /* if they told us "none", then dump the list */
+        signals = opal_argv_split(forwarded_signals, ',');
+        for (i=0; NULL != signals[i]; i++) {
+            sval = 0;
+            if (0 != strncmp(signals[i], "SIG", 3)) {
+                /* treat it like a number */
+                errno = 0;
+                sval = strtoul(signals[i], &tmp, 10);
+                if (0 != errno || '\0' != *tmp) {
+                    orte_show_help("help-ess-base.txt", "ess-base:unknown-signal",
+                                   true, signals[i], forwarded_signals);
+                    opal_argv_free(signals);
+                    return OPAL_ERR_SILENT;
+                }
+            }
+
+            /* see if it is one we already covered */
+            ignore = false;
+            OPAL_LIST_FOREACH(sig, &orte_ess_base_signals, orte_ess_base_signal_t) {
+                if (0 == strcasecmp(signals[i], sig->signame) || sval == sig->signal) {
+                    /* got it - we will ignore */
+                    ignore = true;
+                    break;
+                }
+            }
+
+            if (ignore) {
+                continue;
+            }
+
+            /* see if they gave us a signal name */
+            found = false;
+            for (int j = 0 ; known_signals[j].signame ; ++j) {
+                if (0 == strcasecmp (signals[i], known_signals[j].signame) || sval == known_signals[j].signal) {
+                    if (!known_signals[j].can_forward) {
+                        orte_show_help("help-ess-base.txt", "ess-base:cannot-forward",
+                                       true, known_signals[j].signame, forwarded_signals);
+                        opal_argv_free(signals);
+                        return OPAL_ERR_SILENT;
+                    }
+                    found = true;
+                    ESS_ADDSIGNAL(known_signals[j].signal, known_signals[j].signame);
+                    break;
+                }
+            }
+
+            if (!found) {
+                if (0 == strncmp(signals[i], "SIG", 3)) {
+                    orte_show_help("help-ess-base.txt", "ess-base:unknown-signal",
+                                   true, signals[i], forwarded_signals);
+                    opal_argv_free(signals);
+                    return OPAL_ERR_SILENT;
+                }
+
+                ESS_ADDSIGNAL(sval, signals[i]);
+            }
+        }
+        opal_argv_free (signals);
+    }
+    return ORTE_SUCCESS;
+}
+
+/* instantiate the class */
+static void scon(orte_ess_base_signal_t *t)
+{
+    t->signame = NULL;
+}
+static void sdes(orte_ess_base_signal_t *t)
+{
+    if (NULL != t->signame) {
+        free(t->signame);
+    }
+}
+OBJ_CLASS_INSTANCE(orte_ess_base_signal_t,
+                   opal_list_item_t,
+                   scon, sdes);

orte/mca/ess/base/help-ess-base.txt

@@ -10,6 +10,7 @@
 #                         University of Stuttgart.  All rights reserved.
 # Copyright (c) 2004-2005 The Regents of the University of California.
 #                         All rights reserved.
+# Copyright (c) 2017      Intel, Inc. All rights reserved.
 # $COPYRIGHT$
 #
 # Additional copyrights may follow
@@ -32,3 +33,19 @@ and got the error %s.
 This could mean that your PATH or executable name is wrong, or that you do not
 have the necessary permissions.  Please ensure that the executable is able to be
 found and executed as it is required for singleton operations.
+[ess-base:cannot-forward]
+The system does not support trapping and forwarding of the
+specified signal:
+
+  signal: %s
+  param:  %s
+
+Please remove that signal from the ess_base_forward_signals MCA parameter.
+[ess-base:unknown-signal]
+The following signal was included in the ess_base_forward_signals
+MCA parameter:
+
+  signal: %s
+  param:  %s
+
+This is not a recognized signal value. Please fix or remove it.

orte/mca/ess/hnp/Makefile.am

@@ -20,8 +20,6 @@
 # $HEADER$
 #
 
-dist_ortedata_DATA = help-ess-hnp.txt
-
 sources = \
         ess_hnp.h \
         ess_hnp_component.c \

orte/mca/ess/hnp/ess_hnp.h

@@ -28,19 +28,8 @@ BEGIN_C_DECLS
 /*
  * Module open / close
  */
-typedef struct {
-    opal_list_item_t super;
-    char *signame;
-    int signal;
-} ess_hnp_signal_t;
-OBJ_CLASS_DECLARATION(ess_hnp_signal_t);
 
-typedef struct {
-    orte_ess_base_component_t base;
-    opal_list_t signals;
-} orte_ess_hnp_component_t;
-
-ORTE_MODULE_DECLSPEC extern orte_ess_hnp_component_t mca_ess_hnp_component;
+ORTE_MODULE_DECLSPEC extern orte_ess_base_component_t mca_ess_hnp_component;
 
 END_C_DECLS