From d28ee62a966a017b5bad50f85b2ea85f4ea44004 Mon Sep 17 00:00:00 2001
From: Matias A Cabral <matias.a.cabral@intel.com>
Date: Mon, 9 May 2016 12:08:44 -0700
Subject: [PATCH] Update in PSM and PSM2 MTLs to detect entries created by
 drivers for Intel TrueScale and Intel OmniPath, and detect a link in ACTIVE
 state. This fix addresses the scenario reported in the below OMPI users
 email, including formerly named Qlogic IB, now Intel True scale. Given the
 nature of the PSM/PSM2 mtls this fix applies to OmniPath:
 https://www.open-mpi.org/community/lists/users/2016/04/29018.php

---
 ompi/mca/mtl/psm/mtl_psm_component.c   | 38 ++++++++++++++++++++---
 ompi/mca/mtl/psm2/mtl_psm2_component.c | 42 +++++++++++++++++++++++---
 2 files changed, 71 insertions(+), 9 deletions(-)

diff --git a/ompi/mca/mtl/psm/mtl_psm_component.c b/ompi/mca/mtl/psm/mtl_psm_component.c
index b4b71b2f99..d0bc358cc0 100644
--- a/ompi/mca/mtl/psm/mtl_psm_component.c
+++ b/ompi/mca/mtl/psm/mtl_psm_component.c
@@ -37,6 +37,7 @@
 #include <sys/types.h>
 #include <sys/stat.h>
 #include <unistd.h>
+#include <glob.h>
 
 static int param_priority;
 
@@ -185,12 +186,41 @@ ompi_mtl_psm_component_open(void)
     }
 
   /* Component available only if Truescale hardware is present */
-  if (0 == stat("/dev/ipath", &st)) {
-    return OMPI_SUCCESS;
-  }
-  else {
+  if (0 != stat("/dev/ipath", &st)) {
     return OPAL_ERR_NOT_AVAILABLE;
   }
+
+  /* Component available only if at least one qib port is ACTIVE */
+  bool foundOnlineQibPort = false;
+  size_t i;
+  char portState[128];
+  FILE *devFile;
+  glob_t globbuf;
+  globbuf.gl_offs = 0;
+  if (glob("/sys/class/infiniband/qib*/ports/*/state",
+        GLOB_DOOFFS, NULL, &globbuf) != 0) {
+    return OPAL_ERR_NOT_AVAILABLE;
+  }
+
+  for (i=0;i < globbuf.gl_pathc; i++) {
+    devFile = fopen(globbuf.gl_pathv[i], "r");
+    fgets(portState, sizeof(portState), devFile);
+    fclose(devFile);
+
+    if (strstr(portState, "ACTIVE") != NULL) {
+      /* Found at least one ACTIVE port */
+      foundOnlineQibPort = true;
+      break;
+    }
+  }
+
+  globfree(&globbuf);
+
+  if (!foundOnlineQibPort) {
+    return OPAL_ERR_NOT_AVAILABLE;
+  }
+
+  return OMPI_SUCCESS;
 }
 
 static int
diff --git a/ompi/mca/mtl/psm2/mtl_psm2_component.c b/ompi/mca/mtl/psm2/mtl_psm2_component.c
index 3ae4151360..e46d993bd7 100644
--- a/ompi/mca/mtl/psm2/mtl_psm2_component.c
+++ b/ompi/mca/mtl/psm2/mtl_psm2_component.c
@@ -37,6 +37,7 @@
 #include <sys/types.h>
 #include <sys/stat.h>
 #include <unistd.h>
+#include <glob.h>
 
 static int param_priority;
 
@@ -101,15 +102,46 @@ ompi_mtl_psm2_component_register(void)
 static int
 ompi_mtl_psm2_component_open(void)
 {
-  struct stat st;
+  glob_t globbuf;
+  globbuf.gl_offs = 0;
 
   /* Component available only if Omni-Path hardware is present */
-  if (0 == stat("/dev/hfi1", &st)) {
-    return OMPI_SUCCESS;
-  }
-  else {
+  if ((glob("/dev/hfi1_[0-9]", GLOB_DOOFFS, NULL, &globbuf) != 0) &&
+      (glob("/dev/hfi1_[0-9][0-9]", GLOB_APPEND, NULL, &globbuf) != 0)) {
     return OPAL_ERR_NOT_AVAILABLE;
   }
+
+  globfree(&globbuf);
+
+  /* Component available only if at least one hfi1 port is ACTIVE */
+  bool foundOnlineHfi1Port = false;
+  size_t i;
+  char portState[128];
+  FILE *devFile;
+  if (glob("/sys/class/infiniband/hfi1_*/ports/*/state",
+        GLOB_DOOFFS, NULL, &globbuf) != 0) {
+    return OPAL_ERR_NOT_AVAILABLE;
+  }
+
+  for (i=0;i < globbuf.gl_pathc; i++) {
+    devFile = fopen(globbuf.gl_pathv[i], "r");
+    fgets(portState, sizeof(portState), devFile);
+    fclose(devFile);
+
+    if (strstr(portState, "ACTIVE") != NULL) {
+      /* Found at least one ACTIVE port */
+      foundOnlineHfi1Port = true;
+      break;
+    }
+  }
+
+  globfree(&globbuf);
+
+  if (!foundOnlineHfi1Port) {
+    return OPAL_ERR_NOT_AVAILABLE;
+  }
+
+  return OMPI_SUCCESS;
 }
 
 static int