PSM/PSM2: Disable signal handler hijacking by default

Per discussion on open-mpi#1767 (and some
subsequent phone calls and off-issue email discussions), the PSM and
PSM2 libraries are hijacking signal handlers by default.
Specifically: unless the environment variables `IPATH_NO_BACKTRACE=1`
(for PSM / Intel TrueScale) and `HFI_NO_BACKTRACE=1` (for PSM2 / Intel
OmniPath) are set, the library constructors for these two libraries
will hijack various signal handlers for the purpose of invoking their
own error reporting mechanisms.

This may be a bit *surprising*, but is not a *problem*, per se.  The
real problem is that older versions of at least the PSM library do not
unregister these signal handlers upon being unloaded from memory.
Hence, a segv can actually result in a double segv (i.e., the original
segv and then another segv when the now-non-existent signal handler is
invoked).

This is further compounded by the fact that the PSM / PSM2 libraries
can be loaded by the OFI MTL and the usNIC BTL (because they are
loaded by libfabric), even when there is no Intel networking hardware
present.  Having the PSM libraries behave this way when no Intel
hardware is present is clearly undesirable (and is likely to be fixed
in future releases of the PSM/PSM2 libraries).

Finally, this signal hijacking subverts Open MPI's own signal
reporting mechanism, which may be a bit surprising for some users
(particularly those who do not have Intel TrueScale/OmniPath
hardware).  As such, we disable it by default so that Open MPI's own
error-reporting mechanisms are used.

This commit will set the following two environment variables to
disable the signal hijacking from the PSM/PSM2 libraries (if they are
not already set):

* IPATH_NO_BACKTRACE=1
* HFI_NO_BACKTRACE=1

If the user has set these variables before invoking Open MPI, we will
not override their values (i.e., their preferences will be honored).

Signed-off-by: Jeff Squyres <[email protected]>

Author: jsquyres

ompi/mpi/java/c/mpi_MPI.c

@@ -12,7 +12,7 @@
  *                         All rights reserved.
  * Copyright (c) 2015      Los Alamos National Security, LLC.  All rights
  *                         reserved.
- * Copyright (c) 2015      Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2015-2016 Cisco Systems, Inc.  All rights reserved.
  * Copyright (c) 2015      Intel, Inc. All rights reserved.
  * Copyright (c) 2015      Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
@@ -131,6 +131,10 @@ OBJ_CLASS_INSTANCE(ompi_java_buffer_t,
  */
 jint JNI_OnLoad(JavaVM *vm, void *reserved)
 {
+    // Ensure that PSM signal hijacking is disabled *before* loading
+    // the library (see comment in the function for more detail).
+    opal_init_psm();
+
     libmpi = dlopen("libmpi." OPAL_DYN_LIB_SUFFIX, RTLD_NOW | RTLD_GLOBAL);
 
 #if defined(HAVE_DL_INFO) && defined(HAVE_LIBGEN_H)

opal/runtime/opal.h

@@ -10,7 +10,7 @@
  * Copyright (c) 2004-2005 The Regents of the University of California.
  *                         All rights reserved.
  * Copyright (c) 2008	   Sun Microsystems, Inc.  All rights reserved.
- * Copyright (c) 2010-2012 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2010-2016 Cisco Systems, Inc.  All rights reserved.
  * Copyright (c) 2014      Intel, Inc. All rights reserved.
  * $COPYRIGHT$
  *
@@ -76,6 +76,13 @@ OPAL_DECLSPEC int opal_finalize(void);
  */
 OPAL_DECLSPEC int opal_init_util(int* pargc, char*** pargv);
 
+/**
+ * Disable PSM/PSM2 signal hijacking.
+ *
+ * See comment in the function for more detail.
+ */
+OPAL_DECLSPEC int opal_init_psm(void);
+
 /**
  * Finalize the OPAL layer, excluding the MCA system.
  *

opal/runtime/opal_init.c

@@ -10,7 +10,7 @@
  *                         University of Stuttgart.  All rights reserved.
  * Copyright (c) 2004-2005 The Regents of the University of California.
  *                         All rights reserved.
- * Copyright (c) 2007-2012 Cisco Systems, Inc.  All rights reserved.
+ * Copyright (c) 2007-2016 Cisco Systems, Inc.  All rights reserved.
  * Copyright (c) 2007      Sun Microsystems, Inc.  All rights reserved.
  * Copyright (c) 2009      Oak Ridge National Labs.  All rights reserved.
  * Copyright (c) 2010-2015 Los Alamos National Security, LLC.
@@ -265,6 +265,31 @@ opal_err2str(int errnum, const char **errmsg)
 }
 
 
+int opal_init_psm(void)
+{
+    /* Very early in the init sequence -- before *ANY* MCA components
+       are opened -- we need to disable bad behavior from the PSM and
+       PSM2 libraries: at least some old versions of these libraries
+       hijack signal handlers during their library constructors and
+       then do not un-hijack them when the libraries are unloaded.
+
+       It is a bit of an abstraction break that we have to put
+       vendor/transport-specific code in the OPAL core, but we're
+       out of options, unfortunately.
+
+       NOTE: We only disable this behavior if the corresponding
+       environment variables are not already set (i.e., if the
+       user/environment has indicated a preference for this behavior,
+       we won't override it). */
+    if (NULL == getenv("IPATH_NO_BACKTRACE")) {
+        opal_setenv("IPATH_NO_BACKTRACE", "1", true, &environ);
+    }
+    if (NULL == getenv("HFI_NO_BACKTRACE")) {
+        opal_setenv("HFI_NO_BACKTRACE", "1", true, &environ);
+    }
+}
+
+
 int
 opal_init_util(int* pargc, char*** pargv)
 {
@@ -328,6 +353,10 @@ opal_init_util(int* pargc, char*** pargv)
         goto return_error;
     }
 
+    // Disable PSM signal hijacking (see comment in function for more
+    // details)
+    opal_init_psm();
+
     /* Setup the parameter system */
     if (OPAL_SUCCESS != (ret = mca_base_var_init())) {
         error = "mca_base_var_init";