Update hwloc for msvc.

src/3rdparty/hwloc/NEWS

@@ -1,5 +1,5 @@
 Copyright © 2009 CNRS
-Copyright © 2009-2021 Inria.  All rights reserved.
+Copyright © 2009-2022 Inria.  All rights reserved.
 Copyright © 2009-2013 Université Bordeaux
 Copyright © 2009-2011 Cisco Systems, Inc.  All rights reserved.
 Copyright © 2020 Hewlett Packard Enterprise.  All rights reserved.
@@ -17,6 +17,18 @@ bug fixes (and other actions) for each version of hwloc since version
 0.9.
 
 
+Version 2.7.1
+-------------
+* Workaround crashes when virtual machines report incoherent x86 CPUID
+  information about numbers of cores and threads.
+  Thanks to Peter Bense for the report.
+* Use setenv() instead of putenv() when trying to force enable oneAPI L0
+  support, to avoid issues with applications that touch the environment,
+  thanks to Josh Hursey for the patch.
+* Add some warnings at the end of configure when GPU libraries are
+  missing on the system or their path is missing in the environment.
+
+
 Version 2.7.0
 -------------
 * Backends

src/3rdparty/hwloc/VERSION

@@ -9,7 +9,7 @@
 
 major=2
 minor=7
-release=0
+release=1
 
 # greek is used for alpha or beta release tags.  If it is non-empty,
 # it will be appended to the version number.  It does not have to be
@@ -22,7 +22,7 @@ greek=
 
 # The date when this release was created
 
-date="Dec 06, 2021"
+date="Mar 20, 2022"
 
 # If snapshot=1, then use the value from snapshot_version as the
 # entire hwloc version (i.e., ignore major, minor, release, and
@@ -41,7 +41,7 @@ snapshot_version=${major}.${minor}.${release}${greek}-git
 # 2. Version numbers are described in the Libtool current:revision:age
 # format.
 
-libhwloc_so_version=20:2:5
+libhwloc_so_version=20:3:5
 libnetloc_so_version=0:0:0
 
 # Please also update the <TargetName> lines in contrib/windows/libhwloc.vcxproj

src/3rdparty/hwloc/include/hwloc.h

@@ -29,7 +29,7 @@
  * THAT IS IN THE PDF/HTML THAT IS ***NOT*** IN hwloc.h!
  *
  * There are entire paragraph-length descriptions, discussions, and
- * pretty prictures to explain subtle corner cases, provide concrete
+ * pretty pictures to explain subtle corner cases, provide concrete
  * examples, etc.
  *
  * Please, go read the documentation.  :-)
@@ -517,7 +517,7 @@ struct hwloc_obj {
                                           * objects).
                                           *
                                           * If the ::HWLOC_TOPOLOGY_FLAG_INCLUDE_DISALLOWED configuration flag is set,
-                                          * some of these CPUs may not be allowed for binding,
+                                          * some of these CPUs may be online but not allowed for binding,
                                           * see hwloc_topology_get_allowed_cpuset().
                                           *
 					  * \note All objects have non-NULL CPU and node sets except Misc and I/O objects.
@@ -549,7 +549,7 @@ struct hwloc_obj {
                                           * nodes more precisely.
                                           *
                                           * If the ::HWLOC_TOPOLOGY_FLAG_INCLUDE_DISALLOWED configuration flag is set,
-                                          * some of these nodes may not be allowed for allocation,
+                                          * some of these nodes may be online but not allowed for allocation,
                                           * see hwloc_topology_get_allowed_nodeset().
                                           *
                                           * If there are no NUMA nodes in the machine, all the memory is close to this
@@ -642,7 +642,7 @@ union hwloc_obj_attr_u {
     unsigned char revision;
     float linkspeed; /* in GB/s */
   } pcidev;
-  /** \brief Bridge specific Object Attribues */
+  /** \brief Bridge specific Object Attributes */
   struct hwloc_bridge_attr_s {
     union {
       struct hwloc_pcidev_attr_s pci;
@@ -1089,7 +1089,7 @@ HWLOC_DECLSPEC int hwloc_obj_add_info(hwloc_obj_t obj, const char *name, const c
  *
  * Some operating systems only support binding threads or processes to a single PU.
  * Others allow binding to larger sets such as entire Cores or Packages or
- * even random sets of invididual PUs. In such operating system, the scheduler
+ * even random sets of individual PUs. In such operating system, the scheduler
  * is free to run the task on one of these PU, then migrate it to another PU, etc.
  * It is often useful to call hwloc_bitmap_singlify() on the target CPU set before
  * passing it to the binding function to avoid these expensive migrations.
@@ -1167,7 +1167,7 @@ typedef enum {
    * CPUs are idle, operating systems may execute the thread/process
    * on those other CPUs instead of the designated CPUs, to let them
    * progress anyway.  Strict binding means that the thread/process
-   * will _never_ execute on other cpus than the designated CPUs, even
+   * will _never_ execute on other CPUs than the designated CPUs, even
    * when those are busy with other tasks and other CPUs are idle.
    *
    * \note Depending on the operating system, strict binding may not
@@ -1204,7 +1204,7 @@ typedef enum {
   HWLOC_CPUBIND_NOMEMBIND = (1<<3)
 } hwloc_cpubind_flags_t;
 
-/** \brief Bind current process or thread on cpus given in physical bitmap \p set.
+/** \brief Bind current process or thread on CPUs given in physical bitmap \p set.
  *
  * \return -1 with errno set to ENOSYS if the action is not supported
  * \return -1 with errno set to EXDEV if the binding cannot be enforced
@@ -1219,7 +1219,7 @@ HWLOC_DECLSPEC int hwloc_set_cpubind(hwloc_topology_t topology, hwloc_const_cpus
  */
 HWLOC_DECLSPEC int hwloc_get_cpubind(hwloc_topology_t topology, hwloc_cpuset_t set, int flags);
 
-/** \brief Bind a process \p pid on cpus given in physical bitmap \p set.
+/** \brief Bind a process \p pid on CPUs given in physical bitmap \p set.
  *
  * \note \p hwloc_pid_t is \p pid_t on Unix platforms,
  * and \p HANDLE on native Windows platforms.
@@ -1250,7 +1250,7 @@ HWLOC_DECLSPEC int hwloc_set_proc_cpubind(hwloc_topology_t topology, hwloc_pid_t
 HWLOC_DECLSPEC int hwloc_get_proc_cpubind(hwloc_topology_t topology, hwloc_pid_t pid, hwloc_cpuset_t set, int flags);
 
 #ifdef hwloc_thread_t
-/** \brief Bind a thread \p thread on cpus given in physical bitmap \p set.
+/** \brief Bind a thread \p thread on CPUs given in physical bitmap \p set.
  *
  * \note \p hwloc_thread_t is \p pthread_t on Unix platforms,
  * and \p HANDLE on native Windows platforms.
@@ -1914,8 +1914,9 @@ HWLOC_DECLSPEC int hwloc_topology_set_components(hwloc_topology_t __hwloc_restri
 enum hwloc_topology_flags_e {
  /** \brief Detect the whole system, ignore reservations, include disallowed objects.
    *
-   * Gather all resources, even if some were disabled by the administrator.
+   * Gather all online resources, even if some were disabled by the administrator.
    * For instance, ignore Linux Cgroup/Cpusets and gather all processors and memory nodes.
+   * However offline PUs and NUMA nodes are still ignored.
    *
    * When this flag is not set, PUs and NUMA nodes that are disallowed are not added to the topology.
    * Parent objects (package, core, cache, etc.) are added only if some of their children are allowed.
@@ -2066,10 +2067,12 @@ enum hwloc_topology_flags_e {
  *
  * Set a OR'ed set of ::hwloc_topology_flags_e onto a topology that was not yet loaded.
  *
- * If this function is called multiple times, the last invokation will erase
+ * If this function is called multiple times, the last invocation will erase
  * and replace the set of flags that was previously set.
  *
- * The flags set in a topology may be retrieved with hwloc_topology_get_flags()
+ * By default, no flags are set (\c 0).
+ *
+ * The flags set in a topology may be retrieved with hwloc_topology_get_flags().
  */
 HWLOC_DECLSPEC int hwloc_topology_set_flags (hwloc_topology_t topology, unsigned long flags);
 
@@ -2077,6 +2080,9 @@ HWLOC_DECLSPEC int hwloc_topology_set_flags (hwloc_topology_t topology, unsigned
  *
  * Get the OR'ed set of ::hwloc_topology_flags_e of a topology.
  *
+ * If hwloc_topology_set_flags() was not called earlier,
+ * no flags are set (\c 0 is returned).
+ *
  * \return the flags previously set with hwloc_topology_set_flags().
  */
 HWLOC_DECLSPEC unsigned long hwloc_topology_get_flags (hwloc_topology_t topology);

src/3rdparty/hwloc/include/hwloc/autogen/config.h

@@ -11,10 +11,10 @@
 #ifndef HWLOC_CONFIG_H
 #define HWLOC_CONFIG_H
 
-#define HWLOC_VERSION "2.7.0"
+#define HWLOC_VERSION "2.7.1"
 #define HWLOC_VERSION_MAJOR 2
 #define HWLOC_VERSION_MINOR 7
-#define HWLOC_VERSION_RELEASE 0
+#define HWLOC_VERSION_RELEASE 1
 #define HWLOC_VERSION_GREEK ""
 
 #define __hwloc_restrict

src/3rdparty/hwloc/include/hwloc/memattrs.h

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2019-2020 Inria.  All rights reserved.
+ * Copyright © 2019-2022 Inria.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
@@ -354,7 +354,7 @@ hwloc_memattr_register(hwloc_topology_t topology,
  * \p flags must be \c 0 for now.
  *
  * \note The initiator \p initiator should be of type ::HWLOC_LOCATION_TYPE_CPUSET
- * when refering to accesses performed by CPU cores.
+ * when referring to accesses performed by CPU cores.
  * ::HWLOC_LOCATION_TYPE_OBJECT is currently unused internally by hwloc,
  * but users may for instance use it to provide custom information about
  * host memory accesses performed by GPUs.
@@ -398,7 +398,7 @@ hwloc_memattr_set_value(hwloc_topology_t topology,
  * values.
  *
  * \note The initiator \p initiator should be of type ::HWLOC_LOCATION_TYPE_CPUSET
- * when refering to accesses performed by CPU cores.
+ * when referring to accesses performed by CPU cores.
  * ::HWLOC_LOCATION_TYPE_OBJECT is currently unused internally by hwloc,
  * but users may for instance use it to provide custom information about
  * host memory accesses performed by GPUs.
@@ -408,7 +408,7 @@ hwloc_memattr_get_targets(hwloc_topology_t topology,
                           hwloc_memattr_id_t attribute,
                           struct hwloc_location *initiator,
                           unsigned long flags,
-                          unsigned *nrp, hwloc_obj_t *targets, hwloc_uint64_t *values);
+                          unsigned *nr, hwloc_obj_t *targets, hwloc_uint64_t *values);
 
 /** \brief Return the initiators that have values for a given attribute for a specific target NUMA node.
  *

src/3rdparty/hwloc/src/topology-x86.c

@@ -614,10 +614,13 @@ static void look_proc(struct hwloc_backend *backend, struct procinfo *infos, uns
   eax = 0x01;
   cpuid_or_from_dump(&eax, &ebx, &ecx, &edx, src_cpuiddump);
   infos->apicid = ebx >> 24;
-  if (edx & (1 << 28))
+  if (edx & (1 << 28)) {
     legacy_max_log_proc = 1 << hwloc_flsl(((ebx >> 16) & 0xff) - 1);
-  else
+  } else {
+    hwloc_debug("HTT bit not set in CPUID 0x01.edx, assuming legacy_max_log_proc = 1\n");
     legacy_max_log_proc = 1;
+  }
+
   hwloc_debug("APIC ID 0x%02x legacy_max_log_proc %u\n", infos->apicid, legacy_max_log_proc);
   infos->ids[PKG] = infos->apicid / legacy_max_log_proc;
   legacy_log_proc_id = infos->apicid % legacy_max_log_proc;
@@ -680,12 +683,23 @@ static void look_proc(struct hwloc_backend *backend, struct procinfo *infos, uns
       unsigned max_nbcores;
       unsigned max_nbthreads;
       unsigned threadid __hwloc_attribute_unused;
+      hwloc_debug("Trying to get core/thread IDs from 0x04...\n");
       max_nbcores = ((eax >> 26) & 0x3f) + 1;
-      max_nbthreads = legacy_max_log_proc / max_nbcores;
+      hwloc_debug("found %u cores max\n", max_nbcores);
-      hwloc_debug("thus %u threads\n", max_nbthreads);
+      /* some VMs (e.g. issue#525) don't report valid information, check things before dividing by 0. */
-      threadid = legacy_log_proc_id % max_nbthreads;
+      if (!max_nbcores) {
-      infos->ids[CORE] = legacy_log_proc_id / max_nbthreads;
+        hwloc_debug("cannot detect core/thread IDs from 0x04 without a valid max of cores\n");
-      hwloc_debug("this is thread %u of core %u\n", threadid, infos->ids[CORE]);
+      } else {
+        max_nbthreads = legacy_max_log_proc / max_nbcores;
+        hwloc_debug("found %u threads max\n", max_nbthreads);
+        if (!max_nbthreads) {
+          hwloc_debug("cannot detect core/thread IDs from 0x04 without a valid max of threads\n");
+        } else {
+          threadid = legacy_log_proc_id % max_nbthreads;
+          infos->ids[CORE] = legacy_log_proc_id / max_nbthreads;
+          hwloc_debug("this is thread %u of core %u\n", threadid, infos->ids[CORE]);
+        }
+      }
     }
   }
 

src/3rdparty/hwloc/src/topology.c

@@ -3,6 +3,7 @@
  * Copyright © 2009-2021 Inria.  All rights reserved.
  * Copyright © 2009-2012, 2020 Université Bordeaux
  * Copyright © 2009-2011 Cisco Systems, Inc.  All rights reserved.
+ * Copyright © 2022 IBM Corporation.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
@@ -52,29 +53,41 @@
 #include <windows.h>
 #endif
 
+
+#ifdef HWLOC_HAVE_LEVELZERO
 /*
  * Define ZES_ENABLE_SYSMAN=1 early so that the LevelZero backend gets Sysman enabled.
+ *
+ * Only if the levelzero was enabled in this build so that we don't enable sysman
+ * for external levelzero users when hwloc doesn't need it. If somebody ever loads
+ * an external levelzero plugin in a hwloc library built without levelzero (unlikely),
+ * he may have to manually set ZES_ENABLE_SYSMAN=1.
+ *
  * Use the constructor if supported and/or the Windows DllMain callback.
  * Do it in the main hwloc library instead of the levelzero component because
  * the latter could be loaded later as a plugin.
  *
  * L0 seems to be using getenv() to check this variable on Windows
  * (at least in the Intel Compute-Runtime of March 2021),
- * so use putenv() to set the variable.
+ * but setenv() doesn't seem to exist on Windows, hence use putenv() to set the variable.
  *
  * For the record, Get/SetEnvironmentVariable() is not exactly the same as getenv/putenv():
  * - getenv() doesn't see what was set with SetEnvironmentVariable()
  * - GetEnvironmentVariable() doesn't see putenv() in cygwin (while it does in MSVC and MinGW).
  * Hence, if L0 ever switches from getenv() to GetEnvironmentVariable(),
  * it will break in cygwin, we'll have to use both putenv() and SetEnvironmentVariable().
- * Hopefully L0 will be provide a way to enable Sysman without env vars before it happens.
+ * Hopefully L0 will provide a way to enable Sysman without env vars before it happens.
  */
 #if HWLOC_HAVE_ATTRIBUTE_CONSTRUCTOR
 static void hwloc_constructor(void) __attribute__((constructor));
 static void hwloc_constructor(void)
 {
   if (!getenv("ZES_ENABLE_SYSMAN"))
-    putenv((char *) "ZES_ENABLE_SYSMAN=1");
+#ifdef HWLOC_WIN_SYS
+    putenv("ZES_ENABLE_SYSMAN=1");
+#else
+    setenv("ZES_ENABLE_SYSMAN", "1", 1);
+#endif
 }
 #endif
 #ifdef HWLOC_WIN_SYS
@@ -82,11 +95,14 @@ BOOL WINAPI DllMain(HINSTANCE hinstDLL, DWORD fdwReason, LPVOID lpReserved)
 {
   if (fdwReason == DLL_PROCESS_ATTACH) {
     if (!getenv("ZES_ENABLE_SYSMAN"))
+      /* Windows does not have a setenv, so use putenv. */
       putenv((char *) "ZES_ENABLE_SYSMAN=1");
   }
   return TRUE;
 }
 #endif
+#endif /* HWLOC_HAVE_LEVELZERO */
+
 
 unsigned hwloc_get_api_version(void)
 {