1
1
Граф коммитов

25898 Коммитов

Автор SHA1 Сообщение Дата
Valentin Petrov
9747a9ea9b coll/hcoll: ialltoallv interface 2016-10-10 15:09:07 +03:00
rhc54
b703f2e167 Merge pull request #2195 from rhc54/topic/notify
Implement the backend support for process-generated event notification
2016-10-08 13:55:19 -05:00
Ralph Castain
5b1484a836 Implement the backend support for process-generated event notification 2016-10-08 09:24:28 -07:00
Gilles Gouaillardet
315a622723 ompi: invokes opal_cleanup() if ompi_mpi_finalize() when possible
As long as it is illegal to call MPI_T_init_thread() after MPI_Finalize(),
be gentle and release as much memory as possible in MPI_Finalize().
opal_cleanup() will be invoked again by the OPAL destructor, but will
do nothing since classes was set to NULL
2016-10-08 16:58:20 +09:00
Gilles Gouaillardet
0d24fad307 opal: always run opal_class_finalize in the opal_cleanup destructor
if MPI_Init[_thread]/MPI_Finalize and MPI_T_init_thread/MPI_T_finalize
are balanced, opal_initialized is zero, and hence opal_cleanup destructor
never invokes opal_class_finalize.
if MPI_Init[_thread] nor MPI_T_init_thread have been called, classes is NULL,
so opal_class_finalize does nothing
2016-10-08 16:58:20 +09:00
Gilles Gouaillardet
b55dd2442a libevent2022: rename _event_strlcpy 2016-10-08 16:58:20 +09:00
Gilles Gouaillardet
c92e9a5406 use the new OPAL_HASH_TABLE_FOREACH convenience macro 2016-10-08 16:58:20 +09:00
Gilles Gouaillardet
23a8f764bd opal: add the OPAL_HASH_TABLE_FOREACH macro
this is a convenience macro similar to the OPAL_LIST_FOREACH macro,
that can be used to iterate on all the key/value pairs of an opal_hash_table_t
2016-10-08 16:58:20 +09:00
Gilles Gouaillardet
014f917462 opal: fix comment in OPAL_LIST_FOREACH macro. no code change. 2016-10-08 16:58:19 +09:00
Gilles Gouaillardet
13d49c135f Merge pull request #2193 from ggouaillardet/topic/pmix_misc_plugs_and_fixes
pmix3x: plugs misc memory leaks and misc fixes
2016-10-08 16:57:06 +09:00
Gilles Gouaillardet
f1f1fb15eb pmix3x: configury: output major, minor and release version after checking them
and hence fix the configure output

(back-ported from upstream commit pmix/master@7b7cdda2de)
2016-10-08 13:01:28 +09:00
Gilles Gouaillardet
f3af799608 pmix3x: misc fixes to get pmix build on Solaris
- replace MAXHOSTNAMELEN with hardcoded 1024.
  unlike Linux, Solaris #define MAXHOSTNAMELEN in <netdb.h>,
  so use a hard coded value to keep the test simpl
- stdout cannot be assigned on Solaris, so use freopen instead

(back-ported from upstream commit pmix/master@a63f6e53f4)
2016-10-08 13:01:28 +09:00
Gilles Gouaillardet
5cbfddb8f1 pmix3x: fix misc memory leaks
(back-ported from upstream commit pmix/master@1eff526929)
2016-10-08 13:01:28 +09:00
Gilles Gouaillardet
b4e4e4a5f1 pmix3x: enhance pmix_nspace_t destructor
PMIX_RELEASE all elements stored in the internal and modex hash tables

(back-ported from upstream commit pmix/master@b90674fc52)
2016-10-08 13:01:27 +09:00
Gilles Gouaillardet
f1dc033767 pmix3x: add the PMIX_HASH_TABLE_FOREACH macro
this is a convenience macro similar to the PMIX_LIST_FOREACH macro,
that can be used to iterate on all the key/value pairs of a pmix_hash_table_t

(back-ported from upstream commit pmix/master@349971c68c)
2016-10-08 13:01:27 +09:00
rhc54
73298ad4e2 Merge pull request #2192 from rhc54/topic/showhelp
Send show_help out thru stderr
2016-10-07 22:43:44 -05:00
Ralph Castain
51b2bb1d41 Send show_help out thru stderr 2016-10-07 19:23:52 -07:00
Joshua Ladd
4f1b63d9a2 Merge pull request #2188 from jladd-mlnx/topic/oshmem-bump-to-v1.3
OSHMEM Specification version: Bump to v1.3.
2016-10-07 07:15:43 -04:00
Gilles Gouaillardet
1ef2ad029f fs: do not build the fs components configured with --disable-io-ompio 2016-10-07 13:15:04 +09:00
Joshua Ladd
fe2b8b7e06 OSHMEM Specification version: Bump to v1.3. 2016-10-06 22:12:07 +03:00
Gilles Gouaillardet
6c6e35bb40 ompi/communicator: silence warnings 2016-10-06 15:03:06 +09:00
Gilles Gouaillardet
b95e243f83 ompi/errhandler: silence warnings
ISO C forbids mixing object pointer and function pointer
2016-10-06 13:20:51 +09:00
Gilles Gouaillardet
95e63d7803 cxx bindings: fix support for --disable-mpi-io configure option
Fixes open-mpi/ompi#2179
2016-10-06 09:53:59 +09:00
Jeff Squyres
67684be7c9 usnic: fix one last stray fabric_attr->name --> linux_device_name
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-10-04 18:17:38 -07:00
Todd Kordenbrock
54c46ca14e Merge pull request #2156 from tkordenbrock/topic/raccumulate.offset.fix
osc-portals4: fix offset bug in raccumulate()
2016-10-04 20:11:54 -05:00
Gilles Gouaillardet
ebb43c59d9 travis: cope with brew upgrade failure
also try to brew install gcc before brew upgrade gcc
2016-10-05 09:50:34 +09:00
Gilles Gouaillardet
bc6724567f README: document --disable-io-romio and --disable-io-ompio configure options 2016-10-05 09:14:32 +09:00
Nathan Hjelm
6cdbdceee6 Merge pull request #2064 from hjelmn/cxx_isolation
mpi/cxx: isolate internal headers from C++ bindings
2016-10-04 12:54:11 -06:00
Nathan Hjelm
c6464cae37 mpi/cxx: isolate internal headers from C++ bindings
This commit adds some glue code to support the C++ bindings and
updates the bindings to use the new glue code. This protects our
internal headers (which are C99) from C++. This is done as a quick
workaround to compilation errors when the legacy C++ bindings are
requested.

Fixes open-mpi/ompi#2055

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-10-04 11:13:25 -06:00
Todd Kordenbrock
c536e11cf3 osc-portals4: fix offset bug in raccumulate()
This commit fixes a bug where the remote offset was used as both
the local and remote offset.

Thanks to @PDeveze for the patch.
2016-10-04 09:09:17 -05:00
Jeff Squyres
f3144c7a55 Merge pull request #2152 from jsquyres/pr/usnic-improvements
usNIC BTL improvements w.r.t. libfabric bootstrapping
2016-10-03 16:46:41 -04:00
Jeff Squyres
8b77359cac usnic: remove some legacy libfabric 1.0/1.1 code
We only support running with libfabric v1.3 or greater.  So it's safe
to remove the legacy/adaptive cq_readerr() behavior.

Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-10-03 11:59:41 -07:00
Jeff Squyres
345c07a252 usnic: require libfabric >= v1.3 at run time
There are critical usnic libfabric AV insert bugs before v1.3, so
don't allow any version prior to v1.3 at run time (still allow
*compiling* with earlier versions, though, since the ABI guarantees
allow us to compile with an earlier libfabric and run with a later
libfabric).

Switch to using fi_version() to check the version (instead of calling
fi_getinfo()) as a potentially lighter-weight / simpler solution.
This allows us to only call fi_getinfo() once.

Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-10-03 11:59:41 -07:00
Jeff Squyres
b13813810f usnic: print a helpful message invoke PML error callback
The previous message was unhelpful / confusing.

Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-10-03 11:59:41 -07:00
Gilles Gouaillardet
7601e783cc pmix3x: sec/munge: add a missing include file
(cherry picked from upstream pmix/master@f7cfb11f6b)
2016-10-03 16:09:10 +09:00
rhc54
507ef670e5 Merge pull request #2148 from rhc54/topic/showhelp
Put show_help thru the PMIx "log" API.
2016-10-02 19:04:56 -05:00
Ralph Castain
e773c17cf3 Put show_help thru the PMIx "log" API. This pushes the show_help output from apps into the pmix thread, thus avoiding conflicts in the RML thread, which should help with thread lock situations. 2016-10-02 16:02:23 -07:00
Jeff Squyres
2975c7f5b8 Merge pull request #2143 from jsquyres/pr/usnic-fix-cagent-max-msg-size
usnic cagent: correctly compute the "large" ping message size
2016-09-30 20:57:21 -04:00
Jeff Squyres
545d8f2e66 usnic cagent: correctly compute the "large" ping message size
The (effective) "+42" computation was, in fact, the incorrect answer
in this case (gasp!).

We should just take the max_msg_size from the command (which came from
the libfabric endpoint max_msg_size attribute in the client) and
subtract off the max header size: 68 (which is explained in the
comment).  This will result in a "large" message size which is likely
slightly smaller than the MTU, but still right up near the MTU, and
therefore good enough.

Note: the old computation (i.e., -(68-42)) worked fine when we asked
for Libfabric API v1.1 because the usnic provider would return a
max_msg_size that was already less than the MTU due to FI_PREFIX
behavior shenanigans.  Once we started asking for Libfabric API v1.4,
the usnic Libfabric provider started returning (MTU + prefix_size),
and the -(68-42) computation started giving a value that was over the
MTU.  This caused sendto() on the connectivity checker UDP socket
to fail.

This commit also removes an old/misleading comment.

Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-09-30 17:01:05 -07:00
Joshua Hursey
fc3cf994db build: Custom libmpi_FOO name fix for wrapper compilers
* In open-mpi/ompi@f6f24a4f67 I missed
   updating the library references for the wrapper compilers.
 * Fixes the CXX wrapper compiler and CXX library is renamed as needed.
 * Fixes the Java wrapper compiler and the Java library is renamed as needed.
2016-09-30 16:40:56 -05:00
Joshua Hursey
f6f24a4f67 build: Custom libmpi(_FOO) name option in configure
* Add a configure time option to rename libmpi(_FOO).*
   - `--with-libmpi-name=STRING`
 * This commit only impacts the installed libraries.
   Internal, temporary libraries have not been renamed to limit the
   scope of the patch to only what is needed.

For example:
```shell
shell$ ./configure --with-libmpi-name=wookie
...
shell$ find . -name "libmpi*"
shell$ find . -name "libwookie*"
./lib/libwookie.so.0.0.0
./lib/libwookie.so.0
./lib/libwookie.so
./lib/libwookie.la
./lib/libwookie_mpifh.so.0.0.0
./lib/libwookie_mpifh.so.0
./lib/libwookie_mpifh.so
./lib/libwookie_mpifh.la
./lib/libwookie_usempi.so.0.0.0
./lib/libwookie_usempi.so.0
./lib/libwookie_usempi.so
./lib/libwookie_usempi.la
shell$
```
2016-09-29 21:47:24 -05:00
Gilles Gouaillardet
871ade9231 pmix/{cray,s1,s2}: make pmi_opcaddy_t class static
theses three pmix components use the same class name,
declare it as static so Open MPI can be built with --disable-dlopen

Thanks Limin Gu for the report
2016-09-28 09:18:36 +09:00
Jeff Squyres
08fd26093c Merge pull request #2125 from bertwesarg/autogen-contrib-typos
autogen.pl: Fix debug output for the MPI contribs.
2016-09-27 13:11:24 -04:00
Jeff Squyres
1a5a5fb400 Merge pull request #1861 from bharatpotnuri/master
btl/openib: Disqualify rdmacm CPC if MPI_THREAD_MULTIPLE
2016-09-27 13:03:35 -04:00
Ralph Castain
3acbc92efd If everyone is going to start using this script, then let's at least line up the entries 2016-09-27 09:05:05 -07:00
Joshua Ladd
4b0b7fd18e Merge pull request #2089 from artpol84/fix_pmix_barrier
ompi/mpi_init: fix barrier
2016-09-27 09:43:25 -04:00
Potnuri Bharat Teja
740b636dbe btl/openib: Disqualify rdmacm CPC if MPI_THREAD_MULTIPLE
The rdmacm CPC in the openib BTL is not thread safe. The rdmacm CPC
should disqualify itself (instead of failing in random ways) if
MPI_THREAD_MULTIPLE is the thread level.

Signed-off-by: Potnuri Bharat Teja <bharat@chelsio.com>
2016-09-27 14:20:59 +05:30
Gilles Gouaillardet
80beb30c58 coll/sync: plug a memory leak 2016-09-27 16:29:57 +09:00
Bert Wesarg
8c9ca83210 autogen.pl: Fix debug output for the MPI contribs. 2016-09-27 09:13:14 +02:00
Gilles Gouaillardet
0931d09afa ess/singleton: silence a valgrind warning
initialize a pointer and keep valgrind happy about it
2016-09-27 15:22:39 +09:00