1
1
Граф коммитов

24985 Коммитов

Автор SHA1 Сообщение Дата
Joshua Hursey
788cf1a9fe asm/powerpc: Fix empty colon list in asm for XL compiler on power
Thanks to Paul Hargrove for reporting the problem, and submitting patch.
 * https://www.open-mpi.org/community/lists/devel/2016/05/18886.php
2016-05-04 14:14:33 -05:00
Joshua Ladd
4771c9ece6 Merge pull request #1617 from jladd-mlnx/topic/disable-hcoll-barrier-in-finalize-ompi-trunk
HCOLL: fix hang in hcoll barrier called from finalize for MXM/yalla
2016-05-04 10:12:34 -04:00
Nathan Hjelm
0f54a95408 Merge pull request #1626 from hjelmn/vader_32
btl/vader: fix compilation on 32-bit systems
2016-05-03 16:39:46 -06:00
rhc54
ff8518853e Merge pull request #1604 from rhc54/topic/psm2
Improve the transport key print statement to ensure that we don't get…
2016-05-03 13:43:10 -07:00
Nathan Hjelm
4a740e9f27 Merge pull request #1619 from hjelmn/ext_verbs_fix
btl/openib: fix check for exp verbs struct members
2016-05-03 14:16:17 -06:00
Nathan Hjelm
c110d96687 Merge pull request #1622 from hjelmn/asm_fix
patcher: do not clobber ebx
2016-05-03 10:58:47 -06:00
Nathan Hjelm
e7ccbdee27 btl/vader: fix compilation on 32-bit systems
This commit fixes a compile/link issue caused by vader. The vader btl
was using OPAL_THREAD_ADD64 to increment a counter which may not be
available on 32-bit systems. Changed to use OPAL_THREAD_ADD_SIZE_T
which will be 64-bit or 32-bit depending on the system.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-05-03 10:14:44 -06:00
Aurélien Bouteiller
8344d00418 use-mpi extensions do not have a .la lib, so the fortran module should not depend on them. 2016-05-03 11:54:35 -04:00
Todd Kordenbrock
3498bed650 Merge pull request #1555 from shawone/check_reduce_ret
coll-portals4: check return value from reduce kary tree functions
2016-05-03 10:17:23 -05:00
Jeff Squyres
33dd8ca81e osc_rdma_peer: properly include ompi_config.h
Thanks to Paul Hargrove for reporting.

Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-05-03 07:39:55 -07:00
Nathan Hjelm
2d0e2b6233 patcher: do not clobber ebx
ebx can not be clobbered when using -fPIC so save and restore the
register instead of allowing it to be clobbered.

Signed-off-by: Nathan Hjelm <hjelmn@me.com>
2016-05-03 08:24:33 -06:00
Jeff Squyres
cd6ab54fdf opal_mca.m4: fix help message typo
Thanks to Paul Hargrove for reporting.

Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-05-03 07:24:06 -07:00
Jeff Squyres
cc53feaa66 Merge pull request #1621 from bgoglin/hwloc-for-2.0
linux: actually enable libudev based on the result of AC_CHECK_LIB
2016-05-03 06:34:44 -04:00
Brice Goglin
a2a721f961 linux: actually enable libudev based on the result of AC_CHECK_LIB
instead of doing AC_CHECK_HEADERS+AC_CHECK_LIB and only using the result of the former.

Thanks to Paul Hargrove for reporting the issue (OMPI build with -m32).

(cherry picked from open-mpi/hwloc@9549fd59af)
2016-05-03 10:00:40 +02:00
Nathan Hjelm
da695a6ce6 Merge pull request #1618 from hjelmn/new_hooks_update
More hook updates
2016-05-02 18:12:50 -06:00
Nathan Hjelm
566850e4b5 Merge pull request #1616 from hjelmn/rdma_fixes
osc/rdma: clear all_sync object when obtaining a lock
2016-05-02 17:58:43 -06:00
Nathan Hjelm
a65af6d079 btl/openib: fix check for exp verbs struct members
This commit fixes a compilation issue with some versions of exp
verbs. In some cases struct ibv_exp_device_attr does not have either
the exp_atom or exp_atomic_cap fields. It is fine to drop one check
and fall back to the non-exp attribute check on the other.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-05-02 17:13:33 -06:00
Nathan Hjelm
1ff79656dd patcher: remove debug fprintf
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-05-02 17:11:00 -06:00
Nathan Hjelm
581e47c271 patcher: check for clflush
Add a feature check for clflush before trying to use the clflush
instruction. As far as I can tell there is no equivalent before the
SSE2 instruction set.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-05-02 17:10:42 -06:00
Nathan Hjelm
67fd6fa6eb Merge pull request #1615 from hjelmn/new_hooks_update
memory/patcher: add #if check for MREMAP_FIXED
2016-05-02 16:26:58 -06:00
Devendar Bureddy
cafd55f18c HCOLL: fix hang in hcoll barrier called from finalize for MXM/yalla
tear down

HCOLL barrier may not complete if HCOLL progress is not called periodically.
which is the case in HCOLL teardown progress in the finalize.
(cherry picked from commit 793244d75dd94d1d5e0243bcccf6d04318750f3f)
2016-05-03 00:49:57 +03:00
Nathan Hjelm
d3d779f6d9 osc/rdma: clear all_sync object when obtaining a lock
This commit fixes a bad synchronization detection bug that occurs when
mixing MPI_Win_fence() and MPI_Win_lock(). If no communication has
occurred in the fence epoch it is safe to just clear the all_sync
object (it was set up by fence).

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-05-02 15:28:47 -06:00
Nathan Hjelm
eb14b34f04 memory/patcher: fix compilation on BSDs
The function signature of mremap on BSD (NetBSD, FreeBSD) differs from
the linux version. Added support for the BSD style of mremap.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-05-02 14:54:08 -06:00
Nathan Hjelm
52edb43bdc memory/patcher: check for linux/mman.h
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-05-02 14:29:46 -06:00
Nathan Hjelm
f8b3be6236 patcher/overwrite: fix ia64 compilation
Fixed a couple of typos in ia64 code.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-05-02 14:10:34 -06:00
Nathan Hjelm
14c34ae9f0 memory/patcher: add #if check for MREMAP_FIXED
This commit fixes a compile error when the system has mremap but not
MREMAP_FIXED. In this case we do not care about the value of
new_address as the argument does not exist.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-05-02 13:58:51 -06:00
Jeff Squyres
f3e3b800e9 opal_check_pmi.m4: remove stale code and comments
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-05-02 12:51:01 -04:00
rhc54
648043597a Merge pull request #1612 from ggouaillardet/poc/pmix_external_configury
pmix/external: revamp external pmix package detection
2016-05-02 09:46:05 -07:00
Jeff Squyres
8e0167422e NEWS: improve bullet about MPI_THREAD_MULTIPLE
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-05-02 10:24:19 -04:00
Jeff Squyres
5f258f06cb NEWS: Add bullet about MPI C++ bindings
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-05-02 06:47:38 -07:00
Jeff Squyres
265e5b9795 Merge pull request #1552 from kmroz/wip-hostname-len-cleanup-1
ompi/opal/orte/oshmem/test: max hostname length cleanup
2016-05-02 09:44:18 -04:00
Gilles Gouaillardet
45f9a47d77 pmix/external: fix typo and silence a warning 2016-05-02 17:15:52 +09:00
Gilles Gouaillardet
08d91b9a03 pmix/external: revamp external pmix package detection 2016-05-02 16:23:31 +09:00
Gilles Gouaillardet
cdac9e813d configury: fix MPI Java bindings message in summary 2016-05-02 16:21:52 +09:00
rhc54
2fa8b6c6ac Merge pull request #1525 from rhc54/topic/schizo
Extend the schizo framework
2016-05-01 15:09:08 -07:00
rhc54
c239ef5c94 Merge pull request #1606 from rhc54/topic/univ
Ensure consistency between max_procs and univ_size values - since ort…
2016-05-01 12:18:42 -07:00
Ralph Castain
6ac7929bd0 Extend the schizo framework to allow definition of CLI options by environment. Refactor orterun to mesh with the orted_submit code, thus improving code reuse. Eliminate the orte-submit tool as orterun can now meet that need.
Cleanups per @jjhursey review
2016-05-01 11:30:25 -07:00
George Bosilca
3445577f4c Avoid race conditions during BTP TCP handshake.
In some rare cases when a process receives the connect ack while
locally updating the peer endpoint structure, we could drop the
incomming connect ack due to the fact that the send handler is
protected with a try lock (on the endpoint) and our initial send
event was not persistent. Making the send event persistent solves
all issues.
2016-05-01 14:19:29 -04:00
Ralph Castain
0f05893952 Ensure consistency between max_procs and univ_size values - since orte wants max_procs, have the proc get that value instead of univ_size
Make the singleton module consistent as well
2016-05-01 11:13:33 -07:00
George Bosilca
6e6ed62a3c Allow NULL arrays for emoty datatypes.
When building an empty datatype (aka. size = 0) because the count of
included datatypes is 0, be less strict on what the arguments are
(allow NULL pointers).
2016-05-01 12:37:02 -04:00
George Bosilca
702f80ad7e Remove "signed vs. unsigned" warnings. 2016-05-01 11:45:48 -04:00
rhc54
737c62d34c Merge pull request #1608 from rhc54/topic/pmixapp
Fix minor typo in PMIx packing of pmix_app_t
2016-04-29 10:52:02 -07:00
Ralph Castain
42d9d861fc Fix minor typo in PMIx packing of pmix_app_t - thanks to Gilles for pointing it out 2016-04-29 08:55:46 -07:00
Ralph Castain
29bc24bdd5 Improve the transport key print statement to ensure that we don't get zero fields as this can be a problem for PSM 2016-04-28 20:11:12 -07:00
Nathan Hjelm
ec66a6a1f8 Merge pull request #1605 from hjelmn/rdma_fixes
osc/rdma: fix global index array calculation
2016-04-28 20:41:36 -06:00
Nathan Hjelm
7bda3eb2dc osc/rdma: fix global index array calculation
This commit fixes a bug that occurs when ranks are either not mapped
evenly or by something other than core.

Fixes open-mpi/ompi#1599

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2016-04-28 19:11:11 -06:00
rhc54
f24a6bd0eb Merge pull request #1603 from rhc54/topic/attach
Ensure the personality gets set for the debugger job launch when attaching
2016-04-28 17:04:15 -07:00
Ralph Castain
fac409d094 Ensure the personality gets set for the debugger job launch when attaching 2016-04-28 15:28:55 -07:00
Jeff Squyres
ae81f5c845 NEWS: sync to v2.x NEWS
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
2016-04-28 17:59:19 -04:00
rhc54
30cc878672 Merge pull request #1601 from rhc54/topic/oversub
Up-port of change for 2.x: if user directs oversubscribe, then do not bind as we will otherwise overload resources
2016-04-28 14:58:29 -07:00