1
1
Граф коммитов

7247 Коммитов

Автор SHA1 Сообщение Дата
Jeff Squyres
a3acc49688 usnic_component.c: don't complain if there are no usNIC devices
cmr=v1.8.2:reviewer=dgoodell

This commit was SVN r31468.
2014-04-21 19:28:48 +00:00
George Bosilca
399c8b18d2 Fix a parameter check. Thanks Lisandro Dalcin for the bug report.
More info @ http://www.open-mpi.org/community/lists/devel/2014/04/14561.php

cmr=v1.8.2:reviewer=ompi-rm1.8

This commit was SVN r31467.
2014-04-21 17:00:56 +00:00
George Bosilca
7f5314eed9 Fix the handling of the displacement array for HINDEXED_BLOCK
datatype creation.

This commit was SVN r31466.
2014-04-21 16:43:58 +00:00
Mike Dubman
6f057e57ba MXM: enable on demand mapping for only MPI mxm context
fixed by Devender, reviewed by Yossi

cmr=v1.8.2:reviewer=ompi-rm1.8

This commit was SVN r31463.
2014-04-20 09:15:37 +00:00
Jeff Squyres
a28d7af262 Remove set-but-unused variable.
This commit was SVN r31457.
2014-04-19 12:51:51 +00:00
Jeff Squyres
10e2137b97 mpi_f08 fixes from Craig's review:
* Superfluous use of MPI_User_function in comm_create_keyval_f08.F90
* Missed adding "value" keyword to function pointer arguments in pmpi
  C interfaces

Submitted by Craig, reviewed by Jeff.

Refs trac:4512

This commit was SVN r31455.

The following Trac tickets were found above:
  Ticket 4512 --> https://svn.open-mpi.org/trac/ompi/ticket/4512
2014-04-19 12:06:24 +00:00
Oscar Vega-Gisbert
f1133e2520 Java: initialize free list to only 2 buffers
This commit was SVN r31454.
2014-04-19 12:03:52 +00:00
Oscar Vega-Gisbert
19803d6605 Java - neighborhood collective communication: get buffers according topology information
This commit was SVN r31452.
2014-04-19 11:12:38 +00:00
Rolf vandeVaart
1fab9bb37f Fixes per review by jsquyres. Piggy back on btl_base_verbose rather than using my own special MCA var.
This commit was SVN r31427.
2014-04-18 18:09:09 +00:00
Nathan Hjelm
a88b24ce21 Per comments from Jeff makes some changes to the communicator changes.
Changed:

 - Use ompi_mpi_group_null instead of MPI_GROUP_NULL.

 - Asserts don't always quiet the clang static analyser. Change them to
   ifs to really quite the warnings.

cmr=v1.8.1:ticket=trac:4527:reviewer=jsquyres

This commit was SVN r31424.

The following Trac tickets were found above:
  Ticket 4527 --> https://svn.open-mpi.org/trac/ompi/ticket/4527
2014-04-18 17:23:46 +00:00
Rolf vandeVaart
a6a245b5b5 More efficient way of waiting for asynchronous copy to complete.
This commit was SVN r31420.
2014-04-17 15:18:50 +00:00
Nathan Hjelm
a03b11c20e bcol/basesmuma: fix broken allgather algorithm
The algorithm was failing ibm/collective/allgather and iallgather. I
cleaned up the code to eliminate duplicate code paths and tracked the
issue down to an error in the way extra nodes in the knomial exchange
are handled. The new code is more compact and has been tested with up
to 64 ranks with the ibm test suite.

cmr=v1.8.1:reviewer=manjugv

This commit was SVN r31419.
2014-04-16 22:43:52 +00:00
Nathan Hjelm
e125bbe347 coll/ml: clean out apparently stale code
The file coll_ml_ibarrier.c wasn't included in coll/ml's Makefile.am
and the setup code from coll_ml_hier_algorithms_ibarrier.c was not
being called. It looks like this code is stale and has long since been
replaced by the code in coll_ml_barrier.c

Once all these little CMRs are approved I may make it into one roll-up
CMR to make it easier on the RM.

cmr=v1.8.1:reviewer=manjugv

This commit was SVN r31418.
2014-04-16 22:43:43 +00:00
Nathan Hjelm
484a3f6147 coll/ml: fix issues identified by the clang static analyser and fix
a segmentation fault in the reduce cleanup

Some of the changes address false warnings produced by scan-build. I
added asserts and changed some malloc calls to calloc to silence these
warnings.

The was one issue in cleanup for reduce since the component_functions
member is changed by the allreduce call. There may be other issues
with how this code works but releasing the allocated
component_functions after setting up the static functions addresses
the primary issue (SIGSEGV).

cmr=v1.8.1:reviewer=manjugv

This commit was SVN r31417.
2014-04-16 22:43:35 +00:00
Nathan Hjelm
f80aece271 Silence warnings identified by the clang static analyzer in the
communicator code.

Many of the warnings were false warnings. These were silenced by
adding the appropriate asserts. Other warnings identified some
potential issues in error paths that should now be resolved.

cmr=v1.8.1:reviewer=jsquyres

This commit was SVN r31416.
2014-04-16 22:43:20 +00:00
Nathan Hjelm
a64bd4035c Fix bugs in intercomm creation and comm split.
This commit addresses bugs discovered by ggouaillardet.

 - Fix hang when creating an intercommunicator

 - Fix memory leak

 - Fix coverity warning cid70288

 - Fix false coverity warning cid1196589

Fixes trac:4507
Fixes trac:4522

cmr=v1.8.1:reviewer=jsquyres

This commit was SVN r31415.

The following Trac tickets were found above:
  Ticket 4507 --> https://svn.open-mpi.org/trac/ompi/ticket/4507
  Ticket 4522 --> https://svn.open-mpi.org/trac/ompi/ticket/4522
2014-04-16 22:43:12 +00:00
Jeff Squyres
ae43ae4d17 MPI_Type_indexed.3in: fix array_of_displacements type
This man page contains the prototype and descriptions for both
MPI_TYPE_INDEXED and MPI_TYPE_CREATE_HINDEXED.  Bastian Beischer
noticed that the type of the array_of_displacements argument in the
MPI_TYPE_CREATE_HINDEXED was wrong.

Also, a minor update to MPI_Type_hindexed.3in: indicate that the C
type is MPI_Aint and the Fortran type is INTEGER (which is why this
function was deprecated and then deleted by the MPI Forum!).

cmr=v1.8.1:reviewer=dgoodell

This commit was SVN r31411.
2014-04-16 15:25:37 +00:00
Rolf vandeVaart
8897e2f5bb Fix typo error in commit r31388.
This commit was SVN r31398.

The following SVN revision numbers were found above:
  r31388 --> open-mpi/ompi@ccb33ff811
2014-04-15 19:50:54 +00:00
Jeff Squyres
d46db26721 ompi_config.h: no need to AC_CONFIG_HEADER this file
Nothing is generated in this file; this commit essentially just
renames ompi_config.h.in -> ompi_config.h.

cmr=v1.8.1:reviewer=dgoodell

This commit was SVN r31395.
2014-04-15 15:38:49 +00:00
Oscar Vega-Gisbert
5a21790caa Java: improve access to Java arrays.
This commit was SVN r31391.
2014-04-14 21:29:13 +00:00
Jeff Squyres
c0f3d80c98 Change this filename back (i.e., revert r31375).
It's a singular filename because there's only 1 interface in the
file.  Also, r31375 missed updating the name in a few places, and
broke the build for compilers that supported the mpi_f08 interface.

This commit was SVN r31389.

The following SVN revision numbers were found above:
  r31375 --> open-mpi/ompi@fe1935de14
2014-04-14 19:51:06 +00:00
Nathan Hjelm
ccb33ff811 btl: Use C99 sub-object naming when initializing BTL components
Two things to note:

 - This change will allow us to expand the BTL interface without
   having to worry about modifying BTLs that will not support the new
   interfaces. More on this will come later this year as part of the
   1.9 series.

 - C99 guarantees that uninitialed members of structs declared outside
   of functions (DATA binary section) will be initialized with
   0's. This allows us to drop stuff like .btl_flags = 0, or .btl_get
   = NULL.

This commit was SVN r31388.
2014-04-14 19:29:26 +00:00
Yossi Etigin
7efb724d7b osc/rdma: fix deadlock with put_long protocol.
When sending PUT_LONG, the data is sent before headers, and sometimes 
the header is not flushed immediately. This creates a lot of unexpected 
receives in the peer, since it would posts a receive only when gets the 
header, which makes it run out of receive buffers. When the sender 
eventually flushes the window, the receiver already has no buffers to 
receive the header, which causes a deadlock.

The fix is to always flush the headers when doing put_long.

cmr=v1.8.1:reviewer=hjelmn

This commit was SVN r31378.
2014-04-13 16:24:56 +00:00
Ralph Castain
fe1935de14 Rename file to match pattern used in other names and in Makefile.am
Don't know if this needs to go to 1.8.1, but will cmr it and leave for Jeff to decide

cmr=v1.8.1:reviewer=jsquyres

This commit was SVN r31375.
2014-04-12 14:43:20 +00:00
Jeff Squyres
ed32143dc5 fortran: remove warning that compiling the "mpi" module takes a long time
Since we only builds the "small" size of the "mpi" module any more, it
does not take a long time to compile.  So remove the warning that is
emitted.

Also remove a vestage of Windows support that was leftover in the
Fortran area (i.e., building mpi.obj).

This commit was SVN r31374.
2014-04-11 21:06:43 +00:00
Jeff Squyres
5551e78275 fortran: Separate pre-defined attribute and conversion interfaces
Differentiate the pre-defined attribute and conversion interfaces into
those with INTEGER handles and those with TYPE(MPI_*) handles.

Refs trac:4157

cmr=v1.8.1:ticket=trac:4512

This commit was SVN r31372.

The following Trac tickets were found above:
  Ticket 4157 --> https://svn.open-mpi.org/trac/ompi/ticket/4157
  Ticket 4512 --> https://svn.open-mpi.org/trac/ompi/ticket/4512
2014-04-11 21:01:16 +00:00
Jeff Squyres
65b95e2844 mpi_f08: pass function pointers from Fortran to C properly
Use type(c_funptr) to "cast" the fortran function pointers to
arbitrary C pointers.  In C, we then pick up the appropriate function
pointer type.

Tested with ifort 14.0.2 and gfortran 4.9 snapshot (which is what
identified that the previous method of passing function pointers was
not Fortran'08-compliant).

Refs trac:4157

This commit was SVN r31371.

The following Trac tickets were found above:
  Ticket 4157 --> https://svn.open-mpi.org/trac/ompi/ticket/4157
2014-04-11 20:56:11 +00:00
Jeff Squyres
0ceebe16f2 mpi.h.in: We're long past the v1.5 series; this comment can go away
This commit was SVN r31369.
2014-04-11 15:17:48 +00:00
Jeff Squyres
619dd71a92 mpi.h.in: fix comment to match code
Thanks to Lisandro Dalcin for noticing.

This commit was SVN r31366.
2014-04-10 17:29:14 +00:00
Jeff Squyres
12e82daa78 mpi_f08: Fix param name and string length for MPI_GET_LIBRARY_VERSION
Junchao Zhang pointed out to me that we had the wrong parameter name
and string length specification for the "version" parameter.  This
matters because Fortran allows passing by parameter name
(vs. parameter ordering).  Specifically, we had the interface as:

{{{
subroutine MPI_Get_library_version_f08(name,resultlen,ierror)
   character(len=MPI_MAX_PROCESSOR_NAME), intent(out) :: name
...etc.
}}}

but it should be:

{{{
subroutine MPI_Get_library_version_f08(version,resultlen,ierror)
   character(len=MPI_MAX_LIBRARY_VERSION_STRING), intent(out) :: version
...etc.
}}}

Thankfully, MPI_MAX_PROCESSOR_NAME and MPI_MAX_LIBRARY_VERSION_STRING
are both 255 in OMPI, so there's no ABI issue caused by changing the
length from MMPN --> MMLVS.

The ABI is also unaffected by the parameter name change: if you
compile/link an MPI application calling MPI_GET_LIBRARY_VERSION with
1.8, it'll still run-time link with this change.

However, if an MPI program compiled using parameter name passing with
the old/incorrect parameter name ("name"), it won't be able to compile
with the new/correct parameter name ("version").  But this will only
happen for an incorrect MPI application (because the MPI-3 mandated
parameter name is "version", not "name"), so they deserve what they
get.

cmr=v1.8.1:reviewer=dgoodell

This commit was SVN r31365.
2014-04-10 14:45:36 +00:00
Jeff Squyres
6521dcc4f1 Trivial defensive programming/style update: use {}, even for 1-line blocks.
This commit was SVN r31361.
2014-04-09 16:28:31 +00:00
Nathan Hjelm
7aece0a7fd osc/sm: fix bugs in both the passive and active target paths
While testing one-sided on LANL systems I found a couple more OSC
bugs that were not caught during the initial testing:

 - In the passive target code we read the read lock count as a
   char instead of the intended uint32_t. This causes lock to
   lockup when using shared locks after 127 iterations.

 - The post code used the wrong group when trying to increment post
   counters. This causes a segmentation fault.

 - Both the post and wait code used the wrong check in the inner
   loop leading to an infinite loop.

cmr=v1.8.1:reviewer=jsquyres

This commit was SVN r31354.
2014-04-08 21:55:00 +00:00
Nathan Hjelm
a31bfbeb2c osc/rdma: fix typo in get accumulate path
There was a typo in the ompi_osc_gacc_long_start that was causing a
segmentation fault when executing long get accumulate operations.

cmr=v1.8.1:reviewer=jsquyres

This commit was SVN r31353.
2014-04-08 21:54:52 +00:00
Ryan Grant
ca0a7b1a9a Correct typo in r31332, mtl_portals_enpoint.h -> mtl_portals_endpoint.h
This commit was SVN r31338.

The following SVN revision numbers were found above:
  r31332 --> open-mpi/ompi@b12ee27b3d
2014-04-08 14:41:51 +00:00
Ralph Castain
b12ee27b3d Add missing files - thanks to Mr. Anonymous for reporting them as missing from the 1.8 tarball
cmr=v1.8.1:reviewer=jsquyres:subject=add missing portals4 files

This commit was SVN r31332.
2014-04-08 02:55:14 +00:00
Jeff Squyres
16f90acbaf btl usnic: Add some SHOW_HELP: tokens and remove 2 unused help messages
This commit was SVN r31322.
2014-04-07 15:40:19 +00:00
George Bosilca
95a4f219ea This commit fixes some of the Coverity reported warnings. I addressed
some of the collective modules, the shared memory and the profiling
interface. I left out VT, dynamic fcoll and seq rmaps.

cmr=v1.8.1:reviewer=jsquyres:subject=silence Coverity reported warnings

This commit was SVN r31309.
2014-04-06 18:23:49 +00:00
Nathan Hjelm
9112977d86 btl/openib/udcm: fix two race conditions
This commit fixes two nasty races:

 - One can occur if the connection request message and connection completion
   message arrive out of order. This can happen normally when adaptive routing
   is used and also in a timeout situation where a UD message is lost.

 - One occurs when handling an ack at the same time as we are handling the
   message timeout. In this case we can not free the message or the timeout
   will be operating on invalid data. This fix is a band-aid until I can come
   up with a better approach. Instead of freeing the message it is marked
   as inactive and the event callback is triggered immediately (this has no
   affect if the callback is already active). The callback then frees the
   message if it is inactive.

cmr=v1.8.1:reviewer=pasha

This commit was SVN r31305.
2014-04-02 15:09:50 +00:00
Oscar Vega-Gisbert
f133b6b693 Java: move buffer methods from Comm.c to MPI.c
This commit was SVN r31303.
2014-04-01 21:34:51 +00:00
Oscar Vega-Gisbert
3be348e34a Java: create an MCA parameter: ompi_mpi_java_eager
This commit was SVN r31301.
2014-04-01 20:57:12 +00:00
Oscar Vega-Gisbert
3ef08328e7 Java - Info, Datatype: it is recommended to use dup() instead of clone()
This commit was SVN r31300.
2014-03-31 21:30:23 +00:00
Oscar Vega-Gisbert
aa3e2f7afd New Java method in Comm: iDup
This commit was SVN r31299.
2014-03-31 21:15:26 +00:00
Oscar Vega-Gisbert
c9e6f09af1 Java: neighborhood collective communications
This commit was SVN r31296.
2014-03-30 19:55:56 +00:00
Oscar Vega-Gisbert
a1004bf5cd Java: remove useless native datatype initialization
This commit was SVN r31293.
2014-03-29 19:08:39 +00:00
Oscar Vega-Gisbert
0e177597c3 Java: Comm.set/getInfo
This commit was SVN r31289.
2014-03-29 18:57:51 +00:00
Oscar Vega-Gisbert
845818de59 Avoid use Status member in Request.
This commit was SVN r31287.
2014-03-29 18:13:38 +00:00
Nathan Hjelm
71bdb8c439 coll/ml: fix some warnings identified by clang
cmr=v1.8.1:reviewer=manjugv

This commit was SVN r31285.
2014-03-28 22:31:41 +00:00
Nathan Hjelm
fdf4c3b900 osc/rdma: really fix active message support
The last fix prevented a hang but had some cases where the results were
wrong. Fixed. Tested with armci, openmpi/ibm, openmpi/onesided.

cmr=v1.8:reviewer=jsquyres

This commit was SVN r31284.
2014-03-28 22:06:16 +00:00
Nathan Hjelm
6913a0f3cf osc/base: defensive programming. handle one more possible datatype case
It might be possible (don't know) for a datatype to made of a contiguous block
of a primitive datatype and have an lb. If this is ever the case the code
would have done the wrong thing. Add the lb in to be safe.

cmr=v1.8:reviewer=jsquyres

This commit was SVN r31283.
2014-03-28 22:06:05 +00:00
Nathan Hjelm
459431622b Revert "coll/ml: there is no reason not to enable coll/ml when a process in not"
Discussed this with Manju and we decided to back this one out until a later time.

This reverts commit r31188 and closes trac:4435

This commit was SVN r31282.

The following SVN revision numbers were found above:
  r31188 --> open-mpi/ompi@f1dd589092

The following Trac tickets were found above:
  Ticket 4435 --> https://svn.open-mpi.org/trac/ompi/ticket/4435
2014-03-28 21:16:34 +00:00