GitLab

Xin Zhao authored Apr 30, 2015 and Pavan Balaji committed May 30, 2015

This reverts commit 19f29078

.
Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored Apr 29, 2015 and Pavan Balaji committed May 30, 2015

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored May 18, 2015 and Pavan Balaji committed May 30, 2015

Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored May 18, 2015 and Pavan Balaji committed May 30, 2015

In RMA communication, when data to be sent is derived datatype but
contiguous, we can avoid copy/packing but directly issue it from
the user buffer. However, the data may have non-zero lb and we should
add the true lb when calculating the starting address. This patch
fixes this issue.
Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored May 28, 2015 and Pavan Balaji committed May 30, 2015

Originally the implementation of sendNoncontig() in Nemesis
assume that req->dev.segment_first is 0 and req->dev.segment_size
is the size of data, which potentially asssumes that the data passed
from upper layer must be the entire data, prohibiting the possibility
of streaming that data in the upper layer. The general way to use them
should be that req->dev.segment_first specifies the current starting
location and req->dev.segment_size specifies the current ending location.

We fixed this issue in commit 5132e070

, but
there are some places missed in that fixing. Here we fixed those places.
Signed-off-by: Pavan Balaji <balaji@anl.gov>

Xin Zhao authored May 28, 2015 and Pavan Balaji committed May 30, 2015

Note that here we move the definition of streaming unit size to
mpidimpl.h in CH3, so that Nemesis can see it and do assert check on it.
Signed-off-by: Pavan Balaji <balaji@anl.gov>

Removes unnecessary thread yielding in threaded nonblocking context id
allocation algorithm. The error was introduced by "copy-pasting" from
the blocking context id allocation algorithm
(MPIR_Get_contextid_sparse_group) when implementing the nonblocking
algorithm. Note the subtle difference on thread handling between the
two. In the blocking algorithm, yield is needed to allow another thread
to make progress. In nonblocking algorithm, there is no need to yield to
another thread because this thread will not block the progress. On the
contrary, unnecessary yield will allow other threads to execute and
insert wrong order of entries to the nonblocking schedule and cause
errors.

Fixes #2183
Signed-off-by: Junchao Zhang <jczhang@mcs.anl.gov>

Two tag bits are reserved for error propagation. We need to make sure
they are ignored by the network matching capabilities. Refs #2260

No reviewer.

Halim Amer authored May 26, 2015 and Kenneth Raffenetti committed May 27, 2015

If the destination of a LMT nonblocking send, which
follows a rendez-vous protocol, receives the RTS packet to the
unexpected queue followed by a cancelation request, the request in
the unexpected queue never gets freed. To solve the issue, we
forcefully free the rendez-vous request by adding an additional
request release operation. Refs #287
Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

Frees memory allocated to track large sends when those sends are
successfully cancelled.

No reviewer.

Junchao Zhang authored May 26, 2015 and Kenneth Raffenetti committed May 26, 2015

The configure test was dropped long time ago in MPICH, and to keep ABI compatible,
we don't want to restore it.

Refs #1877
Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

Change the bindings to make it as if MPI_BOTTOM in Fortran is
at address zero as it in C. See discussion on p.652 of MPI-3.0

Fixes #1877
Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

It was created but never destroyed.

Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

Simplifies request completion logic and ensures operations are properly
accounted for in the cancel send case. Removes pt2pt send tracking in the
VC, as requests are enough to guarantee completion.
Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

This is currently a workaround to avoid weird errors, e.g., stack
fault, occurred in the NBC I/O implementation on Linux. When the host
OS is Linux and aio-lite is not used, a blocking ADIO function is used
in the NBC I/O implementation.

Fixes #2201
Signed-off-by: Rob Latham <robl@mcs.anl.gov>

configure defines ROMIO_RUN_ON_LINUX as 1 when the host OS is Linux.
Signed-off-by: Rob Latham <robl@mcs.anl.gov>

Defines usleep when needed. This build error was discovered via the
random config tests in jenkins.
Signed-off-by: Rob Latham <robl@mcs.anl.gov>

Pavan Balaji authored May 13, 2015 and Kenneth Raffenetti committed May 13, 2015

Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

Pavan Balaji authored May 12, 2015 and Kenneth Raffenetti committed May 13, 2015

Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

we were not correctly handling the case where the error handler
MPI_ERRORS_ARE_FATAL was called.  credit Lisandro Dalcin with finidng
the issue.
Signed-off-by: Junchao Zhang <jczhang@mcs.anl.gov>

Lisandro Dalcin found we returned the wrong value from
MPI_File_call_errhandler.  it should return MPI_SUCCESS unless something
goes wrong with MPI_File_call_errhandler itself.

Closes #2259
Signed-off-by: Junchao Zhang <jczhang@mcs.anl.gov>

Signed-off-by: Jithin Jose <jithin.jose@intel.com>
Signed-off-by: Charles J Archer <charles.j.archer@intel.com>

Signed-off-by: Jithin Jose <jithin.jose@intel.com>
Signed-off-by: Charles J Archer <charles.j.archer@intel.com>

No reviewer

If the user passes the -static flag, we disable interlibrary
dependencies, since the linker would pick the static versions of the
libraries in that case.

We are still making an assumption that the default mode of the linker
is shared, and the user can specify static library builds with
-static.  However, this is not always true.  On BG/Q, for example, the
default is static.

Fixes #2190.
Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

This includes several changes:

1. Merged WRAPPER and EXTERNAL LIBS.  There is no reason to maintain
two names for these flags.  These are eventually appended and added to
the compiler wrappers anyway.

2. Updated mpicc and friend to add the necessary flags directly
instead of trying to merge these flags in configure.
Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

Min Si authored Apr 26, 2015 and Pavan Balaji committed Apr 28, 2015

On FreeBSD-10.1, function hstrerror is implemented in libc but its
declaration is disabled in netdb.h by default. Thus error "implicit
declaration of function hstrerrori" was reported when compiling hydra.
This patch adds two steps in hydra configure in order to transparently
fix this issue:
1. Check if function hstrerror exists (in libc).
2. Add the declaration of hstrerror if it exists in libc but the
declaration is disabled.
Signed-off-by: Pavan Balaji <balaji@anl.gov>

Min Si authored Apr 26, 2015 and Pavan Balaji committed Apr 28, 2015

Signed-off-by: Pavan Balaji <balaji@anl.gov>

MPID_nem_ofi_data_callback used to check sreq->cc in order to track progress of
the RTS/CTS/DATA protocol. The was an implicit assumption that fi_tsend with RTS
completes first. However this would cause a hang if fi_trecv completed earlier.
The fix is: don't rely on the cc but rather check the tag bits explicitly.
Note, the RTS/CTS/DATA bits are no longer accumulated (i.e., no more
"wc->tag | CTS/DATA").
Signed-off-by: Charles J Archer <charles.j.archer@intel.com>

It uses nemesis shared memory which is already cleaned up at this stage.
However, w/o any synchronization a hang in the close protocol is possible
since rts/cts/data messages may be on the fly. This change fixes the issue.
Signed-off-by: Charles J Archer <charles.j.archer@intel.com>

Signed-off-by: Charles J Archer <charles.j.archer@intel.com>

Signed-off-by: Charles J Archer <charles.j.archer@intel.com>

This patch modifies the OFI netmod to support large tag layouts, while preserving the old
tag layout.  OFI defines a 64 bit tag, but also provides for a 64 bit tag and immediate data.
In some OFI providers, we may want to select different tag layouts.  This patch currently
does not query for the proper tag layout or attempt to make a choice of the optimal layout,
it provides macro/templatized support for different tag formats.  Additional selection
criteria will be added in subsequent patches.

  * Tag layout is moved to a separate file.
    Added init_sendtag_M2, init_recvtag_M2 (M2 stands for MODE #2, i.e. the mode
    that uses fi_tsenddata and does not pack source into tag).

  * Created a template file for ofi_tagged.c
    Moved do_isend into template file which is included twice into ofi_tagged.c thus providing for the two
    versions of do_isend and do_isend_2 corresponding to the two API sets.

  * All send functions are available in two versions.
    Added macro that declares a function for the two API sets. The first set has the namings inherited from
    the previous netmod version. The functions of the second API set have the "_2" suffix.

  * Recv_posted, anysource_posted, recv_callback, ofi_probe  are templatized.

  * ofi_tag_to_vc renamed ofi_wc_to_vc
    Note, for the API_SET_2 the pgid is stored in the imm data while
    psource and port will be packed the same way as in API_SET_1.

  * Adds api_set member in gl_data struct.  Initialize routines based on api_set

  * Added RCD (RtsCtsData) protocol identifiers

  * Added support for OFI MEM_TAG_FORMAT

  * PGID placement modified
Signed-off-by: Charles J Archer <charles.j.archer@intel.com>

The atomic datatype check macro is meant to use the datatype itself
to create an error string. This removes the unused "name" argument.

Fixes #1906
Signed-off-by: Pavan Balaji <balaji@anl.gov>

We were originally returning "int" from these functions because the
idea was that they'd never read/write more than INT_MAX.  However, in
the case where the upper layer passes an IOV, where the first element
is smaller than INT_MAX, but the next element is larger than INT_MAX,
the total number of bytes sent might be larger than INT_MAX.

This was occassionally showing up as an error on Mac OS for large
message communication.
Signed-off-by: Antonio J. Pena <apenya@mcs.anl.gov>

Igor Ivanov authored Apr 23, 2015 and Pavan Balaji committed Apr 24, 2015

- Fixed hcoll broken by #98c76f78

 commit;
- Resolved warnings reported by hcoll supported code;
Signed-off-by: Igor Ivanov <Igor.Ivanov@itseez.com>
Signed-off-by: Pavan Balaji <balaji@anl.gov>

Igor Ivanov authored Apr 21, 2015 and Pavan Balaji committed Apr 24, 2015

Signed-off-by: Igor Ivanov <Igor.Ivanov@itseez.com>

Paul Coffman authored Apr 23, 2015 and Rob Latham committed Apr 24, 2015

Because the PAMI_Rput_typed / PAMI_Rget_typed are used with user defined
derived types that could be freed between the time of the MPID_Put / Get
and when the pami context actually executes the put or get (say in the
win_fence) add a ref count to the MPID_Datatype at the time of the
MPID_Put / Get and release the ref in the MPIDI_Win_DoneCB callback once
the get or put has actually executed and it is ok to free the datatype.
Signed-off-by: Rob Latham <robl@mcs.anl.gov>