No reviewer.

No reviewer.

No reviewer.

Fine-grained multithreading still breaks dynamic processes in a pretty big way,
but at least the test suite passes a bit more reliably with these in place.

No reviewer.

No reviewer.

This suppresses warnings about defined but not used.

No reviewer.

No reviewer.

This seems like it is a real bug, but there's no obvious fix right now.

No reviewer.

No reviewer.

No reviewer.

Racy accesses don't happen any longer, although I think this code now runs
slower in some cases.  Also, unrelated to this change, this algorithm probably
scales poorly with the number of contending threads.

No reviewer.

This commit begins to parallelize the progress engine slightly by making
it possible to signal request completion without acquiring the (bloated
and largely serial) MPIDCOMM critical section.

No reviewer.

No reviewer.

No reviewer.

When compiled for fine-grained threading, the completion counter serves
as a form of lockfree signalling.  As such, atomic access and memory
barriers must be used to ensure correctness.

In per-object mode, this code also contains valgrind client request annotations
to inform Helgrind/DRD/TSan about the lockfree signalling pattern.

No reviewer.

This code is still really, really nasty and shouldn't be doing what it
is doing.  The progress engine is indirectly called recursively, even
with a single thread.

No reviewer.

Reviewed by buntinas@.

No reviewer.

No reviewer.

No reviewer.

No reviewer.

I don't like this commit, but it's what we've got for the moment.
Sensible improvements are welcome.

No reviewer.

The LMT code should definitely be able to get better parallelism than this, but it
will require careful study to get it right.  The LMT critical section is just held
in a far too broad and clumsy way right now.  There are two major
problems:

1) the LMT code must call "up" to the main CH3 code in order to send and
   receive helper packets
2) MPIDCOMM and LMT both lock code regions, rather than data.  Together
   with (1) above, this makes it difficult to avoid deadlocks.

No reviewer.

Otherwise certain changes in the common object header cause surprising bugs.

No reviewer.

No reviewer.

A small helper function to handle the common-ish case of running the
progress engine until a particular request is complete.

No reviewer.

Valgrind output is much easier to understand when these functions aren't
inlined.

No reviewer.

[svn-r6903] Fixed some warnings, fixed compile failure with --enable-fast, fixed places where MPIR_Err_return_* was not being called.  No reviewer.

[svn-r6902] relaxed the check for autoconf version so that both 2.63 and 2.64 are allowed in updatefiles.

[svn-r6899] Making changes to windows configure script related to r6852. (1) All *F90* vars are now renamed as *FC* (2) The complete set of fort integer models is now represented by a new var - *ALL_INTEGER_MODELS and the older *INTEGER_MODEL var is now used similar to other dt MODELs

[svn-r6898] fix to r6897 so that top-level configure's enable-option-checking is followed in the subconfigure.

[svn-r6897] disable option checking in mpich2's configure of subpackage (no more warnings of unrecognize --enable or --with options) so it behaves the same as AC_CONFIG_SUBDIRS.

to connect back to the server, ask the user to check for
firewalls. Also, added some code to error out on the server side if
the proxy terminates badly.

bootstraps, pass a NULL stdin socket.