Change Summary for MPICH2 for week of Aug 20, 2006

This list was automatically generated on Mon Aug 21 04:11:53 2006 .

Information on all weekly change summaries is available.

List of revisions

Files for each revision

*** empty log message *** src/include/mpidbg.h src/include/mpidu_func_nesting.h src/include/mpifunclog.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_finalize.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_impl.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_impl.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_impl.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_init.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_init.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_init.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_init.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_queue.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_send.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_send.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_utility.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_utility.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.h src/util/dbg/dbg_printf.c
actually made things work out ok. fixing the test resulted in segfaultsfurther down in the code. Since the broken (and apparently debugged) versionof the test is never true (and confusing to at least me), just remove it. src/mpi/romio/adio/ad_pvfs/ad_pvfs_write.c src/mpi/romio/adio/ad_pvfs2/ad_pvfs2_read.c src/mpi/romio/adio/ad_pvfs2/ad_pvfs2_write.c
Add missing initialization of channel-specific fields in request src/mpid/ch3/channels/shm/include/mpidi_ch3_pre.h src/mpid/ch3/channels/shm/src/shm.c src/mpid/ch3/channels/ssm/include/mpidi_ch3_pre.h src/mpid/ch3/include/mpidimpl.h src/mpid/ch3/src/ch3u_request.c
added a -subset flag to skip the tests we know won't work with pvfs2 src/mpi/romio/test/runtests.in
Added comment to suppress warning message from extractstates src/mpid/globus/mpidpost.h
Added cvsignore file src/mpid/ch3/channels/nemesis/nemesis/net_mod/sctp_module/.cvsignore
Added found/unexpected versions of Post_data_receive src/mpid/ch3/include/mpidimpl.h src/mpid/ch3/src/ch3u_handle_recv_pkt.c src/mpid/ch3/src/ch3u_handle_recv_req.c
Added initialization for otherwise common symbol src/mpid/ch3/channels/ssm/src/ch3_progress.c
Added missing targets for clean and distclean to ROMIO src/mpi/romio/Makefile.in
Added option for messages about datatype used in put test/mpi/rma/putfence1.c
added partial delivery support and fixed some related bugs. src/mpid/ch3/channels/nemesis/nemesis/net_mod/sctp_module/sctp_module_init.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/sctp_module/sctp_module_poll.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/sctp_module/sctp_module_send.c
Added step to remove CVS directories that may be added by the updatefiles step Makefile.sm
Added support for F90 MPI_SIZEOF configure.in src/binding/f90/Makefile.sm src/binding/f90/mpi.f90
Added tests for MPI_SIZEOF test/mpi/configure.in test/mpi/f90/Makefile.sm test/mpi/f90/testlist.in test/mpi/f90/misc/testlist
changed to use dynamic memory test/mpi/rma/transpose1.c test/mpi/rma/transpose5.c test/mpi/rma/transpose7.c
Current working version of the elan module src/mpid/ch3/channels/nemesis/nemesis/net_mod/elan_module/elan_module_impl.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/elan_module/elan_module_init.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/elan_module/elan_module_poll.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/elan_module/elan_module_poll.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/elan_module/elan_module_send.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/elan_module/elan_module_send.c
Fix No Errors output test/mpi/datatype/simple-pack-external.c test/mpi/datatype/triangular-pack.c test/mpi/pt2pt/waitany-null.c
Fix comment src/pm/mpd/mpdcleanup.py
Fix for recursive ops with only a few directories maint/simplemake.in
Fix minor typos in comments src/util/mem/handlemem.c
Fix typo in debug statement src/mpid/ch3/util/shm/ch3i_shm_bootstrapq.c
Fixed handling of comments when extracting states maint/extractstates.in
Fixed handling of sizeof for F90 types configure.in src/binding/f90/Makefile.sm
further down in the code. Since the broken (and apparently debugged) versionof the test is never true (and confusing to at least me), just remove it. src/mpi/romio/adio/ad_pvfs2/ad_pvfs2_read.c src/mpi/romio/adio/ad_pvfs2/ad_pvfs2_write.c
Improved error message from extract states maint/extractstates.in
nemesis: newtcp src/mpi/errhan/errnames.txt src/mpid/ch3/channels/nemesis/nemesis/include/mpid_nem_inline.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_init.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/newtcp_module_send.c src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.c
nemesis: newtcp module src/mpid/ch3/channels/nemesis/nemesis/net_mod/newtcp_module/socksm.c
nemesis: removed references to ca field of request from receive requests src/mpid/ch3/channels/nemesis/include/mpidi_ch3_pre.h src/mpid/ch3/channels/nemesis/nemesis/net_mod/tcp_module/tcp_module_init.c src/mpid/ch3/channels/nemesis/src/ch3_isendv.c src/mpid/ch3/channels/nemesis/src/ch3_progress.c
Now enqueue cell to process free Q in the case of failure.Created real mpi_errno's (internal error). src/mpid/ch3/channels/nemesis/nemesis/net_mod/sctp_module/sctp_module_send.c
null-out a pointer after freeing it so we do the right thing when re-enteringopen a 2nd time (as when ADIO_Open tests for EXCL) src/mpi/romio/adio/ad_pvfs2/ad_pvfs2_open.c
of the test is never true (and confusing to at least me), just remove it. src/mpi/romio/adio/ad_pvfs2/ad_pvfs2_write.c
Split Post_receive into found, unexpected, and zero cases src/mpid/ch3/channels/shm/src/shm.c src/mpid/ch3/channels/sshm/src/ch3_shm.c src/mpid/ch3/channels/ssm/src/ch3_shm.c src/mpid/ch3/src/ch3u_eager.c src/mpid/ch3/src/ch3u_eagersync.c src/mpid/ch3/src/ch3u_handle_recv_pkt.c src/mpid/ch3/src/ch3u_rndv.c
Switch to newer MPIU_DBG_MSG macros src/mpid/ch3/util/shm/ch3i_shm_bootstrapq.c
Switch to using an array to dispatch packet handlers (instead of a big case statement) and to using a completion function (onDataAvail) in the request instaed of a big case statement in handle_recv_req src/mpid/ch3/channels/nemesis/src/ch3_progress.c src/mpid/ch3/channels/shm/include/mpidi_ch3_impl.h src/mpid/ch3/channels/shm/src/ch3_istartmsg.c src/mpid/ch3/channels/shm/src/ch3_istartmsgv.c src/mpid/ch3/channels/shm/src/ch3_progress.c src/mpid/ch3/channels/shm/src/shm.c src/mpid/ch3/channels/sock/src/ch3_istartmsg.c src/mpid/ch3/channels/sock/src/ch3_istartmsgv.c src/mpid/ch3/channels/sock/src/ch3_progress.c src/mpid/ch3/channels/ssm/src/ch3_istartmsg.c src/mpid/ch3/channels/ssm/src/ch3_istartmsgv.c src/mpid/ch3/channels/ssm/src/ch3_progress.c src/mpid/ch3/channels/ssm/src/ch3_progress_sock.c src/mpid/ch3/channels/ssm/src/ch3_shm.c src/mpid/ch3/include/mpidimpl.h src/mpid/ch3/include/mpidpost.h src/mpid/ch3/include/mpidpre.h src/mpid/ch3/src/ch3u_eager.c src/mpid/ch3/src/ch3u_eagersync.c src/mpid/ch3/src/ch3u_handle_recv_pkt.c src/mpid/ch3/src/ch3u_handle_recv_req.c src/mpid/ch3/src/ch3u_handle_send_req.c src/mpid/ch3/src/ch3u_recvq.c src/mpid/ch3/src/ch3u_request.c src/mpid/ch3/src/ch3u_rndv.c src/mpid/ch3/src/mpid_irsend.c src/mpid/ch3/src/mpid_isend.c src/mpid/ch3/src/mpid_issend.c
Switched to newer debugging macros (MPIU_DBG_MSG) src/mpi/datatype/type_get_name.c src/mpid/common/datatype/mpid_type_commit.c src/mpid/common/datatype/mpid_type_contiguous.c src/mpid/common/datatype/mpid_type_create_resized.c src/mpid/common/datatype/mpid_type_dup.c src/mpid/common/datatype/mpid_type_vector.c
sync up with HEAD src/mpi/romio/configure.in src/mpi/romio/adio/ad_nfs/ad_nfs_resize.c src/mpi/romio/adio/common/ad_write_coll.c src/mpi/romio/adio/common/cb_config_list.c src/mpi/romio/adio/common/flatten.c src/mpi/romio/adio/include/adioi_error.h src/mpi/romio/mpi-io/mpioimpl.h src/mpi/romio/test/fcoll_test.f.in src/mpi/romio/test/fmisc.f.in src/mpi/romio/test/fperf.f.in src/mpi/romio/test/large_file.c.in src/mpi/romio/test/misc.c.in src/mpi/romio/test/pfcoll_test.f.in
Turns out the "a == a == b" test, while broken, was broken in a way thatactually made things work out ok. fixing the test resulted in segfaultsfurther down in the code. Since the broken (and apparently debugged) versionof the test is never true (and confusing to at least me), just remove it. src/mpi/romio/adio/ad_pvfs/ad_pvfs_read.c src/mpi/romio/adio/ad_pvfs/ad_pvfs_write.c src/mpi/romio/adio/ad_pvfs2/ad_pvfs2_read.c src/mpi/romio/adio/ad_pvfs2/ad_pvfs2_write.c
we are working with the flattend file type, not memory type here src/mpi/romio/adio/ad_pvfs2/ad_pvfs2_read.c src/mpi/romio/adio/ad_pvfs2/ad_pvfs2_write.c