Open MPI logo

Open MPI Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Development mailing list

Subject: Re: [OMPI devel] [OMPI svn-full] svn:open-mpi r25093
From: George Bosilca (bosilca_at_[hidden])
Date: 2011-08-26 20:32:15


I see. This should be fixed in r25098. Thanks for your patience.

  george.

On Aug 26, 2011, at 19:47 , Ralph Castain wrote:

> Has nothing to do with version, George - it's a problem of ORTE_ENABLE_EPOCH not being included in an AC_DEFINE. It is solely defined via AM_CONDITIONAL, but then used in .h files - which is simply wrong.
>
> Please fix it.
>
>
> On Aug 26, 2011, at 5:41 PM, George Bosilca wrote:
>
>> We can't reproduce this. It compiles and runs without troubles on our macs. However, it might depend on the Mac OS X version, we recently moved to Lion.
>>
>> Thanks,
>> george.
>>
>> On Aug 26, 2011, at 19:19 , Ralph Castain wrote:
>>
>>> Hate to say this, but the trunk is broken - won't build on Mac with that disabled. I'll try to dig into it later :-(
>>>
>>>
>>> On Aug 26, 2011, at 4:18 PM, Wesley Bland wrote:
>>>
>>>> The epoch and resilient rote code is now macro'd away. To enable use
>>>>
>>>> --enable-resilient-orte
>>>>
>>>> which defines:
>>>>
>>>> ORTE_ENABLE_EPOCH
>>>> ORTE_RESIL_ORTE
>>>>
>>>> --
>>>>
>>>> Wesley
>>>>
>>>> On Aug 26, 2011, at 6:16 PM, wbland_at_[hidden] wrote:
>>>>
>>>>> Author: wbland
>>>>> Date: 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> New Revision: 25093
>>>>> URL: https://svn.open-mpi.org/trac/ompi/changeset/25093
>>>>>
>>>>> Log:
>>>>> By popular demand the epoch code is now disabled by default.
>>>>>
>>>>> To enable the epochs and the resilient orte code, use the configure flag:
>>>>>
>>>>> --enable-resilient-orte
>>>>>
>>>>> This will define both:
>>>>>
>>>>> ORTE_ENABLE_EPOCH
>>>>> ORTE_RESIL_ORTE
>>>>>
>>>>> Text files modified:
>>>>> trunk/ompi/mca/btl/openib/connect/btl_openib_connect_xoob.c | 12 ++++
>>>>> trunk/ompi/mca/coll/sm2/coll_sm2_module.c | 3
>>>>> trunk/ompi/mca/crcp/bkmrk/crcp_bkmrk_pml.c | 49 ++++++++----------
>>>>> trunk/ompi/mca/dpm/orte/dpm_orte.c | 2
>>>>> trunk/ompi/mca/pml/bfo/pml_bfo_failover.c | 10 +--
>>>>> trunk/ompi/mca/pml/bfo/pml_bfo_hdr.h | 6 --
>>>>> trunk/ompi/proc/proc.c | 6 +-
>>>>> trunk/opal/config/opal_configure_options.m4 | 8 +++
>>>>> trunk/orte/include/orte/types.h | 24 +++++++++
>>>>> trunk/orte/mca/db/daemon/db_daemon.c | 2
>>>>> trunk/orte/mca/errmgr/app/errmgr_app.c | 19 ++++++-
>>>>> trunk/orte/mca/errmgr/base/errmgr_base_fns.c | 12 ++--
>>>>> trunk/orte/mca/errmgr/base/errmgr_base_tool.c | 6 +-
>>>>> trunk/orte/mca/errmgr/hnp/errmgr_hnp.c | 99 +++++++++++++++++++++++++++------------
>>>>> trunk/orte/mca/errmgr/hnp/errmgr_hnp_autor.c | 6 +-
>>>>> trunk/orte/mca/errmgr/hnp/errmgr_hnp_crmig.c | 6 +-
>>>>> trunk/orte/mca/errmgr/orted/errmgr_orted.c | 71 +++++++++++++++++++++-------
>>>>> trunk/orte/mca/ess/alps/ess_alps_module.c | 4
>>>>> trunk/orte/mca/ess/base/base.h | 4 +
>>>>> trunk/orte/mca/ess/base/ess_base_select.c | 14 ++---
>>>>> trunk/orte/mca/ess/env/ess_env_module.c | 3
>>>>> trunk/orte/mca/ess/ess.h | 4 +
>>>>> trunk/orte/mca/ess/generic/ess_generic_module.c | 6 +-
>>>>> trunk/orte/mca/ess/hnp/ess_hnp_module.c | 2
>>>>> trunk/orte/mca/ess/lsf/ess_lsf_module.c | 3
>>>>> trunk/orte/mca/ess/singleton/ess_singleton_module.c | 2
>>>>> trunk/orte/mca/ess/slave/ess_slave_module.c | 3
>>>>> trunk/orte/mca/ess/slurm/ess_slurm_module.c | 3
>>>>> trunk/orte/mca/ess/slurmd/ess_slurmd_module.c | 4
>>>>> trunk/orte/mca/ess/tm/ess_tm_module.c | 2
>>>>> trunk/orte/mca/filem/rsh/filem_rsh_module.c | 6 +-
>>>>> trunk/orte/mca/grpcomm/base/grpcomm_base_coll.c | 21 ++-----
>>>>> trunk/orte/mca/grpcomm/hier/grpcomm_hier_module.c | 8 +-
>>>>> trunk/orte/mca/iof/base/base.h | 8 +-
>>>>> trunk/orte/mca/iof/base/iof_base_open.c | 2
>>>>> trunk/orte/mca/iof/hnp/iof_hnp.c | 7 +-
>>>>> trunk/orte/mca/iof/hnp/iof_hnp_receive.c | 6 +-
>>>>> trunk/orte/mca/iof/orted/iof_orted.c | 2
>>>>> trunk/orte/mca/odls/base/odls_base_default_fns.c | 7 +-
>>>>> trunk/orte/mca/odls/base/odls_base_open.c | 5 -
>>>>> trunk/orte/mca/odls/base/odls_base_state.c | 6 +-
>>>>> trunk/orte/mca/oob/tcp/oob_tcp_msg.c | 2
>>>>> trunk/orte/mca/oob/tcp/oob_tcp_peer.c | 5 ++
>>>>> trunk/orte/mca/plm/base/plm_base_jobid.c | 4
>>>>> trunk/orte/mca/plm/base/plm_base_launch_support.c | 3
>>>>> trunk/orte/mca/plm/base/plm_base_orted_cmds.c | 8 +--
>>>>> trunk/orte/mca/plm/base/plm_base_receive.c | 7 ++
>>>>> trunk/orte/mca/plm/base/plm_base_rsh_support.c | 4 +
>>>>> trunk/orte/mca/rmaps/base/rmaps_base_support_fns.c | 23 +++++----
>>>>> trunk/orte/mca/rmaps/rank_file/rmaps_rank_file.c | 3
>>>>> trunk/orte/mca/rmaps/seq/rmaps_seq.c | 3
>>>>> trunk/orte/mca/rmcast/base/rmcast_base_open.c | 6 +-
>>>>> trunk/orte/mca/rmcast/tcp/rmcast_tcp.c | 4
>>>>> trunk/orte/mca/rmcast/udp/rmcast_udp.c | 4
>>>>> trunk/orte/mca/rml/base/rml_base_components.c | 5 +
>>>>> trunk/orte/mca/rml/rml_types.h | 6 +
>>>>> trunk/orte/mca/routed/base/routed_base_components.c | 6 +-
>>>>> trunk/orte/mca/routed/base/routed_base_register_sync.c | 4 +
>>>>> trunk/orte/mca/routed/binomial/routed_binomial.c | 54 ++++++++++++---------
>>>>> trunk/orte/mca/routed/cm/routed_cm.c | 19 +++----
>>>>> trunk/orte/mca/routed/direct/routed_direct.c | 3
>>>>> trunk/orte/mca/routed/linear/routed_linear.c | 17 +++---
>>>>> trunk/orte/mca/routed/radix/routed_radix.c | 22 ++++----
>>>>> trunk/orte/mca/routed/slave/routed_slave.c | 6 +-
>>>>> trunk/orte/mca/sensor/file/sensor_file.c | 2
>>>>> trunk/orte/mca/snapc/base/snapc_base_fns.c | 4
>>>>> trunk/orte/mca/snapc/full/snapc_full_global.c | 12 ++--
>>>>> trunk/orte/mca/snapc/full/snapc_full_local.c | 6 +-
>>>>> trunk/orte/mca/snapc/full/snapc_full_module.c | 4
>>>>> trunk/orte/mca/sstore/base/sstore_base_fns.c | 6 +-
>>>>> trunk/orte/mca/sstore/central/sstore_central_global.c | 3
>>>>> trunk/orte/mca/sstore/central/sstore_central_local.c | 6 +-
>>>>> trunk/orte/mca/sstore/stage/sstore_stage_global.c | 7 +-
>>>>> trunk/orte/mca/sstore/stage/sstore_stage_local.c | 12 ++--
>>>>> trunk/orte/orted/orted_comm.c | 20 ++++----
>>>>> trunk/orte/orted/orted_main.c | 7 +-
>>>>> trunk/orte/runtime/data_type_support/orte_dt_compare_fns.c | 4 +
>>>>> trunk/orte/runtime/data_type_support/orte_dt_copy_fns.c | 4 +
>>>>> trunk/orte/runtime/data_type_support/orte_dt_packing_fns.c | 6 ++
>>>>> trunk/orte/runtime/data_type_support/orte_dt_print_fns.c | 19 +++++++
>>>>> trunk/orte/runtime/data_type_support/orte_dt_size_fns.c | 2
>>>>> trunk/orte/runtime/data_type_support/orte_dt_support.h | 11 ++++
>>>>> trunk/orte/runtime/data_type_support/orte_dt_unpacking_fns.c | 10 +++
>>>>> trunk/orte/runtime/orte_data_server.c | 2
>>>>> trunk/orte/runtime/orte_globals.c | 4 +
>>>>> trunk/orte/runtime/orte_init.c | 9 +++
>>>>> trunk/orte/runtime/orte_wait.h | 6 +-
>>>>> trunk/orte/test/system/oob_stress.c | 3
>>>>> trunk/orte/test/system/orte_ring.c | 6 -
>>>>> trunk/orte/test/system/orte_spawn.c | 4
>>>>> trunk/orte/tools/orte-ps/orte-ps.c | 10 +++
>>>>> trunk/orte/tools/orte-top/orte-top.c | 2
>>>>> trunk/orte/util/comm/comm.c | 7 ++
>>>>> trunk/orte/util/comm/comm.h | 5 +
>>>>> trunk/orte/util/hnp_contact.c | 3
>>>>> trunk/orte/util/name_fns.c | 47 ++++++++++++++----
>>>>> trunk/orte/util/name_fns.h | 30 ++++++++++++
>>>>> trunk/orte/util/nidmap.c | 13 ++++
>>>>> trunk/orte/util/nidmap.h | 11 ++++
>>>>> trunk/orte/util/proc_info.c | 14 ++++-
>>>>> trunk/test/util/orte_session_dir.c | 2
>>>>> 101 files changed, 652 insertions(+), 362 deletions(-)
>>>>>
>>>>> Modified: trunk/ompi/mca/btl/openib/connect/btl_openib_connect_xoob.c
>>>>> ==============================================================================
>>>>> --- trunk/ompi/mca/btl/openib/connect/btl_openib_connect_xoob.c (original)
>>>>> +++ trunk/ompi/mca/btl/openib/connect/btl_openib_connect_xoob.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -693,8 +693,16 @@
>>>>> bool found = false;
>>>>>
>>>>> BTL_VERBOSE(("Searching for ep and proc with follow parameters:"
>>>>> - "jobid %d, vpid %d, epoch %d, sid %" PRIx64 ", lid %d",
>>>>> - process_name->jobid, process_name->vpid, process_name->epoch, subnet_id, lid));
>>>>> + "jobid %d, vpid %d, "
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> + "epoch %d, "
>>>>> +#endif
>>>>> + "sid %" PRIx64 ", lid %d",
>>>>> + process_name->jobid, process_name->vpid,
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> + process_name->epoch,
>>>>> +#endif
>>>>> + subnet_id, lid));
>>>>> /* find ibproc */
>>>>> OPAL_THREAD_LOCK(&mca_btl_openib_component.ib_lock);
>>>>> for (ib_proc = (mca_btl_openib_proc_t*)
>>>>>
>>>>> Modified: trunk/ompi/mca/coll/sm2/coll_sm2_module.c
>>>>> ==============================================================================
>>>>> --- trunk/ompi/mca/coll/sm2/coll_sm2_module.c (original)
>>>>> +++ trunk/ompi/mca/coll/sm2/coll_sm2_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -1208,7 +1208,8 @@
>>>>> peer = OBJ_NEW(orte_namelist_t);
>>>>> peer->name.jobid = comm->c_local_group->grp_proc_pointers[i]->proc_name.jobid;
>>>>> peer->name.vpid = comm->c_local_group->grp_proc_pointers[i]->proc_name.vpid;
>>>>> - peer->name.epoch = comm->c_local_group->grp_proc_pointers[i]->proc_name.epoch;
>>>>> + ORTE_EPOCH_SET(peer->name.epoch,comm->c_local_group->grp_proc_pointers[i]->proc_name.epoch);
>>>>> +
>>>>> opal_list_append(&peers, &peer->item);
>>>>> }
>>>>> /* prepare send data */
>>>>>
>>>>> Modified: trunk/ompi/mca/crcp/bkmrk/crcp_bkmrk_pml.c
>>>>> ==============================================================================
>>>>> --- trunk/ompi/mca/crcp/bkmrk/crcp_bkmrk_pml.c (original)
>>>>> +++ trunk/ompi/mca/crcp/bkmrk/crcp_bkmrk_pml.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -702,7 +702,7 @@
>>>>> void ompi_crcp_bkmrk_pml_peer_ref_construct(ompi_crcp_bkmrk_pml_peer_ref_t *peer_ref) {
>>>>> peer_ref->proc_name.jobid = ORTE_JOBID_INVALID;
>>>>> peer_ref->proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - peer_ref->proc_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(peer_ref->proc_name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> OBJ_CONSTRUCT(&peer_ref->send_list, opal_list_t);
>>>>> OBJ_CONSTRUCT(&peer_ref->isend_list, opal_list_t);
>>>>> @@ -730,7 +730,7 @@
>>>>>
>>>>> peer_ref->proc_name.jobid = ORTE_JOBID_INVALID;
>>>>> peer_ref->proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - peer_ref->proc_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(peer_ref->proc_name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> while( NULL != (item = opal_list_remove_first(&peer_ref->send_list)) ) {
>>>>> HOKE_TRAFFIC_MSG_REF_RETURN(item);
>>>>> @@ -840,7 +840,7 @@
>>>>>
>>>>> msg_ref->proc_name.jobid = ORTE_JOBID_INVALID;
>>>>> msg_ref->proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - msg_ref->proc_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(msg_ref->proc_name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> msg_ref->matched = INVALID_INT;
>>>>> msg_ref->done = INVALID_INT;
>>>>> @@ -868,7 +868,7 @@
>>>>>
>>>>> msg_ref->proc_name.jobid = ORTE_JOBID_INVALID;
>>>>> msg_ref->proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - msg_ref->proc_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(msg_ref->proc_name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> msg_ref->matched = INVALID_INT;
>>>>> msg_ref->done = INVALID_INT;
>>>>> @@ -902,7 +902,7 @@
>>>>>
>>>>> msg_ref->proc_name.jobid = ORTE_JOBID_INVALID;
>>>>> msg_ref->proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - msg_ref->proc_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(msg_ref->proc_name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> msg_ref->done = INVALID_INT;
>>>>> msg_ref->active = INVALID_INT;
>>>>> @@ -934,7 +934,7 @@
>>>>>
>>>>> msg_ref->proc_name.jobid = ORTE_JOBID_INVALID;
>>>>> msg_ref->proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - msg_ref->proc_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(msg_ref->proc_name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> msg_ref->done = INVALID_INT;
>>>>> msg_ref->active = INVALID_INT;
>>>>> @@ -954,7 +954,7 @@
>>>>>
>>>>> msg_ack_ref->peer.jobid = ORTE_JOBID_INVALID;
>>>>> msg_ack_ref->peer.vpid = ORTE_VPID_INVALID;
>>>>> - msg_ack_ref->peer.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(msg_ack_ref->peer.epoch,ORTE_EPOCH_MIN);
>>>>> }
>>>>>
>>>>> void ompi_crcp_bkmrk_pml_drain_message_ack_ref_destruct( ompi_crcp_bkmrk_pml_drain_message_ack_ref_t *msg_ack_ref) {
>>>>> @@ -962,7 +962,7 @@
>>>>>
>>>>> msg_ack_ref->peer.jobid = ORTE_JOBID_INVALID;
>>>>> msg_ack_ref->peer.vpid = ORTE_VPID_INVALID;
>>>>> - msg_ack_ref->peer.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(msg_ack_ref->peer.epoch,ORTE_EPOCH_MIN);
>>>>> }
>>>>>
>>>>>
>>>>> @@ -1015,7 +1015,7 @@
>>>>> }
>>>>>
>>>>>
>>>>> -#define CREATE_NEW_MSG(msg_ref, v_type, v_count, v_ddt_size, v_tag, v_rank, v_comm, p_jobid, p_vpid, p_epoch) \
>>>>> +#define CREATE_NEW_MSG(msg_ref, v_type, v_count, v_ddt_size, v_tag, v_rank, v_comm, p_jobid, p_vpid) \
>>>>> { \
>>>>> HOKE_TRAFFIC_MSG_REF_ALLOC(msg_ref, ret); \
>>>>> \
>>>>> @@ -1034,7 +1034,7 @@
>>>>> \
>>>>> msg_ref->proc_name.jobid = p_jobid; \
>>>>> msg_ref->proc_name.vpid = p_vpid; \
>>>>> - msg_ref->proc_name.epoch = p_epoch; \
>>>>> + ORTE_EPOCH_SET(msg_ref->proc_name.epoch,orte_ess.proc_get_epoch(&(msg_ref->proc_name))); \
>>>>> \
>>>>> msg_ref->matched = 0; \
>>>>> msg_ref->done = 0; \
>>>>> @@ -1043,7 +1043,7 @@
>>>>> msg_ref->active_drain = 0; \
>>>>> }
>>>>>
>>>>> -#define CREATE_NEW_DRAIN_MSG(msg_ref, v_type, v_count, v_ddt_size, v_tag, v_rank, v_comm, p_jobid, p_vpid, p_epoch) \
>>>>> +#define CREATE_NEW_DRAIN_MSG(msg_ref, v_type, v_count, v_ddt_size, v_tag, v_rank, v_comm, p_jobid, p_vpid) \
>>>>> { \
>>>>> HOKE_DRAIN_MSG_REF_ALLOC(msg_ref, ret); \
>>>>> \
>>>>> @@ -1063,7 +1063,7 @@
>>>>> \
>>>>> msg_ref->proc_name.jobid = p_jobid; \
>>>>> msg_ref->proc_name.vpid = p_vpid; \
>>>>> - msg_ref->proc_name.epoch = p_epoch; \
>>>>> + ORTE_EPOCH_SET(msg_ref->proc_name.epoch,orte_ess.proc_get_epoch(&(msg_ref->proc_name))); \
>>>>> }
>>>>>
>>>>>
>>>>> @@ -1466,7 +1466,7 @@
>>>>>
>>>>> new_peer_ref->proc_name.jobid = procs[i]->proc_name.jobid;
>>>>> new_peer_ref->proc_name.vpid = procs[i]->proc_name.vpid;
>>>>> - new_peer_ref->proc_name.epoch = procs[i]->proc_name.epoch;
>>>>> + ORTE_EPOCH_SET(new_peer_ref->proc_name.epoch,procs[i]->proc_name.epoch);
>>>>>
>>>>> opal_list_append(&ompi_crcp_bkmrk_pml_peer_refs, &(new_peer_ref->super));
>>>>> }
>>>>> @@ -3237,13 +3237,11 @@
>>>>> CREATE_NEW_MSG((*msg_ref), msg_type,
>>>>> count, ddt_size, tag, dest, comm,
>>>>> peer_ref->proc_name.jobid,
>>>>> - peer_ref->proc_name.vpid,
>>>>> - peer_ref->proc_name.epoch);
>>>>> + peer_ref->proc_name.vpid);
>>>>> } else {
>>>>> CREATE_NEW_MSG((*msg_ref), msg_type,
>>>>> count, ddt_size, tag, dest, comm,
>>>>> - ORTE_JOBID_INVALID, ORTE_VPID_INVALID,
>>>>> - ORTE_EPOCH_INVALID);
>>>>> + ORTE_JOBID_INVALID, ORTE_VPID_INVALID);
>>>>> }
>>>>>
>>>>> if( msg_type == COORD_MSG_TYPE_P_SEND ||
>>>>> @@ -3377,7 +3375,7 @@
>>>>> if( NULL == from_peer_ref && NULL != to_peer_ref ) {
>>>>> (*new_msg_ref)->proc_name.jobid = to_peer_ref->proc_name.jobid;
>>>>> (*new_msg_ref)->proc_name.vpid = to_peer_ref->proc_name.vpid;
>>>>> - (*new_msg_ref)->proc_name.epoch = to_peer_ref->proc_name.epoch;
>>>>> + ORTE_EPOCH_SET((*new_msg_ref)->proc_name.epoch,to_peer_ref->proc_name.epoch);
>>>>> }
>>>>>
>>>>> return exit_status;
>>>>> @@ -3808,8 +3806,7 @@
>>>>> CREATE_NEW_DRAIN_MSG((*msg_ref), msg_type,
>>>>> count, NULL, tag, dest, comm,
>>>>> peer_ref->proc_name.jobid,
>>>>> - peer_ref->proc_name.vpid,
>>>>> - peer_ref->proc_name.epoch);
>>>>> + peer_ref->proc_name.vpid);
>>>>>
>>>>> (*msg_ref)->done = 0;
>>>>> (*msg_ref)->active = 0;
>>>>> @@ -5284,8 +5281,7 @@
>>>>> */
>>>>> peer_name.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> peer_name.vpid = peer_idx;
>>>>> - peer_name.epoch = ORTE_EPOCH_INVALID;
>>>>> - peer_name.epoch = orte_ess.proc_get_epoch(&peer_name);
>>>>> + ORTE_EPOCH_SET(peer_name.epoch,orte_ess.proc_get_epoch(&peer_name));
>>>>>
>>>>> if( NULL == (peer_ref = find_peer(peer_name))) {
>>>>> opal_output(mca_crcp_bkmrk_component.super.output_handle,
>>>>> @@ -5346,8 +5342,7 @@
>>>>>
>>>>> peer_name.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> peer_name.vpid = peer_idx;
>>>>> - peer_name.epoch = ORTE_EPOCH_INVALID;
>>>>> - peer_name.epoch = orte_ess.proc_get_epoch(&peer_name);
>>>>> + ORTE_EPOCH_SET(peer_name.epoch,orte_ess.proc_get_epoch(&peer_name));
>>>>>
>>>>> if ( 0 > (ret = orte_rml.recv_buffer_nb(&peer_name,
>>>>> OMPI_CRCP_COORD_BOOKMARK_TAG,
>>>>> @@ -5529,7 +5524,8 @@
>>>>> HOKE_DRAIN_ACK_MSG_REF_ALLOC(d_msg_ack, ret);
>>>>> d_msg_ack->peer.jobid = peer_ref->proc_name.jobid;
>>>>> d_msg_ack->peer.vpid = peer_ref->proc_name.vpid;
>>>>> - d_msg_ack->peer.epoch = peer_ref->proc_name.epoch;
>>>>> + ORTE_EPOCH_SET(d_msg_ack->peer.epoch,peer_ref->proc_name.epoch);
>>>>> +
>>>>> d_msg_ack->complete = false;
>>>>> opal_list_append(&drained_msg_ack_list, &(d_msg_ack->super));
>>>>> OPAL_OUTPUT_VERBOSE((10, mca_crcp_bkmrk_component.super.output_handle,
>>>>> @@ -6169,8 +6165,7 @@
>>>>> count, datatype_size, tag, rank,
>>>>> ompi_comm_lookup(comm_id),
>>>>> peer_ref->proc_name.jobid,
>>>>> - peer_ref->proc_name.vpid,
>>>>> - peer_ref->proc_name.epoch);
>>>>> + peer_ref->proc_name.vpid);
>>>>>
>>>>> traffic_message_create_drain_message(true, num_left_unresolved,
>>>>> peer_ref,
>>>>>
>>>>> Modified: trunk/ompi/mca/dpm/orte/dpm_orte.c
>>>>> ==============================================================================
>>>>> --- trunk/ompi/mca/dpm/orte/dpm_orte.c (original)
>>>>> +++ trunk/ompi/mca/dpm/orte/dpm_orte.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -1130,7 +1130,7 @@
>>>>> /* flag the identity of the remote proc */
>>>>> carport.jobid = mev->sender.jobid;
>>>>> carport.vpid = mev->sender.vpid;
>>>>> - carport.epoch = mev->sender.epoch;
>>>>> + ORTE_EPOCH_SET(carport.epoch,mev->sender.epoch);
>>>>>
>>>>> /* release the event */
>>>>> OBJ_RELEASE(mev);
>>>>>
>>>>> Modified: trunk/ompi/mca/pml/bfo/pml_bfo_failover.c
>>>>> ==============================================================================
>>>>> --- trunk/ompi/mca/pml/bfo/pml_bfo_failover.c (original)
>>>>> +++ trunk/ompi/mca/pml/bfo/pml_bfo_failover.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -1,8 +1,5 @@
>>>>> /*
>>>>> * Copyright (c) 2010 Oracle and/or its affiliates. All rights reserved.
>>>>> - * Copyright (c) 2004-2011 The University of Tennessee and The University
>>>>> - * of Tennessee Research Foundation. All rights
>>>>> - * reserved.
>>>>> * $COPYRIGHT$
>>>>> *
>>>>> * Additional copyrights may follow
>>>>> @@ -398,13 +395,13 @@
>>>>> (hdr->hdr_match.hdr_seq != (uint16_t)recvreq->req_msgseq)) {
>>>>> orte_proc.jobid = hdr->hdr_restart.hdr_jobid;
>>>>> orte_proc.vpid = hdr->hdr_restart.hdr_vpid;
>>>>> - orte_proc.epoch = hdr->hdr_restart.hdr_epoch;
>>>>> +
>>>>> ompi_proc = ompi_proc_find(&orte_proc);
>>>>> opal_output_verbose(20, mca_pml_bfo_output,
>>>>> "RNDVRESTARTNOTIFY: received: does not match request, sending NACK back "
>>>>> "PML:req=%d,hdr=%d CTX:req=%d,hdr=%d SRC:req=%d,hdr=%d "
>>>>> "RQS:req=%d,hdr=%d src_req=%p, dst_req=%p, peer=%d, hdr->hdr_jobid=%d, "
>>>>> - "hdr->hdr_vpid=%d, hdr->hdr_epoch=%d, ompi_proc->proc_hostname=%s",
>>>>> + "hdr->hdr_vpid=%d, ompi_proc->proc_hostname=%s",
>>>>> (uint16_t)recvreq->req_msgseq, hdr->hdr_match.hdr_seq,
>>>>> recvreq->req_recv.req_base.req_comm->c_contextid, hdr->hdr_match.hdr_ctx,
>>>>> recvreq->req_recv.req_base.req_ompi.req_status.MPI_SOURCE,
>>>>> @@ -413,7 +410,7 @@
>>>>> recvreq->remote_req_send.pval, (void *)recvreq,
>>>>> recvreq->req_recv.req_base.req_ompi.req_status.MPI_SOURCE,
>>>>> hdr->hdr_restart.hdr_jobid, hdr->hdr_restart.hdr_vpid,
>>>>> - hdr->hdr_restart.hdr_epoch, ompi_proc->proc_hostname);
>>>>> + ompi_proc->proc_hostname);
>>>>> mca_pml_bfo_recv_request_rndvrestartnack(des, ompi_proc, false);
>>>>> return;
>>>>> }
>>>>> @@ -715,7 +712,6 @@
>>>>> restart->hdr_dst_rank = sendreq->req_send.req_base.req_peer; /* Needed for NACKs */
>>>>> restart->hdr_jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> restart->hdr_vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - restart->hdr_epoch = ORTE_PROC_MY_NAME->epoch;
>>>>>
>>>>> bfo_hdr_hton(restart, MCA_PML_BFO_HDR_TYPE_RNDVRESTARTNOTIFY, proc);
>>>>>
>>>>>
>>>>> Modified: trunk/ompi/mca/pml/bfo/pml_bfo_hdr.h
>>>>> ==============================================================================
>>>>> --- trunk/ompi/mca/pml/bfo/pml_bfo_hdr.h (original)
>>>>> +++ trunk/ompi/mca/pml/bfo/pml_bfo_hdr.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -2,9 +2,6 @@
>>>>> * Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
>>>>> * University Research and Technology
>>>>> * Corporation. All rights reserved.
>>>>> - * Copyright (c) 2004-2011 The University of Tennessee and The University
>>>>> - * of Tennessee Research Foundation. All rights
>>>>> - * reserved.
>>>>> * Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
>>>>> * University of Stuttgart. All rights reserved.
>>>>> * Copyright (c) 2004-2005 The Regents of the University of California.
>>>>> @@ -415,7 +412,6 @@
>>>>> int32_t hdr_dst_rank; /**< needed to send NACK */
>>>>> uint32_t hdr_jobid; /**< needed to send NACK */
>>>>> uint32_t hdr_vpid; /**< needed to send NACK */
>>>>> - uint32_t hdr_epoch; /**< needed to send NACK */
>>>>> };
>>>>> typedef struct mca_pml_bfo_restart_hdr_t mca_pml_bfo_restart_hdr_t;
>>>>>
>>>>> @@ -428,7 +424,6 @@
>>>>> (h).hdr_dst_rank = ntohl((h).hdr_dst_rank); \
>>>>> (h).hdr_jobid = ntohl((h).hdr_jobid); \
>>>>> (h).hdr_vpid = ntohl((h).hdr_vpid); \
>>>>> - (h).hdr_epoch = ntohl((h).hdr_epoch); \
>>>>> } while (0)
>>>>>
>>>>> #define MCA_PML_BFO_RESTART_HDR_HTON(h) \
>>>>> @@ -437,7 +432,6 @@
>>>>> (h).hdr_dst_rank = htonl((h).hdr_dst_rank); \
>>>>> (h).hdr_jobid = htonl((h).hdr_jobid); \
>>>>> (h).hdr_vpid = htonl((h).hdr_vpid); \
>>>>> - (h).hdr_epoch = htonl((h).hdr_epoch); \
>>>>> } while (0)
>>>>>
>>>>> #endif /* PML_BFO */
>>>>>
>>>>> Modified: trunk/ompi/proc/proc.c
>>>>> ==============================================================================
>>>>> --- trunk/ompi/proc/proc.c (original)
>>>>> +++ trunk/ompi/proc/proc.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -108,7 +108,8 @@
>>>>>
>>>>> proc->proc_name.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> proc->proc_name.vpid = i;
>>>>> - proc->proc_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(proc->proc_name.epoch,ORTE_EPOCH_MIN);
>>>>> +
>>>>> if (i == ORTE_PROC_MY_NAME->vpid) {
>>>>> ompi_proc_local_proc = proc;
>>>>> proc->proc_flags = OPAL_PROC_ALL_LOCAL;
>>>>> @@ -362,8 +363,7 @@
>>>>>
>>>>> /* Does not change: proc->proc_name.vpid */
>>>>> proc->proc_name.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> - proc->proc_name.epoch = ORTE_EPOCH_INVALID;
>>>>> - proc->proc_name.epoch = orte_ess.proc_get_epoch(&proc->proc_name);
>>>>> + ORTE_EPOCH_SET(proc->proc_name.epoch,orte_ess.proc_get_epoch(&proc->proc_name));
>>>>>
>>>>> /* Make sure to clear the local flag before we set it below */
>>>>> proc->proc_flags = 0;
>>>>>
>>>>> Modified: trunk/opal/config/opal_configure_options.m4
>>>>> ==============================================================================
>>>>> --- trunk/opal/config/opal_configure_options.m4 (original)
>>>>> +++ trunk/opal/config/opal_configure_options.m4 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -416,6 +416,14 @@
>>>>> AM_CONDITIONAL(WANT_FT_CR, test "$opal_want_ft_cr" = "1")
>>>>>
>>>>> #
>>>>> +# Compile in resilient runtime code
>>>>> +#
>>>>> +AC_ARG_ENABLE(resilient-orte,
>>>>> + [AC_HELP_STRING([--enable-resilient-orte], [Enable the resilient runtime code.])])
>>>>> +AM_CONDITIONAL(ORTE_RESIL_ORTE, [test "$enable_resilient_orte" = "yes"])
>>>>> +AM_CONDITIONAL(ORTE_ENABLE_EPOCH, [test "$enable_resilient_orte" = "yes"])
>>>>> +
>>>>> +#
>>>>> # Do we want to install binaries?
>>>>> #
>>>>> AC_ARG_ENABLE([binaries],
>>>>>
>>>>> Modified: trunk/orte/include/orte/types.h
>>>>> ==============================================================================
>>>>> --- trunk/orte/include/orte/types.h (original)
>>>>> +++ trunk/orte/include/orte/types.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -81,24 +81,43 @@
>>>>> #define ORTE_VPID_T OPAL_UINT32
>>>>> #define ORTE_VPID_MAX UINT32_MAX-2
>>>>> #define ORTE_VPID_MIN 0
>>>>> +
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> typedef uint32_t orte_epoch_t;
>>>>> #define ORTE_EPOCH_T OPAL_UINT32
>>>>> #define ORTE_EPOCH_MAX UINT32_MAX-2
>>>>> #define ORTE_EPOCH_MIN 0
>>>>> +#endif
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> #define ORTE_PROCESS_NAME_HTON(n) \
>>>>> do { \
>>>>> n.jobid = htonl(n.jobid); \
>>>>> n.vpid = htonl(n.vpid); \
>>>>> n.epoch = htonl(n.epoch); \
>>>>> } while (0)
>>>>> +#else
>>>>> +#define ORTE_PROCESS_NAME_HTON(n) \
>>>>> +do { \
>>>>> + n.jobid = htonl(n.jobid); \
>>>>> + n.vpid = htonl(n.vpid); \
>>>>> +} while (0)
>>>>> +#endif
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> #define ORTE_PROCESS_NAME_NTOH(n) \
>>>>> do { \
>>>>> n.jobid = ntohl(n.jobid); \
>>>>> n.vpid = ntohl(n.vpid); \
>>>>> n.epoch = ntohl(n.epoch); \
>>>>> } while (0)
>>>>> +#else
>>>>> +#define ORTE_PROCESS_NAME_NTOH(n) \
>>>>> +do { \
>>>>> + n.jobid = ntohl(n.jobid); \
>>>>> + n.vpid = ntohl(n.vpid); \
>>>>> +} while (0)
>>>>> +#endif
>>>>>
>>>>> #define ORTE_NAME_ARGS(n) \
>>>>> (unsigned long) ((NULL == n) ? (unsigned long)ORTE_JOBID_INVALID : (unsigned long)(n)->jobid), \
>>>>> @@ -127,6 +146,7 @@
>>>>> struct orte_process_name_t {
>>>>> orte_jobid_t jobid; /**< Job number */
>>>>> orte_vpid_t vpid; /**< Process id - equivalent to rank */
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> orte_epoch_t epoch; /**< Epoch - used to measure the generation of a recovered process.
>>>>> * The epoch will start at ORTE_EPOCH_MIN and
>>>>> * increment every time the process is detected as
>>>>> @@ -135,6 +155,7 @@
>>>>> * processes that did not directly detect the
>>>>> * failure to increment their epochs.
>>>>> */
>>>>> +#endif
>>>>> };
>>>>> typedef struct orte_process_name_t orte_process_name_t;
>>>>>
>>>>> @@ -157,7 +178,10 @@
>>>>> #define ORTE_NAME (OPAL_DSS_ID_DYNAMIC + 2) /**< an orte_process_name_t */
>>>>> #define ORTE_VPID (OPAL_DSS_ID_DYNAMIC + 3) /**< a vpid */
>>>>> #define ORTE_JOBID (OPAL_DSS_ID_DYNAMIC + 4) /**< a jobid */
>>>>> +
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> #define ORTE_EPOCH (OPAL_DSS_ID_DYNAMIC + 5) /**< an epoch */
>>>>> +#endif
>>>>>
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>> /* State-related types */
>>>>>
>>>>> Modified: trunk/orte/mca/db/daemon/db_daemon.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/db/daemon/db_daemon.c (original)
>>>>> +++ trunk/orte/mca/db/daemon/db_daemon.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -386,7 +386,7 @@
>>>>> dat = OBJ_NEW(orte_db_data_t);
>>>>> dat->name.jobid = sender->jobid;
>>>>> dat->name.vpid = sender->vpid;
>>>>> - dat->name.epoch= sender->epoch;
>>>>> + ORTE_EPOCH_SET(dat->name.epoch,sender->epoch);
>>>>> dat->key = key;
>>>>> count=1;
>>>>> opal_dss.unpack(buf, &dat->size, &count, OPAL_INT32);
>>>>>
>>>>> Modified: trunk/orte/mca/errmgr/app/errmgr_app.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/errmgr/app/errmgr_app.c (original)
>>>>> +++ trunk/orte/mca/errmgr/app/errmgr_app.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -82,8 +82,10 @@
>>>>> NULL,
>>>>> NULL,
>>>>> NULL,
>>>>> - orte_errmgr_base_register_migration_warning,
>>>>> - orte_errmgr_base_set_fault_callback
>>>>> + orte_errmgr_base_register_migration_warning
>>>>> +#if ORTE_RESIL_ORTE
>>>>> + ,orte_errmgr_base_set_fault_callback
>>>>> +#endif
>>>>> };
>>>>>
>>>>> /************************
>>>>> @@ -93,18 +95,23 @@
>>>>> {
>>>>> int ret = ORTE_SUCCESS;
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> ret = orte_rml.recv_buffer_nb(ORTE_NAME_WILDCARD,
>>>>> ORTE_RML_TAG_EPOCH_CHANGE,
>>>>> ORTE_RML_PERSISTENT,
>>>>> epoch_change_recv,
>>>>> NULL);
>>>>> +#endif
>>>>> +
>>>>> return ret;
>>>>> }
>>>>>
>>>>> static int finalize(void)
>>>>> {
>>>>> +#if ORTE_RESIL_ORTE
>>>>> orte_rml.recv_cancel(ORTE_NAME_WILDCARD,
>>>>> ORTE_RML_TAG_EPOCH_CHANGE);
>>>>> +#endif
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> @@ -151,6 +158,7 @@
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> void epoch_change_recv(int status,
>>>>> orte_process_name_t *sender,
>>>>> opal_buffer_t *buffer,
>>>>> @@ -209,15 +217,20 @@
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME)));
>>>>>
>>>>> (*fault_cbfunc)(procs);
>>>>> + } else if (NULL == fault_cbfunc) {
>>>>> + OPAL_OUTPUT_VERBOSE((1, orte_errmgr_base.output,
>>>>> + "%s errmgr:app Calling fault callback failed (NULL pointer)!",
>>>>> + ORTE_NAME_PRINT(ORTE_PROC_MY_NAME)));
>>>>> } else {
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_errmgr_base.output,
>>>>> - "%s errmgr:app Calling fault callback failed!",
>>>>> + "%s errmgr:app Calling fault callback failed (num_dead <= 0)!",
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME)));
>>>>> }
>>>>>
>>>>> free(proc);
>>>>> OBJ_RELEASE(procs);
>>>>> }
>>>>> +#endif
>>>>>
>>>>> static int orte_errmgr_app_abort_peers(orte_process_name_t *procs, orte_std_cntr_t num_procs)
>>>>> {
>>>>>
>>>>> Modified: trunk/orte/mca/errmgr/base/errmgr_base_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/errmgr/base/errmgr_base_fns.c (original)
>>>>> +++ trunk/orte/mca/errmgr/base/errmgr_base_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -97,13 +97,13 @@
>>>>> {
>>>>> item->proc_name.vpid = ORTE_VPID_INVALID;
>>>>> item->proc_name.jobid = ORTE_JOBID_INVALID;
>>>>> - item->proc_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(item->proc_name.epoch,ORTE_EPOCH_MIN);
>>>>> }
>>>>>
>>>>> void orte_errmgr_predicted_proc_destruct( orte_errmgr_predicted_proc_t *item)
>>>>> {
>>>>> item->proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - item->proc_name.epoch = ORTE_EPOCH_INVALID;
>>>>> + ORTE_EPOCH_SET(item->proc_name.epoch,ORTE_EPOCH_INVALID);
>>>>> item->proc_name.jobid = ORTE_JOBID_INVALID;
>>>>> }
>>>>>
>>>>> @@ -139,13 +139,13 @@
>>>>> void orte_errmgr_predicted_map_construct(orte_errmgr_predicted_map_t *item)
>>>>> {
>>>>> item->proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - item->proc_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(item->proc_name.epoch,ORTE_EPOCH_MIN);
>>>>> item->proc_name.jobid = ORTE_JOBID_INVALID;
>>>>>
>>>>> item->node_name = NULL;
>>>>>
>>>>> item->map_proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - item->map_proc_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(item->map_proc_name.epoch,ORTE_EPOCH_MIN);
>>>>> item->map_proc_name.jobid = ORTE_JOBID_INVALID;
>>>>>
>>>>> item->map_node_name = NULL;
>>>>> @@ -156,7 +156,7 @@
>>>>> void orte_errmgr_predicted_map_destruct( orte_errmgr_predicted_map_t *item)
>>>>> {
>>>>> item->proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - item->proc_name.epoch = ORTE_EPOCH_INVALID;
>>>>> + ORTE_EPOCH_SET(item->proc_name.epoch,ORTE_EPOCH_INVALID);
>>>>> item->proc_name.jobid = ORTE_JOBID_INVALID;
>>>>>
>>>>> if( NULL != item->node_name ) {
>>>>> @@ -165,7 +165,7 @@
>>>>> }
>>>>>
>>>>> item->map_proc_name.vpid = ORTE_VPID_INVALID;
>>>>> - item->map_proc_name.epoch = ORTE_EPOCH_INVALID;
>>>>> + ORTE_EPOCH_SET(item->map_proc_name.epoch,ORTE_EPOCH_INVALID);
>>>>> item->map_proc_name.jobid = ORTE_JOBID_INVALID;
>>>>>
>>>>> if( NULL != item->map_node_name ) {
>>>>>
>>>>> Modified: trunk/orte/mca/errmgr/base/errmgr_base_tool.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/errmgr/base/errmgr_base_tool.c (original)
>>>>> +++ trunk/orte/mca/errmgr/base/errmgr_base_tool.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -267,7 +267,7 @@
>>>>> */
>>>>> errmgr_cmdline_sender.jobid = ORTE_JOBID_INVALID;
>>>>> errmgr_cmdline_sender.vpid = ORTE_VPID_INVALID;
>>>>> - errmgr_cmdline_sender.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(errmgr_cmdline_sender.epoch,ORTE_EPOCH_MIN);
>>>>> if (ORTE_SUCCESS != (ret = orte_rml.recv_buffer_nb(ORTE_NAME_WILDCARD,
>>>>> ORTE_RML_TAG_MIGRATE,
>>>>> 0,
>>>>> @@ -379,14 +379,14 @@
>>>>> if( OPAL_EQUAL != orte_util_compare_name_fields(ORTE_NS_CMP_ALL, ORTE_NAME_INVALID, &errmgr_cmdline_sender) ) {
>>>>> swap_dest.jobid = errmgr_cmdline_sender.jobid;
>>>>> swap_dest.vpid = errmgr_cmdline_sender.vpid;
>>>>> - swap_dest.epoch = errmgr_cmdline_sender.epoch;
>>>>> + ORTE_EPOCH_SET(swap_dest.epoch,errmgr_cmdline_sender.epoch);
>>>>>
>>>>> errmgr_cmdline_sender = *sender;
>>>>> orte_errmgr_base_migrate_update(ORTE_ERRMGR_MIGRATE_STATE_ERR_INPROGRESS);
>>>>>
>>>>> errmgr_cmdline_sender.jobid = swap_dest.jobid;
>>>>> errmgr_cmdline_sender.vpid = swap_dest.vpid;
>>>>> - errmgr_cmdline_sender.epoch = swap_dest.epoch;
>>>>> + ORTE_EPOCH_SET(errmgr_cmdline_sender.epoch,swap_dest.epoch);
>>>>>
>>>>> goto cleanup;
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/errmgr/hnp/errmgr_hnp.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/errmgr/hnp/errmgr_hnp.c (original)
>>>>> +++ trunk/orte/mca/errmgr/hnp/errmgr_hnp.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -53,6 +53,7 @@
>>>>> #include "orte/runtime/orte_globals.h"
>>>>> #include "orte/runtime/orte_locks.h"
>>>>> #include "orte/runtime/orte_quit.h"
>>>>> +#include "orte/runtime/data_type_support/orte_dt_support.h"
>>>>>
>>>>> #include "orte/mca/errmgr/errmgr.h"
>>>>> #include "orte/mca/errmgr/base/base.h"
>>>>> @@ -83,9 +84,11 @@
>>>>> orte_errmgr_hnp_global_suggest_map_targets,
>>>>> /* FT Event hook */
>>>>> orte_errmgr_hnp_global_ft_event,
>>>>> - orte_errmgr_base_register_migration_warning,
>>>>> + orte_errmgr_base_register_migration_warning
>>>>> +#if ORTE_RESIL_ORTE
>>>>> /* Set the callback */
>>>>> - orte_errmgr_base_set_fault_callback
>>>>> + ,orte_errmgr_base_set_fault_callback
>>>>> +#endif
>>>>> };
>>>>>
>>>>>
>>>>> @@ -97,14 +100,16 @@
>>>>> static void update_local_procs_in_job(orte_job_t *jdata, orte_job_state_t jobstate,
>>>>> orte_proc_state_t state, orte_exit_code_t exit_code);
>>>>> static void check_job_complete(orte_job_t *jdata);
>>>>> -static void killprocs(orte_jobid_t job, orte_vpid_t vpid, orte_epoch_t epoch);
>>>>> +static void killprocs(orte_jobid_t job, orte_vpid_t vpid);
>>>>> static int hnp_relocate(orte_job_t *jdata, orte_process_name_t *proc,
>>>>> orte_proc_state_t state, orte_exit_code_t exit_code);
>>>>> static orte_odls_child_t* proc_is_local(orte_process_name_t *proc);
>>>>> +#if ORTE_RESIL_ORTE
>>>>> static int send_to_local_applications(opal_pointer_array_t *dead_names);
>>>>> static void failure_notification(int status, orte_process_name_t* sender,
>>>>> opal_buffer_t *buffer, orte_rml_tag_t tag,
>>>>> void* cbdata);
>>>>> +#endif
>>>>>
>>>>> /************************
>>>>> * API Definitions
>>>>> @@ -380,16 +385,21 @@
>>>>> **********************/
>>>>> int orte_errmgr_hnp_base_global_init(void)
>>>>> {
>>>>> - int ret;
>>>>> + int ret = ORTE_SUCCESS;
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> ret = orte_rml.recv_buffer_nb(ORTE_NAME_WILDCARD, ORTE_RML_TAG_FAILURE_NOTICE,
>>>>> ORTE_RML_PERSISTENT, failure_notification, NULL);
>>>>> +#endif
>>>>> +
>>>>> return ret;
>>>>> }
>>>>>
>>>>> int orte_errmgr_hnp_base_global_finalize(void)
>>>>> {
>>>>> +#if ORTE_RESIL_ORTE
>>>>> orte_rml.recv_cancel(ORTE_NAME_WILDCARD, ORTE_RML_TAG_FAILURE_NOTICE);
>>>>> +#endif
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> @@ -406,6 +416,7 @@
>>>>> orte_odls_child_t *child;
>>>>> int rc;
>>>>> orte_app_context_t *app;
>>>>> + orte_proc_t *pdat;
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_errmgr_base.output,
>>>>> "%s errmgr:hnp: job %s reported state %s"
>>>>> @@ -538,7 +549,7 @@
>>>>> ORTE_PROC_STATE_SENSOR_BOUND_EXCEEDED,
>>>>> exit_code);
>>>>> /* order all local procs for this job to be killed */
>>>>> - killprocs(jdata->jobid, ORTE_VPID_WILDCARD, ORTE_EPOCH_WILDCARD);
>>>>> + killprocs(jdata->jobid, ORTE_VPID_WILDCARD);
>>>>> check_job_complete(jdata); /* set the local proc states */
>>>>> /* the job object for this job will have been NULL'd
>>>>> * in the array if the job was solely local. If it isn't
>>>>> @@ -550,7 +561,7 @@
>>>>> break;
>>>>> case ORTE_JOB_STATE_COMM_FAILED:
>>>>> /* order all local procs for this job to be killed */
>>>>> - killprocs(jdata->jobid, ORTE_VPID_WILDCARD, ORTE_EPOCH_WILDCARD);
>>>>> + killprocs(jdata->jobid, ORTE_VPID_WILDCARD);
>>>>> check_job_complete(jdata); /* set the local proc states */
>>>>> /* the job object for this job will have been NULL'd
>>>>> * in the array if the job was solely local. If it isn't
>>>>> @@ -562,7 +573,7 @@
>>>>> break;
>>>>> case ORTE_JOB_STATE_HEARTBEAT_FAILED:
>>>>> /* order all local procs for this job to be killed */
>>>>> - killprocs(jdata->jobid, ORTE_VPID_WILDCARD, ORTE_EPOCH_WILDCARD);
>>>>> + killprocs(jdata->jobid, ORTE_VPID_WILDCARD);
>>>>> check_job_complete(jdata); /* set the local proc states */
>>>>> /* the job object for this job will have been NULL'd
>>>>> * in the array if the job was solely local. If it isn't
>>>>> @@ -632,10 +643,6 @@
>>>>> }
>>>>> }
>>>>>
>>>>> - if (ORTE_PROC_STATE_ABORTED_BY_SIG == state) {
>>>>> - exit_code = 0;
>>>>> - }
>>>>> -
>>>>> orte_errmgr_hnp_update_proc(jdata, proc, state, pid, exit_code);
>>>>> check_job_complete(jdata); /* need to set the job state */
>>>>> /* the job object for this job will have been NULL'd
>>>>> @@ -679,7 +686,7 @@
>>>>>
>>>>> case ORTE_PROC_STATE_SENSOR_BOUND_EXCEEDED:
>>>>> if (jdata->enable_recovery) {
>>>>> - killprocs(proc->jobid, proc->vpid, proc->epoch);
>>>>> + killprocs(proc->jobid, proc->vpid);
>>>>> /* is this a local proc */
>>>>> if (NULL != (child = proc_is_local(proc))) {
>>>>> /* local proc - see if it has reached its restart limit */
>>>>> @@ -778,18 +785,37 @@
>>>>> opal_output(0, "%s UNABLE TO RELOCATE PROCS FROM FAILED DAEMON %s",
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), ORTE_NAME_PRINT(proc));
>>>>> /* kill all local procs */
>>>>> - killprocs(ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD, ORTE_EPOCH_WILDCARD);
>>>>> + killprocs(ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD);
>>>>> /* kill all jobs */
>>>>> hnp_abort(ORTE_JOBID_WILDCARD, exit_code);
>>>>> /* check if all is complete so we can terminate */
>>>>> check_job_complete(jdata);
>>>>> }
>>>>> } else {
>>>>> +#if !ORTE_RESIL_ORTE
>>>>> + if (NULL == (pdat = (orte_proc_t*)opal_pointer_array_get_item(jdata->procs, proc->vpid))) {
>>>>> + ORTE_ERROR_LOG(ORTE_ERR_NOT_FOUND);
>>>>> + orte_show_help("help-orte-errmgr-hnp.txt", "errmgr-hnp:daemon-died", true,
>>>>> + ORTE_VPID_PRINT(proc->vpid), "Unknown");
>>>>> + } else {
>>>>> + orte_show_help("help-orte-errmgr-hnp.txt", "errmgr-hnp:daemon-died", true,
>>>>> + ORTE_VPID_PRINT(proc->vpid),
>>>>> + (NULL == pdat->node) ? "Unknown" :
>>>>> + ((NULL == pdat->node->name) ? "Unknown" : pdat->node->name));
>>>>> + }
>>>>> +#endif
>>>>> if (ORTE_SUCCESS != orte_errmgr_hnp_record_dead_process(proc)) {
>>>>> /* The process is already dead so don't keep trying to do
>>>>> * this stuff. */
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> +
>>>>> +#if !ORTE_RESIL_ORTE
>>>>> + /* kill all local procs */
>>>>> + killprocs(ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD);
>>>>> + /* kill all jobs */
>>>>> + hnp_abort(ORTE_JOBID_WILDCARD, exit_code);
>>>>> +#endif
>>>>> /* We'll check if the job was complete when we get the
>>>>> * message back from the HNP notifying us of the dead
>>>>> * process */
>>>>> @@ -805,7 +831,7 @@
>>>>> } else {
>>>>> orte_errmgr_hnp_record_dead_process(proc);
>>>>> /* kill all local procs */
>>>>> - killprocs(ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD, ORTE_EPOCH_WILDCARD);
>>>>> + killprocs(ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD);
>>>>> /* kill all jobs */
>>>>> hnp_abort(ORTE_JOBID_WILDCARD, exit_code);
>>>>> return ORTE_ERR_UNRECOVERABLE;
>>>>> @@ -824,6 +850,7 @@
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> static void failure_notification(int status, orte_process_name_t* sender,
>>>>> opal_buffer_t *buffer, orte_rml_tag_t tag,
>>>>> void* cbdata)
>>>>> @@ -984,6 +1011,7 @@
>>>>>
>>>>> OBJ_RELEASE(dead_names);
>>>>> }
>>>>> +#endif
>>>>>
>>>>> /*****************
>>>>> * Local Functions
>>>>> @@ -1354,7 +1382,6 @@
>>>>> ORTE_UPDATE_EXIT_STATUS(proc->exit_code);
>>>>> }
>>>>> break;
>>>>> -#if 0
>>>>> case ORTE_PROC_STATE_ABORTED_BY_SIG:
>>>>> OPAL_OUTPUT_VERBOSE((5, orte_errmgr_base.output,
>>>>> "%s errmgr:hnp:check_job_completed proc %s aborted by signal",
>>>>> @@ -1370,7 +1397,6 @@
>>>>> ORTE_UPDATE_EXIT_STATUS(proc->exit_code);
>>>>> }
>>>>> break;
>>>>> -#endif
>>>>> case ORTE_PROC_STATE_TERM_WO_SYNC:
>>>>> OPAL_OUTPUT_VERBOSE((5, orte_errmgr_base.output,
>>>>> "%s errmgr:hnp:check_job_completed proc %s terminated without sync",
>>>>> @@ -1393,7 +1419,6 @@
>>>>> }
>>>>> break;
>>>>> case ORTE_PROC_STATE_COMM_FAILED:
>>>>> -#if 1
>>>>> if (!jdata->abort) {
>>>>> jdata->state = ORTE_JOB_STATE_COMM_FAILED;
>>>>> /* point to the lowest rank to cause the problem */
>>>>> @@ -1403,7 +1428,6 @@
>>>>> jdata->abort = true;
>>>>> ORTE_UPDATE_EXIT_STATUS(proc->exit_code);
>>>>> }
>>>>> -#endif
>>>>> break;
>>>>> case ORTE_PROC_STATE_SENSOR_BOUND_EXCEEDED:
>>>>> if (!jdata->abort) {
>>>>> @@ -1530,9 +1554,6 @@
>>>>> */
>>>>> CHECK_DAEMONS:
>>>>> if (jdata == NULL || jdata->jobid == ORTE_PROC_MY_NAME->jobid) {
>>>>> -#if 0
>>>>> - if ((jdata->num_procs - 1) <= jdata->num_terminated) { /* Subtract one for the HNP */
>>>>> -#endif
>>>>> if (0 == orte_routed.num_routes()) {
>>>>> /* orteds are done! */
>>>>> OPAL_OUTPUT_VERBOSE((5, orte_errmgr_base.output,
>>>>> @@ -1696,7 +1717,7 @@
>>>>> }
>>>>> }
>>>>>
>>>>> -static void killprocs(orte_jobid_t job, orte_vpid_t vpid, orte_epoch_t epoch)
>>>>> +static void killprocs(orte_jobid_t job, orte_vpid_t vpid)
>>>>> {
>>>>> opal_pointer_array_t cmd;
>>>>> orte_proc_t proc;
>>>>> @@ -1707,7 +1728,9 @@
>>>>> orte_sensor.stop(job);
>>>>> }
>>>>>
>>>>> - if (ORTE_JOBID_WILDCARD == job && ORTE_VPID_WILDCARD == vpid && ORTE_EPOCH_WILDCARD == epoch) {
>>>>> + if (ORTE_JOBID_WILDCARD == job
>>>>> + && ORTE_VPID_WILDCARD == vpid
>>>>> + && ORTE_EPOCH_CMP(ORTE_EPOCH_WILDCARD,epoch)) {
>>>>> if (ORTE_SUCCESS != (rc = orte_odls.kill_local_procs(NULL))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> }
>>>>> @@ -1718,7 +1741,7 @@
>>>>> OBJ_CONSTRUCT(&proc, orte_proc_t);
>>>>> proc.name.jobid = job;
>>>>> proc.name.vpid = vpid;
>>>>> - proc.name.epoch = epoch;
>>>>> + ORTE_EPOCH_SET(proc.name.epoch,epoch);
>>>>> opal_pointer_array_add(&cmd, &proc);
>>>>> if (ORTE_SUCCESS != (rc = orte_odls.kill_local_procs(&cmd))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> @@ -1913,13 +1936,15 @@
>>>>> }
>>>>>
>>>>> if (NULL != (pdat = (orte_proc_t*)opal_pointer_array_get_item(jdat->procs, proc->vpid)) &&
>>>>> - ORTE_PROC_STATE_TERMINATED < pdat->state) {
>>>>> + ORTE_PROC_STATE_TERMINATED > pdat->state) {
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* Make sure that the epochs match. */
>>>>> if (proc->epoch != pdat->name.epoch) {
>>>>> opal_output(1, "The epoch does not match the current epoch. Throwing the request out.");
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> dead_names = OBJ_NEW(opal_pointer_array_t);
>>>>>
>>>>> @@ -1935,6 +1960,7 @@
>>>>> }
>>>>> }
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> if (!mca_errmgr_hnp_component.term_in_progress) {
>>>>> /*
>>>>> * Send a message to the other daemons so they know that a daemon has
>>>>> @@ -1949,7 +1975,7 @@
>>>>> OBJ_RELEASE(buffer);
>>>>> } else {
>>>>>
>>>>> - /* Iterate of the list of dead procs and send them along with
>>>>> + /* Iterate over the list of dead procs and send them along with
>>>>> * the rest. The HNP needs this info so it can tell the other
>>>>> * ORTEDs and they can inform the appropriate applications.
>>>>> */
>>>>> @@ -1973,6 +1999,9 @@
>>>>> } else {
>>>>> orte_errmgr_hnp_global_mark_processes_as_dead(dead_names);
>>>>> }
>>>>> +#else
>>>>> + orte_errmgr_hnp_global_mark_processes_as_dead(dead_names);
>>>>> +#endif
>>>>> }
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> @@ -2011,6 +2040,7 @@
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
>>>>> ORTE_NAME_PRINT(&pdat->name)));
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> /* Make sure the epochs match, if not it probably means that we
>>>>> * already reported this failure. */
>>>>> if (name_item->epoch != pdat->name.epoch) {
>>>>> @@ -2018,6 +2048,7 @@
>>>>> }
>>>>>
>>>>> orte_util_set_epoch(name_item, name_item->epoch + 1);
>>>>> +#endif
>>>>>
>>>>> /* Remove it from the job array */
>>>>> opal_pointer_array_set_item(jdat->procs, name_item->vpid, NULL);
>>>>> @@ -2034,6 +2065,7 @@
>>>>>
>>>>> OBJ_RELEASE(pdat);
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> /* Create a new proc object that will keep track of the epoch
>>>>> * information */
>>>>> pdat = OBJ_NEW(orte_proc_t);
>>>>> @@ -2041,14 +2073,15 @@
>>>>> pdat->name.vpid = name_item->vpid;
>>>>> pdat->name.epoch = name_item->epoch + 1;
>>>>>
>>>>> - /* Set the state as terminated so we'll know the process isn't
>>>>> - * actually there. */
>>>>> - pdat->state = ORTE_PROC_STATE_TERMINATED;
>>>>> -
>>>>> opal_pointer_array_set_item(jdat->procs, name_item->vpid, pdat);
>>>>> jdat->num_procs++;
>>>>> jdat->num_terminated++;
>>>>> +#endif
>>>>> + /* Set the state as terminated so we'll know the process isn't
>>>>> + * actually there. */
>>>>> + pdat->state = ORTE_PROC_STATE_TERMINATED;
>>>>> } else {
>>>>> +#if ORTE_RESIL_ORTE
>>>>> opal_output(0, "Proc data not found for %s", ORTE_NAME_PRINT(name_item));
>>>>> /* Create a new proc object that will keep track of the epoch
>>>>> * information */
>>>>> @@ -2064,11 +2097,13 @@
>>>>> opal_pointer_array_set_item(jdat->procs, name_item->vpid, pdat);
>>>>> jdat->num_procs++;
>>>>> jdat->num_terminated++;
>>>>> +#endif
>>>>> }
>>>>>
>>>>> check_job_complete(jdat);
>>>>> }
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> if (!orte_orteds_term_ordered) {
>>>>> /* Need to update the orted routing module. */
>>>>> orte_routed.update_routing_tree(ORTE_PROC_MY_NAME->jobid);
>>>>> @@ -2077,10 +2112,12 @@
>>>>> (*fault_cbfunc)(dead_procs);
>>>>> }
>>>>> }
>>>>> +#endif
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> int send_to_local_applications(opal_pointer_array_t *dead_names) {
>>>>> opal_buffer_t *buf;
>>>>> int ret = ORTE_SUCCESS;
>>>>> @@ -2121,3 +2158,5 @@
>>>>>
>>>>> return ret;
>>>>> }
>>>>> +#endif
>>>>> +
>>>>>
>>>>> Modified: trunk/orte/mca/errmgr/hnp/errmgr_hnp_autor.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/errmgr/hnp/errmgr_hnp_autor.c (original)
>>>>> +++ trunk/orte/mca/errmgr/hnp/errmgr_hnp_autor.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -522,7 +522,7 @@
>>>>> wp_item = OBJ_NEW(errmgr_autor_wp_item_t);
>>>>> wp_item->name.jobid = proc->jobid;
>>>>> wp_item->name.vpid = proc->vpid;
>>>>> - wp_item->name.epoch = proc->epoch;
>>>>> + ORTE_EPOCH_SET(wp_item->name.epoch,proc->epoch);
>>>>> wp_item->state = state;
>>>>>
>>>>> opal_list_append(procs_pending_recovery, &(wp_item->super));
>>>>> @@ -626,7 +626,7 @@
>>>>> {
>>>>> wp->name.jobid = ORTE_JOBID_INVALID;
>>>>> wp->name.vpid = ORTE_VPID_INVALID;
>>>>> - wp->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(wp->name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> wp->state = 0;
>>>>> }
>>>>> @@ -635,7 +635,7 @@
>>>>> {
>>>>> wp->name.jobid = ORTE_JOBID_INVALID;
>>>>> wp->name.vpid = ORTE_VPID_INVALID;
>>>>> - wp->name.epoch = ORTE_EPOCH_INVALID;
>>>>> + ORTE_EPOCH_SET(wp->name.epoch,ORTE_EPOCH_INVALID);
>>>>>
>>>>> wp->state = 0;
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/errmgr/hnp/errmgr_hnp_crmig.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/errmgr/hnp/errmgr_hnp_crmig.c (original)
>>>>> +++ trunk/orte/mca/errmgr/hnp/errmgr_hnp_crmig.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -750,7 +750,7 @@
>>>>> close_iof_stdin = true;
>>>>> iof_name.jobid = proc->name.jobid;
>>>>> iof_name.vpid = proc->name.vpid;
>>>>> - iof_name.epoch = proc->name.epoch;
>>>>> + ORTE_EPOCH_SET(iof_name.epoch,proc->name.epoch);
>>>>> }
>>>>> }
>>>>> }
>>>>> @@ -807,7 +807,7 @@
>>>>> close_iof_stdin = true;
>>>>> iof_name.jobid = proc->name.jobid;
>>>>> iof_name.vpid = proc->name.vpid;
>>>>> - iof_name.epoch = proc->name.epoch;
>>>>> + ORTE_EPOCH_SET(iof_name.epoch,proc->name.epoch);
>>>>> }
>>>>> }
>>>>> }
>>>>> @@ -855,7 +855,7 @@
>>>>> close_iof_stdin = true;
>>>>> iof_name.jobid = proc->name.jobid;
>>>>> iof_name.vpid = proc->name.vpid;
>>>>> - iof_name.epoch = proc->name.epoch;
>>>>> + ORTE_EPOCH_SET(iof_name.epoch,proc->name.epoch);
>>>>> }
>>>>> }
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/errmgr/orted/errmgr_orted.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/errmgr/orted/errmgr_orted.c (original)
>>>>> +++ trunk/orte/mca/errmgr/orted/errmgr_orted.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -34,6 +34,7 @@
>>>>> #include "orte/util/show_help.h"
>>>>> #include "orte/util/nidmap.h"
>>>>> #include "orte/runtime/orte_globals.h"
>>>>> +#include "orte/runtime/data_type_support/orte_dt_support.h"
>>>>> #include "orte/mca/rml/rml.h"
>>>>> #include "orte/mca/odls/odls.h"
>>>>> #include "orte/mca/odls/base/base.h"
>>>>> @@ -41,7 +42,9 @@
>>>>> #include "orte/mca/plm/plm_types.h"
>>>>> #include "orte/mca/routed/routed.h"
>>>>> #include "orte/mca/sensor/sensor.h"
>>>>> +#include "orte/mca/ess/ess.h"
>>>>> #include "orte/runtime/orte_quit.h"
>>>>> +#include "orte/runtime/orte_globals.h"
>>>>>
>>>>> #include "orte/mca/errmgr/errmgr.h"
>>>>> #include "orte/mca/errmgr/base/base.h"
>>>>> @@ -59,13 +62,15 @@
>>>>> static void update_local_children(orte_odls_job_t *jobdat,
>>>>> orte_job_state_t jobstate,
>>>>> orte_proc_state_t state);
>>>>> -static void killprocs(orte_jobid_t job, orte_vpid_t vpid, orte_epoch_t epoch);
>>>>> +static void killprocs(orte_jobid_t job, orte_vpid_t vpid);
>>>>> static int record_dead_process(orte_process_name_t *proc);
>>>>> -static int send_to_local_applications(opal_pointer_array_t *dead_names);
>>>>> static int mark_processes_as_dead(opal_pointer_array_t *dead_procs);
>>>>> +#if ORTE_RESIL_ORTE
>>>>> +static int send_to_local_applications(opal_pointer_array_t *dead_names);
>>>>> static void failure_notification(int status, orte_process_name_t* sender,
>>>>> opal_buffer_t *buffer, orte_rml_tag_t tag,
>>>>> void* cbdata);
>>>>> +#endif
>>>>>
>>>>> /*
>>>>> * Module functions: Global
>>>>> @@ -104,8 +109,10 @@
>>>>> predicted_fault,
>>>>> suggest_map_targets,
>>>>> ft_event,
>>>>> - orte_errmgr_base_register_migration_warning,
>>>>> - orte_errmgr_base_set_fault_callback /* Set callback function */
>>>>> + orte_errmgr_base_register_migration_warning
>>>>> +#if ORTE_RESIL_ORTE
>>>>> + ,orte_errmgr_base_set_fault_callback /* Set callback function */
>>>>> +#endif
>>>>> };
>>>>>
>>>>> /************************
>>>>> @@ -113,16 +120,22 @@
>>>>> ************************/
>>>>> static int init(void)
>>>>> {
>>>>> - int ret;
>>>>> + int ret = ORTE_SUCCESS;
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> ret = orte_rml.recv_buffer_nb(ORTE_NAME_WILDCARD, ORTE_RML_TAG_FAILURE_NOTICE,
>>>>> ORTE_RML_PERSISTENT, failure_notification, NULL);
>>>>> +#endif
>>>>> +
>>>>> return ret;
>>>>> }
>>>>>
>>>>> static int finalize(void)
>>>>> {
>>>>> +#if ORTE_RESIL_ORTE
>>>>> orte_rml.recv_cancel(ORTE_NAME_WILDCARD, ORTE_RML_TAG_FAILURE_NOTICE);
>>>>> +#endif
>>>>> +
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> @@ -228,10 +241,10 @@
>>>>> /* update all procs in job */
>>>>> update_local_children(jobdat, jobstate, ORTE_PROC_STATE_SENSOR_BOUND_EXCEEDED);
>>>>> /* order all local procs for this job to be killed */
>>>>> - killprocs(jobdat->jobid, ORTE_VPID_WILDCARD, ORTE_EPOCH_WILDCARD);
>>>>> + killprocs(jobdat->jobid, ORTE_VPID_WILDCARD);
>>>>> case ORTE_JOB_STATE_COMM_FAILED:
>>>>> /* kill all local procs */
>>>>> - killprocs(ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD, ORTE_EPOCH_WILDCARD);
>>>>> + killprocs(ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD);
>>>>> /* tell the caller we can't recover */
>>>>> return ORTE_ERR_UNRECOVERABLE;
>>>>> break;
>>>>> @@ -276,7 +289,7 @@
>>>>> /* see if this was a lifeline */
>>>>> if (ORTE_SUCCESS != orte_routed.route_lost(proc)) {
>>>>> /* kill our children */
>>>>> - killprocs(ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD, ORTE_EPOCH_WILDCARD);
>>>>> + killprocs(ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD);
>>>>> /* terminate - our routed children will see
>>>>> * us leave and automatically die
>>>>> */
>>>>> @@ -290,10 +303,18 @@
>>>>> if (0 == orte_routed.num_routes() &&
>>>>> 0 == opal_list_get_size(&orte_local_children)) {
>>>>> orte_quit();
>>>>> + } else {
>>>>> + OPAL_OUTPUT_VERBOSE((5, orte_errmgr_base.output,
>>>>> + "%s errmgr:orted not exiting, num_routes() == %d, num children == %d",
>>>>> + ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
>>>>> + orte_routed.num_routes(),
>>>>> + opal_list_get_size(&orte_local_children)));
>>>>> }
>>>>> }
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> record_dead_process(proc);
>>>>> +#endif
>>>>>
>>>>> /* if not, then indicate we can continue */
>>>>> return ORTE_SUCCESS;
>>>>> @@ -344,7 +365,7 @@
>>>>> /* Decrement the number of local procs */
>>>>> jobdat->num_local_procs--;
>>>>> /* kill this proc */
>>>>> - killprocs(proc->jobid, proc->vpid, proc->epoch);
>>>>> + killprocs(proc->jobid, proc->vpid);
>>>>> }
>>>>> app = (orte_app_context_t*)opal_pointer_array_get_item(&jobdat->apps, child->app_idx);
>>>>> if( jobdat->enable_recovery && child->restarts < app->max_restarts ) {
>>>>> @@ -526,10 +547,12 @@
>>>>> ORTE_ERROR_LOG(rc);
>>>>> goto FINAL_CLEANUP;
>>>>> }
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> if (ORTE_SUCCESS != (rc = opal_dss.pack(alert, &child->name->epoch, 1, ORTE_EPOCH))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> goto FINAL_CLEANUP;
>>>>> }
>>>>> +#endif
>>>>> }
>>>>> }
>>>>> /* pack an invalid marker */
>>>>> @@ -660,7 +683,7 @@
>>>>> continue;
>>>>> }
>>>>>
>>>>> - if (name_item->epoch < orte_util_lookup_epoch(name_item)) {
>>>>> + if (0 < ORTE_EPOCH_CMP(name_item->epoch,orte_ess.proc_get_epoch(name_item))) {
>>>>> continue;
>>>>> }
>>>>>
>>>>> @@ -669,9 +692,11 @@
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
>>>>> ORTE_NAME_PRINT(name_item)));
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* Increment the epoch */
>>>>> orte_util_set_proc_state(name_item, ORTE_PROC_STATE_TERMINATED);
>>>>> orte_util_set_epoch(name_item, name_item->epoch + 1);
>>>>> +#endif
>>>>>
>>>>> OPAL_THREAD_LOCK(&orte_odls_globals.mutex);
>>>>>
>>>>> @@ -706,6 +731,7 @@
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> static void failure_notification(int status, orte_process_name_t* sender,
>>>>> opal_buffer_t *buffer, orte_rml_tag_t tag,
>>>>> void* cbdata)
>>>>> @@ -714,7 +740,7 @@
>>>>> orte_std_cntr_t n;
>>>>> int ret = ORTE_SUCCESS, num_failed;
>>>>> int32_t i;
>>>>> - orte_process_name_t *name_item, proc;
>>>>> + orte_process_name_t *name_item;
>>>>>
>>>>> dead_names = OBJ_NEW(opal_pointer_array_t);
>>>>>
>>>>> @@ -746,7 +772,7 @@
>>>>> /* There shouldn't be an issue of receiving this message multiple
>>>>> * times but it doesn't hurt to double check.
>>>>> */
>>>>> - if (proc.epoch < orte_util_lookup_epoch(name_item)) {
>>>>> + if (0 < ORTE_EPOCH_CMP(name_item->epoch,orte_ess.proc_get_epoch(name_item))) {
>>>>> opal_output(1, "Received from proc %s local epoch %d", ORTE_NAME_PRINT(name_item), orte_util_lookup_epoch(name_item));
>>>>> continue;
>>>>> }
>>>>> @@ -767,6 +793,7 @@
>>>>> free(name_item);
>>>>> }
>>>>> }
>>>>> +#endif
>>>>>
>>>>> /*****************
>>>>> * Local Functions
>>>>> @@ -948,11 +975,13 @@
>>>>> ORTE_ERROR_LOG(rc);
>>>>> return rc;
>>>>> }
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* Pack the child's epoch. */
>>>>> if (ORTE_SUCCESS != (rc = opal_dss.pack(buf, &(child->name->epoch), 1, ORTE_EPOCH))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> return rc;
>>>>> }
>>>>> +#endif
>>>>> /* pack the contact info */
>>>>> if (ORTE_SUCCESS != (rc = opal_dss.pack(buf, &child->rml_uri, 1, OPAL_STRING))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> @@ -1015,7 +1044,7 @@
>>>>> }
>>>>> }
>>>>>
>>>>> -static void killprocs(orte_jobid_t job, orte_vpid_t vpid, orte_epoch_t epoch)
>>>>> +static void killprocs(orte_jobid_t job, orte_vpid_t vpid)
>>>>> {
>>>>> opal_pointer_array_t cmd;
>>>>> orte_proc_t proc;
>>>>> @@ -1026,7 +1055,9 @@
>>>>> orte_sensor.stop(job);
>>>>> }
>>>>>
>>>>> - if (ORTE_JOBID_WILDCARD == job && ORTE_VPID_WILDCARD == vpid && ORTE_EPOCH_WILDCARD == epoch) {
>>>>> + if (ORTE_JOBID_WILDCARD == job
>>>>> + && ORTE_VPID_WILDCARD == vpid
>>>>> + && 0 == ORTE_EPOCH_CMP(ORTE_EPOCH_WILDCARD,epoch)) {
>>>>> if (ORTE_SUCCESS != (rc = orte_odls.kill_local_procs(NULL))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> }
>>>>> @@ -1037,7 +1068,7 @@
>>>>> OBJ_CONSTRUCT(&proc, orte_proc_t);
>>>>> proc.name.jobid = job;
>>>>> proc.name.vpid = vpid;
>>>>> - proc.name.epoch = epoch;
>>>>> + ORTE_EPOCH_SET(proc.name.epoch,epoch);
>>>>> opal_pointer_array_add(&cmd, &proc);
>>>>> if (ORTE_SUCCESS != (rc = orte_odls.kill_local_procs(&cmd))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> @@ -1082,20 +1113,21 @@
>>>>> return rc;
>>>>> }
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> int send_to_local_applications(opal_pointer_array_t *dead_names) {
>>>>> opal_buffer_t *buf;
>>>>> int ret;
>>>>> orte_process_name_t *name_item;
>>>>> int size, i;
>>>>>
>>>>> - OPAL_OUTPUT_VERBOSE((10, orte_errmgr_base.output,
>>>>> - "%s Sending failure to local applications.",
>>>>> - ORTE_NAME_PRINT(ORTE_PROC_MY_NAME)));
>>>>> -
>>>>> buf = OBJ_NEW(opal_buffer_t);
>>>>>
>>>>> size = opal_pointer_array_get_size(dead_names);
>>>>>
>>>>> + OPAL_OUTPUT_VERBOSE((10, orte_errmgr_base.output,
>>>>> + "%s Sending %d failure(s) to local applications.",
>>>>> + ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), size));
>>>>> +
>>>>> if (ORTE_SUCCESS != (ret = opal_dss.pack(buf, &size, 1, ORTE_VPID))) {
>>>>> ORTE_ERROR_LOG(ret);
>>>>> OBJ_RELEASE(buf);
>>>>> @@ -1122,4 +1154,5 @@
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> +#endif
>>>>>
>>>>>
>>>>> Modified: trunk/orte/mca/ess/alps/ess_alps_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/alps/ess_alps_module.c (original)
>>>>> +++ trunk/orte/mca/ess/alps/ess_alps_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -363,8 +363,8 @@
>>>>>
>>>>> ORTE_PROC_MY_NAME->jobid = jobid;
>>>>> ORTE_PROC_MY_NAME->vpid = (orte_vpid_t) cnos_get_rank() + starting_vpid;
>>>>> - ORTE_PROC_MY_NAME->epoch = ORTE_EPOCH_INVALID;
>>>>> - ORTE_PROC_MY_NAME->epoch = orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME);
>>>>> + ORTE_EPOCH_PRINT(ORTE_PROC_MY_NAME->epoch,ORTE_EPOCH_INVALID);
>>>>> + ORTE_EPOCH_PRINT(ORTE_PROC_MY_NAME->epoch,orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME));
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_ess_base_output,
>>>>> "ess:alps set name to %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME)));
>>>>>
>>>>> Modified: trunk/orte/mca/ess/base/base.h
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/base/base.h (original)
>>>>> +++ trunk/orte/mca/ess/base/base.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -57,7 +57,11 @@
>>>>>
>>>>> ORTE_DECLSPEC extern opal_list_t orte_ess_base_components_available;
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> ORTE_DECLSPEC orte_epoch_t orte_ess_base_proc_get_epoch(orte_process_name_t *proc);
>>>>> +#else
>>>>> +ORTE_DECLSPEC int orte_ess_base_proc_get_epoch(orte_process_name_t *proc);
>>>>> +#endif
>>>>>
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>>
>>>>>
>>>>> Modified: trunk/orte/mca/ess/base/ess_base_select.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/base/ess_base_select.c (original)
>>>>> +++ trunk/orte/mca/ess/base/ess_base_select.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -36,21 +36,19 @@
>>>>> * Generic function to retrieve the epoch of a specific process
>>>>> * from the job data.
>>>>> */
>>>>> +#if !ORTE_ENABLE_EPOCH
>>>>> +int orte_ess_base_proc_get_epoch(orte_process_name_t *proc) {
>>>>> + return 0;
>>>>> +}
>>>>> +#else
>>>>> orte_epoch_t orte_ess_base_proc_get_epoch(orte_process_name_t *proc) {
>>>>> orte_epoch_t epoch = ORTE_EPOCH_INVALID;
>>>>>
>>>>> -#if !ORTE_DISABLE_FULL_SUPPORT
>>>>> epoch = orte_util_lookup_epoch(proc);
>>>>> -#endif
>>>>> -
>>>>> - OPAL_OUTPUT_VERBOSE((2, orte_ess_base_output,
>>>>> - "%s ess:generic: proc %s has epoch %d",
>>>>> - ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
>>>>> - ORTE_NAME_PRINT(proc),
>>>>> - epoch));
>>>>>
>>>>> return epoch;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> int
>>>>> orte_ess_base_select(void)
>>>>>
>>>>> Modified: trunk/orte/mca/ess/env/ess_env_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/env/ess_env_module.c (original)
>>>>> +++ trunk/orte/mca/ess/env/ess_env_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -392,8 +392,7 @@
>>>>>
>>>>> ORTE_PROC_MY_NAME->jobid = jobid;
>>>>> ORTE_PROC_MY_NAME->vpid = vpid;
>>>>> - ORTE_PROC_MY_NAME->epoch = ORTE_EPOCH_INVALID;
>>>>> - ORTE_PROC_MY_NAME->epoch = orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME);
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_NAME->epoch,orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME));
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_ess_base_output,
>>>>> "ess:env set name to %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME)));
>>>>>
>>>>> Modified: trunk/orte/mca/ess/ess.h
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/ess.h (original)
>>>>> +++ trunk/orte/mca/ess/ess.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -111,7 +111,11 @@
>>>>> * will get the most up to date version stored within the orte_proc_t struct.
>>>>> * Obviously the epoch of the proc that is passed in will be ignored.
>>>>> */
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> typedef orte_epoch_t (*orte_ess_base_module_proc_get_epoch_fn_t)(orte_process_name_t *proc);
>>>>> +#else
>>>>> +typedef int (*orte_ess_base_module_proc_get_epoch_fn_t)(orte_process_name_t *proc);
>>>>> +#endif
>>>>>
>>>>> /**
>>>>> * Update the pidmap
>>>>>
>>>>> Modified: trunk/orte/mca/ess/generic/ess_generic_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/generic/ess_generic_module.c (original)
>>>>> +++ trunk/orte/mca/ess/generic/ess_generic_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -155,7 +155,7 @@
>>>>> goto error;
>>>>> }
>>>>> ORTE_PROC_MY_NAME->vpid = strtol(envar, NULL, 10);
>>>>> - ORTE_PROC_MY_NAME->epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_NAME->epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_ess_base_output,
>>>>> "%s completed name definition",
>>>>> @@ -273,7 +273,7 @@
>>>>> if (vpid == ORTE_PROC_MY_NAME->vpid) {
>>>>> ORTE_PROC_MY_DAEMON->jobid = 0;
>>>>> ORTE_PROC_MY_DAEMON->vpid = i;
>>>>> - ORTE_PROC_MY_DAEMON->epoch = ORTE_PROC_MY_NAME->epoch;
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_DAEMON->epoch,ORTE_PROC_MY_NAME->epoch);
>>>>> }
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_ess_base_output,
>>>>> "%s node %d name %s rank %s",
>>>>> @@ -304,7 +304,7 @@
>>>>> if (vpid == ORTE_PROC_MY_NAME->vpid) {
>>>>> ORTE_PROC_MY_DAEMON->jobid = 0;
>>>>> ORTE_PROC_MY_DAEMON->vpid = i;
>>>>> - ORTE_PROC_MY_DAEMON->epoch = ORTE_PROC_MY_NAME->epoch;
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_DAEMON->epoch,ORTE_PROC_MY_NAME->epoch);
>>>>> }
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_ess_base_output,
>>>>> "%s node %d name %s rank %d",
>>>>>
>>>>> Modified: trunk/orte/mca/ess/hnp/ess_hnp_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/hnp/ess_hnp_module.c (original)
>>>>> +++ trunk/orte/mca/ess/hnp/ess_hnp_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -494,7 +494,7 @@
>>>>> proc = OBJ_NEW(orte_proc_t);
>>>>> proc->name.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> proc->name.vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - proc->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> proc->pid = orte_process_info.pid;
>>>>> proc->rml_uri = orte_rml.get_contact_info();
>>>>>
>>>>> Modified: trunk/orte/mca/ess/lsf/ess_lsf_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/lsf/ess_lsf_module.c (original)
>>>>> +++ trunk/orte/mca/ess/lsf/ess_lsf_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -357,8 +357,7 @@
>>>>>
>>>>> ORTE_PROC_MY_NAME->jobid = jobid;
>>>>> ORTE_PROC_MY_NAME->vpid = vpid;
>>>>> - ORTE_PROC_MY_NAME->epoch = ORTE_EPOCH_INVALID;
>>>>> - ORTE_PROC_MY_NAME->epoch = orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME);
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_NAME->epoch,orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME));
>>>>>
>>>>> /* fix up the base name and make it the "real" name */
>>>>> lsf_nodeid = atoi(getenv("LSF_PM_TASKID"));
>>>>>
>>>>> Modified: trunk/orte/mca/ess/singleton/ess_singleton_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/singleton/ess_singleton_module.c (original)
>>>>> +++ trunk/orte/mca/ess/singleton/ess_singleton_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -188,7 +188,7 @@
>>>>> /* set the name */
>>>>> ORTE_PROC_MY_NAME->jobid = 0xffff0000 & ((uint32_t)jobfam << 16);
>>>>> ORTE_PROC_MY_NAME->vpid = 0;
>>>>> - ORTE_PROC_MY_NAME->epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_NAME->epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> } else {
>>>>> /*
>>>>>
>>>>> Modified: trunk/orte/mca/ess/slave/ess_slave_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/slave/ess_slave_module.c (original)
>>>>> +++ trunk/orte/mca/ess/slave/ess_slave_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -280,8 +280,7 @@
>>>>>
>>>>> ORTE_PROC_MY_NAME->jobid = jobid;
>>>>> ORTE_PROC_MY_NAME->vpid = vpid;
>>>>> - ORTE_PROC_MY_NAME->epoch = ORTE_EPOCH_INVALID;
>>>>> - ORTE_PROC_MY_NAME->epoch = orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME);
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_NAME->epoch,orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME));
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_ess_base_output,
>>>>> "ess:slave set name to %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME)));
>>>>>
>>>>> Modified: trunk/orte/mca/ess/slurm/ess_slurm_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/slurm/ess_slurm_module.c (original)
>>>>> +++ trunk/orte/mca/ess/slurm/ess_slurm_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -368,8 +368,7 @@
>>>>> /* fix up the vpid and make it the "real" vpid */
>>>>> slurm_nodeid = atoi(getenv("SLURM_NODEID"));
>>>>> ORTE_PROC_MY_NAME->vpid = vpid + slurm_nodeid;
>>>>> - ORTE_PROC_MY_NAME->epoch = ORTE_EPOCH_INVALID;
>>>>> - ORTE_PROC_MY_NAME->epoch = orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME);
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_NAME->epoch,orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME));
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_ess_base_output,
>>>>> "ess:slurm set name to %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME)));
>>>>>
>>>>> Modified: trunk/orte/mca/ess/slurmd/ess_slurmd_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/slurmd/ess_slurmd_module.c (original)
>>>>> +++ trunk/orte/mca/ess/slurmd/ess_slurmd_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -195,7 +195,7 @@
>>>>> }
>>>>> ORTE_PROC_MY_NAME->vpid = strtol(envar, NULL, 10);
>>>>> #endif
>>>>> - ORTE_PROC_MY_NAME->epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_NAME->epoch,ORTE_EPOCH_MIN);
>>>>> /* get our local rank */
>>>>> if (NULL == (envar = getenv("SLURM_LOCALID"))) {
>>>>> error = "could not get SLURM_LOCALID";
>>>>> @@ -260,7 +260,7 @@
>>>>> nodeid = strtol(envar, NULL, 10);
>>>>> ORTE_PROC_MY_DAEMON->jobid = 0;
>>>>> ORTE_PROC_MY_DAEMON->vpid = nodeid;
>>>>> - ORTE_PROC_MY_DAEMON->epoch = ORTE_PROC_MY_NAME->epoch;
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_DAEMON->epoch,ORTE_PROC_MY_NAME->epoch);
>>>>>
>>>>> /* get the number of ppn */
>>>>> if (NULL == (tasks_per_node = getenv("SLURM_STEP_TASKS_PER_NODE"))) {
>>>>>
>>>>> Modified: trunk/orte/mca/ess/tm/ess_tm_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/ess/tm/ess_tm_module.c (original)
>>>>> +++ trunk/orte/mca/ess/tm/ess_tm_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -364,7 +364,7 @@
>>>>>
>>>>> ORTE_PROC_MY_NAME->jobid = jobid;
>>>>> ORTE_PROC_MY_NAME->vpid = vpid;
>>>>> - ORTE_PROC_MY_NAME->epoch = orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME);
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_NAME->epoch,orte_ess.proc_get_epoch(ORTE_PROC_MY_NAME));
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_ess_base_output,
>>>>> "ess:tm set name to %s", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME)));
>>>>>
>>>>> Modified: trunk/orte/mca/filem/rsh/filem_rsh_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/filem/rsh/filem_rsh_module.c (original)
>>>>> +++ trunk/orte/mca/filem/rsh/filem_rsh_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -1097,11 +1097,11 @@
>>>>> if( NULL != proc_set ) {
>>>>> wp_item->proc_set.source.jobid = proc_set->source.jobid;
>>>>> wp_item->proc_set.source.vpid = proc_set->source.vpid;
>>>>> - wp_item->proc_set.source.epoch = proc_set->source.epoch;
>>>>> + ORTE_EPOCH_SET(wp_item->proc_set.source.epoch,proc_set->source.epoch);
>>>>>
>>>>> wp_item->proc_set.sink.jobid = proc_set->sink.jobid;
>>>>> wp_item->proc_set.sink.vpid = proc_set->sink.vpid;
>>>>> - wp_item->proc_set.sink.epoch = proc_set->sink.epoch;
>>>>> + ORTE_EPOCH_SET(wp_item->proc_set.sink.epoch,proc_set->sink.epoch);
>>>>> }
>>>>> /* Copy the File Set */
>>>>> if( NULL != file_set ) {
>>>>> @@ -1396,7 +1396,7 @@
>>>>> wp_item = OBJ_NEW(orte_filem_rsh_work_pool_item_t);
>>>>> wp_item->proc_set.source.jobid = sender->jobid;
>>>>> wp_item->proc_set.source.vpid = sender->vpid;
>>>>> - wp_item->proc_set.source.epoch = sender->epoch;
>>>>> + ORTE_EPOCH_SET(wp_item->proc_set.source.epoch,sender->epoch);
>>>>>
>>>>> opal_list_append(&work_pool_waiting, &(wp_item->super));
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/grpcomm/base/grpcomm_base_coll.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/grpcomm/base/grpcomm_base_coll.c (original)
>>>>> +++ trunk/orte/mca/grpcomm/base/grpcomm_base_coll.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -168,8 +168,7 @@
>>>>> if (vpids[0] == ORTE_PROC_MY_NAME->vpid) {
>>>>> /* I send first */
>>>>> peer.vpid = vpids[1];
>>>>> - peer.epoch = ORTE_EPOCH_INVALID;
>>>>> - peer.epoch = orte_ess.proc_get_epoch(&peer);
>>>>> + ORTE_EPOCH_SET(peer.epoch,orte_ess.proc_get_epoch(&peer));
>>>>>
>>>>> /* setup a temp buffer so I can inform the other side as to the
>>>>> * number of entries in my buffer
>>>>> @@ -226,8 +225,7 @@
>>>>> opal_dss.pack(&buf, &num_entries, 1, OPAL_INT32);
>>>>> opal_dss.copy_payload(&buf, sendbuf);
>>>>> peer.vpid = vpids[0];
>>>>> - peer.epoch = ORTE_EPOCH_INVALID;
>>>>> - peer.epoch = orte_ess.proc_get_epoch(&peer);
>>>>> + ORTE_EPOCH_SET(peer.epoch,orte_ess.proc_get_epoch(&peer));
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((5, orte_grpcomm_base.output,
>>>>> "%s grpcomm:coll:two-proc sending to %s",
>>>>> @@ -320,8 +318,7 @@
>>>>> /* first send my current contents */
>>>>> nv = (rank - distance + np) % np;
>>>>> peer.vpid = vpids[nv];
>>>>> - peer.epoch = ORTE_EPOCH_INVALID;
>>>>> - peer.epoch = orte_ess.proc_get_epoch(&peer);
>>>>> + ORTE_EPOCH_SET(peer.epoch,orte_ess.proc_get_epoch(&peer));
>>>>>
>>>>> OBJ_CONSTRUCT(&buf, opal_buffer_t);
>>>>> opal_dss.pack(&buf, &total_entries, 1, OPAL_INT32);
>>>>> @@ -340,8 +337,7 @@
>>>>> num_recvd = 0;
>>>>> nv = (rank + distance) % np;
>>>>> peer.vpid = vpids[nv];
>>>>> - peer.epoch = ORTE_EPOCH_INVALID;
>>>>> - peer.epoch = orte_ess.proc_get_epoch(&peer);
>>>>> + ORTE_EPOCH_SET(peer.epoch,orte_ess.proc_get_epoch(&peer));
>>>>>
>>>>> OBJ_CONSTRUCT(&bucket, opal_buffer_t);
>>>>> if (ORTE_SUCCESS != (rc = orte_rml.recv_buffer_nb(&peer,
>>>>> @@ -439,8 +435,7 @@
>>>>> /* first send my current contents */
>>>>> nv = rank ^ distance;
>>>>> peer.vpid = vpids[nv];
>>>>> - peer.epoch = ORTE_EPOCH_INVALID;
>>>>> - peer.epoch = orte_ess.proc_get_epoch(&peer);
>>>>> + ORTE_EPOCH_SET(peer.epoch,orte_ess.proc_get_epoch(&peer));
>>>>>
>>>>> OBJ_CONSTRUCT(&buf, opal_buffer_t);
>>>>> opal_dss.pack(&buf, &total_entries, 1, OPAL_INT32);
>>>>> @@ -646,8 +641,7 @@
>>>>> proc.jobid = jobid;
>>>>> proc.vpid = 0;
>>>>> while (proc.vpid < jobdat->num_procs && 0 < opal_list_get_size(&daemon_tree)) {
>>>>> - proc.epoch = ORTE_EPOCH_INVALID;
>>>>> - proc.epoch = orte_ess.proc_get_epoch(&proc);
>>>>> + ORTE_EPOCH_SET(proc.epoch,orte_ess.proc_get_epoch(&proc));
>>>>>
>>>>> /* get the daemon that hosts this proc */
>>>>> daemonvpid = orte_ess.proc_get_daemon(&proc);
>>>>> @@ -713,8 +707,7 @@
>>>>> /* send it */
>>>>> my_parent.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> my_parent.vpid = orte_routed.get_routing_tree(NULL);
>>>>> - my_parent.epoch = ORTE_EPOCH_INVALID;
>>>>> - my_parent.epoch = orte_ess.proc_get_epoch(&my_parent);
>>>>> + ORTE_EPOCH_SET(my_parent.epoch,orte_ess.proc_get_epoch(&my_parent));
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((5, orte_grpcomm_base.output,
>>>>> "%s grpcomm:base:daemon_coll: daemon collective not the HNP - sending to parent %s",
>>>>>
>>>>> Modified: trunk/orte/mca/grpcomm/hier/grpcomm_hier_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/grpcomm/hier/grpcomm_hier_module.c (original)
>>>>> +++ trunk/orte/mca/grpcomm/hier/grpcomm_hier_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -95,7 +95,7 @@
>>>>>
>>>>> my_local_rank_zero_proc.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> my_local_rank_zero_proc.vpid = ORTE_VPID_INVALID;
>>>>> - my_local_rank_zero_proc.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(my_local_rank_zero_proc.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> if (ORTE_SUCCESS != (rc = orte_grpcomm_base_modex_init())) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> @@ -270,7 +270,7 @@
>>>>> proc.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> for (v=0; v < orte_process_info.num_procs; v++) {
>>>>> proc.vpid = v;
>>>>> - proc.epoch = orte_util_lookup_epoch(&proc);
>>>>> + ORTE_EPOCH_SET(proc.epoch,orte_util_lookup_epoch(&proc));
>>>>>
>>>>> /* is this proc local_rank=0 on its node? */
>>>>> if (0 == my_local_rank && 0 == orte_ess.get_local_rank(&proc)) {
>>>>> @@ -285,7 +285,7 @@
>>>>> nm = OBJ_NEW(orte_namelist_t);
>>>>> nm->name.jobid = proc.jobid;
>>>>> nm->name.vpid = proc.vpid;
>>>>> - nm->name.epoch = proc.epoch;
>>>>> + ORTE_EPOCH_SET(nm->name.epoch,proc.epoch);
>>>>>
>>>>> opal_list_append(&my_local_peers, &nm->item);
>>>>> /* if I am not local_rank=0, is this one? */
>>>>> @@ -293,7 +293,7 @@
>>>>> 0 == orte_ess.get_local_rank(&proc)) {
>>>>> my_local_rank_zero_proc.jobid = proc.jobid;
>>>>> my_local_rank_zero_proc.vpid = proc.vpid;
>>>>> - my_local_rank_zero_proc.epoch = proc.epoch;
>>>>> + ORTE_EPOCH_SET(my_local_rank_zero_proc.epoch,proc.epoch);
>>>>> }
>>>>> }
>>>>>
>>>>>
>>>>> Modified: trunk/orte/mca/iof/base/base.h
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/iof/base/base.h (original)
>>>>> +++ trunk/orte/mca/iof/base/base.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -135,7 +135,7 @@
>>>>> ep = OBJ_NEW(orte_iof_sink_t); \
>>>>> ep->name.jobid = (nm)->jobid; \
>>>>> ep->name.vpid = (nm)->vpid; \
>>>>> - ep->name.epoch = (nm)->epoch; \
>>>>> + ORTE_EPOCH_SET(ep->name.epoch,(nm)->epoch); \
>>>>> ep->tag = (tg); \
>>>>> if (0 <= (fid)) { \
>>>>> ep->wev->fd = (fid); \
>>>>> @@ -169,7 +169,7 @@
>>>>> rev = OBJ_NEW(orte_iof_read_event_t); \
>>>>> rev->name.jobid = (nm)->jobid; \
>>>>> rev->name.vpid = (nm)->vpid; \
>>>>> - rev->name.epoch = (nm)->epoch; \
>>>>> + ORTE_EPOCH_SET(rev->name.epoch,(nm)->epoch); \
>>>>> rev->tag = (tg); \
>>>>> rev->fd = (fid); \
>>>>> *(rv) = rev; \
>>>>> @@ -194,7 +194,7 @@
>>>>> ep = OBJ_NEW(orte_iof_sink_t); \
>>>>> ep->name.jobid = (nm)->jobid; \
>>>>> ep->name.vpid = (nm)->vpid; \
>>>>> - ep->name.epoch = (nm)->epoch; \
>>>>> + ORTE_EPOCH_SET(ep->name.epoch,(nm)->epoch); \
>>>>> ep->tag = (tg); \
>>>>> if (0 <= (fid)) { \
>>>>> ep->wev->fd = (fid); \
>>>>> @@ -215,7 +215,7 @@
>>>>> rev = OBJ_NEW(orte_iof_read_event_t); \
>>>>> rev->name.jobid = (nm)->jobid; \
>>>>> rev->name.vpid = (nm)->vpid; \
>>>>> - rev->name.epoch= (nm)->epoch; \
>>>>> + ORTE_EPOCH_SET(rev->name.epoch,(nm)->epoch); \
>>>>> rev->tag = (tg); \
>>>>> *(rv) = rev; \
>>>>> opal_event_set(opal_event_base, \
>>>>>
>>>>> Modified: trunk/orte/mca/iof/base/iof_base_open.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/iof/base/iof_base_open.c (original)
>>>>> +++ trunk/orte/mca/iof/base/iof_base_open.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -91,7 +91,7 @@
>>>>> {
>>>>> ptr->daemon.jobid = ORTE_JOBID_INVALID;
>>>>> ptr->daemon.vpid = ORTE_VPID_INVALID;
>>>>> - ptr->daemon.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ptr->daemon.epoch,ORTE_EPOCH_MIN);
>>>>> ptr->wev = OBJ_NEW(orte_iof_write_event_t);
>>>>> }
>>>>> static void orte_iof_base_sink_destruct(orte_iof_sink_t* ptr)
>>>>>
>>>>> Modified: trunk/orte/mca/iof/hnp/iof_hnp.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/iof/hnp/iof_hnp.c (original)
>>>>> +++ trunk/orte/mca/iof/hnp/iof_hnp.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -186,7 +186,7 @@
>>>>> proct = OBJ_NEW(orte_iof_proc_t);
>>>>> proct->name.jobid = dst_name->jobid;
>>>>> proct->name.vpid = dst_name->vpid;
>>>>> - proct->name.epoch = dst_name->epoch;
>>>>> + ORTE_EPOCH_SET(proct->name.epoch,dst_name->epoch);
>>>>> opal_list_append(&mca_iof_hnp_component.procs, &proct->super);
>>>>> /* see if we are to output to a file */
>>>>> if (NULL != orte_output_filename) {
>>>>> @@ -281,8 +281,7 @@
>>>>> &mca_iof_hnp_component.sinks);
>>>>> sink->daemon.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> sink->daemon.vpid = proc->node->daemon->name.vpid;
>>>>> - sink->daemon.epoch = ORTE_EPOCH_INVALID;
>>>>> - sink->daemon.epoch = orte_ess.proc_get_epoch(&sink->daemon);
>>>>> + ORTE_EPOCH_SET(sink->daemon.epoch,orte_ess.proc_get_epoch(&sink->daemon));
>>>>> }
>>>>> }
>>>>>
>>>>> @@ -389,7 +388,7 @@
>>>>> &mca_iof_hnp_component.sinks);
>>>>> sink->daemon.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> sink->daemon.vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - sink->daemon.epoch = ORTE_PROC_MY_NAME->epoch;
>>>>> + ORTE_EPOCH_SET(sink->daemon.epoch,ORTE_PROC_MY_NAME->epoch);
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/iof/hnp/iof_hnp_receive.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/iof/hnp/iof_hnp_receive.c (original)
>>>>> +++ trunk/orte/mca/iof/hnp/iof_hnp_receive.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -109,21 +109,21 @@
>>>>> NULL, &mca_iof_hnp_component.sinks);
>>>>> sink->daemon.jobid = mev->sender.jobid;
>>>>> sink->daemon.vpid = mev->sender.vpid;
>>>>> - sink->daemon.epoch = mev->sender.epoch;
>>>>> + ORTE_EPOCH_SET(sink->daemon.epoch,mev->sender.epoch);
>>>>> }
>>>>> if (ORTE_IOF_STDERR & stream) {
>>>>> ORTE_IOF_SINK_DEFINE(&sink, &origin, -1, ORTE_IOF_STDERR,
>>>>> NULL, &mca_iof_hnp_component.sinks);
>>>>> sink->daemon.jobid = mev->sender.jobid;
>>>>> sink->daemon.vpid = mev->sender.vpid;
>>>>> - sink->daemon.epoch = mev->sender.epoch;
>>>>> + ORTE_EPOCH_SET(sink->daemon.epoch,mev->sender.epoch);
>>>>> }
>>>>> if (ORTE_IOF_STDDIAG & stream) {
>>>>> ORTE_IOF_SINK_DEFINE(&sink, &origin, -1, ORTE_IOF_STDDIAG,
>>>>> NULL, &mca_iof_hnp_component.sinks);
>>>>> sink->daemon.jobid = mev->sender.jobid;
>>>>> sink->daemon.vpid = mev->sender.vpid;
>>>>> - sink->daemon.epoch = mev->sender.epoch;
>>>>> + ORTE_EPOCH_SET(sink->daemon.epoch,mev->sender.epoch);
>>>>> }
>>>>> goto CLEAN_RETURN;
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/iof/orted/iof_orted.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/iof/orted/iof_orted.c (original)
>>>>> +++ trunk/orte/mca/iof/orted/iof_orted.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -163,7 +163,7 @@
>>>>> proct = OBJ_NEW(orte_iof_proc_t);
>>>>> proct->name.jobid = dst_name->jobid;
>>>>> proct->name.vpid = dst_name->vpid;
>>>>> - proct->name.epoch = dst_name->epoch;
>>>>> + ORTE_EPOCH_SET(proct->name.epoch,dst_name->epoch);
>>>>> opal_list_append(&mca_iof_orted_component.procs, &proct->super);
>>>>> /* see if we are to output to a file */
>>>>> if (NULL != orte_output_filename) {
>>>>>
>>>>> Modified: trunk/orte/mca/odls/base/odls_base_default_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/odls/base/odls_base_default_fns.c (original)
>>>>> +++ trunk/orte/mca/odls/base/odls_base_default_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -734,8 +734,7 @@
>>>>> proc.jobid = jobdat->jobid;
>>>>> for (j=0; j < jobdat->num_procs; j++) {
>>>>> proc.vpid = j;
>>>>> - proc.epoch = ORTE_EPOCH_INVALID;
>>>>> - proc.epoch = orte_ess.proc_get_epoch(&proc);
>>>>> + ORTE_EPOCH_SET(proc.epoch,orte_ess.proc_get_epoch(&proc));
>>>>> /* get the vpid of the daemon that is to host this proc */
>>>>> if (ORTE_VPID_INVALID == (host_daemon = orte_ess.proc_get_daemon(&proc))) {
>>>>> ORTE_ERROR_LOG(ORTE_ERR_NOT_FOUND);
>>>>> @@ -1044,6 +1043,7 @@
>>>>> free(param);
>>>>> free(value);
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* setup the epoch */
>>>>> if (ORTE_SUCCESS != (rc = orte_util_convert_epoch_to_string(&value, child->name->epoch))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> @@ -1057,6 +1057,7 @@
>>>>> opal_setenv(param, value, true, env);
>>>>> free(param);
>>>>> free(value);
>>>>> +#endif
>>>>>
>>>>> /* setup the vpid */
>>>>> if (ORTE_SUCCESS != (rc = orte_util_convert_vpid_to_string(&value, child->name->vpid))) {
>>>>> @@ -2721,7 +2722,7 @@
>>>>> OBJ_CONSTRUCT(&proctmp, orte_proc_t);
>>>>> proctmp.name.jobid = ORTE_JOBID_WILDCARD;
>>>>> proctmp.name.vpid = ORTE_VPID_WILDCARD;
>>>>> - proctmp.name.epoch = ORTE_EPOCH_WILDCARD;
>>>>> + ORTE_EPOCH_SET(proctmp.name.epoch,ORTE_EPOCH_WILDCARD);
>>>>> opal_pointer_array_add(&procarray, &proctmp);
>>>>> procptr = &procarray;
>>>>> do_cleanup = true;
>>>>>
>>>>> Modified: trunk/orte/mca/odls/base/odls_base_open.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/odls/base/odls_base_open.c (original)
>>>>> +++ trunk/orte/mca/odls/base/odls_base_open.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -187,7 +187,7 @@
>>>>> if (-1 == rank) {
>>>>> /* wildcard */
>>>>> nm->name.vpid = ORTE_VPID_WILDCARD;
>>>>> - nm->name.epoch = ORTE_EPOCH_WILDCARD;
>>>>> + ORTE_EPOCH_SET(nm->name.epoch,ORTE_EPOCH_WILDCARD);
>>>>> } else if (rank < 0) {
>>>>> /* error out on bozo case */
>>>>> orte_show_help("help-odls-base.txt",
>>>>> @@ -200,8 +200,7 @@
>>>>> * will be in the job - we'll check later
>>>>> */
>>>>> nm->name.vpid = rank;
>>>>> - nm->name.epoch = ORTE_EPOCH_INVALID;
>>>>> - nm->name.epoch = orte_ess.proc_get_epoch(&nm->name);
>>>>> + ORTE_EPOCH_SET(nm->name.epoch,orte_ess.proc_get_epoch(&nm->name));
>>>>> }
>>>>> opal_list_append(&orte_odls_globals.xterm_ranks, &nm->item);
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/odls/base/odls_base_state.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/odls/base/odls_base_state.c (original)
>>>>> +++ trunk/orte/mca/odls/base/odls_base_state.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -77,17 +77,17 @@
>>>>> /* if I am the HNP, then use me as the source */
>>>>> p_set->source.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> p_set->source.vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - p_set->source.epoch = ORTE_PROC_MY_NAME->epoch;
>>>>> + ORTE_EPOCH_SET(p_set->source.epoch,ORTE_PROC_MY_NAME->epoch);
>>>>> }
>>>>> else {
>>>>> /* otherwise, set the HNP as the source */
>>>>> p_set->source.jobid = ORTE_PROC_MY_HNP->jobid;
>>>>> p_set->source.vpid = ORTE_PROC_MY_HNP->vpid;
>>>>> - p_set->source.epoch = ORTE_PROC_MY_HNP->epoch;
>>>>> + ORTE_EPOCH_SET(p_set->source.epoch,ORTE_PROC_MY_HNP->epoch);
>>>>> }
>>>>> p_set->sink.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> p_set->sink.vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - p_set->sink.epoch = ORTE_PROC_MY_NAME->epoch;
>>>>> + ORTE_EPOCH_SET(p_set->sink.epoch,ORTE_PROC_MY_NAME->epoch);
>>>>>
>>>>> opal_list_append(&(filem_request->process_sets), &(p_set->super) );
>>>>>
>>>>>
>>>>> Modified: trunk/orte/mca/oob/tcp/oob_tcp_msg.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/oob/tcp/oob_tcp_msg.c (original)
>>>>> +++ trunk/orte/mca/oob/tcp/oob_tcp_msg.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -137,6 +137,7 @@
>>>>> bool mca_oob_tcp_msg_send_handler(mca_oob_tcp_msg_t* msg, struct mca_oob_tcp_peer_t * peer)
>>>>> {
>>>>> int rc;
>>>>> +
>>>>> while(1) {
>>>>> rc = writev(peer->peer_sd, msg->msg_rwptr, msg->msg_rwnum);
>>>>> if(rc < 0) {
>>>>> @@ -338,6 +339,7 @@
>>>>> orte_process_name_t src = msg->msg_hdr.msg_src;
>>>>>
>>>>> OPAL_THREAD_LOCK(&mca_oob_tcp_component.tcp_lock);
>>>>> +
>>>>> if (orte_util_compare_name_fields(ORTE_NS_CMP_ALL, &peer->peer_name, &src) != OPAL_EQUAL) {
>>>>> opal_hash_table_remove_value_uint64(&mca_oob_tcp_component.tcp_peers,
>>>>> orte_util_hash_name(&peer->peer_name));
>>>>>
>>>>> Modified: trunk/orte/mca/oob/tcp/oob_tcp_peer.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/oob/tcp/oob_tcp_peer.c (original)
>>>>> +++ trunk/orte/mca/oob/tcp/oob_tcp_peer.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -903,6 +903,11 @@
>>>>> static void mca_oob_tcp_peer_recv_handler(int sd, short flags, void* user)
>>>>> {
>>>>> mca_oob_tcp_peer_t* peer = (mca_oob_tcp_peer_t *)user;
>>>>> +
>>>>> + if (orte_abnormal_term_ordered) {
>>>>> + return;
>>>>> + }
>>>>> +
>>>>> OPAL_THREAD_LOCK(&peer->peer_lock);
>>>>> switch(peer->peer_state) {
>>>>> case MCA_OOB_TCP_CONNECT_ACK:
>>>>>
>>>>> Modified: trunk/orte/mca/plm/base/plm_base_jobid.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/plm/base/plm_base_jobid.c (original)
>>>>> +++ trunk/orte/mca/plm/base/plm_base_jobid.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -62,12 +62,12 @@
>>>>> /* set the name */
>>>>> ORTE_PROC_MY_NAME->jobid = 0xffff0000 & ((uint32_t)jobfam << 16);
>>>>> ORTE_PROC_MY_NAME->vpid = 0;
>>>>> - ORTE_PROC_MY_NAME->epoch= ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_NAME->epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> /* copy it to the HNP field */
>>>>> ORTE_PROC_MY_HNP->jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> ORTE_PROC_MY_HNP->vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - ORTE_PROC_MY_HNP->epoch = ORTE_PROC_MY_NAME->epoch;
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_HNP->epoch,ORTE_PROC_MY_NAME->epoch);
>>>>>
>>>>> /* done */
>>>>> return ORTE_SUCCESS;
>>>>>
>>>>> Modified: trunk/orte/mca/plm/base/plm_base_launch_support.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/plm/base/plm_base_launch_support.c (original)
>>>>> +++ trunk/orte/mca/plm/base/plm_base_launch_support.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -377,8 +377,7 @@
>>>>> /* push stdin - the IOF will know what to do with the specified target */
>>>>> name.jobid = job;
>>>>> name.vpid = jdata->stdin_target;
>>>>> - name.epoch = ORTE_EPOCH_INVALID;
>>>>> - name.epoch = orte_ess.proc_get_epoch(&name);
>>>>> + ORTE_EPOCH_SET(name.epoch,orte_ess.proc_get_epoch(&name));
>>>>>
>>>>> if (ORTE_SUCCESS != (rc = orte_iof.push(&name, ORTE_IOF_STDIN, 0))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>>
>>>>> Modified: trunk/orte/mca/plm/base/plm_base_orted_cmds.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/plm/base/plm_base_orted_cmds.c (original)
>>>>> +++ trunk/orte/mca/plm/base/plm_base_orted_cmds.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -163,8 +163,7 @@
>>>>> continue;
>>>>> }
>>>>> peer.vpid = v;
>>>>> - peer.epoch = ORTE_EPOCH_INVALID;
>>>>> - peer.epoch = orte_ess.proc_get_epoch(&peer);
>>>>> + ORTE_EPOCH_SET(peer.epoch,orte_ess.proc_get_epoch(&peer));
>>>>>
>>>>> /* don't worry about errors on the send here - just
>>>>> * issue it and keep going
>>>>> @@ -242,7 +241,7 @@
>>>>> OBJ_CONSTRUCT(&proc, orte_proc_t);
>>>>> proc.name.jobid = jobid;
>>>>> proc.name.vpid = ORTE_VPID_WILDCARD;
>>>>> - proc.name.epoch = ORTE_EPOCH_WILDCARD;
>>>>> + ORTE_EPOCH_SET(proc.name.epoch,ORTE_EPOCH_WILDCARD);
>>>>> opal_pointer_array_add(&procs, &proc);
>>>>> if (ORTE_SUCCESS != (rc = orte_plm_base_orted_kill_local_procs(&procs))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> @@ -340,8 +339,7 @@
>>>>> continue;
>>>>> }
>>>>> peer.vpid = v;
>>>>> - peer.epoch = ORTE_EPOCH_INVALID;
>>>>> - peer.epoch = orte_ess.proc_get_epoch(&peer);
>>>>> + ORTE_EPOCH_SET(peer.epoch,orte_ess.proc_get_epoch(&peer));
>>>>> /* check to see if this daemon is known to be "dead" */
>>>>> if (proc->state > ORTE_PROC_STATE_UNTERMINATED) {
>>>>> /* don't try to send this */
>>>>>
>>>>> Modified: trunk/orte/mca/plm/base/plm_base_receive.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/plm/base/plm_base_receive.c (original)
>>>>> +++ trunk/orte/mca/plm/base/plm_base_receive.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -146,7 +146,9 @@
>>>>> orte_job_t *jdata, *parent;
>>>>> opal_buffer_t answer;
>>>>> orte_vpid_t vpid;
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> orte_epoch_t epoch;
>>>>> +#endif
>>>>> orte_proc_t *proc;
>>>>> orte_proc_state_t state;
>>>>> orte_exit_code_t exit_code;
>>>>> @@ -394,8 +396,7 @@
>>>>> break;
>>>>> }
>>>>> name.vpid = vpid;
>>>>> - name.epoch = ORTE_EPOCH_INVALID;
>>>>> - name.epoch = orte_ess.proc_get_epoch(&name);
>>>>> + ORTE_EPOCH_SET(name.epoch,orte_ess.proc_get_epoch(&name));
>>>>>
>>>>> /* unpack the pid */
>>>>> count = 1;
>>>>> @@ -488,9 +489,11 @@
>>>>> }
>>>>> name.vpid = vpid;
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> count=1;
>>>>> opal_dss.unpack(msgpkt->buffer, &epoch, &count, ORTE_EPOCH);
>>>>> name.epoch = epoch;
>>>>> +#endif
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((5, orte_plm_globals.output,
>>>>> "%s plm:base:receive Described rank %s",
>>>>>
>>>>> Modified: trunk/orte/mca/plm/base/plm_base_rsh_support.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/plm/base/plm_base_rsh_support.c (original)
>>>>> +++ trunk/orte/mca/plm/base/plm_base_rsh_support.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -1527,7 +1527,9 @@
>>>>> {
>>>>> char *param, *path, *tmp, *cmd, *basename, *dest_dir;
>>>>> int i;
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> orte_epoch_t epoch;
>>>>> +#endif
>>>>> orte_process_name_t proc;
>>>>>
>>>>> /* if a prefix is set, pass it to the bootproxy in a special way */
>>>>> @@ -1638,6 +1640,7 @@
>>>>> opal_setenv("OMPI_COMM_WORLD_RANK", cmd, true, argv);
>>>>> free(cmd);
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* set the epoch */
>>>>> proc.jobid = jobid;
>>>>> proc.vpid = vpid;
>>>>> @@ -1648,6 +1651,7 @@
>>>>> opal_setenv(param, cmd, true, argv);
>>>>> free(param);
>>>>> free(cmd);
>>>>> +#endif
>>>>>
>>>>> /* set the number of procs */
>>>>> asprintf(&cmd, "%d", (int)num_procs);
>>>>>
>>>>> Modified: trunk/orte/mca/rmaps/base/rmaps_base_support_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/rmaps/base/rmaps_base_support_fns.c (original)
>>>>> +++ trunk/orte/mca/rmaps/base/rmaps_base_support_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -33,12 +33,14 @@
>>>>> #include "orte/mca/ess/ess.h"
>>>>> #include "opal/mca/sysinfo/sysinfo_types.h"
>>>>>
>>>>> +#include "orte/types.h"
>>>>> #include "orte/util/show_help.h"
>>>>> #include "orte/util/name_fns.h"
>>>>> #include "orte/runtime/orte_globals.h"
>>>>> #include "orte/util/hostfile/hostfile.h"
>>>>> #include "orte/util/dash_host/dash_host.h"
>>>>> #include "orte/mca/errmgr/errmgr.h"
>>>>> +#include "orte/runtime/data_type_support/orte_dt_support.h"
>>>>>
>>>>> #include "orte/mca/rmaps/base/rmaps_private.h"
>>>>> #include "orte/mca/rmaps/base/base.h"
>>>>> @@ -454,7 +456,7 @@
>>>>> */
>>>>>
>>>>> /* We do set the epoch here since they all start with the same value. */
>>>>> - proc->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> proc->app_idx = app_idx;
>>>>> OPAL_OUTPUT_VERBOSE((5, orte_rmaps_base.rmaps_output,
>>>>> @@ -559,11 +561,12 @@
>>>>> }
>>>>> }
>>>>> proc->name.vpid = vpid;
>>>>> - proc->name.epoch = ORTE_EPOCH_INVALID;
>>>>> - proc->name.epoch = orte_ess.proc_get_epoch(&proc->name);
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,ORTE_EPOCH_INVALID);
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,orte_ess.proc_get_epoch(&proc->name));
>>>>> +
>>>>> /* If there is an invalid epoch here, it's because it doesn't exist yet. */
>>>>> - if (ORTE_NODE_RANK_INVALID == proc->name.epoch) {
>>>>> - proc->name.epoch = ORTE_EPOCH_MIN;
>>>>> + if (0 == ORTE_EPOCH_CMP(ORTE_EPOCH_INVALID,proc->name.epoch)) {
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,ORTE_EPOCH_MIN);
>>>>> }
>>>>> }
>>>>> if (NULL == opal_pointer_array_get_item(jdata->procs, proc->name.vpid)) {
>>>>> @@ -601,8 +604,8 @@
>>>>> }
>>>>> }
>>>>> proc->name.vpid = vpid;
>>>>> - proc->name.epoch = ORTE_EPOCH_INVALID;
>>>>> - proc->name.epoch = orte_ess.proc_get_epoch(&proc->name);
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,ORTE_EPOCH_INVALID);
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,orte_ess.proc_get_epoch(&proc->name));
>>>>> }
>>>>> if (NULL == opal_pointer_array_get_item(jdata->procs, proc->name.vpid)) {
>>>>> if (ORTE_SUCCESS != (rc = opal_pointer_array_set_item(jdata->procs, proc->name.vpid, proc))) {
>>>>> @@ -835,7 +838,7 @@
>>>>> return ORTE_ERR_OUT_OF_RESOURCE;
>>>>> }
>>>>> proc->name.vpid = daemons->num_procs; /* take the next available vpid */
>>>>> - proc->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,ORTE_EPOCH_MIN);
>>>>> proc->node = node;
>>>>> proc->nodename = node->name;
>>>>> OPAL_OUTPUT_VERBOSE((5, orte_rmaps_base.rmaps_output,
>>>>> @@ -1014,8 +1017,8 @@
>>>>> return ORTE_ERR_OUT_OF_RESOURCE;
>>>>> }
>>>>> proc->name.vpid = jdata->num_procs; /* take the next available vpid */
>>>>> - proc->name.epoch = ORTE_EPOCH_INVALID;
>>>>> - proc->name.epoch = orte_ess.proc_get_epoch(&proc->name);
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,ORTE_EPOCH_INVALID);
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,orte_ess.proc_get_epoch(&proc->name));
>>>>> proc->node = node;
>>>>> proc->nodename = node->name;
>>>>> OPAL_OUTPUT_VERBOSE((5, orte_rmaps_base.rmaps_output,
>>>>>
>>>>> Modified: trunk/orte/mca/rmaps/rank_file/rmaps_rank_file.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/rmaps/rank_file/rmaps_rank_file.c (original)
>>>>> +++ trunk/orte/mca/rmaps/rank_file/rmaps_rank_file.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -502,8 +502,7 @@
>>>>> }
>>>>> proc->name.vpid = rank;
>>>>> /* Either init or update the epoch. */
>>>>> - proc->name.epoch = ORTE_EPOCH_INVALID;
>>>>> - proc->name.epoch = orte_ess.proc_get_epoch(&proc->name);
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,orte_ess.proc_get_epoch(&proc->name));
>>>>>
>>>>> proc->slot_list = strdup(rfmap->slot_list);
>>>>> /* insert the proc into the proper place */
>>>>>
>>>>> Modified: trunk/orte/mca/rmaps/seq/rmaps_seq.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/rmaps/seq/rmaps_seq.c (original)
>>>>> +++ trunk/orte/mca/rmaps/seq/rmaps_seq.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -235,8 +235,7 @@
>>>>> }
>>>>> /* assign the vpid */
>>>>> proc->name.vpid = vpid++;
>>>>> - proc->name.epoch = ORTE_EPOCH_INVALID;
>>>>> - proc->name.epoch = orte_ess.proc_get_epoch(&proc->name);
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,orte_ess.proc_get_epoch(&proc->name));
>>>>>
>>>>> /* add to the jdata proc array */
>>>>> if (ORTE_SUCCESS != (rc = opal_pointer_array_set_item(jdata->procs, proc->name.vpid, proc))) {
>>>>>
>>>>> Modified: trunk/orte/mca/rmcast/base/rmcast_base_open.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/rmcast/base/rmcast_base_open.c (original)
>>>>> +++ trunk/orte/mca/rmcast/base/rmcast_base_open.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -341,7 +341,7 @@
>>>>> {
>>>>> ptr->name.jobid = ORTE_JOBID_INVALID;
>>>>> ptr->name.vpid = ORTE_VPID_INVALID;
>>>>> - ptr->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ptr->name.epoch,ORTE_EPOCH_MIN);
>>>>> ptr->channel = ORTE_RMCAST_INVALID_CHANNEL;
>>>>> OBJ_CONSTRUCT(&ptr->ctl, orte_thread_ctl_t);
>>>>> ptr->seq_num = ORTE_RMCAST_SEQ_INVALID;
>>>>> @@ -430,7 +430,7 @@
>>>>> {
>>>>> ptr->name.jobid = ORTE_JOBID_INVALID;
>>>>> ptr->name.vpid = ORTE_VPID_INVALID;
>>>>> - ptr->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ptr->name.epoch,ORTE_EPOCH_MIN);
>>>>> OBJ_CONSTRUCT(&ptr->last_msg, opal_list_t);
>>>>> }
>>>>> static void recvlog_destruct(rmcast_recv_log_t *ptr)
>>>>> @@ -439,7 +439,7 @@
>>>>>
>>>>> ptr->name.jobid = ORTE_JOBID_INVALID;
>>>>> ptr->name.vpid = ORTE_VPID_INVALID;
>>>>> - ptr->name.epoch = ORTE_EPOCH_INVALID;
>>>>> + ORTE_EPOCH_SET(ptr->name.epoch,ORTE_EPOCH_INVALID);
>>>>> while (NULL != (item = opal_list_remove_first(&ptr->last_msg))) {
>>>>> OBJ_RELEASE(item);
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/rmcast/tcp/rmcast_tcp.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/rmcast/tcp/rmcast_tcp.c (original)
>>>>> +++ trunk/orte/mca/rmcast/tcp/rmcast_tcp.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -681,7 +681,7 @@
>>>>> /* caller requested id of sender */
>>>>> name->jobid = recvptr->name.jobid;
>>>>> name->vpid = recvptr->name.vpid;
>>>>> - name->epoch= recvptr->name.epoch;
>>>>> + ORTE_EPOCH_SET(name->epoch,recvptr->name.epoch);
>>>>> }
>>>>> *seq_num = recvptr->seq_num;
>>>>> *msg = recvptr->iovec_array;
>>>>> @@ -776,7 +776,7 @@
>>>>> /* caller requested id of sender */
>>>>> name->jobid = recvptr->name.jobid;
>>>>> name->vpid = recvptr->name.vpid;
>>>>> - name->epoch= recvptr->name.epoch;
>>>>> + ORTE_EPOCH_SET(name->epoch,recvptr->name.epoch);
>>>>> }
>>>>> *seq_num = recvptr->seq_num;
>>>>> if (ORTE_SUCCESS != (ret = opal_dss.copy_payload(buf, recvptr->buf))) {
>>>>>
>>>>> Modified: trunk/orte/mca/rmcast/udp/rmcast_udp.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/rmcast/udp/rmcast_udp.c (original)
>>>>> +++ trunk/orte/mca/rmcast/udp/rmcast_udp.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -460,7 +460,7 @@
>>>>> /* caller requested id of sender */
>>>>> name->jobid = recvptr->name.jobid;
>>>>> name->vpid = recvptr->name.vpid;
>>>>> - name->epoch= recvptr->name.epoch;
>>>>> + ORTE_EPOCH_SET(name->epoch,recvptr->name.epoch);
>>>>> }
>>>>> *seq_num = recvptr->seq_num;
>>>>> *msg = recvptr->iovec_array;
>>>>> @@ -553,7 +553,7 @@
>>>>> /* caller requested id of sender */
>>>>> name->jobid = recvptr->name.jobid;
>>>>> name->vpid = recvptr->name.vpid;
>>>>> - name->epoch= recvptr->name.epoch;
>>>>> + ORTE_EPOCH_SET(name->epoch,recvptr->name.epoch);
>>>>> }
>>>>> *seq_num = recvptr->seq_num;
>>>>> if (ORTE_SUCCESS != (ret = opal_dss.copy_payload(buf, recvptr->buf))) {
>>>>>
>>>>> Modified: trunk/orte/mca/rml/base/rml_base_components.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/rml/base/rml_base_components.c (original)
>>>>> +++ trunk/orte/mca/rml/base/rml_base_components.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -20,6 +20,7 @@
>>>>> #include "opal/util/output.h"
>>>>>
>>>>> #include "orte/mca/rml/rml.h"
>>>>> +#include "orte/util/name_fns.h"
>>>>>
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>>
>>>>> @@ -67,14 +68,14 @@
>>>>> {
>>>>> pkt->sender.jobid = ORTE_JOBID_INVALID;
>>>>> pkt->sender.vpid = ORTE_VPID_INVALID;
>>>>> - pkt->sender.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(pkt->sender.epoch,ORTE_EPOCH_MIN);
>>>>> pkt->buffer = NULL;
>>>>> }
>>>>> static void msg_pkt_destructor(orte_msg_packet_t *pkt)
>>>>> {
>>>>> pkt->sender.jobid = ORTE_JOBID_INVALID;
>>>>> pkt->sender.vpid = ORTE_VPID_INVALID;
>>>>> - pkt->sender.epoch = ORTE_EPOCH_INVALID;
>>>>> + ORTE_EPOCH_SET(pkt->sender.epoch,ORTE_EPOCH_INVALID);
>>>>> if (NULL != pkt->buffer) {
>>>>> OBJ_RELEASE(pkt->buffer);
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/rml/rml_types.h
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/rml/rml_types.h (original)
>>>>> +++ trunk/orte/mca/rml/rml_types.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -62,7 +62,7 @@
>>>>> pkt = OBJ_NEW(orte_msg_packet_t); \
>>>>> pkt->sender.jobid = (sndr)->jobid; \
>>>>> pkt->sender.vpid = (sndr)->vpid; \
>>>>> - pkt->sender.epoch = (sndr)->epoch; \
>>>>> + ORTE_EPOCH_SET(pkt->sender.epoch,(sndr)->epoch); \
>>>>> if ((crt)) { \
>>>>> pkt->buffer = OBJ_NEW(opal_buffer_t); \
>>>>> opal_dss.copy_payload(pkt->buffer, *(buf)); \
>>>>> @@ -85,7 +85,7 @@
>>>>> pkt = OBJ_NEW(orte_msg_packet_t); \
>>>>> pkt->sender.jobid = (sndr)->jobid; \
>>>>> pkt->sender.vpid = (sndr)->vpid; \
>>>>> - pkt->sender.epoch = (sndr)->epoch; \
>>>>> + ORTE_EPOCH_SET(pkt->sender.epoch,(sndr)->epoch); \
>>>>> if ((crt)) { \
>>>>> pkt->buffer = OBJ_NEW(opal_buffer_t); \
>>>>> opal_dss.copy_payload(pkt->buffer, *(buf)); \
>>>>> @@ -191,8 +191,10 @@
>>>>>
>>>>> #define ORTE_RML_TAG_SUBSCRIBE 46
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* For Epoch Updates */
>>>>> #define ORTE_RML_TAG_EPOCH_CHANGE 47
>>>>> +#endif
>>>>>
>>>>> /* Notify of failed processes */
>>>>> #define ORTE_RML_TAG_FAILURE_NOTICE 48
>>>>>
>>>>> Modified: trunk/orte/mca/routed/base/routed_base_components.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/routed/base/routed_base_components.c (original)
>>>>> +++ trunk/orte/mca/routed/base/routed_base_components.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -65,7 +65,7 @@
>>>>> {
>>>>> ptr->route.jobid = ORTE_JOBID_INVALID;
>>>>> ptr->route.vpid = ORTE_VPID_INVALID;
>>>>> - ptr->route.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ptr->route.epoch,ORTE_EPOCH_MIN);
>>>>> ptr->hnp_uri = NULL;
>>>>> }
>>>>> static void jfamdest(orte_routed_jobfam_t *ptr)
>>>>> @@ -117,7 +117,7 @@
>>>>> jfam = OBJ_NEW(orte_routed_jobfam_t);
>>>>> jfam->route.jobid = ORTE_PROC_MY_HNP->jobid;
>>>>> jfam->route.vpid = ORTE_PROC_MY_HNP->vpid;
>>>>> - jfam->route.epoch = ORTE_PROC_MY_HNP->epoch;
>>>>> + ORTE_EPOCH_SET(jfam->route.epoch,ORTE_PROC_MY_HNP->epoch);
>>>>> jfam->job_family = ORTE_JOB_FAMILY(ORTE_PROC_MY_NAME->jobid);
>>>>> if (NULL != orte_process_info.my_hnp_uri) {
>>>>> jfam->hnp_uri = strdup(orte_process_info.my_hnp_uri);
>>>>> @@ -252,7 +252,7 @@
>>>>> jfam->job_family = jobfamily;
>>>>> jfam->route.jobid = name.jobid;
>>>>> jfam->route.vpid = name.vpid;
>>>>> - jfam->route.epoch = name.epoch;
>>>>> + ORTE_EPOCH_SET(jfam->route.epoch,name.epoch);
>>>>> jfam->hnp_uri = strdup(uri);
>>>>> done:
>>>>> free(uri);
>>>>>
>>>>> Modified: trunk/orte/mca/routed/base/routed_base_register_sync.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/routed/base/routed_base_register_sync.c (original)
>>>>> +++ trunk/orte/mca/routed/base/routed_base_register_sync.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -127,7 +127,9 @@
>>>>> orte_std_cntr_t cnt;
>>>>> char *rml_uri;
>>>>> orte_vpid_t vpid;
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> orte_epoch_t epoch;
>>>>> +#endif
>>>>> int rc;
>>>>>
>>>>> if (ORTE_JOB_FAMILY(job) == ORTE_JOB_FAMILY(ORTE_PROC_MY_NAME->jobid)) {
>>>>> @@ -146,11 +148,13 @@
>>>>> cnt = 1;
>>>>> while (ORTE_SUCCESS == (rc = opal_dss.unpack(buffer, &vpid, &cnt, ORTE_VPID))) {
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> cnt = 1;
>>>>> if (ORTE_SUCCESS != (rc = opal_dss.unpack(buffer, &epoch, &cnt, ORTE_EPOCH))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>> continue;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> if (ORTE_SUCCESS != (rc = opal_dss.unpack(buffer, &rml_uri, &cnt, OPAL_STRING))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>>
>>>>> Modified: trunk/orte/mca/routed/binomial/routed_binomial.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/routed/binomial/routed_binomial.c (original)
>>>>> +++ trunk/orte/mca/routed/binomial/routed_binomial.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -33,6 +33,7 @@
>>>>> #include "orte/runtime/orte_globals.h"
>>>>> #include "orte/runtime/orte_wait.h"
>>>>> #include "orte/runtime/runtime.h"
>>>>> +#include "orte/runtime/data_type_support/orte_dt_support.h"
>>>>>
>>>>> #include "orte/mca/rml/base/rml_contact.h"
>>>>>
>>>>> @@ -147,7 +148,7 @@
>>>>>
>>>>> if (proc->jobid == ORTE_JOBID_INVALID ||
>>>>> proc->vpid == ORTE_VPID_INVALID ||
>>>>> - proc->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(proc->epoch,ORTE_EPOCH_INVALID)) {
>>>>> return ORTE_ERR_BAD_PARAM;
>>>>> }
>>>>>
>>>>> @@ -216,7 +217,7 @@
>>>>>
>>>>> if (target->jobid == ORTE_JOBID_INVALID ||
>>>>> target->vpid == ORTE_VPID_INVALID ||
>>>>> - target->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(target->epoch,ORTE_EPOCH_INVALID)) {
>>>>> return ORTE_ERR_BAD_PARAM;
>>>>> }
>>>>>
>>>>> @@ -274,8 +275,7 @@
>>>>> ORTE_NAME_PRINT(route)));
>>>>> jfam->route.jobid = route->jobid;
>>>>> jfam->route.vpid = route->vpid;
>>>>> - jfam->route.epoch = ORTE_EPOCH_INVALID;
>>>>> - jfam->route.epoch = orte_ess.proc_get_epoch(&jfam->route);
>>>>> + ORTE_EPOCH_SET(jfam->route.epoch,orte_ess.proc_get_epoch(&jfam->route));
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> @@ -290,8 +290,7 @@
>>>>> jfam->job_family = jfamily;
>>>>> jfam->route.jobid = route->jobid;
>>>>> jfam->route.vpid = route->vpid;
>>>>> - jfam->route.epoch = ORTE_EPOCH_INVALID;
>>>>> - jfam->route.epoch = orte_ess.proc_get_epoch(&jfam->route);
>>>>> + ORTE_EPOCH_SET(jfam->route.epoch,orte_ess.proc_get_epoch(&jfam->route));
>>>>>
>>>>> opal_pointer_array_add(&orte_routed_jobfams, jfam);
>>>>> return ORTE_SUCCESS;
>>>>> @@ -317,11 +316,21 @@
>>>>> /* initialize */
>>>>> daemon.jobid = ORTE_PROC_MY_DAEMON->jobid;
>>>>> daemon.vpid = ORTE_PROC_MY_DAEMON->vpid;
>>>>> - daemon.epoch = ORTE_PROC_MY_DAEMON->epoch;
>>>>> + ORTE_EPOCH_SET(daemon.epoch,ORTE_PROC_MY_DAEMON->epoch);
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> if (target->jobid == ORTE_JOBID_INVALID ||
>>>>> target->vpid == ORTE_VPID_INVALID ||
>>>>> target->epoch == ORTE_EPOCH_INVALID) {
>>>>> +#else
>>>>> + if (target->jobid == ORTE_JOBID_INVALID ||
>>>>> + target->vpid == ORTE_VPID_INVALID) {
>>>>> +#endif
>>>>> + ret = ORTE_NAME_INVALID;
>>>>> + goto found;
>>>>> + }
>>>>> +
>>>>> + if (0 > ORTE_EPOCH_CMP(target->epoch, orte_ess.proc_get_epoch(target))) {
>>>>> ret = ORTE_NAME_INVALID;
>>>>> goto found;
>>>>> }
>>>>> @@ -443,7 +452,7 @@
>>>>>
>>>>> /* If the daemon to which we should be routing is dead, then update
>>>>> * the routing tree and start over. */
>>>>> - if (!orte_util_proc_is_running(&daemon)) {
>>>>> + if (!PROC_IS_RUNNING(&daemon)) {
>>>>> update_routing_tree(daemon.jobid);
>>>>> goto startover;
>>>>> }
>>>>> @@ -461,8 +470,7 @@
>>>>> ret = &daemon;
>>>>>
>>>>> found:
>>>>> - daemon.epoch = ORTE_EPOCH_INVALID;
>>>>> - daemon.epoch = orte_ess.proc_get_epoch(&daemon);
>>>>> + ORTE_EPOCH_SET(daemon.epoch,orte_ess.proc_get_epoch(&daemon));
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_routed_base_output,
>>>>> "%s routed_binomial_get(%s) --> %s",
>>>>> @@ -879,7 +887,7 @@
>>>>> */
>>>>> local_lifeline.jobid = proc->jobid;
>>>>> local_lifeline.vpid = proc->vpid;
>>>>> - local_lifeline.epoch = proc->epoch;
>>>>> + ORTE_EPOCH_SET(local_lifeline.epoch,proc->epoch);
>>>>> lifeline = &local_lifeline;
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> @@ -924,11 +932,11 @@
>>>>> * that process so we can check it's state.
>>>>> */
>>>>> proc_name.vpid = peer;
>>>>> - proc_name.epoch = orte_util_lookup_epoch(&proc_name);
>>>>> + ORTE_EPOCH_SET(proc_name.epoch,orte_util_lookup_epoch(&proc_name));
>>>>>
>>>>> - if (!orte_util_proc_is_running(&proc_name)
>>>>> - && ORTE_EPOCH_MIN < proc_name.epoch
>>>>> - && ORTE_EPOCH_INVALID != proc_name.epoch) {
>>>>> + if (!PROC_IS_RUNNING(&proc_name)
>>>>> + && 0 < ORTE_EPOCH_CMP(ORTE_EPOCH_MIN,proc_name.epoch)
>>>>> + && 0 != ORTE_EPOCH_CMP(ORTE_EPOCH_INVALID,proc_name.epoch)) {
>>>>> OPAL_OUTPUT_VERBOSE((3, orte_routed_base_output,
>>>>> "%s routed:binomial child %s is dead",
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
>>>>> @@ -967,7 +975,7 @@
>>>>> }
>>>>>
>>>>> /* find the children of this rank */
>>>>> - OPAL_OUTPUT_VERBOSE((3, orte_routed_base_output,
>>>>> + OPAL_OUTPUT_VERBOSE((5, orte_routed_base_output,
>>>>> "%s routed:binomial find children of rank %d",
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), rank));
>>>>> bitmap = opal_cube_dim(num_procs);
>>>>> @@ -977,24 +985,25 @@
>>>>>
>>>>> for (i = hibit + 1, mask = 1 << i; i <= bitmap; ++i, mask <<= 1) {
>>>>> peer = rank | mask;
>>>>> - OPAL_OUTPUT_VERBOSE((3, orte_routed_base_output,
>>>>> + OPAL_OUTPUT_VERBOSE((5, orte_routed_base_output,
>>>>> "%s routed:binomial find children checking peer %d",
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), peer));
>>>>> if (peer < num_procs) {
>>>>> - OPAL_OUTPUT_VERBOSE((3, orte_routed_base_output,
>>>>> + OPAL_OUTPUT_VERBOSE((5, orte_routed_base_output,
>>>>> "%s routed:binomial find children computing tree",
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME)));
>>>>> /* execute compute on this child */
>>>>> if (0 <= (found = binomial_tree(peer, rank, me, num_procs, nchildren, childrn, relatives, mine, jobid))) {
>>>>> proc_name.vpid = found;
>>>>>
>>>>> - if (!orte_util_proc_is_running(&proc_name) && ORTE_EPOCH_MIN < orte_util_lookup_epoch(&proc_name)) {
>>>>> - OPAL_OUTPUT_VERBOSE((3, orte_routed_base_output,
>>>>> + if (!PROC_IS_RUNNING(&proc_name)
>>>>> + && 0 < ORTE_EPOCH_CMP(ORTE_EPOCH_MIN,orte_util_lookup_epoch(&proc_name))) {
>>>>> + OPAL_OUTPUT_VERBOSE((5, orte_routed_base_output,
>>>>> "%s routed:binomial find children proc out of date - returning parent %d",
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), parent));
>>>>> return parent;
>>>>> }
>>>>> - OPAL_OUTPUT_VERBOSE((3, orte_routed_base_output,
>>>>> + OPAL_OUTPUT_VERBOSE((5, orte_routed_base_output,
>>>>> "%s routed:binomial find children returning found value %d",
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), found));
>>>>> return found;
>>>>> @@ -1029,8 +1038,7 @@
>>>>> ORTE_PROC_MY_PARENT->vpid = binomial_tree(0, 0, ORTE_PROC_MY_NAME->vpid,
>>>>> orte_process_info.max_procs,
>>>>> &num_children, &my_children, NULL, true, jobid);
>>>>> - ORTE_PROC_MY_PARENT->epoch = ORTE_EPOCH_INVALID;
>>>>> - ORTE_PROC_MY_PARENT->epoch = orte_ess.proc_get_epoch(ORTE_PROC_MY_PARENT);
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_PARENT->epoch,orte_ess.proc_get_epoch(ORTE_PROC_MY_PARENT));
>>>>>
>>>>> if (0 < opal_output_get_verbosity(orte_routed_base_output)) {
>>>>> opal_output(0, "%s: parent %d num_children %d", ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), ORTE_PROC_MY_PARENT->vpid, num_children);
>>>>>
>>>>> Modified: trunk/orte/mca/routed/cm/routed_cm.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/routed/cm/routed_cm.c (original)
>>>>> +++ trunk/orte/mca/routed/cm/routed_cm.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -35,6 +35,7 @@
>>>>> #include "orte/runtime/orte_globals.h"
>>>>> #include "orte/runtime/orte_wait.h"
>>>>> #include "orte/runtime/runtime.h"
>>>>> +#include "orte/runtime/data_type_support/orte_dt_support.h"
>>>>>
>>>>> #include "orte/mca/rml/base/rml_contact.h"
>>>>>
>>>>> @@ -139,7 +140,7 @@
>>>>>
>>>>> if (proc->jobid == ORTE_JOBID_INVALID ||
>>>>> proc->vpid == ORTE_VPID_INVALID ||
>>>>> - proc->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(proc->epoch,ORTE_EPOCH_INVALID)) {
>>>>> return ORTE_ERR_BAD_PARAM;
>>>>> }
>>>>>
>>>>> @@ -200,7 +201,7 @@
>>>>>
>>>>> if (target->jobid == ORTE_JOBID_INVALID ||
>>>>> target->vpid == ORTE_VPID_INVALID ||
>>>>> - target->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(target->epoch,ORTE_EPOCH_INVALID)) {
>>>>> return ORTE_ERR_BAD_PARAM;
>>>>> }
>>>>>
>>>>> @@ -257,8 +258,7 @@
>>>>> ORTE_NAME_PRINT(route)));
>>>>> jfam->route.jobid = route->jobid;
>>>>> jfam->route.vpid = route->vpid;
>>>>> - jfam->route.epoch = ORTE_EPOCH_INVALID;
>>>>> - jfam->route.epoch = orte_ess.proc_get_epoch(&jfam->route);
>>>>> + ORTE_EPOCH_SET(jfam->route.epoch,orte_ess.proc_get_epoch(&jfam->route));
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> @@ -273,8 +273,7 @@
>>>>> jfam->job_family = jfamily;
>>>>> jfam->route.jobid = route->jobid;
>>>>> jfam->route.vpid = route->vpid;
>>>>> - jfam->route.epoch = ORTE_EPOCH_INVALID;
>>>>> - jfam->route.epoch = orte_ess.proc_get_epoch(&jfam->route);
>>>>> + ORTE_EPOCH_SET(jfam->route.epoch,orte_ess.proc_get_epoch(&jfam->route));
>>>>>
>>>>> opal_pointer_array_add(&orte_routed_jobfams, jfam);
>>>>> return ORTE_SUCCESS;
>>>>> @@ -299,7 +298,7 @@
>>>>>
>>>>> if (target->jobid == ORTE_JOBID_INVALID ||
>>>>> target->vpid == ORTE_VPID_INVALID ||
>>>>> - target->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(target->epoch,ORTE_EPOCH_INVALID)) {
>>>>> ret = ORTE_NAME_INVALID;
>>>>> goto found;
>>>>> }
>>>>> @@ -367,8 +366,7 @@
>>>>> }
>>>>>
>>>>> /* Initialize daemon's epoch, based on its current vpid/jobid */
>>>>> - daemon.epoch = ORTE_EPOCH_INVALID;
>>>>> - daemon.epoch = orte_ess.proc_get_epoch(&daemon);
>>>>> + ORTE_EPOCH_SET(daemon.epoch,orte_ess.proc_get_epoch(&daemon));
>>>>>
>>>>> /* if the daemon is me, then send direct to the target! */
>>>>> if (ORTE_PROC_MY_NAME->vpid == daemon.vpid) {
>>>>> @@ -814,8 +812,7 @@
>>>>> */
>>>>> local_lifeline.jobid = proc->jobid;
>>>>> local_lifeline.vpid = proc->vpid;
>>>>> - local_lifeline.epoch = ORTE_EPOCH_INVALID;
>>>>> - local_lifeline.epoch = orte_ess.proc_get_epoch(&local_lifeline);
>>>>> + ORTE_EPOCH_SET(local_lifeline.epoch,orte_ess.proc_get_epoch(&local_lifeline));
>>>>>
>>>>> lifeline = &local_lifeline;
>>>>>
>>>>>
>>>>> Modified: trunk/orte/mca/routed/direct/routed_direct.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/routed/direct/routed_direct.c (original)
>>>>> +++ trunk/orte/mca/routed/direct/routed_direct.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -24,6 +24,7 @@
>>>>> #include "orte/util/name_fns.h"
>>>>> #include "orte/util/proc_info.h"
>>>>> #include "orte/runtime/orte_globals.h"
>>>>> +#include "orte/runtime/data_type_support/orte_dt_support.h"
>>>>>
>>>>> #include "orte/mca/rml/base/rml_contact.h"
>>>>>
>>>>> @@ -135,7 +136,7 @@
>>>>>
>>>>> if (target->jobid == ORTE_JOBID_INVALID ||
>>>>> target->vpid == ORTE_VPID_INVALID ||
>>>>> - target->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(target->epoch,ORTE_EPOCH_INVALID)) {
>>>>> ret = ORTE_NAME_INVALID;
>>>>> } else {
>>>>> /* all routes are direct */
>>>>>
>>>>> Modified: trunk/orte/mca/routed/linear/routed_linear.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/routed/linear/routed_linear.c (original)
>>>>> +++ trunk/orte/mca/routed/linear/routed_linear.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -31,6 +31,7 @@
>>>>> #include "orte/runtime/orte_globals.h"
>>>>> #include "orte/runtime/orte_wait.h"
>>>>> #include "orte/runtime/runtime.h"
>>>>> +#include "orte/runtime/data_type_support/orte_dt_support.h"
>>>>>
>>>>> #include "orte/mca/rml/base/rml_contact.h"
>>>>>
>>>>> @@ -132,7 +133,7 @@
>>>>>
>>>>> if (proc->jobid == ORTE_JOBID_INVALID ||
>>>>> proc->vpid == ORTE_VPID_INVALID ||
>>>>> - proc->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(proc->epoch,ORTE_EPOCH_INVALID)) {
>>>>> return ORTE_ERR_BAD_PARAM;
>>>>> }
>>>>>
>>>>> @@ -201,7 +202,7 @@
>>>>>
>>>>> if (target->jobid == ORTE_JOBID_INVALID ||
>>>>> target->vpid == ORTE_VPID_INVALID ||
>>>>> - target->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(target->epoch,ORTE_EPOCH_INVALID)) {
>>>>> return ORTE_ERR_BAD_PARAM;
>>>>> }
>>>>>
>>>>> @@ -259,7 +260,7 @@
>>>>> ORTE_NAME_PRINT(route)));
>>>>> jfam->route.jobid = route->jobid;
>>>>> jfam->route.vpid = route->vpid;
>>>>> - jfam->route.epoch = route->epoch;
>>>>> + ORTE_EPOCH_SET(jfam->route.epoch,route->epoch);
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> }
>>>>> @@ -273,7 +274,7 @@
>>>>> jfam->job_family = jfamily;
>>>>> jfam->route.jobid = route->jobid;
>>>>> jfam->route.vpid = route->vpid;
>>>>> - jfam->route.epoch = route->epoch;
>>>>> + ORTE_EPOCH_SET(jfam->route.epoch,route->epoch);
>>>>> opal_pointer_array_add(&orte_routed_jobfams, jfam);
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> @@ -373,8 +374,7 @@
>>>>> }
>>>>>
>>>>> /* Initialize daemon's epoch, based on its current vpid/jobid */
>>>>> - daemon.epoch = ORTE_EPOCH_INVALID;
>>>>> - daemon.epoch = orte_ess.proc_get_epoch(&daemon);
>>>>> + ORTE_EPOCH_SET(daemon.epoch,orte_ess.proc_get_epoch(&daemon));
>>>>>
>>>>> /* if the daemon is me, then send direct to the target! */
>>>>> if (ORTE_PROC_MY_NAME->vpid == daemon.vpid) {
>>>>> @@ -395,8 +395,7 @@
>>>>> /* we are at end of chain - wrap around */
>>>>> daemon.vpid = 0;
>>>>> }
>>>>> - daemon.epoch = ORTE_EPOCH_INVALID;
>>>>> - daemon.epoch = orte_ess.proc_get_epoch(&daemon);
>>>>> + ORTE_EPOCH_SET(daemon.epoch,orte_ess.proc_get_epoch(&daemon));
>>>>> ret = &daemon;
>>>>> }
>>>>> }
>>>>> @@ -741,7 +740,7 @@
>>>>> */
>>>>> local_lifeline.jobid = proc->jobid;
>>>>> local_lifeline.vpid = proc->vpid;
>>>>> - local_lifeline.epoch = proc->epoch;
>>>>> + ORTE_EPOCH_SET(local_lifeline.epoch,proc->epoch);
>>>>> lifeline = &local_lifeline;
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>>
>>>>> Modified: trunk/orte/mca/routed/radix/routed_radix.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/routed/radix/routed_radix.c (original)
>>>>> +++ trunk/orte/mca/routed/radix/routed_radix.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -31,6 +31,7 @@
>>>>> #include "orte/runtime/orte_globals.h"
>>>>> #include "orte/runtime/orte_wait.h"
>>>>> #include "orte/runtime/runtime.h"
>>>>> +#include "orte/runtime/data_type_support/orte_dt_support.h"
>>>>>
>>>>> #include "orte/mca/rml/base/rml_contact.h"
>>>>>
>>>>> @@ -145,7 +146,7 @@
>>>>>
>>>>> if (proc->jobid == ORTE_JOBID_INVALID ||
>>>>> proc->vpid == ORTE_VPID_INVALID ||
>>>>> - proc->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(proc->epoch,ORTE_EPOCH_INVALID)) {
>>>>> return ORTE_ERR_BAD_PARAM;
>>>>> }
>>>>>
>>>>> @@ -214,7 +215,7 @@
>>>>>
>>>>> if (target->jobid == ORTE_JOBID_INVALID ||
>>>>> target->vpid == ORTE_VPID_INVALID ||
>>>>> - target->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(target->epoch,ORTE_EPOCH_INVALID)) {
>>>>> return ORTE_ERR_BAD_PARAM;
>>>>> }
>>>>>
>>>>> @@ -272,7 +273,7 @@
>>>>> ORTE_NAME_PRINT(route)));
>>>>> jfam->route.jobid = route->jobid;
>>>>> jfam->route.vpid = route->vpid;
>>>>> - jfam->route.epoch = route->epoch;
>>>>> + ORTE_EPOCH_SET(jfam->route.epoch,route->epoch);
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> }
>>>>> @@ -286,7 +287,7 @@
>>>>> jfam->job_family = jfamily;
>>>>> jfam->route.jobid = route->jobid;
>>>>> jfam->route.vpid = route->vpid;
>>>>> - jfam->route.epoch = route->epoch;
>>>>> + ORTE_EPOCH_SET(jfam->route.epoch,route->epoch);
>>>>> opal_pointer_array_add(&orte_routed_jobfams, jfam);
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> @@ -310,7 +311,7 @@
>>>>>
>>>>> if (target->jobid == ORTE_JOBID_INVALID ||
>>>>> target->vpid == ORTE_VPID_INVALID ||
>>>>> - target->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(target->epoch,ORTE_EPOCH_INVALID)) {
>>>>> ret = ORTE_NAME_INVALID;
>>>>> goto found;
>>>>> }
>>>>> @@ -413,8 +414,7 @@
>>>>> if (opal_bitmap_is_set_bit(&child->relatives, daemon.vpid)) {
>>>>> /* yep - we need to step through this child */
>>>>> daemon.vpid = child->vpid;
>>>>> - daemon.epoch = ORTE_EPOCH_INVALID;
>>>>> - daemon.epoch = orte_ess.proc_get_epoch(&daemon);
>>>>> + ORTE_EPOCH_SET(daemon.epoch,orte_ess.proc_get_epoch(&daemon));
>>>>> ret = &daemon;
>>>>> goto found;
>>>>> }
>>>>> @@ -425,8 +425,7 @@
>>>>> * any of our children, so we have to step up through our parent
>>>>> */
>>>>> daemon.vpid = ORTE_PROC_MY_PARENT->vpid;
>>>>> - daemon.epoch = ORTE_EPOCH_INVALID;
>>>>> - daemon.epoch = orte_ess.proc_get_epoch(&daemon);
>>>>> + ORTE_EPOCH_SET(daemon.epoch,orte_ess.proc_get_epoch(&daemon));
>>>>>
>>>>> ret = &daemon;
>>>>>
>>>>> @@ -788,7 +787,7 @@
>>>>> */
>>>>> local_lifeline.jobid = proc->jobid;
>>>>> local_lifeline.vpid = proc->vpid;
>>>>> - local_lifeline.epoch = proc->epoch;
>>>>> + ORTE_EPOCH_SET(local_lifeline.epoch,proc->epoch);
>>>>> lifeline = &local_lifeline;
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> @@ -881,8 +880,7 @@
>>>>> ORTE_PROC_MY_PARENT->vpid = (Ii-Sum) % NInPrevLevel;
>>>>> ORTE_PROC_MY_PARENT->vpid += (Sum - NInPrevLevel);
>>>>> }
>>>>> - ORTE_PROC_MY_PARENT->epoch = ORTE_EPOCH_INVALID;
>>>>> - ORTE_PROC_MY_PARENT->epoch = orte_ess.proc_get_epoch(ORTE_PROC_MY_PARENT);
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_PARENT->epoch,orte_ess.proc_get_epoch(ORTE_PROC_MY_PARENT));
>>>>>
>>>>> /* compute my direct children and the bitmap that shows which vpids
>>>>> * lie underneath their branch
>>>>>
>>>>> Modified: trunk/orte/mca/routed/slave/routed_slave.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/routed/slave/routed_slave.c (original)
>>>>> +++ trunk/orte/mca/routed/slave/routed_slave.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -26,6 +26,7 @@
>>>>> #include "orte/runtime/orte_globals.h"
>>>>> #include "orte/runtime/orte_wait.h"
>>>>> #include "orte/runtime/runtime.h"
>>>>> +#include "orte/runtime/data_type_support/orte_dt_support.h"
>>>>>
>>>>> #include "orte/mca/rml/base/rml_contact.h"
>>>>>
>>>>> @@ -134,7 +135,7 @@
>>>>>
>>>>> if (target->jobid == ORTE_JOBID_INVALID ||
>>>>> target->vpid == ORTE_VPID_INVALID ||
>>>>> - target->epoch == ORTE_EPOCH_INVALID) {
>>>>> + 0 == ORTE_EPOCH_CMP(target->epoch,ORTE_EPOCH_INVALID)) {
>>>>> ret = ORTE_NAME_INVALID;
>>>>> } else {
>>>>> /* a slave must always route via its parent daemon */
>>>>> @@ -275,8 +276,7 @@
>>>>> */
>>>>> local_lifeline.jobid = proc->jobid;
>>>>> local_lifeline.vpid = proc->vpid;
>>>>> - local_lifeline.epoch = ORTE_EPOCH_INVALID;
>>>>> - local_lifeline.epoch = orte_ess.proc_get_epoch(&local_lifeline);
>>>>> + ORTE_EPOCH_SET(local_lifeline.epoch,orte_ess.proc_get_epoch(&local_lifeline));
>>>>>
>>>>> lifeline = &local_lifeline;
>>>>>
>>>>>
>>>>> Modified: trunk/orte/mca/sensor/file/sensor_file.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/sensor/file/sensor_file.c (original)
>>>>> +++ trunk/orte/mca/sensor/file/sensor_file.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -70,7 +70,9 @@
>>>>> opal_list_item_t super;
>>>>> orte_jobid_t jobid;
>>>>> orte_vpid_t vpid;
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> orte_epoch_t epoch;
>>>>> +#endif
>>>>> char *file;
>>>>> int tick;
>>>>> bool check_size;
>>>>>
>>>>> Modified: trunk/orte/mca/snapc/base/snapc_base_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/snapc/base/snapc_base_fns.c (original)
>>>>> +++ trunk/orte/mca/snapc/base/snapc_base_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -81,7 +81,7 @@
>>>>> {
>>>>> snapshot->process_name.jobid = 0;
>>>>> snapshot->process_name.vpid = 0;
>>>>> - snapshot->process_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(snapshot->process_name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> snapshot->state = ORTE_SNAPC_CKPT_STATE_NONE;
>>>>>
>>>>> @@ -92,7 +92,7 @@
>>>>> {
>>>>> snapshot->process_name.jobid = 0;
>>>>> snapshot->process_name.vpid = 0;
>>>>> - snapshot->process_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(snapshot->process_name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> snapshot->state = ORTE_SNAPC_CKPT_STATE_NONE;
>>>>>
>>>>>
>>>>> Modified: trunk/orte/mca/snapc/full/snapc_full_global.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/snapc/full/snapc_full_global.c (original)
>>>>> +++ trunk/orte/mca/snapc/full/snapc_full_global.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -427,7 +427,7 @@
>>>>> new_proc = OBJ_NEW(orte_proc_t);
>>>>> new_proc->name.jobid = proc->name.jobid;
>>>>> new_proc->name.vpid = proc->name.vpid;
>>>>> - new_proc->name.epoch = proc->name.epoch;
>>>>> + ORTE_EPOCH_SET(new_proc->name.epoch,proc->name.epoch);
>>>>> new_proc->node = OBJ_NEW(orte_node_t);
>>>>> new_proc->node->name = proc->node->name;
>>>>> opal_list_append(migrating_procs, &new_proc->super);
>>>>> @@ -618,7 +618,7 @@
>>>>>
>>>>> orted_snapshot->process_name.jobid = cur_node->daemon->name.jobid;
>>>>> orted_snapshot->process_name.vpid = cur_node->daemon->name.vpid;
>>>>> - orted_snapshot->process_name.epoch = cur_node->daemon->name.epoch;
>>>>> + ORTE_EPOCH_SET(orted_snapshot->process_name.epoch,cur_node->daemon->name.epoch);
>>>>>
>>>>> mask = ORTE_NS_CMP_JOBID;
>>>>>
>>>>> @@ -636,7 +636,7 @@
>>>>>
>>>>> app_snapshot->process_name.jobid = procs[p]->name.jobid;
>>>>> app_snapshot->process_name.vpid = procs[p]->name.vpid;
>>>>> - app_snapshot->process_name.epoch = procs[p]->name.epoch;
>>>>> + ORTE_EPOCH_SET(app_snapshot->process_name.epoch,procs[p]->name.epoch);
>>>>>
>>>>> opal_list_append(&(orted_snapshot->super.local_snapshots), &(app_snapshot->super));
>>>>> }
>>>>> @@ -800,7 +800,7 @@
>>>>>
>>>>> app_snapshot->process_name.jobid = procs[p]->name.jobid;
>>>>> app_snapshot->process_name.vpid = procs[p]->name.vpid;
>>>>> - app_snapshot->process_name.epoch = procs[p]->name.epoch;
>>>>> + ORTE_EPOCH_SET(app_snapshot->process_name.epoch,procs[p]->name.epoch);
>>>>>
>>>>> opal_list_append(&(orted_snapshot->super.local_snapshots), &(app_snapshot->super));
>>>>> }
>>>>> @@ -816,7 +816,7 @@
>>>>>
>>>>> orted_snapshot->process_name.jobid = cur_node->daemon->name.jobid;
>>>>> orted_snapshot->process_name.vpid = cur_node->daemon->name.vpid;
>>>>> - orted_snapshot->process_name.epoch = cur_node->daemon->name.epoch;
>>>>> + ORTE_EPOCH_SET(orted_snapshot->process_name.epoch,cur_node->daemon->name.epoch);
>>>>>
>>>>> mask = ORTE_NS_CMP_ALL;
>>>>>
>>>>> @@ -837,7 +837,7 @@
>>>>>
>>>>> app_snapshot->process_name.jobid = procs[p]->name.jobid;
>>>>> app_snapshot->process_name.vpid = procs[p]->name.vpid;
>>>>> - app_snapshot->process_name.epoch = procs[p]->name.epoch;
>>>>> + ORTE_EPOCH_SET(app_snapshot->process_name.epoch,procs[p]->name.epoch);
>>>>>
>>>>> opal_list_append(&(orted_snapshot->super.local_snapshots), &(app_snapshot->super));
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/snapc/full/snapc_full_local.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/snapc/full/snapc_full_local.c (original)
>>>>> +++ trunk/orte/mca/snapc/full/snapc_full_local.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -2033,7 +2033,7 @@
>>>>> vpid_snapshot->process_pid = child->pid;
>>>>> vpid_snapshot->super.process_name.jobid = child->name->jobid;
>>>>> vpid_snapshot->super.process_name.vpid = child->name->vpid;
>>>>> - vpid_snapshot->super.process_name.epoch = child->name->epoch;
>>>>> + ORTE_EPOCH_SET(vpid_snapshot->super.process_name.epoch,child->name->epoch);
>>>>> }
>>>>> }
>>>>>
>>>>> @@ -2095,7 +2095,7 @@
>>>>> vpid_snapshot->process_pid = child->pid;
>>>>> vpid_snapshot->super.process_name.jobid = child->name->jobid;
>>>>> vpid_snapshot->super.process_name.vpid = child->name->vpid;
>>>>> - vpid_snapshot->super.process_name.epoch = child->name->epoch;
>>>>> + ORTE_EPOCH_SET(vpid_snapshot->super.process_name.epoch,child->name->epoch);
>>>>> /*vpid_snapshot->migrating = true;*/
>>>>>
>>>>> opal_list_append(&(local_global_snapshot.local_snapshots), &(vpid_snapshot->super.super));
>>>>> @@ -2111,7 +2111,7 @@
>>>>> vpid_snapshot->process_pid = child->pid;
>>>>> vpid_snapshot->super.process_name.jobid = child->name->jobid;
>>>>> vpid_snapshot->super.process_name.vpid = child->name->vpid;
>>>>> - vpid_snapshot->super.process_name.epoch = child->name->epoch;
>>>>> + ORTE_EPOCH_SET(vpid_snapshot->super.process_name.epoch,child->name->epoch);
>>>>> }
>>>>> }
>>>>>
>>>>>
>>>>> Modified: trunk/orte/mca/snapc/full/snapc_full_module.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/snapc/full/snapc_full_module.c (original)
>>>>> +++ trunk/orte/mca/snapc/full/snapc_full_module.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -83,7 +83,7 @@
>>>>> void orte_snapc_full_orted_construct(orte_snapc_full_orted_snapshot_t *snapshot) {
>>>>> snapshot->process_name.jobid = 0;
>>>>> snapshot->process_name.vpid = 0;
>>>>> - snapshot->process_name.epoch = 0;
>>>>> + ORTE_EPOCH_SET(snapshot->process_name.epoch,0);
>>>>>
>>>>> snapshot->state = ORTE_SNAPC_CKPT_STATE_NONE;
>>>>> }
>>>>> @@ -91,7 +91,7 @@
>>>>> void orte_snapc_full_orted_destruct( orte_snapc_full_orted_snapshot_t *snapshot) {
>>>>> snapshot->process_name.jobid = 0;
>>>>> snapshot->process_name.vpid = 0;
>>>>> - snapshot->process_name.epoch = 0;
>>>>> + ORTE_EPOCH_SET(snapshot->process_name.epoch,0);
>>>>>
>>>>> snapshot->state = ORTE_SNAPC_CKPT_STATE_NONE;
>>>>> }
>>>>>
>>>>> Modified: trunk/orte/mca/sstore/base/sstore_base_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/sstore/base/sstore_base_fns.c (original)
>>>>> +++ trunk/orte/mca/sstore/base/sstore_base_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -62,7 +62,7 @@
>>>>> {
>>>>> snapshot->process_name.jobid = 0;
>>>>> snapshot->process_name.vpid = 0;
>>>>> - snapshot->process_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(snapshot->process_name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> snapshot->crs_comp = NULL;
>>>>> snapshot->compress_comp = NULL;
>>>>> @@ -76,7 +76,7 @@
>>>>> {
>>>>> snapshot->process_name.jobid = 0;
>>>>> snapshot->process_name.vpid = 0;
>>>>> - snapshot->process_name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(snapshot->process_name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> if( NULL != snapshot->crs_comp ) {
>>>>> free(snapshot->crs_comp);
>>>>> @@ -637,7 +637,7 @@
>>>>>
>>>>> vpid_snapshot->process_name.jobid = proc.jobid;
>>>>> vpid_snapshot->process_name.vpid = proc.vpid;
>>>>> - vpid_snapshot->process_name.epoch = proc.epoch;
>>>>> + ORTE_EPOCH_SET(vpid_snapshot->process_name.epoch,proc.epoch);
>>>>> }
>>>>> else if(0 == strncmp(token, SSTORE_METADATA_LOCAL_CRS_COMP_STR, strlen(SSTORE_METADATA_LOCAL_CRS_COMP_STR))) {
>>>>> vpid_snapshot->crs_comp = strdup(value);
>>>>>
>>>>> Modified: trunk/orte/mca/sstore/central/sstore_central_global.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/sstore/central/sstore_central_global.c (original)
>>>>> +++ trunk/orte/mca/sstore/central/sstore_central_global.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -1216,8 +1216,7 @@
>>>>>
>>>>> vpid_snapshot->process_name.jobid = handle_info->jobid;
>>>>> vpid_snapshot->process_name.vpid = i;
>>>>> - vpid_snapshot->process_name.epoch = ORTE_EPOCH_INVALID;
>>>>> - vpid_snapshot->process_name.epoch = orte_ess.proc_get_epoch(&vpid_snapshot->process_name);
>>>>> + ORTE_EPOCH_SET(vpid_snapshot->process_name.epoch,orte_ess.proc_get_epoch(&vpid_snapshot->process_name));
>>>>>
>>>>> vpid_snapshot->crs_comp = NULL;
>>>>> global_snapshot->start_time = NULL;
>>>>>
>>>>> Modified: trunk/orte/mca/sstore/central/sstore_central_local.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/sstore/central/sstore_central_local.c (original)
>>>>> +++ trunk/orte/mca/sstore/central/sstore_central_local.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -210,7 +210,7 @@
>>>>> {
>>>>> info->name.jobid = ORTE_JOBID_INVALID;
>>>>> info->name.vpid = ORTE_VPID_INVALID;
>>>>> - info->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(info->name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> info->local_location = NULL;
>>>>> info->metadata_filename = NULL;
>>>>> @@ -222,7 +222,7 @@
>>>>> {
>>>>> info->name.jobid = ORTE_JOBID_INVALID;
>>>>> info->name.vpid = ORTE_VPID_INVALID;
>>>>> - info->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(info->name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> if( NULL != info->local_location ) {
>>>>> free(info->local_location);
>>>>> @@ -535,7 +535,7 @@
>>>>>
>>>>> app_info->name.jobid = name->jobid;
>>>>> app_info->name.vpid = name->vpid;
>>>>> - app_info->name.epoch = name->epoch;
>>>>> + ORTE_EPOCH_SET(app_info->name.epoch,name->epoch);
>>>>>
>>>>> opal_list_append(handle_info->app_info_handle, &(app_info->super));
>>>>>
>>>>>
>>>>> Modified: trunk/orte/mca/sstore/stage/sstore_stage_global.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/sstore/stage/sstore_stage_global.c (original)
>>>>> +++ trunk/orte/mca/sstore/stage/sstore_stage_global.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -1218,10 +1218,10 @@
>>>>> p_set = OBJ_NEW(orte_filem_base_process_set_t);
>>>>> p_set->source.jobid = peer->jobid;
>>>>> p_set->source.vpid = peer->vpid;
>>>>> - p_set->source.epoch = peer->epoch;
>>>>> + ORTE_EPOCH_SET(p_set->source.epoch,peer->epoch);
>>>>> p_set->sink.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> p_set->sink.vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - p_set->sink.epoch = ORTE_PROC_MY_NAME->epoch;
>>>>> + ORTE_EPOCH_SET(p_set->sink.epoch,ORTE_PROC_MY_NAME->epoch);
>>>>> opal_list_append(&(filem_request->process_sets), &(p_set->super) );
>>>>> }
>>>>>
>>>>> @@ -1706,8 +1706,7 @@
>>>>>
>>>>> vpid_snapshot->process_name.jobid = handle_info->jobid;
>>>>> vpid_snapshot->process_name.vpid = i;
>>>>> - vpid_snapshot->process_name.epoch = ORTE_EPOCH_INVALID;
>>>>> - vpid_snapshot->process_name.epoch = orte_ess.proc_get_epoch(&vpid_snapshot->process_name);
>>>>> + ORTE_EPOCH_SET(vpid_snapshot->process_name.epoch,orte_ess.proc_get_epoch(&vpid_snapshot->process_name));
>>>>>
>>>>> /* JJH: Currently we do not have this information since we do not save
>>>>> * individual vpid info in the Global SStore. It is in the metadata
>>>>>
>>>>> Modified: trunk/orte/mca/sstore/stage/sstore_stage_local.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/mca/sstore/stage/sstore_stage_local.c (original)
>>>>> +++ trunk/orte/mca/sstore/stage/sstore_stage_local.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -287,7 +287,7 @@
>>>>> {
>>>>> info->name.jobid = ORTE_JOBID_INVALID;
>>>>> info->name.vpid = ORTE_VPID_INVALID;
>>>>> - info->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(info->name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> info->local_location = NULL;
>>>>> info->compressed_local_location = NULL;
>>>>> @@ -302,7 +302,7 @@
>>>>> {
>>>>> info->name.jobid = ORTE_JOBID_INVALID;
>>>>> info->name.vpid = ORTE_VPID_INVALID;
>>>>> - info->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(info->name.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> if( NULL != info->local_location ) {
>>>>> free(info->local_location);
>>>>> @@ -1014,7 +1014,7 @@
>>>>>
>>>>> app_info->name.jobid = name->jobid;
>>>>> app_info->name.vpid = name->vpid;
>>>>> - app_info->name.epoch = name->epoch;
>>>>> + ORTE_EPOCH_SET(app_info->name.epoch,name->epoch);
>>>>>
>>>>> opal_list_append(handle_info->app_info_handle, &(app_info->super));
>>>>>
>>>>> @@ -2057,17 +2057,17 @@
>>>>> /* if I am the HNP, then use me as the source */
>>>>> p_set->source.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> p_set->source.vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - p_set->source.epoch = ORTE_PROC_MY_NAME->epoch;
>>>>> + ORTE_EPOCH_SET(p_set->source.epoch,ORTE_PROC_MY_NAME->epoch);
>>>>> }
>>>>> else {
>>>>> /* otherwise, set the HNP as the source */
>>>>> p_set->source.jobid = ORTE_PROC_MY_HNP->jobid;
>>>>> p_set->source.vpid = ORTE_PROC_MY_HNP->vpid;
>>>>> - p_set->source.epoch = ORTE_PROC_MY_HNP->epoch;
>>>>> + ORTE_EPOCH_SET(p_set->source.epoch,ORTE_PROC_MY_HNP->epoch);
>>>>> }
>>>>> p_set->sink.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> p_set->sink.vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - p_set->sink.epoch = ORTE_PROC_MY_NAME->epoch;
>>>>> + ORTE_EPOCH_SET(p_set->sink.epoch,ORTE_PROC_MY_NAME->epoch);
>>>>> opal_list_append(&(filem_request->process_sets), &(p_set->super) );
>>>>>
>>>>> /* Define the file set */
>>>>>
>>>>> Modified: trunk/orte/orted/orted_comm.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/orted/orted_comm.c (original)
>>>>> +++ trunk/orte/orted/orted_comm.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -123,18 +123,13 @@
>>>>> nm = (orte_routed_tree_t*)item;
>>>>>
>>>>> target.vpid = nm->vpid;
>>>>> - target.epoch = orte_util_lookup_epoch(&target);
>>>>> + ORTE_EPOCH_SET(target.epoch,orte_ess.proc_get_epoch(&target));
>>>>>
>>>>> - if (!orte_util_proc_is_running(&target)) {
>>>>> + if (!PROC_IS_RUNNING(&target)) {
>>>>> continue;
>>>>> }
>>>>>
>>>>> - target.epoch = ORTE_EPOCH_INVALID;
>>>>> - if (ORTE_NODE_RANK_INVALID == (target.epoch = orte_ess.proc_get_epoch(&target))) {
>>>>> - /* If we are trying to send to a previously failed process it's
>>>>> - * better to fail silently. */
>>>>> - continue;
>>>>> - }
>>>>> + ORTE_EPOCH_SET(target.epoch,orte_ess.proc_get_epoch(&target));
>>>>>
>>>>> OPAL_OUTPUT_VERBOSE((1, orte_debug_output,
>>>>> "%s orte:daemon:send_relay sending relay msg to %s",
>>>>> @@ -422,7 +417,8 @@
>>>>> proct = OBJ_NEW(orte_proc_t);
>>>>> proct->name.jobid = proc.jobid;
>>>>> proct->name.vpid = proc.vpid;
>>>>> - proct->name.epoch = proc.epoch;
>>>>> + ORTE_EPOCH_SET(proct->name.epoch,proc.epoch);
>>>>> +
>>>>> opal_pointer_array_add(&procarray, proct);
>>>>> num_replies++;
>>>>> }
>>>>> @@ -1059,7 +1055,9 @@
>>>>> orte_job_t *jdata;
>>>>> orte_proc_t *proc;
>>>>> orte_vpid_t vpid;
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> orte_epoch_t epoch;
>>>>> +#endif
>>>>> int32_t i, num_procs;
>>>>>
>>>>> /* setup the answer */
>>>>> @@ -1086,12 +1084,14 @@
>>>>> goto CLEANUP;
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* unpack the epoch */
>>>>> n = 1;
>>>>> if (ORTE_SUCCESS != (ret = opal_dss.unpack(buffer, &epoch, &n, ORTE_EPOCH))) {
>>>>> ORTE_ERROR_LOG(ret);
>>>>> goto CLEANUP;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> /* if they asked for a specific proc, then just get that info */
>>>>> if (ORTE_VPID_WILDCARD != vpid) {
>>>>> @@ -1201,7 +1201,7 @@
>>>>> /* loop across all daemons */
>>>>> proc2.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> for (proc2.vpid=1; proc2.vpid < orte_process_info.num_procs; proc2.vpid++) {
>>>>> - proc2.epoch = orte_util_lookup_epoch(&proc2);
>>>>> + ORTE_EPOCH_SET(proc2.epoch,orte_util_lookup_epoch(&proc2));
>>>>>
>>>>> /* setup the cmd */
>>>>> relay_msg = OBJ_NEW(opal_buffer_t);
>>>>>
>>>>> Modified: trunk/orte/orted/orted_main.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/orted/orted_main.c (original)
>>>>> +++ trunk/orte/orted/orted_main.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -388,14 +388,14 @@
>>>>> orte_process_info.my_daemon_uri = orte_rml.get_contact_info();
>>>>> ORTE_PROC_MY_DAEMON->jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> ORTE_PROC_MY_DAEMON->vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - ORTE_PROC_MY_DAEMON->epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_DAEMON->epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> /* if I am also the hnp, then update that contact info field too */
>>>>> if (ORTE_PROC_IS_HNP) {
>>>>> orte_process_info.my_hnp_uri = orte_rml.get_contact_info();
>>>>> ORTE_PROC_MY_HNP->jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> ORTE_PROC_MY_HNP->vpid = ORTE_PROC_MY_NAME->vpid;
>>>>> - ORTE_PROC_MY_HNP->epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ORTE_PROC_MY_HNP->epoch,ORTE_EPOCH_MIN);
>>>>> }
>>>>>
>>>>> /* setup the primary daemon command receive function */
>>>>> @@ -495,7 +495,8 @@
>>>>> proc = OBJ_NEW(orte_proc_t);
>>>>> proc->name.jobid = jdata->jobid;
>>>>> proc->name.vpid = 0;
>>>>> - proc->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,ORTE_EPOCH_MIN);
>>>>> +
>>>>> proc->state = ORTE_PROC_STATE_RUNNING;
>>>>> proc->app_idx = 0;
>>>>> proc->node = nodes[0]; /* hnp node must be there */
>>>>>
>>>>> Modified: trunk/orte/runtime/data_type_support/orte_dt_compare_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/data_type_support/orte_dt_compare_fns.c (original)
>>>>> +++ trunk/orte/runtime/data_type_support/orte_dt_compare_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -76,6 +76,7 @@
>>>>> }
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /** check the epochs - if one of them is WILDCARD, then ignore
>>>>> * this field since anything is okay
>>>>> */
>>>>> @@ -87,6 +88,7 @@
>>>>> return OPAL_VALUE1_GREATER;
>>>>> }
>>>>> }
>>>>> +#endif
>>>>>
>>>>> /** only way to get here is if all fields are equal or WILDCARD */
>>>>> return OPAL_EQUAL;
>>>>> @@ -122,6 +124,7 @@
>>>>> return OPAL_EQUAL;
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> int orte_dt_compare_epoch(orte_epoch_t *value1,
>>>>> orte_epoch_t *value2,
>>>>> opal_data_type_t type)
>>>>> @@ -136,6 +139,7 @@
>>>>>
>>>>> return OPAL_EQUAL;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>> /**
>>>>>
>>>>> Modified: trunk/orte/runtime/data_type_support/orte_dt_copy_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/data_type_support/orte_dt_copy_fns.c (original)
>>>>> +++ trunk/orte/runtime/data_type_support/orte_dt_copy_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -61,7 +61,7 @@
>>>>>
>>>>> val->jobid = src->jobid;
>>>>> val->vpid = src->vpid;
>>>>> - val->epoch = src->epoch;
>>>>> + ORTE_EPOCH_SET(val->epoch,src->epoch);
>>>>>
>>>>> *dest = val;
>>>>> return ORTE_SUCCESS;
>>>>> @@ -105,6 +105,7 @@
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /*
>>>>> * EPOCH
>>>>> */
>>>>> @@ -123,6 +124,7 @@
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>>
>>>>>
>>>>> Modified: trunk/orte/runtime/data_type_support/orte_dt_packing_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/data_type_support/orte_dt_packing_fns.c (original)
>>>>> +++ trunk/orte/runtime/data_type_support/orte_dt_packing_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -58,7 +58,9 @@
>>>>> orte_process_name_t* proc;
>>>>> orte_jobid_t *jobid;
>>>>> orte_vpid_t *vpid;
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> orte_epoch_t *epoch;
>>>>> +#endif
>>>>>
>>>>> /* collect all the jobids in a contiguous array */
>>>>> jobid = (orte_jobid_t*)malloc(num_vals * sizeof(orte_jobid_t));
>>>>> @@ -100,6 +102,7 @@
>>>>> }
>>>>> free(vpid);
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* Collect all the epochs in a contiguous array */
>>>>> epoch = (orte_epoch_t *) malloc(num_vals * sizeof(orte_epoch_t));
>>>>> if (NULL == epoch) {
>>>>> @@ -118,6 +121,7 @@
>>>>> return rc;
>>>>> }
>>>>> free(epoch);
>>>>> +#endif
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> @@ -156,6 +160,7 @@
>>>>> return ret;
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /*
>>>>> * EPOCH
>>>>> */
>>>>> @@ -171,6 +176,7 @@
>>>>>
>>>>> return ret;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>> /*
>>>>>
>>>>> Modified: trunk/orte/runtime/data_type_support/orte_dt_print_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/data_type_support/orte_dt_print_fns.c (original)
>>>>> +++ trunk/orte/runtime/data_type_support/orte_dt_print_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -125,8 +125,10 @@
>>>>> orte_dt_quick_print(output, "ORTE_STD_CNTR", prefix, src, ORTE_STD_CNTR_T);
>>>>> break;
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> case ORTE_EPOCH:
>>>>> orte_dt_quick_print(output, "ORTE_EPOCH", prefix, src, ORTE_EPOCH_T);
>>>>> +#endif
>>>>>
>>>>> case ORTE_VPID:
>>>>> orte_dt_quick_print(output, "ORTE_VPID", prefix, src, ORTE_VPID_T);
>>>>> @@ -478,11 +480,21 @@
>>>>> if (orte_xml_output) {
>>>>> /* need to create the output in XML format */
>>>>> if (0 == src->pid) {
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> asprintf(output, "%s<process rank=\"%s\" status=\"%s\" epoch=\"%s\"/>\n", pfx2,
>>>>> ORTE_VPID_PRINT(src->name.vpid), orte_proc_state_to_str(src->state), ORTE_EPOCH_PRINT(src->name.epoch));
>>>>> +#else
>>>>> + asprintf(output, "%s<process rank=\"%s\" status=\"%s\"/>\n", pfx2,
>>>>> + ORTE_VPID_PRINT(src->name.vpid), orte_proc_state_to_str(src->state));
>>>>> +#endif
>>>>> } else {
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> asprintf(output, "%s<process rank=\"%s\" pid=\"%d\" status=\"%s\" epoch=\"%s\"/>\n", pfx2,
>>>>> ORTE_VPID_PRINT(src->name.vpid), (int)src->pid, orte_proc_state_to_str(src->state), ORTE_EPOCH_PRINT(src->name.epoch));
>>>>> +#else
>>>>> + asprintf(output, "%s<process rank=\"%s\" pid=\"%d\" status=\"%s\"/>\n", pfx2,
>>>>> + ORTE_VPID_PRINT(src->name.vpid), (int)src->pid, orte_proc_state_to_str(src->state));
>>>>> +#endif
>>>>> }
>>>>> free(pfx2);
>>>>> return ORTE_SUCCESS;
>>>>> @@ -490,10 +502,17 @@
>>>>>
>>>>> if (!orte_devel_level_output) {
>>>>> /* just print a very simple output for users */
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> asprintf(&tmp, "\n%sProcess OMPI jobid: %s Process rank: %s Epoch: %s", pfx2,
>>>>> ORTE_JOBID_PRINT(src->name.jobid),
>>>>> ORTE_VPID_PRINT(src->name.vpid),
>>>>> ORTE_EPOCH_PRINT(src->name.epoch));
>>>>> +#else
>>>>> + asprintf(&tmp, "\n%sProcess OMPI jobid: %s Process rank: %s Epoch: %s", pfx2,
>>>>> + ORTE_JOBID_PRINT(src->name.jobid),
>>>>> + ORTE_VPID_PRINT(src->name.vpid));
>>>>> +#endif
>>>>> +
>>>>> /* set the return */
>>>>> *output = tmp;
>>>>> free(pfx2);
>>>>>
>>>>> Modified: trunk/orte/runtime/data_type_support/orte_dt_size_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/data_type_support/orte_dt_size_fns.c (original)
>>>>> +++ trunk/orte/runtime/data_type_support/orte_dt_size_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -45,9 +45,11 @@
>>>>> *size = sizeof(orte_std_cntr_t);
>>>>> break;
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> case ORTE_EPOCH:
>>>>> *size = sizeof(orte_epoch_t);
>>>>> break;
>>>>> +#endif
>>>>>
>>>>> case ORTE_VPID:
>>>>> *size = sizeof(orte_vpid_t);
>>>>>
>>>>> Modified: trunk/orte/runtime/data_type_support/orte_dt_support.h
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/data_type_support/orte_dt_support.h (original)
>>>>> +++ trunk/orte/runtime/data_type_support/orte_dt_support.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -52,9 +52,14 @@
>>>>> int orte_dt_compare_vpid(orte_vpid_t *value1,
>>>>> orte_vpid_t *value2,
>>>>> opal_data_type_t type);
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> int orte_dt_compare_epoch(orte_epoch_t *value1,
>>>>> orte_epoch_t *value2,
>>>>> opal_data_type_t type);
>>>>> +#define ORTE_EPOCH_CMP(n,m) ( (m) - (n) )
>>>>> +#else
>>>>> +#define ORTE_EPOCH_CMP(n,m) ( 0 )
>>>>> +#endif
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>> int orte_dt_compare_job(orte_job_t *value1, orte_job_t *value2, opal_data_type_t type);
>>>>> int orte_dt_compare_node(orte_node_t *value1, orte_node_t *value2, opal_data_type_t type);
>>>>> @@ -86,7 +91,9 @@
>>>>> int orte_dt_copy_name(orte_process_name_t **dest, orte_process_name_t *src, opal_data_type_t type);
>>>>> int orte_dt_copy_jobid(orte_jobid_t **dest, orte_jobid_t *src, opal_data_type_t type);
>>>>> int orte_dt_copy_vpid(orte_vpid_t **dest, orte_vpid_t *src, opal_data_type_t type);
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> int orte_dt_copy_epoch(orte_epoch_t **dest, orte_epoch_t *src, opal_data_type_t type);
>>>>> +#endif
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>> int orte_dt_copy_job(orte_job_t **dest, orte_job_t *src, opal_data_type_t type);
>>>>> int orte_dt_copy_node(orte_node_t **dest, orte_node_t *src, opal_data_type_t type);
>>>>> @@ -116,8 +123,10 @@
>>>>> int32_t num_vals, opal_data_type_t type);
>>>>> int orte_dt_pack_vpid(opal_buffer_t *buffer, const void *src,
>>>>> int32_t num_vals, opal_data_type_t type);
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> int orte_dt_pack_epoch(opal_buffer_t *buffer, const void *src,
>>>>> int32_t num_vals, opal_data_type_t type);
>>>>> +#endif
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>> int orte_dt_pack_job(opal_buffer_t *buffer, const void *src,
>>>>> int32_t num_vals, opal_data_type_t type);
>>>>> @@ -185,8 +194,10 @@
>>>>> int32_t *num_vals, opal_data_type_t type);
>>>>> int orte_dt_unpack_vpid(opal_buffer_t *buffer, void *dest,
>>>>> int32_t *num_vals, opal_data_type_t type);
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> int orte_dt_unpack_epoch(opal_buffer_t *buffer, void *dest,
>>>>> int32_t *num_vals, opal_data_type_t type);
>>>>> +#endif
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>> int orte_dt_unpack_job(opal_buffer_t *buffer, void *dest,
>>>>> int32_t *num_vals, opal_data_type_t type);
>>>>>
>>>>> Modified: trunk/orte/runtime/data_type_support/orte_dt_unpacking_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/data_type_support/orte_dt_unpacking_fns.c (original)
>>>>> +++ trunk/orte/runtime/data_type_support/orte_dt_unpacking_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -54,7 +54,9 @@
>>>>> orte_process_name_t* proc;
>>>>> orte_jobid_t *jobid;
>>>>> orte_vpid_t *vpid;
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> orte_epoch_t *epoch;
>>>>> +#endif
>>>>>
>>>>> num = *num_vals;
>>>>>
>>>>> @@ -92,6 +94,7 @@
>>>>> return rc;
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* collect all the epochs in a contiguous array */
>>>>> epoch= (orte_epoch_t*)malloc(num * sizeof(orte_epoch_t));
>>>>> if (NULL == epoch) {
>>>>> @@ -109,18 +112,21 @@
>>>>> free(jobid);
>>>>> return rc;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> /* build the names from the jobid/vpid/epoch arrays */
>>>>> proc = (orte_process_name_t*)dest;
>>>>> for (i=0; i < num; i++) {
>>>>> proc->jobid = jobid[i];
>>>>> proc->vpid = vpid[i];
>>>>> - proc->epoch = epoch[i];
>>>>> + ORTE_EPOCH_SET(proc->epoch,epoch[i]);
>>>>> proc++;
>>>>> }
>>>>>
>>>>> /* cleanup */
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> free(epoch);
>>>>> +#endif
>>>>> free(vpid);
>>>>> free(jobid);
>>>>>
>>>>> @@ -159,6 +165,7 @@
>>>>> return ret;
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /*
>>>>> * EPOCH
>>>>> */
>>>>> @@ -174,6 +181,7 @@
>>>>>
>>>>> return ret;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>> /*
>>>>>
>>>>> Modified: trunk/orte/runtime/orte_data_server.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/orte_data_server.c (original)
>>>>> +++ trunk/orte/runtime/orte_data_server.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -220,7 +220,7 @@
>>>>> data->port = port_name;
>>>>> data->owner.jobid = sender->jobid;
>>>>> data->owner.vpid = sender->vpid;
>>>>> - data->owner.epoch = sender->epoch;
>>>>> + ORTE_EPOCH_SET(data->owner.epoch,sender->epoch);
>>>>>
>>>>> /* store the data */
>>>>> data->index = opal_pointer_array_add(orte_data_server_store, data);
>>>>>
>>>>> Modified: trunk/orte/runtime/orte_globals.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/orte_globals.c (original)
>>>>> +++ trunk/orte/runtime/orte_globals.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -277,6 +277,7 @@
>>>>> return rc;
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> tmp = ORTE_EPOCH;
>>>>> if (ORTE_SUCCESS != (rc = opal_dss.register_type(orte_dt_pack_epoch,
>>>>> orte_dt_unpack_epoch,
>>>>> @@ -290,6 +291,7 @@
>>>>> ORTE_ERROR_LOG(rc);
>>>>> return rc;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> #if !ORTE_DISABLE_FULL_SUPPORT
>>>>> tmp = ORTE_JOB;
>>>>> @@ -933,7 +935,7 @@
>>>>> proc->beat = 0;
>>>>> OBJ_CONSTRUCT(&proc->stats, opal_ring_buffer_t);
>>>>> opal_ring_buffer_init(&proc->stats, orte_stat_history_size);
>>>>> - proc->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(proc->name.epoch,ORTE_EPOCH_MIN);
>>>>> #if OPAL_ENABLE_FT_CR == 1
>>>>> proc->ckpt_state = 0;
>>>>> proc->ckpt_snapshot_ref = NULL;
>>>>>
>>>>> Modified: trunk/orte/runtime/orte_init.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/orte_init.c (original)
>>>>> +++ trunk/orte/runtime/orte_init.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -57,8 +57,17 @@
>>>>> char *orte_prohibited_session_dirs = NULL;
>>>>> bool orte_create_session_dirs = true;
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> +orte_process_name_t orte_name_wildcard = {ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD, ORTE_EPOCH_WILDCARD};
>>>>> +#else
>>>>> orte_process_name_t orte_name_wildcard = {ORTE_JOBID_WILDCARD, ORTE_VPID_WILDCARD};
>>>>> +#endif
>>>>> +
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> +orte_process_name_t orte_name_invalid = {ORTE_JOBID_INVALID, ORTE_VPID_INVALID, ORTE_EPOCH_INVALID};
>>>>> +#else
>>>>> orte_process_name_t orte_name_invalid = {ORTE_JOBID_INVALID, ORTE_VPID_INVALID};
>>>>> +#endif
>>>>>
>>>>>
>>>>> #if OPAL_CC_USE_PRAGMA_IDENT
>>>>>
>>>>> Modified: trunk/orte/runtime/orte_wait.h
>>>>> ==============================================================================
>>>>> --- trunk/orte/runtime/orte_wait.h (original)
>>>>> +++ trunk/orte/runtime/orte_wait.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -204,7 +204,7 @@
>>>>> mev = OBJ_NEW(orte_message_event_t); \
>>>>> mev->sender.jobid = (sndr)->jobid; \
>>>>> mev->sender.vpid = (sndr)->vpid; \
>>>>> - mev->sender.epoch = (sndr)->epoch; \
>>>>> + ORTE_EPOCH_SET(mev->sender.epoch,(sndr)->epoch); \
>>>>> opal_dss.copy_payload(mev->buffer, (buf)); \
>>>>> mev->tag = (tg); \
>>>>> mev->file = strdup((buf)->parent.cls_init_file_name); \
>>>>> @@ -228,7 +228,7 @@
>>>>> mev = OBJ_NEW(orte_message_event_t); \
>>>>> mev->sender.jobid = (sndr)->jobid; \
>>>>> mev->sender.vpid = (sndr)->vpid; \
>>>>> - mev->sender.epoch = (sndr)->epoch; \
>>>>> + ORTE_EPOCH_SET(mev->sender.epoch,(sndr)->epoch); \
>>>>> opal_dss.copy_payload(mev->buffer, (buf)); \
>>>>> mev->tag = (tg); \
>>>>> opal_event_evtimer_set(opal_event_base, \
>>>>> @@ -258,7 +258,7 @@
>>>>> tmp = OBJ_NEW(orte_notify_event_t); \
>>>>> tmp->proc.jobid = (data)->jobid; \
>>>>> tmp->proc.vpid = (data)->vpid; \
>>>>> - tmp->proc.epoch = (data)->epoch; \
>>>>> + ORTE_EPOCH_SET(tmp->proc.epoch,(data)->epoch); \
>>>>> opal_event.evtimer_set(opal_event_base, \
>>>>> tmp->ev, (cbfunc), tmp); \
>>>>> now.tv_sec = 0; \
>>>>>
>>>>> Modified: trunk/orte/test/system/oob_stress.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/test/system/oob_stress.c (original)
>>>>> +++ trunk/orte/test/system/oob_stress.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -74,8 +74,7 @@
>>>>>
>>>>> for (j=1; j < count+1; j++) {
>>>>> peer.vpid = (ORTE_PROC_MY_NAME->vpid + j) % orte_process_info.num_procs;
>>>>> - peer.epoch = ORTE_EPOCH_INVALID;
>>>>> - peer.epoch = orte_ess.proc_get_epoch(&peer);
>>>>> + ORTE_EPOCH_SET(peer.epoch,orte_ess.proc_get_epoch(&peer));
>>>>>
>>>>> /* rank0 starts ring */
>>>>> if (ORTE_PROC_MY_NAME->vpid == 0) {
>>>>>
>>>>> Modified: trunk/orte/test/system/orte_ring.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/test/system/orte_ring.c (original)
>>>>> +++ trunk/orte/test/system/orte_ring.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -41,16 +41,14 @@
>>>>> if( right_peer_orte_name.vpid >= num_peers ) {
>>>>> right_peer_orte_name.vpid = 0;
>>>>> }
>>>>> - right_peer_orte_name.epoch = ORTE_EPOCH_INVALID;
>>>>> - right_peer_orte_name.epoch = orte_ess.proc_get_epoch(&right_peer_orte_name);
>>>>> + ORTE_EPOCH_SET(right_peer_orte_name.epoch,orte_ess.proc_get_epoch(&right_peer_orte_name));
>>>>>
>>>>> left_peer_orte_name.jobid = ORTE_PROC_MY_NAME->jobid;
>>>>> left_peer_orte_name.vpid = ORTE_PROC_MY_NAME->vpid - 1;
>>>>> if( ORTE_PROC_MY_NAME->vpid == 0 ) {
>>>>> left_peer_orte_name.vpid = num_peers - 1;
>>>>> }
>>>>> - left_peer_orte_name.epoch = ORTE_EPOCH_INVALID;
>>>>> - left_peer_orte_name.epoch = orte_ess.proc_get_epoch(&left_peer_orte_name);
>>>>> + ORTE_EPOCH_SET(left_peer_orte_name.epoch,orte_ess.proc_get_epoch(&left_peer_orte_name));
>>>>>
>>>>> printf("My name is: %s -- PID %d\tMy Left Peer is %s\tMy Right Peer is %s\n",
>>>>> ORTE_NAME_PRINT(ORTE_PROC_MY_NAME), getpid(),
>>>>>
>>>>> Modified: trunk/orte/test/system/orte_spawn.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/test/system/orte_spawn.c (original)
>>>>> +++ trunk/orte/test/system/orte_spawn.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -74,8 +74,8 @@
>>>>> for (i=0; i < app->num_procs; i++) {
>>>>> name.vpid = i;
>>>>>
>>>>> - name.epoch = ORTE_EPOCH_INVALID;
>>>>> - name.epoch = orte_ess.proc_get_epoch(&name);
>>>>> + ORTE_EPOCH_SET(name.epoch,orte_ess.proc_get_epoch(&name));
>>>>> +
>>>>> fprintf(stderr, "Parent: sending message to child %s\n", ORTE_NAME_PRINT(&name));
>>>>> if (0 > (rc = orte_rml.send(&name, &msg, 1, MY_TAG, 0))) {
>>>>> ORTE_ERROR_LOG(rc);
>>>>>
>>>>> Modified: trunk/orte/tools/orte-ps/orte-ps.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/tools/orte-ps/orte-ps.c (original)
>>>>> +++ trunk/orte/tools/orte-ps/orte-ps.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -869,8 +869,14 @@
>>>>> }
>>>>>
>>>>> /* query the HNP for info on the procs in this job */
>>>>> - if (ORTE_SUCCESS != (ret = orte_util_comm_query_proc_info(&(hnpinfo->hnp->name), job->jobid,
>>>>> - ORTE_VPID_WILDCARD, ORTE_EPOCH_WILDCARD, &cnt, &procs))) {
>>>>> + if (ORTE_SUCCESS != (ret = orte_util_comm_query_proc_info(&(hnpinfo->hnp->name),
>>>>> + job->jobid,
>>>>> + ORTE_VPID_WILDCARD,
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> + ORTE_EPOCH_WILDCARD,
>>>>> +#endif
>>>>> + &cnt,
>>>>> + &procs))) {
>>>>> ORTE_ERROR_LOG(ret);
>>>>> }
>>>>> job->procs->addr = (void**)procs;
>>>>>
>>>>> Modified: trunk/orte/tools/orte-top/orte-top.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/tools/orte-top/orte-top.c (original)
>>>>> +++ trunk/orte/tools/orte-top/orte-top.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -471,7 +471,7 @@
>>>>> if (NULL == ranks) {
>>>>> /* take all ranks */
>>>>> proc.vpid = ORTE_VPID_WILDCARD;
>>>>> - proc.epoch = ORTE_EPOCH_WILDCARD;
>>>>> + ORTE_EPOCH_SET(proc.epoch,ORTE_EPOCH_WILDCARD);
>>>>> if (ORTE_SUCCESS != (ret = opal_dss.pack(&cmdbuf, &proc, 1, ORTE_NAME))) {
>>>>> ORTE_ERROR_LOG(ret);
>>>>> goto cleanup;
>>>>>
>>>>> Modified: trunk/orte/util/comm/comm.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/util/comm/comm.c (original)
>>>>> +++ trunk/orte/util/comm/comm.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -433,8 +433,13 @@
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> int orte_util_comm_query_proc_info(const orte_process_name_t *hnp, orte_jobid_t job, orte_vpid_t vpid,
>>>>> orte_epoch_t epoch, int *num_procs, orte_proc_t ***proc_info_array)
>>>>> +#else
>>>>> +int orte_util_comm_query_proc_info(const orte_process_name_t *hnp, orte_jobid_t job, orte_vpid_t vpid,
>>>>> + int *num_procs, orte_proc_t ***proc_info_array)
>>>>> +#endif
>>>>> {
>>>>> int ret;
>>>>> int32_t cnt, cnt_procs, n;
>>>>> @@ -463,11 +468,13 @@
>>>>> OBJ_RELEASE(cmd);
>>>>> return ret;
>>>>> }
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> if (ORTE_SUCCESS != (ret = opal_dss.pack(cmd, &epoch, 1, ORTE_EPOCH))) {
>>>>> ORTE_ERROR_LOG(ret);
>>>>> OBJ_RELEASE(cmd);
>>>>> return ret;
>>>>> }
>>>>> +#endif
>>>>> /* define a max time to wait for send to complete */
>>>>> timer_fired = false;
>>>>> error_exit = ORTE_SUCCESS;
>>>>>
>>>>> Modified: trunk/orte/util/comm/comm.h
>>>>> ==============================================================================
>>>>> --- trunk/orte/util/comm/comm.h (original)
>>>>> +++ trunk/orte/util/comm/comm.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -52,7 +52,10 @@
>>>>> int *num_nodes, orte_node_t ***node_info_array);
>>>>>
>>>>> ORTE_DECLSPEC int orte_util_comm_query_proc_info(const orte_process_name_t *hnp, orte_jobid_t job, orte_vpid_t vpid,
>>>>> - orte_epoch_t epoch, int *num_procs, orte_proc_t ***proc_info_array);
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> + orte_epoch_t epoch,
>>>>> +#endif
>>>>> + int *num_procs, orte_proc_t ***proc_info_array);
>>>>>
>>>>> ORTE_DECLSPEC int orte_util_comm_spawn_job(const orte_process_name_t *hnp, orte_job_t *jdata);
>>>>>
>>>>>
>>>>> Modified: trunk/orte/util/hnp_contact.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/util/hnp_contact.c (original)
>>>>> +++ trunk/orte/util/hnp_contact.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -55,7 +55,8 @@
>>>>> {
>>>>> ptr->name.jobid = ORTE_JOBID_INVALID;
>>>>> ptr->name.vpid = ORTE_VPID_INVALID;
>>>>> - ptr->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(ptr->name.epoch,ORTE_EPOCH_MIN);
>>>>> +
>>>>> ptr->rml_uri = NULL;
>>>>> }
>>>>> static void orte_hnp_contact_destruct(orte_hnp_contact_t *ptr)
>>>>>
>>>>> Modified: trunk/orte/util/name_fns.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/util/name_fns.c (original)
>>>>> +++ trunk/orte/util/name_fns.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -46,7 +46,7 @@
>>>>> {
>>>>> list->name.jobid = ORTE_JOBID_INVALID;
>>>>> list->name.vpid = ORTE_VPID_INVALID;
>>>>> - list->name.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(list->name.epoch,ORTE_EPOCH_MIN);
>>>>> }
>>>>>
>>>>> /* destructor - used to free any resources held by instance */
>>>>> @@ -116,7 +116,10 @@
>>>>> char* orte_util_print_name_args(const orte_process_name_t *name)
>>>>> {
>>>>> orte_print_args_buffers_t *ptr;
>>>>> - char *job, *vpid, *epoch;
>>>>> + char *job, *vpid;
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> + char *epoch;
>>>>> +#endif
>>>>>
>>>>> /* protect against NULL names */
>>>>> if (NULL == name) {
>>>>> @@ -141,7 +144,7 @@
>>>>> */
>>>>> job = orte_util_print_jobids(name->jobid);
>>>>> vpid = orte_util_print_vpids(name->vpid);
>>>>> - epoch = orte_util_print_epoch(name->epoch);
>>>>> + ORTE_EPOCH_SET(epoch,orte_util_print_epoch(name->epoch));
>>>>>
>>>>> /* get the next buffer */
>>>>> ptr = get_print_name_buffer();
>>>>> @@ -156,9 +159,15 @@
>>>>> ptr->cntr = 0;
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> snprintf(ptr->buffers[ptr->cntr++],
>>>>> ORTE_PRINT_NAME_ARGS_MAX_SIZE,
>>>>> "[%s,%s,%s]", job, vpid, epoch);
>>>>> +#else
>>>>> + snprintf(ptr->buffers[ptr->cntr++],
>>>>> + ORTE_PRINT_NAME_ARGS_MAX_SIZE,
>>>>> + "[%s,%s]", job, vpid);
>>>>> +#endif
>>>>>
>>>>> return ptr->buffers[ptr->cntr-1];
>>>>> }
>>>>> @@ -282,6 +291,7 @@
>>>>> return ptr->buffers[ptr->cntr-1];
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> char* orte_util_print_epoch(const orte_epoch_t epoch)
>>>>> {
>>>>> orte_print_args_buffers_t *ptr;
>>>>> @@ -309,6 +319,7 @@
>>>>> }
>>>>> return ptr->buffers[ptr->cntr-1];
>>>>> }
>>>>> +#endif
>>>>>
>>>>>
>>>>>
>>>>> @@ -403,6 +414,7 @@
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> int orte_util_convert_epoch_to_string(char **epoch_string, const orte_epoch_t epoch)
>>>>> {
>>>>> /* check for wildcard value - handle appropriately */
>>>>> @@ -425,7 +437,6 @@
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> -
>>>>> int orte_util_convert_string_to_epoch(orte_epoch_t *epoch, const char* epoch_string)
>>>>> {
>>>>> if (NULL == epoch_string) { /* got an error */
>>>>> @@ -450,6 +461,7 @@
>>>>>
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> int orte_util_convert_string_to_process_name(orte_process_name_t *name,
>>>>> const char* name_string)
>>>>> @@ -457,13 +469,15 @@
>>>>> char *temp, *token;
>>>>> orte_jobid_t job;
>>>>> orte_vpid_t vpid;
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> orte_epoch_t epoch;
>>>>> +#endif
>>>>> int return_code=ORTE_SUCCESS;
>>>>> -
>>>>> +
>>>>> /* set default */
>>>>> name->jobid = ORTE_JOBID_INVALID;
>>>>> name->vpid = ORTE_VPID_INVALID;
>>>>> - name->epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(name->epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> /* check for NULL string - error */
>>>>> if (NULL == name_string) {
>>>>> @@ -510,6 +524,7 @@
>>>>> vpid = strtoul(token, NULL, 10);
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> token = strtok(NULL, ORTE_SCHEMA_DELIMITER_STRING); /** get next field -> epoch*/
>>>>>
>>>>> /* check for error */
>>>>> @@ -528,10 +543,11 @@
>>>>> } else {
>>>>> epoch = strtoul(token, NULL, 10);
>>>>> }
>>>>> +#endif
>>>>>
>>>>> name->jobid = job;
>>>>> name->vpid = vpid;
>>>>> - name->epoch = epoch;
>>>>> + ORTE_EPOCH_SET(name->epoch,epoch);
>>>>>
>>>>> free(temp);
>>>>>
>>>>> @@ -568,6 +584,7 @@
>>>>> asprintf(&tmp2, "%s%c%lu", tmp, ORTE_SCHEMA_DELIMITER_CHAR, (unsigned long)name->vpid);
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> if (ORTE_EPOCH_WILDCARD == name->epoch) {
>>>>> asprintf(name_string, "%s%c%s", tmp2, ORTE_SCHEMA_DELIMITER_CHAR, ORTE_SCHEMA_WILDCARD_STRING);
>>>>> } else if (ORTE_EPOCH_INVALID == name->epoch) {
>>>>> @@ -575,6 +592,10 @@
>>>>> } else {
>>>>> asprintf(name_string, "%s%c%lu", tmp2, ORTE_SCHEMA_DELIMITER_CHAR, (unsigned long)name->epoch);
>>>>> }
>>>>> +#else
>>>>> + asprintf(name_string, "%s", tmp2);
>>>>> +#endif
>>>>> +
>>>>> free(tmp);
>>>>> free(tmp2);
>>>>>
>>>>> @@ -585,8 +606,11 @@
>>>>> /**** CREATE PROCESS NAME ****/
>>>>> int orte_util_create_process_name(orte_process_name_t **name,
>>>>> orte_jobid_t job,
>>>>> - orte_vpid_t vpid,
>>>>> - orte_epoch_t epoch)
>>>>> + orte_vpid_t vpid
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> + ,orte_epoch_t epoch
>>>>> +#endif
>>>>> + )
>>>>> {
>>>>> *name = NULL;
>>>>>
>>>>> @@ -598,7 +622,8 @@
>>>>>
>>>>> (*name)->jobid = job;
>>>>> (*name)->vpid = vpid;
>>>>> - (*name)->epoch = epoch;
>>>>> + ORTE_EPOCH_SET((*name)->epoch,epoch);
>>>>> +
>>>>> return ORTE_SUCCESS;
>>>>> }
>>>>>
>>>>> @@ -655,6 +680,7 @@
>>>>> }
>>>>> }
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* Get here if jobid's and vpid's are equal, or not being checked.
>>>>> * Now check epoch.
>>>>> */
>>>>> @@ -666,6 +692,7 @@
>>>>> return OPAL_VALUE1_GREATER;
>>>>> }
>>>>> }
>>>>> +#endif
>>>>>
>>>>> /* only way to get here is if all fields are being checked and are equal,
>>>>> * or jobid not checked, but vpid equal,
>>>>>
>>>>> Modified: trunk/orte/util/name_fns.h
>>>>> ==============================================================================
>>>>> --- trunk/orte/util/name_fns.h (original)
>>>>> +++ trunk/orte/util/name_fns.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -61,9 +61,13 @@
>>>>> #define ORTE_VPID_PRINT(n) \
>>>>> orte_util_print_vpids(n)
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> ORTE_DECLSPEC char* orte_util_print_epoch(const orte_epoch_t epoch);
>>>>> #define ORTE_EPOCH_PRINT(n) \
>>>>> orte_util_print_epoch(n)
>>>>> +#else
>>>>> +#define ORTE_EPOCH_PRINT(n)
>>>>> +#endif
>>>>>
>>>>> ORTE_DECLSPEC char* orte_util_print_job_family(const orte_jobid_t job);
>>>>> #define ORTE_JOB_FAMILY_PRINT(n) \
>>>>> @@ -104,6 +108,24 @@
>>>>> #define ORTE_JOBID_IS_DAEMON(n) \
>>>>> !((n) & 0x0000ffff)
>>>>>
>>>>> +/* Macro for getting the epoch out of the process name */
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> +#define ORTE_EPOCH_GET(n) \
>>>>> + ((n)->epoch)
>>>>> +#else
>>>>> +#define ORTE_EPOCH_GET(n)
>>>>> +#endif
>>>>> +
>>>>> +/* Macro for setting the epoch in the process name */
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> +#define ORTE_EPOCH_SET(n,m) \
>>>>> + ( (n) = (m) )
>>>>> +#else
>>>>> +#define ORTE_EPOCH_SET(n,m) \
>>>>> + do { \
>>>>> + } while(0);
>>>>> +#endif
>>>>> +
>>>>> /* List of names for general use */
>>>>> struct orte_namelist_t {
>>>>> opal_list_item_t item; /**< Allows this item to be placed on a list */
>>>>> @@ -117,16 +139,24 @@
>>>>> ORTE_DECLSPEC int orte_util_convert_string_to_jobid(orte_jobid_t *jobid, const char* jobidstring);
>>>>> ORTE_DECLSPEC int orte_util_convert_vpid_to_string(char **vpid_string, const orte_vpid_t vpid);
>>>>> ORTE_DECLSPEC int orte_util_convert_string_to_vpid(orte_vpid_t *vpid, const char* vpidstring);
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> ORTE_DECLSPEC int orte_util_convert_epoch_to_string(char **epoch_string, const orte_epoch_t epoch);
>>>>> ORTE_DECLSPEC int orte_util_convert_string_to_epoch(orte_vpid_t *epoch, const char* epochstring);
>>>>> +#endif
>>>>> ORTE_DECLSPEC int orte_util_convert_string_to_process_name(orte_process_name_t *name,
>>>>> const char* name_string);
>>>>> ORTE_DECLSPEC int orte_util_convert_process_name_to_string(char** name_string,
>>>>> const orte_process_name_t *name);
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> ORTE_DECLSPEC int orte_util_create_process_name(orte_process_name_t **name,
>>>>> orte_jobid_t job,
>>>>> orte_vpid_t vpid,
>>>>> orte_epoch_t epoch);
>>>>> +#else
>>>>> +ORTE_DECLSPEC int orte_util_create_process_name(orte_process_name_t **name,
>>>>> + orte_jobid_t job,
>>>>> + orte_vpid_t vpid);
>>>>> +#endif
>>>>> ORTE_DECLSPEC int orte_util_compare_name_fields(orte_ns_cmp_bitmask_t fields,
>>>>> const orte_process_name_t* name1,
>>>>> const orte_process_name_t* name2);
>>>>>
>>>>> Modified: trunk/orte/util/nidmap.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/util/nidmap.c (original)
>>>>> +++ trunk/orte/util/nidmap.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -249,7 +249,7 @@
>>>>> */
>>>>> /* construct the URI */
>>>>> proc.vpid = node->daemon;
>>>>> - proc.epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(proc.epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> orte_util_convert_process_name_to_string(&proc_name, &proc);
>>>>> asprintf(&uri, "%s;tcp://%s:%d", proc_name, addr, (int)orte_process_info.my_port);
>>>>> @@ -1001,6 +1001,7 @@
>>>>> }
>>>>> #endif
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /* Look up the current epoch value that we have stored locally.
>>>>> *
>>>>> * Note that this will not ping the HNP to get the most up to date epoch stored
>>>>> @@ -1023,7 +1024,9 @@
>>>>> /*print_orte_job_data();*/
>>>>> return e;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> +#if ORTE_RESIL_ORTE
>>>>> bool orte_util_proc_is_running(orte_process_name_t *proc) {
>>>>> int i;
>>>>> unsigned int j;
>>>>> @@ -1078,7 +1081,9 @@
>>>>>
>>>>> return ORTE_ERROR;
>>>>> }
>>>>> +#endif
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> /*
>>>>> * This function performs both the get and set operations on the epoch for a
>>>>> * sepcific process name. If the epoch passed into the function is
>>>>> @@ -1091,6 +1096,11 @@
>>>>> orte_job_t *jdata;
>>>>> orte_proc_t *pdata;
>>>>>
>>>>> + if (ORTE_JOBID_INVALID == proc->jobid ||
>>>>> + ORTE_VPID_INVALID == proc->vpid) {
>>>>> + return ORTE_EPOCH_INVALID;
>>>>> + }
>>>>> +
>>>>> /* Sanity check just to make sure we don't overwrite our existing
>>>>> * orte_job_data.
>>>>> */
>>>>> @@ -1165,4 +1175,5 @@
>>>>> return ORTE_EPOCH_MIN;
>>>>> }
>>>>> }
>>>>> +#endif
>>>>>
>>>>>
>>>>> Modified: trunk/orte/util/nidmap.h
>>>>> ==============================================================================
>>>>> --- trunk/orte/util/nidmap.h (original)
>>>>> +++ trunk/orte/util/nidmap.h 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -48,11 +48,19 @@
>>>>> ORTE_DECLSPEC orte_pmap_t* orte_util_lookup_pmap(orte_process_name_t *proc);
>>>>> ORTE_DECLSPEC orte_nid_t* orte_util_lookup_nid(orte_process_name_t *proc);
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> ORTE_DECLSPEC orte_epoch_t orte_util_lookup_epoch(orte_process_name_t *proc);
>>>>> ORTE_DECLSPEC orte_epoch_t orte_util_set_epoch(orte_process_name_t *proc, orte_epoch_t epoch);
>>>>> +#endif
>>>>>
>>>>> ORTE_DECLSPEC int orte_util_set_proc_state(orte_process_name_t *proc, orte_proc_state_t state);
>>>>> +
>>>>> +#if ORTE_RESIL_ORTE
>>>>> +#define PROC_IS_RUNNING(n) orte_util_proc_is_running(n)
>>>>> ORTE_DECLSPEC bool orte_util_proc_is_running(orte_process_name_t *proc);
>>>>> +#else
>>>>> +#define PROC_IS_RUNNING(n) ( true )
>>>>> +#endif
>>>>>
>>>>> ORTE_DECLSPEC int orte_util_encode_nodemap(opal_byte_object_t *boptr);
>>>>> ORTE_DECLSPEC int orte_util_decode_nodemap(opal_byte_object_t *boptr);
>>>>> @@ -72,5 +80,8 @@
>>>>> END_C_DECLS
>>>>>
>>>>> /* Local functions */
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> orte_epoch_t get_epoch_from_orte_job_data(orte_process_name_t *proc, orte_epoch_t epoch);
>>>>> #endif
>>>>> +
>>>>> +#endif
>>>>>
>>>>> Modified: trunk/orte/util/proc_info.c
>>>>> ==============================================================================
>>>>> --- trunk/orte/util/proc_info.c (original)
>>>>> +++ trunk/orte/util/proc_info.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -36,13 +36,19 @@
>>>>>
>>>>> #include "orte/util/proc_info.h"
>>>>>
>>>>> +#if ORTE_ENABLE_EPOCH
>>>>> +#define ORTE_NAME_INVALID {ORTE_JOBID_INVALID, ORTE_VPID_INVALID, ORTE_EPOCH_MIN}
>>>>> +#else
>>>>> +#define ORTE_NAME_INVALID {ORTE_JOBID_INVALID, ORTE_VPID_INVALID}
>>>>> +#endif
>>>>> +
>>>>> ORTE_DECLSPEC orte_proc_info_t orte_process_info = {
>>>>> - /* .my_name = */ {ORTE_JOBID_INVALID, ORTE_VPID_INVALID, ORTE_EPOCH_MIN},
>>>>> - /* .my_daemon = */ {ORTE_JOBID_INVALID, ORTE_VPID_INVALID, ORTE_EPOCH_MIN},
>>>>> + /* .my_name = */ ORTE_NAME_INVALID,
>>>>> + /* .my_daemon = */ ORTE_NAME_INVALID,
>>>>> /* .my_daemon_uri = */ NULL,
>>>>> - /* .my_hnp = */ {ORTE_JOBID_INVALID, ORTE_VPID_INVALID, ORTE_EPOCH_MIN},
>>>>> + /* .my_hnp = */ ORTE_NAME_INVALID,
>>>>> /* .my_hnp_uri = */ NULL,
>>>>> - /* .my_parent = */ {ORTE_JOBID_INVALID, ORTE_VPID_INVALID, ORTE_EPOCH_MIN},
>>>>> + /* .my_parent = */ ORTE_NAME_INVALID,
>>>>> /* .hnp_pid = */ 0,
>>>>> /* .app_num = */ 0,
>>>>> /* .num_procs = */ 1,
>>>>>
>>>>> Modified: trunk/test/util/orte_session_dir.c
>>>>> ==============================================================================
>>>>> --- trunk/test/util/orte_session_dir.c (original)
>>>>> +++ trunk/test/util/orte_session_dir.c 2011-08-26 18:16:14 EDT (Fri, 26 Aug 2011)
>>>>> @@ -57,7 +57,7 @@
>>>>> orte_process_info.my_name->cellid = 0;
>>>>> orte_process_info.my_name->jobid = 0;
>>>>> orte_process_info.my_name->vpid = 0;
>>>>> - orte_process_info.my_name->epoch = ORTE_EPOCH_MIN;
>>>>> + ORTE_EPOCH_SET(orte_process_info.my_name->epoch,ORTE_EPOCH_MIN);
>>>>>
>>>>> test_init("orte_session_dir_t");
>>>>> test_out = fopen( "test_session_dir_out", "w+" );
>>>>> _______________________________________________
>>>>> svn-full mailing list
>>>>> svn-full_at_[hidden]
>>>>> http://www.open-mpi.org/mailman/listinfo.cgi/svn-full
>>>>
>>>>
>>>> _______________________________________________
>>>> devel mailing list
>>>> devel_at_[hidden]
>>>> http://www.open-mpi.org/mailman/listinfo.cgi/devel
>>>
>>>
>>> _______________________________________________
>>> devel mailing list
>>> devel_at_[hidden]
>>> http://www.open-mpi.org/mailman/listinfo.cgi/devel
>>
>>
>> _______________________________________________
>> devel mailing list
>> devel_at_[hidden]
>> http://www.open-mpi.org/mailman/listinfo.cgi/devel
>
>
> _______________________________________________
> devel mailing list
> devel_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/devel