1
1
openmpi/ompi/mca/pml/bfo/pml_bfo_failover.h
Brian Barrett 16a1166884 Remove the proc_pml and proc_bml fields from ompi_proc_t and replace with a
configure-time dynamic allocation of flags.  The net result for platforms
which only support BTL-based communication is a reduction of 8*nprocs bytes
per process.  Platforms which support both MTLs and BTLs will not see
a space reduction, but will now be able to safely run both the MTL and BTL
side-by-side, which will prove useful.

This commit was SVN r29100.
2013-08-30 16:54:55 +00:00

399 строки
23 KiB
C

/*
* Copyright (c) 2010 Oracle and/or its affiliates. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
*
* $HEADER$
*/
/**
* @file
* Functions that implement failover capabilities.
*/
#ifndef MCA_PML_BFO_FAILOVER_H
#define MCA_PML_BFO_FAILOVER_H
#include "ompi/mca/btl/btl.h"
#include "pml_bfo_hdr.h"
BEGIN_C_DECLS
bool mca_pml_bfo_is_duplicate_msg(mca_pml_bfo_comm_proc_t* proc,
mca_pml_bfo_match_hdr_t *hdr);
bool mca_pml_bfo_is_duplicate_fin(mca_pml_bfo_hdr_t* hdr, mca_btl_base_descriptor_t* rdma,
mca_btl_base_module_t* btl);
mca_pml_bfo_recv_request_t* mca_pml_bfo_get_request(mca_pml_bfo_match_hdr_t *hdr);
void mca_pml_bfo_send_request_restart(mca_pml_bfo_send_request_t* sendreq,
bool repost, mca_btl_base_tag_t tag);
void mca_pml_bfo_send_request_rndvrestartnotify(mca_pml_bfo_send_request_t* sendreq,
bool repost, mca_btl_base_tag_t tag, int status,
mca_btl_base_module_t* btl);
void
mca_pml_bfo_rndvrestartnotify_completion(mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* ep,
struct mca_btl_base_descriptor_t* des,
int status);
void
mca_pml_bfo_check_recv_ctl_completion_status(mca_btl_base_module_t* btl,
struct mca_btl_base_descriptor_t* des,
int status);
/* Reset a receive request to the beginning */
void mca_pml_bfo_recv_request_reset(mca_pml_bfo_recv_request_t* recvreq);
/* Notify sender that receiver detected an error */
void mca_pml_bfo_recv_request_recverrnotify(mca_pml_bfo_recv_request_t* recvreq,
mca_btl_base_tag_t tag, int status);
/* Ack the RNDVRESTARTNOTIFY message */
void mca_pml_bfo_recv_request_rndvrestartack(mca_pml_bfo_recv_request_t* recvreq,
mca_btl_base_tag_t tag, int status,
mca_btl_base_module_t* btl);
/* Nack the RNDVRESTARTNOTIFY message */
void mca_pml_bfo_recv_request_rndvrestartnack(mca_btl_base_descriptor_t* olddes,
ompi_proc_t* ompi_proc, bool repost);
void mca_pml_bfo_recv_restart_completion(mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* ep,
struct mca_btl_base_descriptor_t* des,
int status);
void mca_pml_bfo_failover_error_handler(struct mca_btl_base_module_t* btl,
int32_t flags, ompi_proc_t *errproc, char *btlname);
void mca_pml_bfo_repost_match_fragment(struct mca_btl_base_descriptor_t* des);
void mca_pml_bfo_repost_fin(struct mca_btl_base_descriptor_t* des);
void mca_pml_bfo_map_out_btl(struct mca_btl_base_module_t* btl,
ompi_proc_t *errproc, char *btlname);
extern void mca_pml_bfo_map_out( mca_btl_base_module_t *btl,
mca_btl_base_tag_t tag,
mca_btl_base_descriptor_t* descriptor,
void* cbdata );
int mca_pml_bfo_register_callbacks(void);
void mca_pml_bfo_update_rndv_fields(mca_pml_bfo_hdr_t* hdr,
mca_pml_bfo_send_request_t*, char *type);
void mca_pml_bfo_update_bml_btl(mca_bml_base_btl_t** bml_btl, mca_btl_base_module_t* btl,
struct mca_btl_base_descriptor_t* des);
void mca_pml_bfo_find_recvreq_eager_bml_btl(mca_bml_base_btl_t** bml_btl,
mca_btl_base_module_t* btl,
mca_pml_bfo_recv_request_t* recvreq,
char* type);
void mca_pml_bfo_find_sendreq_eager_bml_btl(mca_bml_base_btl_t** bml_btl,
mca_btl_base_module_t* btl,
mca_pml_bfo_send_request_t* sendreq,
char* type);
void mca_pml_bfo_find_sendreq_rdma_bml_btl(mca_bml_base_btl_t** bml_btl,
mca_btl_base_module_t* btl,
mca_pml_bfo_send_request_t* sendreq,
char* type);
void mca_pml_bfo_update_eager_bml_btl_recv_ctl(mca_bml_base_btl_t** bml_btl,
mca_btl_base_module_t* btl,
struct mca_btl_base_descriptor_t* des);
void mca_pml_bfo_find_recvreq_rdma_bml_btl(mca_bml_base_btl_t** bml_btl,
mca_btl_base_module_t* btl,
mca_pml_bfo_recv_request_t* recvreq,
char* type);
bool mca_pml_bfo_rndv_completion_status_error(struct mca_btl_base_descriptor_t* des,
mca_pml_bfo_send_request_t* sendreq);
void mca_pml_bfo_send_ctl_completion_status_error(struct mca_btl_base_descriptor_t* des);
void mca_pml_bfo_completion_sendreq_has_error(mca_pml_bfo_send_request_t* sendreq,
int status,
mca_btl_base_module_t* btl,
int type,
char *description);
/**
* Four new callbacks for the four new message types.
*/
extern void mca_pml_bfo_recv_frag_callback_rndvrestartnotify( mca_btl_base_module_t *btl,
mca_btl_base_tag_t tag,
mca_btl_base_descriptor_t* descriptor,
void* cbdata );
extern void mca_pml_bfo_recv_frag_callback_rndvrestartack( mca_btl_base_module_t *btl,
mca_btl_base_tag_t tag,
mca_btl_base_descriptor_t* descriptor,
void* cbdata );
extern void mca_pml_bfo_recv_frag_callback_rndvrestartnack( mca_btl_base_module_t *btl,
mca_btl_base_tag_t tag,
mca_btl_base_descriptor_t* descriptor,
void* cbdata );
extern void mca_pml_bfo_recv_frag_callback_recverrnotify( mca_btl_base_module_t *btl,
mca_btl_base_tag_t tag,
mca_btl_base_descriptor_t* descriptor,
void* cbdata );
/**
* A bunch of macros to help isolate failover code from regular ob1 code.
*/
/* Drop any ACK fragments if request is in error state. Do not want
* to initiate any more activity. */
#define MCA_PML_BFO_ERROR_CHECK_ON_ACK_CALLBACK(sendreq) \
if( OPAL_UNLIKELY((sendreq)->req_error)) { \
opal_output_verbose(20, mca_pml_bfo_output, \
"ACK: received: dropping because request in error, " \
"PML=%d, RQS=%d, src_req=%p, dst_req=%p, peer=%d", \
(uint16_t)(sendreq)->req_send.req_base.req_sequence, \
(sendreq)->req_restartseq, \
(void *)(sendreq), (sendreq)->req_recv.pval, \
(sendreq)->req_send.req_base.req_peer); \
return; \
}
/* Drop any FRAG fragments if request is in error state. Do not want
* to initiate any more activity. */
#define MCA_PML_BFO_ERROR_CHECK_ON_FRAG_CALLBACK(recvreq) \
if( OPAL_UNLIKELY((recvreq)->req_errstate)) { \
opal_output_verbose(20, mca_pml_bfo_output, \
"FRAG: received: dropping because request in error, " \
"PML=%d, src_req=%p, dst_req=%p, peer=%d, offset=%d", \
(uint16_t)(recvreq)->req_msgseq, \
(recvreq)->remote_req_send.pval, \
(void *)(recvreq), \
(recvreq)->req_recv.req_base.req_ompi.req_status.MPI_SOURCE, \
(int)hdr->hdr_frag.hdr_frag_offset); \
return; \
}
/* Drop any PUT fragments if request is in error state. Do not want
* to initiate any more activity. */
#define MCA_PML_BFO_ERROR_CHECK_ON_PUT_CALLBACK(sendreq) \
if( OPAL_UNLIKELY((sendreq)->req_error)) { \
opal_output_verbose(20, mca_pml_bfo_output, \
"PUT: received: dropping because request in error, " \
"PML=%d, src_req=%p, dst_req=%p, peer=%d", \
(uint16_t)(sendreq)->req_send.req_base.req_sequence, \
(void *)(sendreq), (sendreq)->req_recv.pval, \
(sendreq)->req_send.req_base.req_peer); \
return; \
}
/**
* Macros for pml_bfo_recvreq.c file.
*/
/* This can happen if a FIN message arrives after the request was
* marked in error. So, just drop the message. Note that the status
* field is not being checked. That is because the status field is the
* value returned in the FIN hdr.hdr_fail field and may be used for
* other things. Note that we allow the various fields to be updated
* in case this actually completes the request and the sending side
* thinks it is done. */
#define MCA_PML_BFO_ERROR_CHECK_ON_FIN_FOR_PUT(recvreq) \
if( OPAL_UNLIKELY((recvreq)->req_errstate)) { \
opal_output_verbose(20, mca_pml_bfo_output, \
"FIN: received on broken request, skipping, " \
"PML=%d, RQS=%d, src_req=%p, dst_req=%p, peer=%d", \
(recvreq)->req_msgseq, (recvreq)->req_restartseq, \
(recvreq)->remote_req_send.pval, (void *)(recvreq), \
(recvreq)->req_recv.req_base.req_ompi.req_status.MPI_SOURCE); \
/* Even though in error, it still might complete. */ \
recv_request_pml_complete_check(recvreq); \
return; \
}
#define MCA_PML_BFO_ERROR_CHECK_ON_RDMA_READ_COMPLETION(recvreq) \
if ((recvreq)->req_errstate) { \
opal_output_verbose(30, mca_pml_bfo_output, \
"RDMA read: completion failed, error already seen, " \
"PML=%d, RQS=%d, src_req=%lx, dst_req=%lx, peer=%d", \
(recvreq)->req_msgseq, (recvreq)->req_restartseq, \
(unsigned long)(recvreq)->remote_req_send.pval, \
(unsigned long)(recvreq), \
(recvreq)->req_recv.req_base.req_ompi.req_status.MPI_SOURCE); \
return; \
} else { \
opal_output_verbose(30, mca_pml_bfo_output, \
"RDMA read: completion failed, sending RECVERRNOTIFY to " \
"sender, PML=%d, RQS=%d, src_req=%lx, dst_req=%lx, peer=%d", \
(recvreq)->req_msgseq, (recvreq)->req_restartseq, \
(unsigned long)(recvreq)->remote_req_send.pval, \
(unsigned long)(recvreq), \
(recvreq)->req_recv.req_base.req_ompi.req_status.MPI_SOURCE); \
mca_pml_bfo_recv_request_recverrnotify(recvreq, MCA_PML_BFO_HDR_TYPE_RGET, status); \
}
#define MCA_PML_BFO_SECOND_ERROR_CHECK_ON_RDMA_READ_COMPLETION(recvreq, status, btl) \
/* See if the request has received a RNDVRESTARTNOTIFY */ \
if( OPAL_UNLIKELY(recvreq->req_errstate)) { \
if (recvreq->req_errstate & RECVREQ_RNDVRESTART_RECVED) { \
opal_output_verbose(30, mca_pml_bfo_output, \
"RDMA read: completion: recvreq has error, outstanding events=%d " \
"PML=%d, RQS=%d, src_req=%lx, dst_req=%lx, status=%d, peer=%d", \
recvreq->req_events, recvreq->req_msgseq, recvreq->req_restartseq, \
(unsigned long)recvreq->remote_req_send.pval, \
(unsigned long)recvreq, status, \
recvreq->req_recv.req_base.req_ompi.req_status.MPI_SOURCE); \
if (0 == recvreq->req_events) { \
mca_pml_bfo_recv_request_rndvrestartack(recvreq, MCA_PML_BFO_HDR_TYPE_RGET, \
status, btl); \
} \
} \
MCA_PML_BFO_RDMA_FRAG_RETURN(frag); \
return; \
}
/**
* Macros for pml_bfo_sendreq.c file.
*/
/* This macro is called on the sending side after receiving
* a PUT message. There is a chance that this PUT message
* has shown up and is attempting to modify the state of
* the req_state, but the req_state is no longer being tracked
* because the RNDV message has turned into a RGET message
* because it got an error on the RNDV completion.
*/
#define MCA_PML_BFO_VERIFY_SENDREQ_REQ_STATE_VALUE(sendreq) \
if (sendreq->req_state == -1) { \
OPAL_THREAD_ADD32(&sendreq->req_state, 1); \
}
/* Now check the error state. This request can be in error if the
* RNDV message made it over, but the receiver got an error trying to
* send the ACK back and therefore sent a RECVERRNOTIFY message. In
* that case, we want to start the restart dance as the receiver has
* matched this message already. Only restart if there are no
* outstanding events on send request. */
#define MCA_PML_BFO_RNDV_COMPLETION_SENDREQ_ERROR_CHECK(sendreq, status, btl, type, description) \
if( OPAL_UNLIKELY ((sendreq)->req_error)) { \
mca_pml_bfo_completion_sendreq_has_error(sendreq, status, \
btl, type, description); \
return; \
}
/**
* This macro is called within the frag completion function in two
* places. It is called to see if any errors occur prior to the
* completion event on the frag. It is then called a second time
* after the scheduling routine is called as the scheduling routine
* may have detected that a BTL that was cached on the request had
* been removed and therefore marked the request in error. In that
* case, the scheduling of fragments can no longer proceed properly,
* and if there are no outstanding events, iniated the restart dance.
*/
#define MCA_PML_BFO_FRAG_COMPLETION_SENDREQ_ERROR_CHECK(sendreq, status, btl, type, description) \
if( OPAL_UNLIKELY((sendreq)->req_error)) { \
mca_pml_bfo_completion_sendreq_has_error(sendreq, status, \
btl, type, description); \
return; \
}
/* This can happen if a FIN message arrives after the request was
* marked in error. So, just drop the message. Note that the status
* field is not checked here. That is because that is the value
* returned in the FIN hdr.hdr_fail field and may be used for other
* things. */
#define MCA_PML_BFO_RGET_COMPLETION_SENDREQ_ERROR_CHECK(sendreq, btl, des) \
if( OPAL_UNLIKELY(sendreq->req_error)) { \
opal_output_verbose(30, mca_pml_bfo_output, \
"FIN: received on broken request, skipping, " \
"PML=%d, src_req=%lx, dst_req=%lx, peer=%d", \
(uint16_t)sendreq->req_send.req_base.req_sequence, \
(unsigned long)sendreq, (unsigned long)sendreq->req_recv.pval, \
sendreq->req_send.req_base.req_peer); \
btl->btl_free(btl, des); \
return; \
}
/* Check if there has been an error on the send request when we get
* a completion event on the RDMA write. */
#define MCA_PML_BFO_PUT_COMPLETION_SENDREQ_ERROR_CHECK(sendreq, status, btl) \
if ( OPAL_UNLIKELY(sendreq->req_error)) { \
mca_pml_bfo_completion_sendreq_has_error(sendreq, status, btl, \
MCA_PML_BFO_HDR_TYPE_PUT, "RDMA write"); \
MCA_PML_BFO_RDMA_FRAG_RETURN(frag); \
return; \
}
#define MCA_PML_BFO_CHECK_FOR_RNDV_RESTART(hdr, sendreq, type) \
if (0 < sendreq->req_restartseq) { \
mca_pml_bfo_update_rndv_fields(hdr, sendreq, type); \
}
/* If a bml_btl gets mapped out, then we need to adjust it based
* on the btl from the callback function. These macros are called on
* every callback to make sure things are copacetic.
*/
#define MCA_PML_BFO_CHECK_EAGER_BML_BTL_ON_FIN_COMPLETION(bml_btl, btl, des) \
if (bml_btl->btl != btl) { \
ompi_proc_t *proc = (ompi_proc_t*) des->des_cbdata; \
mca_bml_base_endpoint_t* bml_endpoint = (mca_bml_base_endpoint_t*) proc->proc_endpoints[OMPI_PROC_ENDPOINT_TAG_BML]; \
bml_btl = mca_bml_base_btl_array_find(&bml_endpoint->btl_eager, btl); \
}
#define MCA_PML_BFO_CHECK_SENDREQ_EAGER_BML_BTL(bml_btl, btl, sendreq, type) \
if (bml_btl->btl != btl) { \
mca_pml_bfo_find_sendreq_eager_bml_btl(&bml_btl, btl, sendreq, type); \
}
#define MCA_PML_BFO_CHECK_SENDREQ_RDMA_BML_BTL(bml_btl, btl, sendreq, type) \
if (bml_btl->btl != btl) { \
mca_pml_bfo_find_sendreq_rdma_bml_btl(&bml_btl, btl, sendreq, type); \
}
#define MCA_PML_BFO_CHECK_RECVREQ_EAGER_BML_BTL(bml_btl, btl, recvreq, type) \
if (bml_btl->btl != btl) { \
mca_pml_bfo_find_recvreq_eager_bml_btl(&bml_btl, btl, recvreq, type); \
}
#define MCA_PML_BFO_CHECK_RECVREQ_RDMA_BML_BTL(bml_btl, btl, recvreq, type) \
if (bml_btl->btl != btl) { \
mca_pml_bfo_find_recvreq_rdma_bml_btl(&bml_btl, btl, recvreq, type); \
}
#define MCA_PML_BFO_CHECK_RECVREQ_EAGER_BML_BTL_RECV_CTL(bml_btl, btl, des) \
if (bml_btl->btl != btl) { \
mca_pml_bfo_update_eager_bml_btl_recv_ctl(&bml_btl, btl, des); \
}
#define MCA_PML_BFO_CHECK_FOR_REMOVED_BML(sendreq, frag, btl) \
if( OPAL_UNLIKELY(NULL == frag->rdma_bml) ) { \
opal_output_verbose(30, mca_pml_bfo_output, \
"PUT received: no matching BTL to RDMA write to, oustanding " \
"events=%d, PML=%d, RQS=%d, src_req=%p, dst_req=%p, peer=%d", \
sendreq->req_events, \
(uint16_t)sendreq->req_send.req_base.req_sequence, \
sendreq->req_restartseq, (void *)sendreq, \
sendreq->req_recv.pval, sendreq->req_send.req_base.req_peer); \
MCA_PML_BFO_RDMA_FRAG_RETURN(frag); \
sendreq->req_error++; \
if (0 == sendreq->req_events) { \
mca_pml_bfo_send_request_rndvrestartnotify(sendreq, false, \
MCA_PML_BFO_HDR_TYPE_PUT, \
OMPI_ERROR, btl); \
} \
return; \
}
/* This macro checks to see if the cached number of BTLs in the
* send request still matches the value from the endpoint.
* If it does not, this means that a BTL was removed from the
* available list. In this case, start the request over.
*/
#define MCA_PML_BFO_CHECK_FOR_REMOVED_BTL(sendreq, range) \
if ((int)mca_bml_base_btl_array_get_size(&sendreq->req_endpoint->btl_send) \
!= range->range_btl_cnt) { \
sendreq->req_error++; \
return OMPI_ERROR; \
}
END_C_DECLS
#endif