1
1
openmpi/ompi/mca/btl/sctp/btl_sctp.c
George Bosilca e361bcb64c Send optimizations.
1. The send path get shorter. The BTL is allowed to return > 0 to specify that the
   descriptor was pushed to the networks, and that the memory attached to it is 
   available again for the upper layer. The MCA_BTL_DES_SEND_ALWAYS_CALLBACK flag
   can be used by the PML to force the BTL to always trigger the callback.
   Unmodified BTL will continue to work as expected, as they will return OMPI_SUCCESS
   which force the PML to have exactly the same behavior as before. Some BTLs have
   been modified: self, sm, tcp, mx.
2. Add send immediate interface to BTL.
   The idea is to have a mechanism of allowing the BTL to take advantage of
   send optimizations such as the ability to deliver data "inline". Some
   network APIs such as Portals allow data to be sent using a "thin" event
   without packing data into a memory descriptor. This interface change
   allows the BTL to use such capabilities and allows for other optimizations
   in the future. All existing BTLs except for Portals and sm have this interface
   set to NULL.

This commit was SVN r18551.
2008-05-30 03:58:39 +00:00

497 строки
15 KiB
C

/*
* Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
* University Research and Technology
* Corporation. All rights reserved.
* Copyright (c) 2004-2008 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.
* All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
*
* $HEADER$
*/
#include "ompi_config.h"
#include <string.h>
#include "orte/util/output.h"
#include "opal/util/if.h"
#include "ompi/mca/pml/pml.h"
#include "ompi/mca/btl/btl.h"
#include "btl_sctp.h"
#include "btl_sctp_frag.h"
#include "btl_sctp_proc.h"
#include "btl_sctp_endpoint.h"
#include "ompi/datatype/convertor.h"
#include "ompi/datatype/datatype.h"
#include "ompi/mca/mpool/base/base.h"
#include "ompi/mca/mpool/mpool.h"
#include "ompi/proc/proc.h"
mca_btl_sctp_module_t mca_btl_sctp_module = {
{
&mca_btl_sctp_component.super,
0, /* max size of first fragment */
0, /* min send fragment size */
0, /* max send fragment size */
0, /* btl_rdma_pipeline_send_length */
0, /* btl_rdma_pipeline_frag_size */
0, /* btl_min_rdma_pipeline_size */
0, /* exclusivity */
0, /* latency */
0, /* bandwidth */
0, /* flags */
mca_btl_sctp_add_procs,
mca_btl_sctp_del_procs,
NULL,
mca_btl_sctp_finalize,
mca_btl_sctp_alloc,
mca_btl_sctp_free,
mca_btl_sctp_prepare_src,
mca_btl_sctp_prepare_dst,
mca_btl_sctp_send,
NULL, /* send immediate */
mca_btl_sctp_put,
NULL, /* get */
mca_btl_base_dump,
NULL, /* mpool */
NULL /* register error */
}
};
/**
*
*/
int mca_btl_sctp_add_procs(
struct mca_btl_base_module_t* btl,
size_t nprocs,
struct ompi_proc_t **ompi_procs,
struct mca_btl_base_endpoint_t** peers,
ompi_bitmap_t* reachable)
{
mca_btl_sctp_module_t* sctp_btl = (mca_btl_sctp_module_t*)btl;
ompi_proc_t* my_proc; /* pointer to caller's proc structure */
int i, rc;
/* get pointer to my proc structure */
my_proc = ompi_proc_local();
if( NULL == my_proc ) {
return OMPI_ERR_OUT_OF_RESOURCE;
}
for(i = 0; i < (int) nprocs; i++) {
struct ompi_proc_t* ompi_proc = ompi_procs[i];
mca_btl_sctp_proc_t* sctp_proc;
mca_btl_base_endpoint_t* sctp_endpoint;
/* Do not create loopback SCTP connections */
if( my_proc == ompi_proc ) {
continue;
}
if(NULL == (sctp_proc = mca_btl_sctp_proc_create(ompi_proc))) {
return OMPI_ERR_OUT_OF_RESOURCE;
}
/*
* Check to make sure that the peer has at least as many interface
* addresses exported as we are trying to use. If not, then
* don't bind this BTL instance to the proc.
*/
OPAL_THREAD_LOCK(&sctp_proc->proc_lock);
/* The btl_proc datastructure is shared by all SCTP BTL
* instances that are trying to reach this destination.
* Cache the peer instance on the btl_proc.
*/
sctp_endpoint = OBJ_NEW(mca_btl_sctp_endpoint_t);
if(NULL == sctp_endpoint) {
OPAL_THREAD_UNLOCK(&sctp_proc->proc_lock);
return OMPI_ERR_OUT_OF_RESOURCE;
}
sctp_endpoint->endpoint_btl = sctp_btl;
rc = mca_btl_sctp_proc_insert(sctp_proc, sctp_endpoint);
if(rc != OMPI_SUCCESS) {
OBJ_RELEASE(sctp_endpoint);
OPAL_THREAD_UNLOCK(&sctp_proc->proc_lock);
continue;
}
ompi_bitmap_set_bit(reachable, i);
OPAL_THREAD_UNLOCK(&sctp_proc->proc_lock);
peers[i] = sctp_endpoint;
opal_list_append(&sctp_btl->sctp_endpoints, (opal_list_item_t*)sctp_endpoint);
/* we increase the count of MPI users of the event library
once per peer, so that we are used until we aren't
connected to a peer */
opal_progress_event_users_increment();
}
return OMPI_SUCCESS;
}
int mca_btl_sctp_del_procs(struct mca_btl_base_module_t* btl,
size_t nprocs,
struct ompi_proc_t **procs,
struct mca_btl_base_endpoint_t ** endpoints)
{
mca_btl_sctp_module_t* sctp_btl = (mca_btl_sctp_module_t*)btl;
size_t i;
for(i=0; i<nprocs; i++) {
mca_btl_sctp_endpoint_t* sctp_endpoint = endpoints[i];
if(sctp_endpoint->endpoint_proc != mca_btl_sctp_proc_local()) {
opal_list_remove_item(&sctp_btl->sctp_endpoints, (opal_list_item_t*)sctp_endpoint);
OBJ_RELEASE(sctp_endpoint);
}
opal_progress_event_users_decrement();
}
return OMPI_SUCCESS;
}
/**
* Allocate a segment.
*
* @param btl (IN) BTL module
* @param size (IN) Request segment size.
*/
mca_btl_base_descriptor_t* mca_btl_sctp_alloc(
struct mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* endpoint,
uint8_t order,
size_t size,
uint32_t flags)
{
mca_btl_sctp_frag_t* frag;
int rc;
if(size <= btl->btl_eager_limit) {
MCA_BTL_SCTP_FRAG_ALLOC_EAGER(frag, rc);
frag->segments[0].seg_len = size;
} else if (size <= btl->btl_max_send_size) {
MCA_BTL_SCTP_FRAG_ALLOC_MAX(frag, rc);
frag->segments[0].seg_len = size;
} else {
return NULL;
}
frag->segments[0].seg_addr.pval = frag+1;
frag->base.des_src = frag->segments;
frag->base.des_src_cnt = 1;
frag->base.des_dst = NULL;
frag->base.des_dst_cnt = 0;
frag->base.des_flags = flags;
frag->base.order = MCA_BTL_NO_ORDER;
frag->btl = (mca_btl_sctp_module_t*)btl;
return (mca_btl_base_descriptor_t*)frag;
}
/**
* Return a segment
*/
int mca_btl_sctp_free(
struct mca_btl_base_module_t* btl,
mca_btl_base_descriptor_t* des)
{
mca_btl_sctp_frag_t* frag = (mca_btl_sctp_frag_t*)des;
MCA_BTL_SCTP_FRAG_RETURN(frag);
return OMPI_SUCCESS;
}
/**
* Pack data and return a descriptor that can be
* used for send/put.
*
* @param btl (IN) BTL module
* @param peer (IN) BTL peer addressing
*/
mca_btl_base_descriptor_t* mca_btl_sctp_prepare_src(
struct mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* endpoint,
struct mca_mpool_base_registration_t* registration,
struct ompi_convertor_t* convertor,
uint8_t order,
size_t reserve,
size_t* size,
uint32_t flags)
{
mca_btl_sctp_frag_t* frag;
struct iovec iov;
uint32_t iov_count = 1;
size_t max_data = *size;
int rc;
/*
* if we aren't pinning the data and the requested size is less
* than the eager limit pack into a fragment from the eager pool
*/
if (max_data+reserve <= btl->btl_eager_limit) {
MCA_BTL_SCTP_FRAG_ALLOC_EAGER(frag, rc);
}
/*
* otherwise pack as much data as we can into a fragment
* that is the max send size.
*/
else {
MCA_BTL_SCTP_FRAG_ALLOC_MAX(frag, rc);
}
if(NULL == frag) {
return NULL;
}
if(max_data == 0) {
frag->segments[0].seg_addr.pval = (frag + 1);
frag->segments[0].seg_len = reserve;
frag->base.des_src_cnt = 1;
} else if(ompi_convertor_need_buffers(convertor)) {
if (max_data + reserve > frag->size) {
max_data = frag->size - reserve;
}
iov.iov_len = max_data;
iov.iov_base = (IOVBASE_TYPE*)(((unsigned char*)(frag+1)) + reserve);
rc = ompi_convertor_pack(convertor, &iov, &iov_count, &max_data );
if( rc < 0 ) {
mca_btl_sctp_free(btl, &frag->base);
return NULL;
}
frag->segments[0].seg_addr.pval = (frag + 1);
frag->segments[0].seg_len = max_data + reserve;
frag->base.des_src_cnt = 1;
} else {
iov.iov_len = max_data;
iov.iov_base = NULL;
rc = ompi_convertor_pack(convertor, &iov, &iov_count, &max_data );
if( rc < 0 ) {
mca_btl_sctp_free(btl, &frag->base);
return NULL;
}
frag->segments[0].seg_addr.pval = frag+1;
frag->segments[0].seg_len = reserve;
frag->segments[1].seg_addr.pval = iov.iov_base;
frag->segments[1].seg_len = max_data;
frag->base.des_src_cnt = 2;
}
frag->base.des_src = frag->segments;
frag->base.des_dst = NULL;
frag->base.des_dst_cnt = 0;
frag->base.des_flags = flags;
*size = max_data;
return &frag->base;
}
/**
* Prepare a descriptor for send/rdma using the supplied
* convertor. If the convertor references data that is contigous,
* the descriptor may simply point to the user buffer. Otherwise,
* this routine is responsible for allocating buffer space and
* packing if required.
*
* @param btl (IN) BTL module
* @param endpoint (IN) BTL peer addressing
* @param convertor (IN) Data type convertor
* @param reserve (IN) Additional bytes requested by upper layer to precede user data
* @param size (IN/OUT) Number of bytes to prepare (IN), number of bytes actually prepared (OUT)
*/
mca_btl_base_descriptor_t* mca_btl_sctp_prepare_dst(
struct mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* endpoint,
struct mca_mpool_base_registration_t* registration,
struct ompi_convertor_t* convertor,
uint8_t order,
size_t reserve,
size_t* size,
uint32_t flags)
{
mca_btl_sctp_frag_t* frag;
ptrdiff_t lb;
int rc;
MCA_BTL_SCTP_FRAG_ALLOC_USER(frag, rc);
if(NULL == frag) {
return NULL;
}
ompi_ddt_type_lb(convertor->pDesc, &lb);
frag->segments->seg_len = *size;
frag->segments->seg_addr.pval = convertor->pBaseBuf + lb + convertor->bConverted;
frag->base.des_src = NULL;
frag->base.des_src_cnt = 0;
frag->base.des_dst = frag->segments;
frag->base.des_dst_cnt = 1;
frag->base.des_flags = flags;
return &frag->base;
}
/**
* Initiate an asynchronous send.
*
* @param btl (IN) BTL module
* @param endpoint (IN) BTL addressing information
* @param descriptor (IN) Description of the data to be transfered
* @param tag (IN) The tag value used to notify the peer.
*/
int mca_btl_sctp_send(
struct mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* endpoint,
struct mca_btl_base_descriptor_t* descriptor,
mca_btl_base_tag_t tag)
{
mca_btl_sctp_module_t* sctp_btl = (mca_btl_sctp_module_t*) btl;
mca_btl_sctp_frag_t* frag = (mca_btl_sctp_frag_t*)descriptor;
size_t i;
frag->btl = sctp_btl;
frag->endpoint = endpoint;
frag->rc = 0;
frag->iov_idx = 0;
frag->iov_cnt = 1;
frag->iov_ptr = frag->iov;
frag->iov[0].iov_base = (IOVBASE_TYPE*)&frag->hdr;
frag->iov[0].iov_len = sizeof(frag->hdr);
frag->hdr.size = 0;
for(i=0; i<frag->base.des_src_cnt; i++) {
frag->hdr.size += frag->segments[i].seg_len;
frag->iov[i+1].iov_len = frag->segments[i].seg_len;
frag->iov[i+1].iov_base = (IOVBASE_TYPE*)frag->segments[i].seg_addr.pval;
frag->iov_cnt++;
}
frag->hdr.base.tag = tag;
frag->hdr.type = MCA_BTL_SCTP_HDR_TYPE_SEND;
frag->hdr.count = 0;
if (endpoint->endpoint_nbo) {
MCA_BTL_SCTP_HDR_HTON(frag->hdr);
}
return mca_btl_sctp_endpoint_send(endpoint,frag);
}
/**
* Initiate an asynchronous put.
*
* @param btl (IN) BTL module
* @param endpoint (IN) BTL addressing information
* @param descriptor (IN) Description of the data to be transferred
*/
int mca_btl_sctp_put(
mca_btl_base_module_t* btl,
mca_btl_base_endpoint_t* endpoint,
mca_btl_base_descriptor_t* descriptor)
{
mca_btl_sctp_module_t* sctp_btl = (mca_btl_sctp_module_t*) btl;
mca_btl_sctp_frag_t* frag = (mca_btl_sctp_frag_t*)descriptor;
size_t i;
frag->btl = sctp_btl;
frag->endpoint = endpoint;
frag->rc = 0;
frag->iov_idx = 0;
frag->hdr.size = 0;
frag->iov_cnt = 2;
frag->iov_ptr = frag->iov;
frag->iov[0].iov_base = (IOVBASE_TYPE*)&frag->hdr;
frag->iov[0].iov_len = sizeof(frag->hdr);
frag->iov[1].iov_base = (IOVBASE_TYPE*)frag->base.des_dst;
frag->iov[1].iov_len = frag->base.des_dst_cnt * sizeof(mca_btl_base_segment_t);
for(i=0; i<frag->base.des_src_cnt; i++) {
frag->hdr.size += frag->segments[i].seg_len;
frag->iov[i+2].iov_len = frag->segments[i].seg_len;
frag->iov[i+2].iov_base = (IOVBASE_TYPE*)frag->segments[i].seg_addr.pval;
frag->iov_cnt++;
}
frag->hdr.base.tag = MCA_BTL_TAG_BTL;
frag->hdr.type = MCA_BTL_SCTP_HDR_TYPE_PUT;
frag->hdr.count = frag->base.des_dst_cnt;
if (endpoint->endpoint_nbo) {
MCA_BTL_SCTP_HDR_HTON(frag->hdr);
}
return mca_btl_sctp_endpoint_send(endpoint,frag);
}
/**
* Initiate an asynchronous get.
*
* @param btl (IN) BTL module
* @param endpoint (IN) BTL addressing information
* @param descriptor (IN) Description of the data to be transferred
*
*/
int mca_btl_sctp_get(
mca_btl_base_module_t* btl,
mca_btl_base_endpoint_t* endpoint,
mca_btl_base_descriptor_t* descriptor)
{
mca_btl_sctp_module_t* sctp_btl = (mca_btl_sctp_module_t*) btl;
mca_btl_sctp_frag_t* frag = (mca_btl_sctp_frag_t*)descriptor;
frag->btl = sctp_btl;
frag->endpoint = endpoint;
frag->rc = 0;
frag->iov_idx = 0;
frag->hdr.size = 0;
frag->iov_cnt = 2;
frag->iov_ptr = frag->iov;
frag->iov[0].iov_base = (IOVBASE_TYPE*)&frag->hdr;
frag->iov[0].iov_len = sizeof(frag->hdr);
frag->iov[1].iov_base = (IOVBASE_TYPE*)frag->base.des_src;
frag->iov[1].iov_len = frag->base.des_src_cnt * sizeof(mca_btl_base_segment_t);
frag->hdr.base.tag = MCA_BTL_TAG_BTL;
frag->hdr.type = MCA_BTL_SCTP_HDR_TYPE_GET;
frag->hdr.count = frag->base.des_src_cnt;
if (endpoint->endpoint_nbo) MCA_BTL_SCTP_HDR_HTON(frag->hdr);
return mca_btl_sctp_endpoint_send(endpoint,frag);
}
/*
* Cleanup/release module resources.
*/
int mca_btl_sctp_finalize(struct mca_btl_base_module_t* btl)
{
mca_btl_sctp_module_t* sctp_btl = (mca_btl_sctp_module_t*) btl;
opal_list_item_t* item;
for( item = opal_list_remove_first(&sctp_btl->sctp_endpoints);
item != NULL;
item = opal_list_remove_first(&sctp_btl->sctp_endpoints)) {
mca_btl_sctp_endpoint_t *endpoint = (mca_btl_sctp_endpoint_t*)item;
OBJ_RELEASE(endpoint);
opal_progress_event_users_decrement();
}
free(sctp_btl);
return OMPI_SUCCESS;
}