1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
|
/* -*- Mode: C; c-basic-offset:4 ; indent-tabs-mode:nil -*- */
/*
* 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 (c) 2017 Research Organization for Information Science
* and Technology (RIST). All rights reserved.
* Copyright (c) 2022 Amazon.com, Inc. or its affiliates. All Rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
*
* $HEADER$
*/
#include "ompi_config.h"
#include "pml_ob1.h"
#include "pml_ob1_sendreq.h"
#include "pml_ob1_accelerator.h"
#include "ompi/mca/bml/base/base.h"
#include "pml_ob1_recvreq.h"
#include "opal/runtime/opal_params.h"
/**
* Return the number of completed events allowing the upper level
* to know when no pending events are expected so that it can
* unregister the progress function.
*/
static inline int mca_pml_ob1_process_pending_accelerator_async_copies(void)
{
mca_btl_base_descriptor_t *frag;
int progress, count = 0;
do {
progress = mca_pml_ob1_progress_one_htod_event(&frag);
if (1 == progress) {
/* Call the finish function to make progress. */
mca_pml_ob1_recv_request_frag_copy_finished(NULL, NULL, frag, 0);
count++;
}
} while (progress > 0);
/* Consider progressing dtoh events here in future */
return count;
}
static opal_atomic_int32_t mca_pml_ob1_progress_needed = 0;
int mca_pml_ob1_enable_progress(int32_t count)
{
int32_t progress_count = OPAL_ATOMIC_ADD_FETCH32(&mca_pml_ob1_progress_needed, count);
if( 1 < progress_count )
return 0; /* progress was already on */
opal_progress_register(mca_pml_ob1_progress);
return 1;
}
int mca_pml_ob1_progress(void)
{
int i, queue_length = opal_list_get_size(&mca_pml_ob1.send_pending);
int j, completed_requests = 0;
bool send_succeeded;
completed_requests += mca_pml_ob1_process_pending_accelerator_async_copies();
for( i = 0; i < queue_length; i++ ) {
mca_pml_ob1_send_pending_t pending_type = MCA_PML_OB1_SEND_PENDING_NONE;
mca_pml_ob1_send_request_t* sendreq;
mca_bml_base_endpoint_t* endpoint;
sendreq = get_request_from_send_pending(&pending_type);
if(OPAL_UNLIKELY(NULL == sendreq))
break;
switch(pending_type) {
case MCA_PML_OB1_SEND_PENDING_NONE:
assert(0);
return 0;
case MCA_PML_OB1_SEND_PENDING_SCHEDULE:
if( mca_pml_ob1_send_request_schedule_exclusive(sendreq) ==
OMPI_ERR_OUT_OF_RESOURCE ) {
return 0;
}
completed_requests++;
break;
case MCA_PML_OB1_SEND_PENDING_START:
MCA_PML_OB1_SEND_REQUEST_RESET(sendreq);
endpoint = sendreq->req_endpoint;
send_succeeded = false;
for(j = 0; j < (int)mca_bml_base_btl_array_get_size(&endpoint->btl_eager); j++) {
mca_bml_base_btl_t* bml_btl;
int rc;
/* select a btl */
bml_btl = mca_bml_base_btl_array_get_next(&endpoint->btl_eager);
rc = mca_pml_ob1_send_request_start_btl(sendreq, bml_btl);
if( OPAL_LIKELY(OMPI_SUCCESS == rc) ) {
send_succeeded = true;
completed_requests++;
break;
}
}
if( false == send_succeeded ) {
add_request_to_send_pending(sendreq, MCA_PML_OB1_SEND_PENDING_START, true);
}
}
}
if( 0 != completed_requests ) {
j = OPAL_ATOMIC_ADD_FETCH32(&mca_pml_ob1_progress_needed, -completed_requests);
if( 0 == j ) {
opal_progress_unregister(mca_pml_ob1_progress);
}
}
return completed_requests;
}
|
