summaryrefslogtreecommitdiff
path: root/src/cpu/trace/trace_cpu.cc
blob: 2e989f6ff416385b91147e947da74c9ac6a0e899 (plain)
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
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
/*
 * Copyright (c) 2013 - 2015 ARM Limited
 * All rights reserved
 *
 * The license below extends only to copyright in the software and shall
 * not be construed as granting a license to any other intellectual
 * property including but not limited to intellectual property relating
 * to a hardware implementation of the functionality of the software
 * licensed hereunder.  You may use the software subject to the license
 * terms below provided that you ensure that this notice is replicated
 * unmodified and in its entirety in all distributions of the software,
 * modified or unmodified, in source code or in binary form.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are
 * met: redistributions of source code must retain the above copyright
 * notice, this list of conditions and the following disclaimer;
 * redistributions in binary form must reproduce the above copyright
 * notice, this list of conditions and the following disclaimer in the
 * documentation and/or other materials provided with the distribution;
 * neither the name of the copyright holders nor the names of its
 * contributors may be used to endorse or promote products derived from
 * this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 * Authors: Radhika Jagtap
 *          Andreas Hansson
 *          Thomas Grass
 */

#include "cpu/trace/trace_cpu.hh"

#include "sim/sim_exit.hh"

// Declare and initialize the static counter for number of trace CPUs.
int TraceCPU::numTraceCPUs = 0;

TraceCPU::TraceCPU(TraceCPUParams *params)
    :   BaseCPU(params),
        icachePort(this),
        dcachePort(this),
        instMasterID(params->system->getMasterId(name() + ".inst")),
        dataMasterID(params->system->getMasterId(name() + ".data")),
        instTraceFile(params->instTraceFile),
        dataTraceFile(params->dataTraceFile),
        icacheGen(*this, ".iside", icachePort, instMasterID, instTraceFile),
        dcacheGen(*this, ".dside", dcachePort, dataMasterID, dataTraceFile,
                    params->sizeROB, params->sizeStoreBuffer,
                    params->sizeLoadBuffer),
        icacheNextEvent(this),
        dcacheNextEvent(this),
        oneTraceComplete(false),
        firstFetchTick(0),
        execCompleteEvent(nullptr)
{
    // Increment static counter for number of Trace CPUs.
    ++TraceCPU::numTraceCPUs;

    // Check that the python parameters for sizes of ROB, store buffer and load
    // buffer do not overflow the corresponding C++ variables.
    fatal_if(params->sizeROB > UINT16_MAX, "ROB size set to %d exceeds the "
                "max. value of %d.\n", params->sizeROB, UINT16_MAX);
    fatal_if(params->sizeStoreBuffer > UINT16_MAX, "ROB size set to %d "
                "exceeds the max. value of %d.\n", params->sizeROB,
                UINT16_MAX);
    fatal_if(params->sizeLoadBuffer > UINT16_MAX, "Load buffer size set to"
                " %d exceeds the max. value of %d.\n",
                params->sizeLoadBuffer, UINT16_MAX);
}

TraceCPU::~TraceCPU()
{

}

TraceCPU*
TraceCPUParams::create()
{
    return new TraceCPU(this);
}

void
TraceCPU::takeOverFrom(BaseCPU *oldCPU)
{
    // Unbind the ports of the old CPU and bind the ports of the TraceCPU.
    assert(!getInstPort().isConnected());
    assert(oldCPU->getInstPort().isConnected());
    BaseSlavePort &inst_peer_port = oldCPU->getInstPort().getSlavePort();
    oldCPU->getInstPort().unbind();
    getInstPort().bind(inst_peer_port);

    assert(!getDataPort().isConnected());
    assert(oldCPU->getDataPort().isConnected());
    BaseSlavePort &data_peer_port = oldCPU->getDataPort().getSlavePort();
    oldCPU->getDataPort().unbind();
    getDataPort().bind(data_peer_port);
}

void
TraceCPU::init()
{
    DPRINTF(TraceCPUInst, "Instruction fetch request trace file is \"%s\"."
            "\n", instTraceFile);
    DPRINTF(TraceCPUData, "Data memory request trace file is \"%s\".\n",
            dataTraceFile);

    BaseCPU::init();

    // Get the send tick of the first instruction read request and schedule
    // icacheNextEvent at that tick.
    Tick first_icache_tick = icacheGen.init();
    schedule(icacheNextEvent, first_icache_tick);

    // Get the send tick of the first data read/write request and schedule
    // dcacheNextEvent at that tick.
    Tick first_dcache_tick = dcacheGen.init();
    schedule(dcacheNextEvent, first_dcache_tick);

    // The static counter for number of Trace CPUs is correctly set at this
    // point so create an event and pass it.
    execCompleteEvent = new CountedExitEvent("end of all traces reached.",
                                                numTraceCPUs);
    // Save the first fetch request tick to dump it as tickOffset
    firstFetchTick = first_icache_tick;
}

void
TraceCPU::schedIcacheNext()
{
    DPRINTF(TraceCPUInst, "IcacheGen event.\n");

    // Try to send the current packet or a retry packet if there is one
    bool sched_next = icacheGen.tryNext();
    // If packet sent successfully, schedule next event
    if (sched_next) {
        DPRINTF(TraceCPUInst, "Scheduling next icacheGen event "
                "at %d.\n", curTick() + icacheGen.tickDelta());
        schedule(icacheNextEvent, curTick() + icacheGen.tickDelta());
        ++numSchedIcacheEvent;
    } else {
        // check if traceComplete. If not, do nothing because sending failed
        // and next event will be scheduled via RecvRetry()
        if (icacheGen.isTraceComplete()) {
            // If this is the first trace to complete, set the variable. If it
            // is already set then both traces are complete to exit sim.
            checkAndSchedExitEvent();
        }
    }
    return;
}

void
TraceCPU::schedDcacheNext()
{
    DPRINTF(TraceCPUData, "DcacheGen event.\n");

    dcacheGen.execute();
    if (dcacheGen.isExecComplete()) {
        checkAndSchedExitEvent();
    }
}

void
TraceCPU::checkAndSchedExitEvent()
{
    if (!oneTraceComplete) {
        oneTraceComplete = true;
    } else {
        // Schedule event to indicate execution is complete as both
        // instruction and data access traces have been played back.
        inform("%s: Execution complete.\n", name());

        // Record stats which are computed at the end of simulation
        tickOffset = firstFetchTick;
        numCycles = (clockEdge() - firstFetchTick) / clockPeriod();
        numOps = dcacheGen.getMicroOpCount();
        schedule(*execCompleteEvent, curTick());
    }
}

void
TraceCPU::regStats()
{

    BaseCPU::regStats();

    numSchedDcacheEvent
    .name(name() + ".numSchedDcacheEvent")
    .desc("Number of events scheduled to trigger data request generator")
    ;

    numSchedIcacheEvent
    .name(name() + ".numSchedIcacheEvent")
    .desc("Number of events scheduled to trigger instruction request generator")
    ;

    numOps
    .name(name() + ".numOps")
    .desc("Number of micro-ops simulated by the Trace CPU")
    ;

    cpi
    .name(name() + ".cpi")
    .desc("Cycles per micro-op used as a proxy for CPI")
    .precision(6)
    ;
    cpi = numCycles/numOps;

    tickOffset
    .name(name() + ".tickOffset")
    .desc("The first execution tick for the root node of elastic traces")
    ;

    icacheGen.regStats();
    dcacheGen.regStats();
}

void
TraceCPU::ElasticDataGen::regStats()
{
    using namespace Stats;

    maxDependents
    .name(name() + ".maxDependents")
    .desc("Max number of dependents observed on a node")
    ;

    maxReadyListSize
    .name(name() + ".maxReadyListSize")
    .desc("Max size of the ready list observed")
    ;

    numSendAttempted
    .name(name() + ".numSendAttempted")
    .desc("Number of first attempts to send a request")
    ;

    numSendSucceeded
    .name(name() + ".numSendSucceeded")
    .desc("Number of successful first attempts")
    ;

    numSendFailed
    .name(name() + ".numSendFailed")
    .desc("Number of failed first attempts")
    ;

    numRetrySucceeded
    .name(name() + ".numRetrySucceeded")
    .desc("Number of successful retries")
    ;

    numSplitReqs
    .name(name() + ".numSplitReqs")
    .desc("Number of split requests")
    ;

    numSOLoads
    .name(name() + ".numSOLoads")
    .desc("Number of strictly ordered loads")
    ;

    numSOStores
    .name(name() + ".numSOStores")
    .desc("Number of strictly ordered stores")
    ;

    dataLastTick
    .name(name() + ".dataLastTick")
    .desc("Last tick simulated from the elastic data trace")
    ;
}

Tick
TraceCPU::ElasticDataGen::init()
{
    DPRINTF(TraceCPUData, "Initializing data memory request generator "
            "DcacheGen: elastic issue with retry.\n");

    if (!readNextWindow())
        panic("Trace has %d elements. It must have at least %d elements.\n",
              depGraph.size(), 2 * windowSize);
    DPRINTF(TraceCPUData, "After 1st read, depGraph size:%d.\n",
            depGraph.size());

    if (!readNextWindow())
        panic("Trace has %d elements. It must have at least %d elements.\n",
              depGraph.size(), 2 * windowSize);
    DPRINTF(TraceCPUData, "After 2st read, depGraph size:%d.\n",
            depGraph.size());

    // Print readyList
    if (DTRACE(TraceCPUData)) {
        printReadyList();
    }
    auto free_itr = readyList.begin();
    DPRINTF(TraceCPUData, "Execute tick of the first dependency free node %lli"
            " is %d.\n", free_itr->seqNum, free_itr->execTick);
    // Return the execute tick of the earliest ready node so that an event
    // can be scheduled to call execute()
    return (free_itr->execTick);
}

void
TraceCPU::ElasticDataGen::exit()
{
    trace.reset();
}

bool
TraceCPU::ElasticDataGen::readNextWindow()
{

    // Read and add next window
    DPRINTF(TraceCPUData, "Reading next window from file.\n");

    if (traceComplete) {
        // We are at the end of the file, thus we have no more records.
        // Return false.
        return false;
    }

    DPRINTF(TraceCPUData, "Start read: Size of depGraph is %d.\n",
            depGraph.size());

    uint32_t num_read = 0;
    while (num_read != windowSize) {

        // Create a new graph node
        GraphNode* new_node = new GraphNode;

        // Read the next line to get the next record. If that fails then end of
        // trace has been reached and traceComplete needs to be set in addition
        // to returning false.
        if (!trace.read(new_node)) {
            DPRINTF(TraceCPUData, "\tTrace complete!\n");
            traceComplete = true;
            return false;
        }

        // Annotate the ROB dependencies of the new node onto the parent nodes.
        addDepsOnParent(new_node, new_node->robDep, new_node->numRobDep);
        // Annotate the register dependencies of the new node onto the parent
        // nodes.
        addDepsOnParent(new_node, new_node->regDep, new_node->numRegDep);

        num_read++;
        // Add to map
        depGraph[new_node->seqNum] = new_node;
        if (new_node->numRobDep == 0 && new_node->numRegDep == 0) {
            // Source dependencies are already complete, check if resources
            // are available and issue. The execution time is approximated
            // to current time plus the computational delay.
            checkAndIssue(new_node);
        }
    }

    DPRINTF(TraceCPUData, "End read: Size of depGraph is %d.\n",
            depGraph.size());
    return true;
}

template<typename T> void
TraceCPU::ElasticDataGen::addDepsOnParent(GraphNode *new_node,
                                            T& dep_array, uint8_t& num_dep)
{
    for (auto& a_dep : dep_array) {
        // The convention is to set the dependencies starting with the first
        // index in the ROB and register dependency arrays. Thus, when we reach
        // a dependency equal to the initialisation value of zero, we know have
        // iterated over all dependencies and can break.
        if (a_dep == 0)
            break;
        // We look up the valid dependency, i.e. the parent of this node
        auto parent_itr = depGraph.find(a_dep);
        if (parent_itr != depGraph.end()) {
            // If the parent is found, it is yet to be executed. Append a
            // pointer to the new node to the dependents list of the parent
            // node.
            parent_itr->second->dependents.push_back(new_node);
            auto num_depts = parent_itr->second->dependents.size();
            maxDependents = std::max<double>(num_depts, maxDependents.value());
        } else {
            // The dependency is not found in the graph. So consider
            // the execution of the parent is complete, i.e. remove this
            // dependency.
            a_dep = 0;
            num_dep--;
        }
    }
}

void
TraceCPU::ElasticDataGen::execute()
{
    DPRINTF(TraceCPUData, "Execute start occupancy:\n");
    DPRINTFR(TraceCPUData, "\tdepGraph = %d, readyList = %d, "
            "depFreeQueue = %d ,", depGraph.size(), readyList.size(),
            depFreeQueue.size());
    hwResource.printOccupancy();

    // Read next window to make sure that dependents of all dep-free nodes
    // are in the depGraph
    if (nextRead) {
        readNextWindow();
        nextRead = false;
    }

    // First attempt to issue the pending dependency-free nodes held
    // in depFreeQueue. If resources have become available for a node,
    // then issue it, i.e. add the node to readyList.
    while (!depFreeQueue.empty()) {
        if (checkAndIssue(depFreeQueue.front(), false)) {
            DPRINTF(TraceCPUData, "Removing from depFreeQueue: seq. num "
                "%lli.\n", (depFreeQueue.front())->seqNum);
            depFreeQueue.pop();
        } else {
            break;
        }
    }
    // Proceed to execute from readyList
    auto graph_itr = depGraph.begin();
    auto free_itr = readyList.begin();
    // Iterate through readyList until the next free node has its execute
    // tick later than curTick or the end of readyList is reached
    while (free_itr->execTick <= curTick() && free_itr != readyList.end()) {

        // Get pointer to the node to be executed
        graph_itr = depGraph.find(free_itr->seqNum);
        assert(graph_itr != depGraph.end());
        GraphNode* node_ptr = graph_itr->second;

        // If there is a retryPkt send that else execute the load
        if (retryPkt) {
            // The retryPkt must be the request that was created by the
            // first node in the readyList.
            if (retryPkt->req->getReqInstSeqNum() != node_ptr->seqNum) {
                panic("Retry packet's seqence number does not match "
                      "the first node in the readyList.\n");
            }
            if (port.sendTimingReq(retryPkt)) {
                ++numRetrySucceeded;
                retryPkt = nullptr;
            }
        } else if (node_ptr->isLoad || node_ptr->isStore) {
            // If there is no retryPkt, attempt to send a memory request in
            // case of a load or store node. If the send fails, executeMemReq()
            // returns a packet pointer, which we save in retryPkt. In case of
            // a comp node we don't do anything and simply continue as if the
            // execution of the comp node succedded.
            retryPkt = executeMemReq(node_ptr);
        }
        // If the retryPkt or a new load/store node failed, we exit from here
        // as a retry from cache will bring the control to execute(). The
        // first node in readyList then, will be the failed node.
        if (retryPkt) {
            break;
        }

        // Proceed to remove dependencies for the successfully executed node.
        // If it is a load which is not strictly ordered and we sent a
        // request for it successfully, we do not yet mark any register
        // dependencies complete. But as per dependency modelling we need
        // to mark ROB dependencies of load and non load/store nodes which
        // are based on successful sending of the load as complete.
        if (node_ptr->isLoad && !node_ptr->isStrictlyOrdered()) {
            // If execute succeeded mark its dependents as complete
            DPRINTF(TraceCPUData, "Node seq. num %lli sent. Waking up "
                    "dependents..\n", node_ptr->seqNum);

            auto child_itr = (node_ptr->dependents).begin();
            while (child_itr != (node_ptr->dependents).end()) {
                // ROB dependency of a store on a load must not be removed
                // after load is sent but after response is received
                if (!(*child_itr)->isStore &&
                    (*child_itr)->removeRobDep(node_ptr->seqNum)) {

                    // Check if the child node has become dependency free
                    if ((*child_itr)->numRobDep == 0 &&
                        (*child_itr)->numRegDep == 0) {

                        // Source dependencies are complete, check if
                        // resources are available and issue
                        checkAndIssue(*child_itr);
                    }
                    // Remove this child for the sent load and point to new
                    // location of the element following the erased element
                    child_itr = node_ptr->dependents.erase(child_itr);
                } else {
                    // This child is not dependency-free, point to the next
                    // child
                    child_itr++;
                }
            }
        } else {
            // If it is a strictly ordered load mark its dependents as complete
            // as we do not send a request for this case. If it is a store or a
            // comp node we also mark all its dependents complete.
            DPRINTF(TraceCPUData, "Node seq. num %lli done. Waking"
                    " up dependents..\n", node_ptr->seqNum);

            for (auto child : node_ptr->dependents) {
                // If the child node is dependency free removeDepOnInst()
                // returns true.
                if (child->removeDepOnInst(node_ptr->seqNum)) {
                    // Source dependencies are complete, check if resources
                    // are available and issue
                    checkAndIssue(child);
                }
            }
        }

        // After executing the node, remove from readyList and delete node.
        readyList.erase(free_itr);
        // If it is a cacheable load which was sent, don't delete
        // just yet.  Delete it in completeMemAccess() after the
        // response is received. If it is an strictly ordered
        // load, it was not sent and all dependencies were simply
        // marked complete. Thus it is safe to delete it. For
        // stores and non load/store nodes all dependencies were
        // marked complete so it is safe to delete it.
        if (!node_ptr->isLoad || node_ptr->isStrictlyOrdered()) {
            // Release all resources occupied by the completed node
            hwResource.release(node_ptr);
            // clear the dynamically allocated set of dependents
            (node_ptr->dependents).clear();
            // delete node
            delete node_ptr;
            // remove from graph
            depGraph.erase(graph_itr);
        }
        // Point to first node to continue to next iteration of while loop
        free_itr = readyList.begin();
    } // end of while loop

    // Print readyList, sizes of queues and resource status after updating
    if (DTRACE(TraceCPUData)) {
        printReadyList();
        DPRINTF(TraceCPUData, "Execute end occupancy:\n");
        DPRINTFR(TraceCPUData, "\tdepGraph = %d, readyList = %d, "
                "depFreeQueue = %d ,", depGraph.size(), readyList.size(),
                depFreeQueue.size());
        hwResource.printOccupancy();
    }

    if (retryPkt) {
        DPRINTF(TraceCPUData, "Not scheduling an event as expecting a retry"
                "event from the cache for seq. num %lli.\n",
                retryPkt->req->getReqInstSeqNum());
        return;
    }
    // If the size of the dependency graph is less than the dependency window
    // then read from the trace file to populate the graph next time we are in
    // execute.
    if (depGraph.size() < windowSize && !traceComplete)
        nextRead = true;

    // If cache is not blocked, schedule an event for the first execTick in
    // readyList else retry from cache will schedule the event. If the ready
    // list is empty then check if the next pending node has resources
    // available to issue. If yes, then schedule an event for the next cycle.
    if (!readyList.empty()) {
        Tick next_event_tick = std::max(readyList.begin()->execTick,
                                        curTick());
        DPRINTF(TraceCPUData, "Attempting to schedule @%lli.\n",
                next_event_tick);
        owner.schedDcacheNextEvent(next_event_tick);
    } else if (readyList.empty() && !depFreeQueue.empty() &&
                hwResource.isAvailable(depFreeQueue.front())) {
        DPRINTF(TraceCPUData, "Attempting to schedule @%lli.\n",
                owner.clockEdge(Cycles(1)));
        owner.schedDcacheNextEvent(owner.clockEdge(Cycles(1)));
    }

    // If trace is completely read, readyList is empty and depGraph is empty,
    // set execComplete to true
    if (depGraph.empty() && readyList.empty() && traceComplete &&
        !hwResource.awaitingResponse()) {
        DPRINTF(TraceCPUData, "\tExecution Complete!\n");
        execComplete = true;
        dataLastTick = curTick();
    }
}

PacketPtr
TraceCPU::ElasticDataGen::executeMemReq(GraphNode* node_ptr)
{

    DPRINTF(TraceCPUData, "Executing memory request %lli (addr %d, pc %#x, "
            "size %d, flags %d).\n", node_ptr->seqNum, node_ptr->addr,
            node_ptr->pc, node_ptr->size, node_ptr->flags);

    // If the request is strictly ordered, do not send it. Just return nullptr
    // as if it was succesfully sent.
    if (node_ptr->isStrictlyOrdered()) {
        node_ptr->isLoad ? ++numSOLoads : ++numSOStores;
        DPRINTF(TraceCPUData, "Skipping strictly ordered request %lli.\n",
                node_ptr->seqNum);
        return nullptr;
    }

    // Check if the request spans two cache lines as this condition triggers
    // an assert fail in the L1 cache. If it does then truncate the size to
    // access only until the end of that line and ignore the remainder. The
    // stat counting this is useful to keep a check on how frequently this
    // happens. If required the code could be revised to mimick splitting such
    // a request into two.
    unsigned blk_size = owner.cacheLineSize();
    Addr blk_offset = (node_ptr->addr & (Addr)(blk_size - 1));
    if (!(blk_offset + node_ptr->size <= blk_size)) {
        node_ptr->size = blk_size - blk_offset;
        ++numSplitReqs;
    }

    // Create a request and the packet containing request
    Request* req = new Request(node_ptr->addr, node_ptr->size, node_ptr->flags,
                               masterID, node_ptr->seqNum,
                               ContextID(0), ThreadID(0));
    req->setPC(node_ptr->pc);
    PacketPtr pkt;
    uint8_t* pkt_data = new uint8_t[req->getSize()];
    if (node_ptr->isLoad) {
        pkt = Packet::createRead(req);
    } else {
        pkt = Packet::createWrite(req);
        memset(pkt_data, 0xA, req->getSize());
    }
    pkt->dataDynamic(pkt_data);

    // Call MasterPort method to send a timing request for this packet
    bool success = port.sendTimingReq(pkt);
    ++numSendAttempted;

    if (!success) {
        // If it fails, return the packet to retry when a retry is signalled by
        // the cache
        ++numSendFailed;
        DPRINTF(TraceCPUData, "Send failed. Saving packet for retry.\n");
        return pkt;
    } else {
        // It is succeeds, return nullptr
        ++numSendSucceeded;
        return nullptr;
    }
}

bool
TraceCPU::ElasticDataGen::checkAndIssue(const GraphNode* node_ptr, bool first)
{
    // Assert the node is dependency-free
    assert(node_ptr->numRobDep == 0 && node_ptr->numRegDep == 0);

    // If this is the first attempt, print a debug message to indicate this.
    if (first) {
        DPRINTFR(TraceCPUData, "\t\tseq. num %lli(%s) with rob num %lli is now"
            " dependency free.\n", node_ptr->seqNum,
            node_ptr->isLoad ? "L" : (node_ptr->isStore ? "S" : "C"),
            node_ptr->robNum);
    }

    // Check if resources are available to issue the specific node
    if (hwResource.isAvailable(node_ptr)) {
        // If resources are free only then add to readyList
        DPRINTFR(TraceCPUData, "\t\tResources available for seq. num %lli. Adding"
            " to readyList, occupying resources.\n", node_ptr->seqNum);
        // Compute the execute tick by adding the compute delay for the node
        // and add the ready node to the ready list
        addToSortedReadyList(node_ptr->seqNum,
                                owner.clockEdge() + node_ptr->compDelay);
        // Account for the resources taken up by this issued node.
        hwResource.occupy(node_ptr);
        return true;

    } else {
        if (first) {
            // Although dependencies are complete, resources are not available.
            DPRINTFR(TraceCPUData, "\t\tResources unavailable for seq. num %lli."
                " Adding to depFreeQueue.\n", node_ptr->seqNum);
            depFreeQueue.push(node_ptr);
        } else {
            DPRINTFR(TraceCPUData, "\t\tResources unavailable for seq. num %lli. "
                "Still pending issue.\n", node_ptr->seqNum);
        }
        return false;
    }
}

void
TraceCPU::ElasticDataGen::completeMemAccess(PacketPtr pkt)
{
    // Release the resources for this completed node.
    if (pkt->isWrite()) {
        // Consider store complete.
        hwResource.releaseStoreBuffer();
        // If it is a store response then do nothing since we do not model
        // dependencies on store completion in the trace. But if we were
        // blocking execution due to store buffer fullness, we need to schedule
        // an event and attempt to progress.
    } else {
        // If it is a load response then release the dependents waiting on it.
        // Get pointer to the completed load
        auto graph_itr = depGraph.find(pkt->req->getReqInstSeqNum());
        assert(graph_itr != depGraph.end());
        GraphNode* node_ptr = graph_itr->second;

        // Release resources occupied by the load
        hwResource.release(node_ptr);

        DPRINTF(TraceCPUData, "Load seq. num %lli response received. Waking up"
                " dependents..\n", node_ptr->seqNum);

        for (auto child : node_ptr->dependents) {
            if (child->removeDepOnInst(node_ptr->seqNum)) {
                checkAndIssue(child);
            }
        }

        // clear the dynamically allocated set of dependents
        (node_ptr->dependents).clear();
        // delete node
        delete node_ptr;
        // remove from graph
        depGraph.erase(graph_itr);
    }

    if (DTRACE(TraceCPUData)) {
        printReadyList();
    }

    // If the size of the dependency graph is less than the dependency window
    // then read from the trace file to populate the graph next time we are in
    // execute.
    if (depGraph.size() < windowSize && !traceComplete)
        nextRead = true;

    // If not waiting for retry, attempt to schedule next event
    if (!retryPkt) {
        // We might have new dep-free nodes in the list which will have execute
        // tick greater than or equal to curTick. But a new dep-free node might
        // have its execute tick earlier. Therefore, attempt to reschedule. It
        // could happen that the readyList is empty and we got here via a
        // last remaining response. So, either the trace is complete or there
        // are pending nodes in the depFreeQueue. The checking is done in the
        // execute() control flow, so schedule an event to go via that flow.
        Tick next_event_tick = readyList.empty() ? owner.clockEdge(Cycles(1)) :
            std::max(readyList.begin()->execTick, owner.clockEdge(Cycles(1)));
        DPRINTF(TraceCPUData, "Attempting to schedule @%lli.\n",
                next_event_tick);
        owner.schedDcacheNextEvent(next_event_tick);
    }
}

void
TraceCPU::ElasticDataGen::addToSortedReadyList(NodeSeqNum seq_num,
                                                    Tick exec_tick)
{
    ReadyNode ready_node;
    ready_node.seqNum = seq_num;
    ready_node.execTick = exec_tick;

    // Iterator to readyList
    auto itr = readyList.begin();

    // If the readyList is empty, simply insert the new node at the beginning
    // and return
    if (itr == readyList.end()) {
        readyList.insert(itr, ready_node);
        maxReadyListSize = std::max<double>(readyList.size(),
                                              maxReadyListSize.value());
        return;
    }

    // If the new node has its execution tick equal to the first node in the
    // list then go to the next node. If the first node in the list failed
    // to execute, its position as the first is thus maintained.
    if (retryPkt)
        if (retryPkt->req->getReqInstSeqNum() == itr->seqNum)
            itr++;

    // Increment the iterator and compare the node pointed to by it to the new
    // node till the position to insert the new node is found.
    bool found = false;
    while (!found && itr != readyList.end()) {
        // If the execution tick of the new node is less than the node then
        // this is the position to insert
        if (exec_tick < itr->execTick)
            found = true;
        // If the execution tick of the new node is equal to the node then
        // sort in ascending order of sequence numbers
        else if (exec_tick == itr->execTick) {
            // If the sequence number of the new node is less than the node
            // then this is the position to insert
            if (seq_num < itr->seqNum)
                found = true;
            // Else go to next node
            else
                itr++;
        }
        // If the execution tick of the new node is greater than the node then
        // go to the next node
        else
            itr++;
    }
    readyList.insert(itr, ready_node);
    // Update the stat for max size reached of the readyList
    maxReadyListSize = std::max<double>(readyList.size(),
                                          maxReadyListSize.value());
}

void
TraceCPU::ElasticDataGen::printReadyList() {

    auto itr = readyList.begin();
    if (itr == readyList.end()) {
        DPRINTF(TraceCPUData, "readyList is empty.\n");
        return;
    }
    DPRINTF(TraceCPUData, "Printing readyList:\n");
    while (itr != readyList.end()) {
        auto graph_itr = depGraph.find(itr->seqNum);
        GraphNode* node_ptr M5_VAR_USED = graph_itr->second;
        DPRINTFR(TraceCPUData, "\t%lld(%s), %lld\n", itr->seqNum,
            node_ptr->isLoad ? "L" : (node_ptr->isStore ? "S" : "C"),
            itr->execTick);
        itr++;
    }
}

TraceCPU::ElasticDataGen::HardwareResource::HardwareResource(
    uint16_t max_rob, uint16_t max_stores, uint16_t max_loads)
  : sizeROB(max_rob),
    sizeStoreBuffer(max_stores),
    sizeLoadBuffer(max_loads),
    oldestInFlightRobNum(UINT64_MAX),
    numInFlightLoads(0),
    numInFlightStores(0)
{}

void
TraceCPU::ElasticDataGen::HardwareResource::occupy(const GraphNode* new_node)
{
    // Occupy ROB entry for the issued node
    // Merely maintain the oldest node, i.e. numerically least robNum by saving
    // it in the variable oldestInFLightRobNum.
    inFlightNodes[new_node->seqNum] = new_node->robNum;
    oldestInFlightRobNum = inFlightNodes.begin()->second;

    // Occupy Load/Store Buffer entry for the issued node if applicable
    if (new_node->isLoad) {
        ++numInFlightLoads;
    } else if (new_node->isStore) {
        ++numInFlightStores;
    } // else if it is a non load/store node, no buffer entry is occupied

    printOccupancy();
}

void
TraceCPU::ElasticDataGen::HardwareResource::release(const GraphNode* done_node)
{
    assert(!inFlightNodes.empty());
    DPRINTFR(TraceCPUData, "\tClearing done seq. num %d from inFlightNodes..\n",
        done_node->seqNum);

    assert(inFlightNodes.find(done_node->seqNum) != inFlightNodes.end());
    inFlightNodes.erase(done_node->seqNum);

    if (inFlightNodes.empty()) {
        // If we delete the only in-flight node and then the
        // oldestInFlightRobNum is set to it's initialized (max) value.
        oldestInFlightRobNum = UINT64_MAX;
    } else {
        // Set the oldest in-flight node rob number equal to the first node in
        // the inFlightNodes since that will have the numerically least value.
        oldestInFlightRobNum = inFlightNodes.begin()->second;
    }

    DPRINTFR(TraceCPUData, "\tCleared. inFlightNodes.size() = %d, "
        "oldestInFlightRobNum = %d\n", inFlightNodes.size(),
        oldestInFlightRobNum);

    // A store is considered complete when a request is sent, thus ROB entry is
    // freed. But it occupies an entry in the Store Buffer until its response
    // is received. A load is considered complete when a response is received,
    // thus both ROB and Load Buffer entries can be released.
    if (done_node->isLoad) {
        assert(numInFlightLoads != 0);
        --numInFlightLoads;
    }
    // For normal writes, we send the requests out and clear a store buffer
    // entry on response. For writes which are strictly ordered, for e.g.
    // writes to device registers, we do that within release() which is called
    // when node is executed and taken off from readyList.
    if (done_node->isStore && done_node->isStrictlyOrdered()) {
        releaseStoreBuffer();
    }
}

void
TraceCPU::ElasticDataGen::HardwareResource::releaseStoreBuffer()
{
    assert(numInFlightStores != 0);
    --numInFlightStores;
}

bool
TraceCPU::ElasticDataGen::HardwareResource::isAvailable(
    const GraphNode* new_node) const
{
    uint16_t num_in_flight_nodes;
    if (inFlightNodes.empty()) {
        num_in_flight_nodes = 0;
        DPRINTFR(TraceCPUData, "\t\tChecking resources to issue seq. num %lli:"
            " #in-flight nodes = 0", new_node->seqNum);
    } else if (new_node->robNum > oldestInFlightRobNum) {
        // This is the intuitive case where new dep-free node is younger
        // instruction than the oldest instruction in-flight. Thus we make sure
        // in_flight_nodes does not overflow.
        num_in_flight_nodes = new_node->robNum - oldestInFlightRobNum;
        DPRINTFR(TraceCPUData, "\t\tChecking resources to issue seq. num %lli:"
            " #in-flight nodes = %d - %d =  %d", new_node->seqNum,
             new_node->robNum, oldestInFlightRobNum, num_in_flight_nodes);
    } else {
        // This is the case where an instruction older than the oldest in-
        // flight instruction becomes dep-free. Thus we must have already
        // accounted for the entry in ROB for this new dep-free node.
        // Immediately after this check returns true, oldestInFlightRobNum will
        // be updated in occupy(). We simply let this node issue now.
        num_in_flight_nodes = 0;
        DPRINTFR(TraceCPUData, "\t\tChecking resources to issue seq. num %lli:"
            " new oldestInFlightRobNum = %d, #in-flight nodes ignored",
            new_node->seqNum, new_node->robNum);
    }
    DPRINTFR(TraceCPUData, ", LQ = %d/%d, SQ  = %d/%d.\n",
        numInFlightLoads, sizeLoadBuffer,
        numInFlightStores, sizeStoreBuffer);
    // Check if resources are available to issue the specific node
    if (num_in_flight_nodes >= sizeROB) {
        return false;
    }
    if (new_node->isLoad && numInFlightLoads >= sizeLoadBuffer) {
        return false;
    }
    if (new_node->isStore && numInFlightStores >= sizeStoreBuffer) {
        return false;
    }
    return true;
}

bool
TraceCPU::ElasticDataGen::HardwareResource::awaitingResponse() const {
    // Return true if there is at least one read or write request in flight
    return (numInFlightStores != 0 || numInFlightLoads != 0);
}

void
TraceCPU::ElasticDataGen::HardwareResource::printOccupancy() {
    DPRINTFR(TraceCPUData, "oldestInFlightRobNum = %d, "
            "LQ = %d/%d, SQ  = %d/%d.\n",
            oldestInFlightRobNum,
            numInFlightLoads, sizeLoadBuffer,
            numInFlightStores, sizeStoreBuffer);
}

void
TraceCPU::FixedRetryGen::regStats()
{
    using namespace Stats;

    numSendAttempted
    .name(name() + ".numSendAttempted")
    .desc("Number of first attempts to send a request")
    ;

    numSendSucceeded
    .name(name() + ".numSendSucceeded")
    .desc("Number of successful first attempts")
    ;

    numSendFailed
    .name(name() + ".numSendFailed")
    .desc("Number of failed first attempts")
    ;

    numRetrySucceeded
    .name(name() + ".numRetrySucceeded")
    .desc("Number of successful retries")
    ;

    instLastTick
    .name(name() + ".instLastTick")
    .desc("Last tick simulated from the fixed inst trace")
    ;
}

Tick
TraceCPU::FixedRetryGen::init()
{
    DPRINTF(TraceCPUInst, "Initializing instruction fetch request generator"
            " IcacheGen: fixed issue with retry.\n");

    if (nextExecute()) {
        DPRINTF(TraceCPUInst, "\tFirst tick = %d.\n", currElement.tick);
        return currElement.tick;
    } else {
        panic("Read of first message in the trace failed.\n");
        return MaxTick;
    }
}

bool
TraceCPU::FixedRetryGen::tryNext()
{
    // If there is a retry packet, try to send it
    if (retryPkt) {

        DPRINTF(TraceCPUInst, "Trying to send retry packet.\n");

        if (!port.sendTimingReq(retryPkt)) {
            // Still blocked! This should never occur.
            DPRINTF(TraceCPUInst, "Retry packet sending failed.\n");
            return false;
        }
        ++numRetrySucceeded;
    } else {

        DPRINTF(TraceCPUInst, "Trying to send packet for currElement.\n");

        // try sending current element
        assert(currElement.isValid());

        ++numSendAttempted;

        if (!send(currElement.addr, currElement.blocksize,
                    currElement.cmd, currElement.flags, currElement.pc)) {
            DPRINTF(TraceCPUInst, "currElement sending failed.\n");
            ++numSendFailed;
            // return false to indicate not to schedule next event
            return false;
        } else {
            ++numSendSucceeded;
        }
    }
    // If packet was sent successfully, either retryPkt or currElement, return
    // true to indicate to schedule event at current Tick plus delta. If packet
    // was sent successfully and there is no next packet to send, return false.
    DPRINTF(TraceCPUInst, "Packet sent successfully, trying to read next "
        "element.\n");
    retryPkt = nullptr;
    // Read next element into currElement, currElement gets cleared so save the
    // tick to calculate delta
    Tick last_tick = currElement.tick;
    if (nextExecute()) {
        assert(currElement.tick >= last_tick);
        delta = currElement.tick - last_tick;
    }
    return !traceComplete;
}

void
TraceCPU::FixedRetryGen::exit()
{
    trace.reset();
}

bool
TraceCPU::FixedRetryGen::nextExecute()
{
    if (traceComplete)
        // We are at the end of the file, thus we have no more messages.
        // Return false.
        return false;


    //Reset the currElement to the default values
    currElement.clear();

    // Read the next line to get the next message. If that fails then end of
    // trace has been reached and traceComplete needs to be set in addition
    // to returning false. If successful then next message is in currElement.
    if (!trace.read(&currElement)) {
        traceComplete = true;
        instLastTick = curTick();
        return false;
    }

    DPRINTF(TraceCPUInst, "inst fetch: %c addr %d pc %#x size %d tick %d\n",
            currElement.cmd.isRead() ? 'r' : 'w',
            currElement.addr,
            currElement.pc,
            currElement.blocksize,
            currElement.tick);

    return true;
}

bool
TraceCPU::FixedRetryGen::send(Addr addr, unsigned size, const MemCmd& cmd,
              Request::FlagsType flags, Addr pc)
{

    // Create new request
    Request* req = new Request(addr, size, flags, masterID);
    req->setPC(pc);

    // If this is not done it triggers assert in L1 cache for invalid contextId
    req->setThreadContext(ContextID(0), ThreadID(0));

    // Embed it in a packet
    PacketPtr pkt = new Packet(req, cmd);

    uint8_t* pkt_data = new uint8_t[req->getSize()];
    pkt->dataDynamic(pkt_data);

    if (cmd.isWrite()) {
        memset(pkt_data, 0xA, req->getSize());
    }

    // Call MasterPort method to send a timing request for this packet
    bool success = port.sendTimingReq(pkt);
    if (!success) {
        // If it fails, save the packet to retry when a retry is signalled by
        // the cache
        retryPkt = pkt;
    }
    return success;
}

void
TraceCPU::icacheRetryRecvd()
{
    // Schedule an event to go through the control flow in the same tick as
    // retry is received
    DPRINTF(TraceCPUInst, "Icache retry received. Scheduling next IcacheGen"
            " event @%lli.\n", curTick());
    schedule(icacheNextEvent, curTick());
}

void
TraceCPU::dcacheRetryRecvd()
{
    // Schedule an event to go through the execute flow in the same tick as
    // retry is received
    DPRINTF(TraceCPUData, "Dcache retry received. Scheduling next DcacheGen"
            " event @%lli.\n", curTick());
    schedule(dcacheNextEvent, curTick());
}

void
TraceCPU::schedDcacheNextEvent(Tick when)
{
    if (!dcacheNextEvent.scheduled()) {
        DPRINTF(TraceCPUData, "Scheduling next DcacheGen event at %lli.\n",
                when);
        schedule(dcacheNextEvent, when);
        ++numSchedDcacheEvent;
    } else if (when < dcacheNextEvent.when()) {
        DPRINTF(TraceCPUData, "Re-scheduling next dcache event from %lli"
                " to %lli.\n", dcacheNextEvent.when(), when);
        reschedule(dcacheNextEvent, when);
    }

}

bool
TraceCPU::IcachePort::recvTimingResp(PacketPtr pkt)
{
    // All responses on the instruction fetch side are ignored. Simply delete
    // the request and packet to free allocated memory
    delete pkt->req;
    delete pkt;

    return true;
}

void
TraceCPU::IcachePort::recvReqRetry()
{
    owner->icacheRetryRecvd();
}

void
TraceCPU::dcacheRecvTimingResp(PacketPtr pkt)
{
    DPRINTF(TraceCPUData, "Received timing response from Dcache.\n");
    dcacheGen.completeMemAccess(pkt);
}

bool
TraceCPU::DcachePort::recvTimingResp(PacketPtr pkt)
{
    // Handle the responses for data memory requests which is done inside the
    // elastic data generator
    owner->dcacheRecvTimingResp(pkt);
    // After processing the response delete the request and packet to free
    // memory
    delete pkt->req;
    delete pkt;

    return true;
}

void
TraceCPU::DcachePort::recvReqRetry()
{
    owner->dcacheRetryRecvd();
}

TraceCPU::ElasticDataGen::InputStream::InputStream(const std::string& filename)
    : trace(filename),
      microOpCount(0)
{
    // Create a protobuf message for the header and read it from the stream
    ProtoMessage::InstDepRecordHeader header_msg;
    if (!trace.read(header_msg)) {
        panic("Failed to read packet header from %s\n", filename);

        if (header_msg.tick_freq() != SimClock::Frequency) {
            panic("Trace %s was recorded with a different tick frequency %d\n",
                  header_msg.tick_freq());
        }
    } else {
        // Assign window size equal to the field in the trace that was recorded
        // when the data dependency trace was captured in the o3cpu model
        windowSize = header_msg.window_size();
    }
}

void
TraceCPU::ElasticDataGen::InputStream::reset()
{
    trace.reset();
}

bool
TraceCPU::ElasticDataGen::InputStream::read(GraphNode* element)
{
    ProtoMessage::InstDepRecord pkt_msg;
    if (trace.read(pkt_msg)) {
        // Required fields
        element->seqNum = pkt_msg.seq_num();
        element->isLoad = pkt_msg.load();
        element->isStore = pkt_msg.store();
        element->compDelay = pkt_msg.comp_delay();

        // Repeated field robDepList
        element->clearRobDep();
        assert((pkt_msg.rob_dep()).size() <= element->maxRobDep);
        for (int i = 0; i < (pkt_msg.rob_dep()).size(); i++) {
            element->robDep[element->numRobDep] = pkt_msg.rob_dep(i);
            element->numRobDep += 1;
        }

        // Repeated field
        element->clearRegDep();
        assert((pkt_msg.reg_dep()).size() <= TheISA::MaxInstSrcRegs);
        for (int i = 0; i < (pkt_msg.reg_dep()).size(); i++) {
            // There is a possibility that an instruction has both, a register
            // and order dependency on an instruction. In such a case, the
            // register dependency is omitted
            bool duplicate = false;
            for (int j = 0; j < element->numRobDep; j++) {
                duplicate |= (pkt_msg.reg_dep(i) == element->robDep[j]);
            }
            if (!duplicate) {
                element->regDep[element->numRegDep] = pkt_msg.reg_dep(i);
                element->numRegDep += 1;
            }
        }

        // Optional fields
        if (pkt_msg.has_addr())
            element->addr = pkt_msg.addr();
        else
            element->addr = 0;

        if (pkt_msg.has_size())
            element->size = pkt_msg.size();
        else
            element->size = 0;

        if (pkt_msg.has_flags())
            element->flags = pkt_msg.flags();
        else
            element->flags = 0;

        if (pkt_msg.has_pc())
            element->pc = pkt_msg.pc();
        else
            element->pc = 0;

        // ROB occupancy number
        ++microOpCount;
        if (pkt_msg.has_weight()) {
            microOpCount += pkt_msg.weight();
        }
        element->robNum = microOpCount;
        return true;
    }

    // We have reached the end of the file
    return false;
}

bool
TraceCPU::ElasticDataGen::GraphNode::removeRegDep(NodeSeqNum reg_dep)
{
    for (auto& own_reg_dep : regDep) {
        if (own_reg_dep == reg_dep) {
            // If register dependency is found, make it zero and return true
            own_reg_dep = 0;
            --numRegDep;
            assert(numRegDep >= 0);
            DPRINTFR(TraceCPUData, "\tFor %lli: Marking register dependency %lli "
                    "done.\n", seqNum, reg_dep);
            return true;
        }
    }

    // Return false if the dependency is not found
    return false;
}

bool
TraceCPU::ElasticDataGen::GraphNode::removeRobDep(NodeSeqNum rob_dep)
{
    for (auto& own_rob_dep : robDep) {
        if (own_rob_dep == rob_dep) {
            // If the rob dependency is found, make it zero and return true
            own_rob_dep = 0;
            --numRobDep;
            assert(numRobDep >= 0);
            DPRINTFR(TraceCPUData, "\tFor %lli: Marking ROB dependency %lli "
                "done.\n", seqNum, rob_dep);
            return true;
        }
    }
    return false;
}

void
TraceCPU::ElasticDataGen::GraphNode::clearRegDep() {
    for (auto& own_reg_dep : regDep) {
        own_reg_dep = 0;
    }
    numRegDep = 0;
}

void
TraceCPU::ElasticDataGen::GraphNode::clearRobDep() {
    for (auto& own_rob_dep : robDep) {
        own_rob_dep = 0;
    }
    numRobDep = 0;
}

bool
TraceCPU::ElasticDataGen::GraphNode::removeDepOnInst(NodeSeqNum done_seq_num)
{
    // If it is an rob dependency then remove it
    if (!removeRobDep(done_seq_num)) {
        // If it is not an rob dependency then it must be a register dependency
        // If the register dependency is not found, it violates an assumption
        // and must be caught by assert.
        bool regdep_found M5_VAR_USED = removeRegDep(done_seq_num);
        assert(regdep_found);
    }
    // Return true if the node is dependency free
    return (numRobDep == 0 && numRegDep == 0);
}

void
TraceCPU::ElasticDataGen::GraphNode::writeElementAsTrace() const
{
    DPRINTFR(TraceCPUData, "%lli", seqNum);
    DPRINTFR(TraceCPUData, ",%s", (isLoad ? "True" : "False"));
    DPRINTFR(TraceCPUData, ",%s", (isStore ? "True" : "False"));
    if (isLoad || isStore) {
        DPRINTFR(TraceCPUData, ",%i", addr);
        DPRINTFR(TraceCPUData, ",%i", size);
        DPRINTFR(TraceCPUData, ",%i", flags);
    }
    DPRINTFR(TraceCPUData, ",%lli", compDelay);
    int i = 0;
    DPRINTFR(TraceCPUData, "robDep:");
    while (robDep[i] != 0) {
        DPRINTFR(TraceCPUData, ",%lli", robDep[i]);
        i++;
    }
    i = 0;
    DPRINTFR(TraceCPUData, "regDep:");
    while (regDep[i] != 0) {
        DPRINTFR(TraceCPUData, ",%lli", regDep[i]);
        i++;
    }
    auto child_itr = dependents.begin();
    DPRINTFR(TraceCPUData, "dependents:");
    while (child_itr != dependents.end()) {
        DPRINTFR(TraceCPUData, ":%lli", (*child_itr)->seqNum);
        child_itr++;
    }

    DPRINTFR(TraceCPUData, "\n");
}

TraceCPU::FixedRetryGen::InputStream::InputStream(const std::string& filename)
    : trace(filename)
{
    // Create a protobuf message for the header and read it from the stream
    ProtoMessage::PacketHeader header_msg;
    if (!trace.read(header_msg)) {
        panic("Failed to read packet header from %s\n", filename);

        if (header_msg.tick_freq() != SimClock::Frequency) {
            panic("Trace %s was recorded with a different tick frequency %d\n",
                  header_msg.tick_freq());
        }
    }
}

void
TraceCPU::FixedRetryGen::InputStream::reset()
{
    trace.reset();
}

bool
TraceCPU::FixedRetryGen::InputStream::read(TraceElement* element)
{
    ProtoMessage::Packet pkt_msg;
    if (trace.read(pkt_msg)) {
        element->cmd = pkt_msg.cmd();
        element->addr = pkt_msg.addr();
        element->blocksize = pkt_msg.size();
        element->tick = pkt_msg.tick();
        element->flags = pkt_msg.has_flags() ? pkt_msg.flags() : 0;
        element->pc = pkt_msg.has_pc() ? pkt_msg.pc() : 0;
        return true;
    }

    // We have reached the end of the file
    return false;
}