summaryrefslogtreecommitdiff
path: root/src/cpu/simple/atomic.cc
blob: 3001241fe390e9c599621e5c2d6eef127b6a9b9b (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
/*
 * Copyright (c) 2002-2005 The Regents of The University of Michigan
 * All rights reserved.
 *
 * 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: Steve Reinhardt
 */

#include "arch/locked_mem.hh"
#include "arch/mmaped_ipr.hh"
#include "arch/utility.hh"
#include "base/bigint.hh"
#include "cpu/exetrace.hh"
#include "cpu/simple/atomic.hh"
#include "mem/packet.hh"
#include "mem/packet_access.hh"
#include "sim/builder.hh"
#include "sim/system.hh"

using namespace std;
using namespace TheISA;

AtomicSimpleCPU::TickEvent::TickEvent(AtomicSimpleCPU *c)
    : Event(&mainEventQueue, CPU_Tick_Pri), cpu(c)
{
}


void
AtomicSimpleCPU::TickEvent::process()
{
    cpu->tick();
}

const char *
AtomicSimpleCPU::TickEvent::description()
{
    return "AtomicSimpleCPU tick event";
}

Port *
AtomicSimpleCPU::getPort(const std::string &if_name, int idx)
{
    if (if_name == "dcache_port")
        return &dcachePort;
    else if (if_name == "icache_port")
        return &icachePort;
    else
        panic("No Such Port\n");
}

void
AtomicSimpleCPU::init()
{
    BaseCPU::init();
#if FULL_SYSTEM
    for (int i = 0; i < threadContexts.size(); ++i) {
        ThreadContext *tc = threadContexts[i];

        // initialize CPU, including PC
        TheISA::initCPU(tc, tc->readCpuId());
    }
#endif
}

bool
AtomicSimpleCPU::CpuPort::recvTiming(PacketPtr pkt)
{
    panic("AtomicSimpleCPU doesn't expect recvTiming callback!");
    return true;
}

Tick
AtomicSimpleCPU::CpuPort::recvAtomic(PacketPtr pkt)
{
    //Snooping a coherence request, just return
    return 0;
}

void
AtomicSimpleCPU::CpuPort::recvFunctional(PacketPtr pkt)
{
    //No internal storage to update, just return
    return;
}

void
AtomicSimpleCPU::CpuPort::recvStatusChange(Status status)
{
    if (status == RangeChange) {
        if (!snoopRangeSent) {
            snoopRangeSent = true;
            sendStatusChange(Port::RangeChange);
        }
        return;
    }

    panic("AtomicSimpleCPU doesn't expect recvStatusChange callback!");
}

void
AtomicSimpleCPU::CpuPort::recvRetry()
{
    panic("AtomicSimpleCPU doesn't expect recvRetry callback!");
}


AtomicSimpleCPU::AtomicSimpleCPU(Params *p)
    : BaseSimpleCPU(p), tickEvent(this),
      width(p->width), simulate_stalls(p->simulate_stalls),
      icachePort(name() + "-iport", this), dcachePort(name() + "-iport", this)
{
    _status = Idle;

    icachePort.snoopRangeSent = false;
    dcachePort.snoopRangeSent = false;

    ifetch_req = new Request();
    ifetch_req->setThreadContext(p->cpu_id, 0); // Add thread ID if we add MT
    ifetch_pkt = new Packet(ifetch_req, MemCmd::ReadReq, Packet::Broadcast);
    ifetch_pkt->dataStatic(&inst);

    data_read_req = new Request();
    data_read_req->setThreadContext(p->cpu_id, 0); // Add thread ID here too
    data_read_pkt = new Packet(data_read_req, MemCmd::ReadReq,
                               Packet::Broadcast);
    data_read_pkt->dataStatic(&dataReg);

    data_write_req = new Request();
    data_write_req->setThreadContext(p->cpu_id, 0); // Add thread ID here too
    data_write_pkt = new Packet(data_write_req, MemCmd::WriteReq,
                                Packet::Broadcast);
    data_swap_pkt = new Packet(data_write_req, MemCmd::SwapReq,
                                Packet::Broadcast);
}


AtomicSimpleCPU::~AtomicSimpleCPU()
{
}

void
AtomicSimpleCPU::serialize(ostream &os)
{
    SimObject::State so_state = SimObject::getState();
    SERIALIZE_ENUM(so_state);
    Status _status = status();
    SERIALIZE_ENUM(_status);
    BaseSimpleCPU::serialize(os);
    nameOut(os, csprintf("%s.tickEvent", name()));
    tickEvent.serialize(os);
}

void
AtomicSimpleCPU::unserialize(Checkpoint *cp, const string &section)
{
    SimObject::State so_state;
    UNSERIALIZE_ENUM(so_state);
    UNSERIALIZE_ENUM(_status);
    BaseSimpleCPU::unserialize(cp, section);
    tickEvent.unserialize(cp, csprintf("%s.tickEvent", section));
}

void
AtomicSimpleCPU::resume()
{
    if (_status != SwitchedOut && _status != Idle) {
        assert(system->getMemoryMode() == System::Atomic);

        changeState(SimObject::Running);
        if (thread->status() == ThreadContext::Active) {
            if (!tickEvent.scheduled()) {
                tickEvent.schedule(nextCycle());
            }
        }
    }
}

void
AtomicSimpleCPU::switchOut()
{
    assert(status() == Running || status() == Idle);
    _status = SwitchedOut;

    tickEvent.squash();
}


void
AtomicSimpleCPU::takeOverFrom(BaseCPU *oldCPU)
{
    BaseCPU::takeOverFrom(oldCPU);

    assert(!tickEvent.scheduled());

    // if any of this CPU's ThreadContexts are active, mark the CPU as
    // running and schedule its tick event.
    for (int i = 0; i < threadContexts.size(); ++i) {
        ThreadContext *tc = threadContexts[i];
        if (tc->status() == ThreadContext::Active && _status != Running) {
            _status = Running;
            tickEvent.schedule(nextCycle());
            break;
        }
    }
    if (_status != Running) {
        _status = Idle;
    }
}


void
AtomicSimpleCPU::activateContext(int thread_num, int delay)
{
    assert(thread_num == 0);
    assert(thread);

    assert(_status == Idle);
    assert(!tickEvent.scheduled());

    notIdleFraction++;

#if FULL_SYSTEM
    // Connect the ThreadContext's memory ports (Functional/Virtual
    // Ports)
    tc->connectMemPorts();
#endif

    //Make sure ticks are still on multiples of cycles
    tickEvent.schedule(nextCycle(curTick + cycles(delay)));
    _status = Running;
}


void
AtomicSimpleCPU::suspendContext(int thread_num)
{
    assert(thread_num == 0);
    assert(thread);

    assert(_status == Running);

    // tick event may not be scheduled if this gets called from inside
    // an instruction's execution, e.g. "quiesce"
    if (tickEvent.scheduled())
        tickEvent.deschedule();

    notIdleFraction--;
    _status = Idle;
}


template <class T>
Fault
AtomicSimpleCPU::read(Addr addr, T &data, unsigned flags)
{
    // use the CPU's statically allocated read request and packet objects
    Request *req = data_read_req;
    PacketPtr pkt = data_read_pkt;

    req->setVirt(0, addr, sizeof(T), flags, thread->readPC());

    if (traceData) {
        traceData->setAddr(addr);
    }

    // translate to physical address
    Fault fault = thread->translateDataReadReq(req);

    // Now do the access.
    if (fault == NoFault) {
        pkt->reinitFromRequest();

        if (req->isMmapedIpr())
            dcache_latency = TheISA::handleIprRead(thread->getTC(),pkt);
        else
            dcache_latency = dcachePort.sendAtomic(pkt);
        dcache_access = true;
#if !defined(NDEBUG)
        if (pkt->result != Packet::Success)
            panic("Unable to find responder for address pa = %#X va = %#X\n",
                    pkt->req->getPaddr(), pkt->req->getVaddr());
#endif
        data = pkt->get<T>();

        if (req->isLocked()) {
            TheISA::handleLockedRead(thread, req);
        }
    }

    // This will need a new way to tell if it has a dcache attached.
    if (req->isUncacheable())
        recordEvent("Uncached Read");

    return fault;
}

#ifndef DOXYGEN_SHOULD_SKIP_THIS

template
Fault
AtomicSimpleCPU::read(Addr addr, Twin64_t &data, unsigned flags);

template
Fault
AtomicSimpleCPU::read(Addr addr, uint64_t &data, unsigned flags);

template
Fault
AtomicSimpleCPU::read(Addr addr, uint32_t &data, unsigned flags);

template
Fault
AtomicSimpleCPU::read(Addr addr, uint16_t &data, unsigned flags);

template
Fault
AtomicSimpleCPU::read(Addr addr, uint8_t &data, unsigned flags);

#endif //DOXYGEN_SHOULD_SKIP_THIS

template<>
Fault
AtomicSimpleCPU::read(Addr addr, double &data, unsigned flags)
{
    return read(addr, *(uint64_t*)&data, flags);
}

template<>
Fault
AtomicSimpleCPU::read(Addr addr, float &data, unsigned flags)
{
    return read(addr, *(uint32_t*)&data, flags);
}


template<>
Fault
AtomicSimpleCPU::read(Addr addr, int32_t &data, unsigned flags)
{
    return read(addr, (uint32_t&)data, flags);
}


template <class T>
Fault
AtomicSimpleCPU::write(T data, Addr addr, unsigned flags, uint64_t *res)
{
    // use the CPU's statically allocated write request and packet objects
    Request *req = data_write_req;
    PacketPtr pkt;

    req->setVirt(0, addr, sizeof(T), flags, thread->readPC());

    if (req->isSwap())
        pkt = data_swap_pkt;
    else
        pkt = data_write_pkt;

    if (traceData) {
        traceData->setAddr(addr);
    }

    // translate to physical address
    Fault fault = thread->translateDataWriteReq(req);

    // Now do the access.
    if (fault == NoFault) {
        bool do_access = true;  // flag to suppress cache access

        if (req->isLocked()) {
            do_access = TheISA::handleLockedWrite(thread, req);
        }
        if (req->isCondSwap()) {
             assert(res);
             req->setExtraData(*res);
        }


        if (do_access) {
            pkt->reinitFromRequest();
            pkt->dataStatic(&data);

            if (req->isMmapedIpr()) {
                dcache_latency = TheISA::handleIprWrite(thread->getTC(), pkt);
            } else {
                data = htog(data);
                dcache_latency = dcachePort.sendAtomic(pkt);
            }
            dcache_access = true;

#if !defined(NDEBUG)
            if (pkt->result != Packet::Success)
                panic("Unable to find responder for address pa = %#X va = %#X\n",
                        pkt->req->getPaddr(), pkt->req->getVaddr());
#endif
        }

        if (req->isSwap()) {
            assert(res);
            *res = pkt->get<T>();
        } else if (res) {
            *res = req->getExtraData();
        }
    }

    // This will need a new way to tell if it's hooked up to a cache or not.
    if (req->isUncacheable())
        recordEvent("Uncached Write");

    // If the write needs to have a fault on the access, consider calling
    // changeStatus() and changing it to "bad addr write" or something.
    return fault;
}


#ifndef DOXYGEN_SHOULD_SKIP_THIS
template
Fault
AtomicSimpleCPU::write(uint64_t data, Addr addr,
                       unsigned flags, uint64_t *res);

template
Fault
AtomicSimpleCPU::write(uint32_t data, Addr addr,
                       unsigned flags, uint64_t *res);

template
Fault
AtomicSimpleCPU::write(uint16_t data, Addr addr,
                       unsigned flags, uint64_t *res);

template
Fault
AtomicSimpleCPU::write(uint8_t data, Addr addr,
                       unsigned flags, uint64_t *res);

#endif //DOXYGEN_SHOULD_SKIP_THIS

template<>
Fault
AtomicSimpleCPU::write(double data, Addr addr, unsigned flags, uint64_t *res)
{
    return write(*(uint64_t*)&data, addr, flags, res);
}

template<>
Fault
AtomicSimpleCPU::write(float data, Addr addr, unsigned flags, uint64_t *res)
{
    return write(*(uint32_t*)&data, addr, flags, res);
}


template<>
Fault
AtomicSimpleCPU::write(int32_t data, Addr addr, unsigned flags, uint64_t *res)
{
    return write((uint32_t)data, addr, flags, res);
}


void
AtomicSimpleCPU::tick()
{
    Tick latency = cycles(1); // instruction takes one cycle by default

    for (int i = 0; i < width; ++i) {
        numCycles++;

        if (!curStaticInst || !curStaticInst->isDelayedCommit())
            checkForInterrupts();

        Fault fault = setupFetchRequest(ifetch_req);

        if (fault == NoFault) {
            ifetch_pkt->reinitFromRequest();

            Tick icache_latency = icachePort.sendAtomic(ifetch_pkt);
            // ifetch_req is initialized to read the instruction directly
            // into the CPU object's inst field.

            dcache_access = false; // assume no dcache access
            preExecute();

            fault = curStaticInst->execute(this, traceData);
            postExecute();

            // @todo remove me after debugging with legion done
            if (curStaticInst && (!curStaticInst->isMicroOp() ||
                        curStaticInst->isFirstMicroOp()))
                instCnt++;

            if (simulate_stalls) {
                Tick icache_stall = icache_latency - cycles(1);
                Tick dcache_stall =
                    dcache_access ? dcache_latency - cycles(1) : 0;
                Tick stall_cycles = (icache_stall + dcache_stall) / cycles(1);
                if (cycles(stall_cycles) < (icache_stall + dcache_stall))
                    latency += cycles(stall_cycles+1);
                else
                    latency += cycles(stall_cycles);
            }

        }

        advancePC(fault);
    }

    if (_status != Idle)
        tickEvent.schedule(curTick + latency);
}


////////////////////////////////////////////////////////////////////////
//
//  AtomicSimpleCPU Simulation Object
//
BEGIN_DECLARE_SIM_OBJECT_PARAMS(AtomicSimpleCPU)

    Param<Counter> max_insts_any_thread;
    Param<Counter> max_insts_all_threads;
    Param<Counter> max_loads_any_thread;
    Param<Counter> max_loads_all_threads;
    Param<Tick> progress_interval;
    SimObjectParam<System *> system;
    Param<int> cpu_id;

#if FULL_SYSTEM
    SimObjectParam<TheISA::ITB *> itb;
    SimObjectParam<TheISA::DTB *> dtb;
    Param<Tick> profile;

    Param<bool> do_quiesce;
    Param<bool> do_checkpoint_insts;
    Param<bool> do_statistics_insts;
#else
    SimObjectParam<Process *> workload;
#endif // FULL_SYSTEM

    Param<int> clock;
    Param<int> phase;

    Param<bool> defer_registration;
    Param<int> width;
    Param<bool> function_trace;
    Param<Tick> function_trace_start;
    Param<bool> simulate_stalls;

END_DECLARE_SIM_OBJECT_PARAMS(AtomicSimpleCPU)

BEGIN_INIT_SIM_OBJECT_PARAMS(AtomicSimpleCPU)

    INIT_PARAM(max_insts_any_thread,
               "terminate when any thread reaches this inst count"),
    INIT_PARAM(max_insts_all_threads,
               "terminate when all threads have reached this inst count"),
    INIT_PARAM(max_loads_any_thread,
               "terminate when any thread reaches this load count"),
    INIT_PARAM(max_loads_all_threads,
               "terminate when all threads have reached this load count"),
    INIT_PARAM(progress_interval, "Progress interval"),
    INIT_PARAM(system, "system object"),
    INIT_PARAM(cpu_id, "processor ID"),

#if FULL_SYSTEM
    INIT_PARAM(itb, "Instruction TLB"),
    INIT_PARAM(dtb, "Data TLB"),
    INIT_PARAM(profile, ""),
    INIT_PARAM(do_quiesce, ""),
    INIT_PARAM(do_checkpoint_insts, ""),
    INIT_PARAM(do_statistics_insts, ""),
#else
    INIT_PARAM(workload, "processes to run"),
#endif // FULL_SYSTEM

    INIT_PARAM(clock, "clock speed"),
    INIT_PARAM_DFLT(phase, "clock phase", 0),
    INIT_PARAM(defer_registration, "defer system registration (for sampling)"),
    INIT_PARAM(width, "cpu width"),
    INIT_PARAM(function_trace, "Enable function trace"),
    INIT_PARAM(function_trace_start, "Cycle to start function trace"),
    INIT_PARAM(simulate_stalls, "Simulate cache stall cycles")

END_INIT_SIM_OBJECT_PARAMS(AtomicSimpleCPU)


CREATE_SIM_OBJECT(AtomicSimpleCPU)
{
    AtomicSimpleCPU::Params *params = new AtomicSimpleCPU::Params();
    params->name = getInstanceName();
    params->numberOfThreads = 1;
    params->max_insts_any_thread = max_insts_any_thread;
    params->max_insts_all_threads = max_insts_all_threads;
    params->max_loads_any_thread = max_loads_any_thread;
    params->max_loads_all_threads = max_loads_all_threads;
    params->progress_interval = progress_interval;
    params->deferRegistration = defer_registration;
    params->phase = phase;
    params->clock = clock;
    params->functionTrace = function_trace;
    params->functionTraceStart = function_trace_start;
    params->width = width;
    params->simulate_stalls = simulate_stalls;
    params->system = system;
    params->cpu_id = cpu_id;

#if FULL_SYSTEM
    params->itb = itb;
    params->dtb = dtb;
    params->profile = profile;
    params->do_quiesce = do_quiesce;
    params->do_checkpoint_insts = do_checkpoint_insts;
    params->do_statistics_insts = do_statistics_insts;
#else
    params->process = workload;
#endif

    AtomicSimpleCPU *cpu = new AtomicSimpleCPU(params);
    return cpu;
}

REGISTER_SIM_OBJECT("AtomicSimpleCPU", AtomicSimpleCPU)