summaryrefslogtreecommitdiff
path: root/cpu/ozone/cpu.hh
blob: fa849bb0936eeb4b39dd4d50d0cb1c1a9c53f04e (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
/*
 * Copyright (c) 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.
 */

#ifndef __CPU_OOO_CPU_OOO_CPU_HH__
#define __CPU_OOO_CPU_OOO_CPU_HH__

#include "base/statistics.hh"
#include "config/full_system.hh"
#include "cpu/base.hh"
#include "cpu/exec_context.hh"
#include "encumbered/cpu/full/fu_pool.hh"
#include "cpu/ooo_cpu/ea_list.hh"
#include "cpu/pc_event.hh"
#include "cpu/static_inst.hh"
#include "mem/mem_interface.hh"
#include "sim/eventq.hh"

// forward declarations
#if FULL_SYSTEM
class Processor;
class AlphaITB;
class AlphaDTB;
class PhysicalMemory;

class RemoteGDB;
class GDBListener;

#else

class Process;

#endif // FULL_SYSTEM

class Checkpoint;
class MemInterface;

namespace Trace {
    class InstRecord;
}

/**
 * Declaration of Out-of-Order CPU class.  Basically it is a SimpleCPU with
 * simple out-of-order capabilities added to it.  It is still a 1 CPI machine
 * (?), but is capable of handling cache misses.  Basically it models having
 * a ROB/IQ by only allowing a certain amount of instructions to execute while
 * the cache miss is outstanding.
 */

template <class Impl>
class OoOCPU : public BaseCPU
{
  private:
    typedef typename Impl::DynInst DynInst;
    typedef typename Impl::DynInstPtr DynInstPtr;

  public:
    // main simulation loop (one cycle)
    void tick();

  private:
    struct TickEvent : public Event
    {
        OoOCPU *cpu;
        int width;

        TickEvent(OoOCPU *c, int w);
        void process();
        const char *description();
    };

    TickEvent tickEvent;

    /// Schedule tick event, regardless of its current state.
    void scheduleTickEvent(int delay)
    {
        if (tickEvent.squashed())
            tickEvent.reschedule(curTick + delay);
        else if (!tickEvent.scheduled())
            tickEvent.schedule(curTick + delay);
    }

    /// Unschedule tick event, regardless of its current state.
    void unscheduleTickEvent()
    {
        if (tickEvent.scheduled())
            tickEvent.squash();
    }

  private:
    Trace::InstRecord *traceData;

    template<typename T>
    void trace_data(T data);

  public:
    //
    enum Status {
        Running,
        Idle,
        IcacheMiss,
        IcacheMissComplete,
        DcacheMissStall,
        SwitchedOut
    };

  private:
    Status _status;

  public:
    void post_interrupt(int int_num, int index);

    void zero_fill_64(Addr addr) {
        static int warned = 0;
        if (!warned) {
            warn ("WH64 is not implemented");
            warned = 1;
        }
    };

    struct Params : public BaseCPU::Params
    {
        MemInterface *icache_interface;
        MemInterface *dcache_interface;
        int width;
#if FULL_SYSTEM
        AlphaITB *itb;
        AlphaDTB *dtb;
        FunctionalMemory *mem;
#else
        Process *process;
#endif
        int issueWidth;
    };

    OoOCPU(Params *params);

    virtual ~OoOCPU();

    void init();

  private:
    void copyFromXC();

  public:
    // execution context
    ExecContext *xc;

    void switchOut();
    void takeOverFrom(BaseCPU *oldCPU);

#if FULL_SYSTEM
    Addr dbg_vtophys(Addr addr);

    bool interval_stats;
#endif

    // L1 instruction cache
    MemInterface *icacheInterface;

    // L1 data cache
    MemInterface *dcacheInterface;

    FuncUnitPool *fuPool;

    // Refcounted pointer to the one memory request.
    MemReqPtr cacheMemReq;

    class ICacheCompletionEvent : public Event
    {
      private:
        OoOCPU *cpu;

      public:
        ICacheCompletionEvent(OoOCPU *_cpu);

        virtual void process();
        virtual const char *description();
    };

    // Will need to create a cache completion event upon any memory miss.
    ICacheCompletionEvent iCacheCompletionEvent;

    class DCacheCompletionEvent;

    typedef typename
    std::list<DCacheCompletionEvent>::iterator DCacheCompEventIt;

    class DCacheCompletionEvent : public Event
    {
      private:
        OoOCPU *cpu;
        DynInstPtr inst;
        DCacheCompEventIt dcceIt;

      public:
        DCacheCompletionEvent(OoOCPU *_cpu, DynInstPtr &_inst,
                              DCacheCompEventIt &_dcceIt);

        virtual void process();
        virtual const char *description();
    };

    friend class DCacheCompletionEvent;

  protected:
    std::list<DCacheCompletionEvent> dCacheCompList;
    DCacheCompEventIt dcceIt;

  private:
    Status status() const { return _status; }

    virtual void activateContext(int thread_num, int delay);
    virtual void suspendContext(int thread_num);
    virtual void deallocateContext(int thread_num);
    virtual void haltContext(int thread_num);

    // statistics
    virtual void regStats();
    virtual void resetStats();

    // number of simulated instructions
    Counter numInst;
    Counter startNumInst;
    Stats::Scalar<> numInsts;

    virtual Counter totalInstructions() const
    {
        return numInst - startNumInst;
    }

    // number of simulated memory references
    Stats::Scalar<> numMemRefs;

    // number of simulated loads
    Counter numLoad;
    Counter startNumLoad;

    // number of idle cycles
    Stats::Average<> notIdleFraction;
    Stats::Formula idleFraction;

    // number of cycles stalled for I-cache misses
    Stats::Scalar<> icacheStallCycles;
    Counter lastIcacheStall;

    // number of cycles stalled for D-cache misses
    Stats::Scalar<> dcacheStallCycles;
    Counter lastDcacheStall;

    void processICacheCompletion();

  public:

    virtual void serialize(std::ostream &os);
    virtual void unserialize(Checkpoint *cp, const std::string &section);

#if FULL_SYSTEM
    bool validInstAddr(Addr addr) { return true; }
    bool validDataAddr(Addr addr) { return true; }
    int getInstAsid() { return xc->regs.instAsid(); }
    int getDataAsid() { return xc->regs.dataAsid(); }

    Fault translateInstReq(MemReqPtr &req)
    {
        return itb->translate(req);
    }

    Fault translateDataReadReq(MemReqPtr &req)
    {
        return dtb->translate(req, false);
    }

    Fault translateDataWriteReq(MemReqPtr &req)
    {
        return dtb->translate(req, true);
    }

#else
    bool validInstAddr(Addr addr)
    { return xc->validInstAddr(addr); }

    bool validDataAddr(Addr addr)
    { return xc->validDataAddr(addr); }

    int getInstAsid() { return xc->asid; }
    int getDataAsid() { return xc->asid; }

    Fault dummyTranslation(MemReqPtr &req)
    {
#if 0
        assert((req->vaddr >> 48 & 0xffff) == 0);
#endif

        // put the asid in the upper 16 bits of the paddr
        req->paddr = req->vaddr & ~((Addr)0xffff << sizeof(Addr) * 8 - 16);
        req->paddr = req->paddr | (Addr)req->asid << sizeof(Addr) * 8 - 16;
        return NoFault;
    }
    Fault translateInstReq(MemReqPtr &req)
    {
        return dummyTranslation(req);
    }
    Fault translateDataReadReq(MemReqPtr &req)
    {
        return dummyTranslation(req);
    }
    Fault translateDataWriteReq(MemReqPtr &req)
    {
        return dummyTranslation(req);
    }

#endif

    template <class T>
    Fault read(Addr addr, T &data, unsigned flags, DynInstPtr inst);

    template <class T>
    Fault write(T data, Addr addr, unsigned flags,
                uint64_t *res, DynInstPtr inst);

    void prefetch(Addr addr, unsigned flags)
    {
        // need to do this...
    }

    void writeHint(Addr addr, int size, unsigned flags)
    {
        // need to do this...
    }

    Fault copySrcTranslate(Addr src);

    Fault copy(Addr dest);

  private:
    bool executeInst(DynInstPtr &inst);

    void renameInst(DynInstPtr &inst);

    void addInst(DynInstPtr &inst);

    void commitHeadInst();

    bool getOneInst();

    Fault fetchCacheLine();

    InstSeqNum getAndIncrementInstSeq();

    bool ambigMemAddr;

  private:
    InstSeqNum globalSeqNum;

    DynInstPtr renameTable[TheISA::TotalNumRegs];
    DynInstPtr commitTable[TheISA::TotalNumRegs];

    // Might need a table of the shadow registers as well.
#if FULL_SYSTEM
    DynInstPtr palShadowTable[TheISA::NumIntRegs];
#endif

  public:
    // The register accessor methods provide the index of the
    // instruction's operand (e.g., 0 or 1), not the architectural
    // register index, to simplify the implementation of register
    // renaming.  We find the architectural register index by indexing
    // into the instruction's own operand index table.  Note that a
    // raw pointer to the StaticInst is provided instead of a
    // ref-counted StaticInstPtr to redice overhead.  This is fine as
    // long as these methods don't copy the pointer into any long-term
    // storage (which is pretty hard to imagine they would have reason
    // to do).

    // In the OoO case these shouldn't read from the XC but rather from the
    // rename table of DynInsts.  Also these likely shouldn't be called very
    // often, other than when adding things into the xc during say a syscall.

    uint64_t readIntReg(StaticInst *si, int idx)
    {
        return xc->readIntReg(si->srcRegIdx(idx));
    }

    FloatReg readFloatReg(StaticInst *si, int idx, width)
    {
        int reg_idx = si->srcRegIdx(idx) - TheISA::FP_Base_DepTag;
        return xc->readFloatReg(reg_idx, width);
    }

    FloatReg readFloatReg(StaticInst *si, int idx)
    {
        int reg_idx = si->srcRegIdx(idx) - TheISA::FP_Base_DepTag;
        return xc->readFloatReg(reg_idx);
    }

    FloatRegBits readFloatRegBits(StaticInst *si, int idx, int width)
    {
        int reg_idx = si->srcRegIdx(idx) - TheISA::FP_Base_DepTag;
        return xc->readFloatRegBits(reg_idx, width);
    }

    FloatRegBits readFloatRegBits(StaticInst *si, int idx)
    {
        int reg_idx = si->srcRegIdx(idx) - TheISA::FP_Base_DepTag;
        return xc->readFloatRegBits(reg_idx);
    }

    void setIntReg(StaticInst *si, int idx, uint64_t val)
    {
        xc->setIntReg(si->destRegIdx(idx), val);
    }

    void setFloatReg(StaticInst *si, int idx, FloatReg val, int width)
    {
        int reg_idx = si->destRegIdx(idx) - TheISA::FP_Base_DepTag;
        xc->setFloatReg(reg_idx, val, width);
    }

    void setFloatReg(StaticInst *si, int idx, FloatReg val)
    {
        int reg_idx = si->destRegIdx(idx) - TheISA::FP_Base_DepTag;
        xc->setFloatReg(reg_idx, val);
    }

    void setFloatRegBits(StaticInst *si, int idx, FloatRegBits val, int width)
    {
        int reg_idx = si->destRegIdx(idx) - TheISA::FP_Base_DepTag;
        xc->setFloatRegBits(reg_idx, val, width);
    }

    void setFloatRegBits(StaticInst *si, int idx, FloatRegBits val)
    {
        int reg_idx = si->destRegIdx(idx) - TheISA::FP_Base_DepTag;
        xc->setFloatRegBits(reg_idx, val);
    }

    uint64_t readPC() { return PC; }
    void setNextPC(Addr val) { nextPC = val; }

  private:
    Addr PC;
    Addr nextPC;

    unsigned issueWidth;

    bool fetchRedirExcp;
    bool fetchRedirBranch;

    /** Mask to get a cache block's address. */
    Addr cacheBlkMask;

    unsigned cacheBlkSize;

    Addr cacheBlkPC;

    /** The cache line being fetched. */
    uint8_t *cacheData;

  protected:
    bool cacheBlkValid;

  private:

    // Align an address (typically a PC) to the start of an I-cache block.
    // We fold in the PISA 64- to 32-bit conversion here as well.
    Addr icacheBlockAlignPC(Addr addr)
    {
        addr = TheISA::realPCToFetchPC(addr);
        return (addr & ~(cacheBlkMask));
    }

    unsigned instSize;

    // ROB tracking stuff.
    DynInstPtr robHeadPtr;
    DynInstPtr robTailPtr;
    unsigned robSize;
    unsigned robInsts;

    // List of outstanding EA instructions.
  protected:
    EAList eaList;

  public:
    void branchToTarget(Addr val)
    {
        if (!fetchRedirExcp) {
            fetchRedirBranch = true;
            PC = val;
        }
    }

    // ISA stuff:
    uint64_t readUniq() { return xc->readUniq(); }
    void setUniq(uint64_t val) { xc->setUniq(val); }

    uint64_t readFpcr() { return xc->readFpcr(); }
    void setFpcr(uint64_t val) { xc->setFpcr(val); }

#if FULL_SYSTEM
    uint64_t readIpr(int idx, Fault &fault) { return xc->readIpr(idx, fault); }
    Fault setIpr(int idx, uint64_t val) { return xc->setIpr(idx, val); }
    Fault hwrei() { return xc->hwrei(); }
    int readIntrFlag() { return xc->readIntrFlag(); }
    void setIntrFlag(int val) { xc->setIntrFlag(val); }
    bool inPalMode() { return xc->inPalMode(); }
    void trap(Fault fault) { fault->invoke(xc); }
    bool simPalCheck(int palFunc) { return xc->simPalCheck(palFunc); }
#else
    void syscall() { xc->syscall(); }
#endif

    ExecContext *xcBase() { return xc; }
};


// precise architected memory state accessor macros
template <class Impl>
template <class T>
Fault
OoOCPU<Impl>::read(Addr addr, T &data, unsigned flags, DynInstPtr inst)
{
    MemReqPtr readReq = new MemReq();
    readReq->xc = xc;
    readReq->asid = 0;
    readReq->data = new uint8_t[64];

    readReq->reset(addr, sizeof(T), flags);

    // translate to physical address - This might be an ISA impl call
    Fault fault = translateDataReadReq(readReq);

    // do functional access
    if (fault == NoFault)
        fault = xc->mem->read(readReq, data);
#if 0
    if (traceData) {
        traceData->setAddr(addr);
        if (fault == NoFault)
            traceData->setData(data);
    }
#endif

    // if we have a cache, do cache access too
    if (fault == NoFault && dcacheInterface) {
        readReq->cmd = Read;
        readReq->completionEvent = NULL;
        readReq->time = curTick;
        /*MemAccessResult result = */dcacheInterface->access(readReq);

        if (dcacheInterface->doEvents()) {
            readReq->completionEvent = new DCacheCompletionEvent(this, inst,
                                                                 dcceIt);
        }
    }

    if (!dcacheInterface && (readReq->flags & UNCACHEABLE))
        recordEvent("Uncached Read");

    return fault;
}

template <class Impl>
template <class T>
Fault
OoOCPU<Impl>::write(T data, Addr addr, unsigned flags,
                    uint64_t *res, DynInstPtr inst)
{
    MemReqPtr writeReq = new MemReq();
    writeReq->xc = xc;
    writeReq->asid = 0;
    writeReq->data = new uint8_t[64];

#if 0
    if (traceData) {
        traceData->setAddr(addr);
        traceData->setData(data);
    }
#endif

    writeReq->reset(addr, sizeof(T), flags);

    // translate to physical address
    Fault fault = translateDataWriteReq(writeReq);

    // do functional access
    if (fault == NoFault)
        fault = xc->write(writeReq, data);

    if (fault == NoFault && dcacheInterface) {
        writeReq->cmd = Write;
        memcpy(writeReq->data,(uint8_t *)&data,writeReq->size);
        writeReq->completionEvent = NULL;
        writeReq->time = curTick;
        /*MemAccessResult result = */dcacheInterface->access(writeReq);

        if (dcacheInterface->doEvents()) {
            writeReq->completionEvent = new DCacheCompletionEvent(this, inst,
                                                                  dcceIt);
        }
    }

    if (res && (fault == NoFault))
        *res = writeReq->result;

    if (!dcacheInterface && (writeReq->flags & UNCACHEABLE))
        recordEvent("Uncached Write");

    return fault;
}


#endif // __CPU_OOO_CPU_OOO_CPU_HH__