summaryrefslogtreecommitdiff
path: root/src/arch/x86/decoder.cc
blob: 930c2b951054ca496804935bca45d87febee8187 (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
/*
 * Copyright (c) 2011 Google
 * 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: Gabe Black
 */

#include "arch/x86/decoder.hh"

#include "arch/x86/regs/misc.hh"
#include "base/misc.hh"
#include "base/trace.hh"
#include "base/types.hh"
#include "debug/Decoder.hh"

namespace X86ISA
{

Decoder::State
Decoder::doResetState()
{
    origPC = basePC + offset;
    DPRINTF(Decoder, "Setting origPC to %#x\n", origPC);
    instBytes = &decodePages->lookup(origPC);
    chunkIdx = 0;

    emi.rex = 0;
    emi.legacy = 0;
    emi.vex = 0;

    emi.opcode.type = BadOpcode;
    emi.opcode.op = 0;

    immediateCollected = 0;
    emi.immediate = 0;
    emi.displacement = 0;
    emi.dispSize = 0;

    emi.modRM = 0;
    emi.sib = 0;

    if (instBytes->si) {
        return FromCacheState;
    } else {
        instBytes->chunks.clear();
        return PrefixState;
    }
}

void
Decoder::process()
{
    //This function drives the decoder state machine.

    //Some sanity checks. You shouldn't try to process more bytes if
    //there aren't any, and you shouldn't overwrite an already
    //decoder ExtMachInst.
    assert(!outOfBytes);
    assert(!instDone);

    if (state == ResetState)
        state = doResetState();
    if (state == FromCacheState) {
        state = doFromCacheState();
    } else {
        instBytes->chunks.push_back(fetchChunk);
    }

    //While there's still something to do...
    while (!instDone && !outOfBytes) {
        uint8_t nextByte = getNextByte();
        switch (state) {
          case PrefixState:
            state = doPrefixState(nextByte);
            break;
          case Vex2Of2State:
            state = doVex2Of2State(nextByte);
            break;
          case Vex2Of3State:
            state = doVex2Of3State(nextByte);
            break;
          case Vex3Of3State:
            state = doVex3Of3State(nextByte);
            break;
          case VexOpcodeState:
            state = doVexOpcodeState(nextByte);
            break;
          case OneByteOpcodeState:
            state = doOneByteOpcodeState(nextByte);
            break;
          case TwoByteOpcodeState:
            state = doTwoByteOpcodeState(nextByte);
            break;
          case ThreeByte0F38OpcodeState:
            state = doThreeByte0F38OpcodeState(nextByte);
            break;
          case ThreeByte0F3AOpcodeState:
            state = doThreeByte0F3AOpcodeState(nextByte);
            break;
          case ModRMState:
            state = doModRMState(nextByte);
            break;
          case SIBState:
            state = doSIBState(nextByte);
            break;
          case DisplacementState:
            state = doDisplacementState();
            break;
          case ImmediateState:
            state = doImmediateState();
            break;
          case ErrorState:
            panic("Went to the error state in the decoder.\n");
          default:
            panic("Unrecognized state! %d\n", state);
        }
    }
}

Decoder::State
Decoder::doFromCacheState()
{
    DPRINTF(Decoder, "Looking at cache state.\n");
    if ((fetchChunk & instBytes->masks[chunkIdx]) !=
            instBytes->chunks[chunkIdx]) {
        DPRINTF(Decoder, "Decode cache miss.\n");
        // The chached chunks didn't match what was fetched. Fall back to the
        // predecoder.
        instBytes->chunks[chunkIdx] = fetchChunk;
        instBytes->chunks.resize(chunkIdx + 1);
        instBytes->si = NULL;
        chunkIdx = 0;
        fetchChunk = instBytes->chunks[0];
        offset = origPC % sizeof(MachInst);
        basePC = origPC - offset;
        return PrefixState;
    } else if (chunkIdx == instBytes->chunks.size() - 1) {
        // We matched the cache, so use its value.
        instDone = true;
        offset = instBytes->lastOffset;
        if (offset == sizeof(MachInst))
            outOfBytes = true;
        return ResetState;
    } else {
        // We matched so far, but need to check more chunks.
        chunkIdx++;
        outOfBytes = true;
        return FromCacheState;
    }
}

//Either get a prefix and record it in the ExtMachInst, or send the
//state machine on to get the opcode(s).
Decoder::State
Decoder::doPrefixState(uint8_t nextByte)
{
    uint8_t prefix = Prefixes[nextByte];
    State nextState = PrefixState;
    // REX prefixes are only recognized in 64 bit mode.
    if (prefix == RexPrefix && emi.mode.submode != SixtyFourBitMode)
        prefix = 0;
    if (prefix)
        consumeByte();
    switch(prefix)
    {
        //Operand size override prefixes
      case OperandSizeOverride:
        DPRINTF(Decoder, "Found operand size override prefix.\n");
        emi.legacy.op = true;
        break;
      case AddressSizeOverride:
        DPRINTF(Decoder, "Found address size override prefix.\n");
        emi.legacy.addr = true;
        break;
        //Segment override prefixes
      case CSOverride:
      case DSOverride:
      case ESOverride:
      case FSOverride:
      case GSOverride:
      case SSOverride:
        DPRINTF(Decoder, "Found segment override.\n");
        emi.legacy.seg = prefix;
        break;
      case Lock:
        DPRINTF(Decoder, "Found lock prefix.\n");
        emi.legacy.lock = true;
        break;
      case Rep:
        DPRINTF(Decoder, "Found rep prefix.\n");
        emi.legacy.rep = true;
        break;
      case Repne:
        DPRINTF(Decoder, "Found repne prefix.\n");
        emi.legacy.repne = true;
        break;
      case RexPrefix:
        DPRINTF(Decoder, "Found Rex prefix %#x.\n", nextByte);
        emi.rex = nextByte;
        break;
      case Vex2Prefix:
        DPRINTF(Decoder, "Found VEX two-byte prefix %#x.\n", nextByte);
        emi.vex.present = 1;
        nextState = Vex2Of2State;
        break;
      case Vex3Prefix:
        DPRINTF(Decoder, "Found VEX three-byte prefix %#x.\n", nextByte);
        emi.vex.present = 1;
        nextState = Vex2Of3State;
        break;
      case 0:
        nextState = OneByteOpcodeState;
        break;

      default:
        panic("Unrecognized prefix %#x\n", nextByte);
    }
    return nextState;
}

Decoder::State
Decoder::doVex2Of2State(uint8_t nextByte)
{
    consumeByte();
    Vex2Of2 vex = nextByte;

    emi.rex.r = !vex.r;

    emi.vex.l = vex.l;
    emi.vex.v = ~vex.v;

    switch (vex.p) {
      case 0:
        break;
      case 1:
        emi.legacy.op = 1;
        break;
      case 2:
        emi.legacy.rep = 1;
        break;
      case 3:
        emi.legacy.repne = 1;
        break;
    }

    emi.opcode.type = TwoByteOpcode;

    return VexOpcodeState;
}

Decoder::State
Decoder::doVex2Of3State(uint8_t nextByte)
{
    if (emi.mode.submode != SixtyFourBitMode && bits(nextByte, 7, 6) == 0x3) {
        // This was actually an LDS instruction. Reroute to that path.
        emi.vex.present = 0;
        emi.opcode.type = OneByteOpcode;
        emi.opcode.op = 0xC4;
        return processOpcode(ImmediateTypeOneByte, UsesModRMOneByte,
                             nextByte >= 0xA0 && nextByte <= 0xA3);
    }

    consumeByte();
    Vex2Of3 vex = nextByte;

    emi.rex.r = !vex.r;
    emi.rex.x = !vex.x;
    emi.rex.b = !vex.b;

    switch (vex.m) {
      case 1:
        emi.opcode.type = TwoByteOpcode;
        break;
      case 2:
        emi.opcode.type = ThreeByte0F38Opcode;
        break;
      case 3:
        emi.opcode.type = ThreeByte0F3AOpcode;
        break;
      default:
        // These encodings are reserved. Pretend this was an undefined
        // instruction so the main decoder will behave correctly, and stop
        // trying to interpret bytes.
        emi.opcode.type = TwoByteOpcode;
        emi.opcode.op = 0x0B;
        instDone = true;
        return ResetState;
    }
    return Vex3Of3State;
}

Decoder::State
Decoder::doVex3Of3State(uint8_t nextByte)
{
    if (emi.mode.submode != SixtyFourBitMode && bits(nextByte, 7, 6) == 0x3) {
        // This was actually an LES instruction. Reroute to that path.
        emi.vex.present = 0;
        emi.opcode.type = OneByteOpcode;
        emi.opcode.op = 0xC5;
        return processOpcode(ImmediateTypeOneByte, UsesModRMOneByte,
                             nextByte >= 0xA0 && nextByte <= 0xA3);
    }

    consumeByte();
    Vex3Of3 vex = nextByte;

    emi.rex.w = vex.w;

    emi.vex.l = vex.l;
    emi.vex.v = ~vex.v;

    switch (vex.p) {
      case 0:
        break;
      case 1:
        emi.legacy.op = 1;
        break;
      case 2:
        emi.legacy.rep = 1;
        break;
      case 3:
        emi.legacy.repne = 1;
        break;
    }

    return VexOpcodeState;
}

Decoder::State
Decoder::doVexOpcodeState(uint8_t nextByte)
{
    DPRINTF(Decoder, "Found VEX opcode %#x.\n", nextByte);

    emi.opcode.op = nextByte;

    switch (emi.opcode.type) {
      case TwoByteOpcode:
        return processOpcode(ImmediateTypeTwoByte, UsesModRMTwoByte);
      case ThreeByte0F38Opcode:
        return processOpcode(ImmediateTypeThreeByte0F38,
                             UsesModRMThreeByte0F38);
      case ThreeByte0F3AOpcode:
        return processOpcode(ImmediateTypeThreeByte0F3A,
                             UsesModRMThreeByte0F3A);
      default:
        panic("Unrecognized opcode type %d.\n", emi.opcode.type);
    }
}

// Load the first opcode byte. Determine if there are more opcode bytes, and
// if not, what immediate and/or ModRM is needed.
Decoder::State
Decoder::doOneByteOpcodeState(uint8_t nextByte)
{
    State nextState = ErrorState;
    consumeByte();

    if (nextByte == 0x0f) {
        DPRINTF(Decoder, "Found opcode escape byte %#x.\n", nextByte);
        nextState = TwoByteOpcodeState;
    } else {
        DPRINTF(Decoder, "Found one byte opcode %#x.\n", nextByte);
        emi.opcode.type = OneByteOpcode;
        emi.opcode.op = nextByte;

        nextState = processOpcode(ImmediateTypeOneByte, UsesModRMOneByte,
                                  nextByte >= 0xA0 && nextByte <= 0xA3);
    }
    return nextState;
}

// Load the second opcode byte. Determine if there are more opcode bytes, and
// if not, what immediate and/or ModRM is needed.
Decoder::State
Decoder::doTwoByteOpcodeState(uint8_t nextByte)
{
    State nextState = ErrorState;
    consumeByte();
    if (nextByte == 0x38) {
        nextState = ThreeByte0F38OpcodeState;
        DPRINTF(Decoder, "Found opcode escape byte %#x.\n", nextByte);
    } else if (nextByte == 0x3a) {
        nextState = ThreeByte0F3AOpcodeState;
        DPRINTF(Decoder, "Found opcode escape byte %#x.\n", nextByte);
    } else {
        DPRINTF(Decoder, "Found two byte opcode %#x.\n", nextByte);
        emi.opcode.type = TwoByteOpcode;
        emi.opcode.op = nextByte;

        nextState = processOpcode(ImmediateTypeTwoByte, UsesModRMTwoByte);
    }
    return nextState;
}

// Load the third opcode byte and determine what immediate and/or ModRM is
// needed.
Decoder::State
Decoder::doThreeByte0F38OpcodeState(uint8_t nextByte)
{
    consumeByte();

    DPRINTF(Decoder, "Found three byte 0F38 opcode %#x.\n", nextByte);
    emi.opcode.type = ThreeByte0F38Opcode;
    emi.opcode.op = nextByte;

    return processOpcode(ImmediateTypeThreeByte0F38, UsesModRMThreeByte0F38);
}

// Load the third opcode byte and determine what immediate and/or ModRM is
// needed.
Decoder::State
Decoder::doThreeByte0F3AOpcodeState(uint8_t nextByte)
{
    consumeByte();

    DPRINTF(Decoder, "Found three byte 0F3A opcode %#x.\n", nextByte);
    emi.opcode.type = ThreeByte0F3AOpcode;
    emi.opcode.op = nextByte;

    return processOpcode(ImmediateTypeThreeByte0F3A, UsesModRMThreeByte0F3A);
}

// Generic opcode processing which determines the immediate size, and whether
// or not there's a modrm byte.
Decoder::State
Decoder::processOpcode(ByteTable &immTable, ByteTable &modrmTable,
                       bool addrSizedImm)
{
    State nextState = ErrorState;
    const uint8_t opcode = emi.opcode.op;

    //Figure out the effective operand size. This can be overriden to
    //a fixed value at the decoder level.
    int logOpSize;
    if (emi.rex.w)
        logOpSize = 3; // 64 bit operand size
    else if (emi.legacy.op)
        logOpSize = altOp;
    else
        logOpSize = defOp;

    //Set the actual op size
    emi.opSize = 1 << logOpSize;

    //Figure out the effective address size. This can be overriden to
    //a fixed value at the decoder level.
    int logAddrSize;
    if (emi.legacy.addr)
        logAddrSize = altAddr;
    else
        logAddrSize = defAddr;

    //Set the actual address size
    emi.addrSize = 1 << logAddrSize;

    //Figure out the effective stack width. This can be overriden to
    //a fixed value at the decoder level.
    emi.stackSize = 1 << stack;

    //Figure out how big of an immediate we'll retreive based
    //on the opcode.
    int immType = immTable[opcode];
    if (addrSizedImm)
        immediateSize = SizeTypeToSize[logAddrSize - 1][immType];
    else
        immediateSize = SizeTypeToSize[logOpSize - 1][immType];

    //Determine what to expect next
    if (modrmTable[opcode]) {
        nextState = ModRMState;
    } else {
        if (immediateSize) {
            nextState = ImmediateState;
        } else {
            instDone = true;
            nextState = ResetState;
        }
    }
    return nextState;
}

//Get the ModRM byte and determine what displacement, if any, there is.
//Also determine whether or not to get the SIB byte, displacement, or
//immediate next.
Decoder::State
Decoder::doModRMState(uint8_t nextByte)
{
    State nextState = ErrorState;
    ModRM modRM = nextByte;
    DPRINTF(Decoder, "Found modrm byte %#x.\n", nextByte);
    if (defOp == 1) {
        //figure out 16 bit displacement size
        if ((modRM.mod == 0 && modRM.rm == 6) || modRM.mod == 2)
            displacementSize = 2;
        else if (modRM.mod == 1)
            displacementSize = 1;
        else
            displacementSize = 0;
    } else {
        //figure out 32/64 bit displacement size
        if ((modRM.mod == 0 && modRM.rm == 5) || modRM.mod == 2)
            displacementSize = 4;
        else if (modRM.mod == 1)
            displacementSize = 1;
        else
            displacementSize = 0;
    }

    // The "test" instruction in group 3 needs an immediate, even though
    // the other instructions with the same actual opcode don't.
    if (emi.opcode.type == OneByteOpcode && (modRM.reg & 0x6) == 0) {
       if (emi.opcode.op == 0xF6)
           immediateSize = 1;
       else if (emi.opcode.op == 0xF7)
           immediateSize = (emi.opSize == 8) ? 4 : emi.opSize;
    }

    //If there's an SIB, get that next.
    //There is no SIB in 16 bit mode.
    if (modRM.rm == 4 && modRM.mod != 3) {
            // && in 32/64 bit mode)
        nextState = SIBState;
    } else if (displacementSize) {
        nextState = DisplacementState;
    } else if (immediateSize) {
        nextState = ImmediateState;
    } else {
        instDone = true;
        nextState = ResetState;
    }
    //The ModRM byte is consumed no matter what
    consumeByte();
    emi.modRM = modRM;
    return nextState;
}

//Get the SIB byte. We don't do anything with it at this point, other
//than storing it in the ExtMachInst. Determine if we need to get a
//displacement or immediate next.
Decoder::State
Decoder::doSIBState(uint8_t nextByte)
{
    State nextState = ErrorState;
    emi.sib = nextByte;
    DPRINTF(Decoder, "Found SIB byte %#x.\n", nextByte);
    consumeByte();
    if (emi.modRM.mod == 0 && emi.sib.base == 5)
        displacementSize = 4;
    if (displacementSize) {
        nextState = DisplacementState;
    } else if (immediateSize) {
        nextState = ImmediateState;
    } else {
        instDone = true;
        nextState = ResetState;
    }
    return nextState;
}

//Gather up the displacement, or at least as much of it
//as we can get.
Decoder::State
Decoder::doDisplacementState()
{
    State nextState = ErrorState;

    getImmediate(immediateCollected,
            emi.displacement,
            displacementSize);

    DPRINTF(Decoder, "Collecting %d byte displacement, got %d bytes.\n",
            displacementSize, immediateCollected);

    if (displacementSize == immediateCollected) {
        //Reset this for other immediates.
        immediateCollected = 0;
        //Sign extend the displacement
        switch(displacementSize)
        {
          case 1:
            emi.displacement = sext<8>(emi.displacement);
            break;
          case 2:
            emi.displacement = sext<16>(emi.displacement);
            break;
          case 4:
            emi.displacement = sext<32>(emi.displacement);
            break;
          default:
            panic("Undefined displacement size!\n");
        }
        DPRINTF(Decoder, "Collected displacement %#x.\n",
                emi.displacement);
        if (immediateSize) {
            nextState = ImmediateState;
        } else {
            instDone = true;
            nextState = ResetState;
        }

        emi.dispSize = displacementSize;
    }
    else
        nextState = DisplacementState;
    return nextState;
}

//Gather up the immediate, or at least as much of it
//as we can get
Decoder::State
Decoder::doImmediateState()
{
    State nextState = ErrorState;

    getImmediate(immediateCollected,
            emi.immediate,
            immediateSize);

    DPRINTF(Decoder, "Collecting %d byte immediate, got %d bytes.\n",
            immediateSize, immediateCollected);

    if (immediateSize == immediateCollected)
    {
        //Reset this for other immediates.
        immediateCollected = 0;

        //XXX Warning! The following is an observed pattern and might
        //not always be true!

        //Instructions which use 64 bit operands but 32 bit immediates
        //need to have the immediate sign extended to 64 bits.
        //Instructions which use true 64 bit immediates won't be
        //affected, and instructions that use true 32 bit immediates
        //won't notice.
        switch(immediateSize)
        {
          case 4:
            emi.immediate = sext<32>(emi.immediate);
            break;
          case 1:
            emi.immediate = sext<8>(emi.immediate);
        }

        DPRINTF(Decoder, "Collected immediate %#x.\n",
                emi.immediate);
        instDone = true;
        nextState = ResetState;
    }
    else
        nextState = ImmediateState;
    return nextState;
}

Decoder::InstBytes Decoder::dummy;
Decoder::InstCacheMap Decoder::instCacheMap;

StaticInstPtr
Decoder::decode(ExtMachInst mach_inst, Addr addr)
{
    DecodeCache::InstMap::iterator iter = instMap->find(mach_inst);
    if (iter != instMap->end())
        return iter->second;

    StaticInstPtr si = decodeInst(mach_inst);
    (*instMap)[mach_inst] = si;
    return si;
}

StaticInstPtr
Decoder::decode(PCState &nextPC)
{
    if (!instDone)
        return NULL;
    instDone = false;
    updateNPC(nextPC);

    StaticInstPtr &si = instBytes->si;
    if (si)
        return si;

    // We didn't match in the AddrMap, but we still populated an entry. Fix
    // up its byte masks.
    const int chunkSize = sizeof(MachInst);

    instBytes->lastOffset = offset;

    Addr firstBasePC = basePC - (instBytes->chunks.size() - 1) * chunkSize;
    Addr firstOffset = origPC - firstBasePC;
    Addr totalSize = instBytes->lastOffset - firstOffset +
        (instBytes->chunks.size() - 1) * chunkSize;
    int start = firstOffset;
    instBytes->masks.clear();

    while (totalSize) {
        int end = start + totalSize;
        end = (chunkSize < end) ? chunkSize : end;
        int size = end - start;
        int idx = instBytes->masks.size();

        MachInst maskVal = mask(size * 8) << (start * 8);
        assert(maskVal);

        instBytes->masks.push_back(maskVal);
        instBytes->chunks[idx] &= instBytes->masks[idx];
        totalSize -= size;
        start = 0;
    }

    si = decode(emi, origPC);
    return si;
}

}