aboutsummaryrefslogtreecommitdiff
path: root/disas/libvixl/a64/assembler-a64.h
blob: 1e2947b2834a141be96998c3062219e1c64e4f39 (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
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
// Copyright 2013, ARM Limited
// 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 ARM Limited 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 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 VIXL_A64_ASSEMBLER_A64_H_
#define VIXL_A64_ASSEMBLER_A64_H_

#include <list>

#include "globals.h"
#include "utils.h"
#include "a64/instructions-a64.h"

namespace vixl {

typedef uint64_t RegList;
static const int kRegListSizeInBits = sizeof(RegList) * 8;


// Registers.

// Some CPURegister methods can return Register and FPRegister types, so we
// need to declare them in advance.
class Register;
class FPRegister;


class CPURegister {
 public:
  enum RegisterType {
    // The kInvalid value is used to detect uninitialized static instances,
    // which are always zero-initialized before any constructors are called.
    kInvalid = 0,
    kRegister,
    kFPRegister,
    kNoRegister
  };

  CPURegister() : code_(0), size_(0), type_(kNoRegister) {
    VIXL_ASSERT(!IsValid());
    VIXL_ASSERT(IsNone());
  }

  CPURegister(unsigned code, unsigned size, RegisterType type)
      : code_(code), size_(size), type_(type) {
    VIXL_ASSERT(IsValidOrNone());
  }

  unsigned code() const {
    VIXL_ASSERT(IsValid());
    return code_;
  }

  RegisterType type() const {
    VIXL_ASSERT(IsValidOrNone());
    return type_;
  }

  RegList Bit() const {
    VIXL_ASSERT(code_ < (sizeof(RegList) * 8));
    return IsValid() ? (static_cast<RegList>(1) << code_) : 0;
  }

  unsigned size() const {
    VIXL_ASSERT(IsValid());
    return size_;
  }

  int SizeInBytes() const {
    VIXL_ASSERT(IsValid());
    VIXL_ASSERT(size() % 8 == 0);
    return size_ / 8;
  }

  int SizeInBits() const {
    VIXL_ASSERT(IsValid());
    return size_;
  }

  bool Is32Bits() const {
    VIXL_ASSERT(IsValid());
    return size_ == 32;
  }

  bool Is64Bits() const {
    VIXL_ASSERT(IsValid());
    return size_ == 64;
  }

  bool IsValid() const {
    if (IsValidRegister() || IsValidFPRegister()) {
      VIXL_ASSERT(!IsNone());
      return true;
    } else {
      VIXL_ASSERT(IsNone());
      return false;
    }
  }

  bool IsValidRegister() const {
    return IsRegister() &&
           ((size_ == kWRegSize) || (size_ == kXRegSize)) &&
           ((code_ < kNumberOfRegisters) || (code_ == kSPRegInternalCode));
  }

  bool IsValidFPRegister() const {
    return IsFPRegister() &&
           ((size_ == kSRegSize) || (size_ == kDRegSize)) &&
           (code_ < kNumberOfFPRegisters);
  }

  bool IsNone() const {
    // kNoRegister types should always have size 0 and code 0.
    VIXL_ASSERT((type_ != kNoRegister) || (code_ == 0));
    VIXL_ASSERT((type_ != kNoRegister) || (size_ == 0));

    return type_ == kNoRegister;
  }

  bool Aliases(const CPURegister& other) const {
    VIXL_ASSERT(IsValidOrNone() && other.IsValidOrNone());
    return (code_ == other.code_) && (type_ == other.type_);
  }

  bool Is(const CPURegister& other) const {
    VIXL_ASSERT(IsValidOrNone() && other.IsValidOrNone());
    return Aliases(other) && (size_ == other.size_);
  }

  inline bool IsZero() const {
    VIXL_ASSERT(IsValid());
    return IsRegister() && (code_ == kZeroRegCode);
  }

  inline bool IsSP() const {
    VIXL_ASSERT(IsValid());
    return IsRegister() && (code_ == kSPRegInternalCode);
  }

  inline bool IsRegister() const {
    return type_ == kRegister;
  }

  inline bool IsFPRegister() const {
    return type_ == kFPRegister;
  }

  const Register& W() const;
  const Register& X() const;
  const FPRegister& S() const;
  const FPRegister& D() const;

  inline bool IsSameSizeAndType(const CPURegister& other) const {
    return (size_ == other.size_) && (type_ == other.type_);
  }

 protected:
  unsigned code_;
  unsigned size_;
  RegisterType type_;

 private:
  bool IsValidOrNone() const {
    return IsValid() || IsNone();
  }
};


class Register : public CPURegister {
 public:
  explicit Register() : CPURegister() {}
  inline explicit Register(const CPURegister& other)
      : CPURegister(other.code(), other.size(), other.type()) {
    VIXL_ASSERT(IsValidRegister());
  }
  explicit Register(unsigned code, unsigned size)
      : CPURegister(code, size, kRegister) {}

  bool IsValid() const {
    VIXL_ASSERT(IsRegister() || IsNone());
    return IsValidRegister();
  }

  static const Register& WRegFromCode(unsigned code);
  static const Register& XRegFromCode(unsigned code);

  // V8 compatibility.
  static const int kNumRegisters = kNumberOfRegisters;
  static const int kNumAllocatableRegisters = kNumberOfRegisters - 1;

 private:
  static const Register wregisters[];
  static const Register xregisters[];
};


class FPRegister : public CPURegister {
 public:
  inline FPRegister() : CPURegister() {}
  inline explicit FPRegister(const CPURegister& other)
      : CPURegister(other.code(), other.size(), other.type()) {
    VIXL_ASSERT(IsValidFPRegister());
  }
  inline FPRegister(unsigned code, unsigned size)
      : CPURegister(code, size, kFPRegister) {}

  bool IsValid() const {
    VIXL_ASSERT(IsFPRegister() || IsNone());
    return IsValidFPRegister();
  }

  static const FPRegister& SRegFromCode(unsigned code);
  static const FPRegister& DRegFromCode(unsigned code);

  // V8 compatibility.
  static const int kNumRegisters = kNumberOfFPRegisters;
  static const int kNumAllocatableRegisters = kNumberOfFPRegisters - 1;

 private:
  static const FPRegister sregisters[];
  static const FPRegister dregisters[];
};


// No*Reg is used to indicate an unused argument, or an error case. Note that
// these all compare equal (using the Is() method). The Register and FPRegister
// variants are provided for convenience.
const Register NoReg;
const FPRegister NoFPReg;
const CPURegister NoCPUReg;


#define DEFINE_REGISTERS(N)  \
const Register w##N(N, kWRegSize);  \
const Register x##N(N, kXRegSize);
REGISTER_CODE_LIST(DEFINE_REGISTERS)
#undef DEFINE_REGISTERS
const Register wsp(kSPRegInternalCode, kWRegSize);
const Register sp(kSPRegInternalCode, kXRegSize);


#define DEFINE_FPREGISTERS(N)  \
const FPRegister s##N(N, kSRegSize);  \
const FPRegister d##N(N, kDRegSize);
REGISTER_CODE_LIST(DEFINE_FPREGISTERS)
#undef DEFINE_FPREGISTERS


// Registers aliases.
const Register ip0 = x16;
const Register ip1 = x17;
const Register lr = x30;
const Register xzr = x31;
const Register wzr = w31;


// AreAliased returns true if any of the named registers overlap. Arguments
// set to NoReg are ignored. The system stack pointer may be specified.
bool AreAliased(const CPURegister& reg1,
                const CPURegister& reg2,
                const CPURegister& reg3 = NoReg,
                const CPURegister& reg4 = NoReg,
                const CPURegister& reg5 = NoReg,
                const CPURegister& reg6 = NoReg,
                const CPURegister& reg7 = NoReg,
                const CPURegister& reg8 = NoReg);


// AreSameSizeAndType returns true if all of the specified registers have the
// same size, and are of the same type. The system stack pointer may be
// specified. Arguments set to NoReg are ignored, as are any subsequent
// arguments. At least one argument (reg1) must be valid (not NoCPUReg).
bool AreSameSizeAndType(const CPURegister& reg1,
                        const CPURegister& reg2,
                        const CPURegister& reg3 = NoCPUReg,
                        const CPURegister& reg4 = NoCPUReg,
                        const CPURegister& reg5 = NoCPUReg,
                        const CPURegister& reg6 = NoCPUReg,
                        const CPURegister& reg7 = NoCPUReg,
                        const CPURegister& reg8 = NoCPUReg);


// Lists of registers.
class CPURegList {
 public:
  inline explicit CPURegList(CPURegister reg1,
                             CPURegister reg2 = NoCPUReg,
                             CPURegister reg3 = NoCPUReg,
                             CPURegister reg4 = NoCPUReg)
      : list_(reg1.Bit() | reg2.Bit() | reg3.Bit() | reg4.Bit()),
        size_(reg1.size()), type_(reg1.type()) {
    VIXL_ASSERT(AreSameSizeAndType(reg1, reg2, reg3, reg4));
    VIXL_ASSERT(IsValid());
  }

  inline CPURegList(CPURegister::RegisterType type, unsigned size, RegList list)
      : list_(list), size_(size), type_(type) {
    VIXL_ASSERT(IsValid());
  }

  inline CPURegList(CPURegister::RegisterType type, unsigned size,
                    unsigned first_reg, unsigned last_reg)
      : size_(size), type_(type) {
    VIXL_ASSERT(((type == CPURegister::kRegister) &&
                 (last_reg < kNumberOfRegisters)) ||
                ((type == CPURegister::kFPRegister) &&
                 (last_reg < kNumberOfFPRegisters)));
    VIXL_ASSERT(last_reg >= first_reg);
    list_ = (UINT64_C(1) << (last_reg + 1)) - 1;
    list_ &= ~((UINT64_C(1) << first_reg) - 1);
    VIXL_ASSERT(IsValid());
  }

  inline CPURegister::RegisterType type() const {
    VIXL_ASSERT(IsValid());
    return type_;
  }

  // Combine another CPURegList into this one. Registers that already exist in
  // this list are left unchanged. The type and size of the registers in the
  // 'other' list must match those in this list.
  void Combine(const CPURegList& other) {
    VIXL_ASSERT(IsValid());
    VIXL_ASSERT(other.type() == type_);
    VIXL_ASSERT(other.RegisterSizeInBits() == size_);
    list_ |= other.list();
  }

  // Remove every register in the other CPURegList from this one. Registers that
  // do not exist in this list are ignored. The type and size of the registers
  // in the 'other' list must match those in this list.
  void Remove(const CPURegList& other) {
    VIXL_ASSERT(IsValid());
    VIXL_ASSERT(other.type() == type_);
    VIXL_ASSERT(other.RegisterSizeInBits() == size_);
    list_ &= ~other.list();
  }

  // Variants of Combine and Remove which take a single register.
  inline void Combine(const CPURegister& other) {
    VIXL_ASSERT(other.type() == type_);
    VIXL_ASSERT(other.size() == size_);
    Combine(other.code());
  }

  inline void Remove(const CPURegister& other) {
    VIXL_ASSERT(other.type() == type_);
    VIXL_ASSERT(other.size() == size_);
    Remove(other.code());
  }

  // Variants of Combine and Remove which take a single register by its code;
  // the type and size of the register is inferred from this list.
  inline void Combine(int code) {
    VIXL_ASSERT(IsValid());
    VIXL_ASSERT(CPURegister(code, size_, type_).IsValid());
    list_ |= (UINT64_C(1) << code);
  }

  inline void Remove(int code) {
    VIXL_ASSERT(IsValid());
    VIXL_ASSERT(CPURegister(code, size_, type_).IsValid());
    list_ &= ~(UINT64_C(1) << code);
  }

  inline RegList list() const {
    VIXL_ASSERT(IsValid());
    return list_;
  }

  inline void set_list(RegList new_list) {
    VIXL_ASSERT(IsValid());
    list_ = new_list;
  }

  // Remove all callee-saved registers from the list. This can be useful when
  // preparing registers for an AAPCS64 function call, for example.
  void RemoveCalleeSaved();

  CPURegister PopLowestIndex();
  CPURegister PopHighestIndex();

  // AAPCS64 callee-saved registers.
  static CPURegList GetCalleeSaved(unsigned size = kXRegSize);
  static CPURegList GetCalleeSavedFP(unsigned size = kDRegSize);

  // AAPCS64 caller-saved registers. Note that this includes lr.
  static CPURegList GetCallerSaved(unsigned size = kXRegSize);
  static CPURegList GetCallerSavedFP(unsigned size = kDRegSize);

  inline bool IsEmpty() const {
    VIXL_ASSERT(IsValid());
    return list_ == 0;
  }

  inline bool IncludesAliasOf(const CPURegister& other) const {
    VIXL_ASSERT(IsValid());
    return (type_ == other.type()) && ((other.Bit() & list_) != 0);
  }

  inline bool IncludesAliasOf(int code) const {
    VIXL_ASSERT(IsValid());
    return ((code & list_) != 0);
  }

  inline int Count() const {
    VIXL_ASSERT(IsValid());
    return CountSetBits(list_, kRegListSizeInBits);
  }

  inline unsigned RegisterSizeInBits() const {
    VIXL_ASSERT(IsValid());
    return size_;
  }

  inline unsigned RegisterSizeInBytes() const {
    int size_in_bits = RegisterSizeInBits();
    VIXL_ASSERT((size_in_bits % 8) == 0);
    return size_in_bits / 8;
  }

  inline unsigned TotalSizeInBytes() const {
    VIXL_ASSERT(IsValid());
    return RegisterSizeInBytes() * Count();
  }

 private:
  RegList list_;
  unsigned size_;
  CPURegister::RegisterType type_;

  bool IsValid() const;
};


// AAPCS64 callee-saved registers.
extern const CPURegList kCalleeSaved;
extern const CPURegList kCalleeSavedFP;


// AAPCS64 caller-saved registers. Note that this includes lr.
extern const CPURegList kCallerSaved;
extern const CPURegList kCallerSavedFP;


// Operand.
class Operand {
 public:
  // #<immediate>
  // where <immediate> is int64_t.
  // This is allowed to be an implicit constructor because Operand is
  // a wrapper class that doesn't normally perform any type conversion.
  Operand(int64_t immediate);           // NOLINT(runtime/explicit)

  // rm, {<shift> #<shift_amount>}
  // where <shift> is one of {LSL, LSR, ASR, ROR}.
  //       <shift_amount> is uint6_t.
  // This is allowed to be an implicit constructor because Operand is
  // a wrapper class that doesn't normally perform any type conversion.
  Operand(Register reg,
          Shift shift = LSL,
          unsigned shift_amount = 0);   // NOLINT(runtime/explicit)

  // rm, {<extend> {#<shift_amount>}}
  // where <extend> is one of {UXTB, UXTH, UXTW, UXTX, SXTB, SXTH, SXTW, SXTX}.
  //       <shift_amount> is uint2_t.
  explicit Operand(Register reg, Extend extend, unsigned shift_amount = 0);

  bool IsImmediate() const;
  bool IsShiftedRegister() const;
  bool IsExtendedRegister() const;
  bool IsZero() const;

  // This returns an LSL shift (<= 4) operand as an equivalent extend operand,
  // which helps in the encoding of instructions that use the stack pointer.
  Operand ToExtendedRegister() const;

  int64_t immediate() const {
    VIXL_ASSERT(IsImmediate());
    return immediate_;
  }

  Register reg() const {
    VIXL_ASSERT(IsShiftedRegister() || IsExtendedRegister());
    return reg_;
  }

  Shift shift() const {
    VIXL_ASSERT(IsShiftedRegister());
    return shift_;
  }

  Extend extend() const {
    VIXL_ASSERT(IsExtendedRegister());
    return extend_;
  }

  unsigned shift_amount() const {
    VIXL_ASSERT(IsShiftedRegister() || IsExtendedRegister());
    return shift_amount_;
  }

 private:
  int64_t immediate_;
  Register reg_;
  Shift shift_;
  Extend extend_;
  unsigned shift_amount_;
};


// MemOperand represents the addressing mode of a load or store instruction.
class MemOperand {
 public:
  explicit MemOperand(Register base,
                      ptrdiff_t offset = 0,
                      AddrMode addrmode = Offset);
  explicit MemOperand(Register base,
                      Register regoffset,
                      Shift shift = LSL,
                      unsigned shift_amount = 0);
  explicit MemOperand(Register base,
                      Register regoffset,
                      Extend extend,
                      unsigned shift_amount = 0);
  explicit MemOperand(Register base,
                      const Operand& offset,
                      AddrMode addrmode = Offset);

  const Register& base() const { return base_; }
  const Register& regoffset() const { return regoffset_; }
  ptrdiff_t offset() const { return offset_; }
  AddrMode addrmode() const { return addrmode_; }
  Shift shift() const { return shift_; }
  Extend extend() const { return extend_; }
  unsigned shift_amount() const { return shift_amount_; }
  bool IsImmediateOffset() const;
  bool IsRegisterOffset() const;
  bool IsPreIndex() const;
  bool IsPostIndex() const;

 private:
  Register base_;
  Register regoffset_;
  ptrdiff_t offset_;
  AddrMode addrmode_;
  Shift shift_;
  Extend extend_;
  unsigned shift_amount_;
};


class Label {
 public:
  Label() : is_bound_(false), link_(NULL), target_(NULL) {}
  ~Label() {
    // If the label has been linked to, it needs to be bound to a target.
    VIXL_ASSERT(!IsLinked() || IsBound());
  }

  inline Instruction* link() const { return link_; }
  inline Instruction* target() const { return target_; }

  inline bool IsBound() const { return is_bound_; }
  inline bool IsLinked() const { return link_ != NULL; }

  inline void set_link(Instruction* new_link) { link_ = new_link; }

  static const int kEndOfChain = 0;

 private:
  // Indicates if the label has been bound, ie its location is fixed.
  bool is_bound_;
  // Branches instructions branching to this label form a chained list, with
  // their offset indicating where the next instruction is located.
  // link_ points to the latest branch instruction generated branching to this
  // branch.
  // If link_ is not NULL, the label has been linked to.
  Instruction* link_;
  // The label location.
  Instruction* target_;

  friend class Assembler;
};


// TODO: Obtain better values for these, based on real-world data.
const int kLiteralPoolCheckInterval = 4 * KBytes;
const int kRecommendedLiteralPoolRange = 2 * kLiteralPoolCheckInterval;


// Control whether a branch over the literal pool should also be emitted. This
// is needed if the literal pool has to be emitted in the middle of the JITted
// code.
enum LiteralPoolEmitOption {
  JumpRequired,
  NoJumpRequired
};


// Literal pool entry.
class Literal {
 public:
  Literal(Instruction* pc, uint64_t imm, unsigned size)
      : pc_(pc), value_(imm), size_(size) {}

 private:
  Instruction* pc_;
  int64_t value_;
  unsigned size_;

  friend class Assembler;
};


// Assembler.
class Assembler {
 public:
  Assembler(byte* buffer, unsigned buffer_size);

  // The destructor asserts that one of the following is true:
  //  * The Assembler object has not been used.
  //  * Nothing has been emitted since the last Reset() call.
  //  * Nothing has been emitted since the last FinalizeCode() call.
  ~Assembler();

  // System functions.

  // Start generating code from the beginning of the buffer, discarding any code
  // and data that has already been emitted into the buffer.
  //
  // In order to avoid any accidental transfer of state, Reset ASSERTs that the
  // constant pool is not blocked.
  void Reset();

  // Finalize a code buffer of generated instructions. This function must be
  // called before executing or copying code from the buffer.
  void FinalizeCode();

  // Label.
  // Bind a label to the current PC.
  void bind(Label* label);
  int UpdateAndGetByteOffsetTo(Label* label);
  inline int UpdateAndGetInstructionOffsetTo(Label* label) {
    VIXL_ASSERT(Label::kEndOfChain == 0);
    return UpdateAndGetByteOffsetTo(label) >> kInstructionSizeLog2;
  }


  // Instruction set functions.

  // Branch / Jump instructions.
  // Branch to register.
  void br(const Register& xn);

  // Branch with link to register.
  void blr(const Register& xn);

  // Branch to register with return hint.
  void ret(const Register& xn = lr);

  // Unconditional branch to label.
  void b(Label* label);

  // Conditional branch to label.
  void b(Label* label, Condition cond);

  // Unconditional branch to PC offset.
  void b(int imm26);

  // Conditional branch to PC offset.
  void b(int imm19, Condition cond);

  // Branch with link to label.
  void bl(Label* label);

  // Branch with link to PC offset.
  void bl(int imm26);

  // Compare and branch to label if zero.
  void cbz(const Register& rt, Label* label);

  // Compare and branch to PC offset if zero.
  void cbz(const Register& rt, int imm19);

  // Compare and branch to label if not zero.
  void cbnz(const Register& rt, Label* label);

  // Compare and branch to PC offset if not zero.
  void cbnz(const Register& rt, int imm19);

  // Test bit and branch to label if zero.
  void tbz(const Register& rt, unsigned bit_pos, Label* label);

  // Test bit and branch to PC offset if zero.
  void tbz(const Register& rt, unsigned bit_pos, int imm14);

  // Test bit and branch to label if not zero.
  void tbnz(const Register& rt, unsigned bit_pos, Label* label);

  // Test bit and branch to PC offset if not zero.
  void tbnz(const Register& rt, unsigned bit_pos, int imm14);

  // Address calculation instructions.
  // Calculate a PC-relative address. Unlike for branches the offset in adr is
  // unscaled (i.e. the result can be unaligned).

  // Calculate the address of a label.
  void adr(const Register& rd, Label* label);

  // Calculate the address of a PC offset.
  void adr(const Register& rd, int imm21);

  // Data Processing instructions.
  // Add.
  void add(const Register& rd,
           const Register& rn,
           const Operand& operand);

  // Add and update status flags.
  void adds(const Register& rd,
            const Register& rn,
            const Operand& operand);

  // Compare negative.
  void cmn(const Register& rn, const Operand& operand);

  // Subtract.
  void sub(const Register& rd,
           const Register& rn,
           const Operand& operand);

  // Subtract and update status flags.
  void subs(const Register& rd,
            const Register& rn,
            const Operand& operand);

  // Compare.
  void cmp(const Register& rn, const Operand& operand);

  // Negate.
  void neg(const Register& rd,
           const Operand& operand);

  // Negate and update status flags.
  void negs(const Register& rd,
            const Operand& operand);

  // Add with carry bit.
  void adc(const Register& rd,
           const Register& rn,
           const Operand& operand);

  // Add with carry bit and update status flags.
  void adcs(const Register& rd,
            const Register& rn,
            const Operand& operand);

  // Subtract with carry bit.
  void sbc(const Register& rd,
           const Register& rn,
           const Operand& operand);

  // Subtract with carry bit and update status flags.
  void sbcs(const Register& rd,
            const Register& rn,
            const Operand& operand);

  // Negate with carry bit.
  void ngc(const Register& rd,
           const Operand& operand);

  // Negate with carry bit and update status flags.
  void ngcs(const Register& rd,
            const Operand& operand);

  // Logical instructions.
  // Bitwise and (A & B).
  void and_(const Register& rd,
            const Register& rn,
            const Operand& operand);

  // Bitwise and (A & B) and update status flags.
  void ands(const Register& rd,
            const Register& rn,
            const Operand& operand);

  // Bit test and set flags.
  void tst(const Register& rn, const Operand& operand);

  // Bit clear (A & ~B).
  void bic(const Register& rd,
           const Register& rn,
           const Operand& operand);

  // Bit clear (A & ~B) and update status flags.
  void bics(const Register& rd,
            const Register& rn,
            const Operand& operand);

  // Bitwise or (A | B).
  void orr(const Register& rd, const Register& rn, const Operand& operand);

  // Bitwise nor (A | ~B).
  void orn(const Register& rd, const Register& rn, const Operand& operand);

  // Bitwise eor/xor (A ^ B).
  void eor(const Register& rd, const Register& rn, const Operand& operand);

  // Bitwise enor/xnor (A ^ ~B).
  void eon(const Register& rd, const Register& rn, const Operand& operand);

  // Logical shift left by variable.
  void lslv(const Register& rd, const Register& rn, const Register& rm);

  // Logical shift right by variable.
  void lsrv(const Register& rd, const Register& rn, const Register& rm);

  // Arithmetic shift right by variable.
  void asrv(const Register& rd, const Register& rn, const Register& rm);

  // Rotate right by variable.
  void rorv(const Register& rd, const Register& rn, const Register& rm);

  // Bitfield instructions.
  // Bitfield move.
  void bfm(const Register& rd,
           const Register& rn,
           unsigned immr,
           unsigned imms);

  // Signed bitfield move.
  void sbfm(const Register& rd,
            const Register& rn,
            unsigned immr,
            unsigned imms);

  // Unsigned bitfield move.
  void ubfm(const Register& rd,
            const Register& rn,
            unsigned immr,
            unsigned imms);

  // Bfm aliases.
  // Bitfield insert.
  inline void bfi(const Register& rd,
                  const Register& rn,
                  unsigned lsb,
                  unsigned width) {
    VIXL_ASSERT(width >= 1);
    VIXL_ASSERT(lsb + width <= rn.size());
    bfm(rd, rn, (rd.size() - lsb) & (rd.size() - 1), width - 1);
  }

  // Bitfield extract and insert low.
  inline void bfxil(const Register& rd,
                    const Register& rn,
                    unsigned lsb,
                    unsigned width) {
    VIXL_ASSERT(width >= 1);
    VIXL_ASSERT(lsb + width <= rn.size());
    bfm(rd, rn, lsb, lsb + width - 1);
  }

  // Sbfm aliases.
  // Arithmetic shift right.
  inline void asr(const Register& rd, const Register& rn, unsigned shift) {
    VIXL_ASSERT(shift < rd.size());
    sbfm(rd, rn, shift, rd.size() - 1);
  }

  // Signed bitfield insert with zero at right.
  inline void sbfiz(const Register& rd,
                    const Register& rn,
                    unsigned lsb,
                    unsigned width) {
    VIXL_ASSERT(width >= 1);
    VIXL_ASSERT(lsb + width <= rn.size());
    sbfm(rd, rn, (rd.size() - lsb) & (rd.size() - 1), width - 1);
  }

  // Signed bitfield extract.
  inline void sbfx(const Register& rd,
                   const Register& rn,
                   unsigned lsb,
                   unsigned width) {
    VIXL_ASSERT(width >= 1);
    VIXL_ASSERT(lsb + width <= rn.size());
    sbfm(rd, rn, lsb, lsb + width - 1);
  }

  // Signed extend byte.
  inline void sxtb(const Register& rd, const Register& rn) {
    sbfm(rd, rn, 0, 7);
  }

  // Signed extend halfword.
  inline void sxth(const Register& rd, const Register& rn) {
    sbfm(rd, rn, 0, 15);
  }

  // Signed extend word.
  inline void sxtw(const Register& rd, const Register& rn) {
    sbfm(rd, rn, 0, 31);
  }

  // Ubfm aliases.
  // Logical shift left.
  inline void lsl(const Register& rd, const Register& rn, unsigned shift) {
    unsigned reg_size = rd.size();
    VIXL_ASSERT(shift < reg_size);
    ubfm(rd, rn, (reg_size - shift) % reg_size, reg_size - shift - 1);
  }

  // Logical shift right.
  inline void lsr(const Register& rd, const Register& rn, unsigned shift) {
    VIXL_ASSERT(shift < rd.size());
    ubfm(rd, rn, shift, rd.size() - 1);
  }

  // Unsigned bitfield insert with zero at right.
  inline void ubfiz(const Register& rd,
                    const Register& rn,
                    unsigned lsb,
                    unsigned width) {
    VIXL_ASSERT(width >= 1);
    VIXL_ASSERT(lsb + width <= rn.size());
    ubfm(rd, rn, (rd.size() - lsb) & (rd.size() - 1), width - 1);
  }

  // Unsigned bitfield extract.
  inline void ubfx(const Register& rd,
                   const Register& rn,
                   unsigned lsb,
                   unsigned width) {
    VIXL_ASSERT(width >= 1);
    VIXL_ASSERT(lsb + width <= rn.size());
    ubfm(rd, rn, lsb, lsb + width - 1);
  }

  // Unsigned extend byte.
  inline void uxtb(const Register& rd, const Register& rn) {
    ubfm(rd, rn, 0, 7);
  }

  // Unsigned extend halfword.
  inline void uxth(const Register& rd, const Register& rn) {
    ubfm(rd, rn, 0, 15);
  }

  // Unsigned extend word.
  inline void uxtw(const Register& rd, const Register& rn) {
    ubfm(rd, rn, 0, 31);
  }

  // Extract.
  void extr(const Register& rd,
            const Register& rn,
            const Register& rm,
            unsigned lsb);

  // Conditional select: rd = cond ? rn : rm.
  void csel(const Register& rd,
            const Register& rn,
            const Register& rm,
            Condition cond);

  // Conditional select increment: rd = cond ? rn : rm + 1.
  void csinc(const Register& rd,
             const Register& rn,
             const Register& rm,
             Condition cond);

  // Conditional select inversion: rd = cond ? rn : ~rm.
  void csinv(const Register& rd,
             const Register& rn,
             const Register& rm,
             Condition cond);

  // Conditional select negation: rd = cond ? rn : -rm.
  void csneg(const Register& rd,
             const Register& rn,
             const Register& rm,
             Condition cond);

  // Conditional set: rd = cond ? 1 : 0.
  void cset(const Register& rd, Condition cond);

  // Conditional set mask: rd = cond ? -1 : 0.
  void csetm(const Register& rd, Condition cond);

  // Conditional increment: rd = cond ? rn + 1 : rn.
  void cinc(const Register& rd, const Register& rn, Condition cond);

  // Conditional invert: rd = cond ? ~rn : rn.
  void cinv(const Register& rd, const Register& rn, Condition cond);

  // Conditional negate: rd = cond ? -rn : rn.
  void cneg(const Register& rd, const Register& rn, Condition cond);

  // Rotate right.
  inline void ror(const Register& rd, const Register& rs, unsigned shift) {
    extr(rd, rs, rs, shift);
  }

  // Conditional comparison.
  // Conditional compare negative.
  void ccmn(const Register& rn,
            const Operand& operand,
            StatusFlags nzcv,
            Condition cond);

  // Conditional compare.
  void ccmp(const Register& rn,
            const Operand& operand,
            StatusFlags nzcv,
            Condition cond);

  // Multiply.
  void mul(const Register& rd, const Register& rn, const Register& rm);

  // Negated multiply.
  void mneg(const Register& rd, const Register& rn, const Register& rm);

  // Signed long multiply: 32 x 32 -> 64-bit.
  void smull(const Register& rd, const Register& rn, const Register& rm);

  // Signed multiply high: 64 x 64 -> 64-bit <127:64>.
  void smulh(const Register& xd, const Register& xn, const Register& xm);

  // Multiply and accumulate.
  void madd(const Register& rd,
            const Register& rn,
            const Register& rm,
            const Register& ra);

  // Multiply and subtract.
  void msub(const Register& rd,
            const Register& rn,
            const Register& rm,
            const Register& ra);

  // Signed long multiply and accumulate: 32 x 32 + 64 -> 64-bit.
  void smaddl(const Register& rd,
              const Register& rn,
              const Register& rm,
              const Register& ra);

  // Unsigned long multiply and accumulate: 32 x 32 + 64 -> 64-bit.
  void umaddl(const Register& rd,
              const Register& rn,
              const Register& rm,
              const Register& ra);

  // Signed long multiply and subtract: 64 - (32 x 32) -> 64-bit.
  void smsubl(const Register& rd,
              const Register& rn,
              const Register& rm,
              const Register& ra);

  // Unsigned long multiply and subtract: 64 - (32 x 32) -> 64-bit.
  void umsubl(const Register& rd,
              const Register& rn,
              const Register& rm,
              const Register& ra);

  // Signed integer divide.
  void sdiv(const Register& rd, const Register& rn, const Register& rm);

  // Unsigned integer divide.
  void udiv(const Register& rd, const Register& rn, const Register& rm);

  // Bit reverse.
  void rbit(const Register& rd, const Register& rn);

  // Reverse bytes in 16-bit half words.
  void rev16(const Register& rd, const Register& rn);

  // Reverse bytes in 32-bit words.
  void rev32(const Register& rd, const Register& rn);

  // Reverse bytes.
  void rev(const Register& rd, const Register& rn);

  // Count leading zeroes.
  void clz(const Register& rd, const Register& rn);

  // Count leading sign bits.
  void cls(const Register& rd, const Register& rn);

  // Memory instructions.
  // Load integer or FP register.
  void ldr(const CPURegister& rt, const MemOperand& src);

  // Store integer or FP register.
  void str(const CPURegister& rt, const MemOperand& dst);

  // Load word with sign extension.
  void ldrsw(const Register& rt, const MemOperand& src);

  // Load byte.
  void ldrb(const Register& rt, const MemOperand& src);

  // Store byte.
  void strb(const Register& rt, const MemOperand& dst);

  // Load byte with sign extension.
  void ldrsb(const Register& rt, const MemOperand& src);

  // Load half-word.
  void ldrh(const Register& rt, const MemOperand& src);

  // Store half-word.
  void strh(const Register& rt, const MemOperand& dst);

  // Load half-word with sign extension.
  void ldrsh(const Register& rt, const MemOperand& src);

  // Load integer or FP register pair.
  void ldp(const CPURegister& rt, const CPURegister& rt2,
           const MemOperand& src);

  // Store integer or FP register pair.
  void stp(const CPURegister& rt, const CPURegister& rt2,
           const MemOperand& dst);

  // Load word pair with sign extension.
  void ldpsw(const Register& rt, const Register& rt2, const MemOperand& src);

  // Load integer or FP register pair, non-temporal.
  void ldnp(const CPURegister& rt, const CPURegister& rt2,
            const MemOperand& src);

  // Store integer or FP register pair, non-temporal.
  void stnp(const CPURegister& rt, const CPURegister& rt2,
            const MemOperand& dst);

  // Load literal to register.
  void ldr(const Register& rt, uint64_t imm);

  // Load double precision floating point literal to FP register.
  void ldr(const FPRegister& ft, double imm);

  // Load single precision floating point literal to FP register.
  void ldr(const FPRegister& ft, float imm);

  // Move instructions. The default shift of -1 indicates that the move
  // instruction will calculate an appropriate 16-bit immediate and left shift
  // that is equal to the 64-bit immediate argument. If an explicit left shift
  // is specified (0, 16, 32 or 48), the immediate must be a 16-bit value.
  //
  // For movk, an explicit shift can be used to indicate which half word should
  // be overwritten, eg. movk(x0, 0, 0) will overwrite the least-significant
  // half word with zero, whereas movk(x0, 0, 48) will overwrite the
  // most-significant.

  // Move immediate and keep.
  void movk(const Register& rd, uint64_t imm, int shift = -1) {
    MoveWide(rd, imm, shift, MOVK);
  }

  // Move inverted immediate.
  void movn(const Register& rd, uint64_t imm, int shift = -1) {
    MoveWide(rd, imm, shift, MOVN);
  }

  // Move immediate.
  void movz(const Register& rd, uint64_t imm, int shift = -1) {
    MoveWide(rd, imm, shift, MOVZ);
  }

  // Misc instructions.
  // Monitor debug-mode breakpoint.
  void brk(int code);

  // Halting debug-mode breakpoint.
  void hlt(int code);

  // Move register to register.
  void mov(const Register& rd, const Register& rn);

  // Move inverted operand to register.
  void mvn(const Register& rd, const Operand& operand);

  // System instructions.
  // Move to register from system register.
  void mrs(const Register& rt, SystemRegister sysreg);

  // Move from register to system register.
  void msr(SystemRegister sysreg, const Register& rt);

  // System hint.
  void hint(SystemHint code);

  // Data memory barrier.
  void dmb(BarrierDomain domain, BarrierType type);

  // Data synchronization barrier.
  void dsb(BarrierDomain domain, BarrierType type);

  // Instruction synchronization barrier.
  void isb();

  // Alias for system instructions.
  // No-op.
  void nop() {
    hint(NOP);
  }

  // FP instructions.
  // Move double precision immediate to FP register.
  void fmov(const FPRegister& fd, double imm);

  // Move single precision immediate to FP register.
  void fmov(const FPRegister& fd, float imm);

  // Move FP register to register.
  void fmov(const Register& rd, const FPRegister& fn);

  // Move register to FP register.
  void fmov(const FPRegister& fd, const Register& rn);

  // Move FP register to FP register.
  void fmov(const FPRegister& fd, const FPRegister& fn);

  // FP add.
  void fadd(const FPRegister& fd, const FPRegister& fn, const FPRegister& fm);

  // FP subtract.
  void fsub(const FPRegister& fd, const FPRegister& fn, const FPRegister& fm);

  // FP multiply.
  void fmul(const FPRegister& fd, const FPRegister& fn, const FPRegister& fm);

  // FP fused multiply and add.
  void fmadd(const FPRegister& fd,
             const FPRegister& fn,
             const FPRegister& fm,
             const FPRegister& fa);

  // FP fused multiply and subtract.
  void fmsub(const FPRegister& fd,
             const FPRegister& fn,
             const FPRegister& fm,
             const FPRegister& fa);

  // FP fused multiply, add and negate.
  void fnmadd(const FPRegister& fd,
              const FPRegister& fn,
              const FPRegister& fm,
              const FPRegister& fa);

  // FP fused multiply, subtract and negate.
  void fnmsub(const FPRegister& fd,
              const FPRegister& fn,
              const FPRegister& fm,
              const FPRegister& fa);

  // FP divide.
  void fdiv(const FPRegister& fd, const FPRegister& fn, const FPRegister& fm);

  // FP maximum.
  void fmax(const FPRegister& fd, const FPRegister& fn, const FPRegister& fm);

  // FP minimum.
  void fmin(const FPRegister& fd, const FPRegister& fn, const FPRegister& fm);

  // FP maximum number.
  void fmaxnm(const FPRegister& fd, const FPRegister& fn, const FPRegister& fm);

  // FP minimum number.
  void fminnm(const FPRegister& fd, const FPRegister& fn, const FPRegister& fm);

  // FP absolute.
  void fabs(const FPRegister& fd, const FPRegister& fn);

  // FP negate.
  void fneg(const FPRegister& fd, const FPRegister& fn);

  // FP square root.
  void fsqrt(const FPRegister& fd, const FPRegister& fn);

  // FP round to integer (nearest with ties to away).
  void frinta(const FPRegister& fd, const FPRegister& fn);

  // FP round to integer (toward minus infinity).
  void frintm(const FPRegister& fd, const FPRegister& fn);

  // FP round to integer (nearest with ties to even).
  void frintn(const FPRegister& fd, const FPRegister& fn);

  // FP round to integer (towards zero).
  void frintz(const FPRegister& fd, const FPRegister& fn);

  // FP compare registers.
  void fcmp(const FPRegister& fn, const FPRegister& fm);

  // FP compare immediate.
  void fcmp(const FPRegister& fn, double value);

  // FP conditional compare.
  void fccmp(const FPRegister& fn,
             const FPRegister& fm,
             StatusFlags nzcv,
             Condition cond);

  // FP conditional select.
  void fcsel(const FPRegister& fd,
             const FPRegister& fn,
             const FPRegister& fm,
             Condition cond);

  // Common FP Convert function.
  void FPConvertToInt(const Register& rd,
                      const FPRegister& fn,
                      FPIntegerConvertOp op);

  // FP convert between single and double precision.
  void fcvt(const FPRegister& fd, const FPRegister& fn);

  // Convert FP to signed integer (nearest with ties to away).
  void fcvtas(const Register& rd, const FPRegister& fn);

  // Convert FP to unsigned integer (nearest with ties to away).
  void fcvtau(const Register& rd, const FPRegister& fn);

  // Convert FP to signed integer (round towards -infinity).
  void fcvtms(const Register& rd, const FPRegister& fn);

  // Convert FP to unsigned integer (round towards -infinity).
  void fcvtmu(const Register& rd, const FPRegister& fn);

  // Convert FP to signed integer (nearest with ties to even).
  void fcvtns(const Register& rd, const FPRegister& fn);

  // Convert FP to unsigned integer (nearest with ties to even).
  void fcvtnu(const Register& rd, const FPRegister& fn);

  // Convert FP to signed integer (round towards zero).
  void fcvtzs(const Register& rd, const FPRegister& fn);

  // Convert FP to unsigned integer (round towards zero).
  void fcvtzu(const Register& rd, const FPRegister& fn);

  // Convert signed integer or fixed point to FP.
  void scvtf(const FPRegister& fd, const Register& rn, unsigned fbits = 0);

  // Convert unsigned integer or fixed point to FP.
  void ucvtf(const FPRegister& fd, const Register& rn, unsigned fbits = 0);

  // Emit generic instructions.
  // Emit raw instructions into the instruction stream.
  inline void dci(Instr raw_inst) { Emit(raw_inst); }

  // Emit 32 bits of data into the instruction stream.
  inline void dc32(uint32_t data) { EmitData(&data, sizeof(data)); }

  // Emit 64 bits of data into the instruction stream.
  inline void dc64(uint64_t data) { EmitData(&data, sizeof(data)); }

  // Copy a string into the instruction stream, including the terminating NULL
  // character. The instruction pointer (pc_) is then aligned correctly for
  // subsequent instructions.
  void EmitStringData(const char * string) {
    VIXL_ASSERT(string != NULL);

    size_t len = strlen(string) + 1;
    EmitData(string, len);

    // Pad with NULL characters until pc_ is aligned.
    const char pad[] = {'\0', '\0', '\0', '\0'};
    VIXL_STATIC_ASSERT(sizeof(pad) == kInstructionSize);
    Instruction* next_pc = AlignUp(pc_, kInstructionSize);
    EmitData(&pad, next_pc - pc_);
  }

  // Code generation helpers.

  // Register encoding.
  static Instr Rd(CPURegister rd) {
    VIXL_ASSERT(rd.code() != kSPRegInternalCode);
    return rd.code() << Rd_offset;
  }

  static Instr Rn(CPURegister rn) {
    VIXL_ASSERT(rn.code() != kSPRegInternalCode);
    return rn.code() << Rn_offset;
  }

  static Instr Rm(CPURegister rm) {
    VIXL_ASSERT(rm.code() != kSPRegInternalCode);
    return rm.code() << Rm_offset;
  }

  static Instr Ra(CPURegister ra) {
    VIXL_ASSERT(ra.code() != kSPRegInternalCode);
    return ra.code() << Ra_offset;
  }

  static Instr Rt(CPURegister rt) {
    VIXL_ASSERT(rt.code() != kSPRegInternalCode);
    return rt.code() << Rt_offset;
  }

  static Instr Rt2(CPURegister rt2) {
    VIXL_ASSERT(rt2.code() != kSPRegInternalCode);
    return rt2.code() << Rt2_offset;
  }

  // These encoding functions allow the stack pointer to be encoded, and
  // disallow the zero register.
  static Instr RdSP(Register rd) {
    VIXL_ASSERT(!rd.IsZero());
    return (rd.code() & kRegCodeMask) << Rd_offset;
  }

  static Instr RnSP(Register rn) {
    VIXL_ASSERT(!rn.IsZero());
    return (rn.code() & kRegCodeMask) << Rn_offset;
  }

  // Flags encoding.
  static Instr Flags(FlagsUpdate S) {
    if (S == SetFlags) {
      return 1 << FlagsUpdate_offset;
    } else if (S == LeaveFlags) {
      return 0 << FlagsUpdate_offset;
    }
    VIXL_UNREACHABLE();
    return 0;
  }

  static Instr Cond(Condition cond) {
    return cond << Condition_offset;
  }

  // PC-relative address encoding.
  static Instr ImmPCRelAddress(int imm21) {
    VIXL_ASSERT(is_int21(imm21));
    Instr imm = static_cast<Instr>(truncate_to_int21(imm21));
    Instr immhi = (imm >> ImmPCRelLo_width) << ImmPCRelHi_offset;
    Instr immlo = imm << ImmPCRelLo_offset;
    return (immhi & ImmPCRelHi_mask) | (immlo & ImmPCRelLo_mask);
  }

  // Branch encoding.
  static Instr ImmUncondBranch(int imm26) {
    VIXL_ASSERT(is_int26(imm26));
    return truncate_to_int26(imm26) << ImmUncondBranch_offset;
  }

  static Instr ImmCondBranch(int imm19) {
    VIXL_ASSERT(is_int19(imm19));
    return truncate_to_int19(imm19) << ImmCondBranch_offset;
  }

  static Instr ImmCmpBranch(int imm19) {
    VIXL_ASSERT(is_int19(imm19));
    return truncate_to_int19(imm19) << ImmCmpBranch_offset;
  }

  static Instr ImmTestBranch(int imm14) {
    VIXL_ASSERT(is_int14(imm14));
    return truncate_to_int14(imm14) << ImmTestBranch_offset;
  }

  static Instr ImmTestBranchBit(unsigned bit_pos) {
    VIXL_ASSERT(is_uint6(bit_pos));
    // Subtract five from the shift offset, as we need bit 5 from bit_pos.
    unsigned b5 = bit_pos << (ImmTestBranchBit5_offset - 5);
    unsigned b40 = bit_pos << ImmTestBranchBit40_offset;
    b5 &= ImmTestBranchBit5_mask;
    b40 &= ImmTestBranchBit40_mask;
    return b5 | b40;
  }

  // Data Processing encoding.
  static Instr SF(Register rd) {
      return rd.Is64Bits() ? SixtyFourBits : ThirtyTwoBits;
  }

  static Instr ImmAddSub(int64_t imm) {
    VIXL_ASSERT(IsImmAddSub(imm));
    if (is_uint12(imm)) {  // No shift required.
      return imm << ImmAddSub_offset;
    } else {
      return ((imm >> 12) << ImmAddSub_offset) | (1 << ShiftAddSub_offset);
    }
  }

  static inline Instr ImmS(unsigned imms, unsigned reg_size) {
    VIXL_ASSERT(((reg_size == kXRegSize) && is_uint6(imms)) ||
           ((reg_size == kWRegSize) && is_uint5(imms)));
    USE(reg_size);
    return imms << ImmS_offset;
  }

  static inline Instr ImmR(unsigned immr, unsigned reg_size) {
    VIXL_ASSERT(((reg_size == kXRegSize) && is_uint6(immr)) ||
           ((reg_size == kWRegSize) && is_uint5(immr)));
    USE(reg_size);
    VIXL_ASSERT(is_uint6(immr));
    return immr << ImmR_offset;
  }

  static inline Instr ImmSetBits(unsigned imms, unsigned reg_size) {
    VIXL_ASSERT((reg_size == kWRegSize) || (reg_size == kXRegSize));
    VIXL_ASSERT(is_uint6(imms));
    VIXL_ASSERT((reg_size == kXRegSize) || is_uint6(imms + 3));
    USE(reg_size);
    return imms << ImmSetBits_offset;
  }

  static inline Instr ImmRotate(unsigned immr, unsigned reg_size) {
    VIXL_ASSERT((reg_size == kWRegSize) || (reg_size == kXRegSize));
    VIXL_ASSERT(((reg_size == kXRegSize) && is_uint6(immr)) ||
           ((reg_size == kWRegSize) && is_uint5(immr)));
    USE(reg_size);
    return immr << ImmRotate_offset;
  }

  static inline Instr ImmLLiteral(int imm19) {
    VIXL_ASSERT(is_int19(imm19));
    return truncate_to_int19(imm19) << ImmLLiteral_offset;
  }

  static inline Instr BitN(unsigned bitn, unsigned reg_size) {
    VIXL_ASSERT((reg_size == kWRegSize) || (reg_size == kXRegSize));
    VIXL_ASSERT((reg_size == kXRegSize) || (bitn == 0));
    USE(reg_size);
    return bitn << BitN_offset;
  }

  static Instr ShiftDP(Shift shift) {
    VIXL_ASSERT(shift == LSL || shift == LSR || shift == ASR || shift == ROR);
    return shift << ShiftDP_offset;
  }

  static Instr ImmDPShift(unsigned amount) {
    VIXL_ASSERT(is_uint6(amount));
    return amount << ImmDPShift_offset;
  }

  static Instr ExtendMode(Extend extend) {
    return extend << ExtendMode_offset;
  }

  static Instr ImmExtendShift(unsigned left_shift) {
    VIXL_ASSERT(left_shift <= 4);
    return left_shift << ImmExtendShift_offset;
  }

  static Instr ImmCondCmp(unsigned imm) {
    VIXL_ASSERT(is_uint5(imm));
    return imm << ImmCondCmp_offset;
  }

  static Instr Nzcv(StatusFlags nzcv) {
    return ((nzcv >> Flags_offset) & 0xf) << Nzcv_offset;
  }

  // MemOperand offset encoding.
  static Instr ImmLSUnsigned(int imm12) {
    VIXL_ASSERT(is_uint12(imm12));
    return imm12 << ImmLSUnsigned_offset;
  }

  static Instr ImmLS(int imm9) {
    VIXL_ASSERT(is_int9(imm9));
    return truncate_to_int9(imm9) << ImmLS_offset;
  }

  static Instr ImmLSPair(int imm7, LSDataSize size) {
    VIXL_ASSERT(((imm7 >> size) << size) == imm7);
    int scaled_imm7 = imm7 >> size;
    VIXL_ASSERT(is_int7(scaled_imm7));
    return truncate_to_int7(scaled_imm7) << ImmLSPair_offset;
  }

  static Instr ImmShiftLS(unsigned shift_amount) {
    VIXL_ASSERT(is_uint1(shift_amount));
    return shift_amount << ImmShiftLS_offset;
  }

  static Instr ImmException(int imm16) {
    VIXL_ASSERT(is_uint16(imm16));
    return imm16 << ImmException_offset;
  }

  static Instr ImmSystemRegister(int imm15) {
    VIXL_ASSERT(is_uint15(imm15));
    return imm15 << ImmSystemRegister_offset;
  }

  static Instr ImmHint(int imm7) {
    VIXL_ASSERT(is_uint7(imm7));
    return imm7 << ImmHint_offset;
  }

  static Instr ImmBarrierDomain(int imm2) {
    VIXL_ASSERT(is_uint2(imm2));
    return imm2 << ImmBarrierDomain_offset;
  }

  static Instr ImmBarrierType(int imm2) {
    VIXL_ASSERT(is_uint2(imm2));
    return imm2 << ImmBarrierType_offset;
  }

  static LSDataSize CalcLSDataSize(LoadStoreOp op) {
    VIXL_ASSERT((SizeLS_offset + SizeLS_width) == (kInstructionSize * 8));
    return static_cast<LSDataSize>(op >> SizeLS_offset);
  }

  // Move immediates encoding.
  static Instr ImmMoveWide(uint64_t imm) {
    VIXL_ASSERT(is_uint16(imm));
    return imm << ImmMoveWide_offset;
  }

  static Instr ShiftMoveWide(int64_t shift) {
    VIXL_ASSERT(is_uint2(shift));
    return shift << ShiftMoveWide_offset;
  }

  // FP Immediates.
  static Instr ImmFP32(float imm);
  static Instr ImmFP64(double imm);

  // FP register type.
  static Instr FPType(FPRegister fd) {
    return fd.Is64Bits() ? FP64 : FP32;
  }

  static Instr FPScale(unsigned scale) {
    VIXL_ASSERT(is_uint6(scale));
    return scale << FPScale_offset;
  }

  // Size of the code generated in bytes
  uint64_t SizeOfCodeGenerated() const {
    VIXL_ASSERT((pc_ >= buffer_) && (pc_ < (buffer_ + buffer_size_)));
    return pc_ - buffer_;
  }

  // Size of the code generated since label to the current position.
  uint64_t SizeOfCodeGeneratedSince(Label* label) const {
    VIXL_ASSERT(label->IsBound());
    VIXL_ASSERT((pc_ >= label->target()) && (pc_ < (buffer_ + buffer_size_)));
    return pc_ - label->target();
  }


  inline void BlockLiteralPool() {
    literal_pool_monitor_++;
  }

  inline void ReleaseLiteralPool() {
    if (--literal_pool_monitor_ == 0) {
      // Has the literal pool been blocked for too long?
      VIXL_ASSERT(literals_.empty() ||
             (pc_ < (literals_.back()->pc_ + kMaxLoadLiteralRange)));
    }
  }

  inline bool IsLiteralPoolBlocked() {
    return literal_pool_monitor_ != 0;
  }

  void CheckLiteralPool(LiteralPoolEmitOption option = JumpRequired);
  void EmitLiteralPool(LiteralPoolEmitOption option = NoJumpRequired);
  size_t LiteralPoolSize();

 protected:
  inline const Register& AppropriateZeroRegFor(const CPURegister& reg) const {
    return reg.Is64Bits() ? xzr : wzr;
  }


  void LoadStore(const CPURegister& rt,
                 const MemOperand& addr,
                 LoadStoreOp op);
  static bool IsImmLSUnscaled(ptrdiff_t offset);
  static bool IsImmLSScaled(ptrdiff_t offset, LSDataSize size);

  void Logical(const Register& rd,
               const Register& rn,
               const Operand& operand,
               LogicalOp op);
  void LogicalImmediate(const Register& rd,
                        const Register& rn,
                        unsigned n,
                        unsigned imm_s,
                        unsigned imm_r,
                        LogicalOp op);
  static bool IsImmLogical(uint64_t value,
                           unsigned width,
                           unsigned* n,
                           unsigned* imm_s,
                           unsigned* imm_r);

  void ConditionalCompare(const Register& rn,
                          const Operand& operand,
                          StatusFlags nzcv,
                          Condition cond,
                          ConditionalCompareOp op);
  static bool IsImmConditionalCompare(int64_t immediate);

  void AddSubWithCarry(const Register& rd,
                       const Register& rn,
                       const Operand& operand,
                       FlagsUpdate S,
                       AddSubWithCarryOp op);

  static bool IsImmFP32(float imm);
  static bool IsImmFP64(double imm);

  // Functions for emulating operands not directly supported by the instruction
  // set.
  void EmitShift(const Register& rd,
                 const Register& rn,
                 Shift shift,
                 unsigned amount);
  void EmitExtendShift(const Register& rd,
                       const Register& rn,
                       Extend extend,
                       unsigned left_shift);

  void AddSub(const Register& rd,
              const Register& rn,
              const Operand& operand,
              FlagsUpdate S,
              AddSubOp op);
  static bool IsImmAddSub(int64_t immediate);

  // Find an appropriate LoadStoreOp or LoadStorePairOp for the specified
  // registers. Only simple loads are supported; sign- and zero-extension (such
  // as in LDPSW_x or LDRB_w) are not supported.
  static LoadStoreOp LoadOpFor(const CPURegister& rt);
  static LoadStorePairOp LoadPairOpFor(const CPURegister& rt,
                                       const CPURegister& rt2);
  static LoadStoreOp StoreOpFor(const CPURegister& rt);
  static LoadStorePairOp StorePairOpFor(const CPURegister& rt,
                                        const CPURegister& rt2);
  static LoadStorePairNonTemporalOp LoadPairNonTemporalOpFor(
    const CPURegister& rt, const CPURegister& rt2);
  static LoadStorePairNonTemporalOp StorePairNonTemporalOpFor(
    const CPURegister& rt, const CPURegister& rt2);


 private:
  // Instruction helpers.
  void MoveWide(const Register& rd,
                uint64_t imm,
                int shift,
                MoveWideImmediateOp mov_op);
  void DataProcShiftedRegister(const Register& rd,
                               const Register& rn,
                               const Operand& operand,
                               FlagsUpdate S,
                               Instr op);
  void DataProcExtendedRegister(const Register& rd,
                                const Register& rn,
                                const Operand& operand,
                                FlagsUpdate S,
                                Instr op);
  void LoadStorePair(const CPURegister& rt,
                     const CPURegister& rt2,
                     const MemOperand& addr,
                     LoadStorePairOp op);
  void LoadStorePairNonTemporal(const CPURegister& rt,
                                const CPURegister& rt2,
                                const MemOperand& addr,
                                LoadStorePairNonTemporalOp op);
  void LoadLiteral(const CPURegister& rt, uint64_t imm, LoadLiteralOp op);
  void ConditionalSelect(const Register& rd,
                         const Register& rn,
                         const Register& rm,
                         Condition cond,
                         ConditionalSelectOp op);
  void DataProcessing1Source(const Register& rd,
                             const Register& rn,
                             DataProcessing1SourceOp op);
  void DataProcessing3Source(const Register& rd,
                             const Register& rn,
                             const Register& rm,
                             const Register& ra,
                             DataProcessing3SourceOp op);
  void FPDataProcessing1Source(const FPRegister& fd,
                               const FPRegister& fn,
                               FPDataProcessing1SourceOp op);
  void FPDataProcessing2Source(const FPRegister& fd,
                               const FPRegister& fn,
                               const FPRegister& fm,
                               FPDataProcessing2SourceOp op);
  void FPDataProcessing3Source(const FPRegister& fd,
                               const FPRegister& fn,
                               const FPRegister& fm,
                               const FPRegister& fa,
                               FPDataProcessing3SourceOp op);

  void RecordLiteral(int64_t imm, unsigned size);

  // Emit the instruction at pc_.
  void Emit(Instr instruction) {
    VIXL_STATIC_ASSERT(sizeof(*pc_) == 1);
    VIXL_STATIC_ASSERT(sizeof(instruction) == kInstructionSize);
    VIXL_ASSERT((pc_ + sizeof(instruction)) <= (buffer_ + buffer_size_));

#ifdef DEBUG
    finalized_ = false;
#endif

    memcpy(pc_, &instruction, sizeof(instruction));
    pc_ += sizeof(instruction);
    CheckBufferSpace();
  }

  // Emit data inline in the instruction stream.
  void EmitData(void const * data, unsigned size) {
    VIXL_STATIC_ASSERT(sizeof(*pc_) == 1);
    VIXL_ASSERT((pc_ + size) <= (buffer_ + buffer_size_));

#ifdef DEBUG
    finalized_ = false;
#endif

    // TODO: Record this 'instruction' as data, so that it can be disassembled
    // correctly.
    memcpy(pc_, data, size);
    pc_ += size;
    CheckBufferSpace();
  }

  inline void CheckBufferSpace() {
    VIXL_ASSERT(pc_ < (buffer_ + buffer_size_));
    if (pc_ > next_literal_pool_check_) {
      CheckLiteralPool();
    }
  }

  // The buffer into which code and relocation info are generated.
  Instruction* buffer_;
  // Buffer size, in bytes.
  unsigned buffer_size_;
  Instruction* pc_;
  std::list<Literal*> literals_;
  Instruction* next_literal_pool_check_;
  unsigned literal_pool_monitor_;

  friend class BlockLiteralPoolScope;

#ifdef DEBUG
  bool finalized_;
#endif
};

class BlockLiteralPoolScope {
 public:
  explicit BlockLiteralPoolScope(Assembler* assm) : assm_(assm) {
    assm_->BlockLiteralPool();
  }

  ~BlockLiteralPoolScope() {
    assm_->ReleaseLiteralPool();
  }

 private:
  Assembler* assm_;
};
}  // namespace vixl

#endif  // VIXL_A64_ASSEMBLER_A64_H_