aboutsummaryrefslogtreecommitdiffstats
path: root/src/3rdparty/masm/assembler/ARM64Assembler.h
blob: 8ec9823f3309421c2ac8a7aa764180d48ce0ac4f (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
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
/*
 * Copyright (C) 2012, 2014 Apple Inc. All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. 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.
 *
 * THIS SOFTWARE IS PROVIDED BY APPLE INC. ``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 APPLE INC. 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 ARM64Assembler_h
#define ARM64Assembler_h

#if ENABLE(ASSEMBLER) && CPU(ARM64)

#include "AssemblerBuffer.h"
#include <limits.h>
#include <wtf/Assertions.h>
#include <wtf/Vector.h>
#include <stdint.h>

#if OS(IOS)
#include <libkern/OSCacheControl.h>
#endif

#define CHECK_DATASIZE_OF(datasize) ASSERT(datasize == 32 || datasize == 64)
#define DATASIZE_OF(datasize) ((datasize == 64) ? Datasize_64 : Datasize_32)
#define MEMOPSIZE_OF(datasize) ((datasize == 8 || datasize == 128) ? MemOpSize_8_or_128 : (datasize == 16) ? MemOpSize_16 : (datasize == 32) ? MemOpSize_32 : MemOpSize_64)
#define CHECK_DATASIZE() CHECK_DATASIZE_OF(datasize)
#define CHECK_VECTOR_DATASIZE() ASSERT(datasize == 64 || datasize == 128)
#define DATASIZE DATASIZE_OF(datasize)
#define MEMOPSIZE MEMOPSIZE_OF(datasize)
#define CHECK_FP_MEMOP_DATASIZE() ASSERT(datasize == 8 || datasize == 16 || datasize == 32 || datasize == 64 || datasize == 128)
#define MEMPAIROPSIZE_INT(datasize) ((datasize == 64) ? MemPairOp_64 : MemPairOp_32)
#define MEMPAIROPSIZE_FP(datasize) ((datasize == 128) ? MemPairOp_V128 : (datasize == 64) ? MemPairOp_V64 : MemPairOp_32)

namespace JSC {

ALWAYS_INLINE bool isInt9(int32_t value)
{
    return value == ((value << 23) >> 23);
}

template<typename Type>
ALWAYS_INLINE bool isUInt12(Type value)
{
    return !(value & ~static_cast<Type>(0xfff));
}

template<int datasize>
ALWAYS_INLINE bool isValidScaledUImm12(int32_t offset)
{
    int32_t maxPImm = 4095 * (datasize / 8);
    if (offset < 0)
        return false;
    if (offset > maxPImm)
        return false;
    if (offset & ((datasize / 8) - 1))
        return false;
    return true;
}

ALWAYS_INLINE bool isValidSignedImm9(int32_t value)
{
    return isInt9(value);
}

ALWAYS_INLINE bool isInt7(int32_t value)
{
    return value == ((value << 25) >> 25);
}

ALWAYS_INLINE bool isInt11(int32_t value)
{
    return value == ((value << 21) >> 21);
}

ALWAYS_INLINE bool isUInt5(int32_t value)
{
    return !(value & ~0x1f);
}

class UInt5 {
public:
    explicit UInt5(int value)
        : m_value(value)
    {
        ASSERT(isUInt5(value));
    }

    operator int() { return m_value; }

private:
    int m_value;
};

class UInt12 {
public:
    explicit UInt12(int value)
        : m_value(value)
    {
        ASSERT(isUInt12(value));
    }

    operator int() { return m_value; }

private:
    int m_value;
};

class PostIndex {
public:
    explicit PostIndex(int value)
        : m_value(value)
    {
        ASSERT(isInt9(value));
    }

    operator int() { return m_value; }

private:
    int m_value;
};

class PreIndex {
public:
    explicit PreIndex(int value)
        : m_value(value)
    {
        ASSERT(isInt9(value));
    }

    operator int() { return m_value; }

private:
    int m_value;
};

class PairPostIndex {
public:
    explicit PairPostIndex(int value)
        : m_value(value)
    {
        ASSERT(isInt11(value));
    }

    operator int() { return m_value; }

private:
    int m_value;
};

class PairPreIndex {
public:
    explicit PairPreIndex(int value)
        : m_value(value)
    {
        ASSERT(isInt11(value));
    }

    operator int() { return m_value; }

private:
    int m_value;
};

class LogicalImmediate {
public:
    static LogicalImmediate create32(uint32_t value)
    {
        // Check for 0, -1 - these cannot be encoded.
        if (!value || !~value)
            return InvalidLogicalImmediate;

        // First look for a 32-bit pattern, then for repeating 16-bit
        // patterns, 8-bit, 4-bit, and finally 2-bit.

        unsigned hsb, lsb;
        bool inverted;
        if (findBitRange<32>(value, hsb, lsb, inverted))
            return encodeLogicalImmediate<32>(hsb, lsb, inverted);

        if ((value & 0xffff) != (value >> 16))
            return InvalidLogicalImmediate;
        value &= 0xffff;

        if (findBitRange<16>(value, hsb, lsb, inverted))
            return encodeLogicalImmediate<16>(hsb, lsb, inverted);

        if ((value & 0xff) != (value >> 8))
            return InvalidLogicalImmediate;
        value &= 0xff;

        if (findBitRange<8>(value, hsb, lsb, inverted))
            return encodeLogicalImmediate<8>(hsb, lsb, inverted);

        if ((value & 0xf) != (value >> 4))
            return InvalidLogicalImmediate;
        value &= 0xf;

        if (findBitRange<4>(value, hsb, lsb, inverted))
            return encodeLogicalImmediate<4>(hsb, lsb, inverted);

        if ((value & 0x3) != (value >> 2))
            return InvalidLogicalImmediate;
        value &= 0x3;

        if (findBitRange<2>(value, hsb, lsb, inverted))
            return encodeLogicalImmediate<2>(hsb, lsb, inverted);

        return InvalidLogicalImmediate;
    }

    static LogicalImmediate create64(uint64_t value)
    {
        // Check for 0, -1 - these cannot be encoded.
        if (!value || !~value)
            return InvalidLogicalImmediate;

        // Look for a contiguous bit range.
        unsigned hsb, lsb;
        bool inverted;
        if (findBitRange<64>(value, hsb, lsb, inverted))
            return encodeLogicalImmediate<64>(hsb, lsb, inverted);

        // If the high & low 32 bits are equal, we can try for a 32-bit (or narrower) pattern.
        if (static_cast<uint32_t>(value) == static_cast<uint32_t>(value >> 32))
            return create32(static_cast<uint32_t>(value));
        return InvalidLogicalImmediate;
    }

    int value() const
    {
        ASSERT(isValid());
        return m_value;
    }

    bool isValid() const
    {
        return m_value != InvalidLogicalImmediate;
    }

    bool is64bit() const
    {
        return m_value & (1 << 12);
    }

private:
    LogicalImmediate(int value)
        : m_value(value)
    {
    }

    // Generate a mask with bits in the range hsb..0 set, for example:
    //   hsb:63 = 0xffffffffffffffff
    //   hsb:42 = 0x000007ffffffffff
    //   hsb: 0 = 0x0000000000000001
    static uint64_t mask(unsigned hsb)
    {
        ASSERT(hsb < 64);
        return 0xffffffffffffffffull >> (63 - hsb);
    }

    template<unsigned N>
    static void partialHSB(uint64_t& value, unsigned&result)
    {
        if (value & (0xffffffffffffffffull << N)) {
            result += N;
            value >>= N;
        }
    }

    // Find the bit number of the highest bit set in a non-zero value, for example:
    //   0x8080808080808080 = hsb:63
    //   0x0000000000000001 = hsb: 0
    //   0x000007ffffe00000 = hsb:42
    static unsigned highestSetBit(uint64_t value)
    {
        ASSERT(value);
        unsigned hsb = 0;
        partialHSB<32>(value, hsb);
        partialHSB<16>(value, hsb);
        partialHSB<8>(value, hsb);
        partialHSB<4>(value, hsb);
        partialHSB<2>(value, hsb);
        partialHSB<1>(value, hsb);
        return hsb;
    }

    // This function takes a value and a bit width, where value obeys the following constraints:
    //   * bits outside of the width of the value must be zero.
    //   * bits within the width of value must neither be all clear or all set.
    // The input is inspected to detect values that consist of either two or three contiguous
    // ranges of bits. The output range hsb..lsb will describe the second range of the value.
    // if the range is set, inverted will be false, and if the range is clear, inverted will
    // be true. For example (with width 8):
    //   00001111 = hsb:3, lsb:0, inverted:false
    //   11110000 = hsb:3, lsb:0, inverted:true
    //   00111100 = hsb:5, lsb:2, inverted:false
    //   11000011 = hsb:5, lsb:2, inverted:true
    template<unsigned width>
    static bool findBitRange(uint64_t value, unsigned& hsb, unsigned& lsb, bool& inverted)
    {
        ASSERT(value & mask(width - 1));
        ASSERT(value != mask(width - 1));
        ASSERT(!(value & ~mask(width - 1)));

        // Detect cases where the top bit is set; if so, flip all the bits & set invert.
        // This halves the number of patterns we need to look for.
        const uint64_t msb = 1ull << (width - 1);
        if ((inverted = (value & msb)))
            value ^= mask(width - 1);

        // Find the highest set bit in value, generate a corresponding mask & flip all
        // bits under it.
        hsb = highestSetBit(value);
        value ^= mask(hsb);
        if (!value) {
            // If this cleared the value, then the range hsb..0 was all set.
            lsb = 0;
            return true;
        }

        // Try making one more mask, and flipping the bits!
        lsb = highestSetBit(value);
        value ^= mask(lsb);
        if (!value) {
            // Success - but lsb actually points to the hsb of a third range - add one
            // to get to the lsb of the mid range.
            ++lsb;
            return true;
        }

        return false;
    }

    // Encodes the set of immN:immr:imms fields found in a logical immediate.
    template<unsigned width>
    static int encodeLogicalImmediate(unsigned hsb, unsigned lsb, bool inverted)
    {
        // Check width is a power of 2!
        ASSERT(!(width & (width -1)));
        ASSERT(width <= 64 && width >= 2);
        ASSERT(hsb >= lsb);
        ASSERT(hsb < width);

        int immN = 0;
        int imms = 0;
        int immr = 0;

        // For 64-bit values this is easy - just set immN to true, and imms just
        // contains the bit number of the highest set bit of the set range. For
        // values with narrower widths, these are encoded by a leading set of
        // one bits, followed by a zero bit, followed by the remaining set of bits
        // being the high bit of the range. For a 32-bit immediate there are no
        // leading one bits, just a zero followed by a five bit number. For a
        // 16-bit immediate there is one one bit, a zero bit, and then a four bit
        // bit-position, etc.
        if (width == 64)
            immN = 1;
        else
            imms = 63 & ~(width + width - 1);

        if (inverted) {
            // if width is 64 & hsb is 62, then we have a value something like:
            //   0x80000000ffffffff (in this case with lsb 32).
            // The ror should be by 1, imms (effectively set width minus 1) is
            // 32. Set width is full width minus cleared width.
            immr = (width - 1) - hsb;
            imms |= (width - ((hsb - lsb) + 1)) - 1;
        } else {
            // if width is 64 & hsb is 62, then we have a value something like:
            //   0x7fffffff00000000 (in this case with lsb 32).
            // The value is effectively rol'ed by lsb, which is equivalent to
            // a ror by width - lsb (or 0, in the case where lsb is 0). imms
            // is hsb - lsb.
            immr = (width - lsb) & (width - 1);
            imms |= hsb - lsb;
        }

        return immN << 12 | immr << 6 | imms;
    }

    static const int InvalidLogicalImmediate = -1;

    int m_value;
};

inline uint16_t getHalfword(uint64_t value, int which)
{
    return value >> (which << 4);
}

namespace ARM64Registers {

#define FOR_EACH_CPU_REGISTER(V) \
    FOR_EACH_CPU_GPREGISTER(V) \
    FOR_EACH_CPU_SPECIAL_REGISTER(V) \
    FOR_EACH_CPU_FPREGISTER(V)

// The following are defined as pairs of the following value:
// 1. type of the storage needed to save the register value by the JIT probe.
// 2. name of the register.
#define FOR_EACH_CPU_GPREGISTER(V) \
    /* Parameter/result registers */ \
    V(void*, x0) \
    V(void*, x1) \
    V(void*, x2) \
    V(void*, x3) \
    V(void*, x4) \
    V(void*, x5) \
    V(void*, x6) \
    V(void*, x7) \
    /* Indirect result location register */ \
    V(void*, x8) \
    /* Temporary registers */ \
    V(void*, x9) \
    V(void*, x10) \
    V(void*, x11) \
    V(void*, x12) \
    V(void*, x13) \
    V(void*, x14) \
    V(void*, x15) \
    /* Intra-procedure-call scratch registers (temporary) */ \
    V(void*, x16) \
    V(void*, x17) \
    /* Platform Register (temporary) */ \
    V(void*, x18) \
    /* Callee-saved */ \
    V(void*, x19) \
    V(void*, x20) \
    V(void*, x21) \
    V(void*, x22) \
    V(void*, x23) \
    V(void*, x24) \
    V(void*, x25) \
    V(void*, x26) \
    V(void*, x27) \
    V(void*, x28) \
    /* Special */ \
    V(void*, fp) \
    V(void*, lr) \
    V(void*, sp)

#define FOR_EACH_CPU_SPECIAL_REGISTER(V) \
    V(void*, pc) \
    V(void*, nzcv) \
    V(void*, fpsr) \

// ARM64 always has 32 FPU registers 128-bits each. See http://llvm.org/devmtg/2012-11/Northover-AArch64.pdf
// and Section 5.1.2 in http://infocenter.arm.com/help/topic/com.arm.doc.ihi0055b/IHI0055B_aapcs64.pdf.
// However, we only use them for 64-bit doubles.
#define FOR_EACH_CPU_FPREGISTER(V) \
    /* Parameter/result registers */ \
    V(double, q0) \
    V(double, q1) \
    V(double, q2) \
    V(double, q3) \
    V(double, q4) \
    V(double, q5) \
    V(double, q6) \
    V(double, q7) \
    /* Callee-saved (up to 64-bits only!) */ \
    V(double, q8) \
    V(double, q9) \
    V(double, q10) \
    V(double, q11) \
    V(double, q12) \
    V(double, q13) \
    V(double, q14) \
    V(double, q15) \
    /* Temporary registers */ \
    V(double, q16) \
    V(double, q17) \
    V(double, q18) \
    V(double, q19) \
    V(double, q20) \
    V(double, q21) \
    V(double, q22) \
    V(double, q23) \
    V(double, q24) \
    V(double, q25) \
    V(double, q26) \
    V(double, q27) \
    V(double, q28) \
    V(double, q29) \
    V(double, q30) \
    V(double, q31)

typedef enum {
    #define DECLARE_REGISTER(_type, _regName) _regName,
    FOR_EACH_CPU_GPREGISTER(DECLARE_REGISTER)
    #undef DECLARE_REGISTER

    ip0 = x16,
    ip1 = x17,
    x29 = fp,
    x30 = lr,
    zr = 0x3f,
} RegisterID;

typedef enum {
    #define DECLARE_REGISTER(_type, _regName) _regName,
    FOR_EACH_CPU_FPREGISTER(DECLARE_REGISTER)
    #undef DECLARE_REGISTER
} FPRegisterID;

static constexpr bool isSp(RegisterID reg) { return reg == sp; }
static constexpr bool isZr(RegisterID reg) { return reg == zr; }

} // namespace ARM64Registers

class ARM64Assembler {
public:
    typedef ARM64Registers::RegisterID RegisterID;
    typedef ARM64Registers::FPRegisterID FPRegisterID;
    
    static constexpr RegisterID firstRegister() { return ARM64Registers::x0; }
    static constexpr RegisterID lastRegister() { return ARM64Registers::sp; }
    
    static constexpr FPRegisterID firstFPRegister() { return ARM64Registers::q0; }
    static constexpr FPRegisterID lastFPRegister() { return ARM64Registers::q31; }

private:
    static constexpr bool isSp(RegisterID reg) { return ARM64Registers::isSp(reg); }
    static constexpr bool isZr(RegisterID reg) { return ARM64Registers::isZr(reg); }

public:
    ARM64Assembler()
        : m_indexOfLastWatchpoint(INT_MIN)
        , m_indexOfTailOfLastWatchpoint(INT_MIN)
    {
    }
    
    AssemblerBuffer& buffer() { return m_buffer; }

    // (HS, LO, HI, LS) -> (AE, B, A, BE)
    // (VS, VC) -> (O, NO)
    typedef enum {
        ConditionEQ,
        ConditionNE,
        ConditionHS, ConditionCS = ConditionHS,
        ConditionLO, ConditionCC = ConditionLO,
        ConditionMI,
        ConditionPL,
        ConditionVS,
        ConditionVC,
        ConditionHI,
        ConditionLS,
        ConditionGE,
        ConditionLT,
        ConditionGT,
        ConditionLE,
        ConditionAL,
        ConditionInvalid
    } Condition;

    static Condition invert(Condition cond)
    {
        return static_cast<Condition>(cond ^ 1);
    }

    typedef enum {
        LSL,
        LSR,
        ASR,
        ROR
    } ShiftType;

    typedef enum {
        UXTB,
        UXTH,
        UXTW,
        UXTX,
        SXTB,
        SXTH,
        SXTW,
        SXTX
    } ExtendType;

    enum SetFlags {
        DontSetFlags,
        S
    };

#define JUMP_ENUM_WITH_SIZE(index, value) (((value) << 4) | (index))
#define JUMP_ENUM_SIZE(jump) ((jump) >> 4)
    enum JumpType { JumpFixed = JUMP_ENUM_WITH_SIZE(0, 0),
        JumpNoCondition = JUMP_ENUM_WITH_SIZE(1, 1 * sizeof(uint32_t)),
        JumpCondition = JUMP_ENUM_WITH_SIZE(2, 2 * sizeof(uint32_t)),
        JumpCompareAndBranch = JUMP_ENUM_WITH_SIZE(3, 2 * sizeof(uint32_t)),
        JumpTestBit = JUMP_ENUM_WITH_SIZE(4, 2 * sizeof(uint32_t)),
        JumpNoConditionFixedSize = JUMP_ENUM_WITH_SIZE(5, 1 * sizeof(uint32_t)),
        JumpConditionFixedSize = JUMP_ENUM_WITH_SIZE(6, 2 * sizeof(uint32_t)),
        JumpCompareAndBranchFixedSize = JUMP_ENUM_WITH_SIZE(7, 2 * sizeof(uint32_t)),
        JumpTestBitFixedSize = JUMP_ENUM_WITH_SIZE(8, 2 * sizeof(uint32_t)),
    };
    enum JumpLinkType {
        LinkInvalid = JUMP_ENUM_WITH_SIZE(0, 0),
        LinkJumpNoCondition = JUMP_ENUM_WITH_SIZE(1, 1 * sizeof(uint32_t)),
        LinkJumpConditionDirect = JUMP_ENUM_WITH_SIZE(2, 1 * sizeof(uint32_t)),
        LinkJumpCondition = JUMP_ENUM_WITH_SIZE(3, 2 * sizeof(uint32_t)),
        LinkJumpCompareAndBranch = JUMP_ENUM_WITH_SIZE(4, 2 * sizeof(uint32_t)),
        LinkJumpCompareAndBranchDirect = JUMP_ENUM_WITH_SIZE(5, 1 * sizeof(uint32_t)),
        LinkJumpTestBit = JUMP_ENUM_WITH_SIZE(6, 2 * sizeof(uint32_t)),
        LinkJumpTestBitDirect = JUMP_ENUM_WITH_SIZE(7, 1 * sizeof(uint32_t)),
    };

    class LinkRecord {
    public:
        LinkRecord(intptr_t from, intptr_t to, JumpType type, Condition condition)
        {
            data.realTypes.m_from = from;
            data.realTypes.m_to = to;
            data.realTypes.m_type = type;
            data.realTypes.m_linkType = LinkInvalid;
            data.realTypes.m_condition = condition;
        }
        LinkRecord(intptr_t from, intptr_t to, JumpType type, Condition condition, bool is64Bit, RegisterID compareRegister)
        {
            data.realTypes.m_from = from;
            data.realTypes.m_to = to;
            data.realTypes.m_type = type;
            data.realTypes.m_linkType = LinkInvalid;
            data.realTypes.m_condition = condition;
            data.realTypes.m_is64Bit = is64Bit;
            data.realTypes.m_compareRegister = compareRegister;
        }
        LinkRecord(intptr_t from, intptr_t to, JumpType type, Condition condition, unsigned bitNumber, RegisterID compareRegister)
        {
            data.realTypes.m_from = from;
            data.realTypes.m_to = to;
            data.realTypes.m_type = type;
            data.realTypes.m_linkType = LinkInvalid;
            data.realTypes.m_condition = condition;
            data.realTypes.m_bitNumber = bitNumber;
            data.realTypes.m_compareRegister = compareRegister;
        }
        void operator=(const LinkRecord& other)
        {
            data.copyTypes.content[0] = other.data.copyTypes.content[0];
            data.copyTypes.content[1] = other.data.copyTypes.content[1];
            data.copyTypes.content[2] = other.data.copyTypes.content[2];
        }
        intptr_t from() const { return data.realTypes.m_from; }
        void setFrom(intptr_t from) { data.realTypes.m_from = from; }
        intptr_t to() const { return data.realTypes.m_to; }
        JumpType type() const { return data.realTypes.m_type; }
        JumpLinkType linkType() const { return data.realTypes.m_linkType; }
        void setLinkType(JumpLinkType linkType) { ASSERT(data.realTypes.m_linkType == LinkInvalid); data.realTypes.m_linkType = linkType; }
        Condition condition() const { return data.realTypes.m_condition; }
        bool is64Bit() const { return data.realTypes.m_is64Bit; }
        unsigned bitNumber() const { return data.realTypes.m_bitNumber; }
        RegisterID compareRegister() const { return data.realTypes.m_compareRegister; }

    private:
        union {
            struct RealTypes {
                intptr_t m_from : 48;
                intptr_t m_to : 48;
                JumpType m_type : 8;
                JumpLinkType m_linkType : 8;
                Condition m_condition : 4;
                unsigned m_bitNumber : 6;
                RegisterID m_compareRegister : 6;
                bool m_is64Bit : 1;
            } realTypes;
            struct CopyTypes {
                uint64_t content[3];
            } copyTypes;
            COMPILE_ASSERT(sizeof(RealTypes) == sizeof(CopyTypes), LinkRecordCopyStructSizeEqualsRealStruct);
        } data;
    };

    // bits(N) VFPExpandImm(bits(8) imm8);
    //
    // Encoding of floating point immediates is a litte complicated. Here's a
    // high level description:
    //     +/-m*2-n where m and n are integers, 16 <= m <= 31, 0 <= n <= 7
    // and the algirithm for expanding to a single precision float:
    //     return imm8<7>:NOT(imm8<6>):Replicate(imm8<6>,5):imm8<5:0>:Zeros(19);
    //
    // The trickiest bit is how the exponent is handled. The following table
    // may help clarify things a little:
    //     654
    //     100 01111100 124 -3 1020 01111111100
    //     101 01111101 125 -2 1021 01111111101
    //     110 01111110 126 -1 1022 01111111110
    //     111 01111111 127  0 1023 01111111111
    //     000 10000000 128  1 1024 10000000000
    //     001 10000001 129  2 1025 10000000001
    //     010 10000010 130  3 1026 10000000010
    //     011 10000011 131  4 1027 10000000011
    // The first column shows the bit pattern stored in bits 6-4 of the arm
    // encoded immediate. The second column shows the 8-bit IEEE 754 single
    // -precision exponent in binary, the third column shows the raw decimal
    // value. IEEE 754 single-precision numbers are stored with a bias of 127
    // to the exponent, so the fourth column shows the resulting exponent.
    // From this was can see that the exponent can be in the range -3..4,
    // which agrees with the high level description given above. The fifth
    // and sixth columns shows the value stored in a IEEE 754 double-precision
    // number to represent these exponents in decimal and binary, given the
    // bias of 1023.
    //
    // Ultimately, detecting doubles that can be encoded as immediates on arm
    // and encoding doubles is actually not too bad. A floating point value can
    // be encoded by retaining the sign bit, the low three bits of the exponent
    // and the high 4 bits of the mantissa. To validly be able to encode an
    // immediate the remainder of the mantissa must be zero, and the high part
    // of the exponent must match the top bit retained, bar the highest bit
    // which must be its inverse.
    static bool canEncodeFPImm(double d)
    {
        // Discard the sign bit, the low two bits of the exponent & the highest
        // four bits of the mantissa.
        uint64_t masked = bitwise_cast<uint64_t>(d) & 0x7fc0ffffffffffffull;
        return (masked == 0x3fc0000000000000ull) || (masked == 0x4000000000000000ull);
    }

    template<int datasize>
    static bool canEncodePImmOffset(int32_t offset)
    {
        return isValidScaledUImm12<datasize>(offset);
    }

    static bool canEncodeSImmOffset(int32_t offset)
    {
        return isValidSignedImm9(offset);
    }

private:
    int encodeFPImm(double d)
    {
        ASSERT(canEncodeFPImm(d));
        uint64_t u64 = bitwise_cast<uint64_t>(d);
        return (static_cast<int>(u64 >> 56) & 0x80) | (static_cast<int>(u64 >> 48) & 0x7f);
    }

    template<int datasize>
    int encodeShiftAmount(int amount)
    {
        ASSERT(!amount || datasize == (8 << amount));
        return amount;
    }

    template<int datasize>
    static int encodePositiveImmediate(unsigned pimm)
    {
        ASSERT(!(pimm & ((datasize / 8) - 1)));
        return pimm / (datasize / 8);
    }

    enum Datasize {
        Datasize_32,
        Datasize_64,
        Datasize_64_top,
        Datasize_16
    };

    enum MemOpSize {
        MemOpSize_8_or_128,
        MemOpSize_16,
        MemOpSize_32,
        MemOpSize_64,
    };

    enum BranchType {
        BranchType_JMP,
        BranchType_CALL,
        BranchType_RET
    };

    enum AddOp {
        AddOp_ADD,
        AddOp_SUB
    };

    enum BitfieldOp {
        BitfieldOp_SBFM,
        BitfieldOp_BFM,
        BitfieldOp_UBFM
    };

    enum DataOp1Source {
        DataOp_RBIT,
        DataOp_REV16,
        DataOp_REV32,
        DataOp_REV64,
        DataOp_CLZ,
        DataOp_CLS
    };

    enum DataOp2Source {
        DataOp_UDIV = 2,
        DataOp_SDIV = 3,
        DataOp_LSLV = 8,
        DataOp_LSRV = 9,
        DataOp_ASRV = 10,
        DataOp_RORV = 11
    };

    enum DataOp3Source {
        DataOp_MADD = 0,
        DataOp_MSUB = 1,
        DataOp_SMADDL = 2,
        DataOp_SMSUBL = 3,
        DataOp_SMULH = 4,
        DataOp_UMADDL = 10,
        DataOp_UMSUBL = 11,
        DataOp_UMULH = 12
    };

    enum ExcepnOp {
        ExcepnOp_EXCEPTION = 0,
        ExcepnOp_BREAKPOINT = 1,
        ExcepnOp_HALT = 2,
        ExcepnOp_DCPS = 5
    };

    enum FPCmpOp {
        FPCmpOp_FCMP = 0x00,
        FPCmpOp_FCMP0 = 0x08,
        FPCmpOp_FCMPE = 0x10,
        FPCmpOp_FCMPE0 = 0x18
    };

    enum FPCondCmpOp {
        FPCondCmpOp_FCMP,
        FPCondCmpOp_FCMPE
    };

    enum FPDataOp1Source {
        FPDataOp_FMOV = 0,
        FPDataOp_FABS = 1,
        FPDataOp_FNEG = 2,
        FPDataOp_FSQRT = 3,
        FPDataOp_FCVT_toSingle = 4,
        FPDataOp_FCVT_toDouble = 5,
        FPDataOp_FCVT_toHalf = 7,
        FPDataOp_FRINTN = 8,
        FPDataOp_FRINTP = 9,
        FPDataOp_FRINTM = 10,
        FPDataOp_FRINTZ = 11,
        FPDataOp_FRINTA = 12,
        FPDataOp_FRINTX = 14,
        FPDataOp_FRINTI = 15
    };

    enum FPDataOp2Source {
        FPDataOp_FMUL,
        FPDataOp_FDIV,
        FPDataOp_FADD,
        FPDataOp_FSUB,
        FPDataOp_FMAX,
        FPDataOp_FMIN,
        FPDataOp_FMAXNM,
        FPDataOp_FMINNM,
        FPDataOp_FNMUL
    };

    enum SIMD3Same {
        SIMD_LogicalOp_AND = 0x03
    };

    enum FPIntConvOp {
        FPIntConvOp_FCVTNS = 0x00,
        FPIntConvOp_FCVTNU = 0x01,
        FPIntConvOp_SCVTF = 0x02,
        FPIntConvOp_UCVTF = 0x03,
        FPIntConvOp_FCVTAS = 0x04,
        FPIntConvOp_FCVTAU = 0x05,
        FPIntConvOp_FMOV_QtoX = 0x06,
        FPIntConvOp_FMOV_XtoQ = 0x07,
        FPIntConvOp_FCVTPS = 0x08,
        FPIntConvOp_FCVTPU = 0x09,
        FPIntConvOp_FMOV_QtoX_top = 0x0e,
        FPIntConvOp_FMOV_XtoQ_top = 0x0f,
        FPIntConvOp_FCVTMS = 0x10,
        FPIntConvOp_FCVTMU = 0x11,
        FPIntConvOp_FCVTZS = 0x18,
        FPIntConvOp_FCVTZU = 0x19,
    };

    enum LogicalOp {
        LogicalOp_AND,
        LogicalOp_ORR,
        LogicalOp_EOR,
        LogicalOp_ANDS
    };

    enum MemOp {
        MemOp_STORE,
        MemOp_LOAD,
        MemOp_STORE_V128, 
        MemOp_LOAD_V128,
        MemOp_PREFETCH = 2, // size must be 3
        MemOp_LOAD_signed64 = 2, // size may be 0, 1 or 2
        MemOp_LOAD_signed32 = 3 // size may be 0 or 1
    };

    enum MemPairOpSize {
        MemPairOp_32 = 0,
        MemPairOp_LoadSigned_32 = 1,
        MemPairOp_64 = 2,

        MemPairOp_V32 = MemPairOp_32,
        MemPairOp_V64 = 1,
        MemPairOp_V128 = 2
    };

    enum MoveWideOp {
        MoveWideOp_N = 0,
        MoveWideOp_Z = 2,
        MoveWideOp_K = 3 
    };

    enum LdrLiteralOp {
        LdrLiteralOp_32BIT = 0,
        LdrLiteralOp_64BIT = 1,
        LdrLiteralOp_LDRSW = 2,
        LdrLiteralOp_128BIT = 2
    };

    static unsigned memPairOffsetShift(bool V, MemPairOpSize size)
    {
        // return the log2 of the size in bytes, e.g. 64 bit size returns 3
        if (V)
            return size + 2;
        return (size >> 1) + 2;
    }

public:
    // Integer Instructions:

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void adc(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        CHECK_DATASIZE();
        insn(addSubtractWithCarry(DATASIZE, AddOp_ADD, setFlags, rm, rn, rd));
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void add(RegisterID rd, RegisterID rn, UInt12 imm12, int shift = 0)
    {
        CHECK_DATASIZE();
        ASSERT(!shift || shift == 12);
        insn(addSubtractImmediate(DATASIZE, AddOp_ADD, setFlags, shift == 12, imm12, rn, rd));
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void add(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        add<datasize, setFlags>(rd, rn, rm, LSL, 0);
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void add(RegisterID rd, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        CHECK_DATASIZE();
        insn(addSubtractExtendedRegister(DATASIZE, AddOp_ADD, setFlags, rm, extend, amount, rn, rd));
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void add(RegisterID rd, RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        CHECK_DATASIZE();
        if (isSp(rd) || isSp(rn)) {
            ASSERT(shift == LSL);
            ASSERT(!isSp(rm));
            add<datasize, setFlags>(rd, rn, rm, UXTX, amount);
        } else
            insn(addSubtractShiftedRegister(DATASIZE, AddOp_ADD, setFlags, shift, rm, amount, rn, rd));
    }

    ALWAYS_INLINE void adr(RegisterID rd, int offset)
    {
        insn(pcRelative(false, offset, rd));
    }

    ALWAYS_INLINE void adrp(RegisterID rd, int offset)
    {
        ASSERT(!(offset & 0xfff));
        insn(pcRelative(true, offset >> 12, rd));
        nopCortexA53Fix843419();
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void and_(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        and_<datasize, setFlags>(rd, rn, rm, LSL, 0);
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void and_(RegisterID rd, RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        CHECK_DATASIZE();
        insn(logicalShiftedRegister(DATASIZE, setFlags ? LogicalOp_ANDS : LogicalOp_AND, shift, false, rm, amount, rn, rd));
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void and_(RegisterID rd, RegisterID rn, LogicalImmediate imm)
    {
        CHECK_DATASIZE();
        insn(logicalImmediate(DATASIZE, setFlags ? LogicalOp_ANDS : LogicalOp_AND, imm.value(), rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void asr(RegisterID rd, RegisterID rn, int shift)
    {
        ASSERT(shift < datasize);
        sbfm<datasize>(rd, rn, shift, datasize - 1);
    }

    template<int datasize>
    ALWAYS_INLINE void asr(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        asrv<datasize>(rd, rn, rm);
    }

    template<int datasize>
    ALWAYS_INLINE void asrv(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        CHECK_DATASIZE();
        insn(dataProcessing2Source(DATASIZE, rm, DataOp_ASRV, rn, rd));
    }

    ALWAYS_INLINE void b(int32_t offset = 0)
    {
        ASSERT(!(offset & 3));
        offset >>= 2;
        ASSERT(offset == (offset << 6) >> 6);
        insn(unconditionalBranchImmediate(false, offset));
    }

    ALWAYS_INLINE void b_cond(Condition cond, int32_t offset = 0)
    {
        ASSERT(!(offset & 3));
        offset >>= 2;
        ASSERT(offset == (offset << 13) >> 13);
        insn(conditionalBranchImmediate(offset, cond));
    }

    template<int datasize>
    ALWAYS_INLINE void bfi(RegisterID rd, RegisterID rn, int lsb, int width)
    {
        bfm<datasize>(rd, rn, (datasize - lsb) & (datasize - 1), width - 1);
    }

    template<int datasize>
    ALWAYS_INLINE void bfm(RegisterID rd, RegisterID rn, int immr, int imms)
    {
        CHECK_DATASIZE();
        insn(bitfield(DATASIZE, BitfieldOp_BFM, immr, imms, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void bfxil(RegisterID rd, RegisterID rn, int lsb, int width)
    {
        bfm<datasize>(rd, rn, lsb, lsb + width - 1);
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void bic(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        bic<datasize, setFlags>(rd, rn, rm, LSL, 0);
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void bic(RegisterID rd, RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        CHECK_DATASIZE();
        insn(logicalShiftedRegister(DATASIZE, setFlags ? LogicalOp_ANDS : LogicalOp_AND, shift, true, rm, amount, rn, rd));
    }

    ALWAYS_INLINE void bl(int32_t offset = 0)
    {
        ASSERT(!(offset & 3));
        offset >>= 2;
        insn(unconditionalBranchImmediate(true, offset));
    }

    ALWAYS_INLINE void blr(RegisterID rn)
    {
        insn(unconditionalBranchRegister(BranchType_CALL, rn));
    }

    ALWAYS_INLINE void br(RegisterID rn)
    {
        insn(unconditionalBranchRegister(BranchType_JMP, rn));
    }

    ALWAYS_INLINE void brk(uint16_t imm)
    {
        insn(excepnGeneration(ExcepnOp_BREAKPOINT, imm, 0));
    }
    
    template<int datasize>
    ALWAYS_INLINE void cbnz(RegisterID rt, int32_t offset = 0)
    {
        CHECK_DATASIZE();
        ASSERT(!(offset & 3));
        offset >>= 2;
        insn(compareAndBranchImmediate(DATASIZE, true, offset, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void cbz(RegisterID rt, int32_t offset = 0)
    {
        CHECK_DATASIZE();
        ASSERT(!(offset & 3));
        offset >>= 2;
        insn(compareAndBranchImmediate(DATASIZE, false, offset, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ccmn(RegisterID rn, RegisterID rm, int nzcv, Condition cond)
    {
        CHECK_DATASIZE();
        insn(conditionalCompareRegister(DATASIZE, AddOp_ADD, rm, cond, rn, nzcv));
    }

    template<int datasize>
    ALWAYS_INLINE void ccmn(RegisterID rn, UInt5 imm, int nzcv, Condition cond)
    {
        CHECK_DATASIZE();
        insn(conditionalCompareImmediate(DATASIZE, AddOp_ADD, imm, cond, rn, nzcv));
    }

    template<int datasize>
    ALWAYS_INLINE void ccmp(RegisterID rn, RegisterID rm, int nzcv, Condition cond)
    {
        CHECK_DATASIZE();
        insn(conditionalCompareRegister(DATASIZE, AddOp_SUB, rm, cond, rn, nzcv));
    }

    template<int datasize>
    ALWAYS_INLINE void ccmp(RegisterID rn, UInt5 imm, int nzcv, Condition cond)
    {
        CHECK_DATASIZE();
        insn(conditionalCompareImmediate(DATASIZE, AddOp_SUB, imm, cond, rn, nzcv));
    }

    template<int datasize>
    ALWAYS_INLINE void cinc(RegisterID rd, RegisterID rn, Condition cond)
    {
        csinc<datasize>(rd, rn, rn, invert(cond));
    }

    template<int datasize>
    ALWAYS_INLINE void cinv(RegisterID rd, RegisterID rn, Condition cond)
    {
        csinv<datasize>(rd, rn, rn, invert(cond));
    }

    template<int datasize>
    ALWAYS_INLINE void cls(RegisterID rd, RegisterID rn)
    {
        CHECK_DATASIZE();
        insn(dataProcessing1Source(DATASIZE, DataOp_CLS, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void clz(RegisterID rd, RegisterID rn)
    {
        CHECK_DATASIZE();
        insn(dataProcessing1Source(DATASIZE, DataOp_CLZ, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void cmn(RegisterID rn, UInt12 imm12, int shift = 0)
    {
        add<datasize, S>(ARM64Registers::zr, rn, imm12, shift);
    }

    template<int datasize>
    ALWAYS_INLINE void cmn(RegisterID rn, RegisterID rm)
    {
        add<datasize, S>(ARM64Registers::zr, rn, rm);
    }

    template<int datasize>
    ALWAYS_INLINE void cmn(RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        add<datasize, S>(ARM64Registers::zr, rn, rm, extend, amount);
    }

    template<int datasize>
    ALWAYS_INLINE void cmn(RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        add<datasize, S>(ARM64Registers::zr, rn, rm, shift, amount);
    }

    template<int datasize>
    ALWAYS_INLINE void cmp(RegisterID rn, UInt12 imm12, int shift = 0)
    {
        sub<datasize, S>(ARM64Registers::zr, rn, imm12, shift);
    }

    template<int datasize>
    ALWAYS_INLINE void cmp(RegisterID rn, RegisterID rm)
    {
        sub<datasize, S>(ARM64Registers::zr, rn, rm);
    }

    template<int datasize>
    ALWAYS_INLINE void cmp(RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        sub<datasize, S>(ARM64Registers::zr, rn, rm, extend, amount);
    }

    template<int datasize>
    ALWAYS_INLINE void cmp(RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        sub<datasize, S>(ARM64Registers::zr, rn, rm, shift, amount);
    }

    template<int datasize>
    ALWAYS_INLINE void cneg(RegisterID rd, RegisterID rn, Condition cond)
    {
        csneg<datasize>(rd, rn, rn, invert(cond));
    }

    template<int datasize>
    ALWAYS_INLINE void csel(RegisterID rd, RegisterID rn, RegisterID rm, Condition cond)
    {
        CHECK_DATASIZE();
        insn(conditionalSelect(DATASIZE, false, rm, cond, false, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void cset(RegisterID rd, Condition cond)
    {
        csinc<datasize>(rd, ARM64Registers::zr, ARM64Registers::zr, invert(cond));
    }

    template<int datasize>
    ALWAYS_INLINE void csetm(RegisterID rd, Condition cond)
    {
        csinv<datasize>(rd, ARM64Registers::zr, ARM64Registers::zr, invert(cond));
    }

    template<int datasize>
    ALWAYS_INLINE void csinc(RegisterID rd, RegisterID rn, RegisterID rm, Condition cond)
    {
        CHECK_DATASIZE();
        insn(conditionalSelect(DATASIZE, false, rm, cond, true, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void csinv(RegisterID rd, RegisterID rn, RegisterID rm, Condition cond)
    {
        CHECK_DATASIZE();
        insn(conditionalSelect(DATASIZE, true, rm, cond, false, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void csneg(RegisterID rd, RegisterID rn, RegisterID rm, Condition cond)
    {
        CHECK_DATASIZE();
        insn(conditionalSelect(DATASIZE, true, rm, cond, true, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void eon(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        eon<datasize>(rd, rn, rm, LSL, 0);
    }

    template<int datasize>
    ALWAYS_INLINE void eon(RegisterID rd, RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        CHECK_DATASIZE();
        insn(logicalShiftedRegister(DATASIZE, LogicalOp_EOR, shift, true, rm, amount, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void eor(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        eor<datasize>(rd, rn, rm, LSL, 0);
    }

    template<int datasize>
    ALWAYS_INLINE void eor(RegisterID rd, RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        CHECK_DATASIZE();
        insn(logicalShiftedRegister(DATASIZE, LogicalOp_EOR, shift, false, rm, amount, rn, rd));
    }
    
    template<int datasize>
    ALWAYS_INLINE void eor(RegisterID rd, RegisterID rn, LogicalImmediate imm)
    {
        CHECK_DATASIZE();
        insn(logicalImmediate(DATASIZE, LogicalOp_EOR, imm.value(), rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void extr(RegisterID rd, RegisterID rn, RegisterID rm, int lsb)
    {
        CHECK_DATASIZE();
        insn(extract(DATASIZE, rm, lsb, rn, rd));
    }

    ALWAYS_INLINE void hint(int imm)
    {
        insn(hintPseudo(imm));
    }

    ALWAYS_INLINE void hlt(uint16_t imm)
    {
        insn(excepnGeneration(ExcepnOp_HALT, imm, 0));
    }

    template<int datasize>
    ALWAYS_INLINE void ldp(RegisterID rt, RegisterID rt2, RegisterID rn, PairPostIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPairPostIndex(MEMPAIROPSIZE_INT(datasize), false, MemOp_LOAD, simm, rn, rt, rt2));
    }

    template<int datasize>
    ALWAYS_INLINE void ldp(RegisterID rt, RegisterID rt2, RegisterID rn, PairPreIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPairPreIndex(MEMPAIROPSIZE_INT(datasize), false, MemOp_LOAD, simm, rn, rt, rt2));
    }

    template<int datasize>
    ALWAYS_INLINE void ldr(RegisterID rt, RegisterID rn, RegisterID rm)
    {
        ldr<datasize>(rt, rn, rm, UXTX, 0);
    }

    template<int datasize>
    ALWAYS_INLINE void ldr(RegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterRegisterOffset(MEMOPSIZE, false, MemOp_LOAD, rm, extend, encodeShiftAmount<datasize>(amount), rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldr(RegisterID rt, RegisterID rn, unsigned pimm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterUnsignedImmediate(MEMOPSIZE, false, MemOp_LOAD, encodePositiveImmediate<datasize>(pimm), rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldr(RegisterID rt, RegisterID rn, PostIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPostIndex(MEMOPSIZE, false, MemOp_LOAD, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldr(RegisterID rt, RegisterID rn, PreIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPreIndex(MEMOPSIZE, false, MemOp_LOAD, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldr_literal(RegisterID rt, int offset = 0)
    {
        CHECK_DATASIZE();
        ASSERT(!(offset & 3));
        insn(loadRegisterLiteral(datasize == 64 ? LdrLiteralOp_64BIT : LdrLiteralOp_32BIT, false, offset >> 2, rt));
    }

    ALWAYS_INLINE void ldrb(RegisterID rt, RegisterID rn, RegisterID rm)
    {
        // Not calling the 5 argument form of ldrb, since is amount is ommitted S is false.
        insn(loadStoreRegisterRegisterOffset(MemOpSize_8_or_128, false, MemOp_LOAD, rm, UXTX, false, rn, rt));
    }

    ALWAYS_INLINE void ldrb(RegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        ASSERT_UNUSED(amount, !amount);
        insn(loadStoreRegisterRegisterOffset(MemOpSize_8_or_128, false, MemOp_LOAD, rm, extend, true, rn, rt));
    }

    ALWAYS_INLINE void ldrb(RegisterID rt, RegisterID rn, unsigned pimm)
    {
        insn(loadStoreRegisterUnsignedImmediate(MemOpSize_8_or_128, false, MemOp_LOAD, encodePositiveImmediate<8>(pimm), rn, rt));
    }

    ALWAYS_INLINE void ldrb(RegisterID rt, RegisterID rn, PostIndex simm)
    {
        insn(loadStoreRegisterPostIndex(MemOpSize_8_or_128, false, MemOp_LOAD, simm, rn, rt));
    }

    ALWAYS_INLINE void ldrb(RegisterID rt, RegisterID rn, PreIndex simm)
    {
        insn(loadStoreRegisterPreIndex(MemOpSize_8_or_128, false, MemOp_LOAD, simm, rn, rt));
    }

    ALWAYS_INLINE void ldrh(RegisterID rt, RegisterID rn, RegisterID rm)
    {
        ldrh(rt, rn, rm, UXTX, 0);
    }

    ALWAYS_INLINE void ldrh(RegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        ASSERT(!amount || amount == 1);
        insn(loadStoreRegisterRegisterOffset(MemOpSize_16, false, MemOp_LOAD, rm, extend, amount == 1, rn, rt));
    }

    ALWAYS_INLINE void ldrh(RegisterID rt, RegisterID rn, unsigned pimm)
    {
        insn(loadStoreRegisterUnsignedImmediate(MemOpSize_16, false, MemOp_LOAD, encodePositiveImmediate<16>(pimm), rn, rt));
    }

    ALWAYS_INLINE void ldrh(RegisterID rt, RegisterID rn, PostIndex simm)
    {
        insn(loadStoreRegisterPostIndex(MemOpSize_16, false, MemOp_LOAD, simm, rn, rt));
    }

    ALWAYS_INLINE void ldrh(RegisterID rt, RegisterID rn, PreIndex simm)
    {
        insn(loadStoreRegisterPreIndex(MemOpSize_16, false, MemOp_LOAD, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldrsb(RegisterID rt, RegisterID rn, RegisterID rm)
    {
        CHECK_DATASIZE();
        // Not calling the 5 argument form of ldrsb, since is amount is ommitted S is false.
        insn(loadStoreRegisterRegisterOffset(MemOpSize_8_or_128, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, rm, UXTX, false, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldrsb(RegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        CHECK_DATASIZE();
        ASSERT_UNUSED(amount, !amount);
        insn(loadStoreRegisterRegisterOffset(MemOpSize_8_or_128, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, rm, extend, true, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldrsb(RegisterID rt, RegisterID rn, unsigned pimm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterUnsignedImmediate(MemOpSize_8_or_128, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, encodePositiveImmediate<8>(pimm), rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldrsb(RegisterID rt, RegisterID rn, PostIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPostIndex(MemOpSize_8_or_128, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldrsb(RegisterID rt, RegisterID rn, PreIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPreIndex(MemOpSize_8_or_128, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldrsh(RegisterID rt, RegisterID rn, RegisterID rm)
    {
        ldrsh<datasize>(rt, rn, rm, UXTX, 0);
    }

    template<int datasize>
    ALWAYS_INLINE void ldrsh(RegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        CHECK_DATASIZE();
        ASSERT(!amount || amount == 1);
        insn(loadStoreRegisterRegisterOffset(MemOpSize_16, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, rm, extend, amount == 1, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldrsh(RegisterID rt, RegisterID rn, unsigned pimm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterUnsignedImmediate(MemOpSize_16, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, encodePositiveImmediate<16>(pimm), rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldrsh(RegisterID rt, RegisterID rn, PostIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPostIndex(MemOpSize_16, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldrsh(RegisterID rt, RegisterID rn, PreIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPreIndex(MemOpSize_16, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, simm, rn, rt));
    }

    ALWAYS_INLINE void ldrsw(RegisterID rt, RegisterID rn, RegisterID rm)
    {
        ldrsw(rt, rn, rm, UXTX, 0);
    }

    ALWAYS_INLINE void ldrsw(RegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        ASSERT(!amount || amount == 2);
        insn(loadStoreRegisterRegisterOffset(MemOpSize_32, false, MemOp_LOAD_signed64, rm, extend, amount == 2, rn, rt));
    }

    ALWAYS_INLINE void ldrsw(RegisterID rt, RegisterID rn, unsigned pimm)
    {
        insn(loadStoreRegisterUnsignedImmediate(MemOpSize_32, false, MemOp_LOAD_signed64, encodePositiveImmediate<32>(pimm), rn, rt));
    }

    ALWAYS_INLINE void ldrsw(RegisterID rt, RegisterID rn, PostIndex simm)
    {
        insn(loadStoreRegisterPostIndex(MemOpSize_32, false, MemOp_LOAD_signed64, simm, rn, rt));
    }

    ALWAYS_INLINE void ldrsw(RegisterID rt, RegisterID rn, PreIndex simm)
    {
        insn(loadStoreRegisterPreIndex(MemOpSize_32, false, MemOp_LOAD_signed64, simm, rn, rt));
    }

    ALWAYS_INLINE void ldrsw_literal(RegisterID rt, int offset = 0)
    {
        ASSERT(!(offset & 3));
        insn(loadRegisterLiteral(LdrLiteralOp_LDRSW, false, offset >> 2, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldur(RegisterID rt, RegisterID rn, int simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterUnscaledImmediate(MEMOPSIZE, false, MemOp_LOAD, simm, rn, rt));
    }

    ALWAYS_INLINE void ldurb(RegisterID rt, RegisterID rn, int simm)
    {
        insn(loadStoreRegisterUnscaledImmediate(MemOpSize_8_or_128, false, MemOp_LOAD, simm, rn, rt));
    }

    ALWAYS_INLINE void ldurh(RegisterID rt, RegisterID rn, int simm)
    {
        insn(loadStoreRegisterUnscaledImmediate(MemOpSize_16, false, MemOp_LOAD, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldursb(RegisterID rt, RegisterID rn, int simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterUnscaledImmediate(MemOpSize_8_or_128, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldursh(RegisterID rt, RegisterID rn, int simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterUnscaledImmediate(MemOpSize_16, false, (datasize == 64) ? MemOp_LOAD_signed64 : MemOp_LOAD_signed32, simm, rn, rt));
    }

    ALWAYS_INLINE void ldursw(RegisterID rt, RegisterID rn, int simm)
    {
        insn(loadStoreRegisterUnscaledImmediate(MemOpSize_32, false, MemOp_LOAD_signed64, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void lsl(RegisterID rd, RegisterID rn, int shift)
    {
        ASSERT(shift < datasize);
        ubfm<datasize>(rd, rn, (datasize - shift) & (datasize - 1), datasize - 1 - shift);
    }

    template<int datasize>
    ALWAYS_INLINE void lsl(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        lslv<datasize>(rd, rn, rm);
    }

    template<int datasize>
    ALWAYS_INLINE void lslv(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        CHECK_DATASIZE();
        insn(dataProcessing2Source(DATASIZE, rm, DataOp_LSLV, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void lsr(RegisterID rd, RegisterID rn, int shift)
    {
        ASSERT(shift < datasize);
        ubfm<datasize>(rd, rn, shift, datasize - 1);
    }

    template<int datasize>
    ALWAYS_INLINE void lsr(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        lsrv<datasize>(rd, rn, rm);
    }

    template<int datasize>
    ALWAYS_INLINE void lsrv(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        CHECK_DATASIZE();
        insn(dataProcessing2Source(DATASIZE, rm, DataOp_LSRV, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void madd(RegisterID rd, RegisterID rn, RegisterID rm, RegisterID ra)
    {
        CHECK_DATASIZE();
        nopCortexA53Fix835769<datasize>();
        insn(dataProcessing3Source(DATASIZE, DataOp_MADD, rm, ra, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void mneg(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        msub<datasize>(rd, rn, rm, ARM64Registers::zr);
    }

    template<int datasize>
    ALWAYS_INLINE void mov(RegisterID rd, RegisterID rm)
    {
        if (isSp(rd) || isSp(rm))
            add<datasize>(rd, rm, UInt12(0));
        else
            orr<datasize>(rd, ARM64Registers::zr, rm);
    }

    template<int datasize>
    ALWAYS_INLINE void movi(RegisterID rd, LogicalImmediate imm)
    {
        orr<datasize>(rd, ARM64Registers::zr, imm);
    }

    template<int datasize>
    ALWAYS_INLINE void movk(RegisterID rd, uint16_t value, int shift = 0)
    {
        CHECK_DATASIZE();
        ASSERT(!(shift & 0xf));
        insn(moveWideImediate(DATASIZE, MoveWideOp_K, shift >> 4, value, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void movn(RegisterID rd, uint16_t value, int shift = 0)
    {
        CHECK_DATASIZE();
        ASSERT(!(shift & 0xf));
        insn(moveWideImediate(DATASIZE, MoveWideOp_N, shift >> 4, value, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void movz(RegisterID rd, uint16_t value, int shift = 0)
    {
        CHECK_DATASIZE();
        ASSERT(!(shift & 0xf));
        insn(moveWideImediate(DATASIZE, MoveWideOp_Z, shift >> 4, value, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void msub(RegisterID rd, RegisterID rn, RegisterID rm, RegisterID ra)
    {
        CHECK_DATASIZE();
        nopCortexA53Fix835769<datasize>();
        insn(dataProcessing3Source(DATASIZE, DataOp_MSUB, rm, ra, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void mul(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        madd<datasize>(rd, rn, rm, ARM64Registers::zr);
    }

    template<int datasize>
    ALWAYS_INLINE void mvn(RegisterID rd, RegisterID rm)
    {
        orn<datasize>(rd, ARM64Registers::zr, rm);
    }

    template<int datasize>
    ALWAYS_INLINE void mvn(RegisterID rd, RegisterID rm, ShiftType shift, int amount)
    {
        orn<datasize>(rd, ARM64Registers::zr, rm, shift, amount);
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void neg(RegisterID rd, RegisterID rm)
    {
        sub<datasize, setFlags>(rd, ARM64Registers::zr, rm);
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void neg(RegisterID rd, RegisterID rm, ShiftType shift, int amount)
    {
        sub<datasize, setFlags>(rd, ARM64Registers::zr, rm, shift, amount);
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void ngc(RegisterID rd, RegisterID rm)
    {
        sbc<datasize, setFlags>(rd, ARM64Registers::zr, rm);
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void ngc(RegisterID rd, RegisterID rm, ShiftType shift, int amount)
    {
        sbc<datasize, setFlags>(rd, ARM64Registers::zr, rm, shift, amount);
    }

    ALWAYS_INLINE void nop()
    {
        insn(nopPseudo());
    }
    
    static void fillNops(void* base, size_t size)
    {
        RELEASE_ASSERT(!(size % sizeof(int32_t)));
        size_t n = size / sizeof(int32_t);
        for (int32_t* ptr = static_cast<int32_t*>(base); n--;)
            *ptr++ = nopPseudo();
    }
    
    ALWAYS_INLINE void dmbSY()
    {
        insn(0xd5033fbf);
    }

    template<int datasize>
    ALWAYS_INLINE void orn(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        orn<datasize>(rd, rn, rm, LSL, 0);
    }

    template<int datasize>
    ALWAYS_INLINE void orn(RegisterID rd, RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        CHECK_DATASIZE();
        insn(logicalShiftedRegister(DATASIZE, LogicalOp_ORR, shift, true, rm, amount, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void orr(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        orr<datasize>(rd, rn, rm, LSL, 0);
    }

    template<int datasize>
    ALWAYS_INLINE void orr(RegisterID rd, RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        CHECK_DATASIZE();
        insn(logicalShiftedRegister(DATASIZE, LogicalOp_ORR, shift, false, rm, amount, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void orr(RegisterID rd, RegisterID rn, LogicalImmediate imm)
    {
        CHECK_DATASIZE();
        insn(logicalImmediate(DATASIZE, LogicalOp_ORR, imm.value(), rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void rbit(RegisterID rd, RegisterID rn)
    {
        CHECK_DATASIZE();
        insn(dataProcessing1Source(DATASIZE, DataOp_RBIT, rn, rd));
    }

    ALWAYS_INLINE void ret(RegisterID rn = ARM64Registers::lr)
    {
        insn(unconditionalBranchRegister(BranchType_RET, rn));
    }

    template<int datasize>
    ALWAYS_INLINE void rev(RegisterID rd, RegisterID rn)
    {
        CHECK_DATASIZE();
        if (datasize == 32) // 'rev' mnemonic means REV32 or REV64 depending on the operand width.
            insn(dataProcessing1Source(Datasize_32, DataOp_REV32, rn, rd));
        else
            insn(dataProcessing1Source(Datasize_64, DataOp_REV64, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void rev16(RegisterID rd, RegisterID rn)
    {
        CHECK_DATASIZE();
        insn(dataProcessing1Source(DATASIZE, DataOp_REV16, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void rev32(RegisterID rd, RegisterID rn)
    {
        ASSERT(datasize == 64); // 'rev32' only valid with 64-bit operands.
        insn(dataProcessing1Source(Datasize_64, DataOp_REV32, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void ror(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        rorv<datasize>(rd, rn, rm);
    }

    template<int datasize>
    ALWAYS_INLINE void ror(RegisterID rd, RegisterID rs, int shift)
    {
        extr<datasize>(rd, rs, rs, shift);
    }

    template<int datasize>
    ALWAYS_INLINE void rorv(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        CHECK_DATASIZE();
        insn(dataProcessing2Source(DATASIZE, rm, DataOp_RORV, rn, rd));
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void sbc(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        CHECK_DATASIZE();
        insn(addSubtractWithCarry(DATASIZE, AddOp_SUB, setFlags, rm, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void sbfiz(RegisterID rd, RegisterID rn, int lsb, int width)
    {
        sbfm<datasize>(rd, rn, (datasize - lsb) & (datasize - 1), width - 1);
    }

    template<int datasize>
    ALWAYS_INLINE void sbfm(RegisterID rd, RegisterID rn, int immr, int imms)
    {
        CHECK_DATASIZE();
        insn(bitfield(DATASIZE, BitfieldOp_SBFM, immr, imms, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void sbfx(RegisterID rd, RegisterID rn, int lsb, int width)
    {
        sbfm<datasize>(rd, rn, lsb, lsb + width - 1);
    }

    template<int datasize>
    ALWAYS_INLINE void sdiv(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        CHECK_DATASIZE();
        insn(dataProcessing2Source(DATASIZE, rm, DataOp_SDIV, rn, rd));
    }

    ALWAYS_INLINE void smaddl(RegisterID rd, RegisterID rn, RegisterID rm, RegisterID ra)
    {
        nopCortexA53Fix835769<64>();
        insn(dataProcessing3Source(Datasize_64, DataOp_SMADDL, rm, ra, rn, rd));
    }

    ALWAYS_INLINE void smnegl(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        smsubl(rd, rn, rm, ARM64Registers::zr);
    }

    ALWAYS_INLINE void smsubl(RegisterID rd, RegisterID rn, RegisterID rm, RegisterID ra)
    {
        nopCortexA53Fix835769<64>();
        insn(dataProcessing3Source(Datasize_64, DataOp_SMSUBL, rm, ra, rn, rd));
    }

    ALWAYS_INLINE void smulh(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        insn(dataProcessing3Source(Datasize_64, DataOp_SMULH, rm, ARM64Registers::zr, rn, rd));
    }

    ALWAYS_INLINE void smull(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        smaddl(rd, rn, rm, ARM64Registers::zr);
    }

    template<int datasize>
    ALWAYS_INLINE void stp(RegisterID rt, RegisterID rt2, RegisterID rn, PairPostIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPairPostIndex(MEMPAIROPSIZE_INT(datasize), false, MemOp_STORE, simm, rn, rt, rt2));
    }

    template<int datasize>
    ALWAYS_INLINE void stp(RegisterID rt, RegisterID rt2, RegisterID rn, PairPreIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPairPreIndex(MEMPAIROPSIZE_INT(datasize), false, MemOp_STORE, simm, rn, rt, rt2));
    }

    template<int datasize>
    ALWAYS_INLINE void str(RegisterID rt, RegisterID rn, RegisterID rm)
    {
        str<datasize>(rt, rn, rm, UXTX, 0);
    }

    template<int datasize>
    ALWAYS_INLINE void str(RegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterRegisterOffset(MEMOPSIZE, false, MemOp_STORE, rm, extend, encodeShiftAmount<datasize>(amount), rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void str(RegisterID rt, RegisterID rn, unsigned pimm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterUnsignedImmediate(MEMOPSIZE, false, MemOp_STORE, encodePositiveImmediate<datasize>(pimm), rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void str(RegisterID rt, RegisterID rn, PostIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPostIndex(MEMOPSIZE, false, MemOp_STORE, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void str(RegisterID rt, RegisterID rn, PreIndex simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterPreIndex(MEMOPSIZE, false, MemOp_STORE, simm, rn, rt));
    }

    ALWAYS_INLINE void strb(RegisterID rt, RegisterID rn, RegisterID rm)
    {
        // Not calling the 5 argument form of strb, since is amount is ommitted S is false.
        insn(loadStoreRegisterRegisterOffset(MemOpSize_8_or_128, false, MemOp_STORE, rm, UXTX, false, rn, rt));
    }

    ALWAYS_INLINE void strb(RegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        ASSERT_UNUSED(amount, !amount);
        insn(loadStoreRegisterRegisterOffset(MemOpSize_8_or_128, false, MemOp_STORE, rm, extend, true, rn, rt));
    }

    ALWAYS_INLINE void strb(RegisterID rt, RegisterID rn, unsigned pimm)
    {
        insn(loadStoreRegisterUnsignedImmediate(MemOpSize_8_or_128, false, MemOp_STORE, encodePositiveImmediate<8>(pimm), rn, rt));
    }

    ALWAYS_INLINE void strb(RegisterID rt, RegisterID rn, PostIndex simm)
    {
        insn(loadStoreRegisterPostIndex(MemOpSize_8_or_128, false, MemOp_STORE, simm, rn, rt));
    }

    ALWAYS_INLINE void strb(RegisterID rt, RegisterID rn, PreIndex simm)
    {
        insn(loadStoreRegisterPreIndex(MemOpSize_8_or_128, false, MemOp_STORE, simm, rn, rt));
    }

    ALWAYS_INLINE void strh(RegisterID rt, RegisterID rn, RegisterID rm)
    {
        strh(rt, rn, rm, UXTX, 0);
    }

    ALWAYS_INLINE void strh(RegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        ASSERT(!amount || amount == 1);
        insn(loadStoreRegisterRegisterOffset(MemOpSize_16, false, MemOp_STORE, rm, extend, amount == 1, rn, rt));
    }

    ALWAYS_INLINE void strh(RegisterID rt, RegisterID rn, unsigned pimm)
    {
        insn(loadStoreRegisterUnsignedImmediate(MemOpSize_16, false, MemOp_STORE, encodePositiveImmediate<16>(pimm), rn, rt));
    }

    ALWAYS_INLINE void strh(RegisterID rt, RegisterID rn, PostIndex simm)
    {
        insn(loadStoreRegisterPostIndex(MemOpSize_16, false, MemOp_STORE, simm, rn, rt));
    }

    ALWAYS_INLINE void strh(RegisterID rt, RegisterID rn, PreIndex simm)
    {
        insn(loadStoreRegisterPreIndex(MemOpSize_16, false, MemOp_STORE, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void stur(RegisterID rt, RegisterID rn, int simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterUnscaledImmediate(MEMOPSIZE, false, MemOp_STORE, simm, rn, rt));
    }

    ALWAYS_INLINE void sturb(RegisterID rt, RegisterID rn, int simm)
    {
        insn(loadStoreRegisterUnscaledImmediate(MemOpSize_8_or_128, false, MemOp_STORE, simm, rn, rt));
    }

    ALWAYS_INLINE void sturh(RegisterID rt, RegisterID rn, int simm)
    {
        insn(loadStoreRegisterUnscaledImmediate(MemOpSize_16, false, MemOp_STORE, simm, rn, rt));
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void sub(RegisterID rd, RegisterID rn, UInt12 imm12, int shift = 0)
    {
        CHECK_DATASIZE();
        ASSERT(!shift || shift == 12);
        insn(addSubtractImmediate(DATASIZE, AddOp_SUB, setFlags, shift == 12, imm12, rn, rd));
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void sub(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        ASSERT_WITH_MESSAGE(!isSp(rd) || setFlags == DontSetFlags, "SUBS with shifted register does not support SP for Xd, it uses XZR for the register 31. SUBS with extended register support SP for Xd, but only if SetFlag is not used, otherwise register 31 is Xd.");
        ASSERT_WITH_MESSAGE(!isSp(rm), "No encoding of SUBS supports SP for the third operand.");

        if (isSp(rd) || isSp(rn))
            sub<datasize, setFlags>(rd, rn, rm, UXTX, 0);
        else
            sub<datasize, setFlags>(rd, rn, rm, LSL, 0);
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void sub(RegisterID rd, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        CHECK_DATASIZE();
        insn(addSubtractExtendedRegister(DATASIZE, AddOp_SUB, setFlags, rm, extend, amount, rn, rd));
    }

    template<int datasize, SetFlags setFlags = DontSetFlags>
    ALWAYS_INLINE void sub(RegisterID rd, RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        CHECK_DATASIZE();
        ASSERT(!isSp(rd) && !isSp(rn) && !isSp(rm));
        insn(addSubtractShiftedRegister(DATASIZE, AddOp_SUB, setFlags, shift, rm, amount, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void sxtb(RegisterID rd, RegisterID rn)
    {
        sbfm<datasize>(rd, rn, 0, 7);
    }

    template<int datasize>
    ALWAYS_INLINE void sxth(RegisterID rd, RegisterID rn)
    {
        sbfm<datasize>(rd, rn, 0, 15);
    }

    ALWAYS_INLINE void sxtw(RegisterID rd, RegisterID rn)
    {
        sbfm<64>(rd, rn, 0, 31);
    }

    ALWAYS_INLINE void tbz(RegisterID rt, int imm, int offset = 0)
    {
        ASSERT(!(offset & 3));
        offset >>= 2;
        insn(testAndBranchImmediate(false, imm, offset, rt));
    }

    ALWAYS_INLINE void tbnz(RegisterID rt, int imm, int offset = 0)
    {
        ASSERT(!(offset & 3));
        offset >>= 2;
        insn(testAndBranchImmediate(true, imm, offset, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void tst(RegisterID rn, RegisterID rm)
    {
        and_<datasize, S>(ARM64Registers::zr, rn, rm);
    }

    template<int datasize>
    ALWAYS_INLINE void tst(RegisterID rn, RegisterID rm, ShiftType shift, int amount)
    {
        and_<datasize, S>(ARM64Registers::zr, rn, rm, shift, amount);
    }

    template<int datasize>
    ALWAYS_INLINE void tst(RegisterID rn, LogicalImmediate imm)
    {
        and_<datasize, S>(ARM64Registers::zr, rn, imm);
    }

    template<int datasize>
    ALWAYS_INLINE void ubfiz(RegisterID rd, RegisterID rn, int lsb, int width)
    {
        ubfm<datasize>(rd, rn, (datasize - lsb) & (datasize - 1), width - 1);
    }

    template<int datasize>
    ALWAYS_INLINE void ubfm(RegisterID rd, RegisterID rn, int immr, int imms)
    {
        CHECK_DATASIZE();
        insn(bitfield(DATASIZE, BitfieldOp_UBFM, immr, imms, rn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void ubfx(RegisterID rd, RegisterID rn, int lsb, int width)
    {
        ubfm<datasize>(rd, rn, lsb, lsb + width - 1);
    }

    template<int datasize>
    ALWAYS_INLINE void udiv(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        CHECK_DATASIZE();
        insn(dataProcessing2Source(DATASIZE, rm, DataOp_UDIV, rn, rd));
    }

    ALWAYS_INLINE void umaddl(RegisterID rd, RegisterID rn, RegisterID rm, RegisterID ra)
    {
        nopCortexA53Fix835769<64>();
        insn(dataProcessing3Source(Datasize_64, DataOp_UMADDL, rm, ra, rn, rd));
    }

    ALWAYS_INLINE void umnegl(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        umsubl(rd, rn, rm, ARM64Registers::zr);
    }

    ALWAYS_INLINE void umsubl(RegisterID rd, RegisterID rn, RegisterID rm, RegisterID ra)
    {
        nopCortexA53Fix835769<64>();
        insn(dataProcessing3Source(Datasize_64, DataOp_UMSUBL, rm, ra, rn, rd));
    }

    ALWAYS_INLINE void umulh(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        insn(dataProcessing3Source(Datasize_64, DataOp_UMULH, rm, ARM64Registers::zr, rn, rd));
    }

    ALWAYS_INLINE void umull(RegisterID rd, RegisterID rn, RegisterID rm)
    {
        umaddl(rd, rn, rm, ARM64Registers::zr);
    }

    template<int datasize>
    ALWAYS_INLINE void uxtb(RegisterID rd, RegisterID rn)
    {
        ubfm<datasize>(rd, rn, 0, 7);
    }

    template<int datasize>
    ALWAYS_INLINE void uxth(RegisterID rd, RegisterID rn)
    {
        ubfm<datasize>(rd, rn, 0, 15);
    }

    ALWAYS_INLINE void uxtw(RegisterID rd, RegisterID rn)
    {
        ubfm<64>(rd, rn, 0, 31);
    }

    // Floating Point Instructions:

    template<int datasize>
    ALWAYS_INLINE void fabs(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FABS, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fadd(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing2Source(DATASIZE, vm, FPDataOp_FADD, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fccmp(FPRegisterID vn, FPRegisterID vm, int nzcv, Condition cond)
    {
        CHECK_DATASIZE();
        insn(floatingPointConditionalCompare(DATASIZE, vm, cond, vn, FPCondCmpOp_FCMP, nzcv));
    }

    template<int datasize>
    ALWAYS_INLINE void fccmpe(FPRegisterID vn, FPRegisterID vm, int nzcv, Condition cond)
    {
        CHECK_DATASIZE();
        insn(floatingPointConditionalCompare(DATASIZE, vm, cond, vn, FPCondCmpOp_FCMPE, nzcv));
    }

    template<int datasize>
    ALWAYS_INLINE void fcmp(FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointCompare(DATASIZE, vm, vn, FPCmpOp_FCMP));
    }

    template<int datasize>
    ALWAYS_INLINE void fcmp_0(FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointCompare(DATASIZE, static_cast<FPRegisterID>(0), vn, FPCmpOp_FCMP0));
    }

    template<int datasize>
    ALWAYS_INLINE void fcmpe(FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointCompare(DATASIZE, vm, vn, FPCmpOp_FCMPE));
    }

    template<int datasize>
    ALWAYS_INLINE void fcmpe_0(FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointCompare(DATASIZE, static_cast<FPRegisterID>(0), vn, FPCmpOp_FCMPE0));
    }

    template<int datasize>
    ALWAYS_INLINE void fcsel(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm, Condition cond)
    {
        CHECK_DATASIZE();
        insn(floatingPointConditionalSelect(DATASIZE, vm, cond, vn, vd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvt(FPRegisterID vd, FPRegisterID vn)
    {
        ASSERT(dstsize == 16 || dstsize == 32 || dstsize == 64);
        ASSERT(srcsize == 16 || srcsize == 32 || srcsize == 64);
        ASSERT(dstsize != srcsize);
        Datasize type = (srcsize == 64) ? Datasize_64 : (srcsize == 32) ? Datasize_32 : Datasize_16;
        FPDataOp1Source opcode = (dstsize == 64) ? FPDataOp_FCVT_toDouble : (dstsize == 32) ? FPDataOp_FCVT_toSingle : FPDataOp_FCVT_toHalf;
        insn(floatingPointDataProcessing1Source(type, opcode, vn, vd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvtas(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(dstsize), DATASIZE_OF(srcsize), FPIntConvOp_FCVTAS, vn, rd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvtau(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(dstsize), DATASIZE_OF(srcsize), FPIntConvOp_FCVTAU, vn, rd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvtms(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(dstsize), DATASIZE_OF(srcsize), FPIntConvOp_FCVTMS, vn, rd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvtmu(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(dstsize), DATASIZE_OF(srcsize), FPIntConvOp_FCVTMU, vn, rd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvtns(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(dstsize), DATASIZE_OF(srcsize), FPIntConvOp_FCVTNS, vn, rd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvtnu(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(dstsize), DATASIZE_OF(srcsize), FPIntConvOp_FCVTNU, vn, rd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvtps(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(dstsize), DATASIZE_OF(srcsize), FPIntConvOp_FCVTPS, vn, rd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvtpu(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(dstsize), DATASIZE_OF(srcsize), FPIntConvOp_FCVTPU, vn, rd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvtzs(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(dstsize), DATASIZE_OF(srcsize), FPIntConvOp_FCVTZS, vn, rd));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void fcvtzu(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(dstsize), DATASIZE_OF(srcsize), FPIntConvOp_FCVTZU, vn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void fdiv(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing2Source(DATASIZE, vm, FPDataOp_FDIV, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fmadd(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm, FPRegisterID va)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing3Source(DATASIZE, false, vm, AddOp_ADD, va, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fmax(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing2Source(DATASIZE, vm, FPDataOp_FMAX, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fmaxnm(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing2Source(DATASIZE, vm, FPDataOp_FMAXNM, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fmin(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing2Source(DATASIZE, vm, FPDataOp_FMIN, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fminnm(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing2Source(DATASIZE, vm, FPDataOp_FMINNM, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fmov(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FMOV, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fmov(FPRegisterID vd, RegisterID rn)
    {
        CHECK_DATASIZE();
        insn(floatingPointIntegerConversions(DATASIZE, DATASIZE, FPIntConvOp_FMOV_XtoQ, rn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fmov(RegisterID rd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointIntegerConversions(DATASIZE, DATASIZE, FPIntConvOp_FMOV_QtoX, vn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void fmov(FPRegisterID vd, double imm)
    {
        CHECK_DATASIZE();
        insn(floatingPointImmediate(DATASIZE, encodeFPImm(imm), vd));
    }

    ALWAYS_INLINE void fmov_top(FPRegisterID vd, RegisterID rn)
    {
        insn(floatingPointIntegerConversions(Datasize_64, Datasize_64, FPIntConvOp_FMOV_XtoQ_top, rn, vd));
    }

    ALWAYS_INLINE void fmov_top(RegisterID rd, FPRegisterID vn)
    {
        insn(floatingPointIntegerConversions(Datasize_64, Datasize_64, FPIntConvOp_FMOV_QtoX_top, vn, rd));
    }

    template<int datasize>
    ALWAYS_INLINE void fmsub(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm, FPRegisterID va)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing3Source(DATASIZE, false, vm, AddOp_SUB, va, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fmul(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing2Source(DATASIZE, vm, FPDataOp_FMUL, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fneg(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FNEG, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fnmadd(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm, FPRegisterID va)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing3Source(DATASIZE, true, vm, AddOp_ADD, va, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fnmsub(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm, FPRegisterID va)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing3Source(DATASIZE, true, vm, AddOp_SUB, va, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fnmul(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing2Source(DATASIZE, vm, FPDataOp_FNMUL, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void vand(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_VECTOR_DATASIZE();
        insn(vectorDataProcessing2Source(SIMD_LogicalOp_AND, vm, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void frinta(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FRINTA, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void frinti(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FRINTI, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void frintm(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FRINTM, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void frintn(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FRINTN, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void frintp(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FRINTP, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void frintx(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FRINTX, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void frintz(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FRINTZ, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fsqrt(FPRegisterID vd, FPRegisterID vn)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing1Source(DATASIZE, FPDataOp_FSQRT, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void fsub(FPRegisterID vd, FPRegisterID vn, FPRegisterID vm)
    {
        CHECK_DATASIZE();
        insn(floatingPointDataProcessing2Source(DATASIZE, vm, FPDataOp_FSUB, vn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void ldr(FPRegisterID rt, RegisterID rn, RegisterID rm)
    {
        ldr<datasize>(rt, rn, rm, UXTX, 0);
    }

    template<int datasize>
    ALWAYS_INLINE void ldr(FPRegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        CHECK_FP_MEMOP_DATASIZE();
        insn(loadStoreRegisterRegisterOffset(MEMOPSIZE, true, datasize == 128 ? MemOp_LOAD_V128 : MemOp_LOAD, rm, extend, encodeShiftAmount<datasize>(amount), rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldr(FPRegisterID rt, RegisterID rn, unsigned pimm)
    {
        CHECK_FP_MEMOP_DATASIZE();
        insn(loadStoreRegisterUnsignedImmediate(MEMOPSIZE, true, datasize == 128 ? MemOp_LOAD_V128 : MemOp_LOAD, encodePositiveImmediate<datasize>(pimm), rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldr(FPRegisterID rt, RegisterID rn, PostIndex simm)
    {
        CHECK_FP_MEMOP_DATASIZE();
        insn(loadStoreRegisterPostIndex(MEMOPSIZE, true, datasize == 128 ? MemOp_LOAD_V128 : MemOp_LOAD, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldr(FPRegisterID rt, RegisterID rn, PreIndex simm)
    {
        CHECK_FP_MEMOP_DATASIZE();
        insn(loadStoreRegisterPreIndex(MEMOPSIZE, true, datasize == 128 ? MemOp_LOAD_V128 : MemOp_LOAD, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldr_literal(FPRegisterID rt, int offset = 0)
    {
        CHECK_FP_MEMOP_DATASIZE();
        ASSERT(datasize >= 32);
        ASSERT(!(offset & 3));
        insn(loadRegisterLiteral(datasize == 128 ? LdrLiteralOp_128BIT : datasize == 64 ? LdrLiteralOp_64BIT : LdrLiteralOp_32BIT, true, offset >> 2, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void ldur(FPRegisterID rt, RegisterID rn, int simm)
    {
        CHECK_FP_MEMOP_DATASIZE();
        insn(loadStoreRegisterUnscaledImmediate(MEMOPSIZE, true, datasize == 128 ? MemOp_LOAD_V128 : MemOp_LOAD, simm, rn, rt));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void scvtf(FPRegisterID vd, RegisterID rn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(srcsize), DATASIZE_OF(dstsize), FPIntConvOp_SCVTF, rn, vd));
    }

    template<int datasize>
    ALWAYS_INLINE void str(FPRegisterID rt, RegisterID rn, RegisterID rm)
    {
        str<datasize>(rt, rn, rm, UXTX, 0);
    }

    template<int datasize>
    ALWAYS_INLINE void str(FPRegisterID rt, RegisterID rn, RegisterID rm, ExtendType extend, int amount)
    {
        CHECK_FP_MEMOP_DATASIZE();
        insn(loadStoreRegisterRegisterOffset(MEMOPSIZE, true, datasize == 128 ? MemOp_STORE_V128 : MemOp_STORE, rm, extend, encodeShiftAmount<datasize>(amount), rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void str(FPRegisterID rt, RegisterID rn, unsigned pimm)
    {
        CHECK_FP_MEMOP_DATASIZE();
        insn(loadStoreRegisterUnsignedImmediate(MEMOPSIZE, true, datasize == 128 ? MemOp_STORE_V128 : MemOp_STORE, encodePositiveImmediate<datasize>(pimm), rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void str(FPRegisterID rt, RegisterID rn, PostIndex simm)
    {
        CHECK_FP_MEMOP_DATASIZE();
        insn(loadStoreRegisterPostIndex(MEMOPSIZE, true, datasize == 128 ? MemOp_STORE_V128 : MemOp_STORE, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void str(FPRegisterID rt, RegisterID rn, PreIndex simm)
    {
        CHECK_FP_MEMOP_DATASIZE();
        insn(loadStoreRegisterPreIndex(MEMOPSIZE, true, datasize == 128 ? MemOp_STORE_V128 : MemOp_STORE, simm, rn, rt));
    }

    template<int datasize>
    ALWAYS_INLINE void stur(FPRegisterID rt, RegisterID rn, int simm)
    {
        CHECK_DATASIZE();
        insn(loadStoreRegisterUnscaledImmediate(MEMOPSIZE, true, datasize == 128 ? MemOp_STORE_V128 : MemOp_STORE, simm, rn, rt));
    }

    template<int dstsize, int srcsize>
    ALWAYS_INLINE void ucvtf(FPRegisterID vd, RegisterID rn)
    {
        CHECK_DATASIZE_OF(dstsize);
        CHECK_DATASIZE_OF(srcsize);
        insn(floatingPointIntegerConversions(DATASIZE_OF(srcsize), DATASIZE_OF(dstsize), FPIntConvOp_UCVTF, rn, vd));
    }

    // Admin methods:

    AssemblerLabel labelIgnoringWatchpoints()
    {
        return m_buffer.label();
    }

    AssemblerLabel labelForWatchpoint()
    {
        AssemblerLabel result = m_buffer.label();
        if (static_cast<int>(result.m_offset) != m_indexOfLastWatchpoint)
            result = label();
        m_indexOfLastWatchpoint = result.m_offset;
        m_indexOfTailOfLastWatchpoint = result.m_offset + maxJumpReplacementSize();
        return result;
    }

    AssemblerLabel label()
    {
        AssemblerLabel result = m_buffer.label();
        while (UNLIKELY(static_cast<int>(result.m_offset) < m_indexOfTailOfLastWatchpoint)) {
            nop();
            result = m_buffer.label();
        }
        return result;
    }

    AssemblerLabel align(int alignment)
    {
        ASSERT(!(alignment & 3));
        while (!m_buffer.isAligned(alignment))
            brk(0);
        return label();
    }
    
    static void* getRelocatedAddress(void* code, AssemblerLabel label)
    {
        ASSERT(label.isSet());
        return reinterpret_cast<void*>(reinterpret_cast<ptrdiff_t>(code) + label.m_offset);
    }
    
    static int getDifferenceBetweenLabels(AssemblerLabel a, AssemblerLabel b)
    {
        return b.m_offset - a.m_offset;
    }

    void* unlinkedCode() { return m_buffer.data(); }
    size_t codeSize() const { return m_buffer.codeSize(); }

    static unsigned getCallReturnOffset(AssemblerLabel call)
    {
        ASSERT(call.isSet());
        return call.m_offset;
    }

    // Linking & patching:
    //
    // 'link' and 'patch' methods are for use on unprotected code - such as the code
    // within the AssemblerBuffer, and code being patched by the patch buffer. Once
    // code has been finalized it is (platform support permitting) within a non-
    // writable region of memory; to modify the code in an execute-only execuable
    // pool the 'repatch' and 'relink' methods should be used.

    void linkJump(AssemblerLabel from, AssemblerLabel to, JumpType type, Condition condition)
    {
        ASSERT(to.isSet());
        ASSERT(from.isSet());
        m_jumpsToLink.append(LinkRecord(from.m_offset, to.m_offset, type, condition));
    }

    void linkJump(AssemblerLabel from, AssemblerLabel to, JumpType type, Condition condition, bool is64Bit, RegisterID compareRegister)
    {
        ASSERT(to.isSet());
        ASSERT(from.isSet());
        m_jumpsToLink.append(LinkRecord(from.m_offset, to.m_offset, type, condition, is64Bit, compareRegister));
    }

    void linkJump(AssemblerLabel from, AssemblerLabel to, JumpType type, Condition condition, unsigned bitNumber, RegisterID compareRegister)
    {
        ASSERT(to.isSet());
        ASSERT(from.isSet());
        m_jumpsToLink.append(LinkRecord(from.m_offset, to.m_offset, type, condition, bitNumber, compareRegister));
    }

    void linkJump(AssemblerLabel from, AssemblerLabel to)
    {
        ASSERT(from.isSet());
        ASSERT(to.isSet());
        relinkJumpOrCall<false>(addressOf(from), addressOf(to));
    }
    
    static void linkJump(void* code, AssemblerLabel from, void* to)
    {
        ASSERT(from.isSet());
        relinkJumpOrCall<false>(addressOf(code, from), to);
    }

    static void linkCall(void* code, AssemblerLabel from, void* to)
    {
        ASSERT(from.isSet());
        linkJumpOrCall<true>(addressOf(code, from) - 1, to);
    }

    static void linkPointer(void* code, AssemblerLabel where, void* valuePtr)
    {
        linkPointer(addressOf(code, where), valuePtr);
    }

    static void replaceWithJump(void* where, void* to)
    {
        intptr_t offset = (reinterpret_cast<intptr_t>(to) - reinterpret_cast<intptr_t>(where)) >> 2;
        ASSERT(static_cast<int>(offset) == offset);
        *static_cast<int*>(where) = unconditionalBranchImmediate(false, static_cast<int>(offset));
        cacheFlush(where, sizeof(int));
    }
    
    static ptrdiff_t maxJumpReplacementSize()
    {
        return 4;
    }
    
    static void replaceWithLoad(void* where)
    {
        Datasize sf;
        AddOp op;
        SetFlags S;
        int shift;
        int imm12;
        RegisterID rn;
        RegisterID rd;
        if (disassembleAddSubtractImmediate(where, sf, op, S, shift, imm12, rn, rd)) {
            ASSERT(sf == Datasize_64);
            ASSERT(op == AddOp_ADD);
            ASSERT(!S);
            ASSERT(!shift);
            ASSERT(!(imm12 & ~0xff8));
            *static_cast<int*>(where) = loadStoreRegisterUnsignedImmediate(MemOpSize_64, false, MemOp_LOAD, encodePositiveImmediate<64>(imm12), rn, rd);
            cacheFlush(where, sizeof(int));
        }
#if !ASSERT_DISABLED
        else {
            MemOpSize size;
            bool V;
            MemOp opc;
            int imm12;
            RegisterID rn;
            RegisterID rt;
            ASSERT(disassembleLoadStoreRegisterUnsignedImmediate(where, size, V, opc, imm12, rn, rt));
            ASSERT(size == MemOpSize_64);
            ASSERT(!V);
            ASSERT(opc == MemOp_LOAD);
            ASSERT(!(imm12 & ~0x1ff));
        }
#endif
    }

    static void replaceWithAddressComputation(void* where)
    {
        MemOpSize size;
        bool V;
        MemOp opc;
        int imm12;
        RegisterID rn;
        RegisterID rt;
        if (disassembleLoadStoreRegisterUnsignedImmediate(where, size, V, opc, imm12, rn, rt)) {
            ASSERT(size == MemOpSize_64);
            ASSERT(!V);
            ASSERT(opc == MemOp_LOAD);
            ASSERT(!(imm12 & ~0x1ff));
            *static_cast<int*>(where) = addSubtractImmediate(Datasize_64, AddOp_ADD, DontSetFlags, 0, imm12 * sizeof(void*), rn, rt);
            cacheFlush(where, sizeof(int));
        }
#if !ASSERT_DISABLED
        else {
            Datasize sf;
            AddOp op;
            SetFlags S;
            int shift;
            int imm12;
            RegisterID rn;
            RegisterID rd;
            ASSERT(disassembleAddSubtractImmediate(where, sf, op, S, shift, imm12, rn, rd));
            ASSERT(sf == Datasize_64);
            ASSERT(op == AddOp_ADD);
            ASSERT(!S);
            ASSERT(!shift);
            ASSERT(!(imm12 & ~0xff8));
        }
#endif
    }

    static void repatchPointer(void* where, void* valuePtr)
    {
        linkPointer(static_cast<int*>(where), valuePtr, true);
    }

    static void setPointer(int* address, void* valuePtr, RegisterID rd, bool flush)
    {
        uintptr_t value = reinterpret_cast<uintptr_t>(valuePtr);
        address[0] = moveWideImediate(Datasize_64, MoveWideOp_Z, 0, getHalfword(value, 0), rd);
        address[1] = moveWideImediate(Datasize_64, MoveWideOp_K, 1, getHalfword(value, 1), rd);
        address[2] = moveWideImediate(Datasize_64, MoveWideOp_K, 2, getHalfword(value, 2), rd);

        if (flush)
            cacheFlush(address, sizeof(int) * 3);
    }

    static void repatchInt32(void* where, int32_t value)
    {
        int* address = static_cast<int*>(where);

        Datasize sf;
        MoveWideOp opc;
        int hw;
        uint16_t imm16;
        RegisterID rd;
        bool expected = disassembleMoveWideImediate(address, sf, opc, hw, imm16, rd);
        ASSERT_UNUSED(expected, expected && !sf && (opc == MoveWideOp_Z || opc == MoveWideOp_N) && !hw);
        ASSERT(checkMovk<Datasize_32>(address[1], 1, rd));

        if (value >= 0) {
            address[0] = moveWideImediate(Datasize_32, MoveWideOp_Z, 0, getHalfword(value, 0), rd);
            address[1] = moveWideImediate(Datasize_32, MoveWideOp_K, 1, getHalfword(value, 1), rd);
        } else {
            address[0] = moveWideImediate(Datasize_32, MoveWideOp_N, 0, ~getHalfword(value, 0), rd);
            address[1] = moveWideImediate(Datasize_32, MoveWideOp_K, 1, getHalfword(value, 1), rd);
        }

        cacheFlush(where, sizeof(int) * 2);
    }

    static void* readPointer(void* where)
    {
        int* address = static_cast<int*>(where);

        Datasize sf;
        MoveWideOp opc;
        int hw;
        uint16_t imm16;
        RegisterID rdFirst, rd;

        bool expected = disassembleMoveWideImediate(address, sf, opc, hw, imm16, rdFirst);
        ASSERT_UNUSED(expected, expected && sf && opc == MoveWideOp_Z && !hw);
        uintptr_t result = imm16;

        expected = disassembleMoveWideImediate(address + 1, sf, opc, hw, imm16, rd);
        ASSERT_UNUSED(expected, expected && sf && opc == MoveWideOp_K && hw == 1 && rd == rdFirst);
        result |= static_cast<uintptr_t>(imm16) << 16;

        expected = disassembleMoveWideImediate(address + 2, sf, opc, hw, imm16, rd);
        ASSERT_UNUSED(expected, expected && sf && opc == MoveWideOp_K && hw == 2 && rd == rdFirst);
        result |= static_cast<uintptr_t>(imm16) << 32;

        return reinterpret_cast<void*>(result);
    }

    static void* readCallTarget(void* from)
    {
        return readPointer(reinterpret_cast<int*>(from) - 4);
    }

    static void relinkJump(void* from, void* to)
    {
        relinkJumpOrCall<false>(reinterpret_cast<int*>(from), to);
        cacheFlush(from, sizeof(int));
    }
    
    static void relinkCall(void* from, void* to)
    {
        relinkJumpOrCall<true>(reinterpret_cast<int*>(from) - 1, to);
        cacheFlush(reinterpret_cast<int*>(from) - 1, sizeof(int));
    }
    
    static void repatchCompact(void* where, int32_t value)
    {
        ASSERT(!(value & ~0x3ff8));

        MemOpSize size;
        bool V;
        MemOp opc;
        int imm12;
        RegisterID rn;
        RegisterID rt;
        bool expected = disassembleLoadStoreRegisterUnsignedImmediate(where, size, V, opc, imm12, rn, rt);
        ASSERT_UNUSED(expected, expected && size >= MemOpSize_32 && !V && opc == MemOp_LOAD); // expect 32/64 bit load to GPR.

        if (size == MemOpSize_32)
            imm12 = encodePositiveImmediate<32>(value);
        else
            imm12 = encodePositiveImmediate<64>(value);
        *static_cast<int*>(where) = loadStoreRegisterUnsignedImmediate(size, V, opc, imm12, rn, rt);

        cacheFlush(where, sizeof(int));
    }

    unsigned debugOffset() { return m_buffer.debugOffset(); }

    void recordLinkOffsets(int32_t regionStart, int32_t regionEnd, int32_t offset)
    {
        int32_t ptr = regionStart / sizeof(int32_t);
        const int32_t end = regionEnd / sizeof(int32_t);
        int32_t* offsets = static_cast<int32_t*>(m_buffer.data());
        while (ptr < end)
            offsets[ptr++] = offset;
    }

    int executableOffsetFor(int location)
    {
        if (!location)
            return 0;
        return static_cast<int32_t*>(m_buffer.data())[location / sizeof(int32_t) - 1];
    }

#if OS(LINUX) && COMPILER(GCC_OR_CLANG)
    static inline void linuxPageFlush(uintptr_t begin, uintptr_t end)
    {
        __builtin___clear_cache(reinterpret_cast<char*>(begin), reinterpret_cast<char*>(end));
    }
#endif

    static void cacheFlush(void* code, size_t size)
    {
#if OS(IOS)
        sys_cache_control(kCacheFunctionPrepareForExecution, code, size);
#elif OS(LINUX)
        size_t page = pageSize();
        uintptr_t current = reinterpret_cast<uintptr_t>(code);
        uintptr_t end = current + size;
        uintptr_t firstPageEnd = (current & ~(page - 1)) + page;

        if (end <= firstPageEnd) {
            linuxPageFlush(current, end);
            return;
        }

        linuxPageFlush(current, firstPageEnd);

        for (current = firstPageEnd; current + page < end; current += page)
            linuxPageFlush(current, current + page);

        linuxPageFlush(current, end);
#else
#error "The cacheFlush support is missing on this platform."
#endif
    }

    // Assembler admin methods:

    static int jumpSizeDelta(JumpType jumpType, JumpLinkType jumpLinkType) { return JUMP_ENUM_SIZE(jumpType) - JUMP_ENUM_SIZE(jumpLinkType); }

    static ALWAYS_INLINE bool linkRecordSourceComparator(const LinkRecord& a, const LinkRecord& b)
    {
        return a.from() < b.from();
    }

    static bool canCompact(JumpType jumpType)
    {
        // Fixed jumps cannot be compacted
        return (jumpType == JumpNoCondition) || (jumpType == JumpCondition) || (jumpType == JumpCompareAndBranch) || (jumpType == JumpTestBit);
    }

    static JumpLinkType computeJumpType(JumpType jumpType, const uint8_t* from, const uint8_t* to)
    {
        switch (jumpType) {
        case JumpFixed:
            return LinkInvalid;
        case JumpNoConditionFixedSize:
            return LinkJumpNoCondition;
        case JumpConditionFixedSize:
            return LinkJumpCondition;
        case JumpCompareAndBranchFixedSize:
            return LinkJumpCompareAndBranch;
        case JumpTestBitFixedSize:
            return LinkJumpTestBit;
        case JumpNoCondition:
            return LinkJumpNoCondition;
        case JumpCondition: {
            ASSERT(!(reinterpret_cast<intptr_t>(from) & 0x3));
            ASSERT(!(reinterpret_cast<intptr_t>(to) & 0x3));
            intptr_t relative = reinterpret_cast<intptr_t>(to) - (reinterpret_cast<intptr_t>(from));

            if (((relative << 43) >> 43) == relative)
                return LinkJumpConditionDirect;

            return LinkJumpCondition;
            }
        case JumpCompareAndBranch:  {
            ASSERT(!(reinterpret_cast<intptr_t>(from) & 0x3));
            ASSERT(!(reinterpret_cast<intptr_t>(to) & 0x3));
            intptr_t relative = reinterpret_cast<intptr_t>(to) - (reinterpret_cast<intptr_t>(from));

            if (((relative << 43) >> 43) == relative)
                return LinkJumpCompareAndBranchDirect;

            return LinkJumpCompareAndBranch;
        }
        case JumpTestBit:   {
            ASSERT(!(reinterpret_cast<intptr_t>(from) & 0x3));
            ASSERT(!(reinterpret_cast<intptr_t>(to) & 0x3));
            intptr_t relative = reinterpret_cast<intptr_t>(to) - (reinterpret_cast<intptr_t>(from));

            if (((relative << 50) >> 50) == relative)
                return LinkJumpTestBitDirect;

            return LinkJumpTestBit;
        }
        default:
            ASSERT_NOT_REACHED();
        }

        return LinkJumpNoCondition;
    }

    static JumpLinkType computeJumpType(LinkRecord& record, const uint8_t* from, const uint8_t* to)
    {
        JumpLinkType linkType = computeJumpType(record.type(), from, to);
        record.setLinkType(linkType);
        return linkType;
    }

    Vector<LinkRecord, 0, UnsafeVectorOverflow>& jumpsToLink()
    {
        std::sort(m_jumpsToLink.begin(), m_jumpsToLink.end(), linkRecordSourceComparator);
        return m_jumpsToLink;
    }

    static void ALWAYS_INLINE link(LinkRecord& record, uint8_t* from, uint8_t* to)
    {
        switch (record.linkType()) {
        case LinkJumpNoCondition:
            linkJumpOrCall<false>(reinterpret_cast<int*>(from), to);
            break;
        case LinkJumpConditionDirect:
            linkConditionalBranch<true>(record.condition(), reinterpret_cast<int*>(from), to);
            break;
        case LinkJumpCondition:
            linkConditionalBranch<false>(record.condition(), reinterpret_cast<int*>(from) - 1, to);
            break;
        case LinkJumpCompareAndBranchDirect:
            linkCompareAndBranch<true>(record.condition(), record.is64Bit(), record.compareRegister(), reinterpret_cast<int*>(from), to);
            break;
        case LinkJumpCompareAndBranch:
            linkCompareAndBranch<false>(record.condition(), record.is64Bit(), record.compareRegister(), reinterpret_cast<int*>(from) - 1, to);
            break;
        case LinkJumpTestBitDirect:
            linkTestAndBranch<true>(record.condition(), record.bitNumber(), record.compareRegister(), reinterpret_cast<int*>(from), to);
            break;
        case LinkJumpTestBit:
            linkTestAndBranch<false>(record.condition(), record.bitNumber(), record.compareRegister(), reinterpret_cast<int*>(from) - 1, to);
            break;
        default:
            ASSERT_NOT_REACHED();
            break;
        }
    }

private:
    template<Datasize size>
    static bool checkMovk(int insn, int _hw, RegisterID _rd)
    {
        Datasize sf;
        MoveWideOp opc;
        int hw;
        uint16_t imm16;
        RegisterID rd;
        bool expected = disassembleMoveWideImediate(&insn, sf, opc, hw, imm16, rd);

        return expected
            && sf == size
            && opc == MoveWideOp_K
            && hw == _hw
            && rd == _rd;
    }

    static void linkPointer(int* address, void* valuePtr, bool flush = false)
    {
        Datasize sf;
        MoveWideOp opc;
        int hw;
        uint16_t imm16;
        RegisterID rd;
        bool expected = disassembleMoveWideImediate(address, sf, opc, hw, imm16, rd);
        ASSERT_UNUSED(expected, expected && sf && opc == MoveWideOp_Z && !hw);
        ASSERT(checkMovk<Datasize_64>(address[1], 1, rd));
        ASSERT(checkMovk<Datasize_64>(address[2], 2, rd));

        setPointer(address, valuePtr, rd, flush);
    }

    template<bool isCall>
    static void linkJumpOrCall(int* from, void* to)
    {
        bool link;
        int imm26;
        bool isUnconditionalBranchImmediateOrNop = disassembleUnconditionalBranchImmediate(from, link, imm26) || disassembleNop(from);

        ASSERT_UNUSED(isUnconditionalBranchImmediateOrNop, isUnconditionalBranchImmediateOrNop);
        ASSERT_UNUSED(isCall, (link == isCall) || disassembleNop(from));
        ASSERT(!(reinterpret_cast<intptr_t>(from) & 3));
        ASSERT(!(reinterpret_cast<intptr_t>(to) & 3));
        intptr_t offset = (reinterpret_cast<intptr_t>(to) - reinterpret_cast<intptr_t>(from)) >> 2;
        ASSERT(static_cast<int>(offset) == offset);

        *from = unconditionalBranchImmediate(isCall, static_cast<int>(offset));
    }

    template<bool isDirect>
    static void linkCompareAndBranch(Condition condition, bool is64Bit, RegisterID rt, int* from, void* to)
    {
        ASSERT(!(reinterpret_cast<intptr_t>(from) & 3));
        ASSERT(!(reinterpret_cast<intptr_t>(to) & 3));
        intptr_t offset = (reinterpret_cast<intptr_t>(to) - reinterpret_cast<intptr_t>(from)) >> 2;
        ASSERT(((offset << 38) >> 38) == offset);

        bool useDirect = ((offset << 45) >> 45) == offset; // Fits in 19 bits
        ASSERT(!isDirect || useDirect);

        if (useDirect || isDirect) {
            *from = compareAndBranchImmediate(is64Bit ? Datasize_64 : Datasize_32, condition == ConditionNE, static_cast<int>(offset), rt);
            if (!isDirect)
                *(from + 1) = nopPseudo();
        } else {
            *from = compareAndBranchImmediate(is64Bit ? Datasize_64 : Datasize_32, invert(condition) == ConditionNE, 2, rt);
            linkJumpOrCall<false>(from + 1, to);
        }
    }

    template<bool isDirect>
    static void linkConditionalBranch(Condition condition, int* from, void* to)
    {
        ASSERT(!(reinterpret_cast<intptr_t>(from) & 3));
        ASSERT(!(reinterpret_cast<intptr_t>(to) & 3));
        intptr_t offset = (reinterpret_cast<intptr_t>(to) - reinterpret_cast<intptr_t>(from)) >> 2;
        ASSERT(((offset << 38) >> 38) == offset);

        bool useDirect = ((offset << 45) >> 45) == offset; // Fits in 19 bits
        ASSERT(!isDirect || useDirect);

        if (useDirect || isDirect) {
            *from = conditionalBranchImmediate(static_cast<int>(offset), condition);
            if (!isDirect)
                *(from + 1) = nopPseudo();
        } else {
            *from = conditionalBranchImmediate(2, invert(condition));
            linkJumpOrCall<false>(from + 1, to);
        }
    }

    template<bool isDirect>
    static void linkTestAndBranch(Condition condition, unsigned bitNumber, RegisterID rt, int* from, void* to)
    {
        ASSERT(!(reinterpret_cast<intptr_t>(from) & 3));
        ASSERT(!(reinterpret_cast<intptr_t>(to) & 3));
        intptr_t offset = (reinterpret_cast<intptr_t>(to) - reinterpret_cast<intptr_t>(from)) >> 2;
        ASSERT(static_cast<int>(offset) == offset);
        ASSERT(((offset << 38) >> 38) == offset);

        bool useDirect = ((offset << 50) >> 50) == offset; // Fits in 14 bits
        ASSERT(!isDirect || useDirect);

        if (useDirect || isDirect) {
            *from = testAndBranchImmediate(condition == ConditionNE, static_cast<int>(bitNumber), static_cast<int>(offset), rt);
            if (!isDirect)
                *(from + 1) = nopPseudo();
        } else {
            *from = testAndBranchImmediate(invert(condition) == ConditionNE, static_cast<int>(bitNumber), 2, rt);
            linkJumpOrCall<false>(from + 1, to);
        }
    }

    template<bool isCall>
    static void relinkJumpOrCall(int* from, void* to)
    {
        if (!isCall && disassembleNop(from)) {
            unsigned op01;
            int imm19;
            Condition condition;
            bool isConditionalBranchImmediate = disassembleConditionalBranchImmediate(from - 1, op01, imm19, condition);

            if (isConditionalBranchImmediate) {
                ASSERT_UNUSED(op01, !op01);
                ASSERT_UNUSED(isCall, !isCall);

                if (imm19 == 8)
                    condition = invert(condition);

                linkConditionalBranch<false>(condition, from - 1, to);
                return;
            }

            Datasize opSize;
            bool op;
            RegisterID rt;
            bool isCompareAndBranchImmediate = disassembleCompareAndBranchImmediate(from - 1, opSize, op, imm19, rt);

            if (isCompareAndBranchImmediate) {
                if (imm19 == 8)
                    op = !op;

                linkCompareAndBranch<false>(op ? ConditionNE : ConditionEQ, opSize == Datasize_64, rt, from - 1, to);
                return;
            }

            int imm14;
            unsigned bitNumber;
            bool isTestAndBranchImmediate = disassembleTestAndBranchImmediate(from - 1, op, bitNumber, imm14, rt);

            if (isTestAndBranchImmediate) {
                if (imm14 == 8)
                    op = !op;

                linkTestAndBranch<false>(op ? ConditionNE : ConditionEQ, bitNumber, rt, from - 1, to);
                return;
            }
        }

        linkJumpOrCall<isCall>(from, to);
    }

    static int* addressOf(void* code, AssemblerLabel label)
    {
        return reinterpret_cast<int*>(static_cast<char*>(code) + label.m_offset);
    }

    int* addressOf(AssemblerLabel label)
    {
        return addressOf(m_buffer.data(), label);
    }

    static RegisterID disassembleXOrSp(int reg) { return reg == 31 ? ARM64Registers::sp : static_cast<RegisterID>(reg); }
    static RegisterID disassembleXOrZr(int reg) { return reg == 31 ? ARM64Registers::zr : static_cast<RegisterID>(reg); }
    static RegisterID disassembleXOrZrOrSp(bool useZr, int reg) { return reg == 31 ? (useZr ? ARM64Registers::zr : ARM64Registers::sp) : static_cast<RegisterID>(reg); }

    static bool disassembleAddSubtractImmediate(void* address, Datasize& sf, AddOp& op, SetFlags& S, int& shift, int& imm12, RegisterID& rn, RegisterID& rd)
    {
        int insn = *static_cast<int*>(address);
        sf = static_cast<Datasize>((insn >> 31) & 1);
        op = static_cast<AddOp>((insn >> 30) & 1);
        S = static_cast<SetFlags>((insn >> 29) & 1);
        shift = (insn >> 22) & 3;
        imm12 = (insn >> 10) & 0x3ff;
        rn = disassembleXOrSp((insn >> 5) & 0x1f);
        rd = disassembleXOrZrOrSp(S, insn & 0x1f);
        return (insn & 0x1f000000) == 0x11000000;
    }

    static bool disassembleLoadStoreRegisterUnsignedImmediate(void* address, MemOpSize& size, bool& V, MemOp& opc, int& imm12, RegisterID& rn, RegisterID& rt)
    {
        int insn = *static_cast<int*>(address);
        size = static_cast<MemOpSize>((insn >> 30) & 3);
        V = (insn >> 26) & 1;
        opc = static_cast<MemOp>((insn >> 22) & 3);
        imm12 = (insn >> 10) & 0xfff;
        rn = disassembleXOrSp((insn >> 5) & 0x1f);
        rt = disassembleXOrZr(insn & 0x1f);
        return (insn & 0x3b000000) == 0x39000000;
    }

    static bool disassembleMoveWideImediate(void* address, Datasize& sf, MoveWideOp& opc, int& hw, uint16_t& imm16, RegisterID& rd)
    {
        int insn = *static_cast<int*>(address);
        sf = static_cast<Datasize>((insn >> 31) & 1);
        opc = static_cast<MoveWideOp>((insn >> 29) & 3);
        hw = (insn >> 21) & 3;
        imm16 = insn >> 5;
        rd = disassembleXOrZr(insn & 0x1f);
        return (insn & 0x1f800000) == 0x12800000;
    }

    static bool disassembleNop(void* address)
    {
        unsigned insn = *static_cast<unsigned*>(address);
        return insn == 0xd503201f;
    }

    static bool disassembleCompareAndBranchImmediate(void* address, Datasize& sf, bool& op, int& imm19, RegisterID& rt)
    {
        int insn = *static_cast<int*>(address);
        sf = static_cast<Datasize>((insn >> 31) & 1);
        op = (insn >> 24) & 0x1;
        imm19 = (insn << 8) >> 13;
        rt = static_cast<RegisterID>(insn & 0x1f);
        return (insn & 0x7e000000) == 0x34000000;
        
    }

    static bool disassembleConditionalBranchImmediate(void* address, unsigned& op01, int& imm19, Condition &condition)
    {
        int insn = *static_cast<int*>(address);
        op01 = ((insn >> 23) & 0x2) | ((insn >> 4) & 0x1);
        imm19 = (insn << 8) >> 13;
        condition = static_cast<Condition>(insn & 0xf);
        return (insn & 0xfe000000) == 0x54000000;
    }

    static bool disassembleTestAndBranchImmediate(void* address, bool& op, unsigned& bitNumber, int& imm14, RegisterID& rt)
    {
        int insn = *static_cast<int*>(address);
        op = (insn >> 24) & 0x1;
        imm14 = (insn << 13) >> 18;
        bitNumber = static_cast<unsigned>((((insn >> 26) & 0x20)) | ((insn >> 19) & 0x1f));
        rt = static_cast<RegisterID>(insn & 0x1f);
        return (insn & 0x7e000000) == 0x36000000;
        
    }

    static bool disassembleUnconditionalBranchImmediate(void* address, bool& op, int& imm26)
    {
        int insn = *static_cast<int*>(address);
        op = (insn >> 31) & 1;
        imm26 = (insn << 6) >> 6;
        return (insn & 0x7c000000) == 0x14000000;
    }

    static int xOrSp(RegisterID reg) { ASSERT(!isZr(reg)); return reg; }
    static int xOrZr(RegisterID reg) { ASSERT(!isSp(reg)); return reg & 31; }
    static FPRegisterID xOrZrAsFPR(RegisterID reg) { return static_cast<FPRegisterID>(xOrZr(reg)); }
    static int xOrZrOrSp(bool useZr, RegisterID reg) { return useZr ? xOrZr(reg) : xOrSp(reg); }

    ALWAYS_INLINE void insn(int instruction)
    {
        m_buffer.putInt(instruction);
    }

    ALWAYS_INLINE static int addSubtractExtendedRegister(Datasize sf, AddOp op, SetFlags S, RegisterID rm, ExtendType option, int imm3, RegisterID rn, RegisterID rd)
    {
        ASSERT(imm3 < 5);
        // The only allocated values for opt is 0.
        const int opt = 0;
        return (0x0b200000 | sf << 31 | op << 30 | S << 29 | opt << 22 | xOrZr(rm) << 16 | option << 13 | (imm3 & 0x7) << 10 | xOrSp(rn) << 5 | xOrZrOrSp(S, rd));
    }

    ALWAYS_INLINE static int addSubtractImmediate(Datasize sf, AddOp op, SetFlags S, int shift, int imm12, RegisterID rn, RegisterID rd)
    {
        ASSERT(shift < 2);
        ASSERT(isUInt12(imm12));
        return (0x11000000 | sf << 31 | op << 30 | S << 29 | shift << 22 | (imm12 & 0xfff) << 10 | xOrSp(rn) << 5 | xOrZrOrSp(S, rd));
    }

    ALWAYS_INLINE static int addSubtractShiftedRegister(Datasize sf, AddOp op, SetFlags S, ShiftType shift, RegisterID rm, int imm6, RegisterID rn, RegisterID rd)
    {
        ASSERT(shift < 3);
        ASSERT(!(imm6 & (sf ? ~63 : ~31)));
        return (0x0b000000 | sf << 31 | op << 30 | S << 29 | shift << 22 | xOrZr(rm) << 16 | (imm6 & 0x3f) << 10 | xOrZr(rn) << 5 | xOrZr(rd));
    }

    ALWAYS_INLINE static int addSubtractWithCarry(Datasize sf, AddOp op, SetFlags S, RegisterID rm, RegisterID rn, RegisterID rd)
    {
        const int opcode2 = 0;
        return (0x1a000000 | sf << 31 | op << 30 | S << 29 | xOrZr(rm) << 16 | opcode2 << 10 | xOrZr(rn) << 5 | xOrZr(rd));
    }

    ALWAYS_INLINE static int bitfield(Datasize sf, BitfieldOp opc, int immr, int imms, RegisterID rn, RegisterID rd)
    {
        ASSERT(immr < (sf ? 64 : 32));
        ASSERT(imms < (sf ? 64 : 32));
        const int N = sf;
        return (0x13000000 | sf << 31 | opc << 29 | N << 22 | immr << 16 | imms << 10 | xOrZr(rn) << 5 | xOrZr(rd));
    }

    // 'op' means negate
    ALWAYS_INLINE static int compareAndBranchImmediate(Datasize sf, bool op, int32_t imm19, RegisterID rt)
    {
        ASSERT(imm19 == (imm19 << 13) >> 13);
        return (0x34000000 | sf << 31 | op << 24 | (imm19 & 0x7ffff) << 5 | xOrZr(rt));
    }

    ALWAYS_INLINE static int conditionalBranchImmediate(int32_t imm19, Condition cond)
    {
        ASSERT(imm19 == (imm19 << 13) >> 13);
        ASSERT(!(cond & ~15));
        // The only allocated values for o1 & o0 are 0.
        const int o1 = 0;
        const int o0 = 0;
        return (0x54000000 | o1 << 24 | (imm19 & 0x7ffff) << 5 | o0 << 4 | cond);
    }

    ALWAYS_INLINE static int conditionalCompareImmediate(Datasize sf, AddOp op, int imm5, Condition cond, RegisterID rn, int nzcv)
    {
        ASSERT(!(imm5 & ~0x1f));
        ASSERT(nzcv < 16);
        const int S = 1;
        const int o2 = 0;
        const int o3 = 0;
        return (0x1a400800 | sf << 31 | op << 30 | S << 29 | (imm5 & 0x1f) << 16 | cond << 12 | o2 << 10 | xOrZr(rn) << 5 | o3 << 4 | nzcv);
    }

    ALWAYS_INLINE static int conditionalCompareRegister(Datasize sf, AddOp op, RegisterID rm, Condition cond, RegisterID rn, int nzcv)
    {
        ASSERT(nzcv < 16);
        const int S = 1;
        const int o2 = 0;
        const int o3 = 0;
        return (0x1a400000 | sf << 31 | op << 30 | S << 29 | xOrZr(rm) << 16 | cond << 12 | o2 << 10 | xOrZr(rn) << 5 | o3 << 4 | nzcv);
    }

    // 'op' means negate
    // 'op2' means increment
    ALWAYS_INLINE static int conditionalSelect(Datasize sf, bool op, RegisterID rm, Condition cond, bool op2, RegisterID rn, RegisterID rd)
    {
        const int S = 0;
        return (0x1a800000 | sf << 31 | op << 30 | S << 29 | xOrZr(rm) << 16 | cond << 12 | op2 << 10 | xOrZr(rn) << 5 | xOrZr(rd));
    }

    ALWAYS_INLINE static int dataProcessing1Source(Datasize sf, DataOp1Source opcode, RegisterID rn, RegisterID rd)
    {
        const int S = 0;
        const int opcode2 = 0;
        return (0x5ac00000 | sf << 31 | S << 29 | opcode2 << 16 | opcode << 10 | xOrZr(rn) << 5 | xOrZr(rd));
    }

    ALWAYS_INLINE static int dataProcessing2Source(Datasize sf, RegisterID rm, DataOp2Source opcode, RegisterID rn, RegisterID rd)
    {
        const int S = 0;
        return (0x1ac00000 | sf << 31 | S << 29 | xOrZr(rm) << 16 | opcode << 10 | xOrZr(rn) << 5 | xOrZr(rd));
    }

    ALWAYS_INLINE static int dataProcessing3Source(Datasize sf, DataOp3Source opcode, RegisterID rm, RegisterID ra, RegisterID rn, RegisterID rd)
    {
        int op54 = opcode >> 4;
        int op31 = (opcode >> 1) & 7;
        int op0 = opcode & 1;
        return (0x1b000000 | sf << 31 | op54 << 29 | op31 << 21 | xOrZr(rm) << 16 | op0 << 15 | xOrZr(ra) << 10 | xOrZr(rn) << 5 | xOrZr(rd));
    }

    ALWAYS_INLINE static int excepnGeneration(ExcepnOp opc, uint16_t imm16, int LL)
    {
        ASSERT((opc == ExcepnOp_BREAKPOINT || opc == ExcepnOp_HALT) ? !LL : (LL && (LL < 4)));
        const int op2 = 0;
        return (0xd4000000 | opc << 21 | imm16 << 5 | op2 << 2 | LL);
    }

    ALWAYS_INLINE static int extract(Datasize sf, RegisterID rm, int imms, RegisterID rn, RegisterID rd)
    {
        ASSERT(imms < (sf ? 64 : 32));
        const int op21 = 0;
        const int N = sf;
        const int o0 = 0;
        return (0x13800000 | sf << 31 | op21 << 29 | N << 22 | o0 << 21 | xOrZr(rm) << 16 | imms << 10 | xOrZr(rn) << 5 | xOrZr(rd));
    }

    ALWAYS_INLINE static int floatingPointCompare(Datasize type, FPRegisterID rm, FPRegisterID rn, FPCmpOp opcode2)
    {
        const int M = 0;
        const int S = 0;
        const int op = 0;
        return (0x1e202000 | M << 31 | S << 29 | type << 22 | rm << 16 | op << 14 | rn << 5 | opcode2);
    }

    ALWAYS_INLINE static int floatingPointConditionalCompare(Datasize type, FPRegisterID rm, Condition cond, FPRegisterID rn, FPCondCmpOp op, int nzcv)
    {
        ASSERT(nzcv < 16);
        const int M = 0;
        const int S = 0;
        return (0x1e200400 | M << 31 | S << 29 | type << 22 | rm << 16 | cond << 12 | rn << 5 | op << 4 | nzcv);
    }

    ALWAYS_INLINE static int floatingPointConditionalSelect(Datasize type, FPRegisterID rm, Condition cond, FPRegisterID rn, FPRegisterID rd)
    {
        const int M = 0;
        const int S = 0;
        return (0x1e200c00 | M << 31 | S << 29 | type << 22 | rm << 16 | cond << 12 | rn << 5 | rd);
    }

    ALWAYS_INLINE static int floatingPointImmediate(Datasize type, int imm8, FPRegisterID rd)
    {
        const int M = 0;
        const int S = 0;
        const int imm5 = 0;
        return (0x1e201000 | M << 31 | S << 29 | type << 22 | (imm8 & 0xff) << 13 | imm5 << 5 | rd);
    }

    ALWAYS_INLINE static int floatingPointIntegerConversions(Datasize sf, Datasize type, FPIntConvOp rmodeOpcode, FPRegisterID rn, FPRegisterID rd)
    {
        const int S = 0;
        return (0x1e200000 | sf << 31 | S << 29 | type << 22 | rmodeOpcode << 16 | rn << 5 | rd);
    }

    ALWAYS_INLINE static int floatingPointIntegerConversions(Datasize sf, Datasize type, FPIntConvOp rmodeOpcode, FPRegisterID rn, RegisterID rd)
    {
        return floatingPointIntegerConversions(sf, type, rmodeOpcode, rn, xOrZrAsFPR(rd));
    }

    ALWAYS_INLINE static int floatingPointIntegerConversions(Datasize sf, Datasize type, FPIntConvOp rmodeOpcode, RegisterID rn, FPRegisterID rd)
    {
        return floatingPointIntegerConversions(sf, type, rmodeOpcode, xOrZrAsFPR(rn), rd);
    }

    ALWAYS_INLINE static int floatingPointDataProcessing1Source(Datasize type, FPDataOp1Source opcode, FPRegisterID rn, FPRegisterID rd)
    {
        const int M = 0;
        const int S = 0;
        return (0x1e204000 | M << 31 | S << 29 | type << 22 | opcode << 15 | rn << 5 | rd);
    }

    ALWAYS_INLINE static int floatingPointDataProcessing2Source(Datasize type, FPRegisterID rm, FPDataOp2Source opcode, FPRegisterID rn, FPRegisterID rd)
    {
        const int M = 0;
        const int S = 0;
        return (0x1e200800 | M << 31 | S << 29 | type << 22 | rm << 16 | opcode << 12 | rn << 5 | rd);
    }

    ALWAYS_INLINE static int vectorDataProcessing2Source(SIMD3Same opcode, unsigned size, FPRegisterID vm, FPRegisterID vn, FPRegisterID vd)
    {
        const int Q = 0;
        return (0xe201c00 | Q << 30 | size << 22 | vm << 16 | opcode << 11 | vn << 5 | vd);
    }

    ALWAYS_INLINE static int vectorDataProcessing2Source(SIMD3Same opcode, FPRegisterID vm, FPRegisterID vn, FPRegisterID vd)
    {
        return vectorDataProcessing2Source(opcode, 0, vm, vn, vd);
    }


    // 'o1' means negate
    ALWAYS_INLINE static int floatingPointDataProcessing3Source(Datasize type, bool o1, FPRegisterID rm, AddOp o2, FPRegisterID ra, FPRegisterID rn, FPRegisterID rd)
    {
        const int M = 0;
        const int S = 0;
        return (0x1f000000 | M << 31 | S << 29 | type << 22 | o1 << 21 | rm << 16 | o2 << 15 | ra << 10 | rn << 5 | rd);
    }

    // 'V' means vector
    ALWAYS_INLINE static int loadRegisterLiteral(LdrLiteralOp opc, bool V, int imm19, FPRegisterID rt)
    {
        ASSERT(((imm19 << 13) >> 13) == imm19);
        return (0x18000000 | opc << 30 | V << 26 | (imm19 & 0x7ffff) << 5 | rt);
    }

    ALWAYS_INLINE static int loadRegisterLiteral(LdrLiteralOp opc, bool V, int imm19, RegisterID rt)
    {
        return loadRegisterLiteral(opc, V, imm19, xOrZrAsFPR(rt));
    }

    // 'V' means vector
    ALWAYS_INLINE static int loadStoreRegisterPostIndex(MemOpSize size, bool V, MemOp opc, int imm9, RegisterID rn, FPRegisterID rt)
    {
        ASSERT(!(size && V && (opc & 2))); // Maximum vector size is 128 bits.
        ASSERT(!((size & 2) && !V && (opc == 3))); // signed 32-bit load must be extending from 8/16 bits.
        ASSERT(isInt9(imm9));
        return (0x38000400 | size << 30 | V << 26 | opc << 22 | (imm9 & 0x1ff) << 12 | xOrSp(rn) << 5 | rt);
    }

    ALWAYS_INLINE static int loadStoreRegisterPostIndex(MemOpSize size, bool V, MemOp opc, int imm9, RegisterID rn, RegisterID rt)
    {
        return loadStoreRegisterPostIndex(size, V, opc, imm9, rn, xOrZrAsFPR(rt));
    }

    // 'V' means vector
    ALWAYS_INLINE static int loadStoreRegisterPairPostIndex(MemPairOpSize size, bool V, MemOp opc, int immediate, RegisterID rn, FPRegisterID rt, FPRegisterID rt2)
    {
        ASSERT(size < 3);
        ASSERT(opc == (opc & 1)); // Only load or store, load signed 64 is handled via size.
        ASSERT(V || (size != MemPairOp_LoadSigned_32) || (opc == MemOp_LOAD)); // There isn't an integer store signed.
        unsigned immedShiftAmount = memPairOffsetShift(V, size);
        int imm7 = immediate >> immedShiftAmount;
        ASSERT((imm7 << immedShiftAmount) == immediate && isInt7(imm7));
        return (0x28800000 | size << 30 | V << 26 | opc << 22 | (imm7 & 0x7f) << 15 | rt2 << 10 | xOrSp(rn) << 5 | rt);
    }

    ALWAYS_INLINE static int loadStoreRegisterPairPostIndex(MemPairOpSize size, bool V, MemOp opc, int immediate, RegisterID rn, RegisterID rt, RegisterID rt2)
    {
        return loadStoreRegisterPairPostIndex(size, V, opc, immediate, rn, xOrZrAsFPR(rt), xOrZrAsFPR(rt2));
    }

    // 'V' means vector
    ALWAYS_INLINE static int loadStoreRegisterPreIndex(MemOpSize size, bool V, MemOp opc, int imm9, RegisterID rn, FPRegisterID rt)
    {
        ASSERT(!(size && V && (opc & 2))); // Maximum vector size is 128 bits.
        ASSERT(!((size & 2) && !V && (opc == 3))); // signed 32-bit load must be extending from 8/16 bits.
        ASSERT(isInt9(imm9));
        return (0x38000c00 | size << 30 | V << 26 | opc << 22 | (imm9 & 0x1ff) << 12 | xOrSp(rn) << 5 | rt);
    }

    ALWAYS_INLINE static int loadStoreRegisterPreIndex(MemOpSize size, bool V, MemOp opc, int imm9, RegisterID rn, RegisterID rt)
    {
        return loadStoreRegisterPreIndex(size, V, opc, imm9, rn, xOrZrAsFPR(rt));
    }

    // 'V' means vector
    ALWAYS_INLINE static int loadStoreRegisterPairPreIndex(MemPairOpSize size, bool V, MemOp opc, int immediate, RegisterID rn, FPRegisterID rt, FPRegisterID rt2)
    {
        ASSERT(size < 3);
        ASSERT(opc == (opc & 1)); // Only load or store, load signed 64 is handled via size.
        ASSERT(V || (size != MemPairOp_LoadSigned_32) || (opc == MemOp_LOAD)); // There isn't an integer store signed.
        unsigned immedShiftAmount = memPairOffsetShift(V, size);
        int imm7 = immediate >> immedShiftAmount;
        ASSERT((imm7 << immedShiftAmount) == immediate && isInt7(imm7));
        return (0x29800000 | size << 30 | V << 26 | opc << 22 | (imm7 & 0x7f) << 15 | rt2 << 10 | xOrSp(rn) << 5 | rt);
    }

    ALWAYS_INLINE static int loadStoreRegisterPairPreIndex(MemPairOpSize size, bool V, MemOp opc, int immediate, RegisterID rn, RegisterID rt, RegisterID rt2)
    {
        return loadStoreRegisterPairPreIndex(size, V, opc, immediate, rn, xOrZrAsFPR(rt), xOrZrAsFPR(rt2));
    }

    // 'V' means vector
    // 'S' means shift rm
    ALWAYS_INLINE static int loadStoreRegisterRegisterOffset(MemOpSize size, bool V, MemOp opc, RegisterID rm, ExtendType option, bool S, RegisterID rn, FPRegisterID rt)
    {
        ASSERT(!(size && V && (opc & 2))); // Maximum vector size is 128 bits.
        ASSERT(!((size & 2) && !V && (opc == 3))); // signed 32-bit load must be extending from 8/16 bits.
        ASSERT(option & 2); // The ExtendType for the address must be 32/64 bit, signed or unsigned - not 8/16bit.
        return (0x38200800 | size << 30 | V << 26 | opc << 22 | xOrZr(rm) << 16 | option << 13 | S << 12 | xOrSp(rn) << 5 | rt);
    }

    ALWAYS_INLINE static int loadStoreRegisterRegisterOffset(MemOpSize size, bool V, MemOp opc, RegisterID rm, ExtendType option, bool S, RegisterID rn, RegisterID rt)
    {
        return loadStoreRegisterRegisterOffset(size, V, opc, rm, option, S, rn, xOrZrAsFPR(rt));
    }

    // 'V' means vector
    ALWAYS_INLINE static int loadStoreRegisterUnscaledImmediate(MemOpSize size, bool V, MemOp opc, int imm9, RegisterID rn, FPRegisterID rt)
    {
        ASSERT(!(size && V && (opc & 2))); // Maximum vector size is 128 bits.
        ASSERT(!((size & 2) && !V && (opc == 3))); // signed 32-bit load must be extending from 8/16 bits.
        ASSERT(isInt9(imm9));
        return (0x38000000 | size << 30 | V << 26 | opc << 22 | (imm9 & 0x1ff) << 12 | xOrSp(rn) << 5 | rt);
    }

    ALWAYS_INLINE static int loadStoreRegisterUnscaledImmediate(MemOpSize size, bool V, MemOp opc, int imm9, RegisterID rn, RegisterID rt)
    {
        ASSERT(isInt9(imm9));
        return loadStoreRegisterUnscaledImmediate(size, V, opc, imm9, rn, xOrZrAsFPR(rt));
    }

    // 'V' means vector
    ALWAYS_INLINE static int loadStoreRegisterUnsignedImmediate(MemOpSize size, bool V, MemOp opc, int imm12, RegisterID rn, FPRegisterID rt)
    {
        ASSERT(!(size && V && (opc & 2))); // Maximum vector size is 128 bits.
        ASSERT(!((size & 2) && !V && (opc == 3))); // signed 32-bit load must be extending from 8/16 bits.
        ASSERT(isUInt12(imm12));
        return (0x39000000 | size << 30 | V << 26 | opc << 22 | (imm12 & 0xfff) << 10 | xOrSp(rn) << 5 | rt);
    }

    ALWAYS_INLINE static int loadStoreRegisterUnsignedImmediate(MemOpSize size, bool V, MemOp opc, int imm12, RegisterID rn, RegisterID rt)
    {
        return loadStoreRegisterUnsignedImmediate(size, V, opc, imm12, rn, xOrZrAsFPR(rt));
    }

    ALWAYS_INLINE static int logicalImmediate(Datasize sf, LogicalOp opc, int N_immr_imms, RegisterID rn, RegisterID rd)
    {
        ASSERT(!(N_immr_imms & (sf ? ~0x1fff : ~0xfff)));
        return (0x12000000 | sf << 31 | opc << 29 | N_immr_imms << 10 | xOrZr(rn) << 5 | xOrZrOrSp(opc == LogicalOp_ANDS, rd));
    }

    // 'N' means negate rm
    ALWAYS_INLINE static int logicalShiftedRegister(Datasize sf, LogicalOp opc, ShiftType shift, bool N, RegisterID rm, int imm6, RegisterID rn, RegisterID rd)
    {
        ASSERT(!(imm6 & (sf ? ~63 : ~31)));
        return (0x0a000000 | sf << 31 | opc << 29 | shift << 22 | N << 21 | xOrZr(rm) << 16 | (imm6 & 0x3f) << 10 | xOrZr(rn) << 5 | xOrZr(rd));
    }

    ALWAYS_INLINE static int moveWideImediate(Datasize sf, MoveWideOp opc, int hw, uint16_t imm16, RegisterID rd)
    {
        ASSERT(hw < (sf ? 4 : 2));
        return (0x12800000 | sf << 31 | opc << 29 | hw << 21 | (int)imm16 << 5 | xOrZr(rd));
    }

    // 'op' means link
    ALWAYS_INLINE static int unconditionalBranchImmediate(bool op, int32_t imm26)
    {
        ASSERT(imm26 == (imm26 << 6) >> 6);
        return (0x14000000 | op << 31 | (imm26 & 0x3ffffff));
    }

    // 'op' means page
    ALWAYS_INLINE static int pcRelative(bool op, int32_t imm21, RegisterID rd)
    {
        ASSERT(imm21 == (imm21 << 11) >> 11);
        int32_t immlo = imm21 & 3;
        int32_t immhi = (imm21 >> 2) & 0x7ffff;
        return (0x10000000 | op << 31 | immlo << 29 | immhi << 5 | xOrZr(rd));
    }

    ALWAYS_INLINE static int system(bool L, int op0, int op1, int crn, int crm, int op2, RegisterID rt)
    {
        return (0xd5000000 | L << 21 | op0 << 19 | op1 << 16 | crn << 12 | crm << 8 | op2 << 5 | xOrZr(rt));
    }

    ALWAYS_INLINE static int hintPseudo(int imm)
    {
        ASSERT(!(imm & ~0x7f));
        return system(0, 0, 3, 2, (imm >> 3) & 0xf, imm & 0x7, ARM64Registers::zr);
    }

    ALWAYS_INLINE static int nopPseudo()
    {
        return hintPseudo(0);
    }
    
    // 'op' means negate
    ALWAYS_INLINE static int testAndBranchImmediate(bool op, int b50, int imm14, RegisterID rt)
    {
        ASSERT(!(b50 & ~0x3f));
        ASSERT(imm14 == (imm14 << 18) >> 18);
        int b5 = b50 >> 5;
        int b40 = b50 & 0x1f;
        return (0x36000000 | b5 << 31 | op << 24 | b40 << 19 | (imm14 & 0x3fff) << 5 | xOrZr(rt));
    }

    ALWAYS_INLINE static int unconditionalBranchRegister(BranchType opc, RegisterID rn)
    {
        // The only allocated values for op2 is 0x1f, for op3 & op4 are 0.
        const int op2 = 0x1f;
        const int op3 = 0;
        const int op4 = 0;
        return (0xd6000000 | opc << 21 | op2 << 16 | op3 << 10 | xOrZr(rn) << 5 | op4);
    }

    // Workaround for Cortex-A53 erratum (835769). Emit an extra nop if the
    // last instruction in the buffer is a load, store or prefetch. Needed
    // before 64-bit multiply-accumulate instructions.
    template<int datasize>
    ALWAYS_INLINE void nopCortexA53Fix835769()
    {
#if CPU(ARM64_CORTEXA53)
        CHECK_DATASIZE();
        if (datasize == 64) {
            if (LIKELY(m_buffer.codeSize() >= sizeof(int32_t))) {
                // From ARMv8 Reference Manual, Section C4.1: the encoding of the
                // instructions in the Loads and stores instruction group is:
                // ---- 1-0- ---- ---- ---- ---- ---- ----
                if (UNLIKELY((*reinterpret_cast_ptr<int32_t*>(reinterpret_cast_ptr<char*>(m_buffer.data()) + m_buffer.codeSize() - sizeof(int32_t)) & 0x0a000000) == 0x08000000))
                    nop();
            }
        }
#endif
    }

    // Workaround for Cortex-A53 erratum (843419). Emit extra nops to avoid
    // wrong address access after ADRP instruction.
    ALWAYS_INLINE void nopCortexA53Fix843419()
    {
#if CPU(ARM64_CORTEXA53)
        nop();
        nop();
        nop();
#endif
    }

    AssemblerBuffer m_buffer;
    Vector<LinkRecord, 0, UnsafeVectorOverflow> m_jumpsToLink;
    int m_indexOfLastWatchpoint;
    int m_indexOfTailOfLastWatchpoint;
};

} // namespace JSC

#undef CHECK_DATASIZE_OF
#undef DATASIZE_OF
#undef MEMOPSIZE_OF
#undef CHECK_DATASIZE
#undef DATASIZE
#undef MEMOPSIZE
#undef CHECK_FP_MEMOP_DATASIZE

#endif // ENABLE(ASSEMBLER) && CPU(ARM64)

#endif // ARM64Assembler_h