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
// SPDX-License-Identifier: MIT
// Copyright (C) 2018-present iced project and contributors

pub(super) mod enums;
mod fmt_data;
mod fmt_tbl;
mod mem_size_tbl;
mod options;
mod pseudo_ops_fast;
mod regs;
#[cfg(test)]
mod tests;
mod trait_options;
mod trait_options_fast_fmt;

use crate::formatter::fast::enums::*;
use crate::formatter::fast::fmt_tbl::FMT_DATA;
use crate::formatter::fast::mem_size_tbl::MEM_SIZE_TBL;
pub use crate::formatter::fast::options::*;
use crate::formatter::fast::pseudo_ops_fast::get_pseudo_ops;
use crate::formatter::fast::regs::REGS_TBL;
pub use crate::formatter::fast::trait_options::*;
pub use crate::formatter::fast::trait_options_fast_fmt::*;
use crate::formatter::fmt_utils_all::*;
use crate::formatter::instruction_internal::{self, get_address_size_in_bytes};
use crate::formatter::*;
use crate::iced_constants::IcedConstants;
use crate::*;
use alloc::boxed::Box;
use alloc::vec::Vec;
use core::marker::PhantomData;
use core::{mem, ptr, slice};

// full fmt'd str = "prefixes mnemonic op0<decorators1>, op1, op2, op3, op4<decorators2>"
// prefixes = "es xacquire xrelease lock notrack repe repne "
// mnemonic = "prefetch_exclusive"
// op sep = ", "
// op = "fpustate108 ptr fs:[rax+zmm31*8+0x12345678]"
//		- longest 'xxxx ptr' and longest memory operand
// op = "0x123456789ABCDEF0"
// op = "0x1234:0x12345678"
// op = "zmm31"
// op = "offset symbol (123456789ABCDEF0)"
//		- symbol can have any length
// <decorators1> = "{k3}{z}"
// <decorators2> = "{rn-sae}"
// symbol = any length + optional address " (123456789ABCDEF0)"
// full = "es xacquire xrelease lock notrack repe repne prefetch_exclusive fpustate108 ptr fs:[rax+zmm31*8+0x12345678]{k3}{z}, fpustate108 ptr fs:[rax+zmm31*8+0x12345678], fpustate108 ptr fs:[rax+zmm31*8+0x12345678], fpustate108 ptr fs:[rax+zmm31*8+0x12345678], fpustate108 ptr fs:[rax+zmm31*8+0x12345678]{rn-sae}"
//		- it's not possible to have 5 `fpustate108 ptr fs:[rax+zmm31*8+0x12345678]` operands
//		  so we'll never get a formatted string this long if there's no symbol resolver.
#[allow(dead_code)]
const MAX_FMT_INSTR_LEN: usize = {
	const MAX_PREFIXES_LEN: usize = "es xacquire xrelease lock notrack repe repne ".len();
	const MAX_OPERAND_LEN: usize = "fpustate108 ptr fs:[rax+zmm31*8+0x12345678]".len();
	const MAX_DECORATOR1_LEN: usize = "{k3}{z}{eh}".len();
	const MAX_DECORATOR2_LEN: usize = "{rn-sae}{float16}".len();

	MAX_PREFIXES_LEN
	+ crate::formatter::strings_data::MAX_STRING_LEN
	+ MAX_DECORATOR1_LEN
	+ (IcedConstants::MAX_OP_COUNT * (2/*", "*/ + MAX_OPERAND_LEN)) - 1/*','*/
	+ MAX_DECORATOR2_LEN
};
const _: () = assert!(
	MAX_FMT_INSTR_LEN ==
	// Max mnemonic len
	crate::formatter::strings_data::MAX_STRING_LEN
		+ "es xacquire xrelease lock notrack repe repne  \
			fpustate108 ptr fs:[rax+zmm31*8+0x12345678]{k3}{z}{eh}, \
			fpustate108 ptr fs:[rax+zmm31*8+0x12345678], \
			fpustate108 ptr fs:[rax+zmm31*8+0x12345678], \
			fpustate108 ptr fs:[rax+zmm31*8+0x12345678], \
			fpustate108 ptr fs:[rax+zmm31*8+0x12345678]{rn-sae}{float16}"
			.len()
);
// Make sure it doesn't grow too much without us knowing about it (eg. if more operands are added)
const _: () = assert!(MAX_FMT_INSTR_LEN < 350);

// Creates a fast string type. It contains one ptr to the len (u8) + valid utf8 string.
// The utf8 string has enough bytes following it (eg. padding or the next fast str instance)
// so it's possible to read up to Self::SIZE bytes without crashing or causing a UB.
// Since the compiler knows that Self::SIZE is a constant, it can optimize the string copy,
// eg. if Self::SIZE == 8, it can read one unaligned u64 and write one unaligned u64.
macro_rules! mk_fast_str_ty {
	($ty_name:ident, $size:literal) => {
		#[repr(transparent)]
		#[derive(Copy, Clone)]
		struct $ty_name {
			// offset 0: u8, length in bytes of utf8 string
			// offset 1: [u8; SIZE] SIZE bytes can be read but only the first len() bytes are part of the string
			len_data: *const u8,
		}
		impl $ty_name {
			const SIZE: usize = $size;

			#[allow(dead_code)]
			fn new(len_data: *const u8) -> Self {
				debug_assert!(unsafe { *len_data as usize <= <$ty_name>::SIZE });
				Self { len_data }
			}

			fn len(self) -> usize {
				unsafe { *self.len_data as usize }
			}

			fn utf8_data(self) -> *const u8 {
				unsafe { self.len_data.add(1) }
			}

			#[allow(dead_code)]
			fn get_slice(self) -> &'static [u8] {
				unsafe { slice::from_raw_parts(self.utf8_data(), self.len()) }
			}
		}
		// SAFETY: The ptr field points to a static immutable u8 array.
		unsafe impl Send for $ty_name {}
		unsafe impl Sync for $ty_name {}
	};
}
// FastString2 isn't used since the code needs a 66h prefix (if target CPU is x86)
mk_fast_str_ty! {FastString4, 4} // ld 4
mk_fast_str_ty! {FastString8, 8} // ld 8
mk_fast_str_ty! {FastString12, 12} // ld 8 + ld 4
mk_fast_str_ty! {FastString16, 16} // ld 16
mk_fast_str_ty! {FastString20, 20} // ld 16 + ld 4

type FastStringMnemonic = FastString20;
type FastStringMemorySize = FastString16;
type FastStringRegister = FastString8;

// It doesn't seem to be possible to const-verify the arg (string literal) in a const fn so we create it with this macro
macro_rules! mk_const_fast_str {
	// $fast_ty = FastStringN where N is some integer
	// $str = padded string. First byte is the string len and the rest is the utf8 data
	//		  of $fast_ty::SIZE bytes padded with any bytes if needed
	($fast_ty:tt, $str:literal) => {{
		const STR: &str = $str;
		const _: () = assert!(STR.len() == 1 + <$fast_ty>::SIZE);
		const _: () = assert!(STR.as_bytes()[0] as usize <= <$fast_ty>::SIZE);
		$fast_ty { len_data: STR.as_ptr() }
	}};
}

macro_rules! verify_output_has_enough_bytes_left {
	($dst:ident, $dst_next_p:ident, $num_bytes:expr) => {
		// SAFETY: This is an opt out feature so if this returns `false`, they know what they're doing.
		if unsafe { TraitOptions::verify_output_has_enough_bytes_left() } {
			// Verify that there's enough bytes left. This should never fail (because we've called
			// `$dst.reserve(MAX_FMT_INSTR_LEN)`).
			iced_assert!($dst.capacity() - ($dst_next_p as usize - $dst.as_ptr() as usize) >= $num_bytes);
		}
	};
}

macro_rules! write_fast_str {
	// $dst = dest vector (from output.as_mut_vec())
	// $dst_next_p = next ptr to write in $dst
	// $source_ty = source fast string type
	// $source = source fast string instance, must be the same type as $source_ty (compiler will give an error if it's not the same type)
	($dst:ident, $dst_next_p:ident, $source_ty:ty, $source:ident) => {{
		const DATA_LEN: usize = <$source_ty>::SIZE;
		verify_output_has_enough_bytes_left!($dst, $dst_next_p, DATA_LEN);
		// SAFETY:
		// - $source is a valid utf8 string and it points to DATA_LEN readable bytes
		//   ($source is never from user code)
		// - $source is not in $dst ($source is static)
		// - $dst is writable with at least DATA_LEN bytes left (see assert above)
		// - $dst is at a valid utf8 char boundary (we're appending bytes)
		unsafe {
			ptr::copy_nonoverlapping(<$source_ty>::utf8_data($source), $dst_next_p, DATA_LEN);
		}
		debug_assert!(<$source_ty>::len($source) <= DATA_LEN);
		// SAFETY:
		// - $source.len() <= DATA_LEN so the new ptr is valid
		$dst_next_p = unsafe { $dst_next_p.add(<$source_ty>::len($source)) };
	}};
}

static HEX_GROUP2_UPPER: &str = "\
	000102030405060708090A0B0C0D0E0F\
	101112131415161718191A1B1C1D1E1F\
	202122232425262728292A2B2C2D2E2F\
	303132333435363738393A3B3C3D3E3F\
	404142434445464748494A4B4C4D4E4F\
	505152535455565758595A5B5C5D5E5F\
	606162636465666768696A6B6C6D6E6F\
	707172737475767778797A7B7C7D7E7F\
	808182838485868788898A8B8C8D8E8F\
	909192939495969798999A9B9C9D9E9F\
	A0A1A2A3A4A5A6A7A8A9AAABACADAEAF\
	B0B1B2B3B4B5B6B7B8B9BABBBCBDBEBF\
	C0C1C2C3C4C5C6C7C8C9CACBCCCDCECF\
	D0D1D2D3D4D5D6D7D8D9DADBDCDDDEDF\
	E0E1E2E3E4E5E6E7E8E9EAEBECEDEEEF\
	F0F1F2F3F4F5F6F7F8F9FAFBFCFDFEFF\
	__"; // Padding so we can read 4 bytes at every index 0-0xFF inclusive

macro_rules! write_fast_hex2_rw_4bytes {
	($dst:ident, $dst_next_p:ident, $value:ident, $lower_or_value:ident, $check_limit:literal) => {{
		const DATA_LEN: usize = 4;
		const REAL_LEN: usize = 2;
		if $check_limit {
			verify_output_has_enough_bytes_left!($dst, $dst_next_p, DATA_LEN);
		}
		// We'll read DATA_LEN (4) bytes so we must be able to access up to and including offset 0x201
		debug_assert_eq!(HEX_GROUP2_UPPER.len(), 0xFF * REAL_LEN + DATA_LEN);
		debug_assert!($value < 0x100);
		// $lower_or_value == 0 if we should use uppercase hex digits or 0x2020_2020 to use lowercase hex digits.
		// If LE, we need xxxx2020 and if BE, we need 2020xxxx.
		debug_assert!($lower_or_value == 0 || $lower_or_value == 0x2020_2020);
		// SAFETY:
		// - HEX_GROUP2_UPPER is a valid utf8 string and every valid 2-digit hex number
		//	 0-0xFF can be used as an index * REAL_LEN (2) to read DATA_LEN (4) bytes.
		// - $dst is writable with at least DATA_LEN bytes left (see assert above)
		// - $dst is at a valid utf8 char boundary (we're appending bytes)
		#[allow(trivial_numeric_casts)]
		unsafe {
			let src_ptr = HEX_GROUP2_UPPER.as_ptr().add(($value as usize) * REAL_LEN) as *const u32;
			ptr::write_unaligned($dst_next_p as *mut u32, ptr::read_unaligned(src_ptr) | $lower_or_value);
		}
		const _: () = assert!(REAL_LEN <= DATA_LEN);
		// SAFETY:
		// - REAL_LEN <= DATA_LEN so the new ptr is valid since there's at least DATA_LEN bytes available in $dst
		$dst_next_p = unsafe { $dst_next_p.add(REAL_LEN) };
	}};
}

macro_rules! write_fast_ascii_char {
	// $dst = dest vector (from output.as_mut_vec())
	// $dst_next_p = next ptr to write in $dst
	// $ch = char to write (must be ASCII)
	($dst:ident, $dst_next_p:ident, $ch:expr, $check_limit:literal) => {{
		const DATA_LEN: usize = 1;
		if $check_limit {
			verify_output_has_enough_bytes_left!($dst, $dst_next_p, DATA_LEN);
		}
		#[allow(trivial_numeric_casts)]
		{
			debug_assert!($ch as u32 <= 0x7F);
		}
		// SAFETY:
		// - $ch is ASCII (valid 1-byte utf8 char)
		// - $dst is writable with at least DATA_LEN bytes left (see assert above)
		// - $dst is at a valid utf8 char boundary (we're appending bytes)
		#[allow(trivial_numeric_casts)]
		unsafe {
			*$dst_next_p = $ch as u8;
		}
		// SAFETY: There's at least one byte left so the new ptr is valid
		$dst_next_p = unsafe { $dst_next_p.add(1) };
	}};
}

macro_rules! write_fast_ascii_char_lit {
	// $dst = dest vector (from output.as_mut_vec())
	// $dst_next_p = next ptr to write in $dst
	// $ch = char to write (must be ASCII)
	($dst:ident, $dst_next_p:ident, $ch:tt, $check_limit:literal) => {{
		const _: () = assert!($ch as u32 <= 0x7F);
		write_fast_ascii_char!($dst, $dst_next_p, $ch, $check_limit);
	}};
}

macro_rules! update_vec_len {
	// $dst = dest vector (from output.as_mut_vec())
	// $dst_next_p = next ptr to write in $dst
	($dst:ident, $dst_next_p:ident) => {
		// SAFETY:
		// - we only write valid utf8 strings and ASCII chars to vec
		// - We've written all chars up to but not including $dst_next_p so all visible data have been initialized
		// - $dst_next_p points to a valid location inside the vec or at most 1 byte past the last valid byte
		unsafe {
			$dst.set_len($dst_next_p as usize - $dst.as_ptr() as usize);
		}
	};
}

macro_rules! use_dst_only_now {
	// $dst = dest vector (from output.as_mut_vec())
	// $dst_next_p = next ptr to write in $dst
	($dst:ident, $dst_next_p:ident) => {
		update_vec_len!($dst, $dst_next_p);
		// Make sure we don't use it accidentally
		#[allow(unused_variables)]
		let $dst_next_p: () = ();
	};
}
macro_rules! use_dst_next_p_now {
	// $dst = dest vector (from output.as_mut_vec())
	// $dst_next_p = next ptr to write in $dst
	($dst:ident, $dst_next_p:ident) => {
		// Need to make sure we have enough bytes available again because we could've
		// written a very long symbol name.
		$dst.reserve(MAX_FMT_INSTR_LEN);
		// Restore variable
		let mut $dst_next_p = unsafe { $dst.as_mut_ptr().add($dst.len()) };
	};
}

// Macros to safely call the methods (make sure the return value is stored back in dst_next_p)
macro_rules! call_format_register {
	($slf:ident, $dst:ident, $dst_next_p:ident, $reg:expr) => {{
		$dst_next_p = $slf.format_register($dst, $dst_next_p, $reg);
	}};
}
macro_rules! call_format_number {
	($slf:ident, $dst:ident, $dst_next_p:ident, $imm:expr) => {{
		$dst_next_p = $slf.format_number($dst, $dst_next_p, $imm);
	}};
}
macro_rules! call_write_symbol {
	($slf:ident, $dst:ident, $dst_next_p:ident, $imm:expr, $sym:expr) => {{
		$dst_next_p = $slf.write_symbol($dst, $dst_next_p, $imm, $sym);
	}};
}
macro_rules! call_write_symbol2 {
	($slf:ident, $dst:ident, $dst_next_p:ident, $imm:expr, $sym:expr, $write_minus_if_signed:literal) => {{
		$dst_next_p = $slf.write_symbol2($dst, $dst_next_p, $imm, $sym, $write_minus_if_signed);
	}};
}
macro_rules! format_memory_else_block {
	($slf:ident, $dst:ident, $dst_next_p:ident, $need_plus:ident, $displ_size:ident, $displ:ident, $addr_size:ident) => {
		if !$need_plus || ($displ_size != 0 && $displ != 0) {
			if $need_plus {
				let c = if $addr_size == 8 {
					if $displ < 0 {
						$displ = $displ.wrapping_neg();
						'-'
					} else {
						'+'
					}
				} else if $addr_size == 4 {
					if ($displ as i32) < 0 {
						$displ = ($displ as i32).wrapping_neg() as u32 as i64;
						'-'
					} else {
						'+'
					}
				} else {
					debug_assert_eq!($addr_size, 2);
					if ($displ as i16) < 0 {
						$displ = ($displ as i16).wrapping_neg() as u16 as i64;
						'-'
					} else {
						'+'
					}
				};
				write_fast_ascii_char!($dst, $dst_next_p, c, true);
			}
			call_format_number!($slf, $dst, $dst_next_p, $displ as u64);
		}
	};
}
// Only one caller has variable args starting from $seg_reg so this is a macro. The compiler is able
// to remove lots of code in all the other cases with literal macro args.
macro_rules! format_memory_code {
	($slf:ident, $dst:ident, $dst_next_p:ident, $instruction:ident, $operand:expr, $seg_reg:expr, $base_reg:expr, $index_reg:expr,
		$scale:expr, $displ_size:expr, $displ:expr, $addr_size:expr) => {
		#[allow(trivial_numeric_casts)]
		{
			let mut base_reg = $base_reg;
			let mut displ_size: u32 = $displ_size;
			let mut displ: i64 = $displ;
			debug_assert!(get_address_size_in_bytes(base_reg, $index_reg, displ_size, $instruction.code_size()) == $addr_size);

			let abs_addr;
			if base_reg == Register::RIP {
				abs_addr = displ as u64;
				if TraitOptions::rip_relative_addresses(&$slf.d.options) {
					displ = displ.wrapping_sub($instruction.next_ip() as i64);
				} else {
					debug_assert_eq!($index_reg, Register::None);
					base_reg = Register::None;
				}
				displ_size = 8;
			} else if base_reg == Register::EIP {
				abs_addr = displ as u32 as u64;
				if TraitOptions::rip_relative_addresses(&$slf.d.options) {
					displ = (displ as u32).wrapping_sub($instruction.next_ip32()) as i32 as i64;
				} else {
					debug_assert_eq!($index_reg, Register::None);
					base_reg = Register::None;
				}
				displ_size = 4;
			} else {
				abs_addr = displ as u64;
			}

			let show_mem_size = TraitOptions::always_show_memory_size(&$slf.d.options) || {
				let flags = $slf.d.code_flags[$instruction.code() as usize];
				(flags & (FastFmtFlags::FORCE_MEM_SIZE as u8)) != 0 || $instruction.is_broadcast()
			};
			if show_mem_size {
				let keywords = $slf.d.all_memory_sizes[$instruction.memory_size() as usize];
				write_fast_str!($dst, $dst_next_p, FastStringMemorySize, keywords);
			}

			let seg_override;
			if TraitOptions::always_show_segment_register(&$slf.d.options)
				|| ({
					seg_override = $instruction.segment_prefix();
					seg_override != Register::None
				} && !{
					let notrack_prefix = seg_override == Register::DS && is_notrack_prefix_branch($instruction.code()) && {
						let code_size = $instruction.code_size();
						!((code_size == CodeSize::Code16 || code_size == CodeSize::Code32)
							&& (base_reg == Register::BP || base_reg == Register::EBP || base_reg == Register::ESP))
					};
					notrack_prefix
				} && (SpecializedFormatter::<TraitOptions>::SHOW_USELESS_PREFIXES
					|| show_segment_prefix_bool(Register::None, $instruction, SpecializedFormatter::<TraitOptions>::SHOW_USELESS_PREFIXES)))
			{
				call_format_register!($slf, $dst, $dst_next_p, $seg_reg);
				write_fast_ascii_char_lit!($dst, $dst_next_p, ':', true);
			}
			write_fast_ascii_char_lit!($dst, $dst_next_p, '[', true);

			let mut need_plus = if base_reg != Register::None {
				call_format_register!($slf, $dst, $dst_next_p, base_reg);
				true
			} else {
				false
			};

			if $index_reg != Register::None {
				if need_plus {
					write_fast_ascii_char_lit!($dst, $dst_next_p, '+', true);
				}
				need_plus = true;

				call_format_register!($slf, $dst, $dst_next_p, $index_reg);

				// [rsi] = base reg, [rsi*1] = index reg
				if $addr_size != 2 && ($scale != 0 || base_reg == Register::None) {
					let scale_str = SCALE_NUMBERS[$scale as usize];
					write_fast_str!($dst, $dst_next_p, FastString4, scale_str);
				}
			}

			if TraitOptions::ENABLE_SYMBOL_RESOLVER {
				// See OpKind::NearBranch16 in format() for why we clone the symbols
				let mut vec: Vec<SymResTextPart<'_>> = Vec::new();
				if let Some(ref symbol) = if let Some(ref mut symbol_resolver) = $slf.symbol_resolver {
					to_owned(symbol_resolver.symbol($instruction, $operand, Some($operand), abs_addr, $addr_size), &mut vec)
				} else {
					None
				} {
					if need_plus {
						let c = if (symbol.flags & SymbolFlags::SIGNED) != 0 { '-' } else { '+' };
						write_fast_ascii_char!($dst, $dst_next_p, c, true);
					} else if (symbol.flags & SymbolFlags::SIGNED) != 0 {
						write_fast_ascii_char_lit!($dst, $dst_next_p, '-', true);
					}

					call_write_symbol2!($slf, $dst, $dst_next_p, abs_addr, symbol, false);
				} else {
					let addr_size = $addr_size;
					format_memory_else_block!($slf, $dst, $dst_next_p, need_plus, displ_size, displ, addr_size);
				}
			} else {
				let addr_size = $addr_size;
				format_memory_else_block!($slf, $dst, $dst_next_p, need_plus, displ_size, displ, addr_size);
			}

			write_fast_ascii_char_lit!($dst, $dst_next_p, ']', true);
		}
	};
}
macro_rules! call_format_memory {
	($slf:ident, $dst:ident, $dst_next_p:ident, $instruction:ident, $operand:ident, $seg_reg:expr, $base_reg:tt,
		$index_reg:tt, $scale:tt, $displ_size:tt, $displ:tt, $addr_size:tt $(,)?) => {
		// This speeds up SpecializedFormatter but slows down FastFormatter so detect which
		// formatter it is. Both paths are tested (same tests).
		// This is fugly but the whole point of this formatter is to be fast which can result in ugly code.
		{
			if TraitOptions::__IS_FAST_FORMATTER {
				// Less code: call a method
				$dst_next_p = $slf.format_memory(
					$dst,
					$dst_next_p,
					$instruction,
					$operand,
					$seg_reg,
					$base_reg,
					$index_reg,
					$scale,
					$displ_size,
					$displ,
					$addr_size,
				)
			} else {
				// The options are all most likely hard coded so inline and specialize the 'method call'
				format_memory_code!(
					$slf,
					$dst,
					$dst_next_p,
					$instruction,
					$operand,
					$seg_reg,
					$base_reg,
					$index_reg,
					$scale,
					$displ_size,
					$displ,
					$addr_size
				)
			}
		}
	};
}

static SCALE_NUMBERS: [FastString4; 4] = [
	mk_const_fast_str!(FastString4, "\x02*1  "),
	mk_const_fast_str!(FastString4, "\x02*2  "),
	mk_const_fast_str!(FastString4, "\x02*4  "),
	mk_const_fast_str!(FastString4, "\x02*8  "),
];
const _: () = assert!(RoundingControl::None as u32 == 0);
const _: () = assert!(RoundingControl::RoundToNearest as u32 == 1);
const _: () = assert!(RoundingControl::RoundDown as u32 == 2);
const _: () = assert!(RoundingControl::RoundUp as u32 == 3);
const _: () = assert!(RoundingControl::RoundTowardZero as u32 == 4);
static RC_SAE_STRINGS: [FastString8; IcedConstants::ROUNDING_CONTROL_ENUM_COUNT] = [
	mk_const_fast_str!(FastString8, "\x00        "),
	mk_const_fast_str!(FastString8, "\x08{rn-sae}"),
	mk_const_fast_str!(FastString8, "\x08{rd-sae}"),
	mk_const_fast_str!(FastString8, "\x08{ru-sae}"),
	mk_const_fast_str!(FastString8, "\x08{rz-sae}"),
];
static RC_STRINGS: [FastString4; IcedConstants::ROUNDING_CONTROL_ENUM_COUNT] = [
	mk_const_fast_str!(FastString4, "\x00    "),
	mk_const_fast_str!(FastString4, "\x04{rn}"),
	mk_const_fast_str!(FastString4, "\x04{rd}"),
	mk_const_fast_str!(FastString4, "\x04{ru}"),
	mk_const_fast_str!(FastString4, "\x04{rz}"),
];
#[cfg(feature = "mvex")]
static MVEX_REG_MEM_CONSTS_32: [FastString12; IcedConstants::MVEX_REG_MEM_CONV_ENUM_COUNT] = [
	mk_const_fast_str!(FastString12, "\x00            "),
	mk_const_fast_str!(FastString12, "\x00            "),
	mk_const_fast_str!(FastString12, "\x06{cdab}      "),
	mk_const_fast_str!(FastString12, "\x06{badc}      "),
	mk_const_fast_str!(FastString12, "\x06{dacb}      "),
	mk_const_fast_str!(FastString12, "\x06{aaaa}      "),
	mk_const_fast_str!(FastString12, "\x06{bbbb}      "),
	mk_const_fast_str!(FastString12, "\x06{cccc}      "),
	mk_const_fast_str!(FastString12, "\x06{dddd}      "),
	mk_const_fast_str!(FastString12, "\x00            "),
	mk_const_fast_str!(FastString12, "\x07{1to16}     "),
	mk_const_fast_str!(FastString12, "\x07{4to16}     "),
	mk_const_fast_str!(FastString12, "\x09{float16}   "),
	mk_const_fast_str!(FastString12, "\x07{uint8}     "),
	mk_const_fast_str!(FastString12, "\x07{sint8}     "),
	mk_const_fast_str!(FastString12, "\x08{uint16}    "),
	mk_const_fast_str!(FastString12, "\x08{sint16}    "),
];
#[cfg(feature = "mvex")]
static MVEX_REG_MEM_CONSTS_64: [FastString12; IcedConstants::MVEX_REG_MEM_CONV_ENUM_COUNT] = [
	mk_const_fast_str!(FastString12, "\x00            "),
	mk_const_fast_str!(FastString12, "\x00            "),
	mk_const_fast_str!(FastString12, "\x06{cdab}      "),
	mk_const_fast_str!(FastString12, "\x06{badc}      "),
	mk_const_fast_str!(FastString12, "\x06{dacb}      "),
	mk_const_fast_str!(FastString12, "\x06{aaaa}      "),
	mk_const_fast_str!(FastString12, "\x06{bbbb}      "),
	mk_const_fast_str!(FastString12, "\x06{cccc}      "),
	mk_const_fast_str!(FastString12, "\x06{dddd}      "),
	mk_const_fast_str!(FastString12, "\x00            "),
	mk_const_fast_str!(FastString12, "\x06{1to8}      "),
	mk_const_fast_str!(FastString12, "\x06{4to8}      "),
	mk_const_fast_str!(FastString12, "\x09{float16}   "),
	mk_const_fast_str!(FastString12, "\x07{uint8}     "),
	mk_const_fast_str!(FastString12, "\x07{sint8}     "),
	mk_const_fast_str!(FastString12, "\x08{uint16}    "),
	mk_const_fast_str!(FastString12, "\x08{sint16}    "),
];

struct FmtTableData {
	mnemonics: Box<[FastStringMnemonic; IcedConstants::CODE_ENUM_COUNT]>,
	flags: Box<[u8; IcedConstants::CODE_ENUM_COUNT]>, // FastFmtFlags
}

/// Fast specialized formatter with less formatting options and with a masm-like syntax.
/// Use it if formatting speed is more important than being able to re-assemble formatted instructions.
///
/// The `TraitOptions` generic parameter is a [`SpecializedFormatterTraitOptions`] trait. It can
/// be used to hard code options so the compiler can create a smaller and faster formatter.
/// See also [`FastFormatter`] which allows changing the options at runtime at the cost of
/// being a little bit slower and using a little bit more code.
///
/// This formatter is ~3.3x faster than the gas/intel/masm/nasm formatters (the time includes decoding + formatting).
///
/// [`SpecializedFormatterTraitOptions`]: trait.SpecializedFormatterTraitOptions.html
/// [`FastFormatter`]: type.FastFormatter.html
///
/// # Examples
///
/// ```
/// use iced_x86::*;
///
/// let bytes = b"\x62\xF2\x4F\xDD\x72\x50\x01";
/// let mut decoder = Decoder::new(64, bytes, DecoderOptions::NONE);
/// let instr = decoder.decode();
///
/// // If you like the default options, you can also use DefaultSpecializedFormatterTraitOptions
/// // instead of impl the options trait.
/// struct MyTraitOptions;
/// impl SpecializedFormatterTraitOptions for MyTraitOptions {
///     fn space_after_operand_separator(_options: &FastFormatterOptions) -> bool {
///         // We hard code the value to `true` which means it's not possible to
///         // change this option at runtime, i.e., this will do nothing:
///         //      formatter.options_mut().set_space_after_operand_separator(false);
///         true
///     }
///     fn rip_relative_addresses(options: &FastFormatterOptions) -> bool {
///         // Since we return the input, we can change this value at runtime, i.e.,
///         // this works:
///         //      formatter.options_mut().set_rip_relative_addresses(false);
///         options.rip_relative_addresses()
///     }
/// }
/// type MyFormatter = SpecializedFormatter<MyTraitOptions>;
///
/// let mut output = String::new();
/// let mut formatter = MyFormatter::new();
/// formatter.format(&instr, &mut output);
/// assert_eq!(output, "vcvtne2ps2bf16 zmm2{k5}{z}, zmm6, dword bcst [rax+0x4]");
/// ```
///
/// # Fastest possible disassembly
///
/// For fastest possible disassembly, you should *not* enable the `db` feature (or you should set [`ENABLE_DB_DW_DD_DQ`] to `false`)
/// and you should also override the unsafe [`verify_output_has_enough_bytes_left()`] and return `false`.
///
/// [`ENABLE_DB_DW_DD_DQ`]: trait.SpecializedFormatterTraitOptions.html#associatedconstant.ENABLE_DB_DW_DD_DQ
/// [`verify_output_has_enough_bytes_left()`]: trait.SpecializedFormatterTraitOptions.html#method.verify_output_has_enough_bytes_left
///
/// ```
/// use iced_x86::*;
///
/// struct MyTraitOptions;
/// impl SpecializedFormatterTraitOptions for MyTraitOptions {
///     // If you never create a db/dw/dd/dq 'instruction', we don't need this feature.
///     const ENABLE_DB_DW_DD_DQ: bool = false;
///     // For a few percent faster code, you can also override `verify_output_has_enough_bytes_left()` and return `false`
///     // unsafe fn verify_output_has_enough_bytes_left() -> bool {
///     //     false
///     // }
/// }
/// type MyFormatter = SpecializedFormatter<MyTraitOptions>;
///
/// // Assume this is a big slice and not just one instruction
/// let bytes = b"\x62\xF2\x4F\xDD\x72\x50\x01";
/// let mut decoder = Decoder::new(64, bytes, DecoderOptions::NONE);
///
/// let mut output = String::new();
/// let mut instruction = Instruction::default();
/// let mut formatter = MyFormatter::new();
/// while decoder.can_decode() {
///     decoder.decode_out(&mut instruction);
///     output.clear();
///     formatter.format(&instruction, &mut output);
///     // do something with 'output' here, eg.:
///     //     println!("{}", output);
/// }
/// ```
///
/// Also add this to your `Cargo.toml` file:
///
/// ```toml
/// [profile.release]
/// codegen-units = 1
/// lto = true
/// opt-level = 3
/// ```
///
/// # Using a symbol resolver
///
/// The symbol resolver is disabled by default, but it's easy to enable it (or you can just use [`FastFormatter`])
///
/// ```
/// use iced_x86::*;
/// use std::collections::HashMap;
///
/// let bytes = b"\x48\x8B\x8A\xA5\x5A\xA5\x5A";
/// let mut decoder = Decoder::new(64, bytes, DecoderOptions::NONE);
/// let instr = decoder.decode();
///
/// struct MyTraitOptions;
/// impl SpecializedFormatterTraitOptions for MyTraitOptions {
///     const ENABLE_SYMBOL_RESOLVER: bool = true;
/// }
/// type MyFormatter = SpecializedFormatter<MyTraitOptions>;
///
/// struct MySymbolResolver { map: HashMap<u64, String> }
/// impl SymbolResolver for MySymbolResolver {
///     fn symbol(&mut self, _instruction: &Instruction, _operand: u32, _instruction_operand: Option<u32>,
///          address: u64, _address_size: u32) -> Option<SymbolResult> {
///         if let Some(symbol_string) = self.map.get(&address) {
///             // The 'address' arg is the address of the symbol and doesn't have to be identical
///             // to the 'address' arg passed to symbol(). If it's different from the input
///             // address, the formatter will add +N or -N, eg. '[rax+symbol+123]'
///             Some(SymbolResult::with_str(address, symbol_string.as_str()))
///         } else {
///             None
///         }
///     }
/// }
///
/// // Hard code the symbols, it's just an example!😄
/// let mut sym_map: HashMap<u64, String> = HashMap::new();
/// sym_map.insert(0x5AA55AA5, String::from("my_data"));
///
/// let mut output = String::new();
/// let resolver = Box::new(MySymbolResolver { map: sym_map });
/// let mut formatter = MyFormatter::try_with_options(Some(resolver)).unwrap();
/// formatter.format(&instr, &mut output);
/// assert_eq!("mov rcx,[rdx+my_data]", output);
/// ```
#[allow(missing_debug_implementations)]
pub struct SpecializedFormatter<TraitOptions: SpecializedFormatterTraitOptions> {
	d: SelfData,
	symbol_resolver: Option<Box<dyn SymbolResolver>>,
	_required_by_rustc: PhantomData<fn() -> TraitOptions>,
}

impl<TraitOptions: SpecializedFormatterTraitOptions> Default for SpecializedFormatter<TraitOptions> {
	#[must_use]
	#[inline]
	fn default() -> Self {
		SpecializedFormatter::<TraitOptions>::new()
	}
}

// Read-only data which is needed a couple of times due to borrow checker
struct SelfData {
	options: FastFormatterOptions,
	all_registers: &'static [FastStringRegister; IcedConstants::REGISTER_ENUM_COUNT],
	code_mnemonics: &'static [FastStringMnemonic; IcedConstants::CODE_ENUM_COUNT],
	code_flags: &'static [u8; IcedConstants::CODE_ENUM_COUNT],
	all_memory_sizes: &'static [FastStringMemorySize; IcedConstants::MEMORY_SIZE_ENUM_COUNT],
}

impl<TraitOptions: SpecializedFormatterTraitOptions> SpecializedFormatter<TraitOptions> {
	const SHOW_USELESS_PREFIXES: bool = true;

	/// Creates a new instance of this formatter
	#[must_use]
	#[inline]
	#[allow(clippy::unwrap_used)]
	pub fn new() -> Self {
		// This never panics
		SpecializedFormatter::<TraitOptions>::try_with_options(None).unwrap()
	}

	/// Creates a new instance of this formatter
	///
	/// # Errors
	///
	/// Fails if [`TraitOptions::ENABLE_SYMBOL_RESOLVER`] is `false` and `symbol_resolver.is_some()`
	///
	/// [`TraitOptions::ENABLE_SYMBOL_RESOLVER`]: trait.SpecializedFormatterTraitOptions.html#associatedconstant.ENABLE_SYMBOL_RESOLVER
	///
	/// # Arguments
	///
	/// - `symbol_resolver`: Symbol resolver or `None`
	#[allow(clippy::missing_inline_in_public_items)]
	pub fn try_with_options(symbol_resolver: Option<Box<dyn SymbolResolver>>) -> Result<Self, IcedError> {
		if !TraitOptions::ENABLE_SYMBOL_RESOLVER && symbol_resolver.is_some() {
			Err(IcedError::new(concat!(stringify!(TraitOptions::ENABLE_SYMBOL_RESOLVER), " is disabled so symbol resolvers aren't supported")))
		} else {
			Ok(Self {
				d: SelfData {
					options: FastFormatterOptions::new(),
					all_registers: &REGS_TBL,
					code_mnemonics: &FMT_DATA.mnemonics,
					code_flags: &FMT_DATA.flags,
					all_memory_sizes: &MEM_SIZE_TBL,
				},
				symbol_resolver,
				_required_by_rustc: PhantomData,
			})
		}
	}

	/// Gets the formatter options (immutable)
	///
	/// Note that the `TraitOptions` generic parameter can override any option and hard code them,
	/// see [`SpecializedFormatterTraitOptions`]
	///
	/// [`SpecializedFormatterTraitOptions`]: trait.SpecializedFormatterTraitOptions.html
	#[must_use]
	#[inline]
	pub fn options(&self) -> &FastFormatterOptions {
		&self.d.options
	}

	/// Gets the formatter options (mutable)
	///
	/// Note that the `TraitOptions` generic parameter can override any option and hard code them,
	/// see [`SpecializedFormatterTraitOptions`]
	///
	/// [`SpecializedFormatterTraitOptions`]: trait.SpecializedFormatterTraitOptions.html
	#[must_use]
	#[inline]
	pub fn options_mut(&mut self) -> &mut FastFormatterOptions {
		&mut self.d.options
	}

	/// Formats the whole instruction: prefixes, mnemonic, operands
	///
	/// # Arguments
	///
	/// - `instruction`: Instruction
	/// - `output`: Output
	#[allow(clippy::missing_inline_in_public_items)]
	#[allow(clippy::let_unit_value)]
	#[allow(clippy::useless_let_if_seq)]
	pub fn format(&mut self, instruction: &Instruction, output: &mut String) {
		// SAFETY: We only append data that come from a `&str`, a `String` or ASCII chars so the data is always valid utf8
		let dst = unsafe { output.as_mut_vec() };
		// The code assumes there's enough bytes (or it will panic) so reserve enough bytes here
		dst.reserve(MAX_FMT_INSTR_LEN);
		// SAFETY:
		// - ptr is in bounds (after last valid byte)
		// - it's reloaded when using 'dst' to write to the vector
		let mut dst_next_p = unsafe { dst.as_mut_ptr().add(dst.len()) };

		let code = instruction.code();
		let mut mnemonic = self.d.code_mnemonics[code as usize];
		let mut op_count = instruction.op_count();
		if TraitOptions::use_pseudo_ops(&self.d.options) {
			let flags = self.d.code_flags[code as usize];
			let pseudo_ops_num = flags >> FastFmtFlags::PSEUDO_OPS_KIND_SHIFT;
			if pseudo_ops_num != 0 && instruction.op_kind(op_count - 1) == OpKind::Immediate8 {
				let mut index = instruction.immediate8() as usize;
				// SAFETY: the generator generates only valid values (1-based)
				let pseudo_ops_kind: PseudoOpsKind = unsafe { mem::transmute(pseudo_ops_num - 1) };
				// Not enough bits to store all values so some are mapped to the same value. Fix that here.
				let pseudo_ops_kind = if pseudo_ops_kind == PseudoOpsKind::vpcmpd6 {
					match code {
						#[cfg(feature = "mvex")]
						Code::MVEX_Vpcmpud_kr_k1_zmm_zmmmt_imm8 => PseudoOpsKind::vpcmpud6,
						_ => pseudo_ops_kind,
					}
				} else {
					pseudo_ops_kind
				};
				let pseudo_ops = get_pseudo_ops(pseudo_ops_kind);
				if pseudo_ops_kind == PseudoOpsKind::pclmulqdq || pseudo_ops_kind == PseudoOpsKind::vpclmulqdq {
					if index <= 1 {
						// nothing
					} else if index == 0x10 {
						index = 2;
					} else if index == 0x11 {
						index = 3;
					} else {
						index = usize::MAX;
					}
				}
				if let Some(&pseudo_op_mnemonic) = pseudo_ops.get(index) {
					mnemonic = pseudo_op_mnemonic;
					op_count -= 1;
				}
			}
		}

		let prefix_seg = instruction_internal::internal_segment_prefix_raw(instruction);
		const _: () = assert!(Register::None as u32 == 0);
		if prefix_seg < 6 || instruction_internal::internal_has_any_of_lock_rep_repne_prefix(instruction) != 0 {
			const DS_REG: u32 = Register::DS as u32 - Register::ES as u32;
			let has_notrack_prefix = prefix_seg == DS_REG && is_notrack_prefix_branch(code);
			if !has_notrack_prefix && prefix_seg < 6 && SpecializedFormatter::<TraitOptions>::show_segment_prefix(instruction, op_count) {
				let prefix_seg = unsafe { mem::transmute((Register::ES as u32 + prefix_seg) as RegisterUnderlyingType) };
				call_format_register!(self, dst, dst_next_p, prefix_seg);
				write_fast_ascii_char_lit!(dst, dst_next_p, ' ', true);
			}

			let mut has_xacquire_xrelease = false;
			if instruction.has_xacquire_prefix() {
				const FAST_STR: FastString12 = mk_const_fast_str!(FastString12, "\x09xacquire    ");
				write_fast_str!(dst, dst_next_p, FastString12, FAST_STR);
				has_xacquire_xrelease = true;
			}
			if instruction.has_xrelease_prefix() {
				const FAST_STR: FastString12 = mk_const_fast_str!(FastString12, "\x09xrelease    ");
				write_fast_str!(dst, dst_next_p, FastString12, FAST_STR);
				has_xacquire_xrelease = true;
			}
			if instruction.has_lock_prefix() {
				const FAST_STR: FastString8 = mk_const_fast_str!(FastString8, "\x05lock    ");
				write_fast_str!(dst, dst_next_p, FastString8, FAST_STR);
			}
			if has_notrack_prefix {
				const FAST_STR: FastString8 = mk_const_fast_str!(FastString8, "\x08notrack ");
				write_fast_str!(dst, dst_next_p, FastString8, FAST_STR);
			}
			if !has_xacquire_xrelease {
				if instruction.has_repe_prefix()
					&& (SpecializedFormatter::<TraitOptions>::SHOW_USELESS_PREFIXES
						|| show_rep_or_repe_prefix_bool(code, SpecializedFormatter::<TraitOptions>::SHOW_USELESS_PREFIXES))
				{
					if is_repe_or_repne_instruction(code) {
						const FAST_STR: FastString8 = mk_const_fast_str!(FastString8, "\x05repe    ");
						write_fast_str!(dst, dst_next_p, FastString8, FAST_STR);
					} else {
						const FAST_STR: FastString4 = mk_const_fast_str!(FastString4, "\x04rep ");
						write_fast_str!(dst, dst_next_p, FastString4, FAST_STR);
					}
				}
				if instruction.has_repne_prefix() {
					if (Code::Retnw_imm16 <= code && code <= Code::Retnq)
						|| (Code::Call_rel16 <= code && code <= Code::Jmp_rel32_64)
						|| (Code::Call_rm16 <= code && code <= Code::Call_rm64)
						|| (Code::Jmp_rm16 <= code && code <= Code::Jmp_rm64)
						|| code.is_jcc_short_or_near()
					{
						const FAST_STR: FastString4 = mk_const_fast_str!(FastString4, "\x04bnd ");
						write_fast_str!(dst, dst_next_p, FastString4, FAST_STR);
					} else if SpecializedFormatter::<TraitOptions>::SHOW_USELESS_PREFIXES
						|| show_repne_prefix_bool(code, SpecializedFormatter::<TraitOptions>::SHOW_USELESS_PREFIXES)
					{
						const FAST_STR: FastString8 = mk_const_fast_str!(FastString8, "\x06repne   ");
						write_fast_str!(dst, dst_next_p, FastString8, FAST_STR);
					}
				}
			}
		}

		write_fast_str!(dst, dst_next_p, FastStringMnemonic, mnemonic);

		let is_declare_data;
		let declare_data_kind = if !TraitOptions::ENABLE_DB_DW_DD_DQ {
			is_declare_data = false;
			OpKind::Register
		} else if (code as u32).wrapping_sub(Code::DeclareByte as u32) <= (Code::DeclareQword as u32 - Code::DeclareByte as u32) {
			op_count = instruction.declare_data_len() as u32;
			is_declare_data = true;
			match code {
				Code::DeclareByte => OpKind::Immediate8,
				Code::DeclareWord => OpKind::Immediate16,
				Code::DeclareDword => OpKind::Immediate32,
				_ => {
					debug_assert_eq!(code, Code::DeclareQword);
					OpKind::Immediate64
				}
			}
		} else {
			is_declare_data = false;
			OpKind::Register
		};

		if op_count > 0 {
			write_fast_ascii_char_lit!(dst, dst_next_p, ' ', true);

			#[cfg(feature = "mvex")]
			let mvex_rm_operand = {
				if IcedConstants::is_mvex(instruction.code()) {
					debug_assert_ne!(op_count, 0);
					if instruction.op_kind(op_count.wrapping_sub(1)) == OpKind::Immediate8 {
						op_count.wrapping_sub(2)
					} else {
						op_count.wrapping_sub(1)
					}
				} else {
					u32::MAX
				}
			};

			let mut operand = 0;
			loop {
				let imm8;
				let imm16;
				let imm32;
				let imm64;
				let imm_size;
				let op_kind = if TraitOptions::ENABLE_DB_DW_DD_DQ && is_declare_data { declare_data_kind } else { instruction.op_kind(operand) };

				// Share as much code as possible so put these in macros
				macro_rules! fmt_near_branch {
					($slf:ident, $dst:ident, $dst_next_p:ident, $instruction:ident, $imm_size:expr, $imm:ident) => {{
						if TraitOptions::ENABLE_SYMBOL_RESOLVER {
							// PERF: Symbols should be rare when using fast fmt with a symbol resolver so clone
							// the symbol (forced by borrowck).
							// This results in slightly faster code when we do NOT support a symbol resolver since
							// we don't need to pass in the options to various methods and can instead pass in &Self
							// (i.e., use a method instead of a func).
							let mut vec: Vec<SymResTextPart<'_>> = Vec::new();
							if let Some(ref symbol) = if let Some(ref mut symbol_resolver) = $slf.symbol_resolver {
								to_owned(symbol_resolver.symbol($instruction, operand, Some(operand), $imm, $imm_size), &mut vec)
							} else {
								None
							} {
								call_write_symbol!($slf, $dst, $dst_next_p, $imm, symbol);
							} else {
								call_format_number!($slf, $dst, $dst_next_p, $imm);
							}
						} else {
							call_format_number!($slf, $dst, $dst_next_p, $imm);
						}
					}};
				}
				macro_rules! fmt_far_branch {
					($slf:ident, $dst:ident, $dst_next_p:ident, $instruction:ident, $op_kind:ident, $imm_size_ident:ident, $imm64:ident) => {{
						if $op_kind == OpKind::FarBranch32 {
							$imm_size_ident = 4;
							$imm64 = $instruction.far_branch32() as u64;
						} else {
							$imm_size_ident = 2;
							$imm64 = $instruction.far_branch16() as u64;
						}
						if TraitOptions::ENABLE_SYMBOL_RESOLVER {
							// See fmt_near_branch!() above for why we clone the symbols
							let mut vec: Vec<SymResTextPart<'_>> = Vec::new();
							let mut vec2: Vec<SymResTextPart<'_>> = Vec::new();
							if let Some(ref symbol) = if let Some(ref mut symbol_resolver) = $slf.symbol_resolver {
								to_owned(
									symbol_resolver.symbol($instruction, operand, Some(operand), $imm64 as u32 as u64, $imm_size_ident),
									&mut vec,
								)
							} else {
								None
							} {
								debug_assert!(operand + 1 == 1);
								let selector_symbol = if let Some(ref mut symbol_resolver) = $slf.symbol_resolver {
									to_owned(
										symbol_resolver.symbol(
											$instruction,
											operand + 1,
											Some(operand),
											$instruction.far_branch_selector() as u64,
											2,
										),
										&mut vec2,
									)
								} else {
									None
								};
								if let Some(ref selector_symbol) = selector_symbol {
									call_write_symbol!($slf, $dst, $dst_next_p, $instruction.far_branch_selector() as u64, selector_symbol);
								} else {
									call_format_number!($slf, $dst, $dst_next_p, $instruction.far_branch_selector() as u64);
								}
								write_fast_ascii_char_lit!(dst, dst_next_p, ':', true);
								call_write_symbol!($slf, $dst, $dst_next_p, $imm64, symbol);
							} else {
								call_format_number!($slf, $dst, $dst_next_p, $instruction.far_branch_selector() as u64);
								write_fast_ascii_char_lit!(dst, dst_next_p, ':', true);
								call_format_number!($slf, $dst, $dst_next_p, $imm64);
							}
						} else {
							call_format_number!($slf, $dst, $dst_next_p, $instruction.far_branch_selector() as u64);
							write_fast_ascii_char_lit!(dst, dst_next_p, ':', true);
							call_format_number!($slf, $dst, $dst_next_p, $imm64);
						}
					}};
				}
				macro_rules! fmt_imm {
					($slf:ident, $dst:ident, $dst_next_p:ident, $instruction:ident, $imm:ident, $imm_size:literal) => {
						#[allow(trivial_numeric_casts)]
						{
							if TraitOptions::ENABLE_SYMBOL_RESOLVER {
								// See fmt_near_branch!() above for why we clone the symbols
								let mut vec: Vec<SymResTextPart<'_>> = Vec::new();
								if let Some(ref symbol) = if let Some(ref mut symbol_resolver) = $slf.symbol_resolver {
									to_owned(symbol_resolver.symbol($instruction, operand, Some(operand), $imm as u64, $imm_size), &mut vec)
								} else {
									None
								} {
									if (symbol.flags & SymbolFlags::RELATIVE) == 0 {
										const FAST_STR: FastString8 = mk_const_fast_str!(FastString8, "\x07offset  ");
										write_fast_str!($dst, $dst_next_p, FastString8, FAST_STR);
									}
									call_write_symbol!($slf, $dst, $dst_next_p, $imm as u64, symbol);
								} else {
									call_format_number!($slf, $dst, $dst_next_p, $imm as u64);
								}
							} else {
								call_format_number!($slf, $dst, $dst_next_p, $imm as u64);
							}
						}
					};
				}
				macro_rules! fmt_register {
					() => {{
						call_format_register!(self, dst, dst_next_p, instruction_internal::internal_op_register(instruction, operand))
					}};
				}
				macro_rules! fmt_far_br_16_32 {
					() => {{
						fmt_far_branch!(self, dst, dst_next_p, instruction, op_kind, imm_size, imm64)
					}};
				}
				macro_rules! fmt_memory_seg_si {
					() => {{
						call_format_memory!(
							self,
							dst,
							dst_next_p,
							instruction,
							operand,
							(instruction.memory_segment()),
							(Register::SI),
							(Register::None),
							0,
							0,
							0,
							2,
						)
					}};
				}
				macro_rules! fmt_memory_seg_esi {
					() => {{
						call_format_memory!(
							self,
							dst,
							dst_next_p,
							instruction,
							operand,
							(instruction.memory_segment()),
							(Register::ESI),
							(Register::None),
							0,
							0,
							0,
							4,
						)
					}};
				}
				macro_rules! fmt_memory_seg_rsi {
					() => {{
						call_format_memory!(
							self,
							dst,
							dst_next_p,
							instruction,
							operand,
							(instruction.memory_segment()),
							(Register::RSI),
							(Register::None),
							0,
							0,
							0,
							8,
						)
					}};
				}
				macro_rules! fmt_memory_seg_di {
					() => {{
						call_format_memory!(
							self,
							dst,
							dst_next_p,
							instruction,
							operand,
							(instruction.memory_segment()),
							(Register::DI),
							(Register::None),
							0,
							0,
							0,
							2,
						)
					}};
				}
				macro_rules! fmt_memory_seg_edi {
					() => {{
						call_format_memory!(
							self,
							dst,
							dst_next_p,
							instruction,
							operand,
							(instruction.memory_segment()),
							(Register::EDI),
							(Register::None),
							0,
							0,
							0,
							4,
						)
					}};
				}
				macro_rules! fmt_memory_seg_rdi {
					() => {{
						call_format_memory!(
							self,
							dst,
							dst_next_p,
							instruction,
							operand,
							(instruction.memory_segment()),
							(Register::RDI),
							(Register::None),
							0,
							0,
							0,
							8,
						)
					}};
				}
				macro_rules! fmt_memory_es_di {
					() => {{
						call_format_memory!(self, dst, dst_next_p, instruction, operand, (Register::ES), (Register::DI), (Register::None), 0, 0, 0, 2)
					}};
				}
				macro_rules! fmt_memory_es_edi {
					() => {{
						call_format_memory!(
							self,
							dst,
							dst_next_p,
							instruction,
							operand,
							(Register::ES),
							(Register::EDI),
							(Register::None),
							0,
							0,
							0,
							4
						)
					}};
				}
				macro_rules! fmt_memory_es_rdi {
					() => {{
						call_format_memory!(
							self,
							dst,
							dst_next_p,
							instruction,
							operand,
							(Register::ES),
							(Register::RDI),
							(Register::None),
							0,
							0,
							0,
							8
						)
					}};
				}
				macro_rules! fmt_memory {
					() => {{
						let displ_size = instruction.memory_displ_size();
						let base_reg = instruction.memory_base();
						let mut index_reg = instruction.memory_index();
						let addr_size = get_address_size_in_bytes(base_reg, index_reg, displ_size, instruction.code_size());
						let displ =
							if addr_size == 8 { instruction.memory_displacement64() as i64 } else { instruction.memory_displacement32() as i64 };
						if code == Code::Xlat_m8 {
							index_reg = Register::None;
						}
						call_format_memory!(
							self,
							dst,
							dst_next_p,
							instruction,
							operand,
							(instruction.memory_segment()),
							base_reg,
							index_reg,
							(instruction_internal::internal_get_memory_index_scale(instruction)),
							displ_size,
							displ,
							addr_size,
						);
						#[cfg(feature = "mvex")]
						if instruction.is_mvex_eviction_hint() {
							const FAST_STR: FastString4 = mk_const_fast_str!(FastString4, "\x04{eh}");
							write_fast_str!(dst, dst_next_p, FastString4, FAST_STR);
						}
					}};
				}

				// This speeds up SpecializedFormatter since every option is hard coded, but makes FastFormatter
				// slower because every option is dynamic (more code). Detect FastFormatter and generate smaller
				// and faster code. Both paths are tested (same tests).
				// The whole point of this formatter is to be fast so unfortunately it can result in fugly code...
				if TraitOptions::__IS_FAST_FORMATTER {
					match op_kind {
						OpKind::Register => fmt_register!(),

						OpKind::NearBranch16 | OpKind::NearBranch32 | OpKind::NearBranch64 => {
							if op_kind == OpKind::NearBranch64 {
								imm_size = 8;
								imm64 = instruction.near_branch64();
							} else if op_kind == OpKind::NearBranch32 {
								imm_size = 4;
								imm64 = instruction.near_branch32() as u64;
							} else {
								imm_size = 2;
								imm64 = instruction.near_branch16() as u64;
							}
							fmt_near_branch!(self, dst, dst_next_p, instruction, imm_size, imm64);
						}

						OpKind::FarBranch16 | OpKind::FarBranch32 => fmt_far_br_16_32!(),

						OpKind::Immediate8 | OpKind::Immediate8_2nd => {
							if TraitOptions::ENABLE_DB_DW_DD_DQ && is_declare_data {
								imm8 = instruction.get_declare_byte_value(operand as usize);
							} else if op_kind == OpKind::Immediate8 {
								imm8 = instruction.immediate8();
							} else {
								debug_assert_eq!(op_kind, OpKind::Immediate8_2nd);
								imm8 = instruction.immediate8_2nd();
							}
							fmt_imm!(self, dst, dst_next_p, instruction, imm8, 1);
						}

						OpKind::Immediate16 | OpKind::Immediate8to16 => {
							if TraitOptions::ENABLE_DB_DW_DD_DQ && is_declare_data {
								imm16 = instruction.get_declare_word_value(operand as usize);
							} else if op_kind == OpKind::Immediate16 {
								imm16 = instruction.immediate16();
							} else {
								debug_assert_eq!(op_kind, OpKind::Immediate8to16);
								imm16 = instruction.immediate8to16() as u16;
							}
							fmt_imm!(self, dst, dst_next_p, instruction, imm16, 2)
						}

						OpKind::Immediate32 | OpKind::Immediate8to32 => {
							if TraitOptions::ENABLE_DB_DW_DD_DQ && is_declare_data {
								imm32 = instruction.get_declare_dword_value(operand as usize);
							} else if op_kind == OpKind::Immediate32 {
								imm32 = instruction.immediate32();
							} else {
								debug_assert_eq!(op_kind, OpKind::Immediate8to32);
								imm32 = instruction.immediate8to32() as u32;
							}
							fmt_imm!(self, dst, dst_next_p, instruction, imm32, 4)
						}

						OpKind::Immediate64 | OpKind::Immediate8to64 | OpKind::Immediate32to64 => {
							if TraitOptions::ENABLE_DB_DW_DD_DQ && is_declare_data {
								imm64 = instruction.get_declare_qword_value(operand as usize);
							} else if op_kind == OpKind::Immediate32to64 {
								imm64 = instruction.immediate32to64() as u64;
							} else if op_kind == OpKind::Immediate8to64 {
								imm64 = instruction.immediate8to64() as u64;
							} else {
								debug_assert_eq!(op_kind, OpKind::Immediate64);
								imm64 = instruction.immediate64();
							}
							fmt_imm!(self, dst, dst_next_p, instruction, imm64, 8)
						}

						OpKind::MemorySegSI => fmt_memory_seg_si!(),
						OpKind::MemorySegESI => fmt_memory_seg_esi!(),
						OpKind::MemorySegRSI => fmt_memory_seg_rsi!(),
						OpKind::MemorySegDI => fmt_memory_seg_di!(),
						OpKind::MemorySegEDI => fmt_memory_seg_edi!(),
						OpKind::MemorySegRDI => fmt_memory_seg_rdi!(),
						OpKind::MemoryESDI => fmt_memory_es_di!(),
						OpKind::MemoryESEDI => fmt_memory_es_edi!(),
						OpKind::MemoryESRDI => fmt_memory_es_rdi!(),
						OpKind::Memory => fmt_memory!(),
					}
				} else {
					match op_kind {
						OpKind::Register => fmt_register!(),

						OpKind::NearBranch16 => {
							imm64 = instruction.near_branch16() as u64;
							fmt_near_branch!(self, dst, dst_next_p, instruction, 2, imm64);
						}

						OpKind::NearBranch32 => {
							imm64 = instruction.near_branch32() as u64;
							fmt_near_branch!(self, dst, dst_next_p, instruction, 4, imm64);
						}

						OpKind::NearBranch64 => {
							imm64 = instruction.near_branch64();
							fmt_near_branch!(self, dst, dst_next_p, instruction, 8, imm64);
						}

						OpKind::FarBranch16 | OpKind::FarBranch32 => fmt_far_br_16_32!(),

						OpKind::Immediate8 => {
							imm8 = if TraitOptions::ENABLE_DB_DW_DD_DQ && is_declare_data {
								instruction.get_declare_byte_value(operand as usize)
							} else {
								instruction.immediate8()
							};
							fmt_imm!(self, dst, dst_next_p, instruction, imm8, 1);
						}

						OpKind::Immediate8_2nd => {
							imm8 = instruction.immediate8_2nd();
							fmt_imm!(self, dst, dst_next_p, instruction, imm8, 1);
						}

						OpKind::Immediate16 => {
							imm16 = if TraitOptions::ENABLE_DB_DW_DD_DQ && is_declare_data {
								instruction.get_declare_word_value(operand as usize)
							} else {
								instruction.immediate16()
							};
							fmt_imm!(self, dst, dst_next_p, instruction, imm16, 2)
						}

						OpKind::Immediate8to16 => {
							imm16 = instruction.immediate8to16() as u16;
							fmt_imm!(self, dst, dst_next_p, instruction, imm16, 2)
						}

						OpKind::Immediate32 => {
							imm32 = if TraitOptions::ENABLE_DB_DW_DD_DQ && is_declare_data {
								instruction.get_declare_dword_value(operand as usize)
							} else {
								instruction.immediate32()
							};
							fmt_imm!(self, dst, dst_next_p, instruction, imm32, 4)
						}

						OpKind::Immediate8to32 => {
							imm32 = instruction.immediate8to32() as u32;
							fmt_imm!(self, dst, dst_next_p, instruction, imm32, 4)
						}

						OpKind::Immediate64 => {
							imm64 = if TraitOptions::ENABLE_DB_DW_DD_DQ && is_declare_data {
								instruction.get_declare_qword_value(operand as usize)
							} else {
								instruction.immediate64()
							};
							fmt_imm!(self, dst, dst_next_p, instruction, imm64, 8)
						}

						OpKind::Immediate8to64 => {
							imm64 = instruction.immediate8to64() as u64;
							fmt_imm!(self, dst, dst_next_p, instruction, imm64, 8)
						}

						OpKind::Immediate32to64 => {
							imm64 = instruction.immediate32to64() as u64;
							fmt_imm!(self, dst, dst_next_p, instruction, imm64, 8)
						}

						OpKind::MemorySegSI => fmt_memory_seg_si!(),
						OpKind::MemorySegESI => fmt_memory_seg_esi!(),
						OpKind::MemorySegRSI => fmt_memory_seg_rsi!(),
						OpKind::MemorySegDI => fmt_memory_seg_di!(),
						OpKind::MemorySegEDI => fmt_memory_seg_edi!(),
						OpKind::MemorySegRDI => fmt_memory_seg_rdi!(),
						OpKind::MemoryESDI => fmt_memory_es_di!(),
						OpKind::MemoryESEDI => fmt_memory_es_edi!(),
						OpKind::MemoryESRDI => fmt_memory_es_rdi!(),
						OpKind::Memory => fmt_memory!(),
					}
				}

				if operand == 0 && instruction_internal::internal_has_op_mask_or_zeroing_masking(instruction) {
					if instruction.has_op_mask() {
						write_fast_ascii_char_lit!(dst, dst_next_p, '{', true);
						call_format_register!(self, dst, dst_next_p, instruction.op_mask());
						write_fast_ascii_char_lit!(dst, dst_next_p, '}', true);
					}
					if instruction.zeroing_masking() {
						const FAST_STR: FastString4 = mk_const_fast_str!(FastString4, "\x03{z} ");
						write_fast_str!(dst, dst_next_p, FastString4, FAST_STR);
					}
				}
				#[cfg(feature = "mvex")]
				if mvex_rm_operand == operand {
					let conv = instruction.mvex_reg_mem_conv();
					if conv != MvexRegMemConv::None {
						let mvex = crate::mvex::get_mvex_info(instruction.code());
						if mvex.conv_fn != MvexConvFn::None {
							let tbl = if mvex.is_conv_fn_32() { &MVEX_REG_MEM_CONSTS_32 } else { &MVEX_REG_MEM_CONSTS_64 };
							let s = tbl[conv as usize];
							write_fast_str!(dst, dst_next_p, FastString12, s);
						}
					}
				}

				operand += 1;
				if operand >= op_count {
					break;
				}

				if TraitOptions::space_after_operand_separator(&self.d.options) {
					const FAST_STR: FastString4 = mk_const_fast_str!(FastString4, "\x02,   ");
					write_fast_str!(dst, dst_next_p, FastString4, FAST_STR);
				} else {
					write_fast_ascii_char_lit!(dst, dst_next_p, ',', true);
				}
			}
			if instruction_internal::internal_has_rounding_control_or_sae(instruction) {
				let rc = instruction.rounding_control();
				if rc != RoundingControl::None {
					if IcedConstants::is_mvex(instruction.code()) && !instruction.suppress_all_exceptions() {
						let fast_str = RC_STRINGS[rc as usize];
						write_fast_str!(dst, dst_next_p, FastString4, fast_str);
					} else {
						let fast_str = RC_SAE_STRINGS[rc as usize];
						write_fast_str!(dst, dst_next_p, FastString8, fast_str);
					}
				} else {
					debug_assert!(instruction.suppress_all_exceptions());
					const FAST_STR: FastString8 = mk_const_fast_str!(FastString8, "\x05{sae}   ");
					write_fast_str!(dst, dst_next_p, FastString8, FAST_STR);
				}
			}
		}

		update_vec_len!(dst, dst_next_p);
	}

	// Only one caller so inline it
	#[must_use]
	#[inline]
	fn show_segment_prefix(instruction: &Instruction, op_count: u32) -> bool {
		for i in 0..op_count {
			match instruction.op_kind(i) {
				OpKind::Register
				| OpKind::NearBranch16
				| OpKind::NearBranch32
				| OpKind::NearBranch64
				| OpKind::FarBranch16
				| OpKind::FarBranch32
				| OpKind::Immediate8
				| OpKind::Immediate8_2nd
				| OpKind::Immediate16
				| OpKind::Immediate32
				| OpKind::Immediate64
				| OpKind::Immediate8to16
				| OpKind::Immediate8to32
				| OpKind::Immediate8to64
				| OpKind::Immediate32to64
				| OpKind::MemoryESDI
				| OpKind::MemoryESEDI
				| OpKind::MemoryESRDI => {}

				OpKind::MemorySegSI
				| OpKind::MemorySegESI
				| OpKind::MemorySegRSI
				| OpKind::MemorySegDI
				| OpKind::MemorySegEDI
				| OpKind::MemorySegRDI
				| OpKind::Memory => return false,
			}
		}

		SpecializedFormatter::<TraitOptions>::SHOW_USELESS_PREFIXES
	}

	#[inline]
	#[must_use]
	fn format_register(&self, dst: &mut Vec<u8>, mut dst_next_p: *mut u8, register: Register) -> *mut u8 {
		let reg_str = self.d.all_registers[register as usize];
		write_fast_str!(dst, dst_next_p, FastStringRegister, reg_str);
		dst_next_p
	}

	#[must_use]
	fn format_number(&self, dst: &mut Vec<u8>, mut dst_next_p: *mut u8, value: u64) -> *mut u8 {
		macro_rules! format_number_impl {
			($dst:ident, $dst_next_p:ident, $value:ident, $uppercase_hex:literal, $use_hex_prefix:literal) => {{
				if $use_hex_prefix {
					const FAST_STR: FastString4 = mk_const_fast_str!(FastString4, "\x020x  ");
					write_fast_str!($dst, $dst_next_p, FastString4, FAST_STR);
				}

				if $value < 0x10 {
					if $use_hex_prefix {
						let hex_table = if $uppercase_hex { b"0123456789ABCDEF" } else { b"0123456789abcdef" };
						// SAFETY: 0<=$value<=0xF and hex_table.len() == 0x10
						let c = unsafe { *hex_table.get_unchecked($value as usize) };
						write_fast_ascii_char!($dst, $dst_next_p, c, true);

						$dst_next_p
					} else {
						// 1 (possible '0' prefix) + 1 (hex digit) + 1 ('h' suffix)
						verify_output_has_enough_bytes_left!($dst, $dst_next_p, 1 + 1 + 1);
						if $value > 9 {
							write_fast_ascii_char_lit!($dst, $dst_next_p, '0', false);
						}

						let hex_table = if $uppercase_hex { b"0123456789ABCDEF" } else { b"0123456789abcdef" };
						// SAFETY: 0<=$value<=0xF and hex_table.len() == 0x10
						let c = unsafe { *hex_table.get_unchecked($value as usize) };
						write_fast_ascii_char!($dst, $dst_next_p, c, false);
						write_fast_ascii_char_lit!($dst, $dst_next_p, 'h', false);

						$dst_next_p
					}
				} else if $value < 0x100 {
					if $use_hex_prefix {
						let lower_or_value = if $uppercase_hex { 0 } else { 0x2020_2020 };
						write_fast_hex2_rw_4bytes!($dst, $dst_next_p, $value, lower_or_value, true);

						$dst_next_p
					} else {
						// 1 (possible '0' prefix) + 2 (hex digits) + 2 since
						// write_fast_hex2_rw_4bytes!() reads/writes 4 bytes and not 2.
						// '+2' also includes the 'h' suffix.
						verify_output_has_enough_bytes_left!($dst, $dst_next_p, 1 + 2 + 2);
						if $value > 0x9F {
							write_fast_ascii_char_lit!($dst, $dst_next_p, '0', false);
						}

						let lower_or_value = if $uppercase_hex { 0 } else { 0x2020_2020 };
						write_fast_hex2_rw_4bytes!($dst, $dst_next_p, $value, lower_or_value, false);
						write_fast_ascii_char_lit!($dst, $dst_next_p, 'h', false);

						$dst_next_p
					}
				} else {
					let mut rshift = ((64 - u64::leading_zeros($value) + 3) & !3) as usize;

					// The first '1' is an optional '0' prefix.
					// `rshift / 4` == number of hex digits to copy. The last `+ 2` is the extra padding needed
					// since the write_fast_hex2_rw_4bytes!() macro reads and writes 4 bytes (2 hex digits + 2 bytes padding).
					// '+2' also includes the 'h' suffix.
					verify_output_has_enough_bytes_left!($dst, $dst_next_p, 1 + rshift / 4 + 2);

					if !$use_hex_prefix && (($value >> (rshift - 4)) & 0xF) > 9 {
						write_fast_ascii_char_lit!($dst, $dst_next_p, '0', false);
					}

					// If odd number of hex digits
					if (rshift & 4) != 0 {
						rshift -= 4;
						let hex_table = if $uppercase_hex { b"0123456789ABCDEF" } else { b"0123456789abcdef" };
						let digit = (($value >> rshift) & 0xF) as usize;
						// SAFETY: 0<=digit<=0xF and hex_table.len() == 0x10
						let c = unsafe { *hex_table.get_unchecked(digit) };
						write_fast_ascii_char!($dst, $dst_next_p, c, false);
					}

					// If we're here, $value >= 0x100 so rshift >= 8
					debug_assert!(rshift >= 8);
					let lower_or_value = if $uppercase_hex { 0 } else { 0x2020_2020 };
					loop {
						rshift -= 8;
						let digits2 = (($value >> rshift) & 0xFF) as usize;
						write_fast_hex2_rw_4bytes!($dst, $dst_next_p, digits2, lower_or_value, false);

						if rshift == 0 {
							break;
						}
					}

					if !$use_hex_prefix {
						// We've verified that `dst` had `1 + rshift / 4 + 2` bytes left (see above).
						// The last `+2` is the padding that needed to be there. That's where
						// this 'h' gets written so we don't need to verify the vec len here
						// because it has at least 2 more bytes left.
						write_fast_ascii_char_lit!($dst, $dst_next_p, 'h', false);
					}

					$dst_next_p
				}
			}};
		}

		if TraitOptions::uppercase_hex(&self.d.options) {
			if TraitOptions::use_hex_prefix(&self.d.options) {
				// 0x12AB
				format_number_impl!(dst, dst_next_p, value, true, true)
			} else {
				// 12ABh
				format_number_impl!(dst, dst_next_p, value, true, false)
			}
		} else {
			if TraitOptions::use_hex_prefix(&self.d.options) {
				// 0x12ab
				format_number_impl!(dst, dst_next_p, value, false, true)
			} else {
				// 12abh
				format_number_impl!(dst, dst_next_p, value, false, false)
			}
		}
	}

	#[inline]
	#[must_use]
	fn write_symbol(&self, dst: &mut Vec<u8>, mut dst_next_p: *mut u8, address: u64, symbol: &SymbolResult<'_>) -> *mut u8 {
		call_write_symbol2!(self, dst, dst_next_p, address, symbol, true);
		dst_next_p
	}

	#[cold]
	#[must_use]
	fn write_symbol2(
		&self, dst: &mut Vec<u8>, mut dst_next_p: *mut u8, address: u64, symbol: &SymbolResult<'_>, write_minus_if_signed: bool,
	) -> *mut u8 {
		let mut displ = address.wrapping_sub(symbol.address) as i64;
		if (symbol.flags & SymbolFlags::SIGNED) != 0 {
			if write_minus_if_signed {
				write_fast_ascii_char_lit!(dst, dst_next_p, '-', true);
			}
			displ = displ.wrapping_neg();
		}

		// Write the symbol. The symbol can be any length and is a `&'a str` so we must
		// write using `dst`. The macro will invalidate `dst_next_p` and will restore
		// it after the match statement.
		use_dst_only_now!(dst, dst_next_p);
		match symbol.text {
			SymResTextInfo::Text(ref part) => {
				let s = match &part.text {
					&SymResString::Str(s) => s,
					SymResString::String(s) => s.as_str(),
				};
				dst.extend_from_slice(s.as_bytes());
			}

			SymResTextInfo::TextVec(v) => {
				for part in v {
					let s = match &part.text {
						&SymResString::Str(s) => s,
						SymResString::String(s) => s.as_str(),
					};
					dst.extend_from_slice(s.as_bytes());
				}
			}
		}
		use_dst_next_p_now!(dst, dst_next_p);

		if displ != 0 {
			let c = if displ < 0 {
				displ = displ.wrapping_neg();
				'-'
			} else {
				'+'
			};
			write_fast_ascii_char!(dst, dst_next_p, c, true);
			call_format_number!(self, dst, dst_next_p, displ as u64);
		}
		if TraitOptions::show_symbol_address(&self.d.options) {
			const FAST_STR: FastString4 = mk_const_fast_str!(FastString4, "\x02 (  ");
			write_fast_str!(dst, dst_next_p, FastString4, FAST_STR);
			call_format_number!(self, dst, dst_next_p, address);
			write_fast_ascii_char_lit!(dst, dst_next_p, ')', true);
		}

		dst_next_p
	}

	#[must_use]
	fn format_memory(
		&mut self, dst: &mut Vec<u8>, mut dst_next_p: *mut u8, instruction: &Instruction, operand: u32, seg_reg: Register, base_reg: Register,
		index_reg: Register, scale: u32, displ_size: u32, displ: i64, addr_size: u32,
	) -> *mut u8 {
		format_memory_code!(self, dst, dst_next_p, instruction, operand, seg_reg, base_reg, index_reg, scale, displ_size, displ, addr_size);
		dst_next_p
	}
}

/// Fast formatter with less formatting options and with a masm-like syntax.
/// Use it if formatting speed is more important than being able to re-assemble formatted instructions.
///
/// This is a variant of [`SpecializedFormatter<TraitOptions>`] and allows changing the
/// formatter options at runtime and the use of a symbol resolver. For fastest possible
/// disassembly and smallest code, the options should be hard coded, so see [`SpecializedFormatter<TraitOptions>`].
///
/// This formatter is ~2.8x faster than the gas/intel/masm/nasm formatters (the time includes decoding + formatting).
///
/// [`SpecializedFormatter<TraitOptions>`]: struct.SpecializedFormatter.html
///
/// # Examples
///
/// ```
/// use iced_x86::*;
///
/// let bytes = b"\x62\xF2\x4F\xDD\x72\x50\x01";
/// let mut decoder = Decoder::new(64, bytes, DecoderOptions::NONE);
/// let instr = decoder.decode();
///
/// let mut output = String::new();
/// let mut formatter = FastFormatter::new();
/// formatter.options_mut().set_space_after_operand_separator(true);
/// formatter.format(&instr, &mut output);
/// assert_eq!(output, "vcvtne2ps2bf16 zmm2{k5}{z}, zmm6, dword bcst [rax+4h]");
/// ```
///
/// # Using a symbol resolver
///
/// ```
/// use iced_x86::*;
/// use std::collections::HashMap;
///
/// let bytes = b"\x48\x8B\x8A\xA5\x5A\xA5\x5A";
/// let mut decoder = Decoder::new(64, bytes, DecoderOptions::NONE);
/// let instr = decoder.decode();
///
/// struct MySymbolResolver { map: HashMap<u64, String> }
/// impl SymbolResolver for MySymbolResolver {
///     fn symbol(&mut self, _instruction: &Instruction, _operand: u32, _instruction_operand: Option<u32>,
///          address: u64, _address_size: u32) -> Option<SymbolResult> {
///         if let Some(symbol_string) = self.map.get(&address) {
///             // The 'address' arg is the address of the symbol and doesn't have to be identical
///             // to the 'address' arg passed to symbol(). If it's different from the input
///             // address, the formatter will add +N or -N, eg. '[rax+symbol+123]'
///             Some(SymbolResult::with_str(address, symbol_string.as_str()))
///         } else {
///             None
///         }
///     }
/// }
///
/// // Hard code the symbols, it's just an example!😄
/// let mut sym_map: HashMap<u64, String> = HashMap::new();
/// sym_map.insert(0x5AA55AA5, String::from("my_data"));
///
/// let mut output = String::new();
/// let resolver = Box::new(MySymbolResolver { map: sym_map });
/// let mut formatter = FastFormatter::try_with_options(Some(resolver)).unwrap();
/// formatter.format(&instr, &mut output);
/// assert_eq!("mov rcx,[rdx+my_data]", output);
/// ```
pub type FastFormatter = SpecializedFormatter<DefaultFastFormatterTraitOptions>;

/// Default [`SpecializedFormatter<TraitOptions>`] options. It doesn't override any `const` or `fn`
///
/// [`SpecializedFormatter<TraitOptions>`]: struct.SpecializedFormatter.html
#[allow(missing_copy_implementations)]
#[allow(missing_debug_implementations)]
pub struct DefaultSpecializedFormatterTraitOptions;
impl SpecializedFormatterTraitOptions for DefaultSpecializedFormatterTraitOptions {}