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
|
//===-- TargetLowering.cpp - Implement the TargetLowering class -----------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This implements the TargetLowering class.
//
//===----------------------------------------------------------------------===//
#include "llvm/Target/TargetAsmInfo.h"
#include "llvm/Target/TargetLowering.h"
#include "llvm/Target/TargetSubtarget.h"
#include "llvm/Target/TargetData.h"
#include "llvm/Target/TargetMachine.h"
#include "llvm/Target/TargetRegisterInfo.h"
#include "llvm/GlobalVariable.h"
#include "llvm/DerivedTypes.h"
#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/SelectionDAG.h"
#include "llvm/ADT/StringExtras.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/Support/MathExtras.h"
using namespace llvm;
namespace llvm {
TLSModel::Model getTLSModel(const GlobalValue *GV, Reloc::Model reloc) {
bool isLocal = GV->hasLocalLinkage();
bool isDeclaration = GV->isDeclaration();
// FIXME: what should we do for protected and internal visibility?
// For variables, is internal different from hidden?
bool isHidden = GV->hasHiddenVisibility();
if (reloc == Reloc::PIC_) {
if (isLocal || isHidden)
return TLSModel::LocalDynamic;
else
return TLSModel::GeneralDynamic;
} else {
if (!isDeclaration || isHidden)
return TLSModel::LocalExec;
else
return TLSModel::InitialExec;
}
}
}
/// InitLibcallNames - Set default libcall names.
///
static void InitLibcallNames(const char **Names) {
Names[RTLIB::SHL_I16] = "__ashli16";
Names[RTLIB::SHL_I32] = "__ashlsi3";
Names[RTLIB::SHL_I64] = "__ashldi3";
Names[RTLIB::SHL_I128] = "__ashlti3";
Names[RTLIB::SRL_I16] = "__lshri16";
Names[RTLIB::SRL_I32] = "__lshrsi3";
Names[RTLIB::SRL_I64] = "__lshrdi3";
Names[RTLIB::SRL_I128] = "__lshrti3";
Names[RTLIB::SRA_I16] = "__ashri16";
Names[RTLIB::SRA_I32] = "__ashrsi3";
Names[RTLIB::SRA_I64] = "__ashrdi3";
Names[RTLIB::SRA_I128] = "__ashrti3";
Names[RTLIB::MUL_I16] = "__muli16";
Names[RTLIB::MUL_I32] = "__mulsi3";
Names[RTLIB::MUL_I64] = "__muldi3";
Names[RTLIB::MUL_I128] = "__multi3";
Names[RTLIB::SDIV_I32] = "__divsi3";
Names[RTLIB::SDIV_I64] = "__divdi3";
Names[RTLIB::SDIV_I128] = "__divti3";
Names[RTLIB::UDIV_I32] = "__udivsi3";
Names[RTLIB::UDIV_I64] = "__udivdi3";
Names[RTLIB::UDIV_I128] = "__udivti3";
Names[RTLIB::SREM_I32] = "__modsi3";
Names[RTLIB::SREM_I64] = "__moddi3";
Names[RTLIB::SREM_I128] = "__modti3";
Names[RTLIB::UREM_I32] = "__umodsi3";
Names[RTLIB::UREM_I64] = "__umoddi3";
Names[RTLIB::UREM_I128] = "__umodti3";
Names[RTLIB::NEG_I32] = "__negsi2";
Names[RTLIB::NEG_I64] = "__negdi2";
Names[RTLIB::ADD_F32] = "__addsf3";
Names[RTLIB::ADD_F64] = "__adddf3";
Names[RTLIB::ADD_F80] = "__addxf3";
Names[RTLIB::ADD_PPCF128] = "__gcc_qadd";
Names[RTLIB::SUB_F32] = "__subsf3";
Names[RTLIB::SUB_F64] = "__subdf3";
Names[RTLIB::SUB_F80] = "__subxf3";
Names[RTLIB::SUB_PPCF128] = "__gcc_qsub";
Names[RTLIB::MUL_F32] = "__mulsf3";
Names[RTLIB::MUL_F64] = "__muldf3";
Names[RTLIB::MUL_F80] = "__mulxf3";
Names[RTLIB::MUL_PPCF128] = "__gcc_qmul";
Names[RTLIB::DIV_F32] = "__divsf3";
Names[RTLIB::DIV_F64] = "__divdf3";
Names[RTLIB::DIV_F80] = "__divxf3";
Names[RTLIB::DIV_PPCF128] = "__gcc_qdiv";
Names[RTLIB::REM_F32] = "fmodf";
Names[RTLIB::REM_F64] = "fmod";
Names[RTLIB::REM_F80] = "fmodl";
Names[RTLIB::REM_PPCF128] = "fmodl";
Names[RTLIB::POWI_F32] = "__powisf2";
Names[RTLIB::POWI_F64] = "__powidf2";
Names[RTLIB::POWI_F80] = "__powixf2";
Names[RTLIB::POWI_PPCF128] = "__powitf2";
Names[RTLIB::SQRT_F32] = "sqrtf";
Names[RTLIB::SQRT_F64] = "sqrt";
Names[RTLIB::SQRT_F80] = "sqrtl";
Names[RTLIB::SQRT_PPCF128] = "sqrtl";
Names[RTLIB::LOG_F32] = "logf";
Names[RTLIB::LOG_F64] = "log";
Names[RTLIB::LOG_F80] = "logl";
Names[RTLIB::LOG_PPCF128] = "logl";
Names[RTLIB::LOG2_F32] = "log2f";
Names[RTLIB::LOG2_F64] = "log2";
Names[RTLIB::LOG2_F80] = "log2l";
Names[RTLIB::LOG2_PPCF128] = "log2l";
Names[RTLIB::LOG10_F32] = "log10f";
Names[RTLIB::LOG10_F64] = "log10";
Names[RTLIB::LOG10_F80] = "log10l";
Names[RTLIB::LOG10_PPCF128] = "log10l";
Names[RTLIB::EXP_F32] = "expf";
Names[RTLIB::EXP_F64] = "exp";
Names[RTLIB::EXP_F80] = "expl";
Names[RTLIB::EXP_PPCF128] = "expl";
Names[RTLIB::EXP2_F32] = "exp2f";
Names[RTLIB::EXP2_F64] = "exp2";
Names[RTLIB::EXP2_F80] = "exp2l";
Names[RTLIB::EXP2_PPCF128] = "exp2l";
Names[RTLIB::SIN_F32] = "sinf";
Names[RTLIB::SIN_F64] = "sin";
Names[RTLIB::SIN_F80] = "sinl";
Names[RTLIB::SIN_PPCF128] = "sinl";
Names[RTLIB::COS_F32] = "cosf";
Names[RTLIB::COS_F64] = "cos";
Names[RTLIB::COS_F80] = "cosl";
Names[RTLIB::COS_PPCF128] = "cosl";
Names[RTLIB::POW_F32] = "powf";
Names[RTLIB::POW_F64] = "pow";
Names[RTLIB::POW_F80] = "powl";
Names[RTLIB::POW_PPCF128] = "powl";
Names[RTLIB::CEIL_F32] = "ceilf";
Names[RTLIB::CEIL_F64] = "ceil";
Names[RTLIB::CEIL_F80] = "ceill";
Names[RTLIB::CEIL_PPCF128] = "ceill";
Names[RTLIB::TRUNC_F32] = "truncf";
Names[RTLIB::TRUNC_F64] = "trunc";
Names[RTLIB::TRUNC_F80] = "truncl";
Names[RTLIB::TRUNC_PPCF128] = "truncl";
Names[RTLIB::RINT_F32] = "rintf";
Names[RTLIB::RINT_F64] = "rint";
Names[RTLIB::RINT_F80] = "rintl";
Names[RTLIB::RINT_PPCF128] = "rintl";
Names[RTLIB::NEARBYINT_F32] = "nearbyintf";
Names[RTLIB::NEARBYINT_F64] = "nearbyint";
Names[RTLIB::NEARBYINT_F80] = "nearbyintl";
Names[RTLIB::NEARBYINT_PPCF128] = "nearbyintl";
Names[RTLIB::FLOOR_F32] = "floorf";
Names[RTLIB::FLOOR_F64] = "floor";
Names[RTLIB::FLOOR_F80] = "floorl";
Names[RTLIB::FLOOR_PPCF128] = "floorl";
Names[RTLIB::FPEXT_F32_F64] = "__extendsfdf2";
Names[RTLIB::FPROUND_F64_F32] = "__truncdfsf2";
Names[RTLIB::FPROUND_F80_F32] = "__truncxfsf2";
Names[RTLIB::FPROUND_PPCF128_F32] = "__trunctfsf2";
Names[RTLIB::FPROUND_F80_F64] = "__truncxfdf2";
Names[RTLIB::FPROUND_PPCF128_F64] = "__trunctfdf2";
Names[RTLIB::FPTOSINT_F32_I32] = "__fixsfsi";
Names[RTLIB::FPTOSINT_F32_I64] = "__fixsfdi";
Names[RTLIB::FPTOSINT_F32_I128] = "__fixsfti";
Names[RTLIB::FPTOSINT_F64_I32] = "__fixdfsi";
Names[RTLIB::FPTOSINT_F64_I64] = "__fixdfdi";
Names[RTLIB::FPTOSINT_F64_I128] = "__fixdfti";
Names[RTLIB::FPTOSINT_F80_I32] = "__fixxfsi";
Names[RTLIB::FPTOSINT_F80_I64] = "__fixxfdi";
Names[RTLIB::FPTOSINT_F80_I128] = "__fixxfti";
Names[RTLIB::FPTOSINT_PPCF128_I32] = "__fixtfsi";
Names[RTLIB::FPTOSINT_PPCF128_I64] = "__fixtfdi";
Names[RTLIB::FPTOSINT_PPCF128_I128] = "__fixtfti";
Names[RTLIB::FPTOUINT_F32_I32] = "__fixunssfsi";
Names[RTLIB::FPTOUINT_F32_I64] = "__fixunssfdi";
Names[RTLIB::FPTOUINT_F32_I128] = "__fixunssfti";
Names[RTLIB::FPTOUINT_F64_I32] = "__fixunsdfsi";
Names[RTLIB::FPTOUINT_F64_I64] = "__fixunsdfdi";
Names[RTLIB::FPTOUINT_F64_I128] = "__fixunsdfti";
Names[RTLIB::FPTOUINT_F80_I32] = "__fixunsxfsi";
Names[RTLIB::FPTOUINT_F80_I64] = "__fixunsxfdi";
Names[RTLIB::FPTOUINT_F80_I128] = "__fixunsxfti";
Names[RTLIB::FPTOUINT_PPCF128_I32] = "__fixunstfsi";
Names[RTLIB::FPTOUINT_PPCF128_I64] = "__fixunstfdi";
Names[RTLIB::FPTOUINT_PPCF128_I128] = "__fixunstfti";
Names[RTLIB::SINTTOFP_I32_F32] = "__floatsisf";
Names[RTLIB::SINTTOFP_I32_F64] = "__floatsidf";
Names[RTLIB::SINTTOFP_I32_F80] = "__floatsixf";
Names[RTLIB::SINTTOFP_I32_PPCF128] = "__floatsitf";
Names[RTLIB::SINTTOFP_I64_F32] = "__floatdisf";
Names[RTLIB::SINTTOFP_I64_F64] = "__floatdidf";
Names[RTLIB::SINTTOFP_I64_F80] = "__floatdixf";
Names[RTLIB::SINTTOFP_I64_PPCF128] = "__floatditf";
Names[RTLIB::SINTTOFP_I128_F32] = "__floattisf";
Names[RTLIB::SINTTOFP_I128_F64] = "__floattidf";
Names[RTLIB::SINTTOFP_I128_F80] = "__floattixf";
Names[RTLIB::SINTTOFP_I128_PPCF128] = "__floattitf";
Names[RTLIB::UINTTOFP_I32_F32] = "__floatunsisf";
Names[RTLIB::UINTTOFP_I32_F64] = "__floatunsidf";
Names[RTLIB::UINTTOFP_I32_F80] = "__floatunsixf";
Names[RTLIB::UINTTOFP_I32_PPCF128] = "__floatunsitf";
Names[RTLIB::UINTTOFP_I64_F32] = "__floatundisf";
Names[RTLIB::UINTTOFP_I64_F64] = "__floatundidf";
Names[RTLIB::UINTTOFP_I64_F80] = "__floatundixf";
Names[RTLIB::UINTTOFP_I64_PPCF128] = "__floatunditf";
Names[RTLIB::UINTTOFP_I128_F32] = "__floatuntisf";
Names[RTLIB::UINTTOFP_I128_F64] = "__floatuntidf";
Names[RTLIB::UINTTOFP_I128_F80] = "__floatuntixf";
Names[RTLIB::UINTTOFP_I128_PPCF128] = "__floatuntitf";
Names[RTLIB::OEQ_F32] = "__eqsf2";
Names[RTLIB::OEQ_F64] = "__eqdf2";
Names[RTLIB::UNE_F32] = "__nesf2";
Names[RTLIB::UNE_F64] = "__nedf2";
Names[RTLIB::OGE_F32] = "__gesf2";
Names[RTLIB::OGE_F64] = "__gedf2";
Names[RTLIB::OLT_F32] = "__ltsf2";
Names[RTLIB::OLT_F64] = "__ltdf2";
Names[RTLIB::OLE_F32] = "__lesf2";
Names[RTLIB::OLE_F64] = "__ledf2";
Names[RTLIB::OGT_F32] = "__gtsf2";
Names[RTLIB::OGT_F64] = "__gtdf2";
Names[RTLIB::UO_F32] = "__unordsf2";
Names[RTLIB::UO_F64] = "__unorddf2";
Names[RTLIB::O_F32] = "__unordsf2";
Names[RTLIB::O_F64] = "__unorddf2";
}
/// getFPEXT - Return the FPEXT_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getFPEXT(MVT OpVT, MVT RetVT) {
if (OpVT == MVT::f32) {
if (RetVT == MVT::f64)
return FPEXT_F32_F64;
}
return UNKNOWN_LIBCALL;
}
/// getFPROUND - Return the FPROUND_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getFPROUND(MVT OpVT, MVT RetVT) {
if (RetVT == MVT::f32) {
if (OpVT == MVT::f64)
return FPROUND_F64_F32;
if (OpVT == MVT::f80)
return FPROUND_F80_F32;
if (OpVT == MVT::ppcf128)
return FPROUND_PPCF128_F32;
} else if (RetVT == MVT::f64) {
if (OpVT == MVT::f80)
return FPROUND_F80_F64;
if (OpVT == MVT::ppcf128)
return FPROUND_PPCF128_F64;
}
return UNKNOWN_LIBCALL;
}
/// getFPTOSINT - Return the FPTOSINT_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getFPTOSINT(MVT OpVT, MVT RetVT) {
if (OpVT == MVT::f32) {
if (RetVT == MVT::i32)
return FPTOSINT_F32_I32;
if (RetVT == MVT::i64)
return FPTOSINT_F32_I64;
if (RetVT == MVT::i128)
return FPTOSINT_F32_I128;
} else if (OpVT == MVT::f64) {
if (RetVT == MVT::i32)
return FPTOSINT_F64_I32;
if (RetVT == MVT::i64)
return FPTOSINT_F64_I64;
if (RetVT == MVT::i128)
return FPTOSINT_F64_I128;
} else if (OpVT == MVT::f80) {
if (RetVT == MVT::i32)
return FPTOSINT_F80_I32;
if (RetVT == MVT::i64)
return FPTOSINT_F80_I64;
if (RetVT == MVT::i128)
return FPTOSINT_F80_I128;
} else if (OpVT == MVT::ppcf128) {
if (RetVT == MVT::i32)
return FPTOSINT_PPCF128_I32;
if (RetVT == MVT::i64)
return FPTOSINT_PPCF128_I64;
if (RetVT == MVT::i128)
return FPTOSINT_PPCF128_I128;
}
return UNKNOWN_LIBCALL;
}
/// getFPTOUINT - Return the FPTOUINT_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getFPTOUINT(MVT OpVT, MVT RetVT) {
if (OpVT == MVT::f32) {
if (RetVT == MVT::i32)
return FPTOUINT_F32_I32;
if (RetVT == MVT::i64)
return FPTOUINT_F32_I64;
if (RetVT == MVT::i128)
return FPTOUINT_F32_I128;
} else if (OpVT == MVT::f64) {
if (RetVT == MVT::i32)
return FPTOUINT_F64_I32;
if (RetVT == MVT::i64)
return FPTOUINT_F64_I64;
if (RetVT == MVT::i128)
return FPTOUINT_F64_I128;
} else if (OpVT == MVT::f80) {
if (RetVT == MVT::i32)
return FPTOUINT_F80_I32;
if (RetVT == MVT::i64)
return FPTOUINT_F80_I64;
if (RetVT == MVT::i128)
return FPTOUINT_F80_I128;
} else if (OpVT == MVT::ppcf128) {
if (RetVT == MVT::i32)
return FPTOUINT_PPCF128_I32;
if (RetVT == MVT::i64)
return FPTOUINT_PPCF128_I64;
if (RetVT == MVT::i128)
return FPTOUINT_PPCF128_I128;
}
return UNKNOWN_LIBCALL;
}
/// getSINTTOFP - Return the SINTTOFP_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getSINTTOFP(MVT OpVT, MVT RetVT) {
if (OpVT == MVT::i32) {
if (RetVT == MVT::f32)
return SINTTOFP_I32_F32;
else if (RetVT == MVT::f64)
return SINTTOFP_I32_F64;
else if (RetVT == MVT::f80)
return SINTTOFP_I32_F80;
else if (RetVT == MVT::ppcf128)
return SINTTOFP_I32_PPCF128;
} else if (OpVT == MVT::i64) {
if (RetVT == MVT::f32)
return SINTTOFP_I64_F32;
else if (RetVT == MVT::f64)
return SINTTOFP_I64_F64;
else if (RetVT == MVT::f80)
return SINTTOFP_I64_F80;
else if (RetVT == MVT::ppcf128)
return SINTTOFP_I64_PPCF128;
} else if (OpVT == MVT::i128) {
if (RetVT == MVT::f32)
return SINTTOFP_I128_F32;
else if (RetVT == MVT::f64)
return SINTTOFP_I128_F64;
else if (RetVT == MVT::f80)
return SINTTOFP_I128_F80;
else if (RetVT == MVT::ppcf128)
return SINTTOFP_I128_PPCF128;
}
return UNKNOWN_LIBCALL;
}
/// getUINTTOFP - Return the UINTTOFP_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getUINTTOFP(MVT OpVT, MVT RetVT) {
if (OpVT == MVT::i32) {
if (RetVT == MVT::f32)
return UINTTOFP_I32_F32;
else if (RetVT == MVT::f64)
return UINTTOFP_I32_F64;
else if (RetVT == MVT::f80)
return UINTTOFP_I32_F80;
else if (RetVT == MVT::ppcf128)
return UINTTOFP_I32_PPCF128;
} else if (OpVT == MVT::i64) {
if (RetVT == MVT::f32)
return UINTTOFP_I64_F32;
else if (RetVT == MVT::f64)
return UINTTOFP_I64_F64;
else if (RetVT == MVT::f80)
return UINTTOFP_I64_F80;
else if (RetVT == MVT::ppcf128)
return UINTTOFP_I64_PPCF128;
} else if (OpVT == MVT::i128) {
if (RetVT == MVT::f32)
return UINTTOFP_I128_F32;
else if (RetVT == MVT::f64)
return UINTTOFP_I128_F64;
else if (RetVT == MVT::f80)
return UINTTOFP_I128_F80;
else if (RetVT == MVT::ppcf128)
return UINTTOFP_I128_PPCF128;
}
return UNKNOWN_LIBCALL;
}
/// InitCmpLibcallCCs - Set default comparison libcall CC.
///
static void InitCmpLibcallCCs(ISD::CondCode *CCs) {
memset(CCs, ISD::SETCC_INVALID, sizeof(ISD::CondCode)*RTLIB::UNKNOWN_LIBCALL);
CCs[RTLIB::OEQ_F32] = ISD::SETEQ;
CCs[RTLIB::OEQ_F64] = ISD::SETEQ;
CCs[RTLIB::UNE_F32] = ISD::SETNE;
CCs[RTLIB::UNE_F64] = ISD::SETNE;
CCs[RTLIB::OGE_F32] = ISD::SETGE;
CCs[RTLIB::OGE_F64] = ISD::SETGE;
CCs[RTLIB::OLT_F32] = ISD::SETLT;
CCs[RTLIB::OLT_F64] = ISD::SETLT;
CCs[RTLIB::OLE_F32] = ISD::SETLE;
CCs[RTLIB::OLE_F64] = ISD::SETLE;
CCs[RTLIB::OGT_F32] = ISD::SETGT;
CCs[RTLIB::OGT_F64] = ISD::SETGT;
CCs[RTLIB::UO_F32] = ISD::SETNE;
CCs[RTLIB::UO_F64] = ISD::SETNE;
CCs[RTLIB::O_F32] = ISD::SETEQ;
CCs[RTLIB::O_F64] = ISD::SETEQ;
}
TargetLowering::TargetLowering(TargetMachine &tm)
: TM(tm), TD(TM.getTargetData()) {
// All operations default to being supported.
memset(OpActions, 0, sizeof(OpActions));
memset(LoadExtActions, 0, sizeof(LoadExtActions));
memset(TruncStoreActions, 0, sizeof(TruncStoreActions));
memset(IndexedModeActions, 0, sizeof(IndexedModeActions));
memset(ConvertActions, 0, sizeof(ConvertActions));
memset(CondCodeActions, 0, sizeof(CondCodeActions));
// Set default actions for various operations.
for (unsigned VT = 0; VT != (unsigned)MVT::LAST_VALUETYPE; ++VT) {
// Default all indexed load / store to expand.
for (unsigned IM = (unsigned)ISD::PRE_INC;
IM != (unsigned)ISD::LAST_INDEXED_MODE; ++IM) {
setIndexedLoadAction(IM, (MVT::SimpleValueType)VT, Expand);
setIndexedStoreAction(IM, (MVT::SimpleValueType)VT, Expand);
}
// These operations default to expand.
setOperationAction(ISD::FGETSIGN, (MVT::SimpleValueType)VT, Expand);
}
// Most targets ignore the @llvm.prefetch intrinsic.
setOperationAction(ISD::PREFETCH, MVT::Other, Expand);
// ConstantFP nodes default to expand. Targets can either change this to
// Legal, in which case all fp constants are legal, or use addLegalFPImmediate
// to optimize expansions for certain constants.
setOperationAction(ISD::ConstantFP, MVT::f32, Expand);
setOperationAction(ISD::ConstantFP, MVT::f64, Expand);
setOperationAction(ISD::ConstantFP, MVT::f80, Expand);
// These library functions default to expand.
setOperationAction(ISD::FLOG , MVT::f64, Expand);
setOperationAction(ISD::FLOG2, MVT::f64, Expand);
setOperationAction(ISD::FLOG10,MVT::f64, Expand);
setOperationAction(ISD::FEXP , MVT::f64, Expand);
setOperationAction(ISD::FEXP2, MVT::f64, Expand);
setOperationAction(ISD::FLOG , MVT::f32, Expand);
setOperationAction(ISD::FLOG2, MVT::f32, Expand);
setOperationAction(ISD::FLOG10,MVT::f32, Expand);
setOperationAction(ISD::FEXP , MVT::f32, Expand);
setOperationAction(ISD::FEXP2, MVT::f32, Expand);
// Default ISD::TRAP to expand (which turns it into abort).
setOperationAction(ISD::TRAP, MVT::Other, Expand);
IsLittleEndian = TD->isLittleEndian();
UsesGlobalOffsetTable = false;
ShiftAmountTy = PointerTy = getValueType(TD->getIntPtrType());
ShiftAmtHandling = Undefined;
memset(RegClassForVT, 0,MVT::LAST_VALUETYPE*sizeof(TargetRegisterClass*));
memset(TargetDAGCombineArray, 0, array_lengthof(TargetDAGCombineArray));
maxStoresPerMemset = maxStoresPerMemcpy = maxStoresPerMemmove = 8;
allowUnalignedMemoryAccesses = false;
UseUnderscoreSetJmp = false;
UseUnderscoreLongJmp = false;
SelectIsExpensive = false;
IntDivIsCheap = false;
Pow2DivIsCheap = false;
StackPointerRegisterToSaveRestore = 0;
ExceptionPointerRegister = 0;
ExceptionSelectorRegister = 0;
BooleanContents = UndefinedBooleanContent;
SchedPreferenceInfo = SchedulingForLatency;
JumpBufSize = 0;
JumpBufAlignment = 0;
IfCvtBlockSizeLimit = 2;
IfCvtDupBlockSizeLimit = 0;
PrefLoopAlignment = 0;
InitLibcallNames(LibcallRoutineNames);
InitCmpLibcallCCs(CmpLibcallCCs);
// Tell Legalize whether the assembler supports DEBUG_LOC.
const TargetAsmInfo *TASM = TM.getTargetAsmInfo();
if (!TASM || !TASM->hasDotLocAndDotFile())
setOperationAction(ISD::DEBUG_LOC, MVT::Other, Expand);
}
TargetLowering::~TargetLowering() {}
/// computeRegisterProperties - Once all of the register classes are added,
/// this allows us to compute derived properties we expose.
void TargetLowering::computeRegisterProperties() {
assert(MVT::LAST_VALUETYPE <= 32 &&
"Too many value types for ValueTypeActions to hold!");
// Everything defaults to needing one register.
for (unsigned i = 0; i != MVT::LAST_VALUETYPE; ++i) {
NumRegistersForVT[i] = 1;
RegisterTypeForVT[i] = TransformToType[i] = (MVT::SimpleValueType)i;
}
// ...except isVoid, which doesn't need any registers.
NumRegistersForVT[MVT::isVoid] = 0;
// Find the largest integer register class.
unsigned LargestIntReg = MVT::LAST_INTEGER_VALUETYPE;
for (; RegClassForVT[LargestIntReg] == 0; --LargestIntReg)
assert(LargestIntReg != MVT::i1 && "No integer registers defined!");
// Every integer value type larger than this largest register takes twice as
// many registers to represent as the previous ValueType.
for (unsigned ExpandedReg = LargestIntReg + 1; ; ++ExpandedReg) {
MVT EVT = (MVT::SimpleValueType)ExpandedReg;
if (!EVT.isInteger())
break;
NumRegistersForVT[ExpandedReg] = 2*NumRegistersForVT[ExpandedReg-1];
RegisterTypeForVT[ExpandedReg] = (MVT::SimpleValueType)LargestIntReg;
TransformToType[ExpandedReg] = (MVT::SimpleValueType)(ExpandedReg - 1);
ValueTypeActions.setTypeAction(EVT, Expand);
}
// Inspect all of the ValueType's smaller than the largest integer
// register to see which ones need promotion.
unsigned LegalIntReg = LargestIntReg;
for (unsigned IntReg = LargestIntReg - 1;
IntReg >= (unsigned)MVT::i1; --IntReg) {
MVT IVT = (MVT::SimpleValueType)IntReg;
if (isTypeLegal(IVT)) {
LegalIntReg = IntReg;
} else {
RegisterTypeForVT[IntReg] = TransformToType[IntReg] =
(MVT::SimpleValueType)LegalIntReg;
ValueTypeActions.setTypeAction(IVT, Promote);
}
}
// ppcf128 type is really two f64's.
if (!isTypeLegal(MVT::ppcf128)) {
NumRegistersForVT[MVT::ppcf128] = 2*NumRegistersForVT[MVT::f64];
RegisterTypeForVT[MVT::ppcf128] = MVT::f64;
TransformToType[MVT::ppcf128] = MVT::f64;
ValueTypeActions.setTypeAction(MVT::ppcf128, Expand);
}
// Decide how to handle f64. If the target does not have native f64 support,
// expand it to i64 and we will be generating soft float library calls.
if (!isTypeLegal(MVT::f64)) {
NumRegistersForVT[MVT::f64] = NumRegistersForVT[MVT::i64];
RegisterTypeForVT[MVT::f64] = RegisterTypeForVT[MVT::i64];
TransformToType[MVT::f64] = MVT::i64;
ValueTypeActions.setTypeAction(MVT::f64, Expand);
}
// Decide how to handle f32. If the target does not have native support for
// f32, promote it to f64 if it is legal. Otherwise, expand it to i32.
if (!isTypeLegal(MVT::f32)) {
if (isTypeLegal(MVT::f64)) {
NumRegistersForVT[MVT::f32] = NumRegistersForVT[MVT::f64];
RegisterTypeForVT[MVT::f32] = RegisterTypeForVT[MVT::f64];
TransformToType[MVT::f32] = MVT::f64;
ValueTypeActions.setTypeAction(MVT::f32, Promote);
} else {
NumRegistersForVT[MVT::f32] = NumRegistersForVT[MVT::i32];
RegisterTypeForVT[MVT::f32] = RegisterTypeForVT[MVT::i32];
TransformToType[MVT::f32] = MVT::i32;
ValueTypeActions.setTypeAction(MVT::f32, Expand);
}
}
// Loop over all of the vector value types to see which need transformations.
for (unsigned i = MVT::FIRST_VECTOR_VALUETYPE;
i <= (unsigned)MVT::LAST_VECTOR_VALUETYPE; ++i) {
MVT VT = (MVT::SimpleValueType)i;
if (!isTypeLegal(VT)) {
MVT IntermediateVT, RegisterVT;
unsigned NumIntermediates;
NumRegistersForVT[i] =
getVectorTypeBreakdown(VT,
IntermediateVT, NumIntermediates,
RegisterVT);
RegisterTypeForVT[i] = RegisterVT;
// Determine if there is a legal wider type.
bool IsLegalWiderType = false;
MVT EltVT = VT.getVectorElementType();
unsigned NElts = VT.getVectorNumElements();
for (unsigned nVT = i+1; nVT <= MVT::LAST_VECTOR_VALUETYPE; ++nVT) {
MVT SVT = (MVT::SimpleValueType)nVT;
if (isTypeLegal(SVT) && SVT.getVectorElementType() == EltVT &&
SVT.getVectorNumElements() > NElts) {
TransformToType[i] = SVT;
ValueTypeActions.setTypeAction(VT, Promote);
IsLegalWiderType = true;
break;
}
}
if (!IsLegalWiderType) {
MVT NVT = VT.getPow2VectorType();
if (NVT == VT) {
// Type is already a power of 2. The default action is to split.
TransformToType[i] = MVT::Other;
ValueTypeActions.setTypeAction(VT, Expand);
} else {
TransformToType[i] = NVT;
ValueTypeActions.setTypeAction(VT, Promote);
}
}
}
}
}
const char *TargetLowering::getTargetNodeName(unsigned Opcode) const {
return NULL;
}
MVT TargetLowering::getSetCCResultType(MVT VT) const {
return getValueType(TD->getIntPtrType());
}
/// getVectorTypeBreakdown - Vector types are broken down into some number of
/// legal first class types. For example, MVT::v8f32 maps to 2 MVT::v4f32
/// with Altivec or SSE1, or 8 promoted MVT::f64 values with the X86 FP stack.
/// Similarly, MVT::v2i64 turns into 4 MVT::i32 values with both PPC and X86.
///
/// This method returns the number of registers needed, and the VT for each
/// register. It also returns the VT and quantity of the intermediate values
/// before they are promoted/expanded.
///
unsigned TargetLowering::getVectorTypeBreakdown(MVT VT,
MVT &IntermediateVT,
unsigned &NumIntermediates,
MVT &RegisterVT) const {
// Figure out the right, legal destination reg to copy into.
unsigned NumElts = VT.getVectorNumElements();
MVT EltTy = VT.getVectorElementType();
unsigned NumVectorRegs = 1;
// FIXME: We don't support non-power-of-2-sized vectors for now. Ideally we
// could break down into LHS/RHS like LegalizeDAG does.
if (!isPowerOf2_32(NumElts)) {
NumVectorRegs = NumElts;
NumElts = 1;
}
// Divide the input until we get to a supported size. This will always
// end with a scalar if the target doesn't support vectors.
while (NumElts > 1 && !isTypeLegal(MVT::getVectorVT(EltTy, NumElts))) {
NumElts >>= 1;
NumVectorRegs <<= 1;
}
NumIntermediates = NumVectorRegs;
MVT NewVT = MVT::getVectorVT(EltTy, NumElts);
if (!isTypeLegal(NewVT))
NewVT = EltTy;
IntermediateVT = NewVT;
MVT DestVT = getRegisterType(NewVT);
RegisterVT = DestVT;
if (DestVT.bitsLT(NewVT)) {
// Value is expanded, e.g. i64 -> i16.
return NumVectorRegs*(NewVT.getSizeInBits()/DestVT.getSizeInBits());
} else {
// Otherwise, promotion or legal types use the same number of registers as
// the vector decimated to the appropriate level.
return NumVectorRegs;
}
return 1;
}
/// getWidenVectorType: given a vector type, returns the type to widen to
/// (e.g., v7i8 to v8i8). If the vector type is legal, it returns itself.
/// If there is no vector type that we want to widen to, returns MVT::Other
/// When and where to widen is target dependent based on the cost of
/// scalarizing vs using the wider vector type.
MVT TargetLowering::getWidenVectorType(MVT VT) const {
assert(VT.isVector());
if (isTypeLegal(VT))
return VT;
// Default is not to widen until moved to LegalizeTypes
return MVT::Other;
}
/// getByValTypeAlignment - Return the desired alignment for ByVal aggregate
/// function arguments in the caller parameter area. This is the actual
/// alignment, not its logarithm.
unsigned TargetLowering::getByValTypeAlignment(const Type *Ty) const {
return TD->getCallFrameTypeAlignment(Ty);
}
SDValue TargetLowering::getPICJumpTableRelocBase(SDValue Table,
SelectionDAG &DAG) const {
if (usesGlobalOffsetTable())
return DAG.getGLOBAL_OFFSET_TABLE(getPointerTy());
return Table;
}
bool
TargetLowering::isOffsetFoldingLegal(const GlobalAddressSDNode *GA) const {
// Assume that everything is safe in static mode.
if (getTargetMachine().getRelocationModel() == Reloc::Static)
return true;
// In dynamic-no-pic mode, assume that known defined values are safe.
if (getTargetMachine().getRelocationModel() == Reloc::DynamicNoPIC &&
GA &&
!GA->getGlobal()->isDeclaration() &&
!GA->getGlobal()->isWeakForLinker())
return true;
// Otherwise assume nothing is safe.
return false;
}
//===----------------------------------------------------------------------===//
// Optimization Methods
//===----------------------------------------------------------------------===//
/// ShrinkDemandedConstant - Check to see if the specified operand of the
/// specified instruction is a constant integer. If so, check to see if there
/// are any bits set in the constant that are not demanded. If so, shrink the
/// constant and return true.
bool TargetLowering::TargetLoweringOpt::ShrinkDemandedConstant(SDValue Op,
const APInt &Demanded) {
DebugLoc dl = Op.getDebugLoc();
// FIXME: ISD::SELECT, ISD::SELECT_CC
switch (Op.getOpcode()) {
default: break;
case ISD::XOR:
case ISD::AND:
case ISD::OR: {
ConstantSDNode *C = dyn_cast<ConstantSDNode>(Op.getOperand(1));
if (!C) return false;
if (Op.getOpcode() == ISD::XOR &&
(C->getAPIntValue() | (~Demanded)).isAllOnesValue())
return false;
// if we can expand it to have all bits set, do it
if (C->getAPIntValue().intersects(~Demanded)) {
MVT VT = Op.getValueType();
SDValue New = DAG.getNode(Op.getOpcode(), dl, VT, Op.getOperand(0),
DAG.getConstant(Demanded &
C->getAPIntValue(),
VT));
return CombineTo(Op, New);
}
break;
}
}
return false;
}
/// ShrinkDemandedOp - Convert x+y to (VT)((SmallVT)x+(SmallVT)y) if the
/// casts are free. This uses isZExtFree and ZERO_EXTEND for the widening
/// cast, but it could be generalized for targets with other types of
/// implicit widening casts.
bool
TargetLowering::TargetLoweringOpt::ShrinkDemandedOp(SDValue Op,
unsigned BitWidth,
const APInt &Demanded,
DebugLoc dl) {
assert(Op.getNumOperands() == 2 &&
"ShrinkDemandedOp only supports binary operators!");
assert(Op.getNode()->getNumValues() == 1 &&
"ShrinkDemandedOp only supports nodes with one result!");
// Don't do this if the node has another user, which may require the
// full value.
if (!Op.getNode()->hasOneUse())
return false;
// Search for the smallest integer type with free casts to and from
// Op's type. For expedience, just check power-of-2 integer types.
const TargetLowering &TLI = DAG.getTargetLoweringInfo();
unsigned SmallVTBits = BitWidth - Demanded.countLeadingZeros();
if (!isPowerOf2_32(SmallVTBits))
SmallVTBits = NextPowerOf2(SmallVTBits);
for (; SmallVTBits < BitWidth; SmallVTBits = NextPowerOf2(SmallVTBits)) {
MVT SmallVT = MVT::getIntegerVT(SmallVTBits);
if (TLI.isTruncateFree(Op.getValueType(), SmallVT) &&
TLI.isZExtFree(SmallVT, Op.getValueType())) {
// We found a type with free casts.
SDValue X = DAG.getNode(Op.getOpcode(), dl, SmallVT,
DAG.getNode(ISD::TRUNCATE, dl, SmallVT,
Op.getNode()->getOperand(0)),
DAG.getNode(ISD::TRUNCATE, dl, SmallVT,
Op.getNode()->getOperand(1)));
SDValue Z = DAG.getNode(ISD::ZERO_EXTEND, dl, Op.getValueType(), X);
return CombineTo(Op, Z);
}
}
return false;
}
/// SimplifyDemandedBits - Look at Op. At this point, we know that only the
/// DemandedMask bits of the result of Op are ever used downstream. If we can
/// use this information to simplify Op, create a new simplified DAG node and
/// return true, returning the original and new nodes in Old and New. Otherwise,
/// analyze the expression and return a mask of KnownOne and KnownZero bits for
/// the expression (used to simplify the caller). The KnownZero/One bits may
/// only be accurate for those bits in the DemandedMask.
bool TargetLowering::SimplifyDemandedBits(SDValue Op,
const APInt &DemandedMask,
APInt &KnownZero,
APInt &KnownOne,
TargetLoweringOpt &TLO,
unsigned Depth) const {
unsigned BitWidth = DemandedMask.getBitWidth();
assert(Op.getValueSizeInBits() == BitWidth &&
"Mask size mismatches value type size!");
APInt NewMask = DemandedMask;
DebugLoc dl = Op.getDebugLoc();
// Don't know anything.
KnownZero = KnownOne = APInt(BitWidth, 0);
// Other users may use these bits.
if (!Op.getNode()->hasOneUse()) {
if (Depth != 0) {
// If not at the root, Just compute the KnownZero/KnownOne bits to
// simplify things downstream.
TLO.DAG.ComputeMaskedBits(Op, DemandedMask, KnownZero, KnownOne, Depth);
return false;
}
// If this is the root being simplified, allow it to have multiple uses,
// just set the NewMask to all bits.
NewMask = APInt::getAllOnesValue(BitWidth);
} else if (DemandedMask == 0) {
// Not demanding any bits from Op.
if (Op.getOpcode() != ISD::UNDEF)
return TLO.CombineTo(Op, TLO.DAG.getUNDEF(Op.getValueType()));
return false;
} else if (Depth == 6) { // Limit search depth.
return false;
}
APInt KnownZero2, KnownOne2, KnownZeroOut, KnownOneOut;
switch (Op.getOpcode()) {
case ISD::Constant:
// We know all of the bits for a constant!
KnownOne = cast<ConstantSDNode>(Op)->getAPIntValue() & NewMask;
KnownZero = ~KnownOne & NewMask;
return false; // Don't fall through, will infinitely loop.
case ISD::AND:
// If the RHS is a constant, check to see if the LHS would be zero without
// using the bits from the RHS. Below, we use knowledge about the RHS to
// simplify the LHS, here we're using information from the LHS to simplify
// the RHS.
if (ConstantSDNode *RHSC = dyn_cast<ConstantSDNode>(Op.getOperand(1))) {
APInt LHSZero, LHSOne;
TLO.DAG.ComputeMaskedBits(Op.getOperand(0), NewMask,
LHSZero, LHSOne, Depth+1);
// If the LHS already has zeros where RHSC does, this and is dead.
if ((LHSZero & NewMask) == (~RHSC->getAPIntValue() & NewMask))
return TLO.CombineTo(Op, Op.getOperand(0));
// If any of the set bits in the RHS are known zero on the LHS, shrink
// the constant.
if (TLO.ShrinkDemandedConstant(Op, ~LHSZero & NewMask))
return true;
}
if (SimplifyDemandedBits(Op.getOperand(1), NewMask, KnownZero,
KnownOne, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
if (SimplifyDemandedBits(Op.getOperand(0), ~KnownZero & NewMask,
KnownZero2, KnownOne2, TLO, Depth+1))
return true;
assert((KnownZero2 & KnownOne2) == 0 && "Bits known to be one AND zero?");
// If all of the demanded bits are known one on one side, return the other.
// These bits cannot contribute to the result of the 'and'.
if ((NewMask & ~KnownZero2 & KnownOne) == (~KnownZero2 & NewMask))
return TLO.CombineTo(Op, Op.getOperand(0));
if ((NewMask & ~KnownZero & KnownOne2) == (~KnownZero & NewMask))
return TLO.CombineTo(Op, Op.getOperand(1));
// If all of the demanded bits in the inputs are known zeros, return zero.
if ((NewMask & (KnownZero|KnownZero2)) == NewMask)
return TLO.CombineTo(Op, TLO.DAG.getConstant(0, Op.getValueType()));
// If the RHS is a constant, see if we can simplify it.
if (TLO.ShrinkDemandedConstant(Op, ~KnownZero2 & NewMask))
return true;
// If the operation can be done in a smaller type, do so.
if (TLO.ShrinkDemandedOp(Op, BitWidth, NewMask, dl))
return true;
// Output known-1 bits are only known if set in both the LHS & RHS.
KnownOne &= KnownOne2;
// Output known-0 are known to be clear if zero in either the LHS | RHS.
KnownZero |= KnownZero2;
break;
case ISD::OR:
if (SimplifyDemandedBits(Op.getOperand(1), NewMask, KnownZero,
KnownOne, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
if (SimplifyDemandedBits(Op.getOperand(0), ~KnownOne & NewMask,
KnownZero2, KnownOne2, TLO, Depth+1))
return true;
assert((KnownZero2 & KnownOne2) == 0 && "Bits known to be one AND zero?");
// If all of the demanded bits are known zero on one side, return the other.
// These bits cannot contribute to the result of the 'or'.
if ((NewMask & ~KnownOne2 & KnownZero) == (~KnownOne2 & NewMask))
return TLO.CombineTo(Op, Op.getOperand(0));
if ((NewMask & ~KnownOne & KnownZero2) == (~KnownOne & NewMask))
return TLO.CombineTo(Op, Op.getOperand(1));
// If all of the potentially set bits on one side are known to be set on
// the other side, just use the 'other' side.
if ((NewMask & ~KnownZero & KnownOne2) == (~KnownZero & NewMask))
return TLO.CombineTo(Op, Op.getOperand(0));
if ((NewMask & ~KnownZero2 & KnownOne) == (~KnownZero2 & NewMask))
return TLO.CombineTo(Op, Op.getOperand(1));
// If the RHS is a constant, see if we can simplify it.
if (TLO.ShrinkDemandedConstant(Op, NewMask))
return true;
// If the operation can be done in a smaller type, do so.
if (TLO.ShrinkDemandedOp(Op, BitWidth, NewMask, dl))
return true;
// Output known-0 bits are only known if clear in both the LHS & RHS.
KnownZero &= KnownZero2;
// Output known-1 are known to be set if set in either the LHS | RHS.
KnownOne |= KnownOne2;
break;
case ISD::XOR:
if (SimplifyDemandedBits(Op.getOperand(1), NewMask, KnownZero,
KnownOne, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
if (SimplifyDemandedBits(Op.getOperand(0), NewMask, KnownZero2,
KnownOne2, TLO, Depth+1))
return true;
assert((KnownZero2 & KnownOne2) == 0 && "Bits known to be one AND zero?");
// If all of the demanded bits are known zero on one side, return the other.
// These bits cannot contribute to the result of the 'xor'.
if ((KnownZero & NewMask) == NewMask)
return TLO.CombineTo(Op, Op.getOperand(0));
if ((KnownZero2 & NewMask) == NewMask)
return TLO.CombineTo(Op, Op.getOperand(1));
// If the operation can be done in a smaller type, do so.
if (TLO.ShrinkDemandedOp(Op, BitWidth, NewMask, dl))
return true;
// If all of the unknown bits are known to be zero on one side or the other
// (but not both) turn this into an *inclusive* or.
// e.g. (A & C1)^(B & C2) -> (A & C1)|(B & C2) iff C1&C2 == 0
if ((NewMask & ~KnownZero & ~KnownZero2) == 0)
return TLO.CombineTo(Op, TLO.DAG.getNode(ISD::OR, dl, Op.getValueType(),
Op.getOperand(0),
Op.getOperand(1)));
// Output known-0 bits are known if clear or set in both the LHS & RHS.
KnownZeroOut = (KnownZero & KnownZero2) | (KnownOne & KnownOne2);
// Output known-1 are known to be set if set in only one of the LHS, RHS.
KnownOneOut = (KnownZero & KnownOne2) | (KnownOne & KnownZero2);
// If all of the demanded bits on one side are known, and all of the set
// bits on that side are also known to be set on the other side, turn this
// into an AND, as we know the bits will be cleared.
// e.g. (X | C1) ^ C2 --> (X | C1) & ~C2 iff (C1&C2) == C2
if ((NewMask & (KnownZero|KnownOne)) == NewMask) { // all known
if ((KnownOne & KnownOne2) == KnownOne) {
MVT VT = Op.getValueType();
SDValue ANDC = TLO.DAG.getConstant(~KnownOne & NewMask, VT);
return TLO.CombineTo(Op, TLO.DAG.getNode(ISD::AND, dl, VT,
Op.getOperand(0), ANDC));
}
}
// If the RHS is a constant, see if we can simplify it.
// for XOR, we prefer to force bits to 1 if they will make a -1.
// if we can't force bits, try to shrink constant
if (ConstantSDNode *C = dyn_cast<ConstantSDNode>(Op.getOperand(1))) {
APInt Expanded = C->getAPIntValue() | (~NewMask);
// if we can expand it to have all bits set, do it
if (Expanded.isAllOnesValue()) {
if (Expanded != C->getAPIntValue()) {
MVT VT = Op.getValueType();
SDValue New = TLO.DAG.getNode(Op.getOpcode(), dl,VT, Op.getOperand(0),
TLO.DAG.getConstant(Expanded, VT));
return TLO.CombineTo(Op, New);
}
// if it already has all the bits set, nothing to change
// but don't shrink either!
} else if (TLO.ShrinkDemandedConstant(Op, NewMask)) {
return true;
}
}
KnownZero = KnownZeroOut;
KnownOne = KnownOneOut;
break;
case ISD::SELECT:
if (SimplifyDemandedBits(Op.getOperand(2), NewMask, KnownZero,
KnownOne, TLO, Depth+1))
return true;
if (SimplifyDemandedBits(Op.getOperand(1), NewMask, KnownZero2,
KnownOne2, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
assert((KnownZero2 & KnownOne2) == 0 && "Bits known to be one AND zero?");
// If the operands are constants, see if we can simplify them.
if (TLO.ShrinkDemandedConstant(Op, NewMask))
return true;
// Only known if known in both the LHS and RHS.
KnownOne &= KnownOne2;
KnownZero &= KnownZero2;
break;
case ISD::SELECT_CC:
if (SimplifyDemandedBits(Op.getOperand(3), NewMask, KnownZero,
KnownOne, TLO, Depth+1))
return true;
if (SimplifyDemandedBits(Op.getOperand(2), NewMask, KnownZero2,
KnownOne2, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
assert((KnownZero2 & KnownOne2) == 0 && "Bits known to be one AND zero?");
// If the operands are constants, see if we can simplify them.
if (TLO.ShrinkDemandedConstant(Op, NewMask))
return true;
// Only known if known in both the LHS and RHS.
KnownOne &= KnownOne2;
KnownZero &= KnownZero2;
break;
case ISD::SHL:
if (ConstantSDNode *SA = dyn_cast<ConstantSDNode>(Op.getOperand(1))) {
unsigned ShAmt = SA->getZExtValue();
SDValue InOp = Op.getOperand(0);
// If the shift count is an invalid immediate, don't do anything.
if (ShAmt >= BitWidth)
break;
// If this is ((X >>u C1) << ShAmt), see if we can simplify this into a
// single shift. We can do this if the bottom bits (which are shifted
// out) are never demanded.
if (InOp.getOpcode() == ISD::SRL &&
isa<ConstantSDNode>(InOp.getOperand(1))) {
if (ShAmt && (NewMask & APInt::getLowBitsSet(BitWidth, ShAmt)) == 0) {
unsigned C1= cast<ConstantSDNode>(InOp.getOperand(1))->getZExtValue();
unsigned Opc = ISD::SHL;
int Diff = ShAmt-C1;
if (Diff < 0) {
Diff = -Diff;
Opc = ISD::SRL;
}
SDValue NewSA =
TLO.DAG.getConstant(Diff, Op.getOperand(1).getValueType());
MVT VT = Op.getValueType();
return TLO.CombineTo(Op, TLO.DAG.getNode(Opc, dl, VT,
InOp.getOperand(0), NewSA));
}
}
if (SimplifyDemandedBits(Op.getOperand(0), NewMask.lshr(ShAmt),
KnownZero, KnownOne, TLO, Depth+1))
return true;
KnownZero <<= SA->getZExtValue();
KnownOne <<= SA->getZExtValue();
// low bits known zero.
KnownZero |= APInt::getLowBitsSet(BitWidth, SA->getZExtValue());
}
break;
case ISD::SRL:
if (ConstantSDNode *SA = dyn_cast<ConstantSDNode>(Op.getOperand(1))) {
MVT VT = Op.getValueType();
unsigned ShAmt = SA->getZExtValue();
unsigned VTSize = VT.getSizeInBits();
SDValue InOp = Op.getOperand(0);
// If the shift count is an invalid immediate, don't do anything.
if (ShAmt >= BitWidth)
break;
// If this is ((X << C1) >>u ShAmt), see if we can simplify this into a
// single shift. We can do this if the top bits (which are shifted out)
// are never demanded.
if (InOp.getOpcode() == ISD::SHL &&
isa<ConstantSDNode>(InOp.getOperand(1))) {
if (ShAmt && (NewMask & APInt::getHighBitsSet(VTSize, ShAmt)) == 0) {
unsigned C1= cast<ConstantSDNode>(InOp.getOperand(1))->getZExtValue();
unsigned Opc = ISD::SRL;
int Diff = ShAmt-C1;
if (Diff < 0) {
Diff = -Diff;
Opc = ISD::SHL;
}
SDValue NewSA =
TLO.DAG.getConstant(Diff, Op.getOperand(1).getValueType());
return TLO.CombineTo(Op, TLO.DAG.getNode(Opc, dl, VT,
InOp.getOperand(0), NewSA));
}
}
// Compute the new bits that are at the top now.
if (SimplifyDemandedBits(InOp, (NewMask << ShAmt),
KnownZero, KnownOne, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
KnownZero = KnownZero.lshr(ShAmt);
KnownOne = KnownOne.lshr(ShAmt);
APInt HighBits = APInt::getHighBitsSet(BitWidth, ShAmt);
KnownZero |= HighBits; // High bits known zero.
}
break;
case ISD::SRA:
// If this is an arithmetic shift right and only the low-bit is set, we can
// always convert this into a logical shr, even if the shift amount is
// variable. The low bit of the shift cannot be an input sign bit unless
// the shift amount is >= the size of the datatype, which is undefined.
if (DemandedMask == 1)
return TLO.CombineTo(Op, TLO.DAG.getNode(ISD::SRL, dl, Op.getValueType(),
Op.getOperand(0), Op.getOperand(1)));
if (ConstantSDNode *SA = dyn_cast<ConstantSDNode>(Op.getOperand(1))) {
MVT VT = Op.getValueType();
unsigned ShAmt = SA->getZExtValue();
// If the shift count is an invalid immediate, don't do anything.
if (ShAmt >= BitWidth)
break;
APInt InDemandedMask = (NewMask << ShAmt);
// If any of the demanded bits are produced by the sign extension, we also
// demand the input sign bit.
APInt HighBits = APInt::getHighBitsSet(BitWidth, ShAmt);
if (HighBits.intersects(NewMask))
InDemandedMask |= APInt::getSignBit(VT.getSizeInBits());
if (SimplifyDemandedBits(Op.getOperand(0), InDemandedMask,
KnownZero, KnownOne, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
KnownZero = KnownZero.lshr(ShAmt);
KnownOne = KnownOne.lshr(ShAmt);
// Handle the sign bit, adjusted to where it is now in the mask.
APInt SignBit = APInt::getSignBit(BitWidth).lshr(ShAmt);
// If the input sign bit is known to be zero, or if none of the top bits
// are demanded, turn this into an unsigned shift right.
if (KnownZero.intersects(SignBit) || (HighBits & ~NewMask) == HighBits) {
return TLO.CombineTo(Op, TLO.DAG.getNode(ISD::SRL, dl, VT,
Op.getOperand(0),
Op.getOperand(1)));
} else if (KnownOne.intersects(SignBit)) { // New bits are known one.
KnownOne |= HighBits;
}
}
break;
case ISD::SIGN_EXTEND_INREG: {
MVT EVT = cast<VTSDNode>(Op.getOperand(1))->getVT();
// Sign extension. Compute the demanded bits in the result that are not
// present in the input.
APInt NewBits = APInt::getHighBitsSet(BitWidth,
BitWidth - EVT.getSizeInBits()) &
NewMask;
// If none of the extended bits are demanded, eliminate the sextinreg.
if (NewBits == 0)
return TLO.CombineTo(Op, Op.getOperand(0));
APInt InSignBit = APInt::getSignBit(EVT.getSizeInBits());
InSignBit.zext(BitWidth);
APInt InputDemandedBits = APInt::getLowBitsSet(BitWidth,
EVT.getSizeInBits()) &
NewMask;
// Since the sign extended bits are demanded, we know that the sign
// bit is demanded.
InputDemandedBits |= InSignBit;
if (SimplifyDemandedBits(Op.getOperand(0), InputDemandedBits,
KnownZero, KnownOne, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
// If the sign bit of the input is known set or clear, then we know the
// top bits of the result.
// If the input sign bit is known zero, convert this into a zero extension.
if (KnownZero.intersects(InSignBit))
return TLO.CombineTo(Op,
TLO.DAG.getZeroExtendInReg(Op.getOperand(0),dl,EVT));
if (KnownOne.intersects(InSignBit)) { // Input sign bit known set
KnownOne |= NewBits;
KnownZero &= ~NewBits;
} else { // Input sign bit unknown
KnownZero &= ~NewBits;
KnownOne &= ~NewBits;
}
break;
}
case ISD::ZERO_EXTEND: {
unsigned OperandBitWidth = Op.getOperand(0).getValueSizeInBits();
APInt InMask = NewMask;
InMask.trunc(OperandBitWidth);
// If none of the top bits are demanded, convert this into an any_extend.
APInt NewBits =
APInt::getHighBitsSet(BitWidth, BitWidth - OperandBitWidth) & NewMask;
if (!NewBits.intersects(NewMask))
return TLO.CombineTo(Op, TLO.DAG.getNode(ISD::ANY_EXTEND, dl,
Op.getValueType(),
Op.getOperand(0)));
if (SimplifyDemandedBits(Op.getOperand(0), InMask,
KnownZero, KnownOne, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
KnownZero.zext(BitWidth);
KnownOne.zext(BitWidth);
KnownZero |= NewBits;
break;
}
case ISD::SIGN_EXTEND: {
MVT InVT = Op.getOperand(0).getValueType();
unsigned InBits = InVT.getSizeInBits();
APInt InMask = APInt::getLowBitsSet(BitWidth, InBits);
APInt InSignBit = APInt::getBitsSet(BitWidth, InBits - 1, InBits);
APInt NewBits = ~InMask & NewMask;
// If none of the top bits are demanded, convert this into an any_extend.
if (NewBits == 0)
return TLO.CombineTo(Op,TLO.DAG.getNode(ISD::ANY_EXTEND, dl,
Op.getValueType(),
Op.getOperand(0)));
// Since some of the sign extended bits are demanded, we know that the sign
// bit is demanded.
APInt InDemandedBits = InMask & NewMask;
InDemandedBits |= InSignBit;
InDemandedBits.trunc(InBits);
if (SimplifyDemandedBits(Op.getOperand(0), InDemandedBits, KnownZero,
KnownOne, TLO, Depth+1))
return true;
KnownZero.zext(BitWidth);
KnownOne.zext(BitWidth);
// If the sign bit is known zero, convert this to a zero extend.
if (KnownZero.intersects(InSignBit))
return TLO.CombineTo(Op, TLO.DAG.getNode(ISD::ZERO_EXTEND, dl,
Op.getValueType(),
Op.getOperand(0)));
// If the sign bit is known one, the top bits match.
if (KnownOne.intersects(InSignBit)) {
KnownOne |= NewBits;
KnownZero &= ~NewBits;
} else { // Otherwise, top bits aren't known.
KnownOne &= ~NewBits;
KnownZero &= ~NewBits;
}
break;
}
case ISD::ANY_EXTEND: {
unsigned OperandBitWidth = Op.getOperand(0).getValueSizeInBits();
APInt InMask = NewMask;
InMask.trunc(OperandBitWidth);
if (SimplifyDemandedBits(Op.getOperand(0), InMask,
KnownZero, KnownOne, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
KnownZero.zext(BitWidth);
KnownOne.zext(BitWidth);
break;
}
case ISD::TRUNCATE: {
// Simplify the input, using demanded bit information, and compute the known
// zero/one bits live out.
APInt TruncMask = NewMask;
TruncMask.zext(Op.getOperand(0).getValueSizeInBits());
if (SimplifyDemandedBits(Op.getOperand(0), TruncMask,
KnownZero, KnownOne, TLO, Depth+1))
return true;
KnownZero.trunc(BitWidth);
KnownOne.trunc(BitWidth);
// If the input is only used by this truncate, see if we can shrink it based
// on the known demanded bits.
if (Op.getOperand(0).getNode()->hasOneUse()) {
SDValue In = Op.getOperand(0);
unsigned InBitWidth = In.getValueSizeInBits();
switch (In.getOpcode()) {
default: break;
case ISD::SRL:
// Shrink SRL by a constant if none of the high bits shifted in are
// demanded.
if (ConstantSDNode *ShAmt = dyn_cast<ConstantSDNode>(In.getOperand(1))){
APInt HighBits = APInt::getHighBitsSet(InBitWidth,
InBitWidth - BitWidth);
HighBits = HighBits.lshr(ShAmt->getZExtValue());
HighBits.trunc(BitWidth);
if (ShAmt->getZExtValue() < BitWidth && !(HighBits & NewMask)) {
// None of the shifted in bits are needed. Add a truncate of the
// shift input, then shift it.
SDValue NewTrunc = TLO.DAG.getNode(ISD::TRUNCATE, dl,
Op.getValueType(),
In.getOperand(0));
return TLO.CombineTo(Op, TLO.DAG.getNode(ISD::SRL, dl,
Op.getValueType(),
NewTrunc,
In.getOperand(1)));
}
}
break;
}
}
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
break;
}
case ISD::AssertZext: {
MVT VT = cast<VTSDNode>(Op.getOperand(1))->getVT();
APInt InMask = APInt::getLowBitsSet(BitWidth,
VT.getSizeInBits());
if (SimplifyDemandedBits(Op.getOperand(0), InMask & NewMask,
KnownZero, KnownOne, TLO, Depth+1))
return true;
assert((KnownZero & KnownOne) == 0 && "Bits known to be one AND zero?");
KnownZero |= ~InMask & NewMask;
break;
}
case ISD::BIT_CONVERT:
#if 0
// If this is an FP->Int bitcast and if the sign bit is the only thing that
// is demanded, turn this into a FGETSIGN.
if (NewMask == MVT::getIntegerVTSignBit(Op.getValueType()) &&
MVT::isFloatingPoint(Op.getOperand(0).getValueType()) &&
!MVT::isVector(Op.getOperand(0).getValueType())) {
// Only do this xform if FGETSIGN is valid or if before legalize.
if (!TLO.AfterLegalize ||
isOperationLegal(ISD::FGETSIGN, Op.getValueType())) {
// Make a FGETSIGN + SHL to move the sign bit into the appropriate
// place. We expect the SHL to be eliminated by other optimizations.
SDValue Sign = TLO.DAG.getNode(ISD::FGETSIGN, Op.getValueType(),
Op.getOperand(0));
unsigned ShVal = Op.getValueType().getSizeInBits()-1;
SDValue ShAmt = TLO.DAG.getConstant(ShVal, getShiftAmountTy());
return TLO.CombineTo(Op, TLO.DAG.getNode(ISD::SHL, Op.getValueType(),
Sign, ShAmt));
}
}
#endif
break;
case ISD::ADD:
case ISD::MUL:
case ISD::SUB: {
// Add, Sub, and Mul don't demand any bits in positions beyond that
// of the highest bit demanded of them.
APInt LoMask = APInt::getLowBitsSet(BitWidth,
BitWidth - NewMask.countLeadingZeros());
if (SimplifyDemandedBits(Op.getOperand(0), LoMask, KnownZero2,
KnownOne2, TLO, Depth+1))
return true;
if (SimplifyDemandedBits(Op.getOperand(1), LoMask, KnownZero2,
KnownOne2, TLO, Depth+1))
return true;
// See if the operation should be performed at a smaller bit width.
if (TLO.ShrinkDemandedOp(Op, BitWidth, NewMask, dl))
return true;
}
// FALL THROUGH
default:
// Just use ComputeMaskedBits to compute output bits.
TLO.DAG.ComputeMaskedBits(Op, NewMask, KnownZero, KnownOne, Depth);
break;
}
// If we know the value of all of the demanded bits, return this as a
// constant.
if ((NewMask & (KnownZero|KnownOne)) == NewMask)
return TLO.CombineTo(Op, TLO.DAG.getConstant(KnownOne, Op.getValueType()));
return false;
}
/// computeMaskedBitsForTargetNode - Determine which of the bits specified
/// in Mask are known to be either zero or one and return them in the
/// KnownZero/KnownOne bitsets.
void TargetLowering::computeMaskedBitsForTargetNode(const SDValue Op,
const APInt &Mask,
APInt &KnownZero,
APInt &KnownOne,
const SelectionDAG &DAG,
unsigned Depth) const {
assert((Op.getOpcode() >= ISD::BUILTIN_OP_END ||
Op.getOpcode() == ISD::INTRINSIC_WO_CHAIN ||
Op.getOpcode() == ISD::INTRINSIC_W_CHAIN ||
Op.getOpcode() == ISD::INTRINSIC_VOID) &&
"Should use MaskedValueIsZero if you don't know whether Op"
" is a target node!");
KnownZero = KnownOne = APInt(Mask.getBitWidth(), 0);
}
/// ComputeNumSignBitsForTargetNode - This method can be implemented by
/// targets that want to expose additional information about sign bits to the
/// DAG Combiner.
unsigned TargetLowering::ComputeNumSignBitsForTargetNode(SDValue Op,
unsigned Depth) const {
assert((Op.getOpcode() >= ISD::BUILTIN_OP_END ||
Op.getOpcode() == ISD::INTRINSIC_WO_CHAIN ||
Op.getOpcode() == ISD::INTRINSIC_W_CHAIN ||
Op.getOpcode() == ISD::INTRINSIC_VOID) &&
"Should use ComputeNumSignBits if you don't know whether Op"
" is a target node!");
return 1;
}
/// ValueHasExactlyOneBitSet - Test if the given value is known to have exactly
/// one bit set. This differs from ComputeMaskedBits in that it doesn't need to
/// determine which bit is set.
///
static bool ValueHasExactlyOneBitSet(SDValue Val, const SelectionDAG &DAG) {
// A left-shift of a constant one will have exactly one bit set, because
// shifting the bit off the end is undefined.
if (Val.getOpcode() == ISD::SHL)
if (ConstantSDNode *C =
dyn_cast<ConstantSDNode>(Val.getNode()->getOperand(0)))
if (C->getAPIntValue() == 1)
return true;
// Similarly, a right-shift of a constant sign-bit will have exactly
// one bit set.
if (Val.getOpcode() == ISD::SRL)
if (ConstantSDNode *C =
dyn_cast<ConstantSDNode>(Val.getNode()->getOperand(0)))
if (C->getAPIntValue().isSignBit())
return true;
// More could be done here, though the above checks are enough
// to handle some common cases.
// Fall back to ComputeMaskedBits to catch other known cases.
MVT OpVT = Val.getValueType();
unsigned BitWidth = OpVT.getSizeInBits();
APInt Mask = APInt::getAllOnesValue(BitWidth);
APInt KnownZero, KnownOne;
DAG.ComputeMaskedBits(Val, Mask, KnownZero, KnownOne);
return (KnownZero.countPopulation() == BitWidth - 1) &&
(KnownOne.countPopulation() == 1);
}
/// SimplifySetCC - Try to simplify a setcc built with the specified operands
/// and cc. If it is unable to simplify it, return a null SDValue.
SDValue
TargetLowering::SimplifySetCC(MVT VT, SDValue N0, SDValue N1,
ISD::CondCode Cond, bool foldBooleans,
DAGCombinerInfo &DCI, DebugLoc dl) const {
SelectionDAG &DAG = DCI.DAG;
// These setcc operations always fold.
switch (Cond) {
default: break;
case ISD::SETFALSE:
case ISD::SETFALSE2: return DAG.getConstant(0, VT);
case ISD::SETTRUE:
case ISD::SETTRUE2: return DAG.getConstant(1, VT);
}
if (ConstantSDNode *N1C = dyn_cast<ConstantSDNode>(N1.getNode())) {
const APInt &C1 = N1C->getAPIntValue();
if (isa<ConstantSDNode>(N0.getNode())) {
return DAG.FoldSetCC(VT, N0, N1, Cond, dl);
} else {
// If the LHS is '(srl (ctlz x), 5)', the RHS is 0/1, and this is an
// equality comparison, then we're just comparing whether X itself is
// zero.
if (N0.getOpcode() == ISD::SRL && (C1 == 0 || C1 == 1) &&
N0.getOperand(0).getOpcode() == ISD::CTLZ &&
N0.getOperand(1).getOpcode() == ISD::Constant) {
unsigned ShAmt = cast<ConstantSDNode>(N0.getOperand(1))->getZExtValue();
if ((Cond == ISD::SETEQ || Cond == ISD::SETNE) &&
ShAmt == Log2_32(N0.getValueType().getSizeInBits())) {
if ((C1 == 0) == (Cond == ISD::SETEQ)) {
// (srl (ctlz x), 5) == 0 -> X != 0
// (srl (ctlz x), 5) != 1 -> X != 0
Cond = ISD::SETNE;
} else {
// (srl (ctlz x), 5) != 0 -> X == 0
// (srl (ctlz x), 5) == 1 -> X == 0
Cond = ISD::SETEQ;
}
SDValue Zero = DAG.getConstant(0, N0.getValueType());
return DAG.getSetCC(dl, VT, N0.getOperand(0).getOperand(0),
Zero, Cond);
}
}
// If the LHS is '(and load, const)', the RHS is 0,
// the test is for equality or unsigned, and all 1 bits of the const are
// in the same partial word, see if we can shorten the load.
if (DCI.isBeforeLegalize() &&
N0.getOpcode() == ISD::AND && C1 == 0 &&
N0.getNode()->hasOneUse() &&
isa<LoadSDNode>(N0.getOperand(0)) &&
N0.getOperand(0).getNode()->hasOneUse() &&
isa<ConstantSDNode>(N0.getOperand(1))) {
LoadSDNode *Lod = cast<LoadSDNode>(N0.getOperand(0));
uint64_t bestMask = 0;
unsigned bestWidth = 0, bestOffset = 0;
if (!Lod->isVolatile() && Lod->isUnindexed() &&
// FIXME: This uses getZExtValue() below so it only works on i64 and
// below.
N0.getValueType().getSizeInBits() <= 64) {
unsigned origWidth = N0.getValueType().getSizeInBits();
// We can narrow (e.g.) 16-bit extending loads on 32-bit target to
// 8 bits, but have to be careful...
if (Lod->getExtensionType() != ISD::NON_EXTLOAD)
origWidth = Lod->getMemoryVT().getSizeInBits();
uint64_t Mask =cast<ConstantSDNode>(N0.getOperand(1))->getZExtValue();
for (unsigned width = origWidth / 2; width>=8; width /= 2) {
uint64_t newMask = (1ULL << width) - 1;
for (unsigned offset=0; offset<origWidth/width; offset++) {
if ((newMask & Mask) == Mask) {
if (!TD->isLittleEndian())
bestOffset = (origWidth/width - offset - 1) * (width/8);
else
bestOffset = (uint64_t)offset * (width/8);
bestMask = Mask >> (offset * (width/8) * 8);
bestWidth = width;
break;
}
newMask = newMask << width;
}
}
}
if (bestWidth) {
MVT newVT = MVT::getIntegerVT(bestWidth);
if (newVT.isRound()) {
MVT PtrType = Lod->getOperand(1).getValueType();
SDValue Ptr = Lod->getBasePtr();
if (bestOffset != 0)
Ptr = DAG.getNode(ISD::ADD, dl, PtrType, Lod->getBasePtr(),
DAG.getConstant(bestOffset, PtrType));
unsigned NewAlign = MinAlign(Lod->getAlignment(), bestOffset);
SDValue NewLoad = DAG.getLoad(newVT, dl, Lod->getChain(), Ptr,
Lod->getSrcValue(),
Lod->getSrcValueOffset() + bestOffset,
false, NewAlign);
return DAG.getSetCC(dl, VT,
DAG.getNode(ISD::AND, dl, newVT, NewLoad,
DAG.getConstant(bestMask, newVT)),
DAG.getConstant(0LL, newVT), Cond);
}
}
}
// If the LHS is a ZERO_EXTEND, perform the comparison on the input.
if (N0.getOpcode() == ISD::ZERO_EXTEND) {
unsigned InSize = N0.getOperand(0).getValueType().getSizeInBits();
// If the comparison constant has bits in the upper part, the
// zero-extended value could never match.
if (C1.intersects(APInt::getHighBitsSet(C1.getBitWidth(),
C1.getBitWidth() - InSize))) {
switch (Cond) {
case ISD::SETUGT:
case ISD::SETUGE:
case ISD::SETEQ: return DAG.getConstant(0, VT);
case ISD::SETULT:
case ISD::SETULE:
case ISD::SETNE: return DAG.getConstant(1, VT);
case ISD::SETGT:
case ISD::SETGE:
// True if the sign bit of C1 is set.
return DAG.getConstant(C1.isNegative(), VT);
case ISD::SETLT:
case ISD::SETLE:
// True if the sign bit of C1 isn't set.
return DAG.getConstant(C1.isNonNegative(), VT);
default:
break;
}
}
// Otherwise, we can perform the comparison with the low bits.
switch (Cond) {
case ISD::SETEQ:
case ISD::SETNE:
case ISD::SETUGT:
case ISD::SETUGE:
case ISD::SETULT:
case ISD::SETULE:
return DAG.getSetCC(dl, VT, N0.getOperand(0),
DAG.getConstant(APInt(C1).trunc(InSize),
N0.getOperand(0).getValueType()),
Cond);
default:
break; // todo, be more careful with signed comparisons
}
} else if (N0.getOpcode() == ISD::SIGN_EXTEND_INREG &&
(Cond == ISD::SETEQ || Cond == ISD::SETNE)) {
MVT ExtSrcTy = cast<VTSDNode>(N0.getOperand(1))->getVT();
unsigned ExtSrcTyBits = ExtSrcTy.getSizeInBits();
MVT ExtDstTy = N0.getValueType();
unsigned ExtDstTyBits = ExtDstTy.getSizeInBits();
// If the extended part has any inconsistent bits, it cannot ever
// compare equal. In other words, they have to be all ones or all
// zeros.
APInt ExtBits =
APInt::getHighBitsSet(ExtDstTyBits, ExtDstTyBits - ExtSrcTyBits);
if ((C1 & ExtBits) != 0 && (C1 & ExtBits) != ExtBits)
return DAG.getConstant(Cond == ISD::SETNE, VT);
SDValue ZextOp;
MVT Op0Ty = N0.getOperand(0).getValueType();
if (Op0Ty == ExtSrcTy) {
ZextOp = N0.getOperand(0);
} else {
APInt Imm = APInt::getLowBitsSet(ExtDstTyBits, ExtSrcTyBits);
ZextOp = DAG.getNode(ISD::AND, dl, Op0Ty, N0.getOperand(0),
DAG.getConstant(Imm, Op0Ty));
}
if (!DCI.isCalledByLegalizer())
DCI.AddToWorklist(ZextOp.getNode());
// Otherwise, make this a use of a zext.
return DAG.getSetCC(dl, VT, ZextOp,
DAG.getConstant(C1 & APInt::getLowBitsSet(
ExtDstTyBits,
ExtSrcTyBits),
ExtDstTy),
Cond);
} else if ((N1C->isNullValue() || N1C->getAPIntValue() == 1) &&
(Cond == ISD::SETEQ || Cond == ISD::SETNE)) {
// SETCC (SETCC), [0|1], [EQ|NE] -> SETCC
if (N0.getOpcode() == ISD::SETCC) {
bool TrueWhenTrue = (Cond == ISD::SETEQ) ^ (N1C->getZExtValue() != 1);
if (TrueWhenTrue)
return N0;
// Invert the condition.
ISD::CondCode CC = cast<CondCodeSDNode>(N0.getOperand(2))->get();
CC = ISD::getSetCCInverse(CC,
N0.getOperand(0).getValueType().isInteger());
return DAG.getSetCC(dl, VT, N0.getOperand(0), N0.getOperand(1), CC);
}
if ((N0.getOpcode() == ISD::XOR ||
(N0.getOpcode() == ISD::AND &&
N0.getOperand(0).getOpcode() == ISD::XOR &&
N0.getOperand(1) == N0.getOperand(0).getOperand(1))) &&
isa<ConstantSDNode>(N0.getOperand(1)) &&
cast<ConstantSDNode>(N0.getOperand(1))->getAPIntValue() == 1) {
// If this is (X^1) == 0/1, swap the RHS and eliminate the xor. We
// can only do this if the top bits are known zero.
unsigned BitWidth = N0.getValueSizeInBits();
if (DAG.MaskedValueIsZero(N0,
APInt::getHighBitsSet(BitWidth,
BitWidth-1))) {
// Okay, get the un-inverted input value.
SDValue Val;
if (N0.getOpcode() == ISD::XOR)
Val = N0.getOperand(0);
else {
assert(N0.getOpcode() == ISD::AND &&
N0.getOperand(0).getOpcode() == ISD::XOR);
// ((X^1)&1)^1 -> X & 1
Val = DAG.getNode(ISD::AND, dl, N0.getValueType(),
N0.getOperand(0).getOperand(0),
N0.getOperand(1));
}
return DAG.getSetCC(dl, VT, Val, N1,
Cond == ISD::SETEQ ? ISD::SETNE : ISD::SETEQ);
}
}
}
APInt MinVal, MaxVal;
unsigned OperandBitSize = N1C->getValueType(0).getSizeInBits();
if (ISD::isSignedIntSetCC(Cond)) {
MinVal = APInt::getSignedMinValue(OperandBitSize);
MaxVal = APInt::getSignedMaxValue(OperandBitSize);
} else {
MinVal = APInt::getMinValue(OperandBitSize);
MaxVal = APInt::getMaxValue(OperandBitSize);
}
// Canonicalize GE/LE comparisons to use GT/LT comparisons.
if (Cond == ISD::SETGE || Cond == ISD::SETUGE) {
if (C1 == MinVal) return DAG.getConstant(1, VT); // X >= MIN --> true
// X >= C0 --> X > (C0-1)
return DAG.getSetCC(dl, VT, N0,
DAG.getConstant(C1-1, N1.getValueType()),
(Cond == ISD::SETGE) ? ISD::SETGT : ISD::SETUGT);
}
if (Cond == ISD::SETLE || Cond == ISD::SETULE) {
if (C1 == MaxVal) return DAG.getConstant(1, VT); // X <= MAX --> true
// X <= C0 --> X < (C0+1)
return DAG.getSetCC(dl, VT, N0,
DAG.getConstant(C1+1, N1.getValueType()),
(Cond == ISD::SETLE) ? ISD::SETLT : ISD::SETULT);
}
if ((Cond == ISD::SETLT || Cond == ISD::SETULT) && C1 == MinVal)
return DAG.getConstant(0, VT); // X < MIN --> false
if ((Cond == ISD::SETGE || Cond == ISD::SETUGE) && C1 == MinVal)
return DAG.getConstant(1, VT); // X >= MIN --> true
if ((Cond == ISD::SETGT || Cond == ISD::SETUGT) && C1 == MaxVal)
return DAG.getConstant(0, VT); // X > MAX --> false
if ((Cond == ISD::SETLE || Cond == ISD::SETULE) && C1 == MaxVal)
return DAG.getConstant(1, VT); // X <= MAX --> true
// Canonicalize setgt X, Min --> setne X, Min
if ((Cond == ISD::SETGT || Cond == ISD::SETUGT) && C1 == MinVal)
return DAG.getSetCC(dl, VT, N0, N1, ISD::SETNE);
// Canonicalize setlt X, Max --> setne X, Max
if ((Cond == ISD::SETLT || Cond == ISD::SETULT) && C1 == MaxVal)
return DAG.getSetCC(dl, VT, N0, N1, ISD::SETNE);
// If we have setult X, 1, turn it into seteq X, 0
if ((Cond == ISD::SETLT || Cond == ISD::SETULT) && C1 == MinVal+1)
return DAG.getSetCC(dl, VT, N0,
DAG.getConstant(MinVal, N0.getValueType()),
ISD::SETEQ);
// If we have setugt X, Max-1, turn it into seteq X, Max
else if ((Cond == ISD::SETGT || Cond == ISD::SETUGT) && C1 == MaxVal-1)
return DAG.getSetCC(dl, VT, N0,
DAG.getConstant(MaxVal, N0.getValueType()),
ISD::SETEQ);
// If we have "setcc X, C0", check to see if we can shrink the immediate
// by changing cc.
// SETUGT X, SINTMAX -> SETLT X, 0
if (Cond == ISD::SETUGT &&
C1 == APInt::getSignedMaxValue(OperandBitSize))
return DAG.getSetCC(dl, VT, N0,
DAG.getConstant(0, N1.getValueType()),
ISD::SETLT);
// SETULT X, SINTMIN -> SETGT X, -1
if (Cond == ISD::SETULT &&
C1 == APInt::getSignedMinValue(OperandBitSize)) {
SDValue ConstMinusOne =
DAG.getConstant(APInt::getAllOnesValue(OperandBitSize),
N1.getValueType());
return DAG.getSetCC(dl, VT, N0, ConstMinusOne, ISD::SETGT);
}
// Fold bit comparisons when we can.
if ((Cond == ISD::SETEQ || Cond == ISD::SETNE) &&
VT == N0.getValueType() && N0.getOpcode() == ISD::AND)
if (ConstantSDNode *AndRHS =
dyn_cast<ConstantSDNode>(N0.getOperand(1))) {
MVT ShiftTy = DCI.isBeforeLegalize() ?
getPointerTy() : getShiftAmountTy();
if (Cond == ISD::SETNE && C1 == 0) {// (X & 8) != 0 --> (X & 8) >> 3
// Perform the xform if the AND RHS is a single bit.
if (isPowerOf2_64(AndRHS->getZExtValue())) {
return DAG.getNode(ISD::SRL, dl, VT, N0,
DAG.getConstant(Log2_64(AndRHS->getZExtValue()),
ShiftTy));
}
} else if (Cond == ISD::SETEQ && C1 == AndRHS->getZExtValue()) {
// (X & 8) == 8 --> (X & 8) >> 3
// Perform the xform if C1 is a single bit.
if (C1.isPowerOf2()) {
return DAG.getNode(ISD::SRL, dl, VT, N0,
DAG.getConstant(C1.logBase2(), ShiftTy));
}
}
}
}
} else if (isa<ConstantSDNode>(N0.getNode())) {
// Ensure that the constant occurs on the RHS.
return DAG.getSetCC(dl, VT, N1, N0, ISD::getSetCCSwappedOperands(Cond));
}
if (isa<ConstantFPSDNode>(N0.getNode())) {
// Constant fold or commute setcc.
SDValue O = DAG.FoldSetCC(VT, N0, N1, Cond, dl);
if (O.getNode()) return O;
} else if (ConstantFPSDNode *CFP = dyn_cast<ConstantFPSDNode>(N1.getNode())) {
// If the RHS of an FP comparison is a constant, simplify it away in
// some cases.
if (CFP->getValueAPF().isNaN()) {
// If an operand is known to be a nan, we can fold it.
switch (ISD::getUnorderedFlavor(Cond)) {
default: assert(0 && "Unknown flavor!");
case 0: // Known false.
return DAG.getConstant(0, VT);
case 1: // Known true.
return DAG.getConstant(1, VT);
case 2: // Undefined.
return DAG.getUNDEF(VT);
}
}
// Otherwise, we know the RHS is not a NaN. Simplify the node to drop the
// constant if knowing that the operand is non-nan is enough. We prefer to
// have SETO(x,x) instead of SETO(x, 0.0) because this avoids having to
// materialize 0.0.
if (Cond == ISD::SETO || Cond == ISD::SETUO)
return DAG.getSetCC(dl, VT, N0, N0, Cond);
}
if (N0 == N1) {
// We can always fold X == X for integer setcc's.
if (N0.getValueType().isInteger())
return DAG.getConstant(ISD::isTrueWhenEqual(Cond), VT);
unsigned UOF = ISD::getUnorderedFlavor(Cond);
if (UOF == 2) // FP operators that are undefined on NaNs.
return DAG.getConstant(ISD::isTrueWhenEqual(Cond), VT);
if (UOF == unsigned(ISD::isTrueWhenEqual(Cond)))
return DAG.getConstant(UOF, VT);
// Otherwise, we can't fold it. However, we can simplify it to SETUO/SETO
// if it is not already.
ISD::CondCode NewCond = UOF == 0 ? ISD::SETO : ISD::SETUO;
if (NewCond != Cond)
return DAG.getSetCC(dl, VT, N0, N1, NewCond);
}
if ((Cond == ISD::SETEQ || Cond == ISD::SETNE) &&
N0.getValueType().isInteger()) {
if (N0.getOpcode() == ISD::ADD || N0.getOpcode() == ISD::SUB ||
N0.getOpcode() == ISD::XOR) {
// Simplify (X+Y) == (X+Z) --> Y == Z
if (N0.getOpcode() == N1.getOpcode()) {
if (N0.getOperand(0) == N1.getOperand(0))
return DAG.getSetCC(dl, VT, N0.getOperand(1), N1.getOperand(1), Cond);
if (N0.getOperand(1) == N1.getOperand(1))
return DAG.getSetCC(dl, VT, N0.getOperand(0), N1.getOperand(0), Cond);
if (DAG.isCommutativeBinOp(N0.getOpcode())) {
// If X op Y == Y op X, try other combinations.
if (N0.getOperand(0) == N1.getOperand(1))
return DAG.getSetCC(dl, VT, N0.getOperand(1), N1.getOperand(0),
Cond);
if (N0.getOperand(1) == N1.getOperand(0))
return DAG.getSetCC(dl, VT, N0.getOperand(0), N1.getOperand(1),
Cond);
}
}
if (ConstantSDNode *RHSC = dyn_cast<ConstantSDNode>(N1)) {
if (ConstantSDNode *LHSR = dyn_cast<ConstantSDNode>(N0.getOperand(1))) {
// Turn (X+C1) == C2 --> X == C2-C1
if (N0.getOpcode() == ISD::ADD && N0.getNode()->hasOneUse()) {
return DAG.getSetCC(dl, VT, N0.getOperand(0),
DAG.getConstant(RHSC->getAPIntValue()-
LHSR->getAPIntValue(),
N0.getValueType()), Cond);
}
// Turn (X^C1) == C2 into X == C1^C2 iff X&~C1 = 0.
if (N0.getOpcode() == ISD::XOR)
// If we know that all of the inverted bits are zero, don't bother
// performing the inversion.
if (DAG.MaskedValueIsZero(N0.getOperand(0), ~LHSR->getAPIntValue()))
return
DAG.getSetCC(dl, VT, N0.getOperand(0),
DAG.getConstant(LHSR->getAPIntValue() ^
RHSC->getAPIntValue(),
N0.getValueType()),
Cond);
}
// Turn (C1-X) == C2 --> X == C1-C2
if (ConstantSDNode *SUBC = dyn_cast<ConstantSDNode>(N0.getOperand(0))) {
if (N0.getOpcode() == ISD::SUB && N0.getNode()->hasOneUse()) {
return
DAG.getSetCC(dl, VT, N0.getOperand(1),
DAG.getConstant(SUBC->getAPIntValue() -
RHSC->getAPIntValue(),
N0.getValueType()),
Cond);
}
}
}
// Simplify (X+Z) == X --> Z == 0
if (N0.getOperand(0) == N1)
return DAG.getSetCC(dl, VT, N0.getOperand(1),
DAG.getConstant(0, N0.getValueType()), Cond);
if (N0.getOperand(1) == N1) {
if (DAG.isCommutativeBinOp(N0.getOpcode()))
return DAG.getSetCC(dl, VT, N0.getOperand(0),
DAG.getConstant(0, N0.getValueType()), Cond);
else if (N0.getNode()->hasOneUse()) {
assert(N0.getOpcode() == ISD::SUB && "Unexpected operation!");
// (Z-X) == X --> Z == X<<1
SDValue SH = DAG.getNode(ISD::SHL, dl, N1.getValueType(),
N1,
DAG.getConstant(1, getShiftAmountTy()));
if (!DCI.isCalledByLegalizer())
DCI.AddToWorklist(SH.getNode());
return DAG.getSetCC(dl, VT, N0.getOperand(0), SH, Cond);
}
}
}
if (N1.getOpcode() == ISD::ADD || N1.getOpcode() == ISD::SUB ||
N1.getOpcode() == ISD::XOR) {
// Simplify X == (X+Z) --> Z == 0
if (N1.getOperand(0) == N0) {
return DAG.getSetCC(dl, VT, N1.getOperand(1),
DAG.getConstant(0, N1.getValueType()), Cond);
} else if (N1.getOperand(1) == N0) {
if (DAG.isCommutativeBinOp(N1.getOpcode())) {
return DAG.getSetCC(dl, VT, N1.getOperand(0),
DAG.getConstant(0, N1.getValueType()), Cond);
} else if (N1.getNode()->hasOneUse()) {
assert(N1.getOpcode() == ISD::SUB && "Unexpected operation!");
// X == (Z-X) --> X<<1 == Z
SDValue SH = DAG.getNode(ISD::SHL, dl, N1.getValueType(), N0,
DAG.getConstant(1, getShiftAmountTy()));
if (!DCI.isCalledByLegalizer())
DCI.AddToWorklist(SH.getNode());
return DAG.getSetCC(dl, VT, SH, N1.getOperand(0), Cond);
}
}
}
// Simplify x&y == y to x&y != 0 if y has exactly one bit set.
// Note that where y is variable and is known to have at most
// one bit set (for example, if it is z&1) we cannot do this;
// the expressions are not equivalent when y==0.
if (N0.getOpcode() == ISD::AND)
if (N0.getOperand(0) == N1 || N0.getOperand(1) == N1) {
if (ValueHasExactlyOneBitSet(N1, DAG)) {
Cond = ISD::getSetCCInverse(Cond, /*isInteger=*/true);
SDValue Zero = DAG.getConstant(0, N1.getValueType());
return DAG.getSetCC(dl, VT, N0, Zero, Cond);
}
}
if (N1.getOpcode() == ISD::AND)
if (N1.getOperand(0) == N0 || N1.getOperand(1) == N0) {
if (ValueHasExactlyOneBitSet(N0, DAG)) {
Cond = ISD::getSetCCInverse(Cond, /*isInteger=*/true);
SDValue Zero = DAG.getConstant(0, N0.getValueType());
return DAG.getSetCC(dl, VT, N1, Zero, Cond);
}
}
}
// Fold away ALL boolean setcc's.
SDValue Temp;
if (N0.getValueType() == MVT::i1 && foldBooleans) {
switch (Cond) {
default: assert(0 && "Unknown integer setcc!");
case ISD::SETEQ: // X == Y -> ~(X^Y)
Temp = DAG.getNode(ISD::XOR, dl, MVT::i1, N0, N1);
N0 = DAG.getNOT(dl, Temp, MVT::i1);
if (!DCI.isCalledByLegalizer())
DCI.AddToWorklist(Temp.getNode());
break;
case ISD::SETNE: // X != Y --> (X^Y)
N0 = DAG.getNode(ISD::XOR, dl, MVT::i1, N0, N1);
break;
case ISD::SETGT: // X >s Y --> X == 0 & Y == 1 --> ~X & Y
case ISD::SETULT: // X <u Y --> X == 0 & Y == 1 --> ~X & Y
Temp = DAG.getNOT(dl, N0, MVT::i1);
N0 = DAG.getNode(ISD::AND, dl, MVT::i1, N1, Temp);
if (!DCI.isCalledByLegalizer())
DCI.AddToWorklist(Temp.getNode());
break;
case ISD::SETLT: // X <s Y --> X == 1 & Y == 0 --> ~Y & X
case ISD::SETUGT: // X >u Y --> X == 1 & Y == 0 --> ~Y & X
Temp = DAG.getNOT(dl, N1, MVT::i1);
N0 = DAG.getNode(ISD::AND, dl, MVT::i1, N0, Temp);
if (!DCI.isCalledByLegalizer())
DCI.AddToWorklist(Temp.getNode());
break;
case ISD::SETULE: // X <=u Y --> X == 0 | Y == 1 --> ~X | Y
case ISD::SETGE: // X >=s Y --> X == 0 | Y == 1 --> ~X | Y
Temp = DAG.getNOT(dl, N0, MVT::i1);
N0 = DAG.getNode(ISD::OR, dl, MVT::i1, N1, Temp);
if (!DCI.isCalledByLegalizer())
DCI.AddToWorklist(Temp.getNode());
break;
case ISD::SETUGE: // X >=u Y --> X == 1 | Y == 0 --> ~Y | X
case ISD::SETLE: // X <=s Y --> X == 1 | Y == 0 --> ~Y | X
Temp = DAG.getNOT(dl, N1, MVT::i1);
N0 = DAG.getNode(ISD::OR, dl, MVT::i1, N0, Temp);
break;
}
if (VT != MVT::i1) {
if (!DCI.isCalledByLegalizer())
DCI.AddToWorklist(N0.getNode());
// FIXME: If running after legalize, we probably can't do this.
N0 = DAG.getNode(ISD::ZERO_EXTEND, dl, VT, N0);
}
return N0;
}
// Could not fold it.
return SDValue();
}
/// isGAPlusOffset - Returns true (and the GlobalValue and the offset) if the
/// node is a GlobalAddress + offset.
bool TargetLowering::isGAPlusOffset(SDNode *N, GlobalValue* &GA,
int64_t &Offset) const {
if (isa<GlobalAddressSDNode>(N)) {
GlobalAddressSDNode *GASD = cast<GlobalAddressSDNode>(N);
GA = GASD->getGlobal();
Offset += GASD->getOffset();
return true;
}
if (N->getOpcode() == ISD::ADD) {
SDValue N1 = N->getOperand(0);
SDValue N2 = N->getOperand(1);
if (isGAPlusOffset(N1.getNode(), GA, Offset)) {
ConstantSDNode *V = dyn_cast<ConstantSDNode>(N2);
if (V) {
Offset += V->getSExtValue();
return true;
}
} else if (isGAPlusOffset(N2.getNode(), GA, Offset)) {
ConstantSDNode *V = dyn_cast<ConstantSDNode>(N1);
if (V) {
Offset += V->getSExtValue();
return true;
}
}
}
return false;
}
/// isConsecutiveLoad - Return true if LD (which must be a LoadSDNode) is
/// loading 'Bytes' bytes from a location that is 'Dist' units away from the
/// location that the 'Base' load is loading from.
bool TargetLowering::isConsecutiveLoad(SDNode *LD, SDNode *Base,
unsigned Bytes, int Dist,
const MachineFrameInfo *MFI) const {
if (LD->getOperand(0).getNode() != Base->getOperand(0).getNode())
return false;
MVT VT = LD->getValueType(0);
if (VT.getSizeInBits() / 8 != Bytes)
return false;
SDValue Loc = LD->getOperand(1);
SDValue BaseLoc = Base->getOperand(1);
if (Loc.getOpcode() == ISD::FrameIndex) {
if (BaseLoc.getOpcode() != ISD::FrameIndex)
return false;
int FI = cast<FrameIndexSDNode>(Loc)->getIndex();
int BFI = cast<FrameIndexSDNode>(BaseLoc)->getIndex();
int FS = MFI->getObjectSize(FI);
int BFS = MFI->getObjectSize(BFI);
if (FS != BFS || FS != (int)Bytes) return false;
return MFI->getObjectOffset(FI) == (MFI->getObjectOffset(BFI) + Dist*Bytes);
}
GlobalValue *GV1 = NULL;
GlobalValue *GV2 = NULL;
int64_t Offset1 = 0;
int64_t Offset2 = 0;
bool isGA1 = isGAPlusOffset(Loc.getNode(), GV1, Offset1);
bool isGA2 = isGAPlusOffset(BaseLoc.getNode(), GV2, Offset2);
if (isGA1 && isGA2 && GV1 == GV2)
return Offset1 == (Offset2 + Dist*Bytes);
return false;
}
SDValue TargetLowering::
PerformDAGCombine(SDNode *N, DAGCombinerInfo &DCI) const {
// Default implementation: no optimization.
return SDValue();
}
//===----------------------------------------------------------------------===//
// Inline Assembler Implementation Methods
//===----------------------------------------------------------------------===//
TargetLowering::ConstraintType
TargetLowering::getConstraintType(const std::string &Constraint) const {
// FIXME: lots more standard ones to handle.
if (Constraint.size() == 1) {
switch (Constraint[0]) {
default: break;
case 'r': return C_RegisterClass;
case 'm': // memory
case 'o': // offsetable
case 'V': // not offsetable
return C_Memory;
case 'i': // Simple Integer or Relocatable Constant
case 'n': // Simple Integer
case 's': // Relocatable Constant
case 'X': // Allow ANY value.
case 'I': // Target registers.
case 'J':
case 'K':
case 'L':
case 'M':
case 'N':
case 'O':
case 'P':
return C_Other;
}
}
if (Constraint.size() > 1 && Constraint[0] == '{' &&
Constraint[Constraint.size()-1] == '}')
return C_Register;
return C_Unknown;
}
/// LowerXConstraint - try to replace an X constraint, which matches anything,
/// with another that has more specific requirements based on the type of the
/// corresponding operand.
const char *TargetLowering::LowerXConstraint(MVT ConstraintVT) const{
if (ConstraintVT.isInteger())
return "r";
if (ConstraintVT.isFloatingPoint())
return "f"; // works for many targets
return 0;
}
/// LowerAsmOperandForConstraint - Lower the specified operand into the Ops
/// vector. If it is invalid, don't add anything to Ops.
void TargetLowering::LowerAsmOperandForConstraint(SDValue Op,
char ConstraintLetter,
bool hasMemory,
std::vector<SDValue> &Ops,
SelectionDAG &DAG) const {
switch (ConstraintLetter) {
default: break;
case 'X': // Allows any operand; labels (basic block) use this.
if (Op.getOpcode() == ISD::BasicBlock) {
Ops.push_back(Op);
return;
}
// fall through
case 'i': // Simple Integer or Relocatable Constant
case 'n': // Simple Integer
case 's': { // Relocatable Constant
// These operands are interested in values of the form (GV+C), where C may
// be folded in as an offset of GV, or it may be explicitly added. Also, it
// is possible and fine if either GV or C are missing.
ConstantSDNode *C = dyn_cast<ConstantSDNode>(Op);
GlobalAddressSDNode *GA = dyn_cast<GlobalAddressSDNode>(Op);
// If we have "(add GV, C)", pull out GV/C
if (Op.getOpcode() == ISD::ADD) {
C = dyn_cast<ConstantSDNode>(Op.getOperand(1));
GA = dyn_cast<GlobalAddressSDNode>(Op.getOperand(0));
if (C == 0 || GA == 0) {
C = dyn_cast<ConstantSDNode>(Op.getOperand(0));
GA = dyn_cast<GlobalAddressSDNode>(Op.getOperand(1));
}
if (C == 0 || GA == 0)
C = 0, GA = 0;
}
// If we find a valid operand, map to the TargetXXX version so that the
// value itself doesn't get selected.
if (GA) { // Either &GV or &GV+C
if (ConstraintLetter != 'n') {
int64_t Offs = GA->getOffset();
if (C) Offs += C->getZExtValue();
Ops.push_back(DAG.getTargetGlobalAddress(GA->getGlobal(),
Op.getValueType(), Offs));
return;
}
}
if (C) { // just C, no GV.
// Simple constants are not allowed for 's'.
if (ConstraintLetter != 's') {
// gcc prints these as sign extended. Sign extend value to 64 bits
// now; without this it would get ZExt'd later in
// ScheduleDAGSDNodes::EmitNode, which is very generic.
Ops.push_back(DAG.getTargetConstant(C->getAPIntValue().getSExtValue(),
MVT::i64));
return;
}
}
break;
}
}
}
std::vector<unsigned> TargetLowering::
getRegClassForInlineAsmConstraint(const std::string &Constraint,
MVT VT) const {
return std::vector<unsigned>();
}
std::pair<unsigned, const TargetRegisterClass*> TargetLowering::
getRegForInlineAsmConstraint(const std::string &Constraint,
MVT VT) const {
if (Constraint[0] != '{')
return std::pair<unsigned, const TargetRegisterClass*>(0, 0);
assert(*(Constraint.end()-1) == '}' && "Not a brace enclosed constraint?");
// Remove the braces from around the name.
std::string RegName(Constraint.begin()+1, Constraint.end()-1);
// Figure out which register class contains this reg.
const TargetRegisterInfo *RI = TM.getRegisterInfo();
for (TargetRegisterInfo::regclass_iterator RCI = RI->regclass_begin(),
E = RI->regclass_end(); RCI != E; ++RCI) {
const TargetRegisterClass *RC = *RCI;
// If none of the the value types for this register class are valid, we
// can't use it. For example, 64-bit reg classes on 32-bit targets.
bool isLegal = false;
for (TargetRegisterClass::vt_iterator I = RC->vt_begin(), E = RC->vt_end();
I != E; ++I) {
if (isTypeLegal(*I)) {
isLegal = true;
break;
}
}
if (!isLegal) continue;
for (TargetRegisterClass::iterator I = RC->begin(), E = RC->end();
I != E; ++I) {
if (StringsEqualNoCase(RegName, RI->get(*I).AsmName))
return std::make_pair(*I, RC);
}
}
return std::pair<unsigned, const TargetRegisterClass*>(0, 0);
}
//===----------------------------------------------------------------------===//
// Constraint Selection.
/// isMatchingInputConstraint - Return true of this is an input operand that is
/// a matching constraint like "4".
bool TargetLowering::AsmOperandInfo::isMatchingInputConstraint() const {
assert(!ConstraintCode.empty() && "No known constraint!");
return isdigit(ConstraintCode[0]);
}
/// getMatchedOperand - If this is an input matching constraint, this method
/// returns the output operand it matches.
unsigned TargetLowering::AsmOperandInfo::getMatchedOperand() const {
assert(!ConstraintCode.empty() && "No known constraint!");
return atoi(ConstraintCode.c_str());
}
/// getConstraintGenerality - Return an integer indicating how general CT
/// is.
static unsigned getConstraintGenerality(TargetLowering::ConstraintType CT) {
switch (CT) {
default: assert(0 && "Unknown constraint type!");
case TargetLowering::C_Other:
case TargetLowering::C_Unknown:
return 0;
case TargetLowering::C_Register:
return 1;
case TargetLowering::C_RegisterClass:
return 2;
case TargetLowering::C_Memory:
return 3;
}
}
/// ChooseConstraint - If there are multiple different constraints that we
/// could pick for this operand (e.g. "imr") try to pick the 'best' one.
/// This is somewhat tricky: constraints fall into four classes:
/// Other -> immediates and magic values
/// Register -> one specific register
/// RegisterClass -> a group of regs
/// Memory -> memory
/// Ideally, we would pick the most specific constraint possible: if we have
/// something that fits into a register, we would pick it. The problem here
/// is that if we have something that could either be in a register or in
/// memory that use of the register could cause selection of *other*
/// operands to fail: they might only succeed if we pick memory. Because of
/// this the heuristic we use is:
///
/// 1) If there is an 'other' constraint, and if the operand is valid for
/// that constraint, use it. This makes us take advantage of 'i'
/// constraints when available.
/// 2) Otherwise, pick the most general constraint present. This prefers
/// 'm' over 'r', for example.
///
static void ChooseConstraint(TargetLowering::AsmOperandInfo &OpInfo,
bool hasMemory, const TargetLowering &TLI,
SDValue Op, SelectionDAG *DAG) {
assert(OpInfo.Codes.size() > 1 && "Doesn't have multiple constraint options");
unsigned BestIdx = 0;
TargetLowering::ConstraintType BestType = TargetLowering::C_Unknown;
int BestGenerality = -1;
// Loop over the options, keeping track of the most general one.
for (unsigned i = 0, e = OpInfo.Codes.size(); i != e; ++i) {
TargetLowering::ConstraintType CType =
TLI.getConstraintType(OpInfo.Codes[i]);
// If this is an 'other' constraint, see if the operand is valid for it.
// For example, on X86 we might have an 'rI' constraint. If the operand
// is an integer in the range [0..31] we want to use I (saving a load
// of a register), otherwise we must use 'r'.
if (CType == TargetLowering::C_Other && Op.getNode()) {
assert(OpInfo.Codes[i].size() == 1 &&
"Unhandled multi-letter 'other' constraint");
std::vector<SDValue> ResultOps;
TLI.LowerAsmOperandForConstraint(Op, OpInfo.Codes[i][0], hasMemory,
ResultOps, *DAG);
if (!ResultOps.empty()) {
BestType = CType;
BestIdx = i;
break;
}
}
// This constraint letter is more general than the previous one, use it.
int Generality = getConstraintGenerality(CType);
if (Generality > BestGenerality) {
BestType = CType;
BestIdx = i;
BestGenerality = Generality;
}
}
OpInfo.ConstraintCode = OpInfo.Codes[BestIdx];
OpInfo.ConstraintType = BestType;
}
/// ComputeConstraintToUse - Determines the constraint code and constraint
/// type to use for the specific AsmOperandInfo, setting
/// OpInfo.ConstraintCode and OpInfo.ConstraintType.
void TargetLowering::ComputeConstraintToUse(AsmOperandInfo &OpInfo,
SDValue Op,
bool hasMemory,
SelectionDAG *DAG) const {
assert(!OpInfo.Codes.empty() && "Must have at least one constraint");
// Single-letter constraints ('r') are very common.
if (OpInfo.Codes.size() == 1) {
OpInfo.ConstraintCode = OpInfo.Codes[0];
OpInfo.ConstraintType = getConstraintType(OpInfo.ConstraintCode);
} else {
ChooseConstraint(OpInfo, hasMemory, *this, Op, DAG);
}
// 'X' matches anything.
if (OpInfo.ConstraintCode == "X" && OpInfo.CallOperandVal) {
// Labels and constants are handled elsewhere ('X' is the only thing
// that matches labels).
if (isa<BasicBlock>(OpInfo.CallOperandVal) ||
isa<ConstantInt>(OpInfo.CallOperandVal))
return;
// Otherwise, try to resolve it to something we know about by looking at
// the actual operand type.
if (const char *Repl = LowerXConstraint(OpInfo.ConstraintVT)) {
OpInfo.ConstraintCode = Repl;
OpInfo.ConstraintType = getConstraintType(OpInfo.ConstraintCode);
}
}
}
//===----------------------------------------------------------------------===//
// Loop Strength Reduction hooks
//===----------------------------------------------------------------------===//
/// isLegalAddressingMode - Return true if the addressing mode represented
/// by AM is legal for this target, for a load/store of the specified type.
bool TargetLowering::isLegalAddressingMode(const AddrMode &AM,
const Type *Ty) const {
// The default implementation of this implements a conservative RISCy, r+r and
// r+i addr mode.
// Allows a sign-extended 16-bit immediate field.
if (AM.BaseOffs <= -(1LL << 16) || AM.BaseOffs >= (1LL << 16)-1)
return false;
// No global is ever allowed as a base.
if (AM.BaseGV)
return false;
// Only support r+r,
switch (AM.Scale) {
case 0: // "r+i" or just "i", depending on HasBaseReg.
break;
case 1:
if (AM.HasBaseReg && AM.BaseOffs) // "r+r+i" is not allowed.
return false;
// Otherwise we have r+r or r+i.
break;
case 2:
if (AM.HasBaseReg || AM.BaseOffs) // 2*r+r or 2*r+i is not allowed.
return false;
// Allow 2*r as r+r.
break;
}
return true;
}
struct mu {
APInt m; // magic number
bool a; // add indicator
unsigned s; // shift amount
};
/// magicu - calculate the magic numbers required to codegen an integer udiv as
/// a sequence of multiply, add and shifts. Requires that the divisor not be 0.
static mu magicu(const APInt& d) {
unsigned p;
APInt nc, delta, q1, r1, q2, r2;
struct mu magu;
magu.a = 0; // initialize "add" indicator
APInt allOnes = APInt::getAllOnesValue(d.getBitWidth());
APInt signedMin = APInt::getSignedMinValue(d.getBitWidth());
APInt signedMax = APInt::getSignedMaxValue(d.getBitWidth());
nc = allOnes - (-d).urem(d);
p = d.getBitWidth() - 1; // initialize p
q1 = signedMin.udiv(nc); // initialize q1 = 2p/nc
r1 = signedMin - q1*nc; // initialize r1 = rem(2p,nc)
q2 = signedMax.udiv(d); // initialize q2 = (2p-1)/d
r2 = signedMax - q2*d; // initialize r2 = rem((2p-1),d)
do {
p = p + 1;
if (r1.uge(nc - r1)) {
q1 = q1 + q1 + 1; // update q1
r1 = r1 + r1 - nc; // update r1
}
else {
q1 = q1+q1; // update q1
r1 = r1+r1; // update r1
}
if ((r2 + 1).uge(d - r2)) {
if (q2.uge(signedMax)) magu.a = 1;
q2 = q2+q2 + 1; // update q2
r2 = r2+r2 + 1 - d; // update r2
}
else {
if (q2.uge(signedMin)) magu.a = 1;
q2 = q2+q2; // update q2
r2 = r2+r2 + 1; // update r2
}
delta = d - 1 - r2;
} while (p < d.getBitWidth()*2 &&
(q1.ult(delta) || (q1 == delta && r1 == 0)));
magu.m = q2 + 1; // resulting magic number
magu.s = p - d.getBitWidth(); // resulting shift
return magu;
}
// Magic for divide replacement
struct ms {
APInt m; // magic number
unsigned s; // shift amount
};
/// magic - calculate the magic numbers required to codegen an integer sdiv as
/// a sequence of multiply and shifts. Requires that the divisor not be 0, 1,
/// or -1.
static ms magic(const APInt& d) {
unsigned p;
APInt ad, anc, delta, q1, r1, q2, r2, t;
APInt allOnes = APInt::getAllOnesValue(d.getBitWidth());
APInt signedMin = APInt::getSignedMinValue(d.getBitWidth());
APInt signedMax = APInt::getSignedMaxValue(d.getBitWidth());
struct ms mag;
ad = d.abs();
t = signedMin + (d.lshr(d.getBitWidth() - 1));
anc = t - 1 - t.urem(ad); // absolute value of nc
p = d.getBitWidth() - 1; // initialize p
q1 = signedMin.udiv(anc); // initialize q1 = 2p/abs(nc)
r1 = signedMin - q1*anc; // initialize r1 = rem(2p,abs(nc))
q2 = signedMin.udiv(ad); // initialize q2 = 2p/abs(d)
r2 = signedMin - q2*ad; // initialize r2 = rem(2p,abs(d))
do {
p = p + 1;
q1 = q1<<1; // update q1 = 2p/abs(nc)
r1 = r1<<1; // update r1 = rem(2p/abs(nc))
if (r1.uge(anc)) { // must be unsigned comparison
q1 = q1 + 1;
r1 = r1 - anc;
}
q2 = q2<<1; // update q2 = 2p/abs(d)
r2 = r2<<1; // update r2 = rem(2p/abs(d))
if (r2.uge(ad)) { // must be unsigned comparison
q2 = q2 + 1;
r2 = r2 - ad;
}
delta = ad - r2;
} while (q1.ule(delta) || (q1 == delta && r1 == 0));
mag.m = q2 + 1;
if (d.isNegative()) mag.m = -mag.m; // resulting magic number
mag.s = p - d.getBitWidth(); // resulting shift
return mag;
}
/// BuildSDIVSequence - Given an ISD::SDIV node expressing a divide by constant,
/// return a DAG expression to select that will generate the same value by
/// multiplying by a magic number. See:
/// <http://the.wall.riscom.net/books/proc/ppc/cwg/code2.html>
SDValue TargetLowering::BuildSDIV(SDNode *N, SelectionDAG &DAG,
std::vector<SDNode*>* Created) const {
MVT VT = N->getValueType(0);
DebugLoc dl= N->getDebugLoc();
// Check to see if we can do this.
// FIXME: We should be more aggressive here.
if (!isTypeLegal(VT))
return SDValue();
APInt d = cast<ConstantSDNode>(N->getOperand(1))->getAPIntValue();
ms magics = magic(d);
// Multiply the numerator (operand 0) by the magic value
// FIXME: We should support doing a MUL in a wider type
SDValue Q;
if (isOperationLegalOrCustom(ISD::MULHS, VT))
Q = DAG.getNode(ISD::MULHS, dl, VT, N->getOperand(0),
DAG.getConstant(magics.m, VT));
else if (isOperationLegalOrCustom(ISD::SMUL_LOHI, VT))
Q = SDValue(DAG.getNode(ISD::SMUL_LOHI, dl, DAG.getVTList(VT, VT),
N->getOperand(0),
DAG.getConstant(magics.m, VT)).getNode(), 1);
else
return SDValue(); // No mulhs or equvialent
// If d > 0 and m < 0, add the numerator
if (d.isStrictlyPositive() && magics.m.isNegative()) {
Q = DAG.getNode(ISD::ADD, dl, VT, Q, N->getOperand(0));
if (Created)
Created->push_back(Q.getNode());
}
// If d < 0 and m > 0, subtract the numerator.
if (d.isNegative() && magics.m.isStrictlyPositive()) {
Q = DAG.getNode(ISD::SUB, dl, VT, Q, N->getOperand(0));
if (Created)
Created->push_back(Q.getNode());
}
// Shift right algebraic if shift value is nonzero
if (magics.s > 0) {
Q = DAG.getNode(ISD::SRA, dl, VT, Q,
DAG.getConstant(magics.s, getShiftAmountTy()));
if (Created)
Created->push_back(Q.getNode());
}
// Extract the sign bit and add it to the quotient
SDValue T =
DAG.getNode(ISD::SRL, dl, VT, Q, DAG.getConstant(VT.getSizeInBits()-1,
getShiftAmountTy()));
if (Created)
Created->push_back(T.getNode());
return DAG.getNode(ISD::ADD, dl, VT, Q, T);
}
/// BuildUDIVSequence - Given an ISD::UDIV node expressing a divide by constant,
/// return a DAG expression to select that will generate the same value by
/// multiplying by a magic number. See:
/// <http://the.wall.riscom.net/books/proc/ppc/cwg/code2.html>
SDValue TargetLowering::BuildUDIV(SDNode *N, SelectionDAG &DAG,
std::vector<SDNode*>* Created) const {
MVT VT = N->getValueType(0);
DebugLoc dl = N->getDebugLoc();
// Check to see if we can do this.
// FIXME: We should be more aggressive here.
if (!isTypeLegal(VT))
return SDValue();
// FIXME: We should use a narrower constant when the upper
// bits are known to be zero.
ConstantSDNode *N1C = cast<ConstantSDNode>(N->getOperand(1));
mu magics = magicu(N1C->getAPIntValue());
// Multiply the numerator (operand 0) by the magic value
// FIXME: We should support doing a MUL in a wider type
SDValue Q;
if (isOperationLegalOrCustom(ISD::MULHU, VT))
Q = DAG.getNode(ISD::MULHU, dl, VT, N->getOperand(0),
DAG.getConstant(magics.m, VT));
else if (isOperationLegalOrCustom(ISD::UMUL_LOHI, VT))
Q = SDValue(DAG.getNode(ISD::UMUL_LOHI, dl, DAG.getVTList(VT, VT),
N->getOperand(0),
DAG.getConstant(magics.m, VT)).getNode(), 1);
else
return SDValue(); // No mulhu or equvialent
if (Created)
Created->push_back(Q.getNode());
if (magics.a == 0) {
assert(magics.s < N1C->getAPIntValue().getBitWidth() &&
"We shouldn't generate an undefined shift!");
return DAG.getNode(ISD::SRL, dl, VT, Q,
DAG.getConstant(magics.s, getShiftAmountTy()));
} else {
SDValue NPQ = DAG.getNode(ISD::SUB, dl, VT, N->getOperand(0), Q);
if (Created)
Created->push_back(NPQ.getNode());
NPQ = DAG.getNode(ISD::SRL, dl, VT, NPQ,
DAG.getConstant(1, getShiftAmountTy()));
if (Created)
Created->push_back(NPQ.getNode());
NPQ = DAG.getNode(ISD::ADD, dl, VT, NPQ, Q);
if (Created)
Created->push_back(NPQ.getNode());
return DAG.getNode(ISD::SRL, dl, VT, NPQ,
DAG.getConstant(magics.s-1, getShiftAmountTy()));
}
}
/// IgnoreHarmlessInstructions - Ignore instructions between a CALL and RET
/// node that don't prevent tail call optimization.
static SDValue IgnoreHarmlessInstructions(SDValue node) {
// Found call return.
if (node.getOpcode() == ISD::CALL) return node;
// Ignore MERGE_VALUES. Will have at least one operand.
if (node.getOpcode() == ISD::MERGE_VALUES)
return IgnoreHarmlessInstructions(node.getOperand(0));
// Ignore ANY_EXTEND node.
if (node.getOpcode() == ISD::ANY_EXTEND)
return IgnoreHarmlessInstructions(node.getOperand(0));
if (node.getOpcode() == ISD::TRUNCATE)
return IgnoreHarmlessInstructions(node.getOperand(0));
// Any other node type.
return node;
}
bool TargetLowering::CheckTailCallReturnConstraints(CallSDNode *TheCall,
SDValue Ret) {
unsigned NumOps = Ret.getNumOperands();
// ISD::CALL results:(value0, ..., valuen, chain)
// ISD::RET operands:(chain, value0, flag0, ..., valuen, flagn)
// Value return:
// Check that operand of the RET node sources from the CALL node. The RET node
// has at least two operands. Operand 0 holds the chain. Operand 1 holds the
// value.
if (NumOps > 1 &&
IgnoreHarmlessInstructions(Ret.getOperand(1)) == SDValue(TheCall,0))
return true;
// void return: The RET node has the chain result value of the CALL node as
// input.
if (NumOps == 1 &&
Ret.getOperand(0) == SDValue(TheCall, TheCall->getNumValues()-1))
return true;
return false;
}
|