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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=riscv32 -verify-machineinstrs < %s \
; RUN: | FileCheck %s -check-prefixes=RV32
; RUN: llc -mtriple=riscv64 -verify-machineinstrs < %s \
; RUN: | FileCheck %s -check-prefixes=RV64
;Copy tests from llvm/tests/Transforms/CodeGenPrepare/X86/overflow-intrinsics.ll
;to test shouldFormOverflowOp on RISCV
define i64 @uaddo1_overflow_used(i64 %a, i64 %b) nounwind ssp {
; RV32-LABEL: uaddo1_overflow_used:
; RV32: # %bb.0:
; RV32-NEXT: add a5, a3, a1
; RV32-NEXT: add a4, a2, a0
; RV32-NEXT: sltu a6, a4, a2
; RV32-NEXT: add a5, a5, a6
; RV32-NEXT: beq a5, a1, .LBB0_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a0, a5, a1
; RV32-NEXT: beqz a0, .LBB0_3
; RV32-NEXT: j .LBB0_4
; RV32-NEXT: .LBB0_2:
; RV32-NEXT: sltu a0, a4, a0
; RV32-NEXT: bnez a0, .LBB0_4
; RV32-NEXT: .LBB0_3:
; RV32-NEXT: li a2, 42
; RV32-NEXT: .LBB0_4:
; RV32-NEXT: neg a1, a0
; RV32-NEXT: and a1, a1, a3
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo1_overflow_used:
; RV64: # %bb.0:
; RV64-NEXT: add a2, a1, a0
; RV64-NEXT: bltu a2, a0, .LBB0_2
; RV64-NEXT: # %bb.1:
; RV64-NEXT: li a1, 42
; RV64-NEXT: .LBB0_2:
; RV64-NEXT: mv a0, a1
; RV64-NEXT: ret
%add = add i64 %b, %a
%cmp = icmp ult i64 %add, %a
%Q = select i1 %cmp, i64 %b, i64 42
ret i64 %Q
}
define i64 @uaddo1_math_overflow_used(i64 %a, i64 %b, ptr %res) nounwind ssp {
; RV32-LABEL: uaddo1_math_overflow_used:
; RV32: # %bb.0:
; RV32-NEXT: add a5, a3, a1
; RV32-NEXT: add a0, a2, a0
; RV32-NEXT: sltu a1, a0, a2
; RV32-NEXT: add a5, a5, a1
; RV32-NEXT: beq a5, a3, .LBB1_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a1, a5, a3
; RV32-NEXT: .LBB1_2:
; RV32-NEXT: bnez a1, .LBB1_4
; RV32-NEXT: # %bb.3:
; RV32-NEXT: li a2, 42
; RV32-NEXT: .LBB1_4:
; RV32-NEXT: neg a1, a1
; RV32-NEXT: and a1, a1, a3
; RV32-NEXT: sw a0, 0(a4)
; RV32-NEXT: sw a5, 4(a4)
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo1_math_overflow_used:
; RV64: # %bb.0:
; RV64-NEXT: add a0, a1, a0
; RV64-NEXT: bltu a0, a1, .LBB1_2
; RV64-NEXT: # %bb.1:
; RV64-NEXT: li a1, 42
; RV64-NEXT: .LBB1_2:
; RV64-NEXT: sd a0, 0(a2)
; RV64-NEXT: mv a0, a1
; RV64-NEXT: ret
%add = add i64 %b, %a
%cmp = icmp ult i64 %add, %a
%Q = select i1 %cmp, i64 %b, i64 42
store i64 %add, ptr %res
ret i64 %Q
}
define i64 @uaddo2_overflow_used(i64 %a, i64 %b) nounwind ssp {
; RV32-LABEL: uaddo2_overflow_used:
; RV32: # %bb.0:
; RV32-NEXT: add a1, a3, a1
; RV32-NEXT: add a0, a2, a0
; RV32-NEXT: sltu a0, a0, a2
; RV32-NEXT: add a1, a1, a0
; RV32-NEXT: beq a1, a3, .LBB2_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a0, a1, a3
; RV32-NEXT: .LBB2_2:
; RV32-NEXT: bnez a0, .LBB2_4
; RV32-NEXT: # %bb.3:
; RV32-NEXT: li a2, 42
; RV32-NEXT: .LBB2_4:
; RV32-NEXT: neg a1, a0
; RV32-NEXT: and a1, a1, a3
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo2_overflow_used:
; RV64: # %bb.0:
; RV64-NEXT: add a0, a1, a0
; RV64-NEXT: bltu a0, a1, .LBB2_2
; RV64-NEXT: # %bb.1:
; RV64-NEXT: li a1, 42
; RV64-NEXT: .LBB2_2:
; RV64-NEXT: mv a0, a1
; RV64-NEXT: ret
%add = add i64 %b, %a
%cmp = icmp ult i64 %add, %b
%Q = select i1 %cmp, i64 %b, i64 42
ret i64 %Q
}
define i64 @uaddo2_math_overflow_used(i64 %a, i64 %b, ptr %res) nounwind ssp {
; RV32-LABEL: uaddo2_math_overflow_used:
; RV32: # %bb.0:
; RV32-NEXT: add a5, a3, a1
; RV32-NEXT: add a0, a2, a0
; RV32-NEXT: sltu a1, a0, a2
; RV32-NEXT: add a5, a5, a1
; RV32-NEXT: beq a5, a3, .LBB3_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a1, a5, a3
; RV32-NEXT: .LBB3_2:
; RV32-NEXT: bnez a1, .LBB3_4
; RV32-NEXT: # %bb.3:
; RV32-NEXT: li a2, 42
; RV32-NEXT: .LBB3_4:
; RV32-NEXT: neg a1, a1
; RV32-NEXT: and a1, a1, a3
; RV32-NEXT: sw a0, 0(a4)
; RV32-NEXT: sw a5, 4(a4)
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo2_math_overflow_used:
; RV64: # %bb.0:
; RV64-NEXT: add a0, a1, a0
; RV64-NEXT: bltu a0, a1, .LBB3_2
; RV64-NEXT: # %bb.1:
; RV64-NEXT: li a1, 42
; RV64-NEXT: .LBB3_2:
; RV64-NEXT: sd a0, 0(a2)
; RV64-NEXT: mv a0, a1
; RV64-NEXT: ret
%add = add i64 %b, %a
%cmp = icmp ult i64 %add, %b
%Q = select i1 %cmp, i64 %b, i64 42
store i64 %add, ptr %res
ret i64 %Q
}
define i64 @uaddo3_overflow_used(i64 %a, i64 %b) nounwind ssp {
; RV32-LABEL: uaddo3_overflow_used:
; RV32: # %bb.0:
; RV32-NEXT: add a1, a3, a1
; RV32-NEXT: add a0, a2, a0
; RV32-NEXT: sltu a0, a0, a2
; RV32-NEXT: add a1, a1, a0
; RV32-NEXT: beq a3, a1, .LBB4_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a0, a1, a3
; RV32-NEXT: .LBB4_2:
; RV32-NEXT: bnez a0, .LBB4_4
; RV32-NEXT: # %bb.3:
; RV32-NEXT: li a2, 42
; RV32-NEXT: .LBB4_4:
; RV32-NEXT: neg a1, a0
; RV32-NEXT: and a1, a1, a3
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo3_overflow_used:
; RV64: # %bb.0:
; RV64-NEXT: add a0, a1, a0
; RV64-NEXT: bltu a0, a1, .LBB4_2
; RV64-NEXT: # %bb.1:
; RV64-NEXT: li a1, 42
; RV64-NEXT: .LBB4_2:
; RV64-NEXT: mv a0, a1
; RV64-NEXT: ret
%add = add i64 %b, %a
%cmp = icmp ugt i64 %b, %add
%Q = select i1 %cmp, i64 %b, i64 42
ret i64 %Q
}
define i64 @uaddo3_math_overflow_used(i64 %a, i64 %b, ptr %res) nounwind ssp {
; RV32-LABEL: uaddo3_math_overflow_used:
; RV32: # %bb.0:
; RV32-NEXT: add a5, a3, a1
; RV32-NEXT: add a0, a2, a0
; RV32-NEXT: sltu a1, a0, a2
; RV32-NEXT: add a5, a5, a1
; RV32-NEXT: beq a5, a3, .LBB5_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a1, a5, a3
; RV32-NEXT: .LBB5_2:
; RV32-NEXT: bnez a1, .LBB5_4
; RV32-NEXT: # %bb.3:
; RV32-NEXT: li a2, 42
; RV32-NEXT: .LBB5_4:
; RV32-NEXT: neg a1, a1
; RV32-NEXT: and a1, a1, a3
; RV32-NEXT: sw a0, 0(a4)
; RV32-NEXT: sw a5, 4(a4)
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo3_math_overflow_used:
; RV64: # %bb.0:
; RV64-NEXT: add a0, a1, a0
; RV64-NEXT: bltu a0, a1, .LBB5_2
; RV64-NEXT: # %bb.1:
; RV64-NEXT: li a1, 42
; RV64-NEXT: .LBB5_2:
; RV64-NEXT: sd a0, 0(a2)
; RV64-NEXT: mv a0, a1
; RV64-NEXT: ret
%add = add i64 %b, %a
%cmp = icmp ugt i64 %b, %add
%Q = select i1 %cmp, i64 %b, i64 42
store i64 %add, ptr %res
ret i64 %Q
}
; TODO? CGP sinks the compare before we have a chance to form the overflow intrinsic.
define i64 @uaddo4(i64 %a, i64 %b, i1 %c) nounwind ssp {
; RV32-LABEL: uaddo4:
; RV32: # %bb.0: # %entry
; RV32-NEXT: andi a4, a4, 1
; RV32-NEXT: beqz a4, .LBB6_6
; RV32-NEXT: # %bb.1: # %next
; RV32-NEXT: add a1, a3, a1
; RV32-NEXT: add a0, a2, a0
; RV32-NEXT: sltu a0, a0, a2
; RV32-NEXT: add a1, a1, a0
; RV32-NEXT: beq a3, a1, .LBB6_3
; RV32-NEXT: # %bb.2: # %next
; RV32-NEXT: sltu a0, a1, a3
; RV32-NEXT: .LBB6_3: # %next
; RV32-NEXT: bnez a0, .LBB6_5
; RV32-NEXT: # %bb.4: # %next
; RV32-NEXT: li a2, 42
; RV32-NEXT: .LBB6_5: # %next
; RV32-NEXT: neg a1, a0
; RV32-NEXT: and a1, a1, a3
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
; RV32-NEXT: .LBB6_6: # %exit
; RV32-NEXT: li a0, 0
; RV32-NEXT: li a1, 0
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo4:
; RV64: # %bb.0: # %entry
; RV64-NEXT: andi a2, a2, 1
; RV64-NEXT: beqz a2, .LBB6_4
; RV64-NEXT: # %bb.1: # %next
; RV64-NEXT: add a0, a1, a0
; RV64-NEXT: bltu a0, a1, .LBB6_3
; RV64-NEXT: # %bb.2: # %next
; RV64-NEXT: li a1, 42
; RV64-NEXT: .LBB6_3: # %next
; RV64-NEXT: mv a0, a1
; RV64-NEXT: ret
; RV64-NEXT: .LBB6_4: # %exit
; RV64-NEXT: li a0, 0
; RV64-NEXT: ret
entry:
%add = add i64 %b, %a
%cmp = icmp ugt i64 %b, %add
br i1 %c, label %next, label %exit
next:
%Q = select i1 %cmp, i64 %b, i64 42
ret i64 %Q
exit:
ret i64 0
}
define i64 @uaddo5(i64 %a, i64 %b, ptr %ptr, i1 %c) nounwind ssp {
; RV32-LABEL: uaddo5:
; RV32: # %bb.0: # %entry
; RV32-NEXT: andi a5, a5, 1
; RV32-NEXT: add a1, a3, a1
; RV32-NEXT: add a6, a2, a0
; RV32-NEXT: sltu a0, a6, a2
; RV32-NEXT: add a1, a1, a0
; RV32-NEXT: sw a6, 0(a4)
; RV32-NEXT: sw a1, 4(a4)
; RV32-NEXT: beqz a5, .LBB7_6
; RV32-NEXT: # %bb.1: # %next
; RV32-NEXT: beq a3, a1, .LBB7_3
; RV32-NEXT: # %bb.2: # %next
; RV32-NEXT: sltu a0, a1, a3
; RV32-NEXT: .LBB7_3: # %next
; RV32-NEXT: bnez a0, .LBB7_5
; RV32-NEXT: # %bb.4: # %next
; RV32-NEXT: li a2, 42
; RV32-NEXT: .LBB7_5: # %next
; RV32-NEXT: neg a1, a0
; RV32-NEXT: and a1, a1, a3
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
; RV32-NEXT: .LBB7_6: # %exit
; RV32-NEXT: li a0, 0
; RV32-NEXT: li a1, 0
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo5:
; RV64: # %bb.0: # %entry
; RV64-NEXT: andi a3, a3, 1
; RV64-NEXT: add a0, a1, a0
; RV64-NEXT: sd a0, 0(a2)
; RV64-NEXT: beqz a3, .LBB7_4
; RV64-NEXT: # %bb.1: # %next
; RV64-NEXT: bltu a0, a1, .LBB7_3
; RV64-NEXT: # %bb.2: # %next
; RV64-NEXT: li a1, 42
; RV64-NEXT: .LBB7_3: # %next
; RV64-NEXT: mv a0, a1
; RV64-NEXT: ret
; RV64-NEXT: .LBB7_4: # %exit
; RV64-NEXT: li a0, 0
; RV64-NEXT: ret
entry:
%add = add i64 %b, %a
store i64 %add, ptr %ptr
%cmp = icmp ugt i64 %b, %add
br i1 %c, label %next, label %exit
next:
%Q = select i1 %cmp, i64 %b, i64 42
ret i64 %Q
exit:
ret i64 0
}
; Instcombine folds (a + b <u a) to (a ^ -1 <u b). Make sure we match this
; pattern as well.
define i64 @uaddo6_xor(i64 %a, i64 %b) {
; RV32-LABEL: uaddo6_xor:
; RV32: # %bb.0:
; RV32-NEXT: not a1, a1
; RV32-NEXT: beq a1, a3, .LBB8_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a0, a1, a3
; RV32-NEXT: beqz a0, .LBB8_3
; RV32-NEXT: j .LBB8_4
; RV32-NEXT: .LBB8_2:
; RV32-NEXT: not a0, a0
; RV32-NEXT: sltu a0, a0, a2
; RV32-NEXT: bnez a0, .LBB8_4
; RV32-NEXT: .LBB8_3:
; RV32-NEXT: li a2, 42
; RV32-NEXT: .LBB8_4:
; RV32-NEXT: neg a1, a0
; RV32-NEXT: and a1, a1, a3
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo6_xor:
; RV64: # %bb.0:
; RV64-NEXT: not a2, a0
; RV64-NEXT: mv a0, a1
; RV64-NEXT: bltu a2, a1, .LBB8_2
; RV64-NEXT: # %bb.1:
; RV64-NEXT: li a0, 42
; RV64-NEXT: .LBB8_2:
; RV64-NEXT: ret
%x = xor i64 %a, -1
%cmp = icmp ult i64 %x, %b
%Q = select i1 %cmp, i64 %b, i64 42
ret i64 %Q
}
define i64 @uaddo6_xor_commuted(i64 %a, i64 %b) {
; RV32-LABEL: uaddo6_xor_commuted:
; RV32: # %bb.0:
; RV32-NEXT: not a1, a1
; RV32-NEXT: beq a1, a3, .LBB9_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a0, a1, a3
; RV32-NEXT: beqz a0, .LBB9_3
; RV32-NEXT: j .LBB9_4
; RV32-NEXT: .LBB9_2:
; RV32-NEXT: not a0, a0
; RV32-NEXT: sltu a0, a0, a2
; RV32-NEXT: bnez a0, .LBB9_4
; RV32-NEXT: .LBB9_3:
; RV32-NEXT: li a2, 42
; RV32-NEXT: .LBB9_4:
; RV32-NEXT: neg a1, a0
; RV32-NEXT: and a1, a1, a3
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo6_xor_commuted:
; RV64: # %bb.0:
; RV64-NEXT: not a2, a0
; RV64-NEXT: mv a0, a1
; RV64-NEXT: bltu a2, a1, .LBB9_2
; RV64-NEXT: # %bb.1:
; RV64-NEXT: li a0, 42
; RV64-NEXT: .LBB9_2:
; RV64-NEXT: ret
%x = xor i64 %a, -1
%cmp = icmp ult i64 %x, %b
%Q = select i1 %cmp, i64 %b, i64 42
ret i64 %Q
}
declare void @use(i64)
define i64 @uaddo6_xor_multi_use(i64 %a, i64 %b) {
; RV32-LABEL: uaddo6_xor_multi_use:
; RV32: # %bb.0:
; RV32-NEXT: addi sp, sp, -16
; RV32-NEXT: .cfi_def_cfa_offset 16
; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
; RV32-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
; RV32-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
; RV32-NEXT: .cfi_offset ra, -4
; RV32-NEXT: .cfi_offset s0, -8
; RV32-NEXT: .cfi_offset s1, -12
; RV32-NEXT: mv s0, a2
; RV32-NEXT: not a1, a1
; RV32-NEXT: not a0, a0
; RV32-NEXT: beq a1, a3, .LBB10_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a2, a1, a3
; RV32-NEXT: beqz a2, .LBB10_3
; RV32-NEXT: j .LBB10_4
; RV32-NEXT: .LBB10_2:
; RV32-NEXT: sltu a2, a0, s0
; RV32-NEXT: bnez a2, .LBB10_4
; RV32-NEXT: .LBB10_3:
; RV32-NEXT: li s0, 42
; RV32-NEXT: .LBB10_4:
; RV32-NEXT: neg s1, a2
; RV32-NEXT: and s1, s1, a3
; RV32-NEXT: call use
; RV32-NEXT: mv a0, s0
; RV32-NEXT: mv a1, s1
; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
; RV32-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
; RV32-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
; RV32-NEXT: addi sp, sp, 16
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo6_xor_multi_use:
; RV64: # %bb.0:
; RV64-NEXT: addi sp, sp, -16
; RV64-NEXT: .cfi_def_cfa_offset 16
; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
; RV64-NEXT: sd s0, 0(sp) # 8-byte Folded Spill
; RV64-NEXT: .cfi_offset ra, -8
; RV64-NEXT: .cfi_offset s0, -16
; RV64-NEXT: not a0, a0
; RV64-NEXT: mv s0, a1
; RV64-NEXT: bltu a0, a1, .LBB10_2
; RV64-NEXT: # %bb.1:
; RV64-NEXT: li s0, 42
; RV64-NEXT: .LBB10_2:
; RV64-NEXT: call use
; RV64-NEXT: mv a0, s0
; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
; RV64-NEXT: ld s0, 0(sp) # 8-byte Folded Reload
; RV64-NEXT: addi sp, sp, 16
; RV64-NEXT: ret
%x = xor i64 -1, %a
%cmp = icmp ult i64 %x, %b
%Q = select i1 %cmp, i64 %b, i64 42
call void @use(i64 %x)
ret i64 %Q
}
; Make sure we do not use the XOR binary operator as insert point, as it may
; come before the second operand of the overflow intrinsic.
define i1 @uaddo6_xor_op_after_XOR(i32 %a, ptr %b.ptr) {
; RV32-LABEL: uaddo6_xor_op_after_XOR:
; RV32: # %bb.0:
; RV32-NEXT: lw a1, 0(a1)
; RV32-NEXT: not a0, a0
; RV32-NEXT: sltu a0, a0, a1
; RV32-NEXT: xori a0, a0, 1
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo6_xor_op_after_XOR:
; RV64: # %bb.0:
; RV64-NEXT: lw a1, 0(a1)
; RV64-NEXT: not a0, a0
; RV64-NEXT: sext.w a0, a0
; RV64-NEXT: sltu a0, a0, a1
; RV64-NEXT: xori a0, a0, 1
; RV64-NEXT: ret
%x = xor i32 %a, -1
%b = load i32, ptr %b.ptr, align 8
%cmp14 = icmp ugt i32 %b, %x
%ov = xor i1 %cmp14, true
ret i1 %ov
}
; When adding 1, the general pattern for add-overflow may be different due to icmp canonicalization.
; PR31754: https://bugs.llvm.org/show_bug.cgi?id=31754
define i1 @uaddo_i64_increment(i64 %x, ptr %p) {
; RV32-LABEL: uaddo_i64_increment:
; RV32: # %bb.0:
; RV32-NEXT: addi a3, a0, 1
; RV32-NEXT: seqz a0, a3
; RV32-NEXT: add a1, a1, a0
; RV32-NEXT: or a0, a3, a1
; RV32-NEXT: seqz a0, a0
; RV32-NEXT: sw a3, 0(a2)
; RV32-NEXT: sw a1, 4(a2)
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo_i64_increment:
; RV64: # %bb.0:
; RV64-NEXT: addi a2, a0, 1
; RV64-NEXT: seqz a0, a2
; RV64-NEXT: sd a2, 0(a1)
; RV64-NEXT: ret
%a = add i64 %x, 1
%ov = icmp eq i64 %a, 0
store i64 %a, ptr %p
ret i1 %ov
}
define i1 @uaddo_i8_increment_noncanonical_1(i8 %x, ptr %p) {
; RV32-LABEL: uaddo_i8_increment_noncanonical_1:
; RV32: # %bb.0:
; RV32-NEXT: addi a2, a0, 1
; RV32-NEXT: andi a0, a2, 255
; RV32-NEXT: seqz a0, a0
; RV32-NEXT: sb a2, 0(a1)
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo_i8_increment_noncanonical_1:
; RV64: # %bb.0:
; RV64-NEXT: addi a2, a0, 1
; RV64-NEXT: andi a0, a2, 255
; RV64-NEXT: seqz a0, a0
; RV64-NEXT: sb a2, 0(a1)
; RV64-NEXT: ret
%a = add i8 1, %x ; commute
%ov = icmp eq i8 %a, 0
store i8 %a, ptr %p
ret i1 %ov
}
define i1 @uaddo_i32_increment_noncanonical_2(i32 %x, ptr %p) {
; RV32-LABEL: uaddo_i32_increment_noncanonical_2:
; RV32: # %bb.0:
; RV32-NEXT: addi a2, a0, 1
; RV32-NEXT: seqz a0, a2
; RV32-NEXT: sw a2, 0(a1)
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo_i32_increment_noncanonical_2:
; RV64: # %bb.0:
; RV64-NEXT: addiw a2, a0, 1
; RV64-NEXT: seqz a0, a2
; RV64-NEXT: sw a2, 0(a1)
; RV64-NEXT: ret
%a = add i32 %x, 1
%ov = icmp eq i32 0, %a ; commute
store i32 %a, ptr %p
ret i1 %ov
}
define i1 @uaddo_i16_increment_noncanonical_3(i16 %x, ptr %p) {
; RV32-LABEL: uaddo_i16_increment_noncanonical_3:
; RV32: # %bb.0:
; RV32-NEXT: addi a2, a0, 1
; RV32-NEXT: slli a0, a2, 16
; RV32-NEXT: srli a0, a0, 16
; RV32-NEXT: seqz a0, a0
; RV32-NEXT: sh a2, 0(a1)
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo_i16_increment_noncanonical_3:
; RV64: # %bb.0:
; RV64-NEXT: addi a2, a0, 1
; RV64-NEXT: slli a0, a2, 48
; RV64-NEXT: srli a0, a0, 48
; RV64-NEXT: seqz a0, a0
; RV64-NEXT: sh a2, 0(a1)
; RV64-NEXT: ret
%a = add i16 1, %x ; commute
%ov = icmp eq i16 0, %a ; commute
store i16 %a, ptr %p
ret i1 %ov
}
; The overflow check may be against the input rather than the sum.
define i1 @uaddo_i64_increment_alt(i64 %x, ptr %p) {
; RV32-LABEL: uaddo_i64_increment_alt:
; RV32: # %bb.0:
; RV32-NEXT: addi a3, a0, 1
; RV32-NEXT: seqz a0, a3
; RV32-NEXT: add a1, a1, a0
; RV32-NEXT: or a0, a3, a1
; RV32-NEXT: seqz a0, a0
; RV32-NEXT: sw a3, 0(a2)
; RV32-NEXT: sw a1, 4(a2)
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo_i64_increment_alt:
; RV64: # %bb.0:
; RV64-NEXT: addi a2, a0, 1
; RV64-NEXT: seqz a0, a2
; RV64-NEXT: sd a2, 0(a1)
; RV64-NEXT: ret
%a = add i64 %x, 1
store i64 %a, ptr %p
%ov = icmp eq i64 %x, -1
ret i1 %ov
}
; Make sure insertion is done correctly based on dominance.
define i1 @uaddo_i64_increment_alt_dom(i64 %x, ptr %p) {
; RV32-LABEL: uaddo_i64_increment_alt_dom:
; RV32: # %bb.0:
; RV32-NEXT: addi a3, a0, 1
; RV32-NEXT: seqz a0, a3
; RV32-NEXT: add a1, a1, a0
; RV32-NEXT: or a0, a3, a1
; RV32-NEXT: seqz a0, a0
; RV32-NEXT: sw a3, 0(a2)
; RV32-NEXT: sw a1, 4(a2)
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo_i64_increment_alt_dom:
; RV64: # %bb.0:
; RV64-NEXT: addi a2, a0, 1
; RV64-NEXT: seqz a0, a2
; RV64-NEXT: sd a2, 0(a1)
; RV64-NEXT: ret
%ov = icmp eq i64 %x, -1
%a = add i64 %x, 1
store i64 %a, ptr %p
ret i1 %ov
}
; The overflow check may be against the input rather than the sum.
define i1 @uaddo_i32_decrement_alt(i32 signext %x, ptr %p) {
; RV32-LABEL: uaddo_i32_decrement_alt:
; RV32: # %bb.0:
; RV32-NEXT: snez a2, a0
; RV32-NEXT: addi a0, a0, -1
; RV32-NEXT: sw a0, 0(a1)
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo_i32_decrement_alt:
; RV64: # %bb.0:
; RV64-NEXT: snez a2, a0
; RV64-NEXT: addi a0, a0, -1
; RV64-NEXT: sw a0, 0(a1)
; RV64-NEXT: mv a0, a2
; RV64-NEXT: ret
%a = add i32 %x, -1
store i32 %a, ptr %p
%ov = icmp ne i32 %x, 0
ret i1 %ov
}
define i1 @uaddo_i64_decrement_alt(i64 %x, ptr %p) {
; RV32-LABEL: uaddo_i64_decrement_alt:
; RV32: # %bb.0:
; RV32-NEXT: or a3, a0, a1
; RV32-NEXT: snez a3, a3
; RV32-NEXT: seqz a4, a0
; RV32-NEXT: sub a1, a1, a4
; RV32-NEXT: addi a0, a0, -1
; RV32-NEXT: sw a0, 0(a2)
; RV32-NEXT: sw a1, 4(a2)
; RV32-NEXT: mv a0, a3
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo_i64_decrement_alt:
; RV64: # %bb.0:
; RV64-NEXT: snez a2, a0
; RV64-NEXT: addi a0, a0, -1
; RV64-NEXT: sd a0, 0(a1)
; RV64-NEXT: mv a0, a2
; RV64-NEXT: ret
%a = add i64 %x, -1
store i64 %a, ptr %p
%ov = icmp ne i64 %x, 0
ret i1 %ov
}
; Make sure insertion is done correctly based on dominance.
define i1 @uaddo_i64_decrement_alt_dom(i64 %x, ptr %p) {
; RV32-LABEL: uaddo_i64_decrement_alt_dom:
; RV32: # %bb.0:
; RV32-NEXT: or a3, a0, a1
; RV32-NEXT: snez a3, a3
; RV32-NEXT: seqz a4, a0
; RV32-NEXT: sub a1, a1, a4
; RV32-NEXT: addi a0, a0, -1
; RV32-NEXT: sw a0, 0(a2)
; RV32-NEXT: sw a1, 4(a2)
; RV32-NEXT: mv a0, a3
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo_i64_decrement_alt_dom:
; RV64: # %bb.0:
; RV64-NEXT: snez a2, a0
; RV64-NEXT: addi a0, a0, -1
; RV64-NEXT: sd a0, 0(a1)
; RV64-NEXT: mv a0, a2
; RV64-NEXT: ret
%ov = icmp ne i64 %x, 0
%a = add i64 %x, -1
store i64 %a, ptr %p
ret i1 %ov
}
; No transform for illegal types.
define i1 @uaddo_i42_increment_illegal_type(i42 %x, ptr %p) {
; RV32-LABEL: uaddo_i42_increment_illegal_type:
; RV32: # %bb.0:
; RV32-NEXT: addi a3, a0, 1
; RV32-NEXT: seqz a0, a3
; RV32-NEXT: add a0, a1, a0
; RV32-NEXT: andi a1, a0, 1023
; RV32-NEXT: or a0, a3, a1
; RV32-NEXT: seqz a0, a0
; RV32-NEXT: sw a3, 0(a2)
; RV32-NEXT: sh a1, 4(a2)
; RV32-NEXT: ret
;
; RV64-LABEL: uaddo_i42_increment_illegal_type:
; RV64: # %bb.0:
; RV64-NEXT: addi a2, a0, 1
; RV64-NEXT: slli a0, a2, 22
; RV64-NEXT: srli a3, a0, 22
; RV64-NEXT: seqz a0, a3
; RV64-NEXT: sw a2, 0(a1)
; RV64-NEXT: srli a3, a3, 32
; RV64-NEXT: sh a3, 4(a1)
; RV64-NEXT: ret
%a = add i42 %x, 1
%ov = icmp eq i42 %a, 0
store i42 %a, ptr %p
ret i1 %ov
}
define i1 @usubo_ult_i64_overflow_used(i64 %x, i64 %y, ptr %p) {
; RV32-LABEL: usubo_ult_i64_overflow_used:
; RV32: # %bb.0:
; RV32-NEXT: beq a1, a3, .LBB22_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a0, a1, a3
; RV32-NEXT: ret
; RV32-NEXT: .LBB22_2:
; RV32-NEXT: sltu a0, a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_ult_i64_overflow_used:
; RV64: # %bb.0:
; RV64-NEXT: sltu a0, a0, a1
; RV64-NEXT: ret
%s = sub i64 %x, %y
%ov = icmp ult i64 %x, %y
ret i1 %ov
}
define i1 @usubo_ult_i64_math_overflow_used(i64 %x, i64 %y, ptr %p) {
; RV32-LABEL: usubo_ult_i64_math_overflow_used:
; RV32: # %bb.0:
; RV32-NEXT: mv a5, a0
; RV32-NEXT: sltu a0, a0, a2
; RV32-NEXT: sub a6, a1, a3
; RV32-NEXT: sub a6, a6, a0
; RV32-NEXT: sub a5, a5, a2
; RV32-NEXT: sw a5, 0(a4)
; RV32-NEXT: sw a6, 4(a4)
; RV32-NEXT: beq a1, a3, .LBB23_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: sltu a0, a1, a3
; RV32-NEXT: .LBB23_2:
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_ult_i64_math_overflow_used:
; RV64: # %bb.0:
; RV64-NEXT: sub a3, a0, a1
; RV64-NEXT: sltu a0, a0, a1
; RV64-NEXT: sd a3, 0(a2)
; RV64-NEXT: ret
%s = sub i64 %x, %y
store i64 %s, ptr %p
%ov = icmp ult i64 %x, %y
ret i1 %ov
}
; Verify insertion point for single-BB. Toggle predicate.
define i1 @usubo_ugt_i32(i32 %x, i32 %y, ptr %p) {
; RV32-LABEL: usubo_ugt_i32:
; RV32: # %bb.0:
; RV32-NEXT: sltu a3, a0, a1
; RV32-NEXT: sub a0, a0, a1
; RV32-NEXT: sw a0, 0(a2)
; RV32-NEXT: mv a0, a3
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_ugt_i32:
; RV64: # %bb.0:
; RV64-NEXT: sext.w a3, a1
; RV64-NEXT: sext.w a4, a0
; RV64-NEXT: sltu a3, a4, a3
; RV64-NEXT: subw a0, a0, a1
; RV64-NEXT: sw a0, 0(a2)
; RV64-NEXT: mv a0, a3
; RV64-NEXT: ret
%ov = icmp ugt i32 %y, %x
%s = sub i32 %x, %y
store i32 %s, ptr %p
ret i1 %ov
}
; Constant operand should match.
define i1 @usubo_ugt_constant_op0_i8(i8 %x, ptr %p) {
; RV32-LABEL: usubo_ugt_constant_op0_i8:
; RV32: # %bb.0:
; RV32-NEXT: andi a2, a0, 255
; RV32-NEXT: li a3, 42
; RV32-NEXT: sub a3, a3, a0
; RV32-NEXT: sltiu a0, a2, 43
; RV32-NEXT: xori a0, a0, 1
; RV32-NEXT: sb a3, 0(a1)
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_ugt_constant_op0_i8:
; RV64: # %bb.0:
; RV64-NEXT: andi a2, a0, 255
; RV64-NEXT: li a3, 42
; RV64-NEXT: subw a3, a3, a0
; RV64-NEXT: sltiu a0, a2, 43
; RV64-NEXT: xori a0, a0, 1
; RV64-NEXT: sb a3, 0(a1)
; RV64-NEXT: ret
%s = sub i8 42, %x
%ov = icmp ugt i8 %x, 42
store i8 %s, ptr %p
ret i1 %ov
}
; Compare with constant operand 0 is canonicalized by commuting, but verify match for non-canonical form.
define i1 @usubo_ult_constant_op0_i16(i16 %x, ptr %p) {
; RV32-LABEL: usubo_ult_constant_op0_i16:
; RV32: # %bb.0:
; RV32-NEXT: slli a2, a0, 16
; RV32-NEXT: srli a2, a2, 16
; RV32-NEXT: li a3, 43
; RV32-NEXT: sub a3, a3, a0
; RV32-NEXT: sltiu a0, a2, 44
; RV32-NEXT: xori a0, a0, 1
; RV32-NEXT: sh a3, 0(a1)
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_ult_constant_op0_i16:
; RV64: # %bb.0:
; RV64-NEXT: slli a2, a0, 48
; RV64-NEXT: srli a2, a2, 48
; RV64-NEXT: li a3, 43
; RV64-NEXT: subw a3, a3, a0
; RV64-NEXT: sltiu a0, a2, 44
; RV64-NEXT: xori a0, a0, 1
; RV64-NEXT: sh a3, 0(a1)
; RV64-NEXT: ret
%s = sub i16 43, %x
%ov = icmp ult i16 43, %x
store i16 %s, ptr %p
ret i1 %ov
}
; Subtract with constant operand 1 is canonicalized to add.
define i1 @usubo_ult_constant_op1_i16(i16 %x, ptr %p) {
; RV32-LABEL: usubo_ult_constant_op1_i16:
; RV32: # %bb.0:
; RV32-NEXT: slli a2, a0, 16
; RV32-NEXT: srli a2, a2, 16
; RV32-NEXT: addi a3, a0, -44
; RV32-NEXT: sltiu a0, a2, 44
; RV32-NEXT: sh a3, 0(a1)
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_ult_constant_op1_i16:
; RV64: # %bb.0:
; RV64-NEXT: slli a2, a0, 48
; RV64-NEXT: srli a2, a2, 48
; RV64-NEXT: addi a3, a0, -44
; RV64-NEXT: sltiu a0, a2, 44
; RV64-NEXT: sh a3, 0(a1)
; RV64-NEXT: ret
%s = add i16 %x, -44
%ov = icmp ult i16 %x, 44
store i16 %s, ptr %p
ret i1 %ov
}
define i1 @usubo_ugt_constant_op1_i8(i8 %x, ptr %p) {
; RV32-LABEL: usubo_ugt_constant_op1_i8:
; RV32: # %bb.0:
; RV32-NEXT: andi a2, a0, 255
; RV32-NEXT: sltiu a2, a2, 45
; RV32-NEXT: addi a0, a0, -45
; RV32-NEXT: sb a0, 0(a1)
; RV32-NEXT: mv a0, a2
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_ugt_constant_op1_i8:
; RV64: # %bb.0:
; RV64-NEXT: andi a2, a0, 255
; RV64-NEXT: sltiu a2, a2, 45
; RV64-NEXT: addi a0, a0, -45
; RV64-NEXT: sb a0, 0(a1)
; RV64-NEXT: mv a0, a2
; RV64-NEXT: ret
%ov = icmp ugt i8 45, %x
%s = add i8 %x, -45
store i8 %s, ptr %p
ret i1 %ov
}
; Special-case: subtract 1 changes the compare predicate and constant.
define i1 @usubo_eq_constant1_op1_i32(i32 %x, ptr %p) {
; RV32-LABEL: usubo_eq_constant1_op1_i32:
; RV32: # %bb.0:
; RV32-NEXT: addi a2, a0, -1
; RV32-NEXT: seqz a0, a0
; RV32-NEXT: sw a2, 0(a1)
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_eq_constant1_op1_i32:
; RV64: # %bb.0:
; RV64-NEXT: sext.w a2, a0
; RV64-NEXT: addi a3, a0, -1
; RV64-NEXT: seqz a0, a2
; RV64-NEXT: sw a3, 0(a1)
; RV64-NEXT: ret
%s = add i32 %x, -1
%ov = icmp eq i32 %x, 0
store i32 %s, ptr %p
ret i1 %ov
}
; Special-case: subtract from 0 (negate) changes the compare predicate.
define i1 @usubo_ne_constant0_op1_i32(i32 %x, ptr %p) {
; RV32-LABEL: usubo_ne_constant0_op1_i32:
; RV32: # %bb.0:
; RV32-NEXT: neg a2, a0
; RV32-NEXT: snez a0, a0
; RV32-NEXT: sw a2, 0(a1)
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_ne_constant0_op1_i32:
; RV64: # %bb.0:
; RV64-NEXT: sext.w a2, a0
; RV64-NEXT: negw a3, a0
; RV64-NEXT: snez a0, a2
; RV64-NEXT: sw a3, 0(a1)
; RV64-NEXT: ret
%s = sub i32 0, %x
%ov = icmp ne i32 %x, 0
store i32 %s, ptr %p
ret i1 %ov
}
; This used to verify insertion point for multi-BB, but now we just bail out.
declare void @call(i1)
define i1 @usubo_ult_sub_dominates_i64(i64 %x, i64 %y, ptr %p, i1 %cond) {
; RV32-LABEL: usubo_ult_sub_dominates_i64:
; RV32: # %bb.0: # %entry
; RV32-NEXT: andi a6, a5, 1
; RV32-NEXT: beqz a6, .LBB31_5
; RV32-NEXT: # %bb.1: # %t
; RV32-NEXT: mv a7, a0
; RV32-NEXT: sltu a0, a0, a2
; RV32-NEXT: sub t0, a1, a3
; RV32-NEXT: sub t0, t0, a0
; RV32-NEXT: sub a2, a7, a2
; RV32-NEXT: sw a2, 0(a4)
; RV32-NEXT: sw t0, 4(a4)
; RV32-NEXT: beqz a6, .LBB31_5
; RV32-NEXT: # %bb.2: # %end
; RV32-NEXT: beq a1, a3, .LBB31_4
; RV32-NEXT: # %bb.3: # %end
; RV32-NEXT: sltu a0, a1, a3
; RV32-NEXT: .LBB31_4: # %end
; RV32-NEXT: ret
; RV32-NEXT: .LBB31_5: # %f
; RV32-NEXT: mv a0, a5
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_ult_sub_dominates_i64:
; RV64: # %bb.0: # %entry
; RV64-NEXT: andi a4, a3, 1
; RV64-NEXT: beqz a4, .LBB31_3
; RV64-NEXT: # %bb.1: # %t
; RV64-NEXT: sub a5, a0, a1
; RV64-NEXT: sd a5, 0(a2)
; RV64-NEXT: beqz a4, .LBB31_3
; RV64-NEXT: # %bb.2: # %end
; RV64-NEXT: sltu a0, a0, a1
; RV64-NEXT: ret
; RV64-NEXT: .LBB31_3: # %f
; RV64-NEXT: mv a0, a3
; RV64-NEXT: ret
entry:
br i1 %cond, label %t, label %f
t:
%s = sub i64 %x, %y
store i64 %s, ptr %p
br i1 %cond, label %end, label %f
f:
ret i1 %cond
end:
%ov = icmp ult i64 %x, %y
ret i1 %ov
}
define i1 @usubo_ult_cmp_dominates_i64(i64 %x, i64 %y, ptr %p, i1 %cond) {
; RV32-LABEL: usubo_ult_cmp_dominates_i64:
; RV32: # %bb.0: # %entry
; RV32-NEXT: addi sp, sp, -32
; RV32-NEXT: .cfi_def_cfa_offset 32
; RV32-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
; RV32-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
; RV32-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
; RV32-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
; RV32-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
; RV32-NEXT: sw s4, 8(sp) # 4-byte Folded Spill
; RV32-NEXT: sw s5, 4(sp) # 4-byte Folded Spill
; RV32-NEXT: sw s6, 0(sp) # 4-byte Folded Spill
; RV32-NEXT: .cfi_offset ra, -4
; RV32-NEXT: .cfi_offset s0, -8
; RV32-NEXT: .cfi_offset s1, -12
; RV32-NEXT: .cfi_offset s2, -16
; RV32-NEXT: .cfi_offset s3, -20
; RV32-NEXT: .cfi_offset s4, -24
; RV32-NEXT: .cfi_offset s5, -28
; RV32-NEXT: .cfi_offset s6, -32
; RV32-NEXT: mv s2, a5
; RV32-NEXT: andi a5, a5, 1
; RV32-NEXT: beqz a5, .LBB32_8
; RV32-NEXT: # %bb.1: # %t
; RV32-NEXT: mv s0, a4
; RV32-NEXT: mv s3, a3
; RV32-NEXT: mv s1, a2
; RV32-NEXT: mv s5, a1
; RV32-NEXT: mv s4, a0
; RV32-NEXT: beq a1, a3, .LBB32_3
; RV32-NEXT: # %bb.2: # %t
; RV32-NEXT: sltu s6, s5, s3
; RV32-NEXT: j .LBB32_4
; RV32-NEXT: .LBB32_3:
; RV32-NEXT: sltu s6, s4, s1
; RV32-NEXT: .LBB32_4: # %t
; RV32-NEXT: mv a0, s6
; RV32-NEXT: call call
; RV32-NEXT: beqz s6, .LBB32_8
; RV32-NEXT: # %bb.5: # %end
; RV32-NEXT: sltu a1, s4, s1
; RV32-NEXT: mv a0, a1
; RV32-NEXT: beq s5, s3, .LBB32_7
; RV32-NEXT: # %bb.6: # %end
; RV32-NEXT: sltu a0, s5, s3
; RV32-NEXT: .LBB32_7: # %end
; RV32-NEXT: sub a2, s5, s3
; RV32-NEXT: sub a2, a2, a1
; RV32-NEXT: sub a1, s4, s1
; RV32-NEXT: sw a1, 0(s0)
; RV32-NEXT: sw a2, 4(s0)
; RV32-NEXT: j .LBB32_9
; RV32-NEXT: .LBB32_8: # %f
; RV32-NEXT: mv a0, s2
; RV32-NEXT: .LBB32_9: # %f
; RV32-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
; RV32-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
; RV32-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
; RV32-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
; RV32-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
; RV32-NEXT: lw s4, 8(sp) # 4-byte Folded Reload
; RV32-NEXT: lw s5, 4(sp) # 4-byte Folded Reload
; RV32-NEXT: lw s6, 0(sp) # 4-byte Folded Reload
; RV32-NEXT: addi sp, sp, 32
; RV32-NEXT: ret
;
; RV64-LABEL: usubo_ult_cmp_dominates_i64:
; RV64: # %bb.0: # %entry
; RV64-NEXT: addi sp, sp, -48
; RV64-NEXT: .cfi_def_cfa_offset 48
; RV64-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
; RV64-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
; RV64-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
; RV64-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
; RV64-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
; RV64-NEXT: sd s4, 0(sp) # 8-byte Folded Spill
; RV64-NEXT: .cfi_offset ra, -8
; RV64-NEXT: .cfi_offset s0, -16
; RV64-NEXT: .cfi_offset s1, -24
; RV64-NEXT: .cfi_offset s2, -32
; RV64-NEXT: .cfi_offset s3, -40
; RV64-NEXT: .cfi_offset s4, -48
; RV64-NEXT: mv s0, a3
; RV64-NEXT: andi a3, a3, 1
; RV64-NEXT: beqz a3, .LBB32_3
; RV64-NEXT: # %bb.1: # %t
; RV64-NEXT: mv s1, a2
; RV64-NEXT: mv s2, a1
; RV64-NEXT: mv s3, a0
; RV64-NEXT: sltu s4, a0, a1
; RV64-NEXT: mv a0, s4
; RV64-NEXT: call call
; RV64-NEXT: bgeu s3, s2, .LBB32_3
; RV64-NEXT: # %bb.2: # %end
; RV64-NEXT: sub a0, s3, s2
; RV64-NEXT: sd a0, 0(s1)
; RV64-NEXT: mv a0, s4
; RV64-NEXT: j .LBB32_4
; RV64-NEXT: .LBB32_3: # %f
; RV64-NEXT: mv a0, s0
; RV64-NEXT: .LBB32_4: # %f
; RV64-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
; RV64-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
; RV64-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
; RV64-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
; RV64-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
; RV64-NEXT: ld s4, 0(sp) # 8-byte Folded Reload
; RV64-NEXT: addi sp, sp, 48
; RV64-NEXT: ret
entry:
br i1 %cond, label %t, label %f
t:
%ov = icmp ult i64 %x, %y
call void @call(i1 %ov)
br i1 %ov, label %end, label %f
f:
ret i1 %cond
end:
%s = sub i64 %x, %y
store i64 %s, ptr %p
ret i1 %ov
}
; Verify that crazy/non-canonical code does not crash.
define void @bar() {
; RV32-LABEL: bar:
; RV32: # %bb.0:
;
; RV64-LABEL: bar:
; RV64: # %bb.0:
%cmp = icmp eq i64 1, -1
%frombool = zext i1 %cmp to i8
unreachable
}
define void @foo() {
; RV32-LABEL: foo:
; RV32: # %bb.0:
;
; RV64-LABEL: foo:
; RV64: # %bb.0:
%sub = add nsw i64 1, 1
%conv = trunc i64 %sub to i32
unreachable
}
; Similarly for usubo.
define i1 @bar2() {
; RV32-LABEL: bar2:
; RV32: # %bb.0:
; RV32-NEXT: li a0, 0
; RV32-NEXT: ret
;
; RV64-LABEL: bar2:
; RV64: # %bb.0:
; RV64-NEXT: li a0, 0
; RV64-NEXT: ret
%cmp = icmp eq i64 1, 0
ret i1 %cmp
}
define i64 @foo2(ptr %p) {
; RV32-LABEL: foo2:
; RV32: # %bb.0:
; RV32-NEXT: li a0, 0
; RV32-NEXT: li a1, 0
; RV32-NEXT: ret
;
; RV64-LABEL: foo2:
; RV64: # %bb.0:
; RV64-NEXT: li a0, 0
; RV64-NEXT: ret
%sub = add nsw i64 1, -1
ret i64 %sub
}
; Avoid hoisting a math op into a dominating block which would
; increase the critical path.
define void @PR41129(ptr %p64) {
; RV32-LABEL: PR41129:
; RV32: # %bb.0: # %entry
; RV32-NEXT: lw a2, 4(a0)
; RV32-NEXT: lw a1, 0(a0)
; RV32-NEXT: or a3, a1, a2
; RV32-NEXT: beqz a3, .LBB37_2
; RV32-NEXT: # %bb.1: # %false
; RV32-NEXT: andi a1, a1, 7
; RV32-NEXT: sw zero, 4(a0)
; RV32-NEXT: sw a1, 0(a0)
; RV32-NEXT: ret
; RV32-NEXT: .LBB37_2: # %true
; RV32-NEXT: seqz a3, a1
; RV32-NEXT: sub a2, a2, a3
; RV32-NEXT: addi a1, a1, -1
; RV32-NEXT: sw a1, 0(a0)
; RV32-NEXT: sw a2, 4(a0)
; RV32-NEXT: ret
;
; RV64-LABEL: PR41129:
; RV64: # %bb.0: # %entry
; RV64-NEXT: ld a1, 0(a0)
; RV64-NEXT: beqz a1, .LBB37_2
; RV64-NEXT: # %bb.1: # %false
; RV64-NEXT: andi a1, a1, 7
; RV64-NEXT: sd a1, 0(a0)
; RV64-NEXT: ret
; RV64-NEXT: .LBB37_2: # %true
; RV64-NEXT: addi a1, a1, -1
; RV64-NEXT: sd a1, 0(a0)
; RV64-NEXT: ret
entry:
%key = load i64, ptr %p64, align 8
%cond17 = icmp eq i64 %key, 0
br i1 %cond17, label %true, label %false
false:
%andval = and i64 %key, 7
store i64 %andval, ptr %p64
br label %exit
true:
%svalue = add i64 %key, -1
store i64 %svalue, ptr %p64
br label %exit
exit:
ret void
}
define i16 @overflow_not_used(i16 %a, i16 %b, ptr %res) {
; RV32-LABEL: overflow_not_used:
; RV32: # %bb.0:
; RV32-NEXT: lui a3, 16
; RV32-NEXT: addi a3, a3, -1
; RV32-NEXT: and a4, a1, a3
; RV32-NEXT: add a0, a1, a0
; RV32-NEXT: and a3, a0, a3
; RV32-NEXT: bltu a3, a4, .LBB38_2
; RV32-NEXT: # %bb.1:
; RV32-NEXT: li a1, 42
; RV32-NEXT: .LBB38_2:
; RV32-NEXT: sh a0, 0(a2)
; RV32-NEXT: mv a0, a1
; RV32-NEXT: ret
;
; RV64-LABEL: overflow_not_used:
; RV64: # %bb.0:
; RV64-NEXT: lui a3, 16
; RV64-NEXT: addiw a3, a3, -1
; RV64-NEXT: and a4, a1, a3
; RV64-NEXT: add a0, a1, a0
; RV64-NEXT: and a3, a0, a3
; RV64-NEXT: bltu a3, a4, .LBB38_2
; RV64-NEXT: # %bb.1:
; RV64-NEXT: li a1, 42
; RV64-NEXT: .LBB38_2:
; RV64-NEXT: sh a0, 0(a2)
; RV64-NEXT: mv a0, a1
; RV64-NEXT: ret
%add = add i16 %b, %a
%cmp = icmp ult i16 %add, %b
%Q = select i1 %cmp, i16 %b, i16 42
store i16 %add, ptr %res
ret i16 %Q
}
|