Skip to content

Commit 9462abd

Browse files
committed
[SLP]Fix PR89187: fixx assertion check.
Need to use proper index variable to fix a crash.
1 parent c26e9bf commit 9462abd

File tree

2 files changed

+72
-1
lines changed

2 files changed

+72
-1
lines changed

llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -13362,7 +13362,7 @@ Value *BoUpSLP::vectorizeTree(
1336213362
do {
1336313363
IEBase = cast<InsertElementInst>(Base);
1336413364
int IEIdx = *getInsertIndex(IEBase);
13365-
assert(Mask[Idx] == PoisonMaskElem &&
13365+
assert(Mask[IEIdx] == PoisonMaskElem &&
1336613366
"InsertElementInstruction used already.");
1336713367
Mask[IEIdx] = IEIdx;
1336813368
Base = IEBase->getOperand(0);
Lines changed: 71 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,71 @@
1+
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
2+
; RUN: opt -S --passes=slp-vectorizer < %s | FileCheck %s
3+
4+
define void @test(i8 %0, i8 %1) {
5+
; CHECK-LABEL: define void @test(
6+
; CHECK-SAME: i8 [[TMP0:%.*]], i8 [[TMP1:%.*]]) {
7+
; CHECK-NEXT: entry:
8+
; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i8>, ptr getelementptr (i8, ptr null, i32 8), align 1
9+
; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <8 x i8> [[TMP2]], <8 x i8> poison, <16 x i32> <i32 7, i32 6, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 poison, i32 poison>
10+
; CHECK-NEXT: [[LUPTO132421:%.*]] = shufflevector <16 x i8> zeroinitializer, <16 x i8> [[TMP3]], <16 x i32> <i32 16, i32 17, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 14, i32 15>
11+
; CHECK-NEXT: [[TMP4:%.*]] = insertelement <16 x i8> [[LUPTO132421]], i8 [[TMP0]], i32 0
12+
; CHECK-NEXT: [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[TMP1]], i32 1
13+
; CHECK-NEXT: [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 0, i32 7
14+
; CHECK-NEXT: [[TMP7:%.*]] = shufflevector <8 x i8> [[TMP2]], <8 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
15+
; CHECK-NEXT: [[TMP8:%.*]] = shufflevector <16 x i8> [[TMP6]], <16 x i8> [[TMP7]], <16 x i32> <i32 16, i32 17, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 17, i32 poison, i32 poison, i32 poison, i32 poison, i32 14, i32 15>
16+
; CHECK-NEXT: [[TMP9:%.*]] = shufflevector <16 x i8> [[TMP8]], <16 x i8> poison, <16 x i32> <i32 0, i32 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 7, i32 0, i32 9, i32 0, i32 0, i32 0, i32 0, i32 0, i32 15>
17+
; CHECK-NEXT: [[TMP10:%.*]] = icmp ne <16 x i8> zeroinitializer, [[TMP9]]
18+
; CHECK-NEXT: ret void
19+
;
20+
entry:
21+
%li8 = load i8, ptr getelementptr (i8, ptr null, i32 8), align 1
22+
%li9163 = load i8, ptr getelementptr (i8, ptr null, i32 9), align 1
23+
%li10 = load i8, ptr getelementptr (i8, ptr null, i32 10), align 1
24+
%li11164 = load i8, ptr getelementptr (i8, ptr null, i32 11), align 1
25+
%li12 = load i8, ptr getelementptr (i8, ptr null, i32 12), align 1
26+
%li13 = load i8, ptr getelementptr (i8, ptr null, i32 13), align 1
27+
%li14 = load i8, ptr getelementptr (i8, ptr null, i32 14), align 1
28+
%li15165 = load i8, ptr getelementptr (i8, ptr null, i32 15), align 1
29+
%lupto8237 = insertelement <16 x i8> zeroinitializer, i8 %li8, i64 8
30+
%lupto9238 = insertelement <16 x i8> %lupto8237, i8 %li9163, i64 9
31+
%lupto10239 = insertelement <16 x i8> %lupto9238, i8 %li10, i64 10
32+
%lupto11240 = insertelement <16 x i8> %lupto10239, i8 %li11164, i64 11
33+
%lupto12241 = insertelement <16 x i8> %lupto11240, i8 %li12, i64 12
34+
%lupto13242 = insertelement <16 x i8> %lupto12241, i8 %li13, i64 13
35+
%lupto14243 = insertelement <16 x i8> %lupto13242, i8 %li14, i64 1
36+
%l = insertelement <16 x i8> %lupto14243, i8 %li15165, i64 0
37+
%li15 = extractelement <16 x i8> %l, i64 15
38+
%2 = icmp ne i8 %0, 0
39+
%3 = icmp ne i8 %1, 0
40+
%4 = icmp ne i8 %0, 0
41+
%.i3 = icmp ne i8 %0, 0
42+
%.i4 = icmp ne i8 %0, 0
43+
%5 = icmp ne i8 %0, 0
44+
%6 = icmp ne i8 %0, 0
45+
%7 = icmp ne i8 0, 0
46+
%8 = icmp ne i8 %0, 0
47+
%.i9 = icmp ne i8 %li9163, 0
48+
%9 = icmp ne i8 %0, 0
49+
%10 = icmp ne i8 %0, 0
50+
%11 = icmp ne i8 %0, 0
51+
%12 = icmp ne i8 %0, 0
52+
%13 = icmp ne i8 %0, 0
53+
%.i15 = icmp ne i8 %li15, 0
54+
%i0244 = insertelement <16 x i1> zeroinitializer, i1 %2, i64 0
55+
%i1245 = insertelement <16 x i1> %i0244, i1 %3, i64 1
56+
%i2246 = insertelement <16 x i1> %i1245, i1 %4, i64 2
57+
%i3247 = insertelement <16 x i1> %i2246, i1 %.i3, i64 3
58+
%i4248 = insertelement <16 x i1> %i3247, i1 %.i4, i64 4
59+
%i5249 = insertelement <16 x i1> %i4248, i1 %5, i64 5
60+
%i6250 = insertelement <16 x i1> %i5249, i1 %6, i64 6
61+
%i7251 = insertelement <16 x i1> %i6250, i1 %7, i64 7
62+
%i8252 = insertelement <16 x i1> %i7251, i1 %8, i64 8
63+
%i9253 = insertelement <16 x i1> %i8252, i1 %.i9, i64 9
64+
%i10254 = insertelement <16 x i1> %i9253, i1 %9, i64 10
65+
%i11255 = insertelement <16 x i1> %i10254, i1 %10, i64 11
66+
%i12256 = insertelement <16 x i1> %i11255, i1 %11, i64 12
67+
%i13257 = insertelement <16 x i1> %i12256, i1 %12, i64 13
68+
%i14 = insertelement <16 x i1> %i13257, i1 %13, i64 14
69+
%14 = insertelement <16 x i1> %i14, i1 %.i15, i64 15
70+
ret void
71+
}

0 commit comments

Comments
 (0)