|
| 1 | + |
| 2 | +; RUN: llvm-profdata merge %S/Inputs/misexpect-switch.proftext -o %t.profdata |
| 3 | + |
| 4 | +; RUN: opt < %s -lower-expect -pgo-instr-use -pgo-test-profile-file=%t.profdata -S -pgo-warn-misexpect 2>&1 | FileCheck %s --check-prefix=WARNING |
| 5 | +; RUN: opt < %s -lower-expect -pgo-instr-use -pgo-test-profile-file=%t.profdata -S -pass-remarks=misexpect 2>&1 | FileCheck %s --check-prefix=REMARK |
| 6 | +; RUN: opt < %s -lower-expect -pgo-instr-use -pgo-test-profile-file=%t.profdata -S -pgo-warn-misexpect -pass-remarks=misexpect 2>&1 | FileCheck %s --check-prefix=BOTH |
| 7 | +; RUN: opt < %s -lower-expect -pgo-instr-use -pgo-test-profile-file=%t.profdata -S 2>&1 | FileCheck %s --check-prefix=DISABLED |
| 8 | + |
| 9 | +; New PM |
| 10 | +; RUN: opt < %s -passes="function(lower-expect),pgo-instr-use" -pgo-test-profile-file=%t.profdata -pgo-warn-misexpect -S 2>&1 | FileCheck %s --check-prefix=WARNING |
| 11 | +; RUN: opt < %s -passes="function(lower-expect),pgo-instr-use" -pgo-test-profile-file=%t.profdata -pass-remarks=misexpect -S 2>&1 | FileCheck %s --check-prefix=REMARK |
| 12 | +; RUN: opt < %s -passes="function(lower-expect),pgo-instr-use" -pgo-test-profile-file=%t.profdata -pgo-warn-misexpect -pass-remarks=misexpect -S 2>&1 | FileCheck %s --check-prefix=BOTH |
| 13 | +; RUN: opt < %s -passes="function(lower-expect),pgo-instr-use" -pgo-test-profile-file=%t.profdata -S 2>&1 | FileCheck %s --check-prefix=DISABLED |
| 14 | + |
| 15 | +; WARNING-DAG: warning: <unknown>:0:0: 0.00% |
| 16 | +; WARNING-NOT: remark: <unknown>:0:0: Potential performance regression from use of the llvm.expect intrinsic: Annotation was correct on 0.00% (0 / 27943) of profiled executions. |
| 17 | + |
| 18 | +; REMARK-NOT: warning: <unknown>:0:0: 0.00% |
| 19 | +; REMARK-DAG: remark: <unknown>:0:0: Potential performance regression from use of the llvm.expect intrinsic: Annotation was correct on 0.00% (0 / 27943) of profiled executions. |
| 20 | + |
| 21 | +; BOTH-DAG: warning: <unknown>:0:0: 0.00% |
| 22 | +; BOTH-DAG: remark: <unknown>:0:0: Potential performance regression from use of the llvm.expect intrinsic: Annotation was correct on 0.00% (0 / 27943) of profiled executions. |
| 23 | + |
| 24 | +; DISABLED-NOT: warning: <unknown>:0:0: 0.00% |
| 25 | +; DISABLED-NOT: remark: <unknown>:0:0: Potential performance regression from use of the llvm.expect intrinsic: Annotation was correct on 0.00% (0 / 27943) of profiled executions. |
| 26 | + |
| 27 | +; DISABLED-NOT: warning: <unknown>:0:0: 0.00% |
| 28 | +; DISABLED-NOT: remark: <unknown>:0:0: Potential performance regression from use of the llvm.expect intrinsic: Annotation was correct on 0.00% (0 / 27943) of profiled executions. |
| 29 | + |
| 30 | +; CORRECT-NOT: warning: {{.*}} |
| 31 | +; CORRECT-NOT: remark: {{.*}} |
| 32 | +; CHECK-DAG: !{!"misexpect", i64 0, i64 2000, i64 1} |
| 33 | + |
| 34 | + |
| 35 | + |
| 36 | +; ModuleID = 'misexpect-switch.c' |
| 37 | +source_filename = "misexpect-switch.c" |
| 38 | +target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" |
| 39 | +target triple = "x86_64-unknown-linux-gnu" |
| 40 | + |
| 41 | +@inner_loop = dso_local constant i32 1000, align 4 |
| 42 | +@outer_loop = dso_local constant i32 20, align 4 |
| 43 | +@arry_size = dso_local constant i32 25, align 4 |
| 44 | +@arry = dso_local global [25 x i32] zeroinitializer, align 16 |
| 45 | + |
| 46 | +; Function Attrs: nounwind uwtable |
| 47 | +define dso_local void @init_arry() #0 { |
| 48 | +entry: |
| 49 | + %i = alloca i32, align 4 |
| 50 | + %0 = bitcast i32* %i to i8* |
| 51 | + call void @llvm.lifetime.start.p0i8(i64 4, i8* %0) #6 |
| 52 | + store i32 0, i32* %i, align 4, !tbaa !4 |
| 53 | + br label %for.cond |
| 54 | + |
| 55 | +for.cond: ; preds = %for.inc, %entry |
| 56 | + %1 = load i32, i32* %i, align 4, !tbaa !4 |
| 57 | + %cmp = icmp slt i32 %1, 25 |
| 58 | + br i1 %cmp, label %for.body, label %for.end |
| 59 | + |
| 60 | +for.body: ; preds = %for.cond |
| 61 | + %call = call i32 @rand() #6 |
| 62 | + %rem = srem i32 %call, 10 |
| 63 | + %2 = load i32, i32* %i, align 4, !tbaa !4 |
| 64 | + %idxprom = sext i32 %2 to i64 |
| 65 | + %arrayidx = getelementptr inbounds [25 x i32], [25 x i32]* @arry, i64 0, i64 %idxprom |
| 66 | + store i32 %rem, i32* %arrayidx, align 4, !tbaa !4 |
| 67 | + br label %for.inc |
| 68 | + |
| 69 | +for.inc: ; preds = %for.body |
| 70 | + %3 = load i32, i32* %i, align 4, !tbaa !4 |
| 71 | + %inc = add nsw i32 %3, 1 |
| 72 | + store i32 %inc, i32* %i, align 4, !tbaa !4 |
| 73 | + br label %for.cond |
| 74 | + |
| 75 | +for.end: ; preds = %for.cond |
| 76 | + %4 = bitcast i32* %i to i8* |
| 77 | + call void @llvm.lifetime.end.p0i8(i64 4, i8* %4) #6 |
| 78 | + ret void |
| 79 | +} |
| 80 | + |
| 81 | +; Function Attrs: argmemonly nounwind willreturn |
| 82 | +declare void @llvm.lifetime.start.p0i8(i64 immarg, i8* nocapture) #1 |
| 83 | + |
| 84 | +; Function Attrs: nounwind readnone speculatable willreturn |
| 85 | +declare void @llvm.dbg.declare(metadata, metadata, metadata) #2 |
| 86 | + |
| 87 | +; Function Attrs: nounwind |
| 88 | +declare dso_local i32 @rand() #3 |
| 89 | + |
| 90 | +; Function Attrs: argmemonly nounwind willreturn |
| 91 | +declare void @llvm.lifetime.end.p0i8(i64 immarg, i8* nocapture) #1 |
| 92 | + |
| 93 | +; Function Attrs: nounwind uwtable |
| 94 | +define dso_local i32 @main() #0 { |
| 95 | +entry: |
| 96 | + %retval = alloca i32, align 4 |
| 97 | + %val = alloca i32, align 4 |
| 98 | + %j = alloca i32, align 4 |
| 99 | + %condition = alloca i32, align 4 |
| 100 | + store i32 0, i32* %retval, align 4 |
| 101 | + call void @init_arry() |
| 102 | + %0 = bitcast i32* %val to i8* |
| 103 | + call void @llvm.lifetime.start.p0i8(i64 4, i8* %0) #6 |
| 104 | + store i32 0, i32* %val, align 4, !tbaa !4 |
| 105 | + %1 = bitcast i32* %j to i8* |
| 106 | + call void @llvm.lifetime.start.p0i8(i64 4, i8* %1) #6 |
| 107 | + store i32 0, i32* %j, align 4, !tbaa !4 |
| 108 | + br label %for.cond |
| 109 | + |
| 110 | +for.cond: ; preds = %for.inc, %entry |
| 111 | + %2 = load i32, i32* %j, align 4, !tbaa !4 |
| 112 | + %cmp = icmp slt i32 %2, 20000 |
| 113 | + br i1 %cmp, label %for.body, label %for.end |
| 114 | + |
| 115 | +for.body: ; preds = %for.cond |
| 116 | + %3 = bitcast i32* %condition to i8* |
| 117 | + call void @llvm.lifetime.start.p0i8(i64 4, i8* %3) #6 |
| 118 | + %call = call i32 @rand() #6 |
| 119 | + %rem = srem i32 %call, 5 |
| 120 | + store i32 %rem, i32* %condition, align 4, !tbaa !4 |
| 121 | + %4 = load i32, i32* %condition, align 4, !tbaa !4 |
| 122 | + %conv = zext i32 %4 to i64 |
| 123 | + %expval = call i64 @llvm.expect.i64(i64 %conv, i64 6) |
| 124 | + switch i64 %expval, label %sw.default [ |
| 125 | + i64 0, label %sw.bb |
| 126 | + i64 1, label %sw.bb2 |
| 127 | + i64 2, label %sw.bb2 |
| 128 | + i64 3, label %sw.bb2 |
| 129 | + i64 4, label %sw.bb3 |
| 130 | + ] |
| 131 | + |
| 132 | +sw.bb: ; preds = %for.body |
| 133 | + %call1 = call i32 @sum(i32* getelementptr inbounds ([25 x i32], [25 x i32]* @arry, i64 0, i64 0), i32 25) |
| 134 | + %5 = load i32, i32* %val, align 4, !tbaa !4 |
| 135 | + %add = add nsw i32 %5, %call1 |
| 136 | + store i32 %add, i32* %val, align 4, !tbaa !4 |
| 137 | + br label %sw.epilog |
| 138 | + |
| 139 | +sw.bb2: ; preds = %for.body, %for.body, %for.body |
| 140 | + br label %sw.epilog |
| 141 | + |
| 142 | +sw.bb3: ; preds = %for.body |
| 143 | + %call4 = call i32 @random_sample(i32* getelementptr inbounds ([25 x i32], [25 x i32]* @arry, i64 0, i64 0), i32 25) |
| 144 | + %6 = load i32, i32* %val, align 4, !tbaa !4 |
| 145 | + %add5 = add nsw i32 %6, %call4 |
| 146 | + store i32 %add5, i32* %val, align 4, !tbaa !4 |
| 147 | + br label %sw.epilog |
| 148 | + |
| 149 | +sw.default: ; preds = %for.body |
| 150 | + unreachable |
| 151 | + |
| 152 | +sw.epilog: ; preds = %sw.bb3, %sw.bb2, %sw.bb |
| 153 | + %7 = bitcast i32* %condition to i8* |
| 154 | + call void @llvm.lifetime.end.p0i8(i64 4, i8* %7) #6 |
| 155 | + br label %for.inc |
| 156 | + |
| 157 | +for.inc: ; preds = %sw.epilog |
| 158 | + %8 = load i32, i32* %j, align 4, !tbaa !4 |
| 159 | + %inc = add nsw i32 %8, 1 |
| 160 | + store i32 %inc, i32* %j, align 4, !tbaa !4 |
| 161 | + br label %for.cond |
| 162 | + |
| 163 | +for.end: ; preds = %for.cond |
| 164 | + %9 = bitcast i32* %j to i8* |
| 165 | + call void @llvm.lifetime.end.p0i8(i64 4, i8* %9) #6 |
| 166 | + %10 = bitcast i32* %val to i8* |
| 167 | + call void @llvm.lifetime.end.p0i8(i64 4, i8* %10) #6 |
| 168 | + ret i32 0 |
| 169 | +} |
| 170 | + |
| 171 | +; Function Attrs: nounwind readnone willreturn |
| 172 | +declare i64 @llvm.expect.i64(i64, i64) #4 |
| 173 | + |
| 174 | +declare dso_local i32 @sum(i32*, i32) #5 |
| 175 | + |
| 176 | +declare dso_local i32 @random_sample(i32*, i32) #5 |
| 177 | + |
| 178 | +attributes #0 = { nounwind uwtable "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="none" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" } |
| 179 | +attributes #1 = { argmemonly nounwind willreturn } |
| 180 | +attributes #2 = { nounwind readnone speculatable willreturn } |
| 181 | +attributes #3 = { nounwind "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="none" "less-precise-fpmad"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" } |
| 182 | +attributes #4 = { nounwind readnone willreturn } |
| 183 | +attributes #5 = { "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="none" "less-precise-fpmad"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" } |
| 184 | +attributes #6 = { nounwind } |
| 185 | + |
| 186 | +!llvm.module.flags = !{!0, !1, !2} |
| 187 | +!llvm.ident = !{!3} |
| 188 | + |
| 189 | +!0 = !{i32 2, !"Dwarf Version", i32 4} |
| 190 | +!1 = !{i32 2, !"Debug Info Version", i32 3} |
| 191 | +!2 = !{i32 1, !"wchar_size", i32 4} |
| 192 | +!3 = !{!"clang version 10.0.0 (60b79b85b1763d3d25630261e5cd1adb7f0835bc)"} |
| 193 | +!4 = !{!5, !5, i64 0} |
| 194 | +!5 = !{!"int", !6, i64 0} |
| 195 | +!6 = !{!"omnipotent char", !7, i64 0} |
| 196 | +!7 = !{!"Simple C/C++ TBAA"} |
0 commit comments