diff options
| author | Derek Schuff <dschuff@chromium.org> | 2012-11-06 10:23:47 -0800 |
|---|---|---|
| committer | Derek Schuff <dschuff@chromium.org> | 2012-11-06 10:23:47 -0800 |
| commit | 5bcab54cfde18b4b11f163d7d916711df70cbebf (patch) | |
| tree | c5774bfc00faa412178497d9ae92dea73d717a7c /test/Transforms | |
| parent | 96cb06677afe87ea958bf986ca2b9fb87daa2da1 (diff) | |
| parent | cfe09ed28d8a65b671e8b7a716a933e98e810e32 (diff) | |
Merge commit 'cfe09ed28d8a65b671e8b7a716a933e98e810e32'
Conflicts:
lib/Target/ARM/ARMFrameLowering.cpp
lib/Target/Mips/MipsRegisterInfo.cpp
lib/Target/X86/X86ISelLowering.cpp
lib/Transforms/IPO/ExtractGV.cpp
tools/Makefile
tools/gold/gold-plugin.cpp
The only interesting conflict was X86ISelLowering.ccp, which
meant I had to essentially revert r167104. The problem is that we are
using ESP as the stack pointer in X86ISelLowering and RSP as the
stack pointer in X86FrameLowering, and that revision made them
both consistently use X86RegisterInfo to determine which to use.
Diffstat (limited to 'test/Transforms')
60 files changed, 2473 insertions, 487 deletions
diff --git a/test/Transforms/BBVectorize/X86/loop1.ll b/test/Transforms/BBVectorize/X86/loop1.ll new file mode 100644 index 0000000000..493f23b098 --- /dev/null +++ b/test/Transforms/BBVectorize/X86/loop1.ll @@ -0,0 +1,53 @@ +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-unknown-linux-gnu" +; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 -bb-vectorize -bb-vectorize-req-chain-depth=3 -instcombine -gvn -S | FileCheck %s +; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 -basicaa -loop-unroll -unroll-threshold=45 -unroll-allow-partial -bb-vectorize -bb-vectorize-req-chain-depth=3 -instcombine -gvn -S | FileCheck %s -check-prefix=CHECK-UNRL +; The second check covers the use of alias analysis (with loop unrolling). + +define void @test1(double* noalias %out, double* noalias %in1, double* noalias %in2) nounwind uwtable { +entry: + br label %for.body +; CHECK: @test1 +; CHECK-UNRL: @test1 + +for.body: ; preds = %for.body, %entry + %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ] + %arrayidx = getelementptr inbounds double* %in1, i64 %indvars.iv + %0 = load double* %arrayidx, align 8 + %arrayidx2 = getelementptr inbounds double* %in2, i64 %indvars.iv + %1 = load double* %arrayidx2, align 8 + %mul = fmul double %0, %0 + %mul3 = fmul double %0, %1 + %add = fadd double %mul, %mul3 + %add4 = fadd double %1, %1 + %add5 = fadd double %add4, %0 + %mul6 = fmul double %0, %add5 + %add7 = fadd double %add, %mul6 + %mul8 = fmul double %1, %1 + %add9 = fadd double %0, %0 + %add10 = fadd double %add9, %0 + %mul11 = fmul double %mul8, %add10 + %add12 = fadd double %add7, %mul11 + %arrayidx14 = getelementptr inbounds double* %out, i64 %indvars.iv + store double %add12, double* %arrayidx14, align 8 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, 10 + br i1 %exitcond, label %for.end, label %for.body +; CHECK-NOT: <2 x double> +; CHECK-UNRL: %mul = fmul <2 x double> %2, %2 +; CHECK-UNRL: %mul3 = fmul <2 x double> %2, %3 +; CHECK-UNRL: %add = fadd <2 x double> %mul, %mul3 +; CHECK-UNRL: %add4 = fadd <2 x double> %3, %3 +; CHECK-UNRL: %add5 = fadd <2 x double> %add4, %2 +; CHECK-UNRL: %mul6 = fmul <2 x double> %2, %add5 +; CHECK-UNRL: %add7 = fadd <2 x double> %add, %mul6 +; CHECK-UNRL: %mul8 = fmul <2 x double> %3, %3 +; CHECK-UNRL: %add9 = fadd <2 x double> %2, %2 +; CHECK-UNRL: %add10 = fadd <2 x double> %add9, %2 +; CHECK-UNRL: %mul11 = fmul <2 x double> %mul8, %add10 +; CHECK-UNRL: %add12 = fadd <2 x double> %add7, %mul11 + +for.end: ; preds = %for.body + ret void +} diff --git a/test/Transforms/BBVectorize/X86/simple-ldstr.ll b/test/Transforms/BBVectorize/X86/simple-ldstr.ll new file mode 100644 index 0000000000..0124399bad --- /dev/null +++ b/test/Transforms/BBVectorize/X86/simple-ldstr.ll @@ -0,0 +1,29 @@ +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 -bb-vectorize -bb-vectorize-req-chain-depth=3 -instcombine -gvn -S | FileCheck %s + +; Simple 3-pair chain with loads and stores +define void @test1(double* %a, double* %b, double* %c) nounwind uwtable readonly { +entry: + %i0 = load double* %a, align 8 + %i1 = load double* %b, align 8 + %mul = fmul double %i0, %i1 + %arrayidx3 = getelementptr inbounds double* %a, i64 1 + %i3 = load double* %arrayidx3, align 8 + %arrayidx4 = getelementptr inbounds double* %b, i64 1 + %i4 = load double* %arrayidx4, align 8 + %mul5 = fmul double %i3, %i4 + store double %mul, double* %c, align 8 + %arrayidx5 = getelementptr inbounds double* %c, i64 1 + store double %mul5, double* %arrayidx5, align 8 + ret void +; CHECK: @test1 +; CHECK: %i0.v.i0 = bitcast double* %a to <2 x double>* +; CHECK: %i1.v.i0 = bitcast double* %b to <2 x double>* +; CHECK: %i0 = load <2 x double>* %i0.v.i0, align 8 +; CHECK: %i1 = load <2 x double>* %i1.v.i0, align 8 +; CHECK: %mul = fmul <2 x double> %i0, %i1 +; CHECK: %0 = bitcast double* %c to <2 x double>* +; CHECK: store <2 x double> %mul, <2 x double>* %0, align 8 +; CHECK: ret void +} + diff --git a/test/Transforms/BBVectorize/X86/simple.ll b/test/Transforms/BBVectorize/X86/simple.ll new file mode 100644 index 0000000000..0113e38bb1 --- /dev/null +++ b/test/Transforms/BBVectorize/X86/simple.ll @@ -0,0 +1,103 @@ +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128" +; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 -bb-vectorize -bb-vectorize-req-chain-depth=3 -instcombine -gvn -S | FileCheck %s + +; Basic depth-3 chain +define double @test1(double %A1, double %A2, double %B1, double %B2) { + %X1 = fsub double %A1, %B1 + %X2 = fsub double %A2, %B2 + %Y1 = fmul double %X1, %A1 + %Y2 = fmul double %X2, %A2 + %Z1 = fadd double %Y1, %B1 + %Z2 = fadd double %Y2, %B2 + %R = fmul double %Z1, %Z2 + ret double %R +; CHECK: @test1 +; CHECK-NOT: fmul <2 x double> +; CHECK: ret double %R +} + +; Basic chain +define double @test1a(double %A1, double %A2, double %B1, double %B2) { + %X1 = fsub double %A1, %B1 + %X2 = fsub double %A2, %B2 + %Y1 = fmul double %X1, %A1 + %Y2 = fmul double %X2, %A2 + %Z1 = fadd double %Y1, %B1 + %Z2 = fadd double %Y2, %B2 + %W1 = fadd double %Y1, %Z1 + %W2 = fadd double %Y2, %Z2 + %V1 = fadd double %W1, %Z1 + %V2 = fadd double %W2, %Z2 + %Q1 = fadd double %W1, %V1 + %Q2 = fadd double %W2, %V2 + %S1 = fadd double %W1, %Q1 + %S2 = fadd double %W2, %Q2 + %R = fmul double %S1, %S2 + ret double %R +; CHECK: @test1a +; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 +; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 +; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 +; CHECK: %X1 = fsub <2 x double> %X1.v.i0.2, %X1.v.i1.2 +; CHECK: %Y1 = fmul <2 x double> %X1, %X1.v.i0.2 +; CHECK: %Z1 = fadd <2 x double> %Y1, %X1.v.i1.2 +; CHECK: %W1 = fadd <2 x double> %Y1, %Z1 +; CHECK: %V1 = fadd <2 x double> %W1, %Z1 +; CHECK: %Q1 = fadd <2 x double> %W1, %V1 +; CHECK: %S1 = fadd <2 x double> %W1, %Q1 +; CHECK: %S1.v.r1 = extractelement <2 x double> %S1, i32 0 +; CHECK: %S1.v.r2 = extractelement <2 x double> %S1, i32 1 +; CHECK: %R = fmul double %S1.v.r1, %S1.v.r2 +; CHECK: ret double %R +} + +; Basic depth-3 chain (last pair permuted) +define double @test2(double %A1, double %A2, double %B1, double %B2) { + %X1 = fsub double %A1, %B1 + %X2 = fsub double %A2, %B2 + %Y1 = fmul double %X1, %A1 + %Y2 = fmul double %X2, %A2 + %Z1 = fadd double %Y2, %B1 + %Z2 = fadd double %Y1, %B2 + %R = fmul double %Z1, %Z2 + ret double %R +; CHECK: @test2 +; CHECK-NOT: fmul <2 x double> +; CHECK: ret double %R +} + +; Basic depth-4 chain (internal permutation) +define double @test4(double %A1, double %A2, double %B1, double %B2) { + %X1 = fsub double %A1, %B1 + %X2 = fsub double %A2, %B2 + %Y1 = fmul double %X1, %A1 + %Y2 = fmul double %X2, %A2 + %Z1 = fadd double %Y2, %B1 + %Z2 = fadd double %Y1, %B2 + %W1 = fadd double %Y2, %Z1 + %W2 = fadd double %Y1, %Z2 + %R = fmul double %Z1, %Z2 + ret double %R +; CHECK: @test4 +; CHECK-NOT: fmul <2 x double> +; CHECK: ret double %R +} + +; Basic chain with shuffles +define <8 x i8> @test6(<8 x i8> %A1, <8 x i8> %A2, <8 x i8> %B1, <8 x i8> %B2) { + %X1 = sub <8 x i8> %A1, %B1 + %X2 = sub <8 x i8> %A2, %B2 + %Y1 = mul <8 x i8> %X1, %A1 + %Y2 = mul <8 x i8> %X2, %A2 + %Z1 = add <8 x i8> %Y1, %B1 + %Z2 = add <8 x i8> %Y2, %B2 + %Q1 = shufflevector <8 x i8> %Z1, <8 x i8> %Z2, <8 x i32> <i32 15, i32 8, i32 6, i32 1, i32 13, i32 10, i32 4, i32 3> + %Q2 = shufflevector <8 x i8> %Z2, <8 x i8> %Z2, <8 x i32> <i32 6, i32 7, i32 0, i32 1, i32 2, i32 4, i32 4, i32 1> + %R = mul <8 x i8> %Q1, %Q2 + ret <8 x i8> %R +; CHECK: @test6 +; CHECK-NOT: sub <16 x i8> +; CHECK: ret <8 x i8> +} + diff --git a/test/Transforms/BBVectorize/X86/vs-cast.ll b/test/Transforms/BBVectorize/X86/vs-cast.ll new file mode 100644 index 0000000000..be3efca925 --- /dev/null +++ b/test/Transforms/BBVectorize/X86/vs-cast.ll @@ -0,0 +1,12 @@ +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-unknown-linux-gnu" +; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 -bb-vectorize -S | FileCheck %s + +define void @main() nounwind uwtable { +entry: + %0 = bitcast <2 x i64> undef to i128 + %1 = bitcast <2 x i64> undef to i128 + ret void +; CHECK: @main +} + diff --git a/test/Transforms/BBVectorize/cycle.ll b/test/Transforms/BBVectorize/cycle.ll index 32a91ceee0..e8e82ce024 100644 --- a/test/Transforms/BBVectorize/cycle.ll +++ b/test/Transforms/BBVectorize/cycle.ll @@ -107,6 +107,6 @@ done: ret void ; CHECK: @test1 ; CHECK: go: -; CHECK-NEXT: %conv.v.i0.1 = insertelement <2 x i32> undef, i32 %n.0, i32 0 +; CHECK: %conv.v.i0.1 = insertelement <2 x i32> undef, i32 %n.0, i32 0 ; FIXME: When tree pruning is deterministic, include the entire output. } diff --git a/test/Transforms/BBVectorize/lit.local.cfg b/test/Transforms/BBVectorize/lit.local.cfg index 19eebc0ac7..a8ad0f1a28 100644 --- a/test/Transforms/BBVectorize/lit.local.cfg +++ b/test/Transforms/BBVectorize/lit.local.cfg @@ -1 +1,6 @@ config.suffixes = ['.ll', '.c', '.cpp'] + +targets = set(config.root.targets_to_build.split()) +if not 'X86' in targets: + config.unsupported = True + diff --git a/test/Transforms/BBVectorize/loop1.ll b/test/Transforms/BBVectorize/loop1.ll index bebc91ad91..c22ea5852a 100644 --- a/test/Transforms/BBVectorize/loop1.ll +++ b/test/Transforms/BBVectorize/loop1.ll @@ -42,8 +42,8 @@ for.body: ; preds = %for.body, %entry ; CHECK: %mul = fmul double %0, %0 ; CHECK: %mul3 = fmul double %0, %1 ; CHECK: %add = fadd double %mul, %mul3 -; CHECK: %add4.v.i1.1 = insertelement <2 x double> undef, double %1, i32 0 ; CHECK: %mul8 = fmul double %1, %1 +; CHECK: %add4.v.i1.1 = insertelement <2 x double> undef, double %1, i32 0 ; CHECK: %add4.v.i1.2 = insertelement <2 x double> %add4.v.i1.1, double %0, i32 1 ; CHECK: %add4 = fadd <2 x double> %add4.v.i1.2, %add4.v.i1.2 ; CHECK: %add5.v.i1.1 = insertelement <2 x double> undef, double %0, i32 0 diff --git a/test/Transforms/BBVectorize/search-limit.ll b/test/Transforms/BBVectorize/search-limit.ll index d9945b5630..aeaf98865b 100644 --- a/test/Transforms/BBVectorize/search-limit.ll +++ b/test/Transforms/BBVectorize/search-limit.ll @@ -7,8 +7,8 @@ define double @test1(double %A1, double %A2, double %B1, double %B2) { ; CHECK-SL4: @test1 ; CHECK-SL4-NOT: <2 x double> ; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 -; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 %X1 = fsub double %A1, %B1 %X2 = fsub double %A2, %B2 diff --git a/test/Transforms/BBVectorize/simple-int.ll b/test/Transforms/BBVectorize/simple-int.ll index 6844977143..ae1d63bfd8 100644 --- a/test/Transforms/BBVectorize/simple-int.ll +++ b/test/Transforms/BBVectorize/simple-int.ll @@ -17,8 +17,8 @@ define double @test1(double %A1, double %A2, double %B1, double %B2, double %C1, ret double %R ; CHECK: @test1 ; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 -; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 ; CHECK: %X1 = fsub <2 x double> %X1.v.i0.2, %X1.v.i1.2 ; CHECK: %Y1.v.i2.1 = insertelement <2 x double> undef, double %C1, i32 0 @@ -43,8 +43,8 @@ define double @test2(double %A1, double %A2, double %B1, double %B2) { ret double %R ; CHECK: @test2 ; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 -; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 ; CHECK: %X1 = fsub <2 x double> %X1.v.i0.2, %X1.v.i1.2 ; CHECK: %Y1 = call <2 x double> @llvm.cos.v2f64(<2 x double> %X1) @@ -68,8 +68,8 @@ define double @test3(double %A1, double %A2, double %B1, double %B2, i32 %P) { ret double %R ; CHECK: @test3 ; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 -; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 ; CHECK: %X1 = fsub <2 x double> %X1.v.i0.2, %X1.v.i1.2 ; CHECK: %Y1 = call <2 x double> @llvm.powi.v2f64(<2 x double> %X1, i32 %P) diff --git a/test/Transforms/BBVectorize/simple-ldstr-ptrs.ll b/test/Transforms/BBVectorize/simple-ldstr-ptrs.ll index 6294543cd8..d46f7692b6 100644 --- a/test/Transforms/BBVectorize/simple-ldstr-ptrs.ll +++ b/test/Transforms/BBVectorize/simple-ldstr-ptrs.ll @@ -2,6 +2,9 @@ target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f3 ; RUN: opt < %s -bb-vectorize -bb-vectorize-req-chain-depth=3 -instcombine -gvn -S | FileCheck %s ; RUN: opt < %s -bb-vectorize -bb-vectorize-req-chain-depth=3 -bb-vectorize-aligned-only -instcombine -gvn -S | FileCheck %s -check-prefix=CHECK-AO +; FIXME: re-enable this once pointer vectors work properly +; XFAIL: * + ; Simple 3-pair chain also with loads and stores (using ptrs and gep) define double @test1(i64* %a, i64* %b, i64* %c) nounwind uwtable readonly { entry: diff --git a/test/Transforms/BBVectorize/simple-ldstr.ll b/test/Transforms/BBVectorize/simple-ldstr.ll index c1e6a09eea..7dd77c933f 100644 --- a/test/Transforms/BBVectorize/simple-ldstr.ll +++ b/test/Transforms/BBVectorize/simple-ldstr.ll @@ -94,13 +94,13 @@ entry: ; CHECK-AO: @test3 ; CHECK-AO: %i0 = load double* %a, align 8 ; CHECK-AO: %i1 = load double* %b, align 8 -; CHECK-AO: %mul.v.i1.1 = insertelement <2 x double> undef, double %i1, i32 0 -; CHECK-AO: %mul.v.i0.1 = insertelement <2 x double> undef, double %i0, i32 0 ; CHECK-AO: %arrayidx3 = getelementptr inbounds double* %a, i64 1 ; CHECK-AO: %i3 = load double* %arrayidx3, align 8 ; CHECK-AO: %arrayidx4 = getelementptr inbounds double* %b, i64 1 ; CHECK-AO: %i4 = load double* %arrayidx4, align 8 +; CHECK-AO: %mul.v.i1.1 = insertelement <2 x double> undef, double %i1, i32 0 ; CHECK-AO: %mul.v.i1.2 = insertelement <2 x double> %mul.v.i1.1, double %i4, i32 1 +; CHECK-AO: %mul.v.i0.1 = insertelement <2 x double> undef, double %i0, i32 0 ; CHECK-AO: %mul.v.i0.2 = insertelement <2 x double> %mul.v.i0.1, double %i3, i32 1 ; CHECK-AO: %mul = fmul <2 x double> %mul.v.i0.2, %mul.v.i1.2 ; CHECK-AO: %mulf = fptrunc <2 x double> %mul to <2 x float> @@ -140,3 +140,31 @@ if.end: ; CHECK-AO-NOT: <2 x double> } +; Simple 3-pair chain with loads and stores +define void @test5(double* %a, double* %b, double* %c) nounwind uwtable readonly { +entry: + %i0 = load double* %a, align 8 + %i1 = load double* %b, align 8 + %mul = fmul double %i0, %i1 + %arrayidx3 = getelementptr inbounds double* %a, i64 1 + %i3 = load double* %arrayidx3, align 8 + %arrayidx4 = getelementptr inbounds double* %b, i64 1 + %i4 = load double* %arrayidx4, align 8 + %mul5 = fmul double %i3, %i4 + %arrayidx5 = getelementptr inbounds double* %c, i64 1 + store double %mul5, double* %arrayidx5, align 8 + store double %mul, double* %c, align 4 + ret void +; CHECK: @test5 +; CHECK: %i0.v.i0 = bitcast double* %a to <2 x double>* +; CHECK: %i1.v.i0 = bitcast double* %b to <2 x double>* +; CHECK: %i0 = load <2 x double>* %i0.v.i0, align 8 +; CHECK: %i1 = load <2 x double>* %i1.v.i0, align 8 +; CHECK: %mul = fmul <2 x double> %i0, %i1 +; CHECK: %0 = bitcast double* %c to <2 x double>* +; CHECK: store <2 x double> %mul, <2 x double>* %0, align 4 +; CHECK: ret void +; CHECK-AO: @test5 +; CHECK-AO-NOT: <2 x double> +} + diff --git a/test/Transforms/BBVectorize/simple-sel.ll b/test/Transforms/BBVectorize/simple-sel.ll index 325792a5dc..15ecb59702 100644 --- a/test/Transforms/BBVectorize/simple-sel.ll +++ b/test/Transforms/BBVectorize/simple-sel.ll @@ -6,8 +6,8 @@ target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f3 define double @test1(double %A1, double %A2, double %B1, double %B2, i1 %C1, i1 %C2) { ; CHECK: @test1 ; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 -; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 %X1 = fsub double %A1, %B1 %X2 = fsub double %A2, %B2 @@ -33,8 +33,8 @@ define double @test2(double %A1, double %A2, double %B1, double %B2) { ; CHECK: @test2 ; CHECK-NB: @test2 ; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 -; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 %X1 = fsub double %A1, %B1 %X2 = fsub double %A2, %B2 diff --git a/test/Transforms/BBVectorize/simple.ll b/test/Transforms/BBVectorize/simple.ll index 88eb9c90f7..d9a12eebed 100644 --- a/test/Transforms/BBVectorize/simple.ll +++ b/test/Transforms/BBVectorize/simple.ll @@ -5,8 +5,8 @@ target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f3 define double @test1(double %A1, double %A2, double %B1, double %B2) { ; CHECK: @test1 ; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 -; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 %X1 = fsub double %A1, %B1 %X2 = fsub double %A2, %B2 @@ -29,8 +29,8 @@ define double @test1(double %A1, double %A2, double %B1, double %B2) { define double @test2(double %A1, double %A2, double %B1, double %B2) { ; CHECK: @test2 ; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 -; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 %X1 = fsub double %A1, %B1 %X2 = fsub double %A2, %B2 @@ -40,12 +40,13 @@ define double @test2(double %A1, double %A2, double %B1, double %B2) { ; CHECK: %Y1 = fmul <2 x double> %X1, %X1.v.i0.2 %Z1 = fadd double %Y2, %B1 %Z2 = fadd double %Y1, %B2 -; CHECK: %Z1.v.i0 = shufflevector <2 x double> %Y1, <2 x double> undef, <2 x i32> <i32 1, i32 0> -; CHECK: %Z1 = fadd <2 x double> %Z1.v.i0, %X1.v.i1.2 +; CHECK: %Z1.v.i1.1 = insertelement <2 x double> undef, double %B2, i32 0 +; CHECK: %Z1.v.i1.2 = insertelement <2 x double> %Z1.v.i1.1, double %B1, i32 1 +; CHECK: %Z2 = fadd <2 x double> %Y1, %Z1.v.i1.2 %R = fmul double %Z1, %Z2 -; CHECK: %Z1.v.r1 = extractelement <2 x double> %Z1, i32 0 -; CHECK: %Z1.v.r2 = extractelement <2 x double> %Z1, i32 1 -; CHECK: %R = fmul double %Z1.v.r1, %Z1.v.r2 +; CHECK: %Z2.v.r1 = extractelement <2 x double> %Z2, i32 0 +; CHECK: %Z2.v.r2 = extractelement <2 x double> %Z2, i32 1 +; CHECK: %R = fmul double %Z2.v.r2, %Z2.v.r1 ret double %R ; CHECK: ret double %R } @@ -54,8 +55,8 @@ define double @test2(double %A1, double %A2, double %B1, double %B2) { define double @test3(double %A1, double %A2, double %B1, double %B2) { ; CHECK: @test3 ; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 -; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 %X1 = fsub double %A1, %B1 %X2 = fsub double %A2, %B2 @@ -79,8 +80,8 @@ define double @test3(double %A1, double %A2, double %B1, double %B2) { define double @test4(double %A1, double %A2, double %B1, double %B2) { ; CHECK: @test4 ; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 -; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 ; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 %X1 = fsub double %A1, %B1 %X2 = fsub double %A2, %B2 @@ -148,4 +149,27 @@ define <8 x i8> @test6(<8 x i8> %A1, <8 x i8> %A2, <8 x i8> %B1, <8 x i8> %B2) { ; CHECK: ret <8 x i8> %R } +; Basic depth-3 chain (flipped order) +define double @test7(double %A1, double %A2, double %B1, double %B2) { +; CHECK: @test7 +; CHECK: %X1.v.i1.1 = insertelement <2 x double> undef, double %B1, i32 0 +; CHECK: %X1.v.i1.2 = insertelement <2 x double> %X1.v.i1.1, double %B2, i32 1 +; CHECK: %X1.v.i0.1 = insertelement <2 x double> undef, double %A1, i32 0 +; CHECK: %X1.v.i0.2 = insertelement <2 x double> %X1.v.i0.1, double %A2, i32 1 + %X1 = fsub double %A1, %B1 + %X2 = fsub double %A2, %B2 +; CHECK: %X1 = fsub <2 x double> %X1.v.i0.2, %X1.v.i1.2 + %Y1 = fmul double %X1, %A1 + %Y2 = fmul double %X2, %A2 +; CHECK: %Y1 = fmul <2 x double> %X1, %X1.v.i0.2 + %Z2 = fadd double %Y2, %B2 + %Z1 = fadd double %Y1, %B1 +; CHECK: %Z1 = fadd <2 x double> %Y1, %X1.v.i1.2 + %R = fmul double %Z1, %Z2 +; CHECK: %Z1.v.r1 = extractelement <2 x double> %Z1, i32 0 +; CHECK: %Z1.v.r2 = extractelement <2 x double> %Z1, i32 1 +; CHECK: %R = fmul double %Z1.v.r1, %Z1.v.r2 + ret double %R +; CHECK: ret double %R +} diff --git a/test/Transforms/IndVarSimplify/verify-scev.ll b/test/Transforms/IndVarSimplify/verify-scev.ll new file mode 100644 index 0000000000..019f5830d5 --- /dev/null +++ b/test/Transforms/IndVarSimplify/verify-scev.ll @@ -0,0 +1,421 @@ +; RUN: opt < %s -S -indvars -verify-scev +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +define void @test1() nounwind uwtable ssp { +entry: + br i1 undef, label %for.end, label %for.body + +for.body: ; preds = %for.body, %entry + br i1 false, label %for.end, label %for.body + +for.end: ; preds = %for.body, %entry + br i1 undef, label %for.end11, label %for.body3 + +for.body3: ; preds = %for.end + unreachable + +for.end11: ; preds = %for.end + br i1 undef, label %while.body, label %while.end + +while.body: ; preds = %for.end11 + unreachable + +while.end: ; preds = %for.end11 + br i1 undef, label %if.end115, label %for.cond109 + +for.cond109: ; preds = %while.end + unreachable + +if.end115: ; preds = %while.end + br i1 undef, label %while.body119.lr.ph.lr.ph, label %for.cond612 + +while.body119.lr.ph.lr.ph: ; preds = %if.end115 + br i1 undef, label %for.cond612, label %if.end123.us + +if.end123.us: ; preds = %while.body119.lr.ph.lr.ph + br label %for.cond132.us + +for.cond132.us: ; preds = %for.cond132.us, %if.end123.us + br i1 undef, label %if.then136.us, label %for.cond132.us + +if.then136.us: ; preds = %for.cond132.us + br i1 undef, label %while.end220, label %while.body211 + +while.body211: ; preds = %while.body211, %if.then136.us + br i1 undef, label %while.end220, label %while.body211 + +while.end220: ; preds = %while.body211, %if.then136.us + br label %for.cond246.outer + +for.cond246.outer: ; preds = %for.inc558, %for.cond394.preheader, %if.then274, %for.cond404.preheader, %while.end220 + br label %for.cond246 + +for.cond246: ; preds = %for.cond372.loopexit, %for.cond246.outer + br i1 undef, label %for.end562, label %if.end250 + +if.end250: ; preds = %for.cond246 + br i1 undef, label %if.end256, label %for.end562 + +if.end256: ; preds = %if.end250 + %cmp272 = icmp eq i32 undef, undef + br i1 %cmp272, label %if.then274, label %for.cond404.preheader + +for.cond404.preheader: ; preds = %if.end256 + br i1 undef, label %for.cond246.outer, label %for.body409.lr.ph + +for.body409.lr.ph: ; preds = %for.cond404.preheader + br label %for.body409 + +if.then274: ; preds = %if.end256 + br i1 undef, label %for.cond246.outer, label %if.end309 + +if.end309: ; preds = %if.then274 + br i1 undef, label %for.cond372.loopexit, label %for.body361 + +for.body361: ; preds = %for.body361, %if.end309 + br i1 undef, label %for.cond372.loopexit, label %for.body361 + +for.cond372.loopexit: ; preds = %for.body361, %if.end309 + br i1 undef, label %for.cond394.preheader, label %for.cond246 + +for.cond394.preheader: ; preds = %for.cond372.loopexit + br i1 undef, label %for.cond246.outer, label %for.body397 + +for.body397: ; preds = %for.cond394.preheader + unreachable + +for.body409: ; preds = %for.inc558, %for.body409.lr.ph + %k.029 = phi i32 [ 1, %for.body409.lr.ph ], [ %inc559, %for.inc558 ] + br i1 undef, label %if.then412, label %if.else433 + +if.then412: ; preds = %for.body409 + br label %if.end440 + +if.else433: ; preds = %for.body409 + br label %if.end440 + +if.end440: ; preds = %if.else433, %if.then412 + br i1 undef, label %for.inc558, label %if.end461 + +if.end461: ; preds = %if.end440 + br i1 undef, label %for.cond528.loopexit, label %for.body517 + +for.body517: ; preds = %for.body517, %if.end461 + br i1 undef, label %for.cond528.loopexit, label %for.body517 + +for.cond528.loopexit: ; preds = %for.body517, %if.end461 + br label %for.inc558 + +for.inc558: ; preds = %for.cond528.loopexit, %if.end440 + %inc559 = add nsw i32 %k.029, 1 + %cmp407 = icmp sgt i32 %inc559, undef + br i1 %cmp407, label %for.cond246.outer, label %for.body409 + +for.end562: ; preds = %if.end250, %for.cond246 + unreachable + +for.cond612: ; preds = %while.body119.lr.ph.lr.ph, %if.end115 + unreachable +} + +define void @test2() nounwind uwtable ssp { +entry: + br i1 undef, label %for.end, label %for.body + +for.body: ; preds = %for.body, %entry + br i1 undef, label %for.end, label %for.body + +for.end: ; preds = %for.body, %entry + br i1 undef, label %for.end11, label %for.body3 + +for.body3: ; preds = %for.end + unreachable + +for.end11: ; preds = %for.end + br i1 undef, label %while.body, label %while.end + +while.body: ; preds = %for.end11 + unreachable + +while.end: ; preds = %for.end11 + br i1 undef, label %if.end115, label %for.cond109 + +for.cond109: ; preds = %while.end + unreachable + +if.end115: ; preds = %while.end + br i1 undef, label %while.body119.lr.ph.lr.ph, label %for.cond612 + +while.body119.lr.ph.lr.ph: ; preds = %if.end115 + br i1 undef, label %for.cond612, label %if.end123.us + +if.end123.us: ; preds = %while.body119.lr.ph.lr.ph + br label %for.cond132.us + +for.cond132.us: ; preds = %for.cond132.us, %if.end123.us + br i1 undef, label %if.then136.us, label %for.cond132.us + +if.then136.us: ; preds = %for.cond132.us + br i1 undef, label %while.end220, label %while.body211 + +while.body211: ; preds = %while.body211, %if.then136.us + br i1 undef, label %while.end220, label %while.body211 + +while.end220: ; preds = %while.body211, %if.then136.us + br label %for.cond246.outer + +for.cond246.outer: ; preds = %for.inc558, %for.cond394.preheader, %if.then274, %for.cond404.preheader, %while.end220 + br label %for.cond246 + +for.cond246: ; preds = %for.cond372.loopexit, %for.cond246.outer + br i1 undef, label %for.end562, label %if.end250 + +if.end250: ; preds = %for.cond246 + br i1 undef, label %if.end256, label %for.end562 + +if.end256: ; preds = %if.end250 + %0 = load i32* undef, align 4 + br i1 undef, label %if.then274, label %for.cond404.preheader + +for.cond404.preheader: ; preds = %if.end256 + %add406 = add i32 0, %0 + br i1 undef, label %for.cond246.outer, label %for.body409.lr.ph + +for.body409.lr.ph: ; preds = %for.cond404.preheader + br label %for.body409 + +if.then274: ; preds = %if.end256 + br i1 undef, label %for.cond246.outer, label %if.end309 + +if.end309: ; preds = %if.then274 + br i1 undef, label %for.cond372.loopexit, label %for.body361 + +for.body361: ; preds = %for.body361, %if.end309 + br i1 undef, label %for.cond372.loopexit, label %for.body361 + +for.cond372.loopexit: ; preds = %for.body361, %if.end309 + br i1 undef, label %for.cond394.preheader, label %for.cond246 + +for.cond394.preheader: ; preds = %for.cond372.loopexit + br i1 undef, label %for.cond246.outer, label %for.body397 + +for.body397: ; preds = %for.cond394.preheader + unreachable + +for.body409: ; preds = %for.inc558, %for.body409.lr.ph + %k.029 = phi i32 [ 1, %for.body409.lr.ph ], [ %inc559, %for.inc558 ] + br i1 undef, label %if.then412, label %if.else433 + +if.then412: ; preds = %for.body409 + br label %if.end440 + +if.else433: ; preds = %for.body409 + br label %if.end440 + +if.end440: ; preds = %if.else433, %if.then412 + br i1 undef, label %for.inc558, label %if.end461 + +if.end461: ; preds = %if.end440 + br i1 undef, label %for.cond528.loopexit, label %for.body517 + +for.body517: ; preds = %for.body517, %if.end461 + br i1 undef, label %for.cond528.loopexit, label %for.body517 + +for.cond528.loopexit: ; preds = %for.body517, %if.end461 + br label %for.inc558 + +for.inc558: ; preds = %for.cond528.loopexit, %if.end440 + %inc559 = add nsw i32 %k.029, 1 + %cmp407 = icmp sgt i32 %inc559, %add406 + br i1 %cmp407, label %for.cond246.outer, label %for.body409 + +for.end562: ; preds = %if.end250, %for.cond246 + unreachable + +for.cond612: ; preds = %while.body119.lr.ph.lr.ph, %if.end115 + unreachable +} + +define void @test3() nounwind uwtable ssp { +entry: + br i1 undef, label %for.end, label %for.body + +for.body: ; preds = %for.body, %entry + br i1 undef, label %for.end, label %for.body + +for.end: ; preds = %for.body, %entry + br i1 undef, label %for.end11, label %for.body3 + +for.body3: ; preds = %for.end + unreachable + +for.end11: ; preds = %for.end + br i1 undef, label %while.body, label %while.end + +while.body: ; preds = %for.end11 + unreachable + +while.end: ; preds = %for.end11 + br i1 undef, label %if.end115, label %for.cond109 + +for.cond109: ; preds = %while.end + unreachable + +if.end115: ; preds = %while.end + br i1 undef, label %while.body119.lr.ph.lr.ph, label %for.cond612 + +while.body119.lr.ph.lr.ph: ; preds = %if.end115 + br i1 undef, label %for.cond612, label %if.end123.us + +if.end123.us: ; preds = %while.body119.lr.ph.lr.ph + br label %for.cond132.us + +for.cond132.us: ; preds = %for.cond132.us, %if.end123.us + br i1 undef, label %if.then136.us, label %for.cond132.us + +if.then136.us: ; preds = %for.cond132.us + br i1 undef, label %while.end220, label %while.body211 + +while.body211: ; preds = %while.body211, %if.then136.us + br i1 undef, label %while.end220, label %while.body211 + +while.end220: ; preds = %while.body211, %if.then136.us + br label %for.cond246.outer + +for.cond246.outer: ; preds = %for.inc558, %for.cond394.preheader, %if.then274, %for.cond404.preheader, %while.end220 + br label %for.cond246 + +for.cond246: ; preds = %for.cond372.loopexit, %for.cond246.outer + br i1 undef, label %for.end562, label %if.end250 + +if.end250: ; preds = %for.cond246 + br i1 undef, label %if.end256, label %for.end562 + +if.end256: ; preds = %if.end250 + br i1 undef, label %if.then274, label %for.cond404.preheader + +for.cond404.preheader: ; preds = %if.end256 + br i1 undef, label %for.cond246.outer, label %for.body409.lr.ph + +for.body409.lr.ph: ; preds = %for.cond404.preheader + br label %for.body409 + +if.then274: ; preds = %if.end256 + br i1 undef, label %for.cond246.outer, label %if.end309 + +if.end309: ; preds = %if.then274 + br i1 undef, label %for.cond372.loopexit, label %for.body361 + +for.body361: ; preds = %for.body361, %if.end309 + br i1 undef, label %for.cond372.loopexit, label %for.body361 + +for.cond372.loopexit: ; preds = %for.body361, %if.end309 + br i1 undef, label %for.cond394.preheader, label %for.cond246 + +for.cond394.preheader: ; preds = %for.cond372.loopexit + br i1 undef, label %for.cond246.outer, label %for.body397 + +for.body397: ; preds = %for.cond394.preheader + unreachable + +for.body409: ; preds = %for.inc558, %for.body409.lr.ph + br i1 undef, label %if.then412, label %if.else433 + +if.then412: ; preds = %for.body409 + br label %if.end440 + +if.else433: ; preds = %for.body409 + br label %if.end440 + +if.end440: ; preds = %if.else433, %if.then412 + br i1 undef, label %for.inc558, label %if.end461 + +if.end461: ; preds = %if.end440 + br i1 undef, label %for.cond528.loopexit, label %for.body517 + +for.body517: ; preds = %for.body517, %if.end461 + br i1 undef, label %for.cond528.loopexit, label %for.body517 + +for.cond528.loopexit: ; preds = %for.body517, %if.end461 + br label %for.inc558 + +for.inc558: ; preds = %for.cond528.loopexit, %if.end440 + br i1 undef, label %for.cond246.outer, label %for.body409 + +for.end562: ; preds = %if.end250, %for.cond246 + unreachable + +for.cond612: ; preds = %while.body119.lr.ph.lr.ph, %if.end115 + unreachable +} + +define void @test4() nounwind uwtable ssp { +entry: + br i1 undef, label %if.end8, label %if.else + +if.else: ; preds = %entry + br label %if.end8 + +if.end8: ; preds = %if.else, %entry + br i1 undef, label %if.end26, label %if.else22 + +if.else22: ; preds = %if.end8 + br label %if.end26 + +if.end26: ; preds = %if.else22, %if.end8 + br i1 undef, label %if.end35, label %if.else31 + +if.else31: ; preds = %if.end26 + br label %if.end35 + +if.end35: ; preds = %if.else31, %if.end26 + br i1 undef, label %for.end226, label %for.body.lr.ph + +for.body.lr.ph: ; preds = %if.end35 + br label %for.body48 + +for.body48: ; preds = %for.inc221, %for.body.lr.ph + br i1 undef, label %for.inc221, label %for.body65.lr.ph + +for.body65.lr.ph: ; preds = %for.body48 + %0 = load i32* undef, align 4 + br label %for.body65.us + +for.body65.us: ; preds = %for.inc219.us, %for.body65.lr.ph + %k.09.us = phi i32 [ %inc.us, %for.inc219.us ], [ 1, %for.body65.lr.ph ] + %idxprom66.us = sext i32 %k.09.us to i64 + br i1 undef, label %for.inc219.us, label %if.end72.us + +if.end72.us: ; preds = %for.body65.us + br i1 undef, label %if.end93.us, label %if.then76.us + +if.then76.us: ; preds = %if.end72.us + br label %if.end93.us + +if.end93.us: ; preds = %if.then76.us, %if.end72.us + br i1 undef, label %if.end110.us, label %for.inc219.us + +if.end110.us: ; preds = %if.end93.us + br i1 undef, label %for.inc219.us, label %for.body142.us + +for.body142.us: ; preds = %for.cond139.loopexit.us, %if.end110.us + br label %for.cond152.us + +for.cond152.us: ; preds = %for.cond152.us, %for.body142.us + br i1 undef, label %for.cond139.loopexit.us, label %for.cond152.us + +for.inc219.us: ; preds = %for.cond139.loopexit.us, %if.end110.us, %if.end93.us, %for.body65.us + %inc.us = add nsw i32 %k.09.us, 1 + %cmp64.us = icmp sgt i32 %inc.us, %0 + br i1 %cmp64.us, label %for.inc221, label %for.body65.us + +for.cond139.loopexit.us: ; preds = %for.cond152.us + br i1 undef, label %for.inc219.us, label %for.body142.us + +for.inc221: ; preds = %for.inc219.us, %for.body48 + br label %for.body48 + +for.end226: ; preds = %if.end35 + ret void +} diff --git a/test/Transforms/InstCombine/2012-10-25-vector-of-pointers.ll b/test/Transforms/InstCombine/2012-10-25-vector-of-pointers.ll new file mode 100644 index 0000000000..20ea282687 --- /dev/null +++ b/test/Transforms/InstCombine/2012-10-25-vector-of-pointers.ll @@ -0,0 +1,51 @@ +; RUN: opt < %s -instcombine -S + +; Make sure that we don't crash when optimizing the vectors of pointers. + +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +%struct.hoge = type { double*, double*, double*, double** } + +define void @widget(%struct.hoge* nocapture %arg) nounwind uwtable ssp { +bb: + %tmp = getelementptr inbounds %struct.hoge* %arg, i64 0, i32 0 + br i1 undef, label %bb1, label %bb17 + +bb1: ; preds = %bb + br i1 undef, label %bb2, label %bb3 + +bb2: ; preds = %bb1 + br label %bb17 + +bb3: ; preds = %bb1 + %tmp4 = bitcast double** %tmp to <2 x double*>* + %tmp5 = load <2 x double*>* %tmp4, align 8 + %tmp6 = ptrtoint <2 x double*> %tmp5 to <2 x i64> + %tmp7 = sub <2 x i64> zeroinitializer, %tmp6 + %tmp8 = ashr exact <2 x i64> %tmp7, <i64 3, i64 3> + %tmp9 = extractelement <2 x i64> %tmp8, i32 0 + %tmp10 = add nsw i64 undef, %tmp9 + br i1 undef, label %bb11, label %bb12 + +bb11: ; preds = %bb3 + br label %bb13 + +bb12: ; preds = %bb3 + br label %bb13 + +bb13: ; preds = %bb12, %bb11 + br i1 undef, label %bb16, label %bb14 + +bb14: ; preds = %bb13 + br i1 undef, label %bb16, label %bb15 + +bb15: ; preds = %bb14 + br label %bb16 + +bb16: ; preds = %bb15, %bb14, %bb13 + unreachable + +bb17: ; preds = %bb2, %bb + ret void +} diff --git a/test/Transforms/InstCombine/constant-fold-gep-as-0.ll b/test/Transforms/InstCombine/constant-fold-gep-as-0.ll deleted file mode 100644 index 74fe316137..0000000000 --- a/test/Transforms/InstCombine/constant-fold-gep-as-0.ll +++ /dev/null @@ -1,235 +0,0 @@ -; "PLAIN" - No optimizations. This tests the target-independent -; constant folder. -; RUN: opt -S -o - < %s | FileCheck --check-prefix=PLAIN %s - -target datalayout = "e-p:128:128:128-p1:32:32:32-p2:8:8:8-p3:16:16:16-p4:64:64:64-p5:96:96:96-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:32" - -; PLAIN: ModuleID = '<stdin>' - -; The automatic constant folder in opt does not have targetdata access, so -; it can't fold gep arithmetic, in general. However, the constant folder run -; from instcombine and global opt can use targetdata. -; PLAIN: @G8 = global i8 addrspace(1)* getelementptr (i8 addrspace(1)* inttoptr (i32 1 to i8 addrspace(1)*), i32 -1) -@G8 = global i8 addrspace(1)* getelementptr (i8 addrspace(1)* inttoptr (i32 1 to i8 addrspace(1)*), i32 -1) -; PLAIN: @G1 = global i1 addrspace(2)* getelementptr (i1 addrspace(2)* inttoptr (i8 1 to i1 addrspace(2)*), i8 -1) -@G1 = global i1 addrspace(2)* getelementptr (i1 addrspace(2)* inttoptr (i8 1 to i1 addrspace(2)*), i8 -1) -; PLAIN: @F8 = global i8 addrspace(1)* getelementptr (i8 addrspace(1)* inttoptr (i32 1 to i8 addrspace(1)*), i32 -2) -@F8 = global i8 addrspace(1)* getelementptr (i8 addrspace(1)* inttoptr (i32 1 to i8 addrspace(1)*), i32 -2) -; PLAIN: @F1 = global i1 addrspace(2)* getelementptr (i1 addrspace(2)* inttoptr (i8 1 to i1 addrspace(2)*), i8 -2) -@F1 = global i1 addrspace(2)* getelementptr (i1 addrspace(2)* inttoptr (i8 1 to i1 addrspace(2)*), i8 -2) -; PLAIN: @H8 = global i8 addrspace(1)* getelementptr (i8 addrspace(1)* null, i32 -1) -@H8 = global i8 addrspace(1)* getelementptr (i8 addrspace(1)* inttoptr (i32 0 to i8 addrspace(1)*), i32 -1) -; PLAIN: @H1 = global i1 addrspace(2)* getelementptr (i1 addrspace(2)* null, i8 -1) -@H1 = global i1 addrspace(2)* getelementptr (i1 addrspace(2)* inttoptr (i8 0 to i1 addrspace(2)*), i8 -1) - - -; The target-independent folder should be able to do some clever -; simplifications on sizeof, alignof, and offsetof expressions. The -; target-dependent folder should fold these down to constants. -; PLAIN-X: @a = constant i64 mul (i64 ptrtoint (double addrspace(4)* getelementptr (double addrspace(4)* null, i32 1) to i64), i64 2310) -@a = constant i64 mul (i64 3, i64 mul (i64 ptrtoint ({[7 x double], [7 x double]} addrspace(4)* getelementptr ({[7 x double], [7 x double]} addrspace(4)* null, i64 11) to i64), i64 5)) - -; PLAIN-X: @b = constant i64 ptrtoint (double addrspace(4)* getelementptr ({ i1, double }* null, i64 0, i32 1) to i64) -@b = constant i64 ptrtoint ([13 x double] addrspace(4)* getelementptr ({i1, [13 x double]} addrspace(4)* null, i64 0, i32 1) to i64) - -; PLAIN-X: @c = constant i64 mul nuw (i64 ptrtoint (double addrspace(4)* getelementptr (double addrspace(4)* null, i32 1) to i64), i64 2) -@c = constant i64 ptrtoint (double addrspace(4)* getelementptr ({double, double, double, double} addrspace(4)* null, i64 0, i32 2) to i64) - -; PLAIN-X: @d = constant i64 mul nuw (i64 ptrtoint (double addrspace(4)* getelementptr (double addrspace(4)* null, i32 1) to i64), i64 11) -@d = constant i64 ptrtoint (double addrspace(4)* getelementptr ([13 x double] addrspace(4)* null, i64 0, i32 11) to i64) - -; PLAIN-X: @e = constant i64 ptrtoint (double addrspace(4)* getelementptr ({ double, float, double, double }* null, i64 0, i32 2) to i64) -@e = constant i64 ptrtoint (double addrspace(4)* getelementptr ({double, float, double, double} addrspace(4)* null, i64 0, i32 2) to i64) - -; PLAIN-X: @f = constant i64 1 -@f = constant i64 ptrtoint (<{ i16, i128 }> addrspace(4)* getelementptr ({i1, <{ i16, i128 }>} addrspace(4)* null, i64 0, i32 1) to i64) - -; PLAIN-X: @g = constant i64 ptrtoint (double addrspace(4)* getelementptr ({ i1, double }* null, i64 0, i32 1) to i64) -@g = constant i64 ptrtoint ({double, double} addrspace(4)* getelementptr ({i1, {double, double}} addrspace(4)* null, i64 0, i32 1) to i64) - -; PLAIN-X: @h = constant i64 ptrtoint (i1 addrspace(2)* getelementptr (i1 addrspace(2)* null, i32 1) to i64) -@h = constant i64 ptrtoint (double addrspace(4)* getelementptr (double addrspace(4)* null, i64 1) to i64) - -; PLAIN-X: @i = constant i64 ptrtoint (i1 addrspace(2)* getelementptr ({ i1, i1 addrspace(2)* }* null, i64 0, i32 1) to i64) -@i = constant i64 ptrtoint (double addrspace(4)* getelementptr ({i1, double} addrspace(4)* null, i64 0, i32 1) to i64) - -; The target-dependent folder should cast GEP indices to integer-sized pointers. - -; PLAIN: @M = constant i64 addrspace(5)* getelementptr (i64 addrspace(5)* null, i32 1) -; PLAIN: @N = constant i64 addrspace(5)* getelementptr ({ i64, i64 } addrspace(5)* null, i32 0, i32 1) -; PLAIN: @O = constant i64 addrspace(5)* getelementptr ([2 x i64] addrspace(5)* null, i32 0, i32 1) - -@M = constant i64 addrspace(5)* getelementptr (i64 addrspace(5)* null, i32 1) -@N = constant i64 addrspace(5)* getelementptr ({ i64, i64 } addrspace(5)* null, i32 0, i32 1) -@O = constant i64 addrspace(5)* getelementptr ([2 x i64] addrspace(5)* null, i32 0, i32 1) - -; Fold GEP of a GEP. Very simple cases are folded. - -; PLAIN-X: @Y = global [3 x { i32, i32 }]addrspace(3)* getelementptr inbounds ([3 x { i32, i32 }]addrspace(3)* @ext, i64 2) -@ext = external addrspace(3) global [3 x { i32, i32 }] -@Y = global [3 x { i32, i32 }]addrspace(3)* getelementptr inbounds ([3 x { i32, i32 }]addrspace(3)* getelementptr inbounds ([3 x { i32, i32 }]addrspace(3)* @ext, i64 1), i64 1) - -; PLAIN-X: @Z = global i32addrspace(3)* getelementptr inbounds (i32addrspace(3)* getelementptr inbounds ([3 x { i32, i32 }]addrspace(3)* @ext, i64 0, i64 1, i32 0), i64 1) -@Z = global i32addrspace(3)* getelementptr inbounds (i32addrspace(3)* getelementptr inbounds ([3 x { i32, i32 }]addrspace(3)* @ext, i64 0, i64 1, i32 0), i64 1) - - -; Duplicate all of the above as function return values rather than -; global initializers. - -; PLAIN: define i8 addrspace(1)* @goo8() nounwind { -; PLAIN: %t = bitcast i8 addrspace(1)* getelementptr (i8 addrspace(1)* inttoptr (i32 1 to i8 addrspace(1)*), i32 -1) to i8 addrspace(1)* -; PLAIN: ret i8 addrspace(1)* %t -; PLAIN: } -; PLAIN: define i1 addrspace(2)* @goo1() nounwind { -; PLAIN: %t = bitcast i1 addrspace(2)* getelementptr (i1 addrspace(2)* inttoptr (i32 1 to i1 addrspace(2)*), i32 -1) to i1 addrspace(2)* -; PLAIN: ret i1 addrspace(2)* %t -; PLAIN: } -; PLAIN: define i8 addrspace(1)* @foo8() nounwind { -; PLAIN: %t = bitcast i8 addrspace(1)* getelementptr (i8 addrspace(1)* inttoptr (i32 1 to i8 addrspace(1)*), i32 -2) to i8 addrspace(1)* -; PLAIN: ret i8 addrspace(1)* %t -; PLAIN: } -; PLAIN: define i1 addrspace(2)* @foo1() nounwind { -; PLAIN: %t = bitcast i1 addrspace(2)* getelementptr (i1 addrspace(2)* inttoptr (i32 1 to i1 addrspace(2)*), i32 -2) to i1 addrspace(2)* -; PLAIN: ret i1 addrspace(2)* %t -; PLAIN: } -; PLAIN: define i8 addrspace(1)* @hoo8() nounwind { -; PLAIN: %t = bitcast i8 addrspace(1)* getelementptr (i8 addrspace(1)* null, i32 -1) to i8 addrspace(1)* -; PLAIN: ret i8 addrspace(1)* %t -; PLAIN: } -; PLAIN: define i1 addrspace(2)* @hoo1() nounwind { -; PLAIN: %t = bitcast i1 addrspace(2)* getelementptr (i1 addrspace(2)* null, i32 -1) to i1 addrspace(2)* -; PLAIN: ret i1 addrspace(2)* %t -; PLAIN: } -define i8 addrspace(1)* @goo8() nounwind { - %t = bitcast i8 addrspace(1)* getelementptr (i8 addrspace(1)* inttoptr (i32 1 to i8 addrspace(1)*), i32 -1) to i8 addrspace(1)* - ret i8 addrspace(1)* %t -} -define i1 addrspace(2)* @goo1() nounwind { - %t = bitcast i1 addrspace(2)* getelementptr (i1 addrspace(2)* inttoptr (i32 1 to i1 addrspace(2)*), i32 -1) to i1 addrspace(2)* - ret i1 addrspace(2)* %t -} -define i8 addrspace(1)* @foo8() nounwind { - %t = bitcast i8 addrspace(1)* getelementptr (i8 addrspace(1)* inttoptr (i32 1 to i8 addrspace(1)*), i32 -2) to i8 addrspace(1)* - ret i8 addrspace(1)* %t -} -define i1 addrspace(2)* @foo1() nounwind { - %t = bitcast i1 addrspace(2)* getelementptr (i1 addrspace(2)* inttoptr (i32 1 to i1 addrspace(2)*), i32 -2) to i1 addrspace(2)* - ret i1 addrspace(2)* %t -} -define i8 addrspace(1)* @hoo8() nounwind { - %t = bitcast i8 addrspace(1)* getelementptr (i8 addrspace(1)* inttoptr (i32 0 to i8 addrspace(1)*), i32 -1) to i8 addrspace(1)* - ret i8 addrspace(1)* %t -} -define i1 addrspace(2)* @hoo1() nounwind { - %t = bitcast i1 addrspace(2)* getelementptr (i1 addrspace(2)* inttoptr (i32 0 to i1 addrspace(2)*), i32 -1) to i1 addrspace(2)* - ret i1 addrspace(2)* %t -} - -; PLAIN-X: define i64 @fa() nounwind { -; PLAIN-X: %t = bitcast i64 mul (i64 ptrtoint (double addrspace(4)* getelementptr (double addrspace(4)* null, i32 1) to i64), i64 2310) to i64 -; PLAIN-X: ret i64 %t -; PLAIN-X: } -; PLAIN-X: define i64 @fb() nounwind { -; PLAIN-X: %t = bitcast i64 ptrtoint (double addrspace(4)* getelementptr ({ i1, double }* null, i64 0, i32 1) to i64) to i64 -; PLAIN-X: ret i64 %t -; PLAIN-X: } -; PLAIN-X: define i64 @fc() nounwind { -; PLAIN-X: %t = bitcast i64 mul nuw (i64 ptrtoint (double addrspace(4)* getelementptr (double addrspace(4)* null, i32 1) to i64), i64 2) to i64 -; PLAIN-X: ret i64 %t -; PLAIN-X: } -; PLAIN-X: define i64 @fd() nounwind { -; PLAIN-X: %t = bitcast i64 mul nuw (i64 ptrtoint (double addrspace(4)* getelementptr (double addrspace(4)* null, i32 1) to i64), i64 11) to i64 -; PLAIN-X: ret i64 %t -; PLAIN-X: } -; PLAIN-X: define i64 @fe() nounwind { -; PLAIN-X: %t = bitcast i64 ptrtoint (double addrspace(4)* getelementptr ({ double, float, double, double }* null, i64 0, i32 2) to i64) to i64 -; PLAIN-X: ret i64 %t -; PLAIN-X: } -; PLAIN-X: define i64 @ff() nounwind { -; PLAIN-X: %t = bitcast i64 1 to i64 -; PLAIN-X: ret i64 %t -; PLAIN-X: } -; PLAIN-X: define i64 @fg() nounwind { -; PLAIN-X: %t = bitcast i64 ptrtoint (double addrspace(4)* getelementptr ({ i1, double }* null, i64 0, i32 1) to i64) to i64 -; PLAIN-X: ret i64 %t -; PLAIN-X: } -; PLAIN-X: define i64 @fh() nounwind { -; PLAIN-X: %t = bitcast i64 ptrtoint (i1 addrspace(2)* getelementptr (i1 addrspace(2)* null, i32 1) to i64) to i64 -; PLAIN-X: ret i64 %t -; PLAIN-X: } -; PLAIN-X: define i64 @fi() nounwind { -; PLAIN-X: %t = bitcast i64 ptrtoint (i1 addrspace(2)* getelementptr ({ i1, i1 addrspace(2)* }* null, i64 0, i32 1) to i64) to i64 -; PLAIN-X: ret i64 %t -; PLAIN-X: } -define i64 @fa() nounwind { - %t = bitcast i64 mul (i64 3, i64 mul (i64 ptrtoint ({[7 x double], [7 x double]}* getelementptr ({[7 x double], [7 x double]}* null, i64 11) to i64), i64 5)) to i64 - ret i64 %t -} -define i64 @fb() nounwind { - %t = bitcast i64 ptrtoint ([13 x double] addrspace(4)* getelementptr ({i1, [13 x double]} addrspace(4)* null, i64 0, i32 1) to i64) to i64 - ret i64 %t -} -define i64 @fc() nounwind { - %t = bitcast i64 ptrtoint (double addrspace(4)* getelementptr ({double, double, double, double} addrspace(4)* null, i64 0, i32 2) to i64) to i64 - ret i64 %t -} -define i64 @fd() nounwind { - %t = bitcast i64 ptrtoint (double addrspace(4)* getelementptr ([13 x double] addrspace(4)* null, i64 0, i32 11) to i64) to i64 - ret i64 %t -} -define i64 @fe() nounwind { - %t = bitcast i64 ptrtoint (double addrspace(4)* getelementptr ({double, float, double, double} addrspace(4)* null, i64 0, i32 2) to i64) to i64 - ret i64 %t -} -define i64 @ff() nounwind { - %t = bitcast i64 ptrtoint (<{ i16, i128 }> addrspace(4)* getelementptr ({i1, <{ i16, i128 }>} addrspace(4)* null, i64 0, i32 1) to i64) to i64 - ret i64 %t -} -define i64 @fg() nounwind { - %t = bitcast i64 ptrtoint ({double, double} addrspace(4)* getelementptr ({i1, {double, double}} addrspace(4)* null, i64 0, i32 1) to i64) to i64 - ret i64 %t -} -define i64 @fh() nounwind { - %t = bitcast i64 ptrtoint (double addrspace(4)* getelementptr (double addrspace(4)* null, i32 1) to i64) to i64 - ret i64 %t -} -define i64 @fi() nounwind { - %t = bitcast i64 ptrtoint (double addrspace(4)* getelementptr ({i1, double}addrspace(4)* null, i64 0, i32 1) to i64) to i64 - ret i64 %t -} - -; PLAIN: define i64* @fM() nounwind { -; PLAIN: %t = bitcast i64* getelementptr (i64* null, i32 1) to i64* -; PLAIN: ret i64* %t -; PLAIN: } -; PLAIN: define i64* @fN() nounwind { -; PLAIN: %t = bitcast i64* getelementptr ({ i64, i64 }* null, i32 0, i32 1) to i64* -; PLAIN: ret i64* %t -; PLAIN: } -; PLAIN: define i64* @fO() nounwind { -; PLAIN: %t = bitcast i64* getelementptr ([2 x i64]* null, i32 0, i32 1) to i64* -; PLAIN: ret i64* %t -; PLAIN: } - -define i64* @fM() nounwind { - %t = bitcast i64* getelementptr (i64* null, i32 1) to i64* - ret i64* %t -} -define i64* @fN() nounwind { - %t = bitcast i64* getelementptr ({ i64, i64 }* null, i32 0, i32 1) to i64* - ret i64* %t -} -define i64* @fO() nounwind { - %t = bitcast i64* getelementptr ([2 x i64]* null, i32 0, i32 1) to i64* - ret i64* %t -} - -; PLAIN: define i32 addrspace(1)* @fZ() nounwind { -; PLAIN: %t = bitcast i32 addrspace(1)* getelementptr inbounds (i32 addrspace(1)* getelementptr inbounds ([3 x { i32, i32 }] addrspace(1)* @ext2, i64 0, i64 1, i32 0), i64 1) to i32 addrspace(1)* -; PLAIN: ret i32 addrspace(1)* %t -; PLAIN: } -@ext2 = external addrspace(1) global [3 x { i32, i32 }] -define i32 addrspace(1)* @fZ() nounwind { - %t = bitcast i32 addrspace(1)* getelementptr inbounds (i32 addrspace(1)* getelementptr inbounds ([3 x { i32, i32 }] addrspace(1)* @ext2, i64 0, i64 1, i32 0), i64 1) to i32 addrspace(1)* - ret i32 addrspace(1)* %t -} diff --git a/test/Transforms/InstCombine/fcmp.ll b/test/Transforms/InstCombine/fcmp.ll index fc1ced052a..376fa079d2 100644 --- a/test/Transforms/InstCombine/fcmp.ll +++ b/test/Transforms/InstCombine/fcmp.ll @@ -54,9 +54,8 @@ define i1 @test7(float %x) nounwind readnone ssp noredzone { %ext = fpext float %x to ppc_fp128 %cmp = fcmp ogt ppc_fp128 %ext, 0xM00000000000000000000000000000000 ret i1 %cmp -; Can't convert ppc_fp128 ; CHECK: @test7 -; CHECK-NEXT: fpext float %x to ppc_fp128 +; CHECK-NEXT: fcmp ogt float %x, 0.000000e+00 } define float @test8(float %x) nounwind readnone optsize ssp { diff --git a/test/Transforms/InstCombine/stpcpy-1.ll b/test/Transforms/InstCombine/stpcpy-1.ll new file mode 100644 index 0000000000..8b6bb0e0d5 --- /dev/null +++ b/test/Transforms/InstCombine/stpcpy-1.ll @@ -0,0 +1,46 @@ +; Test that the stpcpy library call simplifier works correctly. +; RUN: opt < %s -instcombine -S | FileCheck %s +; +; This transformation requires the pointer size, as it assumes that size_t is +; the size of a pointer. +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128-n8:16:32" + +@hello = constant [6 x i8] c"hello\00" +@a = common global [32 x i8] zeroinitializer, align 1 +@b = common global [32 x i8] zeroinitializer, align 1 + +declare i8* @stpcpy(i8*, i8*) + +define i8* @test_simplify1() { +; CHECK: @test_simplify1 + + %dst = getelementptr [32 x i8]* @a, i32 0, i32 0 + %src = getelementptr [6 x i8]* @hello, i32 0, i32 0 + + %ret = call i8* @stpcpy(i8* %dst, i8* %src) +; CHECK: @llvm.memcpy.p0i8.p0i8.i32 +; CHECK-NEXT: getelementptr inbounds ([32 x i8]* @a, i32 0, i32 5) + ret i8* %ret +} + +define i8* @test_simplify2() { +; CHECK: @test_simplify2 + + %dst = getelementptr [32 x i8]* @a, i32 0, i32 0 + + %ret = call i8* @stpcpy(i8* %dst, i8* %dst) +; CHECK: [[LEN:%[a-z]+]] = call i32 @strlen +; CHECK-NEXT: getelementptr inbounds [32 x i8]* @a, i32 0, i32 [[LEN]] + ret i8* %ret +} + +define i8* @test_no_simplify1() { +; CHECK: @test_no_simplify1 + + %dst = getelementptr [32 x i8]* @a, i32 0, i32 0 + %src = getelementptr [32 x i8]* @b, i32 0, i32 0 + + %ret = call i8* @stpcpy(i8* %dst, i8* %src) +; CHECK: call i8* @stpcpy + ret i8* %ret +} diff --git a/test/Transforms/InstCombine/stpcpy-2.ll b/test/Transforms/InstCombine/stpcpy-2.ll new file mode 100644 index 0000000000..2e92c0895e --- /dev/null +++ b/test/Transforms/InstCombine/stpcpy-2.ll @@ -0,0 +1,22 @@ +; Test that the stpcpy library call simplifier works correctly. +; RUN: opt < %s -instcombine -S | FileCheck %s +; +; This transformation requires the pointer size, as it assumes that size_t is +; the size of a pointer. +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128-n8:16:32" + +@hello = constant [6 x i8] c"hello\00" +@a = common global [32 x i8] zeroinitializer, align 1 + +declare i16* @stpcpy(i8*, i8*) + +define void @test_no_simplify1() { +; CHECK: @test_no_simplify1 + + %dst = getelementptr [32 x i8]* @a, i32 0, i32 0 + %src = getelementptr [6 x i8]* @hello, i32 0, i32 0 + + call i16* @stpcpy(i8* %dst, i8* %src) +; CHECK: call i16* @stpcpy + ret void +} diff --git a/test/Transforms/InstCombine/stpcpy_chk-1.ll b/test/Transforms/InstCombine/stpcpy_chk-1.ll new file mode 100644 index 0000000000..05603918c6 --- /dev/null +++ b/test/Transforms/InstCombine/stpcpy_chk-1.ll @@ -0,0 +1,96 @@ +; Test lib call simplification of __stpcpy_chk calls with various values +; for src, dst, and slen. +; +; RUN: opt < %s -instcombine -S | FileCheck %s + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" + +@a = common global [60 x i8] zeroinitializer, align 1 +@b = common global [60 x i8] zeroinitializer, align 1 +@.str = private constant [12 x i8] c"abcdefghijk\00" + +; Check cases where slen >= strlen (src). + +define void @test_simplify1() { +; CHECK: @test_simplify1 + %dst = getelementptr inbounds [60 x i8]* @a, i32 0, i32 0 + %src = getelementptr inbounds [12 x i8]* @.str, i32 0, i32 0 + +; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32 + call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 60) + ret void +} + +define void @test_simplify2() { +; CHECK: @test_simplify2 + %dst = getelementptr inbounds [60 x i8]* @a, i32 0, i32 0 + %src = getelementptr inbounds [12 x i8]* @.str, i32 0, i32 0 + +; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32 + call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 12) + ret void +} + +define void @test_simplify3() { +; CHECK: @test_simplify3 + %dst = getelementptr inbounds [60 x i8]* @a, i32 0, i32 0 + %src = getelementptr inbounds [12 x i8]* @.str, i32 0, i32 0 + +; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32 + call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 -1) + ret void +} + +; Check cases where there are no string constants. + +define void @test_simplify4() { +; CHECK: @test_simplify4 + %dst = getelementptr inbounds [60 x i8]* @a, i32 0, i32 0 + %src = getelementptr inbounds [60 x i8]* @b, i32 0, i32 0 + +; CHECK-NEXT: call i8* @stpcpy + call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 -1) + ret void +} + +; Check case where the string length is not constant. + +define i8* @test_simplify5() { +; CHECK: @test_simplify5 + %dst = getelementptr inbounds [60 x i8]* @a, i32 0, i32 0 + %src = getelementptr inbounds [12 x i8]* @.str, i32 0, i32 0 + +; CHECK: @__memcpy_chk + %len = call i32 @llvm.objectsize.i32(i8* %dst, i1 false) + %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 %len) +; CHECK: ret i8* getelementptr inbounds ([60 x i8]* @a, i32 0, i32 11) + ret i8* %ret +} + +; Check case where the source and destination are the same. + +define i8* @test_simplify6() { +; CHECK: @test_simplify6 + %dst = getelementptr inbounds [60 x i8]* @a, i32 0, i32 0 + +; CHECK: [[LEN:%[a-z]+]] = call i32 @strlen +; CHECK-NEXT: getelementptr inbounds [60 x i8]* @a, i32 0, i32 [[LEN]] + %len = call i32 @llvm.objectsize.i32(i8* %dst, i1 false) + %ret = call i8* @__stpcpy_chk(i8* %dst, i8* %dst, i32 %len) + ret i8* %ret +} + +; Check case where slen < strlen (src). + +define void @test_no_simplify1() { +; CHECK: @test_no_simplify1 + %dst = getelementptr inbounds [60 x i8]* @a, i32 0, i32 0 + %src = getelementptr inbounds [60 x i8]* @b, i32 0, i32 0 + +; CHECK-NEXT: call i8* @__stpcpy_chk + call i8* @__stpcpy_chk(i8* %dst, i8* %src, i32 8) + ret void +} + +declare i8* @__stpcpy_chk(i8*, i8*, i32) nounwind +declare i32 @llvm.objectsize.i32(i8*, i1) nounwind readonly diff --git a/test/Transforms/InstCombine/stpcpy_chk-2.ll b/test/Transforms/InstCombine/stpcpy_chk-2.ll new file mode 100644 index 0000000000..46c2139276 --- /dev/null +++ b/test/Transforms/InstCombine/stpcpy_chk-2.ll @@ -0,0 +1,21 @@ +; Test that lib call simplification doesn't simplify __stpcpy_chk calls +; with the wrong prototype. +; +; RUN: opt < %s -instcombine -S | FileCheck %s + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" + +@a = common global [60 x i16] zeroinitializer, align 1 +@.str = private constant [8 x i8] c"abcdefg\00" + +define void @test_no_simplify() { +; CHECK: @test_no_simplify + %dst = getelementptr inbounds [60 x i16]* @a, i32 0, i32 0 + %src = getelementptr inbounds [8 x i8]* @.str, i32 0, i32 0 + +; CHECK-NEXT: call i16* @__strcpy_chk + call i16* @__strcpy_chk(i16* %dst, i8* %src, i32 8) + ret void +} + +declare i16* @__strcpy_chk(i16*, i8*, i32) diff --git a/test/Transforms/InstCombine/strlen-1.ll b/test/Transforms/InstCombine/strlen-1.ll new file mode 100644 index 0000000000..6d7464a4cc --- /dev/null +++ b/test/Transforms/InstCombine/strlen-1.ll @@ -0,0 +1,97 @@ +; Test that the strlen library call simplifier works correctly. +; +; RUN: opt < %s -instcombine -S | FileCheck %s + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" + +@hello = constant [6 x i8] c"hello\00" +@null = constant [1 x i8] zeroinitializer +@null_hello = constant [7 x i8] c"\00hello\00" +@nullstring = constant i8 0 +@a = common global [32 x i8] zeroinitializer, align 1 + +declare i32 @strlen(i8*) + +; Check strlen(string constant) -> integer constant. + +define i32 @test_simplify1() { +; CHECK: @test_simplify1 + %hello_p = getelementptr [6 x i8]* @hello, i32 0, i32 0 + %hello_l = call i32 @strlen(i8* %hello_p) + ret i32 %hello_l +; CHECK-NEXT: ret i32 5 +} + +define i32 @test_simplify2() { +; CHECK: @test_simplify2 + %null_p = getelementptr [1 x i8]* @null, i32 0, i32 0 + %null_l = call i32 @strlen(i8* %null_p) + ret i32 %null_l +; CHECK-NEXT: ret i32 0 +} + +define i32 @test_simplify3() { +; CHECK: @test_simplify3 + %null_hello_p = getelementptr [7 x i8]* @null_hello, i32 0, i32 0 + %null_hello_l = call i32 @strlen(i8* %null_hello_p) + ret i32 %null_hello_l +; CHECK-NEXT: ret i32 0 +} + +define i32 @test_simplify4() { +; CHECK: @test_simplify4 + %len = tail call i32 @strlen(i8* @nullstring) nounwind + ret i32 %len +; CHECK-NEXT: ret i32 0 +} + +; Check strlen(x) == 0 --> *x == 0. + +define i1 @test_simplify5() { +; CHECK: @test_simplify5 + %hello_p = getelementptr [6 x i8]* @hello, i32 0, i32 0 + %hello_l = call i32 @strlen(i8* %hello_p) + %eq_hello = icmp eq i32 %hello_l, 0 + ret i1 %eq_hello +; CHECK-NEXT: ret i1 false +} + +define i1 @test_simplify6() { +; CHECK: @test_simplify6 + %null_p = getelementptr [1 x i8]* @null, i32 0, i32 0 + %null_l = call i32 @strlen(i8* %null_p) + %eq_null = icmp eq i32 %null_l, 0 + ret i1 %eq_null +; CHECK-NEXT: ret i1 true +} + +; Check strlen(x) != 0 --> *x != 0. + +define i1 @test_simplify7() { +; CHECK: @test_simplify7 + %hello_p = getelementptr [6 x i8]* @hello, i32 0, i32 0 + %hello_l = call i32 @strlen(i8* %hello_p) + %ne_hello = icmp ne i32 %hello_l, 0 + ret i1 %ne_hello +; CHECK-NEXT: ret i1 true +} + +define i1 @test_simplify8() { +; CHECK: @test_simplify8 + %null_p = getelementptr [1 x i8]* @null, i32 0, i32 0 + %null_l = call i32 @strlen(i8* %null_p) + %ne_null = icmp ne i32 %null_l, 0 + ret i1 %ne_null +; CHECK-NEXT: ret i1 false +} + +; Check cases that shouldn't be simplified. + +define i32 @test_no_simplify1() { +; CHECK: @test_no_simplify1 + %a_p = getelementptr [32 x i8]* @a, i32 0, i32 0 + %a_l = call i32 @strlen(i8* %a_p) +; CHECK-NEXT: %a_l = call i32 @strlen + ret i32 %a_l +; CHECK-NEXT: ret i32 %a_l +} diff --git a/test/Transforms/InstCombine/strlen-2.ll b/test/Transforms/InstCombine/strlen-2.ll new file mode 100644 index 0000000000..c4fd54c06d --- /dev/null +++ b/test/Transforms/InstCombine/strlen-2.ll @@ -0,0 +1,18 @@ +; Test that the strlen library call simplifier works correctly. +; +; RUN: opt < %s -instcombine -S | FileCheck %s + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" + +@hello = constant [6 x i8] c"hello\00" + +declare i32 @strlen(i8*, i32) + +define i32 @test_no_simplify1() { +; CHECK: @test_no_simplify1 + %hello_p = getelementptr [6 x i8]* @hello, i32 0, i32 0 + %hello_l = call i32 @strlen(i8* %hello_p, i32 187) +; CHECK-NEXT: %hello_l = call i32 @strlen + ret i32 %hello_l +; CHECK-NEXT: ret i32 %hello_l +} diff --git a/test/Transforms/InstCombine/strncpy-1.ll b/test/Transforms/InstCombine/strncpy-1.ll new file mode 100644 index 0000000000..3ce2b9b5ee --- /dev/null +++ b/test/Transforms/InstCombine/strncpy-1.ll @@ -0,0 +1,95 @@ +; Test that the strncpy library call simplifier works correctly. +; +; RUN: opt < %s -instcombine -S | FileCheck %s + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" + +@hello = constant [6 x i8] c"hello\00" +@null = constant [1 x i8] zeroinitializer +@null_hello = constant [7 x i8] c"\00hello\00" +@a = common global [32 x i8] zeroinitializer, align 1 +@b = common global [32 x i8] zeroinitializer, align 1 + +declare i8* @strncpy(i8*, i8*, i32) +declare i32 @puts(i8*) + +; Check a bunch of strncpy invocations together. + +define i32 @test_simplify1() { +; CHECK: @test_simplify1 +; CHECK-NOT: call i8* @strncpy +; CHECK: call i32 @puts + %target = alloca [1024 x i8] + %arg1 = getelementptr [1024 x i8]* %target, i32 0, i32 0 + store i8 0, i8* %arg1 + + %arg2 = getelementptr [6 x i8]* @hello, i32 0, i32 0 + %rslt1 = call i8* @strncpy(i8* %arg1, i8* %arg2, i32 6) + + %arg3 = getelementptr [1 x i8]* @null, i32 0, i32 0 + %rslt2 = call i8* @strncpy(i8* %rslt1, i8* %arg3, i32 42) + + %arg4 = getelementptr [7 x i8]* @null_hello, i32 0, i32 0 + %rslt3 = call i8* @strncpy(i8* %rslt2, i8* %arg4, i32 42) + + call i32 @puts( i8* %rslt3 ) + ret i32 0 +} + +; Check strncpy(x, "", y) -> memset(x, '\0', y, 1). + +define void @test_simplify2() { +; CHECK: @test_simplify2 + %dst = getelementptr [32 x i8]* @a, i32 0, i32 0 + %src = getelementptr [1 x i8]* @null, i32 0, i32 0 + + call i8* @strncpy(i8* %dst, i8* %src, i32 32) +; CHECK: call void @llvm.memset.p0i8.i32 + ret void +} + +; Check strncpy(x, y, 0) -> x. + +define i8* @test_simplify3() { +; CHECK: @test_simplify3 + %dst = getelementptr [32 x i8]* @a, i32 0, i32 0 + %src = getelementptr [6 x i8]* @hello, i32 0, i32 0 + + %ret = call i8* @strncpy(i8* %dst, i8* %src, i32 0) + ret i8* %ret +; CHECK: ret i8* getelementptr inbounds ([32 x i8]* @a, i32 0, i32 0) +} + +; Check strncpy(x, s, c) -> memcpy(x, s, c, 1) [s and c are constant]. + +define void @test_simplify4() { +; CHECK: @test_simplify4 + %dst = getelementptr [32 x i8]* @a, i32 0, i32 0 + %src = getelementptr [6 x i8]* @hello, i32 0, i32 0 + + call i8* @strncpy(i8* %dst, i8* %src, i32 6) +; CHECK: call void @llvm.memcpy.p0i8.p0i8.i32 + ret void +} + +; Check cases that shouldn't be simplified. + +define void @test_no_simplify1() { +; CHECK: @test_no_simplify1 + %dst = getelementptr [32 x i8]* @a, i32 0, i32 0 + %src = getelementptr [32 x i8]* @b, i32 0, i32 0 + + call i8* @strncpy(i8* %dst, i8* %src, i32 32) +; CHECK: call i8* @strncpy + ret void +} + +define void @test_no_simplify2() { +; CHECK: @test_no_simplify2 + %dst = getelementptr [32 x i8]* @a, i32 0, i32 0 + %src = getelementptr [6 x i8]* @hello, i32 0, i32 0 + + call i8* @strncpy(i8* %dst, i8* %src, i32 8) +; CHECK: call i8* @strncpy + ret void +} diff --git a/test/Transforms/InstCombine/strncpy-2.ll b/test/Transforms/InstCombine/strncpy-2.ll new file mode 100644 index 0000000000..ac28ea6550 --- /dev/null +++ b/test/Transforms/InstCombine/strncpy-2.ll @@ -0,0 +1,22 @@ +; Test that the strncpy library call simplifier works correctly. +; +; RUN: opt < %s -instcombine -S | FileCheck %s + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" + +@hello = constant [6 x i8] c"hello\00" +@a = common global [32 x i8] zeroinitializer, align 1 + +declare i16* @strncpy(i8*, i8*, i32) + +; Check that 'strncpy' functions with the wrong prototype aren't simplified. + +define void @test_no_simplify1() { +; CHECK: @test_no_simplify1 + %dst = getelementptr [32 x i8]* @a, i32 0, i32 0 + %src = getelementptr [6 x i8]* @hello, i32 0, i32 0 + + call i16* @strncpy(i8* %dst, i8* %src, i32 6) +; CHECK: call i16* @strncpy + ret void +} diff --git a/test/Transforms/InstCombine/strncpy_chk-1.ll b/test/Transforms/InstCombine/strncpy_chk-1.ll index ae7e2fb5f1..aadff4268e 100644 --- a/test/Transforms/InstCombine/strncpy_chk-1.ll +++ b/test/Transforms/InstCombine/strncpy_chk-1.ll @@ -7,27 +7,27 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f3 @a = common global [60 x i8] zeroinitializer, align 1 @b = common global [60 x i8] zeroinitializer, align 1 -@.str = private constant [8 x i8] c"abcdefg\00" +@.str = private constant [12 x i8] c"abcdefghijk\00" ; Check cases where dstlen >= len define void @test_simplify1() { ; CHECK: @test_simplify1 %dst = getelementptr inbounds [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [8 x i8]* @.str, i32 0, i32 0 + %src = getelementptr inbounds [12 x i8]* @.str, i32 0, i32 0 -; CHECK-NEXT: call i8* @strncpy - call i8* @__strncpy_chk(i8* %dst, i8* %src, i32 8, i32 60) +; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32 + call i8* @__strncpy_chk(i8* %dst, i8* %src, i32 12, i32 60) ret void } define void @test_simplify2() { ; CHECK: @test_simplify2 %dst = getelementptr inbounds [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [8 x i8]* @.str, i32 0, i32 0 + %src = getelementptr inbounds [12 x i8]* @.str, i32 0, i32 0 -; CHECK-NEXT: call i8* @strncpy - call i8* @__strncpy_chk(i8* %dst, i8* %src, i32 8, i32 8) +; CHECK-NEXT: call void @llvm.memcpy.p0i8.p0i8.i32 + call i8* @__strncpy_chk(i8* %dst, i8* %src, i32 12, i32 12) ret void } @@ -37,7 +37,7 @@ define void @test_simplify3() { %src = getelementptr inbounds [60 x i8]* @b, i32 0, i32 0 ; CHECK-NEXT: call i8* @strncpy - call i8* @__strncpy_chk(i8* %dst, i8* %src, i32 8, i32 60) + call i8* @__strncpy_chk(i8* %dst, i8* %src, i32 12, i32 60) ret void } @@ -46,7 +46,7 @@ define void @test_simplify3() { define void @test_no_simplify1() { ; CHECK: @test_no_simplify1 %dst = getelementptr inbounds [60 x i8]* @a, i32 0, i32 0 - %src = getelementptr inbounds [8 x i8]* @.str, i32 0, i32 0 + %src = getelementptr inbounds [12 x i8]* @.str, i32 0, i32 0 ; CHECK-NEXT: call i8* @__strncpy_chk call i8* @__strncpy_chk(i8* %dst, i8* %src, i32 8, i32 4) diff --git a/test/Transforms/InstCombine/strpbrk-1.ll b/test/Transforms/InstCombine/strpbrk-1.ll new file mode 100644 index 0000000000..a5d0d86501 --- /dev/null +++ b/test/Transforms/InstCombine/strpbrk-1.ll @@ -0,0 +1,68 @@ +; Test that the strpbrk library call simplifier works correctly. +; +; RUN: opt < %s -instcombine -S | FileCheck %s + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" + +@hello = constant [12 x i8] c"hello world\00" +@w = constant [2 x i8] c"w\00" +@null = constant [1 x i8] zeroinitializer + +declare i8* @strpbrk(i8*, i8*) + +; Check strpbrk(s, "") -> NULL. + +define i8* @test_simplify1(i8* %str) { +; CHECK: @test_simplify1 + %pat = getelementptr [1 x i8]* @null, i32 0, i32 0 + + %ret = call i8* @strpbrk(i8* %str, i8* %pat) + ret i8* %ret +; CHECK-NEXT: ret i8* null +} + +; Check strpbrk("", s) -> NULL. + +define i8* @test_simplify2(i8* %pat) { +; CHECK: @test_simplify2 + %str = getelementptr [1 x i8]* @null, i32 0, i32 0 + + %ret = call i8* @strpbrk(i8* %str, i8* %pat) + ret i8* %ret +; CHECK-NEXT: ret i8* null +} + +; Check strpbrk(s1, s2), where s1 and s2 are constants. + +define i8* @test_simplify3() { +; CHECK: @test_simplify3 + %str = getelementptr [12 x i8]* @hello, i32 0, i32 0 + %pat = getelementptr [2 x i8]* @w, i32 0, i32 0 + + %ret = call i8* @strpbrk(i8* %str, i8* %pat) + ret i8* %ret +; CHECK-NEXT: ret i8* getelementptr inbounds ([12 x i8]* @hello, i32 0, i32 6) +} + +; Check strpbrk(s, "a") -> strchr(s, 'a'). + +define i8* @test_simplify4(i8* %str) { +; CHECK: @test_simplify4 + %pat = getelementptr [2 x i8]* @w, i32 0, i32 0 + + %ret = call i8* @strpbrk(i8* %str, i8* %pat) +; CHECK-NEXT: [[VAR:%[a-z]+]] = call i8* @strchr(i8* %str, i32 119) + ret i8* %ret +; CHECK-NEXT: ret i8* [[VAR]] +} + +; Check cases that shouldn't be simplified. + +define i8* @test_no_simplify1(i8* %str, i8* %pat) { +; CHECK: @test_no_simplify1 + + %ret = call i8* @strpbrk(i8* %str, i8* %pat) +; CHECK-NEXT: %ret = call i8* @strpbrk(i8* %str, i8* %pat) + ret i8* %ret +; CHECK-NEXT: ret i8* %ret +} diff --git a/test/Transforms/InstCombine/strpbrk-2.ll b/test/Transforms/InstCombine/strpbrk-2.ll new file mode 100644 index 0000000000..31ac2905df --- /dev/null +++ b/test/Transforms/InstCombine/strpbrk-2.ll @@ -0,0 +1,23 @@ +; Test that the strpbrk library call simplifier works correctly. +; +; RUN: opt < %s -instcombine -S | FileCheck %s + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" + +@hello = constant [12 x i8] c"hello world\00" +@w = constant [2 x i8] c"w\00" + +declare i16* @strpbrk(i8*, i8*) + +; Check that 'strpbrk' functions with the wrong prototype aren't simplified. + +define i16* @test_no_simplify1() { +; CHECK: @test_no_simplify1 + %str = getelementptr [12 x i8]* @hello, i32 0, i32 0 + %pat = getelementptr [2 x i8]* @w, i32 0, i32 0 + + %ret = call i16* @strpbrk(i8* %str, i8* %pat) +; CHECK-NEXT: %ret = call i16* @strpbrk + ret i16* %ret +; CHECK-NEXT: ret i16* %ret +} diff --git a/test/Transforms/InstCombine/strto-1.ll b/test/Transforms/InstCombine/strto-1.ll new file mode 100644 index 0000000000..16c0c67970 --- /dev/null +++ b/test/Transforms/InstCombine/strto-1.ll @@ -0,0 +1,82 @@ +; Test that the strto* library call simplifiers works correctly. +; +; RUN: opt < %s -instcombine -S | FileCheck %s + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" + +declare i64 @strtol(i8* %s, i8** %endptr, i32 %base) +; CHECK: declare i64 @strtol(i8*, i8**, i32) + +declare double @strtod(i8* %s, i8** %endptr, i32 %base) +; CHECK: declare double @strtod(i8*, i8**, i32) + +declare float @strtof(i8* %s, i8** %endptr, i32 %base) +; CHECK: declare float @strtof(i8*, i8**, i32) + +declare i64 @strtoul(i8* %s, i8** %endptr, i32 %base) +; CHECK: declare i64 @strtoul(i8*, i8**, i32) + +declare i64 @strtoll(i8* %s, i8** %endptr, i32 %base) +; CHECK: declare i64 @strtoll(i8*, i8**, i32) + +declare double @strtold(i8* %s, i8** %endptr) +; CHECK: declare double @strtold(i8*, i8**) + +declare i64 @strtoull(i8* %s, i8** %endptr, i32 %base) +; CHECK: declare i64 @strtoull(i8*, i8**, i32) + +define void @test_simplify1(i8* %x, i8** %endptr) { +; CHECK: @test_simplify1 + call i64 @strtol(i8* %x, i8** null, i32 10) +; CHECK-NEXT: call i64 @strtol(i8* nocapture %x, i8** null, i32 10) + ret void +} + +define void @test_simplify2(i8* %x, i8** %endptr) { +; CHECK: @test_simplify2 + call double @strtod(i8* %x, i8** null, i32 10) +; CHECK-NEXT: call double @strtod(i8* nocapture %x, i8** null, i32 10) + ret void +} + +define void @test_simplify3(i8* %x, i8** %endptr) { +; CHECK: @test_simplify3 + call float @strtof(i8* %x, i8** null, i32 10) +; CHECK-NEXT: call float @strtof(i8* nocapture %x, i8** null, i32 10) + ret void +} + +define void @test_simplify4(i8* %x, i8** %endptr) { +; CHECK: @test_simplify4 + call i64 @strtoul(i8* %x, i8** null, i32 10) +; CHECK-NEXT: call i64 @strtoul(i8* nocapture %x, i8** null, i32 10) + ret void +} + +define void @test_simplify5(i8* %x, i8** %endptr) { +; CHECK: @test_simplify5 + call i64 @strtoll(i8* %x, i8** null, i32 10) +; CHECK-NEXT: call i64 @strtoll(i8* nocapture %x, i8** null, i32 10) + ret void +} + +define void @test_simplify6(i8* %x, i8** %endptr) { +; CHECK: @test_simplify6 + call double @strtold(i8* %x, i8** null) +; CHECK-NEXT: call double @strtold(i8* nocapture %x, i8** null) + ret void +} + +define void @test_simplify7(i8* %x, i8** %endptr) { +; CHECK: @test_simplify7 + call i64 @strtoull(i8* %x, i8** null, i32 10) +; CHECK-NEXT: call i64 @strtoull(i8* nocapture %x, i8** null, i32 10) + ret void +} + +define void @test_no_simplify1(i8* %x, i8** %endptr) { +; CHECK: @test_no_simplify1 + call i64 @strtol(i8* %x, i8** %endptr, i32 10) +; CHECK-NEXT: call i64 @strtol(i8* %x, i8** %endptr, i32 10) + ret void +} diff --git a/test/Transforms/InstCombine/vector_gep2.ll b/test/Transforms/InstCombine/vector_gep2.ll new file mode 100644 index 0000000000..20165b1100 --- /dev/null +++ b/test/Transforms/InstCombine/vector_gep2.ll @@ -0,0 +1,11 @@ +; RUN: opt < %s -instcombine -S | FileCheck %s + +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-unknown-linux-gnu" + +define <2 x i8*> @testa(<2 x i8*> %a) { +; CHECK: @testa + %g = getelementptr <2 x i8*> %a, <2 x i32> <i32 0, i32 1> +; CHECK: getelementptr <2 x i8*> %a, <2 x i64> <i64 0, i64 1> + ret <2 x i8*> %g +} diff --git a/test/Transforms/Internalize/2008-05-09-AllButMain.ll b/test/Transforms/Internalize/2008-05-09-AllButMain.ll index a85e834582..c07abb0c63 100644 --- a/test/Transforms/Internalize/2008-05-09-AllButMain.ll +++ b/test/Transforms/Internalize/2008-05-09-AllButMain.ll @@ -1,27 +1,55 @@ -; No arguments means internalize all but main -; RUN: opt < %s -internalize -S | grep internal | count 4 +; No arguments means internalize everything +; RUN: opt < %s -internalize -S | FileCheck --check-prefix=NOARGS %s + ; Internalize all but foo and j -; RUN: opt < %s -internalize -internalize-public-api-list foo -internalize-public-api-list j -S | grep internal | count 3 -; Non existent files should be treated as if they were empty (so internalize all but main) -; RUN: opt < %s -internalize -internalize-public-api-file /nonexistent/file 2> /dev/null -S | grep internal | count 4 -; RUN: opt < %s -internalize -internalize-public-api-list bar -internalize-public-api-list foo -internalize-public-api-file /nonexistent/file 2> /dev/null -S | grep internal | count 3 +; RUN: opt < %s -internalize -internalize-public-api-list foo -internalize-public-api-list j -S | FileCheck --check-prefix=LIST %s + +; Non existent files should be treated as if they were empty (so internalize +; everything) +; RUN: opt < %s -internalize -internalize-public-api-file /nonexistent/file 2> /dev/null -S | FileCheck --check-prefix=EMPTYFILE %s + +; RUN: opt < %s -S -internalize -internalize-public-api-list bar -internalize-public-api-list foo -internalize-public-api-file /nonexistent/file 2> /dev/null | FileCheck --check-prefix=LIST2 %s + ; -file and -list options should be merged, the .apifile contains foo and j -; RUN: opt < %s -internalize -internalize-public-api-list bar -internalize-public-api-file %s.apifile -S | grep internal | count 2 +; RUN: opt < %s -internalize -internalize-public-api-list bar -internalize-public-api-file %s.apifile -S | FileCheck --check-prefix=MERGE %s + +; NOARGS: @i = internal global +; LIST: @i = internal global +; EMPTYFILE: @i = internal global +; LIST2: @i = internal global +; MERGE: @i = internal global +@i = global i32 0 -@i = weak global i32 0 ; <i32*> [#uses=0] -@j = weak global i32 0 ; <i32*> [#uses=0] +; NOARGS: @j = internal global +; LIST: @j = global +; EMPTYFILE: @j = internal global +; LIST2: @j = internal global +; MERGE: @j = global +@j = global i32 0 -define void @main(...) { -entry: +; NOARGS: define internal void @main +; LIST: define internal void @main +; EMPTYFILE: define internal void @main +; LIST2: define internal void @main +; MERGE: define internal void @main +define void @main() { ret void } -define void @foo(...) { -entry: +; NOARGS: define internal void @foo +; LIST: define void @foo +; EMPTYFILE: define internal void @foo +; LIST2: define void @foo +; MERGE: define void @foo +define void @foo() { ret void } -define void @bar(...) { -entry: +; NOARGS: define internal void @bar +; LIST: define internal void @bar +; EMPTYFILE: define internal void @bar +; LIST2: define void @bar +; MERGE: define void @bar +define void @bar() { ret void } diff --git a/test/Transforms/Internalize/2009-01-05-InternalizeAliases.ll b/test/Transforms/Internalize/2009-01-05-InternalizeAliases.ll index 7b18a04e11..47cf3f0373 100644 --- a/test/Transforms/Internalize/2009-01-05-InternalizeAliases.ll +++ b/test/Transforms/Internalize/2009-01-05-InternalizeAliases.ll @@ -1,4 +1,4 @@ -; RUN: opt < %s -internalize -S | grep internal | count 3 +; RUN: opt < %s -internalize -internalize-public-api-list main -S | grep internal | count 3 @A = global i32 0 @B = alias i32* @A diff --git a/test/Transforms/JumpThreading/crash.ll b/test/Transforms/JumpThreading/crash.ll index b9c03544db..2fe87464c1 100644 --- a/test/Transforms/JumpThreading/crash.ll +++ b/test/Transforms/JumpThreading/crash.ll @@ -511,3 +511,56 @@ lbl_260: ; preds = %for.cond, %entry if.end: ; preds = %for.cond ret void } + +define void @PR14233(i1 %cmp, i1 %cmp2, i1 %cmp3, i1 %cmp4) { +entry: + br i1 %cmp, label %cond.true, label %cond.false + +cond.true: + br label %if.end + +cond.false: + br label %if.end + +if.end: + %A = phi i64 [ 0, %cond.true ], [ 1, %cond.false ] + br i1 %cmp2, label %bb, label %if.end2 + +bb: + br label %if.end2 + +if.end2: + %B = phi i64 [ ptrtoint (i8* ()* @PR14233.f1 to i64), %bb ], [ %A, %if.end ] + %cmp.ptr = icmp eq i64 %B, ptrtoint (i8* ()* @PR14233.f2 to i64) + br i1 %cmp.ptr, label %cond.true2, label %if.end3 + +cond.true2: + br i1 %cmp3, label %bb2, label %ur + +bb2: + br i1 %cmp4, label %if.end4, label %if.end3 + +if.end4: + unreachable + +if.end3: + %cmp.ptr2 = icmp eq i64 %B, ptrtoint (i8* ()* @PR14233.f2 to i64) + br i1 %cmp.ptr2, label %ur, label %if.then601 + +if.then601: + %C = icmp eq i64 %B, 0 + br i1 %C, label %bb3, label %bb4 + +bb3: + unreachable + +bb4: + unreachable + +ur: + unreachable +} + +declare i8* @PR14233.f1() + +declare i8* @PR14233.f2() diff --git a/test/Transforms/LICM/2003-12-11-SinkingToPHI.ll b/test/Transforms/LICM/2003-12-11-SinkingToPHI.ll index 67c3951d74..fe8d445313 100644 --- a/test/Transforms/LICM/2003-12-11-SinkingToPHI.ll +++ b/test/Transforms/LICM/2003-12-11-SinkingToPHI.ll @@ -1,4 +1,4 @@ -; RUN: opt < %s -licm | lli +; RUN: opt < %s -licm | lli %defaultjit define i32 @main() { entry: diff --git a/test/Transforms/LoopIdiom/basic.ll b/test/Transforms/LoopIdiom/basic.ll index 46ab7e5542..06a5bd9086 100644 --- a/test/Transforms/LoopIdiom/basic.ll +++ b/test/Transforms/LoopIdiom/basic.ll @@ -383,4 +383,37 @@ for.end: ; preds = %for.inc } +define void @PR14241(i32* %s, i64 %size) { +; Ensure that we don't form a memcpy for strided loops. Briefly, when we taught +; LoopIdiom about memmove and strided loops, this got miscompiled into a memcpy +; instead of a memmove. If we get the memmove transform back, this will catch +; regressions. +; +; CHECK: @PR14241 +entry: + %end.idx = add i64 %size, -1 + %end.ptr = getelementptr inbounds i32* %s, i64 %end.idx + br label %while.body +; CHECK-NOT: memcpy +; +; FIXME: When we regain the ability to form a memmove here, this test should be +; reversed and turned into a positive assertion. +; CHECK-NOT: memmove + +while.body: + %phi.ptr = phi i32* [ %s, %entry ], [ %next.ptr, %while.body ] + %src.ptr = getelementptr inbounds i32* %phi.ptr, i64 1 + %val = load i32* %src.ptr, align 4 +; CHECK: load + %dst.ptr = getelementptr inbounds i32* %phi.ptr, i64 0 + store i32 %val, i32* %dst.ptr, align 4 +; CHECK: store + %next.ptr = getelementptr inbounds i32* %phi.ptr, i64 1 + %cmp = icmp eq i32* %next.ptr, %end.ptr + br i1 %cmp, label %exit, label %while.body + +exit: + ret void +; CHECK: ret void +} diff --git a/test/Transforms/LoopIdiom/crash.ll b/test/Transforms/LoopIdiom/crash.ll new file mode 100644 index 0000000000..969adbcd76 --- /dev/null +++ b/test/Transforms/LoopIdiom/crash.ll @@ -0,0 +1,25 @@ +; RUN: opt -basicaa -loop-idiom -S < %s +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +; Don't crash inside DependenceAnalysis +; PR14219 +define void @test1(i64* %iwork, i64 %x) { +bb0: + %mul116 = mul nsw i64 %x, %x + %incdec.ptr6.sum175 = add i64 42, %x + %arrayidx135 = getelementptr inbounds i64* %iwork, i64 %incdec.ptr6.sum175 + br label %bb1 +bb1: + %storemerge4226 = phi i64 [ 0, %bb0 ], [ %inc139, %bb1 ] + store i64 1, i64* %arrayidx135, align 8 + %incdec.ptr6.sum176 = add i64 %mul116, %storemerge4226 + %arrayidx137 = getelementptr inbounds i64* %iwork, i64 %incdec.ptr6.sum176 + store i64 1, i64* %arrayidx137, align 8 + %inc139 = add nsw i64 %storemerge4226, 1 + %cmp131 = icmp sgt i64 %storemerge4226, 42 + br i1 %cmp131, label %bb2, label %bb1 +bb2: + ret void +} + diff --git a/test/Transforms/LoopIdiom/scev-invalidation.ll b/test/Transforms/LoopIdiom/scev-invalidation.ll new file mode 100644 index 0000000000..a244d9a280 --- /dev/null +++ b/test/Transforms/LoopIdiom/scev-invalidation.ll @@ -0,0 +1,74 @@ +; RUN: opt -S -indvars -loop-idiom < %s +; PR14214 +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-unknown-linux-gnu" + +define i32 @quote_arg() nounwind { +entry: + br label %for.cond + +for.cond: ; preds = %for.inc, %entry + %backslashes.0 = phi i32 [ undef, %entry ], [ %backslashes.2, %for.inc ] + %p.0 = phi i8* [ undef, %entry ], [ %incdec.ptr3, %for.inc ] + %q.0 = phi i8* [ undef, %entry ], [ %q.2, %for.inc ] + %0 = load i8* %p.0, align 1 + switch i8 %0, label %while.cond.preheader [ + i8 0, label %for.cond4.preheader + i8 92, label %for.inc + ] + +while.cond.preheader: ; preds = %for.cond + %tobool210 = icmp eq i32 %backslashes.0, 0 + br i1 %tobool210, label %for.inc.loopexit, label %while.body.lr.ph + +while.body.lr.ph: ; preds = %while.cond.preheader + %1 = add i32 %backslashes.0, -1 + %2 = zext i32 %1 to i64 + br label %while.body + +for.cond4.preheader: ; preds = %for.cond + %tobool57 = icmp eq i32 %backslashes.0, 0 + br i1 %tobool57, label %for.end10, label %for.body6.lr.ph + +for.body6.lr.ph: ; preds = %for.cond4.preheader + br label %for.body6 + +while.body: ; preds = %while.body.lr.ph, %while.body + %q.112 = phi i8* [ %q.0, %while.body.lr.ph ], [ %incdec.ptr, %while.body ] + %backslashes.111 = phi i32 [ %backslashes.0, %while.body.lr.ph ], [ %dec, %while.body ] + %incdec.ptr = getelementptr inbounds i8* %q.112, i64 1 + store i8 92, i8* %incdec.ptr, align 1 + %dec = add nsw i32 %backslashes.111, -1 + %tobool2 = icmp eq i32 %dec, 0 + br i1 %tobool2, label %while.cond.for.inc.loopexit_crit_edge, label %while.body + +while.cond.for.inc.loopexit_crit_edge: ; preds = %while.body + %scevgep.sum = add i64 %2, 1 + %scevgep13 = getelementptr i8* %q.0, i64 %scevgep.sum + br label %for.inc.loopexit + +for.inc.loopexit: ; preds = %while.cond.for.inc.loopexit_crit_edge, %while.cond.preheader + %q.1.lcssa = phi i8* [ %scevgep13, %while.cond.for.inc.loopexit_crit_edge ], [ %q.0, %while.cond.preheader ] + br label %for.inc + +for.inc: ; preds = %for.inc.loopexit, %for.cond + %backslashes.2 = phi i32 [ %backslashes.0, %for.cond ], [ 0, %for.inc.loopexit ] + %q.2 = phi i8* [ %q.0, %for.cond ], [ %q.1.lcssa, %for.inc.loopexit ] + %incdec.ptr3 = getelementptr inbounds i8* %p.0, i64 1 + br label %for.cond + +for.body6: ; preds = %for.body6.lr.ph, %for.body6 + %q.39 = phi i8* [ %q.0, %for.body6.lr.ph ], [ %incdec.ptr7, %for.body6 ] + %backslashes.38 = phi i32 [ %backslashes.0, %for.body6.lr.ph ], [ %dec9, %for.body6 ] + %incdec.ptr7 = getelementptr inbounds i8* %q.39, i64 1 + store i8 92, i8* %incdec.ptr7, align 1 + %dec9 = add nsw i32 %backslashes.38, -1 + %tobool5 = icmp eq i32 %dec9, 0 + br i1 %tobool5, label %for.cond4.for.end10_crit_edge, label %for.body6 + +for.cond4.for.end10_crit_edge: ; preds = %for.body6 + br label %for.end10 + +for.end10: ; preds = %for.cond4.for.end10_crit_edge, %for.cond4.preheader + ret i32 undef +} diff --git a/test/Transforms/LoopVectorize/2012-10-22-isconsec.ll b/test/Transforms/LoopVectorize/2012-10-22-isconsec.ll index 12e8e55983..2516e248bc 100644 --- a/test/Transforms/LoopVectorize/2012-10-22-isconsec.ll +++ b/test/Transforms/LoopVectorize/2012-10-22-isconsec.ll @@ -1,4 +1,4 @@ -; RUN: opt < %s -loop-vectorize -dce +; RUN: opt < %s -loop-vectorize -dce -force-vector-width=4 ; Check that we don't crash. diff --git a/test/Transforms/LoopVectorize/X86/avx1.ll b/test/Transforms/LoopVectorize/X86/avx1.ll new file mode 100644 index 0000000000..a2d176a534 --- /dev/null +++ b/test/Transforms/LoopVectorize/X86/avx1.ll @@ -0,0 +1,49 @@ +; RUN: opt < %s -loop-vectorize -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx -S | FileCheck %s + +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +;CHECK: @read_mod_write_single_ptr +;CHECK: load <8 x float> +;CHECK: ret i32 +define i32 @read_mod_write_single_ptr(float* nocapture %a, i32 %n) nounwind uwtable ssp { + %1 = icmp sgt i32 %n, 0 + br i1 %1, label %.lr.ph, label %._crit_edge + +.lr.ph: ; preds = %0, %.lr.ph + %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 0, %0 ] + %2 = getelementptr inbounds float* %a, i64 %indvars.iv + %3 = load float* %2, align 4 + %4 = fmul float %3, 3.000000e+00 + store float %4, float* %2, align 4 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, %n + br i1 %exitcond, label %._crit_edge, label %.lr.ph + +._crit_edge: ; preds = %.lr.ph, %0 + ret i32 undef +} + + +;CHECK: @read_mod_i64 +;CHECK: load <8 x i64> +;CHECK: ret i32 +define i32 @read_mod_i64(i64* nocapture %a, i32 %n) nounwind uwtable ssp { + %1 = icmp sgt i32 %n, 0 + br i1 %1, label %.lr.ph, label %._crit_edge + +.lr.ph: ; preds = %0, %.lr.ph + %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 0, %0 ] + %2 = getelementptr inbounds i64* %a, i64 %indvars.iv + %3 = load i64* %2, align 4 + %4 = mul i64 %3, 3 + store i64 %4, i64* %2, align 4 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, %n + br i1 %exitcond, label %._crit_edge, label %.lr.ph + +._crit_edge: ; preds = %.lr.ph, %0 + ret i32 undef +} diff --git a/test/Transforms/LoopVectorize/X86/conversion-cost.ll b/test/Transforms/LoopVectorize/X86/conversion-cost.ll new file mode 100644 index 0000000000..8582613617 --- /dev/null +++ b/test/Transforms/LoopVectorize/X86/conversion-cost.ll @@ -0,0 +1,48 @@ +; RUN: opt < %s -loop-vectorize -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx -S | FileCheck %s + +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +;CHECK: @conversion_cost1 +;CHECK: store <8 x i8> +;CHECK: ret +define i32 @conversion_cost1(i32 %n, i8* nocapture %A, float* nocapture %B) nounwind uwtable ssp { + %1 = icmp sgt i32 %n, 3 + br i1 %1, label %.lr.ph, label %._crit_edge + +.lr.ph: ; preds = %0, %.lr.ph + %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 3, %0 ] + %2 = trunc i64 %indvars.iv to i8 + %3 = getelementptr inbounds i8* %A, i64 %indvars.iv + store i8 %2, i8* %3, align 1 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, %n + br i1 %exitcond, label %._crit_edge, label %.lr.ph + +._crit_edge: ; preds = %.lr.ph, %0 + ret i32 undef +} + +;CHECK: @conversion_cost2 +;CHECK: store <8 x float> +;CHECK: ret +define i32 @conversion_cost2(i32 %n, i8* nocapture %A, float* nocapture %B) nounwind uwtable ssp { + %1 = icmp sgt i32 %n, 9 + br i1 %1, label %.lr.ph, label %._crit_edge + +.lr.ph: ; preds = %0, %.lr.ph + %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 9, %0 ] + %2 = add nsw i64 %indvars.iv, 3 + %3 = trunc i64 %2 to i32 + %4 = sitofp i32 %3 to float + %5 = getelementptr inbounds float* %B, i64 %indvars.iv + store float %4, float* %5, align 4 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, %n + br i1 %exitcond, label %._crit_edge, label %.lr.ph + +._crit_edge: ; preds = %.lr.ph, %0 + ret i32 undef +} diff --git a/test/Transforms/LoopVectorize/cost-model.ll b/test/Transforms/LoopVectorize/X86/cost-model.ll index 18abf2885e..40e660855b 100644 --- a/test/Transforms/LoopVectorize/cost-model.ll +++ b/test/Transforms/LoopVectorize/X86/cost-model.ll @@ -1,4 +1,4 @@ -; RUN: opt < %s -loop-vectorize -mtriple=x86_64-apple-macosx10.8.0 -S | FileCheck %s +; RUN: opt < %s -loop-vectorize -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx -S | FileCheck %s target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" target triple = "x86_64-apple-macosx10.8.0" diff --git a/test/Transforms/LoopVectorize/X86/gcc-examples.ll b/test/Transforms/LoopVectorize/X86/gcc-examples.ll new file mode 100644 index 0000000000..574c529834 --- /dev/null +++ b/test/Transforms/LoopVectorize/X86/gcc-examples.ll @@ -0,0 +1,62 @@ +; RUN: opt < %s -loop-vectorize -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7 -dce -instcombine -licm -S | FileCheck %s + +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +@b = common global [2048 x i32] zeroinitializer, align 16 +@c = common global [2048 x i32] zeroinitializer, align 16 +@a = common global [2048 x i32] zeroinitializer, align 16 + +; Select VF = 8; +;CHECK: @example1 +;CHECK: load <8 x i32> +;CHECK: add nsw <8 x i32> +;CHECK: store <8 x i32> +;CHECK: ret void +define void @example1() nounwind uwtable ssp { + br label %1 + +; <label>:1 ; preds = %1, %0 + %indvars.iv = phi i64 [ 0, %0 ], [ %indvars.iv.next, %1 ] + %2 = getelementptr inbounds [2048 x i32]* @b, i64 0, i64 %indvars.iv + %3 = load i32* %2, align 4 + %4 = getelementptr inbounds [2048 x i32]* @c, i64 0, i64 %indvars.iv + %5 = load i32* %4, align 4 + %6 = add nsw i32 %5, %3 + %7 = getelementptr inbounds [2048 x i32]* @a, i64 0, i64 %indvars.iv + store i32 %6, i32* %7, align 4 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, 256 + br i1 %exitcond, label %8, label %1 + +; <label>:8 ; preds = %1 + ret void +} + + +; Select VF=4 because sext <8 x i1> to <8 x i32> is expensive. +;CHECK: @example10b +;CHECK: load <4 x i16> +;CHECK: sext <4 x i16> +;CHECK: store <4 x i32> +;CHECK: ret void +define void @example10b(i16* noalias nocapture %sa, i16* noalias nocapture %sb, i16* noalias nocapture %sc, i32* noalias nocapture %ia, i32* noalias nocapture %ib, i32* noalias nocapture %ic) nounwind uwtable ssp { + br label %1 + +; <label>:1 ; preds = %1, %0 + %indvars.iv = phi i64 [ 0, %0 ], [ %indvars.iv.next, %1 ] + %2 = getelementptr inbounds i16* %sb, i64 %indvars.iv + %3 = load i16* %2, align 2 + %4 = sext i16 %3 to i32 + %5 = getelementptr inbounds i32* %ia, i64 %indvars.iv + store i32 %4, i32* %5, align 4 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, 1024 + br i1 %exitcond, label %6, label %1 + +; <label>:6 ; preds = %1 + ret void +} + diff --git a/test/Transforms/LoopVectorize/X86/lit.local.cfg b/test/Transforms/LoopVectorize/X86/lit.local.cfg new file mode 100644 index 0000000000..a8ad0f1a28 --- /dev/null +++ b/test/Transforms/LoopVectorize/X86/lit.local.cfg @@ -0,0 +1,6 @@ +config.suffixes = ['.ll', '.c', '.cpp'] + +targets = set(config.root.targets_to_build.split()) +if not 'X86' in targets: + config.unsupported = True + diff --git a/test/Transforms/LoopVectorize/flags.ll b/test/Transforms/LoopVectorize/flags.ll new file mode 100644 index 0000000000..2f22a76457 --- /dev/null +++ b/test/Transforms/LoopVectorize/flags.ll @@ -0,0 +1,53 @@ +; RUN: opt < %s -loop-vectorize -force-vector-width=4 -dce -instcombine -licm -S | FileCheck %s + +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +;CHECK: @flags1 +;CHECK: load <4 x i32> +;CHECK: mul nsw <4 x i32> +;CHECK: store <4 x i32> +;CHECK: ret i32 +define i32 @flags1(i32 %n, i32* nocapture %A) nounwind uwtable ssp { + %1 = icmp sgt i32 %n, 9 + br i1 %1, label %.lr.ph, label %._crit_edge + +.lr.ph: ; preds = %0, %.lr.ph + %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 9, %0 ] + %2 = getelementptr inbounds i32* %A, i64 %indvars.iv + %3 = load i32* %2, align 4 + %4 = mul nsw i32 %3, 3 + store i32 %4, i32* %2, align 4 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, %n + br i1 %exitcond, label %._crit_edge, label %.lr.ph + +._crit_edge: ; preds = %.lr.ph, %0 + ret i32 undef +} + + +;CHECK: @flags2 +;CHECK: load <4 x i32> +;CHECK: mul <4 x i32> +;CHECK: store <4 x i32> +;CHECK: ret i32 +define i32 @flags2(i32 %n, i32* nocapture %A) nounwind uwtable ssp { + %1 = icmp sgt i32 %n, 9 + br i1 %1, label %.lr.ph, label %._crit_edge + +.lr.ph: ; preds = %0, %.lr.ph + %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 9, %0 ] + %2 = getelementptr inbounds i32* %A, i64 %indvars.iv + %3 = load i32* %2, align 4 + %4 = mul i32 %3, 3 + store i32 %4, i32* %2, align 4 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, %n + br i1 %exitcond, label %._crit_edge, label %.lr.ph + +._crit_edge: ; preds = %.lr.ph, %0 + ret i32 undef +} diff --git a/test/Transforms/LoopVectorize/gcc-examples.ll b/test/Transforms/LoopVectorize/gcc-examples.ll index d8942ac861..fce29d2404 100644 --- a/test/Transforms/LoopVectorize/gcc-examples.ll +++ b/test/Transforms/LoopVectorize/gcc-examples.ll @@ -21,7 +21,7 @@ target triple = "x86_64-apple-macosx10.8.0" ;CHECK: @example1 ;CHECK: load <4 x i32> -;CHECK: add <4 x i32> +;CHECK: add nsw <4 x i32> ;CHECK: store <4 x i32> ;CHECK: ret void define void @example1() nounwind uwtable ssp { @@ -227,6 +227,8 @@ define i32 @example9() nounwind uwtable readonly ssp { } ;CHECK: @example10a +;CHECK: load <4 x i32> +;CHECK: add nsw <4 x i32> ;CHECK: load <4 x i16> ;CHECK: add <4 x i16> ;CHECK: store <4 x i16> diff --git a/test/Transforms/LoopVectorize/increment.ll b/test/Transforms/LoopVectorize/increment.ll index 069b7ea031..71ea7689fc 100644 --- a/test/Transforms/LoopVectorize/increment.ll +++ b/test/Transforms/LoopVectorize/increment.ll @@ -11,7 +11,7 @@ target triple = "x86_64-apple-macosx10.8.0" ; } ;CHECK: @inc ;CHECK: load <4 x i32> -;CHECK: add <4 x i32> +;CHECK: add nsw <4 x i32> ;CHECK: store <4 x i32> ;CHECK: ret void define void @inc(i32 %n) nounwind uwtable noinline ssp { diff --git a/test/Transforms/LoopVectorize/non-const-n.ll b/test/Transforms/LoopVectorize/non-const-n.ll index 7727b0a2dc..1a6c15ed96 100644 --- a/test/Transforms/LoopVectorize/non-const-n.ll +++ b/test/Transforms/LoopVectorize/non-const-n.ll @@ -11,7 +11,7 @@ target triple = "x86_64-apple-macosx10.8.0" ;CHECK: shl i32 ;CHECK: zext i32 ;CHECK: load <4 x i32> -;CHECK: add <4 x i32> +;CHECK: add nsw <4 x i32> ;CHECK: store <4 x i32> ;CHECK: ret void define void @example1(i32 %n) nounwind uwtable ssp { diff --git a/test/Transforms/LoopVectorize/reduction.ll b/test/Transforms/LoopVectorize/reduction.ll index 746a08c3ea..c1848b35fc 100644 --- a/test/Transforms/LoopVectorize/reduction.ll +++ b/test/Transforms/LoopVectorize/reduction.ll @@ -66,7 +66,7 @@ define i32 @reduction_prod(i32 %n, i32* noalias nocapture %A, i32* noalias nocap ;CHECK: @reduction_mix ;CHECK: phi <4 x i32> ;CHECK: load <4 x i32> -;CHECK: mul <4 x i32> +;CHECK: mul nsw <4 x i32> ;CHECK: ret i32 define i32 @reduction_mix(i32 %n, i32* noalias nocapture %A, i32* noalias nocapture %B) nounwind uwtable readonly noinline ssp { %1 = icmp sgt i32 %n, 0 @@ -151,6 +151,7 @@ for.end: ; preds = %for.body, %entry ;CHECK: @reduction_and ;CHECK: and <4 x i32> +;CHECK: <i32 -1, i32 -1, i32 -1, i32 -1> ;CHECK: ret i32 define i32 @reduction_and(i32 %n, i32* nocapture %A, i32* nocapture %B) nounwind uwtable readonly { entry: diff --git a/test/Transforms/LoopVectorize/small-loop.ll b/test/Transforms/LoopVectorize/small-loop.ll new file mode 100644 index 0000000000..4a6e4b231d --- /dev/null +++ b/test/Transforms/LoopVectorize/small-loop.ll @@ -0,0 +1,33 @@ +; RUN: opt < %s -loop-vectorize -force-vector-width=4 -dce -instcombine -licm -S | FileCheck %s + +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +@a = common global [2048 x i32] zeroinitializer, align 16 +@b = common global [2048 x i32] zeroinitializer, align 16 +@c = common global [2048 x i32] zeroinitializer, align 16 + +;CHECK: @example1 +;CHECK-NOT: load <4 x i32> +;CHECK: ret void +define void @example1() nounwind uwtable ssp { + br label %1 + +; <label>:1 ; preds = %1, %0 + %indvars.iv = phi i64 [ 0, %0 ], [ %indvars.iv.next, %1 ] + %2 = getelementptr inbounds [2048 x i32]* @b, i64 0, i64 %indvars.iv + %3 = load i32* %2, align 4 + %4 = getelementptr inbounds [2048 x i32]* @c, i64 0, i64 %indvars.iv + %5 = load i32* %4, align 4 + %6 = add nsw i32 %5, %3 + %7 = getelementptr inbounds [2048 x i32]* @a, i64 0, i64 %indvars.iv + store i32 %6, i32* %7, align 4 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, 8 ; <----- A really small trip count. + br i1 %exitcond, label %8, label %1 + +; <label>:8 ; preds = %1 + ret void +} + diff --git a/test/Transforms/LoopVectorize/start-non-zero.ll b/test/Transforms/LoopVectorize/start-non-zero.ll new file mode 100644 index 0000000000..5aa3bc034d --- /dev/null +++ b/test/Transforms/LoopVectorize/start-non-zero.ll @@ -0,0 +1,35 @@ +; RUN: opt < %s -loop-vectorize -force-vector-width=4 -instcombine -S | FileCheck %s + +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +;CHECK: @start_at_nonzero +;CHECK: mul nuw <4 x i32> +;CHECK: ret i32 +define i32 @start_at_nonzero(i32* nocapture %a, i32 %start, i32 %end) nounwind uwtable ssp { +entry: + %cmp3 = icmp slt i32 %start, %end + br i1 %cmp3, label %for.body.lr.ph, label %for.end + +for.body.lr.ph: ; preds = %entry + %0 = sext i32 %start to i64 + br label %for.body + +for.body: ; preds = %for.body.lr.ph, %for.body + %indvars.iv = phi i64 [ %0, %for.body.lr.ph ], [ %indvars.iv.next, %for.body ] + %arrayidx = getelementptr inbounds i32* %a, i64 %indvars.iv + %1 = load i32* %arrayidx, align 4, !tbaa !0 + %mul = mul nuw i32 %1, 333 + store i32 %mul, i32* %arrayidx, align 4, !tbaa !0 + %indvars.iv.next = add i64 %indvars.iv, 1 + %2 = trunc i64 %indvars.iv.next to i32 + %cmp = icmp slt i32 %2, %end + br i1 %cmp, label %for.body, label %for.end + +for.end: ; preds = %for.body, %entry + ret i32 4 +} + +!0 = metadata !{metadata !"int", metadata !1} +!1 = metadata !{metadata !"omnipotent char", metadata !2} +!2 = metadata !{metadata !"Simple C/C++ TBAA"} diff --git a/test/Transforms/LoopVectorize/write-only.ll b/test/Transforms/LoopVectorize/write-only.ll new file mode 100644 index 0000000000..eb02760413 --- /dev/null +++ b/test/Transforms/LoopVectorize/write-only.ll @@ -0,0 +1,26 @@ +; RUN: opt < %s -loop-vectorize -force-vector-width=4 -dce -instcombine -licm -S | FileCheck %s + +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" +target triple = "x86_64-apple-macosx10.8.0" + +;CHECK: @read_mod_write_single_ptr +;CHECK: load <4 x float> +;CHECK: ret i32 +define i32 @read_mod_write_single_ptr(float* nocapture %a, i32 %n) nounwind uwtable ssp { + %1 = icmp sgt i32 %n, 0 + br i1 %1, label %.lr.ph, label %._crit_edge + +.lr.ph: ; preds = %0, %.lr.ph + %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 0, %0 ] + %2 = getelementptr inbounds float* %a, i64 %indvars.iv + %3 = load float* %2, align 4 + %4 = fmul float %3, 3.000000e+00 + store float %4, float* %2, align 4 + %indvars.iv.next = add i64 %indvars.iv, 1 + %lftr.wideiv = trunc i64 %indvars.iv.next to i32 + %exitcond = icmp eq i32 %lftr.wideiv, %n + br i1 %exitcond, label %._crit_edge, label %.lr.ph + +._crit_edge: ; preds = %.lr.ph, %0 + ret i32 undef +} diff --git a/test/Transforms/SROA/vector-promotion.ll b/test/Transforms/SROA/vector-promotion.ll index 92051c62a7..02e084bf11 100644 --- a/test/Transforms/SROA/vector-promotion.ll +++ b/test/Transforms/SROA/vector-promotion.ll @@ -205,3 +205,18 @@ define i64 @test6(<4 x i64> %x, <4 x i64> %y, i64 %n) { %res = load i64* %addr, align 4 ret i64 %res } + +define i32 @PR14212() { +; CHECK: @PR14212 +; This caused a crash when "splitting" the load of the i32 in order to promote +; the store of <3 x i8> properly. Heavily reduced from an OpenCL test case. +entry: + %retval = alloca <3 x i8>, align 4 +; CHECK-NOT: alloca + + store <3 x i8> undef, <3 x i8>* %retval, align 4 + %cast = bitcast <3 x i8>* %retval to i32* + %load = load i32* %cast, align 4 + ret i32 %load +; CHECK: ret i32 +} diff --git a/test/Transforms/SimplifyCFG/SPARC/lit.local.cfg b/test/Transforms/SimplifyCFG/SPARC/lit.local.cfg new file mode 100644 index 0000000000..786fee9e66 --- /dev/null +++ b/test/Transforms/SimplifyCFG/SPARC/lit.local.cfg @@ -0,0 +1,6 @@ +config.suffixes = ['.ll', '.c', '.cpp'] + +targets = set(config.root.targets_to_build.split()) +if not 'Sparc' in targets: + config.unsupported = True + diff --git a/test/Transforms/SimplifyCFG/SPARC/switch_to_lookup_table.ll b/test/Transforms/SimplifyCFG/SPARC/switch_to_lookup_table.ll new file mode 100644 index 0000000000..9d1568557f --- /dev/null +++ b/test/Transforms/SimplifyCFG/SPARC/switch_to_lookup_table.ll @@ -0,0 +1,32 @@ +; RUN: opt < %s -simplifycfg -S -mtriple=sparc-unknown-unknown | FileCheck %s + +; Check that switches are not turned into lookup tables, as this is not +; considered profitable on the target. + +define i32 @f(i32 %c) nounwind uwtable readnone { +entry: + switch i32 %c, label %sw.default [ + i32 42, label %return + i32 43, label %sw.bb1 + i32 44, label %sw.bb2 + i32 45, label %sw.bb3 + i32 46, label %sw.bb4 + i32 47, label %sw.bb5 + i32 48, label %sw.bb6 + ] + +sw.bb1: br label %return +sw.bb2: br label %return +sw.bb3: br label %return +sw.bb4: br label %return +sw.bb5: br label %return +sw.bb6: br label %return +sw.default: br label %return +return: + %retval.0 = phi i32 [ 15, %sw.default ], [ 1, %sw.bb6 ], [ 62, %sw.bb5 ], [ 27, %sw.bb4 ], [ -1, %sw.bb3 ], [ 0, %sw.bb2 ], [ 123, %sw.bb1 ], [ 55, %entry ] + ret i32 %retval.0 + +; CHECK: @f +; CHECK-NOT: getelementptr +; CHECK: switch i32 %c +} diff --git a/test/Transforms/SimplifyCFG/switch_to_lookup_table.ll b/test/Transforms/SimplifyCFG/switch_to_lookup_table.ll index aa48ec6481..d358350554 100644 --- a/test/Transforms/SimplifyCFG/switch_to_lookup_table.ll +++ b/test/Transforms/SimplifyCFG/switch_to_lookup_table.ll @@ -15,12 +15,21 @@ target triple = "x86_64-unknown-linux-gnu" ; The table for @earlyreturncrash ; CHECK: @switch.table3 = private unnamed_addr constant [4 x i32] [i32 42, i32 9, i32 88, i32 5] +; The table for @large. +; CHECK: @switch.table4 = private unnamed_addr constant [199 x i32] [i32 1, i32 4, i32 9, + +; The table for @cprop +; CHECK: @switch.table5 = private unnamed_addr constant [7 x i32] [i32 5, i32 42, i32 126, i32 -452, i32 128, i32 6, i32 7] + +; The table for @unreachable +; CHECK: @switch.table6 = private unnamed_addr constant [5 x i32] [i32 0, i32 0, i32 0, i32 1, i32 -1] + ; A simple int-to-int selection switch. ; It is dense enough to be replaced by table lookup. ; The result is directly by a ret from an otherwise empty bb, ; so we return early, directly from the lookup bb. -define i32 @f(i32 %c) nounwind uwtable readnone { +define i32 @f(i32 %c) { entry: switch i32 %c, label %sw.default [ i32 42, label %return @@ -233,7 +242,7 @@ lor.end: } ; PR13946 -define i32 @overflow(i32 %type) nounwind { +define i32 @overflow(i32 %type) { entry: switch i32 %type, label %sw.default [ i32 -2147483648, label %sw.bb @@ -244,23 +253,12 @@ entry: i32 3, label %sw.bb3 ] -sw.bb: - br label %if.end - -sw.bb1: - br label %if.end - -sw.bb2: - br label %if.end - -sw.bb3: - br label %if.end - -sw.default: - br label %if.end - -if.else: - br label %if.end +sw.bb: br label %if.end +sw.bb1: br label %if.end +sw.bb2: br label %if.end +sw.bb3: br label %if.end +sw.default: br label %if.end +if.else: br label %if.end if.end: %dirent_type.0 = phi i32 [ 3, %sw.default ], [ 6, %sw.bb3 ], [ 5, %sw.bb2 ], [ 0, %sw.bb1 ], [ 3, %sw.bb ], [ 0, %if.else ] @@ -271,7 +269,7 @@ if.end: } ; PR13985 -define i1 @undef(i32 %tmp) uwtable ssp { +define i1 @undef(i32 %tmp) { bb: switch i32 %tmp, label %bb3 [ i32 0, label %bb1 @@ -280,16 +278,502 @@ bb: i32 8, label %bb2 ] -bb1: ; preds = %bb, %bb - br label %bb3 - -bb2: ; preds = %bb, %bb - br label %bb3 +bb1: br label %bb3 +bb2: br label %bb3 -bb3: ; preds = %bb2, %bb1, %bb +bb3: %tmp4 = phi i1 [ undef, %bb ], [ false, %bb2 ], [ true, %bb1 ] ret i1 %tmp4 ; CHECK: define i1 @undef ; CHECK: %switch.cast = trunc i32 %switch.tableidx to i9 ; CHECK: %switch.downshift = lshr i9 3, %switch.shiftamt } + +; Also handle large switches that would be rejected by +; isValueEqualityComparison() +; CHECK: large +; CHECK-NOT: switch i32 +define i32 @large(i32 %x) { +entry: + %cmp = icmp slt i32 %x, 0 + br i1 %cmp, label %if.then, label %if.end + +if.then: + %mul = mul i32 %x, -10 + br label %if.end + +if.end: + %x.addr.0 = phi i32 [ %mul, %if.then ], [ %x, %entry ] + switch i32 %x.addr.0, label %return [ + i32 199, label %sw.bb203 + i32 1, label %sw.bb1 + i32 2, label %sw.bb2 + i32 3, label %sw.bb3 + i32 4, label %sw.bb4 + i32 5, label %sw.bb5 + i32 6, label %sw.bb6 + i32 7, label %sw.bb7 + i32 8, label %sw.bb8 + i32 9, label %sw.bb9 + i32 10, label %sw.bb10 + i32 11, label %sw.bb11 + i32 12, label %sw.bb12 + i32 13, label %sw.bb13 + i32 14, label %sw.bb14 + i32 15, label %sw.bb15 + i32 16, label %sw.bb16 + i32 17, label %sw.bb17 + i32 18, label %sw.bb18 + i32 19, label %sw.bb19 + i32 20, label %sw.bb20 + i32 21, label %sw.bb21 + i32 22, label %sw.bb22 + i32 23, label %sw.bb23 + i32 24, label %sw.bb24 + i32 25, label %sw.bb25 + i32 26, label %sw.bb26 + i32 27, label %sw.bb27 + i32 28, label %sw.bb28 + i32 29, label %sw.bb29 + i32 30, label %sw.bb30 + i32 31, label %sw.bb31 + i32 32, label %sw.bb32 + i32 33, label %sw.bb33 + i32 34, label %sw.bb34 + i32 35, label %sw.bb35 + i32 36, label %sw.bb37 + i32 37, label %sw.bb38 + i32 38, label %sw.bb39 + i32 39, label %sw.bb40 + i32 40, label %sw.bb41 + i32 41, label %sw.bb42 + i32 42, label %sw.bb43 + i32 43, label %sw.bb44 + i32 44, label %sw.bb45 + i32 45, label %sw.bb47 + i32 46, label %sw.bb48 + i32 47, label %sw.bb49 + i32 48, label %sw.bb50 + i32 49, label %sw.bb51 + i32 50, label %sw.bb52 + i32 51, label %sw.bb53 + i32 52, label %sw.bb54 + i32 53, label %sw.bb55 + i32 54, label %sw.bb56 + i32 55, label %sw.bb58 + i32 56, label %sw.bb59 + i32 57, label %sw.bb60 + i32 58, label %sw.bb61 + i32 59, label %sw.bb62 + i32 60, label %sw.bb63 + i32 61, label %sw.bb64 + i32 62, label %sw.bb65 + i32 63, label %sw.bb66 + i32 64, label %sw.bb67 + i32 65, label %sw.bb68 + i32 66, label %sw.bb69 + i32 67, label %sw.bb70 + i32 68, label %sw.bb71 + i32 69, label %sw.bb72 + i32 70, label %sw.bb73 + i32 71, label %sw.bb74 + i32 72, label %sw.bb76 + i32 73, label %sw.bb77 + i32 74, label %sw.bb78 + i32 75, label %sw.bb79 + i32 76, label %sw.bb80 + i32 77, label %sw.bb81 + i32 78, label %sw.bb82 + i32 79, label %sw.bb83 + i32 80, label %sw.bb84 + i32 81, label %sw.bb85 + i32 82, label %sw.bb86 + i32 83, label %sw.bb87 + i32 84, label %sw.bb88 + i32 85, label %sw.bb89 + i32 86, label %sw.bb90 + i32 87, label %sw.bb91 + i32 88, label %sw.bb92 + i32 89, label %sw.bb93 + i32 90, label %sw.bb94 + i32 91, label %sw.bb95 + i32 92, label %sw.bb96 + i32 93, label %sw.bb97 + i32 94, label %sw.bb98 + i32 95, label %sw.bb99 + i32 96, label %sw.bb100 + i32 97, label %sw.bb101 + i32 98, label %sw.bb102 + i32 99, label %sw.bb103 + i32 100, label %sw.bb104 + i32 101, label %sw.bb105 + i32 102, label %sw.bb106 + i32 103, label %sw.bb107 + i32 104, label %sw.bb108 + i32 105, label %sw.bb109 + i32 106, label %sw.bb110 + i32 107, label %sw.bb111 + i32 108, label %sw.bb112 + i32 109, label %sw.bb113 + i32 110, label %sw.bb114 + i32 111, label %sw.bb115 + i32 112, label %sw.bb116 + i32 113, label %sw.bb117 + i32 114, label %sw.bb118 + i32 115, label %sw.bb119 + i32 116, label %sw.bb120 + i32 117, label %sw.bb121 + i32 118, label %sw.bb122 + i32 119, label %sw.bb123 + i32 120, label %sw.bb124 + i32 121, label %sw.bb125 + i32 122, label %sw.bb126 + i32 123, label %sw.bb127 + i32 124, label %sw.bb128 + i32 125, label %sw.bb129 + i32 126, label %sw.bb130 + i32 127, label %sw.bb131 + i32 128, label %sw.bb132 + i32 129, label %sw.bb133 + i32 130, label %sw.bb134 + i32 131, label %sw.bb135 + i32 132, label %sw.bb136 + i32 133, label %sw.bb137 + i32 134, label %sw.bb138 + i32 135, label %sw.bb139 + i32 136, label %sw.bb140 + i32 137, label %sw.bb141 + i32 138, label %sw.bb142 + i32 139, label %sw.bb143 + i32 140, label %sw.bb144 + i32 141, label %sw.bb145 + i32 142, label %sw.bb146 + i32 143, label %sw.bb147 + i32 144, label %sw.bb148 + i32 145, label %sw.bb149 + i32 146, label %sw.bb150 + i32 147, label %sw.bb151 + i32 148, label %sw.bb152 + i32 149, label %sw.bb153 + i32 150, label %sw.bb154 + i32 151, label %sw.bb155 + i32 152, label %sw.bb156 + i32 153, label %sw.bb157 + i32 154, label %sw.bb158 + i32 155, label %sw.bb159 + i32 156, label %sw.bb160 + i32 157, label %sw.bb161 + i32 158, label %sw.bb162 + i32 159, label %sw.bb163 + i32 160, label %sw.bb164 + i32 161, label %sw.bb165 + i32 162, label %sw.bb166 + i32 163, label %sw.bb167 + i32 164, label %sw.bb168 + i32 165, label %sw.bb169 + i32 166, label %sw.bb170 + i32 167, label %sw.bb171 + i32 168, label %sw.bb172 + i32 169, label %sw.bb173 + i32 170, label %sw.bb174 + i32 171, label %sw.bb175 + i32 172, label %sw.bb176 + i32 173, label %sw.bb177 + i32 174, label %sw.bb178 + i32 175, label %sw.bb179 + i32 176, label %sw.bb180 + i32 177, label %sw.bb181 + i32 178, label %sw.bb182 + i32 179, label %sw.bb183 + i32 180, label %sw.bb184 + i32 181, label %sw.bb185 + i32 182, label %sw.bb186 + i32 183, label %sw.bb187 + i32 184, label %sw.bb188 + i32 185, label %sw.bb189 + i32 186, label %sw.bb190 + i32 187, label %sw.bb191 + i32 188, label %sw.bb192 + i32 189, label %sw.bb193 + i32 190, label %sw.bb194 + i32 191, label %sw.bb195 + i32 192, label %sw.bb196 + i32 193, label %sw.bb197 + i32 194, label %sw.bb198 + i32 195, label %sw.bb199 + i32 196, label %sw.bb200 + i32 197, label %sw.bb201 + i32 198, label %sw.bb202 + ] + +sw.bb1: br label %return +sw.bb2: br label %return +sw.bb3: br label %return +sw.bb4: br label %return +sw.bb5: br label %return +sw.bb6: br label %return +sw.bb7: br label %return +sw.bb8: br label %return +sw.bb9: br label %return +sw.bb10: br label %return +sw.bb11: br label %return +sw.bb12: br label %return +sw.bb13: br label %return +sw.bb14: br label %return +sw.bb15: br label %return +sw.bb16: br label %return +sw.bb17: br label %return +sw.bb18: br label %return +sw.bb19: br label %return +sw.bb20: br label %return +sw.bb21: br label %return +sw.bb22: br label %return +sw.bb23: br label %return +sw.bb24: br label %return +sw.bb25: br label %return +sw.bb26: br label %return +sw.bb27: br label %return +sw.bb28: br label %return +sw.bb29: br label %return +sw.bb30: br label %return +sw.bb31: br label %return +sw.bb32: br label %return +sw.bb33: br label %return +sw.bb34: br label %return +sw.bb35: br label %return +sw.bb37: br label %return +sw.bb38: br label %return +sw.bb39: br label %return +sw.bb40: br label %return +sw.bb41: br label %return +sw.bb42: br label %return +sw.bb43: br label %return +sw.bb44: br label %return +sw.bb45: br label %return +sw.bb47: br label %return +sw.bb48: br label %return +sw.bb49: br label %return +sw.bb50: br label %return +sw.bb51: br label %return +sw.bb52: br label %return +sw.bb53: br label %return +sw.bb54: br label %return +sw.bb55: br label %return +sw.bb56: br label %return +sw.bb58: br label %return +sw.bb59: br label %return +sw.bb60: br label %return +sw.bb61: br label %return +sw.bb62: br label %return +sw.bb63: br label %return +sw.bb64: br label %return +sw.bb65: br label %return +sw.bb66: br label %return +sw.bb67: br label %return +sw.bb68: br label %return +sw.bb69: br label %return +sw.bb70: br label %return +sw.bb71: br label %return +sw.bb72: br label %return +sw.bb73: br label %return +sw.bb74: br label %return +sw.bb76: br label %return +sw.bb77: br label %return +sw.bb78: br label %return +sw.bb79: br label %return +sw.bb80: br label %return +sw.bb81: br label %return +sw.bb82: br label %return +sw.bb83: br label %return +sw.bb84: br label %return +sw.bb85: br label %return +sw.bb86: br label %return +sw.bb87: br label %return +sw.bb88: br label %return +sw.bb89: br label %return +sw.bb90: br label %return +sw.bb91: br label %return +sw.bb92: br label %return +sw.bb93: br label %return +sw.bb94: br label %return +sw.bb95: br label %return +sw.bb96: br label %return +sw.bb97: br label %return +sw.bb98: br label %return +sw.bb99: br label %return +sw.bb100: br label %return +sw.bb101: br label %return +sw.bb102: br label %return +sw.bb103: br label %return +sw.bb104: br label %return +sw.bb105: br label %return +sw.bb106: br label %return +sw.bb107: br label %return +sw.bb108: br label %return +sw.bb109: br label %return +sw.bb110: br label %return +sw.bb111: br label %return +sw.bb112: br label %return +sw.bb113: br label %return +sw.bb114: br label %return +sw.bb115: br label %return +sw.bb116: br label %return +sw.bb117: br label %return +sw.bb118: br label %return +sw.bb119: br label %return +sw.bb120: br label %return +sw.bb121: br label %return +sw.bb122: br label %return +sw.bb123: br label %return +sw.bb124: br label %return +sw.bb125: br label %return +sw.bb126: br label %return +sw.bb127: br label %return +sw.bb128: br label %return +sw.bb129: br label %return +sw.bb130: br label %return +sw.bb131: br label %return +sw.bb132: br label %return +sw.bb133: br label %return +sw.bb134: br label %return +sw.bb135: br label %return +sw.bb136: br label %return +sw.bb137: br label %return +sw.bb138: br label %return +sw.bb139: br label %return +sw.bb140: br label %return +sw.bb141: br label %return +sw.bb142: br label %return +sw.bb143: br label %return +sw.bb144: br label %return +sw.bb145: br label %return +sw.bb146: br label %return +sw.bb147: br label %return +sw.bb148: br label %return +sw.bb149: br label %return +sw.bb150: br label %return +sw.bb151: br label %return +sw.bb152: br label %return +sw.bb153: br label %return +sw.bb154: br label %return +sw.bb155: br label %return +sw.bb156: br label %return +sw.bb157: br label %return +sw.bb158: br label %return +sw.bb159: br label %return +sw.bb160: br label %return +sw.bb161: br label %return +sw.bb162: br label %return +sw.bb163: br label %return +sw.bb164: br label %return +sw.bb165: br label %return +sw.bb166: br label %return +sw.bb167: br label %return +sw.bb168: br label %return +sw.bb169: br label %return +sw.bb170: br label %return +sw.bb171: br label %return +sw.bb172: br label %return +sw.bb173: br label %return +sw.bb174: br label %return +sw.bb175: br label %return +sw.bb176: br label %return +sw.bb177: br label %return +sw.bb178: br label %return +sw.bb179: br label %return +sw.bb180: br label %return +sw.bb181: br label %return +sw.bb182: br label %return +sw.bb183: br label %return +sw.bb184: br label %return +sw.bb185: br label %return +sw.bb186: br label %return +sw.bb187: br label %return +sw.bb188: br label %return +sw.bb189: br label %return +sw.bb190: br label %return +sw.bb191: br label %return +sw.bb192: br label %return +sw.bb193: br label %return +sw.bb194: br label %return +sw.bb195: br label %return +sw.bb196: br label %return +sw.bb197: br label %return +sw.bb198: br label %return +sw.bb199: br label %return +sw.bb200: br label %return +sw.bb201: br label %return +sw.bb202: br label %return +sw.bb203: br label %return + +return: + %retval.0 = phi i32 [ 39204, %sw.bb202 ], [ 38809, %sw.bb201 ], [ 38416, %sw.bb200 ], [ 38025, %sw.bb199 ], [ 37636, %sw.bb198 ], [ 37249, %sw.bb197 ], [ 36864, %sw.bb196 ], [ 36481, %sw.bb195 ], [ 36100, %sw.bb194 ], [ 35721, %sw.bb193 ], [ 35344, %sw.bb192 ], [ 34969, %sw.bb191 ], [ 34596, %sw.bb190 ], [ 34225, %sw.bb189 ], [ 33856, %sw.bb188 ], [ 33489, %sw.bb187 ], [ 33124, %sw.bb186 ], [ 32761, %sw.bb185 ], [ 32400, %sw.bb184 ], [ 32041, %sw.bb183 ], [ 31684, %sw.bb182 ], [ 31329, %sw.bb181 ], [ 30976, %sw.bb180 ], [ 30625, %sw.bb179 ], [ 30276, %sw.bb178 ], [ 29929, %sw.bb177 ], [ 29584, %sw.bb176 ], [ 29241, %sw.bb175 ], [ 28900, %sw.bb174 ], [ 28561, %sw.bb173 ], [ 28224, %sw.bb172 ], [ 27889, %sw.bb171 ], [ 27556, %sw.bb170 ], [ 27225, %sw.bb169 ], [ 26896, %sw.bb168 ], [ 26569, %sw.bb167 ], [ 26244, %sw.bb166 ], [ 25921, %sw.bb165 ], [ 25600, %sw.bb164 ], [ 25281, %sw.bb163 ], [ 24964, %sw.bb162 ], [ 24649, %sw.bb161 ], [ 24336, %sw.bb160 ], [ 24025, %sw.bb159 ], [ 23716, %sw.bb158 ], [ 23409, %sw.bb157 ], [ 23104, %sw.bb156 ], [ 22801, %sw.bb155 ], [ 22500, %sw.bb154 ], [ 22201, %sw.bb153 ], [ 21904, %sw.bb152 ], [ 21609, %sw.bb151 ], [ 21316, %sw.bb150 ], [ 21025, %sw.bb149 ], [ 20736, %sw.bb148 ], [ 20449, %sw.bb147 ], [ 20164, %sw.bb146 ], [ 19881, %sw.bb145 ], [ 19600, %sw.bb144 ], [ 19321, %sw.bb143 ], [ 19044, %sw.bb142 ], [ 18769, %sw.bb141 ], [ 18496, %sw.bb140 ], [ 18225, %sw.bb139 ], [ 17956, %sw.bb138 ], [ 17689, %sw.bb137 ], [ 17424, %sw.bb136 ], [ 17161, %sw.bb135 ], [ 16900, %sw.bb134 ], [ 16641, %sw.bb133 ], [ 16384, %sw.bb132 ], [ 16129, %sw.bb131 ], [ 15876, %sw.bb130 ], [ 15625, %sw.bb129 ], [ 15376, %sw.bb128 ], [ 15129, %sw.bb127 ], [ 14884, %sw.bb126 ], [ 14641, %sw.bb125 ], [ 14400, %sw.bb124 ], [ 14161, %sw.bb123 ], [ 13924, %sw.bb122 ], [ 13689, %sw.bb121 ], [ 13456, %sw.bb120 ], [ 13225, %sw.bb119 ], [ 12996, %sw.bb118 ], [ 12769, %sw.bb117 ], [ 12544, %sw.bb116 ], [ 12321, %sw.bb115 ], [ 12100, %sw.bb114 ], [ 11881, %sw.bb113 ], [ 11664, %sw.bb112 ], [ 11449, %sw.bb111 ], [ 11236, %sw.bb110 ], [ 11025, %sw.bb109 ], [ 10816, %sw.bb108 ], [ 10609, %sw.bb107 ], [ 10404, %sw.bb106 ], [ 10201, %sw.bb105 ], [ 10000, %sw.bb104 ], [ 9801, %sw.bb103 ], [ 9604, %sw.bb102 ], [ 9409, %sw.bb101 ], [ 9216, %sw.bb100 ], [ 9025, %sw.bb99 ], [ 8836, %sw.bb98 ], [ 8649, %sw.bb97 ], [ 8464, %sw.bb96 ], [ 8281, %sw.bb95 ], [ 8100, %sw.bb94 ], [ 7921, %sw.bb93 ], [ 7744, %sw.bb92 ], [ 7569, %sw.bb91 ], [ 7396, %sw.bb90 ], [ 7225, %sw.bb89 ], [ 7056, %sw.bb88 ], [ 6889, %sw.bb87 ], [ 6724, %sw.bb86 ], [ 6561, %sw.bb85 ], [ 6400, %sw.bb84 ], [ 6241, %sw.bb83 ], [ 6084, %sw.bb82 ], [ 5929, %sw.bb81 ], [ 5776, %sw.bb80 ], [ 5625, %sw.bb79 ], [ 5476, %sw.bb78 ], [ 5329, %sw.bb77 ], [ 5184, %sw.bb76 ], [ 5112, %sw.bb74 ], [ 4900, %sw.bb73 ], [ 4761, %sw.bb72 ], [ 4624, %sw.bb71 ], [ 4489, %sw.bb70 ], [ 4356, %sw.bb69 ], [ 4225, %sw.bb68 ], [ 4096, %sw.bb67 ], [ 3969, %sw.bb66 ], [ 3844, %sw.bb65 ], [ 3721, %sw.bb64 ], [ 3600, %sw.bb63 ], [ 3481, %sw.bb62 ], [ 3364, %sw.bb61 ], [ 3249, %sw.bb60 ], [ 3136, %sw.bb59 ], [ 3025, %sw.bb58 ], [ 2970, %sw.bb56 ], [ 2809, %sw.bb55 ], [ 2704, %sw.bb54 ], [ 2601, %sw.bb53 ], [ 2500, %sw.bb52 ], [ 2401, %sw.bb51 ], [ 2304, %sw.bb50 ], [ 2209, %sw.bb49 ], [ 2116, %sw.bb48 ], [ 2025, %sw.bb47 ], [ 1980, %sw.bb45 ], [ 1849, %sw.bb44 ], [ 1764, %sw.bb43 ], [ 1681, %sw.bb42 ], [ 1600, %sw.bb41 ], [ 1521, %sw.bb40 ], [ 1444, %sw.bb39 ], [ 1369, %sw.bb38 ], [ 1296, %sw.bb37 ], [ 1260, %sw.bb35 ], [ 1156, %sw.bb34 ], [ 1089, %sw.bb33 ], [ 1024, %sw.bb32 ], [ 961, %sw.bb31 ], [ 900, %sw.bb30 ], [ 841, %sw.bb29 ], [ 784, %sw.bb28 ], [ 729, %sw.bb27 ], [ 676, %sw.bb26 ], [ 625, %sw.bb25 ], [ 576, %sw.bb24 ], [ 529, %sw.bb23 ], [ 484, %sw.bb22 ], [ 441, %sw.bb21 ], [ 400, %sw.bb20 ], [ 361, %sw.bb19 ], [ 342, %sw.bb18 ], [ 289, %sw.bb17 ], [ 256, %sw.bb16 ], [ 225, %sw.bb15 ], [ 196, %sw.bb14 ], [ 169, %sw.bb13 ], [ 144, %sw.bb12 ], [ 121, %sw.bb11 ], [ 100, %sw.bb10 ], [ 81, %sw.bb9 ], [ 64, %sw.bb8 ], [ 49, %sw.bb7 ], [ 36, %sw.bb6 ], [ 25, %sw.bb5 ], [ 16, %sw.bb4 ], [ 9, %sw.bb3 ], [ 4, %sw.bb2 ], [ 1, %sw.bb1 ], [ 39601, %sw.bb203 ], [ 0, %if.end ] + ret i32 %retval.0 +} + +define i32 @cprop(i32 %x) { +entry: + switch i32 %x, label %sw.default [ + i32 1, label %return + i32 2, label %sw.bb1 + i32 3, label %sw.bb2 + i32 4, label %sw.bb2 + i32 5, label %sw.bb2 + i32 6, label %sw.bb3 + i32 7, label %sw.bb3 + ] + +sw.bb1: br label %return + +sw.bb2: + %and = and i32 %x, 1 + %tobool = icmp ne i32 %and, 0 + %cond = select i1 %tobool, i32 -123, i32 456 + %sub = sub nsw i32 %x, %cond + br label %return + +sw.bb3: + %trunc = trunc i32 %x to i8 + %sext = sext i8 %trunc to i32 + br label %return + +sw.default: + br label %return + +return: + %retval.0 = phi i32 [ 123, %sw.default ], [ %sext, %sw.bb3 ], [ %sub, %sw.bb2 ], [ 42, %sw.bb1 ], [ 5, %entry ] + ret i32 %retval.0 + +; CHECK: @cprop +; CHECK: switch.lookup: +; CHECK: %switch.gep = getelementptr inbounds [7 x i32]* @switch.table5, i32 0, i32 %switch.tableidx +} + +define i32 @unreachable(i32 %x) { +entry: + switch i32 %x, label %sw.default [ + i32 0, label %sw.bb + i32 1, label %sw.bb + i32 2, label %sw.bb + i32 3, label %sw.bb1 + i32 4, label %sw.bb2 + i32 5, label %sw.bb3 + i32 6, label %sw.bb3 + i32 7, label %sw.bb3 + i32 8, label %sw.bb3 + ] + +sw.bb: br label %return +sw.bb1: unreachable +sw.bb2: br label %return +sw.bb3: br label %return +sw.default: unreachable + +return: + %retval.0 = phi i32 [ 1, %sw.bb3 ], [ -1, %sw.bb2 ], [ 1, %sw.bb1 ], [ 0, %sw.bb ] + ret i32 %retval.0 + +; CHECK: @unreachable +; CHECK: switch.lookup: +; CHECK: getelementptr inbounds [5 x i32]* @switch.table6, i32 0, i32 %switch.tableidx +} diff --git a/test/Transforms/SimplifyLibCalls/2009-02-12-StrTo.ll b/test/Transforms/SimplifyLibCalls/2009-02-12-StrTo.ll deleted file mode 100644 index 2717228f7e..0000000000 --- a/test/Transforms/SimplifyLibCalls/2009-02-12-StrTo.ll +++ /dev/null @@ -1,14 +0,0 @@ -; RUN: opt < %s -simplify-libcalls -S | FileCheck %s - -; Test that we add nocapture to the declaration, and to the second call only. - -; CHECK: declare float @strtol(i8*, i8** nocapture, i32) nounwind -declare float @strtol(i8* %s, i8** %endptr, i32 %base) - -define void @foo(i8* %x, i8** %endptr) { -; CHECK: call float @strtol(i8* %x, i8** %endptr, i32 10) - call float @strtol(i8* %x, i8** %endptr, i32 10) -; CHECK: %2 = call float @strtol(i8* nocapture %x, i8** null, i32 10) - call float @strtol(i8* %x, i8** null, i32 10) - ret void -} diff --git a/test/Transforms/SimplifyLibCalls/StpCpy.ll b/test/Transforms/SimplifyLibCalls/StpCpy.ll deleted file mode 100644 index 914b0955bc..0000000000 --- a/test/Transforms/SimplifyLibCalls/StpCpy.ll +++ /dev/null @@ -1,43 +0,0 @@ -; Test that the StpCpyOptimizer works correctly -; RUN: opt < %s -simplify-libcalls -S | FileCheck %s - -; This transformation requires the pointer size, as it assumes that size_t is -; the size of a pointer. -target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128-n8:16:32" - -@hello = constant [6 x i8] c"hello\00" - -declare i8* @stpcpy(i8*, i8*) - -declare i8* @__stpcpy_chk(i8*, i8*, i32) nounwind - -declare i32 @llvm.objectsize.i32(i8*, i1) nounwind readonly - -define i32 @t1() { -; CHECK: @t1 - %target = alloca [1024 x i8] - %arg1 = getelementptr [1024 x i8]* %target, i32 0, i32 0 - %arg2 = getelementptr [6 x i8]* @hello, i32 0, i32 0 - %rslt1 = call i8* @stpcpy( i8* %arg1, i8* %arg2 ) -; CHECK: @llvm.memcpy.p0i8.p0i8.i32 - ret i32 0 -} - -define i32 @t2() { -; CHECK: @t2 - %target = alloca [1024 x i8] - %arg1 = getelementptr [1024 x i8]* %target, i32 0, i32 0 - %arg2 = getelementptr [6 x i8]* @hello, i32 0, i32 0 - %tmp1 = call i32 @llvm.objectsize.i32(i8* %arg1, i1 false) - %rslt1 = call i8* @__stpcpy_chk(i8* %arg1, i8* %arg2, i32 %tmp1) -; CHECK: @__memcpy_chk - ret i32 0 -} - -define i8* @t3(i8* %arg) { -; CHECK: @t3 - %stpcpy = tail call i8* @stpcpy(i8* %arg, i8* %arg) -; CHECK: [[LEN:%[a-z]+]] = call i32 @strlen(i8* %arg) -; CHECK-NEXT: getelementptr inbounds i8* %arg, i32 [[LEN]] - ret i8* %stpcpy -} diff --git a/test/Transforms/SimplifyLibCalls/StrLen.ll b/test/Transforms/SimplifyLibCalls/StrLen.ll deleted file mode 100644 index 4a20bbd2ce..0000000000 --- a/test/Transforms/SimplifyLibCalls/StrLen.ll +++ /dev/null @@ -1,62 +0,0 @@ -; Test that the StrCatOptimizer works correctly -; RUN: opt < %s -simplify-libcalls -S | \ -; RUN: not grep "call.*strlen" - -target datalayout = "e-p:32:32" -@hello = constant [6 x i8] c"hello\00" ; <[6 x i8]*> [#uses=3] -@null = constant [1 x i8] zeroinitializer ; <[1 x i8]*> [#uses=3] -@null_hello = constant [7 x i8] c"\00hello\00" ; <[7 x i8]*> [#uses=1] -@nullstring = constant i8 0 - -declare i32 @strlen(i8*) - -define i32 @test1() { - %hello_p = getelementptr [6 x i8]* @hello, i32 0, i32 0 ; <i8*> [#uses=1] - %hello_l = call i32 @strlen( i8* %hello_p ) ; <i32> [#uses=1] - ret i32 %hello_l -} - -define i32 @test2() { - %null_p = getelementptr [1 x i8]* @null, i32 0, i32 0 ; <i8*> [#uses=1] - %null_l = call i32 @strlen( i8* %null_p ) ; <i32> [#uses=1] - ret i32 %null_l -} - -define i32 @test3() { - %null_hello_p = getelementptr [7 x i8]* @null_hello, i32 0, i32 0 ; <i8*> [#uses=1] - %null_hello_l = call i32 @strlen( i8* %null_hello_p ) ; <i32> [#uses=1] - ret i32 %null_hello_l -} - -define i1 @test4() { - %hello_p = getelementptr [6 x i8]* @hello, i32 0, i32 0 ; <i8*> [#uses=1] - %hello_l = call i32 @strlen( i8* %hello_p ) ; <i32> [#uses=1] - %eq_hello = icmp eq i32 %hello_l, 0 ; <i1> [#uses=1] - ret i1 %eq_hello -} - -define i1 @test5() { - %null_p = getelementptr [1 x i8]* @null, i32 0, i32 0 ; <i8*> [#uses=1] - %null_l = call i32 @strlen( i8* %null_p ) ; <i32> [#uses=1] - %eq_null = icmp eq i32 %null_l, 0 ; <i1> [#uses=1] - ret i1 %eq_null -} - -define i1 @test6() { - %hello_p = getelementptr [6 x i8]* @hello, i32 0, i32 0 ; <i8*> [#uses=1] - %hello_l = call i32 @strlen( i8* %hello_p ) ; <i32> [#uses=1] - %ne_hello = icmp ne i32 %hello_l, 0 ; <i1> [#uses=1] - ret i1 %ne_hello -} - -define i1 @test7() { - %null_p = getelementptr [1 x i8]* @null, i32 0, i32 0 ; <i8*> [#uses=1] - %null_l = call i32 @strlen( i8* %null_p ) ; <i32> [#uses=1] - %ne_null = icmp ne i32 %null_l, 0 ; <i1> [#uses=1] - ret i1 %ne_null -} - -define i32 @test8() { - %len = tail call i32 @strlen(i8* @nullstring) nounwind - ret i32 %len -} diff --git a/test/Transforms/SimplifyLibCalls/StrNCpy.ll b/test/Transforms/SimplifyLibCalls/StrNCpy.ll deleted file mode 100644 index 4e47b31a6a..0000000000 --- a/test/Transforms/SimplifyLibCalls/StrNCpy.ll +++ /dev/null @@ -1,29 +0,0 @@ -; Test that the StrNCpyOptimizer works correctly -; RUN: opt < %s -simplify-libcalls -S | \ -; RUN: not grep "call.*strncpy" - -; This transformation requires the pointer size, as it assumes that size_t is -; the size of a pointer. -target datalayout = "-p:64:64:64" - -@hello = constant [6 x i8] c"hello\00" ; <[6 x i8]*> [#uses=1] -@null = constant [1 x i8] zeroinitializer ; <[1 x i8]*> [#uses=1] -@null_hello = constant [7 x i8] c"\00hello\00" ; <[7 x i8]*> [#uses=1] - -declare i8* @strncpy(i8*, i8*, i32) - -declare i32 @puts(i8*) - -define i32 @main() { - %target = alloca [1024 x i8] ; <[1024 x i8]*> [#uses=1] - %arg1 = getelementptr [1024 x i8]* %target, i32 0, i32 0 ; <i8*> [#uses=2] - store i8 0, i8* %arg1 - %arg2 = getelementptr [6 x i8]* @hello, i32 0, i32 0 ; <i8*> [#uses=1] - %rslt1 = call i8* @strncpy( i8* %arg1, i8* %arg2, i32 6 ) ; <i8*> [#uses=1] - %arg3 = getelementptr [1 x i8]* @null, i32 0, i32 0 ; <i8*> [#uses=1] - %rslt2 = call i8* @strncpy( i8* %rslt1, i8* %arg3, i32 42 ) ; <i8*> [#uses=1] - %arg4 = getelementptr [7 x i8]* @null_hello, i32 0, i32 0 ; <i8*> [#uses=1] - %rslt3 = call i8* @strncpy( i8* %rslt2, i8* %arg4, i32 42 ) ; <i8*> [#uses=1] - call i32 @puts( i8* %rslt3 ) ; <i32>:1 [#uses=0] - ret i32 0 -} diff --git a/test/Transforms/SimplifyLibCalls/StrPBrk.ll b/test/Transforms/SimplifyLibCalls/StrPBrk.ll deleted file mode 100644 index 29c3b7477b..0000000000 --- a/test/Transforms/SimplifyLibCalls/StrPBrk.ll +++ /dev/null @@ -1,25 +0,0 @@ -; RUN: opt < %s -simplify-libcalls -S | FileCheck %s - -target datalayout = "-p:64:64:64" - -@hello = constant [12 x i8] c"hello world\00" -@w = constant [2 x i8] c"w\00" -@null = constant [1 x i8] zeroinitializer - -declare i8* @strpbrk(i8*, i8*) - -define void @test(i8* %s1, i8* %s2) { - %hello_p = getelementptr [12 x i8]* @hello, i32 0, i32 0 - %w_p = getelementptr [2 x i8]* @w, i32 0, i32 0 - %null_p = getelementptr [1 x i8]* @null, i32 0, i32 0 - %test1 = call i8* @strpbrk(i8* %null_p, i8* %s2) - %test2 = call i8* @strpbrk(i8* %s1, i8* %null_p) -; CHECK-NOT: call i8* @strpbrk - %test3 = call i8* @strpbrk(i8* %s1, i8* %w_p) -; CHECK: call i8* @strchr(i8* %s1, i32 119) - %test4 = call i8* @strpbrk(i8* %hello_p, i8* %w_p) -; CHECK: getelementptr i8* %hello_p, i64 6 - %test5 = call i8* @strpbrk(i8* %s1, i8* %s2) -; CHECK: call i8* @strpbrk(i8* %s1, i8* %s2) - ret void -} |