From 931c52b84ca0ecf8f1736a080ac851e09628e5a2 Mon Sep 17 00:00:00 2001
From: Hal Finkel <hfinkel@anl.gov>
Date: Mon, 22 Oct 2012 18:00:55 +0000
Subject: [PATCH] BBVectorize should ignore unreachable blocks.

Unreachable blocks can have invalid instructions. For example,
jump threading can produce self-referential instructions in
unreachable blocks. Also, we should not be spending time
optimizing unreachable code. Fixes PR14133.

llvm-svn: 166423
---
 llvm/lib/Transforms/Vectorize/BBVectorize.cpp    | 13 ++++++++++
 llvm/test/Transforms/BBVectorize/simple-ldstr.ll | 32 ++++++++++++++++++++++++
 2 files changed, 45 insertions(+)

diff --git a/llvm/lib/Transforms/Vectorize/BBVectorize.cpp b/llvm/lib/Transforms/Vectorize/BBVectorize.cpp
index 1296ef1..81125f2 100644
--- a/llvm/lib/Transforms/Vectorize/BBVectorize.cpp
+++ b/llvm/lib/Transforms/Vectorize/BBVectorize.cpp
@@ -34,6 +34,7 @@
 #include "llvm/ADT/StringExtras.h"
 #include "llvm/Analysis/AliasAnalysis.h"
 #include "llvm/Analysis/AliasSetTracker.h"
+#include "llvm/Analysis/Dominators.h"
 #include "llvm/Analysis/ScalarEvolution.h"
 #include "llvm/Analysis/ScalarEvolutionExpressions.h"
 #include "llvm/Analysis/ValueTracking.h"
@@ -177,6 +178,7 @@ namespace {
     BBVectorize(Pass *P, const VectorizeConfig &C)
       : BasicBlockPass(ID), Config(C) {
       AA = &P->getAnalysis<AliasAnalysis>();
+      DT = &P->getAnalysis<DominatorTree>();
       SE = &P->getAnalysis<ScalarEvolution>();
       TD = P->getAnalysisIfAvailable<DataLayout>();
     }
@@ -191,6 +193,7 @@ namespace {
                 VPPIteratorPair;
 
     AliasAnalysis *AA;
+    DominatorTree *DT;
     ScalarEvolution *SE;
     DataLayout *TD;
 
@@ -330,6 +333,12 @@ namespace {
     void combineMetadata(Instruction *K, const Instruction *J);
 
     bool vectorizeBB(BasicBlock &BB) {
+      if (!DT->isReachableFromEntry(&BB)) {
+        DEBUG(dbgs() << "BBV: skipping unreachable " << BB.getName() <<
+              " in " << BB.getParent()->getName() << "\n");
+        return false;
+      }
+
       bool changed = false;
       // Iterate a sufficient number of times to merge types of size 1 bit,
       // then 2 bits, then 4, etc. up to half of the target vector width of the
@@ -363,6 +372,7 @@ namespace {
 
     virtual bool runOnBasicBlock(BasicBlock &BB) {
       AA = &getAnalysis<AliasAnalysis>();
+      DT = &getAnalysis<DominatorTree>();
       SE = &getAnalysis<ScalarEvolution>();
       TD = getAnalysisIfAvailable<DataLayout>();
 
@@ -372,8 +382,10 @@ namespace {
     virtual void getAnalysisUsage(AnalysisUsage &AU) const {
       BasicBlockPass::getAnalysisUsage(AU);
       AU.addRequired<AliasAnalysis>();
+      AU.addRequired<DominatorTree>();
       AU.addRequired<ScalarEvolution>();
       AU.addPreserved<AliasAnalysis>();
+      AU.addPreserved<DominatorTree>();
       AU.addPreserved<ScalarEvolution>();
       AU.setPreservesCFG();
     }
@@ -2397,6 +2409,7 @@ char BBVectorize::ID = 0;
 static const char bb_vectorize_name[] = "Basic-Block Vectorization";
 INITIALIZE_PASS_BEGIN(BBVectorize, BBV_NAME, bb_vectorize_name, false, false)
 INITIALIZE_AG_DEPENDENCY(AliasAnalysis)
+INITIALIZE_PASS_DEPENDENCY(DominatorTree)
 INITIALIZE_PASS_DEPENDENCY(ScalarEvolution)
 INITIALIZE_PASS_END(BBVectorize, BBV_NAME, bb_vectorize_name, false, false)
 
diff --git a/llvm/test/Transforms/BBVectorize/simple-ldstr.ll b/llvm/test/Transforms/BBVectorize/simple-ldstr.ll
index a5397ee..c1e6a09 100644
--- a/llvm/test/Transforms/BBVectorize/simple-ldstr.ll
+++ b/llvm/test/Transforms/BBVectorize/simple-ldstr.ll
@@ -108,3 +108,35 @@ entry:
 ; CHECK-AO: store <2 x float> %mulf, <2 x float>* %0, align 8
 ; CHECK-AO: ret void
 }
+
+; Simple 3-pair chain with loads and stores (unreachable)
+define void @test4(i1 %bool, double* %a, double* %b, double* %c) nounwind uwtable readonly {
+entry:
+  br i1 %bool, label %if.then1, label %if.end
+
+if.then1:
+  unreachable
+  br label %if.then
+
+if.then:
+  %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
+  br label %if.end
+
+if.end:
+  ret void
+; CHECK: @test4
+; CHECK-NOT: <2 x double>
+; CHECK-AO: @test4
+; CHECK-AO-NOT: <2 x double>
+}
+
-- 
2.7.4