Index: llvm/trunk/include/llvm/Analysis/ScalarEvolutionExpander.h =================================================================== --- llvm/trunk/include/llvm/Analysis/ScalarEvolutionExpander.h +++ llvm/trunk/include/llvm/Analysis/ScalarEvolutionExpander.h @@ -27,9 +27,16 @@ class TargetTransformInfo; /// Return true if the given expression is safe to expand in the sense that - /// all materialized values are safe to speculate. + /// all materialized values are safe to speculate anywhere their operands are + /// defined. bool isSafeToExpand(const SCEV *S, ScalarEvolution &SE); + /// Return true if the given expression is safe to expand in the sense that + /// all materialized values are defined and safe to speculate at the specified + /// location and their operands are defined at this location. + bool isSafeToExpandAt(const SCEV *S, const Instruction *InsertionPoint, + ScalarEvolution &SE); + /// This class uses information about analyze scalars to rewrite expressions /// in canonical form. /// Index: llvm/trunk/lib/Analysis/ScalarEvolutionExpander.cpp =================================================================== --- llvm/trunk/lib/Analysis/ScalarEvolutionExpander.cpp +++ llvm/trunk/lib/Analysis/ScalarEvolutionExpander.cpp @@ -2293,4 +2293,9 @@ visitAll(S, Search); return !Search.IsUnsafe; } + +bool isSafeToExpandAt(const SCEV *S, const Instruction *InsertionPoint, + ScalarEvolution &SE) { + return isSafeToExpand(S, SE) && SE.dominates(S, InsertionPoint->getParent()); +} } Index: llvm/trunk/unittests/Analysis/ScalarEvolutionTest.cpp =================================================================== --- llvm/trunk/unittests/Analysis/ScalarEvolutionTest.cpp +++ llvm/trunk/unittests/Analysis/ScalarEvolutionTest.cpp @@ -1113,5 +1113,76 @@ EXPECT_EQ(Expr, ZeroConst); } +// Check that we can correctly identify the points at which the SCEV of the +// AddRec can be expanded. +TEST_F(ScalarEvolutionsTest, SCEVExpanderIsSafeToExpandAt) { + /* + * Create the following code: + * func(i64 addrspace(10)* %arg) + * top: + * br label %L.ph + * L.ph: + * br label %L + * L: + * %phi = phi i64 [i64 0, %L.ph], [ %add, %L2 ] + * %add = add i64 %phi2, 1 + * %cond = icmp slt i64 %add, 1000; then becomes 2000. + * br i1 %cond, label %post, label %L2 + * post: + * ret void + * + */ + + // Create a module with non-integral pointers in it's datalayout + Module NIM("nonintegral", Context); + std::string DataLayout = M.getDataLayoutStr(); + if (!DataLayout.empty()) + DataLayout += "-"; + DataLayout += "ni:10"; + NIM.setDataLayout(DataLayout); + + Type *T_int64 = Type::getInt64Ty(Context); + Type *T_pint64 = T_int64->getPointerTo(10); + + FunctionType *FTy = + FunctionType::get(Type::getVoidTy(Context), {T_pint64}, false); + Function *F = cast(NIM.getOrInsertFunction("foo", FTy)); + + BasicBlock *Top = BasicBlock::Create(Context, "top", F); + BasicBlock *LPh = BasicBlock::Create(Context, "L.ph", F); + BasicBlock *L = BasicBlock::Create(Context, "L", F); + BasicBlock *Post = BasicBlock::Create(Context, "post", F); + + IRBuilder<> Builder(Top); + Builder.CreateBr(LPh); + + Builder.SetInsertPoint(LPh); + Builder.CreateBr(L); + + Builder.SetInsertPoint(L); + PHINode *Phi = Builder.CreatePHI(T_int64, 2); + auto *Add = cast( + Builder.CreateAdd(Phi, ConstantInt::get(T_int64, 1), "add")); + auto *Limit = ConstantInt::get(T_int64, 1000); + auto *Cond = cast( + Builder.CreateICmp(ICmpInst::ICMP_SLT, Add, Limit, "cond")); + Builder.CreateCondBr(Cond, L, Post); + Phi->addIncoming(ConstantInt::get(T_int64, 0), LPh); + Phi->addIncoming(Add, L); + + Builder.SetInsertPoint(Post); + Builder.CreateRetVoid(); + + ScalarEvolution SE = buildSE(*F); + const SCEV *S = SE.getSCEV(Phi); + EXPECT_TRUE(isa(S)); + const SCEVAddRecExpr *AR = cast(S); + EXPECT_TRUE(AR->isAffine()); + EXPECT_FALSE(isSafeToExpandAt(AR, Top->getTerminator(), SE)); + EXPECT_FALSE(isSafeToExpandAt(AR, LPh->getTerminator(), SE)); + EXPECT_TRUE(isSafeToExpandAt(AR, L->getTerminator(), SE)); + EXPECT_TRUE(isSafeToExpandAt(AR, Post->getTerminator(), SE)); +} + } // end anonymous namespace } // end namespace llvm