Index: lib/Transforms/InstCombine/InstCombineVectorOps.cpp =================================================================== --- lib/Transforms/InstCombine/InstCombineVectorOps.cpp +++ lib/Transforms/InstCombine/InstCombineVectorOps.cpp @@ -1171,7 +1171,14 @@ SmallVector NewOps; bool NeedsRebuild = (Mask.size() != I->getType()->getVectorNumElements()); for (int i = 0, e = I->getNumOperands(); i != e; ++i) { - Value *V = evaluateInDifferentElementOrder(I->getOperand(i), Mask); + Value *V; + // Recursively call evaluateInDifferentElementOrder on vector arguments + // as well. E.g. GetElementPtr may have scalar operands even if the + // return value is a vector, so we need to examine the operand type. + if (I->getOperand(i)->getType()->isVectorTy()) + V = evaluateInDifferentElementOrder(I->getOperand(i), Mask); + else + V = I->getOperand(i); NewOps.push_back(V); NeedsRebuild |= (V != I->getOperand(i)); } Index: test/Transforms/InstCombine/vec_gep_scalar_arg.ll =================================================================== --- /dev/null +++ test/Transforms/InstCombine/vec_gep_scalar_arg.ll @@ -0,0 +1,16 @@ +; NOTE: Assertions have been autogenerated by utils/update_test_checks.py +; RUN: opt -instcombine -S < %s | FileCheck %s + +define <4 x i16*> @PR41270([4 x i16]* %x) { +; CHECK-LABEL: @PR41270( +; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x [4 x i16]*> undef, [4 x i16]* [[X:%.*]], i32 0 +; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds [4 x i16], <4 x [4 x i16]*> [[TMP1]], i64 0, i64 3 +; CHECK-NEXT: ret <4 x i16*> [[TMP2]] +; + %ins = insertelement <4 x [4 x i16]*> undef, [4 x i16]* %x, i32 0 + %splat = shufflevector <4 x [4 x i16]*> %ins, <4 x [4 x i16]*> undef, <4 x i32> zeroinitializer + %t2 = getelementptr inbounds [4 x i16], <4 x [4 x i16]*> %splat, i32 0, i32 3 + %t3 = extractelement <4 x i16*> %t2, i32 3 + %ins2 = insertelement <4 x i16*> undef, i16* %t3, i32 0 + ret <4 x i16*> %ins2 +}