[ConstantFold] Return the correct type when folding a GEP with vector indices

[davide]

Summary

As Eli pointed out (and I got wrong in the first place), langref says: "The getelementptr returns a vector of pointers, instead of a single address, when one or more of its arguments is a vector. In such cases, all vector arguments should have the same number of elements, and every scalar argument will be effectively broadcast into a vector during address calculation."

Costantfold for gep doesn't really take in account this paragraph, returning a pointer instead of a vector of pointer which triggers an assertion in RAUW, as we're trying to replace values with mistmatching types.

[efriedma]

This looks right.

lib/IR/ConstantFold.cpp

2074 ↗

(On Diff #115478)

Redundant call to PointerType::get()?

[davide]

lib/IR/ConstantFold.cpp

2074 ↗

(On Diff #115478)

Yes, let me fix that before committing.

llvm/trunk/lib/IR/ConstantFold.cpp

for (unsigned i = 0, e = Idxs.size(); i != e; ++i)
         isNull = false;
         break;
       }
     if (isNull) {
       PointerType *PtrTy = cast<PointerType>(C->getType()->getScalarType());
       Type *Ty = GetElementPtrInst::getIndexedType(PointeeTy, Idxs);
 
       assert(Ty && "Invalid indices for GEP!");
+      Type *OrigGEPTy = PointerType::get(Ty, PtrTy->getAddressSpace());
       Type *GEPTy = PointerType::get(Ty, PtrTy->getAddressSpace());
       if (VectorType *VT = dyn_cast<VectorType>(C->getType()))
-        GEPTy = VectorType::get(GEPTy, VT->getNumElements());
+        GEPTy = VectorType::get(OrigGEPTy, VT->getNumElements());
+
+      // The GEP returns a vector of pointers when one of more of
+      // its arguments is a vector.
+      for (unsigned i = 0, e = Idxs.size(); i != e; ++i) {
+        if (auto *VT = dyn_cast<VectorType>(Idxs[i]->getType())) {
+          GEPTy = VectorType::get(OrigGEPTy, VT->getNumElements());
+          break;
+        }
+      }
+
       return Constant::getNullValue(GEPTy);
     }
   }
 
   if (ConstantExpr *CE = dyn_cast<ConstantExpr>(C)) {
     // Combine Indices - If the source pointer to this getelementptr instruction
     // is a getelementptr instruction, combine the indices of the two
     // getelementptr instructions into a single instruction.

llvm/trunk/test/Transforms/InstCombine/pr34627.ll

+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -S -instcombine %s |FileCheck %s
+
+define <2 x i16> @patatino() {
+; CHECK-LABEL: @patatino(
+; CHECK-NEXT:    ret <2 x i16> zeroinitializer
+;
+  %tmp2 = getelementptr inbounds [1 x i16], [1 x i16]* null, i16 0, <2 x i16> undef
+  %tmp3 = ptrtoint <2 x i16*> %tmp2 to <2 x i16>
+  ret <2 x i16> %tmp3
+}