Don't pretend AMDGPU backend knows how to custom-lower UDIVREM for vector types; it can't

[tyomitch]

Summary

This recovers the rest of the tests broken by http://reviews.llvm.org/D13732

[arsenm]

LGTM

llvm/trunk/lib/Target/AMDGPU/AMDGPUISelLowering.cpp

for (MVT VT : VectorIntTypes) {
     setOperationAction(ISD::UINT_TO_FP, VT, Expand);
     setOperationAction(ISD::SDIV, VT, Expand);
     setOperationAction(ISD::UDIV, VT, Expand);
     setOperationAction(ISD::SREM, VT, Expand);
     setOperationAction(ISD::UREM, VT, Expand);
     setOperationAction(ISD::SMUL_LOHI, VT, Expand);
     setOperationAction(ISD::UMUL_LOHI, VT, Expand);
     setOperationAction(ISD::SDIVREM, VT, Custom);
-    setOperationAction(ISD::UDIVREM, VT, Custom);
+    setOperationAction(ISD::UDIVREM, VT, Expand);
     setOperationAction(ISD::ADDC, VT, Expand);
     setOperationAction(ISD::SUBC, VT, Expand);
     setOperationAction(ISD::ADDE, VT, Expand);
     setOperationAction(ISD::SUBE, VT, Expand);
     setOperationAction(ISD::SELECT, VT, Expand);
     setOperationAction(ISD::VSELECT, VT, Expand);
     setOperationAction(ISD::SELECT_CC, VT, Expand);
     setOperationAction(ISD::XOR,  VT, Expand);