Handle ARM "ret void" and function alignment with proper padding.

src/IceTargetLoweringARM32.h

Index: src/IceTargetLoweringARM32.h
diff --git a/src/IceTargetLoweringARM32.h b/src/IceTargetLoweringARM32.h
index e973652660c906e9c4ab2e4d2167bf094af7157c..c1862d59226a8d3244715e83866922179439fc4b 100644
--- a/src/IceTargetLoweringARM32.h
+++ b/src/IceTargetLoweringARM32.h
@@ -16,6 +16,7 @@
 #define SUBZERO_SRC_ICETARGETLOWERINGARM32_H
 
 #include "IceDefs.h"
+#include "IceInstARM32.h"
 #include "IceRegistersARM32.h"
 #include "IceTargetLowering.h"
 
@@ -91,6 +92,14 @@ protected:
 
   static Type stackSlotType();
 
+  // The following are helpers that insert lowered ARM32 instructions
+  // with minimal syntactic overhead, so that the lowering code can
+  // look as close to assembly as practical.
+
+  void _ret(Variable *LR, Variable *Src0 = nullptr) {
+    Context.insert(InstARM32Ret::create(Func, LR, Src0));
+  }
+
   bool UsesFramePointer;
   bool NeedsStackAlignment;
   llvm::SmallBitVector TypeToRegisterSet[IceType_NUM];