[AArch64] Teach AArch64TargetLowering::getOptimalMemOpType to consider alignment

restrictions when choosing a type for small-memcpy inlining in
SelectionDAGBuilder.

This ensures that the loads and stores output for the memcpy won't be further
expanded during legalization, which would cause the total number of instructions
for the memcpy to exceed (often significantly) the inlining thresholds.

<rdar://problem/17829180>

llvm-svn: 234462

lib/Target/AArch64/AArch64ISelLowering.cpp

@@ -6664,7 +6664,17 @@ EVT AArch64TargetLowering::getOptimalMemOpType(uint64_t Size, unsigned DstAlign,
        (allowsMisalignedMemoryAccesses(MVT::f128, 0, 1, &Fast) && Fast)))
     return MVT::f128;
 
-  return Size >= 8 ? MVT::i64 : MVT::i32;
+  if (Size >= 8 &&
+      (memOpAlign(SrcAlign, DstAlign, 8) ||
+       (allowsMisalignedMemoryAccesses(MVT::i64, 0, 1, &Fast) && Fast)))
+    return MVT::i64;
+
+  if (Size >= 4 &&
+      (memOpAlign(SrcAlign, DstAlign, 4) ||
+       (allowsMisalignedMemoryAccesses(MVT::i32, 0, 1, &Fast) && Fast)))
+    return MVT::i64;
+
+  return MVT::Other;
 }
 
 // 12-bit optionally shifted immediates are legal for adds.

test/CodeGen/AArch64/arm64-misaligned-memcpy-inline.ll

@@ -0,0 +1,14 @@
+; RUN: llc -march=arm64 -aarch64-strict-align < %s | FileCheck %s
+
+; Small (16-bytes here) unaligned memcpys should stay memcpy calls if
+; strict-alignment is turned on.
+define void @t0(i8* %out, i8* %in) {
+; CHECK-LABEL: t0:
+; CHECK:         orr w2, wzr, #0x10
+; CHECK-NEXT:    bl _memcpy
+entry:
+  call void @llvm.memcpy.p0i8.p0i8.i64(i8* %out, i8* %in, i64 16, i32 1, i1 false)
+  ret void
+}
+
+declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture readonly, i64, i32, i1)