[AArch64][GlobalISel] Inline tiny memcpy et al at -O0.

FastISel already does this since the initial arm64 port was upstreamed, so
it seems there are no issues with doing this at -O0 for very small memcpys.

Gives a 0.2% geomean code size improvement on CTMark.

Differential Revision: https://reviews.llvm.org/D65758

llvm-svn: 367919

1 files changed, 5 insertions, 2 deletions

diff --git a/llvm/lib/Target/AArch64/AArch64PreLegalizerCombiner.cpp b/llvm/lib/Target/AArch64/AArch64PreLegalizerCombiner.cpp
index 5ec209ada17..835fcf09f0a 100644
--- a/llvm/lib/Target/AArch64/AArch64PreLegalizerCombiner.cpp
+++ b/llvm/lib/Target/AArch64/AArch64PreLegalizerCombiner.cpp
@@ -56,9 +56,12 @@ bool AArch64PreLegalizerCombinerInfo::combine(GISelChangeObserver &Observer,
     case Intrinsic::memcpy:
     case Intrinsic::memmove:
     case Intrinsic::memset: {
+      // If we're at -O0 set a maxlen of 32 to inline, otherwise let the other
+      // heuristics decide.
+      unsigned MaxLen = EnableOpt ? 0 : 32;
       // Try to inline memcpy type calls if optimizations are enabled.
-      return (EnableOpt && !EnableOptSize) ? Helper.tryCombineMemCpyFamily(MI)
-                                           : false;
+      return (!EnableOptSize) ? Helper.tryCombineMemCpyFamily(MI, MaxLen)
+                              : false;
     }
     default:
       break;