[PATCH] D12573: check for fastness before merging in DAGCombiner::MergeConsecutiveStores()

[Matt Arsenault via llvm-commits]

arsenm added inline comments.

================
Comment at: lib/Target/AMDGPU/SIISelLowering.cpp:412-415
@@ -411,3 +411,6 @@
     // with adjacent offsets.
-    return Align % 4 == 0;
+    bool AlignedBy4 = (Align % 4 == 0);
+    if (IsFast)
+      *IsFast = AlignedBy4;
+    return AlignedBy4;
   }
----------------
This looks fine to me


http://reviews.llvm.org/D12573