Index: llvm/trunk/lib/Transforms/Utils/LoopUtils.cpp
===================================================================
--- llvm/trunk/lib/Transforms/Utils/LoopUtils.cpp
+++ llvm/trunk/lib/Transforms/Utils/LoopUtils.cpp
@@ -565,7 +565,8 @@
     auto *I = Phi->user_back();
     if (I->isCast() && (I->getParent() == Phi->getParent()) && I->hasOneUse() &&
         DT->dominates(Previous, I->user_back())) {
-      SinkAfter[I] = Previous;
+      if (!DT->dominates(Previous, I)) // Otherwise we're good w/o sinking.
+        SinkAfter[I] = Previous;
       return true;
     }
   }
Index: llvm/trunk/lib/Transforms/Vectorize/LoopVectorize.cpp
===================================================================
--- llvm/trunk/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ llvm/trunk/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -7655,7 +7655,7 @@
   // 2. Copy and widen instructions from the old loop into the new loop.
 
   // Move instructions to handle first-order recurrences.
-  DenseMap<Instruction *, Instruction *> SinkAfter = Legal->getSinkAfter();
+  DenseMap<Instruction *, Instruction *> &SinkAfter = Legal->getSinkAfter();
   for (auto &Entry : SinkAfter) {
     Entry.first->removeFromParent();
     Entry.first->insertAfter(Entry.second);
Index: llvm/trunk/test/Transforms/LoopVectorize/first-order-recurrence.ll
===================================================================
--- llvm/trunk/test/Transforms/LoopVectorize/first-order-recurrence.ll
+++ llvm/trunk/test/Transforms/LoopVectorize/first-order-recurrence.ll
@@ -140,7 +140,10 @@
 ; CHECK:       vector.body:
 ; CHECK:         %vector.recur = phi <4 x i16> [ %vector.recur.init, %vector.ph ], [ [[L1:%[a-zA-Z0-9.]+]], %vector.body ]
 ; CHECK:         [[L1]] = load <4 x i16>
-; CHECK:         {{.*}} = shufflevector <4 x i16> %vector.recur, <4 x i16> [[L1]], <4 x i32> <i32 3, i32 4, i32 5, i32 6>
+; CHECK:         [[SHUF:%[a-zA-Z0-9.]+]] = shufflevector <4 x i16> %vector.recur, <4 x i16> [[L1]], <4 x i32> <i32 3, i32 4, i32 5, i32 6>
+; Check also that the casts were not moved needlessly.
+; CHECK:         sitofp <4 x i16> [[L1]] to <4 x double>
+; CHECK:         sitofp <4 x i16> [[SHUF]] to <4 x double> 
 ; CHECK:       middle.block:
 ; CHECK:         %vector.recur.extract = extractelement <4 x i16> [[L1]], i32 3
 ; CHECK:       scalar.ph: