diff --git a/lib/Transforms/Scalar/TailRecursionElimination.cpp b/lib/Transforms/Scalar/TailRecursionElimination.cpp
index 65b1f142821..f3c3e3054b6 100644
--- a/lib/Transforms/Scalar/TailRecursionElimination.cpp
+++ b/lib/Transforms/Scalar/TailRecursionElimination.cpp
@@ -404,18 +404,28 @@ bool TailCallElim::runTRE(Function &F) {
   // alloca' is changed from being a static alloca to being a dynamic alloca.
   // Until this is resolved, disable this transformation if that would ever
   // happen.  This bug is PR962.
+  SmallVector<BasicBlock*, 8> BBToErase;
   for (Function::iterator BB = F.begin(), E = F.end(); BB != E; ++BB) {
     if (ReturnInst *Ret = dyn_cast<ReturnInst>(BB->getTerminator())) {
       bool Change = ProcessReturningBlock(Ret, OldEntry, TailCallsAreMarkedTail,
                                           ArgumentPHIs, !CanTRETailMarkedCall);
-      if (!Change && BB->getFirstNonPHIOrDbg() == Ret)
+      if (!Change && BB->getFirstNonPHIOrDbg() == Ret) {
         Change = FoldReturnAndProcessPred(BB, Ret, OldEntry,
                                           TailCallsAreMarkedTail, ArgumentPHIs,
                                           !CanTRETailMarkedCall);
+        // FoldReturnAndProcessPred may have emptied some BB. Remember to
+        // erase them.
+        if (Change && BB->empty())
+          BBToErase.push_back(BB);
+
+      }
       MadeChange |= Change;
     }
   }
 
+  for (auto BB: BBToErase)
+    BB->eraseFromParent();
+
   // If we eliminated any tail recursions, it's possible that we inserted some
   // silly PHI nodes which just merge an initial value (the incoming operand)
   // with themselves.  Check to see if we did and clean up our mess if so.  This
@@ -823,8 +833,20 @@ bool TailCallElim::FoldReturnAndProcessPred(BasicBlock *BB,
     if (CallInst *CI = FindTRECandidate(BI, CannotTailCallElimCallsMarkedTail)){
       DEBUG(dbgs() << "FOLDING: " << *BB
             << "INTO UNCOND BRANCH PRED: " << *Pred);
-      EliminateRecursiveTailCall(CI, FoldReturnIntoUncondBranch(Ret, BB, Pred),
-                                 OldEntry, TailCallsAreMarkedTail, ArgumentPHIs,
+      ReturnInst *RI = FoldReturnIntoUncondBranch(Ret, BB, Pred);
+
+      // Cleanup: if all predecessors of BB have been eliminated by
+      // FoldReturnIntoUncondBranch, we would like to delete it, but we
+      // can not just nuke it as it is being used as an iterator by our caller.
+      // Just empty it, and the caller will erase it when it is safe to do so.
+      // It is important to empty it, because the ret instruction in there is
+      // still using a value which EliminateRecursiveTailCall will attempt
+      // to remove.
+      if (!BB->hasAddressTaken() && pred_begin(BB) == pred_end(BB))
+        BB->getInstList().clear();
+
+      EliminateRecursiveTailCall(CI, RI, OldEntry, TailCallsAreMarkedTail,
+                                 ArgumentPHIs,
                                  CannotTailCallElimCallsMarkedTail);
       ++NumRetDuped;
       Change = true;
diff --git a/test/Transforms/TailCallElim/EraseBB.ll b/test/Transforms/TailCallElim/EraseBB.ll
new file mode 100644
index 00000000000..c8290d7ea95
--- /dev/null
+++ b/test/Transforms/TailCallElim/EraseBB.ll
@@ -0,0 +1,26 @@
+; RUN: opt -tailcallelim -S < %s 2>&1 | FileCheck %s
+
+; CHECK: add nsw i32
+; CHECK-NEXT: br label
+; CHECK: add nsw i32
+; CHECK-NEXT: br label
+; CHECK-NOT: Uses remain when a value is destroyed
+define i32 @test(i32 %n) {
+entry:
+  %cmp = icmp slt i32 %n, 2
+  br i1 %cmp, label %if.then, label %if.else
+
+if.then:                                          ; preds = %entry
+  %v1 = add nsw i32 %n, -2
+  %call1 = tail call i32 @test(i32 %v1)
+  br label %return
+
+if.else:                                          ; preds = %entry
+  %v2 = add nsw i32 %n, 4
+  %call2 = tail call i32 @test(i32 %v2)
+  br label %return
+
+return:                                           ; preds = %if.end, %if.else
+  %retval = phi i32 [ %call1, %if.then ], [ %call2, %if.else ]
+  ret i32 %retval
+}