Index: llvm/lib/Transforms/Scalar/LoopFuse.cpp
===================================================================
--- llvm/lib/Transforms/Scalar/LoopFuse.cpp
+++ llvm/lib/Transforms/Scalar/LoopFuse.cpp
@@ -1536,7 +1536,10 @@
     // Update DT/PDT
     DTU.applyUpdates(TreeUpdates);
 
+    LI.removeBlock(FC1GuardBlock);
     LI.removeBlock(FC1.Preheader);
+    LI.removeBlock(FC0.ExitBlock);
+    DTU.deleteBB(FC1GuardBlock);
     DTU.deleteBB(FC1.Preheader);
     DTU.deleteBB(FC0.ExitBlock);
     DTU.flush();
Index: llvm/test/Transforms/LoopFusion/double-loop-nest-inner-guard.ll
===================================================================
--- /dev/null
+++ llvm/test/Transforms/LoopFusion/double-loop-nest-inner-guard.ll
@@ -0,0 +1,80 @@
+; RUN: opt -S -loop-fusion < %s 2>&1 | FileCheck %s
+
+; TODO: Add CHECK rules
+
+@a = dso_local local_unnamed_addr global [10 x [10 x i32]] zeroinitializer
+@b = dso_local local_unnamed_addr global [10 x [10 x i32]] zeroinitializer
+@c = dso_local local_unnamed_addr global [10 x [10 x i32]] zeroinitializer
+
+define dso_local i32 @foo(i32 %m, i32 %n, i32 %M, i32 %N) local_unnamed_addr {
+entry:
+  %cmp63 = icmp sgt i32 %m, 0
+  br i1 %cmp63, label %for.cond1.preheader.lr.ph, label %for.cond.cleanup17
+
+for.cond1.preheader.lr.ph:
+  %cmp261 = icmp sgt i32 %n, 0
+  %wide.trip.count76 = zext i32 %m to i64
+  %wide.trip.count72 = zext i32 %n to i64
+  br label %for.cond1.preheader
+
+for.cond1.preheader:
+  %iv74 = phi i64 [ 0, %for.cond1.preheader.lr.ph ], [ %iv.next75, %for.cond.cleanup3 ]
+  br i1 %cmp261, label %for.body4.preheader, label %for.cond.cleanup3
+
+for.body4.preheader:
+  br label %for.body4
+
+for.body4:
+  %iv70 = phi i64 [ %iv.next71, %for.body4 ], [ 0, %for.body4.preheader ]
+  %idx6 = getelementptr inbounds [10 x [10 x i32]], [10 x [10 x i32]]* @a, i64 0, i64 %iv74, i64 %iv70
+  %0 = load i32, i32* %idx6
+  %add = add nsw i32 %0, 2
+  %idx10 = getelementptr inbounds [10 x [10 x i32]], [10 x [10 x i32]]* @b, i64 0, i64 %iv74, i64 %iv70
+  store i32 %add, i32* %idx10
+  %iv.next71 = add nuw nsw i64 %iv70, 1
+  %exitcond73 = icmp eq i64 %iv.next71, %wide.trip.count72
+  br i1 %exitcond73, label %for.cond.cleanup3.loopexit, label %for.body4
+
+for.cond.cleanup3.loopexit:
+  br label %for.cond.cleanup3
+
+for.cond.cleanup3:
+  %iv.next75 = add nuw nsw i64 %iv74, 1
+  %exitcond77 = icmp eq i64 %iv.next75, %wide.trip.count76
+  br i1 %exitcond77, label %for.cond20.preheader.preheader, label %for.cond1.preheader
+
+for.cond20.preheader.preheader:
+  br label %for.cond20.preheader
+
+for.cond20.preheader:
+  %iv66 = phi i64 [ %iv.next67, %for.cond.cleanup22 ], [ 0, %for.cond20.preheader.preheader ]
+  br i1 %cmp261, label %for.body23.preheader, label %for.cond.cleanup22
+
+for.body23.preheader:
+  br label %for.body23
+
+for.body23:
+  %iv = phi i64 [ %iv.next, %for.body23 ], [ 0, %for.body23.preheader ]
+  %idx27 = getelementptr inbounds [10 x [10 x i32]], [10 x [10 x i32]]* @a, i64 0, i64 %iv66, i64 %iv
+  %1 = load i32, i32* %idx27
+  %mul = shl nsw i32 %1, 1
+  %idx31 = getelementptr inbounds [10 x [10 x i32]], [10 x [10 x i32]]* @c, i64 0, i64 %iv66, i64 %iv
+  store i32 %mul, i32* %idx31
+  %iv.next = add nuw nsw i64 %iv, 1
+  %exitcond = icmp eq i64 %iv.next, %wide.trip.count72
+  br i1 %exitcond, label %for.cond.cleanup22.loopexit, label %for.body23
+
+for.cond.cleanup22.loopexit:
+  br label %for.cond.cleanup22
+
+for.cond.cleanup22:
+  %iv.next67 = add nuw nsw i64 %iv66, 1
+  %exitcond69 = icmp eq i64 %iv.next67, %wide.trip.count76
+  br i1 %exitcond69, label %for.cond.cleanup17.loopexit, label %for.cond20.preheader
+
+for.cond.cleanup17.loopexit:
+  br label %for.cond.cleanup17
+
+for.cond.cleanup17:
+  ret i32 undef
+}
Index: llvm/test/Transforms/LoopFusion/triple-loop-nest-inner-guard.ll
===================================================================
--- /dev/null
+++ llvm/test/Transforms/LoopFusion/triple-loop-nest-inner-guard.ll
@@ -0,0 +1,112 @@
+; RUN: opt -S -loop-fusion < %s 2>&1 | FileCheck %s
+
+; TODO: Add CHECK rules
+
+@a = dso_local local_unnamed_addr global [10 x [10 x [10 x i32]]] zeroinitializer
+@b = dso_local local_unnamed_addr global [10 x [10 x [10 x i32]]] zeroinitializer
+@c = dso_local local_unnamed_addr global [10 x [10 x [10 x i32]]] zeroinitializer
+
+define dso_local i32 @foo(i32 %m, i32 %n, i32 %M, i32 %N) local_unnamed_addr {
+entry:
+  %cmp101 = icmp sgt i32 %m, 0
+  br i1 %cmp101, label %for.cond1.preheader.lr.ph, label %for.cond.cleanup28
+
+for.cond1.preheader.lr.ph:
+  %cmp298 = icmp sgt i32 %n, 0
+  %cmp696 = icmp sgt i32 %M, 0
+  %wide.trip.count122 = zext i32 %m to i64
+  %wide.trip.count118 = zext i32 %n to i64
+  %wide.trip.count114 = zext i32 %M to i64
+  br label %for.cond1.preheader
+
+for.cond1.preheader:
+  %iv120 = phi i64 [ 0, %for.cond1.preheader.lr.ph ], [ %iv.next121, %for.cond.cleanup3 ]
+  br i1 %cmp298, label %for.cond5.preheader.preheader, label %for.cond.cleanup3
+
+for.cond5.preheader.preheader:
+  br label %for.cond5.preheader
+
+for.cond5.preheader:
+  %iv116 = phi i64 [ %iv.next117, %for.cond.cleanup7 ], [ 0, %for.cond5.preheader.preheader ]
+  br i1 %cmp696, label %for.body8.preheader, label %for.cond.cleanup7
+
+for.body8.preheader:
+  br label %for.body8
+
+for.cond.cleanup3.loopexit:
+  br label %for.cond.cleanup3
+
+for.cond.cleanup3:
+  %iv.next121 = add nuw nsw i64 %iv120, 1
+  %exitcond123 = icmp eq i64 %iv.next121, %wide.trip.count122
+  br i1 %exitcond123, label %for.cond31.preheader.preheader, label %for.cond1.preheader
+
+for.cond31.preheader.preheader:
+  br label %for.cond31.preheader
+
+for.cond.cleanup7.loopexit:
+  br label %for.cond.cleanup7
+
+for.cond.cleanup7:
+  %iv.next117 = add nuw nsw i64 %iv116, 1
+  %exitcond119 = icmp eq i64 %iv.next117, %wide.trip.count118
+  br i1 %exitcond119, label %for.cond.cleanup3.loopexit, label %for.cond5.preheader
+
+for.body8:
+  %iv112 = phi i64 [ %iv.next113, %for.body8 ], [ 0, %for.body8.preheader ]
+  %idx12 = getelementptr inbounds [10 x [10 x [10 x i32]]], [10 x [10 x [10 x i32]]]* @a, i64 0, i64 %iv120, i64 %iv116, i64 %iv112
+  %0 = load i32, i32* %idx12
+  %add = add nsw i32 %0, 2
+  %idx18 = getelementptr inbounds [10 x [10 x [10 x i32]]], [10 x [10 x [10 x i32]]]* @b, i64 0, i64 %iv120, i64 %iv116, i64 %iv112
+  store i32 %add, i32* %idx18
+  %iv.next113 = add nuw nsw i64 %iv112, 1
+  %exitcond115 = icmp eq i64 %iv.next113, %wide.trip.count114
+  br i1 %exitcond115, label %for.cond.cleanup7.loopexit, label %for.body8
+
+for.cond31.preheader:
+  %iv108 = phi i64 [ %iv.next109, %for.cond.cleanup33 ], [ 0, %for.cond31.preheader.preheader ]
+  br i1 %cmp298, label %for.cond36.preheader.preheader, label %for.cond.cleanup33
+
+for.cond36.preheader.preheader:
+  br label %for.cond36.preheader
+
+for.cond.cleanup28.loopexit:
+  br label %for.cond.cleanup28
+
+for.cond.cleanup28:
+  ret i32 undef
+
+for.cond36.preheader:
+  %iv104 = phi i64 [ %iv.next105, %for.cond.cleanup38 ], [ 0, %for.cond36.preheader.preheader ]
+  br i1 %cmp696, label %for.body39.preheader, label %for.cond.cleanup38
+
+for.body39.preheader:
+  br label %for.body39
+
+for.cond.cleanup33.loopexit:
+  br label %for.cond.cleanup33
+
+for.cond.cleanup33:
+  %iv.next109 = add nuw nsw i64 %iv108, 1
+  %exitcond111 = icmp eq i64 %iv.next109, %wide.trip.count122
+  br i1 %exitcond111, label %for.cond.cleanup28.loopexit, label %for.cond31.preheader
+
+for.cond.cleanup38.loopexit:
+  br label %for.cond.cleanup38
+
+for.cond.cleanup38:
+  %iv.next105 = add nuw nsw i64 %iv104, 1
+  %exitcond107 = icmp eq i64 %iv.next105, %wide.trip.count118
+  br i1 %exitcond107, label %for.cond.cleanup33.loopexit, label %for.cond36.preheader
+
+for.body39:
+  %iv = phi i64 [ %iv.next, %for.body39 ], [ 0, %for.body39.preheader ]
+  %idx45 = getelementptr inbounds [10 x [10 x [10 x i32]]], [10 x [10 x [10 x i32]]]* @a, i64 0, i64 %iv108, i64 %iv104, i64 %iv
+  %1 = load i32, i32* %idx45
+  %mul = shl nsw i32 %1, 1
+  %idx51 = getelementptr inbounds [10 x [10 x [10 x i32]]], [10 x [10 x [10 x i32]]]* @c, i64 0, i64 %iv108, i64 %iv104, i64 %iv
+  store i32 %mul, i32* %idx51
+  %iv.next = add nuw nsw i64 %iv, 1
+  %exitcond = icmp eq i64 %iv.next, %wide.trip.count114
+  br i1 %exitcond, label %for.cond.cleanup38.loopexit, label %for.body39
+}