Index: lib/CodeGen/BranchFolding.cpp
===================================================================
--- lib/CodeGen/BranchFolding.cpp
+++ lib/CodeGen/BranchFolding.cpp
@@ -1074,28 +1074,24 @@
   if (!EnableTailMerge) return MadeChange;
 
   // First find blocks with no successors.
-  // Block placement does not create new tail merging opportunities for these
-  // blocks.
-  if (!AfterBlockPlacement) {
-    MergePotentials.clear();
-    for (MachineBasicBlock &MBB : MF) {
-      if (MergePotentials.size() == TailMergeThreshold)
-        break;
-      if (!TriedMerging.count(&MBB) && MBB.succ_empty())
-        MergePotentials.push_back(MergePotentialsElt(HashEndOfMBB(MBB), &MBB));
-    }
-
-    // If this is a large problem, avoid visiting the same basic blocks
-    // multiple times.
+  MergePotentials.clear();
+  for (MachineBasicBlock &MBB : MF) {
     if (MergePotentials.size() == TailMergeThreshold)
-      for (unsigned i = 0, e = MergePotentials.size(); i != e; ++i)
-        TriedMerging.insert(MergePotentials[i].getBlock());
-
-    // See if we can do any tail merging on those.
-    if (MergePotentials.size() >= 2)
-      MadeChange |= TryTailMergeBlocks(nullptr, nullptr, MinCommonTailLength);
+      break;
+    if (!TriedMerging.count(&MBB) && MBB.succ_empty())
+      MergePotentials.push_back(MergePotentialsElt(HashEndOfMBB(MBB), &MBB));
   }
 
+  // If this is a large problem, avoid visiting the same basic blocks
+  // multiple times.
+  if (MergePotentials.size() == TailMergeThreshold)
+    for (unsigned i = 0, e = MergePotentials.size(); i != e; ++i)
+      TriedMerging.insert(MergePotentials[i].getBlock());
+
+  // See if we can do any tail merging on those.
+  if (MergePotentials.size() >= 2)
+    MadeChange |= TryTailMergeBlocks(nullptr, nullptr, MinCommonTailLength);
+
   // Look at blocks (IBB) with multiple predecessors (PBB).
   // We change each predecessor to a canonical form, by
   // (1) temporarily removing any unconditional branch from the predecessor
Index: test/CodeGen/Thumb2/v8_IT_3.ll
===================================================================
--- test/CodeGen/Thumb2/v8_IT_3.ll
+++ test/CodeGen/Thumb2/v8_IT_3.ll
@@ -62,8 +62,7 @@
 ; CHECK-PIC-NEXT: beq
 ; CHECK-PIC: %bb6
 ; CHECK-PIC-NEXT: movs
-; CHECK-PIC-NEXT: add
-; CHECK-PIC-NEXT: pop
+; CHECK-PIC-NEXT: b
   ret i32 0
 
 bb6:
Index: test/CodeGen/WebAssembly/cfg-stackify.ll
===================================================================
--- test/CodeGen/WebAssembly/cfg-stackify.ll
+++ test/CodeGen/WebAssembly/cfg-stackify.ll
@@ -540,13 +540,8 @@
 ; OPT:      br          1{{$}}
 ; OPT-NEXT: .LBB13_3:
 ; OPT-NEXT: end_block{{$}}
-; OPT-NEXT: block   {{$}}
 ; OPT:      br_if       0, $pop{{[0-9]+}}{{$}}
-; OPT:      br_if       1, $pop{{[0-9]+}}{{$}}
-; OPT-NEXT: .LBB13_5:
-; OPT-NEXT: end_block{{$}}
-; OPT-NEXT: return{{$}}
-; OPT-NEXT: .LBB13_6:
+; OPT:      .LBB13_5:
 ; OPT-NEXT: end_block{{$}}
 ; OPT-NEXT: return{{$}}
 define void @test4(i32 %t) {
Index: test/CodeGen/WinEH/wineh-noret-cleanup.ll
===================================================================
--- test/CodeGen/WinEH/wineh-noret-cleanup.ll
+++ test/CodeGen/WinEH/wineh-noret-cleanup.ll
@@ -68,13 +68,13 @@
 ; SEH-NEXT:    .long   .Ltmp0@IMGREL+1
 ; SEH-NEXT:    .long   .Ltmp1@IMGREL+1
 ; SEH-NEXT:    .long   dummy_filter@IMGREL
-; SEH-NEXT:    .long   .LBB0_5@IMGREL
+; SEH-NEXT:    .long   .LBB0_2@IMGREL
 ; SEH-NEXT:    .long   .Ltmp2@IMGREL+1
 ; SEH-NEXT:    .long   .Ltmp3@IMGREL+1
-; SEH-NEXT:    .long   "?dtor$2@?0?test@4HA"@IMGREL
+; SEH-NEXT:    .long   "?dtor$5@?0?test@4HA"@IMGREL
 ; SEH-NEXT:    .long   0
 ; SEH-NEXT:    .long   .Ltmp2@IMGREL+1
 ; SEH-NEXT:    .long   .Ltmp3@IMGREL+1
 ; SEH-NEXT:    .long   dummy_filter@IMGREL
-; SEH-NEXT:    .long   .LBB0_5@IMGREL
+; SEH-NEXT:    .long   .LBB0_2@IMGREL
 ; SEH-NEXT:  .Llsda_end0:
Index: test/CodeGen/X86/conditional-tailcall.ll
===================================================================
--- test/CodeGen/X86/conditional-tailcall.ll
+++ test/CodeGen/X86/conditional-tailcall.ll
@@ -134,7 +134,7 @@
 
 ; Make sure Machine Copy Propagation doesn't delete the mov to %ecx becaue it
 ; thinks the conditional tail call clobbers it.
-; CHECK64-LABEL: .LBB3_11:
+; CHECK64-LABEL: .LBB3_9:
 ; CHECK64:       movzbl  (%rdi), %ecx
 ; CHECK64-NEXT:  addl    $-48, %ecx
 ; CHECK64-NEXT:  cmpl    $10, %ecx
Index: test/CodeGen/X86/loop-search.ll
===================================================================
--- test/CodeGen/X86/loop-search.ll
+++ test/CodeGen/X86/loop-search.ll
@@ -8,26 +8,22 @@
 ; CHECK-LABEL: search:
 ; CHECK:       ## %bb.0: ## %entry
 ; CHECK-NEXT:    testl %edx, %edx
-; CHECK-NEXT:    jle LBB0_1
-; CHECK-NEXT:  ## %bb.4: ## %for.body.preheader
+; CHECK-NEXT:    jle LBB0_5
+; CHECK-NEXT:  ## %bb.1: ## %for.body.preheader
 ; CHECK-NEXT:    movslq %edx, %rax
 ; CHECK-NEXT:    xorl %ecx, %ecx
 ; CHECK-NEXT:    .p2align 4, 0x90
-; CHECK-NEXT:  LBB0_5: ## %for.body
+; CHECK-NEXT:  LBB0_2: ## %for.body
 ; CHECK-NEXT:    ## =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    cmpl %edi, (%rsi,%rcx,4)
 ; CHECK-NEXT:    je LBB0_6
-; CHECK-NEXT:  ## %bb.2: ## %for.cond
-; CHECK-NEXT:    ## in Loop: Header=BB0_5 Depth=1
+; CHECK-NEXT:  ## %bb.3: ## %for.cond
+; CHECK-NEXT:    ## in Loop: Header=BB0_2 Depth=1
 ; CHECK-NEXT:    incq %rcx
 ; CHECK-NEXT:    cmpq %rax, %rcx
-; CHECK-NEXT:    jl LBB0_5
+; CHECK-NEXT:    jl LBB0_2
 ;            ### FIXME: %bb.3 and LBB0_1 should be merged
-; CHECK-NEXT:  ## %bb.3:
-; CHECK-NEXT:    xorl %eax, %eax
-; CHECK-NEXT:    ## kill: def $al killed $al killed $eax
-; CHECK-NEXT:    retq
-; CHECK-NEXT:  LBB0_1:
+; CHECK-NEXT:  LBB0_5:
 ; CHECK-NEXT:    xorl %eax, %eax
 ; CHECK-NEXT:    ## kill: def $al killed $al killed $eax
 ; CHECK-NEXT:    retq
Index: test/CodeGen/X86/machine-cp.ll
===================================================================
--- test/CodeGen/X86/machine-cp.ll
+++ test/CodeGen/X86/machine-cp.ll
@@ -8,22 +8,21 @@
 ; CHECK:       ## %bb.0: ## %entry
 ; CHECK-NEXT:    movl %edi, %eax
 ; CHECK-NEXT:    testl %esi, %esi
-; CHECK-NEXT:    je LBB0_1
-; CHECK-NEXT:  ## %bb.2: ## %while.body.preheader
+; CHECK-NEXT:    je LBB0_4
+; CHECK-NEXT:  ## %bb.1: ## %while.body.preheader
 ; CHECK-NEXT:    movl %esi, %edx
 ; CHECK-NEXT:    .p2align 4, 0x90
-; CHECK-NEXT:  LBB0_3: ## %while.body
+; CHECK-NEXT:  LBB0_2: ## %while.body
 ; CHECK-NEXT:    ## =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    movl %edx, %ecx
 ; CHECK-NEXT:    cltd
 ; CHECK-NEXT:    idivl %ecx
 ; CHECK-NEXT:    testl %edx, %edx
 ; CHECK-NEXT:    movl %ecx, %eax
-; CHECK-NEXT:    jne LBB0_3
-; CHECK-NEXT:  ## %bb.4: ## %while.end
+; CHECK-NEXT:    jne LBB0_2
+; CHECK-NEXT:  ## %bb.3: ## %while.end
 ; CHECK-NEXT:    movl %ecx, %eax
-; CHECK-NEXT:    retq
-; CHECK-NEXT:  LBB0_1:
+; CHECK-NEXT:  LBB0_4:
 ; CHECK-NEXT:    retq
 entry:
   %cmp1 = icmp eq i32 %b, 0
@@ -60,22 +59,21 @@
 ; CHECK:       ## %bb.0: ## %entry
 ; CHECK-NEXT:    movq %rdi, %rax
 ; CHECK-NEXT:    testq %rsi, %rsi
-; CHECK-NEXT:    je LBB2_1
-; CHECK-NEXT:  ## %bb.2: ## %while.body.preheader
+; CHECK-NEXT:    je LBB2_4
+; CHECK-NEXT:  ## %bb.1: ## %while.body.preheader
 ; CHECK-NEXT:    movq %rsi, %rdx
 ; CHECK-NEXT:    .p2align 4, 0x90
-; CHECK-NEXT:  LBB2_3: ## %while.body
+; CHECK-NEXT:  LBB2_2: ## %while.body
 ; CHECK-NEXT:    ## =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    movq %rdx, %rcx
 ; CHECK-NEXT:    cqto
 ; CHECK-NEXT:    idivq %rcx
 ; CHECK-NEXT:    testq %rdx, %rdx
 ; CHECK-NEXT:    movq %rcx, %rax
-; CHECK-NEXT:    jne LBB2_3
-; CHECK-NEXT:  ## %bb.4: ## %while.end
+; CHECK-NEXT:    jne LBB2_2
+; CHECK-NEXT:  ## %bb.3: ## %while.end
 ; CHECK-NEXT:    movl %ecx, %eax
-; CHECK-NEXT:    retq
-; CHECK-NEXT:  LBB2_1:
+; CHECK-NEXT:  LBB2_4:
 ; CHECK-NEXT:    retq
 entry:
   %cmp1 = icmp eq i64 %b, 0
Index: test/CodeGen/X86/mul-constant-result.ll
===================================================================
--- test/CodeGen/X86/mul-constant-result.ll
+++ test/CodeGen/X86/mul-constant-result.ll
@@ -28,7 +28,7 @@
 ; X86-NEXT:  .LBB0_4:
 ; X86-NEXT:    decl %ecx
 ; X86-NEXT:    cmpl $31, %ecx
-; X86-NEXT:    ja .LBB0_39
+; X86-NEXT:    ja .LBB0_7
 ; X86-NEXT:  # %bb.5:
 ; X86-NEXT:    jmpl *.LJTI0_0(,%ecx,4)
 ; X86-NEXT:  .LBB0_6:
@@ -36,209 +36,149 @@
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_39:
-; X86-NEXT:    .cfi_def_cfa_offset 8
-; X86-NEXT:    xorl %eax, %eax
-; X86-NEXT:  .LBB0_40:
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
 ; X86-NEXT:  .LBB0_7:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
-; X86-NEXT:    leal (%eax,%eax,2), %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
+; X86-NEXT:    xorl %eax, %eax
 ; X86-NEXT:  .LBB0_8:
-; X86-NEXT:    .cfi_def_cfa_offset 8
-; X86-NEXT:    shll $2, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_9:
-; X86-NEXT:    .cfi_def_cfa_offset 8
-; X86-NEXT:    leal (%eax,%eax,4), %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
 ; X86-NEXT:  .LBB0_10:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
-; X86-NEXT:    addl %eax, %eax
-; X86-NEXT:    leal (%eax,%eax,2), %eax
+; X86-NEXT:    shll $2, %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_11:
+; X86-NEXT:  .LBB0_12:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
-; X86-NEXT:    leal (,%eax,8), %ecx
-; X86-NEXT:    jmp .LBB0_12
+; X86-NEXT:    addl %eax, %eax
+; X86-NEXT:    jmp .LBB0_9
 ; X86-NEXT:  .LBB0_13:
-; X86-NEXT:    shll $3, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
+; X86-NEXT:    leal (,%eax,8), %ecx
+; X86-NEXT:    jmp .LBB0_41
 ; X86-NEXT:  .LBB0_14:
-; X86-NEXT:    .cfi_def_cfa_offset 8
-; X86-NEXT:    leal (%eax,%eax,8), %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_15:
-; X86-NEXT:    .cfi_def_cfa_offset 8
-; X86-NEXT:    addl %eax, %eax
-; X86-NEXT:    leal (%eax,%eax,4), %eax
+; X86-NEXT:    shll $3, %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
 ; X86-NEXT:  .LBB0_16:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
-; X86-NEXT:    leal (%eax,%eax,4), %ecx
-; X86-NEXT:    leal (%eax,%ecx,2), %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
+; X86-NEXT:    addl %eax, %eax
+; X86-NEXT:    jmp .LBB0_11
 ; X86-NEXT:  .LBB0_17:
-; X86-NEXT:    .cfi_def_cfa_offset 8
+; X86-NEXT:    leal (%eax,%eax,4), %ecx
+; X86-NEXT:    jmp .LBB0_18
+; X86-NEXT:  .LBB0_19:
 ; X86-NEXT:    shll $2, %eax
-; X86-NEXT:    leal (%eax,%eax,2), %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_18:
-; X86-NEXT:    .cfi_def_cfa_offset 8
+; X86-NEXT:    jmp .LBB0_9
+; X86-NEXT:  .LBB0_20:
 ; X86-NEXT:    leal (%eax,%eax,2), %ecx
-; X86-NEXT:    leal (%eax,%ecx,4), %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_19:
-; X86-NEXT:    .cfi_def_cfa_offset 8
+; X86-NEXT:    jmp .LBB0_21
+; X86-NEXT:  .LBB0_22:
 ; X86-NEXT:    movl %eax, %ecx
 ; X86-NEXT:    shll $4, %ecx
 ; X86-NEXT:    subl %eax, %ecx
-; X86-NEXT:    jmp .LBB0_12
-; X86-NEXT:  .LBB0_21:
+; X86-NEXT:    jmp .LBB0_41
+; X86-NEXT:  .LBB0_23:
 ; X86-NEXT:    leal (%eax,%eax,4), %eax
-; X86-NEXT:    leal (%eax,%eax,2), %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_22:
-; X86-NEXT:    .cfi_def_cfa_offset 8
+; X86-NEXT:    jmp .LBB0_9
+; X86-NEXT:  .LBB0_24:
 ; X86-NEXT:    shll $4, %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_23:
+; X86-NEXT:  .LBB0_25:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
 ; X86-NEXT:    movl %eax, %ecx
 ; X86-NEXT:    shll $4, %ecx
-; X86-NEXT:    addl %ecx, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_24:
-; X86-NEXT:    .cfi_def_cfa_offset 8
+; X86-NEXT:    jmp .LBB0_26
+; X86-NEXT:  .LBB0_27:
 ; X86-NEXT:    addl %eax, %eax
+; X86-NEXT:  .LBB0_15:
 ; X86-NEXT:    leal (%eax,%eax,8), %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_25:
+; X86-NEXT:  .LBB0_28:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
 ; X86-NEXT:    leal (%eax,%eax,8), %ecx
+; X86-NEXT:  .LBB0_18:
 ; X86-NEXT:    leal (%eax,%ecx,2), %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_26:
+; X86-NEXT:  .LBB0_29:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
 ; X86-NEXT:    shll $2, %eax
-; X86-NEXT:    leal (%eax,%eax,4), %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_27:
-; X86-NEXT:    .cfi_def_cfa_offset 8
+; X86-NEXT:    jmp .LBB0_11
+; X86-NEXT:  .LBB0_30:
 ; X86-NEXT:    leal (%eax,%eax,4), %ecx
+; X86-NEXT:  .LBB0_21:
 ; X86-NEXT:    leal (%eax,%ecx,4), %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_28:
+; X86-NEXT:  .LBB0_31:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
 ; X86-NEXT:    leal (%eax,%eax,4), %ecx
 ; X86-NEXT:    leal (%eax,%ecx,4), %ecx
-; X86-NEXT:    addl %ecx, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_29:
-; X86-NEXT:    .cfi_def_cfa_offset 8
+; X86-NEXT:    jmp .LBB0_26
+; X86-NEXT:  .LBB0_32:
 ; X86-NEXT:    leal (%eax,%eax,2), %ecx
 ; X86-NEXT:    shll $3, %ecx
-; X86-NEXT:    jmp .LBB0_12
-; X86-NEXT:  .LBB0_30:
+; X86-NEXT:    jmp .LBB0_41
+; X86-NEXT:  .LBB0_33:
 ; X86-NEXT:    shll $3, %eax
-; X86-NEXT:    leal (%eax,%eax,2), %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_31:
-; X86-NEXT:    .cfi_def_cfa_offset 8
+; X86-NEXT:    jmp .LBB0_9
+; X86-NEXT:  .LBB0_34:
 ; X86-NEXT:    leal (%eax,%eax,4), %eax
+; X86-NEXT:  .LBB0_11:
 ; X86-NEXT:    leal (%eax,%eax,4), %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_32:
+; X86-NEXT:  .LBB0_35:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
 ; X86-NEXT:    leal (%eax,%eax,4), %ecx
 ; X86-NEXT:    leal (%ecx,%ecx,4), %ecx
-; X86-NEXT:    addl %ecx, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_33:
-; X86-NEXT:    .cfi_def_cfa_offset 8
+; X86-NEXT:    jmp .LBB0_26
+; X86-NEXT:  .LBB0_36:
 ; X86-NEXT:    leal (%eax,%eax,8), %eax
+; X86-NEXT:  .LBB0_9:
 ; X86-NEXT:    leal (%eax,%eax,2), %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_34:
+; X86-NEXT:  .LBB0_37:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
 ; X86-NEXT:    leal (%eax,%eax,8), %ecx
 ; X86-NEXT:    leal (%ecx,%ecx,2), %ecx
-; X86-NEXT:    addl %ecx, %eax
-; X86-NEXT:    popl %esi
-; X86-NEXT:    .cfi_def_cfa_offset 4
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_35:
-; X86-NEXT:    .cfi_def_cfa_offset 8
+; X86-NEXT:    jmp .LBB0_26
+; X86-NEXT:  .LBB0_38:
 ; X86-NEXT:    leal (%eax,%eax,8), %ecx
 ; X86-NEXT:    leal (%ecx,%ecx,2), %ecx
 ; X86-NEXT:    addl %eax, %ecx
+; X86-NEXT:  .LBB0_26:
 ; X86-NEXT:    addl %ecx, %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_36:
+; X86-NEXT:  .LBB0_39:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
 ; X86-NEXT:    movl %eax, %ecx
 ; X86-NEXT:    shll $5, %ecx
 ; X86-NEXT:    subl %eax, %ecx
-; X86-NEXT:    jmp .LBB0_12
-; X86-NEXT:  .LBB0_37:
+; X86-NEXT:    jmp .LBB0_41
+; X86-NEXT:  .LBB0_40:
 ; X86-NEXT:    movl %eax, %ecx
 ; X86-NEXT:    shll $5, %ecx
-; X86-NEXT:  .LBB0_12:
+; X86-NEXT:  .LBB0_41:
 ; X86-NEXT:    subl %eax, %ecx
 ; X86-NEXT:    movl %ecx, %eax
 ; X86-NEXT:    popl %esi
 ; X86-NEXT:    .cfi_def_cfa_offset 4
 ; X86-NEXT:    retl
-; X86-NEXT:  .LBB0_38:
+; X86-NEXT:  .LBB0_42:
 ; X86-NEXT:    .cfi_def_cfa_offset 8
 ; X86-NEXT:    shll $5, %eax
 ; X86-NEXT:    popl %esi
@@ -256,158 +196,149 @@
 ; X64-HSW-NEXT:    cmovel %ecx, %eax
 ; X64-HSW-NEXT:    addl $-1, %edi
 ; X64-HSW-NEXT:    cmpl $31, %edi
-; X64-HSW-NEXT:    ja .LBB0_36
+; X64-HSW-NEXT:    ja .LBB0_3
 ; X64-HSW-NEXT:  # %bb.1:
 ; X64-HSW-NEXT:    jmpq *.LJTI0_0(,%rdi,8)
 ; X64-HSW-NEXT:  .LBB0_2:
 ; X64-HSW-NEXT:    addl %eax, %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_36:
-; X64-HSW-NEXT:    xorl %eax, %eax
-; X64-HSW-NEXT:  .LBB0_37:
-; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
-; X64-HSW-NEXT:    retq
 ; X64-HSW-NEXT:  .LBB0_3:
-; X64-HSW-NEXT:    leal (%rax,%rax,2), %eax
-; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
-; X64-HSW-NEXT:    retq
+; X64-HSW-NEXT:    xorl %eax, %eax
 ; X64-HSW-NEXT:  .LBB0_4:
-; X64-HSW-NEXT:    shll $2, %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_5:
-; X64-HSW-NEXT:    leal (%rax,%rax,4), %eax
+; X64-HSW-NEXT:  .LBB0_6:
+; X64-HSW-NEXT:    shll $2, %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_6:
+; X64-HSW-NEXT:  .LBB0_8:
 ; X64-HSW-NEXT:    addl %eax, %eax
+; X64-HSW-NEXT:  .LBB0_5:
 ; X64-HSW-NEXT:    leal (%rax,%rax,2), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_7:
-; X64-HSW-NEXT:    leal (,%rax,8), %ecx
-; X64-HSW-NEXT:    jmp .LBB0_8
 ; X64-HSW-NEXT:  .LBB0_9:
-; X64-HSW-NEXT:    shll $3, %eax
-; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
-; X64-HSW-NEXT:    retq
+; X64-HSW-NEXT:    leal (,%rax,8), %ecx
+; X64-HSW-NEXT:    jmp .LBB0_37
 ; X64-HSW-NEXT:  .LBB0_10:
-; X64-HSW-NEXT:    leal (%rax,%rax,8), %eax
+; X64-HSW-NEXT:    shll $3, %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_11:
+; X64-HSW-NEXT:  .LBB0_12:
 ; X64-HSW-NEXT:    addl %eax, %eax
+; X64-HSW-NEXT:  .LBB0_7:
 ; X64-HSW-NEXT:    leal (%rax,%rax,4), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_12:
+; X64-HSW-NEXT:  .LBB0_13:
 ; X64-HSW-NEXT:    leal (%rax,%rax,4), %ecx
 ; X64-HSW-NEXT:    leal (%rax,%rcx,2), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_13:
+; X64-HSW-NEXT:  .LBB0_15:
 ; X64-HSW-NEXT:    shll $2, %eax
 ; X64-HSW-NEXT:    leal (%rax,%rax,2), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_14:
+; X64-HSW-NEXT:  .LBB0_16:
 ; X64-HSW-NEXT:    leal (%rax,%rax,2), %ecx
 ; X64-HSW-NEXT:    leal (%rax,%rcx,4), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_15:
+; X64-HSW-NEXT:  .LBB0_18:
 ; X64-HSW-NEXT:    movl %eax, %ecx
 ; X64-HSW-NEXT:    shll $4, %ecx
 ; X64-HSW-NEXT:    subl %eax, %ecx
-; X64-HSW-NEXT:    jmp .LBB0_8
-; X64-HSW-NEXT:  .LBB0_17:
+; X64-HSW-NEXT:    jmp .LBB0_37
+; X64-HSW-NEXT:  .LBB0_19:
 ; X64-HSW-NEXT:    leal (%rax,%rax,4), %eax
 ; X64-HSW-NEXT:    leal (%rax,%rax,2), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_18:
+; X64-HSW-NEXT:  .LBB0_20:
 ; X64-HSW-NEXT:    shll $4, %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_19:
+; X64-HSW-NEXT:  .LBB0_21:
 ; X64-HSW-NEXT:    movl %eax, %ecx
 ; X64-HSW-NEXT:    shll $4, %ecx
-; X64-HSW-NEXT:    jmp .LBB0_20
-; X64-HSW-NEXT:  .LBB0_21:
+; X64-HSW-NEXT:    jmp .LBB0_34
+; X64-HSW-NEXT:  .LBB0_22:
 ; X64-HSW-NEXT:    addl %eax, %eax
+; X64-HSW-NEXT:  .LBB0_11:
 ; X64-HSW-NEXT:    leal (%rax,%rax,8), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_22:
+; X64-HSW-NEXT:  .LBB0_23:
 ; X64-HSW-NEXT:    leal (%rax,%rax,8), %ecx
 ; X64-HSW-NEXT:    leal (%rax,%rcx,2), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_23:
+; X64-HSW-NEXT:  .LBB0_24:
 ; X64-HSW-NEXT:    shll $2, %eax
 ; X64-HSW-NEXT:    leal (%rax,%rax,4), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_24:
+; X64-HSW-NEXT:  .LBB0_25:
 ; X64-HSW-NEXT:    leal (%rax,%rax,4), %ecx
 ; X64-HSW-NEXT:    leal (%rax,%rcx,4), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_25:
+; X64-HSW-NEXT:  .LBB0_26:
 ; X64-HSW-NEXT:    leal (%rax,%rax,4), %ecx
 ; X64-HSW-NEXT:    leal (%rax,%rcx,4), %ecx
-; X64-HSW-NEXT:    jmp .LBB0_20
-; X64-HSW-NEXT:  .LBB0_26:
+; X64-HSW-NEXT:    jmp .LBB0_34
+; X64-HSW-NEXT:  .LBB0_27:
 ; X64-HSW-NEXT:    leal (%rax,%rax,2), %ecx
 ; X64-HSW-NEXT:    shll $3, %ecx
-; X64-HSW-NEXT:    jmp .LBB0_8
-; X64-HSW-NEXT:  .LBB0_27:
+; X64-HSW-NEXT:    jmp .LBB0_37
+; X64-HSW-NEXT:  .LBB0_28:
 ; X64-HSW-NEXT:    shll $3, %eax
 ; X64-HSW-NEXT:    leal (%rax,%rax,2), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_28:
+; X64-HSW-NEXT:  .LBB0_29:
 ; X64-HSW-NEXT:    leal (%rax,%rax,4), %eax
 ; X64-HSW-NEXT:    leal (%rax,%rax,4), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_29:
+; X64-HSW-NEXT:  .LBB0_30:
 ; X64-HSW-NEXT:    leal (%rax,%rax,4), %ecx
 ; X64-HSW-NEXT:    leal (%rcx,%rcx,4), %ecx
-; X64-HSW-NEXT:    jmp .LBB0_20
-; X64-HSW-NEXT:  .LBB0_30:
+; X64-HSW-NEXT:    jmp .LBB0_34
+; X64-HSW-NEXT:  .LBB0_31:
 ; X64-HSW-NEXT:    leal (%rax,%rax,8), %eax
 ; X64-HSW-NEXT:    leal (%rax,%rax,2), %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_31:
+; X64-HSW-NEXT:  .LBB0_32:
 ; X64-HSW-NEXT:    leal (%rax,%rax,8), %ecx
 ; X64-HSW-NEXT:    leal (%rcx,%rcx,2), %ecx
-; X64-HSW-NEXT:    jmp .LBB0_20
-; X64-HSW-NEXT:  .LBB0_32:
+; X64-HSW-NEXT:    jmp .LBB0_34
+; X64-HSW-NEXT:  .LBB0_33:
 ; X64-HSW-NEXT:    leal (%rax,%rax,8), %ecx
 ; X64-HSW-NEXT:    leal (%rcx,%rcx,2), %ecx
 ; X64-HSW-NEXT:    addl %eax, %ecx
-; X64-HSW-NEXT:  .LBB0_20:
+; X64-HSW-NEXT:  .LBB0_34:
 ; X64-HSW-NEXT:    addl %eax, %ecx
 ; X64-HSW-NEXT:    movl %ecx, %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_33:
+; X64-HSW-NEXT:  .LBB0_35:
 ; X64-HSW-NEXT:    movl %eax, %ecx
 ; X64-HSW-NEXT:    shll $5, %ecx
 ; X64-HSW-NEXT:    subl %eax, %ecx
-; X64-HSW-NEXT:    jmp .LBB0_8
-; X64-HSW-NEXT:  .LBB0_34:
+; X64-HSW-NEXT:    jmp .LBB0_37
+; X64-HSW-NEXT:  .LBB0_36:
 ; X64-HSW-NEXT:    movl %eax, %ecx
 ; X64-HSW-NEXT:    shll $5, %ecx
-; X64-HSW-NEXT:  .LBB0_8:
+; X64-HSW-NEXT:  .LBB0_37:
 ; X64-HSW-NEXT:    subl %eax, %ecx
 ; X64-HSW-NEXT:    movl %ecx, %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
-; X64-HSW-NEXT:  .LBB0_35:
+; X64-HSW-NEXT:  .LBB0_39:
 ; X64-HSW-NEXT:    shll $5, %eax
 ; X64-HSW-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-HSW-NEXT:    retq
Index: test/CodeGen/X86/tail-merge-after-mbp.mir
===================================================================
--- test/CodeGen/X86/tail-merge-after-mbp.mir
+++ test/CodeGen/X86/tail-merge-after-mbp.mir
@@ -4,28 +4,28 @@
 # check loop bb.7 is not merged with bb.10, bb.13
 # check loop bb.9 is not merged with bb.12
 # CHECK:  bb.2:
-# CHECK-NEXT:    successors: %bb.3(0x30000000), %bb.4(0x50000000)
+# CHECK-NEXT:    successors: %bb.1(0x30000000), %bb.3(0x50000000)
 # CHECK:    $rax = MOV64rm $r14, 1, $noreg, 0, $noreg
 # CHECK-NEXT:    TEST64rr $rax, $rax
-# CHECK-NEXT:    JE_1 %bb.3
-# CHECK:  bb.4:
-# CHECK-NEXT:    successors: %bb.5(0x30000000), %bb.10(0x50000000)
+# CHECK-NEXT:    JE_1 %bb.1
+# CHECK:  bb.3:
+# CHECK-NEXT:    successors: %bb.4(0x30000000), %bb.8(0x50000000)
 # CHECK:    CMP64mi8 killed $rax, 1, $noreg, 8, $noreg, 0
-# CHECK-NEXT:    JNE_1 %bb.10
-# CHECK:  bb.5:
-# CHECK-NEXT:    successors: %bb.6(0x30000000), %bb.7(0x50000000)
+# CHECK-NEXT:    JNE_1 %bb.8
+# CHECK:  bb.4:
+# CHECK-NEXT:    successors: %bb.1(0x30000000), %bb.5(0x50000000)
 # CHECK:    $rax = MOV64rm $r14, 1, $noreg, 0, $noreg
 # CHECK-NEXT:    TEST64rr $rax, $rax
-# CHECK-NEXT:    JE_1 %bb.6
-# CHECK:  bb.7
-# CHECK-NEXT:    successors: %bb.8(0x71555555), %bb.10(0x0eaaaaab)
+# CHECK-NEXT:    JE_1 %bb.1
+# CHECK:  bb.5
+# CHECK-NEXT:    successors: %bb.6(0x71555555), %bb.8(0x0eaaaaab)
 # CHECK:    CMP64mi8 killed $rax, 1, $noreg, 8, $noreg, 0
-# CHECK-NEXT:    JNE_1 %bb.10
-# CHECK:  bb.8:
-# CHECK-NEXT:    successors: %bb.9(0x04000000), %bb.7(0x7c000000)
+# CHECK-NEXT:    JNE_1 %bb.8
+# CHECK:  bb.6:
+# CHECK-NEXT:    successors: %bb.1(0x04000000), %bb.5(0x7c000000)
 # CHECK:    $rax = MOV64rm $r14, 1, $noreg, 0, $noreg
 # CHECK-NEXT:    TEST64rr $rax, $rax
-# CHECK-NEXT:    JNE_1 %bb.7
+# CHECK-NEXT:    JNE_1 %bb.5
 
 name:            foo
 body:             |
Index: test/CodeGen/X86/tail-opts.ll
===================================================================
--- test/CodeGen/X86/tail-opts.ll
+++ test/CodeGen/X86/tail-opts.ll
@@ -435,9 +435,8 @@
   ret void
 
 bbx:
-  switch i32 undef, label %bb12 [
-    i32 128, label %return
-  ]
+  %1 = icmp eq i32 undef, 0
+  br i1 %1, label %return, label %bb12
 
 bb12:
   store volatile i32 0, i32* @XYZ
Index: test/CodeGen/X86/tail-threshold.ll
===================================================================
--- test/CodeGen/X86/tail-threshold.ll
+++ test/CodeGen/X86/tail-threshold.ll
@@ -5,7 +5,6 @@
 
 ; CHECK: 	callq	bar
 ; CHECK:	callq	bar
-; CHECK:	callq	bar
 ; CHECK-NOT:    callq
 
 declare void @bar()
Index: test/CodeGen/X86/test-shrink-bug.ll
===================================================================
--- test/CodeGen/X86/test-shrink-bug.ll
+++ test/CodeGen/X86/test-shrink-bug.ll
@@ -52,15 +52,13 @@
 ; CHECK-X86-NEXT:    cmpb $123, {{[0-9]+}}(%esp)
 ; CHECK-X86-NEXT:    sete %al
 ; CHECK-X86-NEXT:    testl $263, %ecx ## imm = 0x107
-; CHECK-X86-NEXT:    je LBB1_2
+; CHECK-X86-NEXT:    je LBB1_3
 ; CHECK-X86-NEXT:  ## %bb.1:
 ; CHECK-X86-NEXT:    testb %al, %al
-; CHECK-X86-NEXT:    jne LBB1_2
-; CHECK-X86-NEXT:  ## %bb.3: ## %no
+; CHECK-X86-NEXT:    jne LBB1_3
+; CHECK-X86-NEXT:  ## %bb.2: ## %no
 ; CHECK-X86-NEXT:    calll _bar
-; CHECK-X86-NEXT:    addl $12, %esp
-; CHECK-X86-NEXT:    retl
-; CHECK-X86-NEXT:  LBB1_2: ## %yes
+; CHECK-X86-NEXT:  LBB1_3: ## %yes
 ; CHECK-X86-NEXT:    addl $12, %esp
 ; CHECK-X86-NEXT:    retl
 ;
@@ -69,7 +67,7 @@
 ; CHECK-X64-NEXT:    pushq %rax
 ; CHECK-X64-NEXT:    .cfi_def_cfa_offset 16
 ; CHECK-X64-NEXT:    testw $263, %di # imm = 0x107
-; CHECK-X64-NEXT:    je .LBB1_2
+; CHECK-X64-NEXT:    je .LBB1_3
 ; CHECK-X64-NEXT:  # %bb.1:
 ; CHECK-X64-NEXT:    pand {{.*}}(%rip), %xmm0
 ; CHECK-X64-NEXT:    pcmpeqd {{.*}}(%rip), %xmm0
@@ -77,14 +75,10 @@
 ; CHECK-X64-NEXT:    pand %xmm0, %xmm1
 ; CHECK-X64-NEXT:    pextrw $4, %xmm1, %eax
 ; CHECK-X64-NEXT:    testb $1, %al
-; CHECK-X64-NEXT:    jne .LBB1_2
-; CHECK-X64-NEXT:  # %bb.3: # %no
+; CHECK-X64-NEXT:    jne .LBB1_3
+; CHECK-X64-NEXT:  # %bb.2: # %no
 ; CHECK-X64-NEXT:    callq bar
-; CHECK-X64-NEXT:    popq %rax
-; CHECK-X64-NEXT:    .cfi_def_cfa_offset 8
-; CHECK-X64-NEXT:    retq
-; CHECK-X64-NEXT:  .LBB1_2: # %yes
-; CHECK-X64-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-X64-NEXT:  .LBB1_3: # %yes
 ; CHECK-X64-NEXT:    popq %rax
 ; CHECK-X64-NEXT:    .cfi_def_cfa_offset 8
 ; CHECK-X64-NEXT:    retq
Index: test/Transforms/LoopStrengthReduce/AArch64/small-constant.ll
===================================================================
--- test/Transforms/LoopStrengthReduce/AArch64/small-constant.ll
+++ test/Transforms/LoopStrengthReduce/AArch64/small-constant.ll
@@ -19,25 +19,22 @@
 ; CHECK-LABEL: test1:
 ; CHECK:       // %bb.0: // %entry
 ; CHECK-NEXT:    fmov s2, #-7.00000000
-; CHECK-NEXT:    cbz x1, .LBB0_5
+; CHECK-NEXT:    cbz x1, .LBB0_4
 ; CHECK-NEXT:  // %bb.1: // %for.body.preheader
 ; CHECK-NEXT:    add x8, x0, #28 // =28
 ; CHECK-NEXT:  .LBB0_2: // %for.body
 ; CHECK-NEXT:    // =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    ldr s1, [x8, x1, lsl #2]
 ; CHECK-NEXT:    fcmp s1, s0
-; CHECK-NEXT:    b.gt .LBB0_6
+; CHECK-NEXT:    b.gt .LBB0_5
 ; CHECK-NEXT:  // %bb.3: // %for.cond
 ; CHECK-NEXT:    // in Loop: Header=BB0_2 Depth=1
 ; CHECK-NEXT:    add x1, x1, #1 // =1
 ; CHECK-NEXT:    cbnz x1, .LBB0_2
-; CHECK-NEXT:  // %bb.4:
+; CHECK-NEXT:  .LBB0_4:
 ; CHECK-NEXT:    mov v0.16b, v2.16b
 ; CHECK-NEXT:    ret
-; CHECK-NEXT:  .LBB0_5:
-; CHECK-NEXT:    mov v0.16b, v2.16b
-; CHECK-NEXT:    ret
-; CHECK-NEXT:  .LBB0_6: // %cleanup2
+; CHECK-NEXT:  .LBB0_5: // %cleanup2
 ; CHECK-NEXT:    mov v0.16b, v1.16b
 ; CHECK-NEXT:    ret
 entry:
@@ -68,7 +65,7 @@
 ; CHECK-LABEL: test2:
 ; CHECK:       // %bb.0: // %entry
 ; CHECK-NEXT:    fmov s2, #-7.00000000
-; CHECK-NEXT:    cbz x1, .LBB1_5
+; CHECK-NEXT:    cbz x1, .LBB1_4
 ; CHECK-NEXT:  // %bb.1: // %for.body.preheader
 ; CHECK-NEXT:    add x8, x0, #28 // =28
 ; CHECK-NEXT:  .LBB1_2: // %for.body
@@ -77,18 +74,15 @@
 ; CHECK-NEXT:    scvtf s3, x1
 ; CHECK-NEXT:    fadd s3, s3, s0
 ; CHECK-NEXT:    fcmp s1, s3
-; CHECK-NEXT:    b.gt .LBB1_6
+; CHECK-NEXT:    b.gt .LBB1_5
 ; CHECK-NEXT:  // %bb.3: // %for.cond
 ; CHECK-NEXT:    // in Loop: Header=BB1_2 Depth=1
 ; CHECK-NEXT:    add x1, x1, #1 // =1
 ; CHECK-NEXT:    cbnz x1, .LBB1_2
-; CHECK-NEXT:  // %bb.4:
-; CHECK-NEXT:    mov v0.16b, v2.16b
-; CHECK-NEXT:    ret
-; CHECK-NEXT:  .LBB1_5:
+; CHECK-NEXT:  .LBB1_4:
 ; CHECK-NEXT:    mov v0.16b, v2.16b
 ; CHECK-NEXT:    ret
-; CHECK-NEXT:  .LBB1_6: // %cleanup4
+; CHECK-NEXT:  .LBB1_5: // %cleanup4
 ; CHECK-NEXT:    mov v0.16b, v1.16b
 ; CHECK-NEXT:    ret
 entry: