Index: llvm/lib/CodeGen/RegisterCoalescer.cpp
===================================================================
--- llvm/lib/CodeGen/RegisterCoalescer.cpp
+++ llvm/lib/CodeGen/RegisterCoalescer.cpp
@@ -305,7 +305,11 @@
     /// number if it is not zero. If DstReg is a physical register and the
     /// existing subregister number of the def / use being updated is not zero,
     /// make sure to set it to the correct physical subregister.
-    void updateRegDefsUses(Register SrcReg, Register DstReg, unsigned SubIdx);
+    ///
+    /// If \p IsSubregToReg, we are coalescing a DstReg = SUBREG_TO_REG
+    /// SrcReg. This introduces an implicit-def of DstReg on coalesced users.
+    void updateRegDefsUses(Register SrcReg, Register DstReg, unsigned SubIdx,
+                           bool IsSubregToReg);
 
     /// If the given machine operand reads only undefined lanes add an undef
     /// flag.
@@ -1323,8 +1327,7 @@
     if (DstReg.isPhysical()) {
       Register NewDstReg = DstReg;
 
-      unsigned NewDstIdx = TRI->composeSubRegIndices(CP.getSrcIdx(),
-                                              DefMI->getOperand(0).getSubReg());
+      unsigned NewDstIdx = TRI->composeSubRegIndices(CP.getSrcIdx(), DefSubIdx);
       if (NewDstIdx)
         NewDstReg = TRI->getSubReg(DstReg, NewDstIdx);
 
@@ -1467,7 +1470,7 @@
     MRI->setRegClass(DstReg, NewRC);
 
     // Update machine operands and add flags.
-    updateRegDefsUses(DstReg, DstReg, DstIdx);
+    updateRegDefsUses(DstReg, DstReg, DstIdx, false);
     NewMI.getOperand(0).setSubReg(NewIdx);
     // updateRegDefUses can add an "undef" flag to the definition, since
     // it will replace DstReg with DstReg.DstIdx. If NewIdx is 0, make
@@ -1782,7 +1785,7 @@
 }
 
 void RegisterCoalescer::updateRegDefsUses(Register SrcReg, Register DstReg,
-                                          unsigned SubIdx) {
+                                          unsigned SubIdx, bool IsSubregToReg) {
   bool DstIsPhys = DstReg.isPhysical();
   LiveInterval *DstInt = DstIsPhys ? nullptr : &LIS->getInterval(DstReg);
 
@@ -1822,6 +1825,8 @@
     if (DstInt && !Reads && SubIdx && !UseMI->isDebugInstr())
       Reads = DstInt->liveAt(LIS->getInstructionIndex(*UseMI));
 
+    bool FullDef = true;
+
     // Replace SrcReg with DstReg in all UseMI operands.
     for (unsigned i = 0, e = Ops.size(); i != e; ++i) {
       MachineOperand &MO = UseMI->getOperand(Ops[i]);
@@ -1829,9 +1834,13 @@
       // Adjust <undef> flags in case of sub-register joins. We don't want to
       // turn a full def into a read-modify-write sub-register def and vice
       // versa.
-      if (SubIdx && MO.isDef())
+      if (SubIdx && MO.isDef()) {
         MO.setIsUndef(!Reads);
 
+        if (!Reads)
+          FullDef = false;
+      }
+
       // A subreg use of a partially undef (super) register may be a complete
       // undef use now and then has to be marked that way.
       if (MO.isUse() && !DstIsPhys) {
@@ -1863,6 +1872,25 @@
         MO.substVirtReg(DstReg, SubIdx, *TRI);
     }
 
+    if (IsSubregToReg && !FullDef) {
+      // If the coalesed instruction doesn't fully define the register, we need
+      // to preserve the original super register liveness for SUBREG_TO_REG.
+      //
+      // We pretended SUBREG_TO_REG was a regular copy for coalescing purposes,
+      // but it introduces liveness for other subregisters. Downstream users may
+      // have been relying on those bits, so we need to ensure their liveness is
+      // captured with a def of other lanes.
+
+      // FIXME: Need to add new subrange if tracking subranges. We could also
+      // skip adding this if we knew the other lanes are dead, and only for
+      // other lanes.
+
+      assert(!MRI->shouldTrackSubRegLiveness(DstReg) &&
+             "this should update subranges");
+      MachineInstrBuilder MIB(*MF, UseMI);
+      MIB.addReg(DstReg, RegState::ImplicitDefine);
+    }
+
     LLVM_DEBUG({
       dbgs() << "\t\tupdated: ";
       if (!UseMI->isDebugInstr())
@@ -2062,6 +2090,8 @@
     });
   }
 
+  const bool IsSubregToReg = CopyMI->isSubregToReg();
+
   ShrinkMask = LaneBitmask::getNone();
   ShrinkMainRange = false;
 
@@ -2129,9 +2159,12 @@
 
   // Rewrite all SrcReg operands to DstReg.
   // Also update DstReg operands to include DstIdx if it is set.
-  if (CP.getDstIdx())
-    updateRegDefsUses(CP.getDstReg(), CP.getDstReg(), CP.getDstIdx());
-  updateRegDefsUses(CP.getSrcReg(), CP.getDstReg(), CP.getSrcIdx());
+  if (CP.getDstIdx()) {
+    assert(!IsSubregToReg && "can this happen?");
+    updateRegDefsUses(CP.getDstReg(), CP.getDstReg(), CP.getDstIdx(), false);
+  }
+  updateRegDefsUses(CP.getSrcReg(), CP.getDstReg(), CP.getSrcIdx(),
+                    IsSubregToReg);
 
   // Shrink subregister ranges if necessary.
   if (ShrinkMask.any()) {
Index: llvm/test/CodeGen/X86/bswap.ll
===================================================================
--- llvm/test/CodeGen/X86/bswap.ll
+++ llvm/test/CodeGen/X86/bswap.ll
@@ -226,7 +226,7 @@
 ; CHECK64:       # %bb.0:
 ; CHECK64-NEXT:    movzwl var16(%rip), %ecx
 ; CHECK64-NEXT:    movzbl %cl, %eax
-; CHECK64-NEXT:    # kill: def $ecx killed $ecx killed $rcx def $rcx
+; CHECK64-NEXT:    # kill: def $ecx killed $ecx def $rcx killed $rcx
 ; CHECK64-NEXT:    shrl $8, %ecx
 ; CHECK64-NEXT:    shlq $8, %rax
 ; CHECK64-NEXT:    orq %rcx, %rax
Index: llvm/test/CodeGen/X86/coalescer-breaks-subreg-to-reg-liveness.ll
===================================================================
--- /dev/null
+++ llvm/test/CodeGen/X86/coalescer-breaks-subreg-to-reg-liveness.ll
@@ -0,0 +1,182 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
+; RUN: llc -mtriple=x86_64-grtev4-linux-gnu < %s | FileCheck %s
+
+%struct.wibble = type { %struct.wombat }
+%struct.wombat = type { %struct.ham, [3 x i8] }
+%struct.ham = type { %struct.zot }
+%struct.zot = type { %struct.blam }
+%struct.blam = type { %struct.ham.0 }
+%struct.ham.0 = type { %struct.bar }
+%struct.bar = type { %struct.bar.1 }
+%struct.bar.1 = type { %struct.baz, i8 }
+%struct.baz = type { %struct.snork }
+%struct.snork = type <{ %struct.spam, i8, [3 x i8] }>
+%struct.spam = type { %struct.snork.2, %struct.snork.2 }
+%struct.snork.2 = type { i32 }
+%struct.snork.3 = type { %struct.baz, i8, [3 x i8] }
+
+define void @foo(ptr %arg, ptr %arg1, i40 %arg2, ptr %arg3, i32 %arg4) #0 {
+; CHECK-LABEL: foo:
+; CHECK:       # %bb.0: # %bb
+; CHECK-NEXT:    pushq %rbp
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    .cfi_offset %rbp, -16
+; CHECK-NEXT:    movq %rsp, %rbp
+; CHECK-NEXT:    .cfi_def_cfa_register %rbp
+; CHECK-NEXT:    pushq %r15
+; CHECK-NEXT:    pushq %r14
+; CHECK-NEXT:    pushq %r13
+; CHECK-NEXT:    pushq %r12
+; CHECK-NEXT:    pushq %rbx
+; CHECK-NEXT:    subq $24, %rsp
+; CHECK-NEXT:    .cfi_offset %rbx, -56
+; CHECK-NEXT:    .cfi_offset %r12, -48
+; CHECK-NEXT:    .cfi_offset %r13, -40
+; CHECK-NEXT:    .cfi_offset %r14, -32
+; CHECK-NEXT:    .cfi_offset %r15, -24
+; CHECK-NEXT:    movl %r8d, %r14d
+; CHECK-NEXT:    movq %rcx, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
+; CHECK-NEXT:    movq %rdx, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
+; CHECK-NEXT:    movq %rsi, %r12
+; CHECK-NEXT:    movq %rdi, %r15
+; CHECK-NEXT:    incl %r14d
+; CHECK-NEXT:    xorl %ebx, %ebx
+; CHECK-NEXT:    # implicit-def: $rax
+; CHECK-NEXT:    # kill: killed $rax
+; CHECK-NEXT:    jmp .LBB0_3
+; CHECK-NEXT:    .p2align 4, 0x90
+; CHECK-NEXT:  .LBB0_1: # %bb17
+; CHECK-NEXT:    # in Loop: Header=BB0_3 Depth=1
+; CHECK-NEXT:    movq %r15, %r13
+; CHECK-NEXT:    xorl %r15d, %r15d
+; CHECK-NEXT:    testq %rbx, %rbx
+; CHECK-NEXT:    sete %r15b
+; CHECK-NEXT:    xorl %edi, %edi
+; CHECK-NEXT:    callq _Znwm@PLT
+; CHECK-NEXT:    shlq $4, %r15
+; CHECK-NEXT:    addq {{[-0-9]+}}(%r{{[sb]}}p), %r15 # 8-byte Folded Reload
+; CHECK-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rdx # 8-byte Reload
+; CHECK-NEXT:    movq %rdx, %rcx
+; CHECK-NEXT:    shrq $32, %rcx
+; CHECK-NEXT:    movb %cl, 12(%rax)
+; CHECK-NEXT:    movl %edx, 8(%rax)
+; CHECK-NEXT:    movq %r15, %rbx
+; CHECK-NEXT:    movq %r13, %r15
+; CHECK-NEXT:    decl %r14d
+; CHECK-NEXT:    je .LBB0_7
+; CHECK-NEXT:  .LBB0_3: # %bb7
+; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
+; CHECK-NEXT:    callq widget@PLT
+; CHECK-NEXT:    cmpb $-5, (%r12)
+; CHECK-NEXT:    jb .LBB0_5
+; CHECK-NEXT:  # %bb.4: # %bb12
+; CHECK-NEXT:    # in Loop: Header=BB0_3 Depth=1
+; CHECK-NEXT:    movq 0, %rax
+; CHECK-NEXT:    movq 8, %rax
+; CHECK-NEXT:    movq {{[-0-9]+}}(%r{{[sb]}}p), %rax # 8-byte Reload
+; CHECK-NEXT:    movq %rax, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
+; CHECK-NEXT:  .LBB0_5: # %bb14
+; CHECK-NEXT:    # in Loop: Header=BB0_3 Depth=1
+; CHECK-NEXT:    cmpq %r15, %rbx
+; CHECK-NEXT:    jbe .LBB0_1
+; CHECK-NEXT:  # %bb.6: # in Loop: Header=BB0_3 Depth=1
+; CHECK-NEXT:    xorl %eax, %eax
+; CHECK-NEXT:    xorl %ebx, %ebx
+; CHECK-NEXT:    decl %r14d
+; CHECK-NEXT:    jne .LBB0_3
+; CHECK-NEXT:  .LBB0_7: # %bb21
+; CHECK-NEXT:    cmpb $0, 12(%rax)
+; CHECK-NEXT:    jne .LBB0_9
+; CHECK-NEXT:  # %bb.8: # %bb26
+; CHECK-NEXT:    addq $24, %rsp
+; CHECK-NEXT:    popq %rbx
+; CHECK-NEXT:    popq %r12
+; CHECK-NEXT:    popq %r13
+; CHECK-NEXT:    popq %r14
+; CHECK-NEXT:    popq %r15
+; CHECK-NEXT:    popq %rbp
+; CHECK-NEXT:    .cfi_def_cfa %rsp, 8
+; CHECK-NEXT:    retq
+; CHECK-NEXT:  .LBB0_9: # %bb25
+; CHECK-NEXT:    .cfi_def_cfa %rbp, 16
+; CHECK-NEXT:    movq %r15, %rdi
+; CHECK-NEXT:    callq pluto@PLT
+bb:
+  br label %bb7
+
+bb5:                                              ; preds = %bb17, %bb14
+  %phi = phi ptr [ %call19, %bb17 ], [ null, %bb14 ]
+  %phi6 = phi ptr [ %getelementptr, %bb17 ], [ null, %bb14 ]
+  %add = add i32 %phi9, 1
+  %icmp = icmp eq i32 %phi9, %arg4
+  br i1 %icmp, label %bb21, label %bb7
+
+bb7:                                              ; preds = %bb5, %bb
+  %phi8 = phi ptr [ null, %bb ], [ %phi6, %bb5 ]
+  %phi9 = phi i32 [ 0, %bb ], [ %add, %bb5 ]
+  %phi10 = phi i40 [ undef, %bb ], [ %phi15, %bb5 ]
+  %call = call ptr @widget()
+  %load = load i8, ptr %arg1, align 8
+  %icmp11 = icmp ult i8 %load, -5
+  %and = and i40 %phi10, 4294967295
+  br i1 %icmp11, label %bb14, label %bb12
+
+bb12:                                             ; preds = %bb7
+  %load13 = load volatile { i64, i64 }, ptr null, align 4294967296
+  br label %bb14
+
+bb14:                                             ; preds = %bb12, %bb7
+  %phi15 = phi i40 [ %and, %bb7 ], [ %arg2, %bb12 ]
+  %icmp16 = icmp ugt ptr %phi8, %arg
+  br i1 %icmp16, label %bb5, label %bb17
+
+bb17:                                             ; preds = %bb14
+  %icmp18 = icmp eq ptr %phi8, null
+  %zext = zext i1 %icmp18 to i64
+  %call19 = call ptr @_Znwm(i64 0)
+  %getelementptr = getelementptr %struct.wibble, ptr %arg3, i64 %zext
+  %getelementptr20 = getelementptr i8, ptr %call19, i64 8
+  store i40 %phi15, ptr %getelementptr20, align 4
+  br label %bb5
+
+bb21:                                             ; preds = %bb5
+  %getelementptr22 = getelementptr %struct.snork.3, ptr %phi, i64 0, i32 1
+  %load23 = load i8, ptr %getelementptr22, align 4
+  %icmp24 = icmp eq i8 %load23, 0
+  br i1 %icmp24, label %bb26, label %bb25
+
+bb25:                                             ; preds = %bb21
+  call void @pluto(ptr %arg)
+  unreachable
+
+bb26:                                             ; preds = %bb21
+  ret void
+}
+
+define void @eggs(ptr %arg, ptr %arg1) {
+; CHECK-LABEL: eggs:
+; CHECK:       # %bb.0: # %bb
+; CHECK-NEXT:    pushq %rax
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    movq %rdi, %rax
+; CHECK-NEXT:    movq %rsi, %rdi
+; CHECK-NEXT:    movq %rax, %rsi
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    xorl %ecx, %ecx
+; CHECK-NEXT:    xorl %r8d, %r8d
+; CHECK-NEXT:    callq foo@PLT
+; CHECK-NEXT:    popq %rax
+; CHECK-NEXT:    .cfi_def_cfa_offset 8
+; CHECK-NEXT:    retq
+bb:
+  call void @foo(ptr %arg1, ptr %arg, i40 0, ptr null, i32 0)
+  ret void
+}
+
+declare ptr @widget()
+
+declare void @pluto(ptr)
+
+declare ptr @_Znwm(i64)
+
+attributes #0 = { noinline "frame-pointer"="all" }
Index: llvm/test/CodeGen/X86/coalescing-subreg-to-reg-requires-subrange-update.mir
===================================================================
--- /dev/null
+++ llvm/test/CodeGen/X86/coalescing-subreg-to-reg-requires-subrange-update.mir
@@ -0,0 +1,47 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 3
+# RUN: llc -mtriple=x86_64-- -run-pass=register-coalescer -enable-subreg-liveness -verify-coalescing -o - %s | FileCheck %s
+
+
+# FIXME: Need to handle subrange updates when coalescing with subreg_to_reg
+# This will fail if x86 enables subregister liveness.
+---
+name: requires_new_subrange_coalesce_subreg_to_reg
+tracksRegLiveness: true
+body:             |
+  ; CHECK-LABEL: name: requires_new_subrange_coalesce_subreg_to_reg
+  ; CHECK: bb.0:
+  ; CHECK-NEXT:   successors: %bb.2(0x40000000), %bb.1(0x40000000)
+  ; CHECK-NEXT:   liveins: $eax
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   undef %a.sub_32bit:gr64_with_sub_8bit = COPY $eax
+  ; CHECK-NEXT:   %b:gr32 = IMPLICIT_DEF
+  ; CHECK-NEXT:   %c:gr64 = INSERT_SUBREG %a, %b, %subreg.sub_32bit
+  ; CHECK-NEXT:   JCC_1 %bb.2, 4, implicit undef $eflags
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT: bb.1:
+  ; CHECK-NEXT:   successors: %bb.2(0x80000000)
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   undef %a.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags
+  ; CHECK-NEXT:   %c.sub_32bit:gr64 = COPY %a
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT: bb.2:
+  ; CHECK-NEXT:   %c.sub_32bit:gr64 = SUBREG_TO_REG %a, %b, %subreg.sub_32bit
+  ; CHECK-NEXT:   RET 0, implicit %c
+  bb.0:
+    liveins: $eax
+    %init_eax:gr32 = COPY $eax
+    %a:gr64 = SUBREG_TO_REG 0, %init_eax, %subreg.sub_32bit
+    %b:gr32 = IMPLICIT_DEF
+    %c:gr64 = INSERT_SUBREG %a, %b, %subreg.sub_32bit
+    JCC_1 %bb.2, 4, implicit undef $eflags
+
+  bb.1:
+    %imm0:gr32 = MOV32r0 implicit-def dead $eflags
+    %a = SUBREG_TO_REG 0, %imm0, %subreg.sub_32bit
+    %c.sub_32bit = COPY %a
+
+  bb.2:
+    %c.sub_32bit = SUBREG_TO_REG %a, %b, %subreg.sub_32bit
+    RET 0, implicit %c
+
+...
Index: llvm/test/CodeGen/X86/subreg-to-reg-coalescing.mir
===================================================================
--- /dev/null
+++ llvm/test/CodeGen/X86/subreg-to-reg-coalescing.mir
@@ -0,0 +1,348 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 2
+# RUN: llc -mtriple=x86_64-- -run-pass=register-coalescer -o - %s | FileCheck %s
+
+# We cannot lose the liveness of the high subregister of %1 when
+# coalesced with %0, so introduce an implicit-def of the super
+# register on the MOV.
+
+---
+name: coalesce_mov32r0_into_subreg_to_reg64
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    ; CHECK-LABEL: name: coalesce_mov32r0_into_subreg_to_reg64
+    ; CHECK: ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: undef %1.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags, implicit-def %1
+    ; CHECK-NEXT: dead $edi = MOV32r0 implicit-def dead $eflags, implicit-def $rdi
+    ; CHECK-NEXT: CALL64r %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ; CHECK-NEXT: ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: RET 0
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    %0:gr32 = MOV32r0 implicit-def dead $eflags
+    %1:gr64 = SUBREG_TO_REG 0, killed %0, %subreg.sub_32bit
+    $rdi = COPY %1
+    CALL64r killed %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    RET 0
+
+...
+
+---
+name: subreg_to_reg_folds_to_undef
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $rax
+
+    ; CHECK-LABEL: name: subreg_to_reg_folds_to_undef
+    ; CHECK: liveins: $rax
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:gr64_with_sub_8bit = COPY $rax
+    ; CHECK-NEXT: undef %4.sub_32bit:gr64_with_sub_8bit = MOV32rr [[COPY]].sub_32bit, implicit-def %4
+    ; CHECK-NEXT: RET 0, implicit %4
+    %0:gr64 = COPY killed $rax
+    %1:gr32 = COPY killed %0.sub_32bit
+    %2:gr32 = MOV32rr killed %1
+    %3:gr64 = SUBREG_TO_REG 0, killed %2, %subreg.sub_32bit
+    %4:gr64 = COPY killed %3
+    RET 0, implicit %4
+
+...
+
+---
+name: coalesce_mov32r0_subreg_def_into_subreg_to_reg64
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    ; CHECK-LABEL: name: coalesce_mov32r0_subreg_def_into_subreg_to_reg64
+    ; CHECK: ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: undef %1.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags
+    ; CHECK-NEXT: dead $edi = MOV32r0 implicit-def dead $eflags, implicit-def $rdi
+    ; CHECK-NEXT: CALL64r %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ; CHECK-NEXT: ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: RET 0
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    undef %0.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags
+    %1:gr64 = SUBREG_TO_REG 0, killed %0.sub_32bit, %subreg.sub_32bit
+    $rdi = COPY %1
+    CALL64r killed %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    RET 0
+
+...
+
+---
+name: coalesce_mov32r0_into_subreg_def_with_super_def_to_reg64
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    ; CHECK-LABEL: name: coalesce_mov32r0_into_subreg_def_with_super_def_to_reg64
+    ; CHECK: ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: undef %1.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags, implicit-def %1
+    ; CHECK-NEXT: dead $edi = MOV32r0 implicit-def dead $eflags, implicit-def $rdi
+    ; CHECK-NEXT: CALL64r %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ; CHECK-NEXT: ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: RET 0
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    undef %0.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags, implicit-def %0
+    %1:gr64 = SUBREG_TO_REG 0, killed %0.sub_32bit, %subreg.sub_32bit
+    $rdi = COPY %1
+    CALL64r killed %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    RET 0
+
+...
+
+---
+name: coalesce_mov32r0_into_subreg_to_reg64_already_defs_other_subreg
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    ; CHECK-LABEL: name: coalesce_mov32r0_into_subreg_to_reg64_already_defs_other_subreg
+    ; CHECK: ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: undef %1.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags, implicit-def undef %1.sub_8bit, implicit-def %1
+    ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, implicit %1
+    ; CHECK-NEXT: CALL64r %1, csr_64, implicit $rsp, implicit $ssp, implicit undef $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ; CHECK-NEXT: ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: RET 0
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    %0:gr32 = MOV32r0 implicit-def dead $eflags, implicit-def undef %0.sub_8bit
+    %1:gr64 = SUBREG_TO_REG 0, killed %0, %subreg.sub_32bit
+    INLINEASM &"", 0, implicit %1
+    CALL64r killed %1, csr_64, implicit $rsp, implicit $ssp, implicit undef $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    RET 0
+
+...
+
+
+# Reduced realistic case which was asserting after introducing new implicit-defs
+---
+name: coalesce_needs_implicit_defs
+tracksRegLiveness: true
+body:             |
+  ; CHECK-LABEL: name: coalesce_needs_implicit_defs
+  ; CHECK: bb.0:
+  ; CHECK-NEXT:   successors: %bb.1(0x80000000)
+  ; CHECK-NEXT:   liveins: $rdi
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   [[COPY:%[0-9]+]]:gr64 = COPY $rdi
+  ; CHECK-NEXT:   undef %2.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags, implicit-def %2
+  ; CHECK-NEXT:   undef %3.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags, implicit-def %3
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT: bb.1:
+  ; CHECK-NEXT:   successors: %bb.1(0x80000000)
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   undef %10.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags
+  ; CHECK-NEXT:   TEST64rr %3, %3, implicit-def $eflags
+  ; CHECK-NEXT:   %10.sub_8bit:gr64_with_sub_8bit = SETCCr 4, implicit killed $eflags
+  ; CHECK-NEXT:   ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+  ; CHECK-NEXT:   dead $edi = MOV32r0 implicit-def dead $eflags, implicit-def $rdi
+  ; CHECK-NEXT:   CALL64r %2, csr_64, implicit $rsp, implicit $ssp, implicit $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+  ; CHECK-NEXT:   ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+  ; CHECK-NEXT:   [[SHL64ri:%[0-9]+]]:gr64_with_sub_8bit = SHL64ri [[SHL64ri]], 4, implicit-def dead $eflags
+  ; CHECK-NEXT:   [[ADD64rr:%[0-9]+]]:gr64_with_sub_8bit = ADD64rr [[ADD64rr]], [[COPY]], implicit-def dead $eflags
+  ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:gr64_with_sub_8bit = COPY [[ADD64rr]]
+  ; CHECK-NEXT:   JMP_1 %bb.1
+  bb.0:
+    liveins: $rdi
+
+    %0:gr64 = COPY killed $rdi
+    %1:gr32 = MOV32r0 implicit-def dead $eflags
+    %2:gr64 = SUBREG_TO_REG 0, %1, %subreg.sub_32bit
+    %3:gr64 = COPY killed %2
+
+  bb.1:
+    %4:gr64 = COPY killed %3
+    %5:gr32 = MOV32r0 implicit-def dead $eflags
+    TEST64rr killed %4, %4, implicit-def $eflags
+    %6:gr8 = SETCCr 4, implicit killed $eflags
+    %7:gr32 = COPY killed %5
+    %7.sub_8bit:gr32 = COPY killed %6
+    %8:gr64 = SUBREG_TO_REG 0, killed %7, %subreg.sub_32bit
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    %9:gr64 = SUBREG_TO_REG 0, %1, %subreg.sub_32bit
+    $rdi = COPY %9
+    CALL64r killed %9, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    %10:gr64 = COPY killed %8
+    %10:gr64 = SHL64ri %10, 4, implicit-def dead $eflags
+    %11:gr64 = COPY killed %10
+    %11:gr64 = ADD64rr %11, %0, implicit-def dead $eflags
+    %3:gr64 = COPY killed %11
+    JMP_1 %bb.1
+
+...
+
+---
+name: coalesce_mov32r0_into_subreg_to_reg64_physreg_def
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    ; CHECK-LABEL: name: coalesce_mov32r0_into_subreg_to_reg64_physreg_def
+    ; CHECK: ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: dead $edi = MOV32r0 implicit-def dead $eflags, implicit-def $rdi
+    ; CHECK-NEXT: CALL64r killed $rdi, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ; CHECK-NEXT: ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: RET 0
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    %0:gr32 = MOV32r0 implicit-def dead $eflags
+    $rdi = SUBREG_TO_REG 0, killed %0, %subreg.sub_32bit
+    CALL64r killed $rdi, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    RET 0
+
+...
+
+---
+name: coalesce_mov32r0_into_subreg_to_reg64_physreg_use
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $eax
+    ; CHECK-LABEL: name: coalesce_mov32r0_into_subreg_to_reg64_physreg_use
+    ; CHECK: liveins: $eax
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: $eax = MOV32r0 implicit-def dead $eflags
+    ; CHECK-NEXT: [[SUBREG_TO_REG:%[0-9]+]]:gr64 = SUBREG_TO_REG 0, $eax, %subreg.sub_32bit
+    ; CHECK-NEXT: $rdi = COPY [[SUBREG_TO_REG]]
+    ; CHECK-NEXT: CALL64r [[SUBREG_TO_REG]], csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ; CHECK-NEXT: ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: RET 0
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    $eax = MOV32r0 implicit-def dead $eflags
+    %1:gr64 = SUBREG_TO_REG 0, killed $eax, %subreg.sub_32bit
+    $rdi = COPY %1
+    CALL64r killed %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    RET 0
+
+...
+
+# Coalesced instruction is a copy with other implicit operands
+---
+name: coalesce_copy_into_subreg_to_reg64
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    liveins: $eax
+    ; CHECK-LABEL: name: coalesce_copy_into_subreg_to_reg64
+    ; CHECK: liveins: $eax
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: undef %1.sub_32bit:gr64_with_sub_8bit = COPY $eax, implicit-def dead $eflags, implicit-def %1
+    ; CHECK-NEXT: $rdi = COPY %1
+    ; CHECK-NEXT: CALL64r %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ; CHECK-NEXT: ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: RET 0
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    %0:gr32 = COPY $eax, implicit-def dead $eflags
+    %1:gr64 = SUBREG_TO_REG 0, killed %0, %subreg.sub_32bit
+    $rdi = COPY %1
+    CALL64r killed %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    RET 0
+
+...
+
+---
+name: coalesce_mov32r0_into_subreg_to_reg64_multiple_redef_value
+tracksRegLiveness: true
+body:             |
+  bb.0:
+    ; CHECK-LABEL: name: coalesce_mov32r0_into_subreg_to_reg64_multiple_redef_value
+    ; CHECK: ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: undef %1.sub_32bit:gr64_with_sub_8bit = MOV32r0 implicit-def dead $eflags, implicit-def %1
+    ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, implicit-def %1.sub_32bit, implicit %1.sub_32bit
+    ; CHECK-NEXT: $rdi = COPY %1
+    ; CHECK-NEXT: CALL64r %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ; CHECK-NEXT: ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    ; CHECK-NEXT: RET 0
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    %0:gr32 = MOV32r0 implicit-def dead $eflags
+    INLINEASM &"", 0, implicit-def %0, implicit %0
+    %1:gr64 = SUBREG_TO_REG 0, killed %0, %subreg.sub_32bit
+    $rdi = COPY %1
+    CALL64r killed %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    RET 0
+
+...
+
+---
+name: coalesce_mov32r0_into_subreg_to_reg64_def_is_block_liveout
+tracksRegLiveness: true
+body:             |
+  ; CHECK-LABEL: name: coalesce_mov32r0_into_subreg_to_reg64_def_is_block_liveout
+  ; CHECK: bb.0:
+  ; CHECK-NEXT:   successors: %bb.1(0x40000000), %bb.2(0x40000000)
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   INLINEASM &"", 0 /* attdialect */, implicit-def undef %1.sub_32bit, implicit-def %1
+  ; CHECK-NEXT:   JCC_1 %bb.1, 4, implicit undef $eflags
+  ; CHECK-NEXT:   JMP_1 %bb.2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT: bb.1:
+  ; CHECK-NEXT:   $rdi = COPY %1
+  ; CHECK-NEXT:   ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+  ; CHECK-NEXT:   CALL64r %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+  ; CHECK-NEXT:   ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+  ; CHECK-NEXT:   RET 0
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT: bb.2:
+  bb.0:
+    INLINEASM &"", 0, implicit-def %0:gr32
+    JCC_1 %bb.1, 4, implicit undef $eflags
+    JMP_1 %bb.2
+
+  bb.1:
+    %1:gr64 = SUBREG_TO_REG 0, killed %0, %subreg.sub_32bit
+    $rdi = COPY %1
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    CALL64r killed %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    RET 0
+
+  bb.2:
+
+...
+
+---
+name: coalesce_mov32r0_into_subreg_to_reg64_def_is_phi_def
+tracksRegLiveness: true
+body:             |
+  ; CHECK-LABEL: name: coalesce_mov32r0_into_subreg_to_reg64_def_is_phi_def
+  ; CHECK: bb.0:
+  ; CHECK-NEXT:   successors: %bb.1(0x40000000), %bb.2(0x40000000)
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   INLINEASM &"", 0 /* attdialect */, implicit-def undef %1.sub_32bit, implicit-def %1
+  ; CHECK-NEXT:   JCC_1 %bb.1, 4, implicit undef $eflags
+  ; CHECK-NEXT:   JMP_1 %bb.2
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT: bb.1:
+  ; CHECK-NEXT:   successors: %bb.1(0x80000000)
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT:   $rdi = COPY %1
+  ; CHECK-NEXT:   ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+  ; CHECK-NEXT:   CALL64r %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+  ; CHECK-NEXT:   ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+  ; CHECK-NEXT:   JMP_1 %bb.1
+  ; CHECK-NEXT: {{  $}}
+  ; CHECK-NEXT: bb.2:
+  bb.0:
+
+    INLINEASM &"", 0, implicit-def %0:gr32
+    JCC_1 %bb.1, 4, implicit undef $eflags
+    JMP_1 %bb.2
+
+  bb.1:
+    %1:gr64 = SUBREG_TO_REG 0, %0, %subreg.sub_32bit
+    $rdi = COPY %1
+    ADJCALLSTACKDOWN64 0, 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    CALL64r %1, csr_64, implicit $rsp, implicit $ssp, implicit killed $rdi, implicit-def $rsp, implicit-def $ssp, implicit-def dead $rax
+    ADJCALLSTACKUP64 0, 0, implicit-def dead $rsp, implicit-def dead $eflags, implicit-def dead $ssp, implicit $rsp, implicit $ssp
+    JMP_1 %bb.1
+
+  bb.2:
+
+...