[AMDGPU] Add True16 register classes.

[foad]

llvm/lib/Target/AMDGPU/SIRegisterInfo.td

593–594

Do VGPR_LO16 and VGPR_HI16 need to define BaseClassOrder? From the comments below it sounds like the base class of a 16-bit VGPR will always be VGPR_16_Lo128 or VGPR_16.

1033

Use Reg16Types.types here and below?

[kosarev]

Removed extra BaseClassOrder assignments and use Reg16Types.types instead of [i16, f16].

[kosarev]

llvm/lib/Target/AMDGPU/SIRegisterInfo.td

593–594

Indeed these are now base classes to no registers. Nice catch, thanks.

[rampitec]

llvm/lib/Target/AMDGPU/SIRegisterInfo.td

1034

Does this really work when you have an RC with registers of different size?

[kosarev]

llvm/lib/Target/AMDGPU/SIRegisterInfo.td

1034

Does this really work when you have an RC with registers of different size?

It seems so, yes. Downstream I see us generating instructions like v_and_b16 v0.l, s2, v0.l without problem.

[rampitec]

LGTM

llvm/lib/Target/AMDGPU/SIRegisterInfo.td

1034

OK, thanks! I used to have problems with this in the past, thus the question.

[Joe_Nash]

llvm/lib/Target/AMDGPU/SIRegisterInfo.td

602

I think the comment downstream about which instruction types use which types of Registers is helpful.

llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir

31

Unrelated to this patch, but I don't understand the purpose of these tests. It looks like we are referencing register classes by number in the test input. Therefore any time we modify the number of register classes the tests changes their meaning.

[arsenm]

llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir

31

The raw number thing is a syntactical inconvenience from how inlineasm is represented. Ideally the mir printer/parser would produce something nicer. I think the purpose of the test is reasonably clear from the comment at the top

[Joe_Nash]

llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir

31

How can one determine if the change in output is still correct? The different RC choices in coalescer-early-clobber-subreg.mir are even more abstract. Is the RC to be ignored?

[arsenm]

llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir

31

the printed class name should be correct, you just get the added bonus random number. Fur the purpose of this test the class doesn't really matter. However there are no 16-bit values in this test, so I don't understand how this ended up with any vgpr16 references in the first place

[kosarev]

Added comments for the VGPR_16 and VGPR_16_Lo128 definitions.

[kosarev]

llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir

31

The InlineAsm instruction uses the numeric register class ids when computing these flags for register operands. These ids may change as we add or remove register classes. D78088 made the meaning of these flags visible in MIR tests, but there is still no support on the parser side for such operands.

[Joe_Nash]

llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir

31

These ids may change as we add or remove register classes.

Exactly.
The IDs in the input have not been updated but the output has been auto-updated about 10 times when someone previously added/removed a register class. Originally I think the register class was VRegOrLds_32. That is why we see 16-bit register classes in a test without 16-bit values. IMO this test format is not maintainable.

[Joe_Nash]

LGTM

[kosarev]

Rebased.

llvm/lib/Target/AMDGPU/SIRegisterInfo.td

 def Reg16Types : RegisterTypes<[i16, f16]>;
 def Reg32Types : RegisterTypes<[i32, f32, v2i16, v2f16, p2, p3, p5, p6]>;
 
 let HasVGPR = 1 in {
 def VGPR_LO16 : SIRegisterClass<"AMDGPU", Reg16Types.types, 16,
                               (add (sequence "VGPR%u_LO16", 0, 255))> {
   let AllocationPriority = 0;
   let Size = 16;
   let GeneratePressureSet = 0;
-  let BaseClassOrder = 16;
 }
 
 def VGPR_HI16 : SIRegisterClass<"AMDGPU", Reg16Types.types, 16,
                               (add (sequence "VGPR%u_HI16", 0, 255))> {
   let AllocationPriority = 0;
   let Size = 16;
   let GeneratePressureSet = 0;
+}
+
+def VGPR_16 : SIRegisterClass<"AMDGPU",  Reg16Types.types, 16,
+                            (add (interleave (sequence "VGPR%u_LO16", 0, 255),
+                                             (sequence "VGPR%u_HI16", 0, 255)))> {
+  let AllocationPriority = 2;
+  let Size = 16;
+  let GeneratePressureSet = 0;
+
+  // This is the base class for VGPR{128..255}_{LO16,HI16}.
   let BaseClassOrder = 17;
 }
 
+// The order of registers in the class determines order of allocation, so it is
+// important to interleave lo and hi registers.
+def VGPR_16_Lo128 : SIRegisterClass<"AMDGPU",  Reg16Types.types, 16,
+                            (add (interleave (sequence "VGPR%u_LO16", 0, 127),
+                                             (sequence "VGPR%u_HI16", 0, 127)))> {
+  let Size = 16;
+  let GeneratePressureSet = 0;
+  let isAllocatable = 0;
+
+  // This is the base class for VGPR{0..127}_{LO16,HI16}.
+  let BaseClassOrder = 16;
+}
+
 // VGPR 32-bit registers
 // i16/f16 only on VI+
 def VGPR_32 : SIRegisterClass<"AMDGPU", !listconcat(Reg32Types.types, Reg16Types.types), 32,
                             (add (sequence "VGPR%u", 0, 255))> {
   let AllocationPriority = 0;
   let Size = 32;
   let Weight = 1;
   let BaseClassOrder = 32;

 // SelectionDAG that should always be lowered by SILowerI1Copies.  TableGen
 // sorts register classes based on the number of registers in them so this is
 // sorted to the end and not preferred over VGPR_32.
 def VReg_1 : SIRegisterClass<"AMDGPU", [i1], 32, (add)> {
   let Size = 1;
   let HasVGPR = 1;
 }
 
+def VS_16 : SIRegisterClass<"AMDGPU", Reg16Types.types, 16,
+                          (add VGPR_16, SReg_32, LDS_DIRECT_CLASS)> {
+  let isAllocatable = 0;
+  let HasVGPR = 1;
+}
+
+def VS_16_Lo128 : SIRegisterClass<"AMDGPU", Reg16Types.types, 16,
+                          (add VGPR_16_Lo128, SReg_32, LDS_DIRECT_CLASS)> {
+  let isAllocatable = 0;
+  let HasVGPR = 1;
+}
+
 def VS_32 : SIRegisterClass<"AMDGPU", [i32, f32, i16, f16, v2i16, v2f16], 32,
                           (add VGPR_32, SReg_32, LDS_DIRECT_CLASS)> {
   let isAllocatable = 0;
   let HasVGPR = 1;
   let HasSGPR = 1;
 }
 
 def VS_32_Lo128 : SIRegisterClass<"AMDGPU", [i32, f32, i16, f16, v2i16, v2f16], 32,

llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-inline-asm.ll

define amdgpu_kernel void @asm_simple_agpr_clobber() {
   ; CHECK-NEXT:   S_ENDPGM 0
   call void asm sideeffect "; def a0", "~{a0}"(), !srcloc !0
   ret void
 }
 
 define i32 @asm_vgpr_early_clobber() {
   ; CHECK-LABEL: name: asm_vgpr_early_clobber
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 7; v_mov_b32 $1, 7", 1 /* sideeffect attdialect */, 1769483 /* regdef-ec:VGPR_32 */, def early-clobber %8, 1769483 /* regdef-ec:VGPR_32 */, def early-clobber %9, !0
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 7; v_mov_b32 $1, 7", 1 /* sideeffect attdialect */, 2490379 /* regdef-ec:VGPR_32 */, def early-clobber %8, 2490379 /* regdef-ec:VGPR_32 */, def early-clobber %9, !0
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %8
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %9
   ; CHECK-NEXT:   [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[COPY1]]
   ; CHECK-NEXT:   $vgpr0 = COPY [[ADD]](s32)
   ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   call { i32, i32 } asm sideeffect "v_mov_b32 $0, 7; v_mov_b32 $1, 7", "=&v,=&v"(), !srcloc !0
   %asmresult = extractvalue { i32, i32 } %1, 0
   %asmresult1 = extractvalue { i32, i32 } %1, 1

 entry:
   %0 = tail call i32 asm "v_mov_b32 v1, 7", "={v1}"() nounwind
   ret i32 %0
 }
 
 define i32 @test_single_vgpr_output() nounwind {
   ; CHECK-LABEL: name: test_single_vgpr_output
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 7", 0 /* attdialect */, 1769482 /* regdef:VGPR_32 */, def %8
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 7", 0 /* attdialect */, 2490378 /* regdef:VGPR_32 */, def %8
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %8
   ; CHECK-NEXT:   $vgpr0 = COPY [[COPY]](s32)
   ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %0 = tail call i32 asm "v_mov_b32 $0, 7", "=v"() nounwind
   ret i32 %0
 }
 
 define i32 @test_single_sgpr_output_s32() nounwind {
   ; CHECK-LABEL: name: test_single_sgpr_output_s32
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 1900554 /* regdef:SReg_32 */, def %8
+  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 2621450 /* regdef:SReg_32 */, def %8
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %8
   ; CHECK-NEXT:   $vgpr0 = COPY [[COPY]](s32)
   ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %0 = tail call i32 asm "s_mov_b32 $0, 7", "=s"() nounwind
   ret i32 %0
 }
 
 ; Check support for returning several floats
 define float @test_multiple_register_outputs_same() #0 {
   ; CHECK-LABEL: name: test_multiple_register_outputs_same
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 0; v_mov_b32 $1, 1", 0 /* attdialect */, 1769482 /* regdef:VGPR_32 */, def %8, 1769482 /* regdef:VGPR_32 */, def %9
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 0; v_mov_b32 $1, 1", 0 /* attdialect */, 2490378 /* regdef:VGPR_32 */, def %8, 2490378 /* regdef:VGPR_32 */, def %9
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %8
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %9
   ; CHECK-NEXT:   [[FADD:%[0-9]+]]:_(s32) = G_FADD [[COPY]], [[COPY1]]
   ; CHECK-NEXT:   $vgpr0 = COPY [[FADD]](s32)
   ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %1 = call { float, float } asm "v_mov_b32 $0, 0; v_mov_b32 $1, 1", "=v,=v"()
   %asmresult = extractvalue { float, float } %1, 0
   %asmresult1 = extractvalue { float, float } %1, 1
   %add = fadd float %asmresult, %asmresult1
   ret float %add
 }
 
 ; Check support for returning several floats
 define double @test_multiple_register_outputs_mixed() #0 {
   ; CHECK-LABEL: name: test_multiple_register_outputs_mixed
   ; CHECK: bb.1 (%ir-block.0):
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 0; v_add_f64 $1, 0, 0", 0 /* attdialect */, 1769482 /* regdef:VGPR_32 */, def %8, 3080202 /* regdef:VReg_64 */, def %9
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, 0; v_add_f64 $1, 0, 0", 0 /* attdialect */, 2490378 /* regdef:VGPR_32 */, def %8, 3801098 /* regdef:VReg_64 */, def %9
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %8
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s64) = COPY %9
   ; CHECK-NEXT:   [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
   ; CHECK-NEXT:   $vgpr0 = COPY [[UV]](s32)
   ; CHECK-NEXT:   $vgpr1 = COPY [[UV1]](s32)
   ; CHECK-NEXT:   SI_RETURN implicit $vgpr0, implicit $vgpr1
   %1 = call { float, double } asm "v_mov_b32 $0, 0; v_add_f64 $1, 0, 0", "=v,=v"()
   %asmresult = extractvalue { float, double } %1, 1

 define amdgpu_kernel void @test_input_vgpr_imm() {
   ; CHECK-LABEL: name: test_input_vgpr_imm
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr8_sgpr9
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p4) = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 42
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[C]](s32)
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 v0, $0", 1 /* sideeffect attdialect */, 1769481 /* reguse:VGPR_32 */, [[COPY1]]
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 v0, $0", 1 /* sideeffect attdialect */, 2490377 /* reguse:VGPR_32 */, [[COPY1]]
   ; CHECK-NEXT:   S_ENDPGM 0
   call void asm sideeffect "v_mov_b32 v0, $0", "v"(i32 42)
   ret void
 }
 
 define amdgpu_kernel void @test_input_sgpr_imm() {
   ; CHECK-LABEL: name: test_input_sgpr_imm
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $sgpr8_sgpr9
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p4) = COPY $sgpr8_sgpr9
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 42
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY [[C]](s32)
-  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 s0, $0", 1 /* sideeffect attdialect */, 1900553 /* reguse:SReg_32 */, [[COPY1]]
+  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 s0, $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[COPY1]]
   ; CHECK-NEXT:   S_ENDPGM 0
   call void asm sideeffect "s_mov_b32 s0, $0", "s"(i32 42)
   ret void
 }
 
 define amdgpu_kernel void @test_input_imm() {
   ; CHECK-LABEL: name: test_input_imm
   ; CHECK: bb.1 (%ir-block.0):

 
 define float @test_input_vgpr(i32 %src) nounwind {
   ; CHECK-LABEL: name: test_input_vgpr
   ; CHECK: bb.1.entry:
   ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
-  ; CHECK-NEXT:   INLINEASM &"v_add_f32 $0, 1.0, $1", 0 /* attdialect */, 1769482 /* regdef:VGPR_32 */, def %9, 1769481 /* reguse:VGPR_32 */, [[COPY1]]
+  ; CHECK-NEXT:   INLINEASM &"v_add_f32 $0, 1.0, $1", 0 /* attdialect */, 2490378 /* regdef:VGPR_32 */, def %9, 2490377 /* reguse:VGPR_32 */, [[COPY1]]
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY %9
   ; CHECK-NEXT:   $vgpr0 = COPY [[COPY2]](s32)
   ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %0 = tail call float asm "v_add_f32 $0, 1.0, $1", "=v,v"(i32 %src) nounwind
   ret float %0
 }
 
 define i32 @test_memory_constraint(ptr addrspace(3) %a) nounwind {
   ; CHECK-LABEL: name: test_memory_constraint
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
-  ; CHECK-NEXT:   INLINEASM &"ds_read_b32 $0, $1", 8 /* mayload attdialect */, 1769482 /* regdef:VGPR_32 */, def %9, 262158 /* mem:m */, [[COPY]](p3)
+  ; CHECK-NEXT:   INLINEASM &"ds_read_b32 $0, $1", 8 /* mayload attdialect */, 2490378 /* regdef:VGPR_32 */, def %9, 262158 /* mem:m */, [[COPY]](p3)
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %9
   ; CHECK-NEXT:   $vgpr0 = COPY [[COPY1]](s32)
   ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %1 = tail call i32 asm "ds_read_b32 $0, $1", "=v,*m"(ptr addrspace(3) elementtype(i32) %a)
   ret i32 %1
 }
 
 define i32 @test_vgpr_matching_constraint(i32 %a) nounwind {
   ; CHECK-LABEL: name: test_vgpr_matching_constraint
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $vgpr0
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
   ; CHECK-NEXT:   [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY]], [[C]]
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[AND]](s32)
-  ; CHECK-NEXT:   INLINEASM &";", 1 /* sideeffect attdialect */, 1769482 /* regdef:VGPR_32 */, def %11, 2147483657 /* reguse tiedto:$0 */, [[COPY1]](tied-def 3)
+  ; CHECK-NEXT:   INLINEASM &";", 1 /* sideeffect attdialect */, 2490378 /* regdef:VGPR_32 */, def %11, 2147483657 /* reguse tiedto:$0 */, [[COPY1]](tied-def 3)
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY %11
   ; CHECK-NEXT:   $vgpr0 = COPY [[COPY2]](s32)
   ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
   %and = and i32 %a, 1
   %asm = call i32 asm sideeffect ";", "=v,0"(i32 %and)
   ret i32 %asm
 }
 
 define i32 @test_sgpr_matching_constraint() nounwind {
   ; CHECK-LABEL: name: test_sgpr_matching_constraint
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 1900554 /* regdef:SReg_32 */, def %8
+  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 2621450 /* regdef:SReg_32 */, def %8
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %8
-  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 8", 0 /* attdialect */, 1900554 /* regdef:SReg_32 */, def %10
+  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 8", 0 /* attdialect */, 2621450 /* regdef:SReg_32 */, def %10
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY %10
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY [[COPY]](s32)
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY [[COPY1]](s32)
-  ; CHECK-NEXT:   INLINEASM &"s_add_u32 $0, $1, $2", 0 /* attdialect */, 1900554 /* regdef:SReg_32 */, def %12, 1900553 /* reguse:SReg_32 */, [[COPY2]], 2147483657 /* reguse tiedto:$0 */, [[COPY3]](tied-def 3)
+  ; CHECK-NEXT:   INLINEASM &"s_add_u32 $0, $1, $2", 0 /* attdialect */, 2621450 /* regdef:SReg_32 */, def %12, 2621449 /* reguse:SReg_32 */, [[COPY2]], 2147483657 /* reguse tiedto:$0 */, [[COPY3]](tied-def 3)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:_(s32) = COPY %12
   ; CHECK-NEXT:   $vgpr0 = COPY [[COPY4]](s32)
   ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %asm0 = tail call i32 asm "s_mov_b32 $0, 7", "=s"() nounwind
   %asm1 = tail call i32 asm "s_mov_b32 $0, 8", "=s"() nounwind
   %asm2 = tail call i32 asm "s_add_u32 $0, $1, $2", "=s,s,0"(i32 %asm0, i32 %asm1) nounwind
   ret i32 %asm2
 }
 
 define void @test_many_matching_constraints(i32 %a, i32 %b, i32 %c) nounwind {
   ; CHECK-LABEL: name: test_many_matching_constraints
   ; CHECK: bb.1 (%ir-block.0):
   ; CHECK-NEXT:   liveins: $vgpr0, $vgpr1, $vgpr2
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY $vgpr2
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF
   ; CHECK-NEXT:   [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[COPY2]](s32)
   ; CHECK-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
   ; CHECK-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[COPY1]](s32)
-  ; CHECK-NEXT:   INLINEASM &"; ", 1 /* sideeffect attdialect */, 1769482 /* regdef:VGPR_32 */, def %11, 1769482 /* regdef:VGPR_32 */, def %12, 1769482 /* regdef:VGPR_32 */, def %13, 2147483657 /* reguse tiedto:$0 */, [[COPY3]](tied-def 3), 2147614729 /* reguse tiedto:$2 */, [[COPY4]](tied-def 7), 2147549193 /* reguse tiedto:$1 */, [[COPY5]](tied-def 5)
+  ; CHECK-NEXT:   INLINEASM &"; ", 1 /* sideeffect attdialect */, 2490378 /* regdef:VGPR_32 */, def %11, 2490378 /* regdef:VGPR_32 */, def %12, 2490378 /* regdef:VGPR_32 */, def %13, 2147483657 /* reguse tiedto:$0 */, [[COPY3]](tied-def 3), 2147614729 /* reguse tiedto:$2 */, [[COPY4]](tied-def 7), 2147549193 /* reguse tiedto:$1 */, [[COPY5]](tied-def 5)
   ; CHECK-NEXT:   [[COPY6:%[0-9]+]]:_(s32) = COPY %11
   ; CHECK-NEXT:   [[COPY7:%[0-9]+]]:_(s32) = COPY %12
   ; CHECK-NEXT:   [[COPY8:%[0-9]+]]:_(s32) = COPY %13
   ; CHECK-NEXT:   G_STORE [[COPY6]](s32), [[DEF]](p1) :: (store (s32) into `ptr addrspace(1) undef`, addrspace 1)
   ; CHECK-NEXT:   G_STORE [[COPY7]](s32), [[DEF]](p1) :: (store (s32) into `ptr addrspace(1) undef`, addrspace 1)
   ; CHECK-NEXT:   G_STORE [[COPY8]](s32), [[DEF]](p1) :: (store (s32) into `ptr addrspace(1) undef`, addrspace 1)
   ; CHECK-NEXT:   SI_RETURN
   %asm = call {i32, i32, i32} asm sideeffect "; ", "=v,=v,=v,0,2,1"(i32 %c, i32 %a, i32 %b)
   %asmresult0 = extractvalue  {i32, i32, i32} %asm, 0
   store i32 %asmresult0, ptr addrspace(1) undef
   %asmresult1 = extractvalue  {i32, i32, i32} %asm, 1
   store i32 %asmresult1, ptr addrspace(1) undef
   %asmresult2 = extractvalue  {i32, i32, i32} %asm, 2
   store i32 %asmresult2, ptr addrspace(1) undef
   ret void
 }
 
 define i32 @test_sgpr_to_vgpr_move_matching_constraint() nounwind {
   ; CHECK-LABEL: name: test_sgpr_to_vgpr_move_matching_constraint
   ; CHECK: bb.1.entry:
-  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 1900554 /* regdef:SReg_32 */, def %8
+  ; CHECK-NEXT:   INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 2621450 /* regdef:SReg_32 */, def %8
   ; CHECK-NEXT:   [[COPY:%[0-9]+]]:_(s32) = COPY %8
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
-  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, $1", 0 /* attdialect */, 1769482 /* regdef:VGPR_32 */, def %10, 2147483657 /* reguse tiedto:$0 */, [[COPY1]](tied-def 3)
+  ; CHECK-NEXT:   INLINEASM &"v_mov_b32 $0, $1", 0 /* attdialect */, 2490378 /* regdef:VGPR_32 */, def %10, 2147483657 /* reguse tiedto:$0 */, [[COPY1]](tied-def 3)
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:_(s32) = COPY %10
   ; CHECK-NEXT:   $vgpr0 = COPY [[COPY2]](s32)
   ; CHECK-NEXT:   SI_RETURN implicit $vgpr0
 entry:
   %asm0 = tail call i32 asm "s_mov_b32 $0, 7", "=s"() nounwind
   %asm1 = tail call i32 asm "v_mov_b32 $0, $1", "=v,0"(i32 %asm0) nounwind
   ret i32 %asm1
 }

llvm/test/CodeGen/AMDGPU/branch-relax-indirect-branch.mir

body:             |
   ; CHECK-NEXT:   $sgpr1 = V_READLANE_B32 killed $vgpr2, 1
   ; CHECK-NEXT:   $vgpr2 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.0, addrspace 5)
   ; CHECK-NEXT:   $exec = S_NOT_B64 $exec, implicit-def dead $scc, implicit killed $vgpr2
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.2:
   ; CHECK-NEXT:   successors: %bb.3(0x04000000), %bb.7(0x7c000000)
   ; CHECK-NEXT:   liveins: $vcc_hi, $vcc_lo, $sgpr5, $sgpr6, $sgpr7, $sgpr8, $sgpr9, $sgpr10, $sgpr11, $sgpr12, $sgpr13, $sgpr14, $sgpr15, $sgpr16, $sgpr17, $sgpr18, $sgpr19, $sgpr20, $sgpr21, $sgpr22, $sgpr23, $sgpr24, $sgpr25, $sgpr26, $sgpr27, $sgpr28, $sgpr29, $sgpr30, $sgpr31, $sgpr34, $sgpr35, $sgpr36, $sgpr37, $sgpr38, $sgpr39, $sgpr40, $sgpr41, $sgpr42, $sgpr43, $sgpr44, $sgpr45, $sgpr46, $sgpr47, $sgpr48, $sgpr49, $sgpr50, $sgpr51, $sgpr52, $sgpr53, $sgpr54, $sgpr55, $sgpr56, $sgpr57, $sgpr58, $sgpr59, $sgpr60, $sgpr61, $sgpr62, $sgpr63, $sgpr64, $sgpr65, $sgpr66, $sgpr67, $sgpr68, $sgpr69, $sgpr70, $sgpr71, $sgpr72, $sgpr73, $sgpr74, $sgpr75, $sgpr76, $sgpr77, $sgpr78, $sgpr79, $sgpr80, $sgpr81, $sgpr82, $sgpr83, $sgpr84, $sgpr85, $sgpr86, $sgpr87, $sgpr88, $sgpr89, $sgpr90, $sgpr91, $sgpr92, $sgpr93, $sgpr94, $sgpr95, $sgpr96, $sgpr97, $sgpr98, $sgpr99, $sgpr100, $sgpr101, $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   INLINEASM &"v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64", 1 /* sideeffect attdialect */, 2097162 /* regdef:SReg_32_XEXEC */, def renamable $sgpr4
+  ; CHECK-NEXT:   INLINEASM &"v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64", 1 /* sideeffect attdialect */, 2097162 /* regdef:VS_32 */, def renamable $sgpr4
   ; CHECK-NEXT:   S_CMP_LG_U32 killed renamable $sgpr4, 0, implicit-def $scc
   ; CHECK-NEXT:   S_CBRANCH_SCC0 %bb.3, implicit killed $scc
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.7:
   ; CHECK-NEXT:   successors: %bb.8(0x80000000)
   ; CHECK-NEXT:   liveins: $vcc_hi, $vcc_lo, $sgpr5, $sgpr6, $sgpr7, $sgpr8, $sgpr9, $sgpr10, $sgpr11, $sgpr12, $sgpr13, $sgpr14, $sgpr15, $sgpr16, $sgpr17, $sgpr18, $sgpr19, $sgpr20, $sgpr21, $sgpr22, $sgpr23, $sgpr24, $sgpr25, $sgpr26, $sgpr27, $sgpr28, $sgpr29, $sgpr30, $sgpr31, $sgpr34, $sgpr35, $sgpr36, $sgpr37, $sgpr38, $sgpr39, $sgpr40, $sgpr41, $sgpr42, $sgpr43, $sgpr44, $sgpr45, $sgpr46, $sgpr47, $sgpr48, $sgpr49, $sgpr50, $sgpr51, $sgpr52, $sgpr53, $sgpr54, $sgpr55, $sgpr56, $sgpr57, $sgpr58, $sgpr59, $sgpr60, $sgpr61, $sgpr62, $sgpr63, $sgpr64, $sgpr65, $sgpr66, $sgpr67, $sgpr68, $sgpr69, $sgpr70, $sgpr71, $sgpr72, $sgpr73, $sgpr74, $sgpr75, $sgpr76, $sgpr77, $sgpr78, $sgpr79, $sgpr80, $sgpr81, $sgpr82, $sgpr83, $sgpr84, $sgpr85, $sgpr86, $sgpr87, $sgpr88, $sgpr89, $sgpr90, $sgpr91, $sgpr92, $sgpr93, $sgpr94, $sgpr95, $sgpr96, $sgpr97, $sgpr98, $sgpr99, $sgpr100, $sgpr101, $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   $exec = S_NOT_B64 $exec, implicit-def dead $scc, implicit-def $vgpr2

llvm/test/CodeGen/AMDGPU/branch-relax-no-terminators.mir

body:             |
   ; CHECK-NEXT:   $sgpr1 = V_READLANE_B32 killed $vgpr2, 1
   ; CHECK-NEXT:   $vgpr2 = BUFFER_LOAD_DWORD_OFFSET $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr32, 0, 0, 0, implicit $exec :: (load (s32) from %stack.0, addrspace 5)
   ; CHECK-NEXT:   $exec = S_NOT_B64 $exec, implicit-def dead $scc, implicit killed $vgpr2
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.2:
   ; CHECK-NEXT:   successors: %bb.3(0x04000000), %bb.7(0x7c000000)
   ; CHECK-NEXT:   liveins: $vcc_hi, $vcc_lo, $sgpr5, $sgpr6, $sgpr7, $sgpr8, $sgpr9, $sgpr10, $sgpr11, $sgpr12, $sgpr13, $sgpr14, $sgpr15, $sgpr16, $sgpr17, $sgpr18, $sgpr19, $sgpr20, $sgpr21, $sgpr22, $sgpr23, $sgpr24, $sgpr25, $sgpr26, $sgpr27, $sgpr28, $sgpr29, $sgpr30, $sgpr31, $sgpr34, $sgpr35, $sgpr36, $sgpr37, $sgpr38, $sgpr39, $sgpr40, $sgpr41, $sgpr42, $sgpr43, $sgpr44, $sgpr45, $sgpr46, $sgpr47, $sgpr48, $sgpr49, $sgpr50, $sgpr51, $sgpr52, $sgpr53, $sgpr54, $sgpr55, $sgpr56, $sgpr57, $sgpr58, $sgpr59, $sgpr60, $sgpr61, $sgpr62, $sgpr63, $sgpr64, $sgpr65, $sgpr66, $sgpr67, $sgpr68, $sgpr69, $sgpr70, $sgpr71, $sgpr72, $sgpr73, $sgpr74, $sgpr75, $sgpr76, $sgpr77, $sgpr78, $sgpr79, $sgpr80, $sgpr81, $sgpr82, $sgpr83, $sgpr84, $sgpr85, $sgpr86, $sgpr87, $sgpr88, $sgpr89, $sgpr90, $sgpr91, $sgpr92, $sgpr93, $sgpr94, $sgpr95, $sgpr96, $sgpr97, $sgpr98, $sgpr99, $sgpr100, $sgpr101, $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   INLINEASM &"v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64", 1 /* sideeffect attdialect */, 2097162 /* regdef:SReg_32_XEXEC */, def renamable $sgpr4
+  ; CHECK-NEXT:   INLINEASM &"v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64\0A    v_nop_e64", 1 /* sideeffect attdialect */, 2097162 /* regdef:VS_32 */, def renamable $sgpr4
   ; CHECK-NEXT:   S_CMP_LG_U32 killed renamable $sgpr4, 0, implicit-def $scc
   ; CHECK-NEXT:   S_CBRANCH_SCC0 %bb.3, implicit killed $scc
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.7:
   ; CHECK-NEXT:   successors: %bb.8(0x80000000)
   ; CHECK-NEXT:   liveins: $vcc_hi, $vcc_lo, $sgpr5, $sgpr6, $sgpr7, $sgpr8, $sgpr9, $sgpr10, $sgpr11, $sgpr12, $sgpr13, $sgpr14, $sgpr15, $sgpr16, $sgpr17, $sgpr18, $sgpr19, $sgpr20, $sgpr21, $sgpr22, $sgpr23, $sgpr24, $sgpr25, $sgpr26, $sgpr27, $sgpr28, $sgpr29, $sgpr30, $sgpr31, $sgpr34, $sgpr35, $sgpr36, $sgpr37, $sgpr38, $sgpr39, $sgpr40, $sgpr41, $sgpr42, $sgpr43, $sgpr44, $sgpr45, $sgpr46, $sgpr47, $sgpr48, $sgpr49, $sgpr50, $sgpr51, $sgpr52, $sgpr53, $sgpr54, $sgpr55, $sgpr56, $sgpr57, $sgpr58, $sgpr59, $sgpr60, $sgpr61, $sgpr62, $sgpr63, $sgpr64, $sgpr65, $sgpr66, $sgpr67, $sgpr68, $sgpr69, $sgpr70, $sgpr71, $sgpr72, $sgpr73, $sgpr74, $sgpr75, $sgpr76, $sgpr77, $sgpr78, $sgpr79, $sgpr80, $sgpr81, $sgpr82, $sgpr83, $sgpr84, $sgpr85, $sgpr86, $sgpr87, $sgpr88, $sgpr89, $sgpr90, $sgpr91, $sgpr92, $sgpr93, $sgpr94, $sgpr95, $sgpr96, $sgpr97, $sgpr98, $sgpr99, $sgpr100, $sgpr101, $vgpr0, $vgpr1
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   $exec = S_NOT_B64 $exec, implicit-def dead $scc, implicit-def $vgpr2

llvm/test/CodeGen/AMDGPU/coalescer-early-clobber-subreg.mir

 tracksRegLiveness: true
 body:             |
   bb.0:
     liveins: $vgpr0_vgpr1
 
     ; CHECK-LABEL: name: foo1
     ; CHECK: liveins: $vgpr0_vgpr1
     ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835018 /* regdef:SRegOrLds_32 */, def undef %2.sub0, 1835019 /* regdef-ec:SRegOrLds_32 */, def undef early-clobber %2.sub1
+    ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835018 /* regdef:VS_16_and_SReg_1_with_lo16_in_TTMP_LO16 */, def undef %2.sub0, 1835019 /* regdef-ec:VS_16_and_SReg_1_with_lo16_in_TTMP_LO16 */, def undef early-clobber %2.sub1
     ; CHECK-NEXT: FLAT_STORE_DWORDX2 $vgpr0_vgpr1, %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
     ; CHECK-NEXT: S_ENDPGM 0
     INLINEASM &"", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %0:vgpr_32, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %1:vgpr_32
     undef %2.sub0:vreg_64 = COPY killed %0
     %2.sub1:vreg_64 = COPY killed %1
     FLAT_STORE_DWORDX2 killed $vgpr0_vgpr1, killed %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
     S_ENDPGM 0
 
 ...
 
 ---
 name:            foo2
 tracksRegLiveness: true
 body:             |
   bb.0:
     liveins: $vgpr0_vgpr1
 
     ; CHECK-LABEL: name: foo2
     ; CHECK: liveins: $vgpr0_vgpr1
     ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835019 /* regdef-ec:SRegOrLds_32 */, def undef early-clobber %2.sub1, 1835018 /* regdef:SRegOrLds_32 */, def undef %2.sub0
+    ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835019 /* regdef-ec:VS_16_and_SReg_1_with_lo16_in_TTMP_LO16 */, def undef early-clobber %2.sub1, 1835018 /* regdef:VS_16_and_SReg_1_with_lo16_in_TTMP_LO16 */, def undef %2.sub0
     ; CHECK-NEXT: FLAT_STORE_DWORDX2 $vgpr0_vgpr1, %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
     ; CHECK-NEXT: S_ENDPGM 0
     INLINEASM &"", 0 /* attdialect */, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %1:vgpr_32, 1835018 /* regdef:VGPR_32 */, def %0:vgpr_32
     undef %2.sub0:vreg_64 = COPY killed %0
     %2.sub1:vreg_64 = COPY killed %1
     FLAT_STORE_DWORDX2 killed $vgpr0_vgpr1, killed %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
     S_ENDPGM 0
 
 ...
 
 ---
 name:            foo3
 tracksRegLiveness: true
 body:             |
   bb.0:
     liveins: $vgpr0_vgpr1
 
     ; CHECK-LABEL: name: foo3
     ; CHECK: liveins: $vgpr0_vgpr1
     ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835018 /* regdef:SRegOrLds_32 */, def undef %2.sub0, 1835019 /* regdef-ec:SRegOrLds_32 */, def undef early-clobber %2.sub1
+    ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835018 /* regdef:VS_16_and_SReg_1_with_lo16_in_TTMP_LO16 */, def undef %2.sub0, 1835019 /* regdef-ec:VS_16_and_SReg_1_with_lo16_in_TTMP_LO16 */, def undef early-clobber %2.sub1
     ; CHECK-NEXT: FLAT_STORE_DWORDX2 $vgpr0_vgpr1, %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
     ; CHECK-NEXT: S_ENDPGM 0
     INLINEASM &"", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %1:vgpr_32, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %0:vgpr_32
     undef %2.sub0:vreg_64 = COPY killed %1
     %2.sub1:vreg_64 = COPY killed %0
     FLAT_STORE_DWORDX2 killed $vgpr0_vgpr1, killed %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
     S_ENDPGM 0
 
 ...
 
 ---
 name:            foo4
 tracksRegLiveness: true
 body:             |
   bb.0:
     liveins: $vgpr0_vgpr1
 
     ; CHECK-LABEL: name: foo4
     ; CHECK: liveins: $vgpr0_vgpr1
     ; CHECK-NEXT: {{  $}}
-    ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835019 /* regdef-ec:SRegOrLds_32 */, def undef early-clobber %2.sub1, 1835018 /* regdef:SRegOrLds_32 */, def undef %2.sub0
+    ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835019 /* regdef-ec:VS_16_and_SReg_1_with_lo16_in_TTMP_LO16 */, def undef early-clobber %2.sub1, 1835018 /* regdef:VS_16_and_SReg_1_with_lo16_in_TTMP_LO16 */, def undef %2.sub0
     ; CHECK-NEXT: FLAT_STORE_DWORDX2 $vgpr0_vgpr1, %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
     ; CHECK-NEXT: S_ENDPGM 0
     INLINEASM &"", 0 /* attdialect */, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %0:vgpr_32, 1835018 /* regdef:VGPR_32 */, def %1:vgpr_32
     undef %2.sub0:vreg_64 = COPY killed %1
     %2.sub1:vreg_64 = COPY killed %0
     FLAT_STORE_DWORDX2 killed $vgpr0_vgpr1, killed %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
     S_ENDPGM 0
 
 ...

llvm/test/CodeGen/AMDGPU/inline-asm.i128.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -stop-after=finalize-isel -o - %s | FileCheck -check-prefix=GFX908 %s
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a -stop-after=finalize-isel -o - %s | FileCheck -check-prefix=GFX90A %s
 
 ; Make sure we only use one 128-bit register instead of 2 for i128 asm
 ; constraints
 
 define amdgpu_kernel void @s_input_output_i128() {
   ; GFX908-LABEL: name: s_input_output_i128
   ; GFX908: bb.0 (%ir-block.0):
-  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6881290 /* regdef:SGPR_128 */, def %4
+  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7602186 /* regdef:SGPR_128 */, def %4
   ; GFX908-NEXT:   [[COPY:%[0-9]+]]:sgpr_128 = COPY %4
-  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6881289 /* reguse:SGPR_128 */, [[COPY]]
+  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7602185 /* reguse:SGPR_128 */, [[COPY]]
   ; GFX908-NEXT:   S_ENDPGM 0
+  ;
   ; GFX90A-LABEL: name: s_input_output_i128
   ; GFX90A: bb.0 (%ir-block.0):
-  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6881290 /* regdef:SGPR_128 */, def %4
+  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7602186 /* regdef:SGPR_128 */, def %4
   ; GFX90A-NEXT:   [[COPY:%[0-9]+]]:sgpr_128 = COPY %4
-  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6881289 /* reguse:SGPR_128 */, [[COPY]]
+  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7602185 /* reguse:SGPR_128 */, [[COPY]]
   ; GFX90A-NEXT:   S_ENDPGM 0
   %val = tail call i128 asm sideeffect "; def $0", "=s"()
   call void asm sideeffect "; use $0", "s"(i128 %val)
   ret void
 }
 
 define amdgpu_kernel void @v_input_output_i128() {
   ; GFX908-LABEL: name: v_input_output_i128
   ; GFX908: bb.0 (%ir-block.0):
-  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5767178 /* regdef:VReg_128 */, def %4
+  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6488074 /* regdef:VReg_128 */, def %4
   ; GFX908-NEXT:   [[COPY:%[0-9]+]]:vreg_128 = COPY %4
-  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5767177 /* reguse:VReg_128 */, [[COPY]]
+  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6488073 /* reguse:VReg_128 */, [[COPY]]
   ; GFX908-NEXT:   S_ENDPGM 0
+  ;
   ; GFX90A-LABEL: name: v_input_output_i128
   ; GFX90A: bb.0 (%ir-block.0):
-  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6094858 /* regdef:VReg_128_Align2 */, def %4
+  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6815754 /* regdef:VReg_128_Align2 */, def %4
   ; GFX90A-NEXT:   [[COPY:%[0-9]+]]:vreg_128_align2 = COPY %4
-  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6094857 /* reguse:VReg_128_Align2 */, [[COPY]]
+  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6815753 /* reguse:VReg_128_Align2 */, [[COPY]]
   ; GFX90A-NEXT:   S_ENDPGM 0
   %val = tail call i128 asm sideeffect "; def $0", "=v"()
   call void asm sideeffect "; use $0", "v"(i128 %val)
   ret void
 }
 
 define amdgpu_kernel void @a_input_output_i128() {
   ; GFX908-LABEL: name: a_input_output_i128
   ; GFX908: bb.0 (%ir-block.0):
-  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5701642 /* regdef:AReg_128 */, def %4
+  ; GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6422538 /* regdef:AReg_128 */, def %4
   ; GFX908-NEXT:   [[COPY:%[0-9]+]]:areg_128 = COPY %4
-  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5701641 /* reguse:AReg_128 */, [[COPY]]
+  ; GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6422537 /* reguse:AReg_128 */, [[COPY]]
   ; GFX908-NEXT:   S_ENDPGM 0
+  ;
   ; GFX90A-LABEL: name: a_input_output_i128
   ; GFX90A: bb.0 (%ir-block.0):
-  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5963786 /* regdef:AReg_128_Align2 */, def %4
+  ; GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6684682 /* regdef:AReg_128_Align2 */, def %4
   ; GFX90A-NEXT:   [[COPY:%[0-9]+]]:areg_128_align2 = COPY %4
-  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5963785 /* reguse:AReg_128_Align2 */, [[COPY]]
+  ; GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6684681 /* reguse:AReg_128_Align2 */, [[COPY]]
   ; GFX90A-NEXT:   S_ENDPGM 0
   %val = call i128 asm sideeffect "; def $0", "=a"()
   call void asm sideeffect "; use $0", "a"(i128 %val)
   ret void
 }

llvm/test/CodeGen/AMDGPU/partial-regcopy-and-spill-missed-at-regalloc.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ;RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 --stop-after=greedy,1 -verify-machineinstrs < %s | FileCheck -check-prefix=REGALLOC-GFX908 %s
 ;RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 --stop-after=prologepilog -verify-machineinstrs < %s | FileCheck -check-prefix=PEI-GFX908 %s
 ;RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a --stop-after=greedy,1 -verify-machineinstrs < %s | FileCheck -check-prefix=REGALLOC-GFX90A %s
 ;RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a --stop-after=prologepilog -verify-machineinstrs < %s | FileCheck -check-prefix=PEI-GFX90A %s
 
 ; Partial reg copy and spill missed during regalloc handled later at frame lowering.
 define amdgpu_kernel void @partial_copy(<4 x i32> %arg) #0 {
   ; REGALLOC-GFX908-LABEL: name: partial_copy
   ; REGALLOC-GFX908: bb.0 (%ir-block.0):
   ; REGALLOC-GFX908-NEXT:   liveins: $sgpr4_sgpr5
   ; REGALLOC-GFX908-NEXT: {{  $}}
-  ; REGALLOC-GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1703945 /* reguse:AGPR_32 */, undef %5:agpr_32
-  ; REGALLOC-GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5767178 /* regdef:VReg_128 */, def %26
+  ; REGALLOC-GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef %5:agpr_32
+  ; REGALLOC-GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6488074 /* regdef:VReg_128 */, def %26
   ; REGALLOC-GFX908-NEXT:   [[COPY:%[0-9]+]]:av_128 = COPY %26
-  ; REGALLOC-GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3080202 /* regdef:VReg_64 */, def %23
+  ; REGALLOC-GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3801098 /* regdef:VReg_64 */, def %23
   ; REGALLOC-GFX908-NEXT:   SI_SPILL_V64_SAVE %23, %stack.0, $sgpr32, 0, implicit $exec :: (store (s64) into %stack.0, align 4, addrspace 5)
   ; REGALLOC-GFX908-NEXT:   [[COPY1:%[0-9]+]]:vreg_128 = COPY [[COPY]]
   ; REGALLOC-GFX908-NEXT:   GLOBAL_STORE_DWORDX4 undef %14:vreg_64, [[COPY1]], 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) undef`, addrspace 1)
   ; REGALLOC-GFX908-NEXT:   renamable $sgpr0_sgpr1_sgpr2_sgpr3 = S_LOAD_DWORDX4_IMM killed renamable $sgpr4_sgpr5, 0, 0 :: (dereferenceable invariant load (s128) from %ir.arg.kernarg.offset1, addrspace 4)
   ; REGALLOC-GFX908-NEXT:   [[COPY2:%[0-9]+]]:areg_128 = COPY killed renamable $sgpr0_sgpr1_sgpr2_sgpr3
   ; REGALLOC-GFX908-NEXT:   [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 1, implicit $exec
   ; REGALLOC-GFX908-NEXT:   [[V_MOV_B32_e32_1:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 2, implicit $exec
   ; REGALLOC-GFX908-NEXT:   [[V_MFMA_I32_4X4X4I8_e64_:%[0-9]+]]:areg_128 = V_MFMA_I32_4X4X4I8_e64 [[V_MOV_B32_e32_]], [[V_MOV_B32_e32_1]], [[COPY2]], 0, 0, 0, implicit $mode, implicit $exec
   ; REGALLOC-GFX908-NEXT:   [[SI_SPILL_V64_RESTORE:%[0-9]+]]:vreg_64 = SI_SPILL_V64_RESTORE %stack.0, $sgpr32, 0, implicit $exec :: (load (s64) from %stack.0, align 4, addrspace 5)
   ; REGALLOC-GFX908-NEXT:   GLOBAL_STORE_DWORDX2 undef %16:vreg_64, [[SI_SPILL_V64_RESTORE]], 0, 0, implicit $exec :: (volatile store (s64) into `ptr addrspace(1) undef`, addrspace 1)
   ; REGALLOC-GFX908-NEXT:   [[COPY3:%[0-9]+]]:vreg_128 = COPY [[V_MFMA_I32_4X4X4I8_e64_]]
   ; REGALLOC-GFX908-NEXT:   GLOBAL_STORE_DWORDX4 undef %18:vreg_64, [[COPY3]], 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) undef`, addrspace 1)
   ; REGALLOC-GFX908-NEXT:   S_ENDPGM 0
+  ;
   ; PEI-GFX908-LABEL: name: partial_copy
   ; PEI-GFX908: bb.0 (%ir-block.0):
   ; PEI-GFX908-NEXT:   liveins: $agpr4, $sgpr4_sgpr5, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr7
   ; PEI-GFX908-NEXT: {{  $}}
   ; PEI-GFX908-NEXT:   $sgpr8_sgpr9_sgpr10_sgpr11 = COPY killed $sgpr0_sgpr1_sgpr2_sgpr3
   ; PEI-GFX908-NEXT:   $sgpr8 = S_ADD_U32 $sgpr8, $sgpr7, implicit-def $scc, implicit-def $sgpr8_sgpr9_sgpr10_sgpr11
   ; PEI-GFX908-NEXT:   $sgpr9 = S_ADDC_U32 $sgpr9, 0, implicit-def dead $scc, implicit $scc, implicit-def $sgpr8_sgpr9_sgpr10_sgpr11
-  ; PEI-GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1703945 /* reguse:AGPR_32 */, undef renamable $agpr0
-  ; PEI-GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 5767178 /* regdef:VReg_128 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
+  ; PEI-GFX908-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef renamable $agpr0
+  ; PEI-GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6488074 /* regdef:VReg_128 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
   ; PEI-GFX908-NEXT:   renamable $agpr0_agpr1_agpr2_agpr3 = COPY killed renamable $vgpr0_vgpr1_vgpr2_vgpr3, implicit $exec
-  ; PEI-GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3080202 /* regdef:VReg_64 */, def renamable $vgpr0_vgpr1
+  ; PEI-GFX908-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3801098 /* regdef:VReg_64 */, def renamable $vgpr0_vgpr1
   ; PEI-GFX908-NEXT:   BUFFER_STORE_DWORD_OFFSET killed $vgpr0, $sgpr8_sgpr9_sgpr10_sgpr11, 0, 4, 0, 0, implicit $exec, implicit-def $vgpr0_vgpr1, implicit $vgpr0_vgpr1 :: (store (s32) into %stack.0, addrspace 5)
   ; PEI-GFX908-NEXT:   $agpr4 = V_ACCVGPR_WRITE_B32_e64 killed $vgpr1, implicit $exec, implicit killed $vgpr0_vgpr1
   ; PEI-GFX908-NEXT:   renamable $vgpr0_vgpr1_vgpr2_vgpr3 = COPY killed renamable $agpr0_agpr1_agpr2_agpr3, implicit $exec
   ; PEI-GFX908-NEXT:   GLOBAL_STORE_DWORDX4 undef renamable $vgpr0_vgpr1, killed renamable $vgpr0_vgpr1_vgpr2_vgpr3, 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) undef`, addrspace 1)
   ; PEI-GFX908-NEXT:   renamable $sgpr0_sgpr1_sgpr2_sgpr3 = S_LOAD_DWORDX4_IMM killed renamable $sgpr4_sgpr5, 0, 0 :: (dereferenceable invariant load (s128) from %ir.arg.kernarg.offset1, addrspace 4)
   ; PEI-GFX908-NEXT:   renamable $agpr0_agpr1_agpr2_agpr3 = COPY killed renamable $sgpr0_sgpr1_sgpr2_sgpr3, implicit $exec
   ; PEI-GFX908-NEXT:   renamable $vgpr0 = V_MOV_B32_e32 1, implicit $exec
   ; PEI-GFX908-NEXT:   renamable $vgpr1 = V_MOV_B32_e32 2, implicit $exec
   ; PEI-GFX908-NEXT:   renamable $agpr0_agpr1_agpr2_agpr3 = V_MFMA_I32_4X4X4I8_e64 killed $vgpr0, killed $vgpr1, killed $agpr0_agpr1_agpr2_agpr3, 0, 0, 0, implicit $mode, implicit $exec
   ; PEI-GFX908-NEXT:   $vgpr0 = BUFFER_LOAD_DWORD_OFFSET $sgpr8_sgpr9_sgpr10_sgpr11, 0, 4, 0, 0, implicit $exec, implicit-def $vgpr0_vgpr1 :: (load (s32) from %stack.0, addrspace 5)
   ; PEI-GFX908-NEXT:   $vgpr1 = V_ACCVGPR_READ_B32_e64 $agpr4, implicit $exec, implicit $vgpr0_vgpr1
   ; PEI-GFX908-NEXT:   GLOBAL_STORE_DWORDX2 undef renamable $vgpr0_vgpr1, killed renamable $vgpr0_vgpr1, 0, 0, implicit $exec :: (volatile store (s64) into `ptr addrspace(1) undef`, addrspace 1)
   ; PEI-GFX908-NEXT:   renamable $vgpr0_vgpr1_vgpr2_vgpr3 = COPY killed renamable $agpr0_agpr1_agpr2_agpr3, implicit $exec
   ; PEI-GFX908-NEXT:   GLOBAL_STORE_DWORDX4 undef renamable $vgpr0_vgpr1, killed renamable $vgpr0_vgpr1_vgpr2_vgpr3, 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) undef`, addrspace 1)
   ; PEI-GFX908-NEXT:   S_ENDPGM 0
+  ;
   ; REGALLOC-GFX90A-LABEL: name: partial_copy
   ; REGALLOC-GFX90A: bb.0 (%ir-block.0):
   ; REGALLOC-GFX90A-NEXT:   liveins: $sgpr4_sgpr5
   ; REGALLOC-GFX90A-NEXT: {{  $}}
-  ; REGALLOC-GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1703945 /* reguse:AGPR_32 */, undef %5:agpr_32
-  ; REGALLOC-GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6094858 /* regdef:VReg_128_Align2 */, def %25
+  ; REGALLOC-GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef %5:agpr_32
+  ; REGALLOC-GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6815754 /* regdef:VReg_128_Align2 */, def %25
   ; REGALLOC-GFX90A-NEXT:   [[COPY:%[0-9]+]]:av_128_align2 = COPY %25
-  ; REGALLOC-GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3407882 /* regdef:VReg_64_Align2 */, def %23
+  ; REGALLOC-GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 4128778 /* regdef:VReg_64_Align2 */, def %23
   ; REGALLOC-GFX90A-NEXT:   SI_SPILL_V64_SAVE %23, %stack.0, $sgpr32, 0, implicit $exec :: (store (s64) into %stack.0, align 4, addrspace 5)
   ; REGALLOC-GFX90A-NEXT:   GLOBAL_STORE_DWORDX4 undef %14:vreg_64_align2, [[COPY]], 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) undef`, addrspace 1)
   ; REGALLOC-GFX90A-NEXT:   renamable $sgpr0_sgpr1_sgpr2_sgpr3 = S_LOAD_DWORDX4_IMM killed renamable $sgpr4_sgpr5, 0, 0 :: (dereferenceable invariant load (s128) from %ir.arg.kernarg.offset1, addrspace 4)
   ; REGALLOC-GFX90A-NEXT:   [[COPY1:%[0-9]+]]:areg_128_align2 = COPY killed renamable $sgpr0_sgpr1_sgpr2_sgpr3
   ; REGALLOC-GFX90A-NEXT:   [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 1, implicit $exec
   ; REGALLOC-GFX90A-NEXT:   [[V_MOV_B32_e32_1:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 2, implicit $exec
   ; REGALLOC-GFX90A-NEXT:   [[V_MFMA_I32_4X4X4I8_e64_:%[0-9]+]]:areg_128_align2 = V_MFMA_I32_4X4X4I8_e64 [[V_MOV_B32_e32_]], [[V_MOV_B32_e32_1]], [[COPY1]], 0, 0, 0, implicit $mode, implicit $exec
   ; REGALLOC-GFX90A-NEXT:   [[SI_SPILL_AV64_RESTORE:%[0-9]+]]:av_64_align2 = SI_SPILL_AV64_RESTORE %stack.0, $sgpr32, 0, implicit $exec :: (load (s64) from %stack.0, align 4, addrspace 5)
   ; REGALLOC-GFX90A-NEXT:   GLOBAL_STORE_DWORDX2 undef %16:vreg_64_align2, [[SI_SPILL_AV64_RESTORE]], 0, 0, implicit $exec :: (volatile store (s64) into `ptr addrspace(1) undef`, addrspace 1)
   ; REGALLOC-GFX90A-NEXT:   GLOBAL_STORE_DWORDX4 undef %18:vreg_64_align2, [[V_MFMA_I32_4X4X4I8_e64_]], 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) undef`, addrspace 1)
   ; REGALLOC-GFX90A-NEXT:   S_ENDPGM 0
+  ;
   ; PEI-GFX90A-LABEL: name: partial_copy
   ; PEI-GFX90A: bb.0 (%ir-block.0):
   ; PEI-GFX90A-NEXT:   liveins: $agpr4, $sgpr4_sgpr5, $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr7
   ; PEI-GFX90A-NEXT: {{  $}}
   ; PEI-GFX90A-NEXT:   $sgpr8_sgpr9_sgpr10_sgpr11 = COPY killed $sgpr0_sgpr1_sgpr2_sgpr3
   ; PEI-GFX90A-NEXT:   $sgpr8 = S_ADD_U32 $sgpr8, $sgpr7, implicit-def $scc, implicit-def $sgpr8_sgpr9_sgpr10_sgpr11
   ; PEI-GFX90A-NEXT:   $sgpr9 = S_ADDC_U32 $sgpr9, 0, implicit-def dead $scc, implicit $scc, implicit-def $sgpr8_sgpr9_sgpr10_sgpr11
-  ; PEI-GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1703945 /* reguse:AGPR_32 */, undef renamable $agpr0
-  ; PEI-GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6094858 /* regdef:VReg_128_Align2 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
+  ; PEI-GFX90A-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef renamable $agpr0
+  ; PEI-GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 6815754 /* regdef:VReg_128_Align2 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
   ; PEI-GFX90A-NEXT:   renamable $agpr0_agpr1_agpr2_agpr3 = COPY killed renamable $vgpr0_vgpr1_vgpr2_vgpr3, implicit $exec
-  ; PEI-GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3407882 /* regdef:VReg_64_Align2 */, def renamable $vgpr0_vgpr1
+  ; PEI-GFX90A-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 4128778 /* regdef:VReg_64_Align2 */, def renamable $vgpr0_vgpr1
   ; PEI-GFX90A-NEXT:   BUFFER_STORE_DWORD_OFFSET killed $vgpr0, $sgpr8_sgpr9_sgpr10_sgpr11, 0, 4, 0, 0, implicit $exec, implicit-def $vgpr0_vgpr1, implicit $vgpr0_vgpr1 :: (store (s32) into %stack.0, addrspace 5)
   ; PEI-GFX90A-NEXT:   $agpr4 = V_ACCVGPR_WRITE_B32_e64 killed $vgpr1, implicit $exec, implicit killed $vgpr0_vgpr1
   ; PEI-GFX90A-NEXT:   GLOBAL_STORE_DWORDX4 undef renamable $vgpr0_vgpr1, killed renamable $agpr0_agpr1_agpr2_agpr3, 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) undef`, addrspace 1)
   ; PEI-GFX90A-NEXT:   renamable $sgpr0_sgpr1_sgpr2_sgpr3 = S_LOAD_DWORDX4_IMM killed renamable $sgpr4_sgpr5, 0, 0 :: (dereferenceable invariant load (s128) from %ir.arg.kernarg.offset1, addrspace 4)
   ; PEI-GFX90A-NEXT:   renamable $agpr0_agpr1_agpr2_agpr3 = COPY killed renamable $sgpr0_sgpr1_sgpr2_sgpr3, implicit $exec
   ; PEI-GFX90A-NEXT:   renamable $vgpr0 = V_MOV_B32_e32 1, implicit $exec
   ; PEI-GFX90A-NEXT:   renamable $vgpr1 = V_MOV_B32_e32 2, implicit $exec
   ; PEI-GFX90A-NEXT:   renamable $agpr0_agpr1_agpr2_agpr3 = V_MFMA_I32_4X4X4I8_e64 killed $vgpr0, killed $vgpr1, killed $agpr0_agpr1_agpr2_agpr3, 0, 0, 0, implicit $mode, implicit $exec

llvm/test/CodeGen/AMDGPU/sched-assert-dead-def-subreg-use-other-subreg.mir

body:             |
   ; CHECK-NEXT:   successors: %bb.1(0x80000000)
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   BUFFER_STORE_DWORD_OFFEN %0.sub3, undef %5:vgpr_32, $sgpr24_sgpr25_sgpr26_sgpr27, $sgpr32, 0, 0, 0, implicit $exec :: (store (s32), align 8, addrspace 5)
   ; CHECK-NEXT:   dead [[DS_READ_B32_gfx9_:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 undef %7:vgpr_32, 0, 0, implicit $exec
   ; CHECK-NEXT:   dead [[DS_READ_B64_gfx9_:%[0-9]+]]:vreg_64 = DS_READ_B64_gfx9 [[V_MOV_B32_e32_]], 0, 0, implicit $exec
   ; CHECK-NEXT:   dead [[DS_READ_B128_gfx9_:%[0-9]+]]:vreg_128 = DS_READ_B128_gfx9 [[V_ADD_U32_e32_]], 0, 0, implicit $exec
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[COPY]].sub0
   ; CHECK-NEXT:   undef %11.sub1:vreg_512 = COPY [[COPY]].sub1
-  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_HI16 */, def dead [[COPY1]], 851978 /* regdef:VGPR_HI16 */, def dead [[COPY]].sub1, 2147483657 /* reguse tiedto:$0 */, [[COPY1]], 2147549193 /* reguse tiedto:$1 */, [[COPY]].sub1
+  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_16 */, def dead [[COPY1]], 851978 /* regdef:VGPR_16 */, def dead [[COPY]].sub1, 2147483657 /* reguse tiedto:$0 */, [[COPY1]], 2147549193 /* reguse tiedto:$1 */, [[COPY]].sub1
   ; CHECK-NEXT:   %11.sub0:vreg_512 = COPY [[COPY]].sub0
   ; CHECK-NEXT:   %11.sub3:vreg_512 = COPY [[COPY]].sub3
   ; CHECK-NEXT:   [[V_MOV_B32_e32_1:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
   ; CHECK-NEXT:   dead [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 4, [[V_MOV_B32_e32_1]], implicit-def dead $vcc, implicit $exec
   ; CHECK-NEXT:   %11.sub2:vreg_512 = COPY undef [[V_MOV_B32_e32_]]
   ; CHECK-NEXT:   %11.sub5:vreg_512 = COPY undef [[V_MOV_B32_e32_]]
   ; CHECK-NEXT:   [[COPY2:%[0-9]+]]:vreg_512 = COPY %11
   ; CHECK-NEXT:   S_BRANCH %bb.1

llvm/test/CodeGen/AMDGPU/sched-handleMoveUp-subreg-def-across-subreg-def.mir

body:             |
   ; CHECK-NEXT:   [[V_MOV_B32_e32_1:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 -4, implicit $exec
   ; CHECK-NEXT:   [[DEF:%[0-9]+]]:vreg_64 = IMPLICIT_DEF
   ; CHECK-NEXT:   [[DEF1:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
   ; CHECK-NEXT:   [[DEF2:%[0-9]+]]:vgpr_32 = IMPLICIT_DEF
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.1:
   ; CHECK-NEXT:   successors: %bb.1(0x80000000)
   ; CHECK-NEXT: {{  $}}
-  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_HI16 */, def dead %11
+  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_16 */, def dead %11
   ; CHECK-NEXT:   GLOBAL_STORE_DWORD undef %12:vreg_64, [[BUFFER_LOAD_DWORD_OFFEN]], 0, 0, implicit $exec :: (store (s32), addrspace 1)
   ; CHECK-NEXT:   [[V_MOV_B32_e32_2:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
   ; CHECK-NEXT:   %5.sub1:vreg_64 = COPY [[V_MOV_B32_e32_]]
   ; CHECK-NEXT:   [[DS_READ_B64_gfx9_:%[0-9]+]]:vreg_64 = DS_READ_B64_gfx9 undef %14:vgpr_32, 0, 0, implicit $exec :: (load (s64), addrspace 3)
-  ; CHECK-NEXT:   INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_HI16 */, def %15, 851978 /* regdef:VGPR_HI16 */, def %16
+  ; CHECK-NEXT:   INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_16 */, def %15, 851978 /* regdef:VGPR_16 */, def %16
   ; CHECK-NEXT:   [[DS_READ_B32_gfx9_:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 [[V_MOV_B32_e32_]], 0, 0, implicit $exec
   ; CHECK-NEXT:   [[DS_READ_B32_gfx9_1:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 [[V_MOV_B32_e32_1]], 0, 0, implicit $exec
   ; CHECK-NEXT:   [[DS_READ_B32_gfx9_2:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 undef %20:vgpr_32, 0, 0, implicit $exec
-  ; CHECK-NEXT:   INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_HI16 */, def %21, 851978 /* regdef:VGPR_HI16 */, def %22
+  ; CHECK-NEXT:   INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_16 */, def %21, 851978 /* regdef:VGPR_16 */, def %22
   ; CHECK-NEXT:   [[DS_READ_B32_gfx9_3:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 [[V_MOV_B32_e32_1]], 0, 0, implicit $exec
   ; CHECK-NEXT:   [[V_MOV_B32_e32_3:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
-  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_HI16 */, def dead [[V_MOV_B32_e32_2]], 851978 /* regdef:VGPR_HI16 */, def dead [[V_MOV_B32_e32_3]], 851977 /* reguse:VGPR_HI16 */, [[DS_READ_B64_gfx9_]].sub0, 2147483657 /* reguse tiedto:$0 */, [[V_MOV_B32_e32_2]](tied-def 3), 2147549193 /* reguse tiedto:$1 */, [[V_MOV_B32_e32_3]](tied-def 5), 851977 /* reguse:VGPR_HI16 */, %15, 851977 /* reguse:VGPR_HI16 */, %16, 851977 /* reguse:VGPR_HI16 */, [[DS_READ_B32_gfx9_1]], 851977 /* reguse:VGPR_HI16 */, [[DS_READ_B32_gfx9_]], 851977 /* reguse:VGPR_HI16 */, [[DS_READ_B32_gfx9_3]], 851977 /* reguse:VGPR_HI16 */, [[DS_READ_B32_gfx9_2]]
+  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_16 */, def dead [[V_MOV_B32_e32_2]], 851978 /* regdef:VGPR_16 */, def dead [[V_MOV_B32_e32_3]], 851977 /* reguse:VGPR_16 */, [[DS_READ_B64_gfx9_]].sub0, 2147483657 /* reguse tiedto:$0 */, [[V_MOV_B32_e32_2]](tied-def 3), 2147549193 /* reguse tiedto:$1 */, [[V_MOV_B32_e32_3]](tied-def 5), 851977 /* reguse:VGPR_16 */, %15, 851977 /* reguse:VGPR_16 */, %16, 851977 /* reguse:VGPR_16 */, [[DS_READ_B32_gfx9_1]], 851977 /* reguse:VGPR_16 */, [[DS_READ_B32_gfx9_]], 851977 /* reguse:VGPR_16 */, [[DS_READ_B32_gfx9_3]], 851977 /* reguse:VGPR_16 */, [[DS_READ_B32_gfx9_2]]
   ; CHECK-NEXT:   DS_WRITE_B32_gfx9 undef %28:vgpr_32, %21, 0, 0, implicit $exec :: (store (s32), addrspace 3)
   ; CHECK-NEXT:   DS_WRITE_B32_gfx9 undef %29:vgpr_32, %22, 0, 0, implicit $exec :: (store (s32), addrspace 3)
   ; CHECK-NEXT:   DS_WRITE_B64_gfx9 undef %30:vgpr_32, %5, 0, 0, implicit $exec :: (store (s64), addrspace 3)
   ; CHECK-NEXT:   undef %31.sub1:vreg_64 = FLAT_LOAD_DWORD undef %32:vreg_64, 0, 0, implicit $exec, implicit $flat_scr :: (load (s32))
   ; CHECK-NEXT:   [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 1, [[DEF2]], implicit $exec
   ; CHECK-NEXT:   [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_LOAD_DWORDX2_IMM]].sub1
   ; CHECK-NEXT:   [[DEF]].sub1:vreg_64 = COPY [[V_MOV_B32_e32_]]
   ; CHECK-NEXT:   [[V_MUL_LO_U32_e64_:%[0-9]+]]:vgpr_32 = V_MUL_LO_U32_e64 [[V_ADD_U32_e32_]], [[S_MOV_B32_]], implicit $exec

llvm/test/CodeGen/AMDGPU/spill-vector-superclass.ll

 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx908 -stop-after=greedy,1 -verify-machineinstrs -o - %s | FileCheck -check-prefix=GCN %s
 ; Convert AV spills into VGPR spills by introducing appropriate copies in between.
 
 define amdgpu_kernel void @test_spill_av_class(<4 x i32> %arg) #0 {
   ; GCN-LABEL: name: test_spill_av_class
   ; GCN: bb.0 (%ir-block.0):
   ; GCN-NEXT:   liveins: $sgpr4_sgpr5
   ; GCN-NEXT: {{  $}}
   ; GCN-NEXT:   renamable $sgpr0_sgpr1_sgpr2_sgpr3 = S_LOAD_DWORDX4_IMM killed renamable $sgpr4_sgpr5, 0, 0 :: (dereferenceable invariant load (s128) from %ir.arg.kernarg.offset1, addrspace 4)
   ; GCN-NEXT:   [[COPY:%[0-9]+]]:areg_128 = COPY killed renamable $sgpr0_sgpr1_sgpr2_sgpr3
   ; GCN-NEXT:   [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 1, implicit $exec
   ; GCN-NEXT:   [[V_MOV_B32_e32_1:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 2, implicit $exec
   ; GCN-NEXT:   [[V_MFMA_I32_4X4X4I8_e64_:%[0-9]+]]:areg_128 = V_MFMA_I32_4X4X4I8_e64 [[V_MOV_B32_e32_]], [[V_MOV_B32_e32_1]], [[COPY]], 0, 0, 0, implicit $mode, implicit $exec
-  ; GCN-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 1769482 /* regdef:VGPR_32 */, def undef %24.sub0
+  ; GCN-NEXT:   INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 2490378 /* regdef:VGPR_32 */, def undef %24.sub0
   ; GCN-NEXT:   SI_SPILL_V64_SAVE %24, %stack.0, $sgpr32, 0, implicit $exec :: (store (s64) into %stack.0, align 4, addrspace 5)
   ; GCN-NEXT:   [[COPY1:%[0-9]+]]:vreg_128 = COPY [[V_MFMA_I32_4X4X4I8_e64_]]
   ; GCN-NEXT:   GLOBAL_STORE_DWORDX4 undef %16:vreg_64, [[COPY1]], 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) undef`, addrspace 1)
   ; GCN-NEXT:   [[SI_SPILL_V64_RESTORE:%[0-9]+]]:vreg_64 = SI_SPILL_V64_RESTORE %stack.0, $sgpr32, 0, implicit $exec :: (load (s64) from %stack.0, align 4, addrspace 5)
-  ; GCN-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3080201 /* reguse:VReg_64 */, [[SI_SPILL_V64_RESTORE]]
+  ; GCN-NEXT:   INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3801097 /* reguse:VReg_64 */, [[SI_SPILL_V64_RESTORE]]
   ; GCN-NEXT:   S_ENDPGM 0
   %v0 = call i32 asm sideeffect "; def $0", "=v"()
   %tmp = insertelement <2 x i32> undef, i32 %v0, i32 0
   %mai = tail call <4 x i32> @llvm.amdgcn.mfma.i32.4x4x4i8(i32 1, i32 2, <4 x i32> %arg, i32 0, i32 0, i32 0)
   store volatile <4 x i32> %mai, ptr addrspace(1) undef
   call void asm sideeffect "; use $0", "v"(<2 x i32> %tmp);
   ret void
 }
 
 declare <4 x i32> @llvm.amdgcn.mfma.i32.4x4x4i8(i32, i32, <4 x i32>, i32, i32, i32)
 
 attributes #0 = { nounwind "amdgpu-num-vgpr"="5" }

llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir

body:             |
   ; CHECK-NEXT:   undef %0.sub0:vreg_64 = V_MOV_B32_e32 0, implicit $exec
   ; CHECK-NEXT:   %0.sub1:vreg_64 = V_MOV_B32_e32 0, implicit $exec
   ; CHECK-NEXT:   [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.1:
   ; CHECK-NEXT:   successors: %bb.1(0x80000000)
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[DS_READ_B32_gfx9_:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 [[V_MOV_B32_e32_]], 0, 0, implicit $exec :: (load (s32), addrspace 3)
-  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_HI16 */, def %0, 2147549193 /* reguse tiedto:$1 */, %0(tied-def 3)
-  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851977 /* reguse:VGPR_HI16 */, [[DS_READ_B32_gfx9_]]
-  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_HI16 */, def undef %0.sub0, 851978 /* regdef:VGPR_HI16 */, def undef %0.sub1
+  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_16 */, def %0, 2147549193 /* reguse tiedto:$1 */, %0(tied-def 3)
+  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851977 /* reguse:VGPR_16 */, [[DS_READ_B32_gfx9_]]
+  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_16 */, def undef %0.sub0, 851978 /* regdef:VGPR_16 */, def undef %0.sub1
   ; CHECK-NEXT:   S_NOP 0, implicit %0.sub1
   ; CHECK-NEXT:   $sgpr10 = S_MOV_B32 -1
   ; CHECK-NEXT:   S_BRANCH %bb.1
   bb.0:
     undef %0.sub0:vreg_64 = V_MOV_B32_e32 0, implicit $exec
     %0.sub1:vreg_64 = V_MOV_B32_e32 0, implicit $exec
     %1:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
 
body:             |
   ; CHECK-NEXT:   undef %0.sub0:vreg_64 = V_MOV_B32_e32 0, implicit $exec
   ; CHECK-NEXT:   %0.sub1:vreg_64 = V_MOV_B32_e32 0, implicit $exec
   ; CHECK-NEXT:   [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT: bb.1:
   ; CHECK-NEXT:   successors: %bb.1(0x80000000)
   ; CHECK-NEXT: {{  $}}
   ; CHECK-NEXT:   [[DS_READ_B32_gfx9_:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 [[V_MOV_B32_e32_]], 0, 0, implicit $exec :: (load (s32), addrspace 3)
-  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_HI16 */, def %0, 2147549193 /* reguse tiedto:$1 */, %0(tied-def 3)
-  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851977 /* reguse:VGPR_HI16 */, [[DS_READ_B32_gfx9_]]
-  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_HI16 */, def undef %0.sub1, 851978 /* regdef:VGPR_HI16 */, def undef %0.sub0
+  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_16 */, def %0, 2147549193 /* reguse tiedto:$1 */, %0(tied-def 3)
+  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851977 /* reguse:VGPR_16 */, [[DS_READ_B32_gfx9_]]
+  ; CHECK-NEXT:   INLINEASM &"", 1 /* sideeffect attdialect */, 851978 /* regdef:VGPR_16 */, def undef %0.sub1, 851978 /* regdef:VGPR_16 */, def undef %0.sub0
   ; CHECK-NEXT:   S_NOP 0, implicit %0.sub1
   ; CHECK-NEXT:   $sgpr10 = S_MOV_B32 -1
   ; CHECK-NEXT:   S_BRANCH %bb.1
   bb.0:
     undef %0.sub0:vreg_64 = V_MOV_B32_e32 0, implicit $exec
     %0.sub1:vreg_64 = V_MOV_B32_e32 0, implicit $exec
     %1:vgpr_32 = V_MOV_B32_e32 0, implicit $exec