AMDGPU: Fix expensive_checks machine verifier errors in new tests

Author: arsenm

llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-insert-extract.mir

@@ -18,6 +18,7 @@ body: |
  ; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:areg_64_align2 = GLOBAL_LOAD_DWORDX2 [[COPY]], 0, 0, implicit $exec :: (load (s64), addrspace 1)
  ; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
  ; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]]
+ ; CHECK-NEXT: [[COPY3:%[0-9]+]].sub2_sub3:areg_128_align2 = IMPLICIT_DEF
  ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, [[COPY3]]
  ; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
  ; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], [[COPY3]].sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
@@ -28,6 +29,7 @@ body: |
  %3:vreg_64_align2 = GLOBAL_LOAD_DWORDX2 %0, 0, 0, implicit $exec :: (load (s64), addrspace 1)
  %4:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
  undef %5.sub0_sub1:areg_128_align2 = COPY %4
+ %5.sub2_sub3 = IMPLICIT_DEF
  INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, %5
  GLOBAL_STORE_DWORDX4 %0, %5, 0, 0, implicit $exec :: (store (s128), addrspace 1)
  GLOBAL_STORE_DWORDX2 %0, %5.sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
@@ -136,15 +138,17 @@ body: |
  ; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
  ; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
  ; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:areg_64_align2 = GLOBAL_LOAD_DWORDX2 [[COPY]], 0, 0, implicit $exec :: (load (s64), addrspace 1)
- ; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
- ; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr_32 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub0
+ ; CHECK-NEXT: undef [[DEF:%[0-9]+]].sub0:areg_128_align2 = IMPLICIT_DEF
+ ; CHECK-NEXT: [[DEF:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
+ ; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr_32 = COPY [[DEF]].sub0
  ; CHECK-NEXT: GLOBAL_STORE_DWORD [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s32), addrspace 1)
  ; CHECK-NEXT: SI_RETURN
  %0:vreg_64_align2 = COPY $vgpr4_vgpr5
  %1:av_64_align2 = COPY $vgpr0_vgpr1
  %2:av_64_align2 = COPY $vgpr2_vgpr3
  %3:vreg_64_align2 = GLOBAL_LOAD_DWORDX2 %0, 0, 0, implicit $exec :: (load (s64), addrspace 1)
- undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
+ undef %4.sub0 = IMPLICIT_DEF
+ %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
  %5:agpr_32 = COPY %4.sub0
  GLOBAL_STORE_DWORD %0, %5, 0, 0, implicit $exec :: (store (s32), addrspace 1)
  SI_RETURN
@@ -167,6 +171,7 @@ body: |
  ; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:areg_64_align2 = GLOBAL_LOAD_DWORDX2 [[COPY]], 0, 0, implicit $exec :: (load (s64), addrspace 1)
  ; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
  ; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
+ ; CHECK-NEXT: [[COPY3:%[0-9]+]].sub2_sub3:areg_128_align2 = IMPLICIT_DEF
  ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, [[COPY3]]
  ; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
  ; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], [[COPY3]].sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
@@ -177,6 +182,7 @@ body: |
  %3:vreg_64_align2 = GLOBAL_LOAD_DWORDX2 %0, 0, 0, implicit $exec :: (load (s64), addrspace 1)
  undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
  undef %5.sub0_sub1:areg_128_align2 = COPY %4.sub2_sub3
+ %5.sub2_sub3 = IMPLICIT_DEF
  INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, %5
  GLOBAL_STORE_DWORDX4 %0, %5, 0, 0, implicit $exec :: (store (s128), addrspace 1)
  GLOBAL_STORE_DWORDX2 %0, %5.sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
@@ -201,6 +207,7 @@ body: |
  ; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:vreg_64_align2 = GLOBAL_LOAD_DWORDX2 [[COPY]], 0, 0, implicit $exec :: (load (s64), addrspace 1)
  ; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_vgprcd_e64_:%[0-9]+]].sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
  ; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_vgprcd_e64_]].sub2
+ ; CHECK-NEXT: [[COPY3:%[0-9]+]].sub2_sub3:areg_128_align2 = IMPLICIT_DEF
  ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, [[COPY3]]
  ; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
  ; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], [[COPY3]].sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
@@ -211,6 +218,7 @@ body: |
  %3:vreg_64_align2 = GLOBAL_LOAD_DWORDX2 %0, 0, 0, implicit $exec :: (load (s64), addrspace 1)
  undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
  undef %5.sub1:areg_128_align2 = COPY %4.sub2
+ %5.sub2_sub3 = IMPLICIT_DEF
  INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, %5
  GLOBAL_STORE_DWORDX4 %0, %5, 0, 0, implicit $exec :: (store (s128), addrspace 1)
  GLOBAL_STORE_DWORDX2 %0, %5.sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)

llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-src2-chain.mir

@@ -225,6 +225,7 @@ body: |
  ; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
  ; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
  ; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
+ ; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub0_sub1:areg_128_align2 = IMPLICIT_DEF
  ; CHECK-NEXT: dead %other_use0:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub0_sub1
  ; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub0_sub1:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
  ; CHECK-NEXT: dead %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
@@ -238,6 +239,7 @@ body: |
  %2:av_64_align2 = COPY $vgpr2_vgpr3
  %3:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
  undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
+ %4.sub0_sub1 = IMPLICIT_DEF
  %other_use0:vreg_64_align2 = COPY %4.sub0_sub1
  %4.sub0_sub1:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
  %other_use1:vreg_64_align2 = COPY %4.sub2_sub3
@@ -265,6 +267,7 @@ body: |
  ; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
  ; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
  ; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
+ ; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub0_sub1:areg_128_align2 = IMPLICIT_DEF
  ; CHECK-NEXT: dead %other_use0:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub0_sub1
  ; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub0_sub1:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
  ; CHECK-NEXT: %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
@@ -278,6 +281,7 @@ body: |
  %2:av_64_align2 = COPY $vgpr2_vgpr3
  %3:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
  undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
+ %4.sub0_sub1 = IMPLICIT_DEF
  %other_use0:vreg_64_align2 = COPY %4.sub0_sub1
  %4.sub0_sub1:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
  %other_use1:vreg_64_align2 = COPY %4.sub2_sub3
@@ -301,13 +305,14 @@ body: |
  ; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr4_vgpr5
  ; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
  ; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
- ; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
- ; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
- ; CHECK-NEXT: dead %other_use0:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub0_sub1
- ; CHECK-NEXT: dead [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
- ; CHECK-NEXT: %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
- ; CHECK-NEXT: dead %other_use2:vreg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
- ; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
+ ; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
+ ; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_vgprcd_e64_:%[0-9]+]].sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
+ ; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_vgprcd_e64_:%[0-9]+]].sub0_sub1:vreg_128_align2 = IMPLICIT_DEF
+ ; CHECK-NEXT: dead [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]].sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_vgprcd_e64_]].sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
+ ; CHECK-NEXT: dead %other_use0:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_vgprcd_e64_]].sub0_sub1
+ ; CHECK-NEXT: %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_vgprcd_e64_]].sub2_sub3
+ ; CHECK-NEXT: dead %other_use2:vreg_64 = COPY [[V_MFMA_F64_4X4X4F64_vgprcd_e64_]].sub1_sub2
+ ; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_64 = COPY [[V_MFMA_F64_4X4X4F64_vgprcd_e64_]].sub1_sub2
  ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3473417 /* reguse:AReg_64 */, [[COPY3]]
  ; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], %other_use1, 0, 0, implicit $exec :: (store (s64), addrspace 1)
  ; CHECK-NEXT: SI_RETURN
@@ -316,8 +321,9 @@ body: |
  %2:av_64_align2 = COPY $vgpr2_vgpr3
  %3:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
  undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
- %other_use0:vreg_64_align2 = COPY %4.sub0_sub1
+ %4.sub0_sub1 = IMPLICIT_DEF
  %3.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %4.sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
+ %other_use0:vreg_64_align2 = COPY %4.sub0_sub1
  %other_use1:vreg_64_align2 = COPY %4.sub2_sub3
  %other_use2:vreg_64 = COPY %4.sub1_sub2
  %6:areg_64 = COPY %4.sub1_sub2