fix lit tests after rebase.

Author: cdevadas

llvm/test/CodeGen/AMDGPU/fp-atomics-gfx940.ll

@@ -416,298 +416,4 @@ define <2 x i16> @local_atomic_fadd_v2bf16_rtn(ptr addrspace(3) %ptr, <2 x i16>
   ret <2 x i16> %ret
 }
 
-define float @flat_atomic_fadd_f32_intrinsic_ret__posoffset(ptr %ptr, float %data) {
-; GFX940-LABEL: flat_atomic_fadd_f32_intrinsic_ret__posoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    flat_atomic_add_f32 v0, v[0:1], v2 offset:4092 sc0
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_f32_intrinsic_ret__posoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_add_f32 v0, v[0:1], v2 offset:4092 th:TH_ATOMIC_RETURN
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr float, ptr %ptr, i64 1023
-  %result = call float @llvm.amdgcn.flat.atomic.fadd.f32.p0.f32(ptr %gep, float %data)
-  ret float %result
-}
-
-define float @flat_atomic_fadd_f32_intrinsic_ret__negoffset(ptr %ptr, float %data) {
-; GFX940-LABEL: flat_atomic_fadd_f32_intrinsic_ret__negoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    v_add_co_u32_e32 v0, vcc, 0xfffffc00, v0
-; GFX940-NEXT:    s_nop 1
-; GFX940-NEXT:    v_addc_co_u32_e32 v1, vcc, -1, v1, vcc
-; GFX940-NEXT:    flat_atomic_add_f32 v0, v[0:1], v2 sc0
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_f32_intrinsic_ret__negoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_add_f32 v0, v[0:1], v2 offset:-1024 th:TH_ATOMIC_RETURN
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr float, ptr %ptr, i64 -256
-  %result = call float @llvm.amdgcn.flat.atomic.fadd.f32.p0.f32(ptr %gep, float %data)
-  ret float %result
-}
-
-define void @flat_atomic_fadd_f32_intrinsic_noret__posoffset(ptr %ptr, float %data) {
-; GFX940-LABEL: flat_atomic_fadd_f32_intrinsic_noret__posoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    flat_atomic_add_f32 v[0:1], v2 offset:4092
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_f32_intrinsic_noret__posoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_add_f32 v[0:1], v2 offset:4092
-; GFX12-NEXT:    s_wait_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr float, ptr %ptr, i64 1023
-  %unused = call float @llvm.amdgcn.flat.atomic.fadd.f32.p0.f32(ptr %gep, float %data)
-  ret void
-}
-
-define void @flat_atomic_fadd_f32_intrinsic_noret__negoffset(ptr %ptr, float %data) {
-; GFX940-LABEL: flat_atomic_fadd_f32_intrinsic_noret__negoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    v_add_co_u32_e32 v0, vcc, 0xfffffc00, v0
-; GFX940-NEXT:    s_nop 1
-; GFX940-NEXT:    v_addc_co_u32_e32 v1, vcc, -1, v1, vcc
-; GFX940-NEXT:    flat_atomic_add_f32 v[0:1], v2
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_f32_intrinsic_noret__negoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_add_f32 v[0:1], v2 offset:-1024
-; GFX12-NEXT:    s_wait_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr float, ptr %ptr, i64 -256
-  %unused = call float @llvm.amdgcn.flat.atomic.fadd.f32.p0.f32(ptr %gep, float %data)
-  ret void
-}
-
-define <2 x half> @flat_atomic_fadd_v2f16_intrinsic_ret__posoffset(ptr %ptr, <2 x half> %data) {
-; GFX940-LABEL: flat_atomic_fadd_v2f16_intrinsic_ret__posoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    flat_atomic_pk_add_f16 v0, v[0:1], v2 offset:4092 sc0
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_v2f16_intrinsic_ret__posoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_pk_add_f16 v0, v[0:1], v2 offset:4092 th:TH_ATOMIC_RETURN
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr <2 x half>, ptr %ptr, i64 1023
-  %result = call <2 x half> @llvm.amdgcn.flat.atomic.fadd.v2f16.p0.v2f16(ptr %gep, <2 x half> %data)
-  ret <2 x half> %result
-}
-
-define <2 x half> @flat_atomic_fadd_v2f16_intrinsic_ret__negoffset(ptr %ptr, <2 x half> %data) {
-; GFX940-LABEL: flat_atomic_fadd_v2f16_intrinsic_ret__negoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    v_add_co_u32_e32 v0, vcc, 0xfffffc00, v0
-; GFX940-NEXT:    s_nop 1
-; GFX940-NEXT:    v_addc_co_u32_e32 v1, vcc, -1, v1, vcc
-; GFX940-NEXT:    flat_atomic_pk_add_f16 v0, v[0:1], v2 sc0
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_v2f16_intrinsic_ret__negoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_pk_add_f16 v0, v[0:1], v2 offset:-1024 th:TH_ATOMIC_RETURN
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr <2 x half>, ptr %ptr, i64 -256
-  %result = call <2 x half> @llvm.amdgcn.flat.atomic.fadd.v2f16.p0.v2f16(ptr %gep, <2 x half> %data)
-  ret <2 x half> %result
-}
-
-define void @flat_atomic_fadd_v2f16_intrinsic_noret__posoffset(ptr %ptr, <2 x half> %data) {
-; GFX940-LABEL: flat_atomic_fadd_v2f16_intrinsic_noret__posoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    flat_atomic_pk_add_f16 v[0:1], v2 offset:4092
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_v2f16_intrinsic_noret__posoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_pk_add_f16 v[0:1], v2 offset:4092
-; GFX12-NEXT:    s_wait_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr <2 x half>, ptr %ptr, i64 1023
-  %unused = call <2 x half> @llvm.amdgcn.flat.atomic.fadd.v2f16.p0.v2f16(ptr %gep, <2 x half> %data)
-  ret void
-}
-
-define void @flat_atomic_fadd_v2f16_intrinsic_noret__negoffset(ptr %ptr, <2 x half> %data) {
-; GFX940-LABEL: flat_atomic_fadd_v2f16_intrinsic_noret__negoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    v_add_co_u32_e32 v0, vcc, 0xfffffc00, v0
-; GFX940-NEXT:    s_nop 1
-; GFX940-NEXT:    v_addc_co_u32_e32 v1, vcc, -1, v1, vcc
-; GFX940-NEXT:    flat_atomic_pk_add_f16 v[0:1], v2
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_v2f16_intrinsic_noret__negoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_pk_add_f16 v[0:1], v2 offset:-1024
-; GFX12-NEXT:    s_wait_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr <2 x half>, ptr %ptr, i64 -256
-  %unused = call <2 x half> @llvm.amdgcn.flat.atomic.fadd.v2f16.p0.v2f16(ptr %gep, <2 x half> %data)
-  ret void
-}
-
-define <2 x i16> @flat_atomic_fadd_v2bf16_intrinsic_ret__posoffset(ptr %ptr, <2 x i16> %data) {
-; GFX940-LABEL: flat_atomic_fadd_v2bf16_intrinsic_ret__posoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    flat_atomic_pk_add_bf16 v0, v[0:1], v2 offset:4092 sc0
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_v2bf16_intrinsic_ret__posoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_pk_add_bf16 v0, v[0:1], v2 offset:4092 th:TH_ATOMIC_RETURN
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr <2 x i16>, ptr %ptr, i64 1023
-  %result = call <2 x i16> @llvm.amdgcn.flat.atomic.fadd.v2bf16.p0.v2bf16(ptr %gep, <2 x i16> %data)
-  ret <2 x i16> %result
-}
-
-define <2 x i16> @flat_atomic_fadd_v2bf16_intrinsic_ret__negoffset(ptr %ptr, <2 x i16> %data) {
-; GFX940-LABEL: flat_atomic_fadd_v2bf16_intrinsic_ret__negoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    v_add_co_u32_e32 v0, vcc, 0xfffffc00, v0
-; GFX940-NEXT:    s_nop 1
-; GFX940-NEXT:    v_addc_co_u32_e32 v1, vcc, -1, v1, vcc
-; GFX940-NEXT:    flat_atomic_pk_add_bf16 v0, v[0:1], v2 sc0
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_v2bf16_intrinsic_ret__negoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_pk_add_bf16 v0, v[0:1], v2 offset:-1024 th:TH_ATOMIC_RETURN
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr <2 x i16>, ptr %ptr, i64 -256
-  %result = call <2 x i16> @llvm.amdgcn.flat.atomic.fadd.v2bf16.p0.v2bf16(ptr %gep, <2 x i16> %data)
-  ret <2 x i16> %result
-}
-
-define void @flat_atomic_fadd_v2bf16_intrinsic_noret__posoffset(ptr %ptr, <2 x i16> %data) {
-; GFX940-LABEL: flat_atomic_fadd_v2bf16_intrinsic_noret__posoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    flat_atomic_pk_add_bf16 v[0:1], v2 offset:4092
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_v2bf16_intrinsic_noret__posoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_pk_add_bf16 v[0:1], v2 offset:4092
-; GFX12-NEXT:    s_wait_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr <2 x i16>, ptr %ptr, i64 1023
-  %unused = call <2 x i16> @llvm.amdgcn.flat.atomic.fadd.v2bf16.p0.v2bf16(ptr %gep, <2 x i16> %data)
-  ret void
-}
-
-define void @flat_atomic_fadd_v2bf16_intrinsic_noret__negoffset(ptr %ptr, <2 x i16> %data) {
-; GFX940-LABEL: flat_atomic_fadd_v2bf16_intrinsic_noret__negoffset:
-; GFX940:       ; %bb.0:
-; GFX940-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    v_add_co_u32_e32 v0, vcc, 0xfffffc00, v0
-; GFX940-NEXT:    s_nop 1
-; GFX940-NEXT:    v_addc_co_u32_e32 v1, vcc, -1, v1, vcc
-; GFX940-NEXT:    flat_atomic_pk_add_bf16 v[0:1], v2
-; GFX940-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
-; GFX940-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX12-LABEL: flat_atomic_fadd_v2bf16_intrinsic_noret__negoffset:
-; GFX12:       ; %bb.0:
-; GFX12-NEXT:    s_wait_loadcnt_dscnt 0x0
-; GFX12-NEXT:    s_wait_expcnt 0x0
-; GFX12-NEXT:    s_wait_samplecnt 0x0
-; GFX12-NEXT:    s_wait_bvhcnt 0x0
-; GFX12-NEXT:    s_wait_kmcnt 0x0
-; GFX12-NEXT:    flat_atomic_pk_add_bf16 v[0:1], v2 offset:-1024
-; GFX12-NEXT:    s_wait_dscnt 0x0
-; GFX12-NEXT:    s_setpc_b64 s[30:31]
-  %gep = getelementptr <2 x i16>, ptr %ptr, i64 -256
-  %unused = call <2 x i16> @llvm.amdgcn.flat.atomic.fadd.v2bf16.p0.v2bf16(ptr %gep, <2 x i16> %data)
-  ret void
-}
-
 attributes #0 = { "denormal-fp-math-f32"="ieee,ieee" }

llvm/test/CodeGen/AMDGPU/memory_clause.ll

@@ -446,7 +446,7 @@ define amdgpu_kernel void @flat_scratch_load(float %a, float %b, <8 x i32> %desc
   %val = call <2 x float> @llvm.amdgcn.image.sample.2d.v2f32.f32(i32 9, float %a, float %b, <8 x i32> %desc, <4 x i32> <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 0>, i1 false, i32 0, i32 0)
   %val0 = extractelement <2 x float> %val, i32 0
   %valadd = fadd float %load, %val0
-  call void @llvm.amdgcn.exp.f32(i32 immarg 0, i32 immarg 1, float %valadd, float undef, float undef, float undef, i1 immarg true, i1 immarg true)
+  call void @llvm.amdgcn.exp.f32(i32 0, i32 1, float %valadd, float undef, float undef, float undef, i1 true, i1 true)
   ret void
 }
 
@@ -504,7 +504,7 @@ define amdgpu_kernel void @flat_scratch_load_clause(float %a, float %b, <8 x i32
   %load0 = load float, ptr addrspace(5) %alloca
   %load1 = load float, ptr addrspace(5) %alloca2
   %valadd = fadd float %load0, %load1
-  call void @llvm.amdgcn.exp.f32(i32 immarg 0, i32 immarg 1, float %valadd, float undef, float undef, float undef, i1 immarg true, i1 immarg true)
+  call void @llvm.amdgcn.exp.f32(i32 0, i32 1, float %valadd, float undef, float undef, float undef, i1 true, i1 true)
   ret void
 }