; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py ; RUN: llc -global-isel -global-isel-abort=2 -mtriple=amdgcn -mcpu=tonga -mattr=-flat-for-global < %s | FileCheck -check-prefix=GCN %s define amdgpu_kernel void @set_inactive(ptr addrspace(1) %out, i32 %in) { ; GCN-LABEL: set_inactive: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dword s3, s[4:5], 0x2c ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v1, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, 42, v1, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v1, v0 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dword v1, off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call i32 @llvm.amdgcn.set.inactive.i32(i32 %in, i32 42) #0 %tmp = call i32 @llvm.amdgcn.strict.wwm.i32(i32 %tmp.0) store i32 %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_imm_poison(ptr addrspace(1) %out) { ; GCN-LABEL: set_inactive_imm_poison: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: v_mov_b32_e32 v0, 1 ; GCN-NEXT: v_mov_b32_e32 v0, v0 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: buffer_store_dword v0, off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call i32 @llvm.amdgcn.set.inactive.i32(i32 1, i32 poison) #0 %tmp = call i32 @llvm.amdgcn.strict.wwm.i32(i32 %tmp.0) store i32 %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_64(ptr addrspace(1) %out, i64 %in) { ; GCN-LABEL: set_inactive_64: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v2, s2 ; GCN-NEXT: v_mov_b32_e32 v3, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, 0, v2, s[4:5] ; GCN-NEXT: v_cndmask_b32_e64 v1, 0, v3, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v2, v0 ; GCN-NEXT: v_mov_b32_e32 v3, v1 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dwordx2 v[2:3], off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call i64 @llvm.amdgcn.set.inactive.i64(i64 %in, i64 0) #0 %tmp = call i64 @llvm.amdgcn.strict.wwm.i64(i64 %tmp.0) store i64 %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_imm_poison_64(ptr addrspace(1) %out) { ; GCN-LABEL: set_inactive_imm_poison_64: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: v_mov_b32_e32 v0, 1 ; GCN-NEXT: v_mov_b32_e32 v1, 0 ; GCN-NEXT: v_mov_b32_e32 v0, v0 ; GCN-NEXT: v_mov_b32_e32 v1, v1 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: buffer_store_dwordx2 v[0:1], off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call i64 @llvm.amdgcn.set.inactive.i64(i64 1, i64 poison) #0 %tmp = call i64 @llvm.amdgcn.strict.wwm.i64(i64 %tmp.0) store i64 %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_scc(ptr addrspace(1) %out, i32 %in, <4 x i32> inreg %desc) { ; GCN-LABEL: set_inactive_scc: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x34 ; GCN-NEXT: s_load_dword s6, s[4:5], 0x2c ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: s_buffer_load_dword s7, s[0:3], 0x0 ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: v_mov_b32_e32 v1, s6 ; GCN-NEXT: s_or_saveexec_b64 s[2:3], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, 42, v1, s[2:3] ; GCN-NEXT: s_mov_b64 exec, s[2:3] ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: s_cmp_lg_u32 s7, 56 ; GCN-NEXT: v_mov_b32_e32 v1, v0 ; GCN-NEXT: s_mov_b32 s2, 1 ; GCN-NEXT: s_cbranch_scc0 .LBB4_2 ; GCN-NEXT: ; %bb.1: ; %.one ; GCN-NEXT: v_add_u32_e32 v2, vcc, 1, v1 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dword v2, off, s[0:3], 0 ; GCN-NEXT: s_mov_b32 s2, 0 ; GCN-NEXT: .LBB4_2: ; %Flow ; GCN-NEXT: s_xor_b32 s2, s2, 1 ; GCN-NEXT: s_cmp_lg_u32 s2, 0 ; GCN-NEXT: s_cbranch_scc1 .LBB4_4 ; GCN-NEXT: ; %bb.3: ; %.zero ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dword v1, off, s[0:3], 0 ; GCN-NEXT: .LBB4_4: ; %.exit ; GCN-NEXT: s_endpgm %val = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %desc, i32 0, i32 0) %cmp = icmp eq i32 %val, 56 %tmp.0 = call i32 @llvm.amdgcn.set.inactive.i32(i32 %in, i32 42) #0 %tmp = call i32 @llvm.amdgcn.strict.wwm.i32(i32 %tmp.0) br i1 %cmp, label %.zero, label %.one .zero: store i32 %tmp, ptr addrspace(1) %out br label %.exit .one: %tmp.1 = add i32 %tmp, 1 store i32 %tmp.1, ptr addrspace(1) %out br label %.exit .exit: ret void } define amdgpu_kernel void @set_inactive_f32(ptr addrspace(1) %out, float %in) { ; GCN-LABEL: set_inactive_f32: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dword s6, s[4:5], 0x2c ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: s_or_saveexec_b64 s[2:3], -1 ; GCN-NEXT: v_mov_b32_e32 v0, 0x40400000 ; GCN-NEXT: s_mov_b64 exec, s[2:3] ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v1, s6 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, v0, v1, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v1, v0 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dword v1, off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call float @llvm.amdgcn.set.inactive.f32(float %in, float 3.0) #0 %tmp = call float @llvm.amdgcn.strict.wwm.f32(float %tmp.0) store float %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_f64(ptr addrspace(1) %out, double %in) { ; GCN-LABEL: set_inactive_f64: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_mov_b32_e32 v0, 0xcccccccd ; GCN-NEXT: v_mov_b32_e32 v1, 0x4010cccc ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v2, s2 ; GCN-NEXT: v_mov_b32_e32 v3, s3 ; GCN-NEXT: s_or_saveexec_b64 s[2:3], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, v0, v2, s[2:3] ; GCN-NEXT: v_cndmask_b32_e64 v1, v1, v3, s[2:3] ; GCN-NEXT: s_mov_b64 exec, s[2:3] ; GCN-NEXT: v_mov_b32_e32 v2, v0 ; GCN-NEXT: v_mov_b32_e32 v3, v1 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dwordx2 v[2:3], off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call double @llvm.amdgcn.set.inactive.f64(double %in, double 4.2) #0 %tmp = call double @llvm.amdgcn.strict.wwm.f64(double %tmp.0) store double %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_v2i16(ptr addrspace(1) %out, <2 x i16> %in) { ; GCN-LABEL: set_inactive_v2i16: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dword s6, s[4:5], 0x2c ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: s_or_saveexec_b64 s[2:3], -1 ; GCN-NEXT: v_mov_b32_e32 v0, 0x10001 ; GCN-NEXT: s_mov_b64 exec, s[2:3] ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v1, s6 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, v0, v1, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v1, v0 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dword v1, off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call <2 x i16> @llvm.amdgcn.set.inactive.v2i16(<2 x i16> %in, <2 x i16> ) #0 %tmp = call <2 x i16> @llvm.amdgcn.strict.wwm.v2i16(<2 x i16> %tmp.0) store <2 x i16> %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_v2f16(ptr addrspace(1) %out, <2 x half> %in) { ; GCN-LABEL: set_inactive_v2f16: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dword s6, s[4:5], 0x2c ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: s_or_saveexec_b64 s[2:3], -1 ; GCN-NEXT: v_mov_b32_e32 v0, 0x3c003c00 ; GCN-NEXT: s_mov_b64 exec, s[2:3] ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v1, s6 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, v0, v1, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v1, v0 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dword v1, off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call <2 x half> @llvm.amdgcn.set.inactive.v2f16(<2 x half> %in, <2 x half> ) #0 %tmp = call <2 x half> @llvm.amdgcn.strict.wwm.v2i16(<2 x half> %tmp.0) store <2 x half> %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_v2i32(ptr addrspace(1) %out, <2 x i32> %in) { ; GCN-LABEL: set_inactive_v2i32: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v2, s2 ; GCN-NEXT: v_mov_b32_e32 v3, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, 1, v2, s[4:5] ; GCN-NEXT: v_cndmask_b32_e64 v1, 1, v3, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v2, v0 ; GCN-NEXT: v_mov_b32_e32 v3, v1 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dwordx2 v[2:3], off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call <2 x i32> @llvm.amdgcn.set.inactive.v2i32(<2 x i32> %in, <2 x i32> ) #0 %tmp = call <2 x i32> @llvm.amdgcn.strict.wwm.v2i32(<2 x i32> %tmp.0) store <2 x i32> %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_v2f32(ptr addrspace(1) %out, <2 x float> %in) { ; GCN-LABEL: set_inactive_v2f32: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v2, s2 ; GCN-NEXT: v_mov_b32_e32 v3, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, 1.0, v2, s[4:5] ; GCN-NEXT: v_cndmask_b32_e64 v1, 1.0, v3, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v2, v0 ; GCN-NEXT: v_mov_b32_e32 v3, v1 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dwordx2 v[2:3], off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call <2 x float> @llvm.amdgcn.set.inactive.v2f32(<2 x float> %in, <2 x float> ) #0 %tmp = call <2 x float> @llvm.amdgcn.strict.wwm.v2f32(<2 x float> %tmp.0) store <2 x float> %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_v2bf16(ptr addrspace(1) %out, <2 x bfloat> %in) { ; GCN-LABEL: set_inactive_v2bf16: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dword s6, s[4:5], 0x2c ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v1, s6 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_mov_b32_e32 v0, 0x3f803f80 ; GCN-NEXT: v_cndmask_b32_e64 v0, v0, v1, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v1, v0 ; GCN-NEXT: buffer_store_dword v1, off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call <2 x bfloat> @llvm.amdgcn.set.inactive.v2bf16(<2 x bfloat> %in, <2 x bfloat> ) #0 %tmp = call <2 x bfloat> @llvm.amdgcn.strict.wwm.v2bf16(<2 x bfloat> %tmp.0) store <2 x bfloat> %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_v4i16(ptr addrspace(1) %out, <4 x i16> %in) { ; GCN-LABEL: set_inactive_v4i16: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_mov_b32_e32 v0, 0x10001 ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v3, s2 ; GCN-NEXT: v_mov_b32_e32 v4, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: v_cndmask_b32_e64 v1, v0, v3, s[4:5] ; GCN-NEXT: v_cndmask_b32_e64 v2, v0, v4, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v3, v1 ; GCN-NEXT: v_mov_b32_e32 v4, v2 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dwordx2 v[3:4], off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call <4 x i16> @llvm.amdgcn.set.inactive.v4i16(<4 x i16> %in, <4 x i16> ) #0 %tmp = call <4 x i16> @llvm.amdgcn.strict.wwm.v4i16(<4 x i16> %tmp.0) store <4 x i16> %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_v4f16(ptr addrspace(1) %out, <4 x half> %in) { ; GCN-LABEL: set_inactive_v4f16: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_mov_b32_e32 v0, 0x3c003c00 ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v3, s2 ; GCN-NEXT: v_mov_b32_e32 v4, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: v_cndmask_b32_e64 v1, v0, v3, s[4:5] ; GCN-NEXT: v_cndmask_b32_e64 v2, v0, v4, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v3, v1 ; GCN-NEXT: v_mov_b32_e32 v4, v2 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dwordx2 v[3:4], off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call <4 x half> @llvm.amdgcn.set.inactive.v4f16(<4 x half> %in, <4 x half> ) #0 %tmp = call <4 x half> @llvm.amdgcn.strict.wwm.v4f16(<4 x half> %tmp.0) store <4 x half> %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_v4bf16(ptr addrspace(1) %out, <4 x bfloat> %in) { ; GCN-LABEL: set_inactive_v4bf16: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24 ; GCN-NEXT: s_mov_b32 s7, 0xf000 ; GCN-NEXT: s_mov_b32 s6, -1 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: s_mov_b32 s4, s0 ; GCN-NEXT: s_mov_b32 s5, s1 ; GCN-NEXT: v_mov_b32_e32 v2, s3 ; GCN-NEXT: s_or_saveexec_b64 s[0:1], -1 ; GCN-NEXT: v_mov_b32_e32 v0, 0x3f803f80 ; GCN-NEXT: v_cndmask_b32_e64 v1, v0, v2, s[0:1] ; GCN-NEXT: s_mov_b64 exec, s[0:1] ; GCN-NEXT: v_mov_b32_e32 v2, s2 ; GCN-NEXT: s_or_saveexec_b64 s[0:1], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, v0, v2, s[0:1] ; GCN-NEXT: s_mov_b64 exec, s[0:1] ; GCN-NEXT: v_mov_b32_e32 v2, v0 ; GCN-NEXT: v_mov_b32_e32 v3, v1 ; GCN-NEXT: buffer_store_dwordx2 v[2:3], off, s[4:7], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call <4 x bfloat> @llvm.amdgcn.set.inactive.v4bf16(<4 x bfloat> %in, <4 x bfloat> ) #0 %tmp = call <4 x bfloat> @llvm.amdgcn.strict.wwm.v4bf16(<4 x bfloat> %tmp.0) store <4 x bfloat> %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_p0(ptr addrspace(1) %out, ptr %in) { ; GCN-LABEL: set_inactive_p0: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x24 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v2, s2 ; GCN-NEXT: v_mov_b32_e32 v3, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, 0, v2, s[4:5] ; GCN-NEXT: v_cndmask_b32_e64 v1, 0, v3, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v2, v0 ; GCN-NEXT: v_mov_b32_e32 v3, v1 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dwordx2 v[2:3], off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call ptr @llvm.amdgcn.set.inactive.p0(ptr %in, ptr null) #0 %tmp = call ptr @llvm.amdgcn.strict.wwm.p0(ptr %tmp.0) store ptr %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_p2(ptr addrspace(1) %out, ptr addrspace(2) %in) { ; GCN-LABEL: set_inactive_p2: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dword s3, s[4:5], 0x2c ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v1, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, 0, v1, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v1, v0 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dword v1, off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call ptr addrspace(2) @llvm.amdgcn.set.inactive.p2(ptr addrspace(2) %in, ptr addrspace(2) null) #0 %tmp = call ptr addrspace(2) @llvm.amdgcn.strict.wwm.p2(ptr addrspace(2) %tmp.0) store ptr addrspace(2) %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_p3(ptr addrspace(1) %out, ptr addrspace(3) %in) { ; GCN-LABEL: set_inactive_p3: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dword s3, s[4:5], 0x2c ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v1, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, 0, v1, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v1, v0 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dword v1, off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call ptr addrspace(3) @llvm.amdgcn.set.inactive.p3(ptr addrspace(3) %in, ptr addrspace(3) null) #0 %tmp = call ptr addrspace(3) @llvm.amdgcn.strict.wwm.p3(ptr addrspace(3) %tmp.0) store ptr addrspace(3) %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_p5(ptr addrspace(1) %out, ptr addrspace(5) %in) { ; GCN-LABEL: set_inactive_p5: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dword s3, s[4:5], 0x2c ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v1, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, 0, v1, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v1, v0 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dword v1, off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call ptr addrspace(5) @llvm.amdgcn.set.inactive.p5(ptr addrspace(5) %in, ptr addrspace(5) null) #0 %tmp = call ptr addrspace(5) @llvm.amdgcn.strict.wwm.p5(ptr addrspace(5) %tmp.0) store ptr addrspace(5) %tmp, ptr addrspace(1) %out ret void } define amdgpu_kernel void @set_inactive_p6(ptr addrspace(1) %out, ptr addrspace(6) %in) { ; GCN-LABEL: set_inactive_p6: ; GCN: ; %bb.0: ; GCN-NEXT: s_load_dword s3, s[4:5], 0x2c ; GCN-NEXT: s_load_dwordx2 s[0:1], s[4:5], 0x24 ; GCN-NEXT: s_mov_b32 s2, -1 ; GCN-NEXT: s_waitcnt lgkmcnt(0) ; GCN-NEXT: v_mov_b32_e32 v1, s3 ; GCN-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GCN-NEXT: v_cndmask_b32_e64 v0, 0, v1, s[4:5] ; GCN-NEXT: s_mov_b64 exec, s[4:5] ; GCN-NEXT: v_mov_b32_e32 v1, v0 ; GCN-NEXT: s_mov_b32 s3, 0xf000 ; GCN-NEXT: buffer_store_dword v1, off, s[0:3], 0 ; GCN-NEXT: s_endpgm %tmp.0 = call ptr addrspace(6) @llvm.amdgcn.set.inactive.p6(ptr addrspace(6) %in, ptr addrspace(6) null) #0 %tmp = call ptr addrspace(6) @llvm.amdgcn.strict.wwm.p6(ptr addrspace(6) %tmp.0) store ptr addrspace(6) %tmp, ptr addrspace(1) %out ret void } declare i32 @llvm.amdgcn.set.inactive.i32(i32, i32) #0 declare i64 @llvm.amdgcn.set.inactive.i64(i64, i64) #0 declare i32 @llvm.amdgcn.strict.wwm.i32(i32) #1 declare i64 @llvm.amdgcn.strict.wwm.i64(i64) #1 declare i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32>, i32, i32) attributes #0 = { convergent readnone } attributes #1 = { convergent nounwind readnone speculatable willreturn }