[llvm] 578a26a - [AMDGPU] Relax restrictions on amdgcn.cs.chain intrinsic (#169785)
via llvm-commits
llvm-commits at lists.llvm.org
Wed Dec 10 02:12:49 PST 2025
Author: Diana Picus
Date: 2025-12-10T11:12:46+01:00
New Revision: 578a26ada2e51e90733b50d048b0d74ac48ab21c
URL: https://github.com/llvm/llvm-project/commit/578a26ada2e51e90733b50d048b0d74ac48ab21c
DIFF: https://github.com/llvm/llvm-project/commit/578a26ada2e51e90733b50d048b0d74ac48ab21c.diff
LOG: [AMDGPU] Relax restrictions on amdgcn.cs.chain intrinsic (#169785)
We have a new use-case for chain functions, so slightly relax the
restriction on which calling conventions may contain calls to chain
functions.
Added:
Modified:
llvm/lib/IR/Verifier.cpp
llvm/test/CodeGen/AMDGPU/isel-amdgcn-cs-chain-intrinsic-w32.ll
llvm/test/Verifier/AMDGPU/intrinsic-amdgpu-cs-chain.ll
Removed:
################################################################################
diff --git a/llvm/lib/IR/Verifier.cpp b/llvm/lib/IR/Verifier.cpp
index 10e6c452680b0..040e6efe4c6e3 100644
--- a/llvm/lib/IR/Verifier.cpp
+++ b/llvm/lib/IR/Verifier.cpp
@@ -6761,11 +6761,15 @@ void Verifier::visitIntrinsicCall(Intrinsic::ID ID, CallBase &Call) {
case CallingConv::AMDGPU_CS:
case CallingConv::AMDGPU_CS_Chain:
case CallingConv::AMDGPU_CS_ChainPreserve:
+ case CallingConv::AMDGPU_ES:
+ case CallingConv::AMDGPU_GS:
+ case CallingConv::AMDGPU_HS:
+ case CallingConv::AMDGPU_LS:
+ case CallingConv::AMDGPU_VS:
break;
default:
- CheckFailed("Intrinsic can only be used from functions with the "
- "amdgpu_cs, amdgpu_cs_chain or amdgpu_cs_chain_preserve "
- "calling conventions",
+ CheckFailed("Intrinsic cannot be called from functions with this "
+ "calling convention",
&Call);
break;
}
diff --git a/llvm/test/CodeGen/AMDGPU/isel-amdgcn-cs-chain-intrinsic-w32.ll b/llvm/test/CodeGen/AMDGPU/isel-amdgcn-cs-chain-intrinsic-w32.ll
index 43ba2925914a0..d16d6a260821b 100644
--- a/llvm/test/CodeGen/AMDGPU/isel-amdgcn-cs-chain-intrinsic-w32.ll
+++ b/llvm/test/CodeGen/AMDGPU/isel-amdgcn-cs-chain-intrinsic-w32.ll
@@ -259,6 +259,636 @@ define amdgpu_cs void @cs_to_chain(<3 x i32> inreg %sgpr, { i32, ptr addrspace(5
unreachable
}
+define amdgpu_es void @es_to_chain(<3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr) {
+ ; GISEL-GFX11-LABEL: name: es_to_chain
+ ; GISEL-GFX11: bb.1 (%ir-block.0):
+ ; GISEL-GFX11-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; GISEL-GFX11-NEXT: {{ $}}
+ ; GISEL-GFX11-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+ ; GISEL-GFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
+ ; GISEL-GFX11-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
+ ; GISEL-GFX11-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; GISEL-GFX11-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; GISEL-GFX11-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; GISEL-GFX11-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; GISEL-GFX11-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; GISEL-GFX11-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY1]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; GISEL-GFX11-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; GISEL-GFX11-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; GISEL-GFX11-NEXT: $vgpr9 = COPY [[COPY4]]
+ ; GISEL-GFX11-NEXT: $vgpr10 = COPY [[COPY5]]
+ ; GISEL-GFX11-NEXT: $vgpr11 = COPY [[COPY6]]
+ ; GISEL-GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; GISEL-GFX11-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; GISEL-GFX11-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[S_MOV_B32_]], %subreg.sub0, [[S_MOV_B32_1]], %subreg.sub1
+ ; GISEL-GFX11-NEXT: [[COPY10:%[0-9]+]]:sgpr_64 = COPY [[REG_SEQUENCE]]
+ ; GISEL-GFX11-NEXT: SI_CS_CHAIN_TC_W32 [[COPY10]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ ;
+ ; GISEL-GFX10-LABEL: name: es_to_chain
+ ; GISEL-GFX10: bb.1 (%ir-block.0):
+ ; GISEL-GFX10-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; GISEL-GFX10-NEXT: {{ $}}
+ ; GISEL-GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+ ; GISEL-GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
+ ; GISEL-GFX10-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
+ ; GISEL-GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; GISEL-GFX10-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; GISEL-GFX10-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; GISEL-GFX10-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; GISEL-GFX10-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; GISEL-GFX10-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY1]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; GISEL-GFX10-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; GISEL-GFX10-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; GISEL-GFX10-NEXT: $vgpr9 = COPY [[COPY4]]
+ ; GISEL-GFX10-NEXT: $vgpr10 = COPY [[COPY5]]
+ ; GISEL-GFX10-NEXT: $vgpr11 = COPY [[COPY6]]
+ ; GISEL-GFX10-NEXT: [[COPY10:%[0-9]+]]:sgpr_128 = COPY $sgpr100_sgpr101_sgpr102_sgpr103
+ ; GISEL-GFX10-NEXT: $sgpr48_sgpr49_sgpr50_sgpr51 = COPY [[COPY10]]
+ ; GISEL-GFX10-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; GISEL-GFX10-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; GISEL-GFX10-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[S_MOV_B32_]], %subreg.sub0, [[S_MOV_B32_1]], %subreg.sub1
+ ; GISEL-GFX10-NEXT: [[COPY11:%[0-9]+]]:sgpr_64 = COPY [[REG_SEQUENCE]]
+ ; GISEL-GFX10-NEXT: SI_CS_CHAIN_TC_W32 [[COPY11]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $sgpr48_sgpr49_sgpr50_sgpr51
+ ;
+ ; DAGISEL-GFX11-LABEL: name: es_to_chain
+ ; DAGISEL-GFX11: bb.0 (%ir-block.0):
+ ; DAGISEL-GFX11-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; DAGISEL-GFX11-NEXT: {{ $}}
+ ; DAGISEL-GFX11-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; DAGISEL-GFX11-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; DAGISEL-GFX11-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; DAGISEL-GFX11-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; DAGISEL-GFX11-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr2
+ ; DAGISEL-GFX11-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr1
+ ; DAGISEL-GFX11-NEXT: [[COPY6:%[0-9]+]]:sgpr_32 = COPY $sgpr0
+ ; DAGISEL-GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; DAGISEL-GFX11-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; DAGISEL-GFX11-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE killed [[S_MOV_B32_1]], %subreg.sub0, killed [[S_MOV_B32_]], %subreg.sub1
+ ; DAGISEL-GFX11-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY4]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; DAGISEL-GFX11-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; DAGISEL-GFX11-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; DAGISEL-GFX11-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; DAGISEL-GFX11-NEXT: $vgpr9 = COPY [[COPY2]]
+ ; DAGISEL-GFX11-NEXT: $vgpr10 = COPY [[COPY1]]
+ ; DAGISEL-GFX11-NEXT: $vgpr11 = COPY [[COPY]]
+ ; DAGISEL-GFX11-NEXT: SI_CS_CHAIN_TC_W32 killed [[REG_SEQUENCE]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ ;
+ ; DAGISEL-GFX10-LABEL: name: es_to_chain
+ ; DAGISEL-GFX10: bb.0 (%ir-block.0):
+ ; DAGISEL-GFX10-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; DAGISEL-GFX10-NEXT: {{ $}}
+ ; DAGISEL-GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; DAGISEL-GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; DAGISEL-GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; DAGISEL-GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; DAGISEL-GFX10-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr2
+ ; DAGISEL-GFX10-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr1
+ ; DAGISEL-GFX10-NEXT: [[COPY6:%[0-9]+]]:sgpr_32 = COPY $sgpr0
+ ; DAGISEL-GFX10-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; DAGISEL-GFX10-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; DAGISEL-GFX10-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE killed [[S_MOV_B32_1]], %subreg.sub0, killed [[S_MOV_B32_]], %subreg.sub1
+ ; DAGISEL-GFX10-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY4]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY10:%[0-9]+]]:sgpr_128 = COPY $sgpr100_sgpr101_sgpr102_sgpr103
+ ; DAGISEL-GFX10-NEXT: $sgpr48_sgpr49_sgpr50_sgpr51 = COPY [[COPY10]]
+ ; DAGISEL-GFX10-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; DAGISEL-GFX10-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; DAGISEL-GFX10-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; DAGISEL-GFX10-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; DAGISEL-GFX10-NEXT: $vgpr9 = COPY [[COPY2]]
+ ; DAGISEL-GFX10-NEXT: $vgpr10 = COPY [[COPY1]]
+ ; DAGISEL-GFX10-NEXT: $vgpr11 = COPY [[COPY]]
+ ; DAGISEL-GFX10-NEXT: SI_CS_CHAIN_TC_W32 killed [[REG_SEQUENCE]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr48_sgpr49_sgpr50_sgpr51, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ call void(ptr, i32, <3 x i32>, { i32, ptr addrspace(5), i32, i32 }, i32, ...) @llvm.amdgcn.cs.chain(ptr @callee, i32 -1, <3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr, i32 0)
+ unreachable
+}
+
+define amdgpu_gs void @gs_to_chain(<3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr) {
+ ; GISEL-GFX11-LABEL: name: gs_to_chain
+ ; GISEL-GFX11: bb.1 (%ir-block.0):
+ ; GISEL-GFX11-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; GISEL-GFX11-NEXT: {{ $}}
+ ; GISEL-GFX11-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+ ; GISEL-GFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
+ ; GISEL-GFX11-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
+ ; GISEL-GFX11-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; GISEL-GFX11-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; GISEL-GFX11-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; GISEL-GFX11-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; GISEL-GFX11-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; GISEL-GFX11-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY1]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; GISEL-GFX11-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; GISEL-GFX11-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; GISEL-GFX11-NEXT: $vgpr9 = COPY [[COPY4]]
+ ; GISEL-GFX11-NEXT: $vgpr10 = COPY [[COPY5]]
+ ; GISEL-GFX11-NEXT: $vgpr11 = COPY [[COPY6]]
+ ; GISEL-GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; GISEL-GFX11-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; GISEL-GFX11-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[S_MOV_B32_]], %subreg.sub0, [[S_MOV_B32_1]], %subreg.sub1
+ ; GISEL-GFX11-NEXT: [[COPY10:%[0-9]+]]:sgpr_64 = COPY [[REG_SEQUENCE]]
+ ; GISEL-GFX11-NEXT: SI_CS_CHAIN_TC_W32 [[COPY10]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ ;
+ ; GISEL-GFX10-LABEL: name: gs_to_chain
+ ; GISEL-GFX10: bb.1 (%ir-block.0):
+ ; GISEL-GFX10-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; GISEL-GFX10-NEXT: {{ $}}
+ ; GISEL-GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+ ; GISEL-GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
+ ; GISEL-GFX10-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
+ ; GISEL-GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; GISEL-GFX10-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; GISEL-GFX10-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; GISEL-GFX10-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; GISEL-GFX10-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; GISEL-GFX10-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY1]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; GISEL-GFX10-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; GISEL-GFX10-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; GISEL-GFX10-NEXT: $vgpr9 = COPY [[COPY4]]
+ ; GISEL-GFX10-NEXT: $vgpr10 = COPY [[COPY5]]
+ ; GISEL-GFX10-NEXT: $vgpr11 = COPY [[COPY6]]
+ ; GISEL-GFX10-NEXT: [[COPY10:%[0-9]+]]:sgpr_128 = COPY $sgpr100_sgpr101_sgpr102_sgpr103
+ ; GISEL-GFX10-NEXT: $sgpr48_sgpr49_sgpr50_sgpr51 = COPY [[COPY10]]
+ ; GISEL-GFX10-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; GISEL-GFX10-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; GISEL-GFX10-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[S_MOV_B32_]], %subreg.sub0, [[S_MOV_B32_1]], %subreg.sub1
+ ; GISEL-GFX10-NEXT: [[COPY11:%[0-9]+]]:sgpr_64 = COPY [[REG_SEQUENCE]]
+ ; GISEL-GFX10-NEXT: SI_CS_CHAIN_TC_W32 [[COPY11]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $sgpr48_sgpr49_sgpr50_sgpr51
+ ;
+ ; DAGISEL-GFX11-LABEL: name: gs_to_chain
+ ; DAGISEL-GFX11: bb.0 (%ir-block.0):
+ ; DAGISEL-GFX11-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; DAGISEL-GFX11-NEXT: {{ $}}
+ ; DAGISEL-GFX11-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; DAGISEL-GFX11-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; DAGISEL-GFX11-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; DAGISEL-GFX11-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; DAGISEL-GFX11-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr2
+ ; DAGISEL-GFX11-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr1
+ ; DAGISEL-GFX11-NEXT: [[COPY6:%[0-9]+]]:sgpr_32 = COPY $sgpr0
+ ; DAGISEL-GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; DAGISEL-GFX11-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; DAGISEL-GFX11-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE killed [[S_MOV_B32_1]], %subreg.sub0, killed [[S_MOV_B32_]], %subreg.sub1
+ ; DAGISEL-GFX11-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY4]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; DAGISEL-GFX11-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; DAGISEL-GFX11-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; DAGISEL-GFX11-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; DAGISEL-GFX11-NEXT: $vgpr9 = COPY [[COPY2]]
+ ; DAGISEL-GFX11-NEXT: $vgpr10 = COPY [[COPY1]]
+ ; DAGISEL-GFX11-NEXT: $vgpr11 = COPY [[COPY]]
+ ; DAGISEL-GFX11-NEXT: SI_CS_CHAIN_TC_W32 killed [[REG_SEQUENCE]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ ;
+ ; DAGISEL-GFX10-LABEL: name: gs_to_chain
+ ; DAGISEL-GFX10: bb.0 (%ir-block.0):
+ ; DAGISEL-GFX10-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; DAGISEL-GFX10-NEXT: {{ $}}
+ ; DAGISEL-GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; DAGISEL-GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; DAGISEL-GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; DAGISEL-GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; DAGISEL-GFX10-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr2
+ ; DAGISEL-GFX10-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr1
+ ; DAGISEL-GFX10-NEXT: [[COPY6:%[0-9]+]]:sgpr_32 = COPY $sgpr0
+ ; DAGISEL-GFX10-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; DAGISEL-GFX10-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; DAGISEL-GFX10-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE killed [[S_MOV_B32_1]], %subreg.sub0, killed [[S_MOV_B32_]], %subreg.sub1
+ ; DAGISEL-GFX10-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY4]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY10:%[0-9]+]]:sgpr_128 = COPY $sgpr100_sgpr101_sgpr102_sgpr103
+ ; DAGISEL-GFX10-NEXT: $sgpr48_sgpr49_sgpr50_sgpr51 = COPY [[COPY10]]
+ ; DAGISEL-GFX10-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; DAGISEL-GFX10-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; DAGISEL-GFX10-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; DAGISEL-GFX10-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; DAGISEL-GFX10-NEXT: $vgpr9 = COPY [[COPY2]]
+ ; DAGISEL-GFX10-NEXT: $vgpr10 = COPY [[COPY1]]
+ ; DAGISEL-GFX10-NEXT: $vgpr11 = COPY [[COPY]]
+ ; DAGISEL-GFX10-NEXT: SI_CS_CHAIN_TC_W32 killed [[REG_SEQUENCE]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr48_sgpr49_sgpr50_sgpr51, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ call void(ptr, i32, <3 x i32>, { i32, ptr addrspace(5), i32, i32 }, i32, ...) @llvm.amdgcn.cs.chain(ptr @callee, i32 -1, <3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr, i32 0)
+ unreachable
+}
+
+define amdgpu_hs void @hs_to_chain(<3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr) {
+ ; GISEL-GFX11-LABEL: name: hs_to_chain
+ ; GISEL-GFX11: bb.1 (%ir-block.0):
+ ; GISEL-GFX11-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; GISEL-GFX11-NEXT: {{ $}}
+ ; GISEL-GFX11-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+ ; GISEL-GFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
+ ; GISEL-GFX11-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
+ ; GISEL-GFX11-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; GISEL-GFX11-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; GISEL-GFX11-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; GISEL-GFX11-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; GISEL-GFX11-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; GISEL-GFX11-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY1]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; GISEL-GFX11-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; GISEL-GFX11-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; GISEL-GFX11-NEXT: $vgpr9 = COPY [[COPY4]]
+ ; GISEL-GFX11-NEXT: $vgpr10 = COPY [[COPY5]]
+ ; GISEL-GFX11-NEXT: $vgpr11 = COPY [[COPY6]]
+ ; GISEL-GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; GISEL-GFX11-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; GISEL-GFX11-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[S_MOV_B32_]], %subreg.sub0, [[S_MOV_B32_1]], %subreg.sub1
+ ; GISEL-GFX11-NEXT: [[COPY10:%[0-9]+]]:sgpr_64 = COPY [[REG_SEQUENCE]]
+ ; GISEL-GFX11-NEXT: SI_CS_CHAIN_TC_W32 [[COPY10]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ ;
+ ; GISEL-GFX10-LABEL: name: hs_to_chain
+ ; GISEL-GFX10: bb.1 (%ir-block.0):
+ ; GISEL-GFX10-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; GISEL-GFX10-NEXT: {{ $}}
+ ; GISEL-GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+ ; GISEL-GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
+ ; GISEL-GFX10-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
+ ; GISEL-GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; GISEL-GFX10-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; GISEL-GFX10-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; GISEL-GFX10-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; GISEL-GFX10-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; GISEL-GFX10-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY1]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; GISEL-GFX10-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; GISEL-GFX10-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; GISEL-GFX10-NEXT: $vgpr9 = COPY [[COPY4]]
+ ; GISEL-GFX10-NEXT: $vgpr10 = COPY [[COPY5]]
+ ; GISEL-GFX10-NEXT: $vgpr11 = COPY [[COPY6]]
+ ; GISEL-GFX10-NEXT: [[COPY10:%[0-9]+]]:sgpr_128 = COPY $sgpr100_sgpr101_sgpr102_sgpr103
+ ; GISEL-GFX10-NEXT: $sgpr48_sgpr49_sgpr50_sgpr51 = COPY [[COPY10]]
+ ; GISEL-GFX10-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; GISEL-GFX10-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; GISEL-GFX10-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[S_MOV_B32_]], %subreg.sub0, [[S_MOV_B32_1]], %subreg.sub1
+ ; GISEL-GFX10-NEXT: [[COPY11:%[0-9]+]]:sgpr_64 = COPY [[REG_SEQUENCE]]
+ ; GISEL-GFX10-NEXT: SI_CS_CHAIN_TC_W32 [[COPY11]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $sgpr48_sgpr49_sgpr50_sgpr51
+ ;
+ ; DAGISEL-GFX11-LABEL: name: hs_to_chain
+ ; DAGISEL-GFX11: bb.0 (%ir-block.0):
+ ; DAGISEL-GFX11-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; DAGISEL-GFX11-NEXT: {{ $}}
+ ; DAGISEL-GFX11-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; DAGISEL-GFX11-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; DAGISEL-GFX11-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; DAGISEL-GFX11-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; DAGISEL-GFX11-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr2
+ ; DAGISEL-GFX11-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr1
+ ; DAGISEL-GFX11-NEXT: [[COPY6:%[0-9]+]]:sgpr_32 = COPY $sgpr0
+ ; DAGISEL-GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; DAGISEL-GFX11-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; DAGISEL-GFX11-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE killed [[S_MOV_B32_1]], %subreg.sub0, killed [[S_MOV_B32_]], %subreg.sub1
+ ; DAGISEL-GFX11-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY4]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; DAGISEL-GFX11-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; DAGISEL-GFX11-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; DAGISEL-GFX11-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; DAGISEL-GFX11-NEXT: $vgpr9 = COPY [[COPY2]]
+ ; DAGISEL-GFX11-NEXT: $vgpr10 = COPY [[COPY1]]
+ ; DAGISEL-GFX11-NEXT: $vgpr11 = COPY [[COPY]]
+ ; DAGISEL-GFX11-NEXT: SI_CS_CHAIN_TC_W32 killed [[REG_SEQUENCE]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ ;
+ ; DAGISEL-GFX10-LABEL: name: hs_to_chain
+ ; DAGISEL-GFX10: bb.0 (%ir-block.0):
+ ; DAGISEL-GFX10-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; DAGISEL-GFX10-NEXT: {{ $}}
+ ; DAGISEL-GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; DAGISEL-GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; DAGISEL-GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; DAGISEL-GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; DAGISEL-GFX10-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr2
+ ; DAGISEL-GFX10-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr1
+ ; DAGISEL-GFX10-NEXT: [[COPY6:%[0-9]+]]:sgpr_32 = COPY $sgpr0
+ ; DAGISEL-GFX10-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; DAGISEL-GFX10-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; DAGISEL-GFX10-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE killed [[S_MOV_B32_1]], %subreg.sub0, killed [[S_MOV_B32_]], %subreg.sub1
+ ; DAGISEL-GFX10-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY4]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY10:%[0-9]+]]:sgpr_128 = COPY $sgpr100_sgpr101_sgpr102_sgpr103
+ ; DAGISEL-GFX10-NEXT: $sgpr48_sgpr49_sgpr50_sgpr51 = COPY [[COPY10]]
+ ; DAGISEL-GFX10-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; DAGISEL-GFX10-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; DAGISEL-GFX10-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; DAGISEL-GFX10-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; DAGISEL-GFX10-NEXT: $vgpr9 = COPY [[COPY2]]
+ ; DAGISEL-GFX10-NEXT: $vgpr10 = COPY [[COPY1]]
+ ; DAGISEL-GFX10-NEXT: $vgpr11 = COPY [[COPY]]
+ ; DAGISEL-GFX10-NEXT: SI_CS_CHAIN_TC_W32 killed [[REG_SEQUENCE]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr48_sgpr49_sgpr50_sgpr51, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ call void(ptr, i32, <3 x i32>, { i32, ptr addrspace(5), i32, i32 }, i32, ...) @llvm.amdgcn.cs.chain(ptr @callee, i32 -1, <3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr, i32 0)
+ unreachable
+}
+
+define amdgpu_ls void @ls_to_chain(<3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr) {
+ ; GISEL-GFX11-LABEL: name: ls_to_chain
+ ; GISEL-GFX11: bb.1 (%ir-block.0):
+ ; GISEL-GFX11-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; GISEL-GFX11-NEXT: {{ $}}
+ ; GISEL-GFX11-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+ ; GISEL-GFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
+ ; GISEL-GFX11-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
+ ; GISEL-GFX11-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; GISEL-GFX11-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; GISEL-GFX11-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; GISEL-GFX11-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; GISEL-GFX11-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; GISEL-GFX11-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY1]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; GISEL-GFX11-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; GISEL-GFX11-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; GISEL-GFX11-NEXT: $vgpr9 = COPY [[COPY4]]
+ ; GISEL-GFX11-NEXT: $vgpr10 = COPY [[COPY5]]
+ ; GISEL-GFX11-NEXT: $vgpr11 = COPY [[COPY6]]
+ ; GISEL-GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; GISEL-GFX11-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; GISEL-GFX11-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[S_MOV_B32_]], %subreg.sub0, [[S_MOV_B32_1]], %subreg.sub1
+ ; GISEL-GFX11-NEXT: [[COPY10:%[0-9]+]]:sgpr_64 = COPY [[REG_SEQUENCE]]
+ ; GISEL-GFX11-NEXT: SI_CS_CHAIN_TC_W32 [[COPY10]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ ;
+ ; GISEL-GFX10-LABEL: name: ls_to_chain
+ ; GISEL-GFX10: bb.1 (%ir-block.0):
+ ; GISEL-GFX10-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; GISEL-GFX10-NEXT: {{ $}}
+ ; GISEL-GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+ ; GISEL-GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
+ ; GISEL-GFX10-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
+ ; GISEL-GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; GISEL-GFX10-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; GISEL-GFX10-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; GISEL-GFX10-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; GISEL-GFX10-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; GISEL-GFX10-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY1]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; GISEL-GFX10-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; GISEL-GFX10-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; GISEL-GFX10-NEXT: $vgpr9 = COPY [[COPY4]]
+ ; GISEL-GFX10-NEXT: $vgpr10 = COPY [[COPY5]]
+ ; GISEL-GFX10-NEXT: $vgpr11 = COPY [[COPY6]]
+ ; GISEL-GFX10-NEXT: [[COPY10:%[0-9]+]]:sgpr_128 = COPY $sgpr100_sgpr101_sgpr102_sgpr103
+ ; GISEL-GFX10-NEXT: $sgpr48_sgpr49_sgpr50_sgpr51 = COPY [[COPY10]]
+ ; GISEL-GFX10-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; GISEL-GFX10-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; GISEL-GFX10-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[S_MOV_B32_]], %subreg.sub0, [[S_MOV_B32_1]], %subreg.sub1
+ ; GISEL-GFX10-NEXT: [[COPY11:%[0-9]+]]:sgpr_64 = COPY [[REG_SEQUENCE]]
+ ; GISEL-GFX10-NEXT: SI_CS_CHAIN_TC_W32 [[COPY11]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $sgpr48_sgpr49_sgpr50_sgpr51
+ ;
+ ; DAGISEL-GFX11-LABEL: name: ls_to_chain
+ ; DAGISEL-GFX11: bb.0 (%ir-block.0):
+ ; DAGISEL-GFX11-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; DAGISEL-GFX11-NEXT: {{ $}}
+ ; DAGISEL-GFX11-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; DAGISEL-GFX11-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; DAGISEL-GFX11-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; DAGISEL-GFX11-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; DAGISEL-GFX11-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr2
+ ; DAGISEL-GFX11-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr1
+ ; DAGISEL-GFX11-NEXT: [[COPY6:%[0-9]+]]:sgpr_32 = COPY $sgpr0
+ ; DAGISEL-GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; DAGISEL-GFX11-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; DAGISEL-GFX11-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE killed [[S_MOV_B32_1]], %subreg.sub0, killed [[S_MOV_B32_]], %subreg.sub1
+ ; DAGISEL-GFX11-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY4]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; DAGISEL-GFX11-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; DAGISEL-GFX11-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; DAGISEL-GFX11-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; DAGISEL-GFX11-NEXT: $vgpr9 = COPY [[COPY2]]
+ ; DAGISEL-GFX11-NEXT: $vgpr10 = COPY [[COPY1]]
+ ; DAGISEL-GFX11-NEXT: $vgpr11 = COPY [[COPY]]
+ ; DAGISEL-GFX11-NEXT: SI_CS_CHAIN_TC_W32 killed [[REG_SEQUENCE]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ ;
+ ; DAGISEL-GFX10-LABEL: name: ls_to_chain
+ ; DAGISEL-GFX10: bb.0 (%ir-block.0):
+ ; DAGISEL-GFX10-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; DAGISEL-GFX10-NEXT: {{ $}}
+ ; DAGISEL-GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; DAGISEL-GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; DAGISEL-GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; DAGISEL-GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; DAGISEL-GFX10-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr2
+ ; DAGISEL-GFX10-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr1
+ ; DAGISEL-GFX10-NEXT: [[COPY6:%[0-9]+]]:sgpr_32 = COPY $sgpr0
+ ; DAGISEL-GFX10-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; DAGISEL-GFX10-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; DAGISEL-GFX10-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE killed [[S_MOV_B32_1]], %subreg.sub0, killed [[S_MOV_B32_]], %subreg.sub1
+ ; DAGISEL-GFX10-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY4]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY10:%[0-9]+]]:sgpr_128 = COPY $sgpr100_sgpr101_sgpr102_sgpr103
+ ; DAGISEL-GFX10-NEXT: $sgpr48_sgpr49_sgpr50_sgpr51 = COPY [[COPY10]]
+ ; DAGISEL-GFX10-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; DAGISEL-GFX10-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; DAGISEL-GFX10-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; DAGISEL-GFX10-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; DAGISEL-GFX10-NEXT: $vgpr9 = COPY [[COPY2]]
+ ; DAGISEL-GFX10-NEXT: $vgpr10 = COPY [[COPY1]]
+ ; DAGISEL-GFX10-NEXT: $vgpr11 = COPY [[COPY]]
+ ; DAGISEL-GFX10-NEXT: SI_CS_CHAIN_TC_W32 killed [[REG_SEQUENCE]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr48_sgpr49_sgpr50_sgpr51, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ call void(ptr, i32, <3 x i32>, { i32, ptr addrspace(5), i32, i32 }, i32, ...) @llvm.amdgcn.cs.chain(ptr @callee, i32 -1, <3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr, i32 0)
+ unreachable
+}
+
+define amdgpu_vs void @vs_to_chain(<3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr) {
+ ; GISEL-GFX11-LABEL: name: vs_to_chain
+ ; GISEL-GFX11: bb.1 (%ir-block.0):
+ ; GISEL-GFX11-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; GISEL-GFX11-NEXT: {{ $}}
+ ; GISEL-GFX11-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+ ; GISEL-GFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
+ ; GISEL-GFX11-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
+ ; GISEL-GFX11-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; GISEL-GFX11-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; GISEL-GFX11-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; GISEL-GFX11-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; GISEL-GFX11-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; GISEL-GFX11-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY1]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; GISEL-GFX11-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
+ ; GISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; GISEL-GFX11-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; GISEL-GFX11-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; GISEL-GFX11-NEXT: $vgpr9 = COPY [[COPY4]]
+ ; GISEL-GFX11-NEXT: $vgpr10 = COPY [[COPY5]]
+ ; GISEL-GFX11-NEXT: $vgpr11 = COPY [[COPY6]]
+ ; GISEL-GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; GISEL-GFX11-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; GISEL-GFX11-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[S_MOV_B32_]], %subreg.sub0, [[S_MOV_B32_1]], %subreg.sub1
+ ; GISEL-GFX11-NEXT: [[COPY10:%[0-9]+]]:sgpr_64 = COPY [[REG_SEQUENCE]]
+ ; GISEL-GFX11-NEXT: SI_CS_CHAIN_TC_W32 [[COPY10]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ ;
+ ; GISEL-GFX10-LABEL: name: vs_to_chain
+ ; GISEL-GFX10: bb.1 (%ir-block.0):
+ ; GISEL-GFX10-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; GISEL-GFX10-NEXT: {{ $}}
+ ; GISEL-GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+ ; GISEL-GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
+ ; GISEL-GFX10-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
+ ; GISEL-GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; GISEL-GFX10-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; GISEL-GFX10-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; GISEL-GFX10-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; GISEL-GFX10-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; GISEL-GFX10-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY1]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; GISEL-GFX10-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
+ ; GISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; GISEL-GFX10-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; GISEL-GFX10-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; GISEL-GFX10-NEXT: $vgpr9 = COPY [[COPY4]]
+ ; GISEL-GFX10-NEXT: $vgpr10 = COPY [[COPY5]]
+ ; GISEL-GFX10-NEXT: $vgpr11 = COPY [[COPY6]]
+ ; GISEL-GFX10-NEXT: [[COPY10:%[0-9]+]]:sgpr_128 = COPY $sgpr100_sgpr101_sgpr102_sgpr103
+ ; GISEL-GFX10-NEXT: $sgpr48_sgpr49_sgpr50_sgpr51 = COPY [[COPY10]]
+ ; GISEL-GFX10-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; GISEL-GFX10-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; GISEL-GFX10-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[S_MOV_B32_]], %subreg.sub0, [[S_MOV_B32_1]], %subreg.sub1
+ ; GISEL-GFX10-NEXT: [[COPY11:%[0-9]+]]:sgpr_64 = COPY [[REG_SEQUENCE]]
+ ; GISEL-GFX10-NEXT: SI_CS_CHAIN_TC_W32 [[COPY11]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $sgpr48_sgpr49_sgpr50_sgpr51
+ ;
+ ; DAGISEL-GFX11-LABEL: name: vs_to_chain
+ ; DAGISEL-GFX11: bb.0 (%ir-block.0):
+ ; DAGISEL-GFX11-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; DAGISEL-GFX11-NEXT: {{ $}}
+ ; DAGISEL-GFX11-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; DAGISEL-GFX11-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; DAGISEL-GFX11-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; DAGISEL-GFX11-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; DAGISEL-GFX11-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr2
+ ; DAGISEL-GFX11-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr1
+ ; DAGISEL-GFX11-NEXT: [[COPY6:%[0-9]+]]:sgpr_32 = COPY $sgpr0
+ ; DAGISEL-GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; DAGISEL-GFX11-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; DAGISEL-GFX11-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE killed [[S_MOV_B32_1]], %subreg.sub0, killed [[S_MOV_B32_]], %subreg.sub1
+ ; DAGISEL-GFX11-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY4]]
+ ; DAGISEL-GFX11-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; DAGISEL-GFX11-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; DAGISEL-GFX11-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; DAGISEL-GFX11-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; DAGISEL-GFX11-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; DAGISEL-GFX11-NEXT: $vgpr9 = COPY [[COPY2]]
+ ; DAGISEL-GFX11-NEXT: $vgpr10 = COPY [[COPY1]]
+ ; DAGISEL-GFX11-NEXT: $vgpr11 = COPY [[COPY]]
+ ; DAGISEL-GFX11-NEXT: SI_CS_CHAIN_TC_W32 killed [[REG_SEQUENCE]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ ;
+ ; DAGISEL-GFX10-LABEL: name: vs_to_chain
+ ; DAGISEL-GFX10: bb.0 (%ir-block.0):
+ ; DAGISEL-GFX10-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
+ ; DAGISEL-GFX10-NEXT: {{ $}}
+ ; DAGISEL-GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr3
+ ; DAGISEL-GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
+ ; DAGISEL-GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+ ; DAGISEL-GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+ ; DAGISEL-GFX10-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr2
+ ; DAGISEL-GFX10-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr1
+ ; DAGISEL-GFX10-NEXT: [[COPY6:%[0-9]+]]:sgpr_32 = COPY $sgpr0
+ ; DAGISEL-GFX10-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-hi) @callee
+ ; DAGISEL-GFX10-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @callee
+ ; DAGISEL-GFX10-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE killed [[S_MOV_B32_1]], %subreg.sub0, killed [[S_MOV_B32_]], %subreg.sub1
+ ; DAGISEL-GFX10-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[COPY4]]
+ ; DAGISEL-GFX10-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32_xm0 = V_READFIRSTLANE_B32 [[COPY9]], implicit $exec
+ ; DAGISEL-GFX10-NEXT: [[COPY10:%[0-9]+]]:sgpr_128 = COPY $sgpr100_sgpr101_sgpr102_sgpr103
+ ; DAGISEL-GFX10-NEXT: $sgpr48_sgpr49_sgpr50_sgpr51 = COPY [[COPY10]]
+ ; DAGISEL-GFX10-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
+ ; DAGISEL-GFX10-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
+ ; DAGISEL-GFX10-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
+ ; DAGISEL-GFX10-NEXT: $vgpr8 = COPY [[COPY3]]
+ ; DAGISEL-GFX10-NEXT: $vgpr9 = COPY [[COPY2]]
+ ; DAGISEL-GFX10-NEXT: $vgpr10 = COPY [[COPY1]]
+ ; DAGISEL-GFX10-NEXT: $vgpr11 = COPY [[COPY]]
+ ; DAGISEL-GFX10-NEXT: SI_CS_CHAIN_TC_W32 killed [[REG_SEQUENCE]], @callee, 0, -1, amdgpu_allvgprs, implicit $sgpr48_sgpr49_sgpr50_sgpr51, implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11
+ call void(ptr, i32, <3 x i32>, { i32, ptr addrspace(5), i32, i32 }, i32, ...) @llvm.amdgcn.cs.chain(ptr @callee, i32 -1, <3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr, i32 0)
+ unreachable
+}
+
define amdgpu_cs_chain void @chain_to_chain_preserve(<3 x i32> inreg %sgpr, { i32, ptr addrspace(5), i32, i32 } %vgpr) {
; GISEL-GFX11-LABEL: name: chain_to_chain_preserve
; GISEL-GFX11: bb.1 (%ir-block.0):
diff --git a/llvm/test/Verifier/AMDGPU/intrinsic-amdgpu-cs-chain.ll b/llvm/test/Verifier/AMDGPU/intrinsic-amdgpu-cs-chain.ll
index 6a99f77cf63a1..7e22ab21779e4 100644
--- a/llvm/test/Verifier/AMDGPU/intrinsic-amdgpu-cs-chain.ll
+++ b/llvm/test/Verifier/AMDGPU/intrinsic-amdgpu-cs-chain.ll
@@ -44,7 +44,7 @@ define void @bad_caller_default_cc(ptr %fn, i32 %exec, <4 x i32> inreg %sgpr, {
; CHECK-NEXT: @llvm.amdgcn.set.inactive.chain.arg
%unused = call i32 @llvm.amdgcn.set.inactive.chain.arg(i32 0, i32 1)
- ; CHECK: Intrinsic can only be used from functions with the amdgpu_cs, amdgpu_cs_chain or amdgpu_cs_chain_preserve calling conventions
+ ; CHECK: Intrinsic cannot be called from functions with this calling convention
; CHECK-NEXT: @llvm.amdgcn.cs.chain
call void(ptr, i32, <4 x i32>, { ptr, <3 x i32> }, i32, ...) @llvm.amdgcn.cs.chain(ptr %fn, i32 %exec, <4 x i32> %sgpr, { ptr, <3 x i32> } %vgpr, i32 0)
unreachable
@@ -55,7 +55,7 @@ define amdgpu_kernel void @bad_caller_amdgpu_kernel(ptr %fn, i32 %exec, <4 x i32
; CHECK-NEXT: @llvm.amdgcn.set.inactive.chain.arg
%unused = call i32 @llvm.amdgcn.set.inactive.chain.arg(i32 0, i32 1)
- ; CHECK: Intrinsic can only be used from functions with the amdgpu_cs, amdgpu_cs_chain or amdgpu_cs_chain_preserve calling conventions
+ ; CHECK: Intrinsic cannot be called from functions with this calling convention
; CHECK-NEXT: @llvm.amdgcn.cs.chain
call void(ptr, i32, <4 x i32>, { ptr, <3 x i32> }, i32, ...) @llvm.amdgcn.cs.chain(ptr %fn, i32 %exec, <4 x i32> %sgpr, { ptr, <3 x i32> } %vgpr, i32 0)
unreachable
@@ -66,7 +66,7 @@ define amdgpu_gfx void @bad_caller_amdgpu_gfx(ptr %fn, i32 %exec, <4 x i32> inre
; CHECK-NEXT: @llvm.amdgcn.set.inactive.chain.arg
%unused = call i32 @llvm.amdgcn.set.inactive.chain.arg(i32 0, i32 1)
- ; CHECK: Intrinsic can only be used from functions with the amdgpu_cs, amdgpu_cs_chain or amdgpu_cs_chain_preserve calling conventions
+ ; CHECK: Intrinsic cannot be called from functions with this calling convention
; CHECK-NEXT: @llvm.amdgcn.cs.chain
call void(ptr, i32, <4 x i32>, { ptr, <3 x i32> }, i32, ...) @llvm.amdgcn.cs.chain(ptr %fn, i32 %exec, <4 x i32> %sgpr, { ptr, <3 x i32> } %vgpr, i32 0)
unreachable
@@ -77,10 +77,8 @@ define amdgpu_vs void @bad_caller_amdgpu_vs(ptr %fn, i32 %exec, <4 x i32> inreg
; CHECK-NEXT: @llvm.amdgcn.set.inactive.chain.arg
%unused = call i32 @llvm.amdgcn.set.inactive.chain.arg(i32 0, i32 1)
- ; CHECK: Intrinsic can only be used from functions with the amdgpu_cs, amdgpu_cs_chain or amdgpu_cs_chain_preserve calling conventions
- ; CHECK-NEXT: @llvm.amdgcn.cs.chain
- call void(ptr, i32, <4 x i32>, { ptr, <3 x i32> }, i32, ...) @llvm.amdgcn.cs.chain(ptr %fn, i32 %exec, <4 x i32> %sgpr, { ptr, <3 x i32> } %vgpr, i32 0)
- unreachable
+ ; Unlike llvm.amdgcn.set.inactive.chain.arg, llvm.amdgcn.cs.chain may be called from amdgpu_vs functions.
+ ret void
}
define amdgpu_cs void @bad_caller_amdgpu_cs(ptr %fn, i32 %exec, <4 x i32> inreg %sgpr, { ptr, <3 x i32> } %vgpr) {
More information about the llvm-commits
mailing list