diff options
Diffstat (limited to 'contrib/llvm/lib/Target/AMDGPU')
12 files changed, 119 insertions, 11 deletions
diff --git a/contrib/llvm/lib/Target/AMDGPU/AMDGPU.h b/contrib/llvm/lib/Target/AMDGPU/AMDGPU.h index bb7801c172f6..55668867cc8e 100644 --- a/contrib/llvm/lib/Target/AMDGPU/AMDGPU.h +++ b/contrib/llvm/lib/Target/AMDGPU/AMDGPU.h @@ -254,7 +254,7 @@ namespace AMDGPUAS { FLAT_ADDRESS = 0, ///< Address space for flat memory. GLOBAL_ADDRESS = 1, ///< Address space for global memory (RAT0, VTX0). - REGION_ADDRESS = 2, ///< Address space for region memory. + REGION_ADDRESS = 2, ///< Address space for region memory. (GDS) CONSTANT_ADDRESS = 4, ///< Address space for constant memory (VTX2) LOCAL_ADDRESS = 3, ///< Address space for local memory. diff --git a/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp b/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp index 6951c915b177..8d36511a2830 100644 --- a/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp +++ b/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp @@ -4192,6 +4192,7 @@ const char* AMDGPUTargetLowering::getTargetNodeName(unsigned Opcode) const { NODE_NAME_CASE(TBUFFER_STORE_FORMAT_D16) NODE_NAME_CASE(TBUFFER_LOAD_FORMAT) NODE_NAME_CASE(TBUFFER_LOAD_FORMAT_D16) + NODE_NAME_CASE(DS_ORDERED_COUNT) NODE_NAME_CASE(ATOMIC_CMP_SWAP) NODE_NAME_CASE(ATOMIC_INC) NODE_NAME_CASE(ATOMIC_DEC) diff --git a/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.h b/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.h index 0d22cb2e3e20..d4a751d00a50 100644 --- a/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.h +++ b/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.h @@ -474,6 +474,7 @@ enum NodeType : unsigned { TBUFFER_STORE_FORMAT_D16, TBUFFER_LOAD_FORMAT, TBUFFER_LOAD_FORMAT_D16, + DS_ORDERED_COUNT, ATOMIC_CMP_SWAP, ATOMIC_INC, ATOMIC_DEC, diff --git a/contrib/llvm/lib/Target/AMDGPU/AMDGPUSearchableTables.td b/contrib/llvm/lib/Target/AMDGPU/AMDGPUSearchableTables.td index 9dbd7751b4d8..4d0962f65fdc 100644 --- a/contrib/llvm/lib/Target/AMDGPU/AMDGPUSearchableTables.td +++ b/contrib/llvm/lib/Target/AMDGPU/AMDGPUSearchableTables.td @@ -72,6 +72,8 @@ def : SourceOfDivergence<int_amdgcn_buffer_atomic_xor>; def : SourceOfDivergence<int_amdgcn_buffer_atomic_cmpswap>; def : SourceOfDivergence<int_amdgcn_ps_live>; def : SourceOfDivergence<int_amdgcn_ds_swizzle>; +def : SourceOfDivergence<int_amdgcn_ds_ordered_add>; +def : SourceOfDivergence<int_amdgcn_ds_ordered_swap>; foreach intr = AMDGPUImageDimAtomicIntrinsics in def : SourceOfDivergence<intr>; diff --git a/contrib/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp b/contrib/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp index 11e4ba4b5010..62e7e44ddb80 100644 --- a/contrib/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp +++ b/contrib/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp @@ -308,6 +308,8 @@ bool GCNTTIImpl::getTgtMemIntrinsic(IntrinsicInst *Inst, switch (Inst->getIntrinsicID()) { case Intrinsic::amdgcn_atomic_inc: case Intrinsic::amdgcn_atomic_dec: + case Intrinsic::amdgcn_ds_ordered_add: + case Intrinsic::amdgcn_ds_ordered_swap: case Intrinsic::amdgcn_ds_fadd: case Intrinsic::amdgcn_ds_fmin: case Intrinsic::amdgcn_ds_fmax: { diff --git a/contrib/llvm/lib/Target/AMDGPU/DSInstructions.td b/contrib/llvm/lib/Target/AMDGPU/DSInstructions.td index 31d2ebef481d..9c7097e9a520 100644 --- a/contrib/llvm/lib/Target/AMDGPU/DSInstructions.td +++ b/contrib/llvm/lib/Target/AMDGPU/DSInstructions.td @@ -817,6 +817,11 @@ defm : DSAtomicRetPat_mc<DS_MAX_RTN_U64, i64, "atomic_load_umax_local">; defm : DSAtomicCmpXChg_mc<DS_CMPST_RTN_B64, i64, "atomic_cmp_swap_local">; +def : Pat < + (SIds_ordered_count i32:$value, i16:$offset), + (DS_ORDERED_COUNT $value, (as_i16imm $offset)) +>; + //===----------------------------------------------------------------------===// // Real instructions //===----------------------------------------------------------------------===// diff --git a/contrib/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp b/contrib/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp index c6396de89c4f..69ddbfb53958 100644 --- a/contrib/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp +++ b/contrib/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp @@ -88,14 +88,28 @@ static bool isSMovRel(unsigned Opcode) { } } -static bool isSendMsgTraceDataOrGDS(const MachineInstr &MI) { +static bool isSendMsgTraceDataOrGDS(const SIInstrInfo &TII, + const MachineInstr &MI) { + if (TII.isAlwaysGDS(MI.getOpcode())) + return true; + switch (MI.getOpcode()) { case AMDGPU::S_SENDMSG: case AMDGPU::S_SENDMSGHALT: case AMDGPU::S_TTRACEDATA: return true; + // These DS opcodes don't support GDS. + case AMDGPU::DS_NOP: + case AMDGPU::DS_PERMUTE_B32: + case AMDGPU::DS_BPERMUTE_B32: + return false; default: - // TODO: GDS + if (TII.isDS(MI.getOpcode())) { + int GDS = AMDGPU::getNamedOperandIdx(MI.getOpcode(), + AMDGPU::OpName::gds); + if (MI.getOperand(GDS).getImm()) + return true; + } return false; } } @@ -145,7 +159,7 @@ GCNHazardRecognizer::getHazardType(SUnit *SU, int Stalls) { checkReadM0Hazards(MI) > 0) return NoopHazard; - if (ST.hasReadM0SendMsgHazard() && isSendMsgTraceDataOrGDS(*MI) && + if (ST.hasReadM0SendMsgHazard() && isSendMsgTraceDataOrGDS(TII, *MI) && checkReadM0Hazards(MI) > 0) return NoopHazard; @@ -199,7 +213,7 @@ unsigned GCNHazardRecognizer::PreEmitNoops(MachineInstr *MI) { isSMovRel(MI->getOpcode()))) return std::max(WaitStates, checkReadM0Hazards(MI)); - if (ST.hasReadM0SendMsgHazard() && isSendMsgTraceDataOrGDS(*MI)) + if (ST.hasReadM0SendMsgHazard() && isSendMsgTraceDataOrGDS(TII, *MI)) return std::max(WaitStates, checkReadM0Hazards(MI)); return WaitStates; diff --git a/contrib/llvm/lib/Target/AMDGPU/SIISelLowering.cpp b/contrib/llvm/lib/Target/AMDGPU/SIISelLowering.cpp index 0ba921647097..12113fcc1fcb 100644 --- a/contrib/llvm/lib/Target/AMDGPU/SIISelLowering.cpp +++ b/contrib/llvm/lib/Target/AMDGPU/SIISelLowering.cpp @@ -910,6 +910,8 @@ bool SITargetLowering::getTgtMemIntrinsic(IntrinsicInfo &Info, switch (IntrID) { case Intrinsic::amdgcn_atomic_inc: case Intrinsic::amdgcn_atomic_dec: + case Intrinsic::amdgcn_ds_ordered_add: + case Intrinsic::amdgcn_ds_ordered_swap: case Intrinsic::amdgcn_ds_fadd: case Intrinsic::amdgcn_ds_fmin: case Intrinsic::amdgcn_ds_fmax: { @@ -937,6 +939,8 @@ bool SITargetLowering::getAddrModeArguments(IntrinsicInst *II, switch (II->getIntrinsicID()) { case Intrinsic::amdgcn_atomic_inc: case Intrinsic::amdgcn_atomic_dec: + case Intrinsic::amdgcn_ds_ordered_add: + case Intrinsic::amdgcn_ds_ordered_swap: case Intrinsic::amdgcn_ds_fadd: case Intrinsic::amdgcn_ds_fmin: case Intrinsic::amdgcn_ds_fmax: { @@ -5438,6 +5442,63 @@ SDValue SITargetLowering::LowerINTRINSIC_W_CHAIN(SDValue Op, SDLoc DL(Op); switch (IntrID) { + case Intrinsic::amdgcn_ds_ordered_add: + case Intrinsic::amdgcn_ds_ordered_swap: { + MemSDNode *M = cast<MemSDNode>(Op); + SDValue Chain = M->getOperand(0); + SDValue M0 = M->getOperand(2); + SDValue Value = M->getOperand(3); + unsigned OrderedCountIndex = M->getConstantOperandVal(7); + unsigned WaveRelease = M->getConstantOperandVal(8); + unsigned WaveDone = M->getConstantOperandVal(9); + unsigned ShaderType; + unsigned Instruction; + + switch (IntrID) { + case Intrinsic::amdgcn_ds_ordered_add: + Instruction = 0; + break; + case Intrinsic::amdgcn_ds_ordered_swap: + Instruction = 1; + break; + } + + if (WaveDone && !WaveRelease) + report_fatal_error("ds_ordered_count: wave_done requires wave_release"); + + switch (DAG.getMachineFunction().getFunction().getCallingConv()) { + case CallingConv::AMDGPU_CS: + case CallingConv::AMDGPU_KERNEL: + ShaderType = 0; + break; + case CallingConv::AMDGPU_PS: + ShaderType = 1; + break; + case CallingConv::AMDGPU_VS: + ShaderType = 2; + break; + case CallingConv::AMDGPU_GS: + ShaderType = 3; + break; + default: + report_fatal_error("ds_ordered_count unsupported for this calling conv"); + } + + unsigned Offset0 = OrderedCountIndex << 2; + unsigned Offset1 = WaveRelease | (WaveDone << 1) | (ShaderType << 2) | + (Instruction << 4); + unsigned Offset = Offset0 | (Offset1 << 8); + + SDValue Ops[] = { + Chain, + Value, + DAG.getTargetConstant(Offset, DL, MVT::i16), + copyToM0(DAG, Chain, DL, M0).getValue(1), // Glue + }; + return DAG.getMemIntrinsicNode(AMDGPUISD::DS_ORDERED_COUNT, DL, + M->getVTList(), Ops, M->getMemoryVT(), + M->getMemOperand()); + } case Intrinsic::amdgcn_atomic_inc: case Intrinsic::amdgcn_atomic_dec: case Intrinsic::amdgcn_ds_fadd: diff --git a/contrib/llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp b/contrib/llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp index afc0b4467610..3c13bccd94fa 100644 --- a/contrib/llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp +++ b/contrib/llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp @@ -536,10 +536,13 @@ void WaitcntBrackets::updateByEvent(const SIInstrInfo *TII, CurrScore); } if (Inst.mayStore()) { - setExpScore( - &Inst, TII, TRI, MRI, - AMDGPU::getNamedOperandIdx(Inst.getOpcode(), AMDGPU::OpName::data0), - CurrScore); + if (AMDGPU::getNamedOperandIdx(Inst.getOpcode(), + AMDGPU::OpName::data0) != -1) { + setExpScore( + &Inst, TII, TRI, MRI, + AMDGPU::getNamedOperandIdx(Inst.getOpcode(), AMDGPU::OpName::data0), + CurrScore); + } if (AMDGPU::getNamedOperandIdx(Inst.getOpcode(), AMDGPU::OpName::data1) != -1) { setExpScore(&Inst, TII, TRI, MRI, @@ -1093,7 +1096,8 @@ void SIInsertWaitcnts::updateEventWaitcntAfter(MachineInstr &Inst, // bracket and the destination operand scores. // TODO: Use the (TSFlags & SIInstrFlags::LGKM_CNT) property everywhere. if (TII->isDS(Inst) && TII->usesLGKM_CNT(Inst)) { - if (TII->hasModifiersSet(Inst, AMDGPU::OpName::gds)) { + if (TII->isAlwaysGDS(Inst.getOpcode()) || + TII->hasModifiersSet(Inst, AMDGPU::OpName::gds)) { ScoreBrackets->updateByEvent(TII, TRI, MRI, GDS_ACCESS, Inst); ScoreBrackets->updateByEvent(TII, TRI, MRI, GDS_GPR_LOCK, Inst); } else { diff --git a/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp b/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp index 2370d5fa7b27..7f7f1807987a 100644 --- a/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp +++ b/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp @@ -2390,6 +2390,16 @@ bool SIInstrInfo::isSchedulingBoundary(const MachineInstr &MI, changesVGPRIndexingMode(MI); } +bool SIInstrInfo::isAlwaysGDS(uint16_t Opcode) const { + return Opcode == AMDGPU::DS_ORDERED_COUNT || + Opcode == AMDGPU::DS_GWS_INIT || + Opcode == AMDGPU::DS_GWS_SEMA_V || + Opcode == AMDGPU::DS_GWS_SEMA_BR || + Opcode == AMDGPU::DS_GWS_SEMA_P || + Opcode == AMDGPU::DS_GWS_SEMA_RELEASE_ALL || + Opcode == AMDGPU::DS_GWS_BARRIER; +} + bool SIInstrInfo::hasUnwantedEffectsWhenEXECEmpty(const MachineInstr &MI) const { unsigned Opcode = MI.getOpcode(); @@ -2403,7 +2413,8 @@ bool SIInstrInfo::hasUnwantedEffectsWhenEXECEmpty(const MachineInstr &MI) const // EXEC = 0, but checking for that case here seems not worth it // given the typical code patterns. if (Opcode == AMDGPU::S_SENDMSG || Opcode == AMDGPU::S_SENDMSGHALT || - Opcode == AMDGPU::EXP || Opcode == AMDGPU::EXP_DONE) + Opcode == AMDGPU::EXP || Opcode == AMDGPU::EXP_DONE || + Opcode == AMDGPU::DS_ORDERED_COUNT) return true; if (MI.isInlineAsm()) diff --git a/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.h b/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.h index 5b1a05f3785e..8847fd6babb3 100644 --- a/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.h +++ b/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.h @@ -450,6 +450,8 @@ public: return get(Opcode).TSFlags & SIInstrFlags::DS; } + bool isAlwaysGDS(uint16_t Opcode) const; + static bool isMIMG(const MachineInstr &MI) { return MI.getDesc().TSFlags & SIInstrFlags::MIMG; } diff --git a/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.td b/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.td index 13afa4d4974b..180a7b0601d7 100644 --- a/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.td +++ b/contrib/llvm/lib/Target/AMDGPU/SIInstrInfo.td @@ -45,6 +45,11 @@ def SIsbuffer_load : SDNode<"AMDGPUISD::SBUFFER_LOAD", [SDNPMayLoad, SDNPMemOperand] >; +def SIds_ordered_count : SDNode<"AMDGPUISD::DS_ORDERED_COUNT", + SDTypeProfile<1, 2, [SDTCisVT<0, i32>, SDTCisVT<1, i32>, SDTCisVT<2, i16>]>, + [SDNPMayLoad, SDNPMayStore, SDNPMemOperand, SDNPHasChain, SDNPInGlue] +>; + def SIatomic_inc : SDNode<"AMDGPUISD::ATOMIC_INC", SDTAtomic2, [SDNPMayLoad, SDNPMayStore, SDNPMemOperand, SDNPHasChain] >; |