+//===----------------------------------------------------------------------===//
+// GenericScheduler - Generic implementation of MachineSchedStrategy.
+//===----------------------------------------------------------------------===//
+
+namespace {
+/// Base class for GenericScheduler. This class maintains information about
+/// scheduling candidates based on TargetSchedModel making it easy to implement
+/// heuristics for either preRA or postRA scheduling.
+class GenericSchedulerBase : public MachineSchedStrategy {
+public:
+ /// Represent the type of SchedCandidate found within a single queue.
+ /// pickNodeBidirectional depends on these listed by decreasing priority.
+ enum CandReason {
+ NoCand, PhysRegCopy, RegExcess, RegCritical, Stall, Cluster, Weak, RegMax,
+ ResourceReduce, ResourceDemand, BotHeightReduce, BotPathReduce,
+ TopDepthReduce, TopPathReduce, NextDefUse, NodeOrder};
+
+#ifndef NDEBUG
+ static const char *getReasonStr(GenericSchedulerBase::CandReason Reason);
+#endif
+
+ /// Policy for scheduling the next instruction in the candidate's zone.
+ struct CandPolicy {
+ bool ReduceLatency;
+ unsigned ReduceResIdx;
+ unsigned DemandResIdx;
+
+ CandPolicy(): ReduceLatency(false), ReduceResIdx(0), DemandResIdx(0) {}
+ };
+
+ /// Status of an instruction's critical resource consumption.
+ struct SchedResourceDelta {
+ // Count critical resources in the scheduled region required by SU.
+ unsigned CritResources;
+
+ // Count critical resources from another region consumed by SU.
+ unsigned DemandedResources;
+
+ SchedResourceDelta(): CritResources(0), DemandedResources(0) {}
+
+ bool operator==(const SchedResourceDelta &RHS) const {
+ return CritResources == RHS.CritResources
+ && DemandedResources == RHS.DemandedResources;
+ }
+ bool operator!=(const SchedResourceDelta &RHS) const {
+ return !operator==(RHS);
+ }
+ };
+
+ /// Store the state used by GenericScheduler heuristics, required for the
+ /// lifetime of one invocation of pickNode().
+ struct SchedCandidate {
+ CandPolicy Policy;
+
+ // The best SUnit candidate.
+ SUnit *SU;
+
+ // The reason for this candidate.
+ CandReason Reason;
+
+ // Set of reasons that apply to multiple candidates.
+ uint32_t RepeatReasonSet;
+
+ // Register pressure values for the best candidate.
+ RegPressureDelta RPDelta;
+
+ // Critical resource consumption of the best candidate.
+ SchedResourceDelta ResDelta;
+
+ SchedCandidate(const CandPolicy &policy)
+ : Policy(policy), SU(NULL), Reason(NoCand), RepeatReasonSet(0) {}
+
+ bool isValid() const { return SU; }
+
+ // Copy the status of another candidate without changing policy.
+ void setBest(SchedCandidate &Best) {
+ assert(Best.Reason != NoCand && "uninitialized Sched candidate");
+ SU = Best.SU;
+ Reason = Best.Reason;
+ RPDelta = Best.RPDelta;
+ ResDelta = Best.ResDelta;
+ }
+
+ bool isRepeat(CandReason R) { return RepeatReasonSet & (1 << R); }
+ void setRepeat(CandReason R) { RepeatReasonSet |= (1 << R); }
+
+ void initResourceDelta(const ScheduleDAGMI *DAG,
+ const TargetSchedModel *SchedModel);
+ };
+
+protected:
+ const MachineSchedContext *Context;
+ const TargetSchedModel *SchedModel;
+ const TargetRegisterInfo *TRI;
+
+ SchedRemainder Rem;
+protected:
+ GenericSchedulerBase(const MachineSchedContext *C):
+ Context(C), SchedModel(0), TRI(0) {}
+
+ void setPolicy(CandPolicy &Policy, bool IsPostRA, SchedBoundary &CurrZone,
+ SchedBoundary *OtherZone);
+
+#ifndef NDEBUG
+ void traceCandidate(const SchedCandidate &Cand);
+#endif
+};
+} // namespace
+
+void GenericSchedulerBase::SchedCandidate::
+initResourceDelta(const ScheduleDAGMI *DAG,
+ const TargetSchedModel *SchedModel) {
+ if (!Policy.ReduceResIdx && !Policy.DemandResIdx)
+ return;
+
+ const MCSchedClassDesc *SC = DAG->getSchedClass(SU);
+ for (TargetSchedModel::ProcResIter
+ PI = SchedModel->getWriteProcResBegin(SC),
+ PE = SchedModel->getWriteProcResEnd(SC); PI != PE; ++PI) {
+ if (PI->ProcResourceIdx == Policy.ReduceResIdx)
+ ResDelta.CritResources += PI->Cycles;
+ if (PI->ProcResourceIdx == Policy.DemandResIdx)
+ ResDelta.DemandedResources += PI->Cycles;
+ }
+}
+
+/// Set the CandPolicy given a scheduling zone given the current resources and
+/// latencies inside and outside the zone.
+void GenericSchedulerBase::setPolicy(CandPolicy &Policy,
+ bool IsPostRA,
+ SchedBoundary &CurrZone,
+ SchedBoundary *OtherZone) {
+ // Apply preemptive heuristics based on the the total latency and resources
+ // inside and outside this zone. Potential stalls should be considered before
+ // following this policy.
+
+ // Compute remaining latency. We need this both to determine whether the
+ // overall schedule has become latency-limited and whether the instructions
+ // outside this zone are resource or latency limited.
+ //
+ // The "dependent" latency is updated incrementally during scheduling as the
+ // max height/depth of scheduled nodes minus the cycles since it was
+ // scheduled:
+ // DLat = max (N.depth - (CurrCycle - N.ReadyCycle) for N in Zone
+ //
+ // The "independent" latency is the max ready queue depth:
+ // ILat = max N.depth for N in Available|Pending
+ //
+ // RemainingLatency is the greater of independent and dependent latency.
+ unsigned RemLatency = CurrZone.getDependentLatency();
+ RemLatency = std::max(RemLatency,
+ CurrZone.findMaxLatency(CurrZone.Available.elements()));
+ RemLatency = std::max(RemLatency,
+ CurrZone.findMaxLatency(CurrZone.Pending.elements()));
+
+ // Compute the critical resource outside the zone.
+ unsigned OtherCritIdx = 0;
+ unsigned OtherCount =
+ OtherZone ? OtherZone->getOtherResourceCount(OtherCritIdx) : 0;
+
+ bool OtherResLimited = false;
+ if (SchedModel->hasInstrSchedModel()) {
+ unsigned LFactor = SchedModel->getLatencyFactor();
+ OtherResLimited = (int)(OtherCount - (RemLatency * LFactor)) > (int)LFactor;
+ }
+ // Schedule aggressively for latency in PostRA mode. We don't check for
+ // acyclic latency during PostRA, and highly out-of-order processors will
+ // skip PostRA scheduling.
+ if (!OtherResLimited) {
+ if (IsPostRA || (RemLatency + CurrZone.getCurrCycle() > Rem.CriticalPath)) {
+ Policy.ReduceLatency |= true;
+ DEBUG(dbgs() << " " << CurrZone.Available.getName()
+ << " RemainingLatency " << RemLatency << " + "
+ << CurrZone.getCurrCycle() << "c > CritPath "
+ << Rem.CriticalPath << "\n");
+ }
+ }
+ // If the same resource is limiting inside and outside the zone, do nothing.
+ if (CurrZone.getZoneCritResIdx() == OtherCritIdx)
+ return;
+
+ DEBUG(
+ if (CurrZone.isResourceLimited()) {
+ dbgs() << " " << CurrZone.Available.getName() << " ResourceLimited: "
+ << SchedModel->getResourceName(CurrZone.getZoneCritResIdx())
+ << "\n";
+ }
+ if (OtherResLimited)
+ dbgs() << " RemainingLimit: "
+ << SchedModel->getResourceName(OtherCritIdx) << "\n";
+ if (!CurrZone.isResourceLimited() && !OtherResLimited)
+ dbgs() << " Latency limited both directions.\n");
+
+ if (CurrZone.isResourceLimited() && !Policy.ReduceResIdx)
+ Policy.ReduceResIdx = CurrZone.getZoneCritResIdx();
+
+ if (OtherResLimited)
+ Policy.DemandResIdx = OtherCritIdx;
+}
+
+#ifndef NDEBUG
+const char *GenericSchedulerBase::getReasonStr(
+ GenericSchedulerBase::CandReason Reason) {
+ switch (Reason) {
+ case NoCand: return "NOCAND ";
+ case PhysRegCopy: return "PREG-COPY";
+ case RegExcess: return "REG-EXCESS";
+ case RegCritical: return "REG-CRIT ";
+ case Stall: return "STALL ";
+ case Cluster: return "CLUSTER ";
+ case Weak: return "WEAK ";
+ case RegMax: return "REG-MAX ";
+ case ResourceReduce: return "RES-REDUCE";
+ case ResourceDemand: return "RES-DEMAND";
+ case TopDepthReduce: return "TOP-DEPTH ";
+ case TopPathReduce: return "TOP-PATH ";
+ case BotHeightReduce:return "BOT-HEIGHT";
+ case BotPathReduce: return "BOT-PATH ";
+ case NextDefUse: return "DEF-USE ";
+ case NodeOrder: return "ORDER ";
+ };
+ llvm_unreachable("Unknown reason!");
+}
+
+void GenericSchedulerBase::traceCandidate(const SchedCandidate &Cand) {
+ PressureChange P;
+ unsigned ResIdx = 0;
+ unsigned Latency = 0;
+ switch (Cand.Reason) {
+ default:
+ break;
+ case RegExcess:
+ P = Cand.RPDelta.Excess;
+ break;
+ case RegCritical:
+ P = Cand.RPDelta.CriticalMax;
+ break;
+ case RegMax:
+ P = Cand.RPDelta.CurrentMax;
+ break;
+ case ResourceReduce:
+ ResIdx = Cand.Policy.ReduceResIdx;
+ break;
+ case ResourceDemand:
+ ResIdx = Cand.Policy.DemandResIdx;
+ break;
+ case TopDepthReduce:
+ Latency = Cand.SU->getDepth();
+ break;
+ case TopPathReduce:
+ Latency = Cand.SU->getHeight();
+ break;
+ case BotHeightReduce:
+ Latency = Cand.SU->getHeight();
+ break;
+ case BotPathReduce:
+ Latency = Cand.SU->getDepth();
+ break;
+ }
+ dbgs() << " SU(" << Cand.SU->NodeNum << ") " << getReasonStr(Cand.Reason);
+ if (P.isValid())
+ dbgs() << " " << TRI->getRegPressureSetName(P.getPSet())
+ << ":" << P.getUnitInc() << " ";
+ else
+ dbgs() << " ";
+ if (ResIdx)
+ dbgs() << " " << SchedModel->getProcResource(ResIdx)->Name << " ";
+ else
+ dbgs() << " ";
+ if (Latency)
+ dbgs() << " " << Latency << " cycles ";
+ else
+ dbgs() << " ";
+ dbgs() << '\n';
+}
+#endif
+
+/// Return true if this heuristic determines order.
+static bool tryLess(int TryVal, int CandVal,
+ GenericSchedulerBase::SchedCandidate &TryCand,
+ GenericSchedulerBase::SchedCandidate &Cand,
+ GenericSchedulerBase::CandReason Reason) {
+ if (TryVal < CandVal) {
+ TryCand.Reason = Reason;
+ return true;
+ }
+ if (TryVal > CandVal) {
+ if (Cand.Reason > Reason)
+ Cand.Reason = Reason;
+ return true;
+ }
+ Cand.setRepeat(Reason);
+ return false;
+}
+
+static bool tryGreater(int TryVal, int CandVal,
+ GenericSchedulerBase::SchedCandidate &TryCand,
+ GenericSchedulerBase::SchedCandidate &Cand,
+ GenericSchedulerBase::CandReason Reason) {
+ if (TryVal > CandVal) {
+ TryCand.Reason = Reason;
+ return true;
+ }
+ if (TryVal < CandVal) {
+ if (Cand.Reason > Reason)
+ Cand.Reason = Reason;
+ return true;
+ }
+ Cand.setRepeat(Reason);
+ return false;
+}
+
+static bool tryLatency(GenericSchedulerBase::SchedCandidate &TryCand,
+ GenericSchedulerBase::SchedCandidate &Cand,
+ SchedBoundary &Zone) {
+ if (Zone.isTop()) {
+ if (Cand.SU->getDepth() > Zone.getScheduledLatency()) {
+ if (tryLess(TryCand.SU->getDepth(), Cand.SU->getDepth(),
+ TryCand, Cand, GenericSchedulerBase::TopDepthReduce))
+ return true;
+ }
+ if (tryGreater(TryCand.SU->getHeight(), Cand.SU->getHeight(),
+ TryCand, Cand, GenericSchedulerBase::TopPathReduce))
+ return true;
+ }
+ else {
+ if (Cand.SU->getHeight() > Zone.getScheduledLatency()) {
+ if (tryLess(TryCand.SU->getHeight(), Cand.SU->getHeight(),
+ TryCand, Cand, GenericSchedulerBase::BotHeightReduce))
+ return true;
+ }
+ if (tryGreater(TryCand.SU->getDepth(), Cand.SU->getDepth(),
+ TryCand, Cand, GenericSchedulerBase::BotPathReduce))
+ return true;
+ }
+ return false;
+}
+
+static void tracePick(const GenericSchedulerBase::SchedCandidate &Cand,
+ bool IsTop) {
+ DEBUG(dbgs() << "Pick " << (IsTop ? "Top " : "Bot ")
+ << GenericSchedulerBase::getReasonStr(Cand.Reason) << '\n');
+}
+
+namespace {
+/// GenericScheduler shrinks the unscheduled zone using heuristics to balance
+/// the schedule.
+class GenericScheduler : public GenericSchedulerBase {
+ ScheduleDAGMILive *DAG;
+
+ // State of the top and bottom scheduled instruction boundaries.
+ SchedBoundary Top;
+ SchedBoundary Bot;
+
+ MachineSchedPolicy RegionPolicy;
+public:
+ GenericScheduler(const MachineSchedContext *C):
+ GenericSchedulerBase(C), DAG(0), Top(SchedBoundary::TopQID, "TopQ"),
+ Bot(SchedBoundary::BotQID, "BotQ") {}
+
+ virtual void initPolicy(MachineBasicBlock::iterator Begin,
+ MachineBasicBlock::iterator End,
+ unsigned NumRegionInstrs) override;
+
+ virtual bool shouldTrackPressure() const override {
+ return RegionPolicy.ShouldTrackPressure;
+ }
+
+ virtual void initialize(ScheduleDAGMI *dag) override;
+
+ virtual SUnit *pickNode(bool &IsTopNode) override;
+
+ virtual void schedNode(SUnit *SU, bool IsTopNode) override;
+
+ virtual void releaseTopNode(SUnit *SU) override {
+ Top.releaseTopNode(SU);
+ }
+
+ virtual void releaseBottomNode(SUnit *SU) override {
+ Bot.releaseBottomNode(SU);
+ }
+
+ virtual void registerRoots() override;
+
+protected:
+ void checkAcyclicLatency();
+
+ void tryCandidate(SchedCandidate &Cand,
+ SchedCandidate &TryCand,
+ SchedBoundary &Zone,
+ const RegPressureTracker &RPTracker,
+ RegPressureTracker &TempTracker);
+
+ SUnit *pickNodeBidirectional(bool &IsTopNode);
+
+ void pickNodeFromQueue(SchedBoundary &Zone,
+ const RegPressureTracker &RPTracker,
+ SchedCandidate &Candidate);
+
+ void reschedulePhysRegCopies(SUnit *SU, bool isTop);
+};
+} // namespace
+
+void GenericScheduler::initialize(ScheduleDAGMI *dag) {
+ assert(dag->hasVRegLiveness() &&
+ "(PreRA)GenericScheduler needs vreg liveness");
+ DAG = static_cast<ScheduleDAGMILive*>(dag);
+ SchedModel = DAG->getSchedModel();
+ TRI = DAG->TRI;
+
+ Rem.init(DAG, SchedModel);
+ Top.init(DAG, SchedModel, &Rem);
+ Bot.init(DAG, SchedModel, &Rem);
+
+ // Initialize resource counts.
+
+ // Initialize the HazardRecognizers. If itineraries don't exist, are empty, or
+ // are disabled, then these HazardRecs will be disabled.
+ const InstrItineraryData *Itin = SchedModel->getInstrItineraries();
+ const TargetMachine &TM = DAG->MF.getTarget();
+ if (!Top.HazardRec) {
+ Top.HazardRec =
+ TM.getInstrInfo()->CreateTargetMIHazardRecognizer(Itin, DAG);
+ }
+ if (!Bot.HazardRec) {
+ Bot.HazardRec =
+ TM.getInstrInfo()->CreateTargetMIHazardRecognizer(Itin, DAG);
+ }
+}
+
+/// Initialize the per-region scheduling policy.
+void GenericScheduler::initPolicy(MachineBasicBlock::iterator Begin,
+ MachineBasicBlock::iterator End,
+ unsigned NumRegionInstrs) {
+ const TargetMachine &TM = Context->MF->getTarget();
+ const TargetLowering *TLI = TM.getTargetLowering();
+
+ // Avoid setting up the register pressure tracker for small regions to save
+ // compile time. As a rough heuristic, only track pressure when the number of
+ // schedulable instructions exceeds half the integer register file.
+ RegionPolicy.ShouldTrackPressure = true;
+ for (unsigned VT = MVT::i32; VT > (unsigned)MVT::i1; --VT) {
+ MVT::SimpleValueType LegalIntVT = (MVT::SimpleValueType)VT;
+ if (TLI->isTypeLegal(LegalIntVT)) {
+ unsigned NIntRegs = Context->RegClassInfo->getNumAllocatableRegs(
+ TLI->getRegClassFor(LegalIntVT));
+ RegionPolicy.ShouldTrackPressure = NumRegionInstrs > (NIntRegs / 2);
+ }
+ }
+
+ // For generic targets, we default to bottom-up, because it's simpler and more
+ // compile-time optimizations have been implemented in that direction.
+ RegionPolicy.OnlyBottomUp = true;
+
+ // Allow the subtarget to override default policy.
+ const TargetSubtargetInfo &ST = TM.getSubtarget<TargetSubtargetInfo>();
+ ST.overrideSchedPolicy(RegionPolicy, Begin, End, NumRegionInstrs);
+
+ // After subtarget overrides, apply command line options.
+ if (!EnableRegPressure)
+ RegionPolicy.ShouldTrackPressure = false;
+
+ // Check -misched-topdown/bottomup can force or unforce scheduling direction.
+ // e.g. -misched-bottomup=false allows scheduling in both directions.
+ assert((!ForceTopDown || !ForceBottomUp) &&
+ "-misched-topdown incompatible with -misched-bottomup");
+ if (ForceBottomUp.getNumOccurrences() > 0) {
+ RegionPolicy.OnlyBottomUp = ForceBottomUp;
+ if (RegionPolicy.OnlyBottomUp)
+ RegionPolicy.OnlyTopDown = false;
+ }
+ if (ForceTopDown.getNumOccurrences() > 0) {
+ RegionPolicy.OnlyTopDown = ForceTopDown;
+ if (RegionPolicy.OnlyTopDown)
+ RegionPolicy.OnlyBottomUp = false;
+ }
+}
+
+/// Set IsAcyclicLatencyLimited if the acyclic path is longer than the cyclic
+/// critical path by more cycles than it takes to drain the instruction buffer.
+/// We estimate an upper bounds on in-flight instructions as:
+///
+/// CyclesPerIteration = max( CyclicPath, Loop-Resource-Height )
+/// InFlightIterations = AcyclicPath / CyclesPerIteration
+/// InFlightResources = InFlightIterations * LoopResources
+///
+/// TODO: Check execution resources in addition to IssueCount.
+void GenericScheduler::checkAcyclicLatency() {
+ if (Rem.CyclicCritPath == 0 || Rem.CyclicCritPath >= Rem.CriticalPath)
+ return;
+
+ // Scaled number of cycles per loop iteration.
+ unsigned IterCount =
+ std::max(Rem.CyclicCritPath * SchedModel->getLatencyFactor(),
+ Rem.RemIssueCount);
+ // Scaled acyclic critical path.
+ unsigned AcyclicCount = Rem.CriticalPath * SchedModel->getLatencyFactor();
+ // InFlightCount = (AcyclicPath / IterCycles) * InstrPerLoop
+ unsigned InFlightCount =
+ (AcyclicCount * Rem.RemIssueCount + IterCount-1) / IterCount;
+ unsigned BufferLimit =
+ SchedModel->getMicroOpBufferSize() * SchedModel->getMicroOpFactor();