gem5/cpu/ozone/front_end.hh
Kevin Lim 21df09cf7a Fixes for ozone CPU to successfully boot and run linux.
cpu/base_dyn_inst.hh:
    Remove snoop function (did not mean to commit it).
cpu/ozone/back_end_impl.hh:
    Set instruction as having its result ready, not completed.
cpu/ozone/cpu.hh:
    Fixes for store conditionals.  Use an additional lock addr list to make sure that the access is valid.  I don't know if this is fully necessary, but it gives me a peace of mind (at some performance cost).
    Make sure to schedule for cycles(1) and not just 1 cycle in the future as tick = 1ps.
    Also support the new Checker.
cpu/ozone/cpu_builder.cc:
    Add parameter for maxOutstandingMemOps so it can be set through the config.
    Also add in the checker.  Right now it's a BaseCPU simobject, but that may change in the future.
cpu/ozone/cpu_impl.hh:
    Add support for the checker.  For now there's a dynamic cast to convert the simobject passed back from the builder to the proper Checker type.  It's ugly, but only happens at startup, and is probably a justified use of dynamic cast.

    Support switching out/taking over from other CPUs.

    Correct indexing problem for float registers.
cpu/ozone/dyn_inst.hh:
    Add ability for instructions to wait on memory instructions in addition to source register instructions.  This is needed for memory dependence predictors and memory barriers.
cpu/ozone/dyn_inst_impl.hh:
    Support waiting on memory operations.
    Use "resultReady" to differentiate an instruction having its registers produced vs being totally completed.
cpu/ozone/front_end.hh:
    Support switching out.
    Also record if an interrupt is pending.
cpu/ozone/front_end_impl.hh:
    Support switching out.  Also support stalling the front end if an interrupt is pending.
cpu/ozone/lw_back_end.hh:
    Add checker in.
    Support switching out.
    Support memory barriers.
cpu/ozone/lw_back_end_impl.hh:
    Lots of changes to get things to work right.
    Faults, traps, interrupts all wait until all stores have written back (important).
    Memory barriers are supported, as is the general ability for instructions to be dependent on other memory instructions.
cpu/ozone/lw_lsq.hh:
    Support switching out.
    Also use store writeback events in all cases, not just dcache misses.
cpu/ozone/lw_lsq_impl.hh:
    Support switching out.
    Also use store writeback events in all cases, not just dcache misses.
    Support the checker CPU.  Marks instructions as completed once the functional access is done (which has to be done for the checker to be able to verify results).
cpu/ozone/simple_params.hh:
    Add max outstanding mem ops parameter.
python/m5/objects/OzoneCPU.py:
    Add max outstanding mem ops, checker.

--HG--
extra : convert_revision : f4d408e1bb1f25836a097b6abe3856111e950c59
2006-05-11 19:18:36 -04:00

262 lines
6.3 KiB
C++

#ifndef __CPU_OZONE_FRONT_END_HH__
#define __CPU_OZONE_FRONT_END_HH__
#include <deque>
//#include "cpu/ozone/cpu.hh"
#include "cpu/inst_seq.hh"
#include "cpu/o3/bpred_unit.hh"
#include "cpu/ozone/rename_table.hh"
//#include "cpu/ozone/thread_state.hh"
#include "mem/mem_req.hh"
#include "sim/eventq.hh"
#include "sim/stats.hh"
class ExecContext;
class MemInterface;
template <class>
class OzoneThreadState;
class PageTable;
template <class>
class TimeBuffer;
template <class Impl>
class FrontEnd
{
public:
typedef typename Impl::Params Params;
typedef typename Impl::DynInst DynInst;
typedef typename Impl::DynInstPtr DynInstPtr;
typedef typename Impl::FullCPU FullCPU;
typedef typename Impl::BackEnd BackEnd;
typedef typename Impl::FullCPU::OzoneXC OzoneXC;
typedef typename Impl::FullCPU::CommStruct CommStruct;
FrontEnd(Params *params);
std::string name() const;
void setCPU(FullCPU *cpu_ptr)
{ cpu = cpu_ptr; }
void setBackEnd(BackEnd *back_end_ptr)
{ backEnd = back_end_ptr; }
void setCommBuffer(TimeBuffer<CommStruct> *_comm);
void setXC(ExecContext *xc_ptr);
void setThreadState(OzoneThreadState<Impl> *thread_ptr)
{ thread = thread_ptr; }
void regStats();
void tick();
Fault fetchCacheLine();
void processInst(DynInstPtr &inst);
void squash(const InstSeqNum &squash_num, const Addr &next_PC,
const bool is_branch = false, const bool branch_taken = false);
DynInstPtr getInst();
void processCacheCompletion(MemReqPtr &req);
void addFreeRegs(int num_freed);
bool isEmpty() { return instBuffer.empty(); }
void switchOut();
void takeOverFrom(ExecContext *old_xc = NULL);
bool isSwitchedOut() { return switchedOut; }
bool switchedOut;
private:
bool updateStatus();
void checkBE();
DynInstPtr getInstFromCacheline();
void renameInst(DynInstPtr &inst);
// Returns true if we need to stop the front end this cycle
bool processBarriers(DynInstPtr &inst);
void handleFault(Fault &fault);
public:
Fault getFault() { return fetchFault; }
private:
Fault fetchFault;
// Align an address (typically a PC) to the start of an I-cache block.
// We fold in the PISA 64- to 32-bit conversion here as well.
Addr icacheBlockAlignPC(Addr addr)
{
addr = TheISA::realPCToFetchPC(addr);
return (addr & ~(cacheBlkMask));
}
InstSeqNum getAndIncrementInstSeq()
{ return cpu->globalSeqNum++; }
public:
FullCPU *cpu;
BackEnd *backEnd;
ExecContext *xc;
OzoneThreadState<Impl> *thread;
enum Status {
Running,
Idle,
IcacheMissStall,
IcacheMissComplete,
SerializeBlocked,
SerializeComplete,
RenameBlocked,
QuiescePending,
BEBlocked
};
Status status;
private:
TimeBuffer<CommStruct> *comm;
typename TimeBuffer<CommStruct>::wire fromCommit;
typedef typename Impl::BranchPred BranchPred;
// Typedef for semi-opaque type that holds any information the branch
// predictor needs to update itself. Only two fields are used outside of
// branch predictor, nextPC and isTaken.
// typedef typename BranchPred::BPredInfo BPredInfo;
BranchPred branchPred;
class ICacheCompletionEvent : public Event
{
private:
MemReqPtr req;
FrontEnd *frontEnd;
public:
ICacheCompletionEvent(MemReqPtr &_req, FrontEnd *_fe);
virtual void process();
virtual const char *description();
};
MemInterface *icacheInterface;
#if !FULL_SYSTEM
PageTable *pTable;
#endif
MemReqPtr memReq;
/** Mask to get a cache block's address. */
Addr cacheBlkMask;
unsigned cacheBlkSize;
Addr cacheBlkPC;
/** The cache line being fetched. */
uint8_t *cacheData;
bool fetchCacheLineNextCycle;
bool cacheBlkValid;
public:
RenameTable<Impl> renameTable;
private:
Addr PC;
Addr nextPC;
public:
void setPC(Addr val) { PC = val; }
void setNextPC(Addr val) { nextPC = val; }
void wakeFromQuiesce();
void dumpInsts();
private:
typedef typename std::deque<DynInstPtr> InstBuff;
typedef typename InstBuff::iterator InstBuffIt;
InstBuff instBuffer;
int instBufferSize;
int maxInstBufferSize;
int width;
int freeRegs;
int numPhysRegs;
bool serializeNext;
DynInstPtr barrierInst;
public:
bool interruptPending;
private:
// number of idle cycles
/*
Stats::Average<> notIdleFraction;
Stats::Formula idleFraction;
*/
// @todo: Consider making these vectors and tracking on a per thread basis.
/** Stat for total number of cycles stalled due to an icache miss. */
Stats::Scalar<> icacheStallCycles;
/** Stat for total number of fetched instructions. */
Stats::Scalar<> fetchedInsts;
Stats::Scalar<> fetchedBranches;
/** Stat for total number of predicted branches. */
Stats::Scalar<> predictedBranches;
/** Stat for total number of cycles spent fetching. */
Stats::Scalar<> fetchCycles;
Stats::Scalar<> fetchIdleCycles;
/** Stat for total number of cycles spent squashing. */
Stats::Scalar<> fetchSquashCycles;
/** Stat for total number of cycles spent blocked due to other stages in
* the pipeline.
*/
Stats::Scalar<> fetchBlockedCycles;
/** Stat for total number of fetched cache lines. */
Stats::Scalar<> fetchedCacheLines;
Stats::Scalar<> fetchIcacheSquashes;
/** Distribution of number of instructions fetched each cycle. */
Stats::Distribution<> fetchNisnDist;
// Stats::Vector<> qfull_iq_occupancy;
// Stats::VectorDistribution<> qfull_iq_occ_dist_;
Stats::Formula idleRate;
Stats::Formula branchRate;
Stats::Formula fetchRate;
Stats::Scalar<> IFQCount; // cumulative IFQ occupancy
Stats::Formula IFQOccupancy;
Stats::Formula IFQLatency;
Stats::Scalar<> IFQFcount; // cumulative IFQ full count
Stats::Formula IFQFullRate;
Stats::Scalar<> dispatchCountStat;
Stats::Scalar<> dispatchedSerializing;
Stats::Scalar<> dispatchedTempSerializing;
Stats::Scalar<> dispatchSerializeStallCycles;
Stats::Formula dispatchRate;
Stats::Formula regIntFull;
Stats::Formula regFpFull;
};
#endif // __CPU_OZONE_FRONT_END_HH__