gem5/src/cpu/o3/probe/elastic_trace.hh

563 lines
22 KiB
C++
Raw Normal View History

proto, probe: Add elastic trace probe to o3 cpu The elastic trace is a type of probe listener and listens to probe points in multiple stages of the O3CPU. The notify method is called on a probe point typically when an instruction successfully progresses through that stage. As different listener methods mapped to the different probe points execute, relevant information about the instruction, e.g. timestamps and register accesses, are captured and stored in temporary InstExecInfo class objects. When the instruction progresses through the commit stage, the timing and the dependency information about the instruction is finalised and encapsulated in a struct called TraceInfo. TraceInfo objects are collected in a list instead of writing them out to the trace file one a time. This is required as the trace is processed in chunks to evaluate order dependencies and computational delay in case an instruction does not have any register dependencies. By this we achieve a simpler algorithm during replay because every record in the trace can be hooked onto a record in its past. The instruction dependency trace is written out as a protobuf format file. A second trace containing fetch requests at absolute timestamps is written to a separate protobuf format file. If the instruction is not executed then it is not added to the trace. The code checks if the instruction had a fault, if it predicated false and thus previous register values were restored or if it was a load/store that did not have a request (e.g. when the size of the request is zero). In all these cases the instruction is set as executed by the Execute stage and is picked up by the commit probe listener. But a request is not issued and registers are not written. So practically, skipping these should not hurt the dependency modelling. If squashing results in squashing younger instructions, it may happen that the squash probe discards the inst and removes it from the temporary store but execute stage deals with the instruction in the next cycle which results in the execute probe seeing this inst as 'new' inst. A sequence number of the last processed trace record is used to trap these cases and not add to the temporary store. The elastic instruction trace and fetch request trace can be read in and played back by the TraceCPU.
2015-12-07 23:42:15 +01:00
/*
* Copyright (c) 2013 - 2015 ARM Limited
* All rights reserved
*
* The license below extends only to copyright in the software and shall
* not be construed as granting a license to any other intellectual
* property including but not limited to intellectual property relating
* to a hardware implementation of the functionality of the software
* licensed hereunder. You may use the software subject to the license
* terms below provided that you ensure that this notice is replicated
* unmodified and in its entirety in all distributions of the software,
* modified or unmodified, in source code or in binary form.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are
* met: redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer;
* redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution;
* neither the name of the copyright holders nor the names of its
* contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Authors: Radhika Jagtap
* Andreas Hansson
* Thomas Grass
*/
/**
* @file This file describes a trace component which is a cpu probe listener
* used to generate elastic cpu traces. It registers listeners to probe points
* in the fetch, rename, iew and commit stages of the O3CPU. It processes the
* dependency graph of the cpu execution and writes out a protobuf trace. It
* also generates a protobuf trace of the instruction fetch requests.
*/
#ifndef __CPU_O3_PROBE_ELASTIC_TRACE_HH__
#define __CPU_O3_PROBE_ELASTIC_TRACE_HH__
#include <set>
#include <unordered_map>
#include <utility>
#include "cpu/o3/dyn_inst.hh"
#include "cpu/o3/impl.hh"
#include "mem/request.hh"
#include "params/ElasticTrace.hh"
#include "proto/inst_dep_record.pb.h"
#include "proto/packet.pb.h"
#include "proto/protoio.hh"
#include "sim/eventq.hh"
#include "sim/probe/probe.hh"
/**
* The elastic trace is a type of probe listener and listens to probe points
* in multiple stages of the O3CPU. The notify method is called on a probe
* point typically when an instruction successfully progresses through that
* stage.
*
* As different listener methods mapped to the different probe points execute,
* relevant information about the instruction, e.g. timestamps and register
* accesses, are captured and stored in temporary data structures. When the
* instruction progresses through the commit stage, the timing as well as
* dependency information about the instruction is finalised and encapsulated in
* a struct called TraceInfo. TraceInfo objects are collected in a list instead
* of writing them out to the trace file one a time. This is required as the
* trace is processed in chunks to evaluate order dependencies and computational
* delay in case an instruction does not have any register dependencies. By this
* we achieve a simpler algorithm during replay because every record in the
* trace can be hooked onto a record in its past. The trace is written out as
* a protobuf format output file.
*
* The output trace can be read in and played back by the TraceCPU.
*/
class ElasticTrace : public ProbeListenerObject
{
public:
typedef typename O3CPUImpl::DynInstPtr DynInstPtr;
typedef typename std::pair<InstSeqNum, PhysRegIndex> SeqNumRegPair;
/** Trace record types corresponding to instruction node types */
typedef ProtoMessage::InstDepRecord::RecordType RecordType;
typedef ProtoMessage::InstDepRecord Record;
proto, probe: Add elastic trace probe to o3 cpu The elastic trace is a type of probe listener and listens to probe points in multiple stages of the O3CPU. The notify method is called on a probe point typically when an instruction successfully progresses through that stage. As different listener methods mapped to the different probe points execute, relevant information about the instruction, e.g. timestamps and register accesses, are captured and stored in temporary InstExecInfo class objects. When the instruction progresses through the commit stage, the timing and the dependency information about the instruction is finalised and encapsulated in a struct called TraceInfo. TraceInfo objects are collected in a list instead of writing them out to the trace file one a time. This is required as the trace is processed in chunks to evaluate order dependencies and computational delay in case an instruction does not have any register dependencies. By this we achieve a simpler algorithm during replay because every record in the trace can be hooked onto a record in its past. The instruction dependency trace is written out as a protobuf format file. A second trace containing fetch requests at absolute timestamps is written to a separate protobuf format file. If the instruction is not executed then it is not added to the trace. The code checks if the instruction had a fault, if it predicated false and thus previous register values were restored or if it was a load/store that did not have a request (e.g. when the size of the request is zero). In all these cases the instruction is set as executed by the Execute stage and is picked up by the commit probe listener. But a request is not issued and registers are not written. So practically, skipping these should not hurt the dependency modelling. If squashing results in squashing younger instructions, it may happen that the squash probe discards the inst and removes it from the temporary store but execute stage deals with the instruction in the next cycle which results in the execute probe seeing this inst as 'new' inst. A sequence number of the last processed trace record is used to trap these cases and not add to the temporary store. The elastic instruction trace and fetch request trace can be read in and played back by the TraceCPU.
2015-12-07 23:42:15 +01:00
/** Constructor */
ElasticTrace(const ElasticTraceParams *params);
/**
* Register the probe listeners that is the methods called on a probe point
* notify() call.
*/
void regProbeListeners();
/** Register all listeners. */
void regEtraceListeners();
/** Returns the name of the trace probe listener. */
const std::string name() const;
/**
* Process any outstanding trace records, flush them out to the protobuf
* output streams and delete the streams at simulation exit.
*/
void flushTraces();
/**
* Take the fields of the request class object that are relevant to create
* an instruction fetch request. It creates a protobuf message containing
* the request fields and writes it to instTraceStream.
*
* @param req pointer to the fetch request
*/
void fetchReqTrace(const RequestPtr &req);
/**
* Populate the execute timestamp field in an InstExecInfo object for an
* instruction in flight.
*
* @param dyn_inst pointer to dynamic instruction in flight
*/
void recordExecTick(const DynInstPtr &dyn_inst);
/**
* Populate the timestamp field in an InstExecInfo object for an
* instruction in flight when it is execution is complete and it is ready
* to commit.
*
* @param dyn_inst pointer to dynamic instruction in flight
*/
void recordToCommTick(const DynInstPtr &dyn_inst);
/**
* Record a Read After Write physical register dependency if there has
* been a write to the source register and update the physical register
* map. For this look up the physRegDepMap with this instruction as the
* writer of its destination register. If the dependency falls outside the
* window it is assumed as already complete. Duplicate entries are avoided.
*
* @param dyn_inst pointer to dynamic instruction in flight
*/
void updateRegDep(const DynInstPtr &dyn_inst);
/**
* When an instruction gets squashed the destination register mapped to it
* is freed up in the rename stage. Remove the register entry from the
* physRegDepMap as well to avoid dependencies on squashed instructions.
*
* @param inst_reg_pair pair of inst. sequence no. and the register
*/
void removeRegDepMapEntry(const SeqNumRegPair &inst_reg_pair);
/**
* Add an instruction that is at the head of the ROB and is squashed only
* if it is a load and a request was sent for it.
*
* @param head_inst pointer to dynamic instruction to be squashed
*/
void addSquashedInst(const DynInstPtr &head_inst);
/**
* Add an instruction that is at the head of the ROB and is committed.
*
* @param head_inst pointer to dynamic instruction to be committed
*/
void addCommittedInst(const DynInstPtr &head_inst);
/** Register statistics for the elastic trace. */
void regStats();
/** Event to trigger registering this listener for all probe points. */
EventWrapper<ElasticTrace,
&ElasticTrace::regEtraceListeners> regEtraceListenersEvent;
private:
/**
* Used for checking the first window for processing and writing of
* dependency trace. At the start of the program there can be dependency-
* free instructions and such cases are handled differently.
*/
bool firstWin;
/**
* @defgroup InstExecInfo Struct for storing information before an
* instruction reaches the commit stage, e.g. execute timestamp.
*/
struct InstExecInfo
{
/**
* @ingroup InstExecInfo
* @{
*/
/** Timestamp when instruction was first processed by execute stage */
Tick executeTick;
/**
* Timestamp when instruction execution is completed in execute stage
* and instruction is marked as ready to commit
*/
Tick toCommitTick;
/**
* Set of instruction sequence numbers that this instruction depends on
* due to Read After Write data dependency based on physical register.
*/
std::set<InstSeqNum> physRegDepSet;
/** @} */
/** Constructor */
InstExecInfo()
: executeTick(MaxTick),
toCommitTick(MaxTick)
{ }
};
/**
* Temporary store of InstExecInfo objects. Later on when an instruction
* is processed for commit or retire, if it is chosen to be written to
* the output trace then this information is looked up using the instruction
* sequence number as the key. If it is not chosen then the entry for it in
* the store is cleared.
*/
std::unordered_map<InstSeqNum, InstExecInfo*> tempStore;
/**
* The last cleared instruction sequence number used to free up the memory
* allocated in the temporary store.
*/
InstSeqNum lastClearedSeqNum;
/**
* Map for recording the producer of a physical register to check Read
* After Write dependencies. The key is the renamed physical register and
* the value is the instruction sequence number of its last producer.
*/
std::unordered_map<PhysRegIndex, InstSeqNum> physRegDepMap;
/**
* @defgroup TraceInfo Struct for a record in the instruction dependency
* trace. All information required to process and calculate the
* computational delay is stored in TraceInfo objects. The memory request
* fields for a load or store instruction are also included here. Note
* that the structure TraceInfo does not store pointers to children
* or parents. The dependency trace is maintained as an ordered collection
* of records for writing to the output trace and not as a tree data
* structure.
*/
struct TraceInfo
{
/**
* @ingroup TraceInfo
* @{
*/
/* Instruction sequence number. */
InstSeqNum instNum;
/** The type of trace record for the instruction node */
RecordType type;
proto, probe: Add elastic trace probe to o3 cpu The elastic trace is a type of probe listener and listens to probe points in multiple stages of the O3CPU. The notify method is called on a probe point typically when an instruction successfully progresses through that stage. As different listener methods mapped to the different probe points execute, relevant information about the instruction, e.g. timestamps and register accesses, are captured and stored in temporary InstExecInfo class objects. When the instruction progresses through the commit stage, the timing and the dependency information about the instruction is finalised and encapsulated in a struct called TraceInfo. TraceInfo objects are collected in a list instead of writing them out to the trace file one a time. This is required as the trace is processed in chunks to evaluate order dependencies and computational delay in case an instruction does not have any register dependencies. By this we achieve a simpler algorithm during replay because every record in the trace can be hooked onto a record in its past. The instruction dependency trace is written out as a protobuf format file. A second trace containing fetch requests at absolute timestamps is written to a separate protobuf format file. If the instruction is not executed then it is not added to the trace. The code checks if the instruction had a fault, if it predicated false and thus previous register values were restored or if it was a load/store that did not have a request (e.g. when the size of the request is zero). In all these cases the instruction is set as executed by the Execute stage and is picked up by the commit probe listener. But a request is not issued and registers are not written. So practically, skipping these should not hurt the dependency modelling. If squashing results in squashing younger instructions, it may happen that the squash probe discards the inst and removes it from the temporary store but execute stage deals with the instruction in the next cycle which results in the execute probe seeing this inst as 'new' inst. A sequence number of the last processed trace record is used to trap these cases and not add to the temporary store. The elastic instruction trace and fetch request trace can be read in and played back by the TraceCPU.
2015-12-07 23:42:15 +01:00
/* Tick when instruction was in execute stage. */
Tick executeTick;
/* Tick when instruction was marked ready and sent to commit stage. */
Tick toCommitTick;
/* Tick when instruction was committed. */
Tick commitTick;
/* If instruction was committed, as against squashed. */
bool commit;
proto, probe: Add elastic trace probe to o3 cpu The elastic trace is a type of probe listener and listens to probe points in multiple stages of the O3CPU. The notify method is called on a probe point typically when an instruction successfully progresses through that stage. As different listener methods mapped to the different probe points execute, relevant information about the instruction, e.g. timestamps and register accesses, are captured and stored in temporary InstExecInfo class objects. When the instruction progresses through the commit stage, the timing and the dependency information about the instruction is finalised and encapsulated in a struct called TraceInfo. TraceInfo objects are collected in a list instead of writing them out to the trace file one a time. This is required as the trace is processed in chunks to evaluate order dependencies and computational delay in case an instruction does not have any register dependencies. By this we achieve a simpler algorithm during replay because every record in the trace can be hooked onto a record in its past. The instruction dependency trace is written out as a protobuf format file. A second trace containing fetch requests at absolute timestamps is written to a separate protobuf format file. If the instruction is not executed then it is not added to the trace. The code checks if the instruction had a fault, if it predicated false and thus previous register values were restored or if it was a load/store that did not have a request (e.g. when the size of the request is zero). In all these cases the instruction is set as executed by the Execute stage and is picked up by the commit probe listener. But a request is not issued and registers are not written. So practically, skipping these should not hurt the dependency modelling. If squashing results in squashing younger instructions, it may happen that the squash probe discards the inst and removes it from the temporary store but execute stage deals with the instruction in the next cycle which results in the execute probe seeing this inst as 'new' inst. A sequence number of the last processed trace record is used to trap these cases and not add to the temporary store. The elastic instruction trace and fetch request trace can be read in and played back by the TraceCPU.
2015-12-07 23:42:15 +01:00
/* List of order dependencies. */
std::list<InstSeqNum> robDepList;
/* List of physical register RAW dependencies. */
std::list<InstSeqNum> physRegDepList;
/**
* Computational delay after the last dependent inst. completed.
* A value of -1 which means instruction has no dependencies.
*/
int64_t compDelay;
/* Number of dependents. */
uint32_t numDepts;
/* The instruction PC for a load, store or non load/store. */
Addr pc;
/* Request flags in case of a load/store instruction */
Request::FlagsType reqFlags;
/* Request physical address in case of a load/store instruction */
Addr physAddr;
/* Request virtual address in case of a load/store instruction */
Addr virtAddr;
/* Address space id in case of a load/store instruction */
uint32_t asid;
proto, probe: Add elastic trace probe to o3 cpu The elastic trace is a type of probe listener and listens to probe points in multiple stages of the O3CPU. The notify method is called on a probe point typically when an instruction successfully progresses through that stage. As different listener methods mapped to the different probe points execute, relevant information about the instruction, e.g. timestamps and register accesses, are captured and stored in temporary InstExecInfo class objects. When the instruction progresses through the commit stage, the timing and the dependency information about the instruction is finalised and encapsulated in a struct called TraceInfo. TraceInfo objects are collected in a list instead of writing them out to the trace file one a time. This is required as the trace is processed in chunks to evaluate order dependencies and computational delay in case an instruction does not have any register dependencies. By this we achieve a simpler algorithm during replay because every record in the trace can be hooked onto a record in its past. The instruction dependency trace is written out as a protobuf format file. A second trace containing fetch requests at absolute timestamps is written to a separate protobuf format file. If the instruction is not executed then it is not added to the trace. The code checks if the instruction had a fault, if it predicated false and thus previous register values were restored or if it was a load/store that did not have a request (e.g. when the size of the request is zero). In all these cases the instruction is set as executed by the Execute stage and is picked up by the commit probe listener. But a request is not issued and registers are not written. So practically, skipping these should not hurt the dependency modelling. If squashing results in squashing younger instructions, it may happen that the squash probe discards the inst and removes it from the temporary store but execute stage deals with the instruction in the next cycle which results in the execute probe seeing this inst as 'new' inst. A sequence number of the last processed trace record is used to trap these cases and not add to the temporary store. The elastic instruction trace and fetch request trace can be read in and played back by the TraceCPU.
2015-12-07 23:42:15 +01:00
/* Request size in case of a load/store instruction */
unsigned size;
/** Default Constructor */
TraceInfo()
: type(Record::INVALID)
{ }
/** Is the record a load */
bool isLoad() const { return (type == Record::LOAD); }
/** Is the record a store */
bool isStore() const { return (type == Record::STORE); }
/** Is the record a fetch triggering an Icache request */
bool isComp() const { return (type == Record::COMP); }
/** Return string specifying the type of the node */
const std::string& typeToStr() const;
proto, probe: Add elastic trace probe to o3 cpu The elastic trace is a type of probe listener and listens to probe points in multiple stages of the O3CPU. The notify method is called on a probe point typically when an instruction successfully progresses through that stage. As different listener methods mapped to the different probe points execute, relevant information about the instruction, e.g. timestamps and register accesses, are captured and stored in temporary InstExecInfo class objects. When the instruction progresses through the commit stage, the timing and the dependency information about the instruction is finalised and encapsulated in a struct called TraceInfo. TraceInfo objects are collected in a list instead of writing them out to the trace file one a time. This is required as the trace is processed in chunks to evaluate order dependencies and computational delay in case an instruction does not have any register dependencies. By this we achieve a simpler algorithm during replay because every record in the trace can be hooked onto a record in its past. The instruction dependency trace is written out as a protobuf format file. A second trace containing fetch requests at absolute timestamps is written to a separate protobuf format file. If the instruction is not executed then it is not added to the trace. The code checks if the instruction had a fault, if it predicated false and thus previous register values were restored or if it was a load/store that did not have a request (e.g. when the size of the request is zero). In all these cases the instruction is set as executed by the Execute stage and is picked up by the commit probe listener. But a request is not issued and registers are not written. So practically, skipping these should not hurt the dependency modelling. If squashing results in squashing younger instructions, it may happen that the squash probe discards the inst and removes it from the temporary store but execute stage deals with the instruction in the next cycle which results in the execute probe seeing this inst as 'new' inst. A sequence number of the last processed trace record is used to trap these cases and not add to the temporary store. The elastic instruction trace and fetch request trace can be read in and played back by the TraceCPU.
2015-12-07 23:42:15 +01:00
/** @} */
/**
* Get the execute tick of the instruction.
*
* @return Tick when instruction was executed
*/
Tick getExecuteTick() const;
};
/**
* The instruction dependency trace containing TraceInfo objects. The
* container implemented is sequential as dependencies obey commit
* order (program order). For example, if B is dependent on A then B must
* be committed after A. Thus records are updated with dependency
* information and written to the trace in commit order. This ensures that
* when a graph is reconstructed from the trace during replay, all the
* dependencies are stored in the graph before the dependent itself is
* added. This facilitates creating a tree data structure during replay,
* i.e. adding children as records are read from the trace in an efficient
* manner.
*/
std::vector<TraceInfo*> depTrace;
/**
* Map where the instruction sequence number is mapped to the pointer to
* the TraceInfo object.
*/
std::unordered_map<InstSeqNum, TraceInfo*> traceInfoMap;
/** Typedef of iterator to the instruction dependency trace. */
typedef typename std::vector<TraceInfo*>::iterator depTraceItr;
/** Typedef of the reverse iterator to the instruction dependency trace. */
typedef typename std::reverse_iterator<depTraceItr> depTraceRevItr;
/**
* The maximum distance for a dependency and is set by a top level
* level parameter. It must be equal to or greater than the number of
* entries in the ROB. This variable is used as the length of the sliding
* window for processing the dependency trace.
*/
uint32_t depWindowSize;
/** Protobuf output stream for data dependency trace */
ProtoOutputStream* dataTraceStream;
/** Protobuf output stream for instruction fetch trace. */
ProtoOutputStream* instTraceStream;
/** Number of instructions after which to enable tracing. */
const InstSeqNum startTraceInst;
/**
* Whther the elastic trace listener has been registered for all probes.
*
* When enabling tracing after a specified number of instructions have
* committed, check this to prevent re-registering the listener.
*/
bool allProbesReg;
/** Whether to trace virtual addresses for memory requests. */
const bool traceVirtAddr;
proto, probe: Add elastic trace probe to o3 cpu The elastic trace is a type of probe listener and listens to probe points in multiple stages of the O3CPU. The notify method is called on a probe point typically when an instruction successfully progresses through that stage. As different listener methods mapped to the different probe points execute, relevant information about the instruction, e.g. timestamps and register accesses, are captured and stored in temporary InstExecInfo class objects. When the instruction progresses through the commit stage, the timing and the dependency information about the instruction is finalised and encapsulated in a struct called TraceInfo. TraceInfo objects are collected in a list instead of writing them out to the trace file one a time. This is required as the trace is processed in chunks to evaluate order dependencies and computational delay in case an instruction does not have any register dependencies. By this we achieve a simpler algorithm during replay because every record in the trace can be hooked onto a record in its past. The instruction dependency trace is written out as a protobuf format file. A second trace containing fetch requests at absolute timestamps is written to a separate protobuf format file. If the instruction is not executed then it is not added to the trace. The code checks if the instruction had a fault, if it predicated false and thus previous register values were restored or if it was a load/store that did not have a request (e.g. when the size of the request is zero). In all these cases the instruction is set as executed by the Execute stage and is picked up by the commit probe listener. But a request is not issued and registers are not written. So practically, skipping these should not hurt the dependency modelling. If squashing results in squashing younger instructions, it may happen that the squash probe discards the inst and removes it from the temporary store but execute stage deals with the instruction in the next cycle which results in the execute probe seeing this inst as 'new' inst. A sequence number of the last processed trace record is used to trap these cases and not add to the temporary store. The elastic instruction trace and fetch request trace can be read in and played back by the TraceCPU.
2015-12-07 23:42:15 +01:00
/** Pointer to the O3CPU that is this listener's parent a.k.a. manager */
FullO3CPU<O3CPUImpl>* cpu;
/**
* Add a record to the dependency trace depTrace which is a sequential
* container. A record is inserted per committed instruction and in the same
* order as the order in which instructions are committed.
*
* @param head_inst Pointer to the instruction which is head of the
* ROB and ready to commit
* @param exec_info_ptr Pointer to InstExecInfo for that instruction
* @param commit True if instruction is committed, false if squashed
*/
void addDepTraceRecord(const DynInstPtr &head_inst,
InstExecInfo* exec_info_ptr, bool commit);
/**
* Clear entries in the temporary store of execution info objects to free
* allocated memory until the present instruction being added to the trace.
*
* @param head_inst pointer to dynamic instruction
*/
void clearTempStoreUntil(const DynInstPtr head_inst);
/**
* Calculate the computational delay between an instruction and a
* subsequent instruction that has an ROB (order) dependency on it
*
* @param past_record Pointer to instruction
*
* @param new_record Pointer to subsequent instruction having an ROB
* dependency on the instruction pointed to by
* past_record
*/
void compDelayRob(TraceInfo* past_record, TraceInfo* new_record);
/**
* Calculate the computational delay between an instruction and a
* subsequent instruction that has a Physical Register (data) dependency on
* it.
*
* @param past_record Pointer to instruction
*
* @param new_record Pointer to subsequent instruction having a Physical
* Register dependency on the instruction pointed to
* by past_record
*/
void compDelayPhysRegDep(TraceInfo* past_record, TraceInfo* new_record);
/**
* Write out given number of records to the trace starting with the first
* record in depTrace and iterating through the trace in sequence. A
* record is deleted after it is written.
*
* @param num_to_write Number of records to write to the trace
*/
void writeDepTrace(uint32_t num_to_write);
/**
* Reverse iterate through the graph, search for a store-after-store or
* store-after-load dependency and update the new node's Rob dependency list.
*
* If a dependency is found, then call the assignRobDep() method that
* updates the store with the dependency information. This function is only
* called when a new store node is added to the trace.
*
* @param new_record pointer to new store record
* @param find_load_not_store true for searching store-after-load and false
* for searching store-after-store dependency
*/
void updateCommitOrderDep(TraceInfo* new_record, bool find_load_not_store);
/**
* Reverse iterate through the graph, search for an issue order dependency
* for a new node and update the new node's Rob dependency list.
*
* If a dependency is found, call the assignRobDep() method that updates
* the node with its dependency information. This function is called in
* case a new node to be added to the trace is dependency-free or its
* dependency got discarded because the dependency was outside the window.
*
* @param new_record pointer to new record to be added to the trace
*/
void updateIssueOrderDep(TraceInfo* new_record);
/**
* The new_record has an order dependency on a past_record, thus update the
* new record's Rob dependency list and increment the number of dependents
* of the past record.
*
* @param new_record pointer to new record
* @param past_record pointer to record that new_record has a rob
* dependency on
*/
void assignRobDep(TraceInfo* past_record, TraceInfo* new_record);
/**
* Check if past record is a store sent earlier than the execute tick.
*
* @param past_record pointer to past store
* @param execute_tick tick with which to compare past store's commit tick
*
* @return true if past record is store sent earlier
*/
bool hasStoreCommitted(TraceInfo* past_record, Tick execute_tick) const;
/**
* Check if past record is a load that completed earlier than the execute
* tick.
*
* @param past_record pointer to past load
* @param execute_tick tick with which to compare past load's complete
* tick
*
* @return true if past record is load completed earlier
*/
bool hasLoadCompleted(TraceInfo* past_record, Tick execute_tick) const;
/**
* Check if past record is a load sent earlier than the execute tick.
*
* @param past_record pointer to past load
* @param execute_tick tick with which to compare past load's send tick
*
* @return true if past record is load sent earlier
*/
bool hasLoadBeenSent(TraceInfo* past_record, Tick execute_tick) const;
/**
* Check if past record is a comp node that completed earlier than the
* execute tick.
*
* @param past_record pointer to past comp node
* @param execute_tick tick with which to compare past comp node's
* completion tick
*
* @return true if past record is comp completed earlier
*/
bool hasCompCompleted(TraceInfo* past_record, Tick execute_tick) const;
/** Number of register dependencies recorded during tracing */
Stats::Scalar numRegDep;
/**
* Number of stores that got assigned a commit order dependency
* on a past load/store.
*/
Stats::Scalar numOrderDepStores;
/**
* Number of load insts that got assigned an issue order dependency
* because they were dependency-free.
*/
Stats::Scalar numIssueOrderDepLoads;
/**
* Number of store insts that got assigned an issue order dependency
* because they were dependency-free.
*/
Stats::Scalar numIssueOrderDepStores;
/**
* Number of non load/store insts that got assigned an issue order
* dependency because they were dependency-free.
*/
Stats::Scalar numIssueOrderDepOther;
/** Number of filtered nodes */
Stats::Scalar numFilteredNodes;
/** Maximum number of dependents on any instruction */
Stats::Scalar maxNumDependents;
/**
* Maximum size of the temporary store mostly useful as a check that it is
* not growing
*/
Stats::Scalar maxTempStoreSize;
/**
* Maximum size of the map that holds the last writer to a physical
* register.
* */
Stats::Scalar maxPhysRegDepMapSize;
};
#endif//__CPU_O3_PROBE_ELASTIC_TRACE_HH__