gem5/src/mem/packet.hh

1005 lines
33 KiB
C++
Raw Normal View History

/*
* Copyright (c) 2012-2015 ARM Limited
MEM: Remove the Broadcast destination from the packet This patch simplifies the packet by removing the broadcast flag and instead more firmly relying on (and enforcing) the semantics of transactions in the classic memory system, i.e. request packets are routed from a master to a slave based on the address, and when they are created they have neither a valid source, nor destination. On their way to the slave, the request packet is updated with a source field for all modules that multiplex packets from multiple master (e.g. a bus). When a request packet is turned into a response packet (at the final slave), it moves the potentially populated source field to the destination field, and the response packet is routed through any multiplexing components back to the master based on the destination field. Modules that connect multiplexing components, such as caches and bridges store any existing source and destination field in the sender state as a stack (just as before). The packet constructor is simplified in that there is no longer a need to pass the Packet::Broadcast as the destination (this was always the case for the classic memory system). In the case of Ruby, rather than using the parameter to the constructor we now rely on setDest, as there is already another three-argument constructor in the packet class. In many places where the packet information was printed as part of DPRINTFs, request packets would be printed with a numeric "dest" that would always be -1 (Broadcast) and that field is now removed from the printing.
2012-04-14 11:45:55 +02:00
* 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.
*
* Copyright (c) 2006 The Regents of The University of Michigan
2010-08-25 23:08:27 +02:00
* Copyright (c) 2010 Advanced Micro Devices, Inc.
* All rights reserved.
*
* 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: Ron Dreslinski
* Steve Reinhardt
* Ali Saidi
* Andreas Hansson
*/
/**
* @file
* Declaration of the Packet class.
*/
#ifndef __MEM_PACKET_HH__
#define __MEM_PACKET_HH__
2011-04-15 19:44:06 +02:00
#include <bitset>
#include <cassert>
#include <list>
2008-11-10 20:51:17 +01:00
#include "base/cast.hh"
#include "base/compiler.hh"
2008-11-10 20:51:17 +01:00
#include "base/flags.hh"
#include "base/misc.hh"
#include "base/printable.hh"
#include "base/types.hh"
2009-05-17 23:34:52 +02:00
#include "mem/request.hh"
#include "sim/core.hh"
class Packet;
typedef Packet *PacketPtr;
typedef uint8_t* PacketDataPtr;
typedef std::list<PacketPtr> PacketList;
class MemCmd
{
2008-11-10 20:51:17 +01:00
friend class Packet;
2008-11-10 20:51:17 +01:00
public:
/**
* List of all commands associated with a packet.
*/
enum Command
{
InvalidCmd,
ReadReq,
ReadResp,
ReadRespWithInvalidate,
WriteReq,
WriteResp,
Writeback,
mem: Add clean evicts to improve snoop filter tracking This patch adds eviction notices to the caches, to provide accurate tracking of cache blocks in snoop filters. We add the CleanEvict message to the memory heirarchy and use both CleanEvicts and Writebacks with BLOCK_CACHED flags to propagate notice of clean and dirty evictions respectively, down the memory hierarchy. Note that the BLOCK_CACHED flag indicates whether there exist any copies of the evicted block in the caches above the evicting cache. The purpose of the CleanEvict message is to notify snoop filters of silent evictions in the relevant caches. The CleanEvict message behaves much like a Writeback. CleanEvict is a write and a request but unlike a Writeback, CleanEvict does not have data and does not need exclusive access to the block. The cache generates the CleanEvict message on a fill resulting in eviction of a clean block. Before travelling downwards CleanEvict requests generate zero-time snoop requests to check if the same block is cached in upper levels of the memory heirarchy. If the block exists, the cache discards the CleanEvict message. The snoops check the tags, writeback queue and the MSHRs of upper level caches in a manner similar to snoops generated from HardPFReqs. Currently CleanEvicts keep travelling towards main memory unless they encounter the block corresponding to their address or reach main memory (since we have no well defined point of serialisation). Main memory simply discards CleanEvict messages. We have modified the behavior of Writebacks, such that they generate snoops to check for the presence of blocks in upper level caches. It is possible in our current implmentation for a lower level cache to be writing back a block while a shared copy of the same block exists in the upper level cache. If the snoops find the same block in upper level caches, we set the BLOCK_CACHED flag in the Writeback message. We have also added logic to account for interaction of other message types with CleanEvicts waiting in the writeback queue. A simple example is of a response arriving at a cache removing any CleanEvicts to the same address from the cache's writeback queue.
2015-07-03 16:14:37 +02:00
CleanEvict,
SoftPFReq,
HardPFReq,
SoftPFResp,
HardPFResp,
WriteLineReq,
UpgradeReq,
SCUpgradeReq, // Special "weak" upgrade for StoreCond
UpgradeResp,
SCUpgradeFailReq, // Failed SCUpgradeReq in MSHR (never sent)
UpgradeFailResp, // Valid for SCUpgradeReq only
ReadExReq,
ReadExResp,
ReadCleanReq,
ReadSharedReq,
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
LoadLockedReq,
StoreCondReq,
StoreCondFailReq, // Failed StoreCondReq in MSHR (never sent)
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
StoreCondResp,
rename store conditional stuff as extra data so it can be used for conditional swaps as well Add support for a twin 64 bit int load Add Memory barrier and write barrier flags as appropriate Make atomic memory ops atomic src/arch/alpha/isa/mem.isa: src/arch/alpha/locked_mem.hh: src/cpu/base_dyn_inst.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_impl.hh: rename store conditional stuff as extra data so it can be used for conditional swaps as well src/arch/alpha/types.hh: src/arch/mips/types.hh: src/arch/sparc/types.hh: add a largest read data type for statically allocating read buffers in atomic simple cpu src/arch/isa_parser.py: Add support for a twin 64 bit int load src/arch/sparc/isa/decoder.isa: Make atomic memory ops atomic Add Memory barrier and write barrier flags as appropriate src/arch/sparc/isa/formats/mem/basicmem.isa: add post access code block and define a twinload format for twin loads src/arch/sparc/isa/formats/mem/blockmem.isa: remove old microcoded twin load coad src/arch/sparc/isa/formats/mem/mem.isa: swap.isa replaces the code in loadstore.isa src/arch/sparc/isa/formats/mem/util.isa: add a post access code block src/arch/sparc/isa/includes.isa: need bigint.hh for Twin64_t src/arch/sparc/isa/operands.isa: add a twin 64 int type src/cpu/simple/atomic.cc: src/cpu/simple/atomic.hh: src/cpu/simple/base.hh: src/cpu/simple/timing.cc: add support for twinloads add support for swap and conditional swap instructions rename store conditional stuff as extra data so it can be used for conditional swaps as well src/mem/packet.cc: src/mem/packet.hh: Add support for atomic swap memory commands src/mem/packet_access.hh: Add endian conversion function for Twin64_t type src/mem/physical.cc: src/mem/physical.hh: src/mem/request.hh: Add support for atomic swap memory commands Rename sc code to extradata --HG-- extra : convert_revision : 69d908512fb34a4e28b29a6e58b807fb1a6b1656
2007-02-12 19:06:30 +01:00
SwapReq,
SwapResp,
MessageReq,
MessageResp,
// Error responses
// @TODO these should be classified as responses rather than
// requests; coding them as requests initially for backwards
// compatibility
InvalidDestError, // packet dest field invalid
BadAddressError, // memory address invalid
FunctionalReadError, // unable to fulfill functional read
FunctionalWriteError, // unable to fulfill functional write
// Fake simulator-only commands
PrintReq, // Print state matching address
FlushReq, //request for a cache flush
InvalidateReq, // request for address to be invalidated
InvalidateResp,
NUM_MEM_CMDS
};
private:
2008-11-10 20:51:17 +01:00
/**
* List of command attributes.
*/
enum Attribute
{
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
IsRead, //!< Data flows from responder to requester
IsWrite, //!< Data flows from requester to responder
IsUpgrade,
IsInvalidate,
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
NeedsExclusive, //!< Requires exclusive copy to complete in-cache
IsRequest, //!< Issued by requester
IsResponse, //!< Issue by responder
NeedsResponse, //!< Requester needs response from target
IsSWPrefetch,
IsHWPrefetch,
IsLlsc, //!< Alpha/MIPS LL or SC access
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
HasData, //!< There is an associated payload
IsError, //!< Error response
IsPrint, //!< Print state matching address (for debugging)
IsFlush, //!< Flush the address from caches
NUM_COMMAND_ATTRIBUTES
};
2008-11-10 20:51:17 +01:00
/**
* Structure that defines attributes and other data associated
* with a Command.
*/
struct CommandInfo
{
/// Set of attribute flags.
const std::bitset<NUM_COMMAND_ATTRIBUTES> attributes;
2008-11-10 20:51:17 +01:00
/// Corresponding response for requests; InvalidCmd if no
/// response is applicable.
const Command response;
2008-11-10 20:51:17 +01:00
/// String representation (for printing)
const std::string str;
};
2008-11-10 20:51:17 +01:00
/// Array to map Command enum to associated info.
static const CommandInfo commandInfo[];
private:
Command cmd;
2008-11-10 20:51:17 +01:00
bool
testCmdAttrib(MemCmd::Attribute attrib) const
{
return commandInfo[cmd].attributes[attrib] != 0;
}
public:
bool isRead() const { return testCmdAttrib(IsRead); }
bool isWrite() const { return testCmdAttrib(IsWrite); }
bool isUpgrade() const { return testCmdAttrib(IsUpgrade); }
bool isRequest() const { return testCmdAttrib(IsRequest); }
bool isResponse() const { return testCmdAttrib(IsResponse); }
bool needsExclusive() const { return testCmdAttrib(NeedsExclusive); }
bool needsResponse() const { return testCmdAttrib(NeedsResponse); }
bool isInvalidate() const { return testCmdAttrib(IsInvalidate); }
/**
* Check if this particular packet type carries payload data. Note
* that this does not reflect if the data pointer of the packet is
* valid or not.
*/
bool hasData() const { return testCmdAttrib(HasData); }
2009-04-20 06:44:15 +02:00
bool isLLSC() const { return testCmdAttrib(IsLlsc); }
bool isSWPrefetch() const { return testCmdAttrib(IsSWPrefetch); }
bool isHWPrefetch() const { return testCmdAttrib(IsHWPrefetch); }
bool isPrefetch() const { return testCmdAttrib(IsSWPrefetch) ||
testCmdAttrib(IsHWPrefetch); }
bool isError() const { return testCmdAttrib(IsError); }
bool isPrint() const { return testCmdAttrib(IsPrint); }
bool isFlush() const { return testCmdAttrib(IsFlush); }
2008-11-10 20:51:17 +01:00
const Command
responseCommand() const
{
return commandInfo[cmd].response;
}
2008-11-10 20:51:17 +01:00
/// Return the string to a cmd given by idx.
const std::string &toString() const { return commandInfo[cmd].str; }
int toInt() const { return (int)cmd; }
2008-11-10 20:51:17 +01:00
MemCmd(Command _cmd) : cmd(_cmd) { }
MemCmd(int _cmd) : cmd((Command)_cmd) { }
MemCmd() : cmd(InvalidCmd) { }
2008-11-10 20:51:17 +01:00
bool operator==(MemCmd c2) const { return (cmd == c2.cmd); }
bool operator!=(MemCmd c2) const { return (cmd != c2.cmd); }
};
/**
* A Packet is used to encapsulate a transfer between two objects in
* the memory system (e.g., the L1 and L2 cache). (In contrast, a
* single Request travels all the way from the requester to the
* ultimate destination and back, possibly being conveyed by several
* different Packets along the way.)
*/
class Packet : public Printable
{
public:
2008-11-10 20:51:17 +01:00
typedef uint32_t FlagsType;
typedef ::Flags<FlagsType> Flags;
private:
static const FlagsType PUBLIC_FLAGS = 0x00000000;
static const FlagsType PRIVATE_FLAGS = 0x00007F0F;
static const FlagsType COPY_FLAGS = 0x0000000F;
static const FlagsType SHARED = 0x00000001;
// Special control flags
/// Special timing-mode atomic snoop for multi-level coherence.
static const FlagsType EXPRESS_SNOOP = 0x00000002;
/// Does supplier have exclusive copy?
/// Useful for multi-level coherence.
static const FlagsType SUPPLY_EXCLUSIVE = 0x00000004;
// Snoop response flags
static const FlagsType MEM_INHIBIT = 0x00000008;
/// Are the 'addr' and 'size' fields valid?
static const FlagsType VALID_ADDR = 0x00000100;
static const FlagsType VALID_SIZE = 0x00000200;
/// Is the data pointer set to a value that shouldn't be freed
/// when the packet is destroyed?
static const FlagsType STATIC_DATA = 0x00001000;
/// The data pointer points to a value that should be freed when
/// the packet is destroyed. The pointer is assumed to be pointing
/// to an array, and delete [] is consequently called
2008-11-10 20:51:17 +01:00
static const FlagsType DYNAMIC_DATA = 0x00002000;
/// suppress the error if this packet encounters a functional
/// access failure.
static const FlagsType SUPPRESS_FUNC_ERROR = 0x00008000;
// Signal block present to squash prefetch and cache evict packets
// through express snoop flag
static const FlagsType BLOCK_CACHED = 0x00010000;
2008-11-10 20:51:17 +01:00
Flags flags;
2008-11-10 20:51:17 +01:00
public:
typedef MemCmd::Command Command;
2008-11-10 20:51:17 +01:00
/// The command field of the packet.
MemCmd cmd;
2008-11-10 20:51:17 +01:00
/// A pointer to the original request.
const RequestPtr req;
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
private:
2008-11-10 20:51:17 +01:00
/**
* A pointer to the data being transfered. It can be differnt
* sizes at each level of the heirarchy so it belongs in the
* packet, not request. This may or may not be populated when a
* responder recieves the packet. If not populated it memory should
* be allocated.
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
*/
PacketDataPtr data;
2008-11-10 20:51:17 +01:00
/// The address of the request. This address could be virtual or
/// physical, depending on the system configuration.
Addr addr;
/// True if the request targets the secure memory space.
bool _isSecure;
2008-11-10 20:51:17 +01:00
/// The size of the request or transfer.
unsigned size;
2008-11-10 20:51:17 +01:00
/**
* The original value of the command field. Only valid when the
* current command field is an error condition; in that case, the
* previous contents of the command field are copied here. This
* field is *not* set on non-error responses.
*/
MemCmd origCmd;
/**
* Track the bytes found that satisfy a functional read.
*/
std::vector<bool> bytesValid;
public:
/**
* The extra delay from seeing the packet until the header is
* transmitted. This delay is used to communicate the crossbar
* forwarding latency to the neighbouring object (e.g. a cache)
* that actually makes the packet wait. As the delay is relative,
* a 32-bit unsigned should be sufficient.
*/
uint32_t headerDelay;
/**
* The extra pipelining delay from seeing the packet until the end of
* payload is transmitted by the component that provided it (if
* any). This includes the header delay. Similar to the header
* delay, this is used to make up for the fact that the
* crossbar does not make the packet wait. As the delay is
* relative, a 32-bit unsigned should be sufficient.
*/
uint32_t payloadDelay;
2008-11-10 20:51:17 +01:00
/**
* A virtual base opaque structure used to hold state associated
* with the packet (e.g., an MSHR), specific to a MemObject that
* sees the packet. A pointer to this state is returned in the
* packet's response so that the MemObject in question can quickly
* look up the state needed to process it. A specific subclass
* would be derived from this to carry state specific to a
* particular sending device.
*
* As multiple MemObjects may add their SenderState throughout the
* memory system, the SenderStates create a stack, where a
* MemObject can add a new Senderstate, as long as the
* predecessing SenderState is restored when the response comes
* back. For this reason, the predecessor should always be
* populated with the current SenderState of a packet before
* modifying the senderState field in the request packet.
2008-11-10 20:51:17 +01:00
*/
struct SenderState
{
SenderState* predecessor;
SenderState() : predecessor(NULL) {}
virtual ~SenderState() {}
};
/**
* Object used to maintain state of a PrintReq. The senderState
* field of a PrintReq should always be of this type.
*/
class PrintReqState : public SenderState
2008-11-10 20:51:17 +01:00
{
private:
/**
* An entry in the label stack.
*/
struct LabelStackEntry
{
const std::string label;
std::string *prefix;
bool labelPrinted;
2008-11-10 20:51:17 +01:00
LabelStackEntry(const std::string &_label, std::string *_prefix);
};
typedef std::list<LabelStackEntry> LabelStack;
LabelStack labelStack;
std::string *curPrefixPtr;
public:
std::ostream &os;
const int verbosity;
PrintReqState(std::ostream &os, int verbosity = 0);
~PrintReqState();
2008-11-10 20:51:17 +01:00
/**
* Returns the current line prefix.
*/
const std::string &curPrefix() { return *curPrefixPtr; }
2008-11-10 20:51:17 +01:00
/**
* Push a label onto the label stack, and prepend the given
* prefix string onto the current prefix. Labels will only be
2008-11-10 20:51:17 +01:00
* printed if an object within the label's scope is printed.
*/
void pushLabel(const std::string &lbl,
const std::string &prefix = " ");
2008-11-10 20:51:17 +01:00
/**
* Pop a label off the label stack.
*/
void popLabel();
2008-11-10 20:51:17 +01:00
/**
* Print all of the pending unprinted labels on the
* stack. Called by printObj(), so normally not called by
2008-11-10 20:51:17 +01:00
* users unless bypassing printObj().
*/
void printLabels();
2008-11-10 20:51:17 +01:00
/**
* Print a Printable object to os, because it matched the
* address on a PrintReq.
*/
void printObj(Printable *obj);
};
2008-11-10 20:51:17 +01:00
/**
* This packet's sender state. Devices should use dynamic_cast<>
* to cast to the state appropriate to the sender. The intent of
* this variable is to allow a device to attach extra information
* to a request. A response packet must return the sender state
2008-11-10 20:51:17 +01:00
* that was attached to the original request (even if a new packet
* is created).
*/
SenderState *senderState;
/**
* Push a new sender state to the packet and make the current
* sender state the predecessor of the new one. This should be
* prefered over direct manipulation of the senderState member
* variable.
*
* @param sender_state SenderState to push at the top of the stack
*/
void pushSenderState(SenderState *sender_state);
/**
* Pop the top of the state stack and return a pointer to it. This
* assumes the current sender state is not NULL. This should be
* preferred over direct manipulation of the senderState member
* variable.
*
* @return The current top of the stack
*/
SenderState *popSenderState();
/**
* Go through the sender state stack and return the first instance
* that is of type T (as determined by a dynamic_cast). If there
* is no sender state of type T, NULL is returned.
*
* @return The topmost state of type T
*/
template <typename T>
T * findNextSenderState() const
{
T *t = NULL;
SenderState* sender_state = senderState;
while (t == NULL && sender_state != NULL) {
t = dynamic_cast<T*>(sender_state);
sender_state = sender_state->predecessor;
}
return t;
}
2008-11-10 20:51:17 +01:00
/// Return the string name of the cmd field (for debugging and
/// tracing).
const std::string &cmdString() const { return cmd.toString(); }
2008-11-10 20:51:17 +01:00
/// Return the index of this command.
inline int cmdToIndex() const { return cmd.toInt(); }
bool isRead() const { return cmd.isRead(); }
bool isWrite() const { return cmd.isWrite(); }
bool isUpgrade() const { return cmd.isUpgrade(); }
bool isRequest() const { return cmd.isRequest(); }
bool isResponse() const { return cmd.isResponse(); }
bool needsExclusive() const { return cmd.needsExclusive(); }
bool needsResponse() const { return cmd.needsResponse(); }
bool isInvalidate() const { return cmd.isInvalidate(); }
bool hasData() const { return cmd.hasData(); }
bool isLLSC() const { return cmd.isLLSC(); }
bool isError() const { return cmd.isError(); }
bool isPrint() const { return cmd.isPrint(); }
bool isFlush() const { return cmd.isFlush(); }
// Snoop flags
void assertMemInhibit()
{
assert(isRequest());
assert(!flags.isSet(MEM_INHIBIT));
flags.set(MEM_INHIBIT);
}
bool memInhibitAsserted() const { return flags.isSet(MEM_INHIBIT); }
void assertShared() { flags.set(SHARED); }
bool sharedAsserted() const { return flags.isSet(SHARED); }
// Special control flags
void setExpressSnoop() { flags.set(EXPRESS_SNOOP); }
bool isExpressSnoop() const { return flags.isSet(EXPRESS_SNOOP); }
void setSupplyExclusive() { flags.set(SUPPLY_EXCLUSIVE); }
void clearSupplyExclusive() { flags.clear(SUPPLY_EXCLUSIVE); }
bool isSupplyExclusive() const { return flags.isSet(SUPPLY_EXCLUSIVE); }
void setSuppressFuncError() { flags.set(SUPPRESS_FUNC_ERROR); }
bool suppressFuncError() const { return flags.isSet(SUPPRESS_FUNC_ERROR); }
void setBlockCached() { flags.set(BLOCK_CACHED); }
bool isBlockCached() const { return flags.isSet(BLOCK_CACHED); }
mem: Add clean evicts to improve snoop filter tracking This patch adds eviction notices to the caches, to provide accurate tracking of cache blocks in snoop filters. We add the CleanEvict message to the memory heirarchy and use both CleanEvicts and Writebacks with BLOCK_CACHED flags to propagate notice of clean and dirty evictions respectively, down the memory hierarchy. Note that the BLOCK_CACHED flag indicates whether there exist any copies of the evicted block in the caches above the evicting cache. The purpose of the CleanEvict message is to notify snoop filters of silent evictions in the relevant caches. The CleanEvict message behaves much like a Writeback. CleanEvict is a write and a request but unlike a Writeback, CleanEvict does not have data and does not need exclusive access to the block. The cache generates the CleanEvict message on a fill resulting in eviction of a clean block. Before travelling downwards CleanEvict requests generate zero-time snoop requests to check if the same block is cached in upper levels of the memory heirarchy. If the block exists, the cache discards the CleanEvict message. The snoops check the tags, writeback queue and the MSHRs of upper level caches in a manner similar to snoops generated from HardPFReqs. Currently CleanEvicts keep travelling towards main memory unless they encounter the block corresponding to their address or reach main memory (since we have no well defined point of serialisation). Main memory simply discards CleanEvict messages. We have modified the behavior of Writebacks, such that they generate snoops to check for the presence of blocks in upper level caches. It is possible in our current implmentation for a lower level cache to be writing back a block while a shared copy of the same block exists in the upper level cache. If the snoops find the same block in upper level caches, we set the BLOCK_CACHED flag in the Writeback message. We have also added logic to account for interaction of other message types with CleanEvicts waiting in the writeback queue. A simple example is of a response arriving at a cache removing any CleanEvicts to the same address from the cache's writeback queue.
2015-07-03 16:14:37 +02:00
void clearBlockCached() { flags.clear(BLOCK_CACHED); }
// Network error conditions... encapsulate them as methods since
// their encoding keeps changing (from result field to command
// field, etc.)
2008-11-10 20:51:17 +01:00
void
setBadAddress()
{
assert(isResponse());
cmd = MemCmd::BadAddressError;
}
bool hadBadAddress() const { return cmd == MemCmd::BadAddressError; }
void copyError(Packet *pkt) { assert(pkt->isError()); cmd = pkt->cmd; }
Addr getAddr() const { assert(flags.isSet(VALID_ADDR)); return addr; }
/**
* Update the address of this packet mid-transaction. This is used
* by the address mapper to change an already set address to a new
* one based on the system configuration. It is intended to remap
* an existing address, so it asserts that the current address is
* valid.
*/
void setAddr(Addr _addr) { assert(flags.isSet(VALID_ADDR)); addr = _addr; }
unsigned getSize() const { assert(flags.isSet(VALID_SIZE)); return size; }
2008-11-10 20:51:17 +01:00
Addr getOffset(int blkSize) const { return getAddr() & (Addr)(blkSize - 1); }
bool isSecure() const
{
assert(flags.isSet(VALID_ADDR));
return _isSecure;
}
/**
* It has been determined that the SC packet should successfully update
* memory. Therefore, convert this SC packet to a normal write.
*/
void
convertScToWrite()
{
assert(isLLSC());
assert(isWrite());
cmd = MemCmd::WriteReq;
}
/**
* When ruby is in use, Ruby will monitor the cache line and thus M5
* phys memory should treat LL ops as normal reads.
*/
void
convertLlToRead()
{
assert(isLLSC());
assert(isRead());
cmd = MemCmd::ReadReq;
}
2008-11-10 20:51:17 +01:00
/**
* Constructor. Note that a Request object must be constructed
* first, but the Requests's physical address and size fields need
* not be valid. The command must be supplied.
2008-11-10 20:51:17 +01:00
*/
Packet(const RequestPtr _req, MemCmd _cmd)
: cmd(_cmd), req(_req), data(nullptr), addr(0), _isSecure(false),
size(0), headerDelay(0), payloadDelay(0),
senderState(NULL)
{
if (req->hasPaddr()) {
addr = req->getPaddr();
flags.set(VALID_ADDR);
_isSecure = req->isSecure();
}
if (req->hasSize()) {
size = req->getSize();
flags.set(VALID_SIZE);
}
}
2008-11-10 20:51:17 +01:00
/**
* Alternate constructor if you are trying to create a packet with
* a request that is for a whole block, not the address from the
* req. this allows for overriding the size/addr of the req.
*/
Packet(const RequestPtr _req, MemCmd _cmd, int _blkSize)
: cmd(_cmd), req(_req), data(nullptr), addr(0), _isSecure(false),
headerDelay(0), payloadDelay(0),
senderState(NULL)
{
if (req->hasPaddr()) {
addr = req->getPaddr() & ~(_blkSize - 1);
flags.set(VALID_ADDR);
_isSecure = req->isSecure();
}
size = _blkSize;
flags.set(VALID_SIZE);
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
}
2008-11-10 20:51:17 +01:00
/**
* Alternate constructor for copying a packet. Copy all fields
* *except* if the original packet's data was dynamic, don't copy
* that, as we can't guarantee that the new packet's lifetime is
* less than that of the original packet. In this case the new
2008-11-10 20:51:17 +01:00
* packet should allocate its own data.
*/
Packet(const PacketPtr pkt, bool clear_flags, bool alloc_data)
2008-11-10 20:51:17 +01:00
: cmd(pkt->cmd), req(pkt->req),
data(nullptr),
addr(pkt->addr), _isSecure(pkt->_isSecure), size(pkt->size),
bytesValid(pkt->bytesValid),
headerDelay(pkt->headerDelay),
payloadDelay(pkt->payloadDelay),
senderState(pkt->senderState)
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
{
if (!clear_flags)
2008-11-10 20:51:17 +01:00
flags.set(pkt->flags & COPY_FLAGS);
flags.set(pkt->flags & (VALID_ADDR|VALID_SIZE));
// should we allocate space for data, or not, the express
// snoops do not need to carry any data as they only serve to
// co-ordinate state changes
if (alloc_data) {
// even if asked to allocate data, if the original packet
// holds static data, then the sender will not be doing
// any memcpy on receiving the response, thus we simply
// carry the pointer forward
if (pkt->flags.isSet(STATIC_DATA)) {
data = pkt->data;
flags.set(STATIC_DATA);
} else {
allocate();
}
}
}
/**
* Generate the appropriate read MemCmd based on the Request flags.
*/
static MemCmd
makeReadCmd(const RequestPtr req)
{
if (req->isLLSC())
return MemCmd::LoadLockedReq;
else if (req->isPrefetch())
return MemCmd::SoftPFReq;
else
return MemCmd::ReadReq;
}
/**
* Generate the appropriate write MemCmd based on the Request flags.
*/
static MemCmd
makeWriteCmd(const RequestPtr req)
{
if (req->isLLSC())
return MemCmd::StoreCondReq;
else if (req->isSwap())
return MemCmd::SwapReq;
else
return MemCmd::WriteReq;
}
/**
* Constructor-like methods that return Packets based on Request objects.
* Fine-tune the MemCmd type if it's not a vanilla read or write.
*/
static PacketPtr
createRead(const RequestPtr req)
{
return new Packet(req, makeReadCmd(req));
}
static PacketPtr
createWrite(const RequestPtr req)
{
return new Packet(req, makeWriteCmd(req));
}
2008-11-10 20:51:17 +01:00
/**
* clean up packet variables
*/
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
~Packet()
{
// Delete the request object if this is a request packet which
// does not need a response, because the requester will not get
// a chance. If the request packet needs a response then the
// request will be deleted on receipt of the response
// packet. We also make sure to never delete the request for
// express snoops, even for cases when responses are not
// needed (CleanEvict and Writeback), since the snoop packet
// re-uses the same request.
if (req && isRequest() && !needsResponse() &&
!isExpressSnoop()) {
delete req;
}
2008-11-10 20:51:17 +01:00
deleteData();
}
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
/**
* Take a request packet and modify it in place to be suitable for
* returning as a response to that request.
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
*/
2008-11-10 20:51:17 +01:00
void
makeResponse()
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
{
assert(needsResponse());
assert(isRequest());
origCmd = cmd;
cmd = cmd.responseCommand();
// responses are never express, even if the snoop that
// triggered them was
flags.clear(EXPRESS_SNOOP);
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
}
2008-11-10 20:51:17 +01:00
void
makeAtomicResponse()
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
{
makeResponse();
}
2008-11-10 20:51:17 +01:00
void
makeTimingResponse()
{
makeResponse();
}
void
setFunctionalResponseStatus(bool success)
{
if (!success) {
if (isWrite()) {
cmd = MemCmd::FunctionalWriteError;
} else {
cmd = MemCmd::FunctionalReadError;
}
}
}
2010-03-13 02:31:08 +01:00
void
setSize(unsigned size)
{
assert(!flags.isSet(VALID_SIZE));
this->size = size;
flags.set(VALID_SIZE);
}
/**
* Set the data pointer to the following value that should not be
* freed. Static data allows us to do a single memcpy even if
* multiple packets are required to get from source to destination
* and back. In essence the pointer is set calling dataStatic on
* the original packet, and whenever this packet is copied and
* forwarded the same pointer is passed on. When a packet
* eventually reaches the destination holding the data, it is
* copied once into the location originally set. On the way back
* to the source, no copies are necessary.
*/
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
template <typename T>
void
dataStatic(T *p)
{
assert(flags.noneSet(STATIC_DATA|DYNAMIC_DATA));
data = (PacketDataPtr)p;
2008-11-10 20:51:17 +01:00
flags.set(STATIC_DATA);
}
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
/**
* Set the data pointer to the following value that should not be
* freed. This version of the function allows the pointer passed
* to us to be const. To avoid issues down the line we cast the
* constness away, the alternative would be to keep both a const
* and non-const data pointer and cleverly choose between
* them. Note that this is only allowed for static data.
*/
template <typename T>
void
dataStaticConst(const T *p)
{
assert(flags.noneSet(STATIC_DATA|DYNAMIC_DATA));
data = const_cast<PacketDataPtr>(p);
flags.set(STATIC_DATA);
}
/**
* Set the data pointer to a value that should have delete []
* called on it. Dynamic data is local to this packet, and as the
* packet travels from source to destination, forwarded packets
* will allocate their own data. When a packet reaches the final
* destination it will populate the dynamic data of that specific
* packet, and on the way back towards the source, memcpy will be
* invoked in every step where a new packet was created e.g. in
* the caches. Ultimately when the response reaches the source a
* final memcpy is needed to extract the data from the packet
* before it is deallocated.
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
*/
template <typename T>
void
dataDynamic(T *p)
{
assert(flags.noneSet(STATIC_DATA|DYNAMIC_DATA));
data = (PacketDataPtr)p;
2008-11-10 20:51:17 +01:00
flags.set(DYNAMIC_DATA);
}
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
2008-11-10 20:51:17 +01:00
/**
* get a pointer to the data ptr.
*/
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
template <typename T>
T*
getPtr()
{
assert(flags.isSet(STATIC_DATA|DYNAMIC_DATA));
return (T*)data;
}
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
template <typename T>
const T*
getConstPtr() const
{
assert(flags.isSet(STATIC_DATA|DYNAMIC_DATA));
return (const T*)data;
}
2008-11-10 20:51:17 +01:00
/**
* return the value of what is pointed to in the packet.
*/
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
template <typename T>
T get() const;
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
2008-11-10 20:51:17 +01:00
/**
* set the value in the data pointer to v.
*/
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
template <typename T>
void set(T v);
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
/**
* Copy data into the packet from the provided pointer.
*/
2008-11-10 20:51:17 +01:00
void
setData(const uint8_t *p)
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
{
// we should never be copying data onto itself, which means we
// must idenfity packets with static data, as they carry the
// same pointer from source to destination and back
assert(p != getPtr<uint8_t>() || flags.isSet(STATIC_DATA));
if (p != getPtr<uint8_t>())
// for packet with allocated dynamic data, we copy data from
// one to the other, e.g. a forwarded response to a response
std::memcpy(getPtr<uint8_t>(), p, getSize());
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
}
/**
* Copy data into the packet from the provided block pointer,
* which is aligned to the given block size.
*/
2008-11-10 20:51:17 +01:00
void
setDataFromBlock(const uint8_t *blk_data, int blkSize)
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
{
setData(blk_data + getOffset(blkSize));
}
/**
* Copy data from the packet to the provided block pointer, which
* is aligned to the given block size.
*/
2008-11-10 20:51:17 +01:00
void
writeData(uint8_t *p) const
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
{
std::memcpy(p, getConstPtr<uint8_t>(), getSize());
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
}
/**
* Copy data from the packet to the memory at the provided pointer.
*/
2008-11-10 20:51:17 +01:00
void
writeDataToBlock(uint8_t *blk_data, int blkSize) const
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
{
writeData(blk_data + getOffset(blkSize));
}
/**
* delete the data pointed to in the data pointer. Ok to call to
* matter how data was allocted.
*/
2008-11-10 20:51:17 +01:00
void
deleteData()
{
if (flags.isSet(DYNAMIC_DATA))
2008-11-10 20:51:17 +01:00
delete [] data;
flags.clear(STATIC_DATA|DYNAMIC_DATA);
2008-11-10 20:51:17 +01:00
data = NULL;
}
Mostly done with all device models for new memory system. Still need to get timing packets working and get sinic working after merge from head. Checkpointing may need some work now. Endian-happiness still not complete. SConscript: add all devices back into make file base/inet.hh: dev/etherbus.cc: dev/etherbus.hh: dev/etherdump.cc: dev/etherdump.hh: dev/etherint.hh: dev/etherlink.cc: dev/etherlink.hh: dev/etherpkt.cc: dev/etherpkt.hh: dev/ethertap.cc: dev/ethertap.hh: dev/pktfifo.cc: dev/pktfifo.hh: rename PacketPtr EthPacketPtr so it doesn't conflict with the PacketPtr type in the memory system configs/test/fs.py: add nics to fs.py cpu/cpu_exec_context.cc: remove this check, as it's not valid. We may want to add something else back in to make sure that no one can delete the static virtual ports in the exec context cpu/simple/cpu.cc: cpu/simple/cpu.hh: dev/alpha_console.cc: dev/ide_ctrl.cc: use new methods for accessing packet data dev/ide_disk.cc: add some more dprintfs dev/io_device.cc: delete packets when we are done with them. Update for new packet methods to access data dev/isa_fake.cc: dev/pciconfigall.cc: dev/tsunami_cchip.cc: dev/tsunami_io.cc: dev/tsunami_pchip.cc: dev/uart8250.cc: dev/uart8250.hh: mem/physical.cc: mem/port.cc: dUpdate for new packet methods to access data dev/ns_gige.cc: Update for new memory system dev/ns_gige.hh: python/m5/objects/Ethernet.py: update for new memory system dev/sinic.cc: dev/sinic.hh: Update for new memory system. Untested as need to merge in head because of kernel driver differences between versions mem/packet.hh: Add methods to access data instead of accessing it directly. --HG-- extra : convert_revision : 223f43876afd404e68337270cd9a5e44d0bf553e
2006-04-25 01:31:50 +02:00
/** Allocate memory for the packet. */
2008-11-10 20:51:17 +01:00
void
allocate()
{
assert(flags.noneSet(STATIC_DATA|DYNAMIC_DATA));
flags.set(DYNAMIC_DATA);
data = new uint8_t[getSize()];
2008-11-10 20:51:17 +01:00
}
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
/**
* Check a functional request against a memory value stored in
* another packet (i.e. an in-transit request or
* response). Returns true if the current packet is a read, and
* the other packet provides the data, which is then copied to the
* current packet. If the current packet is a write, and the other
* packet intersects this one, then we update the data
* accordingly.
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
*/
2008-11-10 20:51:17 +01:00
bool
checkFunctional(PacketPtr other)
2008-11-10 20:51:17 +01:00
{
// all packets that are carrying a payload should have a valid
// data pointer
return checkFunctional(other, other->getAddr(), other->isSecure(),
other->getSize(),
other->hasData() ?
other->getPtr<uint8_t>() : NULL);
More major reorg of cache. Seems to work for atomic mode now, timing mode still broken. configs/example/memtest.py: Revamp options. src/cpu/memtest/memtest.cc: No need for memory initialization. No need to make atomic response... memory system should do that now. src/cpu/memtest/memtest.hh: MemTest really doesn't want to snoop. src/mem/bridge.cc: checkFunctional() cleanup. src/mem/bus.cc: src/mem/bus.hh: src/mem/cache/base_cache.cc: src/mem/cache/base_cache.hh: src/mem/cache/cache.cc: src/mem/cache/cache.hh: src/mem/cache/cache_blk.hh: src/mem/cache/cache_builder.cc: src/mem/cache/cache_impl.hh: src/mem/cache/coherence/coherence_protocol.cc: src/mem/cache/coherence/coherence_protocol.hh: src/mem/cache/coherence/simple_coherence.hh: src/mem/cache/miss/SConscript: src/mem/cache/miss/mshr.cc: src/mem/cache/miss/mshr.hh: src/mem/cache/miss/mshr_queue.cc: src/mem/cache/miss/mshr_queue.hh: src/mem/cache/prefetch/base_prefetcher.cc: src/mem/cache/tags/fa_lru.cc: src/mem/cache/tags/fa_lru.hh: src/mem/cache/tags/iic.cc: src/mem/cache/tags/iic.hh: src/mem/cache/tags/lru.cc: src/mem/cache/tags/lru.hh: src/mem/cache/tags/split.cc: src/mem/cache/tags/split.hh: src/mem/cache/tags/split_lifo.cc: src/mem/cache/tags/split_lifo.hh: src/mem/cache/tags/split_lru.cc: src/mem/cache/tags/split_lru.hh: src/mem/packet.cc: src/mem/packet.hh: src/mem/physical.cc: src/mem/physical.hh: src/mem/tport.cc: More major reorg. Seems to work for atomic mode now, timing mode still broken. --HG-- extra : convert_revision : 7e70dfc4a752393b911880ff028271433855ae87
2007-06-18 02:27:53 +02:00
}
mem: Add clean evicts to improve snoop filter tracking This patch adds eviction notices to the caches, to provide accurate tracking of cache blocks in snoop filters. We add the CleanEvict message to the memory heirarchy and use both CleanEvicts and Writebacks with BLOCK_CACHED flags to propagate notice of clean and dirty evictions respectively, down the memory hierarchy. Note that the BLOCK_CACHED flag indicates whether there exist any copies of the evicted block in the caches above the evicting cache. The purpose of the CleanEvict message is to notify snoop filters of silent evictions in the relevant caches. The CleanEvict message behaves much like a Writeback. CleanEvict is a write and a request but unlike a Writeback, CleanEvict does not have data and does not need exclusive access to the block. The cache generates the CleanEvict message on a fill resulting in eviction of a clean block. Before travelling downwards CleanEvict requests generate zero-time snoop requests to check if the same block is cached in upper levels of the memory heirarchy. If the block exists, the cache discards the CleanEvict message. The snoops check the tags, writeback queue and the MSHRs of upper level caches in a manner similar to snoops generated from HardPFReqs. Currently CleanEvicts keep travelling towards main memory unless they encounter the block corresponding to their address or reach main memory (since we have no well defined point of serialisation). Main memory simply discards CleanEvict messages. We have modified the behavior of Writebacks, such that they generate snoops to check for the presence of blocks in upper level caches. It is possible in our current implmentation for a lower level cache to be writing back a block while a shared copy of the same block exists in the upper level cache. If the snoops find the same block in upper level caches, we set the BLOCK_CACHED flag in the Writeback message. We have also added logic to account for interaction of other message types with CleanEvicts waiting in the writeback queue. A simple example is of a response arriving at a cache removing any CleanEvicts to the same address from the cache's writeback queue.
2015-07-03 16:14:37 +02:00
/**
* Is this request notification of a clean or dirty eviction from the cache.
**/
bool
evictingBlock() const
{
return (cmd == MemCmd::Writeback ||
cmd == MemCmd::CleanEvict);
}
/**
* Does the request need to check for cached copies of the same block
* in the memory hierarchy above.
**/
bool
mustCheckAbove() const
{
return (cmd == MemCmd::HardPFReq ||
evictingBlock());
}
/**
* Check a functional request against a memory value represented
* by a base/size pair and an associated data array. If the
* current packet is a read, it may be satisfied by the memory
* value. If the current packet is a write, it may update the
* memory value.
*/
bool
checkFunctional(Printable *obj, Addr base, bool is_secure, int size,
uint8_t *_data);
/**
* Push label for PrintReq (safe to call unconditionally).
*/
2008-11-10 20:51:17 +01:00
void
pushLabel(const std::string &lbl)
{
if (isPrint())
safe_cast<PrintReqState*>(senderState)->pushLabel(lbl);
}
/**
* Pop label for PrintReq (safe to call unconditionally).
*/
2008-11-10 20:51:17 +01:00
void
popLabel()
{
if (isPrint())
safe_cast<PrintReqState*>(senderState)->popLabel();
}
void print(std::ostream &o, int verbosity = 0,
const std::string &prefix = "") const;
/**
* A no-args wrapper of print(std::ostream...)
* meant to be invoked from DPRINTFs
* avoiding string overheads in fast mode
* @return string with the request's type and start<->end addresses
*/
std::string print() const;
};
#endif //__MEM_PACKET_HH