gem5/src/arch/x86/pagetable_walker.cc

694 lines
21 KiB
C++
Raw Normal View History

/*
MEM: Separate snoops and normal memory requests/responses This patch introduces port access methods that separates snoop request/responses from normal memory request/responses. The differentiation is made for functional, atomic and timing accesses and builds on the introduction of master and slave ports. Before the introduction of this patch, the packets belonging to the different phases of the protocol (request -> [forwarded snoop request -> snoop response]* -> response) all use the same port access functions, even though the snoop packets flow in the opposite direction to the normal packet. That is, a coherent master sends normal request and receives responses, but receives snoop requests and sends snoop responses (vice versa for the slave). These two distinct phases now use different access functions, as described below. Starting with the functional access, a master sends a request to a slave through sendFunctional, and the request packet is turned into a response before the call returns. In a system without cache coherence, this is all that is needed from the functional interface. For the cache-coherent scenario, a slave also sends snoop requests to coherent masters through sendFunctionalSnoop, with responses returned within the same packet pointer. This is currently used by the bus and caches, and the LSQ of the O3 CPU. The send/recvFunctional and send/recvFunctionalSnoop are moved from the Port super class to the appropriate subclass. Atomic accesses follow the same flow as functional accesses, with request being sent from master to slave through sendAtomic. In the case of cache-coherent ports, a slave can send snoop requests to a master through sendAtomicSnoop. Just as for the functional access methods, the atomic send and receive member functions are moved to the appropriate subclasses. The timing access methods are different from the functional and atomic in that requests and responses are separated in time and send/recvTiming are used for both directions. Hence, a master uses sendTiming to send a request to a slave, and a slave uses sendTiming to send a response back to a master, at a later point in time. Snoop requests and responses travel in the opposite direction, similar to what happens in functional and atomic accesses. With the introduction of this patch, it is possible to determine the direction of packets in the bus, and no longer necessary to look for both a master and a slave port with the requested port id. In contrast to the normal recvFunctional, recvAtomic and recvTiming that are pure virtual functions, the recvFunctionalSnoop, recvAtomicSnoop and recvTimingSnoop have a default implementation that calls panic. This is to allow non-coherent master and slave ports to not implement these functions.
2012-04-14 11:45:07 +02:00
* Copyright (c) 2012 ARM Limited
* All rights reserved.
*
* The license below extends only to copyright in the software and shall
* not be construed as granting a license to any other intellectual
* property including but not limited to intellectual property relating
* to a hardware implementation of the functionality of the software
* licensed hereunder. You may use the software subject to the license
* terms below provided that you ensure that this notice is replicated
* unmodified and in its entirety in all distributions of the software,
* modified or unmodified, in source code or in binary form.
*
* Copyright (c) 2007 The Hewlett-Packard Development Company
* All rights reserved.
*
* The license below extends only to copyright in the software and shall
* not be construed as granting a license to any other intellectual
* property including but not limited to intellectual property relating
* to a hardware implementation of the functionality of the software
* licensed hereunder. You may use the software subject to the license
* terms below provided that you ensure that this notice is replicated
* unmodified and in its entirety in all distributions of the software,
* modified or unmodified, in source code or in binary form.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are
* met: redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer;
* redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution;
* neither the name of the copyright holders nor the names of its
* contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Authors: Gabe Black
*/
#include "arch/x86/pagetable.hh"
#include "arch/x86/pagetable_walker.hh"
#include "arch/x86/tlb.hh"
#include "arch/x86/vtophys.hh"
#include "base/bitfield.hh"
#include "base/trie.hh"
#include "cpu/base.hh"
2011-04-15 19:44:06 +02:00
#include "cpu/thread_context.hh"
#include "debug/PageTableWalker.hh"
#include "mem/packet_access.hh"
#include "mem/request.hh"
namespace X86ISA {
// Unfortunately, the placement of the base field in a page table entry is
// very erratic and would make a mess here. It might be moved here at some
// point in the future.
BitUnion64(PageTableEntry)
Bitfield<63> nx;
Bitfield<11, 9> avl;
Bitfield<8> g;
Bitfield<7> ps;
Bitfield<6> d;
Bitfield<5> a;
Bitfield<4> pcd;
Bitfield<3> pwt;
Bitfield<2> u;
Bitfield<1> w;
Bitfield<0> p;
EndBitUnion(PageTableEntry)
Fault
Walker::start(ThreadContext * _tc, BaseTLB::Translation *_translation,
RequestPtr _req, BaseTLB::Mode _mode)
{
// TODO: in timing mode, instead of blocking when there are other
// outstanding requests, see if this request can be coalesced with
// another one (i.e. either coalesce or start walk)
WalkerState * newState = new WalkerState(this, _translation, _req);
newState->initState(_tc, _mode, sys->getMemoryMode() == Enums::timing);
if (currStates.size()) {
assert(newState->isTiming());
DPRINTF(PageTableWalker, "Walks in progress: %d\n", currStates.size());
currStates.push_back(newState);
return NoFault;
} else {
currStates.push_back(newState);
Fault fault = newState->startWalk();
if (!newState->isTiming()) {
currStates.pop_front();
delete newState;
}
return fault;
}
}
Fault
Walker::startFunctional(ThreadContext * _tc, Addr &addr, unsigned &logBytes,
BaseTLB::Mode _mode)
{
funcState.initState(_tc, _mode);
return funcState.startFunctional(addr, logBytes);
}
bool
MEM: Separate requests and responses for timing accesses This patch moves send/recvTiming and send/recvTimingSnoop from the Port base class to the MasterPort and SlavePort, and also splits them into separate member functions for requests and responses: send/recvTimingReq, send/recvTimingResp, and send/recvTimingSnoopReq, send/recvTimingSnoopResp. A master port sends requests and receives responses, and also receives snoop requests and sends snoop responses. A slave port has the reciprocal behaviour as it receives requests and sends responses, and sends snoop requests and receives snoop responses. For all MemObjects that have only master ports or slave ports (but not both), e.g. a CPU, or a PIO device, this patch merely adds more clarity to what kind of access is taking place. For example, a CPU port used to call sendTiming, and will now call sendTimingReq. Similarly, a response previously came back through recvTiming, which is now recvTimingResp. For the modules that have both master and slave ports, e.g. the bus, the behaviour was previously relying on branches based on pkt->isRequest(), and this is now replaced with a direct call to the apprioriate member function depending on the type of access. Please note that send/recvRetry is still shared by all the timing accessors and remains in the Port base class for now (to maintain the current bus functionality and avoid changing the statistics of all regressions). The packet queue is split into a MasterPort and SlavePort version to facilitate the use of the new timing accessors. All uses of the PacketQueue are updated accordingly. With this patch, the type of packet (request or response) is now well defined for each type of access, and asserts on pkt->isRequest() and pkt->isResponse() are now moved to the appropriate send member functions. It is also worth noting that sendTimingSnoopReq no longer returns a boolean, as the semantics do not alow snoop requests to be rejected or stalled. All these assumptions are now excplicitly part of the port interface itself.
2012-05-01 19:40:42 +02:00
Walker::WalkerPort::recvTimingResp(PacketPtr pkt)
{
MEM: Separate requests and responses for timing accesses This patch moves send/recvTiming and send/recvTimingSnoop from the Port base class to the MasterPort and SlavePort, and also splits them into separate member functions for requests and responses: send/recvTimingReq, send/recvTimingResp, and send/recvTimingSnoopReq, send/recvTimingSnoopResp. A master port sends requests and receives responses, and also receives snoop requests and sends snoop responses. A slave port has the reciprocal behaviour as it receives requests and sends responses, and sends snoop requests and receives snoop responses. For all MemObjects that have only master ports or slave ports (but not both), e.g. a CPU, or a PIO device, this patch merely adds more clarity to what kind of access is taking place. For example, a CPU port used to call sendTiming, and will now call sendTimingReq. Similarly, a response previously came back through recvTiming, which is now recvTimingResp. For the modules that have both master and slave ports, e.g. the bus, the behaviour was previously relying on branches based on pkt->isRequest(), and this is now replaced with a direct call to the apprioriate member function depending on the type of access. Please note that send/recvRetry is still shared by all the timing accessors and remains in the Port base class for now (to maintain the current bus functionality and avoid changing the statistics of all regressions). The packet queue is split into a MasterPort and SlavePort version to facilitate the use of the new timing accessors. All uses of the PacketQueue are updated accordingly. With this patch, the type of packet (request or response) is now well defined for each type of access, and asserts on pkt->isRequest() and pkt->isResponse() are now moved to the appropriate send member functions. It is also worth noting that sendTimingSnoopReq no longer returns a boolean, as the semantics do not alow snoop requests to be rejected or stalled. All these assumptions are now excplicitly part of the port interface itself.
2012-05-01 19:40:42 +02:00
return walker->recvTimingResp(pkt);
}
bool
MEM: Separate requests and responses for timing accesses This patch moves send/recvTiming and send/recvTimingSnoop from the Port base class to the MasterPort and SlavePort, and also splits them into separate member functions for requests and responses: send/recvTimingReq, send/recvTimingResp, and send/recvTimingSnoopReq, send/recvTimingSnoopResp. A master port sends requests and receives responses, and also receives snoop requests and sends snoop responses. A slave port has the reciprocal behaviour as it receives requests and sends responses, and sends snoop requests and receives snoop responses. For all MemObjects that have only master ports or slave ports (but not both), e.g. a CPU, or a PIO device, this patch merely adds more clarity to what kind of access is taking place. For example, a CPU port used to call sendTiming, and will now call sendTimingReq. Similarly, a response previously came back through recvTiming, which is now recvTimingResp. For the modules that have both master and slave ports, e.g. the bus, the behaviour was previously relying on branches based on pkt->isRequest(), and this is now replaced with a direct call to the apprioriate member function depending on the type of access. Please note that send/recvRetry is still shared by all the timing accessors and remains in the Port base class for now (to maintain the current bus functionality and avoid changing the statistics of all regressions). The packet queue is split into a MasterPort and SlavePort version to facilitate the use of the new timing accessors. All uses of the PacketQueue are updated accordingly. With this patch, the type of packet (request or response) is now well defined for each type of access, and asserts on pkt->isRequest() and pkt->isResponse() are now moved to the appropriate send member functions. It is also worth noting that sendTimingSnoopReq no longer returns a boolean, as the semantics do not alow snoop requests to be rejected or stalled. All these assumptions are now excplicitly part of the port interface itself.
2012-05-01 19:40:42 +02:00
Walker::recvTimingResp(PacketPtr pkt)
{
MEM: Separate snoops and normal memory requests/responses This patch introduces port access methods that separates snoop request/responses from normal memory request/responses. The differentiation is made for functional, atomic and timing accesses and builds on the introduction of master and slave ports. Before the introduction of this patch, the packets belonging to the different phases of the protocol (request -> [forwarded snoop request -> snoop response]* -> response) all use the same port access functions, even though the snoop packets flow in the opposite direction to the normal packet. That is, a coherent master sends normal request and receives responses, but receives snoop requests and sends snoop responses (vice versa for the slave). These two distinct phases now use different access functions, as described below. Starting with the functional access, a master sends a request to a slave through sendFunctional, and the request packet is turned into a response before the call returns. In a system without cache coherence, this is all that is needed from the functional interface. For the cache-coherent scenario, a slave also sends snoop requests to coherent masters through sendFunctionalSnoop, with responses returned within the same packet pointer. This is currently used by the bus and caches, and the LSQ of the O3 CPU. The send/recvFunctional and send/recvFunctionalSnoop are moved from the Port super class to the appropriate subclass. Atomic accesses follow the same flow as functional accesses, with request being sent from master to slave through sendAtomic. In the case of cache-coherent ports, a slave can send snoop requests to a master through sendAtomicSnoop. Just as for the functional access methods, the atomic send and receive member functions are moved to the appropriate subclasses. The timing access methods are different from the functional and atomic in that requests and responses are separated in time and send/recvTiming are used for both directions. Hence, a master uses sendTiming to send a request to a slave, and a slave uses sendTiming to send a response back to a master, at a later point in time. Snoop requests and responses travel in the opposite direction, similar to what happens in functional and atomic accesses. With the introduction of this patch, it is possible to determine the direction of packets in the bus, and no longer necessary to look for both a master and a slave port with the requested port id. In contrast to the normal recvFunctional, recvAtomic and recvTiming that are pure virtual functions, the recvFunctionalSnoop, recvAtomicSnoop and recvTimingSnoop have a default implementation that calls panic. This is to allow non-coherent master and slave ports to not implement these functions.
2012-04-14 11:45:07 +02:00
WalkerSenderState * senderState =
dynamic_cast<WalkerSenderState *>(pkt->senderState);
pkt->senderState = senderState->saved;
WalkerState * senderWalk = senderState->senderWalk;
bool walkComplete = senderWalk->recvPacket(pkt);
delete senderState;
if (walkComplete) {
std::list<WalkerState *>::iterator iter;
for (iter = currStates.begin(); iter != currStates.end(); iter++) {
WalkerState * walkerState = *(iter);
if (walkerState == senderWalk) {
iter = currStates.erase(iter);
break;
}
}
MEM: Separate snoops and normal memory requests/responses This patch introduces port access methods that separates snoop request/responses from normal memory request/responses. The differentiation is made for functional, atomic and timing accesses and builds on the introduction of master and slave ports. Before the introduction of this patch, the packets belonging to the different phases of the protocol (request -> [forwarded snoop request -> snoop response]* -> response) all use the same port access functions, even though the snoop packets flow in the opposite direction to the normal packet. That is, a coherent master sends normal request and receives responses, but receives snoop requests and sends snoop responses (vice versa for the slave). These two distinct phases now use different access functions, as described below. Starting with the functional access, a master sends a request to a slave through sendFunctional, and the request packet is turned into a response before the call returns. In a system without cache coherence, this is all that is needed from the functional interface. For the cache-coherent scenario, a slave also sends snoop requests to coherent masters through sendFunctionalSnoop, with responses returned within the same packet pointer. This is currently used by the bus and caches, and the LSQ of the O3 CPU. The send/recvFunctional and send/recvFunctionalSnoop are moved from the Port super class to the appropriate subclass. Atomic accesses follow the same flow as functional accesses, with request being sent from master to slave through sendAtomic. In the case of cache-coherent ports, a slave can send snoop requests to a master through sendAtomicSnoop. Just as for the functional access methods, the atomic send and receive member functions are moved to the appropriate subclasses. The timing access methods are different from the functional and atomic in that requests and responses are separated in time and send/recvTiming are used for both directions. Hence, a master uses sendTiming to send a request to a slave, and a slave uses sendTiming to send a response back to a master, at a later point in time. Snoop requests and responses travel in the opposite direction, similar to what happens in functional and atomic accesses. With the introduction of this patch, it is possible to determine the direction of packets in the bus, and no longer necessary to look for both a master and a slave port with the requested port id. In contrast to the normal recvFunctional, recvAtomic and recvTiming that are pure virtual functions, the recvFunctionalSnoop, recvAtomicSnoop and recvTimingSnoop have a default implementation that calls panic. This is to allow non-coherent master and slave ports to not implement these functions.
2012-04-14 11:45:07 +02:00
delete senderWalk;
// Since we block requests when another is outstanding, we
// need to check if there is a waiting request to be serviced
if (currStates.size()) {
WalkerState * newState = currStates.front();
if (!newState->wasStarted())
newState->startWalk();
}
}
return true;
}
void
Walker::WalkerPort::recvRetry()
{
walker->recvRetry();
}
void
Walker::recvRetry()
{
std::list<WalkerState *>::iterator iter;
for (iter = currStates.begin(); iter != currStates.end(); iter++) {
WalkerState * walkerState = *(iter);
if (walkerState->isRetrying()) {
walkerState->retry();
}
}
}
bool Walker::sendTiming(WalkerState* sendingState, PacketPtr pkt)
{
pkt->senderState = new WalkerSenderState(sendingState, pkt->senderState);
MEM: Separate requests and responses for timing accesses This patch moves send/recvTiming and send/recvTimingSnoop from the Port base class to the MasterPort and SlavePort, and also splits them into separate member functions for requests and responses: send/recvTimingReq, send/recvTimingResp, and send/recvTimingSnoopReq, send/recvTimingSnoopResp. A master port sends requests and receives responses, and also receives snoop requests and sends snoop responses. A slave port has the reciprocal behaviour as it receives requests and sends responses, and sends snoop requests and receives snoop responses. For all MemObjects that have only master ports or slave ports (but not both), e.g. a CPU, or a PIO device, this patch merely adds more clarity to what kind of access is taking place. For example, a CPU port used to call sendTiming, and will now call sendTimingReq. Similarly, a response previously came back through recvTiming, which is now recvTimingResp. For the modules that have both master and slave ports, e.g. the bus, the behaviour was previously relying on branches based on pkt->isRequest(), and this is now replaced with a direct call to the apprioriate member function depending on the type of access. Please note that send/recvRetry is still shared by all the timing accessors and remains in the Port base class for now (to maintain the current bus functionality and avoid changing the statistics of all regressions). The packet queue is split into a MasterPort and SlavePort version to facilitate the use of the new timing accessors. All uses of the PacketQueue are updated accordingly. With this patch, the type of packet (request or response) is now well defined for each type of access, and asserts on pkt->isRequest() and pkt->isResponse() are now moved to the appropriate send member functions. It is also worth noting that sendTimingSnoopReq no longer returns a boolean, as the semantics do not alow snoop requests to be rejected or stalled. All these assumptions are now excplicitly part of the port interface itself.
2012-05-01 19:40:42 +02:00
return port.sendTimingReq(pkt);
}
BaseMasterPort &
Walker::getMasterPort(const std::string &if_name, PortID idx)
{
if (if_name == "port")
return port;
else
return MemObject::getMasterPort(if_name, idx);
}
void
Walker::WalkerState::initState(ThreadContext * _tc,
BaseTLB::Mode _mode, bool _isTiming)
{
assert(state == Ready);
started = false;
tc = _tc;
mode = _mode;
timing = _isTiming;
}
Fault
Walker::WalkerState::startWalk()
{
Fault fault = NoFault;
assert(started == false);
started = true;
setupWalk(req->getVaddr());
if (timing) {
nextState = state;
state = Waiting;
timingFault = NoFault;
sendPackets();
} else {
do {
walker->port.sendAtomic(read);
PacketPtr write = NULL;
fault = stepWalk(write);
assert(fault == NoFault || read == NULL);
state = nextState;
nextState = Ready;
if (write)
walker->port.sendAtomic(write);
} while(read);
state = Ready;
nextState = Waiting;
}
return fault;
}
Fault
Walker::WalkerState::startFunctional(Addr &addr, unsigned &logBytes)
{
Fault fault = NoFault;
assert(started == false);
started = true;
setupWalk(addr);
do {
walker->port.sendFunctional(read);
// On a functional access (page table lookup), writes should
// not happen so this pointer is ignored after stepWalk
PacketPtr write = NULL;
fault = stepWalk(write);
assert(fault == NoFault || read == NULL);
state = nextState;
nextState = Ready;
} while(read);
logBytes = entry.logBytes;
addr = entry.paddr;
return fault;
}
Fault
Walker::WalkerState::stepWalk(PacketPtr &write)
{
assert(state != Ready && state != Waiting);
Fault fault = NoFault;
write = NULL;
PageTableEntry pte;
if (dataSize == 8)
pte = read->get<uint64_t>();
else
pte = read->get<uint32_t>();
VAddr vaddr = entry.vaddr;
bool uncacheable = pte.pcd;
Addr nextRead = 0;
bool doWrite = false;
bool doTLBInsert = false;
bool doEndWalk = false;
bool badNX = pte.nx && mode == BaseTLB::Execute && enableNX;
switch(state) {
case LongPML4:
DPRINTF(PageTableWalker,
"Got long mode PML4 entry %#016x.\n", (uint64_t)pte);
nextRead = ((uint64_t)pte & (mask(40) << 12)) + vaddr.longl3 * dataSize;
doWrite = !pte.a;
pte.a = 1;
entry.writable = pte.w;
entry.user = pte.u;
if (badNX || !pte.p) {
doEndWalk = true;
fault = pageFault(pte.p);
break;
}
entry.noExec = pte.nx;
nextState = LongPDP;
break;
case LongPDP:
DPRINTF(PageTableWalker,
"Got long mode PDP entry %#016x.\n", (uint64_t)pte);
nextRead = ((uint64_t)pte & (mask(40) << 12)) + vaddr.longl2 * dataSize;
doWrite = !pte.a;
pte.a = 1;
entry.writable = entry.writable && pte.w;
entry.user = entry.user && pte.u;
if (badNX || !pte.p) {
doEndWalk = true;
fault = pageFault(pte.p);
break;
}
nextState = LongPD;
break;
case LongPD:
DPRINTF(PageTableWalker,
"Got long mode PD entry %#016x.\n", (uint64_t)pte);
doWrite = !pte.a;
pte.a = 1;
entry.writable = entry.writable && pte.w;
entry.user = entry.user && pte.u;
if (badNX || !pte.p) {
doEndWalk = true;
fault = pageFault(pte.p);
break;
}
if (!pte.ps) {
// 4 KB page
entry.logBytes = 12;
nextRead =
((uint64_t)pte & (mask(40) << 12)) + vaddr.longl1 * dataSize;
nextState = LongPTE;
break;
} else {
// 2 MB page
entry.logBytes = 21;
entry.paddr = (uint64_t)pte & (mask(31) << 21);
entry.uncacheable = uncacheable;
entry.global = pte.g;
entry.patBit = bits(pte, 12);
entry.vaddr = entry.vaddr & ~((2 * (1 << 20)) - 1);
doTLBInsert = true;
doEndWalk = true;
break;
}
case LongPTE:
DPRINTF(PageTableWalker,
"Got long mode PTE entry %#016x.\n", (uint64_t)pte);
doWrite = !pte.a;
pte.a = 1;
entry.writable = entry.writable && pte.w;
entry.user = entry.user && pte.u;
if (badNX || !pte.p) {
doEndWalk = true;
fault = pageFault(pte.p);
break;
}
entry.paddr = (uint64_t)pte & (mask(40) << 12);
entry.uncacheable = uncacheable;
entry.global = pte.g;
entry.patBit = bits(pte, 12);
entry.vaddr = entry.vaddr & ~((4 * (1 << 10)) - 1);
doTLBInsert = true;
doEndWalk = true;
break;
case PAEPDP:
DPRINTF(PageTableWalker,
"Got legacy mode PAE PDP entry %#08x.\n", (uint32_t)pte);
nextRead = ((uint64_t)pte & (mask(40) << 12)) + vaddr.pael2 * dataSize;
if (!pte.p) {
doEndWalk = true;
fault = pageFault(pte.p);
break;
}
nextState = PAEPD;
break;
case PAEPD:
DPRINTF(PageTableWalker,
"Got legacy mode PAE PD entry %#08x.\n", (uint32_t)pte);
doWrite = !pte.a;
pte.a = 1;
entry.writable = pte.w;
entry.user = pte.u;
if (badNX || !pte.p) {
doEndWalk = true;
fault = pageFault(pte.p);
break;
}
if (!pte.ps) {
// 4 KB page
entry.logBytes = 12;
nextRead = ((uint64_t)pte & (mask(40) << 12)) + vaddr.pael1 * dataSize;
nextState = PAEPTE;
break;
} else {
// 2 MB page
entry.logBytes = 21;
entry.paddr = (uint64_t)pte & (mask(31) << 21);
entry.uncacheable = uncacheable;
entry.global = pte.g;
entry.patBit = bits(pte, 12);
entry.vaddr = entry.vaddr & ~((2 * (1 << 20)) - 1);
doTLBInsert = true;
doEndWalk = true;
break;
}
case PAEPTE:
DPRINTF(PageTableWalker,
"Got legacy mode PAE PTE entry %#08x.\n", (uint32_t)pte);
doWrite = !pte.a;
pte.a = 1;
entry.writable = entry.writable && pte.w;
entry.user = entry.user && pte.u;
if (badNX || !pte.p) {
doEndWalk = true;
fault = pageFault(pte.p);
break;
}
entry.paddr = (uint64_t)pte & (mask(40) << 12);
entry.uncacheable = uncacheable;
entry.global = pte.g;
entry.patBit = bits(pte, 7);
entry.vaddr = entry.vaddr & ~((4 * (1 << 10)) - 1);
doTLBInsert = true;
doEndWalk = true;
break;
case PSEPD:
DPRINTF(PageTableWalker,
"Got legacy mode PSE PD entry %#08x.\n", (uint32_t)pte);
doWrite = !pte.a;
pte.a = 1;
entry.writable = pte.w;
entry.user = pte.u;
if (!pte.p) {
doEndWalk = true;
fault = pageFault(pte.p);
break;
}
if (!pte.ps) {
// 4 KB page
entry.logBytes = 12;
nextRead =
((uint64_t)pte & (mask(20) << 12)) + vaddr.norml2 * dataSize;
nextState = PTE;
break;
} else {
// 4 MB page
entry.logBytes = 21;
entry.paddr = bits(pte, 20, 13) << 32 | bits(pte, 31, 22) << 22;
entry.uncacheable = uncacheable;
entry.global = pte.g;
entry.patBit = bits(pte, 12);
entry.vaddr = entry.vaddr & ~((4 * (1 << 20)) - 1);
doTLBInsert = true;
doEndWalk = true;
break;
}
case PD:
DPRINTF(PageTableWalker,
"Got legacy mode PD entry %#08x.\n", (uint32_t)pte);
doWrite = !pte.a;
pte.a = 1;
entry.writable = pte.w;
entry.user = pte.u;
if (!pte.p) {
doEndWalk = true;
fault = pageFault(pte.p);
break;
}
// 4 KB page
entry.logBytes = 12;
nextRead = ((uint64_t)pte & (mask(20) << 12)) + vaddr.norml2 * dataSize;
nextState = PTE;
break;
case PTE:
DPRINTF(PageTableWalker,
"Got legacy mode PTE entry %#08x.\n", (uint32_t)pte);
doWrite = !pte.a;
pte.a = 1;
entry.writable = pte.w;
entry.user = pte.u;
if (!pte.p) {
doEndWalk = true;
fault = pageFault(pte.p);
break;
}
entry.paddr = (uint64_t)pte & (mask(20) << 12);
entry.uncacheable = uncacheable;
entry.global = pte.g;
entry.patBit = bits(pte, 7);
entry.vaddr = entry.vaddr & ~((4 * (1 << 10)) - 1);
doTLBInsert = true;
doEndWalk = true;
break;
default:
panic("Unknown page table walker state %d!\n");
}
if (doEndWalk) {
if (doTLBInsert)
if (!functional)
walker->tlb->insert(entry.vaddr, entry);
endWalk();
} else {
PacketPtr oldRead = read;
//If we didn't return, we're setting up another read.
Request::Flags flags = oldRead->req->getFlags();
flags.set(Request::UNCACHEABLE, uncacheable);
RequestPtr request =
new Request(nextRead, oldRead->getSize(), flags, walker->masterId);
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
read = new Packet(request, MemCmd::ReadReq);
read->allocate();
// If we need to write, adjust the read packet to write the modified
// value back to memory.
if (doWrite) {
write = oldRead;
write->set<uint64_t>(pte);
write->cmd = MemCmd::WriteReq;
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
write->clearDest();
} else {
write = NULL;
delete oldRead->req;
delete oldRead;
}
}
return fault;
}
void
Walker::WalkerState::endWalk()
{
nextState = Ready;
delete read->req;
delete read;
read = NULL;
}
void
Walker::WalkerState::setupWalk(Addr vaddr)
{
VAddr addr = vaddr;
CR3 cr3 = tc->readMiscRegNoEffect(MISCREG_CR3);
// Check if we're in long mode or not
Efer efer = tc->readMiscRegNoEffect(MISCREG_EFER);
dataSize = 8;
Addr topAddr;
if (efer.lma) {
// Do long mode.
state = LongPML4;
topAddr = (cr3.longPdtb << 12) + addr.longl4 * dataSize;
enableNX = efer.nxe;
} else {
// We're in some flavor of legacy mode.
CR4 cr4 = tc->readMiscRegNoEffect(MISCREG_CR4);
if (cr4.pae) {
// Do legacy PAE.
state = PAEPDP;
topAddr = (cr3.paePdtb << 5) + addr.pael3 * dataSize;
enableNX = efer.nxe;
} else {
dataSize = 4;
topAddr = (cr3.pdtb << 12) + addr.norml2 * dataSize;
if (cr4.pse) {
// Do legacy PSE.
state = PSEPD;
} else {
// Do legacy non PSE.
state = PD;
}
enableNX = false;
}
}
nextState = Ready;
entry.vaddr = vaddr;
Request::Flags flags = Request::PHYSICAL;
if (cr3.pcd)
flags.set(Request::UNCACHEABLE);
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
RequestPtr request = new Request(topAddr, dataSize, flags,
walker->masterId);
read = new Packet(request, MemCmd::ReadReq);
read->allocate();
}
bool
Walker::WalkerState::recvPacket(PacketPtr pkt)
{
MEM: Separate snoops and normal memory requests/responses This patch introduces port access methods that separates snoop request/responses from normal memory request/responses. The differentiation is made for functional, atomic and timing accesses and builds on the introduction of master and slave ports. Before the introduction of this patch, the packets belonging to the different phases of the protocol (request -> [forwarded snoop request -> snoop response]* -> response) all use the same port access functions, even though the snoop packets flow in the opposite direction to the normal packet. That is, a coherent master sends normal request and receives responses, but receives snoop requests and sends snoop responses (vice versa for the slave). These two distinct phases now use different access functions, as described below. Starting with the functional access, a master sends a request to a slave through sendFunctional, and the request packet is turned into a response before the call returns. In a system without cache coherence, this is all that is needed from the functional interface. For the cache-coherent scenario, a slave also sends snoop requests to coherent masters through sendFunctionalSnoop, with responses returned within the same packet pointer. This is currently used by the bus and caches, and the LSQ of the O3 CPU. The send/recvFunctional and send/recvFunctionalSnoop are moved from the Port super class to the appropriate subclass. Atomic accesses follow the same flow as functional accesses, with request being sent from master to slave through sendAtomic. In the case of cache-coherent ports, a slave can send snoop requests to a master through sendAtomicSnoop. Just as for the functional access methods, the atomic send and receive member functions are moved to the appropriate subclasses. The timing access methods are different from the functional and atomic in that requests and responses are separated in time and send/recvTiming are used for both directions. Hence, a master uses sendTiming to send a request to a slave, and a slave uses sendTiming to send a response back to a master, at a later point in time. Snoop requests and responses travel in the opposite direction, similar to what happens in functional and atomic accesses. With the introduction of this patch, it is possible to determine the direction of packets in the bus, and no longer necessary to look for both a master and a slave port with the requested port id. In contrast to the normal recvFunctional, recvAtomic and recvTiming that are pure virtual functions, the recvFunctionalSnoop, recvAtomicSnoop and recvTimingSnoop have a default implementation that calls panic. This is to allow non-coherent master and slave ports to not implement these functions.
2012-04-14 11:45:07 +02:00
assert(pkt->isResponse());
assert(inflight);
assert(state == Waiting);
assert(!read);
inflight--;
if (pkt->isRead()) {
state = nextState;
nextState = Ready;
PacketPtr write = NULL;
read = pkt;
timingFault = stepWalk(write);
state = Waiting;
assert(timingFault == NoFault || read == NULL);
if (write) {
writes.push_back(write);
}
sendPackets();
MEM: Separate snoops and normal memory requests/responses This patch introduces port access methods that separates snoop request/responses from normal memory request/responses. The differentiation is made for functional, atomic and timing accesses and builds on the introduction of master and slave ports. Before the introduction of this patch, the packets belonging to the different phases of the protocol (request -> [forwarded snoop request -> snoop response]* -> response) all use the same port access functions, even though the snoop packets flow in the opposite direction to the normal packet. That is, a coherent master sends normal request and receives responses, but receives snoop requests and sends snoop responses (vice versa for the slave). These two distinct phases now use different access functions, as described below. Starting with the functional access, a master sends a request to a slave through sendFunctional, and the request packet is turned into a response before the call returns. In a system without cache coherence, this is all that is needed from the functional interface. For the cache-coherent scenario, a slave also sends snoop requests to coherent masters through sendFunctionalSnoop, with responses returned within the same packet pointer. This is currently used by the bus and caches, and the LSQ of the O3 CPU. The send/recvFunctional and send/recvFunctionalSnoop are moved from the Port super class to the appropriate subclass. Atomic accesses follow the same flow as functional accesses, with request being sent from master to slave through sendAtomic. In the case of cache-coherent ports, a slave can send snoop requests to a master through sendAtomicSnoop. Just as for the functional access methods, the atomic send and receive member functions are moved to the appropriate subclasses. The timing access methods are different from the functional and atomic in that requests and responses are separated in time and send/recvTiming are used for both directions. Hence, a master uses sendTiming to send a request to a slave, and a slave uses sendTiming to send a response back to a master, at a later point in time. Snoop requests and responses travel in the opposite direction, similar to what happens in functional and atomic accesses. With the introduction of this patch, it is possible to determine the direction of packets in the bus, and no longer necessary to look for both a master and a slave port with the requested port id. In contrast to the normal recvFunctional, recvAtomic and recvTiming that are pure virtual functions, the recvFunctionalSnoop, recvAtomicSnoop and recvTimingSnoop have a default implementation that calls panic. This is to allow non-coherent master and slave ports to not implement these functions.
2012-04-14 11:45:07 +02:00
} else {
sendPackets();
}
if (inflight == 0 && read == NULL && writes.size() == 0) {
state = Ready;
nextState = Waiting;
if (timingFault == NoFault) {
/*
* Finish the translation. Now that we now the right entry is
* in the TLB, this should work with no memory accesses.
* There could be new faults unrelated to the table walk like
* permissions violations, so we'll need the return value as
* well.
*/
bool delayedResponse;
Fault fault = walker->tlb->translate(req, tc, NULL, mode,
delayedResponse, true);
assert(!delayedResponse);
// Let the CPU continue.
translation->finish(fault, req, tc, mode);
} else {
// There was a fault during the walk. Let the CPU know.
translation->finish(timingFault, req, tc, mode);
}
return true;
}
return false;
}
void
Walker::WalkerState::sendPackets()
{
//If we're already waiting for the port to become available, just return.
if (retrying)
return;
//Reads always have priority
if (read) {
PacketPtr pkt = read;
read = NULL;
inflight++;
if (!walker->sendTiming(this, pkt)) {
retrying = true;
read = pkt;
inflight--;
return;
}
}
//Send off as many of the writes as we can.
while (writes.size()) {
PacketPtr write = writes.back();
writes.pop_back();
inflight++;
if (!walker->sendTiming(this, write)) {
retrying = true;
writes.push_back(write);
inflight--;
return;
}
}
}
bool
Walker::WalkerState::isRetrying()
{
return retrying;
}
bool
Walker::WalkerState::isTiming()
{
return timing;
}
bool
Walker::WalkerState::wasStarted()
{
return started;
}
void
Walker::WalkerState::retry()
{
retrying = false;
sendPackets();
}
Fault
Walker::WalkerState::pageFault(bool present)
{
DPRINTF(PageTableWalker, "Raising page fault.\n");
HandyM5Reg m5reg = tc->readMiscRegNoEffect(MISCREG_M5_REG);
2009-04-09 07:21:27 +02:00
if (mode == BaseTLB::Execute && !enableNX)
mode = BaseTLB::Read;
return new PageFault(entry.vaddr, present, mode, m5reg.cpl == 3, false);
}
/* end namespace X86ISA */ }
X86ISA::Walker *
X86PagetableWalkerParams::create()
{
return new X86ISA::Walker(this);
}