First pass at snooping stuff that compiles and doesn't break.

Still need:
-Handle NACK's on the recieve side
-Distinguish top level caches
-Handle repsonses from caches failing the fast path
-Handle BusError and propogate it
-Fix the invalidate packet associated with snooping in the cache

src/mem/bus.cc:
    Make sure to snoop on functional accesses
src/mem/cache/base_cache.cc:
    Wait to make a request into a response until it is ready to be issued
src/mem/cache/base_cache.hh:
    Support range changes for snoops
    Set up snoop responses for cache->cache transfers
src/mem/cache/cache_impl.hh:
    Only access the cache if it wasn't satisfied by cache->cache transfer
    Handle snoop phases (detect block, then snoop)
    Fix functional access to work properly (still need to fix snoop path for functional accesses)

--HG--
extra : convert_revision : 4c25f11d7a996c1f56f4f7b55dde87a344e5fdf8

6 files changed, 237 insertions, 20 deletions

diff --git a/src/mem/bus.cc b/src/mem/bus.cc
index cf9e54e62..3c5283a77 100644
--- a/src/mem/bus.cc
+++ b/src/mem/bus.cc
@@ -252,6 +252,7 @@ Bus::recvFunctional(Packet *pkt)
     DPRINTF(Bus, "recvFunctional: packet src %d dest %d addr 0x%x cmd %s\n",
             pkt->getSrc(), pkt->getDest(), pkt->getAddr(), pkt->cmdString());
     assert(pkt->getDest() == Packet::Broadcast);
+    atomicSnoop(pkt);
     findPort(pkt->getAddr(), pkt->getSrc())->sendFunctional(pkt);
 }
 
diff --git a/src/mem/cache/base_cache.cc b/src/mem/cache/base_cache.cc
index a172847df..e6138e320 100644
--- a/src/mem/cache/base_cache.cc
+++ b/src/mem/cache/base_cache.cc
@@ -199,7 +199,9 @@ BaseCache::CacheEvent::process()
         return;
     }
     //Response
-    //Know the packet to send, no need to mark in service (must succed)
+    //Know the packet to send
+    pkt->result = Packet::Success;
+    pkt->makeTimingResponse();
     assert(cachePort->sendTiming(pkt));
 }
 
diff --git a/src/mem/cache/base_cache.hh b/src/mem/cache/base_cache.hh
index 069dbab58..49999dcb4 100644
--- a/src/mem/cache/base_cache.hh
+++ b/src/mem/cache/base_cache.hh
@@ -127,6 +127,8 @@ class BaseCache : public MemObject
     CachePort *cpuSidePort;
     CachePort *memSidePort;
 
+    bool snoopRangesSent;
+
   public:
     virtual Port *getPort(const std::string &if_name, int idx = -1);
 
@@ -149,17 +151,22 @@ class BaseCache : public MemObject
 
     void recvStatusChange(Port::Status status, bool isCpuSide)
     {
-        if (status == Port::RangeChange)
-        {
-            if (!isCpuSide)
-            {
+        if (status == Port::RangeChange){
+            if (!isCpuSide) {
                 cpuSidePort->sendStatusChange(Port::RangeChange);
+                if (topLevelCache && !snoopRangesSent) {
+                    snoopRangesSent = true;
+                    memSidePort->sendStatusChange(Port::RangeChange);
+                }
             }
-            else
-            {
+            else {
                 memSidePort->sendStatusChange(Port::RangeChange);
             }
         }
+        else if (status == Port::SnoopSquash) {
+            assert(snoopPhase2);
+            snoopPhase2 = false;
+        }
     }
 
     virtual Packet *getPacket()
@@ -205,6 +212,10 @@ class BaseCache : public MemObject
     /** True if this cache is connected to the CPU. */
     bool topLevelCache;
 
+
+    /** True if we are now in phase 2 of the snoop process. */
+    bool snoopPhase2;
+
     /** Stores time the cache blocked for statistics. */
     Tick blockedCycle;
 
@@ -332,6 +343,7 @@ class BaseCache : public MemObject
         //Start ports at null if more than one is created we should panic
         cpuSidePort = NULL;
         memSidePort = NULL;
+        snoopRangesSent = false;
     }
 
     virtual void init();
@@ -519,8 +531,6 @@ class BaseCache : public MemObject
         if (!pkt->req->isUncacheable()) {
             missLatency[pkt->cmdToIndex()][pkt->req->getThreadNum()] += time - pkt->time;
         }
-        pkt->makeTimingResponse();
-        pkt->result = Packet::Success;
         CacheEvent *reqCpu = new CacheEvent(cpuSidePort, pkt);
         reqCpu->schedule(time);
     }
@@ -529,10 +539,12 @@ class BaseCache : public MemObject
      * Suppliess the data if cache to cache transfers are enabled.
      * @param pkt The bus transaction to fulfill.
      */
-    void respondToSnoop(Packet *pkt)
+    void respondToSnoop(Packet *pkt, Tick time)
     {
-        assert("Implement\n" && 0);
+//        assert("Implement\n" && 0);
 //	mi->respond(pkt,curTick + hitLatency);
+        CacheEvent *reqMem = new CacheEvent(memSidePort, pkt);
+        reqMem->schedule(time);
     }
 
     /**
@@ -551,6 +563,16 @@ class BaseCache : public MemObject
         else
         {
             //This is where snoops get updated
+            AddrRangeList dummy;
+            if (!topLevelCache)
+            {
+                cpuSidePort->getPeerAddressRanges(dummy, snoop);
+            }
+            else
+            {
+                snoop.push_back(RangeSize(0,-1));
+            }
+
             return;
         }
     }
diff --git a/src/mem/cache/cache_impl.hh b/src/mem/cache/cache_impl.hh
index 11cd84e88..bea495f9f 100644
--- a/src/mem/cache/cache_impl.hh
+++ b/src/mem/cache/cache_impl.hh
@@ -63,14 +63,26 @@ doTimingAccess(Packet *pkt, CachePort *cachePort, bool isCpuSide)
         if (pkt->isWrite() && (pkt->req->getFlags() & LOCKED)) {
             pkt->req->setScResult(1);
         }
-        access(pkt);
+        if (!(pkt->flags & SATISFIED)) {
+            access(pkt);
+        }
     }
     else
     {
         if (pkt->isResponse())
             handleResponse(pkt);
-        else
-            snoop(pkt);
+        else {
+            //Check if we are in phase1
+            if (!snoopPhase2) {
+                snoopPhase2 = true;
+            }
+            else {
+                //Check if we should do the snoop
+                if (pkt->flags && SNOOP_COMMIT)
+                    snoop(pkt);
+                snoopPhase2 = false;
+            }
+        }
     }
     return true;
 }
@@ -117,7 +129,7 @@ doFunctionalAccess(Packet *pkt, bool isCpuSide)
             assert("Can't handle LL/SC on functional path\n");
         }
 
-        probe(pkt, true);
+        probe(pkt, false);
         //TEMP ALWAYS SUCCESFUL FOR NOW
         pkt->result = Packet::Success;
     }
@@ -126,7 +138,7 @@ doFunctionalAccess(Packet *pkt, bool isCpuSide)
         if (pkt->isResponse())
             handleResponse(pkt);
         else
-            snoopProbe(pkt, true);
+            snoopProbe(pkt, false);
     }
 }
 
@@ -372,7 +384,7 @@ template<class TagStore, class Buffering, class Coherence>
 void
 Cache<TagStore,Buffering,Coherence>::snoop(Packet * &pkt)
 {
-
+    DPRINTF(Cache, "SNOOPING");
     Addr blk_addr = pkt->getAddr() & ~(Addr(blkSize-1));
     BlkType *blk = tags->findBlock(pkt);
     MSHR *mshr = missQueue->findMSHR(blk_addr);
@@ -385,7 +397,10 @@ Cache<TagStore,Buffering,Coherence>::snoop(Packet * &pkt)
                     //If the outstanding request was an invalidate (upgrade,readex,..)
                     //Then we need to ACK the request until we get the data
                     //Also NACK if the outstanding request is not a cachefill (writeback)
+                    pkt->flags |= SATISFIED;
                     pkt->flags |= NACKED_LINE;
+                    assert("Don't detect these on the other side yet\n");
+                    respondToSnoop(pkt, curTick + hitLatency);
                     return;
                 }
                 else {
@@ -398,6 +413,7 @@ Cache<TagStore,Buffering,Coherence>::snoop(Packet * &pkt)
                     //@todo Make it so that a read to a pending read can't be exclusive now.
 
                     //Set the address so find match works
+                    assert("Don't have invalidates yet\n");
                     invalidatePkt->addrOverride(pkt->getAddr());
 
                     //Append the invalidate on
@@ -433,7 +449,7 @@ Cache<TagStore,Buffering,Coherence>::snoop(Packet * &pkt)
                         assert(offset + pkt->getSize() <=blkSize);
                         memcpy(pkt->getPtr<uint8_t>(), mshr->pkt->getPtr<uint8_t>() + offset, pkt->getSize());
 
-                        respondToSnoop(pkt);
+                        respondToSnoop(pkt, curTick + hitLatency);
                     }
 
                     if (pkt->isInvalidate()) {
@@ -449,7 +465,7 @@ Cache<TagStore,Buffering,Coherence>::snoop(Packet * &pkt)
     bool satisfy = coherence->handleBusRequest(pkt,blk,mshr, new_state);
     if (satisfy) {
         tags->handleSnoop(blk, new_state, pkt);
-        respondToSnoop(pkt);
+        respondToSnoop(pkt, curTick + hitLatency);
         return;
     }
     tags->handleSnoop(blk, new_state);
@@ -517,7 +533,7 @@ Cache<TagStore,Buffering,Coherence>::probe(Packet * &pkt, bool update)
         missQueue->findWrites(blk_addr, writes);
 
         if (!update) {
-            memSidePort->sendFunctional(pkt);
+                memSidePort->sendFunctional(pkt);
             // Check for data in MSHR and writebuffer.
             if (mshr) {
                 warn("Found outstanding miss on an non-update probe");
diff --git a/tests/configs/o3-timing-mp.py b/tests/configs/o3-timing-mp.py
new file mode 100644
index 000000000..881c23156
--- /dev/null
+++ b/tests/configs/o3-timing-mp.py
@@ -0,0 +1,90 @@
+# Copyright (c) 2006 The Regents of The University of Michigan
+# 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
+
+import m5
+from m5.objects import *
+m5.AddToPath('../configs/common')
+from FullO3Config import *
+
+# --------------------
+# Base L1 Cache
+# ====================
+
+class L1(BaseCache):
+    latency = 1
+    block_size = 64
+    mshrs = 4
+    tgts_per_mshr = 8
+    protocol = CoherenceProtocol(protocol='moesi')
+
+# ----------------------
+# Base L2 Cache
+# ----------------------
+
+class L2(BaseCache):
+    block_size = 64
+    latency = 100
+    mshrs = 92
+    tgts_per_mshr = 16
+    write_buffers = 8
+
+nb_cores = 4
+cpus = [ DetailedO3CPU() for i in xrange(nb_cores) ]
+
+# system simulated
+system = System(cpu = cpus, physmem = PhysicalMemory(), membus =
+Bus())
+
+# l2cache & bus
+system.toL2Bus = Bus()
+system.l2c = L2(size='4MB', assoc=8)
+system.l2c.cpu_side = system.toL2Bus.port
+
+# connect l2c to membus
+system.l2c.mem_side = system.membus.port
+
+# add L1 caches
+for cpu in cpus:
+    cpu.addPrivateSplitL1Caches(L1(size = '32kB', assoc = 1),
+                                L1(size = '32kB', assoc = 4))
+    cpu.mem = cpu.dcache
+    # connect cpu level-1 caches to shared level-2 cache
+    cpu.connectMemPorts(system.toL2Bus)
+
+# connect memory to membus
+system.physmem.port = system.membus.port
+
+
+# -----------------------
+# run simulation
+# -----------------------
+
+root = Root( system = system )
+root.system.mem_mode = 'timing'
+root.trace.flags="Bus Cache"
+#root.trace.flags = "BusAddrRanges"
diff --git a/tests/configs/simple-atomic-mp.py b/tests/configs/simple-atomic-mp.py
new file mode 100644
index 000000000..cc1a36dda
--- /dev/null
+++ b/tests/configs/simple-atomic-mp.py
@@ -0,0 +1,86 @@
+# Copyright (c) 2006 The Regents of The University of Michigan
+# 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
+
+import m5
+from m5.objects import *
+
+# --------------------
+# Base L1 Cache
+# ====================
+
+class L1(BaseCache):
+    latency = 1
+    block_size = 64
+    mshrs = 4
+    tgts_per_mshr = 8
+    protocol = CoherenceProtocol(protocol='moesi')
+
+# ----------------------
+# Base L2 Cache
+# ----------------------
+
+class L2(BaseCache):
+    block_size = 64
+    latency = 100
+    mshrs = 92
+    tgts_per_mshr = 16
+    write_buffers = 8
+
+nb_cores = 4
+cpus = [ AtomicSimpleCPU() for i in xrange(nb_cores) ]
+
+# system simulated
+system = System(cpu = cpus, physmem = PhysicalMemory(), membus =
+Bus())
+
+# l2cache & bus
+system.toL2Bus = Bus()
+system.l2c = L2(size='4MB', assoc=8)
+system.l2c.cpu_side = system.toL2Bus.port
+
+# connect l2c to membus
+system.l2c.mem_side = system.membus.port
+
+# add L1 caches
+for cpu in cpus:
+    cpu.addPrivateSplitL1Caches(L1(size = '32kB', assoc = 1),
+                                L1(size = '32kB', assoc = 4))
+    cpu.mem = cpu.dcache
+    # connect cpu level-1 caches to shared level-2 cache
+    cpu.connectMemPorts(system.toL2Bus)
+
+# connect memory to membus
+system.physmem.port = system.membus.port
+
+
+# -----------------------
+# run simulation
+# -----------------------
+
+root = Root( system = system )
+root.system.mem_mode = 'atomic'