ruby: call setMRU from L1 controllers, not from sequencer

Currently the sequencer calls the function setMRU that updates the replacement
policy structures with the first level caches.  While functionally this is
correct, the problem is that this requires calling findTagInSet() which is an
expensive function.  This patch removes the calls to setMRU from the sequencer.
All controllers should now update the replacement policy on their own.

The set and the way index for a given cache entry can be found within the
AbstractCacheEntry structure. Use these indicies to update the replacement
policy structures.

1 files changed, 24 insertions, 11 deletions

diff --git a/src/mem/protocol/MOESI_hammer-cache.sm b/src/mem/protocol/MOESI_hammer-cache.sm
index 0b7acb701..269e47dfd 100644
--- a/src/mem/protocol/MOESI_hammer-cache.sm
+++ b/src/mem/protocol/MOESI_hammer-cache.sm
@@ -857,9 +857,18 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
     }
   }
 
-  action(h_load_hit, "h", desc="Notify sequencer the load completed.") {
+  action(h_load_hit, "hd", desc="Notify sequencer the load completed.") {
     assert(is_valid(cache_entry));
     DPRINTF(RubySlicc, "%s\n", cache_entry.DataBlk);
+    L1Dcache.setMRU(cache_entry);
+    sequencer.readCallback(address, cache_entry.DataBlk, false,
+                           testAndClearLocalHit(cache_entry));
+  }
+
+  action(h_ifetch_hit, "hi", desc="Notify sequencer the ifetch completed.") {
+    assert(is_valid(cache_entry));
+    DPRINTF(RubySlicc, "%s\n", cache_entry.DataBlk);
+    L1Icache.setMRU(cache_entry);
     sequencer.readCallback(address, cache_entry.DataBlk, false,
                            testAndClearLocalHit(cache_entry));
   }
@@ -869,7 +878,8 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
     assert(is_valid(tbe));
     DPRINTF(RubySlicc, "%s\n", cache_entry.DataBlk);
     peek(responseToCache_in, ResponseMsg) {
-
+      L1Icache.setMRU(address);
+      L1Dcache.setMRU(address);
       sequencer.readCallback(address, cache_entry.DataBlk, true,
                  machineIDToMachineType(in_msg.Sender), tbe.InitialRequestTime,
                  tbe.ForwardRequestTime, tbe.FirstResponseTime);
@@ -880,6 +890,7 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
     assert(is_valid(cache_entry));
     DPRINTF(RubySlicc, "%s\n", cache_entry.DataBlk);
     peek(mandatoryQueue_in, RubyRequest) {
+      L1Dcache.setMRU(cache_entry);
       sequencer.writeCallback(address, cache_entry.DataBlk, false,
                               testAndClearLocalHit(cache_entry));
 
@@ -901,7 +912,8 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
     assert(is_valid(tbe));
     DPRINTF(RubySlicc, "%s\n", cache_entry.DataBlk);
     peek(responseToCache_in, ResponseMsg) {
-
+      L1Icache.setMRU(address);
+      L1Dcache.setMRU(address);
       sequencer.writeCallback(address, cache_entry.DataBlk, true,
               machineIDToMachineType(in_msg.Sender), tbe.InitialRequestTime,
               tbe.ForwardRequestTime, tbe.FirstResponseTime);
@@ -914,7 +926,8 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
     assert(is_valid(cache_entry));
     assert(is_valid(tbe));
     DPRINTF(RubySlicc, "%s\n", cache_entry.DataBlk);
-
+    L1Icache.setMRU(address);
+    L1Dcache.setMRU(address);
     sequencer.writeCallback(address, cache_entry.DataBlk, true,
             machineIDToMachineType(tbe.LastResponder), tbe.InitialRequestTime,
             tbe.ForwardRequestTime, tbe.FirstResponseTime);
@@ -1508,7 +1521,7 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
   }
 
   transition({S, SM, ISM}, Ifetch) {
-    h_load_hit;
+    h_ifetch_hit;
     uu_profileL1InstHit;
     k_popMandatoryQueue;
   }
@@ -1522,7 +1535,7 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
   }
 
   transition(SR, Ifetch, S) {
-    h_load_hit;
+    h_ifetch_hit;
     uu_profileL1InstMiss;
     uu_profileL2Hit;
     k_popMandatoryQueue;
@@ -1570,7 +1583,7 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
   }
 
   transition({O, OM, SS, MM_W, M_W}, {Ifetch}) {
-    h_load_hit;
+    h_ifetch_hit;
     uu_profileL1InstHit;
     k_popMandatoryQueue;
   }
@@ -1584,7 +1597,7 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
   }
 
   transition(OR, Ifetch, O) {
-    h_load_hit;
+    h_ifetch_hit;
     uu_profileL1InstMiss;
     uu_profileL2Hit;
     k_popMandatoryQueue;
@@ -1635,7 +1648,7 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
 
   // Transitions from Modified
   transition({MM, M}, {Ifetch}) {
-    h_load_hit;
+    h_ifetch_hit;
     uu_profileL1InstHit;
     k_popMandatoryQueue;
   }
@@ -1661,7 +1674,7 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
   }
 
   transition(MMR, Ifetch, MM) {
-    h_load_hit;
+    h_ifetch_hit;
     uu_profileL1InstMiss;
     uu_profileL2Hit;
     k_popMandatoryQueue;
@@ -1742,7 +1755,7 @@ machine({L1Cache, L2Cache}, "AMD Hammer-like protocol")
   }
 
   transition(MR, Ifetch, M) {
-    h_load_hit;
+    h_ifetch_hit;
     uu_profileL1InstMiss;
     uu_profileL2Hit;
     k_popMandatoryQueue;