Make profiling work with multiple capabilities (+RTS -N)

This means that both time and heap profiling work for parallel
programs.  Main internal changes:

  - CCCS is no longer a global variable; it is now another
    pseudo-register in the StgRegTable struct.  Thus every
    Capability has its own CCCS.

  - There is a new built-in CCS called "IDLE", which records ticks for
    Capabilities in the idle state.  If you profile a single-threaded
    program with +RTS -N2, you'll see about 50% of time in "IDLE".

  - There is appropriate locking in rts/Profiling.c to protect the
    shared cost-centre-stack data structures.

This patch does enough to get it working, I have cut one big corner:
the cost-centre-stack data structure is still shared amongst all
Capabilities, which means that multiple Capabilities will race when
updating the "allocations" and "entries" fields of a CCS.  Not only
does this give unpredictable results, but it runs very slowly due to
cache line bouncing.

It is strongly recommended that you use -fno-prof-count-entries to
disable the "entries" count when profiling parallel programs. (I shall
add a note to this effect to the docs).

1 files changed, 11 insertions, 11 deletions

diff --git a/compiler/codeGen/CgProf.hs b/compiler/codeGen/CgProf.hs
index 13667c399a..3e247ff4d6 100644
--- a/compiler/codeGen/CgProf.hs
+++ b/compiler/codeGen/CgProf.hs
@@ -21,7 +21,7 @@ module CgProf (
         enterCostCentreThunk,
         enterCostCentreFun,
         costCentreFrom,
-	curCCS, curCCSAddr,
+        curCCS, storeCurCCS,
 	emitCostCentreDecl, emitCostCentreStackDecl, 
         emitSetCCC,
 
@@ -66,11 +66,10 @@ import Control.Monad
 
 -- Expression representing the current cost centre stack
 curCCS :: CmmExpr
-curCCS = CmmLoad curCCSAddr bWord
+curCCS = CmmReg (CmmGlobal CCCS)
 
--- Address of current CCS variable, for storing into
-curCCSAddr :: CmmExpr
-curCCSAddr = CmmLit (CmmLabel (mkCmmDataLabel rtsPackageId (fsLit "CCCS")))
+storeCurCCS :: CmmExpr -> CmmStmt
+storeCurCCS e = CmmAssign (CmmGlobal CCCS) e
 
 mkCCostCentre :: CostCentre -> CmmLit
 mkCCostCentre cc = CmmLabel (mkCCLabel cc)
@@ -135,14 +134,15 @@ profAlloc words ccs
 enterCostCentreThunk :: CmmExpr -> Code
 enterCostCentreThunk closure = 
   ifProfiling $ do 
-    stmtC $ CmmStore curCCSAddr (costCentreFrom closure)
+    stmtC $ storeCurCCS (costCentreFrom closure)
 
-enterCostCentreFun :: CostCentreStack -> CmmExpr -> Code
-enterCostCentreFun ccs closure =
+enterCostCentreFun :: CostCentreStack -> CmmExpr -> [GlobalReg] -> Code
+enterCostCentreFun ccs closure vols =
   ifProfiling $ do
     if isCurrentCCS ccs
-       then emitRtsCall rtsPackageId (fsLit "enterFunCCS")
-               [CmmHinted (costCentreFrom closure) AddrHint]
+       then emitRtsCallWithVols rtsPackageId (fsLit "enterFunCCS")
+               [CmmHinted (CmmReg (CmmGlobal BaseReg)) AddrHint,
+                CmmHinted (costCentreFrom closure) AddrHint] vols
        else return () -- top-level function, nothing to do
 
 ifProfiling :: Code -> Code
@@ -226,7 +226,7 @@ emitSetCCC cc tick push
     tmp <- newTemp bWord -- TODO FIXME NOW
     pushCostCentre tmp curCCS cc
     when tick $ stmtC (bumpSccCount (CmmReg (CmmLocal tmp)))
-    when push $ stmtC (CmmStore curCCSAddr (CmmReg (CmmLocal tmp)))
+    when push $ stmtC (storeCurCCS (CmmReg (CmmLocal tmp)))
 
 pushCostCentre :: LocalReg -> CmmExpr -> CostCentre -> Code
 pushCostCentre result ccs cc