Fwd: RFC: omit write barrier instructions unless '-threaded'

26 Feb 2013

From: Gabor Greif 
Date: Tue, 26 Feb 2013 10:45:52 +0100
Subject: Re: RFC: omit write barrier instructions unless '-threaded'
To: Simon Marlow 

Cool,

pushed

https://github.com/ghc/ghc/commit/b7ae3cd9273da044384de4a327a4a5d71190129e

Cheers,

   Gabor

On 2/26/13, Simon Marlow  wrote:
...
On 25/02/13 22:07, Gabor Greif wrote:
...
Thanks, this helped!
Something like this?
$ git show c3f5312
commit c3f53126fb0835cb21fc085e9b7cb5acb1ea1fba
Author: Gabor Greif 
Date:   Mon Feb 25 23:03:23 2013 +0100
Only emit %write_barrier primitive for THREADED_RTS

diff --git a/includes/Cmm.h b/includes/Cmm.h
index 1505b1c..bc5702c 100644
--- a/includes/Cmm.h
+++ b/includes/Cmm.h
@@ -583,6 +583,12 @@
  #define OVERWRITING_CLOSURE(c) /* nothing */
  #endif
+#ifdef THREADED_RTS
+#define prim_write_barrier prim %write_barrier()
+#else
+#define prim_write_barrier /* nothing */
+#endif
+
  /*
-----------------------------------------------------------------------------
     Ticky macros
--------------------------------------------------------------------------
*/
diff --git a/rts/Updates.h b/rts/Updates.h
index b4ff7d1..e6a2eb9 100644
--- a/rts/Updates.h
+++ b/rts/Updates.h
@@ -46,7 +46,7 @@
                                                                 \
      OVERWRITING_CLOSURE(p1);                                    \
      StgInd_indirectee(p1) = p2;                                 \
-    prim %write_barrier();                                      \
+    prim_write_barrier;                                         \
      SET_INFO(p1, stg_BLACKHOLE_info);                           \
      LDV_RECORD_CREATE(p1);                                      \
      bd = Bdescr(p1);                                           \
Okay to commit?
Yes, ok.
Cheers,
  Simon
...
Cheers,
Gabor
On 2/25/13, Simon Marlow  wrote:
...
On 25/02/13 14:53, Gabor Greif wrote:
...
On 2/25/13, Simon Marlow  wrote:
...
On 24/02/13 20:40, Gabor Greif wrote:
...
Hi all,
from what I gathered so far no emission of write barriers is needed
when
    - running on a uniprocessor (-threaded or not)
    - running on a multiprocessor sans having linked with -threaded.
Below patch suppresses the emission of 'lwsync' when no '-threaded'
is
specified on PPC only. So it does not cover both criteria above.
It helps me a lot since I have a uniprocessor target that does not
understand the 'lwsync' instruction (instruction is newer than core).
Still, I have some doubts:
    o do we want to extend this approach to other archs?
    o possibly suppress the emission of MO_WriteBarrier in
compiler/codeGen/StgCmmBind.hs
       (more care is needed to also cover compiler/cmm/CmmParse.y).
Anyway this should be a safe first step, and I'd like to push it.
When
we find a general solution be can back this commit out.
What do you think?
I don't think this is the right way to do it.  The -threaded flag is
not
supposed to affect code generation, it only changes the RTS that gets
linked in.  That is, you can compile all your code without -threaded
and
then just add -threaded at the link step, and it will work.
Hi Simon,
thanks for your review!
...
Instead, use conditional compilation in the RTS so that the
write_barrer() calls are only present when THREADED_RTS is on.
This is already done the way you suggest:
#define write_barrier()      /* nothing */
in includes/stg/SMP.h:368
Works perfectly, 'lwsync' only appears in .thr_*.o files.
My problem stems from another call:
emitPrimCall [] MO_WriteBarrier []
in compiler/codeGen/StgCmmBind.hs:614
or possibly
callishMachOps = listToUFM $
          map (\(x, y) -> (mkFastString x, y)) [
          ( "write_barrier", MO_WriteBarrier ),
          ( "memcpy", MO_Memcpy ),
          ( "memset", MO_Memset ),
          ( "memmove", MO_Memmove )
          -- ToDo: the rest, maybe
      ]
in compiler/cmm/CmmParse.y:920
I doubt these files should be compiled with a dependency on
THREADED_RTS, and indeed this symbol is only ever defined
(-optc-DTHREADED_RTS) for the C compiler, and thus not available when
compiling haskell source.
So I am still stumped.
The function where "lwsync" surfaces is:
stg_marked_upd_frame_info which is rts/Updates.cmm:46.
These come from rts/Updates.h:updateWithIndirection(), which expands to
some Cmm code that uses "prim %write_barrier()".  So you want to use
some conditional compilation to make that prim %write_barrier()
disappear when THREADED_RTS is off.  I suggest replacing it with
write_barrier(), and then #define write_barrier() in includes/Cmm.h.
Cheers,
  Simon
...
This one is creating a blackhole (IISC) by means of StgCmmBind.hs:614.
How can we go on from here?
Cheers,
Gabor
...
Cheers,
  Simon
...
Cheers,
Gabor
$ git show c0d682fb98f32e4ce5d27ff3a30f43b6cd70733e
commit c0d682fb98f32e4ce5d27ff3a30f43b6cd70733e
Author: Gabor Greif 
Date:   Fri Feb 22 18:47:00 2013 +0100
Do not emit barriers on PPC unless we go the threaded way
diff --git a/compiler/nativeGen/PPC/CodeGen.hs
b/compiler/nativeGen/PPC/CodeGen.hs
index 92eff36..6c33cca 100644
--- a/compiler/nativeGen/PPC/CodeGen.hs
+++ b/compiler/nativeGen/PPC/CodeGen.hs
@@ -1,9 +1,8 @@
-
-----------------------------------------------------------------------------
    --
    -- Generating machine code (instruction selection)
    --
--- (c) The University of Glasgow 1996-2004
+-- (c) The University of Glasgow 1996-2013
    --
-----------------------------------------------------------------------------
@@ -906,8 +905,10 @@ genCCall'
    -}
-genCCall' _ _ (PrimTarget MO_WriteBarrier) _ _
- = return $ unitOL LWSYNC
+genCCall' dflags _ (PrimTarget MO_WriteBarrier) _ _
+ = return (if WayThreaded `elem` ways dflags
+           then unitOL LWSYNC
+           else nilOL)
genCCall' _ _ (PrimTarget MO_Touch) _ _
     = return $ nilOL
_______________________________________________
ghc-devs mailing list
ghc-devs@haskell.org
http://www.haskell.org/mailman/listinfo/ghc-devs


    