Save a word in the info table on x86_64

Summary:
An info table with an SRT normally looks like this:

    StgWord64 srt_offset
    StgClosureInfo layout
    StgWord32 layout
    StgWord32 has_srt

But we only need 32 bits for srt_offset on x86_64, because the small
memory model requires that code segments are at most 2GB. So we can
optimise this to

    StgClosureInfo layout
    StgWord32 layout
    StgWord32 srt_offset

saving a word.  We can tell whether the info table has an SRT or not,
because zero is not a valid srt_offset, so zero still indicates that
there's no SRT.

Test Plan:
* validate
* For results, see D4632.

Reviewers: bgamari, niteria, osa1, erikd

Subscribers: thomie, carter

Differential Revision: https://phabricator.haskell.org/D4634

3 files changed, 54 insertions, 25 deletions

diff --git a/compiler/cmm/CmmBuildInfoTables.hs b/compiler/cmm/CmmBuildInfoTables.hs
index 498fded724..043f62f811 100644
--- a/compiler/cmm/CmmBuildInfoTables.hs
+++ b/compiler/cmm/CmmBuildInfoTables.hs
@@ -102,8 +102,6 @@ In each case, the info table points to the SRT.
 - info->srt is zero if there's no SRT, otherwise:
 - info->srt == 1 and info->f.srt_offset points to the SRT
 
-(but see TODO below, we can improve this)
-
 e.g. for a FUN with an SRT:
 
 StgFunInfoTable       +------+
@@ -115,6 +113,23 @@ StgStdInfoTable       +------+
   info->type          | ...  |
                       |------|
 
+On x86_64, we optimise the info table representation further.  The
+offset to the SRT can be stored in 32 bits (all code lives within a
+2GB region in x86_64's small memory model), so we can save a word in
+the info table by storing the srt_offset in the srt field, which is
+half a word.
+
+On x86_64 with TABLES_NEXT_TO_CODE:
+
+- info->srt is zero if there's no SRT, otherwise:
+- info->srt is an offset from the info pointer to the SRT object
+
+StgStdInfoTable       +------+
+  info->layout.ptrs   |      |
+  info->layout.nptrs  |      |
+  info->srt           |  ------------> offset to SRT object
+                      |------|
+
 
 EXAMPLE
 ^^^^^^^
@@ -281,9 +296,6 @@ implemented.
 As an alternative to [FUN]: we could merge the FUN's SRT with the FUN
 object itself.
 
-TODO: make info->srt be an offset to the SRT, or zero if none (save
-one word per info table that has an SRT)
-
 Note that there are many other optimisations that we could do, but
 aren't implemented. In general, we could omit any reference from an
 SRT if everything reachable from it is also reachable from the other
@@ -297,7 +309,6 @@ B = {Y,Z}
 C = {X,B}
 
 Here we could use C = {A} and therefore [Shortcut] C = A.
-
 -}
 
 -- ---------------------------------------------------------------------
diff --git a/compiler/cmm/CmmExpr.hs b/compiler/cmm/CmmExpr.hs
index 80ca1b1ef2..46f772731b 100644
--- a/compiler/cmm/CmmExpr.hs
+++ b/compiler/cmm/CmmExpr.hs
@@ -189,9 +189,13 @@ data CmmLit
         -- It is also used inside the NCG during when generating
         -- position-independent code.
   | CmmLabelDiffOff CLabel CLabel Int Width -- label1 - label2 + offset
-        -- The supported Widths depend on the architecture.  wordWidth
-        -- is supported on all architectures. Additionally W32 is
-        -- supported on x86_64 when using the small memory model.
+        -- In an expression, the width just has the effect of MO_SS_Conv
+        -- from wordWidth to the desired width.
+        --
+        -- In a static literal, the supported Widths depend on the
+        -- architecture: wordWidth is supported on all
+        -- architectures. Additionally W32 is supported on x86_64 when
+        -- using the small memory model.
 
   | CmmBlock {-# UNPACK #-} !BlockId     -- Code label
         -- Invariant: must be a continuation BlockId
@@ -224,7 +228,7 @@ cmmLitType cflags (CmmVec (l:ls))      = let ty = cmmLitType cflags l
                                             else panic "cmmLitType: CmmVec"
 cmmLitType dflags (CmmLabel lbl)       = cmmLabelType dflags lbl
 cmmLitType dflags (CmmLabelOff lbl _)  = cmmLabelType dflags lbl
-cmmLitType dflags (CmmLabelDiffOff _ _ _ width) = cmmBits width
+cmmLitType _      (CmmLabelDiffOff _ _ _ width) = cmmBits width
 cmmLitType dflags (CmmBlock _)         = bWord dflags
 cmmLitType dflags (CmmHighStackMark)   = bWord dflags
 
diff --git a/compiler/cmm/CmmInfo.hs b/compiler/cmm/CmmInfo.hs
index ea7923264f..4201fda36a 100644
--- a/compiler/cmm/CmmInfo.hs
+++ b/compiler/cmm/CmmInfo.hs
@@ -45,6 +45,7 @@ import Stream (Stream)
 import qualified Stream
 import Hoopl.Collections
 
+import Platform
 import Maybes
 import DynFlags
 import Panic
@@ -188,7 +189,7 @@ mkInfoTableContents dflags
 
   | StackRep frame <- smrep
   = do { (prof_lits, prof_data) <- mkProfLits dflags prof
-       ; let (srt_label, srt_bitmap) = mkSRTLit dflags srt
+       ; let (srt_label, srt_bitmap) = mkSRTLit dflags info_lbl srt
        ; (liveness_lit, liveness_data) <- mkLivenessBits dflags frame
        ; let
              std_info = mkStdInfoTable dflags prof_lits rts_tag srt_bitmap liveness_lit
@@ -201,7 +202,7 @@ mkInfoTableContents dflags
   | HeapRep _ ptrs nonptrs closure_type <- smrep
   = do { let layout  = packIntsCLit dflags ptrs nonptrs
        ; (prof_lits, prof_data) <- mkProfLits dflags prof
-       ; let (srt_label, srt_bitmap) = mkSRTLit dflags srt
+       ; let (srt_label, srt_bitmap) = mkSRTLit dflags info_lbl srt
        ; (mb_srt_field, mb_layout, extra_bits, ct_data)
                                 <- mk_pieces closure_type srt_label
        ; let std_info = mkStdInfoTable dflags prof_lits
@@ -211,20 +212,22 @@ mkInfoTableContents dflags
        ; return (prof_data ++ ct_data, (std_info, extra_bits)) }
   where
     mk_pieces :: ClosureTypeInfo -> [CmmLit]
-              -> UniqSM ( Maybe StgHalfWord  -- Override the SRT field with this
-                        , Maybe CmmLit       -- Override the layout field with this
+              -> UniqSM ( Maybe CmmLit  -- Override the SRT field with this
+                        , Maybe CmmLit  -- Override the layout field with this
                         , [CmmLit]           -- "Extra bits" for info table
                         , [RawCmmDecl])      -- Auxiliary data decls
     mk_pieces (Constr con_tag con_descr) _no_srt    -- A data constructor
       = do { (descr_lit, decl) <- newStringLit con_descr
-           ; return ( Just (toStgHalfWord dflags (fromIntegral con_tag))
+           ; return ( Just (CmmInt (fromIntegral con_tag)
+                                   (halfWordWidth dflags))
                     , Nothing, [descr_lit], [decl]) }
 
     mk_pieces Thunk srt_label
       = return (Nothing, Nothing, srt_label, [])
 
     mk_pieces (ThunkSelector offset) _no_srt
-      = return (Just (toStgHalfWord dflags 0), Just (mkWordCLit dflags (fromIntegral offset)), [], [])
+      = return (Just (CmmInt 0 (halfWordWidth dflags)),
+                Just (mkWordCLit dflags (fromIntegral offset)), [], [])
          -- Layout known (one free var); we use the layout field for offset
 
     mk_pieces (Fun arity (ArgSpec fun_type)) srt_label
@@ -235,8 +238,9 @@ mkInfoTableContents dflags
       = do { (liveness_lit, liveness_data) <- mkLivenessBits dflags arg_bits
            ; let fun_type | null liveness_data = aRG_GEN
                           | otherwise          = aRG_GEN_BIG
-                 extra_bits = [ packIntsCLit dflags fun_type arity
-                              , srt_lit, liveness_lit, slow_entry ]
+                 extra_bits = [ packIntsCLit dflags fun_type arity ]
+                           ++ (if inlineSRT dflags then [] else [ srt_lit ])
+                           ++ [ liveness_lit, slow_entry ]
            ; return (Nothing, Nothing, extra_bits, liveness_data) }
       where
         slow_entry = CmmLabel (toSlowEntryLbl info_lbl)
@@ -255,11 +259,21 @@ packIntsCLit dflags a b = packHalfWordsCLit dflags
 
 
 mkSRTLit :: DynFlags
+         -> CLabel
          -> Maybe CLabel
          -> ([CmmLit],    -- srt_label, if any
-             StgHalfWord) -- srt_bitmap
-mkSRTLit dflags Nothing    = ([], toStgHalfWord dflags 0)
-mkSRTLit dflags (Just lbl) = ([CmmLabel lbl], toStgHalfWord dflags 1)
+             CmmLit)      -- srt_bitmap
+mkSRTLit dflags info_lbl (Just lbl)
+  | inlineSRT dflags
+  = ([], CmmLabelDiffOff lbl info_lbl 0 (halfWordWidth dflags))
+mkSRTLit dflags _ Nothing    = ([], CmmInt 0 (halfWordWidth dflags))
+mkSRTLit dflags _ (Just lbl) = ([CmmLabel lbl], CmmInt 1 (halfWordWidth dflags))
+
+
+-- | is the SRT offset field inline in the info table on this platform?
+inlineSRT :: DynFlags -> Bool
+inlineSRT dflags = platformArch (targetPlatform dflags) == ArchX86_64
+  && tablesNextToCode dflags
 
 -------------------------------------------------------------------------
 --
@@ -367,23 +381,23 @@ mkStdInfoTable
    :: DynFlags
    -> (CmmLit,CmmLit)   -- Closure type descr and closure descr  (profiling)
    -> Int               -- Closure RTS tag
-   -> StgHalfWord       -- SRT length
+   -> CmmLit            -- SRT length
    -> CmmLit            -- layout field
    -> [CmmLit]
 
-mkStdInfoTable dflags (type_descr, closure_descr) cl_type srt_len layout_lit
+mkStdInfoTable dflags (type_descr, closure_descr) cl_type srt layout_lit
  =      -- Parallel revertible-black hole field
     prof_info
         -- Ticky info (none at present)
         -- Debug info (none at present)
- ++ [layout_lit, type_lit]
+ ++ [layout_lit, tag, srt]
 
  where
     prof_info
         | gopt Opt_SccProfilingOn dflags = [type_descr, closure_descr]
         | otherwise = []
 
-    type_lit = packHalfWordsCLit dflags (toStgHalfWord dflags (fromIntegral cl_type)) srt_len
+    tag = CmmInt (fromIntegral cl_type) (halfWordWidth dflags)
 
 -------------------------------------------------------------------------
 --