Completely new treatment of INLINE pragmas (big patch)

[ghc-hetmet.git] / compiler / simplCore / OccurAnal.lhs

%
% (c) The GRASP/AQUA Project, Glasgow University, 1992-1998
%
%************************************************************************
%*                                                                      *
\section[OccurAnal]{Occurrence analysis pass}
%*                                                                      *
%************************************************************************

The occurrence analyser re-typechecks a core expression, returning a new
core expression with (hopefully) improved usage information.

\begin{code}
module OccurAnal (
        occurAnalysePgm, occurAnalyseExpr
    ) where

#include "HsVersions.h"

import CoreSyn
import CoreFVs
import CoreUtils        ( exprIsTrivial, isDefaultAlt )
import Coercion         ( mkSymCoercion )
import Id
import BasicTypes

import VarSet
import VarEnv

import Maybes           ( orElse )
import Digraph          ( SCC(..), stronglyConnCompFromEdgedVerticesR )
import PrelNames        ( buildIdKey, foldrIdKey, runSTRepIdKey, augmentIdKey )
import Unique           ( Unique )
import UniqFM           ( keysUFM, intersectUFM_C, foldUFM_Directly )
import Util             ( mapAndUnzip )
import Outputable

import Data.List
\end{code}


%************************************************************************
%*                                                                      *
\subsection[OccurAnal-main]{Counting occurrences: main function}
%*                                                                      *
%************************************************************************

Here's the externally-callable interface:

\begin{code}
occurAnalysePgm :: [CoreBind] -> [CoreBind]
occurAnalysePgm binds
  = snd (go initOccEnv binds)
  where
    go :: OccEnv -> [CoreBind] -> (UsageDetails, [CoreBind])
    go _ []
        = (emptyDetails, [])
    go env (bind:binds)
        = (final_usage, bind' ++ binds')
        where
           (bs_usage, binds')   = go env binds
           (final_usage, bind') = occAnalBind env bind bs_usage

occurAnalyseExpr :: CoreExpr -> CoreExpr
        -- Do occurrence analysis, and discard occurence info returned
occurAnalyseExpr expr = snd (occAnal initOccEnv expr)
\end{code}


%************************************************************************
%*                                                                      *
\subsection[OccurAnal-main]{Counting occurrences: main function}
%*                                                                      *
%************************************************************************

Bindings
~~~~~~~~

\begin{code}
occAnalBind :: OccEnv
            -> CoreBind
            -> UsageDetails             -- Usage details of scope
            -> (UsageDetails,           -- Of the whole let(rec)
                [CoreBind])

occAnalBind env (NonRec binder rhs) body_usage
  | isTyVar binder                      -- A type let; we don't gather usage info
  = (body_usage, [NonRec binder rhs])

  | not (binder `usedIn` body_usage)    -- It's not mentioned
  = (body_usage, [])

  | otherwise                   -- It's mentioned in the body
  = (body_usage' +++ addRuleUsage rhs_usage binder,     -- Note [Rules are extra RHSs]
     [NonRec tagged_binder rhs'])
  where
    (body_usage', tagged_binder) = tagBinder body_usage binder
    (rhs_usage, rhs')            = occAnalRhs env tagged_binder rhs
\end{code}

Note [Dead code]
~~~~~~~~~~~~~~~~
Dropping dead code for recursive bindings is done in a very simple way:

        the entire set of bindings is dropped if none of its binders are
        mentioned in its body; otherwise none are.

This seems to miss an obvious improvement.

        letrec  f = ...g...
                g = ...f...
        in
        ...g...
===>
        letrec f = ...g...
               g = ...(...g...)...
        in
        ...g...

Now 'f' is unused! But it's OK!  Dependency analysis will sort this
out into a letrec for 'g' and a 'let' for 'f', and then 'f' will get
dropped.  It isn't easy to do a perfect job in one blow.  Consider

        letrec f = ...g...
               g = ...h...
               h = ...k...
               k = ...m...
               m = ...m...
        in
        ...m...


Note [Loop breaking and RULES]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Loop breaking is surprisingly subtle.  First read the section 4 of
"Secrets of the GHC inliner".  This describes our basic plan.

However things are made quite a bit more complicated by RULES.  Remember

  * Note [Rules are extra RHSs]
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~
    A RULE for 'f' is like an extra RHS for 'f'. That way the "parent"
    keeps the specialised "children" alive.  If the parent dies
    (because it isn't referenced any more), then the children will die
    too (unless they are already referenced directly).

    To that end, we build a Rec group for each cyclic strongly
    connected component,
        *treating f's rules as extra RHSs for 'f'*.

    When we make the Rec groups we include variables free in *either*
    LHS *or* RHS of the rule.  The former might seems silly, but see
    Note [Rule dependency info].

    So in Example [eftInt], eftInt and eftIntFB will be put in the
    same Rec, even though their 'main' RHSs are both non-recursive.

  * Note [Rules are visible in their own rec group]
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    We want the rules for 'f' to be visible in f's right-hand side.
    And we'd like them to be visible in other functions in f's Rec
    group.  E.g. in Example [Specialisation rules] we want f' rule
    to be visible in both f's RHS, and fs's RHS.

    This means that we must simplify the RULEs first, before looking
    at any of the definitions.  This is done by Simplify.simplRecBind,
    when it calls addLetIdInfo.

  * Note [Choosing loop breakers]
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    We avoid infinite inlinings by choosing loop breakers, and
    ensuring that a loop breaker cuts each loop.  But what is a
    "loop"?  In particular, a RULE is like an equation for 'f' that
    is *always* inlined if it is applicable.  We do *not* disable
    rules for loop-breakers.  It's up to whoever makes the rules to
    make sure that the rules themselves alwasys terminate.  See Note
    [Rules for recursive functions] in Simplify.lhs

    Hence, if
        f's RHS mentions g, and
        g has a RULE that mentions h, and
        h has a RULE that mentions f

    then we *must* choose f to be a loop breaker.  In general, take the
    free variables of f's RHS, and augment it with all the variables
    reachable by RULES from those starting points.  That is the whole
    reason for computing rule_fv_env in occAnalBind.  (Of course we
    only consider free vars that are also binders in this Rec group.)

    Note that when we compute this rule_fv_env, we only consider variables
    free in the *RHS* of the rule, in contrast to the way we build the
    Rec group in the first place (Note [Rule dependency info])

    Note that in Example [eftInt], *neither* eftInt *nor* eftIntFB is
    chosen as a loop breaker, because their RHSs don't mention each other.
    And indeed both can be inlined safely.

    Note that the edges of the graph we use for computing loop breakers
    are not the same as the edges we use for computing the Rec blocks.
    That's why we compute
        rec_edges          for the Rec block analysis
        loop_breaker_edges for the loop breaker analysis


  * Note [Weak loop breakers]
    ~~~~~~~~~~~~~~~~~~~~~~~~~
    There is a last nasty wrinkle.  Suppose we have

        Rec { f = f_rhs
              RULE f [] = g

              h = h_rhs
              g = h
              ...more...
        }

    Remmber that we simplify the RULES before any RHS (see Note
    [Rules are visible in their own rec group] above).

    So we must *not* postInlineUnconditionally 'g', even though
    its RHS turns out to be trivial.  (I'm assuming that 'g' is
    not choosen as a loop breaker.)

    We "solve" this by making g a "weak" or "rules-only" loop breaker,
    with OccInfo = IAmLoopBreaker True.  A normal "strong" loop breaker
    has IAmLoopBreaker False.  So

                                Inline  postInlineUnconditinoally
        IAmLoopBreaker False    no      no
        IAmLoopBreaker True     yes     no
        other                   yes     yes

    The **sole** reason for this kind of loop breaker is so that
    postInlineUnconditionally does not fire.  Ugh.

  * Note [Rule dependency info]
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~
    The VarSet in a SpecInfo is used for dependency analysis in the
    occurrence analyser.  We must track free vars in *both* lhs and rhs.  
    Hence use of idRuleVars, rather than idRuleRhsVars in addRuleUsage.  
    Why both? Consider
        x = y
        RULE f x = 4
    Then if we substitute y for x, we'd better do so in the
    rule's LHS too, so we'd better ensure the dependency is respected


Example [eftInt]
~~~~~~~~~~~~~~~
Example (from GHC.Enum):

  eftInt :: Int# -> Int# -> [Int]
  eftInt x y = ...(non-recursive)...

  {-# INLINE [0] eftIntFB #-}
  eftIntFB :: (Int -> r -> r) -> r -> Int# -> Int# -> r
  eftIntFB c n x y = ...(non-recursive)...

  {-# RULES
  "eftInt"  [~1] forall x y. eftInt x y = build (\ c n -> eftIntFB c n x y)
  "eftIntList"  [1] eftIntFB  (:) [] = eftInt
   #-}

Example [Specialisation rules]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Consider this group, which is typical of what SpecConstr builds:

   fs a = ....f (C a)....
   f  x = ....f (C a)....
   {-# RULE f (C a) = fs a #-}

So 'f' and 'fs' are in the same Rec group (since f refers to fs via its RULE).

But watch out!  If 'fs' is not chosen as a loop breaker, we may get an infinite loop:
        - the RULE is applied in f's RHS (see Note [Self-recursive rules] in Simplify
        - fs is inlined (say it's small)
        - now there's another opportunity to apply the RULE

This showed up when compiling Control.Concurrent.Chan.getChanContents.


\begin{code}
occAnalBind env (Rec pairs) body_usage
  = foldr occAnalRec (body_usage, []) sccs
        -- For a recursive group, we 
        --      * occ-analyse all the RHSs
        --      * compute strongly-connected components
        --      * feed those components to occAnalRec
  where
    -------------Dependency analysis ------------------------------
    bndr_set = mkVarSet (map fst pairs)

    sccs :: [SCC (Node Details)]
    sccs = {-# SCC "occAnalBind.scc" #-} stronglyConnCompFromEdgedVerticesR rec_edges

    rec_edges :: [Node Details]
    rec_edges = {-# SCC "occAnalBind.assoc" #-}  map make_node pairs
    
    make_node (bndr, rhs)
        = (ND bndr rhs' rhs_usage rhs_fvs, idUnique bndr, out_edges)
        where
          (rhs_usage, rhs') = occAnalRhs env bndr rhs
          rhs_fvs = intersectUFM_C (\b _ -> b) bndr_set rhs_usage
          out_edges = keysUFM (rhs_fvs `unionVarSet` idRuleVars bndr)
        -- (a -> b) means a mentions b
        -- Given the usage details (a UFM that gives occ info for each free var of
        -- the RHS) we can get the list of free vars -- or rather their Int keys --
        -- by just extracting the keys from the finite map.  Grimy, but fast.
        -- Previously we had this:
        --      [ bndr | bndr <- bndrs,
        --               maybeToBool (lookupVarEnv rhs_usage bndr)]
        -- which has n**2 cost, and this meant that edges_from alone
        -- consumed 10% of total runtime!

-----------------------------
occAnalRec :: SCC (Node Details) -> (UsageDetails, [CoreBind])
                                 -> (UsageDetails, [CoreBind])

        -- The NonRec case is just like a Let (NonRec ...) above
occAnalRec (AcyclicSCC (ND bndr rhs rhs_usage _, _, _)) (body_usage, binds)
  | not (bndr `usedIn` body_usage) 
  = (body_usage, binds)

  | otherwise                   -- It's mentioned in the body
  = (body_usage' +++ addRuleUsage rhs_usage bndr,       -- Note [Rules are extra RHSs]
     NonRec tagged_bndr rhs : binds)
  where
    (body_usage', tagged_bndr) = tagBinder body_usage bndr


        -- The Rec case is the interesting one
        -- See Note [Loop breaking]
occAnalRec (CyclicSCC nodes) (body_usage, binds)
  | not (any (`usedIn` body_usage) bndrs)       -- NB: look at body_usage, not total_usage
  = (body_usage, binds)                         -- Dead code

  | otherwise   -- At this point we always build a single Rec
  = (final_usage, Rec pairs : binds)

  where
    bndrs    = [b | (ND b _ _ _, _, _) <- nodes]
    bndr_set = mkVarSet bndrs

        ----------------------------
        -- Tag the binders with their occurrence info
    total_usage = foldl add_usage body_usage nodes
    add_usage body_usage (ND bndr _ rhs_usage _, _, _)
        = body_usage +++ addRuleUsage rhs_usage bndr
    (final_usage, tagged_nodes) = mapAccumL tag_node total_usage nodes

    tag_node :: UsageDetails -> Node Details -> (UsageDetails, Node Details)
        -- (a) Tag the binders in the details with occ info
        -- (b) Mark the binder with "weak loop-breaker" OccInfo 
        --      saying "no preInlineUnconditionally" if it is used
        --      in any rule (lhs or rhs) of the recursive group
        --      See Note [Weak loop breakers]
    tag_node usage (ND bndr rhs rhs_usage rhs_fvs, k, ks)
      = (usage `delVarEnv` bndr, (ND bndr2 rhs rhs_usage rhs_fvs, k, ks))
      where
        bndr2 | bndr `elemVarSet` all_rule_fvs = makeLoopBreaker True bndr1
              | otherwise                      = bndr1
        bndr1 = setBinderOcc usage bndr
    all_rule_fvs = bndr_set `intersectVarSet` foldr (unionVarSet . idRuleVars) 
                                                    emptyVarSet bndrs

        ----------------------------
        -- Now reconstruct the cycle
    pairs | no_rules  = reOrderCycle tagged_nodes
          | otherwise = concatMap reOrderRec (stronglyConnCompFromEdgedVerticesR loop_breaker_edges)

        -- See Note [Choosing loop breakers] for looop_breaker_edges
    loop_breaker_edges = map mk_node tagged_nodes
    mk_node (details@(ND _ _ _ rhs_fvs), k, _) = (details, k, new_ks)
        where
          new_ks = keysUFM (extendFvs rule_fv_env rhs_fvs rhs_fvs)

    ------------------------------------
    rule_fv_env :: IdEnv IdSet  -- Variables from this group mentioned in RHS of rules
                                -- Domain is *subset* of bound vars (others have no rule fvs)
    rule_fv_env = rule_loop init_rule_fvs

    no_rules      = null init_rule_fvs
    init_rule_fvs = [(b, rule_fvs)
                    | b <- bndrs
                    , let rule_fvs = idRuleRhsVars b `intersectVarSet` bndr_set
                    , not (isEmptyVarSet rule_fvs)]

    rule_loop :: [(Id,IdSet)] -> IdEnv IdSet    -- Finds fixpoint
    rule_loop fv_list
        | no_change = env
        | otherwise = rule_loop new_fv_list
        where
          env = mkVarEnv init_rule_fvs
          (no_change, new_fv_list) = mapAccumL bump True fv_list
          bump no_change (b,fvs)
                | new_fvs `subVarSet` fvs = (no_change, (b,fvs))
                | otherwise               = (False,     (b,new_fvs `unionVarSet` fvs))
                where
                  new_fvs = extendFvs env emptyVarSet fvs

extendFvs :: IdEnv IdSet -> IdSet -> IdSet -> IdSet
-- (extendFVs env fvs s) returns (fvs `union` env(s))
extendFvs env fvs id_set
  = foldUFM_Directly add fvs id_set
  where
    add uniq _ fvs
        = case lookupVarEnv_Directly env uniq  of
            Just fvs' -> fvs' `unionVarSet` fvs
            Nothing   -> fvs
\end{code}

@reOrderRec@ is applied to the list of (binder,rhs) pairs for a cyclic
strongly connected component (there's guaranteed to be a cycle).  It returns the
same pairs, but
        a) in a better order,
        b) with some of the Ids having a IAmALoopBreaker pragma

The "loop-breaker" Ids are sufficient to break all cycles in the SCC.  This means
that the simplifier can guarantee not to loop provided it never records an inlining
for these no-inline guys.

Furthermore, the order of the binds is such that if we neglect dependencies
on the no-inline Ids then the binds are topologically sorted.  This means
that the simplifier will generally do a good job if it works from top bottom,
recording inlinings for any Ids which aren't marked as "no-inline" as it goes.

==============
[June 98: I don't understand the following paragraphs, and I've
          changed the a=b case again so that it isn't a special case any more.]

Here's a case that bit me:

        letrec
                a = b
                b = \x. BIG
        in
        ...a...a...a....

Re-ordering doesn't change the order of bindings, but there was no loop-breaker.

My solution was to make a=b bindings record b as Many, rather like INLINE bindings.
Perhaps something cleverer would suffice.
===============


\begin{code}
type Node details = (details, Unique, [Unique]) -- The Ints are gotten from the Unique,
                                                -- which is gotten from the Id.
data Details = ND Id            -- Binder
                  CoreExpr      -- RHS
                  UsageDetails  -- Full usage from RHS (*not* including rules)
                  IdSet         -- Other binders from this Rec group mentioned on RHS
                                -- (derivable from UsageDetails but cached here)

reOrderRec :: SCC (Node Details)
           -> [(Id,CoreExpr)]
-- Sorted into a plausible order.  Enough of the Ids have
--      IAmALoopBreaker pragmas that there are no loops left.
reOrderRec (AcyclicSCC (ND bndr rhs _ _, _, _)) = [(bndr, rhs)]
reOrderRec (CyclicSCC cycle)                    = reOrderCycle cycle

reOrderCycle :: [Node Details] -> [(Id,CoreExpr)]
reOrderCycle []
  = panic "reOrderCycle"
reOrderCycle [bind]     -- Common case of simple self-recursion
  = [(makeLoopBreaker False bndr, rhs)]
  where
    (ND bndr rhs _ _, _, _) = bind

reOrderCycle (bind : binds)
  =     -- Choose a loop breaker, mark it no-inline,
        -- do SCC analysis on the rest, and recursively sort them out
    concatMap reOrderRec (stronglyConnCompFromEdgedVerticesR unchosen) ++
    [(makeLoopBreaker False bndr, rhs)]

  where
    (chosen_bind, unchosen) = choose_loop_breaker bind (score bind) [] binds
    ND bndr rhs _ _ = chosen_bind

        -- This loop looks for the bind with the lowest score
        -- to pick as the loop  breaker.  The rest accumulate in
    choose_loop_breaker (details,_,_) _loop_sc acc []
        = (details, acc)        -- Done

    choose_loop_breaker loop_bind loop_sc acc (bind : binds)
        | sc < loop_sc  -- Lower score so pick this new one
        = choose_loop_breaker bind sc (loop_bind : acc) binds

        | otherwise     -- No lower so don't pick it
        = choose_loop_breaker loop_bind loop_sc (bind : acc) binds
        where
          sc = score bind

    score :: Node Details -> Int        -- Higher score => less likely to be picked as loop breaker
    score (ND bndr rhs _ _, _, _)
        | isInlineRule (idUnfolding bndr) = 10
                -- Note [INLINE pragmas]

        | exprIsTrivial rhs        = 5  -- Practically certain to be inlined
                -- Used to have also: && not (isExportedId bndr)
                -- But I found this sometimes cost an extra iteration when we have
                --      rec { d = (a,b); a = ...df...; b = ...df...; df = d }
                -- where df is the exported dictionary. Then df makes a really
                -- bad choice for loop breaker

        | is_con_app rhs = 3    -- Data types help with cases
                -- Note [Constructor applictions]

-- If an Id is marked "never inline" then it makes a great loop breaker
-- The only reason for not checking that here is that it is rare
-- and I've never seen a situation where it makes a difference,
-- so it probably isn't worth the time to test on every binder
--      | isNeverActive (idInlinePragma bndr) = -10

        | isOneOcc (idOccInfo bndr) = 1  -- Likely to be inlined

        | canUnfold (idUnfolding bndr) = 1
                -- the Id has some kind of unfolding

        | otherwise = 0

        -- Checking for a constructor application
        -- Cheap and cheerful; the simplifer moves casts out of the way
        -- The lambda case is important to spot x = /\a. C (f a)
        -- which comes up when C is a dictionary constructor and
        -- f is a default method.
        -- Example: the instance for Show (ST s a) in GHC.ST
        --
        -- However we *also* treat (\x. C p q) as a con-app-like thing,
        --      Note [Closure conversion]
    is_con_app (Var v)    = isDataConWorkId v
    is_con_app (App f _)  = is_con_app f
    is_con_app (Lam _ e)  = is_con_app e
    is_con_app (Note _ e) = is_con_app e
    is_con_app _          = False

makeLoopBreaker :: Bool -> Id -> Id
-- Set the loop-breaker flag: see Note [Weak loop breakers]
makeLoopBreaker weak bndr = setIdOccInfo bndr (IAmALoopBreaker weak)
\end{code}

Note [INLINE pragmas]
~~~~~~~~~~~~~~~~~~~~~
Never choose a function with an INLINE pramga as the loop breaker!  
If such a function is mutually-recursive with a non-INLINE thing,
then the latter should be the loop-breaker.

A particular case is wrappers generated by the demand analyser.
If you make then into a loop breaker you may get an infinite 
inlining loop.  For example:
  rec {
        $wfoo x = ....foo x....

        {-loop brk-} foo x = ...$wfoo x...
  }
The interface file sees the unfolding for $wfoo, and sees that foo is
strict (and hence it gets an auto-generated wrapper).  Result: an
infinite inlining in the importing scope.  So be a bit careful if you
change this.  A good example is Tree.repTree in
nofib/spectral/minimax. If the repTree wrapper is chosen as the loop
breaker then compiling Game.hs goes into an infinite loop (this
happened when we gave is_con_app a lower score than inline candidates).

Note [Constructor applications]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
It's really really important to inline dictionaries.  Real
example (the Enum Ordering instance from GHC.Base):

     rec     f = \ x -> case d of (p,q,r) -> p x
             g = \ x -> case d of (p,q,r) -> q x
             d = (v, f, g)

Here, f and g occur just once; but we can't inline them into d.
On the other hand we *could* simplify those case expressions if
we didn't stupidly choose d as the loop breaker.
But we won't because constructor args are marked "Many".
Inlining dictionaries is really essential to unravelling
the loops in static numeric dictionaries, see GHC.Float.

Note [Closure conversion]
~~~~~~~~~~~~~~~~~~~~~~~~~
We treat (\x. C p q) as a high-score candidate in the letrec scoring algorithm.
The immediate motivation came from the result of a closure-conversion transformation
which generated code like this:

    data Clo a b = forall c. Clo (c -> a -> b) c

    ($:) :: Clo a b -> a -> b
    Clo f env $: x = f env x

    rec { plus = Clo plus1 ()

        ; plus1 _ n = Clo plus2 n

        ; plus2 Zero     n = n
        ; plus2 (Succ m) n = Succ (plus $: m $: n) }

If we inline 'plus' and 'plus1', everything unravels nicely.  But if
we choose 'plus1' as the loop breaker (which is entirely possible
otherwise), the loop does not unravel nicely.


@occAnalRhs@ deals with the question of bindings where the Id is marked
by an INLINE pragma.  For these we record that anything which occurs
in its RHS occurs many times.  This pessimistically assumes that ths
inlined binder also occurs many times in its scope, but if it doesn't
we'll catch it next time round.  At worst this costs an extra simplifier pass.
ToDo: try using the occurrence info for the inline'd binder.

[March 97] We do the same for atomic RHSs.  Reason: see notes with reOrderRec.
[June 98, SLPJ]  I've undone this change; I don't understand it.  See notes with reOrderRec.


\begin{code}
occAnalRhs :: OccEnv
           -> Id -> CoreExpr    -- Binder and rhs
                                -- For non-recs the binder is alrady tagged
                                -- with occurrence info
           -> (UsageDetails, CoreExpr)

occAnalRhs env id rhs
  = occAnal ctxt rhs
  where
    ctxt | certainly_inline id = env
         | otherwise           = rhsCtxt env
        -- Note that we generally use an rhsCtxt.  This tells the occ anal n
        -- that it's looking at an RHS, which has an effect in occAnalApp
        --
        -- But there's a problem.  Consider
        --      x1 = a0 : []
        --      x2 = a1 : x1
        --      x3 = a2 : x2
        --      g  = f x3
        -- First time round, it looks as if x1 and x2 occur as an arg of a
        -- let-bound constructor ==> give them a many-occurrence.
        -- But then x3 is inlined (unconditionally as it happens) and
        -- next time round, x2 will be, and the next time round x1 will be
        -- Result: multiple simplifier iterations.  Sigh.
        -- Crude solution: use rhsCtxt for things that occur just once...

    certainly_inline id = case idOccInfo id of
                            OneOcc in_lam one_br _ -> not in_lam && one_br
                            _                      -> False
\end{code}



\begin{code}
addRuleUsage :: UsageDetails -> Id -> UsageDetails
-- Add the usage from RULES in Id to the usage
addRuleUsage usage id
  = foldVarSet add usage (idRuleVars id)
        -- idRuleVars here: see Note [Rule dependency info]
  where
    add v u = addOneOcc u v NoOccInfo
        -- Give a non-committal binder info (i.e manyOcc) because
        --   a) Many copies of the specialised thing can appear
        --   b) We don't want to substitute a BIG expression inside a RULE
        --      even if that's the only occurrence of the thing
        --      (Same goes for INLINE.)
\end{code}

Expressions
~~~~~~~~~~~
\begin{code}
occAnal :: OccEnv
        -> CoreExpr
        -> (UsageDetails,       -- Gives info only about the "interesting" Ids
            CoreExpr)

occAnal _   (Type t)  = (emptyDetails, Type t)
occAnal env (Var v)   = (mkOneOcc env v False, Var v)
    -- At one stage, I gathered the idRuleVars for v here too,
    -- which in a way is the right thing to do.
    -- But that went wrong right after specialisation, when
    -- the *occurrences* of the overloaded function didn't have any
    -- rules in them, so the *specialised* versions looked as if they
    -- weren't used at all.
\end{code}

We regard variables that occur as constructor arguments as "dangerousToDup":

\begin{verbatim}
module A where
f x = let y = expensive x in
      let z = (True,y) in
      (case z of {(p,q)->q}, case z of {(p,q)->q})
\end{verbatim}

We feel free to duplicate the WHNF (True,y), but that means
that y may be duplicated thereby.

If we aren't careful we duplicate the (expensive x) call!
Constructors are rather like lambdas in this way.

\begin{code}
occAnal _   expr@(Lit _) = (emptyDetails, expr)
\end{code}

\begin{code}
occAnal env (Note note@(SCC _) body)
  = case occAnal env body of { (usage, body') ->
    (mapVarEnv markInsideSCC usage, Note note body')
    }

occAnal env (Note note body)
  = case occAnal env body of { (usage, body') ->
    (usage, Note note body')
    }

occAnal env (Cast expr co)
  = case occAnal env expr of { (usage, expr') ->
    (markRhsUds env True usage, Cast expr' co)
        -- If we see let x = y `cast` co
        -- then mark y as 'Many' so that we don't
        -- immediately inline y again.
    }
\end{code}

\begin{code}
occAnal env app@(App _ _)
  = occAnalApp env (collectArgs app)

-- Ignore type variables altogether
--   (a) occurrences inside type lambdas only not marked as InsideLam
--   (b) type variables not in environment

occAnal env (Lam x body) | isTyVar x
  = case occAnal env body of { (body_usage, body') ->
    (body_usage, Lam x body')
    }

-- For value lambdas we do a special hack.  Consider
--      (\x. \y. ...x...)
-- If we did nothing, x is used inside the \y, so would be marked
-- as dangerous to dup.  But in the common case where the abstraction
-- is applied to two arguments this is over-pessimistic.
-- So instead, we just mark each binder with its occurrence
-- info in the *body* of the multiple lambda.
-- Then, the simplifier is careful when partially applying lambdas.

occAnal env expr@(Lam _ _)
  = case occAnal env_body body of { (body_usage, body') ->
    let
        (final_usage, tagged_binders) = tagBinders body_usage binders
        --      URGH!  Sept 99: we don't seem to be able to use binders' here, because
        --      we get linear-typed things in the resulting program that we can't handle yet.
        --      (e.g. PrelShow)  TODO

        really_final_usage = if linear then
                                final_usage
                             else
                                mapVarEnv markInsideLam final_usage
    in
    (really_final_usage,
     mkLams tagged_binders body') }
  where
    env_body        = vanillaCtxt env        -- Body is (no longer) an RhsContext
    (binders, body) = collectBinders expr
    binders'        = oneShotGroup env binders
    linear          = all is_one_shot binders'
    is_one_shot b   = isId b && isOneShotBndr b

occAnal env (Case scrut bndr ty alts)
  = case occ_anal_scrut scrut alts     of { (scrut_usage, scrut') ->
    case mapAndUnzip occ_anal_alt alts of { (alts_usage_s, alts')   ->
    let
        alts_usage  = foldr1 combineAltsUsageDetails alts_usage_s
        alts_usage' = addCaseBndrUsage alts_usage
        (alts_usage1, tagged_bndr) = tagBinder alts_usage' bndr
        total_usage = scrut_usage +++ alts_usage1
    in
    total_usage `seq` (total_usage, Case scrut' tagged_bndr ty alts') }}
  where
        -- Note [Case binder usage]     
        -- ~~~~~~~~~~~~~~~~~~~~~~~~
        -- The case binder gets a usage of either "many" or "dead", never "one".
        -- Reason: we like to inline single occurrences, to eliminate a binding,
        -- but inlining a case binder *doesn't* eliminate a binding.
        -- We *don't* want to transform
        --      case x of w { (p,q) -> f w }
        -- into
        --      case x of w { (p,q) -> f (p,q) }
    addCaseBndrUsage usage = case lookupVarEnv usage bndr of
                                Nothing -> usage
                                Just _  -> extendVarEnv usage bndr NoOccInfo

    alt_env = mkAltEnv env bndr_swap
        -- Consider     x = case v of { True -> (p,q); ... }
        -- Then it's fine to inline p and q

    bndr_swap = case scrut of
                  Var v           -> Just (v, Var bndr)
                  Cast (Var v) co -> Just (v, Cast (Var bndr) (mkSymCoercion co))
                  _other          -> Nothing

    occ_anal_alt = occAnalAlt alt_env bndr bndr_swap

    occ_anal_scrut (Var v) (alt1 : other_alts)
        | not (null other_alts) || not (isDefaultAlt alt1)
        = (mkOneOcc env v True, Var v)  -- The 'True' says that the variable occurs
                                        -- in an interesting context; the case has
                                        -- at least one non-default alternative
    occ_anal_scrut scrut _alts  
        = occAnal (vanillaCtxt env) scrut    -- No need for rhsCtxt

occAnal env (Let bind body)
  = case occAnal env body                of { (body_usage, body') ->
    case occAnalBind env bind body_usage of { (final_usage, new_binds) ->
       (final_usage, mkLets new_binds body') }}

occAnalArgs :: OccEnv -> [CoreExpr] -> (UsageDetails, [CoreExpr])
occAnalArgs env args
  = case mapAndUnzip (occAnal arg_env) args of  { (arg_uds_s, args') ->
    (foldr (+++) emptyDetails arg_uds_s, args')}
  where
    arg_env = vanillaCtxt env
\end{code}

Applications are dealt with specially because we want
the "build hack" to work.

\begin{code}
occAnalApp :: OccEnv
           -> (Expr CoreBndr, [Arg CoreBndr])
           -> (UsageDetails, Expr CoreBndr)
occAnalApp env (Var fun, args)
  = case args_stuff of { (args_uds, args') ->
    let
        final_args_uds = markRhsUds env is_pap args_uds
    in
    (fun_uds +++ final_args_uds, mkApps (Var fun) args') }
  where
    fun_uniq = idUnique fun
    fun_uds  = mkOneOcc env fun (valArgCount args > 0)
    is_pap = isDataConWorkId fun || valArgCount args < idArity fun

                -- Hack for build, fold, runST
    args_stuff  | fun_uniq == buildIdKey    = appSpecial env 2 [True,True]  args
                | fun_uniq == augmentIdKey  = appSpecial env 2 [True,True]  args
                | fun_uniq == foldrIdKey    = appSpecial env 3 [False,True] args
                | fun_uniq == runSTRepIdKey = appSpecial env 2 [True]       args
                        -- (foldr k z xs) may call k many times, but it never
                        -- shares a partial application of k; hence [False,True]
                        -- This means we can optimise
                        --      foldr (\x -> let v = ...x... in \y -> ...v...) z xs
                        -- by floating in the v

                | otherwise = occAnalArgs env args


occAnalApp env (fun, args)
  = case occAnal (addAppCtxt env args) fun of   { (fun_uds, fun') ->
        -- The addAppCtxt is a bit cunning.  One iteration of the simplifier
        -- often leaves behind beta redexs like
        --      (\x y -> e) a1 a2
        -- Here we would like to mark x,y as one-shot, and treat the whole
        -- thing much like a let.  We do this by pushing some True items
        -- onto the context stack.

    case occAnalArgs env args of        { (args_uds, args') ->
    let
        final_uds = fun_uds +++ args_uds
    in
    (final_uds, mkApps fun' args') }}


markRhsUds :: OccEnv            -- Check if this is a RhsEnv
           -> Bool              -- and this is true
           -> UsageDetails      -- The do markMany on this
           -> UsageDetails
-- We mark the free vars of the argument of a constructor or PAP
-- as "many", if it is the RHS of a let(rec).
-- This means that nothing gets inlined into a constructor argument
-- position, which is what we want.  Typically those constructor
-- arguments are just variables, or trivial expressions.
--
-- This is the *whole point* of the isRhsEnv predicate
markRhsUds env is_pap arg_uds
  | isRhsEnv env && is_pap = mapVarEnv markMany arg_uds
  | otherwise              = arg_uds


appSpecial :: OccEnv
           -> Int -> CtxtTy     -- Argument number, and context to use for it
           -> [CoreExpr]
           -> (UsageDetails, [CoreExpr])
appSpecial env n ctxt args
  = go n args
  where
    arg_env = vanillaCtxt env

    go _ [] = (emptyDetails, [])        -- Too few args

    go 1 (arg:args)                     -- The magic arg
      = case occAnal (setCtxtTy arg_env ctxt) arg of    { (arg_uds, arg') ->
        case occAnalArgs env args of                    { (args_uds, args') ->
        (arg_uds +++ args_uds, arg':args') }}

    go n (arg:args)
      = case occAnal arg_env arg of     { (arg_uds, arg') ->
        case go (n-1) args of           { (args_uds, args') ->
        (arg_uds +++ args_uds, arg':args') }}
\end{code}


Note [Binder swap]
~~~~~~~~~~~~~~~~~~
We do these two transformations right here:

 (1)   case x of b { pi -> ri }
    ==>
      case x of b { pi -> let x=b in ri }

 (2)  case (x |> co) of b { pi -> ri }
    ==>
      case (x |> co) of b { pi -> let x = b |> sym co in ri }

    Why (2)?  See Note [Case of cast]

In both cases, in a particular alternative (pi -> ri), we only 
add the binding if
  (a) x occurs free in (pi -> ri)
        (ie it occurs in ri, but is not bound in pi)
  (b) the pi does not bind b (or the free vars of co)
We need (a) and (b) for the inserted binding to be correct.

For the alternatives where we inject the binding, we can transfer
all x's OccInfo to b.  And that is the point.

Notice that 
  * The deliberate shadowing of 'x'. 
  * That (a) rapidly becomes false, so no bindings are injected.

The reason for doing these transformations here is because it allows
us to adjust the OccInfo for 'x' and 'b' as we go.

  * Suppose the only occurrences of 'x' are the scrutinee and in the
    ri; then this transformation makes it occur just once, and hence
    get inlined right away.

  * If we do this in the Simplifier, we don't know whether 'x' is used
    in ri, so we are forced to pessimistically zap b's OccInfo even
    though it is typically dead (ie neither it nor x appear in the
    ri).  There's nothing actually wrong with zapping it, except that
    it's kind of nice to know which variables are dead.  My nose
    tells me to keep this information as robustly as possible.

The Maybe (Id,CoreExpr) passed to occAnalAlt is the extra let-binding
{x=b}; it's Nothing if the binder-swap doesn't happen.

Note [Binder swap on GlobalId scrutinees]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
When the scrutinee is a GlobalId we must take care in two ways

 i) In order to *know* whether 'x' occurs free in the RHS, we need its
    occurrence info. BUT, we don't gather occurrence info for
    GlobalIds.  That's what the (small) occ_scrut_ids set in OccEnv is
    for: it says "gather occurrence info for these.

 ii) We must call localiseId on 'x' first, in case it's a GlobalId, or
     has an External Name. See, for example, SimplEnv Note [Global Ids in
     the substitution].

Note [Case of cast]
~~~~~~~~~~~~~~~~~~~
Consider        case (x `cast` co) of b { I# ->
                ... (case (x `cast` co) of {...}) ...
We'd like to eliminate the inner case.  That is the motivation for
equation (2) in Note [Binder swap].  When we get to the inner case, we
inline x, cancel the casts, and away we go.

Note [Binders in case alternatives]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Consider
    case x of y { (a,b) -> f y }
We treat 'a', 'b' as dead, because they don't physically occur in the
case alternative.  (Indeed, a variable is dead iff it doesn't occur in
its scope in the output of OccAnal.)  This invariant is It really
helpe to know when binders are unused.  See esp the call to
isDeadBinder in Simplify.mkDupableAlt

In this example, though, the Simplifier will bring 'a' and 'b' back to
life, beause it binds 'y' to (a,b) (imagine got inlined and
scrutinised y).

\begin{code}
occAnalAlt :: OccEnv
           -> CoreBndr
           -> Maybe (Id, CoreExpr)  -- Note [Binder swap]
           -> CoreAlt
           -> (UsageDetails, Alt IdWithOccInfo)
occAnalAlt env case_bndr mb_scrut_var (con, bndrs, rhs)
  = case occAnal env rhs of { (rhs_usage, rhs') ->
    let
        (alt_usg, tagged_bndrs) = tagBinders rhs_usage bndrs
        bndrs' = tagged_bndrs      -- See Note [Binders in case alternatives]
    in
    case mb_scrut_var of
        Just (scrut_var, scrut_rhs)             -- See Note [Binder swap]
          | scrut_var `localUsedIn` alt_usg     -- (a) Fast path, usually false
          , not (any shadowing bndrs)           -- (b) 
          -> (addOneOcc usg_wo_scrut case_bndr NoOccInfo,
                        -- See Note [Case binder usage] for the NoOccInfo
              (con, bndrs', Let (NonRec scrut_var' scrut_rhs) rhs'))
          where
           (usg_wo_scrut, scrut_var') = tagBinder alt_usg (localiseId scrut_var)
                        -- Note the localiseId; we're making a new binding
                        -- for it, and it might have an External Name, or
                        -- even be a GlobalId; Note [Binder swap on GlobalId scrutinees]
           shadowing bndr = bndr `elemVarSet` rhs_fvs
           rhs_fvs = exprFreeVars scrut_rhs

        _other -> (alt_usg, (con, bndrs', rhs')) }
\end{code}


%************************************************************************
%*                                                                      *
\subsection[OccurAnal-types]{OccEnv}
%*                                                                      *
%************************************************************************

\begin{code}
data OccEnv
  = OccEnv { occ_encl      :: !OccEncl      -- Enclosing context information
           , occ_ctxt      :: !CtxtTy       -- Tells about linearity
           , occ_scrut_ids :: !GblScrutIds }

type GblScrutIds = IdSet  -- GlobalIds that are scrutinised, and for which
                          -- we want to gather occurence info; see
                          -- Note [Binder swap for GlobalId scrutinee]
                          -- No need to prune this if there's a shadowing binding
                          -- because it's OK for it to be too big

-- OccEncl is used to control whether to inline into constructor arguments
-- For example:
--      x = (p,q)               -- Don't inline p or q
--      y = /\a -> (p a, q a)   -- Still don't inline p or q
--      z = f (p,q)             -- Do inline p,q; it may make a rule fire
-- So OccEncl tells enought about the context to know what to do when
-- we encounter a contructor application or PAP.

data OccEncl
  = OccRhs              -- RHS of let(rec), albeit perhaps inside a type lambda
                        -- Don't inline into constructor args here
  | OccVanilla          -- Argument of function, body of lambda, scruintee of case etc.
                        -- Do inline into constructor args here

type CtxtTy = [Bool]
        -- []           No info
        --
        -- True:ctxt    Analysing a function-valued expression that will be
        --                      applied just once
        --
        -- False:ctxt   Analysing a function-valued expression that may
        --                      be applied many times; but when it is,
        --                      the CtxtTy inside applies

initOccEnv :: OccEnv
initOccEnv = OccEnv { occ_encl = OccRhs
                    , occ_ctxt = []
                    , occ_scrut_ids = emptyVarSet }

vanillaCtxt :: OccEnv -> OccEnv
vanillaCtxt env = OccEnv { occ_encl = OccVanilla, occ_ctxt = []
                         , occ_scrut_ids = occ_scrut_ids env }

rhsCtxt :: OccEnv -> OccEnv
rhsCtxt env = OccEnv { occ_encl = OccRhs, occ_ctxt = []
                     , occ_scrut_ids = occ_scrut_ids env }

mkAltEnv :: OccEnv -> Maybe (Id, CoreExpr) -> OccEnv
-- Does two things: a) makes the occ_ctxt = OccVanilla
--                  b) extends the scrut_ids if necessary
mkAltEnv env (Just (scrut_id, _))
  | not (isLocalId scrut_id) 
  = OccEnv { occ_encl      = OccVanilla
           , occ_scrut_ids = extendVarSet (occ_scrut_ids env) scrut_id
           , occ_ctxt      = occ_ctxt env }
mkAltEnv env _
  | isRhsEnv env = env { occ_encl = OccVanilla }
  | otherwise    = env

setCtxtTy :: OccEnv -> CtxtTy -> OccEnv
setCtxtTy env ctxt = env { occ_ctxt = ctxt }

isRhsEnv :: OccEnv -> Bool
isRhsEnv (OccEnv { occ_encl = OccRhs })     = True
isRhsEnv (OccEnv { occ_encl = OccVanilla }) = False

oneShotGroup :: OccEnv -> [CoreBndr] -> [CoreBndr]
        -- The result binders have one-shot-ness set that they might not have had originally.
        -- This happens in (build (\cn -> e)).  Here the occurrence analyser
        -- linearity context knows that c,n are one-shot, and it records that fact in
        -- the binder. This is useful to guide subsequent float-in/float-out tranformations

oneShotGroup (OccEnv { occ_ctxt = ctxt }) bndrs
  = go ctxt bndrs []
  where
    go _ [] rev_bndrs = reverse rev_bndrs

    go (lin_ctxt:ctxt) (bndr:bndrs) rev_bndrs
        | isId bndr = go ctxt bndrs (bndr':rev_bndrs)
        where
          bndr' | lin_ctxt  = setOneShotLambda bndr
                | otherwise = bndr

    go ctxt (bndr:bndrs) rev_bndrs = go ctxt bndrs (bndr:rev_bndrs)

addAppCtxt :: OccEnv -> [Arg CoreBndr] -> OccEnv
addAppCtxt env@(OccEnv { occ_ctxt = ctxt }) args
  = env { occ_ctxt = replicate (valArgCount args) True ++ ctxt }
\end{code}

%************************************************************************
%*                                                                      *
\subsection[OccurAnal-types]{OccEnv}
%*                                                                      *
%************************************************************************

\begin{code}
type UsageDetails = IdEnv OccInfo       -- A finite map from ids to their usage
                -- INVARIANT: never IAmDead
                -- (Deadness is signalled by not being in the map at all)

(+++), combineAltsUsageDetails
        :: UsageDetails -> UsageDetails -> UsageDetails

(+++) usage1 usage2
  = plusVarEnv_C addOccInfo usage1 usage2

combineAltsUsageDetails usage1 usage2
  = plusVarEnv_C orOccInfo usage1 usage2

addOneOcc :: UsageDetails -> Id -> OccInfo -> UsageDetails
addOneOcc usage id info
  = plusVarEnv_C addOccInfo usage (unitVarEnv id info)
        -- ToDo: make this more efficient

emptyDetails :: UsageDetails
emptyDetails = (emptyVarEnv :: UsageDetails)

localUsedIn, usedIn :: Id -> UsageDetails -> Bool
v `localUsedIn` details = v `elemVarEnv` details
v `usedIn`      details =  isExportedId v || v `localUsedIn` details

type IdWithOccInfo = Id

tagBinders :: UsageDetails          -- Of scope
           -> [Id]                  -- Binders
           -> (UsageDetails,        -- Details with binders removed
              [IdWithOccInfo])    -- Tagged binders

tagBinders usage binders
 = let
     usage' = usage `delVarEnvList` binders
     uss    = map (setBinderOcc usage) binders
   in
   usage' `seq` (usage', uss)

tagBinder :: UsageDetails           -- Of scope
          -> Id                     -- Binders
          -> (UsageDetails,         -- Details with binders removed
              IdWithOccInfo)        -- Tagged binders

tagBinder usage binder
 = let
     usage'  = usage `delVarEnv` binder
     binder' = setBinderOcc usage binder
   in
   usage' `seq` (usage', binder')

setBinderOcc :: UsageDetails -> CoreBndr -> CoreBndr
setBinderOcc usage bndr
  | isTyVar bndr      = bndr
  | isExportedId bndr = case idOccInfo bndr of
                          NoOccInfo -> bndr
                          _         -> setIdOccInfo bndr NoOccInfo
            -- Don't use local usage info for visible-elsewhere things
            -- BUT *do* erase any IAmALoopBreaker annotation, because we're
            -- about to re-generate it and it shouldn't be "sticky"

  | otherwise = setIdOccInfo bndr occ_info
  where
    occ_info = lookupVarEnv usage bndr `orElse` IAmDead
\end{code}


%************************************************************************
%*                                                                      *
\subsection{Operations over OccInfo}
%*                                                                      *
%************************************************************************

\begin{code}
mkOneOcc :: OccEnv -> Id -> InterestingCxt -> UsageDetails
mkOneOcc env id int_cxt
  | isLocalId id = unitVarEnv id (OneOcc False True int_cxt)
  | id `elemVarSet` occ_scrut_ids env = unitVarEnv id NoOccInfo
  | otherwise                         = emptyDetails

markMany, markInsideLam, markInsideSCC :: OccInfo -> OccInfo

markMany _  = NoOccInfo

markInsideSCC occ = markMany occ

markInsideLam (OneOcc _ one_br int_cxt) = OneOcc True one_br int_cxt
markInsideLam occ                       = occ

addOccInfo, orOccInfo :: OccInfo -> OccInfo -> OccInfo

addOccInfo a1 a2  = ASSERT( not (isDeadOcc a1 || isDeadOcc a2) )
                    NoOccInfo   -- Both branches are at least One
                                -- (Argument is never IAmDead)

-- (orOccInfo orig new) is used
-- when combining occurrence info from branches of a case

orOccInfo (OneOcc in_lam1 _ int_cxt1)
          (OneOcc in_lam2 _ int_cxt2)
  = OneOcc (in_lam1 || in_lam2)
           False        -- False, because it occurs in both branches
           (int_cxt1 && int_cxt2)
orOccInfo a1 a2 = ASSERT( not (isDeadOcc a1 || isDeadOcc a2) )
                  NoOccInfo
\end{code}