2 % (c) The GRASP/AQUA Project, Glasgow University, 1992-1998
4 \section[SpecConstr]{Specialise over constructors}
11 #include "HsVersions.h"
14 import CoreLint ( showPass, endPass )
15 import CoreUtils ( exprType, mkPiTypes )
16 import CoreFVs ( exprsFreeVars )
17 import CoreTidy ( tidyRules )
18 import PprCore ( pprRules )
19 import WwLib ( mkWorkerArgs )
20 import DataCon ( dataConRepArity, dataConUnivTyVars )
21 import Type ( Type, tyConAppArgs )
22 import Coercion ( coercionKind )
23 import Rules ( matchN )
24 import Id ( Id, idName, idType, isDataConWorkId_maybe,
25 mkUserLocal, mkSysLocal, idUnfolding, isLocalId )
29 import Name ( nameOccName, nameSrcLoc )
30 import Rules ( addIdSpecialisations, mkLocalRule, rulesOfBinds )
31 import OccName ( mkSpecOcc )
32 import ErrUtils ( dumpIfSet_dyn )
33 import DynFlags ( DynFlags, DynFlag(..) )
34 import BasicTypes ( Activation(..) )
35 import Maybes ( orElse, catMaybes, isJust )
36 import Util ( zipWithEqual, lengthAtLeast, notNull )
37 import List ( nubBy, partition )
44 -----------------------------------------------------
46 -----------------------------------------------------
51 drop n (x:xs) = drop (n-1) xs
53 After the first time round, we could pass n unboxed. This happens in
54 numerical code too. Here's what it looks like in Core:
56 drop n xs = case xs of
61 _ -> drop (I# (n# -# 1#)) xs
63 Notice that the recursive call has an explicit constructor as argument.
64 Noticing this, we can make a specialised version of drop
66 RULE: drop (I# n#) xs ==> drop' n# xs
68 drop' n# xs = let n = I# n# in ...orig RHS...
70 Now the simplifier will apply the specialisation in the rhs of drop', giving
72 drop' n# xs = case xs of
76 _ -> drop (n# -# 1#) xs
80 We'd also like to catch cases where a parameter is carried along unchanged,
81 but evaluated each time round the loop:
83 f i n = if i>0 || i>n then i else f (i*2) n
85 Here f isn't strict in n, but we'd like to avoid evaluating it each iteration.
86 In Core, by the time we've w/wd (f is strict in i) we get
88 f i# n = case i# ># 0 of
90 True -> case n of n' { I# n# ->
93 True -> f (i# *# 2#) n'
95 At the call to f, we see that the argument, n is know to be (I# n#),
96 and n is evaluated elsewhere in the body of f, so we can play the same
102 We must be careful not to allocate the same constructor twice. Consider
103 f p = (...(case p of (a,b) -> e)...p...,
104 ...let t = (r,s) in ...t...(f t)...)
105 At the recursive call to f, we can see that t is a pair. But we do NOT want
106 to make a specialised copy:
107 f' a b = let p = (a,b) in (..., ...)
108 because now t is allocated by the caller, then r and s are passed to the
109 recursive call, which allocates the (r,s) pair again.
112 (a) the argument p is used in other than a case-scrutinsation way.
113 (b) the argument to the call is not a 'fresh' tuple; you have to
114 look into its unfolding to see that it's a tuple
116 Hence the "OR" part of Note [Good arguments] below.
118 ALTERNATIVE 2: pass both boxed and unboxed versions. This no longer saves
119 allocation, but does perhaps save evals. In the RULE we'd have
122 f (I# x#) = f' (I# x#) x#
124 If at the call site the (I# x) was an unfolding, then we'd have to
125 rely on CSE to eliminate the duplicate allocation.... This alternative
126 doesn't look attractive enough to pursue.
128 ALTERNATIVE 3: ignore the reboxing problem. The trouble is that
129 the conservative reboxing story prevents many useful functions from being
130 specialised. Example:
131 foo :: Maybe Int -> Int -> Int
133 foo x@(Just m) n = foo x (n-m)
134 Here the use of 'x' will clearly not require boxing in the specialised function.
136 The strictness analyser has the same problem, in fact. Example:
138 If we pass just 'a' and 'b' to the worker, it might need to rebox the
139 pair to create (a,b). A more sophisticated analysis might figure out
140 precisely the cases in which this could happen, but the strictness
141 analyser does no such analysis; it just passes 'a' and 'b', and hopes
144 So my current choice is to make SpecConstr similarly aggressive, and
145 ignore the bad potential of reboxing.
148 Note [Good arguments]
149 ~~~~~~~~~~~~~~~~~~~~~
152 * A self-recursive function. Ignore mutual recursion for now,
153 because it's less common, and the code is simpler for self-recursion.
157 a) At a recursive call, one or more parameters is an explicit
158 constructor application
160 That same parameter is scrutinised by a case somewhere in
161 the RHS of the function
165 b) At a recursive call, one or more parameters has an unfolding
166 that is an explicit constructor application
168 That same parameter is scrutinised by a case somewhere in
169 the RHS of the function
171 Those are the only uses of the parameter (see Note [Reboxing])
174 What to abstract over
175 ~~~~~~~~~~~~~~~~~~~~~
176 There's a bit of a complication with type arguments. If the call
179 f p = ...f ((:) [a] x xs)...
181 then our specialised function look like
183 f_spec x xs = let p = (:) [a] x xs in ....as before....
185 This only makes sense if either
186 a) the type variable 'a' is in scope at the top of f, or
187 b) the type variable 'a' is an argument to f (and hence fs)
189 Actually, (a) may hold for value arguments too, in which case
190 we may not want to pass them. Supose 'x' is in scope at f's
191 defn, but xs is not. Then we'd like
193 f_spec xs = let p = (:) [a] x xs in ....as before....
195 Similarly (b) may hold too. If x is already an argument at the
196 call, no need to pass it again.
198 Finally, if 'a' is not in scope at the call site, we could abstract
199 it as we do the term variables:
201 f_spec a x xs = let p = (:) [a] x xs in ...as before...
203 So the grand plan is:
205 * abstract the call site to a constructor-only pattern
206 e.g. C x (D (f p) (g q)) ==> C s1 (D s2 s3)
208 * Find the free variables of the abstracted pattern
210 * Pass these variables, less any that are in scope at
211 the fn defn. But see Note [Shadowing] below.
214 NOTICE that we only abstract over variables that are not in scope,
215 so we're in no danger of shadowing variables used in "higher up"
221 In this pass we gather up usage information that may mention variables
222 that are bound between the usage site and the definition site; or (more
223 seriously) may be bound to something different at the definition site.
226 f x = letrec g y v = let x = ...
229 Since 'x' is in scope at the call site, we may make a rewrite rule that
231 RULE forall a,b. g (a,b) x = ...
232 But this rule will never match, because it's really a different 'x' at
233 the call site -- and that difference will be manifest by the time the
234 simplifier gets to it. [A worry: the simplifier doesn't *guarantee*
235 no-shadowing, so perhaps it may not be distinct?]
237 Anyway, the rule isn't actually wrong, it's just not useful. One possibility
238 is to run deShadowBinds before running SpecConstr, but instead we run the
239 simplifier. That gives the simplest possible program for SpecConstr to
240 chew on; and it virtually guarantees no shadowing.
242 Note [Specialising for constant parameters]
243 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
244 This one is about specialising on a *constant* (but not necessarily
245 constructor) argument
247 foo :: Int -> (Int -> Int) -> Int
249 foo m f = foo (f m) (+1)
253 lvl_rmV :: GHC.Base.Int -> GHC.Base.Int
255 \ (ds_dlk :: GHC.Base.Int) ->
256 case ds_dlk of wild_alH { GHC.Base.I# x_alG ->
257 GHC.Base.I# (GHC.Prim.+# x_alG 1)
259 T.$wfoo :: GHC.Prim.Int# -> (GHC.Base.Int -> GHC.Base.Int) ->
262 \ (ww_sme :: GHC.Prim.Int#) (w_smg :: GHC.Base.Int -> GHC.Base.Int) ->
263 case ww_sme of ds_Xlw {
265 case w_smg (GHC.Base.I# ds_Xlw) of w1_Xmo { GHC.Base.I# ww1_Xmz ->
266 T.$wfoo ww1_Xmz lvl_rmV
271 The recursive call has lvl_rmV as its argument, so we could create a specialised copy
272 with that argument baked in; that is, not passed at all. Now it can perhaps be inlined.
274 When is this worth it? Call the constant 'lvl'
275 - If 'lvl' has an unfolding that is a constructor, see if the corresponding
276 parameter is scrutinised anywhere in the body.
278 - If 'lvl' has an unfolding that is a inlinable function, see if the corresponding
279 parameter is applied (...to enough arguments...?)
281 Also do this is if the function has RULES?
285 Note [Specialising for lambda parameters]
286 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
287 foo :: Int -> (Int -> Int) -> Int
289 foo m f = foo (f m) (\n -> n-m)
291 This is subtly different from the previous one in that we get an
292 explicit lambda as the argument:
294 T.$wfoo :: GHC.Prim.Int# -> (GHC.Base.Int -> GHC.Base.Int) ->
297 \ (ww_sm8 :: GHC.Prim.Int#) (w_sma :: GHC.Base.Int -> GHC.Base.Int) ->
298 case ww_sm8 of ds_Xlr {
300 case w_sma (GHC.Base.I# ds_Xlr) of w1_Xmf { GHC.Base.I# ww1_Xmq ->
303 (\ (n_ad3 :: GHC.Base.Int) ->
304 case n_ad3 of wild_alB { GHC.Base.I# x_alA ->
305 GHC.Base.I# (GHC.Prim.-# x_alA ds_Xlr)
311 I wonder if SpecConstr couldn't be extended to handle this? After all,
312 lambda is a sort of constructor for functions and perhaps it already
313 has most of the necessary machinery?
315 Furthermore, there's an immediate win, because you don't need to allocate the lamda
316 at the call site; and if perchance it's called in the recursive call, then you
317 may avoid allocating it altogether. Just like for constructors.
319 Looks cool, but probably rare...but it might be easy to implement.
322 Note [SpecConstr for casts]
323 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
326 data instance T Int = T Int
331 go (T n) = go (T (n-1))
333 The recursive call ends up looking like
334 go (T (I# ...) `cast` g)
335 So we want to spot the construtor application inside the cast.
336 That's why we have the Cast case in argToPat
339 -----------------------------------------------------
340 Stuff not yet handled
341 -----------------------------------------------------
343 Here are notes arising from Roman's work that I don't want to lose.
349 foo :: Int -> T Int -> Int
351 foo x t | even x = case t of { T n -> foo (x-n) t }
352 | otherwise = foo (x-1) t
354 SpecConstr does no specialisation, because the second recursive call
355 looks like a boxed use of the argument. A pity.
357 $wfoo_sFw :: GHC.Prim.Int# -> T.T GHC.Base.Int -> GHC.Prim.Int#
359 \ (ww_sFo [Just L] :: GHC.Prim.Int#) (w_sFq [Just L] :: T.T GHC.Base.Int) ->
360 case ww_sFo of ds_Xw6 [Just L] {
362 case GHC.Prim.remInt# ds_Xw6 2 of wild1_aEF [Dead Just A] {
363 __DEFAULT -> $wfoo_sFw (GHC.Prim.-# ds_Xw6 1) w_sFq;
365 case w_sFq of wild_Xy [Just L] { T.T n_ad5 [Just U(L)] ->
366 case n_ad5 of wild1_aET [Just A] { GHC.Base.I# y_aES [Just L] ->
367 $wfoo_sFw (GHC.Prim.-# ds_Xw6 y_aES) wild_Xy
373 data a :*: b = !a :*: !b
376 foo :: (Int :*: T Int) -> Int
378 foo (x :*: t) | even x = case t of { T n -> foo ((x-n) :*: t) }
379 | otherwise = foo ((x-1) :*: t)
381 Very similar to the previous one, except that the parameters are now in
382 a strict tuple. Before SpecConstr, we have
384 $wfoo_sG3 :: GHC.Prim.Int# -> T.T GHC.Base.Int -> GHC.Prim.Int#
386 \ (ww_sFU [Just L] :: GHC.Prim.Int#) (ww_sFW [Just L] :: T.T
388 case ww_sFU of ds_Xws [Just L] {
390 case GHC.Prim.remInt# ds_Xws 2 of wild1_aEZ [Dead Just A] {
392 case ww_sFW of tpl_B2 [Just L] { T.T a_sFo [Just A] ->
393 $wfoo_sG3 (GHC.Prim.-# ds_Xws 1) tpl_B2 -- $wfoo1
396 case ww_sFW of wild_XB [Just A] { T.T n_ad7 [Just S(L)] ->
397 case n_ad7 of wild1_aFd [Just L] { GHC.Base.I# y_aFc [Just L] ->
398 $wfoo_sG3 (GHC.Prim.-# ds_Xws y_aFc) wild_XB -- $wfoo2
402 We get two specialisations:
403 "SC:$wfoo1" [0] __forall {a_sFB :: GHC.Base.Int sc_sGC :: GHC.Prim.Int#}
404 Foo.$wfoo sc_sGC (Foo.T @ GHC.Base.Int a_sFB)
405 = Foo.$s$wfoo1 a_sFB sc_sGC ;
406 "SC:$wfoo2" [0] __forall {y_aFp :: GHC.Prim.Int# sc_sGC :: GHC.Prim.Int#}
407 Foo.$wfoo sc_sGC (Foo.T @ GHC.Base.Int (GHC.Base.I# y_aFp))
408 = Foo.$s$wfoo y_aFp sc_sGC ;
410 But perhaps the first one isn't good. After all, we know that tpl_B2 is
411 a T (I# x) really, because T is strict and Int has one constructor. (We can't
412 unbox the strict fields, becuase T is polymorphic!)
416 %************************************************************************
418 \subsection{Top level wrapper stuff}
420 %************************************************************************
423 specConstrProgram :: DynFlags -> UniqSupply -> [CoreBind] -> IO [CoreBind]
424 specConstrProgram dflags us binds
426 showPass dflags "SpecConstr"
428 let (binds', _) = initUs us (go emptyScEnv binds)
430 endPass dflags "SpecConstr" Opt_D_dump_spec binds'
432 dumpIfSet_dyn dflags Opt_D_dump_rules "Top-level specialisations"
433 (pprRules (tidyRules emptyTidyEnv (rulesOfBinds binds')))
437 go env [] = returnUs []
438 go env (bind:binds) = scBind env bind `thenUs` \ (env', _, bind') ->
439 go env' binds `thenUs` \ binds' ->
440 returnUs (bind' : binds')
444 %************************************************************************
446 \subsection{Environment: goes downwards}
448 %************************************************************************
451 data ScEnv = SCE { scope :: InScopeEnv,
452 -- Binds all non-top-level variables in scope
457 type InScopeEnv = VarEnv HowBound
459 type ConstrEnv = IdEnv ConValue
460 data ConValue = CV AltCon [CoreArg]
461 -- Variables known to be bound to a constructor
462 -- in a particular case alternative
465 instance Outputable ConValue where
466 ppr (CV con args) = ppr con <+> interpp'SP args
468 emptyScEnv = SCE { scope = emptyVarEnv, cons = emptyVarEnv }
470 data HowBound = RecFun -- These are the recursive functions for which
471 -- we seek interesting call patterns
473 | RecArg -- These are those functions' arguments, or their sub-components;
474 -- we gather occurrence information for these
476 | Other -- We track all others so we know what's in scope
477 -- This is used in spec_one to check what needs to be
478 -- passed as a parameter and what is in scope at the
479 -- function definition site
481 instance Outputable HowBound where
482 ppr RecFun = text "RecFun"
483 ppr RecArg = text "RecArg"
484 ppr Other = text "Other"
486 lookupScopeEnv env v = lookupVarEnv (scope env) v
488 extendBndrs env bndrs = env { scope = extendVarEnvList (scope env) [(b,Other) | b <- bndrs] }
489 extendBndr env bndr = env { scope = extendVarEnv (scope env) bndr Other }
494 -- we want to bind b, and perhaps scrut too, to (C x y)
495 extendCaseBndrs :: ScEnv -> Id -> CoreExpr -> AltCon -> [Var] -> ScEnv
496 extendCaseBndrs env case_bndr scrut con alt_bndrs
499 LitAlt lit -> extendCons env1 scrut case_bndr (CV con [])
500 DataAlt dc -> extend_data_con dc
502 cur_scope = scope env
503 env1 = env { scope = extendVarEnvList cur_scope
504 [(b,how_bound) | b <- case_bndr:alt_bndrs] }
506 -- Record RecArg for the components iff the scrutinee is RecArg
507 -- I think the only reason for this is to keep the usage envt small
508 -- so is it worth it at all?
509 -- [This comment looks plain wrong to me, so I'm ignoring it
510 -- "Also forget if the scrutinee is a RecArg, because we're
511 -- now in the branch of a case, and we don't want to
512 -- record a non-scrutinee use of v if we have
513 -- case v of { (a,b) -> ...(f v)... }" ]
514 how_bound = get_how scrut
516 get_how (Var v) = lookupVarEnv cur_scope v `orElse` Other
517 get_how (Cast e _) = get_how e
518 get_how (Note _ e) = get_how e
519 get_how other = Other
521 extend_data_con data_con =
522 extendCons env1 scrut case_bndr (CV con vanilla_args)
524 vanilla_args = map Type (tyConAppArgs (idType case_bndr)) ++
525 varsToCoreExprs alt_bndrs
527 extendCons :: ScEnv -> CoreExpr -> Id -> ConValue -> ScEnv
528 extendCons env scrut case_bndr val
530 Var v -> env { cons = extendVarEnv cons1 v val }
531 other -> env { cons = cons1 }
533 cons1 = extendVarEnv (cons env) case_bndr val
535 -- When we encounter a recursive function binding
537 -- we want to extend the scope env with bindings
538 -- that record that f is a RecFn and x,y are RecArgs
539 extendRecBndr env fn bndrs
540 = env { scope = scope env `extendVarEnvList`
541 ((fn,RecFun): [(bndr,RecArg) | bndr <- bndrs]) }
545 %************************************************************************
547 \subsection{Usage information: flows upwards}
549 %************************************************************************
554 calls :: !(IdEnv ([Call])), -- Calls
555 -- The functions are a subset of the
556 -- RecFuns in the ScEnv
558 occs :: !(IdEnv ArgOcc) -- Information on argument occurrences
559 } -- The variables are a subset of the
560 -- RecArg in the ScEnv
562 type Call = (ConstrEnv, [CoreArg])
563 -- The arguments of the call, together with the
564 -- env giving the constructor bindings at the call site
566 nullUsage = SCU { calls = emptyVarEnv, occs = emptyVarEnv }
568 combineUsage u1 u2 = SCU { calls = plusVarEnv_C (++) (calls u1) (calls u2),
569 occs = plusVarEnv_C combineOcc (occs u1) (occs u2) }
571 combineUsages [] = nullUsage
572 combineUsages us = foldr1 combineUsage us
574 lookupOcc :: ScUsage -> Var -> (ScUsage, ArgOcc)
575 lookupOcc (SCU { calls = sc_calls, occs = sc_occs }) bndr
576 = (SCU {calls = sc_calls, occs = delVarEnv sc_occs bndr},
577 lookupVarEnv sc_occs bndr `orElse` NoOcc)
579 lookupOccs :: ScUsage -> [Var] -> (ScUsage, [ArgOcc])
580 lookupOccs (SCU { calls = sc_calls, occs = sc_occs }) bndrs
581 = (SCU {calls = sc_calls, occs = delVarEnvList sc_occs bndrs},
582 [lookupVarEnv sc_occs b `orElse` NoOcc | b <- bndrs])
584 data ArgOcc = NoOcc -- Doesn't occur at all; or a type argument
585 | UnkOcc -- Used in some unknown way
587 | ScrutOcc (UniqFM [ArgOcc]) -- See Note [ScrutOcc]
589 | BothOcc -- Definitely taken apart, *and* perhaps used in some other way
593 An occurrence of ScrutOcc indicates that the thing, or a `cast` version of the thing,
594 is *only* taken apart or applied.
596 Functions, literal: ScrutOcc emptyUFM
597 Data constructors: ScrutOcc subs,
599 where (subs :: UniqFM [ArgOcc]) gives usage of the *pattern-bound* components,
600 The domain of the UniqFM is the Unique of the data constructor
602 The [ArgOcc] is the occurrences of the *pattern-bound* components
603 of the data structure. E.g.
604 data T a = forall b. MkT a b (b->a)
605 A pattern binds b, x::a, y::b, z::b->a, but not 'a'!
609 instance Outputable ArgOcc where
610 ppr (ScrutOcc xs) = ptext SLIT("scrut-occ") <> ppr xs
611 ppr UnkOcc = ptext SLIT("unk-occ")
612 ppr BothOcc = ptext SLIT("both-occ")
613 ppr NoOcc = ptext SLIT("no-occ")
615 -- Experimentally, this vesion of combineOcc makes ScrutOcc "win", so
616 -- that if the thing is scrutinised anywhere then we get to see that
617 -- in the overall result, even if it's also used in a boxed way
618 -- This might be too agressive; see Note [Reboxing] Alternative 3
619 combineOcc NoOcc occ = occ
620 combineOcc occ NoOcc = occ
621 combineOcc (ScrutOcc xs) (ScrutOcc ys) = ScrutOcc (plusUFM_C combineOccs xs ys)
622 combineOcc occ (ScrutOcc ys) = ScrutOcc ys
623 combineOcc (ScrutOcc xs) occ = ScrutOcc xs
624 combineOcc UnkOcc UnkOcc = UnkOcc
625 combineOcc _ _ = BothOcc
627 combineOccs :: [ArgOcc] -> [ArgOcc] -> [ArgOcc]
628 combineOccs xs ys = zipWithEqual "combineOccs" combineOcc xs ys
630 conArgOccs :: ArgOcc -> AltCon -> [ArgOcc]
631 -- Find usage of components of data con; returns [UnkOcc...] if unknown
632 -- See Note [ScrutOcc] for the extra UnkOccs in the vanilla datacon case
634 conArgOccs (ScrutOcc fm) (DataAlt dc)
635 | Just pat_arg_occs <- lookupUFM fm dc
636 = [UnkOcc | tv <- dataConUnivTyVars dc] ++ pat_arg_occs
638 conArgOccs other con = repeat UnkOcc
642 %************************************************************************
644 \subsection{The main recursive function}
646 %************************************************************************
648 The main recursive function gathers up usage information, and
649 creates specialised versions of functions.
652 scExpr :: ScEnv -> CoreExpr -> UniqSM (ScUsage, CoreExpr)
653 -- The unique supply is needed when we invent
654 -- a new name for the specialised function and its args
656 scExpr env e@(Type t) = returnUs (nullUsage, e)
657 scExpr env e@(Lit l) = returnUs (nullUsage, e)
658 scExpr env e@(Var v) = returnUs (varUsage env v UnkOcc, e)
659 scExpr env (Note n e) = scExpr env e `thenUs` \ (usg,e') ->
660 returnUs (usg, Note n e')
661 scExpr env (Cast e co)= scExpr env e `thenUs` \ (usg,e') ->
662 returnUs (usg, Cast e' co)
663 scExpr env (Lam b e) = scExpr (extendBndr env b) e `thenUs` \ (usg,e') ->
664 returnUs (usg, Lam b e')
666 scExpr env (Case scrut b ty alts)
667 = do { (alt_usgs, alt_occs, alts') <- mapAndUnzip3Us sc_alt alts
668 ; let (alt_usg, b_occ) = lookupOcc (combineUsages alt_usgs) b
669 scrut_occ = foldr combineOcc b_occ alt_occs
670 -- The combined usage of the scrutinee is given
671 -- by scrut_occ, which is passed to scScrut, which
672 -- in turn treats a bare-variable scrutinee specially
673 ; (scrut_usg, scrut') <- scScrut env scrut scrut_occ
674 ; return (alt_usg `combineUsage` scrut_usg,
675 Case scrut' b ty alts') }
678 = do { let env1 = extendCaseBndrs env b scrut con bs
679 ; (usg,rhs') <- scExpr env1 rhs
680 ; let (usg', arg_occs) = lookupOccs usg bs
681 scrut_occ = case con of
682 DataAlt dc -> ScrutOcc (unitUFM dc arg_occs)
683 other -> ScrutOcc emptyUFM
684 ; return (usg', scrut_occ, (con,bs,rhs')) }
686 scExpr env (Let bind body)
687 = scBind env bind `thenUs` \ (env', bind_usg, bind') ->
688 scExpr env' body `thenUs` \ (body_usg, body') ->
689 returnUs (bind_usg `combineUsage` body_usg, Let bind' body')
691 scExpr env e@(App _ _)
692 = do { let (fn, args) = collectArgs e
693 ; (fn_usg, fn') <- scScrut env fn (ScrutOcc emptyUFM)
694 -- Process the function too. It's almost always a variable,
695 -- but not always. In particular, if this pass follows float-in,
696 -- which it may, we can get
697 -- (let f = ...f... in f) arg1 arg2
698 -- We use scScrut to record the fact that the function is called
699 -- Perhpas we should check that it has at least one value arg,
700 -- but currently we don't bother
702 ; (arg_usgs, args') <- mapAndUnzipUs (scExpr env) args
703 ; let call_usg = case fn of
704 Var f | Just RecFun <- lookupScopeEnv env f
705 -> SCU { calls = unitVarEnv f [(cons env, args)],
708 ; return (combineUsages arg_usgs `combineUsage` fn_usg
709 `combineUsage` call_usg,
713 ----------------------
714 scScrut :: ScEnv -> CoreExpr -> ArgOcc -> UniqSM (ScUsage, CoreExpr)
715 -- Used for the scrutinee of a case,
716 -- or the function of an application.
717 -- Remember to look through casts
718 scScrut env e@(Var v) occ = returnUs (varUsage env v occ, e)
719 scScrut env (Cast e co) occ = do { (usg, e') <- scScrut env e occ
720 ; returnUs (usg, Cast e' co) }
721 scScrut env e occ = scExpr env e
724 ----------------------
725 scBind :: ScEnv -> CoreBind -> UniqSM (ScEnv, ScUsage, CoreBind)
726 scBind env (Rec [(fn,rhs)])
728 = scExpr env_fn_body body `thenUs` \ (usg, body') ->
729 specialise env fn bndrs body' usg `thenUs` \ (rules, spec_prs) ->
730 -- Note body': the specialised copies should be based on the
731 -- optimised version of the body, in case there were
732 -- nested functions inside.
734 SCU { calls = calls, occs = occs } = usg
736 returnUs (extendBndr env fn, -- For the body of the letrec, just
737 -- extend the env with Other to record
738 -- that it's in scope; no funny RecFun business
739 SCU { calls = calls `delVarEnv` fn, occs = occs `delVarEnvList` val_bndrs},
740 Rec ((fn `addIdSpecialisations` rules, mkLams bndrs body') : spec_prs))
742 (bndrs,body) = collectBinders rhs
743 val_bndrs = filter isId bndrs
744 env_fn_body = extendRecBndr env fn bndrs
747 = mapAndUnzipUs do_one prs `thenUs` \ (usgs, prs') ->
748 returnUs (extendBndrs env (map fst prs), combineUsages usgs, Rec prs')
750 do_one (bndr,rhs) = scExpr env rhs `thenUs` \ (usg, rhs') ->
751 returnUs (usg, (bndr,rhs'))
753 scBind env (NonRec bndr rhs)
754 = scExpr env rhs `thenUs` \ (usg, rhs') ->
755 returnUs (extendBndr env bndr, usg, NonRec bndr rhs')
757 ----------------------
759 | Just RecArg <- lookupScopeEnv env v = SCU { calls = emptyVarEnv,
760 occs = unitVarEnv v use }
761 | otherwise = nullUsage
765 %************************************************************************
767 \subsection{The specialiser}
769 %************************************************************************
774 -> [CoreBndr] -> CoreExpr -- Its RHS
775 -> ScUsage -- Info on usage
776 -> UniqSM ([CoreRule], -- Rules
777 [(Id,CoreExpr)]) -- Bindings
779 specialise env fn bndrs body body_usg
780 = do { let (_, bndr_occs) = lookupOccs body_usg bndrs
781 all_calls = lookupVarEnv (calls body_usg) fn `orElse` []
783 ; mb_pats <- mapM (callToPats (scope env) bndr_occs) all_calls
785 ; let good_pats :: [([Var], [CoreArg])]
786 good_pats = catMaybes mb_pats
787 in_scope = mkInScopeSet $ unionVarSets $
788 [ exprsFreeVars pats `delVarSetList` vs
789 | (vs,pats) <- good_pats ]
790 uniq_pats = nubBy (same_pat in_scope) good_pats
791 ; -- pprTrace "specialise" (vcat [ppr fn <+> ppr bndrs <+> ppr bndr_occs,
792 -- text "calls" <+> ppr all_calls,
793 -- text "good pats" <+> ppr good_pats,
794 -- text "uniq pats" <+> ppr uniq_pats]) $
795 mapAndUnzipUs (spec_one env fn (mkLams bndrs body))
796 (uniq_pats `zip` [1..]) }
798 -- Two pats are the same if they match both ways
799 same_pat in_scope (vs1,as1)(vs2,as2)
800 = isJust (matchN in_scope vs1 as1 as2)
801 && isJust (matchN in_scope vs2 as2 as1)
803 callToPats :: InScopeEnv -> [ArgOcc] -> Call
804 -> UniqSM (Maybe ([Var], [CoreExpr]))
805 -- The VarSet is the variables to quantify over in the rule
806 -- The [CoreExpr] are the argument patterns for the rule
807 callToPats in_scope bndr_occs (con_env, args)
808 | length args < length bndr_occs -- Check saturated
811 = do { prs <- argsToPats in_scope con_env (args `zip` bndr_occs)
812 ; let (good_pats, pats) = unzip prs
813 pat_fvs = varSetElems (exprsFreeVars pats)
814 qvars = filter (not . (`elemVarEnv` in_scope)) pat_fvs
815 -- Quantify over variables that are not in sccpe
816 -- See Note [Shadowing] at the top
818 ; -- pprTrace "callToPats" (ppr args $$ ppr prs $$ ppr bndr_occs) $
820 then return (Just (qvars, pats))
821 else return Nothing }
823 ---------------------
826 -> CoreExpr -- Rhs of the original function
827 -> (([Var], [CoreArg]), Int)
828 -> UniqSM (CoreRule, (Id,CoreExpr)) -- Rule and binding
830 -- spec_one creates a specialised copy of the function, together
831 -- with a rule for using it. I'm very proud of how short this
832 -- function is, considering what it does :-).
838 f = /\b \y::[(a,b)] -> ....f (b,c) ((:) (a,(b,c)) (x,v) (h w))...
839 [c::*, v::(b,c) are presumably bound by the (...) part]
841 f_spec = /\ b c \ v::(b,c) hw::[(a,(b,c))] ->
842 (...entire RHS of f...) (b,c) ((:) (a,(b,c)) (x,v) hw)
844 RULE: forall b::* c::*, -- Note, *not* forall a, x
848 f (b,c) ((:) (a,(b,c)) (x,v) hw) = f_spec b c v hw
851 spec_one env fn rhs ((vars_to_bind, pats), rule_number)
852 = getUniqueUs `thenUs` \ spec_uniq ->
855 fn_loc = nameSrcLoc fn_name
856 spec_occ = mkSpecOcc (nameOccName fn_name)
858 -- Put the type variables first; the type of a term
859 -- variable may mention a type variable
860 (tvs, ids) = partition isTyVar vars_to_bind
862 spec_body = mkApps rhs pats
863 body_ty = exprType spec_body
865 (spec_lam_args, spec_call_args) = mkWorkerArgs bndrs body_ty
866 -- Usual w/w hack to avoid generating
867 -- a spec_rhs of unlifted type and no args
869 rule_name = mkFastString ("SC:" ++ showSDoc (ppr fn <> int rule_number))
870 spec_rhs = mkLams spec_lam_args spec_body
871 spec_id = mkUserLocal spec_occ spec_uniq (mkPiTypes spec_lam_args body_ty) fn_loc
872 rule_rhs = mkVarApps (Var spec_id) spec_call_args
873 rule = mkLocalRule rule_name specConstrActivation fn_name bndrs pats rule_rhs
875 returnUs (rule, (spec_id, spec_rhs))
877 -- In which phase should the specialise-constructor rules be active?
878 -- Originally I made them always-active, but Manuel found that
879 -- this defeated some clever user-written rules. So Plan B
880 -- is to make them active only in Phase 0; after all, currently,
881 -- the specConstr transformation is only run after the simplifier
882 -- has reached Phase 0. In general one would want it to be
883 -- flag-controllable, but for now I'm leaving it baked in
885 specConstrActivation :: Activation
886 specConstrActivation = ActiveAfter 0 -- Baked in; see comments above
889 %************************************************************************
891 \subsection{Argument analysis}
893 %************************************************************************
895 This code deals with analysing call-site arguments to see whether
896 they are constructor applications.
900 -- argToPat takes an actual argument, and returns an abstracted
901 -- version, consisting of just the "constructor skeleton" of the
902 -- argument, with non-constructor sub-expression replaced by new
903 -- placeholder variables. For example:
904 -- C a (D (f x) (g y)) ==> C p1 (D p2 p3)
906 argToPat :: InScopeEnv -- What's in scope at the fn defn site
907 -> ConstrEnv -- ConstrEnv at the call site
908 -> CoreArg -- A call arg (or component thereof)
910 -> UniqSM (Bool, CoreArg)
911 -- Returns (interesting, pat),
912 -- where pat is the pattern derived from the argument
913 -- intersting=True if the pattern is non-trivial (not a variable or type)
914 -- E.g. x:xs --> (True, x:xs)
915 -- f xs --> (False, w) where w is a fresh wildcard
916 -- (f xs, 'c') --> (True, (w, 'c')) where w is a fresh wildcard
917 -- \x. x+y --> (True, \x. x+y)
918 -- lvl7 --> (True, lvl7) if lvl7 is bound
919 -- somewhere further out
921 argToPat in_scope con_env arg@(Type ty) arg_occ
922 = return (False, arg)
924 argToPat in_scope con_env (Let _ arg) arg_occ
925 = argToPat in_scope con_env arg arg_occ
926 -- Look through let expressions
927 -- e.g. f (let v = rhs in \y -> ...v...)
928 -- Here we can specialise for f (\y -> ...)
929 -- because the rule-matcher will look through the let.
931 argToPat in_scope con_env (Cast arg co) arg_occ
932 = do { (interesting, arg') <- argToPat in_scope con_env arg arg_occ
933 ; if interesting then
934 return (interesting, Cast arg' co)
936 wildCardPat (snd (coercionKind co)) }
938 argToPat in_scope con_env arg arg_occ
942 is_value_lam (Lam v e) -- Spot a value lambda, even if
943 | isId v = True -- it is inside a type lambda
944 | otherwise = is_value_lam e
945 is_value_lam other = False
947 -- Check for a constructor application
948 -- NB: this *precedes* the Var case, so that we catch nullary constrs
949 argToPat in_scope con_env arg arg_occ
950 | Just (CV dc args) <- is_con_app_maybe con_env arg
952 ScrutOcc _ -> True -- Used only by case scrutinee
953 BothOcc -> case arg of -- Used elsewhere
954 App {} -> True -- see Note [Reboxing]
956 other -> False -- No point; the arg is not decomposed
957 = do { args' <- argsToPats in_scope con_env (args `zip` conArgOccs arg_occ dc)
958 ; return (True, mk_con_app dc (map snd args')) }
960 -- Check if the argument is a variable that
961 -- is in scope at the function definition site
962 -- It's worth specialising on this if
963 -- (a) it's used in an interesting way in the body
964 -- (b) we know what its value is
965 argToPat in_scope con_env (Var v) arg_occ
966 | not (isLocalId v) || v `elemVarEnv` in_scope,
967 case arg_occ of { UnkOcc -> False; other -> True }, -- (a)
968 isValueUnfolding (idUnfolding v) -- (b)
969 = return (True, Var v)
971 -- Check for a variable bound inside the function.
972 -- Don't make a wild-card, because we may usefully share
973 -- e.g. f a = let x = ... in f (x,x)
974 -- NB: this case follows the lambda and con-app cases!!
975 argToPat in_scope con_env (Var v) arg_occ
976 = return (False, Var v)
978 -- The default case: make a wild-card
979 argToPat in_scope con_env arg arg_occ
980 = wildCardPat (exprType arg)
982 wildCardPat :: Type -> UniqSM (Bool, CoreArg)
983 wildCardPat ty = do { uniq <- getUniqueUs
984 ; let id = mkSysLocal FSLIT("sc") uniq ty
985 ; return (False, Var id) }
987 argsToPats :: InScopeEnv -> ConstrEnv
988 -> [(CoreArg, ArgOcc)]
989 -> UniqSM [(Bool, CoreArg)]
990 argsToPats in_scope con_env args
993 do_one (arg,occ) = argToPat in_scope con_env arg occ
998 is_con_app_maybe :: ConstrEnv -> CoreExpr -> Maybe ConValue
999 is_con_app_maybe env (Lit lit)
1000 = Just (CV (LitAlt lit) [])
1002 is_con_app_maybe env expr -- Maybe it's a constructor application
1003 | (Var fun, args) <- collectArgs expr,
1004 Just con <- isDataConWorkId_maybe fun,
1005 args `lengthAtLeast` dataConRepArity con
1006 -- Might be > because the arity excludes type args
1007 = Just (CV (DataAlt con) args)
1009 is_con_app_maybe env (Var v)
1010 | Just stuff <- lookupVarEnv env v
1011 = Just stuff -- You might think we could look in the idUnfolding here
1012 -- but that doesn't take account of which branch of a
1013 -- case we are in, which is the whole point
1015 | isCheapUnfolding unf
1016 = is_con_app_maybe env (unfoldingTemplate unf)
1019 -- However we do want to consult the unfolding
1020 -- as well, for let-bound constructors!
1022 is_con_app_maybe env expr = Nothing
1024 mk_con_app :: AltCon -> [CoreArg] -> CoreExpr
1025 mk_con_app (LitAlt lit) [] = Lit lit
1026 mk_con_app (DataAlt con) args = mkConApp con args
1027 mk_con_app other args = panic "SpecConstr.mk_con_app"