Commit ce64b397 authored by Sebastian Graf's avatar Sebastian Graf Committed by Marge Bot
Browse files

`exprOkForSpeculation` for Note [IO hack in the demand analyser]

In #14998 I realised that the notion of speculative execution
*exactly matches* eager evaluation of expressions in a case alternative
where the scrutinee is an IO action.

Normally we have to `deferIO` any result from that single case
alternative to prevent this speculative execution, so we had a special
case in place in the demand analyser that would check if the scrutinee
was a prim-op, in which case we assumed that it would be ok to do the
eager evaluation.

Now we just check if the scrutinee is `exprOkForSpeculation`,
corresponding to the notion that we want to push evaluation of the
scrutinee *after* eagerly evaluating stuff from the case alternative.

This fixes #14988, because it resolves the last open Item 4 there.
parent c5d888d4
......@@ -333,10 +333,7 @@ io_hack_reqd scrut con bndrs
| (bndr:_) <- bndrs
, con == tupleDataCon Unboxed 2
, idType bndr `eqType` realWorldStatePrimTy
, (fun, _) <- collectArgs scrut
= case fun of
Var f -> not (isPrimOpId f)
_ -> True
= not (exprOkForSpeculation scrut)
| otherwise
= False
......@@ -387,15 +384,18 @@ getMaskingState# is not going to diverge or throw an exception! This
situation actually arises in GHC.IO.Handle.Internals.wantReadableHandle
(on an MVar not an Int), and made a material difference.
So if the scrutinee is a primop call, we *don't* apply the
state hack:
So if the scrutinee is ok-for-speculation, we *don't* apply the state hack,
because we are free to push evaluation of the scrutinee after evaluation of
expressions from the (single) case alternative.
A few examples for different scrutinees:
- If it is a simple, terminating one like getMaskingState,
applying the hack is over-conservative.
- If the primop is raise# then it returns bottom, so
the case alternatives are already discarded.
applying the hack would be over-conservative.
- If the primop is raise# then it returns bottom (so not ok-for-speculation),
but the result from the case alternatives are discarded anyway.
- If the primop can raise a non-IO exception, like
divide by zero or seg-fault (eg writing an array
out of bounds) then we don't mind evaluating 'x' first.
divide by zero (so not ok-for-speculation), then we are also bottoming out
anyway and don't mind evaluating 'x' first.
Note [Demand on the scrutinee of a product case]
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment