#4514: System.Timeout cannot properly cancel IO actions with new IO manager
---------------------------------+------------------------------------------
Reporter: adept | Owner: tibbe
Type: bug | Status: new
Priority: normal | Component: Runtime System
Version: 7.0.1 | Keywords:
Testcase: | Blockedby:
Os: Unknown/Multiple | Blocking:
Architecture: Unknown/Multiple | Failure: Incorrect result at runtime
---------------------------------+------------------------------------------
Comment(by adept):
Ok.
After chat with tibbe it seems like the root cause lies here:
{{{
registerFd_ :: EventManager -> IOCallback -> Fd -> Event
-> IO (FdKey, Bool)
registerFd_ EventManager{..} cb fd evs = do
u <- newUnique emUniqueSource
modifyMVar emFds $ \oldMap -> do
let fd' = fromIntegral fd
reg = FdKey fd u
!fdd = FdData reg evs cb
(!newMap, (oldEvs, newEvs)) =
case IM.insertWith (++) fd' [fdd] oldMap of
(Nothing, n) -> (n, (mempty, evs))
(Just prev, n) -> (n, pairEvents prev newMap fd')
modify = oldEvs /= newEvs
when modify $ I.modifyFd emBackend fd oldEvs newEvs
return (newMap, (reg, modify))
}}}
After reconnect, when `recv` is called, it checks for data with C `recv`
and (since there is nothing there yet) calls `threadWaitRead` on the
approprate descriptor. This in turn should call `registerFd`.
Since there is handler for the same FD in `emFds` already, `pairEvent`
would try to compute the combined event mask. In this case `oldEvs ==
newEvs`, so `I.modifyFd` would not be called, and this must be the source
of the problem.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4514#comment:4>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
