Re: Alexandre Julliard : server: Add support for queuing a user APC upon async I/O completion.
Alexandre Julliard a écrit :
Module: wine Branch: master Commit: c16eb8efd942bfbb0bb7f3da4c740a2e483cce28 URL: http://source.winehq.org/git/wine.git/?a=commit;h=c16eb8efd942bfbb0bb7f3da4c...
Author: Alexandre Julliard <julliard(a)winehq.org> Date: Tue Mar 27 16:42:27 2007 +0200
server: Add support for queuing a user APC upon async I/O completion.
do you have a plan, when the apc routine is used, to free the allocated async_fileio (in dlls/ntdll/file.c) ? A+ -- Eric Pouech "The problem with designing something completely foolproof is to underestimate the ingenuity of a complete idiot." (Douglas Adams)
Eric Pouech <eric.pouech(a)wanadoo.fr> writes:
do you have a plan, when the apc routine is used, to free the allocated async_fileio (in dlls/ntdll/file.c) ?
The async_fileio is freed by the async callback that performs the actual I/O, that's not going to change. The APC here is the user APC called after completion, currently called directly by fileio_terminate (which is wrong since we may not be in an alertable wait at this point). -- Alexandre Julliard julliard(a)winehq.org
Alexandre Julliard a écrit :
Eric Pouech <eric.pouech(a)wanadoo.fr> writes:
do you have a plan, when the apc routine is used, to free the allocated async_fileio (in dlls/ntdll/file.c) ?
The async_fileio is freed by the async callback that performs the actual I/O, that's not going to change. The APC here is the user APC called after completion, currently called directly by fileio_terminate (which is wrong since we may not be in an alertable wait at this point).
ok, I didn't see you where queueing in fact two APCs A+ -- Eric Pouech "The problem with designing something completely foolproof is to underestimate the ingenuity of a complete idiot." (Douglas Adams)
participants (2)
-
Alexandre Julliard -
Eric Pouech