Hi Harmut,
Because the functions modify thread structures, the lock needs to be
held, IMO.
Also, according to Microsoft documents, they suggest to hold the lock
(they do in any case).
Best reagards,
Alex Ionescu
Hartmut Birr wrote:
Hi,
I think some of your changes in apc.c are wrong. Apc's are always thread
local. They have nothing to do with the dispatcher database lock. The only
locking object is the ApcQueueLock from the thread structure.
- Hartmut
-----Original Message-----
From: ros-cvs-bounces(a)reactos.com
[mailto:ros-cvs-bounces@reactos.com] On Behalf Of ion(a)cvs.reactos.com
Sent: Thursday, November 11, 2004 2:24 PM
To: ros-cvs(a)reactos.com
Subject: [ros-cvs] CVS Update: reactos
CVSROOT: /CVS/ReactOS
Module name: reactos
Repository: reactos/ntoskrnl/ps/
Changes by: ion(a)mok.osexperts.com 04/11/11 14:23:53
Modified files:
reactos/include/ddk/: kefuncs.h ketypes.h
reactos/ntoskrnl/include/internal/: ke.h ps.h
reactos/ntoskrnl/ke/: apc.c process.c wait.c
reactos/ntoskrnl/mm/: marea.c virtual.c
reactos/ntoskrnl/mm/i386/: page.c
reactos/ntoskrnl/ps/: psmgr.c
Log message:
Kernel:
- apc.c: Fixed some bugs and used Dispatcher Lock.
Simplified some functions and renamed/inlined others.
- wait.c: Fixed some assumptions about Dispatcher Lock
and Waiting/Alertability
- process.c: Fully implemented KeStackAttachProcess and
KeStackDetachProcess. Made all functions use PKPROCESS and
not PEPROCESS.
Memory Manager: Fixed calls to
KeAttachProcess/DetachProcess to typecast PKPROCESS
Process Manager: Removed call to empty function in apc.c
_______________________________________________
Ros-cvs mailing list
Ros-cvs(a)reactos.com
http://reactos.com/mailman/listinfo/ros-cvs
_______________________________________________
Ros-dev mailing list
Ros-dev(a)reactos.com
http://reactos.com:8080/mailman/listinfo/ros-dev