November 2007 - Wine-Devel - list.winehq.org

Re: bits: Description of patch set (resend 2) (0/26)

by Juan Lang

> I looked into switching svchost to ansi since there is no unicode > version of GetProcAddress (used within svchost). In general Unicode is the way to go, but there's no reason you can't use Unicode for those APIs that are, and ANSI for those that aren't, specifically, when querying the function name to be used with GetProcAddress. --Juan

Nov. 20, 2007

Re: Unimplemented function KERNEL32.dll.FindFirstVolumeW

by Pavel Troller

> mhhh - since it cost me more then one hour to find out why my machine reboots and triggering a reboot by just _reading_ a file on my system, device node or not is at least very weird behaviour from a user`s perspective. > > >waiting for an user-space daemon to start handling of /dev/watchdog file (opening it and manipulating it from time to time to show that it's still alive. > >When you cat the file, kernel recognizes it as that such an app is taking over control of the watchdog and starts it. > it`s looks a little bit too simple for me. triggering a reboot just by some simple open() and read() and close() ? > let`s wait what the maintainer has to tell. > > regards > roland > OK, Roland, but it's Linux :-). Isn't "rm -rf /" a simple command ? Yes, it is :-). Less keystrokes than "cat /dev/watchdog". And its impact is even more catastrophic :-). It's the reason why working as root is so much discouraged. With regards, Pavel Troller

Nov. 20, 2007

Re: server: Change the get_dll_info server request to allowretrievingthe image file name of a process.

by Robert Shearman

Dmitry Timoshkov wrote: > "Robert Shearman" <rob(a)codeweavers.com> wrote: > >>> Main executable has its own base address which should work just fine >>> in find_process_dll(). >> >> Really? http://msdn2.microsoft.com/en-us/library/f7f5138s(VS.71).aspx > > I don't see how the page above is related to what I said. Main exe module > is being added to the process dlls list at the end of the initialization > process by load_dll server call with its real base address. I don't see how that is relevant either. We don't know the main exe module base address in anything other than the server and that one process. -- Rob Shearman

Nov. 20, 2007

Re: server: Change the get_dll_info server request to allowretrievingthe image file name of a process.

by Dmitry Timoshkov

"Robert Shearman" <rob(a)codeweavers.com> wrote: >> Main executable has its own base address which should work just fine >> in find_process_dll(). > > Really? http://msdn2.microsoft.com/en-us/library/f7f5138s(VS.71).aspx I don't see how the page above is related to what I said. Main exe module is being added to the process dlls list at the end of the initialization process by load_dll server call with its real base address. -- Dmitry.

Nov. 20, 2007

Re: server: Change the get_dll_info server request to allow retrievingthe image file name of a process.

by Alexandre Julliard

Robert Shearman <rob(a)codeweavers.com> writes: > Dmitry Timoshkov wrote: >> Is it really necessary to introduce the exe_module request parameter? > > Yes. > >> Main executable has its own base address which should work just fine >> in find_process_dll(). > > Really? http://msdn2.microsoft.com/en-us/library/f7f5138s(VS.71).aspx The usual convention is that a NULL module address means the main exe, I think that would be appropriate here. -- Alexandre Julliard julliard(a)winehq.org

Nov. 20, 2007

Re: d3dx8 [patch 1/10] Implement D3DXPlaneFromPointNormal

by David.Adam＠math.cnrs.fr

Hello, In the beginning, I thought to share the code between d3drm and d3dx8, but unfortunelaty, it was not a good idea, since the structure quaternion is not the same in the both dll (definitively 4 floats, but 4 floats in d3dx8, 1 float and a d3drmvector in d3drm). So, it was not pretty. Moreover, Once I asked to Alexandre what was the best method to implement d3dx9__.dll. He answered to implement the maximum of functions in d3dx8 and to call these functions from d3dx9__.dll. I proposed the idea of wined3dx.dll and he found that it was not a good idea. So, the basis dll is d3dx8 and we consider d3dx9__.dll as a wrapper (at least, for the math functions) Thunderbird was interested in implementing the other functions. Anyway, I am interested in it too. And also, I am interesting in completing d3drm since there still are a few bugs related to this dll. To finish my post: today (or tommorrow) I shall send the last patch for math-functions in d3dx8. The real job will begin then :D :D David

Nov. 20, 2007

Re: server: Change the get_dll_info server request to allow retrievingthe image file name of a process.

by Robert Shearman

Dmitry Timoshkov wrote: > "Robert Shearman" <rob(a)codeweavers.com> wrote: > >> + if (req->exe_module) >> + dll = list_head( &process->dlls ) ? >> + LIST_ENTRY(list_head( &process->dlls ), struct >> process_dll, entry) : NULL; >> + else >> + dll = find_process_dll( process, req->base_address ); > > Is it really necessary to introduce the exe_module request parameter? Yes. > Main executable has its own base address which should work just fine > in find_process_dll(). Really? http://msdn2.microsoft.com/en-us/library/f7f5138s(VS.71).aspx -- Rob Shearman

Nov. 20, 2007

Re: Unimplemented function KERNEL32.dll.FindFirstVolumeW

by devzero＠web.de

mhhh - since it cost me more then one hour to find out why my machine reboots and triggering a reboot by just _reading_ a file on my system, device node or not is at least very weird behaviour from a user`s perspective. >waiting for an user-space daemon to start handling of /dev/watchdog file (opening it and manipulating it from time to time to show that it's still alive. >When you cat the file, kernel recognizes it as that such an app is taking over control of the watchdog and starts it. it`s looks a little bit too simple for me. triggering a reboot just by some simple open() and read() and close() ? let`s wait what the maintainer has to tell. regards roland > Hi! > I think it's not a bug, it's a normal behaviour. > The kernel watchdog is normally not active when the machine is booted, > waiting for an user-space daemon to start handling of /dev/watchdog file > (opening it and manipulating it from time to time to show that it's still > alive). When you cat the file, kernel recognizes it as that such an app > is taking over control of the watchdog and starts it. However, because > cat closes the file immediately, it is not handled anymore and kernel > recognizes it as a userspace problem, thus rebooting the machine after > the timeout expires. > With regards, Pavel Troller > > -----Ursprüngliche Nachricht----- > Von: Pavel Troller <patrol(a)sinus.cz> > Gesendet: 20.11.07 08:49:53 > An: devzero(a)web.de > CC: Hans Leidekker <hans(a)it.vu.nl>, wine-devel(a)winehq.org > Betreff: Re: Unimplemented function KERNEL32.dll.FindFirstVolumeW > > > i`ve tracked this down to a watchdog issue. > > > > whenever i do "cat /dev/watchdog" - my system reboots after ~1 minute. > > > > i`ve contacted the maintainer of the watchdog subsystem. > > > > looks like a bug to me. > > > > thanks > > roland > > > Hi! > I think it's not a bug, it's a normal behaviour. > The kernel watchdog is normally not active when the machine is booted, > waiting for an user-space daemon to start handling of /dev/watchdog file > (opening it and manipulating it from time to time to show that it's still > alive). When you cat the file, kernel recognizes it as that such an app > is taking over control of the watchdog and starts it. However, because > cat closes the file immediately, it is not handled anymore and kernel > recognizes it as a userspace problem, thus rebooting the machine after > the timeout expires. > With regards, Pavel Troller > ______________________________________________________________________________ Jetzt neu! Im riesigen WEB.DE Club SmartDrive Dateien freigeben und mit Freunden teilen! http://www.freemail.web.de/club/smartdrive_ttc.htm/?mc=021134

Nov. 20, 2007

Re: Unimplemented function KERNEL32.dll.FindFirstVolumeW

by Pavel Troller

> i`ve tracked this down to a watchdog issue. > > whenever i do "cat /dev/watchdog" - my system reboots after ~1 minute. > > i`ve contacted the maintainer of the watchdog subsystem. > > looks like a bug to me. > > thanks > roland > Hi! I think it's not a bug, it's a normal behaviour. The kernel watchdog is normally not active when the machine is booted, waiting for an user-space daemon to start handling of /dev/watchdog file (opening it and manipulating it from time to time to show that it's still alive). When you cat the file, kernel recognizes it as that such an app is taking over control of the watchdog and starts it. However, because cat closes the file immediately, it is not handled anymore and kernel recognizes it as a userspace problem, thus rebooting the machine after the timeout expires. With regards, Pavel Troller

Nov. 20, 2007

Re: kernel32 [1/5]: Export part of UnhandledExceptionHandler as aninternal Wine function.

by Dmitry Timoshkov

"Dan Hipschman" <dsh(a)linux.ucla.edu> wrote: > This patch exports the part of UnhandledExceptionFilter that sets a page > as either writable or executable in some cases to allow continued execution > as an internal wine function. This is used in the next patch to implement > exception handling in code generated by WIDL. It needs to be able to do > this because if we can continue execution in these cases then we shouldn't > wind down the handler stack. We also shouldn't execute an "except" block > if we can continue normally. There is no need to introduce Wine-specific exception handlers, there is a sample in dlls/ntdll/tests/exception.c how to add an exception handler without Wine-specific hacks. A DIB exception handler in dlls/winex11.drv/dib.c is another one. -- Dmitry.

Nov. 20, 2007