April 2002 - Wine-Devel - list.winehq.org

Re: ReadFile and WriteFile in overlapped mode

by Martin Wilck

Hi Jaques, Please try the following patches and see if they change the results you see: http://www.winehq.com/hypermail/wine-patches/2002/04/0020.html http://www.winehq.com/hypermail/wine-patches/2002/04/0021.html (the second one is the relevant one in this context, but it requires the other to be applied first). Apart from that, I have a few remarks on your posting: > I have been looking at those two functions and some specs. My understanding > is that those function do not work at all in overlapped (async mode). I have done a fair bit of testing on these very recently and think they do work. On what sort of file (regular, serial port, pipe, ...) are you reading/writing ? > Currently (latest stuff in CVS), both function handle overlap mode like this This has been so for quite a while. > a) Give it a try in synchronous mode using a combination > of pread/read and pwrite/write. > > b) If not enough bytes were read or written, a request > is sent to wineserver to perform and asynchronous > operation. This request always fail. (don't know why yet). What do you mean by "fail" ? > There is two problem here. If step a succeed, no event will ever be > generated since step be is not performed. Application doing the following > sequence That's right. This is what the MSDN specs say: <quote from ReadFile specs> If hFile was opened with FILE_FLAG_OVERLAPPED and lpOverlapped is not NULL, the read operation starts at the offset specified in the OVERLAPPED structure and ReadFile may return before the read operation has been completed. In this case, ReadFile returns FALSE and the GetLastError function returns ERROR_IO_PENDING. This allows the calling process to continue while the read operation finishes. The event specified in the OVERLAPPED structure is set to the signaled state upon completion of the read operation. </quote> The function may, but doesn't have to, return before the operation completes. Whether the event is signaled upon immediate completion is unclear. But then, compare the GetOverlappedResult spec: <quote from GetOverlappedResult spec> The results reported by the GetOverlappedResult function are those of the specified handle's last overlapped operation to which the specified OVERLAPPED structure was provided, -->and for which the operation's results were pending<--. A pending operation is indicated when the function that started the operation returns FALSE, and the GetLastError function returns ERROR_IO_PENDING. When an I/O operation is pending, the function that started the operation resets the hEvent member of the OVERLAPPED structure to the nonsignaled state. Then when the pending operation has been completed, the system sets the event object to the signaled state. </quote> (-->arrows inserted by me<--). This is a bit clearer and indicates that the following code you quote: > ReadFile > .. > WaitOnSingleObject is wrong unless the return state and error code of ReadFile() are checked between the two statements - WaitForSingleObject() should only be called if ReadFile fails with ERROR_IO_PENDING. > My understanding is that step a should never be done and only step b. > Is that correct ? No, but see the patches above - things can be done differently. > Now, why step b is failing is another story. Indeed, that should be further investigated. Martin -- Martin Wilck Phone: +49 5251 8 15113 Fujitsu Siemens Computers Fax: +49 5251 8 20409 Heinz-Nixdorf-Ring 1 mailto:Martin.Wilck(a)Fujitsu-Siemens.com D-33106 Paderborn http://www.fujitsu-siemens.com/primergy

April 5, 2002

Re: ReadFile and WriteFile in overlapped mode

by Martin Wilck

On Fri, 5 Apr 2002, Tony Bryant wrote: > I fixed this a while ago, but never finished submitting the patch. The > problem is that wine uses async callbacks to signal the event in the > overlapped structure, which is a problem, because async callbacks can only > happen when WaitOnSingleObjectEx specifically allows them. This is not correct. The async IO requests are system APCs which are called if the wait status is SELECT_INTERRUPTIBLE, which is always the case with WaitForMultipleObjectsEx() and related functions (see server/thread.c and scheduler/synchro.c). Only the "Completion function" for ReadFileEx() and friends needs an alertable wait, which is perfectly in accord with the MS specs. > The solution is to > have the wine directly trigger the event without going through the async > callback system. I think the current code does things right. If you look at the MSDN specs for GetOverlappedResult(), it says clearly that you can only rely on event notification if the function that created the request (e.g. ReadFile() returned FALSE with error code ERROR_IO_PENDING. The current implemengtation does set the event when the async IO is finished, and no sooner. Martin -- Martin Wilck Phone: +49 5251 8 15113 Fujitsu Siemens Computers Fax: +49 5251 8 20409 Heinz-Nixdorf-Ring 1 mailto:Martin.Wilck(a)Fujitsu-Siemens.com D-33106 Paderborn http://www.fujitsu-siemens.com/primergy

April 5, 2002

Re: DeviceIoControl

by Eric Pouech

Laurent Pinchart a écrit : > > > > I'll add a check for the 0x10000 and 0x20000 bits, and will dispatch the > > > call to either a VxD or a device handler. Is that ok ?Can I use the same > > > VxD list as for (HIWORD(dwIoControlCode) == 0 ) ? > > > > I don't follow you, you only need the VxD list when the hiword is 0 > > I'm trying to implement support for a WDM driver (.sys). The driver registers > itself by calling IoCreateDevice with DeviceType set to 0xef00. I don't think we should support loadable WDM drivers for now. So a hard coded table, as we do for VxD should be just fine. However, the internal structures may be reviewed, so we can install (but not load) devices with the right device types (in fact checking that the device type in the io request matches the type of the device) A+

April 5, 2002

Re: Review - Winnt.h warning patch for mingw32

by Dmitry Timoshkov

"Francois Gouget" <fgouget(a)free.fr> wrote: > > +#if !defined(_stdcall) > > #define _stdcall __stdcall > > #define _fastcall __stdcall > > #define __fastcall __stdcall > > +#endif > > Is _stdcall going to be a macro? I would rather expect it to be a > reserved keyword like in Visual C++. Certainly I don't know whether _stdcall is a macro or not, but the construction below compiles correctly without warnings by both cl and gcc. #if !defined(_stdcall) #define _stdcall __stdcall #endif #define WINAPI _stdcall typedef unsigned long DWORD; extern DWORD WINAPI GetLastError(void); int main(void) { return GetLastError() != 0xdeadbeef; } -- Dmitry.

April 5, 2002

Re: SetCooperativeLevel

by Ove Kaaven

On Thu, 4 Apr 2002, Tels wrote: > ddraw/user.c has comments like these: > > /* GetDisplayMode: generic */ > /* GetFourCCCodes: generic */ > /* GetGDISurface: ??? */ > /* GetMonitorFrequency: generic */ > /* GetScanLine: generic */ > /* GetSurfaceFromDC: generic */ > /* GetVerticalBlankStatus: generic */ > /* Initialize: generic */ > /* RestoreAllSurfaces: generic */ > /* RestoreDisplayMode: generic */ > /* SetCooperativeLevel: ??? */ > > * What does the "generic" stand for? Return value? Function class? Simply that the generic implementations in ddraw/main.c are used instead of reimplementing them in ddraw/user.c. > * It seems that SetCooperativeLevel associates a "value" (or more than one) > with the givenwindow handle and TestCooperativeLevel let's you later check > this status. Where would I store such a "global" value? Somewhere in the IDirectDrawImpl structure. Most of this is already implemented in WineX, though. You can merge that code (but only the 2D parts) into main Wine if you want... I think I never got around to do it myself, we've worked in the next WineX release for far too long. > * If I understand it correctly, the directx stuff is done by transgaming. > Their licence differs from the "normal" wine. (I am really unsure > how all these things play together even after readin codeweavers.com, > winehq.com and transgaming.com) > So, if I send in a patch for that, should I send it to wine-devel-patches? > wine-devel? winex-devel? Under what license would it fall? Or does that > not matter? Each list has a "default license", I suppose... the default license for both used to be the X11 license, until WineHQ changed licenses. Any code you want to see in WineX must be submitted under the X11 license. (Submitting only under the AFPL won't work since we want to release WineX under the X11 once we have enough subscribers.) The X11 license is the license that allows sublicensing to any other license, including LGPL and AFPL, so both Wine and WineX can use it. > * Also, since I am a bit CVS challanged (newbie, I told you ;), is it okay > to add my files directly in the CVS source tree with vi(m) and then send > you a diff -u patch? Or do I need some CVS magic to do locally? Or just > keep the file outside the CVS and then send a patch? You can modify the source files and run "cvs diff -u" to diff against the repository itself, without keeping the original files around.

April 5, 2002

cool font viewer... might make good test case

by Dan Kegel

FYI... My wife is a Windows user who loves fonts. She never found good free font viewing utilities until today; now she's raving about 'font xplorer lite'. Since it's a free download, it might make a good testcase for Wine's font support. It's at http://www.moonsoftware.com/freeware.asp but I haven't tried it yet. - Dan

April 5, 2002

posting wineconf info to slashdot?

by Michael Cardenas

Now that we have the video feeds, some of the presentations, and some pictures, does anyone think we should post links to slashdot so everyone who couldn't attend can see what went on?

April 5, 2002

Crossover/WINE Test In Large Rollout

by Dave Richards

All- I am not on this list, so if you would like to respond to me you will have to reply to me directly. This is a real world situation where WINE could be used for a massive sever rollout (hundreds of users) and is having some problems in our environment. The City of Largo, Florida is a large Linux and thin client shop that has deployed a PC-less network for running out software. We are currently doing some R&D into options for running Lotus Notes in our environment. Our first choice obviously would be a native client running on Linux---it doesn't exist. We haven't looked at WINE in about 2 years and today the kind folks at CodeWeavers allowed us to test their new release with the Lotus Client. Here are the problems--> 1) Endian Bug. WINE is not able to detect that the runtime Endian order might be different than the server. Our Xterminals are RISC based and obviously WINE is running on Intel. All of the colors display incorrectly because of this. 2) Performance. On remote displays WINE is *much* slower than it is on the console. The Lotus Notes client took over 5 minutes to run remotely whereas on the console it comes up in about 20 seconds. I am assuming that it is making use of some of the XFree extensions for performance and that perhaps they have to timeout and drop through code in order to run on our Xservers. Once it is running, performance is much better, but the initial startup make it undeployable. We are thankful for everyone that has gotten WINE this far and that it works as well as it does considering the complexity of this task. If these areas are ever addressed, we would be happy to explore the product again! Dave Richards City of Largo, FL Systems Administrator

April 5, 2002

Re: ReadFile and WriteFile in overlapped mode

by Tony Bryant

On Fri, 05 Apr 2002 06:01, Jacques Gelinas wrote: > I have been looking at those two functions and some specs. My understanding > is that those function do not work at all in overlapped (async mode). > Currently (latest stuff in CVS), both function handle overlap mode like > this > > a) Give it a try in synchronous mode using a combination > of pread/read and pwrite/write. > > b) If not enough bytes were read or written, a request > is sent to wineserver to perform and asynchronous > operation. This request always fail. (don't know why yet). > > There is two problem here. If step a succeed, no event will ever be > generated since step be is not performed. Application doing the following > sequence > > ReadFile > .. > WaitOnSingleObject > > will wait forever. Some borland Builder classes do this > > My understanding is that step a should never be done and only step b. > Is that correct ? > > Now, why step b is failing is another story. > I fixed this a while ago, but never finished submitting the patch. The problem is that wine uses async callbacks to signal the event in the overlapped structure, which is a problem, because async callbacks can only happen when WaitOnSingleObjectEx specifically allows them. The solution is to have the wine directly trigger the event without going through the async callback system.

April 4, 2002

Re: ReadFile and WriteFile in overlapped mode

by Mike McCormack

Hi, > I have been looking at those two functions and some specs. > My understanding is that those function do not work at all I haven't tested the current CVS lately, but i can assure you that ReadFile and WriteFile did work at some fairly recent time :-) You may be correct that they do not work in *exactly* the right way, but for most cases they do work. > ever be generated since step be is not performed. Application > doing the following sequence > ReadFile > .. > WaitOnSingleObject > > will wait forever. Some borland Builder classes do this No, it won't because asynchronous reads can be processed during on-alertable waits. The application must call WaitForSingleObject eventually to see f the overlapped operation has finished. > My understanding is that step a should never be done and only > step b. Is that correct ? We need to check if data is pending before returning RROR_IO_PENDING because some applications do this: WaitCommEvent(hComm,&ev,&ov); WaitForSingleObject(ov.hEvent); ClearCommError(hComm,NULL,&cs) len = cs.cbInQue; if(!ReadFile(hComm,buf,len,&count,&ov)) FATAL() You're right that the way wine currently handles this is not erfect, but it will only fail if multiple overlapped operations re done at the same time. Martin Wilck is currently fixing this. > Now, why step b is failing is another story. > > Comments ? It's not perfect, and you're welcome to fix it if you want. Mike ------------------------------------------ mailto:Mike_McCormack(a)start.com.au ph +82 16 430 0425 __________________________________________________________________ Get your free Australian email account at http://www.Looksmart.com.au

April 4, 2002

Re: SetCooperativeLevel

by Tels

-----BEGIN PGP SIGNED MESSAGE----- Moin, On 04-Apr-02 Dimitrie O. Paun carved into stone: > On April 4, 2002 12:55 pm, Tels wrote: > >> wine-devel-patches? wine-devel? winex-devel? Under what license would it >> fall? Or does that not matter? > > Patches for the public tree go to wine-patches and fall under LGPL unless > you relicense them with another license. Okay. LGPL for these parts then. Fine by me. > The transgaming tree is proprietary, > and no code from it can go to the public tree unless transgaming > donates it. Okay. But that doesn't really answer my qustion, since it that case *I* am writing the code and not Transmeta (well, not sure if this matters ;). Does my code fall automatically under the transgaming licence when I patch insise their CVS tree? Can arrange that my code goes automatically public, or do I just have to hope that someday transmeta will release stuff (meaning that I work for free for transmeta, and they just hold on to my patches/work?) Thanx in advance! >> * Also, since I am a bit CVS challanged (newbie, I told you ;), is it [snip] > If you just modify some file, simply modify them in place, then do a Thank you for your explanations! Cheers, Tels - -- "Why do you go so slowly? Do you think this is some kind of game?" PGP key available on http://bloodgate.com/tels.asc or via email. perl -MDev::Bollocks -e'print Dev::Bollocks->rand(),"\n"' vitalistically harness industry-wide e-business -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl. iQEVAwUBPKzIhncLPEOTuEwVAQGsXwf+PL1QL/WwnaDdytdLROjEyCcHRB+J6N1t w2AlSOepo3AyN8VWYws9p5eZb/5zVkSaLcscmI0PruTndHKnJ/qhDoKrwFtuLjZM Ox6JfKW8Aw6eVsTRXxBP17HckETwMaUHrs7u/GYzOwfOKLBZ20i4Rh1iSuCavvEP kK9Aji/UF/ztDhLjXezPouGj7O4F9Kgz4wy29tuacFNUxbNntv4+l827HccB2i0D 9RXcfCM9Teg+f+1ebZet7iwtokkyUYdYRL3MgBItYjCbzMZM/oKOCMj+XEwpzHhn deRHKZF71un44EorpXdP8fFIIJWbTywOzScv4I5kcezaabhmulLdFQ== =xL49 -----END PGP SIGNATURE-----

April 4, 2002

Re: Review - Winnt.h warning patch for mingw32

by Francois Gouget

On Fri, 5 Apr 2002, Dmitry Timoshkov wrote: > "Steven Edwards" <Steven_Ed4153(a)yahoo.com> wrote: > > > diff -u -r1.104 winnt.h > > --- winnt.h 2 Apr 2002 19:37:15 -0000 1.104 > > +++ winnt.h 4 Apr 2002 08:15:55 -0000 > > @@ -80,9 +80,11 @@ > > #ifndef __WINE__ > > #define pascal __stdcall > > #define _pascal __stdcall > > +#if ! defined(__WINE__) && !defined(__MINGW__) > > #define _stdcall __stdcall > > #define _fastcall __stdcall > > #define __fastcall __stdcall > > +#endif > > #define __export __stdcall > > #define cdecl __cdecl > > #define _cdecl __cdecl > > Why don't you use something like this: > > +#if !defined(_stdcall) > #define _stdcall __stdcall > #define _fastcall __stdcall > #define __fastcall __stdcall > +#endif Is _stdcall going to be a macro? I would rather expect it to be a reserved keyword like in Visual C++. Steven's patch looks good to me. -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ Advice is what we ask for when we already know the answer but wish we didn't -- Eric Jong

April 4, 2002

Re: [PATCH] async IO APCs call check_async_list() directly

by Mike McCormack

Hi Martin, I think this is a better way to do things, but it might be better to submit this patch against the current CVS rather than against CVS+previous patch. Mike Original message from: Martin Wilck <Martin.Wilck(a)fujitsu-siemens.com> > > >PATCH: async-func.diff > >This is an improvement over my previous patch (async-struct.diff) >in that it does not need the static function check_async_lists >in several source files anymore and still achieves DLL separation. > >The idea is that all async I/O requests use check_async_list as callback function >anyway, so that there is no need to pass a function pointer. >I discussed the idea with Mike, and he seemed to like it. > >Patch against: Wine CVS 2002/04/04, with my async-struct patch applied. >Test status: Compiles warnings. > Tested regular file async IO (ok). > Serial port async IO: a test would be desirable. > ------------------------------------------ mailto:Mike_McCormack(a)start.com.au ph +82 16 430 0425 __________________________________________________________________ Get your free Australian email account at http://www.Looksmart.com.au

April 4, 2002

Re: [PATCH] async IO APCs call check_async_list() directly

by Mike McCormack

Hi Martin, I think this is a better way to do things, but it might be better to submit this patch against the current CVS rather than against CVS+previous patch. Mike Original message from: Martin Wilck <Martin.Wilck(a)fujitsu-siemens.com> > > >PATCH: async-func.diff > >This is an improvement over my previous patch (async-struct.diff) >in that it does not need the static function check_async_lists >in several source files anymore and still achieves DLL separation. > >The idea is that all async I/O requests use check_async_list as callback function >anyway, so that there is no need to pass a function pointer. >I discussed the idea with Mike, and he seemed to like it. > >Patch against: Wine CVS 2002/04/04, with my async-struct patch applied. >Test status: Compiles warnings. > Tested regular file async IO (ok). > Serial port async IO: a test would be desirable. > ------------------------------------------ mailto:Mike_McCormack(a)start.com.au ph +82 16 430 0425 __________________________________________________________________ Get your free Australian email account at http://www.Looksmart.com.au

April 4, 2002

Re: DeviceIoControl

by Laurent Pinchart

> > I'll add a check for the 0x10000 and 0x20000 bits, and will dispatch the > > call to either a VxD or a device handler. Is that ok ?Can I use the same > > VxD list as for (HIWORD(dwIoControlCode) == 0 ) ? > > I don't follow you, you only need the VxD list when the hiword is 0 I'm trying to implement support for a WDM driver (.sys). The driver registers itself by calling IoCreateDevice with DeviceType set to 0xef00. The user program calls CreateFile( "\\\\.\\MyDriver", ... ) and uses the returned handle in the call to DeviceIoControl, with dwIoControlCode set to 0xef002407. The dwIoControlCode higword is not 0. Shouldn't we have a handler list, similar to the VxD list in this case ? Laurent Pinchart

April 4, 2002

RE: posting wineconf info to slashdot?

by Steven Edwards

When is the reactos presentation going to be put up? I know it wasn't the greatest in the world but we were on the guest list. =P > -----Original Message----- > From: wine-devel-admin(a)winehq.com > [mailto:wine-devel-admin(a)winehq.com] On Behalf Of Michael Cardenas > Sent: Thursday, April 04, 2002 8:27 PM > To: wine-devel(a)winehq.com > Subject: posting wineconf info to slashdot? > > > Now that we have the video feeds, some of the presentations, and some > pictures, does anyone think we should post links to slashdot > so everyone > who couldn't attend can see what went on? > > > > > _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com

April 4, 2002

Re: Startup fix for scanbin 6.0

by Dimitrie O. Paun

On April 4, 2002 01:48 pm, Stefan Leichter wrote: > i thought the NULL - pointer is the result of a missing feature or a bug in > wine. For me i does not make sense to call the functions DrawText* without > a text. Nevertheless here is the patch without the ERR. While it is true that some Wine code call DrawText, so does a lot of user programs... -- Dimi.

April 4, 2002

Re: DeviceIoControl

by Eric Pouech

Laurent Pinchart a écrit : > >the control on "device" handles is done when the high word of the > >control code is non 0. > > > >as of today (but it's a bit hacky) handles to "device" and VxD:s are > You're refering to both VxDs and WDM (.sys) drivers, right ? just VxDs > How does the server find the correct low word of the client ID ? The > driver I want to implement has a device type of 0xef00 (value passed to > IoCreateDevice). I added a field to the VxD table with id set to 0xef00. > Does wine, when I call CreateFile( "\\\\.\\MyDevice", ... ), check that > table and set the client ID low word to the proper value ? this is poorly supported at the moment. only a few of the proper \\.\ syntax is handled (see files/file.c for the details) > > >If the 0x20000 bit is set, it's in fact a handle a device. Driver A: > >through Z: are mapped to 0 to 25 values (in the low word) > > > >Otherwise, it's a handle to the old DOS named devices (with their old > >IDs too) > > > I think we should handler those cases in a new device handler (which > would just test if the device is a CDROM for now). all of this needs to be revisited, but this is not a simple task > >So, back to DeviceIoControl when the high word of the io control code is > >0. We expect device handle in that case (but don't test the 0x20000 bit for > >error reporting), and then try to check if the device (from it's driver > >letter) is a CDROM, and if so apply the io control on it. Support of io > > > >control on other types of device (hard disks, storage...) isn't done at the moment. > > > You mean when the high word is not 0, right ? yes > I'll add a check for the 0x10000 and 0x20000 bits, and will dispatch the > call to either a VxD or a device handler. Is that ok ?Can I use the same > VxD list as for (HIWORD(dwIoControlCode) == 0 ) ? I don't follow you, you only need the VxD list when the hiword is 0 A+

April 4, 2002

Re: SetCooperativeLevel

by Dimitrie O. Paun

On April 4, 2002 12:55 pm, Tels wrote: > wine-devel-patches? wine-devel? winex-devel? Under what license would it > fall? Or does that not matter? Patches for the public tree go to wine-patches and fall under LGPL unless you relicense them with another license. The transgaming tree is proprietary, and no code from it can go to the public tree unless transgaming donates it. > * Also, since I am a bit CVS challanged (newbie, I told you ;), is it okay > to add my files directly in the CVS source tree with vi(m) and then send > you a diff -u patch? Or do I need some CVS magic to do locally? Or just > keep the file outside the CVS and then send a patch? If you just modify some file, simply modify them in place, then do a cvs diff -u > mypatch.diff If you also add some files, add them directly with vi, then do a diff -uN path/to/your/new/file >> mypatch.diff Then send mypatch.diff to wine-patches with an appropriate ChangeLog entry. -- Dimi.

April 4, 2002

Re: posting wineconf info to slashdot?

by Dimitrie O. Paun

On April 4, 2002 08:26 pm, Michael Cardenas wrote: > Now that we have the video feeds, some of the presentations, and some > pictures, does anyone think we should post links to slashdot so everyone > who couldn't attend can see what went on? I'm all for it. Before we do, it seems like we have 3 group photos, but only one is annotated. We should annotate them all, and post them together, this way people that are not in one, are in the other one... -- Dimi.

April 4, 2002

ReadFile and WriteFile in overlapped mode

by Jacques Gelinas

I have been looking at those two functions and some specs. My understanding is that those function do not work at all in overlapped (async mode). Currently (latest stuff in CVS), both function handle overlap mode like this a) Give it a try in synchronous mode using a combination of pread/read and pwrite/write. b) If not enough bytes were read or written, a request is sent to wineserver to perform and asynchronous operation. This request always fail. (don't know why yet). There is two problem here. If step a succeed, no event will ever be generated since step be is not performed. Application doing the following sequence ReadFile .. WaitOnSingleObject will wait forever. Some borland Builder classes do this My understanding is that step a should never be done and only step b. Is that correct ? Now, why step b is failing is another story. Comments ? --------------------------------------------------------- Jacques Gelinas <jack(a)solucorp.qc.ca> vserver: run general purpose virtual servers on one box, full speed! http://www.solucorp.qc.ca/miscprj/s_context.hc

April 4, 2002

SetCooperativeLevel

by Tels

-----BEGIN PGP SIGNED MESSAGE----- Moin, [crosspost to both wine lists] I am preparing a patch for DirectDraw::SetCooperativeLevel and friends: Some questions from a newbie, so please bear with me: ddraw/user.c has comments like these: /* GetDisplayMode: generic */ /* GetFourCCCodes: generic */ /* GetGDISurface: ??? */ /* GetMonitorFrequency: generic */ /* GetScanLine: generic */ /* GetSurfaceFromDC: generic */ /* GetVerticalBlankStatus: generic */ /* Initialize: generic */ /* RestoreAllSurfaces: generic */ /* RestoreDisplayMode: generic */ /* SetCooperativeLevel: ??? */ * What does the "generic" stand for? Return value? Function class? * It seems that SetCooperativeLevel associates a "value" (or more than one) with the givenwindow handle and TestCooperativeLevel let's you later check this status. Where would I store such a "global" value? * If I understand it correctly, the directx stuff is done by transgaming. Their licence differs from the "normal" wine. (I am really unsure how all these things play together even after readin codeweavers.com, winehq.com and transgaming.com) So, if I send in a patch for that, should I send it to wine-devel-patches? wine-devel? winex-devel? Under what license would it fall? Or does that not matter? I have no idea what hidden snatches are in the transmeta licence, so I would like to use the LGPL. ;) Is there a problem with that for the transmeta guys or for you wine guys? Or would it fall automatically under the transmeta licence, since I am patching their source anyway? (You can tell I am confused ;) * Also, since I am a bit CVS challanged (newbie, I told you ;), is it okay to add my files directly in the CVS source tree with vi(m) and then send you a diff -u patch? Or do I need some CVS magic to do locally? Or just keep the file outside the CVS and then send a patch? Thanx in advance, Te"Wine, I am coooominng!"ls - -- "Why do you go so slowly? Do you think this is some kind of game?" PGP key available on http://bloodgate.com/tels.asc or via email. perl -MDev::Bollocks -e'print Dev::Bollocks->rand(),"\n"' conveniently consolidate out-of-the-box ROI -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl. iQEVAwUBPKyTS3cLPEOTuEwVAQElUgf/Sjq2aXSUmuXz7W1FDBaAZQvNjEHovAB0 uZty72+LFRToxiGGvwU85X02Ug7Yc3zdT4R1cqtpGjjuGsXnBoE2h33Nz9rtDPhq wvZEIdxqv+MU3yVXWo5yRjSZZKpsDP78hiHoIZ+pCwtapqsZWdAOMIPwgVwNFwNv gJaG2OEtk9arlDNcN0uibzbq0lONh08BbqT+i1DvYYDRHJw6pNnUNKslV6b4sZdG bGjH0td8coj2vSM0RdRHaTYdjlY9BpJclqJI/7YGtS9cn3HCrWgAhCHwXLoLOL76 GcSfx/35eFrvmZM8LuJlXpK+abo2J4/+24CpuDEwwaI6Tqyuu+PS6A== =8o1v -----END PGP SIGNATURE-----

April 4, 2002

[PATCH] async IO APCs call check_async_list() directly

by Martin Wilck

PATCH: async-func.diff This is an improvement over my previous patch (async-struct.diff) in that it does not need the static function check_async_lists in several source files anymore and still achieves DLL separation. The idea is that all async I/O requests use check_async_list as callback function anyway, so that there is no need to pass a function pointer. I discussed the idea with Mike, and he seemed to like it. Patch against: Wine CVS 2002/04/04, with my async-struct patch applied. Test status: Compiles warnings. Tested regular file async IO (ok). Serial port async IO: a test would be desirable. Modified files: include : async.h (remove check_async_list) server : protocol.def (add APC_ASYNC_IO APC type, remove func field for create_async) async.h (remove func field in async struct) async.c (remove references to func field) thread.c (allow NULL function pointer for APC_ASYNC_IO) scheduler : synchro.c (add check_async_list; call from call_apcs for APC_ASYNC_IO) -- Martin diff -ruNX ignore MW/wine/include/async.h TMP/wine/include/async.h --- MW/wine/include/async.h Thu Apr 4 18:07:36 2002 +++ TMP/wine/include/async.h Thu Apr 4 18:08:56 2002 @@ -56,8 +56,6 @@ /* All functions declared static for Dll separation purposes */ -static void WINAPI check_async_list(async_private *asp, DWORD status); - inline static void finish_async( async_private *ovp ) { if(ovp->prev) @@ -87,7 +85,6 @@ req->overlapped = ovp; req->type = ovp->type; req->count = ovp->ops->get_count( ovp ); - req->func = check_async_list; req->status = status; ret = wine_server_call( req ); } @@ -115,29 +112,6 @@ NtCurrentTeb()->pending_list = ovp; return __register_async( ovp, status ); -} - -static void WINAPI check_async_list(async_private *asp, DWORD status) -{ - async_private *ovp; - DWORD ovp_status; - - for( ovp = NtCurrentTeb()->pending_list; ovp && ovp != asp; ovp = ovp->next ); - - if(!ovp) - return; - - if( status != STATUS_ALERTED ) - { - ovp_status = status; - ovp->ops->set_status (ovp, status); - } - else ovp_status = ovp->ops->get_status (ovp); - - if( ovp_status == STATUS_PENDING ) ovp->func( ovp ); /* ovp->func may change status */ - - /* This will destroy all but PENDING requests */ - register_old_async( ovp ); } #endif /* __WINE_ASYNC_H */ diff -ruNX ignore MW/wine/include/wine/server_protocol.h TMP/wine/include/wine/server_protocol.h --- MW/wine/include/wine/server_protocol.h Thu Apr 4 18:07:36 2002 +++ TMP/wine/include/wine/server_protocol.h Thu Apr 4 17:49:50 2002 @@ -502,7 +502,7 @@ int type; /* VARARG(args,ptrs); */ }; -enum apc_type { APC_NONE, APC_USER, APC_TIMER, APC_ASYNC }; +enum apc_type { APC_NONE, APC_USER, APC_TIMER, APC_ASYNC, APC_ASYNC_IO }; @@ -2278,7 +2278,6 @@ { struct request_header __header; handle_t handle; - void* func; int type; void* overlapped; int count; @@ -3200,6 +3199,6 @@ struct get_window_properties_reply get_window_properties_reply; }; -#define SERVER_PROTOCOL_VERSION 78 +#define SERVER_PROTOCOL_VERSION 81 #endif /* __WINE_WINE_SERVER_PROTOCOL_H */ diff -ruNX ignore MW/wine/scheduler/synchro.c TMP/wine/scheduler/synchro.c --- MW/wine/scheduler/synchro.c Thu Apr 4 18:07:36 2002 +++ TMP/wine/scheduler/synchro.c Thu Apr 4 17:54:20 2002 @@ -30,7 +30,7 @@ #include "thread.h" #include "winerror.h" #include "wine/server.h" - +#include "async.h" /*********************************************************************** * get_timeout @@ -90,6 +90,34 @@ /*********************************************************************** + * check_async_list + * + * Handle async I/O requests. + */ +static void WINAPI check_async_list(async_private *asp, DWORD status) +{ + async_private *ovp; + DWORD ovp_status; + + for( ovp = NtCurrentTeb()->pending_list; ovp && ovp != asp; ovp = ovp->next ); + + if(!ovp) + return; + + if( status != STATUS_ALERTED ) + { + ovp_status = status; + ovp->ops->set_status (ovp, status); + } + else ovp_status = ovp->ops->get_status (ovp); + + if( ovp_status == STATUS_PENDING ) ovp->func( ovp ); + + /* This will destroy all but PENDING requests */ + register_old_async( ovp ); +} + +/*********************************************************************** * call_apcs * * Call outstanding APCs. @@ -129,6 +157,9 @@ /* convert sec/usec to NT time */ DOSFS_UnixTimeToFileTime( (time_t)args[0], &ft, (DWORD)args[1] * 10 ); proc( args[2], ft.dwLowDateTime, ft.dwHighDateTime ); + break; + case APC_ASYNC_IO: + check_async_list ( args[0], (DWORD) args[1]); break; default: server_protocol_error( "get_apc_request: bad type %d\n", type ); diff -ruNX ignore MW/wine/server/async.c TMP/wine/server/async.c --- MW/wine/server/async.c Thu Apr 4 18:07:36 2002 +++ TMP/wine/server/async.c Thu Apr 4 17:50:45 2002 @@ -64,7 +64,7 @@ { /* fprintf(stderr,"notifying %p!\n",async->overlapped); */ async->status = status; - thread_queue_apc(async->thread, NULL, async->func, APC_ASYNC, 1, 2, async->overlapped, status); + thread_queue_apc(async->thread, NULL, NULL, APC_ASYNC_IO, 1, 2, async->overlapped, status); } void destroy_async_queue( struct async_queue *q ) @@ -116,7 +116,7 @@ destroy_async(async); } -struct async *create_async(struct object *obj, struct thread *thread, void *func, +struct async *create_async(struct object *obj, struct thread *thread, void *overlapped) { struct async *async = (struct async *) malloc(sizeof(struct async)); @@ -128,7 +128,6 @@ async->obj = obj; async->thread = thread; - async->func = func; async->overlapped = overlapped; async->next = NULL; async->prev = NULL; @@ -165,7 +164,7 @@ if(req->status==STATUS_PENDING) { if(!async) - async = create_async(obj, current, req->func, req->overlapped); + async = create_async(obj, current, req->overlapped); if(async) { diff -ruNX ignore MW/wine/server/async.h TMP/wine/server/async.h --- MW/wine/server/async.h Tue Apr 2 16:51:50 2002 +++ TMP/wine/server/async.h Thu Apr 4 17:50:57 2002 @@ -30,7 +30,6 @@ { struct object *obj; struct thread *thread; - void *func; void *overlapped; unsigned int status; struct timeval when; @@ -51,7 +50,7 @@ struct async *find_async(struct async_queue *q, struct thread *thread, void *overlapped); void async_insert(struct async_queue *q, struct async *async); struct async *create_async(struct object *obj, struct thread *thread, - void *func, void *overlapped); + void *overlapped); void async_add_timeout(struct async *async, int timeout); static inline void init_async_queue(struct async_queue *q) { diff -ruNX ignore MW/wine/server/protocol.def TMP/wine/server/protocol.def --- MW/wine/server/protocol.def Thu Apr 4 18:07:36 2002 +++ TMP/wine/server/protocol.def Thu Apr 4 17:49:45 2002 @@ -412,7 +412,7 @@ int type; /* function type */ VARARG(args,ptrs); /* function arguments */ @END -enum apc_type { APC_NONE, APC_USER, APC_TIMER, APC_ASYNC }; +enum apc_type { APC_NONE, APC_USER, APC_TIMER, APC_ASYNC, APC_ASYNC_IO }; /* Close a handle for the current process */ @@ -1618,7 +1618,6 @@ /* Create / reschedule an async I/O */ @REQ(register_async) handle_t handle; /* handle to comm port, socket or file */ - void* func; int type; void* overlapped; int count; diff -ruNX ignore MW/wine/server/thread.c TMP/wine/server/thread.c --- MW/wine/server/thread.c Tue Apr 2 16:51:51 2002 +++ TMP/wine/server/thread.c Thu Apr 4 17:37:22 2002 @@ -959,8 +959,9 @@ } /* Optimization: ignore APCs that have a NULL func; they are only used * to wake up a thread, but since we got here the thread woke up already. + * Exception: for APC_ASYNC_IO, func == NULL is legal. */ - if (apc->func) break; + if (apc->func || apc->type == APC_ASYNC_IO) break; free( apc ); } size = apc->nb_args * sizeof(apc->args[0]); diff -ruNX ignore MW/wine/server/trace.c TMP/wine/server/trace.c --- MW/wine/server/trace.c Thu Apr 4 18:07:36 2002 +++ TMP/wine/server/trace.c Thu Apr 4 17:49:50 2002 @@ -1846,7 +1846,6 @@ static void dump_register_async_request( const struct register_async_request *req ) { fprintf( stderr, " handle=%d,", req->handle ); - fprintf( stderr, " func=%p,", req->func ); fprintf( stderr, " type=%d,", req->type ); fprintf( stderr, " overlapped=%p,", req->overlapped ); fprintf( stderr, " count=%d,", req->count );

April 4, 2002

Re: Review - Winnt.h warning patch for mingw32

by Dmitry Timoshkov

"Steven Edwards" <Steven_Ed4153(a)yahoo.com> wrote: > diff -u -r1.104 winnt.h > --- winnt.h 2 Apr 2002 19:37:15 -0000 1.104 > +++ winnt.h 4 Apr 2002 08:15:55 -0000 > @@ -80,9 +80,11 @@ > #ifndef __WINE__ > #define pascal __stdcall > #define _pascal __stdcall > +#if ! defined(__WINE__) && !defined(__MINGW__) > #define _stdcall __stdcall > #define _fastcall __stdcall > #define __fastcall __stdcall > +#endif > #define __export __stdcall > #define cdecl __cdecl > #define _cdecl __cdecl Why don't you use something like this: +#if !defined(_stdcall) #define _stdcall __stdcall #define _fastcall __stdcall #define __fastcall __stdcall +#endif -- Dmitry.

April 4, 2002

Re: freetype.c:1852: conflicting types for `WineEngGetTextExtentPointI'

by Martin Wilck

This one-liner fixes it: --- dlls/gdi/freetype.c~ Thu Apr 4 09:17:14 2002 +++ dlls/gdi/freetype.c Thu Apr 4 09:30:48 2002 @@ -1847,7 +1847,7 @@ return FALSE; } -BOOL WineEngGetTextExtentPointI(GdiFont font, LPWORD indices, INT count, +BOOL WineEngGetTextExtentPointI(GdiFont font, const WORD *indices, INT count, LPSIZE size) { ERR("called but we don't have FreeType\n"); -- Martin Wilck Phone: +49 5251 8 15113 Fujitsu Siemens Computers Fax: +49 5251 8 20409 Heinz-Nixdorf-Ring 1 mailto:Martin.Wilck(a)Fujitsu-Siemens.com D-33106 Paderborn http://www.fujitsu-siemens.com/primergy

April 4, 2002