- Wine-Devel - list.winehq.org

More overlapped I/O issues (important)
by Martin Wilck Nov. 9, 2001

Nov. 9, 2001

Hi Mike, all, I see a severe problem with the current overlapped I/O implementation, especially wrt extending it to Winsock2 overlapped I/O. The problem I see is with buffer ordering. For normal files, the file position from which to read is on the OVERLAPPED struct. This was not in the Wine implementation, the patch in my last email fixes it. For sockets and other FIFOs, this makes no sense. In this case Windows specifies that buffers are read / written **in the same order they were submitted**. This matters only if users spawn several asynchronous I/O requests at the same time, but this is allowed and actually beneficial for application performance (at least on Windows). Unless I oversee something essential, the current wine implementation does not guarantee such behaviour. Actually, since new requests are inserted at the head of the NtCurrentTeb() list, they will be considered first in subsequent asynchronous reads - i.e. the order will be reversed. If I/O requests don't complete in a single read() or write() call, the ordering will be basically random. IMO this implies that for sockets, asynchronous reads and writes must be scheduled such that only one request for a given socket is actually reading at a time, and all others are blocked until this request is finished. Again, this holds only for FIFO-type files. With the current structure, it seems to be very complex to sort this out. One could recheck the fd list in check_async and remove duplicate fds. One could also try to "lock" the fd somehow. Still it would be necessary to ascertain that the correct request gets the lock first. With this min mind, it seems more natural for me to organize the asynchronous requests in the "file" data structure itself. One would have a reader's and a writers's list there, and make sure that for FIFOS only the first list element can actually issue a read(), write(), send(), or recv() system call. For normal files all elements in the lists could be doing IO at the same time, question is if that makes sense or not. As I see it, this would mean both the readers and writers lists must be stored in the file data structure in the wine server. This would of course strongly affect the current implementation. Note that with Winsock2 overlapped I/O in place, it must be possible to use also ReadFile(), ReadFileEx() etc. on overlapped sockets, and get correct behaviour. Thus the ReadFile() etc. code must be able to deal with both sockets and ordinary files, and it makes no sense to intriduce a different approach for sockets only. Please give me your opinion on these issues ASAP. Regards 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

1 0

[PATCH:] fix overlapped I/O problems
by Martin Wilck Nov. 9, 2001

Nov. 9, 2001

Hi, The included patch (against CVS 2001-11-08) fixes two problems with overlapped I/O: 1) Overlapped ReadFile() didn't work on normal files because file_get_info() never returned FD_TYPE_OVERLAPPED. 2) According to the MS docs, the lpOverlapped argument's Offset and OffsetHigh fields contain the offset from which to read. This was not honoured by the current implementation which only did sequential reads. Regards, 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 Index: files/file.c =================================================================== RCS file: /home/wine/wine/files/file.c,v retrieving revision 1.119 diff -u -r1.119 file.c --- files/file.c 2001/10/24 00:23:25 1.119 +++ files/file.c 2001/11/09 17:35:29 @@ -1275,6 +1275,7 @@ { LPOVERLAPPED lpOverlapped = ovp->lpOverlapped; int result, r; + off_t ofs; TRACE("%p %p %08x\n", lpOverlapped, ovp->buffer, events ); @@ -1294,9 +1295,18 @@ goto async_end; } + ofs = lpOverlapped->Offset + ((off_t)lpOverlapped->OffsetHigh << 32) + + lpOverlapped->InternalHigh; + /* check to see if the data is ready (non-blocking) */ - result = read(ovp->fd, &ovp->buffer[lpOverlapped->InternalHigh], - ovp->count - lpOverlapped->InternalHigh); + /* Use the offset as given in the Overlapped structure */ + result = pread (ovp->fd, &ovp->buffer[lpOverlapped->InternalHigh], + ovp->count - lpOverlapped->InternalHigh, ofs); + + /* If pread fails, try a normal read */ + if (result == -1 && errno == ESPIPE) + result = read (ovp->fd, &ovp->buffer[lpOverlapped->InternalHigh], + ovp->count - lpOverlapped->InternalHigh); if ( (result<0) && ((errno == EAGAIN) || (errno == EINTR))) { @@ -1444,6 +1454,7 @@ { int unix_handle, result; DWORD type; + off_t ofs; TRACE("%d %p %ld %p %p\n", hFile, buffer, bytesToRead, bytesRead, overlapped ); @@ -1466,7 +1477,11 @@ } /* see if we can read some data already (this shouldn't block) */ - result = read( unix_handle, buffer, bytesToRead ); + ofs = overlapped->Offset + ((off_t)overlapped->OffsetHigh << 32); + /* Try using the user-given offset first, fall back to read otherwise */ + result = pread( unix_handle, buffer, bytesToRead, ofs); + if (result == -1 && errno == ESPIPE) + result = read( unix_handle, buffer, bytesToRead ); close(unix_handle); if(result<0) Index: server/file.c =================================================================== RCS file: /home/wine/wine/server/file.c,v retrieving revision 1.48 diff -u -r1.48 file.c --- server/file.c 2001/10/24 00:23:26 1.48 +++ server/file.c 2001/11/09 17:35:29 @@ -307,6 +307,10 @@ req->index_low = st.st_ino; req->serial = 0; /* FIXME */ } + + if (file->flags & FILE_FLAG_OVERLAPPED) + return FD_TYPE_OVERLAPPED; + return FD_TYPE_DEFAULT; }

1 0

queries on internals...
by Roger Fujii Nov. 9, 2001

Nov. 9, 2001

ok... getting reasonable output, but now I have some questions. First, it seems like wine can't find the audio codecs, though it seems to find the video ones. Is this a to-be-implemented or a bug and any clues on where to look? The next are portability / solaris questions... 1) For some unknown reason, the solaris version segfaults when it tries to load win87em.dll, but the linux one doesn't. I presume that this is trying to load the fpu emulator, but I'm not certain. In any case, what controls the loading of this in the first place? 2) regapi uses strsep, which solaris doesn't have. Is its usage in wine compatible with strtok, or should I cut/paste strsep into library/ports.c? 3) In win32/console.c, it seems that master/slave is reversed through the fork. This is only going off the docs, but if it is, how is this working under linux? That's all for now.... -r -- Roger Fujii <rmf(a)lookhere.com> Underemployed, and trying to keep it that way....

4 8

Need suggestions for audio patch testing.
by Simon Britnell Nov. 8, 2001

Nov. 8, 2001

I've posted my patch against wine cvs audio.c 1.44 to wine patches, but I've not yet been able to give it a thorough test. Anyone using freeware/shareware/opensource programs with sound under wine (not winex) please send me URLs for the programs you use (specially if they break) so I can use them for testing my patch. Thanks. __________________________________________________ Do You Yahoo!? Find a job, post your resume. http://careers.yahoo.com

1 0

Re: Two answers/questions from yesterdays and todays digests
by Roger Fujii Nov. 8, 2001

Nov. 8, 2001

Robert Lunnon <bob(a)yarrabee.net.au> wrote: > >First, it seems like wine can't find the audio codecs, > >though it seems to find the video ones. Is this a > >to-be-implemented or a bug and any clues on where to look? > > Is this in Solaris ?, yes. > if so you wiould have to have OSS installed and even > then I am not sure it would work properly. I do have OSS installed (like Sun gives you a choice as far as audio drivers go), and it doesn't. > Investigating audio on Solaris I found that ACM was not happy unless > the audio driver was available and installed properly otherwise it > returns Device Busy. I know there is a problem with 4front's devaudio emulation such that it will return an EBUSY if you open the device for READ and then WRITE (but it doesn't fail for WRITE, then READ) - they're too busy to fix the problem. > I do have a working > Waveout Solution for native Solaris based on my libaudioio project. I've not > tackled Direct Sound or Wavein yet. As yet this solution is not integrated > with Wines Build environment so you can't set it up with configure. If you > are prepared to grub through the relevant mods I can send you a patch kit. > actually, my main interest is to get the compressor(s) to work - don't care too much currently to get the audio working (mainly due to the fact that the audio drivers in solx86 is in such a sorry state). I'll be interested once I get the current CVS to work again. -r -- Roger Fujii <rmf(a)lookhere.com> Underemployed, and trying to keep it that way....

1 0

Re: Winsock2 & overlapped IO
by Mike McCormack Nov. 8, 2001

Nov. 8, 2001

Hi Martin, i don't know much about Winsock, but i can tell you how overlapped I/O is supposed to work in Wine. The bulk of the implementation is in files/file.c and scheduler/synchro.c. When ReadFile is called on a handle, the following happens: 1. check that the handle is valid and was opened with FILE_FLAG_OVERLAPPED. 2. attempt to read any data that is available immediately, and perhaps return if we read enough already. 3. queue the operation onto CurrentTeb()->pending_list When the thread waits upon any server object, all (unix) fds with pending overlapped operations are polled on. This is a minor problem in the design; if the thread never waits upon an object the overlapped can never complete, however the thread must wait upon the overlapped's hEvent to check whether it has completed... The only part of the serial comms overlapped implementation the lives in the wineserver is the calculation of overlapped timeouts. So to get winsock to do overlapped, you may have to modify the winsock code to work like that... but the winsock code uses the service thread to wait upon sockets. The task is definitely not impossible, but it might take a bit of work :-) Maybe Ove K. can give you some advice? Mike > In particular, it uses the Winsock 2 feature of "Overlapped IO" > on sockets, with user-supplied callbacks that are called when > data is received on a socket etc. > > Browsing through the Wine Winsock code, I saw that this part of > Winsock2 is obviously unimplemented. ------------------------------------------ mailto:Mike_McCormack(a)start.com.au ph +82 16 430 0425 __________________________________________________________________ Get your free Australian email account at http://www.Looksmart.com.au

3 2

Re: audio.c cleanup
by Alexandre Julliard Nov. 8, 2001

Nov. 8, 2001

Simon Britnell <ughbash(a)yahoo.com> writes: > This patch is based on winex cvs 1.14 and tidys up the > wodPlayer functionality to make it easier to follow. The WineX tree is not under the Wine license, so you cannot grab stuff from there. While I guess the Transgaming folks wouldn't object in this case, I'd prefer that you leave it to them to submit their patches back to the main tree, so that the license situation is clear. -- Alexandre Julliard julliard(a)winehq.com

2 1

wineshelllink
by Johan Gill Nov. 8, 2001

Nov. 8, 2001

Are there any known issues with wineshelllink and KDE1 at the moment? It seems quite a few of my install programs hang when the shortcuts are about to be put in place. /Johan Gill, johane(a)lysator.liu.se

1 0

Re: Extra info in scrollbar can be an handle.
by Alexandre Julliard Nov. 8, 2001

Nov. 8, 2001

Marco Bizzarri <m.bizzarri(a)icube.it> writes: > The included patch solves the problem by making an explicit GlobalLock, > and returning that value. I'm not sure this is the correct solution, so > comments are welcome. If you want to GlobalLock it, it should be allocated by GlobalAlloc. Plus the block layout will have to be compatible with what the app expects. Could you check what the app does with the memory block? Also what kind of brain-damaged app is this? -- Alexandre Julliard julliard(a)winehq.com

2 1

Two answers/questions from yesterdays and todays digests
by Robert Lunnon Nov. 8, 2001

Nov. 8, 2001

Answers Now... >Message: 1 >Date: Tue, 06 Nov 2001 12:53:56 -0500 >From: Roger Fujii <rmf(a)lookhere.com> >To: wine-devel(a)winehq.com >Subject: queries on internals.... > >ok... getting reasonable output, but now I have some >questions. > >First, it seems like wine can't find the audio codecs, >though it seems to find the video ones. Is this a >to-be-implemented or a bug and any clues on where to look? Is this in Solaris ?, if so you wiould have to have OSS installed and even then I am not sure it would work properly. Investigating audio on Solaris I found that ACM was not happy unless the audio driver was available and installed properly otherwise it returns Device Busy. I do have a working Waveout Solution for native Solaris based on my libaudioio project. I've not tackled Direct Sound or Wavein yet. As yet this solution is not integrated with Wines Build environment so you can't set it up with configure. If you are prepared to grub through the relevant mods I can send you a patch kit. > Message: 1 > Date: Wed, 7 Nov 2001 20:50:11 +0100 > From: Andreas Mohr <andi(a)rhlx01.fht-esslingen.de> > To: Francois Gouget <fgouget(a)free.fr> > Cc: Martin Wilck <Martin.Wilck(a)fujitsu-siemens.com>, > wine-devel mailing list <wine-devel(a)winehq.com> > Subject: Re: [EXPERIMENTAL PATCH]: Winsock2 > Reply-To: andi(a)rhlx01.fht-esslingen.de > > On Wed, Nov 07, 2001 at 10:54:02AM -0800, Francois Gouget wrote: > > * To answer your question: > > /* Code copied largely from WSOCK32_accept() */ > > /* FIXME: Can we simply call WSOCK32_accept() instead ? */ > > /* I wasn't sure because of the AsyncSelect stuff at the end, > > which presumably should only be called if the connection is > > accepted */ > > > > Winsock2 should not call the wsock32 library. It should be the > > opposite. One of the things on my todo list (but whether I will have > > time to actually do it is another matter) is to reorganize the winsock > > libraries so that wsock32 and winsock are together in dlls/wsock32. The > > reason is that they both implement the winsock1 API except one is 16bit > > and the other 32bits. Then in dlls/winsock we would be left with just > > ws2_32 which implements the winsock2 API. So having ws2_32 call to > > wsock32 is not going in the right direction :-) > > Hmm ? > Sorry, but this directory stuff sounds rather dainbread to me. > IMHO winsock/wsock32 should be in winsock/, and ws2_32 should be in > winsock2/ Or, maybe even better, use winsock1/ instead of winsock/. > > WTH would you want to do it in a different way ? Whoever is playing with Winsock can you please ensure that the solutions you come up with are not linux centric. There is a lot of code in there that does not run on solaris because it grubs around in PROC getting TCP structures. I have made a bit of a start in cleaning this up, Here it is, absolutely untested non working code, but a start nevertheless, I put this into a separate file and link it with winsock and wsock32 from a common source, these functions replace the helper functions of similar name in the two dll subdirectories.

1 0