November 2001 - Wine-Devel - list.winehq.org

Re: fix for ugly things...
by Francois Gouget Nov. 9, 2001

Nov. 9, 2001

On Fri, 9 Nov 2001, Andreas Mohr wrote: > [ok, mailed for the *3rd* time now, since WineHQ has been out of space *again*] > > Hi all, > > - fix potentially problematic stuff (see Apple installer desaster, e.g. on /.) [...] That's good but while you're at it I think you should fix the other things like: Index: documentation/db2html-winehq [...] cat $TMPDIR/* Index: documentation/make_winehq [...] mv $WWWDIR $WWWDIR.old [...] mkdir $WWWDIR Otherwise the scripts will still not work and make a mess (even if data loss is less likely, unless we also have 'cmd... >$xxx'). -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ RFC 2549: ftp://ftp.isi.edu/in-notes/rfc2549.txt IP over Avian Carriers with Quality of Service

1 0

Re: OSS audio race fix
by eric pouech Nov. 9, 2001

Nov. 9, 2001

eric pouech wrote: > > as reported by Simon Brunnel oops... I meant Britnell (sorry for the misspell Simon) A+ -- --------------- Eric Pouech (http://perso.wanadoo.fr/eric.pouech/) "The future will be better tomorrow", Vice President Dan Quayle

1 0

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. 9, 2001

Nov. 9, 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

Re: File record locking (exclusive, nowait)
by eric pouech Nov. 8, 2001

Nov. 8, 2001

"Medland, Bill" wrote: > > <<diff29.txt>> didn't you forget to add the server/fd.c file ? A+ -- --------------- Eric Pouech (http://perso.wanadoo.fr/eric.pouech/) "The future will be better tomorrow", Vice President Dan Quayle

1 0

linking in .lib files
by Chris Dawson Nov. 8, 2001

Nov. 8, 2001

Is it possible to link in .lib files using winelib and winemaker? If possible, what switch do I give winemaker? My application when compiled under Windows linked in a .lib file and I would rather not modify the source code. I've tried using the -l switch, but this is obviously only for *nix libraries even when compiling with winelib. Or, other suggestions to resolve all the undefined reference errors? Thanks in advance, Chris

3 4

[PROBLEM REPORT] Focus problem for mdi children
by Gerard Patel Nov. 8, 2001

Nov. 8, 2001

Mdi child window activation is broken since ChangeSet ID: 10050857000162942344322801 CVSROOT: /opt/cvs-commit Module name: wine Changes by: julliard(a)wine2. 01/11/06 17:28:20 Modified files: dlls/ttydrv : wnd.c dlls/x11drv : winpos.c include : winpos.h windows : message.c winpos.c Log message: Window activation cleanups. The change that broke it is : --- /opt/cvs-commit/wine/dlls/x11drv/winpos.c 2001/10/22 19:08:34 1.31 +++ /opt/cvs-commit/wine/dlls/x11drv/winpos.c 2001/11/06 22:28:19 1.32 @@ -955,8 +955,7 @@ ShowCaret(winpos->hwnd); } - if (!(winpos->flags & SWP_NOACTIVATE)) - WINPOS_ChangeActiveWindow( winpos->hwnd, FALSE ); + if (!(winpos->flags & SWP_NOACTIVATE)) SetActiveWindow( winpos->hwnd ); /* And last, send the WM_WINDOWPOSCHANGED message */ Cause of the problem : the WM_CHILDACTIVATE message is not sent anymore, so the mdi children are not activated (caption does not change color, more annoyingly the active child is not set). The problem can be seen with Eudora or ForteFreeAgent. Using this patch : --- winpos.c.orig Wed Nov 7 10:41:32 2001 +++ winpos.c Thu Nov 8 00:44:29 2001 @@ -955,7 +955,20 @@ ShowCaret(winpos->hwnd); } - if (!(winpos->flags & SWP_NOACTIVATE)) SetActiveWindow( winpos->hwnd ); + if (!(winpos->flags & SWP_NOACTIVATE)) + { + SetActiveWindow( winpos->hwnd ); + if (!(wndPtr = WIN_FindWndPtr(winpos->hwnd))) return FALSE; + + /* child windows get WM_CHILDACTIVATE message */ + if( (wndPtr->dwStyle & (WS_CHILD | WS_POPUP)) == WS_CHILD ) + { + WIN_ReleaseWndPtr(wndPtr); + SendMessageA(winpos->hwnd, WM_CHILDACTIVATE, 0, 0L); + } + else + WIN_ReleaseWndPtr(wndPtr); + } /* And last, send the WM_WINDOWPOSCHANGED message */ seems to fix the problem. It may be too simple though :-) Gerard

1 0

Re: [PROBLEM REPORT] Focus problem for mdi children
by Alexandre Julliard Nov. 8, 2001

Nov. 8, 2001

Gerard Patel <gerard.patel(a)nerim.net> writes: > Using this patch : > --- winpos.c.orig Wed Nov 7 10:41:32 2001 > +++ winpos.c Thu Nov 8 00:44:29 2001 > @@ -955,7 +955,20 @@ > ShowCaret(winpos->hwnd); > } > > - if (!(winpos->flags & SWP_NOACTIVATE)) SetActiveWindow( winpos->hwnd ); > + if (!(winpos->flags & SWP_NOACTIVATE)) > + { > + SetActiveWindow( winpos->hwnd ); > + if (!(wndPtr = WIN_FindWndPtr(winpos->hwnd))) return FALSE; > + > + /* child windows get WM_CHILDACTIVATE message */ > + if( (wndPtr->dwStyle & (WS_CHILD | WS_POPUP)) == WS_CHILD ) > + { > + WIN_ReleaseWndPtr(wndPtr); > + SendMessageA(winpos->hwnd, WM_CHILDACTIVATE, 0, 0L); > + } > + else > + WIN_ReleaseWndPtr(wndPtr); > + } > > /* And last, send the WM_WINDOWPOSCHANGED message */ > > > seems to fix the problem. It may be too simple though :-) No actually it's pretty much the right thing. I have applied a slightly simplified version of your patch, thanks for the report. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: RtlNtStatusToDosError issue
by Alexandre Julliard Nov. 7, 2001

Nov. 7, 2001

"Bill Medland" <medbi01_1(a)accpac.com> writes: > What is the correct solution? To choose a status at random that translates > to ERROR_LOCK_VIOLATION? Yes. (but it's OK to pick one that makes some sense ;-) -- Alexandre Julliard julliard(a)winehq.com

1 0

RtlNtStatusToDosError issue
by Bill Medland Nov. 7, 2001

Nov. 7, 2001

Can I solve this through documentation or do I have to use debug? I am implementing the LockFile stuff based on what Alexandre did to the CorelWine last year. Currently I have it returning error codes such as ERROR_LOCK_VIOLATION. However that causes RtlNtStatusToDosError to complain about not recognising it (no big surprise) during the server call. What is the correct solution? To choose a status at random that translates to ERROR_LOCK_VIOLATION? Bill

1 0

What's happened to www.winehq.com/ident
by Bill Medland Nov. 7, 2001

Nov. 7, 2001

Anyone know what has happened to the LXR stuff? I used to use it to find where files are implemented but now the symbols don't seem to be indexed any longer. Bill

1 0

[EXPERIMENTAL PATCH]: Winsock2
by Martin Wilck Nov. 7, 2001

Nov. 7, 2001

Hi, this patch tries to implement a small part of the extended Winsock 2 functionality in Wine. It is very experimental, I have only made sure everything compiles. It is _NOT_ intended for people to try and run. I am submitting this patch because I am pretty new to Wine hacking and I'd appreciate some Wine gurus to audit this for obvious stupidity, style incompatiblities, etc. Features implemented: - WSAAccept(), including callback functionality. - WSASocket(): support for dwFlags parameter. The real difficult stuff will follow when I start looking into WSARecv(), WSASend(), and WSAGetOverlappedResults(). Thus it'd be nice if someone pointed out my misconceptions before I put real work into the latter. Thanks & 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: dlls/winsock/socket.c =================================================================== RCS file: /home/wine/wine/dlls/winsock/socket.c,v retrieving revision 1.65 diff -u -r1.65 socket.c --- dlls/winsock/socket.c 2001/10/14 16:25:47 1.65 +++ dlls/winsock/socket.c 2001/11/07 16:36:53 @@ -154,6 +154,7 @@ int WS_dup_pe(struct protoent* p_pe, int flag); int WS_dup_se(struct servent* p_se, int flag); int WINAPI WSOCK32_getpeername(SOCKET s, ws_sockaddr *name, int *namelen); +int WINAPI WSOCK32_getsockname(SOCKET s, ws_sockaddr *name, int *namelen); typedef void WIN_hostent; typedef void WIN_protoent; @@ -938,6 +939,131 @@ } /*********************************************************************** + * WSAAccept (WS2_32.26) + */ +SOCKET WINAPI WSAAccept(SOCKET s, + ws_sockaddr* addr, + LPINT addrlen32, + LPCONDITIONPROC cond_cb, + DWORD cb_data ) +{ + + /* 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 */ + + SOCKET as; + int fd = _get_sock_fd(s); + unsigned omask; + + TRACE("%04x cond=%08x data=%08x\n", (UINT16) s, (UINT) cond_cb, (UINT) cb_data); + + if (fd == -1) + { + SetLastError(WSAENOTSOCK); + return INVALID_SOCKET; + } + + if (_is_blocking(s)) + { + /* block here */ + do_block(fd, 5); + _sync_sock_state(s); /* let wineserver notice connection */ + /* retrieve any error codes from it */ + SetLastError(_get_sock_error(s, FD_ACCEPT_BIT)); + /* FIXME: care about the error? */ + } + + close(fd); + + SERVER_START_REQ( accept_socket ) + { + req->lhandle = s; + req->access = GENERIC_READ|GENERIC_WRITE|SYNCHRONIZE; + req->inherit = TRUE; + set_error( SERVER_CALL() ); + as = (SOCKET)req->handle; + } + SERVER_END_REQ; + + if (!as) + { + SetLastError(WSAENOTSOCK); + return INVALID_SOCKET; + } + + if (cond_cb) + { + /* We need our own buffers for Caller and Callee addresses. */ + /* FIXME: large enough / too large ? */ + const int SOCKADDR_LEN=128; + INT caller_len = SOCKADDR_LEN, callee_len = SOCKADDR_LEN; + char _caller_addr[SOCKADDR_LEN], _callee_addr[SOCKADDR_LEN]; + ws_sockaddr *caller_addr = (ws_sockaddr*) _caller_addr; + ws_sockaddr *callee_addr = (ws_sockaddr*) _callee_addr; + WSABUF caller_buf = { 0, _caller_addr }; + WSABUF callee_buf = { 0, _callee_addr }; + WSABUF callee_data_buf = { 0, NULL }; + GROUP group = 0; + int cb_result; + + WSOCK32_getpeername (as, caller_addr, &caller_len); + WSOCK32_getsockname (as, callee_addr, &callee_len); + caller_buf.len = caller_len; + callee_buf.len = callee_len; + + TRACE ("Calling callback\n"); + cb_result = cond_cb (&caller_buf, /* Caller ID */ + NULL, /* Caller data unsupported */ + NULL, NULL, /* QOS unsupported */ + &callee_buf, /* Callee id */ + &callee_data_buf, /* Callee data unsupported */ + &group, /* Group unsupported */ + cb_data); + + switch (cb_result) + { + case CF_ACCEPT: + TRACE ("Connection accepted\n"); + break; + case CF_REJECT: + closesocket (as); + TRACE ("Connection refused\n"); + SetLastError (WSAECONNREFUSED); + return INVALID_SOCKET; + case CF_DEFER: + TRACE ("Connection deferred\n"); + SERVER_START_REQ (set_socket_deferred) + { + req->handle = s; + req->deferred = as; + SERVER_CALL (); + } + SERVER_END_REQ; + SetLastError (WSATRY_AGAIN); + return INVALID_SOCKET; + default: + WARN ("Invalid return value\n"); + SetLastError (WSAEINVAL); + return INVALID_SOCKET; + } + + if (addrlen32 && addr && *addrlen32 >= caller_len) + memcpy (addr, caller_addr, caller_len); + + } else + WSOCK32_getpeername (as, addr, addrlen32); + + omask = _get_sock_mask( s ); + if (omask & FD_WINE_SERVEVENT) + ws2_async_accept(s, as); + + return as; +} + + +/*********************************************************************** * bind (WS2_32.2) */ int WINAPI WSOCK32_bind(SOCKET s, const ws_sockaddr* name, int namelen) @@ -1016,7 +1142,21 @@ */ INT WINAPI WSOCK32_closesocket(SOCKET s) { + SOCKET deferred; TRACE("socket %08x\n", s); + + /* Close a deferred socket if present */ + SERVER_START_REQ ( get_socket_deferred ) + { + req->handle = s; + SERVER_CALL (); + deferred = req->deferred; + } + SERVER_END_REQ; + + if (deferred) + WSOCK32_closesocket (deferred); + if (CloseHandle(s)) return 0; return SOCKET_ERROR; } @@ -2898,10 +3038,24 @@ g, dwFlags) are ignored. */ - TRACE("af=%d type=%d protocol=%d protocol_info=%p group=%d flags=0x%lx\n", + SOCKET s; + + TRACE("af=%d type=%d protocol=%d protocol_info=%p group=%d flags=0x%lx\n", af, type, protocol, lpProtocolInfo, g, dwFlags ); - return ( WSOCK32_socket (af, type, protocol) ); + s = WSOCK32_socket (af, type, protocol); + + if (dwFlags && s != INVALID_SOCKET) + { + SERVER_START_REQ ( set_socket_flags ) + { + req->handle = s; + req->flags = dwFlags; + SERVER_CALL(); + } + SERVER_END_REQ; + } + return s; } Index: dlls/winsock/ws2_32.spec =================================================================== RCS file: /home/wine/wine/dlls/winsock/ws2_32.spec,v retrieving revision 1.13 diff -u -r1.13 ws2_32.spec --- dlls/winsock/ws2_32.spec 2001/10/02 17:46:59 1.13 +++ dlls/winsock/ws2_32.spec 2001/11/07 16:36:53 @@ -38,7 +38,7 @@ 23 stdcall socket(long long long) WSOCK32_socket 24 stdcall WSApSetPostRoutine(ptr) WSApSetPostRoutine 25 stub WPUCompleteOverlappedRequest -26 stub WSAAccept +26 stdcall WSAAccept(long ptr ptr ptr long) WSAAccept 27 stub WSAAddressToStringA 28 stub WSAAddressToStringW 29 stdcall WSACloseEvent(long) WSACloseEvent Index: include/winsock2.h =================================================================== RCS file: /home/wine/wine/include/winsock2.h,v retrieving revision 1.10 diff -u -r1.10 winsock2.h --- include/winsock2.h 2001/10/02 17:46:59 1.10 +++ include/winsock2.h 2001/11/07 16:36:54 @@ -225,10 +225,44 @@ CHAR* buf; } WSABUF, *LPWSABUF; -typedef struct _OVERLAPPED * LPWSAOVERLAPPED; typedef HANDLE WSAEVENT; + +typedef struct _WSAOVERLAPPED { + DWORD Internal; + DWORD InternalHigh; + DWORD Offset; + DWORD OffsetHigh; + WSAEVENT hEvent; +} WSAOVERLAPPED, *LPWSAOVERLAPPED; + +#define WSA_IO_PENDING (WSAEWOULDBLOCK) +#define WSA_IO_INCOMPLETE (WSAEWOULDBLOCK) +#define WSA_INVALID_HANDLE (WSAENOTSOCK) +#define WSA_INVALID_PARAMETER (WSAEINVAL) +#define WSA_NOT_ENOUGH_MEMORY (WSAENOBUFS) +#define WSA_OPERATION_ABORTED (WSAEINTR) + +#define WSA_INVALID_EVENT ((WSAEVENT)NULL) +#define WSA_MAXIMUM_WAIT_EVENTS (MAXIMUM_WAIT_OBJECTS) +#define WSA_WAIT_FAILED ((DWORD)-1L) +#define WSA_WAIT_EVENT_0 ((DWORD)0) +#define WSA_WAIT_TIMEOUT ((DWORD)0x102L) +#define WSA_INFINITE ((DWORD)-1L) + +/* + * WinSock 2 extension -- manifest constants for WSASocket() + */ +#define WSA_FLAG_OVERLAPPED 0x01 +#define WSA_FLAG_MULTIPOINT_C_ROOT 0x02 +#define WSA_FLAG_MULTIPOINT_C_LEAF 0x04 +#define WSA_FLAG_MULTIPOINT_D_ROOT 0x08 +#define WSA_FLAG_MULTIPOINT_D_LEAF 0x10 + typedef unsigned int GROUP; +/* FIXME: We don't support QOS */ +typedef DWORD QOS, *LPQOS; + typedef void CALLBACK (*LPWSAOVERLAPPED_COMPLETION_ROUTINE) ( DWORD dwError, @@ -237,7 +271,27 @@ DWORD dwFlags ); +/* + * WinSock 2 extension -- manifest constants for return values of the condition function + */ +#define CF_ACCEPT 0x0000 +#define CF_REJECT 0x0001 +#define CF_DEFER 0x0002 +/* FIXME: The lpCallerData / lpCalleeData arguments are unsupported */ +typedef int CALLBACK (*LPCONDITIONPROC) +( + LPWSABUF lpCallerId, + LPWSABUF lpCallerData, + LPQOS lpSQOS, + LPQOS lpGQOS, + LPWSABUF lpCalleeId, + LPWSABUF lpCalleeData, + GROUP *g, + DWORD dwCallbackData +); + + /* Function declarations */ int WINAPI WSAEnumNetworkEvents(SOCKET s, WSAEVENT hEventObject, LPWSANETWORKEVENTS lpNetworkEvents); int WINAPI WSAEventSelect(SOCKET s, WSAEVENT hEventObject, long lNetworkEvents); @@ -251,6 +305,9 @@ LPWSAPROTOCOL_INFOA lpProtocolInfo, GROUP g, DWORD dwFlags); INT WINAPI ioctlsocket(SOCKET s, LONG cmd, ULONG *argp); + +/* FIXME: omitted the deviating WIN64 declaration */ +SOCKET WINAPI WSAAccept(SOCKET,ws_sockaddr*,LPINT,LPCONDITIONPROC,DWORD); #include "poppack.h" Index: server/protocol.def =================================================================== RCS file: /home/wine/wine/server/protocol.def,v retrieving revision 1.16 diff -u -r1.16 protocol.def --- server/protocol.def 2001/10/24 00:23:26 1.16 +++ server/protocol.def 2001/11/07 16:36:55 @@ -661,6 +661,31 @@ handle_t event; /* event object */ @END +/* Set Winsock2 flags */ +(a)REQ(set_socket_flags) + handle_t handle; /* handle to the socket */ + unsigned int flags; /* Winsock2 flags */ +(a)END + +/* Get Winsock2 flags */ +(a)REQ(get_socket_flags) + handle_t handle; /* handle to the socket */ +(a)REPLY + unsigned int flags; /* Winsock2 flags */ +(a)END + +/* Get deferred socket to be accept()ed later, for WSAAccept() */ +(a)REQ(get_socket_deferred) + handle_t handle; /* handle to the socket */ +(a)REPLY + handle_t deferred; /* deferred socket */ +(a)END + +/* Set deferred socket to be accept()ed later, for WSAAccept() */ +(a)REQ(set_socket_deferred) + handle_t handle; /* handle to the socket */ + handle_t deferred; /* deferred socket */ +(a)END /* Get socket event parameters */ @REQ(get_socket_event) Index: server/sock.c =================================================================== RCS file: /home/wine/wine/server/sock.c,v retrieving revision 1.21 diff -u -r1.21 sock.c --- server/sock.c 2001/10/05 19:45:45 1.21 +++ server/sock.c 2001/11/07 16:36:55 @@ -47,6 +47,8 @@ unsigned int pmask; /* pending events */ struct event *event; /* event object */ int errors[FD_MAX_EVENTS]; /* event errors */ + unsigned int flags; /* Winsock 2 socket flags */ + handle_t deferred; /* deferred socket to be accept()ed later */ }; static void sock_dump( struct object *obj, int verbose ); @@ -324,6 +326,8 @@ sock->hmask = 0; sock->pmask = 0; sock->event = NULL; + sock->flags = 0; + sock->deferred = 0; sock_reselect( sock ); clear_error(); return &sock->obj; @@ -342,36 +346,51 @@ GENERIC_READ|GENERIC_WRITE|SYNCHRONIZE,&sock_ops); if (!sock) return NULL; - /* Try to accept(2). We can't be safe that this an already connected socket - * or that accept() is allowed on it. In those cases we will get -1/errno - * return. - */ - slen = sizeof(saddr); - acceptfd = accept(sock->obj.fd,&saddr,&slen); - if (acceptfd==-1) { - sock_set_error(); - release_object( sock ); - return NULL; - } - if (!(acceptsock = alloc_object( &sock_ops, -1 ))) - { - release_object( sock ); - return NULL; + + if (sock->deferred) { + + /* If a previous WSAAccept() deferred accepting this request, try it again. */ + /* We do this no matter if called through accept() or WSAAccept(), */ + /* because this is the way Windows does it, too. */ + + acceptsock = (struct sock*)get_handle_obj(current->process,sock->deferred, + GENERIC_READ|GENERIC_WRITE|SYNCHRONIZE,&sock_ops); + sock->deferred = 0; + } else { + /* Try to accept(2). We can't be safe that this an already connected socket + * or that accept() is allowed on it. In those cases we will get -1/errno + * return. + */ + slen = sizeof(saddr); + acceptfd = accept(sock->obj.fd,&saddr,&slen); + if (acceptfd==-1) { + sock_set_error(); + release_object( sock ); + return NULL; + } + if (!(acceptsock = alloc_object( &sock_ops, -1 ))) + { + release_object( sock ); + return NULL; + } + + /* newly created socket gets the same properties of the listening socket */ + fcntl(acceptfd, F_SETFL, O_NONBLOCK); /* make socket nonblocking */ + acceptsock->obj.fd = acceptfd; + acceptsock->state = FD_WINE_CONNECTED|FD_READ|FD_WRITE; + if (sock->state & FD_WINE_NONBLOCKING) + acceptsock->state |= FD_WINE_NONBLOCKING; + acceptsock->mask = sock->mask; + acceptsock->hmask = 0; + acceptsock->pmask = 0; + acceptsock->event = NULL; + acceptsock->flags = sock->flags; + acceptsock->deferred = 0; } - /* newly created socket gets the same properties of the listening socket */ - fcntl(acceptfd, F_SETFL, O_NONBLOCK); /* make socket nonblocking */ - acceptsock->obj.fd = acceptfd; - acceptsock->state = FD_WINE_CONNECTED|FD_READ|FD_WRITE; - if (sock->state & FD_WINE_NONBLOCKING) - acceptsock->state |= FD_WINE_NONBLOCKING; - acceptsock->mask = sock->mask; - acceptsock->hmask = 0; - acceptsock->pmask = 0; - acceptsock->event = NULL; if (sock->event && !(sock->mask & FD_WINE_SERVEVENT)) - acceptsock->event = (struct event *)grab_object( sock->event ); - + acceptsock->event = (struct event *)grab_object( sock->event ); + sock_reselect( acceptsock ); clear_error(); sock->pmask &= ~FD_ACCEPT; @@ -589,3 +608,37 @@ release_object( &sock->obj ); } +DECL_HANDLER(get_socket_deferred) +{ + struct sock *sock; + + sock=(struct sock*)get_handle_obj(current->process,req->handle,GENERIC_READ|GENERIC_WRITE|SYNCHRONIZE,&sock_ops); + req->deferred = sock->deferred; +} + +DECL_HANDLER(set_socket_deferred) +{ + struct sock *sock; + + sock=(struct sock*)get_handle_obj(current->process,req->handle,GENERIC_READ|GENERIC_WRITE|SYNCHRONIZE,&sock_ops); + + sock->deferred = req->deferred; + release_object (&sock->obj); +} + +DECL_HANDLER(get_socket_flags) +{ + struct sock *sock; + + sock=(struct sock*)get_handle_obj(current->process,req->handle,GENERIC_READ|GENERIC_WRITE|SYNCHRONIZE,&sock_ops); + req->flags = sock->flags; +} + +DECL_HANDLER(set_socket_flags) +{ + struct sock *sock; + + sock=(struct sock*)get_handle_obj(current->process,req->handle,GENERIC_READ|GENERIC_WRITE|SYNCHRONIZE,&sock_ops); + sock->flags = req->flags; +} +

3 3

Re: SystemParametersInfo
by Alexandre Julliard Nov. 7, 2001

Nov. 7, 2001

Andriy Palamarchuk <apa3a(a)yahoo.com> writes: > Changed system parameters won't be reset back to X > settings? Not if they are changed permanently; if we want them to be reset then we simply should not set them and always query X. Basically there are many possibilities: 1) get and set in registry, ignore X 2) get and set in registry, if nothing in registry get default value from X 3) always get from X, ignore registry, do nothing on set 4) always get from X, ignore registry, set X value on set (not recommended) and there are even more combinations if you differentiate setting permanently/temporarily. There is no single right answer, it all depends on what the parameter does. IMO most cases should do 1), but there can be exceptions. > One case which comes to mind when such integration > would be useful is system parameter which shows > whether mouse buttons are swapped. Left-handed person > can configure mouse after he ran Wine the first time, > or the settings can be changed temporarily. We need a > way to find out the system setting was changed and > don't rely on saved value. In this case I don't think we should query X at all, since if the buttons are swapped at the X level this is transparent for us. Otherwise the buttons will get swapped twice... -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: SystemParametersInfo
by Alexandre Julliard Nov. 7, 2001

Nov. 7, 2001

Andriy Palamarchuk <apa3a(a)yahoo.com> writes: > I'm going to initialize system parameters from X > settings only on the first wine process startup, so > all other wine processes for the same wine server can > see the changes made by the first process. The X settings should probably be loaded lazily just like the rest. Just check if the registry key exists already to find out if you need to get the value from X. Another possibility is to have the X driver set the registry values on startup; this may actually be cleaner since then you don't need to add new calls to the USER driver. Also I'm not convinced there are many cases where we actually need to get proper values from X; what cases do you have in mind? > Other issue is fixed with this approach - the values > in registry are not changed back to X settings on > startup of new Wine process. > > Alexandre, the submitted patch is not affected by this > issue because none of the system parameters there is > initialized from X settings. > > Do you have any other concerns about the patch? Not really; a few minor issues but I'll fix them when committing the patch. -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: SystemParametersInfo
by Alexandre Julliard Nov. 7, 2001

Nov. 7, 2001

Andreas Mohr <andi(a)rhlx01.fht-esslingen.de> writes: > On Mon, Nov 05, 2001 at 06:20:24AM -0800, Andriy Palamarchuk wrote: > > Is there any way to check: > > 1) whether Wine process is the first process started > > for current Wine server? > Heh, I should really know ;-) > CLIENT_IsBootThread(). No, this shouldn't be used from user32 (besides it won't really do what you want). Why do you need this at all? -- Alexandre Julliard julliard(a)winehq.com

2 1

OLE File Format Documentation
by Charles Wyble Nov. 7, 2001

Nov. 7, 2001

Hello I have obtained a copy of Microsofts Documentation on the Compound File Binary Format. I have posted it on my web site at chicago.sf.net/devel/ole/ff.txt. It is plain text now, but I will be converting it to HTML over the next few days. I hope it helps. _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com

1 0

Re: What does user mean?
by Alexandre Julliard Nov. 6, 2001

Nov. 6, 2001

"Bill Medland" <medbi01_1(a)accpac.com> writes: > I am working on the server code (adding the byte locking stuff) and down in > the select/thread/object stuff I keep coming across the term "user" which > seems to be synonymous with object. What is it supposed to mean? Why is an > object a user? How does it relate to a physical person or a windows process > or whatever? In that context "user" means an object that is a user of the select facilities. It has nothing to do with Unix users or anything else, it's purely internal to the select routines. -- Alexandre Julliard julliard(a)winehq.com

1 0