March 2006 - Wine-Devel - list.winehq.org

Re: Implement THREAD_PRIORITY_TIME_CRITICAL

by Andreas Mohr

Hi, On Fri, Mar 31, 2006 at 02:06:23PM +0100, Mike Hearn wrote: > This patch gives me rock solid audio in Imperium Galactica 2, so now the > game works perfectly. > > There has been discussion around this issue with respect to security in > the past, however, regardless of what approach is adopted this code will > have to be written anyway. Ah, thanks, nice to see that it actually helps. > + scheduler = SCHED_FIFO; IMHO we should really add some smart detection of SCHED_ISO kernel capability and *much* prefer to use that one then. One really wouldn't want to hang the box... Not to mention that SCHED_FIFO requires root, which is an absolute PITA. Nice To Have would be support for stuff such as SCHED_IDLEPRIO and SCHED_BATCH... > + if (result == -EPERM) > + { > + static int need_warning = 1; > + > + if (need_warning) > + { > + fprintf( stderr, "\nwineserver: Failed to promote the priority of a time critical thread.\n" ); > + fprintf( stderr, "Audio may destabilise. To fix this re-run the application as root.\n\n" ); > + need_warning = 0; > + } > + > + return; > + } Make that static int already_warned; /* static: = 0! */ Since static uses 0 as default and you don't want to waste space in the .bss(?) segment for an explicit 1 init. Andreas Mohr -- No programming skills!? Why not help translate many Linux applications! https://launchpad.ubuntu.com/rosetta (or alternatively buy nicely packaged Linux distros/OSS software to help support Linux developers creating shiny new things for you?)

March 31, 2006

Re: [RESEND][x11drv] cleanup: Move x11drv escape codes to one common header file.

by Tomas Carnecky

Alexandre Julliard wrote: > Tomas Carnecky <tom(a)dbservice.com> writes: > >> This patch doesn't change any logic or C sourcecode, it only moves the >> X11DRV escape code definition to one common header file. Currently the >> definitions are spread over several C source files, and even there the >> definitions differ from file to file. I don't see anything wrong with >> this patch unless you *want* to have it difficult to maintain the code. > > Actually, that's more or less what we want. That sort of inter-dll > dependencies should be strongly discouraged, and we don't want to add > more of it than strictly necessary; as far as possible we also don't > want dlls to depend on private headers, so copying a few things across > is the preferred approach. If it becomes a problem it means something > is wrong with our implementation. > It's not that I've added a new inter-dll dependency, at least that's how I think. The dependency already *is* there, all these dlls depend on the ExtEscape() function. What about putitng the x11drv escape codes into the header where the ExtEscape() prototype is? These escape codes belong to the ExtEscape() function, that's what that function understands/supports. tom

March 31, 2006

Re: "Unhandled exception: floating point stack check" early in program?

by Petr Tesarik

Hi, I'm not online too often. :( So, I thought the patch for Linux must already be in CVS. It is not. The patch can be found here: http://www.winehq.org/pipermail/wine-patches/2006-March/025228.html However, the patch only works for Linux. Please, let me know if you're using a different platform, as it is very likely that it can be made to work too, but I won't send patches which I can't test. To the wine-devel list: Why was the patch rejected, anyway? Petr Tesarik On 03/29/06 at 09:35:04 (-0800), Dan Kegel wrote: > On 3/29/06, Petr Tesarik <hat(a)tesarici.cz> wrote: > > No, this is not normal. This problem was introduced on the i386 > > architecture, when floating point registers were added to winedbg. > > Currently, there are two pending patches by me, one which fixes the > > problem on Linux, and another which fixes it on all platforms but is > > most likely wrong. > >... > > Anyway, see also my next email. > > Looking forward to it. > Can you give me the URL of the i386 patche in the wine-patches archive? > Thanks, > Dan >

March 31, 2006

[PATCH] Don't enforce process handle access rights in Win98 mode (fixes dcom98 ncalrpc)

by Mike Hearn

Sending here until wine-patches stops eating my email. IMHO this would best be done in get_obj_from_handle in the server, but this approach is what the registry APIs do and what Rob suggested, so here it is. diff --git a/dlls/kernel/process.c b/dlls/kernel/process.c index 4263a22..d5017c4 100644 --- a/dlls/kernel/process.c +++ b/dlls/kernel/process.c @@ -2265,6 +2265,8 @@ HANDLE WINAPI OpenProcess( DWORD access, attr.SecurityQualityOfService = NULL; attr.ObjectName = NULL; + if (GetVersion() & 0x80000000) access = PROCESS_ALL_ACCESS; + status = NtOpenProcess(&handle, access, &attr, &cid); if (status != STATUS_SUCCESS) {

March 31, 2006

winspool: [1/2] implement EnumMonitorsW

by Detlef Riekenberg

Our Tests will work, when both Patches are in the Tree Changelog: - winspool: implement EnumMonitorsW -- By By ... ... Detlef

March 31, 2006

Re: [WIDL] Implement support for unique strings

by Eric Kohl

Robert Shearman wrote: > Counter-example with the following IDL: > void ConformantString2Out([in] long size, [in, out, string, > size_is(size)] char **pszString); > void Test([in] long size, [in, out, size_is(size)] char **pszArray); > > Both generate NdrPointerFree calls and I believe both are satisfied by > the test you introduced: These example include arrays of strings which I have not tested yet and are not officially supported yet. I'll fix these later. My focus is making the Wine-WIDL compatible with ReactOS-WIDL as quickly as possible. Additional features will be added or fixed later. Regards, Eric

March 31, 2006

Re: [WIDL] Implement support for unique strings

by Robert Shearman

Eric Kohl wrote: > Robert Shearman wrote: > >>> case PASS_OUT: >>> + if (in_attr && has_size && pointer_type == RPC_FC_UP && >>> phase == PHASE_FREE) >>> + break; >>> if (!out_attr) >>> continue; >>> break; >>> >>> >> >> I don't think this is correct. Firstly, it should continue, not >> break. Secondly, I think it only doesn't need to free the string if >> it has a size. I'm pretty sure it shouldn't depend on being a unique >> string. >> > > Robert, > this is exactly what I want! WIDL must generate a call to > NdrPointerFree for 'in', 'unique', sized pointers to strings. Other > 'in' pointers don't need to be freed. > > Why? Because three different versions of MIDL do it this way. > Counter-example with the following IDL: void ConformantString2Out([in] long size, [in, out, string, size_is(size)] char **pszString); void Test([in] long size, [in, out, size_is(size)] char **pszArray); Both generate NdrPointerFree calls and I believe both are satisfied by the test you introduced: void __RPC_STUB Tests_ConformantString2Out( PRPC_MESSAGE _pRpcMessage ) { MIDL_STUB_MESSAGE _StubMsg; unsigned char **pszString; long size; RPC_STATUS _Status; ((void)(_Status)); NdrServerInitializeNew( _pRpcMessage, &_StubMsg, &Tests_StubDesc); size = 0; ( unsigned char ** )pszString = 0; RpcTryFinally { RpcTryExcept { if ( (_pRpcMessage->DataRepresentation & 0X0000FFFFUL) != NDR_LOCAL_DATA_REPRESENTATION ) NdrConvert( (PMIDL_STUB_MESSAGE) &_StubMsg, (PFORMAT_STRING) &__MIDL_ProcFormatString.Format[44] ); _StubMsg.Buffer = (unsigned char *)(((long)_StubMsg.Buffer + 3) & ~ 0x3); if(_StubMsg.Buffer + 4 > _StubMsg.BufferEnd) { RpcRaiseException(RPC_X_BAD_STUB_DATA); } size = *(( long * )_StubMsg.Buffer)++; NdrConformantArrayUnmarshall( (PMIDL_STUB_MESSAGE) &_StubMsg, (unsigned char * *)&pszString, (PFORMAT_STRING) &__MIDL_TypeFormatString.Format[44], (unsigned char)0 ); if(_StubMsg.Buffer > _StubMsg.BufferEnd) { RpcRaiseException(RPC_X_BAD_STUB_DATA); } } RpcExcept( RPC_BAD_STUB_DATA_EXCEPTION_FILTER ) { RpcRaiseException(RPC_X_BAD_STUB_DATA); } RpcEndExcept ConformantString2Out(size,pszString); _StubMsg.BufferLength = 0; _StubMsg.MaxCount = ( unsigned long )size; NdrConformantArrayBufferSize( (PMIDL_STUB_MESSAGE) &_StubMsg, (unsigned char *)pszString, (PFORMAT_STRING) &__MIDL_TypeFormatString.Format[44] ); _pRpcMessage->BufferLength = _StubMsg.BufferLength; _Status = I_RpcGetBuffer( _pRpcMessage ); if ( _Status ) RpcRaiseException( _Status ); _StubMsg.Buffer = (unsigned char *) _pRpcMessage->Buffer; _StubMsg.MaxCount = ( unsigned long )size; NdrConformantArrayMarshall( (PMIDL_STUB_MESSAGE)& _StubMsg, (unsigned char *)pszString, (PFORMAT_STRING) &__MIDL_TypeFormatString.Format[44] ); } RpcFinally { _StubMsg.MaxCount = ( unsigned long )size; NdrPointerFree( &_StubMsg, (unsigned char *)pszString, &__MIDL_TypeFormatString.Format[40] ); } RpcEndFinally _pRpcMessage->BufferLength = (unsigned int)(_StubMsg.Buffer - (unsigned char *)_pRpcMessage->Buffer); } void __RPC_STUB Tests_Test( PRPC_MESSAGE _pRpcMessage ) { MIDL_STUB_MESSAGE _StubMsg; unsigned char **pszArray; long size; RPC_STATUS _Status; ((void)(_Status)); NdrServerInitializeNew( _pRpcMessage, &_StubMsg, &Tests_StubDesc); size = 0; ( unsigned char ** )pszArray = 0; RpcTryFinally { RpcTryExcept { if ( (_pRpcMessage->DataRepresentation & 0X0000FFFFUL) != NDR_LOCAL_DATA_REPRESENTATION ) NdrConvert( (PMIDL_STUB_MESSAGE) &_StubMsg, (PFORMAT_STRING) &__MIDL_ProcFormatString.Format[52] ); _StubMsg.Buffer = (unsigned char *)(((long)_StubMsg.Buffer + 3) & ~ 0x3); if(_StubMsg.Buffer + 4 > _StubMsg.BufferEnd) { RpcRaiseException(RPC_X_BAD_STUB_DATA); } size = *(( long * )_StubMsg.Buffer)++; NdrConformantArrayUnmarshall( (PMIDL_STUB_MESSAGE) &_StubMsg, (unsigned char * *)&pszArray, (PFORMAT_STRING) &__MIDL_TypeFormatString.Format[78], (unsigned char)0 ); if(_StubMsg.Buffer > _StubMsg.BufferEnd) { RpcRaiseException(RPC_X_BAD_STUB_DATA); } } RpcExcept( RPC_BAD_STUB_DATA_EXCEPTION_FILTER ) { RpcRaiseException(RPC_X_BAD_STUB_DATA); } RpcEndExcept Test(size,pszArray); _StubMsg.BufferLength = 0; _StubMsg.MaxCount = ( unsigned long )size; NdrConformantArrayBufferSize( (PMIDL_STUB_MESSAGE) &_StubMsg, (unsigned char *)pszArray, (PFORMAT_STRING) &__MIDL_TypeFormatString.Format[78] ); _pRpcMessage->BufferLength = _StubMsg.BufferLength; _Status = I_RpcGetBuffer( _pRpcMessage ); if ( _Status ) RpcRaiseException( _Status ); _StubMsg.Buffer = (unsigned char *) _pRpcMessage->Buffer; _StubMsg.MaxCount = ( unsigned long )size; NdrConformantArrayMarshall( (PMIDL_STUB_MESSAGE)& _StubMsg, (unsigned char *)pszArray, (PFORMAT_STRING) &__MIDL_TypeFormatString.Format[78] ); } RpcFinally { _StubMsg.MaxCount = ( unsigned long )size; NdrPointerFree( &_StubMsg, (unsigned char *)pszArray, &__MIDL_TypeFormatString.Format[74] ); } RpcEndFinally _pRpcMessage->BufferLength = (unsigned int)(_StubMsg.Buffer - (unsigned char *)_pRpcMessage->Buffer); } -- Rob Shearman

March 31, 2006

Re: advpack: Implement LaunchINFSectionEx [resend]

by Alexandre Julliard

"James Hawkins" <truiken(a)gmail.com> writes: > +static LPWSTR get_parameter(LPWSTR parameters, WCHAR separator, DWORD index) > +{ > + LPWSTR param = NULL, ptr = parameters; > + DWORD i = 0; > + > + while (*parameters && i < index) > + { > + if (*parameters == separator) > + { > + ptr = parameters; > + i++; > + } > + > + parameters++; > + } > + > + if (!*parameters || *parameters == separator) > + return NULL; > + > + param = HeapAlloc(GetProcessHeap(), 0, (parameters - ptr) * sizeof(WCHAR)); > + lstrcpynW(param, parameters, parameters - ptr); > + > + return param; > +} That's a very inefficient way of splitting a string. It's much better to allocate a copy of the whole string and go through it just once, inserting nulls to separate the tokens. -- Alexandre Julliard julliard(a)winehq.org

March 31, 2006

Re: advpack: The Plan

by Markus Amsler

James Hawkins wrote: >Internally, there will be three main install functions: install_init, >which will open the INF, make sure it's legit, and other >initializations, spapi_install, which will call setupapi to install >the base INF commands, and adv_install, which will parse the install >section and run the provided commands. > I'm quite sure RunPreSetupCommands gets executed before the base INF commands. I'm not sure wether there are others (BeginPrompt, BackupReg comes to mind). So you need two install functions like adv_install_pre and adv_install_post. >This is The Plan so far, so if anyone has any ideas or suggestions, >I'd love to hear from you. > > Why not simply use the same mechanism as in setupapi [1]? My patch [2] back in january was using it, and it looked like a good idea. Markus [1] http://source.winehq.org/source/dlls/setupapi/install.c [2] http://www.winehq.org/pipermail/wine-devel/2006-January/044083.html

March 31, 2006

Re: advpack: The Plan

by Alexandre Julliard

"James Hawkins" <truiken(a)gmail.com> writes: > Internally, there will be three main install functions: install_init, > which will open the INF, make sure it's legit, and other > initializations, spapi_install, which will call setupapi to install > the base INF commands, and adv_install, which will parse the install > section and run the provided commands. Each advanced command will be > implemented by a function of similar name, e.g., RegisterOCXs would be > implemented by the function register_ocxs, etc. All the command > functions will be in a hash map, the key being the command name, the > value being a pointer to the function that implements the command. > adv_install will use setupapi to parse the entries of the install > section, calling the function returned by a query to the hash map. > According to INF Web [1], some commands are run before others, so > there will probably be three or four maps, the functions in the maps > grouped according to the order in which the commands should be run. > For example, RunPreSetupCommands is to be executed after the > BeginPrompt command, so begin_prompt would be in the phase 1 map, and > run_pre_setup_commands would be in the phase 2 map. We would call the > parser like so: > > run_adv_commands("InstallSection", phase1_map); > run_adv_commands("InstallSection", phase2_map); > run_adv_commands("InstallSection", phase3_map); > > If a command in the install section is not in the map, we just ignore it. That hash map thing sounds like overkill to me. You have only a few sections you know you will handle, you can just look for these explicitly, like setupapi does. And if you look them up in the order they should be executed you don't need the different phases either. -- Alexandre Julliard julliard(a)winehq.org

March 31, 2006