July 2021 - Wine-Bugs - list.winehq.org

[Bug 51416] New: deferred context memory leak
by WineHQ Bugzilla July 20, 2021

July 20, 2021

https://bugs.winehq.org/show_bug.cgi?id=51416 Bug ID: 51416 Summary: deferred context memory leak Product: Wine Version: 6.12 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: directx-d3d Assignee: wine-bugs(a)winehq.org Reporter: janbraun(a)gmx.net CC: z.figura12(a)gmail.com Distribution: --- This is a followup to bug#51238. While WoWs now starts and is playable with deferred contexts, it leaks LOTS of memory. For comparison, without deferred contexts [either wine < 6.10, or by adding "return E_OUTOFMEMORY;" at the beginning of *_deferred_context_create() in current wine], WoWs sits at a fairly constant 8G virt, 2.5G res memory usage here. With unmodified wine 6.12, WoWs arrives at these values when it finishes loading, but then continues to gobble up an additional approx. 1G virt=res per minute, just for sitting idly in the port UI, backgrounded. Memory usage keeps rising similarly during actual gameplay, though I haven't measured that in detail. Obviously, swapping and general mayhem eventually ensues. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 3

[Bug 51366] New: several wine-mono network tests fail with timeout
by WineHQ Bugzilla July 20, 2021

July 20, 2021

https://bugs.winehq.org/show_bug.cgi?id=51366 Bug ID: 51366 Summary: several wine-mono network tests fail with timeout Product: Wine Version: 6.11 Hardware: x86-64 OS: Linux Status: NEW Severity: normal Priority: P2 Component: winsock Assignee: wine-bugs(a)winehq.org Reporter: madewokherd(a)gmail.com Distribution: --- The following wine-mono test cases fail on the master branch: MonoTests.System.Net.HttpWebRequestTest:TestTimeoutWithEndpointThatDoesntExistThrowsConnectFailureBeforeTimeout MonoTests.System.Net.Sockets.SocketTest:ConnectFailAsync MonoTests.System.Net.WebRequestTest:TestFailedConnection These can be run using run-tests.exe from https://dl.winehq.org/wine/wine-mono/6.2.0/wine-mono-6.2.0-tests.zip Testcase source code: https://github.com/madewokherd/mono/blob/3445531d2550157162ba71a5ff460a4d6a… https://github.com/madewokherd/mono/blob/3445531d2550157162ba71a5ff460a4d6a… https://github.com/madewokherd/mono/blob/3445531d2550157162ba71a5ff460a4d6a… Bisect result: 414b31bc0bbbfe005e90a1946a649082dc303c55 is the first bad commit commit 414b31bc0bbbfe005e90a1946a649082dc303c55 Author: Zebediah Figura <z.figura12(a)gmail.com> Date: Sun Jun 20 11:05:54 2021 -0500 ws2_32: Reimplement select() on top of IOCTL_AFD_POLL. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 4

[Bug 51340] New: d3d10core:d3d10core test_texture_compressed_3d() fails on Nvidia (cw-gtx560)
by WineHQ Bugzilla July 20, 2021

July 20, 2021

https://bugs.winehq.org/show_bug.cgi?id=51340 Bug ID: 51340 Summary: d3d10core:d3d10core test_texture_compressed_3d() fails on Nvidia (cw-gtx560) Product: Wine Version: 6.10 Hardware: x86-64 OS: Windows Status: NEW Severity: normal Priority: P2 Component: directx-d3d Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com test_texture_compressed_3d() fails systematically on cw-gtx560, the TestBot's only Nvidia test configuration: https://test.winehq.org/data/patterns.html#d3d10core:d3d10core d3d10core.c:19081: Test failed: Got unexpected colour 0xff00006a at (251, 9, 0), expected 0xff00006d. This failure happens on all Windows versions and with the 353.82 (win81), 388.13 (win 10 1709 & 1809) and 391.35 (win10 1909 & 2009) driver versions. Also this failure does not happen on any of the other GPUs (cw-rx460 and QEmu VMs). The failing test was introduced in this commit: commit 2aeab85c27a202d0ef7b1e01d7ccab32d9d53c93 Author: Henri Verbeet <hverbeet(a)codeweavers.com> Date: Tue Jun 22 14:36:00 2021 +0200 wined3d: Add decompression information for BC4_UNORM. Signed-off-by: Henri Verbeet <hverbeet(a)codeweavers.com> Signed-off-by: Alexandre Julliard <julliard(a)winehq.org> -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 7

[Bug 41703] New: LdrUnloadDll is not suit for musl( musl lacks unload support though it has a dlclose())
by wine-bugs＠winehq.org July 20, 2021

July 20, 2021

https://bugs.winehq.org/show_bug.cgi?id=41703 Bug ID: 41703 Summary: LdrUnloadDll is not suit for musl(musl lacks unload support though it has a dlclose()) Product: Wine Version: unspecified Hardware: x86 OS: Linux Status: UNCONFIRMED Severity: major Priority: P2 Component: ntdll Assignee: wine-bugs(a)winehq.org Reporter: xw897002528(a)gmail.com Distribution: --- Created attachment 56111 --> https://bugs.winehq.org/attachment.cgi?id=56111 fix for this problem Hey, sorry for my poor english, and here is the problem: First, musl won't unload module even dlclose(). Check here: http://wiki.musl-libc.org/wiki/Functional_differences_from_glibc#Unloading_… Then, musl won't excute init(__wine_spec_init) function if a module is already loaded. That means load_builtin_dll() thought dll is loaded by the current thread since init function was not called, but actually LdrUnloadDll() destroyed them in another thread and nothing is loaded in the current thread. Load_builtin_dll() try to find wm in InLoadOrderModuleList but nothing. After that, app will crash. And I solved this by making LdrUnloadDll() a stub. Patch is uploaded as a attachment. I hope this could be added in official wine by adding a configure option, thx. Here's my tracking details on this problem if you're intereted in how i found this: https://github.com/xhebox/noname-linux/issues/2#issuecomment-258680745 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

2 36

[Bug 51381] New: Battlefield 4 cannot find online matches
by WineHQ Bugzilla July 20, 2021

July 20, 2021

https://bugs.winehq.org/show_bug.cgi?id=51381 Bug ID: 51381 Summary: Battlefield 4 cannot find online matches Product: Wine Version: 6.9 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: winsock Assignee: wine-bugs(a)winehq.org Reporter: GloriousEggroll(a)gmail.com Distribution: --- Created attachment 70236 --> https://bugs.winehq.org/attachment.cgi?id=70236 workaround for broken matchmaking/online functionality in various games Online functionality in several games was broken by: https://discord.com/channels/414894830418657301/414894830418657303/85202511… Notably Battlefield 4 cannot join matches, MK11 cannot go online, NBA2k21 cannot join online matches or see friends, Rocket League cannot join matches/see friends, steam cannot connect online. This was partially fixed upstream (not sure what commit), however Battlefield 4 matchmaking still remains broken. The attached patch (from Zebediah Figura) resolves the BF4 and other relevant game issue. Opening bug for further investigation. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 5

[Bug 51295] New: ntoskrnl.exe:ntoskrnl occasionally hangs in test_overlapped()
by WineHQ Bugzilla July 20, 2021

July 20, 2021

https://bugs.winehq.org/show_bug.cgi?id=51295 Bug ID: 51295 Summary: ntoskrnl.exe:ntoskrnl occasionally hangs in test_overlapped() Product: Wine Version: 6.10 Hardware: x86-64 OS: Linux Status: NEW Keywords: regression, source, testcase Severity: normal Priority: P2 Component: wineserver Assignee: wine-bugs(a)winehq.org Reporter: z.figura12(a)gmail.com Regression SHA1: 97afac469fbe012e22acc1f1045c88b1004a241f Distribution: --- Created attachment 70168 --> https://bugs.winehq.org/attachment.cgi?id=70168 make the problem more reproducible Specifically, it hangs if ntoskrnl is able to signal that the IRP is pending (by calling get_next_device_request) before the user space process calls wait_async(). 97afac469 intends for wait_async() to return STATUS_PENDING, because that is the synchronous return of the ioctl. The problem is that we're trying to have the select request *succeed immediately*, but STATUS_PENDING means that it didn't succeed immediately and that the client needs to wait for a wakeup. This can be triggered pretty reliably by running the relevant tests in a loop, as in the attached patch. I think <https://testbot.winehq.org/JobDetails.pl?Key=92616&f401=win32.report#k401> is an example of this happening on the testbot. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 3

[Bug 51262] New: Regression: all 32bit programs doesn't start since 308a5e7
by WineHQ Bugzilla July 20, 2021

July 20, 2021

https://bugs.winehq.org/show_bug.cgi?id=51262 Bug ID: 51262 Summary: Regression: all 32bit programs doesn't start since 308a5e7 Product: Wine Version: 6.10 Hardware: x86-64 OS: Linux Status: NEW Severity: blocker Priority: P2 Component: ntdll Assignee: wine-bugs(a)winehq.org Reporter: zzhang(a)codeweavers.com Distribution: --- master:wine/win32-build$ wine notepad 002c:err:virtual:virtual_setup_exception stack overflow 1156 bytes in thread 002c addr 0x7bc1f36f stack 0x120b7c (0x120000-0x121000-0x320000) 0024:err:virtual:virtual_setup_exception stack overflow 1156 bytes in thread 0024 addr 0x7bc1f36f stack 0x120b7c (0x120000-0x121000-0x320000) master:wine/win32-build$ which wine wine: aliased to ~/src/wine/win32-build/wine wine is configure with "../configure --enable-silent-rules --disable-tests". Ccache is disabled and wineprefix is new. master:wine/win32-build$ gcc --version gcc (Ubuntu 10.3.0-1ubuntu1) 10.3.0 Copyright (C) 2020 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. master:wine/win32-build$ i686-w64-mingw32-gcc --version i686-w64-mingw32-gcc (GCC) 10-win32 20210408 Copyright (C) 2020 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. master:wine/win32-build$ uname -a Linux eric-pc 5.11.0-18-generic #19-Ubuntu SMP Fri May 7 14:22:03 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux pure 32bit wine, wow32 wine, wow64 wine all crash with stack overflows. Only pure 64bit wine starts fine. Bisect points to 308a5e7c4d01525184b057058949c68c0ee77dba is the first bad commit commit 308a5e7c4d01525184b057058949c68c0ee77dba Author: Alexandre Julliard <julliard(a)winehq.org> Date: Fri Jun 11 10:57:20 2021 +0200 ntdll: Switch to the kernel stack for syscalls on i386. Signed-off-by: Alexandre Julliard <julliard(a)winehq.org> dlls/ntdll/unix/signal_i386.c | 250 ++++++++++++++++++++---------------------- tools/winebuild/import.c | 192 ++++++++++++-------------------- 2 files changed, 186 insertions(+), 256 deletions(-) -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 11

[Bug 50374] New: pCon.planner doesn't finish installation and doesn't open
by WineHQ Bugzilla July 20, 2021

July 20, 2021

https://bugs.winehq.org/show_bug.cgi?id=50374 Bug ID: 50374 Summary: pCon.planner doesn't finish installation and doesn't open Product: Wine Version: 6.0-rc2 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: tinozzo123(a)gmail.com Distribution: --- Created attachment 68979 --> https://bugs.winehq.org/attachment.cgi?id=68979 Contains installation.txt and backtrace.txt pCon.planner 8.4 Update 2 doesn't install correctly: the progress bar of the installer reaches the end, but the installation complete screen doesn't appear. Trying to cancel the installation once it actually started also doesn't do anything (didn't test if it's the same on Windows), so the only way to quit the installation is by killing wineserver. Wine logs during the installation are attached. The files of the program are there though, but trying to execute it results in the error from TLocaleReader "(-2147418113) Line 4 Column 19", and generates a backtrace I attached below. The uninstaller doesn't work either, with the progress bar not progressing at all and a cancel button that doesn't cancel anything. Link for the program download: https://pcon-planner.com/en/download/ -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 5

[Bug 51425] New: Mathearbeit G 5.6 installer reports null pointer exception in module 'shell32.dll' starting with Wine 6.9-142-g98d43c5dcfb
by WineHQ Bugzilla July 20, 2021

July 20, 2021

https://bugs.winehq.org/show_bug.cgi?id=51425 Bug ID: 51425 Summary: Mathearbeit G 5.6 installer reports null pointer exception in module 'shell32.dll' starting with Wine 6.9-142-g98d43c5dcfb Product: Wine Version: 6.12 Hardware: x86-64 OS: Linux Status: NEW Severity: normal Priority: P2 Component: shell32 Assignee: wine-bugs(a)winehq.org Reporter: focht(a)gmx.net Distribution: --- Hello folks, while revisiting bug 38740 ("Mathearbeit G 5.6 installer hangs during installation (ShellFolder attributes for virtual folder 'CLSID_Printers', clsid '{2227a280-3aea-1069-a2de-08002b30309d}' missing in registry)") to validate the claim that the bug has been fixed I've found a regression which actually hides the original bug. Stable download link via Internet Archive: https://web.archive.org/web/20210708082931/http://zahlenbasar.de/MA_DL/setu… --- snip --- ... 0118:trace:shell:_SHGetUserShellFolderPath 0x80000002,(null),L"{00000000-0000-0000-0000-000000000000}",0x32ea98 0118:Call advapi32.RegCreateKeyW(80000002,7daf0dc0 L"Software\\Microsoft\\Windows\\CurrentVersion\\Explorer\\Shell Folders",0032e18c) ret=7daa23be ... 0118:Ret advapi32.RegCreateKeyW() retval=00000000 ret=7daa23be 0118:Call advapi32.RegCreateKeyW(80000002,7daf0d20 L"Software\\Microsoft\\Windows\\CurrentVersion\\Explorer\\User Shell Folders",0032e188) ret=7daa23d8 ... 0118:Ret advapi32.RegCreateKeyW() retval=00000000 ret=7daa23d8 0118:Call advapi32.RegQueryValueExW(000000f0,0032e800 L"{00000000-0000-0000-0000-000000000000}",00000000,0032e190,0032ea98,0032e194) ret=7daa2417 ... 0118:Ret advapi32.RegQueryValueExW() retval=00000002 ret=7daa2417 ... 0118:trace:shell:_SHGetUserShellFolderPath returning 0x80004005 0118:trace:shell:_SHGetDefaultValue 0x45,0x32ea98 0118:trace:seh:dispatch_exception code=c0000005 flags=0 addr=7DAA0EA5 ip=7daa0ea5 tid=0118 0118:trace:seh:dispatch_exception info[0]=00000000 0118:trace:seh:dispatch_exception info[1]=00000000 0118:warn:seh:dispatch_exception EXCEPTION_ACCESS_VIOLATION exception (code=c0000005) raised 0118:trace:seh:dispatch_exception eax=00000000 ebx=00000045 ecx=0032e354 edx=0032e37e esi=0032ea98 edi=00000000 0118:trace:seh:dispatch_exception ebp=0032e7d8 esp=0032e7a0 cs=0023 ds=002b es=002b fs=0063 gs=006b flags=00010202 0118:trace:seh:call_vectored_handlers calling handler at 7B00F750 code=c0000005 flags=0 0118:trace:seh:call_vectored_handlers handler at 7B00F750 returned 0 0118:trace:seh:call_stack_handlers calling handler at 004D512B code=c0000005 flags=0 0118:trace:seh:call_stack_handlers handler at 004D512B returned 1 0118:trace:seh:call_stack_handlers calling handler at 004E2570 code=c0000005 flags=0 0118:trace:seh:call_stack_handlers handler at 004E2570 returned 1 0118:trace:seh:call_stack_handlers calling handler at 0044B389 code=c0000005 flags=0 ... 0118:Call user32.MessageBoxA(0001009a,02dbd314 "Zugriffsverletzung bei Adresse 7DAA0EA5 in Modul 'shell32.dll'. Lesen von Adresse 00000000.",013f5160 "MATHEARBEIT G Setup",00000010) ret=0045273f ... --- snip --- After dismissing the dialog, the installer runs further and indeed completes without live-looping (copies files). Bug 38740 *appears* to be fixed but in fact it's not. Commit https://source.winehq.org/git/wine.git/commitdiff/98d43c5dcfb28183667d874d4… ("shell32: Calculate known folder paths from parent and relative path."), part of Wine 6.10 release, introduced a regression which causes a NULL pointer exception the reworked shell32 code. Wine source: https://source.winehq.org/git/wine.git/blob/HEAD:/dlls/shell32/shellpath.c#… --- snip --- 2446 { /* 0x45 */ 2447 &GUID_NULL, 2448 CSIDL_Type_User, 2449 NULL, 2450 DocumentsW 2451 }, --- snip --- https://source.winehq.org/git/wine.git/blob/HEAD:/dlls/shell32/shellpath.c#… --- snip --- 3470 /* Gets a 'semi-expanded' default value of the CSIDL with index folder into 3471 * pszPath, based on the entries in CSIDL_Data. By semi-expanded, I mean: 3472 * - Depending on the entry's type, the path may begin with an (unexpanded) 3473 * environment variable name. The caller is responsible for expanding 3474 * environment strings if so desired. 3475 * The types that are prepended with environment variables are: 3476 * CSIDL_Type_User: %USERPROFILE% 3477 * CSIDL_Type_AllUsers: %ALLUSERSPROFILE% 3478 * CSIDL_Type_CurrVer: %SystemDrive% 3479 * (Others might make sense too, but as yet are unneeded.) 3480 */ 3481 static HRESULT _SHGetDefaultValue(BYTE folder, LPWSTR pszPath) 3482 { 3483 HRESULT hr; 3484 3485 TRACE("0x%02x,%p\n", folder, pszPath); 3486 3487 if (folder >= ARRAY_SIZE(CSIDL_Data)) 3488 return E_INVALIDARG; 3489 3490 if (!pszPath) 3491 return E_INVALIDARG; 3492 3493 if (!is_win64) 3494 { 3495 BOOL is_wow64; 3496 3497 switch (folder) 3498 { 3499 case CSIDL_PROGRAM_FILES: 3500 case CSIDL_PROGRAM_FILESX86: 3501 IsWow64Process( GetCurrentProcess(), &is_wow64 ); 3502 folder = is_wow64 ? CSIDL_PROGRAM_FILESX86 : CSIDL_PROGRAM_FILES; 3503 break; 3504 case CSIDL_PROGRAM_FILES_COMMON: 3505 case CSIDL_PROGRAM_FILES_COMMONX86: 3506 IsWow64Process( GetCurrentProcess(), &is_wow64 ); 3507 folder = is_wow64 ? CSIDL_PROGRAM_FILES_COMMONX86 : CSIDL_PROGRAM_FILES_COMMON; 3508 break; 3509 } 3510 } 3511 3512 if (IsEqualGUID(CSIDL_Data[folder].fidParent, &GUID_NULL)) 3513 { 3514 /* hit the root, sub in env var */ 3515 switch (CSIDL_Data[folder].type) 3516 { 3517 case CSIDL_Type_User: 3518 strcpyW(pszPath, UserProfileW); 3519 break; 3520 case CSIDL_Type_AllUsers: 3521 strcpyW(pszPath, PublicProfileW); 3522 break; 3523 case CSIDL_Type_ProgramData: 3524 strcpyW(pszPath, ProgramDataVarW); 3525 break; 3526 case CSIDL_Type_CurrVer: 3527 strcpyW(pszPath, SystemDriveW); 3528 break; 3529 default: 3530 ; /* no corresponding env. var, do nothing */ 3531 } 3532 hr = S_OK; 3533 }else{ 3534 /* prepend with parent */ 3535 hr = _SHGetDefaultValue(csidl_from_id(CSIDL_Data[folder].fidParent), pszPath); 3536 } 3537 3538 if (SUCCEEDED(hr)) 3539 append_relative_path(folder, pszPath); 3540 3541 TRACE("returning 0x%08x\n", hr); 3542 return hr; 3543 } --- snip --- Line 3512 -> 'fidParent' member is zero-init for 'CSIDL_Data' 0x45 folder entry, causing null pointer exception. After fixing the regression, the installer runs into bug 38740 $ sha1sum setupgs5_6.exe fbbbae71ce4214e3848ae29399fa3b271bd6763f setupgs5_6.exe $ du -sh setupgs5_6.exe 5.8M setupgs5_6.exe $ wine --version wine-6.12-125-g6763ed84cf0 Regards -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 4

[Bug 45807] New: Incorrect toolbar background in MFC apps, when visual theming enabled
by wine-bugs＠winehq.org July 20, 2021

July 20, 2021

https://bugs.winehq.org/show_bug.cgi?id=45807 Bug ID: 45807 Summary: Incorrect toolbar background in MFC apps, when visual theming enabled Product: Wine Version: 3.15 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: jimbo1qaz(a)gmail.com Distribution: --- Created attachment 62266 --> https://bugs.winehq.org/attachment.cgi?id=62266 Screenshot of bug When visual theming is enabled, MFC toolbar button backgrounds are rendered as "grab handles", and space between buttons is black. (Also visible in screenshot, in 0CC and j0CC, the instrument editor piano is invisible when theming is enabled. Should I file a separate bug?) Tested with ReactOS.iso Lautus theme, and Zune theme (via .msi). Tested with j0CC-Famitracker and OpenMPT (both use MFC), on dirty WINEPREFIX. Also tested with Lautus and j0CC on clean WINEPREFIX with Lautus theme. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

2 4