- wine-gitlab - winehq.org

[PATCH 0/1] MR7657: win32u: Store window styles in a DWORD (Coverity).
by Sven Baars (＠sbaars) 25 Mar '25

25 Mar '25

`WS_POPUP` for instance is `0x80000000`, which overflows a `LONG`. Coverity complained about this. The return type of `get_window_long()` already was `DWORD`. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7657

2 1

[PATCH v6 0/6] MR6866: ntdll: On macOS x86_64, swap GSBASE between the TEB and macOS TSD when entering/leaving PE code.
by Brendan Shanks (＠bshanks) 25 Mar '25

25 Mar '25

x86_64 Windows and macOS both use `%gs` to access thread-specific data (Windows TEB, macOS TSD). To date, Wine has worked around this conflict by filling the most important TEB fields (`0x30`/`Self`, `0x58`/`ThreadLocalStorage`) in the macOS TSD structure (Apple reserved the fields for our use). This was sufficient for most Windows apps. CrossOver's Wine had an additional hack to handle `0x60`/`ProcessEnvironmentBlock`, and binary patches for certain CEF binaries which directly accessed `0x8`/`StackBase`. Additionally, Apple's libd3dshared could activate a special mode in Rosetta 2 where code executing in certain regions would use the Windows TEB when accessing `%gs`. Now that the PE separation is complete, GSBASE can be swapped when entering/exiting PE code. This is done in the syscall dispatcher, unix-call dispatcher, and for user-mode callbacks. GSBASE also needs to be set to the macOS TSD when entering signal handlers (in `init_handler()`), and then restored to the Windows TEB when exiting (in `leave_handler()`). Some changes to the syscall dispatcher were needed to ensure that the TEB is not accessed through `%gs` while on the kernel stack (since a SIGUSR1 while on the kernel stack will result in GSBASE being set to the TSD). --- I've tested this successfully on macOS 15 (Apple Silicon and Intel) and macOS 10.13 with several apps and games, including the `cefclient.exe` CEF sample. Encouragingly, in some simple tests I didn't see a noticeable performance regression from this MR. There are drawbacks though: - libraries which jump directly from PE code into Unix code (expecting that %gs is always pointing to the macOS TSD) will crash. Notable examples are D3DMetal and DXMT. These will need to be changed to use Unix calls. - If Windows code uses the `syscall` instruction directly, the stack pointer likely needs to be valid (which is probably not true on Windows). This is due to the syscall dispatcher saving registers onto the user stack and having to call `_thread_set_tsd_base`. I can't say I've ever seen direct syscalls done with an invalid `%rsp`, but it seems like something anticheat code might do. --- macOS does not have a public API for setting GSBASE, but the private `_thread_set_tsd_base()` works and was added in macOS 10.12. `_thread_set_tsd_base()` is a small thunk that sets `%esi`, `%eax`, and does the `syscall`: https://github.com/apple-oss-distributions/xnu/blob/8d741a5de7ff4191bf97d57…. The syscall instruction itself clobbers `%rcx` and `%r11`. I've tried to save as few registers as possible when calling `_thread_set_tsd_base()`, but there may be room for improvement there. --- I also tested an alternate implementation strategy for this which took advantage of the expanded "full" thread state which is passed to signal handlers when a process has set a user LDT. The full thread state includes GSBASE, so GSBASE is set back to whatever is in the sigcontext on return (like every other field in the context). This would avoid needing to explicitly reset GSBASE in `leave_handler()`. This strategy was simpler, but I'm not using it for 2 reasons: - the "full" thread state is only available starting with macOS 10.15, and we still support 10.13. - more crucially, Rosetta 2 doesn't seem to correctly implement the GS.base field of the full thread state. It's set to 0 on entry, and isn't read on exit. -- v6: ntdll: Remove x86_64 Mac-specific TEB access workarounds that are no longer needed. ntdll: On macOS x86_64, swap GSBASE between the TEB and macOS TSD when entering/leaving PE code. ntdll: Set %rsp to be inside syscall_frame before accessing %gs in x86_64 syscall dispatcher. https://gitlab.winehq.org/wine/wine/-/merge_requests/6866

5 15

[PATCH v15 0/1] MR7607: win32u: Add stub for NtUserSetAdditionalForegroundBoostProcesses.
by Louis Lenders (＠xe) 25 Mar '25

25 Mar '25

Microsoft Edge worked if windows version is set to win8.1 (or win7) but now crashes in current wine due to this unimplemented function. This patch makes it work again in win8.1 mode; (It's still broken in win10 mode, see https://bugs.winehq.org/show_bug.cgi?id=56378 for more info) -- v15: win32u: Add stub for NtUserSetAdditionalForegroundBoostProcesses. https://gitlab.winehq.org/wine/wine/-/merge_requests/7607

2 1

[PATCH v14 0/2] MR7607: win32u: Add stub for NtUserSetAdditionalForegroundBoostProcesses.
by Louis Lenders (＠xe) 25 Mar '25

25 Mar '25

Microsoft Edge worked if windows version is set to win8.1 (or win7) but now crashes in current wine due to this unimplemented function. This patch makes it work again in win8.1 mode; (It's still broken in win10 mode, see https://bugs.winehq.org/show_bug.cgi?id=56378 for more info) -- v14: v2: handle handles inside array and rename variable. win32u: Add stub for NtUserSetAdditionalForegroundBoostProcesses. https://gitlab.winehq.org/wine/wine/-/merge_requests/7607

2 2

[PATCH v4 0/5] MR7540: kernel32/tests: Add tests for maximum path length limitation.
by Yongjie Yao (＠yaoyongjie) 25 Mar '25

25 Mar '25

In CopyFileEx, and DeleteFile functions, by default, the file name and path are limited to MAX_PATH characters. To extend this limit to 32,767 wide characters, need prepend "\\\\?\\" to the path. -- v4: kernelbase: Limit the maximum path length for DeleteFile. kernelbase: Fix DeleteFileA doesn't support long path. kernelbase: Limit the maximum path length for filesystem. ntdll: Load the value of longPathAware from manifest. kernel32/tests: Add tests for maximum path length limitation. https://gitlab.winehq.org/wine/wine/-/merge_requests/7540

2 7

[PATCH v2 0/3] MR7653: windowscodecs/metadata: Always initialize handler pointer.
by Nikolay Sivov (＠nsivov) 25 Mar '25

25 Mar '25

In case of empty block reader we currently return uninitialized pointer that is accessed later. Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> -- v2: windowscodecs/tests: Add some tests for the App0 reader. https://gitlab.winehq.org/wine/wine/-/merge_requests/7653

2 4

[PATCH v2 0/2] MR7649: mf/tests: Add timestamp tests for audio decoders.
by Brendan McGrath (＠redmcg) 25 Mar '25

25 Mar '25

This MR is similar to MR !7563 but targets the MFT audio decoders rather than the video decoders. -- v2: mf/tests: Add timestamp tests for wma audio decoder. mf/tests: Add timestamp tests for aac audio decoder. https://gitlab.winehq.org/wine/wine/-/merge_requests/7649

2 5

[PATCH 0/3] MR7653: windowscodecs/metadata: Always initialize handler pointer.
by Nikolay Sivov (＠nsivov) 25 Mar '25

25 Mar '25

In case of empty block reader we currently return uninitialized pointer that is accessed later. Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7653

3 4

[PATCH v2 0/2] MR7647: Add single point support for GdipGetPathWorldBounds
by Bartosz Kosiorek (＠gang65) 25 Mar '25

25 Mar '25

-- v2: gdiplus: Add single point support for GdipGetPathWorldBounds gdiplus/test: add tests for GdipGetPathWorldBounds with single point https://gitlab.winehq.org/wine/wine/-/merge_requests/7647

3 5

[PATCH v3 0/5] MR7540: kernel32/tests: Add tests for maximum path length limitation.
by Yongjie Yao (＠yaoyongjie) 25 Mar '25

25 Mar '25

In CopyFileEx, and DeleteFile functions, by default, the file name and path are limited to MAX_PATH characters. To extend this limit to 32,767 wide characters, need prepend "\\\\?\\" to the path. -- v3: kernelbase: Limit the maximum path length for DeleteFile. kernelbase: Fix DeleteFileA doesn't support long path. kernelbase: Limit the maximum path length for filesystem. user32: Load the value of longPathAware from manifest. kernel32/tests: Add tests for maximum path length limitation. https://gitlab.winehq.org/wine/wine/-/merge_requests/7540

3 8

[PATCH v5 0/1] MR7642: server: Fall back to epoll_wait if epoll_pwait2 isn't available at run-time.
by William Horvath (＠whrvt) 24 Mar '25

24 Mar '25

This fixes a real problem and improves behavior at the same time. - Fixes: hanging when trying to run Wine built with epoll_pwait2 headers available but epoll_pwait2 is missing in the run-time kernel (i.e. build 5.11->run 5.10) - Improvement: bring the improved timeout resolution to Wine builds which didn't have epoll_pwait2 at compile time, if the run-time kernel supports it (i.e. build 5.10->run 6.14) This last point is especially important in my opinion, as it applies to official Proton builds. Proton is (currently) built in Debian 11 with kernel 5.10, but SteamOS is a moving target with a much newer kernel version being used to run these Wine builds. These builds use the lower resolution timeouts if epoll_pwait2 is only checked for at compile-time, when it can cheaply and easily be done at run-time instead. Fixes: 87ca5db40e2c1b37423bdc25101a5c5e39e67e6f An alternative to !7640 that kills two birds with one stone. -- v5: server: Use dlsym() to check for epoll_pwait2 at run-time. https://gitlab.winehq.org/wine/wine/-/merge_requests/7642

4 6

[PATCH v5 0/6] MR7608: server: Avoid delegating console waits to other objects.
by Elizabeth Figura (＠zfigura) 24 Mar '25

24 Mar '25

Some console objects currently do several unique things: * Delegate waits onto the queue of another object. This is not really a problem for in-process waits, since we can just return the sync object for the delegate. However, it's also unnecessary, adds to the complexity of the server logic, and is one out of one places where this is done. * Make the wait state dependent on the process. This is difficult to emulate with ntsync and would require creating separate server objects for each process, hacking into duplicate_handle. * Fail a wait entirely in certain circumstances. This is pretty much impossible to emulate with in-process waits. Although ntsync has been in development for some time, I have regrettably failed to notice these problems until now. Fortunately, none of these behaviours happen on modern Windows. Although I/O on unbound handles delegates to the console of the current process, the signaled state does not. As the tests here show, the signaled state of a handle depends on the active console of the process in which the handle was created. If that console no longer exists, the signaled state is no longer updated [with one rather inexplicable exception]. Crucially, in current Windows waits never fail, and the state of an object is the same across all process which hold handles to it. Therefore this patch brings our behaviour to closer match current Windows. In theory these are fds and should use default_fd_signaled(). However, the points at which the handles are signaled are completely different, and I/O does not trigger console handles to become signaled when it normally would. Therefore for the time being I've kept the code using custom signaled ops. There is one other oddity related to consoles, which is the existence of console_add_queue(), which seeks to lazily create an input thread when a console is first waited on. This is one out of two places, after this patch, when the wait process is hijacked (the other being message queues). Fortunately this is easy to handle for in-process synchronization objects, by queueing the ioctl from the callback used to retrieve the in-process synchronization object itself. -- v5: server: Track unbound output signaled state based on its original console. server: Track unbound input signaled state based on its original console. server: Allow waiting on an orphaned screen buffer. server: Use a list of screen buffers per console. server: Fail to create an unbound input/output when there is no console. kernel32/tests: Add more tests for orphaned console handles. https://gitlab.winehq.org/wine/wine/-/merge_requests/7608

2 6

[PATCH v3 0/5] MR7636: d3dx9: Implement saving to DDS files in D3DXSaveTextureToFileInMemory().
by Connor McAdams (＠cmcadams) 24 Mar '25

24 Mar '25

The last of the saving patches for d3dx9. :) -- v3: d3dx9: Add support for saving IDirect3DVolumeTexture9 textures to DDS files in D3DXSaveTextureToFileInMemory(). d3dx9: Add support for saving IDirect3DCubeTexture9 textures to DDS files in D3DXSaveTextureToFileInMemory(). d3dx9: Add support for saving IDirect3DTexture9 textures to DDS files in D3DXSaveTextureToFileInMemory(). d3dx9/tests: Add more tests for saving textures to DDS files. https://gitlab.winehq.org/wine/wine/-/merge_requests/7636

3 12

Re: [PATCH v2 0/1] MR7637: wow64: skip memcpy for null pointer - approved
by Alexandre Julliard (＠julliard) 24 Mar '25

24 Mar '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7637

1 0

Re: [PATCH v2 0/1] MR7625: winex11: Use the state tracker for the desktop window _NET_WM_STATE. - approved
by Alexandre Julliard (＠julliard) 24 Mar '25

24 Mar '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7625

1 0

Re: [PATCH v2 0/1] MR7626: winex11: Use the current state when deciding how to reply to WM_TAKE_FOCUS. - approved
by Alexandre Julliard (＠julliard) 24 Mar '25

24 Mar '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7626

1 0

Re: [PATCH v3 0/2] MR7619: comctl32: Implement EVENT_OBJECT_VALUECHANGE for edit controls. - approved
by Alexandre Julliard (＠julliard) 24 Mar '25

24 Mar '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7619

1 0

Re: [PATCH v3 0/1] MR7640: server: Fall back to epoll_wait if epoll_pwait2 doesn't work - approved
by Alexandre Julliard (＠julliard) 24 Mar '25

24 Mar '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7640

1 0

Re: [PATCH v2 0/1] MR7651: msvcrt: Update file position in _flsbuf() in append mode. - approved
by Alexandre Julliard (＠julliard) 24 Mar '25

24 Mar '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7651

1 0

Re: [PATCH v2 0/5] MR7632: server, ntdll: Return the correct base/actual thread/process priority. - approved
by Alexandre Julliard (＠julliard) 24 Mar '25

24 Mar '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7632

1 0

Re: [PATCH v3 0/1] MR7622: comctl32: Avoid segfault in PROPSHEET_DoCommand when psInfo is NULL. - approved
by Alexandre Julliard (＠julliard) 24 Mar '25

24 Mar '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7622

1 0

Re: [PATCH v2 0/4] MR7562: wined3d: Move FFP extra states away from STATE_RENDER, part 1/4. - approved
by Alexandre Julliard (＠julliard) 24 Mar '25

24 Mar '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7562

1 0

Re: [PATCH v2 0/4] MR7648: user32/tests: Add some tests for current selection with LBS_NOSEL. - approved
by Alexandre Julliard (＠julliard) 24 Mar '25

24 Mar '25

This merge request was approved by Alexandre Julliard. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7648

1 0

[PATCH v38 0/8] MR7422: winebus/dinput: Support arbitrary number of axes in the virtual PID descriptor
by Tomasz Pakuła (＠TomaszPakula) 24 Mar '25

24 Mar '25

Related Wine issue: https://bugs.winehq.org/show_bug.cgi?id=52714 I'm starting work on updating/extending Linux force feedback API and one of my goals is to expose the number and an array that contains the device's ffb-enabled axes. As I'm mainly working with USB PID driver, I already have a POC of obtaining this data. What prompted this is that currently Wine always creates virtual joysticks with two PID ffb axes. This causes issues with [Richard Burns Rally](https://github.com/ValveSoftware/Proton/issues/6702#issuecomment-267…, as this game incorrectly initializes a `CONSTANT_FORCE` effect with all the FFB axes that contain `DIDOI_FFACTUATOR` flag and then trying to update `cAxes` and `rgdwAxes` values which leads to `DIERR_INVALIDPARAM`. Now, this flag is set by wine while enumerating virtual device's axes and while `(axis index < FFB axes)`, flag is applied. This means, that every device which has FFB functionality and at least two axes, will report FFB on `X`, `Y`. While updated API would mean a lot of steering wheels would correctly report only 1 axis, a lot of USB PID wheels still include `X` and `Y` in their `AXES_ENABLE` and `DIRECTION` usages. This, again, would lead to broken FFB (in affected games). On Windows, there's a possibility of overwriting some joystick capabilities with registry entries, which ends up removing `DIDOI_FFACTUATOR` flag from a selected axis. This doesn't work in Wine. This MR allows Wine to add an arbitrary number (up to `PID_AXES_MAX`) of axes to the PID descriptor, based on the value of Haptic Axes from SDL or from Linux FF api in the future (I'm working on it). Additionally, to work around RBR and other games with similar, wrong FFB handling, I introduced a [hint](https://github.com/libsdl-org/SDL/issues/12341) to SDL that allows a dynamic overwrite of the number of haptic axes. For the Linux API, I'll figure out a nice way to override when the number of axes will be exposed. With these changes, I was able to successfully get force feedback on my Moza Racing R9 with all axes usable. With additional traces in the dinput code, I confirmed that with one axis defined in the PID descriptor, the game created a constant force effect with just one axis and `cAxes = 1`. -- v38: dinput/tests: Add tests for 6-axis ff joystick https://gitlab.winehq.org/wine/wine/-/merge_requests/7422

2 9

[PATCH 0/8] MR7656: wined3d: Two unrelated fixes.
by Elizabeth Figura (＠zfigura) 24 Mar '25

24 Mar '25

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/7656

2 8

2025

2024

2023

2022

wine-gitlab