September 2023 - Wine-GitLab

[PATCH v2 0/5] MR2781: mfplat: Optimize sample copy to d3d buffer.
by Paul Gofman (＠gofman) Oct. 31, 2023

Oct. 31, 2023

This is ultimately a follow up for introducing d3d manager support in h264 decoder transform. The last patch avoids unneeded GPU to CPU copy when copying a memory sample to GPU sample (in particular, when sample copier is involved). There are a couple of things to fix on the way. Patch 1 fixes the crash which otherwise happens in the test introduced in patch 2 due to memory corruption. Flipping the image instead of using absolute pitch will break the test in patch 2. Patch 3 contradicts to what MS docs says [1] for MFCreate2DMediaBuffer, fBottomUp parameter: "If TRUE, the buffer's IMF2DBuffer::ContiguousCopyTo method copies the buffer into a bottom-up format.". That doesn't match the tests in patch 2. So far I see that fBottomUp parameter only affects what is returned from Lock2D as scanline0 and pitch. Thinking of it, the behaviour of ContiguousCopyFrom makes some sense to me as Lock2D (unlike Lock which converts to contiguous buffer) is not supposed to perform any copies and flips and return the underlying representation as is. And once scanline0 and pitch for that 2d buffer indicate it should be copied bottom up additionally flipping when converting from contiguous buffer wouldn't make much sense. Then, ContiguousCopyTo could flip it actually as the docs suggest, but the test seems to clearly show that it doesn't happen (there is a separate test for clarity showing that ContiguousCopyFrom / ContiguousCopyTo yields the same data). 1. https://learn.microsoft.com/en-us/windows/win32/api/mfapi/nf-mfapi-mfcreate… -- v2: mfplat/sample: Optimize copying to 2d buffer. mfplat/sample: Refactor sample_CopyToBuffer(). mfplat/buffer: Do not flip in memory_2d_buffer_ContiguousCopy{From|To}(). mfplat/tests: Add more tests for copying 2d buffers. mfplat/buffer: Use absolute pitch in memory_1d_2d_buffer_[Un]Lock(). https://gitlab.winehq.org/wine/wine/-/merge_requests/2781

3 6

[PATCH v3 0/2] MR2383: shell32: Avoid stack corruption caused by long URLs
by Stefan Rentsch (＠restet) Oct. 22, 2023

Oct. 22, 2023

Previously long URI/URLs (> 1024 characters) would cause a stack corruption and crash Wine. After this MR, Wine will no longer crash but the URL is not opened either primarily due to `SHELL_Argify` limitations. -- v3: shell32: Increase timeout for long URLs https://gitlab.winehq.org/wine/wine/-/merge_requests/2383

3 4

[PATCH v10 0/4] MR1088: ntdll: Avoid calling pthread_exit on thread exit.
by Rémi Bernon Oct. 20, 2023

Oct. 20, 2023

Using a dedicated exit jmpbuf and removing the need for assembly routines. When Wine handles an exception in unix code, we return to user mode by jumping to the last syscall frame. This can leave some pthread cancel cleanups registered, in the pthread internal linked list, and at the same time later overwrite the stack frame they were registered for. In the same way, jumping to the exit frame on thread exit or abort, can also leave some cleanup handlers registered for invalid stack frames. Depending on the implementation, calling pthread_exit will cause all the registered pthread cleanup handlers to be called, possibly jumping back to now overwritten stack frames and causing segmentation faults. Exiting a pthread normally, by returning from its procedure, or calling exit(0) for the main thread doesn't run pthread_exit and doesn't call cleanup handlers, avoiding that situation. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=52213 ### Additional note: For robustness, we should probably try to execute these cleanup handlers when unwinding the stack frames, as we would otherwise leave pthread objects in a potential problematic state (like a mutex locked, etc). It is however hard to do so when the handlers are registered from some C code: pthread C implementation is done by calling some internal pthread functions to register the handlers, and they aren't registered as standard unwind handlers. Only pthread_cancel and pthread_exit can unwind and call / unregister the C handlers, but interrupting that procedure, for instance calling setjmp / longjmp from withing our own handler isn't supported. >From C++ code, pthread cleanup handlers are registered through C++ class constructors / destructors, and it would then be possible to partially unwind and call them at the same time. -- v10: ntdll: Remove now unnecessary arch-specific exit frame. ntdll: Avoid calling pthread_exit from abort_thread. ntdll: Avoid calling abort_thread from the signal stack. ntdll: Use is_inside_syscall more consistently. https://gitlab.winehq.org/wine/wine/-/merge_requests/1088

2 9

[PATCH v3 0/4] MR3729: crypt32: Avoid truncating unicode chars.
by Dmitry Timoshkov (＠dmitry) Oct. 20, 2023

Oct. 20, 2023

-- v3: cryptui: Use CertNameToStr(CERT_NAME_STR_NO_QUOTING_FLAG) to disable quoting. crypt32: Add support for CERT_NAME_STR_NO_QUOTING_FLAG to CertNameToStr(). crypt32/tests: Add some tests for quoted RDN values. crypt32: Avoid truncating unicode chars. https://gitlab.winehq.org/wine/wine/-/merge_requests/3729

2 5

[PATCH v2 0/1] MR3777: gdi32: Add stub for D3DKMTQueryAdapterInfo.
by Etaash Mathamsetty (＠etaash.mathamsetty) Oct. 16, 2023

Oct. 16, 2023

Rainbow 6 Siege and some other games require this. Type is printed in the FIXME to determine which type the application requires. -- v2: gdi32: Add stub for D3DKMTQueryAdapterInfo. https://gitlab.winehq.org/wine/wine/-/merge_requests/3777

4 3

[PATCH 0/3] MR3839: kernel32: Add structure size checks to CreateActCtx().
by Dmitry Timoshkov (＠dmitry) Oct. 16, 2023

Oct. 16, 2023

Signed-off-by: Dmitry Timoshkov <dmitry(a)baikal.ru> -- https://gitlab.winehq.org/wine/wine/-/merge_requests/3839

4 12

[PATCH 0/2] MR343: Draft: vkd3d-shader/hlsl: Make deref->data_type available for the whole deref's lifetime.
by Francisco Casas (＠fcasas) Oct. 10, 2023

Oct. 10, 2023

I marked this a draft because it may be a somewhat orthogonal step, but IRRC @giomasce suggested it. While a deref's data type can be computed from the variable's data type as long as the deref is still represented as a path (as we did with hlsl_deref_get_type()), we need the deref.data_type field for when the deref is represented as an offset. While we could get rid of this deref.data_type in the future, if we manage to transform the HLSL IR instructions directly to VSIR withot requiring lowering the derefs to their offset representation, this would take a while. So, this patch makes the deref.data_type field available during the whole lifetime of the deref. Which makes deref.data_type easier to understand (since it can be used anytime now) and we no longer have to call hlsl_deref_get_type(). -- https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/343

4 8

[PATCH v2 0/6] MR3810: winegstreamer: Implement add_stream and push_sample for muxer.
by Ziqing Hui (＠zhui) Oct. 9, 2023

Oct. 9, 2023

-- v2: winegstreamer: Implement wg_muxer_push_sample. winegstreamer: Implement ProcessSample for media sink. winegstreamer: Implement wg_muxer_start. winegstreamer: Implement wg_muxer_add_stream. winegstreamer: Introduce link_src_to_sink. winegstreamer: Use NTSTATUS value in wg_muxer_create. https://gitlab.winehq.org/wine/wine/-/merge_requests/3810

3 19

[PATCH v2 0/1] MR3955: mf: Avoid a double free of presentation clock timers.
by Zhiyi Zhang (＠zhiyi) Oct. 9, 2023

Oct. 9, 2023

In clock_change_state() when a clock is running, a timer is removed from clock->timers. The same timer is then used to create an async result, which will eventually calls present_clock_timer_callback_Invoke() and removes the same timer. -- v2: mf: Avoid a double free of presentation clock timers. https://gitlab.winehq.org/wine/wine/-/merge_requests/3955

3 4

[PATCH v4 0/2] MR3806: combase: Create implicit MTA in STA apartment in RoGetActivationFactory().
by Paul Gofman (＠gofman) Oct. 9, 2023

Oct. 9, 2023

Fixes Starfield not being able to take photos in photo mode (due to failing to creating windowscodecs' image factory due to COM apartment being initialized). Turns out on Windows RoGetActivationFactory() initializes implicit MTA apartment when called from STA thread, and so after that called once COM is implicitly uninitialized for any (new) thread until COM is uninitialized in the thread which called RoGetActivationFactory (or that thread exited). This is not the case on Win8 (where the function was first introduced) but looks consistent after that. -- v4: combase: Ensure MTA existence in RoGetActivationFactory(). combase/tests: Add tests for implicit MTA with RoGetActivationFactory(). https://gitlab.winehq.org/wine/wine/-/merge_requests/3806

3 4