- wine-gitlab - winehq.org

[PATCH 0/5] MR2629: msvcrt: Use remaining math functions from the bundled musl library.
by Alexandre Julliard (＠julliard) 11 Apr '23

11 Apr '23

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/2629

3 6

[PATCH v3 0/1] MR2622: ntdll: Use log-linear bucketing for free lists.
by Tatsuyuki Ishi (＠ishitatsuyuki) 11 Apr '23

11 Apr '23

Currently, the free list consists of a "small list" for sizes below 256, which are linearly spaced, and a "large list" which is manually split into a few chunks. This patch replaces it with a single log-linear policy, while expanding the range the large list covers. The old implementation had issues when a lot of large allocations happened. In this case, all the allocations went in the last catch-all bucket in the "large list", and what happens is: 1. The linked list grew in size over time, causing searching cost to skyrocket. 2. With the first-fit allocation policy, fragmentation was also making the problem worse. The new bucketing covers the entire range up until we start allocating large blocks, which will not enter the free list. It also makes the allocation policy closer to best-fit (although not exactly), reducing fragmentation. The increase in number of free lists does incur some cost when it needs to be skipped over, but the improvement in allocation performance outweighs it. For future work, these ideas (mostly from glibc) might or might not benefit performance: - Use an exact best-fit allocation policy. - Add a bitmap for freelist, allowing empty lists to be skipped with a single bit scan. For the benchmark, this drastically improves initial shader loading performance in Overwatch 2. In this workload 78k shaders are passed to DXVK for DXBC -> SPIRV translation, and for each shader a few allocation happens in the 4K – 100K range for the staging buffer. Before this patch, malloc consisted a whooping 43% of overhead. The overhead with log-linear bucketing is drastically lower, resulting in a ~2x improvement in loading time. Overhead for each `FREE_LIST_LINEAR_BITS` is as below: - 0: 7.7% - 1: 2.9% - 2: 1.3% - 3: 0.6% Since performance seems to scale linearly with increase in buckets (up to the point I have tested), I've opted for 3 (8 buckets per doubling) in the current revision of patch. Signed-off-by: Tatsuyuki Ishi <ishitatsuyuki(a)gmail.com> -- v3: ntdll: Use log-linear bucketing for free lists. https://gitlab.winehq.org/wine/wine/-/merge_requests/2622

4 4

[PATCH v5 0/2] MR141: vkd3d-shader/ir: Normalise hull shader control point phase I/O.
by Conor McCarthy (＠cmccarthy) 11 Apr '23

11 Apr '23

To simplify SM 6 support, insert a control point id relative address where needed, and declare control point phase inputs where missing. -- v5: vkd3d-shader/ir: Insert hull shader control point input declarations if no control point phase is defined. https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/141

2 2

[PATCH v4 0/2] MR2032: winex11: Allow replacing either context in wglShareLists.
by Alex Henrie (＠alexhenrie) 11 Apr '23

11 Apr '23

The deleted test is a bad test: The behavior is different on different drivers, for example, the test fails on the Windows 10 and 11 testbot machines that have AMD video cards. Furthermore, MSDN does not say that the destination context "must not" have any display lists yet but rather that it "should not" have any.[1] The Khronos OpenGL Wiki similarly advises against calling wglShareLists after the source or destination context has at least one object, but if you do, it only says that "there is a chance that wglShareLists will fail", not that it will necessarily fail.[2] Since there's no clear "right" behavior here, we can adopt the more permissive behavior that some programs expect, as long as it doesn't corrupt the context. [1] https://learn.microsoft.com/en-us/windows/win32/api/wingdi/nf-wingdi-wglsha… [2] https://www.khronos.org/opengl/wiki/Platform_specifics:_Windows#wglShareLis… Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=11436 -- v4: winex11: Allow replacing either context in wglShareLists. opengl32/tests: Make the wglShareLists tests comprehensive. https://gitlab.winehq.org/wine/wine/-/merge_requests/2032

3 6

[PATCH v2 0/1] MR2591: include: _InterlockedExchangePointer and _InterlockedCompareExchangePointer...
by Stefan Dösinger (＠stefan) 11 Apr '23

11 Apr '23

include: _InterlockedExchangePointer and _InterlockedCompareExchangePointer are intrinsics in x86 msvc. I fixed this issue in ad05f33d67, but a40973f20 regressed this again. I was carrying a patch for quite a while, feeling dejavu. The msvc ver of 1900 is taken from Boost's interlocked.hpp, which matches MSVC 2015 (toolset version v140). Boost has a comment that claims that in msvc 2012 those functions were defined in intrin.h, but those defines are broken with Microsoft's winnt.h. -- v2: include: x86 msvc has _InterlockedExchangePointer and _InterlockedCompareExchangePointer https://gitlab.winehq.org/wine/wine/-/merge_requests/2591

2 1

[PATCH 0/8] MR2627: imm32: Implement input context cache and use INPUTCONTEXT directly in some functions.
by Rémi Bernon (＠rbernon) 11 Apr '23

11 Apr '23

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/2627

3 10

[PATCH v2 0/1] MR2622: ntdll: Use log-linear bucketing for free lists.
by Tatsuyuki Ishi (＠ishitatsuyuki) 11 Apr '23

11 Apr '23

Currently, the free list consists of a "small list" for sizes below 256, which are linearly spaced, and a "large list" which is manually split into a few chunks. This patch replaces it with a single log-linear policy, while expanding the range the large list covers. The old implementation had issues when a lot of large allocations happened. In this case, all the allocations went in the last catch-all bucket in the "large list", and what happens is: 1. The linked list grew in size over time, causing searching cost to skyrocket. 2. With the first-fit allocation policy, fragmentation was also making the problem worse. The new bucketing covers the entire range up until we start allocating large blocks, which will not enter the free list. It also makes the allocation policy closer to best-fit (although not exactly), reducing fragmentation. The increase in number of free lists does incur some cost when it needs to be skipped over, but the improvement in allocation performance outweighs it. For future work, these ideas (mostly from glibc) might or might not benefit performance: - Use an exact best-fit allocation policy. - Add a bitmap for freelist, allowing empty lists to be skipped with a single bit scan. Signed-off-by: Tatsuyuki Ishi <ishitatsuyuki(a)gmail.com> -- v2: ntdll: Use log-linear bucketing for free lists. https://gitlab.winehq.org/wine/wine/-/merge_requests/2622

5 8

[PATCH v2 0/2] MR2620: include: Add windows.ui.composition.interop.idl file.
by Mohamad Al-Jaf (＠maljaf) 11 Apr '23

11 Apr '23

Needed for mingw Firefox build. The WIDL error points to the wrong function, specifically the one after it. include/windows.ui.composition.interop.idl:35:63: error: parameter 'swapchain' of function 'CreateCompositionSurfaceForHandle' cannot derive from void * HRESULT CreateCompositionSurfaceForSwapChain([in] IUnknown *swapchain, [out, retval] ICompositionSurface **result); ^ make[1]: *** [Makefile:163749: include/windows.ui.composition.interop.h] Error 1 -- v2: include: Add windows.ui.composition.interop.idl file. widl: Add support for WinRT HANDLE parameter type. https://gitlab.winehq.org/wine/wine/-/merge_requests/2620

4 6

[PATCH v4 0/2] MR2546: winegstreamer: Create extra decodebin to decode compressed stream.
by Ziqing Hui (＠zhui) 11 Apr '23

11 Apr '23

-- v4: wingstreamer: Create extra decodebin to decode compressed stream. winegstreamer: Implement wg_format_from_caps_video_wmv. https://gitlab.winehq.org/wine/wine/-/merge_requests/2546

2 2

[PATCH 0/4] MR2626: wine{alsa,coreaudio,oss,pulse}: Move common mmdevdrv structs into mmdevapi.
by Davide Beatrici (＠davidebeatrici) 11 Apr '23

11 Apr '23

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/2626

3 11

2025

2024

2023

2022

wine-gitlab