January 2024 - Wine-GitLab - list.winehq.org

[PATCH v6 0/5] MR561: vkd3d-utils: D3DReflect() and reflection tests, part 1/2.
by Zebediah Figura (＠zfigura) Jan. 23, 2024

Jan. 23, 2024

Unlike previous vkd3d-utils interfaces, ID3D12ShaderReflection is rather more large and complex, and will probably end up needing several new individual scan interfaces from vkd3d-shader, which are themselves not exactly trivial to design. Therefore, instead of implementing everything in vkd3d-shader and then hooking up the vkd3d-utils interfaces on top of that, this patch series copies the existing implementation of reflection and then begins the process of moving its implementation to vkd3d-shader. The primary motivation here is to add reflection crosstests (primarily for the benefit of the HLSL compiler) without being blocked on API design. Part 2 of this patch series does this. -- v6: tests: Test signature reflection via D3DReflect(). tests: Test VKD3D_SHADER_COMPILE_OPTION_BACKCOMPAT_MAP_SEMANTIC_NAMES. vkd3d-utils: Implement input and output signature reflection. vkd3d-utils: Add D3DReflect(). include: Add vkd3d_d3d12shader.idl. https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/561

3 6

Re: [PATCH v5 0/9] MR546: vkd3d-shader/spirv: Enable capability ShaderViewportIndexLayerEXT when needed.
by Henri Verbeet (＠hverbeet) Jan. 23, 2024

Jan. 23, 2024

> Isn't the solution there for Wine to be printing vkd3d-shader errors as ERR? I thought I remembered a discussion specifically about this before. Yes. My plan is essentially to wait for DXIL development to settle down a bit, and then do some cleanup to do just that, although I'd use WARN instead of ERR/FIXME. (I.e., errors in the input shader are generally not errors in vkd3d-shader.) -- https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/546#note_58655

1 0

Re: [PATCH v20 0/4] MR4661: gdiplus: Implement Gdiplus effects
by Esme Povirk (＠madewokherd) Jan. 23, 2024

Jan. 23, 2024

Esme Povirk (@madewokherd) commented about dlls/gdiplus/image.c: > { > TRACE("(%p)\n", effect); > > + if (effect->params) > + free(effect->params); Don't check for NULL if you're only calling `free`. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/4661#note_58654

1 0

Re: [PATCH v20 0/4] MR4661: gdiplus: Implement Gdiplus effects
by Esme Povirk (＠madewokherd) Jan. 23, 2024

Jan. 23, 2024

Esme Povirk (@madewokherd) commented about dlls/gdiplus/tests/image.c: > expect(InvalidParameter, stat); > > + stat = pGdipSetEffectParameters(effect, params, 0); > + expect(InvalidParameter, stat); > + > effect = (CGpEffect *)0xdeadbeef; > stat = pGdipCreateEffect(noneffect, &effect); > expect(Win32Error, stat); > ok(effect == NULL, "Expected effect to be NULL\n"); > > + stat = pGdipSetEffectParameters(effect, (void *)0xdeadbeef, 0); > + expect(InvalidParameter, stat); > + stat = pGdipSetEffectParameters(effect, (void *)0xdeadbeef, 8); > + expect(InvalidParameter, stat); > + stat = pGdipSetEffectParameters(effect, (void *)0xdeadbeef, 5); > + expect(InvalidParameter, stat); Don't pass in effect by name if it's expected to be NULL. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/4661#note_58653

1 0

Re: [PATCH v20 0/4] MR4661: gdiplus: Implement Gdiplus effects
by Esme Povirk (＠madewokherd) Jan. 23, 2024

Jan. 23, 2024

Esme Povirk (@madewokherd) commented about dlls/gdiplus/tests/image.c: > stat = pGdipCreateEffect(BlurEffectGuid, NULL); > expect(InvalidParameter, stat); > > + stat = pGdipGetEffectParameterSize(NULL, NULL); > + expect(InvalidParameter, stat); > + > + stat = pGdipGetEffectParameterSize(effect, NULL); `effect` is NULL here, so this test is redundant. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/4661#note_58652

1 0

[PATCH v11 0/1] MR4720: Implement test_continue() for amd64
by Dāvis Mosāns Jan. 23, 2024

Jan. 23, 2024

Currently there is `NtContinue()` test only for `aarch64` so implement it for amd64 aswell. This implementation is very similar to `aarch64` and it's very basic test. It doesn't change/test `ContextFlags` so it won't catch https://bugs.winehq.org/show_bug.cgi?id=56050 but that can be implemented later on top of this. -- v11: ntdll/tests: Implement test_continue() for amd64 https://gitlab.winehq.org/wine/wine/-/merge_requests/4720

2 3

[PATCH v4 0/4] MR4761: ntdll: Implement NtContinueEx()
by Dāvis Mosāns Jan. 23, 2024

Jan. 23, 2024

This MR implements `NtContinueEx(PCONTEXT, PCONTINUE_OPTIONS)` which was added in Windows 10 20H1. Also added basic test reusing existing `test_continue()` (included from !4720) League of Legends game hooks `NtContinue()` and `NtContinueEx()` by putting `jmp` instruction in front of it. Note that LoL doesn't actually use `NtContinueEx()` itself and game will work fine without it but if game doesn't find `NtContinueEx()` it will permanently ban your account due to detecting "scripting program" -- v4: ntdll/tests: Add basic test for NtContinueEx() ntdll: NtContinueEx() handle few error cases ntdll: Implement NtContinueEx() ntdll/tests: Implement test_continue() for amd64 https://gitlab.winehq.org/wine/wine/-/merge_requests/4761

3 10

[PATCH 0/2] MR4883: ntoskrnl: Use internal name for asm exports.
by Gabriel Ivăncescu Jan. 23, 2024

Jan. 23, 2024

Since they're asm functions they conflict when building with LTO (among other things). This is normally hidden when separating translation units, but it's still not correct. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/4883

2 17

[PATCH v5 0/2] MR4022: wined3d: Expose the image view usage to the Vulkan driver.
by Giovanni Mascellani (＠giomasce) Jan. 23, 2024

Jan. 23, 2024

Some driver, e.g. MoltenVK, fail in some cases if the image view usage is too broad or unknown. -- v5: wined3d: Expose the image view usage for non-default views. https://gitlab.winehq.org/wine/wine/-/merge_requests/4022

3 4

Re: [PATCH v2 0/1] MR4834: ntdll: NtQuerySystemInformation(SystemProcessorPerformanceInformation) fix return value
by Dāvis Mosāns Jan. 23, 2024

Jan. 23, 2024

On Tue Jan 23 13:57:14 2024 +0000, Fabian Maurer wrote: > It copies but then gives `STATUS_INFO_LENGTH_MISMATCH`? And returning > `STATUS_ACCESS_VIOLATION` has priority over `STATUS_INFO_LENGTH_MISMATCH`? > Not saying you're wrong, just wondering. Would you mind adding tests for that? This is current behavior on Windows 10 22H2: ``` NULL buffer and size = 0 -> STATUS_INFO_LENGTH_MISMATCH, ReturnLength=cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) NULL buffer and size > 0 -> STATUS_ACCESS_VIOLATION, ReturnLength=0 valid buffer and size = 0 -> STATUS_INFO_LENGTH_MISMATCH, ReturnLength=cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) valid buffer and size > 0 -> STATUS_INFO_LENGTH_MISMATCH, ReturnLength=cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) valid buffer and size < 1*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) -> STATUS_INFO_LENGTH_MISMATCH, ReturnLength=cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) valid buffer and size = 1*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) -> STATUS_SUCCESS, ReturnLength= 1*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) valid buffer and size > 1*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) -> STATUS_INFO_LENGTH_MISMATCH, ReturnLength=cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) valid buffer and size = 2*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) -> STATUS_SUCCESS, ReturnLength= 2*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) valid buffer and size < cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) -> STATUS_INFO_LENGTH_MISMATCH, ReturnLength=cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) valid buffer and size = cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) -> STATUS_SUCCESS, ReturnLength=cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) valid buffer and size > cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) -> STATUS_INFO_LENGTH_MISMATCH, ReturnLength=cpus*sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) ``` Wine has few differences regarding this but here this MR fixes only single case ``` size % sizeof(SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION) -> STATUS_SUCCESS ``` Handling of `STATUS_ACCESS_VIOLATION` and `ReturnLength` is not changed. And regarding test, actually Wine already has such test, see [info.c#752](https://gitlab.winehq.org/wine/wine/-/blob/master/dlls/ntdll/te… ```c ok( status == STATUS_SUCCESS || status == STATUS_INFO_LENGTH_MISMATCH /* vista */, "Expected STATUS_SUCCESS or STATUS_INFO_LENGTH_MISMATCH, got %08lx\n", status); ``` It looks like before Vista it returned `STATUS_SUCCESS` which is current Wine implementation, but since Vista it started to return `STATUS_INFO_LENGTH_MISMATCH` instead -- https://gitlab.winehq.org/wine/wine/-/merge_requests/4834#note_58632

1 0