[PATCH v9 0/10] MR531: Draft: vkd3d: Assorted fixes for MoltenVK.
-- v9: tests: Mark loading from stencil as buggy on MoltenVK. tests: Mark resource arrays as buggy on MoltenVK. tests: Mark unbounded ranges as buggy on MoltenVK. tests: Mark predicated rendering as buggy on MoltenVK. tests: Mark streaming output as buggy on MoltenVK. tests: Mark some sampling tests as buggy on MoltenVK. vkd3d-shader/spirv: Specify behavior for bit field instructions. tests: Use a uint target for rendering uint data. vkd3d: Specify the aspect when creating NULL UAVs. ci: Dumps the d3d12 summary line in the CI log. https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
From: Giovanni Mascellani <gmascellani(a)codeweavers.com> --- gitlab/build-mac | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/gitlab/build-mac b/gitlab/build-mac index 17a8c7a0d..15e398d51 100755 --- a/gitlab/build-mac +++ b/gitlab/build-mac @@ -19,6 +19,10 @@ else touch ../build_failed fi +# d3d12 lumps many tests, so it's useful to have the summary line in +# the CI log +cat tests/d3d12.log | egrep 'd3d12: [0-9]+ tests executed' || true + mkdir -p ../artifacts/$COMMIT rsync -Rr config.log test-suite.log tests/*.log tests/*/*.log ../artifacts/$COMMIT || true -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
From: Giovanni Mascellani <gmascellani(a)codeweavers.com> This fixes a crash on MoltenVK. --- libs/vkd3d/resource.c | 1 + 1 file changed, 1 insertion(+) diff --git a/libs/vkd3d/resource.c b/libs/vkd3d/resource.c index 609c67102..356c2f099 100644 --- a/libs/vkd3d/resource.c +++ b/libs/vkd3d/resource.c @@ -3444,6 +3444,7 @@ static void vkd3d_create_null_uav(struct d3d12_desc *descriptor, vkd3d_desc.miplevel_count = 1; vkd3d_desc.layer_idx = 0; vkd3d_desc.layer_count = 1; + vkd3d_desc.vk_image_aspect = VK_IMAGE_ASPECT_COLOR_BIT; vkd3d_desc.components.r = VK_COMPONENT_SWIZZLE_R; vkd3d_desc.components.g = VK_COMPONENT_SWIZZLE_G; vkd3d_desc.components.b = VK_COMPONENT_SWIZZLE_B; -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
From: Giovanni Mascellani <gmascellani(a)codeweavers.com> This fixes a failure in MoltenVK. --- tests/d3d12.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/d3d12.c b/tests/d3d12.c index 1126d9749..3324eef80 100644 --- a/tests/d3d12.c +++ b/tests/d3d12.c @@ -35376,7 +35376,7 @@ static void test_bufinfo_instruction(void) memset(&desc, 0, sizeof(desc)); desc.rt_width = desc.rt_height = 64; - desc.rt_format = DXGI_FORMAT_R32G32B32A32_FLOAT; + desc.rt_format = DXGI_FORMAT_R32G32B32A32_UINT; desc.no_root_signature = true; desc.no_pipeline = true; if (!init_test_context(&context, &desc)) -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
From: Giovanni Mascellani <gmascellani(a)codeweavers.com> Bit field instructions in SPIR-V do not specify what happens when offset + count exceeds the type bit width. After this commit we refine the emitted code's behavior to match TPF. This fixes a few failures on MoltenVK. --- libs/vkd3d-shader/spirv.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/libs/vkd3d-shader/spirv.c b/libs/vkd3d-shader/spirv.c index 45f46a91c..0e961293c 100644 --- a/libs/vkd3d-shader/spirv.c +++ b/libs/vkd3d-shader/spirv.c @@ -1764,6 +1764,15 @@ static uint32_t vkd3d_spirv_build_op_glsl_std450_max(struct vkd3d_spirv_builder GLSLstd450NMax, operands, ARRAY_SIZE(operands)); } +static uint32_t vkd3d_spirv_build_op_glsl_std450_umin(struct vkd3d_spirv_builder *builder, + uint32_t result_type, uint32_t x, uint32_t y) +{ + uint32_t glsl_std450_id = vkd3d_spirv_get_glsl_std450_instr_set(builder); + uint32_t operands[] = {x, y}; + return vkd3d_spirv_build_op_ext_inst(builder, result_type, glsl_std450_id, + GLSLstd450UMin, operands, ARRAY_SIZE(operands)); +} + static uint32_t vkd3d_spirv_build_op_glsl_std450_nclamp(struct vkd3d_spirv_builder *builder, uint32_t result_type, uint32_t x, uint32_t min, uint32_t max) { @@ -7321,7 +7330,7 @@ static void spirv_compiler_emit_ftou(struct spirv_compiler *compiler, static void spirv_compiler_emit_bitfield_instruction(struct spirv_compiler *compiler, const struct vkd3d_shader_instruction *instruction) { - uint32_t src_ids[4], constituents[VKD3D_VEC4_SIZE], type_id, mask_id; + uint32_t src_ids[4], constituents[VKD3D_VEC4_SIZE], type_id, mask_id, size_id, max_count_id; struct vkd3d_spirv_builder *builder = &compiler->spirv_builder; const struct vkd3d_shader_dst_param *dst = instruction->dst; const struct vkd3d_shader_src_param *src = instruction->src; @@ -7336,6 +7345,7 @@ static void spirv_compiler_emit_bitfield_instruction(struct spirv_compiler *comp component_type = vkd3d_component_type_from_data_type(dst->reg.data_type); type_id = vkd3d_spirv_get_type_id(builder, component_type, 1); mask_id = spirv_compiler_get_constant_uint(compiler, 0x1f); + size_id = spirv_compiler_get_constant_uint(compiler, 0x20); switch (instruction->handler_idx) { @@ -7364,6 +7374,9 @@ static void spirv_compiler_emit_bitfield_instruction(struct spirv_compiler *comp { src_ids[j] = vkd3d_spirv_build_op_and(builder, type_id, src_ids[j], mask_id); } + max_count_id = vkd3d_spirv_build_op_isub(builder, type_id, size_id, src_ids[src_count - 2]); + src_ids[src_count - 1] = vkd3d_spirv_build_op_glsl_std450_umin(builder, type_id, + src_ids[src_count - 1], max_count_id); constituents[k++] = vkd3d_spirv_build_op_trv(builder, &builder->function_stream, op, type_id, src_ids, src_count); -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
From: Giovanni Mascellani <gmascellani(a)codeweavers.com> Apparently Metal doesn't support specifying a bias directly in the sampler, and, with "nearest" mip filtering, it doesn't switch precisely at LOD 0.5 (though still between 0.5 and 0.6). --- tests/d3d12.c | 38 ++++++++++++++++++++++++++------------ 1 file changed, 26 insertions(+), 12 deletions(-) diff --git a/tests/d3d12.c b/tests/d3d12.c index 3324eef80..fb0d4f7a5 100644 --- a/tests/d3d12.c +++ b/tests/d3d12.c @@ -13685,7 +13685,7 @@ static void test_sample_instructions(void) unsigned int x_step, y_step; ID3D12CommandQueue *queue; ID3D12Resource *texture; - unsigned int i, x, y; + unsigned int i, x = 0, y; ID3D12Device *device; HRESULT hr; @@ -13908,6 +13908,7 @@ static void test_sample_instructions(void) float max_lod; float ps_constants[4]; const unsigned int *expected_data; + bool bug_on_mvk; } tests[] = { @@ -13919,7 +13920,7 @@ static void test_sample_instructions(void) {&ps_sample_b, &r8g8b8a8_texture, POINT, 0.0f, 0.0f, MIP_MAX, {0.0f}, r8g8b8a8_data}, {&ps_sample_b, &a8_texture, POINT, 0.0f, 0.0f, MIP_MAX, {0.0f}, a8_expected_data}, {&ps_sample_b, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {0.0f}, rgba_level_0}, - {&ps_sample_b, &rgba_texture, POINT, 8.0f, 0.0f, MIP_MAX, {0.0f}, level_1_colors}, + {&ps_sample_b, &rgba_texture, POINT, 8.0f, 0.0f, MIP_MAX, {0.0f}, level_1_colors, true}, {&ps_sample_b, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {8.0f}, level_1_colors}, {&ps_sample_b, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {8.4f}, level_1_colors}, {&ps_sample_b, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {8.5f}, level_2_colors}, @@ -13937,19 +13938,21 @@ static void test_sample_instructions(void) {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {-1.0f}, rgba_level_0}, {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {0.0f}, rgba_level_0}, {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {0.4f}, rgba_level_0}, - {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {0.5f}, level_1_colors}, + {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {0.5f}, level_1_colors, true}, + {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {0.6f}, level_1_colors}, {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {1.0f}, level_1_colors}, {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {1.4f}, level_1_colors}, - {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {1.5f}, level_2_colors}, + {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {1.5f}, level_2_colors, true}, + {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {1.6f}, level_2_colors}, {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {2.0f}, level_2_colors}, {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {3.0f}, level_2_colors}, {&ps_sample_l, &rgba_texture, POINT, 0.0f, 0.0f, MIP_MAX, {4.0f}, level_2_colors}, {&ps_sample_l, &rgba_texture, POINT_LINEAR, 0.0f, 0.0f, MIP_MAX, {1.5f}, lerp_1_2_colors}, {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 0.0f, MIP_MAX, {-2.0f}, rgba_level_0}, - {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 0.0f, MIP_MAX, {-1.0f}, level_1_colors}, - {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 0.0f, MIP_MAX, {0.0f}, level_2_colors}, - {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 0.0f, MIP_MAX, {1.0f}, level_2_colors}, - {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 0.0f, MIP_MAX, {1.5f}, level_2_colors}, + {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 0.0f, MIP_MAX, {-1.0f}, level_1_colors, true}, + {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 0.0f, MIP_MAX, {0.0f}, level_2_colors, true}, + {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 0.0f, MIP_MAX, {1.0f}, level_2_colors, true}, + {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 0.0f, MIP_MAX, {1.5f}, level_2_colors, true}, {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 2.0f, 2.0f, {-9.0f}, level_2_colors}, {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 2.0f, 2.0f, {-1.0f}, level_2_colors}, {&ps_sample_l, &rgba_texture, POINT_LINEAR, 2.0f, 2.0f, 2.0f, {0.0f}, level_2_colors}, @@ -14015,6 +14018,9 @@ static void test_sample_instructions(void) for (i = 0; i < ARRAY_SIZE(tests); ++i) { + unsigned int color = 0; + bool fail = false; + vkd3d_test_push_context("Test %u", i); memset(&sampler_desc, 0, sizeof(sampler_desc)); @@ -14070,12 +14076,20 @@ static void test_sample_instructions(void) { for (x = 0; x < tests[i].texture->width; ++x) { - unsigned int color = get_readback_uint(&rb.rb, x * x_step + x_step / 2, y * y_step + y_step / 2, 0); - ok(compare_color(color, tests[i].expected_data[tests[i].texture->width * y + x], 1), - "Got color 0x%08x, expected 0x%08x at (%u, %u).\n", - color, tests[i].expected_data[tests[i].texture->width * y + x], x, y); + color = get_readback_uint(&rb.rb, x * x_step + x_step / 2, y * y_step + y_step / 2, 0); + if (!compare_color(color, tests[i].expected_data[tests[i].texture->width * y + x], 1)) + { + fail = true; + break; + } } + + if (fail) + break; } + bug_if(tests[i].bug_on_mvk && is_mvk_device(device)) + ok(!fail, "Got color 0x%08x, expected 0x%08x at (%u, %u).\n", + color, tests[i].expected_data[tests[i].texture->width * y + x], x, y); release_resource_readback(&rb); ID3D12Resource_Release(texture); -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
From: Giovanni Mascellani <gmascellani(a)codeweavers.com> It requires Vulkan transform feedback extension. --- tests/d3d12.c | 40 +++++++++++++++++++++++++++++++++------- 1 file changed, 33 insertions(+), 7 deletions(-) diff --git a/tests/d3d12.c b/tests/d3d12.c index fb0d4f7a5..5967523d6 100644 --- a/tests/d3d12.c +++ b/tests/d3d12.c @@ -30297,7 +30297,9 @@ static void test_domain_shader_inputs(void) struct test_context context; ID3D12Resource *so_buffer; ID3D12CommandQueue *queue; + const float *elems = NULL; unsigned int x, y; + bool fail = false; HRESULT hr; #if 0 @@ -30457,6 +30459,7 @@ static void test_domain_shader_inputs(void) pso_desc.StreamOutput.RasterizedStream = D3D12_SO_NO_RASTERIZED_STREAM; hr = ID3D12Device_CreateGraphicsPipelineState(context.device, &pso_desc, &IID_ID3D12PipelineState, (void **)&context.pipeline_state); + bug_if(is_mvk_device(context.device)) ok(hr == S_OK, "Got unexpected hr %#x.\n", hr); so_buffer = create_default_buffer(context.device, 1024, @@ -30478,13 +30481,21 @@ static void test_domain_shader_inputs(void) get_buffer_readback_with_command_list(so_buffer, DXGI_FORMAT_UNKNOWN, &rb, queue, command_list); for (y = 0; y < ARRAY_SIZE(reference); ++y) { - const float *elems = get_readback_data(&rb.rb, y, 0, 0, stride); + elems = get_readback_data(&rb.rb, y, 0, 0, stride); for (x = 0; x < ARRAY_SIZE(*reference); ++x) { - ok(compare_float(reference[y][x], elems[x], 0), - "Got unexpected value %.8e for [%u][%u], expected %.8e.\n", elems[x], y, x, reference[y][x]); + if (!compare_float(reference[y][x], elems[x], 0)) + { + fail = true; + break; + } } + + if (fail) + break; } + bug_if(is_mvk_device(context.device)) + ok(!fail, "Got unexpected value %.8e for [%u][%u], expected %.8e.\n", elems[x], y, x, reference[y][x]); release_resource_readback(&rb); ID3D12Resource_Release(so_buffer); @@ -30502,7 +30513,9 @@ static void test_domain_shader_one_patch_constant_input(void) struct test_context context; ID3D12Resource *so_buffer; ID3D12CommandQueue *queue; + const float *elems = NULL; unsigned int x, y; + bool fail = false; HRESULT hr; #if 0 @@ -30631,6 +30644,7 @@ static void test_domain_shader_one_patch_constant_input(void) pso_desc.StreamOutput.RasterizedStream = D3D12_SO_NO_RASTERIZED_STREAM; hr = ID3D12Device_CreateGraphicsPipelineState(context.device, &pso_desc, &IID_ID3D12PipelineState, (void **)&context.pipeline_state); + bug_if(is_mvk_device(context.device)) ok(hr == S_OK, "Got unexpected hr %#x.\n", hr); so_buffer = create_default_buffer(context.device, 1024, @@ -30652,13 +30666,21 @@ static void test_domain_shader_one_patch_constant_input(void) get_buffer_readback_with_command_list(so_buffer, DXGI_FORMAT_UNKNOWN, &rb, queue, command_list); for (y = 0; y < ARRAY_SIZE(reference); ++y) { - const float *elems = get_readback_data(&rb.rb, y, 0, 0, stride); + elems = get_readback_data(&rb.rb, y, 0, 0, stride); for (x = 0; x < ARRAY_SIZE(*reference); ++x) { - ok(compare_float(reference[y][x], elems[x], 0), - "Got unexpected value %.8e for [%u][%u], expected %.8e.\n", elems[x], y, x, reference[y][x]); + if (!compare_float(reference[y][x], elems[x], 0)) + { + fail = true; + break; + } } + + if (fail) + break; } + bug_if(is_mvk_device(context.device)) + ok(!fail, "Got unexpected value %.8e for [%u][%u], expected %.8e.\n", elems[x], y, x, reference[y][x]); release_resource_readback(&rb); ID3D12Resource_Release(so_buffer); @@ -36156,6 +36178,7 @@ static void test_hull_shader_relative_addressing(void) pso_desc.StreamOutput.RasterizedStream = D3D12_SO_NO_RASTERIZED_STREAM; hr = ID3D12Device_CreateGraphicsPipelineState(context.device, &pso_desc, &IID_ID3D12PipelineState, (void **)&context.pipeline_state); + bug_if(is_mvk_device(context.device)) ok(hr == S_OK, "Got unexpected hr %#x.\n", hr); vb = create_upload_buffer(context.device, sizeof(vertices), vertices); @@ -36182,7 +36205,8 @@ static void test_hull_shader_relative_addressing(void) transition_resource_state(command_list, so_buffer, D3D12_RESOURCE_STATE_STREAM_OUT, D3D12_RESOURCE_STATE_COPY_SOURCE); get_buffer_readback_with_command_list(so_buffer, DXGI_FORMAT_UNKNOWN, &rb, queue, command_list); - bug_if(is_radv_device(context.device)) check_triangles(&rb.rb, &expected_triangle, 1); + bug_if(is_radv_device(context.device) || is_mvk_device(context.device)) + check_triangles(&rb.rb, &expected_triangle, 1); release_resource_readback(&rb); ID3D12Resource_Release(so_buffer); @@ -36389,6 +36413,7 @@ static void test_hull_shader_patch_constant_inputs(void) pso_desc.StreamOutput.RasterizedStream = D3D12_SO_NO_RASTERIZED_STREAM; hr = ID3D12Device_CreateGraphicsPipelineState(context.device, &pso_desc, &IID_ID3D12PipelineState, (void **)&context.pipeline_state); + bug_if(is_mvk_device(context.device)) ok(hr == S_OK, "Got unexpected hr %#x.\n", hr); vb = create_upload_buffer(context.device, sizeof(vertices), vertices); @@ -36414,6 +36439,7 @@ static void test_hull_shader_patch_constant_inputs(void) transition_resource_state(command_list, so_buffer, D3D12_RESOURCE_STATE_STREAM_OUT, D3D12_RESOURCE_STATE_COPY_SOURCE); get_buffer_readback_with_command_list(so_buffer, DXGI_FORMAT_UNKNOWN, &rb, queue, command_list); + bug_if(is_mvk_device(context.device)) check_triangles(&rb.rb, &expected_triangle, 1); release_resource_readback(&rb); -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
From: Giovanni Mascellani <gmascellani(a)codeweavers.com> It requires Vulkan conditional rendering extension. --- tests/d3d12.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/tests/d3d12.c b/tests/d3d12.c index 5967523d6..42bf82086 100644 --- a/tests/d3d12.c +++ b/tests/d3d12.c @@ -34811,6 +34811,7 @@ static void test_conditional_rendering(void) transition_resource_state(command_list, context.render_target, D3D12_RESOURCE_STATE_RENDER_TARGET, D3D12_RESOURCE_STATE_COPY_SOURCE); + bug_if(is_mvk_device(context.device)) check_sub_resource_uint(context.render_target, 0, queue, command_list, 0xffffffff, 0); reset_command_list(command_list, context.allocator); @@ -34829,6 +34830,7 @@ static void test_conditional_rendering(void) ID3D12GraphicsCommandList_SetPredication(command_list, conditions, sizeof(uint64_t), D3D12_PREDICATION_OP_EQUAL_ZERO); + bug_if(is_mvk_device(context.device)) check_sub_resource_uint(context.render_target, 0, queue, command_list, 0xffffffff, 0); reset_command_list(command_list, context.allocator); @@ -34951,6 +34953,7 @@ static void test_conditional_rendering(void) 0, D3D12_PREDICATION_OP_NOT_EQUAL_ZERO); transition_resource_state(command_list, context.render_target, D3D12_RESOURCE_STATE_RENDER_TARGET, D3D12_RESOURCE_STATE_COPY_SOURCE); + bug_if(is_mvk_device(context.device)) check_sub_resource_uint(context.render_target, 0, queue, command_list, 0xffffffff, 0); reset_command_list(command_list, context.allocator); @@ -34972,6 +34975,7 @@ static void test_conditional_rendering(void) transition_resource_state(command_list, context.render_target, D3D12_RESOURCE_STATE_RENDER_TARGET, D3D12_RESOURCE_STATE_COPY_SOURCE); + bug_if(is_mvk_device(context.device)) check_sub_resource_uint(context.render_target, 0, queue, command_list, 0xffffffff, 0); reset_command_list(command_list, context.allocator); @@ -35146,7 +35150,8 @@ static void test_conditional_rendering(void) get_buffer_readback_with_command_list(buffer, DXGI_FORMAT_R32_UINT, &rb, queue, command_list); value = get_readback_uint(&rb.rb, 0, 0, 0); - ok(value == (!i ? init_value : input.value), "Got %#x, expected %#x.\n", value, input.value); + bug_if(is_mvk_device(context.device)) + ok(value == (!i ? init_value : input.value), "Got %#x, expected %#x.\n", value, !i ? init_value : input.value); release_resource_readback(&rb); reset_command_list(command_list, context.allocator); -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
From: Giovanni Mascellani <gmascellani(a)codeweavers.com> They require Vulkan descriptor indexing extension, which is exposed by MoltenVK but known to be buggy, so we assume it is disabled. --- tests/d3d12.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/tests/d3d12.c b/tests/d3d12.c index 42bf82086..acc2db2ca 100644 --- a/tests/d3d12.c +++ b/tests/d3d12.c @@ -2899,6 +2899,7 @@ static void test_create_root_signature(void) descriptor_ranges[1].RegisterSpace = 0; descriptor_ranges[1].OffsetInDescriptorsFromTableStart = 16; hr = create_root_signature(device, &root_signature_desc, &root_signature); + bug_if(is_mvk_device(device)) ok(hr == E_INVALIDARG, "Got unexpected hr %#x.\n", hr); /* A bounded range overlapping an unbounded one, mapped to the same @@ -2918,6 +2919,7 @@ static void test_create_root_signature(void) descriptor_ranges[1].RangeType = D3D12_DESCRIPTOR_RANGE_TYPE_SRV; descriptor_ranges[1].NumDescriptors = UINT_MAX; hr = create_root_signature(device, &root_signature_desc, &root_signature); + bug_if(is_mvk_device(device)) ok(hr == E_INVALIDARG, "Got unexpected hr %#x.\n", hr); /* And unbounded range overlapping a bounded one, mapped to the same @@ -2926,6 +2928,7 @@ static void test_create_root_signature(void) descriptor_ranges[1].BaseShaderRegister = 0; descriptor_ranges[1].OffsetInDescriptorsFromTableStart = 15; hr = create_root_signature(device, &root_signature_desc, &root_signature); + bug_if(is_mvk_device(device)) ok(hr == E_INVALIDARG, "Got unexpected hr %#x.\n", hr); refcount = ID3D12Device_Release(device); -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
From: Giovanni Mascellani <gmascellani(a)codeweavers.com> They require Vulkan descriptor indexing extension, which is exposed by MoltenVK but known to be buggy, so we assume it is disabled. --- tests/d3d12.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/d3d12.c b/tests/d3d12.c index acc2db2ca..1d4da810d 100644 --- a/tests/d3d12.c +++ b/tests/d3d12.c @@ -36615,6 +36615,7 @@ static void test_resource_arrays(void) get_cpu_descriptor_handle(&context, heap, ARRAY_SIZE(input_buffers) + i)); } + bug_if(is_mvk_device(device)) context.pipeline_state = create_compute_pipeline_state(device, context.root_signature, shader_bytecode(cs_code, sizeof(cs_code))); if (!context.pipeline_state) -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
From: Giovanni Mascellani <gmascellani(a)codeweavers.com> Specifically, MoltenVK seems to be able to load from stencil, but the specific replicating swizzle (repeating the stencil value on all the channels) is not honored. The stencil value is read only on the red channel. --- tests/d3d12.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tests/d3d12.c b/tests/d3d12.c index 1d4da810d..202d29599 100644 --- a/tests/d3d12.c +++ b/tests/d3d12.c @@ -21327,11 +21327,13 @@ static void test_stencil_load(void) transition_sub_resource_state(command_list, context.render_target, 0, D3D12_RESOURCE_STATE_RENDER_TARGET, D3D12_RESOURCE_STATE_COPY_SOURCE); + bug_if(is_mvk_device(context.device)) check_sub_resource_uvec4(context.render_target, 0, queue, command_list, &uvec4); reset_command_list(command_list, context.allocator); transition_sub_resource_state(command_list, texture, 0, D3D12_RESOURCE_STATE_UNORDERED_ACCESS, D3D12_RESOURCE_STATE_COPY_SOURCE); + bug_if(is_mvk_device(context.device)) check_sub_resource_uvec4(texture, 0, queue, command_list, &uvec4); reset_command_list(command_list, context.allocator); -- GitLab https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/531
participants (2)
-
Giovanni Mascellani -
Giovanni Mascellani (＠giomasce)