When checking query wait completion, all queries were previously
checked for Available status, and blocked until all became Available.
However, only queries that were encoded to be written should be checked.
It is okay for queries that are not encoded to be written to remain in
Initial state, and not block the wait.
Support by the MTLDevice for Metal Tier 2 argument buffers has an
impact on both descriptor indexing and buffer device address support.
Add MVKPhysicalDeviceMetalFeatures::argumentBuffersTier, to track
the MTLDevice argument buffers support tier, set it from
[MTLDevice argumentBuffersSupport], and subsequently enable support for
VK_KHR_buffer_device_address and VK_EXT_buffer_device_address extensions,
and set descriptor indexing resource counts, based on it.
Update documentation requirements.
Add MVKPhysicalDeviceMetalFeatures::shaderBarycentricCoordinates,
enable it based on [MTLDevice supportsShaderBarycentricCoordinates],
and subsequently enable support for VK_KHR_fragment_shader_barycentric
and VK_NV_fragment_shader_barycentric extensions, based on it.
- vkCreateSemaphore() create binary semaphore, unless timeline explicitly
requested. Previously, timeline semaphore was always created when
exportObjectType == VK_EXPORT_METAL_OBJECT_TYPE_METAL_SHARED_EVENT_BIT_EXT.
- vkCreateSemaphore() record error when MTLSharedEvent imported or marked for
export, and MoltenVK configured for VkSemaphore using MTLFence or CPU emulation.
- MVKTimelineSemaphore elide unused constructor arguments.
Creating an empty MVKDescriptorSet will read this member and tip off undefined
behavior sanitizer. It shouldn't actually cause any bugs, but with UB you never
know.
- Update to latest SPIRV-Cross to support `SPV_KHR_physical_storage_buffer`
for `VK_KHR_buffer_device_address` and `VK_EXT_buffer_device_address`
- Add support for `VK_EXT_buffer_device_address` extension.
- Advertise support for `VK_KHR_buffer_device_address`
and `VK_EXT_buffer_device_address` on macOS 12.5.
- Add appropriate extension reporting and enablement for
`VkPhysicalDeviceBufferDeviceAddressFeatures`,
`VkPhysicalDeviceBufferDeviceAddressFeaturesEXT`, and
`VkPhysicalDeviceFragmentShaderBarycentricFeaturesKHR`.
- Support reading `VkMemoryAllocateFlagsInfo` to identify memory allocations that
need to support buffer pointer access (in case needed in future on non-shared memory).
- Update `Whats_New.md` and `MoltenVK_Runtime_UserGuide` documents.
- Replace use of deprecated kIOMasterPortDefault with MACH_PORT_NULL.
- Convert an inline VLA with constant length array.
- Add input files to all build phases to trigger dependencies when inputs change.
- Add packaging scripts to per-platform ExternalDependencies targets.
Add dummy output files to each Xcode Run Script build phase that doesn't
already have dependencies set, to avoid setting alwaysOutOfDate flag
(by disabling "Based on dependency analysis" UI flag), which forces
the Xcode project to a version that can't be read by Xcode 11.7.
- Update minimum Xcode deployment targets to macOS 10.13, iOS 11, and tvOS 11,
to avoid Xcode build warnings.
- Add support for MTLLanguageVersion3_0 enumeration.
- Build efficiencies:
- Build scripts create_dylib.sh and gen_moltenvk_rev_hdr.sh
only run if build dependencies require it.
- Packaging and copy_to_staging.sh scripts are too complex to define dependencies,
and are fast, so configured to run every time, to avoid build warning.
- Replace use of deprecated sprintf() with snprintf().
- Replace use of deprecated kIOMasterPortDefault with kIOMainPortDefault.
- Support old-style GPU debug capture only if building for earlier minimum
deployment targets, to avoid deprecation warning.
- Update minimum Xcode deployment targets of Cube demo to macOS 10.14, iOS 12,
and tvOS 12, to avoid Xcode build warning regarding MTLSharedEvent in .
- Update README.md document regarding minimum Xcode deployment targets.
When new descriptor sets are bound, it may be the case that many of the
bindings are not changed from their previous state. Detect this case and avoid
binding the same resources in the Metal comand buffer repeatedly. If possible,
change only the offset. This saves some encoding time.
This patch fixes a regression caused by 1b6b8bc9 when the implicit buffers
were moved to top of Metal buffer index range.
Fix vertex buffer binding counts when binding indexes are not consecutive,
or when additional synthetic buffer bindings are required to accommodate
vertex attributes that are outside the vertex buffer stride values.
Take into consideration that the app may bind more vertex attribute buffers
than a pipeline may consume, and don't allow these to overwrite any implicit
buffers sent to the shader.
MVKResourcesCommandEncoderState::encodeBinding() clear binding dirty flag
before calling lambda function to allow function operation to possibly
override and leave binding marked dirty.
Set tessellation vertex buffer indexes based on platform maximum vertex stage
buffer count, instead of hardcoding them, to preemptively avoid conflict with
implicit buffers should the platform counts ever change.
So far it always returned a 60 Hz refresh on macOS. Now we query the
actual refresh interval for the NSScreen on which our swapChain is likely
displaying.
Tested on a MacBookPro dual-display setup with builtin 60 Hz Retina panel
and external 100 Hz monitor, showing proper reporting depending on which
monitor the window was displaying on. Tested on macOS 10.15.7 only.
renderpass was writing timestamp before renderpass activity was complete.
MVKCommandBuffer tracks whether it contains a stage-based timestamp command,
and MVKCommandEncoder updates the timestamp command fence when ending any
Metal command encoder on such a MVKCommandBuffer.
MVKCommandEncoder reorder member variables to avoid layout gaps (unrelated).
MVKCommandBuffer update _commandCount even for single-use immediate command
encoding (unrelated).
Add VK_KHR_separate_depth_stencil_layouts and enable its features.
Metal generally ignores image layouts, so nothing further needed.
Update Whats_New.md and MoltenVK_Runtime_UserGuide.md documents
with recently-added extensions.
Passes almost all supported separate_layouts CTS tests, with the
few remaining tests failing due to issues with unrelated capabilities.
Unrelated cleanup:
- MVKDevice enable extensions before features.
- MVKDevice reorder ivar declarations to reduce
memory layout gaps, and define default values.
- Rename VkSemaphoreStyle to MVKSemaphoreStyle to
remove potential conflicts with Vulkan Vk name space.
- MVKRenderPass set rendering flags from outside instead of via constructor.
- MVKCommandEncodingContext tracks and manages lifespan of renderpass and
framebuffer, potentially across multiple MVKCommandEncoders.
- When resuming a suspended dynamic renderpass, reuse the renderpass and
framebuffer cached in MVKCommandEncodingContext, otherwise create them.
- MVKCommandBuffer::begin() creates local copy of dynamic rendering color formats.
- MVKDevice track enabling VK_KHR_dynamic_rendering extension features.
- Add MVKCmdBeginRendering and MVKCmdEndRendering command objects
and associated command pools.
- Add MVKCommandEncoder::beginRendering() which dynamically instantiates
temporary MVKRenderpass and (imageless) MVKFramebuffer objects to handle
equivalent renderpass operations.
- MVKCommandEncoder retain and release subpass and framebuffer objects
to allow transient dynamic instances to be created and destroyed.
- Add support functions to create MVKRenderpass and (imageless)
MVKFramebuffer objects from VkRenderingInfo.
- MVKRenderpass track VkRenderingFlags.
- MVKRenderSubpass create and track an internal VkPipelineRenderingCreateInfo.
- MVKRenderPassAttachment support LoadOp and StoreOp behavior for
dynamic rendering suspend and resume.
- Refactor viewMask processing to permit operations on viewMask outside of
MVKGraphicsPipeline without requiring the presence of an MVKSubpass.
- MVKGraphicsPipeline extract VkPipelineRenderingCreateInfo from MVKRenderSubpass
if available, or VkGraphicsPipelineCreateInfo::pNext if not.
- MVKCommandBuffer track VkCommandBufferInheritanceRenderingInfo (currently unused).
- (unrelated) vulkan.mm move location of VK_EXT_sample_locations functions
for layout readability consistency.
buffer binding used by a subsequent pipeline that does not use push constants.
This error was previously introduced in 2a17f75, where a push constants
binding could override the Metal buffer binding 0 of a subsequent pipeline
that does not use push constants.
When pipeline binding is encoded, track which stages use push constants
and only encode push constants if the pipeline and stage uses them.
(unrelated) Make use of MVKResourcesCommandEncoderState::getPipeline() consistent.
