// Copyright (c) 2015-2018 Khronos Group. This work is licensed under a // Creative Commons Attribution 4.0 International License; see // http://creativecommons.org/licenses/by/4.0/ [[synchronization]] = Synchronization and Cache Control Synchronization of access to resources is primarily the responsibility of the application in Vulkan. The order of execution of commands with respect to the host and other commands on the device has few implicit guarantees, and needs to be explicitly specified. Memory caches and other optimizations are also explicitly managed, requiring that the flow of data through the system is largely under application control. Whilst some implicit guarantees exist between commands, five explicit synchronization mechanisms are exposed by Vulkan: <>:: Fences can: be used to communicate to the host that execution of some task on the device has completed. <>:: Semaphores can: be used to control resource access across multiple queues. <>:: Events provide a fine-grained synchronization primitive which can: be signaled either within a command buffer or by the host, and can: be waited upon within a command buffer or queried on the host. <>:: Pipeline barriers also provide synchronization control within a command buffer, but at a single point, rather than with separate signal and wait operations. <>:: Render passes provide a useful synchronization framework for most rendering tasks, built upon the concepts in this chapter. Many cases that would otherwise need an application to use other synchronization primitives can: be expressed more efficiently as part of a render pass. [[synchronization-dependencies]] == Execution and Memory Dependencies An _operation_ is an arbitrary amount of work to be executed on the host, a device, or an external entity such as a presentation engine. Synchronization commands introduce explicit _execution dependencies_, and _memory dependencies_ between two sets of operations defined by the command's two _synchronization scopes_. [[synchronization-dependencies-scopes]] The synchronization scopes define which other operations a synchronization command is able to create execution dependencies with. Any type of operation that is not in a synchronization command's synchronization scopes will not be included in the resulting dependency. For example, for many synchronization commands, the synchronization scopes can: be limited to just operations executing in specific <>, which allows other pipeline stages to be excluded from a dependency. Other scoping options are possible, depending on the particular command. [[synchronization-dependencies-execution]] An _execution dependency_ is a guarantee that for two sets of operations, the first set must: _happen-before_ the second set. If an operation happens-before another operation, then the first operation must: complete before the second operation is initiated. More precisely: * Let *A* and *B* be separate sets of operations. * Let *S* be a synchronization command. * Let *A~S~* and *B~S~* be the synchronization scopes of *S*. * Let *A'* be the intersection of sets *A* and *A~S~*. * Let *B'* be the intersection of sets *B* and *B~S~*. * Submitting *A*, *S* and *B* for execution, in that order, will result in execution dependency *E* between *A'* and *B'*. * Execution dependency *E* guarantees that *A'* happens-before *B'*. [[synchronization-dependencies-chains]] An _execution dependency chain_ is a sequence of execution dependencies that form a happens-before relation between the first dependency's *A'* and the final dependency's *B'*. For each consecutive pair of execution dependencies, a chain exists if the intersection of *B~S~* in the first dependency and *A~S~* in the second dependency is not an empty set. The formation of a single execution dependency from an execution dependency chain can be described by substituting the following in the description of execution dependencies: * Let *S* be a set of synchronization commands that generate an execution dependency chain. * Let *A~S~* be the first synchronization scope of the first command in *S*. * Let *B~S~* be the second synchronization scope of the last command in *S*. [NOTE] .Note ==== An execution dependency is inherently also multiple execution dependencies - a dependency exists between each subset of *A'* and each subset of *B'*, and the same is true for execution dependency chains. For example, a synchronization command with multiple <> in its stage masks effectively generates one dependency between each source stage and each destination stage. This can be useful to think about when considering how execution chains are formed if they do not involve all parts of a synchronization command's dependency. Similarly, any set of adjacent dependencies in an execution dependency chain can: be considered an execution dependency chain in its own right. ==== Execution dependencies alone are not sufficient to guarantee that values resulting from writes in one set of operations can: be read from another set of operations. [[synchronization-dependencies-available-and-visible]] Three additional types of operation are used to control memory access. _Availability operations_ cause the values generated by specified memory write accesses to become _available_ to a memory domain for future access. Any available value remains available until a subsequent write to the same memory location occurs (whether it is made available or not) or the memory is freed. _Memory domain operations_ cause writes that are available to a source memory domain to become available to a destination memory domain (an example of this is making writes available to the host domain available to the device domain). _Visibility operations_ cause values available to a memory domain to become _visible_ to specified memory accesses. ifdef::VK_KHR_vulkan_memory_model[] Availability, visibility, memory domains, and memory domain operations are formally defined in the <> section of the <> chapter. Which API operations perform each of these operations is defined in <>. endif::VK_KHR_vulkan_memory_model[] [[synchronization-dependencies-memory]] A _memory dependency_ is an execution dependency which includes availability and visibility operations such that: * The first set of operations happens-before the availability operation. * The availability operation happens-before the visibility operation. * The visibility operation happens-before the second set of operations. Once written values are made visible to a particular type of memory access, they can: be read or written by that type of memory access. Most synchronization commands in Vulkan define a memory dependency. [[synchronization-dependencies-access-scopes]] The specific memory accesses that are made available and visible are defined by the _access scopes_ of a memory dependency. Any type of access that is in a memory dependency's first access scope and occurs in *A'* is made available. Any type of access that is in a memory dependency's second access scope and occurs in *B'* has any available writes made visible to it. Any type of operation that is not in a synchronization command's access scopes will not be included in the resulting dependency. A memory dependency enforces availability and visibility of memory accesses and execution order between two sets of operations. Adding to the description of <>: * Let *a* be the set of memory accesses performed by *A'*. * Let *b* be the set of memory accesses performed by *B'*. * Let *a~S~* be the first access scope of the first command in *S*. * Let *b~S~* be the second access scope of the last command in *S*. * Let *a'* be the intersection of sets *a* and *a~S~*. * Let *b'* be the intersection of sets *b* and *b~S~*. * Submitting *A*, *S* and *B* for execution, in that order, will result in a memory dependency *m* between *A'* and *B'*. * Memory dependency *m* guarantees that: ** Memory writes in *a'* are made available. ** Available memory writes, including those from *a'*, are made visible to *b'*. [NOTE] .Note ==== Execution and memory dependencies are used to solve data hazards, i.e. to ensure that read and write operations occur in a well-defined order. Write-after-read hazards can be solved with just an execution dependency, but read-after-write and write-after-write hazards need appropriate memory dependencies to be included between them. If an application does not include dependencies to solve these hazards, the results and execution orders of memory accesses are undefined:. ==== [[synchronization-image-layout-transitions]] === Image Layout Transitions Image subresources can: be transitioned from one <> to another as part of a <> (e.g. by using an <>). When a layout transition is specified in a memory dependency, it happens-after the availability operations in the memory dependency, and happens-before the visibility operations. Image layout transitions may: perform read and write accesses on all memory bound to the image subresource range, so applications must: ensure that all memory writes have been made <> before a layout transition is executed. Available memory is automatically made visible to a layout transition, and writes performed by a layout transition are automatically made available. Layout transitions always apply to a particular image subresource range, and specify both an old layout and new layout. If the old layout does not match the new layout, a transition occurs. The old layout must: match the current layout of the image subresource range, with one exception. The old layout can: always be specified as ename:VK_IMAGE_LAYOUT_UNDEFINED, though doing so invalidates the contents of the image subresource range. ifdef::VK_VERSION_1_1,VK_KHR_device_group[] As image layout transitions may: perform read and write accesses on the memory bound to the image, if the image subresource affected by the layout transition is bound to peer memory for any device in the current device mask then the memory heap the bound memory comes from must: support the ename:VK_PEER_MEMORY_FEATURE_GENERIC_SRC_BIT and ename:VK_PEER_MEMORY_FEATURE_GENERIC_DST_BIT capabilities as returned by flink:vkGetDeviceGroupPeerMemoryFeatures. endif::VK_VERSION_1_1,VK_KHR_device_group[] [NOTE] .Note ==== Setting the old layout to ename:VK_IMAGE_LAYOUT_UNDEFINED implies that the contents of the image subresource need not be preserved. Implementations may: use this information to avoid performing expensive data transition operations. ==== [NOTE] .Note ==== Applications must: ensure that layout transitions happen-after all operations accessing the image with the old layout, and happen-before any operations that will access the image with the new layout. Layout transitions are potentially read/write operations, so not defining appropriate memory dependencies to guarantee this will result in a data race. ==== Image layout transitions interact with <>. [[synchronization-pipeline-stages]] === Pipeline Stages The work performed by an <> consists of multiple operations, which are performed as a sequence of logically independent steps known as _pipeline stages_. The exact pipeline stages executed depend on the particular command that is used, and current command buffer state when the command was recorded. <>, <>, <>, <>, and <> all execute in different sets of <>. <> do not execute in a defined pipeline, but do execute ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT and ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT. [NOTE] .Note ==== Operations performed by synchronization commands (e.g. <>) are not executed by a defined pipeline stage. However other commands can still synchronize with them via the ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT and ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT pipeline stages. ==== Execution of operations across pipeline stages must: adhere to <>, particularly including <>. Otherwise, execution across pipeline stages may: overlap or execute out of order with regards to other stages, unless otherwise enforced by an execution dependency. [open,refpage='VkPipelineStageFlagBits',desc='Bitmask specifying pipeline stages',type='enums'] -- Several of the synchronization commands include pipeline stage parameters, restricting the <> for that command to just those stages. This allows fine grained control over the exact execution dependencies and accesses performed by action commands. Implementations should: use these pipeline stages to avoid unnecessary stalls or cache flushing. Bits which can be set, specifying pipeline stages, are: include::../api/enums/VkPipelineStageFlagBits.txt[] * ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT specifies the stage of the pipeline where any commands are initially received by the queue. * ename:VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT specifies the stage of the pipeline where Draw/DispatchIndirect data structures are consumed. ifdef::VK_NVX_device_generated_commands[] This stage also includes reading commands written by flink:vkCmdProcessCommandsNVX. endif::VK_NVX_device_generated_commands[] ifdef::VK_NV_mesh_shader[] * ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT_NV specifies the task shader stage. * ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT_NV specifies the mesh shader stage. endif::VK_NV_mesh_shader[] * ename:VK_PIPELINE_STAGE_VERTEX_INPUT_BIT specifies the stage of the pipeline where vertex and index buffers are consumed. * ename:VK_PIPELINE_STAGE_VERTEX_SHADER_BIT specifies the vertex shader stage. * ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT specifies the tessellation control shader stage. * ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT specifies the tessellation evaluation shader stage. * ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT specifies the geometry shader stage. * ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT specifies the fragment shader stage. * ename:VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT specifies the stage of the pipeline where early fragment tests (depth and stencil tests before fragment shading) are performed. This stage also includes <> for framebuffer attachments with a depth/stencil format. * ename:VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT specifies the stage of the pipeline where late fragment tests (depth and stencil tests after fragment shading) are performed. This stage also includes <> for framebuffer attachments with a depth/stencil format. * ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT specifies the stage of the pipeline after blending where the final color values are output from the pipeline. This stage also includes <> and multisample resolve operations for framebuffer attachments with a color format. * ename:VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT specifies the execution of a compute shader. * [[synchronization-pipeline-stages-transfer]] ename:VK_PIPELINE_STAGE_TRANSFER_BIT specifies the execution of copy commands. This includes the operations resulting from all <>, <> (with the exception of flink:vkCmdClearAttachments), and flink:vkCmdCopyQueryPoolResults. * ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT specifies the final stage in the pipeline where operations generated by all commands complete execution. * ename:VK_PIPELINE_STAGE_HOST_BIT specifies a pseudo-stage indicating execution on the host of reads/writes of device memory. This stage is not invoked by any commands recorded in a command buffer. ifdef::VK_NVX_raytracing[] * ename:VK_PIPELINE_STAGE_RAYTRACING_BIT_NVX specifies the execution of the ray tracing stages. endif::VK_NVX_raytracing[] * ename:VK_PIPELINE_STAGE_ALL_GRAPHICS_BIT specifies the execution of all graphics pipeline stages, and is equivalent to the logical OR of: ** ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT ** ename:VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT ifdef::VK_NV_mesh_shader[] ** ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT_NV ** ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT_NV endif::VK_NV_mesh_shader[] ** ename:VK_PIPELINE_STAGE_VERTEX_INPUT_BIT ** ename:VK_PIPELINE_STAGE_VERTEX_SHADER_BIT ** ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT ** ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT ** ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT ** ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT ** ename:VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT ** ename:VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT ** ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT ** ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT ifdef::VK_EXT_conditional_rendering[] ** ename:VK_PIPELINE_STAGE_CONDITIONAL_RENDERING_BIT_EXT endif::VK_EXT_conditional_rendering[] ifdef::VK_EXT_transform_feedback[] ** ename:VK_PIPELINE_STAGE_TRANSFORM_FEEDBACK_BIT_EXT endif::VK_EXT_transform_feedback[] ifdef::VK_NV_shading_rate_image[] ** ename:VK_PIPELINE_STAGE_SHADING_RATE_IMAGE_BIT_NV endif::VK_NV_shading_rate_image[] * ename:VK_PIPELINE_STAGE_ALL_COMMANDS_BIT is equivalent to the logical OR of every other pipeline stage flag that is supported on the queue it is used with. ifdef::VK_EXT_conditional_rendering[] * ename:VK_PIPELINE_STAGE_CONDITIONAL_RENDERING_BIT_EXT specifies the stage of the pipeline where the predicate of conditional rendering is consumed. endif::VK_EXT_conditional_rendering[] ifdef::VK_EXT_transform_feedback[] * ename:VK_PIPELINE_STAGE_TRANSFORM_FEEDBACK_BIT_EXT specifies the stage of the pipeline where vertex attribute output values are written to the transform feedback buffers. endif::VK_EXT_transform_feedback[] ifdef::VK_NVX_device_generated_commands[] * ename:VK_PIPELINE_STAGE_COMMAND_PROCESS_BIT_NVX specifies the stage of the pipeline where device-side generation of commands via flink:vkCmdProcessCommandsNVX is handled. endif::VK_NVX_device_generated_commands[] ifdef::VK_NV_shading_rate_image[] * ename:VK_PIPELINE_STAGE_SHADING_RATE_IMAGE_BIT_NV specifies the stage of the pipeline where the <> is read to determine the shading rate for portions of a rasterized primitive. endif::VK_NV_shading_rate_image[] [NOTE] .Note ==== An execution dependency with only ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT in the destination stage mask will only prevent that stage from executing in subsequently submitted commands. As this stage does not perform any actual execution, this is not observable - in effect, it does not delay processing of subsequent commands. Similarly an execution dependency with only ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT in the source stage mask will effectively not wait for any prior commands to complete. When defining a memory dependency, using only ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT or ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT would never make any accesses available and/or visible because these stages do not access memory. ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT and ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT are useful for accomplishing layout transitions and queue ownership operations when the required execution dependency is satisfied by other means - for example, semaphore operations between queues. ==== -- [open,refpage='VkPipelineStageFlags',desc='Bitmask of VkPipelineStageFlagBits',type='enums'] -- include::../api/flags/VkPipelineStageFlags.txt[] sname:VkPipelineStageFlags is a bitmask type for setting a mask of zero or more slink:VkPipelineStageFlagBits. -- [[synchronization-pipeline-stages-masks]] If a synchronization command includes a source stage mask, its first <> only includes execution of the pipeline stages specified in that mask, and its first <> only includes memory access performed by pipeline stages specified in that mask. If a synchronization command includes a destination stage mask, its second <> only includes execution of the pipeline stages specified in that mask, and its second <> only includes memory access performed by pipeline stages specified in that mask. [NOTE] .Note ==== Including a particular pipeline stage in the first <> of a command implicitly includes <> pipeline stages in the synchronization scope. Similarly, the second <> includes <> pipeline stages. However, note that <> are not affected in this way - only the precise stages specified are considered part of each access scope. ==== Certain pipeline stages are only available on queues that support a particular set of operations. The following table lists, for each pipeline stage flag, which queue capability flag must: be supported by the queue. When multiple flags are enumerated in the second column of the table, it means that the pipeline stage is supported on the queue if it supports any of the listed capability flags. For further details on queue capabilities see <> and <>. [[synchronization-pipeline-stages-supported]] .Supported pipeline stage flags [cols="60%,40%",options="header"] |==== |Pipeline stage flag | Required queue capability flag |ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT | None required |ename:VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT | ename:VK_QUEUE_GRAPHICS_BIT or ename:VK_QUEUE_COMPUTE_BIT |ename:VK_PIPELINE_STAGE_VERTEX_INPUT_BIT | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_VERTEX_SHADER_BIT | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT | ename:VK_QUEUE_COMPUTE_BIT |ename:VK_PIPELINE_STAGE_TRANSFER_BIT | ename:VK_QUEUE_GRAPHICS_BIT, ename:VK_QUEUE_COMPUTE_BIT or ename:VK_QUEUE_TRANSFER_BIT |ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT | None required |ename:VK_PIPELINE_STAGE_HOST_BIT | None required |ename:VK_PIPELINE_STAGE_ALL_GRAPHICS_BIT | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_ALL_COMMANDS_BIT | None required ifdef::VK_EXT_conditional_rendering[] |ename:VK_PIPELINE_STAGE_CONDITIONAL_RENDERING_BIT_EXT | ename:VK_QUEUE_GRAPHICS_BIT or ename:VK_QUEUE_COMPUTE_BIT endif::VK_EXT_conditional_rendering[] ifdef::VK_EXT_transform_feedback[] |ename:VK_PIPELINE_STAGE_TRANSFORM_FEEDBACK_BIT_EXT | ename:VK_QUEUE_GRAPHICS_BIT endif::VK_EXT_transform_feedback[] ifdef::VK_NVX_device_generated_commands[] |ename:VK_PIPELINE_STAGE_COMMAND_PROCESS_BIT_NVX | ename:VK_QUEUE_GRAPHICS_BIT or ename:VK_QUEUE_COMPUTE_BIT endif::VK_NVX_device_generated_commands[] ifdef::VK_NV_shading_rate_image[] |ename:VK_PIPELINE_STAGE_SHADING_RATE_IMAGE_BIT_NV | ename:VK_QUEUE_GRAPHICS_BIT endif::VK_NV_shading_rate_image[] ifdef::VK_NV_mesh_shader[] |ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT_NV | ename:VK_QUEUE_GRAPHICS_BIT |ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT_NV | ename:VK_QUEUE_GRAPHICS_BIT endif::VK_NV_mesh_shader[] ifdef::VK_NVX_raytracing[] |ename:VK_PIPELINE_STAGE_RAYTRACING_BIT_NVX | ename:VK_QUEUE_COMPUTE_BIT endif::VK_NVX_raytracing[] |==== [[synchronization-pipeline-stages-order]] Pipeline stages that execute as a result of a command logically complete execution in a specific order, such that completion of a logically later pipeline stage must: not happen-before completion of a logically earlier stage. This means that including any stage in the source stage mask for a particular synchronization command also implies that any logically earlier stages are included in *A~S~* for that command. Similarly, initiation of a logically earlier pipeline stage must: not happen-after initiation of a logically later pipeline stage. Including any given stage in the destination stage mask for a particular synchronization command also implies that any logically later stages are included in *B~S~* for that command. [NOTE] .Note ==== Implementations may: not support synchronization at every pipeline stage for every synchronization operation. If a pipeline stage that an implementation does not support synchronization for appears in a source stage mask, it may: substitute any logically later stage in its place for the first synchronization scope. If a pipeline stage that an implementation does not support synchronization for appears in a destination stage mask, it may: substitute any logically earlier stage in its place for the second synchronization scope. For example, if an implementation is unable to signal an event immediately after vertex shader execution is complete, it may: instead signal the event after color attachment output has completed. If an implementation makes such a substitution, it must: not affect the semantics of execution or memory dependencies or image and buffer memory barriers. ==== [[synchronization-pipeline-stages-types]] The order and set of pipeline stages executed by a given command is determined by the command's pipeline type, as described below: ifndef::VK_NV_mesh_shader[] For the graphics pipeline, the following stages occur in this order: endif::VK_NV_mesh_shader[] ifdef::VK_NV_mesh_shader[] For the graphics primitive shading pipeline, the following stages occur in this order: endif::VK_NV_mesh_shader[] * ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT * ename:VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT * ename:VK_PIPELINE_STAGE_VERTEX_INPUT_BIT * ename:VK_PIPELINE_STAGE_VERTEX_SHADER_BIT * ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT * ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT * ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT ifdef::VK_EXT_transform_feedback[] * ename:VK_PIPELINE_STAGE_TRANSFORM_FEEDBACK_BIT_EXT endif::VK_EXT_transform_feedback[] ifdef::VK_NV_shading_rate_image[] * ename:VK_PIPELINE_STAGE_SHADING_RATE_IMAGE_BIT_NV endif::VK_NV_shading_rate_image[] * ename:VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT * ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT * ename:VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT * ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT * ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT ifdef::VK_NV_mesh_shader[] For the graphics mesh shading pipeline, the following stages occur in this order: * ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT * ename:VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT * ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT_NV * ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT_NV ifdef::VK_NV_shading_rate_image[] * ename:VK_PIPELINE_STAGE_SHADING_RATE_IMAGE_BIT_NV endif::VK_NV_shading_rate_image[] * ename:VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT * ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT * ename:VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT * ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT * ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT endif::VK_NV_mesh_shader[] For the compute pipeline, the following stages occur in this order: * ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT * ename:VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT * ename:VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT * ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT ifdef::VK_EXT_conditional_rendering[] The conditional rendering stage is formally part of both the graphics, and the compute pipeline. The pipeline stage where the predicate read happens has unspecified order relative to other stages of these pipelines: * ename:VK_PIPELINE_STAGE_CONDITIONAL_RENDERING_BIT_EXT endif::VK_EXT_conditional_rendering[] For the transfer pipeline, the following stages occur in this order: * ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT * ename:VK_PIPELINE_STAGE_TRANSFER_BIT * ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT For host operations, only one pipeline stage occurs, so no order is guaranteed: * ename:VK_PIPELINE_STAGE_HOST_BIT ifdef::VK_NVX_device_generated_commands[] For the command processing pipeline, the following stages occur in this order: * ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT * ename:VK_PIPELINE_STAGE_COMMAND_PROCESS_BIT_NVX * ename:VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT endif::VK_NVX_device_generated_commands[] ifdef::VK_NVX_raytracing[] For the ray tracing pipeline, only one pipeline stage occurs, so no order is guaranteed: * ename:VK_PIPELINE_STAGE_RAYTRACING_BIT_NVX endif::VK_NVX_raytracing[] [[synchronization-access-types]] === Access Types [open,refpage='VkAccessFlagBits',desc='Bitmask specifying memory access types that will participate in a memory dependency',type='enums'] -- Memory in Vulkan can: be accessed from within shader invocations and via some fixed-function stages of the pipeline. The _access type_ is a function of the <> used, or how a fixed-function stage accesses memory. Each access type corresponds to a bit flag in slink:VkAccessFlagBits. [[synchronization-access-masks]] Some synchronization commands take sets of access types as parameters to define the <> of a memory dependency. If a synchronization command includes a source access mask, its first <> only includes accesses via the access types specified in that mask. Similarly, if a synchronization command includes a destination access mask, its second <> only includes accesses via the access types specified in that mask. Access types that can: be set in an access mask include: include::../api/enums/VkAccessFlagBits.txt[] * ename:VK_ACCESS_INDIRECT_COMMAND_READ_BIT specifies read access to indirect command data read as part of an indirect drawing or dispatch command. * ename:VK_ACCESS_INDEX_READ_BIT specifies read access to an index buffer as part of an indexed drawing command, bound by flink:vkCmdBindIndexBuffer. * ename:VK_ACCESS_VERTEX_ATTRIBUTE_READ_BIT specifies read access to a vertex buffer as part of a drawing command, bound by flink:vkCmdBindVertexBuffers. * ename:VK_ACCESS_UNIFORM_READ_BIT specifies read access to a <>. * ename:VK_ACCESS_INPUT_ATTACHMENT_READ_BIT specifies read access to an <> within a render pass during fragment shading. * ename:VK_ACCESS_SHADER_READ_BIT specifies read access to a <>, <>, <>, <>, or <>. * ename:VK_ACCESS_SHADER_WRITE_BIT specifies write access to a <>, <>, or <>. * ename:VK_ACCESS_COLOR_ATTACHMENT_READ_BIT specifies read access to a <>, such as via <>, <>, or via certain <>. ifdef::VK_EXT_blend_operation_advanced[] It does not include <>. endif::VK_EXT_blend_operation_advanced[] * ename:VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT specifies write access to a <> during a <> or via certain <>. * ename:VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT specifies read access to a <>, via <> or via certain <>. * ename:VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT specifies write access to a <>, via <> or via certain <>. * ename:VK_ACCESS_TRANSFER_READ_BIT specifies read access to an image or buffer in a <> operation. * ename:VK_ACCESS_TRANSFER_WRITE_BIT specifies write access to an image or buffer in a <> or <> operation. * ename:VK_ACCESS_HOST_READ_BIT specifies read access by a host operation. Accesses of this type are not performed through a resource, but directly on memory. * ename:VK_ACCESS_HOST_WRITE_BIT specifies write access by a host operation. Accesses of this type are not performed through a resource, but directly on memory. * ename:VK_ACCESS_MEMORY_READ_BIT specifies read access via non-specific entities. These entities include the Vulkan device and host, but may: also include entities external to the Vulkan device or otherwise not part of the core Vulkan pipeline. When included in a destination access mask, makes all available writes visible to all future read accesses on entities known to the Vulkan device. * ename:VK_ACCESS_MEMORY_WRITE_BIT specifies write access via non-specific entities. These entities include the Vulkan device and host, but may: also include entities external to the Vulkan device or otherwise not part of the core Vulkan pipeline. When included in a source access mask, all writes that are performed by entities known to the Vulkan device are made available. When included in a destination access mask, makes all available writes visible to all future write accesses on entities known to the Vulkan device. ifdef::VK_EXT_conditional_rendering[] * ename:VK_ACCESS_CONDITIONAL_RENDERING_READ_BIT_EXT specifies read access to a predicate as part of conditional rendering. endif::VK_EXT_conditional_rendering[] ifdef::VK_EXT_transform_feedback[] * ename:VK_ACCESS_TRANSFORM_FEEDBACK_WRITE_BIT_EXT specifies write access to a transform feedback buffer made when transform feedback is active. * ename:VK_ACCESS_TRANSFORM_FEEDBACK_COUNTER_READ_BIT_EXT specifies read access to a transform feedback counter buffer which is read when fname:vkCmdBeginTransformFeedbackEXT executes. * ename:VK_ACCESS_TRANSFORM_FEEDBACK_COUNTER_WRITE_BIT_EXT specifies write access to a transform feedback counter buffer which is written when fname:vkCmdEndTransformFeedbackEXT executes. endif::VK_EXT_transform_feedback[] ifdef::VK_NVX_device_generated_commands[] * ename:VK_ACCESS_COMMAND_PROCESS_READ_BIT_NVX specifies reads from sname:VkBuffer inputs to flink:vkCmdProcessCommandsNVX. * ename:VK_ACCESS_COMMAND_PROCESS_WRITE_BIT_NVX specifies writes to the target command buffer in flink:vkCmdProcessCommandsNVX. endif::VK_NVX_device_generated_commands[] ifdef::VK_EXT_blend_operation_advanced[] * ename:VK_ACCESS_COLOR_ATTACHMENT_READ_NONCOHERENT_BIT_EXT is similar to ename:VK_ACCESS_COLOR_ATTACHMENT_READ_BIT, but also includes <>. endif::VK_EXT_blend_operation_advanced[] ifdef::VK_NV_shading_rate_image[] * ename:VK_ACCESS_SHADING_RATE_IMAGE_READ_BIT_NV specifies read access to a shading rate image as part of a drawing command, as bound by flink:vkCmdBindShadingRateImageNV. endif::VK_NV_shading_rate_image[] Certain access types are only performed by a subset of pipeline stages. Any synchronization command that takes both stage masks and access masks uses both to define the <> - only the specified access types performed by the specified stages are included in the access scope. An application must: not specify an access flag in a synchronization command if it does not include a pipeline stage in the corresponding stage mask that is able to perform accesses of that type. The following table lists, for each access flag, which pipeline stages can: perform that type of access. [[synchronization-access-types-supported]] .Supported access types [cols="50,50",options="header"] |==== |Access flag | Supported pipeline stages |ename:VK_ACCESS_INDIRECT_COMMAND_READ_BIT | ename:VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT |ename:VK_ACCESS_INDEX_READ_BIT | ename:VK_PIPELINE_STAGE_VERTEX_INPUT_BIT |ename:VK_ACCESS_VERTEX_ATTRIBUTE_READ_BIT | ename:VK_PIPELINE_STAGE_VERTEX_INPUT_BIT ifndef::VK_NV_mesh_shader[] |ename:VK_ACCESS_UNIFORM_READ_BIT | ename:VK_PIPELINE_STAGE_VERTEX_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT, ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT, ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT, or ename:VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT |ename:VK_ACCESS_SHADER_READ_BIT | ename:VK_PIPELINE_STAGE_VERTEX_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT, ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT, ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT, or ename:VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT |ename:VK_ACCESS_SHADER_WRITE_BIT | ename:VK_PIPELINE_STAGE_VERTEX_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT, ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT, ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT, or ename:VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT endif::VK_NV_mesh_shader[] ifdef::VK_NV_mesh_shader[] |ename:VK_ACCESS_UNIFORM_READ_BIT | ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT, ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT, ename:VK_PIPELINE_STAGE_VERTEX_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT, ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT, ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT, or ename:VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT |ename:VK_ACCESS_SHADER_READ_BIT | ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT, ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT, ename:VK_PIPELINE_STAGE_VERTEX_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT, ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT, ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT, or ename:VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT |ename:VK_ACCESS_SHADER_WRITE_BIT | ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT, ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT, ename:VK_PIPELINE_STAGE_VERTEX_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT, ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT, ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT, ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT, or ename:VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT endif::VK_NV_mesh_shader[] |ename:VK_ACCESS_INPUT_ATTACHMENT_READ_BIT | ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT |ename:VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT |ename:VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT |ename:VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | ename:VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT, or ename:VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT |ename:VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT | ename:VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT, or ename:VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT |ename:VK_ACCESS_TRANSFER_READ_BIT | ename:VK_PIPELINE_STAGE_TRANSFER_BIT |ename:VK_ACCESS_TRANSFER_WRITE_BIT | ename:VK_PIPELINE_STAGE_TRANSFER_BIT |ename:VK_ACCESS_HOST_READ_BIT | ename:VK_PIPELINE_STAGE_HOST_BIT |ename:VK_ACCESS_HOST_WRITE_BIT | ename:VK_PIPELINE_STAGE_HOST_BIT |ename:VK_ACCESS_MEMORY_READ_BIT | N/A |ename:VK_ACCESS_MEMORY_WRITE_BIT | N/A ifdef::VK_EXT_blend_operation_advanced[] |ename:VK_ACCESS_COLOR_ATTACHMENT_READ_NONCOHERENT_BIT_EXT | ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT endif::VK_EXT_blend_operation_advanced[] ifdef::VK_NVX_device_generated_commands[] |ename:VK_ACCESS_COMMAND_PROCESS_READ_BIT_NVX | ename:VK_PIPELINE_STAGE_COMMAND_PROCESS_BIT_NVX |ename:VK_ACCESS_COMMAND_PROCESS_WRITE_BIT_NVX | ename:VK_PIPELINE_STAGE_COMMAND_PROCESS_BIT_NVX endif::VK_NVX_device_generated_commands[] ifdef::VK_EXT_conditional_rendering[] |ename:VK_ACCESS_CONDITIONAL_RENDERING_READ_BIT_EXT | ename:VK_PIPELINE_STAGE_CONDITIONAL_RENDERING_BIT_EXT endif::VK_EXT_conditional_rendering[] ifdef::VK_NV_shading_rate_image[] |ename:VK_ACCESS_SHADING_RATE_IMAGE_BIT_NV | ename:VK_PIPELINE_STAGE_SHADING_RATE_IMAGE_BIT_NV endif::VK_NV_shading_rate_image[] ifdef::VK_EXT_transform_feedback[] |ename:VK_ACCESS_TRANSFORM_FEEDBACK_WRITE_BIT_EXT | ename:VK_PIPELINE_STAGE_TRANSFORM_FEEDBACK_BIT_EXT |ename:VK_ACCESS_TRANSFORM_FEEDBACK_COUNTER_WRITE_BIT_EXT | ename:VK_PIPELINE_STAGE_TRANSFORM_FEEDBACK_BIT_EXT |ename:VK_ACCESS_TRANSFORM_FEEDBACK_COUNTER_READ_BIT_EXT | ename:VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT endif::VK_EXT_transform_feedback[] |==== [[synchronization-host-access-types]] If a memory object does not have the ename:VK_MEMORY_PROPERTY_HOST_COHERENT_BIT property, then flink:vkFlushMappedMemoryRanges must: be called in order to guarantee that writes to the memory object from the host are made available to the host domain, where they can: be further made available to the device domain via a domain operation. Similarly, flink:vkInvalidateMappedMemoryRanges must: be called to guarantee that writes which are available to the host domain are made visible to host operations. If the memory object does have the ename:VK_MEMORY_PROPERTY_HOST_COHERENT_BIT property flag, writes to the memory object from the host are automatically made available to the host domain. Similarly, writes made available to the host domain are automatically made visible to the host. [NOTE] .Note ==== The flink:vkQueueSubmit command <> for all writes performed before the command executes, so in most cases an explicit memory barrier is not needed for this case. In the few circumstances where a submit does not occur between the host write and the device read access, writes can: be made available by using an explicit memory barrier. ==== -- [open,refpage='VkAccessFlags',desc='Bitmask of VkAccessFlagBits',type='enums'] -- include::../api/flags/VkAccessFlags.txt[] sname:VkAccessFlags is a bitmask type for setting a mask of zero or more slink:VkAccessFlagBits. -- [[synchronization-framebuffer-regions]] === Framebuffer Region Dependencies <> that operate on, or with respect to, the framebuffer are collectively the _framebuffer-space_ pipeline stages. These stages are: * ename:VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT * ename:VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT * ename:VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT * ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT For these pipeline stages, an execution or memory dependency from the first set of operations to the second set can: either be a single _framebuffer-global_ dependency, or split into multiple _framebuffer-local_ dependencies. A dependency with non-framebuffer-space pipeline stages is neither framebuffer-global nor framebuffer-local. A _framebuffer region_ is a set of sample (x, y, layer, sample) coordinates that is a subset of the entire framebuffer. Both <> of a framebuffer-local dependency include only the operations performed within corresponding framebuffer regions (as defined below). No ordering guarantees are made between different framebuffer regions for a framebuffer-local dependency. Both <> of a framebuffer-global dependency include operations on all framebuffer-regions. If the first synchronization scope includes operations on pixels/fragments with N samples and the second synchronization scope includes operations on pixels/fragments with M samples, where N does not equal M, then a framebuffer region containing all samples at a given (x, y, layer) coordinate in the first synchronization scope corresponds to a region containing all samples at the same coordinate in the second synchronization scope. In other words, it is a pixel granularity dependency. If N equals M, then a framebuffer region containing a single (x, y, layer, sample) coordinate in the first synchronization scope corresponds to a region containing the same sample at the same coordinate in the second synchronization scope. In other words, it is a sample granularity dependency. [NOTE] .Note ==== Since fragment invocations are not specified to run in any particular groupings, the size of a framebuffer region is implementation-dependent, not known to the application, and must: be assumed to be no larger than specified above. ==== [NOTE] .Note ==== Practically, the pixel vs sample granularity dependency means that if an input attachment has a different number of samples than the pipeline's pname:rasterizationSamples, then a fragment can: access any sample in the input attachment's pixel even if it only uses framebuffer-local dependencies. If the input attachment has the same number of samples, then the fragment can: only access the covered samples in its input code:SampleMask (i.e. the fragment operations happen-after a framebuffer-local dependency for each sample the fragment covers). To access samples that are not covered, a framebuffer-global dependency is required. ==== If a synchronization command includes a pname:dependencyFlags parameter, and specifies the ename:VK_DEPENDENCY_BY_REGION_BIT flag, then it defines framebuffer-local dependencies for the framebuffer-space pipeline stages in that synchronization command, for all framebuffer regions. If no pname:dependencyFlags parameter is included, or the ename:VK_DEPENDENCY_BY_REGION_BIT flag is not specified, then a framebuffer-global dependency is specified for those stages. The ename:VK_DEPENDENCY_BY_REGION_BIT flag does not affect the dependencies between non-framebuffer-space pipeline stages, nor does it affect the dependencies between framebuffer-space and non-framebuffer-space pipeline stages. [NOTE] .Note ==== Framebuffer-local dependencies are more optimal for most architectures; particularly tile-based architectures - which can keep framebuffer-regions entirely in on-chip registers and thus avoid external bandwidth across such a dependency. Including a framebuffer-global dependency in your rendering will usually force all implementations to flush data to memory, or to a higher level cache, breaking any potential locality optimizations. ==== ifdef::VK_VERSION_1_1,VK_KHR_multiview[] [[synchronization-view-local-dependencies]] === View-Local Dependencies In a render pass instance that has <> enabled, dependencies can: be either view-local or view-global. A view-local dependency only includes operations from a single <> from the source subpass in the first synchronization scope, and only includes operations from a single <> from the destination subpass in the second synchronization scope. A view-global dependency includes all views in the view mask of the source and destination subpasses in the corresponding synchronization scopes. If a synchronization command includes a pname:dependencyFlags parameter and specifies the ename:VK_DEPENDENCY_VIEW_LOCAL_BIT flag, then it defines view-local dependencies for that synchronization command, for all views. If no pname:dependencyFlags parameter is included or the ename:VK_DEPENDENCY_VIEW_LOCAL_BIT flag is not specified, then a view-global dependency is specified. endif::VK_VERSION_1_1,VK_KHR_multiview[] ifdef::VK_VERSION_1_1,VK_KHR_device_group[] [[synchronization-device-local-dependencies]] === Device-Local Dependencies Dependencies can: be either device-local or non-device-local. A device-local dependency acts as multiple separate dependencies, one for each physical device that executes the synchronization command, where each dependency only includes operations from that physical device in both synchronization scopes. A non-device-local dependency is a single dependency where both synchronization scopes include operations from all physical devices that participate in the synchronization command. For subpass dependencies, all physical devices in the slink:VkDeviceGroupRenderPassBeginInfo::pname:deviceMask participate in the dependency, and for pipeline barriers all physical devices that are set in the command buffer's current device mask participate in the dependency. If a synchronization command includes a pname:dependencyFlags parameter and specifies the ename:VK_DEPENDENCY_DEVICE_GROUP_BIT flag, then it defines a non-device-local dependency for that synchronization command. If no pname:dependencyFlags parameter is included or the ename:VK_DEPENDENCY_DEVICE_GROUP_BIT flag is not specified, then it defines device-local dependencies for that synchronization command, for all participating physical devices. Semaphore and event dependencies are device-local and only execute on the one physical device that performs the dependency. endif::VK_VERSION_1_1,VK_KHR_device_group[] [[synchronization-implicit]] == Implicit Synchronization Guarantees A small number of implicit ordering guarantees are provided by Vulkan, ensuring that the order in which commands are submitted is meaningful, and avoiding unnecessary complexity in common operations. [[synchronization-submission-order]] _Submission order_ is a fundamental ordering in Vulkan, giving meaning to the order in which <> are recorded and submitted to a single queue. Explicit and implicit ordering guarantees between commands in Vulkan all work on the premise that this ordering is meaningful. This order does not itself define any execution or memory dependencies; synchronization commands and other orderings within the API use this ordering to define their scopes. Submission order for any given set of commands is based on the order in which they were recorded to command buffers and then submitted. This order is determined as follows: . The initial order is determined by the order in which flink:vkQueueSubmit commands are executed on the host, for a single queue, from first to last. . The order in which slink:VkSubmitInfo structures are specified in the pname:pSubmits parameter of flink:vkQueueSubmit, from lowest index to highest. . The order in which command buffers are specified in the pname:pCommandBuffers member of slink:VkSubmitInfo, from lowest index to highest. . The order in which commands were recorded to a command buffer on the host, from first to last: ** For commands recorded outside a render pass, this includes all other commands recorded outside a render pass, including flink:vkCmdBeginRenderPass and flink:vkCmdEndRenderPass commands; it does not directly include commands inside a render pass. ** For commands recorded inside a render pass, this includes all other commands recorded inside the same subpass, including the flink:vkCmdBeginRenderPass and flink:vkCmdEndRenderPass commands that delimit the same render pass instance; it does not include commands recorded to other subpasses. <> recorded to a command buffer execute the ename:VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT pipeline stage in <> - forming an implicit execution dependency between this stage in each command. <> do not execute any operations on the device, instead they set the state of the command buffer when they execute on the host, in the order that they are recorded. <> consume the current state of the command buffer when they are recorded, and will execute state changes on the device as required to match the recorded state. <>, <> and <> provide additional guarantees based on submission order. Execution of <> within a given command also has a loose ordering, dependent only on a single command. [[synchronization-fences]] == Fences [open,refpage='VkFence',desc='Opaque handle to a fence object',type='handles'] -- Fences are a synchronization primitive that can: be used to insert a dependency from a queue to the host. Fences have two states - signaled and unsignaled. A fence can: be signaled as part of the execution of a <> command. Fences can: be unsignaled on the host with flink:vkResetFences. Fences can: be waited on by the host with the flink:vkWaitForFences command, and the current state can: be queried with flink:vkGetFenceStatus. ifdef::VK_VERSION_1_1,VK_KHR_external_fence[] [[synchronization-fences-payloads]] As with most objects in Vulkan, fences are an interface to internal data which is typically opaque to applications. This internal data is referred to as a fence's _payload_. However, in order to enable communication with agents outside of the current device, it is necessary to be able to export that payload to a commonly understood format, and subsequently import from that format as well. The internal data of a fence may: include a reference to any resources and pending work associated with signal or unsignal operations performed on that fence object. Mechanisms to import and export that internal data to and from fences are provided <>. These mechanisms indirectly enable applications to share fence state between two or more fences and other synchronization primitives across process and API boundaries. endif::VK_VERSION_1_1,VK_KHR_external_fence[] Fences are represented by sname:VkFence handles: include::../api/handles/VkFence.txt[] -- [open,refpage='vkCreateFence',desc='Create a new fence object',type='protos'] -- To create a fence, call: include::../api/protos/vkCreateFence.txt[] * pname:device is the logical device that creates the fence. * pname:pCreateInfo is a pointer to an instance of the sname:VkFenceCreateInfo structure which contains information about how the fence is to be created. * pname:pAllocator controls host memory allocation as described in the <> chapter. * pname:pFence points to a handle in which the resulting fence object is returned. include::../validity/protos/vkCreateFence.txt[] -- [open,refpage='VkFenceCreateInfo',desc='Structure specifying parameters of a newly created fence',type='structs'] -- The sname:VkFenceCreateInfo structure is defined as: include::../api/structs/VkFenceCreateInfo.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:flags is a bitmask of elink:VkFenceCreateFlagBits specifying the initial state and behavior of the fence. include::../validity/structs/VkFenceCreateInfo.txt[] -- [open,refpage='VkFenceCreateFlagBits',desc='Bitmask specifying initial state and behavior of a fence',type='enums'] -- include::../api/enums/VkFenceCreateFlagBits.txt[] * ename:VK_FENCE_CREATE_SIGNALED_BIT specifies that the fence object is created in the signaled state. Otherwise, it is created in the unsignaled state. -- [open,refpage='VkFenceCreateFlags',desc='Bitmask of VkFenceCreateFlagBits',type='enums'] -- include::../api/flags/VkFenceCreateFlags.txt[] sname:VkFenceCreateFlags is a bitmask type for setting a mask of zero or more slink:VkFenceCreateFlagBits. -- ifdef::VK_VERSION_1_1,VK_KHR_external_fence[] [open,refpage='VkExportFenceCreateInfo',desc='Structure specifying handle types that can be exported from a fence',type='structs'] -- To create a fence whose payload can: be exported to external handles, add the slink:VkExportFenceCreateInfo structure to the pname:pNext chain of the slink:VkFenceCreateInfo structure. The sname:VkExportFenceCreateInfo structure is defined as: include::../api/structs/VkExportFenceCreateInfo.txt[] ifdef::VK_KHR_external_fence[] or the equivalent include::../api/structs/VkExportFenceCreateInfoKHR.txt[] endif::VK_KHR_external_fence[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:handleTypes is a bitmask of elink:VkExternalFenceHandleTypeFlagBits specifying one or more fence handle types the application can: export from the resulting fence. The application can: request multiple handle types for the same fence. .Valid Usage **** * [[VUID-VkExportFenceCreateInfo-handleTypes-01446]] The bits in pname:handleTypes must be supported and compatible, as reported by slink:VkExternalFenceProperties. **** include::../validity/structs/VkExportFenceCreateInfo.txt[] -- endif::VK_VERSION_1_1,VK_KHR_external_fence[] ifdef::VK_KHR_external_fence_win32[] [open,refpage='VkExportFenceWin32HandleInfoKHR',desc='Structure specifying additional attributes of Windows handles exported from a fence',type='structs'] -- To specify additional attributes of NT handles exported from a fence, add the slink:VkExportFenceWin32HandleInfoKHR structure to the pname:pNext chain of the slink:VkFenceCreateInfo structure. The sname:VkExportFenceWin32HandleInfoKHR structure is defined as: include::../api/structs/VkExportFenceWin32HandleInfoKHR.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:pAttributes is a pointer to a Windows code:SECURITY_ATTRIBUTES structure specifying security attributes of the handle. * pname:dwAccess is a code:DWORD specifying access rights of the handle. * pname:name is a NULL-terminated UTF-16 string to associate with the underlying synchronization primitive referenced by NT handles exported from the created fence. If this structure is not present, or if pname:pAttributes is set to `NULL`, default security descriptor values will be used, and child processes created by the application will not inherit the handle, as described in the MSDN documentation for "`Synchronization Object Security and Access Rights`"^1^. Further, if the structure is not present, the access rights will be code:DXGI_SHARED_RESOURCE_READ | code:DXGI_SHARED_RESOURCE_WRITE for handles of the following types: ename:VK_EXTERNAL_FENCE_HANDLE_TYPE_OPAQUE_WIN32_BIT 1:: https://msdn.microsoft.com/en-us/library/windows/desktop/ms686670.aspx .Valid Usage **** * [[VUID-VkExportFenceWin32HandleInfoKHR-handleTypes-01447]] If slink:VkExportFenceCreateInfo::pname:handleTypes does not include ename:VK_EXTERNAL_FENCE_HANDLE_TYPE_OPAQUE_WIN32_BIT, sname:VkExportFenceWin32HandleInfoKHR must: not be in the pname:pNext chain of slink:VkFenceCreateInfo. **** include::../validity/structs/VkExportFenceWin32HandleInfoKHR.txt[] -- [open,refpage='vkGetFenceWin32HandleKHR',desc='Get a Windows HANDLE for a fence',type='protos'] -- To export a Windows handle representing the state of a fence, call: include::../api/protos/vkGetFenceWin32HandleKHR.txt[] * pname:device is the logical device that created the fence being exported. * pname:pGetWin32HandleInfo is a pointer to an instance of the slink:VkFenceGetWin32HandleInfoKHR structure containing parameters of the export operation. * pname:pHandle will return the Windows handle representing the fence state. For handle types defined as NT handles, the handles returned by fname:vkGetFenceWin32HandleKHR are owned by the application. To avoid leaking resources, the application must: release ownership of them using the fname:CloseHandle system call when they are no longer needed. Exporting a Windows handle from a fence may: have side effects depending on the transference of the specified handle type, as described in <>. include::../validity/protos/vkGetFenceWin32HandleKHR.txt[] -- [open,refpage='VkFenceGetWin32HandleInfoKHR',desc='Structure describing a Win32 handle fence export operation',type='structs'] -- The sname:VkFenceGetWin32HandleInfoKHR structure is defined as: include::../api/structs/VkFenceGetWin32HandleInfoKHR.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:fence is the fence from which state will be exported. * pname:handleType is the type of handle requested. The properties of the handle returned depend on the value of pname:handleType. See elink:VkExternalFenceHandleTypeFlagBits for a description of the properties of the defined external fence handle types. .Valid Usage **** * [[VUID-VkFenceGetWin32HandleInfoKHR-handleType-01448]] pname:handleType must: have been included in slink:VkExportFenceCreateInfo::pname:handleTypes when the pname:fence's current payload was created. * [[VUID-VkFenceGetWin32HandleInfoKHR-handleType-01449]] If pname:handleType is defined as an NT handle, flink:vkGetFenceWin32HandleKHR must: be called no more than once for each valid unique combination of pname:fence and pname:handleType. * [[VUID-VkFenceGetWin32HandleInfoKHR-fence-01450]] pname:fence must: not currently have its payload replaced by an imported payload as described below in <> unless that imported payload's handle type was included in slink:VkExternalFenceProperties::pname:exportFromImportedHandleTypes for pname:handleType. * [[VUID-VkFenceGetWin32HandleInfoKHR-handleType-01451]] If pname:handleType refers to a handle type with copy payload transference semantics, pname:fence must: be signaled, or have an associated <> pending execution. * [[VUID-VkFenceGetWin32HandleInfoKHR-handleType-01452]] pname:handleType must: be defined as an NT handle or a global share handle. **** include::../validity/structs/VkFenceGetWin32HandleInfoKHR.txt[] -- endif::VK_KHR_external_fence_win32[] ifdef::VK_KHR_external_fence_fd[] [open,refpage='vkGetFenceFdKHR',desc='Get a POSIX file descriptor handle for a fence',type='protos'] -- To export a POSIX file descriptor representing the payload of a fence, call: include::../api/protos/vkGetFenceFdKHR.txt[] * pname:device is the logical device that created the fence being exported. * pname:pGetFdInfo is a pointer to an instance of the slink:VkFenceGetFdInfoKHR structure containing parameters of the export operation. * pname:pFd will return the file descriptor representing the fence payload. Each call to fname:vkGetFenceFdKHR must: create a new file descriptor and transfer ownership of it to the application. To avoid leaking resources, the application must: release ownership of the file descriptor when it is no longer needed. [NOTE] .Note ==== Ownership can be released in many ways. For example, the application can call code:close() on the file descriptor, or transfer ownership back to Vulkan by using the file descriptor to import a fence payload. ==== If pname:pGetFdInfo\->pname:handleType is ename:VK_EXTERNAL_FENCE_HANDLE_TYPE_SYNC_FD_BIT and the fence is signaled at the time fname:vkGetFenceFdKHR is called, pname:pFd may: return the value `-1` instead of a valid file descriptor. Where supported by the operating system, the implementation must: set the file descriptor to be closed automatically when an code:execve system call is made. Exporting a file descriptor from a fence may: have side effects depending on the transference of the specified handle type, as described in <>. include::../validity/protos/vkGetFenceFdKHR.txt[] -- [open,refpage='VkFenceGetFdInfoKHR',desc='Structure describing a POSIX FD fence export operation',type='structs'] -- The sname:VkFenceGetFdInfoKHR structure is defined as: include::../api/structs/VkFenceGetFdInfoKHR.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:fence is the fence from which state will be exported. * pname:handleType is the type of handle requested. The properties of the file descriptor returned depend on the value of pname:handleType. See elink:VkExternalFenceHandleTypeFlagBits for a description of the properties of the defined external fence handle types. .Valid Usage **** * [[VUID-VkFenceGetFdInfoKHR-handleType-01453]] pname:handleType must: have been included in slink:VkExportFenceCreateInfo::pname:handleTypes when pname:fence's current payload was created. * [[VUID-VkFenceGetFdInfoKHR-handleType-01454]] If pname:handleType refers to a handle type with copy payload transference semantics, pname:fence must: be signaled, or have an associated <> pending execution. * [[VUID-VkFenceGetFdInfoKHR-fence-01455]] pname:fence must: not currently have its payload replaced by an imported payload as described below in <> unless that imported payload's handle type was included in slink:VkExternalFenceProperties::pname:exportFromImportedHandleTypes for pname:handleType. * [[VUID-VkFenceGetFdInfoKHR-handleType-01456]] pname:handleType must: be defined as a POSIX file descriptor handle. **** include::../validity/structs/VkFenceGetFdInfoKHR.txt[] -- endif::VK_KHR_external_fence_fd[] [open,refpage='vkDestroyFence',desc='Destroy a fence object',type='protos'] -- To destroy a fence, call: include::../api/protos/vkDestroyFence.txt[] * pname:device is the logical device that destroys the fence. * pname:fence is the handle of the fence to destroy. * pname:pAllocator controls host memory allocation as described in the <> chapter. .Valid Usage **** * [[VUID-vkDestroyFence-fence-01120]] All <> commands that refer to pname:fence must: have completed execution * [[VUID-vkDestroyFence-fence-01121]] If sname:VkAllocationCallbacks were provided when pname:fence was created, a compatible set of callbacks must: be provided here * [[VUID-vkDestroyFence-fence-01122]] If no sname:VkAllocationCallbacks were provided when pname:fence was created, pname:pAllocator must: be `NULL` **** include::../validity/protos/vkDestroyFence.txt[] -- [open,refpage='vkGetFenceStatus',desc='Return the status of a fence',type='protos'] -- To query the status of a fence from the host, call: include::../api/protos/vkGetFenceStatus.txt[] * pname:device is the logical device that owns the fence. * pname:fence is the handle of the fence to query. Upon success, fname:vkGetFenceStatus returns the status of the fence object, with the following return codes: .Fence Object Status Codes [width="80%",options="header"] |==== | Status | Meaning | ename:VK_SUCCESS | The fence specified by pname:fence is signaled. | ename:VK_NOT_READY | The fence specified by pname:fence is unsignaled. | ename:VK_ERROR_DEVICE_LOST | The device has been lost. See <>. |==== If a <> command is pending execution, then the value returned by this command may: immediately be out of date. If the device has been lost (see <>), fname:vkGetFenceStatus may: return any of the above status codes. If the device has been lost and fname:vkGetFenceStatus is called repeatedly, it will eventually return either ename:VK_SUCCESS or ename:VK_ERROR_DEVICE_LOST. include::../validity/protos/vkGetFenceStatus.txt[] -- [[synchronization-fences-unsignaling]] [open,refpage='vkResetFences',desc='Resets one or more fence objects',type='protos'] -- To set the state of fences to unsignaled from the host, call: include::../api/protos/vkResetFences.txt[] * pname:device is the logical device that owns the fences. * pname:fenceCount is the number of fences to reset. * pname:pFences is a pointer to an array of fence handles to reset. ifdef::VK_VERSION_1_1,VK_KHR_external_fence[] If any member of pname:pFences currently has its <> with temporary permanence, that fence's prior permanent payload is first restored. The remaining operations described therefore operate on the restored payload. endif::VK_VERSION_1_1,VK_KHR_external_fence[] When flink:vkResetFences is executed on the host, it defines a _fence unsignal operation_ for each fence, which resets the fence to the unsignaled state. If any member of pname:pFences is already in the unsignaled state when flink:vkResetFences is executed, then flink:vkResetFences has no effect on that fence. .Valid Usage **** * [[VUID-vkResetFences-pFences-01123]] Each element of pname:pFences must: not be currently associated with any queue command that has not yet completed execution on that queue **** include::../validity/protos/vkResetFences.txt[] -- [[synchronization-fences-signaling]] When a fence is submitted to a queue as part of a <> command, it defines a memory dependency on the batches that were submitted as part of that command, and defines a _fence signal operation_ which sets the fence to the signaled state. The first <> includes every batch submitted in the same <> command. Fence signal operations that are defined by flink:vkQueueSubmit additionally include in the first synchronization scope all commands that occur earlier in <>. The second <> only includes the fence signal operation. The first <> includes all memory access performed by the device. The second <> is empty. [open,refpage='vkWaitForFences',desc='Wait for one or more fences to become signaled',type='protos'] -- To wait for one or more fences to enter the signaled state on the host, call: include::../api/protos/vkWaitForFences.txt[] * pname:device is the logical device that owns the fences. * pname:fenceCount is the number of fences to wait on. * pname:pFences is a pointer to an array of pname:fenceCount fence handles. * pname:waitAll is the condition that must: be satisfied to successfully unblock the wait. If pname:waitAll is ename:VK_TRUE, then the condition is that all fences in pname:pFences are signaled. Otherwise, the condition is that at least one fence in pname:pFences is signaled. * pname:timeout is the timeout period in units of nanoseconds. pname:timeout is adjusted to the closest value allowed by the implementation-dependent timeout accuracy, which may: be substantially longer than one nanosecond, and may: be longer than the requested period. If the condition is satisfied when fname:vkWaitForFences is called, then fname:vkWaitForFences returns immediately. If the condition is not satisfied at the time fname:vkWaitForFences is called, then fname:vkWaitForFences will block and wait up to pname:timeout nanoseconds for the condition to become satisfied. If pname:timeout is zero, then fname:vkWaitForFences does not wait, but simply returns the current state of the fences. ename:VK_TIMEOUT will be returned in this case if the condition is not satisfied, even though no actual wait was performed. If the specified timeout period expires before the condition is satisfied, fname:vkWaitForFences returns ename:VK_TIMEOUT. If the condition is satisfied before pname:timeout nanoseconds has expired, fname:vkWaitForFences returns ename:VK_SUCCESS. If device loss occurs (see <>) before the timeout has expired, fname:vkWaitForFences must: return in finite time with either ename:VK_SUCCESS or ename:VK_ERROR_DEVICE_LOST. [NOTE] .Note ==== While we guarantee that fname:vkWaitForFences must: return in finite time, no guarantees are made that it returns immediately upon device loss. However, the client can reasonably expect that the delay will be on the order of seconds and that calling fname:vkWaitForFences will not result in a permanently (or seemingly permanently) dead process. ==== include::../validity/protos/vkWaitForFences.txt[] -- [[synchronization-fences-waiting]] An execution dependency is defined by waiting for a fence to become signaled, either via flink:vkWaitForFences or by polling on flink:vkGetFenceStatus. The first <> includes only the fence signal operation. The second <> includes the host operations of flink:vkWaitForFences or flink:vkGetFenceStatus indicating that the fence has become signaled. [NOTE] .Note ==== Signaling a fence and waiting on the host does not guarantee that the results of memory accesses will be visible to the host, as the access scope of a memory dependency defined by a fence only includes device access. A <> or other memory dependency must: be used to guarantee this. See the description of <> for more information. ==== ifdef::VK_EXT_display_control[] include::VK_EXT_display_control/fence_events.txt[] endif::VK_EXT_display_control[] ifdef::VK_VERSION_1_1,VK_KHR_external_fence[] [[synchronization-fences-importing]] === Importing Fence Payloads Applications can: import a fence payload into an existing fence using an external fence handle. The effects of the import operation will be either temporary or permanent, as specified by the application. If the import is temporary, the fence will be _restored_ to its permanent state the next time that fence is passed to flink:vkResetFences. [NOTE] .Note ==== Restoring a fence to its prior permanent payload is a distinct operation from resetting a fence payload. See flink:vkResetFences for more detail. ==== Performing a subsequent temporary import on a fence before resetting it has no effect on this requirement; the next unsignal of the fence must: still restore its last permanent state. A permanent payload import behaves as if the target fence was destroyed, and a new fence was created with the same handle but the imported payload. Because importing a fence payload temporarily or permanently detaches the existing payload from a fence, similar usage restrictions to those applied to fname:vkDestroyFence are applied to any command that imports a fence payload. Which of these import types is used is referred to as the import operation's _permanence_. Each handle type supports either one or both types of permanence. The implementation must: perform the import operation by either referencing or copying the payload referred to by the specified external fence handle, depending on the handle's type. The import method used is referred to as the handle type's _transference_. When using handle types with reference transference, importing a payload to a fence adds the fence to the set of all fences sharing that payload. This set includes the fence from which the payload was exported. Fence signaling, waiting, and resetting operations performed on any fence in the set must: behave as if the set were a single fence. Importing a payload using handle types with copy transference creates a duplicate copy of the payload at the time of import, but makes no further reference to it. Fence signaling, waiting, and resetting operations performed on the target of copy imports must: not affect any other fence or payload. Export operations have the same transference as the specified handle type's import operations. Additionally, exporting a fence payload to a handle with copy transference has the same side effects on the source fence's payload as executing a fence reset operation. If the fence was using a temporarily imported payload, the fence's prior permanent payload will be restored. ifdef::VK_KHR_external_fence_win32,VK_KHR_external_fence_fd[] [NOTE] .Note ==== The ifdef::VK_KHR_external_fence_win32+VK_KHR_external_fence_fd[tables] ifndef::VK_KHR_external_fence_win32+VK_KHR_external_fence_fd[table] ifdef::VK_KHR_external_fence_win32[] <> endif::VK_KHR_external_fence_win32[] ifdef::VK_KHR_external_fence_win32+VK_KHR_external_fence_fd[and] ifdef::VK_KHR_external_fence_fd[] <> endif::VK_KHR_external_fence_fd[] ifdef::VK_KHR_external_fence_win32+VK_KHR_external_fence_fd[define] ifndef::VK_KHR_external_fence_win32+VK_KHR_external_fence_fd[defines] the permanence and transference of each handle type. ==== endif::VK_KHR_external_fence_win32,VK_KHR_external_fence_fd[] <> allows implementations to modify an object's internal state, i.e. payload, without internal synchronization. However, for fences sharing a payload across processes, satisfying the external synchronization requirements of fname:VkFence parameters as if all fences in the set were the same object is sometimes infeasible. Satisfying valid usage constraints on the state of a fence would similarly require impractical coordination or levels of trust between processes. Therefore, these constraints only apply to a specific fence handle, not to its payload. For distinct fence objects which share a payload: * If multiple commands which queue a signal operation, or which unsignal a fence, are called concurrently, behavior will be as if the commands were called in an arbitrary sequential order. * If a queue submission command is called with a fence that is sharing a payload, and the payload is already associated with another queue command that has not yet completed execution, either one or both of the commands will cause the fence to become signaled when they complete execution. * If a fence payload is reset while it is associated with a queue command that has not yet completed execution, the payload will become unsignaled, but may: become signaled again when the command completes execution. * In the preceding cases, any of the devices associated with the fences sharing the payload may: be lost, or any of the queue submission or fence reset commands may: return ename:VK_ERROR_INITIALIZATION_FAILED. Other than these non-deterministic results, behavior is well defined. In particular: * The implementation must: not crash or enter an internally inconsistent state where future valid Vulkan commands might cause undefined: results, * Timeouts on future wait commands on fences sharing the payload must: be effective. [NOTE] .Note ==== These rules allow processes to synchronize access to shared memory without trusting each other. However, such processes must still be cautious not to use the shared fence for more than synchronizing access to the shared memory. For example, a process should not use a fence with shared payload to tell when commands it submitted to a queue have completed and objects used by those commands may be destroyed, since the other process can accidentally or maliciously cause the fence to signal before the commands actually complete. ==== When a fence is using an imported payload, its slink:VkExportFenceCreateInfo::pname:handleTypes value is that specified when creating the fence from which the payload was exported, rather than that specified when creating the fence. Additionally, slink:VkExternalFenceProperties::exportFromImportedHandleTypes restricts which handle types can: be exported from such a fence based on the specific handle type used to import the current payload. ifdef::VK_KHR_swapchain[] Passing a fence to flink:vkAcquireNextImageKHR is equivalent to temporarily importing a fence payload to that fence. [NOTE] .Note ==== Because the exportable handle types of an imported fence correspond to its current imported payload, and flink:vkAcquireNextImageKHR behaves the same as a temporary import operation for which the source fence is opaque to the application, applications have no way of determining whether any external handle types can: be exported from a fence in this state. Therefore, applications must: not attempt to export handles from fences using a temporarily imported payload from flink:vkAcquireNextImageKHR. ==== endif::VK_KHR_swapchain[] When importing a fence payload, it is the responsibility of the application to ensure the external handles meet all valid usage requirements. However, implementations must: perform sufficient validation of external handles to ensure that the operation results in a valid fence which will not cause program termination, device loss, queue stalls, host thread stalls, or corruption of other resources when used as allowed according to its import parameters. If the external handle provided does not meet these requirements, the implementation must: fail the fence payload import operation with the error code ename:VK_ERROR_INVALID_EXTERNAL_HANDLE. endif::VK_VERSION_1_1,VK_KHR_external_fence[] ifdef::VK_KHR_external_fence_win32[] [open,refpage='vkImportFenceWin32HandleKHR',desc='Import a fence from a Windows HANDLE',type='protos'] -- To import a fence payload from a Windows handle, call: include::../api/protos/vkImportFenceWin32HandleKHR.txt[] * pname:device is the logical device that created the fence. * pname:pImportFenceWin32HandleInfo points to a slink:VkImportFenceWin32HandleInfoKHR structure specifying the fence and import parameters. Importing a fence payload from Windows handles does not transfer ownership of the handle to the Vulkan implementation. For handle types defined as NT handles, the application must: release ownership using the fname:CloseHandle system call when the handle is no longer needed. Applications can: import the same fence payload into multiple instances of Vulkan, into the same instance from which it was exported, and multiple times into a given Vulkan instance. include::../validity/protos/vkImportFenceWin32HandleKHR.txt[] -- [open,refpage='VkImportFenceWin32HandleInfoKHR',desc='(None)',type='structs'] -- The sname:VkImportFenceWin32HandleInfoKHR structure is defined as: include::../api/structs/VkImportFenceWin32HandleInfoKHR.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:fence is the fence into which the state will be imported. * pname:flags is a bitmask of elink:VkFenceImportFlagBits specifying additional parameters for the fence payload import operation. * pname:handleType specifies the type of pname:handle. * pname:handle is the external handle to import, or `NULL`. * pname:name is the NULL-terminated UTF-16 string naming the underlying synchronization primitive to import, or `NULL`. The handle types supported by pname:handleType are: [[synchronization-fence-handletypes-win32]] .Handle Types Supported by VkImportFenceWin32HandleInfoKHR [width="80%",options="header"] |==== | Handle Type | Transference | Permanence Supported | ename:VK_EXTERNAL_FENCE_HANDLE_TYPE_OPAQUE_WIN32_BIT | Reference | Temporary,Permanent | ename:VK_EXTERNAL_FENCE_HANDLE_TYPE_OPAQUE_WIN32_KMT_BIT | Reference | Temporary,Permanent |==== .Valid Usage **** * [[VUID-VkImportFenceWin32HandleInfoKHR-handleType-01457]] pname:handleType must: be a value included in the <> table. * [[VUID-VkImportFenceWin32HandleInfoKHR-handleType-01459]] If pname:handleType is not ename:VK_EXTERNAL_FENCE_HANDLE_TYPE_OPAQUE_WIN32_BIT, pname:name must: be `NULL`. * [[VUID-VkImportFenceWin32HandleInfoKHR-handleType-01460]] If pname:handleType is not `0` and pname:handle is `NULL`, pname:name must: name a valid synchronization primitive of the type specified by pname:handleType. * [[VUID-VkImportFenceWin32HandleInfoKHR-handleType-01461]] If pname:handleType is not `0` and pname:name is `NULL`, pname:handle must: be a valid handle of the type specified by pname:handleType. * [[VUID-VkImportFenceWin32HandleInfoKHR-handle-01462]] If pname:handle is not `NULL`, pname:name must be `NULL`. * [[VUID-VkImportFenceWin32HandleInfoKHR-handle-01539]] If pname:handle is not `NULL`, it must: obey any requirements listed for pname:handleType in <>. * [[VUID-VkImportFenceWin32HandleInfoKHR-name-01540]] If pname:name is not `NULL`, it must: obey any requirements listed for pname:handleType in <>. **** include::../validity/structs/VkImportFenceWin32HandleInfoKHR.txt[] -- endif::VK_KHR_external_fence_win32[] ifdef::VK_KHR_external_fence_fd[] [open,refpage='vkImportFenceFdKHR',desc='Import a fence from a POSIX file descriptor',type='protos'] -- To import a fence payload from a POSIX file descriptor, call: include::../api/protos/vkImportFenceFdKHR.txt[] * pname:device is the logical device that created the fence. * pname:pImportFenceFdInfo points to a slink:VkImportFenceFdInfoKHR structure specifying the fence and import parameters. Importing a fence payload from a file descriptor transfers ownership of the file descriptor from the application to the Vulkan implementation. The application must: not perform any operations on the file descriptor after a successful import. Applications can: import the same fence payload into multiple instances of Vulkan, into the same instance from which it was exported, and multiple times into a given Vulkan instance. .Valid Usage **** * [[VUID-vkImportFenceFdKHR-fence-01463]] pname:fence must: not be associated with any queue command that has not yet completed execution on that queue **** include::../validity/protos/vkImportFenceFdKHR.txt[] -- [open,refpage='VkImportFenceFdInfoKHR',desc='(None)',type='structs'] -- The sname:VkImportFenceFdInfoKHR structure is defined as: include::../api/structs/VkImportFenceFdInfoKHR.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:fence is the fence into which the payload will be imported. * pname:flags is a bitmask of elink:VkFenceImportFlagBits specifying additional parameters for the fence payload import operation. * pname:handleType specifies the type of pname:fd. * pname:fd is the external handle to import. The handle types supported by pname:handleType are: [[synchronization-fence-handletypes-fd]] .Handle Types Supported by VkImportFenceFdInfoKHR [width="80%",options="header"] |==== | Handle Type | Transference | Permanence Supported | ename:VK_EXTERNAL_FENCE_HANDLE_TYPE_OPAQUE_FD_BIT | Reference | Temporary,Permanent | ename:VK_EXTERNAL_FENCE_HANDLE_TYPE_SYNC_FD_BIT | Copy | Temporary |==== .Valid Usage **** * [[VUID-VkImportFenceFdInfoKHR-handleType-01464]] pname:handleType must: be a value included in the <> table. * [[VUID-VkImportFenceFdInfoKHR-fd-01541]] pname:fd must: obey any requirements listed for pname:handleType in <>. **** If pname:handleType is ename:VK_EXTERNAL_FENCE_HANDLE_TYPE_SYNC_FD_BIT, the special value `-1` for pname:fd is treated like a valid sync file descriptor referring to an object that has already signaled. The import operation will succeed and the sname:VkFence will have a temporarily imported payload as if a valid file descriptor had been provided. [NOTE] .Note ==== This special behavior for importing an invalid sync file descriptor allows easier interoperability with other system APIs which use the convention that an invalid sync file descriptor represents work that has already completed and does not need to be waited for. It is consistent with the option for implementations to return a `-1` file descriptor when exporting a ename:VK_EXTERNAL_FENCE_HANDLE_TYPE_SYNC_FD_BIT from a sname:VkFence which is signaled. ==== include::../validity/structs/VkImportFenceFdInfoKHR.txt[] -- endif::VK_KHR_external_fence_fd[] ifdef::VK_VERSION_1_1,VK_KHR_external_fence[] ifdef::VK_KHR_external_fence_win32,VK_KHR_external_fence_fd[] [open,refpage='VkFenceImportFlagBits',desc='Bitmask specifying additional parameters of fence payload import',type='enums'] -- Bits which can: be set in ifdef::VK_KHR_external_fence_win32[] slink:VkImportFenceWin32HandleInfoKHR::pname:flags endif::VK_KHR_external_fence_win32[] ifdef::VK_KHR_external_fence_win32+VK_KHR_external_fence_fd[and] ifdef::VK_KHR_external_fence_fd[] slink:VkImportFenceFdInfoKHR::pname:flags endif::VK_KHR_external_fence_fd[] specifying additional parameters of a fence import operation are: include::../api/enums/VkFenceImportFlagBits.txt[] ifdef::VK_KHR_external_fence[] or the equivalent include::../api/enums/VkFenceImportFlagBitsKHR.txt[] endif::VK_KHR_external_fence[] * ename:VK_FENCE_IMPORT_TEMPORARY_BIT specifies that the fence payload will be imported only temporarily, as described in <>, regardless of the permanence of pname:handleType. -- [open,refpage='VkFenceImportFlags',desc='Bitmask of VkFenceImportFlagBits',type='enums'] -- include::../api/flags/VkFenceImportFlags.txt[] ifdef::VK_KHR_external_fence[] or the equivalent include::../api/flags/VkFenceImportFlagsKHR.txt[] endif::VK_KHR_external_fence[] sname:VkFenceImportFlags is a bitmask type for setting a mask of zero or more slink:VkFenceImportFlagBits. -- endif::VK_KHR_external_fence_win32,VK_KHR_external_fence_fd[] endif::VK_VERSION_1_1,VK_KHR_external_fence[] [[synchronization-semaphores]] == Semaphores [open,refpage='VkSemaphore',desc='Opaque handle to a semaphore object',type='handles'] -- Semaphores are a synchronization primitive that can: be used to insert a dependency between batches submitted to queues. Semaphores have two states - signaled and unsignaled. The state of a semaphore can: be signaled after execution of a batch of commands is completed. A batch can: wait for a semaphore to become signaled before it begins execution, and the semaphore is also unsignaled before the batch begins execution. ifdef::VK_VERSION_1_1,VK_KHR_external_semaphore[] [[synchronization-semaphores-payloads]] As with most objects in Vulkan, semaphores are an interface to internal data which is typically opaque to applications. This internal data is referred to as a semaphore's _payload_. However, in order to enable communication with agents outside of the current device, it is necessary to be able to export that payload to a commonly understood format, and subsequently import from that format as well. The internal data of a semaphore may: include a reference to any resources and pending work associated with signal or unsignal operations performed on that semaphore object. Mechanisms to import and export that internal data to and from semaphores are provided <>. These mechanisms indirectly enable applications to share semaphore state between two or more semaphores and other synchronization primitives across process and API boundaries. endif::VK_VERSION_1_1,VK_KHR_external_semaphore[] Semaphores are represented by sname:VkSemaphore handles: include::../api/handles/VkSemaphore.txt[] -- [open,refpage='vkCreateSemaphore',desc='Create a new queue semaphore object',type='protos'] -- To create a semaphore, call: include::../api/protos/vkCreateSemaphore.txt[] * pname:device is the logical device that creates the semaphore. * pname:pCreateInfo is a pointer to an instance of the sname:VkSemaphoreCreateInfo structure which contains information about how the semaphore is to be created. * pname:pAllocator controls host memory allocation as described in the <> chapter. * pname:pSemaphore points to a handle in which the resulting semaphore object is returned. When created, the semaphore is in the unsignaled state. include::../validity/protos/vkCreateSemaphore.txt[] -- [open,refpage='VkSemaphoreCreateInfo',desc='Structure specifying parameters of a newly created semaphore',type='structs'] -- The sname:VkSemaphoreCreateInfo structure is defined as: include::../api/structs/VkSemaphoreCreateInfo.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:flags is reserved for future use. include::../validity/structs/VkSemaphoreCreateInfo.txt[] -- [open,refpage='VkSemaphoreCreateFlags',desc='Reserved for future use',type='enums'] -- include::../api/flags/VkSemaphoreCreateFlags.txt[] sname:VkSemaphoreCreateFlags is a bitmask type for setting a mask, but is currently reserved for future use. -- ifdef::VK_VERSION_1_1,VK_KHR_external_semaphore[] [open,refpage='VkExportSemaphoreCreateInfo',desc='Structure specifying handle types that can be exported from a semaphore',type='structs'] -- To create a semaphore whose payload can: be exported to external handles, add the slink:VkExportSemaphoreCreateInfo structure to the pname:pNext chain of the slink:VkSemaphoreCreateInfo structure. The sname:VkExportSemaphoreCreateInfo structure is defined as: include::../api/structs/VkExportSemaphoreCreateInfo.txt[] ifdef::VK_KHR_external_semaphore[] or the equivalent include::../api/structs/VkExportSemaphoreCreateInfoKHR.txt[] endif::VK_KHR_external_semaphore[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:handleTypes is a bitmask of elink:VkExternalSemaphoreHandleTypeFlagBits specifying one or more semaphore handle types the application can: export from the resulting semaphore. The application can: request multiple handle types for the same semaphore. .Valid Usage **** * [[VUID-VkExportSemaphoreCreateInfo-handleTypes-01124]] The bits in pname:handleTypes must: be supported and compatible, as reported by slink:VkExternalSemaphoreProperties. **** include::../validity/structs/VkExportSemaphoreCreateInfo.txt[] -- endif::VK_VERSION_1_1,VK_KHR_external_semaphore[] ifdef::VK_KHR_external_semaphore_win32[] [open,refpage='VkExportSemaphoreWin32HandleInfoKHR',desc='Structure specifying additional attributes of Windows handles exported from a semaphore',type='structs'] -- To specify additional attributes of NT handles exported from a semaphore, add the sname:VkExportSemaphoreWin32HandleInfoKHR structure to the pname:pNext chain of the slink:VkSemaphoreCreateInfo structure. The sname:VkExportSemaphoreWin32HandleInfoKHR structure is defined as: include::../api/structs/VkExportSemaphoreWin32HandleInfoKHR.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:pAttributes is a pointer to a Windows code:SECURITY_ATTRIBUTES structure specifying security attributes of the handle. * pname:dwAccess is a code:DWORD specifying access rights of the handle. * pname:name is a NULL-terminated UTF-16 string to associate with the underlying synchronization primitive referenced by NT handles exported from the created semaphore. If this structure is not present, or if pname:pAttributes is set to `NULL`, default security descriptor values will be used, and child processes created by the application will not inherit the handle, as described in the MSDN documentation for "`Synchronization Object Security and Access Rights`"^1^. Further, if the structure is not present, the access rights will be code:DXGI_SHARED_RESOURCE_READ | code:DXGI_SHARED_RESOURCE_WRITE for handles of the following types: ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32_BIT And code:GENERIC_ALL for handles of the following types: ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_D3D12_FENCE_BIT 1:: https://msdn.microsoft.com/en-us/library/windows/desktop/ms686670.aspx .Valid Usage **** * [[VUID-VkExportSemaphoreWin32HandleInfoKHR-handleTypes-01125]] If slink:VkExportSemaphoreCreateInfo::pname:handleTypes does not include ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32_BIT or ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_D3D12_FENCE_BIT, sname:VkExportSemaphoreWin32HandleInfoKHR must: not be in the pname:pNext chain of slink:VkSemaphoreCreateInfo. **** include::../validity/structs/VkExportSemaphoreWin32HandleInfoKHR.txt[] -- [open,refpage='vkGetSemaphoreWin32HandleKHR',desc='Get a Windows HANDLE for a semaphore',type='protos'] -- To export a Windows handle representing the payload of a semaphore, call: include::../api/protos/vkGetSemaphoreWin32HandleKHR.txt[] * pname:device is the logical device that created the semaphore being exported. * pname:pGetWin32HandleInfo is a pointer to an instance of the slink:VkSemaphoreGetWin32HandleInfoKHR structure containing parameters of the export operation. * pname:pHandle will return the Windows handle representing the semaphore state. For handle types defined as NT handles, the handles returned by fname:vkGetSemaphoreWin32HandleKHR are owned by the application. To avoid leaking resources, the application must: release ownership of them using the fname:CloseHandle system call when they are no longer needed. Exporting a Windows handle from a semaphore may: have side effects depending on the transference of the specified handle type, as described in <>. include::../validity/protos/vkGetSemaphoreWin32HandleKHR.txt[] -- [open,refpage='VkSemaphoreGetWin32HandleInfoKHR',desc='Structure describing a Win32 handle semaphore export operation',type='structs'] -- The sname:VkSemaphoreGetWin32HandleInfoKHR structure is defined as: include::../api/structs/VkSemaphoreGetWin32HandleInfoKHR.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:semaphore is the semaphore from which state will be exported. * pname:handleType is the type of handle requested. The properties of the handle returned depend on the value of pname:handleType. See elink:VkExternalSemaphoreHandleTypeFlagBits for a description of the properties of the defined external semaphore handle types. .Valid Usage **** * [[VUID-VkSemaphoreGetWin32HandleInfoKHR-handleType-01126]] pname:handleType must: have been included in slink:VkExportSemaphoreCreateInfo::pname:handleTypes when the pname:semaphore's current payload was created. * [[VUID-VkSemaphoreGetWin32HandleInfoKHR-handleType-01127]] If pname:handleType is defined as an NT handle, flink:vkGetSemaphoreWin32HandleKHR must: be called no more than once for each valid unique combination of pname:semaphore and pname:handleType. * [[VUID-VkSemaphoreGetWin32HandleInfoKHR-semaphore-01128]] pname:semaphore must: not currently have its payload replaced by an imported payload as described below in <> unless that imported payload's handle type was included in slink:VkExternalSemaphoreProperties::pname:exportFromImportedHandleTypes for pname:handleType. * [[VUID-VkSemaphoreGetWin32HandleInfoKHR-handleType-01129]] If pname:handleType refers to a handle type with copy payload transference semantics, as defined below in <>, there must: be no queue waiting on pname:semaphore. * [[VUID-VkSemaphoreGetWin32HandleInfoKHR-handleType-01130]] If pname:handleType refers to a handle type with copy payload transference semantics, pname:semaphore must: be signaled, or have an associated <> pending execution. * [[VUID-VkSemaphoreGetWin32HandleInfoKHR-handleType-01131]] pname:handleType must: be defined as an NT handle or a global share handle. **** include::../validity/structs/VkSemaphoreGetWin32HandleInfoKHR.txt[] -- endif::VK_KHR_external_semaphore_win32[] ifdef::VK_KHR_external_semaphore_fd[] [open,refpage='vkGetSemaphoreFdKHR',desc='Get a POSIX file descriptor handle for a semaphore',type='protos'] -- To export a POSIX file descriptor representing the payload of a semaphore, call: include::../api/protos/vkGetSemaphoreFdKHR.txt[] * pname:device is the logical device that created the semaphore being exported. * pname:pGetFdInfo is a pointer to an instance of the slink:VkSemaphoreGetFdInfoKHR structure containing parameters of the export operation. * pname:pFd will return the file descriptor representing the semaphore payload. Each call to fname:vkGetSemaphoreFdKHR must: create a new file descriptor and transfer ownership of it to the application. To avoid leaking resources, the application must: release ownership of the file descriptor when it is no longer needed. [NOTE] .Note ==== Ownership can be released in many ways. For example, the application can call code:close() on the file descriptor, or transfer ownership back to Vulkan by using the file descriptor to import a semaphore payload. ==== Where supported by the operating system, the implementation must: set the file descriptor to be closed automatically when an code:execve system call is made. Exporting a file descriptor from a semaphore may: have side effects depending on the transference of the specified handle type, as described in <>. include::../validity/protos/vkGetSemaphoreFdKHR.txt[] -- [open,refpage='VkSemaphoreGetFdInfoKHR',desc='Structure describing a POSIX FD semaphore export operation',type='structs'] -- The sname:VkSemaphoreGetFdInfoKHR structure is defined as: include::../api/structs/VkSemaphoreGetFdInfoKHR.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:semaphore is the semaphore from which state will be exported. * pname:handleType is the type of handle requested. The properties of the file descriptor returned depend on the value of pname:handleType. See elink:VkExternalSemaphoreHandleTypeFlagBits for a description of the properties of the defined external semaphore handle types. .Valid Usage **** * [[VUID-VkSemaphoreGetFdInfoKHR-handleType-01132]] pname:handleType must: have been included in slink:VkExportSemaphoreCreateInfo::pname:handleTypes when pname:semaphore's current payload was created. * [[VUID-VkSemaphoreGetFdInfoKHR-semaphore-01133]] pname:semaphore must: not currently have its payload replaced by an imported payload as described below in <> unless that imported payload's handle type was included in slink:VkExternalSemaphoreProperties::pname:exportFromImportedHandleTypes for pname:handleType. * [[VUID-VkSemaphoreGetFdInfoKHR-handleType-01134]] If pname:handleType refers to a handle type with copy payload transference semantics, as defined below in <>, there must: be no queue waiting on pname:semaphore. * [[VUID-VkSemaphoreGetFdInfoKHR-handleType-01135]] If pname:handleType refers to a handle type with copy payload transference semantics, pname:semaphore must: be signaled, or have an associated <> pending execution. * [[VUID-VkSemaphoreGetFdInfoKHR-handleType-01136]] pname:handleType must: be defined as a POSIX file descriptor handle. **** include::../validity/structs/VkSemaphoreGetFdInfoKHR.txt[] -- endif::VK_KHR_external_semaphore_fd[] [open,refpage='vkDestroySemaphore',desc='Destroy a semaphore object',type='protos'] -- To destroy a semaphore, call: include::../api/protos/vkDestroySemaphore.txt[] * pname:device is the logical device that destroys the semaphore. * pname:semaphore is the handle of the semaphore to destroy. * pname:pAllocator controls host memory allocation as described in the <> chapter. .Valid Usage **** * [[VUID-vkDestroySemaphore-semaphore-01137]] All submitted batches that refer to pname:semaphore must: have completed execution * [[VUID-vkDestroySemaphore-semaphore-01138]] If sname:VkAllocationCallbacks were provided when pname:semaphore was created, a compatible set of callbacks must: be provided here * [[VUID-vkDestroySemaphore-semaphore-01139]] If no sname:VkAllocationCallbacks were provided when pname:semaphore was created, pname:pAllocator must: be `NULL` **** include::../validity/protos/vkDestroySemaphore.txt[] -- [[synchronization-semaphores-signaling]] === Semaphore Signaling When a batch is submitted to a queue via a <>, and it includes semaphores to be signaled, it defines a memory dependency on the batch, and defines _semaphore signal operations_ which set the semaphores to the signaled state. The first <> includes every command submitted in the same batch. Semaphore signal operations that are defined by flink:vkQueueSubmit additionally include all commands that occur earlier in <>. The second <> includes only the semaphore signal operation. The first <> includes all memory access performed by the device. The second <> is empty. [[synchronization-semaphores-waiting]] === Semaphore Waiting & Unsignaling When a batch is submitted to a queue via a <>, and it includes semaphores to be waited on, it defines a memory dependency between prior semaphore signal operations and the batch, and defines _semaphore unsignal operations_ which set the semaphores to the unsignaled state. The first synchronization scope includes all semaphore signal operations that operate on semaphores waited on in the same batch, and that happen-before the wait completes. The second <> includes every command submitted in the same batch. In the case of flink:vkQueueSubmit, the second synchronization scope is limited to operations on the pipeline stages determined by the <> specified by the corresponding element of pname:pWaitDstStageMask. Also, in the case of flink:vkQueueSubmit, the second synchronization scope additionally includes all commands that occur later in <>. The first <> is empty. The second <> includes all memory access performed by the device. The semaphore unsignal operation happens-after the first set of operations in the execution dependency, and happens-before the second set of operations in the execution dependency. [NOTE] .Note ==== Unlike fences or events, the act of waiting for a semaphore also unsignals that semaphore. If two operations are separately specified to wait for the same semaphore, and there are no other execution dependencies between those operations, behaviour is undefined:. An execution dependency must: be present that guarantees that the semaphore unsignal operation for the first of those waits, happens-before the semaphore is signalled again, and before the second unsignal operation. Semaphore waits and signals should thus occur in discrete 1:1 pairs. ==== ifdef::VK_KHR_swapchain[] [NOTE] .Note ==== A common scenario for using pname:pWaitDstStageMask with values other than ename:VK_PIPELINE_STAGE_ALL_COMMANDS_BIT is when synchronizing a window system presentation operation against subsequent command buffers which render the next frame. In this case, a presentation image must: not be overwritten until the presentation operation completes, but other pipeline stages can: execute without waiting. A mask of ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT prevents subsequent color attachment writes from executing until the semaphore signals. Some implementations may: be able to execute transfer operations and/or vertex processing work before the semaphore is signaled. If an image layout transition needs to be performed on a presentable image before it is used in a framebuffer, that can: be performed as the first operation submitted to the queue after acquiring the image, and should: not prevent other work from overlapping with the presentation operation. For example, a sname:VkImageMemoryBarrier could use: * pname:srcStageMask = ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT * pname:srcAccessMask = 0 * pname:dstStageMask = ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT * pname:dstAccessMask = ename:VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | ename:VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT. * pname:oldLayout = ename:VK_IMAGE_LAYOUT_PRESENT_SRC_KHR * pname:newLayout = ename:VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL Alternatively, pname:oldLayout can: be ename:VK_IMAGE_LAYOUT_UNDEFINED, if the image's contents need not be preserved. This barrier accomplishes a dependency chain between previous presentation operations and subsequent color attachment output operations, with the layout transition performed in between, and does not introduce a dependency between previous work and any vertex processing stages. More precisely, the semaphore signals after the presentation operation completes, the semaphore wait stalls the ename:VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT stage, and there is a dependency from that same stage to itself with the layout transition performed in between. ==== endif::VK_KHR_swapchain[] [[synchronization-semaphores-waiting-state]] === Semaphore State Requirements For Wait Operations Before waiting on a semaphore, the application must: ensure the semaphore is in a valid state for a wait operation. Specifically, when a <> is submitted to a queue: * The semaphore must: be signaled, or have an associated <> that is pending execution. * There must: be no other queue waiting on the same semaphore when the operation executes. ifdef::VK_VERSION_1_1,VK_KHR_external_semaphore[] [[synchronization-semaphores-importing]] === Importing Semaphore Payloads Applications can: import a semaphore payload into an existing semaphore using an external semaphore handle. The effects of the import operation will be either temporary or permanent, as specified by the application. If the import is temporary, the implementation must: restore the semaphore to its prior permanent state after submitting the next semaphore wait operation. Performing a subsequent temporary import on a semaphore before performing a semaphore wait has no effect on this requirement; the next wait submitted on the semaphore must: still restore its last permanent state. A permanent payload import behaves as if the target semaphore was destroyed, and a new semaphore was created with the same handle but the imported payload. Because importing a semaphore payload temporarily or permanently detaches the existing payload from a semaphore, similar usage restrictions to those applied to fname:vkDestroySemaphore are applied to any command that imports a semaphore payload. Which of these import types is used is referred to as the import operation's _permanence_. Each handle type supports either one or both types of permanence. The implementation must: perform the import operation by either referencing or copying the payload referred to by the specified external semaphore handle, depending on the handle's type. The import method used is referred to as the handle type's _transference_. When using handle types with reference transference, importing a payload to a semaphore adds the semaphore to the set of all semaphores sharing that payload. This set includes the semaphore from which the payload was exported. Semaphore signaling and waiting operations performed on any semaphore in the set must: behave as if the set were a single semaphore. Importing a payload using handle types with copy transference creates a duplicate copy of the payload at the time of import, but makes no further reference to it. Semaphore signaling and waiting operations performed on the target of copy imports must: not affect any other semaphore or payload. Export operations have the same transference as the specified handle type's import operations. Additionally, exporting a semaphore payload to a handle with copy transference has the same side effects on the source semaphore's payload as executing a semaphore wait operation. If the semaphore was using a temporarily imported payload, the semaphore's prior permanent payload will be restored. ifdef::VK_KHR_external_semaphore_win32,VK_KHR_external_semaphore_fd[] [NOTE] .Note ==== The ifdef::VK_KHR_external_semaphore_win32+VK_KHR_external_semaphore_fd[tables] ifndef::VK_KHR_external_semaphore_win32+VK_KHR_external_semaphore_fd[table] ifdef::VK_KHR_external_semaphore_win32[] <> endif::VK_KHR_external_semaphore_win32[] ifdef::VK_KHR_external_semaphore_win32+VK_KHR_external_semaphore_fd[and] ifdef::VK_KHR_external_semaphore_fd[] <> endif::VK_KHR_external_semaphore_fd[] ifdef::VK_KHR_external_semaphore_win32+VK_KHR_external_semaphore_fd[define] ifndef::VK_KHR_external_semaphore_win32+VK_KHR_external_semaphore_fd[defines] the permanence and transference of each handle type. ==== endif::VK_KHR_external_semaphore_win32,VK_KHR_external_semaphore_fd[] <> allows implementations to modify an object's internal state, i.e. payload, without internal synchronization. However, for semaphores sharing a payload across processes, satisfying the external synchronization requirements of fname:VkSemaphore parameters as if all semaphores in the set were the same object is sometimes infeasible. Satisfying the <> would similarly require impractical coordination or levels of trust between processes. Therefore, these constraints only apply to a specific semaphore handle, not to its payload. For distinct semaphore objects which share a payload, if the semaphores are passed to separate queue submission commands concurrently, behavior will be as if the commands were called in an arbitrary sequential order. If the <> are violated for the shared payload by a queue submission command, or if a signal operation is queued for a shared payload that is already signaled or has a pending signal operation, effects must: be limited to one or more of the following: * Returning ename:VK_ERROR_INITIALIZATION_FAILED from the command which resulted in the violation. * Losing the logical device on which the violation occured immediately or at a future time, resulting in a ename:VK_ERROR_DEVICE_LOST error from subsequent commands, including the one causing the violation. * Continuing execution of the violating command or operation as if the semaphore wait completed successfully after an implementation-dependent timeout. In this case, the state of the payload becomes undefined:, and future operations on semaphores sharing the payload will be subject to these same rules. The semaphore must: be destroyed or have its payload replaced by an import operation to again have a well-defined state. [NOTE] .Note ==== These rules allow processes to synchronize access to shared memory without trusting each other. However, such processes must still be cautious not to use the shared semaphore for more than synchronizing access to the shared memory. For example, a process should not use a shared semaphore as part of an execution dependency chain that, when complete, leads to objects being destroyed, if it does not trust other processes sharing the semaphore payload. ==== When a semaphore is using an imported payload, its slink:VkExportSemaphoreCreateInfo::pname:handleTypes value is that specified when creating the semaphore from which the payload was exported, rather than that specified when creating the semaphore. Additionally, slink:VkExternalSemaphoreProperties::exportFromImportedHandleTypes restricts which handle types can: be exported from such a semaphore based on the specific handle type used to import the current payload. ifdef::VK_KHR_swapchain[] Passing a semaphore to flink:vkAcquireNextImageKHR is equivalent to temporarily importing a semaphore payload to that semaphore. [NOTE] .Note ==== Because the exportable handle types of an imported semaphore correspond to its current imported payload, and flink:vkAcquireNextImageKHR behaves the same as a temporary import operation for which the source semaphore is opaque to the application, applications have no way of determining whether any external handle types can: be exported from a semaphore in this state. Therefore, applications must: not attempt to export external handles from semaphores using a temporarily imported payload from flink:vkAcquireNextImageKHR. ==== endif::VK_KHR_swapchain[] When importing a semaphore payload, it is the responsibility of the application to ensure the external handles meet all valid usage requirements. However, implementations must: perform sufficient validation of external handles to ensure that the operation results in a valid semaphore which will not cause program termination, device loss, queue stalls, or corruption of other resources when used as allowed according to its import parameters, and excepting those side effects allowed for violations of the <> rules. If the external handle provided does not meet these requirements, the implementation must: fail the semaphore payload import operation with the error code ename:VK_ERROR_INVALID_EXTERNAL_HANDLE. endif::VK_VERSION_1_1,VK_KHR_external_semaphore[] ifdef::VK_KHR_external_semaphore_win32[] [open,refpage='vkImportSemaphoreWin32HandleKHR',desc='Import a semaphore from a Windows HANDLE',type='protos'] -- To import a semaphore payload from a Windows handle, call: include::../api/protos/vkImportSemaphoreWin32HandleKHR.txt[] * pname:device is the logical device that created the semaphore. * pname:pImportSemaphoreWin32HandleInfo points to a slink:VkImportSemaphoreWin32HandleInfoKHR structure specifying the semaphore and import parameters. Importing a semaphore payload from Windows handles does not transfer ownership of the handle to the Vulkan implementation. For handle types defined as NT handles, the application must: release ownership using the fname:CloseHandle system call when the handle is no longer needed. Applications can: import the same semaphore payload into multiple instances of Vulkan, into the same instance from which it was exported, and multiple times into a given Vulkan instance. include::../validity/protos/vkImportSemaphoreWin32HandleKHR.txt[] -- [open,refpage='VkImportSemaphoreWin32HandleInfoKHR',desc='Structure specifying Windows handle to import to a semaphore',type='structs'] -- The sname:VkImportSemaphoreWin32HandleInfoKHR structure is defined as: include::../api/structs/VkImportSemaphoreWin32HandleInfoKHR.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:semaphore is the semaphore into which the payload will be imported. * pname:flags is a bitmask of elink:VkSemaphoreImportFlagBits specifying additional parameters for the semaphore payload import operation. * pname:handleType specifies the type of pname:handle. * pname:handle is the external handle to import, or `NULL`. * pname:name is a NULL-terminated UTF-16 string naming the underlying synchronization primitive to import, or `NULL`. The handle types supported by pname:handleType are: [[synchronization-semaphore-handletypes-win32]] .Handle Types Supported by VkImportSemaphoreWin32HandleInfoKHR [width="80%",options="header"] |==== | Handle Type | Transference | Permanence Supported | ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32_BIT | Reference | Temporary,Permanent | ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32_KMT_BIT | Reference | Temporary,Permanent | ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_D3D12_FENCE_BIT | Reference | Temporary,Permanent |==== .Valid Usage **** * [[VUID-VkImportSemaphoreWin32HandleInfoKHR-handleType-01140]] pname:handleType must: be a value included in the <> table. * [[VUID-VkImportSemaphoreWin32HandleInfoKHR-handleType-01466]] If pname:handleType is not ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32_BIT or ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_D3D12_FENCE_BIT, pname:name must: be `NULL`. * [[VUID-VkImportSemaphoreWin32HandleInfoKHR-handleType-01467]] If pname:handleType is not `0` and pname:handle is `NULL`, pname:name must: name a valid synchronization primitive of the type specified by pname:handleType. * [[VUID-VkImportSemaphoreWin32HandleInfoKHR-handleType-01468]] If pname:handleType is not `0` and pname:name is `NULL`, pname:handle must: be a valid handle of the type specified by pname:handleType. * [[VUID-VkImportSemaphoreWin32HandleInfoKHR-handle-01469]] If pname:handle is not `NULL`, pname:name must be `NULL`. * [[VUID-VkImportSemaphoreWin32HandleInfoKHR-handle-01542]] If pname:handle is not `NULL`, it must: obey any requirements listed for pname:handleType in <>. * [[VUID-VkImportSemaphoreWin32HandleInfoKHR-name-01543]] If pname:name is not `NULL`, it must: obey any requirements listed for pname:handleType in <>. **** include::../validity/structs/VkImportSemaphoreWin32HandleInfoKHR.txt[] -- endif::VK_KHR_external_semaphore_win32[] ifdef::VK_KHR_external_semaphore_fd[] [open,refpage='vkImportSemaphoreFdKHR',desc='Import a semaphore from a POSIX file descriptor',type='protos'] -- To import a semaphore payload from a POSIX file descriptor, call: include::../api/protos/vkImportSemaphoreFdKHR.txt[] * pname:device is the logical device that created the semaphore. * pname:pImportSemaphoreFdInfo points to a slink:VkImportSemaphoreFdInfoKHR structure specifying the semaphore and import parameters. Importing a semaphore payload from a file descriptor transfers ownership of the file descriptor from the application to the Vulkan implementation. The application must: not perform any operations on the file descriptor after a successful import. Applications can: import the same semaphore payload into multiple instances of Vulkan, into the same instance from which it was exported, and multiple times into a given Vulkan instance. .Valid Usage **** * [[VUID-vkImportSemaphoreFdKHR-semaphore-01142]] pname:semaphore must: not be associated with any queue command that has not yet completed execution on that queue **** include::../validity/protos/vkImportSemaphoreFdKHR.txt[] -- [open,refpage='VkImportSemaphoreFdInfoKHR',desc='Structure specifying POSIX file descriptor to import to a semaphore',type='structs'] -- The sname:VkImportSemaphoreFdInfoKHR structure is defined as: include::../api/structs/VkImportSemaphoreFdInfoKHR.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:semaphore is the semaphore into which the payload will be imported. * pname:flags is a bitmask of elink:VkSemaphoreImportFlagBits specifying additional parameters for the semaphore payload import operation. * pname:handleType specifies the type of pname:fd. * pname:fd is the external handle to import. The handle types supported by pname:handleType are: [[synchronization-semaphore-handletypes-fd]] .Handle Types Supported by VkImportSemaphoreFdInfoKHR [width="80%",options="header"] |==== | Handle Type | Transference | Permanence Supported | ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_FD_BIT | Reference | Temporary,Permanent | ename:VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_SYNC_FD_BIT | Copy | Temporary |==== .Valid Usage **** * [[VUID-VkImportSemaphoreFdInfoKHR-handleType-01143]] pname:handleType must: be a value included in the <> table. * [[VUID-VkImportSemaphoreFdInfoKHR-fd-01544]] pname:fd must: obey any requirements listed for pname:handleType in <>. **** include::../validity/structs/VkImportSemaphoreFdInfoKHR.txt[] -- endif::VK_KHR_external_semaphore_fd[] ifdef::VK_VERSION_1_1,VK_KHR_external_semaphore[] ifdef::VK_KHR_external_semaphore_win32,VK_KHR_external_semaphore_fd[] [open,refpage='VkSemaphoreImportFlagBits',desc='Bitmask specifying additional parameters of semaphore payload import',type='enums'] -- Additional parameters of a semaphore import operation are specified by ifdef::VK_KHR_external_semaphore_win32[] slink:VkImportSemaphoreWin32HandleInfoKHR::pname:flags endif::VK_KHR_external_semaphore_win32[] ifdef::VK_KHR_external_semaphore_win32+VK_KHR_external_semaphore_fd[or] ifdef::VK_KHR_external_semaphore_fd[] slink:VkImportSemaphoreFdInfoKHR::pname:flags endif::VK_KHR_external_semaphore_fd[] . Bits which can be set include: include::../api/enums/VkSemaphoreImportFlagBits.txt[] ifdef::VK_KHR_external_semaphore[] or the equivalent include::../api/enums/VkSemaphoreImportFlagBitsKHR.txt[] endif::VK_KHR_external_semaphore[] These bits have the following meanings: * ename:VK_SEMAPHORE_IMPORT_TEMPORARY_BIT specifies that the semaphore payload will be imported only temporarily, as described in <>, regardless of the permanence of pname:handleType. -- [open,refpage='VkSemaphoreImportFlags',desc='Bitmask of VkSemaphoreImportFlagBits',type='enums'] -- include::../api/flags/VkSemaphoreImportFlags.txt[] ifdef::VK_KHR_external_semaphore[] or the equivalent include::../api/flags/VkSemaphoreImportFlagsKHR.txt[] endif::VK_KHR_external_semaphore[] sname:VkSemaphoreImportFlags is a bitmask type for setting a mask of zero or more slink:VkSemaphoreImportFlagBits. -- endif::VK_KHR_external_semaphore_win32,VK_KHR_external_semaphore_fd[] endif::VK_VERSION_1_1,VK_KHR_external_semaphore[] [[synchronization-events]] == Events [open,refpage='VkEvent',desc='Opaque handle to an event object',type='handles'] -- Events are a synchronization primitive that can: be used to insert a fine-grained dependency between commands submitted to the same queue, or between the host and a queue. Events must: not be used to insert a dependency between commands submitted to different queues. Events have two states - signaled and unsignaled. An application can: signal an event, or unsignal it, on either the host or the device. A device can: wait for an event to become signaled before executing further operations. No command exists to wait for an event to become signaled on the host, but the current state of an event can: be queried. Events are represented by sname:VkEvent handles: include::../api/handles/VkEvent.txt[] -- [open,refpage='vkCreateEvent',desc='Create a new event object',type='protos'] -- To create an event, call: include::../api/protos/vkCreateEvent.txt[] * pname:device is the logical device that creates the event. * pname:pCreateInfo is a pointer to an instance of the sname:VkEventCreateInfo structure which contains information about how the event is to be created. * pname:pAllocator controls host memory allocation as described in the <> chapter. * pname:pEvent points to a handle in which the resulting event object is returned. When created, the event object is in the unsignaled state. include::../validity/protos/vkCreateEvent.txt[] -- [open,refpage='VkEventCreateInfo',desc='Structure specifying parameters of a newly created event',type='structs'] -- The sname:VkEventCreateInfo structure is defined as: include::../api/structs/VkEventCreateInfo.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:flags is reserved for future use. include::../validity/structs/VkEventCreateInfo.txt[] -- [open,refpage='VkEventCreateFlags',desc='Reserved for future use',type='enums'] -- include::../api/flags/VkEventCreateFlags.txt[] sname:VkEventCreateFlags is a bitmask type for setting a mask, but is currently reserved for future use. -- [open,refpage='vkDestroyEvent',desc='Destroy an event object',type='protos'] -- To destroy an event, call: include::../api/protos/vkDestroyEvent.txt[] * pname:device is the logical device that destroys the event. * pname:event is the handle of the event to destroy. * pname:pAllocator controls host memory allocation as described in the <> chapter. .Valid Usage **** * [[VUID-vkDestroyEvent-event-01145]] All submitted commands that refer to pname:event must: have completed execution * [[VUID-vkDestroyEvent-event-01146]] If sname:VkAllocationCallbacks were provided when pname:event was created, a compatible set of callbacks must: be provided here * [[VUID-vkDestroyEvent-event-01147]] If no sname:VkAllocationCallbacks were provided when pname:event was created, pname:pAllocator must: be `NULL` **** include::../validity/protos/vkDestroyEvent.txt[] -- [open,refpage='vkGetEventStatus',desc='Retrieve the status of an event object',type='protos'] -- To query the state of an event from the host, call: include::../api/protos/vkGetEventStatus.txt[] * pname:device is the logical device that owns the event. * pname:event is the handle of the event to query. Upon success, fname:vkGetEventStatus returns the state of the event object with the following return codes: .Event Object Status Codes [width="80%",options="header"] |==== | Status | Meaning | ename:VK_EVENT_SET | The event specified by pname:event is signaled. | ename:VK_EVENT_RESET | The event specified by pname:event is unsignaled. |==== If a fname:vkCmdSetEvent or fname:vkCmdResetEvent command is in a command buffer that is in the <>, then the value returned by this command may: immediately be out of date. The state of an event can: be updated by the host. The state of the event is immediately changed, and subsequent calls to fname:vkGetEventStatus will return the new state. If an event is already in the requested state, then updating it to the same state has no effect. include::../validity/protos/vkGetEventStatus.txt[] -- [[synchronization-events-signaling-host]] [open,refpage='vkSetEvent',desc='Set an event to signaled state',type='protos'] -- To set the state of an event to signaled from the host, call: include::../api/protos/vkSetEvent.txt[] * pname:device is the logical device that owns the event. * pname:event is the event to set. When flink:vkSetEvent is executed on the host, it defines an _event signal operation_ which sets the event to the signaled state. If pname:event is already in the signaled state when flink:vkSetEvent is executed, then flink:vkSetEvent has no effect, and no event signal operation occurs. include::../validity/protos/vkSetEvent.txt[] -- [[synchronization-events-unsignaling-host]] [open,refpage='vkResetEvent',desc='Reset an event to non-signaled state',type='protos'] -- To set the state of an event to unsignaled from the host, call: include::../api/protos/vkResetEvent.txt[] * pname:device is the logical device that owns the event. * pname:event is the event to reset. When flink:vkResetEvent is executed on the host, it defines an _event unsignal operation_ which resets the event to the unsignaled state. If pname:event is already in the unsignaled state when flink:vkResetEvent is executed, then flink:vkResetEvent has no effect, and no event unsignal operation occurs. .Valid Usage **** * [[VUID-vkResetEvent-event-01148]] pname:event must: not be waited on by a fname:vkCmdWaitEvents command that is currently executing **** include::../validity/protos/vkResetEvent.txt[] -- The state of an event can: also be updated on the device by commands inserted in command buffers. [[synchronization-events-signaling-device]] [open,refpage='vkCmdSetEvent',desc='Set an event object to signaled state',type='protos'] -- To set the state of an event to signaled from a device, call: include::../api/protos/vkCmdSetEvent.txt[] * pname:commandBuffer is the command buffer into which the command is recorded. * pname:event is the event that will be signaled. * pname:stageMask specifies the <> used to determine when the pname:event is signaled. When flink:vkCmdSetEvent is submitted to a queue, it defines an execution dependency on commands that were submitted before it, and defines an event signal operation which sets the event to the signaled state. The first <> includes all commands that occur earlier in <>. The synchronization scope is limited to operations on the pipeline stages determined by the <> specified by pname:stageMask. The second <> includes only the event signal operation. If pname:event is already in the signaled state when flink:vkCmdSetEvent is executed on the device, then flink:vkCmdSetEvent has no effect, no event signal operation occurs, and no execution dependency is generated. .Valid Usage **** * [[VUID-vkCmdSetEvent-stageMask-01149]] pname:stageMask must: not include ename:VK_PIPELINE_STAGE_HOST_BIT * [[VUID-vkCmdSetEvent-stageMask-01150]] If the <> feature is not enabled, pname:stageMask must: not contain ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT * [[VUID-vkCmdSetEvent-stageMask-01151]] If the <> feature is not enabled, pname:stageMask must: not contain ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT or ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT ifdef::VK_VERSION_1_1,VK_KHR_device_group[] * [[VUID-vkCmdSetEvent-commandBuffer-01152]] pname:commandBuffer's current device mask must: include exactly one physical device. endif::VK_VERSION_1_1,VK_KHR_device_group[] ifdef::VK_NV_mesh_shader[] * [[VUID-vkCmdSetEvent-stageMask-02107]] If the <> feature is not enabled, pname:stageMask must: not contain ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT_NV * [[VUID-vkCmdSetEvent-stageMask-02108]] If the <> feature is not enabled, pname:stageMask must: not contain ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT_NV endif::VK_NV_mesh_shader[] **** include::../validity/protos/vkCmdSetEvent.txt[] -- [[synchronization-events-unsignaling-device]] [open,refpage='vkCmdResetEvent',desc='Reset an event object to non-signaled state',type='protos'] -- To set the state of an event to unsignaled from a device, call: include::../api/protos/vkCmdResetEvent.txt[] * pname:commandBuffer is the command buffer into which the command is recorded. * pname:event is the event that will be unsignaled. * pname:stageMask is a bitmask of elink:VkPipelineStageFlagBits specifying the <> used to determine when the pname:event is unsignaled. When flink:vkCmdResetEvent is submitted to a queue, it defines an execution dependency on commands that were submitted before it, and defines an event unsignal operation which resets the event to the unsignaled state. The first <> includes all commands that occur earlier in <>. The synchronization scope is limited to operations on the pipeline stages determined by the <> specified by pname:stageMask. The second <> includes only the event unsignal operation. If pname:event is already in the unsignaled state when flink:vkCmdResetEvent is executed on the device, then flink:vkCmdResetEvent has no effect, no event unsignal operation occurs, and no execution dependency is generated. .Valid Usage **** * [[VUID-vkCmdResetEvent-stageMask-01153]] pname:stageMask must: not include ename:VK_PIPELINE_STAGE_HOST_BIT * [[VUID-vkCmdResetEvent-stageMask-01154]] If the <> feature is not enabled, pname:stageMask must: not contain ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT * [[VUID-vkCmdResetEvent-stageMask-01155]] If the <> feature is not enabled, pname:stageMask must: not contain ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT or ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT * [[VUID-vkCmdResetEvent-event-01156]] When this command executes, pname:event must: not be waited on by a fname:vkCmdWaitEvents command that is currently executing ifdef::VK_VERSION_1_1,VK_KHR_device_group[] * [[VUID-vkCmdResetEvent-commandBuffer-01157]] pname:commandBuffer's current device mask must: include exactly one physical device. endif::VK_VERSION_1_1,VK_KHR_device_group[] ifdef::VK_NV_mesh_shader[] * [[VUID-vkCmdResetEvent-stageMask-02109]] If the <> feature is not enabled, pname:stageMask must: not contain ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT_NV * [[VUID-vkCmdResetEvent-stageMask-02110]] If the <> feature is not enabled, pname:stageMask must: not contain ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT_NV endif::VK_NV_mesh_shader[] **** include::../validity/protos/vkCmdResetEvent.txt[] -- [open,refpage='vkCmdWaitEvents',desc='Wait for one or more events and insert a set of memory',type='protos'] -- To wait for one or more events to enter the signaled state on a device, call: [[synchronization-events-waiting-device]] include::../api/protos/vkCmdWaitEvents.txt[] * pname:commandBuffer is the command buffer into which the command is recorded. * pname:eventCount is the length of the pname:pEvents array. * pname:pEvents is an array of event object handles to wait on. * pname:srcStageMask is a bitmask of elink:VkPipelineStageFlagBits specifying the <>. * pname:dstStageMask is a bitmask of elink:VkPipelineStageFlagBits specifying the <>. * pname:memoryBarrierCount is the length of the pname:pMemoryBarriers array. * pname:pMemoryBarriers is a pointer to an array of slink:VkMemoryBarrier structures. * pname:bufferMemoryBarrierCount is the length of the pname:pBufferMemoryBarriers array. * pname:pBufferMemoryBarriers is a pointer to an array of slink:VkBufferMemoryBarrier structures. * pname:imageMemoryBarrierCount is the length of the pname:pImageMemoryBarriers array. * pname:pImageMemoryBarriers is a pointer to an array of slink:VkImageMemoryBarrier structures. When fname:vkCmdWaitEvents is submitted to a queue, it defines a memory dependency between prior event signal operations on the same queue or the host, and subsequent commands. fname:vkCmdWaitEvents must: not be used to wait on event signal operations occuring on other queues. The first synchronization scope only includes event signal operations that operate on members of pname:pEvents, and the operations that happened-before the event signal operations. Event signal operations performed by flink:vkCmdSetEvent that occur earlier in <> are included in the first synchronization scope, if the <> pipeline stage in their pname:stageMask parameter is <> than or equal to the <> pipeline stage in pname:srcStageMask. Event signal operations performed by flink:vkSetEvent are only included in the first synchronization scope if ename:VK_PIPELINE_STAGE_HOST_BIT is included in pname:srcStageMask. The second <> includes all commands that occur later in <>. The second synchronization scope is limited to operations on the pipeline stages determined by the <> specified by pname:dstStageMask. The first <> is limited to access in the pipeline stages determined by the <> specified by pname:srcStageMask. Within that, the first access scope only includes the first access scopes defined by elements of the pname:pMemoryBarriers, pname:pBufferMemoryBarriers and pname:pImageMemoryBarriers arrays, which each define a set of <>. If no memory barriers are specified, then the first access scope includes no accesses. The second <> is limited to access in the pipeline stages determined by the <> specified by pname:dstStageMask. Within that, the second access scope only includes the second access scopes defined by elements of the pname:pMemoryBarriers, pname:pBufferMemoryBarriers and pname:pImageMemoryBarriers arrays, which each define a set of <>. If no memory barriers are specified, then the second access scope includes no accesses. [NOTE] .Note ==== flink:vkCmdWaitEvents is used with flink:vkCmdSetEvent to define a memory dependency between two sets of action commands, roughly in the same way as pipeline barriers, but split into two commands such that work between the two may: execute unhindered. ==== [NOTE] .Note ==== Applications should: be careful to avoid race conditions when using events. There is no direct ordering guarantee between a flink:vkCmdResetEvent command and a flink:vkCmdWaitEvents command submitted after it, so some other execution dependency must: be included between these commands (e.g. a semaphore). ==== .Valid Usage **** * [[VUID-vkCmdWaitEvents-srcStageMask-01158]] pname:srcStageMask must: be the bitwise OR of the pname:stageMask parameter used in previous calls to fname:vkCmdSetEvent with any of the members of pname:pEvents and ename:VK_PIPELINE_STAGE_HOST_BIT if any of the members of pname:pEvents was set using fname:vkSetEvent * [[VUID-vkCmdWaitEvents-srcStageMask-01159]] If the <> feature is not enabled, pname:srcStageMask must: not contain ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT * [[VUID-vkCmdWaitEvents-dstStageMask-01160]] If the <> feature is not enabled, pname:dstStageMask must: not contain ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT * [[VUID-vkCmdWaitEvents-srcStageMask-01161]] If the <> feature is not enabled, pname:srcStageMask must: not contain ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT or ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT * [[VUID-vkCmdWaitEvents-dstStageMask-01162]] If the <> feature is not enabled, pname:dstStageMask must: not contain ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT or ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT * [[VUID-vkCmdWaitEvents-pEvents-01163]] If pname:pEvents includes one or more events that will be signaled by fname:vkSetEvent after pname:commandBuffer has been submitted to a queue, then fname:vkCmdWaitEvents must: not be called inside a render pass instance * [[VUID-vkCmdWaitEvents-srcStageMask-01164]] Any pipeline stage included in pname:srcStageMask or pname:dstStageMask must: be supported by the capabilities of the queue family specified by the pname:queueFamilyIndex member of the slink:VkCommandPoolCreateInfo structure that was used to create the sname:VkCommandPool that pname:commandBuffer was allocated from, as specified in the <>. * [[VUID-vkCmdWaitEvents-pMemoryBarriers-01165]] Each element of pname:pMemoryBarriers, pname:pBufferMemoryBarriers or pname:pImageMemoryBarriers must: not have any access flag included in its pname:srcAccessMask member if that bit is not supported by any of the pipeline stages in pname:srcStageMask, as specified in the <>. * [[VUID-vkCmdWaitEvents-pMemoryBarriers-01166]] Each element of pname:pMemoryBarriers, pname:pBufferMemoryBarriers or pname:pImageMemoryBarriers must: not have any access flag included in its pname:dstAccessMask member if that bit is not supported by any of the pipeline stages in pname:dstStageMask, as specified in the <>. ifdef::VK_VERSION_1_1,VK_KHR_device_group[] * [[VUID-vkCmdWaitEvents-commandBuffer-01167]] pname:commandBuffer's current device mask must: include exactly one physical device. endif::VK_VERSION_1_1,VK_KHR_device_group[] ifdef::VK_NV_mesh_shader[] * [[VUID-vkCmdWaitEvents-srcStageMask-02111]] If the <> feature is not enabled, pname:srcStageMask must: not contain ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT_NV * [[VUID-vkCmdWaitEvents-srcStageMask-02112]] If the <> feature is not enabled, pname:srcStageMask must: not contain ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT_NV * [[VUID-vkCmdWaitEvents-dstStageMask-02113]] If the <> feature is not enabled, pname:dstStageMask must: not contain ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT_NV * [[VUID-vkCmdWaitEvents-dstStageMask-02114]] If the <> feature is not enabled, pname:dstStageMask must: not contain ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT_NV endif::VK_NV_mesh_shader[] **** include::../validity/protos/vkCmdWaitEvents.txt[] -- [[synchronization-pipeline-barriers]] == Pipeline Barriers flink:vkCmdPipelineBarrier is a synchronization command that inserts a dependency between commands submitted to the same queue, or between commands in the same subpass. [open,refpage='vkCmdPipelineBarrier',desc='Insert a memory dependency',type='protos'] -- To record a pipeline barrier, call: include::../api/protos/vkCmdPipelineBarrier.txt[] * pname:commandBuffer is the command buffer into which the command is recorded. * pname:srcStageMask is a bitmask of elink:VkPipelineStageFlagBits specifying the <>. * pname:dstStageMask is a bitmask of elink:VkPipelineStageFlagBits specifying the <>. * pname:dependencyFlags is a bitmask of elink:VkDependencyFlagBits specifying how execution and memory dependencies are formed. * pname:memoryBarrierCount is the length of the pname:pMemoryBarriers array. * pname:pMemoryBarriers is a pointer to an array of slink:VkMemoryBarrier structures. * pname:bufferMemoryBarrierCount is the length of the pname:pBufferMemoryBarriers array. * pname:pBufferMemoryBarriers is a pointer to an array of slink:VkBufferMemoryBarrier structures. * pname:imageMemoryBarrierCount is the length of the pname:pImageMemoryBarriers array. * pname:pImageMemoryBarriers is a pointer to an array of slink:VkImageMemoryBarrier structures. When flink:vkCmdPipelineBarrier is submitted to a queue, it defines a memory dependency between commands that were submitted before it, and those submitted after it. If flink:vkCmdPipelineBarrier was recorded outside a render pass instance, the first <> includes all commands that occur earlier in <>. If flink:vkCmdPipelineBarrier was recorded inside a render pass instance, the first synchronization scope includes only commands that occur earlier in <> within the same subpass. In either case, the first synchronization scope is limited to operations on the pipeline stages determined by the <> specified by pname:srcStageMask. If flink:vkCmdPipelineBarrier was recorded outside a render pass instance, the second <> includes all commands that occur later in <>. If flink:vkCmdPipelineBarrier was recorded inside a render pass instance, the second synchronization scope includes only commands that occur later in <> within the same subpass. In either case, the second synchronization scope is limited to operations on the pipeline stages determined by the <> specified by pname:dstStageMask. The first <> is limited to access in the pipeline stages determined by the <> specified by pname:srcStageMask. Within that, the first access scope only includes the first access scopes defined by elements of the pname:pMemoryBarriers, pname:pBufferMemoryBarriers and pname:pImageMemoryBarriers arrays, which each define a set of <>. If no memory barriers are specified, then the first access scope includes no accesses. The second <> is limited to access in the pipeline stages determined by the <> specified by pname:dstStageMask. Within that, the second access scope only includes the second access scopes defined by elements of the pname:pMemoryBarriers, pname:pBufferMemoryBarriers and pname:pImageMemoryBarriers arrays, which each define a set of <>. If no memory barriers are specified, then the second access scope includes no accesses. If pname:dependencyFlags includes ename:VK_DEPENDENCY_BY_REGION_BIT, then any dependency between <> pipeline stages is <> - otherwise it is <>. .Valid Usage **** * [[VUID-vkCmdPipelineBarrier-srcStageMask-01168]] If the <> feature is not enabled, pname:srcStageMask must: not contain ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT * [[VUID-vkCmdPipelineBarrier-dstStageMask-01169]] If the <> feature is not enabled, pname:dstStageMask must: not contain ename:VK_PIPELINE_STAGE_GEOMETRY_SHADER_BIT * [[VUID-vkCmdPipelineBarrier-srcStageMask-01170]] If the <> feature is not enabled, pname:srcStageMask must: not contain ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT or ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT * [[VUID-vkCmdPipelineBarrier-dstStageMask-01171]] If the <> feature is not enabled, pname:dstStageMask must: not contain ename:VK_PIPELINE_STAGE_TESSELLATION_CONTROL_SHADER_BIT or ename:VK_PIPELINE_STAGE_TESSELLATION_EVALUATION_SHADER_BIT * [[VUID-vkCmdPipelineBarrier-pDependencies-02285]] If fname:vkCmdPipelineBarrier is called within a render pass instance, the render pass must: have been created with at least one sname:VkSubpassDependency instance in sname:VkRenderPassCreateInfo::pname:pDependencies that expresses a dependency from the current subpass to itself, and for which pname:srcStageMask contains a subset of the bit values in sname:VkSubpassDependency::pname:srcStageMask, pname:dstStageMask contains a subset of the bit values in sname:VkSubpassDependency::pname:dstStageMask, pname:dependencyFlags is equal to sname:VkSubpassDependency::pname:dependencyFlags, pname:srcAccessMask member of each element of pname:pMemoryBarriers and pname:pImageMemoryBarriers contains a subset of the bit values in sname:VkSubpassDependency::pname:srcAccessMask, and pname:dstAccessMask member of each element of pname:pMemoryBarriers and pname:pImageMemoryBarriers contains a subset of the bit values in sname:VkSubpassDependency::pname:dstAccessMask * [[VUID-vkCmdPipelineBarrier-bufferMemoryBarrierCount-01178]] If fname:vkCmdPipelineBarrier is called within a render pass instance, pname:bufferMemoryBarrierCount must: be `0` * [[VUID-vkCmdPipelineBarrier-image-01179]] If fname:vkCmdPipelineBarrier is called within a render pass instance, the pname:image member of any element of pname:pImageMemoryBarriers must: be equal to one of the elements of pname:pAttachments that the current pname:framebuffer was created with, that is also referred to by one of the elements of the pname:pColorAttachments, pname:pResolveAttachments or pname:pDepthStencilAttachment members of the sname:VkSubpassDescription instance that the current subpass was created with * [[VUID-vkCmdPipelineBarrier-oldLayout-01180]] If fname:vkCmdPipelineBarrier is called within a render pass instance, the pname:oldLayout and pname:newLayout members of any element of pname:pImageMemoryBarriers must: be equal to the pname:layout member of an element of the pname:pColorAttachments, pname:pResolveAttachments or pname:pDepthStencilAttachment members of the sname:VkSubpassDescription instance that the current subpass was created with, that refers to the same pname:image * [[VUID-vkCmdPipelineBarrier-oldLayout-01181]] If fname:vkCmdPipelineBarrier is called within a render pass instance, the pname:oldLayout and pname:newLayout members of an element of pname:pImageMemoryBarriers must: be equal * [[VUID-vkCmdPipelineBarrier-srcQueueFamilyIndex-01182]] If fname:vkCmdPipelineBarrier is called within a render pass instance, the pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex members of any element of pname:pImageMemoryBarriers must: be ename:VK_QUEUE_FAMILY_IGNORED * [[VUID-vkCmdPipelineBarrier-srcStageMask-01183]] Any pipeline stage included in pname:srcStageMask or pname:dstStageMask must: be supported by the capabilities of the queue family specified by the pname:queueFamilyIndex member of the slink:VkCommandPoolCreateInfo structure that was used to create the sname:VkCommandPool that pname:commandBuffer was allocated from, as specified in the <>. * [[VUID-vkCmdPipelineBarrier-pMemoryBarriers-01184]] Each element of pname:pMemoryBarriers, pname:pBufferMemoryBarriers and pname:pImageMemoryBarriers must: not have any access flag included in its pname:srcAccessMask member if that bit is not supported by any of the pipeline stages in pname:srcStageMask, as specified in the <>. * [[VUID-vkCmdPipelineBarrier-pMemoryBarriers-01185]] Each element of pname:pMemoryBarriers, pname:pBufferMemoryBarriers and pname:pImageMemoryBarriers must: not have any access flag included in its pname:dstAccessMask member if that bit is not supported by any of the pipeline stages in pname:dstStageMask, as specified in the <>. ifdef::VK_VERSION_1_1,VK_KHR_multiview[] * [[VUID-vkCmdPipelineBarrier-dependencyFlags-01186]] If fname:vkCmdPipelineBarrier is called outside of a render pass instance, pname:dependencyFlags must: not include ename:VK_DEPENDENCY_VIEW_LOCAL_BIT endif::VK_VERSION_1_1,VK_KHR_multiview[] ifdef::VK_NV_mesh_shader[] * [[VUID-vkCmdPipelineBarrier-srcStageMask-02115]] If the <> feature is not enabled, pname:srcStageMask must: not contain ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT_NV * [[VUID-vkCmdPipelineBarrier-srcStageMask-02116]] If the <> feature is not enabled, pname:srcStageMask must: not contain ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT_NV * [[VUID-vkCmdPipelineBarrier-dstStageMask-02117]] If the <> feature is not enabled, pname:dstStageMask must: not contain ename:VK_PIPELINE_STAGE_MESH_SHADER_BIT_NV * [[VUID-vkCmdPipelineBarrier-dstStageMask-02118]] If the <> feature is not enabled, pname:dstStageMask must: not contain ename:VK_PIPELINE_STAGE_TASK_SHADER_BIT_NV endif::VK_NV_mesh_shader[] **** include::../validity/protos/vkCmdPipelineBarrier.txt[] -- [open,refpage='VkDependencyFlagBits',desc='Bitmask specifying how execution and memory dependencies are formed',type='enums'] -- Bits which can: be set in fname:vkCmdPipelineBarrier::pname:dependencyFlags, specifying how execution and memory dependencies are formed, are: include::../api/enums/VkDependencyFlagBits.txt[] * ename:VK_DEPENDENCY_BY_REGION_BIT specifies that dependencies will be <>. ifdef::VK_VERSION_1_1,VK_KHR_multiview[] * ename:VK_DEPENDENCY_VIEW_LOCAL_BIT specifies that a <>. endif::VK_VERSION_1_1,VK_KHR_multiview[] ifdef::VK_VERSION_1_1,VK_KHR_device_group[] * ename:VK_DEPENDENCY_DEVICE_GROUP_BIT specifies that dependencies are <>. endif::VK_VERSION_1_1,VK_KHR_device_group[] -- [open,refpage='VkDependencyFlags',desc='Bitmask of VkDependencyFlagBits',type='enums'] -- include::../api/flags/VkDependencyFlags.txt[] sname:VkDependencyFlags is a bitmask type for setting a mask of zero or more slink:VkDependencyFlagBits. -- [[synchronization-pipeline-barriers-subpass-self-dependencies]] === Subpass Self-dependency If fname:vkCmdPipelineBarrier is called inside a render pass instance, the following restrictions apply. For a given subpass to allow a pipeline barrier, the render pass must: declare a _self-dependency_ from that subpass to itself. That is, there must: exist a sname:VkSubpassDependency in the subpass dependency list for the render pass with pname:srcSubpass and pname:dstSubpass equal to that subpass index. More than one self-dependency can: be declared for each subpass. Self-dependencies must: only include pipeline stage bits that are graphics stages. Self-dependencies must: not have any earlier pipeline stages depend on any later pipeline stages (according to the order of <>), unless all of the stages are <>. If the source and destination stage masks both include framebuffer-space stages, then pname:dependencyFlags must: include ename:VK_DEPENDENCY_BY_REGION_BIT. ifdef::VK_VERSION_1_1,VK_KHR_multiview[] If the subpass has more than one view, then pname:dependencyFlags must: include ename:VK_DEPENDENCY_VIEW_LOCAL_BIT. endif::VK_VERSION_1_1,VK_KHR_multiview[] A fname:vkCmdPipelineBarrier command inside a render pass instance must: be a _subset_ of one of the self-dependencies of the subpass it is used in, meaning that the stage masks and access masks must: each include only a subset of the bits of the corresponding mask in that self-dependency. If the self-dependency has ename:VK_DEPENDENCY_BY_REGION_BIT ifdef::VK_VERSION_1_1,VK_KHR_multiview[] or ename:VK_DEPENDENCY_VIEW_LOCAL_BIT endif::VK_VERSION_1_1,VK_KHR_multiview[] set, then so must: the pipeline barrier. Pipeline barriers within a render pass instance can: only be types sname:VkMemoryBarrier or sname:VkImageMemoryBarrier. If a sname:VkImageMemoryBarrier is used, the image and image subresource range specified in the barrier must: be a subset of one of the image views used by the framebuffer in the current subpass. Additionally, pname:oldLayout must: be equal to pname:newLayout, and both the pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex must: be ename:VK_QUEUE_FAMILY_IGNORED. [[synchronization-memory-barriers]] == Memory Barriers _Memory barriers_ are used to explicitly control access to buffer and image subresource ranges. Memory barriers are used to <>, <>, and define <>. They explicitly define the <> and buffer and image subresource ranges that are included in the <> of a memory dependency that is created by a synchronization command that includes them. [[synchronization-global-memory-barriers]] === Global Memory Barriers Global memory barriers apply to memory accesses involving all memory objects that exist at the time of its execution. [open,refpage='VkMemoryBarrier',desc='Structure specifying a global memory barrier',type='structs'] -- The sname:VkMemoryBarrier structure is defined as: include::../api/structs/VkMemoryBarrier.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:srcAccessMask is a bitmask of elink:VkAccessFlagBits specifying a <>. * pname:dstAccessMask is a bitmask of elink:VkAccessFlagBits specifying a <>. The first <> is limited to access types in the <> specified by pname:srcAccessMask. The second <> is limited to access types in the <> specified by pname:dstAccessMask. include::../validity/structs/VkMemoryBarrier.txt[] -- [[synchronization-buffer-memory-barriers]] === Buffer Memory Barriers Buffer memory barriers only apply to memory accesses involving a specific buffer range. That is, a memory dependency formed from a buffer memory barrier is <> to access via the specified buffer range. Buffer memory barriers can: also be used to define a <> for the specified buffer range. [open,refpage='VkBufferMemoryBarrier',desc='Structure specifying a buffer memory barrier',type='structs'] -- The sname:VkBufferMemoryBarrier structure is defined as: include::../api/structs/VkBufferMemoryBarrier.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:srcAccessMask is a bitmask of elink:VkAccessFlagBits specifying a <>. * pname:dstAccessMask is a bitmask of elink:VkAccessFlagBits specifying a <>. * pname:srcQueueFamilyIndex is the source queue family for a <>. * pname:dstQueueFamilyIndex is the destination queue family for a <>. * pname:buffer is a handle to the buffer whose backing memory is affected by the barrier. * pname:offset is an offset in bytes into the backing memory for pname:buffer; this is relative to the base offset as bound to the buffer (see flink:vkBindBufferMemory). * pname:size is a size in bytes of the affected area of backing memory for pname:buffer, or ename:VK_WHOLE_SIZE to use the range from pname:offset to the end of the buffer. The first <> is limited to access to memory through the specified buffer range, via access types in the <> specified by pname:srcAccessMask. If pname:srcAccessMask includes ename:VK_ACCESS_HOST_WRITE_BIT, memory writes performed by that access type are also made visible, as that access type is not performed through a resource. The second <> is limited to access to memory through the specified buffer range, via access types in the <>. specified by pname:dstAccessMask. If pname:dstAccessMask includes ename:VK_ACCESS_HOST_WRITE_BIT or ename:VK_ACCESS_HOST_READ_BIT, available memory writes are also made visible to accesses of those types, as those access types are not performed through a resource. If pname:srcQueueFamilyIndex is not equal to pname:dstQueueFamilyIndex, and pname:srcQueueFamilyIndex is equal to the current queue family, then the memory barrier defines a <> for the specified buffer range, and the second access scope includes no access, as if pname:dstAccessMask was `0`. If pname:dstQueueFamilyIndex is not equal to pname:srcQueueFamilyIndex, and pname:dstQueueFamilyIndex is equal to the current queue family, then the memory barrier defines a <> for the specified buffer range, and the first access scope includes no access, as if pname:srcAccessMask was `0`. .Valid Usage **** * [[VUID-VkBufferMemoryBarrier-offset-01187]] pname:offset must: be less than the size of pname:buffer * [[VUID-VkBufferMemoryBarrier-size-01188]] If pname:size is not equal to ename:VK_WHOLE_SIZE, pname:size must: be greater than `0` * [[VUID-VkBufferMemoryBarrier-size-01189]] If pname:size is not equal to ename:VK_WHOLE_SIZE, pname:size must: be less than or equal to than the size of pname:buffer minus pname:offset ifndef::VK_VERSION_1_1,VK_KHR_external_memory[] * [[VUID-VkBufferMemoryBarrier-buffer-01190]] If pname:buffer was created with a sharing mode of ename:VK_SHARING_MODE_CONCURRENT, pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex must: both be ename:VK_QUEUE_FAMILY_IGNORED endif::VK_VERSION_1_1,VK_KHR_external_memory[] ifdef::VK_VERSION_1_1,VK_KHR_external_memory[] * [[VUID-VkBufferMemoryBarrier-buffer-01191]] If pname:buffer was created with a sharing mode of ename:VK_SHARING_MODE_CONCURRENT, at least one of pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex must: be ename:VK_QUEUE_FAMILY_IGNORED * [[VUID-VkBufferMemoryBarrier-buffer-01763]] If pname:buffer was created with a sharing mode of ename:VK_SHARING_MODE_CONCURRENT, and one of pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex is ename:VK_QUEUE_FAMILY_IGNORED, the other must: be ename:VK_QUEUE_FAMILY_IGNORED or a special queue family reserved for external memory ownership transfers, as described in <>. endif::VK_VERSION_1_1,VK_KHR_external_memory[] ifndef::VK_VERSION_1_1,VK_KHR_external_memory[] * [[VUID-VkBufferMemoryBarrier-buffer-01192]] If pname:buffer was created with a sharing mode of ename:VK_SHARING_MODE_EXCLUSIVE, pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex must: either both be ename:VK_QUEUE_FAMILY_IGNORED, or both be a valid queue family (see <>) endif::VK_VERSION_1_1,VK_KHR_external_memory[] ifdef::VK_VERSION_1_1,VK_KHR_external_memory[] * [[VUID-VkBufferMemoryBarrier-buffer-01193]] If pname:buffer was created with a sharing mode of ename:VK_SHARING_MODE_EXCLUSIVE and pname:srcQueueFamilyIndex is ename:VK_QUEUE_FAMILY_IGNORED, pname:dstQueueFamilyIndex must: also be ename:VK_QUEUE_FAMILY_IGNORED * [[VUID-VkBufferMemoryBarrier-buffer-01764]] If pname:buffer was created with a sharing mode of ename:VK_SHARING_MODE_EXCLUSIVE and pname:srcQueueFamilyIndex is not ename:VK_QUEUE_FAMILY_IGNORED, it must: be a valid queue family or a special queue family reserved for external memory transfers, as described in <>. * [[VUID-VkBufferMemoryBarrier-buffer-01765]] If pname:buffer was created with a sharing mode of ename:VK_SHARING_MODE_EXCLUSIVE and pname:dstQueueFamilyIndex is not ename:VK_QUEUE_FAMILY_IGNORED, it must: be a valid queue family or a special queue family reserved for external memory transfers, as described in <>. endif::VK_VERSION_1_1,VK_KHR_external_memory[] * [[VUID-VkBufferMemoryBarrier-buffer-01196]] If pname:buffer was created with a sharing mode of ename:VK_SHARING_MODE_EXCLUSIVE, and pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex are not ename:VK_QUEUE_FAMILY_IGNORED, at least one of them must: be the same as the family of the queue that will execute this barrier * [[VUID-VkBufferMemoryBarrier-buffer-01931]] If pname:buffer is non-sparse then it must: be bound completely and contiguously to a single sname:VkDeviceMemory object **** include::../validity/structs/VkBufferMemoryBarrier.txt[] -- [[synchronization-image-memory-barriers]] === Image Memory Barriers Image memory barriers only apply to memory accesses involving a specific image subresource range. That is, a memory dependency formed from an image memory barrier is <> to access via the specified image subresource range. Image memory barriers can: also be used to define <> or a <> for the specified image subresource range. [open,refpage='VkImageMemoryBarrier',desc='Structure specifying the parameters of an image memory barrier',type='structs'] -- The sname:VkImageMemoryBarrier structure is defined as: include::../api/structs/VkImageMemoryBarrier.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:srcAccessMask is a bitmask of elink:VkAccessFlagBits specifying a <>. * pname:dstAccessMask is a bitmask of elink:VkAccessFlagBits specifying a <>. * pname:oldLayout is the old layout in an <>. * pname:newLayout is the new layout in an <>. * pname:srcQueueFamilyIndex is the source queue family for a <>. * pname:dstQueueFamilyIndex is the destination queue family for a <>. * pname:image is a handle to the image affected by this barrier. * pname:subresourceRange describes the <> within pname:image that is affected by this barrier. The first <> is limited to access to memory through the specified image subresource range, via access types in the <> specified by pname:srcAccessMask. If pname:srcAccessMask includes ename:VK_ACCESS_HOST_WRITE_BIT, memory writes performed by that access type are also made visible, as that access type is not performed through a resource. The second <> is limited to access to memory through the specified image subresource range, via access types in the <> specified by pname:dstAccessMask. If pname:dstAccessMask includes ename:VK_ACCESS_HOST_WRITE_BIT or ename:VK_ACCESS_HOST_READ_BIT, available memory writes are also made visible to accesses of those types, as those access types are not performed through a resource. If pname:srcQueueFamilyIndex is not equal to pname:dstQueueFamilyIndex, and pname:srcQueueFamilyIndex is equal to the current queue family, then the memory barrier defines a <> for the specified image subresource range, and the second access scope includes no access, as if pname:dstAccessMask was `0`. If pname:dstQueueFamilyIndex is not equal to pname:srcQueueFamilyIndex, and pname:dstQueueFamilyIndex is equal to the current queue family, then the memory barrier defines a <> for the specified image subresource range, and the first access scope includes no access, as if pname:srcAccessMask was `0`. If pname:oldLayout is not equal to pname:newLayout, then the memory barrier defines an <> for the specified image subresource range. [[synchronization-image-barrier-layout-transition-order]] Layout transitions that are performed via image memory barriers execute in their entirety in <>, relative to other image layout transitions submitted to the same queue, including those performed by <>. In effect there is an implicit execution dependency from each such layout transition to all layout transitions previously submitted to the same queue. ifdef::VK_EXT_sample_locations[] The image layout of each image subresource of a depth/stencil image created with ename:VK_IMAGE_CREATE_SAMPLE_LOCATIONS_COMPATIBLE_DEPTH_BIT_EXT is dependent on the last sample locations used to render to the image subresource as a depth/stencil attachment, thus when the pname:image member of an sname:VkImageMemoryBarrier is an image created with this flag the application can: chain a slink:VkSampleLocationsInfoEXT structure to the pname:pNext chain of sname:VkImageMemoryBarrier to specify the sample locations to use during the image layout transition. If the sname:VkSampleLocationsInfoEXT structure in the pname:pNext chain of sname:VkImageMemoryBarrier does not match the sample location state last used to render to the image subresource range specified by pname:subresourceRange or if no sname:VkSampleLocationsInfoEXT structure is in the pname:pNext chain of sname:VkImageMemoryBarrier then the contents of the given image subresource range becomes undefined: as if pname:oldLayout would equal ename:VK_IMAGE_LAYOUT_UNDEFINED. endif::VK_EXT_sample_locations[] ifdef::VK_VERSION_1_1,VK_KHR_sampler_ycbcr_conversion[] If pname:image has a multi-planar format and the image is _disjoint_, then including ename:VK_IMAGE_ASPECT_COLOR_BIT in the pname:aspectMask member of pname:subresourceRange is equivalent to including ename:VK_IMAGE_ASPECT_PLANE_0_BIT, ename:VK_IMAGE_ASPECT_PLANE_1_BIT, and (for three-plane formats only) ename:VK_IMAGE_ASPECT_PLANE_2_BIT. endif::VK_VERSION_1_1,VK_KHR_sampler_ycbcr_conversion[] .Valid Usage **** * [[VUID-VkImageMemoryBarrier-oldLayout-01197]] pname:oldLayout must: be ename:VK_IMAGE_LAYOUT_UNDEFINED or the current layout of the image subresources affected by the barrier * [[VUID-VkImageMemoryBarrier-newLayout-01198]] pname:newLayout must: not be ename:VK_IMAGE_LAYOUT_UNDEFINED or ename:VK_IMAGE_LAYOUT_PREINITIALIZED ifndef::VK_VERSION_1_1,VK_KHR_external_memory[] * [[VUID-VkImageMemoryBarrier-image-01199]] If pname:image was created with a sharing mode of ename:VK_SHARING_MODE_CONCURRENT, pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex must: both be ename:VK_QUEUE_FAMILY_IGNORED endif::VK_VERSION_1_1,VK_KHR_external_memory[] ifdef::VK_VERSION_1_1,VK_KHR_external_memory[] * [[VUID-VkImageMemoryBarrier-image-01381]] If pname:image was created with a sharing mode of ename:VK_SHARING_MODE_CONCURRENT, at least one of pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex must: be ename:VK_QUEUE_FAMILY_IGNORED * [[VUID-VkImageMemoryBarrier-image-01766]] If pname:image was created with a sharing mode of ename:VK_SHARING_MODE_CONCURRENT, and one of pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex is ename:VK_QUEUE_FAMILY_IGNORED, the other must: be ename:VK_QUEUE_FAMILY_IGNORED or a special queue family reserved for external memory transfers, as described in <>. endif::VK_VERSION_1_1,VK_KHR_external_memory[] ifndef::VK_VERSION_1_1,VK_KHR_external_memory[] * [[VUID-VkImageMemoryBarrier-image-01200]] If pname:image was created with a sharing mode of ename:VK_SHARING_MODE_EXCLUSIVE, pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex must: either both be ename:VK_QUEUE_FAMILY_IGNORED, or both be a valid queue family (see <>). endif::VK_VERSION_1_1,VK_KHR_external_memory[] ifdef::VK_VERSION_1_1,VK_KHR_external_memory[] * [[VUID-VkImageMemoryBarrier-image-01201]] If pname:image was created with a sharing mode of ename:VK_SHARING_MODE_EXCLUSIVE and pname:srcQueueFamilyIndex is ename:VK_QUEUE_FAMILY_IGNORED, pname:dstQueueFamilyIndex must: also be ename:VK_QUEUE_FAMILY_IGNORED. * [[VUID-VkImageMemoryBarrier-image-01767]] If pname:image was created with a sharing mode of ename:VK_SHARING_MODE_EXCLUSIVE and pname:srcQueueFamilyIndex is not ename:VK_QUEUE_FAMILY_IGNORED, it must: be a valid queue family or a special queue family reserved for external memory transfers, as described in <>. * [[VUID-VkImageMemoryBarrier-image-01768]] If pname:image was created with a sharing mode of ename:VK_SHARING_MODE_EXCLUSIVE and pname:dstQueueFamilyIndex is not ename:VK_QUEUE_FAMILY_IGNORED, it must: be a valid queue family or a special queue family reserved for external memory transfers, as described in <>. endif::VK_VERSION_1_1,VK_KHR_external_memory[] * [[VUID-VkImageMemoryBarrier-image-01205]] If pname:image was created with a sharing mode of ename:VK_SHARING_MODE_EXCLUSIVE, and pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex are not ename:VK_QUEUE_FAMILY_IGNORED, at least one of them must: be the same as the family of the queue that will execute this barrier * [[VUID-VkImageMemoryBarrier-subresourceRange-01486]] pname:subresourceRange.baseMipLevel must: be less than the pname:mipLevels specified in slink:VkImageCreateInfo when pname:image was created * [[VUID-VkImageMemoryBarrier-subresourceRange-01724]] If pname:subresourceRange.levelCount is not ename:VK_REMAINING_MIP_LEVELS, [eq]#pname:subresourceRange.baseMipLevel {plus} pname:subresourceRange.levelCount# must: be less than or equal to the pname:mipLevels specified in slink:VkImageCreateInfo when pname:image was created * [[VUID-VkImageMemoryBarrier-subresourceRange-01488]] pname:subresourceRange.baseArrayLayer must: be less than the pname:arrayLayers specified in slink:VkImageCreateInfo when pname:image was created * [[VUID-VkImageMemoryBarrier-subresourceRange-01725]] If pname:subresourceRange.layerCount is not ename:VK_REMAINING_ARRAY_LAYERS, [eq]#pname:subresourceRange.baseArrayLayer {plus} pname:subresourceRange.layerCount# must: be less than or equal to the pname:arrayLayers specified in slink:VkImageCreateInfo when pname:image was created * [[VUID-VkImageMemoryBarrier-image-01207]] If pname:image has a depth/stencil format with both depth and stencil components, then the pname:aspectMask member of pname:subresourceRange must: include both ename:VK_IMAGE_ASPECT_DEPTH_BIT and ename:VK_IMAGE_ASPECT_STENCIL_BIT ifdef::VK_VERSION_1_1,VK_KHR_sampler_ycbcr_conversion[] * [[VUID-VkImageMemoryBarrier-image-01671]] If pname:image has a single-plane color format or is not _disjoint_, then the pname:aspectMask member of pname:subresourceRange must: be ename:VK_IMAGE_ASPECT_COLOR_BIT * [[VUID-VkImageMemoryBarrier-image-01672]] If pname:image has a multi-planar format and the image is _disjoint_, then the pname:aspectMask member of pname:subresourceRange must: include either at least one of ename:VK_IMAGE_ASPECT_PLANE_0_BIT, ename:VK_IMAGE_ASPECT_PLANE_1_BIT, and ename:VK_IMAGE_ASPECT_PLANE_2_BIT; or must: include ename:VK_IMAGE_ASPECT_COLOR_BIT * [[VUID-VkImageMemoryBarrier-image-01673]] If pname:image has a multi-planar format with only two planes, then the pname:aspectMask member of pname:subresourceRange must: not include ename:VK_IMAGE_ASPECT_PLANE_2_BIT endif::VK_VERSION_1_1,VK_KHR_sampler_ycbcr_conversion[] * [[VUID-VkImageMemoryBarrier-oldLayout-01208]] If either pname:oldLayout or pname:newLayout is ename:VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL then pname:image must: have been created with ename:VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT set * [[VUID-VkImageMemoryBarrier-oldLayout-01209]] If either pname:oldLayout or pname:newLayout is ename:VK_IMAGE_LAYOUT_DEPTH_STENCIL_ATTACHMENT_OPTIMAL then pname:image must: have been created with ename:VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT set * [[VUID-VkImageMemoryBarrier-oldLayout-01210]] If either pname:oldLayout or pname:newLayout is ename:VK_IMAGE_LAYOUT_DEPTH_STENCIL_READ_ONLY_OPTIMAL then pname:image must: have been created with ename:VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT set ifdef::VK_VERSION_1_1,VK_KHR_maintenance2[] * [[VUID-VkImageMemoryBarrier-oldLayout-01658]] If either pname:oldLayout or pname:newLayout is ename:VK_IMAGE_LAYOUT_DEPTH_READ_ONLY_STENCIL_ATTACHMENT_OPTIMAL then pname:image must: have been created with ename:VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT set * [[VUID-VkImageMemoryBarrier-oldLayout-01659]] If either pname:oldLayout or pname:newLayout is ename:VK_IMAGE_LAYOUT_DEPTH_ATTACHMENT_STENCIL_READ_ONLY_OPTIMAL then pname:image must: have been created with ename:VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT set endif::VK_VERSION_1_1,VK_KHR_maintenance2[] * [[VUID-VkImageMemoryBarrier-oldLayout-01211]] If either pname:oldLayout or pname:newLayout is ename:VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL then pname:image must: have been created with ename:VK_IMAGE_USAGE_SAMPLED_BIT or ename:VK_IMAGE_USAGE_INPUT_ATTACHMENT_BIT set * [[VUID-VkImageMemoryBarrier-oldLayout-01212]] If either pname:oldLayout or pname:newLayout is ename:VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL then pname:image must: have been created with ename:VK_IMAGE_USAGE_TRANSFER_SRC_BIT set * [[VUID-VkImageMemoryBarrier-oldLayout-01213]] If either pname:oldLayout or pname:newLayout is ename:VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL then pname:image must: have been created with ename:VK_IMAGE_USAGE_TRANSFER_DST_BIT set * [[VUID-VkImageMemoryBarrier-image-01932]] If pname:image is non-sparse then it must: be bound completely and contiguously to a single sname:VkDeviceMemory object ifdef::VK_NV_shading_rate_image[] * [[VUID-VkImageMemoryBarrier-oldLayout-02088]] If either pname:oldLayout or pname:newLayout is ename:VK_IMAGE_LAYOUT_SHADING_RATE_OPTIMAL_NV then pname:image must: have been created with ename:VK_IMAGE_USAGE_SHADING_RATE_IMAGE_BIT_NV set endif::VK_NV_shading_rate_image[] **** include::../validity/structs/VkImageMemoryBarrier.txt[] -- [[synchronization-queue-transfers]] === Queue Family Ownership Transfer Resources created with a elink:VkSharingMode of ename:VK_SHARING_MODE_EXCLUSIVE must: have their ownership explicitly transferred from one queue family to another in order to access their content in a well-defined manner on a queue in a different queue family. ifdef::VK_VERSION_1_1,VK_KHR_external_memory[] Resources shared with external APIs or instances using external memory must: also explicitly manage ownership transfers between local and external queues (or equivalent constructs in external APIs) regardless of the elink:VkSharingMode specified when creating them. The special queue family index ename:VK_QUEUE_FAMILY_EXTERNAL represents any queue external to the resource's current Vulkan instance, as long as the queue uses the same underlying physical device ifdef::VK_VERSION_1_1,VK_KHR_device_group[] or device group endif::VK_VERSION_1_1,VK_KHR_device_group[] and uses the same driver version as the resource's slink:VkDevice, as indicated by slink:VkPhysicalDeviceIDProperties::pname:deviceUUID and slink:VkPhysicalDeviceIDProperties::pname:driverUUID. ifdef::VK_EXT_queue_family_foreign[] The special queue family index ename:VK_QUEUE_FAMILY_FOREIGN_EXT represents any queue external to the resource's current Vulkan instance, regardless of the queue's underlying physical device or driver version. This includes, for example, queues for fixed-function image processing devices, media codec devices, and display devices, as well as all queues that use the same underlying physical device ifdef::VK_VERSION_1_1,VK_KHR_device_group[] (or device group) endif::VK_VERSION_1_1,VK_KHR_device_group[] and driver version as the resource's slink:VkDevice. endif::VK_EXT_queue_family_foreign[] endif::VK_VERSION_1_1,VK_KHR_external_memory[] If memory dependencies are correctly expressed between uses of such a resource between two queues in different families, but no ownership transfer is defined, the contents of that resource are undefined: for any read accesses performed by the second queue family. [NOTE] .Note ==== If an application does not need the contents of a resource to remain valid when transferring from one queue family to another, then the ownership transfer should: be skipped. ==== ifdef::VK_EXT_queue_family_foreign[] [NOTE] .Note ==== Applications should expect transfers to/from ename:VK_QUEUE_FAMILY_FOREIGN_EXT to be more expensive than transfers to/from ename:VK_QUEUE_FAMILY_EXTERNAL_KHR. ==== endif::VK_EXT_queue_family_foreign[] A queue family ownership transfer consists of two distinct parts: . Release exclusive ownership from the source queue family . Acquire exclusive ownership for the destination queue family An application must: ensure that these operations occur in the correct order by defining an execution dependency between them, e.g. using a semaphore. [[synchronization-queue-transfers-release]] A _release operation_ is used to release exclusive ownership of a range of a buffer or image subresource range. A release operation is defined by executing a <> (for a buffer range) or an <> (for an image subresource range), on a queue from the source queue family. The pname:srcQueueFamilyIndex parameter of the barrier must: be set to the source queue family index, and the pname:dstQueueFamilyIndex parameter to the destination queue family index. pname:dstStageMask is ignored for such a barrier, such that no visibility operation is executed - the value of this mask does not affect the validity of the barrier. The release operation happens-after the availability operation. [[synchronization-queue-transfers-acquire]] An _acquire operation_ is used to acquire exclusive ownership of a range of a buffer or image subresource range. An acquire operation is defined by executing a <> (for a buffer range) or an <> (for an image subresource range), on a queue from the destination queue family. The buffer range or image subresource range specified in an acquire operation must: match exactly that of a previous release operation. The pname:srcQueueFamilyIndex parameter of the barrier must: be set to the source queue family index, and the pname:dstQueueFamilyIndex parameter to the destination queue family index. pname:srcStageMask is ignored for such a barrier, such that no availability operation is executed - the value of this mask does not affect the validity of the barrier. The acquire operation happens-before the visibility operation. [NOTE] .Note ==== Whilst it is not invalid to provide destination or source access masks for memory barriers used for release or acquire operations, respectively, they have no practical effect. Access after a release operation has undefined: results, and so visibility for those accesses has no practical effect. Similarly, write access before an acquire operation will produce undefined: results for future access, so availability of those writes has no practical use. In an earlier version of the specification, these were required to match on both sides - but this was subsequently relaxed. These masks should: be set to 0. ==== If the transfer is via an image memory barrier, and an <> is desired, then the values of pname:oldLayout and pname:newLayout in the release memory barrier must: be equal to values of pname:oldLayout and pname:newLayout in the acquire memory barrier. Although the image layout transition is submitted twice, it will only be executed once. A layout transition specified in this way happens-after the release operation and happens-before the acquire operation. If the values of pname:srcQueueFamilyIndex and pname:dstQueueFamilyIndex are equal, no ownership transfer is performed, and the barrier operates as if they were both set to ename:VK_QUEUE_FAMILY_IGNORED. Queue family ownership transfers may: perform read and write accesses on all memory bound to the image subresource or buffer range, so applications must: ensure that all memory writes have been made <> before a queue family ownership transfer is executed. Available memory is automatically made visible to queue family release and acquire operations, and writes performed by those operations are automatically made available. Once a queue family has acquired ownership of a buffer range or image subresource range of an ename:VK_SHARING_MODE_EXCLUSIVE resource, its contents are undefined: to other queue families unless ownership is transferred. The contents of any portion of another resource which aliases memory that is bound to the transferred buffer or image subresource range are undefined: after a release or acquire operation. [[synchronization-wait-idle]] == Wait Idle Operations [open,refpage='vkQueueWaitIdle',desc='Wait for a queue to become idle',type='protos'] -- To wait on the host for the completion of outstanding queue operations for a given queue, call: include::../api/protos/vkQueueWaitIdle.txt[] * pname:queue is the queue on which to wait. fname:vkQueueWaitIdle is equivalent to submitting a fence to a queue and waiting with an infinite timeout for that fence to signal. include::../validity/protos/vkQueueWaitIdle.txt[] -- [open,refpage='vkDeviceWaitIdle',desc='Wait for a device to become idle',type='protos'] -- To wait on the host for the completion of outstanding queue operations for all queues on a given logical device, call: include::../api/protos/vkDeviceWaitIdle.txt[] * pname:device is the logical device to idle. fname:vkDeviceWaitIdle is equivalent to calling fname:vkQueueWaitIdle for all queues owned by pname:device. include::../validity/protos/vkDeviceWaitIdle.txt[] -- [[synchronization-submission-host-writes]] == Host Write Ordering Guarantees When batches of command buffers are submitted to a queue via flink:vkQueueSubmit, it defines a memory dependency with prior host operations, and execution of command buffers submitted to the queue. The first <> is defined by the host execution model, but includes execution of flink:vkQueueSubmit on the host and anything that happened-before it. The second <> includes all commands submitted in the same <>, and all commands that occur later in <>. The first <> includes all host writes to mappable device memory that are available to the host memory domain. The second <> includes all memory access performed by the device. ifdef::VK_VERSION_1_1,VK_KHR_device_group[] [[synchronization-device-group]] == Synchronization and Multiple Physical Devices If a logical device includes more than one physical device, then fences, semaphores, and events all still have a single instance of the signaled state. A fence becomes signaled when all physical devices complete the necessary queue operations. Semaphore wait and signal operations all include a device index that is the sole physical device that performs the operation. These indices are provided in the slink:VkDeviceGroupSubmitInfo and slink:VkDeviceGroupBindSparseInfo structures. Semaphores are not exclusively owned by any physical device. For example, a semaphore can be signaled by one physical device and then waited on by a different physical device. An event can: only be waited on by the same physical device that signaled it (or the host). endif::VK_VERSION_1_1,VK_KHR_device_group[] ifdef::VK_EXT_calibrated_timestamps[] [[calibrated-timestamps]] == Calibrated timestamps [open,refpage='vkGetCalibratedTimestampsEXT',desc='Query calibrated timestamps',type='protos'] -- In order to be able to correlate the time a particular operation took place at on timelines of different time domains (e.g. a device operation vs a host operation), Vulkan allows querying calibrated timestamps from multiple time domains. To query calibrated timestamps from a set of time domains, call: include::../api/protos/vkGetCalibratedTimestampsEXT.txt[] * pname:device is the logical device used to perform the query. * pname:timestampCount is the number of timestamps to query. * pname:pTimestampInfos is a pointer to an array of pname:timestampCount number of structures of type slink:VkCalibratedTimestampInfoEXT, describing the time domains the calibrated timestamps should be captured from. * pname:pTimestamps is a pointer to an array of pname:timestampCount number of 64-bit unsigned integer values in which the requested calibrated timestamp values are returned. * pname:pMaxDeviation is a pointer to a 64-bit unsigned integer value in which the strictly positive maximum deviation, in nanoseconds, of the calibrated timestamp values is returned. [NOTE] .Note ==== The maximum deviation may: vary between calls to fname:vkGetCalibratedTimestampsEXT even for the same set of time domains due to implementation and platform specific reasons. It's the application's responsibility to assess whether the returned maximum deviation makes the timestamp values suitable for any particular purpose and can: choose to re-issue the timestamp calibration call pursuing a lower devation value. ==== Calibrated timestamp values can: be extrapolated to estimate future coinciding timestamp values, however, depending on the nature of the time domains and other properties of the platform extrapolating values over a sufficiently long period of time may: no longer be accurate enough to fit any particular purpose so applications are expected to re-calibrate the timestamps on a regular basis. include::../validity/protos/vkGetCalibratedTimestampsEXT.txt[] -- [open,refpage='VkCalibratedTimestampInfoEXT',desc='Structure specifying the input parameters of a calibrated timestamp query',type='structs'] -- The sname:VkCalibratedTimestampInfoEXT structure is defined as: include::../api/structs/VkCalibratedTimestampInfoEXT.txt[] * pname:sType is the type of this structure. * pname:pNext is `NULL` or a pointer to an extension-specific structure. * pname:timeDomain is a elink:VkTimeDomainEXT value specifying the time domain from which the calibrated timestamp value should be returned. .Valid Usage **** * [[VUID-VkCalibratedTimestampInfoEXT-timeDomain-02354]] pname:timeDomain must: be one of the elink:VkTimeDomainEXT values returned by flink:vkGetPhysicalDeviceCalibrateableTimeDomainsEXT **** include::../validity/structs/VkCalibratedTimestampInfoEXT.txt[] -- [open,refpage='VkTimeDomainEXT',desc='Supported time domains',type='enums'] -- The set of supported time domains consists of: include::../api/enums/VkTimeDomainEXT.txt[] * ename:VK_TIME_DOMAIN_DEVICE_EXT specifies the device time domain. Timestamp values in this time domain are comparable with device timestamp values captured using flink:vkCmdWriteTimestamp and are defined to be incrementing according to the <> of the device. * ename:VK_TIME_DOMAIN_CLOCK_MONOTONIC_EXT specifies the CLOCK_MONOTONIC time domain available on POSIX platforms. * ename:VK_TIME_DOMAIN_CLOCK_MONOTONIC_RAW_EXT specifies the CLOCK_MONOTONIC_RAW time domain available on POSIX platforms. * ename:VK_TIME_DOMAIN_QUERY_PERFORMANCE_COUNTER_EXT specifies the performance counter (QPC) time domain available on Windows. -- endif::VK_EXT_calibrated_timestamps[]