vkCmdEndRendering

void vkCmdEndRendering(
    VkCommandBuffer                   commandBuffer);

pub fn cmd_end_rendering(
    command_buffer: vk::CommandBuffer,
);

void vkCmdEndRenderingKHR(
    VkCommandBuffer                   commandBuffer);

pub fn cmd_end_rendering_khr(
    command_buffer: vk::CommandBuffer,
);

There is no implicit ordering between separate render passes, even in the same command buffer, and even when the attachments match. Some applications rely on the continuation of rasterization order between multiple render passes with attachments defined in the same way, in order to perform non-rendering operations (such as copies or compute operations) between draw calls, but this has never been required by the specification. There is also no explicit barrier currently in the API that provides the guarantee that applications rely on without additional performance penalties.New applications should avoid relying on this ordering until an appropriate barrier is added to the API.Implementations where applications are performing this splitting are encouraged to continue supporting this guarantee until a suitable barrier is added to the API.Existing applications relying on this ordering should expect that it will continue working on platforms where it currently does. Once a new extension adds support for a new barrier, developers are encouraged to adapt their applications to use this when available.