Vulkan-Hpp/samples/SharedHandles/SharedHandles.cpp
Ilya Doroshenko 5d0d3b7bd9
Shared Handles (#1619)
* stage 1: added shared handles and all traits

* moved deleter to the shared part, now Shared handle weights 2 pointers

* Fixed Destroy Function

* generated headers

* At least no errors.

TODO: get image overload of shared handle back
make an example

* fixed all SFINAE issues, moved image specialization

* Added a sample

* made better example, added specialization for swapchain, added shared handles to readme

* Major update:
Fixed all wishes
Made std::shared_ptr do heavy lifting
Reduced code complexity
Added ParentType to basic handles
Added put and put_native for cross-abi usage

* more readme

* safer release()
removed put() for now

* better synchronization
inspired by STL shared_ptr implementation

* removed test shared pointer

* added forward type for better interop with VMA
added parent checking

* fixed getParent()

* added non-deleting overloads for Queue, DisplayKHR and PhysicalDevice

* Shared non-destoyed type now have handles

* using constructors are made to comply standard

* fixed leak

* shared handles migrated to own header,
made module entries

* header is independent now

* rebased repo, updated the example

* fixed most of the stuff

* renamed parent of, added guide to make shared handles

* vulkansc

* moved destructor to shared handle traits

* resolved issues with CI

* more relaxed memory ordering for release operation

* trying to remove submodule

* renamed to destructorType

* suppressed function cast warning

* only GCC should be affected
2023-09-19 13:04:36 +02:00

378 lines
17 KiB
C++

// Copyright(c) 2023, Ilya Doroshenko. All rights reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// VulkanHpp Samples : SharedHandles
// Draw a textured cube using shared handles for resource management and correct order of destruction
#include "../utils/geometries.hpp"
#include "../utils/math.hpp"
#include "../utils/shaders.hpp"
#include "../utils/utils.hpp"
#include "SPIRV/GlslangToSpv.h"
#include <iostream>
#include <thread>
#include <vulkan/vulkan_shared.hpp>
static char const * AppName = "SharedHandles";
static char const * EngineName = "Vulkan.hpp";
std::vector<vk::SharedFramebuffer> makeSharedFramebuffers( const vk::SharedDevice & device,
const vk::SharedRenderPass & renderPass,
const std::vector<vk::ImageView> & imageViews,
const vk::SharedImageView & depthImageView,
const vk::Extent2D & extent )
{
auto renderPassHandle = renderPass.get(); // lvalue reference is required for the capture below
std::vector<vk::SharedFramebuffer> sharedFramebuffers;
std::vector<vk::Framebuffer> framebuffers = vk::su::createFramebuffers( device.get(), renderPassHandle, imageViews, depthImageView.get(), extent );
sharedFramebuffers.reserve( framebuffers.size() );
for ( auto & framebuffer : framebuffers )
{
sharedFramebuffers.emplace_back( framebuffer, device );
}
return sharedFramebuffers;
}
class Window
{
public:
Window( const char * windowName, vk::Extent2D extent ) : window( vk::su::createWindow( windowName, extent ) ) {}
public:
vk::su::WindowData window;
};
class Engine
{
public:
Engine( const vk::su::WindowData & window )
{
instance = vk::SharedInstance{ vk::su::createInstance( AppName, EngineName, {}, vk::su::getInstanceExtensions() ) };
#if !defined( NDEBUG )
debugUtilsMessenger =
vk::SharedDebugUtilsMessengerEXT{ instance->createDebugUtilsMessengerEXT( vk::su::makeDebugUtilsMessengerCreateInfoEXT() ), instance };
#endif
physicalDevice = instance->enumeratePhysicalDevices().front();
createDeviceAndSwapChain( window );
initialize();
}
void createDeviceAndSwapChain( const vk::su::WindowData & window )
{
VkSurfaceKHR surface;
VkResult err = glfwCreateWindowSurface( static_cast<VkInstance>( instance.get() ), window.handle, nullptr, &surface );
if ( err != VK_SUCCESS )
throw std::runtime_error( "Failed to create window!" );
vk::SharedSurfaceKHR sharedSurface{ surface, instance };
auto graphicsAndPresentQueueFamilyIndex = vk::su::findGraphicsAndPresentQueueFamilyIndex( physicalDevice, sharedSurface.get() );
device = vk::SharedDevice{ vk::su::createDevice( physicalDevice, graphicsAndPresentQueueFamilyIndex.first, vk::su::getDeviceExtensions() ) };
vk::su::SwapChainData swapChainData( physicalDevice,
device.get(),
sharedSurface.get(),
window.extent,
vk::ImageUsageFlagBits::eColorAttachment | vk::ImageUsageFlagBits::eTransferSrc,
{},
graphicsAndPresentQueueFamilyIndex.first,
graphicsAndPresentQueueFamilyIndex.second );
swapChain = vk::SharedSwapchainKHR{ swapChainData.swapChain, device, sharedSurface };
imageViews.reserve( swapChainData.images.size() );
images.reserve( swapChainData.images.size() );
// we don't want to destroy the images, since they are owned by the swapchain,
// but for the consistent representation we might want shared textures
std::transform( swapChainData.images.begin(),
swapChainData.images.end(),
std::back_inserter( images ),
[this]( vk::Image image ) {
return vk::SharedImage{ image, device, vk::SwapchainOwns::yes };
} );
std::transform( swapChainData.imageViews.begin(),
swapChainData.imageViews.end(),
std::back_inserter( imageViews ),
[this]( vk::ImageView imageView ) {
return vk::SharedImageView{ imageView, device };
} );
commandPool =
vk::SharedCommandPool{ device->createCommandPool( { vk::CommandPoolCreateFlagBits::eResetCommandBuffer, graphicsAndPresentQueueFamilyIndex.first } ),
device };
graphicsQueue = vk::SharedQueue{ device->getQueue( graphicsAndPresentQueueFamilyIndex.first, 0 ), device };
presentQueue = vk::SharedQueue{ device->getQueue( graphicsAndPresentQueueFamilyIndex.second, 0 ), device };
depthFormat = vk::Format::eD16Unorm;
vk::su::DepthBufferData depthBufferData( physicalDevice, device.get(), depthFormat, window.extent );
depthImage = vk::SharedImage{ depthBufferData.image, device };
depthImageView = vk::SharedImageView{ depthBufferData.imageView, device };
depthMemory = vk::SharedDeviceMemory{ depthBufferData.deviceMemory, device };
renderPass =
vk::SharedRenderPass{ vk::su::createRenderPass( device.get(),
vk::su::pickSurfaceFormat( physicalDevice.getSurfaceFormatsKHR( swapChain.getSurface().get() ) ).format,
depthFormat ),
device };
framebuffers = makeSharedFramebuffers( device, renderPass, swapChainData.imageViews, depthImageView, window.extent );
imageAcquiredSemaphore = vk::SharedSemaphore{ device->createSemaphore( vk::SemaphoreCreateInfo() ), device };
drawFence = vk::SharedFence{ device->createFence( vk::FenceCreateInfo() ), device };
// We don't need to explicitly keep sharedSurface anymore, it is owned by swapChain now.
}
void initialize()
{
commandBuffer = vk::SharedCommandBuffer{
device->allocateCommandBuffers( vk::CommandBufferAllocateInfo( commandPool.get(), vk::CommandBufferLevel::ePrimary, 1 ) ).front(), device, commandPool
};
auto device_handle = device.get();
descriptorSetLayout = vk::SharedDescriptorSetLayout{ vk::su::createDescriptorSetLayout(
device_handle,
{ { vk::DescriptorType::eUniformBuffer, 1, vk::ShaderStageFlagBits::eVertex },
{ vk::DescriptorType::eCombinedImageSampler, 1, vk::ShaderStageFlagBits::eFragment } } ),
device };
auto dsl = descriptorSetLayout.get();
pipelineLayout = vk::SharedPipelineLayout{ device->createPipelineLayout( vk::PipelineLayoutCreateInfo( vk::PipelineLayoutCreateFlags(), dsl ) ), device };
glslang::InitializeProcess();
vertexShaderModule = vk::SharedShaderModule{ vk::su::createShaderModule( device_handle, vk::ShaderStageFlagBits::eVertex, vertexShaderText_PT_T ), device };
fragmentShaderModule =
vk::SharedShaderModule{ vk::su::createShaderModule( device_handle, vk::ShaderStageFlagBits::eFragment, fragmentShaderText_T_C ), device };
glslang::FinalizeProcess();
descriptorPool = vk::SharedDescriptorPool{
vk::su::createDescriptorPool( device_handle, { { vk::DescriptorType::eUniformBuffer, 1 }, { vk::DescriptorType::eCombinedImageSampler, 1 } } ), device
};
descriptorSetAllocateInfo = vk::DescriptorSetAllocateInfo( descriptorPool.get(), dsl );
descriptorSet = vk::SharedDescriptorSet{ device->allocateDescriptorSets( descriptorSetAllocateInfo ).front(), device, descriptorPool };
pipelineCache = vk::SharedPipelineCache{ device->createPipelineCache( vk::PipelineCacheCreateInfo() ), device };
graphicsPipeline = vk::SharedPipeline{ vk::su::createGraphicsPipeline( device_handle,
pipelineCache.get(),
std::make_pair( vertexShaderModule.get(), nullptr ),
std::make_pair( fragmentShaderModule.get(), nullptr ),
sizeof( texturedCubeData[0] ),
{ { vk::Format::eR32G32B32A32Sfloat, 0 }, { vk::Format::eR32G32Sfloat, 16 } },
vk::FrontFace::eClockwise,
true,
pipelineLayout.get(),
renderPass.get() ),
device };
// Get the index of the next available swapchain image:
vk::ResultValue<uint32_t> currentBufferR = device->acquireNextImageKHR( swapChain.get(), vk::su::FenceTimeout, imageAcquiredSemaphore.get(), nullptr );
assert( currentBufferR.result == vk::Result::eSuccess );
assert( currentBufferR.value < framebuffers.size() );
currentBuffer = currentBufferR.value;
}
void beginFrame( vk::Extent2D extent )
{
std::array<vk::ClearValue, 2> clearValues;
clearValues[0].color = vk::ClearColorValue( 0.2f, 0.2f, 0.2f, 0.2f );
clearValues[1].depthStencil = vk::ClearDepthStencilValue( 1.0f, 0 );
vk::RenderPassBeginInfo renderPassBeginInfo( renderPass.get(), framebuffers[currentBuffer].get(), vk::Rect2D( vk::Offset2D( 0, 0 ), extent ), clearValues );
commandBuffer->begin( vk::CommandBufferBeginInfo() );
commandBuffer->beginRenderPass( renderPassBeginInfo, vk::SubpassContents::eInline );
commandBuffer->bindPipeline( vk::PipelineBindPoint::eGraphics, graphicsPipeline.get() );
commandBuffer->bindDescriptorSets( vk::PipelineBindPoint::eGraphics, pipelineLayout.get(), 0, descriptorSet.get(), nullptr );
commandBuffer->setViewport( 0, vk::Viewport( 0.0f, 0.0f, static_cast<float>( extent.width ), static_cast<float>( extent.height ), 0.0f, 1.0f ) );
commandBuffer->setScissor( 0, vk::Rect2D( vk::Offset2D( 0, 0 ), extent ) );
}
void endFrame()
{
commandBuffer->endRenderPass();
commandBuffer->end();
vk::PipelineStageFlags waitDestinationStageMask( vk::PipelineStageFlagBits::eColorAttachmentOutput );
auto ias = imageAcquiredSemaphore.get();
auto comBuf = commandBuffer.get();
vk::SubmitInfo submitInfo( ias, waitDestinationStageMask, comBuf );
graphicsQueue->submit( submitInfo, drawFence.get() );
while ( vk::Result::eTimeout == device->waitForFences( drawFence.get(), VK_TRUE, vk::su::FenceTimeout ) )
;
auto swap = swapChain.get();
vk::Result result = presentQueue->presentKHR( vk::PresentInfoKHR( {}, swap, currentBuffer ) );
switch ( result )
{
case vk::Result::eSuccess: break;
case vk::Result::eSuboptimalKHR: std::cout << "vk::Queue::presentKHR returned vk::Result::eSuboptimalKHR !\n"; break;
default: assert( false ); // an unexpected result is returned !
}
std::this_thread::sleep_for( std::chrono::milliseconds( 1000 ) );
/* VULKAN_KEY_END */
device->waitIdle();
}
public:
vk::SharedSwapchainKHR swapChain; // swapchain owns surface, that owns Instance, which should be destroyed last
vk::PhysicalDevice physicalDevice; // physical device does not have a shared handle since it is not destroyed
vk::SharedDevice device;
vk::SharedInstance instance; // we don't need to keep the instance, this is just for convenience
vk::SharedDebugUtilsMessengerEXT debugUtilsMessenger;
std::vector<vk::SharedImageView> imageViews;
std::vector<vk::SharedImage> images;
uint32_t currentBuffer = 0;
vk::SharedSemaphore imageAcquiredSemaphore;
// memory still needs to be before the resources that use it in order to get a proper destruction sequence.
vk::SharedDeviceMemory depthMemory;
vk::SharedImage depthImage;
vk::SharedImageView depthImageView;
vk::Format depthFormat;
vk::SharedCommandPool commandPool;
vk::SharedCommandBuffer commandBuffer;
vk::SharedQueue graphicsQueue; // queue is not destroyed, shared handle is purely for consistency
vk::SharedQueue presentQueue;
vk::SharedPipelineCache pipelineCache;
vk::SharedPipelineLayout pipelineLayout;
vk::SharedRenderPass renderPass;
vk::SharedPipeline graphicsPipeline;
vk::SharedDescriptorPool descriptorPool;
vk::DescriptorSetAllocateInfo descriptorSetAllocateInfo;
vk::SharedDescriptorSet descriptorSet;
vk::SharedDescriptorSetLayout descriptorSetLayout;
vk::SharedShaderModule vertexShaderModule;
vk::SharedShaderModule fragmentShaderModule;
std::vector<vk::SharedFramebuffer> framebuffers;
vk::SharedFence drawFence;
};
class Asset
{
public:
Asset( const Engine & engine, vk::Extent2D extent )
{
auto device_handle = engine.device.get();
vk::su::BufferData vertexBufferData( engine.physicalDevice, device_handle, sizeof( texturedCubeData ), vk::BufferUsageFlagBits::eVertexBuffer );
vk::su::copyToDevice( device_handle, vertexBufferData.deviceMemory, texturedCubeData, sizeof( texturedCubeData ) / sizeof( texturedCubeData[0] ) );
vertexBuffer = vk::SharedBuffer{ vertexBufferData.buffer, engine.device };
vertexBufferMemory = vk::SharedDeviceMemory{ vertexBufferData.deviceMemory, engine.device };
engine.commandBuffer->begin( vk::CommandBufferBeginInfo() );
vk::su::BufferData uniformBufferData( engine.physicalDevice, device_handle, sizeof( glm::mat4x4 ), vk::BufferUsageFlagBits::eUniformBuffer );
glm::mat4x4 mvpcMatrix = vk::su::createModelViewProjectionClipMatrix( extent );
vk::su::copyToDevice( device_handle, uniformBufferData.deviceMemory, mvpcMatrix );
uniformBufferMemory = vk::SharedDeviceMemory{ uniformBufferData.deviceMemory, engine.device };
uniformBuffer = vk::SharedBuffer{ uniformBufferData.buffer, engine.device };
vk::su::TextureData textureData( engine.physicalDevice, device_handle );
textureData.setImage( device_handle, engine.commandBuffer.get(), vk::su::CheckerboardImageGenerator() );
textureImage = vk::SharedImage{ textureData.imageData->image, engine.device };
textureImageMemory = vk::SharedDeviceMemory{ textureData.imageData->deviceMemory, engine.device };
textureImageView = vk::SharedImageView{ textureData.imageData->imageView, engine.device };
textureSampler = vk::SharedSampler{ textureData.sampler, engine.device };
vk::su::updateDescriptorSets(
device_handle, engine.descriptorSet.get(), { { vk::DescriptorType::eUniformBuffer, uniformBufferData.buffer, VK_WHOLE_SIZE, {} } }, textureData );
engine.commandBuffer->end();
vk::su::submitAndWait( device_handle, engine.graphicsQueue.get(), engine.commandBuffer.get() );
}
void draw( vk::CommandBuffer commandBuffer )
{
commandBuffer.bindVertexBuffers( 0, vertexBuffer.get(), { 0 } );
commandBuffer.draw( 12 * 3, 1, 0, 0 );
}
vk::SharedDeviceMemory vertexBufferMemory;
vk::SharedBuffer vertexBuffer;
vk::SharedDeviceMemory uniformBufferMemory;
vk::SharedBuffer uniformBuffer;
vk::SharedDeviceMemory textureImageMemory;
vk::SharedImage textureImage;
vk::SharedImageView textureImageView;
vk::SharedSampler textureSampler;
};
class Application
{
public:
Application() : window( AppName, vk::Extent2D( 500, 500 ) ), engine( window.window ), asset( engine, vk::Extent2D( 500, 500 ) ) {}
void renderFrame()
{
engine.beginFrame( vk::Extent2D( 500, 500 ) );
asset.draw( engine.commandBuffer.get() );
engine.endFrame();
}
int start()
{
renderFrame();
return 0;
}
private:
// order of window, engine and asset variables is important !
Window window;
Engine engine;
Asset asset;
};
int main( int /*argc*/, char ** /*argv*/ )
{
try
{
return Application{}.start();
}
catch ( vk::SystemError & err )
{
std::cout << "vk::SystemError: " << err.what() << std::endl;
exit( -1 );
}
catch ( std::exception & err )
{
std::cout << "std::exception: " << err.what() << std::endl;
exit( -1 );
}
catch ( ... )
{
std::cout << "unknown error\n";
exit( -1 );
}
}