vulkan/wsi/vktWsiSharedPresentableImageTests.cpp

/*-------------------------------------------------------------------------
 * Vulkan Conformance Tests
 * ------------------------
 *
 * Copyright (c) 2017 Google Inc.
 *
 * 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.
 *
 *//*!
 * \file
 * \brief Tests for shared presentable image extension
 *//*--------------------------------------------------------------------*/

#include "vktWsiSharedPresentableImageTests.hpp"
#include "vktCustomInstancesDevices.hpp"

#include "vktTestCaseUtil.hpp"
#include "vktTestGroupUtil.hpp"
#include "vkRefUtil.hpp"
#include "vkWsiPlatform.hpp"
#include "vkWsiUtil.hpp"
#include "vkQueryUtil.hpp"
#include "vkDeviceUtil.hpp"
#include "vkPlatform.hpp"
#include "vkTypeUtil.hpp"
#include "vkPrograms.hpp"
#include "vkCmdUtil.hpp"
#include "vkObjUtil.hpp"

#include "vkWsiUtil.hpp"

#include "tcuPlatform.hpp"
#include "tcuResultCollector.hpp"
#include "tcuTestLog.hpp"
#include "tcuCommandLine.hpp"

#include <vector>
#include <string>

using std::vector;
using std::string;

using tcu::Maybe;
using tcu::UVec2;
using tcu::TestLog;

namespace vkt
{
namespace wsi
{
namespace
{
enum Scaling
{
	SCALING_NONE,
	SCALING_UP,
	SCALING_DOWN
};

typedef vector<vk::VkExtensionProperties> Extensions;

void checkAllSupported (const Extensions& supportedExtensions, const vector<string>& requiredExtensions)
{
	for (vector<string>::const_iterator requiredExtName = requiredExtensions.begin();
		 requiredExtName != requiredExtensions.end();
		 ++requiredExtName)
	{
		if (!isExtensionStructSupported(supportedExtensions, vk::RequiredExtension(*requiredExtName)))
			TCU_THROW(NotSupportedError, (*requiredExtName + " is not supported").c_str());
	}
}

CustomInstance createInstanceWithWsi (Context&							context,
									  const Extensions&					supportedExtensions,
									  vk::wsi::Type						wsiType)
{
	const deUint32	version = context.getUsedApiVersion();
	vector<string>	extensions;

	if (!vk::isCoreInstanceExtension(version, "VK_KHR_get_physical_device_properties2"))
		extensions.push_back("VK_KHR_get_physical_device_properties2");

	extensions.push_back("VK_KHR_surface");
	extensions.push_back("VK_KHR_get_surface_capabilities2");
	// Required for device extension to expose new physical device bits (in this
	// case, presentation mode enums)
	extensions.push_back(getExtensionName(wsiType));
	if (isDisplaySurface(wsiType))
		extensions.push_back("VK_KHR_display");

	checkAllSupported(supportedExtensions, extensions);

	return vkt::createCustomInstanceWithExtensions(context, extensions);
}

vk::VkPhysicalDeviceFeatures getDeviceNullFeatures (void)
{
	vk::VkPhysicalDeviceFeatures features;
	deMemset(&features, 0, sizeof(features));
	return features;
}

vk::Move<vk::VkDevice> createDeviceWithWsi (const vk::PlatformInterface&		vkp,
											vk::VkInstance						instance,
											const vk::InstanceInterface&		vki,
											vk::VkPhysicalDevice				physicalDevice,
											const Extensions&					supportedExtensions,
											const deUint32						queueFamilyIndex,
											bool								requiresSharedPresentableImage,
											bool								validationEnabled,
											const vk::VkAllocationCallbacks*	pAllocator = DE_NULL)
{
	const float							queuePriorities[]	= { 1.0f };
	const vk::VkDeviceQueueCreateInfo	queueInfos[]		=
	{
		{
			vk::VK_STRUCTURE_TYPE_DEVICE_QUEUE_CREATE_INFO,
			DE_NULL,
			(vk::VkDeviceQueueCreateFlags)0,
			queueFamilyIndex,
			DE_LENGTH_OF_ARRAY(queuePriorities),
			&queuePriorities[0]
		}
	};
	const vk::VkPhysicalDeviceFeatures	features		= getDeviceNullFeatures();
	const char* const					extensions[]	=
	{
		"VK_KHR_swapchain",
		"VK_KHR_shared_presentable_image"
	};

	const vk::VkDeviceCreateInfo		deviceParams	=
	{
		vk::VK_STRUCTURE_TYPE_DEVICE_CREATE_INFO,
		DE_NULL,
		(vk::VkDeviceCreateFlags)0,
		DE_LENGTH_OF_ARRAY(queueInfos),
		&queueInfos[0],
		0u,
		DE_NULL,
		requiresSharedPresentableImage ? 2u : 1u,
		DE_ARRAY_BEGIN(extensions),
		&features
	};

	for (int ndx = 0; ndx < DE_LENGTH_OF_ARRAY(extensions); ++ndx)
	{
		if (!isExtensionStructSupported(supportedExtensions, vk::RequiredExtension(extensions[ndx])))
			TCU_THROW(NotSupportedError, (string(extensions[ndx]) + " is not supported").c_str());
	}

	return createCustomDevice(validationEnabled, vkp, instance, vki, physicalDevice, &deviceParams, pAllocator);
}

de::MovePtr<vk::wsi::Display> createDisplay (const vk::Platform&	platform,
											 const Extensions&		supportedExtensions,
											 vk::wsi::Type			wsiType)
{
	try
	{
		return de::MovePtr<vk::wsi::Display>(platform.createWsiDisplay(wsiType));
	}
	catch (const tcu::NotSupportedError& e)
	{
		if (isExtensionStructSupported(supportedExtensions, vk::RequiredExtension(getExtensionName(wsiType))) &&
		    platform.hasDisplay(wsiType))
		{
			// If VK_KHR_{platform}_surface was supported, vk::Platform implementation
			// must support creating native display & window for that WSI type.
			throw tcu::TestError(e.getMessage());
		}
		else
			throw;
	}
}

de::MovePtr<vk::wsi::Window> createWindow (const vk::wsi::Display& display, const Maybe<UVec2>& initialSize)
{
	try
	{
		return de::MovePtr<vk::wsi::Window>(display.createWindow(initialSize));
	}
	catch (const tcu::NotSupportedError& e)
	{
		// See createDisplay - assuming that wsi::Display was supported platform port
		// should also support creating a window.
		throw tcu::TestError(e.getMessage());
	}
}

bool wsiTypeSupportsScaling (vk::wsi::Type wsiType)
{
	return vk::wsi::getPlatformProperties(wsiType).swapchainExtent == vk::wsi::PlatformProperties::SWAPCHAIN_EXTENT_SCALED_TO_WINDOW_SIZE;
}

void initSemaphores (const vk::DeviceInterface&		vkd,
					 vk::VkDevice					device,
					 std::vector<vk::VkSemaphore>&	semaphores)
{
	for (size_t ndx = 0; ndx < semaphores.size(); ndx++)
		semaphores[ndx] = createSemaphore(vkd, device).disown();
}

void deinitSemaphores (const vk::DeviceInterface&	vkd,
					 vk::VkDevice					device,
					 std::vector<vk::VkSemaphore>&	semaphores)
{
	for (size_t ndx = 0; ndx < semaphores.size(); ndx++)
	{
		if (semaphores[ndx] != (vk::VkSemaphore)0)
			vkd.destroySemaphore(device, semaphores[ndx], DE_NULL);

		semaphores[ndx] = (vk::VkSemaphore)0;
	}

	semaphores.clear();
}

void initFences (const vk::DeviceInterface&	vkd,
				 vk::VkDevice				device,
				 std::vector<vk::VkFence>&	fences)
{
	for (size_t ndx = 0; ndx < fences.size(); ndx++)
		fences[ndx] = createFence(vkd, device).disown();
}

void deinitFences (const vk::DeviceInterface&	vkd,
				   vk::VkDevice					device,
				   std::vector<vk::VkFence>&	fences)
{
	for (size_t ndx = 0; ndx < fences.size(); ndx++)
	{
		if (fences[ndx] != (vk::VkFence)0)
			vkd.destroyFence(device, fences[ndx], DE_NULL);

		fences[ndx] = (vk::VkFence)0;
	}

	fences.clear();
}

void cmdRenderFrame (const vk::DeviceInterface&	vkd,
					 vk::VkCommandBuffer		commandBuffer,
					 vk::VkPipelineLayout		pipelineLayout,
					 vk::VkPipeline				pipeline,
					 size_t						frameNdx,
					 deUint32					quadCount)
{
	const deUint32 frameNdxValue  =	(deUint32)frameNdx;

	vkd.cmdPushConstants(commandBuffer, pipelineLayout, vk::VK_SHADER_STAGE_FRAGMENT_BIT, 0u, 4u, &frameNdxValue);
	vkd.cmdBindPipeline(commandBuffer, vk::VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline);
	vkd.cmdDraw(commandBuffer, quadCount * 6u, 1u, 0u, 0u);
}

vk::Move<vk::VkCommandBuffer> createCommandBuffer (const vk::DeviceInterface&	vkd,
												   vk::VkDevice					device,
												   vk::VkCommandPool			commandPool,
												   vk::VkPipelineLayout			pipelineLayout,
												   vk::VkRenderPass				renderPass,
												   vk::VkFramebuffer			framebuffer,
												   vk::VkPipeline				pipeline,
												   size_t						frameNdx,
												   deUint32						quadCount,
												   deUint32						imageWidth,
												   deUint32						imageHeight)
{
	const vk::VkCommandBufferAllocateInfo allocateInfo =
	{
		vk::VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO,
		DE_NULL,

		commandPool,
		vk::VK_COMMAND_BUFFER_LEVEL_PRIMARY,
		1
	};

	vk::Move<vk::VkCommandBuffer>	commandBuffer	(vk::allocateCommandBuffer(vkd, device, &allocateInfo));
	beginCommandBuffer(vkd, *commandBuffer, 0u);

	beginRenderPass(vkd, *commandBuffer, renderPass, framebuffer, vk::makeRect2D(0, 0, imageWidth, imageHeight), tcu::Vec4(0.25f, 0.5f, 0.75f, 1.0f));

	cmdRenderFrame(vkd, *commandBuffer, pipelineLayout, pipeline, frameNdx, quadCount);

	endRenderPass(vkd, *commandBuffer);

	endCommandBuffer(vkd, *commandBuffer);
	return commandBuffer;
}

void deinitCommandBuffers (const vk::DeviceInterface&			vkd,
						   vk::VkDevice							device,
						   vk::VkCommandPool					commandPool,
						   std::vector<vk::VkCommandBuffer>&	commandBuffers)
{
	for (size_t ndx = 0; ndx < commandBuffers.size(); ndx++)
	{
		if (commandBuffers[ndx] != (vk::VkCommandBuffer)0)
			vkd.freeCommandBuffers(device, commandPool, 1u,  &commandBuffers[ndx]);

		commandBuffers[ndx] = (vk::VkCommandBuffer)0;
	}

	commandBuffers.clear();
}

vk::Move<vk::VkCommandPool> createCommandPool (const vk::DeviceInterface&	vkd,
											   vk::VkDevice					device,
											   deUint32						queueFamilyIndex)
{
	const vk::VkCommandPoolCreateInfo createInfo =
	{
		vk::VK_STRUCTURE_TYPE_COMMAND_POOL_CREATE_INFO,
		DE_NULL,
		0u,
		queueFamilyIndex
	};

	return vk::createCommandPool(vkd, device, &createInfo);
}

vk::Move<vk::VkFramebuffer>	createFramebuffer (const vk::DeviceInterface&	vkd,
											   vk::VkDevice					device,
											   vk::VkRenderPass				renderPass,
											   vk::VkImageView				imageView,
											   deUint32						width,
											   deUint32						height)
{
	const vk::VkFramebufferCreateInfo createInfo =
	{
		vk::VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO,
		DE_NULL,

		0u,
		renderPass,
		1u,
		&imageView,
		width,
		height,
		1u
	};

	return vk::createFramebuffer(vkd, device, &createInfo);
}

vk::Move<vk::VkImageView> createImageView (const vk::DeviceInterface&	vkd,
										   vk::VkDevice					device,
										   vk::VkImage					image,
										   vk::VkFormat					format)
{
	const vk::VkImageViewCreateInfo	createInfo =
	{
		vk::VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO,
		DE_NULL,

		0u,
		image,
		vk::VK_IMAGE_VIEW_TYPE_2D,
		format,
		vk::makeComponentMappingRGBA(),
		{
			vk::VK_IMAGE_ASPECT_COLOR_BIT,
			0u,
			1u,
			0u,
			1u
		}
	};

	return vk::createImageView(vkd, device, &createInfo, DE_NULL);
}

vk::Move<vk::VkRenderPass> createRenderPass (const vk::DeviceInterface&	vkd,
											 vk::VkDevice				device,
											 vk::VkFormat				format)
{
	const vk::VkAttachmentDescription	attachments[]			=
	{
		{
			0u,
			format,
			vk::VK_SAMPLE_COUNT_1_BIT,

			vk::VK_ATTACHMENT_LOAD_OP_LOAD,
			vk::VK_ATTACHMENT_STORE_OP_STORE,

			vk::VK_ATTACHMENT_LOAD_OP_DONT_CARE,
			vk::VK_ATTACHMENT_STORE_OP_DONT_CARE,

			// This differs from the usual layout handling in that the
			// swapchain image remains in IMAGE_LAYOUT_SHARED_PRESENT_KHR all
			// the time. We should not ever transition it away (or discard the
			// contents with a transition from UNDEFINED) as the PE is accessing
			// the image concurrently with our rendering.
			vk::VK_IMAGE_LAYOUT_SHARED_PRESENT_KHR,
			vk::VK_IMAGE_LAYOUT_SHARED_PRESENT_KHR
		}
	};
	const vk::VkAttachmentReference		colorAttachmentRefs[]	=
	{
		{
			0u,
			vk::VK_IMAGE_LAYOUT_SHARED_PRESENT_KHR
		}
	};
	const vk::VkSubpassDescription		subpasses[]				=
	{
		{
			0u,
			vk::VK_PIPELINE_BIND_POINT_GRAPHICS,
			0u,
			DE_NULL,

			DE_LENGTH_OF_ARRAY(colorAttachmentRefs),
			colorAttachmentRefs,
			DE_NULL,

			DE_NULL,
			0u,
			DE_NULL
		}
	};

	const vk::VkRenderPassCreateInfo	createInfo				=
	{
		vk::VK_STRUCTURE_TYPE_RENDER_PASS_CREATE_INFO,
		DE_NULL,
		0u,

		DE_LENGTH_OF_ARRAY(attachments),
		attachments,

		DE_LENGTH_OF_ARRAY(subpasses),
		subpasses,

		0u,
		DE_NULL
	};

	return vk::createRenderPass(vkd, device, &createInfo);
}

vk::Move<vk::VkPipeline> createPipeline (const vk::DeviceInterface&	vkd,
										 vk::VkDevice				device,
										 vk::VkRenderPass			renderPass,
										 vk::VkPipelineLayout		layout,
										 vk::VkShaderModule			vertexShaderModule,
										 vk::VkShaderModule			fragmentShaderModule,
										 deUint32					width,
										 deUint32					height)
{
	const vk::VkPipelineVertexInputStateCreateInfo	vertexInputState	=
	{
		vk::VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO,
		DE_NULL,
		0u,
		0u,
		DE_NULL,
		0u,
		DE_NULL
	};
	const std::vector<vk::VkViewport>				viewports			(1, vk::makeViewport(tcu::UVec2(width, height)));
	const std::vector<vk::VkRect2D>					scissors			(1, vk::makeRect2D(tcu::UVec2(width, height)));

	return vk::makeGraphicsPipeline(vkd,										// const DeviceInterface&                        vk
									device,										// const VkDevice                                device
									layout,										// const VkPipelineLayout                        pipelineLayout
									vertexShaderModule,							// const VkShaderModule                          vertexShaderModule
									DE_NULL,									// const VkShaderModule                          tessellationControlShaderModule
									DE_NULL,									// const VkShaderModule                          tessellationEvalShaderModule
									DE_NULL,									// const VkShaderModule                          geometryShaderModule
									fragmentShaderModule,						// const VkShaderModule                          fragmentShaderModule
									renderPass,									// const VkRenderPass                            renderPass
									viewports,									// const std::vector<VkViewport>&                viewports
									scissors,									// const std::vector<VkRect2D>&                  scissors
									vk::VK_PRIMITIVE_TOPOLOGY_TRIANGLE_LIST,	// const VkPrimitiveTopology                     topology
									0u,											// const deUint32                                subpass
									0u,											// const deUint32                                patchControlPoints
									&vertexInputState);							// const VkPipelineVertexInputStateCreateInfo*   vertexInputStateCreateInfo
}

vk::Move<vk::VkPipelineLayout> createPipelineLayout (const vk::DeviceInterface&	vkd,
													 vk::VkDevice				device)
{
	const vk::VkPushConstantRange			pushConstants[] =
	{
		{
			vk::VK_SHADER_STAGE_FRAGMENT_BIT,
			0u,
			4u
		}
	};
	const vk::VkPipelineLayoutCreateInfo	createInfo	=
	{
		vk::VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO,
		DE_NULL,
		0u,

		0u,
		DE_NULL,

		DE_LENGTH_OF_ARRAY(pushConstants),
		pushConstants,
	};

	return vk::createPipelineLayout(vkd, device, &createInfo);
}

struct TestConfig
{
	vk::wsi::Type					wsiType;
	Scaling							scaling;
	bool							useSharedPresentableImage;
	vk::VkPresentModeKHR			presentMode;
	vk::VkSurfaceTransformFlagsKHR	transform;
	vk::VkCompositeAlphaFlagsKHR	alpha;
};

class SharedPresentableImageTestInstance : public TestInstance
{
public:
													SharedPresentableImageTestInstance	(Context& context, const TestConfig& testConfig);
													~SharedPresentableImageTestInstance	(void);

	tcu::TestStatus									iterate								(void);

private:
	const TestConfig								m_testConfig;
	const deUint32									m_quadCount;
	const vk::PlatformInterface&					m_vkp;
	const Extensions								m_instanceExtensions;
	const CustomInstance							m_instance;
	const vk::InstanceDriver&						m_vki;
	const vk::VkPhysicalDevice						m_physicalDevice;
	const de::UniquePtr<vk::wsi::Display>			m_nativeDisplay;
	const de::UniquePtr<vk::wsi::Window>			m_nativeWindow;
	const vk::Unique<vk::VkSurfaceKHR>				m_surface;

	const deUint32									m_queueFamilyIndex;
	const Extensions								m_deviceExtensions;
	const vk::Unique<vk::VkDevice>					m_device;
	const vk::DeviceDriver							m_vkd;
	const vk::VkQueue								m_queue;

	const vk::Unique<vk::VkCommandPool>				m_commandPool;
	const vk::Unique<vk::VkShaderModule>			m_vertexShaderModule;
	const vk::Unique<vk::VkShaderModule>			m_fragmentShaderModule;
	const vk::Unique<vk::VkPipelineLayout>			m_pipelineLayout;

	vk::VkImageUsageFlags							m_supportedUsageFlags;
	const vk::VkSurfaceCapabilitiesKHR				m_surfaceProperties;
	const vector<vk::VkSurfaceFormatKHR>			m_surfaceFormats;
	const vector<vk::VkPresentModeKHR>				m_presentModes;

	tcu::ResultCollector							m_resultCollector;

	vk::Move<vk::VkSwapchainKHR>					m_swapchain;
	vk::VkImage										m_swapchainImage;			  // NOTE: not owning. lifetime managed by swapchain
	vk::Move<vk::VkImageView>						m_swapchainImageView;
	vk::Move<vk::VkFramebuffer>						m_framebuffer;

	vk::Move<vk::VkRenderPass>						m_renderPass;
	vk::Move<vk::VkPipeline>						m_pipeline;

	std::vector<vk::VkCommandBuffer>				m_commandBuffers;
	std::vector<vk::VkSemaphore>					m_renderSemaphores;
	std::vector<vk::VkFence>						m_fences;

	std::vector<vk::VkSwapchainCreateInfoKHR>		m_swapchainConfigs;
	size_t											m_swapchainConfigNdx;

	const size_t									m_frameCount;
	size_t											m_frameNdx;

	const size_t									m_maxOutOfDateCount;
	size_t											m_outOfDateCount;

	void											initSwapchainResources		(void);
	void											deinitSwapchainResources	(void);
	void											render						(void);
};

std::vector<vk::VkSwapchainCreateInfoKHR> generateSwapchainConfigs (const vk::InstanceDriver&				vki,
																	const vk::VkPhysicalDevice				physicalDevice,
																	vk::VkSurfaceKHR						surface,
																	deUint32								queueFamilyIndex,
																	Scaling									scaling,
																	const vk::VkSurfaceCapabilitiesKHR&		properties,
																	const vector<vk::VkSurfaceFormatKHR>&	formats,
																	const vector<vk::VkPresentModeKHR>&		presentModes,
																	vk::VkPresentModeKHR					presentMode,
																	vk::VkImageUsageFlags					supportedImageUsage,
																	const vk::VkSurfaceTransformFlagsKHR	transform,
																	const vk::VkCompositeAlphaFlagsKHR		alpha)
{
	const deUint32							imageLayers			= 1u;
	const vk::VkImageUsageFlags				imageUsage			= properties.supportedUsageFlags & supportedImageUsage;
	const vk::VkBool32						clipped				= VK_FALSE;
	vector<vk::VkSwapchainCreateInfoKHR>	createInfos;

	const deUint32				currentWidth		= properties.currentExtent.width != 0xFFFFFFFFu
												? properties.currentExtent.width
												: de::min(1024u, properties.minImageExtent.width + ((properties.maxImageExtent.width - properties.minImageExtent.width) / 2));
	const deUint32				currentHeight		= properties.currentExtent.height != 0xFFFFFFFFu
												? properties.currentExtent.height
												: de::min(1024u, properties.minImageExtent.height + ((properties.maxImageExtent.height - properties.minImageExtent.height) / 2));

	const deUint32				imageWidth		= scaling == SCALING_NONE
												? currentWidth
												: (scaling == SCALING_UP
													? de::max(31u, properties.minImageExtent.width)
													: de::min(deSmallestGreaterOrEquallPowerOfTwoU32(currentWidth+1), properties.maxImageExtent.width));
	const deUint32				imageHeight		= scaling == SCALING_NONE
												? currentHeight
												: (scaling == SCALING_UP
													? de::max(31u, properties.minImageExtent.height)
													: de::min(deSmallestGreaterOrEquallPowerOfTwoU32(currentHeight+1), properties.maxImageExtent.height));
	const vk::VkExtent2D		imageSize		= { imageWidth, imageHeight };

	{
		size_t presentModeNdx;

		for (presentModeNdx = 0; presentModeNdx < presentModes.size(); presentModeNdx++)
		{
			if (presentModes[presentModeNdx] == presentMode)
				break;
		}

		if (presentModeNdx == presentModes.size())
			TCU_THROW(NotSupportedError, "Present mode not supported");

		if ((properties.supportedTransforms & transform) == 0)
			TCU_THROW(NotSupportedError, "Transform not supported");

		if ((properties.supportedCompositeAlpha & alpha) == 0)
			TCU_THROW(NotSupportedError, "Composite alpha not supported");
	}

	for (size_t formatNdx = 0; formatNdx < formats.size(); formatNdx++)
	{
		const vk::VkSurfaceTransformFlagBitsKHR	preTransform	= (vk::VkSurfaceTransformFlagBitsKHR)transform;
		const vk::VkCompositeAlphaFlagBitsKHR	compositeAlpha	= (vk::VkCompositeAlphaFlagBitsKHR)alpha;
		const vk::VkFormat						imageFormat		= formats[formatNdx].format;
		const vk::VkColorSpaceKHR				imageColorSpace	= formats[formatNdx].colorSpace;
		const vk::VkSwapchainCreateInfoKHR		createInfo		=
		{
			vk::VK_STRUCTURE_TYPE_SWAPCHAIN_CREATE_INFO_KHR,
			DE_NULL,
			0u,
			surface,
			1,														// Always 1 image for a shared presentable image swapchain.
			imageFormat,
			imageColorSpace,
			imageSize,
			imageLayers,
			imageUsage,
			vk::VK_SHARING_MODE_EXCLUSIVE,
			1u,
			&queueFamilyIndex,
			preTransform,
			compositeAlpha,
			presentMode,
			clipped,
			(vk::VkSwapchainKHR)0
		};

		{
			vk::VkImageFormatProperties			imageFormatProperties;

			deMemset(&imageFormatProperties, 0, sizeof(imageFormatProperties));

			vk::VkResult result = vki.getPhysicalDeviceImageFormatProperties(physicalDevice, imageFormat, vk::VK_IMAGE_TYPE_2D, vk::VK_IMAGE_TILING_OPTIMAL, imageUsage, 0, &imageFormatProperties);

			if (result == vk::VK_ERROR_FORMAT_NOT_SUPPORTED)
				continue;
		}

		createInfos.push_back(createInfo);
	}

	return createInfos;
}

vk::VkSurfaceCapabilitiesKHR getPhysicalDeviceSurfaceCapabilities (const vk::InstanceInterface&	vki,
																   vk::VkPhysicalDevice			physicalDevice,
																   vk::VkSurfaceKHR				surface,
																   vk::VkImageUsageFlags*		usage)
{
	const vk::VkPhysicalDeviceSurfaceInfo2KHR	info	=
	{
		vk::VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SURFACE_INFO_2_KHR,
		DE_NULL,

		surface
	};
	vk::VkSharedPresentSurfaceCapabilitiesKHR	sharedCapabilities;
	vk::VkSurfaceCapabilities2KHR				capabilities;

	sharedCapabilities.sType	= vk::VK_STRUCTURE_TYPE_SHARED_PRESENT_SURFACE_CAPABILITIES_KHR;
	sharedCapabilities.pNext	= DE_NULL;

	capabilities.sType			= vk::VK_STRUCTURE_TYPE_SURFACE_CAPABILITIES_2_KHR;
	capabilities.pNext			= &sharedCapabilities;

	VK_CHECK(vki.getPhysicalDeviceSurfaceCapabilities2KHR(physicalDevice, &info, &capabilities));

	TCU_CHECK(sharedCapabilities.sharedPresentSupportedUsageFlags & vk::VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT);
	*usage = sharedCapabilities.sharedPresentSupportedUsageFlags;

	return capabilities.surfaceCapabilities;
}

SharedPresentableImageTestInstance::SharedPresentableImageTestInstance (Context& context, const TestConfig& testConfig)
	: TestInstance				(context)
	, m_testConfig				(testConfig)
	, m_quadCount				(16u)
	, m_vkp						(context.getPlatformInterface())
	, m_instanceExtensions		(vk::enumerateInstanceExtensionProperties(m_vkp, DE_NULL))
	, m_instance				(createInstanceWithWsi(context, m_instanceExtensions, testConfig.wsiType))
	, m_vki						(m_instance.getDriver())
	, m_physicalDevice			(vk::chooseDevice(m_vki, m_instance, context.getTestContext().getCommandLine()))
	, m_nativeDisplay			(createDisplay(context.getTestContext().getPlatform().getVulkanPlatform(), m_instanceExtensions, testConfig.wsiType))
	, m_nativeWindow			(createWindow(*m_nativeDisplay, tcu::Nothing))
	, m_surface					(vk::wsi::createSurface(m_vki, m_instance, testConfig.wsiType, *m_nativeDisplay, *m_nativeWindow, context.getTestContext().getCommandLine()))

	, m_queueFamilyIndex		(vk::wsi::chooseQueueFamilyIndex(m_vki, m_physicalDevice, *m_surface))
	, m_deviceExtensions		(vk::enumerateDeviceExtensionProperties(m_vki, m_physicalDevice, DE_NULL))
	, m_device					(createDeviceWithWsi(m_vkp, m_instance, m_vki, m_physicalDevice, m_deviceExtensions, m_queueFamilyIndex, testConfig.useSharedPresentableImage, context.getTestContext().getCommandLine().isValidationEnabled()))
	, m_vkd						(m_vkp, m_instance, *m_device)
	, m_queue					(getDeviceQueue(m_vkd, *m_device, m_queueFamilyIndex, 0u))

	, m_commandPool				(createCommandPool(m_vkd, *m_device, m_queueFamilyIndex))
	, m_vertexShaderModule		(vk::createShaderModule(m_vkd, *m_device, context.getBinaryCollection().get("quad-vert"), 0u))
	, m_fragmentShaderModule	(vk::createShaderModule(m_vkd, *m_device, context.getBinaryCollection().get("quad-frag"), 0u))
	, m_pipelineLayout			(createPipelineLayout(m_vkd, *m_device))

	, m_supportedUsageFlags		(0u)
	, m_surfaceProperties		(getPhysicalDeviceSurfaceCapabilities(m_vki, m_physicalDevice, *m_surface, &m_supportedUsageFlags))
	, m_surfaceFormats			(vk::wsi::getPhysicalDeviceSurfaceFormats(m_vki, m_physicalDevice, *m_surface))
	, m_presentModes			(vk::wsi::getPhysicalDeviceSurfacePresentModes(m_vki, m_physicalDevice, *m_surface))

	, m_swapchainConfigs		(generateSwapchainConfigs(m_vki, m_physicalDevice, *m_surface, m_queueFamilyIndex, testConfig.scaling, m_surfaceProperties, m_surfaceFormats, m_presentModes, testConfig.presentMode, m_supportedUsageFlags, testConfig.transform, testConfig.alpha))
	, m_swapchainConfigNdx		(0u)

	, m_frameCount				(60u * 5u)
	, m_frameNdx				(0u)

	, m_maxOutOfDateCount		(20u)
	, m_outOfDateCount			(0u)
{
	{
		const tcu::ScopedLogSection surfaceInfo (m_context.getTestContext().getLog(), "SurfaceCapabilities", "SurfaceCapabilities");
		m_context.getTestContext().getLog() << TestLog::Message << m_surfaceProperties << TestLog::EndMessage;
		m_context.getTestContext().getLog() << TestLog::Message << "SharedPresentSupportedUsageFlags: " << m_supportedUsageFlags << TestLog::EndMessage;

	}
}

SharedPresentableImageTestInstance::~SharedPresentableImageTestInstance (void)
{
	deinitSwapchainResources();
}

void SharedPresentableImageTestInstance::initSwapchainResources (void)
{
	const size_t		fenceCount	= 6;
	const deUint32		imageWidth	= m_swapchainConfigs[m_swapchainConfigNdx].imageExtent.width;
	const deUint32		imageHeight	= m_swapchainConfigs[m_swapchainConfigNdx].imageExtent.height;
	const vk::VkFormat	imageFormat	= m_swapchainConfigs[m_swapchainConfigNdx].imageFormat;

	m_swapchain				= vk::createSwapchainKHR(m_vkd, *m_device, &m_swapchainConfigs[m_swapchainConfigNdx]);
	m_swapchainImage		= vk::wsi::getSwapchainImages(m_vkd, *m_device, *m_swapchain).front();

	m_renderPass			= createRenderPass(m_vkd, *m_device, imageFormat);
	m_pipeline				= createPipeline(m_vkd, *m_device, *m_renderPass, *m_pipelineLayout, *m_vertexShaderModule, *m_fragmentShaderModule, imageWidth, imageHeight);

	m_swapchainImageView	= createImageView(m_vkd, *m_device, m_swapchainImage, imageFormat);
	m_framebuffer			= createFramebuffer(m_vkd, *m_device, *m_renderPass, *m_swapchainImageView, imageWidth, imageHeight);

	m_renderSemaphores		= std::vector<vk::VkSemaphore>(fenceCount, (vk::VkSemaphore)0);
	m_fences				= std::vector<vk::VkFence>(fenceCount, (vk::VkFence)0);
	m_commandBuffers		= std::vector<vk::VkCommandBuffer>(m_fences.size(), (vk::VkCommandBuffer)0);

	initSemaphores(m_vkd, *m_device, m_renderSemaphores);

	initFences(m_vkd, *m_device, m_fences);

	// Unlike a traditional swapchain, where we'd acquire a new image from the
	// PE every frame, a shared image swapchain has a single image that is
	// acquired upfront. We acquire it here, transition it to the proper layout,
	// and present it.

	// Acquire the one image
	const deUint64				foreverNs		= 0xFFFFFFFFFFFFFFFFul;
	vk::Move<vk::VkSemaphore>	semaphore(createSemaphore(m_vkd, *m_device));
	deUint32					imageIndex		= 42;	// initialize to junk value

	VK_CHECK(m_vkd.acquireNextImageKHR(*m_device, *m_swapchain, foreverNs, *semaphore, 0u, &imageIndex));
	TCU_CHECK(imageIndex == 0);

	// Transition to IMAGE_LAYOUT_SHARED_PRESENT_KHR
	const vk::VkCommandBufferAllocateInfo allocateInfo =
	{
		vk::VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO,
		DE_NULL,
		*m_commandPool,
		vk::VK_COMMAND_BUFFER_LEVEL_PRIMARY,
		1
	};

	const vk::Unique<vk::VkCommandBuffer>	commandBuffer	(vk::allocateCommandBuffer(m_vkd, *m_device, &allocateInfo));
	beginCommandBuffer(m_vkd, *commandBuffer, 0u);

	const vk::VkImageMemoryBarrier barrier = {
		vk::VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
		DE_NULL,
		0,
		0,
		vk::VK_IMAGE_LAYOUT_UNDEFINED,
		vk::VK_IMAGE_LAYOUT_SHARED_PRESENT_KHR,
		VK_QUEUE_FAMILY_IGNORED,
		VK_QUEUE_FAMILY_IGNORED,
		m_swapchainImage,
		{
			vk::VK_IMAGE_ASPECT_COLOR_BIT,
			0,
			1,
			0,
			1
		},
	};

	m_vkd.cmdPipelineBarrier(*commandBuffer,
							vk::VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT,
							vk::VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT,
							0u,
							0, DE_NULL,
							0, DE_NULL,
							1, &barrier);

	endCommandBuffer(m_vkd, *commandBuffer);

	const vk::VkPipelineStageFlags waitDstStages[] = { vk::VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT };
	const vk::VkSubmitInfo submitInfo =
	{
		vk::VK_STRUCTURE_TYPE_SUBMIT_INFO,
		DE_NULL,
		1, &*semaphore, waitDstStages,
		1, &*commandBuffer,
		0, DE_NULL,
	};

	VK_CHECK(m_vkd.queueSubmit(m_queue, 1u, &submitInfo, (vk::VkFence)0));
	VK_CHECK(m_vkd.queueWaitIdle(m_queue));
}

void SharedPresentableImageTestInstance::deinitSwapchainResources (void)
{
	VK_CHECK(m_vkd.queueWaitIdle(m_queue));

	deinitSemaphores(m_vkd, *m_device, m_renderSemaphores);
	deinitFences(m_vkd, *m_device, m_fences);
	deinitCommandBuffers(m_vkd, *m_device, *m_commandPool, m_commandBuffers);

	m_framebuffer	= vk::Move<vk::VkFramebuffer>();
	m_swapchainImageView = vk::Move<vk::VkImageView>();
	m_swapchainImage = (vk::VkImage)0;

	m_swapchain		= vk::Move<vk::VkSwapchainKHR>();
	m_renderPass	= vk::Move<vk::VkRenderPass>();
	m_pipeline		= vk::Move<vk::VkPipeline>();
}

void SharedPresentableImageTestInstance::render (void)
{
	const deUint64		foreverNs		= 0xFFFFFFFFFFFFFFFFul;
	const vk::VkFence	fence			= m_fences[m_frameNdx % m_fences.size()];
	const deUint32		width			= m_swapchainConfigs[m_swapchainConfigNdx].imageExtent.width;
	const deUint32		height			= m_swapchainConfigs[m_swapchainConfigNdx].imageExtent.height;

	// Throttle execution
	if (m_frameNdx >= m_fences.size())
	{
		VK_CHECK(m_vkd.waitForFences(*m_device, 1u, &fence, VK_TRUE, foreverNs));
		VK_CHECK(m_vkd.resetFences(*m_device, 1u, &fence));

		m_vkd.freeCommandBuffers(*m_device, *m_commandPool, 1u, &m_commandBuffers[m_frameNdx % m_commandBuffers.size()]);
		m_commandBuffers[m_frameNdx % m_commandBuffers.size()] = (vk::VkCommandBuffer)0;
	}

	deUint32			imageIndex = 0;		// There is only one image.
	const vk::VkSemaphore currentRenderSemaphore = m_renderSemaphores[m_frameNdx % m_renderSemaphores.size()];

	const bool willPresent = m_swapchainConfigs[m_swapchainConfigNdx].presentMode == vk::VK_PRESENT_MODE_SHARED_DEMAND_REFRESH_KHR || !m_frameNdx;

	// Create command buffer
	m_commandBuffers[m_frameNdx % m_commandBuffers.size()] = createCommandBuffer(m_vkd, *m_device, *m_commandPool, *m_pipelineLayout, *m_renderPass, *m_framebuffer, *m_pipeline, m_frameNdx, m_quadCount, width, height).disown();

	// Submit command buffer
	{
		const vk::VkSubmitInfo			submitInfo		=
		{
			vk::VK_STRUCTURE_TYPE_SUBMIT_INFO,
			DE_NULL,
			0u,
			DE_NULL,
			DE_NULL,
			1u,
			&m_commandBuffers[m_frameNdx % m_commandBuffers.size()],
			willPresent ? 1u : 0u,	  // Only signal the semaphore if we're going to call QueuePresent.
			&currentRenderSemaphore
		};

		// With a traditional swapchain, we'd fence on completion of
		// AcquireNextImage. We never call that for a shared image swapchain, so
		// fence on completion of the rendering work instead. A real shared
		// image application would want a more substantial pacing mechanism.
		VK_CHECK(m_vkd.queueSubmit(m_queue, 1u, &submitInfo, fence));
	}

	// DEMAND_REFRESH requires us to call QueuePresent whenever we want to be
	// assured the PE has picked up a new frame. The PE /may/ also pick up
	// changes whenever it likes.
	//
	// For CONTINUOUS_REFRESH, we need to just call QueuePresent once on the
	// first frame to kick things off.
	if (willPresent)
	{

		// Present frame
		vk::VkResult result;
		const vk::VkPresentInfoKHR presentInfo =
		{
			vk::VK_STRUCTURE_TYPE_PRESENT_INFO_KHR,
			DE_NULL,
			1u,
			&currentRenderSemaphore,
			1u,
			&*m_swapchain,
			&imageIndex,
			&result
		};

		VK_CHECK_WSI(m_vkd.queuePresentKHR(m_queue, &presentInfo));
		VK_CHECK_WSI(result);

	}

	// With either present mode, we can call GetSwapchainStatus at any time
	// to detect possible OUT_OF_DATE conditions. Let's do that every frame.

	const vk::VkResult swapchainStatus = m_vkd.getSwapchainStatusKHR(*m_device, *m_swapchain);
	VK_CHECK(swapchainStatus);
}

tcu::TestStatus SharedPresentableImageTestInstance::iterate (void)
{
	// Initialize swapchain specific resources
	// Render test
	try
	{
		if (m_frameNdx == 0)
		{
			if (m_outOfDateCount == 0)
				m_context.getTestContext().getLog() << tcu::TestLog::Message << "Swapchain: " << m_swapchainConfigs[m_swapchainConfigNdx] << tcu::TestLog::EndMessage;

			initSwapchainResources();
		}

		render();
	}
	catch (const vk::Error& error)
	{
		if (error.getError() == vk::VK_ERROR_OUT_OF_DATE_KHR)
		{
			m_swapchainConfigs = generateSwapchainConfigs(m_vki, m_physicalDevice, *m_surface, m_queueFamilyIndex, m_testConfig.scaling, m_surfaceProperties, m_surfaceFormats, m_presentModes, m_testConfig.presentMode, m_supportedUsageFlags, m_testConfig.transform, m_testConfig.alpha);

			if (m_outOfDateCount < m_maxOutOfDateCount)
			{
				m_context.getTestContext().getLog() << TestLog::Message << "Frame " << m_frameNdx << ": Swapchain out of date. Recreating resources." << TestLog::EndMessage;
				deinitSwapchainResources();
				m_frameNdx = 0;
				m_outOfDateCount++;

				return tcu::TestStatus::incomplete();
			}
			else
			{
				m_context.getTestContext().getLog() << TestLog::Message << "Frame " << m_frameNdx << ": Swapchain out of date." << TestLog::EndMessage;
				m_resultCollector.fail("Received too many VK_ERROR_OUT_OF_DATE_KHR errors. Received " + de::toString(m_outOfDateCount) + ", max " + de::toString(m_maxOutOfDateCount));
			}
		}
		else
		{
			m_resultCollector.fail(error.what());
		}

		deinitSwapchainResources();

		m_swapchainConfigNdx++;
		m_frameNdx = 0;
		m_outOfDateCount = 0;

		if (m_swapchainConfigNdx >= m_swapchainConfigs.size())
			return tcu::TestStatus(m_resultCollector.getResult(), m_resultCollector.getMessage());
		else
			return tcu::TestStatus::incomplete();
	}

	m_frameNdx++;

	if (m_frameNdx >= m_frameCount)
	{
		m_frameNdx = 0;
		m_outOfDateCount = 0;
		m_swapchainConfigNdx++;

		deinitSwapchainResources();

		if (m_swapchainConfigNdx >= m_swapchainConfigs.size())
			return tcu::TestStatus(m_resultCollector.getResult(), m_resultCollector.getMessage());
		else
			return tcu::TestStatus::incomplete();
	}
	else
		return tcu::TestStatus::incomplete();
}

struct Programs
{
	static void init (vk::SourceCollections& dst, TestConfig)
	{
		dst.glslSources.add("quad-vert") << glu::VertexSource(
			"#version 450\n"
			"out gl_PerVertex {\n"
			"\tvec4 gl_Position;\n"
			"};\n"
			"layout(location = 0) out highp uint quadIndex;\n"
			"highp float;\n"
			"void main (void) {\n"
			"\tgl_Position = vec4(((gl_VertexIndex + 2) / 3) % 2 == 0 ? -1.0 : 1.0,\n"
			"\t                   ((gl_VertexIndex + 1) / 3) % 2 == 0 ? -1.0 : 1.0, 0.0, 1.0);\n"
			"\tquadIndex = gl_VertexIndex / 6;\n"
			"}\n");
		dst.glslSources.add("quad-frag") << glu::FragmentSource(
			"#version 310 es\n"
			"layout(location = 0) flat in highp uint quadIndex;\n"
			"layout(location = 0) out highp vec4 o_color;\n"
			"layout(push_constant) uniform PushConstant {\n"
			"\thighp uint frameNdx;\n"
			"} pushConstants;\n"
			"void main (void)\n"
			"{\n"
			"\thighp uint frameNdx = pushConstants.frameNdx;\n"
			"\thighp uint cellX = bitfieldExtract(uint(gl_FragCoord.x), 7, 10);\n"
			"\thighp uint cellY = bitfieldExtract(uint(gl_FragCoord.y), 7, 10);\n"
			"\thighp uint x = quadIndex ^ (frameNdx + (uint(gl_FragCoord.x) >> cellX));\n"
			"\thighp uint y = quadIndex ^ (frameNdx + (uint(gl_FragCoord.y) >> cellY));\n"
			"\thighp uint r = 128u * bitfieldExtract(x, 0, 1)\n"
			"\t             +  64u * bitfieldExtract(y, 1, 1)\n"
			"\t             +  32u * bitfieldExtract(x, 3, 1);\n"
			"\thighp uint g = 128u * bitfieldExtract(y, 0, 1)\n"
			"\t             +  64u * bitfieldExtract(x, 2, 1)\n"
			"\t             +  32u * bitfieldExtract(y, 3, 1);\n"
			"\thighp uint b = 128u * bitfieldExtract(x, 1, 1)\n"
			"\t             +  64u * bitfieldExtract(y, 2, 1)\n"
			"\t             +  32u * bitfieldExtract(x, 4, 1);\n"
			"\to_color = vec4(float(r) / 255.0, float(g) / 255.0, float(b) / 255.0, 1.0);\n"
			"}\n");
	}
};

} // anonymous

void createSharedPresentableImageTests (tcu::TestCaseGroup* testGroup, vk::wsi::Type wsiType)
{
	const struct
	{
		Scaling		scaling;
		const char*	name;
	} scaling [] =
	{
		{ SCALING_NONE,	"scale_none"	},
		{ SCALING_UP,	"scale_up"		},
		{ SCALING_DOWN, "scale_down"	}
	};
	const struct
	{
		vk::VkPresentModeKHR	mode;
		const char*				name;
	} presentModes[] =
	{
		{ vk::VK_PRESENT_MODE_SHARED_DEMAND_REFRESH_KHR,			"demand"			},
		{ vk::VK_PRESENT_MODE_SHARED_CONTINUOUS_REFRESH_KHR,		"continuous"		},
	};
	const struct
	{
		vk::VkSurfaceTransformFlagsKHR	transform;
		const char*						name;
	} transforms[] =
	{
		{ vk::VK_SURFACE_TRANSFORM_IDENTITY_BIT_KHR,						"identity"						},
		{ vk::VK_SURFACE_TRANSFORM_ROTATE_90_BIT_KHR,						"rotate_90"						},
		{ vk::VK_SURFACE_TRANSFORM_ROTATE_180_BIT_KHR,						"rotate_180"					},
		{ vk::VK_SURFACE_TRANSFORM_ROTATE_270_BIT_KHR,						"rotate_270"					},
		{ vk::VK_SURFACE_TRANSFORM_HORIZONTAL_MIRROR_BIT_KHR,				"horizontal_mirror"				},
		{ vk::VK_SURFACE_TRANSFORM_HORIZONTAL_MIRROR_ROTATE_90_BIT_KHR,		"horizontal_mirror_rotate_90"	},
		{ vk::VK_SURFACE_TRANSFORM_HORIZONTAL_MIRROR_ROTATE_180_BIT_KHR,	"horizontal_mirror_rotate_180"	},
		{ vk::VK_SURFACE_TRANSFORM_HORIZONTAL_MIRROR_ROTATE_270_BIT_KHR,	"horizontal_mirror_rotate_270"	},
		{ vk::VK_SURFACE_TRANSFORM_INHERIT_BIT_KHR,							"inherit"						}
	};
	const struct
	{
		vk::VkCompositeAlphaFlagsKHR	alpha;
		const char*						name;
	} alphas[] =
	{
		{ vk::VK_COMPOSITE_ALPHA_OPAQUE_BIT_KHR,			"opaque"			},
		{ vk::VK_COMPOSITE_ALPHA_PRE_MULTIPLIED_BIT_KHR,	"pre_multiplied"	},
		{ vk::VK_COMPOSITE_ALPHA_POST_MULTIPLIED_BIT_KHR,	"post_multiplied"	},
		{ vk::VK_COMPOSITE_ALPHA_INHERIT_BIT_KHR,			"inherit"			}
	};

	for (size_t scalingNdx = 0; scalingNdx < DE_LENGTH_OF_ARRAY(scaling); scalingNdx++)
	{
		if (scaling[scalingNdx].scaling == SCALING_NONE || wsiTypeSupportsScaling(wsiType))
		{
			de::MovePtr<tcu::TestCaseGroup>	scaleGroup	(new tcu::TestCaseGroup(testGroup->getTestContext(), scaling[scalingNdx].name, scaling[scalingNdx].name));

			for (size_t transformNdx = 0; transformNdx < DE_LENGTH_OF_ARRAY(transforms); transformNdx++)
			{
				de::MovePtr<tcu::TestCaseGroup>	transformGroup	(new tcu::TestCaseGroup(testGroup->getTestContext(), transforms[transformNdx].name, transforms[transformNdx].name));

				for (size_t alphaNdx = 0; alphaNdx < DE_LENGTH_OF_ARRAY(alphas); alphaNdx++)
				{
					de::MovePtr<tcu::TestCaseGroup>	alphaGroup	(new tcu::TestCaseGroup(testGroup->getTestContext(), alphas[alphaNdx].name, alphas[alphaNdx].name));

					for (size_t presentModeNdx = 0; presentModeNdx < DE_LENGTH_OF_ARRAY(presentModes); presentModeNdx++)
					{
						const char* const					name	= presentModes[presentModeNdx].name;
						TestConfig							config;

						config.wsiType						= wsiType;
						config.useSharedPresentableImage	= true;
						config.scaling						= scaling[scalingNdx].scaling;
						config.transform					= transforms[transformNdx].transform;
						config.alpha						= alphas[alphaNdx].alpha;
						config.presentMode					= presentModes[presentModeNdx].mode;

						alphaGroup->addChild(new vkt::InstanceFactory1<SharedPresentableImageTestInstance, TestConfig, Programs>(testGroup->getTestContext(), tcu::NODETYPE_SELF_VALIDATE, name, name, Programs(), config));
					}

					transformGroup->addChild(alphaGroup.release());
				}

				scaleGroup->addChild(transformGroup.release());
			}

			testGroup->addChild(scaleGroup.release());
		}
	}
}

} // wsi
} // vkt