xash3d-fwgs/ref_vk/vk_rtx.c

#include "vk_rtx.h"

#include "vk_core.h"
#include "vk_common.h"
#include "vk_render.h"
#include "vk_buffer.h"

#include "eiface.h"

#define MAX_ACCELS 1024
#define MAX_SCRATCH_BUFFER (16*1024*1024)
#define MAX_ACCELS_BUFFER (64*1024*1024)

// TODO sync with shaders
#define WG_W 16
#define WG_H 8

/*
typedef struct {
	//int lightmap, texture;
	//int render_mode;
	uint32_t element_count;
	uint32_t index_offset, vertex_offset;
	VkBuffer buffer;
} vk_ray_model_t;
*/

static struct {
	VkPipelineLayout pipeline_layout;
	VkPipeline pipeline;
	VkDescriptorSetLayout desc_layout;
	VkDescriptorPool desc_pool;
	VkDescriptorSet desc_set;

	vk_buffer_t accels_buffer;
	vk_buffer_t scratch_buffer;
	VkDeviceAddress accels_buffer_addr, scratch_buffer_addr;

	vk_buffer_t tlas_geom_buffer;

	VkAccelerationStructureKHR accels[MAX_ACCELS];
	VkAccelerationStructureKHR tlas;
} g_rtx;

static struct {
	int num_accels;
	uint32_t scratch_offset, buffer_offset;
} g_rtx_scene;

static VkDeviceAddress getBufferDeviceAddress(VkBuffer buffer) {
	const VkBufferDeviceAddressInfo bdai = {.sType = VK_STRUCTURE_TYPE_BUFFER_DEVICE_ADDRESS_INFO, .buffer = buffer};
	return vkGetBufferDeviceAddress(vk_core.device, &bdai);
}

static VkDeviceAddress getASAddress(VkAccelerationStructureKHR as) {
	VkAccelerationStructureDeviceAddressInfoKHR asdai = {
		.sType = VK_STRUCTURE_TYPE_ACCELERATION_STRUCTURE_DEVICE_ADDRESS_INFO_KHR,
		.accelerationStructure = as,
	};
	return vkGetAccelerationStructureDeviceAddressKHR(vk_core.device, &asdai);
}

static VkAccelerationStructureKHR createAndBuildAccelerationStructure(VkCommandBuffer cmdbuf, const VkAccelerationStructureGeometryKHR *geoms, const uint32_t *max_prim_counts, const VkAccelerationStructureBuildRangeInfoKHR **build_ranges, uint32_t n_geoms, VkAccelerationStructureTypeKHR type) {
	VkAccelerationStructureKHR accel;

	VkAccelerationStructureBuildGeometryInfoKHR build_info = {
		.sType = VK_STRUCTURE_TYPE_ACCELERATION_STRUCTURE_BUILD_GEOMETRY_INFO_KHR,
		.type = type,
		.flags = VK_BUILD_ACCELERATION_STRUCTURE_PREFER_FAST_TRACE_BIT_KHR | VK_BUILD_ACCELERATION_STRUCTURE_ALLOW_UPDATE_BIT_KHR,
		.mode = VK_BUILD_ACCELERATION_STRUCTURE_MODE_BUILD_KHR,
		.geometryCount = n_geoms,
		.pGeometries = geoms,
	};

	VkAccelerationStructureBuildSizesInfoKHR build_size = {
		.sType = VK_STRUCTURE_TYPE_ACCELERATION_STRUCTURE_BUILD_SIZES_INFO_KHR};

	VkAccelerationStructureCreateInfoKHR asci = {
		.sType = VK_STRUCTURE_TYPE_ACCELERATION_STRUCTURE_CREATE_INFO_KHR,
		.buffer = g_rtx.accels_buffer.buffer,
		.offset = g_rtx_scene.buffer_offset,
		.type = type,
	};

	vkGetAccelerationStructureBuildSizesKHR(
		vk_core.device, VK_ACCELERATION_STRUCTURE_BUILD_TYPE_DEVICE_KHR, &build_info, max_prim_counts, &build_size);

	if (0)
	{
		uint32_t max_prims = 0;
		for (int i = 0; i < n_geoms; ++i)
			max_prims += max_prim_counts[i];
		gEngine.Con_Reportf(
			"AS max_prims=%u, n_geoms=%u, build size: %d, scratch size: %d\n", max_prims, n_geoms, build_size.accelerationStructureSize, build_size.buildScratchSize);
	}

	if (MAX_SCRATCH_BUFFER - g_rtx_scene.scratch_offset < build_size.buildScratchSize) {
		gEngine.Con_Printf(S_ERROR "Scratch buffer overflow: left %u bytes, but need %u\n",
			MAX_SCRATCH_BUFFER - g_rtx_scene.scratch_offset,
			build_size.buildScratchSize);
		return VK_NULL_HANDLE;
	}

	if (MAX_ACCELS_BUFFER - g_rtx_scene.buffer_offset < build_size.accelerationStructureSize) {
		gEngine.Con_Printf(S_ERROR "Accels buffer overflow: left %u bytes, but need %u\n",
			MAX_ACCELS_BUFFER - g_rtx_scene.buffer_offset,
			build_size.accelerationStructureSize);
		return VK_NULL_HANDLE;
	}

	asci.size = build_size.accelerationStructureSize;
	XVK_CHECK(vkCreateAccelerationStructureKHR(vk_core.device, &asci, NULL, &accel));

	g_rtx_scene.buffer_offset += build_size.accelerationStructureSize;
	g_rtx_scene.buffer_offset = (g_rtx_scene.buffer_offset + 255) & ~255; // Buffer must be aligned to 256 according to spec

	build_info.dstAccelerationStructure = accel;
	build_info.scratchData.deviceAddress = g_rtx.scratch_buffer_addr + g_rtx_scene.scratch_offset;
	g_rtx_scene.scratch_offset += build_size.buildScratchSize;

	vkCmdBuildAccelerationStructuresKHR(cmdbuf, 1, &build_info, build_ranges);
	return accel;
}

static void cleanupASFIXME(void)
{
	// FIXME we really should not do this; cache ASs per model
	for (int i = 0; i < g_rtx_scene.num_accels; ++i) {
		if (g_rtx.accels[i] != VK_NULL_HANDLE)
			vkDestroyAccelerationStructureKHR(vk_core.device, g_rtx.accels[i], NULL);
	}
	if (g_rtx.tlas != VK_NULL_HANDLE)
		vkDestroyAccelerationStructureKHR(vk_core.device, g_rtx.tlas, NULL);

	g_rtx_scene.num_accels = 0;
}

void VK_RaySceneBegin( void )
{
	ASSERT(vk_core.rtx);

	// FIXME this buffer might have objects that live longer
	g_rtx_scene.buffer_offset = 0;
	g_rtx_scene.scratch_offset = 0;

	cleanupASFIXME();
}

/*
static vk_ray_model_t *getModelByHandle(vk_ray_model_handle_t handle)
{
}
*/

void VK_RayScenePushModel( VkCommandBuffer cmdbuf, const vk_ray_model_create_t *model) // _handle_t model_handle )
{
	VkAccelerationStructureKHR *handle = g_rtx.accels + g_rtx_scene.num_accels;
	if (g_rtx_scene.num_accels > ARRAYSIZE(g_rtx.accels)) {
		gEngine.Con_Printf(S_ERROR "Ran out of AccelerationStructure slots\n");
		return;
	}

	ASSERT(vk_core.rtx);

	{
		//vk_ray_model_t *model = getModelByHandle(model_handle);
		const VkDeviceAddress buffer_addr = getBufferDeviceAddress(model->buffer);
		const uint32_t prim_count = model->element_count / 3;
		const VkAccelerationStructureGeometryKHR geom[] = {
			{
				.sType = VK_STRUCTURE_TYPE_ACCELERATION_STRUCTURE_GEOMETRY_KHR,
				.flags = VK_GEOMETRY_OPAQUE_BIT_KHR,
				.geometryType = VK_GEOMETRY_TYPE_TRIANGLES_KHR,
				.geometry.triangles =
					(VkAccelerationStructureGeometryTrianglesDataKHR){
						.sType = VK_STRUCTURE_TYPE_ACCELERATION_STRUCTURE_GEOMETRY_TRIANGLES_DATA_KHR,
						.indexType = model->index_offset == UINT32_MAX ? VK_INDEX_TYPE_NONE_KHR : VK_INDEX_TYPE_UINT16,
						.maxVertex = model->max_vertex,
						.vertexFormat = VK_FORMAT_R32G32B32_SFLOAT,
						.vertexStride = sizeof(vk_vertex_t),
						.vertexData.deviceAddress = buffer_addr + model->vertex_offset,
						.indexData.deviceAddress = buffer_addr + model->index_offset,
					},
			} };

		const uint32_t max_prim_counts[ARRAYSIZE(geom)] = { prim_count };
		const VkAccelerationStructureBuildRangeInfoKHR build_range_tri = {
			.primitiveCount = prim_count,
		};
		const VkAccelerationStructureBuildRangeInfoKHR* build_ranges[ARRAYSIZE(geom)] = { &build_range_tri };

		*handle = createAndBuildAccelerationStructure(cmdbuf,
			geom, max_prim_counts, build_ranges, ARRAYSIZE(geom), VK_ACCELERATION_STRUCTURE_TYPE_BOTTOM_LEVEL_KHR);

		if (!*handle) {
			gEngine.Con_Printf(S_ERROR "Error building BLAS\n");
			return;
		}

		g_rtx_scene.num_accels++;
	}
}

void VK_RaySceneEnd( VkCommandBuffer cmdbuf, VkImageView img_dst_view, VkImage img_dst, uint32_t w, uint32_t h )
{
	ASSERT(vk_core.rtx);

	// Upload all blas instances references to GPU mem
	{
		VkAccelerationStructureInstanceKHR *inst = g_rtx.tlas_geom_buffer.mapped;
		for (int i = 0; i < g_rtx_scene.num_accels; ++i) {
			inst[i] = (VkAccelerationStructureInstanceKHR){
				.transform = (VkTransformMatrixKHR){1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0},
				.instanceCustomIndex = 0,
				.mask = 0xff,
				.instanceShaderBindingTableRecordOffset = 0,
				.flags = 0,
				.accelerationStructureReference = getASAddress(g_rtx.accels[i]),
			};
		}
	}

	// Barrier for building all BLASes
	// BLAS building is now in cmdbuf, need to synchronize with results
	{
		VkBufferMemoryBarrier bmb[] = { {
			.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER,
			.srcAccessMask = VK_ACCESS_ACCELERATION_STRUCTURE_WRITE_BIT_KHR, // | VK_ACCESS_TRANSFER_WRITE_BIT,
			.dstAccessMask = VK_ACCESS_ACCELERATION_STRUCTURE_READ_BIT_KHR,
			.buffer = g_rtx.accels_buffer.buffer,
			.offset = 0,
			.size = VK_WHOLE_SIZE,
		}};
		vkCmdPipelineBarrier(cmdbuf,
				VK_PIPELINE_STAGE_ACCELERATION_STRUCTURE_BUILD_BIT_KHR,
				VK_PIPELINE_STAGE_ACCELERATION_STRUCTURE_BUILD_BIT_KHR,
				0, 0, NULL, ARRAYSIZE(bmb), bmb, 0, NULL);
	}

	// 2. Create TLAS
	{
		const VkAccelerationStructureGeometryKHR tl_geom[] = {
			{
				.sType = VK_STRUCTURE_TYPE_ACCELERATION_STRUCTURE_GEOMETRY_KHR,
				//.flags = VK_GEOMETRY_OPAQUE_BIT,
				.geometryType = VK_GEOMETRY_TYPE_INSTANCES_KHR,
				.geometry.instances =
					(VkAccelerationStructureGeometryInstancesDataKHR){
						.sType = VK_STRUCTURE_TYPE_ACCELERATION_STRUCTURE_GEOMETRY_INSTANCES_DATA_KHR,
						.data.deviceAddress = getBufferDeviceAddress(g_rtx.tlas_geom_buffer.buffer),
						.arrayOfPointers = VK_FALSE,
					},
			},
		};

		const uint32_t tl_max_prim_counts[ARRAYSIZE(tl_geom)] = {g_rtx_scene.num_accels};
		const VkAccelerationStructureBuildRangeInfoKHR tl_build_range = {
				.primitiveCount = g_rtx_scene.num_accels,
		};
		const VkAccelerationStructureBuildRangeInfoKHR *tl_build_ranges[] = {&tl_build_range};
		g_rtx.tlas = createAndBuildAccelerationStructure(cmdbuf,
			tl_geom, tl_max_prim_counts, tl_build_ranges, ARRAYSIZE(tl_geom), VK_ACCELERATION_STRUCTURE_TYPE_TOP_LEVEL_KHR);
	}

	// 3. Update descriptor sets (bind dest image, tlas, projection matrix)
	{
		const VkDescriptorImageInfo dii = {
			.sampler = VK_NULL_HANDLE,
			.imageView = img_dst_view,
			.imageLayout = VK_IMAGE_LAYOUT_GENERAL,
		};
		/*
		const VkDescriptorBufferInfo dbi = {
			.buffer = g_rtx.tri_buf.buffer,
			.offset = 0,
			.range = VK_WHOLE_SIZE,
		};
		*/
		const VkWriteDescriptorSetAccelerationStructureKHR wdsas = {
			.sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET_ACCELERATION_STRUCTURE_KHR,
			.accelerationStructureCount = 1,
			.pAccelerationStructures = &g_rtx.tlas,
		};
		const VkWriteDescriptorSet wds[] = {
			{
				.sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET,
				.descriptorCount = 1,
				.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE,
				.dstSet = g_rtx.desc_set,
				.dstBinding = 0,
				.dstArrayElement = 0,
				.pImageInfo = &dii,
			},
			/*
			{
				.sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET,
				.descriptorCount = 1,
				.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER,
				.dstSet = g_rtx.desc_set,
				.dstBinding = 2,
				.dstArrayElement = 0,
				.pBufferInfo = &dbi,
			},*/
			{
				.sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET,
				.descriptorCount = 1,
				.descriptorType = VK_DESCRIPTOR_TYPE_ACCELERATION_STRUCTURE_KHR,
				.dstSet = g_rtx.desc_set,
				.dstBinding = 1,
				.dstArrayElement = 0,
				.pNext = &wdsas,
			},
		};

		vkUpdateDescriptorSets(vk_core.device, ARRAYSIZE(wds), wds, 0, NULL);
	}
	

	// 4. Barrier for TLAS build and dest image layout transfer
	{
		VkBufferMemoryBarrier bmb[] = { {
			.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER,
			.srcAccessMask = VK_ACCESS_ACCELERATION_STRUCTURE_WRITE_BIT_KHR,
			.dstAccessMask = VK_ACCESS_SHADER_READ_BIT,
			.buffer = g_rtx.accels_buffer.buffer,
			.offset = 0,
			.size = VK_WHOLE_SIZE,
		}};
		VkImageMemoryBarrier image_barrier[] = { {
			.sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
			.image = img_dst,
			.srcAccessMask = 0,
			.dstAccessMask = VK_ACCESS_SHADER_WRITE_BIT,
			.oldLayout = VK_IMAGE_LAYOUT_UNDEFINED,
			.newLayout = VK_IMAGE_LAYOUT_GENERAL,
			.subresourceRange = (VkImageSubresourceRange) {
				.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
				.baseMipLevel = 0,
				.levelCount = 1,
				.baseArrayLayer = 0,
				.layerCount = 1,
			}} };
		vkCmdPipelineBarrier(cmdbuf, VK_PIPELINE_STAGE_ACCELERATION_STRUCTURE_BUILD_BIT_KHR, VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, 0,
			0, NULL, ARRAYSIZE(bmb), bmb, ARRAYSIZE(image_barrier), image_barrier);
	}

	// 4. dispatch compute
	vkCmdBindPipeline(cmdbuf, VK_PIPELINE_BIND_POINT_COMPUTE, g_rtx.pipeline);
	//vkCmdPushConstants(cmdbuf, g_rtx.pipeline_layout, VK_SHADER_STAGE_COMPUTE_BIT, 0, sizeof(matrix4x4), mvp);
	vkCmdBindDescriptorSets(cmdbuf, VK_PIPELINE_BIND_POINT_COMPUTE, g_rtx.pipeline_layout, 0, 1, &g_rtx.desc_set, 0, NULL);
	vkCmdDispatch(cmdbuf, (w+WG_W-1)/WG_W, (h+WG_H-1)/WG_H, 1);
}

static void createLayouts( void ) {
  VkDescriptorSetLayoutBinding bindings[] = {{
		.binding = 0,
		.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE,
		.descriptorCount = 1,
		.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
	}, {
		.binding = 1,
		.descriptorType = VK_DESCRIPTOR_TYPE_ACCELERATION_STRUCTURE_KHR,
		.descriptorCount = 1,
		.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT,
	/* }, { */ /* 	.binding = 2, */
	/* 	.descriptorType = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, */
	/* 	.descriptorCount = 1, */
	/* 	.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT, */
	},
	};

	VkDescriptorSetLayoutCreateInfo dslci = {.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO, .bindingCount = ARRAYSIZE(bindings), .pBindings = bindings, };

	XVK_CHECK(vkCreateDescriptorSetLayout(vk_core.device, &dslci, NULL, &g_rtx.desc_layout));

	/* VkPushConstantRange push_const = {0}; */
	/* push_const.offset = 0; */
	/* push_const.size = sizeof(matrix4x4); */
	/* push_const.stageFlags = VK_SHADER_STAGE_COMPUTE_BIT; */

	{
		VkPipelineLayoutCreateInfo plci = {0};
		plci.sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO;
		plci.setLayoutCount = 1;
		plci.pSetLayouts = &g_rtx.desc_layout;
		//plci.pushConstantRangeCount = 1;
		//plci.pPushConstantRanges = &push_const;
		XVK_CHECK(vkCreatePipelineLayout(vk_core.device, &plci, NULL, &g_rtx.pipeline_layout));
	}

	{
		VkDescriptorPoolSize pools[] = {
			{.type = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE, .descriptorCount = 1},
			//{.type = VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, .descriptorCount = 1},
			{.type = VK_DESCRIPTOR_TYPE_ACCELERATION_STRUCTURE_KHR, .descriptorCount = 1},
		};

		VkDescriptorPoolCreateInfo dpci = {
			.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_POOL_CREATE_INFO,
			.maxSets = 1, .poolSizeCount = ARRAYSIZE(pools), .pPoolSizes = pools,
		};
		XVK_CHECK(vkCreateDescriptorPool(vk_core.device, &dpci, NULL, &g_rtx.desc_pool));
	}

	{
		VkDescriptorSetAllocateInfo dsai = {
			.sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_ALLOCATE_INFO,
			.descriptorPool = g_rtx.desc_pool,
			.descriptorSetCount = 1,
			.pSetLayouts = &g_rtx.desc_layout,
		};
		XVK_CHECK(vkAllocateDescriptorSets(vk_core.device, &dsai, &g_rtx.desc_set));
	}
}

static void createPipeline( void ) {
	VkComputePipelineCreateInfo cpci = {
		.sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO,
		.layout = g_rtx.pipeline_layout,
		.stage = (VkPipelineShaderStageCreateInfo){
			.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO,
			.stage = VK_SHADER_STAGE_COMPUTE_BIT,
			.module = loadShader("rtx.comp.spv"),
			.pName = "main",
		},
	};

	XVK_CHECK(vkCreateComputePipelines(vk_core.device, VK_NULL_HANDLE, 1, &cpci, NULL, &g_rtx.pipeline));
	vkDestroyShaderModule(vk_core.device, cpci.stage.module, NULL);
}

qboolean VK_RayInit( void )
{
	ASSERT(vk_core.rtx);
	// TODO complain and cleanup on failure
	if (!createBuffer(&g_rtx.accels_buffer, MAX_ACCELS_BUFFER,
		VK_BUFFER_USAGE_ACCELERATION_STRUCTURE_STORAGE_BIT_KHR | VK_BUFFER_USAGE_SHADER_DEVICE_ADDRESS_BIT,
		VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT
		))
	{
		return false;
	}
	g_rtx.accels_buffer_addr = getBufferDeviceAddress(g_rtx.accels_buffer.buffer);

	if (!createBuffer(&g_rtx.scratch_buffer, MAX_SCRATCH_BUFFER,
			VK_BUFFER_USAGE_ACCELERATION_STRUCTURE_STORAGE_BIT_KHR | VK_BUFFER_USAGE_SHADER_DEVICE_ADDRESS_BIT,
		VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT
		)) {
		return false;
	}
	g_rtx.scratch_buffer_addr = getBufferDeviceAddress(g_rtx.scratch_buffer.buffer);

	if (!createBuffer(&g_rtx.tlas_geom_buffer, sizeof(VkAccelerationStructureInstanceKHR) * MAX_ACCELS,
		VK_BUFFER_USAGE_SHADER_DEVICE_ADDRESS_BIT | VK_BUFFER_USAGE_STORAGE_BUFFER_BIT |
		VK_BUFFER_USAGE_ACCELERATION_STRUCTURE_BUILD_INPUT_READ_ONLY_BIT_KHR,
		/* TODO DEVICE_LOCAL */ VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT)) {
		// FIXME complain, handle
		return false;
	}

	createLayouts();
	createPipeline();

	return true;
}

void VK_RayShutdown( void )
{
	ASSERT(vk_core.rtx);

	// TODO dealloc all ASes

	vkDestroyPipeline(vk_core.device, g_rtx.pipeline, NULL);
	vkDestroyDescriptorPool(vk_core.device, g_rtx.desc_pool, NULL);
	vkDestroyPipelineLayout(vk_core.device, g_rtx.pipeline_layout, NULL);
	vkDestroyDescriptorSetLayout(vk_core.device, g_rtx.desc_layout, NULL);

	cleanupASFIXME();

	destroyBuffer(&g_rtx.scratch_buffer);
	destroyBuffer(&g_rtx.accels_buffer);
	destroyBuffer(&g_rtx.tlas_geom_buffer);
}