Remove presentation stuff from device class

Device instance is no more a pointer
Remove desc member in device structs

4 files changed, 258 insertions, 246 deletions

diff --git a/api-test/command-buffer.cpp b/api-test/command-buffer.cpp
index 78b1505..79b8aef 100644
--- a/api-test/command-buffer.cpp
+++ b/api-test/command-buffer.cpp
@@ -18,7 +18,7 @@ void command_buffer::start() const {
 
     VKCHECK(vkBeginCommandBuffer(vk_command_buffer, &begin_info));
 
-    const auto& bindless   = vkdevice::get_render_device()->get_bindingmodel();
+    const auto& bindless   = vkdevice::get_render_device().get_bindingmodel();
     const auto* global_set = &bindless.sets[BindlessSetType::GLOBAL];
     if (queue_type == QueueType::GRAPHICS) {
         vkCmdBindDescriptorSets(vk_command_buffer, VK_PIPELINE_BIND_POINT_GRAPHICS, bindless.layout, 0, 1U, global_set, 0U, nullptr);
@@ -33,13 +33,10 @@ void command_buffer::stop() const {
 }
 
 void command_buffer::barrier(handle<device_texture> texture_handle, VkPipelineStageFlags2 stage, VkAccessFlags2 access, VkImageLayout layout) const {
-    auto&                   texture     = vkdevice::get_render_device()->get_texture(texture_handle);
-    VkImageAspectFlags      aspect_mask = ((texture.desc.usages & VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT) != 0U)
-                                              ? VK_IMAGE_ASPECT_DEPTH_BIT
-                                              : VK_IMAGE_ASPECT_COLOR_BIT;
+    auto&                   texture     = vkdevice::get_render_device().get_texture(texture_handle);
 
     VkImageSubresourceRange subresource_range{
-        .aspectMask     = aspect_mask,
+        .aspectMask     = texture.aspects,
         .baseMipLevel   = 0U,
         .levelCount     = 1U,
         .baseArrayLayer = 0U,
@@ -80,14 +77,12 @@ void command_buffer::barrier(handle<device_texture> texture_handle, VkPipelineSt
 }
 
 void command_buffer::copy(handle<device_buffer> buffer_handle, ::handle<device_texture> texture_handle, VkDeviceSize offset) const {
-    auto*                    device      = vkdevice::get_render_device();
-    const auto&              buffer      = device->get_buffer(buffer_handle);
-    const auto&              texture     = device->get_texture(texture_handle);
-    VkImageAspectFlags       aspect_mask = ((texture.desc.usages & VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT) != 0U)
-                                               ? VK_IMAGE_ASPECT_DEPTH_BIT
-                                               : VK_IMAGE_ASPECT_COLOR_BIT;
+    auto&                    device      = vkdevice::get_render_device();
+    const auto&              buffer      = device.get_buffer(buffer_handle);
+    const auto&              texture     = device.get_texture(texture_handle);
+
     VkImageSubresourceLayers image_subresource_layers{
-        .aspectMask     = aspect_mask,
+        .aspectMask     = texture.aspects,
         .mipLevel       = 0,
         .baseArrayLayer = 0,
         .layerCount     = 1,
@@ -100,9 +95,9 @@ void command_buffer::copy(handle<device_buffer> buffer_handle, ::handle<device_t
         .imageSubresource  = image_subresource_layers,
         .imageOffset       = { .x = 0, .y = 0, .z = 0 },
         .imageExtent       = {
-                  .width  = texture.desc.width,
-                  .height = texture.desc.height,
-                  .depth  = texture.desc.depth,
+                  .width  = texture.width,
+                  .height = texture.height,
+                  .depth  = texture.depth,
         },
     };
 
@@ -110,8 +105,8 @@ void command_buffer::copy(handle<device_buffer> buffer_handle, ::handle<device_t
 }
 
 void graphics_command_buffer::dispatch(const dispatch_params& params) const {
-    const auto& pipeline = vkdevice::get_render_device()->get_pipeline(params.pipeline);
-    const auto& bindless = vkdevice::get_render_device()->get_bindingmodel();
+    const auto& pipeline = vkdevice::get_render_device().get_pipeline(params.pipeline);
+    const auto& bindless = vkdevice::get_render_device().get_bindingmodel();
     vkCmdBindDescriptorSets(vk_command_buffer, VK_PIPELINE_BIND_POINT_COMPUTE, bindless.layout, static_cast<uint32_t>(BindlessSetType::INSTANCES_UNIFORMS), 1U, &bindless.sets[BindlessSetType::INSTANCES_UNIFORMS], 1U, &params.uniforms_offset);
     vkCmdBindDescriptorSets(vk_command_buffer, VK_PIPELINE_BIND_POINT_COMPUTE, bindless.layout, static_cast<uint32_t>(BindlessSetType::DRAWS_UNIFORMS), 1U, &bindless.sets[BindlessSetType::DRAWS_UNIFORMS], 1U, &params.uniforms_offset);
 
@@ -125,7 +120,7 @@ void graphics_command_buffer::begin_renderpass(const renderpass_params& params)
 
     for (auto index{ 0U }; index < params.color_attachments.size(); index++) {
         auto& attachment_info              = attachments_info[index];
-        auto& attachment                   = vkdevice::get_render_device()->get_texture(params.color_attachments[index]);
+        auto& attachment                   = vkdevice::get_render_device().get_texture(params.color_attachments[index]);
 
         attachment_info.sType              = VK_STRUCTURE_TYPE_RENDERING_ATTACHMENT_INFO;
         attachment_info.pNext              = nullptr;
@@ -180,16 +175,16 @@ void graphics_command_buffer::begin_renderpass(const renderpass_params& params)
 }
 
 void graphics_command_buffer::render(const draw_params& params) const {
-    auto* device = vkdevice::get_render_device();
-    const auto& pipeline = device->get_pipeline(params.pipeline);
+    auto&       device   = vkdevice::get_render_device();
+    const auto& pipeline = device.get_pipeline(params.pipeline);
 
-    const auto& bindless = device->get_bindingmodel();
+    const auto& bindless = device.get_bindingmodel();
     vkCmdBindDescriptorSets(vk_command_buffer, pipeline.bind_point, bindless.layout, static_cast<uint32_t>(BindlessSetType::DRAWS_UNIFORMS), 1U, &bindless.sets[BindlessSetType::DRAWS_UNIFORMS], 1U, &params.uniforms_offset);
 
     vkCmdBindPipeline(vk_command_buffer, pipeline.bind_point, pipeline.vk_pipeline);
 
     if (params.index_buffer.is_valid()) {
-        const auto& index_buffer = device->get_buffer(params.index_buffer);
+        const auto& index_buffer = device.get_buffer(params.index_buffer);
         vkCmdBindIndexBuffer(vk_command_buffer, index_buffer.vk_buffer, 0U, VK_INDEX_TYPE_UINT16);
         vkCmdDrawIndexed(vk_command_buffer, params.vertex_count, params.instance_count, params.index_offset, static_cast<int32_t>(params.vertex_offset), 0U);
     } else {
diff --git a/api-test/main.cpp b/api-test/main.cpp
index 7f30fda..807883d 100644
--- a/api-test/main.cpp
+++ b/api-test/main.cpp
@@ -9,6 +9,7 @@
 
 #include "imgui.h"
 
+#include "handle.hpp"
 #include "window.hpp"
 #include "vk-device.hpp"
 #include "utils.hpp"
@@ -85,17 +86,28 @@ int main() {
     io.DisplaySize.x = 1024.f;
     io.DisplaySize.y = 1024.f;
 
-    auto* device = vkdevice::init_device(wnd);
+    auto& device = vkdevice::get_render_device();
+    device.init();
+
+    device.create_surface({
+        .window_handle = wnd.handle,
+        .surface_format = {
+            .format = VK_FORMAT_B8G8R8A8_UNORM,
+            .colorSpace = VK_COLOR_SPACE_SRGB_NONLINEAR_KHR,
+        },
+        .present_mode = VK_PRESENT_MODE_FIFO_KHR,
+        .usages = VK_IMAGE_USAGE_TRANSFER_DST_BIT,
+    });
 
     constexpr size_t image_size = 1024U;
     constexpr size_t image_count = 4U;
-    auto staging_buffer_handle = device->create_buffer({
+    auto staging_buffer_handle = device.create_buffer({
         .size = image_size * image_size * sizeof(uint32_t) * image_count,
         .usages = VK_BUFFER_USAGE_TRANSFER_SRC_BIT,
         .memory_properties = VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT,
         .memory_usage = VMA_MEMORY_USAGE_CPU_TO_GPU,
     });
-    const auto& staging_buffer = device->get_buffer(staging_buffer_handle);
+    const auto& staging_buffer = device.get_buffer(staging_buffer_handle);
 
     int width;
     int height;
@@ -104,7 +116,7 @@ int main() {
     size_t size = static_cast<size_t>(width) * height * 4U;
     memcpy(staging_buffer.mapped_ptr, data, size);
 
-    auto gpu_texture_handle = device->create_texture({
+    auto gpu_texture_handle = device.create_texture({
         .width  = static_cast<uint32_t>(width),
         .height = static_cast<uint32_t>(height),
         .usages = VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_STORAGE_BIT | VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT,
@@ -118,7 +130,7 @@ int main() {
     auto* addr = static_cast<void*>(static_cast<uint8_t*>(staging_buffer.mapped_ptr) + offset);
     memcpy(addr, data, size);
 
-    auto second_texture = device->create_texture({
+    auto second_texture = device.create_texture({
         .width  = static_cast<uint32_t>(width),
         .height = static_cast<uint32_t>(height),
         .usages = VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_STORAGE_BIT,
@@ -126,7 +138,7 @@ int main() {
         .type   = VK_IMAGE_TYPE_2D,
     });
 
-    auto output_texture_handle = device->create_texture({
+    auto output_texture_handle = device.create_texture({
         .width  = static_cast<uint32_t>(width),
         .height = static_cast<uint32_t>(height),
         .usages = VK_IMAGE_USAGE_STORAGE_BIT,
@@ -138,14 +150,14 @@ int main() {
     int atlas_width;
     int atlas_height;
     io.Fonts->GetTexDataAsRGBA32(&pixels, &atlas_width, &atlas_height);
-    const auto ui_texture_handle = device->create_texture({
+    const auto ui_texture_handle = device.create_texture({
         .width = static_cast<uint32_t>(atlas_width),
         .height = static_cast<uint32_t>(atlas_height),
         .usages = VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_SAMPLED_BIT,
         .format = VK_FORMAT_R8G8B8A8_UNORM,
         .type = VK_IMAGE_TYPE_2D,
     });
-    const auto& ui_texture = device->get_texture(ui_texture_handle);
+    const auto& ui_texture = device.get_texture(ui_texture_handle);
     io.Fonts->SetTexID(*(void **)&ui_texture_handle.id);
 
     off_t ui_offset = 2L * offset;
@@ -153,38 +165,40 @@ int main() {
     addr = static_cast<void*>(static_cast<uint8_t*>(staging_buffer.mapped_ptr) + ui_offset);
     memcpy(addr, pixels, size);
 
-    auto vertex_buffer = device->get_buffer(device->create_buffer({
+    auto vertex_buffer = device.get_buffer(device.create_buffer({
         .size = 1024U * 1024U * sizeof(uint32_t),
         .usages = VK_BUFFER_USAGE_STORAGE_BUFFER_BIT | VK_BUFFER_USAGE_SHADER_DEVICE_ADDRESS_BIT,
         .memory_properties = VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT,
         .memory_usage = VMA_MEMORY_USAGE_CPU_TO_GPU,
     }));
 
-    auto index_buffer_handle = device->create_buffer({
+    auto index_buffer_handle = device.create_buffer({
         .size = 1024U * 1024U * sizeof(uint32_t),
         .usages = VK_BUFFER_USAGE_STORAGE_BUFFER_BIT | VK_BUFFER_USAGE_INDEX_BUFFER_BIT,
         .memory_properties = VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT,
         .memory_usage = VMA_MEMORY_USAGE_CPU_TO_GPU,
     });
-    auto index_buffer = device->get_buffer(index_buffer_handle);
+    auto index_buffer = device.get_buffer(index_buffer_handle);
 
     auto code = read_file("shaders/test.comp.spv");
-    auto compute = device->create_pipeline({
+    auto compute = device.create_pipeline({
         .cs_code = code,
     });
 
     std::array<VkFormat, 1U> formats { VK_FORMAT_R8G8B8A8_UNORM };
     auto vertex_code = read_file("shaders/api-test-ui.vert.spv");
     auto fragment_code = read_file("shaders/api-test-ui.frag.spv");
-    auto graphics = device->create_pipeline({
+    auto graphics = device.create_pipeline({
         .vs_code = vertex_code,
         .fs_code = fragment_code,
         .color_attachments_format = formats
     });
 
-    auto semaphore = device->create_semaphore();
+    auto semaphore = device.create_semaphore({
+        .type = VK_SEMAPHORE_TYPE_TIMELINE,
+    });
     std::array<graphics_command_buffer, 2U> command_buffers;
-    device->allocate_command_buffers(command_buffers.data(), 2, QueueType::GRAPHICS);
+    device.allocate_command_buffers(command_buffers.data(), 2, QueueType::GRAPHICS);
 
     RD_START_CAPTURE;
 
@@ -205,11 +219,11 @@ int main() {
         uint32_t input2;
         uint32_t output;
     };
-    const auto& uniform_buffer = device->get_bindingmodel().get_uniform_buffer();
+    const auto& uniform_buffer = device.get_bindingmodel().get_uniform_buffer();
     compute_params params {
-        device->get_texture(second_texture).storage_index,
-        device->get_texture(gpu_texture_handle).storage_index,
-        device->get_texture(output_texture_handle).storage_index,
+        device.get_texture(second_texture).storage_index,
+        device.get_texture(gpu_texture_handle).storage_index,
+        device.get_texture(output_texture_handle).storage_index,
     };
     memcpy(uniform_buffer.mapped_ptr, &params, sizeof(params));
 
@@ -225,7 +239,7 @@ int main() {
     command_buffers[0].stop();
 
     std::array<struct command_buffer, 1U> compute_command_buffers = {command_buffers[0]};
-    device->submit(compute_command_buffers, nullptr, &semaphore);
+    device.submit(compute_command_buffers, {}, semaphore);
 
     RD_END_CAPTURE;
 
@@ -239,7 +253,7 @@ int main() {
         auto* draw_data = ImGui::GetDrawData();
         update_buffers(draw_data, vertex_buffer, index_buffer);
 
-        device->wait(semaphore);
+        device.wait(semaphore);
 
         RD_START_CAPTURE;
         command_buffers[1].start(),
@@ -261,8 +275,6 @@ int main() {
             uint32_t texture_index;
         };
 
-        printf("%llu\n", sizeof(ImDrawVert));
-
         uint32_t draw_index = 0U;
         std::array<draw_params, 100U> draws;
         for (auto index {0U} ; index < draw_data->CmdListsCount; index++) {
@@ -302,10 +314,8 @@ int main() {
 
         command_buffers[1].stop();
 
-        device->submit(ui_command_buffers, nullptr, &semaphore);
+        device.submit(ui_command_buffers, {}, semaphore);
         RD_END_CAPTURE;
-
-        break;
     }
 
     return 0;
diff --git a/api-test/vk-bindless.cpp b/api-test/vk-bindless.cpp
index 654ce0c..c48aea1 100644
--- a/api-test/vk-bindless.cpp
+++ b/api-test/vk-bindless.cpp
@@ -1,20 +1,20 @@
 #include "vk-bindless.hpp"
 
-#include <vma/vk_mem_alloc.h>
+#include <vk_mem_alloc.h>
 
 #include "vk-device.hpp"
 #include "vk-utils.hpp"
 #include "vulkan-loader.hpp"
 
 const device_buffer& bindless_model::get_uniform_buffer() const {
-    auto* render_device = vkdevice::get_render_device();
-    return render_device->get_buffer(draws_uniform_buffer_handle);
+    auto& render_device = vkdevice::get_render_device();
+    return render_device.get_buffer(draws_uniform_buffer_handle);
 }
 
 bindless_model bindless_model::create_bindless_model() {
     bindless_model bindless;
-    auto* render_device = vkdevice::get_render_device();
-    auto* device = render_device->get_device();
+    auto& render_device = vkdevice::get_render_device();
+    auto* device = render_device.get_device();
 #ifndef USE_VK_DESCRIPTOR_BUFFER
     bindless.create_immutable_samplers(device);
     bindless.create_layout(device);
@@ -24,14 +24,14 @@ bindless_model bindless_model::create_bindless_model() {
     VkWriteDescriptorSet writes_descriptor_set[2U];
 
     {
-        bindless.instances_uniform_buffer_handle = render_device->create_buffer({
+        bindless.instances_uniform_buffer_handle = render_device.create_buffer({
             .size = instances_uniform_buffer_size,
             .usages = VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT,
             .memory_properties = VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT,
             .memory_usage = VMA_MEMORY_USAGE_CPU_TO_GPU,
         });
 
-        const auto& buffer = render_device->get_buffer(bindless.instances_uniform_buffer_handle);
+        const auto& buffer = render_device.get_buffer(bindless.instances_uniform_buffer_handle);
         VkDescriptorBufferInfo buffer_info {
             .buffer = buffer.vk_buffer,
             .offset = 0U,
@@ -53,14 +53,14 @@ bindless_model bindless_model::create_bindless_model() {
     }
 
     {
-        bindless.draws_uniform_buffer_handle = render_device->create_buffer({
+        bindless.draws_uniform_buffer_handle = render_device.create_buffer({
             .size = draws_uniform_buffer_size,
             .usages = VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT,
             .memory_properties = VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT,
             .memory_usage = VMA_MEMORY_USAGE_CPU_TO_GPU,
         });
 
-        const auto& buffer = render_device->get_buffer(bindless.draws_uniform_buffer_handle);
+        const auto& buffer = render_device.get_buffer(bindless.draws_uniform_buffer_handle);
         VkDescriptorBufferInfo buffer_info {
             .buffer = buffer.vk_buffer,
             .offset = 0U,
@@ -118,7 +118,7 @@ bindless_model bindless_model::create_bindless_model() {
 }
 
 void bindless_model::destroy_bindless_model(bindless_model& bindless) {
-    auto* device = vkdevice::get_render_device()->get_device();
+    auto* device = vkdevice::get_render_device().get_device();
     vkDestroyDescriptorPool(device, bindless.descriptor_pool, nullptr);
 
     for (auto index{ 0U }; index < BindlessSetType::MAX; index++) {
diff --git a/api-test/vk-device.cpp b/api-test/vk-device.cpp
index 502143a..2258bdb 100644
--- a/api-test/vk-device.cpp
+++ b/api-test/vk-device.cpp
@@ -8,30 +8,26 @@
 
 #include "vk-utils.hpp"
 #include "vulkan-loader.hpp"
-#include "window.hpp"
 
-vkdevice* vkdevice::render_device = nullptr;
+vkdevice vkdevice::render_device = vkdevice();
+
+void vkdevice::init() {
+    queues[0U] = {
+        .usages = VK_QUEUE_GRAPHICS_BIT |
+                  VK_QUEUE_COMPUTE_BIT |
+                  VK_QUEUE_TRANSFER_BIT,
+    };
+   queues[1U] = {
+        .usages = VK_QUEUE_COMPUTE_BIT,
+    };
+    queues[2U] = {
+        .usages = VK_QUEUE_TRANSFER_BIT,
+    };
 
-vkdevice::vkdevice(const window& window)
-        : queues{
-              {
-                  .usages = VK_QUEUE_GRAPHICS_BIT |
-                            VK_QUEUE_COMPUTE_BIT |
-                            VK_QUEUE_TRANSFER_BIT,
-              },
-              {
-                  .usages = VK_QUEUE_COMPUTE_BIT,
-              },
-              {
-                  .usages = VK_QUEUE_TRANSFER_BIT,
-              },
-          } {
     load_vulkan();
 
     create_instance();
 
-    create_surface(window);
-
 #ifdef _DEBUG
     create_debug_layer_callback();
 #endif // _DEBUG
@@ -40,41 +36,32 @@ vkdevice::vkdevice(const window& window)
 
     create_memory_allocator();
 
-    create_swapchain();
-
     bindless = bindless_model::create_bindless_model();
 }
 
-vkdevice* vkdevice::init_device(const window& window) {
-    vkdevice::render_device = static_cast<vkdevice*>(malloc(sizeof(vkdevice)));
-    vkdevice::render_device = new (vkdevice::render_device) vkdevice(window);
-    return render_device;
-}
-
-void vkdevice::free_device() {
-    delete render_device;
-    render_device = nullptr;
-}
-
 vkdevice::~vkdevice() {
     vkDeviceWaitIdle(device);
 
     bindless_model::destroy_bindless_model(bindless);
 
-    for (auto& texture : textures) {
+    for (const auto& texture : textures) {
         destroy_texture(texture);
     }
 
-    for (auto& buffer : buffers) {
+    for (const auto& buffer : buffers) {
         destroy_buffer(buffer);
     }
 
-    for (auto& pipeline : pipelines) {
+    for (const auto& pipeline : pipelines) {
         destroy_pipeline(pipeline);
     }
 
-    if (swapchain != nullptr) {
-        vkDestroySwapchainKHR(device, swapchain, nullptr);
+    for (const auto& semaphore : semaphores) {
+        destroy_semaphore(semaphore);
+    }
+
+    for (const auto& surface : surfaces) {
+        destroy_surface(surface);
     }
 
     vmaDestroyAllocator(gpu_allocator);
@@ -89,15 +76,15 @@ vkdevice::~vkdevice() {
     vkDestroyDebugUtilsMessengerEXT(instance, debug_messenger, nullptr);
 #endif // _DEBUG
 
-    if (surface != nullptr) {
-        vkDestroySurfaceKHR(instance, surface, nullptr);
-    }
     vkDestroyInstance(instance, nullptr);
 }
 
 handle<device_texture> vkdevice::create_texture(const texture_desc& desc) {
     device_texture device_texture{
-        .desc = desc,
+        .format = desc.format,
+        .width = desc.width,
+        .height = desc.height,
+        .mips = desc.mips,
     };
 
     assert(desc.mips <= texture_desc::max_mips);
@@ -135,7 +122,11 @@ handle<device_texture> vkdevice::create_texture(const texture_desc& desc) {
 
     VKCHECK(vmaCreateImage(gpu_allocator, &create_info, &alloc_create_info, &device_texture.vk_image, &device_texture.alloc, nullptr));
 
-    create_views(device_texture);
+    device_texture.aspects = ((desc.usages & VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT) != 0U)
+        ? VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT
+        : VK_IMAGE_ASPECT_COLOR_BIT;
+
+    create_views(desc, device_texture);
 
     // TODO: Handle mips views
     VkDescriptorImageInfo images_info[]{
@@ -187,7 +178,6 @@ handle<device_texture> vkdevice::create_texture(const texture_desc& desc) {
 
 handle<device_buffer> vkdevice::create_buffer(const buffer_desc& desc) {
     device_buffer device_buffer{
-        .desc = desc,
     };
 
     VkBufferCreateInfo create_info{
@@ -234,29 +224,14 @@ handle<device_buffer> vkdevice::create_buffer(const buffer_desc& desc) {
 
 handle<device_pipeline> vkdevice::create_pipeline(const pipeline_desc& desc) {
     device_pipeline device_pipeline{
-        .desc     = desc,
     };
     if (!desc.cs_code.empty()) {
-        VkShaderModule shader_module;
-        {
-            const auto&              code = desc.cs_code;
-            VkShaderModuleCreateInfo create_info{
-                .sType    = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO,
-                .pNext    = nullptr,
-                .flags    = 0U,
-                .codeSize = code.size(),
-                .pCode    = (uint32_t*)code.data(),
-            };
-
-            VKCHECK(vkCreateShaderModule(device, &create_info, nullptr, &shader_module));
-        }
-
         VkPipelineShaderStageCreateInfo stage_create_info{
             .sType               = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO,
             .pNext               = nullptr,
             .flags               = 0U,
             .stage               = VK_SHADER_STAGE_COMPUTE_BIT,
-            .module              = shader_module,
+            .module              = create_shader_module(desc.cs_code),
             .pName               = "main",
             .pSpecializationInfo = nullptr,
         };
@@ -274,7 +249,7 @@ handle<device_pipeline> vkdevice::create_pipeline(const pipeline_desc& desc) {
         device_pipeline.bind_point = VK_PIPELINE_BIND_POINT_COMPUTE;
         VKCHECK(vkCreateComputePipelines(device, nullptr, 1U, &create_info, nullptr, &device_pipeline.vk_pipeline));
 
-        vkDestroyShaderModule(device, shader_module, nullptr);
+        vkDestroyShaderModule(device, stage_create_info.module, nullptr);
     } else {
         VkPipelineShaderStageCreateInfo shader_stages_create_info[2U]{
             {
@@ -282,6 +257,7 @@ handle<device_pipeline> vkdevice::create_pipeline(const pipeline_desc& desc) {
                 .pNext               = nullptr,
                 .flags               = 0U,
                 .stage               = VK_SHADER_STAGE_VERTEX_BIT,
+                .module              = create_shader_module(desc.vs_code),
                 .pName               = "main",
                 .pSpecializationInfo = nullptr,
             },
@@ -290,36 +266,12 @@ handle<device_pipeline> vkdevice::create_pipeline(const pipeline_desc& desc) {
                 .pNext               = nullptr,
                 .flags               = 0U,
                 .stage               = VK_SHADER_STAGE_FRAGMENT_BIT,
+                .module              = create_shader_module(desc.fs_code),
                 .pName               = "main",
                 .pSpecializationInfo = nullptr,
             },
         };
 
-        {
-            const auto&              code = desc.vs_code;
-            VkShaderModuleCreateInfo module_create_info{
-                .sType    = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO,
-                .pNext    = nullptr,
-                .flags    = 0U,
-                .codeSize = code.size(),
-                .pCode    = (uint32_t*)code.data(),
-            };
-
-            VKCHECK(vkCreateShaderModule(device, &module_create_info, nullptr, &shader_stages_create_info[0U].module));
-        }
-        {
-            const auto&              code = desc.fs_code;
-            VkShaderModuleCreateInfo module_create_info{
-                .sType    = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO,
-                .pNext    = nullptr,
-                .flags    = 0U,
-                .codeSize = code.size(),
-                .pCode    = (uint32_t*)code.data(),
-            };
-
-            VKCHECK(vkCreateShaderModule(device, &module_create_info, nullptr, &shader_stages_create_info[1U].module));
-        }
-
         VkPipelineVertexInputStateCreateInfo vertex_input_state_create_info{
             .sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO,
         };
@@ -431,20 +383,49 @@ handle<device_pipeline> vkdevice::create_pipeline(const pipeline_desc& desc) {
 
         device_pipeline.bind_point = VK_PIPELINE_BIND_POINT_GRAPHICS;
         VKCHECK(vkCreateGraphicsPipelines(device, nullptr, 1U, &create_info, nullptr, &device_pipeline.vk_pipeline));
+
+        for(auto& stage_create_info : shader_stages_create_info) {
+            vkDestroyShaderModule(device, stage_create_info.module, nullptr);
+        }
     }
 
     return { pipelines.add(device_pipeline) };
 }
 
-vksemaphore vkdevice::create_semaphore() {
-    vksemaphore semaphore{
-        .value = 0,
+handle<device_surface> vkdevice::create_surface(const surface_desc& desc) {
+    device_surface device_surface {
+    };
+
+    {
+        VkWin32SurfaceCreateInfoKHR create_info{
+            .sType     = VK_STRUCTURE_TYPE_WIN32_SURFACE_CREATE_INFO_KHR,
+            .pNext     = nullptr,
+            .flags     = 0U,
+            .hinstance = GetModuleHandle(nullptr),
+            .hwnd      = desc.window_handle,
+        };
+
+        VKCHECK(vkCreateWin32SurfaceKHR(instance, &create_info, nullptr, &device_surface.vk_surface));
+    }
+
+    create_swapchain(desc, device_surface);
+
+    vkGetSwapchainImagesKHR(device, device_surface.vk_swapchain, &device_surface.image_count, nullptr);
+    std::vector<VkImage> images {device_surface.image_count};
+    vkGetSwapchainImagesKHR(device, device_surface.vk_swapchain, &device_surface.image_count, images.data());
+
+    return { surfaces.add(device_surface) };
+}
+
+handle<device_semaphore> vkdevice::create_semaphore(const semaphore_desc& desc) {
+    device_semaphore semaphore {
+        .value = desc.initial_value,
     };
     VkSemaphoreTypeCreateInfo semaphore_type_create_info{
         .sType         = VK_STRUCTURE_TYPE_SEMAPHORE_TYPE_CREATE_INFO,
         .pNext         = nullptr,
         .semaphoreType = VK_SEMAPHORE_TYPE_TIMELINE,
-        .initialValue  = 0U,
+        .initialValue  = desc.initial_value,
     };
 
     VkSemaphoreCreateInfo create_info{
@@ -455,7 +436,7 @@ vksemaphore vkdevice::create_semaphore() {
 
     VKCHECK(vkCreateSemaphore(device, &create_info, nullptr, &semaphore.vk_semaphore));
 
-    return semaphore;
+    return { semaphores.add(semaphore) };
 }
 
 void vkdevice::destroy_texture(const device_texture& texture) {
@@ -495,7 +476,30 @@ void vkdevice::destroy_pipeline(handle<device_pipeline> handle) {
     vkDestroyPipeline(device, pipeline.vk_pipeline, nullptr);
 }
 
-void vkdevice::wait(vksemaphore& semaphore) {
+void vkdevice::destroy_semaphore(const device_semaphore& semaphore) {
+    destroy_semaphore({ .id = semaphores.find(semaphore)});
+}
+
+void vkdevice::destroy_semaphore(handle<device_semaphore> handle) {
+    auto& semaphore = vkdevice::get_semaphore(handle);
+
+    vkDestroySemaphore(device, semaphore.vk_semaphore, nullptr);
+}
+
+void vkdevice::destroy_surface(const device_surface& surface) {
+    destroy_surface({ .id = surfaces.find(surface)});
+}
+
+void vkdevice::destroy_surface(handle<device_surface> handle) {
+    const auto& surface = vkdevice::get_surface(handle);
+
+    vkDestroySwapchainKHR(device, surface.vk_swapchain, nullptr);
+
+    vkDestroySurfaceKHR(instance, surface.vk_surface, nullptr);
+}
+
+void vkdevice::wait(handle<device_semaphore> semaphore_handle) {
+    const auto& semaphore = vkdevice::get_render_device().get_semaphore(semaphore_handle);
     VkSemaphoreWaitInfo wait_info {
         .sType = VK_STRUCTURE_TYPE_SEMAPHORE_WAIT_INFO,
         .pNext = nullptr,
@@ -507,32 +511,10 @@ void vkdevice::wait(vksemaphore& semaphore) {
     VKCHECK(vkWaitSemaphores(device, &wait_info, -1));
 }
 
-void vkdevice::submit(std::span<command_buffer> buffers, vksemaphore* wait, vksemaphore* signal) {
-#define USE_SUBMIT2
+void vkdevice::submit(std::span<command_buffer> buffers, handle<device_semaphore> wait_handle, handle<device_semaphore> signal_handle) {
     assert((max_submitable_command_buffers - buffers.size()) >= 0);
     const auto queue_type = buffers[0].queue_type;
-#ifndef USE_SUBMIT2
-    VkCommandBuffer to_submit[max_submitable_command_buffers];
-
-    for (auto index{ 0U }; index < count; index++) {
-        assert(queue_type == buffers[index].queue_type);
-        to_submit[index] = buffers[index].vk_command_buffer;
-    }
 
-    VkSubmitInfo submit_info{
-        .sType                = VK_STRUCTURE_TYPE_SUBMIT_INFO,
-        .pNext                = nullptr,
-        .waitSemaphoreCount   = 0U,
-        .pWaitSemaphores      = nullptr,
-        .pWaitDstStageMask    = nullptr,
-        .commandBufferCount   = static_cast<uint32_t>(count),
-        .pCommandBuffers      = to_submit,
-        .signalSemaphoreCount = 0U,
-        .pSignalSemaphores    = nullptr,
-    };
-
-    VKCHECK(vkQueueSubmit(queues[static_cast<uint32_t>(queue_type)].vk_queue, 1, &submit_info, nullptr));
-#else
     VkCommandBufferSubmitInfo command_buffers_info[buffers.size()];
 
     for (auto index{ 0U }; index < buffers.size(); index++) {
@@ -567,38 +549,44 @@ void vkdevice::submit(std::span<command_buffer> buffers, vksemaphore* wait, vkse
         .pCommandBufferInfos      = command_buffers_info,
     };
 
-    if (wait != nullptr) {
-        wait_semaphore_info.semaphore      = wait->vk_semaphore;
-        wait_semaphore_info.value          = wait->value;
+    if (wait_handle.is_valid()) {
+        const auto& wait = vkdevice::get_render_device().get_semaphore(wait_handle);
+        wait_semaphore_info.semaphore      = wait.vk_semaphore;
+        wait_semaphore_info.value          = wait.value;
         submit_info.waitSemaphoreInfoCount = 1U;
         submit_info.pWaitSemaphoreInfos    = &wait_semaphore_info;
     }
 
-    if (signal != nullptr) {
-        signal_semaphore_info.semaphore      = signal->vk_semaphore;
-        signal_semaphore_info.value          = ++signal->value;
+    if (signal_handle.is_valid()) {
+        auto& signal = vkdevice::get_render_device().get_semaphore(signal_handle);
+        signal_semaphore_info.semaphore      = signal.vk_semaphore;
+        signal_semaphore_info.value          = ++signal.value;
 
         submit_info.signalSemaphoreInfoCount = 1U;
         submit_info.pSignalSemaphoreInfos    = &signal_semaphore_info;
     }
 
     VKCHECK(vkQueueSubmit2(queues[static_cast<uint32_t>(queue_type)].vk_queue, 1, &submit_info, nullptr));
-#endif
 }
 
-void vkdevice::present(device_surface& surface, const vksemaphore& semaphore) {
+void vkdevice::present(handle<device_surface> surface_handle, handle<device_semaphore> semaphore_handle) {
+    const auto& surface = vkdevice::get_render_device().get_surface(surface_handle);
     VkResult result;
     VkPresentInfoKHR present_info{
         .sType = VK_STRUCTURE_TYPE_PRESENT_INFO_KHR,
         .pNext = nullptr,
-        .waitSemaphoreCount = 1U,
-        .pWaitSemaphores = &semaphore.vk_semaphore,
         .swapchainCount = 1U,
         .pSwapchains = &surface.vk_swapchain,
         .pImageIndices = &surface.image_index,
         .pResults = &result,
     };
 
+    if (semaphore_handle.is_valid()) {
+        const auto& semaphore = vkdevice::get_render_device().get_semaphore(semaphore_handle);
+        present_info.waitSemaphoreCount = 1U;
+        present_info.pWaitSemaphores = &semaphore.vk_semaphore;
+    }
+
     VKCHECK(vkQueuePresentKHR(queues[static_cast<uint32_t>(QueueType::GRAPHICS)].vk_queue, &present_info));
 
     VKCHECK(result);
@@ -626,17 +614,29 @@ void vkdevice::allocate_command_buffers(command_buffer* buffers, size_t count, Q
     }
 }
 
-void vkdevice::create_views(device_texture& texture) {
-    VkImageAspectFlags    aspect_mask = ((texture.desc.usages & VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT) != 0U)
-                                            ? VK_IMAGE_ASPECT_DEPTH_BIT
-                                            : VK_IMAGE_ASPECT_COLOR_BIT;
+VkShaderModule vkdevice::create_shader_module(const std::span<uint8_t>& shader_code) {
+    VkShaderModule shader_module;
+    const auto&              code = shader_code;
+    VkShaderModuleCreateInfo create_info{
+        .sType    = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO,
+        .pNext    = nullptr,
+        .flags    = 0U,
+        .codeSize = code.size(),
+        .pCode    = reinterpret_cast<uint32_t*>(code.data()),
+    };
+
+    VKCHECK(vkCreateShaderModule(device, &create_info, nullptr, &shader_module));
+    return shader_module;
+}
+
+void vkdevice::create_views(const texture_desc& desc, device_texture& texture) {
     VkImageViewCreateInfo create_info{
         .sType      = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO,
         .pNext      = nullptr,
         .flags      = 0U,
         .image      = texture.vk_image,
         .viewType   = VK_IMAGE_VIEW_TYPE_2D,
-        .format     = texture.desc.format,
+        .format     = texture.format,
         .components = {
             .r = VK_COMPONENT_SWIZZLE_R,
             .g = VK_COMPONENT_SWIZZLE_G,
@@ -644,7 +644,7 @@ void vkdevice::create_views(device_texture& texture) {
             .a = VK_COMPONENT_SWIZZLE_A,
         },
         .subresourceRange = {
-            .aspectMask     = aspect_mask,
+            .aspectMask     = texture.aspects,
             .baseMipLevel   = 0,
             .levelCount     = 1,
             .baseArrayLayer = 0,
@@ -652,41 +652,32 @@ void vkdevice::create_views(device_texture& texture) {
         },
     };
 
-    for (auto mip_index{ 0 }; mip_index < texture.desc.mips; mip_index++) {
+    for (auto mip_index{ 0 }; mip_index < texture.mips; mip_index++) {
         create_info.subresourceRange.baseMipLevel = mip_index;
         VKCHECK(vkCreateImageView(device, &create_info, nullptr, &texture.mips_views[mip_index]));
     }
 
     create_info.subresourceRange.baseMipLevel = 0;
-    create_info.subresourceRange.levelCount   = texture.desc.mips;
+    create_info.subresourceRange.levelCount   = texture.mips;
     VKCHECK(vkCreateImageView(device, &create_info, nullptr, &texture.whole_view));
 }
 
-// Presentation stuff
-void vkdevice::create_surface(const window& window) {
-    VkWin32SurfaceCreateInfoKHR create_info{
-        .sType     = VK_STRUCTURE_TYPE_WIN32_SURFACE_CREATE_INFO_KHR,
-        .pNext     = nullptr,
-        .flags     = 0U,
-        .hinstance = GetModuleHandle(nullptr),
-        .hwnd      = window.handle,
-    };
-
-    VKCHECK(vkCreateWin32SurfaceKHR(instance, &create_info, nullptr, &surface));
-}
+void vkdevice::create_swapchain(const surface_desc& desc, device_surface& surface) {
+    VkSurfaceKHR vk_surface = surface.vk_surface;
 
-void vkdevice::create_swapchain() {
-    VkBool32 supported = VK_FALSE;
     auto&    queue     = queues[static_cast<uint32_t>(QueueType::GRAPHICS)];
-    VKCHECK(vkGetPhysicalDeviceSurfaceSupportKHR(physical_device, queue.index, surface, &supported));
 
-    assert(supported);
+    {
+        VkBool32 supported = VK_FALSE;
+        VKCHECK(vkGetPhysicalDeviceSurfaceSupportKHR(physical_device, queue.index, vk_surface, &supported));
+
+        assert(supported);
+    }
 
-    VkSurfaceCapabilitiesKHR surface_capabilities = {};
-    VKCHECK(vkGetPhysicalDeviceSurfaceCapabilitiesKHR(physical_device, surface, &surface_capabilities));
+    VkSurfaceCapabilitiesKHR surface_capabilities;
+    VKCHECK(vkGetPhysicalDeviceSurfaceCapabilitiesKHR(physical_device, vk_surface, &surface_capabilities));
 
     VkExtent2D extent = surface_capabilities.currentExtent;
-
     if (extent.width < surface_capabilities.minImageExtent.width) {
         extent.width = surface_capabilities.minImageExtent.width;
     } else if (extent.width > surface_capabilities.maxImageExtent.width) {
@@ -699,65 +690,68 @@ void vkdevice::create_swapchain() {
         extent.height = surface_capabilities.maxImageExtent.height;
     }
 
-    auto supported_formats_count = 0U;
-    VKCHECK(vkGetPhysicalDeviceSurfaceFormatsKHR(physical_device, surface, &supported_formats_count, nullptr));
+    {
+        auto supported = false;
+        auto supported_formats_count = 0U;
+        VKCHECK(vkGetPhysicalDeviceSurfaceFormatsKHR(physical_device, vk_surface, &supported_formats_count, nullptr));
 
-    std::vector<VkSurfaceFormatKHR> supported_formats{ supported_formats_count };
-    VKCHECK(vkGetPhysicalDeviceSurfaceFormatsKHR(physical_device, surface, &supported_formats_count, supported_formats.data()));
+        std::vector<VkSurfaceFormatKHR> supported_formats{ supported_formats_count };
+        VKCHECK(vkGetPhysicalDeviceSurfaceFormatsKHR(physical_device, vk_surface, &supported_formats_count, supported_formats.data()));
 
+        for (auto& surface_format : supported_formats) {
+            if (surface_format.format == desc.surface_format.format &&
+                surface_format.colorSpace == desc.surface_format.colorSpace) {
+                supported = true;
+                break;
+            }
+        }
+
+        assert(supported);
+    }
+
+    {
+        auto supported = false;
+        auto supported_present_modes_count = 0U;
+        VKCHECK(vkGetPhysicalDeviceSurfacePresentModesKHR(physical_device, vk_surface, &supported_present_modes_count, nullptr));
 
-    auto supported_present_modes_count = 0U;
-    VKCHECK(vkGetPhysicalDeviceSurfacePresentModesKHR(physical_device, surface, &supported_present_modes_count, nullptr));
+        std::vector<VkPresentModeKHR> supported_present_modes{ supported_present_modes_count };
+        VKCHECK(vkGetPhysicalDeviceSurfacePresentModesKHR(physical_device, vk_surface, &supported_present_modes_count, supported_present_modes.data()));
 
-    std::vector<VkPresentModeKHR> supported_present_modes{ supported_present_modes_count };
+        for (auto& present_mode : supported_present_modes) {
+            if (present_mode == desc.present_mode) {
+                supported = true;
+                break;
+            }
+        }
 
-    VKCHECK(vkGetPhysicalDeviceSurfacePresentModesKHR(physical_device, surface, &supported_present_modes_count, supported_present_modes.data()));
+        assert(supported);
+    }
 
-    auto                     min_image_count = (surface_capabilities.minImageCount <= virtual_frames_count) ? virtual_frames_count : surface_capabilities.minImageCount;
+    assert(desc.image_count >= surface_capabilities.minImageCount);
+    assert(desc.image_count <= surface_capabilities.maxImageCount);
 
     VkSwapchainCreateInfoKHR create_info     = {
             .sType                 = VK_STRUCTURE_TYPE_SWAPCHAIN_CREATE_INFO_KHR,
             .pNext                 = nullptr,
             .flags                 = 0U,
-            .surface               = surface,
-            .minImageCount         = min_image_count,
-            .imageFormat           = supported_formats[0].format,
-            .imageColorSpace       = supported_formats[0].colorSpace,
+            .surface               = vk_surface,
+            .minImageCount         = desc.image_count,
+            .imageFormat           = desc.surface_format.format,
+            .imageColorSpace       = desc.surface_format.colorSpace,
             .imageExtent           = extent,
             .imageArrayLayers      = 1U,
-            .imageUsage            = VK_IMAGE_USAGE_TRANSFER_DST_BIT,
+            .imageUsage            = desc.usages,
             .imageSharingMode      = VK_SHARING_MODE_EXCLUSIVE,
             .queueFamilyIndexCount = 0U,
             .pQueueFamilyIndices   = nullptr,
             .preTransform          = VK_SURFACE_TRANSFORM_IDENTITY_BIT_KHR,
             .compositeAlpha        = VK_COMPOSITE_ALPHA_OPAQUE_BIT_KHR,
-            .presentMode           = supported_present_modes[0],
+            .presentMode           = desc.present_mode,
             .clipped               = VK_TRUE,
             .oldSwapchain          = nullptr,
     };
 
-    VKCHECK(vkCreateSwapchainKHR(device, &create_info, nullptr, &swapchain));
-}
-
-void vkdevice::create_virtual_frames() {
-    for (auto index{ 0U }; index < virtual_frames_count; index++) {
-        {
-            VkSemaphoreTypeCreateInfo semaphore_type_create_info{
-                .sType         = VK_STRUCTURE_TYPE_SEMAPHORE_TYPE_CREATE_INFO,
-                .pNext         = nullptr,
-                .semaphoreType = VK_SEMAPHORE_TYPE_TIMELINE,
-                .initialValue  = 0U,
-            };
-
-            VkSemaphoreCreateInfo create_info{
-                .sType = VK_STRUCTURE_TYPE_SEMAPHORE_CREATE_INFO,
-                .pNext = &semaphore_type_create_info,
-                .flags = 0U,
-            };
-
-            VKCHECK(vkCreateSemaphore(device, &create_info, nullptr, &semaphores[index]));
-        }
-    }
+    VKCHECK(vkCreateSwapchainKHR(device, &create_info, nullptr, &surface.vk_swapchain));
 }
 
 // ----------------- vkadapter -----------------
@@ -989,3 +983,16 @@ void vkdevice::create_debug_layer_callback() {
 
     VKCHECK(vkCreateDebugUtilsMessengerEXT(instance, &create_info, nullptr, &debug_messenger));
 }
+
+
+void device_surface::acquire_image_index(handle<device_semaphore> signal_handle) {
+    auto& render_device = vkdevice::get_render_device();
+    auto* device = render_device.get_device();
+    const auto& signal = render_device.get_semaphore(signal_handle);
+
+    vkAcquireNextImageKHR(device, vk_swapchain, -1, signal.vk_semaphore, nullptr, &image_index);
+}
+
+handle<device_texture> device_surface::get_backbuffer_image() {
+
+}