Implement CanvasGroup and CanvasItem clipping

-Allows merging several 2D objects into a single draw operation
-Use current node to clip children nodes
-Further fixes to Vulkan barriers
-Changed font texture generation to white, fixes dark eges when blurred
-Other small misc fixes to backbuffer code.

1 files changed, 4 insertions, 5 deletions

diff --git a/drivers/vulkan/rendering_device_vulkan.cpp b/drivers/vulkan/rendering_device_vulkan.cpp
index 92d7906b6c..913516889e 100644
--- a/drivers/vulkan/rendering_device_vulkan.cpp
+++ b/drivers/vulkan/rendering_device_vulkan.cpp
@@ -2947,7 +2947,7 @@ Error RenderingDeviceVulkan::texture_clear(RID p_texture, const Color &p_color,
 		image_memory_barrier.srcAccessMask = valid_texture_access;
 		image_memory_barrier.dstAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT;
 		image_memory_barrier.oldLayout = src_tex->layout;
-		image_memory_barrier.oldLayout = clear_layout;
+		image_memory_barrier.newLayout = clear_layout;
 
 		image_memory_barrier.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED;
 		image_memory_barrier.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED;
@@ -6795,13 +6795,12 @@ void RenderingDeviceVulkan::compute_list_add_barrier(ComputeListID p_list) {
 
 void RenderingDeviceVulkan::compute_list_end() {
 	ERR_FAIL_COND(!compute_list);
-	const VkPipelineStageFlags dest_stage_mask = VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT | VK_PIPELINE_STAGE_VERTEX_SHADER_BIT | VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT | VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT | VK_PIPELINE_STAGE_TRANSFER_BIT;
 	for (Set<Texture *>::Element *E = compute_list->state.textures_to_sampled_layout.front(); E; E = E->next()) {
 		VkImageMemoryBarrier image_memory_barrier;
 		image_memory_barrier.sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER;
 		image_memory_barrier.pNext = nullptr;
 		image_memory_barrier.srcAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT;
-		image_memory_barrier.dstAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_INDIRECT_COMMAND_READ_BIT;
+		image_memory_barrier.dstAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_TRANSFER_READ_BIT;
 		image_memory_barrier.oldLayout = E->get()->layout;
 		image_memory_barrier.newLayout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL;
 
@@ -6815,7 +6814,7 @@ void RenderingDeviceVulkan::compute_list_end() {
 		image_memory_barrier.subresourceRange.layerCount = E->get()->layers;
 
 		// TODO: Look at the usages in the compute list and determine tighter dst stage and access masks based on some "final" usage equivalent
-		vkCmdPipelineBarrier(compute_list->command_buffer, VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, dest_stage_mask, 0, 0, nullptr, 0, nullptr, 1, &image_memory_barrier);
+		vkCmdPipelineBarrier(compute_list->command_buffer, VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, VK_PIPELINE_STAGE_VERTEX_SHADER_BIT | VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT | VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT | VK_PIPELINE_STAGE_TRANSFER_BIT, 0, 0, nullptr, 0, nullptr, 1, &image_memory_barrier);
 
 		E->get()->layout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL;
 	}
@@ -6825,7 +6824,7 @@ void RenderingDeviceVulkan::compute_list_end() {
 #ifdef FORCE_FULL_BARRIER
 	_full_barrier(true);
 #else
-	_memory_barrier(VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, dest_stage_mask, VK_ACCESS_SHADER_WRITE_BIT, VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT | VK_ACCESS_INDEX_READ_BIT | VK_ACCESS_VERTEX_ATTRIBUTE_READ_BIT | VK_ACCESS_TRANSFER_READ_BIT | VK_ACCESS_TRANSFER_WRITE_BIT | VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_INDIRECT_COMMAND_READ_BIT, true);
+	_memory_barrier(VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, VK_PIPELINE_STAGE_VERTEX_SHADER_BIT | VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT | VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT | VK_PIPELINE_STAGE_VERTEX_INPUT_BIT | VK_PIPELINE_STAGE_TRANSFER_BIT | VK_PIPELINE_STAGE_DRAW_INDIRECT_BIT, VK_ACCESS_SHADER_WRITE_BIT, VK_ACCESS_INDEX_READ_BIT | VK_ACCESS_VERTEX_ATTRIBUTE_READ_BIT | VK_ACCESS_TRANSFER_READ_BIT | VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_INDIRECT_COMMAND_READ_BIT, true);
 #endif
 }