Vulkan VMA使用致黑屏问题排查：uniform缓冲区配置错误

我在应用中集成Vulkan Memory Allocator（VMA）后出现黑屏，采用延迟渲染+HDR管线，所有附件均通过VMA分配。测试时片段着色器输出纯色显示正常，最初推测是顶点/索引缓冲区为空。

尝试的顶点缓冲区填充方案 #

• 使用暂存缓冲区复制数据：

void* data;
vmaMapMemory(_allocator, stagingAllocation, &data);
memcpy(data, bufferData, bufferSize);
vmaUnmapMemory(_allocator, stagingAllocation);

• 跳过暂存缓冲区，直接配置标志后写入：

allocCreateInfo.flags = VMA_ALLOCATION_CREATE_HOST_ACCESS_SEQUENTIAL_WRITE_BIT | VMA_ALLOCATION_CREATE_MAPPED_BIT;

memcpy(allocInfo.pMappedData, bufferData, bufferSize);

完整缓冲区创建代码 #

顶点缓冲区创建流程 #

VkBuffer stagingBuffer;
VkBuffer vertexBuffer;

VmaAllocation stagingAllocation = createBuffer(bufferSize, VK_BUFFER_USAGE_TRANSFER_SRC_BIT, VMA_MEMORY_USAGE_CPU_ONLY, stagingBuffer);

void* data;
vmaMapMemory(_allocator, stagingAllocation, &data);
memcpy(data, bufferData, bufferSize);
vmaUnmapMemory(_allocator, stagingAllocation);

VmaAllocation allocation = createBuffer(bufferSize, VK_BUFFER_USAGE_TRANSFER_DST_BIT | VK_BUFFER_USAGE_VERTEX_BUFFER_BIT, VMA_MEMORY_USAGE_GPU_ONLY, vertexBuffer);

copyBuffer(stagingBuffer, vertexBuffer, bufferSize);

vmaDestroyBuffer(_allocator, stagingBuffer, stagingAllocation);

createBuffer函数实现 #

VmaAllocation createBuffer(VkDeviceSize size, VkBufferUsageFlags usage, VmaMemoryUsage memoryUsage, VkBuffer &buffer) {
  VkBufferCreateInfo bufferInfo{};
  bufferInfo.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO;
  bufferInfo.size = size;
  bufferInfo.usage = usage;
  bufferInfo.sharingMode = VK_SHARING_MODE_EXCLUSIVE;
  
  VmaAllocationCreateInfo allocCreateInfo = {};
  allocCreateInfo.usage = memoryUsage;
  
  VmaAllocation allocation;
  if (vmaCreateBuffer(_allocation, &bufferInfo, & allocCreateInfo, &buffer, &allocation, nullptr) != VK_SUCCESS) {
    throw std::runtime_error("Failed to create buffer");
  }

  return allocation;
}

其他异常现象 #

• 若将allocCreateInfo.flags设为0，会出现奇怪的闪烁
• 切换到VMA前程序运行正常，仅修改了缓冲区和镜像创建代码（镜像部分无异常）
• 使用环境：MacOS + Vulkan 1.2

编辑1：初步定位问题 #

经过数小时调试，发现问题并非顶点缓冲区，而是uniform缓冲区配置错误：原本使用VMA_MEMORY_USAGE_CPU_TO_GPU标志，改为VMA_MEMORY_USAGE_CPU_ONLY后程序可正常运行，但这只是临时解决方案，不希望将所有uniform缓冲区存储在CPU内存中，求问问题根源。

编辑2：最终解决方案 #

最终发现，当设置usage为VMA_MEMORY_USAGE_CPU_TO_GPU时，VMA不会自动启用VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT或VK_MEMORY_PROPERTY_HOST_COHERENT_BIT标志，必须手动在allocCreateInfo.requiredFlags中添加这两个标志，添加后程序运行正常。但未在VMA文档中找到相关说明，调试过程十分困难。

内容的提问来源于stack exchange，提问作者Samo77