Hello
If a ktx image has k mipLevels,
why do we have only one imageMemoryBarrier variable having k mipLevels inside
and why do we have separate k bufferCopyRegions, each of which has one mipLevel inside ?
Could you please explain to me
why do we do just like this ?
below is a source example
VkImageMemoryBarrier imageMemoryBarrier = vks::initializers::imageMemoryBarrier();
....
imageMemoryBarrier.subresourceRange.baseMipLevel = 0;
imageMemoryBarrier.subresourceRange.levelCount = mipLevels;
....
vkCmdPipelineBarrier(
cmdbuffer,
srcStageMask,
dstStageMask,
0,
0, nullptr,
0, nullptr,
1, &imageMemoryBarrier); <=== only one imageMemoryBarrier
std::vector<VkBufferImageCopy> bufferCopyRegions; <=== separate bufferCopyRegions
for (uint32_t i = 0; i < mipLevels; i++)
{
VkBufferImageCopy bufferCopyRegion = {};
bufferCopyRegion.imageSubresource.mipLevel = i;
........
bufferCopyRegions.push_back(bufferCopyRegion);
}
vkCmdCopyBufferToImage(copyCmd, stagingBuffer, image, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, static_cast<uint32_t>(bufferCopyRegions.size()), bufferCopyRegions.data()); <=== separate bufferCopyRegions
and
another vkCmdPipelineBarrier