If there are still some cases where we wait when we could avoid to do so, we should "fix" those.
Well, looks like index buffers are still mapped through CS. Can we use the chunk allocator for them?
Only when using the GL backend (there is no such limitation with the Vulkan backend). We can't trivially enable suballocation for index buffers on GL because there is no GL API that allows to offset the index buffer, specifically WRT indirect draws.