This paves the way for accelerants to map the framebuffer in user memory, not kernel memory. It uses the ServerMemoryAllocator in order to avoid mapping the framebuffer multiple times in client applications. While at it, clean up some names. Tested with DirectWindowStars and GLTeapot, both work. Change-Id: I4e0c003fae99891044af0a5e6e1d6a4506da6430 Reviewed-on: https://review.haiku-os.org/c/haiku/+/10571 Tested-by: Commit checker robot <[email protected]> Reviewed-by: waddlesplash <[email protected]>