The Linux kernel (at least before using XArrays as far as I'm aware, which to my knowledge are wrappers around Radix Trees anyway) uses radix trees in its address_space structs which every file has. This allows a file to efficiently find all the dirty buffers which it has written to for easy flushing.
However, what is the advantage of a Radix Tree over a Multi-level Bitmap in this scenario? They seem so much easier to implement, is there a difference?
Asked by Mike
(11 rep)
Nov 4, 2024, 04:08 PM