* [LSF/MM/BPF TOPIC] Page cache tracking with the Maple Tree
@ 2026-02-24 17:10 Liam R. Howlett
0 siblings, 0 replies; only message in thread
From: Liam R. Howlett @ 2026-02-24 17:10 UTC (permalink / raw)
To: lsf-pc, linux-mm, linux-fsdevel
Cc: Matthew Wilcox, Johannes Weiner, David Hildenbrand, Jan Kara,
Ryan Roberts, Christian Brauner
Hi,
The page cache currently uses a radix tree to store folios. There are several
well-known deficiencies to this approach:
- Inefficient representation of large folios
- Inefficient representation of sparsely accessed files
- Poor cache locality
- Supports 3 search marks
The Maple Tree solves these problems more efficiently than the XArray. In this
session, I'd like to discuss what needs to happen to change from using the
XArray to the Maple Tree for the page cache.
The Maple Tree needs some enhancements:
- New node type to efficiently support entries of length of 1
- New node type to support marks
- Searching and iterating over marks
- Support for purging shadow entries from the page cache
The motivation, beyond potential performance gains and memory footprint,
is to support more features natively in the data structure.
Thanks,
Liam
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2026-02-24 17:10 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-02-24 17:10 [LSF/MM/BPF TOPIC] Page cache tracking with the Maple Tree Liam R. Howlett
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox