LRU-ness of VM's page cache

["Kirk True" <ktrue@movaris.com>]

Hi all,

The page replacement policy is said to use a LRU scheme, but this isn't really true. The pages kept on the active list are really in a "RU" list, because there's no notion of "least" at all, right?

Without implying that I think it's a good idea :) how would you even write a true LRU implementation? Wouldn't the kernel need to get a callback to execute code on *every* memory access to keep the page references in the list in strict LRU order?

Kirk
--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive:       http://mail.nl.linux.org/kernelnewbies/
FAQ:           http://kernelnewbies.org/faq/