Update FAQs 10.21 and 10.22

[pauamma_gundo.com]

Summary

• remove mentions of cache queue
• introduce laundry queue
• copyedit definition of free memory in FAQ 10.22

PR: 242062

Test Plan

tested with lynx

[debdrup]

Syntactically and lexicographically, this makes sense to me.
As to whether it matters kernel-defined behavior, that's outside my area of expertise.

[markj]

en_US.ISO8859-1/books/faq/book.xml

4029 ↗

(On Diff #70742)

Let's remove the "in the interrupt or process states" qualifier. It is referring to a mechanism we use in the page allocator to reserve free pages for high-priority allocations, but that is just an implementation detail and no longer has anything to do with the "state" being referred to.

4040 ↗

(On Diff #70742)

"active or inactive"

4045 ↗

(On Diff #70742)

Let's replace "whether they are active or inactive" with "regardless of their location in the page queues" or something similar.

4046 ↗

(On Diff #70742)

s/inactive/laundry/

And perhaps update the description of the inactive queue:

"In most common cases, it is best to think of the laundry queue to be a queue of relatively unused, dirty pages that might or might not be in the process of being written to disk. The inactive queue contains a mix of clean and dirty pages; clean pages near the head of the queue are reclaimed immediately to alleviate a free page shortage, and dirty pages are moved to the laundry queue for deferred processing."

Feel free to reword, this is just a suggestion.

4049 ↗

(On Diff #70742)

Again let's remove this qualifier, or perhaps even the whole sentence.

4068 ↗

(On Diff #70742)

Maybe qualify with "some other reserved pages".

[pauamma_gundo.com]

This should address all of markj's comments.

[markj]

The patch looks reasonable to me. Thanks for doing this!

[pauamma_gundo.com]

Thanks for reviewing it. Can a docs committer (I'm not one) take care of it?