= Compressed Caching =
= For 2.6.x kernels =
=== Nitin Gupta ===
----
 This page gives all the details collected together for compressed cache implementation. Many of the ideas have been taken from Rodrigo’s work and this paper: Adaptive Main Memory Compression by Irina Chihaia, Thomas Gross.

Here I will keep track of project progress. Please add any suggestions you have or edit anything you find incorrect.

----
== Introduction ==
Compressed caching is the introduction of new layer in virtual memory hierarchy -- Compressed Cache. It compresses and stores pages that would otherwise have been swapped to slow disks or freed under memory pressure. This effectively increases RAM space and avoids /reduces accesses to slow disks. This basically takes advantage to rapidly increasing CPU power, faster, lower latency memories and sluggish hard-disk speed improvements.

== Work Done ==
Currently the implementation simply replaces a page with a copy of itself (i.e. no actual compression yet) and replaces its reference in page cache with a ‘chunk_head’ (all this is detailed below). On page cache lookup, the original page is obtained by again copying and setting corresponding page cache entry back to this page instead of ‘chunk_head’. Also, one of compression algorithm (WKdm) has been ported to kernel space. The idea is to first have a framework ready for further implementation and gain more familiarity with VMM code.

== General Background ==
The system maintains two LRU lists – active and inactive LRU lists. These lists may contain both page-cache (file backed) and swap-cache (anonymous) pages. When under memory pressure, pages in inactive list are freed as:

 * Swap-cache pages are written out to swap disksusing ‘swapper_space’ writepage() (swap_writepage()).
 * Dirty page-cache pages are flushed to filesystemdisks using filesystem specific writepage().
 * Clean page-cache pages are simply freed.

For compressed cache to be effective, it needs to store both swap-cache and page-cache (clean and dirty) pages. So, a way is needed to transparently (i.e. changes should be required within VMM subsystem only) take these pages in/out of compressed cache.