"Squash FS" is the name of a compressed read-only filesystem for Linux. There are a number of such file systems available for Linux, including ROMFS, CramFS and SquashFS.
A compressed file system is interesting in embedded systems for reducing the overall size (in flash) of the Linux system. Squash FS is reported to have better compression capabilities than CramFS, which is a very popular.
The SquashFS home page is at: Squash FS
SquashFS version 4.0 was accepted into the mainline kernel in January 2009. See http://lwn.net/Articles/314326/ and the kernel change log at: http://kernel.org/pub/linux/kernel/v2.6/testing/ChangeLog-2.6.29-rc1
As of November 2009, Phillip Lougher was working on adding LZMA support to the file system. See http://old.nabble.com/Re%3A-squashfs-4.0-lzma-support-td25132198.html#a26028786
- See the Squash fs download page for patches
- if you want LZMA support, read below
The squashfs file release contains a README, the squashfs patch files, and the squashfs-tools directory (mksquashfs). Please see the INSTALL file for install instructions.
How To Use
See the Squash FS Howto:
Here are brief summaries for 2 large file systems, saved using a variety of file system types.
This information was provided by Phillip Lougher.
Ubuntu liveCD compression results
ext3 uncompressed size 1.4 GB ISO9660 uncompressed size 1.3 GB Zisofs compressed size 589.81 MB Cloop compressed size 471.89 MB Squashfs2.0 compressed size 448.58 MB Squashfs2.1 compressed size 448.58 MB
Damn Small Linux liveCD compression results
ext3 uncompressed size 126 MB CRAMFS compressed size 52.19 MB Squashfs2.0 compressed size 46.52 MB Squashfs2.1 compressed size 46.52 MB
File System Performance
There is performance data for these file systems on the page Squash Fs Comparisons
LZMA is the name of a compression algorithm that can be used to achieve better compression with SquashFS.
Benefits of LZMA:
- Better compression - easily saving 5% on a file system
- Decompression is slower than zlib
Up until Fall of 2009, support for the LZMA compression algorithm in SquashFS was available via a set of external patches. See http://www.squashfs-lzma.org/
In December of 2009, Phillip Lougher submitted some patches to LKML for LZMA support in SquashFS. See http://lkml.org/lkml/2009/12/7/84
(As of December, 2009...)
To obtain a mksquashfs that supports LZMA, see the Squashfs CVS (http://sourceforge.net/projects/squashfs/develop).
You'll need to edit the squashfs-tools Makefile to enable LZMA support. The comments in the Makefile should, hopefully, explain how to build LZMA support into Mksquashfs/Unsquashfs.
Once built-in, LZMA support can be specified using the -comp lzma option, i.e.
mksquashfs dir dir.img -comp lzma
Unsquashfs doesn't need any extra options, it automatically detects which compression has been used.
You can tell which compression algorithms Mksquashfs/Unsquashfs support by just typing the command on its own (i.e. mksquashfs, or unsquashfs). The (de-)compressors available are displayed at the end of the output.
lzma data comparison
Here is some information that was posted recently (?) on the squashfs mailing list by Oleg Vdovikin:
> # du -s target > 7836 target > # ls -l target.* > -rw-r--r-- 1 root root 2842788 Aug 27 17:54 target.cramfs > -rwx------ 1 root root 2449408 Jan 26 13:19 target.sqshfs > -rwx------ 1 root root 2060288 Jan 26 13:21 target.lzmafs > > So, lzma for this filesystem gives 84% of original size. For bigger > filesystem I've got 82%.
Notes on compression
From: John Richard Moser <nigele...@comcast.net> Date: Wed, 29 Sep 2004 18:20:11 +0200 Subject: Re: Compressed filesystems: Better compression? Matti Aarnio wrote: | Compression algorithms are a bit tough to be used in a random access | smallish blocks environments. In long streams where you can use megabytes | worth of buffer spaces there is no problem is achieving good performance. | But do try to do that in an environment where your maximum block size | is, say: 4 kB, and you have to start afresh at every block boundary. Yes of course. I've seen the compressed page cache patch do this and get fair peformance (10-20%), though on double size blocks (8KiB) it manages almost twice as good (20-50%, averaged around 30% IIRC). Not great, but not bad. On compressed filesystems you can work with 64k or 128k blocks. Somewhere around 32-64k is usually optimal; you're not going to see great improvements using 1M blocks instead of 512k blocks. | Whatever algorithms you use, there will always be data sequences that | are of maximum entropy, and won't compress. Rather they will be | presented in streams as is with a few bytes long wrappers around | them. Yes, an intelligent algorithm decides that if the underlying compression algorithm used produces no results, it just marks the block as uncompressed and stores it as such. ZLIB does this if the block gets bigger. LZMA might not; but higher level intrinsics (block headers) could handle that easy (as you said).