JFFS2: Difference between revisions
Ian Bicking (talk | contribs) (added note about files.) |
(Added "Measuring disk usage" section, with snippet from recent thread.) |
||
Line 24: | Line 24: | ||
This is a simple algorithm that often manages to squeeze extra few |
This is a simple algorithm that often manages to squeeze extra few |
||
bytes from data already compressed with gzip |
bytes from data already compressed with gzip |
||
==Measuring disk usage== |
|||
From http://lists.laptop.org/pipermail/library/2007-July/000070.html |
|||
JFFS2 compresses 4K chunks using zlib. So it's not just per-file |
|||
compression, it's compressing bits of a file. It doesn't compress files |
|||
where compression doesn't help. And there's a 68 byte overhead per |
|||
compressed chunk. Plus probably some fixed overhead per file. |
|||
Running "mkfs.jffs2 --root DIR | wc" gives a more accurate picture of |
|||
the size of a directory. Things might change as the compression |
|||
parameters of JFFS2 are changed; specifically LZO compression is better |
|||
than zlib, and there might be an attribute to disable compression on a |
|||
particular file. |
|||
I wrote a little Python script to estimate the size and compare it to |
|||
the actual size from mkfs.jffs2 (if you have that available on your |
|||
system), and to the original size. For small files (500 bytes - 1.5K) |
|||
it's compressed to 50% of the size, for Python source code 40% of the |
|||
size (with .pyc files), the Gnome 2 User Guide (with lots of images) |
|||
gets to about 65% (35% reduction). |
|||
Script at: http://svn.colorstudy.com/home/ianb/olpc/jffs2size.py |
|||
== Links == |
== Links == |
Revision as of 15:20, 26 July 2007
JFFS2 is the Journaling Flash File System version 2, jffs2 is a log-structured file system for flash memory devices and has support for NAND devices, also Hard links and compression algorithms like zlib, rubin, and rtime.
Jffs2 also has a garbage collection algorithm that eliminates unnecessary I/O.
Size of Files
Files in JFFS2 are compressed transparently. This removes some of the need for gzipping and otherwise compressing files to save space. For text files compression rates of 50% or more are possible. du
will not show the compressed sizes of files (nor will ls -l
).
More discussion is available in this email post.
Algorithms
Zlib
Lossless data-compression library,
rubin
This simple algorithm is faster than zlib but not quite as effective.
rtime
This is a simple algorithm that often manages to squeeze extra few bytes from data already compressed with gzip
Measuring disk usage
From http://lists.laptop.org/pipermail/library/2007-July/000070.html
JFFS2 compresses 4K chunks using zlib. So it's not just per-file compression, it's compressing bits of a file. It doesn't compress files where compression doesn't help. And there's a 68 byte overhead per compressed chunk. Plus probably some fixed overhead per file. Running "mkfs.jffs2 --root DIR | wc" gives a more accurate picture of the size of a directory. Things might change as the compression parameters of JFFS2 are changed; specifically LZO compression is better than zlib, and there might be an attribute to disable compression on a particular file. I wrote a little Python script to estimate the size and compare it to the actual size from mkfs.jffs2 (if you have that available on your system), and to the original size. For small files (500 bytes - 1.5K) it's compressed to 50% of the size, for Python source code 40% of the size (with .pyc files), the Gnome 2 User Guide (with lots of images) gets to about 65% (35% reduction). Script at: http://svn.colorstudy.com/home/ianb/olpc/jffs2size.py
Links
- JFFS2 at Wikipedia
- JFFS at Wikipedia
- Summary on JFFS
- Homepage of JFFS at AXIS
- Logfs