I am developing the peer to peer filesystem using fuse.
I am confused with whether to use the various data structures like
directory layout, inodes, superblock or not.
I would like to know whether to follow the Bach Algorithms
Email me at :dnyanesh.hande@...
From: Goswin von Brederlow <goswin-v-b@we...> - 2010-01-30 12:47:02
Dnyanesh Hande <dnyanesh.hande@...> writes:
> I am developing the peer to peer filesystem using fuse.
> I am confused with whether to use the various data structures like
> directory layout, inodes, superblock or not.
> I would like to know whether to follow the Bach Algorithms
> Email me at :dnyanesh.hande@...
Nothing is really neccessary. You only need some way to get or fake the
information needed for getattr, readdir, read, ... You could even write
a filesystem that doesn't support readdir. Personally I think it is a
good idea to have a creation time, filesize, readable filename,
directory structure in a filesystem but that is up to you.
If you do have an on-disk structure for your FS, some cache of
downloaded stuff or files you are seeding then I would put some form of
identification there. A Superblock at the beginning of the disk or file
or a specifically named file if it is a directory. Record your
filesystem name and version in there and any other metadata that should
be kept between mounts, e.g. how much data was up/downloaded, peers you
had connections too, ...
Get latest updates about Open Source Projects, Conferences and News.