File System Implementation

[Greg Freemyer]

Historically filesystem designers knew what legal values of the
various parameters were and they simply looked for legal values.  If
they found any illegal values, they reset them to a legal value.

In many cases, that missed bit-rot situations they wanted to find, so
with many modern filesystems, they add a CRC field to the main
datastructures, then do a CRC verification as appropriate.

XFS is right now in the middle of getting a whole bunch of CRC
functionality added.  I don't know if it is on the generation side, or
just on the verification side.
A quick google found this overview email to a patch series.  The
series may have been posted again since then.

http://comments.gmane.org/gmane.comp.file-systems.xfs.general/51575

Notice there are links to the 3 previous overview emails.  Go back and
read those, they may provide more background.

Greg

On Wed, May 29, 2013 at 10:20 AM, Dibyayan Chakraborty
<dib.coolguy at gmail.com> wrote:
> Hi,
> I was implementing a file system and ran into the following problem.
>
> I wanted to replicate the super block data in order to recover from
> occasional Hard disk failures (bit rot problem etc.). While replicating was
> easy i could not find a way to detect the errors in the first place.
>
> Any ideas and explanation will be help full.
>
> --
> With Regards
> Dibyayan Chakraborty
>
> _______________________________________________
> Kernelnewbies mailing list
> Kernelnewbies at kernelnewbies.org
> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>