RTEMS Filesystem check

Chris Johns chrisj at rtems.org
Wed Dec 4 23:35:24 UTC 2013


On 1/12/2013 10:55 am, Wendell Silva wrote:
> Hello,
>
> I'm testing RTEMS RFS file system using qemu (pc386) and huge virtual
> disk file (10G).
> I've created a rfs partition with 8GB with a block size = 4096, which
> allows for huge single file (~ 21 GiB).
>
> The test program (adapted from fileio example) uses synchronized cache
> and create large files (~1GiB). After creation, the file is deleted and
> during the deletion I just close the qemu window in order to simulate a
> system crash an assess the impact of such event on the file system
> integrity.
>
> After repeating this (destructive) test lots of times, I've found that
> rfs survives well, but the free disc space is inconsistent because of
> the non-deleted (marked as free) blocks.
>
> The question is: is there any program/example on how to "sanitize" an
> RFS filesystem, in order the recover these lost/orphans blocks?
> Something like checkdisk.
>
> If such code does not exist, can you provide me directions on how to do
> that? I think this could be an interesting contribution.
>

There is no fsck.rfs code and it would be welcome.

I need to revisit the specifics of the file system to provide you with 
details on what needs to happen here.

Chris



More information about the users mailing list