[ELDK] [RFSB] mkfs.jffs2 build options

Ron Madrid ron_madrid at sbcglobal.net
Thu Jun 24 22:31:44 CEST 2010


--- On Thu, 6/24/10, Wolfgang Denk <wd at denx.de> wrote:

> From: Wolfgang Denk <wd at denx.de>
> Subject: Re: [ELDK] [RFSB] mkfs.jffs2 build options
> To: "Ron Madrid" <ron_madrid at sbcglobal.net>
> Cc: eldk at lists.denx.de
> Date: Thursday, June 24, 2010, 1:13 PM
> Dear Ron Madrid,
> In message <870334.93379.qm at web83506.mail.sp1.yahoo.com>
> you wrote:
> > Whenever I've built my jffs2 image (manually) I have
> > used a few options that are currently not supported
> by
> > the RFSB.  I am guessing because no one has
> bothered to
> > add these capabilities up to this point.  The
> options
> > that I have been using are -n, --pad, and -s. 
> The fact
> > that these aren't included makes me think that
> perhaps
> > I'm using these options unnecessarily.
> I guess it's more that JFFS2 is not being used that often
> any more.
> It was OK on NOR flash (and the typoical partition sizes we
> had at
> these times), but with NAND it becomes obsolete: it does
> not address
> the issue of NAND block going bad just by reading them, and
> it does
> not scale with typical partition sizes we have now. 
> Did you compare
> YAFFS and UBI/UBIFS (especially UBI is highly recommended
> from the
> reliability point of view).

I saw the information regarding the YAFFS and UBI/UBIFS and
will probably migrate over to that in the future.  Of
course that decision hinges on available resources.  In the
meantime, however, I'm just trying to implement the RFSB to
match our current configuration.  Thanks for the input.
> > If it is simply the case that no one has needed them
> > until now I'd be happy to supply a patch to add these
> > options.  Any feedback would be great.
> Patches are always welcome.

I think I'll send in a patch to implement these arguments
and we'll go from there.


More information about the eldk mailing list