On 04/05/11 10:39, Andrea Gasparini wrote:
> Hi Charles,
>
> Charles Manning wrote, Tuesday 03 May 2011:
>> These were cleaned out of the code but not in the Kconfig. Consider it a
>> benign oversight rather than a bug :-).
>
> I see ;)
>
> But, if I'd like to disable yaffs background processing from boot
> arguments, how is it supposed to be done?
>
>
> And just for completeness, I noticed this from my tests on an old
> filesystem flashed with an HWECC (kernel 2.6.18 on DaVinci dm365, and the
> "OOB_FIRST" pattern).
> I can't anymore got yaffs2 working with a newer kernel/yaffs, given that it
> works only with a software ECC. (but it's really slower)
> Did you have any suggestion also on this?
YAFFS doesn't decide which ECC type to use - that's down to the kernel.
I'm using da830/OMAP-L137, and the kernel is using davinci HW ECC with
OOB_FIRST.
See my post on this list "yaffs erases all files on mount" in which I've
been having trouble with the YAFFS tags SW ECC which only protect 16bytes
of tags stored in the OOB. It is not too slow, since the payload is small.
I think the YAFFS patch by default enables this SW ECC, but the old TI
Montavista kernel disabled it. This creates an incompatibility, but you
can turn off the YAFFS ECC to restore the compatibility.
I leave it up to you to decided if that's a good idea or not :)
Nick.
(text/plain)