[PATCH 4/4] UBI: replace MTD_UBI_BEB_LIMIT with user-space parameter

[Richard Genoud]

2012/8/16 Shmulik Ladkani <shmulik.ladkani at gmail.com>:
> On Thu, 16 Aug 2012 16:28:38 +0300 Artem Bityutskiy <dedekind1 at gmail.com> wrote:
>> On Thu, 2012-08-16 at 11:57 +0300, Shmulik Ladkani wrote:
>> >
>> > For the simplest systems (those having one ubi device) that need a
>> > limit
>> > *other* than the default (20 per 1024), they can simply set the config
>> > to their chosen value, as they were used to.
>> >
>> > With you approach, these system MUST pass the limit parameter via the
>> > ioctl / module-parameter.
>>
>> Yeah, when you change the default, you usually need to do an extra step.
>> It does not feel too bad, and I would not keep additional configuration
>> option for a hypothetical user. If someone suffers, we can add an option
>> to change the default. But I'd start without it. So, I think it is OK to
>> remove this.
>
> Yes, but the main drawback I was referring to is those platforms already
> setting MTD_UBI_BEB_RESERVE other than the default, by means of kernel
> configuration.
> (there's one platform known to do so in its defconfig, that's
> sam9_l9260_defconfig, which uses 3% instead of the "standard" 2%).
I found the board:
https://www.olimex.com/dev/sam9-L9260.html
and the nand datasheet:
http://www.rockbox.org/wiki/pub/Main/LyrePrototype/K9xxG08UXM.pdf
page 11, we can see that the max_bad_bebper1024 is 25 (100 for 4096)

> These platforms must now change their usermode code to either pass a
> module parameter during the insmod or change their attach ioctl of
> their application.
>
> We force these systems to change their usermode because we changed ubi's
> default BEB limit to be 20/1024 _hardcoded_ (instead of kernel
> configurable as previously was).
>
> Is this ok?
well, I don't know, Artem, you're the maintainer :)
I made a quick patch on this, if you decide to apply it.

Richard


-- 
for me, ck means con kolivas and not calvin klein... does it mean I'm a geek ?