[RFC 4/5] mtd: move mtd_oob_mode_t to shared kernel/user space

[Brian Norris]

On Mon, Aug 22, 2011 at 10:30 PM, Artem Bityutskiy <dedekind1 at gmail.com> wrote:
> 32/64 problems - what do you mean? enumeration = 'int' which is always
> 32 bits in all architectures Linux supports.

Oops, got things mixed up here (ints and pointers). Thanks for the
correction. My worry was unfounded...

> 3. IMHO, and I even dare to say that many kernel gurus would agree, if
> we are talking about a simple enum with just few constants inside -
> anonymous enum is the best - the code is simpler and more readable.
...
> So, I'd be happy with 3, can live with 2, and dislike 1 very much :-)

3 sounds perfect, now that I've been straightened out.

> I cannot refer to a section in a standard, but I am sure unitialized
> tags will start with 0 reliably. But yes, it is much less error-prone to
> initialize the tags explicitly, because it makes it a bit more difficult
> to make a stupid mistake by adding a new constant and change values of
> other constants.

It's also useful if we have to kill an old option at some point. I'll
probably include initializers.

>> Sorry if this is a lot of thinking out loud here :)
>
> Yeah, least significant topics usually cause most discussions - I was
> told recently by a colleague that this is called "bike shedding". Notice
> how much I wrote comparing to more important posts of yours :-)

I actually learned/recalled a few things here; I had pieces of the
puzzle but not the whole picture. And the "bike shed" is an important
addition to my vocabulary.

Thanks,
Brian