On Fri, 12 Nov 1999, Keith Owens wrote:
|>On Thu, 11 Nov 1999 22:59:49 -0800 (PST),
|>Matt Robinson <yakker@xxxxxxxxxxxxxxxxxxxx> wrote:
|>>Cool. It'd be nice to utilize the same thing for 2.2.X...
|>
|>Which brings up a question. Why bother with lkcd for 2.2. kernels?
|>IMHO, there is little or no chance that a change of this magnitude will
|>be accepted into the stable, production kernel tree. Especially when
|>the 2.3 freeze followed by the release of 2.4 is getting ever closer.
|>If you concentrate on 2.2, you will only have to redo all the work for
|>2.3, taking into account the 2.3 kernel changes. You might be better
|>off concentrating on 2.3.
We intend to concentrate getting this pushed into the 2.3 tree; however,
we also want to make sure older releases that are going out in more
commercial environments also have as many extended features as possible.
It's more an issue with the direction and concentration of porting effort.
We can do our work in 2.3, sure; but we can't neglect 2.2 for customers
that want this feature and a stable kernel as well.
As a general rule, I agree with you.
|>I know that most distributions are using 2.2 kernels but that is
|>irrelevant for a patch like this. Users who just install from a
|>distribution and run it will not bother getting a new kernel and they
|>definitely will not apply a 26,000 line patch. The sort of user who
|>will apply a big patch, install new programs, change init scripts etc.
|>is almost certainly running 2.3 kernels already. Get the patch into
|>2.3 now so it appears on distributions when 2.4 comes out.
Let's hope Linus lets it in ... We'd love to make it a part of the
general release as soon as possible. And we'd do most of our work
in the 2.3 tree ... Linus?
--Matt
|