Re: [PATCH] Fix the zone reclaim code in 2.6.15

From: Andrew Morton
Date: Tue Jan 03 2006 - 19:42:58 EST


Christoph Lameter <clameter@xxxxxxxxxxxx> wrote:
>
> On Tue, 3 Jan 2006, Andrew Morton wrote:
>
> > Christoph Lameter <clameter@xxxxxxxxxxxx> wrote:
> > >
> > > Some bits for zone reclaim exists in 2.6.15 but they are not usable.
> > > This patch fixes them up, removes unused code and makes zone reclaim usable.
> > >
> >
> > You know that there are over 100 mm/ patches in -mm. If Linus applies this
> > patch, it will cause extensive wreckage. And this patch doesn't vaguely
> > apply to the mm/ patches which I have queued. So it's basically useless.
> >
> > Please try to play along.
>
> Well, this is one case where there is crap in Linus tree that needs to be
> fixed. Its an urgent issue. And the existing patches in mm do not fix
> this issue but remove the code altogether. When I asked you to remove
> these patches, you got mad at me for some reason.

The code in Linus's tree is dead and is unused.

> > yes, it's awkward that there's such a large backlog in that area. We just
> > need to be patient and extra careful.
>
> So you are saying we need to remove this feature and then add it back in
> later?

Yes. It's cleaner to remove all the dead stuff and to then reintroduce the
new feature as a clean, standalone thing. So we can see its effects ore
clearly.

Also, what happens if for some reason we later want to revert your patch?
Such a reversion would reintroduce all the old dead stuff!

Patches should each do a single thing.

> This means another 2 and 3 release cycles with this severe unfixed
> problem?

There's plenty of time to get this into 2.6.16.

Your changelog didn't describe this as a "severe" problem. Things have
been like this for quite some time, haven't they?

And we shouldn't be merging new, untested things ahead of work which has
had considerable amounts of testing.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/