Re: Scalability fixes -- 2.6.31 candidate?

[Balbir Singh]

* Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> [2009-08-20 16:13:25]:

> On Fri, 21 Aug 2009 00:39:42 +0530
> Balbir Singh <balbir@xxxxxxxxxxxxxxxxxx> wrote:
> 
> > Hi, Andrew,
> > 
> > I've been wondering if the scalability fixes for root overhead in
> > memory cgroup is a candidate for 2.6.31?
> 
> These?
> 
> memcg-improve-resource-counter-scalability.patch
> memcg-improve-resource-counter-scalability-checkpatch-fixes.patch
> memcg-improve-resource-counter-scalability-v5.patch
> 
> 
> > They don't change
> > functionality but help immensely using existing accounting features.
> > 
> > Opening up the email for more debate and discussion and thoughts.
> > 
> 
> They don't apply terribly well to mainline:
> 
> patching file mm/memcontrol.c
> Hunk #1 FAILED at 70.
> Hunk #2 FAILED at 479.
> Hunk #3 FAILED at 1295.
> Hunk #4 FAILED at 1359.
> Hunk #5 FAILED at 1432.
> Hunk #6 FAILED at 1514.
> Hunk #7 FAILED at 1534.
> Hunk #8 FAILED at 1605.
> Hunk #9 FAILED at 1798.
> Hunk #10 FAILED at 1826.
> Hunk #11 FAILED at 1883.
> Hunk #12 FAILED at 1981.
> Hunk #13 succeeded at 2091 (offset -405 lines).
> 12 out of 13 hunks FAILED -- saving rejects to file mm/memcontrol.c.rej
> Failed to apply memcg-improve-resource-counter-scalability
> 
> so maybe you're referring to these:
> 
> memcg-remove-the-overhead-associated-with-the-root-cgroup.patch
> memcg-remove-the-overhead-associated-with-the-root-cgroup-fix.patch
> memcg-remove-the-overhead-associated-with-the-root-cgroup-fix-2.patch
> 
> as well.
>

Yes, I was referring to those
 
> But then memcg-improve-resource-counter-scalability.patch still doesn't
> apply.  Maybe memcg-improve-resource-counter-scalability.patch depends
> on memory-controller-soft-limit-*.patch too.  I stopped looking.
> 

Yes, there is some diffs that get picked up due to the soft_limit
feature.


> It's a lot of material and a lot of churn.  I'd be more inclined to
> proceed with a 2.6.32 merge and then perhaps you can see if you can
> come up with a minimal patchset for -stable, see if the -stable
> maintainers can be talked into merging it.
> 

Fair enough.. I do have a backport to 2.6.31-rc5 mainline, but going
the stable route would also work.

Thanks!


-- 
	Balbir
--
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/