Re: VM patches in 2.6.4-rc1-mm2

[Mike Fedyk]

Nick Piggin wrote:
> Mike Fedyk wrote:
>
> > Nick Piggin wrote:
> > > Mainline doesn't put enough pressure on slab with highmem systems. This
> > > creates a lot more ZONE_NORMAL pressure and that causes swapping.
> >
> > Yep, saw that.  Especially with 128MB Highmem (eg, 1G RAM)
> >
> > > Now with the 2.6 VM, you don't do any mapped memory scaning at all
> >
> >
> >
> > You mean 2.6-mm?
>
> Yes, either mm or linus.

Have there been any VM patches merged into mainline?  Or are you saying 
that the imbalance in mainline would be enough to overcome to lack of 
scanning of mapped pages?

> > > while you only have a small amount of memory pressure. This means that
> > > truely inactive mapped pages never get reclaimed.
> >
> > If I have enough pressure, they will be eventually?  But my caches 
> > will still be smaller than optimal, right?
>
> If you get a lot of pressure at one time it should push out your
> inactive mapped pages. Will get most of the really inactive ones,
> but it won't help pages becoming inactive in future.

Ok, I see.  This might be happening, since it is steadily getting more 
into swap.

> > > The patches you are using do not address this. My split active list
> > > patches should do so. Alternatively you can increase
> > > /proc/sys/vm/swappiness, but that isn't a complete solution, and might
> > > make things too swappy. It is a difficult beast to control.
> >
> >
> >
> > Has akpm said that he would be including the active split patch in -mm?
>
> Hasn't looked at it much. Probably not until some of the more basic
> VM patches can get merged into -linus.

Yes, I wonder if the VM patches helped -mm in the reaim tests...

Let's get the fsfaz (free slab for all zones) into mainline asap! :-D

> > Do you have a patch against -mm (you wrote to ask for your latest...)?
>
> Yep...

Let me get back to you sometime next week.  So far, the VM is reacting 
ok with this combined workload.

Mike
-
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/