Re: [PATCH] mm, swap: Make VMA based swap readahead configurable

From: huang ying
Date: Fri Oct 06 2017 - 08:29:06 EST

On Tue, Oct 3, 2017 at 5:49 AM, Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
> On Mon, 2 Oct 2017 08:45:40 -0700 Dave Hansen <dave.hansen@xxxxxxxxx> wrote:
>> On 09/27/2017 06:02 PM, Huang, Ying wrote:
>> > I still think there may be a performance regression for some users
>> > because of the change of the algorithm and the knobs, and the
>> > performance regression can be resolved via setting the new knob. But I
>> > don't think there will be a functionality regression. Do you agree?
>> A performance regression is a regression. I don't understand why we are
>> splitting hairs as to what kind of regression it is.
> Yes.
> Ying, please find us a way of avoiding any disruption to existing
> system setups. One which doesn't require that the operator perform a
> configuration change to restore prior behaviour/performance.

Sorry for late. I am in holiday recently.

OK. For me, I think the most clean way is to use page_cluster to
control both the virtual and physical swap readahead. If you are OK
with that, I will prepare the patch.

> And please let's get this done well in advance of the 4.14 release.


Best Regards,
Huang, Ying