Re: [PATCH -mm -v4 00/21] mm, THP, swap: Swapout/swapin THP in one piece

From: Andrew Morton
Date: Thu Jun 28 2018 - 02:18:49 EST


On Thu, 28 Jun 2018 13:35:15 +0800 "Huang\, Ying" <ying.huang@xxxxxxxxx> wrote:

> Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> writes:
>
> > On Thu, 28 Jun 2018 13:29:09 +0800 "Huang\, Ying" <ying.huang@xxxxxxxxx> wrote:
> >
> >> Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> writes:
> >>
> >> > On Fri, 22 Jun 2018 11:51:30 +0800 "Huang, Ying" <ying.huang@xxxxxxxxx> wrote:
> >> >
> >> >> This is the final step of THP (Transparent Huge Page) swap
> >> >> optimization. After the first and second step, the splitting huge
> >> >> page is delayed from almost the first step of swapout to after swapout
> >> >> has been finished. In this step, we avoid splitting THP for swapout
> >> >> and swapout/swapin the THP in one piece.
> >> >
> >> > It's a tremendously good performance improvement. It's also a
> >> > tremendously large patchset :(
> >> >
> >> > And it depends upon your
> >> > mm-swap-fix-race-between-swapoff-and-some-swap-operations.patch and
> >> > mm-fix-race-between-swapoff-and-mincore.patch, the first of which has
> >> > been floating about since February without adequate review.
> >> >
> >> > I'll give this patchset a spin in -mm to see what happens and will come
> >> > back later to take a closer look. But the best I can do at this time
> >> > is to hopefully cc some possible reviewers :)
> >>
> >> Thanks a lot for your help! Hope more people can review it!
> >
> > I took it out of -mm again, temporarily. Due to a huge tangle with the
> > xarray conversions in linux-next.
>
> No problem. I will rebase the patchset on your latest -mm tree, or the
> next version to be released?

We need to figure that out with Matthew.

Probably the xarray conversions are simpler and more mature so yes,
probably they should be staged first.