Re: [PATCH] mm: shmem: enable thp migration (Re: [PATCH v1] mm: consider non-anonymous thp as unmovable page)

From: Naoya Horiguchi
Date: Mon Apr 09 2018 - 03:15:49 EST


On Fri, Apr 06, 2018 at 09:08:15AM +0200, Michal Hocko wrote:
> On Fri 06-04-18 05:14:53, Naoya Horiguchi wrote:
> > On Fri, Apr 06, 2018 at 03:07:11AM +0000, Horiguchi Naoya(堀口 直也) wrote:
> > ...
> > > -----
> > > From e31ec037701d1cc76b26226e4b66d8c783d40889 Mon Sep 17 00:00:00 2001
> > > From: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>
> > > Date: Fri, 6 Apr 2018 10:58:35 +0900
> > > Subject: [PATCH] mm: enable thp migration for shmem thp
> > >
> > > My testing for the latest kernel supporting thp migration showed an
> > > infinite loop in offlining the memory block that is filled with shmem
> > > thps. We can get out of the loop with a signal, but kernel should
> > > return with failure in this case.
> > >
> > > What happens in the loop is that scan_movable_pages() repeats returning
> > > the same pfn without any progress. That's because page migration always
> > > fails for shmem thps.
> > >
> > > In memory offline code, memory blocks containing unmovable pages should
> > > be prevented from being offline targets by has_unmovable_pages() inside
> > > start_isolate_page_range(). So it's possible to change migratability
> > > for non-anonymous thps to avoid the issue, but it introduces more complex
> > > and thp-specific handling in migration code, so it might not good.
> > >
> > > So this patch is suggesting to fix the issue by enabling thp migration
> > > for shmem thp. Both of anon/shmem thp are migratable so we don't need
> > > precheck about the type of thps.
> > >
> > > Fixes: commit 72b39cfc4d75 ("mm, memory_hotplug: do not fail offlining too early")
> > > Signed-off-by: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>
> > > Cc: stable@xxxxxxxxxxxxxxx # v4.15+
> >
> > ... oh, I don't think this is suitable for stable.
> > Michal's fix in another email can come first with "CC: stable",
> > then this one.
> > Anyway I want to get some feedback on the change of this patch.
>
> My patch is indeed much simpler but it depends on [1] and that doesn't
> sound like a stable material as well because it depends on onether 2
> patches. Maybe we need some other hack for 4.15 if we really care enough.
>
> [1] http://lkml.kernel.org/r/20180103082555.14592-4-mhocko@xxxxxxxxxx

OK, so I like just giving up sending to stable.