Re: [PATCH 05/14] mm: workingset: let cache workingset challenge anon

[Johannes Weiner]

On Mon, Jun 01, 2020 at 11:56:17AM -0400, Johannes Weiner wrote:
> On Mon, Jun 01, 2020 at 03:14:24PM +0900, Joonsoo Kim wrote:
> > 2020ë 5ì 30ì (í) ìì 12:12, Johannes Weiner <hannes@xxxxxxxxxxx>ëì ìì:
> > > However, your example cannot have a completely silent stable state. As
> > > we stop workingset aging, the refault distances will slowly increase
> > > again. We will always have a bit of churn, and rightfully so, because
> > > the workingset *could* go stale.
> > >
> > > That's the same situation in my cache-only example above. Anytime you
> > > have a subset of pages that by itself could fit into memory, but can't
> > > because of an established workingset, ongoing sampling is necessary.
> > >
> > > But the rate definitely needs to reduce as we detect that in-memory
> > > pages are indeed hot. Otherwise we cause more churn than is required
> > > for an appropriate rate of workingset sampling.
> > >
> > > How about the patch below? It looks correct, but I will have to re-run
> > > my tests to make sure I / we are not missing anything.
> > 
> > Much better! It may solve my concern mostly.
> 
> Okay thanks for confirming. I'll send a proper version to Andrew.

I sent the below patch through my battery of tests to make sure it
doesn't break anything fundamental, and they didn't. For the
convergence test, the overload test, and the kernel build, the
difference, if any, is in the noise, which is expected.

Andrew, absent any objections, can you please fold the below into the
original patch?

It's a bit noisy due to the rename of inactive_age now that it's used
for anon as well (frankly, it was a misnomer from the start), and
switching the interface to take an lruvec instead of a memcg and a
pgdat, which works better for the callers (I stole parent_lruvec()
from "mm: vmscan: determine anon/file pressure balance at the reclaim
root" later in the series, so that isn't new code).

The main change is the two bits in __remove_mapping and
move_pages_to_lru, so that shadow entries are aged when anon is
reclaimed or rotated.

---