Re: [PATCH v2 00/16] Multigenerational LRU Framework

From: Yu Zhao <yuzhao@google.com>
To: Rik van Riel <riel@surriel.com>, Ying Huang <ying.huang@intel.com>
Cc: Dave Chinner <david@fromorbit.com>, Jens Axboe <axboe@kernel.dk>,
	SeongJae Park <sj38.park@gmail.com>,
	Linux-MM <linux-mm@kvack.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Benjamin Manes <ben.manes@gmail.com>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	Hillf Danton <hdanton@sina.com>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Jonathan Corbet <corbet@lwn.net>,
	Joonsoo Kim <iamjoonsoo.kim@lge.com>,
	Matthew Wilcox <willy@infradead.org>,
	Mel Gorman <mgorman@suse.de>, Miaohe Lin <linmiaohe@huawei.com>,
	Michael Larabel <michael@michaellarabel.com>,
	Michal Hocko <mhocko@suse.com>,
	Michel Lespinasse <michel@lespinasse.org>,
	Roman Gushchin <guro@fb.com>, Rong Chen <rong.a.chen@intel.com>,
	SeongJae Park <sjpark@amazon.de>,
	Tim Chen <tim.c.chen@linux.intel.com>,
	Vlastimil Babka <vbabka@suse.cz>, Yang Shi <shy828301@gmail.com>,
	Zi Yan <ziy@nvidia.com>,
	linux-kernel <linux-kernel@vger.kernel.org>,
	lkp@lists.01.org,
	Kernel Page Reclaim v2 <page-reclaim@google.com>,
	Andi Kleen <ak@linux.intel.com>
Subject: Re: [PATCH v2 00/16] Multigenerational LRU Framework
Date: Thu, 15 Apr 2021 01:13:13 -0600	[thread overview]
Message-ID: <CAOUHufaAz72+0HJu=XPv-6T8rUAdYo_mb5=H+yF3=LfwoufeYg@mail.gmail.com> (raw)
In-Reply-To: <20210415030002.GX3762101@tassilo.jf.intel.com>

On Wed, Apr 14, 2021 at 9:00 PM Andi Kleen <ak@linux.intel.com> wrote:
>
> > We fall back to the rmap when it's obviously not smart to do so. There
> > is still a lot of room for improvement in this function though, i.e.,
> > it should be per VMA and NUMA aware.
>
> Okay so it's more a question to tune the cross over heuristic. That
> sounds much easier than replacing everything.
>
> Of course long term it might be a problem to maintain too many
> different ways to do things, but I suppose short term it's a reasonable
> strategy.

Hi Rik, Ying,

Sorry for being persistent. I want to make sure we are on the same page:

Page table scanning doesn't replace the existing rmap walk. It is
complementary and only happens when it is likely that most of the
pages on a system under pressure have been referenced, i.e., out of
*inactive* pages, by definition of the existing implementation. Under
such a condition, scanning *active* pages one by one with the rmap is
likely to cost more than scanning them all at once via page tables.
When we evict *inactive* pages, we still use the rmap and share a
common path with the existing code.

Page table scanning falls back to the rmap walk if the page tables of
a process are apparently sparse, i.e., rss < size of the page tables.

I should have clarified this at the very beginning of the discussion.
But it has become so natural to me and I assumed we'd all see it this
way.

Your concern regarding the NUMA optimization is still valid, and it's
a high priority.

Thanks.