Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.32 as permitted sender) client-ip=23.128.96.32;
MIME-Version: 1.0
References: <20231106183159.3562879-1-nphamcs@gmail.com> <CAF8kJuMsXUm9=kiL8qPNVfYPzfyq-JWYSH3KraZadjF+myW-2A@mail.gmail.com>
 <CAKEwX=MNKY0UHbxi6Zfwf0KkepYavFaZo8F6LGe5GyyE3U35Jg@mail.gmail.com>
 <CAF8kJuMx4KT9z2RPy8z+snhM6YUtK=kZ1+BdHjKua2jhwFo-XQ@mail.gmail.com>
 <CAKEwX=OpQZhDmCr-a+O0=c8LfPoO0r8y=abpQoKXWcOP+V6yYg@mail.gmail.com> <CAF8kJuNnM_0jDCaAueseiNA1264-MtA0QiQtfjEN1E6aY56MKQ@mail.gmail.com>
In-Reply-To: <CAF8kJuNnM_0jDCaAueseiNA1264-MtA0QiQtfjEN1E6aY56MKQ@mail.gmail.com>
From:   Chris Li <chrisl@kernel.org>
Date:   Thu, 16 Nov 2023 13:57:39 -0800
Message-ID: <CAF8kJuN5fBBmpOzmR72B5NBmjPNCNk4DALqz=+PKBwQrjvHH2w@mail.gmail.com>
Subject: Re: [PATCH v5 0/6] workload-specific and memory pressure-driven zswap writeback
To:     Nhat Pham <nphamcs@gmail.com>
Cc:     Andrew Morton <akpm@linux-foundation.org>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Domenico Cerasuolo <cerasuolodomenico@gmail.com>,
        Yosry Ahmed <yosryahmed@google.com>,
        Seth Jennings <sjenning@redhat.com>,
        Dan Streetman <ddstreet@ieee.org>,
        Vitaly Wool <vitaly.wool@konsulko.com>, mhocko@kernel.org,
        roman.gushchin@linux.dev, Shakeel Butt <shakeelb@google.com>,
        muchun.song@linux.dev, linux-mm <linux-mm@kvack.org>,
        kernel-team@meta.com, LKML <linux-kernel@vger.kernel.org>,
        cgroups@vger.kernel.org, linux-doc@vger.kernel.org,
        linux-kselftest@vger.kernel.org, shuah@kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hi Nhat,

I want want to share the high level feedback we discussed here in the
mailing list as well.

It is my observation that each memcg LRU list can't compare the page
time order with other memcg.
It works great when the leaf level memcg hits the memory limit and you
want to reclaim from that memcg.
It works less well on the global memory pressure you need to reclaim
from all memcg. You kind of have to
scan each all child memcg to find out the best page to shrink from. It
is less effective to get to the most desirable page quickly.

This can benefit from a design similar to MGLRU. This idea is
suggested by Yu Zhao, credit goes to him not me.
In other words, the current patch is similar to the memcg page list
pre MGLRU world. We can have a MRLRU
like per memcg zswap shrink list.


Chris

On Wed, Nov 8, 2023 at 6:10=E2=80=AFPM Chris Li <chrisl@kernel.org> wrote:
>
> On Wed, Nov 8, 2023 at 4:28=E2=80=AFPM Nhat Pham <nphamcs@gmail.com> wrot=
e:
> >
> > Hmm my guess is that I probably sent this out based on an outdated
> > mm-unstable. There has since been a new zswap selftest merged
> > to mm-unstable (written by no other than myself - oh the irony), so
> > maybe it does not apply cleanly anymore with git am.
>
> $ git am -3 patches/zswap-pool-lru/0005
> Applying: selftests: cgroup: update per-memcg zswap writeback selftest
> Using index info to reconstruct a base tree...
> M       tools/testing/selftests/cgroup/test_zswap.c
> Falling back to patching base and 3-way merge...
> Auto-merging tools/testing/selftests/cgroup/test_zswap.c
> $ git am -3 patches/zswap-pool-lru/0006
> Applying: zswap: shrinks zswap pool based on memory pressure
> error: sha1 information is lacking or useless (mm/zswap.c).
> error: could not build fake ancestor
> Patch failed at 0001 zswap: shrinks zswap pool based on memory pressure
> hint: Use 'git am --show-current-patch=3Ddiff' to see the failed patch
> When you have resolved this problem, run "git am --continue".
> If you prefer to skip this patch, run "git am --skip" instead.
> To restore the original branch and stop patching, run "git am --abort".
>
> I was able to resolve the conflict on patch 6 by hand though. So I am goo=
d now.
>
> Thanks
>
> Chris