From: SJ Park <sj@kernel.org>
To: Aveline Noir <jm5905938@gmail.com>
Cc: SJ Park <sj@kernel.org>,
Andrew Morton <akpm@linux-foundation.org>,
arijit.desaii@crusoe.ai, Jonathan Corbet <corbet@lwn.net>,
damon@lists.linux.dev, linux-doc@vger.kernel.org,
linux-mm@kvack.org, Randy Dunlap <rdunlap@infradead.org>,
sashiko-bot@kernel.org, sashiko-reviews@lists.linux.dev,
Shuah Khan <skhan@linuxfoundation.org>
Subject: Re: [RFC PATCH 2/4] mm/damon/reclaim: set memory PSI 0.05% as default quota goal
Date: Sat, 21 Nov 2026 09:23:47 -0800 [thread overview]
Message-ID: <20261121172348.80312-1-sj@kernel.org> (raw)
In-Reply-To: <179526475633.4.14859465649138957731.b4-reply@b4>
Hi Aveline,
On Sat, 21 Nov 2026 23:39:16 +1100 Aveline Noir <jm5905938@gmail.com> wrote:
> Hi SJ,
>
> On Thu, 19 Nov 2026 22:07:36 -0800 SJ Park wrote:
>
> > Set 0.5 milliseconds per second
> > (0.5%) as the default auto-tuning target memory pressure ratio.
>
> 0.5 ms per second comes out to 0.05%.
>
> > +static unsigned long quota_mem_pressure_us __read_mostly = 500;
>
> With a fixed 500 us target, maybe a longer aggregation interval also
> lengthen the PSI measurement window. For example, if the quota ends up
> being reset every 2 s, should 1000 us of stalls still be compared
> against 500 us? If so, the tuner would drop the effective quota
> to 1 byte, skipping pageout for the next quota window.
DAMOS quota is not reset for aggregation interval. It has a separate interval
for quota reset. It is exposed to users via quotqa_reset_interval_ms
parameter. The default value is 1 second.
The quota reset interval is evaluated for every sampling interval, though. If
the sampling interval is tuned to be larger than 1 second, the problem Aveline
mentioned could happen.
But, I'm not sure if it could be a real problem. Reclaiming more than expected
is a definite problem that can degrade the system performance. Reclaiming less
than expected is less concerning, as long as the extent is not too much. It
doesn't regress something. Anyway this is a best effort proactive reclamation.
We could revisit this if it becomes clear some real users find it as a problem.
Thanks,
SJ
[...]
next prev parent reply other threads:[~2026-11-21 17:23 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-20 6:07 [RFC PATCH 0/4] mm/damon/reclaim: enable auto-tuning features by default SJ Park
2026-11-20 6:07 ` [RFC PATCH 1/4] mm/damon/reclaim: enable monitoring intervals auto-tune " SJ Park
2026-11-20 6:21 ` sashiko-bot
2026-11-20 6:07 ` [RFC PATCH 2/4] mm/damon/reclaim: set memory PSI 0.05% as default quota goal SJ Park
2026-11-20 6:13 ` sashiko-bot
2026-11-21 12:39 ` Aveline Noir
2026-11-21 17:23 ` SJ Park [this message]
2026-11-20 6:07 ` [RFC PATCH 3/4] mm/damon/reclaim: disable watermarks by default SJ Park
2026-11-20 6:11 ` sashiko-bot
2026-11-20 6:07 ` [RFC PATCH 4/4] Docs/admin-guide/mm/damon/reclaim: update for changed default parameters SJ Park
2026-11-20 6:10 ` sashiko-bot
2026-11-21 1:53 ` [RFC PATCH 0/4] mm/damon/reclaim: enable auto-tuning features by default SJ Park
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261121172348.80312-1-sj@kernel.org \
--to=sj@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=arijit.desaii@crusoe.ai \
--cc=corbet@lwn.net \
--cc=damon@lists.linux.dev \
--cc=jm5905938@gmail.com \
--cc=linux-doc@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=rdunlap@infradead.org \
--cc=sashiko-bot@kernel.org \
--cc=sashiko-reviews@lists.linux.dev \
--cc=skhan@linuxfoundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox