Linux-mm Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: SJ Park <sj@kernel.org>
To: Aveline Noir <jm5905938@gmail.com>
Cc: SJ Park <sj@kernel.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	damon@lists.linux.dev, linux-kernel@vger.kernel.org,
	linux-mm@kvack.org, sashiko-bot@kernel.org,
	sashiko-reviews@lists.linux.dev
Subject: Re: [RFC PATCH 14/27] mm/damon/paddr: support REPORT_MADE type probe filter
Date: Mon, 16 Nov 2026 07:05:42 -0800	[thread overview]
Message-ID: <20261116150543.80514-1-sj@kernel.org> (raw)
In-Reply-To: <179481501691.2405628.8067144236288202449.b4-reply@b4>

Hi Aveline,


On Mon, 16 Nov 2026 18:43:36 +1100 Aveline Noir <jm5905938@gmail.com> wrote:

> Hi SJ,
> 
> On Sun, 15 Nov 2026 12:28:32 -0800 SJ Park wrote:
> 
> > +	case DAMON_FILTER_TYPE_REPORT_MADE:
> > +		addr_range.start = pa;
> > +		addr_range.end = pa + 1;
> > +		matched = damon_report_exist(filter->report_type, &addr_range,
> > +				jiffies -
> > +				usecs_to_jiffies(ctx->attrs.sample_interval));
> 
> Since the window is counted in jiffies and rounded up, how many samples
> see a given report depends on HZ.  With the default 5 ms sampling
> interval, the window is 2 jiffies at HZ=250, i.e. 8 to 12 ms, so a
> report is usually seen by two consecutive samples.  At HZ=1000 the
> window is 5 jiffies, but kdamond_usleep() sleeps via
> schedule_timeout_idle() there, and samples end up about 6 ms apart, so
> a report made right after a sample can fall out of the window before
> the next one.
> 
> With the queue indexing fixed, user reports submitted every 50 or
> 250 ms and covering the whole physical address range were counted
> about 2.0 times each on an x86_64 guest at HZ=250, and about 0.84
> times at HZ=1000.

Thank you for sharing your thought and test results.

> 
> Is this approximation intended?

Kind of intended.  DAMON uses many approximations for low overhead and easy
maintenance, as long as the results are practical.

For this case, I'm not that scared about such over-counting.  Users ain't
really care about the report counted just one more time.  I'd rather expect
users repeatedly submit reports to make sure DAMON shows the memory as hot.

Also I believe users could tune intervals, manually or using the auto-tuning
feature.

> If each report is meant to be counted
> once, maybe we could keep a per-context report sequence cursor and
> check only the reports enqueued since the previous probe run.  A quick
> local hack of that counted each report once on both HZ values.  If
> reports are rather meant to stay valid for a while, like a user saying
> a range is hot, a validity period separate from the sampling interval
> might fit better.

If it turns out the approximation is real problem for users, we could consider
this kind of improvements.


Thanks,
SJ

[...]


  reply	other threads:[~2026-11-16 15:05 UTC|newest]

Thread overview: 47+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-15 20:28 [RFC PATCH 00/27] mm/damon: let users report their data attribute findings SJ Park
2026-11-15 20:28 ` [RFC PATCH 01/27] mm/damon/api: introduce struct damon_report SJ Park
2026-11-15 20:34   ` sashiko-bot
2026-11-15 21:09     ` SJ Park
2026-11-15 20:28 ` [RFC PATCH 02/27] mm/damon/core: implement data attribute report queue SJ Park
2026-11-15 20:41   ` sashiko-bot
2026-11-15 21:17     ` SJ Park
2026-11-15 20:28 ` [RFC PATCH 03/27] mm/damon/core: introduce damon_submit_report() SJ Park
2026-11-15 20:42   ` sashiko-bot
2026-11-15 21:23     ` SJ Park
2026-11-15 20:28 ` [RFC PATCH 04/27] mm/damon/sysfs: add admin/reports/ directory SJ Park
2026-11-15 20:28 ` [RFC PATCH 05/27] mm/damon/sysfs: add reports/producers/ directory SJ Park
2026-11-15 20:28 ` [RFC PATCH 06/27] mm/damon/sysfs: add producers/nr_producers file SJ Park
2026-11-15 20:28 ` [RFC PATCH 07/27] mm/damon/sysfs: implement report producer directory SJ Park
2026-11-15 20:28 ` [RFC PATCH 08/27] mm/damon/sysfs: implement report producer type and state files SJ Park
2026-11-15 20:28 ` [RFC PATCH 09/27] Docs/mm/damon/design: update for report system SJ Park
2026-11-15 20:28 ` [RFC PATCH 10/27] Docs/admin-guide/mm/damon/usage: update for report producer sysfs SJ Park
2026-11-15 20:28 ` [RFC PATCH 11/27] Docs/ABI/damon: update for report producer sysfs files SJ Park
2026-11-15 20:28 ` [RFC PATCH 12/27] mm/damon/core: introduce DAMON_FILTER_TYPE_REPORT_MADE SJ Park
2026-11-15 20:36   ` sashiko-bot
2026-11-15 21:27     ` SJ Park
2026-11-15 20:28 ` [RFC PATCH 13/27] mm/damon/core: introduce damon_report_exist() SJ Park
2026-11-15 20:37   ` sashiko-bot
2026-11-15 21:35     ` SJ Park
2026-11-15 20:28 ` [RFC PATCH 14/27] mm/damon/paddr: support REPORT_MADE type probe filter SJ Park
2026-11-15 20:41   ` sashiko-bot
2026-11-15 21:40     ` SJ Park
2026-11-16  7:43   ` Aveline Noir
2026-11-16 15:05     ` SJ Park [this message]
2026-11-15 20:28 ` [RFC PATCH 15/27] mm/damon/sysfs: add filter_report_type file under probe filter directory SJ Park
2026-11-15 20:28 ` [RFC PATCH 16/27] mm/damon/sysfs: support REPORT_MADE type probe filter setup SJ Park
2026-11-15 20:41   ` sashiko-bot
2026-11-15 21:42     ` SJ Park
2026-11-15 20:28 ` [RFC PATCH 17/27] Docs/mm/damon/design: update for report_made probe filter type SJ Park
2026-11-15 20:28 ` [RFC PATCH 18/27] Docs/admin-guide/mm/damon/usage: update for REPORT_MADE probe filter SJ Park
2026-11-15 20:28 ` [RFC PATCH 19/27] Docs/ABI/damon: " SJ Park
2026-11-15 20:28 ` [RFC PATCH 20/27] mm/damon/api: introduce user_report report type SJ Park
2026-11-15 20:28 ` [RFC PATCH 21/27] mm/damon/sysfs: add user_report support for report type input SJ Park
2026-11-15 20:28 ` [RFC PATCH 22/27] mm/damon/sysfs: implement report producer addr_{start,end} files SJ Park
2026-11-15 20:28 ` [RFC PATCH 23/27] mm/damon/sysfs: implement submit command for report producer SJ Park
2026-11-15 20:43   ` sashiko-bot
2026-11-15 21:43     ` SJ Park
2026-11-15 20:28 ` [RFC PATCH 24/27] selftests/damon/sysfs.sh: add basic tests for admin/reports/ SJ Park
2026-11-15 20:28 ` [RFC PATCH 25/27] Docs/mm/damon/design: update for user report support SJ Park
2026-11-15 20:28 ` [RFC PATCH 26/27] Docs/admin-guide/mm/damon/usage: update for user_report " SJ Park
2026-11-15 20:28 ` [RFC PATCH 27/27] Docs/ABI/damon: " SJ Park
2026-11-15 20:55 ` [RFC PATCH 00/27] mm/damon: let users report their data attribute findings 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=20261116150543.80514-1-sj@kernel.org \
    --to=sj@kernel.org \
    --cc=akpm@linux-foundation.org \
    --cc=damon@lists.linux.dev \
    --cc=jm5905938@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=sashiko-bot@kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /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