[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20250807001924.76275-1-sj@kernel.org>
Date: Wed, 6 Aug 2025 17:19:24 -0700
From: SeongJae Park <sj@...nel.org>
To: Bijan Tabatabai <bijan311@...il.com>
Cc: SeongJae Park <sj@...nel.org>,
damon@...ts.linux.dev,
linux-mm@...ck.org,
linux-kernel@...r.kernel.org,
Andrew Morton <akpm@...ux-foundation.org>,
Bijan Tabatabai <bijantabatab@...ron.com>
Subject: Re: [PATCH v2] mm/damon/core: skip needless update of damon_attrs in damon_commit_ctx()
On Wed, 6 Aug 2025 18:42:54 -0500 Bijan Tabatabai <bijan311@...il.com> wrote:
> From: Bijan Tabatabai <bijantabatab@...ron.com>
>
> Currently, damon_commit_ctx() always calls damon_set_attrs() even if the
> attributes have not been changed. This can be problematic when the DAMON
> state is committed relatively frequently because damon_set_attrs() resets
> ctx->next_{aggregation,ops_update}_sis, causing aggregation and ops
> update operations to be needlessly delayed.
>
> This patch avoids this by only calling damon_set_attrs() in
> damon_commit_ctx when the attributes have been changed.
>
> Cc: Bijan Tabatabai <bijan311@...il.com>
Maybe above line is added by a mistake?
> Signed-off-by: Bijan Tabatabai <bijantabatab@...ron.com>
Reviewed-by: SeongJae Park <sj@...nel.org>
I have a trivial comment below, though.
> ---
> Changes from v1[1]:
> - Compare entirety of struct damon_attrs
> - Apply logic in damon_commit_ctx() instead of damon_set_attrs()
Thank you for doing this!
>
> [1] https://lore.kernel.org/all/20250806164316.5728-1-bijan311@gmail.com/
> ---
> mm/damon/core.c | 26 +++++++++++++++++++++++---
> 1 file changed, 23 insertions(+), 3 deletions(-)
>
> diff --git a/mm/damon/core.c b/mm/damon/core.c
> index 6a2fe1f2c952..80aaeb876bf2 100644
> --- a/mm/damon/core.c
> +++ b/mm/damon/core.c
> @@ -570,6 +570,24 @@ void damon_destroy_ctx(struct damon_ctx *ctx)
> kfree(ctx);
> }
>
> +static bool damon_attrs_equals(const struct damon_attrs *attrs1,
> + const struct damon_attrs *attrs2)
> +{
> + const struct damon_intervals_goal *ig1 = &attrs1->intervals_goal;
> + const struct damon_intervals_goal *ig2 = &attrs2->intervals_goal;
> +
> + return attrs1->sample_interval == attrs2->sample_interval &&
> + attrs1->aggr_interval == attrs2->aggr_interval &&
> + attrs1->ops_update_interval == attrs2->ops_update_interval &&
> + attrs1->min_nr_regions == attrs2->min_nr_regions &&
> + attrs1->max_nr_regions == attrs2->max_nr_regions &&
> + ig1->access_bp == ig2->access_bp &&
> + ig1->aggrs == ig2->aggrs &&
> + ig1->min_sample_us == ig2->min_sample_us &&
> + ig1->max_sample_us == ig2->max_sample_us;
> +
Unnecessary blank line?
> +}
> +
> static unsigned int damon_age_for_new_attrs(unsigned int age,
> struct damon_attrs *old_attrs, struct damon_attrs *new_attrs)
> {
> @@ -1198,9 +1216,11 @@ int damon_commit_ctx(struct damon_ctx *dst, struct damon_ctx *src)
> * 2. ops update should be done after pid handling is done (target
> * committing require putting pids).
> */
> - err = damon_set_attrs(dst, &src->attrs);
> - if (err)
> - return err;
> + if (!damon_attrs_equals(&dst->attrs, &src->attrs)) {
> + err = damon_set_attrs(dst, &src->attrs);
> + if (err)
> + return err;
> + }
> dst->ops = src->ops;
>
> return 0;
> --
> 2.43.0
Other than the above trivial things, looks good to me. Thank you again!
Andrew, I assume you would prefer resolving the above nits (removing
unnecessary cc and blank line) on your own? Pleae let us know if not. :)
Thanks,
SJ
Powered by blists - more mailing lists