From: Christian Loehle <christian.loehle@arm.com>
To: "Rafael J. Wysocki" <rafael@kernel.org>,
Linux PM <linux-pm@vger.kernel.org>
Cc: LKML <linux-kernel@vger.kernel.org>,
Reka Norman <rekanorman@chromium.org>
Subject: Re: [PATCH v1 4/4] cpuidle: governors: teo: Decay metrics below DECAY_SHIFT threshold
Date: Wed, 12 Nov 2025 17:29:13 +0000 [thread overview]
Message-ID: <a5de1eca-494e-4624-a86b-bf917e562a08@arm.com> (raw)
In-Reply-To: <3396811.44csPzL39Z@rafael.j.wysocki>
On 11/12/25 16:25, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>
> If a given governor metric falls below a certain value (8 for
> DECAY_SHIFT equal to 3), it will not decay any more due to the
> simplistic decay implementation. This may in some cases lead to
> subtle inconsistencies in the governor behavior, so change the
> decay implementation to take it into account and set the metric
> at hand to 0 in that case.
>
> Suggested-by: Christian Loehle <christian.loehle@arm.com>
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---
> drivers/cpuidle/governors/teo.c | 20 +++++++++++++++-----
> 1 file changed, 15 insertions(+), 5 deletions(-)
>
> --- a/drivers/cpuidle/governors/teo.c
> +++ b/drivers/cpuidle/governors/teo.c
> @@ -148,6 +148,16 @@ struct teo_cpu {
>
> static DEFINE_PER_CPU(struct teo_cpu, teo_cpus);
>
> +static void teo_decay(unsigned int *metric)
> +{
> + unsigned int delta = *metric >> DECAY_SHIFT;
> +
> + if (delta)
> + *metric -= delta;
> + else
> + *metric = 0;
> +}
> +
> /**
> * teo_update - Update CPU metrics after wakeup.
> * @drv: cpuidle driver containing state data.
> @@ -159,7 +169,7 @@ static void teo_update(struct cpuidle_dr
> int i, idx_timer = 0, idx_duration = 0;
> s64 target_residency_ns, measured_ns;
>
> - cpu_data->short_idles -= cpu_data->short_idles >> DECAY_SHIFT;
> + teo_decay(&cpu_data->short_idles);
>
> if (cpu_data->artificial_wakeup) {
> /*
> @@ -195,8 +205,8 @@ static void teo_update(struct cpuidle_dr
> for (i = 0; i < drv->state_count; i++) {
> struct teo_bin *bin = &cpu_data->state_bins[i];
>
> - bin->hits -= bin->hits >> DECAY_SHIFT;
> - bin->intercepts -= bin->intercepts >> DECAY_SHIFT;
> + teo_decay(&bin->hits);
> + teo_decay(&bin->intercepts);
>
> target_residency_ns = drv->states[i].target_residency_ns;
>
> @@ -207,7 +217,7 @@ static void teo_update(struct cpuidle_dr
> }
> }
>
> - cpu_data->tick_intercepts -= cpu_data->tick_intercepts >> DECAY_SHIFT;
> + teo_decay(&cpu_data->tick_intercepts);
> /*
> * If the measured idle duration falls into the same bin as the sleep
> * length, this is a "hit", so update the "hits" metric for that bin.
> @@ -222,7 +232,7 @@ static void teo_update(struct cpuidle_dr
> cpu_data->tick_intercepts += PULSE;
> }
>
> - cpu_data->total -= cpu_data->total >> DECAY_SHIFT;
> + teo_decay(&cpu_data->total);
> cpu_data->total += PULSE;
This will result in total no longer being a strict sum of the bins.
Any reason not to do something like:
-----8<-----
diff --git a/drivers/cpuidle/governors/teo.c b/drivers/cpuidle/governors/teo.c
index e5b795cf3155..ff58d70ee80d 100644
--- a/drivers/cpuidle/governors/teo.c
+++ b/drivers/cpuidle/governors/teo.c
@@ -148,14 +148,19 @@ struct teo_cpu {
static DEFINE_PER_CPU(struct teo_cpu, teo_cpus);
-static void teo_decay(unsigned int *metric)
+static unsigned int teo_decay(unsigned int *metric)
{
unsigned int delta = *metric >> DECAY_SHIFT;
+ unsigned int decay;
- if (delta)
+ if (delta) {
*metric -= delta;
- else
- *metric = 0;
+ return delta;
+ }
+
+ decay = *metric;
+ *metric = 0;
+ return decay;
}
/**
@@ -168,6 +173,7 @@ static void teo_update(struct cpuidle_driver *drv, struct cpuidle_device *dev)
struct teo_cpu *cpu_data = per_cpu_ptr(&teo_cpus, dev->cpu);
int i, idx_timer = 0, idx_duration = 0;
s64 target_residency_ns, measured_ns;
+ unsigned int total_decay = 0;
teo_decay(&cpu_data->short_idles);
@@ -205,8 +211,8 @@ static void teo_update(struct cpuidle_driver *drv, struct cpuidle_device *dev)
for (i = 0; i < drv->state_count; i++) {
struct teo_bin *bin = &cpu_data->state_bins[i];
- teo_decay(&bin->hits);
- teo_decay(&bin->intercepts);
+ total_decay += teo_decay(&bin->hits);
+ total_decay += teo_decay(&bin->intercepts);
target_residency_ns = drv->states[i].target_residency_ns;
@@ -232,7 +238,7 @@ static void teo_update(struct cpuidle_driver *drv, struct cpuidle_device *dev)
cpu_data->tick_intercepts += PULSE;
}
- teo_decay(&cpu_data->total);
+ cpu_data->total -= total_decay;
cpu_data->total += PULSE;
}
next prev parent reply other threads:[~2025-11-12 17:29 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-12 16:21 [PATCH v1 0/4] cpuidle: governors: teo: Assorted improvements Rafael J. Wysocki
2025-11-12 16:22 ` [PATCH v1 1/4] cpuidle: governors: teo: Drop incorrect target residency check Rafael J. Wysocki
2025-11-13 11:32 ` Christian Loehle
2025-11-13 11:41 ` Rafael J. Wysocki
2025-11-13 11:47 ` Rafael J. Wysocki
2025-11-13 13:26 ` Rafael J. Wysocki
2025-11-13 13:24 ` [PATCH v2 1/4] cpuidle: governors: teo: Drop misguided " Rafael J. Wysocki
2025-11-14 9:16 ` Christian Loehle
2025-11-12 16:23 ` [PATCH v1 2/4] cpuidle: governors: teo: Drop redundant function parameter Rafael J. Wysocki
2025-11-13 11:46 ` Christian Loehle
2025-11-12 16:24 ` [PATCH v1 3/4] cpuidle: governors: teo: Use s64 consistently in teo_update() Rafael J. Wysocki
2025-11-13 11:48 ` Christian Loehle
2025-11-12 16:25 ` [PATCH v1 4/4] cpuidle: governors: teo: Decay metrics below DECAY_SHIFT threshold Rafael J. Wysocki
2025-11-12 17:29 ` Christian Loehle [this message]
2025-11-12 17:51 ` Rafael J. Wysocki
2025-11-12 18:00 ` Christian Loehle
2025-11-12 18:03 ` [PATCH v2 " Rafael J. Wysocki
2025-11-13 11:49 ` Christian Loehle
2025-11-13 15:21 ` [PATCH v1 0/4] cpuidle: governors: teo: Assorted improvements Christian Loehle
2025-11-13 15:25 ` Rafael J. Wysocki
2025-11-19 22:52 ` Doug Smythies
2025-11-20 11:02 ` Rafael J. Wysocki
2025-11-20 13:35 ` Christian Loehle
2025-11-20 13:38 ` Rafael J. Wysocki
2025-11-20 13:57 ` Rafael J. Wysocki
2025-11-20 15:21 ` Rafael J. Wysocki
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=a5de1eca-494e-4624-a86b-bf917e562a08@arm.com \
--to=christian.loehle@arm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=rafael@kernel.org \
--cc=rekanorman@chromium.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