Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:3 as permitted sender) client-ip=2620:137:e000::3:3;
Message-ID: <7558dd96-0d96-9463-9a97-7ea8bac2046e@linaro.org>
Date:   Mon, 25 Sep 2023 12:19:18 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.13.0
Subject: Re: [PATCH] thermal: core: Check correct temperature for thermal trip
 notification
Content-Language: en-US
To:     =?UTF-8?B?TsOtY29sYXMgRi4gUi4gQS4gUHJhZG8=?= 
        <nfraprado@collabora.com>, "Rafael J . Wysocki" <rafael@kernel.org>
Cc:     kernel@collabora.com,
        AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@collabora.com>,
        Amit Kucheria <amitk@kernel.org>,
        Zhang Rui <rui.zhang@intel.com>, linux-kernel@vger.kernel.org,
        linux-pm@vger.kernel.org
References: <20230922192724.295129-1-nfraprado@collabora.com>
From:   Daniel Lezcano <daniel.lezcano@linaro.org>
In-Reply-To: <20230922192724.295129-1-nfraprado@collabora.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk


Hi Nicolas,

On 22/09/2023 21:27, Nícolas F. R. A. Prado wrote:
> The thermal trip down notification should be triggered when the
> temperature goes below the trip temperature minus the hysteresis. But
> while the temperature is correctly checked against that, the last
> temperature is instead compared against the trip point temperature. The
> end result is that the notification won't always be triggered, only when
> the temperature happens to drop quick enough so that the last
> temperature was still above the trip point temperature.
> 
> Fix the incorrect check.
> 
> Fixes: 55cdf0a283b8 ("thermal: core: Add notifications call in the framework")
> Signed-off-by: Nícolas F. R. A. Prado <nfraprado@collabora.com>
> 
> ---
> 
>   drivers/thermal/thermal_core.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c
> index 58533ea75cd9..120fcf23b8e5 100644
> --- a/drivers/thermal/thermal_core.c
> +++ b/drivers/thermal/thermal_core.c
> @@ -361,7 +361,7 @@ static void handle_thermal_trip(struct thermal_zone_device *tz, int trip_id)
>   		    tz->temperature >= trip.temperature)
>   			thermal_notify_tz_trip_up(tz->id, trip_id,
>   						  tz->temperature);
> -		if (tz->last_temperature >= trip.temperature &&
> +		if (tz->last_temperature >= (trip.temperature - trip.hysteresis) &&
>   		    tz->temperature < (trip.temperature - trip.hysteresis))
>   			thermal_notify_tz_trip_down(tz->id, trip_id,
>   						    tz->temperature);

We already did a try to fix the thermal trip cross notification but this 
is not sufficient for a full fix.

We are receiving multiple notifications from the same event, all this 
due to the hysteresis.

Let's say, we have a trip point T and a hysteresis H.

There is a trip point crossed the way up when:

	last_temperature < T and temperature >= T

At this point, we send a notification

Now, the temperature decreases but it stays in the hysteresis:

	last_temperature >= T and temperature > (T - H)

And then, the temperature increases again and is greater than T.

	last_temperature > (T - H) and temperature >= T

We send a second notification.

With the mitigation kicking in at temp >= T, we end up with multiple 
events 'temperature crossed the trip point the way up"'. That forces the 
receiver of the events to detect duplicate events in order to ignore them.

More info:

	https://lore.kernel.org/all/20220718145038.1114379-4-daniel.lezcano@linaro.org/
	
We have done a lot of cleanups to use the 'generic trip points' and 
remove those get_trip_* ops. So the trip point structure is a core 
component being reused by the drivers and registered as an array.

Next step is to make sure the trip points are ordered in the array, so 
we can implement the correct trip point crossing detection.


-- 
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog