[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <5f3aae11-f881-78e8-f90e-80b485e252f3@linaro.org>
Date: Fri, 8 Jul 2022 12:48:54 +0200
From: Daniel Lezcano <daniel.lezcano@...aro.org>
To: Zhang Rui <rui.zhang@...el.com>, rafael@...nel.org
Cc: quic_manafm@...cinc.com, Amit Kucheria <amitk@...nel.org>,
"open list:THERMAL" <linux-pm@...r.kernel.org>,
open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 3/3] thermal/core: Fix thermal trip cross point
On 08/07/2022 05:56, Zhang Rui wrote:
[ ... ]
>> + if (tz->last_temperature < trip_temp && tz->temperature >=
>> trip_temp &&
>> + trip != tz->prev_trip) {
>> + thermal_notify_tz_trip_up(tz->id, trip, tz-
>>> temperature);
>> + tz->prev_trip = trip;
>> +
>> + } else if (tz->last_temperature >= trip_low_temp && tz-
>>> temperature < trip_low_temp &&
>> + trip == tz->prev_trip) {
>> + thermal_notify_tz_trip_down(tz->id, trip, tz-
>>> temperature);
>> + tz->prev_trip = trip - 1;
>
> Say, let's assume hysteresis is Zero,
> When the temperature increases and we do thermal_notify_tz_trip_up()
> for trip 0 and trip 1, tz->prev_trip is set to 1 in this case.
> And then the temperature drops below trip 0, we don't have chance to do
> thermal_notify_tz_trip_down() for trip 0, because we always handle the
> trips in ascending order, and tz->prev_trip is 1 when we do
> handle_thermal_trip(0).
Well actually you are right, I reproduced the scenario with temperature
emulation. I'll investigate how to handle this case.
Thanks for spotting this
<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
Powered by blists - more mailing lists