linux-kernel - Re: [RFC PATCH 0/2] Add a generic virtual thermal sensor

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <df9ad6a9-4cf9-dd41-8876-d19bef03d6a5@baylibre.com>
Date:   Mon, 23 Aug 2021 09:35:13 +0200
From:   Alexandre Bailon <abailon@...libre.com>
To:     Daniel Lezcano <daniel.lezcano@...aro.org>, rui.zhang@...el.com,
        amitk@...nel.org
Cc:     linux-pm@...r.kernel.org, linux-kernel@...r.kernel.org,
        ben.tseng@...iatek.com, khilman@...libre.com
Subject: Re: [RFC PATCH 0/2] Add a generic virtual thermal sensor

Hi Daniel,

On 20/08/2021 13:30, Daniel Lezcano wrote:
> Hi Alexandre,
>
> thanks for the proposal.
>
> On 19/08/2021 14:32, Alexandre Bailon wrote:
>> This series add a virtual thermal sensor that uses the hardware thermal sensors,
>> aggregate them to return a temperature.
>>
>> My first aptempt was to do the aggregation in the thermal zone but it was not
>> that easy to do, and, there were some case that would have been conflictual
>> such as setting differents trip for a regular zone and a multisensor zone.
>>
>> Instead, I made a virtual thermal sensor that could registered in a thermal
>> zone, and have its own properties.
>> It could be added in the device tree, with the list of sensors to aggregate,
>> and the type of aggregation to be done.
>>
>> As example:
>>    soc_max_sensor: soc_max_sensor {
>>      compatible = "generic,thermal-aggregator";
>>      #thermal-sensor-cells = <1>;
>>      type = "max";
>>      thermal-sensors = <&lvts 0>, <&lvts 1>, <&lvts 2>, <&lvts 3>,
>>            <&lvts 4>, <&lvts 5>, <&lvts 6>, <&lvts 7>,
>>            <&lvts 8>, <&lvts 9>, <&lvts 10>, <&lvts 11>,
>>            <&lvts 12>, <&lvts 13>, <&lvts 14>, <&lvts 15>,
>>            <&lvts 16>;
>>    };
>>
>> The current series build and work but it would require to be completed
>> aswell a lot of cleanup.
>> Before working on it, I would like to get some feedback and I know if that
>> would an acceptable solution and continue that way.
> Yes, I think it is going to the right direction.
>
> IMO, we can get rid of the thermal_of changes. From a design PoV, the
> patch itself should be the virtual thermal driver without any changes in
> the core code, including thermal_of.
I made that changes in order to be able to get the hw sensors from the 
virtual sensor.
I am not really satisfied of that patch but that the simplest way I 
found to do it.
How would you proceed to get the hw sensor from its device tree phandle 
and id ?

Thanks,
Alexandre

>
> I have some comments on patch 2/2
>
>
>> Follows the following discussion:
>> https://patchwork.kernel.org/project/linux-mediatek/patch/20210617114707.10618-3-ben.tseng@mediatek.com/
>>
>> Alexandre Bailon (2):
>>    thermal: provide a way to get thermal sensor from a device tree node
>>    thermal: add a virtual sensor to aggregate temperatures
>>
>>   drivers/thermal/Kconfig              |   8 ++
>>   drivers/thermal/Makefile             |   1 +
>>   drivers/thermal/thermal_aggregator.c | 134 +++++++++++++++++++++++++++
>>   drivers/thermal/thermal_of.c         |  43 +++++++++
>>   include/linux/thermal.h              |  12 +++
>>   5 files changed, 198 insertions(+)
>>   create mode 100644 drivers/thermal/thermal_aggregator.c
>>
>