linux-kernel - Re: [PATCH 1/6] thermal: hwmon: Replace the call the thermal_cdev

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: <bf81a4db-9687-b9b2-4976-64bdd364b101@roeck-us.net>
Date:   Sat, 11 Apr 2020 10:26:59 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     Daniel Lezcano <daniel.lezcano@...aro.org>, rui.zhang@...el.com
Cc:     amit.kucheria@...durent.com, linux-kernel@...r.kernel.org,
        Kamil Debski <kamil@...as.org>,
        Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>,
        Jean Delvare <jdelvare@...e.com>,
        "open list:PWM FAN DRIVER" <linux-hwmon@...r.kernel.org>
Subject: Re: [PATCH 1/6] thermal: hwmon: Replace the call the
 thermal_cdev_update()

On 4/11/20 9:45 AM, Daniel Lezcano wrote:
> On 11/04/2020 03:32, Guenter Roeck wrote:
>> On 4/10/20 3:12 PM, Daniel Lezcano wrote:
>>> The function thermal_cdev_upadte is called from the throttling
>>
>> misspelled
>>
>>> functions in the governors not from the cooling device itself.
>>>
>>> The cooling device is set to its maximum state and then updated. Even
>>> if I don't get the purpose of probing the pwm-fan to its maximum
>>> cooling state, we can replace the thermal_cdev_update() call to the
>>> internal set_cur_state() function directly.
>>>
>>> Signed-off-by: Daniel Lezcano <daniel.lezcano@...aro.org>
>>> ---
>>>  drivers/hwmon/pwm-fan.c | 3 +--
>>>  1 file changed, 1 insertion(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/hwmon/pwm-fan.c b/drivers/hwmon/pwm-fan.c
>>> index 30b7b3ea8836..a654ecdf21ab 100644
>>> --- a/drivers/hwmon/pwm-fan.c
>>> +++ b/drivers/hwmon/pwm-fan.c
>>> @@ -372,7 +372,6 @@ static int pwm_fan_probe(struct platform_device *pdev)
>>>  	if (ret)
>>>  		return ret;
>>>  
>>> -	ctx->pwm_fan_state = ctx->pwm_fan_max_state;
>>>  	if (IS_ENABLED(CONFIG_THERMAL)) {
>>>  		cdev = devm_thermal_of_cooling_device_register(dev,
>>>  			dev->of_node, "pwm-fan", ctx, &pwm_fan_cooling_ops);
>>> @@ -384,7 +383,7 @@ static int pwm_fan_probe(struct platform_device *pdev)
>>>  			return ret;
>>>  		}
>>>  		ctx->cdev = cdev;
>>> -		thermal_cdev_update(cdev);
>>> +		pwm_fan_set_cur_state(cdev, ctx->pwm_fan_max_state);
>>
>> So far the function would only change the state if the new
>> state is not equal to the old state. This was the case because
>> pwm_fan_state was set to pwm_fan_max_state, and the call to
>> thermal_cdev_update() and thus pwm_fan_set_cur_state() would
>> do nothing except update statistics. The old code _assumed_
>> that the current state is pwm_fan_max_state. The new code
>> enforces it. That is a substantial semantic change, and it
>> is not really reflected in the commit message. Is that really
>> what you want ? If so, the commit message needs to state that
>> and explain the rationale.
> 
> Well, to be honest I'm not getting the rational of calling
> thermal_cdev_update(cdev) right after
> devm_thermal_of_cooling_device_register() neither setting pwm_fan_state
> to pwm_fan_max_state.
> 
Good question. The author might know/recall. Maybe the idea was that
thermal would update the state to a lower state shortly thereafter.

> Do we have the guarantee there is at this point a thermal instance
> making the target state working when thermal_cdev_update is called?
> 
> Are we sure a thermal_cdev_update(cdev) is actually right here?
> 
I don't know. I am not exactly familiar with thermal subsystem
particulars. I do recall seeing similar code in other drivers, though.

Either case, your patch does change functionality, and we should not
do that without understanding its impact.

Thanks
Guenter