linux-kernel - Re: [PATCH v3 1/1] platform/x86: asus-wmi: add support for vivobook fan profiles

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: <373bcabb-5175-4937-88b7-bd0fec579357@app.fastmail.com>
Date: Fri, 21 Jun 2024 22:26:20 +1200
From: "Luke Jones" <luke@...nes.dev>
To: "Devin Bayer" <dev@...bly.so>, "Mohamed Ghanmi" <mohamed.ghanmi@...com.tn>
Cc: "Hans de Goede" <hdegoede@...hat.com>, corentin.chary@...il.com,
 platform-driver-x86@...r.kernel.org, LKML <linux-kernel@...r.kernel.org>,
 Ilpo Järvinen <ilpo.jarvinen@...ux.intel.com>
Subject: Re: [PATCH v3 1/1] platform/x86: asus-wmi: add support for vivobook fan
 profiles

On Fri, 21 Jun 2024, at 9:43 PM, Devin Bayer wrote:
> Hello Mohamed et al,
> 
> On 05/06/2024 09.43, Luke Jones wrote:
> > 
> > I am saying I would like to see ASUS_THROTTLE_THERMAL_POLICY_FULLSPEED removed, or placed somewhere else such as in <sysfs>/asus-nb-wmi/hwmon/hwmon3/pwm1_enable.
> > 
> > It certainly shouldn't be included in platform_profile and I'm not keen on seeing it in thorttle_thermal_policy either.
> > 
> > A lot of this reasoning is now coming from the refactor I've just done of asus-wmi and the "features" such as this one to place them under firmware_attributes class and begin deprecation of them in asus_wmi. From what I've achieved so far with it it is much *much* more suitable than this ad-hoc style of adding features I've been doing here (I'll submit this work soon, it repalces the last patch series I did).
> > 
> > In light of the above I'm considering the possibility of removing throttle_thermal_policy completely to wholly favour the use of platform_profile. It doesn't make all that much sense to have two different things manipulating the same thing - and as such I don't think the "full speed fan" setting fits at all with platform_profile as it is not a performance tweak.
> > 
> > Mohamed, I would be happy to include the Vivo support so long as:
> > 1. the fullspeed setting is removed
> > 2. the modes map correctly to platform_profile
> > 
> > I hope this makes sense. Very sorry about the previous brief response (I was recovering from surgery trauma).
> 
> I recently created my own patch to toggle this control[^1]. I thought
> it was an alternative fan_boost_mode because the dev_id's were one apart. But with
> more testing I've realized it also controls the GPU power and supporting the standard
> platform_profile is better, so I prefer this patch.
> 
> I tested it and it works. However, it has a couple issues:
> 
> 1. This dev_id isn't Vivobook specific. My Zenbook UX3404VC (2023) has this control.

I'm not sure what else to call it. "thermal_throttle_alt" or otherwise, I don't know. The intention should be clear even if only in a comment.

> 2. The Zenbook only supports values 0-2 (standard, quiet and performance).
>    Calling the method with 3 causes the KEYBOARD_KEY event to fire instead of
>    adjusting the GPU power and fan speed.

I linked you to v4 of this patch in one of my responses. The link again is https://lore.kernel.org/platform-driver-x86/20240609144849.2532-1-mohamed.ghanmi@supcom.tn/T/#mcd18e74676084e21d5c15af84bc08d8c6b375fb9

> So I would suggest to remove the fullspeed setting and rename the constants,
> perhaps to ASUS_THROTTLE_THERMAL_POLICY2.
> 
> Mohamed, if you are busy I could also create a new version of this patch with
> the requested fixes.

As mentioned it is done in v4.

> 
> I wonder if the existing fan_boost_mode should also be considered a platform_profile?

No. It tends to be only fans, and usually fullspeed. Platform_profile is intended to control platform related variables. I would also be curious if you tested without the PPD (daemon) as recent versions may also control the energy performance preference and that will skew your results.