Re: [PATCH] platform/x86: alienware-wmi-wmax: Do not enable G-Mode for Dell G-series

["Kurt Borja" <kuurtb@gmail.com>]

On Wed Sep 2, 2026 at 4:40 AM -03, Lewis Clement wrote:
> The alienware-wmi-wmax driver auto-engages EC G-Mode (Game Shift) for
> Dell G-series laptops when the platform profile is set to "performance".
> On models like the Dell G5 SE 5505, this pins both fans at 100% (~5400 RPM)
> permanently on every boot.

Hmmm... This seems like you are trying to fix a userspace problem by
patching the driver. What is preventing you from just not selecting the
"performance" profile if you don't like the fans spinning at 100%? With
this patch you are promoting balanced-performance to performance so you
*could* just do that in userspace right? I'm pretty sure Tuned can do
it.

I actually do agree that fans at 100% can be annoying, but that is
subjective. In Windows pressing the "G-Mode" button selects the maximum
performance profile and ramps up fans to 100% too; this driver tries to
mimic that behavior.

>
> Fix this by:
> 1. Setting generic_quirks (gmode=false) for G-series DMI entries so
>    selecting "performance" no longer toggles G-Mode.

You are removing the G-Mode profile for every G-Series laptop. As I
said earlier, some users may like it and some may not at all so it
qualifies as a regression, which would not be acceptable.

> 2. In awcc_platform_profile_probe(), mapping PLATFORM_PROFILE_PERFORMANCE
>    to balanced-performance when no dedicated performance profile exists,
>    keeping the option available with sane fan behavior.

This is a completely different change that would require a different
patch. I'm not exactly against it because of how PPD usually does
things.

Do you know if the AWCC app in Windows does this? Does it map
"performance" to the balanced-performance mode? Not that we have to
follow everything Windows does but it can be a starting point.

>
> This mirrors commit e2468dc70074 ("Revert "platform/x86:
> alienware-wmi-wmax: Add G-Mode support to Alienware m16 R1"").

This does not mirror that commit. That commit revers G-Mode because it
has less performance than the plain "performance" mode. Does this happen
in your case?

>
> Fixes: 202a86120590 ("platform/x86: alienware-wmi-wmax: Extend support to more laptops")

Not a fix.

> Tested-by: Lewis Clement <lewis@lewisclement.nl>
> Signed-off-by: Lewis Clement <lewis@lewisclement.nl>
> ---
> Tested on a Dell G5 SE 5505 (Ryzen 4600H/4800H + RX 5600M, BIOS 1.24.0),
> kernel 7.1.10 (module built out-of-tree against the running kernel):
>
> - Before: fans pinned at ~5400 RPM on every cold boot with
>   profile=performance; cleared only by suspend/resume.
> - After: fans normal (~2400 RPM, scaling with temperature) with
>   profile=performance; survives cold reboot; "performance" remains in
>   /sys/firmware/acpi/platform_profile_choices.
> - Also validated the intermediate gmode=false-only variant (option
>   disappears -> tuned-ppd write rejected as EOPNOTSUPP), which confirmed
>   G-Mode was the only fan-pinning path.
>
> Cc: platform-driver-x86@vger.kernel.org

You dropped this Cc.

> Cc: Kurt Borja <kuurtb@gmail.com>
> Cc: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
> Cc: stable@vger.kernel.org # 6.15+

[...]

Can you email me the acpidump for this model? I want to check what
profiles are reported by the firmware.

-- 
Thanks,
 ~ Kurt