Re: [GIT PULL] platform-drivers-x86 for v7.2-1

[Linus Torvalds]

On Mon, 22 Jun 2026 at 11:42, Ilpo Järvinen
<ilpo.jarvinen@xxxxxxxxxxxxxxx> wrote:
>
> There is going to be a conflict with the commit 3570cb58e317 ("thermal/of:
> Pass cdev_id and introduce devm registration helper") and a few other
> commits that came from thermal tree effectively resulting in a change in
> devm_thermal_of_cooling_device_register() signature. It needs to be
> addressed in the merge commit or the build will fail.
>
> Mark's linux-next discovery about the build failure:
>
>   https://lore.kernel.org/all/aibZy5rrdX3ltGpr@xxxxxxxxxxxxx/

That resolution makes absolutely no sense.

Now that qcom_ec_probe() goes through all the work to create a name for the fan:

                char name[EC_FAN_NAME_SIZE];

                scnprintf(name, sizeof(name), "qcom_ec_fan_%u", i);

but then the name is actually never *used* after the change.

The *logical* resolution would be to make it use

   devm_thermal_of_child_cooling_device_register()

instead, which that function was renamed to in commit 8e1529e79385
("thermal/of: Rename the devm_thermal_of_cooling_device_register()
function")

I have done that resolution that at least makes sense.

But honestly, this is all a complete mess.

Why the hell was that function renamed, and then a new function
created with the old name - when that new function then has exactly
ZERO users?

Renaming a function, and then re-using the old name immediately for
something else is a BAD IDEA. It causes exactly this kind of horrible
confusion. It's an even worse idea when there are exactly zero new
users that make that confusion worthwhile.

Thermal people - you effed up. Stop doing this kind of crap. This is insanity.

                  Linus