On Wed, Jun 25, 2025 at 03:34:55PM +0000, Limonciello, Mario wrote:

On 6/25/25 10:17 AM, Andy Shevchenko wrote:

quoted

On Wed, Jun 25, 2025 at 03:14:40PM +0000, Limonciello, Mario wrote:

quoted

On 6/25/25 10:10 AM, Andy Shevchenko wrote:

quoted

On Wed, Jun 25, 2025 at 03:02:18PM +0000, Limonciello, Mario wrote:

quoted

On 6/25/25 9:41 AM, Mario Limonciello wrote:

quoted

On 6/25/25 9:31 AM, Hans de Goede wrote:

quoted

On 25-Jun-25 4:09 PM, Mario Limonciello wrote:

quoted

On 6/25/25 4:09 AM, Hans de Goede wrote:

quoted

On 24-Jun-25 10:22 PM, Mario Limonciello wrote:

...

quoted

quoted

quoted

quoted

quoted

quoted

Ok, so specifically the gpiod_set_debounce() call with 50 ms
done by gpio_keys.c is the problem I guess?

Yep.

quoted

So amd_gpio_set_debounce() does accept the 50 ms debounce
passed to it by gpio_keys.c as a valid value and then setting
that breaks the wake from suspend?

That's right.

quoted

quoted

quoted

quoted

Also comparing the GPIO register in Windows (where things work)
Windows never programs a debounce.

So maybe the windows ACPI0011 driver always uses a software-
debounce for the buttons? Windows not debouncing the mechanical
switches at all seems unlikely.

I think the best way to fix this might be to add a no-hw-debounce
flag to the data passed from soc_button_array.c to gpio_keys.c
and have gpio_keys.c not call gpiod_set_debounce()  when the
no-hw-debounce flag is set.

I've checked and both on Bay Trail and Cherry Trail devices
where soc_button_array is used a lot hw-debouncing is already
unused. pinctrl-baytrail.c does not accept 50 ms as a valid
value and pinctrl-cherryview.c does not support hw debounce
at all.

That sounds a like a generally good direction to me.

Thinking a bit more of this, perhaps the HW debounce support flag should be
per-GPIO-descriptor thingy. In such cases we don't need to distinguish the
platforms, the GPIO ACPI lib may simply set that flag based on 0 read from
the ACPI tables. It will also give a clue to any driver that uses GPIOs
(not only gpio-keys).

But 0 doesn't mean hardware debounce support is there, 0 means that
hardware debounce is not required to be programmed for this GPIO.

That is - if another system had a non-zero value in the GpioInt entry I
would expect this to be translated into the GPIO register.

Correct. The question is only about 0. So the flow will look like

1) if the GPIO is defined with 0 debounce, set the flag;
2) if the GPIO is defined with non-zero value, try to apply it;
3) if the step 2) fails, warn and set the flag.

Would it make sense?
Hans?

But so on these problematic BYT/CYT tablets which "layer" should be 
setting the 50ms debounce?
That should still be a quirk at the soc_button_array layer, right?

Because gpio_keys_setup_key() will already fallback to software 
debounce, and the goal here is that both of those only use the 50ms 
specifically with software debouncing.

Probably gpiod_set_debounce() should become a no-op in this case?

-- 
With Best Regards,
Andy Shevchenko