Re: intel-gpio interrupts stop firing with Focaltech I2C-HID touchpad

[Chris Chiu]

On Fri, Nov 17, 2017 at 9:52 PM, Mika Westerberg
<mika.westerberg@xxxxxxxxxxxxxxx> wrote:
> On Fri, Nov 17, 2017 at 03:35:18PM +0200, Mika Westerberg wrote:
>> > pin 18 (GPIO_18) GPIO 0x40900100 0x00024075
>
> Hmm,
>
> If I decode 0x40900100 correctly PADCFG0_GPIROUTIOXAPIC (BIT 20) flag is
> set for the pin. This means the interrupt is routed to IO-APIC instead.
>
> Now, we do clear that flag along with the others when the GPIO is
> requested through gpio_request() but in this case we go through irqchip
> instead so the flag is not cleared.
>
> I wonder if the following patch changes anything?
>
> diff --git a/drivers/pinctrl/intel/pinctrl-intel.c b/drivers/pinctrl/intel/pinctrl-intel.c
> index 71df0f70b61f..06af096f5768 100644
> --- a/drivers/pinctrl/intel/pinctrl-intel.c
> +++ b/drivers/pinctrl/intel/pinctrl-intel.c
> @@ -939,6 +939,10 @@ static int intel_gpio_irq_type(struct irq_data *d, unsigned type)
>
>         value &= ~(PADCFG0_RXEVCFG_MASK | PADCFG0_RXINV);
>
> +       /* Disable SCI/SMI/NMI generation */
> +       value &= ~(PADCFG0_GPIROUTIOXAPIC | PADCFG0_GPIROUTSCI);
> +       value &= ~(PADCFG0_GPIROUTSMI | PADCFG0_GPIROUTNMI);
> +
>         if ((type & IRQ_TYPE_EDGE_BOTH) == IRQ_TYPE_EDGE_BOTH) {
>                 value |= PADCFG0_RXEVCFG_EDGE_BOTH << PADCFG0_RXEVCFG_SHIFT;
>         } else if (type & IRQ_TYPE_EDGE_FALLING) {

It does change after the patch. I didn't have the touchpad stopping
things after a long time test.
The INT3452:00/pins under /sys/kernel/debug/pinctrl shows the
following for GPIO 18.
pin 18 (GPIO_18) GPIO 0x40800102 0x00024075

Anything else you need from me?