Re: [RFC PATCH] pinctrl: cherryview: Introduce quirk to mask GPE during suspend

[Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx>]

On Fri, May 04, 2018 at 12:18:46PM +0100, Carlo Caione wrote:
> From: Carlo Caione <carlo@xxxxxxxxxxxx>
> 
> This code is not for inclusion, it's just an ugly patch to sparkle the
> discussion on how to properly fix this issue.
> 
> We have a problem with a Cherrytrail laptop (x5-Z8350) that is waking up
> from s2idle after a few seconds without any external input.
> 
> This laptop doesn't support (apparently) S3 and the Low Power Idle S0
> _DSM interface is missing functions 5 and 6.
> 
> >From '/sys/power/pm_wakeup_irq' we know that the IRQ waking up the
> laptop is coming from the GPIO controller and that the IRQ is handled as
> ACPI event:
> 
> 115:  0  0  0  94  chv-gpio    5  ACPI:Event
> 
> This IRQ is being triggered every few seconds, when it is received in
> s2idle, this is causing the laptop to wake up.
> 
> Looking in the DSDT table (full dump at [0]) we have:
> 
> Method (_AEI, 0, NotSerialized)  // _AEI: ACPI Event Interrupts
> {
>     Name (RBUF, ResourceTemplate ()
>     {
>         GpioInt (Edge, ActiveBoth, ExclusiveAndWake, PullUp, 0x0000,
>             "\\_SB.GPO1", 0x00, ResourceConsumer, ,
>             )
>             {   // Pin list
>                 0x0005
>             }
>     })
>     Return (RBUF) /* \_SB_.GPO1._AEI.RBUF */
> }
> 
> Method (_E05, 0, NotSerialized)  // _Exx: Edge-Triggered GPE
> {
>     Local0 = Zero
>     If (CondRefOf (\_SB.PCI0.I2C3.BATC, Local1))
>     {
>         Local0 = ^^PCI0.I2C3.BATC.INTR ()
>         If (0xFF == Local0)
>         {
>             ADBG ("INTR RD FAIL")
>             Return (Zero)
>         }
> 
>         If (Zero == Local0)
>         {
>             Return (Zero)
>         }
> 
>         ADBG ("ULPMC INTR")
>         ADBG (Local0)
>     }
> ...
> 
> So, the IRQ is basically probing the battery status every few seconds
> but since the GpioInt has argument ExclusiveAndWake, this is waking up
> the laptop after a few seconds.
> 
> The patch in this RFC gets the idea from commit 76380636280b4 ("ACPI /
> EC: Add parameter to force disable the GPE on suspend") adding a quirk
> to disable the IRQ (and then the GPE) triggered by a specific GPIO/pin.

Since this is not specific to the cherryview driver, I wonder if it
makes more sense to handle this in gpiolib-acpi.c::acpi_gpiochip_request_interrupt()?

	if (agpio->wake_capable == ACPI_WAKE_CAPABLE &&
	    !dmi_check_system(acpi_gpio_event_ignore_wake))
		enable_irq_wake(irq);
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html