On Tue, Sep 20, 2022 at 6:32 AM Andy Shevchenko
<andriy.shevchenko@xxxxxxxxxxxxxxx> wrote:
>
> On Mon, Sep 19, 2022 at 09:59:09AM -0600, Raul E Rangel wrote:
> > Device tree already has a mechanism to pass the wake_irq. It does this
> > by looking for the wakeup-source property and setting the
> > I2C_CLIENT_WAKE flag. This CL adds the ACPI equivalent. It uses the
> > ACPI interrupt wake flag to determine if the interrupt can be used to
> > wake the system. Previously the i2c drivers had to make assumptions and
> > blindly enable the wake IRQ. This can cause spurious wake events. e.g.,
> > If there is a device with an Active Low interrupt and the device gets
> > powered off while suspending, the interrupt line will go low since it's
> > no longer powered and wakes the system. For this reason we should
> > respect the board designers wishes and honor the wake bit defined on the
> > interrupt.
>
> ...
>
> > + if (irq_ctx.irq == -ENOENT)
> > + irq_ctx.irq = acpi_dev_gpio_irq_wake_get(adev, 0, &irq_ctx.wake_capable);
>
> I just realized, that there is an inconsistency on how we fill the wake_capable
> parameter. In some cases we check for IRQ for an error condition (IRQ not found)
> and in some the wake_capable still be filled.
>
> Here the best approach I believe is to add
>
> if (irq_ctx.irq < 0)
> return irq_ctx.irq;
>
> I.o.w. we apply the rule "do not fill the output parameters when it's known
> to be an error condition".
>
> > + if (wake_capable)
> > + *wake_capable = irq_ctx.wake_capable;
>
> > + return irq_ctx.irq;
>
I applied the following:
diff --git a/drivers/i2c/i2c-core-acpi.c b/drivers/i2c/i2c-core-acpi.c
index ba64e505183595..1618f5619d5ed9 100644
--- a/drivers/i2c/i2c-core-acpi.c
+++ b/drivers/i2c/i2c-core-acpi.c
@@ -220,7 +220,7 @@ int i2c_acpi_get_irq(struct i2c_client *client,
bool *wake_capable)
if (irq_ctx.irq == -ENOENT)
irq_ctx.irq = acpi_dev_gpio_irq_wake_get(adev, 0,
&irq_ctx.wake_capable);
- if (wake_capable)
+ if (irq_ctx.irq > 0 && wake_capable)
*wake_capable = irq_ctx.wake_capable;
return irq_ctx.irq;
Thanks!
