* Rafael J. Wysocki <rjw@xxxxxxxxxxxxx> [171013 06:32]:
> [+cc Tony]
>
> On Friday, October 13, 2017 5:04:41 AM CEST Bjorn Helgaas wrote:
> > [+cc Rafael, linux-pm]
> > On Mon, Sep 11, 2017 at 11:10:27PM +0800, Jeffy Chen wrote:
> > > Add support for PCIE_WAKE pin in rockchip pcie driver.
...
> > > @@ -995,6 +996,15 @@ static int rockchip_pcie_setup_irq(struct rockchip_pcie *rockchip)
> > > return err;
> > > }
> > >
> > > + /* Must init wakeup before setting dedicated wakeup irq. */
> > > + device_init_wakeup(dev, true);
> > > + irq = platform_get_irq_byname(pdev, "wakeup");
> > > + if (irq >= 0) {
> > > + err = dev_pm_set_dedicated_wake_irq(dev, irq);
> >
> > I'm a little skeptical about dev_pm_set_dedicated_wake_irq(), not
> > because I know anything at all about it, but because there are only
> > five callers in the whole tree, three of which are in UART code, and
> > none in anything resembling PCI code.
Well it should work for any device that can provide an out of band wakeup
interrupt such as a dedicated GPIO line.
> > Is Rockchip really that special, or are we going about this the wrong
> > way?
Maybe this can be set up in a generic way somewhere similar to what
the I2C bus is doing?
See i2c_device_probe() for dev_pm_set_dedicated_wake_irq() in
drivers/i2c/i2c-core-base.c.
Regards,
Tony
