On Thursday 07 January 2010, Rafael J. Wysocki wrote:
> On Wednesday 06 January 2010, Jesse Barnes wrote:
> > On Sun, 27 Dec 2009 21:06:26 +0100
> > "Rafael J. Wysocki" <rjw@xxxxxxx> wrote:
> > > /**
> > > + * acpi_pm_wakeup_power - Enable/disable device wake-up power.
> > > + * @dev: ACPI device to handle.
> > > + * @enable: Whether to enable or disable the wake-up power of the
> > > device.
> > > + */
> > > +int acpi_pm_wakeup_power(struct acpi_device *dev, bool enable)
> > > +{
> >
> > I know we've got these all over now, but functions that just take a
> > bool are generally hard to read when you just look at the call site.
> > If it was called "acpi_pm_set_wakeup_power" and then took an on/off
> > enum it would be really easy to see, from the callsite, what was going
> > on.
> >
> > It's a fairly minor complaint, but it's something that's always bugged
> > me about the PCI PM code in particular.
>
> Well, in this particular case acpi_pm_wakeup_power() uses a bool, because
> acpi_pm_device_sleep_wake() (which is a caller of it) does. IMO it won't
> be logical to use something else just here.
>
> Also, as you noticed above, this follows a convention used not only in the
> PCI PM, but generally in the core PM code. Although we could change this
> convention, I'm not really sure that would be worth the effort.
That said, it looks like we can drop this (ie. [9/12]) patch altogether if the
next one is changed slightly.
I'll post the modified [10/12] shortly.
Rafael
_______________________________________________
linux-pm mailing list
linux-pm@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linux-foundation.org/mailman/listinfo/linux-pm
