Messages in this thread | | | From | Arnd Bergmann <> | Subject | Re: [PATCH 2/3] PWM: vt8500: Update vt8500 PWM driver support | Date | Mon, 22 Oct 2012 11:50:21 +0000 |
| |
On Monday 22 October 2012, Thierry Reding wrote: > On Mon, Oct 22, 2012 at 07:51:52PM +1300, Tony Prisk wrote: > > Replies to your comments inline: > > > > On Mon, 2012-10-22 at 08:34 +0200, Thierry Reding wrote: > > ... > > > > -static int __devinit pwm_probe(struct platform_device *pdev) > > > > +static const struct of_device_id vt8500_pwm_dt_ids[] = { > > > > + { .compatible = "via,vt8500-pwm", }, > > > > + { /* Sentinel */ } > > > > +}; > > > > + > > > > +static int __devinit vt8500_pwm_probe(struct platform_device *pdev) > > > > > > Since you're changing this line anyway, maybe you should drop __devinit > > > (and __devexit for the .remove() callback). HOTPLUG is always enabled > > > nowadays and will go away eventually, in which case these will need to > > > be removed anyway. > > > > Will do. I must say the inconstancy among comments is rather > > frustrating. In another patch I sent out a few days ago (completely > > unrelated to this), I told to add __devexit to a remove() function :\ > > This is a rather recent development, so maybe not everyone knows about > it yet. You can look at the following commit for the details: > > 45f035ab9b8f45aaf1eb2213218b7e9c14af3fc2 > > It's been in linux-next for about 6 weeks and has also gone into > 3.7-rc1.
As long as we get build warnings for leaving out the __devinit/__devexit annotations, I would generally recommend putting them in. If we do a patch to remove all of them, a couple extra instances will not cause any more troubles than we already have.
Arnd
| |