On Thu, May 04, 2017 at 10:45:11PM +0800, Icenowy Zheng wrote:
> 
> 
> 于 2017年5月4日 GMT+08:00 下午10:41:52, Maxime Ripard <maxime.ripard@free-electrons.com> 写到:
> >On Thu, May 04, 2017 at 10:03:26PM +0800, Icenowy Zheng wrote:
> >> 
> >> 
> >> 于 2017年5月4日 GMT+08:00 下午10:02:20, Maxime Ripard
> ><maxime.ripard@free-electrons.com> 写到:
> >> >On Thu, May 04, 2017 at 09:49:59PM +0800, Icenowy Zheng wrote:
> >> >>  static const struct of_device_id sun4i_a10_pinctrl_match[] = {
> >> >> -	{ .compatible = "allwinner,sun4i-a10-pinctrl", },
> >> >> +	{
> >> >> +		.compatible = "allwinner,sun4i-a10-pinctrl",
> >> >> +		.data = (void *)PINCTRL_SUN4I_A10
> >> >> +	},
> >> >> +	{
> >> >> +		.compatible = "allwinner,sun7i-a20-pinctrl",
> >> >> +		.data = (void *)PINCTRL_SUN7I_A20
> >> >> +	},
> >> >>  	{}
> >> >>  };
> >> >
> >> >This means you're going to have two drivers probing for the same
> >> >device: this driver, and the old one. This is not ok, and probably
> >> >introduces some corner cases.
> >> 
> >> The A20 driver is dropped in the next commit.
> >
> >Yes, and that is an issue for bisectability. Anyone that will checkout
> >between those two commits will have a broken kernel, and that is not
> >ok. Any commit should be in a working state.
> 
> Squash them together?

I would simply remove the A20's def_bool in that patch.

Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com