Re: [PATCH] clk: renesas: r9a06g032: Avoid needless probe deferring

[Geert Uytterhoeven]

Hi Phil,

CC Jacopo

On Fri, Jul 20, 2018 at 3:57 PM Phil Edworthy <phil.edworthy@xxxxxxxxxxx> wrote:
> On 20 July 2018 13:41, Geert Uytterhoeven wrote:
> > On Fri, Jul 20, 2018 at 2:26 PM Phil Edworthy wrote:
> > > On 20 July 2018 13:12, Geert Uytterhoeven wrote:
> > > > On Fri, Jul 20, 2018 at 2:06 PM Phil Edworthy  wrote:
> > > > > On 20 July 2018 12:21, Geert Uytterhoeven wrote:
> > > > > > On Wed, Jul 18, 2018 at 4:34 PM Phil Edworthy wrote:
> > > > > > > To avoid all SoC peripheral drivers deferring their probes,
> > > > > > > both clock and pinctrl drivers should already be probed. Since
> > > > > > > the pinctrl driver requires a clock to access the registers,
> > > > > > > the clock driver should be probed before the pinctrl driver.
> > > > > > >
> > > > > > > Therefore, move the clock driver from subsys_initcall to
> > core_initcall.
> > > > > > >
> > > > > > > Signed-off-by: Phil Edworthy <phil.edworthy@xxxxxxxxxxx>
> > > > > >
> > > > > > Thanks for your patch!
> > > > > Thanks for your review!
> > > > >
> > > > > > The (not yet upstreamed) pinctrl driver uses postcore_initcall(), right?
> > > > > No, the pinctrl driver uses subsys_initcall, but postcore_initcall
> > > > > or arch_initcall may be better to make it clear about the dependencies.
> > > >
> > > > if the pinctrl driver uses subsys_initcall(), ...
> >
> > > > > > > -subsys_initcall(r9a06g032_clocks_init);
> > > > > > > +core_initcall(r9a06g032_clocks_init);
> > > >
> > > > ... using postcore_initcall() or arch_initcall() here, should work
> > > > with platform_driver_probe()?
> > > Nope, you have to use platform_driver_register() for DT based drivers.
> > > subsys_initcall is the earliest you can use platform_driver_probe().
> >
> > So drivers/misc/atmel_tclib.c and drivers/pinctrl/pinctrl-coh901.c, which use
> > arch_initcall(), cannot work?
> How those drivers work I do not know. However, I tried with the rzn1 clock
> driver and it does not work.
>
> > If that is really true, you can still use subsys_initcall() in the clock driver, and
> > subsys_initcall_sync() in the pinctrl driver.
> True, I'm not sure how people decide which initcall to use and whether the
> _sync variants of the initcalls have a special meaning or intention. As far as
> I know they were introduced for threaded probes, so are we supposed to
> use them for driver dependencies like this?
>
> Do you know why the rza1 pinctrl driver is running from core_initcall()?

Probably because it can?
Note that drivers/clk/renesas/clk-rz.c uses CLK_OF_DECLARE().

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds