Re: [PATCH v1 15/30] clk: starfive: Use regmap APIs to operate registers

[Hal Feng]

On Sun, 23 Oct 2022 11:25:26 +0100, Conor Dooley wrote:
> On 23 October 2022 05:11:41 IST, Hal Feng <hal.feng@xxxxxxxxxxxxxxxxxxxxxx> wrote:
> >On Wed, 12 Oct 2022 16:05:23 -0700, Stephen Boyd wrote:
> >> Quoting Emil Renner Berthing (2022-10-05 06:14:44)
> >> > > > @@ -295,11 +296,13 @@ static int __init clk_starfive_jh7100_probe(struct platform_device *pdev)
> >> > > >         if (!priv)
> >> > > >                 return -ENOMEM;
> >> > > >
> >> > > > -       spin_lock_init(&priv->rmw_lock);
> >> > > >         priv->dev = &pdev->dev;
> >> > > > -       priv->base = devm_platform_ioremap_resource(pdev, 0);
> >> > > > -       if (IS_ERR(priv->base))
> >> > > > -               return PTR_ERR(priv->base);
> >> > > > +       priv->regmap = device_node_to_regmap(priv->dev->of_node);
> >> > >
> >> > > This is sad. Why do we need to make a syscon? Can we instead use the
> >> > > auxiliary bus to make a reset device that either gets a regmap made here
> >> > > in this driver or uses a void __iomem * mapped with ioremap
> >> > > (priv->base)?
> >> > 
> >> > In my original code the clock driver just registers the resets too
> >> > similar to other combined clock and reset drivers. I wonder what you
> >> > think about that approach:
> >> > https://github.com/esmil/linux/commit/36f15e1b827b02d7f493dc5fce31060b21976e68
> >> > and
> >> > https://github.com/esmil/linux/commit/4ccafadb72968480aa3dd28c227fcccae411c13b#diff-ffec81f902f810cb210012c25e8d88217ea5b4021419a4206d1fd4dd19edfce8R471
> >> 
> >> I think we should use auxiliary bus and split the driver logically into
> >> a reset driver in drivers/reset and a clk driver in drivers/clk. That
> >> way the appropriate maintainers can review the code. There is only one
> >> platform device with a single reg property and node in DT, but there are
> >> two drivers. 
> >
> >Yes, I agree that the reset driver and the clock driver should be split.
> >However, I think using auxiliary bus is a little bit complicated in this
> >case, because the reset is not a part of functionality of the clock in 
> >JH7110. They just share a common register base address. I think it is 
> >better to use ioremap for the same address, and the dt will be like
> >
> >syscrg_clk: clock-controller@13020000 {
> >	compatible = "starfive,jh7110-clkgen-sys";
> >	reg = <0x0 0x13020000 0x0 0x10000>;
> >	...
> >};
> >syscrg_rst: reset-controller@13020000 {
> >	compatible = "starfive,jh7110-reset-sys";
> >	reg = <0x0 0x13020000 0x0 0x10000>;
> >	...
> >};
> >
> >What do you think of this approach? I would appreciate your suggestions.
> 
> No, the dtb checks will all start warning for this.
> Aux bus is not that difficult, you can likely copy much of what I did recently in clk-mpfs.c

Thanks for reminding and your helpful example.

Best regards,
Hal