RE: [PATCH 2/9] clk: realtek: Add basic reset support

["Yu-Chun Lin [林祐君]" <eleanor.lin@realtek.com>]

> On Mon, Dec 29, 2025 at 03:53:06PM +0800, Yu-Chun Lin wrote:
> > +
> > +int rtk_reset_controller_add(struct device *dev,
> > +                          struct rtk_reset_initdata *initdata) {
> > +     struct rtk_reset_data *data;
> > +
> > +     data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL);
> > +
> 
> What sort of coding style do you use? There is absolutely never a blank line
> between kzalloc and the if().
> 

Will remove the blank line.

> > +     if (!data)
> > +             return -ENOMEM;
> > +
> > +     data->dev = dev;
> > +     data->num_banks = initdata->num_banks;
> > +     data->banks = initdata->banks;
> > +     data->regmap = initdata->regmap;
> > +     data->rcdev.owner = THIS_MODULE;
> 
> THIS_MODULE? so which module is exactly the owner - clk-rtk.ko or actual
> driver?
> 
> This feels buggy, but I did not check your Makefile.
> 

The reset support and the clock driver are designed to be linked together into
a single kernel module (clk-rtk.ko). Therefore, THIS_MODULE is the correct owner.

> > +     data->rcdev.ops = &rtk_reset_ops;
> > +     data->rcdev.dev = dev;
> > +     data->rcdev.of_node = dev->of_node;
> > +     data->rcdev.nr_resets = initdata->num_banks * 0x100;
> > +     data->rcdev.of_xlate = rtk_of_reset_xlate;
> > +     data->rcdev.of_reset_n_cells = 1;
> > +
> > +     return devm_reset_controller_register(dev, &data->rcdev); }
> > +EXPORT_SYMBOL_GPL(rtk_reset_controller_add);
> > +
> > diff --git a/drivers/clk/realtek/reset.h b/drivers/clk/realtek/reset.h
> > new file mode 100644 index 000000000000..cd446b098429
> > --- /dev/null
> > +++ b/drivers/clk/realtek/reset.h
> > @@ -0,0 +1,36 @@
> > +/* SPDX-License-Identifier: GPL-2.0-only */
> > +/*
> > + * Copyright (C) 2019 Realtek Semiconductor Corporation
> 
> Ah, so that's why you sent probe() from ~10 years ago...
> 
> Best regards,
> Krzysztof