Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Mattijs Korpershoek <mkorpershoek@baylibre.com>
To:     Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@collabora.com>
Cc:     Dmitry Torokhov <dmitry.torokhov@gmail.com>,
        Marco Felsch <m.felsch@pengutronix.de>,
        Rob Herring <robh+dt@kernel.org>,
        Matthias Brugger <matthias.bgg@gmail.com>,
        Fengping Yu <fengping.yu@mediatek.com>,
        Yingjoe Chen <yingjoe.chen@mediatek.com>,
        Fabien Parent <fparent@baylibre.com>,
        Kevin Hilman <khilman@baylibre.com>,
        linux-input@vger.kernel.org, devicetree@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org,
        linux-mediatek@lists.infradead.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v21 2/3] Input: mt6779-keypad - Add MediaTek keypad driver
In-Reply-To: <YiIINpi4zbJq/AZE@smile.fi.intel.com>
References: <20220303154302.252041-1-mkorpershoek@baylibre.com>
 <20220303154302.252041-3-mkorpershoek@baylibre.com>
 <300114e2-6794-db3c-a51c-3f900b6476f9@collabora.com>
 <YiIINpi4zbJq/AZE@smile.fi.intel.com>
Date:   Mon, 14 Mar 2022 09:00:18 +0100
Message-ID: <87ee35rmjx.fsf@baylibre.com>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

On ven., mars 04, 2022 at 14:38, Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote:

> On Fri, Mar 04, 2022 at 11:31:38AM +0100, AngeloGioacchino Del Regno wrote:
>> Il 03/03/22 16:43, Mattijs Korpershoek ha scritto:
>> > From: "fengping.yu" <fengping.yu@mediatek.com>
>> > 
>> > This patch adds matrix keypad support for Mediatek SoCs.
>
>> > +struct mt6779_keypad {
>> > +	struct regmap *regmap;
>> > +	struct input_dev *input_dev;
>> > +	struct clk *clk;
>
>> > +	void __iomem *base;
>
> Not sure why you need this here.
You are right. There is no point of keeping this __iomem region as part
of the structure, since it's only used in the probe() function to create
the regmap.

Will send an improvement part of a later series.

>
>> > +	u32 n_rows;
>> > +	u32 n_cols;
>> > +	DECLARE_BITMAP(keymap_state, MTK_KPD_NUM_BITS);
>> > +};
>> > +
>> > +static const struct regmap_config mt6779_keypad_regmap_cfg = {
>> > +	.reg_bits = 32,
>> > +	.val_bits = 32,
>> > +	.reg_stride = sizeof(u32),
>> > +	.max_register = 36,
>> 
>> Are you sure that you can't use .fast_io = true?
>> 
>> Another version for the same question:
>> Are you sure that you need to lock with a mutex here, and not with a spinlock?
>> 
>> Since you're performing reads over a MMIO, I think that there's a very good
>> chance that you can use fast_io.
>> 
>> > +};
>
> ...
>
>> Please use dev_err_probe() to simplify error handling in probe functions: you've
>> done a great job with adding a devm action for the error cases, avoiding gotos to
>> get out cleanly.. it would be a pity to not finish this to perfection.
>> 
>> I'll give you two examples for this, so that you'll be all set.
>> 
>> 	if (IS_ERR(keypad->regmap))
>> 		return dev_err_probe(&pdev->dev, PTR_ERR(keypad->regmap),
>> 				     "regmap init failed\n");
>> 
>> P.S.: No need for %pe here, as dev_err_probe prints the error number for you!
>
> Maintainer of the input subsystem is strongly against dev_err_probe() API. See
> other files there. Ditto for other cases you mentioned below.
>
> -- 
> With Best Regards,
> Andy Shevchenko