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;
MIME-Version: 1.0
References: <20220806152517.78159-1-andriy.shevchenko@linux.intel.com> <20220818232040.GA3505561@roeck-us.net>
In-Reply-To: <20220818232040.GA3505561@roeck-us.net>
From:   Andy Shevchenko <andy.shevchenko@gmail.com>
Date:   Fri, 19 Aug 2022 12:56:42 +0300
Message-ID: <CAHp75Vf+-4U6OW4sf+mKsPt8WoRL_jN-sYtzOMzhhH8--+OU1w@mail.gmail.com>
Subject: Re: [PATCH v1 1/3] hwmon: (pwm-fan) Make use of device properties
To:     Guenter Roeck <linux@roeck-us.net>
Cc:     Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        =?UTF-8?Q?Uwe_Kleine=2DK=C3=B6nig?= 
        <u.kleine-koenig@pengutronix.de>,
        Linux Documentation List <linux-doc@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        linux-pwm@vger.kernel.org, linux-hwmon@vger.kernel.org,
        Jonathan Corbet <corbet@lwn.net>,
        Thierry Reding <thierry.reding@gmail.com>,
        Jean Delvare <jdelvare@suse.com>,
        Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Fri, Aug 19, 2022 at 2:41 AM Guenter Roeck <linux@roeck-us.net> wrote:
> On Sat, Aug 06, 2022 at 06:25:15PM +0300, Andy Shevchenko wrote:
> > Convert the module to be property provider agnostic and allow
> > it to be used on non-OF platforms.
> >
> > Add mod_devicetable.h include.
> >
> > Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> > Acked-by: Uwe Kleine-K=C3=B6nig <u.kleine-koenig@pengutronix.de>
>
> I had another look at this patch. A substantial part of the changes
> is because device properties don't support of_property_read_u32_index(),
> reworking the code to use device_property_read_u32_array() instead.
> Sorry, I don't like it, it results in a substantial number of unnecessary
> changes. Device properties should support the equivalent of
> of_property_read_u32_index() instead to simplify conversions.

Not all (device property) providers can have such API available. Are
you suggesting to
 a) alloc memory for entire array;
 b) cache one for a given index;
 c) free a memory;
 d) loop as many times as index op is called.

Sorry, this is way too far and non-optimal in comparison to the
substantial number of unnecessary changes (two or three small
refactorings?).

Another way is to provide a pwm-fan-acpi, which will be the copy of
the driver after this patch is applied. I don't think it's a very
bright idea either.

--=20
With Best Regards,
Andy Shevchenko