Re: [PATCH v4 2/2] iio: adc: ad7173: add AD7173 driver

From: Andy Shevchenko <andy@kernel.org>
To: Ceclan Dumitru <mitrutzceclan@gmail.com>
Cc: linus.walleij@linaro.org, brgl@bgdev.pl,
	linux-gpio@vger.kernel.org,
	"Lars-Peter Clausen" <lars@metafoo.de>,
	"Jonathan Cameron" <jic23@kernel.org>,
	"Rob Herring" <robh+dt@kernel.org>,
	"Krzysztof Kozlowski" <krzysztof.kozlowski+dt@linaro.org>,
	"Conor Dooley" <conor+dt@kernel.org>,
	"Michael Walle" <michael@walle.cc>,
	"Arnd Bergmann" <arnd@arndb.de>,
	"ChiaEn Wu" <chiaen_wu@richtek.com>,
	"Niklas Schnelle" <schnelle@linux.ibm.com>,
	"Leonard Göhrs" <l.goehrs@pengutronix.de>,
	"Mike Looijmans" <mike.looijmans@topic.nl>,
	"Haibo Chen" <haibo.chen@nxp.com>,
	"Hugo Villeneuve" <hvilleneuve@dimonoff.com>,
	"Ceclan Dumitru" <dumitru.ceclan@analog.com>,
	linux-iio@vger.kernel.org, devicetree@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v4 2/2] iio: adc: ad7173: add AD7173 driver
Date: Mon, 20 Nov 2023 18:21:05 +0200	[thread overview]
Message-ID: <ZVuHcaUQL6xnchK-@smile.fi.intel.com> (raw)
In-Reply-To: <5cf3824f-6375-4c76-86ff-c5389ddf4196@gmail.com>

On Mon, Nov 20, 2023 at 05:55:12PM +0200, Ceclan Dumitru wrote:
> On 11/20/23 15:00, Andy Shevchenko wrote:
> > On Thu, Nov 16, 2023 at 03:46:55PM +0200, mitrutzceclan wrote:

...

> >> +struct ad7173_channel_config {
> >> +	bool live;
> >> +	u8 cfg_slot;
> >> +	/* Following fields are used to compare equality. Bipolar must be first */
> >> +	bool bipolar;
> >> +	bool input_buf;
> >> +	u8 odr;
> >> +	u8 ref_sel;
> > 
> > If you group better by types, it might save a few bytes on the architectures /
> > compilers where bool != byte.
> >
> Grouping by type will result in not being able to use memcmp() for
> comparing configs. But then there is the issue that I was under the
> assumption that bool=byte. If that is not the case, the config equality
> check might be comparing padding bytes.

It's most likely the case, BUT... it is not guaranteed by the C standard.

> In this case what do you suggest:
> - using the packed attribute
> - using only u8
> - drop memcmp, manually compare fields

Use struct_group() to show explicitly the group of members. If it's not an ABI
in any sense, then memcmp() is fine.

...

> >> +	cmp_size = sizeof(*cfg) - offset;
> > 
> > sizeof_field() from the above mentioned header?
> 
> This computes the size of multiple fields, following cfg_slot. Better to
> group the fields that need to be compared into another struct then use
> sizeof_field()?

See above.

...

> >> +	return vref / (MICRO/MILLI);
> > 
> > What does the denominator mean and why you can't simply use MILL?
> 
> Original vref values are in micro, I considered that it was adequate to
> represent the conversion from MICRO to MILLI as a fraction.
> 
> >> +		*val = st->info->sinc5_data_rates[reg] / (MICRO/MILLI);
> >> +		*val2 = (st->info->sinc5_data_rates[reg] % MILLI) * (MICRO/MILLI);
> > 
> > Same Q about denominator.
> > 
> Here, a misunderstanding on my part of a suggestion from Jonathan in V2,
> will be removed.

You need to clarify with him that.

-- 
With Best Regards,
Andy Shevchenko