Re: [PATCH v4 1/2] iio: pressure: dps310: Refactor startup procedure

From: Eddie James <eajames@linux.ibm.com>
To: Jonathan Cameron <jic23@kernel.org>,
	Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: linux-iio <linux-iio@vger.kernel.org>,
	Lars-Peter Clausen <lars@metafoo.de>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Joel Stanley <joel@jms.id.au>
Subject: Re: [PATCH v4 1/2] iio: pressure: dps310: Refactor startup procedure
Date: Tue, 6 Sep 2022 14:48:20 -0500	[thread overview]
Message-ID: <85280d48-4251-2811-b66d-092f4153fbb5@linux.ibm.com> (raw)
In-Reply-To: <20220820124915.5dd5b745@jic23-huawei>

On 8/20/22 06:49, Jonathan Cameron wrote:
> On Fri, 19 Aug 2022 12:42:00 +0300
> Andy Shevchenko <andy.shevchenko@gmail.com> wrote:
>
>> On Mon, Aug 15, 2022 at 4:42 PM Eddie James <eajames@linux.ibm.com> wrote:
>>> On 8/12/22 17:03, Andy Shevchenko wrote:
>>>> On Wed, Aug 10, 2022 at 12:12 AM Eddie James <eajames@linux.ibm.com> wrote:
>> ...
>>
>>>>> +       rc = regmap_write(data->regmap, 0x0e, 0xA5);
>>>>> +       if (rc)
>>>>> +               return rc;
>>>>> +
>>>>> +       rc = regmap_write(data->regmap, 0x0f, 0x96);
>>>>> +       if (rc)
>>>>> +               return rc;
>>>> This code already exists, but still want to ask, is it really
>>>> byte-registers here and not be16/le16 one? In such a case perhaps bulk
>>>> write can be used to reflect it better?
>>> The temperature and pressure regs are 24 bits big endian, and all the
>>> rest are 8 bits. I think the existing approach is best.
>> It doesn't look like you got what I was meaning... Or I misunderstood
>> what you said.
>>
>> The code above writes two byte values to two sequential registers
>> which make me think that they are 16-bit registers at offset 0x0e.
> Given they are undocumented, this is guessing territory.
> Probably best to just leave them as is.
> You could do a bulk write on an array though as that implies
> nothing about what's in the registers -just that they happen
> to be next to each other.

Indeed. Is it worth it to switch to bulk write for two 2-byte writes? 
I'm inclined to say no and will leave this as-is for v6, but if you 
think it is, I can switch it.

Thanks,

Eddie

>
>> ...
>>
>>>>> +       rc = regmap_write(data->regmap, 0x0e, 0x00);
>>>>> +       if (rc)
>>>>> +               return rc;
>>>>> +
>>>>> +       return regmap_write(data->regmap, 0x0f, 0x00);
>> Ditto.
>>