Received-SPF: pass (google.com: domain of linux-kernel+bounces-199400-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Message-ID: <2df46968-ff5f-43bc-98fd-506840c1aaa9@baylibre.com>
Date: Mon, 3 Jun 2024 11:00:12 -0500
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH v4 3/6] iio: adc: ad7173: refactor ain and vref selection
To: "Ceclan, Dumitru" <mitrutzceclan@gmail.com>,
 =?UTF-8?Q?Nuno_S=C3=A1?= <noname.nuno@gmail.com>,
 Jonathan Cameron <jic23@kernel.org>,
 Dumitru Ceclan via B4 Relay <devnull+dumitru.ceclan.analog.com@kernel.org>
Cc: dumitru.ceclan@analog.com, Lars-Peter Clausen <lars@metafoo.de>,
 Michael Hennerich <Michael.Hennerich@analog.com>,
 Rob Herring <robh@kernel.org>,
 Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
 Conor Dooley <conor+dt@kernel.org>, linux-iio@vger.kernel.org,
 devicetree@vger.kernel.org, linux-kernel@vger.kernel.org
References: <20240531-ad4111-v4-0-64607301c057@analog.com>
 <20240531-ad4111-v4-3-64607301c057@analog.com>
 <20240601194925.23123071@jic23-huawei>
 <e9ade241e57383d5342d377bc865046e612a7033.camel@gmail.com>
 <d2370ad2-5fed-41b3-bdd5-c6c895283c18@gmail.com>
Content-Language: en-US
From: David Lechner <dlechner@baylibre.com>
In-Reply-To: <d2370ad2-5fed-41b3-bdd5-c6c895283c18@gmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

On 6/3/24 8:08 AM, Ceclan, Dumitru wrote:
> On 03/06/2024 16:00, Nuno Sá wrote:
>> On Sat, 2024-06-01 at 19:49 +0100, Jonathan Cameron wrote:
>>> On Fri, 31 May 2024 22:42:29 +0300
>>> Dumitru Ceclan via B4 Relay <devnull+dumitru.ceclan.analog.com@kernel.org> wrote:
>>>
>>>> From: Dumitru Ceclan <dumitru.ceclan@analog.com>
>>>>
>>>> Move validation of analog inputs and reference voltage selection to
>>>> separate functions to reduce the size of the channel config parsing
>>>> function and improve readability.
>>>> Add defines for the number of analog inputs in a channel.
>>>>
>>>> Reviewed-by: David Lechner <dlechner@baylibre.com>
>>>> Signed-off-by: Dumitru Ceclan <dumitru.ceclan@analog.com>
>>>> ---
>>>>  drivers/iio/adc/ad7173.c | 71 ++++++++++++++++++++++++++++++++++--------------
>>>>  1 file changed, 50 insertions(+), 21 deletions(-)
>>>>
>>>> diff --git a/drivers/iio/adc/ad7173.c b/drivers/iio/adc/ad7173.c
>>>> index 6e249628bc64..a20831d99aa5 100644
>>>> --- a/drivers/iio/adc/ad7173.c
>>>> +++ b/drivers/iio/adc/ad7173.c
>>>> @@ -60,6 +60,7 @@
>>>>  #define AD7173_CH_SETUP_AINPOS_MASK	GENMASK(9, 5)
>>>>  #define AD7173_CH_SETUP_AINNEG_MASK	GENMASK(4, 0)
>>>>  
>>>> +#define AD7173_NO_AINS_PER_CHANNEL	2
>>>>  #define AD7173_CH_ADDRESS(pos, neg) \
>>>>  	(FIELD_PREP(AD7173_CH_SETUP_AINPOS_MASK, pos) | \
>>>>  	 FIELD_PREP(AD7173_CH_SETUP_AINNEG_MASK, neg))
>>>> @@ -623,6 +624,7 @@ static int ad7173_setup(struct iio_dev *indio_dev)
>>>>  static unsigned int ad7173_get_ref_voltage_milli(struct ad7173_state *st,
>>>>  						 u8 reference_select)
>>>>  {
>>>> +	struct device *dev = &st->sd.spi->dev;
>>>>  	int vref;
>>>>  
>>>>  	switch (reference_select) {
>>>> @@ -646,9 +648,11 @@ static unsigned int ad7173_get_ref_voltage_milli(struct
>>>> ad7173_state *st,
>>>>  		return -EINVAL;
>>>>  	}
>>>>  
>>>> -	if (vref < 0)
>>>> +	if (vref < 0) {
>>>> +		dev_err(dev, "Cannot use reference %u. Error:%d\n",
>>>> +			reference_select, vref);
>>>>  		return vref;
>>>> -
>>>> +	}
>>>>  	return vref / (MICRO / MILLI);
>>>>  }
>>>>  
>>>> @@ -905,13 +909,50 @@ static int ad7173_register_clk_provider(struct iio_dev
>>>> *indio_dev)
>>>>  					   &st->int_clk_hw);
>>>>  }
>>>>  
>>>> +static int ad7173_validate_voltage_ain_inputs(struct ad7173_state *st,
>>>> +					      const unsigned int
>>>> ain[AD7173_NO_AINS_PER_CHANNEL])
>>> I was late to the game in replying to previous thread.
>>>
>>> This is neater without the loop and with 2 parameters.  Anyhow see reply to v3.
>>>
>>
>> Yeps, even more given that we're passing/copying the complete array which always
>> fells awkward to me :)
>>
>> - Nuno Sá
>>
>>
> 
> I rewrote the function, but it feels a bit awkward, perhaps I could get a bit of
> advice before sending V5:

Maybe we could make this easier to read with macros?

> 
> static int ad7173_validate_voltage_ain_inputs(struct ad7173_state *st,
> 					      unsigned int ain0, unsigned int ain1)
> {
> 	struct device *dev = &st->sd.spi->dev;
> 	bool special_input0, special_input1;
> 
> 	special_input0 = ain0 == AD7173_AIN_REF_POS || ain0 == AD7173_AIN_REF_NEG ||
> 			 ((ain0 == AD7173_AIN_COM_IN_POS || ain0 == AD7173_AIN_COM_IN_NEG) &&
> 			 (st->info->has_common_input)) || ain0 == AD4111_VINCOM_INPUT;
> 	special_input1 = (ain1 == AD7173_AIN_REF_POS || ain1 == AD7173_AIN_REF_NEG) ||
> 			 ((ain1 == AD7173_AIN_COM_IN_POS || ain1 == AD7173_AIN_COM_IN_NEG) &&
> 			 (st->info->has_common_input)) || ain1 == AD4111_VINCOM_INPUT;
> 

	special_input0 = AD7173_IS_SPECIAL_INPUT(ain0);
	special_input1 = AD7173_IS_SPECIAL_INPUT(ain1);

> 	if (st->info->has_vincom_input) {
> 		if (ain0 == AD4111_VINCOM_INPUT &&
> 		    ain1 < st->info->num_voltage_in && /* Normal input */
> 		    ain1 >= st->info->num_voltage_in_div) /* Input without divider */
> 			return dev_err_probe(dev, -EINVAL,
> 				"VINCOM must be paired with inputs having divider.\n");
> 
> 		if (ain1 == AD4111_VINCOM_INPUT &&
> 		    ain0 < st->info->num_voltage_in && /* Normal input */
> 		    ain0 >= st->info->num_voltage_in_div) /* Input without divider */
> 			return dev_err_probe(dev, -EINVAL,
> 				"VINCOM must be paired with inputs having divider.\n");

		if (AD7173_IS_VINCOM_MISMATCH(ain0, ain1) ||
		    AD7173_IS_VINCOM_MISMATCH(ain1, ain0)) {
 			return dev_err_probe(dev, -EINVAL,
 				"VINCOM must be paired with inputs having divider.\n");

> 	}
> 
> 	if ((ain0 >= st->info->num_voltage_in && !special_input0) ||
> 	    (ain1 >= st->info->num_voltage_in && !special_input1))
> 		return dev_err_probe(dev, -EINVAL,
> 				     "Input pin number out of range for pair (%d %d).\n",
> 				     ain0, ain1);
> 
> 	if (!special_input0 && !special_input1 &&
> 	    ((ain0 >= st->info->num_voltage_in_div) !=
> 	     (ain1 >= st->info->num_voltage_in_div)))
> 		return dev_err_probe(dev, -EINVAL,
> 			"Both inputs must either have a voltage divider or not have: (%d %d).\n",
> 			ain0, ain1);

These last two don't seem so bad.

> 
> 	return 0;
> }
> 
> It feels a bit too verbose, but I could not come up with a better way to
> incorporate all those cases.