linux-kernel - Re: [PATCH v4 2/2] i2c: mux: ltc4306: LTC4306 and LTC4305 I2C multiplexer/switch

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Tue, 11 Apr 2017 11:29:03 +0200
From:   Michael Hennerich <michael.hennerich@...log.com>
To:     Peter Rosin <peda@...ntia.se>, <wsa@...-dreams.de>,
        <robh+dt@...nel.org>, <mark.rutland@....com>,
        <linus.walleij@...aro.org>
CC:     <linux-i2c@...r.kernel.org>, <devicetree@...r.kernel.org>,
        <linux-gpio@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v4 2/2] i2c: mux: ltc4306: LTC4306 and LTC4305 I2C
 multiplexer/switch

On 06.04.2017 22:03, Peter Rosin wrote:
> On 2017-04-06 13:31, Michael Hennerich wrote:
>> On 06.04.2017 10:39, Peter Rosin wrote:

Hi Peter,

again - thanks for your review.
Comments below.

>>>> +static const struct regmap_config ltc4306_regmap_config = {
>>>> +	.reg_bits = 8,
>>>> +	.val_bits = 8,
>>>> +	.max_register = LTC_REG_SWITCH,
>>>> +	.volatile_reg = ltc4306_is_volatile_reg,
>>>> +	.cache_type = REGCACHE_RBTREE,
>>>
>>> Did you consider REGCACHE_FLAT? There are very few registers and no hole
>>> in the map, and maintaining a tree seems like total overkill.
>>
>> There is no reason to use REGCACHE_FLAT, in our case it will be a single
>> node.
>
> Ok, so that makes me wonder what need REGCACHE_FLAT satisfies? To me,
> a flat regmap seems like a perfect fit here. Oh well...

https://lkml.org/lkml/2012/12/19/172

It's not worth arguing - if you prefer FLAT - then it's FLAT
While it's still round :-)


>>>> +static int ltc4306_gpio_init(struct ltc4306 *data)
>>>> +{
>>>> +	struct device *dev = regmap_get_device(data->regmap);
>>>> +
>>>> +	if (!data->chip->num_gpios)
>>>> +		return 0;
>>>> +
>>>> +	data->gpiochip.label = dev_name(dev);
>>>> +	data->gpiochip.base = -1;
>>>> +	data->gpiochip.ngpio = data->chip->num_gpios;
>>>> +	data->gpiochip.parent = dev;
>>>> +	data->gpiochip.can_sleep = true;
>>>> +	data->gpiochip.direction_input = ltc4306_gpio_direction_input;
>>>> +	data->gpiochip.direction_output = ltc4306_gpio_direction_output;
>>>
>>> I'm missing a get_direction op?
>>
>> No - its purely optional - the vast majority of gpiochips don't
>> implement it.
>>
>> linux/drivers/gpio$ grep -lr --include \*.c get_direction | wc
>>       36      36     523
>> linux/drivers/gpio$ grep -Lr --include \*.c get_direction | wc
>>      101     101    1461
>
> Ok, but while optional, why not provide it? The implementation would
> be about as simple as ltc4306_gpio_get, no?

ok - convinced me.


I'll send version 5 shortly.


-- 
Greetings,
Michael

--
Analog Devices GmbH      Otl-Aicher Strasse 60-64      80807 München
Sitz der Gesellschaft München, Registergericht München HRB 40368,
Geschäftsführer: Peter Kolberg, Ali Raza Husain, Eileen Wynne