[Gta04-owner] [PATCH 5/6] phy: twl4030-usb: add support for reading resistor on ID pin.

[Dr. H. Nikolaus Schaller]

Hi,

Am 02.06.2015 um 22:11 schrieb Pavel Machek <pavel at ucw.cz>:

> On Tue 2015-06-02 16:06:47, Dr. H. Nikolaus Schaller wrote:
>> Hi,
>> 
>> Am 02.06.2015 um 15:49 schrieb Kishon Vijay Abraham I <kishon at ti.com>:
>> 
>>> Hi,
>>> 
>>> On Tuesday 02 June 2015 03:07 AM, NeilBrown wrote:
>>>> On Mon, 1 Jun 2015 19:06:52 +0530 Kishon Vijay Abraham I <kishon at ti.com>
>>>> wrote:
>>>> 
>>>>> Hi,
>>>>> 
>>>>> On Thursday 16 April 2015 01:33 PM, NeilBrown wrote:
>>>>>> From: NeilBrown <neilb at suse.de>
>>>>>> 
>>>>>> The twl4030 phy can measure, with low precision, the
>>>>>> resistance-to-ground of the ID pin.
>>>>>> 
>>>>>> Add a function to read the value, and export the result
>>>>>> via sysfs.
>>>>> 
>>>>> Little sceptical about adding new sysfs entries. Do you have a good reason to
>>>>> add this?
>>>> 
>>>> The hardware can report the value, so why not present it to user-space?
>>>> 
>>>> I originally used this with a udev rule which would configure the maximum
>>>> current based on the resistance measure - to work with the particular charger
>>>> hardware I have.
>>>> 
>>>> More recent patches try to do all of the max-current configuration in the
>>>> kernel, so I could live without exporting the value via sysfs if that is a
>>>> show-stopper.
>>>> 
>>>> I can't see where the scepticism comes from though.  It is a well defined
>>>> and cleary documented feature of the hardware.  Why not expose it?
> 
> Is it well defined enough that it will work on other chargers, too?

It reports the resistance of the charger’s ID pin. So that works for all chargers connected
to a twl4030. As long as the ID pin goes to a 5 pin USB socket.

Other charger drivers do not need to expose a similar attribute since each twl4030 has its
unique path within the /sys tree.

> 
>>> ABI can never be removed or modified later. So should be really careful before adding it.
>> 
>> Is /sys considered ABI?
> 
> Yes.

You are right: https://lwn.net/Articles/172986/

But I am as well with my doubts: https://lwn.net/Articles/173093/

> 
>> User space developers are always reminded not to rely on /sys nodes.
> 
> No.

Then please explain why I have the impression that it is quite unstable.

BR,
Nikolaus