Re: [PATCH 6/6] eeprom: 93xx46: extend driver to plug into the NVMEM framework

[Srinivas Kandagatla <srinivas.kandagatla@linaro.org>]

On 15/12/15 10:37, Andrew Lunn wrote:
> On Tue, Dec 15, 2015 at 10:26:47AM +0000, Srinivas Kandagatla wrote:
>> + adding Anatolij
>>
>> On 15/12/15 10:17, Andrew Lunn wrote:
>>>> Is there a reason for this driver to be using the old style?
>>>> I can understand the issues with at24/at25 but does this driver also
>>>> suffer from such issues?
>>>
>>> In order to keep backwards compatibility, we need the older file in
>>> /sys. The only other option is to remove it and see if anybody
>>> complains about us breaking the ABI.
>> We should atleast attempt to pitch in this direction, and ask if
>> somebody really cares if the location of the eeprom/nvmem file
>> matters to them?
>
> I expect it does matter.
>
> This driver does not implement the in kernel API for accessing the
> EEPROM. That means all users are in user space. And if this file
> moves, it seems very likely these user space users break.

We have no choice I guess, other than adding the NEMEM_COMPAT flag to 
this driver too :-)



>
>         Andrew
>