Re: [GIT PULL] MMC fixes for v.4.12 rc3 - take 2/2

[Daniel Lezcano]

On 02/06/2017 02:05, Olof Johansson wrote:

[ ... ]

>> For the story, this mmc/sdio is for the WiFi and this one was working for some
>> version of the bootloader which initialized the clock and set the enable line
>> for the chip. That is bad because the WiFi is working as a side effect from the
>> boot loader.
>>
>> In order to properly make support for the WiFi in the kernel, we need to sort
>> out this by:
>>
>>  - Reset the mmc
>>  - Implement the clock at the pmic level
>>  - Add the mfd clock cell
>>  - Add the DT definition
> 
> Ok, that's all fine but it's not really what we would call a fix. Hardware
> bringup and enabling of major drivers is usually done through the release
> cycle and merged in the merge window, not as "fixes" after it's closed.

Yes, I second you on that. That is the reason why I was not really
convinced about this series for -rc.

> It's not causing problems in this case, but if it happens often it runs
> the risk of causing a mess due to conflicting contributions from different
> subsystem maintainers (i.e. in particular the DT contents updates),
> and in general circumvents the development model we're keeping where
> things are tested in -next before merge, etc.
> 
> Anyway, water under the bridge but it's stuff to think about for next time
> around.

Copy that. Thanks!

  -- Daniel


-- 
 <http://www.linaro.org/> Linaro.org â Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog