[PATCH] ARM: imx_v6_v7_defconfig: enable driver of the LTC3676 PMIC

[Marek Vasut]

On 8/14/21 1:24 PM, Shawn Guo wrote:
> On Sat, Aug 14, 2021 at 09:45:03AM +0200, Marek Vasut wrote:
>> On 8/14/21 7:40 AM, Shawn Guo wrote:
>>> On Tue, Aug 10, 2021 at 07:36:24AM +0200, Marek Vasut wrote:
>>>> On 8/5/21 3:50 PM, Pascal Zimmermann wrote:
>>>>> The LTC3676 is a PMIC which is used on some i.MX6 based boards (like the
>>>>> DHCOM i.MX6 Quad SoM), it was first used on the GW Ventana board, enable
>>>>> LTC3676 driver in imx_v6_v7_defconfig.
>>>>>
>>>>> Fixes: e3946fe805053 ("ARM: dts: add Gateworks Ventana support")
>>>
>>> I don't think the Fixes tag is appropriate here.  Dropped it and applied
>>> patch.
>>
>> Can you please elaborate why do you think the Fixes tag is not appropriate ?
>>
>> Neither system can work properly without the PMIC driver, so I believe the
>> Fixes tag is correct and this should be backported to older LTS kernel
>> releases.
> 
> It's enabling a feature (driver) rather than fixing a bug or regression.
> If you want to enable it for LTS kernel, go ahead to back port it.
> Do not abuse the Fixes tag.

I disagree with this. Without the PMIC driver, the system is basically 
unusable. That is because the internal regulators in the MX6 depend on 
the PMIC, so those fail to probe, and in turn all the other components 
which depend on those internal regulators also fail to probe (that is 
everything which depends on anatop regulators).

If I take linux-stable, compile it for a system that is supported in 
there, with the imx_v6_v7_defconfig which is part of it, I would expect 
the result to work properly, not just barely boot and be barely usable.

Therefore, I believe the Fixes tag here is justified.