Re: Applied "regulator: wm8994: Pass descriptor instead of GPIO number" to the regulator tree

From: Marek Szyprowski <m.szyprowski@samsung.com>
To: Richard Fitzgerald <rf@opensource.cirrus.com>,
	Charles Keepax <ckeepax@opensource.cirrus.com>
Cc: Mark Brown <broonie@kernel.org>,
	Linus Walleij <linus.walleij@linaro.org>,
	Lee Jones <lee.jones@linaro.org>,
	Liam Girdwood <lgirdwood@gmail.com>,
	linux-kernel@vger.kernel.org, patches@opensource.cirrus.com
Subject: Re: Applied "regulator: wm8994: Pass descriptor instead of GPIO number" to the regulator tree
Date: Tue, 20 Nov 2018 18:12:58 +0100	[thread overview]
Message-ID: <b7edc4cc-c433-a154-3d70-d13d930f6278@samsung.com> (raw)
In-Reply-To: <ea2b959c-0d04-6a50-4d69-3f03576fdac9@opensource.cirrus.com>

Hi Richard,

On 2018-11-20 17:57, Richard Fitzgerald wrote:
> On 20/11/18 16:34, Marek Szyprowski wrote:
>> Hi Richard,
>>
>> On 2018-11-20 17:16, Richard Fitzgerald wrote:
>>> On 20/11/18 15:56, Marek Szyprowski wrote:
>>>> Hi Charles,
>>>>
>>>> On 2018-11-20 16:36, Charles Keepax wrote:
>>>>> On Tue, Nov 20, 2018 at 03:32:15PM +0000, Charles Keepax wrote:
>>>>>> On Tue, Nov 20, 2018 at 03:58:59PM +0100, Marek Szyprowski wrote:
>>>>>>> On 2018-11-20 15:47, Charles Keepax wrote:
>>>>>>>> On Tue, Nov 20, 2018 at 02:43:32PM +0100, Marek Szyprowski wrote:
>>>>>>>>> On 2018-05-17 18:41, Mark Brown wrote:
>>>>>>>>>> Subject: [PATCH] regulator: wm8994: Pass descriptor instead of
>>>>>>>>>> GPIO number
>>>>>>>>> This patch causes following kernel warning on Samsung Exynos4412
>>>>>>>>> based
>>>>>>>>> Trats2 board:
>>>>>>>>>
>>>>>>>>> wm8994 4-001a: Failed to get supply 'DBVDD1': -517
>>>>>>>>> wm8994 4-001a: Failed to get supplies: -517
>>>>>> This is really weird, because the error in your log relates to
>>>>>> DBVDD1 which is an independent regulator supplied by a separate
>>>>>> regulator. I am really having some difficulty seeing how the
>>>>>> patch interfers. It is definitely that patch which causes the
>>>>>> issue, like you revert it and things work again?
>>>>> Wait does the board still boot just you have an extra probe defer
>>>>> now? Or does it actually fail?
>>>>
>>>> The board boots fine. The only new thing is the mentioned warning,
>>>> which
>>>> I would
>>>>
>>>> like to have fixed.
>>>>
>>>>
>>>> Best regards
>>>>
>>>
>>> -517 is EPROBE_DEFER. This isn't something  that needs "fixing" unless
>>> the
>>> driver is never able to probe.
>>>
>>> If the wm8994 eventually probes ok after retries it's not a problem,
>>> it's normal kernel behaviour.
>>>
>>> If the wm8994 driver never manages to probe successfully it should
>>> mean that
>>> the driver which supplies DBVDD1 isn't available.
>>
>> Deferred probe was there already. This patch however introduced the
>> warning from gpiolib and I would like to have it fixed somehow. In both
>
> I don't follow what it is you want, are you asking that it shouldn't
> probe
> defer, or that it shouldn't log the reason why it deferred?
>
>> cases (with this patch and before it) the wm8994 driver probes okay -
>> when the required regulators are finally available.
>
> Sounds like all is ok and working as expected.
> If this is causing you a problem you'll need to provide more
> explanation of
> what problem you have so we can understand.

I'm asking for fixing the code (or giving a hint how to fix it) in a way
that gpiolib will not complain. My initial reply [1] had a gpiolib
warning, which is the issue. Deferred probe is the way to trigger it. My
fault that I didn't explain it literally what is the issue.

[1] https://lkml.org/lkml/2018/11/20/997

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland