Re: [platforms: PATCH 01/13] Marvell/Armada: Introduce platform initialization driver

[Marcin Wojtas <mw@semihalf.com>]

2017-10-10 22:36 GMT+02:00 Ard Biesheuvel <ard.biesheuvel@linaro.org>:
> On 10 October 2017 at 16:26, Leif Lindholm <leif.lindholm@linaro.org> wrote:
>> On Tue, Oct 10, 2017 at 05:06:42PM +0200, Marcin Wojtas wrote:
>>> 2017-10-10 17:03 GMT+02:00 Leif Lindholm <leif.lindholm@linaro.org>:
>>> > On Tue, Oct 10, 2017 at 04:45:10PM +0200, Marcin Wojtas wrote:
>>> >> Hi Leif,
>>> >>
>>> >> 2017-10-10 16:37 GMT+02:00 Leif Lindholm <leif.lindholm@linaro.org>:
>>> >> > On Mon, Oct 09, 2017 at 07:00:50PM +0200, Marcin Wojtas wrote:
>>> >> >> In order to enable modification of dynamic PCD's for the libraries
>>> >> >> and DXE drivers, this patch introduces new driver. It is
>>> >> >> executed prior to other drivers. Mpp, ComPhy and Utmi libraries
>>> >> >> initialization were moved from PrePi stage to DXE.
>>> >> >>
>>> >> >> To force the correct driver dispatch sequence, introduce a protocol GUID
>>> >> >> and install the protocol as a NULL protocol when PlatInitDxe executes.
>>> >> >>
>>> >> >> Contributed-under: TianoCore Contribution Agreement 1.1
>>> >> >> Signed-off-by: Marcin Wojtas <mw@semihalf.com>
>>> >> >> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>>> >> >
>>> >> > What does Ard's Signed-off-by signify here?
>>> >> > (I know the authorship on some of these is a bit blurred, since you've
>>> >> > been working together, but I'd like to be clear.)
>>> >>
>>> >> These were the lines, introducing/installing protocol GUID stuff. It
>>> >> was in a small separate patch, but I squashed it into bigger one.
>>> >
>>> > Personally, I would in this instance do:
>>> > <me>
>>> > Ard
>>> > <me>
>>> >
>>> > It's verbose, but reasonably clear.
>>> >
>>>
>>> How about:
>>>
>>> In order to enable modification of dynamic PCD's for the libraries
>>> and DXE drivers, this patch introduces new driver. It is
>>> executed prior to other drivers. Mpp, ComPhy and Utmi libraries
>>> initialization were moved from PrePi stage to DXE.
>>>
>>> Signed-off-by: Marcin Wojtas <mw@semihalf.com>
>>>
>>> To force the correct driver dispatch sequence, introduce a protocol GUID
>>> and install the protocol as a NULL protocol when PlatInitDxe executes.
>>>
>>> Contributed-under: TianoCore Contribution Agreement 1.1
>>> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>>> Signed-off-by: Marcin Wojtas <mw@semihalf.com>
>>>
>>> ?
>>>
>>> Was that, what you meant?
>>
>> I think Contibuted-under: still needs to come first.
>>
>> I don't think we have an explicit policy for how to deal with
>> multi-contributor patches. The ones we do see tend to just keep a
>> single commit message and list the contributors.
>>
>> In Linux. it would be something like
>> Signed-off-by: Marcin Wojtas <mw@semihalf.com>
>> [Introduce protocol GUID to force correct driver dispatch order]
>> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>> Signed-off-by: Marcin Wojtas <mw@semihalf.com>
>>
>> I would be quite happy to use the same format here.
>>
>
> Well, Tianocore still conflates authorship with a statement regarding
> the origin of the contribution. I wonder how this is supposed to work
> when Linaro engineers such as myself contribute code that was authored
> by engineers working in member companies, e.g., Socionext. The license
> and the contract that company has with Linaro give me the right to
> contribute that code, but that does not make me the author, and I
> cannot add a Signed-off-by that wasn't present when we received the
> code (even if I knew the name of the author)

I think it's fairly easy thing, needlessly twisted... How does above
reflect the requirement to add contributor sign-off to someone else's
patch (with his authorship and original sign-off - should they be
removed?)?

Anyway, let's make a quick decision here - should I submit patch with
linux-like signatures and description? Or should I split the patches?

Best regards,
Marcin