Re: [edk2-platforms: PATCH v2 03/10] Marvell/Cn9130Db: Add DeviceTree

["Laszlo Ersek" <lersek@redhat.com>]

On 08/19/19 19:14, Leif Lindholm wrote:
> +Stewards (don't panic!)
> 
> On Fri, Aug 16, 2019 at 11:03:42PM +0200, Marcin Wojtas wrote:
>> Hi Leif,
>>
>> pt., 16 sie 2019 o 19:10 Leif Lindholm <leif.lindholm@linaro.org> napisał(a):
>>>
>>> On Thu, Aug 15, 2019 at 04:54:07AM +0200, Marcin Wojtas wrote:
>>>> This patch adds device tree sources which are common for Cn913x SoCs
>>>> and the CN9130 development board (variant A). Wiring up of support
>>>> will be done in the follow-up commits.
>>>>
>>>> Signed-off-by: Marcin Wojtas <mw@semihalf.com>
>>>
>>> Again, I've not reviewed the contents extensively, *but* this patch
>>> cannot go into edk2-platforms - it has GPL licensed content.
>>> Was it meant for edk2-non-osi? Because we could take it there.
>>
>> Thanks for catching this - my oversight. The top-level DT files should
>> also be dual licenced (GPL2.0/MIT), same as recently submitted to
>> Linux. So IMO no need to go for edk2-non-osi.
> 
> Right. MIT *is* an acceptable license, but no other licenses than
> BSD+Patent are blanket approved - it's a case by case basis.
> 
> We still have a gap with regards to the explicit patent grant since
> the dropping of the TianoCore Contribution Agreement. While this is
> less problematic for something like .dts files, I don't think it can
> be argued that the problem does not exist for them.
> 
> My preference for resolving this would be that anything merged into
> !edk2-non-osi (and, yes, we really ought to create an
> edk2-other-license or whatever) needs to use a license compatible with
> BSD+Patent (MIT is) and is registered as:
> 
> SPDX-License-Identifier: BSD-2-Clause-Patent AND <other license>
> 
> TL;DR: if you want to get the set merged quickly, edk2-non-osi is the
> best route for these files.

no particular comments from me; what Leif suggests looks fine to me
(from orbit anyway)

thanks
Laszlo