Re: [PATCH V2 2/4] UefiCpuPkg/CpuDxe: Implement Cpu2 protocol

["Leif Lindholm" <leif.lindholm@linaro.org>]

On Thu, Jul 11, 2019 at 12:24:46PM +0200, Laszlo Ersek wrote:
> Every single source file must have both copyright notice(s) and
> licensing information.
> 
> - Licensing information without copyright holder(s) makes zero sense --
> the licensing terms are offered / dictated *by* the copyright holders.
> 
> - A copyright notice in itself, without a suitable license (including
> the "no license" case) does not provide us with the necessary rights to
> carry the source file in the edk2 project.

Agree with all of the above.

> Regarding "EnableInterrupts.nasm". This looks like a tricky situation.
> Under the above link, the file is currently missing a Microsoft (C)
> notice (the last commit to modify the file is apparently from Microsoft,
> b621971). So that should be addressed in Project Mu first, in my opinion.
> 
> Then, Microsoft should please relicense the file under BSD+Patent, from
> pure 2-BSD. If they disagree, then the 2-BSD is still acceptable for edk2.
> 
> Then, the file could be imported into edk2, carrying
> - the MS (C) notice,
> - the Intel (C) notice (extended to 2019),
> - the license -- BSD+Patent, or else 2-BSD --, expressed with an SPDX ID
> (not open-coded license text).

While I agree it would be better if Microsoft relicensed it as
BSD+Patent, BSD+Patent is a strict superset of 2-clause BSD - so there
is no point in TianoCore to carry 2-Clause BSD code.

We can just add the explicit patent grant on contribution (i.e. change
the license header for a BSD+Patent SPDX tag). And we should, because
otherwise without the tianocore contribution agreement, we open
ourselves up to submarine contributions.

The downside to doing that would be if we wanted the code to be able
to flow in both directions, which is why it would be good if
Microsoft relicensed (and also because then they would be making an
explicit statement that they weren't trying to submarine TianoCore).

Best Regards,

Leif