Re: [PATCH 1/8] crypto: Convert dual BSD 3-Clause/GPL 2.0 boilerplate to SPDX identifier

[Ard Biesheuvel]

On Wed, 7 Jun 2023 at 16:38, Richard Fontana <rfontana@xxxxxxxxxx> wrote:
>
> On Wed, Jun 7, 2023 at 1:42 AM Bagas Sanjaya <bagasdotme@xxxxxxxxx> wrote:
> >
> > Replace license boilerplate for dual BSD-3-Clause/GPL 2.0 (only or
> > later) with corresponding SPDX license identifier.
>
> This is at least the fourth or fifth time (I'm losing track) where you
> have incorrectly assumed a particular non-GPL license text matches a
> particular SPDX identifier without (apparently) checking.
>

What exactly does 'checking' entail here? There is no guidance in
Documentation/process/license-rules.rst on how to perform this
comparison.

Also, checkpatch now complains about missing SPDX identifiers, which
is what triggered this effort. Should it stop doing that?

> Bagas, I urge that you learn more about the nature of SPDX identifiers
> before submitting any further patches at least involving replacement
> of non-GPL notices with SPDX identifiers. For this unprecedented
> license notice replacement initiative to have any legitimacy it must
> attempt to apply SPDX identifiers correctly.
>

Since we're in language pedantic mode: it must do more than attempt,
it must apply them correctly, period.

Arguably, this is an 'attempt to apply SPDX identifiers correctly' on
Bagas's part, which apparently falls short (and I may be guilty of the
same for some arch crypto code)

So what is the ambition here: do we just leave the ambiguous ones as-is?