lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAK7LNAQ=M0ejV3C8bgjuMxdRR9v=2-GRdXeUjFR6URrrtYPCnA@mail.gmail.com>
Date:   Thu, 23 May 2019 11:49:38 +0900
From:   Masahiro Yamada <yamada.masahiro@...ionext.com>
To:     Joe Perches <joe@...ches.com>
Cc:     Greg KH <gregkh@...uxfoundation.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        linux-spdx@...r.kernel.org
Subject: Re: [GIT PULL] SPDX update for 5.2-rc1 - round 1

On Wed, May 22, 2019 at 3:37 PM Joe Perches <joe@...ches.com> wrote:
>
> On Wed, 2019-05-22 at 13:32 +0900, Masahiro Yamada wrote:
> > On Tue, May 21, 2019 at 10:34 PM Greg KH <gregkh@...uxfoundation.org> wrote:
> []
> > >  - Add GPL-2.0-only or GPL-2.0-or-later tags to files where our scan
> > >     tools can determine the license text in the file itself.  Where this
> > >     happens, the license text is removed, in order to cut down on the
> > >     700+ different ways we have in the kernel today, in a quest to get
> > >     rid of all of these.
> []
> > I have been wondering for a while
> > which version of spdx tags I should use in my work.
> >
> > I know the 'GPL-2.0' tag is already deprecated.
> > (https://spdx.org/licenses/GPL-2.0.html)
> >
> > But, I saw negative reaction to this:
> > https://lore.kernel.org/patchwork/patch/975394/
> >
> > Nor "-only" / "-or-later" are documented in
> > Documentation/process/license-rules.rst
> >
> > In this patch series, Thomas used 'GPL-2.0-only' and 'GPL-2.0-or-later'
> > instead of 'GPL-2.0' and 'GPL-2.0+'.
> >
> > Now, we have a great number of users of spdx v3 tags.
> > $ git grep -P 'SPDX-License-Identifier.*(?:-or-later|-only)'| wc -l
> > 4135
> > So, what I understood is:
> >
> >   For newly added tags, '*-only' and '*-or-later' are preferred.
> >
> > (But, we do not convert existing spdx v2 tags globally.)
> >
> >
> > "
> > Joe's patch was not merged, but at least
> > Documentation/process/license-rules.rst
> > should be updated in my opinion.
> >
> > (Perhaps, checkpatch.pl can suggest newer tags in case
> > patch submitters do not even know that deprecation.)
>
> I'd still prefer the kernel use of a single SPDX style.
>
> I don't know why the -only and -or-later forms were
> used for this patch, but I like it.
>
> I believe the -only and -or-later are more intelligible
> as a trivial reading of
>
>         SPDX-License-Identifier: GPL-2.0
>
> would generally mean to me the original
> GPL-2.0 license without the elision of the
> (or at your option, any later version) bits
>
> whereas
>
>         SPDX-License-Identifier: GPL-2.0-only
>
> seems fairly descriptive.
>
> Is it agreed that the GPL-<v>-only and GPL-<v>-or-later
> forms should be preferred for new SPDX identifiers?


I agree.


> If so, I'll submit a checkpatch patch.

That will be nice.


> I could also wire up a patch to checkpatch and docs to
> remove the /* */
> requirement for .h files and prefer
> the generic // form for both .c and
> .h files as the
> current minimum tooling versions now all allow //
> comments
> .

We have control for minimal tool versions for building the kernel,
so I think // will be OK for in-kernel headers.


On the other hand, I am not quite sure about UAPI headers.
We cannot define minimum tool versions
for building user-space.
Perhaps, using // in UAPI headers causes a problem
if an ancient compiler is used?


BTW, if we allow to use // in header files,
we have no reason to forbid // in assembly files either.

We use *.S files (assembly that should be preprocessed)
instead of *.s files.

So, comments are ripped off by CPP anyway
whichever comment style is used.


--
Best Regards
Masahiro Yamada

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ