Re: [PATCH] media: v4l: xilinx: Use SPDX-License-Identifier

From: Mauro Carvalho Chehab
Date: Fri Dec 15 2017 - 04:34:31 EST


Em Fri, 15 Dec 2017 00:02:21 +0200
Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx> escreveu:

> Hi Mauro,
>
> On Thursday, 14 December 2017 23:50:03 EET Mauro Carvalho Chehab wrote:
> > Em Thu, 14 Dec 2017 21:57:06 +0100 Greg KH escreveu:
> > > On Thu, Dec 14, 2017 at 10:44:16PM +0200, Laurent Pinchart wrote:
> > >> On Thursday, 14 December 2017 22:08:51 EET Greg KH wrote:
> > >>> On Thu, Dec 14, 2017 at 09:05:27PM +0200, Laurent Pinchart wrote:
> > >>>> On Thursday, 14 December 2017 20:54:39 EET Joe Perches wrote:
> > >>>>> On Thu, 2017-12-14 at 20:37 +0200, Laurent Pinchart wrote:
> > >>>>>> On Thursday, 14 December 2017 20:32:20 EET Joe Perches wrote:
> > >>>>>>> On Thu, 2017-12-14 at 20:28 +0200, Laurent Pinchart wrote:
> > >>>>>>>> On Thursday, 14 December 2017 19:05:27 EET Mauro Carvalho Chehab
> > >>>>>>>> wrote:
> > >>>>>>>>> Em Fri, 8 Dec 2017 18:05:37 +0530 Dhaval Shah escreveu:
> > >>>>>>>>>> SPDX-License-Identifier is used for the Xilinx Video IP and
> > >>>>>>>>>> related drivers.
> > >>>>>>>>>>
> > >>>>>>>>>> Signed-off-by: Dhaval Shah <dhaval23031987@xxxxxxxxx>
> > >>>>>>>>>
> > >>>>>>>>> Hi Dhaval,
> > >>>>>>>>>
> > >>>>>>>>> You're not listed as one of the Xilinx driver maintainers. I'm
> > >>>>>>>>> afraid that, without their explicit acks, sent to the ML, I
> > >>>>>>>>> can't accept a patch touching at the driver's license tags.
> > >>>>>>>>
> > >>>>>>>> The patch doesn't change the license, I don't see why it would
> > >>>>>>>> cause any issue. Greg isn't listed as the maintainer or copyright
> > >>>>>>>> holder of any of the 10k+ files to which he added an SPDX license
> > >>>>>>>> header in the last kernel release.
> > >>>>>>>
> > >>>>>>> Adding a comment line that describes an implicit or
> > >>>>>>> explicit license is different than removing the license
> > >>>>>>> text itself.
> > >>>>>>
> > >>>>>> The SPDX license header is meant to be equivalent to the license
> > >>>>>> text.
> > >>>>>
> > >>>>> I understand that.
> > >>>>> At a minimum, removing BSD license text is undesirable
> > >>>>>
> > >>>>> as that license states:
> > >>>>> * * Redistributions of source code must retain the above copyright
> > >>>>> * notice, this list of conditions and the following disclaimer.
> > >>>>>
> > >>>>> etc...
> > >>>>
> > >>>> But this patch only removes the following text:
> > >>>>
> > >>>> - * This program is free software; you can redistribute it and/or
> > >>>> modify
> > >>>> - * it under the terms of the GNU General Public License version 2 as
> > >>>> - * published by the Free Software Foundation.
> > >>>>
> > >>>> and replaces it by the corresponding SPDX header.
> > >>>>
> > >>>>>> The only reason why the large SPDX patch didn't touch the whole
> > >>>>>> kernel in one go was that it was easier to split in in multiple
> > >>>>>> chunks.
> > >>>>>
> > >>>>> Not really, it was scripted.
> > >>>>
> > >>>> But still manually reviewed as far as I know.
> > >>>>
> > >>>>>> This is no different than not including the full GPL license in
> > >>>>>> every header file but only pointing to it through its name and
> > >>>>>> reference, as every kernel source file does.
> > >>>>>
> > >>>>> Not every kernel source file had a license text
> > >>>>> or a reference to another license file.
> > >>>>
> > >>>> Correct, but the files touched by this patch do.
> > >>>>
> > >>>> This issue is in no way specific to linux-media and should be
> > >>>> decided upon at the top level, not on a per-subsystem basis. Greg,
> > >>>> could you comment on this ?
> > >>>
> > >>> Comment on what exactly? I don't understand the problem here, care to
> > >>> summarize it?
> > >>
> > >> In a nutshell (if I understand it correctly), Dhaval Shah submitted
> > >> https:// patchwork.kernel.org/patch/10102451/ which replaces
> > >>
> > >> +// SPDX-License-Identifier: GPL-2.0
> > >> [...]
> > >> - *
> > >> - * This program is free software; you can redistribute it and/or modify
> > >> - * it under the terms of the GNU General Public License version 2 as
> > >> - * published by the Free Software Foundation.
> > >>
> > >> in all .c and .h files of the Xilinx V4L2 driver
> > >> (drivers/media/platform/
> > >> xilinx). I have reviewed the patch and acked it. Mauro then rejected it,
> > >> stating that he can't accept a change to license text without an
> > >> explicit ack from the official driver's maintainers. My position is
> > >> that such a change doesn't change the license and thus doesn't need to
> > >> track all copyright holders, and can be merged without an explicit ack
> > >> from the respective maintainers.
> > >
> > > Yes, I agree with you, no license is being changed here, and no
> > > copyright is either.
> > >
> > > BUT, I know that most major companies are reviewing this process right
> > > now. We have gotten approval from almost all of the major kernel
> > > developer companies to do this, which is great, and supports this work
> > > as being acceptable.
> > >
> > > So it's nice to ask Xilinx if they object to this happening, which I
> > > guess Mauro is trying to say here (in not so many words...) To at least
> > > give them the heads-up that this is what is going to be going on
> > > throughout the kernel tree soon, and if they object, it would be good to
> > > speak up as to why (and if they do, I can put their lawyers in contact
> > > with some lawyers to explain it all to them.)
> >
> > Yes, that's basically what I'm saying.
> >
> > I don't feel comfortable on signing a patch changing the license text
> > without giving the copyright owners an opportunity and enough time
> > to review it and approve, or otherwise comment about such changes.
>
> If I understand you and Greg correctly, you would like to get a general
> approval from Xilinx for SPDX-related changes, but that would be a blanket
> approval that would cover this and all subsequent similar patches. Is that
> correct ? That is reasonable for me.

I doubt any Company's legal department would give a blanket approval for
others to touch on their licensing text.

>
> In that case, could the fact that commit
>
> commit 5fd54ace4721fc5ce2bb5aef6318fcf17f421460
> Author: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Date: Fri Nov 3 11:28:30 2017 +0100
>
> USB: add SPDX identifiers to all remaining files in drivers/usb/
>
> add SPDX headers to several Xilinx-authored source files constitute such a
> blanket approval ?

If you look at this patch's summary:

651 files changed, 651 insertions(+)

And on the patch contents itself, it is just adding a SPDX header.
It doesn't remove any text from the license.

On Dhaval's patch, it is not only adding SPDX header. It is also removing
the legal text from it:

diff --git a/drivers/media/platform/xilinx/xilinx-dma.c b/drivers/media/platform/xilinx/xilinx-dma.c
index 522cdfdd3345..2e5daf7dba1a 100644
--- a/drivers/media/platform/xilinx/xilinx-dma.c
+++ b/drivers/media/platform/xilinx/xilinx-dma.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: GPL-2.0
/*
* Xilinx Video DMA
*
@@ -6,10 +7,6 @@
*
* Contacts: Hyun Kwon <hyun.kwon@xxxxxxxxxx>
* Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
*/

And that's the part I'm more concerned about: we should give Xilinx
enough time to review and approve such change.

Thanks,
Mauro