Re: clarification on -only and -or-later

[Thomas Gleixner <tglx@linutronix.de>]

[-- Attachment #1: Type: text/plain, Size: 1727 bytes --]

On Wed, 22 May 2019, Greg KH wrote:
> On Wed, May 22, 2019 at 09:53:48AM -0400, Allison Randal wrote:
> > On 5/22/19 2:23 PM, Greg KH wrote:
> > > On Tue, May 21, 2019 at 12:05:37PM -0600, J Lovejoy wrote:
> > >>
> > >> I think what I was looking for here, was confirmation as to whether we
> > >> want to do the “literal” GPL-1.0-or-later option that the license
> > >> provides for, or trigger the option to “choose any version” and go
> > >> with GPL-2.0-or-later for consistency of v2 across the kernel and for
> > >> other reasons I believe you raised regarding GPL-1.0
> > > 
> > > I don't understand.  Can you point to any files in the kernel where we
> > > have used the "GPL-1.0+" marking incorrectly?
> > 
> > Jilayne's question wasn't about current usage in the kernel, it was
> > about what we should do in this cleanup process when we get to files
> > where the license notice doesn't have an explicit GPL version number or
> > include the "or later" text. Thomas hasn't gotten to those patterns yet
> > in his batch processing.
> 
> Ah, ok.
> 
> But note, we have already marked such files as "GPL-1.0+" in the past,
> so any change in that behavior would require us go back and change what
> we did, showing the justification for that.
> 
> I would stick to the rule of what we have already done in these cases,
> it's simpler and seems to make sense of a crazy situation.

Yes and no. There is a good reason to remove the GPL1+ crap completely and
I already got permission from Redhat to change their GPL/GPL'ed notices to
GPL-2.0-or-later. So once we get to that pile we might at least try to talk
with the copyright holders and clarify it. Each odd license which gets
removed is a win.

Thanks,

	tglx