[libre-riscv-dev] power pc

[Jacob Lifshay]

On Tue, Oct 22, 2019, 17:21 Luke Kenneth Casson Leighton <lkcl at lkcl.net>
wrote:

> just had a very quick 15 min call with hugh, he had a little bit of
> time.

Yay!

> he explained that the OpenPower Foundation is literally in the
> middle of deciding how to accommodate both libre/open and commercial
> interests.
>
hopefully they pick a good model

>
> he said that something akin to ISANX/ISAMUX had actually been
> considered, as a way to allow people to do extensions that don't
> interfere with the main system.  i explained that that's exactly what
> we need... but in an "endorsed" fashion, i.e. it needs to actually
> have an "official sanction" behind it.  they totally get it.
>
yay!

>
> also, very interestingly - bear in mind that this is all fluid - one
> idea being considered is that "Compliant" implementations get access
> to IBM's relevant Patent Portfolio (with a royalty-free non-assert
> agreement).  if you *don't* produce a "Compliant" implementation, that
> patent pool access is *not* granted.  this gives a "Big Stick"
> incentive.
>

one important consideration is if someone suing ibm for something totally
unrelated to power invalidates their patent licenses under above agreement
due to no-assert clause.


> now, all of this is completely under discussion, nothing guaranteed,
> it's quite literally in the middle of being sorted out, and not
> "announced" or finalised.  however it all sounds extremely positive.
>

sounds good.

>
> that's all we're going to get for the next couple of weeks because of
> the Conference in Europe.  he'll let me know when he's available next,
> which should be... mid-november.
>

ok.

will be responding to email about power's atomic operations, since I'm
pretty sure it doesn't work quite like you described.

Jacob

>