[openstack-dev] [nova] Testing concerns around boot from UEFI spec

James Bottomley James.Bottomley at HansenPartnership.com
Wed Dec 16 19:10:11 UTC 2015


On Fri, 2015-12-04 at 08:46 -0500, Sean Dague wrote:
> On 12/04/2015 08:34 AM, Daniel P. Berrange wrote:
> > On Fri, Dec 04, 2015 at 07:43:41AM -0500, Sean Dague wrote:
> > > That seems weird enough that I'd rather push back on our Platinum
> > > Board
> > > member to fix the licensing before we let this in. Especially as
> > > this
> > > feature is being drive by Intel.
> > 
> > As copyright holder, Intel could choose to change the license of
> > their
> > code to make it free software avoiding all the problems. None the
> > less,
> > as above, I don't think this is a blocker for inclusion of the
> > feature
> > in Nova, nor our testing of it.

Actually, it's a bit over simplified to claim this.  The origins of
this clause are in the covenants not to sue in the FAT spec:

http://download.microsoft.com/download/1/6/1/161ba512-40e2-4cc9-843a-92
3143f3456c/fatgen103.doc

It's clause 1(e).  The reason for the clause is a complex negotiation
over the UEFI spec (Microsoft committed to a royalty free
implementation and UEFI needed to use FAT for backward compatibility
with older BIOS).  The problem is that the litigation history no longer
supports claiming the patents are invalid:

http://en.swpat.org/wiki/Microsoft_FAT_patents

As you can see, they're mostly expired (in the US) but the last one
will expire in 2020 (if I calculate the date correctly).  No
corporation (including Intel) can safely release a driver under a
licence that doesn't respect the FAT covenant not to sue without being
subject to potential accusations of contributory infringement.  So,
you're right, Intel could release the FAT 32 driver under a non
-restricted licence as you say but only if they effectively take on
liability for potential infringement for every downstream user ...
amazingly enough they don't want to do that.  Red Hat could do the
same, of course: just strip the additional restrictions clause; Intel
won't enforce it; then Red Hat would take on all the liability ...

The FAT driver is fully separated from the EDKII source:

https://github.com/tianocore/tianocore.github.io/wiki/Edk2-fat-driver

So it can easily be replaced.  The problem is how when every UEFI
driver or update comes on a FAT32 format system.

> That's fair. However we could also force having this conversation 
> again, and pay it forward to the larger open source community by 
> getting this ridiculous licensing fixed. We did the same thing with 
> some other libraries in the past.

The only way to "fix" the licence is to either get Microsoft to extend
the covenant not to sue to all open source projects (I suppose not
impossible given they're making friendlier open source noises) or wait
for the patents to expire.

James




More information about the OpenStack-dev mailing list