*countable infinities only

Thu Jun 14 19:48:51 UTC 2012

On Thu, 14 Jun 2012, Jay Sulzberger wrote:

> Please forgive this top posting.
>
> I will not answer now your radical defense of Microsoft, except to
> say two things:
>
> 1. Your defense would apply also to the decades long fraud of
> Microsoft saying in their EULA that, if you do not run the
> Microsoft OS installed at point of sale of the hardware, you get
> a refund for the OS.  But Microsoft and the hardware vendors
> systematically refused refunds.
>
> 2. Does your defense apply to the case of Microsoft certified devices?

Ah, sorry for typo in above.  Correct line is:

2. Does your defense apply to the case of Microsoft certified ARM devices?

oo--JS.

>
> Your long careful post deserves a careful answer.  I hope to make
> one before one week has gone by.
>
> Adam, thank you for your clear presentation!
>
> oo--JS.
>
>
> On Thu, 14 Jun 2012, Adam Williamson <awilliam at redhat.com> wrote:
>
>> On Thu, 2012-06-14 at 15:03 -0400, Jay Sulzberger wrote:
>> On Thu, 14 Jun 2012, Peter Jones <pjones at redhat.com> wrote:
>> 
>> > On 06/14/2012 01:56 PM, Jay Sulzberger wrote:
>> >
>> >> If Fedora appears to accept that Microsoft should have the
>> >> Hardware Root Key, our side's arguments, in several arenas, are
>> >> weakened.
>> >
>> > Okay, first off, quit hijacking fedora-devel-list for your unrelated DMCA
>> > stuff. It's entirely the wrong place for that.
>> 
>> No.  You intend to grant to Microsoft the power to impede
>> installation of Fedora.  The DMCA can today be used to threaten
>> those who go around the impediment with jail time.
>
> This is, at minimum, arguable. It would require Secure Boot to meet the
> definition of a 'technological protection measure'. According to
> chillingeffects.org, these are defined as:
>
> a measure which "in the ordinary course of its operation, requires the
> application of information, or a process or a treatment, with the
> authority of the copyright owner, to gain access to the work."
>
> I don't immediately see how this can be held to apply to secure boot, as
> it is not intended as a copy protection measure and, as I understand it,
> is not necessarily or indeed often deployed by a copyright holder.
> Especially as the secure boot specification explicitly allows for the
> deployment of user keys, and the disabling (not circumvention) of secure
> boot.
>
>> > Aside from that, you've still got the facts wrong.  What you call the
>> > "Hardware Root Key" the specification calls the "Platform Key" or "PK". 
>> PK
>> > serves a couple of functions - it is the ultimate arbiter of what can and
>> > can't add keys to the system, and it is the determining factor as to 
>> whether
>> > the Secure Boot feature is enabled.  PK will probably not ever be 
>> Microsoft's
>> > key on any system. It'll be a unique to each hardware vendor, or possibly
>> > even unique to various business units within a hardware vendor, or 
>> anything
>> > else they happen to choose. It's completely their decision as to how they
>> > ship this, and nothing we can do will ever change that.
>> 
>> The specification's words are carefully designed to mislead.  As
>> pointed out, if Microsoft has the Hardware Root Key, then
>> "SecureBoot" is not a method of securely booting the hardware you
>> own.
>> 
>> You agree that the key in question is the Hardware Root Key.  You
>> just wrote:
>> 
>> > [the PK] is the ultimate arbiter of what can and can't add keys
>> > to the system, and it is the determining factor as to whether
>> > the Secure Boot feature is enabled.
>> >
>> > The contents of PK are not and have not ever been the question in this > 
>> thread.
>> 
>> Yes, of course, who has the Hardware Root Key is the issue here.
>
> No, it isn't. You are fundamentally misunderstanding secure boot. Peter
> specifically stated that the "hardware root key" (as you call it; the
> platform key, as it is correctly called) is not the key that Microsoft
> will control. As Peter said, hardware manufacturers will control the
> hardware root key for their hardware. What Microsoft is pushing for (and
> requiring for compliance with its certification scheme) is that systems
> are shipped with Microsoft's signing key - not platform key.
>
> Microsoft do not require that Microsoft's be the _only_ signing key. Per
> their certification, it'd be perfectly fine to ship a system with
> Microsoft's key and 500 others. Signing keys are not a 'There Can Be
> Only One' proposition. It's therefore hard to argue that the setup is
> giving Microsoft any kind of exclusive control over anything. There is
> in theory nothing to stop any other organization from acting as a
> signing authority and persuading hardware vendors to install their
> signing key in addition to Microsoft's. The problems with this approach
> are discussed in mjg59's blog post. None of the problems with it is
> 'Microsoft don't want it to happen', because that isn't the case.
>
>> If there is no issue as to who has the Hardware Root Key, why do
>> you propose having Microsoft sign a Fedora key which allows for
>> more convenient installation of Fedora?
>
> Read the initial blog post. Because in practice, no-one else besides
> Microsoft actually wants to go to the considerable trouble and expense
> of acting as a signing authority. _In theory_ any number of bodies could
> do so. _In practice_, no-one has yet showed up with the will and ability
> to do so, and apparently (I am not privy to any private planning in this
> regard) Red Hat doesn't want to either act as one in itself or lead a
> consortium to do so. Given that only Microsoft has committed to being a
> signing authority, and we aren't going to do so ourselves (either 'we'
> as in Red Hat or 'we' as in Fedora), the choices for secure boot boil
> down to either 'don't support it' or 'get our code signed by Microsoft'.
> But it's hard to blame Microsoft, exactly, for no-one else wanting to be
> a signing authority. Microsoft have certainly not done anything to
> preclude the possibility of any other body acting as a signing authority
> and getting their keys on hardware. The only thing you can fairly
> 'blame' Microsoft for is using their influence to effectively enforce
> the default activation of Secure Boot with _indifference_ - but not
> active malice - to the effects on other operating systems. I think
> Microsoft would have had to have tried much harder to preclude other
> people from acting as signing authorities to justify a charge of active
> malice on their part.
> -- 
> Adam Williamson
> Fedora QA Community Monkey
> IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
> http://www.happyassassin.net
>
> -- 
> devel mailing list
> devel at lists.fedoraproject.org
> https://admin.fedoraproject.org/mailman/listinfo/devel
>
>