[coreboot] Re: Web site revamp

[Timothy Pearson]

----- Original Message -----
> From: "Patrick Georgi" 
> To: "Timothy Pearson" 
> Cc: "ron minnich" , "David Hendricks" , "coreboot"
>
> Sent: Monday, September 2, 2019 3:26:45 AM
> Subject: Re: [coreboot] Re: Web site revamp
> Am Mo., 2. Sept. 2019 um 10:17 Uhr schrieb Timothy Pearson <
> [e-mail]>:
> > this text needs to be completely rewritten to clearly show where the
> > limits are on modern x86 platforms.
>
> No, you want it rewritten to be able to better advertise Talos. This
> project is not in the business of advertising your project, even though we
> do (It's mentioned in the distributions list despite there being no
> coreboot port for it yet).

No, that's not what I want at all.  Don't conflate the honest desire to deal with what I consider a significant issue on the x86 side with something else.  Please do note that nowhere did I say Talos, in fact I specifically indicated several systems with open ISA and owner controlled CPUs (RISC-V, MIPS, and yes POWER).  *Any* of these actually meet the goal of an owner controlled, open source boot right here and now, versus hoping one of the x86 vendors will release ME/PSP unlocked CPUs.  I'd be quite happy if even one of the RISC-V boards (that Raptor doesn't make) was the poster child for coreboot at this point -- anything other than the signed binary situation we have now with x86.
> > (as a result) what coreboot cannot fix without a change of direction from
> > the silicon vendors.
>
> You're promoting a system that comes with native NVLink support. Where's
> the nvidia GPU that runs with unsigned firmware?
> Get off your high horse, it limps.

I think this is the point where rational discussion has stopped and personal attacks have started, so I'll take a break and see where things go in my absence.  Just for the record, we don't (and never have) promoted NVLink, our systems don't have the required firmware binaries installed to make it work, *and* we preferentially promote and ship open driver AMD GPUs.  We even waited to introduce CAPI until OpenCAPI with its open firmware was available.  The only thing I see limping here is the x86 horse, hobbled by its signed, proprietary, black box coprocessors and NDA-restricted documentation.