Uncorrectable freedom and security issues on x86 platforms

[Timothy Pearson]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 04/04/2016 01:46 PM, Paul Boddie wrote:
> On Monday 4. April 2016 19.00.17 Timothy Pearson wrote:
>>
>> ARM is very nice for low end / remote work and embedded systems, there's
>> no question about that.  However, at least in my experience, it is
>> woefully inadequate for non-trivial content creation tasks (editing
>> video, engineering design work, etc.).
> 
> I don't know whether the more recent 64-bit products will make a big 
> difference here. I remarked in another forum that one benefit of a fully-64-
> bit architecture is the ability to address more memory, but then I got flamed 
> for saying that people might need more than 2GB RAM (as I sit and write this, 
> and that earlier message, on a machine with only 1GB RAM).
> 
> Various widely- and wildly- ;-) hyped 64-bit ARM products (Raspberry Pi 3, for 
> example, plus Allwinner A64 stuff) don't appear to physically support more 
> than 2GB RAM, and I've only seen some Freescale i.MX devices and products 
> supporting 4GB, and those were 32-bit devices with extended addressing 
> features.

I was fairly excited about the upcoming 64-bit ARM server products until
the third party benchmarks came out indicating performance less than an
Intel Atom processor (!).  That for a $3k+ price tag and higher power
consumption than Atom if you want server-grade features such as ECC and
PCIe.

> Of course, AMD are getting round to delivering ARM-based products, which might 
> bring more emphasis on performance, but then it's AMD with all the objections 
> you've raised.

AMD has locked away all the documentation for the A1100; given their
track record with the PSP it is highly doubtful they will allow unsigned
(FOSS) boot firmware.
>> That statement was based on actual in-house benchmarks:
>> https://raptorengineeringinc.com/TALOS/power_advantages.php
> 
> Here's the blog post I mentioned:
> 
> https://seravo.fi/2016/linux-and-power8-microprocessors

Interesting; I actually had not seen that one before.

> It could be that people have become aware of POWER again and the "news" 
> travels fast. I suppose that a lot of people thought that POWER was going the 
> way of PA-RISC and Itanium.

That would be my best guess.  Intel is also helping this by keeping
their Xeon prices sky high and their platforms locked down.

> [...]
> 
>> Yes, it's a risk.  However, everything seems to be risky right now.  x86
>> cannot be trusted, POWER itself has somewhat poor power efficiency, and
> 
> That genuinely surprises me. I thought that Motorola/Freescale/NXP had pitched 
> POWER at embedded markets and that apart from IBM's own direction with POWER, 
> it had largely followed ARM into the embedded space.

To be fair, the power efficiency conclusions are based on both direct
measurements of the POWER8 chips and IBM's failure years ago to provide
power efficient chips to Apple for their laptops.  I imagine Freescale
and similar have done a fairly good job in lowering the power
consumption of the embedded POWER architecture.

> [...]
> 
>> Yeah, MIPS seems mostly targeted at embedded systems, and honestly this
>> is an area where RISCV could shine while it tries to ramp up to
>> ARM-level performance.  Regarding POWER, bear in mind the architecture
>> was only opened up for licensing / third party manufacture a couple of
>> years ago, and China is already manufacturing custom CPUs using POWER8.
> 
> One difference here might be that Chinese manufacturers were already 
> apparently making MIPS-compatible designs without an architecture licence, 
> probably because they weren't infringing any patents, or perhaps because it 
> wasn't worth MIPS going after those manufacturers. (OpenRISC was able to exist 
> precisely because the original MIPS patents had expired, as I understand it.)

This is quite possible, yes.  It also puts US/EU users of said chips on
shaky ground unless it can be proven the MIPS architecture is actually
free of patent restrictions.

> IBM, meanwhile, has historically been aggressive with microarchitecture 
> patents: I recall some remarks about litigation against Sun and ARM, with the 
> latter successfully defending because one of the originators of the ARM 
> architecture dug up some old software from the Acorn era that supposedly 
> demonstrated prior art on whichever patents were involved.

It looks like IBM has been changing their tune ever since they realised
they were mishandling POWER to the point that x86 and ARM would become
permanently dominant.  Now that a third party foundation controls POWER
architecture licensing and patents I don't think this will be as much of
a concern (at least not any more than ARM is right now).

> But anyway, I see that the laptop project that Tobias linked to intends to use 
> Freescale products, which means that a reliable manufacturer is standing 
> behind the architecture and actually making products available, so maybe there 
> is some mileage in choosing POWER-based products for open hardware designs.

This is good to know as well.  I always thought it would be POWER on
workstations and servers with ARM laptops remoting in; perhaps it could
actually be various flavors of POWER throughout the entire ecosystem.
At least there could be some real competition in the laptop sphere,
unlike the workstation / server realm as it stands today.

- -- 
Timothy Pearson
Raptor Engineering
+1 (415) 727-8645 (direct line)
+1 (512) 690-0200 (switchboard)
http://www.raptorengineeringinc.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJXArpOAAoJEK+E3vEXDOFbQ5gH/2xW7+zrXDWLgO0QuzW80oCE
tBaSXYDpIJN7+zhcHFNWV2AMN0bWeZD7oxDbsqTCEs5cGi3SdlA8leuNoheJTxGF
N1wAb/sQFP+WNu1PgHxFMcsJRpFMYvAuX3PyiA/ldbSwcHOoo/1tZZBtnZ8adhG3
cQUnZC5CaseIVLwllmer2uENPlZrkDihPIGbEK9v6ITfSzGMNkvkgYuO0TGzufCw
i36HaUigZq5ToVsvnGlwr2DDoLBa6UVeO1gYn1Uoy/sMGY8hOTyjZa1r7ECD2sgU
5i8kQ5IunGXrEDd/5ju4NhrjYnc+XeNFQRlwFlJxOku220477E6fq1C2Cyn5p2A=
=AC4q
-----END PGP SIGNATURE-----