The Tricky World of Securing Firmware
blogs.intel.com
blogs.intel.com
As you clearly say, your job is to serve your clients, the manufacturers who do not want their BIOSes to be replaced with something else like coreboot.
You say "Intel has to consider the needs of the system manufacturer along with the end user", but the balance is obviously slanted in one direction. The reason is simple: we end users are not your customers.
As pointed out before, the trickiest issues are with Intel ME and AMT, who can not be removed except on the X200. Otherwise, the luckiest machines will reboot every 30 minutes only. The others won't boot.
You say "Making firmware more open and more secure is an interesting balancing act, and I hope we work towards getting it right". I'm sorry but you don't. Having a backdoor that the physical owner of the machine can not remove is neither secure nor open. If there was a jumper than could be used by the knowledgeable end user to override all his, yes. There is no such things, only barriers.
This is a nice PR attempt to deflect the heat towards Lenovo (for the thinkpads), but sorry, that's vacuous.
There is a moral responsibility in a given design.
At the moment, this is at best the moral equivalent of making weapons without a safety and saying it's not your fault if they can harm the end user, and that the end user should direct the complains to whoever is selling these defective guns. Sorry, but I disagree.
So while Intel could make a principled stand here (in favor of end-users), it's ODMs/OEMs that have to work with primarily. CPUs directly bought by end-users are probably a pretty minor part of their revenue, too.
[edit to add:] Of course, I'd still prefer them to get rid of Verified Boot, and have OEMs decide between Measured Boot or no verification at all.
I can easily imagine EU to pass new law, which would require "signed bootloader bypass" switch on all newly sold laptops and workstation. Great way to milk Intel, MS, Apple and other manufactures for couple of billions euro.
Why should that be true? Source?
I have a couple of decade-old laptops that are perfectly happy to run Linux, and Windows 7 won't even install on them because they don't have enough memory.
It's also conceivable that a different OS would have a lower thermal profile, resulting in longer hardware lifetime on average.
2. Newer applications tend to not be installable on older OSes. For example, newer IE won't run on Windows XP.
3. Therefore, older OSes become useless after a while, if you're talking to the outside world.
4. Therefore, being able to install a new OS on your hardware, can make that hardware useful longer.
5. Newer OSes may not always support older hardware. Also, companies that produce OSes (Microsoft, Apple, Red Hat, etc) may go out of business.
6. Therefore, being able to choose from a wider variety of OSes to install, increases the chances that you'll be able to find one that you can use.
7. Therefore, having more choice in OSes makes your hardware likely to be useful longer.
(I'm not saying definitively that one is better than the other, only offering both perspectives.)
Firefox has been doing it, Chrome has been doing it, damn near everyone is doing it.
How many individuals really need the first amendment? The Saudi kids in my undergrad PoliSci classes loved to point this out. "Who needs freedom? I'll take my mercedes."
By default, the firmware is locked, but there's a (warranty voiding if you're outside of the EU, which I don't mind) screw on the bottom of the machine, which allows you to flash your own firmware.
This solves one of the problems(upgrading firmware when vendor support runs out). One of the things that would be nice to explore is creating "flashing keys" on first boot. Maybe put them on a password-protected USB key or cloud share that can be held on to until you want to upgrade or modify firmware.
Another solution might be the ability to take out that screw, load your own key, then reinsert the screw(tripping some sort of hardware fuse to lock it to that key, maybe).
I don't like the tone of this article, it downplays the importance of hobbyists, and presents a false dichotomy: firmware which can't be modified by the owner of the device, versus firmware that can be modified by anyone, including remote attackers.
We can have our cake and eat it too, we just need to stop giving up so early.
Is it not the case that these days, harddisks and who-knows-what-other-subsystem have their own firmware which can be upgraded and so infected, separately, right? Infected harddisk firmware being mode of attack described for the Equations Group as far as I can tell.
[1] http://www.networkworld.com/article/2885814/security0/has-eq...