SUSE Slowly Shows UEFI Secure Boot Plan

itwbennett writes "One blog post at a time, SUSE is revealing its plan for getting SUSE Linux Enterprise Server (SLES) to boot on machines with UEFI Secure Boot. The short version: 'For now, it seems, SLES will implement an approach similar to that used by Fedora,' writes Brian Proffitt. '[Director of the SUSE Linux Enterprise Olaf] Kirch's first blog entry on Tuesday merely introduced the problem of UEFI Secure Boot. Today's blog only specified the use of the shim bootloader.' Just dying to know what's next? Tune in to the SUSE blog."

Comments

Slashdot has gone batsh*t crazy

[Anonymous Coward]

I'm used to a little bit of healthy paranoia here, but the amount of FUD and flat-out misinformation in Slashdot's UEFI reporting is frankly astonishing. Let's get a few things straight.

UEFI is not a Microsoft technology. It is an industry standard intended intended to replace the archaic x86 BIOS. Microsoft participated in the standard, as did Slashdot favorites Red Hat, Canonical, IBM, and AMD. You can freely download the full specification from the uefi.org website.

Secure Boot is part of the larger UEFI specification. See section 27 for the technical details. Of particular interest to Slashdot readers will be section 27.7 which describes the key update mechanism.

Secure Boot is intended to solve the real-world security problem of boot-time malware. No operating system can defend against malware at boot-time; this would be equivalent to defending against the hardware itself. If it helps, imagine how you would defeat a keylogger embedded in your keyboard.

Secure Boot uses code-signing to defeat boot-time malware. This is the optimal solution and should be full-proof provided (1) the machine is physically secured, and (2) the private keys are secure. (I am defining "full-proof" here to mean the keys and hashes involved are adequately difficuly to brute-force with modern hardware. I am also explicitly discounting scenarios outside of UEFI's area-of-responsibility, such as vulnerabilities in the operating system's signed image.)

For some real irony, see the Slashdot article Windows 8 Secure Boot Defeated. Both the headline and much of the discussion in this article were flat-out wrong. The exploit in question targetted the legacy BIOS and MBR. This is exactly the problem that Secure Boot addresses, and it reinforces the need for this technology.

Secure Boot is not a DRM scheme, nor it is explicitly a tool for Microsoft lock-in. Remember that on x86 platforms, the end-user can edit the key database, and can disable Secure Boot entirely. I concur that Microsoft's treatment of ARM is a dick move, but is also typical for other vendors in that market segment. In either case, remember that Secure Boot is a logical solution to a real-world problem affecting all operating systems, and evaluate it on this merit first.

Just because the technology can be mis-used is no reason to completely boycott it. For my part, I intend to use Secure Boot when it becomes generally available, but only buy parts that allow me to edit the key database.

Links:
UEFI membership list: http://www.uefi.org/join/list/
UEFI specification: http://www.uefi.org/specs/agreement

Re:Slashdot has gone batsh*t crazy

[gomiam]

UEFI is not a Microsoft technology. It is an industry standard intended intended to replace the archaic x86 BIOS.

OOXMLz is a standard as well. Your point being?

Secure Boot uses code-signing to defeat boot-time malware. This is the optimal solution and should be full-proof provided (1) the machine is physically secured, and (2) the private keys are secure.

I guess you meant fool-proof. And it is. It is fool-proof against all those fools who want to decide to run their own code on the computer without having to ask permission beforehand.

Secure Boot is not a DRM scheme, nor it is explicitly a tool for Microsoft lock-in.

True, and yet... it can be used as such. Excuse me, I meant it is already being used as such (see Windows 8 on ARM).

Just because the technology can be mis-used is no reason to completely boycott it. For my part, I intend to use Secure Boot when it becomes generally available, but only buy parts that allow me to edit the key database.

You are free to decide what to use. Just tell me: what will you do when the parts that allow you to edit the key database stop being manufactured? What will you do when, say, the graphics cards you want to use require UEFI to protect their HDMI hardware? It will happen, and rather sooner than later.

Remember: it's not paranoia when they are out to get you. And they are, oh how they are.

Re:what is the point again?

[gomiam]

Theory is closer to practice in theory than in practice. The facts are clear: UEFI lets someone else decide what you can or can not run in your computer.

Think you can disable it? Think again: who is going to care about your being able to disable it when, eventually, Microsoft requires it to be always on on Intel versions of Windows just like they have done on ARM?

It might be easy enough for us....

[complete loony]

Disabling secure boot, or manually installing a new vendor key, may be easy enough for us. But it adds another large hurdle for joe average user to try another operating system. That alone is reason enough to complain about it and object to it.

As it stands now the UEFI standard doesn't specify how the user can install a custom trusted key.

IMHO, hardware vendors should be required to leave the trusted key set empty from the factory. UEFI should then have a standard prompt to enable secure boot and install a key found on bootable media. If Microsoft were forced to guide the user through the same process that a linux installation would require, this process would get the attention it deserves to make it as user friendly and standardised as possible.