News for hackers. Stuff that matters.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by Voultapher on Wednesday December 06, 2023 @12:21PM (#38546810)
I'm shocked!

Firmware vendors, the last bastion of high quality memory-safe C code, solving complex problems. Code firmly audited and formally verified to the highest degree. And now you tell me these people, with the highest known software quality got "parsing image formats" wrong?

It's truly zebras all the way down https://youtu.be/fE2KDzZaxvE

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by bcantrill on Wednesday December 06, 2023 @04:37PM (#38549850)
Indeed -- and not coincidentally, this is why we don't have UEFI at all on the Oxide compute sled.[0]

[0] https://www.osfc.io/2022/talks/i-have-come-to-bury-the-bios-...

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by Voultapher on Wednesday December 06, 2023 @05:24PM (#38550332)
Hey, big fan of your work. Keep up the great work.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by gpderetta on Wednesday December 06, 2023 @01:27PM (#38547696)
Firmware and BIOS vendors, world renown for the quality of their software.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by jsight on Wednesday December 06, 2023 @03:56PM (#38549352)
Of course their software is the best. Some of these vendors are subject to regulatory, contractual legal requirements.

That always produces the best software, unlike the "move fast, break things" people in other domains.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by imoverclocked on Wednesday December 06, 2023 @03:58PM (#38549374)
I’ve been in the move fast and break things domain. Unsurprisingly, things are often broken there too.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by jsight on Wednesday December 06, 2023 @04:06PM (#38549464)
I've yet to run into place where they aren't. But run fast, fix fast often works better than constant debate and red tape.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by a_subsystem on Wednesday December 06, 2023 @04:27PM (#38549707)
Currently working on a new federal report requiring us to send 7 fields to them yearly. They have provided a 1,670 page document to 'help' us send them the data.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by jacquesm on Wednesday December 06, 2023 @08:28PM (#38551749)
Hehe. "We're from the government and we're here to help you!"
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by sonicanatidae on Friday December 08, 2023 @04:57PM (#38575124)
I'm somehow positive that the folks that fix the things that break are thrilled on the daily that someone did something quickly, without really understanding the ramifications, when they could have asked about it first. ;)

You have a valid point about red tape, but does it really amount to 2 steps forward, or is it the same one step, but 2 forward, 1 back?

I work in environments where moving fast is usually bad (Enterprise/Healthcare), so I'm honestly asking.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by imoverclocked on Tuesday December 12, 2023 @02:21AM (#38610617)
IMO, the step analogy breaks down fast. Almost like comparing dancing to hiking. Eg: you can take a lot more steps dancing and get nowhere…

Knowing where you want to go and keeping your focus on that direction is probably the most important aspect of any environment. Iterating fast is a double edged sword that both allows quick fixes to happen while also encouraging an environment where those quick fixes are needed.

Iterating more meticulously feels a lot slower but can produce just as fast… sometimes a lot faster. For me, there is also an inherent stability component to slower iteration.

Perhaps, the best is neither approach on their own but a clever combination of the two?

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by sonicanatidae on Tuesday December 12, 2023 @09:39AM (#38612710)
Personally, I prefer uptime. Uptime is king, in my world and I can't maintain that if people are haphazardly breaking shit throughout the Enterprise.

I work in Healthcare, so the needs are a little different.

I'm glad I work in the field I do. I'm not sure my personality lends itself to move fast and break shit.

Like all things, YMMV. :)

Thanks for answering my question.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by WesolyKubeczek on Wednesday December 06, 2023 @04:05PM (#38549451)
Since it's always turtles^Wsubcontractors all the way down, I wouldn't be surprised in the least if some of the actual mission critical stuff turned out to be written by some unpaid intern at an underpaid outsourcing shop somewhere in Elbonia. And if it would be in fact the most solid and reliable piece of the firmware.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by toyg on Wednesday December 06, 2023 @04:38PM (#38549867)
> Elbonia

Showing your age there, pal. And yes, the '90s were awesome.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by dingnuts on Wednesday December 06, 2023 @03:23PM (#38548984)
genuine question: is this sarcasm? how dumb am i for not being able to tell?
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by sonicanatidae on Wednesday December 06, 2023 @03:40PM (#38549206)
You're not dumb. You simply lack the specific body of knowledge and industry experience that provides the context for you to determine this.

That's just life. :)

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by jacquesm on Wednesday December 06, 2023 @03:30PM (#38549072)
Yes.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by nine_k on Wednesday December 06, 2023 @01:18PM (#38547572)
On topic: does the open-source Coreboot fare any better here? (I hope it does; the article does not mention it.)
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by est31 on Wednesday December 06, 2023 @01:28PM (#38547712)
coreboot just initializes the hardware, the logo is something that the payload displays: https://www.coreboot.org/Payloads

The most typically used payload is u-boot: https://docs.u-boot.org/en/latest/

u-boot supports specifying splash screens via "splashfile", but it seems only bmp and maybe some raw image format are supported: https://github.com/u-boot/u-boot/blob/2f0282922b2c458eea7f85...

In other words, no support for png, which this exploit uses :). That doesn't mean that coreboot/u-boot aren't written in C though which is a language known for its vulnerabilities.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by a1o on Wednesday December 06, 2023 @04:43PM (#38549921)
> which is a language known for its vulnerabilities

C is not known for its vulnerabilities, it's known for being fast, portable, and low on abstractions. It's a perfect fine language to write this kind of code. There are techniques like fuzzing and using things like valgrind that can catch invalid input and memory issues quite quickly.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by nine_k on Wednesday December 06, 2023 @07:34PM (#38551394)
Rather, C is know for lack of guard rails, like no bounds checking by default, and some unfortunate early design decisions, like the handling of null, or zero-terminated strings. These are partly mitigated by modern compilers, but can't be done with completely.

I'm anticipating the time when Rust and Zig will displace C as the default language for writing hardware-interfacing code. It's not going to happen tomorrow, though.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by zlg_codes on Wednesday December 06, 2023 @06:10PM (#38550755)
Yeah that stuck out to me, too. The fucking kernel is written in C. 90% or more of the networking stack is written in C. Graphics APIs are written in C. Compilers, databases, firmwares, older video games, everything.

C is not known for vulnerabilities, it simply makes it easy to shoot yourself in the foot. The tooling exists to improve code safety, but generally if you don't do anything crazy, nothing crazy happens.

Lasers, chainsaws and pressure washers are all easy to hurt one's self with but I wouldn't say they're known for creating injuries. It's more like people are stupid enough to hurt themselves with tools regularly.

It's not hard to check a value before dereferencing or freeing it. It's hard to remember to do it.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by yencabulator on Thursday December 07, 2023 @01:01PM (#38559560)
Lasers are so known for creating eye injuries they come with warning stickers.

Chainsaws are so known for accidents they come with warning pamphlets instructing you in how to use them in ways that minimize the risk -- including recommendations for wearing reinforced chaps to protect your legs.

So, following your comparison, it should be perfectly valid to apply a warning sticker to C that says "Warning: Known to state of California to cause data corruption and accidental code execution".

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by zlg_codes on Thursday December 07, 2023 @01:50PM (#38560167)
Except all programming languages can corrupt data. Even knives with sheaths can be misused to harm one's self.

Are we gonna include warnings for SQL injection risk? This is a stupid premise.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by yencabulator on Thursday December 07, 2023 @01:56PM (#38560257)
> Are we gonna include warnings for SQL injection risk?

No, we've moved to calling conventions that avoid SQL injections: bind params. That argument speaks for minimizing use of C, not in its defense.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by zlg_codes on Thursday December 07, 2023 @02:44PM (#38560874)
No, all I'm seeing from you is "it's possible to break in C therefore C bad."

It's okay to admit you don't like a programming language.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by yencabulator on Thursday December 07, 2023 @10:18PM (#38565043)
Constructing SQL by string manipulation is hard to get right, hence we largely stopped doing that. Same story with C, really. It's okay to admit you're irrationally attached to a tool, and calling things stupid because of it.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by pgeorgi on Wednesday December 06, 2023 @02:25PM (#38548341)
coreboot has a rather fragile JPEG parser for splash screens but the saving grace is that barely anybody uses it because that's normally the payload's job (the jpeg + display thing was mostly added for a proof of concept "look how fast we can boot" demo in 2005)

With payloads, some are more at risk than others. edk2 might run into issues here, grub2 has tons of features where stuff like this might hide in, other payloads (e.g. ChromeOS' depthcharge) are more limited in their functionality and therefore in their attack surface, too.

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by anitil on Wednesday December 06, 2023 @07:55PM (#38551553)
My time in firmware has thoroughly broken my brain. All those things that we rely on to live our lives? I don't think people understand The Horrors that live within them
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by seemaze on Wednesday December 06, 2023 @12:46PM (#38547147)
Thank you for that reference. As someone outside of tech, I still found Bryan Cantrill's presentation to thoroughly engaging and entertaining.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by kbenson on Wednesday December 06, 2023 @03:46PM (#38549268)
You'd enjoy most his talks, even if you don't understand the source material. They're half anecdotes and bashing Oracle in a very entertaining way.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by themaninthedark on Wednesday December 06, 2023 @01:49PM (#38547969)
We need UEFI! With out it everything is ancient. And it needs to be updateable from the internet and the OS.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by loxias on Wednesday December 06, 2023 @05:46PM (#38550541)
Yeah, and it should (somehow) involve javascript/wasm and Big YAML.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by forgetfreeman on Wednesday December 06, 2023 @02:54PM (#38548679)
Not sure if sarcasm...
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by baz00 on Wednesday December 06, 2023 @01:53PM (#38548006)
As a former firmware developer I would like to point out that I lacked the competence and interest to have done a good enough job of it past getting paid at the end of the month.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by danparsonson on Wednesday December 06, 2023 @06:29PM (#38550918)
Well that's a relief - I was worried I'd never be good enough to get firmware dev job.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by sonicanatidae on Friday December 08, 2023 @04:53PM (#38575056)
Do you breathe air? Are you capable of making noises, intelligible or not?

Welcome aboard!

I kid...or do I?

Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by pwdisswordfishc on Wednesday December 06, 2023 @01:02PM (#38547368)
I fear your sarcasm may be a bit too subtle.
Re:LogoFAIL: Secure Boot bypass with manipulated boot logos by chx on Wednesday December 06, 2023 @01:23PM (#38547651)
C code? Nah, my friend, firmware is written in SPARK to achieve the highest possible quality. /s

(and yes it should have been written in it, arduous it is, it's a "you had one job" situation)