4 ms·
I know nothing about the IronChat service, but if what is shown in that archived page is the real device then it's pretty obvious that it was doomed to be crack
by squarefoot 8y ago
I know nothing about the IronChat service, but if what is shown in that archived page is the real device then it's pretty obvious that it was doomed to be cracked. A cellphone? Seriously... a freaking cellphone?
Cellphones -all of them- use binary closed blobs to manage device drivers, and to date there is not a single cellphone in the known universe which is free of proprietary closed code. That includes also the Librem5, which is a wonderful step in the right direction, still not completely free of closed blobs, hence not secure.
So what's the problem with (closed) device drivers? Well, they run all the time, they run at maximum privileges (higher than root) and they cannot be audited to spot any malicious code, which makes them the most effective place to hide spyware code. If any government tells a hardware manufacturer to "put our spyware into your driver or your business ends tomorrow" they comply, nobody can spot the code and there's no anti malware software that will detect it.
But why one should care if all text is end-to-end encrypted? Well, on a bugged phone there's no such thing as safe encryption. Let me be more clear: if you tap the text on the virtual keyboard or any device connected to say the USB port or through bluetooth, the text is read by the relevant drivers (higher priority, closed, not auditable) before it reaches the encryption code (lower priority user app) then it can be stored, transmitted (network drivers are closed too) etc. Closed device drivers can be used on most platforms (including PCs) to build a covert channel where information (text, sound, images etc.) travels completely unbeknownst to the user, so a platform can't be considered as secure until every single bit of software and firmware contained can be checked.
So, how did the police decrypt that traffic? I can only speculate that they confiscated one of these devices, then built a bugged driver for some vital devices within it, then got to the manufacturer and forced them to inject that tampered driver as an online update for that given model of phone, possibly installing only if some conditions were verified to be sure it was one of the targets.
If that scenario is half true, then there is not a single piece of computer hardware in the world one can safely assume to be secure. An Arduino-like board, maybe, until the day they'll build faster ones around bigger chips carrying closed blobs inside.
- errantspark 8y agoWhile what you're saying is technically correct, I think you vastly overestimate the resources available and likely to be committed by law enforcement, and the competence of the people making/selling these phones.
- mmirate 8y agoWhen you overestimate a foe, you waste some resources; when you underestimate them, you die or get tortured or etc.
- brokenmachine 8y agoIt's sad that the police are considered a foe. Aren't they meant to be the good guys?
- mmirate 8y agoEver heard of something called an unjust law? Today there are very few of them in the US, and fewer still of them (if any at all) that actually matter. But it's not hard to find other nations where the story is very different.
- dddddaviddddd 8y agoPolice today, maybe organized crime tomorrow? — encryption protects either without prejudice or not at all.
- TheDong 8y agoI don't think this comment is an accurate assessment of what makes or breaks security. To address specific points: > Cellphones -all of them- use binary closed blobs to manage device drivers ... not completely free of closed blobs, hence not secure. Closed source software can be secure. Whether there are binary blobs or not is not really relevant. > So what's the problem with (closed) device drivers? Well, they run all the time, they run at maximum privileges (higher than root) and they cannot be audited ... Modern phones have baseband/AP separation; closed source components are often running more like a peripheral on good phones. For any of those components to exfiltrate the data it would have to somehow get access to the network, persistently store it, or use some other side-channel... Yeah, I'm sure those tiny bluetooth chips can do all that over the limited peripheral interface they use to communicate with the kernel. > So, how did the police decrypt that traffic? I can only speculate that they confiscated one of these devices, then built a bugged driver for some vital devices within it, then got to the manufacturer and forced them to inject that tampered driver as an online update for that given model of phone, possibly installing only if some conditions were verified to be sure it was one of the targets. Absolutely ridiculous. Why would you spend the work to create a malicious driver when you could just update the app code itself if you can push updates to the phones? There's no reason anyone would use malicious drivers when they could use malicious application code; the latter is a darn sight easier to manage. The most likely scenario, however, is that there was a bug in the cryptography that assumed the servers to be trusted or assumed some specific key had authority to mint new keys (e.g. a trusted CA that the police got the private key to). Your post is a rant against closed-source driver blobs, when the reality is they're a difficult to exploit vector, at best. I would like to direct you to tptacek's comment on the Librem5 [0] where he indicates that he, a security professional, believes the iPhone to be the most secure because of the level of auditing and security work they've put into it. > ... there is not a single piece of computer hardware in the world one can safely assume to be secure ... Thanks are not either 'secure or not'. They are secure against something. Security is a continuum. An iPhone (with secure enclave, good disk encryption) is more secure than my laptop, which in turn is probably more secure than the average wordpress server. I fully believe that my iPhone will withstand even motivated attackers with physical access. I don't think my laptop will. I don't know how it would fare against a nation-state specifically targetting me, but I don't really have to worry about that. [0]: https://news.ycombinator.com/item?id=17913148 https://news.ycombinator.com/item?id=17913148