[chatmasta]

I would be very concerned with someone pwning my phone with a $5 wireless device, but that's not what's happening here. This is a DoS attack. It could never be perfectly mitigated, as long as any mitigation depends on the (arguably) fundamentally impossible task of verifying an external device is a "real" Apple device. It's possible to design security protocols that allow me to verify my device is a real Apple device, but likely not to verify a packet was sent by someone else's "real" Apple device. So I'm less concerned about the latter.

Also note that I specified I'm more concerned with verification of trusted hardware on my own device. Because the repercussions of malicious hardware implanted in my own device cannot be mitigated purely in software. Whereas verifying the integrity of an external device inherently depends solely on software, since there is no hardware interaction. I'm still concerned about it, in the sense that I'd like my OS to take best efforts to only "trust" external devices insofar as it can verify they're trustable, but I also accept that those devices are outside of my control and so any protocol for trusting them will have holes in it. My main requirement is that I should be able to opt out of the system if possible (by e.g. disabling bluetooth).

[AshamedCaptain]

> This is a DoS attack. It could never be perfectly mitigated

Bullshit.

Flooding the waves with radio interference (something that Bluetooth is particularly resistant to) would at most "deny service" of another device trying to connect to my iPhone through Bluetooth. It should NOT deny service of the _entire_ iPhone, which is what is discussed here. This is 100% preventable crap.

> the (arguably) fundamentally impossible task of verifying an external device is a "real" Apple device

Bullshit... and egregious considering you apparently think it is doable for replacement parts, but "fundamentally impossible" for networking devices. SSL is about 30 years old by now.

> I'm still concerned about it, in the sense that I'd like my OS to take best efforts to only "trust" external devices insofar as it can verify they're trustable, but I also accept that those devices are outside of my control and so any protocol for trusting them will have holes in it.

Also bullshit. All these holes are because of the proprietary extensions Apple puts on top of Bluetooth, which are exploited to no end. Notice my original post is about Apple not being able to identify when it is a (real vs fake) Apple device that is trying to initiate a connection. The protocol is 100% controlled by Apple.

Normal Bluetooth protocols and devices (which do not identify as Apple devices and are therefore subject to the standard Bluetooth pairing UI) are almost never the problem.

[chatmasta]

I agree with you, but that doesn't negate what I said about this being correctable in software. Whereas if someone implants a malicious HSM in my iPhone, or a screen that has a secondary chip connected to it recording everything I touch, then that's not correctable in software.

It also does not qualify as "pwning" your device, at least for my interpretation of the word "pwn."

[lxgr]

> that doesn't negate what I said about this being correctable in software

“My house is on fire, but that is easily correctable by the fire department using water, a cheap and widely available commodity. The real concern is alien abductions in my neighborhood. We are defenseless against these!”

> It also does not qualify as "pwning" your device, at least for my interpretation of the word "pwn."

Random people on the same train as me being able to crash my phone fits my definition of “pwned”. And so does me having to use wired headphones as a countermeasure.

[AshamedCaptain]

Your threat model is still ridiculously upside down. You are literally arguing that you are more worried about the possibility that someone subjects you to some type of maid attack (which requires an almost implausible level of dedication) rather than someone with a 5$ atmel chip claiming to be an Apple TV, automatically pairing with your device, and afterwards doing god knows what with it (including leaking more data than _anything you could do_ with a even country-agent level trojanized replacement screen). All from the comfort of their car and with so little targeting they could practically wardrive with it.

[chatmasta]

I'm assuming I can opt-into the threat, i.e. it's possible for me to disable Bluetooth to remove my exposure to this class of attacks. When I turn on my WiFi I know that I'm subjecting myself to de-auth attacks, for example.

I can't opt out of a hardware attack once a malicious repair shop has replaced a critical module in my phone with their own.

Like I said, I'm more concerned with the latter. It doesn't mean I'm not concerned about attacks from external devices too.

[AshamedCaptain]

> can't opt out of a hardware attack once a malicious repair shop

So apparently you forever disable Bluetooth out of concern but at the same time think it is unavoidable to leave your iPhone unattended at random repair shops? At least the maid stuff (even if astronaut-level engineering) is remotely plausible.

[lxgr]

Since when can de-auth attacks crash devices? That is what’s happening here!