[walterbell]

If an iOS app did not have "Background App Refresh" permission, could it still have exploited this vulnerability?

Can physical microphones be removed from Apple devices by a repair shop, while still allowing use of wired/wireless headsets?

We need Purism-style hardware kill switches for microphones, cameras and radios.

[SamuelAdams]

> Can physical microphones be removed from Apple devices by a repair shop, while still allowing use of wired/wireless headsets?

Yes, this is what I do. The mike is actually still in the laptop but it's disconnected from the motherboard. On a 2021 M1 Macbook pro all you need to do is pop off the back cover and disconnect one cable on the right side of the motherboard. All in all takes about 10 minutes of work.

[hoorible]

There actually is a physical microphone disconnect for new Mac laptops (~2019 and later). When the clamshell is closed, the mic’s connection to the MLB is physically severed.

I actually just learned this exists on new iPad models too, with any MFi-compliant case!

I know this isn’t strictly relevant, since the vulnerability discussed here is during active use, just thought you might find it interesting.

https://support.apple.com/en-ca/guide/security/secbbd20b00b/...

[judge2020]

My only problem is knowing whether it’s on or off - IIRC the indicator next to the camera is indeed a dumb LED wired right into the camera, but the microphone doesn’t have a hardware indicator.

[fsflover]

It's not a physical disconnect. More like firmware disconnect. Purism have a physical kill switch for the microphone in their laptop.

[hoorible]

They pretty clearly state the disconnect is “implemented in hardware alone” about three times in the support article, how would that not be a physical implementation?

[fsflover]

What are the physically moving parts in the iPad? How is the T2 connected to the laptop lid, where does the contact breaks? It's pure corporate doublespeak. Especially without schematics, unlike with Purism.

[reubenmorais]

A disconnect doesn't have to move to be implemented in hardware. They describe the implementation in the article linked above:

> In each product with a hardware microphone cutoff, one or more lid sensors detect the physical closure of the lid or case using some physical property (for example, a Hall effect sensor or a hinge angle sensor) of the interaction. For sensors where calibration is necessary, parameters are set during production of the device and the calibration process includes a nonreversible hardware lock out of any subsequent changes to sensitive parameters on the sensor. These sensors emit a direct hardware signal that goes through a simple set of nonreprogrammable hardware logic. This logic provides debounce, hysteresis, and/or a delay of up to 500 ms before disabling the microphone. Depending on the product, this signal can be implemented either by disabling the lines transporting data between the microphone and the System on Chip (SoC) or by disabling one of the input lines to the microphone module that’s allowing it to be active—for example, the clock line or a similar effective control.

[9935c101ab17a66]

This is a strange way to reason about electronics. A processor is a “physically moving part.”

Hardware !== clunky obvious tactile contraptions

Anyway, I actually have an answer for you, at least for apple portable computers — most new macs come with a sophisticated lid angle sensor used to detect the display angle with high precision. (Previous models used more conventional Hall effect sensors which didn’t live to apples standards for a hardware cut off

Why is it importable to know the precise angle of the display in relation to the top case of the computer? Because you can detect when it’s closed with a high degree of confidence!

> How is the T2 connected to the laptop lid, where does the contact breaks?

So, your comment clearly indicates that you have very little experience with apple products and industrial design and engineer. You also made no effort to look into the matter see if you could find an answer yourself (it took me 4 mins of googling to connect the new LAS to the mic cutoff when computer is closed.

So you made a spurious allegation that apple was lying, when pushed on it, you followed up with an even more nonsensical comment that made your lack of hardware experience more obvious.

And after all of that, rather than having the self-awareness to drop the matter, or at very least, do some of your own research, you instead decide to sign off by accusing apple of misleading their users.

Super weird.

[javajosh]

It's a good idea, but I can imagine how frustrating it would be if someone called and I didn't have my headset. The EV of avoiding that experience seems slightly higher than the EV of avoiding risk of being eavesdropped on by a wayward smartphone process by disabling the internal mic.

[f1refly]

> but I can imagine how frustrating it would be if someone called and I didn't have my headset.

"Sorry, I don't have a headset right now, we'll have to talk later". It's not that "frustrating", really.

[shivsta]

“Why aren’t you able to dial in with just your computer?”

“Oh, you know, I’m a bit paranoid about my microphone being hacked so I disconnected the microphone internally. If you give me 10 minutes, I can put it back together real quick”

Not the vibe I personally wanna be giving friends and colleagues.

[walterbell]

Some people use headsets to cancel environmental noise, improving the listening experience of friends and colleagues.

[f1refly]

If you are too weak to admit this is important to you, just go the "it's broken" route.

[walterbell]

Indeed, many audio/video calls are preceded by text communication.

[walterbell]

Excellent, thanks for the field report.

[cosentiyes]

> We need Purism-style hardware kill switches for microphones, cameras and radios.

And accelerometers and ...

[fsflover]

https://puri.sm/posts/lockdown-mode-on-the-librem-5-beyond-h...

[MBCook]

Note this Bluetooth only.

[walterbell]

Yes, the question is how to permanently restrict the attack surface / time windows for audio and video surveillance attacks.

[ASalazarMX]

Instead of Bluetooth defaulting to on, and re-enabling itself next day if you turn it off from the control center, I'd like for Bluetooth to default to off. You'd have to enable it from the control center, and it would disable itself after a certain period of inactivity.

I suppose that won't happen, as it would wreck the Find My network if it depends solely on Bluetooth.

[varenc]

You can create a Shortcut (in the Shortcuts app) that actually turns off Bluetooth completely. Then you can add automation to run your "turn bluetooth off" shortcut multiple times a day. Haven't look into it but you might be able to create another shortcut that turns Bluetooth on, but then sets a background timer for X minutes after which it'll then turn Bluetooth off again.

Not that elegant of course, but sort of makes it possible.

[byteduck]

When you turn off bluetooth from CC, it’s not even turning it off. The radio is still on - it just doesn’t make any new connections. You have to turn it off in preferences for that.

[sroussey]

Right. If you want it off, use Settings. Then it stays off.

[Gigachad]

It's more than find my. Bluetooth just being on is core to the just works experience for airpods, apple watch, and a bunch of other smaller features. And the reasons for turning it off are vanishingly small for the average person.

For security this is probably something that could be brought in to lockdown mode for people who want absolute security over convenience.

[walterbell]

That would be a good safety-first default. If Control Center could have buttons linked to iOS Automations for radio state, then advanced users could control this behavior with custom scripts.

> wreck the Find Me network if it depends solely on Bluetooth

Find Me presumably uses all identifiable radios, including BT, UWB, Wi-Fi.

[dontbenebby]

It's not really a question, hardware switches work and companies refuse to put them in so they can... shrink the profile of devices in ways that rely on rare earth minerals to an unsustainable degree when combined with the typical replacement rate.

[walterbell]

Hopefully legislated right-to-repair can open the door to aftermarket mods, including phone body with new switches that can electrically disconnect specific sensors.

[dontbenebby]

Ehhhh... is right to repair the right phrasing?

I worry about requiring switches in the same way one can require a universal standard for power delivery. (The EU did that recently... good move IMO, though I can understand the delay since discussions about amperages and whatnot do take time.[0])

Maybe requiring anyone who wants to contract with the US government to offer such a model, and that said model be available for consumer purchase as well, would be a simple solution.

They sometimes won't let say, Russia, buy the same stuff as say... Canada... but that's usually stuff like night vision goggles. The exact same phone or laptop, just slightly larger with more switches shouldn't have any... I think the word is "export controls"?

Please keep in mind, I am not a lawyer, and I'm very stupid -- I only have a master's degree -- so sometimes the things I say are wrong... please only credit me for the times I'm right. Thx!

I'm off to do more drugs now... have a nice Thursday!!

- Greg from Pennsylvania

[0] https://www.npr.org/2022/10/07/1127543116/eu-mandate-for-a-s...

[walterbell]

"Right to Repair" is the terminology that has gained the most legal traction (e.g. some narrowly-scoped legislation) in the US and EU. https://www.repair.org/

[LaputanMachine]

>Can physical microphones be removed from Apple devices by a repair shop, while still allowing use of wired/wireless headsets?

There have been reports that the 2020 iPhone SE cannot be used without a microphone:

https://repair.wiki/w/IPhone_SE_(2020)