Live data from Hacker News

Bricked iPhone 16 Can Be Restored Wirelessly Using Another iPhone

macrumors.com

111–120 of 127 posts

Re: Bricked iPhone 16 Can Be Restored Wirelessly Using Another iPhone

#112

I still don't understand why Apple haven't changed their update model to A/B partitions, so if an update can't boot, it boots the previous partition. Android has been doing this for a while and Fedora Silverblue/IoT does the same thing, both of which I've found really useful in the past. If communicated well, it would elviate the update anxiety many people I know have with their iPhones.

How often does a deployment rollback go smoothly? Almost never. The priority is roll-forward and reduced error scopes.

Re: Bricked iPhone 16 Can Be Restored Wirelessly Using Another iPhone

#114
post #70

Earlier quoted context omitted.

Why have three Ctrl keys? Why not use an empty modifier and have your own shortcut keys that you can guarantee no other program uses?

> Why have three Ctrl keys? You could just swap caps with one. > Why not use an empty modifier and have your own shortcut keys that you can guarantee no other program uses? Because that would require messy tinkering with multiple layers of software.

From memory, Linux has spare modifiers. You can see them using xmodmap. It's easy to assign shortcuts using them.

Re: Bricked iPhone 16 Can Be Restored Wirelessly Using Another iPhone

#115
post #103

Earlier quoted context omitted.

> Does anyone offer a tethering plan that's rate limited but not data limited? T-Mobile in the US; they give you a set amount of high-speed tethering based on your plan, then it rate limits severely until the end of the billing cycle unless you upgrade your plan or buy a a pack of data.

Okay, 600kbps is acceptable. Beats a lot of networks. Things are improving since I last checked. Though with how happily and loudly they've made 1.5Mbps video excluded from data caps, it would be better if the throttled speed was at least 1.5Mbps.

Verizon after 60GB is 600kbps on 5G/4G but up to 3Mbps on UWB. Pretty reasonable IMO. It’s AT&T that’s the holdout, 128k regardless of plan, after 60GB.

Re: Bricked iPhone 16 Can Be Restored Wirelessly Using Another iPhone

#117
post #100

I still don't understand why Apple haven't changed their update model to A/B partitions, so if an update can't boot, it boots the previous partition. Android has been doing this for a while and Fedora Silverblue/IoT does the same thing, both of which I've found really useful in the past. If communicated well, it would elviate the update anxiety many people I know have with their iPhones.

It will not be able to reliably work with the A partition once B boots the first time - the update process even for many minor releases will trigger local and cloud data upgrades, and there are security reasons to not leave the option to leave older/unpatched versions runnable. I don't think I've had a software update fail since iOS 7 beta 1 - and that was an issue with updating local data on first boot, so any attem…

I believe the point isn't to let you rollback at will, but it'll automatically rollback if it fails a boot check, so before anything is written to the filesystem. However, it's worth noting that Fedora Silverblue/IoT let's you rollback at will without problem.

Re: Bricked iPhone 16 Can Be Restored Wirelessly Using Another iPhone

#118
post #50

Earlier quoted context omitted.

It isn’t for most people here. But the most valuable tool for fixing a broken computer is a working one, and a lot of people don’t have more than one anymore. Some don’t even have that. My wife texted me on Friday as she was visiting her mother six hours’ drive away. She needed her laptop. They do not own a desktop, and her dad was at our house on a business trip with his laptop in tow. Do not underestimate the utili…

You would be amazed how many normal people dont have a computer - they have an iPad and an iPhone (or android similars), and yeah - its like that now.

You didn't have to be the owner of the computer.

Previously, you just needed temporary access to a computer with an internet connection.

Re: Bricked iPhone 16 Can Be Restored Wirelessly Using Another iPhone

#119
post #92
post #83

Earlier quoted context omitted.

Well, IF you have an exploit capable of bypassing all iPhone security you could now use it by setting something on someone’s phone for a few minutes instead of plugging into it. Is that useful? Maybe. No fingerprints? Of course you still need that billion dollar exploit first.

You can't bypass "all iPhone security" - this is a restore path: the phone is already in a state where it cannot boot, and cannot access user data. Once an iPhone is in a restore state (restore mode or DFU which afaict are only marginally different) it cannot do anything except install a new OS. If an iPhone is in a restore mode it is not in an "update the device mode" it's in a "the OS is dead and the user data is g…

Right. I was saying hypothetically if you had a big enough chain of exploits for all the various protections Apple has then this may let you use them without physically touching the device, only being really close.

But if you had that chain of exploits, who cares. You already won the game.

So I don’t think this matters to security.

Re: Bricked iPhone 16 Can Be Restored Wirelessly Using Another iPhone

#120
post #106
post #13

Earlier quoted context omitted.

I would argue the opposite: AirPods are a pure "client" device — they need to be connected to by something, they don't connect to something; and they have no display, input method, or other means to initiate/change Bluetooth pairing, if the state of the firmware somehow gets in a mucked-up state and they need to be flashed. The only possible way to recover broken AirPods is via tethered recovery. An iPhone is the opp…

> if the state of the firmware somehow gets in a mucked-up state and they need to be flashed. The only possible way to recover broken AirPods is via tethered recovery. There is no process to reflash AirPods, at least one known outside of Apple's official repair channel. But there's also no way you can diagnose you _need_ such a software recovery mechanism. USB-C might help if there was a feature to trigger a firmware…

> Any diagnostic and recovery tools available at boot time are going to be designed to be non-privileged, so there are a number of things (such as validating the integrity of the encrypted user data partition) that just aren't going to be possible. That locks you down to basically rudimentary hardware validation.

Eh, sort of. Remember that most users have an iCloud Recovery Key enabled. Meaning that Apple could design a non-tethered restore process for iOS that looks like:

1. The device boots into recovery mode; asks the user to configure a wi-fi connection (if none of the ones it has persisted in NVRAM are available.)

2. The device reaches out to Apple, hitting an "I'm a broken device; help me!" endpoint, referencing the iCloud account the device is signed into, and signing the request with the device's root TPM key.

3. Apple validates that the message is from a device that is registered to the given user's iCloud account.

4. Apple sends back a signed payload containing: a device-root-TPM-key-encrypted copy of the device's own disk encryption key (the iCloud Recovery Key), and some amount of "privileged restore" logic (at minimum, just an XPC daemon for the restore image to load and run as root, with a code-signed capabilities manifest that grants the device-shipped client-side of that XPC connection the authority to make changes through it; or, at maximum, a whole second-stage recovery firmware for the restore image to reboot into.)

5. The resulting boot would have the same privileges to modify the on-disk system that an OS firmware update does. But, unlike an OS update, the boot that results by combining this Apple-delivered privileged payload, with the existing restore image, would be interactive. Just like macOS's Remote Recovery image, it would give the user access to some of the same management functionality that Apple would use to repair the filesystem, forensically archive + wipe the disk, repair user permission databases, etc.

6. However, this privileged recovery experience, would still ultimately be a DRMed kiosk experience. (The tool would have full power to access or modify your on-disk system, but the tool wouldn't give you any way to modify the running tool itself. Within the OS of the recovery experience, you're still an unprivileged user. Like running VM software on ChromeOS, where you have root in the VM but not on the host.) Especially, since the disk is decrypted enough to get at the keychain, regular authentication methods could be used to enforce that the device's owner is present and consenting to the repair — i.e. the device could require a "sign-in" with iCloud password + TouchID/FaceID/etc. — just like when first making a purchase on the App Store on a new device. (If the disk is so corrupted that this info is unavailable, then the device could fall back to "authorize with another device on your iCloud account" auth; and if that too wouldn't work, then it'd probably tell you to stop what you're doing and go to an Apple Store so they can identity-verify you.)

7. At the same time, Apple could add special handling into the Secure Enclave for a sort of PKI latch storage type: "anyone with the pubkey can set latch k to a value v; but only a message signed by the associated privkey can modify/delete the latch k once set." Apple could then make this privileged recovery experience do two things once booted:

- Tell the Apple servers "consider me de-activated for now; and require an extended activation — with full remote attestation — to reactivate!" Boot would not proceed until the Apple servers confirm the de-activation.

- Set an on-device TPM latch, using an Apple-provided pubkey.

(These two actions can be atomic, as a "start non-tethered repair session" endpoint, that both marks the device as de-activated, and returns the pubkey for a new dynamically-generated keypair associated with that repair session in Apple's backend.)

8. The latch means that the regular on-disk OS won't be allowed to boot again until this recovery process "considers itself complete" — so you can't break out of it mid-job by hard-resetting the device. And the remote de-activation means that, if you were trying to use non-tethered repair on "some random grey-market iPhone" to wipe it and/or scrap it for parts, the device and/or parts would still be considered stolen.

9. And "extended activation with full remote attestation" means that, when you do finish the recovery process, you wouldn't boot into the regular on-disk OS, but into a special "remote attester" mode (essentially, another type of small-ish factory-shipped recovery volume, that the repair system either has no right to modify, or always puts back after a reformat) that carefully examines and validates the OS to ensure everything is as it should be to enforce security (no rootkits, no jailbreaks) — sending that info to Apple, who then send back activation only if they're happy with what they see. (And if you somehow trick the device into booting into the on-disk OS instead of the remote attester, the on-disk OS will find that it can't activate with Apple; will notice the latch; and will just reboot back to the remote attester to try extended activation.)

Under this tight set of constraints, you can do whatever you need to do during recovery — you have root on the phone! But you're also forced to put everything back into a valid state, before Apple will be willing to "release" you from that mode back into regular use of the phone.

(Amusingly enough, Apple could even let this repair image totally break the "paradigm" that iOS normally operates under. The repair image could expose Terminal.app. The repair image could display a desktop when connected to HDMI. Heck, for the iPhone 16, the repair image could be a somewhat-neutered copy of macOS [similar to macOS Recovery] running on the M-series processor, where you're expected to connect a keyboard/mouse/display to it using a USB-C dock! As long as that sort of stuff is only available in this repair experience — and this repair experience is in turn useless as a phone [or as a general-purpose computer] — then exposing these capabilities during repair wouldn't be sales-cannibalizing.)

Also, if you think about it, this is in spirit basically the same "workflow" as Apple's Self Service [Hardware] Repair process — but without the series of phone calls being required to validate that you're actually someone who was asked to repair that device (instead of some rando who bought the phone on the grey market) and to sign off on your repair "session" as completed at the end.

In this case, you're repairing your own device that Apple knows you own because it's listed on your iCloud account, so Apple already has all the info they need to auto-authorize the "repair"; and you're purely doing a firmware/OS repair, not a hardware repair, so the device itself can do the "signing off" at the end.

Post reply on HN