No new iPhone? No secure iOS: Looking at an unfixed iOS vulnerability
joshua.hu
joshua.hu
To answer three questions:
1) this was not reported in the context of any bug bounty[0], and the total conversation between me and Apple is 4 emails (1: hello do you plan to fix this? 2: can you reproduce this on the newest ios17? 3: no. 4: if you are able to reproduce it on ios17 let us know)
2) exfiltration is obviously possible, I’m not sure why I would even need to specify that any page is able to read its own contents using JavaScript
3) the iPhone 6s and 15.8 are still seemingly supported by Apple.
[0] and you won’t find me on any bug bounty websites except for where I try to get contact with humans, see https://joshua.hu/slack-is-broken-with-noscript
It's sad that one has to open a bug bounty request to get an engineer to look into an issue like the one you described in your linked blog post.
How much time would you estimate goes into researching? And do you have any pointers for someone which want to dip their toes into this vast sea of exploration?
>And do you have any pointers for someone which want to dip their toes into this vast sea of exploration?
Can you expand on what you mean by this exactly? I got into security by accident when I was a child and have never followed any type of routine when it comes to learning.
Probably the only pointer that may help is: take something that you know and use often (some piece of software) and ask "what if?" or "why can't I or someone do X?" ¯\_(ツ)_/¯
Things that are obvious to you may be non-obvious to other people, including the readers of your blog.
It's not reasonable to expect the author to explain _every_ underlying technology involved since that was likely not the scope of the post.
> Apple continues to support nearly-decade-old devices like the iPhone 6S, which and iOS 15.8 is still powering those devices, seemingly with official Apple support, with the latest update from October the 25th, which addressed some security vulnerabilities.
With: try to update your Apple Watch with a device running the latest iOS 16
This photo was taken the day the blog post was published.
So, to the author's assertion that "Apple does a good job supporting old hardware", they may give some security patches but interoperability between current and previous versions is seemingly deliberately broken.
The list is widely published: https://support.apple.com/guide/iphone/models-compatible-wit...
Any model more than 5 years old (Xr and Xs) are essentially not being updated and not secure. So an iPhone 1, 3, 5, 6, 7, 8 and X are all not secure and most people who use the iPhone are totally aware of this.
It's like writing an article that Windows 7 is insecure and Microsoft isn't patching it. This is essentially their policy in most cases.
That reads very much like X's are not supported. The 2nd sentence even says it explicitly.
So at a glance one could assume that the XS is lumped in with the X just like the 6S is lumped in with the 6.
That is some Windows 98 nonsense right there.
Doesn't sound better today does it?
At least on Windows they didn't prevent you from installing a different browser.
Windows still ships Internet Explorer for rendering some old components, although I think they are moving to Edge (chromium) based web view. I can't find a better source right now, but something to start with: https://www.reddit.com/r/Windows11/comments/11n79xc/why_does...
Android has a chromium based WebView that ships with the system and is updated via Play Store: https://www.techtarget.com/searchsecurity/definition/Android...
My 10 year old android phone still runs the latest versions of both Chrome and Firefox.
Androids WebView being on the play store even allows that part of the system to be updated independently of the OS.
So why should I assume that latest iOS 16 isn't completely patched? I think it's a shame to say at least that Apple has no public policy of how which OS versions are supported and which are not, it's just guesswork. Whereas I definitely know how long Microsoft supports Windows versions, e.g. Windows 10 until October 14th, 2025: https://learn.microsoft.com/en-us/lifecycle/products/windows...
I think if something is a relatively easy fix and high severity that they will fix it. I don't think they view security updates as a tool to force people to buy new products. The low hanging fruit for large numbers of users gets fixed. The underlying software however, should not be trusted or viewed as secure.
Even though these applications are bundled with the operating system, they are probably separate code bases and if they believe the patch can be accomplished across the versions with minimal work like fixing the same line of code in the old version it probably goes out. If they have to do a major overhall of the old operating system and port the new browser version to the old software, it probably doesn't.
I wish the author included a full proof of concept
Also not sure I agree with the implication that Apple shouldn’t publish which vulnerabilities they’ve patched (the only logical conclusion because the alternative, patching every version in perpetuity, is unrealistic).
I do wish legislators forced Apple and Google to give users a path to install an alternative OS on their device. That would enable old iPhones(and Androids) to have their lifetime extended further.
All Google phones allow installation of alternative OSes (and are pretty much the only phones that allow resigning the boot loader so they're still secure - which is why they're chosen by projects like GrapheneOS). Why do you think they need to be forced into anything? You can buy a Pixel right now and run an alternative OS.
I think this is also the answer to "the app store question". It feels unfair to force Apple to change a core value proposition by forcing them to allow side loading, but in the spirit of "It's my device" it should be possible to install an OS that makes a different tradeoff e.g. Android
* If you don't use a super popular model, you would have very few choices for custom ROMs * Often custom ROMs are still at the mercy of original manufacturer for certain hardware support -- e.g. they need to release "base AOSP image" or something like that * Custom ROMs often had random things not working, e.g. NFC not functional or cell/WiFi signal weaker than stock ROM * Some come with certain crappy preinstalled apps. Sure you can remove them, but still annoying * Battling SafetyNet was a cat-and-mouse game which I gave up * Browsing xda-developers forum and following the latest reply of a 10-page, 200-post thread like back in the early 2000s was the only way to get updates of a certain ROM. OTA updates were mostly out of the question * ... and many more added to this
I don't know how much has changed, hopefully a lot. I don't doubt if you have a phone of a popular model, you can find a custom ROM that does not make compromises and is much cleaner and better than the stock ROM. But these days I simply don't have any time for tasking the risk and messing with these things and worry WiFi might not work in some cases.
What's stopping you from keeping your Android 6 years making it an event better value? Most people I know don't throw away their Androids after 3 years but keep them as long as iPhones. Basically until it breaks/dies. So far I don't know anyone who got hacked and suffered damages for using an Android that stopped getting updates.
Obsession with zero day patches and security hypochondria is mostly a HN/tech-workers thing as if they're under attack by state actors using Pegasus. The average joe doesn't care not is he very likely to be impacted, unless using Android Jellybean or something, since most malware in the wild out to get them is script kiddie level not state-actor level.
And anyway, someone correct me if I'm wrong, but from my knowledge and experience with Android security, most Android critical vulns that can impact the day to day security of the Average Joe are covered by updating Google Play services and Chrome or whatever browser you use as those are the main attack surfaces for (non state sponsored) malware based on what people do with their phones, and those services keep getting updates from Google long after the manufacturers stopped pushing OS updates.
That’s a “hope for the best” approach to security, for me it’d be irresponsible to recommend it to even friends and family.
But if you want to do it with knowledge of the potential problems — go for it. It’s your phone and your data.
Here's a though exercise: Most people use their device for browsing the web and messaging people, right? So as long as you keep your Play Services, browser and messenger apps up to date, how will malware get to the outdated layers of your OS to PWN you? Especially that modern web browsers and Andorid use sandboxing for apps and web tabs. I'm talking about realistic documented scenarios from the wild that have happened and are likely to repeat again, not state actors or scenarios from research labs where they hack you through the firmware vulns of the baseband modem.
Like I said, I'm not recommending you still use unpatched devices, but the realistic risk from using an Andorid device that stopped getting updates a year or so is relatively minimal in practice, otherwise there would be mass hacks and credential thefts left and right on a daily basis considering how many unpatched Android phones are out there.
[0] https://www.lineageos.org/ [1] https://grapheneos.org/ [2] https://calyxos.org/ [3] https://download.lineageos.org/devices/sailfish/changes
People who’re comfortable (or can be bothered) installing alternative OSes on their phones have an entirely different view on device obsolescence. Statistically they’re also a rounding error in the total mobile-using population.
Understood, but that has no bearing on the point being argued. You have no control over Apple IOS hardware after Apple stops supporting it. The fact that there is that "rounding error" is good for everyone as it is a force against closing that ecosystem which currently exists. It matters.
I’d be very curious to see HOW they contacted Apple. Depending on if you’re reaching out to security or just filing a standard radar I’d expect a very different answer.
Also, was it reported to the WebKit team? If that is where the bug is, perhaps that’s who should be taking the report?
[^1] It would be completely reasonable to say "Earth-shaking? Really? You expect security backports for a decade?" I've been in mobile my whole career, iOS for 7 years, starting from jailbreaking the original iPhone, then worked on Android itself for 7 years. I am sure significant decisions were made assuming this was the case.
Google no longer offers security patches directly, but since you control the phone sufficiently to install your own OS, the community can come together and keep security updates flowing. You could do it yourself if you wanted.
Apple devices make this sort of community maintainership effectively impossible.
I know this means practically nothing since only nerds can actually install a third-party ROM, so for the general populace only the "default" security patch window matters, but for the hacker news crowd it seems like it might be a meaningful difference.
Can we just choose to stop suggesting it as a legitimate solution cause outside of this bubble, it absolutely is not.
The bigger problem is that a huge bunch of software running on the phone is fully proprietary and closed source, and there are many many different versions for different phones around - making it virtually impossible to do any meaningful reverse engineering. So sure, your main OS may be up to date, but the baseband OS and virtually all of the device drivers will be left vulnerable, and they have just as much if not more access to the data on your device.
Come on.
You don't have to agree but resell value of older iPhone being much-much higher than Android tells customer values the support and quality of iPhone.
Pixel 8: released in 2023, updates through 2030 Pixel 5: released in 2020, stopped getting updates in October 2023.
3 years worth of updates is pretty shit… my son’s iPhone 5c got updates for over 5 years (and I think there were some security issues they patched after that even)
At the moment I’ve got a perfectly usable Pixel 4a that I’m going to have to replace as it’s not secure enough for work related stuff anymore
Don’t buy brand new old phones new from Apple, they’re a ripoff. If you buy either an iPhone 12 or 13 used for $250-350 you can basically plan on a $50 a year budget to have a smartphone that always has the latest OS judging by their expected remaining lifespans.
I think the big flaw with the status quo is e-waste more than cost to the consumer. I think an iPhone 6S or 7 are incredibly slow and outdated devices for today’s usage but in 5 years I don’t think we will be able to say the same thing about an iPhone 12 or 13. Smartphone hardware is far more mature now than it was even 6 generations deep into the iPhone product line.
We should be able to replace batteries for $20 and replace things like broken screens for not much more, and Apple should be enthusiastic about it considering how services are their bread and butter moving forward. Apple should be happy to produce fewer phones and keep more consumer dollars allocated toward the purchase of high margin digital goods.
The wildcard here is local LLM use cases and any new hardware that increases their speed by orders of magnitude.
I also don’t see any indication that there will be impactful local LLM silicon at the smartphone scale anytime soon.
The 2020 SE is available from a wide variety of sources for 200USD (still new in box); it'll be supported until 2027. The 2022 SE is 400USD, supported until 2029.
By comparison the Android phones at this price point functionally went out of support 2 years before they even existed- not only is there zero support for them, but they ship with outdated OS versions to begin with. And no, "but I can go to XDA and get a shitty ROM at the cost of my camera" doesn't count as support.
This isn't a debate about what company is better. The word "now" is used for Google's promises for a reason.
Let’s talk again in 5 years, once they had the opportunity to prove their plans. So far, it’s all just talk.
Over a decade of Nexus then Pixel devices being flashable has not moved any needle of Apple doing the same. Google promising 7 years is in line with Apple's 10 year track record of providing 6-8 years of updates, so it's more like Google aligning with Apple, not Google pushing Apple.
Still, a vague† promise in a blog post or keynote address is not going to fit the bill, at the very least it should be in the EULA or other contractually enforceable document, otherwise the promise is worth nothing.
Ideally I wish software would be treated as with e.g automotive or washing machine manufacturers, who in the EU have a legal requirement to provide parts for 10 years.
† I mean the promise is clearly worded but bears no weight, especially when pitted against Google's track record over the last decade of making grand announcements then puling the rug down the road.
Yes, but that is only one component of a modern phone. Basebands and system bootloaders, among other firmwares, don't receive updates. Those are regularly attacked.
It's good that they do but it's not enough.
"not everyone" is an understatement.
That's a solution for you (and the dozens - dozens! - of people doing the same), in practice it is not for 99% of Android users, therefore, again in practice, there's a huge fleet of devices with out-of-date software out there.
> But Android also lets you run custom builds
That's not even counting that:
- many Android manufacturers make it non-trivial† to root/unlock/flash a build and/or make it blow a warranty fuse, and that's if it's even possible at all.
- usually the camera goes ape shit, and often loudspeaker audio quality too.
- unless you relock the bootloader it immediately compromises security and makes bootloader updates nontrivial as unlocking again clears the device.
Mind you, this is a fine, intellectually satisfying strategy for you and me to be able to flash open builds, but it's by and large an extremely fringe strategy, and it's been shown over a decade that it's staying that way.
† Often involving downloading random flashing tools from obscure forums, that run only on Windows, some of these being one shot and requiring to plug in magic numbers corresponding to your exact device, and if you screw it up the device is bricked (e.g Samsung). Or the unlocking is on a low-write-count chip and once you exceed that limit the device is bricked (e.g OnePlus). I know, I've been there, bricked a few, recovered only one through JTAG.
And how many are there, then?
Google: promises
you're being disingenuous
Google is not promising this out of the goodness of their heart. They're just getting ahead of what the EU is planning to mandate [0], and doing that to get some good marketing while they're at it.
So, while Google's track record leaves a lot to be desired, in this, I think they'll keep their promise, either because they actually care, or because the EU will force them to. Either way, we, the end users, will benefit from it.
And this will apply to all electronic device makers. That's probably why Samsung also increased their updates policy to five years as well.
[0]: https://www.insideprivacy.com/cybersecurity-2/eu-publishes-d...
> "The requirements apply for the lifetime of a product or five years from its placement on the market, whichever is shorter."
If that was the case then why did Google exceed the requirement by 2 years? Additionally, Google is providing 7 years of OS upgrades and 7 years of security updates. Google could have easily just do what they did with the Pixel 7 and offer 3 years of OS upgrades and 5 years of security updates, thus, meeting these EU requirement of 5 years of updates. So to claim that Google offering an industry leading 7 years of OS upgrades and 7 years of security updates is not out of the "goodness of their heart" is being disingenuous IMO.
Do you really think the cost of the class action lawsuit and settlement and the bad publicity for not adhering to their 7 years of support would not exponentially exceed the cost of a team of engineers tasked with supporting updates for their Pixel phones?
As for "promises" - why hasn't any other OEM matched or exceeded Google? Apple should have been the first one to step up the very next day.
I’ll believe it at 6 years in, maybe.
While Google promises, Apple actually has a decade long track record of updating older phones for 5 or more years. We don’t know if Google will actually follow through on their promises or the execs in charge in 5 years will feel differently. But I personally bet $1000 that the iPhone 13 will get 5 years of OS updates minimum.
It means they’re shoehorning another issue into this discussion.
> seems very unethical
I agree. Clearly Apple should do the only ethical thing and immediately ban all unsupported devices from new App Store purchases.When Apple announces this (very ethical) policy change, I expect you to full-throatedly defend Apple from the mass shrieking cries of "but muh planned obsolescence!!" Right? Right?? ;-)
In fact, one could argue it's also "very unethical" for Apple to (negligently) allow someone to use a potentially unsafe or hackable product. To be maximally ethical, Apple should be remotely bricking any Apple device the day it loses security support. Better to be on the safe ("ethical") side and remote wipe too, to protect people's data privacy.
Can you see how Argument From I Merely Assert XYZ Is Unethical Then Demand Some Arbitrary Relief can quickly break down?
I don't think there is a real concern about app store economics collapsing, the app marketplace business is very lucrative. We can see this in related cases: you can avoid certain iOS taxes by purchasing your subscriptions on the web: Twitter Blue is $11 on iOS and $8 on the web. Spotify used to be $12.99 for iOS sign-ups and $9.99 on the web.
Why should users pay full bundled iOS tax that supports security updates, if they are getting none?
I don't think I've ever seen someone express the expectation that Apple's App Store fees are for the purpose of supporting iOS development and maintenance. Mostly I've seen and heard the expectation that those fees are connected to running the App Store itself (payment processing, hosting, app review, etc.) and beyond that, vague profiteering. iOS itself isn't a subscription service, and Apple seems quite happy to sell you devices even if you don't spend money in the App Store. So you seem to be stretching a bit by attributing those fees to iOS maintenance and then turning around to say that unsupported iOS versions should get a discount on the fees for any services that still work.
> iOS 12.5.7
> Released January 23, 2023
I do wish apple would follow google's example and commit to a service lifetime upfront, but other than that, I don't object to their model. in practice, it vastly exceeds the level of support for any android phone other than the pixel 8, and we have yet to see whether google actually follows through on that.
This device had 256Mb RAM and 400Mhz 32 bit processor. Should Apple still support this with security updates?
Like the battery issue, I feel the whole issue is communication. Apple needs to communicate when they EOL OS versions. You don’t otherwise know it, partly because EOL OS’s, including this phone’s, still get security updates, just not all of them.
Is the implication that once a device is EOL that a vendor should never release an update for that device again?
It seems typical for vendors use "EOL" to refer to end of support life, not merely discontinuing sales of the produce. Most notably, that's how Microsoft generally frames EOL for major Windows releases, hence expectation of jumps in PC sales corresponding to EOL of XP, 7, and 10.
I feel like Apple changed the dynamics of smartphone market from company-issued devices like BlackBerry to BYO with the iPhone essentially on purpose so they don’t get stuck providing decades of enterprise support promises like companies like Microsoft.
Companies purchasing bulk orders of hardware probably wouldn’t tolerate a vendor unwilling to make any sort of concrete support promise for the contract. But a company who employs iPhone users can basically put the responsibility on the user and simply block access to non-compliant devices.
No, I’m not implying there’s something wrong with shipping the occasional update to EOL devices.
There is nothing fundamentally incompatible about the last couple of generation of iphones. ARMv8 CPU, PowerVR derived GPU. If the mobile computing space weren't driven by greed, this would be a non issue.
A Sandy Bridge era intel machine deployed in 2011 is easily capable of running the latest Linux, BSD or win10. And in the case of the first two, I'd wager it will continue to be viable for the foreseeable future.
- Linux dropping support for old graphics drivers (Nov 2023) - https://www.phoronix.com/news/Linux-Drop-Old-UMS-DRM-Infra
- Linux Kernel Developers Discuss Dropping A Bunch Of Old CPUs (Jan 2021) - https://www.phoronix.com/news/2021-Linux-Drop-Old-CPUs
Supporting all of these is work. It makes development of new features harder, because it has to account for quirks of older hardware. Older hardware is also harder to get in the hands of developers and harder to test on. That’s why Linux has dropped support for 386, 486, IA-64 and other architectures.
There’s no point saying trillion dollar corporation etc. It comes down to some basic fact - phones must be built with SoCs, that’s the easiest way. The PC way doesn’t work at scale. Now that we are on SoCs you have to draw the line on support somewhere. Just because the costs imposed on future development aren’t obvious to us doesn’t mean they don’t exist.
I think 5 years minimum (and sometimes more) of OS updates is pretty good, FWIW.
Don't sell crap you can't support for a decent amount of time. Stop ruining this planet we live on by creating immense amounts of e-waste every few years.
We both know your argument is dishonest or at least naive, though. They could easily support updates if they want to. But it's about money. This way they are forcing people to buy a new phone every few years. It's clever, shame about the planet.
Dishonest? You're saying I'm lying to support a trillion dollar corporation I have no financial stake in and never have? Is such an accusation really in the spirit of this forum?
I suggest you review the guidelines - https://news.ycombinator.com/newsguidelines.html
Don't be snarky. Don't sneer. Assume good faith.
Also, your claim that it's "not economical" is entirely unproven and arguably false. iPhones are still architecturally supported by Linux and will continue to be for a while (even longer on BSD). Other Apple products (eg. Apple Silicon) received community driver support entirely from donations and volunteer time. There's no reason to assume that iPhones lack community interest, especially since Apple has never given the iPhone community the same leverage they had on Mac.
If that's the sum of both arguments, then you're mostly just leveraging FOMO to support an unproven concept. At best you're jumping the gun, at worst you're twisting the facts to preclude discussion of open iPhone software alternatives.
Think about how many devices Microsoft has to support in Windows, it’s orders of magnitude more.
Apple doesn’t want to support older devices because they don’t see a benefit to themselves.
5 years of support is pitifully short. Pretty much everything I own lasts longer than 5 years, my phone is one of the things I have to replace most often, not because the hardware is broken, but because it stops receiving updates.
Again, bugs are this are not hardware specific. You are not supporting "devices". You are supporting the OS which all of them run. Ideally (I'm not familiar with OSX/iOS internals) all they have to do is push out an update that contains the newly fixed libwebkit.so or whatever. They control everything on their own platform so they don't even have to deal with glibc breaking backcompat like we have to in the GNU/Linux world.
If they can't figure out a way to make changes like this universal across devices, it's either deliberate negligence or incompetence.
Microsoft is trying to fix this. Win 11 wants a TPM. /s
The more recent SE models are iPhone 8 bodies with upgraded internals (being an iPhone 11 and 13 inside, respectively) and Apple no longer offers a Mini version of their newer phones (not that the A16 and A17 are significant improvements on the A15, but still).
And the Mini is still larger than the first SE by about a half inch in either direction; the 8 (and the newer phones) add another half inch on top of that.
Of course, the problem with using a first-gen SE today is that because information density on mobile is abysmal, the added vertical space has been very welcome to app developers (who also might not test on the smallest screens). So while the smaller phones were ergonomically far superior to the larger ones, and this is to a point true for the Minis as well, that's since been "balanced" by inherently worse UI/UX on said phones.
With each phone I have always taken steps to reduce battery usage, disabled background refresh, and manually turn on low power mode after charging every single time.
It’s a feature called Reachability.
https://support.apple.com/guide/iphone/use-reachability-iph1...
Even iOS 12 had a security update in 2023 still.
But don’t get too excited.
If they really wanted to update phones that are full they could move the images/video to their server until the update is done.
If it weren't for a friend giving me his old iPhone XS as thanks for a favor, I'd probably still be using my old iPhone 6s--and I would not have worried about it from a security perspective purely because (as the article notes) Apple is still releasing security fixes for iOS 15. I'd feel differently if Apple had publicly stated that all iOS 15 security fixes from now on will be on a minimum effort basis only.
Windows Phone ? I bet nobody writes exploits for it anymore. /s
The phone is almost 10 years old, the only thing Apple should do is send out a push notification warning users their device is no longer supported.
Sure you can argue that it's "easy" to install Linux on it, but that's out of scope for the typical user.
> you can argue that it's "easy" to install Linux on it
In my experience, the average Linux install wizard is less technically involved than the MacOS one. So yeah, maybe I would argue that.
I wouldn't expect the typical person to be able to do it.
Plus, you still haven't touched the central point; tightly coupled software and hardware creates more e-waste. What people do today doesn't matter if they're never presented a serious alternative; if ignorance was a basis for removing functionality there would be nothing left of the modern smartphone.
> iOS 15.8 is still powering those devices, seemingly with official Apple support, with the latest update from October the 25th, which addressed some security vulnerabilities.
> In reality, however, Apple seems to only be addressing some security issues in its older supported devices.
The author is surprised, reasonably IMO, that if a device is getting security updates, you are still exposed to known vulnerabilities.
For a tech competent user this is worse than clearly being out of support in which case you would retire the device.
No reasonable person is asking for this.
The linked piece specifically refers to iPhone 6S and iOS 15.8.
> A new exploit targeting the iOS (among other OS’) Bluetooth stack has also been left unpatched by Apple in all versions except iOS 17.
You'll probably always be safe from remote code execution on supported versions of iOS/macOS, but exploits like these aren't always fixed. Unfortunately, you can't install an alternative browser engine on iOS/iPadOS, so I guess you should expect all websites to be able to read all of your cookies just in case.
...more precisely, for you to see; this needs to be combined with something to send back data (JS?) to be truly exploited.
The attack assumes that the victim is visiting an attacker-controlled web server.
If the attacker can put secret data in the DOM within the victim's browser, the attacker can also add JS on the same page that POSTs the DOM contents to the server once they're populated with secrets.
If the page body can read a file, then it can just execute an XmlHttpRequest to send that data to the origin server, which is the attacker in this scenario. This is just how the web works, nothing more to say about it, and no need to prove it.
I hate that I wrote that lol. It reeks of the kind of cybersecurity whataboutism that leads to people inconveniencing the SHIT out of themselves for the sake of security.
Edit: Though to keep things fair, it must be said that that particular argument only really applies to old devices. If you have a new device (or rather, one still getting regular security updates from the vendor) and a trustworthy vendor, a person could reasonably argue for staying on the stock ROM.
Installing random ROMs from random developers is an interesting take on security.