Remix.run Logo
▲ kodebach 4 hours ago

I can't prove it, but I still think this is (partly) due to a recent EU regulation majorly backfiring. I can't remember which one it was, but one of the recent regulations mandates that devices get updates for a certain period. Within that period a manufacturer must provide each update for all their devices in a short timeframe. The intention was to prevent flagships getting updates immediately and lower end devices having to wait months. However, due to some (also well intentioned) wording about open source, the release of the source code is counted as the first release of an update and starts the timer. I'm pretty sure Google is delaying public AOSP updates to give their partners (mainly Samsung I suppose) more time to finish building their updates on top of AOSP.

▲geraneum 4 hours ago | parent | next [-]

> EU regulation majorly backfiring

I wouldn't be surprised if large corporations like Apple and Google intentionally use these tactics to pressure regulators through their European users. Siri AI is a prime example. While it's a complicated issue, I certainly wouldn't put such maneuvers past them.

▲ahartmetz 4 hours ago | parent | next [-]

It's funny that they keep trying, it has never once worked. I think US companies are projecting their dislike for regulation and keep being surprised that EU citizens don't agree. We do not love the EU government, but it sometimes does a good thing and we do like these, especially versus the anti-consumer crap that US companies like to pull.

▲mangecoeur a few seconds ago | parent | prev | next [-]

They absolutely do this.

▲xvokcarts 20 minutes ago | parent | prev | next [-]

Do you think it's right if a group of people (the EU) use their power to achieve goals, but somehow wrong wheh another group does the same?

▲sirtaj 13 minutes ago | parent | next [-]

Whether it's "right" depends on the goal of course, but in this case I'm not understanding your point because nobody is prosecuting those companies for doing this.

In any case, these kinds of blanket statements are not very useful, because each case tends to be judged on its own merits. For example, yes, enough people believe that monopolistic tactics are not "right" and laws reflect that.

▲mgol94 14 minutes ago | parent | prev | next [-]

yes, because that second "group" is actually an international corporation leeching on society to increase its value for small group of investors

▲foepys 13 minutes ago | parent | prev [-]

Is it okay for Russia to influence elections in the EU or the US by spreading misinformation?

This is also an instance of a group using its power.

▲gvfsa 3 hours ago | parent | prev [-]

Giving their partners the time they need to finalise their releases legally seems like a... completely normal and correct thing to do? Why does everything have to be a big conspiracy?

▲dns_snek 2 hours ago | parent [-]

Why would that be a normal and correct thing to do? They're delaying releases for everyone else.

https://grapheneos.social/@GrapheneOS/115164133992525834

https://news.ycombinator.com/item?id=49946698

▲xvokcarts 13 minutes ago | parent [-]

So it is not normal/correct to have an agreement with a partner that excludes others?

▲hobofan 3 hours ago | parent | prev | next [-]

It's an interesting theory. However, Samsung has already started to roll out One UI 9, which is built on top of Android 17 a month ago, so their clock has already started ticking.

I'm not familiar with the concrete wording of the regulation, but I would wager that the all the customization layers (like "One UI") that OEMs put on top of AOSP is the component for which the update clock should tick, and they should be unaffected of an underlying open source component (AOSP) being released.

▲microtonal 4 hours ago | parent | prev | next [-]

As far as I understand, most OEMs are not allowed to release QPR1 or QPR3 either (not that OEMs outside maybe Samsung would be able to). So part of it also reeks of giving Pixel benefits over phones of other manufacturers.

▲dathinab 4 hours ago | parent | next [-]

the more I learn about how Google handles releases the more this looks like there being a need for anti trust actions

▲izacus 3 hours ago | parent | prev [-]

> As far as I understand, most OEMs are not allowed to release QPR1 or QPR3 either (not that OEMs outside maybe Samsung would be able to).

This is not true at all though. OEMs can release anything whenever they want.

▲microtonal 2 hours ago | parent | next [-]

I have read in multiple sources that OEMs cannot roll out QPR1/3. E.g. see this thread:

https://grapheneos.social/@GrapheneOS/117282080803799576

Since they are working with two OEMs (one for GrapheneOS support and one for getting embargoed security patches), I think they know what they are taking about.

▲izacus 6 minutes ago | parent [-]

Nothing of that has to do with OEMs being forbidden from shipping QPRs though.

Graphene isn't OEM so they don't get access to Android repository before AOSP drops. OEMs - like Motorola for example - do have access to partner repository and can ship QPRs if they wish to do so. Most often OEMs do not want to push updates (and it took years to make them push regular major Android updates).

> I think they know what they are taking about.

Err, if you look at that account you'll find that the person running it has a penchant for overexaggeration and selective truth telling when they feel aggrieved.

▲LoganDark 3 hours ago | parent | prev [-]

Except for phones preinstalled with GrapheneOS (unless you are Samsung)

▲dathinab 4 hours ago | parent | prev | next [-]

if that is the case then it should be brought to the respective overseen regulatory body, weather it is for them to clarify that this is not a valid interpretation of the law (it's at least clearly in conflict with the spirit of the law) and intentionally delaying public releases this way is a no-go or weather it is to amend the law to make it more explicitly illegal

either way if that is true it looks like a very intentional abuse/misinterpretation of existing regulations

▲eddythompson80 4 hours ago | parent [-]

What makes you think it’s misinterpretation?

▲pas 2 hours ago | parent [-]

FAQ https://ec.europa.eu/newsroom/dae/redirection/document/12233... [PDF]

https://digital-strategy.ec.europa.eu/en/policies/cra-manufa...

▲dns_snek 2 hours ago | parent [-]

It's not a CRA obligation, but a rule from ecodesign requirements for smartphones (Reg. 2023/1670)

> security updates or corrective updates mentioned under point (a) need to be available to the user at the latest 4 months after the public release of the source code of an update of the underlying operating system or, if the source code is not publicly released, after an update of the same operating system is released by the operating system provider or on any other product of the same brand;

https://eur-lex.europa.eu/eli/reg/2023/1670/oj/eng#anx_II

So by delaying AOSP releases by 4 months, Google is allowing OEMs to leave their users exposed to known vulnerabilities for up to 8 months.

▲notrealyme123 4 hours ago | parent | prev | next [-]

Interesting theory. Is there a distinction between phone release and os release there? I would imagine the time starts when the model goes in sale, not when the os releases.

Otherwise I would expect Google would also wait with pixel updates

▲aboringusername 4 hours ago | parent | prev | next [-]

I posted about this 11 months ago [1]. The law you are referring to [2] absolutely makes distinctions between open/closed source. This is 100% why Google closed their source and now does security releases privately.

[1]: https://news.ycombinator.com/item?id=45911673

[2]: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=OJ%3AJOL...

> Clause 1.2:

> (6) Operating system updates:

> (a)

> from the date of end of placement on the market to at least 5 years after that date, manufacturers, importers or authorised representatives shall, if they provide security updates, corrective updates or functionality updates to an operating system, make such updates available at no cost for all units of a product model with the same operating system;

> (b)

> the requirement referred to in point (a) shall apply both to operating system updates offered voluntarily by manufacturers, importers or authorised representatives and to operating system updates provided to comply with Union law;

> (c)

> security updates or corrective updates mentioned under point (a) need to be available to the user at the latest 4 months after the public release of the source code of an update of the underlying operating system or, if the source code is not publicly released, after an update of the same operating system is released by the operating system provider or on any other product of the same brand;

> (d)

> functionality updates mentioned under point (a) need to be available to the user at the latest 6 months after the public release of the source code of an update of the underlying operating system or, if the source code is not publicly released, after an update of the same operating system is released by the operating system provider or on any other product of the same brand;

> (e) an operating system update may combine security, corrective and functionality updates.

▲yorwba an hour ago | parent [-]

(c) and (d) treat binary releases as equivalent to source code releases, so that's rather the opposite of making a distinction between open and closed source. It would serve as an explanation if Google had started to release source code and binaries simultaneously, but it doesn't explain why the source code release is delayed even beyond that.

EDIT: I see that was already brought up multiple times when you posted about it 11 months ago.

▲izacus 4 hours ago | parent | prev | next [-]

It's Cyber Resilience Act (Regulation (EU) 2024/2847) - if there's a security patch that shows up in public, then the manufacturers are fined if they don't immediately roll it out as an update ("without undue delay").

I think that's the one that made a lot of OEMs pull back on opensource publishing, because their legal departments want to batch the CVE monitoring and reporting obligations into smaller batches.

▲microtonal 4 hours ago | parent | next [-]

Aren’t most OEMs in violation anyway because Linux (as in the kernel) gets regular security patches and most OEMs are on ancient kernel versions?

Also, major Android releases and QPR2s have many security fixes (anything not marked high/critical), yet OEMs are not rolling them out (QPR2) or very late (major releases).

Just to say that OEMs were probably violating the act already before Google started withholding source code.

I also think it’s an attempt of Google closing up AOSP more, otherwise they wouldn’t make it so darn hard to get updated Pixel kernel trees.

▲dathinab 4 hours ago | parent | next [-]

probably yes

withholding security updates seems to avoid a law which requires rolling out security updates seems very much like something which needs to be fixed one way or another

▲imtringued 3 hours ago | parent | prev | next [-]

This reminds me of the OpenAI hack against Hugging Face.

Regulators present an impossible goal to the companies to implement, instead of implementing the goal, the companies try to reward hack their way around it.

The end result is that they basically try to make it impossible to know about vulnerabilities because that would start the clock.

▲microtonal 2 hours ago | parent [-]

But doing timely security updates is not impossible. Apple does it, GrapheneOS does it, Google does it with Pixel, Samsung is nearly doing it on flagships.

But many Android OEMs are probably terrible at software engineering and have their attention spread across multiple SoCs, etc. It’s their own fault.

▲hardbass 43 minutes ago | parent [-]

Then they should leave it to others. Release specs and sources and let whoever wants, maintain the firmwares.

▲izacus 3 hours ago | parent | prev [-]

I think it's mostly the "you need to monitor CVEs and write report within 24/72 hours" that causes OEMs pain and they prefer to have less work than more.

Also - the reality of what you think is ok and what a corporate legal department think its ok is... rarely the same -_-

▲themagic80 38 minutes ago | parent [-]

Law: please do the minimum to protect users. Corps: screaming

▲arghwhat 4 hours ago | parent | prev | next [-]

nit: "without undue delay" does not mean "immediately" under any legal interpretation, only that the timeline should not be longer than necessary. This leaves a lot of wiggle-room in favor of the manufacturer, and mainly rules out that they go "eh, we'll look at all these in half a year maybe".

As long as they start the process and it is not absurd from an industry standards perspective, they should be good. They can have slow processes, including being a good bit slower than their peers, without running afoul of "without undue delay".

▲imtringued 3 hours ago | parent | next [-]

They have literal 24 hour and 72 hour deadlines. Obviously, they might put a blind eye to it if you are reliable but delayed the reporting by a day, but the 24 hour and 72 hour deadlines are a fact. You have to initiate the reporting immediately.

▲arghwhat 2 hours ago | parent | next [-]

You are mixing unrelated requirements together here.

> The CRA requires manufacturers to notify actively exploited vulnerabilities and severe incidents having an impact on the security of their product with digital elements. They need to submit an early warning within 24 hours of becoming aware, and a full notification within 72 hours.

These deadlines are entirely unrelated to the availability of security fixes and delays in deliverying them.

▲alerighi 3 hours ago | parent | prev [-]

This is impossible to happen.

Security flaws are discovered almost every day (now that they are automating it with AI there are a ton of reports), pretending the manufacturer releases an update in 1 day is nonsense.

Just building the code alone it would take a lot of time, and of course you have to test it at least a bit before giving it out to the end users, you can't pretend to rollout on a global scale a nightly build without first at least some level of QA risking bricking user devices.

And you have to consider that phone manufacturer doesn't work h24 and 7/7 of course to give out patches.

I think that we are fortunate if the whole process takes less than a month, not a day.

▲izacus 3 hours ago | parent | prev [-]

> nit: "without undue delay" does not mean "immediately" under any legal interpretation, only that the timeline should not be longer than necessary. This leaves a lot of wiggle-room in favor of the manufacturer, and mainly rules out that they go "eh, we'll look at all these in half a year maybe".

Well, the corporate lawyers think otherwise at those corporations, so I'm not sure how much your interpretation counts :)

▲arghwhat 2 hours ago | parent [-]

I do not believe that any corporate lawyer would think that. It does not read like that in neither plain english nor legalese.

Maybe the misunderstanding stems from the related common misuse of "as soon as possible" as "immediately"?

A corporate lawyer would, however, always suggest to minimize/reduce/delay any legal obligations so as to minimize risk of liability, regardless of whether said obligation is easy to comply with or not.

▲aboringusername 3 hours ago | parent | prev [-]

From what I see [1] if you search for 'undue delay', it's only:

> an early warning notification of an actively exploited vulnerability, without undue delay and in any event within 24 hours of the manufacturer becoming aware of it

> without undue delay and in any event within 72 hours of the manufacturer becoming aware of the actively

> without undue delay and in any event within 24 hours of the manufacturer becoming aware of it

> without undue delay and in any event within 72 hours of the manufacturer becoming aware of the incident

"Sorry your honour, we only became 'aware of it' 2 years later, pinky promise ;)".

So any competant authority would have to prove when they 'became aware of it', whatever the test to pass that threshold means.

[1]: https://eur-lex.europa.eu/eli/reg/2024/2847/oj/eng

▲soraminazuki 2 hours ago | parent | prev | next [-]

So it's the regulator's fault that a trillion dollar company is acting against the public interest? That's some unproven theory indeed.

▲_nobody 2 hours ago | parent | prev [-]

> due to a recent EU regulation majorly backfiring

And why should a regulation in EU affect rest of the world. Or for that matter, any US regulation. The arrogance of these nations and the ignorance of tech companies is ruining the entire Internet.

▲sham1 an hour ago | parent [-]

Because making things everywhere compliant is cheaper than having both a non-EU and EU versions of software. See also: Brussels effect[0].

It's simple economics. The companies are also allowed to either do special versions to comply with local laws or even to pull out of the market. They just choose not to for reasons.

[0]: <https://en.wikipedia.org/wiki/Brussels_effect>