←back to thread

845 points the-anarchist | 1 comments | | HN request time: 0.233s | source
Show context
userbinator ◴[] No.44334486[source]
making it nearly impossible for regular users to uninstall it without root access, which voids warranties and poses security risks

Stop parroting the corporate propaganda that put us into this stupid situation in the first place. Having root access on devices you own should be a fundamental right, as otherwise it's not ownership.

replies(12): >>44334515 #>>44334549 #>>44334577 #>>44334616 #>>44334661 #>>44334912 #>>44335283 #>>44335463 #>>44335597 #>>44336211 #>>44336257 #>>44336433 #
ulrikrasmussen ◴[] No.44335283[source]
We need regulation which defines that any hardware device capable of running software developed by a third party different from the hardware manufacturer qualifies as a general purpose computing device, and that any such device is disallowed to put cryptographic or other restrictions on what software the user wants to execute. This pertains to all programmable components on the device, including low-level hardware controllers.

These restrictions extend outside the particular device. It must also be illegal as a commercial entity to enforce security schemes which involve remote attestation of the software stack on the client device such that service providers can refuse to service clients based on failing attestation. Service providers have other means of protecting themselves, taking away users control of their own devices is a heavy handed and unnecessarily draconian approach which ultimately only benefits the ad company that happens to make the software stack since they also benefit from restricting what software users can run. Hypothetically, they might be interested in making it impossible to modify video players to skip ads.

replies(3): >>44335513 #>>44335681 #>>44335780 #
akoboldfrying ◴[] No.44335513[source]
> any such device is disallowed to put cryptographic or other restrictions on what software the user wants to execute

Won't this also forbid virus scanners that quarantine files?

> This pertains to all programmable components on the device, including low-level hardware controllers.

I don't think it's reasonable to expect any manufacturer to uphold a warranty if making unlimited changes to the system is permitted.

replies(3): >>44335670 #>>44335696 #>>44335749 #
fc417fc802 ◴[] No.44335670[source]
It wouldn't forbid shipping the device with a virus scanner. It would only forbid refusing the user control over what software does and does not run.

There might be a couple messy edge cases if applied at the software level but I think it would work well.

Applied at the hardware level it would be very clear cut. It would simply outlaw technical measures taken to prevent the user from installing an arbitrary OS on the device.

Regarding warranties, what's so difficult about flashing a stock image to a device being serviced? At least in the US wasn't this already settled long ago by Magnuson-Moss? https://en.wikipedia.org/wiki/Magnuson%E2%80%93Moss_Warranty...

replies(1): >>44342726 #
akoboldfrying ◴[] No.44342726[source]
> what's so difficult about flashing a stock image to a device being serviced?

Yes, I think that would cover most cases if we take it to its logical conclusion of wiping all device state (hard disk). OTOH, a few points:

1. I would accept the need to wipe the hard disk if I had messed with firmware or the OS, but not if a couple of keys on the keyboard had stopped working. This implies that (for me at least) a meaningful distinction remains between these two "levels" of warranty service. Do you agree?

2. Activities like overclocking or overvolting a CPU have the potential to cause lasting damage that can't be reversed by re-flashing. Under the policy you're suggesting, it would be illegal for manufacturers to offer users the option "You can pull this pin low to overclock outside the supported range, but you will void the warranty by doing so", and too expensive for them to endlessly replace parts damaged by these activities for free under warranty, so that consumer option, rare as it already is, would go away completely.

3. I still think there may be some devices that are impractical to completely re-flash. According to this 2021 Porsche article [0], modern cars contain 70-100 ECUs (microcontrollers), each of which will have its own flash/EEPROM.

[0]: https://medium.com/next-level-german-engineering/porsche-fut...

replies(1): >>44350184 #
fc417fc802 ◴[] No.44350184[source]
I think there's somewhat of a misunderstanding. The suggestion is that warranties not be denied due to user modifications that are unrelated or trivial to revert. It's the logical extension of Magnusun-Moss from aftermarket parts to aftermarket software packages.

1. I expect wiping any given component to be entirely up to the manufacturer's discretion. If doing so is not trivial and is legitimately required for the repair to proceed then I'd expect the user to be charged for the additional service.

2. Violating manufacturer specifications and being at fault for damages are sometimes distinct. A manufacturer arbitrarily saying "you must not do X" should not necessarily mean that doing X will void the warranty. It might though. Discretion is obviously required.

3. If your car stops working after you mess with the firmware and you take it in to the dealer I imagine they'd charge you to reflash things since the issue was caused by your own actions. That doesn't mean they should be able to decline to cover entirely unrelated defects.

Also I don't think vehicle firmware would be caught up by the original proposal in the first place since cars aren't generally intended to run third party software. There's a grey area with infotainment systems that have an app store depending on if those are viewed as standalone or part of the larger vehicle. However reframing the proposal to revolve around intent would likely leave the firmware on unrelated embedded components in the clear to be locked down so long as those components don't interfere with the ability to freely use the general purpose computing element.

Personally I'd like vehicle firmware to be covered by similar protections but I recognize that falls outside the scope of a proposal about products intended for use as general purpose computing devices.

replies(1): >>44360393 #
1. akoboldfrying ◴[] No.44360393[source]
I like the concept of unrelated modifications remaining under warranty. That seems reasonable and also, crucially, amenable to the usual "Would a reasonable person say that damage X increases the chances of damage Y occurring?" type of test.

I don't like the "intended for general purpose computing" concept so much. For one, it seems to offer lots of easy wiggle room to manufacturers: Just say that your product is not intended for that, but for something marginally more specific. For another, it's not clear to me why general purpose computing ought to enjoy consumer protections that other manufactured devices do not. (One exception I'd grant is for safety reasons: If tinkering with a device could make it cause injury, fine, that device can be in a different class.)