Hacker News Reader

Comment by sixhobbits | Hacker News Reader

what tech does this even use? Do they mean using Pegasus or similar malware that the govt has to first get onto the suspect's devices, or is this via Google/Apple or the device manufacturers that makes 'remotely and secretly activating a microphone' even possible?

View on HN

This bill AFIK only covers the "they are allowed to do it" part but not the "how do they gain the capability to do it" part.

But spyware which can do so exists in endless amounts, including from companies focused on selling it to governments.

Hence also why in recent years physical microphone switches, or e.g. stuff like (I think it was) Apple laptops "physically" disconnecting the microphone/camera if you close the lid have been become increasingly more common and in demand. (Through the demand comes more from bad actors using it then from people being afraid the government spies on them AFIK, but technically there is 100% no difference)

Also, I recall that the green light on the apple camera is controlled by the camera's firmware, making it more difficult to turn on without the light come on. (You'd need to overwrite the camera firmware.)

Why does this even need to be controlled with firmware? Can't it be a simple circuit? (I know nothing about this)

It could be simple, just hook the led to the reset line of the camera and you'll know if the camera is in use. Downside is that the led will blink on boot when linux probes the driver.

> Downside is that the led will blink on boot when linux probes the driver.

Sounds like an upside to me.

I should have included the caveat that my comment was from memory.

I just did some brief research, and it seems that it was the case prior to some point in time between 2008 and 2019, but it is now hard-wired. It is discussed near the end of https://daringfireball.net/2019/02/on_covering_webcams

The engineer quoted in the article pointed out that it may be possible (at least in 2019) to briefly turn the camera on - flashing the led too quickly for someone to see.

I didn't find a discussion of this issue in the May 2022 Apple Platform Security guide.

iSight cameras had an LED that was supposedly unbypassable. Turns out that was bullshit and it was trivial to rewrite the iSight module's firmware.

Could you reflash it without access to the device?

Probably, if you hacked a device to a point you have root access or worse it's not that uncommon that you are able to reflash firmware to custom firmware.

A lot of firmware, and firmware signature validation, is ... not very well done.

Practically it's often not worth it (you already had root+ access. It's doable, but not simple and less uniform applicable.).

But there have been known cases of viruses tryign to persist themself in the firmware of connected devices (which in this context includes all internal devices including the motherboard).

That is the difference between something being firmware controlled and hardwired.

And for new macs the off switch for microphone/mic if the lid is closed is hard wired using some "dump circuit logic switch" AFIK.

The problem with lights (even if hard wired) is that you might still find ways to brake them, like finding a way to fry them without braking the camera or switching on/off so fast that it isn't really visible but you still get some image/sound you might be able to post process to a point where it's usable even if not grate. You probably can prevent this with further hard wired circuits, like a hard wired "super slow de-bouncer" which in a on-1->off-2->on makes sure 2 is prevented if not at least Xs passed since 1. And which in a off-2->on-3->off cycle makes sure the LED is on for at least Ys (the camera can be switched off faster, just not on again). But I don't think apple has something like that. At that points the question is why not put in physical switches like e.g. on Framework laptops or some older ThinkPads.

> The bill allows the geolocation of crime suspects, covering other devices like laptops, cars and connected devices, just as it could be remotely activated to record sound and images of people suspected of terror offences, as well as delinquency and organised crime.

Not just phones looks like.

> cars

For the very strange who accept driving the new "smartphones with wheels".

Including, note, the cars with the embedded telephone as mandated by the european union past 2018 - the e-call systems. Some articles went "there could be privacy issues, but it is a remote eventuality": now you see that someone could push as normal an eavesdropper in your car.

"A remote eventuality" my ass, I bet that's the primary reason behind the push.

Sometimes, Stavros, it is a painful choice to decide in assessment between malice and the various shades of stupidity - the perspective of the latter can be a tad more difficult to digest.

But the last time I read reported unhandable blabbering devoid of awareness and mental competence, attributed to a "minister of the republic", was just minutes ago. When you hear utterings like "it will save lives" - completely alien to conscience of quality of life, good reason, cleanness and propriety, rejection of absurdity, collaterals etc. - I am afraid they may actually "believe" that (or, better, just "sit on that", "hold on that", "cling to that" in some internal economy).

Hm, yeah, you're right. It's a difficult choice because both options are so dire, but I guess they have to be when you have a decision like this.

Is this even possible on an iPhone?

My memory may be failing me or confusing things so please correct me, but I seem to recall reading somewhere that the baseband lives segregated (with only a narrow communication cannel, kinda as if it were a remote machine) from the remainder of the hardware, so while it could be made to run stuff itself it has no way to physically access to main cpu, ram, mic, nor cams (barring, of course, any vulnerability on the comm channel that would land an exploit in the main OS+hardware).

GPS is another matter, but then again it's baseband so it gotta communicates with towers, so that's a done deal already that does not even require baseband access.

Apple is good at /marketing/ their security, not being good at security.

I take pride in striving to form my opinions based on verifiable facts as discovered by third party sources, such as via work done by the Asahi Linux team†.

If I were buying into marketing I would not even have asked the question above in hopes of getting a hard reference to these kinds of details.

† e.g the following tidbits:

> The SEP is designed to be mutually isolated from the main CPU (AP); neither can compromise the other.

> No runtime blobs are designed to have total system access (no ME, no PSP, no TrustZone, nothing of that sort). Almost all blobs are running behind IOMMUs or similar firewalls, with the sole exception of the GPU firmware*. All code running on the main CPUs is under the control of the OS.

> AS machines use a large number of auxiliary firmware blobs, each dedicated to a specific purpose and running on a separate CPU core. This is better than having a smaller number of kitchen sink blobs (like Intel ME), since each blob can only affect a particular subsystem (e.g. display, storage, camera), which makes it harder for multiple blobs to collude in order to compromise the user in a meaningful way. For example, the blob running inside the keyboard controller has no mechanism to communicate with the blob running on the WiFi card, and thus cannot implement a keylogger surreptitiously; the blob running on the display controller similarly has no way to communicate with the network, and thus can't implement a secret screen scraper.

> From a security perspective, these machines may possibly qualify as the most secure general purpose computers available to the public which support third-party OSes, in terms of resistance to attack by non-owners. This is, of course, predicated on some level of trust in Apple, but some level of trust in the manufacturer is required for any system (there is no way to prove the non-existence of hardware backdoors on any machine, so this is not as much of a sticking point as it might initially seem).

> *it's worth pointing out that this firmware is not particularly large, is shipped in plain text and even with some symbols, does not have any functionality to talk via questionable interfaces (network, etc.), and is optional and not running when the OS boots (the OS must explicitly start it)

Which is an entirely different league than, say, Intel ME which completely owns the machine at the design level.

https://github.com/AsahiLinux/docs/wiki/Introduction-to-Appl...

https://github.com/AsahiLinux/docs/wiki/Apple-Platform-Secur...

"Apple complies with local laws in each of the countries in which it operates."

Looks like any app can be installed silently behind your back with GTalkService that is running on your Android phone

https://github.com/CellularPrivacy/Android-IMSI-Catcher-Dete...

https://jon.oberheide.org/blog/2010/06/28/a-peek-inside-the-...

---------------------

Google Play Services spyware discussion

https://forum.xda-developers.com/t/guide-insanely-better-bat...

https://forum.xda-developers.com/t/app-disable-service-guide...

---------------------

"...the cellular carrier can send blobs of FORTH code right to the radio. The radio firmware also seems to have an IP stack (with TCP) so it can do its own interesting things (both bad and good)..." https://boston.conman.org/2013/01/22.2

"...easily spotted loads and loads of bugs, scattered all over the place, each and every one of which could lead to exploits – crashing the device, and even allowing the attacker to remotely execute code. Remember: all over the air. One of the exploits he found required nothing more but a 73 byte message to get remote code execution. Over the air..."

"... It’s kind of a sobering thought that mobile communications, the cornerstone of the modern world in both developed and developing regions, pivots around software that is of dubious quality, poorly understood, entirely proprietary, and wholly insecure by design." https://www.osnews.com/story/27416/the-second-operating-syst... (archive: https://archive.is/FOR5V)

https://news.ycombinator.com/item?id=6722539

https://news.ycombinator.com/item?id=6722732

https://news.ycombinator.com/item?id=6722648

https://news.ycombinator.com/item?id=6738066

https://news.ycombinator.com/item?id=6724034 <-- Seems to be higher risk with Qualcomm basebands where everything is integrated

-------------------

SIM card reader chips have their own operating system https://en.wikipedia.org/wiki/SIM_card#Design

Rooting SIM cards https://archive.is/3ZohQ

https://news.ycombinator.com/item?id=6722896

https://news.ycombinator.com/item?id=6724215

https://news.ycombinator.com/item?id=6723236

-------------------

It's insecure by design because the NSA has for decades purposefully degraded the security of everything they can get their hands on to make it easier for them and law enforcement to spy on.

They don't want you listening in on John Q. Senator's phone calls, but they sure do...

The workaround is to carry around a small insulated faraday cage for your phone which blocks cellular reception, transmits bogus GPS signals locating yourself as a random wanderer, and plays television series reruns 24/7 at low volume into the microphone.

Replying so I can read all of these later.

part of it is already being done: judges in France just ask companies like FAANG for location data, including live location. E2E encryption is the only way for companies to be able to refuse (and this is why there’s been a strong governmental push against it).

The scary new part is the turning on the camera/mic.

E.U. is planning to purchase Pegasus for law enforcement purposes:

https://www.europarl.europa.eu/news/en/press-room/20230609IP...

"They want EU rules on the use of spyware by law enforcement, which should only be authorised in exceptional cases for a pre-defined purpose and a limited time."

> what tech does this even use? Do they mean using Pegasus or similar malware that the govt has to first get onto the suspect's devices, or is this via Google/Apple or the device manufacturers that makes 'remotely and secretly activating a microphone' even possible?

It would have to be after compromise, which would mean its likely only used on a very small number of cases due to the sensitivity and cost of the technology involved.

Right.

Is it expensive to compromise a phone?

Is it sensitive to compromise a phone, now that there is a national law allowing it, passed through a democratic process?

This might be quibbling but true democratic process could be interpreted as direct democracy where the law is directly voted for and people give a specific assent to the law (which ancient Athens did do). This is more of a democratic republic where officials elected through democratic process then make decisions, which usually works more smoothly than a pure democracy, but it does mean that the population at large does not necessarily support this specific law (though they might).

> but it does mean that the population at large does not necessarily support this specific law

If the population does not support the law, the government could be potentially replaced.

But the most probable conclusion is that the population at large would not care.

> Is it expensive to compromise a phone?

2-2.5+ million USD depending for a full chain with persistence (lower end of range is for iOS higher end is for Android).

That is probably the lower end of the rough cost to buy that capability which you can use as many times as you want.

https://zerodium.com/images/zerodium_prices_mobiles.png

> Is it sensitive to compromise a phone, now that there is a national law allowing it, passed through a democratic process?

The technology itself is sensitive, when you buy a full chain exploit like the ones that have the public bounty price above, if it gets burnt it's useless for everyone else who bought it after its patched.

Generally exploit brokers don't like it when you burn their exploits.

So, on the order of a dollar per person, assuming a town/municipality of a couple million people?

Is that a lot?

In the US the major cities have police budgets of between $200 million and $5 billion

> In the US the major cities have police budgets of between $200 million and $5 billion

Yeah and most didn’t have this capability generally considering both the costs and the little amount of use you'd have for it.

You'd expect something more federal like the NSA, CIA or the FBI to have this kind of capability which is why its kinda a big deal when normal cops get it.

>Yeah and most didn’t have this capability generally considering both the costs and the little amount of use you'd have for it.

I would say most didn't have this capability because it's illegal. In France it is now legal

This bill could be step one. Step two could be a requirement for SW/HW manufacturers to add a backdoor since this is not really effective to fight crime this way. I'm not saying it is planned or there's some conspiracy for establishing a totalitarian state.

But we can't really predict the future and more loose rules could be introduced by the next government with a totally different agenda who might thank for the previous one for creating this legal framework.

Also, this section is weird too:

> They said sensitive professions, including doctors, journalists, lawyers, judges and MPs, would not be legitimate targets.

Apparently software engineering is not a sensitive job.

> Step two could be a requirement for SW/HW manufacturers to add a backdoor since this is not really effective to fight crime this way

We have already been there, it's just like good old phone calls. They can be intercepted under the proper judicial supervision.

I don't have criteria to answer if this is a necessary evil or directly a blatant totalitarian push, but one needs to consider: does the French government have a bad track record of abusing the capability of eavesdropping any phone communication? has this capability been useful for law enforcement?

The slippery slope argument can be overused.

But any surveillance capability results in “good” as well as bad actors, frustrated with the current limits, asking for just a little more formal power or finding ways to justify informal drift in practice.

The only way to stop surveillance creep is openness and clear principles more coherent than a recursive “well they have been responsible so far, so far as we know” argument.

Good point, most people however don't expect their phone calls to be private for this very reason.

> does the French government have a bad track record of abusing the capability of eavesdropping any phone communication?

I'd rather look into possible future scenarios, I don't know about past cases, but I don't really follow french politics. I know discussing politics is HN is not encouraged, so I'd just rather suggest to look up what's coming for France (and EU) if the current government fails and major powers switch. Sorry if you already know this.

> has this capability been useful for law enforcement?

In Hungary, Pegasus was used to eavesdrop on many citizens who were simply in an opposing position. Law enforcement is a good cover for data collection and of course can be effective, but we need transparency and safety to see if these tools are abused or not. There are no good answers I'm afraid.

Reply on news.ycombinator.com