Your phone doesn't emit any signals at all when turned off. If it did, people from the FAA to every geek with a spectrum analyzer at work would be screaming bloody murder.
Can the phone be made to receive while "powered off" -- even sporadically?
They wouldn't want to tip their hand re:transmission capabilities. They'd probably only activate the capability on-demand, and only for high-value targets (ie. by sending a special packet to be received by the "powered off" handset). It's sorta like turning your phone into a listening "bug" while powered on... you wouldn't want to constantly stream data and tip your hand -- you'd only activate the capability when there is a strong likelihood of a valuable conversation occurring.
It probably wouldn't take noticeable power to run voice-activated microphone audio compression and storage. Then the recorded audio could be transmitted later when the battery was back in.
Any attempt to do stuff like that will still require a clock signal somewhere. That clock will be noticed. An attempt to do it in the analog domain with no synchronous logic would be even more apparent to anyone who studies the internals of the phone.
In short, no, it ain't gonna happen. It would show up on Bunnie Huang's blog one day and there would be blood in the streets the next.
I'm looking at my phone and the power button is conspicuously not a toggle switch. You press the same button to turn it off as to turn it on. That means the phone has to have some logic enabled (even when it's "off") to distinguish whether closing the circuit on that button is supposed to turn the device on or turn it off (or wake up the screen or lock the device etc.)
Is it so hard to imagine part of that logic including the mobile phone equivalent of wake-on-lan?
Is it so hard to imagine part of that logic including the mobile phone equivalent of wake-on-lan?
Yes. You don't have to power up the cellular transceiver to poll a power button. That's usually handled by a dedicated power-management chip with few/no other capabilities (and an internal clock in the kHz range at most, I'd expect.)
Also you don't need to poll a power button - it's perfectly possible to build a momentary switch which would connect a circuit long enough to turn on a FET to the battery, which could then be monitored by digital logic to use the same connection to power off.
> Any attempt to do stuff like that will still require a clock signal somewhere. That clock will be noticed.
I wouldn't expect to get away with it on all phones all the time. But certain phones targeted by surveillance malware? Sure.
Note that powered-off phones still have real time clocks running, for things like keeping time.
> An attempt to do it in the analog domain with no synchronous logic would be even more apparent to anyone who studies the internals of the phone.
I agree, there's little chance of an analog tape recorder being in the phone.
> In short, no, it ain't gonna happen. It would show up on Bunnie Huang's blog one day
Go ask Bunnie if he thinks a well-funded attacker could implement it, having access to the baseband design docs. Then ask him how long it could be used in the wild against unsophisticated targets before it ended up on his blog.
> and there would be blood in the streets the next.
This surveillance malware is being found on phones and in some places there is, in fact, literal blood in the streets.
"The FBI can access cell phones and modify them remotely without ever having to physically handle them," James Atkinson, a counterintelligence security consultant, told ABC News.
They issue the update while the phone is powered on. They can then change the behavior of the "shut down" element in the UI to black out the screen and refuse calls but continue to transmit GPS or audio on the radios.
