Hacker News Reader

Comment by defen | Hacker News Reader

"Mijangos told me that he'd figured out how to turn off a camera's LED, cloaking himself completely."

Anyone know if this is actually possible? I don't know anything about laptop cameras, but it seems like you wouldn't want the LED to even be under software/firmware control - just put it in series with the camera circuit. An LED has to have a significantly lower failure rate than a camera, right?

View on HN

There's some interesting answers to this question on security.stackexchange http://security.stackexchange.com/questions/6758/can-webcams... . Short answer is that it looks to be possible with some webcams but not others, depending on the design.

I know that with my Logitech Pro 9000, the Logitech software that comes with it gives you the ability to turn off the LED. Their software also has remote viewing built in. I'm sure designs vary, but it is at least possible with this camera.

I believe some Thinkpads' drivers expose a /sys interface to disable/blink that light.

I can't find where I read this, but I believe the reset line of the camera module is typically tied to the LED, so it can't be in a powered-on state without the LED being active.

"typically"? Perhaps there are a significant number of models for which this isn't the case.

That was interesting. I also have a hard time believing his claim about infecting a phone with a text.

There would have to be a buffer overrun in the text message handling code plus a way to exploit it in 160 characters. That sounds difficult, unless I'm wrong about the fact that the carrier enforces the limit.

I could maybe see doing it with MMS or iMessage. The more I think about it, the more interesting this question is...

Has anyone heard of any exploitable flaws in a phone's SMS software?

From what I have seen longer messages get sent as multiple segmented texts.

This might be of interest: Fuzzing the Phone in your Phone [26C3]

http://www.youtube.com/watch?v=qBWc67iy4zI

CVE-2009-2204 was a vulnerability in iOS' SMS handling (versions < 3.0.1).

http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2009-2204

It's not just buffer overruns that can cause issues. You can crash the font display system by sending characters that aren't handled properly. You may also be able to direct the phone to download a hacked firmware update via SMS (AIUI carriers sometimes use specially-coded SMS messages to tell phones to update their software, PRL, etc.).

On windows mobile 6.0, you could send wap pushes that linked to signed apps which would auto-execute/install.

right now, there aren't any vulns which are similar in danger that I am aware of. SMS isn't a super friendly medium for stack manipulation, and most modern mobile OS'es implement ASLR.

The browser is the more likely vector today.

Georgia Weidman has a botnet C&C (Command and Control) network running via hidden SMS. But I don't think it can infect via text message. http://georgiaweidman.com/wordpress/?cat=10

See my comment.

Reply on news.ycombinator.com