Slashdot Mirror
Ask Slashdot: Is GNU/Linux Malware a Real Threat?
New submitter m.alessandrini writes "I've been using Debian for a long time, and I'm not a novice at all; I install system updates almost daily, I avoid risky behaviors on Internet, and like all Linux users I always felt safe. Yesterday my webcam suddenly turned on, and turned off after several minutes. I'm pretty sure it was nothing serious, but I started thinking about malware. At work I use noscript and other tools, but at home I have a more relaxed browser to be used by other family members, too. Here I'm not talking about rootkits or privilege escalation (I trust Debian), I think more of normal user compromise. For example, these days much malware come from malicious scripts in sites, even in advertising banners inside trusted sites, and this is more 'cross-platform' than normal viruses. So, what about non-root user malware? How much could this be real? And how can you diagnose it?"
Your webcam turned on, then off, and you didn't ask it to? I think you need to figure out what happened first.
It was just Skynet checking out what you were up to. Or maybe the ATF. Or Russian Mafia. Or...
As for security, ~5 years ago read someone's account of watching while someone on the internet installed a root kit on his Linux box in a matter of minutes.
Presumably some platforms/applications are less likely to be compromised than others, but the safest assumption is that everything is compromised, or would be if the experts wanted it.
Sheesh, evil *and* a jerk. -- Jade
As long as you have people on Ubuntu forums posting "sudo apt-get " as the solution to everything without explaining what they do, and as long as you have people willing to copy/paste the commands without understanding what they are doing, then malware is a threat.
The same groupthink plagues the Arch Linux forums. Blindly copy/pasting commands that someone else put on a wiki does not make you elite, it makes you an idiot.
The same issue exists in adding repositories from untrusted sources. What's the point of running an enterprise-class operating system if the first thing you do is add a third party repo from Russia and update the kernel with something ending -kmod?
The critical mass of idiot users still reside in Windows, where things like UAC and walled gardens exist to protect them somewhat. At least there, you have to know the administrator password to do real damage. Ubuntu and all the new user-friendly distros are content to put every new account in /etc/sudoers and allow you to use your own password to gain root access. Any operating system is prone to malware so long as people are willing to bend security practices.
Linux is much easier to exploit than Windows. All of its internals are well understood, and there are more things one can do with shell access.
2003 is calling. They want their FUD back.
"I don't know, therefore Aliens" Wafflebox1
That is what SELinux and AppArmor are for. They might not be 100% (as there were some kernel exploits that could be used to bypass those), but with proper policies in place, something getting UID 0 would be pretty limited in what it can accomplish.
OS X also has a similar mechanism in place.
Linux also has a bunch of different distributions. A bug that causes SSL keys to be very weak in Ubuntu is not going to affect RedHat systems.
This doesn't mean Linux is worry-free, but it is more secure than people think. To cite an anecdotal example, the proof is in the pudding -- look at all the amateurish Apache servers and LAMP stacks out there. If Linux had major issues in general, there would be major screaming on almost every forum how insecure the OS is.