Why removing rootkits is such a pain

Someone once made the comment, “Most people, I think, don’t even know what a rootkit is, so why should they care about it?”

That was in in 2005, when rootkits were an unknown menace for most users. Nowadays, that isn’t quite the case any more, as the number of rootkit infections have exploded in the last few years and lead to more media coverage. In any case, you know a malware has reached evil superstar status when it warrants its own ‘For Dummies’ book.

In the beginning (as in the late 1980s), rootkits were standalone toolkits that allowed hackers to gain root, or administrative access to a computer system (hence the name). Today, the term is usually used to mean programs, codes or techniques that are used to hide malware on an computer.

I’m not going to dwell much on their history or workings (though if you’re interested, Alisa Shevchenko over on Securelist has an excellent article on rootkit history). Instead, I’m going to focus on one particular aspect of rootkits that’s been irritating the daylights out of our Support and Analyst folks recently – why are they so difficult to remove?

Why worry?

Media reports tend to hype ‘rootkits’ as the next big evil in computing, but it’s a bit more complicated than that. For one thing, rootkit tools, coding or techniques aren’t strictly illegal, or even undesirable – perfectly legitimate commercial applications use them to the benefit of users. It also doesn’t help that security vendors don’t have a uniform approach to rootkits; some consider all rootkits as a type of malware, while others shade their evaluations depending on whether the rootkit-like behavior is in a commercial software (in which case, the program may just be potentially unwanted).

Personally, I find it more useful to think of rootkits as operating system controllers. Their entire purpose is to burrow deep into the operating system’s files and subroutines, latching onto and modifying specific processes to gain control over the system. The processes targeted will vary depending on the system and the rootkit in question, but the end result is the same – the rootkit is now in a position to direct the system’s actions for its own ends; it’s become the puppeteer to the computer’s marrionette.

Rootkits have been around a long time, but they only really became a major concern for most users when malware authors found ways to incorporate rootkits into their malicious programs. And for most security professionals, rootkits are considered one of the most troublesome threats to deal with.

How does the rootkit gain so much control?

A rootkit’s defining characteristic is that it has administrative access – its commands are accepted by the operating system as though they were its own. How this access is gained is another story – a separate trojan may exploit a vulnerability to gain access to a administrator account, or a worm might steal the necessary passwords, any number of things. However the access is gained, the end result is that the rootkit is installed with admin rights, and from there proceeds to do its dirty work.

Rootkits use their privileged access to control the operating system itself, mainly by intercepting and modifying the commands it sends to other programs and basic system activities. Slightly more technically, rootkits usually manipulate various application programming interfaces (APIs), or the subroutines used by the operating system to direct operations (at least, in Windows).

An important point to remember is that these APIs are a built-in features of the operating system. They may be undocumented, or rarely used – but commands made through them are perfectly legitimate, and recognized and treated as such. These APIs can involve and affect every activity performed on the computer, from the mundane (e.g., displaying a folder) to the most fundamental  (e.g., booting up).

There are various types of rootkits based on how deeply they can penetrate the operating system to control its most basic processes (if you want to get more technical,  Joanna Rutkowska has a good article), but in every case, the key idea is the same – commands sent by the operating system can be viewed and countermanded by the rootkit, if necessary; likewise, requests coming from other programs or system processes are checked and filtered by the rootkit before they reach the operating system.

Not like other malware

To illustrate why a rootkit’s manipulation of APIs is significant, let’s compare it to other malwares. When a trojan or virus infects a computer, its interactions with the operating system will usually fall into one of two strategies:

  • Strategy 1: Uses the operating system’s standard procedures to run it
  • Strategy 2: Exploits a vulnerability (a flaw or loophole) to execute malicious code

Note that strategy 1 involves the malware functioning just like any other program – its processes and files are visible, the instructions between operating system and program are ‘standard’, and so on. Strategy 2 usually involves some novel technique that forces the system to behave in an unintended manner – ‘breaking  the system’, if you like.

Rootkits on the other hand, doesn’t do either. Unlike trojans or viruses, the rootkit doesn’t behave like a separate program being run on top of the operating system; instead, the rootkit acts more like a driver, or one of the operating system’s own components, giving directions on how other programs should be handled. The rootkit also doesn’t exploit any vulnerabilities – it simply uses the operating system’s own features for its own ends.

The thing is, malwares that use Strategies 1 & 2 can be defeated with fairly standard countermeasures: for example, software vendors can release patches to close vulnerabilities, and users can uninstall malicious programs. Rootkits however don’t suffer either problem: there’s no vulnerability that can be patched, and because a rootkit’s first action is usually to hide itself, the rootkit can effectively prevent the user or the operating system from detecting its presence at all, let alone uninstalling it.

Why are rootkits so difficult to remove?

The highly technical reason for this is: you can’t remove a file you can’t find. Remember, the rootkit is in control. If the user starts looking through system folders for suspicious files, or starts an antivirus scan, a sophisticated rootkit can display a clean ‘image’ of the infected folder rather than the actual infected one, or move the infected file to another location for the duration of the scan; it can stop the antivirus from running, or force it to report false scan results; anything, really, to prevent detection.

Malware authors really want their creations stay installed and active on your computer, and they can use the rootkit to perform any number of actions to prevent their malwares – or the rootkit itself – from being detected. Some of the tricks they can use to get their way include:

  • Renaming their files to match a legitimate system file
  • Burying their processes and files deep within the driver and kernel
  • Installing in such a way that they reinstall again if the computer is rebooted
  • Actively altering its behavior while antiviruses are running to prevent detection
  • Actively changing its own code to make it appear to be a new, unknown program
  • Prevent AV/spyware removal programs from opening at all

Heck, about the only thing they don’t do is say they love you and will still respect you in the morning.

How does an AV detect and remove rootkits, then?

Antivirus programs have historically had a difficult time dealing with rootkits, precisely because of how they operate: by using the operating system itself to evade detection and prevent removal. In the case of simpler rootkits, it was possible to look for telltale signs – odd changes, missing or alter folders, etc, to determine a rootkit was present. With more sophisticated threats though, detection meant deactivating the rootkit entirely before it could start active evasion; because once it was active, detection and removal became well nigh impossible.

That status quo has changed somewhat in the last few years, as more antivirus vendors have developed the necessary tools to combat the threat. As rootkits themselves vary in complexity, detecting and removing them requires a multi-layered approach:

  • First Line of Defense: Heuristic Scanning
    This preliminary defense can deal with the more obvious rootkits, those that make easy-to-spot changes or ham-fistedly modify normally untouched components. Most antivirus products nowadays include heuristic or behavior-based scanning, which examines each program to evaluate how potentially damaging its actions may be. If the rootkit (or the malware it’s hiding) is found, the AV may be able to find and remove them as usual.
  • Second Line of Defense: Specific Malware Removal
    Even with heuristic scanning, standard scanning engines may not detect more sophisticated or devious rootkits.  At this point human ingenuity enters the picture, in the form of Malware Analysts, who analyse the threat and create specific removal scripts designed to find and remove a particular rootkit. These scripts are also called on to scan the computer, looking for specific threats to complement the more general, automated checks.
  • Third Line of Defense: Offline Scanning
    Sometimes, a rootkit can compromise a computer so thoroughly that any detection program running on the infected system is hopelessly outfoxed by the wily rootkit. In that event, the safest bet is to perform offline scanning – shutting down the computer so that the rootkit can’t actively hide itself, then scanning the system using an antivirus program or rootkit detection tool that runs off a CD or USB drive.
  • Fourth Line of Defense: Manual Removal
    As a last resort, some antivirus vendors will recommend specific manual removal procedures, which only apply for particular rootkits. Generally, this type of removal is considered quite advanced for an average user, and is best left to an IT technician or at least to someone more experienced. Some vendors also develop and publish removal utility programs, either for general or specific rootkit removal.

These detection and removal methods will probably catch most of the rootkits out there, but none of them are 100% certain. In some cases, the fastest, easiest and cheapest possible solution is to simply format and reinstall the entire operating system (assuming of course you have backups of your important files). Determining whether that applies in your case really depends on your personal evaluation of the costs and benefits though, so it’s hard to state any hard and fast rule about this.

Unfortunately, malware authors are ingenious at finding ways to get where they’re not wanted, and the highly complex, multi-layered nature of computing tilts the odds in their favour more than it does to ensuring computer security. Then again, to be fair, humans have lived in houses for thousands of years, and we still haven’t figured out how to totally prevent burglars from invading our homes, so you could probably also credit a natural human genius for finding ways to inconvenience their fellows.

More

If you’re still interested, here are few other articles with more details (some technical, others less so) about rootkits:

Also partially available in Google Books:

  • Rootkits for Dummies By Larry Stevenson, Nancy Altholz
  • The Rootkit Arsenal: Escape and Evasion in the Dark Corners of the System By Bill Blunden
  • Rootkits: Subverting the Windows Kernel By Greg Hoglund, James Butler

More posts from this topic

IMG_3395

5 ways to get ready to ask Mikko anything

It's like a press conference anyone can join from anywhere. And even if you don't have a question, you can upvote the ones you don't like and downvote the ones you do. President Obama did one. Snoop Dogg/Snoop Lion did one. An astronaut did one from outer space. And our Mikko Hypponen will sit down for his second Reddit AMA on December 2 at 9 AM ET. If you have something you've wanted to ask him about online security, great. If not, here are five resources that document some of Mikko's more than two decades in the security industry to prod you or prepare you. 1. Check out this 2004 profile of his work from Vanity Fair. 2. Watch his 3 talks that have been featured on TED.com. [protected-iframe id="7579bbf790267cc081ac7d92d951262c-10874323-9129869" info="https://embed-ssl.ted.com/talks/mikko_hypponen_fighting_viruses_defending_the_net.html" width="640" height="360" frameborder="0" scrolling="no" webkitallowfullscreen="" mozallowfullscreen="" allowfullscreen=""] [protected-iframe id="fdf818f4afa2f7dcb179c5516c44918c-10874323-9129869" info="https://embed-ssl.ted.com/talks/mikko_hypponen_three_types_of_online_attack.html" width="640" height="360" frameborder="0" scrolling="no" webkitallowfullscreen="" mozallowfullscreen="" allowfullscreen=""] [protected-iframe id="54be2fe9bce28ae991becbe3d4291e56-10874323-9129869" info="https://embed-ssl.ted.com/talks/mikko_hypponen_how_the_nsa_betrayed_the_world_s_trust_time_to_act.html" width="640" height="360" frameborder="0" scrolling="no" webkitallowfullscreen="" mozallowfullscreen="" allowfullscreen=""] 3. Check out his first AMA, which took place just after his first talk at TEDglobal was published. 4. Take a trip to Pakistan with Mikko to meet the creators of the first PC virus. [protected-iframe id="8c0605f62076aa901ed165dbd3f4fcd7-10874323-9129869" info="//www.youtube-nocookie.com/v/lnedOWfPKT0?version=3&hl=en_US&rel=0" width="640" height="360"] 5. To get a sense of what he's been thinking about recently, watch his most recent talk at Black Hat "Governments as Malware Creators". [protected-iframe id="54b24406f022e81b15ad6dadf2adfc93-10874323-9129869" info="//www.youtube-nocookie.com/v/txknsq5Z5-8?hl=en_US&version=3&rel=0" width="640" height="360"] BONUS: Make sure you follow him on Twitter to get a constant stream of insight about online security, privacy and classic arcade games. Cheers, Sandra

Nov 14, 2014
bash

Shellshock only concerns server admins – WRONG

Yet another high-profile vulnerability in the headlines, Shellshock. This one could be a big issue. The crap could really hit the fan big time if someone creates a worm that infects servers, and that is possible. But the situation seems to be brighter for us ordinary users. The affected component is the Unix/Linux command shell Bash, which is only used by nerdy admins. It is present in Macs as well, but they seem to be unaffected. Linux-based Android does not use Bash and Windows is a totally different world. So we ordinary users can relax and forget about this one. We are not affected. Right? WRONG! Where is your cloud content stored? What kind of software is used to protect your login and password, credit card number, your mail correspondence, your social media updates and all other personal info you store in web-based systems? Exactly. A significant part of that may be on systems that are vulnerable to Shellshock, and that makes you vulnerable. The best protection against vulnerabilities on your own devices is to make sure the automatic update services are enabled and working. That is like outsourcing the worries to professionals, they will create and distribute fixes when vulnerabilities are found. But what about the servers? You have no way to affect how they are managed, and you don’t even know if the services you use are affected. Is there anything you can do? Yes, but only indirectly. This issue is an excellent reminder of some very basic security principles. We have repeated them over and over, but they deserve to be repeated once again now. You can’t control how your web service providers manage their servers, but you can choose which providers you trust. Prefer services that are managed professionally. Remember that you always can, and should, demand more from services you pay for. Never reuse your password on different services. This will not prevent intrusions, but it will limit the damage when someone breaks into the system. You may still be hurt by a Shellshock-based intrusion even if you do this, but the risk should be small and the damage limited. Anyway, you know you have done your part, and its bad luck if an incident hurts you despite that. Safe surfing, Micke   PS. The best way to evaluate a service provider’s security practices is to see how they deal with security incidents. It tells a lot about their attitude, which is crucial in all security work. An incident is bad, but a swift, accurate and open response is very good.   Addition on September 30th. Contrary to what's stated above, Mac computers seem to be affected and Apple has released a patch. It's of course important to keep your device patched, but this does not really affect the main point of this article. Your cloud content is valuable and part of that may be on vulnerable servers.  

Sep 26, 2014
BY 
APPLE WATCH

Why your Apple Watch will probably never be infected by malware

On Tuesday Apple announced its latest iPhone models and a new piece of wearable technology some have been anxiously waiting for -- Apple Watch. TechRadar describes the latest innovation from Cupertino as "An iOS 8-friendly watch that plays nice with your iPhone." And if it works like your iPhone, you can expect that it will free of all mobile malware threats, unless you decide to "jailbreak" it. The latest F-Secure Labs Threat Report clears up one big misconception about iOS malware: It does exist, barely. In the first half of 2014, 295 new families and variants or mobile malware were discovered – 294 on Android and one on iOS.  iPhone users can face phishing scams and Wi-Fi hijacking, which is why we created our Freedome VPN, but the threat of getting a bad app on your iOS device is almost non-existent. "Unlike Android, malware on iOS have so far only been effective against jailbroken devices, making the jailbreak tools created by various hacker outfits (and which usually work by exploiting undocumented bugs in the platform) of interest to security researchers," the report explains. The iOS threat that was found earlier this year, Unflod Baby Panda, was designed to listen to outgoing SSL connections in order to steal the device’s Apple ID and password details. Apple ID and passwords have been in the news recently as they may have played a role in a series of hacks of celebrity iCloud accounts that led to the posting of dozens of private photos. Our Mikko Hypponen explained in our latest Threat Report Webinar that many users have been using these accounts for years, mostly to purchase items in the iTunes store, without realizing how much data they were actually protecting. But Unflod Baby Panda is very unlikely to have played any role in the celebrity hacks, as "jailbreaking" a device is still very rare. Few users know about the hack that gives up the protection of the "closed garden" approach of the iOS app store, which has been incredibly successful in keeping malware off the platform, especially compared to the more open Android landscape. The official Play store has seen some infiltration by bad apps, adware and spamware -- as has the iOS app store to a far lesser degree -- but the majority of Android threats come from third-party marketplaces, which is why F-Secure Labs recommends you avoid them. The vast majority of iPhone owners have never had to worry about malware -- and if the Apple Watch employs the some tight restrictions on apps, the device will likely be free of security concerns. However, having a watch with the power of a smartphone attached to your body nearly twenty-four hours a day promises to introduce privacy questions few have ever considered.    

Sep 9, 2014
BY