Researchers Propose Software Mitigations for Rowhammer Attacks

A team of researchers has proposed two software-based methods that could be used to mitigate Rowhammer, a type of attack that exploits weaknesses in the design of dynamic random-access memory (DRAM).

Rowhammer attacks are possible due to increasing DRAM density, which has led to memory cells being physically smaller and closer together. If a row is accessed repeatedly, it causes bit flips in adjacent memory rows.

The first working privilege escalation exploits leveraging Rowhammer were disclosed by Google researchers in March 2015. Experts later created a JavaScript implementation, and they recently demonstrated that the attack can even be used to root some Android devices.

Rowhammer attacks are not easy to mitigate, particularly using software. The most efficient mitigation involves redesigning DRAM modules. However, researchers from the Technische Universität Darmstadt and the University of Duisburg-Essen in Germany have now come up with what they call “practical and generic software-only defenses.”

The first method, dubbed B-CATT, doesn’t require any changes to the OS and it can be used on all x86 systems. B-CATT is a bootloader extension that locates and disables vulnerable physical memory. Vulnerable memory addresses are identified using existing Rowhammer exploitation tools. Since OSs are designed to handle unavailable memory regions, B-CATT should not break any system functionality, researchers said.

The second mitigation, G-CATT (Generic-CATT), aims to prevent bit flips from affecting memory locations belonging to high-privileged security domains, such as the kernel and co-located virtual machines. It does this by ensuring that memory between the row controlled by the attacker and the row storing the targeted data are separated by at least one row.

“Our extensive evaluation shows that both mitigation schemes can stop available real-world Rowhammer attacks, impose virtually no run-time overhead for common user and kernel benchmarks as well as commonly used applications, and do not affect the stability of the overall system,” the researchers said.

Experts comment on the B-CATT and G-CATT mitigations

At the recent 2016 Black Hat Europe conference, Anders Fogh, principal malware analyst at G DATA Advanced Analytics, and Michael Schwarz, a Ph.D. student at the Graz University of Technology in Austria, detailed a new DRAM attack and demonstrated an improvement to Rowhammer.

Contacted by SecurityWeek, both Schwartz and Fogh agreed that the paper describing software defenses for Rowhammer attacks is an important contribution, but they believe there are several issues that still need to be addressed.

“I haven't read the paper in detail yet, but as far as I've seen it, it does not break Rowhammer completely, it only prevents the strong double-sided Rowhammer,” Schwartz said. “Furthermore, an attack from one application to another is also not prevented, only from user applications to the kernel.”

Schwartz believes that while these mitigations make it more difficult to mount successful Rowhammer attacks, such attacks are still possible.

Fogh shared a more detailed analysis of the paper, but pointed out that some of his comments are speculative given the short timeframe.

The researcher believes the B-CATT idea is “fairly good,” but there are some weak points, including that the prevalence of bit flips is underestimated. Fogh said the following about this method:

1) Might black list too much memory

a. Kim et al. mention the idea and concluded that it’s likely to be prohibitively expensive in terms of blacklisted memory – though the data they present does not allow strong conclusions. I think Kim et al. is close to the truth if we want to protect the entire population of DRAM, but for moderately vulnerable DRAM this solution could make a serious dent in the attack surface;

b. The authors use a sample of 3 systems for their evaluation, which has a large variance of 133 vs. 23 affected page frames. Consequently, their result is unlikely to say much about the “population of DRAM”;

c. I’m not aware of other sources on the number of affected page frames, but we do have a solid indication of a large variance. For instance, Kim et al. reported upwards of 105 errors in 109 cells on some DIMMs. That’s more than 6 million bit flips in an 8Gb DIMM, which should be significantly more than 133 page frames. Another data point is that the row hammer patient 0 (the first computer where an exploit was running) is lovingly called “Flippy the laptop” by Thomas Dullien;

d. Bhattacharya et al. actually does target row hammering – that is on a known, fixed location. That speaks for that vulnerable page frames is significantly more common that the 0,0something% that the paper suggests.

2) The methodology of finding bit flips

a. The authors mention “On the other hand, it’s not yet known whether memory vulnerable to bit flip change over time or under different experiment conditions (e.g. temperature, methodology to identify bit flips)”;

b. Unfortunately, Kim et al. concludes that temperature does matter (though not much);

c. Mark Lanteigne shows that data patterns matter a lot and coins the term “killer data patterns” and speculates “We believe that there is likely a killer data pattern specific and unique to every system, but that discovering a single pattern to find more issues on”;

d. Mark Lanteigne shows that bit flips are best found by local hammering as opposed to the global approach by Seaborn and Dullien, which this paper adopts. Also, Mark Lanteigne shows that multi-core hamming outperforms the single-core hammering of Seaborn and Dullien;

e. There are good reasons to think that DRAM ages and DRAM can become vulnerable with age. And potentially even be made to age artificially. I speculated on the subject in October last year. I did some experiments on the subject with inconclusive results, however Al-Ars provides some data from DRAM from 2005. The author notices this and says that regular rescans would be required – certainly a significant overhead and associated downtime on a system. With an evil race condition between updating and getting pwned.

Section one essentially adds up to “We need more research and despite the result in this paper, in my opinion, we are likely to end up with unsatisfying results for some DIMMs”. Section two is not an argument that the B-CATT is broken, rather an argument that the issue is more complex than it first looks. Again, a call for more research seems prudent, but there is also this nagging feeling that building a truly safe solution across systems and use cases may not be possible. Essentially, even with B-CATT running, we may still run into situations where bit flips are possible.

As for G-CATT, Fogh pointed out that the idea is not new – it was described last year at the Chaos Communication Congress by Daniel Gruss and Clementiné Maurice. While the B-CATT mitigation is independent of the work presented by Fogh and Schwartz at Black Hat Europe, G-CATT “draws heavily” on their research. According to Fogh,

the method will be really effective, but it comes at a price. The memory allocation process loses a degree of freedom and becomes more rigid. With just a few security domains, say a single-user computer with just kernel protected from user mode, this is probably not going to be a big factor. It gets much much worse in say a 64-user cloud computer, which each has a kernel and it’s consequently 128 different security domains. Especially if physical memory requirements of the clients change often, this will become a problem.

The model they suggest for a per process protection would require blocking no less than 12 pages to allocate a single page on skylake, when it cannot be allocate next to existing allocations. That’s a fairly big overhead. Allocating next to existing allocations however will come at a performance penalty through lacking bank parallelism for that process.

They present no test data on their per process model. There is also the problem of sandboxes within processes, such as the chrome NaCl sandbox. Here the very code you are protecting in a security domain is trying to modify itself to break out – that seems like a particularly nasty scenario. I’m skeptical that such a problem can easily be solved at a reasonable overhead even with a custom “malloc” function.

in

50 Million Potentially Impacted by AirDroid Vulnerabilities

Vulnerabilities in the Android remote management tool AirDroid potentially impact over 50 million devices, security researchers at Zimperium zLabs warn.

The application has seen between 10 and 50 million downloads through the official Google Play software portal, but the security firm says that its device base is larger than that. According to Zimperium (the security firm that discovered the Stagefright flaw in Android), vulnerabilities in AirDroid allow an attacker to exploit built-in features and use them against the application’s users.

The issue, the security researchers say, is that AirDroid uses insecure communication channels, which means that the application’s millions of users are exposed to Man-in-the-Middle (MitM) attacks and other types of nefarious acts. Information leakage is also possible, along with remote hijacking of update APKs, which could result in remote code execution.

While analyzing AirDroid, the security researchers discovered that the communication channels employed to send authentication data to the statistics server are insecure. While the requests are encrypted with the Data Encryption Standard (DES) symmetric-key block cipher in Electronic Codebook (ECB) mode, the encryption key is hardcoded inside the application, meaning that the attacker knows it.

Armed with these details, an actor on the same network with the target device could execute MitM attacks to grab authentication credentials from the very first HTTP request the application performs, and can then impersonate the user for further requests, Zimperium’s Simone Margaritelli explains.

“This HTTP request can be decrypted at runtime using the 890jklms key hardcoded inside the application and the authentication fields parsed from the resulting JSON. Having this information, the attacker can now impersonate the victim’s device and perform various HTTP or HTTPS requests on its behalf to the AirDroid API endpoints,” the researcher notes.

An attacker could craft a payload encrypted in DES with the same exact key to trick the server into spewing user information, which will result in the email and password hash being exposed.

What’s more, an attacker performing MitM could also redirect HTTP traffic to a malicious transparent proxy, thus being able to modify the response for the /phone/vncupgraderequest, which the application normally uses to check for addons updates. Thus, by injecting a fake update, the actor could execute malicious code remotely on the compromised device, because AirDroid notifies the user about an update, downloads the RCE.apk package, and might even prompt the user to install it.

Zimperium notes that most of the AirDroid’s functionalities are carried out using secure HTTPS API endpoints, but insecure channels are used for specific tasks, such as the sending of statistics to a remote server. For security, the app relies on DES encrypted JSON payloads for this type of communication, but the use of a hardcoded key (determined to be 890jklms) nullifies the security measure.

To resolve these issues, researchers say, the application should use only secure communication channels (HTTPS), should verify the remote public key (key pinning) to avoid SSL MitM, should use safe key exchange mechanisms, and should leverage and verify digital signatures for update files.

The company also explains that these issues were discovered in May this year, when the developer was notified. However, although two updates were released this week in the form of AirDroid 4.0.0 and AirDroid 4.0.1, the application continues to be vulnerable, putting tens of millions of users at risk.

To stay protected, Zimperium says, users are advised to uninstall or disable AirDroid until a fix is available. They should also install and maintain an anti-malware solution on their Android devices.

FBI, GCHQ Get Foreign Hacking Authority

Changes to Rule 41 of the federal rules of criminal procedure come into force today, giving the FBI (with a judicially granted search warrant) authority to hack computers in any jurisdiction, and potentially overseas. This happened just two days after the UK's Investigatory Powers Act (IPA) was granted royal assent and became law. The latter gives Britain's Government Communications Headquarters (GCHQ) the legal authority to 'mass hack' outside of the UK.

Democratic Senator Ron Wyden repeatedly tried to delay the changes to Rule 41. His last attempt in the Senate failed on Tuesday. Speaking from the floor he described the changes as "one of the biggest mistakes in surveillance policy in years," adding that the government would have "unprecedented authority to hack into Americans' personal phones, computers and other devices."

A major concern for Wyden is that these changes effectively came via the backdoor without requiring congressional approval. It evolved from a regular review of criminal procedure conducted by a conference of federal judges. After several years considering the rule with a public comment period, the conference submitted the suggested change to the Supreme Court -- which approved it to go into effect today.

The changes to the Rule 41 effectively expands the US government's hacking powers, allowing even magistrate judges to grant investigative agencies with the authority to conduct mass hacking operations on computers potentially located in any part of the globe. Previously, Rule 41 laws restricted such authorizations, with magistrate judges only authorized to issue warrants within the jurisdiction of their court.

Although this is a major expansion of FBI authority, it is merely an expansion of existing authority. This is not the case with the UK's Investigatory Powers Act. It has been known since Snowden's revelations that GCHQ hacks into computers; but it had been doing so illegally.

"The investigatory powers tribunal," reported the Guardian in October, "which is the only court that hears complaints against MI5, MI6 and GCHQ, said the security services operated an illegal regime to collect vast amounts of communications data, tracking individual phone and web use and other confidential personal information, without adequate safeguards or supervision for 17 years."

The new IPA (PDF) now makes this legal. It does not use the term 'hacking' but describes it as 'equipment interference'. It requires judicial oversight with a warrant from both the Secretary of State and a panel of judges. There are two categories: targeted and bulk. Bulk hacking is only authorized for foreign targets, but could involve hacking whole towns or even taking down whole geographic areas if perceived as a threat.

Targeted local hacking is also a bit of a misnomer. The Act contains the concept of a 'general warrant' where the target could be a group or an organization or a location rather than an individual even within the UK. If GCHQ believed a terrorist threat was imminent in a particular town, it could obtain a warrant that could effectively cover everyone in that town.

Hacking, however, is not the only new or expanded surveillance capability provided by the IPA. Two that are causing particular concern are the retention of everybody's internet data by ISPs for 12 months, and the provision for what amounts to a general encryption backdoor.

Critics point to recent ISP breaches, including Three Mobile and TalkTalk. With so much more personal data being held, the ISPs will become prime targets for cyber criminals and even foreign states -- and the suggestion is that some of these ISPs will undoubtedly and inevitably be breached.

The encryption backdoor is fittingly backdoored into the legislation. Tucked away in Part 9 under the title 'Miscellaneous and General Provisions' subtitled 'Technical capability notices' is the statement, "The Secretary of State may give a relevant operator a technical capability notice" where the obligations include "the removal by a relevant operator of electronic protection applied by or on behalf of that operator to any communications or data..."

Any such notice would be given with an accompanying gagging order. What this would mean in practice is that an FBI/Apple San Bernardino tussle would not happen in the UK. In theory, the UK government could simply legally require the device operator to provide assistance in gaining access to an encrypted device, while simultaneously preventing that operator from disclosing the fact.

"In general," Sophos senior security advisor John Shier told SecurityWeek, "governments are trying to formalize in law mass surveillance and hacking in an attempt to better protect states and citizens. Some believe this is fine since it 'might' help catch hackers, terrorists and other criminals. Others," he added, "believe this is gross government overreach. Political (and personal) ideology will determine which side is determined to be 'right.' Our concern is more practical than political. This storage of personal data and opening of special access through backdoors only gives the massive cybercrime industry more opportunity to steal it, and places an increased burden on those companies required to collect the data to protect it. High profile data leaks occur all too often, so why put more data at risk?"

Global Police Smash Huge Online Crime Network: Europol

In one of the biggest takedowns to date, police across the globe have smashed a massive criminal network providing online services including malware attacks that infected half a million computers worldwide, Europol said Thursday.

Known as "Avalanche", the criminal network "was used as a delivery platform to launch and manage mass global malware attacks and money mule recruiting campaigns," Europe's policing agency said in a statement.

The network would be contacted by other criminal groups to send emails to specific victims containing malware to steal bank details and passwords as well as to conduct so-called distributed denial of service (DDos) spam attacks.

"Victims of the malware infections were identified in more than 180 countries," it said from its fortress-like headquarters in The Hague.

The damage caused ran into "hundreds of millions of euros" and in Germany alone -- where the gang's activities were first detected -- some six million euros ($6.3 million) in losses were reported as a result of "concentrated cyberattacks on online banking systems."

Police from as far afield as Australia, Belize, India and Singapore, together with 25 other countries pounced on Wednesday after a four-year operation, arresting five top network bosses, seizing 37 servers and searching 39 premises. Europol would not say in which countries the five were held.

"Avalanche functioned like a company -- and we have arrested its 'chief executive' and 'board members'," said Fernando Ruiz, operations chief at Europol's EC3 Cybercrime Centre.

"They had different branches that catered for different crimes -- and even had a customer service section to help their clients," Ruiz told AFP.

Law officials also shut down 221 other servers and seized or blocked some 800,000 internet domains in the "unprecedented" operation.

The operation was supported by Europol, the continent's judicial cooperation agency Eurojust, Interpol, the US Attorney's Office, the FBI and private companies.

The suspects face charges in different countries all related to being involved in organized crime.

Eurojust president Michele Coninsx said the operation "marked a significant moment in the fight against organized cybercrime."

"Avalanche, one of the world's largest and most malicious botnet infrastructures has been decisively neutralized in one of the biggest takedowns to date," she said.

Bug Allows Activation Lock Bypass on iPhone, iPad

Researchers have found a bug that can be used to bypass Apple’s Activation Lock feature and gain access to the homescreen of locked iPhones and iPads running the latest version of iOS.

There seem to be at least two variations of the vulnerability – one of them works on iOS 10.1 and the second has also been reproduced on the latest 10.1.1 version of Apple’s mobile operating system.

The Find My iPhone service allows users to activate Lost Mode on their iPhone, iPad or iPod if the device is lost or stolen. When Lost Mode is used, it automatically enables Activation Lock, a feature that prevents reactivation of the device without the legitimate owner’s permission.

When a locked device is started, users are prompted to connect to a Wi-Fi network. If the “Other Network” option is selected, the user must enter the name of the network and choose a security protocol (e.g. WEP, WPA2, etc.). Depending on which protocol is selected, a username and/or a password must also be entered.

The problem, as researchers discovered, is that there is no limitation on the number of characters that can be entered into the name, username and password fields. By entering very long strings into these fields, an attacker can trigger a crash that exposes the device’s homescreen.

The crash can be triggered using at least two methods. The first involves Apple’s iPad smart cases, which cause the device to wake or sleep when the case is opened or closed.

The flaw was first discovered by India-based security enthusiast Hemant Joseph, who started analyzing the Activation Lock feature after purchasing a locked iPad from eBay. He managed to trigger the crash that gave him access to the homescreen by closing and opening the smart case.

This method worked on iOS 10.1, but the bug was addressed in iOS 10.1.1, which Apple released on October 31.

Researchers at Vulnerability Lab also analyzed the issue and discovered that it can be reproduced on iOS 10.1.1 using the screen rotation feature and Night Shift mode. In the video published by Vulnerability Lab, the homescreen is only visible for a second, but the company’s founder, Benjamin Kunz-Mejri, told SecurityWeek that access can be maintained by quickly pressing the power button.

The videos published by Vulnerability Lab and Joseph show how the bug is triggered on an iPad, but Kunz-Mejri said the bug affects iPhones as well.

"Gooligan" Android Malware Steals Authentication Tokens to Hack User Accounts

"Gooligan" Android Malware Steals Authentication Tokens to Compromise More Than 1 Million Google User Accounts

Researchers from Check Point Software Technologies shared details on Wednesday of new Android malware that has compromised more than a million Google Accounts.

Dubbed Gooligan by the security firm, the malware targets devices running Android 4 and 5, which represent nearly 74 percent of Android devices currently in use.

According to Check Point, the mobile malware can steal authentication tokens stored on devices which can be used to access sensitive data from Gmail, Google Photos, Google Docs and other services, including G Suite.

Check Point’s research team originally discovered Gooligan's code in a malicious app called SnapPea last year. They discovered a new variant in August 2016 which they say is infecting 13,000 Android devices per day, with approximately 57 percent of infected devices located in Asia and about nine percent in Europe.

"The infection begins when a user downloads and installs a Gooligan-infected app on a vulnerable Android device, or by clicking on malicious links in phishing attack messages," Check Point explained in a blog post.

After gaining control over the Android device, the cybercriminals behind Gooligan make money by fraudulently installing apps from Google Play and rating them on behalf of the victim, Check Point said. Gooligan installs at least 30,000 apps daily on compromised devices, totaling more than 2 million apps since the campaign first kicked off.

“If your account has been breached, a clean installation of an operating system on your mobile device is required. This complex process is called flashing, and we recommend powering off your device, and approaching a certified technician or your mobile service provider, to re-flash your device,” said Michael Shaulov, Check Point's head of mobile products.

Check Point has made available a free online tool that allows users to check if their account has been breached by Gooligan. 

In related Android security news, Palo Alto Networks shared details on a recently discovered Android Trojan dubbed “PluginPhantom” that abuses a legitimate plugin framework to update itself and evade static detection. According to to the network security firm, PluginPhantom focuses on data theft and is capable of stealing files, contacts, location data and Wi-Fi information, while also being able to take photos, capture screenshots, intercept and send SMS messages, record audio and log keystrokes.

"PluginPhantom" Android Trojan Uses Plugins to Evade Detection

A recently discovered Android Trojan dubbed “PluginPhantom” abuses a legitimate plugin framework to update itself and evade static detection, Palo Alto Networks reported on Wednesday.

PluginPhantom focuses on data theft and it’s capable of stealing files, contacts, location data and Wi-Fi information. The threat can also take photos, capture screenshots, intercept and send SMS messages, record audio and log keystrokes.

PluginPhantom is believed to be a successor of Android.Trojan.Ihide, a piece of malware analyzed by TrustLock in July. However, unlike other Android Trojans and its predecessor, PluginPhantom is based on a type of design architecture where the malicious application is divided into a main host app and multiple plugins.

To achieve this, it uses DroidPlugin, a plugin framework developed by Chinese security firm Qihoo 360. DroidPlugin enables a host application to run multiple plugins without the need to install them. This has allowed PluginPhantom developers to implement various malicious functions in different plugins, which can be loaded and launched by the host app.

PluginPhantom has nine plugins embedded in the host app as asset files. These include three core plugins, designed for command and control (C&C) server communications and updates, and six plugins focusing on data theft.

In addition to the functionality provided by the plugins, the main app includes keylogging capabilities obtained by abusing Accessibility features.

Palo Alto Networks researchers told SecurityWeek that they have no reason to believe the malware has made it onto Google Play, but they don’t have any information on how the threat has been distributed.

There is also no information on who might be targeted, but the security firm pointed out that the location data collected by the malware is translated to coordinate systems used by Baidu Maps and Amap Maps, navigation apps that are highly popular in China.

By hiding the malicious functionality in plugins, the Android Trojan increases its chances of evading static detection mechanisms. PluginPhantom appears to be the first Android Trojan to leverage this method, but experts believe other malware developers may start using it and it’s possible that it will end up replacing the widely used repackaging techniques.

“Since the plugin development pattern is generic and the plugin SDK can be easily embedded, the plugin architecture could be a trend among Android malware in the future,” Palo Alto Networks researchers said.

In related news, researchers from Check Point Software Technologies shared details on Wednesday of new Android malware that has compromised more than a million Google Accounts. Dubbed Gooligan by the security firm, the malware targets devices running Android 4 and 5 and can steal authentication tokens stored on devices which can be used to access sensitive data from Gmail, Google Photos, Google Docs and other services, including G Suite.