The “hits” keep on coming for Android’s mediaserver component. Trend Micro discovered yet another Android mediaserver vulnerability, which can be exploited to perform attacks involving arbitrary code execution. With this new vulnerability, an attacker would be able to run their code with the same permissions that the mediaserver program already has as part of its normal routines.
This vulnerability has been designated as CVE-2015-3842. While it affects Android versions 2.3 to 5.1.1, Google has fixed and published details this vulnerability via the Android Open Source Project (AOSP). Currently, there are no known active attacks against this vulnerability.
This discovery closely follows three other major vulnerabilities in Android’s mediaserver component that surfaced recently. CVE-2015-3823 may allow attackers to trap phones in endless reboots, ANDROID-21296336 may render devices silent, while CVE-2015-3824 (Stagefright), can be used to install malware through a multimedia message.
How it works
The vulnerability involves AudioEffect, a component of the mediaserver program. It uses an unchecked variable which comes from the client, which is usually an app. For an attack to begin, attackers convince the victim to install an app that doesn’t require any required permissions, giving them a false sense of security.
Possible threat scenarios
This attack can be fully controlled, which means a malicious app can decide when to start the attack and also when to stop. An attacker would be able to run their code with the same permissions that mediaserver already has as part of its normal routines. Since the mediaserver component deals with a lot of media-related tasks including taking pictures, reading MP4 files, and recording videos, the privacy of the victim may be at risk. Devices with customized versions of Android but with no modification made to the mediaserver component are also affected.
A dilemma users may face is that it will be difficult for them to locate the cause once an attack occurs. In our demo, we simply triggered the attack by running an app; this is convenient and intuitive for disclosure. While attacks can be triggered by apps alone, real-world attacks won’t involve apps that are easy to detect. The malicious app will try as much as possible to appear legitimate and use dynamic load technology to remain undetected while triggering the attack several days/months later, either persistently or intermittently, similar to other malware.
End users can block this threat from the onset by downloading Trend Micro Mobile Security (TMMS), which can detect threats trying to use this vulnerability and running any of the scenarios presented. Android users can also reboot their device using safe mode to uninstall the malicious app. However, this method might prove difficult, especially for those unaccustomed to tinkering with their devices. We also recommend that device manufacturers patch their devices regularly to prevent their users from suffering from attacks that use this vulnerability.