In this paper, a new method of hijacking the Linux kernel is presented. It is based on analysing the Linux system call handler, where a proper set of instructions is subsequently replaced by a jump to a different function. The ability to change the execution flow in the middle of an existing function represents a unique approach in Linux kernel hacking. The attack is applicable to all kernels from the 2.6 series on the Intel architecture. Due to this, rootkits based on this kind of technique represent a high risk for Linux administrators.
展开▼
