Method for dynamically detecting memory overflow on GPU based on address compression technology

Abstract

The invention relates to detection on memory overflow in the field of GPU security, and in particular relates to a method for dynamically detecting memory overflow on a GPU based on an address compression technology. The method disclosed by the invention comprises the steps of: enabling a user program to distribute a memory by using a mallocN function, inserting a canary, compressing an address ina compression address set composed of multiple byte arrays, and marking the memory to be released by using a freeN function; and performing garbage collection or dynamic capacity expansion of the compression address set by utilizing a garbage collection module of a detector; decompressing the address by utilizing a memory overflow detection module of the detector, and performing memory overflow check; and releasing the memory marked by the freeN by utilizing a memory releasing module of the detector. As shown in the figure 1, the system can realize memory overflow detection, which is transparent to the user program and little in performance influence when less memory is consumed.

Description

technical field [0001] The invention relates to memory overflow detection in the GPU security field. Specifically, it relates to a method for dynamically detecting memory overflow on a GPU based on an address compression technology. Background technique [0002] Graphics processing units (GPUs) were originally designed for graphics rendering. Today, thanks to its high concurrency and easy-to-use programming framework, GPUs are widely used in various general computing tasks, such as various graphics computing applications , GPU-assisted accelerated routing, etc. In particular, some applications that require high security, such as AES encryption and database systems, use GPUs for acceleration. With the increasing number of GPU application scenarios, more requirements are placed on the GPU programming framework. On the one hand, as cloud computing becomes more and more popular, the technology of GPU virtualization has also become mature. Major cloud service providers, such ...

AI Technical Summary

Problems solved by technology

[0007] (3) Multiple free memory errors (double free). By releasing the same memory twice, program errors will occur. Attackers can even read and write arbitrary memory by using the system's memory management mechanism.

But it does not provide protection for fine-grained memory allocation by calling malloc

[0013] (3) clARMOR can only detect overflow, but cannot detect underflow and doublefree errors

[0014] Not only are these detection tools on the GPU side insufficient, but the memory overflow detection tools deployed on the CPU side, such as Electric Fence and StackGuard, are difficult to integrate into the current closed system because they need to modify the virtual memory system on the device driver or compiler side. source business environment

Images