Skip to main content

Improper Restriction of Operations within the Bounds of a Memory Buffer

CVE-2020-15213

Medium
4/10

Summary

In TensorFlow Lite 2.2.x before versions 2.2.1 and 2.3.x before 2.3.1, models using segment sum can trigger a denial of service by causing an out of memory allocation in the implementation of segment sum. Since code uses the last element of the tensor holding them to determine the dimensionality of output tensor, attackers can use a very large value to trigger a large allocation. The issue is patched in commit 204945b19e44b57906c9344c0d00120eeeae178a and is released in TensorFlow versions 2.2.1, or 2.3.1. A potential workaround would be to add a custom `Verifier` to limit the maximum value in the segment ids tensor. This only handles the case when the segment ids are stored statically in the model, but a similar validation could be done if the segment ids are generated at runtime, between inference steps. However, if the segment ids are generated as outputs of a tensor during inference steps, then there are no possible workaround and users are advised to upgrade to patched code.

  • HIGH
  • NETWORK
  • NONE
  • CHANGED
  • NONE
  • NONE
  • NONE
  • LOW

CWE-119 - Buffer Overflow

Buffer overflow attacks involve data transit and operations exceeding the restricted memory buffer, thereby corrupting or overwriting data in adjacent memory locations. Such overflow allows the attacker to run arbitrary code or manipulate the existing code to cause privilege escalation, data breach, denial of service, system crash and even complete system compromise. Given that languages such as C and C++ lack default safeguards against overwriting or accessing data in their memory, applications utilizing these languages are most susceptible to buffer overflows attacks.

References

Advisory Timeline

  • Published