Vulnerabilities > CVE-2020-15196 - Out-of-bounds Read vulnerability in Google Tensorflow 2.3.0

047910
CVSS 9.9 - CRITICAL
Attack vector
NETWORK
Attack complexity
LOW
Privileges required
LOW
Confidentiality impact
HIGH
Integrity impact
HIGH
Availability impact
HIGH
network
low complexity
google
CWE-125
critical
NVD
Published: 2020-09-25
Updated: 2024-11-21

Summary

In Tensorflow version 2.3.0, the `SparseCountSparseOutput` and `RaggedCountSparseOutput` implementations don't validate that the `weights` tensor has the same shape as the data. The check exists for `DenseCountSparseOutput`, where both tensors are fully specified. In the sparse and ragged count weights are still accessed in parallel with the data. But, since there is no validation, a user passing fewer weights than the values for the tensors can generate a read from outside the bounds of the heap buffer allocated for the weights. The issue is patched in commit 3cbb917b4714766030b28eba9fb41bb97ce9ee02 and is released in TensorFlow version 2.3.1.

Vulnerable Configurations

Part Description Count
Application
Google
1

Common Weakness Enumeration (CWE)

Common Attack Pattern Enumeration and Classification (CAPEC)

  • Overread Buffers
    An adversary attacks a target by providing input that causes an application to read beyond the boundary of a defined buffer. This typically occurs when a value influencing where to start or stop reading is set to reflect positions outside of the valid memory location of the buffer. This type of attack may result in exposure of sensitive information, a system crash, or arbitrary code execution.

References