
Cloud Vulnerability DB
A community-led vulnerabilities database
CVE-2021-29614 is a critical vulnerability in TensorFlow's tf.io.decode_raw implementation that affects versions prior to 2.5.0. The vulnerability was discovered in early 2021 and fixed in TensorFlow versions 2.1.4, 2.2.3, 2.3.3, and 2.4.2. The issue affects the core functionality of TensorFlow's data decoding operations (GitHub Advisory).
The vulnerability stems from a pointer arithmetic error in the implementation of tf.io.decode_raw when combining fixed_length and wider datatypes. The bug occurs due to incorrect pointer advancement, where out_data pointer is moved by fixed_length * sizeof(T) bytes while only copying at most fixed_length bytes from the input. This mismatch in pointer arithmetic leads to memory corruption and potential out-of-bounds writes (GitHub Advisory).
The vulnerability can result in incorrect data decoding, memory corruption, and Python interpreter crashes. More severely, it can be exploited to perform out-of-bounds writes to periodically placed locations in memory, potentially leading to arbitrary code execution (GitHub Advisory).
The issue has been patched in GitHub commit 698e01511f62a3c185754db78ebce0eee1f0184f and included in TensorFlow 2.5.0. Users should upgrade to TensorFlow versions 2.1.4, 2.2.3, 2.3.3, 2.4.2, or later to receive the fix (GitHub Advisory).
Source: This report was generated using AI
Free Vulnerability Assessment
Evaluate your cloud security practices across 9 security domains to benchmark your risk level and identify gaps in your defenses.
Get a personalized demo
"Best User Experience I have ever seen, provides full visibility to cloud workloads."
"Wiz provides a single pane of glass to see what is going on in our cloud environments."
"We know that if Wiz identifies something as critical, it actually is."