In Tensorflow before version 2.4.0, when the `boxes` argument of `tf.image.crop_and_resize` has a very large value, the CPU kernel implementation receives it as a C++ `nan` floating point value. Attempting to operate on this is undefined behavior which later produces a segmentation fault. The issue is patched in eccb7ec454e6617738554a255d77f08e60ee0808 and TensorFlow 2.4.0 will be released containing the patch. TensorFlow nightly packages after this commit will also have the issue resolved.
Source: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2020-15266
Reference (s):
- https://github.com/tensorflow/tensorflow/issues/42129
- URL: https://github.com/tensorflow/tensorflow/issues/42129
- https://github.com/tensorflow/tensorflow/pull/42143/commits/3ade2efec2e90c6237de32a19680caaa3ebc2845
- URL: https://github.com/tensorflow/tensorflow/pull/42143/commits/3ade2efec2e90c6237de32a19680caaa3ebc2845
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-xwhf-g6j5-j5gc