CVE-2022-23585
Memory leak in decoding PNG images in Tensorflow
In short
TensorFlow leaks memory when it tries to decode invalid PNG images. The application fails to free allocated memory in error cases, which could cause memory exhaustion if many malformed images are processed.
Technical detail
A memory leak exists in TensorFlow's PNG decoding function when invalid images trigger error conditions via OP_REQUIRES macros, preventing the cleanup routine png::CommonFreeDecode from executing. This can lead to heap exhaustion when processing untrusted or malformed PNG inputs in applications that repeatedly decode images.
Summary generated and translated by AI from the official description.
Tensorflow is an Open Source Machine Learning Framework. When decoding PNG images TensorFlow can produce a memory leak if the image is invalid. After calling `png::CommonInitDecode(..., &decode)`, the `decode` value contains allocated buffers which can only be freed by calling `png::CommonFreeDecode(&decode)`. However, several error case in the function implementation invoke the `OP_REQUIRES` macro which immediately terminates the execution of the function, without allowing for the memory free to occur. The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L
Affected products
tensorflow · tensorflowWant to know if your infrastructure is exposed to this?
Talk to TrueHacking →References
https://github.com/tensorflow/tensorflow/blob/a1320ec1eac186da1d03f033109191f715b2b130/tensorflow/core/kernels/image/decode_image_op.cc#L322-L416https://github.com/tensorflow/tensorflow/commit/ab51e5b813573dc9f51efa335aebcf2994125ee9https://github.com/tensorflow/tensorflow/security/advisories/GHSA-fq6p-6334-8gr4