Heap out of bounds read in filesystem glob matching in TensorFlow
tensorflow
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI |
|
|
| tensorflow-cpu PyPI |
|
|
| tensorflow-gpu PyPI |
tensorflow
| Product | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI |
|
|
| tensorflow-cpu PyPI |
|
|
| tensorflow-gpu PyPI |
≥ 2.4.0rc0 · < 2.4.0
|
In TensorFlow release candidate versions 2.4.0rc*, the general implementation for matching filesystem paths to globbing pattern is vulnerable to an access out of bounds of the array holding the directories. There are multiple invariants and preconditions that are assumed by the parallel implementation of GetMatchingPaths but are not verified by the PRs introducing it (#40861 and #44310). Thus, we are completely rewriting the implementation to fully specify and validate these. This is patched in version 2.4.0. This issue only impacts master branch and the release candidates for TF version 2.4. The final release of the 2.4 release will be patched.
From the CVE record · GitHub_M
TensorFlow 2.4 release candidates and the master branch have an out-of-bounds array access in filesystem path glob matching. The issue can occur when GetMatchingPaths processes paths; the input does not specify whether an attacker can trigger it over a network, needs authentication, or requires user interaction.
Engineers running PyPI tensorflow, tensorflow-cpu, or tensorflow-gpu versions >=2.4.0rc0 and <2.4.0 should check their installations.