In eager mode, TensorFlow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1 does not set the session state. Hence, calling `tf.raw_ops.GetSessionHandle` or `tf.raw_ops.GetSessionHandleV2` results in a null pointer dereference In linked snippet, in eager mode, `ctx->session_state()` returns `nullptr`. Since code immediately dereferences this, we get a segmentation fault. The issue is patched in commit 9a133d73ae4b4664d22bd1aa6d654fec13c52ee1, and is released in TensorFlow versions 1.15.4, 2.0.3, 2.1.2, 2.2.1, or 2.3.1.
| Package (Ecosystem) | Introduced | Fixed | Limit |
|---|---|---|---|
| tensorflow(PyPI) | 0 | 1.15.4 | N/A |
| tensorflow(PyPI) | 2.0.0 | 2.0.3 | N/A |
| tensorflow(PyPI) | 2.1.0 | 2.1.2 | N/A |
| tensorflow(PyPI) | 2.2.0 | 2.2.1 | N/A |
| tensorflow(PyPI) | 2.3.0 | 2.3.1 | N/A |
| tensorflow-cpu(PyPI) | 0 | 1.15.4 | N/A |
| tensorflow-cpu(PyPI) | 2.0.0 | 2.0.3 | N/A |
| tensorflow-cpu(PyPI) | 2.1.0 | 2.1.2 | N/A |
| tensorflow-cpu(PyPI) | 2.2.0 | 2.2.1 | N/A |
| tensorflow-cpu(PyPI) | 2.3.0 | 2.3.1 | N/A |
| tensorflow-gpu(PyPI) | 0 | 1.15.4 | N/A |
| tensorflow-gpu(PyPI) | 2.0.0 | 2.0.3 | N/A |
| tensorflow-gpu(PyPI) | 2.1.0 | 2.1.2 | N/A |
| tensorflow-gpu(PyPI) | 2.2.0 | 2.2.1 | N/A |
| tensorflow-gpu(PyPI) | 2.3.0 | 2.3.1 | N/A |
CVSS Metrics
