Segfault on strings tensors with mistmatched dimensions, due to Go code
Impact
Under certain conditions, Go code can trigger a segfault in string deallocation.
For string tensors, C.TF_TString_Dealloc is called during garbage collection within a finalizer function. However, tensor structure isn't checked until encoding to avoid a performance penalty. The current method for dealloc assumes that encoding succeeded, but segfaults when a string tensor is garbage collected whose encoding failed (e.g., due to mismatched dimensions).
To fix this, the call to set the finalizer function is deferred until NewTensor returns and, if encoding failed for a string tensor, deallocs are determined based on bytes written.
Patches
We have patched the issue in GitHub commit [8721ba96e5760c229217b594f6d2ba332beedf22](https://github.com/tensorflow/tensorflow/commit/8721ba96e5760c229217b594f6d2ba332beedf22) (merging [#50508](https://github.com/tensorflow/tensorflow/pull/50508)).
The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, which is the other affected version.
For more information
Please consult [our security guide](https://github.com/tensorflow/tensorflow/blob/master/SECURITY.md) for more information regarding the security model and how to contact us with issues and questions.
Attribution
This vulnerability has been reported externally via a [fixing...
Affected versions
| Package | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI | >= 2.5.0rc0, < 2.5.1 | 2.5.1 |
Details and references
### Impact Under certain conditions, Go code can trigger a segfault in string deallocation. For string tensors, `C.TF_TString_Dealloc` is called during garbage collection within a finalizer function. However, tensor structure isn't checked until encoding to avoid a performance penalty. The current method for dealloc assumes that encoding succeeded, but segfaults when a string tensor is garbage collected whose encoding failed (e.g., due to mismatched dimensions). To fix this, the call to set the finalizer function is deferred until `NewTensor` returns and, if encoding failed for a string tensor, deallocs are determined based on bytes written. ### Patches We have patched the issue in GitHub commit [8721ba96e5760c229217b594f6d2ba332beedf22](https://github.com/tensorflow/tensorflow/commit/8721ba96e5760c229217b594f6d2ba332beedf22) (merging [#50508](https://github.com/tensorflow/tensorflow/pull/50508)). The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, which is the other affected version. ### For more information Please consult [our security guide](https://github.com/tensorflow/tensorflow/blob/master/SECURITY.md) for more information regarding the security model and how to contact us with issues and questions. ### Attribution This vulnerability has been reported externally via a [fixing PR](https://github.com/tensorflow/tensorflow/pull/50508).
- CVSS 3.1
- CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
- Severity from
- GitHub (reviewed advisory)
- Weakness
- CWE-20
- Also known as
- BIT-tensorflow-2021-37692, CVE-2021-37692, PYSEC-2021-314, PYSEC-2021-605, PYSEC-2021-803
- github.com/tensorflow/tensorflow/security/advisories/GHSA-cmgw-8vpc-rc59
- nvd.nist.gov/vuln/detail/CVE-2021-37692
- github.com/tensorflow/tensorflow/pull/50508
- github.com/tensorflow/tensorflow/commit/8721ba96e5760c229217b594f6d2ba332beedf22
- github.com/pypa/advisory-database/tree/main/vulns/tensorflow-cpu/PYSEC-2021-605.yaml
- github.com/pypa/advisory-database/tree/main/vulns/tensorflow-gpu/PYSEC-2021-803.yaml
- github.com/pypa/advisory-database/tree/main/vulns/tensorflow/PYSEC-2021-314.yaml
- github.com/tensorflow/tensorflow