CVE-2021-41195: Crash in `tf.math.segment_*` operations
First published: Fri Nov 05 2021(Updated: )
### Impact The implementation of `tf.math.segment_*` operations results in a `CHECK`-fail related abort (and denial of service) if a segment id in `segment_ids` is large. ```python import tensorflow as tf tf.math.segment_max(data=np.ones((1,10,1)), segment_ids=[1676240524292489355]) tf.math.segment_min(data=np.ones((1,10,1)), segment_ids=[1676240524292489355]) tf.math.segment_mean(data=np.ones((1,10,1)), segment_ids=[1676240524292489355]) tf.math.segment_sum(data=np.ones((1,10,1)), segment_ids=[1676240524292489355]) tf.math.segment_prod(data=np.ones((1,10,1)), segment_ids=[1676240524292489355]) ``` This is similar to [CVE-2021-29584](https://github.com/tensorflow/tensorflow/blob/3a74f0307236fe206b046689c4d76f57c9b74eee/tensorflow/security/advisory/tfsa-2021-071.md) (and similar other reported vulnerabilities in TensorFlow, localized to specific APIs): the [implementation](https://github.com/tensorflow/tensorflow/blob/dae66e518c88de9c11718cd0f8f40a0b666a90a0/tensorflow/core/kernels/segment_reduction_ops_impl.h) (both on CPU and GPU) computes the output shape using [`AddDim`](https://github.com/tensorflow/tensorflow/blob/0b6b491d21d6a4eb5fbab1cca565bc1e94ca9543/tensorflow/core/framework/tensor_shape.cc#L395-L408). However, if the number of elements in the tensor overflows an `int64_t` value, `AddDim` results in a `CHECK` failure which provokes a `std::abort`. Instead, code should use [`AddDimWithStatus`](https://github.com/tensorflow/tensorflow/blob/0b6b491d21d6a4eb5fbab1cca565bc1e94ca9543/tensorflow/core/framework/tensor_shape.cc#L410-L440). ### Patches We have patched the issue in GitHub commit [e9c81c1e1a9cd8dd31f4e83676cab61b60658429](https://github.com/tensorflow/tensorflow/commit/e9c81c1e1a9cd8dd31f4e83676cab61b60658429) (merging [#51733](https://github.com/tensorflow/tensorflow/pull/51733)). The fix will be included in TensorFlow 2.7.0. We will also cherrypick this commit on TensorFlow 2.6.1, TensorFlow 2.5.2, and TensorFlow 2.4.4, as these are also affected and still in supported range. ### 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 [GitHub issue](https://github.com/tensorflow/tensorflow/issues/46888).
Credit: security-advisories@github.com security-advisories@github.com
| Affected Software | Affected Version | How to fix |
|---|---|---|
| pip/tensorflow-gpu | <2.4.4 | 2.4.4 |
| pip/tensorflow-gpu | >=2.5.0<2.5.2 | 2.5.2 |
| pip/tensorflow-gpu | >=2.6.0<2.6.1 | 2.6.1 |
| pip/tensorflow-cpu | <2.4.4 | 2.4.4 |
| pip/tensorflow-cpu | >=2.5.0<2.5.2 | 2.5.2 |
| pip/tensorflow-cpu | >=2.6.0<2.6.1 | 2.6.1 |
| pip/tensorflow | <2.4.4 | 2.4.4 |
| pip/tensorflow | >=2.5.0<2.5.2 | 2.5.2 |
| pip/tensorflow | >=2.6.0<2.6.1 | 2.6.1 |
| TensorFlow Keras | <2.4.4 | |
| TensorFlow Keras | >=2.5.0<2.5.2 | |
| TensorFlow Keras | >=2.6.0<2.6.1 |
Never miss a vulnerability like this again
Sign up to SecAlerts for real-time vulnerability data matched to your software, aggregated from hundreds of sources.
Reference Links
- https://github.com/tensorflow/tensorflow/commit/e9c81c1e1a9cd8dd31f4e83676cab61b60658429
- https://github.com/tensorflow/tensorflow/issues/46888
- https://github.com/tensorflow/tensorflow/pull/51733
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-cq76-mxrc-vchh
- https://nvd.nist.gov/vuln/detail/CVE-2021-41195
- https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow-cpu/PYSEC-2021-844.yaml
- https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow-gpu/PYSEC-2021-846.yaml
- https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow/PYSEC-2021-842.yaml
- https://github.com/advisories/GHSA-cq76-mxrc-vchh (Advisory)
Frequently Asked Questions
What is the severity of CVE-2021-41195?
CVE-2021-41195 has a severity impact that can lead to a denial of service due to a CHECK-fail when large segment ids are used.
How do I fix CVE-2021-41195?
To fix CVE-2021-41195, upgrade TensorFlow to version 2.4.4, 2.5.2, or 2.6.1.
Which versions of TensorFlow are affected by CVE-2021-41195?
CVE-2021-41195 affects TensorFlow versions up to 2.4.4 and versions between 2.5.0 and 2.5.2, as well as versions between 2.6.0 and 2.6.1.
What operations are impacted by CVE-2021-41195?
CVE-2021-41195 impacts the implementation of 'tf.math.segment_*' operations in TensorFlow.
Can CVE-2021-41195 cause application crashes?
Yes, CVE-2021-41195 can lead to application crashes and denial of service due to the CHECK-fail related abort.
- agent/type
- collector/mitre-cve
- source/MITRE
- agent/severity
- agent/weakness
- agent/remedy
- agent/title
- agent/first-publish-date
- collector/github-advisory-latest
- source/GitHub
- alias/GHSA-cq76-mxrc-vchh
- alias/CVE-2021-41195
- agent/software-canonical-lookup
- agent/author
- agent/references
- agent/description
- agent/softwarecombine
- agent/event
- collector/github-advisory
- agent/last-modified-date
- agent/trending
- agent/source
- agent/tags
- collector/nvd-cve
- source/NVD
- agent/software-canonical-lookup-request
- collector/nvd-api
- collector/nvd-index
- package-manager/pip
- vendor/google
- canonical/tensorflow keras
- version/tensorflow keras/2.5.0
- version/tensorflow keras/2.6.0