CVE-2021-37654: Heap OOB and CHECK fail in `ResourceGather` in TensorFlow
First published: Thu Aug 12 2021(Updated: )
### Impact An attacker can trigger a crash via a `CHECK`-fail in debug builds of TensorFlow using `tf.raw_ops.ResourceGather` or a read from outside the bounds of heap allocated data in the same API in a release build: ```python import tensorflow as tf tensor = tf.constant(value=[[1,2],[3,4],[5,6]],shape=(3,2),dtype=tf.uint32) v = tf.Variable(tensor) tf.raw_ops.ResourceGather( resource=v.handle, indices=[0], dtype=tf.uint32, batch_dims=10, validate_indices=False) ``` The [implementation](https://github.com/tensorflow/tensorflow/blob/f24faa153ad31a4b51578f8181d3aaab77a1ddeb/tensorflow/core/kernels/resource_variable_ops.cc#L660-L668) does not check that the `batch_dims` value that the user supplies is less than the rank of the input tensor. Since the implementation uses several for loops over the dimensions of `tensor`, this results in reading data from outside the bounds of heap allocated buffer backing the tensor: ```cc // batch_dims_ = > params.dims() (10 > 2) for (int i = 0; i < batch_dims_; ++i) { result_shape.AddDim(params.dim_size(i)); } for (int i = batch_dims_; i < indices.dims(); ++i) { result_shape.AddDim(indices.dim_size(i)); } for (int i = batch_dims_ + 1; i < params.dims(); ++i) { result_shape.AddDim(params.dim_size(i)); } ``` In debug mode, `.dim_size(i)` validates that the argument is less than `.dims()` using a `DCHECK`. But the `DCHECK` is a no-op in release builds. ### Patches We have patched the issue in GitHub commit [bc9c546ce7015c57c2f15c168b3d9201de679a1d](https://github.com/tensorflow/tensorflow/commit/bc9c546ce7015c57c2f15c168b3d9201de679a1d). The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, 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 by members of the Aivul Team from Qihoo 360.
Credit: security-advisories@github.com security-advisories@github.com
| Affected Software | Affected Version | How to fix |
|---|---|---|
| pip/tensorflow-gpu | =2.5.0 | 2.5.1 |
| pip/tensorflow-gpu | >=2.4.0<2.4.3 | 2.4.3 |
| pip/tensorflow-gpu | <2.3.4 | 2.3.4 |
| pip/tensorflow-cpu | =2.5.0 | 2.5.1 |
| pip/tensorflow-cpu | >=2.4.0<2.4.3 | 2.4.3 |
| pip/tensorflow-cpu | <2.3.4 | 2.3.4 |
| pip/tensorflow | =2.5.0 | 2.5.1 |
| pip/tensorflow | >=2.4.0<2.4.3 | 2.4.3 |
| pip/tensorflow | <2.3.4 | 2.3.4 |
| TensorFlow Keras | >=2.3.0<2.3.4 | |
| TensorFlow Keras | >=2.4.0<2.4.3 | |
| TensorFlow Keras | =2.5.0 | |
| TensorFlow Keras | =2.6.0-rc0 | |
| TensorFlow Keras | =2.6.0-rc1 | |
| TensorFlow Keras | =2.6.0-rc2 |
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/bc9c546ce7015c57c2f15c168b3d9201de679a1d
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-2r8p-fg3c-wcj4
- https://nvd.nist.gov/vuln/detail/CVE-2021-37654
- https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow-cpu/PYSEC-2021-567.yaml
- https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow-gpu/PYSEC-2021-765.yaml
- https://github.com/pypa/advisory-database/tree/main/vulns/tensorflow/PYSEC-2021-276.yaml
- https://github.com/advisories/GHSA-2r8p-fg3c-wcj4 (Advisory)
Frequently Asked Questions
What is the severity of CVE-2021-37654?
CVE-2021-37654 is classified as a high severity vulnerability due to its potential to cause crashes and read from outside allocated memory.
How do I fix CVE-2021-37654?
To fix CVE-2021-37654, upgrade to TensorFlow versions 2.3.4, 2.4.3, or 2.5.1, depending on your current version.
What software is affected by CVE-2021-37654?
CVE-2021-37654 affects TensorFlow versions between 2.3.0 and 2.4.3 as well as specific release candidates like 2.6.0-rc0, rc1, and rc2.
What are the potential consequences of CVE-2021-37654?
Exploiting CVE-2021-37654 can lead to application crashes and might allow attackers to read sensitive data from memory.
Is CVE-2021-37654 a remote exploit?
CVE-2021-37654 can potentially be exploited remotely if the vulnerable TensorFlow API is exposed in a web application.
- 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-2r8p-fg3c-wcj4
- alias/CVE-2021-37654
- agent/software-canonical-lookup
- agent/last-modified-date
- agent/references
- agent/description
- agent/event
- agent/softwarecombine
- agent/author
- collector/github-advisory
- agent/trending
- agent/source
- agent/tags
- collector/nvd-cve
- source/NVD
- agent/software-canonical-lookup-request
- collector/nvd-index
- package-manager/pip
- vendor/google
- canonical/tensorflow keras
- version/tensorflow keras/2.3.0
- version/tensorflow keras/2.4.0
- version/tensorflow keras/2.5.0
- version/tensorflow keras/2.6.0-rc0
- version/tensorflow keras/2.6.0-rc1
- version/tensorflow keras/2.6.0-rc2