CVE-2021-37654
CVE-2021-37654 is a high-severity vulnerability in Google Tensorflow with a CVSS 3.x base score of 7.3. It is not currently listed as actively exploited by CISA, and its EPSS exploit-prediction score is low. The underlying weakness is classified as CWE-125.
Key facts
- Severity: High (CVSS 3.x base score 7.3)
- CVSS v2: 3.6
- EPSS exploit prediction: 0% (6th percentile)
- Actively exploited: Not listed in CISA KEV
- Weakness: CWE-125
- Affected product: Google Tensorflow
- Published:
- Last modified:
Description
TensorFlow is an end-to-end open source platform for machine learning. In affected versions 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. 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. We have patched the issue in GitHub 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.
Frequently asked questions
- What is CVE-2021-37654?
- TensorFlow is an end-to-end open source platform for machine learning. In affected versions 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. 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. We have patched the issue in GitHub 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.
- How severe is CVE-2021-37654?
- CVE-2021-37654 has a CVSS 3.x base score of 7.3, rated high severity. It is exploitable over local access with low attack complexity, requires low privileges and no user interaction. Impact on confidentiality is high, integrity low, and availability high.
- Is CVE-2021-37654 being actively exploited?
- It is not currently listed in CISA's KEV catalog. Its EPSS exploit-prediction score is 0% (6th percentile), an estimate of the probability of exploitation in the next 30 days.
- What products are affected by CVE-2021-37654?
- CVE-2021-37654 primarily affects Google Tensorflow. In total, 5 product configurations (CPEs) are listed as vulnerable; see the affected-products list for the exact versions.
- How do I fix CVE-2021-37654?
- Review the linked vendor and NVD advisories for patched versions and mitigations, then upgrade or apply the recommended workaround. Given its high severity, prioritise patching exposed systems.
- When was CVE-2021-37654 published?
- CVE-2021-37654 was published on 2021-08-12 and last updated on 2026-06-17.
References
- https://github.com/tensorflow/tensorflow/commit/bc9c546ce7015c57c2f15c168b3d9201de679a1d
- https://github.com/tensorflow/tensorflow/security/advisories/GHSA-2r8p-fg3c-wcj4
Affected products (5)
- cpe:2.3:a:google:tensorflow:*:*:*:*:*:*:*:*
- cpe:2.3:a:google:tensorflow:2.5.0:*:*:*:*:*:*:*
- cpe:2.3:a:google:tensorflow:2.6.0:rc0:*:*:*:*:*:*
- cpe:2.3:a:google:tensorflow:2.6.0:rc1:*:*:*:*:*:*
- cpe:2.3:a:google:tensorflow:2.6.0:rc2:*:*:*:*:*:*
More vulnerabilities in Google Tensorflow
- CVE-2023-25668 — Critical (CVSS 9.8): TensorFlow is an open source platform for machine learning. Attackers using Tensorflow prior to 2.12.0 or 2.11.1 can…
- CVE-2018-7575 — Critical (CVSS 9.8): Google TensorFlow 1.7.x and earlier is affected by a Buffer Overflow vulnerability. The type of exploitation is…
- CVE-2021-37678 — Critical (CVSS 9.3): TensorFlow is an end-to-end open source platform for machine learning. In affected versions TensorFlow and Keras can be…
- CVE-2021-35958 — Critical (CVSS 9.1): TensorFlow through 2.5.0 allows attackers to overwrite arbitrary files via a crafted archive when…
- CVE-2020-15206 — Critical (CVSS 9.0): In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, changing the TensorFlow's `SavedModel` protocol…
- CVE-2020-15205 — Critical (CVSS 9.0): In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, the `data_splits` argument of…
All CVEs affecting Google Tensorflow →
Other CWE-125 (Out-of-bounds Read) vulnerabilities
- CVE-2026-24826 — Critical (CVSS 10.0): Out-of-bounds Write, Divide By Zero, NULL Pointer Dereference, Use of Uninitialized Resource, Out-of-bounds Read,…
- CVE-2024-22004 — Critical (CVSS 10.0): Due to length check, an attacker with privilege access on a Linux Nonsecure operating system can trigger a…
- CVE-2021-41556 — Critical (CVSS 10.0): sqclass.cpp in Squirrel through 2.2.5 and 3.x through 3.1 allows an out-of-bounds read (in the core interpreter) that…
- CVE-2021-21777 — Critical (CVSS 10.0): An information disclosure vulnerability exists in the Ethernet/IP UDP handler functionality of EIP Stack Group OpENer…
- CVE-2017-14451 — Critical (CVSS 10.0): An exploitable out-of-bounds read vulnerability exists in libevm (Ethereum Virtual Machine) of CPP-Ethereum. A…
- CVE-2013-0767 — Critical (CVSS 10.0): The nsSVGPathElement::GetPathLengthScale function in Mozilla Firefox before 18.0, Firefox ESR 10.x before 10.0.12 and…