CVE-2021-41221— Access to invalid memory during shape inference in `Cudnn*` ops
Quick assessment
Affected
tensorflow tensorflow
Exploitation
No confirmed in-the-wild exploitation; assess based on exposure
Recommended action
Check the vendor advisory and references for a fixed version. If immediate upgrade is impossible, restrict exposure and increase monitoring.
Google TensorFlow是美国谷歌（Google）公司的一套用于机器学习的端到端开源平台。 Google TensorFlow 存在缓冲区错误漏洞，该漏洞源于 Cudnn* 操作的形状推断代码可以通过堆缓冲区溢出被欺骗访问无效内存。
Possible ATT&CK Techniques 1 AI
T1190 · Exploit Public-Facing Application
Get alerts for future matching vulnerabilities
Log in to subscribe
I. Basic Information for CVE-2021-41221
Vulnerability Information
Have questions about the vulnerability? See if Shenlong's analysis helps!
View Shenlong Deep Dive ↗
Although we use advanced large model technology, its output may still contain inaccurate or outdated information.Shenlong tries to ensure data accuracy, but please verify and judge based on the actual situation.
Vulnerability Title
Access to invalid memory during shape inference in `Cudnn*` ops
Source: CVE Program / CVE List V5
Vulnerability Description
TensorFlow is an open source platform for machine learning. In affected versions the shape inference code for the `Cudnn*` operations in TensorFlow can be tricked into accessing invalid memory, via a heap buffer overflow. This occurs because the ranks of the `input`, `input_h` and `input_c` parameters are not validated, but code assumes they have certain values. 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.
Source: CVE Program / CVE List V5
CVSS Information
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
Source: CVE Program / CVE List V5
Vulnerability Type
未进行输入大小检查的缓冲区拷贝(传统缓冲区溢出)
Source: CVE Program / CVE List V5
Vulnerability Title
Google TensorFlow 缓冲区错误漏洞
Source: CNNVD (China National Vulnerability Database)
Vulnerability Description
Google TensorFlow是美国谷歌（Google）公司的一套用于机器学习的端到端开源平台。 Google TensorFlow 存在缓冲区错误漏洞，该漏洞源于 Cudnn* 操作的形状推断代码可以通过堆缓冲区溢出被欺骗访问无效内存。
Source: CNNVD (China National Vulnerability Database)
CVSS Information
N/A
Source: CNNVD (China National Vulnerability Database)
Vulnerability Type
N/A
Source: CNNVD (China National Vulnerability Database)
Affected Products
| Vendor | Product | Affected Versions | CPE | Subscribe |
|---|---|---|---|---|
| tensorflow | tensorflow | >= 2.6.0, < 2.6.1 | - |
|
II. Public POCs for CVE-2021-41221
| # | POC Description | Source Link | Shenlong Link |
|---|
AI-Generated POC
Premium
No public POC found.
Login to generate AI POCIII. Intelligence Information for CVE-2021-41221
请登录查看更多情报信息。
Patches & Fixes for CVE-2021-41221 (1)
Vendor Advisories for CVE-2021-41221 (1)
-
https://github.com/tensorflow/tensorflow/security/advisories/GHSA-cqv6-3phm-hcwx x_refsource_CONFIRM
Same Patch Batch · tensorflow · 2021-11-05 · 34 CVEs total
| CVE-2021-41208 | 8.8 HIGH | Incomplete validation in boosted trees code |
| CVE-2021-41220 | 7.8 HIGH | Use after free in `CollectiveReduceV2` |
| CVE-2021-41214 | 7.8 HIGH | Reference binding to `nullptr` in `tf.ragged.cross` |
| CVE-2021-41219 | 7.8 HIGH | Undefined behavior via `nullptr` reference binding in sparse matrix multiplication |
| CVE-2021-41201 | 7.8 HIGH | Unitialized access in `EinsumHelper::ParseEquation` |
| CVE-2021-41203 | 7.8 HIGH | Missing validation during checkpoint loading |
| CVE-2021-41228 | 7.5 HIGH | Code injection in `saved_model_cli` |
| CVE-2021-41224 | 7.1 HIGH | `SparseFillEmptyRows` heap OOB read |
| CVE-2021-41223 | 7.1 HIGH | Heap OOB read in `FusedBatchNorm` kernels |
| CVE-2021-41226 | 7.1 HIGH | Heap OOB read in `SparseBinCount` |
| CVE-2021-41212 | 7.1 HIGH | Heap OOB read in `tf.ragged.cross` |
| CVE-2021-41211 | 7.1 HIGH | Heap OOB read in shape inference for `QuantizeV2` |
| CVE-2021-41205 | 7.1 HIGH | Heap OOB read in all `tf.raw_ops.QuantizeAndDequantizeV*` ops |
| CVE-2021-41210 | 7.1 HIGH | Heap OOB read in `tf.raw_ops.SparseCountSparseOutput` |
| CVE-2021-41206 | 7.0 HIGH | Incomplete validation of shapes in multiple TF ops |
| CVE-2021-41227 | 6.6 MEDIUM | Arbitrary memory read in `ImmutableConst` |
| CVE-2021-41200 | 5.5 MEDIUM | Incomplete validation in `tf.summary.create_file_writer` |
| CVE-2021-41197 | 5.5 MEDIUM | Crashes due to overflow and `CHECK`-fail in ops with large tensor shapes |
| CVE-2021-41198 | 5.5 MEDIUM | Overflow/crash in `tf.tile` when tiling tensor is large |
| CVE-2021-41199 | 5.5 MEDIUM | Overflow/crash in `tf.image.resize` when size is large |
Showing top 20 of 34 CVEs. View all on vendor page → →
IV. Related Vulnerabilities
Same product: tensorflow
Same vendor: tensorflow
-
CVE-2026-2492
TensorFlow HDF5 Library Uncontrolled Search Path Element Loc -
CVE-2023-33976
TensorFlow segfault in array_ops.upper_bound -
CVE-2024-3660
Arbitrary code injection vulnerability in Keras framework < -
CVE-2023-25661
Denial of Service in TensorFlow -
CVE-2023-25660
TensorFlow vulnerable to seg fault in `tf.raw_ops.Print`
Same weakness: CWE-120
-
CVE-2026-82895
Multiple vulnerabilities affect the Sniffer component as par -
CVE-2026-84198
Multiple vulnerabilities affect the Sniffer component as par -
CVE-2026-84246
Multiple vulnerabilities affect the Sniffer component as par -
CVE-2026-94587
Brocade Fabric OS堆溢出漏洞影响9.2.2d及10.0.0-10.0.0a1 -
CVE-2026-94582
Brocade Fabric OS 10.0.1前版本FSPF内存溢出漏洞
V. Comments for CVE-2021-41221
No comments yet