CVE-2021-37679— Heap OOB in nested `tf.map_fn` with `RaggedTensor`s in TensorFlow
Get alerts for future matching vulnerabilitiesLog in to subscribe
I. Basic Information for CVE-2021-37679
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View Shenlong Deep Dive ↗ Although we use advanced large model technology, its output may still contain inaccurate or outdated information.Shenlong tries to ensure data accuracy, but please verify and judge based on the actual situation.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
Heap OOB in nested `tf.map_fn` with `RaggedTensor`s in TensorFlow
Source: NVD (National Vulnerability Database)
Vulnerability Description
TensorFlow is an end-to-end open source platform for machine learning. In affected versions it is possible to nest a `tf.map_fn` within another `tf.map_fn` call. However, if the input tensor is a `RaggedTensor` and there is no function signature provided, code assumes the output is a fully specified tensor and fills output buffer with uninitialized contents from the heap. The `t` and `z` outputs should be identical, however this is not the case. The last row of `t` contains data from the heap which can be used to leak other memory information. The bug lies in the conversion from a `Variant` tensor to a `RaggedTensor`. The [implementation](https://github.com/tensorflow/tensorflow/blob/460e000de3a83278fb00b61a16d161b1964f15f4/tensorflow/core/kernels/ragged_tensor_from_variant_op.cc#L177-L190) does not check that all inner shapes match and this results in the additional dimensions. The same implementation can result in data loss, if input tensor is tweaked. We have patched the issue in GitHub commit 4e2565483d0ffcadc719bd44893fb7f609bb5f12. 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.
Source: NVD (National Vulnerability Database)
CVSS Information
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:N
Source: NVD (National Vulnerability Database)
Vulnerability Type
跨界内存读
Source: NVD (National Vulnerability Database)
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 中存在缓冲区错误漏洞，该漏洞源于产品可能会泄露内存信息。 以下产品及版本收到影响：TensorFlow 2.5.1、TensorFlow 2.4.3 和 TensorFlow 2.3.4。
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 | >= 2.5.0, < 2.5.1 | - |
II. Public POCs for CVE-2021-37679
| # | POC Description | Source Link | Shenlong Link |
|---|
AI-Generated POCPremium
No public POC found.
Login to generate AI POCIII. Intelligence Information for CVE-2021-37679
请登录查看更多情报信息。
Same Patch Batch · tensorflow · 2021-08-12 · 58 CVEs total
| CVE-2021-37678 | 9.3 CRITICAL | Arbitrary code execution due to YAML deserialization |
| CVE-2021-37639 | 8.4 HIGH | Null pointer dereference and heap OOB read in TensorFlow |
| CVE-2021-37689 | 7.8 HIGH | Null pointer dereference in TensorFlow Lite MLIR optimizations |
| CVE-2021-37665 | 7.8 HIGH | Incomplete validation in MKL requantization in TensorFlow |
| CVE-2021-37663 | 7.8 HIGH | Incomplete validation in `QuantizeV2` in TensorFlow |
| CVE-2021-37681 | 7.8 HIGH | Null pointer exception in TensorFlow Lite |
| CVE-2021-37648 | 7.8 HIGH | Incorrect validation of `SaveV2` inputs in TensorFlow |
| CVE-2021-37688 | 7.8 HIGH | Null pointer dereference in TensorFlow Lite |
| CVE-2021-37676 | 7.8 HIGH | Reference binding to nullptr in shape inference in TensorFlow |
| CVE-2021-37671 | 7.8 HIGH | Reference binding to nullptr in map operations in TensorFlow |
| CVE-2021-37666 | 7.8 HIGH | Reference binding to nullptr in `RaggedTensorToVariant` in TensorFlow |
| CVE-2021-37667 | 7.8 HIGH | Reference binding to nullptr in unicode encoding in TensorFlow |
| CVE-2021-37652 | 7.8 HIGH | Use after free in boosted trees creation in TensorFlow |
| CVE-2021-37650 | 7.8 HIGH | Segfault and heap buffer overflow in `{Experimental,}DatasetToTFRecord` in TensorFlow |
| CVE-2021-37649 | 7.7 HIGH | Null pointer dereference in `UncompressElement` in TensorFlow |
| CVE-2021-37647 | 7.7 HIGH | Null pointer dereference in `SparseTensorSliceDataset` in TensorFlow |
| CVE-2021-37643 | 7.7 HIGH | Null pointer dereference in `MatrixDiagPartOp` in TensorFlow |
| CVE-2021-37637 | 7.7 HIGH | Null pointer dereference in `CompressElement` in TensorFlow |
| CVE-2021-37638 | 7.7 HIGH | Null pointer dereference in `RaggedTensorToTensor` in TensorFlow |
| CVE-2021-37659 | 7.3 HIGH | Out of bounds read via null pointer dereference in TensorFlow |
Showing top 20 of 58 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-125
- CVE-2026-8177
XML::LibXML versions through 2.0210 for Perl read out-of-bou - CVE-2026-6104
Global buffer over-read in mb_convert_encoding() with attack - CVE-2026-7258
Out-of-bounds read in urldecode() on NetBSD - CVE-2026-8186
Open5GS NF client.c ogs_sbi_client_send_via_scp_or_sepp out- - CVE-2026-3508
ASUS System Control Interface 缓冲区错误漏洞
V. Comments for CVE-2021-37679
No comments yet