4.3

CVSS Score

3.1

Basic Information

CVE ID

CVE-2022-21733

GHSA ID

GHSA-98j8-c9q4-r38g

EPSS Score

0.44736%

CWE

CWE-190

Published

2/10/2022

Updated

11/13/2024

KEV Status

Technology

Python

Concerned about an active attack path?

Talk to our security experts and see Miggo in action.

Miggo Vulnerability Database

→

CVE-2022-21733

CVE-2022-21733: Memory exhaustion in Tensorflow

Impact

The implementation of StringNGrams can be used to trigger a denial of service attack by causing an OOM condition after an integer overflow:

import tensorflow as tf

tf.raw_ops.StringNGrams(
  data=['123456'],
  data_splits=[0,1],
  separator='a'*15,
  ngram_widths=[],
  left_pad='',
  right_pad='',
  pad_width=-5, 
  preserve_short_sequences=True)

We are missing a validation on pad_witdh and that result in computing a negative value for ngram_width which is later used to allocate parts of the output.

Patches

We have patched the issue in GitHub commit f68fdab93fb7f4ddb4eb438c8fe052753c9413e8.

The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.

For more information

Please consult our security guide for more information regarding the security model and how to contact us with issues and questions.

Attribution

This vulnerability has been reported by Yu Tian of Qihoo 360 AIVul Team.

(GitHub Advisory)

4.3

CVSS Score

3.1

Basic Information

CVE ID

CVE-2022-21733

GHSA ID

GHSA-98j8-c9q4-r38g

EPSS Score

0.44736%

CWE

CWE-190

Published

2/10/2022

Updated

11/13/2024

KEV Status

Technology

Python

Technical Details

CVSS Vector

CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L

Package Name	Ecosystem	Vulnerable Versions	First Patched Version
tensorflow	pip	< 2.5.3	2.5.3
tensorflow	pip	>= 2.6.0, < 2.6.3	2.6.3
tensorflow	pip	= 2.7.0	2.7.1
tensorflow-cpu	pip	< 2.5.3	2.5.3
tensorflow-cpu	pip	>= 2.6.0, < 2.6.3	2.6.3
tensorflow-cpu	pip	= 2.7.0	2.7.1
tensorflow-gpu	pip	< 2.5.3	2.5.3
tensorflow-gpu	pip	>= 2.6.0, < 2.6.3	2.6.3
tensorflow-gpu	pip	= 2.7.0	2.7.1

Vulnerability Intelligence
Miggo AI

Root Cause Analysis

The vulnerability stems from the StringNGramsOp's Compute function in string_ngrams_op.cc. The code path handling preserve_short_sequences=True with empty ngram_widths used pad_width without proper validation. The patch added an explicit check (OP_REQUIRES) for pad_width >=0 in this context. The original code allowed negative pad_width values, which when combined with data_length in ngram_width calculation (data_length + 2*pad_width), could produce invalid values used for memory allocation. The function's direct role in processing these parameters and the explicit fix in the commit confirm this as the vulnerable location.

Vulnerable functions

Only Mi**o us*rs **n s** t*is s**tion

WAF Protection Rules

WAF Rule

### Imp**t T** [impl*m*nt*tion o* `Strin*N*r*ms`](*ttps://*it*u*.*om/t*nsor*low/t*nsor*low/*lo*/****************************************/t*nsor*low/*or*/k*rn*ls/strin*_n*r*ms_op.**#L**-L***) **n ** us** to tri***r * **ni*l o* s*rvi** *tt**k *y **usi

Reasoning

T** vuln*r**ility st*ms *rom t** Strin*N*r*msOp's *omput* *un*tion in strin*_n*r*ms_op.**. T** *o** p*t* **n*lin* pr*s*rv*_s*ort_s*qu*n**s=Tru* wit* *mpty n*r*m_wi*t*s us** p**_wi*t* wit*out prop*r v*li**tion. T** p*t** ***** *n *xpli*it ****k (OP_R*

4.3

Basic Information

Concerned about an active attack path?

CVE-2022-21733: Memory exhaustion in Tensorflow

Impact

Patches

For more information

Attribution

4.3

Basic Information

Technical Details

Vulnerability IntelligenceMiggo AI

Root Cause Analysis

Vulnerable functions

WAF Protection Rules

WAF Rule

Reasoning

Vulnerability Intelligence
Miggo AI