Miggo Logo

CVE-2021-41207: FPE in `ParallelConcat`

5.5

CVSS Score
3.1

Basic Information

EPSS Score
0.02677%
Published
11/10/2021
Updated
11/13/2024
KEV Status
No
Technology
TechnologyPython

Technical Details

CVSS Vector
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
Package NameEcosystemVulnerable VersionsFirst Patched Version
tensorflowpip>= 2.6.0, < 2.6.12.6.1
tensorflowpip>= 2.5.0, < 2.5.22.5.2
tensorflowpip< 2.4.42.4.4
tensorflow-cpupip>= 0, < 2.4.42.4.4
tensorflow-cpupip>= 2.5.0, < 2.5.22.5.2
tensorflow-cpupip>= 2.6.0, < 2.6.12.6.1
tensorflow-gpupip>= 0, < 2.4.42.4.4
tensorflow-gpupip>= 2.5.0, < 2.5.22.5.2
tensorflow-gpupip>= 2.6.0, < 2.6.12.6.1

Vulnerability Intelligence
Miggo AIMiggo AI

Miggo AIRoot Cause Analysis

The vulnerability stems from missing input validation in ParallelConcat's implementation. The commit f2c3931 added validation checks for tensor rank and 0th dimension size in ParallelConcatUpdate::Compute. The original vulnerable code (before patch) in inplace_ops.cc lines 72-97 did not validate that value.dim_size(0) > loc_, which could result in division by zero when calculating output dimensions. The test case demonstrates triggering this with shape=0, which would fail the added validation checks. The function is directly responsible for handling ParallelConcat operations and contained the unvalidated division operation.

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* `P*r*ll*l*on**t`](*ttps://*it*u*.*om/t*nsor*low/t*nsor*low/*lo*/****************************************/t*nsor*low/*or*/k*rn*ls/inpl***_ops.**#L**-L**) miss*s som* input v*li**tion *n* **n pro*u** * *ivision *y *:

Reasoning

T** vuln*r**ility st*ms *rom missin* input v*li**tion in P*r*ll*l*on**t's impl*m*nt*tion. T** *ommit ******* ***** v*li**tion ****ks *or t*nsor r*nk *n* *t* *im*nsion siz* in P*r*ll*l*on**tUp**t*::*omput*. T** ori*in*l vuln*r**l* *o** (***or* p*t**)