Skip to content
Fencer Vulns
medium 4.4

CVE-2020-26266 json

Uninitialized memory access in Eigen types in TensorFlow

published 2020-12-10 5 y ago modified 2026-09-08 2 mo ago also BIT-tensorflow-2020-26266 , EUVD-2020-0213 , GHSA-qhxx-j73r-qpm2 , PYSEC-2020-254 , PYSEC-2020-297 , PYSEC-2020-332

cvss
4.4
version 3.1, CNA
cisa kev
Not listed
not in CISA's catalog
epss
0.3%
EPSS 0.3% chance of exploitation in 30 days, percentile 18%
public exploits
0
none known

Summary

In affected versions of TensorFlow under certain cases a saved model can trigger use of uninitialized values during code execution. This is caused by having tensor buffers be filled with the default value of the type but forgetting to default initialize the quantized floating point types in Eigen. This is fixed in versions 1.15.5, 2.0.4, 2.1.3, 2.2.2, 2.3.2, and 2.4.0.

Patch / fix

  • pypi tensorflow 1.15.5, 2.0.4, 2.1.3, 2.2.2, 2.3.2
  • pypi tensorflow-cpu 1.15.5, 2.0.4, 2.1.3, 2.2.2, 2.3.2
  • pypi tensorflow-gpu 1.15.5, 2.0.4, 2.1.3, 2.2.2, 2.3.2
  • bitnami tensorflow 1.15.5, 2.0.4, 2.1.3, 2.2.2, 2.3.2
  • google tensorflow 1.15.5, 2.0.4, 2.1.3, 2.2.2, 2.3.2
  • tensorflow2 Package Hub 15 SP3, openSUSE Leap 15.3 2.6.0-bp153.2.3.1
  • bazel3.7 Package Hub 15 SP3, openSUSE Leap 15.3 3.7.2-bp153.4.1
  • tensorflow2-lite Package Hub 15 SP3, openSUSE Leap 15.3 2.6.0-bp153.2.3.1
  • bazel-skylib1.0.3 Package Hub 15 SP3, openSUSE Leap 15.3 1.0.3-bp153.2.1
  • tensorflow2_2_6_0-gnu-hpc Package Hub 15 SP3, openSUSE Leap 15.3 2.6.0-bp153.2.3.1
  • tensorflow2_2_6_0-gnu-openmpi2-hpc Package Hub 15 SP3, openSUSE Leap 15.3 2.6.0-bp153.2.3.1

opensuse advisories openSUSE-SU-2022:10014-1

  • tensorflow Arch Linux 2.4.0-1

arch linux advisories AVG-1348

Exploit prediction

0.26% chance of exploitation in the next 30 days, higher than 18% of all scored CVEs (FIRST EPSS, 2026-11-08).

Exploits and checks

  • github.com https://github.com/tensorflow/tensorflow/security/advisories/GHSA-qhxx-j73r-qpm2 referenced PoC, tagged "exploit" by a source, unverified

CVSS scores

assigned by version score vector
CNA 3.1 4.4 CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:L
GitHub 3.1 4.4 CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:L
NVD 3.1 5.3 CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L
NVD 2.0 4.6 AV:L/AC:L/Au:N/C:P/I:P/A:P
OSV 3.1 5.3 CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L

Affected packages

ecosystem package affected fixed in
pypi tensorflow
<1.15.5
>=2.0.0, <2.0.4
>=2.1.0, <2.1.3
>=2.2.0, <2.2.2
>=2.3.0, <2.3.2
1.15.5
2.0.4
2.1.3
2.2.2
2.3.2
pypi tensorflow-cpu
<1.15.5
>=2.0.0, <2.0.4
>=2.1.0, <2.1.3
>=2.2.0, <2.2.2
>=2.3.0, <2.3.2
1.15.5
2.0.4
2.1.3
2.2.2
2.3.2
pypi tensorflow-gpu
<1.15.5
>=2.0.0, <2.0.4
>=2.1.0, <2.1.3
>=2.2.0, <2.2.2
>=2.3.0, <2.3.2
1.15.5
2.0.4
2.1.3
2.2.2
2.3.2
bitnami tensorflow
<1.15.5
>=2.0.0, <2.0.4
>=2.1.0, <2.1.3
>=2.2.0, <2.2.2
>=2.3.0, <2.3.2
1.15.5
2.0.4
2.1.3
2.2.2
2.3.2

Affected products

  • tensorflow tensorflow

Classification

Weakness
  • CWE-908 Use of Uninitialized Resource

References

Sources

Fencer Vulns keeps one entry per vulnerability: the CVE id wins, then GHSA, and records that alias each other are merged, except that two CVEs always keep separate entries. The same data is available as JSON.