Advisory

GHSA-g8wg-cjwc-xhhp

Also known as: BIT-tensorflow-2021-37679, CVE-2021-37679, PYSEC-2021-301, PYSEC-2021-592, PYSEC-2021-790

Summary

Heap OOB in nested tf.map_fn with RaggedTensors

Details

Impact
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:

import tensorflow as tf
x = tf.ragged.constant([[1,2,3], [4,5], [6]])
t = tf.map_fn(lambda r: tf.map_fn(lambda y: r, r), x)
z = tf.ragged.constant([[[1,2,3],[1,2,3],[1,2,3]],[[4,5],[4,5]],[[6]]])

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 does not check that all inner shapes match and this results in the additional dimensions in the above example.

The same implementation can result in data loss, if input tensor is tweaked:

import tensorflow as tf
x = tf.ragged.constant([[1,2], [3,4,5], [6]])
t = tf.map_fn(lambda r: tf.map_fn(lambda y: r, r), x) 

Here, the output tensor will only have 2 elements for each inner dimension.

Patches
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.

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 Haris Sahovic.

Severity

High Basis: The source's own band: HIGH

Dates

Published
2021-08-25 14:41 UTC
Modified
2026-07-08 06:29 UTC

Affected packages and fixes

tensorflow

PyPI · All systems

Affected

  • introduced 0, fixed 2.3.4 (ecosystem)

Fixed versions

2.3.4

Fix

  • Fixed version 2.3.4
    pip install --upgrade tensorflow==2.3.4

tensorflow

PyPI · All systems

Affected

  • introduced 2.4.0, fixed 2.4.3 (ecosystem)

Fixed versions

2.4.3

Fix

  • Fixed version 2.4.3
    pip install --upgrade tensorflow==2.4.3

tensorflow

PyPI · All systems

Affected

  • introduced 2.5.0, fixed 2.5.1 (ecosystem)

Fixed versions

2.5.1

Fix

  • Fixed version 2.5.1
    pip install --upgrade tensorflow==2.5.1

tensorflow-cpu

PyPI · All systems

Affected

  • introduced 0, fixed 2.3.4 (ecosystem)

Fixed versions

2.3.4

Fix

  • Fixed version 2.3.4
    pip install --upgrade tensorflow-cpu==2.3.4

tensorflow-cpu

PyPI · All systems

Affected

  • introduced 2.4.0, fixed 2.4.3 (ecosystem)

Fixed versions

2.4.3

Fix

  • Fixed version 2.4.3
    pip install --upgrade tensorflow-cpu==2.4.3

tensorflow-cpu

PyPI · All systems

Affected

  • introduced 2.5.0, fixed 2.5.1 (ecosystem)

Fixed versions

2.5.1

Fix

  • Fixed version 2.5.1
    pip install --upgrade tensorflow-cpu==2.5.1

tensorflow-gpu

PyPI · All systems

Affected

  • introduced 0, fixed 2.3.4 (ecosystem)

Fixed versions

2.3.4

Fix

  • Fixed version 2.3.4
    pip install --upgrade tensorflow-gpu==2.3.4

tensorflow-gpu

PyPI · All systems

Affected

  • introduced 2.4.0, fixed 2.4.3 (ecosystem)

Fixed versions

2.4.3

Fix

  • Fixed version 2.4.3
    pip install --upgrade tensorflow-gpu==2.4.3

tensorflow-gpu

PyPI · All systems

Affected

  • introduced 2.5.0, fixed 2.5.1 (ecosystem)

Fixed versions

2.5.1

Fix

  • Fixed version 2.5.1
    pip install --upgrade tensorflow-gpu==2.5.1

Fix commands are the export's own generic commands. Ojuju never runs them. Check them against your setup first.

Check your own computer: guard packages check

View this record on OSV.dev

Source

Source: GitHub Advisory Database (reviewed advisories) · CC BY 4.0

Sources & licences