AI and data stack advisories

Severe, 6 weeks1498Projects326

1498 severe, 6 weeks · 326 projects

TensorFlowGHSA-84mw-34w6-2q43

Null pointer dereference via invalid Ragged Tensors

TensorFlow

CVE-2021-29516 · Published May 21, 2021 · updated Jul 8, 2026

Low2.5
Fix: upgrade to 2.1.4 or later (4 fixed versions below)
GitHub advisory

Impact

Calling tf.raw_ops.RaggedTensorToVariant with arguments specifying an invalid ragged tensor results in a null pointer dereference:

```python

import tensorflow as tf

input_tensor = tf.constant([], shape=[0, 0, 0, 0, 0], dtype=tf.float32)

filter_tensor = tf.constant([], shape=[0, 0, 0, 0, 0], dtype=tf.float32)

tf.raw_ops.Conv3D(input=input_tensor, filter=filter_tensor, strides=[1, 56, 56, 56, 1], padding='VALID', data_format='NDHWC', dilations=[1, 1, 1, 23, 1])

`

```python

import tensorflow as tf

input_tensor = tf.constant([], shape=[2, 2, 2, 2, 0], dtype=tf.float32)

filter_tensor = tf.constant([], shape=[0, 0, 2, 6, 2], dtype=tf.float32)

tf.raw_ops.Conv3D(input=input_tensor, filter=filter_tensor, strides=[1, 56, 39, 34, 1], padding='VALID', data_format='NDHWC', dilations=[1, 1, 1, 1, 1])

`

The implementation of [RaggedTensorToVariant operations](https://github.com/tensorflow/tensorflow/blob/904b3926ed1c6c70380d5313d282d248a776baa1/tensorflow/core/kernels/ragged_tensor_to_variant_op.cc#L39-L40) does not validate that the ragged tensor argument is non-empty:

```cc

int ragged_rank = batched_ragged.ragged_rank();

auto batched_splits_top_vec = batched_ragged.splits(0).vec<SPLIT_TYPE>();

`

Since batched_ragged contains no elements, batched_ragged.splits is a null vector, thus batched_ragged.splits(0) will result in dereferencing nullptr....

Affected versions

PackageAffectedFixed in
tensorflow
PyPI
< 2.1.42.1.4
>= 2.2.0, < 2.2.32.2.3
>= 2.3.0, < 2.3.32.3.3
>= 2.4.0, < 2.4.22.4.2
Details and references

### Impact Calling `tf.raw_ops.RaggedTensorToVariant` with arguments specifying an invalid ragged tensor results in a null pointer dereference: ```python import tensorflow as tf input_tensor = tf.constant([], shape=[0, 0, 0, 0, 0], dtype=tf.float32) filter_tensor = tf.constant([], shape=[0, 0, 0, 0, 0], dtype=tf.float32) tf.raw_ops.Conv3D(input=input_tensor, filter=filter_tensor, strides=[1, 56, 56, 56, 1], padding='VALID', data_format='NDHWC', dilations=[1, 1, 1, 23, 1]) ``` ```python import tensorflow as tf input_tensor = tf.constant([], shape=[2, 2, 2, 2, 0], dtype=tf.float32) filter_tensor = tf.constant([], shape=[0, 0, 2, 6, 2], dtype=tf.float32) tf.raw_ops.Conv3D(input=input_tensor, filter=filter_tensor, strides=[1, 56, 39, 34, 1], padding='VALID', data_format='NDHWC', dilations=[1, 1, 1, 1, 1]) ``` The implementation of [`RaggedTensorToVariant` operations](https://github.com/tensorflow/tensorflow/blob/904b3926ed1c6c70380d5313d282d248a776baa1/tensorflow/core/kernels/ragged_tensor_to_variant_op.cc#L39-L40) does not validate that the ragged tensor argument is non-empty: ```cc int ragged_rank = batched_ragged.ragged_rank(); auto batched_splits_top_vec = batched_ragged.splits(0).vec<SPLIT_TYPE>(); ``` Since `batched_ragged` contains no elements, `batched_ragged.splits` is a null vector, thus `batched_ragged.splits(0)` will result in dereferencing `nullptr`. ### Patches We have patched the issue in GitHub commit [b055b9c474cd376259dde8779908f9eeaf097d93](https://github.com/tensorflow/tensorflow/commit/b055b9c474cd376259dde8779908f9eeaf097d93). The fix will be included in TensorFlow 2.5.0. We will also cherrypick this commit on TensorFlow 2.4.2, TensorFlow 2.3.3, TensorFlow 2.2.3 and TensorFlow 2.1.4, as these are also affected and still in supported range. ### For more information Please consult [our security guide](https://github.com/tensorflow/tensorflow/blob/master/SECURITY.md) for more information regarding the security model and how to contact us with issues and questions. ### Attribution This vulnerability has been reported by Yakun Zhang and Ying Wang of Baidu X-Team.

CVSS 3.1
CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L
Severity from
GitHub (reviewed advisory)
Weakness
CWE-476
Also known as
BIT-tensorflow-2021-29516, CVE-2021-29516, PYSEC-2021-153, PYSEC-2021-444, PYSEC-2021-642

More TensorFlow advisories

All TensorFlow
Low 2.5May 21, 2021Segfault in `tf.raw_ops.SparseCountSparseOutput`CVE-2021-29619fixed in 2.1.4 +3Low 2.5May 21, 2021Crash in `tf.transpose` with complex inputsCVE-2021-29618fixed in 2.1.4 +3Low 2.5May 21, 2021Crash in `tf.strings.substr` due to `CHECK`-failCVE-2021-29617fixed in 2.1.4 +3Low 2.5May 21, 2021Null dereference in Grappler's `TrySimplify`CVE-2021-29616fixed in 2.1.4 +3Low 2.5May 21, 2021Stack overflow in `ParseAttrValue` with nested tensorsCVE-2021-29615fixed in 2.1.4 +3Medium 7.1May 21, 2021Interpreter crash from `tf.io.decode_raw`CVE-2021-29614fixed in 2.1.4 +3