Skip to content
TensorFlowGHSA-c545-c4f9-rf6v

Heap OOB in TFLite

Medium5.5CVE-2021-37685 · Published Aug 25, 2021 · updated Jul 8, 2026

### Impact TFLite's [`expand_dims.cc`](https://github.com/tensorflow/tensorflow/blob/149562d49faa709ea80df1d99fc41d005b81082a/tensorflow/lite/kernels/expand_dims.cc#L36-L50) contains a vulnerability which allows reading one element outside of bounds of heap allocated data: ```cc if (axis < 0) { axis = input_dims.size + 1 + axis; } TF_LITE_ENSURE(context, axis <= input_dims.size); TfLiteIntArray* output_dims = TfLiteIntArrayCreate(input_dims.size + 1); for (int i = 0; i < output_dims->size; ++i) { if (i < axis) { output_dims->data[i] = input_dims.data[i]; } else if (i == axis) { output_dims->data[i] = 1; } else { output_dims->data[i] = input_dims.data[i - 1]; } } ``` If `axis` is a large negative value (e.g., `-100000`), then after the first `if` it would still be negative. The check following the `if` statement will pass and the `for` loop would read one element before the start of `input_dims.data` (when `i = 0`). ### Patches We have patched the issue in GitHub commit [d94ffe08a65400f898241c0374e9edc6fa8ed257](https://github.com/tensorflow/tensorflow/commit/d94ffe08a65400f898241c0374e9edc6fa8ed257). The fix will be included in...

GitHub advisory

Affected versions

PackageAffectedFixed in
tensorflow
PyPI
< 2.3.42.3.4
>= 2.4.0, < 2.4.32.4.3
>= 2.5.0, < 2.5.12.5.1
Details and references

### Impact TFLite's [`expand_dims.cc`](https://github.com/tensorflow/tensorflow/blob/149562d49faa709ea80df1d99fc41d005b81082a/tensorflow/lite/kernels/expand_dims.cc#L36-L50) contains a vulnerability which allows reading one element outside of bounds of heap allocated data: ```cc if (axis < 0) { axis = input_dims.size + 1 + axis; } TF_LITE_ENSURE(context, axis <= input_dims.size); TfLiteIntArray* output_dims = TfLiteIntArrayCreate(input_dims.size + 1); for (int i = 0; i < output_dims->size; ++i) { if (i < axis) { output_dims->data[i] = input_dims.data[i]; } else if (i == axis) { output_dims->data[i] = 1; } else { output_dims->data[i] = input_dims.data[i - 1]; } } ``` If `axis` is a large negative value (e.g., `-100000`), then after the first `if` it would still be negative. The check following the `if` statement will pass and the `for` loop would read one element before the start of `input_dims.data` (when `i = 0`). ### Patches We have patched the issue in GitHub commit [d94ffe08a65400f898241c0374e9edc6fa8ed257](https://github.com/tensorflow/tensorflow/commit/d94ffe08a65400f898241c0374e9edc6fa8ed257). 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](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 of Baidu Security.

CVSS 3.1
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N
Severity from
GitHub (reviewed advisory)
Weakness
CWE-125
Also known as
BIT-tensorflow-2021-37685, CVE-2021-37685, PYSEC-2021-307, PYSEC-2021-598, PYSEC-2021-796

More TensorFlow advisories

All TensorFlow
Advisory
Heap out of bounds access in sparse reduction operations
High7.3Aug 25, 2021
Floating point exception in `SparseDenseCwiseDiv`
Medium5.5Aug 25, 2021
Null pointer dereference in `CompressElement`
High7.7Aug 25, 2021
Null pointer dereference in `RaggedTensorToTensor`
High7.7Aug 25, 2021
Null pointer dereference and heap OOB read in operations restoring tensors
High8.4Aug 25, 2021
Integer division by 0 in sparse reshaping
Medium5.5Aug 25, 2021

Critical advisories by email

Wednesdays: the week’s critical and high advisories in the AI and data stack, with the fixed versions. Only in weeks that have some.

Double opt-in. Unsubscribe any time.