CVE-2021-29533: Denial of Service in TensorFlow-GPU

Background

TensorFlow-GPU is an open-source platform for machine learning that provides a comprehensive ecosystem for developing and deploying machine learning models. It is widely used in various industries and research domains for its powerful capabilities and ease of use. However, a vulnerability has been discovered in TensorFlow-GPU that can lead to a denial of service (DoS) attack.

Vulnerability Detail

The vulnerability (CVE-2021-29533) in TensorFlow-GPU arises from a flaw in the implementation of the tf.raw_ops.DrawBoundingBoxes function. An attacker can trigger a DoS attack by passing an empty image to this function, causing a CHECK failure. The implementation uses CHECK_* assertions instead of OP_REQUIRES to validate user-controlled inputs. While OP_REQUIRES allows returning an error condition to the user, CHECK_* macros result in a program crash if the condition is false. In this case, the height of the image is 0, leading to a negative value for max_box_row_clamp and a falsified assertion, resulting in the program aborting. This vulnerability can be exploited by an attacker to disrupt the availability of TensorFlow-GPU and potentially impact systems relying on it for machine learning tasks.

pip install --upgrade "tensorflow-gpu==2.12.0"

Workaround

As a temporary workaround, users can validate the input images before passing them to the tf.raw_ops.DrawBoundingBoxes function. By ensuring that the height of the image is not zero, the vulnerability can be mitigated. Here is an example code snippet demonstrating the workaround:

# Validate image height
if image.shape[0] == 0:
    raise ValueError('Image height cannot be zero')

# Call the DrawBoundingBoxes function
tf.raw_ops.DrawBoundingBoxes(images, ...)

Conclusion

The vulnerability (CVE-2021-29533) in TensorFlow-GPU can lead to a denial of service (DoS) attack by exploiting a flaw in the tf.raw_ops.DrawBoundingBoxes function. It is crucial for users to update their TensorFlow-GPU installations to the latest patched version once it becomes available. In the meantime, a temporary workaround is available by validating input images to mitigate the vulnerability. For further support and assistance in vulnerability remediation, you can reach out to DevNack.com.

See every vulnerability recorded for tensorflow-gpu →

FAQ