GHSA-RRFP-J2MP-HQ9C Segfault in `tf.quantization.quantize_and_dequantize`

Impact An attacker can pass an invalid axis value to tf.quantization.quantizeanddequantize: python tf.quantization.quantizeanddequantize input=2.5, 2.5, inputmin=0,0, inputmax=1,1, axis=10 This results in accessing a dimension outside the rank of the input tensor in the C++ kernel implementation:...

Segfault in `tf.quantization.quantize_and_dequantize`

Impact An attacker can pass an invalid axis value to tf.quantization.quantizeanddequantize: python tf.quantization.quantizeanddequantize input=2.5, 2.5, inputmin=0,0, inputmax=1,1, axis=10 This results in accessing a dimension outside the rank of the input tensor in the C++ kernel implementation:...

Google TensorFlow Buffer Overflow Vulnerability (CNVD-2020-62802)

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. A security vulnerability exists in Tensorflow versions prior to 2.4.0, which can be exploited by an attacker to pass an invalid axis value to tf.quantization. digit transformation and dequantize...

Security Bulletin: Numerous CVE entires for TensorFlow in Watson Machine Learning Community Edition

Summary An collection of CVE's have been published for TensorFlow. Details of each are listed. Vulnerability Details CVEID: CVE-2020-15206 DESCRIPTION: TensorFlow is vulnerable to a denial of service, caused by a flaw when changing the SavedModel protocol buffer and altering the name of required...

OPENSUSE-SU-2020:1766-1 Security update for tensorflow2

This update for tensorflow2 fixes the following issues: - updated to 2.1.2 with following fixes boo1177022: Fixes an undefined behavior causing a segfault in tf.rawops.Switch CVE-2020-15190 Fixes three vulnerabilities in conversion to DLPack format CVE-2020-15191, CVE-2020-15192, CVE-2020-15193...

Security Bulletin: CVE-2020-15190 for Tensorflow in Watson Machine Learning Community Edition

Summary In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, the tf.rawops.Switch operation takes as input a tensor and a boolean and outputs two tensors. Depending on the boolean value, one of the tensors is exactly the input tensor whereas the other one should be an empty tensor...

Google TensorFlow Buffer Overflow Vulnerability (CNVD-2020-62809)

Google TensorFlow is a suite of end-to-end open source platforms for machine learning from Google USA. A security vulnerability exists in Tensorflow versions prior to 2.4.0, which stems from when the parameter to boxes of tf.image.cropandresize has a very large value, and the CPU kernel...

Denial Of Service (DoS)

tensorflow is vulnerable to denial of service. A segmentation fault in tf.quantization.quantizeanddequantize when an invalid axis value is passed by an attacker results in an application crash and potentially allow arbitrary code execution...

CVE-2020-15266

In Tensorflow before version 2.4.0, when the boxes argument of tf.image.cropandresize has a very large value, the CPU kernel implementation receives it as a C++ nan floating point value. Attempting to operate on this is undefined behavior which later produces a segmentation fault. The issue is...

CVE-2020-15265

In Tensorflow before version 2.4.0, an attacker can pass an invalid axis value to tf.quantization.quantizeanddequantize. This results in accessing a dimension outside the rank of the input tensor in the C++ kernel implementation. However, dimsize only does a DCHECK to validate the argument and th...

CVE-2020-15266

In Tensorflow before version 2.4.0, when the boxes argument of tf.image.cropandresize has a very large value, the CPU kernel implementation receives it as a C++ nan floating point value. Attempting to operate on this is undefined behavior which later produces a segmentation fault. The issue is...

CVE-2020-15265

In Tensorflow before version 2.4.0, an attacker can pass an invalid axis value to tf.quantization.quantizeanddequantize. This results in accessing a dimension outside the rank of the input tensor in the C++ kernel implementation. However, dimsize only does a DCHECK to validate the argument and th...

PYSEC-2020-296

In Tensorflow before version 2.4.0, when the boxes argument of tf.image.cropandresize has a very large value, the CPU kernel implementation receives it as a C++ nan floating point value. Attempting to operate on this is undefined behavior which later produces a segmentation fault. The issue is...

PYSEC-2020-139

In Tensorflow before version 2.4.0, when the boxes argument of tf.image.cropandresize has a very large value, the CPU kernel implementation receives it as a C++ nan floating point value. Attempting to operate on this is undefined behavior which later produces a segmentation fault. The issue is...

PYSEC-2020-330

In Tensorflow before version 2.4.0, an attacker can pass an invalid axis value to tf.quantization.quantizeanddequantize. This results in accessing a dimension outside the rank of the input tensor in the C++ kernel implementation. However, dimsize only does a DCHECK to validate the argument and th...

PYSEC-2020-295

In Tensorflow before version 2.4.0, an attacker can pass an invalid axis value to tf.quantization.quantizeanddequantize. This results in accessing a dimension outside the rank of the input tensor in the C++ kernel implementation. However, dimsize only does a DCHECK to validate the argument and th...

PYSEC-2020-296

In Tensorflow before version 2.4.0, when the boxes argument of tf.image.cropandresize has a very large value, the CPU kernel implementation receives it as a C++ nan floating point value. Attempting to operate on this is undefined behavior which later produces a segmentation fault. The issue is...

Design/Logic Flaw

In Tensorflow before version 2.4.0, an attacker can pass an invalid axis value to tf.quantization.quantizeanddequantize. This results in accessing a dimension outside the rank of the input tensor in the C++ kernel implementation. However, dimsize only does a DCHECK to validate the argument and th...

PYSEC-2020-295

In Tensorflow before version 2.4.0, an attacker can pass an invalid axis value to tf.quantization.quantizeanddequantize. This results in accessing a dimension outside the rank of the input tensor in the C++ kernel implementation. However, dimsize only does a DCHECK to validate the argument and th...

Design/Logic Flaw

In Tensorflow before version 2.4.0, when the boxes argument of tf.image.cropandresize has a very large value, the CPU kernel implementation receives it as a C++ nan floating point value. Attempting to operate on this is undefined behavior which later produces a segmentation fault. The issue is...