PYSEC-2021-568

TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can trigger a read from outside of bounds of heap allocated data by sending invalid arguments to tf.rawops.ResourceScatterUpdate. The implementation has an incomplete validation of the...

PYSEC-2021-567

TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can trigger a crash via a CHECK-fail in debug builds of TensorFlow using tf.rawops.ResourceGather or a read from outside the bounds of heap allocated data in the same API in a release build. Th...

PYSEC-2021-563

TensorFlow is an end-to-end open source platform for machine learning. In affected versions the implementation for tf.rawops.ExperimentalDatasetToTFRecord and tf.rawops.DatasetToTFRecord can trigger heap buffer overflow and segmentation fault. The implementation assumes that all records in the...

PYSEC-2021-564

TensorFlow is an end-to-end open source platform for machine learning. In affected versions the implementation for tf.rawops.FractionalAvgPoolGrad can be tricked into accessing data outside of bounds of heap allocated buffers. The implementation does not validate that the input tensor is non-empt...

Integer overflow

TensorFlow is an end-to-end open source platform for machine learning. In affected versions the implementation of tf.rawops.StringNGrams is vulnerable to an integer overflow issue caused by converting a signed integer value to an unsigned one and then allocating memory based on this value. The...

Input validation

TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can generate undefined behavior via a reference binding to nullptr in BoostedTreesCalculateBestGainsPerFeature and similar attack can occur in BoostedTreesCalculateBestFeatureSplitV2. The...

Design/Logic Flaw

TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can trigger a crash via a CHECK-fail in debug builds of TensorFlow using tf.rawops.ResourceGather or a read from outside the bounds of heap allocated data in the same API in a release build. Th...

PYSEC-2021-577

TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can read from outside of bounds of heap allocated data by sending specially crafted illegal arguments to BoostedTreesSparseCalculateBestFeatureSplit. The implementation needs to validate that...

PYSEC-2021-559

TensorFlow is an end-to-end open source platform for machine learning. In affected versions the implementation of tf.rawops.StringNGrams is vulnerable to an integer overflow issue caused by converting a signed integer value to an unsigned one and then allocating memory based on this value. The...

PYSEC-2021-575

TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can generate undefined behavior via a reference binding to nullptr in BoostedTreesCalculateBestGainsPerFeature and similar attack can occur in BoostedTreesCalculateBestFeatureSplitV2. The...

PYSEC-2021-569

TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can cause undefined behavior via binding a reference to null pointer in tf.rawops.RaggedTensorToSparse. The implementation has an incomplete validation of the splits values: it does not check...

PYSEC-2021-554

TensorFlow is an end-to-end open source platform for machine learning. In affected versions if the arguments to tf.rawops.RaggedGather don't determine a valid ragged tensor code can trigger a read from outside of bounds of heap allocated buffers. The implementation directly reads the first...

PYSEC-2021-572

TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can cause undefined behavior via binding a reference to null pointer in all binary cwise operations that don't require broadcasting e.g., gradients of binary cwise operations. The implementatio...

Null pointer dereference

TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can cause undefined behavior via binding a reference to null pointer in all binary cwise operations that don't require broadcasting e.g., gradients of binary cwise operations. The implementatio...

CVE-2021-37646

CVE-2021-37646 affects TensorFlow: the StringNGrams implementation can overflow when converting a negative, signed ngram_width to an unsigned size during a reserve call, enabling a potential denial-of-service condition. The root cause is a signed-to-unsigned conversion in TF’s string buffer alloc...

CVE-2021-37661 Crash caused by integer conversion to unsigned in TensorFlow

TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can cause a denial of service in boostedtreescreatequantilestreamresource by using negative arguments. The implementation does not validate that numstreams only contains non-negative numbers. I...

CVE-2021-37661

Summary: CVE-2021-37661 affects TensorFlow and causes a denial of service via the boosted_trees_create_quantile_stream_resource path when a negative number of streams is supplied. The code does not validate that num_streams is non-negative before using it to reserve memory, leading to an implicit...

CVE-2021-37645

TensorFlow CVE-2021-37645 affects affected TF versions prior to 2.6.0 and is caused by an integer overflow in tf.raw_ops.QuantizeAndDequantizeV4Grad when converting a signed axis to unsigned for the absl::InlinedVector constructor, leading to memory allocation based on a large value. A GitHub com...

CVE-2021-37651

TensorFlow: The FractionalAvgPoolGrad path has a heap-based buffer overflow when handling empty inputs, caused by not validating that the input tensor is non-empty. The implementation constructs an empty EigenDoubleMatrixMap and accesses out-of-bounds buffers. A patch was committed (0f931751fb20f...

CVE-2021-37650 Segfault and heap buffer overflow in `{Experimental,}DatasetToTFRecord` in TensorFlow

TensorFlow is an end-to-end open source platform for machine learning. In affected versions the implementation for tf.rawops.ExperimentalDatasetToTFRecord and tf.rawops.DatasetToTFRecord can trigger heap buffer overflow and segmentation fault. The implementation assumes that all records in the...