PYSEC-2021-493

TensorFlow is an end-to-end open source platform for machine learning. An attacker can trigger a null pointer dereference in the implementation of tf.rawops.SparseFillEmptyRows. This is because of missing...

PYSEC-2021-510

TensorFlow is an end-to-end open source platform for machine learning. Due to lack of validation in tf.rawops.Dequantize, an attacker can trigger a read from outside of bounds of heap allocated data. The...

PYSEC-2021-460

TensorFlow is an end-to-end open source platform for machine learning. An attacker can force accesses outside the bounds of heap allocated arrays by passing in invalid tensor values to tf.rawops.RaggedCross. This is because the...

PYSEC-2021-457

TensorFlow is an end-to-end open source platform for machine learning. An attacker can trigger a heap buffer overflow in tf.rawops.QuantizedResizeBilinear by manipulating input values so that float rounding results in off-by-one error in accessing image elements. This is because the...

CVE-2021-29513

TensorFlow vulnerability CVE-2021-29513 arises when calling TF operations with tensors of non-numeric types, causing a null pointer dereference due to a type confusion in the Python-to-C++ array conversion (ndarray_tensor.cc). Root cause: PyArray_DESCR_to_TF_DataType path can dereference NULL in ...

CVE-2021-29514

TensorFlow RaggedBincount vulnerability (CVE-2021-29514) in the RaggedBincount kernel can cause a heap out-of-bounds write when the splits argument references an invalid SparseTensor, leading to a write at out(-1, bin). The issue is triggered by setting splits(0) to 7, causing batch_idx to remain...

CVE-2021-29515 Reference binding to null pointer in `MatrixDiag*` ops

TensorFlow is an end-to-end open source platform for machine learning. The implementation of MatrixDiag operationshttps://github.com/tensorflow/tensorflow/blob/4c4f420e68f1cfaf8f4b6e8e3eb857e9e4c3ff33/tensorflow/core/kernels/linalg/matrixdiagop.ccL195-L197 does not validate that the tensor...

CVE-2021-29515

The CVE-2021-29515 issue affects TensorFlow MatrixDiag* ops: input tensors are not validated to be non-empty, which can lead to a null pointer dereference. The root cause is in MatrixDiagV2/V3 path handling inputs, and patches fix the issue (commit a7116dd39…) with the fix slated for TensorFlow 2...

CVE-2021-29516

TensorFlow CVE-2021-29516 describes a null pointer dereference in tf.raw_ops.RaggedTensorToVariant when provided with an invalid ragged tensor. The issue arises because batched_ragged.splits(0) is dereferenced without validating non-emptiness. Affected: TensorFlow and related entries indicate the...

CVE-2021-29516 Null pointer dereference via invalid Ragged Tensors

TensorFlow is an end-to-end open source platform for machine learning. Calling tf.rawops.RaggedTensorToVariant with arguments specifying an invalid ragged tensor results in a null pointer dereference. The implementation of RaggedTensorToVariant...

CVE-2021-29517

CVE-2021-29517 affects TensorFlow Conv3D: division-by-zero in the Conv3D kernel caused by a modulo on user input (fifth filter dimension), potentially triggering an Eigen assertion and a crash. The issue is addressed by a TensorFlow fix in 2.5.0, with cherry-picks to 2.4.2, 2.3.3, 2.2.3 and 2.1.4...

CVE-2021-29518

CVE-2021-29518 describes a vulnerability in TensorFlow where, in eager mode, session operations can dereference a null session_state pointer, leading to undefined behavior. Concrete details from connected documents show the root cause in tensor flow core/kernels/session_ops.cc, where ctx->sess...

CVE-2021-29520

TensorFlow CVE-2021-29520 concerns a heap buffer overflow in Conv3DBackprop* due to missing validation that assumes input, filter_sizes, and out_backprop have identical shapes. Multiple sources (OSV entries and GHSA advisory) corroborate the issue and patch lineage. The vulnerability affects Conv...

CVE-2021-29521

TensorFlow CVE-2021-29521: A bug in tf.raw_ops.SparseCountSparseOutput triggers a segmentation fault when dense_shape contains negative values. Root cause is the implementation assuming the first element of dense_shape is positive to initialize BatchedMap; with multi-element shapes, num_batches d...

CVE-2021-29523

CVE-2021-29523 : TensorFlow vulnerability where a crafted input for AddManySparseToTensorsMap can trigger a denial-of-service via a CHECK failure in TensorShapeInitDims when sparse_shape values overflow. Root cause: legacy TensorShapeBase constructor multiplies dimensions with potential overflow,...

CVE-2021-29524 Division by 0 in `Conv2DBackpropFilter`

TensorFlow is an end-to-end open source platform for machine learning. An attacker can trigger a division by 0 in tf.rawops.Conv2DBackpropFilter. This is because the...

CVE-2021-29524

TensorFlow (Conv2DBackpropFilter) suffers a division-by-zero vulnerability caused by a modulus operation in conv_grad_shape_utils.cc where the divisor is supplied by the caller. The concrete issue has been tracked as CVE-2021-29524 and is documented across multiple sources (OSV and Ghsa advisorie...

CVE-2021-29585 Division by zero in padding computation in TFLite

TensorFlow is an end-to-end open source platform for machine learning. The TFLite computation for size of output after padding, ComputeOutSizehttps://github.com/tensorflow/tensorflow/blob/0c9692ae7b1671c983569e5d3de5565843d500cf/tensorflow/lite/kernels/padding.hL43-L55, does not check that the...

CVE-2021-29585

TensorFlow/TFLite padding compute path has a division-by-zero in ComputeOutSize when stride is 0, enabling a potential denial-of-service scenario via crafted models. The issue affects padding logic in TF Lite; patches were applied in commit 49847ae and a fix is planned for TensorFlow 2.5.0 with c...

CVE-2021-29586

CVE-2021-29586 affects TensorFlow (TFLite pooling) where optimized pooling implementations fail to validate stride values, allowing params->stride_height/width to be zero and cause a division by zero in ComputePaddingHeightWidth. Practically, this is a vulnerability in the pooling path of Tens...