Headline
GHSA-mg66-qvc5-rm93: Missing validation causes denial of service via `SparseTensorToCSRSparseMatrix`
Impact
The implementation of tf.raw_ops.SparseTensorToCSRSparseMatrix does not fully validate the input arguments. This results in a CHECK-failure which can be used to trigger a denial of service attack:
import tensorflow as tf
indices = tf.constant(53, shape=[3], dtype=tf.int64)
values = tf.constant(0.554979503, shape=[218650], dtype=tf.float32)
dense_shape = tf.constant(53, shape=[3], dtype=tf.int64)
tf.raw_ops.SparseTensorToCSRSparseMatrix(
indices=indices,
values=values,
dense_shape=dense_shape)
The code assumes dense_shape is a vector and indices is a matrix (as part of requirements for sparse tensors) but there is no validation for this:
const Tensor& indices = ctx->input(0);
const Tensor& values = ctx->input(1);
const Tensor& dense_shape = ctx->input(2);
const int rank = dense_shape.NumElements();
OP_REQUIRES(ctx, rank == 2 || rank == 3,
errors::InvalidArgument("SparseTensor must have rank 2 or 3; ",
"but indices has rank: ", rank));
auto dense_shape_vec = dense_shape.vec<int64_t>();
// ...
OP_REQUIRES_OK(
ctx,
coo_to_csr(batch_size, num_rows, indices.template matrix<int64_t>(),
batch_ptr.vec<int32>(), csr_row_ptr.vec<int32>(),
csr_col_ind.vec<int32>()));
Patches
We have patched the issue in GitHub commit ea50a40e84f6bff15a0912728e35b657548cef11.
The fix will be included in TensorFlow 2.9.0. We will also cherrypick this commit on TensorFlow 2.8.1, TensorFlow 2.7.2, and TensorFlow 2.6.4, as these are also affected and still in supported range.
For more information
Please consult our security guide for more information regarding the security model and how to contact us with issues and questions.
Attribution
This vulnerability has been reported by Neophytos Christou from Secure Systems Lab at Brown University.
Impact
The implementation of tf.raw_ops.SparseTensorToCSRSparseMatrix does not fully validate the input arguments. This results in a CHECK-failure which can be used to trigger a denial of service attack:
import tensorflow as tf
indices = tf.constant(53, shape=[3], dtype=tf.int64) values = tf.constant(0.554979503, shape=[218650], dtype=tf.float32) dense_shape = tf.constant(53, shape=[3], dtype=tf.int64)
tf.raw_ops.SparseTensorToCSRSparseMatrix( indices=indices, values=values, dense_shape=dense_shape)
The code assumes dense_shape is a vector and indices is a matrix (as part of requirements for sparse tensors) but there is no validation for this:
const Tensor& indices = ctx->input(0);
const Tensor& values = ctx->input(1);
const Tensor& dense\_shape = ctx->input(2);
const int rank = dense\_shape.NumElements();
OP\_REQUIRES(ctx, rank == 2 || rank == 3,
errors::InvalidArgument("SparseTensor must have rank 2 or 3; ",
"but indices has rank: ", rank));
auto dense\_shape\_vec = dense\_shape.vec<int64\_t\>();
// ...
OP\_REQUIRES\_OK(
ctx,
coo\_to\_csr(batch\_size, num\_rows, indices.template matrix<int64\_t\>(),
batch\_ptr.vec<int32>(), csr\_row\_ptr.vec<int32>(),
csr\_col\_ind.vec<int32>()));
Patches
We have patched the issue in GitHub commit ea50a40e84f6bff15a0912728e35b657548cef11.
The fix will be included in TensorFlow 2.9.0. We will also cherrypick this commit on TensorFlow 2.8.1, TensorFlow 2.7.2, and TensorFlow 2.6.4, as these are also affected and still in supported range.
For more information
Please consult our security guide for more information regarding the security model and how to contact us with issues and questions.
Attribution
This vulnerability has been reported by Neophytos Christou from Secure Systems Lab at Brown University.
References
- GHSA-mg66-qvc5-rm93
- https://nvd.nist.gov/vuln/detail/CVE-2022-29198
- tensorflow/tensorflow@ea50a40
- https://github.com/tensorflow/tensorflow/blob/f3b9bf4c3c0597563b289c0512e98d4ce81f886e/tensorflow/core/kernels/sparse/sparse_tensor_to_csr_sparse_matrix_op.cc#L65-L119
- https://github.com/tensorflow/tensorflow/releases/tag/v2.6.4
- https://github.com/tensorflow/tensorflow/releases/tag/v2.7.2
- https://github.com/tensorflow/tensorflow/releases/tag/v2.8.1
- https://github.com/tensorflow/tensorflow/releases/tag/v2.9.0
Related news
TensorFlow is an open source platform for machine learning. Prior to versions 2.9.0, 2.8.1, 2.7.2, and 2.6.4, the implementation of `tf.raw_ops.SparseTensorToCSRSparseMatrix` does not fully validate the input arguments. This results in a `CHECK`-failure which can be used to trigger a denial of service attack. The code assumes `dense_shape` is a vector and `indices` is a matrix (as part of requirements for sparse tensors) but there is no validation for this. Versions 2.9.0, 2.8.1, 2.7.2, and 2.6.4 contain a patch for this issue.
TensorFlow is an open source platform for machine learning. Prior to versions 2.9.0, 2.8.1, 2.7.2, and 2.6.4, multiple TensorFlow operations misbehave in eager mode when the resource handle provided to them is invalid. In graph mode, it would have been impossible to perform these API calls, but migration to TF 2.x eager mode opened up this vulnerability. If the resource handle is empty, then a reference is bound to a null pointer inside TensorFlow codebase (various codepaths). This is undefined behavior. Versions 2.9.0, 2.8.1, 2.7.2, and 2.6.4 contain a patch for this issue.