Headline
CVE-2022-36004: Fix size check for large input shape and rates. · tensorflow/tensorflow@552bfce
TensorFlow is an open source platform for machine learning. When tf.random.gamma receives large input shape and rates, it gives a CHECK fail that can trigger a denial of service attack. We have patched the issue in GitHub commit 552bfced6ce4809db5f3ca305f60ff80dd40c5a3. The fix will be included in TensorFlow 2.10.0. We will also cherrypick this commit on TensorFlow 2.9.1, TensorFlow 2.8.1, and TensorFlow 2.7.2, as these are also affected and still in supported range. There are no known workarounds for this issue.
@@ -16,7 +16,10 @@
import numpy as np
from tensorflow.python.eager import context
from tensorflow.python.framework import constant_op
from tensorflow.python.framework import dtypes
from tensorflow.python.framework import errors
from tensorflow.python.framework import ops
from tensorflow.python.framework import random_seed
from tensorflow.python.framework import test_util
@@ -216,6 +219,16 @@ def testPositive(self):
self.assertEqual(0, math_ops.reduce_sum(math_ops.cast(
math_ops.less_equal(x, 0.), dtype=dtypes.int64)).eval())
def testSizeTooLarge(self):
# Grappler asserts on size overflow, so this error is only caught when
# running eagerly.
if context.executing_eagerly():
with self.assertRaisesRegex((ValueError, errors.InvalidArgumentError),
“overflow”):
rate = constant_op.constant(1.0, shape=(4, 4, 4, 4, 4))
self.evaluate(
random_ops.random_gamma(
shape=[46902, 51188, 34063, 59195], alpha=rate))
if __name__ == "__main__":
test.main()
Related news
### Impact When `tf.random.gamma` receives large input shape and rates, it gives a `CHECK` fail that can trigger a denial of service attack. ```python import tensorflow as tf arg_0=tf.random.uniform(shape=(4,), dtype=tf.int32, maxval=65536) arg_1=tf.random.uniform(shape=(4, 4), dtype=tf.float64, maxval=None) arg_2=tf.random.uniform(shape=(4, 4, 4, 4, 4), dtype=tf.float64, maxval=None) arg_3=tf.float64 arg_4=48 arg_5='None' tf.random.gamma(shape=arg_0, alpha=arg_1, beta=arg_2, dtype=arg_3, seed=arg_4, name=arg_5) ``` ### Patches We have patched the issue in GitHub commit [552bfced6ce4809db5f3ca305f60ff80dd40c5a3](https://github.com/tensorflow/tensorflow/commit/552bfced6ce4809db5f3ca305f60ff80dd40c5a3). The fix will be included in TensorFlow 2.10.0. We will also cherrypick this commit on TensorFlow 2.9.1, TensorFlow 2.8.1, and TensorFlow 2.7.2, as these are also affected and still in supported range. ### For more information Please consult [our security guide](https://github.com/tens...