Impact
Calling tf.raw_ops.ImmutableConst with a dtype of tf.resource or tf.variant results in a segfault in the implementation as code assumes that the tensor contents are pure scalars.
>>> import tensorflow as tf
>>> tf.raw_ops.ImmutableConst(dtype=tf.resource, shape=[], memory_region_name="/tmp/test.txt")
...
Segmentation fault
Patches
We have patched the issue in 4f663d4 and will release TensorFlow 2.5.0 containing the patch. TensorFlow nightly packages after this commit will also have the issue resolved.
Workarounds
If using tf.raw_ops.ImmutableConst in code, you can prevent the segfault by inserting a filter for the dtype argument.
For more information
Please consult our security guide for more information regarding the security model and how to contact us with issues and questions.
Impact
Calling
tf.raw_ops.ImmutableConstwith adtypeoftf.resourceortf.variantresults in a segfault in the implementation as code assumes that the tensor contents are pure scalars.Patches
We have patched the issue in 4f663d4 and will release TensorFlow 2.5.0 containing the patch. TensorFlow nightly packages after this commit will also have the issue resolved.
Workarounds
If using
tf.raw_ops.ImmutableConstin code, you can prevent the segfault by inserting a filter for thedtypeargument.For more information
Please consult our security guide for more information regarding the security model and how to contact us with issues and questions.