TensorFlow is an end-to-end open source platform for machine learning. In affected versions an attacker can trigger a crash via a CHECK-fail in debug builds of TensorFlow using tf.raw_ops.ResourceGather or a read from outside the bounds of heap allocated data in the same API in a release build. The implementation does not check that the batch_dims value that the user supplies is less than the rank of the input tensor. Since the implementation uses several for loops over the dimensions of tensor, this results in reading data from outside the bounds of heap allocated buffer backing the tensor. We have patched the issue in GitHub commit bc9c546ce7015c57c2f15c168b3d9201de679a1d. The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, as these are also affected and still in supported range.
Weakness
The product reads data past the end, or before the beginning, of the intended buffer.
Affected Software
| Name |
Vendor |
Start Version |
End Version |
| Tensorflow |
Google |
2.3.0 (including) |
2.3.4 (excluding) |
| Tensorflow |
Google |
2.4.0 (including) |
2.4.3 (excluding) |
| Tensorflow |
Google |
2.5.0 (including) |
2.5.0 (including) |
| Tensorflow |
Google |
2.6.0-rc0 (including) |
2.6.0-rc0 (including) |
| Tensorflow |
Google |
2.6.0-rc1 (including) |
2.6.0-rc1 (including) |
| Tensorflow |
Google |
2.6.0-rc2 (including) |
2.6.0-rc2 (including) |
Potential Mitigations
- Assume all input is malicious. Use an “accept known good” input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does.
- When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, “boat” may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as “red” or “blue.”
- Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code’s environment changes. This can give attackers enough room to bypass the intended validation. However, denylists can be useful for detecting potential attacks or determining which inputs are so malformed that they should be rejected outright.
- To reduce the likelihood of introducing an out-of-bounds read, ensure that you validate and ensure correct calculations for any length argument, buffer size calculation, or offset. Be especially careful of relying on a sentinel (i.e. special character such as NUL) in untrusted inputs.
References