CVE-2020-15193
published 2020-09-25CVE-2020-15193: In Tensorflow before versions 2.2.1 and 2.3.1, the implementation of `dlpack.to_dlpack` can be made to use uninitialized memory resulting in further memory…
high7.1CVSS 3.1
AVNACLPRLUINSUCNIHAL
In Tensorflow before versions 2.2.1 and 2.3.1, the implementation of `dlpack.to_dlpack` can be made to use uninitialized memory resulting in further memory corruption. This is because the pybind11 glue code assumes that the argument is a tensor. However, there is nothing stopping users from passing in a Python object instead of a tensor. The uninitialized memory address is due to a `reinterpret_cast` Since the `PyObject` is a Python object, not a TensorFlow Tensor, the cast to `EagerTensor` fails. The issue is patched in commit 22e07fb204386768e5bcbea563641ea11f96ceb8 and is released in TensorFlow versions 2.2.1, or 2.3.1.
Affected
11 ranges
| Vendor | Product | Version range | Fixed in |
|---|---|---|---|
| debian | tensorflow | — | — |
| tensorflow | — | — | |
| tensorflow | — | — | |
| intel | optimization_for_tensorflow | >= 0 < 22e07fb204386768e5bcbea563641ea11f96ceb8 | 22e07fb204386768e5bcbea563641ea11f96ceb8 |
| intel | optimization_for_tensorflow | >= 0 < 2.2.1 | 2.2.1 |
| intel | optimization_for_tensorflow | >= 2.2.0 < 2.2.1 | 2.2.1 |
| intel | optimization_for_tensorflow | >= 2.3.0 < 2.3.1 | 2.3.1 |
| intel | optimization_for_tensorflow | >= 2.3.0rc0 < 2.3.1 | 2.3.1 |
| opensuse | leap | — | — |
| tensorflow | tensorflow | — | — |
| tensorflow | tensorflow | — | — |
OSV
Memory corruption in Tensorflow
osv·2020-09-25
CVE-2020-15193 [HIGH] Memory corruption in Tensorflow
Memory corruption in Tensorflow
### Impact
The implementation of `dlpack.to_dlpack` can be made to use uninitialized memory resulting in further memory corruption. This is because the pybind11 glue code assumes that the argument is a tensor:
https://github.com/tensorflow/tensorflow/blob/0e68f4d3295eb0281a517c3662f6698992b7b2cf/tensorflow/python/tfe_wrapper.cc#L1361
However, there is nothing stopping users from passing in a Python object instead of a tensor.
```python
In [2]: tf.experimental.dlpack.to_dlpack([2])
==1720623==WARNING: MemorySanitizer: use-of-uninitialized-value
#0 0x55b0ba5c410a in tensorflow::(anonymous namespace)::GetTensorFromHandle(TFE_TensorHandle*, TF_Status*) third_party/tensorflow/c/eager/dlpack.cc:46:7
#1 0x55b0ba5c38f4 in tensorflow::TFE_HandleToDLPack(TFE_TensorH
OSV
CVE-2020-15193: In Tensorflow before versions 2
osv·2020-09-25
CVE-2020-15193 CVE-2020-15193: In Tensorflow before versions 2
In Tensorflow before versions 2.2.1 and 2.3.1, the implementation of `dlpack.to_dlpack` can be made to use uninitialized memory resulting in further memory corruption. This is because the pybind11 glue code assumes that the argument is a tensor. However, there is nothing stopping users from passing in a Python object instead of a tensor. The uninitialized memory address is due to a `reinterpret_cast` Since the `PyObject` is a Python object, not a TensorFlow Tensor, the cast to `EagerTensor` fails. The issue is patched in commit 22e07fb204386768e5bcbea563641ea11f96ceb8 and is released in TensorFlow versions 2.2.1, or 2.3.1.
GHSA
Memory corruption in Tensorflow
ghsa·2020-09-25
CVE-2020-15193 [HIGH] CWE-908 Memory corruption in Tensorflow
Memory corruption in Tensorflow
### Impact
The implementation of `dlpack.to_dlpack` can be made to use uninitialized memory resulting in further memory corruption. This is because the pybind11 glue code assumes that the argument is a tensor:
https://github.com/tensorflow/tensorflow/blob/0e68f4d3295eb0281a517c3662f6698992b7b2cf/tensorflow/python/tfe_wrapper.cc#L1361
However, there is nothing stopping users from passing in a Python object instead of a tensor.
```python
In [2]: tf.experimental.dlpack.to_dlpack([2])
==1720623==WARNING: MemorySanitizer: use-of-uninitialized-value
#0 0x55b0ba5c410a in tensorflow::(anonymous namespace)::GetTensorFromHandle(TFE_TensorHandle*, TF_Status*) third_party/tensorflow/c/eager/dlpack.cc:46:7
#1 0x55b0ba5c38f4 in tensorflow::TFE_HandleToDLPack(TFE_TensorH
Debian
CVE-2020-15193: tensorflow - In Tensorflow before versions 2.2.1 and 2.3.1, the implementation of `dlpack.to_...
vendor_debian·2020·CVSS 7.1
CVE-2020-15193 [HIGH] CVE-2020-15193: tensorflow - In Tensorflow before versions 2.2.1 and 2.3.1, the implementation of `dlpack.to_...
In Tensorflow before versions 2.2.1 and 2.3.1, the implementation of `dlpack.to_dlpack` can be made to use uninitialized memory resulting in further memory corruption. This is because the pybind11 glue code assumes that the argument is a tensor. However, there is nothing stopping users from passing in a Python object instead of a tensor. The uninitialized memory address is due to a `reinterpret_cast` Since the `PyObject` is a Python object, not a TensorFlow Tensor, the cast to `EagerTensor` fails. The issue is patched in commit 22e07fb204386768e5bcbea563641ea11f96ceb8 and is released in TensorFlow versions 2.2.1, or 2.3.1.
Scope: local
forky: resolved
sid: resolved
No detection rules found.
No public exploits indexed.
No writeups or analysis indexed.
http://lists.opensuse.org/opensuse-security-announce/2020-10/msg00065.htmlhttps://github.com/tensorflow/tensorflow/commit/22e07fb204386768e5bcbea563641ea11f96ceb8https://github.com/tensorflow/tensorflow/releases/tag/v2.3.1https://github.com/tensorflow/tensorflow/security/advisories/GHSA-rjjg-hgv6-h69vhttp://lists.opensuse.org/opensuse-security-announce/2020-10/msg00065.htmlhttps://github.com/tensorflow/tensorflow/commit/22e07fb204386768e5bcbea563641ea11f96ceb8https://github.com/tensorflow/tensorflow/releases/tag/v2.3.1https://github.com/tensorflow/tensorflow/security/advisories/GHSA-rjjg-hgv6-h69v
2020-09-25
Published