CVE-2018-7577 Improper Input Validation in Google TensorFlow
A memcpy parameter overlap in the Snappy library can cause crashes or unauthorized memory reads, affecting applications using TensorFlow.
- Exploited in the wild
- Not known
- Login needed
- Not needed
- Attackable
- Over the internet
- Victim action needed
- A user has to click or open something
- Fix
- Available
What is CVE-2018-7577?
CVE-2018-7577 is a high-severity memory corruption vulnerability in Google Tensorflow, published on 24 Apr 2019. An attacker can exploit it over the internet without logging in, and only if a user opens a file or follows a link.
Official description
Memcpy parameter overlap in Google Snappy library 1.1.4, as used in Google TensorFlow before 1.7.1, could result in a crash or read from other parts of process memory.
Weakness: CWE-20
Which versions are affected by CVE-2018-7577?
7 open source packages ship an affected version, including tensorflow, tensorflow-cpu and tensorflow-gpu. It affects 2 products.
| Package | Affected | Fixed in |
|---|---|---|
tensorflow PyPI | All versions | 1.7.1 |
tensorflow PyPI | From 1.1.0 | 1.7.1 |
tensorflow PyPI | From 1.1.0 | 1.7.1 |
tensorflow-cpu PyPI | All versions | 1.7.1 |
tensorflow-gpu PyPI | All versions | 1.7.1 |
tensorflow-gpu PyPI | From 1.1.0 | 1.7.1 |
1 more packages
tensorflow-gpu PyPI | From 1.1.0 | 1.7.1 |
Products include Google Snappy, Google Tensorflow.
How do I fix CVE-2018-7577?
Upgrade tensorflow to 1.7.1 or later.
Upgrade to the fixed version. Memory-safe languages, compiler hardening and fuzzing prevent most new cases. More about Memory corruption.
Is CVE-2018-7577 being exploited?
There is no known exploitation of CVE-2018-7577 in the wild and no public exploit code.
Sources
Also known as GHSA-qx2v-j445-g354, PYSEC-2019-207, PYSEC-2019-225, PYSEC-2019-232.