Bug 2398198 (CVE-2025-55559)

Summary: CVE-2025-55559 tensorflow: Conv2D with 'valid' padding causes XLA compile crash leading to denial of service
Product: [Other] Security Response Reporter: OSIDB Bzimport <bzimport>
Component: vulnerabilityAssignee: Product Security DevOps Team <prodsec-dev>
Status: NEW --- QA Contact:
Severity: medium Docs Contact:
Priority: medium    
Version: unspecifiedCC: caswilli, jkoehler, kaycoth, lphiri
Target Milestone: ---Keywords: Security
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: ---
Doc Text:
A flaw was found in TensorFlow. When a model uses tf.keras.layers.Conv2D with padding='valid' and is compiled using XLA, the compiler miscalculates the output shape and ends up with a negative dimension. This causes the process to crash during compilation, leading to a denial of service. The same operation works correctly in eager mode because it checks dimensions dynamically, but under XLA the static shape calculation fails and stops execution.
Story Points: ---
Clone Of: Environment:
Last Closed: Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description OSIDB Bzimport 2025-09-25 16:01:46 UTC
An issue was discovered TensorFlow v2.18.0. A Denial of Service (DoS) occurs when padding is set to 'valid' in tf.keras.layers.Conv2D.