Process crash via FlatBuffer-loaded model invocation
Published Apr 2, 2025 · Updated Apr 3, 2025
Memory corruption in torch.jit.jit_module_from_flatbuffer in PyTorch 2.6.0 allows local users to crash a process via a FlatBuffer module. The upstream reproduction saves a scripted model containing BatchNorm2d, reloads it through the affected function, and segfaults when the loaded model processes input. The public record does not identify the corrupting memory operation; low-privileged local code execution is required, and the demonstrated consequence is process termination.
Summary
What happened
Memory corruption in torch.jit.jit_module_from_flatbuffer in PyTorch 2.6.0 allows local users to crash a process via a FlatBuffer module. The upstream reproduction saves a scripted model containing BatchNorm2d, reloads it through the affected function, and segfaults when the loaded model processes input. The public record does not identify the corrupting memory operation; low-privileged local code execution is required, and the demonstrated consequence is process termination.
The record
- CVE
- CVE-2025-3121
- Published
- Apr 2, 2025
- Updated
- Apr 3, 2025
- Vendor
- Linux Foundation
- Product
- PyTorch
- Classifications
- CWE-119, T1499.004
- Attack vector
- local
- Privileges
- authenticated
Timeline
How it unfolded
- Apr 2, 2025CVE publishedPublication date reported by the CVE source.
- Apr 3, 2025Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=2.6.0
What conditions does exploitation require?
What is affected?
Published CVSS scores
CVSS describes severity. EPSS estimates exploitation probability.
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Weakness, pattern, technique
Public exploit references
- Segmentation fault in torch.jit.jit_module_from_flatbufferproof of concept · demonstrated
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo