CVE-2025-4287

A vulnerability was found in PyTorch 2.6.0+cu124. It has been rated as problematic. Affected by this issue is the function torch.cuda.nccl.reduce of the file torch/cuda/nccl.py. The manipulation leads to denial of service. It is possible to launch the attack on the local host. The exploit has been disclosed to the public and may be used. The patch is identified as 5827d2061dcb4acd05ac5f8e65d8693a481ba0f5. It is recommended to apply a patch to fix this issue.

Weakness

The product does not release or incorrectly releases a resource before it is made available for re-use.

Potential Mitigations

• Use a language that does not allow this weakness to occur or provides constructs that make this weakness easier to avoid.
• For example, languages such as Java, Ruby, and Lisp perform automatic garbage collection that releases memory for objects that have been deallocated.

Related Attack Patterns

• https://cwe.mitre.org/data/definitions/125.html
• https://cwe.mitre.org/data/definitions/130.html
• https://cwe.mitre.org/data/definitions/131.html
• https://cwe.mitre.org/data/definitions/494.html
• https://cwe.mitre.org/data/definitions/495.html
• https://cwe.mitre.org/data/definitions/496.html
• https://cwe.mitre.org/data/definitions/666.html

References

• https://github.com/Divigroup-RAP/PYTORCH/commit/5827d2061dcb4acd05ac5f8e65d8693a481ba0f5
• https://github.com/pytorch/pytorch/issues/150836
• https://github.com/pytorch/pytorch/issues/150836#issue-2979097872
• https://github.com/pytorch/pytorch/pull/150923
• https://vuldb.com/?ctiid.307394
• https://vuldb.com/?id.307394
• https://vuldb.com/?submit.553644