Bug 2291225 (CVE-2024-5480)

Summary: CVE-2024-5480 python-pytorch: Remoce Code Execution vulnerability in torch.distributed.rpc framework
Product: [Other] Security Response Reporter: Marco Benatto <mbenatto>
Component: vulnerabilityAssignee: Product Security <prodsec-ir-bot>
Status: NEW --- QA Contact:
Severity: high Docs Contact:
Priority: high    
Version: unspecifiedKeywords: Security
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: pytorch torch.distributed.rpc 2.2.2 Doc Type: If docs needed, set a value
Doc Text:
A vulnerability in PyTorch's torch.distributed.rpc framework, specifically in versions prior to 2.2.2, allows for remote code execution (RCE). The framework, which is used in distributed training scenarios, does not properly verify the functions being called during RPC (Remote Procedure Call) operations. This oversight permits attackers to execute arbitrary commands by leveraging built-in Python functions such as eval during multi-cpu RPC communication. The vulnerability arises from the lack of restriction on function calls when a worker node serializes and sends a PythonUDF (User Defined Function) to the master node, which then deserializes and executes the function without validation. This flaw can be exploited to compromise master nodes initiating distributed training, potentially leading to the theft of sensitive AI-related data.
 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:
Bug Depends On: 2291226    
Bug Blocks:    

Description Marco Benatto 2024-06-10 21:02:34 UTC 
A vulnerability in the PyTorch's torch.distributed.rpc framework, specifically in versions prior to 2.2.2, allows for remote code execution (RCE). The framework, which is used in distributed training scenarios, does not properly verify the functions being called during RPC (Remote Procedure Call) operations. This oversight permits attackers to execute arbitrary commands by leveraging built-in Python functions such as eval during multi-cpu RPC communication. The vulnerability arises from the lack of restriction on function calls when a worker node serializes and sends a PythonUDF (User Defined Function) to the master node, which then deserializes and executes the function without validation. This flaw can be exploited to compromise master nodes initiating distributed training, potentially leading to the theft of sensitive AI-related data.

https://huntr.com/bounties/39811836-c5b3-4999-831e-46fee8fcade3
Comment 1 Marco Benatto 2024-06-10 21:02:46 UTC 
Created python-torch tracking bugs for this issue:

Affects: fedora-40 [bug 2291226]