Line data    Source code

       1              : /**
       2              :  * Copyright (c) 2025 Huawei Technologies Co., Ltd.
       3              :  * This program is free software, you can redistribute it and/or modify it under the terms and conditions of
       4              :  * CANN Open Software License Agreement Version 2.0 (the "License").
       5              :  * Please refer to the License for details. You may not use this file except in compliance with the License.
       6              :  * THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
       7              :  * INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
       8              :  * See LICENSE in the root of the software repository for the full text of the License.
       9              :  */
      10              : 
      11              : #ifndef HCCLV2_RANK_INFO_DETECT_SERVICE_H
      12              : #define HCCLV2_RANK_INFO_DETECT_SERVICE_H
      13              : 
      14              : #include "socket.h"
      15              : #include <unordered_map>
      16              : #include "new_rank_info.h"
      17              : #include "rank_table_info.h"
      18              : #include "json_parser.h"
      19              : #include "internal_exception.h"
      20              : #include "timeout_exception.h"
      21              : #include "socket_exception.h"
      22              : #include "socket_agent.h"
      23              : 
      24              : namespace Hccl {
      25              : 
      26              : class RankInfoDetectService {
      27              : public:
      28           30 :     RankInfoDetectService(u32 devPhyId, std::shared_ptr<Socket> serverSocket, 
      29              :         std::string identifier, vector<RaSocketWhitelist> wlistInfo) 
      30           60 :         : devPhyId_(devPhyId), serverSocket_(serverSocket), hostIp_(serverSocket_->GetLocalIp()),
      31           60 :           identifier_(identifier), wlistInfo_(wlistInfo)
      32              :     {
      33           30 :     }
      34              :     ~RankInfoDetectService();
      35              : 
      36              :     void Setup();
      37              : 
      38              : private:
      39              :     u32                                                      devPhyId_{0};
      40              :     shared_ptr<Socket>                                       serverSocket_{nullptr};
      41              :     std::unordered_map<std::string, std::shared_ptr<Socket>> connSockets_{};
      42              :     IpAddress                                                hostIp_{};
      43              :     RankTableInfo                                            rankTable_{};
      44              :     std::string                                              failedAgentIdList_{};
      45              :     u32                                                      currentStep_{0};
      46              :     std::string                                              identifier_{};
      47              :     vector<RaSocketWhitelist>                                wlistInfo_{};
      48              : 
      49              :     void GetConnections();
      50              :     void GetRankTable();
      51              :     void BroadcastRankTable();
      52              :     void Disconnect();
      53              :     void TearDown();
      54              : 
      55              :     bool RecvRemoteAgentId(SocketAgent &connSocketAgent, std::string &agentId);
      56              :     bool RecvRemoteRankSize(SocketAgent &connSocketAgent, u32 &rankSize);
      57              :     void RecvRankInfoMsg(SocketAgent &connSocketAgent, vector<char> &rankInfoMsg);
      58              :     void SendRankTable(Socket *connSocket);
      59              :     void SortRankTable();
      60              :     void ParseRankTable(vector<char> &rankInfoMsg);
      61              : 
      62              :     // 异常流程处理方法
      63              :     void FailedConnectionAgentIdString(u32 rankSize);
      64              : 
      65              :     // 校验相关方法
      66              :     bool RecvAndVerifyRemoteAgentIdAndRankSize(
      67              :         std::shared_ptr<Socket> connSocket, u32 &expectedSocketNum, u32 &previousRankSize);
      68              :     bool VerifyRemoteRankSize(u32 &previousRankSize, u32 remoteRankSize) const;
      69              : 
      70              :     // DFX相关方法
      71              :     void DisplayConnectedRank(const std::map<std::string, std::shared_ptr<Socket>> &connectSockets);
      72              :     void DisplayConnectingStatus(u32 totalSockets, u32 waitSockets);
      73              :     void DisplayConnectedRanks();
      74              : };
      75              : 
      76              : }  // namespace Hccl
      77              : #endif  // HCCLV2_RANK_INFO_DETECT_SERVICE_H