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_CLIENT_H
      12              : #define HCCLV2_RANK_INFO_DETECT_CLIENT_H
      13              : 
      14              : #include <cstddef>
      15              : #include <functional>
      16              : #include <vector>
      17              : #include "socket.h"
      18              : #include "new_rank_info.h"
      19              : #include "rank_table_info.h"
      20              : #include "json_parser.h"
      21              : #include "internal_exception.h"
      22              : #include "timeout_exception.h"
      23              : #include "socket_exception.h"
      24              : #include "socket_agent.h"
      25              : #include "root_handle_v2.h"
      26              : #include "ip_address.h"
      27              : 
      28              : namespace Hccl {
      29              : 
      30              : #ifdef OPEN_HCCL_TEST
      31              : const u32 WAIT_ERROR_BROADCAST_TIME = 1;
      32              : #else
      33              : const u32 WAIT_ERROR_BROADCAST_TIME = 20;
      34              : #endif
      35              : 
      36              : class RankInfoDetectClient {
      37              : public:
      38           60 :     RankInfoDetectClient(u32 devPhyId, u32 rankSize, u32 rankId, const std::shared_ptr<Socket>& clientSocket)
      39           60 :         : devPhyId_(devPhyId),
      40           60 :           rankSize_(rankSize),
      41           60 :           rankId_(rankId),
      42           60 :           clientSocket_(clientSocket),
      43           60 :           socketAgent_(clientSocket.get())
      44           60 :     {}
      45              :     ~RankInfoDetectClient();
      46              : 
      47              :     void Setup(RankTableInfo& rankTable);
      48              : 
      49              : private:
      50              :     u32 devPhyId_{0};
      51              :     u32 rankSize_{0};
      52              :     u32 rankId_{0};
      53              :     std::shared_ptr<Socket> clientSocket_{nullptr};
      54              :     u32 currentStep_{0};
      55              :     RankTableInfo rankTable_{};
      56              :     SocketAgent socketAgent_;
      57              :     std::mutex hostSocketLock_;
      58              :     std::shared_ptr<Socket> hostSocket_ = nullptr;
      59              :     bool hostSocketRegistered_ = false; // 标记 hostSocket_ 是否已登记到 SocketManager::GetServerSocketMap()
      60              : 
      61              :     void SetupHostListenPort(u32 devLogicId, u32 devPhyId, const IpAddress& hostIp, uint32_t& hostPort);
      62              :     void SocketTearDown(u32 devPhyId);
      63              :     void Connect();
      64              :     void CheckStatus();
      65              :     void SendAgentIdAndRankSize();
      66              :     void SendLocalRankTable(const RankTableInfo& localRankTable);
      67              :     void ConstructRankTable(RankTableInfo& localRankTable);
      68              :     void VerifyRankTable();
      69              :     void RecvRankTable();
      70              :     void RecvRankTableMsg(vector<char>& rankInfoMsg);
      71              :     void ParseRankTable(vector<char>& rankInfoMsg);
      72              :     void GetLocalRankTableJson(const nlohmann::json& parseJson, nlohmann::json& localRankTableJson);
      73              :     void GetLocalDevInfoJson(const nlohmann::json& parseJson, nlohmann::json& localDevInfoJson);
      74              :     void SelectLocalHostBackupAddr(nlohmann::json& localDevInfoJson);
      75              :     void SelectAvailableHostAddr(nlohmann::json& addrJson);
      76              :     void UpdateSelectedHostAddr(
      77              :         nlohmann::json& addrJson, const std::vector<IpAddress>& candidates, std::size_t selectedIndex) const;
      78              :     void ProbeHostRoceAddr(const IpAddress& hostAddr, bool& isAvailable) const;
      79              :     void ConstructSingleRank(RankTableInfo& localRankTable);
      80              :     HcclResult GetLocalTlsStatus(TlsStatus& tlsStatus) const;
      81              :     HcclResult GetLocalHostDpuTlsStatus(TlsStatus& tlsStatus) const;
      82              :     HcclResult VerifyTlsConsistency() const;
      83              :     HcclResult VerifyHostDpuTlsConsistency() const;
      84              :     HcclResult VerifyTlsConsistencyByTlsType(
      85              :         const std::string& tlsType, const std::function<TlsStatus(const NewRankInfo&)>& getTlsStatus) const;
      86              :     void GenerateTlsStatusStr(std::string& tlsStatusStr, const std::vector<u32>& tlsStatusRanks) const;
      87              :     void ReportTlsConfigurationError(
      88              :         const std::string& tlsType, const std::string& tlsInconsistentTlsType, const std::string& tlsEnableRankStr,
      89              :         const std::string& tlsDisableRankStr, const std::string& tlsUnknownRankStr) const;
      90              :     void TearDown();
      91              :     void HostListenPortDetect(NewRankInfo& rankInfo);
      92              : };
      93              : 
      94              : } // namespace Hccl
      95              : #endif // HCCLV2_RANK_INFO_DETECT_CLIENT_H