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              : #ifndef P2P_TRANSPORT_H
      11              : #define P2P_TRANSPORT_H
      12              : 
      13              : #include "base_mem_transport.h"
      14              : #include "virtual_topo.h"
      15              : #include "ipc_remote_notify.h"
      16              : #include "../../../resource/buffer/local_ipc_rma_buffer.h"
      17              : #include "remote_rma_buffer.h"
      18              : 
      19              : namespace Hccl {
      20              : class P2PTransport : public BaseMemTransport {
      21              : public:
      22              :     P2PTransport(CommonLocRes &commonLocRes, Attribution &attr, const LinkData &linkData, const Socket &socket);
      23              : 
      24              :     P2PTransport(CommonLocRes &commonLocRes, Attribution &attr, const LinkData &linkData, const Socket &socket, std::function<void(u32 streamId, u32 taskId, TaskParam taskParam)> callback);
      25              : 
      26           29 :     ~P2PTransport() override = default;
      27              : 
      28              :     std::string Describe() const override;
      29              : 
      30              :     TransportStatus GetStatus() override;
      31              : 
      32              :     std::vector<char> GetUniqueId() override;
      33              : 
      34              :     std::vector<char> GetUniqueIdV2();
      35              : 
      36              :     void Post(u32 index, const Stream &stream) override;
      37              : 
      38              :     void Read(const RmaBufferSlice &locSlice, const RmtRmaBufferSlice &rmtSlice, const Stream &stream) override;
      39              : 
      40              :     void ReadReduce(const RmaBufferSlice &locSlice, const RmtRmaBufferSlice &rmtSlice, const ReduceIn &reduceIn,
      41              :                     const Stream &stream) override;
      42              : 
      43              :     void Write(const RmaBufferSlice &locSlice, const RmtRmaBufferSlice &rmtSlice, const Stream &stream) override;
      44              : 
      45              :     void WriteReduce(const RmaBufferSlice &locSlice, const RmtRmaBufferSlice &rmtSlice, const ReduceIn &reduceIn,
      46              :                      const Stream &stream) override;
      47              : 
      48              :     HcclResult GetRemoteMems(uint32_t *memNum, CommMem **remoteMem, char ***memInfos);
      49              : 
      50              : private:
      51              :     MemoryBuffer GetLocMemBuffer(const RmaBufferSlice &locSlice) const;
      52              :     MemoryBuffer GetRmtMemBuffer(const RmtRmaBufferSlice &rmtSlice) const;
      53              : 
      54           19 :     MAKE_ENUM(P2PStatus, INIT, SOCKET_OK, SEND_PID, RECV_PID, GRANT, SEND_DATA, RECV_DATA, SEND_DATA_SIZE, RECV_DATA_SIZE)
      55              :     P2PStatus p2pStatus{P2PStatus::INIT};
      56              : 
      57              :     u32 pidMsgSize{0};
      58              :     u32 myPid{0};
      59              :     u32 rmtPid{0};
      60              : 
      61              :     bool rmtPidValid{false};
      62              : 
      63              :     std::vector<std::unique_ptr<IpcRemoteNotify>>    rmtNotifyVec;
      64              :     std::vector<std::unique_ptr<RemoteIpcRmaBuffer>> rmtBufferVec;
      65              : 
      66              :     bool                         cacheValid_ = false; // 当前缓存是否有效
      67              :     std::vector<CommMem>         remoteUserMems_;     // 内存基本信息缓存
      68              :     std::vector<std::string>     memInfoCopies_;          // 储存 Tag 字符串副本
      69              :     std::vector<char*>           memInfoPointers_;        // Tag 缓存
      70              :     std::vector<char> sendData;
      71              :     std::vector<char> recvData;
      72              : 
      73              :     bool IsRmtPidValid() const;
      74              :     void SendPid();
      75              :     void RecvPid();
      76              :     void Grant();
      77              :     void PrepareSendData();
      78              :     void RecvDataSize();
      79              :     void SendExchangeData();
      80              :     void RecvExchangeData();
      81              :     void ProcessRecvData();
      82              : 
      83              :     void BufferVecPack(BinaryStream &binaryStream);
      84              : 
      85              :     void RmtNotifyVecUnpackProc(BinaryStream &binaryStream);
      86              :     void RmtBufferVecUnpackProc(BinaryStream &binaryStream);
      87              : 
      88              :     std::vector<char> GetSingleRmtNotifyUniqueId(u64 addr, u64 size, u32 notifyId) const;
      89              :     std::vector<char> GetSingleBufferUniqueId(u64 addr, u64 size) const;
      90              :     std::vector<char> GetNotifyUniqueIds();
      91              :     std::vector<char> GetRmtNotifyUniqueIds() const;
      92              :     std::vector<char> GetLocBufferUniqueIds() const;
      93              :     std::vector<char> GetRmtBufferUniqueIds() const;
      94              : 
      95              :     std::mutex      remoteMemsMutex_; // 远端内存列表互斥锁
      96              : };
      97              : 
      98              : } // namespace Hccl
      99              : 
     100              : #endif