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 AICPU_TS_HCCS_CHANNEL_H
      11              : #define AICPU_TS_HCCS_CHANNEL_H
      12              : #include <memory>
      13              : #include <vector>
      14              : #include "hccl_mem_defs.h"
      15              : #include "transport_mem.h"
      16              : #include "hccl_res.h"
      17              : #include "hccl_socket.h"
      18              : #include "channel.h"
      19              : #include "aicpu_ts_channel_helper.h"
      20              : #include "channel_param.h"
      21              : #include "buffer.h"
      22              : #include "hccl_dispatcher_ctx.h"
      23              : #include "rma_buffer.h"
      24              : #include "hcomm_c_adpt.h"
      25              : #include "../../../endpoints/aicputs_hccs_endpoint.h"
      26              : 
      27              : namespace hcomm {
      28              : /**
      29              :  * @note 职责：Channel的AicpuTs通信引擎、HCCS协议的类派生
      30              :  */
      31              : class AicpuTsHccsChannel : public Channel {
      32              : public:
      33              :     struct HccsExchangeInfo {
      34              :         s32 memNum = 0;
      35              :     };
      36              :     AicpuTsHccsChannel(EndpointHandle endpointHandle, const HcommChannelDesc &channelDesc);
      37              :     virtual ~AicpuTsHccsChannel();
      38              : 
      39              :     HcclResult Init() override;
      40              :     HcclResult GetNotifyNum(uint32_t *notifyNum) const override;
      41              :     HcclResult GetRemoteMems(uint32_t *memNum, CommMem **remoteMem, char ***memInfos) override;
      42              :     ChannelStatus GetStatus() override;
      43              : 
      44              :     std::shared_ptr<hccl::Transport> GetTransport() {return transport_;}
      45              :     HcclResult Clean() override;
      46              :     HcclResult Resume() override;
      47            0 :     const HcommChannelDesc& GetChannelDesc() const override { return channelDesc_; }
      48              : 
      49              :     // for launch channel kernel data
      50              :     HcclResult Serialize(std::shared_ptr<hccl::DeviceMem> &out) override;
      51              :     HcommChannelKind GetChannelKind() const override;
      52              : 
      53              :     // 数据面接口
      54              :     HcclResult NotifyRecord(const uint32_t remoteNotifyIdx) override;
      55              :     HcclResult NotifyWait(const uint32_t localNotifyIdx, const uint32_t timeout) override;
      56              :     HcclResult WriteWithNotify(void *dst, const void *src, const uint64_t len, uint32_t remoteNotifyIdx) override;
      57              :     HcclResult Write(void *dst, const void *src, uint64_t len) override;
      58              :     HcclResult Read(void *dst, const void *src, uint64_t len) override;
      59              :     HcclResult ChannelFence() override;
      60              : 
      61            2 :     AicpuTsChannelHelper *GetAicpuTsHelper() override { return &aicpuTsHelper_; }
      62              : 
      63              : private:
      64              :     AicpuTsChannelHelper aicpuTsHelper_;
      65              :     HcclResult ParseInputParam();
      66              :     HcclResult EnableP2P();
      67              :     void DisableP2P();
      68              :     HcclResult BuildConnection();
      69              :     void DestroyConnection();
      70              :     HcclResult EnableMemAccess();
      71              :     void DisableMemAccess();
      72              :     HcclResult GetFirstIpByPhyId(u32 devicePhyId, u32 superDevId, hccl::HcclIpAddress &ip);
      73              :     HcclResult SetMachinePara(hccl::MachinePara &machinePara);
      74              :     void SetTransportParam(hccl::TransportPara &para);
      75              :     HcclResult TransportInit();
      76              :     void TransportDeInit();
      77              :     HcclResult BuildHcclChannelHccsRes(HcclChannelHccsRes &channelHccsRes);
      78              : 
      79              : private:
      80              :     // --------------------- 入参 ---------------------
      81              :     EndpointHandle                                              endpointHandle_{nullptr};
      82              :     HcommChannelDesc                                            channelDesc_;
      83              : 
      84              :     // --------------------- 转换参数 ---------------------
      85              :     EndpointDesc                                                localEp_{};
      86              :     EndpointDesc                                                remoteEp_{};
      87              :     hccl::HcclIpAddress                                         localIp_;
      88              :     hccl::HcclIpAddress                                         remoteIp_;
      89              :     uint32_t                                                    notifyNum_{0};
      90              :     AicpuTsHccsEndpoint                                         *localEpPtr_{nullptr};
      91              :     uint32_t                                                    serverPort_{AICPU_CHANNEL_DEFAULT_PORT};
      92              :     bool                                                        serverInited_{false};
      93              :     // --------------------- 具体成员 ---------------------
      94              :     std::shared_ptr<hccl::HcclSocket>                           socket_{nullptr};
      95              :     std::string                                                 socketTag_{};
      96              :     bool                                                        isSocketServer_{false};
      97              :     // for create TransportMem
      98              :     HcclDispatcher                                              dispatcher_{nullptr}; // dispatcher放到最后析构
      99              :     DispatcherCtxPtr                                            dispatcherCtx_{nullptr};
     100              :     std::unique_ptr<hccl::NotifyPool>                           notifyPool_;
     101              :     std::shared_ptr<hccl::Transport>                            transport_{nullptr};
     102              : 
     103              :     // for get mem temp
     104              :     std::vector<CommMem>                                        remoteIpcRmaBufferVec_;
     105              :     std::vector<HcclMemEx>                                      localIpcRmaBufferVecEx_;
     106              :     std::vector<HcclMemEx>                                      remoteIpcRmaBufferVecEx_;
     107              : };
     108              : }
     109              : 
     110              : #endif // AICPU_TS_HCCS_CHANNEL_H