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 AICPUTS_ROCE_ENDPOINT_H
      11              : #define AICPUTS_ROCE_ENDPOINT_H
      12              : 
      13              : #include <cstdint>
      14              : #include <functional>
      15              : #include <memory>
      16              : #include <mutex>
      17              : #include <string>
      18              : #include <unordered_map>
      19              : #include <vector>
      20              : #include "hccl_mem_defs.h"
      21              : #include "endpoint.h"
      22              : #include "hccl_socket.h"
      23              : 
      24              : namespace hcomm {
      25              : struct AicpuTsListenSocketSlot {
      26              :     std::shared_ptr<hccl::HcclSocket> socket{};
      27              :     uint32_t refCount{0U};
      28              : };
      29              : 
      30              : struct SocketMapKey {
      31              :     uint32_t devicePhyId;
      32              :     uint32_t port;
      33              : 
      34            7 :     bool operator==(const SocketMapKey& other) const { return devicePhyId == other.devicePhyId && port == other.port; }
      35              : };
      36              : 
      37              : struct SocketMapKeyHash {
      38           13 :     size_t operator()(const SocketMapKey& k) const
      39              :     {
      40           13 :         return std::hash<uint32_t>()(k.devicePhyId) ^ (std::hash<uint32_t>()(k.port) << 1);
      41              :     }
      42              : };
      43              : 
      44              : struct AicpuTsNetDevSlot {
      45              :     HcclNetDev netDev{nullptr};
      46              :     uint32_t refCount{0U};
      47              : };
      48              : 
      49              : class AicpuTsRoceEndpoint : public Endpoint {
      50              : public:
      51              :     explicit AicpuTsRoceEndpoint(const EndpointDesc& endpointDesc);
      52              :     ~AicpuTsRoceEndpoint() override;
      53              : 
      54              :     HcclResult Init() override;
      55              : 
      56              :     HcclResult ServerSocketListen(const uint32_t port) override;
      57              : 
      58              :     HcclResult RegisterMemory(HcommMem mem, const char* memTag, void** memHandle) override;
      59              :     HcclResult UnregisterMemory(void* memHandle) override;
      60              :     HcclResult MemoryExport(void* memHandle, void** memDesc, uint32_t* memDescLen) override;
      61              :     HcclResult MemoryImport(const void* memDesc, uint32_t descLen, HcommMem* outMem) override;
      62              :     HcclResult MemoryUnimport(const void* memDesc, uint32_t descLen) override;
      63              :     HcclResult GetAllMemHandles(void** memHandles, uint32_t* memHandleNum) override;
      64              : 
      65            3 :     HcclNetDev GetNetDev() const { return netDev_; }
      66              : 
      67              :     HcclResult GetSocket(uint32_t port, const std::string& tag, std::shared_ptr<hccl::HcclSocket>& outConnected);
      68              :     HcclResult AcceptDataSocket(
      69              :         uint32_t port, const std::string& tag, std::shared_ptr<hccl::HcclSocket>& outConnected,
      70              :         uint32_t acceptTimeoutMs = 0);
      71              : 
      72              :     HcclResult AddListenSocketWhiteList(uint32_t port, const std::vector<SocketWlistInfo>& wlistInfos);
      73              : 
      74              : private:
      75              :     static std::unordered_map<SocketMapKey, AicpuTsListenSocketSlot, SocketMapKeyHash>& GetServerSocketMap();
      76              :     static std::mutex& ListenSocketMapMutex();
      77              :     bool ReuseListenSocketIfExist(const SocketMapKey& key, const char* logPrefix);
      78              :     void ReleaseListenSocketRefs();
      79              : 
      80              :     static std::unordered_map<uint32_t, AicpuTsNetDevSlot>& GetNetDevMap();
      81              :     static std::mutex& NetDevMapMutex();
      82              :     HcclResult AcquireSharedNetDev(uint32_t devicePhyId, const HcclNetDevInfos& info);
      83              :     void ReleaseSharedNetDev();
      84              :     void ReleaseNicSocketHandle(HcclNetDev netDev);
      85              :     HcclResult AcquireRdmaContext(uint32_t devPhyId, const EndpointDesc& endpointDesc);
      86              : 
      87              :     HcclNetDev netDev_{nullptr};
      88              :     uint32_t netDevRefPhyId_{UINT32_MAX};
      89              :     std::vector<SocketMapKey> listenRefKeys_{};
      90              :     bool hasListenSocketRef_{false};
      91              : };
      92              : } // namespace hcomm
      93              : #endif // AICPUTS_ROCE_ENDPOINT_H