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
      35              :     {
      36            7 :         return devicePhyId == other.devicePhyId && port == other.port;
      37              :     }
      38              : };
      39              : 
      40              : struct SocketMapKeyHash {
      41           13 :     size_t operator()(const SocketMapKey &k) const
      42              :     {
      43           13 :         return std::hash<uint32_t>()(k.devicePhyId) ^ (std::hash<uint32_t>()(k.port) << 1);
      44              :     }
      45              : };
      46              : 
      47              : struct AicpuTsNetDevSlot {
      48              :     HcclNetDev netDev{nullptr};
      49              :     uint32_t refCount{0U};
      50              : };
      51              : 
      52              : class AicpuTsRoceEndpoint : public Endpoint {
      53              : public:
      54              :     explicit AicpuTsRoceEndpoint(const EndpointDesc &endpointDesc);
      55              :     ~AicpuTsRoceEndpoint() override;
      56              : 
      57              :     HcclResult Init() override;
      58              : 
      59              :     HcclResult ServerSocketListen(const uint32_t port) override;
      60              : 
      61              :     HcclResult RegisterMemory(HcommMem mem, const char *memTag, void **memHandle) override;
      62              :     HcclResult UnregisterMemory(void* memHandle) override;
      63              :     HcclResult MemoryExport(void *memHandle, void **memDesc, uint32_t *memDescLen) override;
      64              :     HcclResult MemoryImport(const void *memDesc, uint32_t descLen, HcommMem *outMem) override;
      65              :     HcclResult MemoryUnimport(const void *memDesc, uint32_t descLen) override;
      66              :     HcclResult GetAllMemHandles(void **memHandles, uint32_t *memHandleNum) override;
      67              : 
      68            3 :     HcclNetDev GetNetDev() const { return netDev_; }
      69              : 
      70              :     HcclResult GetSocket(uint32_t port, const std::string &tag,
      71              :         std::shared_ptr<hccl::HcclSocket> &outConnected);
      72              :     HcclResult AcceptDataSocket(uint32_t port, const std::string &tag,
      73              :         std::shared_ptr<hccl::HcclSocket> &outConnected, uint32_t acceptTimeoutMs = 0);
      74              : 
      75              :     HcclResult AddListenSocketWhiteList(uint32_t port, const std::vector<SocketWlistInfo> &wlistInfos);
      76              : 
      77              : private:
      78              :     static std::unordered_map<SocketMapKey, AicpuTsListenSocketSlot, SocketMapKeyHash> &GetServerSocketMap();
      79              :     static std::mutex &ListenSocketMapMutex();
      80              :     bool ReuseListenSocketIfExist(const SocketMapKey &key, const char *logPrefix);
      81              :     void ReleaseListenSocketRefs();
      82              : 
      83              :     static std::unordered_map<uint32_t, AicpuTsNetDevSlot> &GetNetDevMap();
      84              :     static std::mutex &NetDevMapMutex();
      85              :     HcclResult AcquireSharedNetDev(uint32_t devicePhyId, const HcclNetDevInfos &info);
      86              :     void ReleaseSharedNetDev();
      87              :     void ReleaseNicSocketHandle(HcclNetDev netDev);
      88              :     HcclResult AcquireRdmaContext(uint32_t devPhyId, const EndpointDesc &endpointDesc);
      89              : 
      90              :     HcclNetDev netDev_{nullptr};
      91              :     uint32_t netDevRefPhyId_{UINT32_MAX};
      92              :     std::vector<SocketMapKey> listenRefKeys_{};
      93              :     bool hasListenSocketRef_{false};
      94              : };
      95              : }
      96              : #endif // AICPUTS_ROCE_ENDPOINT_H