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 CCU_URMA_CHANNEL_H
      12              : #define CCU_URMA_CHANNEL_H
      13              : 
      14              : #include <memory>
      15              : #include <vector>
      16              : #include <atomic>
      17              : #include <mutex>
      18              : #include <optional>
      19              : 
      20              : #include "../channel.h"
      21              : 
      22              : #include "urma_endpoint.h"
      23              : #include "ccu_transport_.h"
      24              : 
      25              : namespace hcomm {
      26              : 
      27              : class CcuUrmaChannel : public Channel {
      28              : public:
      29              :     // 当前仅支持交换hccl buffer
      30              :     CcuUrmaChannel(const EndpointHandle locEndpointHandle, const HcommChannelDesc& channelDesc);
      31              :     ~CcuUrmaChannel() override;
      32              : 
      33              :     HcclResult Init() override;
      34              :     ChannelStatus GetStatus() override;
      35              : 
      36              :     HcclResult GetNotifyNum(uint32_t* notifyNum) const override;
      37              :     HcclResult GetRemoteMems(uint32_t* memNum, CommMem** remoteMem, char*** memInfos) override;
      38              :     HcclResult UpdateMemInfo(HcommMemHandle* memHandles, uint32_t memHandleNum) override;
      39              : 
      40              :     HcclResult Clean() override;
      41              :     HcclResult Resume() override;
      42              : 
      43              :     HcclResult ChannelFence() override;
      44              :     HcclResult NotifyRecord(const uint32_t remoteNotifyIdx) override;
      45              :     HcclResult NotifyWait(const uint32_t localNotifyIdx, const uint32_t timeout) override;
      46              :     HcclResult WriteWithNotify(void* dst, const void* src, const uint64_t len, uint32_t remoteNotifyIdx) override;
      47              :     HcclResult Write(void* dst, const void* src, uint64_t len) override;
      48              :     HcclResult Read(void* dst, const void* src, uint64_t len) override;
      49              : 
      50              : public:
      51              :     uint32_t GetDieId() const;
      52              :     uint32_t GetChannelId() const;
      53              :     HcclResult GetRmtSignalAddrByIndex(uint32_t index, uint64_t& rmtCkeAddr) const;
      54              :     HcclResult GetRmtCcuBufferTokenInfo(uint32_t& rmtTokenId, uint32_t& rmtTokenValue) const;
      55              :     HcclResult GetRmtVarAddrByIndex(uint32_t index, uint64_t& rmtXnAddr) const;
      56              :     HcclResult GetLocCkeByIndex(const uint32_t index, uint32_t& locCkeId) const;
      57              :     HcclResult GetLocXnByIndex(const uint32_t index, uint32_t& locXnId) const;
      58              : 
      59              :     HcclResult GetRmtCkeByIndex(const uint32_t index, uint32_t& rmtCkeId) const;
      60              :     HcclResult GetRmtXnByIndex(const uint32_t index, uint32_t& rmtXnId) const;
      61              : 
      62              :     HcclResult GetRmtWishCntXnAddr(const std::string& resGroupTag, uint64_t& wishCntXnAddr) const;
      63              : 
      64              :     HcclResult GetRmtBuffer(uint64_t& addr, uint32_t& size, uint32_t& tokenId, uint32_t& tokenValue) const;
      65              : 
      66            0 :     EndpointHandle GetlocEndPointHandle() { return locEndpointHandle_; }
      67           15 :     const HcommChannelDesc& GetChannelDesc() const override { return channelDesc_; }
      68              : 
      69              : private:
      70              :     ChannelStatus TryPrepareAndConstruct();
      71              : 
      72              :     std::atomic<bool> isFirstPrintChannelInfo_{true}; // 是否第一次打印通道建链信息，避免重复打印日志刷屏
      73              :     std::unique_ptr<CcuTransport> impl_{nullptr};
      74              :     ChannelStatus channelStatus_{ChannelStatus::INIT};
      75              :     // 保护 GetStatus 懒建链与状态读写：并发轮询同一 channel 时避免 impl_/channelStatus_ 竞态
      76              :     std::mutex statusMtx_{};
      77              :     EndpointHandle locEndpointHandle_{nullptr};
      78              :     HcommChannelDesc channelDesc_{};
      79              :     UrmaEndpoint* ccuEndpoint_{nullptr};
      80              :     std::optional<Hccl::LinkData> linkData_;
      81              :     Hccl::Socket* socket_{nullptr};
      82              :     std::vector<HcommMemHandle> memHandles_{};
      83              : };
      84              : 
      85              : } // namespace hcomm
      86              : #endif // CCU_URMA_CHANNEL_H