shadesmar

Shadesmar

使用系统共享内存传递消息的IPC库。支持Publish-Subscribe和RPC。

要求：Linux和X86。注意：alpha软件。

特征

• 多个订户和发布者。

• 使用圆形缓冲区在过程之间传递消息。

• 比使用网络堆栈快。大型消息的高吞吐量，低延迟。

• 分散的，没有资源饥饿。

• 使用自定义复印机最小化或优化数据移动。

用法

从源代码生成一个单个标头文件，可以在此处找到。

如果要自己生成单个标头文件，请克隆回购并运行：

 $ cd shadesmar
$ python3 simul/simul.py

这将在include/中生成文件。

---

bublish-subscribe

出版商：

# include < shadesmar/pubsub/publisher.h >

int main () {
    shm::pubsub::Publisher p ( " topic_name " );
    const uint32_t data_size = 1024 ;
    void *data = malloc (data_size);
    
    for ( int i = 0 ; i < 1000 ; ++i) {
        p. publish (data, data_size);
    }
}

订户：

# include < shadesmar/pubsub/subscriber.h >

void callback (shm::memory::Memblock *msg) {
  // `msg->ptr` to access `data`
  // `msg->size` to access `data_size`

  // The memory will be free'd at the end of this callback.
  // Copy to another memory location if you want to persist the data.
  // Alternatively, if you want to avoid the copy, you can call
  // `msg->no_delete()` which prevents the memory from being deleted
  // at the end of the callback.
}

int main () {
    shm::pubsub::Subscriber sub ( " topic_name " , callback);

    // Using `spin_once` with a manual loop
    while ( true ) {
        sub. spin_once ();
    }
    // OR
    // Using `spin`
    sub. spin ();
}

---

RPC

客户：

# include < shadesmar/rpc/client.h >

int main () {
  Client client ( " channel_name " );
  shm::memory::Memblock req, resp;
  // Populate req.
  client. call (req, &resp);
  // Use resp here.

  // resp needs to be explicitly free'd.
  client. free_resp (&resp);
}

服务器：

# include < shadesmar/rpc/server.h >

bool callback ( const shm::memory::Memblock &req,
              shm::memory::Memblock *resp) {
  // resp->ptr is a void ptr, resp->size is the size of the buffer.
  // You can allocate memory here, which can be free'd in the clean-up lambda.
  return true ;
}

void clean_up (shm::memory::Memblock *resp) {
  // This function is called *after* the callback is finished. Any memory
  // allocated for the response can be free'd here. A different copy of the
  // buffer is sent to the client, this can be safely cleaned.
}

int main () {
  shm::rpc::Server server ( " channel_name " , callback, clean_up);

  // Using `serve_once` with a manual loop
  while ( true ) {
    server. serve_once ();
  }
  // OR
  // Using `serve`
  server. serve ();
}