erlang message passing architecture

旧街凉风 提交于 2019-11-28 09:12:48

问题


In erlang if two processes A and B are sending message to a process C simultaneously. Will there be a race condition?

  • C ! {very large message} sent by A
  • C ! {very large message} sent by B

Will C receive the complete message from A and then proceed for the message from B? or is it that C is likely be going to receive chunks of A's message along with chunks of B's message?


回答1:


Message receiving is an atomic operation.

If you are interested how it is done, read the source code of VM. If I simplify it, the sending process is doing those steps:

  1. Allocate a target memory space in the sending process (it's called environment).
  2. Copy the message to that memory space
  3. Take the external lock on the target process
  4. Link message into the mailbox linked list
  5. Release the external lock on the target process

As you can see, copying is done outside (before) critical section and the critical section is pretty fast. It is just juggling with few pointers.



来源:https://stackoverflow.com/questions/30710923/erlang-message-passing-architecture

易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!