Support c++20 coroutine

[wwbmmm]

What problem does this PR solve?

Issue Number:

Problem Summary: Support c++20 coroutine

What is changed and the side effects?

Changed: New feature

Side effects:

• Performance effects(性能影响):

  • Use example/coroutine, start server with taskset -c 1 ./coroutine_server -enable_coroutine=1 -sleep_us=100, start rpc_press with ../../tools/rpc_press/rpc_press -proto=./echo.proto -method=example.EchoService.Proxy -input='{"message":"hello"}' -qps=0 -thread_num=100 -duration=0 -server=0.0.0.0:8000 -timeout_ms=2000 -connection_type=pooled, the result is :
  • When start coroutine_server with -enable_coroutine=0, the result is :
  • Conclusion: the coroutine version's performance is 70%+ better than the classical sync version.

• Breaking backward compatibility(向后兼容性): No

---

Check List:

• Please make sure your changes are compilable(请确保你的更改可以通过编译).
• When providing us with a new feature, it is best to add related tests(如果你向我们增加一个新的功能, 请添加相关测试).
• Please follow Contributor Covenant Code of Conduct.(请遵循贡献者准则).

[chenzhangyi]

@wwbmmm 你可以加一个文档大概说一下实现原理和使用方式么(和当前直接用bthread的区别和注意事项)

[wwbmmm]

@wwbmmm 你可以加一个文档大概说一下实现原理和使用方式么(和当前直接用bthread的区别和注意事项)

好的

[wwbmmm]

@wwbmmm 你可以加一个文档大概说一下实现原理和使用方式么(和当前直接用bthread的区别和注意事项)

已补充文档

[wwbmmm]

加了experimental命名空间

[stdpain]

如果调整系统的mmap_max_count，可以达到coroutine一样的效果吗

[wwbmmm]

如果调整系统的mmap_max_count，可以达到coroutine一样的效果吗

这样只是增加了bthread数量的上限。但是bthread的堆栈内存占用、堆栈切换开销还是会比coroutine更大。