Samuel Williams

Oh that makes total sense. Okay, I'm keen to test this when you are ready. cc @ciconia

I agree, the vast majority of user space code doesn't care about `EAGAIN`, it's like the half way house of concurrency.

I will try this out over the weekend.

I'm hoping we can transparently support this by using the poll fallback - i.e. assume it's going to work, but if we get EAGAIN, go to a fallback path.

> It does not now and for good. Docs should be fixed, what file:line was it? https://unixism.net/loti/ref-liburing/completion.html "If ts is specified, the application need not call io_uring_submit() before calling this...

> Userspace fills SQEs, so it should know when it needs to submit anything or not. On this point, we can certainly keep track of the number of SQEs filled...

> check for io_uring_wait_cqe_timeout(). I suspect this is what you need. I tried this but it failed to work correctly because it does not always call submit apparently. ``` io_uring_submit(&arguments->data->ring);...

We need a function which: - If there are completion events, return immediately with the count, otherwise - Submits if required, then - Waits for at least one event, and...

It would be nice if we could plug in async-http.

I see. Thanks for the explanation.