Re: io_uring backend for NIO

On Mon, Dec 7, 2020 at 3:12 PM Martin Grigorov wrote: > Hi devs, > > Maybe you have also noticed some noise in the web about io_uring > - the better epoll. > > Pros: > - less syscalls > - no copying of data between userland and kernel > > Cons: > - it seems rather

io_uring backend for NIO

Hi devs, Maybe you have also noticed some noise in the web about io_uring - the better epoll. Pros: - less syscalls - no copying of data between userland and kernel Cons: - it seems rather unstable to me. Many articles say that io_uring should be more stable sinc