Introducing io_uring_spawn

Posted Sep 21, 2022 15:30 UTC (Wed) by kees (subscriber, #27264)
In reply to: Introducing io_uring_spawn by josh
Parent article: Introducing io_uring_spawn

On the performance side, Josh and I also carved off like 20% start time for path searches by moving the "open" during execve(2) earlier:
https://lore.kernel.org/lkml/202209161637.9EDAF6B18@keesc...

I'm hoping we can do some more perf examination to check "normal" workloads too.