Re: NFS: 82ms wakeup latency 4.14-rc4

From: Mike Galbraith
Date: Mon Dec 18 2017 - 13:35:25 EST

On Mon, 2017-12-18 at 12:27 -0500, J. Bruce Fields wrote:
> I'd forgotten about throughput/latency tradeoffs--but
> couldn't those in theory be managed by runtime configuration of the
> sceduler, or at least some smaller hammer than turning off preemption
> entirely?

A kernel that has all of the goop required to support preemption can't
possibly perform as well as a kernel that can simply assume everything
is safe.  All that infrastructure costs cycles.