Re: [RFC 1/2] softirq: Defer net rx/tx processing to ksoftirqd context

[Peter Zijlstra]

On Fri, Jan 12, 2018 at 06:23:08AM +0100, Mike Galbraith wrote:
> On Thu, 2018-01-11 at 12:22 -0800, Linus Torvalds wrote:
> > On Thu, Jan 11, 2018 at 12:16 PM, Eric Dumazet <edumazet@xxxxxxxxxx> wrote:
> > >
> > > Note that when I implemented TCP Small queues, I did experiments between
> > > using a work queue or a tasklet, and workqueues added unacceptable P99
> > > latencies, when many user threads are competing with kernel threads.
> > 
> > Yes.
> > 
> > So I think one solution might be to have a hybrid system, where we do
> > the softirq's synchronously normally (which is what you really want
> > for good latency).
> > 
> > But then fall down on a threaded model - but that fallback case should
> > be per-softirq, not global. So if one softirq uses a lot of CPU time,
> > that shouldn't affect the latency of other softirqs.
> > 
> > So maybe we could get rid of the per-cpu ksoftirqd entirely, and
> > replace it with with per-cpu and per-softirq workqueues?
> 
> How would that be better than what RT used to do, and I still do for my
> RT kernels via boot option, namely split ksoftirqd into per-softirq
> threads.

Since we mention it; one of the problems RT has is that all RX is
through a single softirq context, which generates a priority inversion
between devices.

If we're splitting things, it would be awesome if we could get
per-device context.