Re: [PATCH v4 6/8] sched/fair: Add sched group latency support

[Vincent Guittot]

On Wed, 21 Sept 2022 at 18:48, Tejun Heo <tj@xxxxxxxxxx> wrote:
>
> Hello,
>
> On Wed, Sep 21, 2022 at 05:07:38PM +0100, Qais Yousef wrote:
> > Wouldn't cpu.latency.nice be enough? I think the latency_offset is
> > implementation detail that userspace shouldn't be concerned about.
>
> One option could be just using the same mapping as cpu.weight so that 100
> maps to neutral, 10000 maps close to -20, 1 maps close to 19. It isn't great
> that the value can't be interpreted in any intuitive way (e.g. a time
> duration based interface would be a lot easier to grok even if it still is
> best effort) but if that's what the per-task interface is gonna be, it'd be
> best to keep cgroup interface in line.

I would prefer a signed range like the [-1000:1000] as the behavior is
different for sensitive and non sensitive task unlike the cpu.weight
which is reflect that a bigger value get more

>
> As for whether a single value would fit the bill, it's again something which
> should be answered for both task and cgroup based interface at the same
> time. That said, my not-too-throught-through opinion is that a single value
> for per-task / per-cgroup interface + system level knobs to fine tune how
> that actually applies is likely enough and probably better than exposing
> exposing a host of internal details to applications directly.
>
> Thanks.
>
> --
> tejun