Re: [RFC v3 00/12] DRM scheduling cgroup controller

["Michal Koutný" <mkoutny@suse.com>]

[-- Attachment #1: Type: text/plain, Size: 1203 bytes --]

On Fri, Jan 27, 2023 at 11:40:58AM +0000, Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> wrote:
> The main point is, should someone prove me wrong and come up a smarter way
> at some point in the future, then "drm.weight" as an ABI remains compatible
> and the improvement can happen completely under the hood. In the mean time
> users get external control, and _some_ ability to improve the user
> experience with the scenarios such as I described yesterday.

I'm on board now.

(I've done a mental switch of likening this rather to existing IO
control (higher variance of quanta size, worse preemption, limited
effect) than CPU control.)


> Cgroup tree hierarchy modifications being the reason for not converging can
> also happen, but I thought I can hand wave that as not a realistic scenario.
> Perhaps I am not imaginative enough?

My suggestion: simply skip offlined drmcgs instead of restarting whole
iteration. (A respawning cgroup-wrapped job or intentionally adverse
respawner could in my imagination cause that.)

> Under or over-accounting for migrating tasks I don't think can happen since
> I am explicitly handling that.

I'll reply to the patch for better context...

Regards,
Michal

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 228 bytes --]