sched: Remove sched_class::balance()
Get series with b4
b4 am [email protected]Versions newest first
- v1 Latest2 patches 17 repliesIn review
Patches in v1 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/2 | sched/core: Allow newidle for core-schedkernel/ kernel/ | +16 | 3 replies | |
| 2/2 | sched: Remove sched_class::balance()kernel/ kernel/ kernel/ +4 more | +22 | 0 replies |
Reviews trailers per patch
No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.
Cover letter [PATCH 0/2] sched: Remove sched_class::balance()
Hi!
As mentioned [1], I was going to move sched_balance_newidle() into balance_fair(). And while the patch was quickly done, something didn't feel right.
The two things that bugged me were:
- core-sched only calls prev_balance() on one sibling, not allowing the others to pull tasks;
- both ::balance() and ::pick_task() take @rf and do balance operations.
That latter is because of ext; the DSQ move needs to be undone if it turns out a higher class will get ran after all -- this is where ::pick_task() got the rf from and why I moved its balance into pick_task_scx().
After a bit of puzzling and a few wrong turns [2], I think these patches do away with sched_class::balance() instead, less is more and all that.
[1] https://patch.msgid.link/20260611113219.GG187714%40noisy.programming.kicks-ass.net
[2] I'm pretty sure I messed things up a few times, but rt-migration-test never threw a failure on me -- which suggests it isn't very good at finding fails.