| Date | Mon, 19 Aug 2013 18:00:58 +0200 | From | Peter Zijlstra <> | Subject | [PATCH 00/10] Various load-balance cleanups/optimizations -v2 |
| |
After poking at them a little more I feel somewhat more confident.
I found one more bug, but this one was my own fault, we should also clear sds->busiest_stat.avg_load because update_sd_pick_busiest() reads that before we set it.
Moved the memset optimization and the fix for that into a separate patch.
Other than that, there's a sched_domain degenerate fix; and a new way to detect group_imb (which relies on the sched_domain fix).
|