Re: [PATCH RT v2 2/3] sched: Lazy migrate_disable processing

2019-10-18 Thread Scott Wood
On Fri, 2019-10-18 at 14:12 -0400, Waiman Long wrote: > On 10/12/19 2:52 AM, Scott Wood wrote: > > Avoid overhead on the majority of migrate disable/enable sequences by > > only manipulating scheduler data (and grabbing the relevant locks) when > > the task actually schedules while migrate-disabled

Re: [PATCH RT v2 2/3] sched: Lazy migrate_disable processing

2019-10-18 Thread Waiman Long
On 10/12/19 2:52 AM, Scott Wood wrote: > Avoid overhead on the majority of migrate disable/enable sequences by > only manipulating scheduler data (and grabbing the relevant locks) when > the task actually schedules while migrate-disabled. A kernel build > showed around a 10% reduction in system ti

[PATCH RT v2 2/3] sched: Lazy migrate_disable processing

2019-10-11 Thread Scott Wood
Avoid overhead on the majority of migrate disable/enable sequences by only manipulating scheduler data (and grabbing the relevant locks) when the task actually schedules while migrate-disabled. A kernel build showed around a 10% reduction in system time (with CONFIG_NR_CPUS=512). Instead of cpuhp