Re: [PATCH RESEND] sched/fair: Fix wrong cpu selecting from isolated domain

2020-09-24 Thread Xunlei Pang
On 9/24/20 3:18 PM, Vincent Guittot wrote: > On Thu, 24 Sep 2020 at 08:48, Xunlei Pang wrote: >> >> We've met problems that occasionally tasks with full cpumask >> (e.g. by putting it into a cpuset or setting to full affinity) >> were migrated to our isolated cpus in production environment. >> >>

Re: [PATCH RESEND] sched/fair: Fix wrong cpu selecting from isolated domain

2020-09-24 Thread Vincent Guittot
On Thu, 24 Sep 2020 at 08:48, Xunlei Pang wrote: > > We've met problems that occasionally tasks with full cpumask > (e.g. by putting it into a cpuset or setting to full affinity) > were migrated to our isolated cpus in production environment. > > After some analysis, we found that it is due to

[PATCH RESEND] sched/fair: Fix wrong cpu selecting from isolated domain

2020-09-24 Thread Xunlei Pang
We've met problems that occasionally tasks with full cpumask (e.g. by putting it into a cpuset or setting to full affinity) were migrated to our isolated cpus in production environment. After some analysis, we found that it is due to the current select_idle_smt() not considering the sched_domain