Messages in this thread | | | Date | Wed, 24 Oct 2018 16:00:02 +0530 | From | Srikar Dronamraju <> | Subject | Re: [PATCH v2] sched/core: Don't mix isolcpus and housekeeping CPUs |
| |
* Peter Zijlstra <peterz@infradead.org> [2018-10-24 12:03:23]:
> It appears to me the for_each_online_node() iteration in > task_numa_migrate() needs an addition test to see if the selected node > has any CPUs in the relevant sched_domain _at_all_. >
Yes, this should work. Yi Wang does this extra check a little differently.
http://lkml.kernel.org/r/1540177516-38613-1-git-send-email-wang.yi59@zte.com.cn
However the last time I had posted you didn't like that approach. http://lkml.kernel.org/r/20170406073659.y6ubqriyshax4v4m@hirez.programming.kicks-ass.net
Further, I would think the number of times, we would be calling sched_setaffinity would be far less than task_numa_migrate(). In the regular case, where we never have isolcpus, we add this extra check.
Also what does it mean for a task to have its cpu affinity set to a mix of isolcpus and non isolcpus.
Also should we be updating update_numa_stats accordingly? While the problem may not be apparent there but we are counting the load and compute capacity of isolcpus.
-- Thanks and Regards Srikar Dronamraju
| |