Linux scheduler : what is struct sched_entity and why not store task_struct directly

Question

In the Linux scheduler, each CPU has one struct rq, which contains sub-runqueues for each sched_class (CFS, RT, DL, idle). For CFS, the runqueue stores a task’s sched_entity (p->se) instead of the whole task_struct.

What is the purpose of sched_entity in CFS, and why does CFS operate on these entities rather than task_struct directly?

It keeps the source cleaner and this was done nearly 18 years ago (in the 2.6.23 kernel). — Ian Abbott
– Ian Abbott, Commented Aug 14 at 16:44

Marco Bonelli · Accepted Answer · 2025-09-12 12:25:56Z

1

task_struct is huge and the scheduler doesn't really care at all about 90% of the fields it holds. The details in sched_entity are what the scheduler is concerned the most about. Therefore, it makes sense to only keep track of task->se. At the very least, you are saving additional math to retrieve task->se every single time you need to access it, but most importantly you are also making it clear that scheduler code dealing with sched_entity should not deal with other fields of task_struct.

answered Sep 12 at 12:25

Marco Bonelli

71k21 gold badges129 silver badges153 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Linux scheduler : what is struct sched_entity and why not store task_struct directly

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related