diff options
| author |   Dhaval Giani <dhaval@linux.vnet.ibm.com> | 2008-04-19 19:44:59 +0200 |
|---|---|---|
| committer |   Ingo Molnar <mingo@elte.hu> | 2008-04-19 19:44:59 +0200 |
| commit | 354d60c2ff72d86627dfe2089d186824abf4bb8e (patch) | |
| tree | 10cea61ce7036448ed7246820c5575df2a61bb3b /kernel/sched_rt.c | |
| parent | sched: fix checks (diff) | |
| download | linux-dev-354d60c2ff72d86627dfe2089d186824abf4bb8e.tar.xz linux-dev-354d60c2ff72d86627dfe2089d186824abf4bb8e.zip | |
sched: mix tasks and groups
This patch allows tasks and groups to exist in the same cfs_rq. With this
change the CFS group scheduling follows a 1/(M+N) model from a 1/(1+N)
fairness model where M tasks and N groups exist at the cfs_rq level.
[a.p.zijlstra@chello.nl: rt bits and assorted fixes]
Signed-off-by: Dhaval Giani <dhaval@linux.vnet.ibm.com>
Signed-off-by: Srivatsa Vaddagiri <vatsa@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Diffstat (limited to 'kernel/sched_rt.c')
| -rw-r--r-- | kernel/sched_rt.c | 15 |
1 files changed, 9 insertions, 6 deletions
diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 8ff824565e06..201a69382a42 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -374,11 +374,15 @@ static void update_curr_rt(struct rq *rq) curr->se.exec_start = rq->clock; cpuacct_charge(curr, delta_exec); - spin_lock(&rt_rq->rt_runtime_lock); - rt_rq->rt_time += delta_exec; - if (sched_rt_runtime_exceeded(rt_rq)) - resched_task(curr); - spin_unlock(&rt_rq->rt_runtime_lock); + for_each_sched_rt_entity(rt_se) { + rt_rq = rt_rq_of_se(rt_se); + + spin_lock(&rt_rq->rt_runtime_lock); + rt_rq->rt_time += delta_exec; + if (sched_rt_runtime_exceeded(rt_rq)) + resched_task(curr); + spin_unlock(&rt_rq->rt_runtime_lock); + } } static inline @@ -477,7 +481,6 @@ static void dequeue_rt_entity(struct sched_rt_entity *rt_se) * entries, we must remove entries top - down. * * XXX: O(1/2 h^2) because we can only walk up, not down the chain. - * doesn't matter much for now, as h=2 for GROUP_SCHED. */ static void dequeue_rt_stack(struct task_struct *p) { |