Messages in this thread Patch in this message | | | From | Gregory Haskins <> | Subject | [PATCH] sched: fix SCHED_FAIR wake-idle logic error | Date | Thu, 24 Apr 2008 14:08:38 -0400 |
| |
Hi Ingo,
I found this while looking at -rt, but it would in theory be a problem in mainline too given a sufficient number of real-time tasks present. PREEMPT_RT just helps make the likelyhood of RT tasks being running go up ;)
This patch applies to sched-devel and it solves a case where CFS tends to piggy back wakeups on a single core. The issue is that the presence of RT tasks can of course inflate rq->nr_running such that wake_idle() will skip trying to move to an idle core, thinking that we are already balanced. In reality, all the other cores could be idle but we are running RT tasks on the affined core.
Regards, -Greg
-------------------------------------- sched: fix SCHED_FAIR wake-idle logic error
We currently use an optimization to skip the overhead of wake-idle processing if more than one task is assigned to a run-queue. The assumption is that the system must already be load-balanced or we wouldnt be overloaded to begin with.
The problem is that we are looking at rq->nr_running, which may include RT tasks in addition to CFS tasks. Since the presence of RT tasks really has no bearing on the balance status of CFS tasks, this throws the calculation off.
This patch changes the logic to only consider the number of CFS tasks when making the decision to optimze the wake-idle.
Signed-off-by: Gregory Haskins <ghaskins@novell.com> CC: Ingo Molnar <mingo@elte.hu> CC: Peter Zijlstra <peterz@infradead.org> ---
kernel/sched_fair.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 89fa32b..80b7891 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1007,7 +1007,7 @@ static int wake_idle(int cpu, struct task_struct *p) * sibling runqueue info. This will avoid the checks and cache miss * penalities associated with that. */ - if (idle_cpu(cpu) || cpu_rq(cpu)->nr_running > 1) + if (idle_cpu(cpu) || cpu_rq(cpu)->cfs.nr_running > 1) return cpu; for_each_domain(cpu, sd) {
| |