Re: Bug in scheduler when using rt_mutex

[Yong Zhang]

On Thu, Jan 20, 2011 at 12:27 PM, Yong Zhang <yong.zhang0@xxxxxxxxx> wrote:
> on Thu, Jan 20, 2011 at 12:18 PM, Yong Zhang <yong.zhang0@xxxxxxxxx> wrote:
>> On Wed, Jan 19, 2011 at 9:30 PM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
>>> +static void switched_from_fair(struct rq *rq, struct task_struct *p)
>>> +{
>>> + Â Â Â struct sched_entity *se = &p->se;
>>> + Â Â Â struct cfs_rq *cfs_rq = cfs_rq_of(se);
>>> +
>>> + Â Â Â /*
>>> + Â Â Â Â* Ensure the task's vruntime is normalized, so that when its
>>> + Â Â Â Â* switched back to the fair class the enqueue_entity(.flags=0) will
>>> + Â Â Â Â* do the right thing.
>>> + Â Â Â Â*
>>> + Â Â Â Â* If it was on_rq, then the dequeue_entity(.flags=0) will already
>>> + Â Â Â Â* have normalized the vruntime, if it was !on_rq, then only when
>>> + Â Â Â Â* the task is sleeping will it still have non-normalized vruntime.
>>> + Â Â Â Â*/
>>> + Â Â Â if (!se->on_rq && p->state != TASK_RUNNING)
>>> + Â Â Â Â Â Â Â se->vruntime -= cfs_rq->min_vruntime;
>>
>> Here it's possible that se->vruntime is little than cfs_rq->min_vruntime.
>
> Maybe we can:
> Â Â Â Â Â Â Â Â Â Â Â place_entity(cfs_rq, se, 0);
> Â Â Â Â Â Â Â Â Â Â Â se->vruntime -= cfs_rq->min_vruntime;

Hmmm, this doesn't make sense since we don't know how long the task
will sleep after we do this.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/