Messages in this thread | | | Date | Mon, 16 Jan 2012 18:34:45 +0800 | From | Michael Wang <> | Subject | Re: [PATCH] sched: Accelerate "pick_next_entity" under special condition |
| |
Hi, peter
Thanks so much for your reply :)
On 01/16/2012 05:51 PM, Peter Zijlstra wrote:
> On Mon, 2012-01-16 at 17:37 +0800, Michael Wang wrote: >> From: Michael Wang <wangyun@linux.vnet.ibm.com> >> >> We can avoid some useless operation in some special condition. > > This is a pretty empty statement. > >> For example: >> If we have "cfs_rq->next" and it can be use, we just return it directly. > > What it doesn't state is what it actually does, if it affects the common > case and performance numbers (or a good reason for the lack thereof). >
Please help me to understand the logic, I think in the original code, even if we have cfs_rq->next and wakeup_preempt_entity check passed, we still need to do a lot of work (check cfs_rq->last for example) which have no influence on result, will it be better if we skip them and just do what really needed?
>> Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com> >> --- >> kernel/sched/fair.c | 28 +++++++++++++++++----------- >> 1 files changed, 17 insertions(+), 11 deletions(-) >> >> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c >> index 84adb2d..9fc2c3c 100644 >> --- a/kernel/sched/fair.c >> +++ b/kernel/sched/fair.c >> @@ -1295,6 +1295,8 @@ set_next_entity(struct cfs_rq *cfs_rq, struct sched_entity *se) >> static int >> wakeup_preempt_entity(struct sched_entity *curr, struct sched_entity *se); >> >> +#define ENTITY_PREEMPT_ALLOWED(prev,next) (wakeup_preempt_entity(prev, next) < 1) >
> This is just uglification imo, its shouting and it doesn't actually win > you much space. >
I see, sorry for the bad idea.
Best regards, Michael Wang
| |