Re: [PATCH v3 1/3] sched/pelt: Move pelt related code in a dedicated file

2017-12-12 Thread Quentin Perret
On Tuesday 12 Dec 2017 at 12:21:39 (+0100), Vincent Guittot wrote: > Hi Quentin, > > On 11 December 2017 at 15:08, Quentin Perret wrote: > > Hi Vincent, > > > > Although I agree that moving the PELT code in a dedicated file is > > probably the cleanest way to achieve what

Re: [PATCH v3 1/3] sched/pelt: Move pelt related code in a dedicated file

2017-12-12 Thread Quentin Perret
On Tuesday 12 Dec 2017 at 12:21:39 (+0100), Vincent Guittot wrote: > Hi Quentin, > > On 11 December 2017 at 15:08, Quentin Perret wrote: > > Hi Vincent, > > > > Although I agree that moving the PELT code in a dedicated file is > > probably the cleanest way to achieve what you want, I was

Re: [PATCH v3 1/3] sched/pelt: Move pelt related code in a dedicated file

2017-12-12 Thread Vincent Guittot
Hi Quentin, On 11 December 2017 at 15:08, Quentin Perret wrote: > Hi Vincent, > > Although I agree that moving the PELT code in a dedicated file is > probably the cleanest way to achieve what you want, I was wondering if > you were able no measure any overhead due to

Re: [PATCH v3 1/3] sched/pelt: Move pelt related code in a dedicated file

2017-12-12 Thread Vincent Guittot
Hi Quentin, On 11 December 2017 at 15:08, Quentin Perret wrote: > Hi Vincent, > > Although I agree that moving the PELT code in a dedicated file is > probably the cleanest way to achieve what you want, I was wondering if > you were able no measure any overhead due to moving the

Re: [PATCH v3 1/3] sched/pelt: Move pelt related code in a dedicated file

2017-12-11 Thread Quentin Perret
Hi Vincent, Although I agree that moving the PELT code in a dedicated file is probably the cleanest way to achieve what you want, I was wondering if you were able no measure any overhead due to moving the __update_load_avg_*() functions in a different translation unit ? This is introducing

Re: [PATCH v3 1/3] sched/pelt: Move pelt related code in a dedicated file

2017-12-11 Thread Quentin Perret
Hi Vincent, Although I agree that moving the PELT code in a dedicated file is probably the cleanest way to achieve what you want, I was wondering if you were able no measure any overhead due to moving the __update_load_avg_*() functions in a different translation unit ? This is introducing

[PATCH v3 1/3] sched/pelt: Move pelt related code in a dedicated file

2017-11-22 Thread Vincent Guittot
We want to track rt_rq's utilization as a part of the estimation of the whole rq's utilization. This is necessary because rt tasks can steal utilization to cfs tasks and make them lighter than they are. As we want to use the same load tracking mecanism for both and prevent useless dependency

[PATCH v3 1/3] sched/pelt: Move pelt related code in a dedicated file

2017-11-22 Thread Vincent Guittot
We want to track rt_rq's utilization as a part of the estimation of the whole rq's utilization. This is necessary because rt tasks can steal utilization to cfs tasks and make them lighter than they are. As we want to use the same load tracking mecanism for both and prevent useless dependency