Re: [RFC] sched/deadline: Prevent rt_time growth to infinity

[Kirill Tkhai <tkhai@yandex.ru>]

21.02.2014, 14:37, "Peter Zijlstra" <peterz@infradead.org>:
> On Thu, Feb 20, 2014 at 02:16:00AM +0400, Kirill Tkhai wrote:
>
>>  Since deadline tasks share rt bandwidth, we must care about
>>  bandwidth timer set. Otherwise rt_time may grow up to infinity
>>  in update_curr_dl(), if there are no other available RT tasks
>>  on top level bandwidth.
>>
>>  I'm going to decide the problem the way below. Almost untested
>>  because of I skipped almost all of recent patches which haveto be applied from lkml.
>>
>>  Please say, if I skipped anything in idea. Maybe better put
>>  start_top_rt_bandwidth() into set_curr_task_dl()?
>
> How about we only increment rt_time when there's an RT bandwidth timer
> active?

This case RT and DL may eat all the time:

--------------  time ------------------>
|RT's working |DL's working|
----------------------------
|rt_runtime   |            |
----------------------------
|        rt_period         |

Or at least more, than it's allowed.


It looks like, if we want to limit time of high priority classes
execution, we have to set the timer anyway.

> ---
> --- a/kernel/sched/rt.c
> +++ b/kernel/sched/rt.c
> @@ -568,6 +568,12 @@ static inline struct rt_bandwidth *sched

[snipped by dumb mail client]