Re: [RFC] rtmutex: Do not boost fair tasks each other

[Kirill Tkhai <tkhai@yandex.ru>]

Hi, Thomas,

have you seen this version?

Thanks,
Kirill

30.05.2014, 00:52, "Kirill Tkhai" <tkhai@yandex.ru>:
> В Ср, 28/05/2014 в 22:26 +0200, Thomas Gleixner пишет:
>>  On Mon, 5 May 2014, Kirill Tkhai wrote:
>>>  В Сб, 03/05/2014 в 20:54 +0200, Thomas Gleixner пишет:
>>>>  Though exercising that code path as much as we can is not a bad thing
>>>>  either. So I'd like to see that made compile time conditional on one
>>>>  of the lock testing CONFIG items.
>>>  +#ifndef CONFIG_RT_MUTEX_BOOST_ALL
>>  No, not another pointless config option. Read what I said. What's
>>  wrong with using an existing config item, e.g DEBUG_RT_MUTEXES?
>>>  +#define heritable_prio(prio) (rt_prio(prio) || dl_prio(prio))
>>  inheritable please. It's not priority heritance and never will be.
>
> Thanks for comments. Here is new version.
>
> [PATCH] rtmutex: Do not boost owner's prio if waiter is SCHED_OTHER
>
> Higher priority does not provide exclusive privilege
> of one fair class task over the other. In this case
> priority boosting is pointless, and it may worsen
> performance.
>
> This patch makes boosting, which is requested by fair
> class waiters, optional. It's disabled by default, but
> it's possible to enable it for debugging purposes to
> have more cases of priority inheritance.
>
> Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
>
>  kernel/locking/rtmutex.c | 28 +++++++++++++++++++++-------
>  1 file changed, 21 insertions(+), 7 deletions(-)