[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <1227552397.4259.504.camel@twins>
Date: Mon, 24 Nov 2008 19:46:37 +0100
From: Peter Zijlstra <a.p.zijlstra@...llo.nl>
To: vatsa@...ibm.com
Cc: bharata@...ux.vnet.ibm.com, linux-kernel@...r.kernel.org,
Ingo Molnar <mingo@...e.hu>,
Dhaval Giani <dhaval@...ux.vnet.ibm.com>
Subject: Re: [PATCH] sched: Don't allow priority switch to realtime when
the task doesn't belong to init_task_group and when CONFIG_RT_GROUP_SCHED
isn't set
On Mon, 2008-11-24 at 22:35 +0530, Srivatsa Vaddagiri wrote:
> On Mon, Nov 24, 2008 at 09:32:06AM +0100, Peter Zijlstra wrote:
> > > And may be as Ingo
> > > suggested they should be moved to init_task_group.
> >
> > Bzzt, wrong! They should not be moved to any group, they cannot be moved
> > to any group, as they are group invariant.
>
> Though what you say makes sense, the confusion arises from existing
> cgroup<->scheduler interface, which can end up showing the above
> single-set of RT-tasks to be split into multiple sets.
>
> RT Tasks -> {RT0, RT1}
>
> can be shown as:
>
> /a/tasks {RT0, ...}
> /b/tasks {RT1, ...}
>
> Does this cause any problems? Perhaps no, just seems odd ..
>
> Fixing this oddity of representing single RT-tasks set as multiple is not a
> cgroup issue IMHO.
I'm fine with just not showing RT tasks in cgroup:cpu/tasks.
> P.S :- If eventually RT_GROUP_SCHED will be merged with FAIR_GROUP_SCHED, then
> it may make sense for us to just ignore this oddity for timebeing and look
> forward to the options being merged.
Right, that might take a bit of time still as RT_GROUP needs a deadline
scheduler to be complete.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists