Date: Sun, 26 Mar 2006 23:35:52 +0200
From: Ingo Molnar <mingo@elte.hu>
To: Esben Nielsen <simlo@phys.au.dk>
Cc: tglx@linutronix.de, linux-kernel@vger.kernel.org
Subject: Re: Are ALL_TASKS_PI on in 2.6.16-rt7?
Message-ID: <20060326213552.GA1963@elte.hu>
References: <20060326212738.GA32562@elte.hu> <Pine.LNX.4.44L0.0603262231110.8060-100000@lifa03.phys.au.dk>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.44L0.0603262231110.8060-100000@lifa03.phys.au.dk>
User-Agent: Mutt/1.4.2.1i
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1043
Lines: 31


* Esben Nielsen <simlo@phys.au.dk> wrote:

> On Sun, 26 Mar 2006, Ingo Molnar wrote:
> 
> >
> > * Esben Nielsen <simlo@phys.au.dk> wrote:
> >
> > > It just looks like also normal, non-rt tasks are boosting.
> >
> > correct. We'd like to make sure the PI code is correct - and for
> > PI-futex it makes sense anyway.
> >
> 
> It wont work 100% when a task is boosted to a normal, non-rt prio well
> since at scheduler_tick()
>  		p->prio = effective_prio(p);
> can be executed overwriting the boost.

yeah - but that's relatively rare, upon expiration of the timeslice.  
The following would probably solve it: scheduler_tick() could take the 
pi_lock (before taking the rq lock), update normal_prio, and then call 
into rt_mutex_getprio() [just like setscheduler does] to set the 
priority.

	Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/