Re: Real-Time Preemption, -RT-2.6.12-final-V0.7.51-12

[Peter Zijlstra <a.p.zijlstra@chello.nl>]

On Sun, 2005-07-10 at 17:10 +0200, Ingo Molnar wrote:
> * Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:
> 
> > > I can reproduce priority leakage on my SMP system; any userspace process
> > > chrt'ed up and a lot will follow. This makes the system very
> > > unresponsive when doing a make -j5. Verified on 51-{6,18,23}.
> > > 
> > 
> > The following patch seems to fix it for me, YMMV.
> > 
> > --- kernel/sched.c~     2005-07-08 10:27:59.000000000 +0200
> > +++ kernel/sched.c      2005-07-10 13:00:42.000000000 +0200
> > @@ -780,7 +780,8 @@ static void recalc_task_prio(task_t *p,
> >                 }
> >         }
> > 
> > -       p->prio = p->normal_prio = effective_prio(p);
> > +       p->prio = effective_prio(p);
> > +       p->normal_prio = unlikely(rt_prio(p->normal_prio)) ? p->prio : __effective_prio(p);
> >  }
> 
> ahh, indeed, this code did not take boosting into account. Good catch!  
> I'm wondering why this only showed up on SMP. 

I was thinking because of the agressive RT rebalance code this codepath
is more exercised on SMP systems.

> I've fixed it a bit 
> differently in my tree, by making the roles of the various priority 
> fields and functions more obvious, see the delta patch below.  

Yes, much nicer :-)

> I've also 
> released the -51-23 patch with these changes included. Does this fix 
> priority leakage on your SMP system?
> 

-51-24 right? I'll give it a spin.

-- 
Peter Zijlstra <a.p.zijlstra@chello.nl>