linux-kernel - Re: [BUGFIX][PATCH] Fix sched rt group scheduling when hierachy is enabled

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20110304121146.GG2868@balbir.in.ibm.com>
Date:	Fri, 4 Mar 2011 17:41:46 +0530
From:	Balbir Singh <balbir@...ux.vnet.ibm.com>
To:	Yong Zhang <yong.zhang0@...il.com>
Cc:	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Ingo Molnar <mingo@...e.hu>,
	Peter Zijlstra <pzijlstr@...hat.com>,
	Srivatsa Vaddagiri <vatsa@...ux.vnet.ibm.com>,
	Bharata B Rao <bharata.rao@...ibm.com>
Subject: Re: [BUGFIX][PATCH] Fix sched rt group scheduling when hierachy is
 enabled

* Yong Zhang <yong.zhang0@...il.com> [2011-03-04 17:30:58]:

> On Fri, Mar 4, 2011 at 4:59 PM, Balbir Singh <balbir@...ux.vnet.ibm.com> wrote:
> >>>>  deactivate_task(b); /* A.throttled == 1 && A.rt_nr_running == 0 */
> >>>>    do_sched_rt_period_timer(); /* A.run_time == 0 && A.throttled == 0*/
> >>>
> >>> Note at some point rt_time becomes 0 and if enqueue is not set, the
> >>
> >> If the group should be add back, it will be at the first
> >> do_sched_rt_period_timer() which decreases run_time;
> >>
> >
> > As long as idle is 0, the period will continue to run, if it has
> > rt_nr_running or rt_time, the timer will run.
> 
> Yep.
> Should be the first do_sched_rt_period_timer() which meets
> rt_rq->rt_time < runtime :)
> 
> >
> >>> next do_sched_rt_period_timer() is a NOP and does not enqueue back the
> >>> group
> >>
> >> Otherwise it will be added back when a task is attching to it.
> >>
> >> I still can't see how a unthrottled group which has task attched stay
> >> unqueued.
> >
> > The other way of looking at the first change is
> >
> > Can we have rt_time as 0, rt_nr_running >=1, rt_throttled !=0 and
> > still not have the rt_rq enqueued?
> 
> Yeah, the same question. How could we reach that?
>

I based the changes on what I saw during my debugging/test. I
explained it earlier,

Everyone is dequeued

1. child runs first, finds parent throttled, so it does not queue
anything on parent group. child is unthrottled and rt_time now becomes
0, parent's rt_nr_running is not incremented.
2. Parent timer runs, it is unthrottled, its group->rt_nr_running is 0
hence enqueue is not called

 
> IMHO, rt_time == 0 and rt_throttled !=0 can't coexist.
> 
> >
> > If this is not the case, we don't lose much, a quick check for
> > rt_nr_running and on_rt_q
> 
> I don't get what you mean here.
>

I was talking of the overhead of the check I added. 

-- 
	Three Cheers,
	Balbir
--
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/