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]

Date:	Fri, 4 Mar 2011 16:52:26 +0800
From:	Yong Zhang <yong.zhang0@...il.com>
To:	Balbir Singh <balbir@...ux.vnet.ibm.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

On Fri, Mar 4, 2011 at 4:35 PM, Balbir Singh <balbir@...ux.vnet.ibm.com> wrote:
> On Fri, Mar 4, 2011 at 2:02 PM, Yong Zhang <yong.zhang0@...il.com> wrote:
>> On Fri, Mar 4, 2011 at 3:25 PM, Balbir Singh <balbir@...ux.vnet.ibm.com> wrote:
>>> * Yong Zhang <yong.zhang0@...il.com> [2011-03-04 11:43:16]:
>>>
>>>> On Thu, Mar 3, 2011 at 11:29 PM, Balbir Singh <balbir@...ux.vnet.ibm.com> wrote:
>>>> > No, not really :) It is required, it is a backup check to see if we
>>>> > have queued tasks, rt_time of 0 and the runqueue is not throttled, why
>>>> > should it be dequeued?
>>>>
>>>> But I can't see where that kind of rt_rq is dequeued, mind pointing it out?
>>>>
>>>
>>> So here is what I saw
>>>
>>> 1. sched_dequeue_stack called from the dequeue path dequeues the
>>> queues and sets rt_nr_running to 0
>>> 2. Enqueuing fails because rt_throttled is set for the group_rq
>>> (parent who is throttled)
>>> 3. This causes further enqueue to fail, since rt_nr_running did
>>> not increment in step 2
>>
>> Whose rt_nr_running? group_rq's or parent of group_rq?
>> For parent of group_rq, yes.
>> For group_rq's, no, because we don't touch its rt_nr_running.
>>
>
> For parent of group_rq
>
>>
>>> , eventually the timer decrements rt_time
>>> to 0 and the task is never picked up.
>>
>> If I understand correctly, you mean this:
>>
>> For a group(A) which has one task(b) attached but A is throttled,so
>> A is unqueued now
>> A.throttled == 1 && A.rt_nr_running == 1
>>  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;

> 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.

Thanks,
Yong


-- 
Only stand for myself
--
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/