linux-kernel - 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

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:	Mon, 24 Nov 2008 16:36:30 +0100
From:	Peter Zijlstra <a.p.zijlstra@...llo.nl>
To:	Chris Friesen <cfriesen@...tel.com>
Cc:	Dhaval Giani <dhaval@...ux.vnet.ibm.com>,
	bharata@...ux.vnet.ibm.com, linux-kernel@...r.kernel.org,
	Srivatsa Vaddagiri <vatsa@...ux.vnet.ibm.com>,
	Ingo Molnar <mingo@...e.hu>
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 09:24 -0600, Chris Friesen wrote:
> Peter Zijlstra wrote:
> > On Mon, 2008-11-24 at 14:16 +0530, Dhaval Giani wrote:
> > 
> > 
> >>Which would mean the init_task_group becase it contains those tasks which are
> >>not grouped.
> > 
> > 
> > Only because of implementation details (we implement the !group case by
> > having them all part of a single group), conceptually they don't belong
> > to any group, hence talking about moving it to some group is just wrong.
> 
> > Furthermore your statement shows another misconception, a group of
> > ungrouped tasks doesn't make sense.
> 
> Arguably there is such a group, which is "the set of all RT tasks".

Sure, I understand that, and in fact that's how its implemented, no
group is still one group (which is how you can bootstrap math from
group/set theory).

But its not a manageable group in the cgroup sense, its just the
collection of RT tasks.

> Whether or not they should map to the top-level cgroup is a different 
> question.  Maybe in the !group case there should be a second top-level 
> "rt" cgroup?  We could even make the RT sched tuning knobs available there.

I'd rather just not display all that. We don't go make such 'unmanaged'
groups for other not configured controllers either.



--
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/