Re: [PATCH v11]xen: sched: convert RTDS from time to event driven model

From: Dario Faggioli <dario.faggioli@citrix.com>
To: Jan Beulich <JBeulich@suse.com>, Meng Xu <mengxu@cis.upenn.edu>
Cc: "xen-devel@lists.xenproject.org" <xen-devel@lists.xenproject.org>,
	Tianyang Chen <tiche@seas.upenn.edu>,
	George Dunlap <george.dunlap@citrix.com>,
	Dagaen Golomb <dgolomb@seas.upenn.edu>
Subject: Re: [PATCH v11]xen: sched: convert RTDS from time to event driven model
Date: Fri, 18 Mar 2016 10:23:21 +0100	[thread overview]
Message-ID: <1458293001.15374.87.camel@citrix.com> (raw)
In-Reply-To: <56EBC02002000078000DE324@prv-mh.provo.novell.com>

[-- Attachment #1.1: Type: text/plain, Size: 1932 bytes --]

On Fri, 2016-03-18 at 01:45 -0600, Jan Beulich wrote:
> > 
> > > > On 18.03.16 at 05:09, <mengxu@cis.upenn.edu> wrote:
> > Great job! However, we still have 1 mile in the 100-mile journey.
> > :-D
> > 
> > I applied the patch on staging and tried some test cases. One of
> > them
> > is as follows:
> > 
> > I tried to create a cpupool and then migrate a VM to the new
> > cpupool;
>
BTW, Meng:

(XEN)    [<ffff82d08012e182>] schedule_cpu_switch+0x250/0x28a
(XEN)    [<ffff82d080101b49>] cpupool.c#cpupool_assign_cpu_locked+0x31/0x11f

I think you mean "and then move a CPU from a cpupool to another". Or
perhaps what you said is what your script does, and you weren't sure at
what stage it explodes.

Well, let me tell you: it's when you move a CPU between pools that have
schedulers that remaps the scheduler locks (such as Credit2-->RTDS and
vice versa).

> > However, the system triggers the bug as below. I guess this is some
> > kind of bug that are known to us,  and Dario had some uncommitted
> > patch to fix it, IIRC?
> In the context of this patch the most relevant question is: Is this
> an issue with the patch, or one that existed already before? 
>
Exactly!

And the answer is:
 - it's pre-existing
 - it's an even bigger issue than that ASSERT triggering (i.e., there
   are potential races even when things works)
 - I'm taking care of it.

> After
> all that's what we're in need to know whether the change can go
> in. 
>
Yep.

> And skimming over the patch, it doesn't seem to alter code
> related to where you see things blow up.
> 
Indeed it does not.

Thanks and Regards,
Dario
-- 
<<This happens because I choose it to happen!>> (Raistlin Majere)
-----------------------------------------------------------------
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)

[-- Attachment #1.2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

[-- Attachment #2: Type: text/plain, Size: 126 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel