Re: Proposed memcg meeting at October Kernel Summit/European LinuxCon in Prague

[Glauber Costa <glommer@parallels.com>]

On 10/14/2011 02:13 AM, Ying Han wrote:
> On Thu, Oct 13, 2011 at 1:58 PM, James Bottomley
> <James.Bottomley@hansenpartnership.com>  wrote:
>> On Thu, 2011-10-13 at 13:54 -0700, Ying Han wrote:
>>> On Wed, Oct 12, 2011 at 7:14 AM, James Bottomley
>>> <James.Bottomley@hansenpartnership.com>  wrote:
>>>> On Mon, 2011-10-10 at 19:35 -0700, Ying Han wrote:
>>>>> On Thu, Sep 29, 2011 at 2:30 PM, Hugh Dickins<hughd@google.com>  wrote:
>>>>>> On Thu, Sep 29, 2011 at 4:54 AM, Michal Hocko<mhocko@suse.cz>  wrote:
>>>>>>> On Tue 27-09-11 13:16:19, James Bottomley wrote:
>>>>>>>> On Mon, 2011-09-26 at 15:10 +0200, Michal Hocko wrote:
>>>>>>>>> On Thu 22-09-11 12:16:47, James Bottomley wrote:
>>>>>>>>>> Hi All,
>>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> One of the major work items that came out of the Plumbers conference
>>>>>>>>>> containers and Cgroups meeting was the need to work on memcg:
>>>>>>>>>>
>>>>>>>>>> http://www.linuxplumbersconf.org/2011/ocw/events/LPC2011MC/tracks/105
>>>>>>>>>>
>>>>>>>>>> (see etherpad and presentations)
>>>>>>>>>>
>>>>>>>>>> Since almost everyone will be either at KS or LinuxCon, I thought doing
>>>>>>>>>> a small meeting on the Wednesday of Linux Con (so those at KS who might
>>>>>>>>>> not be staying for the whole of LinuxCon could attend) might be a good
>>>>>>>>>> idea.  The object would be to get all the major players to agree on
>>>>>>>>>> who's doing what.  You can see Parallels' direction from the patches
>>>>>>>>>> Glauber has been posting.  Google should shortly be starting work on
>>>>>>>>>> other aspects of the memgc as well.
>>>>>>>>>>
>>>>>>>>>> As a precursor to the meeting (and actually a requirement to make it
>>>>>>>>>> effective) we need to start posting our preliminary patches and design
>>>>>>>>>> ideas to the mm list (hint, Google people, this means you).
>>>>>>>>>>
>>>>>>>>>> I think I've got all of the interested parties in the To: field, but I'm
>>>>>>>>>> sending this to the mm list just in case I missed anyone.  If everyone's
>>>>>>>>>> OK with the idea (and enough people are going to be there) I'll get the
>>>>>>>>>> Linux Foundation to find us a room.
>>>>>>>>>
>>>>>>>>> I am not going to be at KS but I am in Prague. I would be happy to meet
>>>>>>>>> as well if it is possible.
>>>>>>>>
>>>>>>>> Certainly.
>>>>>>>
>>>>>>> OK, then add me as well.
>>>>>>
>>>>>> Please include Ying Han and Hugh Dickins; but regrettably, scheduling
>>>>>> issues will prevent Greg Thelen from attending.
>>>>>
>>>>> Thank you Hugh. I will be in KS as well as the memcg meeting. Sorry
>>>>> for the late reply due to OOO in the past few weeks.
>>>>>
>>>>> James,
>>>>>
>>>>> Thank you so much for organizing this and please keep us informed when
>>>>> the detailed schedule is out :)
>>>>
>>>> We're a bit blocked on this.  We have some proposals, particularly in
>>>> the area of shrinkers, but we know you have patches in this area which
>>>> we haven't seen ... can you post the google memgc patches just for
>>>> reference (they don't have to be final, or even highly polished) just so
>>>> we have a common base to work from?
>>>
>>> sorry for getting back late.
>>>
>>> We are preparing the patches and should be able to send out before the
>>> summit. The patchset itself does the kernel slab accounting in memcg,
>>> and something we would like to discuss in the memcg meeting in
>>> Wednesday as well.
>>
>> Perfect, thanks.  This meshes well because we're trying to recast the
>> Dentry cache patch in terms of slab accounting and we need to make sure
>> our two patches are consistent.
>
> I assume the dentry cache patch is the "[PATCH v3 0/4] Per-container
> dcache limitation "?
>
> Also, I am not sure where to get more information of the proposals
> being sent so far for the meeting, and it would be helpful for us to
> look through them before the day. Sorry if i missed it somewhere in
> linux-mm :)
There were many discussions with no conclusions so far.

My current approach was briefly explained in:

http://www.spinics.net/lists/linux-mm/msg25340.html

(What does glommer think about kmem cgroup ?)
In this particular context, glommer being me.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>