Re: [RFI] Shared accounting for memory resource controller

[Bharata B Rao]

On Wed, Apr 8, 2009 at 1:04 PM, KAMEZAWA Hiroyuki
<kamezawa.hiroyu@xxxxxxxxxxxxxx> wrote:
> On Wed, 8 Apr 2009 13:01:15 +0530
> Bharata B Rao <bharata.rao@xxxxxxxxx> wrote:
>
>> On Wed, Apr 8, 2009 at 12:48 PM, KAMEZAWA Hiroyuki
>> <kamezawa.hiroyu@xxxxxxxxxxxxxx> wrote:
>> >
>> > On Wed, 8 Apr 2009 12:41:15 +0530
>> > Balbir Singh <balbir@xxxxxxxxxxxxxxxxxx> wrote:
>> > > 3. Using the above, we can then try to (using an algorithm you
>> > > proposed), try to do some work for figuring out the shared percentage.
>> > >
>> > This is the point. At last. Why "# of shared pages" is important ?
>> >
>> > I wonder it's better to add new stat file as memory.cacheinfo which helps
>> > following kind of commands.
>> >
>> >  #cacheinfo /cgroups/memory/group01/
>> >       /usr/lib/libc.so.1     30pages
>> >       /var/log/messages      1 pages
>> >       /tmp/xxxxxx            20 pages
>>
>> Can I suggest that we don't add new files for additional stats and try
>> as far as possible to include them in <controller>.stat file. Please
>> note that we have APIs in libcgroup library which can return
>> statistics from controllers associated with a cgroup and these APIs
>> assume that stats are part of <controller>.stat file.
>>
> Hmm ? Is there generic assumption as all cgroup has "stat" file ?

No. But I would think if any controller has any stats to export, it
would do so via <controller>.stat file.

> And libcgroup cause bug if the new entry is added to stat file ?

No. It can cope with new entries being added to stat file as long as
they appear as (name, value) pairs.

> (IOW, libcgroup can't ignore new entry added ?)

It won't ignore, but can read the new stat.

Regards,
Bharata.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/