Re: [PATCH 0/2] page_cgroup cleanups

[Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>]

On 04/05/2013 02:01 PM, Glauber Costa wrote:
> Hi,
> 
> Last time I sent the mem cgroup bypass patches, Kame and Michal pointed out
> that some of it was a bit of cleanup, specifically at the page_cgroup side.
> I've decided to separate those patches and send them separately. After these
> patches are applied, page_cgroup will be initialized together with the root
> cgroup, instead of init/main.c
> 
> When we move cgroup initialization to the first non-root cgroup created, all
> we'll have to do from the page_cgroup side would be to move the initialization
> that now happens at root, to the first child.
> 
> Glauber Costa (2):
>   memcg: consistently use vmalloc for page_cgroup allocations
>   memcg: defer page_cgroup initialization
> 
>  include/linux/page_cgroup.h | 21 +------------------
>  init/main.c                 |  2 --
>  mm/memcontrol.c             |  2 ++
>  mm/page_cgroup.c            | 51 +++++++++++++++------------------------------
>  4 files changed, 20 insertions(+), 56 deletions(-)
> 
FYI: There are kbuild warnings with this. I wanted to send it earlier to
see what people think. If there is no changes requested, please let me
know I will send a new version with just the kbuild fixes folded.