Re: [RFC][PATCH] Remove cgroup member from struct page

[Balbir Singh <balbir@linux.vnet.ibm.com>]

KAMEZAWA Hiroyuki wrote:
> On Mon, 8 Sep 2008 20:58:10 +0530
> Balbir Singh <balbir@linux.vnet.ibm.com> wrote:
>> Sorry for the delay in sending out the new patch, I am traveling and
>> thus a little less responsive. Here is the update patch
>>
>>
> Hmm.. I've considered this approach for a while and my answer is that
> this is not what you really want.
> 
> Because you just moves the placement of pointer from memmap to
> radix_tree both in GFP_KERNEL, total kernel memory usage is not changed.

Agreed, but we do reduce the sizeof(struct page) without adding on to
page_cgroup's size. So why don't we want this?

> So, at least, you have to add some address calculation (as I did in March)
> to getting address of page_cgroup.

What address calculation do we need, sorry I don't recollect it.

 But page_cgroup itself consumes 32bytes
> per page. Then.....
> 
> My proposal to 32bit system is following 
>  - remove page_cgroup completely.
>    - As a result, there is no per-cgroup lru. But it will not be bad
>      bacause the number of cgroups and pages are not big.
>      just a trade-off between kernel-memory-space v.s. speed.

32 bit systems with PAE can support quite a lot of memory, so I am not sure I
agree. I don't like this approach

>    - Removing page_cgroup and just remember address of mem_cgroup per page.
> 

This is on top of the suggested approach above?

> How do you think ?
> 

I don't like the approach.

-- 
	Balbir