Re: [PATCH][RESEND] nommu: add anonymous page memcg accounting

[KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>]

On Fri, 01 Oct 2010 11:41:07 -0500
"Steven J. Magnani" <steve@digidescorp.com> wrote:
> > However, I suppose there's little harm in letting the patch in.  I would guess
> > the additions all optimise away if memcg isn't enabled.
> > 
> > A question for you: why does struct page_cgroup need a page pointer?  If an
> > array of page_cgroup structs is allocated per array of page structs, then you
> > should be able to use the array index to map between them.
> 
No reason. It was not array in the 1st implemenation and ->page still remains. At 2nd
implementation, I didn't know embeded people has any interests on memcg. And I wasn't
sure how
	page_cgroup_to_page() : pfn_to_page(pagec_cgroup_to_pfn(pc))
will be widely used.

Now, we know page_cgroup->page is not used in very critical path _if_ node-id
and zone-id can be directly got from page_cgroup.

I'm now preparing a patch to remove struct page* pointer. I'm wondering
whether it's ok that some architecuture cannot drop struct page pointer.
If SPARSEMEM is used on 32bit arch, I'm not sure whether # of bits isn't enough.
I may have to add overhead to get nid, zid in critical path.
(for example, s390/32bit, x86-32/HIGHMEM, ARM/HIGHMEM?)

Current out priority is supporting dirty_ratio rather than memory usage diet.
Please wait. Removing page_cgroup->page patch will add something a bit complex.


Thanks,
-Kame

--
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/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>