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

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

KAMEZAWA Hiroyuki wrote:
> On Tue, 9 Sep 2008 15:00:10 +1000
> Nick Piggin <nickpiggin@yahoo.com.au> wrote:
>>> maybe a routine like SPARSEMEM is a choice.
>>>
>>> Following is pointer pre-allocation. (just pointer, not page_cgroup itself)
>>> ==
>>> #define PCG_SECTION_SHIFT	(10)
>>> #define PCG_SECTION_SIZE	(1 << PCG_SECTION_SHIFT)
>>>
>>> struct pcg_section {
>>> 	struct page_cgroup **map[PCG_SECTION_SHIFT]; //array of pointer.
>>> };
>>>
>>> struct page_cgroup *get_page_cgroup(unsigned long pfn)
>>> {
>>> 	struct pcg_section *sec;
>>> 	sec = pcg_section[(pfn >> PCG_SECTION_SHIFT)];
>>> 	return *sec->page_cgroup[(pfn & ((1 << PCG_SECTTION_SHIFT) - 1];
>>> }
>>> ==
>>> If we go extreme, we can use kmap_atomic() for pointer array.
>>>
>>> Overhead of pointer-walk is not so bad, maybe.
>>>
>>> For 64bit systems, we can find a way like SPARSEMEM_VMEMMAP.
>> Yes I too think that would be the ideal way to go to get the best of
>> performance in the enabled case. However Balbir I believe is interested
>> in memory savings if not all pages have cgroups... I don't know, I don't
>> care so much about the "enabled" case, so I'll leave you two to fight it
>> out :)
>>
> I'll add a new patch on my set.
> 
> Balbir, are you ok to CONFIG_CGROUP_MEM_RES_CTLR depends on CONFIG_SPARSEMEM ?
> I thinks SPARSEMEM(SPARSEMEM_VMEMMAP) is widely used in various archs now.

Can't we make it more generic. I was thinking of allocating memory for each node
for page_cgroups (of the size of spanned_pages) at initialization time. I've not
yet prototyped the idea. BTW, even with your approach I fail to see why we need
to add a dependency on CONFIG_SPARSEMEM (but again it is 4:30 in the morning and
I might be missing the obvious)

-- 
	Balbir