On Sat, Apr 14, 2012 at 5:41 AM, Hillf Danton <dhillf@xxxxxxxxx> wrote:
> On Thu, Apr 12, 2012 at 5:59 AM, Ying Han <yinghan@xxxxxxxxxx> wrote:
>> This patch reverts all the existing softlimit reclaim implementations.
>>
>> Signed-off-by: Ying Han <yinghan@xxxxxxxxxx>
>> ---
>> include/linux/memcontrol.h | 11 --
>> include/linux/swap.h | 4 -
>> mm/memcontrol.c | 387 --------------------------------------------
>> mm/vmscan.c | 67 --------
>> 4 files changed, 0 insertions(+), 469 deletions(-)
>>
> [...]
>
>> -unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
>> - gfp_t gfp_mask,
>> - unsigned long *total_scanned)
>> -{
>> - unsigned long nr_reclaimed = 0;
>> - struct mem_cgroup_per_zone *mz, *next_mz = NULL;
>> - unsigned long reclaimed;
>> - int loop = 0;
>> - struct mem_cgroup_tree_per_zone *mctz;
>> - unsigned long long excess;
>> - unsigned long nr_scanned;
>> -
>> - if (order > 0)
>> - return 0;
>> -
> Not related to this patch, what is the functionality to check order?
The new implementation doesn't need to check order like before.
The existing soft_limit reclaim skips high order allocation since it's
been best effort before the actual reclaim on global lru. Now the
global lru is gone and all reclaims happens on per-memcg level. The
high order page allocation is being taken care of automatically like
other part of reclaim logic.
--Ying
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href