On Tue, Jul 16, 2013 at 1:58 AM, Greg Thelen <gthelen@xxxxxxxxxx> wrote:
> On Thu, Jul 11, 2013 at 7:56 AM, Michal Hocko <mhocko@xxxxxxx> wrote:
>> On Sat 06-07-13 01:33:43, Sha Zhengju wrote:
>>> From: Sha Zhengju <handai.szj@xxxxxxxxxx>
>>>
>>> If memcg is enabled and no non-root memcg exists, all allocated
>>> pages belongs to root_mem_cgroup and wil go through root memcg
>>> statistics routines. So in order to reduce overheads after adding
>>> memcg dirty/writeback accounting in hot paths, we use jump label to
>>> patch mem_cgroup_{begin,end}_update_page_stat() in or out when not
>>> used.
>>
>> I do not think this is enough. How much do you save? One atomic read.
>> This doesn't seem like a killer.
>
> Given we're already using mem_cgroup_{begin,end}_update_page_stat(),
> this optimization seems independent of memcg dirty/writeback
> accounting. Does this patch help memcg even before dirty/writeback
> accounting? If yes, then we have the option of splitting this
> optimization out of the series.
Set_page_dirty is a hot path, people said I should be careful to the
overhead of adding a new counting, and the optimization is a must
before merging.
But since we have more need of this feature now, if it's blocking
something, I'm willing to split it.
--
Thanks,
Sha
--
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/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>