On Mon, Oct 26, 2020 at 03:47:57PM +0100, Michal Hocko wrote: > On Mon 26-10-20 22:26:35, zhongjiang-ali wrote: > > memcg_page_state will get the specified number in hierarchical memcg, > > It should multiply by HPAGE_PMD_NR rather than an page if the item is > > NR_ANON_THPS. > > I am not sure which tree are you looking at but the current Linus tree > already does have this hunk. This tripped up my pattern matching as well. You go to the code and you see this bit already there. But it's only there for the local stats, not the hierarchical stats, and the code for them is nearly identical - except "%s" vs "total_%s" and memcg_page_state_local() vs memcg_page_state(). Zhongjiang is updating the hierarchical stats a few lines below where you see that hunk.
