KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx> writes:
> based on mmotm and other memory cgroup patches in -mm queue.
> ==
> From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
>
> This patch extracts core logic of mem_cgroup_update_file_mapped() as
> mem_cgroup_update_file_stat() and add a skin.
>
> As a planned future update, memory cgroup has to count dirty pages to implement
> dirty_ratio/limit. And more, the number of dirty pages is required to kick flusher
> thread to start writeback. (Now, no kick.)
>
> This patch is preparation for it and makes other statistics implementation
> clearer. Just a clean up.
>
> Note:
> In previous patch series, I wrote a more complicated patch to make the
> more generic and wanted to avoid using switch(). But now, we found page_mapped()
> check is necessary for updage_file_mapepd().We can't avoid to add some conditions.
> I hope this style is enough easy to read and to maintainance.
>
> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
> ---
> mm/memcontrol.c | 25 ++++++++++++++++++-------
> 1 file changed, 18 insertions(+), 7 deletions(-)
>
> Index: mmotm-0915/mm/memcontrol.c
> ===================================================================
> --- mmotm-0915.orig/mm/memcontrol.c
> +++ mmotm-0915/mm/memcontrol.c
> @@ -1575,7 +1575,8 @@ bool mem_cgroup_handle_oom(struct mem_cg
> * small, we check MEM_CGROUP_ON_MOVE percpu value and detect there are
> * possibility of race condition. If there is, we take a lock.
> */
> -void mem_cgroup_update_file_mapped(struct page *page, int val)
> +
> +static void mem_cgroup_update_file_stat(struct page *page, int idx, int val)
> {
> struct mem_cgroup *mem;
> struct page_cgroup *pc = lookup_page_cgroup(page);
> @@ -1597,13 +1598,18 @@ void mem_cgroup_update_file_mapped(struc
> if (!mem || !PageCgroupUsed(pc))
> goto out;
> }
> - if (val > 0) {
> - this_cpu_inc(mem->stat->count[MEM_CGROUP_STAT_FILE_MAPPED]);
> - SetPageCgroupFileMapped(pc);
> - } else {
> - this_cpu_dec(mem->stat->count[MEM_CGROUP_STAT_FILE_MAPPED]);
> - if (!page_mapped(page)) /* for race between dec->inc counter */
> +
> + this_cpu_add(mem->stat->count[idx], val);
> +
> + switch (idx) {
> + case MEM_CGROUP_STAT_FILE_MAPPED:
> + if (val > 0)
> + SetPageCgroupFileMapped(pc);
> + else if (!page_mapped(page))
> ClearPageCgroupFileMapped(pc);
> + break;
> + default:
> + BUG();
> }
>
> out:
> @@ -1613,6 +1619,11 @@ out:
> return;
> }
>
> +void mem_cgroup_update_file_mapped(struct page *page, int val)
> +{
> + mem_cgroup_update_file_stat(page, MEM_CGROUP_STAT_FILE_MAPPED, val);
> +}
> +
> /*
> * size of first charge trial. "32" comes from vmscan.c's magic value.
> * TODO: maybe necessary to use big numbers in big irons.
Reviewed-by: Greg Thelen <gthelen@xxxxxxxxxx>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>