Re: [PATCH] docs: cgroup: mm: Document why inactive_X + active_X may not equal X

[Michal Hocko]

On Mon 11-11-19 14:49:58, Chris Down wrote:
> This has confused a significant number of people using cgroups inside
> Facebook, and some of those outside as well judging by posts like
> this[0] (although it's not a problem unique to cgroup v2). If shmem
> handling in particular becomes more coherent at some point in the future
> -- although that seems unlikely now -- we can change the wording here.
> 
> [0]: https://unix.stackexchange.com/q/525092/10762
> 
> Signed-off-by: Chris Down <chris@xxxxxxxxxxxxxx>
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: Johannes Weiner <hannes@xxxxxxxxxxx>
> Cc: Tejun Heo <tj@xxxxxxxxxx>
> Cc: Roman Gushchin <guro@xxxxxx>
> Cc: Michal Hocko <mhocko@xxxxxxxxxx>
> Cc: linux-kernel@xxxxxxxxxxxxxxx
> Cc: cgroups@xxxxxxxxxxxxxxx
> Cc: linux-mm@xxxxxxxxx
> Cc: kernel-team@xxxxxx

Acked-by: Michal Hocko <mhocko@xxxxxxxx>

> ---
>  Documentation/admin-guide/cgroup-v2.rst | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
> 
> diff --git a/Documentation/admin-guide/cgroup-v2.rst b/Documentation/admin-guide/cgroup-v2.rst
> index 0704552ed94f..0636bcb60b5a 100644
> --- a/Documentation/admin-guide/cgroup-v2.rst
> +++ b/Documentation/admin-guide/cgroup-v2.rst
> @@ -1289,7 +1289,12 @@ PAGE_SIZE multiple when read back.
>  	  inactive_anon, active_anon, inactive_file, active_file, unevictable
>  		Amount of memory, swap-backed and filesystem-backed,
>  		on the internal memory management lists used by the
> -		page reclaim algorithm
> +		page reclaim algorithm.
> +
> +		As these represent internal list state (eg. shmem pages are on anon
> +		memory management lists), inactive_foo + active_foo may not be equal to

and anon will move to file list after MADV_FREE.

> +		the value for the foo counter, since the foo counter is type-based, not
> +		list-based.
>  
>  	  slab_reclaimable
>  		Part of "slab" that might be reclaimed, such as
> -- 
> 2.24.0

-- 
Michal Hocko
SUSE Labs