Re: [PATCH] memcg: flush lruvec stats in the refault

On Wed, Sep 22, 2021 at 3:50 PM Shakeel Butt [off-list ref] wrote:

From the above result, it seems like the option-2 not only solves the
regression but also improves the performance for at least these
benchmarks.

Feng Tang (intel) ran the aim7 benchmark with these two options and
confirms that option-1 reduces the regression but option-2 removes the
regression.

Michael Larabel (phoronix) ran multiple benchmarks with these options
and reported the results at [3] and it shows for most benchmarks
option-2 removes the regression introduced by the commit aa48e47e3906
("memcg: infrastructure to flush memcg stats").

Ok, I've applied this just to close the issue.

If somebody comes up with more data and the delayed flushing or
something is problematic, we'll revisit, but this looks all sane to me
and fixes the regression.

Thanks,
              Linus