Re: [PATCH V7 1/2] cgroup/rstat: Avoid thundering herd problem by kswapd across NUMA nodes

[Shakeel Butt <shakeel.butt@linux.dev>]

Hi Jesper,

On Wed, Jul 17, 2024 at 06:36:28PM GMT, Jesper Dangaard Brouer wrote:
> 
[...]
> 
> 
> Looking at the production numbers for the time the lock is held for level 0:
> 
> @locked_time_level[0]:
> [4M, 8M)     623 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@               |
> [8M, 16M)    860 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@|
> [16M, 32M)   295 |@@@@@@@@@@@@@@@@@                                   |
> [32M, 64M)   275 |@@@@@@@@@@@@@@@@                                    |
> 

Is it possible to get the above histogram for other levels as well? I
know this is 12 numa node machine, how many total CPUs are there?

> The time is in nanosec, so M corresponds to ms (milliseconds).
> 
> With 36 flushes per second (as shown earlier) this is a flush every
> 27.7ms.  It is not unreasonable (from above data) that the flush time
> also spend 27ms, which means that we spend a full CPU second flushing.
> That is spending too much time flushing.

One idea to further reduce this time is more fine grained flush
skipping. At the moment we either skip the whole flush or not. How
about we make this decision per-cpu? We already have per-cpu updates
data and if it is less than MEMCG_CHARGE_BATCH, skip flush on that cpu.