Re: [PATCH 21/27] mm, vmscan: Only wakeup kswapd once per node for the requested classzone

[Vlastimil Babka]

On 06/09/2016 08:04 PM, Mel Gorman wrote:
> kswapd is woken when zones are below the low watermark but the wakeup
> decision is not taking the classzone into account.  Now that reclaim is
> node-based, it is only required to wake kswapd once per node and only if
> all zones are unbalanced for the requested classzone.
>
> Note that one node might be checked multiple times but there is no cheap
> way of tracking what nodes have already been visited for zoneslists that
> be ordered by either zone or node.

Wouldn't it be possible to optimize for node order as you did in direct reclaim? 
Do the zone_balanced checks when going through zonelist, and once node changes 
in iteration, wake up if no eligible zones visited so far were balanced.

> Signed-off-by: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>
> ---
>  mm/vmscan.c | 13 +++++++++++--
>  1 file changed, 11 insertions(+), 2 deletions(-)
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 7a2d69612231..b9cff9047ac0 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -3389,6 +3389,7 @@ static int kswapd(void *p)
>  void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
>  {
>  	pg_data_t *pgdat;
> +	int z;
>
>  	if (!populated_zone(zone))
>  		return;
> @@ -3402,8 +3403,16 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
>  	pgdat->kswapd_order = max(pgdat->kswapd_order, order);
>  	if (!waitqueue_active(&pgdat->kswapd_wait))
>  		return;
> -	if (zone_balanced(zone, order, 0))
> -		return;
> +
> +	/* Only wake kswapd if all zones are unbalanced */
> +	for (z = 0; z <= classzone_idx; z++) {
> +		zone = pgdat->node_zones + z;
> +		if (!populated_zone(zone))
> +			continue;
> +
> +		if (zone_balanced(zone, order, classzone_idx))
> +			return;
> +	}
>
>  	trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order);
>  	wake_up_interruptible(&pgdat->kswapd_wait);