Re: [PATCH mm-unstable v1 2/4] mm/mglru: try to stop at high watermarks

On Fri, Dec 8, 2023 at 4:00 AM Hillf Danton hdanton@sina.com wrote:

On Thu, 7 Dec 2023 23:14:05 -0700 Yu Zhao yuzhao@google.com

...

-static unsigned long get_nr_to_reclaim(struct scan_control *sc) +static bool should_abort_scan(struct lruvec *lruvec, struct scan_control *sc) {

• int i;
  

• enum zone_watermarks mark;
  

• /* don't abort memcg reclaim to ensure fairness */
  if (!root_reclaim(sc))
  

•         return -1;
  

•         return false;
  
  

• return max(sc->nr_to_reclaim, compact_gap(sc->order));
  

• if (sc->nr_reclaimed >= max(sc->nr_to_reclaim, compact_gap(sc->order)))
  

•         return true;
  

• /* check the order to exclude compaction-induced reclaim */
  

• if (!current_is_kswapd() || sc->order)
  

•         return false;
  

• mark = sysctl_numa_balancing_mode & NUMA_BALANCING_MEMORY_TIERING ?
  

•        WMARK_PROMO : WMARK_HIGH;
  

• for (i = 0; i <= sc->reclaim_idx; i++) {
  

•         struct zone *zone = lruvec_pgdat(lruvec)->node_zones + i;
  

•         unsigned long size = wmark_pages(zone, mark) + MIN_LRU_BATCH;
  

•         if (managed_zone(zone) && !zone_watermark_ok(zone, 0, size, sc->reclaim_idx, 0))
  

•                 return false;
  

• }
  

• /* kswapd should abort if all eligible zones are safe */
  

This comment does not align with 86c79f6b5426 ("mm: vmscan: do not reclaim from kswapd if there is any eligible zone"). Any thing special here?

I don't see how they are not: they essentially say the same thing ("no more than needed") but with different units: zones or pages. IOW, don't reclaim from more zones or pages than needed.