memcg: make memcg kswapd deal with dirty

hulk inclusion
category: feature
bugzilla: https://gitee.com/openeuler/kernel/issues/I4IMAK?from=project-issue
CVE: NA

--------

The memcg kswapd could set dirty state to memcg if current scan find all
pages are unqueued dirty in the memcg. Then kswapd would write out dirty pages.
Signed-off-by: NLu Jialin <lujialin4@huawei.com>
Reviewed-by: NKefeng Wang <wangkefeng.wang@huawei.com>
Reviewed-by: Nweiyang wang <wangweiyang2@huawei.com>
Signed-off-by: NZheng Zengkai <zhengzengkai@huawei.com>

include/linux/mmzone.h

@@ -272,6 +272,10 @@ enum lruvec_flags {
 	LRUVEC_CONGESTED,		/* lruvec has many dirty pages
 					 * backed by a congested BDI
 					 */
+	LRUVEC_DIRTY,			/* reclaim scanning has recently found
+					 * many dirty file pages at the tail of
+					 * the LRU.
+					 */
 };
 
 struct lruvec {
@@ -595,10 +599,6 @@ struct zone {
 } ____cacheline_internodealigned_in_smp;
 
 enum pgdat_flags {
-	PGDAT_DIRTY,			/* reclaim scanning has recently found
-					 * many dirty file pages at the tail
-					 * of the LRU.
-					 */
 	PGDAT_WRITEBACK,		/* reclaim scanning has recently found
 					 * many pages under writeback
 					 */

mm/vmscan.c

@@ -1289,6 +1289,7 @@ static unsigned int shrink_page_list(struct list_head *page_list,
 	LIST_HEAD(free_pages);
 	unsigned int nr_reclaimed = 0;
 	unsigned int pgactivate = 0;
+	struct lruvec *target_lruvec = mem_cgroup_lruvec(sc->target_mem_cgroup, pgdat);
 
 	memset(stat, 0, sizeof(*stat));
 	cond_resched();
@@ -1535,7 +1536,7 @@ static unsigned int shrink_page_list(struct list_head *page_list,
 			 */
 			if (page_is_file_lru(page) &&
 			    (!current_is_kswapd() || !PageReclaim(page) ||
-			     !test_bit(PGDAT_DIRTY, &pgdat->flags))) {
+			     !test_bit(LRUVEC_DIRTY, &target_lruvec->flags))) {
 				/*
 				 * Immediately reclaim when written back.
 				 * Similar in principal to deactivate_page()
@@ -3068,7 +3069,7 @@ static void shrink_node(pg_data_t *pgdat, struct scan_control *sc)
 
 		/* Allow kswapd to start writing pages during reclaim.*/
 		if (sc->nr.unqueued_dirty == sc->nr.file_taken)
-			set_bit(PGDAT_DIRTY, &pgdat->flags);
+			set_bit(LRUVEC_DIRTY, &target_lruvec->flags);
 
 		/*
 		 * If kswapd scans pages marked for immediate
@@ -3088,7 +3089,7 @@ static void shrink_node(pg_data_t *pgdat, struct scan_control *sc)
 	 * Legacy memcg will stall in page writeback so avoid forcibly
 	 * stalling in wait_iff_congested().
 	 */
-	if ((current_is_kswapd() ||
+	if (((current_is_kswapd() && !cgroup_reclaim(sc))||
 	     (cgroup_reclaim(sc) && writeback_throttling_sane(sc))) &&
 	    sc->nr.dirty && sc->nr.dirty == sc->nr.congested)
 		set_bit(LRUVEC_CONGESTED, &target_lruvec->flags);
@@ -3322,6 +3323,8 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist,
 			lruvec = mem_cgroup_lruvec(sc->target_mem_cgroup,
 						   zone->zone_pgdat);
 			clear_bit(LRUVEC_CONGESTED, &lruvec->flags);
+			if (current_is_kswapd())
+				clear_bit(LRUVEC_DIRTY, &lruvec->flags);
 		}
 	}
 
@@ -3712,7 +3715,7 @@ static void clear_pgdat_congested(pg_data_t *pgdat)
 	struct lruvec *lruvec = mem_cgroup_lruvec(NULL, pgdat);
 
 	clear_bit(LRUVEC_CONGESTED, &lruvec->flags);
-	clear_bit(PGDAT_DIRTY, &pgdat->flags);
+	clear_bit(LRUVEC_DIRTY, &pgdat->flags);
 	clear_bit(PGDAT_WRITEBACK, &pgdat->flags);
 }