Received-SPF: pass (google.com: domain of linux-kernel+bounces-144505-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Date: Mon, 15 Apr 2024 11:11:23 +0800 (CST)
Message-ID: <20240415111123924s9IbQkgHF8S4yZv4su8LI@zte.com.cn>
Precedence: bulk
Mime-Version: 1.0
From: <xu.xin16@zte.com.cn>
To: <v-songbaohua@oppo.com>, <mhocko@kernel.org>, <ryan.roberts@arm.com>,
        <david@redhat.com>
Cc: <roman.gushchin@linux.dev>, <shakeel.butt@linux.dev>,
        <muchun.song@linux.dev>, <akpm@linux-foundation.org>,
        <cgroups@vger.kernel.org>, <linux-mm@kvack.org>,
        <linux-kernel@vger.kernel.org>, <ran.xiaokai@zte.com.cn>,
        <xu.xin16@zte.com.cn>, <yang.yang29@zte.com.cn>,
        <lu.zhongjun@zte.com.cn>
Subject: =?UTF-8?B?W1BBVENIXSBtbTogdGhwOiBtYWtlcyB0aGUgbWVtY2cgVEhQIGRlZmVycmVkIHNwbGl0IHNocmlua2VyIGF3YXJlIG9mCgogbm9kZV9pZA==?=
Content-Type: text/plain;
	charset="UTF-8"

From: Ran Xiaokai <ran.xiaokai@zte.com.cn>

Since commit 87eaceb3faa5 ("mm: thp: make deferred split shrinker
memcg aware"), the THP deferred split queue is per memcg but not
per mem_cgroup_per_node. This has two aspects of impact:

Impact1: for kswapd reclaim
=====================
  kswapd
    balance_pgdat
      kswapd_shrink_node
        shrink_node(pgdat, sc);
          shrink_node_memcgs(pgdat, sc);
            shrink_slab(sc->gfp_mask, pgdat->node_id, memcg...);
  the parameter "pgdat->node_id"  does not take effectct for
  THP deferred_split_shrinker, as the deferred_split_queue of
  specified memcg is not for a certain numa node but for all the nodes.
  We want to makes the memcg THP deferred split shrinker aware of
  node_id.

Impact2: thp-deferred_split shrinker debugfs interface
=========================================
   for the "count" file:
   <cgroup inode id> <objects on node 0> <objects on node 1>
     the output is acctually the sum of all numa nodes.
   for the "scan" file:
   <cgroup inode id> <numa id> <number of objects to scan>
     Also the "numa id" input does not take effect here.

This patch makes memcg deferred_split_queue per mem_cgroup_per_node
so try to conform to semantic logic.

Reviewed-by: Lu Zhongjun <lu.zhongjun@zte.com.cn>
Signed-off-by: Ran Xiaokai <ran.xiaokai@zte.com.cn>
Cc: xu xin <xu.xin16@zte.com.cn>
Cc: Yang Yang <yang.yang29@zte.com.cn>
---
 include/linux/memcontrol.h |  7 +++----
 mm/huge_memory.c           |  6 +++---
 mm/memcontrol.c            | 11 +++++------
 3 files changed, 11 insertions(+), 13 deletions(-)

diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index 394fd0a887ae..7282861d5a5d 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -130,6 +130,9 @@ struct mem_cgroup_per_node {
 	bool			on_tree;
 	struct mem_cgroup	*memcg;		/* Back pointer, we cannot */
 						/* use container_of	   */
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+	struct deferred_split deferred_split_queue;
+#endif
 };

 struct mem_cgroup_threshold {
@@ -327,10 +330,6 @@ struct mem_cgroup {
 	struct list_head event_list;
 	spinlock_t event_list_lock;

-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
-	struct deferred_split deferred_split_queue;
-#endif
-
 #ifdef CONFIG_LRU_GEN_WALKS_MMU
 	/* per-memcg mm_struct list */
 	struct lru_gen_mm_list mm_list;
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 9859aa4f7553..338d071070a6 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -774,7 +774,7 @@ struct deferred_split *get_deferred_split_queue(struct folio *folio)
 	struct pglist_data *pgdat = NODE_DATA(folio_nid(folio));

 	if (memcg)
-		return &memcg->deferred_split_queue;
+		return &memcg->nodeinfo[pgdat->node_id]->deferred_split_queue;
 	else
 		return &pgdat->deferred_split_queue;
 }
@@ -3305,7 +3305,7 @@ static unsigned long deferred_split_count(struct shrinker *shrink,

 #ifdef CONFIG_MEMCG
 	if (sc->memcg)
-		ds_queue = &sc->memcg->deferred_split_queue;
+		ds_queue = &sc->memcg->nodeinfo[sc->nid]->deferred_split_queue;
 #endif
 	return READ_ONCE(ds_queue->split_queue_len);
 }
@@ -3322,7 +3322,7 @@ static unsigned long deferred_split_scan(struct shrinker *shrink,

 #ifdef CONFIG_MEMCG
 	if (sc->memcg)
-		ds_queue = &sc->memcg->deferred_split_queue;
+		ds_queue = &sc->memcg->nodeinfo[sc->nid]->deferred_split_queue;
 #endif

 	spin_lock_irqsave(&ds_queue->split_queue_lock, flags);
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index fabce2b50c69..cdf9f5fa3b8e 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -5445,7 +5445,11 @@ static int alloc_mem_cgroup_per_node_info(struct mem_cgroup *memcg, int node)
 		kfree(pn);
 		return 1;
 	}
-
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+	spin_lock_init(&pn->deferred_split_queue.split_queue_lock);
+	INIT_LIST_HEAD(&pn->deferred_split_queue.split_queue);
+	pn->deferred_split_queue.split_queue_len = 0;
+#endif
 	lruvec_init(&pn->lruvec);
 	pn->memcg = memcg;

@@ -5545,11 +5549,6 @@ static struct mem_cgroup *mem_cgroup_alloc(struct mem_cgroup *parent)
 	for (i = 0; i < MEMCG_CGWB_FRN_CNT; i++)
 		memcg->cgwb_frn[i].done =
 			__WB_COMPLETION_INIT(&memcg_cgwb_frn_waitq);
-#endif
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
-	spin_lock_init(&memcg->deferred_split_queue.split_queue_lock);
-	INIT_LIST_HEAD(&memcg->deferred_split_queue.split_queue);
-	memcg->deferred_split_queue.split_queue_len = 0;
 #endif
 	lru_gen_init_memcg(memcg);
 	return memcg;
-- 
2.15.2