Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.31 as permitted sender) client-ip=23.128.96.31;
From:   Nhat Pham <nphamcs@gmail.com>
To:     akpm@linux-foundation.org
Cc:     hannes@cmpxchg.org, cerasuolodomenico@gmail.com,
        yosryahmed@google.com, sjenning@redhat.com, ddstreet@ieee.org,
        vitaly.wool@konsulko.com, mhocko@kernel.org,
        roman.gushchin@linux.dev, shakeelb@google.com,
        muchun.song@linux.dev, chrisl@kernel.org, linux-mm@kvack.org,
        kernel-team@meta.com, linux-kernel@vger.kernel.org,
        cgroups@vger.kernel.org, linux-doc@vger.kernel.org,
        linux-kselftest@vger.kernel.org, shuah@kernel.org
Subject: [PATCH v7 2/6] memcontrol: add a new function to traverse online-only memcg hierarchy
Date:   Mon, 27 Nov 2023 15:45:56 -0800
Message-Id: <20231127234600.2971029-3-nphamcs@gmail.com>
In-Reply-To: <20231127234600.2971029-1-nphamcs@gmail.com>
References: <20231127234600.2971029-1-nphamcs@gmail.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

The new zswap writeback scheme requires an online-only memcg hierarchy
traversal. Add this functionality via the new mem_cgroup_iter_online()
function - the old mem_cgroup_iter() is a special case of this new
function.

Suggested-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Nhat Pham <nphamcs@gmail.com>
---
 include/linux/memcontrol.h | 13 +++++++++++++
 mm/memcontrol.c            | 29 +++++++++++++++++++++++++----
 2 files changed, 38 insertions(+), 4 deletions(-)

diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index 7bdcf3020d7a..5bfe41840e80 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -833,6 +833,10 @@ static inline void mem_cgroup_put(struct mem_cgroup *memcg)
 struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *,
 				   struct mem_cgroup *,
 				   struct mem_cgroup_reclaim_cookie *);
+struct mem_cgroup *mem_cgroup_iter_online(struct mem_cgroup *root,
+				   struct mem_cgroup *prev,
+				   struct mem_cgroup_reclaim_cookie *reclaim,
+				   bool online);
 void mem_cgroup_iter_break(struct mem_cgroup *, struct mem_cgroup *);
 void mem_cgroup_scan_tasks(struct mem_cgroup *memcg,
 			   int (*)(struct task_struct *, void *), void *arg);
@@ -1386,6 +1390,15 @@ mem_cgroup_iter(struct mem_cgroup *root,
 	return NULL;
 }
 
+static inline struct mem_cgroup *
+mem_cgroup_iter_online(struct mem_cgroup *root,
+		struct mem_cgroup *prev,
+		struct mem_cgroup_reclaim_cookie *reclaim,
+		bool online)
+{
+	return NULL;
+}
+
 static inline void mem_cgroup_iter_break(struct mem_cgroup *root,
 					 struct mem_cgroup *prev)
 {
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 470821d1ba1a..b1fb96233fa1 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -1111,10 +1111,11 @@ struct mem_cgroup *get_mem_cgroup_from_current(void)
 }
 
 /**
- * mem_cgroup_iter - iterate over memory cgroup hierarchy
+ * mem_cgroup_iter_online - iterate over memory cgroup hierarchy
  * @root: hierarchy root
  * @prev: previously returned memcg, NULL on first invocation
  * @reclaim: cookie for shared reclaim walks, NULL for full walks
+ * @online: whether to skip offline memcgs
  *
  * Returns references to children of the hierarchy below @root, or
  * @root itself, or %NULL after a full round-trip.
@@ -1123,13 +1124,16 @@ struct mem_cgroup *get_mem_cgroup_from_current(void)
  * invocations for reference counting, or use mem_cgroup_iter_break()
  * to cancel a hierarchy walk before the round-trip is complete.
  *
+ * Caller can skip offline memcgs by passing true for @online.
+ *
  * Reclaimers can specify a node in @reclaim to divide up the memcgs
  * in the hierarchy among all concurrent reclaimers operating on the
  * same node.
  */
-struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root,
+struct mem_cgroup *mem_cgroup_iter_online(struct mem_cgroup *root,
 				   struct mem_cgroup *prev,
-				   struct mem_cgroup_reclaim_cookie *reclaim)
+				   struct mem_cgroup_reclaim_cookie *reclaim,
+				   bool online)
 {
 	struct mem_cgroup_reclaim_iter *iter;
 	struct cgroup_subsys_state *css = NULL;
@@ -1199,7 +1203,8 @@ struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root,
 		 * is provided by the caller, so we know it's alive
 		 * and kicking, and don't take an extra reference.
 		 */
-		if (css == &root->css || css_tryget(css)) {
+		if (css == &root->css || (!online && css_tryget(css)) ||
+				css_tryget_online(css)) {
 			memcg = mem_cgroup_from_css(css);
 			break;
 		}
@@ -1228,6 +1233,22 @@ struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root,
 	return memcg;
 }
 
+/**
+ * mem_cgroup_iter - iterate over memory cgroup hierarchy
+ * @root: hierarchy root
+ * @prev: previously returned memcg, NULL on first invocation
+ * @reclaim: cookie for shared reclaim walks, NULL for full walks
+ *
+ * Perform an iteration on the memory cgroup hierarchy without skipping
+ * offline memcgs.
+ */
+struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root,
+				   struct mem_cgroup *prev,
+				   struct mem_cgroup_reclaim_cookie *reclaim)
+{
+	return mem_cgroup_iter_online(root, prev, reclaim, false);
+}
+
 /**
  * mem_cgroup_iter_break - abort a hierarchy walk prematurely
  * @root: hierarchy root
-- 
2.34.1