[PATCH 3/4] mm: memcg: implement unbalance proactive reclaim

[Huan Yang]

This patch add swappiness arg into proactive reclaim interface.
User can type "bytes swappiness" into memory.reclaim to control
reclaim tendency.
Specially, user can type below to only reclaim anon folios:
`"100M" 200 > memory.reclaim`
Or, type below to only reclaim file folios:
`100M 1 > memory.reclaim`

User can only type key into memory.reclaim to keep original use.

Signed-off-by: Huan Yang <link@xxxxxxxx>
---
 Documentation/admin-guide/cgroup-v2.rst | 16 ++++---
 include/linux/swap.h                    |  2 +
 mm/memcontrol.c                         | 55 +++++++++++++++++++++----
 mm/vmscan.c                             |  2 +
 4 files changed, 63 insertions(+), 12 deletions(-)

diff --git a/Documentation/admin-guide/cgroup-v2.rst b/Documentation/admin-guide/cgroup-v2.rst
index c153d0c75f34..5a471ac7f0c3 100644
--- a/Documentation/admin-guide/cgroup-v2.rst
+++ b/Documentation/admin-guide/cgroup-v2.rst
@@ -1275,16 +1275,22 @@ PAGE_SIZE multiple when read back.
 	This is a simple interface to trigger memory reclaim in the
 	target cgroup.
 
-	This file accepts a single key, the number of bytes to reclaim.
-	No nested keys are currently supported.
+	This file accepts a few key, the number of bytes to reclaim.
+	Few nested keys are currently supported.
 
 	Example::
 
 	  echo "1G" > memory.reclaim
 
-	The interface can be later extended with nested keys to
-	configure the reclaim behavior. For example, specify the
-	type of memory to reclaim from (anon, file, ..).
+	The interface extended with nested keys to configure the
+	reclaim behavior. For example, specify the swappiness of
+	memory to reclaim from (anon, file, ..).
+
+	Example::
+
+	  echo "1G" 200 > memory.reclaim (only reclaim anon)
+	  echo "1G" 0  > memory.reclaim (only reclaim file)
+	  echo "1G" 1  > memory.reclaim (only reclaim file)
 
 	Please note that the kernel can over or under reclaim from
 	the target cgroup. If less bytes are reclaimed than the
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 3ba146ae7cf5..a024194301d4 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -407,6 +407,8 @@ extern unsigned long try_to_free_pages(struct zonelist *zonelist, int order,
 
 #define MEMCG_RECLAIM_MAY_SWAP (1 << 1)
 #define MEMCG_RECLAIM_PROACTIVE (1 << 2)
+#define MEMCG_RECLAIM_ANON (1 << 3)
+#define MEMCG_RECLAIM_FILE (1 << 4)
 extern unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg,
 						  unsigned long nr_pages,
 						  gfp_t gfp_mask,
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index be2ad117515e..a0e460abd41c 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -6953,19 +6953,47 @@ static ssize_t memory_reclaim(struct kernfs_open_file *of, char *buf,
 	unsigned int nr_retries = MAX_RECLAIM_RETRIES;
 	unsigned long nr_to_reclaim, nr_reclaimed = 0;
 	unsigned int reclaim_options;
+	int swappiness = -1, org_swappiness, n;
+	char *tmpbuf;
 	int err;
 
-	buf = strstrip(buf);
-	err = page_counter_memparse(buf, "", &nr_to_reclaim);
+	tmpbuf = kvzalloc(nbytes, GFP_KERNEL);
+	if (unlikely(!tmpbuf))
+		return -ENOMEM;
+
+	buf = skip_spaces(buf);
+	n = sscanf(buf, "%s %d", tmpbuf, &swappiness);
+	if (n < 1) {
+		err = -EINVAL;
+		goto out_free;
+	}
+
+	if (n == 2 && (swappiness > 200 || swappiness < 0)) {
+		err = -EINVAL;
+		goto out_free;
+	}
+
+	err = page_counter_memparse(tmpbuf, "", &nr_to_reclaim);
 	if (err)
-		return err;
+		goto out_free;
 
 	reclaim_options	= MEMCG_RECLAIM_MAY_SWAP | MEMCG_RECLAIM_PROACTIVE;
+	if (swappiness != -1) {
+		org_swappiness = memcg->swappiness;
+		memcg->swappiness = swappiness;
+		if (swappiness == 200)
+			reclaim_options |= MEMCG_RECLAIM_ANON;
+		else if (swappiness == 0 || swappiness == 1)
+			reclaim_options |= MEMCG_RECLAIM_FILE;
+	}
+
 	while (nr_reclaimed < nr_to_reclaim) {
 		unsigned long reclaimed;
 
-		if (signal_pending(current))
-			return -EINTR;
+		if (signal_pending(current)) {
+			err = -EINTR;
+			goto out;
+		}
 
 		/*
 		 * This is the final attempt, drain percpu lru caches in the
@@ -6979,13 +7007,26 @@ static ssize_t memory_reclaim(struct kernfs_open_file *of, char *buf,
 					min(nr_to_reclaim - nr_reclaimed, SWAP_CLUSTER_MAX),
 					GFP_KERNEL, reclaim_options);
 
-		if (!reclaimed && !nr_retries--)
-			return -EAGAIN;
+		if (!reclaimed && !nr_retries--) {
+			err = -EAGAIN;
+			goto out;
+		}
 
 		nr_reclaimed += reclaimed;
 	}
 
+	if (swappiness != -1)
+		memcg->swappiness = org_swappiness;
+
 	return nbytes;
+
+out:
+	if (swappiness != -1)
+		memcg->swappiness = org_swappiness;
+
+out_free:
+	kvfree(tmpbuf);
+	return err;
 }
 
 static struct cftype memory_files[] = {
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 9243a1f0d606..f4221ec833db 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -6505,6 +6505,8 @@ unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg,
 		.may_unmap = 1,
 		.may_swap = !!(reclaim_options & MEMCG_RECLAIM_MAY_SWAP),
 		.proactive = !!(reclaim_options & MEMCG_RECLAIM_PROACTIVE),
+		.unbalance_anon = !!(reclaim_options & MEMCG_RECLAIM_ANON),
+		.unbalance_file = !!(reclaim_options & MEMCG_RECLAIM_FILE),
 	};
 	/*
 	 * Traverse the ZONELIST_FALLBACK zonelist of the current node to put
-- 
2.34.1