Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.34 as permitted sender) client-ip=23.128.96.34;
Received-SPF: Pass (protection.outlook.com: domain of micron.com designates
 137.201.242.130 as permitted sender) receiver=protection.outlook.com;
 client-ip=137.201.242.130; helo=mail.micron.com; pr=C
From:   <sthanneeru.opensrc@micron.com>
To:     <sthanneeru.opensrc@micron.com>, <linux-cxl@vger.kernel.org>,
        <linux-mm@kvack.org>
CC:     <sthanneeru@micron.com>, <aneesh.kumar@linux.ibm.com>,
        <dan.j.williams@intel.com>, <ying.huang@intel.com>,
        <gregory.price@memverge.com>, <mhocko@suse.com>, <tj@kernel.org>,
        <john@jagalactic.com>, <emirakhur@micron.com>,
        <vtavarespetr@micron.com>, <Ravis.OpenSrc@micron.com>,
        <Jonathan.Cameron@huawei.com>, <linux-kernel@vger.kernel.org>,
        Ravi Jonnalagadda <ravis.opensrc@micron.com>
Subject: [PATCH 1/2] base/node: Add sysfs for memtier_override
Date:   Wed, 13 Dec 2023 23:23:28 +0530
Message-ID: <20231213175329.594-2-sthanneeru.opensrc@micron.com>
In-Reply-To: <20231213175329.594-1-sthanneeru.opensrc@micron.com>
References: <20231213175329.594-1-sthanneeru.opensrc@micron.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Type: text/plain
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 13 Dec 2023 17:53:57.2263
 (UTC)
X-MS-Exchange-CrossTenant-Network-Message-Id: 4b99d33b-3297-4022-6904-08dbfc047ad1
X-MS-Exchange-CrossTenant-Id: f38a5ecd-2813-4862-b11b-ac1d563c806f
X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=f38a5ecd-2813-4862-b11b-ac1d563c806f;Ip=[137.201.242.130];Helo=[mail.micron.com]
X-MS-Exchange-CrossTenant-AuthSource: BL6PEPF0001AB54.namprd02.prod.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Anonymous
X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem
X-MS-Exchange-Transport-CrossTenantHeadersStamped: CO6PR08MB7769
X-Spam-Status: No, score=-0.9 required=5.0 tests=DKIM_SIGNED,DKIM_VALID,
	DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS,MAILING_LIST_MULTI,
	SPF_HELO_NONE,SPF_PASS,T_SCC_BODY_TEXT_LINE autolearn=unavailable
	autolearn_force=no version=3.4.6
X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on howler.vger.email
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org
X-Greylist: Sender passed SPF test, not delayed by milter-greylist-4.6.4 (howler.vger.email [0.0.0.0]); Wed, 13 Dec 2023 09:54:21 -0800 (PST)

From: Srinivasulu Thanneeru <sthanneeru.opensrc@micron.com>

This patch introduces a new memtier_override sysfs.

memtier_override is the current memory tier of the node.
To migrate, replace it with the id of the desired memory tier.

adistance_offset is the required offset from memtype to move
the node to the target memory tier(i.e, memtier_override).

Signed-off-by: Srinivasulu Thanneeru <sthanneeru.opensrc@micron.com>
Signed-off-by: Ravi Jonnalagadda <ravis.opensrc@micron.com>
---
 Documentation/ABI/stable/sysfs-devices-node |  7 ++++
 drivers/base/node.c                         | 41 +++++++++++++++++++++
 include/linux/memory-tiers.h                |  6 +++
 include/linux/node.h                        |  6 +++
 mm/memory-tiers.c                           | 19 +++++++++-
 5 files changed, 78 insertions(+), 1 deletion(-)

diff --git a/Documentation/ABI/stable/sysfs-devices-node b/Documentation/ABI/stable/sysfs-devices-node
index 402af4b2b905..447a599cc536 100644
--- a/Documentation/ABI/stable/sysfs-devices-node
+++ b/Documentation/ABI/stable/sysfs-devices-node
@@ -70,6 +70,13 @@ Description:
 		Distance between the node and all the other nodes
 		in the system.
 
+What:		/sys/devices/system/node/nodeX/memtier_overwrite
+Date:		December 2023
+Contact:	Srinivasulu Thanneeru <sthanneeru.opensrc@micron.com>
+Description:
+		The current memory tier of the node.
+		To migrate, replace it with the id of the desired memory tier.
+
 What:		/sys/devices/system/node/nodeX/vmstat
 Date:		October 2002
 Contact:	Linux Memory Management list <linux-mm@kvack.org>
diff --git a/drivers/base/node.c b/drivers/base/node.c
index 493d533f8375..788176b3585a 100644
--- a/drivers/base/node.c
+++ b/drivers/base/node.c
@@ -7,6 +7,7 @@
 #include <linux/init.h>
 #include <linux/mm.h>
 #include <linux/memory.h>
+#include <linux/memory-tiers.h>
 #include <linux/vmstat.h>
 #include <linux/notifier.h>
 #include <linux/node.h>
@@ -569,11 +570,49 @@ static ssize_t node_read_distance(struct device *dev,
 }
 static DEVICE_ATTR(distance, 0444, node_read_distance, NULL);
 
+static ssize_t memtier_override_show(struct device *dev,
+				     struct device_attribute *attr, char *buf)
+{
+	int nid = dev->id;
+	int len = 0;
+
+	len += sysfs_emit(buf, "memory_tier%d\n", node_devices[nid]->memtier);
+	return len;
+}
+
+static ssize_t memtier_override_store(struct device *dev,
+				      struct device_attribute *attr,
+				      const char *buf, size_t size)
+{
+	int nid = dev->id;
+	int ret, memtier;
+
+	ret = kstrtoint(buf, 0, &memtier);
+
+	if (ret)
+		return ret;
+	if (memtier < 0 || memtier > MAX_MEMTIERID)
+		return -EINVAL;
+	if (node_devices[nid]->memtier == memtier)
+		return size;
+	ret = get_memtier_adistance_offset(nid, memtier);
+	node_devices[nid]->adistance_offset = ret;
+
+	return size;
+}
+static DEVICE_ATTR_RW(memtier_override);
+
+void set_node_memtierid(int node, int memtierid)
+{
+	node_devices[node]->memtier = memtierid;
+}
+
 static struct attribute *node_dev_attrs[] = {
 	&dev_attr_meminfo.attr,
 	&dev_attr_numastat.attr,
 	&dev_attr_distance.attr,
 	&dev_attr_vmstat.attr,
+	&dev_attr_memtier_override.attr,
 	NULL
 };
 
@@ -883,6 +922,8 @@ int __register_one_node(int nid)
 
 	INIT_LIST_HEAD(&node_devices[nid]->access_list);
 	node_init_caches(nid);
+	node_devices[nid]->memtier = 0;
+	node_devices[nid]->adistance_offset = 0;
 
 	return error;
 }
diff --git a/include/linux/memory-tiers.h b/include/linux/memory-tiers.h
index 1e39d27bee41..0dba8027e785 100644
--- a/include/linux/memory-tiers.h
+++ b/include/linux/memory-tiers.h
@@ -20,6 +20,11 @@
  */
 #define MEMTIER_ADISTANCE_DRAM	((4 * MEMTIER_CHUNK_SIZE) + (MEMTIER_CHUNK_SIZE >> 1))
 
+/*
+ * Memory tier id is derived from abstract distance(signed 32bits)
+ */
+#define MAX_MEMTIERID (0xFFFFFFFF >> (MEMTIER_CHUNK_BITS + 1))
+
 struct memory_tier;
 struct memory_dev_type {
 	/* list of memory types that are part of same tier as this type */
@@ -48,6 +53,7 @@ int mt_calc_adistance(int node, int *adist);
 int mt_set_default_dram_perf(int nid, struct node_hmem_attrs *perf,
 			     const char *source);
 int mt_perf_to_adistance(struct node_hmem_attrs *perf, int *adist);
+int get_memtier_adistance_offset(int node, int memtier);
 #ifdef CONFIG_MIGRATION
 int next_demotion_node(int node);
 void node_get_allowed_targets(pg_data_t *pgdat, nodemask_t *targets);
diff --git a/include/linux/node.h b/include/linux/node.h
index 427a5975cf40..1c4f4be39db4 100644
--- a/include/linux/node.h
+++ b/include/linux/node.h
@@ -83,6 +83,8 @@ static inline void node_set_perf_attrs(unsigned int nid,
 struct node {
 	struct device	dev;
 	struct list_head access_list;
+	int memtier;
+	int adistance_offset;
 #ifdef CONFIG_HMEM_REPORTING
 	struct list_head cache_attrs;
 	struct device *cache_dev;
@@ -138,6 +140,7 @@ extern void unregister_memory_block_under_nodes(struct memory_block *mem_blk);
 extern int register_memory_node_under_compute_node(unsigned int mem_nid,
 						   unsigned int cpu_nid,
 						   unsigned access);
+extern void set_node_memtierid(int node, int memtierid);
 #else
 static inline void node_dev_init(void)
 {
@@ -165,6 +168,9 @@ static inline int unregister_cpu_under_node(unsigned int cpu, unsigned int nid)
 static inline void unregister_memory_block_under_nodes(struct memory_block *mem_blk)
 {
 }
+static inline void set_node_memtierid(int node, int memtierid)
+{
+}
 #endif
 
 #define to_node(device) container_of(device, struct node, dev)
diff --git a/mm/memory-tiers.c b/mm/memory-tiers.c
index 8d5291add2bc..31ed3c577836 100644
--- a/mm/memory-tiers.c
+++ b/mm/memory-tiers.c
@@ -167,6 +167,21 @@ static const struct attribute_group *memtier_dev_groups[] = {
 	NULL
 };
 
+int get_memtier_adistance_offset(int node, int memtier)
+{
+	struct memory_dev_type *memtype;
+	int adistance_offset;
+
+	memtype = node_memory_types[node].memtype;
+	/*
+	 * Calculate the adistance offset required from memtype
+	 * to move node to target memory tier.
+	 */
+	adistance_offset = (memtier << MEMTIER_CHUNK_BITS) -
+			   memtype->adistance;
+	return adistance_offset;
+}
+
 static struct memory_tier *find_create_memory_tier(struct memory_dev_type *memtype)
 {
 	int ret;
@@ -497,8 +512,10 @@ static struct memory_tier *set_node_memory_tier(int node)
 	memtype = node_memory_types[node].memtype;
 	node_set(node, memtype->nodes);
 	memtier = find_create_memory_tier(memtype);
-	if (!IS_ERR(memtier))
+	if (!IS_ERR(memtier)) {
 		rcu_assign_pointer(pgdat->memtier, memtier);
+		set_node_memtierid(node, memtier->dev.id);
+	}
 	return memtier;
 }
 
-- 
2.25.1