[PATCH 03/25] mm/memcg: Convert to hotplug state machine

[PATCH 03/25] mm/memcg: Convert to hotplug state machine

[Sebastian Andrzej Siewior]

Install the callbacks via the state machine.

Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: cgroups@vger.kernel.org
Cc: linux-mm@kvack.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/cpuhotplug.h |  1 +
 mm/memcontrol.c            | 24 ++++++++----------------
 2 files changed, 9 insertions(+), 16 deletions(-)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 4174083280d7..c622ab349af3 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -32,6 +32,7 @@ enum cpuhp_state {
 	CPUHP_BLK_MQ_DEAD,
 	CPUHP_FS_BUFF_DEAD,
 	CPUHP_PRINTK_DEAD,
+	CPUHP_MM_MEMCQ_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 0f870ba43942..6c2043509fb5 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -1816,22 +1816,13 @@ static void drain_all_stock(struct mem_cgroup *root_memcg)
 	mutex_unlock(&percpu_charge_mutex);
 }
 
-static int memcg_cpu_hotplug_callback(struct notifier_block *nb,
-					unsigned long action,
-					void *hcpu)
+static int memcg_hotplug_cpu_dead(unsigned int cpu)
 {
-	int cpu = (unsigned long)hcpu;
 	struct memcg_stock_pcp *stock;
 
-	if (action == CPU_ONLINE)
-		return NOTIFY_OK;
-
-	if (action != CPU_DEAD && action != CPU_DEAD_FROZEN)
-		return NOTIFY_OK;
-
 	stock = &per_cpu(memcg_stock, cpu);
 	drain_stock(stock);
-	return NOTIFY_OK;
+	return 0;
 }
 
 static void reclaim_high(struct mem_cgroup *memcg,
@@ -5774,16 +5765,17 @@ __setup("cgroup.memory=", cgroup_memory);
 /*
  * subsys_initcall() for memory controller.
  *
- * Some parts like hotcpu_notifier() have to be initialized from this context
- * because of lock dependencies (cgroup_lock -> cpu hotplug) but basically
- * everything that doesn't depend on a specific mem_cgroup structure should
- * be initialized from here.
+ * Some parts like memcg_hotplug_cpu_dead() have to be initialized from this
+ * context because of lock dependencies (cgroup_lock -> cpu hotplug) but
+ * basically everything that doesn't depend on a specific mem_cgroup structure
+ * should be initialized from here.
  */
 static int __init mem_cgroup_init(void)
 {
 	int cpu, node;
 
-	hotcpu_notifier(memcg_cpu_hotplug_callback, 0);
+	cpuhp_setup_state_nocalls(CPUHP_MM_MEMCQ_DEAD, "mm/memctrl:dead", NULL,
+				  memcg_hotplug_cpu_dead);
 
 	for_each_possible_cpu(cpu)
 		INIT_WORK(&per_cpu_ptr(&memcg_stock, cpu)->work,
-- 
2.10.2

[PATCH 06/25] mm/page_alloc: Convert to hotplug state machine

[Sebastian Andrzej Siewior]

Install the callbacks via the state machine.

Cc: linux-mm@kvack.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/cpuhotplug.h |  1 +
 mm/page_alloc.c            | 49 +++++++++++++++++++++++-----------------------
 2 files changed, 26 insertions(+), 24 deletions(-)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 89310fb1031d..31c58f6ec3c6 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -35,6 +35,7 @@ enum cpuhp_state {
 	CPUHP_MM_MEMCQ_DEAD,
 	CPUHP_PERCPU_CNT_DEAD,
 	CPUHP_RADIX_DEAD,
+	CPUHP_PAGE_ALLOC_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 8fd42aa7c4bd..68873a164cc0 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6491,38 +6491,39 @@ void __init free_area_init(unsigned long *zones_size)
 			__pa(PAGE_OFFSET) >> PAGE_SHIFT, NULL);
 }
 
-static int page_alloc_cpu_notify(struct notifier_block *self,
-				 unsigned long action, void *hcpu)
+static int page_alloc_cpu_dead(unsigned int cpu)
 {
-	int cpu = (unsigned long)hcpu;
 
-	if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) {
-		lru_add_drain_cpu(cpu);
-		drain_pages(cpu);
+	lru_add_drain_cpu(cpu);
+	drain_pages(cpu);
 
-		/*
-		 * Spill the event counters of the dead processor
-		 * into the current processors event counters.
-		 * This artificially elevates the count of the current
-		 * processor.
-		 */
-		vm_events_fold_cpu(cpu);
+	/*
+	 * Spill the event counters of the dead processor
+	 * into the current processors event counters.
+	 * This artificially elevates the count of the current
+	 * processor.
+	 */
+	vm_events_fold_cpu(cpu);
 
-		/*
-		 * Zero the differential counters of the dead processor
-		 * so that the vm statistics are consistent.
-		 *
-		 * This is only okay since the processor is dead and cannot
-		 * race with what we are doing.
-		 */
-		cpu_vm_stats_fold(cpu);
-	}
-	return NOTIFY_OK;
+	/*
+	 * Zero the differential counters of the dead processor
+	 * so that the vm statistics are consistent.
+	 *
+	 * This is only okay since the processor is dead and cannot
+	 * race with what we are doing.
+	 */
+	cpu_vm_stats_fold(cpu);
+	return 0;
 }
 
 void __init page_alloc_init(void)
 {
-	hotcpu_notifier(page_alloc_cpu_notify, 0);
+	int ret;
+
+	ret = cpuhp_setup_state_nocalls(CPUHP_PAGE_ALLOC_DEAD,
+					"mm/page_alloc:dead", NULL,
+					page_alloc_cpu_dead);
+	WARN_ON(ret < 0);
 }
 
 /*
-- 
2.10.2

[PATCH 07/25] mm/vmscan: Convert to hotplug state machine

[Sebastian Andrzej Siewior]

Install the callbacks via the state machine.

Cc: linux-mm@kvack.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 mm/vmscan.c | 28 ++++++++++++++--------------
 1 file changed, 14 insertions(+), 14 deletions(-)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 76fda2268148..b8404d32caf0 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3556,24 +3556,21 @@ unsigned long shrink_all_memory(unsigned long nr_to_reclaim)
    not required for correctness.  So if the last cpu in a node goes
    away, we get changed to run anywhere: as the first one comes back,
    restore their cpu bindings. */
-static int cpu_callback(struct notifier_block *nfb, unsigned long action,
-			void *hcpu)
+static int kswapd_cpu_online(unsigned int cpu)
 {
 	int nid;
 
-	if (action == CPU_ONLINE || action == CPU_ONLINE_FROZEN) {
-		for_each_node_state(nid, N_MEMORY) {
-			pg_data_t *pgdat = NODE_DATA(nid);
-			const struct cpumask *mask;
+	for_each_node_state(nid, N_MEMORY) {
+		pg_data_t *pgdat = NODE_DATA(nid);
+		const struct cpumask *mask;
 
-			mask = cpumask_of_node(pgdat->node_id);
+		mask = cpumask_of_node(pgdat->node_id);
 
-			if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids)
-				/* One of our CPUs online: restore mask */
-				set_cpus_allowed_ptr(pgdat->kswapd, mask);
-		}
+		if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids)
+			/* One of our CPUs online: restore mask */
+			set_cpus_allowed_ptr(pgdat->kswapd, mask);
 	}
-	return NOTIFY_OK;
+	return 0;
 }
 
 /*
@@ -3615,12 +3612,15 @@ void kswapd_stop(int nid)
 
 static int __init kswapd_init(void)
 {
-	int nid;
+	int nid, ret;
 
 	swap_setup();
 	for_each_node_state(nid, N_MEMORY)
  		kswapd_run(nid);
-	hotcpu_notifier(cpu_callback, 0);
+	ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN,
+					"mm/vmscan:online", kswapd_cpu_online,
+					NULL);
+	WARN_ON(ret < 0);
 	return 0;
 }
 
-- 
2.10.2