Re: linux-next: Tree for Oct 5 (warnings: a. trace; b. mm/migrate)

[Dave Hansen]

On 10/6/21 9:39 AM, Randy Dunlap wrote:
>>>> ../mm/migrate.c:3216:22: error: 'migrate_on_reclaim_callback' defined
>>>> but not used [-Werror=unused-function]
>>>>   static int __meminit migrate_on_reclaim_callback(struct
>>>> notifier_block
>>>> *self,
>>>>                        ^~~~~~~~~~~~~~~~~~~~~~~~~~~
>>>> ../mm/migrate.c:3197:13: error: 'set_migration_target_nodes' defined
>>>> but
>>>> not used [-Werror=unused-function]
>>>>   static void set_migration_target_nodes(void)
>>>>               ^~~~~~~~~~~~~~~~~~~~~~~~~~
>>>>
>>>>
>>>> (example usage to get the randconfig files:
>>>> KCONFIG_SEED=0xBFBEA13C make [ARCH=x86_64] randconfig
>>>> )

Randy, thanks for the .config!  That did the trick.

Andrew, attached is a replacement patch for

	mm-migrate-add-cpu-hotplug-to-demotion-ifdef.patch

I could do an incremental as well if that would be easier.  But, the fix
here required a bit of a change of tactics from the original and
necessitated a rewrite of the changelog.
Subject: [PATCH] mm/migrate: separate CPU and memory hotplug notifiers
Date: Wed, 6 Oct 2021 11:08:46 -0700
Message-ID: <20211006180846.3321352-1-dave.hansen@xxxxxxxxx>
X-Mailer: git-send-email 2.25.1
In-Reply-To: <20210924161255.E5FE8F7E@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
References: <20210924161255.E5FE8F7E@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx>

From: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>

Once upon a time, the node demotion updates were driven solely by memory
hotplug events.  But now, there are handlers for both CPU and memory
hotplug.

However, the #ifdef around the code checks only memory hotplug.  A system
that has HOTPLUG_CPU=y but MEMORY_HOTPLUG=n would miss CPU hotplug events.

Update the #ifdef around the common code.  Add memory and CPU-specific
option are off.  Move some CPU-hotplug-specific functions to reside under
their specific #ifdef.

Link: https://lkml.kernel.org/r/20210924161255.E5FE8F7E@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Fixes: 884a6e5d1f93 ("mm/migrate: update node demotion order on hotplug events")
Signed-off-by: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
Cc: "Huang, Ying" <ying.huang@xxxxxxxxx>
Cc: Michal Hocko <mhocko@xxxxxxxx>
Cc: Wei Xu <weixugc@xxxxxxxxxx>
Cc: Oscar Salvador <osalvador@xxxxxxx>
Cc: David Rientjes <rientjes@xxxxxxxxxx>
Cc: Dan Williams <dan.j.williams@xxxxxxxxx>
Cc: David Hildenbrand <david@xxxxxxxxxx>
Cc: Greg Thelen <gthelen@xxxxxxxxxx>
Cc: Yang Shi <yang.shi@xxxxxxxxxxxxxxxxx>
Cc: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
---
 mm/migrate.c | 78 +++++++++++++++++++++++++++++-----------------------
 1 file changed, 44 insertions(+), 34 deletions(-)

diff --git a/mm/migrate.c b/mm/migrate.c
index a6311e46f605..5282157575ba 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -3066,7 +3066,7 @@ void migrate_vma_finalize(struct migrate_vma *migrate)
 EXPORT_SYMBOL(migrate_vma_finalize);
 #endif /* CONFIG_DEVICE_PRIVATE */
 
-#if defined(CONFIG_MEMORY_HOTPLUG)
+#if defined(CONFIG_MEMORY_HOTPLUG) || defined(CONFIG_HOTPLUG_CPU)
 /* Disable reclaim-based migration. */
 static void __disable_all_migrate_targets(void)
 {
@@ -3198,35 +3198,7 @@ static void __set_migration_target_nodes(void)
 		goto again;
 }
 
-/*
- * For callers that do not hold get_online_mems() already.
- */
-static void set_migration_target_nodes(void)
-{
-	get_online_mems();
-	__set_migration_target_nodes();
-	put_online_mems();
-}
-
-/*
- * React to hotplug events that might affect the migration targets
- * like events that online or offline NUMA nodes.
- *
- * The ordering is also currently dependent on which nodes have
- * CPUs.  That means we need CPU on/offline notification too.
- */
-static int migration_online_cpu(unsigned int cpu)
-{
-	set_migration_target_nodes();
-	return 0;
-}
-
-static int migration_offline_cpu(unsigned int cpu)
-{
-	set_migration_target_nodes();
-	return 0;
-}
-
+#if defined(CONFIG_MEMORY_HOTPLUG)
 /*
  * This leaves migrate-on-reclaim transiently disabled between
  * the MEM_GOING_OFFLINE and MEM_OFFLINE events.  This runs
@@ -3284,7 +3256,45 @@ static int __meminit migrate_on_reclaim_callback(struct notifier_block *self,
 	return notifier_from_errno(0);
 }
 
-static int __init migrate_on_reclaim_init(void)
+static int __init migrate_on_reclaim_init_memhp(void)
+{
+	hotplug_memory_notifier(migrate_on_reclaim_callback, 100);
+	return 0;
+}
+late_initcall(migrate_on_reclaim_init_memhp);
+#endif /* CONFIG_MEMORY_HOTPLUG */
+
+#ifdef CONFIG_HOTPLUG_CPU
+/*
+ * For callers that do not hold get_online_mems() already.
+ */
+static void set_migration_target_nodes(void)
+{
+	get_online_mems();
+	__set_migration_target_nodes();
+	put_online_mems();
+}
+
+/*
+ * React to hotplug events that might affect the migration targets
+ * like events that online or offline NUMA nodes.
+ *
+ * The ordering is also currently dependent on which nodes have
+ * CPUs.  That means we need CPU on/offline notification too.
+ */
+static int migration_online_cpu(unsigned int cpu)
+{
+	set_migration_target_nodes();
+	return 0;
+}
+
+static int migration_offline_cpu(unsigned int cpu)
+{
+	set_migration_target_nodes();
+	return 0;
+}
+
+static int __init migrate_on_reclaim_init_cpuhp(void)
 {
 	int ret;
 
@@ -3299,8 +3309,8 @@ static int __init migrate_on_reclaim_init(void)
 	 */
 	WARN_ON(ret < 0);
 
-	hotplug_memory_notifier(migrate_on_reclaim_callback, 100);
 	return 0;
 }
-late_initcall(migrate_on_reclaim_init);
-#endif /* CONFIG_MEMORY_HOTPLUG */
+late_initcall(migrate_on_reclaim_init_cpuhp);
+#endif /* CONFIG_HOTPLUG_CPU */
+#endif /* CONFIG_MEMORY_HOTPLUG || CONFIG_HOTPLUG_CPU */
-- 
2.25.1