[PATCH 1/4] mm/hotplug: correctly setup fallback zonelists when creating new pgdat

[PATCH 1/4] mm/hotplug: correctly setup fallback zonelists when creating new pgdat

[Jiang Liu]

When hotadd_new_pgdat() is called to create new pgdat for a new node,
a fallback zonelist should be created for the new node. There's code
to try to achieve that in hotadd_new_pgdat() as below:
	/*
	 * The node we allocated has no zone fallback lists. For avoiding
	 * to access not-initialized zonelist, build here.
	 */
	mutex_lock(&zonelists_mutex);
	build_all_zonelists(pgdat, NULL);
	mutex_unlock(&zonelists_mutex);

But it doesn't work as expected. When hotadd_new_pgdat() is called, the
new node is still in offline state because node_set_online(nid) hasn't
been called yet. And build_all_zonelists() only builds zonelists for
online nodes as:
        for_each_online_node(nid) {
                pg_data_t *pgdat = NODE_DATA(nid);

                build_zonelists(pgdat);
                build_zonelist_cache(pgdat);
        }

Though we hope to create zonelist for the new pgdat, but it doesn't.
So add a new parameter "pgdat" the build_all_zonelists() to build pgdat
for the new pgdat too.

Signed-off-by: Jiang Liu <liuj97@gmail.com>
Signed-off-by: Xishi Qiu <qiuxishi@huawei.com>
---
 include/linux/mmzone.h |    2 +-
 init/main.c            |    2 +-
 kernel/cpu.c           |    2 +-
 mm/memory_hotplug.c    |    4 ++--
 mm/page_alloc.c        |   17 ++++++++++++-----
 5 files changed, 17 insertions(+), 10 deletions(-)

diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 2427706..8ddbfb4 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -718,7 +718,7 @@ typedef struct pglist_data {
 #include <linux/memory_hotplug.h>
 
 extern struct mutex zonelists_mutex;
-void build_all_zonelists(void *data);
+void build_all_zonelists(pg_data_t *pgdat, struct zone *zone);
 void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx);
 bool zone_watermark_ok(struct zone *z, int order, unsigned long mark,
 		int classzone_idx, int alloc_flags);
diff --git a/init/main.c b/init/main.c
index b5cc0a7..622364d 100644
--- a/init/main.c
+++ b/init/main.c
@@ -501,7 +501,7 @@ asmlinkage void __init start_kernel(void)
 	setup_per_cpu_areas();
 	smp_prepare_boot_cpu();	/* arch-specific boot-cpu hooks */
 
-	build_all_zonelists(NULL);
+	build_all_zonelists(NULL, NULL);
 	page_alloc_init();
 
 	printk(KERN_NOTICE "Kernel command line: %s\n", boot_command_line);
diff --git a/kernel/cpu.c b/kernel/cpu.c
index a4eb522..14d3258 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -416,7 +416,7 @@ int __cpuinit cpu_up(unsigned int cpu)
 
 	if (pgdat->node_zonelists->_zonerefs->zone == NULL) {
 		mutex_lock(&zonelists_mutex);
-		build_all_zonelists(NULL);
+		build_all_zonelists(NULL, NULL);
 		mutex_unlock(&zonelists_mutex);
 	}
 #endif
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 0d7e3ec..f93c5b5 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -513,7 +513,7 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages)
 	zone->present_pages += onlined_pages;
 	zone->zone_pgdat->node_present_pages += onlined_pages;
 	if (need_zonelists_rebuild)
-		build_all_zonelists(zone);
+		build_all_zonelists(NULL, zone);
 	else
 		zone_pcp_update(zone);
 
@@ -562,7 +562,7 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
 	 * to access not-initialized zonelist, build here.
 	 */
 	mutex_lock(&zonelists_mutex);
-	build_all_zonelists(NULL);
+	build_all_zonelists(pgdat, NULL);
 	mutex_unlock(&zonelists_mutex);
 
 	return pgdat;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 4403009..ebf319d 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3030,7 +3030,7 @@ int numa_zonelist_order_handler(ctl_table *table, int write,
 			user_zonelist_order = oldval;
 		} else if (oldval != user_zonelist_order) {
 			mutex_lock(&zonelists_mutex);
-			build_all_zonelists(NULL);
+			build_all_zonelists(NULL, NULL);
 			mutex_unlock(&zonelists_mutex);
 		}
 	}
@@ -3413,10 +3413,17 @@ static __init_refok int __build_all_zonelists(void *data)
 {
 	int nid;
 	int cpu;
+	pg_data_t *self = data;
 
 #ifdef CONFIG_NUMA
 	memset(node_load, 0, sizeof(node_load));
 #endif
+
+	if (self && !node_online(self->node_id)) {
+		build_zonelists(self);
+		build_zonelist_cache(self);
+	}
+
 	for_each_online_node(nid) {
 		pg_data_t *pgdat = NODE_DATA(nid);
 
@@ -3461,7 +3468,7 @@ static __init_refok int __build_all_zonelists(void *data)
  * Called with zonelists_mutex held always
  * unless system_state == SYSTEM_BOOTING.
  */
-void __ref build_all_zonelists(void *data)
+void __ref build_all_zonelists(pg_data_t *pgdat, struct zone *zone)
 {
 	set_zonelist_order();
 
@@ -3473,10 +3480,10 @@ void __ref build_all_zonelists(void *data)
 		/* we have to stop all cpus to guarantee there is no user
 		   of zonelist */
 #ifdef CONFIG_MEMORY_HOTPLUG
-		if (data)
-			setup_zone_pageset((struct zone *)data);
+		if (zone)
+			setup_zone_pageset(zone);
 #endif
-		stop_machine(__build_all_zonelists, NULL, NULL);
+		stop_machine(__build_all_zonelists, pgdat, NULL);
 		/* cpuset refresh routine should be here */
 	}
 	vm_total_pages = nr_free_pagecache_pages();
-- 
1.7.1


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 1/4] mm/hotplug: correctly setup fallback zonelists when creating new pgdat

[Jiang Liu]

When hotadd_new_pgdat() is called to create new pgdat for a new node,
a fallback zonelist should be created for the new node. There's code
to try to achieve that in hotadd_new_pgdat() as below:
	/*
	 * The node we allocated has no zone fallback lists. For avoiding
	 * to access not-initialized zonelist, build here.
	 */
	mutex_lock(&zonelists_mutex);
	build_all_zonelists(pgdat, NULL);
	mutex_unlock(&zonelists_mutex);

But it doesn't work as expected. When hotadd_new_pgdat() is called, the
new node is still in offline state because node_set_online(nid) hasn't
been called yet. And build_all_zonelists() only builds zonelists for
online nodes as:
        for_each_online_node(nid) {
                pg_data_t *pgdat = NODE_DATA(nid);

                build_zonelists(pgdat);
                build_zonelist_cache(pgdat);
        }

Though we hope to create zonelist for the new pgdat, but it doesn't.
So add a new parameter "pgdat" the build_all_zonelists() to build pgdat
for the new pgdat too.

Signed-off-by: Jiang Liu <liuj97@gmail.com>
Signed-off-by: Xishi Qiu <qiuxishi@huawei.com>
---
 include/linux/mmzone.h |    2 +-
 init/main.c            |    2 +-
 kernel/cpu.c           |    2 +-
 mm/memory_hotplug.c    |    4 ++--
 mm/page_alloc.c        |   17 ++++++++++++-----
 5 files changed, 17 insertions(+), 10 deletions(-)

diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 2427706..8ddbfb4 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -718,7 +718,7 @@ typedef struct pglist_data {
 #include <linux/memory_hotplug.h>
 
 extern struct mutex zonelists_mutex;
-void build_all_zonelists(void *data);
+void build_all_zonelists(pg_data_t *pgdat, struct zone *zone);
 void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx);
 bool zone_watermark_ok(struct zone *z, int order, unsigned long mark,
 		int classzone_idx, int alloc_flags);
diff --git a/init/main.c b/init/main.c
index b5cc0a7..622364d 100644
--- a/init/main.c
+++ b/init/main.c
@@ -501,7 +501,7 @@ asmlinkage void __init start_kernel(void)
 	setup_per_cpu_areas();
 	smp_prepare_boot_cpu();	/* arch-specific boot-cpu hooks */
 
-	build_all_zonelists(NULL);
+	build_all_zonelists(NULL, NULL);
 	page_alloc_init();
 
 	printk(KERN_NOTICE "Kernel command line: %s\n", boot_command_line);
diff --git a/kernel/cpu.c b/kernel/cpu.c
index a4eb522..14d3258 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -416,7 +416,7 @@ int __cpuinit cpu_up(unsigned int cpu)
 
 	if (pgdat->node_zonelists->_zonerefs->zone == NULL) {
 		mutex_lock(&zonelists_mutex);
-		build_all_zonelists(NULL);
+		build_all_zonelists(NULL, NULL);
 		mutex_unlock(&zonelists_mutex);
 	}
 #endif
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 0d7e3ec..f93c5b5 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -513,7 +513,7 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages)
 	zone->present_pages += onlined_pages;
 	zone->zone_pgdat->node_present_pages += onlined_pages;
 	if (need_zonelists_rebuild)
-		build_all_zonelists(zone);
+		build_all_zonelists(NULL, zone);
 	else
 		zone_pcp_update(zone);
 
@@ -562,7 +562,7 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
 	 * to access not-initialized zonelist, build here.
 	 */
 	mutex_lock(&zonelists_mutex);
-	build_all_zonelists(NULL);
+	build_all_zonelists(pgdat, NULL);
 	mutex_unlock(&zonelists_mutex);
 
 	return pgdat;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 4403009..ebf319d 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3030,7 +3030,7 @@ int numa_zonelist_order_handler(ctl_table *table, int write,
 			user_zonelist_order = oldval;
 		} else if (oldval != user_zonelist_order) {
 			mutex_lock(&zonelists_mutex);
-			build_all_zonelists(NULL);
+			build_all_zonelists(NULL, NULL);
 			mutex_unlock(&zonelists_mutex);
 		}
 	}
@@ -3413,10 +3413,17 @@ static __init_refok int __build_all_zonelists(void *data)
 {
 	int nid;
 	int cpu;
+	pg_data_t *self = data;
 
 #ifdef CONFIG_NUMA
 	memset(node_load, 0, sizeof(node_load));
 #endif
+
+	if (self && !node_online(self->node_id)) {
+		build_zonelists(self);
+		build_zonelist_cache(self);
+	}
+
 	for_each_online_node(nid) {
 		pg_data_t *pgdat = NODE_DATA(nid);
 
@@ -3461,7 +3468,7 @@ static __init_refok int __build_all_zonelists(void *data)
  * Called with zonelists_mutex held always
  * unless system_state == SYSTEM_BOOTING.
  */
-void __ref build_all_zonelists(void *data)
+void __ref build_all_zonelists(pg_data_t *pgdat, struct zone *zone)
 {
 	set_zonelist_order();
 
@@ -3473,10 +3480,10 @@ void __ref build_all_zonelists(void *data)
 		/* we have to stop all cpus to guarantee there is no user
 		   of zonelist */
 #ifdef CONFIG_MEMORY_HOTPLUG
-		if (data)
-			setup_zone_pageset((struct zone *)data);
+		if (zone)
+			setup_zone_pageset(zone);
 #endif
-		stop_machine(__build_all_zonelists, NULL, NULL);
+		stop_machine(__build_all_zonelists, pgdat, NULL);
 		/* cpuset refresh routine should be here */
 	}
 	vm_total_pages = nr_free_pagecache_pages();
-- 
1.7.1

[PATCH 2/4] mm/hotplug: correctly add new zone to all other nodes' zone lists

[Jiang Liu]

When online_pages() is called to add new memory to an empty zone,
it rebuilds all zone lists by calling build_all_zonelists().
But there's a bug which prevents the new zone to be added to other
nodes' zone lists.

online_pages() {
	build_all_zonelists()
	.....
	node_set_state(zone_to_nid(zone), N_HIGH_MEMORY)
}

Here the node of the zone is put into N_HIGH_MEMORY state after calling
build_all_zonelists(), but build_all_zonelists() only adds zones from
nodes in N_HIGH_MEMORY state to the fallback zone lists.
build_all_zonelists()
    ->__build_all_zonelists()
	->build_zonelists()
	    ->find_next_best_node()
		->for_each_node_state(n, N_HIGH_MEMORY)

So memory in the new zone will never be used by other nodes, and it may
cause strange behavor when system is under memory pressure.  So put node
into N_HIGH_MEMORY state before calling build_all_zonelists().

Signed-Off: Jianguo Wu <wujianguo@huawei.com>
Signed-off-by: Jiang Liu <liuj97@gmail.com>
---
 mm/memory_hotplug.c |   15 ++++++++-------
 1 files changed, 8 insertions(+), 7 deletions(-)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index f93c5b5..bce80c7 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -512,19 +512,20 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages)
 
 	zone->present_pages += onlined_pages;
 	zone->zone_pgdat->node_present_pages += onlined_pages;
-	if (need_zonelists_rebuild)
-		build_all_zonelists(NULL, zone);
-	else
-		zone_pcp_update(zone);
+	if (onlined_pages) {
+		node_set_state(zone_to_nid(zone), N_HIGH_MEMORY);
+		if (need_zonelists_rebuild)
+			build_all_zonelists(NULL, zone);
+		else
+			zone_pcp_update(zone);
+	}
 
 	mutex_unlock(&zonelists_mutex);
 
 	init_per_zone_wmark_min();
 
-	if (onlined_pages) {
+	if (onlined_pages)
 		kswapd_run(zone_to_nid(zone));
-		node_set_state(zone_to_nid(zone), N_HIGH_MEMORY);
-	}
 
 	vm_total_pages = nr_free_pagecache_pages();
 
-- 
1.7.1


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 2/4] mm/hotplug: correctly add new zone to all other nodes' zone lists

[Jiang Liu]

When online_pages() is called to add new memory to an empty zone,
it rebuilds all zone lists by calling build_all_zonelists().
But there's a bug which prevents the new zone to be added to other
nodes' zone lists.

online_pages() {
	build_all_zonelists()
	.....
	node_set_state(zone_to_nid(zone), N_HIGH_MEMORY)
}

Here the node of the zone is put into N_HIGH_MEMORY state after calling
build_all_zonelists(), but build_all_zonelists() only adds zones from
nodes in N_HIGH_MEMORY state to the fallback zone lists.
build_all_zonelists()
    ->__build_all_zonelists()
	->build_zonelists()
	    ->find_next_best_node()
		->for_each_node_state(n, N_HIGH_MEMORY)

So memory in the new zone will never be used by other nodes, and it may
cause strange behavor when system is under memory pressure.  So put node
into N_HIGH_MEMORY state before calling build_all_zonelists().

Signed-Off: Jianguo Wu <wujianguo@huawei.com>
Signed-off-by: Jiang Liu <liuj97@gmail.com>
---
 mm/memory_hotplug.c |   15 ++++++++-------
 1 files changed, 8 insertions(+), 7 deletions(-)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index f93c5b5..bce80c7 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -512,19 +512,20 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages)
 
 	zone->present_pages += onlined_pages;
 	zone->zone_pgdat->node_present_pages += onlined_pages;
-	if (need_zonelists_rebuild)
-		build_all_zonelists(NULL, zone);
-	else
-		zone_pcp_update(zone);
+	if (onlined_pages) {
+		node_set_state(zone_to_nid(zone), N_HIGH_MEMORY);
+		if (need_zonelists_rebuild)
+			build_all_zonelists(NULL, zone);
+		else
+			zone_pcp_update(zone);
+	}
 
 	mutex_unlock(&zonelists_mutex);
 
 	init_per_zone_wmark_min();
 
-	if (onlined_pages) {
+	if (onlined_pages)
 		kswapd_run(zone_to_nid(zone));
-		node_set_state(zone_to_nid(zone), N_HIGH_MEMORY);
-	}
 
 	vm_total_pages = nr_free_pagecache_pages();
 
-- 
1.7.1

[PATCH 3/4] mm/hotplug: free zone->pageset when a zone becomes empty

[Jiang Liu]

When a zone becomes empty after memory offlining, free zone->pageset.
Otherwise it will cause memory leak when adding memory to the empty
zone again because build_all_zonelists() will allocate zone->pageset
for an empty zone.

Signed-off-by: Jiang Liu <liuj97@gmail.com>
Signed-off-by: Wei Wang <Bessel.Wang@huawei.com>
---
 include/linux/mm.h  |    1 +
 mm/memory_hotplug.c |    3 +++
 mm/page_alloc.c     |   13 +++++++++++++
 3 files changed, 17 insertions(+), 0 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index b36d08c..f8b62f2 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1331,6 +1331,7 @@ void warn_alloc_failed(gfp_t gfp_mask, int order, const char *fmt, ...);
 extern void setup_per_cpu_pageset(void);
 
 extern void zone_pcp_update(struct zone *zone);
+extern void zone_pcp_reset(struct zone *zone);
 
 /* nommu.c */
 extern atomic_long_t mmap_pages_allocated;
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index bce80c7..998b792 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -966,6 +966,9 @@ repeat:
 
 	init_per_zone_wmark_min();
 
+	if (!populated_zone(zone))
+		zone_pcp_reset(zone);
+
 	if (!node_present_pages(node)) {
 		node_clear_state(node, N_HIGH_MEMORY);
 		kswapd_stop(node);
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index ebf319d..5964b7a 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -5872,6 +5872,19 @@ void free_contig_range(unsigned long pfn, unsigned nr_pages)
 #endif
 
 #ifdef CONFIG_MEMORY_HOTREMOVE
+void zone_pcp_reset(struct zone *zone)
+{
+	unsigned long flags;
+
+	/* avoid races with drain_pages()  */
+	local_irq_save(flags);
+	if (zone->pageset != &boot_pageset) {
+		free_percpu(zone->pageset);
+		zone->pageset = &boot_pageset;
+	}
+	local_irq_restore(flags);
+}
+
 /*
  * All pages in the range must be isolated before calling this.
  */
-- 
1.7.1


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 3/4] mm/hotplug: free zone->pageset when a zone becomes empty

[Jiang Liu]

When a zone becomes empty after memory offlining, free zone->pageset.
Otherwise it will cause memory leak when adding memory to the empty
zone again because build_all_zonelists() will allocate zone->pageset
for an empty zone.

Signed-off-by: Jiang Liu <liuj97@gmail.com>
Signed-off-by: Wei Wang <Bessel.Wang@huawei.com>
---
 include/linux/mm.h  |    1 +
 mm/memory_hotplug.c |    3 +++
 mm/page_alloc.c     |   13 +++++++++++++
 3 files changed, 17 insertions(+), 0 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index b36d08c..f8b62f2 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1331,6 +1331,7 @@ void warn_alloc_failed(gfp_t gfp_mask, int order, const char *fmt, ...);
 extern void setup_per_cpu_pageset(void);
 
 extern void zone_pcp_update(struct zone *zone);
+extern void zone_pcp_reset(struct zone *zone);
 
 /* nommu.c */
 extern atomic_long_t mmap_pages_allocated;
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index bce80c7..998b792 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -966,6 +966,9 @@ repeat:
 
 	init_per_zone_wmark_min();
 
+	if (!populated_zone(zone))
+		zone_pcp_reset(zone);
+
 	if (!node_present_pages(node)) {
 		node_clear_state(node, N_HIGH_MEMORY);
 		kswapd_stop(node);
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index ebf319d..5964b7a 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -5872,6 +5872,19 @@ void free_contig_range(unsigned long pfn, unsigned nr_pages)
 #endif
 
 #ifdef CONFIG_MEMORY_HOTREMOVE
+void zone_pcp_reset(struct zone *zone)
+{
+	unsigned long flags;
+
+	/* avoid races with drain_pages()  */
+	local_irq_save(flags);
+	if (zone->pageset != &boot_pageset) {
+		free_percpu(zone->pageset);
+		zone->pageset = &boot_pageset;
+	}
+	local_irq_restore(flags);
+}
+
 /*
  * All pages in the range must be isolated before calling this.
  */
-- 
1.7.1

Re: [PATCH 3/4] mm/hotplug: free zone->pageset when a zone becomes empty

[Bob Liu]

On Thu, Jul 5, 2012 at 5:45 PM, Jiang Liu <jiang.liu@huawei.com> wrote:
> When a zone becomes empty after memory offlining, free zone->pageset.
> Otherwise it will cause memory leak when adding memory to the empty
> zone again because build_all_zonelists() will allocate zone->pageset
> for an empty zone.
>

What about other area allocated to the zone?  eg. wait_table?

> Signed-off-by: Jiang Liu <liuj97@gmail.com>
> Signed-off-by: Wei Wang <Bessel.Wang@huawei.com>
> ---
>  include/linux/mm.h  |    1 +
>  mm/memory_hotplug.c |    3 +++
>  mm/page_alloc.c     |   13 +++++++++++++
>  3 files changed, 17 insertions(+), 0 deletions(-)
>
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index b36d08c..f8b62f2 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -1331,6 +1331,7 @@ void warn_alloc_failed(gfp_t gfp_mask, int order, const char *fmt, ...);
>  extern void setup_per_cpu_pageset(void);
>
>  extern void zone_pcp_update(struct zone *zone);
> +extern void zone_pcp_reset(struct zone *zone);
>
>  /* nommu.c */
>  extern atomic_long_t mmap_pages_allocated;
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index bce80c7..998b792 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -966,6 +966,9 @@ repeat:
>
>         init_per_zone_wmark_min();
>
> +       if (!populated_zone(zone))
> +               zone_pcp_reset(zone);
> +
>         if (!node_present_pages(node)) {
>                 node_clear_state(node, N_HIGH_MEMORY);
>                 kswapd_stop(node);
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index ebf319d..5964b7a 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -5872,6 +5872,19 @@ void free_contig_range(unsigned long pfn, unsigned nr_pages)
>  #endif
>
>  #ifdef CONFIG_MEMORY_HOTREMOVE
> +void zone_pcp_reset(struct zone *zone)
> +{
> +       unsigned long flags;
> +
> +       /* avoid races with drain_pages()  */
> +       local_irq_save(flags);
> +       if (zone->pageset != &boot_pageset) {
> +               free_percpu(zone->pageset);
> +               zone->pageset = &boot_pageset;
> +       }
> +       local_irq_restore(flags);
> +}
> +
>  /*
>   * All pages in the range must be isolated before calling this.
>   */
> --
> 1.7.1
>
>
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@kvack.org.  For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>



-- 
Regards,
--Bob

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 3/4] mm/hotplug: free zone->pageset when a zone becomes empty

[Bob Liu]

On Thu, Jul 5, 2012 at 5:45 PM, Jiang Liu <jiang.liu@huawei.com> wrote:
> When a zone becomes empty after memory offlining, free zone->pageset.
> Otherwise it will cause memory leak when adding memory to the empty
> zone again because build_all_zonelists() will allocate zone->pageset
> for an empty zone.
>

What about other area allocated to the zone?  eg. wait_table?

> Signed-off-by: Jiang Liu <liuj97@gmail.com>
> Signed-off-by: Wei Wang <Bessel.Wang@huawei.com>
> ---
>  include/linux/mm.h  |    1 +
>  mm/memory_hotplug.c |    3 +++
>  mm/page_alloc.c     |   13 +++++++++++++
>  3 files changed, 17 insertions(+), 0 deletions(-)
>
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index b36d08c..f8b62f2 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -1331,6 +1331,7 @@ void warn_alloc_failed(gfp_t gfp_mask, int order, const char *fmt, ...);
>  extern void setup_per_cpu_pageset(void);
>
>  extern void zone_pcp_update(struct zone *zone);
> +extern void zone_pcp_reset(struct zone *zone);
>
>  /* nommu.c */
>  extern atomic_long_t mmap_pages_allocated;
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index bce80c7..998b792 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -966,6 +966,9 @@ repeat:
>
>         init_per_zone_wmark_min();
>
> +       if (!populated_zone(zone))
> +               zone_pcp_reset(zone);
> +
>         if (!node_present_pages(node)) {
>                 node_clear_state(node, N_HIGH_MEMORY);
>                 kswapd_stop(node);
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index ebf319d..5964b7a 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -5872,6 +5872,19 @@ void free_contig_range(unsigned long pfn, unsigned nr_pages)
>  #endif
>
>  #ifdef CONFIG_MEMORY_HOTREMOVE
> +void zone_pcp_reset(struct zone *zone)
> +{
> +       unsigned long flags;
> +
> +       /* avoid races with drain_pages()  */
> +       local_irq_save(flags);
> +       if (zone->pageset != &boot_pageset) {
> +               free_percpu(zone->pageset);
> +               zone->pageset = &boot_pageset;
> +       }
> +       local_irq_restore(flags);
> +}
> +
>  /*
>   * All pages in the range must be isolated before calling this.
>   */
> --
> 1.7.1
>
>
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@kvack.org.  For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>



-- 
Regards,
--Bob

Re: [PATCH 3/4] mm/hotplug: free zone->pageset when a zone becomes empty

[Jiang Liu]

Hi Bob,
	Yes, we need to handle the wait table too. We has tried to remove the
pgdat and wait table altogether, but found it's really hard to remove pgdat
for empty nodes. I think the candidate solution is to free wait table but
keep pgdat. Any suggestions?
	Thanks!
	Gerry

On 2012-7-19 15:58, Bob Liu wrote:
> On Thu, Jul 5, 2012 at 5:45 PM, Jiang Liu <jiang.liu@huawei.com> wrote:
>> When a zone becomes empty after memory offlining, free zone->pageset.
>> Otherwise it will cause memory leak when adding memory to the empty
>> zone again because build_all_zonelists() will allocate zone->pageset
>> for an empty zone.
>>
> 
> What about other area allocated to the zone?  eg. wait_table?
> 
>> Signed-off-by: Jiang Liu <liuj97@gmail.com>
>> Signed-off-by: Wei Wang <Bessel.Wang@huawei.com>
>> ---
>>  include/linux/mm.h  |    1 +
>>  mm/memory_hotplug.c |    3 +++
>>  mm/page_alloc.c     |   13 +++++++++++++
>>  3 files changed, 17 insertions(+), 0 deletions(-)
>>
>> diff --git a/include/linux/mm.h b/include/linux/mm.h
>> index b36d08c..f8b62f2 100644
>> --- a/include/linux/mm.h
>> +++ b/include/linux/mm.h
>> @@ -1331,6 +1331,7 @@ void warn_alloc_failed(gfp_t gfp_mask, int order, const char *fmt, ...);
>>  extern void setup_per_cpu_pageset(void);
>>
>>  extern void zone_pcp_update(struct zone *zone);
>> +extern void zone_pcp_reset(struct zone *zone);
>>
>>  /* nommu.c */
>>  extern atomic_long_t mmap_pages_allocated;
>> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
>> index bce80c7..998b792 100644
>> --- a/mm/memory_hotplug.c
>> +++ b/mm/memory_hotplug.c
>> @@ -966,6 +966,9 @@ repeat:
>>
>>         init_per_zone_wmark_min();
>>
>> +       if (!populated_zone(zone))
>> +               zone_pcp_reset(zone);
>> +
>>         if (!node_present_pages(node)) {
>>                 node_clear_state(node, N_HIGH_MEMORY);
>>                 kswapd_stop(node);
>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>> index ebf319d..5964b7a 100644
>> --- a/mm/page_alloc.c
>> +++ b/mm/page_alloc.c
>> @@ -5872,6 +5872,19 @@ void free_contig_range(unsigned long pfn, unsigned nr_pages)
>>  #endif
>>
>>  #ifdef CONFIG_MEMORY_HOTREMOVE
>> +void zone_pcp_reset(struct zone *zone)
>> +{
>> +       unsigned long flags;
>> +
>> +       /* avoid races with drain_pages()  */
>> +       local_irq_save(flags);
>> +       if (zone->pageset != &boot_pageset) {
>> +               free_percpu(zone->pageset);
>> +               zone->pageset = &boot_pageset;
>> +       }
>> +       local_irq_restore(flags);
>> +}
>> +
>>  /*
>>   * All pages in the range must be isolated before calling this.
>>   */
>> --
>> 1.7.1
>>
>>
>> --
>> To unsubscribe, send a message with 'unsubscribe linux-mm' in
>> the body to majordomo@kvack.org.  For more info on Linux MM,
>> see: http://www.linux-mm.org/ .
>> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
> 
> 
> 


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 3/4] mm/hotplug: free zone->pageset when a zone becomes empty

[Jiang Liu]

Hi Bob,
	Yes, we need to handle the wait table too. We has tried to remove the
pgdat and wait table altogether, but found it's really hard to remove pgdat
for empty nodes. I think the candidate solution is to free wait table but
keep pgdat. Any suggestions?
	Thanks!
	Gerry

On 2012-7-19 15:58, Bob Liu wrote:
> On Thu, Jul 5, 2012 at 5:45 PM, Jiang Liu <jiang.liu@huawei.com> wrote:
>> When a zone becomes empty after memory offlining, free zone->pageset.
>> Otherwise it will cause memory leak when adding memory to the empty
>> zone again because build_all_zonelists() will allocate zone->pageset
>> for an empty zone.
>>
> 
> What about other area allocated to the zone?  eg. wait_table?
> 
>> Signed-off-by: Jiang Liu <liuj97@gmail.com>
>> Signed-off-by: Wei Wang <Bessel.Wang@huawei.com>
>> ---
>>  include/linux/mm.h  |    1 +
>>  mm/memory_hotplug.c |    3 +++
>>  mm/page_alloc.c     |   13 +++++++++++++
>>  3 files changed, 17 insertions(+), 0 deletions(-)
>>
>> diff --git a/include/linux/mm.h b/include/linux/mm.h
>> index b36d08c..f8b62f2 100644
>> --- a/include/linux/mm.h
>> +++ b/include/linux/mm.h
>> @@ -1331,6 +1331,7 @@ void warn_alloc_failed(gfp_t gfp_mask, int order, const char *fmt, ...);
>>  extern void setup_per_cpu_pageset(void);
>>
>>  extern void zone_pcp_update(struct zone *zone);
>> +extern void zone_pcp_reset(struct zone *zone);
>>
>>  /* nommu.c */
>>  extern atomic_long_t mmap_pages_allocated;
>> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
>> index bce80c7..998b792 100644
>> --- a/mm/memory_hotplug.c
>> +++ b/mm/memory_hotplug.c
>> @@ -966,6 +966,9 @@ repeat:
>>
>>         init_per_zone_wmark_min();
>>
>> +       if (!populated_zone(zone))
>> +               zone_pcp_reset(zone);
>> +
>>         if (!node_present_pages(node)) {
>>                 node_clear_state(node, N_HIGH_MEMORY);
>>                 kswapd_stop(node);
>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>> index ebf319d..5964b7a 100644
>> --- a/mm/page_alloc.c
>> +++ b/mm/page_alloc.c
>> @@ -5872,6 +5872,19 @@ void free_contig_range(unsigned long pfn, unsigned nr_pages)
>>  #endif
>>
>>  #ifdef CONFIG_MEMORY_HOTREMOVE
>> +void zone_pcp_reset(struct zone *zone)
>> +{
>> +       unsigned long flags;
>> +
>> +       /* avoid races with drain_pages()  */
>> +       local_irq_save(flags);
>> +       if (zone->pageset != &boot_pageset) {
>> +               free_percpu(zone->pageset);
>> +               zone->pageset = &boot_pageset;
>> +       }
>> +       local_irq_restore(flags);
>> +}
>> +
>>  /*
>>   * All pages in the range must be isolated before calling this.
>>   */
>> --
>> 1.7.1
>>
>>
>> --
>> To unsubscribe, send a message with 'unsubscribe linux-mm' in
>> the body to majordomo@kvack.org.  For more info on Linux MM,
>> see: http://www.linux-mm.org/ .
>> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
> 
> 
> 

Re: [PATCH 3/4] mm/hotplug: free zone->pageset when a zone becomes empty

[Wen Congyang]

At 07/05/2012 05:45 PM, Jiang Liu Wrote:
> When a zone becomes empty after memory offlining, free zone->pageset.
> Otherwise it will cause memory leak when adding memory to the empty
> zone again because build_all_zonelists() will allocate zone->pageset
> for an empty zone.
> 
> Signed-off-by: Jiang Liu <liuj97@gmail.com>
> Signed-off-by: Wei Wang <Bessel.Wang@huawei.com>
> ---
>  include/linux/mm.h  |    1 +
>  mm/memory_hotplug.c |    3 +++
>  mm/page_alloc.c     |   13 +++++++++++++
>  3 files changed, 17 insertions(+), 0 deletions(-)
> 
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index b36d08c..f8b62f2 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -1331,6 +1331,7 @@ void warn_alloc_failed(gfp_t gfp_mask, int order, const char *fmt, ...);
>  extern void setup_per_cpu_pageset(void);
>  
>  extern void zone_pcp_update(struct zone *zone);
> +extern void zone_pcp_reset(struct zone *zone);
>  
>  /* nommu.c */
>  extern atomic_long_t mmap_pages_allocated;
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index bce80c7..998b792 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -966,6 +966,9 @@ repeat:
>  
>  	init_per_zone_wmark_min();
>  
> +	if (!populated_zone(zone))
> +		zone_pcp_reset(zone);
> +
>  	if (!node_present_pages(node)) {
>  		node_clear_state(node, N_HIGH_MEMORY);
>  		kswapd_stop(node);
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index ebf319d..5964b7a 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -5872,6 +5872,19 @@ void free_contig_range(unsigned long pfn, unsigned nr_pages)
>  #endif
>  
>  #ifdef CONFIG_MEMORY_HOTREMOVE
> +void zone_pcp_reset(struct zone *zone)
> +{
> +	unsigned long flags;
> +
> +	/* avoid races with drain_pages()  */
> +	local_irq_save(flags);

drain_pages() may run on another cpu, so it cann't avoid
races with drain_pages().

Thanks
Wen Congyang

> +	if (zone->pageset != &boot_pageset) {
> +		free_percpu(zone->pageset);
> +		zone->pageset = &boot_pageset;
> +	}
> +	local_irq_restore(flags);
> +}
> +
>  /*
>   * All pages in the range must be isolated before calling this.
>   */

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 3/4] mm/hotplug: free zone->pageset when a zone becomes empty

[Wen Congyang]

At 07/05/2012 05:45 PM, Jiang Liu Wrote:
> When a zone becomes empty after memory offlining, free zone->pageset.
> Otherwise it will cause memory leak when adding memory to the empty
> zone again because build_all_zonelists() will allocate zone->pageset
> for an empty zone.
> 
> Signed-off-by: Jiang Liu <liuj97@gmail.com>
> Signed-off-by: Wei Wang <Bessel.Wang@huawei.com>
> ---
>  include/linux/mm.h  |    1 +
>  mm/memory_hotplug.c |    3 +++
>  mm/page_alloc.c     |   13 +++++++++++++
>  3 files changed, 17 insertions(+), 0 deletions(-)
> 
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index b36d08c..f8b62f2 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -1331,6 +1331,7 @@ void warn_alloc_failed(gfp_t gfp_mask, int order, const char *fmt, ...);
>  extern void setup_per_cpu_pageset(void);
>  
>  extern void zone_pcp_update(struct zone *zone);
> +extern void zone_pcp_reset(struct zone *zone);
>  
>  /* nommu.c */
>  extern atomic_long_t mmap_pages_allocated;
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index bce80c7..998b792 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -966,6 +966,9 @@ repeat:
>  
>  	init_per_zone_wmark_min();
>  
> +	if (!populated_zone(zone))
> +		zone_pcp_reset(zone);
> +
>  	if (!node_present_pages(node)) {
>  		node_clear_state(node, N_HIGH_MEMORY);
>  		kswapd_stop(node);
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index ebf319d..5964b7a 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -5872,6 +5872,19 @@ void free_contig_range(unsigned long pfn, unsigned nr_pages)
>  #endif
>  
>  #ifdef CONFIG_MEMORY_HOTREMOVE
> +void zone_pcp_reset(struct zone *zone)
> +{
> +	unsigned long flags;
> +
> +	/* avoid races with drain_pages()  */
> +	local_irq_save(flags);

drain_pages() may run on another cpu, so it cann't avoid
races with drain_pages().

Thanks
Wen Congyang

> +	if (zone->pageset != &boot_pageset) {
> +		free_percpu(zone->pageset);
> +		zone->pageset = &boot_pageset;
> +	}
> +	local_irq_restore(flags);
> +}
> +
>  /*
>   * All pages in the range must be isolated before calling this.
>   */

[PATCH 4/4] mm/hotplug: mark memory hotplug code in page_alloc.c as __meminit

[Jiang Liu]

Mark functions used by both boot and memory hotplug as __meminit to reduce
memory footprint when memory hotplug is disabled.

Alos guard zone_pcp_update() with CONFIG_MEMORY_HOTPLUG because it's only
used by memory hotplug code.

Signed-off-by: Jiang Liu <liuj97@gmail.com>
---
 mm/page_alloc.c |   66 ++++++++++++++++++++++++++++--------------------------
 1 files changed, 34 insertions(+), 32 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 5964b7a..da08449 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3409,7 +3409,7 @@ static void setup_zone_pageset(struct zone *zone);
 DEFINE_MUTEX(zonelists_mutex);
 
 /* return values int ....just for stop_machine() */
-static __init_refok int __build_all_zonelists(void *data)
+static int __build_all_zonelists(void *data)
 {
 	int nid;
 	int cpu;
@@ -3753,7 +3753,7 @@ static void __meminit zone_init_free_lists(struct zone *zone)
 	memmap_init_zone((size), (nid), (zone), (start_pfn), MEMMAP_EARLY)
 #endif
 
-static int zone_batchsize(struct zone *zone)
+static int __meminit zone_batchsize(struct zone *zone)
 {
 #ifdef CONFIG_MMU
 	int batch;
@@ -3835,7 +3835,7 @@ static void setup_pagelist_highmark(struct per_cpu_pageset *p,
 		pcp->batch = PAGE_SHIFT * 8;
 }
 
-static void setup_zone_pageset(struct zone *zone)
+static void __meminit setup_zone_pageset(struct zone *zone)
 {
 	int cpu;
 
@@ -3908,33 +3908,7 @@ int zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages)
 	return 0;
 }
 
-static int __zone_pcp_update(void *data)
-{
-	struct zone *zone = data;
-	int cpu;
-	unsigned long batch = zone_batchsize(zone), flags;
-
-	for_each_possible_cpu(cpu) {
-		struct per_cpu_pageset *pset;
-		struct per_cpu_pages *pcp;
-
-		pset = per_cpu_ptr(zone->pageset, cpu);
-		pcp = &pset->pcp;
-
-		local_irq_save(flags);
-		free_pcppages_bulk(zone, pcp->count, pcp);
-		setup_pageset(pset, batch);
-		local_irq_restore(flags);
-	}
-	return 0;
-}
-
-void zone_pcp_update(struct zone *zone)
-{
-	stop_machine(__zone_pcp_update, zone, NULL);
-}
-
-static __meminit void zone_pcp_init(struct zone *zone)
+static void __meminit zone_pcp_init(struct zone *zone)
 {
 	/*
 	 * per cpu subsystem is not up at this point. The following code
@@ -3949,7 +3923,7 @@ static __meminit void zone_pcp_init(struct zone *zone)
 					 zone_batchsize(zone));
 }
 
-__meminit int init_currently_empty_zone(struct zone *zone,
+int __meminit init_currently_empty_zone(struct zone *zone,
 					unsigned long zone_start_pfn,
 					unsigned long size,
 					enum memmap_context context)
@@ -4757,7 +4731,7 @@ out:
 }
 
 /* Any regular memory on that node ? */
-static void check_for_regular_memory(pg_data_t *pgdat)
+static void __init check_for_regular_memory(pg_data_t *pgdat)
 {
 #ifdef CONFIG_HIGHMEM
 	enum zone_type zone_type;
@@ -5871,6 +5845,34 @@ void free_contig_range(unsigned long pfn, unsigned nr_pages)
 }
 #endif
 
+#ifdef	CONFIG_MEMORY_HOTPLUG
+static int __meminit __zone_pcp_update(void *data)
+{
+	struct zone *zone = data;
+	int cpu;
+	unsigned long batch = zone_batchsize(zone), flags;
+
+	for_each_possible_cpu(cpu) {
+		struct per_cpu_pageset *pset;
+		struct per_cpu_pages *pcp;
+
+		pset = per_cpu_ptr(zone->pageset, cpu);
+		pcp = &pset->pcp;
+
+		local_irq_save(flags);
+		free_pcppages_bulk(zone, pcp->count, pcp);
+		setup_pageset(pset, batch);
+		local_irq_restore(flags);
+	}
+	return 0;
+}
+
+void __meminit zone_pcp_update(struct zone *zone)
+{
+	stop_machine(__zone_pcp_update, zone, NULL);
+}
+#endif
+
 #ifdef CONFIG_MEMORY_HOTREMOVE
 void zone_pcp_reset(struct zone *zone)
 {
-- 
1.7.1


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 4/4] mm/hotplug: mark memory hotplug code in page_alloc.c as __meminit

[Jiang Liu]

Mark functions used by both boot and memory hotplug as __meminit to reduce
memory footprint when memory hotplug is disabled.

Alos guard zone_pcp_update() with CONFIG_MEMORY_HOTPLUG because it's only
used by memory hotplug code.

Signed-off-by: Jiang Liu <liuj97@gmail.com>
---
 mm/page_alloc.c |   66 ++++++++++++++++++++++++++++--------------------------
 1 files changed, 34 insertions(+), 32 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 5964b7a..da08449 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -3409,7 +3409,7 @@ static void setup_zone_pageset(struct zone *zone);
 DEFINE_MUTEX(zonelists_mutex);
 
 /* return values int ....just for stop_machine() */
-static __init_refok int __build_all_zonelists(void *data)
+static int __build_all_zonelists(void *data)
 {
 	int nid;
 	int cpu;
@@ -3753,7 +3753,7 @@ static void __meminit zone_init_free_lists(struct zone *zone)
 	memmap_init_zone((size), (nid), (zone), (start_pfn), MEMMAP_EARLY)
 #endif
 
-static int zone_batchsize(struct zone *zone)
+static int __meminit zone_batchsize(struct zone *zone)
 {
 #ifdef CONFIG_MMU
 	int batch;
@@ -3835,7 +3835,7 @@ static void setup_pagelist_highmark(struct per_cpu_pageset *p,
 		pcp->batch = PAGE_SHIFT * 8;
 }
 
-static void setup_zone_pageset(struct zone *zone)
+static void __meminit setup_zone_pageset(struct zone *zone)
 {
 	int cpu;
 
@@ -3908,33 +3908,7 @@ int zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages)
 	return 0;
 }
 
-static int __zone_pcp_update(void *data)
-{
-	struct zone *zone = data;
-	int cpu;
-	unsigned long batch = zone_batchsize(zone), flags;
-
-	for_each_possible_cpu(cpu) {
-		struct per_cpu_pageset *pset;
-		struct per_cpu_pages *pcp;
-
-		pset = per_cpu_ptr(zone->pageset, cpu);
-		pcp = &pset->pcp;
-
-		local_irq_save(flags);
-		free_pcppages_bulk(zone, pcp->count, pcp);
-		setup_pageset(pset, batch);
-		local_irq_restore(flags);
-	}
-	return 0;
-}
-
-void zone_pcp_update(struct zone *zone)
-{
-	stop_machine(__zone_pcp_update, zone, NULL);
-}
-
-static __meminit void zone_pcp_init(struct zone *zone)
+static void __meminit zone_pcp_init(struct zone *zone)
 {
 	/*
 	 * per cpu subsystem is not up at this point. The following code
@@ -3949,7 +3923,7 @@ static __meminit void zone_pcp_init(struct zone *zone)
 					 zone_batchsize(zone));
 }
 
-__meminit int init_currently_empty_zone(struct zone *zone,
+int __meminit init_currently_empty_zone(struct zone *zone,
 					unsigned long zone_start_pfn,
 					unsigned long size,
 					enum memmap_context context)
@@ -4757,7 +4731,7 @@ out:
 }
 
 /* Any regular memory on that node ? */
-static void check_for_regular_memory(pg_data_t *pgdat)
+static void __init check_for_regular_memory(pg_data_t *pgdat)
 {
 #ifdef CONFIG_HIGHMEM
 	enum zone_type zone_type;
@@ -5871,6 +5845,34 @@ void free_contig_range(unsigned long pfn, unsigned nr_pages)
 }
 #endif
 
+#ifdef	CONFIG_MEMORY_HOTPLUG
+static int __meminit __zone_pcp_update(void *data)
+{
+	struct zone *zone = data;
+	int cpu;
+	unsigned long batch = zone_batchsize(zone), flags;
+
+	for_each_possible_cpu(cpu) {
+		struct per_cpu_pageset *pset;
+		struct per_cpu_pages *pcp;
+
+		pset = per_cpu_ptr(zone->pageset, cpu);
+		pcp = &pset->pcp;
+
+		local_irq_save(flags);
+		free_pcppages_bulk(zone, pcp->count, pcp);
+		setup_pageset(pset, batch);
+		local_irq_restore(flags);
+	}
+	return 0;
+}
+
+void __meminit zone_pcp_update(struct zone *zone)
+{
+	stop_machine(__zone_pcp_update, zone, NULL);
+}
+#endif
+
 #ifdef CONFIG_MEMORY_HOTREMOVE
 void zone_pcp_reset(struct zone *zone)
 {
-- 
1.7.1

Re: [PATCH 4/4] mm/hotplug: mark memory hotplug code in page_alloc.c as __meminit

[Cong Wang]

On Thu, 05 Jul 2012 at 09:45 GMT, Jiang Liu <jiang.liu@huawei.com> wrote:
> Mark functions used by both boot and memory hotplug as __meminit to reduce
> memory footprint when memory hotplug is disabled.
>
> Alos guard zone_pcp_update() with CONFIG_MEMORY_HOTPLUG because it's only
> used by memory hotplug code.
>

If so, why not move it to mm/memory_hotplug.c and make it static?

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 1/4] mm/hotplug: correctly setup fallback zonelists when creating new pgdat

[Cong Wang]

On Thu, 05 Jul 2012 at 09:45 GMT, Jiang Liu <jiang.liu@huawei.com> wrote:
> +
> +	if (self && !node_online(self->node_id)) {
> +		build_zonelists(self);
> +		build_zonelist_cache(self);
> +	}
> +

You don't need to test !node_online() here, as you are sure it is not
online yet.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>