[PATCH v3 0/2] Fix node meminfo and zoneinfo corruption.

[PATCH v3 0/2] Fix node meminfo and zoneinfo corruption.

[Tang Chen]

There are two problems in memory hot-add progress:

1. When hot-adding a node without onlining any page, node meminfo corrupted:

# hot-add node2 (memory not onlined)
# cat /sys/device/system/node/node2/meminfo
Node 2 MemTotal:       33554432 kB			/* corrupted */
Node 2 MemFree:               0 kB
Node 2 MemUsed:        33554432 kB
Node 2 Active:                0 kB
......


2. When onlining memory on node2, node2 zoneinfo and node3 meminfo corrupted:

# for ((i = 2048; i < 2064; i++)); do echo online_movable > /sys/devices/system/node/node2/memory$i/state; done
# cat /sys/devices/system/node/node2/meminfo
Node 2 MemTotal:       33554432 kB
Node 2 MemFree:        33549092 kB
Node 2 MemUsed:            5340 kB
......
# cat /sys/devices/system/node/node3/meminfo
Node 3 MemTotal:              0 kB
Node 3 MemFree:               248 kB      /* corrupted, should be 0 */
Node 3 MemUsed:               0 kB
......

# cat /proc/zoneinfo
......
Node 2, zone   Movable
......
        spanned  8388608
        present  16777216		/* corrupted, should be 8388608 */
        managed  8388608


Change log v2 -> v3:
1. Also clear pgdat->node_present_pages in patch 2/2.
2. Add Cc: stable@vger.kernel.org # 3.16+ since it is urgent.

Change log v1 -> v2:
1. Replace patch 2/2 with a new one. It provides the simplest way to
   fix problem 2. 

Tang Chen (2):
  mem-hotplug: Reset node managed pages when hot-adding a new pgdat.
  mem-hotplug: Reset node present pages when hot-adding a new pgdat.

 include/linux/bootmem.h |  1 +
 mm/bootmem.c            |  9 +++++----
 mm/memory_hotplug.c     | 26 ++++++++++++++++++++++++++
 mm/nobootmem.c          |  8 +++++---
 4 files changed, 37 insertions(+), 7 deletions(-)

-- 
1.8.3.1

--
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 v3 1/2] mem-hotplug: Reset node managed pages when hot-adding a new pgdat.

[Tang Chen]

In free_area_init_core(), zone->managed_pages is set to an approximate
value for lowmem, and will be adjusted when the bootmem allocator frees
pages into the buddy system. But free_area_init_core() is also called
by hotadd_new_pgdat() when hot-adding memory. As a result, zone->managed_pages
of the newly added node's pgdat is set to an approximate value in the
very beginning. Even if the memory on that node has node been onlined,
/sys/device/system/node/nodeXXX/meminfo has wrong value.

hot-add node2 (memory not onlined)
cat /sys/device/system/node/node2/meminfo
Node 2 MemTotal:       33554432 kB
Node 2 MemFree:               0 kB
Node 2 MemUsed:        33554432 kB
Node 2 Active:                0 kB

This patch fixes this problem by reset node managed pages to 0 after hot-adding
a new node.

1. Move reset_managed_pages_done from reset_node_managed_pages() to reset_all_zones_managed_pages()
2. Make reset_node_managed_pages() non-static
3. Call reset_node_managed_pages() in hotadd_new_pgdat() after pgdat is initialized

Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: stable@vger.kernel.org # 3.16+
---
 include/linux/bootmem.h | 1 +
 mm/bootmem.c            | 9 +++++----
 mm/memory_hotplug.c     | 9 +++++++++
 mm/nobootmem.c          | 8 +++++---
 4 files changed, 20 insertions(+), 7 deletions(-)

diff --git a/include/linux/bootmem.h b/include/linux/bootmem.h
index 4e2bd4c..0995c2d 100644
--- a/include/linux/bootmem.h
+++ b/include/linux/bootmem.h
@@ -46,6 +46,7 @@ extern unsigned long init_bootmem_node(pg_data_t *pgdat,
 extern unsigned long init_bootmem(unsigned long addr, unsigned long memend);
 
 extern unsigned long free_all_bootmem(void);
+extern void reset_node_managed_pages(pg_data_t *pgdat);
 extern void reset_all_zones_managed_pages(void);
 
 extern void free_bootmem_node(pg_data_t *pgdat,
diff --git a/mm/bootmem.c b/mm/bootmem.c
index 8a000ce..477be69 100644
--- a/mm/bootmem.c
+++ b/mm/bootmem.c
@@ -243,13 +243,10 @@ static unsigned long __init free_all_bootmem_core(bootmem_data_t *bdata)
 
 static int reset_managed_pages_done __initdata;
 
-static inline void __init reset_node_managed_pages(pg_data_t *pgdat)
+void reset_node_managed_pages(pg_data_t *pgdat)
 {
 	struct zone *z;
 
-	if (reset_managed_pages_done)
-		return;
-
 	for (z = pgdat->node_zones; z < pgdat->node_zones + MAX_NR_ZONES; z++)
 		z->managed_pages = 0;
 }
@@ -258,8 +255,12 @@ void __init reset_all_zones_managed_pages(void)
 {
 	struct pglist_data *pgdat;
 
+	if (reset_managed_pages_done)
+		return;
+
 	for_each_online_pgdat(pgdat)
 		reset_node_managed_pages(pgdat);
+
 	reset_managed_pages_done = 1;
 }
 
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 29d8693..8aba12b 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -31,6 +31,7 @@
 #include <linux/stop_machine.h>
 #include <linux/hugetlb.h>
 #include <linux/memblock.h>
+#include <linux/bootmem.h>
 
 #include <asm/tlbflush.h>
 
@@ -1096,6 +1097,14 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
 	build_all_zonelists(pgdat, NULL);
 	mutex_unlock(&zonelists_mutex);
 
+	/*
+	 * zone->managed_pages is set to an approximate value in
+	 * free_area_init_core(), which will cause
+	 * /sys/device/system/node/nodeX/meminfo has wrong data.
+	 * So reset it to 0 before any memory is onlined.
+	 */
+	reset_node_managed_pages(pgdat);
+
 	return pgdat;
 }
 
diff --git a/mm/nobootmem.c b/mm/nobootmem.c
index 7c7ab32..90b5046 100644
--- a/mm/nobootmem.c
+++ b/mm/nobootmem.c
@@ -145,12 +145,10 @@ static unsigned long __init free_low_memory_core_early(void)
 
 static int reset_managed_pages_done __initdata;
 
-static inline void __init reset_node_managed_pages(pg_data_t *pgdat)
+void reset_node_managed_pages(pg_data_t *pgdat)
 {
 	struct zone *z;
 
-	if (reset_managed_pages_done)
-		return;
 	for (z = pgdat->node_zones; z < pgdat->node_zones + MAX_NR_ZONES; z++)
 		z->managed_pages = 0;
 }
@@ -159,8 +157,12 @@ void __init reset_all_zones_managed_pages(void)
 {
 	struct pglist_data *pgdat;
 
+	if (reset_managed_pages_done)
+		return;
+
 	for_each_online_pgdat(pgdat)
 		reset_node_managed_pages(pgdat);
+
 	reset_managed_pages_done = 1;
 }
 
-- 
1.8.3.1

--
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 v3 2/2] mem-hotplug: Reset node present pages when hot-adding a new pgdat.

[Tang Chen]

When memory is hot-added, all the memory is in offline state. So
clear all zones' present_pages because they will be updated in
online_pages() and offline_pages(). Otherwise, /proc/zoneinfo
will corrupt:

When the memory of node2 is offline:
# cat /proc/zoneinfo
......
Node 2, zone   Movable
......
        spanned  8388608
        present  8388608
        managed  0

When we online memory on node2:
# cat /proc/zoneinfo
......
Node 2, zone   Movable
......
        spanned  8388608
        present  16777216
        managed  8388608

Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: stable@vger.kernel.org # 3.16+
---
 mm/memory_hotplug.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 8aba12b..d0c9b7c 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -1067,6 +1067,16 @@ out:
 }
 #endif /* CONFIG_MEMORY_HOTPLUG_SPARSE */
 
+static void reset_node_present_pages(pg_data_t *pgdat)
+{
+	struct zone *z;
+
+	for (z = pgdat->node_zones; z < pgdat->node_zones + MAX_NR_ZONES; z++)
+		z->present_pages = 0;
+
+	pgdat->node_present_pages = 0;
+}
+
 /* we are OK calling __meminit stuff here - we have CONFIG_MEMORY_HOTPLUG */
 static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
 {
@@ -1105,6 +1115,13 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
 	 */
 	reset_node_managed_pages(pgdat);
 
+	/*
+	 * When memory is hot-added, all the memory is in offline state. So
+	 * clear all zones' present_pages because they will be updated in
+	 * online_pages() and offline_pages().
+	 */
+	reset_node_present_pages(pgdat);
+
 	return pgdat;
 }
 
-- 
1.8.3.1

--
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 v3 1/2] mem-hotplug: Reset node managed pages when hot-adding a new pgdat.

[Kamezawa Hiroyuki]

(2014/11/12 17:37), Tang Chen wrote:
> In free_area_init_core(), zone->managed_pages is set to an approximate
> value for lowmem, and will be adjusted when the bootmem allocator frees
> pages into the buddy system. But free_area_init_core() is also called
> by hotadd_new_pgdat() when hot-adding memory. As a result, zone->managed_pages
> of the newly added node's pgdat is set to an approximate value in the
> very beginning. Even if the memory on that node has node been onlined,
> /sys/device/system/node/nodeXXX/meminfo has wrong value.
> 
> hot-add node2 (memory not onlined)
> cat /sys/device/system/node/node2/meminfo
> Node 2 MemTotal:       33554432 kB
> Node 2 MemFree:               0 kB
> Node 2 MemUsed:        33554432 kB
> Node 2 Active:                0 kB
> 
> This patch fixes this problem by reset node managed pages to 0 after hot-adding
> a new node.
> 
> 1. Move reset_managed_pages_done from reset_node_managed_pages() to reset_all_zones_managed_pages()
> 2. Make reset_node_managed_pages() non-static
> 3. Call reset_node_managed_pages() in hotadd_new_pgdat() after pgdat is initialized
> 
> Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
> Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
> Cc: stable@vger.kernel.org # 3.16+

(Dropped too many CCs because my mailer tell me some error.)

Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
but I have some comments.

> ---
>   include/linux/bootmem.h | 1 +
>   mm/bootmem.c            | 9 +++++----
>   mm/memory_hotplug.c     | 9 +++++++++
>   mm/nobootmem.c          | 8 +++++---
>   4 files changed, 20 insertions(+), 7 deletions(-)
> 
> diff --git a/include/linux/bootmem.h b/include/linux/bootmem.h
> index 4e2bd4c..0995c2d 100644
> --- a/include/linux/bootmem.h
> +++ b/include/linux/bootmem.h
> @@ -46,6 +46,7 @@ extern unsigned long init_bootmem_node(pg_data_t *pgdat,
>   extern unsigned long init_bootmem(unsigned long addr, unsigned long memend);
>   
>   extern unsigned long free_all_bootmem(void);
> +extern void reset_node_managed_pages(pg_data_t *pgdat);
>   extern void reset_all_zones_managed_pages(void);
>   
>   extern void free_bootmem_node(pg_data_t *pgdat,
> diff --git a/mm/bootmem.c b/mm/bootmem.c
> index 8a000ce..477be69 100644
> --- a/mm/bootmem.c
> +++ b/mm/bootmem.c
> @@ -243,13 +243,10 @@ static unsigned long __init free_all_bootmem_core(bootmem_data_t *bdata)
>   
>   static int reset_managed_pages_done __initdata;
>   
> -static inline void __init reset_node_managed_pages(pg_data_t *pgdat)
> +void reset_node_managed_pages(pg_data_t *pgdat)
>   {
>   	struct zone *z;
>   
> -	if (reset_managed_pages_done)
> -		return;
> -
>   	for (z = pgdat->node_zones; z < pgdat->node_zones + MAX_NR_ZONES; z++)
>   		z->managed_pages = 0;
>   }
> @@ -258,8 +255,12 @@ void __init reset_all_zones_managed_pages(void)
>   {
>   	struct pglist_data *pgdat;
>   
> +	if (reset_managed_pages_done)
> +		return;
> +
>   	for_each_online_pgdat(pgdat)
>   		reset_node_managed_pages(pgdat);
> +

unnecessary change here.

>   	reset_managed_pages_done = 1;
>   }
>   
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index 29d8693..8aba12b 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -31,6 +31,7 @@
>   #include <linux/stop_machine.h>
>   #include <linux/hugetlb.h>
>   #include <linux/memblock.h>
> +#include <linux/bootmem.h>
>   
>   #include <asm/tlbflush.h>
>   
> @@ -1096,6 +1097,14 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
>   	build_all_zonelists(pgdat, NULL);
>   	mutex_unlock(&zonelists_mutex);
>   
> +	/*
> +	 * zone->managed_pages is set to an approximate value in
> +	 * free_area_init_core(), which will cause
> +	 * /sys/device/system/node/nodeX/meminfo has wrong data.
           because online_page() count it again.
> +	 * So reset it to 0 before any memory is onlined.
> +	 */
> +	reset_node_managed_pages(pgdat);
> +
>   	return pgdat;
>   }
>   
> diff --git a/mm/nobootmem.c b/mm/nobootmem.c
> index 7c7ab32..90b5046 100644
> --- a/mm/nobootmem.c
> +++ b/mm/nobootmem.c
> @@ -145,12 +145,10 @@ static unsigned long __init free_low_memory_core_early(void)
>   
>   static int reset_managed_pages_done __initdata;
>   
> -static inline void __init reset_node_managed_pages(pg_data_t *pgdat)
> +void reset_node_managed_pages(pg_data_t *pgdat)
>   {
>   	struct zone *z;
>   
> -	if (reset_managed_pages_done)
> -		return;
>   	for (z = pgdat->node_zones; z < pgdat->node_zones + MAX_NR_ZONES; z++)
>   		z->managed_pages = 0;
>   }
> @@ -159,8 +157,12 @@ void __init reset_all_zones_managed_pages(void)
>   {
>   	struct pglist_data *pgdat;
>   
> +	if (reset_managed_pages_done)
> +		return;
> +
>   	for_each_online_pgdat(pgdat)
>   		reset_node_managed_pages(pgdat);
> +
ditto.

Regards,
-Kame
>   	reset_managed_pages_done = 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 v3 2/2] mem-hotplug: Reset node present pages when hot-adding a new pgdat.

[Kamezawa Hiroyuki]

(2014/11/12 17:37), Tang Chen wrote:
> When memory is hot-added, all the memory is in offline state. So
> clear all zones' present_pages because they will be updated in
> online_pages() and offline_pages(). Otherwise, /proc/zoneinfo
> will corrupt:
> 
> When the memory of node2 is offline:
> # cat /proc/zoneinfo
> ......
> Node 2, zone   Movable
> ......
>          spanned  8388608
>          present  8388608
>          managed  0
> 
> When we online memory on node2:
> # cat /proc/zoneinfo
> ......
> Node 2, zone   Movable
> ......
>          spanned  8388608
>          present  16777216
>          managed  8388608
> 
> Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
> Reviewed-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
> Cc: stable@vger.kernel.org # 3.16+

Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>



--
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>