2013-05-01 23:32:14

by Cody P Schafer

[permalink] [raw]
Subject: [PATCH v2 0/4] misc patches related to resizing nodes & zones

First 2 are comment fixes.
Second 2 add pgdat_resize_lock()/unlock() usage per existing documentation.

--

Since v1 (http://thread.gmane.org/gmane.linux.kernel.mm/99297):
- drop making lock_memory_hotplug() required (old patch #1)
- fix __offline_pages() in the same manner as online_pages() (rientjes)
- make comment regarding pgdat_resize_lock()/unlock() usage more clear (rientjes)

--

Cody P Schafer (4):
mm: fix comment referring to non-existent size_seqlock, change to
span_seqlock
mmzone: note that node_size_lock should be manipulated via
pgdat_resize_lock()
memory_hotplug: use pgdat_resize_lock() in online_pages()
memory_hotplug: use pgdat_resize_lock() in __offline_pages()

include/linux/mmzone.h | 5 ++++-
mm/memory_hotplug.c | 9 +++++++++
2 files changed, 13 insertions(+), 1 deletion(-)

--
1.8.2.2


2013-05-01 23:32:33

by Cody P Schafer

[permalink] [raw]
Subject: [PATCH v2 1/4] mm: fix comment referring to non-existent size_seqlock, change to span_seqlock

Signed-off-by: Cody P Schafer <[email protected]>
---
include/linux/mmzone.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 5c76737..fc859a0c 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -716,7 +716,7 @@ typedef struct pglist_data {
* or node_spanned_pages stay constant. Holding this will also
* guarantee that any pfn_valid() stays that way.
*
- * Nests above zone->lock and zone->size_seqlock.
+ * Nests above zone->lock and zone->span_seqlock
*/
spinlock_t node_size_lock;
#endif
--
1.8.2.2

2013-05-01 23:32:43

by Cody P Schafer

[permalink] [raw]
Subject: [PATCH v2 4/4] memory_hotplug: use pgdat_resize_lock() in __offline_pages()

mmzone.h documents node_size_lock (which pgdat_resize_lock() locks) as
guarding against changes to node_present_pages, so actually lock it when
we update node_present_pages to keep that promise.

Signed-off-by: Cody P Schafer <[email protected]>
---
mm/memory_hotplug.c | 4 ++++
1 file changed, 4 insertions(+)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 0bdca10..b59a695 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -1582,7 +1582,11 @@ repeat:
/* removal success */
zone->managed_pages -= offlined_pages;
zone->present_pages -= offlined_pages;
+
+ pgdat_resize_lock(zone->zone_pgdat, &flags);
zone->zone_pgdat->node_present_pages -= offlined_pages;
+ pgdat_resize_unlock(zone->zone_pgdat, &flags);
+
totalram_pages -= offlined_pages;

init_per_zone_wmark_min();
--
1.8.2.2

2013-05-01 23:32:38

by Cody P Schafer

[permalink] [raw]
Subject: [PATCH v2 3/4] memory_hotplug: use pgdat_resize_lock() in online_pages()

mmzone.h documents node_size_lock (which pgdat_resize_lock() locks) as
guarding against changes to node_present_pages, so actually lock it when
we update node_present_pages to keep that promise.

Signed-off-by: Cody P Schafer <[email protected]>
---
mm/memory_hotplug.c | 5 +++++
1 file changed, 5 insertions(+)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index a221fac..0bdca10 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -915,6 +915,7 @@ static void node_states_set_node(int node, struct memory_notify *arg)

int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_type)
{
+ unsigned long flags;
unsigned long onlined_pages = 0;
struct zone *zone;
int need_zonelists_rebuild = 0;
@@ -993,7 +994,11 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_typ

zone->managed_pages += onlined_pages;
zone->present_pages += onlined_pages;
+
+ pgdat_resize_lock(zone->zone_pgdat, &flags);
zone->zone_pgdat->node_present_pages += onlined_pages;
+ pgdat_resize_unlock(zone->zone_pgdat, &flags);
+
if (onlined_pages) {
node_states_set_node(zone_to_nid(zone), &arg);
if (need_zonelists_rebuild)
--
1.8.2.2

2013-05-01 23:32:30

by Cody P Schafer

[permalink] [raw]
Subject: [PATCH v2 2/4] mmzone: note that node_size_lock should be manipulated via pgdat_resize_lock()

Signed-off-by: Cody P Schafer <[email protected]>
---
include/linux/mmzone.h | 3 +++
1 file changed, 3 insertions(+)

diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index fc859a0c..41557be 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -716,6 +716,9 @@ typedef struct pglist_data {
* or node_spanned_pages stay constant. Holding this will also
* guarantee that any pfn_valid() stays that way.
*
+ * pgdat_resize_lock() and pgdat_resize_unlock() are provided to
+ * manipulate node_size_lock without checking for CONFIG_MEMORY_HOTPLUG.
+ *
* Nests above zone->lock and zone->span_seqlock
*/
spinlock_t node_size_lock;
--
1.8.2.2

2013-05-03 18:58:59

by David Rientjes

[permalink] [raw]
Subject: Re: [PATCH v2 1/4] mm: fix comment referring to non-existent size_seqlock, change to span_seqlock

On Wed, 1 May 2013, Cody P Schafer wrote:

> Signed-off-by: Cody P Schafer <[email protected]>

Acked-by: David Rientjes <[email protected]>

Others in this series aren't needed, in my opinion, but everybody has
their own tastes.