[resend][PATCH 1/2] mm, mem-hotplug: recalculate lowmem_reserve when memory hotplug occur

[resend][PATCH 1/2] mm, mem-hotplug: recalculate lowmem_reserve when memory hotplug occur

[KOSAKI Motohiro]

Currently, memory hotplug call setup_per_zone_wmarks() and
calculate_zone_inactive_ratio(), but don't call setup_per_zone_lowmem_reserve().

It mean number of reserved pages aren't updated even if memory hot plug
occur. This patch fixes it.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
---
 include/linux/mm.h  |    3 +--
 mm/memory_hotplug.c |    9 +++++----
 mm/page_alloc.c     |    6 +++---
 3 files changed, 9 insertions(+), 9 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 721f451..71d1670 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1221,8 +1221,7 @@ extern int __meminit __early_pfn_to_nid(unsigned long pfn);
 extern void set_dma_reserve(unsigned long new_dma_reserve);
 extern void memmap_init_zone(unsigned long, int, unsigned long,
 				unsigned long, enum memmap_context);
-extern void setup_per_zone_wmarks(void);
-extern void calculate_zone_inactive_ratio(struct zone *zone);
+extern int __meminit init_per_zone_wmark_min(void);
 extern void mem_init(void);
 extern void __init mmap_init(void);
 extern void show_mem(void);
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index d8375bb..27d580d 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -437,8 +437,9 @@ int online_pages(unsigned long pfn, unsigned long nr_pages)
 		zone_pcp_update(zone);
 
 	mutex_unlock(&zonelists_mutex);
-	setup_per_zone_wmarks();
-	calculate_zone_inactive_ratio(zone);
+
+	init_per_zone_wmark_min();
+
 	if (onlined_pages) {
 		kswapd_run(zone_to_nid(zone));
 		node_set_state(zone_to_nid(zone), N_HIGH_MEMORY);
@@ -872,8 +873,8 @@ repeat:
 	zone->zone_pgdat->node_present_pages -= offlined_pages;
 	totalram_pages -= offlined_pages;
 
-	setup_per_zone_wmarks();
-	calculate_zone_inactive_ratio(zone);
+	init_per_zone_wmark_min();
+
 	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 b48dea2..14ee899 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -4880,7 +4880,7 @@ static void setup_per_zone_lowmem_reserve(void)
  * Ensures that the watermark[min,low,high] values for each zone are set
  * correctly with respect to min_free_kbytes.
  */
-void setup_per_zone_wmarks(void)
+static void setup_per_zone_wmarks(void)
 {
 	unsigned long pages_min = min_free_kbytes >> (PAGE_SHIFT - 10);
 	unsigned long lowmem_pages = 0;
@@ -4956,7 +4956,7 @@ void setup_per_zone_wmarks(void)
  *    1TB     101        10GB
  *   10TB     320        32GB
  */
-void calculate_zone_inactive_ratio(struct zone *zone)
+static void calculate_zone_inactive_ratio(struct zone *zone)
 {
 	unsigned int gb, ratio;
 
@@ -5002,7 +5002,7 @@ static void __init setup_per_zone_inactive_ratio(void)
  * 8192MB:	11584k
  * 16384MB:	16384k
  */
-static int __init init_per_zone_wmark_min(void)
+int __meminit init_per_zone_wmark_min(void)
 {
 	unsigned long lowmem_kbytes;
 
-- 
1.6.5.2



--
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>

[resend][PATCH 2/2] mm, mem-hotplug: update pcp->stat_threshold when memory hotplug occur

[KOSAKI Motohiro]

Currently, cpu hotplug updates pcp->stat_threashold, but memory
hotplug doesn't. there is no reason.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
---
 include/linux/vmstat.h |    5 ++++-
 mm/page_alloc.c        |    3 +++
 mm/vmstat.c            |    5 ++---
 3 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/include/linux/vmstat.h b/include/linux/vmstat.h
index eaaea37..1997988 100644
--- a/include/linux/vmstat.h
+++ b/include/linux/vmstat.h
@@ -254,6 +254,7 @@ extern void dec_zone_state(struct zone *, enum zone_stat_item);
 extern void __dec_zone_state(struct zone *, enum zone_stat_item);
 
 void refresh_cpu_vm_stats(int);
+void refresh_zone_stat_thresholds(void);
 #else /* CONFIG_SMP */
 
 /*
@@ -299,6 +300,8 @@ static inline void __dec_zone_page_state(struct page *page,
 #define mod_zone_page_state __mod_zone_page_state
 
 static inline void refresh_cpu_vm_stats(int cpu) { }
-#endif
+static inline void refresh_zone_stat_thresholds(void) { }
+
+#endif /* CONFIG_SMP */
 
 #endif /* _LINUX_VMSTAT_H */
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 14ee899..222d8cc 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -51,6 +51,7 @@
 #include <linux/kmemleak.h>
 #include <linux/memory.h>
 #include <linux/compaction.h>
+#include <linux/vmstat.h>
 #include <trace/events/kmem.h>
 #include <linux/ftrace_event.h>
 
@@ -5013,6 +5014,8 @@ int __meminit init_per_zone_wmark_min(void)
 		min_free_kbytes = 128;
 	if (min_free_kbytes > 65536)
 		min_free_kbytes = 65536;
+
+	refresh_zone_stat_thresholds();
 	setup_per_zone_wmarks();
 	setup_per_zone_lowmem_reserve();
 	setup_per_zone_inactive_ratio();
diff --git a/mm/vmstat.c b/mm/vmstat.c
index cd2e42b..b6b7fed 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -133,7 +133,7 @@ static int calculate_threshold(struct zone *zone)
 /*
  * Refresh the thresholds for each zone.
  */
-static void refresh_zone_stat_thresholds(void)
+void refresh_zone_stat_thresholds(void)
 {
 	struct zone *zone;
 	int cpu;
@@ -371,7 +371,7 @@ void refresh_cpu_vm_stats(int cpu)
 			atomic_long_add(global_diff[i], &vm_stat[i]);
 }
 
-#endif
+#endif /* CONFIG_SMP */
 
 #ifdef CONFIG_NUMA
 /*
@@ -1059,7 +1059,6 @@ static int __init setup_vmstat(void)
 #ifdef CONFIG_SMP
 	int cpu;
 
-	refresh_zone_stat_thresholds();
 	register_cpu_notifier(&vmstat_notifier);
 
 	for_each_online_cpu(cpu)
-- 
1.6.5.2



--
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: [resend][PATCH 2/2] mm, mem-hotplug: update pcp->stat_threshold when memory hotplug occur

[Christoph Lameter]

On Tue, 19 Oct 2010, KOSAKI Motohiro wrote:

> Currently, cpu hotplug updates pcp->stat_threashold, but memory
> hotplug doesn't. there is no reason.

Acked-by: Christoph Lameter <cl@linux.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>

Re: [resend][PATCH 2/2] mm, mem-hotplug: update pcp->stat_threshold when memory hotplug occur

[David Rientjes]

On Tue, 19 Oct 2010, KOSAKI Motohiro wrote:

> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 14ee899..222d8cc 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -51,6 +51,7 @@
>  #include <linux/kmemleak.h>
>  #include <linux/memory.h>
>  #include <linux/compaction.h>
> +#include <linux/vmstat.h>
>  #include <trace/events/kmem.h>
>  #include <linux/ftrace_event.h>
>  
> @@ -5013,6 +5014,8 @@ int __meminit init_per_zone_wmark_min(void)
>  		min_free_kbytes = 128;
>  	if (min_free_kbytes > 65536)
>  		min_free_kbytes = 65536;
> +
> +	refresh_zone_stat_thresholds();
>  	setup_per_zone_wmarks();
>  	setup_per_zone_lowmem_reserve();
>  	setup_per_zone_inactive_ratio();

setup_per_zone_wmarks() could change the min and low watermarks for a zone 
when refresh_zone_stat_thresholds() would have used the old value.

--
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: [resend][PATCH 2/2] mm, mem-hotplug: update pcp->stat_threshold when memory hotplug occur

[Andrew Morton]

On Tue, 19 Oct 2010 12:11:09 -0700 (PDT)
David Rientjes <rientjes@google.com> wrote:

> On Tue, 19 Oct 2010, KOSAKI Motohiro wrote:
> 
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index 14ee899..222d8cc 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -51,6 +51,7 @@
> >  #include <linux/kmemleak.h>
> >  #include <linux/memory.h>
> >  #include <linux/compaction.h>
> > +#include <linux/vmstat.h>
> >  #include <trace/events/kmem.h>
> >  #include <linux/ftrace_event.h>
> >  
> > @@ -5013,6 +5014,8 @@ int __meminit init_per_zone_wmark_min(void)
> >  		min_free_kbytes = 128;
> >  	if (min_free_kbytes > 65536)
> >  		min_free_kbytes = 65536;
> > +
> > +	refresh_zone_stat_thresholds();
> >  	setup_per_zone_wmarks();
> >  	setup_per_zone_lowmem_reserve();
> >  	setup_per_zone_inactive_ratio();
> 
> setup_per_zone_wmarks() could change the min and low watermarks for a zone 
> when refresh_zone_stat_thresholds() would have used the old value.

Indeed.

I could make the obvious fix, but then what I'd have wouldn't be
sufficiently tested.

--
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: [resend][PATCH 2/2] mm, mem-hotplug: update pcp->stat_threshold when memory hotplug occur

[KOSAKI Motohiro]

> > > @@ -5013,6 +5014,8 @@ int __meminit init_per_zone_wmark_min(void)
> > >  		min_free_kbytes = 128;
> > >  	if (min_free_kbytes > 65536)
> > >  		min_free_kbytes = 65536;
> > > +
> > > +	refresh_zone_stat_thresholds();
> > >  	setup_per_zone_wmarks();
> > >  	setup_per_zone_lowmem_reserve();
> > >  	setup_per_zone_inactive_ratio();
> > 
> > setup_per_zone_wmarks() could change the min and low watermarks for a zone 
> > when refresh_zone_stat_thresholds() would have used the old value.
> 
> Indeed.
> 
> I could make the obvious fix, but then what I'd have wouldn't be
> sufficiently tested.

Can we review 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: [resend][PATCH 2/2] mm, mem-hotplug: update pcp->stat_threshold when memory hotplug occur

[Andrew Morton]

On Wed, 20 Oct 2010 11:07:33 +0900 (JST) KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> wrote:

> > > > @@ -5013,6 +5014,8 @@ int __meminit init_per_zone_wmark_min(void)
> > > >  		min_free_kbytes = 128;
> > > >  	if (min_free_kbytes > 65536)
> > > >  		min_free_kbytes = 65536;
> > > > +
> > > > +	refresh_zone_stat_thresholds();
> > > >  	setup_per_zone_wmarks();
> > > >  	setup_per_zone_lowmem_reserve();
> > > >  	setup_per_zone_inactive_ratio();
> > > 
> > > setup_per_zone_wmarks() could change the min and low watermarks for a zone 
> > > when refresh_zone_stat_thresholds() would have used the old value.
> > 
> > Indeed.
> > 
> > I could make the obvious fix, but then what I'd have wouldn't be
> > sufficiently tested.
> 
> Can we review this?

It's unclear what you mean?

The patches otherwise look OK to me, but it's pretty easy to make
mistakes in this area.

--
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: [resend][PATCH 2/2] mm, mem-hotplug: update pcp->stat_threshold when memory hotplug occur

[KOSAKI Motohiro]

> On Wed, 20 Oct 2010 11:07:33 +0900 (JST) KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> wrote:
> 
> > > > > @@ -5013,6 +5014,8 @@ int __meminit init_per_zone_wmark_min(void)
> > > > >  		min_free_kbytes = 128;
> > > > >  	if (min_free_kbytes > 65536)
> > > > >  		min_free_kbytes = 65536;
> > > > > +
> > > > > +	refresh_zone_stat_thresholds();
> > > > >  	setup_per_zone_wmarks();
> > > > >  	setup_per_zone_lowmem_reserve();
> > > > >  	setup_per_zone_inactive_ratio();
> > > > 
> > > > setup_per_zone_wmarks() could change the min and low watermarks for a zone 
> > > > when refresh_zone_stat_thresholds() would have used the old value.
> > > 
> > > Indeed.
> > > 
> > > I could make the obvious fix, but then what I'd have wouldn't be
> > > sufficiently tested.
> > 
> > Can we review this?
> 
> It's unclear what you mean?
> 
> The patches otherwise look OK to me, but it's pretty easy to make
> mistakes in this area.

Ah, I misunderstood. I thought you have perfectly different approach patch.
I'll respin.

Thanks.



--
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: [resend][PATCH 2/2] mm, mem-hotplug: update pcp->stat_threshold when memory hotplug occur

[KOSAKI Motohiro]

> On Tue, 19 Oct 2010, KOSAKI Motohiro wrote:
> 
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index 14ee899..222d8cc 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -51,6 +51,7 @@
> >  #include <linux/kmemleak.h>
> >  #include <linux/memory.h>
> >  #include <linux/compaction.h>
> > +#include <linux/vmstat.h>
> >  #include <trace/events/kmem.h>
> >  #include <linux/ftrace_event.h>
> >  
> > @@ -5013,6 +5014,8 @@ int __meminit init_per_zone_wmark_min(void)
> >  		min_free_kbytes = 128;
> >  	if (min_free_kbytes > 65536)
> >  		min_free_kbytes = 65536;
> > +
> > +	refresh_zone_stat_thresholds();
> >  	setup_per_zone_wmarks();
> >  	setup_per_zone_lowmem_reserve();
> >  	setup_per_zone_inactive_ratio();
> 
> setup_per_zone_wmarks() could change the min and low watermarks for a zone 
> when refresh_zone_stat_thresholds() would have used the old value.

Good catch. thanks.

my previous version removed zone->percpu_drift_mark completely. but
current one doesn't. so I need to fix 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>