[PATCH] cpuset: update top cpuset's mems after adding a node

[PATCH] cpuset: update top cpuset's mems after adding a node

[Miao Xie]

After adding a node into the machine, top cpuset's mems isn't updated.

By reviewing the code, we found that the update function
  cpuset_track_online_nodes()
was invoked after node_states[N_ONLINE] changes. It is wrong because N_ONLINE
just means node has pgdat, and if node has/added memory, we use N_HIGH_MEMORY.
So, We should invoke the update function after node_states[N_HIGH_MEMORY]
changes, just like its commit says.

This patch fixes it. And we use notifier of memory hotplug instead of direct
calling of cpuset_track_online_nodes().

Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>

---
 include/linux/cpuset.h |    4 ----
 kernel/cpuset.c        |   19 ++++++++++++++++---
 mm/memory_hotplug.c    |    3 ---
 3 files changed, 16 insertions(+), 10 deletions(-)

diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h
index 2691926..8e540d3 100644
--- a/include/linux/cpuset.h
+++ b/include/linux/cpuset.h
@@ -74,8 +74,6 @@ static inline int cpuset_do_slab_mem_spread(void)
 	return current->flags & PF_SPREAD_SLAB;
 }
 
-extern void cpuset_track_online_nodes(void);
-
 extern int current_cpuset_is_being_rebound(void);
 
 extern void rebuild_sched_domains(void);
@@ -151,8 +149,6 @@ static inline int cpuset_do_slab_mem_spread(void)
 	return 0;
 }
 
-static inline void cpuset_track_online_nodes(void) {}
-
 static inline int current_cpuset_is_being_rebound(void)
 {
 	return 0;
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 3e00526..909a548 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -36,6 +36,7 @@
 #include <linux/list.h>
 #include <linux/mempolicy.h>
 #include <linux/mm.h>
+#include <linux/memory.h>
 #include <linux/module.h>
 #include <linux/mount.h>
 #include <linux/namei.h>
@@ -2011,12 +2012,23 @@ static int cpuset_track_online_cpus(struct notifier_block *unused_nb,
  * Call this routine anytime after node_states[N_HIGH_MEMORY] changes.
  * See also the previous routine cpuset_track_online_cpus().
  */
-void cpuset_track_online_nodes(void)
+static int cpuset_track_online_nodes(struct notifier_block *self,
+				unsigned long action, void *arg)
 {
 	cgroup_lock();
-	top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
-	scan_for_empty_cpusets(&top_cpuset);
+	switch (action) {
+	case MEM_ONLINE:
+		top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
+		break;
+	case MEM_OFFLINE:
+		top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
+		scan_for_empty_cpusets(&top_cpuset);
+		break;
+	default:
+		break;
+	}
 	cgroup_unlock();
+	return NOTIFY_OK;
 }
 #endif
 
@@ -2032,6 +2044,7 @@ void __init cpuset_init_smp(void)
 	top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
 
 	hotcpu_notifier(cpuset_track_online_cpus, 0);
+	hotplug_memory_notifier(cpuset_track_online_nodes, 10);
 }
 
 /**
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 6837a10..b5b2b15 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -22,7 +22,6 @@
 #include <linux/highmem.h>
 #include <linux/vmalloc.h>
 #include <linux/ioport.h>
-#include <linux/cpuset.h>
 #include <linux/delay.h>
 #include <linux/migrate.h>
 #include <linux/page-isolation.h>
@@ -498,8 +497,6 @@ int add_memory(int nid, u64 start, u64 size)
 	/* we online node here. we can't roll back from here. */
 	node_set_online(nid);
 
-	cpuset_track_online_nodes();
-
 	if (new_pgdat) {
 		ret = register_one_node(nid);
 		/*
-- 
1.5.3.8

Re: [PATCH] cpuset: update top cpuset's mems after adding a node

[Yasunori Goto]

Did you test this patch?  Was it OK?

If yes,
Acked-by: Yasunori Goto <y-goto@jp.fujitsu.com>


Thanks.


> After adding a node into the machine, top cpuset's mems isn't updated.
> 
> By reviewing the code, we found that the update function
>   cpuset_track_online_nodes()
> was invoked after node_states[N_ONLINE] changes. It is wrong because N_ONLINE
> just means node has pgdat, and if node has/added memory, we use N_HIGH_MEMORY.
> So, We should invoke the update function after node_states[N_HIGH_MEMORY]
> changes, just like its commit says.
> 
> This patch fixes it. And we use notifier of memory hotplug instead of direct
> calling of cpuset_track_online_nodes().
> 
> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
> Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
> 
> ---
>  include/linux/cpuset.h |    4 ----
>  kernel/cpuset.c        |   19 ++++++++++++++++---
>  mm/memory_hotplug.c    |    3 ---
>  3 files changed, 16 insertions(+), 10 deletions(-)
> 
> diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h
> index 2691926..8e540d3 100644
> --- a/include/linux/cpuset.h
> +++ b/include/linux/cpuset.h
> @@ -74,8 +74,6 @@ static inline int cpuset_do_slab_mem_spread(void)
>  	return current->flags & PF_SPREAD_SLAB;
>  }
>  
> -extern void cpuset_track_online_nodes(void);
> -
>  extern int current_cpuset_is_being_rebound(void);
>  
>  extern void rebuild_sched_domains(void);
> @@ -151,8 +149,6 @@ static inline int cpuset_do_slab_mem_spread(void)
>  	return 0;
>  }
>  
> -static inline void cpuset_track_online_nodes(void) {}
> -
>  static inline int current_cpuset_is_being_rebound(void)
>  {
>  	return 0;
> diff --git a/kernel/cpuset.c b/kernel/cpuset.c
> index 3e00526..909a548 100644
> --- a/kernel/cpuset.c
> +++ b/kernel/cpuset.c
> @@ -36,6 +36,7 @@
>  #include <linux/list.h>
>  #include <linux/mempolicy.h>
>  #include <linux/mm.h>
> +#include <linux/memory.h>
>  #include <linux/module.h>
>  #include <linux/mount.h>
>  #include <linux/namei.h>
> @@ -2011,12 +2012,23 @@ static int cpuset_track_online_cpus(struct notifier_block *unused_nb,
>   * Call this routine anytime after node_states[N_HIGH_MEMORY] changes.
>   * See also the previous routine cpuset_track_online_cpus().
>   */
> -void cpuset_track_online_nodes(void)
> +static int cpuset_track_online_nodes(struct notifier_block *self,
> +				unsigned long action, void *arg)
>  {
>  	cgroup_lock();
> -	top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
> -	scan_for_empty_cpusets(&top_cpuset);
> +	switch (action) {
> +	case MEM_ONLINE:
> +		top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
> +		break;
> +	case MEM_OFFLINE:
> +		top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
> +		scan_for_empty_cpusets(&top_cpuset);
> +		break;
> +	default:
> +		break;
> +	}
>  	cgroup_unlock();
> +	return NOTIFY_OK;
>  }
>  #endif
>  
> @@ -2032,6 +2044,7 @@ void __init cpuset_init_smp(void)
>  	top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
>  
>  	hotcpu_notifier(cpuset_track_online_cpus, 0);
> +	hotplug_memory_notifier(cpuset_track_online_nodes, 10);
>  }
>  
>  /**
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index 6837a10..b5b2b15 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -22,7 +22,6 @@
>  #include <linux/highmem.h>
>  #include <linux/vmalloc.h>
>  #include <linux/ioport.h>
> -#include <linux/cpuset.h>
>  #include <linux/delay.h>
>  #include <linux/migrate.h>
>  #include <linux/page-isolation.h>
> @@ -498,8 +497,6 @@ int add_memory(int nid, u64 start, u64 size)
>  	/* we online node here. we can't roll back from here. */
>  	node_set_online(nid);
>  
> -	cpuset_track_online_nodes();
> -
>  	if (new_pgdat) {
>  		ret = register_one_node(nid);
>  		/*
> -- 
> 1.5.3.8
> 

-- 
Yasunori Goto 

Re: [PATCH] cpuset: update top cpuset's mems after adding a node

[Miao Xie]

on 2008-11-10 14:56 Yasunori Goto wrote:
> Did you test this patch?  Was it OK?

Sorry. I forget.
I will do it as soon as possible. 

> If yes,
> Acked-by: Yasunori Goto <y-goto@jp.fujitsu.com>
> 
> 
> Thanks.
> 
> 
>> After adding a node into the machine, top cpuset's mems isn't updated.
>>
>> By reviewing the code, we found that the update function
>>   cpuset_track_online_nodes()
>> was invoked after node_states[N_ONLINE] changes. It is wrong because N_ONLINE
>> just means node has pgdat, and if node has/added memory, we use N_HIGH_MEMORY.
>> So, We should invoke the update function after node_states[N_HIGH_MEMORY]
>> changes, just like its commit says.
>>
>> This patch fixes it. And we use notifier of memory hotplug instead of direct
>> calling of cpuset_track_online_nodes().
>>
>> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
>> Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
>>
>> ---
>>  include/linux/cpuset.h |    4 ----
>>  kernel/cpuset.c        |   19 ++++++++++++++++---
>>  mm/memory_hotplug.c    |    3 ---
>>  3 files changed, 16 insertions(+), 10 deletions(-)
>>
>> diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h
>> index 2691926..8e540d3 100644
>> --- a/include/linux/cpuset.h
>> +++ b/include/linux/cpuset.h
>> @@ -74,8 +74,6 @@ static inline int cpuset_do_slab_mem_spread(void)
>>  	return current->flags & PF_SPREAD_SLAB;
>>  }
>>  
>> -extern void cpuset_track_online_nodes(void);
>> -
>>  extern int current_cpuset_is_being_rebound(void);
>>  
>>  extern void rebuild_sched_domains(void);
>> @@ -151,8 +149,6 @@ static inline int cpuset_do_slab_mem_spread(void)
>>  	return 0;
>>  }
>>  
>> -static inline void cpuset_track_online_nodes(void) {}
>> -
>>  static inline int current_cpuset_is_being_rebound(void)
>>  {
>>  	return 0;
>> diff --git a/kernel/cpuset.c b/kernel/cpuset.c
>> index 3e00526..909a548 100644
>> --- a/kernel/cpuset.c
>> +++ b/kernel/cpuset.c
>> @@ -36,6 +36,7 @@
>>  #include <linux/list.h>
>>  #include <linux/mempolicy.h>
>>  #include <linux/mm.h>
>> +#include <linux/memory.h>
>>  #include <linux/module.h>
>>  #include <linux/mount.h>
>>  #include <linux/namei.h>
>> @@ -2011,12 +2012,23 @@ static int cpuset_track_online_cpus(struct notifier_block *unused_nb,
>>   * Call this routine anytime after node_states[N_HIGH_MEMORY] changes.
>>   * See also the previous routine cpuset_track_online_cpus().
>>   */
>> -void cpuset_track_online_nodes(void)
>> +static int cpuset_track_online_nodes(struct notifier_block *self,
>> +				unsigned long action, void *arg)
>>  {
>>  	cgroup_lock();
>> -	top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
>> -	scan_for_empty_cpusets(&top_cpuset);
>> +	switch (action) {
>> +	case MEM_ONLINE:
>> +		top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
>> +		break;
>> +	case MEM_OFFLINE:
>> +		top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
>> +		scan_for_empty_cpusets(&top_cpuset);
>> +		break;
>> +	default:
>> +		break;
>> +	}
>>  	cgroup_unlock();
>> +	return NOTIFY_OK;
>>  }
>>  #endif
>>  
>> @@ -2032,6 +2044,7 @@ void __init cpuset_init_smp(void)
>>  	top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
>>  
>>  	hotcpu_notifier(cpuset_track_online_cpus, 0);
>> +	hotplug_memory_notifier(cpuset_track_online_nodes, 10);
>>  }
>>  
>>  /**
>> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
>> index 6837a10..b5b2b15 100644
>> --- a/mm/memory_hotplug.c
>> +++ b/mm/memory_hotplug.c
>> @@ -22,7 +22,6 @@
>>  #include <linux/highmem.h>
>>  #include <linux/vmalloc.h>
>>  #include <linux/ioport.h>
>> -#include <linux/cpuset.h>
>>  #include <linux/delay.h>
>>  #include <linux/migrate.h>
>>  #include <linux/page-isolation.h>
>> @@ -498,8 +497,6 @@ int add_memory(int nid, u64 start, u64 size)
>>  	/* we online node here. we can't roll back from here. */
>>  	node_set_online(nid);
>>  
>> -	cpuset_track_online_nodes();
>> -
>>  	if (new_pgdat) {
>>  		ret = register_one_node(nid);
>>  		/*
>> -- 
>> 1.5.3.8
>>
> 

Re: [PATCH] cpuset: update top cpuset's mems after adding a node

[Miao Xie]

on 2008-11-10 14:56 Yasunori Goto wrote:
> Did you test this patch?  Was it OK?

I did it on IA64 just now. Updating top cpuset's mems is fine.
But I forgot to update the mems_allowed of every task in top cpuset group.
It is difficult to fix it, because I don't know whether updating a memory-bound
task's mem_allowed is good or not.

cpuset_track_online_cpus() has the same problem.

> If yes,
> Acked-by: Yasunori Goto <y-goto@jp.fujitsu.com>
> 
> Thanks.
> 
> 
>> After adding a node into the machine, top cpuset's mems isn't updated.
>>
>> By reviewing the code, we found that the update function
>>   cpuset_track_online_nodes()
>> was invoked after node_states[N_ONLINE] changes. It is wrong because N_ONLINE
>> just means node has pgdat, and if node has/added memory, we use N_HIGH_MEMORY.
>> So, We should invoke the update function after node_states[N_HIGH_MEMORY]
>> changes, just like its commit says.
>>
>> This patch fixes it. And we use notifier of memory hotplug instead of direct
>> calling of cpuset_track_online_nodes().
>>
>> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
>> Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
>>
>> ---
>>  include/linux/cpuset.h |    4 ----
>>  kernel/cpuset.c        |   19 ++++++++++++++++---
>>  mm/memory_hotplug.c    |    3 ---
>>  3 files changed, 16 insertions(+), 10 deletions(-)
>>
>> diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h
>> index 2691926..8e540d3 100644
>> --- a/include/linux/cpuset.h
>> +++ b/include/linux/cpuset.h
>> @@ -74,8 +74,6 @@ static inline int cpuset_do_slab_mem_spread(void)
>>  	return current->flags & PF_SPREAD_SLAB;
>>  }
>>  
>> -extern void cpuset_track_online_nodes(void);
>> -
>>  extern int current_cpuset_is_being_rebound(void);
>>  
>>  extern void rebuild_sched_domains(void);
>> @@ -151,8 +149,6 @@ static inline int cpuset_do_slab_mem_spread(void)
>>  	return 0;
>>  }
>>  
>> -static inline void cpuset_track_online_nodes(void) {}
>> -
>>  static inline int current_cpuset_is_being_rebound(void)
>>  {
>>  	return 0;
>> diff --git a/kernel/cpuset.c b/kernel/cpuset.c
>> index 3e00526..909a548 100644
>> --- a/kernel/cpuset.c
>> +++ b/kernel/cpuset.c
>> @@ -36,6 +36,7 @@
>>  #include <linux/list.h>
>>  #include <linux/mempolicy.h>
>>  #include <linux/mm.h>
>> +#include <linux/memory.h>
>>  #include <linux/module.h>
>>  #include <linux/mount.h>
>>  #include <linux/namei.h>
>> @@ -2011,12 +2012,23 @@ static int cpuset_track_online_cpus(struct notifier_block *unused_nb,
>>   * Call this routine anytime after node_states[N_HIGH_MEMORY] changes.
>>   * See also the previous routine cpuset_track_online_cpus().
>>   */
>> -void cpuset_track_online_nodes(void)
>> +static int cpuset_track_online_nodes(struct notifier_block *self,
>> +				unsigned long action, void *arg)
>>  {
>>  	cgroup_lock();
>> -	top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
>> -	scan_for_empty_cpusets(&top_cpuset);
>> +	switch (action) {
>> +	case MEM_ONLINE:
>> +		top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
>> +		break;
>> +	case MEM_OFFLINE:
>> +		top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
>> +		scan_for_empty_cpusets(&top_cpuset);
>> +		break;
>> +	default:
>> +		break;
>> +	}
>>  	cgroup_unlock();
>> +	return NOTIFY_OK;
>>  }
>>  #endif
>>  
>> @@ -2032,6 +2044,7 @@ void __init cpuset_init_smp(void)
>>  	top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
>>  
>>  	hotcpu_notifier(cpuset_track_online_cpus, 0);
>> +	hotplug_memory_notifier(cpuset_track_online_nodes, 10);
>>  }
>>  
>>  /**
>> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
>> index 6837a10..b5b2b15 100644
>> --- a/mm/memory_hotplug.c
>> +++ b/mm/memory_hotplug.c
>> @@ -22,7 +22,6 @@
>>  #include <linux/highmem.h>
>>  #include <linux/vmalloc.h>
>>  #include <linux/ioport.h>
>> -#include <linux/cpuset.h>
>>  #include <linux/delay.h>
>>  #include <linux/migrate.h>
>>  #include <linux/page-isolation.h>
>> @@ -498,8 +497,6 @@ int add_memory(int nid, u64 start, u64 size)
>>  	/* we online node here. we can't roll back from here. */
>>  	node_set_online(nid);
>>  
>> -	cpuset_track_online_nodes();
>> -
>>  	if (new_pgdat) {
>>  		ret = register_one_node(nid);
>>  		/*
>> -- 
>> 1.5.3.8
>>
> 

Re: [PATCH] cpuset: update top cpuset's mems after adding a node

[David Rientjes]

On Mon, 10 Nov 2008, Miao Xie wrote:

> on 2008-11-10 14:56 Yasunori Goto wrote:
> > Did you test this patch?  Was it OK?
> 
> I did it on IA64 just now. Updating top cpuset's mems is fine.
> But I forgot to update the mems_allowed of every task in top cpuset group.
> It is difficult to fix it, because I don't know whether updating a memory-bound
> task's mem_allowed is good or not.
> 

There's a fix pending in the mmotm tree that calls 
cpuset_update_task_memory_state() in the slow path of the page allocator, 
so if one of these tasks fails to allocate memory, its cached mems_allowed 
nodemask will be updated to include the new node(s).

The patch is available at 
http://userweb.kernel.org/~akpm/mmotm/broken-out/cpusets-update-mems-allowed-in-page-allocator.patch

Re: [PATCH] cpuset: update top cpuset's mems after adding a node

[Yasunori Goto]

> On Mon, 10 Nov 2008, Miao Xie wrote:
> 
> > on 2008-11-10 14:56 Yasunori Goto wrote:
> > > Did you test this patch?  Was it OK?
> > 
> > I did it on IA64 just now. Updating top cpuset's mems is fine.
> > But I forgot to update the mems_allowed of every task in top cpuset group.
> > It is difficult to fix it, because I don't know whether updating a memory-bound
> > task's mem_allowed is good or not.
> > 
> 
> There's a fix pending in the mmotm tree that calls 
> cpuset_update_task_memory_state() in the slow path of the page allocator, 
> so if one of these tasks fails to allocate memory, its cached mems_allowed 
> nodemask will be updated to include the new node(s).
> 
> The patch is available at 
> http://userweb.kernel.org/~akpm/mmotm/broken-out/cpusets-update-mems-allowed-in-page-allocator.patch

Ah, OK. This seems reasonable.

Then Miao-san's patch is fine.
Acked-by: Yasunori Goto <y-goto@jp.fujitsu.com>


Thanks.
-- 
Yasunori Goto