Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1649572 > unrolled thread

[RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction on online/offline

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  [RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction on online/offline Michal Hocko <mhocko@kernel.org> - 2017-05-24 14:30 +0200
    Re: [RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction  on online/offline Vlastimil Babka <vbabka@suse.cz> - 2017-05-24 14:50 +0200
      Re: [RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction  on online/offline Michal Hocko <mhocko@kernel.org> - 2017-05-24 15:00 +0200
    Re: [RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction  on online/offline Reza Arbab <arbab@linux.vnet.ibm.com> - 2017-05-25 00:00 +0200
      Re: [RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction  on online/offline Michal Hocko <mhocko@kernel.org> - 2017-05-25 08:30 +0200

#1649572 — [RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction on online/offline

From: Michal Hocko <mhocko@suse.com>

74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
memory") has added can_offline_normal which checks the amount of
memory in !movable zones as long as CONFIG_MOVABLE_NODE is disable.
It disallows to offline memory if there is nothing left with a
justification that "memory-management acts bad when we have nodes which
is online but don't have any normal memory".

74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
memory") has introduced a restriction that every numa node has to have
at least some memory in !movable zones before a first movable memory
can be onlined if !CONFIG_MOVABLE_NODE with the same justification

While it is true that not having _any_ memory for kernel allocations on
a NUMA node is far from great and such a node would be quite subotimal
because all kernel allocations will have to fallback to another NUMA
node but there is no reason to disallow such a configuration in
principle.

Besides that there is not really a big difference to have one memblock
for ZONE_NORMAL available or none. With 128MB size memblocks the system
might trash on the kernel allocations requests anyway. It is really
hard to draw a line on how much normal memory is really sufficient so
we have to rely on administrator to configure system sanely therefore
drop the artificial restriction and remove can_offline_normal and
can_online_high_movable altogether.

Signed-off-by: Michal Hocko <mhocko@suse.com>

mm, memory_hotplug: drop can_online_high_movable

 because "memory-management acts
bad when we have nodes which is online but don't have any normal memory.

Signed-off-by: Michal Hocko <mhocko@suse.com>
---
 mm/memory_hotplug.c | 58 -----------------------------------------------------
 1 file changed, 58 deletions(-)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 599c675ad538..10052c2fd400 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -763,23 +763,6 @@ static int online_pages_range(unsigned long start_pfn, unsigned long nr_pages,
 	return 0;
 }
 
-#ifdef CONFIG_MOVABLE_NODE
-/*
- * When CONFIG_MOVABLE_NODE, we permit onlining of a node which doesn't have
- * normal memory.
- */
-static bool can_online_high_movable(int nid)
-{
-	return true;
-}
-#else /* CONFIG_MOVABLE_NODE */
-/* ensure every online node has NORMAL memory */
-static bool can_online_high_movable(int nid)
-{
-	return node_state(nid, N_NORMAL_MEMORY);
-}
-#endif /* CONFIG_MOVABLE_NODE */
-
 /* check which state of node_states will be changed when online memory */
 static void node_states_check_changes_online(unsigned long nr_pages,
 	struct zone *zone, struct memory_notify *arg)
@@ -979,9 +962,6 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_typ
 	if (!allow_online_pfn_range(nid, pfn, nr_pages, online_type))
 		return -EINVAL;
 
-	if (online_type == MMOP_ONLINE_MOVABLE && !can_online_high_movable(nid))
-		return -EINVAL;
-
 	/* associate pfn range with the zone */
 	zone = move_pfn_range(online_type, nid, pfn, nr_pages);
 
@@ -1579,41 +1559,6 @@ check_pages_isolated(unsigned long start_pfn, unsigned long end_pfn)
 	return offlined;
 }
 
-#ifdef CONFIG_MOVABLE_NODE
-/*
- * When CONFIG_MOVABLE_NODE, we permit offlining of a node which doesn't have
- * normal memory.
- */
-static bool can_offline_normal(struct zone *zone, unsigned long nr_pages)
-{
-	return true;
-}
-#else /* CONFIG_MOVABLE_NODE */
-/* ensure the node has NORMAL memory if it is still online */
-static bool can_offline_normal(struct zone *zone, unsigned long nr_pages)
-{
-	struct pglist_data *pgdat = zone->zone_pgdat;
-	unsigned long present_pages = 0;
-	enum zone_type zt;
-
-	for (zt = 0; zt <= ZONE_NORMAL; zt++)
-		present_pages += pgdat->node_zones[zt].present_pages;
-
-	if (present_pages > nr_pages)
-		return true;
-
-	present_pages = 0;
-	for (; zt <= ZONE_MOVABLE; zt++)
-		present_pages += pgdat->node_zones[zt].present_pages;
-
-	/*
-	 * we can't offline the last normal memory until all
-	 * higher memory is offlined.
-	 */
-	return present_pages == 0;
-}
-#endif /* CONFIG_MOVABLE_NODE */
-
 static int __init cmdline_parse_movable_node(char *p)
 {
 #ifdef CONFIG_MOVABLE_NODE
@@ -1741,9 +1686,6 @@ static int __ref __offline_pages(unsigned long start_pfn,
 	node = zone_to_nid(zone);
 	nr_pages = end_pfn - start_pfn;
 
-	if (zone_idx(zone) <= ZONE_NORMAL && !can_offline_normal(zone, nr_pages))
-		return -EINVAL;
-
 	/* set above range as isolated */
 	ret = start_isolate_page_range(start_pfn, end_pfn,
 				       MIGRATE_MOVABLE, true);
-- 
2.11.0

[toc] | [next] | [standalone]

#1649590 — Re: [RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction on online/offline

On 05/24/2017 02:24 PM, Michal Hocko wrote:
> From: Michal Hocko <mhocko@suse.com>
> 
> 74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
> memory") has added can_offline_normal which checks the amount of
> memory in !movable zones as long as CONFIG_MOVABLE_NODE is disable.
> It disallows to offline memory if there is nothing left with a
> justification that "memory-management acts bad when we have nodes which
> is online but don't have any normal memory".
> 
> 74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
> memory")

That's the same commit as above... one of them should be different?

> has introduced a restriction that every numa node has to have
> at least some memory in !movable zones before a first movable memory
> can be onlined if !CONFIG_MOVABLE_NODE with the same justification
> 
> While it is true that not having _any_ memory for kernel allocations on
> a NUMA node is far from great and such a node would be quite subotimal
> because all kernel allocations will have to fallback to another NUMA
> node but there is no reason to disallow such a configuration in
> principle.
> 
> Besides that there is not really a big difference to have one memblock
> for ZONE_NORMAL available or none. With 128MB size memblocks the system
> might trash on the kernel allocations requests anyway. It is really
> hard to draw a line on how much normal memory is really sufficient so
> we have to rely on administrator to configure system sanely therefore
> drop the artificial restriction and remove can_offline_normal and
> can_online_high_movable altogether.
> 
> Signed-off-by: Michal Hocko <mhocko@suse.com>

-

> mm, memory_hotplug: drop can_online_high_movable
> 
>  because "memory-management acts
> bad when we have nodes which is online but don't have any normal memory.
> 
> Signed-off-by: Michal Hocko <mhocko@suse.com>

-
Some editing issue?

Otherwise makes sense to me.
Acked-by: Vlastimil Babka <vbabka@suse.cz>

> ---
>  mm/memory_hotplug.c | 58 -----------------------------------------------------
>  1 file changed, 58 deletions(-)
> 
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index 599c675ad538..10052c2fd400 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -763,23 +763,6 @@ static int online_pages_range(unsigned long start_pfn, unsigned long nr_pages,
>  	return 0;
>  }
>  
> -#ifdef CONFIG_MOVABLE_NODE
> -/*
> - * When CONFIG_MOVABLE_NODE, we permit onlining of a node which doesn't have
> - * normal memory.
> - */
> -static bool can_online_high_movable(int nid)
> -{
> -	return true;
> -}
> -#else /* CONFIG_MOVABLE_NODE */
> -/* ensure every online node has NORMAL memory */
> -static bool can_online_high_movable(int nid)
> -{
> -	return node_state(nid, N_NORMAL_MEMORY);
> -}
> -#endif /* CONFIG_MOVABLE_NODE */
> -
>  /* check which state of node_states will be changed when online memory */
>  static void node_states_check_changes_online(unsigned long nr_pages,
>  	struct zone *zone, struct memory_notify *arg)
> @@ -979,9 +962,6 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_typ
>  	if (!allow_online_pfn_range(nid, pfn, nr_pages, online_type))
>  		return -EINVAL;
>  
> -	if (online_type == MMOP_ONLINE_MOVABLE && !can_online_high_movable(nid))
> -		return -EINVAL;
> -
>  	/* associate pfn range with the zone */
>  	zone = move_pfn_range(online_type, nid, pfn, nr_pages);
>  
> @@ -1579,41 +1559,6 @@ check_pages_isolated(unsigned long start_pfn, unsigned long end_pfn)
>  	return offlined;
>  }
>  
> -#ifdef CONFIG_MOVABLE_NODE
> -/*
> - * When CONFIG_MOVABLE_NODE, we permit offlining of a node which doesn't have
> - * normal memory.
> - */
> -static bool can_offline_normal(struct zone *zone, unsigned long nr_pages)
> -{
> -	return true;
> -}
> -#else /* CONFIG_MOVABLE_NODE */
> -/* ensure the node has NORMAL memory if it is still online */
> -static bool can_offline_normal(struct zone *zone, unsigned long nr_pages)
> -{
> -	struct pglist_data *pgdat = zone->zone_pgdat;
> -	unsigned long present_pages = 0;
> -	enum zone_type zt;
> -
> -	for (zt = 0; zt <= ZONE_NORMAL; zt++)
> -		present_pages += pgdat->node_zones[zt].present_pages;
> -
> -	if (present_pages > nr_pages)
> -		return true;
> -
> -	present_pages = 0;
> -	for (; zt <= ZONE_MOVABLE; zt++)
> -		present_pages += pgdat->node_zones[zt].present_pages;
> -
> -	/*
> -	 * we can't offline the last normal memory until all
> -	 * higher memory is offlined.
> -	 */
> -	return present_pages == 0;
> -}
> -#endif /* CONFIG_MOVABLE_NODE */
> -
>  static int __init cmdline_parse_movable_node(char *p)
>  {
>  #ifdef CONFIG_MOVABLE_NODE
> @@ -1741,9 +1686,6 @@ static int __ref __offline_pages(unsigned long start_pfn,
>  	node = zone_to_nid(zone);
>  	nr_pages = end_pfn - start_pfn;
>  
> -	if (zone_idx(zone) <= ZONE_NORMAL && !can_offline_normal(zone, nr_pages))
> -		return -EINVAL;
> -
>  	/* set above range as isolated */
>  	ret = start_isolate_page_range(start_pfn, end_pfn,
>  				       MIGRATE_MOVABLE, true);
> 

[toc] | [prev] | [next] | [standalone]

#1649593 — Re: [RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction on online/offline

On Wed 24-05-17 14:44:34, Vlastimil Babka wrote:
> On 05/24/2017 02:24 PM, Michal Hocko wrote:
> > From: Michal Hocko <mhocko@suse.com>
> > 
> > 74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
> > memory") has added can_offline_normal which checks the amount of
> > memory in !movable zones as long as CONFIG_MOVABLE_NODE is disable.
> > It disallows to offline memory if there is nothing left with a
> > justification that "memory-management acts bad when we have nodes which
> > is online but don't have any normal memory".
> > 
> > 74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
> > memory")
> 
> That's the same commit as above... one of them should be different?

This used to be two different patches which I decided to fold together
and I didn't realize that both online and offline paths were introduced
by the same patch.
[...]
> Some editing issue?

yes result of merging two commits.

> Otherwise makes sense to me.
> Acked-by: Vlastimil Babka <vbabka@suse.cz>

Thanks! Updated version follows
---
From 5196f9e206d780ac473075aa0ad88308409f5acb Mon Sep 17 00:00:00 2001
From: Michal Hocko <mhocko@suse.com>
Date: Tue, 23 May 2017 10:32:04 +0200
Subject: [PATCH] mm, memory_hotplug: drop artificial restriction on
 online/offline

74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
memory") has introduced a restriction that every numa node has to have
at least some memory in !movable zones before a first movable memory can
be onlined if !CONFIG_MOVABLE_NODE. Likewise_offline_normal checks the
amount of normal memory in !movable zones and it disallows to offline
memory if there is no normal memory left with a justification that
"memory-management acts bad when we have nodes which is online but don't
have any normal memory".

While it is true that not having _any_ memory for kernel allocations on
a NUMA node is far from great and such a node would be quite subotimal
because all kernel allocations will have to fallback to another NUMA
node but there is no reason to disallow such a configuration in
principle.

Besides that there is not really a big difference to have one memblock
for ZONE_NORMAL available or none. With 128MB size memblocks the system
might trash on the kernel allocations requests anyway. It is really
hard to draw a line on how much normal memory is really sufficient so
we have to rely on administrator to configure system sanely therefore
drop the artificial restriction and remove can_offline_normal and
can_online_high_movable altogether.

Acked-by: Acked-by: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Michal Hocko <mhocko@suse.com>
---
 mm/memory_hotplug.c | 58 -----------------------------------------------------
 1 file changed, 58 deletions(-)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 599c675ad538..10052c2fd400 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -763,23 +763,6 @@ static int online_pages_range(unsigned long start_pfn, unsigned long nr_pages,
 	return 0;
 }
 
-#ifdef CONFIG_MOVABLE_NODE
-/*
- * When CONFIG_MOVABLE_NODE, we permit onlining of a node which doesn't have
- * normal memory.
- */
-static bool can_online_high_movable(int nid)
-{
-	return true;
-}
-#else /* CONFIG_MOVABLE_NODE */
-/* ensure every online node has NORMAL memory */
-static bool can_online_high_movable(int nid)
-{
-	return node_state(nid, N_NORMAL_MEMORY);
-}
-#endif /* CONFIG_MOVABLE_NODE */
-
 /* check which state of node_states will be changed when online memory */
 static void node_states_check_changes_online(unsigned long nr_pages,
 	struct zone *zone, struct memory_notify *arg)
@@ -979,9 +962,6 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_typ
 	if (!allow_online_pfn_range(nid, pfn, nr_pages, online_type))
 		return -EINVAL;
 
-	if (online_type == MMOP_ONLINE_MOVABLE && !can_online_high_movable(nid))
-		return -EINVAL;
-
 	/* associate pfn range with the zone */
 	zone = move_pfn_range(online_type, nid, pfn, nr_pages);
 
@@ -1579,41 +1559,6 @@ check_pages_isolated(unsigned long start_pfn, unsigned long end_pfn)
 	return offlined;
 }
 
-#ifdef CONFIG_MOVABLE_NODE
-/*
- * When CONFIG_MOVABLE_NODE, we permit offlining of a node which doesn't have
- * normal memory.
- */
-static bool can_offline_normal(struct zone *zone, unsigned long nr_pages)
-{
-	return true;
-}
-#else /* CONFIG_MOVABLE_NODE */
-/* ensure the node has NORMAL memory if it is still online */
-static bool can_offline_normal(struct zone *zone, unsigned long nr_pages)
-{
-	struct pglist_data *pgdat = zone->zone_pgdat;
-	unsigned long present_pages = 0;
-	enum zone_type zt;
-
-	for (zt = 0; zt <= ZONE_NORMAL; zt++)
-		present_pages += pgdat->node_zones[zt].present_pages;
-
-	if (present_pages > nr_pages)
-		return true;
-
-	present_pages = 0;
-	for (; zt <= ZONE_MOVABLE; zt++)
-		present_pages += pgdat->node_zones[zt].present_pages;
-
-	/*
-	 * we can't offline the last normal memory until all
-	 * higher memory is offlined.
-	 */
-	return present_pages == 0;
-}
-#endif /* CONFIG_MOVABLE_NODE */
-
 static int __init cmdline_parse_movable_node(char *p)
 {
 #ifdef CONFIG_MOVABLE_NODE
@@ -1741,9 +1686,6 @@ static int __ref __offline_pages(unsigned long start_pfn,
 	node = zone_to_nid(zone);
 	nr_pages = end_pfn - start_pfn;
 
-	if (zone_idx(zone) <= ZONE_NORMAL && !can_offline_normal(zone, nr_pages))
-		return -EINVAL;
-
 	/* set above range as isolated */
 	ret = start_isolate_page_range(start_pfn, end_pfn,
 				       MIGRATE_MOVABLE, true);
-- 
2.11.0

-- 
Michal Hocko
SUSE Labs

[toc] | [prev] | [next] | [standalone]

#1649983 — Re: [RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction on online/offline

On Wed, May 24, 2017 at 02:24:10PM +0200, Michal Hocko wrote:
>74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
>memory") has added can_offline_normal which checks the amount of
>memory in !movable zones as long as CONFIG_MOVABLE_NODE is disable.
>It disallows to offline memory if there is nothing left with a
>justification that "memory-management acts bad when we have nodes which
>is online but don't have any normal memory".
>
>74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
>memory") has introduced a restriction that every numa node has to have
>at least some memory in !movable zones before a first movable memory
>can be onlined if !CONFIG_MOVABLE_NODE with the same justification
>
>While it is true that not having _any_ memory for kernel allocations on
>a NUMA node is far from great and such a node would be quite subotimal
>because all kernel allocations will have to fallback to another NUMA
>node but there is no reason to disallow such a configuration in
>principle.
>
>Besides that there is not really a big difference to have one memblock
>for ZONE_NORMAL available or none. With 128MB size memblocks the system
>might trash on the kernel allocations requests anyway. It is really
>hard to draw a line on how much normal memory is really sufficient so
>we have to rely on administrator to configure system sanely therefore
>drop the artificial restriction and remove can_offline_normal and
>can_online_high_movable altogether.

I'm really liking all this cleanup of the memory hotplug code. Thanks!  
Much appreciated.

Acked-by: Reza Arbab <arbab@linux.vnet.ibm.com>

-- 
Reza Arbab

[toc] | [prev] | [next] | [standalone]

#1650210 — Re: [RFC PATCH 1/2] mm, memory_hotplug: drop artificial restriction on online/offline

On Wed 24-05-17 16:50:56, Reza Arbab wrote:
> On Wed, May 24, 2017 at 02:24:10PM +0200, Michal Hocko wrote:
> >74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
> >memory") has added can_offline_normal which checks the amount of
> >memory in !movable zones as long as CONFIG_MOVABLE_NODE is disable.
> >It disallows to offline memory if there is nothing left with a
> >justification that "memory-management acts bad when we have nodes which
> >is online but don't have any normal memory".
> >
> >74d42d8fe146 ("memory_hotplug: ensure every online node has NORMAL
> >memory") has introduced a restriction that every numa node has to have
> >at least some memory in !movable zones before a first movable memory
> >can be onlined if !CONFIG_MOVABLE_NODE with the same justification
> >
> >While it is true that not having _any_ memory for kernel allocations on
> >a NUMA node is far from great and such a node would be quite subotimal
> >because all kernel allocations will have to fallback to another NUMA
> >node but there is no reason to disallow such a configuration in
> >principle.
> >
> >Besides that there is not really a big difference to have one memblock
> >for ZONE_NORMAL available or none. With 128MB size memblocks the system
> >might trash on the kernel allocations requests anyway. It is really
> >hard to draw a line on how much normal memory is really sufficient so
> >we have to rely on administrator to configure system sanely therefore
> >drop the artificial restriction and remove can_offline_normal and
> >can_online_high_movable altogether.
> 
> I'm really liking all this cleanup of the memory hotplug code. Thanks!  Much
> appreciated.

I am glad to hear that and more is to come.

> Acked-by: Reza Arbab <arbab@linux.vnet.ibm.com>

Thanks!
-- 
Michal Hocko
SUSE Labs

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web