[PATCH v5 0/4] Refactor free_area_init_core and add free_area_init_core_hotplug

[PATCH v5 0/4] Refactor free_area_init_core and add free_area_init_core_hotplug

[osalvador]

From: Oscar Salvador <osalvador@suse.de>

Changes:

v4 -> v5:
	- Remove __ref from hotadd_new_pgdat and placed it to
	  free_area_init_core_hotplug. (Suggested by Pavel)
	- Since free_area_init_core_hotplug is now allowed to be in a different
	  section (__ref), remove the __paginginit.)
	- Stylecode in free_area_init_core_hotplug (Suggested by Pavel)
	- Replace s/@__paginginit/@__init for free_area_init_node/free_area_init_core
	  as these functions are now only called during early init.
	- Add Reviewd-by from Pavel

v3 -> v4:
        - Unify patch-5 and patch-4.
        - Make free_area_init_core __init (Suggested by Michal).
        - Make zone_init_internals __paginginit (Suggested by Pavel).
        - Add Reviewed-by/Acked-by:

v2 -> v3:
        - Think better about split free_area_init_core for
          memhotplug/early init context (Suggested by Michal).

This patchset does three things:

 1) Clean up/refactor free_area_init_core/free_area_init_node
    by moving the ifdefery out of the functions.
 2) Move the pgdat/zone initialization in free_area_init_core to its
    own function.
 3) Introduce free_area_init_core_hotplug, a small subset of free_area_init_core,
    which is only called from memhotlug code path.
    In this way, we have:

    free_area_init_core: called during early initialization
    free_area_init_core_hotplug: called whenever a new node is allocated/re-used (memhotplug path)

Oscar Salvador (3):
  mm/page_alloc: Move ifdefery out of free_area_init_core
  mm/page_alloc: Inline function to handle
    CONFIG_DEFERRED_STRUCT_PAGE_INIT
  mm/page_alloc: Introduce free_area_init_core_hotplug

Pavel Tatashin (1):
  mm: access zone->node via zone_to_nid() and zone_set_nid()

 include/linux/mm.h     |  15 ++----
 include/linux/mmzone.h |  26 +++++++---
 mm/memory_hotplug.c    |  16 ++----
 mm/mempolicy.c         |   4 +-
 mm/mm_init.c           |   9 +---
 mm/page_alloc.c        | 134 +++++++++++++++++++++++++++++++++++--------------
 6 files changed, 130 insertions(+), 74 deletions(-)

-- 
2.13.6

[PATCH v5 1/4] mm/page_alloc: Move ifdefery out of free_area_init_core

[osalvador]

From: Oscar Salvador <osalvador@suse.de>

Moving the #ifdefs out of the function makes it easier to follow.

Signed-off-by: Oscar Salvador <osalvador@suse.de>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Pavel Tatashin <pasha.tatashin@oracle.com>
---
 mm/page_alloc.c | 50 +++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 37 insertions(+), 13 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index e357189cd24a..8a73305f7c55 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6206,6 +6206,37 @@ static unsigned long __paginginit calc_memmap_size(unsigned long spanned_pages,
 	return PAGE_ALIGN(pages * sizeof(struct page)) >> PAGE_SHIFT;
 }
 
+#ifdef CONFIG_NUMA_BALANCING
+static void pgdat_init_numabalancing(struct pglist_data *pgdat)
+{
+	spin_lock_init(&pgdat->numabalancing_migrate_lock);
+	pgdat->numabalancing_migrate_nr_pages = 0;
+	pgdat->numabalancing_migrate_next_window = jiffies;
+}
+#else
+static void pgdat_init_numabalancing(struct pglist_data *pgdat) {}
+#endif
+
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+static void pgdat_init_split_queue(struct pglist_data *pgdat)
+{
+	spin_lock_init(&pgdat->split_queue_lock);
+	INIT_LIST_HEAD(&pgdat->split_queue);
+	pgdat->split_queue_len = 0;
+}
+#else
+static void pgdat_init_split_queue(struct pglist_data *pgdat) {}
+#endif
+
+#ifdef CONFIG_COMPACTION
+static void pgdat_init_kcompactd(struct pglist_data *pgdat)
+{
+	init_waitqueue_head(&pgdat->kcompactd_wait);
+}
+#else
+static void pgdat_init_kcompactd(struct pglist_data *pgdat) {}
+#endif
+
 /*
  * Set up the zone data structures:
  *   - mark all pages reserved
@@ -6220,21 +6251,14 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
 	int nid = pgdat->node_id;
 
 	pgdat_resize_init(pgdat);
-#ifdef CONFIG_NUMA_BALANCING
-	spin_lock_init(&pgdat->numabalancing_migrate_lock);
-	pgdat->numabalancing_migrate_nr_pages = 0;
-	pgdat->numabalancing_migrate_next_window = jiffies;
-#endif
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
-	spin_lock_init(&pgdat->split_queue_lock);
-	INIT_LIST_HEAD(&pgdat->split_queue);
-	pgdat->split_queue_len = 0;
-#endif
+
+	pgdat_init_numabalancing(pgdat);
+	pgdat_init_split_queue(pgdat);
+	pgdat_init_kcompactd(pgdat);
+
 	init_waitqueue_head(&pgdat->kswapd_wait);
 	init_waitqueue_head(&pgdat->pfmemalloc_wait);
-#ifdef CONFIG_COMPACTION
-	init_waitqueue_head(&pgdat->kcompactd_wait);
-#endif
+
 	pgdat_page_ext_init(pgdat);
 	spin_lock_init(&pgdat->lru_lock);
 	lruvec_init(node_lruvec(pgdat));
-- 
2.13.6

[PATCH v5 2/4] mm: access zone->node via zone_to_nid() and zone_set_nid()

[osalvador]

From: Pavel Tatashin <pasha.tatashin@oracle.com>

zone->node is configured only when CONFIG_NUMA=y, so it is a good idea to
have inline functions to access this field in order to avoid ifdef's in
c files.

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Signed-off-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Acked-by: Michal Hocko <mhocko@suse.com>
---
 include/linux/mm.h     |  9 ---------
 include/linux/mmzone.h | 26 ++++++++++++++++++++------
 mm/mempolicy.c         |  4 ++--
 mm/mm_init.c           |  9 ++-------
 mm/page_alloc.c        | 10 ++++------
 5 files changed, 28 insertions(+), 30 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 726e71475144..6954ad183159 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -940,15 +940,6 @@ static inline int page_zone_id(struct page *page)
 	return (page->flags >> ZONEID_PGSHIFT) & ZONEID_MASK;
 }
 
-static inline int zone_to_nid(struct zone *zone)
-{
-#ifdef CONFIG_NUMA
-	return zone->node;
-#else
-	return 0;
-#endif
-}
-
 #ifdef NODE_NOT_IN_PAGE_FLAGS
 extern int page_to_nid(const struct page *page);
 #else
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index ae1a034c3e2c..17fdff3bfb41 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -842,6 +842,25 @@ static inline bool populated_zone(struct zone *zone)
 	return zone->present_pages;
 }
 
+#ifdef CONFIG_NUMA
+static inline int zone_to_nid(struct zone *zone)
+{
+	return zone->node;
+}
+
+static inline void zone_set_nid(struct zone *zone, int nid)
+{
+	zone->node = nid;
+}
+#else
+static inline int zone_to_nid(struct zone *zone)
+{
+	return 0;
+}
+
+static inline void zone_set_nid(struct zone *zone, int nid) {}
+#endif
+
 extern int movable_zone;
 
 #ifdef CONFIG_HIGHMEM
@@ -957,12 +976,7 @@ static inline int zonelist_zone_idx(struct zoneref *zoneref)
 
 static inline int zonelist_node_idx(struct zoneref *zoneref)
 {
-#ifdef CONFIG_NUMA
-	/* zone_to_nid not available in this context */
-	return zoneref->zone->node;
-#else
-	return 0;
-#endif /* CONFIG_NUMA */
+	return zone_to_nid(zoneref->zone);
 }
 
 struct zoneref *__next_zones_zonelist(struct zoneref *z,
diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index f0fcf70bcec7..8c1c09b3852a 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -1784,7 +1784,7 @@ unsigned int mempolicy_slab_node(void)
 		zonelist = &NODE_DATA(node)->node_zonelists[ZONELIST_FALLBACK];
 		z = first_zones_zonelist(zonelist, highest_zoneidx,
 							&policy->v.nodes);
-		return z->zone ? z->zone->node : node;
+		return z->zone ? zone_to_nid(z->zone) : node;
 	}
 
 	default:
@@ -2326,7 +2326,7 @@ int mpol_misplaced(struct page *page, struct vm_area_struct *vma, unsigned long
 				node_zonelist(numa_node_id(), GFP_HIGHUSER),
 				gfp_zone(GFP_HIGHUSER),
 				&pol->v.nodes);
-		polnid = z->zone->node;
+		polnid = zone_to_nid(z->zone);
 		break;
 
 	default:
diff --git a/mm/mm_init.c b/mm/mm_init.c
index 5b72266b4b03..6838a530789b 100644
--- a/mm/mm_init.c
+++ b/mm/mm_init.c
@@ -53,13 +53,8 @@ void __init mminit_verify_zonelist(void)
 				zone->name);
 
 			/* Iterate the zonelist */
-			for_each_zone_zonelist(zone, z, zonelist, zoneid) {
-#ifdef CONFIG_NUMA
-				pr_cont("%d:%s ", zone->node, zone->name);
-#else
-				pr_cont("0:%s ", zone->name);
-#endif /* CONFIG_NUMA */
-			}
+			for_each_zone_zonelist(zone, z, zonelist, zoneid)
+				pr_cont("%d:%s ", zone_to_nid(zone), zone->name);
 			pr_cont("\n");
 		}
 	}
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 8a73305f7c55..10b754fba5fa 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2909,10 +2909,10 @@ static inline void zone_statistics(struct zone *preferred_zone, struct zone *z)
 	if (!static_branch_likely(&vm_numa_stat_key))
 		return;
 
-	if (z->node != numa_node_id())
+	if (zone_to_nid(z) != numa_node_id())
 		local_stat = NUMA_OTHER;
 
-	if (z->node == preferred_zone->node)
+	if (zone_to_nid(z) == zone_to_nid(preferred_zone))
 		__inc_numa_state(z, NUMA_HIT);
 	else {
 		__inc_numa_state(z, NUMA_MISS);
@@ -5287,7 +5287,7 @@ int local_memory_node(int node)
 	z = first_zones_zonelist(node_zonelist(node, GFP_KERNEL),
 				   gfp_zone(GFP_KERNEL),
 				   NULL);
-	return z->zone->node;
+	return zone_to_nid(z->zone);
 }
 #endif
 
@@ -6311,9 +6311,7 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
 		 * And all highmem pages will be managed by the buddy system.
 		 */
 		zone->managed_pages = freesize;
-#ifdef CONFIG_NUMA
-		zone->node = nid;
-#endif
+		zone_set_nid(zone, nid);
 		zone->name = zone_names[j];
 		zone->zone_pgdat = pgdat;
 		spin_lock_init(&zone->lock);
-- 
2.13.6

[PATCH v5 3/4] mm/page_alloc: Inline function to handle CONFIG_DEFERRED_STRUCT_PAGE_INIT

[osalvador]

From: Oscar Salvador <osalvador@suse.de>

Let us move the code between CONFIG_DEFERRED_STRUCT_PAGE_INIT
to an inline function.
Not having an ifdef in the function makes the code more readable.

Signed-off-by: Oscar Salvador <osalvador@suse.de>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Pavel Tatashin <pasha.tatashin@oracle.com>
---
 mm/page_alloc.c | 25 ++++++++++++++++---------
 1 file changed, 16 insertions(+), 9 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 10b754fba5fa..4e84a17a5030 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6376,6 +6376,21 @@ static void __ref alloc_node_mem_map(struct pglist_data *pgdat)
 static void __ref alloc_node_mem_map(struct pglist_data *pgdat) { }
 #endif /* CONFIG_FLAT_NODE_MEM_MAP */
 
+#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
+static inline void pgdat_set_deferred_range(pg_data_t *pgdat)
+{
+	/*
+	 * We start only with one section of pages, more pages are added as
+	 * needed until the rest of deferred pages are initialized.
+	 */
+	pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION,
+						pgdat->node_spanned_pages);
+	pgdat->first_deferred_pfn = ULONG_MAX;
+}
+#else
+static inline void pgdat_set_deferred_range(pg_data_t *pgdat) {}
+#endif
+
 void __paginginit free_area_init_node(int nid, unsigned long *zones_size,
 		unsigned long node_start_pfn, unsigned long *zholes_size)
 {
@@ -6401,16 +6416,8 @@ void __paginginit free_area_init_node(int nid, unsigned long *zones_size,
 				  zones_size, zholes_size);
 
 	alloc_node_mem_map(pgdat);
+	pgdat_set_deferred_range(pgdat);
 
-#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
-	/*
-	 * We start only with one section of pages, more pages are added as
-	 * needed until the rest of deferred pages are initialized.
-	 */
-	pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION,
-					 pgdat->node_spanned_pages);
-	pgdat->first_deferred_pfn = ULONG_MAX;
-#endif
 	free_area_init_core(pgdat);
 }
 
-- 
2.13.6

[PATCH v5 4/4] mm/page_alloc: Introduce free_area_init_core_hotplug

[osalvador]

From: Oscar Salvador <osalvador@suse.de>

Currently, whenever a new node is created/re-used from the memhotplug path,
we call free_area_init_node()->free_area_init_core().
But there is some code that we do not really need to run when we are coming
from such path.

free_area_init_core() performs the following actions:

1) Initializes pgdat internals, such as spinlock, waitqueues and more.
2) Account # nr_all_pages and # nr_kernel_pages. These values are used later on
   when creating hash tables.
3) Account number of managed_pages per zone, substracting dma_reserved and memmap pages.
4) Initializes some fields of the zone structure data
5) Calls init_currently_empty_zone to initialize all the freelists
6) Calls memmap_init to initialize all pages belonging to certain zone

When called from memhotplug path, free_area_init_core() only performs actions #1 and #4.

Action #2 is pointless as the zones do not have any pages since either the node was freed,
or we are re-using it, eitherway all zones belonging to this node should have 0 pages.
For the same reason, action #3 results always in manages_pages being 0.

Action #5 and #6 are performed later on when onlining the pages:
 online_pages()->move_pfn_range_to_zone()->init_currently_empty_zone()
 online_pages()->move_pfn_range_to_zone()->memmap_init_zone()

This patch does two things:

First, moves the node/zone initializtion to their own function, so it allows us
to create a small version of free_area_init_core, where we only perform:

1) Initialization of pgdat internals, such as spinlock, waitqueues and more
4) Initialization of some fields of the zone structure data

These two functions are: pgdat_init_internals() and zone_init_internals().

The second thing this patch does, is to introduce free_area_init_core_hotplug(),
the memhotplug version of free_area_init_core():

Currently, we call free_area_init_node() from the memhotplug path.
In there, we set some pgdat's fields, and call calculate_node_totalpages().
calculate_node_totalpages() calculates the # of pages the node has.

Since the node is either new, or we are re-using it, the zones belonging to
this node should not have any pages, so there is no point to calculate this now.

Actually, we re-set these values to 0 later on with the calls to:

reset_node_managed_pages()
reset_node_present_pages()

The # of pages per node and the # of pages per zone will be calculated when
onlining the pages:

online_pages()->move_pfn_range()->move_pfn_range_to_zone()->resize_zone_range()
online_pages()->move_pfn_range()->move_pfn_range_to_zone()->resize_pgdat_range()

Also, since free_area_init_core/free_area_init_node will now only get called during early init, let us replace
__paginginit with __init, so their code gets freed up.

Signed-off-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Pavel Tatashin <pasha.tatashin@oracle.com>
---
 include/linux/mm.h  |  6 ++++-
 mm/memory_hotplug.c | 16 ++++--------
 mm/page_alloc.c     | 71 +++++++++++++++++++++++++++++++++++++----------------
 3 files changed, 60 insertions(+), 33 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 6954ad183159..af3222785347 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1998,10 +1998,14 @@ static inline spinlock_t *pud_lock(struct mm_struct *mm, pud_t *pud)
 
 extern void __init pagecache_init(void);
 extern void free_area_init(unsigned long * zones_size);
-extern void free_area_init_node(int nid, unsigned long * zones_size,
+extern void __init free_area_init_node(int nid, unsigned long * zones_size,
 		unsigned long zone_start_pfn, unsigned long *zholes_size);
 extern void free_initmem(void);
 
+#ifdef CONFIG_MEMORY_HOTPLUG
+extern void __ref free_area_init_core_hotplug(int nid);
+#endif
+
 /*
  * Free reserved pages within range [PAGE_ALIGN(start), end & PAGE_MASK)
  * into the buddy system. The freed pages will be poisoned with pattern
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 4eb6e824a80c..9eea6e809a4e 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -982,8 +982,6 @@ static void reset_node_present_pages(pg_data_t *pgdat)
 static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
 {
 	struct pglist_data *pgdat;
-	unsigned long zones_size[MAX_NR_ZONES] = {0};
-	unsigned long zholes_size[MAX_NR_ZONES] = {0};
 	unsigned long start_pfn = PFN_DOWN(start);
 
 	pgdat = NODE_DATA(nid);
@@ -1006,8 +1004,11 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
 
 	/* we can use NODE_DATA(nid) from here */
 
+	pgdat->node_id = nid;
+	pgdat->node_start_pfn = start_pfn;
+
 	/* init node's zones as empty zones, we don't have any present pages.*/
-	free_area_init_node(nid, zones_size, start_pfn, zholes_size);
+	free_area_init_core_hotplug(nid);
 	pgdat->per_cpu_nodestats = alloc_percpu(struct per_cpu_nodestat);
 
 	/*
@@ -1017,18 +1018,11 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
 	build_all_zonelists(pgdat);
 
 	/*
-	 * 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);
-
-	/*
 	 * 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_managed_pages(pgdat);
 	reset_node_present_pages(pgdat);
 
 	return pgdat;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 4e84a17a5030..b2ccade42020 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6237,21 +6237,9 @@ static void pgdat_init_kcompactd(struct pglist_data *pgdat)
 static void pgdat_init_kcompactd(struct pglist_data *pgdat) {}
 #endif
 
-/*
- * Set up the zone data structures:
- *   - mark all pages reserved
- *   - mark all memory queues empty
- *   - clear the memory bitmaps
- *
- * NOTE: pgdat should get zeroed by caller.
- */
-static void __paginginit free_area_init_core(struct pglist_data *pgdat)
+static void __paginginit pgdat_init_internals(struct pglist_data *pgdat)
 {
-	enum zone_type j;
-	int nid = pgdat->node_id;
-
 	pgdat_resize_init(pgdat);
-
 	pgdat_init_numabalancing(pgdat);
 	pgdat_init_split_queue(pgdat);
 	pgdat_init_kcompactd(pgdat);
@@ -6262,7 +6250,54 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
 	pgdat_page_ext_init(pgdat);
 	spin_lock_init(&pgdat->lru_lock);
 	lruvec_init(node_lruvec(pgdat));
+}
+
+static void __paginginit zone_init_internals(struct zone *zone, enum zone_type idx, int nid,
+							unsigned long remaining_pages)
+{
+	zone->managed_pages = remaining_pages;
+	zone_set_nid(zone, nid);
+	zone->name = zone_names[idx];
+	zone->zone_pgdat = NODE_DATA(nid);
+	spin_lock_init(&zone->lock);
+	zone_seqlock_init(zone);
+	zone_pcp_init(zone);
+}
+
+/*
+ * Set up the zone data structures
+ * - init pgdat internals
+ * - init all zones belonging to this node
+ *
+ * NOTE: this function is only called during memory hotplug
+ */
+#ifdef CONFIG_MEMORY_HOTPLUG
+void __ref free_area_init_core_hotplug(int nid)
+{
+	enum zone_type z;
+	pg_data_t *pgdat = NODE_DATA(nid);
+
+	pgdat_init_internals(pgdat);
+	for (z = 0; z < MAX_NR_ZONES; z++)
+		zone_init_internals(&pgdat->node_zones[z], z, nid, 0);
+}
+#endif
+
+/*
+ * Set up the zone data structures:
+ *   - mark all pages reserved
+ *   - mark all memory queues empty
+ *   - clear the memory bitmaps
+ *
+ * NOTE: pgdat should get zeroed by caller.
+ * NOTE: this function is only called during early init.
+ */
+static void __init free_area_init_core(struct pglist_data *pgdat)
+{
+	enum zone_type j;
+	int nid = pgdat->node_id;
 
+	pgdat_init_internals(pgdat);
 	pgdat->per_cpu_nodestats = &boot_nodestats;
 
 	for (j = 0; j < MAX_NR_ZONES; j++) {
@@ -6310,13 +6345,7 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
 		 * when the bootmem allocator frees pages into the buddy system.
 		 * And all highmem pages will be managed by the buddy system.
 		 */
-		zone->managed_pages = freesize;
-		zone_set_nid(zone, nid);
-		zone->name = zone_names[j];
-		zone->zone_pgdat = pgdat;
-		spin_lock_init(&zone->lock);
-		zone_seqlock_init(zone);
-		zone_pcp_init(zone);
+		zone_init_internals(zone, j, nid, freesize);
 
 		if (!size)
 			continue;
@@ -6391,7 +6420,7 @@ static inline void pgdat_set_deferred_range(pg_data_t *pgdat)
 static inline void pgdat_set_deferred_range(pg_data_t *pgdat) {}
 #endif
 
-void __paginginit free_area_init_node(int nid, unsigned long *zones_size,
+void __init free_area_init_node(int nid, unsigned long *zones_size,
 		unsigned long node_start_pfn, unsigned long *zholes_size)
 {
 	pg_data_t *pgdat = NODE_DATA(nid);
-- 
2.13.6

Re: [PATCH v5 4/4] mm/page_alloc: Introduce free_area_init_core_hotplug

[Oscar Salvador]

On Mon, Jul 30, 2018 at 12:17:57PM +0200, osalvador@techadventures.net wrote:
> From: Oscar Salvador <osalvador@suse.de>
...
> Also, since free_area_init_core/free_area_init_node will now only get called during early init, let us replace
> __paginginit with __init, so their code gets freed up.
> 
> Signed-off-by: Oscar Salvador <osalvador@suse.de>
> Reviewed-by: Pavel Tatashin <pasha.tatashin@oracle.com>

Andrew, could you please fold the following cleanup into this patch?
thanks

Pavel, since this has your Reviewed-by, are you ok with the following on top?

set_pageblock_order() is only called from free_area_init_core() and sparse_init().
sparse_init() is only called during early init, and the same applies for free_area_init_core()
from now on (with this patchset)

The same goes for calc_memmap_size().

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index bb11cc23b862..c1cf088607c5 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6142,7 +6142,7 @@ static inline void setup_usemap(struct pglist_data *pgdat, struct zone *zone,
 #ifdef CONFIG_HUGETLB_PAGE_SIZE_VARIABLE
 
 /* Initialise the number of pages represented by NR_PAGEBLOCK_BITS */
-void __paginginit set_pageblock_order(void)
+void __init set_pageblock_order(void)
 {
 	unsigned int order;
 
@@ -6170,13 +6170,13 @@ void __paginginit set_pageblock_order(void)
  * include/linux/pageblock-flags.h for the values of pageblock_order based on
  * the kernel config
  */
-void __paginginit set_pageblock_order(void)
+void __init set_pageblock_order(void)
 {
 }
 
 #endif /* CONFIG_HUGETLB_PAGE_SIZE_VARIABLE */
 
-static unsigned long __paginginit calc_memmap_size(unsigned long spanned_pages,
+static unsigned long __init calc_memmap_size(unsigned long spanned_pages,
 						   unsigned long present_pages)
 {
 	unsigned long pages = spanned_pages;
@@ -6448,7 +6448,7 @@ void __init free_area_init_node(int nid, unsigned long *zones_size,
  * may be accessed (for example page_to_pfn() on some configuration accesses
  * flags). We must explicitly zero those struct pages.
  */
-void __paginginit zero_resv_unavail(void)
+void __init zero_resv_unavail(void)
 {
 	phys_addr_t start, end;
 	unsigned long pfn;

Thanks
-- 
Oscar Salvador
SUSE L3

Re: [PATCH v5 4/4] mm/page_alloc: Introduce free_area_init_core_hotplug

[Pavel Tatashin]

On Tue, Jul 31, 2018 at 6:17 AM Oscar Salvador
<osalvador@techadventures.net> wrote:
>
> On Mon, Jul 30, 2018 at 12:17:57PM +0200, osalvador@techadventures.net wrote:
> > From: Oscar Salvador <osalvador@suse.de>
> ...
> > Also, since free_area_init_core/free_area_init_node will now only get called during early init, let us replace
> > __paginginit with __init, so their code gets freed up.
> >
> > Signed-off-by: Oscar Salvador <osalvador@suse.de>
> > Reviewed-by: Pavel Tatashin <pasha.tatashin@oracle.com>
>
> Andrew, could you please fold the following cleanup into this patch?
> thanks
>
> Pavel, since this has your Reviewed-by, are you ok with the following on top?

Yes, Looks good to me.

Thank you,
Pavel

Re: [PATCH v5 4/4] mm/page_alloc: Introduce free_area_init_core_hotplug

[Oscar Salvador]

On Tue, Jul 31, 2018 at 12:17:52PM +0200, Oscar Salvador wrote:
> On Mon, Jul 30, 2018 at 12:17:57PM +0200, osalvador@techadventures.net wrote:
> > From: Oscar Salvador <osalvador@suse.de>
> ...
> > Also, since free_area_init_core/free_area_init_node will now only get called during early init, let us replace
> > __paginginit with __init, so their code gets freed up.
> > 
> > Signed-off-by: Oscar Salvador <osalvador@suse.de>
> > Reviewed-by: Pavel Tatashin <pasha.tatashin@oracle.com>
> 
> Andrew, could you please fold the following cleanup into this patch?
> thanks

Hi Andrew,

I sent v6, which already includes that cleanup-fixup plus another patch from Pavel and
an Acked-by from Michal Hocko.
So if it looks fine to you, feel free to replace it with the version that is sitting right now in -mm (v5)
 
Thanks
-- 
Oscar Salvador
SUSE L3

Re: [PATCH v5 4/4] mm/page_alloc: Introduce free_area_init_core_hotplug

[Michal Hocko]

On Mon 30-07-18 12:17:57, osalvador@techadventures.net wrote:
> From: Oscar Salvador <osalvador@suse.de>
> 
> Currently, whenever a new node is created/re-used from the memhotplug path,
> we call free_area_init_node()->free_area_init_core().
> But there is some code that we do not really need to run when we are coming
> from such path.
> 
> free_area_init_core() performs the following actions:
> 
> 1) Initializes pgdat internals, such as spinlock, waitqueues and more.
> 2) Account # nr_all_pages and # nr_kernel_pages. These values are used later on
>    when creating hash tables.
> 3) Account number of managed_pages per zone, substracting dma_reserved and memmap pages.
> 4) Initializes some fields of the zone structure data
> 5) Calls init_currently_empty_zone to initialize all the freelists
> 6) Calls memmap_init to initialize all pages belonging to certain zone
> 
> When called from memhotplug path, free_area_init_core() only performs actions #1 and #4.
> 
> Action #2 is pointless as the zones do not have any pages since either the node was freed,
> or we are re-using it, eitherway all zones belonging to this node should have 0 pages.
> For the same reason, action #3 results always in manages_pages being 0.
> 
> Action #5 and #6 are performed later on when onlining the pages:
>  online_pages()->move_pfn_range_to_zone()->init_currently_empty_zone()
>  online_pages()->move_pfn_range_to_zone()->memmap_init_zone()
> 
> This patch does two things:
> 
> First, moves the node/zone initializtion to their own function, so it allows us
> to create a small version of free_area_init_core, where we only perform:
> 
> 1) Initialization of pgdat internals, such as spinlock, waitqueues and more
> 4) Initialization of some fields of the zone structure data
> 
> These two functions are: pgdat_init_internals() and zone_init_internals().
> 
> The second thing this patch does, is to introduce free_area_init_core_hotplug(),
> the memhotplug version of free_area_init_core():
> 
> Currently, we call free_area_init_node() from the memhotplug path.
> In there, we set some pgdat's fields, and call calculate_node_totalpages().
> calculate_node_totalpages() calculates the # of pages the node has.
> 
> Since the node is either new, or we are re-using it, the zones belonging to
> this node should not have any pages, so there is no point to calculate this now.
> 
> Actually, we re-set these values to 0 later on with the calls to:
> 
> reset_node_managed_pages()
> reset_node_present_pages()
> 
> The # of pages per node and the # of pages per zone will be calculated when
> onlining the pages:
> 
> online_pages()->move_pfn_range()->move_pfn_range_to_zone()->resize_zone_range()
> online_pages()->move_pfn_range()->move_pfn_range_to_zone()->resize_pgdat_range()
> 
> Also, since free_area_init_core/free_area_init_node will now only get called during early init, let us replace
> __paginginit with __init, so their code gets freed up.

The split up makes sense to me. Sections attributes can be handled on
top. Btw. free_area_init_core_hotplug declaration could have gone into
include/linux/memory_hotplug.h to save the ifdef

> Signed-off-by: Oscar Salvador <osalvador@suse.de>
> Reviewed-by: Pavel Tatashin <pasha.tatashin@oracle.com>

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

> ---
>  include/linux/mm.h  |  6 ++++-
>  mm/memory_hotplug.c | 16 ++++--------
>  mm/page_alloc.c     | 71 +++++++++++++++++++++++++++++++++++++----------------
>  3 files changed, 60 insertions(+), 33 deletions(-)
> 
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 6954ad183159..af3222785347 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -1998,10 +1998,14 @@ static inline spinlock_t *pud_lock(struct mm_struct *mm, pud_t *pud)
>  
>  extern void __init pagecache_init(void);
>  extern void free_area_init(unsigned long * zones_size);
> -extern void free_area_init_node(int nid, unsigned long * zones_size,
> +extern void __init free_area_init_node(int nid, unsigned long * zones_size,
>  		unsigned long zone_start_pfn, unsigned long *zholes_size);
>  extern void free_initmem(void);
>  
> +#ifdef CONFIG_MEMORY_HOTPLUG
> +extern void __ref free_area_init_core_hotplug(int nid);
> +#endif
> +
>  /*
>   * Free reserved pages within range [PAGE_ALIGN(start), end & PAGE_MASK)
>   * into the buddy system. The freed pages will be poisoned with pattern
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index 4eb6e824a80c..9eea6e809a4e 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -982,8 +982,6 @@ static void reset_node_present_pages(pg_data_t *pgdat)
>  static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
>  {
>  	struct pglist_data *pgdat;
> -	unsigned long zones_size[MAX_NR_ZONES] = {0};
> -	unsigned long zholes_size[MAX_NR_ZONES] = {0};
>  	unsigned long start_pfn = PFN_DOWN(start);
>  
>  	pgdat = NODE_DATA(nid);
> @@ -1006,8 +1004,11 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
>  
>  	/* we can use NODE_DATA(nid) from here */
>  
> +	pgdat->node_id = nid;
> +	pgdat->node_start_pfn = start_pfn;
> +
>  	/* init node's zones as empty zones, we don't have any present pages.*/
> -	free_area_init_node(nid, zones_size, start_pfn, zholes_size);
> +	free_area_init_core_hotplug(nid);
>  	pgdat->per_cpu_nodestats = alloc_percpu(struct per_cpu_nodestat);
>  
>  	/*
> @@ -1017,18 +1018,11 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
>  	build_all_zonelists(pgdat);
>  
>  	/*
> -	 * 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);
> -
> -	/*
>  	 * 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_managed_pages(pgdat);
>  	reset_node_present_pages(pgdat);
>  
>  	return pgdat;
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 4e84a17a5030..b2ccade42020 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -6237,21 +6237,9 @@ static void pgdat_init_kcompactd(struct pglist_data *pgdat)
>  static void pgdat_init_kcompactd(struct pglist_data *pgdat) {}
>  #endif
>  
> -/*
> - * Set up the zone data structures:
> - *   - mark all pages reserved
> - *   - mark all memory queues empty
> - *   - clear the memory bitmaps
> - *
> - * NOTE: pgdat should get zeroed by caller.
> - */
> -static void __paginginit free_area_init_core(struct pglist_data *pgdat)
> +static void __paginginit pgdat_init_internals(struct pglist_data *pgdat)
>  {
> -	enum zone_type j;
> -	int nid = pgdat->node_id;
> -
>  	pgdat_resize_init(pgdat);
> -
>  	pgdat_init_numabalancing(pgdat);
>  	pgdat_init_split_queue(pgdat);
>  	pgdat_init_kcompactd(pgdat);
> @@ -6262,7 +6250,54 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
>  	pgdat_page_ext_init(pgdat);
>  	spin_lock_init(&pgdat->lru_lock);
>  	lruvec_init(node_lruvec(pgdat));
> +}
> +
> +static void __paginginit zone_init_internals(struct zone *zone, enum zone_type idx, int nid,
> +							unsigned long remaining_pages)
> +{
> +	zone->managed_pages = remaining_pages;
> +	zone_set_nid(zone, nid);
> +	zone->name = zone_names[idx];
> +	zone->zone_pgdat = NODE_DATA(nid);
> +	spin_lock_init(&zone->lock);
> +	zone_seqlock_init(zone);
> +	zone_pcp_init(zone);
> +}
> +
> +/*
> + * Set up the zone data structures
> + * - init pgdat internals
> + * - init all zones belonging to this node
> + *
> + * NOTE: this function is only called during memory hotplug
> + */
> +#ifdef CONFIG_MEMORY_HOTPLUG
> +void __ref free_area_init_core_hotplug(int nid)
> +{
> +	enum zone_type z;
> +	pg_data_t *pgdat = NODE_DATA(nid);
> +
> +	pgdat_init_internals(pgdat);
> +	for (z = 0; z < MAX_NR_ZONES; z++)
> +		zone_init_internals(&pgdat->node_zones[z], z, nid, 0);
> +}
> +#endif
> +
> +/*
> + * Set up the zone data structures:
> + *   - mark all pages reserved
> + *   - mark all memory queues empty
> + *   - clear the memory bitmaps
> + *
> + * NOTE: pgdat should get zeroed by caller.
> + * NOTE: this function is only called during early init.
> + */
> +static void __init free_area_init_core(struct pglist_data *pgdat)
> +{
> +	enum zone_type j;
> +	int nid = pgdat->node_id;
>  
> +	pgdat_init_internals(pgdat);
>  	pgdat->per_cpu_nodestats = &boot_nodestats;
>  
>  	for (j = 0; j < MAX_NR_ZONES; j++) {
> @@ -6310,13 +6345,7 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
>  		 * when the bootmem allocator frees pages into the buddy system.
>  		 * And all highmem pages will be managed by the buddy system.
>  		 */
> -		zone->managed_pages = freesize;
> -		zone_set_nid(zone, nid);
> -		zone->name = zone_names[j];
> -		zone->zone_pgdat = pgdat;
> -		spin_lock_init(&zone->lock);
> -		zone_seqlock_init(zone);
> -		zone_pcp_init(zone);
> +		zone_init_internals(zone, j, nid, freesize);
>  
>  		if (!size)
>  			continue;
> @@ -6391,7 +6420,7 @@ static inline void pgdat_set_deferred_range(pg_data_t *pgdat)
>  static inline void pgdat_set_deferred_range(pg_data_t *pgdat) {}
>  #endif
>  
> -void __paginginit free_area_init_node(int nid, unsigned long *zones_size,
> +void __init free_area_init_node(int nid, unsigned long *zones_size,
>  		unsigned long node_start_pfn, unsigned long *zholes_size)
>  {
>  	pg_data_t *pgdat = NODE_DATA(nid);
> -- 
> 2.13.6
> 

-- 
Michal Hocko
SUSE Labs

Re: [PATCH v5 4/4] mm/page_alloc: Introduce free_area_init_core_hotplug

[Oscar Salvador]

On Wed, Aug 01, 2018 at 01:47:26PM +0200, Michal Hocko wrote:
> 
> The split up makes sense to me. Sections attributes can be handled on
> top. Btw. free_area_init_core_hotplug declaration could have gone into
> include/linux/memory_hotplug.h to save the ifdef

You are right, I will fix this up.

> 
> > Signed-off-by: Oscar Salvador <osalvador@suse.de>
> > Reviewed-by: Pavel Tatashin <pasha.tatashin@oracle.com>
> 
> Acked-by: Michal Hocko <mhocko@suse.com>

Thanks Michal!

Since Pavel and I agreed on putting a patch of his which removes __paginginit
into this patchset, I will send out a v6 in a few minutes.

-- 
Oscar Salvador
SUSE L3