The invocation of zone_init_internals(zone, j, nid, zone->present_pages)
initializes zone->managed_pages to zone->present_pages. zone->present_pages
is not 0 when the zone is not empty.
See commit 0ac5e785dcb797 ("mm/mm_init.c: remove meaningless calculation
of zone->managed_pages in free_area_init_core()") for details.
Signed-off-by: Jiwen Qi <jiwen7.qi@xxxxxxxxx>
---
mm/mm_init.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/mm/mm_init.c b/mm/mm_init.c
index 24b68b425afb..48a4f661db98 100644
--- a/mm/mm_init.c
+++ b/mm/mm_init.c
@@ -1567,8 +1567,9 @@ static void __init free_area_init_core(struct pglist_data *pgdat)
unsigned long size = zone->spanned_pages;
/*
- * Initialize zone->managed_pages as 0 , it will be reset
- * when memblock allocator frees pages into buddy system.
+ * Initialize zone->managed_pages to zone->present_pages as a first rough
+ * estimate. memblock_free_all() will reset zone->managed_pages to 0, and
+ * calculate the actual managed pages as they are freed to the buddy.
*/
zone_init_internals(zone, j, nid, zone->present_pages);
base-commit: 4bbf9020becbfd8fc2c3da790855b7042fad455b
--
2.25.1
