postgres.git / summary / log / commit / refs
commit e4e496e88c9a0940abe1ade44337577b26f032a0
Author: Michael Paquier <michael@paquier.xyz>
Date: Tue Oct 21 07:12:30 2025 +0000
First released in: 19.0, 18.1
Fix comment in pg_get_shmem_allocations_numa()
The comment fixed in this commit described the function as dealing with
database blocks, but in reality it processes shared memory allocations.
Author: Bertrand Drouvot <bertranddrouvot.pg@gmail.com>
Discussion: https://postgr.es/m/aH4DDhdiG9Gi0rG7@ip-10-97-1-34.eu-west-3.compute.internal
Backpatch-through: 18
src/backend/storage/ipc/shmem.c | 15 ++++++---------
1 file changed, 6 insertions(+), 9 deletions(-)
diff --git a/src/backend/storage/ipc/shmem.c b/src/backend/storage/ipc/shmem.c
index a0770e86796..b3d0a10792d 100644
--- a/src/backend/storage/ipc/shmem.c
+++ b/src/backend/storage/ipc/shmem.c
@@ -606,16 +606,13 @@ pg_get_shmem_allocations_numa(PG_FUNCTION_ARGS)
nodes = palloc(sizeof(Size) * (max_nodes + 1));
/*
- * Different database block sizes (4kB, 8kB, ..., 32kB) can be used, while
- * the OS may have different memory page sizes.
+ * Shared memory allocations can vary in size and may not align with OS
+ * memory page boundaries, while NUMA queries work on pages.
*
- * To correctly map between them, we need to: 1. Determine the OS memory
- * page size 2. Calculate how many OS pages are used by all buffer blocks
- * 3. Calculate how many OS pages are contained within each database
- * block.
- *
- * This information is needed before calling move_pages() for NUMA memory
- * node inquiry.
+ * To correctly map each allocation to NUMA nodes, we need to: 1.
+ * Determine the OS memory page size. 2. Align each allocation's start/end
+ * addresses to page boundaries. 3. Query NUMA node information for all
+ * pages spanning the allocation.
*/
os_page_size = pg_get_shmem_pagesize();
[parent: 18d261409348]