diff mbox series

[RFC,v2,21/21] mm/vmscan.c: shrink anon list if can migrate to PMEM

Message ID 20181226133352.303666865@intel.com (mailing list archive)
State New, archived
Headers show
Series PMEM NUMA node and hotness accounting/migration | expand

Commit Message

Fengguang Wu Dec. 26, 2018, 1:15 p.m. UTC 
Fix OOM by making in-kernel DRAM=>PMEM migration reachable.

Here we assume these 2 possible demotion paths:
- DRAM migrate to PMEM
- PMEM to swap device

Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
---
 mm/vmscan.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
diff mbox series

Patch

--- linux.orig/mm/vmscan.c	2018-12-23 20:38:44.310446223 +0800
+++ linux/mm/vmscan.c	2018-12-23 20:38:44.306446146 +0800
@@ -2259,7 +2259,7 @@  static bool inactive_list_is_low(struct
 	 * If we don't have swap space, anonymous page deactivation
 	 * is pointless.
 	 */
-	if (!file && !total_swap_pages)
+	if (!file && (is_node_pmem(pgdat->node_id) && !total_swap_pages))
 		return false;
 
 	inactive = lruvec_lru_size(lruvec, inactive_lru, sc->reclaim_idx);
@@ -2340,7 +2340,8 @@  static void get_scan_count(struct lruvec
 	enum lru_list lru;
 
 	/* If we have no swap space, do not bother scanning anon pages. */
-	if (!sc->may_swap || mem_cgroup_get_nr_swap_pages(memcg) <= 0) {
+	if (is_node_pmem(pgdat->node_id) &&
+	    (!sc->may_swap || mem_cgroup_get_nr_swap_pages(memcg) <= 0)) {
 		scan_balance = SCAN_FILE;
 		goto out;
 	}