[PATCH v2 1/2] mm/swapfile: fix list iteration when next node is removed during discard

[Youngjun Park <youngjun.park@lge.com>]

When the next node is removed from the plist (e.g. by swapoff),
plist_del() makes the node point to itself, causing the iteration to
loop on the same entry indefinitely.

Add a plist_node_empty() check to detect this case and restart
iteration, allowing swap_sync_discard() to continue processing
remaining swap devices that still have pending discard entries.

Additionally, switch from swap_avail_lock/swap_avail_head to
swap_lock/swap_active_head so that iteration is only affected by
swapoff operations rather than frequent availability changes,
reducing exceptional condition checks and lock contention.

Fixes: 686ea517f471 ("mm, swap: do not perform synchronous discard during allocation")
Suggested-by: Kairui Song <kasong@tencent.com>
Acked-by: Kairui Song <kasong@tencent.com>
Signed-off-by: Youngjun Park <youngjun.park@lge.com>
---
 mm/swapfile.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/mm/swapfile.c b/mm/swapfile.c
index d12332423a06..8116f36e440b 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -1387,9 +1387,10 @@ static bool swap_sync_discard(void)
 	bool ret = false;
 	struct swap_info_struct *si, *next;
 
-	spin_lock(&swap_avail_lock);
-	plist_for_each_entry_safe(si, next, &swap_avail_head, avail_list) {
-		spin_unlock(&swap_avail_lock);
+	spin_lock(&swap_lock);
+start_over:
+	plist_for_each_entry_safe(si, next, &swap_active_head, list) {
+		spin_unlock(&swap_lock);
 		if (get_swap_device_info(si)) {
 			if (si->flags & SWP_PAGE_DISCARD)
 				ret = swap_do_scheduled_discard(si);
@@ -1397,9 +1398,12 @@ static bool swap_sync_discard(void)
 		}
 		if (ret)
 			return true;
-		spin_lock(&swap_avail_lock);
+
+		spin_lock(&swap_lock);
+		if (plist_node_empty(&next->list))
+			goto start_over;
 	}
-	spin_unlock(&swap_avail_lock);
+	spin_unlock(&swap_lock);
 
 	return false;
 }
-- 
2.34.1