mm: avoid unconditional one-tick sleep when swapcache_prepare fails

Commit 13ddaf26be ("mm/swap: fix race when skipping swapcache")
introduced an unconditional one-tick sleep when `swapcache_prepare()`
fails, which has led to reports of UI stuttering on latency-sensitive
Android devices.  To address this, we can use a waitqueue to wake up tasks
that fail `swapcache_prepare()` sooner, instead of always sleeping for a
full tick.  While tasks may occasionally be woken by an unrelated
`do_swap_page()`, this method is preferable to two scenarios: rapid
re-entry into page faults, which can cause livelocks, and multiple
millisecond sleeps, which visibly degrade user experience.

Oven's testing shows that a single waitqueue resolves the UI stuttering
issue.  If a 'thundering herd' problem becomes apparent later, a waitqueue
hash similar to `folio_wait_table[PAGE_WAIT_TABLE_SIZE]` for page bit
locks can be introduced.

[[email protected]: wake_up only when swapcache_wq waitqueue is active]
  Link: https://lkml.kernel.org/r/[email protected]
Link: https://lkml.kernel.org/r/[email protected]
Fixes: 13ddaf26be ("mm/swap: fix race when skipping swapcache")
Signed-off-by: Barry Song <[email protected]>
Reported-by: Oven Liyang <[email protected]>
Tested-by: Oven Liyang <[email protected]>
Cc: Kairui Song <[email protected]>
Cc: "Huang, Ying" <[email protected]>
Cc: Yu Zhao <[email protected]>
Cc: David Hildenbrand <[email protected]>
Cc: Chris Li <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Matthew Wilcox (Oracle) <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Yosry Ahmed <[email protected]>
Cc: SeongJae Park <[email protected]>
Cc: Kalesh Singh <[email protected]>
Cc: Suren Baghdasaryan <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>

mm/memory.c

@@ -4187,6 +4187,8 @@ static struct folio *alloc_swap_folio(struct vm_fault *vmf)
 }
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
 
+static DECLARE_WAIT_QUEUE_HEAD(swapcache_wq);
+
 /*
  * We enter with non-exclusive mmap_lock (to exclude vma changes,
  * but allow concurrent faults), and pte mapped but not yet locked.
@@ -4199,6 +4201,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 {
 	struct vm_area_struct *vma = vmf->vma;
 	struct folio *swapcache, *folio = NULL;
+	DECLARE_WAITQUEUE(wait, current);
 	struct page *page;
 	struct swap_info_struct *si = NULL;
 	rmap_t rmap_flags = RMAP_NONE;
@@ -4297,7 +4300,9 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 					 * Relax a bit to prevent rapid
 					 * repeated page faults.
 					 */
+					add_wait_queue(&swapcache_wq, &wait);
 					schedule_timeout_uninterruptible(1);
+					remove_wait_queue(&swapcache_wq, &wait);
 					goto out_page;
 				}
 				need_clear_cache = true;
@@ -4604,8 +4609,11 @@ unlock:
 		pte_unmap_unlock(vmf->pte, vmf->ptl);
 out:
 	/* Clear the swap cache pin for direct swapin after PTL unlock */
-	if (need_clear_cache)
+	if (need_clear_cache) {
 		swapcache_clear(si, entry, nr_pages);
+		if (waitqueue_active(&swapcache_wq))
+			wake_up(&swapcache_wq);
+	}
 	if (si)
 		put_swap_device(si);
 	return ret;
@@ -4620,8 +4628,11 @@ out_release:
 		folio_unlock(swapcache);
 		folio_put(swapcache);
 	}
-	if (need_clear_cache)
+	if (need_clear_cache) {
 		swapcache_clear(si, entry, nr_pages);
+		if (waitqueue_active(&swapcache_wq))
+			wake_up(&swapcache_wq);
+	}
 	if (si)
 		put_swap_device(si);
 	return ret;