[PATCH 00/22] Convert the swap code to be more folio-based

[PATCH 00/22] Convert the swap code to be more folio-based

[Matthew Wilcox (Oracle)]

(resend due to email address mistake)

There's still more to do with the swap code, but this reaps a lot of
the folio benefit.  More than 4kB of kernel text saved (with the UEK7
kernel config).  I don't know how much that's going to translate into
CPU savings, but some of those compound_head() calls are on every page
free, so it should be noticable.  It might even be noticable just from
an I-cache consumption perspective.

This is against next-20220617 plus my prior vmscan patch series.  It's in
the middle of an xfstests run right now, so it at least boots.

Matthew Wilcox (Oracle) (22):
  mm: Add folios_put()
  mm/swap: Add folio_batch_move_lru()
  mm/swap: Make __pagevec_lru_add static
  mm/swap: Convert lru_add to a folio_batch
  mm/swap: Convert lru_deactivate_file to a folio_batch
  mm/swap: Convert lru_deactivate to a folio_batch
  mm/swap: Convert lru_lazyfree to a folio_batch
  mm/swap: Convert activate_page to a folio_batch
  mm/swap: Rename lru_pvecs to cpu_fbatches
  mm/swap: Pull the CPU conditional out of __lru_add_drain_all()
  mm/swap: Optimise lru_add_drain_cpu()
  mm/swap: Convert try_to_free_swap to use a folio
  mm/swap: Convert release_pages to use a folio internally
  mm/swap: Convert put_pages_list to use folios
  mm/swap: Convert __put_page() to __folio_put()
  mm/swap: Convert __put_single_page() to __folio_put_small()
  mm/swap: Convert __put_compound_page() to __folio_put_large()
  mm/swap: Convert __page_cache_release() to use a folio
  mm: Convert destroy_compound_page() to destroy_large_folio()
  mm: Convert page_swap_flags to folio_swap_flags
  mm/swap: Convert delete_from_swap_cache() to take a folio
  mm/swap: Convert __delete_from_swap_cache() to a folio

 include/linux/mm.h      |  33 ++-
 include/linux/pagemap.h |   2 -
 include/linux/pagevec.h |   1 -
 mm/memory-failure.c     |   5 +-
 mm/shmem.c              |   4 +-
 mm/swap.c               | 599 +++++++++++++++++++---------------------
 mm/swap.h               |  14 +-
 mm/swap_state.c         |  39 +--
 mm/swapfile.c           |  23 +-
 mm/vmscan.c             |   8 +-
 net/core/page_pool.c    |   2 +-
 11 files changed, 359 insertions(+), 371 deletions(-)

-- 
2.35.1

[PATCH 01/22] mm: Add folios_put()

[Matthew Wilcox (Oracle)]

This is just a wrapper around release_pages() for now.  Place the
prototype in mm.h along with folio_put() and folio_put_refs().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 include/linux/mm.h      | 19 +++++++++++++++++++
 include/linux/pagemap.h |  2 --
 2 files changed, 19 insertions(+), 2 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 82b680261cf0..07a1998c3621 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1251,6 +1251,25 @@ static inline void folio_put_refs(struct folio *folio, int refs)
 		__put_page(&folio->page);
 }
 
+void release_pages(struct page **pages, int nr);
+
+/**
+ * folios_put - Decrement the reference count on an array of folios.
+ * @folios: The folios.
+ * @nr: How many folios there are.
+ *
+ * Like folio_put(), but for an array of folios.  This is more efficient
+ * than writing the loop yourself as it will optimise the locks which
+ * need to be taken if the folios are freed.
+ *
+ * Context: May be called in process or interrupt context, but not in NMI
+ * context.  May be called while holding a spinlock.
+ */
+static inline void folios_put(struct folio **folios, unsigned int nr)
+{
+	release_pages((struct page **)folios, nr);
+}
+
 static inline void put_page(struct page *page)
 {
 	struct folio *folio = page_folio(page);
diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
index 2a67c0ad7348..7c4748a62841 100644
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -345,8 +345,6 @@ static inline void filemap_nr_thps_dec(struct address_space *mapping)
 #endif
 }
 
-void release_pages(struct page **pages, int nr);
-
 struct address_space *page_mapping(struct page *);
 struct address_space *folio_mapping(struct folio *);
 struct address_space *swapcache_mapping(struct folio *);
-- 
2.35.1

[PATCH 02/22] mm/swap: Add folio_batch_move_lru()

[Matthew Wilcox (Oracle)]

Start converting the LRU from pagevecs to folio_batches.

Combine the functionality of pagevec_add_and_need_flush() with
pagevec_lru_move_fn() in the new folio_batch_add_and_move().

Convert the lru_rotate pagevec to a folio_batch.

Adds 223 bytes total to kernel text, because we're duplicating
infrastructure.  This will be more than made up for in future patches.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 78 +++++++++++++++++++++++++++++++++++++++----------------
 1 file changed, 56 insertions(+), 22 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index 275a4ea1bc66..a983a1b93e73 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -46,10 +46,10 @@
 /* How many pages do we try to swap or page in/out together? */
 int page_cluster;
 
-/* Protecting only lru_rotate.pvec which requires disabling interrupts */
+/* Protecting only lru_rotate.fbatch which requires disabling interrupts */
 struct lru_rotate {
 	local_lock_t lock;
-	struct pagevec pvec;
+	struct folio_batch fbatch;
 };
 static DEFINE_PER_CPU(struct lru_rotate, lru_rotate) = {
 	.lock = INIT_LOCAL_LOCK(lock),
@@ -214,18 +214,6 @@ static void pagevec_lru_move_fn(struct pagevec *pvec,
 	pagevec_reinit(pvec);
 }
 
-static void pagevec_move_tail_fn(struct page *page, struct lruvec *lruvec)
-{
-	struct folio *folio = page_folio(page);
-
-	if (!folio_test_unevictable(folio)) {
-		lruvec_del_folio(lruvec, folio);
-		folio_clear_active(folio);
-		lruvec_add_folio_tail(lruvec, folio);
-		__count_vm_events(PGROTATED, folio_nr_pages(folio));
-	}
-}
-
 /* return true if pagevec needs to drain */
 static bool pagevec_add_and_need_flush(struct pagevec *pvec, struct page *page)
 {
@@ -238,6 +226,52 @@ static bool pagevec_add_and_need_flush(struct pagevec *pvec, struct page *page)
 	return ret;
 }
 
+typedef void (*move_fn_t)(struct lruvec *lruvec, struct folio *folio);
+
+static void folio_batch_move_lru(struct folio_batch *fbatch, move_fn_t move_fn)
+{
+	int i;
+	struct lruvec *lruvec = NULL;
+	unsigned long flags = 0;
+
+	for (i = 0; i < folio_batch_count(fbatch); i++) {
+		struct folio *folio = fbatch->folios[i];
+
+		/* block memcg migration while the folio moves between lru */
+		if (!folio_test_clear_lru(folio))
+			continue;
+
+		lruvec = folio_lruvec_relock_irqsave(folio, lruvec, &flags);
+		move_fn(lruvec, folio);
+
+		folio_set_lru(folio);
+	}
+
+	if (lruvec)
+		unlock_page_lruvec_irqrestore(lruvec, flags);
+	folios_put(fbatch->folios, folio_batch_count(fbatch));
+	folio_batch_init(fbatch);
+}
+
+static void folio_batch_add_and_move(struct folio_batch *fbatch,
+		struct folio *folio, move_fn_t move_fn)
+{
+	if (folio_batch_add(fbatch, folio) && !folio_test_large(folio) &&
+	    !lru_cache_disabled())
+		return;
+	folio_batch_move_lru(fbatch, move_fn);
+}
+
+static void lru_move_tail_fn(struct lruvec *lruvec, struct folio *folio)
+{
+	if (!folio_test_unevictable(folio)) {
+		lruvec_del_folio(lruvec, folio);
+		folio_clear_active(folio);
+		lruvec_add_folio_tail(lruvec, folio);
+		__count_vm_events(PGROTATED, folio_nr_pages(folio));
+	}
+}
+
 /*
  * Writeback is about to end against a folio which has been marked for
  * immediate reclaim.  If it still appears to be reclaimable, move it
@@ -249,14 +283,13 @@ void folio_rotate_reclaimable(struct folio *folio)
 {
 	if (!folio_test_locked(folio) && !folio_test_dirty(folio) &&
 	    !folio_test_unevictable(folio) && folio_test_lru(folio)) {
-		struct pagevec *pvec;
+		struct folio_batch *fbatch;
 		unsigned long flags;
 
 		folio_get(folio);
 		local_lock_irqsave(&lru_rotate.lock, flags);
-		pvec = this_cpu_ptr(&lru_rotate.pvec);
-		if (pagevec_add_and_need_flush(pvec, &folio->page))
-			pagevec_lru_move_fn(pvec, pagevec_move_tail_fn);
+		fbatch = this_cpu_ptr(&lru_rotate.fbatch);
+		folio_batch_add_and_move(fbatch, folio, lru_move_tail_fn);
 		local_unlock_irqrestore(&lru_rotate.lock, flags);
 	}
 }
@@ -595,19 +628,20 @@ static void lru_lazyfree_fn(struct page *page, struct lruvec *lruvec)
  */
 void lru_add_drain_cpu(int cpu)
 {
+	struct folio_batch *fbatch;
 	struct pagevec *pvec = &per_cpu(lru_pvecs.lru_add, cpu);
 
 	if (pagevec_count(pvec))
 		__pagevec_lru_add(pvec);
 
-	pvec = &per_cpu(lru_rotate.pvec, cpu);
+	fbatch = &per_cpu(lru_rotate.fbatch, cpu);
 	/* Disabling interrupts below acts as a compiler barrier. */
-	if (data_race(pagevec_count(pvec))) {
+	if (data_race(folio_batch_count(fbatch))) {
 		unsigned long flags;
 
 		/* No harm done if a racing interrupt already did this */
 		local_lock_irqsave(&lru_rotate.lock, flags);
-		pagevec_lru_move_fn(pvec, pagevec_move_tail_fn);
+		folio_batch_move_lru(fbatch, lru_move_tail_fn);
 		local_unlock_irqrestore(&lru_rotate.lock, flags);
 	}
 
@@ -824,7 +858,7 @@ static inline void __lru_add_drain_all(bool force_all_cpus)
 		struct work_struct *work = &per_cpu(lru_add_drain_work, cpu);
 
 		if (pagevec_count(&per_cpu(lru_pvecs.lru_add, cpu)) ||
-		    data_race(pagevec_count(&per_cpu(lru_rotate.pvec, cpu))) ||
+		    data_race(folio_batch_count(&per_cpu(lru_rotate.fbatch, cpu))) ||
 		    pagevec_count(&per_cpu(lru_pvecs.lru_deactivate_file, cpu)) ||
 		    pagevec_count(&per_cpu(lru_pvecs.lru_deactivate, cpu)) ||
 		    pagevec_count(&per_cpu(lru_pvecs.lru_lazyfree, cpu)) ||
-- 
2.35.1

[PATCH 03/22] mm/swap: Make __pagevec_lru_add static

[Matthew Wilcox (Oracle)]

__pagevec_lru_add has no callers outside swap.c, so make it static,
and move it to a more logical position in the file.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 include/linux/pagevec.h |   1 -
 mm/swap.c               | 126 ++++++++++++++++++++--------------------
 2 files changed, 63 insertions(+), 64 deletions(-)

diff --git a/include/linux/pagevec.h b/include/linux/pagevec.h
index 6649154a2115..215eb6c3bdc9 100644
--- a/include/linux/pagevec.h
+++ b/include/linux/pagevec.h
@@ -26,7 +26,6 @@ struct pagevec {
 };
 
 void __pagevec_release(struct pagevec *pvec);
-void __pagevec_lru_add(struct pagevec *pvec);
 unsigned pagevec_lookup_range_tag(struct pagevec *pvec,
 		struct address_space *mapping, pgoff_t *index, pgoff_t end,
 		xa_mark_t tag);
diff --git a/mm/swap.c b/mm/swap.c
index a983a1b93e73..6b015096ef4a 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -228,6 +228,69 @@ static bool pagevec_add_and_need_flush(struct pagevec *pvec, struct page *page)
 
 typedef void (*move_fn_t)(struct lruvec *lruvec, struct folio *folio);
 
+static void __pagevec_lru_add_fn(struct folio *folio, struct lruvec *lruvec)
+{
+	int was_unevictable = folio_test_clear_unevictable(folio);
+	long nr_pages = folio_nr_pages(folio);
+
+	VM_BUG_ON_FOLIO(folio_test_lru(folio), folio);
+
+	folio_set_lru(folio);
+	/*
+	 * Is an smp_mb__after_atomic() still required here, before
+	 * folio_evictable() tests PageMlocked, to rule out the possibility
+	 * of stranding an evictable folio on an unevictable LRU?  I think
+	 * not, because __munlock_page() only clears PageMlocked while the LRU
+	 * lock is held.
+	 *
+	 * (That is not true of __page_cache_release(), and not necessarily
+	 * true of release_pages(): but those only clear PageMlocked after
+	 * put_page_testzero() has excluded any other users of the page.)
+	 */
+	if (folio_evictable(folio)) {
+		if (was_unevictable)
+			__count_vm_events(UNEVICTABLE_PGRESCUED, nr_pages);
+	} else {
+		folio_clear_active(folio);
+		folio_set_unevictable(folio);
+		/*
+		 * folio->mlock_count = !!folio_test_mlocked(folio)?
+		 * But that leaves __mlock_page() in doubt whether another
+		 * actor has already counted the mlock or not.  Err on the
+		 * safe side, underestimate, let page reclaim fix it, rather
+		 * than leaving a page on the unevictable LRU indefinitely.
+		 */
+		folio->mlock_count = 0;
+		if (!was_unevictable)
+			__count_vm_events(UNEVICTABLE_PGCULLED, nr_pages);
+	}
+
+	lruvec_add_folio(lruvec, folio);
+	trace_mm_lru_insertion(folio);
+}
+
+/*
+ * Add the passed pages to the LRU, then drop the caller's refcount
+ * on them.  Reinitialises the caller's pagevec.
+ */
+static void __pagevec_lru_add(struct pagevec *pvec)
+{
+	int i;
+	struct lruvec *lruvec = NULL;
+	unsigned long flags = 0;
+
+	for (i = 0; i < pagevec_count(pvec); i++) {
+		struct folio *folio = page_folio(pvec->pages[i]);
+
+		lruvec = folio_lruvec_relock_irqsave(folio, lruvec, &flags);
+		__pagevec_lru_add_fn(folio, lruvec);
+	}
+	if (lruvec)
+		unlock_page_lruvec_irqrestore(lruvec, flags);
+	release_pages(pvec->pages, pvec->nr);
+	pagevec_reinit(pvec);
+}
+
 static void folio_batch_move_lru(struct folio_batch *fbatch, move_fn_t move_fn)
 {
 	int i;
@@ -1036,69 +1099,6 @@ void __pagevec_release(struct pagevec *pvec)
 }
 EXPORT_SYMBOL(__pagevec_release);
 
-static void __pagevec_lru_add_fn(struct folio *folio, struct lruvec *lruvec)
-{
-	int was_unevictable = folio_test_clear_unevictable(folio);
-	long nr_pages = folio_nr_pages(folio);
-
-	VM_BUG_ON_FOLIO(folio_test_lru(folio), folio);
-
-	folio_set_lru(folio);
-	/*
-	 * Is an smp_mb__after_atomic() still required here, before
-	 * folio_evictable() tests PageMlocked, to rule out the possibility
-	 * of stranding an evictable folio on an unevictable LRU?  I think
-	 * not, because __munlock_page() only clears PageMlocked while the LRU
-	 * lock is held.
-	 *
-	 * (That is not true of __page_cache_release(), and not necessarily
-	 * true of release_pages(): but those only clear PageMlocked after
-	 * put_page_testzero() has excluded any other users of the page.)
-	 */
-	if (folio_evictable(folio)) {
-		if (was_unevictable)
-			__count_vm_events(UNEVICTABLE_PGRESCUED, nr_pages);
-	} else {
-		folio_clear_active(folio);
-		folio_set_unevictable(folio);
-		/*
-		 * folio->mlock_count = !!folio_test_mlocked(folio)?
-		 * But that leaves __mlock_page() in doubt whether another
-		 * actor has already counted the mlock or not.  Err on the
-		 * safe side, underestimate, let page reclaim fix it, rather
-		 * than leaving a page on the unevictable LRU indefinitely.
-		 */
-		folio->mlock_count = 0;
-		if (!was_unevictable)
-			__count_vm_events(UNEVICTABLE_PGCULLED, nr_pages);
-	}
-
-	lruvec_add_folio(lruvec, folio);
-	trace_mm_lru_insertion(folio);
-}
-
-/*
- * Add the passed pages to the LRU, then drop the caller's refcount
- * on them.  Reinitialises the caller's pagevec.
- */
-void __pagevec_lru_add(struct pagevec *pvec)
-{
-	int i;
-	struct lruvec *lruvec = NULL;
-	unsigned long flags = 0;
-
-	for (i = 0; i < pagevec_count(pvec); i++) {
-		struct folio *folio = page_folio(pvec->pages[i]);
-
-		lruvec = folio_lruvec_relock_irqsave(folio, lruvec, &flags);
-		__pagevec_lru_add_fn(folio, lruvec);
-	}
-	if (lruvec)
-		unlock_page_lruvec_irqrestore(lruvec, flags);
-	release_pages(pvec->pages, pvec->nr);
-	pagevec_reinit(pvec);
-}
-
 /**
  * folio_batch_remove_exceptionals() - Prune non-folios from a batch.
  * @fbatch: The batch to prune
-- 
2.35.1

[PATCH 04/22] mm/swap: Convert lru_add to a folio_batch

[Matthew Wilcox (Oracle)]

When adding folios to the LRU for the first time, the LRU flag will
already be clear, so skip the test-and-clear part of moving from one
LRU to another.

Removes 285 bytes from kernel text, mostly due to removing
__pagevec_lru_add().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 75 +++++++++++++++++++------------------------------------
 1 file changed, 26 insertions(+), 49 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index 6b015096ef4a..e9d54cb49cd3 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -61,7 +61,7 @@ static DEFINE_PER_CPU(struct lru_rotate, lru_rotate) = {
  */
 struct lru_pvecs {
 	local_lock_t lock;
-	struct pagevec lru_add;
+	struct folio_batch lru_add;
 	struct pagevec lru_deactivate_file;
 	struct pagevec lru_deactivate;
 	struct pagevec lru_lazyfree;
@@ -228,14 +228,13 @@ static bool pagevec_add_and_need_flush(struct pagevec *pvec, struct page *page)
 
 typedef void (*move_fn_t)(struct lruvec *lruvec, struct folio *folio);
 
-static void __pagevec_lru_add_fn(struct folio *folio, struct lruvec *lruvec)
+static void lru_add_fn(struct lruvec *lruvec, struct folio *folio)
 {
 	int was_unevictable = folio_test_clear_unevictable(folio);
 	long nr_pages = folio_nr_pages(folio);
 
 	VM_BUG_ON_FOLIO(folio_test_lru(folio), folio);
 
-	folio_set_lru(folio);
 	/*
 	 * Is an smp_mb__after_atomic() still required here, before
 	 * folio_evictable() tests PageMlocked, to rule out the possibility
@@ -269,28 +268,6 @@ static void __pagevec_lru_add_fn(struct folio *folio, struct lruvec *lruvec)
 	trace_mm_lru_insertion(folio);
 }
 
-/*
- * Add the passed pages to the LRU, then drop the caller's refcount
- * on them.  Reinitialises the caller's pagevec.
- */
-static void __pagevec_lru_add(struct pagevec *pvec)
-{
-	int i;
-	struct lruvec *lruvec = NULL;
-	unsigned long flags = 0;
-
-	for (i = 0; i < pagevec_count(pvec); i++) {
-		struct folio *folio = page_folio(pvec->pages[i]);
-
-		lruvec = folio_lruvec_relock_irqsave(folio, lruvec, &flags);
-		__pagevec_lru_add_fn(folio, lruvec);
-	}
-	if (lruvec)
-		unlock_page_lruvec_irqrestore(lruvec, flags);
-	release_pages(pvec->pages, pvec->nr);
-	pagevec_reinit(pvec);
-}
-
 static void folio_batch_move_lru(struct folio_batch *fbatch, move_fn_t move_fn)
 {
 	int i;
@@ -301,7 +278,7 @@ static void folio_batch_move_lru(struct folio_batch *fbatch, move_fn_t move_fn)
 		struct folio *folio = fbatch->folios[i];
 
 		/* block memcg migration while the folio moves between lru */
-		if (!folio_test_clear_lru(folio))
+		if (move_fn != lru_add_fn && !folio_test_clear_lru(folio))
 			continue;
 
 		lruvec = folio_lruvec_relock_irqsave(folio, lruvec, &flags);
@@ -473,26 +450,26 @@ static void folio_activate(struct folio *folio)
 
 static void __lru_cache_activate_folio(struct folio *folio)
 {
-	struct pagevec *pvec;
+	struct folio_batch *fbatch;
 	int i;
 
 	local_lock(&lru_pvecs.lock);
-	pvec = this_cpu_ptr(&lru_pvecs.lru_add);
+	fbatch = this_cpu_ptr(&lru_pvecs.lru_add);
 
 	/*
-	 * Search backwards on the optimistic assumption that the page being
-	 * activated has just been added to this pagevec. Note that only
-	 * the local pagevec is examined as a !PageLRU page could be in the
+	 * Search backwards on the optimistic assumption that the folio being
+	 * activated has just been added to this batch. Note that only
+	 * the local batch is examined as a !LRU folio could be in the
 	 * process of being released, reclaimed, migrated or on a remote
-	 * pagevec that is currently being drained. Furthermore, marking
-	 * a remote pagevec's page PageActive potentially hits a race where
-	 * a page is marked PageActive just after it is added to the inactive
+	 * batch that is currently being drained. Furthermore, marking
+	 * a remote batch's folio active potentially hits a race where
+	 * a folio is marked active just after it is added to the inactive
 	 * list causing accounting errors and BUG_ON checks to trigger.
 	 */
-	for (i = pagevec_count(pvec) - 1; i >= 0; i--) {
-		struct page *pagevec_page = pvec->pages[i];
+	for (i = folio_batch_count(fbatch) - 1; i >= 0; i--) {
+		struct folio *batch_folio = fbatch->folios[i];
 
-		if (pagevec_page == &folio->page) {
+		if (batch_folio == folio) {
 			folio_set_active(folio);
 			break;
 		}
@@ -551,16 +528,16 @@ EXPORT_SYMBOL(folio_mark_accessed);
  */
 void folio_add_lru(struct folio *folio)
 {
-	struct pagevec *pvec;
+	struct folio_batch *fbatch;
 
-	VM_BUG_ON_FOLIO(folio_test_active(folio) && folio_test_unevictable(folio), folio);
+	VM_BUG_ON_FOLIO(folio_test_active(folio) &&
+			folio_test_unevictable(folio), folio);
 	VM_BUG_ON_FOLIO(folio_test_lru(folio), folio);
 
 	folio_get(folio);
 	local_lock(&lru_pvecs.lock);
-	pvec = this_cpu_ptr(&lru_pvecs.lru_add);
-	if (pagevec_add_and_need_flush(pvec, &folio->page))
-		__pagevec_lru_add(pvec);
+	fbatch = this_cpu_ptr(&lru_pvecs.lru_add);
+	folio_batch_add_and_move(fbatch, folio, lru_add_fn);
 	local_unlock(&lru_pvecs.lock);
 }
 EXPORT_SYMBOL(folio_add_lru);
@@ -691,11 +668,11 @@ static void lru_lazyfree_fn(struct page *page, struct lruvec *lruvec)
  */
 void lru_add_drain_cpu(int cpu)
 {
-	struct folio_batch *fbatch;
-	struct pagevec *pvec = &per_cpu(lru_pvecs.lru_add, cpu);
+	struct folio_batch *fbatch = &per_cpu(lru_pvecs.lru_add, cpu);
+	struct pagevec *pvec;
 
-	if (pagevec_count(pvec))
-		__pagevec_lru_add(pvec);
+	if (folio_batch_count(fbatch))
+		folio_batch_move_lru(fbatch, lru_add_fn);
 
 	fbatch = &per_cpu(lru_rotate.fbatch, cpu);
 	/* Disabling interrupts below acts as a compiler barrier. */
@@ -920,7 +897,7 @@ static inline void __lru_add_drain_all(bool force_all_cpus)
 	for_each_online_cpu(cpu) {
 		struct work_struct *work = &per_cpu(lru_add_drain_work, cpu);
 
-		if (pagevec_count(&per_cpu(lru_pvecs.lru_add, cpu)) ||
+		if (folio_batch_count(&per_cpu(lru_pvecs.lru_add, cpu)) ||
 		    data_race(folio_batch_count(&per_cpu(lru_rotate.fbatch, cpu))) ||
 		    pagevec_count(&per_cpu(lru_pvecs.lru_deactivate_file, cpu)) ||
 		    pagevec_count(&per_cpu(lru_pvecs.lru_deactivate, cpu)) ||
@@ -1084,8 +1061,8 @@ EXPORT_SYMBOL(release_pages);
  * OK from a correctness point of view but is inefficient - those pages may be
  * cache-warm and we want to give them back to the page allocator ASAP.
  *
- * So __pagevec_release() will drain those queues here.  __pagevec_lru_add()
- * and __pagevec_lru_add_active() call release_pages() directly to avoid
+ * So __pagevec_release() will drain those queues here.
+ * folio_batch_move_lru() calls folios_put() directly to avoid
  * mutual recursion.
  */
 void __pagevec_release(struct pagevec *pvec)
-- 
2.35.1

[PATCH 05/22] mm/swap: Convert lru_deactivate_file to a folio_batch

[Matthew Wilcox (Oracle)]

Use a folio throughout lru_deactivate_file_fn(), removing many hidden
calls to compound_head().  Shrinks the kernel by 864 bytes of text.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 82 ++++++++++++++++++++++++++-----------------------------
 1 file changed, 39 insertions(+), 43 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index e9d54cb49cd3..1ca432163fe4 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -62,7 +62,7 @@ static DEFINE_PER_CPU(struct lru_rotate, lru_rotate) = {
 struct lru_pvecs {
 	local_lock_t lock;
 	struct folio_batch lru_add;
-	struct pagevec lru_deactivate_file;
+	struct folio_batch lru_deactivate_file;
 	struct pagevec lru_deactivate;
 	struct pagevec lru_lazyfree;
 #ifdef CONFIG_SMP
@@ -562,56 +562,57 @@ void lru_cache_add_inactive_or_unevictable(struct page *page,
 }
 
 /*
- * If the page can not be invalidated, it is moved to the
+ * If the folio cannot be invalidated, it is moved to the
  * inactive list to speed up its reclaim.  It is moved to the
  * head of the list, rather than the tail, to give the flusher
  * threads some time to write it out, as this is much more
  * effective than the single-page writeout from reclaim.
  *
- * If the page isn't page_mapped and dirty/writeback, the page
- * could reclaim asap using PG_reclaim.
+ * If the folio isn't mapped and dirty/writeback, the folio
+ * could be reclaimed asap using the reclaim flag.
  *
- * 1. active, mapped page -> none
- * 2. active, dirty/writeback page -> inactive, head, PG_reclaim
- * 3. inactive, mapped page -> none
- * 4. inactive, dirty/writeback page -> inactive, head, PG_reclaim
+ * 1. active, mapped folio -> none
+ * 2. active, dirty/writeback folio -> inactive, head, reclaim
+ * 3. inactive, mapped folio -> none
+ * 4. inactive, dirty/writeback folio -> inactive, head, reclaim
  * 5. inactive, clean -> inactive, tail
  * 6. Others -> none
  *
- * In 4, why it moves inactive's head, the VM expects the page would
- * be write it out by flusher threads as this is much more effective
+ * In 4, it moves to the head of the inactive list so the folio is
+ * written out by flusher threads as this is much more efficient
  * than the single-page writeout from reclaim.
  */
-static void lru_deactivate_file_fn(struct page *page, struct lruvec *lruvec)
+static void lru_deactivate_file_fn(struct lruvec *lruvec, struct folio *folio)
 {
-	bool active = PageActive(page);
-	int nr_pages = thp_nr_pages(page);
+	bool active = folio_test_active(folio);
+	long nr_pages = folio_nr_pages(folio);
 
-	if (PageUnevictable(page))
+	if (folio_test_unevictable(folio))
 		return;
 
-	/* Some processes are using the page */
-	if (page_mapped(page))
+	/* Some processes are using the folio */
+	if (folio_mapped(folio))
 		return;
 
-	del_page_from_lru_list(page, lruvec);
-	ClearPageActive(page);
-	ClearPageReferenced(page);
+	lruvec_del_folio(lruvec, folio);
+	folio_clear_active(folio);
+	folio_clear_referenced(folio);
 
-	if (PageWriteback(page) || PageDirty(page)) {
+	if (folio_test_writeback(folio) || folio_test_dirty(folio)) {
 		/*
-		 * PG_reclaim could be raced with end_page_writeback
-		 * It can make readahead confusing.  But race window
-		 * is _really_ small and  it's non-critical problem.
+		 * Setting the reclaim flag could race with
+		 * folio_end_writeback() and confuse readahead.  But the
+		 * race window is _really_ small and  it's not a critical
+		 * problem.
 		 */
-		add_page_to_lru_list(page, lruvec);
-		SetPageReclaim(page);
+		lruvec_add_folio(lruvec, folio);
+		folio_set_reclaim(folio);
 	} else {
 		/*
-		 * The page's writeback ends up during pagevec
-		 * We move that page into tail of inactive.
+		 * The folio's writeback ended while it was in the batch.
+		 * We move that folio to the tail of the inactive list.
 		 */
-		add_page_to_lru_list_tail(page, lruvec);
+		lruvec_add_folio_tail(lruvec, folio);
 		__count_vm_events(PGROTATED, nr_pages);
 	}
 
@@ -685,9 +686,9 @@ void lru_add_drain_cpu(int cpu)
 		local_unlock_irqrestore(&lru_rotate.lock, flags);
 	}
 
-	pvec = &per_cpu(lru_pvecs.lru_deactivate_file, cpu);
-	if (pagevec_count(pvec))
-		pagevec_lru_move_fn(pvec, lru_deactivate_file_fn);
+	fbatch = &per_cpu(lru_pvecs.lru_deactivate_file, cpu);
+	if (folio_batch_count(fbatch))
+		folio_batch_move_lru(fbatch, lru_deactivate_file_fn);
 
 	pvec = &per_cpu(lru_pvecs.lru_deactivate, cpu);
 	if (pagevec_count(pvec))
@@ -701,32 +702,27 @@ void lru_add_drain_cpu(int cpu)
 }
 
 /**
- * deactivate_file_folio() - Forcefully deactivate a file folio.
+ * deactivate_file_folio() - Deactivate a file folio.
  * @folio: Folio to deactivate.
  *
  * This function hints to the VM that @folio is a good reclaim candidate,
  * for example if its invalidation fails due to the folio being dirty
  * or under writeback.
  *
- * Context: Caller holds a reference on the page.
+ * Context: Caller holds a reference on the folio.
  */
 void deactivate_file_folio(struct folio *folio)
 {
-	struct pagevec *pvec;
+	struct folio_batch *fbatch;
 
-	/*
-	 * In a workload with many unevictable pages such as mprotect,
-	 * unevictable folio deactivation for accelerating reclaim is pointless.
-	 */
+	/* Deactivating an unevictable folio will not accelerate reclaim */
 	if (folio_test_unevictable(folio))
 		return;
 
 	folio_get(folio);
 	local_lock(&lru_pvecs.lock);
-	pvec = this_cpu_ptr(&lru_pvecs.lru_deactivate_file);
-
-	if (pagevec_add_and_need_flush(pvec, &folio->page))
-		pagevec_lru_move_fn(pvec, lru_deactivate_file_fn);
+	fbatch = this_cpu_ptr(&lru_pvecs.lru_deactivate_file);
+	folio_batch_add_and_move(fbatch, folio, lru_deactivate_file_fn);
 	local_unlock(&lru_pvecs.lock);
 }
 
@@ -899,7 +895,7 @@ static inline void __lru_add_drain_all(bool force_all_cpus)
 
 		if (folio_batch_count(&per_cpu(lru_pvecs.lru_add, cpu)) ||
 		    data_race(folio_batch_count(&per_cpu(lru_rotate.fbatch, cpu))) ||
-		    pagevec_count(&per_cpu(lru_pvecs.lru_deactivate_file, cpu)) ||
+		    folio_batch_count(&per_cpu(lru_pvecs.lru_deactivate_file, cpu)) ||
 		    pagevec_count(&per_cpu(lru_pvecs.lru_deactivate, cpu)) ||
 		    pagevec_count(&per_cpu(lru_pvecs.lru_lazyfree, cpu)) ||
 		    need_activate_page_drain(cpu) ||
-- 
2.35.1

[PATCH 06/22] mm/swap: Convert lru_deactivate to a folio_batch

[Matthew Wilcox (Oracle)]

Using folios instead of pages shrinks deactivate_page() and
lru_deactivate_fn() by 778 bytes between them.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 38 ++++++++++++++++++++------------------
 1 file changed, 20 insertions(+), 18 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index 1ca432163fe4..c0323b312a46 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -63,7 +63,7 @@ struct lru_pvecs {
 	local_lock_t lock;
 	struct folio_batch lru_add;
 	struct folio_batch lru_deactivate_file;
-	struct pagevec lru_deactivate;
+	struct folio_batch lru_deactivate;
 	struct pagevec lru_lazyfree;
 #ifdef CONFIG_SMP
 	struct pagevec activate_page;
@@ -623,15 +623,15 @@ static void lru_deactivate_file_fn(struct lruvec *lruvec, struct folio *folio)
 	}
 }
 
-static void lru_deactivate_fn(struct page *page, struct lruvec *lruvec)
+static void lru_deactivate_fn(struct lruvec *lruvec, struct folio *folio)
 {
-	if (PageActive(page) && !PageUnevictable(page)) {
-		int nr_pages = thp_nr_pages(page);
+	if (folio_test_active(folio) && !folio_test_unevictable(folio)) {
+		long nr_pages = folio_nr_pages(folio);
 
-		del_page_from_lru_list(page, lruvec);
-		ClearPageActive(page);
-		ClearPageReferenced(page);
-		add_page_to_lru_list(page, lruvec);
+		lruvec_del_folio(lruvec, folio);
+		folio_clear_active(folio);
+		folio_clear_referenced(folio);
+		lruvec_add_folio(lruvec, folio);
 
 		__count_vm_events(PGDEACTIVATE, nr_pages);
 		__count_memcg_events(lruvec_memcg(lruvec), PGDEACTIVATE,
@@ -690,9 +690,9 @@ void lru_add_drain_cpu(int cpu)
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_deactivate_file_fn);
 
-	pvec = &per_cpu(lru_pvecs.lru_deactivate, cpu);
-	if (pagevec_count(pvec))
-		pagevec_lru_move_fn(pvec, lru_deactivate_fn);
+	fbatch = &per_cpu(lru_pvecs.lru_deactivate, cpu);
+	if (folio_batch_count(fbatch))
+		folio_batch_move_lru(fbatch, lru_deactivate_fn);
 
 	pvec = &per_cpu(lru_pvecs.lru_lazyfree, cpu);
 	if (pagevec_count(pvec))
@@ -736,14 +736,16 @@ void deactivate_file_folio(struct folio *folio)
  */
 void deactivate_page(struct page *page)
 {
-	if (PageLRU(page) && PageActive(page) && !PageUnevictable(page)) {
-		struct pagevec *pvec;
+	struct folio *folio = page_folio(page);
 
+	if (folio_test_lru(folio) && folio_test_active(folio) &&
+	    !folio_test_unevictable(folio)) {
+		struct folio_batch *fbatch;
+
+		folio_get(folio);
 		local_lock(&lru_pvecs.lock);
-		pvec = this_cpu_ptr(&lru_pvecs.lru_deactivate);
-		get_page(page);
-		if (pagevec_add_and_need_flush(pvec, page))
-			pagevec_lru_move_fn(pvec, lru_deactivate_fn);
+		fbatch = this_cpu_ptr(&lru_pvecs.lru_deactivate);
+		folio_batch_add_and_move(fbatch, folio, lru_deactivate_fn);
 		local_unlock(&lru_pvecs.lock);
 	}
 }
@@ -896,7 +898,7 @@ static inline void __lru_add_drain_all(bool force_all_cpus)
 		if (folio_batch_count(&per_cpu(lru_pvecs.lru_add, cpu)) ||
 		    data_race(folio_batch_count(&per_cpu(lru_rotate.fbatch, cpu))) ||
 		    folio_batch_count(&per_cpu(lru_pvecs.lru_deactivate_file, cpu)) ||
-		    pagevec_count(&per_cpu(lru_pvecs.lru_deactivate, cpu)) ||
+		    folio_batch_count(&per_cpu(lru_pvecs.lru_deactivate, cpu)) ||
 		    pagevec_count(&per_cpu(lru_pvecs.lru_lazyfree, cpu)) ||
 		    need_activate_page_drain(cpu) ||
 		    need_mlock_page_drain(cpu) ||
-- 
2.35.1

[PATCH 07/22] mm/swap: Convert lru_lazyfree to a folio_batch

[Matthew Wilcox (Oracle)]

Using folios instead of pages removes several calls to compound_head(),
shrinking the kernel by 1089 bytes of text.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 51 ++++++++++++++++++++++++++-------------------------
 1 file changed, 26 insertions(+), 25 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index c0323b312a46..f2d7b0c3bd54 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -64,7 +64,7 @@ struct lru_pvecs {
 	struct folio_batch lru_add;
 	struct folio_batch lru_deactivate_file;
 	struct folio_batch lru_deactivate;
-	struct pagevec lru_lazyfree;
+	struct folio_batch lru_lazyfree;
 #ifdef CONFIG_SMP
 	struct pagevec activate_page;
 #endif
@@ -639,22 +639,22 @@ static void lru_deactivate_fn(struct lruvec *lruvec, struct folio *folio)
 	}
 }
 
-static void lru_lazyfree_fn(struct page *page, struct lruvec *lruvec)
+static void lru_lazyfree_fn(struct lruvec *lruvec, struct folio *folio)
 {
-	if (PageAnon(page) && PageSwapBacked(page) &&
-	    !PageSwapCache(page) && !PageUnevictable(page)) {
-		int nr_pages = thp_nr_pages(page);
+	if (folio_test_anon(folio) && folio_test_swapbacked(folio) &&
+	    !folio_test_swapcache(folio) && !folio_test_unevictable(folio)) {
+		long nr_pages = folio_nr_pages(folio);
 
-		del_page_from_lru_list(page, lruvec);
-		ClearPageActive(page);
-		ClearPageReferenced(page);
+		lruvec_del_folio(lruvec, folio);
+		folio_clear_active(folio);
+		folio_clear_referenced(folio);
 		/*
-		 * Lazyfree pages are clean anonymous pages.  They have
-		 * PG_swapbacked flag cleared, to distinguish them from normal
-		 * anonymous pages
+		 * Lazyfree folios are clean anonymous folios.  They have
+		 * the swapbacked flag cleared, to distinguish them from normal
+		 * anonymous folios
 		 */
-		ClearPageSwapBacked(page);
-		add_page_to_lru_list(page, lruvec);
+		folio_clear_swapbacked(folio);
+		lruvec_add_folio(lruvec, folio);
 
 		__count_vm_events(PGLAZYFREE, nr_pages);
 		__count_memcg_events(lruvec_memcg(lruvec), PGLAZYFREE,
@@ -670,7 +670,6 @@ static void lru_lazyfree_fn(struct page *page, struct lruvec *lruvec)
 void lru_add_drain_cpu(int cpu)
 {
 	struct folio_batch *fbatch = &per_cpu(lru_pvecs.lru_add, cpu);
-	struct pagevec *pvec;
 
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_add_fn);
@@ -694,9 +693,9 @@ void lru_add_drain_cpu(int cpu)
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_deactivate_fn);
 
-	pvec = &per_cpu(lru_pvecs.lru_lazyfree, cpu);
-	if (pagevec_count(pvec))
-		pagevec_lru_move_fn(pvec, lru_lazyfree_fn);
+	fbatch = &per_cpu(lru_pvecs.lru_lazyfree, cpu);
+	if (folio_batch_count(fbatch))
+		folio_batch_move_lru(fbatch, lru_lazyfree_fn);
 
 	activate_page_drain(cpu);
 }
@@ -759,15 +758,17 @@ void deactivate_page(struct page *page)
  */
 void mark_page_lazyfree(struct page *page)
 {
-	if (PageLRU(page) && PageAnon(page) && PageSwapBacked(page) &&
-	    !PageSwapCache(page) && !PageUnevictable(page)) {
-		struct pagevec *pvec;
+	struct folio *folio = page_folio(page);
 
+	if (folio_test_lru(folio) && folio_test_anon(folio) &&
+	    folio_test_swapbacked(folio) && !folio_test_swapcache(folio) &&
+	    !folio_test_unevictable(folio)) {
+		struct folio_batch *fbatch;
+
+		folio_get(folio);
 		local_lock(&lru_pvecs.lock);
-		pvec = this_cpu_ptr(&lru_pvecs.lru_lazyfree);
-		get_page(page);
-		if (pagevec_add_and_need_flush(pvec, page))
-			pagevec_lru_move_fn(pvec, lru_lazyfree_fn);
+		fbatch = this_cpu_ptr(&lru_pvecs.lru_lazyfree);
+		folio_batch_add_and_move(fbatch, folio, lru_lazyfree_fn);
 		local_unlock(&lru_pvecs.lock);
 	}
 }
@@ -899,7 +900,7 @@ static inline void __lru_add_drain_all(bool force_all_cpus)
 		    data_race(folio_batch_count(&per_cpu(lru_rotate.fbatch, cpu))) ||
 		    folio_batch_count(&per_cpu(lru_pvecs.lru_deactivate_file, cpu)) ||
 		    folio_batch_count(&per_cpu(lru_pvecs.lru_deactivate, cpu)) ||
-		    pagevec_count(&per_cpu(lru_pvecs.lru_lazyfree, cpu)) ||
+		    folio_batch_count(&per_cpu(lru_pvecs.lru_lazyfree, cpu)) ||
 		    need_activate_page_drain(cpu) ||
 		    need_mlock_page_drain(cpu) ||
 		    has_bh_in_lru(cpu, NULL)) {
-- 
2.35.1

[PATCH 08/22] mm/swap: Convert activate_page to a folio_batch

[Matthew Wilcox (Oracle)]

Rename it to just 'activate', saving 696 bytes of text from removals
of compound_page() and the pagevec_lru_move_fn() infrastructure.
Inline need_activate_page_drain() into its only caller.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 81 +++++++++++--------------------------------------------
 1 file changed, 16 insertions(+), 65 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index f2d7b0c3bd54..bba9ed25c56d 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -66,7 +66,7 @@ struct lru_pvecs {
 	struct folio_batch lru_deactivate;
 	struct folio_batch lru_lazyfree;
 #ifdef CONFIG_SMP
-	struct pagevec activate_page;
+	struct folio_batch activate;
 #endif
 };
 static DEFINE_PER_CPU(struct lru_pvecs, lru_pvecs) = {
@@ -188,44 +188,6 @@ int get_kernel_pages(const struct kvec *kiov, int nr_segs, int write,
 }
 EXPORT_SYMBOL_GPL(get_kernel_pages);
 
-static void pagevec_lru_move_fn(struct pagevec *pvec,
-	void (*move_fn)(struct page *page, struct lruvec *lruvec))
-{
-	int i;
-	struct lruvec *lruvec = NULL;
-	unsigned long flags = 0;
-
-	for (i = 0; i < pagevec_count(pvec); i++) {
-		struct page *page = pvec->pages[i];
-		struct folio *folio = page_folio(page);
-
-		/* block memcg migration during page moving between lru */
-		if (!TestClearPageLRU(page))
-			continue;
-
-		lruvec = folio_lruvec_relock_irqsave(folio, lruvec, &flags);
-		(*move_fn)(page, lruvec);
-
-		SetPageLRU(page);
-	}
-	if (lruvec)
-		unlock_page_lruvec_irqrestore(lruvec, flags);
-	release_pages(pvec->pages, pvec->nr);
-	pagevec_reinit(pvec);
-}
-
-/* return true if pagevec needs to drain */
-static bool pagevec_add_and_need_flush(struct pagevec *pvec, struct page *page)
-{
-	bool ret = false;
-
-	if (!pagevec_add(pvec, page) || PageCompound(page) ||
-			lru_cache_disabled())
-		ret = true;
-
-	return ret;
-}
-
 typedef void (*move_fn_t)(struct lruvec *lruvec, struct folio *folio);
 
 static void lru_add_fn(struct lruvec *lruvec, struct folio *folio)
@@ -380,7 +342,7 @@ void lru_note_cost_folio(struct folio *folio)
 			folio_nr_pages(folio));
 }
 
-static void __folio_activate(struct folio *folio, struct lruvec *lruvec)
+static void folio_activate_fn(struct lruvec *lruvec, struct folio *folio)
 {
 	if (!folio_test_active(folio) && !folio_test_unevictable(folio)) {
 		long nr_pages = folio_nr_pages(folio);
@@ -397,41 +359,30 @@ static void __folio_activate(struct folio *folio, struct lruvec *lruvec)
 }
 
 #ifdef CONFIG_SMP
-static void __activate_page(struct page *page, struct lruvec *lruvec)
+static void folio_activate_drain(int cpu)
 {
-	return __folio_activate(page_folio(page), lruvec);
-}
+	struct folio_batch *fbatch = &per_cpu(lru_pvecs.activate, cpu);
 
-static void activate_page_drain(int cpu)
-{
-	struct pagevec *pvec = &per_cpu(lru_pvecs.activate_page, cpu);
-
-	if (pagevec_count(pvec))
-		pagevec_lru_move_fn(pvec, __activate_page);
-}
-
-static bool need_activate_page_drain(int cpu)
-{
-	return pagevec_count(&per_cpu(lru_pvecs.activate_page, cpu)) != 0;
+	if (folio_batch_count(fbatch))
+		folio_batch_move_lru(fbatch, folio_activate_fn);
 }
 
 static void folio_activate(struct folio *folio)
 {
 	if (folio_test_lru(folio) && !folio_test_active(folio) &&
 	    !folio_test_unevictable(folio)) {
-		struct pagevec *pvec;
+		struct folio_batch *fbatch;
 
 		folio_get(folio);
 		local_lock(&lru_pvecs.lock);
-		pvec = this_cpu_ptr(&lru_pvecs.activate_page);
-		if (pagevec_add_and_need_flush(pvec, &folio->page))
-			pagevec_lru_move_fn(pvec, __activate_page);
+		fbatch = this_cpu_ptr(&lru_pvecs.activate);
+		folio_batch_add_and_move(fbatch, folio, folio_activate_fn);
 		local_unlock(&lru_pvecs.lock);
 	}
 }
 
 #else
-static inline void activate_page_drain(int cpu)
+static inline void folio_activate_drain(int cpu)
 {
 }
 
@@ -441,7 +392,7 @@ static void folio_activate(struct folio *folio)
 
 	if (folio_test_clear_lru(folio)) {
 		lruvec = folio_lruvec_lock_irq(folio);
-		__folio_activate(folio, lruvec);
+		folio_activate_fn(lruvec, folio);
 		unlock_page_lruvec_irq(lruvec);
 		folio_set_lru(folio);
 	}
@@ -500,9 +451,9 @@ void folio_mark_accessed(struct folio *folio)
 		 */
 	} else if (!folio_test_active(folio)) {
 		/*
-		 * If the page is on the LRU, queue it for activation via
-		 * lru_pvecs.activate_page. Otherwise, assume the page is on a
-		 * pagevec, mark it active and it'll be moved to the active
+		 * If the folio is on the LRU, queue it for activation via
+		 * lru_pvecs.activate. Otherwise, assume the folio is in a
+		 * folio_batch, mark it active and it'll be moved to the active
 		 * LRU on the next drain.
 		 */
 		if (folio_test_lru(folio))
@@ -697,7 +648,7 @@ void lru_add_drain_cpu(int cpu)
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_lazyfree_fn);
 
-	activate_page_drain(cpu);
+	folio_activate_drain(cpu);
 }
 
 /**
@@ -901,7 +852,7 @@ static inline void __lru_add_drain_all(bool force_all_cpus)
 		    folio_batch_count(&per_cpu(lru_pvecs.lru_deactivate_file, cpu)) ||
 		    folio_batch_count(&per_cpu(lru_pvecs.lru_deactivate, cpu)) ||
 		    folio_batch_count(&per_cpu(lru_pvecs.lru_lazyfree, cpu)) ||
-		    need_activate_page_drain(cpu) ||
+		    folio_batch_count(&per_cpu(lru_pvecs.activate, cpu)) ||
 		    need_mlock_page_drain(cpu) ||
 		    has_bh_in_lru(cpu, NULL)) {
 			INIT_WORK(work, lru_add_drain_per_cpu);
-- 
2.35.1

[PATCH 09/22] mm/swap: Rename lru_pvecs to cpu_fbatches

[Matthew Wilcox (Oracle)]

No change to generated code, but this struct no longer contains any
pagevecs, and not all the folio batches it contains are lru.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 90 ++++++++++++++++++++++++++++---------------------------
 1 file changed, 46 insertions(+), 44 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index bba9ed25c56d..e65a195c184b 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -56,10 +56,10 @@ static DEFINE_PER_CPU(struct lru_rotate, lru_rotate) = {
 };
 
 /*
- * The following struct pagevec are grouped together because they are protected
+ * The following folio batches are grouped together because they are protected
  * by disabling preemption (and interrupts remain enabled).
  */
-struct lru_pvecs {
+struct cpu_fbatches {
 	local_lock_t lock;
 	struct folio_batch lru_add;
 	struct folio_batch lru_deactivate_file;
@@ -69,7 +69,7 @@ struct lru_pvecs {
 	struct folio_batch activate;
 #endif
 };
-static DEFINE_PER_CPU(struct lru_pvecs, lru_pvecs) = {
+static DEFINE_PER_CPU(struct cpu_fbatches, cpu_fbatches) = {
 	.lock = INIT_LOCAL_LOCK(lock),
 };
 
@@ -361,7 +361,7 @@ static void folio_activate_fn(struct lruvec *lruvec, struct folio *folio)
 #ifdef CONFIG_SMP
 static void folio_activate_drain(int cpu)
 {
-	struct folio_batch *fbatch = &per_cpu(lru_pvecs.activate, cpu);
+	struct folio_batch *fbatch = &per_cpu(cpu_fbatches.activate, cpu);
 
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, folio_activate_fn);
@@ -374,10 +374,10 @@ static void folio_activate(struct folio *folio)
 		struct folio_batch *fbatch;
 
 		folio_get(folio);
-		local_lock(&lru_pvecs.lock);
-		fbatch = this_cpu_ptr(&lru_pvecs.activate);
+		local_lock(&cpu_fbatches.lock);
+		fbatch = this_cpu_ptr(&cpu_fbatches.activate);
 		folio_batch_add_and_move(fbatch, folio, folio_activate_fn);
-		local_unlock(&lru_pvecs.lock);
+		local_unlock(&cpu_fbatches.lock);
 	}
 }
 
@@ -404,8 +404,8 @@ static void __lru_cache_activate_folio(struct folio *folio)
 	struct folio_batch *fbatch;
 	int i;
 
-	local_lock(&lru_pvecs.lock);
-	fbatch = this_cpu_ptr(&lru_pvecs.lru_add);
+	local_lock(&cpu_fbatches.lock);
+	fbatch = this_cpu_ptr(&cpu_fbatches.lru_add);
 
 	/*
 	 * Search backwards on the optimistic assumption that the folio being
@@ -426,7 +426,7 @@ static void __lru_cache_activate_folio(struct folio *folio)
 		}
 	}
 
-	local_unlock(&lru_pvecs.lock);
+	local_unlock(&cpu_fbatches.lock);
 }
 
 /*
@@ -452,7 +452,7 @@ void folio_mark_accessed(struct folio *folio)
 	} else if (!folio_test_active(folio)) {
 		/*
 		 * If the folio is on the LRU, queue it for activation via
-		 * lru_pvecs.activate. Otherwise, assume the folio is in a
+		 * cpu_fbatches.activate. Otherwise, assume the folio is in a
 		 * folio_batch, mark it active and it'll be moved to the active
 		 * LRU on the next drain.
 		 */
@@ -474,7 +474,7 @@ EXPORT_SYMBOL(folio_mark_accessed);
  *
  * Queue the folio for addition to the LRU. The decision on whether
  * to add the page to the [in]active [file|anon] list is deferred until the
- * pagevec is drained. This gives a chance for the caller of folio_add_lru()
+ * folio_batch is drained. This gives a chance for the caller of folio_add_lru()
  * have the folio added to the active list using folio_mark_accessed().
  */
 void folio_add_lru(struct folio *folio)
@@ -486,10 +486,10 @@ void folio_add_lru(struct folio *folio)
 	VM_BUG_ON_FOLIO(folio_test_lru(folio), folio);
 
 	folio_get(folio);
-	local_lock(&lru_pvecs.lock);
-	fbatch = this_cpu_ptr(&lru_pvecs.lru_add);
+	local_lock(&cpu_fbatches.lock);
+	fbatch = this_cpu_ptr(&cpu_fbatches.lru_add);
 	folio_batch_add_and_move(fbatch, folio, lru_add_fn);
-	local_unlock(&lru_pvecs.lock);
+	local_unlock(&cpu_fbatches.lock);
 }
 EXPORT_SYMBOL(folio_add_lru);
 
@@ -614,13 +614,13 @@ static void lru_lazyfree_fn(struct lruvec *lruvec, struct folio *folio)
 }
 
 /*
- * Drain pages out of the cpu's pagevecs.
+ * Drain pages out of the cpu's folio_batch.
  * Either "cpu" is the current CPU, and preemption has already been
  * disabled; or "cpu" is being hot-unplugged, and is already dead.
  */
 void lru_add_drain_cpu(int cpu)
 {
-	struct folio_batch *fbatch = &per_cpu(lru_pvecs.lru_add, cpu);
+	struct folio_batch *fbatch = &per_cpu(cpu_fbatches.lru_add, cpu);
 
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_add_fn);
@@ -636,15 +636,15 @@ void lru_add_drain_cpu(int cpu)
 		local_unlock_irqrestore(&lru_rotate.lock, flags);
 	}
 
-	fbatch = &per_cpu(lru_pvecs.lru_deactivate_file, cpu);
+	fbatch = &per_cpu(cpu_fbatches.lru_deactivate_file, cpu);
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_deactivate_file_fn);
 
-	fbatch = &per_cpu(lru_pvecs.lru_deactivate, cpu);
+	fbatch = &per_cpu(cpu_fbatches.lru_deactivate, cpu);
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_deactivate_fn);
 
-	fbatch = &per_cpu(lru_pvecs.lru_lazyfree, cpu);
+	fbatch = &per_cpu(cpu_fbatches.lru_lazyfree, cpu);
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_lazyfree_fn);
 
@@ -670,10 +670,10 @@ void deactivate_file_folio(struct folio *folio)
 		return;
 
 	folio_get(folio);
-	local_lock(&lru_pvecs.lock);
-	fbatch = this_cpu_ptr(&lru_pvecs.lru_deactivate_file);
+	local_lock(&cpu_fbatches.lock);
+	fbatch = this_cpu_ptr(&cpu_fbatches.lru_deactivate_file);
 	folio_batch_add_and_move(fbatch, folio, lru_deactivate_file_fn);
-	local_unlock(&lru_pvecs.lock);
+	local_unlock(&cpu_fbatches.lock);
 }
 
 /*
@@ -693,10 +693,10 @@ void deactivate_page(struct page *page)
 		struct folio_batch *fbatch;
 
 		folio_get(folio);
-		local_lock(&lru_pvecs.lock);
-		fbatch = this_cpu_ptr(&lru_pvecs.lru_deactivate);
+		local_lock(&cpu_fbatches.lock);
+		fbatch = this_cpu_ptr(&cpu_fbatches.lru_deactivate);
 		folio_batch_add_and_move(fbatch, folio, lru_deactivate_fn);
-		local_unlock(&lru_pvecs.lock);
+		local_unlock(&cpu_fbatches.lock);
 	}
 }
 
@@ -717,18 +717,18 @@ void mark_page_lazyfree(struct page *page)
 		struct folio_batch *fbatch;
 
 		folio_get(folio);
-		local_lock(&lru_pvecs.lock);
-		fbatch = this_cpu_ptr(&lru_pvecs.lru_lazyfree);
+		local_lock(&cpu_fbatches.lock);
+		fbatch = this_cpu_ptr(&cpu_fbatches.lru_lazyfree);
 		folio_batch_add_and_move(fbatch, folio, lru_lazyfree_fn);
-		local_unlock(&lru_pvecs.lock);
+		local_unlock(&cpu_fbatches.lock);
 	}
 }
 
 void lru_add_drain(void)
 {
-	local_lock(&lru_pvecs.lock);
+	local_lock(&cpu_fbatches.lock);
 	lru_add_drain_cpu(smp_processor_id());
-	local_unlock(&lru_pvecs.lock);
+	local_unlock(&cpu_fbatches.lock);
 	mlock_page_drain_local();
 }
 
@@ -740,19 +740,19 @@ void lru_add_drain(void)
  */
 static void lru_add_and_bh_lrus_drain(void)
 {
-	local_lock(&lru_pvecs.lock);
+	local_lock(&cpu_fbatches.lock);
 	lru_add_drain_cpu(smp_processor_id());
-	local_unlock(&lru_pvecs.lock);
+	local_unlock(&cpu_fbatches.lock);
 	invalidate_bh_lrus_cpu();
 	mlock_page_drain_local();
 }
 
 void lru_add_drain_cpu_zone(struct zone *zone)
 {
-	local_lock(&lru_pvecs.lock);
+	local_lock(&cpu_fbatches.lock);
 	lru_add_drain_cpu(smp_processor_id());
 	drain_local_pages(zone);
-	local_unlock(&lru_pvecs.lock);
+	local_unlock(&cpu_fbatches.lock);
 	mlock_page_drain_local();
 }
 
@@ -797,8 +797,9 @@ static inline void __lru_add_drain_all(bool force_all_cpus)
 		return;
 
 	/*
-	 * Guarantee pagevec counter stores visible by this CPU are visible to
-	 * other CPUs before loading the current drain generation.
+	 * Guarantee folio_batch counter stores visible by this CPU
+	 * are visible to other CPUs before loading the current drain
+	 * generation.
 	 */
 	smp_mb();
 
@@ -824,8 +825,9 @@ static inline void __lru_add_drain_all(bool force_all_cpus)
 	 * (D) Increment global generation number
 	 *
 	 * Pairs with smp_load_acquire() at (B), outside of the critical
-	 * section. Use a full memory barrier to guarantee that the new global
-	 * drain generation number is stored before loading pagevec counters.
+	 * section. Use a full memory barrier to guarantee that the
+	 * new global drain generation number is stored before loading
+	 * folio_batch counters.
 	 *
 	 * This pairing must be done here, before the for_each_online_cpu loop
 	 * below which drains the page vectors.
@@ -847,12 +849,12 @@ static inline void __lru_add_drain_all(bool force_all_cpus)
 	for_each_online_cpu(cpu) {
 		struct work_struct *work = &per_cpu(lru_add_drain_work, cpu);
 
-		if (folio_batch_count(&per_cpu(lru_pvecs.lru_add, cpu)) ||
+		if (folio_batch_count(&per_cpu(cpu_fbatches.lru_add, cpu)) ||
 		    data_race(folio_batch_count(&per_cpu(lru_rotate.fbatch, cpu))) ||
-		    folio_batch_count(&per_cpu(lru_pvecs.lru_deactivate_file, cpu)) ||
-		    folio_batch_count(&per_cpu(lru_pvecs.lru_deactivate, cpu)) ||
-		    folio_batch_count(&per_cpu(lru_pvecs.lru_lazyfree, cpu)) ||
-		    folio_batch_count(&per_cpu(lru_pvecs.activate, cpu)) ||
+		    folio_batch_count(&per_cpu(cpu_fbatches.lru_deactivate_file, cpu)) ||
+		    folio_batch_count(&per_cpu(cpu_fbatches.lru_deactivate, cpu)) ||
+		    folio_batch_count(&per_cpu(cpu_fbatches.lru_lazyfree, cpu)) ||
+		    folio_batch_count(&per_cpu(cpu_fbatches.activate, cpu)) ||
 		    need_mlock_page_drain(cpu) ||
 		    has_bh_in_lru(cpu, NULL)) {
 			INIT_WORK(work, lru_add_drain_per_cpu);
-- 
2.35.1

[PATCH 10/22] mm/swap: Pull the CPU conditional out of __lru_add_drain_all()

[Matthew Wilcox (Oracle)]

The function is too long, so pull this complicated conditional out into
cpu_needs_drain().  This ends up shrinking the text by 14 bytes,
by allowing GCC to cache the result of calling per_cpu() instead of
relocating each lookup individually.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 24 ++++++++++++++++--------
 1 file changed, 16 insertions(+), 8 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index e65a195c184b..7966aa6bdd3f 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -765,6 +765,21 @@ static void lru_add_drain_per_cpu(struct work_struct *dummy)
 	lru_add_and_bh_lrus_drain();
 }
 
+static bool cpu_needs_drain(unsigned int cpu)
+{
+	struct cpu_fbatches *fbatches = &per_cpu(cpu_fbatches, cpu);
+
+	/* Check these in order of likelihood that they're not zero */
+	return folio_batch_count(&fbatches->lru_add) ||
+		data_race(folio_batch_count(&per_cpu(lru_rotate.fbatch, cpu))) ||
+		folio_batch_count(&fbatches->lru_deactivate_file) ||
+		folio_batch_count(&fbatches->lru_deactivate) ||
+		folio_batch_count(&fbatches->lru_lazyfree) ||
+		folio_batch_count(&fbatches->activate) ||
+		need_mlock_page_drain(cpu) ||
+		has_bh_in_lru(cpu, NULL);
+}
+
 /*
  * Doesn't need any cpu hotplug locking because we do rely on per-cpu
  * kworkers being shut down before our page_alloc_cpu_dead callback is
@@ -849,14 +864,7 @@ static inline void __lru_add_drain_all(bool force_all_cpus)
 	for_each_online_cpu(cpu) {
 		struct work_struct *work = &per_cpu(lru_add_drain_work, cpu);
 
-		if (folio_batch_count(&per_cpu(cpu_fbatches.lru_add, cpu)) ||
-		    data_race(folio_batch_count(&per_cpu(lru_rotate.fbatch, cpu))) ||
-		    folio_batch_count(&per_cpu(cpu_fbatches.lru_deactivate_file, cpu)) ||
-		    folio_batch_count(&per_cpu(cpu_fbatches.lru_deactivate, cpu)) ||
-		    folio_batch_count(&per_cpu(cpu_fbatches.lru_lazyfree, cpu)) ||
-		    folio_batch_count(&per_cpu(cpu_fbatches.activate, cpu)) ||
-		    need_mlock_page_drain(cpu) ||
-		    has_bh_in_lru(cpu, NULL)) {
+		if (cpu_needs_drain(cpu)) {
 			INIT_WORK(work, lru_add_drain_per_cpu);
 			queue_work_on(cpu, mm_percpu_wq, work);
 			__cpumask_set_cpu(cpu, &has_work);
-- 
2.35.1

[PATCH 11/22] mm/swap: Optimise lru_add_drain_cpu()

[Matthew Wilcox (Oracle)]

Do the per-cpu dereferencing of the fbatches once which saves 14 bytes
of text and several percpu relocations.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index 7966aa6bdd3f..5ea6ae795f0d 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -620,7 +620,8 @@ static void lru_lazyfree_fn(struct lruvec *lruvec, struct folio *folio)
  */
 void lru_add_drain_cpu(int cpu)
 {
-	struct folio_batch *fbatch = &per_cpu(cpu_fbatches.lru_add, cpu);
+	struct cpu_fbatches *fbatches = &per_cpu(cpu_fbatches, cpu);
+	struct folio_batch *fbatch = &fbatches->lru_add;
 
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_add_fn);
@@ -636,15 +637,15 @@ void lru_add_drain_cpu(int cpu)
 		local_unlock_irqrestore(&lru_rotate.lock, flags);
 	}
 
-	fbatch = &per_cpu(cpu_fbatches.lru_deactivate_file, cpu);
+	fbatch = &fbatches->lru_deactivate_file;
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_deactivate_file_fn);
 
-	fbatch = &per_cpu(cpu_fbatches.lru_deactivate, cpu);
+	fbatch = &fbatches->lru_deactivate;
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_deactivate_fn);
 
-	fbatch = &per_cpu(cpu_fbatches.lru_lazyfree, cpu);
+	fbatch = &fbatches->lru_lazyfree;
 	if (folio_batch_count(fbatch))
 		folio_batch_move_lru(fbatch, lru_lazyfree_fn);
 
-- 
2.35.1

[PATCH 12/22] mm/swap: Convert try_to_free_swap to use a folio

[Matthew Wilcox (Oracle)]

Save a few calls to compound_head by converting the passed page to
a folio.  Reduces kernel text size by 74 bytes.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swapfile.c | 23 +++++++++++------------
 1 file changed, 11 insertions(+), 12 deletions(-)

diff --git a/mm/swapfile.c b/mm/swapfile.c
index 3fa26f6971e9..4ac6a6c876d6 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -1568,16 +1568,15 @@ static bool swap_page_trans_huge_swapped(struct swap_info_struct *si,
 	return ret;
 }
 
-static bool page_swapped(struct page *page)
+static bool folio_swapped(struct folio *folio)
 {
 	swp_entry_t entry;
 	struct swap_info_struct *si;
 
-	if (!IS_ENABLED(CONFIG_THP_SWAP) || likely(!PageTransCompound(page)))
-		return page_swapcount(page) != 0;
+	if (!IS_ENABLED(CONFIG_THP_SWAP) || likely(!folio_test_large(folio)))
+		return page_swapcount(&folio->page) != 0;
 
-	page = compound_head(page);
-	entry.val = page_private(page);
+	entry = folio_swap_entry(folio);
 	si = _swap_info_get(entry);
 	if (si)
 		return swap_page_trans_huge_swapped(si, entry);
@@ -1590,13 +1589,14 @@ static bool page_swapped(struct page *page)
  */
 int try_to_free_swap(struct page *page)
 {
-	VM_BUG_ON_PAGE(!PageLocked(page), page);
+	struct folio *folio = page_folio(page);
+	VM_BUG_ON_FOLIO(!folio_test_locked(folio), folio);
 
-	if (!PageSwapCache(page))
+	if (!folio_test_swapcache(folio))
 		return 0;
-	if (PageWriteback(page))
+	if (folio_test_writeback(folio))
 		return 0;
-	if (page_swapped(page))
+	if (folio_swapped(folio))
 		return 0;
 
 	/*
@@ -1617,9 +1617,8 @@ int try_to_free_swap(struct page *page)
 	if (pm_suspended_storage())
 		return 0;
 
-	page = compound_head(page);
-	delete_from_swap_cache(page);
-	SetPageDirty(page);
+	delete_from_swap_cache(&folio->page);
+	folio_set_dirty(folio);
 	return 1;
 }
 
-- 
2.35.1

[PATCH 13/22] mm/swap: Convert release_pages to use a folio internally

[Matthew Wilcox (Oracle)]

This function was already calling compound_head(), but now it can
cache the result of calling compound_head() and avoid calling it again.
Saves 299 bytes of text by avoiding various calls to compound_page()
and avoiding checks of PageTail.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 34 ++++++++++++++++------------------
 1 file changed, 16 insertions(+), 18 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index 5ea6ae795f0d..15ed7584ccb9 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -941,8 +941,7 @@ void release_pages(struct page **pages, int nr)
 	unsigned int lock_batch;
 
 	for (i = 0; i < nr; i++) {
-		struct page *page = pages[i];
-		struct folio *folio = page_folio(page);
+		struct folio *folio = page_folio(pages[i]);
 
 		/*
 		 * Make sure the IRQ-safe lock-holding time does not get
@@ -954,35 +953,34 @@ void release_pages(struct page **pages, int nr)
 			lruvec = NULL;
 		}
 
-		page = &folio->page;
-		if (is_huge_zero_page(page))
+		if (is_huge_zero_page(&folio->page))
 			continue;
 
-		if (is_zone_device_page(page)) {
+		if (folio_is_zone_device(folio)) {
 			if (lruvec) {
 				unlock_page_lruvec_irqrestore(lruvec, flags);
 				lruvec = NULL;
 			}
-			if (put_devmap_managed_page(page))
+			if (put_devmap_managed_page(&folio->page))
 				continue;
-			if (put_page_testzero(page))
-				free_zone_device_page(page);
+			if (folio_put_testzero(folio))
+				free_zone_device_page(&folio->page);
 			continue;
 		}
 
-		if (!put_page_testzero(page))
+		if (!folio_put_testzero(folio))
 			continue;
 
-		if (PageCompound(page)) {
+		if (folio_test_large(folio)) {
 			if (lruvec) {
 				unlock_page_lruvec_irqrestore(lruvec, flags);
 				lruvec = NULL;
 			}
-			__put_compound_page(page);
+			__put_compound_page(&folio->page);
 			continue;
 		}
 
-		if (PageLRU(page)) {
+		if (folio_test_lru(folio)) {
 			struct lruvec *prev_lruvec = lruvec;
 
 			lruvec = folio_lruvec_relock_irqsave(folio, lruvec,
@@ -990,8 +988,8 @@ void release_pages(struct page **pages, int nr)
 			if (prev_lruvec != lruvec)
 				lock_batch = 0;
 
-			del_page_from_lru_list(page, lruvec);
-			__clear_page_lru_flags(page);
+			lruvec_del_folio(lruvec, folio);
+			__folio_clear_lru_flags(folio);
 		}
 
 		/*
@@ -1000,13 +998,13 @@ void release_pages(struct page **pages, int nr)
 		 * found set here.  This does not indicate a problem, unless
 		 * "unevictable_pgs_cleared" appears worryingly large.
 		 */
-		if (unlikely(PageMlocked(page))) {
-			__ClearPageMlocked(page);
-			dec_zone_page_state(page, NR_MLOCK);
+		if (unlikely(folio_test_mlocked(folio))) {
+			__folio_clear_mlocked(folio);
+			zone_stat_sub_folio(folio, NR_MLOCK);
 			count_vm_event(UNEVICTABLE_PGCLEARED);
 		}
 
-		list_add(&page->lru, &pages_to_free);
+		list_add(&folio->lru, &pages_to_free);
 	}
 	if (lruvec)
 		unlock_page_lruvec_irqrestore(lruvec, flags);
-- 
2.35.1

[PATCH 14/22] mm/swap: Convert put_pages_list to use folios

[Matthew Wilcox (Oracle)]

Pages linked through the LRU list cannot be tail pages as ->compound_head
is in a union with one of the words of the list_head, and they cannot
be ZONE_DEVICE pages as ->pgmap is in a union with the same word.
Saves 60 bytes of text by removing a call to page_is_fake_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index 15ed7584ccb9..3696dfbff136 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -138,19 +138,19 @@ EXPORT_SYMBOL(__put_page);
  */
 void put_pages_list(struct list_head *pages)
 {
-	struct page *page, *next;
+	struct folio *folio, *next;
 
-	list_for_each_entry_safe(page, next, pages, lru) {
-		if (!put_page_testzero(page)) {
-			list_del(&page->lru);
+	list_for_each_entry_safe(folio, next, pages, lru) {
+		if (!folio_put_testzero(folio)) {
+			list_del(&folio->lru);
 			continue;
 		}
-		if (PageHead(page)) {
-			list_del(&page->lru);
-			__put_compound_page(page);
+		if (folio_test_large(folio)) {
+			list_del(&folio->lru);
+			__put_compound_page(&folio->page);
 			continue;
 		}
-		/* Cannot be PageLRU because it's passed to us using the lru */
+		/* LRU flag must be clear because it's passed using the lru */
 	}
 
 	free_unref_page_list(pages);
-- 
2.35.1

[PATCH 15/22] mm/swap: Convert __put_page() to __folio_put()

[Matthew Wilcox (Oracle)]

Saves 11 bytes of text by removing a check of PageTail.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 include/linux/mm.h   |  6 +++---
 mm/swap.c            | 14 +++++++-------
 net/core/page_pool.c |  2 +-
 3 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 07a1998c3621..764dc0fdae5e 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -886,7 +886,7 @@ static inline struct folio *virt_to_folio(const void *x)
 	return page_folio(page);
 }
 
-void __put_page(struct page *page);
+void __folio_put(struct folio *folio);
 
 void put_pages_list(struct list_head *pages);
 
@@ -1228,7 +1228,7 @@ static inline __must_check bool try_get_page(struct page *page)
 static inline void folio_put(struct folio *folio)
 {
 	if (folio_put_testzero(folio))
-		__put_page(&folio->page);
+		__folio_put(folio);
 }
 
 /**
@@ -1248,7 +1248,7 @@ static inline void folio_put(struct folio *folio)
 static inline void folio_put_refs(struct folio *folio, int refs)
 {
 	if (folio_ref_sub_and_test(folio, refs))
-		__put_page(&folio->page);
+		__folio_put(folio);
 }
 
 void release_pages(struct page **pages, int nr);
diff --git a/mm/swap.c b/mm/swap.c
index 3696dfbff136..0ab7a032a766 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -119,16 +119,16 @@ static void __put_compound_page(struct page *page)
 	destroy_compound_page(page);
 }
 
-void __put_page(struct page *page)
+void __folio_put(struct folio *folio)
 {
-	if (unlikely(is_zone_device_page(page)))
-		free_zone_device_page(page);
-	else if (unlikely(PageCompound(page)))
-		__put_compound_page(page);
+	if (unlikely(folio_is_zone_device(folio)))
+		free_zone_device_page(&folio->page);
+	else if (unlikely(folio_test_large(folio)))
+		__put_compound_page(&folio->page);
 	else
-		__put_single_page(page);
+		__put_single_page(&folio->page);
 }
-EXPORT_SYMBOL(__put_page);
+EXPORT_SYMBOL(__folio_put);
 
 /**
  * put_pages_list() - release a list of pages
diff --git a/net/core/page_pool.c b/net/core/page_pool.c
index f18e6e771993..db70e94c8df2 100644
--- a/net/core/page_pool.c
+++ b/net/core/page_pool.c
@@ -16,7 +16,7 @@
 #include <linux/dma-direction.h>
 #include <linux/dma-mapping.h>
 #include <linux/page-flags.h>
-#include <linux/mm.h> /* for __put_page() */
+#include <linux/mm.h> /* for put_page() */
 #include <linux/poison.h>
 #include <linux/ethtool.h>
 
-- 
2.35.1

[PATCH 16/22] mm/swap: Convert __put_single_page() to __folio_put_small()

[Matthew Wilcox (Oracle)]

Saves 56 bytes of text by removing a call to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index 0ab7a032a766..a7097eaceafb 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -99,11 +99,11 @@ static void __page_cache_release(struct page *page)
 	}
 }
 
-static void __put_single_page(struct page *page)
+static void __folio_put_small(struct folio *folio)
 {
-	__page_cache_release(page);
-	mem_cgroup_uncharge(page_folio(page));
-	free_unref_page(page, 0);
+	__page_cache_release(&folio->page);
+	mem_cgroup_uncharge(folio);
+	free_unref_page(&folio->page, 0);
 }
 
 static void __put_compound_page(struct page *page)
@@ -126,7 +126,7 @@ void __folio_put(struct folio *folio)
 	else if (unlikely(folio_test_large(folio)))
 		__put_compound_page(&folio->page);
 	else
-		__put_single_page(&folio->page);
+		__folio_put_small(folio);
 }
 EXPORT_SYMBOL(__folio_put);
 
-- 
2.35.1

[PATCH 17/22] mm/swap: Convert __put_compound_page() to __folio_put_large()

[Matthew Wilcox (Oracle)]

All the callers now have a folio, so pass it in.  This doesn't
save any text, but it does save a call to compound_head() as
folio_test_hugetlb() does not contain a call like PageHuge() does.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index a7097eaceafb..2570d18e71a5 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -106,7 +106,7 @@ static void __folio_put_small(struct folio *folio)
 	free_unref_page(&folio->page, 0);
 }
 
-static void __put_compound_page(struct page *page)
+static void __folio_put_large(struct folio *folio)
 {
 	/*
 	 * __page_cache_release() is supposed to be called for thp, not for
@@ -114,9 +114,9 @@ static void __put_compound_page(struct page *page)
 	 * (it's never listed to any LRU lists) and no memcg routines should
 	 * be called for hugetlb (it has a separate hugetlb_cgroup.)
 	 */
-	if (!PageHuge(page))
-		__page_cache_release(page);
-	destroy_compound_page(page);
+	if (!folio_test_hugetlb(folio))
+		__page_cache_release(&folio->page);
+	destroy_compound_page(&folio->page);
 }
 
 void __folio_put(struct folio *folio)
@@ -124,7 +124,7 @@ void __folio_put(struct folio *folio)
 	if (unlikely(folio_is_zone_device(folio)))
 		free_zone_device_page(&folio->page);
 	else if (unlikely(folio_test_large(folio)))
-		__put_compound_page(&folio->page);
+		__folio_put_large(folio);
 	else
 		__folio_put_small(folio);
 }
@@ -147,7 +147,7 @@ void put_pages_list(struct list_head *pages)
 		}
 		if (folio_test_large(folio)) {
 			list_del(&folio->lru);
-			__put_compound_page(&folio->page);
+			__folio_put_large(folio);
 			continue;
 		}
 		/* LRU flag must be clear because it's passed using the lru */
@@ -976,7 +976,7 @@ void release_pages(struct page **pages, int nr)
 				unlock_page_lruvec_irqrestore(lruvec, flags);
 				lruvec = NULL;
 			}
-			__put_compound_page(&folio->page);
+			__folio_put_large(folio);
 			continue;
 		}
 
-- 
2.35.1

[PATCH 18/22] mm/swap: Convert __page_cache_release() to use a folio

[Matthew Wilcox (Oracle)]

All the callers now have a folio.  Saves several calls to compound_head,
totalling 502 bytes of text.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.c | 33 ++++++++++++++++-----------------
 1 file changed, 16 insertions(+), 17 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index 2570d18e71a5..f309694d9f84 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -77,31 +77,30 @@ static DEFINE_PER_CPU(struct cpu_fbatches, cpu_fbatches) = {
  * This path almost never happens for VM activity - pages are normally freed
  * via pagevecs.  But it gets used by networking - and for compound pages.
  */
-static void __page_cache_release(struct page *page)
+static void __page_cache_release(struct folio *folio)
 {
-	if (PageLRU(page)) {
-		struct folio *folio = page_folio(page);
+	if (folio_test_lru(folio)) {
 		struct lruvec *lruvec;
 		unsigned long flags;
 
 		lruvec = folio_lruvec_lock_irqsave(folio, &flags);
-		del_page_from_lru_list(page, lruvec);
-		__clear_page_lru_flags(page);
+		lruvec_del_folio(lruvec, folio);
+		__folio_clear_lru_flags(folio);
 		unlock_page_lruvec_irqrestore(lruvec, flags);
 	}
-	/* See comment on PageMlocked in release_pages() */
-	if (unlikely(PageMlocked(page))) {
-		int nr_pages = thp_nr_pages(page);
+	/* See comment on folio_test_mlocked in release_pages() */
+	if (unlikely(folio_test_mlocked(folio))) {
+		long nr_pages = folio_nr_pages(folio);
 
-		__ClearPageMlocked(page);
-		mod_zone_page_state(page_zone(page), NR_MLOCK, -nr_pages);
+		__folio_clear_mlocked(folio);
+		zone_stat_mod_folio(folio, NR_MLOCK, -nr_pages);
 		count_vm_events(UNEVICTABLE_PGCLEARED, nr_pages);
 	}
 }
 
 static void __folio_put_small(struct folio *folio)
 {
-	__page_cache_release(&folio->page);
+	__page_cache_release(folio);
 	mem_cgroup_uncharge(folio);
 	free_unref_page(&folio->page, 0);
 }
@@ -115,7 +114,7 @@ static void __folio_put_large(struct folio *folio)
 	 * be called for hugetlb (it has a separate hugetlb_cgroup.)
 	 */
 	if (!folio_test_hugetlb(folio))
-		__page_cache_release(&folio->page);
+		__page_cache_release(folio);
 	destroy_compound_page(&folio->page);
 }
 
@@ -199,14 +198,14 @@ static void lru_add_fn(struct lruvec *lruvec, struct folio *folio)
 
 	/*
 	 * Is an smp_mb__after_atomic() still required here, before
-	 * folio_evictable() tests PageMlocked, to rule out the possibility
+	 * folio_evictable() tests the mlocked flag, to rule out the possibility
 	 * of stranding an evictable folio on an unevictable LRU?  I think
-	 * not, because __munlock_page() only clears PageMlocked while the LRU
-	 * lock is held.
+	 * not, because __munlock_page() only clears the mlocked flag
+	 * while the LRU lock is held.
 	 *
 	 * (That is not true of __page_cache_release(), and not necessarily
-	 * true of release_pages(): but those only clear PageMlocked after
-	 * put_page_testzero() has excluded any other users of the page.)
+	 * true of release_pages(): but those only clear the mlocked flag after
+	 * folio_put_testzero() has excluded any other users of the folio.)
 	 */
 	if (folio_evictable(folio)) {
 		if (was_unevictable)
-- 
2.35.1

[PATCH 19/22] mm: Convert destroy_compound_page() to destroy_large_folio()

[Matthew Wilcox (Oracle)]

All callers now have a folio, so push the folio->page conversion
down to this function.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 include/linux/mm.h | 8 +++++---
 mm/swap.c          | 2 +-
 mm/vmscan.c        | 4 ++--
 3 files changed, 8 insertions(+), 6 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 764dc0fdae5e..f99e49c4d36a 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -923,10 +923,12 @@ static inline void set_compound_page_dtor(struct page *page,
 	page[1].compound_dtor = compound_dtor;
 }
 
-static inline void destroy_compound_page(struct page *page)
+static inline void destroy_large_folio(struct folio *folio)
 {
-	VM_BUG_ON_PAGE(page[1].compound_dtor >= NR_COMPOUND_DTORS, page);
-	compound_page_dtors[page[1].compound_dtor](page);
+	enum compound_dtor_id dtor = folio_page(folio, 1)->compound_dtor;
+
+	VM_BUG_ON_FOLIO(dtor >= NR_COMPOUND_DTORS, folio);
+	compound_page_dtors[dtor](&folio->page);
 }
 
 static inline int head_compound_pincount(struct page *head)
diff --git a/mm/swap.c b/mm/swap.c
index f309694d9f84..9cee7f6a3809 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -115,7 +115,7 @@ static void __folio_put_large(struct folio *folio)
 	 */
 	if (!folio_test_hugetlb(folio))
 		__page_cache_release(folio);
-	destroy_compound_page(&folio->page);
+	destroy_large_folio(folio);
 }
 
 void __folio_put(struct folio *folio)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 6ba0fdf1dd0a..135f9306fe1e 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1985,7 +1985,7 @@ static unsigned int shrink_page_list(struct list_head *page_list,
 		 * appear not as the counts should be low
 		 */
 		if (unlikely(folio_test_large(folio)))
-			destroy_compound_page(&folio->page);
+			destroy_large_folio(folio);
 		else
 			list_add(&folio->lru, &free_pages);
 		continue;
@@ -2354,7 +2354,7 @@ static unsigned int move_pages_to_lru(struct lruvec *lruvec,
 
 			if (unlikely(folio_test_large(folio))) {
 				spin_unlock_irq(&lruvec->lru_lock);
-				destroy_compound_page(&folio->page);
+				destroy_large_folio(folio);
 				spin_lock_irq(&lruvec->lru_lock);
 			} else
 				list_add(&folio->lru, &folios_to_free);
-- 
2.35.1

[PATCH 20/22] mm: Convert page_swap_flags to folio_swap_flags

[Matthew Wilcox (Oracle)]

The only caller already has a folio, so push the folio->page conversion
down a level.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.h   | 6 +++---
 mm/vmscan.c | 2 +-
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/mm/swap.h b/mm/swap.h
index 0193797b0c92..85453110be8e 100644
--- a/mm/swap.h
+++ b/mm/swap.h
@@ -61,9 +61,9 @@ struct page *swap_cluster_readahead(swp_entry_t entry, gfp_t flag,
 struct page *swapin_readahead(swp_entry_t entry, gfp_t flag,
 			      struct vm_fault *vmf);
 
-static inline unsigned int page_swap_flags(struct page *page)
+static inline unsigned int folio_swap_flags(struct folio *folio)
 {
-	return page_swap_info(page)->flags;
+	return page_swap_info(&folio->page)->flags;
 }
 #else /* CONFIG_SWAP */
 struct swap_iocb;
@@ -149,7 +149,7 @@ static inline void clear_shadow_from_swap_cache(int type, unsigned long begin,
 {
 }
 
-static inline unsigned int page_swap_flags(struct page *page)
+static inline unsigned int folio_swap_flags(struct folio *folio)
 {
 	return 0;
 }
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 135f9306fe1e..1256d6b51b67 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1572,7 +1572,7 @@ static bool may_enter_fs(struct folio *folio, gfp_t gfp_mask)
 	 * but that will never affect SWP_FS_OPS, so the data_race
 	 * is safe.
 	 */
-	return !data_race(page_swap_flags(&folio->page) & SWP_FS_OPS);
+	return !data_race(folio_swap_flags(folio) & SWP_FS_OPS);
 }
 
 /*
-- 
2.35.1

[PATCH 21/22] mm/swap: Convert delete_from_swap_cache() to take a folio

[Matthew Wilcox (Oracle)]

All but one caller already has a folio, so convert it to use a folio.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/memory-failure.c |  5 +++--
 mm/shmem.c          |  4 ++--
 mm/swap.h           |  4 ++--
 mm/swap_state.c     | 16 ++++++++--------
 mm/swapfile.c       |  2 +-
 5 files changed, 16 insertions(+), 15 deletions(-)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index bc031939cf39..bd0e15b4876a 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -1049,12 +1049,13 @@ static int me_swapcache_dirty(struct page_state *ps, struct page *p)
 
 static int me_swapcache_clean(struct page_state *ps, struct page *p)
 {
+	struct folio *folio = page_folio(p);
 	int ret;
 
-	delete_from_swap_cache(p);
+	delete_from_swap_cache(folio);
 
 	ret = delete_from_lru_cache(p) ? MF_FAILED : MF_RECOVERED;
-	unlock_page(p);
+	folio_unlock(folio);
 
 	if (has_extra_refcount(ps, p, false))
 		ret = MF_FAILED;
diff --git a/mm/shmem.c b/mm/shmem.c
index 133c67057d41..7331ed1be014 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1690,7 +1690,7 @@ static void shmem_set_folio_swapin_error(struct inode *inode, pgoff_t index,
 		return;
 
 	folio_wait_writeback(folio);
-	delete_from_swap_cache(&folio->page);
+	delete_from_swap_cache(folio);
 	spin_lock_irq(&info->lock);
 	/*
 	 * Don't treat swapin error folio as alloced. Otherwise inode->i_blocks won't
@@ -1788,7 +1788,7 @@ static int shmem_swapin_folio(struct inode *inode, pgoff_t index,
 	if (sgp == SGP_WRITE)
 		folio_mark_accessed(folio);
 
-	delete_from_swap_cache(&folio->page);
+	delete_from_swap_cache(folio);
 	folio_mark_dirty(folio);
 	swap_free(swap);
 
diff --git a/mm/swap.h b/mm/swap.h
index 85453110be8e..3684f7f4eac9 100644
--- a/mm/swap.h
+++ b/mm/swap.h
@@ -38,7 +38,7 @@ int add_to_swap_cache(struct page *page, swp_entry_t entry,
 		      gfp_t gfp, void **shadowp);
 void __delete_from_swap_cache(struct page *page,
 			      swp_entry_t entry, void *shadow);
-void delete_from_swap_cache(struct page *page);
+void delete_from_swap_cache(struct folio *folio);
 void clear_shadow_from_swap_cache(int type, unsigned long begin,
 				  unsigned long end);
 void free_swap_cache(struct page *page);
@@ -140,7 +140,7 @@ static inline void __delete_from_swap_cache(struct page *page,
 {
 }
 
-static inline void delete_from_swap_cache(struct page *page)
+static inline void delete_from_swap_cache(struct folio *folio)
 {
 }
 
diff --git a/mm/swap_state.c b/mm/swap_state.c
index 41c6a6053d5c..b9464542c064 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -222,22 +222,22 @@ bool add_to_swap(struct folio *folio)
 }
 
 /*
- * This must be called only on pages that have
+ * This must be called only on folios that have
  * been verified to be in the swap cache and locked.
- * It will never put the page into the free list,
- * the caller has a reference on the page.
+ * It will never put the folio into the free list,
+ * the caller has a reference on the folio.
  */
-void delete_from_swap_cache(struct page *page)
+void delete_from_swap_cache(struct folio *folio)
 {
-	swp_entry_t entry = { .val = page_private(page) };
+	swp_entry_t entry = folio_swap_entry(folio);
 	struct address_space *address_space = swap_address_space(entry);
 
 	xa_lock_irq(&address_space->i_pages);
-	__delete_from_swap_cache(page, entry, NULL);
+	__delete_from_swap_cache(&folio->page, entry, NULL);
 	xa_unlock_irq(&address_space->i_pages);
 
-	put_swap_page(page, entry);
-	page_ref_sub(page, thp_nr_pages(page));
+	put_swap_page(&folio->page, entry);
+	folio_ref_sub(folio, folio_nr_pages(folio));
 }
 
 void clear_shadow_from_swap_cache(int type, unsigned long begin,
diff --git a/mm/swapfile.c b/mm/swapfile.c
index 4ac6a6c876d6..91eab0223d9e 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -1617,7 +1617,7 @@ int try_to_free_swap(struct page *page)
 	if (pm_suspended_storage())
 		return 0;
 
-	delete_from_swap_cache(&folio->page);
+	delete_from_swap_cache(folio);
 	folio_set_dirty(folio);
 	return 1;
 }
-- 
2.35.1

[PATCH 22/22] mm/swap: Convert __delete_from_swap_cache() to a folio

[Matthew Wilcox (Oracle)]

All callers now have a folio, so convert the entire function to operate
on folios.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/swap.h       |  4 ++--
 mm/swap_state.c | 25 +++++++++++++------------
 mm/vmscan.c     |  2 +-
 3 files changed, 16 insertions(+), 15 deletions(-)

diff --git a/mm/swap.h b/mm/swap.h
index 3684f7f4eac9..fa0816af4712 100644
--- a/mm/swap.h
+++ b/mm/swap.h
@@ -36,7 +36,7 @@ bool add_to_swap(struct folio *folio);
 void *get_shadow_from_swap_cache(swp_entry_t entry);
 int add_to_swap_cache(struct page *page, swp_entry_t entry,
 		      gfp_t gfp, void **shadowp);
-void __delete_from_swap_cache(struct page *page,
+void __delete_from_swap_cache(struct folio *folio,
 			      swp_entry_t entry, void *shadow);
 void delete_from_swap_cache(struct folio *folio);
 void clear_shadow_from_swap_cache(int type, unsigned long begin,
@@ -135,7 +135,7 @@ static inline int add_to_swap_cache(struct page *page, swp_entry_t entry,
 	return -1;
 }
 
-static inline void __delete_from_swap_cache(struct page *page,
+static inline void __delete_from_swap_cache(struct folio *folio,
 					swp_entry_t entry, void *shadow)
 {
 }
diff --git a/mm/swap_state.c b/mm/swap_state.c
index b9464542c064..e166051566f4 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -133,31 +133,32 @@ int add_to_swap_cache(struct page *page, swp_entry_t entry,
 }
 
 /*
- * This must be called only on pages that have
+ * This must be called only on folios that have
  * been verified to be in the swap cache.
  */
-void __delete_from_swap_cache(struct page *page,
+void __delete_from_swap_cache(struct folio *folio,
 			swp_entry_t entry, void *shadow)
 {
 	struct address_space *address_space = swap_address_space(entry);
-	int i, nr = thp_nr_pages(page);
+	int i;
+	long nr = folio_nr_pages(folio);
 	pgoff_t idx = swp_offset(entry);
 	XA_STATE(xas, &address_space->i_pages, idx);
 
-	VM_BUG_ON_PAGE(!PageLocked(page), page);
-	VM_BUG_ON_PAGE(!PageSwapCache(page), page);
-	VM_BUG_ON_PAGE(PageWriteback(page), page);
+	VM_BUG_ON_FOLIO(!folio_test_locked(folio), folio);
+	VM_BUG_ON_FOLIO(!folio_test_swapcache(folio), folio);
+	VM_BUG_ON_FOLIO(folio_test_writeback(folio), folio);
 
 	for (i = 0; i < nr; i++) {
 		void *entry = xas_store(&xas, shadow);
-		VM_BUG_ON_PAGE(entry != page, entry);
-		set_page_private(page + i, 0);
+		VM_BUG_ON_FOLIO(entry != folio, folio);
+		set_page_private(folio_page(folio, i), 0);
 		xas_next(&xas);
 	}
-	ClearPageSwapCache(page);
+	folio_clear_swapcache(folio);
 	address_space->nrpages -= nr;
-	__mod_node_page_state(page_pgdat(page), NR_FILE_PAGES, -nr);
-	__mod_lruvec_page_state(page, NR_SWAPCACHE, -nr);
+	__node_stat_mod_folio(folio, NR_FILE_PAGES, -nr);
+	__lruvec_stat_mod_folio(folio, NR_SWAPCACHE, -nr);
 }
 
 /**
@@ -233,7 +234,7 @@ void delete_from_swap_cache(struct folio *folio)
 	struct address_space *address_space = swap_address_space(entry);
 
 	xa_lock_irq(&address_space->i_pages);
-	__delete_from_swap_cache(&folio->page, entry, NULL);
+	__delete_from_swap_cache(folio, entry, NULL);
 	xa_unlock_irq(&address_space->i_pages);
 
 	put_swap_page(&folio->page, entry);
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 1256d6b51b67..6a5c34dee00a 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1329,7 +1329,7 @@ static int __remove_mapping(struct address_space *mapping, struct folio *folio,
 		mem_cgroup_swapout(folio, swap);
 		if (reclaimed && !mapping_exiting(mapping))
 			shadow = workingset_eviction(folio, target_memcg);
-		__delete_from_swap_cache(&folio->page, swap, shadow);
+		__delete_from_swap_cache(folio, swap, shadow);
 		xa_unlock_irq(&mapping->i_pages);
 		put_swap_page(&folio->page, swap);
 	} else {
-- 
2.35.1