mm: split a folio in minimum folio order chunks20240423-lbs-works

Set new_order to be at least minimum folio order if it is set in
split_huge_page_to_list() so that we can maintain minimum folio order
requirement in the page cache.

Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

3 files changed, 60 insertions, 15 deletions

diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
index de0c891050769..06748a8fa43be 100644
--- a/include/linux/huge_mm.h
+++ b/include/linux/huge_mm.h
@@ -87,6 +87,8 @@ extern struct kobj_attribute shmem_enabled_attr;
 #define thp_vma_allowable_order(vma, vm_flags, smaps, in_pf, enforce_sysfs, order) \
 	(!!thp_vma_allowable_orders(vma, vm_flags, smaps, in_pf, enforce_sysfs, BIT(order)))
 
+#define split_folio(f) split_folio_to_list(f, NULL)
+
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 #define HPAGE_PMD_SHIFT PMD_SHIFT
 #define HPAGE_PMD_SIZE	((1UL) << HPAGE_PMD_SHIFT)
@@ -267,9 +269,10 @@ void folio_prep_large_rmappable(struct folio *folio);
 bool can_split_folio(struct folio *folio, int *pextra_pins);
 int split_huge_page_to_list_to_order(struct page *page, struct list_head *list,
 		unsigned int new_order);
+int split_folio_to_list(struct folio *folio, struct list_head *list);
 static inline int split_huge_page(struct page *page)
 {
-	return split_huge_page_to_list_to_order(page, NULL, 0);
+	return split_folio(page_folio(page));
 }
 void deferred_split_folio(struct folio *folio);
 
@@ -432,6 +435,10 @@ static inline int split_huge_page(struct page *page)
 {
 	return 0;
 }
+static inline int split_folio_to_list(struct page *page, struct list_head *list)
+{
+	return 0;
+}
 static inline void deferred_split_folio(struct folio *folio) {}
 #define split_huge_pmd(__vma, __pmd, __address)	\
 	do { } while (0)
@@ -532,9 +539,6 @@ static inline int split_folio_to_order(struct folio *folio, int new_order)
 	return split_folio_to_list_to_order(folio, NULL, new_order);
 }
 
-#define split_folio_to_list(f, l) split_folio_to_list_to_order(f, l, 0)
-#define split_folio(f) split_folio_to_order(f, 0)
-
 /*
  * archs that select ARCH_WANTS_THP_SWAP but don't support THP_SWP due to
  * limitations in the implementation like arm64 MTE can override this to
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index e5ec74ca57cd3..0119f05ca3d95 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -3035,6 +3035,9 @@ bool can_split_folio(struct folio *folio, int *pextra_pins)
  * Returns 0 if the hugepage is split successfully.
  * Returns -EBUSY if the page is pinned or if anon_vma disappeared from under
  * us.
+ *
+ * Callers should ensure that the order respects the address space mapping
+ * min-order if one is set.
  */
 int split_huge_page_to_list_to_order(struct page *page, struct list_head *list,
 				     unsigned int new_order)
@@ -3055,6 +3058,27 @@ int split_huge_page_to_list_to_order(struct page *page, struct list_head *list,
 	if (new_order >= folio_order(folio))
 		return -EINVAL;
 
+	if (folio_test_writeback(folio))
+		return -EBUSY;
+
+	if (!folio_test_anon(folio)) {
+		unsigned int min_order;
+
+		/* Truncated ? */
+		if (!folio->mapping) {
+			ret = -EBUSY;
+			goto out;
+		}
+
+		min_order = mapping_min_folio_order(folio->mapping);
+		if (new_order < min_order) {
+			VM_WARN_ONCE(1, "Cannot split mapped folio below min-order: %u",
+				     min_order);
+			ret = -EINVAL;
+			goto out;
+		}
+	}
+
 	/* Cannot split anonymous THP to order-1 */
 	if (new_order == 1 && folio_test_anon(folio)) {
 		VM_WARN_ONCE(1, "Cannot split to order-1 folio");
@@ -3086,9 +3110,6 @@ int split_huge_page_to_list_to_order(struct page *page, struct list_head *list,
 		return -EBUSY;
 	}
 
-	if (folio_test_writeback(folio))
-		return -EBUSY;
-
 	if (folio_test_anon(folio)) {
 		/*
 		 * The caller does not necessarily hold an mmap_lock that would
@@ -3111,12 +3132,6 @@ int split_huge_page_to_list_to_order(struct page *page, struct list_head *list,
 
 		mapping = folio->mapping;
 
-		/* Truncated ? */
-		if (!mapping) {
-			ret = -EBUSY;
-			goto out;
-		}
-
 		/*
 		 * Do not split if mapping has minimum folio order
 		 * requirement.
@@ -3198,6 +3213,7 @@ int split_huge_page_to_list_to_order(struct page *page, struct list_head *list,
 			int nr = folio_nr_pages(folio);
 
 			xas_split(&xas, folio, folio_order(folio));
+			BUG_ON(folio_test_hugetlb(folio));
 			if (folio_test_pmd_mappable(folio) &&
 			    new_order < HPAGE_PMD_ORDER) {
 				if (folio_test_swapbacked(folio)) {
@@ -3236,6 +3252,16 @@ out:
 	return ret;
 }
 
+int split_folio_to_list(struct folio *folio, struct list_head *list)
+{
+	unsigned int min_order = 0;
+
+	if (!folio_test_anon(folio))
+		min_order = mapping_min_folio_order(folio->mapping);
+
+	return split_huge_page_to_list_to_order(&folio->page, list, min_order);
+}
+
 void folio_undo_large_rmappable(struct folio *folio)
 {
 	struct deferred_split *ds_queue;
@@ -3498,6 +3524,15 @@ static int split_huge_pages_pid(int pid, unsigned long vaddr_start,
 		if (new_order >= folio_order(folio))
 			goto next;
 
+		if (!folio_test_anon(folio)) {
+			unsigned int min_order = mapping_min_folio_order(folio->mapping);
+			if (min_order > new_order) {
+				pr_debug("cannot split below min_order: %u\n",
+					 min_order);
+				goto next;
+			}
+		}
+
 		total++;
 		/*
 		 * For folios with private, split_huge_page_to_list_to_order()
@@ -3538,6 +3573,7 @@ static int split_huge_pages_in_file(const char *file_path, pgoff_t off_start,
 	pgoff_t index, max_idx;
 	int nr_pages = 1;
 	unsigned long total = 0, split = 0;
+	unsigned int min_order;
 
 	file = getname_kernel(file_path);
 	if (IS_ERR(file))
@@ -3552,8 +3588,12 @@ static int split_huge_pages_in_file(const char *file_path, pgoff_t off_start,
 	if (off_end > max_idx)
 		off_end = max_idx;
 
-	pr_debug("split file-backed THPs in file: %s, page offset: [0x%lx - 0x%lx]\n",
-		 file_path, off_start, off_end);
+	min_order = mapping_min_folio_order(mapping);
+	if (new_order < min_order)
+		new_order = min_order;
+
+	pr_debug("split file-backed THPs in file: %s, page offset: [0x%lx - 0x%lx] with order: %u\n",
+		 file_path, off_start, off_end, new_order);
 
 	for (index = off_start; index < off_end; index += nr_pages) {
 		struct folio *folio = filemap_get_folio(mapping, index);
diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index 3e19b87049db1..2c554ce039ab3 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -2385,6 +2385,7 @@ static bool folio_prepare_writeback(struct address_space *mapping,
 		folio_wait_writeback(folio);
 	}
 	BUG_ON(folio_test_writeback(folio));
+	BUG_ON(folio->mapping != mapping);
 
 	if (!folio_clear_dirty_for_io(folio))
 		return false;