btrfs: prepare compression folio alloc/free for bs > ps cases

This includes the following preparation for bs > ps cases:

- Always alloc/free the folio directly if bs > ps
  This adds a new @fs_info parameter for btrfs_alloc_compr_folio(), thus
  affecting all compression algorithms.

  For btrfs_free_compr_folio() it needs no parameter for now, as we can
  use the folio size to skip the caching part.

  For now the change is just to passing a @fs_info into the function,
  all the folio size assumption is still based on page size.

- Properly zero the last folio in compress_file_range()
  Since the compressed folios can be larger than a page, we need to
  properly zero the whole folio.

- Use correct folio size for btrfs_add_compressed_bio_folios()
  Instead of page size, use the correct folio size.

- Use correct folio size/shift for btrfs_compress_filemap_get_folio()
  As we are not only using simple page sized folios anymore.

- Use correct folio size for btrfs_decompress()
  There is an ASSERT() making sure the decompressed range is no larger
  than a page, which will be triggered for bs > ps cases.

- Skip readahead for compressed pages
  Similar to subpage cases.

- Make btrfs_alloc_folio_array() to accept a new @order parameter

- Add a helper to calculate the minimal folio size

All those changes should not affect the existing bs <= ps handling.

Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>

Author: adam900710

fs/btrfs/compression.c

@@ -223,10 +223,14 @@ static unsigned long btrfs_compr_pool_scan(struct shrinker *sh, struct shrink_co
 /*
  * Common wrappers for page allocation from compression wrappers
  */
-struct folio *btrfs_alloc_compr_folio(void)
+struct folio *btrfs_alloc_compr_folio(struct btrfs_fs_info *fs_info)
 {
 	struct folio *folio = NULL;
 
+	/* For bs > ps cases, no cached folio pool for now. */
+	if (fs_info->block_min_order)
+		goto alloc;
+
 	spin_lock(&compr_pool.lock);
 	if (compr_pool.count > 0) {
 		folio = list_first_entry(&compr_pool.list, struct folio, lru);
@@ -238,13 +242,18 @@ struct folio *btrfs_alloc_compr_folio(void)
 	if (folio)
 		return folio;
 
-	return folio_alloc(GFP_NOFS, 0);
+alloc:
+	return folio_alloc(GFP_NOFS, fs_info->block_min_order);
 }
 
 void btrfs_free_compr_folio(struct folio *folio)
 {
 	bool do_free = false;
 
+	/* The folio is from bs > ps fs, no cached pool for now. */
+	if (folio_order(folio))
+		goto free;
+
 	spin_lock(&compr_pool.lock);
 	if (compr_pool.count > compr_pool.thresh) {
 		do_free = true;
@@ -257,6 +266,7 @@ void btrfs_free_compr_folio(struct folio *folio)
 	if (!do_free)
 		return;
 
+free:
 	ASSERT(folio_ref_count(folio) == 1);
 	folio_put(folio);
 }
@@ -344,16 +354,19 @@ static void end_bbio_compressed_write(struct btrfs_bio *bbio)
 
 static void btrfs_add_compressed_bio_folios(struct compressed_bio *cb)
 {
+	struct btrfs_fs_info *fs_info = cb->bbio.fs_info;
 	struct bio *bio = &cb->bbio.bio;
 	u32 offset = 0;
 
 	while (offset < cb->compressed_len) {
+		struct folio *folio;
 		int ret;
-		u32 len = min_t(u32, cb->compressed_len - offset, PAGE_SIZE);
+		u32 len = min_t(u32, cb->compressed_len - offset,
+				btrfs_min_folio_size(fs_info));
 
+		folio = cb->compressed_folios[offset >> (PAGE_SHIFT + fs_info->block_min_order)];
 		/* Maximum compressed extent is smaller than bio size limit. */
-		ret = bio_add_folio(bio, cb->compressed_folios[offset >> PAGE_SHIFT],
-				    len, 0);
+		ret = bio_add_folio(bio, folio, len, 0);
 		ASSERT(ret);
 		offset += len;
 	}
@@ -443,6 +456,10 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 	if (fs_info->sectorsize < PAGE_SIZE)
 		return 0;
 
+	/* For bs > ps cases, we don't support readahead for compressed folios for now. */
+	if (fs_info->block_min_order)
+		return 0;
+
 	end_index = (i_size_read(inode) - 1) >> PAGE_SHIFT;
 
 	while (cur < compressed_end) {
@@ -606,14 +623,15 @@ void btrfs_submit_compressed_read(struct btrfs_bio *bbio)
 
 	btrfs_free_extent_map(em);
 
-	cb->nr_folios = DIV_ROUND_UP(compressed_len, PAGE_SIZE);
+	cb->nr_folios = DIV_ROUND_UP(compressed_len, btrfs_min_folio_size(fs_info));
 	cb->compressed_folios = kcalloc(cb->nr_folios, sizeof(struct folio *), GFP_NOFS);
 	if (!cb->compressed_folios) {
 		status = BLK_STS_RESOURCE;
 		goto out_free_bio;
 	}
 
-	ret = btrfs_alloc_folio_array(cb->nr_folios, cb->compressed_folios);
+	ret = btrfs_alloc_folio_array(cb->nr_folios, fs_info->block_min_order,
+				      cb->compressed_folios);
 	if (ret) {
 		status = BLK_STS_RESOURCE;
 		goto out_free_compressed_pages;
@@ -1033,12 +1051,12 @@ int btrfs_compress_filemap_get_folio(struct address_space *mapping, u64 start,
  * - compression algo are 0-3
  * - the level are bits 4-7
  *
- * @out_pages is an in/out parameter, holds maximum number of pages to allocate
- * and returns number of actually allocated pages
+ * @out_folios is an in/out parameter, holds maximum number of folios to allocate
+ * and returns number of actually allocated folios
  *
  * @total_in is used to return the number of bytes actually read.  It
  * may be smaller than the input length if we had to exit early because we
- * ran out of room in the pages array or because we cross the
+ * ran out of room in the folios array or because we cross the
  * max_out threshold.
  *
  * @total_out is an in/out parameter, must be set to the input length and will
@@ -1093,11 +1111,11 @@ int btrfs_decompress(int type, const u8 *data_in, struct folio *dest_folio,
 	int ret;
 
 	/*
-	 * The full destination page range should not exceed the page size.
+	 * The full destination folio range should not exceed the folio size.
 	 * And the @destlen should not exceed sectorsize, as this is only called for
 	 * inline file extents, which should not exceed sectorsize.
 	 */
-	ASSERT(dest_pgoff + destlen <= PAGE_SIZE && destlen <= sectorsize);
+	ASSERT(dest_pgoff + destlen <= folio_size(dest_folio) && destlen <= sectorsize);
 
 	workspace = get_workspace(fs_info, type, 0);
 	ret = compression_decompress(type, workspace, data_in, dest_folio,

fs/btrfs/compression.h

@@ -112,7 +112,7 @@ void btrfs_submit_compressed_read(struct btrfs_bio *bbio);
 
 int btrfs_compress_str2level(unsigned int type, const char *str, int *level_ret);
 
-struct folio *btrfs_alloc_compr_folio(void);
+struct folio *btrfs_alloc_compr_folio(struct btrfs_fs_info *fs_info);
 void btrfs_free_compr_folio(struct folio *folio);
 
 struct workspace_manager {

fs/btrfs/extent_io.c

@@ -626,19 +626,21 @@ static void end_bbio_data_read(struct btrfs_bio *bbio)
  * Populate every free slot in a provided array with folios using GFP_NOFS.
  *
  * @nr_folios:   number of folios to allocate
+ * @order:	 the order of the folios to be allocated
  * @folio_array: the array to fill with folios; any existing non-NULL entries in
  *		 the array will be skipped
  *
  * Return: 0        if all folios were able to be allocated;
  *         -ENOMEM  otherwise, the partially allocated folios would be freed and
  *                  the array slots zeroed
  */
-int btrfs_alloc_folio_array(unsigned int nr_folios, struct folio **folio_array)
+int btrfs_alloc_folio_array(unsigned int nr_folios, unsigned int order,
+			    struct folio **folio_array)
 {
 	for (int i = 0; i < nr_folios; i++) {
 		if (folio_array[i])
 			continue;
-		folio_array[i] = folio_alloc(GFP_NOFS, 0);
+		folio_array[i] = folio_alloc(GFP_NOFS, order);
 		if (!folio_array[i])
 			goto error;
 	}
@@ -647,6 +649,7 @@ int btrfs_alloc_folio_array(unsigned int nr_folios, struct folio **folio_array)
 	for (int i = 0; i < nr_folios; i++) {
 		if (folio_array[i])
 			folio_put(folio_array[i]);
+		folio_array[i] = NULL;
 	}
 	return -ENOMEM;
 }

fs/btrfs/extent_io.h

@@ -366,7 +366,8 @@ void btrfs_clear_buffer_dirty(struct btrfs_trans_handle *trans,
 
 int btrfs_alloc_page_array(unsigned int nr_pages, struct page **page_array,
 			   bool nofail);
-int btrfs_alloc_folio_array(unsigned int nr_folios, struct folio **folio_array);
+int btrfs_alloc_folio_array(unsigned int nr_folios, unsigned int order,
+			    struct folio **folio_array);
 
 #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
 bool find_lock_delalloc_range(struct inode *inode,

fs/btrfs/fs.h

@@ -923,6 +923,12 @@ static inline gfp_t btrfs_alloc_write_mask(struct address_space *mapping)
 	return mapping_gfp_constraint(mapping, ~__GFP_FS);
 }
 
+/* Return the minimal folio size of the fs. */
+static inline unsigned int btrfs_min_folio_size(struct btrfs_fs_info *fs_info)
+{
+	return 1U << (PAGE_SHIFT + fs_info->block_min_order);
+}
+
 static inline u64 btrfs_get_fs_generation(const struct btrfs_fs_info *fs_info)
 {
 	return READ_ONCE(fs_info->generation);

fs/btrfs/inode.c

@@ -854,6 +854,8 @@ static void compress_file_range(struct btrfs_work *work)
 	struct btrfs_inode *inode = async_chunk->inode;
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct address_space *mapping = inode->vfs_inode.i_mapping;
+	const u32 min_folio_shift = PAGE_SHIFT + fs_info->block_min_order;
+	const u32 min_folio_size = btrfs_min_folio_size(fs_info);
 	u64 blocksize = fs_info->sectorsize;
 	u64 start = async_chunk->start;
 	u64 end = async_chunk->end;
@@ -864,7 +866,7 @@ static void compress_file_range(struct btrfs_work *work)
 	unsigned long nr_folios;
 	unsigned long total_compressed = 0;
 	unsigned long total_in = 0;
-	unsigned int poff;
+	unsigned int loff;
 	int i;
 	int compress_type = fs_info->compress_type;
 	int compress_level = fs_info->compress_level;
@@ -902,8 +904,8 @@ static void compress_file_range(struct btrfs_work *work)
 	actual_end = min_t(u64, i_size, end + 1);
 again:
 	folios = NULL;
-	nr_folios = (end >> PAGE_SHIFT) - (start >> PAGE_SHIFT) + 1;
-	nr_folios = min_t(unsigned long, nr_folios, BTRFS_MAX_COMPRESSED_PAGES);
+	nr_folios = (end >> min_folio_shift) - (start >> min_folio_shift) + 1;
+	nr_folios = min_t(unsigned long, nr_folios, BTRFS_MAX_COMPRESSED >> min_folio_shift);
 
 	/*
 	 * we don't want to send crud past the end of i_size through
@@ -965,12 +967,12 @@ static void compress_file_range(struct btrfs_work *work)
 		goto mark_incompressible;
 
 	/*
-	 * Zero the tail end of the last page, as we might be sending it down
+	 * Zero the tail end of the last folio, as we might be sending it down
 	 * to disk.
 	 */
-	poff = offset_in_page(total_compressed);
-	if (poff)
-		folio_zero_range(folios[nr_folios - 1], poff, PAGE_SIZE - poff);
+	loff = (total_compressed & (min_folio_size - 1));
+	if (loff)
+		folio_zero_range(folios[nr_folios - 1], loff, min_folio_size - loff);
 
 	/*
 	 * Try to create an inline extent.

fs/btrfs/lzo.c

@@ -132,13 +132,14 @@ static inline size_t read_compress_length(const char *buf)
  *
  * Will allocate new pages when needed.
  */
-static int copy_compressed_data_to_page(char *compressed_data,
+static int copy_compressed_data_to_page(struct btrfs_fs_info *fs_info,
+					char *compressed_data,
 					size_t compressed_size,
 					struct folio **out_folios,
 					unsigned long max_nr_folio,
-					u32 *cur_out,
-					const u32 sectorsize)
+					u32 *cur_out)
 {
+	const u32 sectorsize = fs_info->sectorsize;
 	u32 sector_bytes_left;
 	u32 orig_out;
 	struct folio *cur_folio;
@@ -156,7 +157,7 @@ static int copy_compressed_data_to_page(char *compressed_data,
 	cur_folio = out_folios[*cur_out / PAGE_SIZE];
 	/* Allocate a new page */
 	if (!cur_folio) {
-		cur_folio = btrfs_alloc_compr_folio();
+		cur_folio = btrfs_alloc_compr_folio(fs_info);
 		if (!cur_folio)
 			return -ENOMEM;
 		out_folios[*cur_out / PAGE_SIZE] = cur_folio;
@@ -182,7 +183,7 @@ static int copy_compressed_data_to_page(char *compressed_data,
 		cur_folio = out_folios[*cur_out / PAGE_SIZE];
 		/* Allocate a new page */
 		if (!cur_folio) {
-			cur_folio = btrfs_alloc_compr_folio();
+			cur_folio = btrfs_alloc_compr_folio(fs_info);
 			if (!cur_folio)
 				return -ENOMEM;
 			out_folios[*cur_out / PAGE_SIZE] = cur_folio;
@@ -217,8 +218,9 @@ int lzo_compress_folios(struct list_head *ws, struct btrfs_inode *inode,
 			u64 start, struct folio **folios, unsigned long *out_folios,
 			unsigned long *total_in, unsigned long *total_out)
 {
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
-	const u32 sectorsize = inode->root->fs_info->sectorsize;
+	const u32 sectorsize = fs_info->sectorsize;
 	struct address_space *mapping = inode->vfs_inode.i_mapping;
 	struct folio *folio_in = NULL;
 	char *sizes_ptr;
@@ -268,9 +270,9 @@ int lzo_compress_folios(struct list_head *ws, struct btrfs_inode *inode,
 			goto out;
 		}
 
-		ret = copy_compressed_data_to_page(workspace->cbuf, out_len,
+		ret = copy_compressed_data_to_page(fs_info, workspace->cbuf, out_len,
 						   folios, max_nr_folio,
-						   &cur_out, sectorsize);
+						   &cur_out);
 		if (ret < 0)
 			goto out;
 

fs/btrfs/zlib.c

@@ -136,6 +136,7 @@ int zlib_compress_folios(struct list_head *ws, struct btrfs_inode *inode,
 			 u64 start, struct folio **folios, unsigned long *out_folios,
 			 unsigned long *total_in, unsigned long *total_out)
 {
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct workspace *workspace = list_entry(ws, struct workspace, list);
 	struct address_space *mapping = inode->vfs_inode.i_mapping;
 	int ret;
@@ -147,7 +148,7 @@ int zlib_compress_folios(struct list_head *ws, struct btrfs_inode *inode,
 	unsigned long len = *total_out;
 	unsigned long nr_dest_folios = *out_folios;
 	const unsigned long max_out = nr_dest_folios * PAGE_SIZE;
-	const u32 blocksize = inode->root->fs_info->sectorsize;
+	const u32 blocksize = fs_info->sectorsize;
 	const u64 orig_end = start + len;
 
 	*out_folios = 0;
@@ -156,7 +157,7 @@ int zlib_compress_folios(struct list_head *ws, struct btrfs_inode *inode,
 
 	ret = zlib_deflateInit(&workspace->strm, workspace->level);
 	if (unlikely(ret != Z_OK)) {
-		btrfs_err(inode->root->fs_info,
+		btrfs_err(fs_info,
 	"zlib compression init failed, error %d root %llu inode %llu offset %llu",
 			  ret, btrfs_root_id(inode->root), btrfs_ino(inode), start);
 		ret = -EIO;
@@ -166,7 +167,7 @@ int zlib_compress_folios(struct list_head *ws, struct btrfs_inode *inode,
 	workspace->strm.total_in = 0;
 	workspace->strm.total_out = 0;
 
-	out_folio = btrfs_alloc_compr_folio();
+	out_folio = btrfs_alloc_compr_folio(fs_info);
 	if (out_folio == NULL) {
 		ret = -ENOMEM;
 		goto out;
@@ -224,7 +225,7 @@ int zlib_compress_folios(struct list_head *ws, struct btrfs_inode *inode,
 
 		ret = zlib_deflate(&workspace->strm, Z_SYNC_FLUSH);
 		if (unlikely(ret != Z_OK)) {
-			btrfs_warn(inode->root->fs_info,
+			btrfs_warn(fs_info,
 		"zlib compression failed, error %d root %llu inode %llu offset %llu",
 				   ret, btrfs_root_id(inode->root), btrfs_ino(inode),
 				   start);
@@ -249,7 +250,7 @@ int zlib_compress_folios(struct list_head *ws, struct btrfs_inode *inode,
 				ret = -E2BIG;
 				goto out;
 			}
-			out_folio = btrfs_alloc_compr_folio();
+			out_folio = btrfs_alloc_compr_folio(fs_info);
 			if (out_folio == NULL) {
 				ret = -ENOMEM;
 				goto out;
@@ -285,7 +286,7 @@ int zlib_compress_folios(struct list_head *ws, struct btrfs_inode *inode,
 				ret = -E2BIG;
 				goto out;
 			}
-			out_folio = btrfs_alloc_compr_folio();
+			out_folio = btrfs_alloc_compr_folio(fs_info);
 			if (out_folio == NULL) {
 				ret = -ENOMEM;
 				goto out;