[PATCHSET v2 0/4] Submit ->readpages() IO as read-ahead

[PATCHSET v2 0/4] Submit ->readpages() IO as read-ahead

[Jens Axboe]

The only caller of ->readpages() is from read-ahead, yet we don't
submit IO flagged with REQ_RAHEAD. This means we don't see it in
blktrace, for instance, which is a shame. We already make assumptions
about ->readpages() just being for read-ahead in the mpage
implementation, using readahead_gfp_mask(mapping) as out GFP mask of
choice.

This small series fixes up mpage_readpages() to submit with
REQ_RAHEAD, which takes care of file systems using mpage_readpages().
The first patch is a prep patch, that makes do_mpage_readpage() take
an argument structure.

Changes since v1:

- Fix ext4_mpage_readpages() also being used for regular reads
- Add prep patch with struct arguments for do_mpage_readpage()

 fs/btrfs/extent_io.c |   2 +-
 fs/ext4/ext4.h       |   2 +-
 fs/ext4/inode.c      |   5 +-
 fs/ext4/readpage.c   |   5 +-
 fs/mpage.c           | 109 ++++++++++++++++++++++---------------------
 5 files changed, 65 insertions(+), 58 deletions(-)

-- 
Jens Axboe

[PATCH 1/4] mpage: add argument structure for do_mpage_readpage()

[Jens Axboe]

We're currently passing 8 arguments to this function, clean it up a
bit by packing the arguments in an args structure we pass to it.

Not intentional functional changes in this patch.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/mpage.c | 106 +++++++++++++++++++++++++++++++------------------------------
 1 file changed, 54 insertions(+), 52 deletions(-)

diff --git a/fs/mpage.c b/fs/mpage.c
index b7e7f570733a..0ecac5c410f4 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -133,6 +133,17 @@ map_buffer_to_page(struct page *page, struct buffer_head *bh, int page_block)
 	} while (page_bh != head);
 }
 
+struct mpage_readpage_args {
+	struct bio *bio;
+	struct page *page;
+	unsigned nr_pages;
+	sector_t last_block_in_bio;
+	struct buffer_head map_bh;
+	unsigned long first_logical_block;
+	get_block_t *get_block;
+	gfp_t gfp;
+};
+
 /*
  * This is the worker routine which does all the work of mapping the disk
  * blocks and constructs largest possible bios, submits them for IO if the
@@ -142,16 +153,14 @@ map_buffer_to_page(struct page *page, struct buffer_head *bh, int page_block)
  * represent the validity of its disk mapping and to decide when to do the next
  * get_block() call.
  */
-static struct bio *
-do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
-		sector_t *last_block_in_bio, struct buffer_head *map_bh,
-		unsigned long *first_logical_block, get_block_t get_block,
-		gfp_t gfp)
+static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 {
+	struct page *page = args->page;
 	struct inode *inode = page->mapping->host;
 	const unsigned blkbits = inode->i_blkbits;
 	const unsigned blocks_per_page = PAGE_SIZE >> blkbits;
 	const unsigned blocksize = 1 << blkbits;
+	struct buffer_head *map_bh = &args->map_bh;
 	sector_t block_in_file;
 	sector_t last_block;
 	sector_t last_block_in_file;
@@ -168,7 +177,7 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 		goto confused;
 
 	block_in_file = (sector_t)page->index << (PAGE_SHIFT - blkbits);
-	last_block = block_in_file + nr_pages * blocks_per_page;
+	last_block = block_in_file + args->nr_pages * blocks_per_page;
 	last_block_in_file = (i_size_read(inode) + blocksize - 1) >> blkbits;
 	if (last_block > last_block_in_file)
 		last_block = last_block_in_file;
@@ -178,9 +187,9 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 	 * Map blocks using the result from the previous get_blocks call first.
 	 */
 	nblocks = map_bh->b_size >> blkbits;
-	if (buffer_mapped(map_bh) && block_in_file > *first_logical_block &&
-			block_in_file < (*first_logical_block + nblocks)) {
-		unsigned map_offset = block_in_file - *first_logical_block;
+	if (buffer_mapped(map_bh) && block_in_file > args->first_logical_block &&
+			block_in_file < (args->first_logical_block + nblocks)) {
+		unsigned map_offset = block_in_file - args->first_logical_block;
 		unsigned last = nblocks - map_offset;
 
 		for (relative_block = 0; ; relative_block++) {
@@ -208,9 +217,9 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 
 		if (block_in_file < last_block) {
 			map_bh->b_size = (last_block-block_in_file) << blkbits;
-			if (get_block(inode, block_in_file, map_bh, 0))
+			if (args->get_block(inode, block_in_file, map_bh, 0))
 				goto confused;
-			*first_logical_block = block_in_file;
+			args->first_logical_block = block_in_file;
 		}
 
 		if (!buffer_mapped(map_bh)) {
@@ -273,43 +282,43 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 	/*
 	 * This page will go to BIO.  Do we need to send this BIO off first?
 	 */
-	if (bio && (*last_block_in_bio != blocks[0] - 1))
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (args->bio && (args->last_block_in_bio != blocks[0] - 1))
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 
 alloc_new:
-	if (bio == NULL) {
+	if (args->bio == NULL) {
 		if (first_hole == blocks_per_page) {
 			if (!bdev_read_page(bdev, blocks[0] << (blkbits - 9),
 								page))
 				goto out;
 		}
-		bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
-				min_t(int, nr_pages, BIO_MAX_PAGES), gfp);
-		if (bio == NULL)
+		args->bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
+				min_t(int, args->nr_pages, BIO_MAX_PAGES), args->gfp);
+		if (args->bio == NULL)
 			goto confused;
 	}
 
 	length = first_hole << blkbits;
-	if (bio_add_page(bio, page, length, 0) < length) {
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (bio_add_page(args->bio, page, length, 0) < length) {
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 		goto alloc_new;
 	}
 
-	relative_block = block_in_file - *first_logical_block;
+	relative_block = block_in_file - args->first_logical_block;
 	nblocks = map_bh->b_size >> blkbits;
 	if ((buffer_boundary(map_bh) && relative_block == nblocks) ||
 	    (first_hole != blocks_per_page))
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 	else
-		*last_block_in_bio = blocks[blocks_per_page - 1];
+		args->last_block_in_bio = blocks[blocks_per_page - 1];
 out:
-	return bio;
+	return args->bio;
 
 confused:
-	if (bio)
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (args->bio)
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 	if (!PageUptodate(page))
-	        block_read_full_page(page, get_block);
+	        block_read_full_page(page, args->get_block);
 	else
 		unlock_page(page);
 	goto out;
@@ -363,15 +372,12 @@ int
 mpage_readpages(struct address_space *mapping, struct list_head *pages,
 				unsigned nr_pages, get_block_t get_block)
 {
-	struct bio *bio = NULL;
+	struct mpage_readpage_args args = {
+		.get_block = get_block,
+		.gfp = readahead_gfp_mask(mapping),
+	};
 	unsigned page_idx;
-	sector_t last_block_in_bio = 0;
-	struct buffer_head map_bh;
-	unsigned long first_logical_block = 0;
-	gfp_t gfp = readahead_gfp_mask(mapping);
 
-	map_bh.b_state = 0;
-	map_bh.b_size = 0;
 	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
 		struct page *page = lru_to_page(pages);
 
@@ -379,18 +385,16 @@ mpage_readpages(struct address_space *mapping, struct list_head *pages,
 		list_del(&page->lru);
 		if (!add_to_page_cache_lru(page, mapping,
 					page->index,
-					gfp)) {
-			bio = do_mpage_readpage(bio, page,
-					nr_pages - page_idx,
-					&last_block_in_bio, &map_bh,
-					&first_logical_block,
-					get_block, gfp);
+					args.gfp)) {
+			args.page = page;
+			args.nr_pages = nr_pages - page_idx;
+			args.bio = do_mpage_readpage(&args);
 		}
 		put_page(page);
 	}
 	BUG_ON(!list_empty(pages));
-	if (bio)
-		mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (args.bio)
+		mpage_bio_submit(REQ_OP_READ, 0, args.bio);
 	return 0;
 }
 EXPORT_SYMBOL(mpage_readpages);
@@ -400,18 +404,16 @@ EXPORT_SYMBOL(mpage_readpages);
  */
 int mpage_readpage(struct page *page, get_block_t get_block)
 {
-	struct bio *bio = NULL;
-	sector_t last_block_in_bio = 0;
-	struct buffer_head map_bh;
-	unsigned long first_logical_block = 0;
-	gfp_t gfp = mapping_gfp_constraint(page->mapping, GFP_KERNEL);
+	struct mpage_readpage_args args = {
+		.page = page,
+		.nr_pages = 1,
+		.get_block = get_block,
+		.gfp = mapping_gfp_constraint(page->mapping, GFP_KERNEL),
+	};
 
-	map_bh.b_state = 0;
-	map_bh.b_size = 0;
-	bio = do_mpage_readpage(bio, page, 1, &last_block_in_bio,
-			&map_bh, &first_logical_block, get_block, gfp);
-	if (bio)
-		mpage_bio_submit(REQ_OP_READ, 0, bio);
+	args.bio = do_mpage_readpage(&args);
+	if (args.bio)
+		mpage_bio_submit(REQ_OP_READ, 0, args.bio);
 	return 0;
 }
 EXPORT_SYMBOL(mpage_readpage);
-- 
2.7.4

[PATCH 2/4] mpage: mpage_readpages() should submit IO as read-ahead

[Jens Axboe]

a_ops->readpages() is only ever used for read-ahead, yet we don't
flag the IO being submitted as such. Fix that up. Any file system
that uses mpage_readpages() as it's ->readpages() implementation
will now get this right.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/mpage.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/fs/mpage.c b/fs/mpage.c
index 0ecac5c410f4..a6344996f924 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -142,6 +142,7 @@ struct mpage_readpage_args {
 	unsigned long first_logical_block;
 	get_block_t *get_block;
 	gfp_t gfp;
+	bool is_readahead;
 };
 
 /*
@@ -170,6 +171,7 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 	struct block_device *bdev = NULL;
 	int length;
 	int fully_mapped = 1;
+	int op_flags = args->is_readahead ? REQ_RAHEAD : 0;
 	unsigned nblocks;
 	unsigned relative_block;
 
@@ -283,7 +285,7 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 	 * This page will go to BIO.  Do we need to send this BIO off first?
 	 */
 	if (args->bio && (args->last_block_in_bio != blocks[0] - 1))
-		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
+		args->bio = mpage_bio_submit(REQ_OP_READ, op_flags, args->bio);
 
 alloc_new:
 	if (args->bio == NULL) {
@@ -300,7 +302,7 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 
 	length = first_hole << blkbits;
 	if (bio_add_page(args->bio, page, length, 0) < length) {
-		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
+		args->bio = mpage_bio_submit(REQ_OP_READ, op_flags, args->bio);
 		goto alloc_new;
 	}
 
@@ -308,7 +310,7 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 	nblocks = map_bh->b_size >> blkbits;
 	if ((buffer_boundary(map_bh) && relative_block == nblocks) ||
 	    (first_hole != blocks_per_page))
-		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
+		args->bio = mpage_bio_submit(REQ_OP_READ, op_flags, args->bio);
 	else
 		args->last_block_in_bio = blocks[blocks_per_page - 1];
 out:
@@ -316,7 +318,7 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 
 confused:
 	if (args->bio)
-		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
+		args->bio = mpage_bio_submit(REQ_OP_READ, op_flags, args->bio);
 	if (!PageUptodate(page))
 	        block_read_full_page(page, args->get_block);
 	else
@@ -375,6 +377,7 @@ mpage_readpages(struct address_space *mapping, struct list_head *pages,
 	struct mpage_readpage_args args = {
 		.get_block = get_block,
 		.gfp = readahead_gfp_mask(mapping),
+		.is_readahead = true,
 	};
 	unsigned page_idx;
 
@@ -394,7 +397,7 @@ mpage_readpages(struct address_space *mapping, struct list_head *pages,
 	}
 	BUG_ON(!list_empty(pages));
 	if (args.bio)
-		mpage_bio_submit(REQ_OP_READ, 0, args.bio);
+		mpage_bio_submit(REQ_OP_READ, REQ_RAHEAD, args.bio);
 	return 0;
 }
 EXPORT_SYMBOL(mpage_readpages);
-- 
2.7.4

[PATCH 3/4] btrfs: readpages() should submit IO as read-ahead

[Jens Axboe]

a_ops->readpages() is only ever used for read-ahead. Ensure that we
pass this information down to the block layer.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/btrfs/extent_io.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index e99b329002cf..32034bc1e2b9 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -3119,7 +3119,7 @@ static inline void __do_contiguous_readpages(struct extent_io_tree *tree,
 
 	for (index = 0; index < nr_pages; index++) {
 		__do_readpage(tree, pages[index], btrfs_get_extent, em_cached,
-				bio, 0, bio_flags, 0, prev_em_start);
+				bio, 0, bio_flags, REQ_RAHEAD, prev_em_start);
 		put_page(pages[index]);
 	}
 }
-- 
2.7.4

[PATCH 4/4] ext4: readpages() should submit IO as read-ahead

[Jens Axboe]

a_ops->readpages() is only ever used for read-ahead. Ensure that we
pass this information down to the block layer.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/ext4/ext4.h     | 2 +-
 fs/ext4/inode.c    | 5 +++--
 fs/ext4/readpage.c | 5 +++--
 3 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index a42e71203e53..62df0322b567 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -3052,7 +3052,7 @@ static inline void ext4_set_de_type(struct super_block *sb,
 /* readpages.c */
 extern int ext4_mpage_readpages(struct address_space *mapping,
 				struct list_head *pages, struct page *page,
-				unsigned nr_pages);
+				unsigned nr_pages, bool is_readahead);
 
 /* symlink.c */
 extern const struct inode_operations ext4_encrypted_symlink_inode_operations;
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 1e50c5efae67..2bb50d184bf5 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -3327,7 +3327,8 @@ static int ext4_readpage(struct file *file, struct page *page)
 		ret = ext4_readpage_inline(inode, page);
 
 	if (ret == -EAGAIN)
-		return ext4_mpage_readpages(page->mapping, NULL, page, 1);
+		return ext4_mpage_readpages(page->mapping, NULL, page, 1,
+						false);
 
 	return ret;
 }
@@ -3342,7 +3343,7 @@ ext4_readpages(struct file *file, struct address_space *mapping,
 	if (ext4_has_inline_data(inode))
 		return 0;
 
-	return ext4_mpage_readpages(mapping, pages, NULL, nr_pages);
+	return ext4_mpage_readpages(mapping, pages, NULL, nr_pages, true);
 }
 
 static void ext4_invalidatepage(struct page *page, unsigned int offset,
diff --git a/fs/ext4/readpage.c b/fs/ext4/readpage.c
index 9ffa6fad18db..28c857d3b309 100644
--- a/fs/ext4/readpage.c
+++ b/fs/ext4/readpage.c
@@ -98,7 +98,7 @@ static void mpage_end_io(struct bio *bio)
 
 int ext4_mpage_readpages(struct address_space *mapping,
 			 struct list_head *pages, struct page *page,
-			 unsigned nr_pages)
+			 unsigned nr_pages, bool is_readahead)
 {
 	struct bio *bio = NULL;
 	sector_t last_block_in_bio = 0;
@@ -259,7 +259,8 @@ int ext4_mpage_readpages(struct address_space *mapping,
 			bio->bi_iter.bi_sector = blocks[0] << (blkbits - 9);
 			bio->bi_end_io = mpage_end_io;
 			bio->bi_private = ctx;
-			bio_set_op_attrs(bio, REQ_OP_READ, 0);
+			bio_set_op_attrs(bio, REQ_OP_READ,
+						is_readahead ? REQ_RAHEAD : 0);
 		}
 
 		length = first_hole << blkbits;
-- 
2.7.4

[PATCHSET v3 0/4] Submit ->readpages() IO as read-ahead

[Jens Axboe]

The only caller of ->readpages() is from read-ahead, yet we don't
submit IO flagged with REQ_RAHEAD. This means we don't see it in
blktrace, for instance, which is a shame. We already make assumptions
about ->readpages() just being for read-ahead in the mpage
implementation, using readahead_gfp_mask(mapping) as out GFP mask of
choice.

This small series fixes up mpage_readpages() to submit with
REQ_RAHEAD, which takes care of file systems using mpage_readpages().
The first patch is a prep patch, that makes do_mpage_readpage() take
an argument structure.

Changes since v2:

- Get rid of 'gfp' passing once we have is_readahead
- Pack struct better, makes it 8 bytes smaller.

Changes since v1:

- Fix ext4_mpage_readpages() also being used for regular reads
- Add prep patch with struct arguments for do_mpage_readpage()

 fs/btrfs/extent_io.c |   2 +-
 fs/ext4/ext4.h       |   2 +-
 fs/ext4/inode.c      |   5 +-
 fs/ext4/readpage.c   |   5 +-
 fs/mpage.c           | 115 ++++++++++++++++++++++++-------------------
 5 files changed, 71 insertions(+), 58 deletions(-)

[PATCH 1/4] mpage: add argument structure for do_mpage_readpage()

[Jens Axboe]

We're currently passing 8 arguments to this function, clean it up a
bit by packing the arguments in an args structure we pass to it.

Not intentional functional changes in this patch.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/mpage.c | 106 +++++++++++++++++++++++++++++++------------------------------
 1 file changed, 54 insertions(+), 52 deletions(-)

diff --git a/fs/mpage.c b/fs/mpage.c
index b7e7f570733a..0ecac5c410f4 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -133,6 +133,17 @@ map_buffer_to_page(struct page *page, struct buffer_head *bh, int page_block)
 	} while (page_bh != head);
 }
 
+struct mpage_readpage_args {
+	struct bio *bio;
+	struct page *page;
+	unsigned nr_pages;
+	sector_t last_block_in_bio;
+	struct buffer_head map_bh;
+	unsigned long first_logical_block;
+	get_block_t *get_block;
+	gfp_t gfp;
+};
+
 /*
  * This is the worker routine which does all the work of mapping the disk
  * blocks and constructs largest possible bios, submits them for IO if the
@@ -142,16 +153,14 @@ map_buffer_to_page(struct page *page, struct buffer_head *bh, int page_block)
  * represent the validity of its disk mapping and to decide when to do the next
  * get_block() call.
  */
-static struct bio *
-do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
-		sector_t *last_block_in_bio, struct buffer_head *map_bh,
-		unsigned long *first_logical_block, get_block_t get_block,
-		gfp_t gfp)
+static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 {
+	struct page *page = args->page;
 	struct inode *inode = page->mapping->host;
 	const unsigned blkbits = inode->i_blkbits;
 	const unsigned blocks_per_page = PAGE_SIZE >> blkbits;
 	const unsigned blocksize = 1 << blkbits;
+	struct buffer_head *map_bh = &args->map_bh;
 	sector_t block_in_file;
 	sector_t last_block;
 	sector_t last_block_in_file;
@@ -168,7 +177,7 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 		goto confused;
 
 	block_in_file = (sector_t)page->index << (PAGE_SHIFT - blkbits);
-	last_block = block_in_file + nr_pages * blocks_per_page;
+	last_block = block_in_file + args->nr_pages * blocks_per_page;
 	last_block_in_file = (i_size_read(inode) + blocksize - 1) >> blkbits;
 	if (last_block > last_block_in_file)
 		last_block = last_block_in_file;
@@ -178,9 +187,9 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 	 * Map blocks using the result from the previous get_blocks call first.
 	 */
 	nblocks = map_bh->b_size >> blkbits;
-	if (buffer_mapped(map_bh) && block_in_file > *first_logical_block &&
-			block_in_file < (*first_logical_block + nblocks)) {
-		unsigned map_offset = block_in_file - *first_logical_block;
+	if (buffer_mapped(map_bh) && block_in_file > args->first_logical_block &&
+			block_in_file < (args->first_logical_block + nblocks)) {
+		unsigned map_offset = block_in_file - args->first_logical_block;
 		unsigned last = nblocks - map_offset;
 
 		for (relative_block = 0; ; relative_block++) {
@@ -208,9 +217,9 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 
 		if (block_in_file < last_block) {
 			map_bh->b_size = (last_block-block_in_file) << blkbits;
-			if (get_block(inode, block_in_file, map_bh, 0))
+			if (args->get_block(inode, block_in_file, map_bh, 0))
 				goto confused;
-			*first_logical_block = block_in_file;
+			args->first_logical_block = block_in_file;
 		}
 
 		if (!buffer_mapped(map_bh)) {
@@ -273,43 +282,43 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 	/*
 	 * This page will go to BIO.  Do we need to send this BIO off first?
 	 */
-	if (bio && (*last_block_in_bio != blocks[0] - 1))
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (args->bio && (args->last_block_in_bio != blocks[0] - 1))
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 
 alloc_new:
-	if (bio == NULL) {
+	if (args->bio == NULL) {
 		if (first_hole == blocks_per_page) {
 			if (!bdev_read_page(bdev, blocks[0] << (blkbits - 9),
 								page))
 				goto out;
 		}
-		bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
-				min_t(int, nr_pages, BIO_MAX_PAGES), gfp);
-		if (bio == NULL)
+		args->bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
+				min_t(int, args->nr_pages, BIO_MAX_PAGES), args->gfp);
+		if (args->bio == NULL)
 			goto confused;
 	}
 
 	length = first_hole << blkbits;
-	if (bio_add_page(bio, page, length, 0) < length) {
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (bio_add_page(args->bio, page, length, 0) < length) {
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 		goto alloc_new;
 	}
 
-	relative_block = block_in_file - *first_logical_block;
+	relative_block = block_in_file - args->first_logical_block;
 	nblocks = map_bh->b_size >> blkbits;
 	if ((buffer_boundary(map_bh) && relative_block == nblocks) ||
 	    (first_hole != blocks_per_page))
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 	else
-		*last_block_in_bio = blocks[blocks_per_page - 1];
+		args->last_block_in_bio = blocks[blocks_per_page - 1];
 out:
-	return bio;
+	return args->bio;
 
 confused:
-	if (bio)
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (args->bio)
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 	if (!PageUptodate(page))
-	        block_read_full_page(page, get_block);
+	        block_read_full_page(page, args->get_block);
 	else
 		unlock_page(page);
 	goto out;
@@ -363,15 +372,12 @@ int
 mpage_readpages(struct address_space *mapping, struct list_head *pages,
 				unsigned nr_pages, get_block_t get_block)
 {
-	struct bio *bio = NULL;
+	struct mpage_readpage_args args = {
+		.get_block = get_block,
+		.gfp = readahead_gfp_mask(mapping),
+	};
 	unsigned page_idx;
-	sector_t last_block_in_bio = 0;
-	struct buffer_head map_bh;
-	unsigned long first_logical_block = 0;
-	gfp_t gfp = readahead_gfp_mask(mapping);
 
-	map_bh.b_state = 0;
-	map_bh.b_size = 0;
 	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
 		struct page *page = lru_to_page(pages);
 
@@ -379,18 +385,16 @@ mpage_readpages(struct address_space *mapping, struct list_head *pages,
 		list_del(&page->lru);
 		if (!add_to_page_cache_lru(page, mapping,
 					page->index,
-					gfp)) {
-			bio = do_mpage_readpage(bio, page,
-					nr_pages - page_idx,
-					&last_block_in_bio, &map_bh,
-					&first_logical_block,
-					get_block, gfp);
+					args.gfp)) {
+			args.page = page;
+			args.nr_pages = nr_pages - page_idx;
+			args.bio = do_mpage_readpage(&args);
 		}
 		put_page(page);
 	}
 	BUG_ON(!list_empty(pages));
-	if (bio)
-		mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (args.bio)
+		mpage_bio_submit(REQ_OP_READ, 0, args.bio);
 	return 0;
 }
 EXPORT_SYMBOL(mpage_readpages);
@@ -400,18 +404,16 @@ EXPORT_SYMBOL(mpage_readpages);
  */
 int mpage_readpage(struct page *page, get_block_t get_block)
 {
-	struct bio *bio = NULL;
-	sector_t last_block_in_bio = 0;
-	struct buffer_head map_bh;
-	unsigned long first_logical_block = 0;
-	gfp_t gfp = mapping_gfp_constraint(page->mapping, GFP_KERNEL);
+	struct mpage_readpage_args args = {
+		.page = page,
+		.nr_pages = 1,
+		.get_block = get_block,
+		.gfp = mapping_gfp_constraint(page->mapping, GFP_KERNEL),
+	};
 
-	map_bh.b_state = 0;
-	map_bh.b_size = 0;
-	bio = do_mpage_readpage(bio, page, 1, &last_block_in_bio,
-			&map_bh, &first_logical_block, get_block, gfp);
-	if (bio)
-		mpage_bio_submit(REQ_OP_READ, 0, bio);
+	args.bio = do_mpage_readpage(&args);
+	if (args.bio)
+		mpage_bio_submit(REQ_OP_READ, 0, args.bio);
 	return 0;
 }
 EXPORT_SYMBOL(mpage_readpage);
-- 
2.7.4

[PATCH v4 0/4] Submit ->readpages() IO as read-ahead

[Jens Axboe]

The only caller of ->readpages() is from read-ahead, yet we don't
submit IO flagged with REQ_RAHEAD. This means we don't see it in
blktrace, for instance, which is a shame. Additionally, it's
preventing further functional changes in the block layer for
deadling with read-ahead more intelligently. We already make assumptions
about ->readpages() just being for read-ahead in the mpage
implementation, using readahead_gfp_mask(mapping) as out GFP mask of
choice.

This small series fixes up mpage_readpages() to submit with REQ_RAHEAD,
which takes care of file systems using mpage_readpages().  The first
patch is a prep patch, that makes do_mpage_readpage() take an argument
structure.

Changes since v3:

- Add comments on ->readpages() purely being read-ahead
- Rebase on current -git

Changes since v2:

- Get rid of 'gfp' passing once we have is_readahead
- Pack struct better, makes it 8 bytes smaller.

Changes since v1:

- Fix ext4_mpage_readpages() also being used for regular reads
- Add prep patch with struct arguments for do_mpage_readpage()

 fs/btrfs/extent_io.c |   2 +-
 fs/ext4/ext4.h       |   2 +-
 fs/ext4/inode.c      |   5 +-
 fs/ext4/readpage.c   |   5 +-
 fs/f2fs/data.c       |   5 ++
 fs/mpage.c           | 115 ++++++++++++++++++++++++-------------------
 include/linux/fs.h   |   4 ++
 7 files changed, 80 insertions(+), 58 deletions(-)

-- 
Jens Axboe

[PATCH 1/4] mpage: add argument structure for do_mpage_readpage()

[Jens Axboe]

We're currently passing 8 arguments to this function, clean it up a
bit by packing the arguments in an args structure we pass to it.

No intentional functional changes in this patch.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/mpage.c | 106 +++++++++++++++++++++++++++--------------------------
 1 file changed, 54 insertions(+), 52 deletions(-)

diff --git a/fs/mpage.c b/fs/mpage.c
index b7e7f570733a..0ecac5c410f4 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -133,6 +133,17 @@ map_buffer_to_page(struct page *page, struct buffer_head *bh, int page_block)
 	} while (page_bh != head);
 }
 
+struct mpage_readpage_args {
+	struct bio *bio;
+	struct page *page;
+	unsigned nr_pages;
+	sector_t last_block_in_bio;
+	struct buffer_head map_bh;
+	unsigned long first_logical_block;
+	get_block_t *get_block;
+	gfp_t gfp;
+};
+
 /*
  * This is the worker routine which does all the work of mapping the disk
  * blocks and constructs largest possible bios, submits them for IO if the
@@ -142,16 +153,14 @@ map_buffer_to_page(struct page *page, struct buffer_head *bh, int page_block)
  * represent the validity of its disk mapping and to decide when to do the next
  * get_block() call.
  */
-static struct bio *
-do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
-		sector_t *last_block_in_bio, struct buffer_head *map_bh,
-		unsigned long *first_logical_block, get_block_t get_block,
-		gfp_t gfp)
+static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 {
+	struct page *page = args->page;
 	struct inode *inode = page->mapping->host;
 	const unsigned blkbits = inode->i_blkbits;
 	const unsigned blocks_per_page = PAGE_SIZE >> blkbits;
 	const unsigned blocksize = 1 << blkbits;
+	struct buffer_head *map_bh = &args->map_bh;
 	sector_t block_in_file;
 	sector_t last_block;
 	sector_t last_block_in_file;
@@ -168,7 +177,7 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 		goto confused;
 
 	block_in_file = (sector_t)page->index << (PAGE_SHIFT - blkbits);
-	last_block = block_in_file + nr_pages * blocks_per_page;
+	last_block = block_in_file + args->nr_pages * blocks_per_page;
 	last_block_in_file = (i_size_read(inode) + blocksize - 1) >> blkbits;
 	if (last_block > last_block_in_file)
 		last_block = last_block_in_file;
@@ -178,9 +187,9 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 	 * Map blocks using the result from the previous get_blocks call first.
 	 */
 	nblocks = map_bh->b_size >> blkbits;
-	if (buffer_mapped(map_bh) && block_in_file > *first_logical_block &&
-			block_in_file < (*first_logical_block + nblocks)) {
-		unsigned map_offset = block_in_file - *first_logical_block;
+	if (buffer_mapped(map_bh) && block_in_file > args->first_logical_block &&
+			block_in_file < (args->first_logical_block + nblocks)) {
+		unsigned map_offset = block_in_file - args->first_logical_block;
 		unsigned last = nblocks - map_offset;
 
 		for (relative_block = 0; ; relative_block++) {
@@ -208,9 +217,9 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 
 		if (block_in_file < last_block) {
 			map_bh->b_size = (last_block-block_in_file) << blkbits;
-			if (get_block(inode, block_in_file, map_bh, 0))
+			if (args->get_block(inode, block_in_file, map_bh, 0))
 				goto confused;
-			*first_logical_block = block_in_file;
+			args->first_logical_block = block_in_file;
 		}
 
 		if (!buffer_mapped(map_bh)) {
@@ -273,43 +282,43 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 	/*
 	 * This page will go to BIO.  Do we need to send this BIO off first?
 	 */
-	if (bio && (*last_block_in_bio != blocks[0] - 1))
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (args->bio && (args->last_block_in_bio != blocks[0] - 1))
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 
 alloc_new:
-	if (bio == NULL) {
+	if (args->bio == NULL) {
 		if (first_hole == blocks_per_page) {
 			if (!bdev_read_page(bdev, blocks[0] << (blkbits - 9),
 								page))
 				goto out;
 		}
-		bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
-				min_t(int, nr_pages, BIO_MAX_PAGES), gfp);
-		if (bio == NULL)
+		args->bio = mpage_alloc(bdev, blocks[0] << (blkbits - 9),
+				min_t(int, args->nr_pages, BIO_MAX_PAGES), args->gfp);
+		if (args->bio == NULL)
 			goto confused;
 	}
 
 	length = first_hole << blkbits;
-	if (bio_add_page(bio, page, length, 0) < length) {
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (bio_add_page(args->bio, page, length, 0) < length) {
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 		goto alloc_new;
 	}
 
-	relative_block = block_in_file - *first_logical_block;
+	relative_block = block_in_file - args->first_logical_block;
 	nblocks = map_bh->b_size >> blkbits;
 	if ((buffer_boundary(map_bh) && relative_block == nblocks) ||
 	    (first_hole != blocks_per_page))
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 	else
-		*last_block_in_bio = blocks[blocks_per_page - 1];
+		args->last_block_in_bio = blocks[blocks_per_page - 1];
 out:
-	return bio;
+	return args->bio;
 
 confused:
-	if (bio)
-		bio = mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (args->bio)
+		args->bio = mpage_bio_submit(REQ_OP_READ, 0, args->bio);
 	if (!PageUptodate(page))
-	        block_read_full_page(page, get_block);
+	        block_read_full_page(page, args->get_block);
 	else
 		unlock_page(page);
 	goto out;
@@ -363,15 +372,12 @@ int
 mpage_readpages(struct address_space *mapping, struct list_head *pages,
 				unsigned nr_pages, get_block_t get_block)
 {
-	struct bio *bio = NULL;
+	struct mpage_readpage_args args = {
+		.get_block = get_block,
+		.gfp = readahead_gfp_mask(mapping),
+	};
 	unsigned page_idx;
-	sector_t last_block_in_bio = 0;
-	struct buffer_head map_bh;
-	unsigned long first_logical_block = 0;
-	gfp_t gfp = readahead_gfp_mask(mapping);
 
-	map_bh.b_state = 0;
-	map_bh.b_size = 0;
 	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
 		struct page *page = lru_to_page(pages);
 
@@ -379,18 +385,16 @@ mpage_readpages(struct address_space *mapping, struct list_head *pages,
 		list_del(&page->lru);
 		if (!add_to_page_cache_lru(page, mapping,
 					page->index,
-					gfp)) {
-			bio = do_mpage_readpage(bio, page,
-					nr_pages - page_idx,
-					&last_block_in_bio, &map_bh,
-					&first_logical_block,
-					get_block, gfp);
+					args.gfp)) {
+			args.page = page;
+			args.nr_pages = nr_pages - page_idx;
+			args.bio = do_mpage_readpage(&args);
 		}
 		put_page(page);
 	}
 	BUG_ON(!list_empty(pages));
-	if (bio)
-		mpage_bio_submit(REQ_OP_READ, 0, bio);
+	if (args.bio)
+		mpage_bio_submit(REQ_OP_READ, 0, args.bio);
 	return 0;
 }
 EXPORT_SYMBOL(mpage_readpages);
@@ -400,18 +404,16 @@ EXPORT_SYMBOL(mpage_readpages);
  */
 int mpage_readpage(struct page *page, get_block_t get_block)
 {
-	struct bio *bio = NULL;
-	sector_t last_block_in_bio = 0;
-	struct buffer_head map_bh;
-	unsigned long first_logical_block = 0;
-	gfp_t gfp = mapping_gfp_constraint(page->mapping, GFP_KERNEL);
+	struct mpage_readpage_args args = {
+		.page = page,
+		.nr_pages = 1,
+		.get_block = get_block,
+		.gfp = mapping_gfp_constraint(page->mapping, GFP_KERNEL),
+	};
 
-	map_bh.b_state = 0;
-	map_bh.b_size = 0;
-	bio = do_mpage_readpage(bio, page, 1, &last_block_in_bio,
-			&map_bh, &first_logical_block, get_block, gfp);
-	if (bio)
-		mpage_bio_submit(REQ_OP_READ, 0, bio);
+	args.bio = do_mpage_readpage(&args);
+	if (args.bio)
+		mpage_bio_submit(REQ_OP_READ, 0, args.bio);
 	return 0;
 }
 EXPORT_SYMBOL(mpage_readpage);
-- 
2.17.1