[PATCH V2 0/4] block: Add support for REQ_OP_ALLOCATE

[PATCH V2 0/4] block: Add support for REQ_OP_ALLOCATE

[Chaitanya Kulkarni]

Hi,

This patch-series is based on the original RFC patch series:-
https://www.spinics.net/lists/linux-block/msg47933.html.

I've designed a rough testcase based on the information present
in the mailing list archive for original RFC, it may need
some corrections from the author.

If anyone is interested, test results are at the end of this patch.

Following is the original cover-letter :-

Information about continuous extent placement may be useful
for some block devices. Say, distributed network filesystems,
which provide block device interface, may use this information
for better blocks placement over the nodes in their cluster,
and for better performance. Block devices, which map a file
on another filesystem (loop), may request the same length extent
on underlining filesystem for less fragmentation and for batching
allocation requests. Also, hypervisors like QEMU may use this
information for optimization of cluster allocations.

This patchset introduces REQ_OP_ALLOCATE, which is going
to be used for forwarding user's fallocate(0) requests into
block device internals. It rather similar to existing
REQ_OP_DISCARD, REQ_OP_WRITE_ZEROES, etc. The corresponding
exported primitive is called blkdev_issue_allocate().
See [1/3] for the details.

Patch [2/3] teaches loop driver to handle REQ_OP_ALLOCATE
requests by calling fallocate(0).

Patch [3/3] makes ext4 to notify a block device about fallocate(0).

Here is a simple test I did:
https://gist.github.com/tkhai/5b788651cdb74c1dbff3500745878856

I attached a file on ext4 to loop. Then, created ext4 partition
on loop device and started the test in the partition. Direct-io
is enabled on loop.

The test fallocates 4G file and writes from some offset with
given step, then it chooses another offset and repeats. After
the test all the blocks in the file become written.

The results shows that batching extents-assigning requests improves
the performance:

Before patchset: real ~ 1min 27sec
After patchset:  real ~ 1min 16sec (18% better)

Ordinary fallocate() before writes improves the performance
by batching the requests. These results just show, the same
is in case of forwarding extents information to underlining
filesystem.

Regards,
Chaitanya

Changes from V1:-

1. Adjust series from using assign range or ASSIGN_RANGE to allocate or
   ALLOCATE. (Martin)
2. Add SECT_FROM_SB_BLK() and NSECTS_FROM_SB_NBLK() macros since it
   avoids weired line breaking and duplication of the code for
   sb_issue_discard(), sb_issue_zeroout() and sb_issue_allocate().
   Should there be a prep patch for this ?

Changes from RFC:-

1. Add missing plumbing for REQ_OP_ASSIGN_RANGE similar to write-zeores.
2. Add a prep patch to create a helper to submit payloadless bios.
3. Design a testcases around the description present in the
   cover-letter.

Chaitanya Kulkarni (1):
  block: create payloadless issue bio helper

Kirill Tkhai (3):
  block: Add support for REQ_OP_ALLOCATE
  loop: Forward REQ_OP_ALLOCATE into fallocate(0)
  ext4: Notify block device about alloc-assigned blk

 block/blk-core.c          |   5 ++
 block/blk-lib.c           | 115 +++++++++++++++++++++++++++++++-------
 block/blk-merge.c         |  21 +++++++
 block/blk-settings.c      |  19 +++++++
 block/blk-zoned.c         |   1 +
 block/bounce.c            |   1 +
 drivers/block/loop.c      |   5 ++
 fs/ext4/ext4.h            |   2 +
 fs/ext4/extents.c         |  12 +++-
 include/linux/bio.h       |   9 ++-
 include/linux/blk_types.h |   2 +
 include/linux/blkdev.h    |  55 ++++++++++++++----
 12 files changed, 210 insertions(+), 37 deletions(-)

-- 
2.22.0

[PATCH V2 1/4] block: create payloadless issue bio helper

[Chaitanya Kulkarni]

This is a prep-patch that creates a helper to submit payloadless bio
with all the required arguments. This is needed to avoid the code
repetition in blk-lib,c so that new payloadless ops can use it.  

Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
---
 block/blk-lib.c | 51 +++++++++++++++++++++++++++++--------------------
 1 file changed, 30 insertions(+), 21 deletions(-)

diff --git a/block/blk-lib.c b/block/blk-lib.c
index 5f2c429d4378..8e53e393703c 100644
--- a/block/blk-lib.c
+++ b/block/blk-lib.c
@@ -209,13 +209,40 @@ int blkdev_issue_write_same(struct block_device *bdev, sector_t sector,
 }
 EXPORT_SYMBOL(blkdev_issue_write_same);
 
+static void __blkdev_issue_payloadless(struct block_device *bdev,
+		unsigned op, sector_t sector, sector_t nr_sects, gfp_t gfp_mask,
+		struct bio **biop, unsigned bio_opf, unsigned int max_sectors)
+{
+	struct bio *bio = *biop;
+
+	while (nr_sects) {
+		bio = blk_next_bio(bio, 0, gfp_mask);
+		bio->bi_iter.bi_sector = sector;
+		bio_set_dev(bio, bdev);
+		bio->bi_opf = op;
+		bio->bi_opf |= bio_opf;
+
+		if (nr_sects > max_sectors) {
+			bio->bi_iter.bi_size = max_sectors << 9;
+			nr_sects -= max_sectors;
+			sector += max_sectors;
+		} else {
+			bio->bi_iter.bi_size = nr_sects << 9;
+			nr_sects = 0;
+		}
+		cond_resched();
+	}
+
+	*biop = bio;
+}
+
 static int __blkdev_issue_write_zeroes(struct block_device *bdev,
 		sector_t sector, sector_t nr_sects, gfp_t gfp_mask,
 		struct bio **biop, unsigned flags)
 {
-	struct bio *bio = *biop;
 	unsigned int max_write_zeroes_sectors;
 	struct request_queue *q = bdev_get_queue(bdev);
+	unsigned int unmap = (flags & BLKDEV_ZERO_NOUNMAP) ? REQ_NOUNMAP : 0;
 
 	if (!q)
 		return -ENXIO;
@@ -229,26 +256,8 @@ static int __blkdev_issue_write_zeroes(struct block_device *bdev,
 	if (max_write_zeroes_sectors == 0)
 		return -EOPNOTSUPP;
 
-	while (nr_sects) {
-		bio = blk_next_bio(bio, 0, gfp_mask);
-		bio->bi_iter.bi_sector = sector;
-		bio_set_dev(bio, bdev);
-		bio->bi_opf = REQ_OP_WRITE_ZEROES;
-		if (flags & BLKDEV_ZERO_NOUNMAP)
-			bio->bi_opf |= REQ_NOUNMAP;
-
-		if (nr_sects > max_write_zeroes_sectors) {
-			bio->bi_iter.bi_size = max_write_zeroes_sectors << 9;
-			nr_sects -= max_write_zeroes_sectors;
-			sector += max_write_zeroes_sectors;
-		} else {
-			bio->bi_iter.bi_size = nr_sects << 9;
-			nr_sects = 0;
-		}
-		cond_resched();
-	}
-
-	*biop = bio;
+	__blkdev_issue_payloadless(bdev, REQ_OP_WRITE_ZEROES, sector, nr_sects,
+		       gfp_mask, biop, unmap, max_write_zeroes_sectors);
 	return 0;
 }
 
-- 
2.22.0

[PATCH V2 2/4] block: Add support for REQ_OP_ALLOCATE

[Chaitanya Kulkarni]

From: Kirill Tkhai <ktkhai@virtuozzo.com>

This operation allows to notify a device about the fact, that some
sectors range was chosen by a filesystem as a single extent, and
the device should try its best to reflect that (keep the range as a
single hunk in its internals, or represent the range as minimal set of
hunks). Speaking directly, the operation is for forwarding fallocate(0)
requests into an essence, on which the device is based.

This may be useful for some distributed network filesystems, providing
block device interface, for optimization of their blocks placement over
the cluster nodes.

Also, block devices mapping a file (like loop) are users of that, since
this allows to allocate more continuous extents and since this batches
blocks allocation requests. In addition, hypervisors like QEMU may use
this for better blocks placement.

This patch adds a new blkdev_issue_allocate() primitive, which is
rather similar to existing blkdev_issue_{*} api. Also, a new queue
limit.max_allocate_sectors is added.

Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
---
 block/blk-core.c          |  5 +++
 block/blk-lib.c           | 64 +++++++++++++++++++++++++++++++++++++++
 block/blk-merge.c         | 21 +++++++++++++
 block/blk-settings.c      | 19 ++++++++++++
 block/blk-zoned.c         |  1 +
 block/bounce.c            |  1 +
 include/linux/bio.h       |  9 ++++--
 include/linux/blk_types.h |  2 ++
 include/linux/blkdev.h    | 55 +++++++++++++++++++++++++--------
 9 files changed, 162 insertions(+), 15 deletions(-)

diff --git a/block/blk-core.c b/block/blk-core.c
index 7e4a1da0715e..3c22197728a4 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -137,6 +137,7 @@ static const char *const blk_op_name[] = {
 	REQ_OP_NAME(ZONE_FINISH),
 	REQ_OP_NAME(WRITE_SAME),
 	REQ_OP_NAME(WRITE_ZEROES),
+	REQ_OP_NAME(ALLOCATE),
 	REQ_OP_NAME(SCSI_IN),
 	REQ_OP_NAME(SCSI_OUT),
 	REQ_OP_NAME(DRV_IN),
@@ -958,6 +959,10 @@ generic_make_request_checks(struct bio *bio)
 		if (!q->limits.max_write_zeroes_sectors)
 			goto not_supported;
 		break;
+	case REQ_OP_ALLOCATE:
+		if (!q->limits.max_allocate_sectors)
+			goto not_supported;
+		break;
 	default:
 		break;
 	}
diff --git a/block/blk-lib.c b/block/blk-lib.c
index 8e53e393703c..d16c09c19010 100644
--- a/block/blk-lib.c
+++ b/block/blk-lib.c
@@ -414,3 +414,67 @@ int blkdev_issue_zeroout(struct block_device *bdev, sector_t sector,
 	return ret;
 }
 EXPORT_SYMBOL(blkdev_issue_zeroout);
+
+static int __blkdev_issue_allocate(struct block_device *bdev,
+		sector_t sector, sector_t nr_sects, gfp_t gfp_mask,
+		struct bio **biop)
+{
+	unsigned int max_allocate_sectors;
+	struct request_queue *q = bdev_get_queue(bdev);
+
+	if (!q)
+		return -ENXIO;
+
+	if (bdev_read_only(bdev))
+		return -EPERM;
+
+	max_allocate_sectors = bdev_allocate_sectors(bdev);
+
+	if (max_allocate_sectors == 0)
+		return -EOPNOTSUPP;
+
+	__blkdev_issue_payloadless(bdev, REQ_OP_ALLOCATE, sector, nr_sects,
+			gfp_mask, biop, 0, max_allocate_sectors);
+	return 0;
+}
+
+/**
+ * __blkdev_issue_allocate - generate number of allocate bios
+ * @bdev:	blockdev to issue
+ * @sector:	start sector
+ * @nr_sects:	number of sectors to write
+ * @gfp_mask:	memory allocation flags (for bio_alloc)
+ * @biop:	pointer to anchor bio
+ *
+ * Description:
+ *  Assign a block range for batched allocation requests. Useful in stacking
+ *  block device on the top of the file system.
+ *
+ */
+int blkdev_issue_allocate(struct block_device *bdev, sector_t sector,
+			sector_t nr_sects, gfp_t gfp_mask)
+{
+	int ret = 0;
+	sector_t bs_mask;
+	struct blk_plug plug;
+	struct bio *bio = NULL;
+
+	if (bdev_allocate_sectors(bdev) == 0)
+		return 0;
+
+	bs_mask = (bdev_logical_block_size(bdev) >> 9) - 1;
+	if ((sector | nr_sects) & bs_mask)
+		return -EINVAL;
+
+	blk_start_plug(&plug);
+	ret = __blkdev_issue_allocate(bdev, sector, nr_sects,
+					  gfp_mask, &bio);
+	if (ret == 0 && bio) {
+		ret = submit_bio_wait(bio);
+		bio_put(bio);
+	}
+	blk_finish_plug(&plug);
+
+	return ret;
+}
+EXPORT_SYMBOL(blkdev_issue_allocate);
diff --git a/block/blk-merge.c b/block/blk-merge.c
index 1534ed736363..4ed5f103a18d 100644
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -116,6 +116,22 @@ static struct bio *blk_bio_write_zeroes_split(struct request_queue *q,
 	return bio_split(bio, q->limits.max_write_zeroes_sectors, GFP_NOIO, bs);
 }
 
+static struct bio *blk_bio_allocate_split(struct request_queue *q,
+					      struct bio *bio,
+					      struct bio_set *bs,
+					      unsigned *nsegs)
+{
+	*nsegs = 0;
+
+	if (!q->limits.max_allocate_sectors)
+		return NULL;
+
+	if (bio_sectors(bio) <= q->limits.max_allocate_sectors)
+		return NULL;
+
+	return bio_split(bio, q->limits.max_allocate_sectors, GFP_NOIO, bs);
+}
+
 static struct bio *blk_bio_write_same_split(struct request_queue *q,
 					    struct bio *bio,
 					    struct bio_set *bs,
@@ -308,6 +324,10 @@ void __blk_queue_split(struct request_queue *q, struct bio **bio,
 		split = blk_bio_write_zeroes_split(q, *bio, &q->bio_split,
 				nr_segs);
 		break;
+	case REQ_OP_ALLOCATE:
+		split = blk_bio_allocate_split(q, *bio, &q->bio_split,
+				nr_segs);
+		break;
 	case REQ_OP_WRITE_SAME:
 		split = blk_bio_write_same_split(q, *bio, &q->bio_split,
 				nr_segs);
@@ -386,6 +406,7 @@ unsigned int blk_recalc_rq_segments(struct request *rq)
 	case REQ_OP_DISCARD:
 	case REQ_OP_SECURE_ERASE:
 	case REQ_OP_WRITE_ZEROES:
+	case REQ_OP_ALLOCATE:
 		return 0;
 	case REQ_OP_WRITE_SAME:
 		return 1;
diff --git a/block/blk-settings.c b/block/blk-settings.c
index 14397b4c4b53..3f8f55e27b9a 100644
--- a/block/blk-settings.c
+++ b/block/blk-settings.c
@@ -48,6 +48,7 @@ void blk_set_default_limits(struct queue_limits *lim)
 	lim->chunk_sectors = 0;
 	lim->max_write_same_sectors = 0;
 	lim->max_write_zeroes_sectors = 0;
+	lim->max_allocate_sectors = 0;
 	lim->max_discard_sectors = 0;
 	lim->max_hw_discard_sectors = 0;
 	lim->discard_granularity = 0;
@@ -83,6 +84,7 @@ void blk_set_stacking_limits(struct queue_limits *lim)
 	lim->max_dev_sectors = UINT_MAX;
 	lim->max_write_same_sectors = UINT_MAX;
 	lim->max_write_zeroes_sectors = UINT_MAX;
+	lim->max_allocate_sectors = UINT_MAX;
 }
 EXPORT_SYMBOL(blk_set_stacking_limits);
 
@@ -221,6 +223,21 @@ void blk_queue_max_write_zeroes_sectors(struct request_queue *q,
 }
 EXPORT_SYMBOL(blk_queue_max_write_zeroes_sectors);
 
+/**
+ * blk_queue_max_allocate_sectors - set max sectors for a single
+ *                                      allocate
+ *
+ * @q:  the request queue for the device
+ * @max_allocate_sectors: maximum number of sectors to allocate per
+ *                            command
+ **/
+void blk_queue_max_allocate_sectors(struct request_queue *q,
+		unsigned int max_allocate_sectors)
+{
+	q->limits.max_allocate_sectors = max_allocate_sectors;
+}
+EXPORT_SYMBOL(blk_queue_max_allocate_sectors);
+
 /**
  * blk_queue_max_segments - set max hw segments for a request for this queue
  * @q:  the request queue for the device
@@ -470,6 +487,8 @@ int blk_stack_limits(struct queue_limits *t, struct queue_limits *b,
 					b->max_write_same_sectors);
 	t->max_write_zeroes_sectors = min(t->max_write_zeroes_sectors,
 					b->max_write_zeroes_sectors);
+	t->max_allocate_sectors = min(t->max_allocate_sectors,
+				    b->max_allocate_sectors);
 	t->bounce_pfn = min_not_zero(t->bounce_pfn, b->bounce_pfn);
 
 	t->seg_boundary_mask = min_not_zero(t->seg_boundary_mask,
diff --git a/block/blk-zoned.c b/block/blk-zoned.c
index f87956e0dcaf..348eef082bc7 100644
--- a/block/blk-zoned.c
+++ b/block/blk-zoned.c
@@ -73,6 +73,7 @@ bool blk_req_needs_zone_write_lock(struct request *rq)
 
 	switch (req_op(rq)) {
 	case REQ_OP_WRITE_ZEROES:
+	case REQ_OP_ALLOCATE:
 	case REQ_OP_WRITE_SAME:
 	case REQ_OP_WRITE:
 		return blk_rq_zone_is_seq(rq);
diff --git a/block/bounce.c b/block/bounce.c
index f8ed677a1bf7..112349dfc235 100644
--- a/block/bounce.c
+++ b/block/bounce.c
@@ -257,6 +257,7 @@ static struct bio *bounce_clone_bio(struct bio *bio_src, gfp_t gfp_mask,
 	case REQ_OP_DISCARD:
 	case REQ_OP_SECURE_ERASE:
 	case REQ_OP_WRITE_ZEROES:
+	case REQ_OP_ALLOCATE:
 		break;
 	case REQ_OP_WRITE_SAME:
 		bio->bi_io_vec[bio->bi_vcnt++] = bio_src->bi_io_vec[0];
diff --git a/include/linux/bio.h b/include/linux/bio.h
index c1c0f9ea4e63..09daf514fd23 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -64,7 +64,8 @@ static inline bool bio_has_data(struct bio *bio)
 	    bio->bi_iter.bi_size &&
 	    bio_op(bio) != REQ_OP_DISCARD &&
 	    bio_op(bio) != REQ_OP_SECURE_ERASE &&
-	    bio_op(bio) != REQ_OP_WRITE_ZEROES)
+	    bio_op(bio) != REQ_OP_WRITE_ZEROES &&
+	    bio_op(bio) != REQ_OP_ALLOCATE)
 		return true;
 
 	return false;
@@ -75,7 +76,8 @@ static inline bool bio_no_advance_iter(struct bio *bio)
 	return bio_op(bio) == REQ_OP_DISCARD ||
 	       bio_op(bio) == REQ_OP_SECURE_ERASE ||
 	       bio_op(bio) == REQ_OP_WRITE_SAME ||
-	       bio_op(bio) == REQ_OP_WRITE_ZEROES;
+	       bio_op(bio) == REQ_OP_WRITE_ZEROES ||
+	       bio_op(bio) == REQ_OP_ALLOCATE;
 }
 
 static inline bool bio_mergeable(struct bio *bio)
@@ -178,7 +180,7 @@ static inline unsigned bio_segments(struct bio *bio)
 	struct bvec_iter iter;
 
 	/*
-	 * We special case discard/write same/write zeroes, because they
+	 * We special case discard/write same/write zeroes/allocate, because
 	 * interpret bi_size differently:
 	 */
 
@@ -186,6 +188,7 @@ static inline unsigned bio_segments(struct bio *bio)
 	case REQ_OP_DISCARD:
 	case REQ_OP_SECURE_ERASE:
 	case REQ_OP_WRITE_ZEROES:
+	case REQ_OP_ALLOCATE:
 		return 0;
 	case REQ_OP_WRITE_SAME:
 		return 1;
diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index 70254ae11769..ef8268c37f93 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -296,6 +296,8 @@ enum req_opf {
 	REQ_OP_ZONE_CLOSE	= 11,
 	/* Transition a zone to full */
 	REQ_OP_ZONE_FINISH	= 12,
+	/* Allocate a sector range */
+	REQ_OP_ALLOCATE	= 15,
 
 	/* SCSI passthrough using struct scsi_request */
 	REQ_OP_SCSI_IN		= 32,
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 32868fbedc9e..6b85973bf2ff 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -336,6 +336,7 @@ struct queue_limits {
 	unsigned int		max_hw_discard_sectors;
 	unsigned int		max_write_same_sectors;
 	unsigned int		max_write_zeroes_sectors;
+	unsigned int		max_allocate_sectors;
 	unsigned int		discard_granularity;
 	unsigned int		discard_alignment;
 
@@ -747,6 +748,9 @@ static inline bool rq_mergeable(struct request *rq)
 	if (req_op(rq) == REQ_OP_WRITE_ZEROES)
 		return false;
 
+	if (req_op(rq) == REQ_OP_ALLOCATE)
+		return false;
+
 	if (rq->cmd_flags & REQ_NOMERGE_FLAGS)
 		return false;
 	if (rq->rq_flags & RQF_NOMERGE_FLAGS)
@@ -1004,6 +1008,10 @@ static inline unsigned int blk_queue_get_max_sectors(struct request_queue *q,
 	if (unlikely(op == REQ_OP_WRITE_ZEROES))
 		return q->limits.max_write_zeroes_sectors;
 
+	if (unlikely(op == REQ_OP_ALLOCATE))
+		return min(q->limits.max_allocate_sectors,
+			   UINT_MAX >> SECTOR_SHIFT);
+
 	return q->limits.max_sectors;
 }
 
@@ -1080,6 +1088,8 @@ extern void blk_queue_max_write_same_sectors(struct request_queue *q,
 		unsigned int max_write_same_sectors);
 extern void blk_queue_max_write_zeroes_sectors(struct request_queue *q,
 		unsigned int max_write_same_sectors);
+extern void blk_queue_max_allocate_sectors(struct request_queue *q,
+		unsigned int max_allocate_sectors);
 extern void blk_queue_logical_block_size(struct request_queue *, unsigned int);
 extern void blk_queue_physical_block_size(struct request_queue *, unsigned int);
 extern void blk_queue_alignment_offset(struct request_queue *q,
@@ -1227,25 +1237,36 @@ extern int __blkdev_issue_zeroout(struct block_device *bdev, sector_t sector,
 extern int blkdev_issue_zeroout(struct block_device *bdev, sector_t sector,
 		sector_t nr_sects, gfp_t gfp_mask, unsigned flags);
 
+#define SECT_FROM_SB_BLK(sb, blk) ((blk) << ((sb)->s_blocksize_bits - 9))
+#define NSECTS_FROM_SB_NBLK(sb, nblk) ((nblk) << ((sb)->s_blocksize_bits - 9))
+
 static inline int sb_issue_discard(struct super_block *sb, sector_t block,
 		sector_t nr_blocks, gfp_t gfp_mask, unsigned long flags)
 {
-	return blkdev_issue_discard(sb->s_bdev,
-				    block << (sb->s_blocksize_bits -
-					      SECTOR_SHIFT),
-				    nr_blocks << (sb->s_blocksize_bits -
-						  SECTOR_SHIFT),
-				    gfp_mask, flags);
+	sector_t sect = SECT_FROM_SB_BLK(sb, block);
+	sector_t nsects = NSECTS_FROM_SB_NBLK(sb, nr_blocks);
+
+	return blkdev_issue_discard(sb->s_bdev, sect, nsects, gfp_mask, flags);
 }
 static inline int sb_issue_zeroout(struct super_block *sb, sector_t block,
 		sector_t nr_blocks, gfp_t gfp_mask)
 {
-	return blkdev_issue_zeroout(sb->s_bdev,
-				    block << (sb->s_blocksize_bits -
-					      SECTOR_SHIFT),
-				    nr_blocks << (sb->s_blocksize_bits -
-						  SECTOR_SHIFT),
-				    gfp_mask, 0);
+	sector_t sect = SECT_FROM_SB_BLK(sb, block);
+	sector_t nsects = NSECTS_FROM_SB_NBLK(sb, nr_blocks);
+
+	return blkdev_issue_zeroout(sb->s_bdev, sect, nsects, gfp_mask, 0);
+}
+
+extern int blkdev_issue_allocate(struct block_device *bdev, sector_t sector,
+		sector_t nr_sects, gfp_t gfp_mask);
+
+static inline int sb_issue_allocate(struct super_block *sb, sector_t block,
+		sector_t nr_blocks, gfp_t gfp_mask)
+{
+	sector_t sect = SECT_FROM_SB_BLK(sb, block);
+	sector_t nsects = NSECTS_FROM_SB_NBLK(sb, nr_blocks);
+
+	return blkdev_issue_allocate(sb->s_bdev, sect, nsects, gfp_mask);
 }
 
 extern int blk_verify_command(unsigned char *cmd, fmode_t mode);
@@ -1429,6 +1450,16 @@ static inline unsigned int bdev_write_zeroes_sectors(struct block_device *bdev)
 	return 0;
 }
 
+static inline unsigned int bdev_allocate_sectors(struct block_device *bdev)
+{
+	struct request_queue *q = bdev_get_queue(bdev);
+
+	if (q)
+		return q->limits.max_allocate_sectors;
+
+	return 0;
+}
+
 static inline enum blk_zoned_model bdev_zoned_model(struct block_device *bdev)
 {
 	struct request_queue *q = bdev_get_queue(bdev);
-- 
2.22.0

[PATCH V2 3/4] loop: Forward REQ_OP_ALLOCATE into fallocate(0)

[Chaitanya Kulkarni]

From: Kirill Tkhai <ktkhai@virtuozzo.com>

Send fallocate(0) request into underlining filesystem after upper
filesystem sent REQ_OP_ALLOCATE request to block device.

Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
[Use blk_queue_max_allocate_sectors() from newly updated previous
 patch.]
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
---
 drivers/block/loop.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index a42c49e04954..26ff16e34f5a 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -611,6 +611,8 @@ static int do_req_filebacked(struct loop_device *lo, struct request *rq)
 				FALLOC_FL_PUNCH_HOLE);
 	case REQ_OP_DISCARD:
 		return lo_fallocate(lo, rq, pos, FALLOC_FL_PUNCH_HOLE);
+	case REQ_OP_ALLOCATE:
+		return lo_fallocate(lo, rq, pos, 0);
 	case REQ_OP_WRITE:
 		if (lo->transfer)
 			return lo_write_transfer(lo, rq, pos);
@@ -878,6 +880,7 @@ static void loop_config_discard(struct loop_device *lo)
 		q->limits.discard_granularity = 0;
 		q->limits.discard_alignment = 0;
 		blk_queue_max_discard_sectors(q, 0);
+		blk_queue_max_allocate_sectors(q, 0);
 		blk_queue_max_write_zeroes_sectors(q, 0);
 		blk_queue_flag_clear(QUEUE_FLAG_DISCARD, q);
 		return;
@@ -888,6 +891,7 @@ static void loop_config_discard(struct loop_device *lo)
 
 	blk_queue_max_discard_sectors(q, UINT_MAX >> 9);
 	blk_queue_max_write_zeroes_sectors(q, UINT_MAX >> 9);
+	blk_queue_max_allocate_sectors(q, UINT_MAX >> 9);
 	blk_queue_flag_set(QUEUE_FLAG_DISCARD, q);
 }
 
@@ -1919,6 +1923,7 @@ static blk_status_t loop_queue_rq(struct blk_mq_hw_ctx *hctx,
 	case REQ_OP_FLUSH:
 	case REQ_OP_DISCARD:
 	case REQ_OP_WRITE_ZEROES:
+	case REQ_OP_ALLOCATE:
 		cmd->use_aio = false;
 		break;
 	default:
-- 
2.22.0

[PATCH V2 4/4] ext4: Notify block device about alloc-assigned blk

[Chaitanya Kulkarni]

From: Kirill Tkhai <ktkhai@virtuozzo.com>

Call sb_issue_allocate() after extent range was allocated on user
request. Hopeful, this helps block device to maintain its internals in
the best way, if this is appliable.

Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
---
 fs/ext4/ext4.h    |  2 ++
 fs/ext4/extents.c | 12 +++++++++++-
 2 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 61b37a052052..0d0fa9904147 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -622,6 +622,8 @@ enum {
 	 * allows jbd2 to avoid submitting data before commit. */
 #define EXT4_GET_BLOCKS_IO_SUBMIT		0x0400
 
+#define EXT4_GET_BLOCKS_SUBMIT_ALLOC		0x0800
+
 /*
  * The bit position of these flags must not overlap with any of the
  * EXT4_GET_BLOCKS_*.  They are used by ext4_find_extent(),
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 954013d6076b..37a60e190349 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -4449,6 +4449,14 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
 		ar.len = allocated;
 
 got_allocated_blocks:
+	if ((flags & EXT4_GET_BLOCKS_SUBMIT_ALLOC) &&
+			inode->i_fop->fallocate) {
+		err = sb_issue_allocate(inode->i_sb, newblock,
+				EXT4_C2B(sbi, allocated_clusters), GFP_NOFS);
+		if (err)
+			goto free_on_err;
+	}
+
 	/* try to insert new extent into found leaf and return */
 	ext4_ext_store_pblock(&newex, newblock + offset);
 	newex.ee_len = cpu_to_le16(ar.len);
@@ -4466,6 +4474,7 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
 		err = ext4_ext_insert_extent(handle, inode, &path,
 					     &newex, flags);
 
+free_on_err:
 	if (err && free_on_err) {
 		int fb_flags = flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE ?
 			EXT4_FREE_BLOCKS_NO_QUOT_UPDATE : 0;
@@ -4733,7 +4742,8 @@ static long ext4_zero_range(struct file *file, loff_t offset,
 			goto out_mutex;
 	}
 
-	flags = EXT4_GET_BLOCKS_CREATE_UNWRIT_EXT;
+	flags = EXT4_GET_BLOCKS_CREATE_UNWRIT_EXT |
+		EXT4_GET_BLOCKS_SUBMIT_ALLOC;
 	if (mode & FALLOC_FL_KEEP_SIZE)
 		flags |= EXT4_GET_BLOCKS_KEEP_SIZE;
 
-- 
2.22.0