block: rework zone reporting

Avoid the need to allocate a potentially large array of struct blk_zone
in the block layer by switching the ->report_zones method interface to
a callback model. Now the caller simply supplies a callback that is
executed on each reported zone, and private data for it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@wdc.com>
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

4 files changed, 115 insertions, 175 deletions

diff --git a/drivers/md/dm-flakey.c b/drivers/md/dm-flakey.c
index 76587e9af0ef..a2cc9e45cbba 100644
--- a/drivers/md/dm-flakey.c
+++ b/drivers/md/dm-flakey.c
@@ -459,21 +459,15 @@ static int flakey_prepare_ioctl(struct dm_target *ti, struct block_device **bdev
 }
 
 #ifdef CONFIG_BLK_DEV_ZONED
-static int flakey_report_zones(struct dm_target *ti, sector_t sector,
-			       struct blk_zone *zones, unsigned int *nr_zones)
+static int flakey_report_zones(struct dm_target *ti,
+		struct dm_report_zones_args *args, unsigned int nr_zones)
 {
 	struct flakey_c *fc = ti->private;
-	int ret;
+	sector_t sector = flakey_map_sector(ti, args->next_sector);
 
-	/* Do report and remap it */
-	ret = blkdev_report_zones(fc->dev->bdev, flakey_map_sector(ti, sector),
-				  zones, nr_zones);
-	if (ret != 0)
-		return ret;
-
-	if (*nr_zones)
-		dm_remap_zone_report(ti, fc->start, zones, nr_zones);
-	return 0;
+	args->start = fc->start;
+	return blkdev_report_zones(fc->dev->bdev, sector, nr_zones,
+				   dm_report_zones_cb, args);
 }
 #endif
 
diff --git a/drivers/md/dm-linear.c b/drivers/md/dm-linear.c
index 97acafd48c85..8d07fdf63a47 100644
--- a/drivers/md/dm-linear.c
+++ b/drivers/md/dm-linear.c
@@ -136,21 +136,15 @@ static int linear_prepare_ioctl(struct dm_target *ti, struct block_device **bdev
 }
 
 #ifdef CONFIG_BLK_DEV_ZONED
-static int linear_report_zones(struct dm_target *ti, sector_t sector,
-			       struct blk_zone *zones, unsigned int *nr_zones)
+static int linear_report_zones(struct dm_target *ti,
+		struct dm_report_zones_args *args, unsigned int nr_zones)
 {
-	struct linear_c *lc = (struct linear_c *) ti->private;
-	int ret;
-
-	/* Do report and remap it */
-	ret = blkdev_report_zones(lc->dev->bdev, linear_map_sector(ti, sector),
-				  zones, nr_zones);
-	if (ret != 0)
-		return ret;
+	struct linear_c *lc = ti->private;
+	sector_t sector = linear_map_sector(ti, args->next_sector);
 
-	if (*nr_zones)
-		dm_remap_zone_report(ti, lc->start, zones, nr_zones);
-	return 0;
+	args->start = lc->start;
+	return blkdev_report_zones(lc->dev->bdev, sector, nr_zones,
+				   dm_report_zones_cb, args);
 }
 #endif
 
diff --git a/drivers/md/dm-zoned-metadata.c b/drivers/md/dm-zoned-metadata.c
index feb4718ce6a6..069e4675da6b 100644
--- a/drivers/md/dm-zoned-metadata.c
+++ b/drivers/md/dm-zoned-metadata.c
@@ -1080,9 +1080,10 @@ static int dmz_load_sb(struct dmz_metadata *zmd)
 /*
  * Initialize a zone descriptor.
  */
-static int dmz_init_zone(struct dmz_metadata *zmd, struct dm_zone *zone,
-			 struct blk_zone *blkz)
+static int dmz_init_zone(struct blk_zone *blkz, unsigned int idx, void *data)
 {
+	struct dmz_metadata *zmd = data;
+	struct dm_zone *zone = &zmd->zones[idx];
 	struct dmz_dev *dev = zmd->dev;
 
 	/* Ignore the eventual last runt (smaller) zone */
@@ -1096,26 +1097,29 @@ static int dmz_init_zone(struct dmz_metadata *zmd, struct dm_zone *zone,
 	atomic_set(&zone->refcount, 0);
 	zone->chunk = DMZ_MAP_UNMAPPED;
 
-	if (blkz->type == BLK_ZONE_TYPE_CONVENTIONAL) {
+	switch (blkz->type) {
+	case BLK_ZONE_TYPE_CONVENTIONAL:
 		set_bit(DMZ_RND, &zone->flags);
 		zmd->nr_rnd_zones++;
-	} else if (blkz->type == BLK_ZONE_TYPE_SEQWRITE_REQ ||
-		   blkz->type == BLK_ZONE_TYPE_SEQWRITE_PREF) {
+		break;
+	case BLK_ZONE_TYPE_SEQWRITE_REQ:
+	case BLK_ZONE_TYPE_SEQWRITE_PREF:
 		set_bit(DMZ_SEQ, &zone->flags);
-	} else
+		break;
+	default:
 		return -ENXIO;
-
-	if (blkz->cond == BLK_ZONE_COND_OFFLINE)
-		set_bit(DMZ_OFFLINE, &zone->flags);
-	else if (blkz->cond == BLK_ZONE_COND_READONLY)
-		set_bit(DMZ_READ_ONLY, &zone->flags);
+	}
 
 	if (dmz_is_rnd(zone))
 		zone->wp_block = 0;
 	else
 		zone->wp_block = dmz_sect2blk(blkz->wp - blkz->start);
 
-	if (!dmz_is_offline(zone) && !dmz_is_readonly(zone)) {
+	if (blkz->cond == BLK_ZONE_COND_OFFLINE)
+		set_bit(DMZ_OFFLINE, &zone->flags);
+	else if (blkz->cond == BLK_ZONE_COND_READONLY)
+		set_bit(DMZ_READ_ONLY, &zone->flags);
+	else {
 		zmd->nr_useable_zones++;
 		if (dmz_is_rnd(zone)) {
 			zmd->nr_rnd_zones++;
@@ -1139,23 +1143,13 @@ static void dmz_drop_zones(struct dmz_metadata *zmd)
 }
 
 /*
- * The size of a zone report in number of zones.
- * This results in 4096*64B=256KB report zones commands.
- */
-#define DMZ_REPORT_NR_ZONES	4096
-
-/*
  * Allocate and initialize zone descriptors using the zone
  * information from disk.
  */
 static int dmz_init_zones(struct dmz_metadata *zmd)
 {
 	struct dmz_dev *dev = zmd->dev;
-	struct dm_zone *zone;
-	struct blk_zone *blkz;
-	unsigned int nr_blkz;
-	sector_t sector = 0;
-	int i, ret = 0;
+	int ret;
 
 	/* Init */
 	zmd->zone_bitmap_size = dev->zone_nr_blocks >> 3;
@@ -1169,54 +1163,38 @@ static int dmz_init_zones(struct dmz_metadata *zmd)
 	dmz_dev_info(dev, "Using %zu B for zone information",
 		     sizeof(struct dm_zone) * dev->nr_zones);
 
-	/* Get zone information */
-	nr_blkz = DMZ_REPORT_NR_ZONES;
-	blkz = kcalloc(nr_blkz, sizeof(struct blk_zone), GFP_KERNEL);
-	if (!blkz) {
-		ret = -ENOMEM;
-		goto out;
-	}
-
 	/*
-	 * Get zone information and initialize zone descriptors.
-	 * At the same time, determine where the super block
-	 * should be: first block of the first randomly writable
-	 * zone.
+	 * Get zone information and initialize zone descriptors.  At the same
+	 * time, determine where the super block should be: first block of the
+	 * first randomly writable zone.
 	 */
-	zone = zmd->zones;
-	while (sector < dev->capacity) {
-		/* Get zone information */
-		nr_blkz = DMZ_REPORT_NR_ZONES;
-		ret = blkdev_report_zones(dev->bdev, sector, blkz, &nr_blkz);
-		if (ret) {
-			dmz_dev_err(dev, "Report zones failed %d", ret);
-			goto out;
-		}
+	ret = blkdev_report_zones(dev->bdev, 0, BLK_ALL_ZONES, dmz_init_zone,
+				  zmd);
+	if (ret < 0) {
+		dmz_drop_zones(zmd);
+		return ret;
+	}
 
-		if (!nr_blkz)
-			break;
+	return 0;
+}
 
-		/* Process report */
-		for (i = 0; i < nr_blkz; i++) {
-			ret = dmz_init_zone(zmd, zone, &blkz[i]);
-			if (ret)
-				goto out;
-			sector += dev->zone_nr_sectors;
-			zone++;
-		}
-	}
+static int dmz_update_zone_cb(struct blk_zone *blkz, unsigned int idx,
+			      void *data)
+{
+	struct dm_zone *zone = data;
 
-	/* The entire zone configuration of the disk should now be known */
-	if (sector < dev->capacity) {
-		dmz_dev_err(dev, "Failed to get correct zone information");
-		ret = -ENXIO;
-	}
-out:
-	kfree(blkz);
-	if (ret)
-		dmz_drop_zones(zmd);
+	clear_bit(DMZ_OFFLINE, &zone->flags);
+	clear_bit(DMZ_READ_ONLY, &zone->flags);
+	if (blkz->cond == BLK_ZONE_COND_OFFLINE)
+		set_bit(DMZ_OFFLINE, &zone->flags);
+	else if (blkz->cond == BLK_ZONE_COND_READONLY)
+		set_bit(DMZ_READ_ONLY, &zone->flags);
 
-	return ret;
+	if (dmz_is_seq(zone))
+		zone->wp_block = dmz_sect2blk(blkz->wp - blkz->start);
+	else
+		zone->wp_block = 0;
+	return 0;
 }
 
 /*
@@ -1224,9 +1202,7 @@ out:
  */
 static int dmz_update_zone(struct dmz_metadata *zmd, struct dm_zone *zone)
 {
-	unsigned int nr_blkz = 1;
 	unsigned int noio_flag;
-	struct blk_zone blkz;
 	int ret;
 
 	/*
@@ -1236,29 +1212,18 @@ static int dmz_update_zone(struct dmz_metadata *zmd, struct dm_zone *zone)
 	 * GFP_NOIO was specified.
 	 */
 	noio_flag = memalloc_noio_save();
-	ret = blkdev_report_zones(zmd->dev->bdev, dmz_start_sect(zmd, zone),
-				  &blkz, &nr_blkz);
+	ret = blkdev_report_zones(zmd->dev->bdev, dmz_start_sect(zmd, zone), 1,
+				  dmz_update_zone_cb, zone);
 	memalloc_noio_restore(noio_flag);
-	if (!nr_blkz)
+
+	if (ret == 0)
 		ret = -EIO;
-	if (ret) {
+	if (ret < 0) {
 		dmz_dev_err(zmd->dev, "Get zone %u report failed",
 			    dmz_id(zmd, zone));
 		return ret;
 	}
 
-	clear_bit(DMZ_OFFLINE, &zone->flags);
-	clear_bit(DMZ_READ_ONLY, &zone->flags);
-	if (blkz.cond == BLK_ZONE_COND_OFFLINE)
-		set_bit(DMZ_OFFLINE, &zone->flags);
-	else if (blkz.cond == BLK_ZONE_COND_READONLY)
-		set_bit(DMZ_READ_ONLY, &zone->flags);
-
-	if (dmz_is_seq(zone))
-		zone->wp_block = dmz_sect2blk(blkz.wp - blkz.start);
-	else
-		zone->wp_block = 0;
-
 	return 0;
 }
 
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 76f4cfdd6b41..e8f9661a10a1 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -440,14 +440,48 @@ static int dm_blk_getgeo(struct block_device *bdev, struct hd_geometry *geo)
 	return dm_get_geometry(md, geo);
 }
 
+#ifdef CONFIG_BLK_DEV_ZONED
+int dm_report_zones_cb(struct blk_zone *zone, unsigned int idx, void *data)
+{
+	struct dm_report_zones_args *args = data;
+	sector_t sector_diff = args->tgt->begin - args->start;
+
+	/*
+	 * Ignore zones beyond the target range.
+	 */
+	if (zone->start >= args->start + args->tgt->len)
+		return 0;
+
+	/*
+	 * Remap the start sector and write pointer position of the zone
+	 * to match its position in the target range.
+	 */
+	zone->start += sector_diff;
+	if (zone->type != BLK_ZONE_TYPE_CONVENTIONAL) {
+		if (zone->cond == BLK_ZONE_COND_FULL)
+			zone->wp = zone->start + zone->len;
+		else if (zone->cond == BLK_ZONE_COND_EMPTY)
+			zone->wp = zone->start;
+		else
+			zone->wp += sector_diff;
+	}
+
+	args->next_sector = zone->start + zone->len;
+	return args->orig_cb(zone, args->zone_idx++, args->orig_data);
+}
+EXPORT_SYMBOL_GPL(dm_report_zones_cb);
+
 static int dm_blk_report_zones(struct gendisk *disk, sector_t sector,
-			       struct blk_zone *zones, unsigned int *nr_zones)
+		unsigned int nr_zones, report_zones_cb cb, void *data)
 {
-#ifdef CONFIG_BLK_DEV_ZONED
 	struct mapped_device *md = disk->private_data;
-	struct dm_target *tgt;
 	struct dm_table *map;
 	int srcu_idx, ret;
+	struct dm_report_zones_args args = {
+		.next_sector = sector,
+		.orig_data = data,
+		.orig_cb = cb,
+	};
 
 	if (dm_suspended_md(md))
 		return -EAGAIN;
@@ -456,32 +490,30 @@ static int dm_blk_report_zones(struct gendisk *disk, sector_t sector,
 	if (!map)
 		return -EIO;
 
-	tgt = dm_table_find_target(map, sector);
-	if (!tgt) {
-		ret = -EIO;
-		goto out;
-	}
+	do {
+		struct dm_target *tgt;
 
-	/*
-	 * If we are executing this, we already know that the block device
-	 * is a zoned device and so each target should have support for that
-	 * type of drive. A missing report_zones method means that the target
-	 * driver has a problem.
-	 */
-	if (WARN_ON(!tgt->type->report_zones)) {
-		ret = -EIO;
-		goto out;
-	}
+		tgt = dm_table_find_target(map, args.next_sector);
+		if (WARN_ON_ONCE(!tgt->type->report_zones)) {
+			ret = -EIO;
+			goto out;
+		}
 
-	ret = tgt->type->report_zones(tgt, sector, zones, nr_zones);
+		args.tgt = tgt;
+		ret = tgt->type->report_zones(tgt, &args, nr_zones);
+		if (ret < 0)
+			goto out;
+	} while (args.zone_idx < nr_zones &&
+		 args.next_sector < get_capacity(disk));
 
+	ret = args.zone_idx;
 out:
 	dm_put_live_table(md, srcu_idx);
 	return ret;
-#else
-	return -ENOTSUPP;
-#endif
 }
+#else
+#define dm_blk_report_zones		NULL
+#endif /* CONFIG_BLK_DEV_ZONED */
 
 static int dm_prepare_ioctl(struct mapped_device *md, int *srcu_idx,
 			    struct block_device **bdev)
@@ -1207,51 +1239,6 @@ void dm_accept_partial_bio(struct bio *bio, unsigned n_sectors)
 }
 EXPORT_SYMBOL_GPL(dm_accept_partial_bio);
 
-/*
- * The zone descriptors obtained with a zone report indicate
- * zone positions within the underlying device of the target. The zone
- * descriptors must be remapped to match their position within the dm device.
- */
-void dm_remap_zone_report(struct dm_target *ti, sector_t start,
-			  struct blk_zone *zones, unsigned int *nr_zones)
-{
-#ifdef CONFIG_BLK_DEV_ZONED
-	struct blk_zone *zone;
-	unsigned int nrz = *nr_zones;
-	int i;
-
-	/*
-	 * Remap the start sector and write pointer position of the zones in
-	 * the array. Since we may have obtained from the target underlying
-	 * device more zones that the target size, also adjust the number
-	 * of zones.
-	 */
-	for (i = 0; i < nrz; i++) {
-		zone = zones + i;
-		if (zone->start >= start + ti->len) {
-			memset(zone, 0, sizeof(struct blk_zone) * (nrz - i));
-			break;
-		}
-
-		zone->start = zone->start + ti->begin - start;
-		if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
-			continue;
-
-		if (zone->cond == BLK_ZONE_COND_FULL)
-			zone->wp = zone->start + zone->len;
-		else if (zone->cond == BLK_ZONE_COND_EMPTY)
-			zone->wp = zone->start;
-		else
-			zone->wp = zone->wp + ti->begin - start;
-	}
-
-	*nr_zones = i;
-#else /* !CONFIG_BLK_DEV_ZONED */
-	*nr_zones = 0;
-#endif
-}
-EXPORT_SYMBOL_GPL(dm_remap_zone_report);
-
 static blk_qc_t __map_bio(struct dm_target_io *tio)
 {
 	int r;