[PATCH v3 0/29] block: Make blkdev_get_by_*() return handle

[PATCH v3 0/29] block: Make blkdev_get_by_*() return handle

[Jan Kara]

Hello,

this is a v3 of the patch series which implements the idea of blkdev_get_by_*()
calls returning bdev_handle which is then passed to blkdev_put() [1]. This
makes the get and put calls for bdevs more obviously matching and allows us to
propagate context from get to put without having to modify all the users
(again!). In particular I need to propagate used open flags to blkdev_put() to
be able count writeable opens and add support for blocking writes to mounted
block devices. I'll send that series separately.

The series is based on Christian's vfs tree as of today as there is quite
some overlap. Patches have passed some reasonable testing - I've tested block
changes, md, dm, bcache, xfs, btrfs, ext4, swap. More testing or review is
always welcome. Thanks! I've pushed out the full branch to:

git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs.git bdev_handle

to ease review / testing. Since there were not many comments for v2 and
Christoph has acked the series I think we should start discussing how to merge
the series. Most collisions with this series seem to happen in the filesystems
area so VFS tree would seem as the least painful way to merge this. Jens,
are you OK with that?

Changes since v2:
* Rebased on top of current vfs tree
* Added some acks
* Reflected minor nits from Christoph
* Added missing conversion of blkdev_put() calls in cramfs and erofs
* Fixed possible leak of bdev handle in xfs if logdev is the same as fs dev

Changes since v1:
* Rebased on top of current vfs tree
* Renamed final functions to bdev_open_by_*() and bdev_release()
* Fixed detection of exclusive open in blkdev_ioctl() and blkdev_fallocate()
* Fixed swap conversion to properly reinitialize swap_info->bdev_handle
* Fixed xfs conversion to not oops with rtdev without logdev
* Couple other minor fixups

								Honza

[1] https://lore.kernel.org/all/ZJGNsVDhZx0Xgs2H@infradead.org

CC: Alasdair Kergon <agk@redhat.com>
CC: Andrew Morton <akpm@linux-foundation.org>
CC: Anna Schumaker <anna@kernel.org>
CC: Chao Yu <chao@kernel.org>
CC: Christian Borntraeger <borntraeger@linux.ibm.com>
CC: Coly Li <colyli@suse.de
CC: "Darrick J. Wong" <djwong@kernel.org>
CC: Dave Kleikamp <shaggy@kernel.org>
CC: David Sterba <dsterba@suse.com>
CC: dm-devel@redhat.com
CC: drbd-dev@lists.linbit.com
CC: Gao Xiang <xiang@kernel.org>
CC: Jack Wang <jinpu.wang@ionos.com>
CC: Jaegeuk Kim <jaegeuk@kernel.org>
CC: jfs-discussion@lists.sourceforge.net
CC: Joern Engel <joern@lazybastard.org>
CC: Joseph Qi <joseph.qi@linux.alibaba.com>
CC: Kent Overstreet <kent.overstreet@gmail.com>
CC: linux-bcache@vger.kernel.org
CC: linux-btrfs@vger.kernel.org
CC: linux-erofs@lists.ozlabs.org
CC: <linux-ext4@vger.kernel.org>
CC: linux-f2fs-devel@lists.sourceforge.net
CC: linux-mm@kvack.org
CC: linux-mtd@lists.infradead.org
CC: linux-nfs@vger.kernel.org
CC: linux-nilfs@vger.kernel.org
CC: linux-nvme@lists.infradead.org
CC: linux-pm@vger.kernel.org
CC: linux-raid@vger.kernel.org
CC: linux-s390@vger.kernel.org
CC: linux-scsi@vger.kernel.org
CC: linux-xfs@vger.kernel.org
CC: "Md. Haris Iqbal" <haris.iqbal@ionos.com>
CC: Mike Snitzer <snitzer@kernel.org>
CC: Minchan Kim <minchan@kernel.org>
CC: ocfs2-devel@oss.oracle.com
CC: reiserfs-devel@vger.kernel.org
CC: Sergey Senozhatsky <senozhatsky@chromium.org>
CC: Song Liu <song@kernel.org>
CC: Sven Schnelle <svens@linux.ibm.com>
CC: target-devel@vger.kernel.org
CC: Ted Tso <tytso@mit.edu>
CC: Trond Myklebust <trond.myklebust@hammerspace.com>
CC: xen-devel@lists.xenproject.org

Previous versions:
Link: http://lore.kernel.org/r/20230629165206.383-1-jack@suse.cz # v1
Link: http://lore.kernel.org/r/20230810171429.31759-1-jack@suse.cz # v2

[PATCH 20/29] btrfs: Convert to bdev_open_by_path()

[Jan Kara]

Convert btrfs to use bdev_open_by_path() and pass the handle around.  We
also drop the holder from struct btrfs_device as it is now not needed
anymore.

CC: David Sterba <dsterba@suse.com>
CC: linux-btrfs@vger.kernel.org
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
---
 fs/btrfs/dev-replace.c |  14 +++---
 fs/btrfs/ioctl.c       |  18 +++----
 fs/btrfs/volumes.c     | 107 +++++++++++++++++++++--------------------
 fs/btrfs/volumes.h     |   6 +--
 4 files changed, 73 insertions(+), 72 deletions(-)

diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index 5f10965fd72b..fec013c5f26c 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -247,6 +247,7 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 {
 	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
 	struct btrfs_device *device;
+	struct bdev_handle *bdev_handle;
 	struct block_device *bdev;
 	u64 devid = BTRFS_DEV_REPLACE_DEVID;
 	int ret = 0;
@@ -257,12 +258,13 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 		return -EINVAL;
 	}
 
-	bdev = blkdev_get_by_path(device_path, BLK_OPEN_WRITE,
-				  fs_info->bdev_holder, NULL);
-	if (IS_ERR(bdev)) {
+	bdev_handle = bdev_open_by_path(device_path, BLK_OPEN_WRITE,
+					fs_info->bdev_holder, NULL);
+	if (IS_ERR(bdev_handle)) {
 		btrfs_err(fs_info, "target device %s is invalid!", device_path);
-		return PTR_ERR(bdev);
+		return PTR_ERR(bdev_handle);
 	}
+	bdev = bdev_handle->bdev;
 
 	if (!btrfs_check_device_zone_type(fs_info, bdev)) {
 		btrfs_err(fs_info,
@@ -313,9 +315,9 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	device->commit_bytes_used = device->bytes_used;
 	device->fs_info = fs_info;
 	device->bdev = bdev;
+	device->bdev_handle = bdev_handle;
 	set_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
 	set_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state);
-	device->holder = fs_info->bdev_holder;
 	device->dev_stats_valid = 1;
 	set_blocksize(device->bdev, BTRFS_BDEV_BLOCKSIZE);
 	device->fs_devices = fs_devices;
@@ -334,7 +336,7 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	return 0;
 
 error:
-	blkdev_put(bdev, fs_info->bdev_holder);
+	bdev_release(bdev_handle);
 	return ret;
 }
 
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index a18ee7b5a166..b4074191fcc7 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -2670,8 +2670,7 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 	struct inode *inode = file_inode(file);
 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
 	struct btrfs_ioctl_vol_args_v2 *vol_args;
-	struct block_device *bdev = NULL;
-	void *holder;
+	struct bdev_handle *bdev_handle = NULL;
 	int ret;
 	bool cancel = false;
 
@@ -2708,7 +2707,7 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 		goto err_drop;
 
 	/* Exclusive operation is now claimed */
-	ret = btrfs_rm_device(fs_info, &args, &bdev, &holder);
+	ret = btrfs_rm_device(fs_info, &args, &bdev_handle);
 
 	btrfs_exclop_finish(fs_info);
 
@@ -2722,8 +2721,8 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 	}
 err_drop:
 	mnt_drop_write_file(file);
-	if (bdev)
-		blkdev_put(bdev, holder);
+	if (bdev_handle)
+		bdev_release(bdev_handle);
 out:
 	btrfs_put_dev_args_from_path(&args);
 	kfree(vol_args);
@@ -2736,8 +2735,7 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
 	struct inode *inode = file_inode(file);
 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
 	struct btrfs_ioctl_vol_args *vol_args;
-	struct block_device *bdev = NULL;
-	void *holder;
+	struct bdev_handle *bdev_handle = NULL;
 	int ret;
 	bool cancel = false;
 
@@ -2764,15 +2762,15 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
 	ret = exclop_start_or_cancel_reloc(fs_info, BTRFS_EXCLOP_DEV_REMOVE,
 					   cancel);
 	if (ret == 0) {
-		ret = btrfs_rm_device(fs_info, &args, &bdev, &holder);
+		ret = btrfs_rm_device(fs_info, &args, &bdev_handle);
 		if (!ret)
 			btrfs_info(fs_info, "disk deleted %s", vol_args->name);
 		btrfs_exclop_finish(fs_info);
 	}
 
 	mnt_drop_write_file(file);
-	if (bdev)
-		blkdev_put(bdev, holder);
+	if (bdev_handle)
+		bdev_release(bdev_handle);
 out:
 	btrfs_put_dev_args_from_path(&args);
 	kfree(vol_args);
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index fc1bd0b03d63..9c4b5b825b5c 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -512,37 +512,39 @@ static struct btrfs_fs_devices *find_fsid_with_metadata_uuid(
 
 static int
 btrfs_get_bdev_and_sb(const char *device_path, blk_mode_t flags, void *holder,
-		      int flush, struct block_device **bdev,
+		      int flush, struct bdev_handle **bdev_handle,
 		      struct btrfs_super_block **disk_super)
 {
+	struct block_device *bdev;
 	int ret;
 
-	*bdev = blkdev_get_by_path(device_path, flags, holder, NULL);
+	*bdev_handle = bdev_open_by_path(device_path, flags, holder, NULL);
 
-	if (IS_ERR(*bdev)) {
-		ret = PTR_ERR(*bdev);
+	if (IS_ERR(*bdev_handle)) {
+		ret = PTR_ERR(*bdev_handle);
 		goto error;
 	}
+	bdev = (*bdev_handle)->bdev;
 
 	if (flush)
-		sync_blockdev(*bdev);
-	ret = set_blocksize(*bdev, BTRFS_BDEV_BLOCKSIZE);
+		sync_blockdev(bdev);
+	ret = set_blocksize(bdev, BTRFS_BDEV_BLOCKSIZE);
 	if (ret) {
-		blkdev_put(*bdev, holder);
+		bdev_release(*bdev_handle);
 		goto error;
 	}
-	invalidate_bdev(*bdev);
-	*disk_super = btrfs_read_dev_super(*bdev);
+	invalidate_bdev(bdev);
+	*disk_super = btrfs_read_dev_super(bdev);
 	if (IS_ERR(*disk_super)) {
 		ret = PTR_ERR(*disk_super);
-		blkdev_put(*bdev, holder);
+		bdev_release(*bdev_handle);
 		goto error;
 	}
 
 	return 0;
 
 error:
-	*bdev = NULL;
+	*bdev_handle = NULL;
 	return ret;
 }
 
@@ -614,7 +616,7 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 			struct btrfs_device *device, blk_mode_t flags,
 			void *holder)
 {
-	struct block_device *bdev;
+	struct bdev_handle *bdev_handle;
 	struct btrfs_super_block *disk_super;
 	u64 devid;
 	int ret;
@@ -625,7 +627,7 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 		return -EINVAL;
 
 	ret = btrfs_get_bdev_and_sb(device->name->str, flags, holder, 1,
-				    &bdev, &disk_super);
+				    &bdev_handle, &disk_super);
 	if (ret)
 		return ret;
 
@@ -649,21 +651,21 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 		clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 		fs_devices->seeding = true;
 	} else {
-		if (bdev_read_only(bdev))
+		if (bdev_read_only(bdev_handle->bdev))
 			clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 		else
 			set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 	}
 
-	if (!bdev_nonrot(bdev))
+	if (!bdev_nonrot(bdev_handle->bdev))
 		fs_devices->rotating = true;
 
-	if (bdev_max_discard_sectors(bdev))
+	if (bdev_max_discard_sectors(bdev_handle->bdev))
 		fs_devices->discardable = true;
 
-	device->bdev = bdev;
+	device->bdev_handle = bdev_handle;
+	device->bdev = bdev_handle->bdev;
 	clear_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
-	device->holder = holder;
 
 	fs_devices->open_devices++;
 	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state) &&
@@ -677,7 +679,7 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 
 error_free_page:
 	btrfs_release_disk_super(disk_super);
-	blkdev_put(bdev, holder);
+	bdev_release(bdev_handle);
 
 	return -EINVAL;
 }
@@ -1067,9 +1069,10 @@ static void __btrfs_free_extra_devids(struct btrfs_fs_devices *fs_devices,
 		if (device->devid == BTRFS_DEV_REPLACE_DEVID)
 			continue;
 
-		if (device->bdev) {
-			blkdev_put(device->bdev, device->holder);
+		if (device->bdev_handle) {
+			bdev_release(device->bdev_handle);
 			device->bdev = NULL;
+			device->bdev_handle = NULL;
 			fs_devices->open_devices--;
 		}
 		if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state)) {
@@ -1114,7 +1117,7 @@ static void btrfs_close_bdev(struct btrfs_device *device)
 		invalidate_bdev(device->bdev);
 	}
 
-	blkdev_put(device->bdev, device->holder);
+	bdev_release(device->bdev_handle);
 }
 
 static void btrfs_close_one_device(struct btrfs_device *device)
@@ -1359,7 +1362,7 @@ struct btrfs_device *btrfs_scan_one_device(const char *path)
 	struct btrfs_super_block *disk_super;
 	bool new_device_added = false;
 	struct btrfs_device *device = NULL;
-	struct block_device *bdev;
+	struct bdev_handle *bdev_handle;
 	u64 bytenr, bytenr_orig;
 	int ret;
 
@@ -1382,18 +1385,19 @@ struct btrfs_device *btrfs_scan_one_device(const char *path)
 	 * values temporarily, as the device paths of the fsid are the only
 	 * required information for assembling the volume.
 	 */
-	bdev = blkdev_get_by_path(path, BLK_OPEN_READ, NULL, NULL);
-	if (IS_ERR(bdev))
-		return ERR_CAST(bdev);
+	bdev_handle = bdev_open_by_path(path, BLK_OPEN_READ, NULL, NULL);
+	if (IS_ERR(bdev_handle))
+		return ERR_CAST(bdev_handle);
 
 	bytenr_orig = btrfs_sb_offset(0);
-	ret = btrfs_sb_log_location_bdev(bdev, 0, READ, &bytenr);
+	ret = btrfs_sb_log_location_bdev(bdev_handle->bdev, 0, READ, &bytenr);
 	if (ret) {
 		device = ERR_PTR(ret);
 		goto error_bdev_put;
 	}
 
-	disk_super = btrfs_read_disk_super(bdev, bytenr, bytenr_orig);
+	disk_super = btrfs_read_disk_super(bdev_handle->bdev, bytenr,
+					   bytenr_orig);
 	if (IS_ERR(disk_super)) {
 		device = ERR_CAST(disk_super);
 		goto error_bdev_put;
@@ -1406,7 +1410,7 @@ struct btrfs_device *btrfs_scan_one_device(const char *path)
 	btrfs_release_disk_super(disk_super);
 
 error_bdev_put:
-	blkdev_put(bdev, NULL);
+	bdev_release(bdev_handle);
 
 	return device;
 }
@@ -2091,7 +2095,7 @@ void btrfs_scratch_superblocks(struct btrfs_fs_info *fs_info,
 
 int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 		    struct btrfs_dev_lookup_args *args,
-		    struct block_device **bdev, void **holder)
+		    struct bdev_handle **bdev_handle)
 {
 	struct btrfs_trans_handle *trans;
 	struct btrfs_device *device;
@@ -2200,7 +2204,7 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 
 	btrfs_assign_next_active_device(device, NULL);
 
-	if (device->bdev) {
+	if (device->bdev_handle) {
 		cur_devices->open_devices--;
 		/* remove sysfs entry */
 		btrfs_sysfs_remove_device(device);
@@ -2216,9 +2220,9 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 	 * free the device.
 	 *
 	 * We cannot call btrfs_close_bdev() here because we're holding the sb
-	 * write lock, and blkdev_put() will pull in the ->open_mutex on the
-	 * block device and it's dependencies.  Instead just flush the device
-	 * and let the caller do the final blkdev_put.
+	 * write lock, and bdev_release() will pull in the ->open_mutex on
+	 * the block device and it's dependencies.  Instead just flush the
+	 * device and let the caller do the final bdev_release.
 	 */
 	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state)) {
 		btrfs_scratch_superblocks(fs_info, device->bdev,
@@ -2229,8 +2233,7 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 		}
 	}
 
-	*bdev = device->bdev;
-	*holder = device->holder;
+	*bdev_handle = device->bdev_handle;
 	synchronize_rcu();
 	btrfs_free_device(device);
 
@@ -2368,7 +2371,7 @@ int btrfs_get_dev_args_from_path(struct btrfs_fs_info *fs_info,
 				 const char *path)
 {
 	struct btrfs_super_block *disk_super;
-	struct block_device *bdev;
+	struct bdev_handle *bdev_handle;
 	int ret;
 
 	if (!path || !path[0])
@@ -2386,7 +2389,7 @@ int btrfs_get_dev_args_from_path(struct btrfs_fs_info *fs_info,
 	}
 
 	ret = btrfs_get_bdev_and_sb(path, BLK_OPEN_READ, NULL, 0,
-				    &bdev, &disk_super);
+				    &bdev_handle, &disk_super);
 	if (ret) {
 		btrfs_put_dev_args_from_path(args);
 		return ret;
@@ -2399,7 +2402,7 @@ int btrfs_get_dev_args_from_path(struct btrfs_fs_info *fs_info,
 	else
 		memcpy(args->fsid, disk_super->fsid, BTRFS_FSID_SIZE);
 	btrfs_release_disk_super(disk_super);
-	blkdev_put(bdev, NULL);
+	bdev_release(bdev_handle);
 	return 0;
 }
 
@@ -2620,7 +2623,7 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	struct btrfs_root *root = fs_info->dev_root;
 	struct btrfs_trans_handle *trans;
 	struct btrfs_device *device;
-	struct block_device *bdev;
+	struct bdev_handle *bdev_handle;
 	struct super_block *sb = fs_info->sb;
 	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
 	struct btrfs_fs_devices *seed_devices = NULL;
@@ -2633,12 +2636,12 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	if (sb_rdonly(sb) && !fs_devices->seeding)
 		return -EROFS;
 
-	bdev = blkdev_get_by_path(device_path, BLK_OPEN_WRITE,
-				  fs_info->bdev_holder, NULL);
-	if (IS_ERR(bdev))
-		return PTR_ERR(bdev);
+	bdev_handle = bdev_open_by_path(device_path, BLK_OPEN_WRITE,
+					fs_info->bdev_holder, NULL);
+	if (IS_ERR(bdev_handle))
+		return PTR_ERR(bdev_handle);
 
-	if (!btrfs_check_device_zone_type(fs_info, bdev)) {
+	if (!btrfs_check_device_zone_type(fs_info, bdev_handle->bdev)) {
 		ret = -EINVAL;
 		goto error;
 	}
@@ -2650,11 +2653,11 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 		locked = true;
 	}
 
-	sync_blockdev(bdev);
+	sync_blockdev(bdev_handle->bdev);
 
 	rcu_read_lock();
 	list_for_each_entry_rcu(device, &fs_devices->devices, dev_list) {
-		if (device->bdev == bdev) {
+		if (device->bdev == bdev_handle->bdev) {
 			ret = -EEXIST;
 			rcu_read_unlock();
 			goto error;
@@ -2670,7 +2673,8 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	}
 
 	device->fs_info = fs_info;
-	device->bdev = bdev;
+	device->bdev_handle = bdev_handle;
+	device->bdev = bdev_handle->bdev;
 	ret = lookup_bdev(device_path, &device->devt);
 	if (ret)
 		goto error_free_device;
@@ -2691,12 +2695,11 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	device->io_align = fs_info->sectorsize;
 	device->sector_size = fs_info->sectorsize;
 	device->total_bytes =
-		round_down(bdev_nr_bytes(bdev), fs_info->sectorsize);
+		round_down(bdev_nr_bytes(device->bdev), fs_info->sectorsize);
 	device->disk_total_bytes = device->total_bytes;
 	device->commit_total_bytes = device->total_bytes;
 	set_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
 	clear_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state);
-	device->holder = fs_info->bdev_holder;
 	device->dev_stats_valid = 1;
 	set_blocksize(device->bdev, BTRFS_BDEV_BLOCKSIZE);
 
@@ -2732,7 +2735,7 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 
 	atomic64_add(device->total_bytes, &fs_info->free_chunk_space);
 
-	if (!bdev_nonrot(bdev))
+	if (!bdev_nonrot(device->bdev))
 		fs_devices->rotating = true;
 
 	orig_super_total_bytes = btrfs_super_total_bytes(fs_info->super_copy);
@@ -2854,7 +2857,7 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 error_free_device:
 	btrfs_free_device(device);
 error:
-	blkdev_put(bdev, fs_info->bdev_holder);
+	bdev_release(bdev_handle);
 	if (locked) {
 		mutex_unlock(&uuid_mutex);
 		up_write(&sb->s_umount);
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index f472d646715e..d92ebeaf1200 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -90,13 +90,11 @@ struct btrfs_device {
 
 	u64 generation;
 
+	struct bdev_handle *bdev_handle;
 	struct block_device *bdev;
 
 	struct btrfs_zoned_device_info *zone_info;
 
-	/* block device holder for blkdev_get/put */
-	void *holder;
-
 	/*
 	 * Device's major-minor number. Must be set even if the device is not
 	 * opened (bdev == NULL), unless the device is missing.
@@ -631,7 +629,7 @@ struct btrfs_device *btrfs_alloc_device(struct btrfs_fs_info *fs_info,
 void btrfs_put_dev_args_from_path(struct btrfs_dev_lookup_args *args);
 int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 		    struct btrfs_dev_lookup_args *args,
-		    struct block_device **bdev, void **holder);
+		    struct bdev_handle **bdev_handle);
 void __exit btrfs_cleanup_fs_uuids(void);
 int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len);
 int btrfs_grow_device(struct btrfs_trans_handle *trans,
-- 
2.35.3

Re: [PATCH 20/29] btrfs: Convert to bdev_open_by_path()

[Christian Brauner]

On Wed, Aug 23, 2023 at 12:48:31PM +0200, Jan Kara wrote:
> Convert btrfs to use bdev_open_by_path() and pass the handle around.  We
> also drop the holder from struct btrfs_device as it is now not needed
> anymore.
> 
> CC: David Sterba <dsterba@suse.com>
> CC: linux-btrfs@vger.kernel.org
> Acked-by: Christoph Hellwig <hch@lst.de>
> Signed-off-by: Jan Kara <jack@suse.cz>
> ---

Looks good to me,
Reviewed-by: Christian Brauner <brauner@kernel.org>

> -	bdev = blkdev_get_by_path(path, BLK_OPEN_READ, NULL, NULL);
> -	if (IS_ERR(bdev))
> -		return ERR_CAST(bdev);
> +	bdev_handle = bdev_open_by_path(path, BLK_OPEN_READ, NULL, NULL);
> +	if (IS_ERR(bdev_handle))
> +		return ERR_CAST(bdev_handle);

Minor merge conflict with what's in vfs.super because we dropped
unrelated btrfs patches like David asked us to. Very easy to fix for me.
Just an fyi.

Re: [PATCH v3 0/29] block: Make blkdev_get_by_*() return handle

[Christian Brauner]

On Wed, Aug 23, 2023 at 12:48:11PM +0200, Jan Kara wrote:
> Hello,
> 
> this is a v3 of the patch series which implements the idea of blkdev_get_by_*()
> calls returning bdev_handle which is then passed to blkdev_put() [1]. This
> makes the get and put calls for bdevs more obviously matching and allows us to
> propagate context from get to put without having to modify all the users
> (again!). In particular I need to propagate used open flags to blkdev_put() to
> be able count writeable opens and add support for blocking writes to mounted
> block devices. I'll send that series separately.
> 
> The series is based on Christian's vfs tree as of today as there is quite
> some overlap. Patches have passed some reasonable testing - I've tested block
> changes, md, dm, bcache, xfs, btrfs, ext4, swap. More testing or review is
> always welcome. Thanks! I've pushed out the full branch to:
> 
> git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs.git bdev_handle
> 
> to ease review / testing. Since there were not many comments for v2 and
> Christoph has acked the series I think we should start discussing how to merge
> the series. Most collisions with this series seem to happen in the filesystems
> area so VFS tree would seem as the least painful way to merge this. Jens,

I really do like this series especially struct bdev_handle and moving
the mode bits in there. I'll happily take this. So far there have only
been minor things that can easily be fixed.

Re: [PATCH v3 0/29] block: Make blkdev_get_by_*() return handle

[Jan Kara]

On Fri 25-08-23 15:32:47, Christian Brauner wrote:
> On Wed, Aug 23, 2023 at 12:48:11PM +0200, Jan Kara wrote:
> > Hello,
> > 
> > this is a v3 of the patch series which implements the idea of blkdev_get_by_*()
> > calls returning bdev_handle which is then passed to blkdev_put() [1]. This
> > makes the get and put calls for bdevs more obviously matching and allows us to
> > propagate context from get to put without having to modify all the users
> > (again!). In particular I need to propagate used open flags to blkdev_put() to
> > be able count writeable opens and add support for blocking writes to mounted
> > block devices. I'll send that series separately.
> > 
> > The series is based on Christian's vfs tree as of today as there is quite
> > some overlap. Patches have passed some reasonable testing - I've tested block
> > changes, md, dm, bcache, xfs, btrfs, ext4, swap. More testing or review is
> > always welcome. Thanks! I've pushed out the full branch to:
> > 
> > git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs.git bdev_handle
> > 
> > to ease review / testing. Since there were not many comments for v2 and
> > Christoph has acked the series I think we should start discussing how to merge
> > the series. Most collisions with this series seem to happen in the filesystems
> > area so VFS tree would seem as the least painful way to merge this. Jens,
> 
> I really do like this series especially struct bdev_handle and moving
> the mode bits in there. I'll happily take this. So far there have only
> been minor things that can easily be fixed.

Thanks. Since Al is fine with just doing a potential conversion to 'struct
file' as a handle on top of this series (it will be dumb Coccinelle
replacement) I think we can go ahead with the series as is. As you said
there will be some conflicts in btrfs and I've learned about f2fs conflicts
as well so I can rebase & repost the series on top of rc1 to make life
easier for you.

								Honza
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Re: [PATCH v3 0/29] block: Make blkdev_get_by_*() return handle

[Christian Brauner]

> replacement) I think we can go ahead with the series as is. As you said
> there will be some conflicts in btrfs and I've learned about f2fs conflicts
> as well so I can rebase & repost the series on top of rc1 to make life
> easier for you.

That is be much appreciated. Thank you!

[PATCH v4 0/29] block: Make blkdev_get_by_*() return handle

[Jan Kara]

Hello,

this is a v3 of the patch series which implements the idea of blkdev_get_by_*()
calls returning bdev_handle which is then passed to blkdev_put() [1]. This
makes the get and put calls for bdevs more obviously matching and allows us to
propagate context from get to put without having to modify all the users
(again!). In particular I need to propagate used open flags to blkdev_put() to
be able count writeable opens and add support for blocking writes to mounted
block devices. I'll send that series separately.

The series is based on Btrfs tree's for-next branch [2] as of today as the
series depends on Christoph's changes to btrfs device handling.  Patches have
passed some reasonable testing - I've tested block changes, md, dm, bcache,
xfs, btrfs, ext4, swap. More testing or review is always welcome. Thanks! I've
pushed out the full branch to:

git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs.git bdev_handle

to ease review / testing. Christian, can you pull the patches to your tree
to get some exposure in linux-next as well? Thanks!

Changes since v3:
* Rebased on top on btrfs tree

Changes since v2:
* Rebased on top of current vfs tree
* Added some acks
* Reflected minor nits from Christoph
* Added missing conversion of blkdev_put() calls in cramfs and erofs
* Fixed possible leak of bdev handle in xfs if logdev is the same as fs dev

Changes since v1:
* Rebased on top of current vfs tree
* Renamed final functions to bdev_open_by_*() and bdev_release()
* Fixed detection of exclusive open in blkdev_ioctl() and blkdev_fallocate()
* Fixed swap conversion to properly reinitialize swap_info->bdev_handle
* Fixed xfs conversion to not oops with rtdev without logdev
* Couple other minor fixups

								Honza

[1] https://lore.kernel.org/all/ZJGNsVDhZx0Xgs2H@infradead.org
[2] git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux.git for-next

CC: Alasdair Kergon <agk@redhat.com>
CC: Andrew Morton <akpm@linux-foundation.org>
CC: Anna Schumaker <anna@kernel.org>
CC: Chao Yu <chao@kernel.org>
CC: Christian Borntraeger <borntraeger@linux.ibm.com>
CC: Coly Li <colyli@suse.de
CC: "Darrick J. Wong" <djwong@kernel.org>
CC: Dave Kleikamp <shaggy@kernel.org>
CC: David Sterba <dsterba@suse.com>
CC: dm-devel@redhat.com
CC: drbd-dev@lists.linbit.com
CC: Gao Xiang <xiang@kernel.org>
CC: Jack Wang <jinpu.wang@ionos.com>
CC: Jaegeuk Kim <jaegeuk@kernel.org>
CC: jfs-discussion@lists.sourceforge.net
CC: Joern Engel <joern@lazybastard.org>
CC: Joseph Qi <joseph.qi@linux.alibaba.com>
CC: Kent Overstreet <kent.overstreet@gmail.com>
CC: linux-bcache@vger.kernel.org
CC: linux-btrfs@vger.kernel.org
CC: linux-erofs@lists.ozlabs.org
CC: <linux-ext4@vger.kernel.org>
CC: linux-f2fs-devel@lists.sourceforge.net
CC: linux-mm@kvack.org
CC: linux-mtd@lists.infradead.org
CC: linux-nfs@vger.kernel.org
CC: linux-nilfs@vger.kernel.org
CC: linux-nvme@lists.infradead.org
CC: linux-pm@vger.kernel.org
CC: linux-raid@vger.kernel.org
CC: linux-s390@vger.kernel.org
CC: linux-scsi@vger.kernel.org
CC: linux-xfs@vger.kernel.org
CC: "Md. Haris Iqbal" <haris.iqbal@ionos.com>
CC: Mike Snitzer <snitzer@kernel.org>
CC: Minchan Kim <minchan@kernel.org>
CC: ocfs2-devel@oss.oracle.com
CC: reiserfs-devel@vger.kernel.org
CC: Sergey Senozhatsky <senozhatsky@chromium.org>
CC: Song Liu <song@kernel.org>
CC: Sven Schnelle <svens@linux.ibm.com>
CC: target-devel@vger.kernel.org
CC: Ted Tso <tytso@mit.edu>
CC: Trond Myklebust <trond.myklebust@hammerspace.com>
CC: xen-devel@lists.xenproject.org

Previous versions:
Link: http://lore.kernel.org/r/20230629165206.383-1-jack@suse.cz # v1
Link: http://lore.kernel.org/r/20230810171429.31759-1-jack@suse.cz # v2

[PATCH 20/29] btrfs: Convert to bdev_open_by_path()

[Jan Kara]

Convert btrfs to use bdev_open_by_path() and pass the handle around.  We
also drop the holder from struct btrfs_device as it is now not needed
anymore.

CC: David Sterba <dsterba@suse.com>
CC: linux-btrfs@vger.kernel.org
Acked-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Jan Kara <jack@suse.cz>
---
 fs/btrfs/dev-replace.c |  14 +++---
 fs/btrfs/ioctl.c       |  18 +++----
 fs/btrfs/volumes.c     | 107 +++++++++++++++++++++--------------------
 fs/btrfs/volumes.h     |   6 +--
 4 files changed, 73 insertions(+), 72 deletions(-)

diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index d3998cad62c2..f9544fda38e9 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -246,6 +246,7 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 {
 	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
 	struct btrfs_device *device;
+	struct bdev_handle *bdev_handle;
 	struct block_device *bdev;
 	u64 devid = BTRFS_DEV_REPLACE_DEVID;
 	int ret = 0;
@@ -256,12 +257,13 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 		return -EINVAL;
 	}
 
-	bdev = blkdev_get_by_path(device_path, BLK_OPEN_WRITE,
-				  fs_info->bdev_holder, NULL);
-	if (IS_ERR(bdev)) {
+	bdev_handle = bdev_open_by_path(device_path, BLK_OPEN_WRITE,
+					fs_info->bdev_holder, NULL);
+	if (IS_ERR(bdev_handle)) {
 		btrfs_err(fs_info, "target device %s is invalid!", device_path);
-		return PTR_ERR(bdev);
+		return PTR_ERR(bdev_handle);
 	}
+	bdev = bdev_handle->bdev;
 
 	if (!btrfs_check_device_zone_type(fs_info, bdev)) {
 		btrfs_err(fs_info,
@@ -312,9 +314,9 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	device->commit_bytes_used = device->bytes_used;
 	device->fs_info = fs_info;
 	device->bdev = bdev;
+	device->bdev_handle = bdev_handle;
 	set_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
 	set_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state);
-	device->holder = fs_info->bdev_holder;
 	device->dev_stats_valid = 1;
 	set_blocksize(device->bdev, BTRFS_BDEV_BLOCKSIZE);
 	device->fs_devices = fs_devices;
@@ -333,7 +335,7 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	return 0;
 
 error:
-	blkdev_put(bdev, fs_info->bdev_holder);
+	bdev_release(bdev_handle);
 	return ret;
 }
 
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 9c05216de36a..59edbf47dbb2 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -2676,8 +2676,7 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 	struct inode *inode = file_inode(file);
 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
 	struct btrfs_ioctl_vol_args_v2 *vol_args;
-	struct block_device *bdev = NULL;
-	void *holder;
+	struct bdev_handle *bdev_handle = NULL;
 	int ret;
 	bool cancel = false;
 
@@ -2720,7 +2719,7 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 		goto err_drop;
 
 	/* Exclusive operation is now claimed */
-	ret = btrfs_rm_device(fs_info, &args, &bdev, &holder);
+	ret = btrfs_rm_device(fs_info, &args, &bdev_handle);
 
 	btrfs_exclop_finish(fs_info);
 
@@ -2734,8 +2733,8 @@ static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
 	}
 err_drop:
 	mnt_drop_write_file(file);
-	if (bdev)
-		blkdev_put(bdev, holder);
+	if (bdev_handle)
+		bdev_release(bdev_handle);
 out:
 	btrfs_put_dev_args_from_path(&args);
 	kfree(vol_args);
@@ -2748,8 +2747,7 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
 	struct inode *inode = file_inode(file);
 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
 	struct btrfs_ioctl_vol_args *vol_args;
-	struct block_device *bdev = NULL;
-	void *holder;
+	struct bdev_handle *bdev_handle = NULL;
 	int ret;
 	bool cancel = false;
 
@@ -2782,15 +2780,15 @@ static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
 	ret = exclop_start_or_cancel_reloc(fs_info, BTRFS_EXCLOP_DEV_REMOVE,
 					   cancel);
 	if (ret == 0) {
-		ret = btrfs_rm_device(fs_info, &args, &bdev, &holder);
+		ret = btrfs_rm_device(fs_info, &args, &bdev_handle);
 		if (!ret)
 			btrfs_info(fs_info, "disk deleted %s", vol_args->name);
 		btrfs_exclop_finish(fs_info);
 	}
 
 	mnt_drop_write_file(file);
-	if (bdev)
-		blkdev_put(bdev, holder);
+	if (bdev_handle)
+		bdev_release(bdev_handle);
 out:
 	btrfs_put_dev_args_from_path(&args);
 	kfree(vol_args);
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index ae64c9885851..d98caa86df4c 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -458,37 +458,39 @@ static noinline struct btrfs_fs_devices *find_fsid(
 
 static int
 btrfs_get_bdev_and_sb(const char *device_path, blk_mode_t flags, void *holder,
-		      int flush, struct block_device **bdev,
+		      int flush, struct bdev_handle **bdev_handle,
 		      struct btrfs_super_block **disk_super)
 {
+	struct block_device *bdev;
 	int ret;
 
-	*bdev = blkdev_get_by_path(device_path, flags, holder, NULL);
+	*bdev_handle = bdev_open_by_path(device_path, flags, holder, NULL);
 
-	if (IS_ERR(*bdev)) {
-		ret = PTR_ERR(*bdev);
+	if (IS_ERR(*bdev_handle)) {
+		ret = PTR_ERR(*bdev_handle);
 		goto error;
 	}
+	bdev = (*bdev_handle)->bdev;
 
 	if (flush)
-		sync_blockdev(*bdev);
-	ret = set_blocksize(*bdev, BTRFS_BDEV_BLOCKSIZE);
+		sync_blockdev(bdev);
+	ret = set_blocksize(bdev, BTRFS_BDEV_BLOCKSIZE);
 	if (ret) {
-		blkdev_put(*bdev, holder);
+		bdev_release(*bdev_handle);
 		goto error;
 	}
-	invalidate_bdev(*bdev);
-	*disk_super = btrfs_read_dev_super(*bdev);
+	invalidate_bdev(bdev);
+	*disk_super = btrfs_read_dev_super(bdev);
 	if (IS_ERR(*disk_super)) {
 		ret = PTR_ERR(*disk_super);
-		blkdev_put(*bdev, holder);
+		bdev_release(*bdev_handle);
 		goto error;
 	}
 
 	return 0;
 
 error:
-	*bdev = NULL;
+	*bdev_handle = NULL;
 	return ret;
 }
 
@@ -563,7 +565,7 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 			struct btrfs_device *device, blk_mode_t flags,
 			void *holder)
 {
-	struct block_device *bdev;
+	struct bdev_handle *bdev_handle;
 	struct btrfs_super_block *disk_super;
 	u64 devid;
 	int ret;
@@ -574,7 +576,7 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 		return -EINVAL;
 
 	ret = btrfs_get_bdev_and_sb(device->name->str, flags, holder, 1,
-				    &bdev, &disk_super);
+				    &bdev_handle, &disk_super);
 	if (ret)
 		return ret;
 
@@ -598,21 +600,21 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 		clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 		fs_devices->seeding = true;
 	} else {
-		if (bdev_read_only(bdev))
+		if (bdev_read_only(bdev_handle->bdev))
 			clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 		else
 			set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 	}
 
-	if (!bdev_nonrot(bdev))
+	if (!bdev_nonrot(bdev_handle->bdev))
 		fs_devices->rotating = true;
 
-	if (bdev_max_discard_sectors(bdev))
+	if (bdev_max_discard_sectors(bdev_handle->bdev))
 		fs_devices->discardable = true;
 
-	device->bdev = bdev;
+	device->bdev_handle = bdev_handle;
+	device->bdev = bdev_handle->bdev;
 	clear_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
-	device->holder = holder;
 
 	fs_devices->open_devices++;
 	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state) &&
@@ -626,7 +628,7 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 
 error_free_page:
 	btrfs_release_disk_super(disk_super);
-	blkdev_put(bdev, holder);
+	bdev_release(bdev_handle);
 
 	return -EINVAL;
 }
@@ -970,9 +972,10 @@ static void __btrfs_free_extra_devids(struct btrfs_fs_devices *fs_devices,
 		if (device->devid == BTRFS_DEV_REPLACE_DEVID)
 			continue;
 
-		if (device->bdev) {
-			blkdev_put(device->bdev, device->holder);
+		if (device->bdev_handle) {
+			bdev_release(device->bdev_handle);
 			device->bdev = NULL;
+			device->bdev_handle = NULL;
 			fs_devices->open_devices--;
 		}
 		if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state)) {
@@ -1017,7 +1020,7 @@ static void btrfs_close_bdev(struct btrfs_device *device)
 		invalidate_bdev(device->bdev);
 	}
 
-	blkdev_put(device->bdev, device->holder);
+	bdev_release(device->bdev_handle);
 }
 
 static void btrfs_close_one_device(struct btrfs_device *device)
@@ -1267,7 +1270,7 @@ struct btrfs_device *btrfs_scan_one_device(const char *path,
 	struct btrfs_super_block *disk_super;
 	bool new_device_added = false;
 	struct btrfs_device *device = NULL;
-	struct block_device *bdev;
+	struct bdev_handle *bdev_handle;
 	u64 bytenr, bytenr_orig;
 	int ret;
 
@@ -1290,18 +1293,19 @@ struct btrfs_device *btrfs_scan_one_device(const char *path,
 	 * values temporarily, as the device paths of the fsid are the only
 	 * required information for assembling the volume.
 	 */
-	bdev = blkdev_get_by_path(path, BLK_OPEN_READ, NULL, NULL);
-	if (IS_ERR(bdev))
-		return ERR_CAST(bdev);
+	bdev_handle = bdev_open_by_path(path, BLK_OPEN_READ, NULL, NULL);
+	if (IS_ERR(bdev_handle))
+		return ERR_CAST(bdev_handle);
 
 	bytenr_orig = btrfs_sb_offset(0);
-	ret = btrfs_sb_log_location_bdev(bdev, 0, READ, &bytenr);
+	ret = btrfs_sb_log_location_bdev(bdev_handle->bdev, 0, READ, &bytenr);
 	if (ret) {
 		device = ERR_PTR(ret);
 		goto error_bdev_put;
 	}
 
-	disk_super = btrfs_read_disk_super(bdev, bytenr, bytenr_orig);
+	disk_super = btrfs_read_disk_super(bdev_handle->bdev, bytenr,
+					   bytenr_orig);
 	if (IS_ERR(disk_super)) {
 		device = ERR_CAST(disk_super);
 		goto error_bdev_put;
@@ -1331,7 +1335,7 @@ struct btrfs_device *btrfs_scan_one_device(const char *path,
 	btrfs_release_disk_super(disk_super);
 
 error_bdev_put:
-	blkdev_put(bdev, NULL);
+	bdev_release(bdev_handle);
 
 	return device;
 }
@@ -2008,7 +2012,7 @@ void btrfs_scratch_superblocks(struct btrfs_fs_info *fs_info,
 
 int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 		    struct btrfs_dev_lookup_args *args,
-		    struct block_device **bdev, void **holder)
+		    struct bdev_handle **bdev_handle)
 {
 	struct btrfs_trans_handle *trans;
 	struct btrfs_device *device;
@@ -2117,7 +2121,7 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 
 	btrfs_assign_next_active_device(device, NULL);
 
-	if (device->bdev) {
+	if (device->bdev_handle) {
 		cur_devices->open_devices--;
 		/* remove sysfs entry */
 		btrfs_sysfs_remove_device(device);
@@ -2133,9 +2137,9 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 	 * free the device.
 	 *
 	 * We cannot call btrfs_close_bdev() here because we're holding the sb
-	 * write lock, and blkdev_put() will pull in the ->open_mutex on the
-	 * block device and it's dependencies.  Instead just flush the device
-	 * and let the caller do the final blkdev_put.
+	 * write lock, and bdev_release() will pull in the ->open_mutex on
+	 * the block device and it's dependencies.  Instead just flush the
+	 * device and let the caller do the final bdev_release.
 	 */
 	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state)) {
 		btrfs_scratch_superblocks(fs_info, device->bdev,
@@ -2146,8 +2150,7 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 		}
 	}
 
-	*bdev = device->bdev;
-	*holder = device->holder;
+	*bdev_handle = device->bdev_handle;
 	synchronize_rcu();
 	btrfs_free_device(device);
 
@@ -2285,7 +2288,7 @@ int btrfs_get_dev_args_from_path(struct btrfs_fs_info *fs_info,
 				 const char *path)
 {
 	struct btrfs_super_block *disk_super;
-	struct block_device *bdev;
+	struct bdev_handle *bdev_handle;
 	int ret;
 
 	if (!path || !path[0])
@@ -2303,7 +2306,7 @@ int btrfs_get_dev_args_from_path(struct btrfs_fs_info *fs_info,
 	}
 
 	ret = btrfs_get_bdev_and_sb(path, BLK_OPEN_READ, NULL, 0,
-				    &bdev, &disk_super);
+				    &bdev_handle, &disk_super);
 	if (ret) {
 		btrfs_put_dev_args_from_path(args);
 		return ret;
@@ -2322,7 +2325,7 @@ int btrfs_get_dev_args_from_path(struct btrfs_fs_info *fs_info,
 	else
 		memcpy(args->fsid, disk_super->fsid, BTRFS_FSID_SIZE);
 	btrfs_release_disk_super(disk_super);
-	blkdev_put(bdev, NULL);
+	bdev_release(bdev_handle);
 	return 0;
 }
 
@@ -2543,7 +2546,7 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	struct btrfs_root *root = fs_info->dev_root;
 	struct btrfs_trans_handle *trans;
 	struct btrfs_device *device;
-	struct block_device *bdev;
+	struct bdev_handle *bdev_handle;
 	struct super_block *sb = fs_info->sb;
 	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
 	struct btrfs_fs_devices *seed_devices = NULL;
@@ -2556,12 +2559,12 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	if (sb_rdonly(sb) && !fs_devices->seeding)
 		return -EROFS;
 
-	bdev = blkdev_get_by_path(device_path, BLK_OPEN_WRITE,
-				  fs_info->bdev_holder, NULL);
-	if (IS_ERR(bdev))
-		return PTR_ERR(bdev);
+	bdev_handle = bdev_open_by_path(device_path, BLK_OPEN_WRITE,
+					fs_info->bdev_holder, NULL);
+	if (IS_ERR(bdev_handle))
+		return PTR_ERR(bdev_handle);
 
-	if (!btrfs_check_device_zone_type(fs_info, bdev)) {
+	if (!btrfs_check_device_zone_type(fs_info, bdev_handle->bdev)) {
 		ret = -EINVAL;
 		goto error;
 	}
@@ -2573,11 +2576,11 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 		locked = true;
 	}
 
-	sync_blockdev(bdev);
+	sync_blockdev(bdev_handle->bdev);
 
 	rcu_read_lock();
 	list_for_each_entry_rcu(device, &fs_devices->devices, dev_list) {
-		if (device->bdev == bdev) {
+		if (device->bdev == bdev_handle->bdev) {
 			ret = -EEXIST;
 			rcu_read_unlock();
 			goto error;
@@ -2593,7 +2596,8 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	}
 
 	device->fs_info = fs_info;
-	device->bdev = bdev;
+	device->bdev_handle = bdev_handle;
+	device->bdev = bdev_handle->bdev;
 	ret = lookup_bdev(device_path, &device->devt);
 	if (ret)
 		goto error_free_device;
@@ -2614,12 +2618,11 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 	device->io_align = fs_info->sectorsize;
 	device->sector_size = fs_info->sectorsize;
 	device->total_bytes =
-		round_down(bdev_nr_bytes(bdev), fs_info->sectorsize);
+		round_down(bdev_nr_bytes(device->bdev), fs_info->sectorsize);
 	device->disk_total_bytes = device->total_bytes;
 	device->commit_total_bytes = device->total_bytes;
 	set_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
 	clear_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state);
-	device->holder = fs_info->bdev_holder;
 	device->dev_stats_valid = 1;
 	set_blocksize(device->bdev, BTRFS_BDEV_BLOCKSIZE);
 
@@ -2655,7 +2658,7 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 
 	atomic64_add(device->total_bytes, &fs_info->free_chunk_space);
 
-	if (!bdev_nonrot(bdev))
+	if (!bdev_nonrot(device->bdev))
 		fs_devices->rotating = true;
 
 	orig_super_total_bytes = btrfs_super_total_bytes(fs_info->super_copy);
@@ -2777,7 +2780,7 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 error_free_device:
 	btrfs_free_device(device);
 error:
-	blkdev_put(bdev, fs_info->bdev_holder);
+	bdev_release(bdev_handle);
 	if (locked) {
 		mutex_unlock(&uuid_mutex);
 		up_write(&sb->s_umount);
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index b5d0586c82af..cb42376ad7b5 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -90,13 +90,11 @@ struct btrfs_device {
 
 	u64 generation;
 
+	struct bdev_handle *bdev_handle;
 	struct block_device *bdev;
 
 	struct btrfs_zoned_device_info *zone_info;
 
-	/* block device holder for blkdev_get/put */
-	void *holder;
-
 	/*
 	 * Device's major-minor number. Must be set even if the device is not
 	 * opened (bdev == NULL), unless the device is missing.
@@ -649,7 +647,7 @@ struct btrfs_device *btrfs_alloc_device(struct btrfs_fs_info *fs_info,
 void btrfs_put_dev_args_from_path(struct btrfs_dev_lookup_args *args);
 int btrfs_rm_device(struct btrfs_fs_info *fs_info,
 		    struct btrfs_dev_lookup_args *args,
-		    struct block_device **bdev, void **holder);
+		    struct bdev_handle **bdev_handle);
 void __exit btrfs_cleanup_fs_uuids(void);
 int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len);
 int btrfs_grow_device(struct btrfs_trans_handle *trans,
-- 
2.35.3

Re: [PATCH v4 0/29] block: Make blkdev_get_by_*() return handle

[Jens Axboe]

On Wed, Sep 27, 2023 at 3:34?AM Jan Kara <jack@suse.cz> wrote:
>
> Hello,
>
> this is a v3 of the patch series which implements the idea of blkdev_get_by_*()

v4?

> calls returning bdev_handle which is then passed to blkdev_put() [1]. This
> makes the get and put calls for bdevs more obviously matching and allows us to
> propagate context from get to put without having to modify all the users
> (again!). In particular I need to propagate used open flags to blkdev_put() to
> be able count writeable opens and add support for blocking writes to mounted
> block devices. I'll send that series separately.
>
> The series is based on Btrfs tree's for-next branch [2] as of today as the
> series depends on Christoph's changes to btrfs device handling.  Patches have
> passed some reasonable testing - I've tested block changes, md, dm, bcache,
> xfs, btrfs, ext4, swap. More testing or review is always welcome. Thanks! I've
> pushed out the full branch to:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs.git bdev_handle
>
> to ease review / testing. Christian, can you pull the patches to your tree
> to get some exposure in linux-next as well? Thanks!

For the block bits:

Acked-by: Jens Axboe <axboe@kernel.dk>

-- 
Jens Axboe

Re: [PATCH v4 0/29] block: Make blkdev_get_by_*() return handle

[Christian Brauner]

On Wed, 27 Sep 2023 11:34:07 +0200, Jan Kara wrote:
> Create struct bdev_handle that contains all parameters that need to be
> passed to blkdev_put() and provide bdev_open_* functions that return
> this structure instead of plain bdev pointer. This will eventually allow
> us to pass one more argument to blkdev_put() (renamed to bdev_release())
> without too much hassle.
> 
> 
> [...]

> to ease review / testing. Christian, can you pull the patches to your tree
> to get some exposure in linux-next as well? Thanks!

Yep. So I did it slighly differently. I pulled in the btrfs prereqs and
then applied your series on top of it so we get all the Link: tags right.
I'm running tests right now. Please double-check.

---

Applied to the vfs.super branch of the vfs/vfs.git tree.
Patches in the vfs.super branch should appear in linux-next soon.

Please report any outstanding bugs that were missed during review in a
new review to the original patch series allowing us to drop it.

It's encouraged to provide Acked-bys and Reviewed-bys even though the
patch has now been applied. If possible patch trailers will be updated.

Note that commit hashes shown below are subject to change due to rebase,
trailer updates or similar. If in doubt, please check the listed branch.

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
branch: vfs.super

[01/29] block: Provide bdev_open_* functions
       https://git.kernel.org/vfs/vfs/c/b7c828aa0b3c
[02/29] block: Use bdev_open_by_dev() in blkdev_open()
        https://git.kernel.org/vfs/vfs/c/d4e36f27b45a
[03/29] block: Use bdev_open_by_dev() in disk_scan_partitions() and blkdev_bszset()
        https://git.kernel.org/vfs/vfs/c/5f9bd6764c7a
[04/29] drdb: Convert to use bdev_open_by_path()
        https://git.kernel.org/vfs/vfs/c/0220ca8e443d
[05/29] pktcdvd: Convert to bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/7af10b889789
[06/29] rnbd-srv: Convert to use bdev_open_by_path()
        https://git.kernel.org/vfs/vfs/c/3d27892a4be7
[07/29] xen/blkback: Convert to bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/26afb0ed10b3
[08/29] zram: Convert to use bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/efc8e3f4c6dc
[09/29] bcache: Convert to bdev_open_by_path()
        https://git.kernel.org/vfs/vfs/c/dc893f51d24a
[10/29] dm: Convert to bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/80c2267c6d07
[11/29] md: Convert to bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/15db36126ca6
[12/29] mtd: block2mtd: Convert to bdev_open_by_dev/path()
        https://git.kernel.org/vfs/vfs/c/4c27234bf3ce
[13/29] nvmet: Convert to bdev_open_by_path()
        https://git.kernel.org/vfs/vfs/c/70cffddcc300
[14/29] s390/dasd: Convert to bdev_open_by_path()
        https://git.kernel.org/vfs/vfs/c/5581d03457f8
[15/29] scsi: target: Convert to bdev_open_by_path()
        https://git.kernel.org/vfs/vfs/c/43de7d844d47
[16/29] PM: hibernate: Convert to bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/105ea4a2fd18
[17/29] PM: hibernate: Drop unused snapshot_test argument
        https://git.kernel.org/vfs/vfs/c/b589a66e3688
[18/29] mm/swap: Convert to use bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/615af8e29233
[19/29] fs: Convert to bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/5173192bcfe6
[20/29] btrfs: Convert to bdev_open_by_path()
        https://git.kernel.org/vfs/vfs/c/8cf64782764f
[21/29] erofs: Convert to use bdev_open_by_path()
        https://git.kernel.org/vfs/vfs/c/4d41880bf249
[22/29] ext4: Convert to bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/f7507612395e
[23/29] f2fs: Convert to bdev_open_by_dev/path()
        https://git.kernel.org/vfs/vfs/c/d9ff8e3b6498
[24/29] jfs: Convert to bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/459dc6376338
[25/29] nfs/blocklayout: Convert to use bdev_open_by_dev/path()
        https://git.kernel.org/vfs/vfs/c/5b1df9a40929
[26/29] ocfs2: Convert to use bdev_open_by_dev()
        https://git.kernel.org/vfs/vfs/c/b6b95acbd943
[27/29] reiserfs: Convert to bdev_open_by_dev/path()
        https://git.kernel.org/vfs/vfs/c/7e3615ff6119
[28/29] xfs: Convert to bdev_open_by_path()
        https://git.kernel.org/vfs/vfs/c/176ccb99e207
[29/29] block: Remove blkdev_get_by_*() functions
        https://git.kernel.org/vfs/vfs/c/953863a5a2ff

Re: [PATCH v4 0/29] block: Make blkdev_get_by_*() return handle

[Jan Kara]

On Wed 27-09-23 18:21:19, Christian Brauner wrote:
> On Wed, 27 Sep 2023 11:34:07 +0200, Jan Kara wrote:
> > Create struct bdev_handle that contains all parameters that need to be
> > passed to blkdev_put() and provide bdev_open_* functions that return
> > this structure instead of plain bdev pointer. This will eventually allow
> > us to pass one more argument to blkdev_put() (renamed to bdev_release())
> > without too much hassle.
> > 
> > 
> > [...]
> 
> > to ease review / testing. Christian, can you pull the patches to your tree
> > to get some exposure in linux-next as well? Thanks!
> 
> Yep. So I did it slighly differently. I pulled in the btrfs prereqs and
> then applied your series on top of it so we get all the Link: tags right.
> I'm running tests right now. Please double-check.

Thanks for picking patches up! I've checked the branch and it looks good to
me. 

								Honza

> 
> ---
> 
> Applied to the vfs.super branch of the vfs/vfs.git tree.
> Patches in the vfs.super branch should appear in linux-next soon.
> 
> Please report any outstanding bugs that were missed during review in a
> new review to the original patch series allowing us to drop it.
> 
> It's encouraged to provide Acked-bys and Reviewed-bys even though the
> patch has now been applied. If possible patch trailers will be updated.
> 
> Note that commit hashes shown below are subject to change due to rebase,
> trailer updates or similar. If in doubt, please check the listed branch.
> 
> tree:   https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
> branch: vfs.super
> 
> [01/29] block: Provide bdev_open_* functions
>        https://git.kernel.org/vfs/vfs/c/b7c828aa0b3c
> [02/29] block: Use bdev_open_by_dev() in blkdev_open()
>         https://git.kernel.org/vfs/vfs/c/d4e36f27b45a
> [03/29] block: Use bdev_open_by_dev() in disk_scan_partitions() and blkdev_bszset()
>         https://git.kernel.org/vfs/vfs/c/5f9bd6764c7a
> [04/29] drdb: Convert to use bdev_open_by_path()
>         https://git.kernel.org/vfs/vfs/c/0220ca8e443d
> [05/29] pktcdvd: Convert to bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/7af10b889789
> [06/29] rnbd-srv: Convert to use bdev_open_by_path()
>         https://git.kernel.org/vfs/vfs/c/3d27892a4be7
> [07/29] xen/blkback: Convert to bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/26afb0ed10b3
> [08/29] zram: Convert to use bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/efc8e3f4c6dc
> [09/29] bcache: Convert to bdev_open_by_path()
>         https://git.kernel.org/vfs/vfs/c/dc893f51d24a
> [10/29] dm: Convert to bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/80c2267c6d07
> [11/29] md: Convert to bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/15db36126ca6
> [12/29] mtd: block2mtd: Convert to bdev_open_by_dev/path()
>         https://git.kernel.org/vfs/vfs/c/4c27234bf3ce
> [13/29] nvmet: Convert to bdev_open_by_path()
>         https://git.kernel.org/vfs/vfs/c/70cffddcc300
> [14/29] s390/dasd: Convert to bdev_open_by_path()
>         https://git.kernel.org/vfs/vfs/c/5581d03457f8
> [15/29] scsi: target: Convert to bdev_open_by_path()
>         https://git.kernel.org/vfs/vfs/c/43de7d844d47
> [16/29] PM: hibernate: Convert to bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/105ea4a2fd18
> [17/29] PM: hibernate: Drop unused snapshot_test argument
>         https://git.kernel.org/vfs/vfs/c/b589a66e3688
> [18/29] mm/swap: Convert to use bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/615af8e29233
> [19/29] fs: Convert to bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/5173192bcfe6
> [20/29] btrfs: Convert to bdev_open_by_path()
>         https://git.kernel.org/vfs/vfs/c/8cf64782764f
> [21/29] erofs: Convert to use bdev_open_by_path()
>         https://git.kernel.org/vfs/vfs/c/4d41880bf249
> [22/29] ext4: Convert to bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/f7507612395e
> [23/29] f2fs: Convert to bdev_open_by_dev/path()
>         https://git.kernel.org/vfs/vfs/c/d9ff8e3b6498
> [24/29] jfs: Convert to bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/459dc6376338
> [25/29] nfs/blocklayout: Convert to use bdev_open_by_dev/path()
>         https://git.kernel.org/vfs/vfs/c/5b1df9a40929
> [26/29] ocfs2: Convert to use bdev_open_by_dev()
>         https://git.kernel.org/vfs/vfs/c/b6b95acbd943
> [27/29] reiserfs: Convert to bdev_open_by_dev/path()
>         https://git.kernel.org/vfs/vfs/c/7e3615ff6119
> [28/29] xfs: Convert to bdev_open_by_path()
>         https://git.kernel.org/vfs/vfs/c/176ccb99e207
> [29/29] block: Remove blkdev_get_by_*() functions
>         https://git.kernel.org/vfs/vfs/c/953863a5a2ff
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR