mbox series

[-next,RFC,00/14] block: don't access bd_inode directly from other modules

Message ID 20231205123728.1866699-1-yukuai1@huaweicloud.com
Headers show
Series block: don't access bd_inode directly from other modules | expand

Message

Yu Kuai Dec. 5, 2023, 12:37 p.m. UTC 
From: Yu Kuai <yukuai3@huawei.com>

Patch 1 add some bdev apis, then follow up patches will use these apis
to avoid access bd_inode directly, and hopefully the field bd_inode can
be removed eventually(after figure out a way for fs/buffer.c).

Yu Kuai (14):
  block: add some bdev apis
  xen/blkback: use bdev api in xen_update_blkif_status()
  bcache: use bdev api in read_super()
  mtd: block2mtd: use bdev apis
  s390/dasd: use bdev api in dasd_format()
  scsicam: use bdev api in scsi_bios_ptable()
  bcachefs: remove dead function bdev_sectors()
  btrfs: use bdev apis
  cramfs: use bdev apis in cramfs_blkdev_read()
  erofs: use bdev api
  ext4: use bdev apis
  jbd2: use bdev apis
  gfs2: use bdev api
  nilfs2: use bdev api in nilfs_attach_log_writer()

 block/bdev.c                       | 116 +++++++++++++++++++++++++++++
 block/bio.c                        |   1 +
 block/blk.h                        |   2 -
 drivers/block/xen-blkback/xenbus.c |   3 +-
 drivers/md/bcache/super.c          |  11 ++-
 drivers/mtd/devices/block2mtd.c    |  80 +++++++++-----------
 drivers/s390/block/dasd_ioctl.c    |   5 +-
 drivers/scsi/scsicam.c             |   3 +-
 fs/bcachefs/util.h                 |   5 --
 fs/btrfs/disk-io.c                 |  68 ++++++++---------
 fs/btrfs/volumes.c                 |  17 ++---
 fs/btrfs/zoned.c                   |  12 ++-
 fs/cramfs/inode.c                  |  35 +++------
 fs/erofs/data.c                    |  17 +++--
 fs/erofs/internal.h                |   1 +
 fs/ext4/dir.c                      |   6 +-
 fs/ext4/ext4_jbd2.c                |   6 +-
 fs/ext4/super.c                    |  27 +------
 fs/gfs2/glock.c                    |   2 +-
 fs/gfs2/ops_fstype.c               |   2 +-
 fs/jbd2/journal.c                  |   3 +-
 fs/jbd2/recovery.c                 |   6 +-
 fs/nilfs2/segment.c                |   2 +-
 include/linux/blkdev.h             |  27 +++++++
 include/linux/buffer_head.h        |   5 +-
 25 files changed, 273 insertions(+), 189 deletions(-)

Comments

Christoph Hellwig Dec. 6, 2023, 5:54 a.m. UTC | #1 
On Tue, Dec 05, 2023 at 08:37:14PM +0800, Yu Kuai wrote:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> Patch 1 add some bdev apis, then follow up patches will use these apis
> to avoid access bd_inode directly, and hopefully the field bd_inode can
> be removed eventually(after figure out a way for fs/buffer.c).

What tree is this against?  It fails to apply to either Jens'
for-6.8/block or Linus tree in the very first patch.
Yu Kuai Dec. 6, 2023, 6:06 a.m. UTC | #2 
Hi,

在 2023/12/06 13:54, Christoph Hellwig 写道:
> On Tue, Dec 05, 2023 at 08:37:14PM +0800, Yu Kuai wrote:
>> From: Yu Kuai <yukuai3@huawei.com>
>>
>> Patch 1 add some bdev apis, then follow up patches will use these apis
>> to avoid access bd_inode directly, and hopefully the field bd_inode can
>> be removed eventually(after figure out a way for fs/buffer.c).
> 
> What tree is this against?  It fails to apply to either Jens'
> for-6.8/block or Linus tree in the very first patch.

It was against linux-next branch, for the tag next-20231201, because I'm
not sure yet if this patchset should be applied to Jans' tree. Please
let me know if I should swith wo Jens' tree for v2.

Thanks,
Kuai
> 
> .
>