Skip to content

btrfs-progs: small cleanups related to subvolume creation

Qu Wenruo linux-btrfs 2 versions – New

Get series with b4
b4 am [email protected]

Versions newest first · linked by author and title

  1. v2 Latest3 patches 0 repliesNew
  2. v12 patches 0 repliesSuperseded

Changes in v2 from the cover letter

- A new patch to cleanup btrfs_mksubvol()
We have a kernel function with the same name, but those two are doing completely different works.

Patches in v2 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/3btrfs-progs: remove fs_info parameter from btrfs_create_tree()kernel-shared/disk-io.c kernel-shared/disk-io.h kernel-shared/free-space-tree.c +1 more+7 -80 replies
2/3btrfs-progs: introduce btrfs_make_subvolume()Makefile check/main.c common/root-tree-utils.c +5 more+141 -1510 replies
3/3btrfs-progs: use btrfs_link_subvolume() to replace btrfs_mksubvol()check/main.c common/root-tree-utils.c common/root-tree-utils.h +3 more+184 -1470 replies

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v2 yet.

Read the v2 thread

Cover letter [PATCH v2 0/3] btrfs-progs: small cleanups related to subvolume creation

[CHANGELOG]
v2:
- A new patch to cleanup btrfs_mksubvol()
We have a kernel function with the same name, but those two are doing completely different works.

The progs version is more like linking a subvolume to a destination.

The new patch removes btrfs_mksubvol(), and move the convert specific code to convert code basis.

Thanks to Mark's new effort to introduce subvolume creation ability, the long existing duplicated subvolume creation problem is exposed again.

The first patch to do a small cleanup for btrfs_create_tree() so that the parameter list matches the kernel one.

The second one is the main dish to fully merge the different functions to create a subvolume.

We have btrfs_create_tree() to properly create an empty tree, and btrfs_make_root_dir() to create the initial root dir.

So use them to create btrfs_make_subvolume():

- Calls btrfs_create_tree() to properly create an empty tree Unlike btrfs_copy_root() used in create_subvol(), which can be unsafe if the source subvolume is not empty.

- Calls btrfs_read_fs_root() to setup the cache and tracking Inside create_data_reloc_tree() we directly added the root to fs_root_tree, which is only safe for data reloc tree.

As we didn't properly set the correct tracking flags.

- Calls btrfs_make_root_dir() to setup the root directory

And finally cleanup a confusing function, btrfs_mksubvol(), which is not really creating a subvolume, but linking a subvolume to a parent inode.

Instead of the confusing name, create a new helper, btrfs_link_subvolume() to do the linkage, and move the convert specific retry behavior into convert code.

Qu Wenruo (3):
btrfs-progs: remove fs_info parameter from btrfs_create_tree()
btrfs-progs: introduce btrfs_make_subvolume()
btrfs-progs: use btrfs_link_subvolume() to replace btrfs_mksubvol()

 Makefile                        |   1 +
 check/main.c                    |   9 +-
 common/root-tree-utils.c        | 215 ++++++++++++++++++++++++++++++++
 common/root-tree-utils.h        |  30 +++++
 convert/main.c                  | 107 +++++++++-------
 kernel-shared/ctree.h           |   8 +-
 kernel-shared/disk-io.c         |   4 +-
 kernel-shared/disk-io.h         |   1 -
 kernel-shared/free-space-tree.c |   2 +-
 kernel-shared/inode.c           | 140 +--------------------
 mkfs/common.c                   |  39 ------
 mkfs/common.h                   |   2 -
 mkfs/main.c                     |  78 +-----------
 13 files changed, 331 insertions(+), 305 deletions(-)
 create mode 100644 common/root-tree-utils.c
 create mode 100644 common/root-tree-utils.h

--
2.45.2