[PATCH v4 0/6] vfs: clean up i_version behavior and expose it via statx

From: Jeff Layton
Date: Wed Sep 07 2022 - 07:33:34 EST


v4: drop xfs patch
revise comment update patch with latest proposed semantics

This is a small revision to the patchset I sent a little over a week
ago [1]. Since then, this has also garnered a LWN article [2], so I
won't go into great detail on the basic premise and rationale.

The biggest change here is that I've dropped the xfs patch. Dave Chinner
stated that they'd need to add a new on-disk field instead of modifying
the behavior of the existing di_changecount field [3]. I'll leave that
to the xfs devs, but this does mean that xfs will have "buggy" behavior
until that's done.

I've also sent a revised manpage patchset separately to make sure that
the semantics are acceptable [4]. That hasn't gotten a lot of comments,
so I'm operating under the assumption that the semantics proposed there
are acceptable to most.

[1]: https://lore.kernel.org/linux-nfs/20220826214703.134870-1-jlayton@xxxxxxxxxx/
[2]: https://lwn.net/Articles/905931/
[3]: https://lore.kernel.org/linux-nfs/20220830000851.GV3600936@xxxxxxxxxxxxxxxxxxx/
[4]: https://lore.kernel.org/linux-nfs/20220907111606.18831-1-jlayton@xxxxxxxxxx/T/#u

Jeff Layton (6):
iversion: update comments with info about atime updates
ext4: fix i_version handling in ext4
ext4: unconditionally enable the i_version counter
vfs: report an inode version in statx for IS_I_VERSION inodes
nfs: report the inode version in statx if requested
ceph: fill in the change attribute in statx requests

fs/ceph/inode.c | 14 +++++++++-----
fs/ext4/inode.c | 15 +++++----------
fs/ext4/ioctl.c | 4 ++++
fs/ext4/move_extent.c | 6 ++++++
fs/ext4/super.c | 13 ++++---------
fs/ext4/xattr.c | 1 +
fs/nfs/inode.c | 7 +++++--
fs/stat.c | 7 +++++++
include/linux/iversion.h | 10 ++++++++--
include/linux/stat.h | 1 +
include/uapi/linux/stat.h | 3 ++-
samples/vfs/test-statx.c | 8 ++++++--
12 files changed, 58 insertions(+), 31 deletions(-)

--
2.37.3