This series enables a new feature to always honour the extent size hint.
It is required for atomic writes support.
It is possible to enable this feature at mkfs time or by changing the
attribute of regular file or directory. For extszhint, it is inherited
by descenent files/directories.
Associated kernel atomic writes patches to follow. We are sending this
support now so that people can try out the kernel support. More testing
is required.
Based on origin/for-next_2023-09-25
Darrick J. Wong (7):
xfs: create a new inode flag to require extsize alignment of file data
space
xfs: allow files to require data mappings to be aligned to extszhint
xfs_db: expose force_align feature and flags
xfs_io: implement lsattr and chattr support for forcealign
xfs_repair: check the force-align flag
mkfs: add an extsize= option that allows units
mkfs: enable the new force-align feature
db/inode.c | 3 +
db/sb.c | 2 +
include/linux.h | 5 +
include/xfs_inode.h | 5 +
include/xfs_mount.h | 2 +
io/attr.c | 5 +-
libxfs/util.c | 2 +
libxfs/xfs_bmap.c | 26 ++++-
libxfs/xfs_format.h | 9 +-
libxfs/xfs_inode_buf.c | 40 ++++++++
libxfs/xfs_inode_buf.h | 3 +
libxfs/xfs_sb.c | 3 +
man/man2/ioctl_xfs_fsgetxattr.2 | 6 ++
man/man8/mkfs.xfs.8.in | 29 ++++++
mkfs/xfs_mkfs.c | 175 ++++++++++++++++++++++++++++++--
repair/dinode.c | 66 ++++++++++++
16 files changed, 367 insertions(+), 14 deletions(-)
--
2.34.1
