On Thu, Jan 17, 2019 at 06:22:21AM -0800, Christoph Hellwig wrote:
> On Mon, Dec 31, 2018 at 06:19:38PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> >
> > Move the kernel-specific parts of xfs_ialloc into a separate function in
> > preparation for hoisting xfs_ialloc to libxfs.
> >
> > Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> > ---
> > fs/xfs/xfs_inode.c | 45 ++++++++++++++++++++++++++++++++++++---------
> > 1 file changed, 36 insertions(+), 9 deletions(-)
> >
> >
> > diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
> > index 403eb8fa2f1f..b635d43caeed 100644
> > --- a/fs/xfs/xfs_inode.c
> > +++ b/fs/xfs/xfs_inode.c
> > @@ -621,6 +621,32 @@ xfs_lookup(
> > return error;
> > }
> >
> > +/* Return an exclusive ILOCK'd in-core inode. */
> > +static int
> > +xfs_ialloc_iget(
> > + struct xfs_mount *mp,
> > + struct xfs_trans *tp,
> > + xfs_ino_t ino,
> > + struct xfs_inode **ipp)
> > +{
> > + return xfs_iget(mp, tp, ino, XFS_IGET_CREATE, XFS_ILOCK_EXCL, ipp);
> > +}
>
> This is just too ugly and pointless. Can you explain (or even better
> show code) why we really need this for xfsprogs?
To avoid having a useless lock_flags argument to xfs_iget in userspace?
It wasn't so long ago that Eric removed it, but I don't mind adding it
back to avoid this kind of ugliness.
> > + xfs_ialloc_platform_init(tp, args, ip);
>
> And that platform naming also is rather horrible.
Hm. Looking at it now, both platform_init() functions basically just
set ctime/mtime/atime/crtime, which means that I really just need to add
a XFS_ICHGTIME_CREATE flag to xfs_trans_ichgtime and replace all the
platform_init stuff with that.
The mkfs protofile platform_init also has some code to set other
attributes from a struct fsxattr, but I think I can just set them once
the inode allocation function returns to the protofile code.
--D
