Re: [PATCH 03/10] generic/710: repurpose this for exchangerange vs. quota testing

["Darrick J. Wong" <djwong@kernel.org>]

On Wed, Jun 19, 2024 at 10:23:18AM -0700, Darrick J. Wong wrote:
> On Tue, Jun 18, 2024 at 11:08:37PM -0700, Christoph Hellwig wrote:
> > On Mon, Jun 17, 2024 at 05:47:32PM -0700, Darrick J. Wong wrote:
> > > From: Darrick J. Wong <djwong@kernel.org>
> > > 
> > > The exchange-range implementation is now completely separate from the
> > > old swapext ioctl.  We're deprecating the old swapext ioctl, so let's
> > > move this test to use exchangerange.
> > 
> > Do we really want to lost the swapext test coverage?  Even if it is
> > deprecated, it will be with us for a long time.  My vote for copy and
> > pasting this into a new test for exchrange.
> 
> Yeah, you're right that we should retain this test for the old swapext
> ioctl.  I'll fork the test into two -- one for swapext, another for
> exchangerange.

...except that the swapext ioctl doesn't support swapping forks if quota
is enabled and any of the user/group/project ids are different:


	/* User/group/project quota ids must match if quotas are enforced. */
	if (XFS_IS_QUOTA_ON(ip->i_mount) &&
	    (!uid_eq(VFS_I(ip)->i_uid, VFS_I(tip)->i_uid) ||
	     !gid_eq(VFS_I(ip)->i_gid, VFS_I(tip)->i_gid) ||
	     ip->i_projid != tip->i_projid))
		return -EINVAL;

I'll amend the commit message:

"There's no point in maintaining this test for the legacy swapext code
because it returns EINVAL if any quota is enabled and the two files have
different user/group/project ids.  Originally I had forward ported the
old swapext ioctl to use commitrange as its backend, but that will be
dropped in favor of porting xfs_fsr to use commitrange directly."

--D