Re: [PATCH v17 15/21] xfs: add fs-verity support

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

On Mon, Oct 05, 2026 at 07:18:35PM +0200, Andrey Albershteyn wrote:

<snip>

> > +/*
> > + * Try to remove all the fsverity metadata after a failed enablement.
> > + */
> > +static int
> > +xfs_fsverity_delete_metadata(
> > +	struct xfs_inode	*ip)
> > +{
> > +	struct xfs_trans	*tp;
> > +	struct xfs_mount	*mp = ip->i_mount;
> > +	int			error;
> > +
> > +	xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
> > +	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp);
> > +	if (error) {
> > +		xfs_iunlock(ip, XFS_MMAPLOCK_EXCL);
> > +		return error;
> > +	}
> > +
> > +	truncate_inode_pages(VFS_I(ip)->i_mapping, XFS_ISIZE(ip));
> 
> Oh I haven't paid enough attention here, it has to be moved before
> xfs_ilock(ip, XFS_MMAPLOCK_EXCL) not just before the lock, to exclude
> ABBA with writeback/reclaim taking folio lock and waiting for trans
> space.

Hrm?  You have to hold IOLOCK and MMAPLOCK, and *not* hold ILOCK, to
truncate the pagecache.  IOWs,

	xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
	truncate_inode_pages(VFS_I(ip)->i_mapping, XFS_ISIZE(ip));

	error = xfs_trans_alloc(...);
	if (error)
		...

	xfs_ilock(ip, XFS_ILOCK_EXCL);
	xfs_trans_ijoin(tp, ip, 0);

--D

> I will send v18 if more feedback will arrive, otherwise, I can
> resend just this patch (or could be fixed during merge?).
> 
> -- 
> - Andrey
>