dounmount: do not release the mount point's reference on the covered vnode

[avg]

Summary

As long as mnt_ref is not zero there can be a consumer that might try
to access mnt_vnodecovered. For this reason the covered vnode must not
be freed until mnt_ref goes to zero.
So, move the release of the covered vnode to vfs_mount_destroy.

[kib]

There is also one use of vfs_mount_destroy() in zfs, but I did not looked at it. You should know it much better.

sys/kern/vfs_mount.c

826

Look at this fragment. I believe that you need vref(vp) before the call to vfs_mount_destroy() at line 822.

[avg]

sys/kern/vfs_mount.c

826

Maybe setting mnt_vnodecovered to NULL would be a better solution?
The mp should not be visible to any external consumers yet.
The same would apply to mount_snapshot in ZFS code.

[kib]

sys/kern/vfs_mount.c

826

Fine with me.

[avg]

clear mnt_vnodecovered before calling vfs_mount_destroy in case of the mount failure,
that's safe to do because the new struct mount is not visible to the outside world