On Mon, Dec 23, 2019 at 3:17 AM Jeff Layton <jlayton@xxxxxxxxxx> wrote:
>
> On Sun, 2019-12-22 at 20:45 +0200, Amir Goldstein wrote:
> > An ino is unsigned so export it as such in /proc/locks.
> >
> > Signed-off-by: Amir Goldstein <amir73il@xxxxxxxxx>
> > ---
> >
> > Hi Jeff,
> >
> > Ran into this while writing tests to verify i_ino == d_ino == st_ino on
> > overlayfs. In some configurations (xino=on) overlayfs sets MSB on i_ino,
> > so /proc/locks reports negative ino values.
> >
> > BTW, the requirement for (i_ino == d_ino) came from nfsd v3 readdirplus.
> >
> > Thanks,
> > Amir.
> >
> > fs/locks.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/fs/locks.c b/fs/locks.c
> > index 6970f55daf54..44b6da032842 100644
> > --- a/fs/locks.c
> > +++ b/fs/locks.c
> > @@ -2853,7 +2853,7 @@ static void lock_get_status(struct seq_file *f, struct file_lock *fl,
> > }
> > if (inode) {
> > /* userspace relies on this representation of dev_t */
> > - seq_printf(f, "%d %02x:%02x:%ld ", fl_pid,
> > + seq_printf(f, "%d %02x:%02x:%lu ", fl_pid,
> > MAJOR(inode->i_sb->s_dev),
> > MINOR(inode->i_sb->s_dev), inode->i_ino);
> > } else {
>
> My that is an old bug! I think that goes back to early v2.x days, if not
> v1.x. I'll queue it up, and maybe we can get this in for v5.6.
I suppose you meant for v5.5?
I'd be happy if we can also mark it for stable (sorry I did not).
Reason is that I have xfstests depending on it, which test overlay
fixes that are marked for stable.
Thanks,
Amir.
