On Thu, Apr 9, 2020 at 1:13 PM Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:
>
> On Thu, Apr 09, 2020 at 11:18:58AM +0200, Miklos Szeredi wrote:
> > @@ -1249,31 +1277,48 @@ struct vfsmount *mnt_clone_internal(cons
> > static void *m_start(struct seq_file *m, loff_t *pos)
> > {
> > struct proc_mounts *p = m->private;
> > + struct mount *mnt;
> >
> > down_read(&namespace_sem);
> > - if (p->cached_event == p->ns->event) {
> > - void *v = p->cached_mount;
> > - if (*pos == p->cached_index)
> > - return v;
> > - if (*pos == p->cached_index + 1) {
> > - v = seq_list_next(v, &p->ns->list, &p->cached_index);
> > - return p->cached_mount = v;
> > + lock_ns_list(p->ns);
> > + if (!*pos) {
> > + list_move(&p->cursor.mnt_list, &p->ns->list);
> > + p->cursor_pos = 0;
> > + } else if (p->cursor_pos != *pos) {
> > + /* Non-zero seek. Could probably just return -ESPIPE... */
>
> Umm, that's not how calling seek() on a seqfile works. I don't think
> you can hit this case. Follow seq_lseek() as it calls traverse().
Ah, okay, f_pos is always a byte offset. That's wonderful, I can get
rid of that wart.
Thanks,
Miklos
