Hi,
On Tue, Mar 21, 2023 at 9:34 AM Andreas Gruenbacher <agrue...@redhat.com> wrote:
>
> On Tue, Mar 21, 2023 at 2:17 PM Alexander Aring <aahri...@redhat.com> wrote:
> > This patch sets another flags variable to zero which is a leftover of
> > commit 0834ed4b ("dlm_controld: initialize waiter->flags").
> > ---
> > dlm_controld/plock.c | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/dlm_controld/plock.c b/dlm_controld/plock.c
> > index ad9b0f27..7f632888 100644
> > --- a/dlm_controld/plock.c
> > +++ b/dlm_controld/plock.c
> > @@ -1956,6 +1956,7 @@ void receive_plocks_data(struct lockspace *ls, struct
> > dlm_header *hd, int len)
> > po->pid = le32_to_cpu(pp->pid);
> > po->nodeid = le32_to_cpu(pp->nodeid);
> > po->ex = pp->ex;
> > + po->flags = 0;
> > list_add_tail(&po->list, &r->locks);
> > } else {
> > w = malloc(sizeof(struct lock_waiter));
>
> ^ And I think w->flags also isn't being initialized.
I added it by changing your patch, but I only checked for
lock_waiter... After pushing those patches I found a missing piece in
posix_lock.
Meanwhile I ran cppcheck on it and it found another issue...
- Alex
