On Wed, Nov 11, 2020 at 09:30:42PM +0100, Paul Bolle wrote: > tip-bot2 for Peter Zijlstra schreef op wo 11-11-2020 om 08:23 [+0000]: > > --- a/kernel/sched/core.c > > +++ b/kernel/sched/core.c > > [...] > > +static void do_balance_callbacks(struct rq *rq, struct callback_head *head) > > +{ > > + void (*func)(struct rq *rq); > > + struct callback_head *next; > > + > > + lockdep_assert_held(&rq->lock); > > + > > + while (head) { > > + func = (void (*)(struct rq *))head->func; > > + next = head->next; > > + head->next = NULL; > > + head = next; > > Naive question: is there some subtle C-issue that is evaded here by setting > head->next to NULL prior to copying over it? > > (I know this piece of code only got copied around in this patch and this is > therefor not something that this patch actually introduced.) It's like list_del_init(), it zeros the entry before unlinking it. queue_balance_callback() relies on this.
![]() |