On Tue, Aug 27, 2024 at 04:26:08PM +0200, Oleg Nesterov wrote:
> On 08/27, Jiri Olsa wrote:
> >
> > On Tue, Aug 27, 2024 at 12:40:52PM +0200, Oleg Nesterov wrote:
> > > static bool
> > > uprobe_multi_link_filter(struct uprobe_consumer *con, struct mm_struct
> > > *mm)
> > > {
> > > struct bpf_uprobe *uprobe;
> > > + struct task_struct *task, *t;
> > > + bool ret = false;
> > >
> > > uprobe = container_of(con, struct bpf_uprobe, consumer);
> > > - return uprobe->link->task->mm == mm;
> > > + task = uprobe->link->task;
> > > +
> > > + rcu_read_lock();
> > > + for_each_thread(task, t) {
> > > + struct mm_struct *mm = READ_ONCE(t->mm);
> > > + if (mm) {
> > > + ret = t->mm == mm;
> > > + break;
> > > + }
> > > + }
> > > + rcu_read_unlock();
> >
> > that seems expensive if there's many threads
>
> many threads with ->mm == NULL? In the likely case for_each_thread()
> stops after the first t->mm check.
aah the mm will be the same.. right, nice
thanks,
jirka
