Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1600555
| From | Peter Zijlstra <peterz@infradead.org> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | Re: perf: use-after-free in perf_release |
| Date | 2017-03-14 16:10 +0100 |
| Message-ID | <tkWfg-1T2-37@gated-at.bofh.it> (permalink) |
| References | (5 earlier) <tkUds-cW-25@gated-at.bofh.it> <tkVjb-1eJ-13@gated-at.bofh.it> <tkVjc-1eJ-41@gated-at.bofh.it> <tkVMd-1qV-13@gated-at.bofh.it> <tkWff-1T2-5@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
On Tue, Mar 14, 2017 at 04:02:41PM +0100, Peter Zijlstra wrote:
> On Tue, Mar 14, 2017 at 03:30:11PM +0100, Oleg Nesterov wrote:
>
> > But. perf_event_init_task() adds child_event to parent_event->child_list.
> >
> > If perf_event_release_kernel(parent_event) is called before copy_process()
> > does perf_event_free_task() which (in particular) removes it from child_list,
> > perf_event_release_kernel() can find this child_event and do get_ctx(ctx)
> > (under the list_for_each_entry(child, &event->child_list, child_list) loop).
>
> Right; the child_list is the only thing that is exposed. And yes, it
> looks like that can interleave just right.
>
> > Then it does put_ctx(ctx), but ctx->task can be already freed by
> > copy_process()->free_task() in this case.
>
>
> Task1 Task2
>
> fork()
> perf_event_init_task()
> /* ... */
> goto bad_fork_$foo;
> /* ... */
> perf_event_free_task()
> mutex_lock(ctx->lock)
> perf_free_event(B)
>
> perf_event_release_kernel(A)
> mutex_lock(A->child_mutex)
> list_for_each_entry(child, ...) {
> /* child == B */
> ctx = B->ctx;
> get_ctx(ctx);
> mutex_unlock(A->child_mutex);
>
> mutex_lock(A->child_mutex)
> list_del_init(B->child_list)
> mutex_unlock(A->child_mutex)
>
> /* ... */
>
> mutex_unlock(ctx->lock);
> put_ctx() /* >0 */
> free_task();
> mutex_lock(ctx->lock);
> mutex_lock(A->child_mutex);
> /* ... */
> mutex_unlock(A->child_mutex);
> mutex_unlock(ctx->lock)
> put_ctx() /* 0 */
> ctx->task && !TOMBSTONE
> put_task_struct() /* UAF */
>
>
> Something like that, right?
>
>
> Let me see if it makes sense to retain perf_event_free_task() at all;
> maybe we should always do perf_event_exit_task().
Do we want a WARN_ON_ONCE(atomic_read(&tsk->usage)); in free_task()?
Because in the above scenario we're freeing it with references on.
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
Re: perf: use-after-free in perf_release Peter Zijlstra <peterz@infradead.org> - 2017-03-14 14:00 +0100
Re: perf: use-after-free in perf_release Oleg Nesterov <oleg@redhat.com> - 2017-03-14 14:30 +0100
Re: perf: use-after-free in perf_release Peter Zijlstra <peterz@infradead.org> - 2017-03-14 14:50 +0100
Re: perf: use-after-free in perf_release Oleg Nesterov <oleg@redhat.com> - 2017-03-14 15:10 +0100
Re: perf: use-after-free in perf_release Peter Zijlstra <peterz@infradead.org> - 2017-03-14 15:10 +0100
Re: perf: use-after-free in perf_release Oleg Nesterov <oleg@redhat.com> - 2017-03-14 15:40 +0100
Re: perf: use-after-free in perf_release Peter Zijlstra <peterz@infradead.org> - 2017-03-14 16:10 +0100
Re: perf: use-after-free in perf_release Peter Zijlstra <peterz@infradead.org> - 2017-03-14 16:10 +0100
Re: perf: use-after-free in perf_release Oleg Nesterov <oleg@redhat.com> - 2017-03-14 16:40 +0100
Re: perf: use-after-free in perf_release Peter Zijlstra <peterz@infradead.org> - 2017-03-14 16:50 +0100
Re: perf: use-after-free in perf_release Peter Zijlstra <peterz@infradead.org> - 2017-03-14 16:30 +0100
Re: perf: use-after-free in perf_release Peter Zijlstra <peterz@infradead.org> - 2017-03-14 18:40 +0100
Re: perf: use-after-free in perf_release Oleg Nesterov <oleg@redhat.com> - 2017-03-15 17:50 +0100
Re: perf: use-after-free in perf_release Peter Zijlstra <peterz@infradead.org> - 2017-03-16 13:10 +0100
Re: perf: use-after-free in perf_release Peter Zijlstra <peterz@infradead.org> - 2017-03-16 15:00 +0100
Re: perf: use-after-free in perf_release Oleg Nesterov <oleg@redhat.com> - 2017-03-16 17:50 +0100
Re: perf: use-after-free in perf_release Oleg Nesterov <oleg@redhat.com> - 2017-03-14 16:30 +0100
csiph-web