Messages in this thread | | | Subject | Re: [RFC][PATCH 2/4] tracing: add rb_event_discard() hack | From | Tom Zanussi <> | Date | Wed, 18 Mar 2009 01:12:05 -0500 |
| |
On Tue, 2009-03-17 at 21:21 -0400, Steven Rostedt wrote: > On Tue, 17 Mar 2009, Tom Zanussi wrote:
> > > > rb_update_read_stamp(cpu_buffer, event); > > @@ -2026,8 +2051,8 @@ rb_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) > > * can have. Nesting 10 deep of interrupts is clearly > > * an anomaly. > > */ > > - if (RB_WARN_ON(cpu_buffer, ++nr_loops > 10)) > > - return NULL; > > +// if (RB_WARN_ON(cpu_buffer, ++nr_loops > 10)) > > +// return NULL; > > Were you hitting this? If so, we need to investigate. >
I was hitting this, but only because my code would always loop around to 'again' for a discarded event. So, 10 discarded events in a row would trigger it.
I noticed that your fix returns instead, so it's no longer a problem.
Thanks,
Tom
| |