On Thu, 23 Jun 2016 10:20:31 -0400 (EDT) Chunyu Hu <[email protected]> wrote:
> I tried the patch, if just for resolving the oops issue, it's indeed enough. > But if we care about just this filter, and compare this with the > trace_graph_entry > filter handling, I found that, the depth 0 item will be submitted to the ring > buffer, not > respecting the filter. I still use the same steps. It's possible to show > schedule entry: Sure, I just wanted to make sure nothing got off, and made the quick patch. Want to send an updated patch that goes this route? That is, make it such that the latency tracers fully support the set_graph_notrace. -- Steve
