Messages in this thread | | | Date | Tue, 29 Apr 2014 18:19:29 -0400 | From | Steven Rostedt <> | Subject | Re: [PATCH v2] tracing: Allow changing default ring buffer size for ftrace instances. |
| |
On Tue, 29 Apr 2014 15:09:15 -0700 Bharath Ravi <rbharath@google.com> wrote:
> We'd actually considered doing just that, in an initial version of the patch. > At the time, we'd decided in favour of allowing the user more explicit > control over the buffer size, rather than auto expanding it. (The > discussion on that is here: > https://lkml.org/lkml/2013/7/1/617)
Yeah, I remember my rational :-)
The thing is, I have been thinking more about it, and I'm now thinking it would be fine. But reading that email again, I may have convinced myself that it should have the option. Gah, I'm now arguing with myself!
> > We could go back to that though if you think that's okay, but I think > your previous reasoning still makes sense: it's useful for the user to > know if there will be an allocation failure before enabling tracing.
Grumble. I need to think about this a bit more. I'm stretching myself a bit thin between different projects. I'll pull in you patches and play with it a bit and then come back with a decision.
Thanks!
-- Steve
| |