On Sun, Mar 2, 2014 at 11:07 AM, zlynus <[email protected]> wrote: > Hi, > I'm a graduate student from China. Due to my last experience using SystemTap > to trace kernel's submit_bio() function activity in my recent research on > disk blocks' correlation, I get interested in system trace tool .So I think > contributing to lttng in GSoC would be a great adventure. >
Absolutely! > According to the idea page, the kernel side tracer has no comparable > filtering mechanism like the one in user-space side however, it is proposed > that the filtering be implemented in the babeltrace , rather than in kernel > side. Can you please tell me the reason? > We have been looking at kernel-level filtering for quite some time and are interested in the feature. The reason we are proposing the implementation of a filtering system in Babeltrace is to both address this shortcoming and provide a non-destructive way of filtering events. It would make it easier for users to search for events in a trace and lower Babeltrace's output volume to speed-up scripts based on it. Ideally, we'd have filtering both in Babeltrace and the kernel, although the latter is out of GSoC's scope. > Nevertheless, I think implementing a usable and relatively optimized filter > is a great work worth working on. I have read the man page and tested the > toolchain. Now I have got a whole picture of it. And in order to get > familiar with you guys and prove my skills I should try to submit patches to > this list, right? Beside that I hope hear your discussion on filter > mechanism. > That would be the way to go. Looking forward to your patches! Thanks, Jérémie > Best regards, > Lingyu Zhu > > > _______________________________________________ > lttng-dev mailing list > [email protected] > http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev > -- Jérémie Galarneau EfficiOS Inc. http://www.efficios.com _______________________________________________ lttng-dev mailing list [email protected] http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev
