[Tracing] Deciding the Yaml configuration for barectf input

Vidushi Vashishth reachvidu at gmail.com
Thu Jun 7 02:11:55 UTC 2018


Hey Guillaume!

>I am wondering what tools are currently used for visualizing the resulting
RTEMS trace, if any. And what are your plan on this aspect once you can
output CTF traces with >RTEMS?

RTEMS wiki pages do suggest using trace compass for the purpose of
visualization of traces [1]. It seems like a good choice.

>There is currently interest in providing support for generic RTOS support
in Trace Compass [1], which I'll be working on soon. If we could coordinate
on this aspect >regarding traced events, I think we could achieve
interesting things.

For sure! I look forward to it.

>Also, my two cent for your question is that you are much better
integrating CTF gen directly in RTEMS if modification to babeltrace source
code would be required. I think it >would be simpler to upstream your
changes that way.

Yes that would be simpler. But at some point in time I will have to decide
on a CTF structure for the output traces. I suspect this knowledge will
also be used when changing the babeltrace source code for RTEMS trace
format conversion. And hence the urgent need for advice :p

Thanks for the input:)

Vidushi

On Wed, Jun 6, 2018 at 11:16 PM, Guillaume Champagne <
guillaume.champagne at polymtl.ca> wrote:

> Hi Vidushi,
>
> I am not answering on the mailing list, because I am not sure if my
> questions were already answered previously.
>
> If I understand correctly, the first part of your GSOC project is to
> integrate CTF traces to kernel level event in RTEMS?
> I am wondering what tools are currently used for visualizing the resulting
> RTEMS trace, if any. And what are your plan on this aspect once you can
> output CTF traces with RTEMS?
>
> There is currently interest in providing support for generic RTOS support
> in Trace Compass [1], which I'll be working on soon. If we could coordinate
> on this aspect regarding traced events, I think we could achieve
> interesting things.
>
> Also, my two cent for your question is that you are much better
> integrating CTF gen directly in RTEMS if modification to babeltrace source
> code would be required. I think it would be simpler to upstream your
> changes that way.
>
> [1] http://tracecompass.org/
>
> Guillaume
>
> Vidushi Vashishth <reachvidu at gmail.com> a écrit :
>
>
> Hi!
>>
>> I have written a blog post  (https://vidushivashishth.gith
>> ub.io/sixthpost/)
>> summarizing two approaches to CTF generation that I can employ at this
>> point. I prefer on working with the barectf approach. The only blocker
>> that
>> exists with this approach is deciding on a YAML configuration [1] for
>> defining the structure of the CTF traces. This will be the input to the
>> barectf. I needed advice on how to go about this.
>>
>> References:
>> [1]
>> https://github.com/efficios/barectf/wiki/Writing-the-YAML-co
>> nfiguration-file
>>
>> Thanks,
>> Vidushi
>>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20180607/0bd9e811/attachment-0002.html>


More information about the devel mailing list