Jérémie, can you have a look at these questions ?

Thanks,

Mathieu

* Salman Rafiq ([email protected]) wrote:
> Hello All,
> 
> In continuation to my previous questions related to reading CTF traces from 
> particular timestamp, I can now further summarize few problems/questions I 
> have:
> 
> 
> 1.       Is it possible to seek CTF file_stream using 'begin' and 'end' 
> iterator position i.e. BT_SEEK_TIME?
> 
> 
> 
> a.       We cannot use babel trace API for CTF iterator creation using 
> defined start and end position based on time. Seems this is in progress?
> 
> b.      The other possible solution I am using, is to create the iterator 
> using defaults, and then later use API to 'create iterator time position 
> based on Raw Timestamp (only takes raw time)' and then 'set iterator position 
> to that timed position'.
> 
> 
> 
> 2.       But, since the area of interest for me was only the time when user 
> application which I am tracing using Lttng-kernel tracer, started recording 
> events (application use external trace recording format).
> 
> User application contains a definition file which tells me the epoch time (us 
> resolution) when it started and stopped tracing. I wanted to use this 
> information to set iterator positions.
> 
> 
> 
> a.       Question is Can I convert this time to raw time which babeltrace API 
> takes as an argument?
> 
> b.      Or Is there any way to read Clock struct from CTF metadata file (is 
> there any scope defined for this), so that I can extract offset variable 
> value to convert epoch time from other trace to raw time?
> 
> 
> It will be really great if I can get few suggestions/help related to these 
> issues.
> Thanking in anticipation!
> 
> 
> Best Regards,
> 
> Salman Rafiq
> Research Engineer
> Industrial Communication
> Fraunhofer-Einrichtung für Systeme der Kommunikationstechnik ESK
> 
> Hansastraße 32 | 80686 München
> Telefon, Fax:  +49 89 547088-356 | +49 89 547088-66-356
> E-Mail:   [email protected]<http://www.esk.fraunhofer.de/>
> 
> Internet:
> http://www.esk.fraunhofer.de<http://www.esk.fraunhofer.de/>
> http://www.facebook.com/FraunhoferESK
> http://www.twitter.com/FraunhoferESK
> 
> Von: Salman Rafiq [mailto:[email protected]]
> Gesendet: Mittwoch, 18. Juli 2012 15:28
> An: [email protected]
> Betreff: [lttng-dev] Reading CTF trace from time T
> 
> Hello All,
> 
> I am tracing a user application using lttng kernel tracer, i.e.,
> 
> Start lttng kernel trace
> ./application
> Stop lttng kernel trace
> 
> Before I was reading kernel CTF trace using babeltrace API's from beginning 
> of trace, e.g.,
> struct  bt_iter_pos  begin_pos
> struct bt_ctf_iter *iterator
> 
> begin_pos.type = BT_SEEK_BEGIN
> iterator = create_iterator(context,  &begin_pos, NULL)
> .. and then read events start from that iterator postion.
> 
> Recently, I have been trying to seek trace using timestamp from when the 
> application started to application end.
> I would like to know if I am doing it right. Would the method below enough to 
> achieve this? Or I am missing something here :-S
> 
> begin_pos.type = BT_SEEK_TIME
> end_pos.type = BT_SEEK_TIME
> 
> begin_pos.u.seek_time = timestamp1 (for now manually taken from CTF kernel 
> trace when application started)
> end_pos.u.seek_time = timestamp2 (for now manually taken from CTF kernel 
> trace when application ended)
> 
> iterator = create_iterator(context,  &begin_pos, &end_pos);
> and then with reading events at iterator position until end_pos.
> 
> 
> I am using babeltrace package version "1.0.0-pre4".
> Any help related to achieving this functionality will be highly appreciated.
> 
> Best Regards,
> Salman
> 
> 
> --
> Salman Rafiq
> Industrial Communication
> Fraunhofer-Einrichtung für Systeme der Kommunikationstechnik ESK
> 
> Hansastraße 32 | 80686 München
> Telefon, Fax:  +49 89 547088-356 | +49 89 547088-66-356
> E-Mail:   [email protected]<http://www.esk.fraunhofer.de/>
> 
> Internet:
> http://www.esk.fraunhofer.de<http://www.esk.fraunhofer.de/>
> http://www.facebook.com/FraunhoferESK
> http://www.twitter.com/FraunhoferESK
> 

Content-Description: ATT00001.txt
> _______________________________________________
> lttng-dev mailing list
> [email protected]
> http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev


-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com

_______________________________________________
lttng-dev mailing list
[email protected]
http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

Reply via email to