[lttng-dev] Babeltrace API timestamp manipulation changes

Mathieu Desnoyers mathieu.desnoyers at efficios.com
Sun Aug 5 13:46:35 EDT 2012


* Yannick Brosseau (yannick.brosseau at gmail.com) wrote:
> On 2012-08-03 15:27, Julien Desfossez wrote:
> > Hi,
> >
> > Previously, the Babeltrace API forced the user to manipulate "raw
> > timestamps" to seek, to get the timestamp begin and end of a trace, etc.
> >
> > This raw timestamp represented the cycle value scaled with the frequency
> > but did not meant anything to the user and was likely to cause problem
> > when dealing with traces recorded with different clock sources.
> >
> > The new patches commited in Babeltrace today removed the notion of the
> > raw clock-source all over the code, and now the users can manipulate
> > real timestamps (with the timezone offset applied) or cycle values if
> > they want more fine-grained control on a specific trace.
> >
> > This means some API changes :
> > - the "seek_time" member of the bt_iter_pos structure now takes real
> > timestamp (with the offset applied)
> > - the functions "bt_ctf_get_timestamp_raw" is removed
> > - the funtion "bt_ctf_get_timestamp" is replaced by
> > "bt_ctf_get_real_timestamp"
> > - the function "bt_ctf_get_cycles_timestamp" is added
> >
> Why not simply calling them
> bt_ctf_get_timestamp and bt_ctf_get_cycles  ?

Yes, it makes sense. Julien, any thoughts on this ?

Thanks,

Mathieu

> 
> Yannick
> 
> _______________________________________________
> lttng-dev mailing list
> lttng-dev at lists.lttng.org
> http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

-- 
Mathieu Desnoyers
Operating System Efficiency R&D Consultant
EfficiOS Inc.
http://www.efficios.com



More information about the lttng-dev mailing list