Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [tracecompass-dev] tracecompass Resources view doesn't make sense

On 2017-02-03 09:19 AM, Matthew Khouzam wrote:
> I might be able to help here.
> The views don't depend on LTTng per-se. The events are mapped to
> intermediate ones like (hardware IRQ entry - exit)
> So it should work, it worked with windows traces. But the mapping needs
> to be correct and it is an area that is untested.
> @Alexandre If you have some perf-ctf traces lying around, and could put
> them in the test git (I don't have access to it), we could better test perf.

We do have a perf trace in the test-traces repo:

Problem is we just do CTF-level validation on it. It could be worth
adding som os.linux-related tests making use of it too.


> BR,
> Matthew.
> On 17-02-03 06:38 AM, Mathieu Desnoyers wrote:
>> Alexandre,
>> How well is the resource view of Trace Compass supposed to
>> deal with events coming from perf ? I know it's been designed
>> to handle lttng-modules events initially.
>> Thanks,
>> Mathieu
>> ----- On Feb 3, 2017, at 4:59 AM, Joel Fernandes joelaf@xxxxxxxxxx wrote:
>>> Hi,
>>> I ran perf to collect some trace events an android device and
>>> converted to CTF just to see what it looks like in tracecompass.
>>> I find that in the 'Resources' view, when I click on the "+" next to
>>> CPUx, the IRQs in the drop down appear to be active for much longer
>>> than how they appear to be on the CPUx row itself. For example, in the
>>> below screen shot [1], you can see this for CPU 0.
>>> Am I doing something wrong?
>>> I happen to enable a lot of events many which are not viewable in
>>> tracecompass, I'm hoping those are ignored and doesn't cause problems.
>>> Thanks,
>>> [1]
> _______________________________________________
> tracecompass-dev mailing list
> tracecompass-dev@xxxxxxxxxxx
> To change your delivery options, retrieve your password, or unsubscribe from this list, visit

Back to the top