Re: [PATCH v7 2/6] perf record: Get the first sample time and last sample time

[Jin, Yao]

On 1/5/2018 8:53 PM, Arnaldo Carvalho de Melo wrote:
> Em Fri, Jan 05, 2018 at 09:15:03AM +0800, Jin, Yao escreveu:
> > On 1/5/2018 3:09 AM, Arnaldo Carvalho de Melo wrote:
> > > Em Fri, Dec 08, 2017 at 09:13:42PM +0800, Jin Yao escreveu:
> > > > In the default 'perf record' configuration, all samples are processed,
> > > > to create the HEADER_BUILD_ID table. So it's very easy to get the
> > > > first/last samples and save the time to perf file header via the
> > > > function write_sample_time().
> > > >
> > > > Later, at post processing time, perf report/script will fetch
> > > > the time from perf file header.
> > >
> > > So, at this point I was expecting that that record would be present on
> > > the perf.data file:
> > >
> > > [acme@jouet perf]$ perf record --timestamp-boundary sleep 1
> > > Cannot read kernel map
> > > [ perf record: Woken up 1 times to write data ]
> > > [ perf record: Captured and wrote 0.001 MB perf.data (7 samples) ]
> > > [acme@jouet perf]$ perf report -D | grep PERF_RECORD_SAMPLE | wc -l
> > > 7
> > > [acme@jouet perf]$ perf report -D | grep PERF_RECORD_SAMPLE_TIME
> > > [acme@jouet perf]$
> > >
> > > What am I doing wrong?
> > >
> > > To clarify, this is with just the first two patches in this series
> > > applied.
> > >
> > > - Arnaldo
> >
> > Hi Arnaldo,
> >
> > The timestamp boundary information is saved in perf file header.
>
> Right, my bad, I somehow thought it would be as PERF_RECORD_SAMPLE_TIME,
> duh.
>
> Will continue reviewing, sorry about that.
>
> - Arnaldo
>   

Thanks for reviewing the patch.

If you see anything I should improve, please let me know.

Thanks
Jin Yao

> > So if we want to look at them, we need to add '--header' in perf report.
> >
> > For example,
> >
> > root@skl:/tmp# perf report -D --header | grep 'time of'
> > # time of first sample : 248333.706656
> > # time of last sample : 248357.215328
> >
> > Thanks
> > Jin Yao