[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAM9d7chXJTcdJ+_WbmXVafUpsMpM7RPZ+vmwg2wkxQJsP8yT-w@mail.gmail.com>
Date: Fri, 8 Apr 2016 00:43:53 -0700
From: Namhyung Kim <namhyung@...nel.org>
To: Arnaldo Carvalho de Melo <acme@...nel.org>
Cc: Jiri Olsa <jolsa@...nel.org>, David Ahern <dsahern@...il.com>,
Wang Nan <wangnan0@...wei.com>,
Adrian Hunter <adrian.hunter@...el.com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: perf report --no-children chopping off tracepoint fields
Hi Arnaldo,
On Thu, Apr 7, 2016 at 7:51 AM, Arnaldo Carvalho de Melo
<acme@...nel.org> wrote:
> Hi Namhyung,
>
> If I do:
>
> # perf record --call dwarf -p 2519 -e syscalls:sys_enter_open
>
> And then run plain 'perf report' I get this on the TUI, perfect:
>
> Samples: 1 of event 'syscalls:sys_enter_open', Event count (approx.): 1
> Children Self Trace output
> + 100.00% 100.00% filename: 0x7efe0cc9a96a, flags: 0x00000000, mode: 0x7fffa3da07f0
>
> But if I use 'perf report --no-children' I get it chopped off at the
> start of the first tp arg value:
>
> Samples: 1 of event 'syscalls:sys_enter_open', Event count (approx.): 1
> Overhead Trace output
> + 100.00% filename: 0x
>
> :-\
Hmm.. I cannot reproduce the problem. All (non-filtered) entries have
trace_output should calculate the column length in the output_resort()
-> hists__calc_col_len().
Btw it seems the --children is meaningless if symbol or dso sort keys
are not used.
Thanks,
Namhyung
Powered by blists - more mailing lists