[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAD=GYpbSh_mu3skV+8nVaObQuJKYgcmBXFmEDWL24kb_tk6ZXw@mail.gmail.com>
Date: Tue, 1 Mar 2016 02:34:35 -0800
From: Joel Fernandes <agnel.joel@...il.com>
To: 平松雅巳 / HIRAMATU,MASAMI
<masami.hiramatsu.pt@...achi.com>
Cc: "ltc-kernel@...yrl.intra.hitachi.co.jp"
<ltc-kernel@...yrl.intra.hitachi.co.jp>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
Taeung Song <treeze.taeung@...il.com>,
Joel Fernandes <joel.opensrc@...il.com>,
"linux-perf-users@...r.kernel.org" <linux-perf-users@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Namhyung Kim <namhyung@...nel.org>,
Jiri Olsa <jolsa@...nel.org>,
sysp-manager <cti.systems-productivity-manager.ts@...achi.com>
Subject: Re: [RFC] perf: probe_finder: continue if atleast one probe point found
Hi Hiramatu,
On Mon, Feb 29, 2016 at 8:29 PM, 平松雅巳 / HIRAMATU,MASAMI
<masami.hiramatsu.pt@...achi.com> wrote:
> BTW,
>
> ./perf probe \
> -k ./vmlinux -s ./ -x /lib/x86_64-linux-gnu/libc.so.6 -a 'malloc $params'
>
> might help your case. $params is expanded to function parameters automatically
> and if there is no parameters, it is just ignored :)
Thanks, I agree this is a better approach.
How about an informative error letting the user know about it? Is the
below Ok? I can send a patch.
----x8--------
diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c
index 4ce5c5e..21f282a 100644
--- a/tools/perf/util/probe-finder.c
+++ b/tools/perf/util/probe-finder.c
@@ -1247,8 +1247,12 @@ static int add_probe_trace_event(Dwarf_Die
*sc_die, struct probe_finder *pf)
pf->tvar = &tev->args[i];
/* Variable should be found from scope DIE */
ret = find_variable(sc_die, pf);
- if (ret != 0)
+ if (ret != 0) {
+ pr_err("Variable couldn't be found, if probe
is on inline function,"
+ "try '%s' or '%s' instead.\n",
+ PROBE_ARG_PARAMS, PROBE_ARG_VARS);
break;
+ }
}
Powered by blists - more mailing lists