[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20170411184435.GH3275@kernel.org>
Date: Tue, 11 Apr 2017 15:44:35 -0300
From: Arnaldo Carvalho de Melo <acme@...nel.org>
To: David Carrillo-Cisneros <davidcc@...gle.com>
Cc: linux-kernel@...r.kernel.org,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Andi Kleen <ak@...ux.intel.com>, Simon Que <sque@...omium.org>,
Wang Nan <wangnan0@...wei.com>, Jiri Olsa <jolsa@...nel.org>,
He Kuang <hekuang@...wei.com>,
Masami Hiramatsu <mhiramat@...nel.org>,
Stephane Eranian <eranian@...gle.com>,
Paul Turner <pjt@...gle.com>
Subject: Re: [PATCH 5/7] perf session: don't rely on evlist in pipe mode
Em Tue, Apr 11, 2017 at 03:40:50PM -0300, Arnaldo Carvalho de Melo escreveu:
> Em Tue, Apr 11, 2017 at 03:40:14PM -0300, Arnaldo Carvalho de Melo escreveu:
> > Em Mon, Apr 10, 2017 at 01:14:30PM -0700, David Carrillo-Cisneros escreveu:
> > > Session sets a number parameters that rely on evlist. These parameters
> > > are not used in pipe-mode and should not be set, since evlist is
> > > unavailable. Fix that.
> >
> > Ditching this one as well, until further investigation, well, file must
> > be null in this case...
This should be enough, right?
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index db554b7461b8..37cf4e6abf9f 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -144,7 +144,7 @@ struct perf_session *perf_session__new(struct perf_data_file *file,
* set session attributes that are present in perf.data
* but not in pipe-mode.
*/
- if (!file->is_pipe) {
+ if (!file || !file->is_pipe) {
perf_session__set_id_hdr_size(session);
perf_session__set_comm_exec(session);
}
@@ -166,7 +166,7 @@ struct perf_session *perf_session__new(struct perf_data_file *file,
* In pipe-mode, evlist is empty until PERF_RECORD_HEADER_ATTR is
* processed, so perf_evlist__sample_id_all is not meaningful here.
*/
- if (!file->is_pipe && tool && tool->ordering_requires_timestamps &&
+ if ((!file || !file->is_pipe) && tool && tool->ordering_requires_timestamps &&
tool->ordered_events && !perf_evlist__sample_id_all(session->evlist)) {
dump_printf("WARNING: No sample_id_all support, falling back to unordered processing\n");
tool->ordered_events = false;
> Ah, this is the problem:
>
> [root@...et ~]# perf top
> perf: Segmentation fault
> -------- backtrace --------
> perf[0x57eddb]
> /lib64/libc.so.6(+0x35990)[0x7fb435d06990]
> perf(perf_session__new+0xb7)[0x4eb547]
> perf(cmd_top+0xde9)[0x44ce89]
> perf[0x49953f]
> perf[0x499831]
> perf(main+0x252)[0x428dd2]
> /lib64/libc.so.6(__libc_start_main+0xf1)[0x7fb435cf1401]
> perf(_start+0x2a)[0x4292da]
> [0x0]
> [root@...et ~]#
>
> > [acme@...et linux]$ git bisect bad
> > 5d71c27204d51492abe9fc69064fd36a0bda410b is the first bad commit
> > commit 5d71c27204d51492abe9fc69064fd36a0bda410b
> > Author: David Carrillo-Cisneros <davidcc@...gle.com>
> > Date: Mon Apr 10 13:14:30 2017 -0700
> >
> > perf session: Don't rely on evlist in pipe mode
> >
> > Session sets a number parameters that rely on evlist. These parameters
> > are not used in pipe-mode and should not be set, since evlist is
> > unavailable. Fix that.
> >
> > Signed-off-by: David Carrillo-Cisneros <davidcc@...gle.com>
> > Acked-by: Jiri Olsa <jolsa@...nel.org>
> > Cc: Alexander Shishkin <alexander.shishkin@...ux.intel.com>
> > Cc: Andi Kleen <ak@...ux.intel.com>
> > Cc: He Kuang <hekuang@...wei.com>
> > Cc: Masami Hiramatsu <mhiramat@...nel.org>
> > Cc: Paul Turner <pjt@...gle.com>
> > Cc: Peter Zijlstra <peterz@...radead.org>
> > Cc: Simon Que <sque@...omium.org>
> > Cc: Stephane Eranian <eranian@...gle.com>
> > Cc: Wang Nan <wangnan0@...wei.com>
> > Link: http://lkml.kernel.org/r/20170410201432.24807-6-davidcc@google.com
> > Signed-off-by: Arnaldo Carvalho de Melo <acme@...hat.com>
> >
> > :040000 040000 7df28fe9d7c9d962b78c0532f4961caea2737b59 34e4fed0042b1401c1a1a247d8ae40df17f9d272 M tools
> > [acme@...et linux]$
> >
> > > Signed-off-by: David Carrillo-Cisneros <davidcc@...gle.com>
> > > ---
> > > tools/perf/util/session.c | 16 +++++++++++++---
> > > 1 file changed, 13 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
> > > index a25302bc55a8..db554b7461b8 100644
> > > --- a/tools/perf/util/session.c
> > > +++ b/tools/perf/util/session.c
> > > @@ -140,8 +140,14 @@ struct perf_session *perf_session__new(struct perf_data_file *file,
> > > if (perf_session__open(session) < 0)
> > > goto out_close;
> > >
> > > - perf_session__set_id_hdr_size(session);
> > > - perf_session__set_comm_exec(session);
> > > + /*
> > > + * set session attributes that are present in perf.data
> > > + * but not in pipe-mode.
> > > + */
> > > + if (!file->is_pipe) {
> > > + perf_session__set_id_hdr_size(session);
> > > + perf_session__set_comm_exec(session);
> > > + }
> > > }
> > > } else {
> > > session->machines.host.env = &perf_env;
> > > @@ -156,7 +162,11 @@ struct perf_session *perf_session__new(struct perf_data_file *file,
> > > pr_warning("Cannot read kernel map\n");
> > > }
> > >
> > > - if (tool && tool->ordering_requires_timestamps &&
> > > + /*
> > > + * In pipe-mode, evlist is empty until PERF_RECORD_HEADER_ATTR is
> > > + * processed, so perf_evlist__sample_id_all is not meaningful here.
> > > + */
> > > + if (!file->is_pipe && tool && tool->ordering_requires_timestamps &&
> > > tool->ordered_events && !perf_evlist__sample_id_all(session->evlist)) {
> > > dump_printf("WARNING: No sample_id_all support, falling back to unordered processing\n");
> > > tool->ordered_events = false;
> > > --
> > > 2.12.2.715.g7642488e1d-goog
Powered by blists - more mailing lists