Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Mon, 24 Aug 2020 22:22:15 +0200
From:   Jiri Olsa <jolsa@redhat.com>
To:     Arnaldo Carvalho de Melo <acme@kernel.org>
Cc:     Jin Yao <yao.jin@linux.intel.com>,
        Thomas Richter <tmricht@linux.ibm.com>,
        Jiri Olsa <jolsa@kernel.org>,
        Namhyung Kim <namhyung@kernel.org>,
        linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
        svens@linux.ibm.com, gor@linux.ibm.com, sumanthk@linux.ibm.com,
        heiko.carstens@de.ibm.com
Subject: Re: [PATCH] Fix s390x compile error on F32 utils/stat-display.c
Message-ID: <20200824202215.GA631468@krava>
References: <20200722092053.22345-1-tmricht@linux.ibm.com>
 <20200812112708.GA13995@kernel.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20200812112708.GA13995@kernel.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Aug 12, 2020 at 08:27:08AM -0300, Arnaldo Carvalho de Melo wrote:
> Em Wed, Jul 22, 2020 at 11:20:53AM +0200, Thomas Richter escreveu:
> > Fix a compile error on F32 and gcc version 10.1 on s390 in file
> > utils/stat-display.c.  The error does not show up with make DEBUG=y.
> > In fact the issue shows up when using both compiler options
> > -O6 and -D_FORTIFY_SOURCE=2 (which are omitted with DEBUG=Y).
> > 
> > This is the offending call chain:
> > print_counter_aggr()
> >   printout(config, -1, 0, ...)  with 2nd parm id set to -1
> >     aggr_printout(config, x, id --> -1, ...) which leads to this code:
> > 		case AGGR_NONE:
> >                 if (evsel->percore && !config->percore_show_thread) {
> >                         ....
> >                 } else {
> >                         fprintf(config->output, "CPU%*d%s",
> >                                 config->csv_output ? 0 : -7,
> >                                 evsel__cpus(evsel)->map[id],
> > 				                        ^^ id is -1 !!!!
> >                                 config->csv_sep);
> >                 }
> > 
> > This is a compiler inlining issue which is detected on s390 but not on
> > other plattforms.
> 
> What is the sequence of events that gets to this? I.e. is it valid to
> get a config->aggr_mode == AGGR_NONE, then have evsel not be percore and
> config->percore_show_thread to be false?
> 
> I wonder if this won't be papering over some bug :-\
> 
> Jin?
> 
> This is where this came from:
> commit 4fc4d8dfa056dfd48afe73b9ea3b7570ceb80b9c (tag: perf-core-for-mingo-5.2-20190517)
> Author: Jin Yao <yao.jin@linux.intel.com>
> Date:   Fri Apr 12 21:59:49 2019 +0800
> 
>     perf stat: Support 'percore' event qualifier
> 
>     With this patch, we can use the 'percore' event qualifier in perf-stat.
> 
> ---
> 
> Also please add at least Jiri and Namhyung on the CC list, having the
> person that added that array usage also helps.
> 
> [acme@quaco perf]$ scripts/get_maintainer.pl tools/perf | grep reviewer
> Mark Rutland <mark.rutland@arm.com> (reviewer:PERFORMANCE EVENTS SUBSYSTEM)
> Alexander Shishkin <alexander.shishkin@linux.intel.com> (reviewer:PERFORMANCE EVENTS SUBSYSTEM)
> Jiri Olsa <jolsa@redhat.com> (reviewer:PERFORMANCE EVENTS SUBSYSTEM)
> Namhyung Kim <namhyung@kernel.org> (reviewer:PERFORMANCE EVENTS SUBSYSTEM)
> [acme@quaco perf]$
> 
> Thanks,
> 
> - Arnaldo
>  
> > Output before:
> >  # make util/stat-display.o
> >     .....
> > 
> >   util/stat-display.c: In function ‘perf_evlist__print_counters’:
> >   util/stat-display.c:121:4: error: array subscript -1 is below array
> >       bounds of ‘int[]’ [-Werror=array-bounds]
> >   121 |    fprintf(config->output, "CPU%*d%s",
> >       |    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >   122 |     config->csv_output ? 0 : -7,
> >       |     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >   123 |     evsel__cpus(evsel)->map[id],
> >       |     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >   124 |     config->csv_sep);
> >       |     ~~~~~~~~~~~~~~~~
> >   In file included from util/evsel.h:13,
> >                  from util/evlist.h:13,
> >                  from util/stat-display.c:9:
> >   /root/linux/tools/lib/perf/include/internal/cpumap.h:10:7:
> >   note: while referencing ‘map’
> >    10 |  int  map[];
> >       |       ^~~
> >   cc1: all warnings being treated as errors
> >   mv: cannot stat 'util/.stat-display.o.tmp': No such file or directory
> >   make[3]: *** [/root/linux/tools/build/Makefile.build:97: util/stat-display.o]
> >   Error 1
> >   make[2]: *** [Makefile.perf:716: util/stat-display.o] Error 2
> >   make[1]: *** [Makefile.perf:231: sub-make] Error 2
> >   make: *** [Makefile:110: util/stat-display.o] Error 2
> >   [root@t35lp46 perf]#

I tested on s390 with rhel8, but did not get this error,
f32's gcc must be smarter

> > 
> > Output after:
> >   # make util/stat-display.o
> >     .....
> >   CC       util/stat-display.o
> >   [root@t35lp46 perf]#
> > 
> > Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
> > ---
> >  tools/perf/util/stat-display.c | 9 +++++----
> >  1 file changed, 5 insertions(+), 4 deletions(-)
> > 
> > diff --git a/tools/perf/util/stat-display.c b/tools/perf/util/stat-display.c
> > index 57d0706e1330..e49e544188e4 100644
> > --- a/tools/perf/util/stat-display.c
> > +++ b/tools/perf/util/stat-display.c
> > @@ -118,10 +118,11 @@ static void aggr_printout(struct perf_stat_config *config,
> >  				config->csv_output ? 0 : -3,
> >  				cpu_map__id_to_cpu(id), config->csv_sep);
> >  		} else {
> > -			fprintf(config->output, "CPU%*d%s",
> > -				config->csv_output ? 0 : -7,
> > -				evsel__cpus(evsel)->map[id],
> > -				config->csv_sep);
> > +			if (id > -1)
> > +				fprintf(config->output, "CPU%*d%s",
> > +					config->csv_output ? 0 : -7,
> > +					evsel__cpus(evsel)->map[id],
> > +					config->csv_sep);

-1 is only through print_counter_aggr which is called for AGGR_GLOBAL
case as you described, but I guess we can have this check as precaution

you could put it together with above else

    } else if (id > -1)

thanks,
jirka