linux-kernel - Re: [PATCH] perf parse-events: add defensive null check

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAP-5=fV9LVmaeSsXaD_oA=SZHjY=nhXi-RgJFe0EXzWzBdjsOQ@mail.gmail.com>
Date:   Wed, 25 Mar 2020 14:35:45 -0700
From:   Ian Rogers <irogers@...gle.com>
To:     Arnaldo Carvalho de Melo <arnaldo.melo@...il.com>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        Ingo Molnar <mingo@...hat.com>,
        Mark Rutland <mark.rutland@....com>,
        Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
        Jiri Olsa <jolsa@...hat.com>,
        Namhyung Kim <namhyung@...nel.org>,
        John Garry <john.garry@...wei.com>,
        Jin Yao <yao.jin@...ux.intel.com>,
        Andi Kleen <ak@...ux.intel.com>,
        Kan Liang <kan.liang@...ux.intel.com>,
        LKML <linux-kernel@...r.kernel.org>,
        clang-built-linux <clang-built-linux@...glegroups.com>,
        Stephane Eranian <eranian@...gle.com>
Subject: Re: [PATCH] perf parse-events: add defensive null check

On Wed, Mar 25, 2020 at 12:25 PM Arnaldo Carvalho de Melo
<arnaldo.melo@...il.com> wrote:
>
> Em Wed, Mar 25, 2020 at 09:40:22AM -0700, Ian Rogers escreveu:
> > Terms may have a null config in which case a strcmp will segv. This can
> > be reproduced with:
> >   perf stat -e '*/event=?,nr/' sleep 1
> > Add a null check to avoid this. This was caught by LLVM's libfuzzer.
>
> Adding the NULL check doesn't hurt, I guess, but I coudln't repro it:
>
>   [root@...enth ~]# perf stat -e '*/event=?,nr/' sleep 1
>   WARNING: multiple event parsing errors
>   event syntax error: '*/event=?,nr/'
>                       \___ 'nr' is not usable in 'perf stat'
>
>   Initial error:
>   event syntax error: '*/event=?,nr/'
>                        \___ Cannot find PMU `*'. Missing kernel support?
>   Run 'perf list' for a list of valid events
>
>    Usage: perf stat [<options>] [<command>]
>
>       -e, --event <event>   event selector. use 'perf list' to list available events
>   [root@...enth ~]#
>
> Does this take place only when libfuzzer is being used?

Good catch! I was driving the repro through the fuzzer and getting the
stack trace below and had assumed this would repro on the command
line. I'm now wondering why it won't reproduce :-) I suspect this
issue can come up in other scenarios so as you say the null check
won't hurt, but the commit message is incorrect.

#0  __interceptor_strcmp ()
    at llvm/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc:448
#1  0x0000555557345eb2 in pmu_resolve_param_term (term=0x607000081190,
head_terms=0x602000007cb0,
    value=0x7ffff4cc0320) at tools/perf/util/pmu.c:994
#2  0x00005555573389a4 in pmu_config_term (formats=0x6080000012c8,
attr=0x7ffff4dd4e20,
    term=0x607000081190, head_terms=0x602000007cb0, zero=false,
err=0x7ffff4c90020)
    at tools/perf/util/pmu.c:1117
#3  0x0000555557337b4e in perf_pmu__config_terms
(formats=0x6080000012c8, attr=0x7ffff4dd4e20,
    head_terms=0x602000007cb0, zero=false, err=0x7ffff4c90020)
    at tools/perf/util/pmu.c:1154
#4  0x0000555557338f7e in perf_pmu__config (pmu=0x6080000012a0,
attr=0x7ffff4dd4e20,
    head_terms=0x602000007cb0, err=0x7ffff4c90020)
    at tools/perf/util/pmu.c:1174
#5  0x0000555557314027 in parse_events_add_pmu
(parse_state=0x7ffff4c900a0, list=0x602000007e10,
    name=0x602000001910 "uncore_cha_1", head_config=0x602000007cb0,
auto_merge_stats=true,
    use_alias=false) at tools/perf/util/parse-events.c:1485
#6  0x0000555556d257f9 in parse_events_parse
(_parse_state=0x7ffff4c900a0, scanner=0x611000000cc0)
    at tools/perf/util/parse-events.y:318
#7  0x000055555731c573 in parse_events__scanner (str=0x7ffff4d85c40
"*/event=?,nr/",
    parse_state=0x7ffff4c900a0, start_token=258)
    at tools/perf/util/parse-events.c:2026
#8  0x000055555731cbc3 in parse_events (evlist=0x61e000000080,
str=0x7ffff4d85c40 "*/event=?,nr/",
    err=0x7ffff4c90020) at tools/perf/util/parse-events.c:2066

Thanks,
Ian

> - Arnaldo
>
> > Signed-off-by: Ian Rogers <irogers@...gle.com>
> > ---
> >  tools/perf/util/pmu.c | 11 +++++------
> >  1 file changed, 5 insertions(+), 6 deletions(-)
> >
> > diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
> > index 616fbda7c3fc..ef6a63f3d386 100644
> > --- a/tools/perf/util/pmu.c
> > +++ b/tools/perf/util/pmu.c
> > @@ -984,12 +984,11 @@ static int pmu_resolve_param_term(struct parse_events_term *term,
> >       struct parse_events_term *t;
> >
> >       list_for_each_entry(t, head_terms, list) {
> > -             if (t->type_val == PARSE_EVENTS__TERM_TYPE_NUM) {
> > -                     if (!strcmp(t->config, term->config)) {
> > -                             t->used = true;
> > -                             *value = t->val.num;
> > -                             return 0;
> > -                     }
> > +             if (t->type_val == PARSE_EVENTS__TERM_TYPE_NUM &&
> > +                 t->config && !strcmp(t->config, term->config)) {
> > +                     t->used = true;
> > +                     *value = t->val.num;
> > +                     return 0;
> >               }
> >       }
> >
> > --
> > 2.25.1.696.g5e7596f4ac-goog
> >
>
> --
>
> - Arnaldo