linux-kernel - Re: [PATCH] perf mem2node: avoid double free related to realloc

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 for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAP-5=fWmJdAC3GtwO3sf+725roY6T8s8GRvDJL4Rfqv+ByysGQ@mail.gmail.com>
Date:   Fri, 20 Mar 2020 11:25:47 -0700
From:   Ian Rogers <irogers@...gle.com>
To:     Jiri Olsa <jolsa@...hat.com>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        Ingo Molnar <mingo@...hat.com>,
        Arnaldo Carvalho de Melo <acme@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
        Namhyung Kim <namhyung@...nel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        clang-built-linux <clang-built-linux@...glegroups.com>,
        Stephane Eranian <eranian@...gle.com>
Subject: Re: [PATCH] perf mem2node: avoid double free related to realloc

On Sun, Mar 15, 2020 at 2:30 AM Jiri Olsa <jolsa@...hat.com> wrote:
>
> On Sat, Mar 14, 2020 at 09:49:37AM -0700, Ian Rogers wrote:
> > On Sat, Mar 14, 2020 at 7:23 AM Jiri Olsa <jolsa@...hat.com> wrote:
> > >
> > > On Fri, Mar 13, 2020 at 09:28:26PM -0700, Ian Rogers wrote:
> > > > Realloc of size zero is a free not an error, avoid this causing a double
> > > > free. Caught by clang's address sanitizer:
> > > >
> > > > ==2634==ERROR: AddressSanitizer: attempting double-free on 0x6020000015f0 in thread T0:
> > > >     #0 0x5649659297fd in free llvm/llvm-project/compiler-rt/lib/asan/asan_malloc_linux.cpp:123:3
> > > >     #1 0x5649659e9251 in __zfree tools/lib/zalloc.c:13:2
> > > >     #2 0x564965c0f92c in mem2node__exit tools/perf/util/mem2node.c:114:2
> > > >     #3 0x564965a08b4c in perf_c2c__report tools/perf/builtin-c2c.c:2867:2
> > > >     #4 0x564965a0616a in cmd_c2c tools/perf/builtin-c2c.c:2989:10
> > > >     #5 0x564965944348 in run_builtin tools/perf/perf.c:312:11
> > > >     #6 0x564965943235 in handle_internal_command tools/perf/perf.c:364:8
> > > >     #7 0x5649659440c4 in run_argv tools/perf/perf.c:408:2
> > > >     #8 0x564965942e41 in main tools/perf/perf.c:538:3
> > > >
> > > > 0x6020000015f0 is located 0 bytes inside of 1-byte region [0x6020000015f0,0x6020000015f1)
> > > > freed by thread T0 here:
> > > >     #0 0x564965929da3 in realloc third_party/llvm/llvm-project/compiler-rt/lib/asan/asan_malloc_linux.cpp:164:3
> > > >     #1 0x564965c0f55e in mem2node__init tools/perf/util/mem2node.c:97:16
> > > >     #2 0x564965a08956 in perf_c2c__report tools/perf/builtin-c2c.c:2803:8
> > > >     #3 0x564965a0616a in cmd_c2c tools/perf/builtin-c2c.c:2989:10
> > > >     #4 0x564965944348 in run_builtin tools/perf/perf.c:312:11
> > > >     #5 0x564965943235 in handle_internal_command tools/perf/perf.c:364:8
> > > >     #6 0x5649659440c4 in run_argv tools/perf/perf.c:408:2
> > > >     #7 0x564965942e41 in main tools/perf/perf.c:538:3
> > > >
> > > > previously allocated by thread T0 here:
> > > >     #0 0x564965929c42 in calloc third_party/llvm/llvm-project/compiler-rt/lib/asan/asan_malloc_linux.cpp:154:3
> > > >     #1 0x5649659e9220 in zalloc tools/lib/zalloc.c:8:9
> > > >     #2 0x564965c0f32d in mem2node__init tools/perf/util/mem2node.c:61:12
> > > >     #3 0x564965a08956 in perf_c2c__report tools/perf/builtin-c2c.c:2803:8
> > > >     #4 0x564965a0616a in cmd_c2c tools/perf/builtin-c2c.c:2989:10
> > > >     #5 0x564965944348 in run_builtin tools/perf/perf.c:312:11
> > > >     #6 0x564965943235 in handle_internal_command tools/perf/perf.c:364:8
> > > >     #7 0x5649659440c4 in run_argv tools/perf/perf.c:408:2
> > > >     #8 0x564965942e41 in main tools/perf/perf.c:538:3
> > > >
> > > > Signed-off-by: Ian Rogers <irogers@...gle.com>
> > > > ---
> > > >  tools/perf/util/mem2node.c | 2 +-
> > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git a/tools/perf/util/mem2node.c b/tools/perf/util/mem2node.c
> > > > index 797d86a1ab09..7f97aa69eb65 100644
> > > > --- a/tools/perf/util/mem2node.c
> > > > +++ b/tools/perf/util/mem2node.c
> > > > @@ -95,7 +95,7 @@ int mem2node__init(struct mem2node *map, struct perf_env *env)
> > > >
> > > >       /* Cut unused entries, due to merging. */
> > > >       tmp_entries = realloc(entries, sizeof(*entries) * j);
> > > > -     if (tmp_entries)
> > > > +     if (j == 0 || tmp_entries)
> > >
> > > nice catch.. I wonder if we should fail in here, or at least
> > > warn that there're no entris.. which is really strange ;-)
> >
> > The workload was the stream benchmark with perf c2c, but stream isn't
> > particularly long running. Not sure how j became 0, there's 2
> > possibilities in the code. The worry with a warning is the spam, so I
> > just wanted to make the code correct.
>
> I was wondering if we should fail completely,
> but we might break some c2c expected behaviour
>
> how about just WARN_ONCE pn j == 0, and then
> the lookup will fail already with:
>
>         if (WARN_ONCE(node < 0, "WARNING: failed to find node\n"))
>                 return;
>
> just get enough hints ;-)

Sorry for the delay. I sent a v2 hopefully doing what you asked :-)
https://lkml.org/lkml/2020/3/20/858

Thanks,
Ian

> jirka
>
> >
> > Thanks,
> > Ian
> >
> > > thanks,
> > > jirka
> > >
> > > >               entries = tmp_entries;
> > > >
> > > >       for (i = 0; i < j; i++) {
> > > > --
> > > > 2.25.1.481.gfbce0eb801-goog
> > > >
> > >
> >
>