linux-kernel - Re: [PATCH] perf top: Fix a race in callchain handling

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

[an error occurred while processing this directive]

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

Message-ID: <87d36gss2m.fsf@sejong.aot.lge.com>
Date:	Mon, 07 May 2012 10:22:41 +0900
From:	Namhyung Kim <namhyung.kim@....com>
To:	Arnaldo Carvalho de Melo <acme@...stprotocols.net>
Cc:	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Paul Mackerras <paulus@...ba.org>,
	Ingo Molnar <mingo@...hat.com>,
	LKML <linux-kernel@...r.kernel.org>,
	Frederic Weisbecker <fweisbec@...il.com>
Subject: Re: [PATCH] perf top: Fix a race in callchain handling

Hi,

On Sun, 6 May 2012 15:07:07 -0300, Arnaldo Carvalho de Melo wrote:
>> 2012-05-05 (토), 20:53 -0300, Arnaldo Carvalho de Melo:
>> > First testing Namhyung patch with -F 100000 tho :-)
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x000000000046a51a in machine__resolve_callchain (self=0x982b20, evsel=0x982440, thread=0x993210, chain=0x7ffff18462d8, parent=0x7fffffffd6f8) at util/session.c:307
> 307				ip = chain->ips[i];
> (gdb) bt
> #0  0x000000000046a51a in machine__resolve_callchain (self=0x982b20, evsel=0x982440, thread=0x993210, chain=0x7ffff18462d8, parent=0x7fffffffd6f8) at util/session.c:307
> #1  0x000000000042b9d2 in perf_event__process_sample (tool=0x7fffffffd8a0, event=0x7ffff18462b8, evsel=0x982440, sample=0x7fffffffd750, machine=0x982b20) at builtin-top.c:777
> #2  0x000000000042bd17 in perf_top__mmap_read_idx (top=0x7fffffffd8a0, idx=0) at builtin-top.c:858
> #3  0x000000000042bdfc in perf_top__mmap_read (top=0x7fffffffd8a0) at builtin-top.c:873
> #4  0x000000000042c45f in __cmd_top (top=0x7fffffffd8a0) at builtin-top.c:1033
> #5  0x000000000042d514 in cmd_top (argc=0, argv=0x7fffffffe630, prefix=0x0) at builtin-top.c:1323
> #6  0x00000000004142c9 in run_builtin (p=0x78a7c8, argc=4, argv=0x7fffffffe630) at perf.c:273
> #7  0x00000000004144cc in handle_internal_command (argc=4, argv=0x7fffffffe630) at perf.c:345
> #8  0x0000000000414618 in run_argv (argcp=0x7fffffffe51c, argv=0x7fffffffe510) at perf.c:389
> #9  0x000000000041489e in main (argc=4, argv=0x7fffffffe630) at perf.c:487
> (gdb) p i
> $1 = 12477348
> (gdb) p *chain
> $2 = {nr = 9007199254740994, ips = 0x7ffff18462d8}
> (gdb) frame 2
> #2  0x000000000042bd17 in perf_top__mmap_read_idx (top=0x7fffffffd8a0, idx=0) at builtin-top.c:858
> 858				perf_event__process_sample(&top->tool, event, evsel,
> (gdb) print event->header
> $3 = {type = 4175878766, misc = 16327, size = 0}
> (gdb) print event->header
> $4 = {type = 4175878766, misc = 16327, size = 0}
> (gdb)
>
> Probably an additional bug :-\
>
> - Arnaldo

Yeah, I think so :(.

The chain->nr has a strang value - it cannot be larger than 255
(PERF_MAX_STACK_DEPTH). Simple fix would be checking ->nr in the
function but it's obviously not fix a root cause.

Anyway, it seems not related to this race problem IMHO.

Thanks,
Namhyung
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/