linux-kernel - Re: [PATCH 6/8] perf tools: Add new comm infrastructure

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: <871u369cq0.fsf@sejong.aot.lge.com>
Date:	Mon, 28 Oct 2013 14:38:31 +0900
From:	Namhyung Kim <namhyung@...nel.org>
To:	David Ahern <dsahern@...il.com>
Cc:	Frederic Weisbecker <fweisbec@...il.com>,
	Arnaldo Carvalho de Melo <acme@...stprotocols.net>,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Paul Mackerras <paulus@...ba.org>,
	Ingo Molnar <mingo@...nel.org>,
	Namhyung Kim <namhyung.kim@....com>,
	LKML <linux-kernel@...r.kernel.org>,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	Jiri Olsa <jolsa@...hat.com>, Ingo Molnar <mingo@...e.hu>,
	Stephane Eranian <eranian@...gle.com>
Subject: Re: [PATCH 6/8] perf tools: Add new comm infrastructure

Hi David,

On Fri, 25 Oct 2013 12:19:07 -0600, David Ahern wrote:
> On 10/25/13 12:12 PM, Frederic Weisbecker wrote:
>> Oh I see. It's possible that my massive conversion to use the comm
>> accessor got blind at some point and left over a few things. I
>> remember that I only lightly tested that new comm infrastructure. I
>> mean I tested a lot of "perf report -s foo,bar" combinations for
>> performance comparisons but I haven't tested the perf script and all
>> the other perf tools.
>>
>> I'll rebase these patches and test that wider before resending.
>
> specifically, I see stuff like perf forking ls and comm still shows as
> perf even though there is COMM record with the rename to ls. I believe
> the test case was something like:
>
> perf sched record -- ls
> perf script

Hmm.. did you try my latest v5 patchset?  I couldn't reproduce the case
at least for the command lines above.

  $ perf script
            perf 24810 [007] 1546517.815809: sched:sched_stat_runtime: comm=perf pid=24810 ru
            perf 24810 [007] 1546517.815909: sched:sched_wakeup: comm=perf pid=24811 prio=120
         swapper     0 [008] 1546517.815913: sched:sched_switch: prev_comm=swapper/8 prev_pid
            perf 24810 [007] 1546517.815953: sched:sched_stat_runtime: comm=perf pid=24810 ru
            perf 24810 [007] 1546517.815957: sched:sched_switch: prev_comm=perf prev_pid=2481
            perf 24811 [008] 1546517.815992: sched:sched_wakeup: comm=migration/8 pid=48 prio
            perf 24811 [008] 1546517.815993: sched:sched_stat_runtime: comm=perf pid=24811 ru
            perf 24811 [008] 1546517.815993: sched:sched_switch: prev_comm=perf prev_pid=2481
     migration/8    48 [008] 1546517.815996: sched:sched_migrate_task: comm=perf pid=24811 pr
     migration/8    48 [008] 1546517.816000: sched:sched_switch: prev_comm=migration/8 prev_p
         swapper     0 [009] 1546517.816002: sched:sched_switch: prev_comm=swapper/9 prev_pid
              ls 24811 [009] 1546517.816808: sched:sched_stat_runtime: comm=ls pid=24811 runt


Here, the process 24811 has only 3 samples before COMM event

  $ perf report -D | grep 24811 | grep -v MMAP | head
  0 0 0x629b0 [0x38]: PERF_RECORD_COMM: perf:24811
  8 1546517815992058 0x65f50 [0x68]: PERF_RECORD_SAMPLE(IP, 1): 24811/24811: 0xffffffff81091512 period: 1 addr: 0
   ... thread: perf:24811
  8 1546517815993189 0x65fb8 [0x70]: PERF_RECORD_SAMPLE(IP, 1): 24811/24811: 0xffffffff81099d25 period: 83314 addr: 0
   ... thread: perf:24811
  8 1546517815993975 0x66028 [0x80]: PERF_RECORD_SAMPLE(IP, 1): 24811/24811: 0xffffffff81659d60 period: 1 addr: 0
   ... thread: perf:24811
  9 1546517816224342 0x66378 [0x38]: PERF_RECORD_COMM: ls:24811
  9 1546517816808637 0x667f0 [0x70]: PERF_RECORD_SAMPLE(IP, 1): 24811/24811: 0xffffffff81099d25 period: 810663 addr: 0
   ... thread: ls:24811

  $ perf version
  perf version 3.11.ge9eb20


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/