mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Namhyung Kim <namhyung.kim@lge.com>
To: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>,
	Paul Mackerras <paulus@samba.org>, Ingo Molnar <mingo@redhat.com>,
	LKML <linux-kernel@vger.kernel.org>,
	David Ahern <dsahern@gmail.com>,
	Stephane Eranian <eranian@google.com>,
	Jiri Olsa <jolsa@redhat.com>
Subject: Re: [RFC 0/5] perf tools: Minimalistic build without libelf dependency
Date: Mon, 21 May 2012 15:00:55 +0900	[thread overview]
Message-ID: <87mx522huw.fsf@sejong.aot.lge.com> (raw)
In-Reply-To: <1336979424-17206-1-git-send-email-namhyung.kim@lge.com> (Namhyung Kim's message of "Mon, 14 May 2012 16:10:19 +0900")

Hi, all

Any other comments for this?


Thanks,
Namhyung


On Mon, 14 May 2012 16:10:19 +0900, Namhyung Kim wrote:
> NOTE: This is a RFC patchset and it contains a lot of rough edges
> and ugly codes.
>
> Current build procedure of perf tools requires a couple of external
> libraries. Although most of them are optional, elfutils' libelf is
> mandatory for resolving symbols and build-id of the binaries being
> profiled. For some reason, it makes embedded guys' life harder who
> want to build and run perf tools on their boards.
>
> This patch set tries to fix the problem by letting the perf could
> be built without libelf. The initial intent was just providing
> perf record command to gather performance events interested on the
> system and copies the data to development box to be used by perf
> report for analysis.
>
> However it's changed to have most of perf commands as is because
> perf can resolve kernel symbols without libelf and many of perf
> commands deal (mostly) with kernel events - so no need to restrict
> capability of perf tools. Therefore, the end result is not so
> minimalistic in fact and only thing it cannot do is "perf probe".
>
> And then I realized that the perf record needs to know about the
> build-id's anyway :(. So I implemented a poor man's version of elf
> parser only for parsing the build-id info.
>
> To build a minimal perf tools, pass MINIMAL=1 to make. The perf
> report will not display symbol names:
>
>  # Samples: 988  of event 'cycles'
>  # Event count (approx.): 3726465542
>  #
>  # Overhead  Command      Shared Object                     Symbol
>  # ........  .......  .................  .........................
>  #
>      98.40%  noploop  noploop            [.] 0x000000000000066d   
>       1.35%  noploop  ld-2.15.so         [.] 0x000000000000bab1   
>       0.10%  noploop  [kernel.kallsyms]  [k] native_write_msr_safe
>       0.10%  noploop  [kernel.kallsyms]  [k] scheduler_tick       
>       0.03%  noploop  [kernel.kallsyms]  [k] bad_range            
>       0.01%  noploop  [kernel.kallsyms]  [k] pipe_read            
>
> But perf buildid-list can show you the build-id's:
>
>  $ ./perf buildid-list
>  ebfc1bdb98999c64d36ddfe394c58df3de6cbcfb [kernel.kallsyms]
>  52f0c5f25b95075e6fd7682763f7c3a779a4e889 /usr/lib64/ld-2.15.so
>  43a7a2b399b6ee2ff29c6bdadbda6bff88712ed4 /home/namhyung/bin/noploop
>
> So it can be packed using perf archive so copied and analized
> with full-fledged perf tools on a development machine.
>
> This patchset is based on commit cb04ff9ac424 ("sched, perf: Use
> a single callback into the scheduler") and tested on x86_64 and arm.
>
> Any comments are welcome, thanks.
> Namhyung
>
>
> Namhyung Kim (5):
>   perf symbols: Introduce symbol__elf_init()
>   perf symbols: Do not use ELF's symbol binding constants
>   perf tools: Split out util/symbol-elf.c
>   perf tools: Support minimal build
>   perf symbols: Implement poor man's ELF parser
>
>  tools/perf/Makefile                 |   29 ++
>  tools/perf/builtin-buildid-list.c   |    4 +-
>  tools/perf/builtin-top.c            |    5 +-
>  tools/perf/command-list.txt         |    2 +-
>  tools/perf/perf.c                   |    2 +
>  tools/perf/ui/browsers/map.c        |    5 +-
>  tools/perf/util/elf-minimal.h       |  153 ++++++++
>  tools/perf/util/generate-cmdlist.sh |   15 +
>  tools/perf/util/symbol-elf.c        |  691 +++++++++++++++++++++++++++++++++
>  tools/perf/util/symbol-minimal.c    |  258 +++++++++++++
>  tools/perf/util/symbol.c            |  712 ++---------------------------------
>  tools/perf/util/symbol.h            |   20 +
>  12 files changed, 1195 insertions(+), 701 deletions(-)
>  create mode 100644 tools/perf/util/elf-minimal.h
>  create mode 100644 tools/perf/util/symbol-elf.c
>  create mode 100644 tools/perf/util/symbol-minimal.c

  parent reply	other threads:[~2012-05-21  6:03 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-05-14  7:10 Namhyung Kim
2012-05-14  7:10 ` [PATCH 1/5] perf symbols: Introduce symbol__elf_init() Namhyung Kim
2012-05-14  7:10 ` [PATCH 2/5] perf symbols: Do not use ELF's symbol binding constants Namhyung Kim
2012-05-14  7:10 ` [PATCH 3/5] perf tools: Split out util/symbol-elf.c Namhyung Kim
2012-05-21 11:45   ` Jiri Olsa
2012-05-22  1:00     ` Namhyung Kim
2012-05-22  8:35       ` Jiri Olsa
2012-05-14  7:10 ` [PATCH 4/5] perf tools: Support minimal build Namhyung Kim
2012-05-14  7:10 ` [PATCH 5/5] perf symbols: Implement poor man's ELF parser Namhyung Kim
2012-05-21 11:47   ` Jiri Olsa
2012-05-22  1:08     ` Namhyung Kim
2012-05-22  7:52     ` Peter Zijlstra
2012-05-22 10:26       ` Namhyung Kim
2012-05-22 10:44         ` Peter Zijlstra
2012-05-14  9:22 ` [RFC 0/5] perf tools: Minimalistic build without libelf dependency Ingo Molnar
2012-05-15  0:40   ` Namhyung Kim
2012-05-21  6:00 ` Namhyung Kim [this message]
2012-05-21 15:03   ` David Ahern
2012-05-22  0:20     ` Namhyung Kim

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87mx522huw.fsf@sejong.aot.lge.com \
    --to=namhyung.kim@lge.com \
    --cc=a.p.zijlstra@chello.nl \
    --cc=acme@ghostprotocols.net \
    --cc=dsahern@gmail.com \
    --cc=eranian@google.com \
    --cc=jolsa@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=paulus@samba.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome