From: Milian Wolff <milian.wolff@kdab.com>
To: Jiri Olsa <jolsa@redhat.com>, ahmadkhorrami <ahmadkhorrami@ut.ac.ir>
Cc: Steven Rostedt <rostedt@goodmis.org>,
Arnaldo Carvalho de Melo <acme@redhat.com>,
Linux-trace Users <linux-trace-users@vger.kernel.org>,
Peter Zijlstra <peterz@infradead.org>,
linux-trace-users-owner@vger.kernel.org,
Jin Yao <yao.jin@linux.intel.com>,
Namhyung Kim <namhyung@kernel.org>,
Andi Kleen <ak@linux.intel.com>
Subject: Re: Wrong Perf Backtraces
Date: Tue, 31 Mar 2020 16:44:18 +0200 [thread overview]
Message-ID: <8573002.CDJkKcVGEf@agathebauer> (raw)
In-Reply-To: <08bf52eda6e699ee6b3070c75eac9123@ut.ac.ir>
[-- Attachment #1: Type: text/plain, Size: 1947 bytes --]
On Dienstag, 31. März 2020 15:39:18 CEST ahmadkhorrami wrote:
> But the addresses do not match. Do you confirm this as a bug in
> libdwarf,...?
>
> So I will ignore addresses without a matching symbol. But they do not
> seem reliable!
>
> Could you tell me the name of the library that generates the raw
> addresses, so that I can try to debug it?
This is a platform specific question. There are multiple ways to unwind a
stack. If you are on x86 then by default the .eh_frame section is available
which holds the information necessary for unwinding. It doesn't depend on
debug symbols, that's only used for symbolization and inline-frame resolution
as Jiri indicated.
That said, in the context of perf, there are multiple scenarios that can lead
to broken unwinding:
a) perf record --call-graph dwarf: unwinding can overflow the stack copy
associated with every sample, so the upper end of the stack will be broken
b) perf record --call-graph $any: when you are sampling on a precise event,
such as cycles:P which is the default afaik, then on Intel with PEBS e.g. the
stack copy may be "wrong". See e.g. https://lkml.org/lkml/2018/11/6/257 and
the overall thread. This is not solved yet afaik and after my initial attempt
at workarounding this issue I stopped looking into it and instead opted for
explicitly sampling on the non-precise events when I record call graphs... You
could try that too: do you see the issue when you run e.g.:
`perf record --call-graph dwarf -e cycles`
This should take the non-precise version for sampling but then at least the
call stacks are correct. I.e. you trade the accuracy of the instruction
pointer to which a sample points with reduced call stack breakage.
c) bugs :)
--
Milian Wolff | milian.wolff@kdab.com | Senior Software Engineer
KDAB (Deutschland) GmbH, a KDAB Group company
Tel: +49-30-521325470
KDAB - The Qt, C++ and OpenGL Experts
[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 3826 bytes --]
next prev parent reply other threads:[~2020-03-31 14:44 UTC|newest]
Thread overview: 67+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <157597d74ff17f781d9de7e7e3defd13@ut.ac.ir>
2020-03-22 20:24 ` Wrong Perf Backtraces ahmadkhorrami
2020-03-23 0:34 ` Steven Rostedt
[not found] ` <21b3df4080709f193d62b159887e2a83@ut.ac.ir>
2020-03-23 8:49 ` Jiri Olsa
2020-03-23 10:03 ` ahmadkhorrami
2020-03-25 15:18 ` ahmadkhorrami
2020-03-25 15:46 ` Jiri Olsa
2020-03-25 18:54 ` ahmadkhorrami
2020-03-25 18:58 ` Arnaldo Carvalho de Melo
2020-03-25 19:10 ` ahmadkhorrami
2020-03-25 19:28 ` Arnaldo Carvalho de Melo
2020-03-25 20:01 ` ahmadkhorrami
2020-03-25 20:39 ` Jiri Olsa
2020-03-25 21:02 ` Jiri Olsa
2020-03-25 21:09 ` Steven Rostedt
2020-03-25 21:37 ` ahmadkhorrami
2020-03-25 21:46 ` Jiri Olsa
2020-03-25 22:21 ` ahmadkhorrami
2020-03-25 23:09 ` ahmadkhorrami
2020-03-26 9:59 ` Jiri Olsa
2020-03-26 13:20 ` ahmadkhorrami
2020-03-26 15:39 ` Jiri Olsa
2020-03-26 18:19 ` ahmadkhorrami
2020-03-26 18:21 ` ahmadkhorrami
2020-03-27 9:20 ` Jiri Olsa
2020-03-27 10:59 ` ahmadkhorrami
2020-03-27 11:04 ` ahmadkhorrami
2020-03-27 12:10 ` Milian Wolff
2020-03-27 12:58 ` ahmadkhorrami
2020-03-27 13:25 ` Milian Wolff
2020-03-27 13:33 ` ahmadkhorrami
2020-03-27 18:43 ` ahmadkhorrami
2020-03-27 22:37 ` Jiri Olsa
2020-03-27 23:12 ` ahmadkhorrami
2020-03-28 23:34 ` Jiri Olsa
2020-03-29 0:43 ` ahmadkhorrami
2020-03-29 1:16 ` ahmadkhorrami
2020-03-29 11:19 ` Jiri Olsa
2020-03-29 11:52 ` ahmadkhorrami
2020-03-29 12:08 ` Jiri Olsa
2020-03-29 12:39 ` ahmadkhorrami
2020-03-29 13:50 ` Milian Wolff
2020-03-29 14:23 ` ahmadkhorrami
2020-03-29 19:20 ` Jiri Olsa
2020-03-30 6:09 ` Milian Wolff
2020-03-30 13:07 ` Jiri Olsa
2020-03-30 13:49 ` ahmadkhorrami
2020-03-30 19:05 ` ahmadkhorrami
2020-03-30 21:05 ` debuginfod-based dwarf downloading, was " Frank Ch. Eigler
2020-03-31 9:26 ` Jiri Olsa
2020-03-31 14:00 ` Frank Ch. Eigler
2020-03-31 4:43 ` ahmadkhorrami
2020-03-31 9:30 ` Jiri Olsa
2020-03-31 11:53 ` ahmadkhorrami
2020-03-31 12:43 ` ahmadkhorrami
2020-03-31 13:20 ` Jiri Olsa
2020-03-31 13:39 ` ahmadkhorrami
2020-03-31 14:44 ` Milian Wolff [this message]
2020-03-31 15:02 ` ahmadkhorrami
2020-03-31 15:05 ` ahmadkhorrami
2020-03-31 15:29 ` Milian Wolff
2020-03-31 16:10 ` Arnaldo Carvalho de Melo
2020-03-31 19:20 ` ahmadkhorrami
2020-03-31 19:17 ` ahmadkhorrami
2020-03-31 20:57 ` ahmadkhorrami
2020-04-04 1:01 ` ahmadkhorrami
2020-04-11 16:42 ` ahmadkhorrami
2020-04-11 21:04 ` ahmadkhorrami
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=8573002.CDJkKcVGEf@agathebauer \
--to=milian.wolff@kdab.com \
--cc=acme@redhat.com \
--cc=ahmadkhorrami@ut.ac.ir \
--cc=ak@linux.intel.com \
--cc=jolsa@redhat.com \
--cc=linux-trace-users-owner@vger.kernel.org \
--cc=linux-trace-users@vger.kernel.org \
--cc=namhyung@kernel.org \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=yao.jin@linux.intel.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.