mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Masami Hiramatsu <mhiramat@kernel.org>
To: Daniel Shaulov <daniel.shaulov@gmail.com>
Cc: Arnaldo Carvalho de Melo <arnaldo.melo@gmail.com>,
	Daniel Shaulov <daniel.shaulov@granulate.io>,
	Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@redhat.com>,
	Mark Rutland <mark.rutland@arm.com>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Jiri Olsa <jolsa@redhat.com>, Namhyung Kim <namhyung@kernel.org>,
	Masami Hiramatsu <mhiramat@kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Thomas Richter <tmricht@linux.ibm.com>,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH] perf probe: Add support for DW_OP_call_frame_cfa vars
Date: Fri, 3 Apr 2020 14:39:56 +0900	[thread overview]
Message-ID: <20200403143956.f6d4f7771a55e8fe97afacee@kernel.org> (raw)
In-Reply-To: <20200402202523.GA55594@5f9ad775e7f6>

Hi Daniel,

On Thu, 2 Apr 2020 20:25:24 +0000
Daniel Shaulov <daniel.shaulov@gmail.com> wrote:

> On Thu, Apr 02, 2020 at 03:49:43PM -0300, Arnaldo Carvalho de Melo wrote:
> > Daniel, next time please consider providing precise steps to show what
> > wasn't working and then how the patch solves the problem, i.e. get some
> > go program, show the command lines used to build, probe, etc.
> >
> > - Arnaldo
> 
> Sure thing:
> 
> ---- contents of example.go ----
> package main
> 
> import "fmt"
> 
> func example(param1 int, param2 int) {
> 	fmt.Println(param1, param2)
> }
> 
> func main() {
> 	example(111, 222)
> }
> ---- end contents of example.go ----
> 
> Build with:
> $ go build example.go
> 
> Note that the output file is named after the directory.
> 
> Then try to probe it with:
> $ perf probe -x ./example 'main.example param1 param2'
> 
> The result is:
> > Target program is compiled without optimization. Skipping prologue.
> > Probe on address 0x4820f0 to force probing at the function entry.
> >
> > Sorry, we don't support this variable location yet.
> >   Error: Failed to add events.
> 
> With my changes the event is added successfuly.
> The reason for the change can be observed in the output of:
> $ readelf --debug-dump=info ./example
> 
> Where we have the following for the main.example function (with go >= 1.10):
>  <1><63dd2>: Abbrev Number: 2 (DW_TAG_subprogram)
>     <63dd3>   DW_AT_name        : main.example
>     <63de0>   DW_AT_low_pc      : 0x4820f0
>     <63de8>   DW_AT_high_pc     : 0x4821c2
>     <63df0>   DW_AT_frame_base  : 1 byte block: 9c 	(DW_OP_call_frame_cfa)
>     <63df2>   DW_AT_decl_file   : 0x1
>     <63df6>   DW_AT_external    : 1
>  <2><63df7>: Abbrev Number: 14 (DW_TAG_formal_parameter)
>     <63df8>   DW_AT_name        : param1
>     <63dff>   DW_AT_variable_parameter: 0
>     <63e00>   DW_AT_decl_line   : 5
>     <63e01>   DW_AT_type        : <0x220f>
>     <63e05>   DW_AT_location    : 1 byte block: 9c 	(DW_OP_call_frame_cfa)
>  <2><63e07>: Abbrev Number: 14 (DW_TAG_formal_parameter)
>     <63e08>   DW_AT_name        : param2
>     <63e0f>   DW_AT_variable_parameter: 0
>     <63e10>   DW_AT_decl_line   : 5
>     <63e11>   DW_AT_type        : <0x220f>
>     <63e15>   DW_AT_location    : 2 byte block: 91 8 	(DW_OP_fbreg: 8)
>  <2><63e18>: Abbrev Number: 0
> 
> The go compiler (after 1.10) uses DW_OP_call_frame_cfa for the first parameter and DW_OP_fbreg for the rest.
> On go versions lower than 1.10, it used DW_OP_call_frame_cfa for all parameters.
> Here is example output of the readelf on main.example, compiled with go1.9.7:
>  <1><2634d>: Abbrev Number: 2 (DW_TAG_subprogram)
>     <2634e>   DW_AT_name        : main.example
>     <2635b>   DW_AT_low_pc      : 0x487db0
>     <26363>   DW_AT_high_pc     : 0x487e99
>     <2636b>   DW_AT_frame_base  : 1 byte block: 9c 	(DW_OP_call_frame_cfa)
>     <2636d>   DW_AT_external    : 1
>  <2><2636e>: Abbrev Number: 5 (DW_TAG_formal_parameter)
>     <2636f>   DW_AT_name        : param1
>     <26376>   DW_AT_location    : 1 byte block: 9c 	(DW_OP_call_frame_cfa)
>     <26378>   DW_AT_type        : <0x27f1c>
>  <2><2637c>: Abbrev Number: 5 (DW_TAG_formal_parameter)
>     <2637d>   DW_AT_name        : param2
>     <26384>   DW_AT_location    : 4 byte block: 9c 11 8 22 	(DW_OP_call_frame_cfa; DW_OP_consts: 8; DW_OP_plus)
>     <26389>   DW_AT_type        : <0x27f1c>
>  <2><2638d>: Abbrev Number: 0
> 
> Hope this makes things easier to understand.

Thanks for the explanation. That is a useful info to understand what go compiler does on dwarf.
OK, so that is same as "DW_OP_fbreg: 0".

Thank you,

> 
> Daniel.
> 


-- 
Masami Hiramatsu <mhiramat@kernel.org>

  reply	other threads:[~2020-04-03  5:40 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-04-01 16:19 Daniel Shaulov
2020-04-02 18:49 ` Arnaldo Carvalho de Melo
2020-04-02 20:25   ` Daniel Shaulov
2020-04-03  5:39     ` Masami Hiramatsu [this message]
2020-04-03 13:02     ` Arnaldo Carvalho de Melo
2020-04-03 17:59       ` Daniel Shaulov
2020-04-03  5:43 ` Masami Hiramatsu

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=20200403143956.f6d4f7771a55e8fe97afacee@kernel.org \
    --to=mhiramat@kernel.org \
    --cc=alexander.shishkin@linux.intel.com \
    --cc=arnaldo.melo@gmail.com \
    --cc=daniel.shaulov@gmail.com \
    --cc=daniel.shaulov@granulate.io \
    --cc=gregkh@linuxfoundation.org \
    --cc=jolsa@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mark.rutland@arm.com \
    --cc=mingo@redhat.com \
    --cc=namhyung@kernel.org \
    --cc=peterz@infradead.org \
    --cc=tglx@linutronix.de \
    --cc=tmricht@linux.ibm.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome