public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Arnaldo Carvalho de Melo <acme@infradead.org>
To: Gleb Natapov <gleb@redhat.com>
Cc: linux-kernel@vger.kernel.org, mingo@elte.hu,
	peterz@infradead.org, paulus@samba.org
Subject: [PATCH] Re: perf annotate segfaults when source code has goto label that looks like hex number
Date: Thu, 22 Jul 2010 11:33:45 -0300	[thread overview]
Message-ID: <20100722143345.GC17631@ghostprotocols.net> (raw)
In-Reply-To: <20100722072044.GD27177@redhat.com>

[-- Attachment #1: Type: text/plain, Size: 2972 bytes --]

Em Thu, Jul 22, 2010 at 10:20:44AM +0300, Gleb Natapov escreveu:
> strtoull() returns valid number when it gets line with label and following
> test passes too. I can't think of a way to unambiguously distinguish between
> label and valid rip. May be running objdump with --prefix-addresses will
> help, but it may make other thing unambiguous.

[root@emilia ~]# objdump --start-address=0x0000000000400474 --stop-address=0x0000000000400496 -dS ./a.out | grep -P ':\t'
  400474:	55                   	push   %rbp
  400475:	48 89 e5             	mov    %rsp,%rbp
  400478:	89 7d ec             	mov    %edi,-0x14(%rbp)
  40047b:	48 89 75 e0          	mov    %rsi,-0x20(%rbp)
  40047f:	eb 01                	jmp    400482 <main+0xe>
  400481:	90                   	nop
  400482:	83 45 fc 01          	addl   $0x1,-0x4(%rbp)
  400486:	81 7d fc 80 96 98 00 	cmpl   $0x989680,-0x4(%rbp)
  40048d:	75 f2                	jne    400481 <main+0xd>
  40048f:	90                   	nop
  400490:	b8 00 00 00 00       	mov    $0x0,%eax
  400495:	c9                   	leaveq 
[root@emilia ~]# objdump --start-address=0x0000000000400474
--stop-address=0x0000000000400496 -dS ./a.out | grep ':$'
Disassembly of section .text:
0000000000400474 <main>:
add:
[root@emilia ~]#

Can you try the attached patch?

With it we get:


[root@emilia ~]# perf annotate

------------------------------------------------
 Percent |	Source code & Disassembly of a.out
------------------------------------------------
         :
         :
         :
         :	Disassembly of section .text:
         :
         :	0000000000400474 <main>:
         :	int main(int argc, char **argv)
         :	{
    0.00 :	  400474:       55                      push   %rbp
    0.00 :	  400475:       48 89 e5                mov    %rsp,%rbp
    0.00 :	  400478:       89 7d ec                mov    %edi,-0x14(%rbp)
    0.00 :	  40047b:       48 89 75 e0             mov    %rsi,-0x20(%rbp)
    0.00 :	  40047f:       eb 01                   jmp    400482 <main+0xe>
         :
         :	      while(1) {
         :	              i++;
         :	              if (i == 10000000)
         :	                      goto add;
         :	      }
   21.05 :	  400481:       90                      nop
         :	int main(int argc, char **argv)
         :	{
         :	      int i;
         :
         :	      while(1) {
         :	              i++;
    0.00 :	  400482:       83 45 fc 01             addl   $0x1,-0x4(%rbp)
         :	              if (i == 10000000)
   15.79 :	  400486:       81 7d fc 80 96 98 00    cmpl   $0x989680,-0x4(%rbp)
   63.16 :	  40048d:       75 f2                   jne    400481 <main+0xd>
         :	                      goto add;
    0.00 :	  40048f:       90                      nop
         :	      }
         :	add:
         :	      return 0;
    0.00 :	  400490:       b8 00 00 00 00          mov    $0x0,%eax
         :	}
    0.00 :	  400495:       c9                      leaveq 

[-- Attachment #2: annotate_fix.patch --]
[-- Type: text/plain, Size: 452 bytes --]

diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index 699cf81..e3486d5 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -976,7 +976,7 @@ static int hist_entry__parse_objdump_line(struct hist_entry *self, FILE *file,
 		 * Parse hexa addresses followed by ':'
 		 */
 		line_ip = strtoull(tmp, &tmp2, 16);
-		if (*tmp2 != ':' || tmp == tmp2)
+		if (*tmp2 != ':' || tmp == tmp2 || tmp2[1] == '\0')
 			line_ip = -1;
 	}
 

  reply	other threads:[~2010-07-22 14:34 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-07-22  7:20 perf annotate segfaults when source code has goto label that looks like hex number Gleb Natapov
2010-07-22 14:33 ` Arnaldo Carvalho de Melo [this message]
2010-07-22 16:38   ` [PATCH] " Gleb Natapov
2010-07-22 16:47     ` Arnaldo Carvalho de Melo
2010-07-22 16:52       ` Arnaldo Carvalho de Melo
2010-07-22 17:05         ` [PATCH v2] " Arnaldo Carvalho de Melo
2010-07-22 18:05           ` Gleb Natapov
2010-07-22 19:11             ` Arnaldo Carvalho de Melo
2010-07-22 19:16               ` Gleb Natapov
2010-08-02  9:02               ` Peter Zijlstra
2010-08-02 14:52                 ` Arnaldo Carvalho de Melo
2010-07-23 12:11           ` [tip:perf/urgent] perf annotate: Fix handling of goto labels that are valid hex numbers tip-bot for Arnaldo Carvalho de Melo

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=20100722143345.GC17631@ghostprotocols.net \
    --to=acme@infradead.org \
    --cc=gleb@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=paulus@samba.org \
    --cc=peterz@infradead.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