Mirror of the gdb mailing list
 help / color / mirror / Atom feed
From: "Avi Gozlan" <avi@checkpoint.com>
To: "'Daniel Jacobowitz'" <drow@false.org>
Cc: <gdb@sourceware.org>
Subject: RE: symbol files and source files of shared objects when debugging remote target
Date: Thu, 21 Dec 2006 12:41:00 -0000	[thread overview]
Message-ID: <200612211241.kBLCf16B014561@michael.checkpoint.com> (raw)
In-Reply-To: <20061212144510.GA17889@nevyn.them.org>

Thanks. I managed to have the symbols loaded using "solib-search-path".
However I encountered several issues as follows:

1) I succeeded steping into a function defined in one of my shared
libraries. However, when I tried to step into another one I got the
following message:

0x007c5050 in __nss_protocols_lookup () from /lib/tls/libc.so.6

It is worth mentioning the stepping into the second function involved
stepping into another function whose symboles were not loaded. Also it is
maybe worth mentioning that libc.so.6 in the target system is different from
libc.so.6 on the host system.

2) When I run 'set solib-search-path path' on the host, I get the following
warning several times on the target:

gdb: error initializing thread_db library: version mismatch between
libthread_db and libpthread.

By the way, when the program symbol file loads, I get the following message:
Using host libthread_db library "/lib/tls/libthread_db.so.1"
But when running "info share" I don't see libthread_db.so.1 but rather
/lib/tls/libpthread.so.0

I'm using gdb and gdbserver of version 6.5. Both are configured and compiled
for i686-pc-linux-gnu. Both the host and target are Linux 2.4.21

3) Why it is necessary to run 'set solib-search-path' only after the program
enters user entry point? Running it beforehand left the so symbols unloaded.
The implication is that the command cannot be executed during gdb
initialization (e.g. in ini file).

Regards,

Avi

-----Original Message-----
From: Daniel Jacobowitz [mailto:drow@false.org] 
Sent: Tuesday, December 12, 2006 4:45 PM
To: Avi Gozlan
Cc: gdb@sourceware.org
Subject: Re: symbol files and source files of shared objects when debugging
remote target

On Tue, Dec 12, 2006 at 01:36:16PM +0200, Avi Gozlan wrote:
> Hello,
>  
> I'm debugging a remote target using gdb/gdbserver.
>  
> This procedure works fine for the source files included in the program 
> symbol table. However, I cannot step into functions included in a 
> shared library. The 'directory' command does not help in this situation.

Does info shared work?  Did you set "solib-absolute-prefix" or
"solib-search-path" to find the shared libraries on your host system?

--
Daniel Jacobowitz
CodeSourcery




  reply	other threads:[~2006-12-21 12:41 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-12-12 11:37 Avi Gozlan
2006-12-12 14:45 ` Daniel Jacobowitz
2006-12-21 12:41   ` Avi Gozlan [this message]
2006-12-23 21:25     ` 'Daniel Jacobowitz'

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=200612211241.kBLCf16B014561@michael.checkpoint.com \
    --to=avi@checkpoint.com \
    --cc=drow@false.org \
    --cc=gdb@sourceware.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