This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [draft patch 0/6] Split FYI and some review notes
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: Aleksandar Ristovski <aristovski at qnx dot com>
- Cc: "gdb-patches at sourceware dot org" <gdb-patches at sourceware dot org>
- Date: Fri, 15 Mar 2013 17:22:25 +0100
- Subject: Re: [draft patch 0/6] Split FYI and some review notes
- References: <51278984.3070208@qnx.com> <20130310210734.GA21130@host2.jankratochvil.net> <51434041.4080309@qnx.com>
On Fri, 15 Mar 2013 16:37:37 +0100, Aleksandar Ristovski wrote:
> Thinking about this, we have to do a real name on this as maps/smaps
> contain name as requested initially, not pathname as resolved.
>
> While we can determine entry that corresponds to the given l_ld, the
> name sent to gdb needs to be realname as that is what gdb expects.
Name sent to gdb is from r_debug->link_map->l_name, gdbserver already sends it
that way and everyone seems to be fine with it
/proc/PID/maps name should be IMO really ignored, it has no use for anything
if I have not forgot about something. What do you want to use /proc/PID/maps
name for?
Maybe gdbserver should transform somehow the name from r_debug or use the name
fro /proc/PID/maps intead etc. but that would be completely unrelated new
patch.
Jan