'next' command broken on sparc

'next' command broken on sparc

[Dieter Stüken]

I frequently observe the 'next' command not to stop
at then next line. Instead the program continues to the end.
Similar problems are reported some time ago in conjunction
with mips, threads or shlibs.

But I'm working on solaris, using:

GNU gdb 6.3
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "sparc-sun-solaris2.8".

and:

 > gcc -v
Reading specs from /share/unix/solaris/8/pakete/gcc-3.4.3-64/lib/gcc/sparc64-sun-solaris2.8/3.4.3/specs
Configured with: ../gcc-3.4.3/configure --prefix=/share/unix/solaris/8/pakete/gcc-3.4.3-64 sparc64-sun-solaris2.8
Thread model: posix
gcc version 3.4.3

It happens in a bigger project, and I could not narrow it down
to special situations. It seems to happen quite randomly to me.
Thus I don't have any small example code by now.

I have to use shared libs, which also require threads
(although I don't use any threads explicitly)

Is this a known problem? Can I avoid it?

Dieter.
-- 
Dieter Stüken, con terra GmbH, Münster
     stueken@conterra.de
     http://www.conterra.de/
     (0)251-7474-501

Re: 'next' command broken on sparc

[Mark Kettenis]

   Date: Tue, 15 Feb 2005 15:35:16 +0100
   From: =?ISO-8859-1?Q?Dieter_St=FCken?= <stueken@conterra.de>

   I frequently observe the 'next' command not to stop
   at then next line. Instead the program continues to the end.
   Similar problems are reported some time ago in conjunction
   with mips, threads or shlibs.

I doubt that this is related to the MIPS issue.  Unfortunately it's
not easy to establish what the cause is.  I noticed that you use gcc
3.4.3.  It might be that the compiler is optimizing in ways that GDB
looses track.  If you can't produce a smallish example that shows the
problem, can you provide me with the disassembled code from the
function where things go wrong, and the functions that are called by
that function around that point?

Mark

Re: 'next' command broken on sparc

[Dieter Stüken]

Mark Kettenis wrote:
>    Date: Tue, 15 Feb 2005 15:35:16 +0100
>    From: =?ISO-8859-1?Q?Dieter_St=FCken?= <stueken@conterra.de>
> 
>    I frequently observe the 'next' command not to stop
>    at then next line. Instead the program continues to the end.
>    Similar problems are reported some time ago in conjunction
>    with mips, threads or shlibs.
> 
> I doubt that this is related to the MIPS issue.  Unfortunately it's
> not easy to establish what the cause is.  I noticed that you use gcc
> 3.4.3.  It might be that the compiler is optimizing in ways that GDB
> looses track....

The problem is not related to the generated code.
But I build gdb as a 64bit binary.
Building gdb using -m32 it seems to work with gcc 3.4.3.

Let me know, if I can do futher tests to narrow that problem.

Dieter.
-- 
Dieter Stüken, con terra GmbH, Münster
     stueken@conterra.de
     http://www.conterra.de/
     (0)251-7474-501