[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

GDB 5.2 Line number out of range error



Hi Aaron:

I checked with addr2line, and it also comes up with the ridiculous line 
number, so I guess the problem is not with gdb :-)

Also something that I find interesting is that it only seems to effect the 
bsp_cleanup routine. Other  addresses for functions work fine, so far it is 
only bsp_cleanup that exhibits this behavior.

When compiling the tests the Makefile is using -g option for debugging.

Today, I am going to roll everything back to RTEMS snapshot 20021118, just 
to make sure I am not going crazy, and verify that this problem has only 
crept in since I have tried upgrading to RTEMS 4.6.0pre1.

Thanks,

Derick

At 11:39 AM 19/03/2003 -0800, Aaron J. Grier wrote:
>On Wed, Mar 19, 2003 at 07:19:21AM +1000, Derick Hammond wrote:
>
> > Hi Aaron:
> >
> > I have done the following to recompiled gcc. Using the
> > m68k-rtems-gcc-3.2.1-newlib-1.11.0.spec file as a basis I compiled my own
> > RPMS:
> >
> >     rtems-base-gcc-gcc3.2.2newlib1.11.0-P1.i386.rpm
> >     m68k-rtems-gcc-gcc3.2.2newlib1.11.0-P1.i386.rpm
> >
> > The only difference in the rpm .spec file is that I use gcc-3.2.2 source
> > unpatched, as opposed to gcc-3.2.1 source patched with
> > gcc-3.2.1-rtems-20021209.diff.
> >
> > Question for the RTEMS Maintainers: Are there any patches required for
> > m68k/Coldfire targets when using gcc-3.2.2? I saw one in the last
> > couple of days, but I thought it was specific to the i386-rtems
> > toolchain.
>
>mmm... I have heard rumors that 3.2.2 should be avoided.  Joel and Ralf
>will know for sure.
>
>I personally use the gcc-3_2-branch via CVS, but I don't know how to
>nicely package that via RPM.
>
> > Anyway, I installed these packages and I am still getting the same
> > "Line number out of range error" when setting a breakpoint on
> > bsp_cleanup.
>
>I'm quickly running out of ideas...   :P
>
> > Question for Chris John: Could this problem be caused by patching
> > gdb-5.2 with the gdb-5.2.1.patch?
>
>does addr2line (built as part of binutils) produce the same rediculous
>line number?  if it looks sane, the problem would appear to be with gdb.
>
> > If you require any more information, please don't hesitate to ask.
>
>what debugging flags are you using when compiling? (IE -g, -gstabs,
>-gdwarf-2, etc...)
>
>--
>   Aaron J. Grier  |   Frye Electronics, Tigard, OR   |  aaron at frye.com