Issue in blackfin toolchain

[Joel Sherrill]

Unfortunately, we probably need to move this to either the
binutils or gdb mailing lists.  I am not sure which.  Probably
the gdb mailing list is more likely to produce an answer.

Just to make sure I understand the bfin-elf gdb configuration
includes support for a debug connection to a target that
bfin-rtems does not?  That doesn't seem correct given
that bfin-*-* is the pattern which matches both rtems
and elf bfin targets in gdb/configure.tgt

And the gdb's are built from the same source tree?

Why do you need the bfin-elf toolchain at all?

--joel

On 05/18/2011 10:55 AM, Rohan Kangralkar wrote:
>
>
> On Wed, May 18, 2011 at 11:44 AM, Rohan Kangralkar 
> <rohan.stays at gmail.com <mailto:rohan.stays at gmail.com>> wrote:
>
>     On Thu, May 12, 2011 at 5:43 PM, Joel Sherrill
>     <joel.sherrill at oarcorp.com <mailto:joel.sherrill at oarcorp.com>> wrote:
>
>         On 05/12/2011 04:41 PM, Rohan Kangralkar wrote:
>
>             Hi,
>
>             I am using the rtems blackfin tools to build and debug my
>             board that contains analog devices BF527 processor.
>
>             The issue i am facing is that I am not able to place
>             breakpoints in some part of my c code. So the debugging is
>             becoming difficult. GDB complains  "No line 291 in file"
>             even though the file has more that 300 lines of code.
>
>
>         What's the name of the file?  There is a chance there are two
>         files with the same name in the executable and gdb is confused.
>         I have seen this before.
>
>
>     The name of the file is tll_fpga.c. I looked for the file in my
>     project and there are no files with duplicate file names.
>
>
>             I use the compiler, debugger provided by rtems4.11 for
>             debugging my application.
>
>             After some debugging and analysis of the mapping the line
>             number in the elf to the C code in the file using objdump
>             i find the following.
>             1.  executing the bfin-elf-objdump (  analog devices
>             toolchain) on the elf the following error is thrown out
>
>             BFD: Dwarf Error: mangled line number section.
>
>             2. but the objdump from rtems does not complain.
>
>         Are they built from the same source?  If not, then random
>         differences.
>
>
>
>     They are built from the same source.
>
>
>
>
>
>             So
>             How do i correct the issue in the rtems blackfin gdb?
>             I also observe that analog devices have their own
>             gdb-proxy but rtems does not provide gpd-proxy
>             should the gdb-proxy also be included in the rtems bfin
>             toolchain?
>
>         Would be good.  Yes.
>
>
>
>
>     I have some more information about the problem. To decouple it
>     from the ADI toolchain, I compile my source using
>     bfin-rtems4.11-gcc. Then i load it to bfin-rtems4.11-gdb and
>     connect it to the gdb bfin simulator.
>     Then i try to put a breakpoint in the file by opening then file. 
>     The debugger complains that there is no such line number. Also
>     there is no optimization set while compiling.
>     This is making my debugging difficult since i have to look at the
>     disassembly for debugging.  Is there any way i can use the C
>     source for debugging.
>
>
>
>
>             Thankyou
>             Rohan Kangralkar
>
>
>
>         -- 
>         Joel Sherrill, Ph.D.             Director of Research&
>          Development
>         joel.sherrill at OARcorp.com        On-Line Applications Research
>         Ask me about RTEMS: a free RTOS  Huntsville AL 35805
>           Support Available (256) 722-9985 <tel:%28256%29%20722-9985>
>
>
>
>
>
>     -- 
>     ___________________________
>     Thanks and regards
>     Rohan Kangralkar
>
>
>
>


-- 
Joel Sherrill, Ph.D.             Director of Research&  Development
joel.sherrill at OARcorp.com        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
    Support Available             (256) 722-9985