[Avrora] Problem with C-Print
Ben L. Titzer
ben.titzer at gmail.com
Wed Oct 20 10:50:10 PDT 2010
It is the last vestige of a time before Avrora could disassemble machine
code and had to parse assembly language as text. It should be deleted.
On Wed, Oct 20, 2010 at 8:58 AM, Daniel Minder <daniel.minder at uni-due.de>wrote:
> > I still don't know the exact
> > reason but I know that using the .elf file directly and skipping the
> > conversion with avr-objdump will make it work.
> As you can see from avrora/syntax/objdump/ObjDumpProgramReader.java only
> the .text and .data sections of a program are loaded. The symbol for the
> c-print monitor (debugbuf1) is located in the .bss section and,
> therefore, Avrora will not know it.
> When adding .bss to the loadable sections Avrora will throw an exception
> because the address of the symbol is too high. This is correct since we
> should not _load_ the whole section but only import the symbols. This is
> what the ELF parser is doing (in ELF the symbol table is separate and
> the .bss section is not flagged als loadable). But to achieve this
> behaviour we would need to modify (at least) the ObjDumpReformatter...
> To be honest: I don't see any advantage of the .od format. The plain exe
> is more flexible and we can skip one working step. Is there any reason
> for using the .od format? If not, I would not investigate this issue any
> Daniel Minder
> University of Duisburg-Essen, Networked Embedded Systems
> Bismarckstr. 90, 47057 Duisburg, Germany
> Skype: d.minder
> Avrora mailing list
> Avrora at lists.ucla.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Avrora