• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 54 subscribers
  • Views 516 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Behavior change in linker labels

Jerry Trantow
Intellectual 600 points

Previous tools (up to v6.0.2) generate a value for __init_end based on the load address.  v6.1.0 generates a value based on the run addess.  Is this the design intent or is this use undefined when load and run addresses are different?  I can use _cinit_loadstart and _cinit_size to get what I need but I'm wondering if older code is going to fail with the new tools.

.cinit:

{
*(.cinit) /* Autointialization initial values for variables in .bss. */
__init_end=.+1; /* cinit adds one extra 0 at end of cinit records. */
} LOAD = FLASHA, PAGE = 0
RUN = L01SARAM, PAGE = 0
LOAD_START(_cinit_loadstart),
RUN_START(_cinit_runstart),
SIZE(_cinit_size)

  • 0
    TI__Guru**** 253380 points

    The C28x assembly language tools book, in the section titled Assigning the SPC to a Symbol, states

    The . symbol refers to the current run address, not the current load address, of the section.

    Thus, if __init_end were ever assigned the load address, that would be a bug.

    Assignments like this are known to have problems.  In the same book see the section titled Why the Dot Operator Does Not Always Work.  Operators such as LOAD_START, RUN_END, etc. are a superior solution.

    I don't understand why you are setting up the .cinit section like this.  If you link with the option --rom_model (-c for short), then the linker sets up the .cinit section exactly as required by the C compiler.  In particular, the extra 0 word is added at the end.  

    Thanks and regards,

    -George

  • 0 in reply to George Mock
    Intellectual 600 points

    Thank you for your reply.  

    The __init_end label was used for a CRC calculation that spans multiple sections.   __init_end=.+1; doesn't allocate space, it just shifts the label by 1 to match what the CRC routine takes.

    I agree that the LOAD_START(), ... operators are superior solutions.  I will change my projects to use the new operators.