• State
  • Locked Locked
  • Replies 10 replies
  • Subscribers 30 subscribers
  • Views 539 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.

const at fixed address

Stefano De Blasiis
Intellectual 530 points

I have a Tiva TM4C123AH6 with CCS 5

I need to write some data information at fixed address in flash.

 const char Driver [19]=  "ABCDEF1000V0\030\003\040\021\025\125";

How can I force this string at address  0x0100.1000 ?


Thanks

  • 0
    Guru 26105 points

    Hi,

    If I am not wrong, the address 0x0100.1000 may be inside TI's ROM area, where the boot loader and driverlib resides - if that is the case, you have no chance since this is ROM, not FLASH.

    If you still need to place something at absolute address, then declare a section in source file and place that string/array there, and then in linker file declare the section placed at the wanted absolute address. 

    Petrei

  • 0 in reply to Petrei
    Intellectual 530 points

    Yes sorry, my address example is wrong, the flas address space start from 0x0000.00000

    I need to place a string at absolute address; can you write a example

  • 0 in reply to Stefano De Blasiis
    Guru 26105 points

    Hi,

    First step is to read the compiler manual - look for #pragma for declaring sections and then the linker chapter.

    Search also the forum, including that for compiler, for same subject.

    I will prepare an example later (although I think I already posted one here)

    Petrei

  • 0 in reply to Petrei
    Guru 27665 points

    Another approach here is to use a spare interrupt vector to point to the memory that hold the information.  This frees you from having to predict the required space to reserve for future uses. 

     

    Robert

  • 0 in reply to Robert Adsett
    Guru 47900 points

    Always love it when a, "need is declared" w/out one iota of such, "need's" justification...

  • 0 in reply to cb1
    Intellectual 530 points

    The justification is write in code the sw and data version.

    This code can be read by serial command that read the code at adress xxxyyy.

    The reason to write a version at fixed address is because the external command read only at fixed addresses

     

  • 0 in reply to cb1
    Guru 26105 points

    Hi,

    Add this to your source file:

    #pragma DATA_SECTION(myvar, ".secret")
    const uint8_t myvar[] = "0123456789ABCDEFGHIJ";

    And modify the .cmd file as below (example):


    SECTIONS
    {
    .intvecs: > APP_BASE
    .text : > FLASH
    .const : > FLASH
    .cinit : > FLASH
    .pinit : > FLASH
    .init_array : > FLASH
    .secret: > APP_BASE+0x0001F000

    .vtable : > RAM_BASE
    .data : > SRAM
    .bss : > SRAM
    .sysmem : > SRAM
    .stack : > SRAM
    }

    Petrei

  • 0 in reply to Stefano De Blasiis
    Guru 27665 points

    Stefano De Blasiis said:

    The justification is write in code the sw and data version.

    This code can be read by serial command that read the code at adress xxxyyy.

    The reason to write a version at fixed address is because the external command read only at fixed addresses

     
    That's a design decision that will cause adverse effects throughout your code.
     
    Robert
  • 0 in reply to Robert Adsett
    Intellectual 530 points
    I do not understand what negative effects can occur if an code area has fixed values ​​of the absolute ... ??
    Anyway I've write the code and it seems to work.
     
    Thanks all !!
     
    :-)
  • 0 in reply to Stefano De Blasiis
    Guru 27665 points

    Stefano De Blasiis said:
    Anyway I've write the code and it seems to work

    Very good

    Stefano De Blasiis said:
    I do not understand what negative effects can occur if an code area has fixed values ​​of the absolute

    That has negative effects too.  You have now commited to manually managing memory.  That is at least a self contained issue.

    I was, however, referring to your communications protocol which uses the processor address in its queries (and by extension probably its sets as well).

    Robert