LINCAN i82527 bug

  • puker_soh

    puker_soh - 2005-10-17


    I like to submit a bug and patch for i82527.c in LINCAN. The msgobj is from 0 to 14 but in i82527_clear_objects(),

        DEBUGMSG("Cleared all message objects on chip\n");

    -    for (i=1; i<=15; i++) {
    +    for (i=0; i<15; i++) {

    This should stop the segmentation fault

    • Pavel Pisa

      Pavel Pisa - 2005-10-17

      We are aware of this problem already. It has been corrected
      together with  some other mistake in the objects counting
      in i82527 interrupt handler in the CVS. Both problems has been
      introduced during driver queues rework during switch to 0.3 version.

      There is some more changes in the CVS and when they are tested,
      I prepare new release. I have not found time to test i82527 code
      after these changes. Please, send report if CVS code works.
      We really need some feedback even in the positive case,
      to know what is correct and what is broken.

      I have even one other report, that the RTR messages sending
      needs to be modified for i82527 to not block on missing replies, but I have
      no code yet. There should be solution in different objects flags
      setup according to datasheet.

    • puker_soh

      puker_soh - 2005-10-18

      I will report once I get enough information of how well the code works.

    • puker_soh

      puker_soh - 2005-10-19

      Hi again,

      Here's the report.

      sending (/dev/can0, message object 1) and receiving(/dev/can1, message object 2) works fine in both standard and extended mode.

      Great job and thanks.


Log in to post a comment.