• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 52 subscribers
  • Views 1786 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.

C6000 7.4.16 clist6x.exe segfault

Andrey Kuznetsov9
Prodigy 45 points

During project compilation one file produces following error:

INTERNAL ERROR: C:\ti\c6000_7.4.16\bin\clist6x.exe experienced a segmentation fault while
                processing file BPI_BPI3.asm line 2254

Error repeatable on 7.4.16 and 7.3.23 versions of CGT.

But same come compiles fine in CSSv4 with 4.9.0 CGT version.

This asm file not written by man it was generated from C source by toolchain itself.

Following error happens always on same line regardless it content, it can be even a comment on some optimizations levels.

I could provide involved asm file if you give me mail or other contact to send it because i'm not authorized to list code publicly.

  • 0
    TI__Guru**** 249900 points

    The clist utility is a helper executable not intended for direct use.  Rather, it is invoked by the compiler shell cl6x.  

    Andrey Kuznetsov9 said:
    This asm file not written by man it was generated from C source by toolchain itself.

    Please preprocess that C source file and send it in.  Note the asm file is not enough to reproduce the problem.  The C source is required.  Also show all the build options exactly as the compiler sees them.

    Andrey Kuznetsov9 said:
    I could provide involved asm file if you give me mail or other contact to send it because i'm not authorized to list code publicly.

    Again, we need the preprocessed C source, not the asm file.  I understand the need to protect the source code.  You can send it to me by private message.  Float your cursor over my avatar.  In the window that pops up, click on Send a private message.  In the message compose interface, click on the paper clip icon to attach the preprocessed file.

    Thanks and regards,

    -George

  • 0 in reply to George Mock
    Prodigy 45 points
    Source was sent via private message
  • 0 in reply to Andrey Kuznetsov9
    TI__Guru**** 249900 points

    Thank you for sending in the test case.  Unfortunately, I am not able to reproduce the error.  That is not too surprising in this case.  The interlist utility clist works with both the original C source file and the just generated assembly file as input.  When I build the preprocessed file, the output assembly file is close to the same, but the C file is quite different.  There must be something about your C file that causes the bug to surface.

    At this point, the best way forward is for you to send in your entire CCS project.  Privately, of course.  Please see this forum post for details on how to have CCS zip up the project.

    Thanks and regards,

    -George

  • 0 in reply to George Mock
    Prodigy 45 points
    Futher investigation lead us to function "get_data_bpi_str" in as source of error. If comment out than function error disappears, but if comment out function body and leave empty function segfault still happens. Also error is position depended and moving this function to end of a file also "resolves" problem.
  • 0 in reply to Andrey Kuznetsov9
    TI__Guru**** 249900 points

    Thank you for sending the test case privately.  I can reproduce the problem while building within CCS 5.5.  I always try to recreate the problem with a command line build, because that makes resolving the problem so much faster.  I am unable to recreate this problem with a command line build.  I can still submit a bug report, but it will probably take much longer to address.  

    In the meantime, here is another thing to consider.  I notice that many of the comments in your source contain non-ASCII characters.  The TI compiler is not able to reliably handle non-ASCII characters.  Please see this thread for more detail.  While I don't know for sure, this could be the cause of your problem.  Are you willing to remove all the non-ASCII characters?

    Thanks and regards,

    -George

  • 0 in reply to George Mock
    TI__Guru**** 249900 points

    I filed SDSCM00052684 in the SDOWP system to have this issue investigated.  You are welcome to follow it with the SDOWP link below in my signature.

    Thanks and regards,

    -George

  • 0 in reply to George Mock
    TI__Guru**** 249900 points

    George Mock said:
    In the meantime, here is another thing to consider.  I notice that many of the comments in your source contain non-ASCII characters.  The TI compiler is not able to reliably handle non-ASCII characters.  Please see this thread for more detail.  While I don't know for sure, this could be the cause of your problem.

    Unfortunately, further investigation has revealed that use of non-ASCII characters is the cause of the problem.  As already established in the thread referenced above, TI compilers do not support non-ASCII characters in any context, including comments.  That being the case, no change will be made to the compiler in response to this report.

    Thanks and regards,

    -George