• Locked Locked
  • Replies 1 reply
  • Subscribers 52 subscribers
  • Views 383 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.

cl6x v7.3.5 (and v7.3.1) have Segmentation Fault

David Negro
Intellectual 330 points

Other Parts Discussed in Thread: OMAP-L138

Hi,

I have a very large project (hundreds of source files) that uses make to drive the compilation process.  I build this project for both OMAP-L138 (c674) and the c667x processors.  During the optimized build of the project, the compiler will have a segmentation fault on a couple of files.  I have tried to rebuild just the file in question, or rebuild just the directory in question without issues.  But if I rebuild the entirety of the project I seem to run into the segmentation fault issue.  I am guessing that the error(s) exhibits itself based on the random state of uninitialized memory.  Indeed, the compilation may even run to completion just fine on some computers although the computers I am currently having troubles with are Windows7 64b.

The sources in question are proprietary, so how can I assist to debug this issue?

Here are my command arguments:

cl6x -c  -DNDEBUG -mv6740 -q -rtti -ma -pden -pds383 -pds694 -pds77 -pds177 -pdse230 -pdse1107  -pds179 -pds552 -pds97 -pdse1 -pdse9 -pdse43 -pdse69 -pdse169 -pdse188 -pdse225 -pdse303  --abi=eabi  --opt_level=2  

Note for the c66 build the -mv6740 argument is replaced with -c6600

Thanks,

-David

  • TI__Guru* 84285 points

    Does the SIGSEGV happen in the same file every time?  Does it fail every time you compile the whole project?  Is the opt6x process running near 2GB of memory when it crashes?  Are there other programs running at the same time?  Can you try it in a Windows XP machine?

    In general, the compiler team needs a complete test case that reproduces the problem; if you cannot reproduce the problem with less than the entire project, then that's pretty much what we'd need to see.  There may be ways to reduce the size of the test case, but we may need more information before we can suggest any such ways.