MISRA 10.3 rule problem

Jernej Skvarc L.

Hello!

I've came accross a piece of code that fails on MISRA 10.3 rule. I'm suspecting a bug in the MISRA checker.

Example of function with MISRA 10.3 violation reported:

void Test1(const int16_t *test_in, int16_t *test_out, int16_t another_in)
{
    *test_out = (int16_t)(((int32_t)(*test_in) * (int32_t)another_in) / (int32_t)32768L); 
}

Example of code without warnings:

void Test2(int16_t test_in, int16_t *test_out, int16_t another_in)
{
    *test_out = (int16_t)(((int32_t)(test_in) * (int32_t)another_in) / (int32_t)32768L);
}

The only difference is, that one variable in a complex expression is passed by pointer. Compiler used is C2000 CGT 6.2.8. Is this a normal behaviour or am I missing something? Same error was reported on compiler version 6.2.1.

Best regards,

Jernej

over 9 years ago

0 Brett Larimore over 9 years ago

TI__Guru 57255 points

Hi Jernej,

I have moved your post to the Compiler forum. I believe that they should be able to help resolve your question.

Thank you,
Brett

0 George Mock over 9 years ago in reply to Brett Larimore

TI__Guru**** 232700 points

Thank you for informing us about this issue. The compiler should handle both of those expressions the same way. Because it doesn't, I filed SDSCM00050815 in the SDOWP system to have this investigated. Feel free to follow it with the SDOWP link below in my signature.

Thanks and regards,

-George

0 Jernej Skvarc L. over 9 years ago in reply to George Mock

Prodigy 40 points

Additionaly, MISRA 10.3 check fails each time variable is passed by a pointer - even when passing structure pointer and using structure->element operator inside complex expression.

Thank you for your support!

Best regards,

Jernej

Code Composer Studio™︎

Code Composer Studio forum

MISRA 10.3 rule problem