RES: about SPARC_HAS_FPU = 1

Jiri Gaisler jiri at gaisler.com
Wed Oct 29 09:02:29 UTC 2008



Joel Sherrill wrote:

> Most importantly though, no one demonstrated the situation(s)
> where gcc would  emit fp instructions for integer only code.
> The FPU is supposed to be disabled on integer only tasks so
> a fault would occur.  I don't recall anyone reporting that occurring.

The problem can be reproduced by creating a C function with
more than 6 parameters. gcc will then use FPU registers to
copy the remaining parameters to and from the stack.

> 
> This is technically a concern on any CPU with an FPU for integer
> only tasks.  We wait until it is a demonstrated problem before
> making all tasks FP.

Making all tasks FP is not an ideal solution - I would recommend
to patch the compiler instead.

> At this point, Jiri has proven it is an issue by identifying the
> place in GCC that does it and by fixing it, he has implemented
> the optimal solution from an RTEMS perspective.

Konrad has generated a patch for gcc-3.4.4. We will test it a
bit further and submit it to Joel for checking ....

Jiri.



More information about the users mailing list