AnsweredAssumed Answered

Inline assembly - floating point register constraints

Question asked by annOnym on Feb 17, 2012
Latest reply on Feb 27, 2012 by annOnym

If I'm using an inline assembly statement in my C/C++ code such as eg.

 

    asm("%0 = %0 * %1, %1 = %2 - %3;" : "=F"(temp) : "F"(a),"F"(b),"F"(c));

 

I run into problem if variables b and c are not in registers f12 - f15 (the Q4 F registers). The problem is that "F" refers to ALL floating point registers (f0 - f15). I am aware that it is possible to constrain to only the Q4 R registers (r12 - r15) by using "l", but I need the floating point registers. Is there any way to achieve this?

 

Thanks in advance!

Outcomes