[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

FPU problem



On 6 Jul 1999 biguo at 263.net wrote:

> Hello,everyone!In i386/cpu.c,what meaning is the note "The No
> RTEMS_WAIt form of coprocessor instruction MUST be used in case there
> is not a coprocessor to wait for"?According to INTEL processor
> manual,prefix "FN" fpu instruction doesn't cause exception.But it can
> cause exception in running.Why?

That would have likely been based on i386DX and i387 documentation.  You
should check the real user's manual for the CPU model you are using.  It
is possible that this causes an exception all the time for CPU models that
can NEVER support floating point.  

Remember the i386 and m68k do not have architectural definitions that
every CPU model in the family must adhere to.  They have evolved over time
and compatibility is what happened.  

In contrast, the SPARC, PowerPC, i960, and HP PA-RISC has architectural
descriptions and requirements that are independent of any CPU model
implementation. 

Basically, I suspect that the statement for the i386DX does not apply on
this model and the "no wait == FPU exception" is just something that
changed.

--joel
Joel Sherrill, Ph.D.             Director of Research & Development
joel at OARcorp.com                 On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
   Support Available             (256) 722-9985