Re: PPC64 Compiler bug !!

[linas at austin dot ibm dot com]

On Mon, Jun 16, 2003 at 10:05:22PM -0700, Mike Stump wrote:
> On Friday, June 13, 2003, at 10:23 AM, linas@austin.ibm.com wrote:
> > Based on the email replies, and on a few tests run over here,
> > -msoft-float also seems to have the undocumented but valuable side
> > effect of preventing the use of FPR's in DImode moves.
> 
> On processors that have specialized registers for the floating point  
> unit, -msoft-float disappears those registers.  This should be  
> documented and certainly can be relied upon.

Well, that belies the point. I think I can safely state that most
programmers, when faced with an all-integer program, would assume
that the compiler generated pure-integer code, and thus it would 
never occur to them to study or take interest in any float-point
flags.

That's what most of this sqwak is about.

> If you want to fix the documentation to be clearer, that'd be good.

I don't have commit access to the documentation source.  Maybe
something along the following lines would be acceptable?


                                                                                
       -msoft-float
       -mhard-float
           Generate code that does not use (uses) the floating-
           point register set.  Software floating point emulation
           is provided if you use the -msoft-float option, and
           pass the option to GCC when linking.
 
           In some cases, gcc will use floating point registers 
           to temporarily hold 64-bit integer values.  Specifying
           -msoft-float will prevent this optimization from being 
           made.


> > My documentation seems to only mention -msoft-float for the DEC Alpha,
> > and not as a PPC flag.
> 
> Odd.  My documentation seems to have it:

Right. Very odd, it seem that man-db 2.4.1 is broken, cause I can 
clearly see the docos in man 2.3.19. The gcc man page is OK, but
the latest man command can't find the string in the man page.  
Ahh, bugs ... 

--linas