A Far Less Ambitous Altivec patch

[Ziemowit Laski]

On 13 Jan, 2004, at 1.31, Hartmut Penner wrote:

> Zem,
>       couldn't you release the changes to the testcases independend
> to the other part of the patch ?

Some of the changes I could, but most of the test cases had to be
cleaned up to conform to fixes in the AltiVec type system.  Why would
it be better to handle these independently?

>
> And, don't change in altivec-10.c
>
> asm volatile ("vor 0,0,0")  to asm volatile ("vor v0,v0,v0") since this
> will
> break on ppc64 linux ...
>
> Maybe change to
> #ifdef __MACH__
> asm volatile ("vor v0,v0,v0");
> #else
> asm volatile ("vor 0,0,0");
> #endif

Ok, I'll do this, although aren't PPC assembler supposed to understand
register names? :-)

--------------------------------------------------------------
Ziemowit Laski                 1 Infinite Loop, MS 301-2K
Mac OS X Compiler Group        Cupertino, CA USA  95014-2083
Apple Computer, Inc.           +1.408.974.6229  Fax .5477