This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: altivec: switch to disable VRSAVE generation


> This is okay for the trunk.  *BUT*, the patch you sent also includes
> 
> + 
> + /* Define to TRUE on a big-endian system when structure arguments
> +    passed (and returned) in registers are passed in a little-endian
> +    manner instead of a big-endian manner.  */
> + 
> + #define FUNCTION_ARG_REG_LITTLE_ENDIAN 1
> 
> which is *not* okay until we get a clear definition of what this macro is
> suppose to affect in the GCC backend.

doh!  my bad.  thanks for catching that.  i'll nuke it.

on another note (CCing mark), i got more mail saying that customers would
really like for 3.1 because hardly anyone uses vrsave instructions (i
think only apple does) so that leaves linux, sysv4+eabi, etc with completely
useless instructions in the prologue and epilogue.  is there any way this 
can go into 3.1?

aldy


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]