This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] SPARC VIS intrinsics, take 2
Eric Botcazou <ebotcazou@libertysurf.fr> writes:
> > Hi, this is an updated patch for SPARC VIS intrinsics. I believe I have
> > addressed all of Eric's comments.
>
> A few remaining nits:
>
> + tree ptr_ftype_ptr_ptr = build_function_type_list (ptr_type_node,
> + ptr_type_node,
> + ptr_type_node, 0);
>
> + def_builtin ("__builtin_vis_alignaddr", CODE_FOR_alignaddrsi_vis,
> + ptr_ftype_ptr_ptr);
>
> What is the prototype of __builtin_vis_alignaddr exactly? I don't think the
> second operand should be a pointer, but an integer with word mode.
>
Ok, I think what you mean is
if (TARGET_ARCH64)
def_builtin ("__builtin_vis_alignaddr", CODE_FOR_alignaddrdi_vis,
ptr_ftype_ptr_di);
else
def_builtin ("__builtin_vis_alignaddr", CODE_FOR_alignaddrsi_vis,
ptr_ftype_ptr_si);
right?
> +;; Note that (vec_merge:V8QI [(V4QI) (V4QI)] (10101010 = 170) doesn't work
> +;; because vec_merge expects all the operands to be the same type.
>
> ...of the same type?
>
>
> OK with these corrections.
>
> --
> Eric Botcazou
>
--
Thanks,
Jim
http://www.student.cs.uwaterloo.ca/~ja2morri/
http://phython.blogspot.com
http://open.nit.ca/wiki/?page=jim