https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108396
Kewen Lin <linkw at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |ASSIGNED
Summary|[12/13 Regression] PPCLE: |[12/13 Regression] PPCLE:
|vec_vsubcuq missing |vec_vsubcuq missing since
| |r12-5752-gd08236359eb229
Assignee|unassigned at gcc dot gnu.org |linkw at gcc dot gnu.org
CC| |linkw at gcc dot gnu.org
--- Comment #2 from Kewen Lin <linkw at gcc dot gnu.org> ---
Yes, it's a typo, which makes the macro definition change to:
#define vec_vsubcuqP __builtin_vec_vsubcuq
Unfortunately we don't have the testing coverage in testsuite for the expected
name vec_vsubcuq (in rs6000-vecdefines.h):
$grep -ir vsubcuq gcc/testsuite/gcc.target/powerpc/
gcc/testsuite/gcc.target/powerpc//p8vector-int128-1.c: return
__builtin_vec_vsubcuq (p, q);
gcc/testsuite/gcc.target/powerpc//p8vector-int128-1.c:/* { dg-final {
scan-assembler "vsubcuq" } } */
gcc/testsuite/gcc.target/powerpc//p8vector-builtin-8.c:/* { dg-final {
scan-assembler-times "vsubcuq" 2 } } */
FWIW, it started to fail from r12-5752-gd08236359eb229 which enables the new
bif framework, though the commit r12-3169 introduced it as Andrew pointed out.
