https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113657
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Ever confirmed|0 |1
Last reconfirmed| |2024-01-30
Status|UNCONFIRMED |NEW
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.
8229 simplify_gen_subreg (TImode, operands[1], V8DImode, offset)
(gdb) p debug_rtx(operands[1])
(mem:V8DI (reg/f:DI 0 x0 [101]) [1 *ptr_2(D)+0 S64 A64])
$4 = void
(gdb) p offset
$5 = 0
That is due to:
/* V8DI mode. */
VECTOR_MODE_WITH_PREFIX (V, INT, DI, 8, 5);
ADJUST_ALIGNMENT (V8DI, 8);
So maybe for strict alignment, we need to loop using DImode instead of TImode
and hope ldp/stp pass optimizes it back into load/store pairs ...