@@ -3047,7 +3047,7 @@ can_find_related_mode_p (machine_mode vector_mode,
scalar_mode element_mode,
GET_MODE_SIZE (element_mode), nunits))
return true;
if (riscv_v_ext_vls_mode_p (vector_mode)
- && multiple_p (TARGET_MIN_VLEN * TARGET_MAX_LMUL,
+ && multiple_p ((TARGET_MIN_VLEN * TARGET_MAX_LMUL) / 8,
GET_MODE_SIZE (element_mode), nunits))
Just a very minor nit but instead of the / 8 could we not just use
GET_MODE_PRECISION like e.g. vls_mode_valid_p?
The rest looks reasonable to me.
--
Regards
Robin
