On 10/24/21 23:01, Richard Henderson wrote: > On 10/23/21 2:47 PM, Philippe Mathieu-Daudé wrote: >> +#define TRANS_DF_W64(NAME, trans_func, gen_func) \ >> + TRANS_CHECK(NAME, check_msa_access(ctx), trans_func, \ >> + DF_HALF, DF_WORD, \ >> + gen_func##_b, gen_func##_h, gen_func##_w, NULL) > > Again with the table. > >> +static bool trans_msa_elm_d64(DisasContext *ctx, arg_msa_elm *a, >> + enum CPUMIPSMSADataFormat df_max32, >> + enum CPUMIPSMSADataFormat df_max64, > > I think you should just arrange for the entries in the table to be null > when the element variant is not defined.
Clever, thank you!