Sorgelig wrote:You are welcome to provide fix.
I suggest to edit gary.v that way:
Code: Select all
assign t_sel_slow = cpu_address_in[23:19]==5'b1100_0 ? |memory_config[3:2] : 1'b0; //$C00000 - $C7FFFF
assign t_sel_slow = cpu_address_in[23:19]==5'b1100_1 ? memory_config : 1'b0; //$C80000 - $CFFFFF
assign t_sel_slow = cpu_address_in[23:19]==5'b1101_0 ? &memory_config[3:2] : 1'b0; //$D00000 - $D7FFFF
assign sel_reg = cpu_address_in[23:21]==3'b110 ? ~(|t_sel_slow | sel_rtc | sel_ide | sel_gayle) : 1'b0; //chip registers at $DF0000 - $DFFFFF
and remove sel_xram completely.
Just checked that in the simulator - all is fine, no more multiple selects asserted.
PS: but i`am really not understand, how current version of minimig could run without immediate crash due to this bug...
That is strange to me...