module command_parser
    (   input   bit         clock
    ,   input   bit         resetn

    ,   output  bit         uart_ready
    ,   input   bit         uart_valid
    ,   input   bit [7:0]   uart_data

    ,   input   bit         echo_ready
    ,   output  bit         echo_valid
    ,   output  bit [7:0]   echo_data

    ,   input   bit         command_ready
    ,   output  bit         command_valid
    ,   output  bit [23:0]  command_address
    ,   output  bit         command_write
    ,   output  bit [15:0]  command_data
    );

    bit         input_byte_valid;
    bit [7:0]   input_byte;

    enum
        {   READ_ADDRESS_OR_COMMAND
        ,   READ_DATA_1
        ,   READ_DATA_2
        ,   READ_DATA_3
        ,   READ_DATA_4
        } state;

    always @(posedge clock) begin
        if (!resetn) begin
            uart_ready = 0;
            command_valid = 0;
            command_address = 0;
            command_write = 0;
            command_data = 0;
            input_byte_valid = 0;
            input_byte = 0;
            state = state.first;
        end else begin
            if (echo_ready) echo_valid = 0;
            if (command_ready) command_valid = 0;
            if (uart_ready && uart_valid) begin
                echo_valid = 1;
                echo_data = uart_data;
                input_byte_valid = 1;
                input_byte = uart_data;
            end

            if (!command_valid && input_byte_valid) begin
                case (state)
                READ_ADDRESS_OR_COMMAND: begin
                    if (input_byte >= "0" && input_byte <= "9") begin
                        command_address = command_address << 4;
                        command_address[3:0] = input_byte - "0";
                    end else if (input_byte >= "a" && input_byte <= "f") begin
                        command_address = command_address << 4;
                        command_address[3:0] = input_byte - "a" + 10;
                    end else if (input_byte >= "A" && input_byte <= "F") begin
                        command_address = command_address << 4;
                        command_address[3:0] = input_byte - "A" + 10;
                    end else if (input_byte == "?") begin
                        command_valid = 1;
                        command_write = 0;
                        command_data = 0;
                    end else if (input_byte == "=") begin
                        command_write = 1;
                        command_data = 0;
                        state = READ_DATA_1;
                    end
                end
                READ_DATA_1, READ_DATA_2, READ_DATA_3, READ_DATA_4: begin
                    if (input_byte >= "0" && input_byte <= "9") begin
                        command_data = command_data << 4;
                        command_data[3:0] = input_byte - "0";
                        state = state.next;
                    end else if (input_byte >= "a" && input_byte <= "f") begin
                        command_data = command_data << 4;
                        command_data[3:0] = input_byte - "a" + 10;
                        state = state.next;
                    end else if (input_byte >= "A" && input_byte <= "F") begin
                        command_data = command_data << 4;
                        command_data[3:0] = input_byte - "A" + 10;
                        state = state.next;
                    end else begin
                        state = state.first;
                    end
                    command_valid = state == state.first;
                end
                endcase
                input_byte_valid = 0;
            end

            uart_ready = !echo_valid && !input_byte_valid;
        end
    end

endmodule

module echo_arbiter
    (   input   bit         clock
    ,   input   bit         resetn

    ,   output  bit         in0_ready
    ,   input   bit         in0_valid
    ,   input   bit [7:0]   in0_data

    ,   output  bit         in1_ready
    ,   input   bit         in1_valid
    ,   input   bit [7:0]   in1_data

    ,   input   bit         out_ready
    ,   output  bit         out_valid
    ,   output  bit [7:0]   out_data
    );

    bit         in0_hold_valid;
    bit [7:0]   in0_hold;

    bit         in1_hold_valid;
    bit [7:0]   in1_hold;

    always @(posedge clock) begin
        if (!resetn) begin
            in0_ready = 0;
            in1_ready = 0;
            out_valid = 0;
            out_data = 0;
            in0_hold_valid = 0;
            in0_hold = 0;
            in1_hold_valid = 0;
            in1_hold = 0;
        end else begin
            if (out_ready) out_valid = 0;
            if (in0_ready && in0_valid) begin
                in0_hold_valid = 1;
                in0_hold = in0_data;
            end
            if (in1_ready && in1_valid) begin
                in1_hold_valid = 1;
                in1_hold = in1_data;
            end

            if (!out_valid) begin
                if (in0_hold_valid) begin
                    out_valid = 1;
                    out_data = in0_hold;
                    in0_hold_valid = 0;
                end else if (in1_hold_valid) begin
                    out_valid = 1;
                    out_data = in1_hold;
                    in1_hold_valid = 0;
                end
            end

            in0_ready = !in0_hold_valid;
            in1_ready = !in1_hold_valid;
        end
    end

endmodule

module ram_controller
    (   input   bit         clock
    ,   input   bit         resetn

    ,   output  bit         command_ready
    ,   input   bit         command_valid
    ,   input   bit [23:0]  command_address
    ,   input   bit         command_write
    ,   input   bit [15:0]  command_data

    ,   input   bit         result_ready
    ,   output  bit         result_valid
    ,   output  bit [15:0]  result_data

    ,   output  bit         ram_resetn
    ,   output  bit [1:0]   ram_csn
    ,   output  bit         ram_clkp
    ,   output  bit         ram_clkn
    ,   output  bit         ram_rwds_oe
    ,   input   bit         ram_rwds_in
    ,   output  bit         ram_rwds_out
    ,   output  bit         ram_data_oe
    ,   input   bit [7:0]   ram_data_in
    ,   output  bit [7:0]   ram_data_out
    );

    assign ram_clkn = !ram_clkp;

    bit         valid;
    bit [23:0]  address;
    bit         write;
    bit [15:0]  data;

    int reset_counter;

    bit         slow;

    enum
        {   SEND_COMMAND_1
        ,   SEND_COMMAND_2
        ,   SEND_COMMAND_3
        ,   SEND_COMMAND_4
        ,   SEND_COMMAND_5
        ,   SEND_COMMAND_6

        ,   LAT2_CYC1
        ,   LAT2_CYC2
        ,   LAT2_CYC3
        ,   LAT2_CYC4
        ,   LAT2_CYC5
        ,   LAT2_CYC6

        ,   LAT1_CYC1
        ,   LAT1_CYC2
        ,   LAT1_CYC3
        ,   LAT1_CYC4
        // Latency blocks are 6 cycle, but they start in the SEND_COMMAND_5 cycle, so block 1 is only 4 real cycles
        //,   LAT1_CYC5
        //,   LAT1_CYC6

        ,   DATA_1
        ,   DATA_2
        } state;

    always @(posedge clock) begin
        if (!resetn || reset_counter != 0) begin
            command_ready = 0;
            result_valid = 0;
            result_data = 0;
            ram_resetn = 0;
            ram_csn[0] = 1;
            ram_csn[1] = 1;
            ram_clkp = 0;
            ram_rwds_oe = 0;
            ram_rwds_out = 0;
            ram_data_oe = 0;
            ram_data_out = 0;
            valid = 0;
            address = 0;
            write = 0;
            data = 0;
            if (!resetn)
                reset_counter = 5; // Spec wants >= 100ns of reset
            else
                reset_counter = reset_counter - 1;
            slow = 0;
            state = state.first;
        end else begin
            ram_resetn = 1;
            ram_rwds_oe = 0;
            ram_data_oe = 0;
            if (result_ready) result_valid = 0;
            if (command_ready && command_valid) begin
                valid = 1;
                address = command_address;
                write = command_write;
                data = command_data;
                state = state.first;
            end

            if (!valid) begin
                ram_csn[0] = 1;
                ram_csn[1] = 1;
                ram_clkp = 0;
            end else begin
                case (state)
                SEND_COMMAND_1: begin
                    ram_clkp = 0;
                    ram_csn[address[23]] = 0;
                    ram_data_oe = 1;
                    ram_data_out = {!write, 1'b0, 1'b0, 5'b0}; // R/W# AS BURST RESERVED
                    state = state.next;
                end
                SEND_COMMAND_2: begin
                    ram_clkp = 1;
                    ram_data_oe = 1;
                    ram_data_out = {4'b0, address[22:19]}; // RESERVED ROW
                    state = state.next;
                end
                SEND_COMMAND_3: begin
                    ram_clkp = 0;
                    ram_data_oe = 1;
                    ram_data_out = {address[18:11]}; // ROW
                    state = state.next;
                end
                SEND_COMMAND_4: begin
                    ram_clkp = 1;
                    ram_data_oe = 1;
                    ram_data_out = {address[10:9], address[8:3]}; // ROW UPPERCOL
                    slow = ram_rwds_in; // Sort of arbitrarily capturing RWDS in the middle of the command string
                    state = state.next;
                end
                SEND_COMMAND_5: begin
                    ram_clkp = 0;
                    ram_data_oe = 1;
                    ram_data_out = {8'b0}; // RESERVED
                    state = state.next;
                end
                SEND_COMMAND_6: begin
                    ram_clkp = 1;
                    ram_data_oe = 1;
                    ram_data_out = {5'b0, address[2:0]}; // RESERVED LOWERCOL
                    if (slow)
                        state = LAT2_CYC1;
                    else
                        state = LAT1_CYC1;
                end
                LAT2_CYC1, LAT2_CYC2, LAT2_CYC3, LAT2_CYC4, LAT2_CYC5, LAT2_CYC6,
                LAT1_CYC1, LAT1_CYC2, LAT1_CYC3, LAT1_CYC4: begin
                    ram_clkp = !ram_clkp;
                    state = state.next;
                end
                DATA_1, DATA_2: begin
                    ram_clkp = !ram_clkp;
                    if (write) begin
                        ram_rwds_oe = 1;
                        ram_rwds_out = 1;
                        ram_data_oe = 1;
                        ram_data_out = data[7:0];
                        data = data >> 8;
                        state = state.next;
                    end else if (ram_rwds_in) begin
                        data = data << 8;
                        data[7:0] = ram_data_in;
                        state = state.next;
                    end
                    if (state == state.first) begin
                        valid = 0;
                        if (!write) begin
                            result_valid = 1;
                            result_data = data;
                        end
                    end
                end
                endcase
            end

            command_ready = !valid && !result_valid;
        end
    end

endmodule

module result_printer
    (   input   bit         clock
    ,   input   bit         resetn

    ,   output  bit         result_ready
    ,   input   bit         result_valid
    ,   input   bit [15:0]  result_data

    ,   input   bit         echo_ready
    ,   output  bit         echo_valid
    ,   output  bit [7:0]   echo_data
    );

    bit         hold_valid;
    bit [15:0]  hold;

    enum
        {   DIGIT_1
        ,   DIGIT_2
        ,   DIGIT_3
        ,   DIGIT_4
        } state;

    always @(posedge clock) begin
        if (!resetn) begin
            result_ready = 0;
            echo_valid = 0;
            echo_data = 0;
            hold_valid = 0;
            hold = 0;
            state = state.first;
        end else begin
            if (echo_ready) echo_valid = 0;
            if (result_ready && result_valid) begin
                hold_valid = 1;
                hold = result_data;
            end

            if (hold_valid && !echo_valid) begin
                echo_valid = 1;
                echo_data = hold[15:12];
                hold = hold << 4;
                if (echo_data < 10)
                    echo_data = echo_data + "0";
                else
                    echo_data = echo_data + "A" - 10;
                state = state.next;
                if (state == state.first) hold_valid = 0;
            end

            result_ready = !hold_valid;
        end
    end

endmodule

module top
    (   input   bit         clock
    ,   input   bit         resetn

    ,   output  bit         ram_resetn
    ,   output  bit [1:0]   ram_csn
    ,   output  bit         ram_clkp
    ,   output  bit         ram_clkn
    ,   inout   bit         ram_rwds
    ,   inout   bit [7:0]   ram_data
    );

    bit         rx_ready;
    bit         rx_valid;
    bit [7:0]   rx_data;

    bit         tx_ready;
    bit         tx_valid;
    bit [7:0]   tx_data;

    alt_jtag_atlantic
        #(  .INSTANCE_ID(0)
        ,   .LOG2_RXFIFO_DEPTH(6)
        ,   .LOG2_TXFIFO_DEPTH(6)
        ,   .SLD_AUTO_INSTANCE_INDEX("NO")
        ) jtag
        (   .clk(clock)
        ,   .rst_n(resetn)

        ,   .r_dat(tx_data)
        ,   .r_val(tx_valid)
        ,   .r_ena(tx_ready)

        ,   .t_dat(rx_data)
        ,   .t_dav(rx_ready)
        ,   .t_ena(rx_valid)
        );

    bit         echo_in0_ready;
    bit         echo_in0_valid;
    bit [7:0]   echo_in0_data;

    bit         echo_in1_ready;
    bit         echo_in1_valid;
    bit [7:0]   echo_in1_data;

    echo_arbiter arb
        (   .clock(clock)
        ,   .resetn(resetn)

        ,   .in0_ready(echo_in0_ready)
        ,   .in0_valid(echo_in0_valid)
        ,   .in0_data(echo_in0_data)

        ,   .in1_ready(echo_in1_ready)
        ,   .in1_valid(echo_in1_valid)
        ,   .in1_data(echo_in1_data)

        ,   .out_ready(tx_ready)
        ,   .out_valid(tx_valid)
        ,   .out_data(tx_data)
        );

    bit         command_ready;
    bit         command_valid;
    bit [23:0]  command_address;
    bit         command_write;
    bit [15:0]  command_data;

    command_parser parser
        (   .clock(clock)
        ,   .resetn(resetn)

        ,   .uart_ready(rx_ready)
        ,   .uart_valid(rx_valid)
        ,   .uart_data(rx_data)

        ,   .echo_ready(echo_in0_ready)
        ,   .echo_valid(echo_in0_valid)
        ,   .echo_data(echo_in0_data)

        ,   .command_ready(command_ready)
        ,   .command_valid(command_valid)
        ,   .command_address(command_address)
        ,   .command_write(command_write)
        ,   .command_data(command_data)
        );

    bit         result_ready;
    bit         result_valid;
    bit [15:0]  result_data;

    bit         ram_rwds_oe;
    bit         ram_rwds_out;
    assign ram_rwds = ram_rwds_oe ? ram_rwds_out : 'bZ;

    bit         ram_data_oe;
    bit [7:0]   ram_data_out;
    assign ram_data = ram_data_oe ? ram_data_out : 'bZ;

    ram_controller ram
        (   .clock(clock)
        ,   .resetn(resetn)

        ,   .command_ready(command_ready)
        ,   .command_valid(command_valid)
        ,   .command_address(command_address)
        ,   .command_write(command_write)
        ,   .command_data(command_data)

        ,   .result_ready(result_ready)
        ,   .result_valid(result_valid)
        ,   .result_data(result_data)

        ,   .ram_resetn(ram_resetn)
        ,   .ram_csn(ram_csn)
        ,   .ram_clkp(ram_clkp)
        ,   .ram_clkn(ram_clkn)
        ,   .ram_rwds_oe(ram_rwds_oe)
        ,   .ram_rwds_in(ram_rwds)
        ,   .ram_rwds_out(ram_rwds_out)
        ,   .ram_data_oe(ram_data_oe)
        ,   .ram_data_in(ram_data)
        ,   .ram_data_out(ram_data_out)
        );

    result_printer print
        (   .clock(clock)
        ,   .resetn(resetn)

        ,   .result_ready(result_ready)
        ,   .result_valid(result_valid)
        ,   .result_data(result_data)

        ,   .echo_ready(echo_in1_ready)
        ,   .echo_valid(echo_in1_valid)
        ,   .echo_data(echo_in1_data)
        );

endmodule