module modexpng_part_recombinator
(
    clk,
    rdy,
    fsm_state_next,
    index_last,
    dsp_x_ce_p, dsp_y_ce_p,
    ena_x,   ena_y,
    dsp_x_p, dsp_y_p,
    col_index, col_index_last, slim_bram_xy_addr,
    fat_bram_xy_bank, fat_bram_xy_addr, fat_bram_x_dout, fat_bram_y_dout, fat_bram_xy_dout_valid
);


    //
    // Headers
    //
    `include "../rtl/modexpng_mmm_fsm.vh"
    `include "../rtl/modexpng_parameters.vh"
    `include "../rtl/modexpng_parameters_x8.vh"


    input                        clk;
    output                       rdy;
    input  [FSM_STATE_WIDTH-1:0] fsm_state_next;
    input [7:0]                  index_last;
    input                        dsp_x_ce_p;
    input                        dsp_y_ce_p;
    input                        ena_x;
    input                        ena_y;
    input  [8*47-1:0] dsp_x_p;
    input  [8*47-1:0] dsp_y_p;
    input  [     4:0] col_index;
    input  [     4:0] col_index_last;
    input  [     7:0] slim_bram_xy_addr;

    output [     2:0] fat_bram_xy_bank;
    output [     7:0] fat_bram_xy_addr;
    output [    17:0] fat_bram_x_dout;
    output [    17:0] fat_bram_y_dout;
    output            fat_bram_xy_dout_valid;


    //
    // Latches
    //
    reg  [1*47-1:0] dsp_x_p_latch[0:7];
    reg  [1*47-1:0] dsp_y_p_latch[0:7];


    //
    // Mapping
    //
    wire [46:0] dsp_x_p_split[0:7];
    wire [46:0] dsp_y_p_split[0:7];
    
    genvar z;
    generate for (z=0; z<NUM_MULTS; z=z+1)
        begin : gen_dsp_xy_p_split
            assign dsp_x_p_split[z] = dsp_x_p[47*z+:47];
            assign dsp_y_p_split[z] = dsp_y_p[47*z+:47];
        end
    endgenerate


    //
    // Delays
    //
    reg dsp_y_ce_p_dly1 = 1'b0;
    reg dsp_x_ce_p_dly1 = 1'b0;

    always @(posedge clk) begin
        //
        {dsp_y_ce_p_dly1, dsp_x_ce_p_dly1} <= {dsp_y_ce_p,      dsp_x_ce_p};
        //
    end


    //
    // Registers
    //
    
    // valid
    reg       x_valid_lsb = 1'b0;
    reg       y_valid_lsb = 1'b0;
    reg       x_valid_msb = 1'b0;
    reg       y_valid_msb = 1'b0;
    
    // bitmap
    reg [7:0] x_bitmap_lsb = {8{1'b0}};
    reg [7:0] y_bitmap_lsb = {8{1'b0}};
    reg [7:0] x_bitmap_msb = {8{1'b0}};
    reg [7:0] y_bitmap_msb = {8{1'b0}};
    
    // index
    reg [2:0] x_index_lsb = 3'dX;
    reg [2:0] y_index_lsb = 3'dX;
    
    // purge
    reg       x_purge_lsb = 1'b0;
    reg       y_purge_lsb = 1'b0;
    reg       x_purge_msb = 1'b0;
    reg       y_purge_msb = 1'b0;
    
    // valid - latch
    reg       x_valid_latch_lsb = 1'b0;
    reg       y_valid_latch_lsb = 1'b0;
    
    // bitmap - latch
    reg [7:0] x_bitmap_latch_lsb = {8{1'b0}};
    reg [7:0] y_bitmap_latch_lsb = {8{1'b0}};
    reg [7:0] x_bitmap_latch_msb = {8{1'b0}};
    reg [7:0] y_bitmap_latch_msb = {8{1'b0}};

    // index - latch
    reg [2:0] x_index_latch_lsb = 3'dX;
    reg [2:0] y_index_latch_lsb = 3'dX;
    
    // purge - index
    reg       x_purge_latch_lsb = 1'b0;
    reg       y_purge_latch_lsb = 1'b0;
    reg       x_purge_latch_msb = 1'b0;
    reg       y_purge_latch_msb = 1'b0;

    // 
    reg       xy_valid_lsb_adv[1:6];
    reg       xy_valid_msb_adv[1:6];
    reg [7:0] xy_bitmap_lsb_adv[1:6];
    reg [7:0] xy_bitmap_msb_adv[1:6];
    reg [2:0] xy_index_lsb_adv[1:6];
    reg [2:0] xy_index_msb_adv[1:6];
    reg       xy_purge_lsb_adv[1:6];
    reg       xy_purge_msb_adv[1:6];
    
    
    integer i;
    initial for (i=1; i<6; i=i+1) begin
        xy_valid_lsb_adv[i] = 1'b0;
        xy_valid_msb_adv[i] = 1'b0;
        xy_bitmap_lsb_adv[i] = {8{1'b0}};
        xy_bitmap_msb_adv[i] = {8{1'b0}};
        xy_index_lsb_adv[i] = 3'dX;
        xy_index_msb_adv[i] = 3'dX;
        xy_purge_lsb_adv[i] = 1'b0;
        xy_purge_msb_adv[i] = 1'b0;
    end
    
    function  [0:0] calc_square_valid_lsb;
        input [4:0] col_index_value;
        input [4:0] col_index_last_value;
        input [7:0] slim_bram_xy_addr_value;
        begin
            //
            if (slim_bram_xy_addr_value[7:3] == col_index_value)
                calc_square_valid_lsb = 1'b1;
            else
                calc_square_valid_lsb = 1'b0;
            //
        end
    endfunction
    
    function  [7:0] calc_square_bitmap_lsb;
        input [4:0] col_index_value;
        input [4:0] col_index_last_value;
        input [7:0] slim_bram_xy_addr_value;
        begin
            //
            if (slim_bram_xy_addr_value[7:3] == col_index_value)
                //
                case (slim_bram_xy_addr_value[2:0])
                    3'b000: calc_square_bitmap_lsb = 8'b00000001;
                    3'b001: calc_square_bitmap_lsb = 8'b00000010;
                    3'b010: calc_square_bitmap_lsb = 8'b00000100;
                    3'b011: calc_square_bitmap_lsb = 8'b00001000;
                    3'b100: calc_square_bitmap_lsb = 8'b00010000;
                    3'b101: calc_square_bitmap_lsb = 8'b00100000;
                    3'b110: calc_square_bitmap_lsb = 8'b01000000;
                    3'b111: calc_square_bitmap_lsb = 8'b10000000;
                endcase
                //
            else
                calc_square_bitmap_lsb = {8{1'b0}};
            //
        end
    endfunction

    function  [2:0] calc_square_index_lsb;
        input [4:0] col_index_value;
        input [4:0] col_index_last_value;
        input [7:0] slim_bram_xy_addr_value;
        begin
            //
            if (slim_bram_xy_addr_value[7:3] == col_index_value)
                //
                case (slim_bram_xy_addr_value[2:0])
                    3'b000: calc_square_index_lsb = 3'd0;
                    3'b001: calc_square_index_lsb = 3'd1;
                    3'b010: calc_square_index_lsb = 3'd2;
                    3'b011: calc_square_index_lsb = 3'd3;
                    3'b100: calc_square_index_lsb = 3'd4;
                    3'b101: calc_square_index_lsb = 3'd5;
                    3'b110: calc_square_index_lsb = 3'd6;
                    3'b111: calc_square_index_lsb = 3'd7;
                endcase
                //
            else
                calc_square_index_lsb = 3'dX;
            //
        end
    endfunction
    
    function        calc_square_purge_lsb;
        input [4:0] col_index_value;
        input [4:0] col_index_last_value;
        input [7:0] slim_bram_xy_addr_value;
        begin
            //
            if (slim_bram_xy_addr_value[7:3] == col_index_value)
                calc_square_purge_lsb = slim_bram_xy_addr_value[7:3] == col_index_last_value;
            else
                calc_square_purge_lsb = 1'b0;
            //
        end
    endfunction

    function        calc_square_valid_msb;
        input [4:0] col_index_value;
        input [4:0] col_index_last_value;
        input [7:0] slim_bram_xy_addr_value;
        input [7:0] index_last_value;
        begin
            //
            if (slim_bram_xy_addr_value == index_last_value)
                calc_square_valid_msb = 1'b1;
            else
                calc_square_valid_msb = 1'b0;
            //
        end
    endfunction
    
    function  [7:0] calc_square_bitmap_msb;
        input [4:0] col_index_value;
        input [4:0] col_index_last_value;
        input [7:0] slim_bram_xy_addr_value;
        input [7:0] index_last_value;
        begin
            //
            if (slim_bram_xy_addr_value == index_last_value) begin
                calc_square_bitmap_msb[7] = col_index_value != col_index_last_value;
                calc_square_bitmap_msb[6:0] = 7'b1111111;
            end else
                calc_square_bitmap_msb[7:0] = 8'b00000000;
            //
        end
    endfunction

    function        calc_square_purge_msb;
        input [4:0] col_index_value;
        input [4:0] col_index_last_value;
        input [7:0] slim_bram_xy_addr_value;
        input [7:0] index_last_value;
        begin
            //
            if (slim_bram_xy_addr_value == index_last_value)
                calc_square_purge_msb = col_index_value == col_index_last_value;
            else
                calc_square_purge_msb = 1'b0;
            //
        end
    endfunction

    
    reg         recomb_lsb_ce = 1'b0;
    reg  [ 2:0] recomb_lsb_ce_purge = 3'b000;
    wire        recomb_lsb_ce_combined = recomb_lsb_ce | recomb_lsb_ce_purge[0];
    reg         recomb_lsb_clr;
    reg         recomb_lsb_vld = 1'b0;

    reg  [46:0] recomb_lsb_din;
    wire [15:0] recomb_lsb_dout;

    reg         recomb_msb_ce = 1'b0;
    reg  [ 1:0] recomb_msb_ce_purge = 2'b00;
    wire        recomb_msb_ce_combined = recomb_msb_ce | recomb_msb_ce_purge[0];
    reg         recomb_msb_clr;
    reg         recomb_msb_vld = 1'b0;
    
    always @(posedge clk)
        //
        {recomb_msb_vld, recomb_lsb_vld} <= {recomb_msb_ce_combined, recomb_lsb_ce_combined};

    reg  [46:0] recomb_msb_din;
    wire [15:0] recomb_msb_dout;
    
    modexpng_recombinator_block recomb_x_lsb
    (
        .clk    (clk),
        .ce     (recomb_lsb_ce_combined),
        .clr    (recomb_lsb_clr),
        .din    (recomb_lsb_din),
        .dout   (recomb_lsb_dout)
    );

    modexpng_recombinator_block recomb_x_msb
    (
        .clk    (clk),
        .ce     (recomb_msb_ce_combined),
        .clr    (recomb_msb_clr),
        .din    (recomb_msb_din),
        .dout   (recomb_msb_dout)
    );

    always @(posedge clk) begin
        //
        recomb_lsb_ce <= x_valid_latch_lsb;
        recomb_msb_ce <= x_bitmap_latch_msb[0];
        //
        if (x_purge_latch_lsb)
            recomb_lsb_ce_purge <= 3'b111;
        else
            recomb_lsb_ce_purge <= {1'b0, recomb_lsb_ce_purge[2:1]};
        //
        if (x_purge_latch_msb && x_bitmap_latch_msb[0] && !x_bitmap_latch_msb[1])
            recomb_msb_ce_purge = 2'b11;
        else
            recomb_msb_ce_purge <= {1'b0, recomb_msb_ce_purge[1]};
        //
    end


    always @(posedge clk)
        //
        if (ena_x & ena_y) begin
            recomb_lsb_clr <= 1'b1;
            recomb_msb_clr <= 1'b1;
        end else begin
            if (recomb_lsb_ce) recomb_lsb_clr <= 1'b0;
            if (recomb_msb_ce) recomb_msb_clr <= 1'b0;
        end

    always @(posedge clk)
        //
        if (x_valid_latch_lsb)
            recomb_lsb_din <= dsp_x_p_latch[x_index_latch_lsb];
        else
            recomb_lsb_din <= {47{1'b0}};

    always @(posedge clk)
        //
        if (x_bitmap_latch_msb[0])
            recomb_msb_din <= dsp_x_p_latch[0];
        else
            recomb_msb_din <= {47{1'b0}};


    always @(posedge clk)
        //
        case (fsm_state_next)
            //
            FSM_STATE_MULT_SQUARE_COL_0_TRIG,
            FSM_STATE_MULT_SQUARE_COL_N_TRIG,
            FSM_STATE_MULT_SQUARE_COL_0_BUSY,
            FSM_STATE_MULT_SQUARE_COL_N_BUSY: begin
                //
                xy_valid_lsb_adv [6] <= calc_square_valid_lsb (col_index, col_index_last, slim_bram_xy_addr);
                xy_bitmap_lsb_adv[6] <= calc_square_bitmap_lsb(col_index, col_index_last, slim_bram_xy_addr);
                xy_index_lsb_adv [6] <= calc_square_index_lsb (col_index, col_index_last, slim_bram_xy_addr);
                xy_purge_lsb_adv [6] <= calc_square_purge_lsb (col_index, col_index_last, slim_bram_xy_addr);
                //
                xy_valid_msb_adv [6] <= calc_square_valid_msb (col_index, col_index_last, slim_bram_xy_addr, index_last);
                xy_bitmap_msb_adv[6] <= calc_square_bitmap_msb(col_index, col_index_last, slim_bram_xy_addr, index_last);
                xy_purge_msb_adv [6] <= calc_square_purge_msb (col_index, col_index_last, slim_bram_xy_addr, index_last);
                //
            end
            //
            default: begin
                //
                xy_valid_lsb_adv [6] <= 1'b0;
                xy_bitmap_lsb_adv[6] <= {8{1'b0}};
                xy_index_lsb_adv [6] <= 3'dX;
                xy_purge_lsb_adv [6] <= 1'b0;
                //
                xy_valid_msb_adv [6] <= 1'b0;
                xy_bitmap_msb_adv[6] <= {8{1'b0}};
                xy_purge_msb_adv [6] <= 1'b0;
                //
            end
            //
        endcase


    always @(posedge clk) begin
        //
        {y_valid_lsb,  x_valid_lsb}  <= {2{xy_valid_lsb_adv [1]}};
        {y_bitmap_lsb, x_bitmap_lsb} <= {2{xy_bitmap_lsb_adv[1]}};
        {y_index_lsb,  x_index_lsb}  <= {2{xy_index_lsb_adv [1]}};
        {y_purge_lsb,  x_purge_lsb}  <= {2{xy_purge_lsb_adv [1]}};
        //
        {y_valid_latch_lsb,  x_valid_latch_lsb}  <= {y_valid_lsb,  x_valid_lsb};
        {y_bitmap_latch_lsb, x_bitmap_latch_lsb} <= {y_bitmap_lsb, x_bitmap_lsb};
        {y_index_latch_lsb,  x_index_latch_lsb}  <= {y_index_lsb,  x_index_lsb};
        {y_purge_latch_lsb,  x_purge_latch_lsb}  <= {y_purge_lsb,  x_purge_lsb};
        //
        {y_valid_msb,  x_valid_msb}  <= {2{xy_valid_msb_adv[1]}};
        {y_bitmap_msb, x_bitmap_msb} <= {2{xy_bitmap_msb_adv[1]}};
        {y_purge_msb,  x_purge_msb}  <= {2{xy_purge_msb_adv[1]}};
        //
        if (x_valid_msb) begin
            x_bitmap_latch_msb <= x_bitmap_msb;
            x_purge_latch_msb  <= x_purge_msb;
        end else begin
            x_bitmap_latch_msb <= {1'b0, x_bitmap_latch_msb[7:1]};
        end
        //
        //
        for (i=1; i<6; i=i+1) begin
            xy_valid_lsb_adv [i] <= xy_valid_lsb_adv [i+1];
            xy_bitmap_lsb_adv[i] <= xy_bitmap_lsb_adv[i+1];
            xy_index_lsb_adv [i] <= xy_index_lsb_adv [i+1];
            xy_purge_lsb_adv [i] <= xy_purge_lsb_adv [i+1];
            //
            xy_valid_msb_adv [i] <= xy_valid_msb_adv [i+1];
            xy_bitmap_msb_adv[i] <= xy_bitmap_msb_adv[i+1];
            xy_purge_msb_adv [i] <= xy_purge_msb_adv [i+1];
        end
        //
    end

    always @(posedge clk)
        //
        if (x_bitmap_latch_msb[1])   // only shift 7 times
            //
            for (i=0; i<8; i=i+1)            
                if (i < 7)
                    dsp_x_p_latch[i] <= dsp_x_p_latch[i+1];
                else
                    dsp_x_p_latch[i] <= {47{1'bX}};
            //
        else if (dsp_x_ce_p_dly1)
            //
            for (i=0; i<8; i=i+1)
                //
                if (x_bitmap_lsb[i])
                    dsp_x_p_latch[i] <= dsp_x_p_split[i];
                else if (x_valid_msb && x_bitmap_msb[i])
                    dsp_x_p_latch[i] <= dsp_x_p_split[i];

    reg recomb_x_lsb_dout_valid = 1'b0;
    reg recomb_x_msb_dout_valid = 1'b0;

    always @(posedge clk) begin
        recomb_x_lsb_dout_valid <= recomb_lsb_ce_combined;
        recomb_x_msb_dout_valid <= recomb_msb_ce_combined;
    end
        


    reg [ 2:0] fat_bram_xy_bank_reg;
    reg [ 7:0] fat_bram_xy_addr_reg;
    reg [ 7:0] fat_bram_xy_cnt_lsb;
    reg [ 7:0] fat_bram_xy_cnt_msb;
    reg [17:0] fat_bram_x_dout_reg;
    reg [17:0] fat_bram_y_dout_reg;
    reg        fat_bram_xy_dout_valid_reg = 1'b0;

    reg [15:0] recomb_msb_dout_carry_0;
    reg [15:0] recomb_msb_dout_carry_1;
    
    reg [15:0] recomb_msb_dout_delay_0;
    reg [15:0] recomb_msb_dout_delay_1;
    reg [15:0] recomb_msb_dout_delay_2;
    
    reg [ 7:0] recomb_msb_cnt_delay_0 = 8'd0;
    reg [ 7:0] recomb_msb_cnt_delay_1 = 8'd0;
    reg [ 7:0] recomb_msb_cnt_delay_2 = 8'd0;

    assign fat_bram_xy_bank       = fat_bram_xy_bank_reg;
    assign fat_bram_xy_addr       = fat_bram_xy_addr_reg;
    assign fat_bram_x_dout        = fat_bram_x_dout_reg;
    assign fat_bram_y_dout        = fat_bram_y_dout_reg;
    assign fat_bram_xy_dout_valid = fat_bram_xy_dout_valid_reg;
    
    reg rdy_reg = 1'b1;
    reg rdy_adv = 1'b1;
    
    assign rdy = rdy_reg;
    
    
    always @(posedge clk)
        //
        if (ena_x & ena_y)
            rdy_reg <= 1'b0;
        else
            rdy_reg <= rdy_adv;

    always @(posedge clk)
        //
        if (ena_x & ena_y) begin
            rdy_adv <= 1'b0;
            fat_bram_xy_cnt_lsb <= 8'd0;
            fat_bram_xy_cnt_msb <= 8'd0;
        end else begin
            //
            case ({recomb_x_msb_dout_valid, recomb_x_lsb_dout_valid})
                //
                2'b00: begin
                    //
                    if (recomb_msb_cnt_delay_2 > 8'd0) begin
                        //
                        rdy_adv <= recomb_msb_cnt_delay_1 == 8'd0;
                        //
                        recomb_msb_dout_delay_0 <= {18{1'bX}};
                        recomb_msb_dout_delay_1 <= recomb_msb_dout_delay_0;
                        recomb_msb_dout_delay_2 <= recomb_msb_dout_delay_1;
                        //
                        recomb_msb_cnt_delay_0 <= 8'd0;
                        recomb_msb_cnt_delay_1 <= recomb_msb_cnt_delay_0;
                        recomb_msb_cnt_delay_2 <= recomb_msb_cnt_delay_1;
                        //
                        fat_bram_xy_bank_reg        <= BANK_FAT_ABH;
                        fat_bram_xy_addr_reg        <= recomb_msb_cnt_delay_2;
                        fat_bram_x_dout_reg         <= recomb_msb_dout_delay_2;
//                      fat_bram_y_dout_reg         <= {18{1'bX}};
                        fat_bram_xy_dout_valid_reg  <= 1'b1;
                        //
                    end else begin
                        //
                        fat_bram_xy_bank_reg        <= 3'bXXX;
                        fat_bram_xy_addr_reg        <= 8'hXX;
                        fat_bram_x_dout_reg         <= {18{1'bX}};
                        fat_bram_y_dout_reg         <= {18{1'bX}};
                        fat_bram_xy_dout_valid_reg  <= 1'b0;
                        //
                    end
                    //
                end
                //
                2'b01: begin
                    //
                    fat_bram_xy_bank_reg        <= BANK_FAT_ABL;
                    fat_bram_xy_addr_reg        <= fat_bram_xy_cnt_lsb;
                    fat_bram_x_dout_reg         <= {2'b00, recomb_lsb_dout};
//                  fat_bram_y_dout_reg
                    fat_bram_xy_dout_valid_reg  <= 1'b1;
                    //
                    fat_bram_xy_cnt_lsb         <= fat_bram_xy_cnt_lsb + 1'b1; 
                    //
                end
                //
                2'b10: begin
                    //
                    if (fat_bram_xy_cnt_msb < 8'd2) begin
                        //
                        recomb_msb_dout_carry_0 <= recomb_msb_dout;
                        recomb_msb_dout_carry_1 <= recomb_msb_dout_carry_0;
                        //
                        fat_bram_xy_bank_reg        <= 3'bXXX;
                        fat_bram_xy_addr_reg        <= 8'hXX;
                        fat_bram_x_dout_reg         <= {18{1'bX}};
    //                  fat_bram_y_dout_reg
                        fat_bram_xy_dout_valid_reg  <= 1'b0;
                        //                        
                    end else begin
                        //
                        fat_bram_xy_bank_reg        <= BANK_FAT_ABH;
                        fat_bram_xy_addr_reg        <= fat_bram_xy_cnt_msb;
                        fat_bram_x_dout_reg         <= {2'b00, recomb_msb_dout};
    //                  fat_bram_y_dout_reg
                        fat_bram_xy_dout_valid_reg  <= 1'b1;                        
                        //
                    end
                    //
                    fat_bram_xy_cnt_msb         <= fat_bram_xy_cnt_msb + 1'b1;
                    //                
                end
                //
                2'b11: begin
                    //
                    if (fat_bram_xy_cnt_lsb == index_last) begin
                        //
                        fat_bram_xy_bank_reg        <= BANK_FAT_ABL;
                        fat_bram_xy_addr_reg        <= fat_bram_xy_cnt_lsb;
                        fat_bram_x_dout_reg         <= {2'b00, recomb_lsb_dout};
//                      fat_bram_y_dout_reg         <= {18{1'bX}};
                        fat_bram_xy_dout_valid_reg  <= 1'b1;
                        //
                        fat_bram_xy_cnt_lsb         <= 8'd0;
                        //
                    end else begin
                        //
                        fat_bram_xy_bank_reg        <= BANK_FAT_ABH;
                        fat_bram_xy_addr_reg        <= fat_bram_xy_cnt_lsb;
                        fat_bram_x_dout_reg         <= {1'b0, {1'b0, recomb_lsb_dout} + {1'b0, recomb_msb_dout_carry_1}};
//                      fat_bram_y_dout_reg         <= {18{1'bX}};
                        fat_bram_xy_dout_valid_reg  <= 1'b1;
                        //
                        fat_bram_xy_cnt_lsb         <= fat_bram_xy_cnt_lsb + 1'b1;
                        //
                        recomb_msb_dout_carry_0 <= {16{1'bX}};
                        recomb_msb_dout_carry_1 <= recomb_msb_dout_carry_0;                        
                        //
                    end
                    //
                    recomb_msb_dout_delay_0 <= recomb_msb_dout;
                    recomb_msb_dout_delay_1 <= recomb_msb_dout_delay_0;
                    recomb_msb_dout_delay_2 <= recomb_msb_dout_delay_1;
                    //
                    recomb_msb_cnt_delay_0 <= fat_bram_xy_cnt_msb;
                    recomb_msb_cnt_delay_1 <= recomb_msb_cnt_delay_0;
                    recomb_msb_cnt_delay_2 <= recomb_msb_cnt_delay_1;
                    //
                    fat_bram_xy_cnt_msb         <= fat_bram_xy_cnt_msb + 1'b1;
                    //
                end
                //
            endcase
            //        
        end
    
    
    
    
endmodule