Subversion Repositories neorv32

-- #################################################################################################

-- # << NEORV32 - Arithmetical/Logical Unit >>                                                     #

-- # ********************************************************************************************* #

-- # Main data and address ALU and co-processor interface/arbiter.                                 #

-- # ********************************************************************************************* #

-- # BSD 3-Clause License                                                                          #

-- #                                                                                               #

-- # Copyright (c) 2021, Stephan Nolting. All rights reserved.                                     #

-- #                                                                                               #

-- # Redistribution and use in source and binary forms, with or without modification, are          #

-- # permitted provided that the following conditions are met:                                     #

-- #                                                                                               #

-- # 1. Redistributions of source code must retain the above copyright notice, this list of        #

-- #    conditions and the following disclaimer.                                                   #

-- #                                                                                               #

-- # 2. Redistributions in binary form must reproduce the above copyright notice, this list of     #

-- #    conditions and the following disclaimer in the documentation and/or other materials        #

-- #    provided with the distribution.                                                            #

-- #                                                                                               #

-- # 3. Neither the name of the copyright holder nor the names of its contributors may be used to  #

-- #    endorse or promote products derived from this software without specific prior written      #

-- #    permission.                                                                                #

-- #                                                                                               #

-- # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS   #

-- # OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF               #

-- # MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE    #

-- # COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,     #

-- # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE #

-- # GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED    #

-- # AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING     #

-- # NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED  #

-- # OF THE POSSIBILITY OF SUCH DAMAGE.                                                            #

-- # ********************************************************************************************* #

-- # The NEORV32 Processor - https://github.com/stnolting/neorv32              (c) Stephan Nolting #

-- #################################################################################################

library ieee;

use ieee.std_logic_1164.all;

use ieee.numeric_std.all;

library neorv32;

use neorv32.neorv32_package.all;

entity neorv32_cpu_alu is

  generic (

    CPU_EXTENSION_RISCV_M : boolean := true;  -- implement muld/div extension?

    FAST_SHIFT_EN         : boolean := false; -- use barrel shifter for shift operations

    TINY_SHIFT_EN         : boolean := false  -- use tiny (single-bit) shifter for shift operations

  );

  port (

    -- global control --

    clk_i       : in  std_ulogic; -- global clock, rising edge

    rstn_i      : in  std_ulogic; -- global reset, low-active, async

    ctrl_i      : in  std_ulogic_vector(ctrl_width_c-1 downto 0); -- main control bus

    -- data input --

    rs1_i       : in  std_ulogic_vector(data_width_c-1 downto 0); -- rf source 1

    rs2_i       : in  std_ulogic_vector(data_width_c-1 downto 0); -- rf source 2

    pc2_i       : in  std_ulogic_vector(data_width_c-1 downto 0); -- delayed PC

    imm_i       : in  std_ulogic_vector(data_width_c-1 downto 0); -- immediate

    -- data output --

    res_o       : out std_ulogic_vector(data_width_c-1 downto 0); -- ALU result

    add_o       : out std_ulogic_vector(data_width_c-1 downto 0); -- address computation result

    -- co-processor interface --

    cp_start_o  : out std_ulogic_vector(7 downto 0); -- trigger co-processor i

    cp_valid_i  : in  std_ulogic_vector(7 downto 0); -- co-processor i done

    cp_result_i : in  cp_data_if_t; -- co-processor result

    -- status --

    wait_o      : out std_ulogic -- busy due to iterative processing units

  );

end neorv32_cpu_alu;

architecture neorv32_cpu_cpu_rtl of neorv32_cpu_alu is

  -- operands --

  signal opa, opb : std_ulogic_vector(data_width_c-1 downto 0);

  -- results --

  signal addsub_res : std_ulogic_vector(data_width_c downto 0);

  --

  signal cp_res     : std_ulogic_vector(data_width_c-1 downto 0);

  signal arith_res  : std_ulogic_vector(data_width_c-1 downto 0);

  signal logic_res  : std_ulogic_vector(data_width_c-1 downto 0);

  -- shifter --

  type shifter_t is record

    cmd     : std_ulogic;

    cmd_ff  : std_ulogic;

    start   : std_ulogic;

    run     : std_ulogic;

    halt    : std_ulogic;

    cnt     : std_ulogic_vector(4 downto 0);

    sreg    : std_ulogic_vector(data_width_c-1 downto 0);

    -- for barrel shifter only --

    bs_a_in : std_ulogic_vector(4 downto 0);

    bs_d_in : std_ulogic_vector(data_width_c-1 downto 0);

  end record;

  signal shifter : shifter_t;

  -- co-processor arbiter and interface --

  type cp_ctrl_t is record

    cmd     : std_ulogic;

    cmd_ff  : std_ulogic;

    busy    : std_ulogic;

    start   : std_ulogic;

    halt    : std_ulogic;

    timeout : std_ulogic_vector(9 downto 0);

  end record;

  signal cp_ctrl : cp_ctrl_t;

begin

  -- Operand Mux ----------------------------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  opa <= pc2_i when (ctrl_i(ctrl_alu_opa_mux_c) = '1') else rs1_i; -- operand a (first ALU input operand), only required for arithmetic ops

  opb <= imm_i when (ctrl_i(ctrl_alu_opb_mux_c) = '1') else rs2_i; -- operand b (second ALU input operand)

  -- Binary Adder/Subtractor ----------------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  binary_arithmetic_core: process(ctrl_i, opa, opb)

    variable cin_v  : std_ulogic_vector(0 downto 0);

    variable op_a_v : std_ulogic_vector(data_width_c downto 0);

    variable op_b_v : std_ulogic_vector(data_width_c downto 0);

    variable op_y_v : std_ulogic_vector(data_width_c downto 0);

    variable res_v  : std_ulogic_vector(data_width_c downto 0);

  begin

    -- operand sign-extension --

    op_a_v := (opa(opa'left) and (not ctrl_i(ctrl_alu_unsigned_c))) & opa;

    op_b_v := (opb(opb'left) and (not ctrl_i(ctrl_alu_unsigned_c))) & opb;

    -- add/sub(slt) select --

    if (ctrl_i(ctrl_alu_addsub_c) = '1') then -- subtraction

      op_y_v   := not op_b_v;

      cin_v(0) := '1';

    else -- addition

      op_y_v   := op_b_v;

      cin_v(0) := '0';

    end if;

    -- adder core (result + carry/borrow) --

    addsub_res <= std_ulogic_vector(unsigned(op_a_v) + unsigned(op_y_v) + unsigned(cin_v(0 downto 0)));

  end process binary_arithmetic_core;

  -- direct output of address result --

  add_o <= addsub_res(data_width_c-1 downto 0);

  -- ALU arithmetic logic core --

  arithmetic_core: process(ctrl_i, addsub_res)

  begin

    if (ctrl_i(ctrl_alu_arith_c) = alu_arith_cmd_addsub_c) then -- ADD/SUB

      arith_res <= addsub_res(data_width_c-1 downto 0);

    else -- SLT

      arith_res <= (others => '0');

      arith_res(0) <= addsub_res(addsub_res'left); -- => carry/borrow

    end if;

  end process arithmetic_core;

  -- Shifter Unit ---------------------------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  shifter_unit: process(clk_i)

    variable bs_input_v   : std_ulogic_vector(data_width_c-1 downto 0);

    variable bs_level_4_v : std_ulogic_vector(data_width_c-1 downto 0);

    variable bs_level_3_v : std_ulogic_vector(data_width_c-1 downto 0);

    variable bs_level_2_v : std_ulogic_vector(data_width_c-1 downto 0);

    variable bs_level_1_v : std_ulogic_vector(data_width_c-1 downto 0);

    variable bs_level_0_v : std_ulogic_vector(data_width_c-1 downto 0);

  begin

    if rising_edge(clk_i) then

      shifter.cmd_ff <= shifter.cmd;

      -- --------------------------------------------------------------------------------

      -- Iterative shifter (small but slow) (default)

      -- --------------------------------------------------------------------------------

      if (FAST_SHIFT_EN = false) then

        if (shifter.start = '1') then -- trigger new shift

          shifter.sreg <= rs1_i; -- shift operand (can only be rs1; opa would also contain pc)

          shifter.cnt  <= opb(index_size_f(data_width_c)-1 downto 0); -- shift amount

        elsif (shifter.run = '1') then -- running shift

          -- coarse shift: multiples of 4 --

          if (TINY_SHIFT_EN = false) and -- use coarse shifts first if TINY SHIFT option is NOT enabled

             (or_all_f(shifter.cnt(shifter.cnt'left downto 2)) = '1') then -- shift amount >= 4

            shifter.cnt <= std_ulogic_vector(unsigned(shifter.cnt) - 4);

            if (ctrl_i(ctrl_alu_shift_dir_c) = '0') then -- SLL: shift left logical

              shifter.sreg <= shifter.sreg(shifter.sreg'left-4 downto 0) & "0000";

            else -- SRL: shift right logical / SRA: shift right arithmetical

              shifter.sreg <= (shifter.sreg(shifter.sreg'left) and ctrl_i(ctrl_alu_shift_ar_c)) &

                              (shifter.sreg(shifter.sreg'left) and ctrl_i(ctrl_alu_shift_ar_c)) &

                              (shifter.sreg(shifter.sreg'left) and ctrl_i(ctrl_alu_shift_ar_c)) &

                              (shifter.sreg(shifter.sreg'left) and ctrl_i(ctrl_alu_shift_ar_c)) & shifter.sreg(shifter.sreg'left downto 4);

            end if;

          -- fine shift: single shifts, 0..3 times; use ONLY single-bit shifts if TINY_SHIFT_EN is enabled (even smaller than the default approach) --

          else

            shifter.cnt <= std_ulogic_vector(unsigned(shifter.cnt) - 1);

            if (ctrl_i(ctrl_alu_shift_dir_c) = '0') then -- SLL: shift left logical

              shifter.sreg <= shifter.sreg(shifter.sreg'left-1 downto 0) & '0';

            else -- SRL: shift right logical / SRA: shift right arithmetical

              shifter.sreg <= (shifter.sreg(shifter.sreg'left) and ctrl_i(ctrl_alu_shift_ar_c)) & shifter.sreg(shifter.sreg'left downto 1);

            end if;

          end if;

        end if;

      -- --------------------------------------------------------------------------------

      -- Barrel shifter (huge but fast)

      -- --------------------------------------------------------------------------------

      else

        -- operands and cycle control --

        if (shifter.start = '1') then -- trigger new shift

          shifter.bs_d_in <= rs1_i; -- shift operand (can only be rs1; opa would also contain pc)

          shifter.bs_a_in <= opb(index_size_f(data_width_c)-1 downto 0); -- shift amount

          shifter.cnt     <= (others => '0');

        end if;

        -- convert left shifts to right shifts --

        if (ctrl_i(ctrl_alu_shift_dir_c) = '0') then -- is left shift?

          bs_input_v := bit_rev_f(shifter.bs_d_in); -- reverse bit order of input operand

        else

          bs_input_v := shifter.bs_d_in;

        end if;

        -- shift >> 16 --

        if (shifter.bs_a_in(4) = '1') then

          bs_level_4_v(31 downto 16) := (others => (bs_input_v(bs_input_v'left) and ctrl_i(ctrl_alu_shift_ar_c)));

          bs_level_4_v(15 downto 00) := (bs_input_v(31 downto 16));

        else

          bs_level_4_v := bs_input_v;

        end if;

        -- shift >> 8 --

        if (shifter.bs_a_in(3) = '1') then

          bs_level_3_v(31 downto 24) := (others => (bs_input_v(bs_input_v'left) and ctrl_i(ctrl_alu_shift_ar_c)));

          bs_level_3_v(23 downto 00) := (bs_level_4_v(31 downto 8));

        else

          bs_level_3_v := bs_level_4_v;

        end if;

        -- shift >> 4 --

        if (shifter.bs_a_in(2) = '1') then

          bs_level_2_v(31 downto 28) := (others => (bs_input_v(bs_input_v'left) and ctrl_i(ctrl_alu_shift_ar_c)));

          bs_level_2_v(27 downto 00) := (bs_level_3_v(31 downto 4));

        else

          bs_level_2_v := bs_level_3_v;

        end if;

        -- shift >> 2 --

        if (shifter.bs_a_in(1) = '1') then

          bs_level_1_v(31 downto 30) := (others => (bs_input_v(bs_input_v'left) and ctrl_i(ctrl_alu_shift_ar_c)));

          bs_level_1_v(29 downto 00) := (bs_level_2_v(31 downto 2));

        else

          bs_level_1_v := bs_level_2_v;

        end if;

        -- shift >> 1 --

        if (shifter.bs_a_in(0) = '1') then

          bs_level_0_v(31 downto 31) := (others => (bs_input_v(bs_input_v'left) and ctrl_i(ctrl_alu_shift_ar_c)));

          bs_level_0_v(30 downto 00) := (bs_level_1_v(31 downto 1));

        else

          bs_level_0_v := bs_level_1_v;

        end if;

        -- re-convert original left shifts --

        if (ctrl_i(ctrl_alu_shift_dir_c) = '0') then

          shifter.sreg <= bit_rev_f(bs_level_0_v);

        else

          shifter.sreg <= bs_level_0_v;

        end if;

      end if;

    end if;

  end process shifter_unit;

  -- is shift operation? --

  shifter.cmd   <= '1' when (ctrl_i(ctrl_alu_func1_c downto ctrl_alu_func0_c) = alu_func_cmd_shift_c) else '0';

  shifter.start <= '1' when (shifter.cmd = '1') and (shifter.cmd_ff = '0') else '0';

  -- shift operation running? --

  shifter.run  <= '1' when (or_all_f(shifter.cnt) = '1') or (shifter.start = '1') else '0';

  shifter.halt <= '1' when (or_all_f(shifter.cnt(shifter.cnt'left downto 1)) = '1') or (shifter.start = '1') else '0';

  -- Co-Processor Arbiter -------------------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  -- Interface:

  -- Co-processor "valid" signal has to be asserted (for one cycle) one cycle before asserting output data

  -- Co-processor "output data" has to be always zero unless co-processor was explicitly triggered

  cp_arbiter: process(rstn_i, clk_i)

  begin

    if (rstn_i = '0') then

      cp_ctrl.cmd_ff  <= '0';

      cp_ctrl.busy    <= '0';

      cp_ctrl.timeout <= (others => '0');

    elsif rising_edge(clk_i) then

      cp_ctrl.cmd_ff <= cp_ctrl.cmd;

      if (or_all_f(cp_valid_i) = '1') then -- cp computation done?

        cp_ctrl.busy <= '0';

      elsif (cp_ctrl.timeout(cp_ctrl.timeout'left) = '1') and (cp_timeout_en_c = true) then -- timeout

        assert false report "NEORV32 CPU CO-PROCESSOR TIMEOUT ERROR!" severity warning;

        cp_ctrl.busy <= '0';

      elsif (cp_ctrl.start = '1') then

        cp_ctrl.busy <= '1';

      end if;

      -- timeout counter --

      if (cp_ctrl.busy = '1') and (cp_timeout_en_c = true) then

        cp_ctrl.timeout <= std_ulogic_vector(unsigned(cp_ctrl.timeout) + 1);

      else

        cp_ctrl.timeout <= (others => '0');

      end if;

    end if;

  end process cp_arbiter;

  -- is co-processor operation? --

  cp_ctrl.cmd   <= '1' when (ctrl_i(ctrl_alu_func1_c downto ctrl_alu_func0_c) = alu_func_cmd_copro_c) else '0';

  cp_ctrl.start <= '1' when (cp_ctrl.cmd = '1') and (cp_ctrl.cmd_ff = '0') else '0';

  -- co-processor select --

  cp_operation_trigger: process(cp_ctrl, ctrl_i)

  begin

    for i in 0 to 7 loop

      if (cp_ctrl.start = '1') and (ctrl_i(ctrl_cp_id_msb_c downto ctrl_cp_id_lsb_c) = std_ulogic_vector(to_unsigned(i, 3))) then

        cp_start_o(i) <= '1';

      else

        cp_start_o(i) <= '0';

      end if;

    end loop; -- i

  end process;

  -- co-processor operation (still) running? --

  cp_ctrl.halt <= (cp_ctrl.busy and (not or_all_f(cp_valid_i))) or cp_ctrl.start;

  -- co-processor result - only the *actually selected* co-processor may output data != 0 --

  cp_res <= cp_result_i(0) or cp_result_i(1) or cp_result_i(2) or cp_result_i(3) or

            cp_result_i(4) or cp_result_i(5) or cp_result_i(6) or cp_result_i(7);

  -- ALU Logic Core -------------------------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  alu_logic_core: process(ctrl_i, rs1_i, opb)

  begin

    case ctrl_i(ctrl_alu_logic1_c downto ctrl_alu_logic0_c) is

      when alu_logic_cmd_movb_c => logic_res <= opb; -- (default)

      when alu_logic_cmd_xor_c  => logic_res <= rs1_i xor opb; -- only rs1 required for logic ops (opa would also contain pc)

      when alu_logic_cmd_or_c   => logic_res <= rs1_i or  opb;

      when alu_logic_cmd_and_c  => logic_res <= rs1_i and opb;

      when others               => logic_res <= opb; -- undefined

    end case;

  end process alu_logic_core;

  -- ALU Function Select --------------------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  alu_function_mux: process(ctrl_i, arith_res, logic_res, shifter.sreg, cp_res)

  begin

    case ctrl_i(ctrl_alu_func1_c downto ctrl_alu_func0_c) is

      when alu_func_cmd_arith_c => res_o <= arith_res; -- (default)

      when alu_func_cmd_logic_c => res_o <= logic_res;

      when alu_func_cmd_shift_c => res_o <= shifter.sreg;

      when alu_func_cmd_copro_c => res_o <= cp_res;

      when others               => res_o <= arith_res; -- undefined

    end case;

  end process alu_function_mux;

  -- ALU Busy -------------------------------------------------------------------------------

  -- -------------------------------------------------------------------------------------------

  wait_o <= shifter.halt or cp_ctrl.halt; -- wait until iterative units have completed

end neorv32_cpu_cpu_rtl;