Subversion Repositories ion

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

-- ion_cpu.vhdl -- MIPS-I(tm) compatible CPU core

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

-- project:       ION (http://www.opencores.org/project,ion_cpu)

-- author:        Jose A. Ruiz (ja_rd@hotmail.com)

-- created:       Jan/11/2011

-- last modified: Jun/05/2011 (ja_rd@hotmail.com)

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

-- Please read file /doc/ion_project.txt for usage instructions.

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

--### MIPS-I things not implemented

--

-- 1.- Most of the R3000 CP0 registers and of course all of the CP1.

-- 2.- External interrupts missing, with CP0.SR IR, NMI and IM7..0 flags.

--

--### Things with provisional implementation

-- 

-- 1.- Load interlocks: the pipeline is stalled for every load instruction, even

--     if the target register is not used in the following instruction. So that

--     every load takes two cycles.

--     The interlock logic should check register indices (@note2)

-- 2.- CP0 SR (status register) bits KUo/IEo & KUP/IEp are missing.

--     This means that EXCEPTIONS CAN'T BE NESTED in this version of the CPU.

-- 3.- Invalid instruction side effects:

--     Invalid opcodes do trap but the logic that prevents bad opcodes from

--     having side affects has not been tested yet.

-- 4.- Kernel/user status.

--     When in user mode, COP* instructions will trigger a 'CpU' exception.

--     BUT there's no address checking and user code can still access kernel 

--     space in this version.

--     Besides, see point 2 above about the missing SR bits.

--

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

-- KNOWN BUGS:

--

-- 1.- The instruction executed right after entering user mode (i.e. the 

--     instruction after the MTC0 or RFE that clears the KU flag) is executed 

--     in kernel mode (instead of user mode). This is a gapping security hole,

--     in case it makes any sense to speak of security in this project at this

--     stage. 

--     This can be easily fixed but is not very urgent.

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

-- Copyright (C) 2011 Jose A. Ruiz

--                                                              

-- This source file may be used and distributed without         

-- restriction provided that this copyright statement is not    

-- removed from the file and that any derivative work contains  

-- the original copyright notice and the associated disclaimer. 

--                                                              

-- This source file is free software; you can redistribute it   

-- and/or modify it under the terms of the GNU Lesser General   

-- Public License as published by the Free Software Foundation; 

-- either version 2.1 of the License, or (at your option) any   

-- later version.                                               

--                                                              

-- This source is distributed in the hope that it will be       

-- useful, but WITHOUT ANY WARRANTY; without even the implied   

-- warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR      

-- PURPOSE.  See the GNU Lesser General Public License for more 

-- details.                                                     

--                                                              

-- You should have received a copy of the GNU Lesser General    

-- Public License along with this source; if not, download it   

-- from http://www.opencores.org/lgpl.shtml

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

library ieee;

use ieee.std_logic_1164.all;

use ieee.std_logic_arith.all;

use ieee.std_logic_unsigned.all;

use work.mips_pkg.all;

entity mips_cpu is

    generic(

        -- Reset vector address minus 4

        RESET_VECTOR_M4 : t_word    := RESET_VECTOR_M4;

        -- Trap vector address

        TRAP_VECTOR : t_word        := TRAP_VECTOR;

        -- Type of memory to be used for register bank in xilinx HW

        XILINX_REGBANK  : string    := "distributed" -- {distributed|block}

    );

    port(

        clk             : in std_logic;

        reset           : in std_logic;

        interrupt       : in std_logic;

        data_addr       : out std_logic_vector(31 downto 0);

        data_rd         : in std_logic_vector(31 downto 0);

        data_rd_vma     : out std_logic;

        byte_we         : out std_logic_vector(3 downto 0);

        data_wr         : out std_logic_vector(31 downto 0);

        code_rd_addr    : out std_logic_vector(31 downto 2);

        code_rd         : in std_logic_vector(31 downto 0);

        code_rd_vma     : out std_logic;

        cache_enable    : out std_logic;

        ic_invalidate   : out std_logic;

        mem_wait        : in std_logic

    );

end; --entity mips_cpu

architecture rtl of mips_cpu is

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

-- Pipeline stage 0

signal p0_pc_reg :          t_pc;

signal p0_pc_restart :      t_pc;

signal p0_pc_incremented :  t_pc;

signal p0_pc_jump :         t_pc;

signal p0_pc_branch :       t_pc;

signal p0_pc_target :       t_pc;

signal p0_pc_next :         t_pc;

signal p0_rs_num :          t_regnum;

signal p0_rt_num :          t_regnum;

signal p0_jump_cond_value : std_logic;

signal p0_rbank_rs_hazard : std_logic;

signal p0_rbank_rt_hazard : std_logic;

signal p0_uses_rs1 :        std_logic;

signal p0_uses_rs2 :        std_logic;

signal p1_rs1_hazard :      std_logic;

signal p1_rs2_hazard :      std_logic;

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

-- Pipeline stage 1

signal p1_rbank :           t_rbank := (others => X"00000000");

-- IMPORTANT: This attribute is used by Xilinx tools to select how to implement

-- the register bank. If we don't use it, by default XST would infer 2 BRAMs for

-- the 1024-bit 3-port reg bank, which you probably don't want.

-- This can take the values {distributed|block}.

attribute ram_style :       string;

attribute ram_style of p1_rbank : signal is XILINX_REGBANK;

signal p1_rs, p1_rt :       t_word;

signal p1_rs_rbank :        t_word;

signal p1_rt_rbank :        t_word;

signal p1_rbank_forward :   t_word;

signal p1_rd_num :          t_regnum;

signal p1_c0_rs_num :       t_regnum;

signal p1_rbank_wr_addr :   t_regnum;

signal p1_rbank_we :        std_logic;

signal p1_rbank_wr_data :   t_word;

signal p1_alu_inp1 :        t_word;

signal p1_alu_inp2 :        t_word;

signal p1_alu_outp :        t_word;

-- ALU control inputs (shortened name for brevity in expressions)

signal p1_ac :              t_alu_control;

-- ALU flag outputs (comparison results)

signal p1_alu_flags :       t_alu_flags;

-- immediate data, sign- or zero-extended as required by IR

signal p1_data_imm :        t_word;

signal p1_branch_offset :   t_pc;

signal p1_branch_offset_sex:std_logic_vector(31 downto 18);

signal p1_rbank_rs_hazard : std_logic;

signal p1_rbank_rt_hazard : std_logic;

signal p1_jump_type_set0 :  std_logic_vector(1 downto 0);

signal p1_jump_type_set1 :  std_logic_vector(1 downto 0);

signal p1_ir_reg :          std_logic_vector(31 downto 0);

signal p1_ir_op :           std_logic_vector(31 downto 26);

signal p1_ir_fn :           std_logic_vector(5 downto 0);

signal p1_op_special :      std_logic;

signal p1_exception :       std_logic;

signal p1_do_reg_jump :     std_logic;

signal p1_do_zero_ext_imm : std_logic;

signal p1_set_cp :          std_logic;

signal p1_get_cp :          std_logic;

signal p1_set_cp0 :         std_logic;

signal p1_get_cp0 :         std_logic;

signal p1_rfe :             std_logic;

signal p1_alu_op2_sel :     std_logic_vector(1 downto 0);

signal p1_alu_op2_sel_set0: std_logic_vector(1 downto 0);

signal p1_alu_op2_sel_set1: std_logic_vector(1 downto 0);

signal p1_do_load :         std_logic;

signal p1_do_store :        std_logic;

signal p1_store_size :      std_logic_vector(1 downto 0);

signal p1_we_control :      std_logic_vector(5 downto 0);

signal p1_load_alu :        std_logic;

signal p1_load_alu_set0 :   std_logic;

signal p1_load_alu_set1 :   std_logic;

signal p1_ld_upper_hword :  std_logic;

signal p1_ld_upper_byte :   std_logic;

signal p1_ld_unsigned :     std_logic;

signal p1_jump_type :       std_logic_vector(1 downto 0);

signal p1_link :            std_logic;

signal p1_jump_cond_sel :   std_logic_vector(2 downto 0);

signal p1_data_addr :       t_addr;

signal p1_data_offset :     t_addr;

signal p1_muldiv_result :   t_word;

signal p1_muldiv_func :     t_mult_function;

signal p1_muldiv_running :  std_logic;

signal p1_muldiv_started :  std_logic;

signal p1_muldiv_stall :    std_logic;

signal p1_unknown_opcode :  std_logic;

signal p1_cp_unavailable :  std_logic;

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

-- Pipeline stage 2

signal p2_muldiv_started :  std_logic;

signal p2_exception :       std_logic;

signal p2_rd_addr :         std_logic_vector(1 downto 0);

signal p2_rd_mux_control :  std_logic_vector(3 downto 0);

signal p2_load_target :     t_regnum;

signal p2_do_load :         std_logic;

signal p2_do_store :        std_logic;

signal p2_ld_upper_hword :  std_logic;

signal p2_ld_upper_byte :   std_logic;

signal p2_ld_unsigned :     std_logic;

signal p2_wback_mux_sel :   std_logic_vector(1 downto 0);

signal p2_data_word_rd :    t_word;

signal p2_data_word_ext :   std_logic;

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

-- Global control signals 

signal load_interlock :     std_logic;

signal stall_pipeline :     std_logic;

-- pipeline is stalled for any reason

signal pipeline_stalled :   std_logic;

signal stalled_memwait :    std_logic;

signal stalled_muldiv :     std_logic;

-- pipeline is stalled because of a load instruction interlock

signal stalled_interlock :  std_logic;

signal reset_done :         std_logic;

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

-- CP0 registers and signals

-- CP0[12]: status register, KUo/IEo & KUP/IEp & KU/IE  bits

signal cp0_status :         std_logic_vector(5 downto 0);

signal cp0_sr_ku_reg :      std_logic;

-- CP0[12]: status register, cache control

signal cp0_cache_control :  std_logic_vector(17 downto 16);

-- Output of CP0 register bank (only a few regs are implemented)

signal cp0_reg_read :       t_word;

-- CP0[14]: EPC register (PC value saved at exceptions)

signal cp0_epc :            t_pc;

-- CP0[13]: 'Cause' register (cause and attributes of exception)

signal cp0_cause :          t_word;

signal cp0_in_delay_slot :  std_logic;

signal cp0_cause_bd :       std_logic;

signal cp0_cause_ce :       std_logic_vector(1 downto 0);

signal cp0_cause_exc_code : std_logic_vector(4 downto 0);

begin

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

-- Register bank & datapath

-- Register indices are 'decoded' out of the instruction word BEFORE loading IR

p0_rs_num <= std_logic_vector(code_rd(25 downto 21));

with p1_ir_reg(31 downto 26) select p1_rd_num <=

    p1_ir_reg(15 downto 11)    when "000000",

    p1_ir_reg(20 downto 16)    when others;

p0_rt_num <= std_logic_vector(code_rd(20 downto 16)); -- also called rs2 in the docs

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

-- Data input shifter & masker (LB,LBU,LH,LHU,LW)

p2_rd_mux_control <= p2_ld_upper_hword & p2_ld_upper_byte & p2_rd_addr;

-- Extension for unused bits will be zero or the sign (bit 7 or bit 15)

p2_data_word_ext <= '0'          when p2_ld_unsigned='1' else

                    -- LH

                    data_rd(31)  when p2_ld_upper_byte='1' and p2_rd_addr="00" else

                    data_rd(15)  when p2_ld_upper_byte='1' and p2_rd_addr="10" else

                    -- LB

                    data_rd(7)  when p2_rd_addr="11" else

                    data_rd(15)  when p2_rd_addr="10" else

                    data_rd(23)  when p2_rd_addr="01" else

                    data_rd(31);

-- byte 0 may come from any of the 4 bytes of the input word

with p2_rd_mux_control select p2_data_word_rd(7 downto 0) <=

    data_rd(31 downto 24)        when "0000",

    data_rd(23 downto 16)        when "0001",

    data_rd(23 downto 16)        when "0100",

    data_rd(15 downto  8)        when "0010",

    data_rd( 7 downto  0)        when others;

-- byte 1 may come from input bytes 1 or 3 or may be extended for LB, LBU

with p2_rd_mux_control select p2_data_word_rd(15 downto 8) <=

    data_rd(31 downto 24)        when "0100",

    data_rd(15 downto  8)        when "0110",

    data_rd(15 downto  8)        when "1100",

    data_rd(15 downto  8)        when "1101",

    data_rd(15 downto  8)        when "1110",

    data_rd(15 downto  8)        when "1111",

    (others => p2_data_word_ext) when others;

-- bytes 2,3 come straight from input or are extended for LH,LHU

with p2_ld_upper_hword select p2_data_word_rd(31 downto 16) <=

    (others => p2_data_word_ext) when '0',

    data_rd(31 downto 16)        when others;

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

-- Reg bank input multiplexor

-- Select which data is to be written back to the reg bank and where

p1_rbank_wr_addr <= p1_rd_num   when p2_do_load='0' and p1_link='0' else

                    "11111"     when p2_do_load='0' and p1_link='1' else

                    p2_load_target;

p2_wback_mux_sel <=

    "00" when p2_do_load='0' and p1_get_cp0='0' and p1_link='0' else

    "01" when p2_do_load='1' and p1_get_cp0='0' and p1_link='0' else

    "10" when p2_do_load='0' and p1_get_cp0='1' and p1_link='0' else

    "11";

with (p2_wback_mux_sel) select p1_rbank_wr_data <=

    p1_alu_outp                when "00",

    p2_data_word_rd            when "01",

    p0_pc_incremented & "00"   when "11",

    cp0_reg_read               when others;

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

-- Register bank RAM & Rbank WE logic

p1_rbank_we <= '1' when (p2_do_load='1' or p1_load_alu='1' or p1_link='1' or

                        -- if mfc0 triggers privilege trap, don't load reg

                        (p1_get_cp0='1' and p1_cp_unavailable='0')) and

                        -- If target register is $zero, ignore write

                        p1_rbank_wr_addr/="00000" and

                        -- if the cache controller keeps the cpu stopped, do

                        -- not writeback

                        mem_wait='0' and

                        -- if stalled because of muldiv, block writeback

                        stalled_muldiv='0' and --@note1

                        -- on exception, abort next instruction (by preventing 

                        -- regbank writeback).

                        p2_exception='0'

                else '0';

-- Register bank as triple-port RAM. Should synth to 2 BRAMs unless you use

-- synth attributes to prevent it (see 'ram_style' attribute above) or your

-- FPGA has 3-port BRAMS, or has none.

synchronous_reg_bank:

process(clk)

begin

    if clk'event and clk='1' then

        if p1_rbank_we='1' then

            p1_rbank(conv_integer(p1_rbank_wr_addr)) <= p1_rbank_wr_data;

        end if;

        -- the rbank read port loads in the same conditions as the IR: don't

        -- update Rs or Rt if the pipeline is frozen

        if stall_pipeline='0' then

            p1_rt_rbank <= p1_rbank(conv_integer(p0_rt_num));

            p1_rs_rbank <= p1_rbank(conv_integer(p0_rs_num));

        end if;

    end if;

end process synchronous_reg_bank;

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

-- Reg bank 'writeback' data forwarding

-- Register writeback data in a DFF in case it needs to be forwarded.

data_forward_register:

process(clk)

begin

    if clk'event and clk='1' then

        if p1_rbank_we='1' then -- no need to check for stall cycles

            p1_rbank_forward <= p1_rbank_wr_data;

        end if;

    end if;

end process data_forward_register;

-- Bypass sync RAM if we're reading and writing to the same address. This saves

-- 1 stall cycle and fixes the data hazard.

p0_rbank_rs_hazard <= '1' when p1_rbank_wr_addr=p0_rs_num and p1_rbank_we='1'

                      else '0';

p0_rbank_rt_hazard <= '1' when p1_rbank_wr_addr=p0_rt_num and p1_rbank_we='1'

                      else '0';

p1_rs <= p1_rs_rbank when p1_rbank_rs_hazard='0' else p1_rbank_forward;

p1_rt <= p1_rt_rbank when p1_rbank_rt_hazard='0' else p1_rbank_forward;

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

-- ALU & ALU input multiplexors

p1_alu_inp1 <= p1_rs;

with p1_alu_op2_sel select p1_alu_inp2 <=

    p1_data_imm         when "11",

    p1_muldiv_result    when "01",

    --p1_muldiv_result    when "10", -- FIXME mux input wasted!

    p1_rt               when others;

alu_inst : entity work.mips_alu

    port map (

        clk             => clk,

        reset           => reset,

        ac              => p1_ac,

        flags           => p1_alu_flags,

        inp1            => p1_alu_inp1,

        inp2            => p1_alu_inp2,

        outp            => p1_alu_outp

    );

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

-- Mul/Div block interface

-- Compute the mdiv block function word. If p1_muldiv_func has any value other

-- than MULT_NOTHING a new mdiv operation will start, truncating whatever other

-- operation that may have been in course.

-- So we encode here the function to be performed and make sure the value stays

-- there for only one cycle (the first ALU cycle of the mul/div instruction).

-- This will be '1' for all mul/div operations other than NOP...

p1_muldiv_func(3) <= '1' when p1_op_special='1' and

                              p1_ir_fn(5 downto 4)="01" and

                              -- ...but only if the mdiv is not already running

                              p2_muldiv_started = '0' and

                              p1_muldiv_running ='0'

                      else '0';

-- When bit(3) is zero, the rest are zeroed too. Otherwise, they come from IR

p1_muldiv_func(2 downto 0) <=

    p1_ir_fn(3) & p1_ir_fn(1 downto 0) when p1_muldiv_func(3)='1'

    else "000";

mult_div: entity work.mips_mult

    port map (

        a           => p1_rs,

        b           => p1_rt,

        c_mult      => p1_muldiv_result,

        pause_out   => p1_muldiv_running,

        mult_func   => p1_muldiv_func,

        clk         => clk,

        reset_in    => reset

    );

-- Active only for the 1st ALU cycle of any mul/div instruction

p1_muldiv_started <= '1' when p1_op_special='1' and

                              p1_ir_fn(5 downto 3)="011" and

                              -- 

                              p1_muldiv_running='0'

                      else '0';

-- Stall the pipeline to enable mdiv operation completion.

-- We need p2_muldiv_started to distinguish the cycle before p1_muldiv_running

-- is asserted and the cycle after it deasserts.

-- Otherwise we would reexecute the same muldiv endlessly instruction after 

-- deassertion of p1_muldiv_running, since the IR was stalled and still contains 

-- the mul opcode...

p1_muldiv_stall <= '1' when

        -- Active for the cycle immediately before p1_muldiv_running asserts

        -- and NOT for the cycle after it deasserts

        (p1_muldiv_started='1' and p2_muldiv_started='0') or

        -- Active until operation is complete

        p1_muldiv_running = '1'

        else '0';

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

-- PC register and branch logic

-- p0_pc_reg will not be incremented on stall cycles

p0_pc_incremented <= p0_pc_reg + (not stall_pipeline);

-- main pc mux: jump or continue

p0_pc_next <=

    p0_pc_target when

        -- We jump on jump instructions whose condition is met...

        ((p1_jump_type(1)='1' and p0_jump_cond_value='1' and

        -- ...except we abort any jump that follows the victim of an exception

          p2_exception='0') or

        -- We jump on exceptions too...

        p1_exception='1')

        -- ... but we only jump at all if the pipeline is not stalled

        and stall_pipeline='0'

    else p0_pc_incremented;

pc_register:

process(clk)

begin

    if clk'event and clk='1' then

        if reset='1' then

            -- reset to <vector>-4 so that 1st fetch addr is <vector>

            p0_pc_reg <= RESET_VECTOR_M4(31 downto 2);

        else

            if reset_done='1' then

            -- p0_pc_reg holds the same value as external sync ram addr register

            p0_pc_reg <= p0_pc_next;

            -- p0_pc_restart = addr saved to EPC on interrupts (@note2)

            -- It's the addr of the instruction triggering the exception,

            -- except when the triggering instruction is in a delay slot. In 

            -- that case, this is the previous jump instruction address.

            -- I.e. all as per the mips-1 specs.

            if (p1_jump_type="00" or p0_jump_cond_value='0') then

                p0_pc_restart <= p0_pc_reg;

                -- remember if we are in a delay slot, in case there's a trap

                cp0_in_delay_slot <= '0'; -- NOT in a delay slot

            else

                cp0_in_delay_slot <= '1'; -- in a delay slot

            end if;

            end if;

        end if;

    end if;

end process pc_register;

-- Common rd/wr address; lowest 2 bits are output as debugging aid only

data_addr <= p1_data_addr(31 downto 0);

-- 'Memory enable' signals for both memory interfaces

data_rd_vma <= p1_do_load and not pipeline_stalled;

code_rd_vma <= (not stall_pipeline) and reset_done;

-- reset_done will be asserted after the reset process is finished, when the

-- CPU can start operating normally.

-- We only use it to make sure code_rd_vma is not asserted prematurely.

wait_for_end_of_reset:

process(clk)

begin

    if clk'event and clk='1' then

        if reset='1' then

            reset_done <= '0';

        else

            reset_done <= '1';

        end if;

    end if;

end process wait_for_end_of_reset;

-- The final value used to access code memory

code_rd_addr <= p0_pc_next;

-- compute target of J/JR instructions

p0_pc_jump <=   p1_rs(31 downto 2) when p1_do_reg_jump='1' else

                p0_pc_reg(31 downto 28) & p1_ir_reg(25 downto 0);

-- compute target of relative branch instructions

p1_branch_offset_sex <= (others => p1_ir_reg(15));

p1_branch_offset <= p1_branch_offset_sex & p1_ir_reg(15 downto 0);

-- p0_pc_reg is the addr of the instruction in delay slot

p0_pc_branch <= p0_pc_reg + p1_branch_offset;

-- decide which jump target is to be used

p0_pc_target <=

    TRAP_VECTOR(31 downto 2)    when p1_exception='1' else

    p0_pc_jump                  when p1_jump_type(0)='1' else

    p0_pc_branch;

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

-- Instruction decoding and IR

instruction_register:

process(clk)

begin

    if clk'event and clk='1' then

        if reset='1' then

            p1_ir_reg <= (others => '0');

        elsif stall_pipeline='0' then

            p1_ir_reg <= code_rd;

        end if;

    end if;

end process instruction_register;

-- Zero extension/Sign extension of instruction immediate data

p1_data_imm(15 downto 0)  <= p1_ir_reg(15 downto 0);

with p1_do_zero_ext_imm select p1_data_imm(31 downto 16) <=

    (others => '0')             when '1',

    (others => p1_ir_reg(15))   when others;

-- 'Extract' main fields from IR, for convenience

p1_ir_op <= p1_ir_reg(31 downto 26);

p1_ir_fn <= p1_ir_reg(5 downto 0);

-- Decode jump type, if any, for instructions with op/=0

with p1_ir_op select p1_jump_type_set0 <=

    -- FIXME verify that we actually weed out ALL invalid instructions

    "10" when "000001", -- BLTZ, BGEZ, BLTZAL, BGTZAL

    "11" when "000010", -- J

    "11" when "000011", -- JAL

    "10" when "000100", -- BEQ

    "10" when "000101", -- BNE

    "10" when "000110", -- BLEZ

    "10" when "000111", -- BGTZ

    "00" when others;   -- no jump

-- Decode jump type, if any, for instructions with op=0

p1_jump_type_set1 <= "11" when p1_op_special='1' and

                               p1_ir_reg(5 downto 1)="00100"

                     else "00";

-- Decode jump type for the instruction in IR (composite of two formats)

p1_jump_type <= p1_jump_type_set0 or p1_jump_type_set1;

p1_link <= '1' when (p1_ir_op="000000" and p1_ir_reg(5 downto 0)="001001") or

                    (p1_ir_op="000001" and p1_ir_reg(20)='1') or

                    (p1_ir_op="000011")

           else '0';

-- Decode jump condition: encode a mux control signal from IR...

p1_jump_cond_sel <=

    "001" when p1_ir_op="000001" and p1_ir_reg(16)='0' else --   op1 < 0   BLTZ*

    "101" when p1_ir_op="000001" and p1_ir_reg(16)='1' else -- !(op1 < 0) BNLTZ*

    "010" when p1_ir_op="000100" else                       --   op1 == op2  BEQ

    "110" when p1_ir_op="000101" else                       -- !(op1 == op2) BNE

    "011" when p1_ir_op="000110" else                       --   op1 <= 0   BLEZ

    "111" when p1_ir_op="000111" else                       -- !(op1 <= 0)  BGTZ

    "000";                                                  -- always

-- ... and use mux control signal to select the condition value

with p1_jump_cond_sel select p0_jump_cond_value <=

        p1_alu_flags.inp1_lt_zero       when "001",

    not p1_alu_flags.inp1_lt_zero       when "101",

        p1_alu_flags.inp1_eq_inp2       when "010",

    not p1_alu_flags.inp1_eq_inp2       when "110",

        (p1_alu_flags.inp1_lt_inp2 or

         p1_alu_flags.inp1_eq_inp2)     when "011",

    not (p1_alu_flags.inp1_lt_inp2 or

         p1_alu_flags.inp1_eq_inp2)     when "111",

    '1'                                 when others;

-- Decode instructions that launch exceptions

p1_exception <= '1' when

    (p1_op_special='1' and p1_ir_reg(5 downto 1)="00110") or -- syscall/break

    p1_unknown_opcode='1' or

    p1_cp_unavailable='1'

    else '0';

-- Decode MTC0/MFC0 instructions (see @note3)

p1_set_cp  <= '1' when p1_ir_reg(31 downto 28)="0100" and

                       p1_ir_reg(25 downto 21)="00100" else '0';

p1_get_cp  <= '1' when p1_ir_reg(31 downto 28)="0100" and

                       p1_ir_reg(25 downto 21)="00000" else '0';

p1_set_cp0 <= '1' when p1_ir_reg(27 downto 26)="00" and p1_set_cp='1' else '0';