243 lines
15 KiB
VHDL
243 lines
15 KiB
VHDL
-- #################################################################################################
|
|
-- # << NEORV32 - Custom Functions Subsystem (CFS) >> #
|
|
-- # ********************************************************************************************* #
|
|
-- # Intended for tightly-coupled, application-specific custom co-processors. This module provides #
|
|
-- # 64x 32-bit memory-mapped interface registers, one interrupt request signal and custom IO #
|
|
-- # conduits for processor-external or chip-external interface. #
|
|
-- # #
|
|
-- # NOTE: This is just an example/illustration template. Modify/replace this file to implement #
|
|
-- # your own custom design logic. #
|
|
-- # ********************************************************************************************* #
|
|
-- # BSD 3-Clause License #
|
|
-- # #
|
|
-- # The NEORV32 RISC-V Processor, https://github.com/stnolting/neorv32 #
|
|
-- # Copyright (c) 2024, 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. #
|
|
-- #################################################################################################
|
|
|
|
library ieee;
|
|
use ieee.std_logic_1164.all;
|
|
use ieee.numeric_std.all;
|
|
|
|
library neorv32;
|
|
use neorv32.neorv32_package.all;
|
|
|
|
entity neorv32_cfs is
|
|
generic (
|
|
CFS_CONFIG : std_ulogic_vector(31 downto 0); -- custom CFS configuration generic
|
|
CFS_IN_SIZE : natural; -- size of CFS input conduit in bits
|
|
CFS_OUT_SIZE : natural -- size of CFS output conduit in bits
|
|
);
|
|
port (
|
|
clk_i : in std_ulogic; -- global clock line
|
|
rstn_i : in std_ulogic; -- global reset line, low-active, use as async
|
|
bus_req_i : in bus_req_t; -- bus request
|
|
bus_rsp_o : out bus_rsp_t := rsp_terminate_c; -- bus response
|
|
clkgen_en_o : out std_ulogic := '0'; -- enable clock generator
|
|
clkgen_i : in std_ulogic_vector(7 downto 0); -- "clock" inputs
|
|
irq_o : out std_ulogic := '0'; -- interrupt request
|
|
cfs_in_i : in std_ulogic_vector(CFS_IN_SIZE-1 downto 0); -- custom inputs
|
|
cfs_out_o : out std_ulogic_vector(CFS_OUT_SIZE-1 downto 0) := (others => '0') -- custom outputs
|
|
);
|
|
end neorv32_cfs;
|
|
|
|
architecture neorv32_cfs_rtl of neorv32_cfs is
|
|
|
|
-- default CFS interface registers --
|
|
type cfs_regs_t is array (0 to 3) of std_ulogic_vector(31 downto 0); -- just implement 4 registers for this example
|
|
signal cfs_reg_wr : cfs_regs_t; -- interface registers for WRITE accesses
|
|
signal cfs_reg_rd : cfs_regs_t; -- interface registers for READ accesses
|
|
|
|
begin
|
|
|
|
-- CFS Generics ---------------------------------------------------------------------------
|
|
-- -------------------------------------------------------------------------------------------
|
|
-- In it's default version the CFS provides three configuration generics:
|
|
-- > CFS_IN_SIZE - configures the size (in bits) of the CFS input conduit cfs_in_i
|
|
-- > CFS_OUT_SIZE - configures the size (in bits) of the CFS output conduit cfs_out_o
|
|
-- > CFS_CONFIG - is a blank 32-bit generic. It is intended as a "generic conduit" to propagate
|
|
-- custom configuration flags from the top entity down to this module.
|
|
|
|
|
|
-- CFS IOs --------------------------------------------------------------------------------
|
|
-- -------------------------------------------------------------------------------------------
|
|
-- By default, the CFS provides two IO signals (cfs_in_i and cfs_out_o) that are available at the processor's top entity.
|
|
-- These are intended as "conduits" to propagate custom signals from this module and the processor top entity.
|
|
|
|
cfs_out_o <= (others => '0'); -- not used for this minimal example
|
|
|
|
|
|
-- Reset System ---------------------------------------------------------------------------
|
|
-- -------------------------------------------------------------------------------------------
|
|
-- The CFS can be reset using the global rstn_i signal. This signal should be used as asynchronous reset and is active-low.
|
|
-- Note that rstn_i can be asserted by a processor-external reset, the on-chip debugger and also by the watchdog.
|
|
--
|
|
-- Most default peripheral devices of the NEORV32 do NOT use a dedicated hardware reset at all. Instead, these units are
|
|
-- reset by writing ZERO to a specific "control register" located right at the beginning of the device's address space
|
|
-- (so this register is cleared at first). The crt0 start-up code writes ZERO to every single address in the processor's
|
|
-- IO space - including the CFS. Make sure that this initial clearing does not cause any unintended CFS actions.
|
|
|
|
|
|
-- Clock System ---------------------------------------------------------------------------
|
|
-- -------------------------------------------------------------------------------------------
|
|
-- The processor top unit implements a clock generator providing 8 "derived clocks".
|
|
-- Actually, these signals should not be used as direct clock signals, but as *clock enable* signals.
|
|
-- clkgen_i is always synchronous to the main system clock (clk_i).
|
|
--
|
|
-- The following clock dividers are available:
|
|
-- > clkgen_i(clk_div2_c) -> MAIN_CLK/2
|
|
-- > clkgen_i(clk_div4_c) -> MAIN_CLK/4
|
|
-- > clkgen_i(clk_div8_c) -> MAIN_CLK/8
|
|
-- > clkgen_i(clk_div64_c) -> MAIN_CLK/64
|
|
-- > clkgen_i(clk_div128_c) -> MAIN_CLK/128
|
|
-- > clkgen_i(clk_div1024_c) -> MAIN_CLK/1024
|
|
-- > clkgen_i(clk_div2048_c) -> MAIN_CLK/2048
|
|
-- > clkgen_i(clk_div4096_c) -> MAIN_CLK/4096
|
|
--
|
|
-- For instance, if you want to drive a clock process at MAIN_CLK/8 clock speed you can use the following construct:
|
|
--
|
|
-- if (rstn_i = '0') then -- async and low-active reset (if required at all)
|
|
-- ...
|
|
-- elsif rising_edge(clk_i) then -- always use the main clock for all clock processes
|
|
-- if (clkgen_i(clk_div8_c) = '1') then -- the div8 "clock" is actually a clock enable
|
|
-- ...
|
|
-- end if;
|
|
-- end if;
|
|
--
|
|
-- The clkgen_i input clocks are available when at least one IO/peripheral device (for example UART0) requires the clocks
|
|
-- generated by the clock generator. The CFS can enable the clock generator by itself by setting the clkgen_en_o signal high.
|
|
-- The CFS cannot ensure to deactivate the clock generator by setting the clkgen_en_o signal low as other peripherals might
|
|
-- still keep the generator activated. Make sure to deactivate the CFS's clkgen_en_o if no clocks are required in here to
|
|
-- reduce dynamic power consumption.
|
|
|
|
clkgen_en_o <= '0'; -- not used for this minimal example
|
|
|
|
|
|
-- Interrupt ------------------------------------------------------------------------------
|
|
-- -------------------------------------------------------------------------------------------
|
|
-- The CFS features a single interrupt signal, which is connected to the CPU's "fast interrupt" channel 1 (FIRQ1).
|
|
-- The interrupt is triggered by a one-cycle high-level. After triggering, the interrupt appears as "pending" in the CPU's
|
|
-- mip CSR ready to trigger execution of the according interrupt handler. It is the task of the application to programmer
|
|
-- to enable/clear the CFS interrupt using the CPU's mie and mip registers when required.
|
|
|
|
irq_o <= '0'; -- not used for this minimal example
|
|
|
|
|
|
-- Read/Write Access ----------------------------------------------------------------------
|
|
-- -------------------------------------------------------------------------------------------
|
|
-- Here we are reading/writing from/to the interface registers of the module and generate the CPU access handshake (bus response).
|
|
--
|
|
-- The CFS provides up to 64 memory-mapped 32-bit interface registers. For instance, these could be used to provide a
|
|
-- <control register> for global control of the unit, a <data register> for reading/writing from/to a data FIFO, a
|
|
-- <command register> for issuing commands and a <status register> for status information.
|
|
--
|
|
-- Following the interface protocol, each read or write access has to be acknowledged in the following cycle using the ack_o
|
|
-- signal (or even later if the module needs additional time). If no ACK is generated at all, the bus access will time out
|
|
-- and cause a bus access fault exception. The current CPU privilege level is available via the 'priv_i' signal (0 = user mode,
|
|
-- 1 = machine mode), which can be used to constrain access to certain registers or features to privileged software only.
|
|
--
|
|
-- This module also provides an optional ERROR signal to indicate a faulty access operation (for example when accessing an
|
|
-- unused, read-only or "locked" CFS register address). This signal may only be set when the module is actually accessed
|
|
-- and is set INSTEAD of the ACK signal. Setting the ERR signal will raise a bus access exception with a "Device Error" qualifier
|
|
-- that can be handled by the application software. Note that the current privilege level should not be exposed to software to
|
|
-- maintain full virtualization. Hence, CFS-based "privilege escalation" should trigger a bus access exception (e.g. by setting 'err_o').
|
|
--
|
|
-- Host access example: Read and write access to the interface registers + bus transfer acknowledge. This example only
|
|
-- implements four physical r/w register (the four lowest CFS registers). The remaining addresses of the CFS are not associated
|
|
-- with any physical registers - any access to those is simply ignored but still acknowledged. Only full-word write accesses are
|
|
-- supported (and acknowledged) by this example. Sub-word write access will not alter any CFS register state and will cause
|
|
-- a "bus store access" exception (with a "Device Timeout" qualifier as not ACK is generated in that case).
|
|
|
|
bus_access: process(rstn_i, clk_i)
|
|
begin
|
|
if (rstn_i = '0') then
|
|
cfs_reg_wr(0) <= (others => '0');
|
|
cfs_reg_wr(1) <= (others => '0');
|
|
cfs_reg_wr(2) <= (others => '0');
|
|
cfs_reg_wr(3) <= (others => '0');
|
|
--
|
|
bus_rsp_o.ack <= '0';
|
|
bus_rsp_o.err <= '0';
|
|
bus_rsp_o.data <= (others => '0');
|
|
elsif rising_edge(clk_i) then -- synchronous interface for read and write accesses
|
|
-- transfer/access acknowledge --
|
|
bus_rsp_o.ack <= bus_req_i.stb;
|
|
|
|
-- tie to zero if not explicitly used --
|
|
bus_rsp_o.err <= '0';
|
|
|
|
-- defaults --
|
|
bus_rsp_o.data <= (others => '0'); -- the output HAS TO BE ZERO if there is no actual (read) access
|
|
|
|
-- bus access --
|
|
if (bus_req_i.stb = '1') then -- valid access cycle, STB is high for one cycle
|
|
|
|
-- write access --
|
|
if (bus_req_i.rw = '1') then
|
|
if (bus_req_i.addr(7 downto 2) = "000000") then -- address size is fixed!
|
|
cfs_reg_wr(0) <= bus_req_i.data;
|
|
end if;
|
|
if (bus_req_i.addr(7 downto 2) = "000001") then
|
|
cfs_reg_wr(1) <= bus_req_i.data;
|
|
end if;
|
|
if (bus_req_i.addr(7 downto 2) = "000010") then
|
|
cfs_reg_wr(2) <= bus_req_i.data;
|
|
end if;
|
|
if (bus_req_i.addr(7 downto 2) = "000011") then
|
|
cfs_reg_wr(3) <= bus_req_i.data;
|
|
end if;
|
|
|
|
-- read access --
|
|
else
|
|
case bus_req_i.addr(7 downto 2) is -- address size is fixed!
|
|
when "000000" => bus_rsp_o.data <= cfs_reg_rd(0);
|
|
when "000001" => bus_rsp_o.data <= cfs_reg_rd(1);
|
|
when "000010" => bus_rsp_o.data <= cfs_reg_rd(2);
|
|
when "000011" => bus_rsp_o.data <= cfs_reg_rd(3);
|
|
when others => bus_rsp_o.data <= (others => '0');
|
|
end case;
|
|
end if;
|
|
|
|
end if;
|
|
end if;
|
|
end process bus_access;
|
|
|
|
|
|
-- CFS Function Core ----------------------------------------------------------------------
|
|
-- -------------------------------------------------------------------------------------------
|
|
|
|
-- This is where the actual functionality can be implemented.
|
|
-- The logic below is just a very simple example that transforms data
|
|
-- from an input register into data in an output register.
|
|
|
|
cfs_reg_rd(0) <= bin_to_gray_f(cfs_reg_wr(0)); -- convert binary to gray code
|
|
cfs_reg_rd(1) <= gray_to_bin_f(cfs_reg_wr(1)); -- convert gray to binary code
|
|
cfs_reg_rd(2) <= bit_rev_f(cfs_reg_wr(2)); -- bit reversal
|
|
cfs_reg_rd(3) <= bswap32_f(cfs_reg_wr(3)); -- byte swap (endianness conversion)
|
|
|
|
|
|
end neorv32_cfs_rtl;
|