///////////////////////////////////////////////////////////////////////////
|
///////////////////////////////////////////////////////////////////////////
|
//
|
//
|
// Filename: cpuops.v
|
// Filename: cpuops.v
|
//
|
//
|
// Project: Zip CPU -- a small, lightweight, RISC CPU soft core
|
// Project: Zip CPU -- a small, lightweight, RISC CPU soft core
|
//
|
//
|
// Purpose: This supports the instruction set reordering of operations
|
// Purpose: This supports the instruction set reordering of operations
|
// created by the second generation instruction set, as well as
|
// created by the second generation instruction set, as well as
|
// the new operations of POPC (population count) and BREV (bit reversal).
|
// the new operations of POPC (population count) and BREV (bit reversal).
|
//
|
//
|
//
|
//
|
// Creator: Dan Gisselquist, Ph.D.
|
// Creator: Dan Gisselquist, Ph.D.
|
// Gisselquist Technology, LLC
|
// Gisselquist Technology, LLC
|
//
|
//
|
///////////////////////////////////////////////////////////////////////////
|
///////////////////////////////////////////////////////////////////////////
|
//
|
//
|
// Copyright (C) 2015, Gisselquist Technology, LLC
|
// Copyright (C) 2015, Gisselquist Technology, LLC
|
//
|
//
|
// This program is free software (firmware): you can redistribute it and/or
|
// This program is free software (firmware): you can redistribute it and/or
|
// modify it under the terms of the GNU General Public License as published
|
// modify it under the terms of the GNU General Public License as published
|
// by the Free Software Foundation, either version 3 of the License, or (at
|
// by the Free Software Foundation, either version 3 of the License, or (at
|
// your option) any later version.
|
// your option) any later version.
|
//
|
//
|
// This program is distributed in the hope that it will be useful, but WITHOUT
|
// This program is distributed in the hope that it will be useful, but WITHOUT
|
// ANY WARRANTY; without even the implied warranty of MERCHANTIBILITY or
|
// ANY WARRANTY; without even the implied warranty of MERCHANTIBILITY or
|
// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
|
// FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
|
// for more details.
|
// for more details.
|
//
|
//
|
// License: GPL, v3, as defined and found on www.gnu.org,
|
// License: GPL, v3, as defined and found on www.gnu.org,
|
// http://www.gnu.org/licenses/gpl.html
|
// http://www.gnu.org/licenses/gpl.html
|
//
|
//
|
//
|
//
|
///////////////////////////////////////////////////////////////////////////
|
///////////////////////////////////////////////////////////////////////////
|
//
|
//
|
`define LONG_MPY
|
`define LONG_MPY
|
module cpuops(i_clk,i_rst, i_ce, i_valid, i_op, i_a, i_b, o_c, o_f, o_valid,
|
module cpuops(i_clk,i_rst, i_ce, i_valid, i_op, i_a, i_b, o_c, o_f, o_valid,
|
o_illegal, o_busy);
|
o_illegal, o_busy);
|
parameter IMPLEMENT_MPY = 1;
|
parameter IMPLEMENT_MPY = 1;
|
input i_clk, i_rst, i_ce;
|
input i_clk, i_rst, i_ce;
|
input [3:0] i_op;
|
input [3:0] i_op;
|
input [31:0] i_a, i_b;
|
input [31:0] i_a, i_b;
|
input i_valid;
|
input i_valid;
|
output reg [31:0] o_c;
|
output reg [31:0] o_c;
|
output wire [3:0] o_f;
|
output wire [3:0] o_f;
|
output reg o_valid;
|
output reg o_valid;
|
output wire o_illegal;
|
output wire o_illegal;
|
output wire o_busy;
|
output wire o_busy;
|
|
|
// Rotate-left pre-logic
|
// Rotate-left pre-logic
|
wire [63:0] w_rol_tmp;
|
wire [63:0] w_rol_tmp;
|
assign w_rol_tmp = { i_a, i_a } << i_b[4:0];
|
assign w_rol_tmp = { i_a, i_a } << i_b[4:0];
|
wire [31:0] w_rol_result;
|
wire [31:0] w_rol_result;
|
assign w_rol_result = w_rol_tmp[63:32]; // Won't set flags
|
assign w_rol_result = w_rol_tmp[63:32]; // Won't set flags
|
|
|
// Shift register pre-logic
|
// Shift register pre-logic
|
wire [32:0] w_lsr_result, w_asr_result;
|
wire [32:0] w_lsr_result, w_asr_result;
|
assign w_asr_result = (|i_b[31:5])? {(33){i_a[31]}}
|
assign w_asr_result = (|i_b[31:5])? {(33){i_a[31]}}
|
: ( {i_a, 1'b0 } >>> (i_b[4:0]) );// ASR
|
: ( {i_a, 1'b0 } >>> (i_b[4:0]) );// ASR
|
assign w_lsr_result = (|i_b[31:5])? 33'h00
|
assign w_lsr_result = (|i_b[31:5])? 33'h00
|
: ( { i_a, 1'b0 } >> (i_b[4:0]) );// LSR
|
: ( { i_a, 1'b0 } >> (i_b[4:0]) );// LSR
|
|
|
// Bit reversal pre-logic
|
// Bit reversal pre-logic
|
wire [31:0] w_brev_result;
|
wire [31:0] w_brev_result;
|
genvar k;
|
genvar k;
|
generate
|
generate
|
for(k=0; k<32; k=k+1)
|
for(k=0; k<32; k=k+1)
|
begin : bit_reversal_cpuop
|
begin : bit_reversal_cpuop
|
assign w_brev_result[k] = i_b[31-k];
|
assign w_brev_result[k] = i_b[31-k];
|
end endgenerate
|
end endgenerate
|
|
|
// Popcount pre-logic
|
// Popcount pre-logic
|
wire [31:0] w_popc_result;
|
wire [31:0] w_popc_result;
|
assign w_popc_result[5:0]=
|
assign w_popc_result[5:0]=
|
({5'h0,i_b[ 0]}+{5'h0,i_b[ 1]}+{5'h0,i_b[ 2]}+{5'h0,i_b[ 3]})
|
({5'h0,i_b[ 0]}+{5'h0,i_b[ 1]}+{5'h0,i_b[ 2]}+{5'h0,i_b[ 3]})
|
+({5'h0,i_b[ 4]}+{5'h0,i_b[ 5]}+{5'h0,i_b[ 6]}+{5'h0,i_b[ 7]})
|
+({5'h0,i_b[ 4]}+{5'h0,i_b[ 5]}+{5'h0,i_b[ 6]}+{5'h0,i_b[ 7]})
|
+({5'h0,i_b[ 8]}+{5'h0,i_b[ 9]}+{5'h0,i_b[10]}+{5'h0,i_b[11]})
|
+({5'h0,i_b[ 8]}+{5'h0,i_b[ 9]}+{5'h0,i_b[10]}+{5'h0,i_b[11]})
|
+({5'h0,i_b[12]}+{5'h0,i_b[13]}+{5'h0,i_b[14]}+{5'h0,i_b[15]})
|
+({5'h0,i_b[12]}+{5'h0,i_b[13]}+{5'h0,i_b[14]}+{5'h0,i_b[15]})
|
+({5'h0,i_b[16]}+{5'h0,i_b[17]}+{5'h0,i_b[18]}+{5'h0,i_b[19]})
|
+({5'h0,i_b[16]}+{5'h0,i_b[17]}+{5'h0,i_b[18]}+{5'h0,i_b[19]})
|
+({5'h0,i_b[20]}+{5'h0,i_b[21]}+{5'h0,i_b[22]}+{5'h0,i_b[23]})
|
+({5'h0,i_b[20]}+{5'h0,i_b[21]}+{5'h0,i_b[22]}+{5'h0,i_b[23]})
|
+({5'h0,i_b[24]}+{5'h0,i_b[25]}+{5'h0,i_b[26]}+{5'h0,i_b[27]})
|
+({5'h0,i_b[24]}+{5'h0,i_b[25]}+{5'h0,i_b[26]}+{5'h0,i_b[27]})
|
+({5'h0,i_b[28]}+{5'h0,i_b[29]}+{5'h0,i_b[30]}+{5'h0,i_b[31]});
|
+({5'h0,i_b[28]}+{5'h0,i_b[29]}+{5'h0,i_b[30]}+{5'h0,i_b[31]});
|
assign w_popc_result[31:6] = 26'h00;
|
assign w_popc_result[31:6] = 26'h00;
|
|
|
// Prelogic for our flags registers
|
// Prelogic for our flags registers
|
wire z, n, v;
|
wire z, n, v;
|
reg c, pre_sign, set_ovfl;
|
reg c, pre_sign, set_ovfl;
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
if (i_ce) // 1 LUT
|
if (i_ce) // 1 LUT
|
set_ovfl =(((i_op==4'h0)&&(i_a[31] != i_b[31]))//SUB&CMP
|
set_ovfl =(((i_op==4'h0)&&(i_a[31] != i_b[31]))//SUB&CMP
|
||((i_op==4'h2)&&(i_a[31] == i_b[31])) // ADD
|
||((i_op==4'h2)&&(i_a[31] == i_b[31])) // ADD
|
||(i_op == 4'h6) // LSL
|
||(i_op == 4'h6) // LSL
|
||(i_op == 4'h5)); // LSR
|
||(i_op == 4'h5)); // LSR
|
|
|
`ifdef LONG_MPY
|
`ifdef LONG_MPY
|
reg mpyhi;
|
reg mpyhi;
|
wire mpybusy;
|
wire mpybusy;
|
`endif
|
`endif
|
|
|
// A 4-way multiplexer can be done in one 6-LUT.
|
// A 4-way multiplexer can be done in one 6-LUT.
|
// A 16-way multiplexer can therefore be done in 4x 6-LUT's with
|
// A 16-way multiplexer can therefore be done in 4x 6-LUT's with
|
// the Xilinx multiplexer fabric that follows.
|
// the Xilinx multiplexer fabric that follows.
|
// Given that we wish to apply this multiplexer approach to 33-bits,
|
// Given that we wish to apply this multiplexer approach to 33-bits,
|
// this will cost a minimum of 132 6-LUTs.
|
// this will cost a minimum of 132 6-LUTs.
|
generate
|
generate
|
if (IMPLEMENT_MPY == 0)
|
if (IMPLEMENT_MPY == 0)
|
begin
|
begin
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
if (i_ce)
|
if (i_ce)
|
begin
|
begin
|
pre_sign <= (i_a[31]);
|
pre_sign <= (i_a[31]);
|
c <= 1'b0;
|
c <= 1'b0;
|
casez(i_op)
|
casez(i_op)
|
4'b0000:{c,o_c } <= {1'b0,i_a}-{1'b0,i_b};// CMP/SUB
|
4'b0000:{c,o_c } <= {1'b0,i_a}-{1'b0,i_b};// CMP/SUB
|
4'b0001: o_c <= i_a & i_b; // BTST/And
|
4'b0001: o_c <= i_a & i_b; // BTST/And
|
4'b0010:{c,o_c } <= i_a + i_b; // Add
|
4'b0010:{c,o_c } <= i_a + i_b; // Add
|
4'b0011: o_c <= i_a | i_b; // Or
|
4'b0011: o_c <= i_a | i_b; // Or
|
4'b0100: o_c <= i_a ^ i_b; // Xor
|
4'b0100: o_c <= i_a ^ i_b; // Xor
|
4'b0101:{o_c,c } <= w_lsr_result[32:0]; // LSR
|
4'b0101:{o_c,c } <= w_lsr_result[32:0]; // LSR
|
4'b0110:{c,o_c } <= (|i_b[31:5])? 33'h00 : {1'b0, i_a } << i_b[4:0]; // LSL
|
4'b0110:{c,o_c } <= (|i_b[31:5])? 33'h00 : {1'b0, i_a } << i_b[4:0]; // LSL
|
4'b0111:{o_c,c } <= w_asr_result[32:0]; // ASR
|
4'b0111:{o_c,c } <= w_asr_result[32:0]; // ASR
|
`ifndef LONG_MPY
|
`ifndef LONG_MPY
|
4'b1000: o_c <= { i_b[15: 0], i_a[15:0] }; // LODIHI
|
4'b1000: o_c <= { i_b[15: 0], i_a[15:0] }; // LODIHI
|
`endif
|
`endif
|
4'b1001: o_c <= { i_a[31:16], i_b[15:0] }; // LODILO
|
4'b1001: o_c <= { i_a[31:16], i_b[15:0] }; // LODILO
|
// 4'h1010: The unimplemented MPYU,
|
// 4'h1010: The unimplemented MPYU,
|
// 4'h1011: and here for the unimplemented MPYS
|
// 4'h1011: and here for the unimplemented MPYS
|
4'b1100: o_c <= w_brev_result; // BREV
|
4'b1100: o_c <= w_brev_result; // BREV
|
4'b1101: o_c <= w_popc_result; // POPC
|
4'b1101: o_c <= w_popc_result; // POPC
|
4'b1110: o_c <= w_rol_result; // ROL
|
4'b1110: o_c <= w_rol_result; // ROL
|
default: o_c <= i_b; // MOV, LDI
|
default: o_c <= i_b; // MOV, LDI
|
endcase
|
endcase
|
end
|
end
|
|
|
assign o_busy = 1'b0;
|
assign o_busy = 1'b0;
|
|
|
reg r_illegal;
|
reg r_illegal;
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
r_illegal <= (i_ce)&&((i_op == 4'ha)||(i_op == 4'hb)
|
r_illegal <= (i_ce)&&((i_op == 4'ha)||(i_op == 4'hb)
|
`ifdef LONG_MPY
|
`ifdef LONG_MPY
|
||(i_op == 4'h8)
|
||(i_op == 4'h8)
|
`endif
|
`endif
|
);
|
);
|
assign o_illegal = r_illegal;
|
assign o_illegal = r_illegal;
|
end else begin
|
end else begin
|
//
|
//
|
// Multiply pre-logic
|
// Multiply pre-logic
|
//
|
//
|
`ifdef LONG_MPY
|
`ifdef LONG_MPY
|
reg [63:0] r_mpy_result;
|
reg [63:0] r_mpy_result;
|
if (IMPLEMENT_MPY == 1)
|
if (IMPLEMENT_MPY == 1)
|
begin // Our two clock option (one clock extra)
|
begin // Our two clock option (one clock extra)
|
reg signed [64:0] r_mpy_a_input, r_mpy_b_input;
|
reg signed [64:0] r_mpy_a_input, r_mpy_b_input;
|
reg mpypipe, x;
|
reg mpypipe, x;
|
initial mpypipe = 1'b0;
|
initial mpypipe = 1'b0;
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
mpypipe <= (i_ce)&&((i_op[3:1]==3'h5)||(i_op[3:0]==4'h8));
|
mpypipe <= (i_ce)&&((i_op[3:1]==3'h5)||(i_op[3:0]==4'h8));
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
if (i_ce)
|
if (i_ce)
|
begin
|
begin
|
r_mpy_a_input <= {{(33){(i_a[31])&(i_op[0])}},
|
r_mpy_a_input <= {{(33){(i_a[31])&(i_op[0])}},
|
i_a[31:0]};
|
i_a[31:0]};
|
r_mpy_b_input <= {{(33){(i_b[31])&(i_op[0])}},
|
r_mpy_b_input <= {{(33){(i_b[31])&(i_op[0])}},
|
i_b[31:0]};
|
i_b[31:0]};
|
end
|
end
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
if (mpypipe)
|
if (mpypipe)
|
{x, r_mpy_result} = r_mpy_a_input
|
{x, r_mpy_result} = r_mpy_a_input
|
* r_mpy_b_input;
|
* r_mpy_b_input;
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
if (i_ce)
|
if (i_ce)
|
mpyhi = i_op[1];
|
mpyhi = i_op[1];
|
assign mpybusy = mpypipe;
|
assign mpybusy = mpypipe;
|
end else if (IMPLEMENT_MPY == 2)
|
end else if (IMPLEMENT_MPY == 2)
|
begin // The three clock option
|
begin // The three clock option
|
reg [31:0] r_mpy_a_input, r_mpy_b_input;
|
reg [31:0] r_mpy_a_input, r_mpy_b_input;
|
reg r_mpy_signed;
|
reg r_mpy_signed;
|
reg [1:0] mpypipe;
|
reg [1:0] mpypipe;
|
|
|
// First clock, latch in the inputs
|
// First clock, latch in the inputs
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
begin
|
begin
|
// mpypipe indicates we have a multiply in the
|
// mpypipe indicates we have a multiply in the
|
// pipeline. In this case, the multiply
|
// pipeline. In this case, the multiply
|
// pipeline is a two stage pipeline, so we need
|
// pipeline is a two stage pipeline, so we need
|
// two bits in the pipe.
|
// two bits in the pipe.
|
mpypipe[0] <= (i_ce)&&((i_op[3:1]==3'h5)
|
mpypipe[0] <= (i_ce)&&((i_op[3:1]==3'h5)
|
||(i_op[3:0]==4'h8));
|
||(i_op[3:0]==4'h8));
|
mpypipe[1] <= mpypipe[0];
|
mpypipe[1] <= mpypipe[0];
|
|
|
if (i_op[0]) // i.e. if signed multiply
|
if (i_op[0]) // i.e. if signed multiply
|
begin
|
begin
|
r_mpy_a_input <= {(~i_a[31]),i_a[30:0]};
|
r_mpy_a_input <= {(~i_a[31]),i_a[30:0]};
|
r_mpy_b_input <= {(~i_b[31]),i_b[30:0]};
|
r_mpy_b_input <= {(~i_b[31]),i_b[30:0]};
|
end else begin
|
end else begin
|
r_mpy_a_input <= i_a[31:0];
|
r_mpy_a_input <= i_a[31:0];
|
r_mpy_b_input <= i_b[31:0];
|
r_mpy_b_input <= i_b[31:0];
|
end
|
end
|
// The signed bit really only matters in the
|
// The signed bit really only matters in the
|
// case of 64 bit multiply. We'll keep track
|
// case of 64 bit multiply. We'll keep track
|
// of it, though, and pretend in all other
|
// of it, though, and pretend in all other
|
// cases.
|
// cases.
|
r_mpy_signed <= i_op[0];
|
r_mpy_signed <= i_op[0];
|
|
|
if (i_ce)
|
if (i_ce)
|
mpyhi = i_op[1];
|
mpyhi = i_op[1];
|
end
|
end
|
|
|
assign mpybusy = |mpypipe;
|
assign mpybusy = |mpypipe;
|
|
|
// Second clock, do the multiplies, get the "partial
|
// Second clock, do the multiplies, get the "partial
|
// products". Here, we break our input up into two
|
// products". Here, we break our input up into two
|
// halves,
|
// halves,
|
//
|
//
|
// A = (2^16 ah + al)
|
// A = (2^16 ah + al)
|
// B = (2^16 bh + bl)
|
// B = (2^16 bh + bl)
|
//
|
//
|
// and use these to compute partial products.
|
// and use these to compute partial products.
|
//
|
//
|
// AB = (2^32 ah*bh + 2^16 (ah*bl + al*bh) + (al*bl)
|
// AB = (2^32 ah*bh + 2^16 (ah*bl + al*bh) + (al*bl)
|
//
|
//
|
// Since we're following the FOIL algorithm to get here,
|
// Since we're following the FOIL algorithm to get here,
|
// we'll name these partial products according to FOIL.
|
// we'll name these partial products according to FOIL.
|
//
|
//
|
// The trick is what happens if A or B is signed. In
|
// The trick is what happens if A or B is signed. In
|
// those cases, the real value of A will not be given by
|
// those cases, the real value of A will not be given by
|
// A = (2^16 ah + al)
|
// A = (2^16 ah + al)
|
// but rather
|
// but rather
|
// A = (2^16 ah[31^] + al) - 2^31
|
// A = (2^16 ah[31^] + al) - 2^31
|
// (where we have flipped the sign bit of A)
|
// (where we have flipped the sign bit of A)
|
// and so ...
|
// and so ...
|
//
|
//
|
// AB= (2^16 ah + al - 2^31) * (2^16 bh + bl - 2^31)
|
// AB= (2^16 ah + al - 2^31) * (2^16 bh + bl - 2^31)
|
// = 2^32(ah*bh)
|
// = 2^32(ah*bh)
|
// +2^16 (ah*bl+al*bh)
|
// +2^16 (ah*bl+al*bh)
|
// +(al*bl)
|
// +(al*bl)
|
// - 2^31 (2^16 bh+bl + 2^16 ah+al)
|
// - 2^31 (2^16 bh+bl + 2^16 ah+al)
|
// - 2^62
|
// - 2^62
|
// = 2^32(ah*bh)
|
// = 2^32(ah*bh)
|
// +2^16 (ah*bl+al*bh)
|
// +2^16 (ah*bl+al*bh)
|
// +(al*bl)
|
// +(al*bl)
|
// - 2^31 (2^16 bh+bl + 2^16 ah+al + 2^31)
|
// - 2^31 (2^16 bh+bl + 2^16 ah+al + 2^31)
|
//
|
//
|
reg [31:0] pp_f, pp_l; // pp_o, pp_i, pp_l;
|
reg [31:0] pp_f, pp_l; // F and L from FOIL
|
reg [32:0] pp_oi;
|
reg [32:0] pp_oi; // The O and I from FOIL
|
reg [32:0] pp_s;
|
reg [32:0] pp_s;
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
begin
|
begin
|
pp_f<=r_mpy_a_input[31:16]*r_mpy_b_input[31:16];
|
pp_f<=r_mpy_a_input[31:16]*r_mpy_b_input[31:16];
|
pp_oi<=r_mpy_a_input[31:16]*r_mpy_b_input[15: 0]
|
pp_oi<=r_mpy_a_input[31:16]*r_mpy_b_input[15: 0]
|
+ r_mpy_a_input[15: 0]*r_mpy_b_input[31:16];
|
+ r_mpy_a_input[15: 0]*r_mpy_b_input[31:16];
|
pp_l<=r_mpy_a_input[15: 0]*r_mpy_b_input[15: 0];
|
pp_l<=r_mpy_a_input[15: 0]*r_mpy_b_input[15: 0];
|
// And a special one for the sign
|
// And a special one for the sign
|
if (r_mpy_signed)
|
if (r_mpy_signed)
|
pp_s <= 32'h8000_0000-(
|
pp_s <= 32'h8000_0000-(
|
r_mpy_a_input[31:0]
|
r_mpy_a_input[31:0]
|
+ r_mpy_b_input[31:0]);
|
+ r_mpy_b_input[31:0]);
|
else
|
else
|
pp_s <= 33'h0;
|
pp_s <= 33'h0;
|
end
|
end
|
|
|
// Third clock, add the results and produce a product
|
// Third clock, add the results and produce a product
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
begin
|
begin
|
r_mpy_result[15:0] <= pp_l[15:0];
|
r_mpy_result[15:0] <= pp_l[15:0];
|
r_mpy_result[63:16] <=
|
r_mpy_result[63:16] <=
|
{ 32'h00, pp_l[31:16] }
|
{ 32'h00, pp_l[31:16] }
|
+ { 15'h00, pp_oi }
|
+ { 15'h00, pp_oi }
|
+ { pp_s, 15'h00 }
|
+ { pp_s, 15'h00 }
|
+ { pp_f, 16'h00 };
|
+ { pp_f, 16'h00 };
|
end
|
end
|
end // Fourth clock -- results are available for writeback.
|
end // Fourth clock -- results are available for writeback.
|
`else
|
`else
|
wire signed [16:0] w_mpy_a_input, w_mpy_b_input;
|
wire signed [16:0] w_mpy_a_input, w_mpy_b_input;
|
wire [33:0] w_mpy_result;
|
wire [33:0] w_mpy_result;
|
reg [31:0] r_mpy_result;
|
reg [31:0] r_mpy_result;
|
assign w_mpy_a_input ={ ((i_a[15])&(i_op[0])), i_a[15:0] };
|
assign w_mpy_a_input ={ ((i_a[15])&(i_op[0])), i_a[15:0] };
|
assign w_mpy_b_input ={ ((i_b[15])&(i_op[0])), i_b[15:0] };
|
assign w_mpy_b_input ={ ((i_b[15])&(i_op[0])), i_b[15:0] };
|
assign w_mpy_result = w_mpy_a_input * w_mpy_b_input;
|
assign w_mpy_result = w_mpy_a_input * w_mpy_b_input;
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
if (i_ce)
|
if (i_ce)
|
r_mpy_result = w_mpy_result[31:0];
|
r_mpy_result = w_mpy_result[31:0];
|
`endif
|
`endif
|
|
|
//
|
//
|
// The master ALU case statement
|
// The master ALU case statement
|
//
|
//
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
if (i_ce)
|
if (i_ce)
|
begin
|
begin
|
pre_sign <= (i_a[31]);
|
pre_sign <= (i_a[31]);
|
c <= 1'b0;
|
c <= 1'b0;
|
casez(i_op)
|
casez(i_op)
|
4'b0000:{c,o_c } <= {1'b0,i_a}-{1'b0,i_b};// CMP/SUB
|
4'b0000:{c,o_c } <= {1'b0,i_a}-{1'b0,i_b};// CMP/SUB
|
4'b0001: o_c <= i_a & i_b; // BTST/And
|
4'b0001: o_c <= i_a & i_b; // BTST/And
|
4'b0010:{c,o_c } <= i_a + i_b; // Add
|
4'b0010:{c,o_c } <= i_a + i_b; // Add
|
4'b0011: o_c <= i_a | i_b; // Or
|
4'b0011: o_c <= i_a | i_b; // Or
|
4'b0100: o_c <= i_a ^ i_b; // Xor
|
4'b0100: o_c <= i_a ^ i_b; // Xor
|
4'b0101:{o_c,c } <= w_lsr_result[32:0]; // LSR
|
4'b0101:{o_c,c } <= w_lsr_result[32:0]; // LSR
|
4'b0110:{c,o_c } <= (|i_b[31:5])? 33'h00 : {1'b0, i_a } << i_b[4:0]; // LSL
|
4'b0110:{c,o_c } <= (|i_b[31:5])? 33'h00 : {1'b0, i_a } << i_b[4:0]; // LSL
|
4'b0111:{o_c,c } <= w_asr_result[32:0]; // ASR
|
4'b0111:{o_c,c } <= w_asr_result[32:0]; // ASR
|
`ifdef LONG_MPY
|
`ifdef LONG_MPY
|
4'b1000: o_c <= r_mpy_result[31:0]; // MPY
|
4'b1000: o_c <= r_mpy_result[31:0]; // MPY
|
`else
|
`else
|
4'b1000: o_c <= { i_b[15: 0], i_a[15:0] }; // LODIHI
|
4'b1000: o_c <= { i_b[15: 0], i_a[15:0] }; // LODIHI
|
`endif
|
`endif
|
4'b1001: o_c <= { i_a[31:16], i_b[15:0] }; // LODILO
|
4'b1001: o_c <= { i_a[31:16], i_b[15:0] }; // LODILO
|
`ifdef LONG_MPY
|
`ifdef LONG_MPY
|
4'b1010: o_c <= r_mpy_result[63:32]; // MPYHU
|
4'b1010: o_c <= r_mpy_result[63:32]; // MPYHU
|
4'b1011: o_c <= r_mpy_result[63:32]; // MPYHS
|
4'b1011: o_c <= r_mpy_result[63:32]; // MPYHS
|
`else
|
`else
|
4'b1010: o_c <= r_mpy_result; // MPYU
|
4'b1010: o_c <= r_mpy_result; // MPYU
|
4'b1011: o_c <= r_mpy_result; // MPYS
|
4'b1011: o_c <= r_mpy_result; // MPYS
|
`endif
|
`endif
|
4'b1100: o_c <= w_brev_result; // BREV
|
4'b1100: o_c <= w_brev_result; // BREV
|
4'b1101: o_c <= w_popc_result; // POPC
|
4'b1101: o_c <= w_popc_result; // POPC
|
4'b1110: o_c <= w_rol_result; // ROL
|
4'b1110: o_c <= w_rol_result; // ROL
|
default: o_c <= i_b; // MOV, LDI
|
default: o_c <= i_b; // MOV, LDI
|
endcase
|
endcase
|
end else if (r_busy)
|
end else if (r_busy)
|
`ifdef LONG_MPY
|
`ifdef LONG_MPY
|
o_c <= (mpyhi)?r_mpy_result[63:32]:r_mpy_result[31:0];
|
o_c <= (mpyhi)?r_mpy_result[63:32]:r_mpy_result[31:0];
|
`else
|
`else
|
o_c <= r_mpy_result;
|
o_c <= r_mpy_result;
|
`endif
|
`endif
|
|
|
reg r_busy;
|
reg r_busy;
|
initial r_busy = 1'b0;
|
initial r_busy = 1'b0;
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
r_busy <= (~i_rst)&&(i_ce)&&(i_valid)
|
r_busy <= (~i_rst)&&(i_ce)&&(i_valid)
|
`ifdef LONG_MPY
|
`ifdef LONG_MPY
|
&&((i_op[3:1] == 3'h5)
|
&&((i_op[3:1] == 3'h5)
|
||(i_op[3:0] == 4'h8))||mpybusy;
|
||(i_op[3:0] == 4'h8))||mpybusy;
|
`else
|
`else
|
&&(i_op[3:1] == 3'h5);
|
&&(i_op[3:1] == 3'h5);
|
`endif
|
`endif
|
|
|
assign o_busy = r_busy;
|
assign o_busy = r_busy;
|
|
|
assign o_illegal = 1'b0;
|
assign o_illegal = 1'b0;
|
end endgenerate
|
end endgenerate
|
|
|
assign z = (o_c == 32'h0000);
|
assign z = (o_c == 32'h0000);
|
assign n = (o_c[31]);
|
assign n = (o_c[31]);
|
assign v = (set_ovfl)&&(pre_sign != o_c[31]);
|
assign v = (set_ovfl)&&(pre_sign != o_c[31]);
|
|
|
assign o_f = { v, n, c, z };
|
assign o_f = { v, n, c, z };
|
|
|
initial o_valid = 1'b0;
|
initial o_valid = 1'b0;
|
always @(posedge i_clk)
|
always @(posedge i_clk)
|
if (i_rst)
|
if (i_rst)
|
o_valid <= 1'b0;
|
o_valid <= 1'b0;
|
else
|
else
|
o_valid <= (i_ce)&&(i_valid)
|
o_valid <= (i_ce)&&(i_valid)
|
`ifdef LONG_MPY
|
`ifdef LONG_MPY
|
&&(i_op[3:1] != 3'h5)&&(i_op[3:0] != 4'h8)
|
&&(i_op[3:1] != 3'h5)&&(i_op[3:0] != 4'h8)
|
||(o_busy)&&(~mpybusy);
|
||(o_busy)&&(~mpybusy);
|
`else
|
`else
|
&&(i_op[3:1] != 3'h5)||(o_busy);
|
&&(i_op[3:1] != 3'h5)||(o_busy);
|
`endif
|
`endif
|
endmodule
|
endmodule
|
|
|
