Subversion Repositories sparc64soc

// ========== Copyright Header Begin ==========================================

// 

// OpenSPARC T1 Processor File: sparc_exu_eclccr.v

// Copyright (c) 2006 Sun Microsystems, Inc.  All Rights Reserved.

// DO NOT ALTER OR REMOVE COPYRIGHT NOTICES.

// 

// The above named program is free software; you can redistribute it and/or

// modify it under the terms of the GNU General Public

// License version 2 as published by the Free Software Foundation.

// 

// The above named program 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

// General Public License for more details.

// 

// You should have received a copy of the GNU General Public

// License along with this work; if not, write to the Free Software

// Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA.

// 

// ========== Copyright Header End ============================================

////////////////////////////////////////////////////////////////////////

/*

//  Module Name: sparc_exu_eclccr

//      Description: 4 bit condition code registers with forwarding.  Takes

//      the e_stage result and writes on the w stage.

*/

module sparc_exu_eclccr (/*AUTOARG*/

   // Outputs

   exu_ifu_cc_d, exu_tlu_ccr0_w, exu_tlu_ccr1_w, exu_tlu_ccr2_w,

   exu_tlu_ccr3_w,

   // Inputs

   clk, se, alu_xcc_e, alu_icc_e, tid_d, thrdec_d, thr_match_dm,

   thr_match_de, tid_w, thr_w, ifu_exu_kill_e, ifu_exu_setcc_d,

   byp_ecl_wrccr_data_w, wb_ccr_wrccr_w, wb_ccr_setcc_g,

   divcntl_ccr_cc_w2, wb_ccr_thr_g, tlu_exu_cwpccr_update_m,

   tlu_exu_ccr_m, ifu_exu_inst_vld_w, ifu_tlu_flush_w, early_flush_w

   ) ;

   input clk;

   input se;

   input [3:0] alu_xcc_e;    // condition codes from the alu

   input [3:0] alu_icc_e;

   input [1:0] tid_d;   // thread for each stage

   input [3:0] thrdec_d;   // decoded tid_d for mux select

   input       thr_match_dm;

   input       thr_match_de;

   input [1:0] tid_w;

   input [3:0] thr_w;        // decoded tid_w

   input       ifu_exu_kill_e;

   input       ifu_exu_setcc_d;

   input [7:0] byp_ecl_wrccr_data_w;// for the WRCCR operation (LSBs of

   input       wb_ccr_wrccr_w; // ALU result) + wen signal

   input       wb_ccr_setcc_g;

   input [7:0] divcntl_ccr_cc_w2;

   input [1:0] wb_ccr_thr_g;

   input       tlu_exu_cwpccr_update_m;

   input [7:0] tlu_exu_ccr_m;

   input       ifu_exu_inst_vld_w;

   input       ifu_tlu_flush_w;

   input       early_flush_w;

   output [7:0] exu_ifu_cc_d;   // condition codes for current thread

   output [7:0] exu_tlu_ccr0_w;

   output [7:0] exu_tlu_ccr1_w;

   output [7:0] exu_tlu_ccr2_w;

   output [7:0] exu_tlu_ccr3_w;

   wire [7:0]   partial_cc_d;   // partial bypassed ccr

   wire [7:0]   alu_cc_e;   // alu combined condition codes

   wire [7:0]   alu_cc_m;   // m stage alu ccs

   wire [7:0]   alu_cc_w;

   wire [7:0]   exu_ifu_cc_w;   // writeback data

   wire         setcc_e;        // from previous stage

   wire         setcc_m;

   wire         setcc_w;

   wire         valid_setcc_e;  // after comparing with kill

   wire         valid_setcc_m;

   wire         valid_setcc_w;

   wire         setcc_w2;

   wire [7:0]   ccrin_thr0;

   wire [7:0]   ccrin_thr1;

   wire [7:0]   ccrin_thr2;

   wire [7:0]   ccrin_thr3;

   wire [7:0]   ccr_d;

   wire [7:0]   ccr_thr0;

   wire [7:0]   ccr_thr1;

   wire [7:0]   ccr_thr2;

   wire [7:0]   ccr_thr3;

   wire         use_alu_cc;

   wire         use_ccr;

   wire         use_cc_e;

   wire         use_cc_m;

   wire         use_cc_w;

   wire  [1:0]   tid_dxorw;

   wire         thr_match_de;

   wire         thrmatch_w;

   wire [1:0]   thr_w2;

   wire          thr0_w2;

   wire          thr1_w2;

   wire          thr2_w2;

   wire          thr3_w2;

   wire          wen_thr0_w;    // write enable for each input/thread

   wire          wen_thr0_w2;

   wire          wen_thr1_w;

   wire          wen_thr1_w2;

   wire          wen_thr2_w;

   wire          wen_thr2_w2;

   wire          wen_thr3_w;

   wire          wen_thr3_w2;

   wire          wen_thr0_l;      // overall write enable for each thread

   wire          wen_thr1_l;

   wire          wen_thr2_l;

   wire          wen_thr3_l;

   wire          bypass_cc_w;

   wire [7:0]    ccr_m;

   // D2E flops

   dff_s dff_setcc_d2e(.din(ifu_exu_setcc_d), .clk(clk), .q(setcc_e),

                     .se(se), .si(), .so());

   // E stage

   assign       alu_cc_e = {alu_xcc_e, alu_icc_e};

   assign       valid_setcc_e = setcc_e & ~ifu_exu_kill_e;

   dff_s #(8) dff_cc_e2m(.din(alu_cc_e[7:0]), .clk(clk), .q(alu_cc_m[7:0]),

                  .se(se), .si(), .so());

   dff_s dff_setcc_e2m(.din(valid_setcc_e), .clk(clk), .q(setcc_m),

                     .se(se), .si(), .so());

   // M stage

   assign       valid_setcc_m = setcc_m | tlu_exu_cwpccr_update_m;

   mux2ds #(8) mux_ccr_m(.dout(ccr_m[7:0]),

                            .in0(alu_cc_m[7:0]),

                            .in1(tlu_exu_ccr_m[7:0]),

                            .sel0(~tlu_exu_cwpccr_update_m),

                            .sel1(tlu_exu_cwpccr_update_m));

   dff_s #(8) dff_cc_m2w(.din(ccr_m[7:0]), .clk(clk), .q(alu_cc_w[7:0]),

                  .se(se), .si(), .so());

   dff_s dff_setcc_m2w(.din(valid_setcc_m), .clk(clk), .q(setcc_w),

                     .se(se), .si(), .so());

   // W stage

   assign bypass_cc_w = ifu_exu_inst_vld_w & setcc_w;

   assign valid_setcc_w = ~ifu_tlu_flush_w & ~early_flush_w & ifu_exu_inst_vld_w & (setcc_w | wb_ccr_wrccr_w);

   // mux with wrccr

   assign        use_alu_cc = ~(wb_ccr_wrccr_w);

   mux2ds #(8) mux_ccrin_cc(.dout(exu_ifu_cc_w[7:0]), .sel0(wb_ccr_wrccr_w),

                          .sel1(use_alu_cc),

                          .in0(byp_ecl_wrccr_data_w[7:0]),

                          .in1(alu_cc_w[7:0]));

   dff_s #(3) setcc_g2w2 (.din({wb_ccr_setcc_g, wb_ccr_thr_g[1:0]}), .clk(clk),

                        .q({setcc_w2, thr_w2[1:0]}),

                        .se(se), .si(), .so());

   /////////////////////////

   // Storage of ccr

   /////////////////////////

`ifdef FPGA_SYN_1THREAD

   assign          thr0_w2 = ~thr_w2[1] & ~thr_w2[0];

   assign          wen_thr0_w = (thr_w[0] & valid_setcc_w & ~wen_thr0_w2);

   assign          wen_thr0_w2 = thr0_w2 & setcc_w2;

   assign          wen_thr0_l = ~(wen_thr0_w | wen_thr0_w2);

   // mux between cc_w, cc_w2, old value, tlu value

   mux3ds #(8) mux_ccrin0(.dout(ccrin_thr0[7:0]), .sel0(wen_thr0_w),

                          .sel1(wen_thr0_w2), .sel2(wen_thr0_l),

                          .in0(exu_ifu_cc_w[7:0]),

                          .in1(divcntl_ccr_cc_w2[7:0]), .in2(ccr_thr0[7:0]));

   // store new value

   dff_s #(8) dff_ccr_thr0(.din(ccrin_thr0[7:0]), .clk(clk), .q(ccr_thr0[7:0]),

                       .se(se), .si(), .so());

   assign          ccr_d[7:0] = ccr_thr0[7:0];

`else // !`ifdef FPGA_SYN_1THREAD

   // decode thr_w2 for mux select

   assign        thr0_w2 = ~thr_w2[1] & ~thr_w2[0];

   assign        thr1_w2 = ~thr_w2[1] & thr_w2[0];

   assign        thr2_w2 = thr_w2[1] & ~thr_w2[0];

   assign        thr3_w2 = thr_w2[1] & thr_w2[0];

   // enable input for each thread

   assign        wen_thr0_w = (thr_w[0] & valid_setcc_w & ~wen_thr0_w2);

   assign        wen_thr0_w2 = thr0_w2 & setcc_w2;

   assign        wen_thr0_l = ~(wen_thr0_w | wen_thr0_w2);

   assign        wen_thr1_w = (thr_w[1] & valid_setcc_w & ~wen_thr1_w2);

   assign        wen_thr1_w2 = (thr1_w2 & setcc_w2);

   assign        wen_thr1_l = ~(wen_thr1_w | wen_thr1_w2);

   assign        wen_thr2_w = (thr_w[2] & valid_setcc_w & ~wen_thr2_w2);

   assign        wen_thr2_w2 = (thr2_w2 & setcc_w2);

   assign        wen_thr2_l = ~(wen_thr2_w | wen_thr2_w2);

   assign        wen_thr3_w = (thr_w[3] & valid_setcc_w & ~wen_thr3_w2);

   assign        wen_thr3_w2 = (thr3_w2 & setcc_w2);

   assign        wen_thr3_l = ~(wen_thr3_w | wen_thr3_w2);

   // mux between cc_w, cc_w2, old value, tlu value

   mux3ds #(8) mux_ccrin0(.dout(ccrin_thr0[7:0]), .sel0(wen_thr0_w),

                          .sel1(wen_thr0_w2), .sel2(wen_thr0_l),

                          .in0(exu_ifu_cc_w[7:0]),

                          .in1(divcntl_ccr_cc_w2[7:0]), .in2(ccr_thr0[7:0]));

   mux3ds #(8) mux_ccrin1(.dout(ccrin_thr1[7:0]), .sel0(wen_thr1_w),

                          .sel1(wen_thr1_w2), .sel2(wen_thr1_l),

                          .in0(exu_ifu_cc_w[7:0]),

                          .in1(divcntl_ccr_cc_w2[7:0]), .in2(ccr_thr1[7:0]));

   mux3ds #(8) mux_ccrin2(.dout(ccrin_thr2[7:0]), .sel0(wen_thr2_w),

                          .sel1(wen_thr2_w2), .sel2(wen_thr2_l),

                          .in0(exu_ifu_cc_w[7:0]),

                          .in1(divcntl_ccr_cc_w2[7:0]), .in2(ccr_thr2[7:0]));

   mux3ds #(8) mux_ccrin3(.dout(ccrin_thr3[7:0]), .sel0(wen_thr3_w),

                          .sel1(wen_thr3_w2), .sel2(wen_thr3_l),

                          .in0(exu_ifu_cc_w[7:0]),

                          .in1(divcntl_ccr_cc_w2[7:0]), .in2(ccr_thr3[7:0]));

   // store new value

   dff_s #(8) dff_ccr_thr0(.din(ccrin_thr0[7:0]), .clk(clk), .q(ccr_thr0[7:0]),

                       .se(se), .si(), .so());

   dff_s #(8) dff_ccr_thr1(.din(ccrin_thr1[7:0]), .clk(clk), .q(ccr_thr1[7:0]),

                       .se(se), .si(), .so());

   dff_s #(8) dff_ccr_thr2(.din(ccrin_thr2[7:0]), .clk(clk), .q(ccr_thr2[7:0]),

                       .se(se), .si(), .so());

   dff_s #(8) dff_ccr_thr3(.din(ccrin_thr3[7:0]), .clk(clk), .q(ccr_thr3[7:0]),

                       .se(se), .si(), .so());

   // mux between the 4 sets of ccrs

   mux4ds #(8) mux_ccr_out(.dout(ccr_d[7:0]), .sel0(thrdec_d[0]),

                         .sel1(thrdec_d[1]), .sel2(thrdec_d[2]),

                         .sel3(thrdec_d[3]), .in0(ccr_thr0[7:0]),

                         .in1(ccr_thr1[7:0]), .in2(ccr_thr2[7:0]),

                         .in3(ccr_thr3[7:0]));

`endif // !`ifdef FPGA_SYN_1THREAD

   // bypass the ccs to the output.  Only alu result needs to be bypassed

   assign        exu_ifu_cc_d[7:0] = (use_cc_e)? alu_cc_e[7:0]: partial_cc_d[7:0];

   mux3ds #(8) mux_ccr_bypass1(.dout(partial_cc_d[7:0]),

                               .sel0(use_ccr),

                               .sel1(use_cc_m),

                               .sel2(use_cc_w),

                               .in0(ccr_d[7:0]),

                               .in1(alu_cc_m[7:0]),

                               .in2(alu_cc_w[7:0]));

   assign        use_cc_e = valid_setcc_e & thr_match_de;

   assign        use_cc_m = setcc_m & thr_match_dm;

   assign        use_cc_w = bypass_cc_w & thrmatch_w & ~use_cc_m;

   assign        use_ccr = ~(use_cc_m | use_cc_w);

   assign        tid_dxorw = tid_w ^ tid_d;

   assign        thrmatch_w = ~(tid_dxorw[1] | tid_dxorw[0]);

   // generate ccr_w for the tlu

   assign        exu_tlu_ccr0_w[7:0] = ccr_thr0[7:0];

   assign        exu_tlu_ccr1_w[7:0] = ccr_thr1[7:0];

   assign        exu_tlu_ccr2_w[7:0] = ccr_thr2[7:0];

   assign        exu_tlu_ccr3_w[7:0] = ccr_thr3[7:0];

endmodule // sparc_exu_eclccr