Subversion Repositories sparc64soc

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

// 

// OpenSPARC T1 Processor File: sparc_ffu_ctl_visctl.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_ffu_ctl_visctl

//      Description: This is the ffu vis control block.

*/

module sparc_ffu_ctl_visctl (/*AUTOARG*/

   // Outputs

   ctl_vis_sel_add, ctl_vis_sel_log, ctl_vis_sel_align,

   ctl_vis_add32, ctl_vis_subtract, ctl_vis_cin, ctl_vis_align0,

   ctl_vis_align2, ctl_vis_align4, ctl_vis_align6, ctl_vis_align_odd,

   ctl_vis_log_sel_pass, ctl_vis_log_sel_nand, ctl_vis_log_sel_nor,

   ctl_vis_log_sel_xor, ctl_vis_log_invert_rs1,

   ctl_vis_log_invert_rs2, ctl_vis_log_constant,

   ctl_vis_log_pass_const, ctl_vis_log_pass_rs1,

   ctl_vis_log_pass_rs2, vis_result, illegal_vis_e, vis_nofrf_e,

   visop_m, visop_w_vld, vis_wen_next, fpu_rnd,

   ffu_exu_rsr_data_hi_m, ffu_exu_rsr_data_mid_m,

   ffu_exu_rsr_data_lo_m, ctl_dp_wsr_data_w2, ctl_dp_gsr_wsr_w2,

   ctl_dp_thr_e,

   // Inputs

   clk, se, reset, opf, tid_w2, tid_e, tid, visop_e, kill_w,

   ifu_tlu_sraddr_d, exu_ffu_wsr_inst_e, exu_ffu_gsr_align_m,

   exu_ffu_gsr_rnd_m, exu_ffu_gsr_mask_m, exu_ffu_gsr_scale_m,

   ifu_ffu_rnd_e, dp_ctl_fsr_rnd, flush_w2, thr_match_mw2,

   thr_match_ww2, ifu_tlu_inst_vld_w, ue_trap_w3, frs1_e, frs2_e,

   frd_e, rollback_c3, rollback_rs2_w2, visop, rollback_rs1_w3,

   dp_ctl_gsr_mask_e, dp_ctl_gsr_scale_e

   ) ;

   input clk;

   input se;

   input reset;

   input [8:0] opf;

   input [1:0] tid_w2;

   input [1:0] tid_e;

   input [1:0] tid;

   input      visop_e;

   input       kill_w;

   input [6:0] ifu_tlu_sraddr_d;

   input       exu_ffu_wsr_inst_e;

   input [2:0] exu_ffu_gsr_align_m;

   input [2:0] exu_ffu_gsr_rnd_m;

   input [31:0] exu_ffu_gsr_mask_m;

   input [4:0]  exu_ffu_gsr_scale_m;

   input [2:0] ifu_ffu_rnd_e;

   input [1:0] dp_ctl_fsr_rnd;

   input      flush_w2;

   input      thr_match_mw2;

   input      thr_match_ww2;

   input      ifu_tlu_inst_vld_w;

   input      ue_trap_w3;

   input [4:0] frs1_e;

   input [4:0] frs2_e;

   input [4:0] frd_e;

   input       rollback_c3;

   input       rollback_rs2_w2;

   input       visop;

   input       rollback_rs1_w3;

   input [31:0] dp_ctl_gsr_mask_e;

   input [4:0]  dp_ctl_gsr_scale_e;

   output      ctl_vis_sel_add;

   output      ctl_vis_sel_log;

   output      ctl_vis_sel_align;

   output      ctl_vis_add32;

   output      ctl_vis_subtract;

   output      ctl_vis_cin;

   output      ctl_vis_align0;

   output      ctl_vis_align2;

   output      ctl_vis_align4;

   output      ctl_vis_align6;

   output      ctl_vis_align_odd;

   output      ctl_vis_log_sel_pass;

   output      ctl_vis_log_sel_nand;

   output      ctl_vis_log_sel_nor;

   output      ctl_vis_log_sel_xor;

   output      ctl_vis_log_invert_rs1;

   output      ctl_vis_log_invert_rs2;

   output      ctl_vis_log_constant;

   output      ctl_vis_log_pass_const;

   output      ctl_vis_log_pass_rs1;

   output      ctl_vis_log_pass_rs2;

   output      vis_result;

   output      illegal_vis_e;

   output      vis_nofrf_e;

   output      visop_m;

   output      visop_w_vld;

   output      vis_wen_next;

   output [1:0] fpu_rnd;

   output [31:0] ffu_exu_rsr_data_hi_m;

   output [2:0]  ffu_exu_rsr_data_mid_m;

   output [7:0] ffu_exu_rsr_data_lo_m;

   output [36:0] ctl_dp_wsr_data_w2;

   output [3:0] ctl_dp_gsr_wsr_w2;

   output [3:0] ctl_dp_thr_e;

   wire         illegal_rs1_e;

   wire         illegal_rs2_e;

   wire         illegal_siam_e;

   wire         rs2_check_nonzero_e;

   wire         rs1_check_nonzero_e;

   wire        visop_e;

   wire        issue_visop_e;

   wire        visop_m;

   wire        visop_w;

   wire        visop_w_vld;

   wire        visop_w2_vld;

   wire        visop_w2;

   wire        visop_w3;

   wire        visop_w3_vld;

   wire        add;

   wire        align;

   wire        logic;

   wire        siam;

   wire        alignaddr;

   wire        opf_log_zero;

   wire        opf_log_one;

   wire        opf_log_src1;

   wire        opf_log_src2;

   wire        opf_log_not1;

   wire        opf_log_not2;

   wire        opf_log_or;

   wire        opf_log_nor;

   wire        opf_log_and;

   wire        opf_log_nand;

   wire        opf_log_xor;

   wire        opf_log_xnor;

   wire        opf_log_ornot1;

   wire        opf_log_ornot2;

   wire        opf_log_andnot1;

   wire        opf_log_andnot2;

   wire        invert_rs1_next;

   wire        invert_rs2_next;

   wire        log_pass_rs1_next;

   wire        log_pass_rs2_next;

   wire        log_pass_rs1;

   wire        log_pass_rs2;

   wire [2:0] t0_gsr_rnd;

   wire [2:0] t1_gsr_rnd;

   wire [2:0] t2_gsr_rnd;

   wire [2:0] t3_gsr_rnd;

   wire [2:0] t0_gsr_align;

   wire [2:0] t1_gsr_align;

   wire [2:0] t2_gsr_align;

   wire [2:0] t3_gsr_align;

   wire [2:0] t0_gsr_rnd_next;

   wire [2:0] t1_gsr_rnd_next;

   wire [2:0] t2_gsr_rnd_next;

   wire [2:0] t3_gsr_rnd_next;

   wire [2:0] t0_gsr_align_next;

   wire [2:0] t1_gsr_align_next;

   wire [2:0] t2_gsr_align_next;

   wire [2:0] t3_gsr_align_next;

   wire [2:0] gsr_rnd_e;

   wire [2:0] gsr_align_e;

   wire       t0_rnd_wen_l;

   wire       t0_gsr_wsr_w2;

   wire       t0_siam_w2;

   wire       t0_align_wen_l;

   wire       t0_alignaddr_w2;

   wire       t1_rnd_wen_l;

   wire       t1_gsr_wsr_w2;

   wire       t1_siam_w2;

   wire       t1_align_wen_l;

   wire       t1_alignaddr_w2;

   wire       t2_rnd_wen_l;

   wire       t2_gsr_wsr_w2;

   wire       t2_siam_w2;

   wire       t2_align_wen_l;

   wire       t2_alignaddr_w2;

   wire       t3_rnd_wen_l;

   wire       t3_gsr_wsr_w2;

   wire       t3_siam_w2;

   wire       t3_align_wen_l;

   wire       t3_alignaddr_w2;

   wire [2:0] siam_rnd;

   wire [3:0] thr_w2;

   wire [3:0] ctl_dp_thr_e;

   wire [3:0] thr_fp;

   wire       gsr_addr_d;

   wire       gsr_addr_e;

   wire       wgsr_e;

   wire       wgsr_m;

   wire       wgsr_w;

   wire       wgsr_vld_m;

   wire       wgsr_vld_w;

   wire       wgsr_vld_w2;

   wire       wgsr_w2;

   wire [2:0] gsr_rnd;

   wire [1:0] fpu_rnd_next;

   wire [2:0]  gsr_align;

   wire [2:0]  gsr_align_d1;

   wire [2:0] align_addr_data_w2;

   wire [2:0] wgsr_align_offset_w;

   wire [2:0] wgsr_rnd_w;

   wire [2:0] wgsr_align_offset_w2;

   wire [2:0] wgsr_rnd_w2;

   wire [36:0] wsr_data_m;

   wire [36:0] wsr_data_w;

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

   // VIS PIPELINE

   //------------------------------------

   // Note: rs2_ce, rs2_ue, rs1_ue will kill vis instruction

   //       in addition to any traps, etc.

   //       These are incorporated into the "kill" signals

   // E: ren rs2

   // M: ren rs1

   // W: rs2 data ready, check rs2 ecc

   // W2: rs1 data ready, check rs1 ecc

   // W3: execute vis operation (result written to rs2/rd flop)

   // W4: gen ecc and write to frf

   dff_s visop_e2m(.din(issue_visop_e), .clk(clk), .q(visop_m), .si(), .so(), .se(se));

   dff_s visop_m2w(.din(visop_m), .clk(clk), .q(visop_w), .si(), .so(), .se(se));

   dff_s visop_w2w2(.din(visop_w_vld), .clk(clk), .q(visop_w2), .si(), .so(), .se(se));

   dff_s visop_w22w3(.din(visop_w2_vld), .clk(clk), .q(visop_w3), .si(), .so(), .se(se));

   assign     issue_visop_e = visop_e | visop & rollback_c3;

   // only check kills in w since they are accumulated into kill_w

   assign      visop_w_vld = visop_w & ~kill_w;

   assign      visop_w2_vld = visop_w2 & ~flush_w2 & ~rollback_rs2_w2;

   assign      visop_w3_vld = visop_w3 & ~ue_trap_w3 & ~rollback_rs1_w3;

   assign      vis_result = visop_w3_vld;

   assign      vis_wen_next = vis_result & ~siam & ~alignaddr;

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

   // Decode opf

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

   assign      add = ~opf[8] & ~opf[7] & opf[6] & ~opf[5] & opf[4] & ~opf[3];

   assign      align = ~opf[8] & ~opf[7] & opf[6] & ~opf[5] & ~opf[4] & opf[3] & ~opf[2] & ~opf[1] & ~opf[0];

   assign      logic = ~opf[8] & ~opf[7] & opf[6] & opf[5];

   assign siam = ~opf[8] & opf[7] & ~opf[6] & ~opf[5] & ~opf[4] & ~opf[3] & ~opf[2] & ~opf[1] & opf[0];

   assign alignaddr = ~opf[8] & ~opf[7] & ~opf[6] & ~opf[5] & opf[4] & opf[3] & ~opf[2] & ~opf[0]; //alignaddress

   assign illegal_vis_e = (visop_e & ~(add | align | logic | siam | alignaddr) |

                           illegal_rs1_e | illegal_rs2_e | illegal_siam_e);

   assign rs1_check_nonzero_e = visop_e & (siam | (logic & (opf_log_zero | opf_log_one | opf_log_src2 | opf_log_not2)));

   assign rs2_check_nonzero_e = visop_e & logic & (opf_log_zero | opf_log_one | opf_log_src1 | opf_log_not1);

   assign illegal_rs1_e = (frs1_e[4:0] != 5'b00000) & rs1_check_nonzero_e;

   assign illegal_rs2_e = (frs2_e[4:0] != 5'b00000) & rs2_check_nonzero_e;

   assign illegal_siam_e = ((frd_e[4:0] != 5'b00000) | frs2_e[4] | frs2_e[3]) & siam & visop_e;

   assign vis_nofrf_e = visop_e & (siam | alignaddr | opf_log_zero | opf_log_one);

   // controls for add

   // Make subtract come out of its own flop for loading purposes (very critical timing)

   dff_s sub_dff(.din(opf[2]), .clk(clk), .q(ctl_vis_subtract), .se(se), .si(), .so());

   assign ctl_vis_cin = opf[2];

   assign ctl_vis_add32 = opf[1];

   // controls for logic

   assign opf_log_zero = ~opf[4] & ~opf[3] & ~opf[2] & ~opf[1];

   assign opf_log_nor = ~opf[4] & ~opf[3] & ~opf[2] & opf[1];

   assign opf_log_andnot2 = ~opf[4] & ~opf[3] & opf[2] & ~opf[1];

   assign opf_log_not2 = ~opf[4] & ~opf[3] & opf[2] & opf[1];

   assign opf_log_andnot1 = ~opf[4] & opf[3] & ~opf[2] & ~opf[1];

   assign opf_log_not1 = ~opf[4] & opf[3] & ~opf[2] & opf[1];

   assign opf_log_xor = ~opf[4] & opf[3] & opf[2] & ~opf[1];

   assign opf_log_nand = ~opf[4] & opf[3] & opf[2] & opf[1];

   assign opf_log_and = opf[4] & ~opf[3] & ~opf[2] & ~opf[1];

   assign opf_log_xnor = opf[4] & ~opf[3] & ~opf[2] & opf[1];

   assign opf_log_src1 = opf[4] & ~opf[3] & opf[2] & ~opf[1];

   assign opf_log_ornot2 = opf[4] & ~opf[3] & opf[2] & opf[1];

   assign opf_log_src2 = opf[4] & opf[3] & ~opf[2] & ~opf[1];

   assign opf_log_ornot1 = opf[4] & opf[3] & ~opf[2] & opf[1];

   assign opf_log_or = opf[4] & opf[3] & opf[2] & ~opf[1];

   assign opf_log_one = opf[4] & opf[3] & opf[2] & opf[1];

   // selects for logic mux

   assign ctl_vis_log_sel_nand = opf_log_or | opf_log_nand | opf_log_ornot1 | opf_log_ornot2;

   assign ctl_vis_log_sel_xor = opf_log_xor | opf_log_xnor;

   assign ctl_vis_log_sel_nor = opf_log_and | opf_log_nor | opf_log_andnot1 | opf_log_andnot2;

   assign ctl_vis_log_sel_pass = (opf_log_zero | opf_log_one | opf_log_src1 | opf_log_src2 |

                                  opf_log_not1 | opf_log_not2);

   assign invert_rs1_next = (opf_log_not1 | opf_log_or | opf_log_and |

                                    opf_log_ornot2 | opf_log_andnot2);

   assign invert_rs2_next = (opf_log_not2 | opf_log_or | opf_log_and |

                                    opf_log_ornot1 | opf_log_andnot1 | opf_log_xnor);

   dff_s invert_rs1_dff(.din(invert_rs1_next), .clk(clk), .q(ctl_vis_log_invert_rs1),

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

   dff_s invert_rs2_dff(.din(invert_rs2_next), .clk(clk), .q(ctl_vis_log_invert_rs2),

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

   // precalculate to help timing

   assign log_pass_rs1_next = opf_log_src1 | opf_log_not1;

   assign log_pass_rs2_next = opf_log_src2 | opf_log_not2;

   dff_s #(2) log_pass_dff(.din({log_pass_rs1_next,log_pass_rs2_next}), .clk(clk),

                         .q({log_pass_rs1,log_pass_rs2}), .se(se), .si(), .so());

   assign ctl_vis_log_pass_rs1 = log_pass_rs1;

   assign ctl_vis_log_pass_rs2 = log_pass_rs2 & ~log_pass_rs1;

   assign ctl_vis_log_constant = opf_log_one;

   assign ctl_vis_log_pass_const = ~(ctl_vis_log_pass_rs1 | ctl_vis_log_pass_rs2);

   // controls for falign

   assign ctl_vis_align0 = ~gsr_align_d1[2] & ~gsr_align_d1[1];

   assign ctl_vis_align2 = ~gsr_align_d1[2] & gsr_align_d1[1];

   assign ctl_vis_align4 = gsr_align_d1[2] & ~gsr_align_d1[1];

   assign ctl_vis_align6 = gsr_align_d1[2] & gsr_align_d1[1];

   assign ctl_vis_align_odd = gsr_align_d1[0];

   // controls for output mux

   assign ctl_vis_sel_add = add;

   assign ctl_vis_sel_align = align;

   assign ctl_vis_sel_log = ~(add | align);

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

   // GSR.alignaddr_offset, GSR.IM, GSR.IRND

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

   mux4ds #(6) curr_gsr_mux(.dout({gsr_rnd[2:0], gsr_align[2:0]}),

                            .in0({t0_gsr_rnd[2:0], t0_gsr_align[2:0]}),

                            .in1({t1_gsr_rnd[2:0], t1_gsr_align[2:0]}),

                            .in2({t2_gsr_rnd[2:0], t2_gsr_align[2:0]}),

                            .in3({t3_gsr_rnd[2:0], t3_gsr_align[2:0]}),

                            .sel0(thr_fp[0]),

                            .sel1(thr_fp[1]),

                            .sel2(thr_fp[2]),

                            .sel3(thr_fp[3]));

   mux4ds #(6) gsr_e_mux(.dout({gsr_rnd_e[2:0], gsr_align_e[2:0]}),

                            .in0({t0_gsr_rnd[2:0], t0_gsr_align[2:0]}),

                            .in1({t1_gsr_rnd[2:0], t1_gsr_align[2:0]}),

                            .in2({t2_gsr_rnd[2:0], t2_gsr_align[2:0]}),

                            .in3({t3_gsr_rnd[2:0], t3_gsr_align[2:0]}),

                            .sel0(ctl_dp_thr_e[0]),

                            .sel1(ctl_dp_thr_e[1]),

                            .sel2(ctl_dp_thr_e[2]),

                            .sel3(ctl_dp_thr_e[3]));

   dff_s #(43) gsr_e2m(.din({dp_ctl_gsr_mask_e[31:0],gsr_rnd_e[2:0],

                          dp_ctl_gsr_scale_e[4:0],gsr_align_e[2:0]}), .clk(clk),

                    .q({ffu_exu_rsr_data_hi_m[31:0],ffu_exu_rsr_data_mid_m[2:0], ffu_exu_rsr_data_lo_m[7:0]}),

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

   dff_s #(3) gsr_align_dff(.din(gsr_align[2:0]), .clk(clk), .q(gsr_align_d1[2:0]), .se(se), .si(), .so());

   // put in to help timing for sending to lsu

   dff_s #(2) fpu_rnd_dff(.din(fpu_rnd_next[1:0]), .clk(clk), .q(fpu_rnd[1:0]), .si(), .so(), .se(se));

   assign      fpu_rnd_next[1:0] = (gsr_rnd[2])? gsr_rnd[1:0]: dp_ctl_fsr_rnd[1:0];

   // if alignaddress_little then write the 2's complement

   assign     align_addr_data_w2[2:0] = (opf[1])? (~wgsr_align_offset_w2[2:0] + 3'b001):

                                                  wgsr_align_offset_w2[2:0];

   assign     gsr_addr_d = (ifu_tlu_sraddr_d[6:0] == 7'b0010011);

   assign     wgsr_e = exu_ffu_wsr_inst_e & gsr_addr_e;

   dff_s gsr_addr_d2e(.din(gsr_addr_d), .clk(clk), .q(gsr_addr_e), .se(se), .si(), .so());

   // need independent kill checks because this isn't killed by new fpop

   assign     wgsr_vld_m = wgsr_m & ~(thr_match_mw2 & flush_w2);

   assign     wgsr_vld_w = wgsr_w & ifu_tlu_inst_vld_w & ~(thr_match_ww2 & flush_w2);

   assign     wgsr_vld_w2 = wgsr_w2 & ~flush_w2;

   dff_s wgsr_e2m(.din(wgsr_e), .clk(clk), .q(wgsr_m), .si(), .so(), .se(se));

   dff_s wgsr_m2w(.din(wgsr_vld_m), .clk(clk), .q(wgsr_w), .si(), .so(), .se(se));

   dff_s wgsr_w2w2(.din(wgsr_vld_w), .clk(clk), .q(wgsr_w2), .si(), .so(), .se(se));

   assign     thr_w2[3] = (tid_w2[1:0] == 2'b11);

   assign     thr_w2[2] = (tid_w2[1:0] == 2'b10);

   assign     thr_w2[1] = (tid_w2[1:0] == 2'b01);

   assign     thr_w2[0] = (tid_w2[1:0] == 2'b00);

   assign     ctl_dp_thr_e[3] = (tid_e[1:0] == 2'b11);

   assign     ctl_dp_thr_e[2] = (tid_e[1:0] == 2'b10);

   assign     ctl_dp_thr_e[1] = (tid_e[1:0] == 2'b01);

   assign     ctl_dp_thr_e[0] = (tid_e[1:0] == 2'b00);

   assign     thr_fp[3] = (tid[1:0] == 2'b11);

   assign     thr_fp[2] = (tid[1:0] == 2'b10);

   assign     thr_fp[1] = (tid[1:0] == 2'b01);

   assign     thr_fp[0] = (tid[1:0] == 2'b00);

   assign     t0_siam_w2 = thr_fp[0] & siam & visop_w2_vld;

   assign     t0_gsr_wsr_w2 = thr_w2[0] & wgsr_vld_w2;

   assign     t0_alignaddr_w2 = thr_fp[0] & alignaddr & visop_w2_vld;

   assign     t0_rnd_wen_l = ~(t0_gsr_wsr_w2 | t0_siam_w2);

   assign     t0_align_wen_l = ~(t0_gsr_wsr_w2 | t0_alignaddr_w2);

   assign     t1_siam_w2 = thr_fp[1] & siam & visop_w2_vld;

   assign     t1_gsr_wsr_w2 = thr_w2[1] & wgsr_vld_w2;

   assign     t1_alignaddr_w2 = thr_fp[1] & alignaddr & visop_w2_vld;

   assign     t1_rnd_wen_l = ~(t1_gsr_wsr_w2 | t1_siam_w2);

   assign     t1_align_wen_l = ~(t1_gsr_wsr_w2 | t1_alignaddr_w2);

   assign     t2_siam_w2 = thr_fp[2] & siam & visop_w2_vld;

   assign     t2_gsr_wsr_w2 = thr_w2[2] & wgsr_vld_w2;

   assign     t2_alignaddr_w2 = thr_fp[2] & alignaddr & visop_w2_vld;

   assign     t2_rnd_wen_l = ~(t2_gsr_wsr_w2 | t2_siam_w2);

   assign     t2_align_wen_l = ~(t2_gsr_wsr_w2 | t2_alignaddr_w2);

   assign     t3_siam_w2 = thr_fp[3] & siam & visop_w2_vld;

   assign     t3_gsr_wsr_w2 = thr_w2[3] & wgsr_vld_w2;

   assign     t3_alignaddr_w2 = thr_fp[3] & alignaddr & visop_w2_vld;

   assign     t3_rnd_wen_l = ~(t3_gsr_wsr_w2 | t3_siam_w2);

   assign     t3_align_wen_l = ~(t3_gsr_wsr_w2 | t3_alignaddr_w2);

   assign     ctl_dp_gsr_wsr_w2[3:0] = {t3_gsr_wsr_w2,t2_gsr_wsr_w2,t1_gsr_wsr_w2,t0_gsr_wsr_w2};

   // Storage flops and muxes

   mux3ds #(3) t0_rnd_mux(.dout(t0_gsr_rnd_next[2:0]),

                          .in0(t0_gsr_rnd[2:0]),

                          .in1(wgsr_rnd_w2[2:0]),

                          .in2(siam_rnd[2:0]),

                          .sel0(t0_rnd_wen_l),

                          .sel1(t0_gsr_wsr_w2),

                          .sel2(t0_siam_w2));

   mux3ds #(3) t0_align_mux(.dout(t0_gsr_align_next[2:0]),

                            .in0(t0_gsr_align[2:0]),

                            .in1(wgsr_align_offset_w2[2:0]),

                            .in2(align_addr_data_w2[2:0]),

                            .sel0(t0_align_wen_l),

                            .sel1(t0_gsr_wsr_w2),

                            .sel2(t0_alignaddr_w2));

   mux3ds #(3) t1_rnd_mux(.dout(t1_gsr_rnd_next[2:0]),

                          .in0(t1_gsr_rnd[2:0]),

                          .in1(wgsr_rnd_w2[2:0]),

                          .in2(siam_rnd[2:0]),

                          .sel0(t1_rnd_wen_l),

                          .sel1(t1_gsr_wsr_w2),

                          .sel2(t1_siam_w2));

   mux3ds #(3) t1_align_mux(.dout(t1_gsr_align_next[2:0]),

                            .in0(t1_gsr_align[2:0]),

                            .in1(wgsr_align_offset_w2[2:0]),

                            .in2(align_addr_data_w2[2:0]),

                            .sel0(t1_align_wen_l),

                            .sel1(t1_gsr_wsr_w2),

                            .sel2(t1_alignaddr_w2));

   mux3ds #(3) t2_rnd_mux(.dout(t2_gsr_rnd_next[2:0]),

                          .in0(t2_gsr_rnd[2:0]),

                          .in1(wgsr_rnd_w2[2:0]),

                          .in2(siam_rnd[2:0]),

                          .sel0(t2_rnd_wen_l),

                          .sel1(t2_gsr_wsr_w2),

                          .sel2(t2_siam_w2));

   mux3ds #(3) t2_align_mux(.dout(t2_gsr_align_next[2:0]),

                            .in0(t2_gsr_align[2:0]),

                            .in1(wgsr_align_offset_w2[2:0]),

                            .in2(align_addr_data_w2[2:0]),

                            .sel0(t2_align_wen_l),

                            .sel1(t2_gsr_wsr_w2),

                            .sel2(t2_alignaddr_w2));

   mux3ds #(3) t3_rnd_mux(.dout(t3_gsr_rnd_next[2:0]),

                          .in0(t3_gsr_rnd[2:0]),

                          .in1(wgsr_rnd_w2[2:0]),

                          .in2(siam_rnd[2:0]),

                          .sel0(t3_rnd_wen_l),

                          .sel1(t3_gsr_wsr_w2),

                          .sel2(t3_siam_w2));

   mux3ds #(3) t3_align_mux(.dout(t3_gsr_align_next[2:0]),

                            .in0(t3_gsr_align[2:0]),

                            .in1(wgsr_align_offset_w2[2:0]),

                            .in2(align_addr_data_w2[2:0]),

                            .sel0(t3_align_wen_l),

                            .sel1(t3_gsr_wsr_w2),

                            .sel2(t3_alignaddr_w2));

   dffr_s #(6) t0_gsr_dff(.din({t0_gsr_rnd_next[2:0], t0_gsr_align_next[2:0]}), .clk(clk),

                       .q({t0_gsr_rnd[2:0], t0_gsr_align[2:0]}), .se(se),

                       .si(), .so(), .rst(reset));

   dffr_s #(6) t1_gsr_dff(.din({t1_gsr_rnd_next[2:0], t1_gsr_align_next[2:0]}), .clk(clk),

                       .q({t1_gsr_rnd[2:0], t1_gsr_align[2:0]}), .se(se),

                       .si(), .so(), .rst(reset));

   dffr_s #(6) t2_gsr_dff(.din({t2_gsr_rnd_next[2:0], t2_gsr_align_next[2:0]}), .clk(clk),

                       .q({t2_gsr_rnd[2:0], t2_gsr_align[2:0]}), .se(se),

                       .si(), .so(), .rst(reset));

   dffr_s #(6) t3_gsr_dff(.din({t3_gsr_rnd_next[2:0], t3_gsr_align_next[2:0]}), .clk(clk),

                       .q({t3_gsr_rnd[2:0], t3_gsr_align[2:0]}), .se(se),