流水線的數據衝突分爲三類:WAR,RAW,WAWhtml
https://wenku.baidu.com/view/e066926d48d7c1c708a14508.html數組
add x5, x4,x6架構
add x4, x3, x2oop
add x5, x4,x6fetch
add x5, x3, x2spa
add x5, x4,x6code
add x4, x5, x2htm
因爲蜂鳥E200系列是按序派遣,按順序寫回的微架構,在指令派遣時候就已經從通用寄存器數組中讀取了源操做數。後續執行的指令寫回regfile的操做不可能影響到前面指令的讀取,因此不可能發生WAR相關性形成的數據衝突。blog
正在派遣的指令處在流水線的第二級,假設以前派遣的指令是單週期指令,則前序指令確定已經完成了執行且將結果寫回了Regfile。所以正在派遣的指令不可能會發生RAW數據衝突。可是假設以前派遣的指令是多週期指令(長指令),因爲指令須要多個週期才能寫回結果。所以正在派遣的指令可能會產生前序相關的RAW相關性。索引
正在派遣的指令處在流水線的第二級,假設以前派遣的指令是單週期指令,則前序指令確定已經完成了執行且將結果寫回了Regfile。所以正在派遣的指令不可能會發生WAW數據衝突。可是假設以前派遣的指令是多週期指令(長指令),因爲指令須要多個週期才能寫回結果。所以正在派遣的指令可能會產生前序相關的WAW相關性。
爲了能檢測出長指令的RAW和WAW相關性,蜂鳥E200使用了一個outstanding instruction track fifo(OITF)模塊。在流水線的派遣(Dispatch)點,每一次派遣一個長指令,則會在OITF中分配一個表項(Entry),在這個表項中會存儲該長指令的結果寄存器索引。在流水線的寫回(Write-back)點,每次按順序寫回一個長指令以後,就會將此指令在OITF中的表項移除。
每條指令派遣時,都會將本指令的源操做數和目的操做數寄存器索引和OITF中的各個表項進行比對,從而判斷本指令是否與已經被派遣出,且還沒有寫回的長指令產生RAW和WAW相關性。若是產生相關性,則stall住當前指令的派遣。若是沒有RAW和WAW相關性,且該指令爲多週期長指令,把該指令寫入OITF,若是OITF是full,則仍要stall住管線,等待OITF釋放空間後,再寫入並派遣。
在writeback模塊,會進行長指令寫回仲裁,長指令寫回regfile後,會釋放OITF中相應的表項。
OITF代碼以下,若是fifo full,則dis_ready=0, 與dispatch模塊握手失敗,不會發送新的dispatch進來。若是不爲空,會發送新的指令進來進行判斷。
`include "e203_defines.v" module e203_exu_oitf ( output dis_ready, input dis_ena, //dispatch a long instruction enable signal input ret_ena, //write back a long instruction enable signal output [`E203_ITAG_WIDTH-1:0] dis_ptr, //write pointer output [`E203_ITAG_WIDTH-1:0] ret_ptr, //read pointer output [`E203_RFIDX_WIDTH-1:0] ret_rdidx, output ret_rdwen, output ret_rdfpu, output [`E203_PC_SIZE-1:0] ret_pc, input disp_i_rs1en, // enable if current dispatch instruction fetch first source operand input disp_i_rs2en, // ... input disp_i_rs3en, // ... input disp_i_rdwen, // enable if current dispatch instruction write back to register input disp_i_rs1fpu, // enable if current dispath instruction need to read float gpr input disp_i_rs2fpu, //... input disp_i_rs3fpu, //... input disp_i_rdfpu, //enable if current dipatch instruction need to write back to float register files. //register index input [`E203_RFIDX_WIDTH-1:0] disp_i_rs1idx, input [`E203_RFIDX_WIDTH-1:0] disp_i_rs2idx, input [`E203_RFIDX_WIDTH-1:0] disp_i_rs3idx, input [`E203_RFIDX_WIDTH-1:0] disp_i_rdidx, input [`E203_PC_SIZE -1:0] disp_i_pc, //pc of current dispatch instruciotn output oitfrd_match_disprs1, //dispatch instruction rs1 is same as any item of result register in oitf output oitfrd_match_disprs2, //... output oitfrd_match_disprs3, //... output oitfrd_match_disprd, //dispatch instruction rd is same as any item of result register in oitf. //if empty, no conflict output oitf_empty, input clk, input rst_n ); wire [`E203_OITF_DEPTH-1:0] vld_set; wire [`E203_OITF_DEPTH-1:0] vld_clr; wire [`E203_OITF_DEPTH-1:0] vld_ena; wire [`E203_OITF_DEPTH-1:0] vld_nxt; wire [`E203_OITF_DEPTH-1:0] vld_r; //if it is valid signal in all item wire [`E203_OITF_DEPTH-1:0] rdwen_r;// if it is write back register in all item wire [`E203_OITF_DEPTH-1:0] rdfpu_r; //result register in all item if are float wire [`E203_RFIDX_WIDTH-1:0] rdidx_r[`E203_OITF_DEPTH-1:0]; //register index in all items // The PC here is to be used at wback stage to track out the // PC of exception of long-pipe instruction wire [`E203_PC_SIZE-1:0] pc_r[`E203_OITF_DEPTH-1:0]; wire alc_ptr_ena = dis_ena; //dispatch a long instruction enable signal, as write pointer enable signal wire ret_ptr_ena = ret_ena; //write back a long instruction enable signal, as read pointer enable signal wire oitf_full ; wire [`E203_ITAG_WIDTH-1:0] alc_ptr_r; //write pointer, long instruction dispatch wire [`E203_ITAG_WIDTH-1:0] ret_ptr_r; //read pointer, long instruction write back generate if(`E203_OITF_DEPTH > 1) begin: depth_gt1//{ //extra mark bit for write full wire alc_ptr_flg_r; wire alc_ptr_flg_nxt = ~alc_ptr_flg_r; wire alc_ptr_flg_ena = (alc_ptr_r == ($unsigned(`E203_OITF_DEPTH-1))) & alc_ptr_ena; sirv_gnrl_dfflr #(1) alc_ptr_flg_dfflrs(alc_ptr_flg_ena, alc_ptr_flg_nxt, alc_ptr_flg_r, clk, rst_n); wire [`E203_ITAG_WIDTH-1:0] alc_ptr_nxt; //if write to fifo depth, write ptr = 0,otherwise write ptr = write ptr + 1 assign alc_ptr_nxt = alc_ptr_flg_ena ? `E203_ITAG_WIDTH'b0 : (alc_ptr_r + 1'b1); sirv_gnrl_dfflr #(`E203_ITAG_WIDTH) alc_ptr_dfflrs(alc_ptr_ena, alc_ptr_nxt, alc_ptr_r, clk, rst_n); //extra mark bit for read empty wire ret_ptr_flg_r; wire ret_ptr_flg_nxt = ~ret_ptr_flg_r; wire ret_ptr_flg_ena = (ret_ptr_r == ($unsigned(`E203_OITF_DEPTH-1))) & ret_ptr_ena; sirv_gnrl_dfflr #(1) ret_ptr_flg_dfflrs(ret_ptr_flg_ena, ret_ptr_flg_nxt, ret_ptr_flg_r, clk, rst_n); wire [`E203_ITAG_WIDTH-1:0] ret_ptr_nxt; //if read to fifo depth, read ptr = 0, otherwise read prt = read prt + 1 assign ret_ptr_nxt = ret_ptr_flg_ena ? `E203_ITAG_WIDTH'b0 : (ret_ptr_r + 1'b1); sirv_gnrl_dfflr #(`E203_ITAG_WIDTH) ret_ptr_dfflrs(ret_ptr_ena, ret_ptr_nxt, ret_ptr_r, clk, rst_n); //empty, full mark assign oitf_empty = (ret_ptr_r == alc_ptr_r) & (ret_ptr_flg_r == alc_ptr_flg_r); assign oitf_full = (ret_ptr_r == alc_ptr_r) & (~(ret_ptr_flg_r == alc_ptr_flg_r)); end//} else begin: depth_eq1//}{ assign alc_ptr_r =1'b0; assign ret_ptr_r =1'b0; assign oitf_empty = ~vld_r[0]; assign oitf_full = vld_r[0]; end//} endgenerate//} assign ret_ptr = ret_ptr_r; assign dis_ptr = alc_ptr_r; //// //// // If the OITF is not full, or it is under retiring, then it is ready to accept new dispatch //// assign dis_ready = (~oitf_full) | ret_ena; // To cut down the loop between ALU write-back valid --> oitf_ret_ena --> oitf_ready ---> dispatch_ready --- > alu_i_valid // we exclude the ret_ena from the ready signal assign dis_ready = (~oitf_full); wire [`E203_OITF_DEPTH-1:0] rd_match_rs1idx; wire [`E203_OITF_DEPTH-1:0] rd_match_rs2idx; wire [`E203_OITF_DEPTH-1:0] rd_match_rs3idx; wire [`E203_OITF_DEPTH-1:0] rd_match_rdidx; genvar i; generate //{ for (i=0; i<`E203_OITF_DEPTH; i=i+1) begin:oitf_entries//{ //every time, assign a item and write pointer same as current i, then //valid set is high assign vld_set[i] = alc_ptr_ena & (alc_ptr_r == i); //every time, assign a item and read pointer same as current i, then //valid clr is high assign vld_clr[i] = ret_ptr_ena & (ret_ptr_r == i); assign vld_ena[i] = vld_set[i] | vld_clr[i]; assign vld_nxt[i] = vld_set[i] | (~vld_clr[i]); sirv_gnrl_dfflr #(1) vld_dfflrs(vld_ena[i], vld_nxt[i], vld_r[i], clk, rst_n); //Payload only set, no need to clear sirv_gnrl_dffl #(`E203_RFIDX_WIDTH) rdidx_dfflrs(vld_set[i], disp_i_rdidx, rdidx_r[i], clk); sirv_gnrl_dffl #(`E203_PC_SIZE ) pc_dfflrs (vld_set[i], disp_i_pc , pc_r[i] , clk); sirv_gnrl_dffl #(1) rdwen_dfflrs(vld_set[i], disp_i_rdwen, rdwen_r[i], clk); sirv_gnrl_dffl #(1) rdfpu_dfflrs(vld_set[i], disp_i_rdfpu, rdfpu_r[i], clk); //compare dispatch source operand with result register in fifo assign rd_match_rs1idx[i] = vld_r[i] & rdwen_r[i] & disp_i_rs1en & (rdfpu_r[i] == disp_i_rs1fpu) & (rdidx_r[i] == disp_i_rs1idx); assign rd_match_rs2idx[i] = vld_r[i] & rdwen_r[i] & disp_i_rs2en & (rdfpu_r[i] == disp_i_rs2fpu) & (rdidx_r[i] == disp_i_rs2idx); assign rd_match_rs3idx[i] = vld_r[i] & rdwen_r[i] & disp_i_rs3en & (rdfpu_r[i] == disp_i_rs3fpu) & (rdidx_r[i] == disp_i_rs3idx); assign rd_match_rdidx [i] = vld_r[i] & rdwen_r[i] & disp_i_rdwen & (rdfpu_r[i] == disp_i_rdfpu ) & (rdidx_r[i] == disp_i_rdidx ); end//} endgenerate//} //rs1 in fifo, so RAW relative assign oitfrd_match_disprs1 = |rd_match_rs1idx; //rs2 in fifo, so RAW relative assign oitfrd_match_disprs2 = |rd_match_rs2idx; //rs3 in fifo, so RAW relative assign oitfrd_match_disprs3 = |rd_match_rs3idx; //rd in fifo, so WAW relative assign oitfrd_match_disprd = |rd_match_rdidx ; assign ret_rdidx = rdidx_r[ret_ptr]; assign ret_pc = pc_r [ret_ptr]; assign ret_rdwen = rdwen_r[ret_ptr]; assign ret_rdfpu = rdfpu_r[ret_ptr]; endmodule