Optimized out aes64im hardware; sharing with aes64d

src/ieu/aes_instructions/aes64d.sv

@@ -28,12 +28,12 @@
 module aes64d(
    input  logic [63:0] rs1,
    input  logic [63:0] rs2,
-   input  logic        finalround,
+   input  logic        finalround, aes64im,
    output logic [63:0] result
 );
    
    logic [127:0] 		    ShiftRowOut;
-   logic [63:0] 		    SboxOut, MixcolOut;
+   logic [63:0] 		    SboxOut, MixcolIn, MixcolOut;
    
    // Apply inverse shiftrows to rs2 and rs1
    aesinvshiftrow srow({rs2, rs1}, ShiftRowOut);
@@ -42,9 +42,11 @@ module aes64d(
    aesinvsboxword invsbox0(ShiftRowOut[31:0],  SboxOut[31:0]);
    aesinvsboxword invsbox1(ShiftRowOut[63:32], SboxOut[63:32]);
    
+    mux2 #(64) mixcolmux(SboxOut, rs1, aes64im, MixcolIn);
+   
    // Apply inverse mixword to sbox outputs
-   aesinvmixcolumns invmw0(SboxOut[31:0], MixcolOut[31:0]);
-   aesinvmixcolumns invmw1(SboxOut[63:32], MixcolOut[63:32]);
+   aesinvmixcolumns invmw0(MixcolIn[31:0], MixcolOut[31:0]);
+   aesinvmixcolumns invmw1(MixcolIn[63:32], MixcolOut[63:32]);
    
    // Final round skips mixcolumns.
    mux2 #(64) resultmux(MixcolOut, SboxOut, finalround, result);

src/ieu/aes_instructions/aes64e.sv

@@ -28,7 +28,7 @@
 module aes64e(
     input  logic [63:0] rs1,
     input  logic [63:0] rs2,
-    input  logic        finalround,
+    input  logic        finalround, 
     output logic [63:0] result
 );
    
@@ -41,9 +41,9 @@ module aes64e(
     // Apply substitution box to 2 lower words
     aessboxword sbox0(ShiftRowOut[31:0],  SboxOut[31:0]);
     aessboxword sbox1(ShiftRowOut[63:32], SboxOut[63:32]);
-   
+
     // Apply mix columns operations
-    aesmixcolumns mw0(SboxOut[31:0], MixcolOut[31:0]);
+    aesmixcolumns mw0(SboxOut[31:0],  MixcolOut[31:0]);
     aesmixcolumns mw1(SboxOut[63:32], MixcolOut[63:32]);    
 
     // Skip mixcolumns on last round

src/ieu/aes_instructions/aes64im.sv

@@ -27,9 +27,9 @@
 
 module aes64im(
    input  logic [63:0] rs1,
-   output logic [63:0] DataOut
+   output logic [63:0] result
 );
 
-   aesinvmixcolumns inv_mw_0(rs1[31:0],  DataOut[31:0]);
-   aesinvmixcolumns inv_mw_1(rs1[63:32], DataOut[63:32]);
+   aesinvmixcolumns inv_mw_0(rs1[31:0],  result[31:0]);
+   aesinvmixcolumns inv_mw_1(rs1[63:32], result[63:32]);
 endmodule

src/ieu/aes_instructions/aes64ks1i.sv

@@ -26,7 +26,7 @@
 ////////////////////////////////////////////////////////////////////////////////////////////////
 
 module aes64ks1i(
-   input  logic [3:0]  roundnum,
+   input  logic [3:0]  round,
    input  logic [63:0] rs1,
    output logic [63:0] result
 );                 
@@ -35,11 +35,10 @@ module aes64ks1i(
    logic [7:0] 			  rcon8;
    logic [31:0] 		     rcon, rs1Rotate, tmp2, SboxOut;
    
-  
-   rconlut128 rc(roundnum, rcon8);                        // Get rcon value from lookup table
+   rconlut128 rc(round, rcon8);                           // Get rcon value from lookup table
    assign rcon = {24'b0, rcon8};                          // Zero-pad RCON 
    assign rs1Rotate = {rs1[39:32], rs1[63:40]};           // Get rotated value fo ruse in tmp2
-   assign finalround = (roundnum == 4'b1010);             // round 10 is the last one 
+   assign finalround = (round == 4'b1010);                // round 10 is the last one 
    assign tmp2 = finalround ? rs1[63:32] : rs1Rotate;     // Don't rotate on the last round
    aessboxword sbox(tmp2, SboxOut);                       // Substitute bytes of value obtained for tmp2 using Rijndael sbox
    assign result[31:0] = SboxOut ^ rcon;

src/ieu/bmu/bitmanipalu.sv

@@ -115,13 +115,13 @@ module bitmanipalu import cvw::*; #(parameter cvw_t P) (
   // ZKND Unit
   if (P.ZKND_SUPPORTED) begin: zknd
     if (P.XLEN == 32) aes32d aes32d(.bs(Funct7[6:5]), .rs1(ABMU), .rs2(BBMU), .finalround(ZBBSelect[2]), .result(ZKNDResult));
-    else              zknd64 #(P.XLEN) ZKND64(.A(ABMU), .B(BBMU), .Funct7, .RNUM(Rs2E[3:0]), .ZKNDSelect(ZBBSelect[2:0]), .ZKNDResult);
+    else              zknd64 #(P.XLEN) ZKND64(.A(ABMU), .B(BBMU), .Funct7, .round(Rs2E[3:0]), .ZKNDSelect(ZBBSelect[3:0]), .ZKNDResult);
   end else assign ZKNDResult = 0;
 
   // ZKNE Unit
   if (P.ZKNE_SUPPORTED) begin: zkne
     if (P.XLEN == 32) aes32e aes32e(.bs(Funct7[6:5]), .rs1(ABMU), .rs2(BBMU), .finalround(ZBBSelect[2]), .result(ZKNEResult));
-    else              zkne64 #(P.XLEN) ZKNE64(.A(ABMU), .B(BBMU), .Funct7, .RNUM(Rs2E[3:0]), .ZKNESelect(ZBBSelect[2:0]), .ZKNEResult);
+    else              zkne64 #(P.XLEN) ZKNE64(.A(ABMU), .B(BBMU), .Funct7, .round(Rs2E[3:0]), .ZKNESelect(ZBBSelect[2:0]), .ZKNEResult);
   end else assign ZKNEResult = 0;
 
   // ZKNH Unit

src/ieu/bmu/bmuctrl.sv

@@ -223,7 +223,7 @@ module bmuctrl import cvw::*;  #(parameter cvw_t P) (
           17'b0110011_0011101_000: BMUControlsD = `BMUCTRLW'b000_0111_0100_1_0_0_1_0_0_0_0_0;  // aes64ds - decrypt final round
           17'b0110011_0011111_000: BMUControlsD = `BMUCTRLW'b000_0111_0000_1_0_0_1_0_0_0_0_0;  // aes64dsm - decrypt mid round
           17'b0010011_0011000_001: if (Rs2D == 5'b00000)
-                                   BMUControlsD = `BMUCTRLW'b000_0111_0011_1_1_0_1_0_0_0_0_0;  // aes64im - decrypt keyschdule mixcolumns
+                                   BMUControlsD = `BMUCTRLW'b000_0111_1000_1_1_0_1_0_0_0_0_0;  // aes64im - decrypt keyschdule mixcolumns
         endcase
     end
 

src/ieu/kmu/zknd64.sv

@@ -29,18 +29,17 @@
 module zknd64 #(parameter WIDTH=32) (
    input  logic [WIDTH-1:0] A, B,
    input  logic [6:0] 	    Funct7,
-   input  logic [3:0] 	    RNUM,
-   input  logic [2:0] 	    ZKNDSelect,
+   input  logic [3:0] 	    round,
+   input  logic [3:0] 	    ZKNDSelect,
    output logic [WIDTH-1:0] ZKNDResult
 );
    
    logic [63:0] 	     aes64dRes, aes64imRes, aes64ks1iRes, aes64ks2Res;
    
    // RV64
-   aes64d    aes64d(.rs1(A), .rs2(B), .finalround(ZKNDSelect[2]), .result(aes64dRes)); // decode AES
-   aes64im   aes64im(.rs1(A), .DataOut(aes64imRes));
-   aes64ks1i aes64ks1i(.roundnum(RNUM), .rs1(A), .result(aes64ks1iRes));
+   aes64d    aes64d(.rs1(A), .rs2(B), .finalround(ZKNDSelect[2]), .aes64im(ZKNDSelect[3]), .result(aes64dRes)); // decode AES
+   aes64ks1i aes64ks1i(.round, .rs1(A), .result(aes64ks1iRes));
    aes64ks2  aes64ks2(.rs2(B), .rs1(A), .result(aes64ks2Res));
    
-   mux4 #(WIDTH) zkndmux(aes64dRes, aes64ks1iRes, aes64ks2Res, aes64imRes, ZKNDSelect[1:0], ZKNDResult);
+   mux3 #(WIDTH) zkndmux(aes64dRes, aes64ks1iRes, aes64ks2Res, ZKNDSelect[1:0], ZKNDResult);
 endmodule

src/ieu/kmu/zkne64.sv

@@ -29,7 +29,7 @@
 module zkne64 #(parameter WIDTH=32) (
    input  logic [WIDTH-1:0] A, B,
    input  logic [6:0] 	    Funct7,
-   input  logic [3:0] 	    RNUM,
+   input  logic [3:0] 	    round,
    input  logic [2:0] 	    ZKNESelect,
    output logic [WIDTH-1:0] ZKNEResult
 );
@@ -38,7 +38,7 @@ module zkne64 #(parameter WIDTH=32) (
    
    // RV64
    aes64e    aes64e(.rs1(A), .rs2(B), .finalround(ZKNESelect[2]), .result(aes64eRes));
-   aes64ks1i aes64ks1i(.roundnum(RNUM), .rs1(A), .result(aes64ks1iRes));
+   aes64ks1i aes64ks1i(.round, .rs1(A), .result(aes64ks1iRes));
    aes64ks2  aes64ks2(.rs2(B), .rs1(A), .result(aes64ks2Res));
    
    // 010 is a placeholder to match the select of ZKND's AES64KS1I since they share some instruction