propagated otfc swap to Rad2 and 4 qslc

pipelined/src/fpu/fdivsqrt/fdivsqrtiter.sv

@@ -120,15 +120,15 @@ module fdivsqrtiter(
   // k=DIVCOPIES of the recurrence logic
   genvar i;
   generate
-    for(i=0; $unsigned(i)<`DIVCOPIES; i++) begin : interations
+    for(i=0; $unsigned(i)<`DIVCOPIES; i++) begin : iterations
       if (`RADIX == 2) begin: stage
-        fdivsqrtstage2 fdivsqrtstage(.D, .DBar, .SqrtM,
+        fdivsqrtstage2 fdivsqrtstage(.D, .DBar, .SqrtM, .OTFCSwap,
         .WS(WS[i]), .WC(WC[i]), .WSNext(WSNext[i]), .WCNext(WCNext[i]),
         .C(C[i]), .U(U[i]), .UM(UM[i]), .CNext(C[i+1]), .UNext(UNext[i]), .UMNext(UMNext[i]), .un(un[i]));
       end else begin: stage
         logic j1;
         assign j1 = (i == 0 & ~C[0][`DIVb-1]);
-        fdivsqrtstage4 fdivsqrtstage(.D, .DBar, .D2, .DBar2, .SqrtM, .j1,
+        fdivsqrtstage4 fdivsqrtstage(.D, .DBar, .D2, .DBar2, .SqrtM, .j1, .OTFCSwap,
         .WS(WS[i]), .WC(WC[i]), .WSNext(WSNext[i]), .WCNext(WCNext[i]), 
         .C(C[i]), .U(U[i]), .UM(UM[i]), .CNext(C[i+1]), .UNext(UNext[i]), .UMNext(UMNext[i]), .un(un[i]));
       end

pipelined/src/fpu/fdivsqrt/fdivsqrtqsel2.sv

@@ -32,11 +32,13 @@
 
 module fdivsqrtqsel2 ( 
   input  logic [3:0] ps, pc, 
+  input  logic  swap,
   output logic  up, uz, un
 );
  
   logic [3:0]  p, g;
   logic        magnitude, sign, cout;
+  logic        pos, neg;
 
   // The quotient selection logic is presented for simplicity, not
   // for efficiency.  You can probably optimize your logic to
@@ -57,7 +59,11 @@ module fdivsqrtqsel2 (
 						(ps[0]&pc[0])))));
 
   // Produce digit = +1, 0, or -1
-  assign up = magnitude & ~sign;
+  assign pos = magnitude & ~sign;
   assign uz  = ~magnitude;
-  assign un = magnitude & sign;
+  assign neg = magnitude & sign;
+
+  // Check for swap (int div only)
+  assign un = swap ? pos : neg;
+  assign up = swap ? neg : pos;
 endmodule

pipelined/src/fpu/fdivsqrt/fdivsqrtqsel4cmp.sv

@@ -34,12 +34,13 @@ module fdivsqrtqsel4cmp (
   input logic [2:0] Dmsbs,
   input logic [4:0] Smsbs,
   input logic [7:0] WSmsbs, WCmsbs,
-  input logic Sqrt, j1,
+  input logic Sqrt, j1, OTFCSwap,
   output logic [3:0] udigit
 );
 	logic [6:0] Wmsbs;
 	logic [7:0] PreWmsbs;
 	logic [2:0] A;
+  logic [3:0] udigitsel, udigitswap;
 
 	assign PreWmsbs = WCmsbs + WSmsbs;
 	assign Wmsbs = PreWmsbs[7:1];
@@ -85,9 +86,12 @@ module fdivsqrtqsel4cmp (
  
   // Compare residual W to selection constants to choose digit
   always_comb 
-    if ($signed(Wmsbs) >= $signed(mk2)) udigit = 4'b1000; // choose 2
+    if ($signed(Wmsbs) >= $signed(mk2)) udigitsel = 4'b1000; // choose 2
-    else if ($signed(Wmsbs) >= $signed(mk1)) udigit = 4'b0100; // choose 1
+    else if ($signed(Wmsbs) >= $signed(mk1)) udigitsel = 4'b0100; // choose 1
-    else if ($signed(Wmsbs) >= $signed(mk0)) udigit = 4'b0000; // choose 0
+    else if ($signed(Wmsbs) >= $signed(mk0)) udigitsel = 4'b0000; // choose 0
-    else if ($signed(Wmsbs) >= $signed(mkm1)) udigit = 4'b0010; // choose -1
+    else if ($signed(Wmsbs) >= $signed(mkm1)) udigitsel = 4'b0010; // choose -1
-    else udigit = 4'b0001; // choose -2	
+    else udigitsel = 4'b0001; // choose -2	
+
+  assign udigitswap = {udigitsel[0], udigitsel[1], udigitsel[2], udigitsel[3]};
+  assign udigit = OTFCSwap ? udigitswap : udigitsel;
 endmodule

pipelined/src/fpu/fdivsqrt/fdivsqrtstage2.sv

@@ -38,6 +38,7 @@ module fdivsqrtstage2 (
   input  logic [`DIVb+3:0]  WS, WC,
   input  logic [`DIVb+1:0] C,
   input  logic SqrtM,
+  input  logic OTFCSwap,
   output logic un,
   output logic [`DIVb+1:0] CNext,
   output logic [`DIVb:0] UNext, UMNext, 
@@ -59,7 +60,7 @@ module fdivsqrtstage2 (
 	// 0000 =  0
 	// 0010 = -1
 	// 0001 = -2
-  fdivsqrtqsel2 qsel2(WS[`DIVb+3:`DIVb], WC[`DIVb+3:`DIVb], up, uz, un);
+  fdivsqrtqsel2 qsel2(WS[`DIVb+3:`DIVb], WC[`DIVb+3:`DIVb], OTFCSwap, up, uz, un);
 
   // Sqrt F generation
   fdivsqrtfgen2 fgen2(.up, .uz, .C(CNext), .U, .UM, .F);

pipelined/src/fpu/fdivsqrt/fdivsqrtstage4.sv

@@ -36,8 +36,8 @@ module fdivsqrtstage4 (
   input  logic [`DIVb:0] U, UM,
   input  logic [`DIVb+3:0]  WS, WC,
   input  logic [`DIVb+1:0] C,
+  input  logic SqrtM, j1, OTFCSwap,
   output logic [`DIVb+1:0] CNext,
-  input logic SqrtM, j1,
   output logic un,
   output logic [`DIVb:0] UNext, UMNext, 
   output logic [`DIVb+3:0]  WSNext, WCNext
@@ -65,8 +65,8 @@ module fdivsqrtstage4 (
   assign WCmsbs = WC[`DIVb+3:`DIVb-4];
   assign WSmsbs = WS[`DIVb+3:`DIVb-4];
 
-  fdivsqrtqsel4cmp qsel4(.Dmsbs, .Smsbs, .WSmsbs, .WCmsbs, .Sqrt(SqrtM), .j1, .udigit);
+  fdivsqrtqsel4cmp qsel4(.Dmsbs, .Smsbs, .WSmsbs, .WCmsbs, .Sqrt(SqrtM), .j1, .udigit, .OTFCSwap);
-  assign un = 0; // unused for radix 4
+  assign un = 1'b0; // unused for radix 4
 
   // F generation logic
   fdivsqrtfgen4 fgen4(.udigit, .C({2'b11, CNext}), .U({3'b000, U}), .UM({3'b000, UM}), .F);