Support shared color mask, implement more shader instructions

Support shared color masks (used by Nouveau and maybe the NVIDIA
driver).
Support draw buffers (also required by OpenGL).
Support viewport transform disable (disabled for now as it breaks some
games).
Fix instanced rendering draw being ignored for multi draw.
Fix IADD and IADD3 immediate shader encodings, that was not matching
some ops.
Implement FFMA32I shader instruction.
Implement IMAD shader instruction.

2 files changed, 64 insertions, 0 deletions

diff --git a/Ryujinx.Graphics.Shader/Instructions/InstEmitAlu.cs b/Ryujinx.Graphics.Shader/Instructions/InstEmitAlu.cs
index 1d3a1101..2a8f00cc 100644
--- a/Ryujinx.Graphics.Shader/Instructions/InstEmitAlu.cs
+++ b/Ryujinx.Graphics.Shader/Instructions/InstEmitAlu.cs
@@ -200,6 +200,50 @@ namespace Ryujinx.Graphics.Shader.Instructions
             // TODO: CC, X, corner cases
         }
 
+        public static void Imad(EmitterContext context)
+        {
+            OpCodeAlu op = (OpCodeAlu)context.CurrOp;
+
+            bool signedA = context.CurrOp.RawOpCode.Extract(48);
+            bool signedB = context.CurrOp.RawOpCode.Extract(53);
+            bool high    = context.CurrOp.RawOpCode.Extract(54);
+
+            Operand srcA = GetSrcA(context);
+            Operand srcB = GetSrcB(context);
+            Operand srcC = GetSrcC(context);
+
+            Operand res;
+
+            if (high)
+            {
+                if (signedA && signedB)
+                {
+                    res = context.MultiplyHighS32(srcA, srcB);
+                }
+                else
+                {
+                    res = context.MultiplyHighU32(srcA, srcB);
+
+                    if (signedA)
+                    {
+                        res = context.IAdd(res, context.IMultiply(srcB, context.ShiftRightS32(srcA, Const(31))));
+                    }
+                    else if (signedB)
+                    {
+                        res = context.IAdd(res, context.IMultiply(srcA, context.ShiftRightS32(srcB, Const(31))));
+                    }
+                }
+            }
+            else
+            {
+                res = context.IMultiply(srcA, srcB);
+            }
+
+            res = context.IAdd(res, srcC);
+
+            context.Copy(GetDest(context), res);
+        }
+
         public static void Imnmx(EmitterContext context)
         {
             OpCodeAlu op = (OpCodeAlu)context.CurrOp;
diff --git a/Ryujinx.Graphics.Shader/Instructions/InstEmitFArith.cs b/Ryujinx.Graphics.Shader/Instructions/InstEmitFArith.cs
index 3b8d7305..63d1efcb 100644
--- a/Ryujinx.Graphics.Shader/Instructions/InstEmitFArith.cs
+++ b/Ryujinx.Graphics.Shader/Instructions/InstEmitFArith.cs
@@ -59,6 +59,26 @@ namespace Ryujinx.Graphics.Shader.Instructions
             SetFPZnFlags(context, dest, op.SetCondCode);
         }
 
+        public static void Ffma32i(EmitterContext context)
+        {
+            IOpCodeFArith op = (IOpCodeFArith)context.CurrOp;
+
+            bool saturate = op.RawOpCode.Extract(55);
+            bool negateA  = op.RawOpCode.Extract(56);
+            bool negateC  = op.RawOpCode.Extract(57);
+
+            Operand srcA = context.FPNegate(GetSrcA(context), negateA);
+            Operand srcC = context.FPNegate(GetDest(context), negateC);
+
+            Operand srcB = GetSrcB(context);
+
+            Operand dest = GetDest(context);
+
+            context.Copy(dest, context.FPSaturate(context.FPFusedMultiplyAdd(srcA, srcB, srcC), saturate));
+
+            SetFPZnFlags(context, dest, op.SetCondCode);
+        }
+
         public static void Fmnmx(EmitterContext context)
         {
             IOpCodeFArith op = (IOpCodeFArith)context.CurrOp;