Merge commit 'origin/master' into gallium-0.2

8 files changed, 297 insertions, 31 deletions

diff --git a/src/mesa/main/ffvertex_prog.c b/src/mesa/main/ffvertex_prog.c
index da2640dd8fc..64a4788aa4b 100644
--- a/src/mesa/main/ffvertex_prog.c
+++ b/src/mesa/main/ffvertex_prog.c
@@ -1306,14 +1306,13 @@ static void build_lighting( struct tnl_program *p )
             }
             else if (!p->state->material_shininess_is_zero) {
                emit_op1(p, OPCODE_LIT, lit, 0, dots);
-               emit_op2(p, OPCODE_ADD, _col0, 0, ambient, _col0);
+               emit_op2(p, OPCODE_ADD, _bfc0, 0, ambient, _bfc0);
             } 
             else {
                emit_degenerate_lit(p, lit, dots);
-               emit_op2(p, OPCODE_ADD, _col0, 0, ambient, _col0);
+               emit_op2(p, OPCODE_ADD, _bfc0, 0, ambient, _bfc0);
             }
 
-	    emit_op2(p, OPCODE_ADD, _bfc0, 0, ambient, _bfc0);
 	    emit_op3(p, OPCODE_MAD, res0, mask0, swizzle1(lit,Y), diffuse, _bfc0);
 	    emit_op3(p, OPCODE_MAD, res1, mask1, swizzle1(lit,Z), specular, _bfc1);
 
diff --git a/src/mesa/shader/prog_execute.c b/src/mesa/shader/prog_execute.c
index cc084c535cc..914056d7456 100644
--- a/src/mesa/shader/prog_execute.c
+++ b/src/mesa/shader/prog_execute.c
@@ -220,6 +220,37 @@ fetch_vector4(const struct prog_src_register *source,
 
 
 /**
+ * Fetch a 4-element uint vector from the given source register.
+ * Apply swizzling but not negation/abs.
+ */
+static void
+fetch_vector4ui(const struct prog_src_register *source,
+                const struct gl_program_machine *machine, GLuint result[4])
+{
+   const GLuint *src = (GLuint *) get_register_pointer(source, machine);
+   ASSERT(src);
+
+   if (source->Swizzle == SWIZZLE_NOOP) {
+      /* no swizzling */
+      COPY_4V(result, src);
+   }
+   else {
+      ASSERT(GET_SWZ(source->Swizzle, 0) <= 3);
+      ASSERT(GET_SWZ(source->Swizzle, 1) <= 3);
+      ASSERT(GET_SWZ(source->Swizzle, 2) <= 3);
+      ASSERT(GET_SWZ(source->Swizzle, 3) <= 3);
+      result[0] = src[GET_SWZ(source->Swizzle, 0)];
+      result[1] = src[GET_SWZ(source->Swizzle, 1)];
+      result[2] = src[GET_SWZ(source->Swizzle, 2)];
+      result[3] = src[GET_SWZ(source->Swizzle, 3)];
+   }
+
+   /* Note: no NegateBase, Abs, NegateAbs here */
+}
+
+
+
+/**
  * Fetch the derivative with respect to X or Y for the given register.
  * XXX this currently only works for fragment program input attribs.
  */
@@ -229,7 +260,8 @@ fetch_vector4_deriv(GLcontext * ctx,
                     const struct gl_program_machine *machine,
                     char xOrY, GLfloat result[4])
 {
-   if (source->File == PROGRAM_INPUT && source->Index < (GLint)machine->NumDeriv) {
+   if (source->File == PROGRAM_INPUT &&
+       source->Index < (GLint) machine->NumDeriv) {
       const GLint col = machine->CurElement;
       const GLfloat w = machine->Attribs[FRAG_ATTRIB_WPOS][col][3];
       const GLfloat invQ = 1.0f / w;
@@ -494,6 +526,89 @@ store_vector4(const struct prog_instruction *inst,
 
 
 /**
+ * Store 4 uints into a register.  Observe the set-condition-code flags.
+ */
+static void
+store_vector4ui(const struct prog_instruction *inst,
+                struct gl_program_machine *machine, const GLuint value[4])
+{
+   const struct prog_dst_register *dest = &(inst->DstReg);
+   GLuint *dstReg;
+   GLuint dummyReg[4];
+   GLuint writeMask = dest->WriteMask;
+
+   switch (dest->File) {
+   case PROGRAM_OUTPUT:
+      ASSERT(dest->Index < MAX_PROGRAM_OUTPUTS);
+      dstReg = (GLuint *) machine->Outputs[dest->Index];
+      break;
+   case PROGRAM_TEMPORARY:
+      ASSERT(dest->Index < MAX_PROGRAM_TEMPS);
+      dstReg = (GLuint *) machine->Temporaries[dest->Index];
+      break;
+   case PROGRAM_WRITE_ONLY:
+      dstReg = dummyReg;
+      return;
+   default:
+      _mesa_problem(NULL, "bad register file in store_vector4(fp)");
+      return;
+   }
+
+   if (dest->CondMask != COND_TR) {
+      /* condition codes may turn off some writes */
+      if (writeMask & WRITEMASK_X) {
+         if (!test_cc(machine->CondCodes[GET_SWZ(dest->CondSwizzle, 0)],
+                      dest->CondMask))
+            writeMask &= ~WRITEMASK_X;
+      }
+      if (writeMask & WRITEMASK_Y) {
+         if (!test_cc(machine->CondCodes[GET_SWZ(dest->CondSwizzle, 1)],
+                      dest->CondMask))
+            writeMask &= ~WRITEMASK_Y;
+      }
+      if (writeMask & WRITEMASK_Z) {
+         if (!test_cc(machine->CondCodes[GET_SWZ(dest->CondSwizzle, 2)],
+                      dest->CondMask))
+            writeMask &= ~WRITEMASK_Z;
+      }
+      if (writeMask & WRITEMASK_W) {
+         if (!test_cc(machine->CondCodes[GET_SWZ(dest->CondSwizzle, 3)],
+                      dest->CondMask))
+            writeMask &= ~WRITEMASK_W;
+      }
+   }
+
+   if (writeMask & WRITEMASK_X)
+      dstReg[0] = value[0];
+   if (writeMask & WRITEMASK_Y)
+      dstReg[1] = value[1];
+   if (writeMask & WRITEMASK_Z)
+      dstReg[2] = value[2];
+   if (writeMask & WRITEMASK_W)
+      dstReg[3] = value[3];
+
+   if (inst->CondUpdate) {
+      if (writeMask & WRITEMASK_X)
+         machine->CondCodes[0] = generate_cc(value[0]);
+      if (writeMask & WRITEMASK_Y)
+         machine->CondCodes[1] = generate_cc(value[1]);
+      if (writeMask & WRITEMASK_Z)
+         machine->CondCodes[2] = generate_cc(value[2]);
+      if (writeMask & WRITEMASK_W)
+         machine->CondCodes[3] = generate_cc(value[3]);
+#if DEBUG_PROG
+      printf("CondCodes=(%s,%s,%s,%s) for:\n",
+             _mesa_condcode_string(machine->CondCodes[0]),
+             _mesa_condcode_string(machine->CondCodes[1]),
+             _mesa_condcode_string(machine->CondCodes[2]),
+             _mesa_condcode_string(machine->CondCodes[3]));
+#endif
+   }
+}
+
+
+
+/**
  * Execute the given vertex/fragment program.
  *
  * \param ctx  rendering context
@@ -572,6 +687,18 @@ _mesa_execute_program(GLcontext * ctx,
             }
          }
          break;
+      case OPCODE_AND:     /* bitwise AND */
+         {
+            GLuint a[4], b[4], result[4];
+            fetch_vector4ui(&inst->SrcReg[0], machine, a);
+            fetch_vector4ui(&inst->SrcReg[1], machine, b);
+            result[0] = a[0] & b[0];
+            result[1] = a[1] & b[1];
+            result[2] = a[2] & b[2];
+            result[3] = a[3] & b[3];
+            store_vector4ui(inst, machine, result);
+         }
+         break;
       case OPCODE_ARL:
          {
             GLfloat t[4];
@@ -650,6 +777,33 @@ _mesa_execute_program(GLcontext * ctx,
             store_vector4(inst, machine, result);
          }
          break;
+      case OPCODE_DP2:
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(&inst->SrcReg[0], machine, a);
+            fetch_vector4(&inst->SrcReg[1], machine, b);
+            result[0] = result[1] = result[2] = result[3] = DOT2(a, b);
+            store_vector4(inst, machine, result);
+            if (DEBUG_PROG) {
+               printf("DP2 %g = (%g %g) . (%g %g)\n",
+                      result[0], a[0], a[1], b[0], b[1]);
+            }
+         }
+         break;
+      case OPCODE_DP2A:
+         {
+            GLfloat a[4], b[4], c, result[4];
+            fetch_vector4(&inst->SrcReg[0], machine, a);
+            fetch_vector4(&inst->SrcReg[1], machine, b);
+            fetch_vector1(&inst->SrcReg[1], machine, &c);
+            result[0] = result[1] = result[2] = result[3] = DOT2(a, b) + c;
+            store_vector4(inst, machine, result);
+            if (DEBUG_PROG) {
+               printf("DP2A %g = (%g %g) . (%g %g) + %g\n",
+                      result[0], a[0], a[1], b[0], b[1], c);
+            }
+         }
+         break;
       case OPCODE_DP3:
          {
             GLfloat a[4], b[4], result[4];
@@ -682,8 +836,7 @@ _mesa_execute_program(GLcontext * ctx,
             GLfloat a[4], b[4], result[4];
             fetch_vector4(&inst->SrcReg[0], machine, a);
             fetch_vector4(&inst->SrcReg[1], machine, b);
-            result[0] = result[1] = result[2] = result[3] =
-               a[0] * b[0] + a[1] * b[1] + a[2] * b[2] + b[3];
+            result[0] = result[1] = result[2] = result[3] = DOT3(a, b) + b[3];
             store_vector4(inst, machine, result);
          }
          break;
@@ -1020,39 +1173,94 @@ _mesa_execute_program(GLcontext * ctx,
          break;
       case OPCODE_NOP:
          break;
-      case OPCODE_PK2H:        /* pack two 16-bit floats in one 32-bit float */
+      case OPCODE_NOT:         /* bitwise NOT */
+         {
+            GLuint a[4], result[4];
+            fetch_vector4ui(&inst->SrcReg[0], machine, a);
+            result[0] = ~a[0];
+            result[1] = ~a[1];
+            result[2] = ~a[2];
+            result[3] = ~a[3];
+            store_vector4ui(inst, machine, result);
+         }
+         break;
+      case OPCODE_NRM3:        /* 3-component normalization */
          {
             GLfloat a[4], result[4];
+            GLfloat tmp;
+            fetch_vector4(&inst->SrcReg[0], machine, a);
+            tmp = a[0] * a[0] + a[1] * a[1] + a[2] * a[2];
+            if (tmp != 0.0F)
+               tmp = 1.0F / tmp;
+            result[0] = tmp * a[0];
+            result[1] = tmp * a[1];
+            result[2] = tmp * a[2];
+            result[3] = 0.0;  /* undefined, but prevent valgrind warnings */
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_NRM4:        /* 4-component normalization */
+         {
+            GLfloat a[4], result[4];
+            GLfloat tmp;
+            fetch_vector4(&inst->SrcReg[0], machine, a);
+            tmp = a[0] * a[0] + a[1] * a[1] + a[2] * a[2] + a[3] * a[3];
+            if (tmp != 0.0F)
+               tmp = 1.0F / tmp;
+            result[0] = tmp * a[0];
+            result[1] = tmp * a[1];
+            result[2] = tmp * a[2];
+            result[3] = tmp * a[3];
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_OR:          /* bitwise OR */
+         {
+            GLuint a[4], b[4], result[4];
+            fetch_vector4ui(&inst->SrcReg[0], machine, a);
+            fetch_vector4ui(&inst->SrcReg[1], machine, b);
+            result[0] = a[0] | b[0];
+            result[1] = a[1] | b[1];
+            result[2] = a[2] | b[2];
+            result[3] = a[3] | b[3];
+            store_vector4ui(inst, machine, result);
+         }
+         break;
+      case OPCODE_PK2H:        /* pack two 16-bit floats in one 32-bit float */
+         {
+            GLfloat a[4];
+            GLuint result[4];
             GLhalfNV hx, hy;
-            GLuint *rawResult = (GLuint *) result;
-            GLuint twoHalves;
             fetch_vector4(&inst->SrcReg[0], machine, a);
             hx = _mesa_float_to_half(a[0]);
             hy = _mesa_float_to_half(a[1]);
-            twoHalves = hx | (hy << 16);
-            rawResult[0] = rawResult[1] = rawResult[2] = rawResult[3]
-               = twoHalves;
-            store_vector4(inst, machine, result);
+            result[0] =
+            result[1] =
+            result[2] =
+            result[3] = hx | (hy << 16);
+            store_vector4ui(inst, machine, result);
          }
          break;
       case OPCODE_PK2US:       /* pack two GLushorts into one 32-bit float */
          {
-            GLfloat a[4], result[4];
-            GLuint usx, usy, *rawResult = (GLuint *) result;
+            GLfloat a[4];
+            GLuint result[4], usx, usy;
             fetch_vector4(&inst->SrcReg[0], machine, a);
             a[0] = CLAMP(a[0], 0.0F, 1.0F);
             a[1] = CLAMP(a[1], 0.0F, 1.0F);
             usx = IROUND(a[0] * 65535.0F);
             usy = IROUND(a[1] * 65535.0F);
-            rawResult[0] = rawResult[1] = rawResult[2] = rawResult[3]
-               = usx | (usy << 16);
-            store_vector4(inst, machine, result);
+            result[0] =
+            result[1] =
+            result[2] =
+            result[3] = usx | (usy << 16);
+            store_vector4ui(inst, machine, result);
          }
          break;
       case OPCODE_PK4B:        /* pack four GLbytes into one 32-bit float */
          {
-            GLfloat a[4], result[4];
-            GLuint ubx, uby, ubz, ubw, *rawResult = (GLuint *) result;
+            GLfloat a[4];
+            GLuint result[4], ubx, uby, ubz, ubw;
             fetch_vector4(&inst->SrcReg[0], machine, a);
             a[0] = CLAMP(a[0], -128.0F / 127.0F, 1.0F);
             a[1] = CLAMP(a[1], -128.0F / 127.0F, 1.0F);
@@ -1062,15 +1270,17 @@ _mesa_execute_program(GLcontext * ctx,
             uby = IROUND(127.0F * a[1] + 128.0F);
             ubz = IROUND(127.0F * a[2] + 128.0F);
             ubw = IROUND(127.0F * a[3] + 128.0F);
-            rawResult[0] = rawResult[1] = rawResult[2] = rawResult[3]
-               = ubx | (uby << 8) | (ubz << 16) | (ubw << 24);
-            store_vector4(inst, machine, result);
+            result[0] =
+            result[1] =
+            result[2] =
+            result[3] = ubx | (uby << 8) | (ubz << 16) | (ubw << 24);
+            store_vector4ui(inst, machine, result);
          }
          break;
       case OPCODE_PK4UB:       /* pack four GLubytes into one 32-bit float */
          {
-            GLfloat a[4], result[4];
-            GLuint ubx, uby, ubz, ubw, *rawResult = (GLuint *) result;
+            GLfloat a[4];
+            GLuint result[4], ubx, uby, ubz, ubw;
             fetch_vector4(&inst->SrcReg[0], machine, a);
             a[0] = CLAMP(a[0], 0.0F, 1.0F);
             a[1] = CLAMP(a[1], 0.0F, 1.0F);
@@ -1080,9 +1290,11 @@ _mesa_execute_program(GLcontext * ctx,
             uby = IROUND(255.0F * a[1]);
             ubz = IROUND(255.0F * a[2]);
             ubw = IROUND(255.0F * a[3]);
-            rawResult[0] = rawResult[1] = rawResult[2] = rawResult[3]
-               = ubx | (uby << 8) | (ubz << 16) | (ubw << 24);
-            store_vector4(inst, machine, result);
+            result[0] =
+            result[1] =
+            result[2] =
+            result[3] = ubx | (uby << 8) | (ubz << 16) | (ubw << 24);
+            store_vector4ui(inst, machine, result);
          }
          break;
       case OPCODE_POW:
@@ -1278,6 +1490,17 @@ _mesa_execute_program(GLcontext * ctx,
             }
          }
          break;
+      case OPCODE_SSG:         /* set sign (-1, 0 or +1) */
+         {
+            GLfloat a[4], result[4];
+            fetch_vector4(&inst->SrcReg[0], machine, a);
+            result[0] = (GLfloat) ((a[0] > 0.0F) - (a[0] < 0.0F));
+            result[1] = (GLfloat) ((a[1] > 0.0F) - (a[1] < 0.0F));
+            result[2] = (GLfloat) ((a[2] > 0.0F) - (a[2] < 0.0F));
+            result[3] = (GLfloat) ((a[3] > 0.0F) - (a[3] < 0.0F));
+            store_vector4(inst, machine, result);
+         }
+         break;
       case OPCODE_STR:         /* set true, operands ignored */
          {
             static const GLfloat result[4] = { 1.0F, 1.0F, 1.0F, 1.0F };
@@ -1476,6 +1699,18 @@ _mesa_execute_program(GLcontext * ctx,
             store_vector4(inst, machine, result);
          }
          break;
+      case OPCODE_XOR:         /* bitwise XOR */
+         {
+            GLuint a[4], b[4], result[4];
+            fetch_vector4ui(&inst->SrcReg[0], machine, a);
+            fetch_vector4ui(&inst->SrcReg[1], machine, b);
+            result[0] = a[0] ^ b[0];
+            result[1] = a[1] ^ b[1];
+            result[2] = a[2] ^ b[2];
+            result[3] = a[3] ^ b[3];
+            store_vector4ui(inst, machine, result);
+         }
+         break;
       case OPCODE_XPD:         /* cross product */
          {
             GLfloat a[4], b[4], result[4];
diff --git a/src/mesa/shader/prog_instruction.c b/src/mesa/shader/prog_instruction.c
index 7e340ce4545..54df3fb2e15 100644
--- a/src/mesa/shader/prog_instruction.c
+++ b/src/mesa/shader/prog_instruction.c
@@ -154,6 +154,7 @@ static const struct instruction_info InstInfo[MAX_OPCODE] = {
    { OPCODE_NOP,    "NOP",     0, 0 },
    { OPCODE_ABS,    "ABS",     1, 1 },
    { OPCODE_ADD,    "ADD",     2, 1 },
+   { OPCODE_AND,    "AND",     2, 1 },
    { OPCODE_ARA,    "ARA",     1, 1 },
    { OPCODE_ARL,    "ARL",     1, 1 },
    { OPCODE_ARL_NV, "ARL",     1, 1 },
@@ -168,6 +169,8 @@ static const struct instruction_info InstInfo[MAX_OPCODE] = {
    { OPCODE_COS,    "COS",     1, 1 },
    { OPCODE_DDX,    "DDX",     1, 1 },
    { OPCODE_DDY,    "DDY",     1, 1 },
+   { OPCODE_DP2,    "DP2",     2, 1 },
+   { OPCODE_DP2A,   "DP2A",    3, 1 },
    { OPCODE_DP3,    "DP3",     2, 1 },
    { OPCODE_DP4,    "DP4",     2, 1 },
    { OPCODE_DPH,    "DPH",     2, 1 },
@@ -193,10 +196,14 @@ static const struct instruction_info InstInfo[MAX_OPCODE] = {
    { OPCODE_MIN,    "MIN",     2, 1 },
    { OPCODE_MOV,    "MOV",     1, 1 },
    { OPCODE_MUL,    "MUL",     2, 1 },
+   { OPCODE_NOT,    "NOT",     1, 1 },
    { OPCODE_NOISE1, "NOISE1",  1, 1 },
    { OPCODE_NOISE2, "NOISE2",  1, 1 },
    { OPCODE_NOISE3, "NOISE3",  1, 1 },
    { OPCODE_NOISE4, "NOISE4",  1, 1 },
+   { OPCODE_OR,     "OR",      2, 1 },
+   { OPCODE_NRM3,   "NRM3",    1, 1 },
+   { OPCODE_NRM4,   "NRM4",    1, 1 },
    { OPCODE_PK2H,   "PK2H",    1, 1 },
    { OPCODE_PK2US,  "PK2US",   1, 1 },
    { OPCODE_PK4B,   "PK4B",    1, 1 },
@@ -235,6 +242,7 @@ static const struct instruction_info InstInfo[MAX_OPCODE] = {
    { OPCODE_UP4B,   "UP4B",    1, 1 },
    { OPCODE_UP4UB,  "UP4UB",   1, 1 },
    { OPCODE_X2D,    "X2D",     3, 1 },
+   { OPCODE_XOR,    "XOR",     2, 1 },
    { OPCODE_XPD,    "XPD",     2, 1 }
 };
 
diff --git a/src/mesa/shader/prog_instruction.h b/src/mesa/shader/prog_instruction.h
index 2be39ac6244..6a0e62221a0 100644
--- a/src/mesa/shader/prog_instruction.h
+++ b/src/mesa/shader/prog_instruction.h
@@ -145,6 +145,7 @@ typedef enum prog_opcode {
    OPCODE_NOP = 0,   /*                                      X   */
    OPCODE_ABS,       /*   X        X       1.1               X   */
    OPCODE_ADD,       /*   X        X       X       X         X   */
+   OPCODE_AND,       /*                                          */
    OPCODE_ARA,       /*                    2                     */
    OPCODE_ARL,       /*   X                X                     */
    OPCODE_ARL_NV,    /*                    2                     */
@@ -159,6 +160,8 @@ typedef enum prog_opcode {
    OPCODE_COS,       /*            X       2       X         X   */
    OPCODE_DDX,       /*                            X         X   */
    OPCODE_DDY,       /*                            X         X   */
+   OPCODE_DP2,       /*                            2             */
+   OPCODE_DP2A,      /*                            2             */
    OPCODE_DP3,       /*   X        X       X       X         X   */
    OPCODE_DP4,       /*   X        X       X       X         X   */
    OPCODE_DPH,       /*   X        X       1.1                   */
@@ -188,6 +191,10 @@ typedef enum prog_opcode {
    OPCODE_NOISE2,    /*                                      X   */
    OPCODE_NOISE3,    /*                                      X   */
    OPCODE_NOISE4,    /*                                      X   */
+   OPCODE_NOT,       /*                                          */
+   OPCODE_NRM3,      /*                                          */
+   OPCODE_NRM4,      /*                                          */
+   OPCODE_OR,        /*                                          */
    OPCODE_PK2H,      /*                            X             */
    OPCODE_PK2US,     /*                            X             */
    OPCODE_PK4B,      /*                            X             */
@@ -226,6 +233,7 @@ typedef enum prog_opcode {
    OPCODE_UP4B,      /*                            X             */
    OPCODE_UP4UB,     /*                            X             */
    OPCODE_X2D,       /*                            X             */
+   OPCODE_XOR,       /*                                          */
    OPCODE_XPD,       /*   X        X                         X   */
    MAX_OPCODE
 } gl_inst_opcode;
diff --git a/src/mesa/shader/slang/slang_codegen.c b/src/mesa/shader/slang/slang_codegen.c
index d83e3b01e6b..ea35d679694 100644
--- a/src/mesa/shader/slang/slang_codegen.c
+++ b/src/mesa/shader/slang/slang_codegen.c
@@ -413,6 +413,9 @@ static slang_asm_info AsmInfo[] = {
    { "vec4_multiply", IR_MUL, 1, 2 },
    { "vec4_dot", IR_DOT4, 1, 2 },
    { "vec3_dot", IR_DOT3, 1, 2 },
+   { "vec2_dot", IR_DOT2, 1, 2 },
+   { "vec3_nrm", IR_NRM3, 1, 1 },
+   { "vec4_nrm", IR_NRM4, 1, 1 },
    { "vec3_cross", IR_CROSS, 1, 2 },
    { "vec4_lrp", IR_LRP, 1, 3 },
    { "vec4_min", IR_MIN, 1, 2 },
diff --git a/src/mesa/shader/slang/slang_emit.c b/src/mesa/shader/slang/slang_emit.c
index 672ec4bd606..827760c917a 100644
--- a/src/mesa/shader/slang/slang_emit.c
+++ b/src/mesa/shader/slang/slang_emit.c
@@ -488,6 +488,9 @@ instruction_annotation(gl_inst_opcode opcode, char *dstAnnot,
    case OPCODE_MUL:
       operator = "*";
       break;
+   case OPCODE_DP2:
+      operator = "DP2";
+      break;
    case OPCODE_DP3:
       operator = "DP3";
       break;
@@ -708,7 +711,7 @@ emit_compare(slang_emit_info *emitInfo, slang_ir_node *n)
       }
       else {
          assert(size == 2);
-         dotOp = OPCODE_DP3;
+         dotOp = OPCODE_DP3; /* XXX use OPCODE_DP2 eventually */
          swizzle = MAKE_SWIZZLE4(SWIZZLE_X, SWIZZLE_Y, SWIZZLE_Y, SWIZZLE_Y);
       }
 
@@ -1893,12 +1896,15 @@ emit(slang_emit_info *emitInfo, slang_ir_node *n)
    case IR_NOISE2:
    case IR_NOISE3:
    case IR_NOISE4:
+   case IR_NRM4:
+   case IR_NRM3:
    /* binary */
    case IR_ADD:
    case IR_SUB:
    case IR_MUL:
    case IR_DOT4:
    case IR_DOT3:
+   case IR_DOT2:
    case IR_CROSS:
    case IR_MIN:
    case IR_MAX:
diff --git a/src/mesa/shader/slang/slang_ir.c b/src/mesa/shader/slang/slang_ir.c
index 20498e8c66a..9d055bf3543 100644
--- a/src/mesa/shader/slang/slang_ir.c
+++ b/src/mesa/shader/slang/slang_ir.c
@@ -27,6 +27,7 @@
 #include "main/context.h"
 #include "slang_ir.h"
 #include "slang_mem.h"
+#include "shader/prog_instruction.h"
 #include "shader/prog_print.h"
 
 
@@ -36,8 +37,11 @@ static const slang_ir_info IrInfo[] = {
    { IR_SUB, "IR_SUB", OPCODE_SUB, 4, 2 },
    { IR_MUL, "IR_MUL", OPCODE_MUL, 4, 2 },
    { IR_DIV, "IR_DIV", OPCODE_NOP, 0, 2 }, /* XXX broke */
-   { IR_DOT4, "IR_DOT_4", OPCODE_DP4, 1, 2 },
-   { IR_DOT3, "IR_DOT_3", OPCODE_DP3, 1, 2 },
+   { IR_DOT4, "IR_DOT4", OPCODE_DP4, 1, 2 },
+   { IR_DOT3, "IR_DOT3", OPCODE_DP3, 1, 2 },
+   { IR_DOT2, "IR_DOT2", OPCODE_DP2, 1, 2 },
+   { IR_NRM4, "IR_NRM4", OPCODE_NRM4, 1, 1 },
+   { IR_NRM3, "IR_NRM3", OPCODE_NRM3, 1, 1 },
    { IR_CROSS, "IR_CROSS", OPCODE_XPD, 3, 2 },
    { IR_LRP, "IR_LRP", OPCODE_LRP, 4, 3 },
    { IR_MIN, "IR_MIN", OPCODE_MIN, 4, 2 },
diff --git a/src/mesa/shader/slang/slang_ir.h b/src/mesa/shader/slang/slang_ir.h
index f64f9a93b78..ab0353c28a5 100644
--- a/src/mesa/shader/slang/slang_ir.h
+++ b/src/mesa/shader/slang/slang_ir.h
@@ -83,6 +83,9 @@ typedef enum
    IR_DIV,
    IR_DOT4,
    IR_DOT3,
+   IR_DOT2,
+   IR_NRM4,
+   IR_NRM3,
    IR_CROSS,   /* vec3 cross product */
    IR_LRP,
    IR_CLAMP,