From abf843a797876b5e3c5c91dbec25b6553d2cc281 Mon Sep 17 00:00:00 2001
From: Eric Anholt <eric@anholt.net>
Date: Tue, 9 Aug 2011 12:30:41 -0700
Subject: [PATCH] i965/vs: Add support for ir_binop_pow.

Fixes vs-pow-float-float.
---
 src/mesa/drivers/dri/i965/brw_vec4.h           | 10 ++++
 src/mesa/drivers/dri/i965/brw_vec4_emit.cpp    | 63 ++++++++++++++++++++++++--
 src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp |  4 +-
 3 files changed, 70 insertions(+), 7 deletions(-)

diff --git a/src/mesa/drivers/dri/i965/brw_vec4.h b/src/mesa/drivers/dri/i965/brw_vec4.h
index b5f442e..0820215 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4.h
+++ b/src/mesa/drivers/dri/i965/brw_vec4.h
@@ -444,12 +444,22 @@ public:
    void generate_vs_instruction(vec4_instruction *inst,
 				struct brw_reg dst,
 				struct brw_reg *src);
+
    void generate_math1_gen4(vec4_instruction *inst,
 			    struct brw_reg dst,
 			    struct brw_reg src);
    void generate_math1_gen6(vec4_instruction *inst,
 			    struct brw_reg dst,
 			    struct brw_reg src);
+   void generate_math2_gen4(vec4_instruction *inst,
+			    struct brw_reg dst,
+			    struct brw_reg src0,
+			    struct brw_reg src1);
+   void generate_math2_gen6(vec4_instruction *inst,
+			    struct brw_reg dst,
+			    struct brw_reg src0,
+			    struct brw_reg src1);
+
    void generate_urb_write(vec4_instruction *inst);
    void generate_oword_dual_block_offsets(struct brw_reg m1,
 					  struct brw_reg index);
diff --git a/src/mesa/drivers/dri/i965/brw_vec4_emit.cpp b/src/mesa/drivers/dri/i965/brw_vec4_emit.cpp
index effc82a..df9521c 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4_emit.cpp
+++ b/src/mesa/drivers/dri/i965/brw_vec4_emit.cpp
@@ -245,6 +245,15 @@ vec4_visitor::generate_math1_gen4(vec4_instruction *inst,
 	    BRW_MATH_PRECISION_FULL);
 }
 
+static void
+check_gen6_math_src_arg(struct brw_reg src)
+{
+   /* Source swizzles are ignored. */
+   assert(!src.abs);
+   assert(!src.negate);
+   assert(src.dw1.bits.swizzle = BRW_SWIZZLE_XYZW);
+}
+
 void
 vec4_visitor::generate_math1_gen6(vec4_instruction *inst,
 				  struct brw_reg dst,
@@ -252,10 +261,7 @@ vec4_visitor::generate_math1_gen6(vec4_instruction *inst,
 {
    /* Can't do writemask because math can't be align16. */
    assert(dst.dw1.bits.writemask == WRITEMASK_XYZW);
-   /* Source swizzles are ignored. */
-   assert(!src.abs);
-   assert(!src.negate);
-   assert(src.dw1.bits.swizzle = BRW_SWIZZLE_XYZW);
+   check_gen6_math_src_arg(src);
 
    brw_set_access_mode(p, BRW_ALIGN_1);
    brw_math(p,
@@ -270,6 +276,49 @@ vec4_visitor::generate_math1_gen6(vec4_instruction *inst,
 }
 
 void
+vec4_visitor::generate_math2_gen6(vec4_instruction *inst,
+				  struct brw_reg dst,
+				  struct brw_reg src0,
+				  struct brw_reg src1)
+{
+   /* Can't do writemask because math can't be align16. */
+   assert(dst.dw1.bits.writemask == WRITEMASK_XYZW);
+   /* Source swizzles are ignored. */
+   check_gen6_math_src_arg(src0);
+   check_gen6_math_src_arg(src1);
+
+   brw_set_access_mode(p, BRW_ALIGN_1);
+   brw_math2(p,
+	     dst,
+	     brw_math_function(inst->opcode),
+	     src0, src1);
+   brw_set_access_mode(p, BRW_ALIGN_16);
+}
+
+void
+vec4_visitor::generate_math2_gen4(vec4_instruction *inst,
+				  struct brw_reg dst,
+				  struct brw_reg src0,
+				  struct brw_reg src1)
+{
+   /* Can't do writemask because math can't be align16. */
+   assert(dst.dw1.bits.writemask == WRITEMASK_XYZW);
+
+   brw_MOV(p, brw_message_reg(inst->base_mrf + 1), src1);
+
+   brw_set_access_mode(p, BRW_ALIGN_1);
+   brw_math(p,
+	    dst,
+	    brw_math_function(inst->opcode),
+	    BRW_MATH_SATURATE_NONE,
+	    inst->base_mrf,
+	    src0,
+	    BRW_MATH_DATA_VECTOR,
+	    BRW_MATH_PRECISION_FULL);
+   brw_set_access_mode(p, BRW_ALIGN_16);
+}
+
+void
 vec4_visitor::generate_urb_write(vec4_instruction *inst)
 {
    brw_urb_WRITE(p,
@@ -442,7 +491,11 @@ vec4_visitor::generate_vs_instruction(vec4_instruction *instruction,
       break;
 
    case SHADER_OPCODE_POW:
-      assert(!"finishme");
+      if (intel->gen >= 6) {
+	 generate_math2_gen6(inst, dst, src[0], src[1]);
+      } else {
+	 generate_math2_gen4(inst, dst, src[0], src[1]);
+      }
       break;
 
    case VS_OPCODE_URB_WRITE:
diff --git a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp b/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
index f4756a9..f9447d7 100644
--- a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
+++ b/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
@@ -188,11 +188,11 @@ vec4_visitor::emit_math2_gen6(enum opcode opcode,
     */
 
    expanded = src_reg(this, glsl_type::vec4_type);
-   emit(BRW_OPCODE_MOV, dst, src0);
+   emit(BRW_OPCODE_MOV, dst_reg(expanded), src0);
    src0 = expanded;
 
    expanded = src_reg(this, glsl_type::vec4_type);
-   emit(BRW_OPCODE_MOV, dst, src1);
+   emit(BRW_OPCODE_MOV, dst_reg(expanded), src1);
    src1 = expanded;
 
    if (dst.writemask != WRITEMASK_XYZW) {
-- 
2.7.4