Merge remote branch 'origin/master' into pipe-video

Conflicts:
	src/gallium/drivers/r600/r600_asm.c
	src/gallium/tests/unit/SConscript

45 files changed, 299 insertions, 228 deletions

diff --git a/src/mesa/drivers/common/driverfuncs.c b/src/mesa/drivers/common/driverfuncs.c
index 5389a500111..854dea94504 100644
--- a/src/mesa/drivers/common/driverfuncs.c
+++ b/src/mesa/drivers/common/driverfuncs.c
@@ -42,6 +42,7 @@
 #include "main/fbobject.h"
 #include "main/texrender.h"
 #include "main/syncobj.h"
+#include "main/texturebarrier.h"
 #include "main/transformfeedback.h"
 
 #include "program/program.h"
@@ -188,6 +189,8 @@ _mesa_init_driver_functions(struct dd_function_table *driver)
 
    driver->BlitFramebuffer = _swrast_BlitFramebuffer;
 
+   _mesa_init_texture_barrier_functions(driver);
+
    /* APPLE_vertex_array_object */
    driver->NewArrayObject = _mesa_new_array_object;
    driver->DeleteArrayObject = _mesa_delete_array_object;
diff --git a/src/mesa/drivers/dri/common/utils.c b/src/mesa/drivers/dri/common/utils.c
index 42be77fd7c4..083edfaa9b4 100644
--- a/src/mesa/drivers/dri/common/utils.c
+++ b/src/mesa/drivers/dri/common/utils.c
@@ -648,6 +648,8 @@ driCreateConfigs(GLenum fb_format, GLenum fb_type,
 			__DRI_ATTRIB_TEXTURE_1D_BIT |
 			__DRI_ATTRIB_TEXTURE_2D_BIT |
 			__DRI_ATTRIB_TEXTURE_RECTANGLE_BIT;
+
+		    modes->sRGBCapable = GL_FALSE;
 		}
 	    }
 	}
@@ -727,6 +729,7 @@ static const struct { unsigned int attrib, offset; } attribMap[] = {
     __ATTRIB(__DRI_ATTRIB_BIND_TO_MIPMAP_TEXTURE,	bindToMipmapTexture),
     __ATTRIB(__DRI_ATTRIB_BIND_TO_TEXTURE_TARGETS,	bindToTextureTargets),
     __ATTRIB(__DRI_ATTRIB_YINVERTED,			yInverted),
+    __ATTRIB(__DRI_ATTRIB_FRAMEBUFFER_SRGB_CAPABLE,	sRGBCapable),
 
     /* The struct field doesn't matter here, these are handled by the
      * switch in driGetConfigAttribIndex.  We need them in the array
diff --git a/src/mesa/drivers/dri/i915/i830_state.c b/src/mesa/drivers/dri/i915/i830_state.c
index ef5b8d971da..9fecab10db9 100644
--- a/src/mesa/drivers/dri/i915/i830_state.c
+++ b/src/mesa/drivers/dri/i915/i830_state.c
@@ -31,6 +31,7 @@
 #include "main/macros.h"
 #include "main/enums.h"
 #include "main/dd.h"
+#include "main/state.h"
 
 #include "texmem.h"
 
@@ -234,7 +235,7 @@ i830EvalLogicOpBlendState(struct gl_context * ctx)
 
    I830_STATECHANGE(i830, I830_UPLOAD_CTX);
 
-   if (RGBA_LOGICOP_ENABLED(ctx)) {
+   if (_mesa_rgba_logicop_enabled(ctx)) {
       i830->state.Ctx[I830_CTXREG_ENABLES_1] &= ~(ENABLE_COLOR_BLEND |
                                                   ENABLE_LOGIC_OP_MASK);
       i830->state.Ctx[I830_CTXREG_ENABLES_1] |= (DISABLE_COLOR_BLEND |
@@ -679,7 +680,7 @@ update_specular(struct gl_context * ctx)
    I830_STATECHANGE(i830, I830_UPLOAD_CTX);
    i830->state.Ctx[I830_CTXREG_ENABLES_1] &= ~ENABLE_SPEC_ADD_MASK;
 
-   if (NEED_SECONDARY_COLOR(ctx))
+   if (_mesa_need_secondary_color(ctx))
       i830->state.Ctx[I830_CTXREG_ENABLES_1] |= ENABLE_SPEC_ADD;
    else
       i830->state.Ctx[I830_CTXREG_ENABLES_1] |= DISABLE_SPEC_ADD;
diff --git a/src/mesa/drivers/dri/i915/i915_fragprog.c b/src/mesa/drivers/dri/i915/i915_fragprog.c
index 25f4fc3c8b1..2bfe665cb65 100644
--- a/src/mesa/drivers/dri/i915/i915_fragprog.c
+++ b/src/mesa/drivers/dri/i915/i915_fragprog.c
@@ -269,7 +269,7 @@ translate_tex_src_target(struct i915_fragment_program *p, GLubyte bit)
 #define EMIT_TEX( OP )						\
 do {								\
    GLuint dim = translate_tex_src_target( p, inst->TexSrcTarget );	\
-   const struct gl_fragment_program *program = p->ctx->FragmentProgram._Current; \
+   const struct gl_fragment_program *program = &p->FragProg;	\
    GLuint unit = program->Base.SamplerUnits[inst->TexSrcUnit];	\
    GLuint sampler = i915_emit_decl(p, REG_TYPE_S,		\
 				   unit, dim);			\
@@ -304,7 +304,7 @@ do {									\
  */
 static void calc_live_regs( struct i915_fragment_program *p )
 {
-    const struct gl_fragment_program *program = p->ctx->FragmentProgram._Current;
+    const struct gl_fragment_program *program = &p->FragProg;
     GLuint regsUsed = 0xffff0000;
     uint8_t live_components[16] = { 0, };
     GLint i;
@@ -344,7 +344,7 @@ static void calc_live_regs( struct i915_fragment_program *p )
 static GLuint get_live_regs( struct i915_fragment_program *p, 
                              const struct prog_instruction *inst )
 {
-    const struct gl_fragment_program *program = p->ctx->FragmentProgram._Current;
+    const struct gl_fragment_program *program = &p->FragProg;
     GLuint nr = inst - program->Base.Instructions;
 
     return p->usedRegs[nr];
@@ -365,8 +365,7 @@ static GLuint get_live_regs( struct i915_fragment_program *p,
 static void
 upload_program(struct i915_fragment_program *p)
 {
-   const struct gl_fragment_program *program =
-      p->ctx->FragmentProgram._Current;
+   const struct gl_fragment_program *program = &p->FragProg;
    const struct prog_instruction *inst = program->Base.Instructions;
 
    if (INTEL_DEBUG & DEBUG_WM)
@@ -1166,7 +1165,7 @@ translate_program(struct i915_fragment_program *p)
 
    if (INTEL_DEBUG & DEBUG_WM) {
       printf("fp:\n");
-      _mesa_print_program(&p->ctx->FragmentProgram._Current->Base);
+      _mesa_print_program(&p->FragProg.Base);
       printf("\n");
    }
 
diff --git a/src/mesa/drivers/dri/i915/i915_state.c b/src/mesa/drivers/dri/i915/i915_state.c
index 58d4f5daffd..3b1af4c455e 100644
--- a/src/mesa/drivers/dri/i915/i915_state.c
+++ b/src/mesa/drivers/dri/i915/i915_state.c
@@ -31,6 +31,7 @@
 #include "main/macros.h"
 #include "main/enums.h"
 #include "main/dd.h"
+#include "main/state.h"
 #include "tnl/tnl.h"
 #include "tnl/t_context.h"
 
@@ -209,7 +210,7 @@ i915EvalLogicOpBlendState(struct gl_context * ctx)
    dw0 = i915->state.Ctx[I915_CTXREG_LIS5];
    dw1 = i915->state.Ctx[I915_CTXREG_LIS6];
 
-   if (RGBA_LOGICOP_ENABLED(ctx)) {
+   if (_mesa_rgba_logicop_enabled(ctx)) {
       dw0 |= S5_LOGICOP_ENABLE;
       dw1 &= ~S6_CBUF_BLEND_ENABLE;
    }
diff --git a/src/mesa/drivers/dri/i915/intel_tris.c b/src/mesa/drivers/dri/i915/intel_tris.c
index cf9291cdfca..7bcb72f42d0 100644
--- a/src/mesa/drivers/dri/i915/intel_tris.c
+++ b/src/mesa/drivers/dri/i915/intel_tris.c
@@ -490,6 +490,7 @@ static void
 intel_wpos_triangle(struct intel_context *intel,
                     intelVertexPtr v0, intelVertexPtr v1, intelVertexPtr v2)
 {
+   const struct gl_framebuffer *fb = intel->ctx.DrawBuffer;
    GLuint offset = intel->wpos_offset;
    GLuint size = intel->wpos_size;
    GLfloat *v0_wpos = (GLfloat *)((char *)v0 + offset);
@@ -500,10 +501,11 @@ intel_wpos_triangle(struct intel_context *intel,
    __memcpy(v1_wpos, v1, size);
    __memcpy(v2_wpos, v2, size);
 
-   v0_wpos[1] = -v0_wpos[1] + intel->ctx.DrawBuffer->Height;
-   v1_wpos[1] = -v1_wpos[1] + intel->ctx.DrawBuffer->Height;
-   v2_wpos[1] = -v2_wpos[1] + intel->ctx.DrawBuffer->Height;
-
+   if (!fb->Name) {
+      v0_wpos[1] = -v0_wpos[1] + fb->Height;
+      v1_wpos[1] = -v1_wpos[1] + fb->Height;
+      v2_wpos[1] = -v2_wpos[1] + fb->Height;
+   }
 
    intel_draw_triangle(intel, v0, v1, v2);
 }
@@ -513,6 +515,7 @@ static void
 intel_wpos_line(struct intel_context *intel,
                 intelVertexPtr v0, intelVertexPtr v1)
 {
+   const struct gl_framebuffer *fb = intel->ctx.DrawBuffer;
    GLuint offset = intel->wpos_offset;
    GLuint size = intel->wpos_size;
    GLfloat *v0_wpos = (GLfloat *)((char *)v0 + offset);
@@ -521,8 +524,10 @@ intel_wpos_line(struct intel_context *intel,
    __memcpy(v0_wpos, v0, size);
    __memcpy(v1_wpos, v1, size);
 
-   v0_wpos[1] = -v0_wpos[1] + intel->ctx.DrawBuffer->Height;
-   v1_wpos[1] = -v1_wpos[1] + intel->ctx.DrawBuffer->Height;
+   if (!fb->Name) {
+      v0_wpos[1] = -v0_wpos[1] + fb->Height;
+      v1_wpos[1] = -v1_wpos[1] + fb->Height;
+   }
 
    intel_draw_line(intel, v0, v1);
 }
@@ -531,12 +536,15 @@ intel_wpos_line(struct intel_context *intel,
 static void
 intel_wpos_point(struct intel_context *intel, intelVertexPtr v0)
 {
+   const struct gl_framebuffer *fb = intel->ctx.DrawBuffer;
    GLuint offset = intel->wpos_offset;
    GLuint size = intel->wpos_size;
    GLfloat *v0_wpos = (GLfloat *)((char *)v0 + offset);
 
    __memcpy(v0_wpos, v0, size);
-   v0_wpos[1] = -v0_wpos[1] + intel->ctx.DrawBuffer->Height;
+
+   if (!fb->Name)
+      v0_wpos[1] = -v0_wpos[1] + fb->Height;
 
    intel_draw_point(intel, v0);
 }
diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index 26a6388f342..7b0551a92bc 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -701,6 +701,9 @@ struct brw_context
    /* Used to give every program string a unique id
     */
    GLuint program_id;
+
+   int num_prepare_atoms, num_emit_atoms;
+   struct brw_tracked_state prepare_atoms[64], emit_atoms[64];
 };
 
 
diff --git a/src/mesa/drivers/dri/i965/brw_defines.h b/src/mesa/drivers/dri/i965/brw_defines.h
index 5496b4fdd3b..6c61aefd7d3 100644
--- a/src/mesa/drivers/dri/i965/brw_defines.h
+++ b/src/mesa/drivers/dri/i965/brw_defines.h
@@ -688,13 +688,13 @@
 #define BRW_SAMPLER_MESSAGE_SIMD8_LD                  3
 #define BRW_SAMPLER_MESSAGE_SIMD16_LD                 3
 
-#define BRW_SAMPLER_MESSAGE_SAMPLE_GEN5            0
-#define BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_GEN5       1
-#define BRW_SAMPLER_MESSAGE_SAMPLE_LOD_GEN5        2
-#define BRW_SAMPLER_MESSAGE_SAMPLE_COMPARE_GEN5    3
-#define BRW_SAMPLER_MESSAGE_SAMPLE_DERIVS_GEN5     4
-#define BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_COMPARE_GEN5 5
-#define BRW_SAMPLER_MESSAGE_SAMPLE_LOD_COMPARE_GEN5 6
+#define GEN5_SAMPLER_MESSAGE_SAMPLE              0
+#define GEN5_SAMPLER_MESSAGE_SAMPLE_BIAS         1
+#define GEN5_SAMPLER_MESSAGE_SAMPLE_LOD          2
+#define GEN5_SAMPLER_MESSAGE_SAMPLE_COMPARE      3
+#define GEN5_SAMPLER_MESSAGE_SAMPLE_DERIVS       4
+#define GEN5_SAMPLER_MESSAGE_SAMPLE_BIAS_COMPARE 5
+#define GEN5_SAMPLER_MESSAGE_SAMPLE_LOD_COMPARE  6
 
 /* for GEN5 only */
 #define BRW_SAMPLER_SIMD_MODE_SIMD4X2                   0
@@ -752,14 +752,14 @@
 #define BRW_DATAPORT_WRITE_MESSAGE_FLUSH_RENDER_CACHE               7
 
 /* GEN6 */
-#define BRW_DATAPORT_WRITE_MESSAGE_DWORD_ATOMIC_WRITE_GEN6		7
-#define BRW_DATAPORT_WRITE_MESSAGE_OWORD_BLOCK_WRITE_GEN6		8
-#define BRW_DATAPORT_WRITE_MESSAGE_OWORD_DUAL_BLOCK_WRITE_GEN6		9
-#define BRW_DATAPORT_WRITE_MESSAGE_MEDIA_BLOCK_WRITE_GEN6		10
-#define BRW_DATAPORT_WRITE_MESSAGE_DWORLD_SCATTERED_WRITE_GEN6		11
-#define BRW_DATAPORT_WRITE_MESSAGE_RENDER_TARGET_WRITE_GEN6		12
-#define BRW_DATAPORT_WRITE_MESSAGE_STREAMED_VB_WRITE_GEN6		13
-#define BRW_DATAPORT_WRITE_MESSAGE_RENDER_TARGET_UNORM_WRITE_GEN6	14
+#define GEN6_DATAPORT_WRITE_MESSAGE_DWORD_ATOMIC_WRITE              7
+#define GEN6_DATAPORT_WRITE_MESSAGE_OWORD_BLOCK_WRITE               8
+#define GEN6_DATAPORT_WRITE_MESSAGE_OWORD_DUAL_BLOCK_WRITE          9
+#define GEN6_DATAPORT_WRITE_MESSAGE_MEDIA_BLOCK_WRITE               10
+#define GEN6_DATAPORT_WRITE_MESSAGE_DWORLD_SCATTERED_WRITE          11
+#define GEN6_DATAPORT_WRITE_MESSAGE_RENDER_TARGET_WRITE             12
+#define GEN6_DATAPORT_WRITE_MESSAGE_STREAMED_VB_WRITE               13
+#define GEN6_DATAPORT_WRITE_MESSAGE_RENDER_TARGET_UNORM_WRITE       14
 
 #define BRW_MATH_FUNCTION_INV                              1
 #define BRW_MATH_FUNCTION_LOG                              2
diff --git a/src/mesa/drivers/dri/i965/brw_draw_upload.c b/src/mesa/drivers/dri/i965/brw_draw_upload.c
index e96c32a93a6..f1d00693168 100644
--- a/src/mesa/drivers/dri/i965/brw_draw_upload.c
+++ b/src/mesa/drivers/dri/i965/brw_draw_upload.c
@@ -291,7 +291,7 @@ static void brw_prepare_vertices(struct brw_context *brw)
    /* Accumulate the list of enabled arrays. */
    brw->vb.nr_enabled = 0;
    while (vs_inputs) {
-      GLuint i = _mesa_ffsll(vs_inputs) - 1;
+      GLuint i = ffs(vs_inputs) - 1;
       struct brw_vertex_element *input = &brw->vb.inputs[i];
 
       vs_inputs &= ~(1 << i);
diff --git a/src/mesa/drivers/dri/i965/brw_eu.h b/src/mesa/drivers/dri/i965/brw_eu.h
index 119ffc72370..2d2ed9de985 100644
--- a/src/mesa/drivers/dri/i965/brw_eu.h
+++ b/src/mesa/drivers/dri/i965/brw_eu.h
@@ -939,8 +939,8 @@ void brw_dp_READ_4_vs_relative(struct brw_compile *p,
  */
 struct brw_instruction *brw_IF(struct brw_compile *p, 
 			       GLuint execute_size);
-struct brw_instruction *brw_IF_gen6(struct brw_compile *p, uint32_t conditional,
-				    struct brw_reg src0, struct brw_reg src1);
+struct brw_instruction *gen6_IF(struct brw_compile *p, uint32_t conditional,
+				struct brw_reg src0, struct brw_reg src1);
 
 struct brw_instruction *brw_ELSE(struct brw_compile *p, 
 				 struct brw_instruction *if_insn);
@@ -958,9 +958,9 @@ struct brw_instruction *brw_WHILE(struct brw_compile *p,
 	       struct brw_instruction *patch_insn);
 
 struct brw_instruction *brw_BREAK(struct brw_compile *p, int pop_count);
-struct brw_instruction *brw_CONT_gen6(struct brw_compile *p,
-				      struct brw_instruction *do_insn);
 struct brw_instruction *brw_CONT(struct brw_compile *p, int pop_count);
+struct brw_instruction *gen6_CONT(struct brw_compile *p,
+				  struct brw_instruction *do_insn);
 /* Forward jumps:
  */
 void brw_land_fwd_jump(struct brw_compile *p, 
diff --git a/src/mesa/drivers/dri/i965/brw_eu_emit.c b/src/mesa/drivers/dri/i965/brw_eu_emit.c
index 88131c432ec..21ce92c9173 100644
--- a/src/mesa/drivers/dri/i965/brw_eu_emit.c
+++ b/src/mesa/drivers/dri/i965/brw_eu_emit.c
@@ -874,8 +874,8 @@ struct brw_instruction *brw_IF(struct brw_compile *p, GLuint execute_size)
 }
 
 struct brw_instruction *
-brw_IF_gen6(struct brw_compile *p, uint32_t conditional,
-	    struct brw_reg src0, struct brw_reg src1)
+gen6_IF(struct brw_compile *p, uint32_t conditional,
+	struct brw_reg src0, struct brw_reg src1)
 {
    struct brw_instruction *insn;
 
@@ -1061,8 +1061,8 @@ struct brw_instruction *brw_BREAK(struct brw_compile *p, int pop_count)
    return insn;
 }
 
-struct brw_instruction *brw_CONT_gen6(struct brw_compile *p,
-				      struct brw_instruction *do_insn)
+struct brw_instruction *gen6_CONT(struct brw_compile *p,
+				  struct brw_instruction *do_insn)
 {
    struct brw_instruction *insn;
    int br = 2;
@@ -1844,7 +1844,7 @@ void brw_fb_WRITE(struct brw_compile *p,
        /* headerless version, just submit color payload */
        src0 = brw_message_reg(msg_reg_nr);
 
-       msg_type = BRW_DATAPORT_WRITE_MESSAGE_RENDER_TARGET_WRITE_GEN6;
+       msg_type = GEN6_DATAPORT_WRITE_MESSAGE_RENDER_TARGET_WRITE;
    } else {
       insn->header.destreg__conditionalmod = msg_reg_nr;
 
diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp b/src/mesa/drivers/dri/i965/brw_fs.cpp
index ce7959b19de..8b3f5adb9f9 100644
--- a/src/mesa/drivers/dri/i965/brw_fs.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
@@ -78,15 +78,6 @@ brw_new_shader_program(struct gl_context *ctx, GLuint name)
 }
 
 GLboolean
-brw_compile_shader(struct gl_context *ctx, struct gl_shader *shader)
-{
-   if (!_mesa_ir_compile_shader(ctx, shader))
-      return GL_FALSE;
-
-   return GL_TRUE;
-}
-
-GLboolean
 brw_link_shader(struct gl_context *ctx, struct gl_shader_program *prog)
 {
    struct brw_context *brw = brw_context(ctx);
@@ -120,6 +111,14 @@ brw_link_shader(struct gl_context *ctx, struct gl_shader_program *prog)
       do_lower_texture_projection(shader->ir);
       do_vec_index_to_cond_assign(shader->ir);
       brw_do_cubemap_normalize(shader->ir);
+      lower_noise(shader->ir);
+      lower_quadop_vector(shader->ir, false);
+      lower_variable_index_to_cond_assign(shader->ir,
+					  GL_TRUE, /* input */
+					  GL_TRUE, /* output */
+					  GL_TRUE, /* temp */
+					  GL_TRUE /* uniform */
+					  );
 
       do {
 	 progress = false;
@@ -134,16 +133,6 @@ brw_link_shader(struct gl_context *ctx, struct gl_shader_program *prog)
 				   ) || progress;
 
 	 progress = do_common_optimization(shader->ir, true, 32) || progress;
-
-	 progress = lower_noise(shader->ir) || progress;
-	 progress =
-	    lower_variable_index_to_cond_assign(shader->ir,
-						GL_TRUE, /* input */
-						GL_TRUE, /* output */
-						GL_TRUE, /* temp */
-						GL_TRUE /* uniform */
-						) || progress;
-	 progress = lower_quadop_vector(shader->ir, false) || progress;
       } while (progress);
 
       validate_ir_tree(shader->ir);
@@ -1452,7 +1441,10 @@ fs_visitor::visit(ir_texture *ir)
    if (ir->shadow_comparitor)
       inst->shadow_compare = true;
 
-   if (c->key.tex_swizzles[inst->sampler] != SWIZZLE_NOOP) {
+   if (ir->type == glsl_type::float_type) {
+      /* Ignore DEPTH_TEXTURE_MODE swizzling. */
+      assert(ir->sampler->type->sampler_shadow);
+   } else if (c->key.tex_swizzles[inst->sampler] != SWIZZLE_NOOP) {
       fs_reg swizzle_dst = fs_reg(this, glsl_type::vec4_type);
 
       for (int i = 0; i < 4; i++) {
@@ -2139,6 +2131,17 @@ fs_visitor::emit_fb_writes()
    }
 
    if (c->key.nr_color_regions == 0) {
+      if (c->key.alpha_test && (this->frag_color || this->frag_data)) {
+	 /* If the alpha test is enabled but there's no color buffer,
+	  * we still need to send alpha out the pipeline to our null
+	  * renderbuffer.
+	  */
+	 color.reg_offset += 3;
+	 emit(fs_inst(BRW_OPCODE_MOV,
+		      fs_reg(MRF, color_mrf + 3),
+		      color));
+      }
+
       fs_inst *inst = emit(fs_inst(FS_OPCODE_FB_WRITE,
 				   reg_undef, reg_undef));
       inst->base_mrf = 0;
@@ -2301,23 +2304,23 @@ fs_visitor::generate_tex(fs_inst *inst, struct brw_reg dst, struct brw_reg src)
       switch (inst->opcode) {
       case FS_OPCODE_TEX:
 	 if (inst->shadow_compare) {
-	    msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_COMPARE_GEN5;
+	    msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_COMPARE;
 	 } else {
-	    msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_GEN5;
+	    msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE;
 	 }
 	 break;
       case FS_OPCODE_TXB:
 	 if (inst->shadow_compare) {
-	    msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_COMPARE_GEN5;
+	    msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_BIAS_COMPARE;
 	 } else {
-	    msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_GEN5;
+	    msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_BIAS;
 	 }
 	 break;
       case FS_OPCODE_TXL:
 	 if (inst->shadow_compare) {
-	    msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_LOD_COMPARE_GEN5;
+	    msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_LOD_COMPARE;
 	 } else {
-	    msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_LOD_GEN5;
+	    msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_LOD;
 	 }
 	 break;
       case FS_OPCODE_TXD:
@@ -3542,7 +3545,7 @@ fs_visitor::generate_code()
       case BRW_OPCODE_IF:
 	 if (inst->src[0].file != BAD_FILE) {
 	    assert(intel->gen >= 6);
-	    if_stack[if_stack_depth] = brw_IF_gen6(p, inst->conditional_mod, src[0], src[1]);
+	    if_stack[if_stack_depth] = gen6_IF(p, inst->conditional_mod, src[0], src[1]);
 	 } else {
 	    if_stack[if_stack_depth] = brw_IF(p, BRW_EXECUTE_8);
 	 }
@@ -3584,7 +3587,7 @@ fs_visitor::generate_code()
       case BRW_OPCODE_CONTINUE:
 	 /* FINISHME: We need to write the loop instruction support still. */
 	 if (intel->gen >= 6)
-	    brw_CONT_gen6(p, loop_stack[loop_stack_depth - 1]);
+	    gen6_CONT(p, loop_stack[loop_stack_depth - 1]);
 	 else
 	    brw_CONT(p, if_depth_in_loop[loop_stack_depth]);
 	 brw_set_predicate_control(p, BRW_PREDICATE_NONE);
diff --git a/src/mesa/drivers/dri/i965/brw_program.c b/src/mesa/drivers/dri/i965/brw_program.c
index 7d653327e30..ee68095fceb 100644
--- a/src/mesa/drivers/dri/i965/brw_program.c
+++ b/src/mesa/drivers/dri/i965/brw_program.c
@@ -243,7 +243,6 @@ void brwInitFragProgFuncs( struct dd_function_table *functions )
 
    functions->NewShader = brw_new_shader;
    functions->NewShaderProgram = brw_new_shader_program;
-   functions->CompileShader = brw_compile_shader;
    functions->LinkShader = brw_link_shader;
 }
 
diff --git a/src/mesa/drivers/dri/i965/brw_state_upload.c b/src/mesa/drivers/dri/i965/brw_state_upload.c
index 2d6fb37355d..6f521be6599 100644
--- a/src/mesa/drivers/dri/i965/brw_state_upload.c
+++ b/src/mesa/drivers/dri/i965/brw_state_upload.c
@@ -104,7 +104,7 @@ static const struct brw_tracked_state *gen4_atoms[] =
    &brw_constant_buffer
 };
 
-const struct brw_tracked_state *gen6_atoms[] =
+static const struct brw_tracked_state *gen6_atoms[] =
 {
    &brw_check_fallback,
 
@@ -169,7 +169,32 @@ const struct brw_tracked_state *gen6_atoms[] =
 
 void brw_init_state( struct brw_context *brw )
 {
+   const struct brw_tracked_state **atoms;
+   int num_atoms;
+
    brw_init_caches(brw);
+
+   if (brw->intel.gen >= 6) {
+      atoms = gen6_atoms;
+      num_atoms = ARRAY_SIZE(gen6_atoms);
+   } else {
+      atoms = gen4_atoms;
+      num_atoms = ARRAY_SIZE(gen4_atoms);
+   }
+
+   while (num_atoms--) {
+      assert((*atoms)->dirty.mesa |
+	     (*atoms)->dirty.brw |
+	     (*atoms)->dirty.cache);
+
+      if ((*atoms)->prepare)
+	 brw->prepare_atoms[brw->num_prepare_atoms++] = **atoms;
+      if ((*atoms)->emit)
+	 brw->emit_atoms[brw->num_emit_atoms++] = **atoms;
+      atoms++;
+   }
+   assert(brw->num_emit_atoms <= ARRAY_SIZE(brw->emit_atoms));
+   assert(brw->num_prepare_atoms <= ARRAY_SIZE(brw->prepare_atoms));
 }
 
 
@@ -186,7 +211,7 @@ static GLuint check_state( const struct brw_state_flags *a,
 {
    return ((a->mesa & b->mesa) |
 	   (a->brw & b->brw) |
-	   (a->cache & b->cache));
+	   (a->cache & b->cache)) != 0;
 }
 
 static void accumulate_state( struct brw_state_flags *a,
@@ -342,9 +367,9 @@ void brw_validate_state( struct brw_context *brw )
    struct gl_context *ctx = &brw->intel.ctx;
    struct intel_context *intel = &brw->intel;
    struct brw_state_flags *state = &brw->state.dirty;
+   const struct brw_tracked_state *atoms = brw->prepare_atoms;
+   int num_atoms = brw->num_prepare_atoms;
    GLuint i;
-   const struct brw_tracked_state **atoms;
-   int num_atoms;
 
    brw_clear_validated_bos(brw);
 
@@ -353,14 +378,6 @@ void brw_validate_state( struct brw_context *brw )
 
    brw_add_validated_bo(brw, intel->batch.bo);
 
-   if (intel->gen >= 6) {
-      atoms = gen6_atoms;
-      num_atoms = ARRAY_SIZE(gen6_atoms);
-   } else {
-      atoms = gen4_atoms;
-      num_atoms = ARRAY_SIZE(gen4_atoms);
-   }
-
    if (brw->emit_state_always) {
       state->mesa |= ~0;
       state->brw |= ~0;
@@ -384,15 +401,13 @@ void brw_validate_state( struct brw_context *brw )
 
    /* do prepare stage for all atoms */
    for (i = 0; i < num_atoms; i++) {
-      const struct brw_tracked_state *atom = atoms[i];
-
-      if (brw->intel.Fallback)
-         break;
+      const struct brw_tracked_state *atom = &atoms[i];
 
       if (check_state(state, &atom->dirty)) {
-         if (atom->prepare) {
-            atom->prepare(brw);
-        }
+	 atom->prepare(brw);
+
+	 if (brw->intel.Fallback)
+	    break;
       }
    }
 
@@ -415,20 +430,11 @@ void brw_validate_state( struct brw_context *brw )
 
 void brw_upload_state(struct brw_context *brw)
 {
-   struct intel_context *intel = &brw->intel;
    struct brw_state_flags *state = &brw->state.dirty;
+   const struct brw_tracked_state *atoms = brw->emit_atoms;
+   int num_atoms = brw->num_emit_atoms;
    int i;
    static int dirty_count = 0;
-   const struct brw_tracked_state **atoms;
-   int num_atoms;
-
-   if (intel->gen >= 6) {
-      atoms = gen6_atoms;
-      num_atoms = ARRAY_SIZE(gen6_atoms);
-   } else {
-      atoms = gen4_atoms;
-      num_atoms = ARRAY_SIZE(gen4_atoms);
-   }
 
    brw_clear_validated_bos(brw);
 
@@ -442,20 +448,14 @@ void brw_upload_state(struct brw_context *brw)
       prev = *state;
 
       for (i = 0; i < num_atoms; i++) {
-	 const struct brw_tracked_state *atom = atoms[i];
+	 const struct brw_tracked_state *atom = &atoms[i];
 	 struct brw_state_flags generated;
 
-	 assert(atom->dirty.mesa ||
-		atom->dirty.brw ||
-		atom->dirty.cache);
-
 	 if (brw->intel.Fallback)
 	    break;
 
 	 if (check_state(state, &atom->dirty)) {
-	    if (atom->emit) {
-	       atom->emit( brw );
-	    }
+	    atom->emit(brw);
 	 }
 
 	 accumulate_state(&examined, &atom->dirty);
@@ -471,15 +471,13 @@ void brw_upload_state(struct brw_context *brw)
    }
    else {
       for (i = 0; i < num_atoms; i++) {
-	 const struct brw_tracked_state *atom = atoms[i];
+	 const struct brw_tracked_state *atom = &atoms[i];
 
 	 if (brw->intel.Fallback)
 	    break;
 
 	 if (check_state(state, &atom->dirty)) {
-	    if (atom->emit) {
-	       atom->emit( brw );
-	    }
+	    atom->emit(brw);
 	 }
       }
    }
diff --git a/src/mesa/drivers/dri/i965/brw_vs_emit.c b/src/mesa/drivers/dri/i965/brw_vs_emit.c
index 6ec62554cc4..acacf374b75 100644
--- a/src/mesa/drivers/dri/i965/brw_vs_emit.c
+++ b/src/mesa/drivers/dri/i965/brw_vs_emit.c
@@ -2109,7 +2109,7 @@ void brw_vs_emit(struct brw_vs_compile *c )
       case OPCODE_CONT:
 	 brw_set_predicate_control(p, get_predicate(inst));
 	 if (intel->gen >= 6) {
-	    brw_CONT_gen6(p, loop_inst[loop_depth - 1]);
+	    gen6_CONT(p, loop_inst[loop_depth - 1]);
 	 } else {
 	    brw_CONT(p, if_depth_in_loop[loop_depth]);
 	 }
diff --git a/src/mesa/drivers/dri/i965/brw_wm.c b/src/mesa/drivers/dri/i965/brw_wm.c
index ee8212f6f16..ca51d1599a4 100644
--- a/src/mesa/drivers/dri/i965/brw_wm.c
+++ b/src/mesa/drivers/dri/i965/brw_wm.c
@@ -284,6 +284,7 @@ static void brw_wm_populate_key( struct brw_context *brw,
    /* Build the index for table lookup
     */
    /* _NEW_COLOR */
+   key->alpha_test = ctx->Color.AlphaEnabled;
    if (fp->program.UsesKill ||
        ctx->Color.AlphaEnabled)
       lookup |= IZ_PS_KILL_ALPHATEST_BIT;
@@ -364,8 +365,6 @@ static void brw_wm_populate_key( struct brw_context *brw,
 	    SWIZZLE_NIL
 	 };
 
-	 key->tex_swizzles[i] = SWIZZLE_NOOP;
-
 	 /* GL_DEPTH_TEXTURE_MODE is normally handled through
 	  * brw_wm_surface_state, but it applies to shadow compares as
 	  * well and our shadow compares always return the result in
@@ -379,9 +378,10 @@ static void brw_wm_populate_key( struct brw_context *brw,
 	    } else if (t->DepthMode == GL_LUMINANCE) {
 	       swizzles[3] = SWIZZLE_ONE;
 	    } else if (t->DepthMode == GL_RED) {
+	       /* See table 3.23 of the GL 3.0 spec. */
 	       swizzles[1] = SWIZZLE_ZERO;
 	       swizzles[2] = SWIZZLE_ZERO;
-	       swizzles[3] = SWIZZLE_ZERO;
+	       swizzles[3] = SWIZZLE_ONE;
 	    }
 	 }
 
diff --git a/src/mesa/drivers/dri/i965/brw_wm.h b/src/mesa/drivers/dri/i965/brw_wm.h
index d9cae75ab5b..c40d7bfae0a 100644
--- a/src/mesa/drivers/dri/i965/brw_wm.h
+++ b/src/mesa/drivers/dri/i965/brw_wm.h
@@ -64,6 +64,7 @@ struct brw_wm_prog_key {
    GLuint linear_color:1;  /**< linear interpolation vs perspective interp */
    GLuint nr_color_regions:5;
    GLuint render_to_fbo:1;
+   GLuint alpha_test:1;
 
    GLbitfield proj_attrib_mask; /**< one bit per fragment program attribute */
    GLuint shadowtex_mask:16;
@@ -467,8 +468,6 @@ void emit_xpd(struct brw_compile *p,
 	      const struct brw_reg *arg0,
 	      const struct brw_reg *arg1);
 
-GLboolean brw_compile_shader(struct gl_context *ctx,
-			     struct gl_shader *shader);
 GLboolean brw_link_shader(struct gl_context *ctx, struct gl_shader_program *prog);
 struct gl_shader *brw_new_shader(struct gl_context *ctx, GLuint name, GLuint type);
 struct gl_shader_program *brw_new_shader_program(struct gl_context *ctx, GLuint name);
diff --git a/src/mesa/drivers/dri/i965/brw_wm_emit.c b/src/mesa/drivers/dri/i965/brw_wm_emit.c
index 2336e27c1ef..ecfd21d4399 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_emit.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_emit.c
@@ -1134,9 +1134,9 @@ void emit_tex(struct brw_wm_compile *c,
 
    if (intel->gen >= 5) {
       if (shadow)
-	 msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_COMPARE_GEN5;
+	 msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_COMPARE;
       else
-	 msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_GEN5;
+	 msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE;
    } else {
       /* Note that G45 and older determines shadow compare and dispatch width
        * from message length for most messages.
@@ -1186,14 +1186,14 @@ void emit_txb(struct brw_wm_compile *c,
     */
    if (c->dispatch_width == 16 || intel->gen < 5) {
       if (intel->gen >= 5)
-	 msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_GEN5;
+	 msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_BIAS;
       else
 	 msg_type = BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_BIAS;
       mrf_per_channel = 2;
       dst_retyped = retype(vec16(dst[0]), BRW_REGISTER_TYPE_UW);
       response_length = 8;
    } else {
-      msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_GEN5;
+      msg_type = GEN5_SAMPLER_MESSAGE_SAMPLE_BIAS;
       mrf_per_channel = 1;
       dst_retyped = retype(vec8(dst[0]), BRW_REGISTER_TYPE_UW);
       response_length = 4;
diff --git a/src/mesa/drivers/dri/intel/intel_pixel_bitmap.c b/src/mesa/drivers/dri/intel/intel_pixel_bitmap.c
index d398775906e..43cdd0d2bac 100644
--- a/src/mesa/drivers/dri/intel/intel_pixel_bitmap.c
+++ b/src/mesa/drivers/dri/intel/intel_pixel_bitmap.c
@@ -208,7 +208,7 @@ do_blit_bitmap( struct gl_context *ctx,
 
    COPY_4V(tmpColor, ctx->Current.RasterColor);
 
-   if (NEED_SECONDARY_COLOR(ctx)) {
+   if (_mesa_need_secondary_color(ctx)) {
        ADD_3V(tmpColor, tmpColor, ctx->Current.RasterSecondaryColor);
    }
 
diff --git a/src/mesa/drivers/dri/intel/intel_regions.c b/src/mesa/drivers/dri/intel/intel_regions.c
index 0857fa8ad70..a4da1ce4fa5 100644
--- a/src/mesa/drivers/dri/intel/intel_regions.c
+++ b/src/mesa/drivers/dri/intel/intel_regions.c
@@ -149,11 +149,6 @@ intel_region_alloc_internal(struct intel_screen *screen,
 {
    struct intel_region *region;
 
-   if (buffer == NULL) {
-      _DBG("%s <-- NULL\n", __FUNCTION__);
-      return NULL;
-   }
-
    region = calloc(sizeof(*region), 1);
    if (region == NULL)
       return region;
@@ -180,6 +175,7 @@ intel_region_alloc(struct intel_screen *screen,
    drm_intel_bo *buffer;
    unsigned long flags = 0;
    unsigned long aligned_pitch;
+   struct intel_region *region;
 
    if (expect_accelerated_upload)
       flags |= BO_ALLOC_FOR_RENDER;
@@ -187,9 +183,17 @@ intel_region_alloc(struct intel_screen *screen,
    buffer = drm_intel_bo_alloc_tiled(screen->bufmgr, "region",
 				     width, height, cpp,
 				     &tiling, &aligned_pitch, flags);
+   if (buffer == NULL)
+      return NULL;
 
-   return intel_region_alloc_internal(screen, cpp, width, height,
-				      aligned_pitch / cpp, tiling, buffer);
+   region = intel_region_alloc_internal(screen, cpp, width, height,
+                                        aligned_pitch / cpp, tiling, buffer);
+   if (region == NULL) {
+      drm_intel_bo_unreference(buffer);
+      return NULL;
+   }
+
+   return region;
 }
 
 GLboolean
diff --git a/src/mesa/drivers/dri/intel/intel_screen.c b/src/mesa/drivers/dri/intel/intel_screen.c
index 5c95c72732e..64a21a147f0 100644
--- a/src/mesa/drivers/dri/intel/intel_screen.c
+++ b/src/mesa/drivers/dri/intel/intel_screen.c
@@ -25,6 +25,7 @@
  * 
  **************************************************************************/
 
+#include <errno.h>
 #include "main/glheader.h"
 #include "main/context.h"
 #include "main/framebuffer.h"
@@ -302,7 +303,8 @@ intel_get_param(__DRIscreen *psp, int param, int *value)
 
    ret = drmCommandWriteRead(psp->fd, DRM_I915_GETPARAM, &gp, sizeof(gp));
    if (ret) {
-      _mesa_warning(NULL, "drm_i915_getparam: %d", ret);
+      if (ret != -EINVAL)
+	 _mesa_warning(NULL, "drm_i915_getparam: %d", ret);
       return GL_FALSE;
    }
 
diff --git a/src/mesa/drivers/dri/intel/intel_tex_subimage.c b/src/mesa/drivers/dri/intel/intel_tex_subimage.c
index 6b7f13ff353..d0f8294113a 100644
--- a/src/mesa/drivers/dri/intel/intel_tex_subimage.c
+++ b/src/mesa/drivers/dri/intel/intel_tex_subimage.c
@@ -90,19 +90,19 @@ intelTexSubimage(struct gl_context * ctx,
 	  intel->gen < 6 && target == GL_TEXTURE_2D &&
 	  drm_intel_bo_busy(dst_bo))
       {
-	 unsigned long pitch;
-	 uint32_t tiling_mode = I915_TILING_NONE;
-	 temp_bo = drm_intel_bo_alloc_tiled(intel->bufmgr,
-					    "subimage blit bo",
-					    width, height,
-					    intelImage->mt->cpp,
-					    &tiling_mode,
-					    &pitch,
-					    0);
-	 drm_intel_gem_bo_map_gtt(temp_bo);
+	 dstRowStride = width * intelImage->mt->cpp;
+         temp_bo = drm_intel_bo_alloc(intel->bufmgr, "subimage blit bo",
+                                      dstRowStride * height, 0);
+         if (!temp_bo)
+            return;
+
+	 if (drm_intel_gem_bo_map_gtt(temp_bo)) {
+            drm_intel_bo_unreference(temp_bo);
+            return;
+         }
+
 	 texImage->Data = temp_bo->virtual;
 	 texImage->ImageOffsets[0] = 0;
-	 dstRowStride = pitch;
 
 	 intel_miptree_get_image_offset(intelImage->mt, level,
 					intelImage->face, 0,
diff --git a/src/mesa/drivers/dri/intel/intel_tex_validate.c b/src/mesa/drivers/dri/intel/intel_tex_validate.c
index 8537e7f3682..a11b07ed09d 100644
--- a/src/mesa/drivers/dri/intel/intel_tex_validate.c
+++ b/src/mesa/drivers/dri/intel/intel_tex_validate.c
@@ -140,6 +140,8 @@ intel_finalize_mipmap_tree(struct intel_context *intel, GLuint unit)
                                           cpp,
                                           comp_byte,
 					  GL_TRUE);
+      if (!intelObj->mt)
+         return GL_FALSE;
    }
 
    /* Pull in any images not in the object's tree:
diff --git a/src/mesa/drivers/dri/mga/mgastate.c b/src/mesa/drivers/dri/mga/mgastate.c
index 2fac2b49cd2..bfc55f4fc6d 100644
--- a/src/mesa/drivers/dri/mga/mgastate.c
+++ b/src/mesa/drivers/dri/mga/mgastate.c
@@ -30,6 +30,7 @@
 #include "main/colormac.h"
 #include "main/dd.h"
 #include "main/mm.h"
+#include "main/state.h"
 
 #include "mgacontext.h"
 #include "mgadd.h"
@@ -114,7 +115,7 @@ static void mgaDDAlphaFunc(struct gl_context *ctx, GLenum func, GLfloat ref)
 static void updateBlendLogicOp(struct gl_context *ctx)
 {
    mgaContextPtr mmesa = MGA_CONTEXT(ctx);
-   GLboolean logicOp = RGBA_LOGICOP_ENABLED(ctx);
+   GLboolean logicOp = _mesa_rgba_logicop_enabled(ctx);
 
    MGA_STATECHANGE( mmesa, MGA_UPLOAD_CONTEXT );
 
@@ -197,7 +198,7 @@ static void mgaDDBlendFuncSeparate( struct gl_context *ctx, GLenum sfactorRGB,
    mmesa->hw.blend_func = (src | dst);
 
    FALLBACK( ctx, MGA_FALLBACK_BLEND,
-             ctx->Color.BlendEnabled && !RGBA_LOGICOP_ENABLED(ctx) &&
+             ctx->Color.BlendEnabled && !_mesa_rgba_logicop_enabled(ctx) &&
              mmesa->hw.blend_func == (AC_src_src_alpha_sat | AC_dst_zero) );
 }
 
@@ -483,7 +484,7 @@ static void updateSpecularLighting( struct gl_context *ctx )
    mgaContextPtr mmesa = MGA_CONTEXT(ctx);
    unsigned int specen;
 
-   specen = NEED_SECONDARY_COLOR(ctx) ? TMC_specen_enable : 0;
+   specen = _mesa_need_secondary_color(ctx) ? TMC_specen_enable : 0;
 
    if ( specen != mmesa->hw.specen ) {
       mmesa->hw.specen = specen;
@@ -962,7 +963,7 @@ void mgaEmitHwStateLocked( mgaContextPtr mmesa )
 	  ? mmesa->hw.zmode : (DC_zmode_nozcmp | DC_atype_i);
 
       mmesa->setup.dwgctl &= DC_bop_MASK;
-      mmesa->setup.dwgctl |= RGBA_LOGICOP_ENABLED(ctx)
+      mmesa->setup.dwgctl |= _mesa_rgba_logicop_enabled(ctx)
 	  ? mmesa->hw.rop : mgarop_NoBLK[ GL_COPY & 0x0f ];
 
       mmesa->setup.alphactrl &= AC_src_MASK & AC_dst_MASK & AC_atmode_MASK
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_driver.h b/src/mesa/drivers/dri/nouveau/nouveau_driver.h
index c5ac1282d0d..158aec820aa 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_driver.h
+++ b/src/mesa/drivers/dri/nouveau/nouveau_driver.h
@@ -31,6 +31,7 @@
 #include "main/mtypes.h"
 #include "main/macros.h"
 #include "main/formats.h"
+#include "main/state.h"
 #include "utils.h"
 #include "dri_util.h"
 
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_texture.c b/src/mesa/drivers/dri/nouveau/nouveau_texture.c
index 988208ff56e..1a1e10e0b3a 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_texture.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_texture.c
@@ -30,6 +30,7 @@
 #include "nouveau_fbo.h"
 #include "nouveau_util.h"
 
+#include "main/pbo.h"
 #include "main/texobj.h"
 #include "main/texstore.h"
 #include "main/texformat.h"
diff --git a/src/mesa/drivers/dri/nouveau/nv04_state_raster.c b/src/mesa/drivers/dri/nouveau/nv04_state_raster.c
index ecfbdfedff6..78d29fc485b 100644
--- a/src/mesa/drivers/dri/nouveau/nv04_state_raster.c
+++ b/src/mesa/drivers/dri/nouveau/nv04_state_raster.c
@@ -277,7 +277,7 @@ nv04_emit_blend(struct gl_context *ctx, int emit)
 			blend |= NV04_MULTITEX_TRIANGLE_BLEND_SHADE_MODE_FLAT;
 
 		/* Secondary color */
-		if (NEED_SECONDARY_COLOR(ctx))
+		if (_mesa_need_secondary_color(ctx))
 			blend |= NV04_MULTITEX_TRIANGLE_BLEND_SPECULAR_ENABLE;
 
 		/* Fog. */
@@ -315,7 +315,7 @@ nv04_emit_blend(struct gl_context *ctx, int emit)
 			blend |= get_texenv_mode(GL_MODULATE);
 
 		/* Secondary color */
-		if (NEED_SECONDARY_COLOR(ctx))
+		if (_mesa_need_secondary_color(ctx))
 			blend |= NV04_TEXTURED_TRIANGLE_BLEND_SPECULAR_ENABLE;
 
 		/* Fog. */
diff --git a/src/mesa/drivers/dri/nouveau/nv10_state_tnl.c b/src/mesa/drivers/dri/nouveau/nv10_state_tnl.c
index e8bd12e6e01..96d1b320d86 100644
--- a/src/mesa/drivers/dri/nouveau/nv10_state_tnl.c
+++ b/src/mesa/drivers/dri/nouveau/nv10_state_tnl.c
@@ -199,7 +199,7 @@ nv10_emit_light_model(struct gl_context *ctx, int emit)
 	BEGIN_RING(chan, celsius, NV10_3D_LIGHT_MODEL, 1);
 	OUT_RING(chan, ((m->LocalViewer ?
 			 NV10_3D_LIGHT_MODEL_LOCAL_VIEWER : 0) |
-			(NEED_SECONDARY_COLOR(ctx) ?
+			(_mesa_need_secondary_color(ctx) ?
 			 NV10_3D_LIGHT_MODEL_SEPARATE_SPECULAR : 0) |
 			(!ctx->Light.Enabled && ctx->Fog.ColorSumEnabled ?
 			 NV10_3D_LIGHT_MODEL_VERTEX_SPECULAR : 0)));
diff --git a/src/mesa/drivers/dri/nouveau/nv20_state_tnl.c b/src/mesa/drivers/dri/nouveau/nv20_state_tnl.c
index 4677198dd02..4f7ddd8e49f 100644
--- a/src/mesa/drivers/dri/nouveau/nv20_state_tnl.c
+++ b/src/mesa/drivers/dri/nouveau/nv20_state_tnl.c
@@ -178,7 +178,7 @@ nv20_emit_light_model(struct gl_context *ctx, int emit)
 	OUT_RING(chan, ((m->LocalViewer ?
 			 NV20_3D_LIGHT_MODEL_VIEWER_LOCAL :
 			 NV20_3D_LIGHT_MODEL_VIEWER_NONLOCAL) |
-			(NEED_SECONDARY_COLOR(ctx) ?
+			(_mesa_need_secondary_color(ctx) ?
 			 NV20_3D_LIGHT_MODEL_SEPARATE_SPECULAR :
 			 0)));
 
diff --git a/src/mesa/drivers/dri/r128/r128_state.c b/src/mesa/drivers/dri/r128/r128_state.c
index d6725cdd0ec..7ce082ead22 100644
--- a/src/mesa/drivers/dri/r128/r128_state.c
+++ b/src/mesa/drivers/dri/r128/r128_state.c
@@ -43,6 +43,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "main/enums.h"
 #include "main/colormac.h"
 #include "main/macros.h"
+#include "main/state.h"
 #include "swrast/swrast.h"
 #include "vbo/vbo.h"
 #include "tnl/tnl.h"
@@ -736,7 +737,7 @@ static void updateSpecularLighting( struct gl_context *ctx )
    r128ContextPtr rmesa = R128_CONTEXT(ctx);
    GLuint t = rmesa->setup.tex_cntl_c;
 
-   if ( NEED_SECONDARY_COLOR( ctx ) ) {
+   if ( _mesa_need_secondary_color( ctx ) ) {
       if (ctx->Light.ShadeModel == GL_FLAT) {
          /* R128 can't do flat-shaded separate specular */
          t &= ~R128_SPEC_LIGHT_ENABLE;
diff --git a/src/mesa/drivers/dri/r300/compiler/r300_fragprog_swizzle.c b/src/mesa/drivers/dri/r300/compiler/r300_fragprog_swizzle.c
index fa906f2fdde..5223aaa71a4 100644
--- a/src/mesa/drivers/dri/r300/compiler/r300_fragprog_swizzle.c
+++ b/src/mesa/drivers/dri/r300/compiler/r300_fragprog_swizzle.c
@@ -98,9 +98,6 @@ static int r300_swizzle_is_native(rc_opcode opcode, struct rc_src_register reg)
 	unsigned int relevant;
 	int j;
 
-	if (reg.Abs)
-		reg.Negate = RC_MASK_NONE;
-
 	if (opcode == RC_OPCODE_KIL ||
 	    opcode == RC_OPCODE_TEX ||
 	    opcode == RC_OPCODE_TXB ||
@@ -140,9 +137,6 @@ static void r300_swizzle_split(
 		struct rc_src_register src, unsigned int mask,
 		struct rc_swizzle_split * split)
 {
-	if (src.Abs)
-		src.Negate = RC_MASK_NONE;
-
 	split->NumPhases = 0;
 
 	while(mask) {
diff --git a/src/mesa/drivers/dri/r300/compiler/r3xx_fragprog.c b/src/mesa/drivers/dri/r300/compiler/r3xx_fragprog.c
index 1616306afbd..9286733635f 100644
--- a/src/mesa/drivers/dri/r300/compiler/r3xx_fragprog.c
+++ b/src/mesa/drivers/dri/r300/compiler/r3xx_fragprog.c
@@ -78,12 +78,32 @@ static void rc_rewrite_depth_out(struct radeon_compiler *cc, void *user)
 	}
 }
 
+static int radeon_saturate_output(
+		struct radeon_compiler * c,
+		struct rc_instruction * inst,
+		void* data)
+{
+	const struct rc_opcode_info *info = rc_get_opcode_info(inst->U.I.Opcode);
+
+	if (!info->HasDstReg || inst->U.I.DstReg.File != RC_FILE_OUTPUT)
+		return 0;
+
+	inst->U.I.SaturateMode = RC_SATURATE_ZERO_ONE;
+	return 1;
+}
+
 void r3xx_compile_fragment_program(struct r300_fragment_program_compiler* c)
 {
 	int is_r500 = c->Base.is_r500;
 	int opt = !c->Base.disable_optimizations;
+	int sat_out = c->state.frag_clamp;
 
 	/* Lists of instruction transformations. */
+	struct radeon_program_transformation saturate_output[] = {
+		{ &radeon_saturate_output, c },
+		{ 0, 0 }
+	};
+
 	struct radeon_program_transformation rewrite_tex[] = {
 		{ &radeonTransformTEX, c },
 		{ 0, 0 }
@@ -113,6 +133,7 @@ void r3xx_compile_fragment_program(struct r300_fragment_program_compiler* c)
 		{"unroll loops",		1, is_r500,	rc_unroll_loops,		NULL},
 		{"transform loops",		1, !is_r500,	rc_transform_loops,		NULL},
 		{"emulate branches",		1, !is_r500,	rc_emulate_branches,		NULL},
+		{"saturate output writes",	1, sat_out,	rc_local_transform,		saturate_output},
 		{"transform TEX",		1, 1,		rc_local_transform,		rewrite_tex},
 		{"native rewrite",		1, is_r500,	rc_local_transform,		native_rewrite_r500},
 		{"native rewrite",		1, !is_r500,	rc_local_transform,		native_rewrite_r300},
diff --git a/src/mesa/drivers/dri/r300/compiler/r500_fragprog.c b/src/mesa/drivers/dri/r300/compiler/r500_fragprog.c
index ef81be48f77..140eeed3de3 100644
--- a/src/mesa/drivers/dri/r300/compiler/r500_fragprog.c
+++ b/src/mesa/drivers/dri/r300/compiler/r500_fragprog.c
@@ -77,9 +77,6 @@ static int r500_swizzle_is_native(rc_opcode opcode, struct rc_src_register reg)
 		if (opcode == RC_OPCODE_KIL && (reg.Swizzle != RC_SWIZZLE_XYZW || reg.Negate != RC_MASK_NONE))
 			return 0;
 
-		if (reg.Negate)
-			reg.Negate ^= RC_MASK_XYZW;
-
 		for(i = 0; i < 4; ++i) {
 			unsigned int swz = GET_SWZ(reg.Swizzle, i);
 			if (swz == RC_SWIZZLE_UNUSED) {
@@ -103,9 +100,6 @@ static int r500_swizzle_is_native(rc_opcode opcode, struct rc_src_register reg)
 		return 0;
 	} else {
 		/* ALU instructions support almost everything */
-		if (reg.Abs)
-			return 1;
-
 		relevant = 0;
 		for(i = 0; i < 3; ++i) {
 			unsigned int swz = GET_SWZ(reg.Swizzle, i);
diff --git a/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c b/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c
index 1febc19cc2d..301b4446693 100644
--- a/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c
+++ b/src/mesa/drivers/dri/r300/compiler/r500_fragprog_emit.c
@@ -372,7 +372,7 @@ static int emit_tex(struct r300_fragment_program_compiler *c, struct rc_sub_inst
 		| (inst->DstReg.WriteMask << 11)
 		| R500_INST_TEX_SEM_WAIT;
 	code->inst[ip].inst1 = R500_TEX_ID(inst->TexSrcUnit)
-		| R500_TEX_SEM_ACQUIRE | R500_TEX_IGNORE_UNCOVERED;
+		| R500_TEX_SEM_ACQUIRE;
 
 	if (inst->TexSrcTarget == RC_TEXTURE_RECT)
 		code->inst[ip].inst1 |= R500_TEX_UNSCALED;
diff --git a/src/mesa/drivers/dri/r300/compiler/radeon_code.h b/src/mesa/drivers/dri/r300/compiler/radeon_code.h
index d1451668947..35360aa70f0 100644
--- a/src/mesa/drivers/dri/r300/compiler/radeon_code.h
+++ b/src/mesa/drivers/dri/r300/compiler/radeon_code.h
@@ -173,6 +173,8 @@ struct r300_fragment_program_external_state {
 		 * RC_STATE_R300_TEXSCALE_FACTOR. */
 		unsigned clamp_and_scale_before_fetch : 1;
 	} unit[16];
+
+	unsigned frag_clamp:1;
 };
 
 
diff --git a/src/mesa/drivers/dri/r300/compiler/radeon_optimize.c b/src/mesa/drivers/dri/r300/compiler/radeon_optimize.c
index 5caff91b00e..c4e6a5e0a1f 100644
--- a/src/mesa/drivers/dri/r300/compiler/radeon_optimize.c
+++ b/src/mesa/drivers/dri/r300/compiler/radeon_optimize.c
@@ -311,7 +311,18 @@ static void constant_folding(struct radeon_compiler * c, struct rc_instruction *
 		struct rc_constant * constant;
 		struct rc_src_register newsrc;
 		int have_real_reference;
+		unsigned int chan;
+
+		/* If there are only 0, 0.5, 1, or _ swizzles, mark the source as a constant. */
+		for (chan = 0; chan < 4; ++chan)
+			if (GET_SWZ(inst->U.I.SrcReg[src].Swizzle, chan) <= 3)
+				break;
+		if (chan == 4) {
+			inst->U.I.SrcReg[src].File = RC_FILE_NONE;
+			continue;
+		}
 
+		/* Convert immediates to swizzles. */
 		if (inst->U.I.SrcReg[src].File != RC_FILE_CONSTANT ||
 		    inst->U.I.SrcReg[src].RelAddr ||
 		    inst->U.I.SrcReg[src].Index >= c->Program.Constants.Count)
@@ -325,7 +336,7 @@ static void constant_folding(struct radeon_compiler * c, struct rc_instruction *
 
 		newsrc = inst->U.I.SrcReg[src];
 		have_real_reference = 0;
-		for(unsigned int chan = 0; chan < 4; ++chan) {
+		for (chan = 0; chan < 4; ++chan) {
 			unsigned int swz = GET_SWZ(newsrc.Swizzle, chan);
 			unsigned int newswz;
 			float imm;
diff --git a/src/mesa/drivers/dri/r300/compiler/radeon_pair_translate.c b/src/mesa/drivers/dri/r300/compiler/radeon_pair_translate.c
index 6d7263b4ab6..9e03eb1aca8 100644
--- a/src/mesa/drivers/dri/r300/compiler/radeon_pair_translate.c
+++ b/src/mesa/drivers/dri/r300/compiler/radeon_pair_translate.c
@@ -214,16 +214,21 @@ static void set_pair_instruction(struct r300_fragment_program_compiler *c,
 		if (needrgb && !istranscendent) {
 			unsigned int srcrgb = 0;
 			unsigned int srcalpha = 0;
+			unsigned int srcmask = 0;
 			int j;
 			/* We don't care about the alpha channel here.  We only
 			 * want the part of the swizzle that writes to rgb,
 			 * since we are creating an rgb instruction. */
 			for(j = 0; j < 3; ++j) {
 				unsigned int swz = GET_SWZ(inst->SrcReg[i].Swizzle, j);
-				if (swz < 3)
+
+				if (swz < RC_SWIZZLE_W)
 					srcrgb = 1;
-				else if (swz < 4)
+				else if (swz == RC_SWIZZLE_W)
 					srcalpha = 1;
+
+				if (swz < RC_SWIZZLE_UNUSED)
+					srcmask |= 1 << j;
 			}
 			source = rc_pair_alloc_source(pair, srcrgb, srcalpha,
 							inst->SrcReg[i].File, inst->SrcReg[i].Index);
@@ -236,7 +241,7 @@ static void set_pair_instruction(struct r300_fragment_program_compiler *c,
 			pair->RGB.Arg[i].Swizzle =
 				rc_init_swizzle(inst->SrcReg[i].Swizzle, 3);
 			pair->RGB.Arg[i].Abs = inst->SrcReg[i].Abs;
-			pair->RGB.Arg[i].Negate = !!(inst->SrcReg[i].Negate & (RC_MASK_X | RC_MASK_Y | RC_MASK_Z));
+			pair->RGB.Arg[i].Negate = !!(srcmask & inst->SrcReg[i].Negate & (RC_MASK_X | RC_MASK_Y | RC_MASK_Z));
 		}
 		if (needalpha) {
 			unsigned int srcrgb = 0;
diff --git a/src/mesa/drivers/dri/r300/compiler/radeon_program_tex.c b/src/mesa/drivers/dri/r300/compiler/radeon_program_tex.c
index f9d9f34b6ad..1cf77d9cf73 100644
--- a/src/mesa/drivers/dri/r300/compiler/radeon_program_tex.c
+++ b/src/mesa/drivers/dri/r300/compiler/radeon_program_tex.c
@@ -33,14 +33,14 @@
 /* Series of transformations to be done on textures. */
 
 static struct rc_src_register shadow_ambient(struct r300_fragment_program_compiler *compiler,
-											 int tmu)
+					     int tmu)
 {
 	struct rc_src_register reg = { 0, };
 
 	if (compiler->enable_shadow_ambient) {
 		reg.File = RC_FILE_CONSTANT;
 		reg.Index = rc_constants_add_state(&compiler->Base.Program.Constants,
-										   RC_STATE_SHADOW_AMBIENT, tmu);
+						   RC_STATE_SHADOW_AMBIENT, tmu);
 		reg.Swizzle = RC_SWIZZLE_WWWW;
 	} else {
 		reg.File = RC_FILE_NONE;
@@ -149,14 +149,11 @@ int radeonTransformTEX(
 
 			return 1;
 		} else {
-			rc_compare_func comparefunc = compiler->state.unit[inst->U.I.TexSrcUnit].texture_compare_func;
 			struct rc_instruction * inst_rcp = NULL;
-			struct rc_instruction * inst_mad;
-			struct rc_instruction * inst_cmp;
+			struct rc_instruction *inst_mul, *inst_add, *inst_cmp;
 			unsigned tmp_texsample;
 			unsigned tmp_sum;
-			unsigned tmp_recip_w = 0;
-			int pass, fail, tex;
+			int pass, fail;
 
 			/* Save the output register. */
 			struct rc_dst_register output_reg = inst->U.I.DstReg;
@@ -167,63 +164,68 @@ int radeonTransformTEX(
 			inst->U.I.DstReg.Index = tmp_texsample;
 			inst->U.I.DstReg.WriteMask = RC_MASK_XYZW;
 
-			if (inst->U.I.Opcode == RC_OPCODE_TXP) {
-				tmp_recip_w = rc_find_free_temporary(c);
+			tmp_sum = rc_find_free_temporary(c);
 
+			if (inst->U.I.Opcode == RC_OPCODE_TXP) {
 				/* Compute 1/W. */
 				inst_rcp = rc_insert_new_instruction(c, inst);
 				inst_rcp->U.I.Opcode = RC_OPCODE_RCP;
 				inst_rcp->U.I.DstReg.File = RC_FILE_TEMPORARY;
-				inst_rcp->U.I.DstReg.Index = tmp_recip_w;
+				inst_rcp->U.I.DstReg.Index = tmp_sum;
 				inst_rcp->U.I.DstReg.WriteMask = RC_MASK_W;
 				inst_rcp->U.I.SrcReg[0] = inst->U.I.SrcReg[0];
 				inst_rcp->U.I.SrcReg[0].Swizzle =
 					RC_MAKE_SWIZZLE_SMEAR(GET_SWZ(inst->U.I.SrcReg[0].Swizzle, 3));
 			}
 
-			/* Perspective-divide Z by W (if it's TXP) and add the texture sample (see below). */
-			tmp_sum = rc_find_free_temporary(c);
-			inst_mad = rc_insert_new_instruction(c, inst_rcp ? inst_rcp : inst);
-			inst_mad->U.I.DstReg.File = RC_FILE_TEMPORARY;
-			inst_mad->U.I.DstReg.Index = tmp_sum;
-			inst_mad->U.I.SrcReg[0] = inst->U.I.SrcReg[0];
-			inst_mad->U.I.SrcReg[0].Swizzle =
+			/* Divide Z by W (if it's TXP) and saturate. */
+			inst_mul = rc_insert_new_instruction(c, inst_rcp ? inst_rcp : inst);
+			inst_mul->U.I.Opcode = inst->U.I.Opcode == RC_OPCODE_TXP ? RC_OPCODE_MUL : RC_OPCODE_MOV;
+			inst_mul->U.I.DstReg.File = RC_FILE_TEMPORARY;
+			inst_mul->U.I.DstReg.Index = tmp_sum;
+			inst_mul->U.I.DstReg.WriteMask = RC_MASK_W;
+			inst_mul->U.I.SaturateMode = RC_SATURATE_ZERO_ONE;
+			inst_mul->U.I.SrcReg[0] = inst->U.I.SrcReg[0];
+			inst_mul->U.I.SrcReg[0].Swizzle =
 				RC_MAKE_SWIZZLE_SMEAR(GET_SWZ(inst->U.I.SrcReg[0].Swizzle, 2));
 			if (inst->U.I.Opcode == RC_OPCODE_TXP) {
-				inst_mad->U.I.Opcode = RC_OPCODE_MAD;
-				inst_mad->U.I.SrcReg[1].File = RC_FILE_TEMPORARY;
-				inst_mad->U.I.SrcReg[1].Index = tmp_recip_w;
-				inst_mad->U.I.SrcReg[1].Swizzle = RC_SWIZZLE_WWWW;
-				tex = 2;
-			} else {
-				inst_mad->U.I.Opcode = RC_OPCODE_ADD;
-				tex = 1;
-			}
-			inst_mad->U.I.SrcReg[tex].File = RC_FILE_TEMPORARY;
-			inst_mad->U.I.SrcReg[tex].Index = tmp_texsample;
-			inst_mad->U.I.SrcReg[tex].Swizzle = compiler->state.unit[inst->U.I.TexSrcUnit].depth_texture_swizzle;
-
-			/* Fake EQUAL/NOTEQUAL, it seems to pass some tests suprisingly. */
-			if (comparefunc == RC_COMPARE_FUNC_EQUAL) {
-				comparefunc = RC_COMPARE_FUNC_GEQUAL;
-			} else if (comparefunc == RC_COMPARE_FUNC_NOTEQUAL) {
-				comparefunc = RC_COMPARE_FUNC_LESS;
+				inst_mul->U.I.SrcReg[1].File = RC_FILE_TEMPORARY;
+				inst_mul->U.I.SrcReg[1].Index = tmp_sum;
+				inst_mul->U.I.SrcReg[1].Swizzle = RC_SWIZZLE_WWWW;
 			}
 
-			/* Recall that SrcReg[0] is r, SrcReg[tex] is tex and:
+			/* Add the depth texture value. */
+			inst_add = rc_insert_new_instruction(c, inst_mul);
+			inst_add->U.I.Opcode = RC_OPCODE_ADD;
+			inst_add->U.I.DstReg.File = RC_FILE_TEMPORARY;
+			inst_add->U.I.DstReg.Index = tmp_sum;
+			inst_add->U.I.DstReg.WriteMask = RC_MASK_W;
+			inst_add->U.I.SrcReg[0].File = RC_FILE_TEMPORARY;
+			inst_add->U.I.SrcReg[0].Index = tmp_sum;
+			inst_add->U.I.SrcReg[0].Swizzle = RC_SWIZZLE_WWWW;
+			inst_add->U.I.SrcReg[1].File = RC_FILE_TEMPORARY;
+			inst_add->U.I.SrcReg[1].Index = tmp_texsample;
+			inst_add->U.I.SrcReg[1].Swizzle = RC_SWIZZLE_XXXX;
+
+			/* Note that SrcReg[0] is r, SrcReg[1] is tex and:
 			 *   LESS:    r  < tex  <=>      -tex+r < 0
 			 *   GEQUAL:  r >= tex  <=> not (-tex+r < 0)
 			 *   GREATER: r  > tex  <=>       tex-r < 0
 			 *   LEQUAL:  r <= tex  <=> not ( tex-r < 0)
-			 *
-			 * This negates either r or tex: */
-			if (comparefunc == RC_COMPARE_FUNC_LESS || comparefunc == RC_COMPARE_FUNC_GEQUAL)
-				inst_mad->U.I.SrcReg[tex].Negate = inst_mad->U.I.SrcReg[tex].Negate ^ RC_MASK_XYZW;
+			 *   EQUAL:   GEQUAL
+			 *   NOTEQUAL:LESS
+			 */
+
+			/* This negates either r or tex: */
+			if (comparefunc == RC_COMPARE_FUNC_LESS || comparefunc == RC_COMPARE_FUNC_GEQUAL ||
+			    comparefunc == RC_COMPARE_FUNC_EQUAL || comparefunc == RC_COMPARE_FUNC_NOTEQUAL)
+				inst_add->U.I.SrcReg[1].Negate = inst_add->U.I.SrcReg[1].Negate ^ RC_MASK_XYZW;
 			else
-				inst_mad->U.I.SrcReg[0].Negate = inst_mad->U.I.SrcReg[0].Negate ^ RC_MASK_XYZW;
+				inst_add->U.I.SrcReg[0].Negate = inst_add->U.I.SrcReg[0].Negate ^ RC_MASK_XYZW;
 
 			/* This negates the whole expresion: */
-			if (comparefunc == RC_COMPARE_FUNC_LESS || comparefunc == RC_COMPARE_FUNC_GREATER) {
+			if (comparefunc == RC_COMPARE_FUNC_LESS || comparefunc == RC_COMPARE_FUNC_GREATER ||
+			    comparefunc == RC_COMPARE_FUNC_NOTEQUAL) {
 				pass = 1;
 				fail = 2;
 			} else {
@@ -231,16 +233,19 @@ int radeonTransformTEX(
 				fail = 1;
 			}
 
-			inst_cmp = rc_insert_new_instruction(c, inst_mad);
+			inst_cmp = rc_insert_new_instruction(c, inst_add);
 			inst_cmp->U.I.Opcode = RC_OPCODE_CMP;
 			inst_cmp->U.I.DstReg = output_reg;
 			inst_cmp->U.I.SrcReg[0].File = RC_FILE_TEMPORARY;
 			inst_cmp->U.I.SrcReg[0].Index = tmp_sum;
+			inst_cmp->U.I.SrcReg[0].Swizzle =
+					combine_swizzles(RC_SWIZZLE_WWWW,
+							 compiler->state.unit[inst->U.I.TexSrcUnit].depth_texture_swizzle);
 			inst_cmp->U.I.SrcReg[pass].File = RC_FILE_NONE;
 			inst_cmp->U.I.SrcReg[pass].Swizzle = RC_SWIZZLE_1111;
 			inst_cmp->U.I.SrcReg[fail] = shadow_ambient(compiler, inst->U.I.TexSrcUnit);
 
-			assert(tmp_texsample != tmp_sum && tmp_sum != tmp_recip_w);
+			assert(tmp_texsample != tmp_sum);
 		}
 	}
 
@@ -420,17 +425,21 @@ int radeonTransformTEX(
 		scale_texcoords(compiler, inst, RC_STATE_R300_TEXSCALE_FACTOR);
 	}
 
-	/* Cannot write texture to output registers (all chips) or with masks (non-r500) */
+	/* Cannot write texture to output registers or with saturate (all chips),
+	 * or with masks (non-r500). */
 	if (inst->U.I.Opcode != RC_OPCODE_KIL &&
 		(inst->U.I.DstReg.File != RC_FILE_TEMPORARY ||
+		 inst->U.I.SaturateMode ||
 		 (!c->is_r500 && inst->U.I.DstReg.WriteMask != RC_MASK_XYZW))) {
 		struct rc_instruction * inst_mov = rc_insert_new_instruction(c, inst);
 
 		inst_mov->U.I.Opcode = RC_OPCODE_MOV;
+		inst_mov->U.I.SaturateMode = inst->U.I.SaturateMode;
 		inst_mov->U.I.DstReg = inst->U.I.DstReg;
 		inst_mov->U.I.SrcReg[0].File = RC_FILE_TEMPORARY;
 		inst_mov->U.I.SrcReg[0].Index = rc_find_free_temporary(c);
 
+		inst->U.I.SaturateMode = 0;
 		inst->U.I.DstReg.File = RC_FILE_TEMPORARY;
 		inst->U.I.DstReg.Index = inst_mov->U.I.SrcReg[0].Index;
 		inst->U.I.DstReg.WriteMask = RC_MASK_XYZW;
diff --git a/src/mesa/drivers/dri/r300/r300_state.c b/src/mesa/drivers/dri/r300/r300_state.c
index 9df9101bcd3..51989c6b224 100644
--- a/src/mesa/drivers/dri/r300/r300_state.c
+++ b/src/mesa/drivers/dri/r300/r300_state.c
@@ -214,7 +214,7 @@ static void r300SetBlendState(struct gl_context * ctx)
 	    (R300_BLEND_GL_ZERO << R300_DST_BLEND_SHIFT);
 	int eqnA = R300_COMB_FCN_ADD_CLAMP;
 
-	if (RGBA_LOGICOP_ENABLED(ctx) || !ctx->Color.BlendEnabled) {
+	if (_mesa_rgba_logicop_enabled(ctx) || !ctx->Color.BlendEnabled) {
 		r300SetBlendCntl(r300, func, eqn, 0, func, eqn);
 		return;
 	}
@@ -335,7 +335,7 @@ static void r300SetLogicOpState(struct gl_context *ctx)
 {
 	r300ContextPtr r300 = R300_CONTEXT(ctx);
 	R300_STATECHANGE(r300, rop);
-	if (RGBA_LOGICOP_ENABLED(ctx)) {
+	if (_mesa_rgba_logicop_enabled(ctx)) {
 		r300->hw.rop.cmd[1] = R300_RB3D_ROPCNTL_ROP_ENABLE |
 			translate_logicop(ctx->Color.LogicOp);
 	} else {
@@ -349,7 +349,7 @@ static void r300SetLogicOpState(struct gl_context *ctx)
  */
 static void r300LogicOpcode(struct gl_context *ctx, GLenum logicop)
 {
-	if (RGBA_LOGICOP_ENABLED(ctx))
+	if (_mesa_rgba_logicop_enabled(ctx))
 		r300SetLogicOpState(ctx);
 }
 
diff --git a/src/mesa/drivers/dri/r600/evergreen_state.c b/src/mesa/drivers/dri/r600/evergreen_state.c
index 006e50007b6..309c93fe088 100644
--- a/src/mesa/drivers/dri/r600/evergreen_state.c
+++ b/src/mesa/drivers/dri/r600/evergreen_state.c
@@ -32,6 +32,7 @@
 #include "main/context.h"
 #include "main/dd.h"
 #include "main/simple_list.h"
+#include "main/state.h"
 
 #include "tnl/tnl.h"
 #include "tnl/t_pipeline.h"
@@ -342,7 +343,7 @@ static void evergreenSetBlendState(struct gl_context * ctx) //diff : CB_COLOR_CO
 
 	EVERGREEN_STATECHANGE(context, cb);
 
-	if (RGBA_LOGICOP_ENABLED(ctx) || !ctx->Color.BlendEnabled) {
+	if (_mesa_rgba_logicop_enabled(ctx) || !ctx->Color.BlendEnabled) {
 		SETfield(blend_reg,
 			 BLEND_ONE, COLOR_SRCBLEND_shift, COLOR_SRCBLEND_mask);
 		SETfield(blend_reg,
@@ -520,7 +521,7 @@ static void evergreenSetLogicOpState(struct gl_context *ctx) //diff : CB_COLOR_C
 
 	EVERGREEN_STATECHANGE(context, cb);
 
-	if (RGBA_LOGICOP_ENABLED(ctx))
+	if (_mesa_rgba_logicop_enabled(ctx))
 		SETfield(evergreen->CB_COLOR_CONTROL.u32All,
 			 evergreen_translate_logicop(ctx->Color.LogicOp), 
              EG_CB_COLOR_CONTROL__ROP3_shift, 
@@ -1148,7 +1149,7 @@ static void evergreenShadeModel(struct gl_context * ctx, GLenum mode) //same
 
 static void evergreenLogicOpcode(struct gl_context *ctx, GLenum logicop) //diff
 {
-	if (RGBA_LOGICOP_ENABLED(ctx))
+	if (_mesa_rgba_logicop_enabled(ctx))
 		evergreenSetLogicOpState(ctx);
 }
 
diff --git a/src/mesa/drivers/dri/r600/r700_state.c b/src/mesa/drivers/dri/r600/r700_state.c
index f8770690ff6..4d285b36489 100644
--- a/src/mesa/drivers/dri/r600/r700_state.c
+++ b/src/mesa/drivers/dri/r600/r700_state.c
@@ -32,6 +32,7 @@
 #include "main/context.h"
 #include "main/dd.h"
 #include "main/simple_list.h"
+#include "main/state.h"
 
 #include "tnl/tnl.h"
 #include "tnl/t_pipeline.h"
@@ -453,7 +454,7 @@ static void r700SetBlendState(struct gl_context * ctx)
 
 	R600_STATECHANGE(context, blnd);
 
-	if (RGBA_LOGICOP_ENABLED(ctx) || !ctx->Color.BlendEnabled) {
+	if (_mesa_rgba_logicop_enabled(ctx) || !ctx->Color.BlendEnabled) {
 		SETfield(blend_reg,
 			 BLEND_ONE, COLOR_SRCBLEND_shift, COLOR_SRCBLEND_mask);
 		SETfield(blend_reg,
@@ -644,7 +645,7 @@ static void r700SetLogicOpState(struct gl_context *ctx)
 
 	R600_STATECHANGE(context, blnd);
 
-	if (RGBA_LOGICOP_ENABLED(ctx))
+	if (_mesa_rgba_logicop_enabled(ctx))
 		SETfield(r700->CB_COLOR_CONTROL.u32All,
 			 translate_logicop(ctx->Color.LogicOp), ROP3_shift, ROP3_mask);
 	else
@@ -657,7 +658,7 @@ static void r700SetLogicOpState(struct gl_context *ctx)
  */
 static void r700LogicOpcode(struct gl_context *ctx, GLenum logicop)
 {
-	if (RGBA_LOGICOP_ENABLED(ctx))
+	if (_mesa_rgba_logicop_enabled(ctx))
 		r700SetLogicOpState(ctx);
 }
 
diff --git a/src/mesa/drivers/dri/radeon/radeon_state.c b/src/mesa/drivers/dri/radeon/radeon_state.c
index e88e984354f..a93e61870a5 100644
--- a/src/mesa/drivers/dri/radeon/radeon_state.c
+++ b/src/mesa/drivers/dri/radeon/radeon_state.c
@@ -40,6 +40,7 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "main/context.h"
 #include "main/framebuffer.h"
 #include "main/simple_list.h"
+#include "main/state.h"
 
 #include "vbo/vbo.h"
 #include "tnl/tnl.h"
@@ -661,7 +662,7 @@ static void radeonUpdateSpecular( struct gl_context *ctx )
 
    TCL_FALLBACK( ctx, RADEON_TCL_FALLBACK_FOGCOORDSPEC, flag);
 
-   if (NEED_SECONDARY_COLOR(ctx)) {
+   if (_mesa_need_secondary_color(ctx)) {
       assert( (p & RADEON_SPECULAR_ENABLE) != 0 );
    } else {
       assert( (p & RADEON_SPECULAR_ENABLE) == 0 );
diff --git a/src/mesa/drivers/dri/savage/savagestate.c b/src/mesa/drivers/dri/savage/savagestate.c
index 1feffa0b0a2..21ebf5dc2b2 100644
--- a/src/mesa/drivers/dri/savage/savagestate.c
+++ b/src/mesa/drivers/dri/savage/savagestate.c
@@ -29,8 +29,9 @@
 #include "main/enums.h"
 #include "main/macros.h"
 #include "main/dd.h"
-
 #include "main/mm.h"
+#include "main/state.h"
+
 #include "savagedd.h"
 #include "savagecontext.h"
 
@@ -869,7 +870,7 @@ static void savageUpdateSpecular_s4(struct gl_context *ctx) {
     savageContextPtr imesa = SAVAGE_CONTEXT( ctx );
     uint32_t drawLocalCtrl = imesa->regs.s4.drawLocalCtrl.ui;
 
-    if (NEED_SECONDARY_COLOR(ctx)) {
+    if (_mesa_need_secondary_color(ctx)) {
 	imesa->regs.s4.drawLocalCtrl.ni.specShadeEn = GL_TRUE;
     } else {
 	imesa->regs.s4.drawLocalCtrl.ni.specShadeEn = GL_FALSE;
@@ -883,7 +884,7 @@ static void savageUpdateSpecular_s3d(struct gl_context *ctx) {
     savageContextPtr imesa = SAVAGE_CONTEXT( ctx );
     uint32_t drawCtrl = imesa->regs.s3d.drawCtrl.ui;
 
-    if (NEED_SECONDARY_COLOR(ctx)) {
+    if (_mesa_need_secondary_color(ctx)) {
 	imesa->regs.s3d.drawCtrl.ni.specShadeEn = GL_TRUE;
     } else {
 	imesa->regs.s3d.drawCtrl.ni.specShadeEn = GL_FALSE;
diff --git a/src/mesa/drivers/dri/sis/sis6326_state.c b/src/mesa/drivers/dri/sis/sis6326_state.c
index 9708f639124..9ca58293b43 100644
--- a/src/mesa/drivers/dri/sis/sis6326_state.c
+++ b/src/mesa/drivers/dri/sis/sis6326_state.c
@@ -35,6 +35,7 @@
 
 #include "main/context.h"
 #include "main/colormac.h"
+#include "main/state.h"
 #include "swrast/swrast.h"
 #include "vbo/vbo.h"
 #include "tnl/tnl.h"
@@ -355,7 +356,7 @@ static void sis6326UpdateSpecular(struct gl_context *ctx)
    sisContextPtr smesa = SIS_CONTEXT(ctx);
    __GLSiSHardware *current = &smesa->current;
 
-   if (NEED_SECONDARY_COLOR(ctx))
+   if (_mesa_need_secondary_color(ctx))
       current->hwCapEnable |= S_ENABLE_Specular;
    else
       current->hwCapEnable &= ~S_ENABLE_Specular;
diff --git a/src/mesa/drivers/dri/sis/sis_state.c b/src/mesa/drivers/dri/sis/sis_state.c
index e53c326441b..828772ed6ef 100644
--- a/src/mesa/drivers/dri/sis/sis_state.c
+++ b/src/mesa/drivers/dri/sis/sis_state.c
@@ -38,6 +38,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "main/context.h"
 #include "main/macros.h"
+#include "main/state.h"
 #include "swrast/swrast.h"
 #include "vbo/vbo.h"
 #include "tnl/tnl.h"
@@ -407,7 +408,7 @@ static void sisUpdateSpecular(struct gl_context *ctx)
    sisContextPtr smesa = SIS_CONTEXT(ctx);
    __GLSiSHardware *current = &smesa->current;
 
-   if (NEED_SECONDARY_COLOR(ctx))
+   if (_mesa_need_secondary_color(ctx))
       current->hwCapEnable |= MASK_SpecularEnable;
    else
       current->hwCapEnable &= ~MASK_SpecularEnable;